Identification of neutralizing epitopes of toxin A and toxin B for the treatment of C. difficile disease

Abstract

The present invention includes methods for generating neutralizing antitoxin directed against clostridial toxins. In particular, the antitoxin directed against these toxins is produced in avian species using soluble recombinant clostridial toxin proteins. This avian antitoxin is designed so as to be orally administrable in therapeutic amounts and may be in any form (i.e., as a solid or in aqueous solution). These antitoxins are useful in the treatment of humans and other animals intoxicated with at least one bacterial toxin.

Description

FIELD OF THE INVENTION

[0002] The present invention relates to clostridial antitoxin therapy for humans and other animals. Antitoxins which neutralize the pathologic effects of clostridial toxins are provided.

BACKGROUND OF THE INVENTION

[0003] The genus Clostridium is comprised of gram-positive, anaerobic, spore-forming bacilli. The natural habitat of these organisms is the environment and the intestinal tracts of humans and other animals. Indeed, clostridia are ubiquitous; they are commonly found in soil, dust, sewage, marine sediments, decaying vegetation, and mud. [See e.g., P. H. A. Sneath et al., “Clostridium,” Bergey's Manual® of Systematic Bacteriology, Vol. 2, pp. 1141-1200, Williams & Wilkins (1986).] Despite the identification of approximately 100 species of Clostridium, only a small number have been recognized as relatively common etiologic agents of medical and veterinary importance. Nonetheless, some of these species are associated with very serious diseases, including botulism, tetanus, anaerobic cellulitis, gas gangrene, bacteremia, pseudomembranous colitis, and clostridial gastroenteritis. Table I lists some of the species of medical and veterinary importance and the diseases with which they are associated. As virtually all of these species have been isolated from fecal samples of apparently healthy persons, some of these isolates may be transient, rather than permanent residents of the colonic flora. Nonetheless, as indicated in Table 1, the

TABLE 1
Clostridium Species of Medical and Veterinary Importance*
Species                  Disease
C. aminovalericum        Bacteriuria (pregnant women)
C. argentinense          Infected wounds; Bacteremia; Botulism; Infections of
                         amniotic fluid
C. baratii               Infected war wounds; Peritonitis; Infectious processes of
                         the eye, ear and prostate
C. beijerinckikii        Infected wounds
C. bifermentans          Infected wounds; Abscesses; Gas Gangrene; Bacteremia
C. botulinum             Food poisoning; Botulism (wound, food, infant)
C. butyricum             Urinary tract, lower respiratory tract, pleural cavity, and
                         abdominal infections; Infected wounds; Abscesses;
                         Bacteremia
C. cadaveris             Abscesses; Infected wounds
C. carnis                Soft tissue infections; Bacteremia
C. chauvoei              Blackleg
C. clostridioforme       Abdominal, cervical, scrotal, pleural, and other
                         infections; Septicemia; Peritonitis; Appendicitis
C. cochlearium           Isolated from human disease processes, but role in
                         disease unknown.
C. difficile             Antimicrobial-associated diarrhea; Pseudomembranous
                         enterocolitis; Bacteremia; Pyogenic infections
C. fallax                Soft tissue infections
C. ghnoii                Soft tissue infections
C. glycolicum            Wound infections; Abscesses; Peritonitis
C. hastiforme            Infected war wounds; Bacteremia; Abscesses
C. histolyticum          Infected war wounds; Gas gangrene; Gingival plaque
                         isolate
C. indolis               Gastrointestinal tract infections
C. innocuum              Gastrointestinal tract infections; Empyema
C. irregulare            Penile lesions
C. leptum                Isolated from human disease processes, but role in
                         disease unknown.
C. limosum               Bacteremia; Peritonitis; Pulmonary infections
C. malenominatum         Various infectious processes
C. novyi                 Infected wounds; Gas gangrene; Blackleg, Big head
                         (ovine); Redwater disease (bovine)
C. oroticum              Urinary tract infections; Rectal abscesses
C. paraputrificum        Bacteremia; Peritonitis; Infected wounds; Appendicitis
C. perfringens           Gas gangrene; Anaerobic cellulitis; Intra-abdominal
                         abscesses; Soft tissue infections; Food poisoning;
                         Necrotizing pneumonia; Empyema; Meningitis;
                         Bacteremia; Uterine Infections; Enteritis necrotans;
                         Lamb dysentery; Struck; Ovine Enterotoxemia;
C. putrefaciens          Bacteriuria (Pregnant women wtih bacteremia)
C. putrificum            Abscesses; Infected wounds; Bacteremia
C. ramosum               Infections of the abdominal cavity, genital tract, lung,
                         and biliary tract; Bacteremia
C. sartagoforme          Isolated from human disease processes, but role in
                         disease unknown.
C. septicum              Gas gangrene; Bacteremia; Suppurative infections;
                         Necrotizing enterocolitis; Braxy
C. sordellii             Gas gangrene; Wound infections; Penile lesions;
                         Bacteremia; Abscesses; Abdominal and vaginal
                         infections
C. sphenoides            Appendicitis; Bacteremia; Bone and soft tissue
                         infections; Intraperitoneal infections; Infected war
                         wounds; Visceral gas gangrene; Renal abscesses
C. sporogenes            Gas gangrene; Bacteremia; Endocarditis; central nervous system
                         and pleuropulmonary infections; Penile lesions;
                         Infected war wounds; Other pyogenic infections
C. subterminale          Bacteremia; Empyema; Biliary tract, soft tissue and
                         bone infections
C. symbiosum             Liver abscesses; Bacteremia; Infections resulting due to
                         bowel flora
C. tertium               Gas gangrene; Appendicitis; Brain abscesses; Intestinal
                         tract and soft tissue infections; Infected war wounds;
                         Periodontitis; Bacteremia
C. tetani                Tetanus; Infected gums and teeth; Corneal ulcerations;
                         Mastoid and middle ear infections; Intraperitoneal
                         infections; Tetanus neonatorum; Postpartum uterine
                         infections; Soft tissue infections, especially related to
                         trauma (including abrasions and lacerations); Infections
                         related to use of contaminated needles
C. thermosaccharolyticum Isolated from human disease processes, but role in
                         disease unknown.
*Compiled from P. G. Engelkirk et al. “Classification”, Principles and Practice of Clinical Anaerobic Bacteriology, pp. 22-23, Star Publishing Co., Belmont, CA (1992); J. Stephen and R. A. Petrowski, “Toxins Which Traverse Membranes and Deregulate Cells,” in Bacterial Toxins, 2d ed., pp. 66-67, American Society for Microbiology (1986); R. Berkow and A. J. Fletcher (eds.),
#“Bacterial Diseases,” Merck Manual of Diagnosis and Therapy, 16th ed., pp. 116-126, Merck Research
#Laboratories, Rahway, N.J. (1992); and O. H. Sigmund and C. M. Fraser (eds.), “Clostridial Infections,” Merck Veterinary Manual, 5th ed., pp. 396-409, Merck & Co., Rahway, N.J. (1979).

[0004] majority of these organisms may be associated with serious and/or debilitating disease.

[0005] In most cases, the pathogenicity of these organisms is related to the release of powerful exotoxins or highly destructive enzymes. Indeed, several species of the genus Clostridium produce toxins and other enzymes of great medical and veterinary significance. [C. L. Hatheway, Clin. Microbiol. Rev. 3:66-98 (1990).]

[0006] Perhaps because of their significance for human and veterinary medicine, much research has been conducted on these toxins, in particular those of C. botulinum and C. difficile.

[0007] C. botulinum

[0008] Several strains of Clostridium botulinum produce toxins of significance to human and animal health. [C. L. Hatheway, Clin. Microbiol. Rev. 3:66-98 (1990).] The effects of these toxins range from diarrheal diseases that can cause destruction of the colon, to paralytic effects that can cause death. Particularly at risk for developing clostridial diseases are neonates and humans and animals in poor health (e.g., those suffering from diseases associated with old age or immunodeficiency diseases).

[0009] Clostridium botulinum produces the most poisonous biological toxin known. The lethal human dose is a mere 10−9 mg/kg bodyweight for toxin in the bloodstream. Botulinal toxin blocks nerve transmission to the muscles, resulting in flaccid paralysis. When the toxin reaches airway and respiratory muscles, it results in respiratory failure that can cause death. [S. Arnon, J. Infect. Dis. 154:201-206 (1986).]

[0010] C. botulinum spores are carried by dust and are found on vegetables taken from the soil, on fresh fruits, and on agricultural products such as honey. Under conditions favorable to the organism, the spores germinate to vegetative cells which produces toxin. [S. Arnon, Ann. Rev. Med. 31:541 (1980).]

[0011] Botulism disease may be grouped into three types, based on the method of introduction of toxin into the bloodstream. Food-borne botulism results from ingesting improperly preserved and inadequately heated food that contains botulinal toxin. There were 355 cases of food-borne botulism in the United States between 1976 and 1984. [K. L. MacDonald et al., Am. J. Epidemiol. 124:794 (1986).] The death rate due to botulinal toxin is 12% and can be higher in particular risk groups. [C. O. Tacket et al., Am. J. Med. 76:794 (1984).] Wound-induced botulism results from C. botulinum penetrating traumatized tissue and producing toxin that is absorbed into the bloodstream. Since 1950, thirty cases of wound botulism have been reported. [M. N. Swartz, “Anaerobic Spore-Forming Bacilli: The Clostridia, ” pp. 633-646, in B. D. Davis et al., (eds.), Microbiology, 4th edition, J.B. Lippincott Co. (1990).] Infectious infant botulism results from C. botulinum colonization of the infant intestine with production of toxin and its absorption into the bloodstream. It is likely that the bacterium gains entry when spores are ingested and subsequently germinate. [S. Arnon, J. Infect. Dis. 154:201 (1986).] There have been 500 cases reported since it was first recognized in 1976. [M. N. Swartz, supra.]

[0012] Infant botulism strikes infants who are three weeks to eleven months old (greater than 90% of the cases are infants less than six months). [S. Arnon, J. Infect. Dis. 154:201 (1986).] It is believed that infants are susceptible, due, in large part, to the absence of the full adult complement of intestinal microflora. The benign microflora present in the adult intestine provide an acidic environment that is not favorable to colonization by C. botulinum. Infants begin life with a sterile intestine which is gradually colonized by microflora. Because of the limited microflora present in early infancy, the intestinal environment is not as acidic, allowing for C. botulinum spore germination, growth, and toxin production. In this regard, some adults who have undergone antibiotic therapy which alters intestinal microflora become more susceptible to botulism.

[0013] An additional factor accounting for infant susceptibility to infectious botulism is the immaturity of the infant immune system. The mature immune system is sensitized to bacterial antigens and produces protective antibodies. Secretory IgA produced in the adult intestine has the ability to agglutinate vegetative cells of C. botulinum. [S. Arnon, J. Infect. Dis. 154:201 (1986).] Secretory IgA may also act by preventing intestinal bacteria and their products from crossing the cells of the intestine. [S. Arnon, Epidemiol. Rev. 3:45 (1981).] The infant immune system is not primed to do this.

[0014] Clinical symptoms of infant botulism range from mild paralysis, to moderate and severe paralysis requiring hospitalization, to fulminant paralysis, leading to sudden death. [S. Arnon, Epidemiol. Rev. 3:45 (1981).]

[0015] The chief therapy for severe infant botulism is ventilatory assistance using a mechanical respirator and concurrent elimination of toxin and bacteria using cathartics, enemas, and gastric lavage. There were 68 hospitalizations in Califormia for infant botulism in a single year with a total cost of over $4 million for treatment. [T. L. Frankovich and S. Arnon, West. J. Med. 154:103 (1991).]

[0016] Different strains of Clostridium botulinum each produce antigenically distinct toxin designated by the letters A-G. Nearly all cases of infant botulism have been caused by bacteria producing either type A or type B toxin. (Exceptionally, one New Mexico case was caused by Clostridium botulinum producing type F toxin and another by Clostridium botulinum producing a type B-type F hybrid.) [S. Arnon, Epidemiol. Rev. 3:45 (1981).] Type C toxin affects waterfowl, type D toxin affects cattle, and type E toxin affects both humans and birds.

[0017] A trivalent antitoxin derived from horse plasma is commercially available from Connaught Industries Ltd. as a therapy for toxin types A, B, and E. However, the antitoxin has several disadvantages. First, extremely large dosages must be injected intravenously and/or intramuscularly. Second, the antitoxin has serious side effects such as acute anaphylaxis which can lead to death, and serum sickness. Finally, the efficacy of the antitoxin is uncertain and the treatment is costly. [T. O. Tacket et al., Am. J. Med. 76:794 (1984).]

[0018] A heptavalent equine botulinal antitoxin which uses only the F(ab′)2 portion of the antibody molecule has been tested by the United States Military. [M. Balady, USAMRDC Newsletter, p. 6 (1991).] This was raised against impure toxoids in those large animals and is not a high titer preparation.

[0019] A pentavalent human antitoxin has been collected from immunized human subjects for use as a treatment for infant botulism. The supply of this antitoxin is limited and cannot be expected to meet the needs of all individuals stricken with botulism disease. In addition, collection of human sera must involve screening out HIV and other potentially serious human pathogens. [P. J. Schwarz and S. S. Arnon, Western J. Med. 156:197 (1992).]

[0020] Infant botulism has been implicated as the cause of mortality in some cases of Sudden Infant Death Syndrome (SIDS, also known as crib death). SIDS is officially recognized as infant death that is sudden and unexpected and that remained unexplained despite complete post-mortem examination. The link of SIDS to infant botulism came when fecal or blood specimens taken at autopsy from SIDS infants were found to contain C. botulinum organisms and/or toxin in 3-4% of cases analyzed. [D. R. Peterson et al., Rev. Infect. Dis. 1:630 (1979).] In contrast, only 1 of 160 healthy infants (0.6%) had C. botulinum organisms in the feces and no botulinal toxin. (S. Arnon et al., Lancet, pp. 1273-76, June 17, 1978.)

[0021] In developed countries, SIDS is the number one cause of death in children between one month and one year old. (S. Arnon et al., Lancet, pp. 1273-77, June 17, 1978.) More children die from SIDS in the first year than from any other single cause of death in the first fourteen years of life. In the United States, there are 8,000-10,000 SIDS victims annually. Id.

[0022] What is needed is an effective therapy against infant botulism that is free of dangerous side effects, is available in large supply at a reasonable price, and can be safely and gently delivered so that prophylactic application to infants is feasible.

[0023] C. difficile

[0024] C. difficile, an organism which gained its name due to difficulties encountered in its isolation, has recently been proven to be an etiologic agent of diarrheal disease. (Sneath et al., p. 1165.). C. difficile is present in the gastrointestinal tract of approximately 3% of healthy adults, and 10-30% of neonates without adverse effect (Swartz, at p. 644); by other estimates, C. difficile is a part of the normal gastrointestinal flora of 2-10% of humans. [G. F. Brooks et al., (eds.) “Infections Caused by Anaerobic Bacteria,” Jawetz, Melnick, & Adelberg's Medical Microbiology, 19th ed., pp. 257-262, Appleton & Lange, San Mateo, Calif. (1991).] As these organisms are relatively resistant to most commonly used antimicrobials, when a patient is treated with antibiotics, the other members of the normal gastrointestinal flora are suppressed and C. difficile flourishes, producing cytopathic toxins and enterotoxins. It has been found in 25% of cases of moderate diarrhea resulting from treatment with antibiotics, especially the cephalosporins, clindamycin, and ampicillin. [M. N. Swartz at 644.]

[0025] Importantly, C. difficile is commonly associated with nosocomial infections. The organism is often present in the hospital and nursing home environments and may be carried on the hands and clothing of hospital personnel who care for debilitated and immunocompromised patients. As many of these patients are being treated with antimicrobials or other chemotherapeutic agents, such transmission of C. difficile represents a significant risk factor for disease. (Engelkirk et al., pp. 64-67.)

[0026] C. difficile is associated with a range of diarrhetic illness, ranging from diarrhea alone to marked diarrhea and necrosis of the gastrointestinal mucosa with the accumulation of inflammatory cells and fibrin, which forms a pseudomembrane in the affected area. (Brooks et al.) It has been found in over 95% of pseudomembranous enterocolitis cases. (Swartz, at p. 644.) This occasionally fatal disease is characterized by diarrhea, multiple small colonic plaques, and toxic megacolon. (Swartz, at p. 644.) Although stool cultures are sometimes used for diagnosis, diagnosis is best made by detection of the heat labile toxins present in fecal filtrates from patients with enterocolitis due to C. difficile. (Swartz, at p. 644-645; and Brooks et al., at p. 260.) C. difficile toxins are cytotoxic for tissue/cell cultures and cause enterocolitis when injected intracecally into hamsters. (Swartz, at p. 644.)

[0027] The enterotoxicity of C. difficile is primarily due to the action of two toxins, designated A and B, each of approximately 300,000 in molecular weight. Both are potent cytotoxins, with toxin A possessing direct enterocytotoxic activity. [Lyerly et al., Infect. Immun. 60:4633 (1992).] Unlike toxin A of C. perfringens, an organism rarely associated with antimicrobial-associated diarrhea, the toxin of C. difficile is not a spore coat constituent and is not produced during sporulation. (Swartz, at p. 644.) C. difficile toxin A causes hemorrhage, fluid accumulation and mucosal damage in rabbit ileal loops and appears to increase the uptake of toxin B by the intestinal mucosa. Toxin B does not cause intestinal fluid accumulation, but it is 1000 times more toxic than toxin A to tissue culture cells and causes membrane damage. Although both toxins induce similar cellular effects such as actin disaggregation, differences in cell specificity occurs.

[0028] Both toxins are important in disease. [Borriello et al., Rev. Infect. Dis., 12(suppl. 2):S185 (1990); Lyerly et al., Infect. Immun., 47:349 (1985); and Rolfe, Infect. Immun., 59:1223 (1990).] Toxin A is thought to act first by binding to brush border receptors, destroying the outer mucosal layer, then allowing toxin B to gain access to the underlying tissue. These steps in pathogenesis would indicate that the production of neutralizing antibodies against toxin A may be sufficient in the prophylactic therapy of CDAD. However, antibodies against toxin B may be a necessary additional component for an effective therapeutic against later stage colonic disease. Indeed, it has been reported that animals require antibodies to both toxin A and toxin B to be completely protected against the disease. [Kim and Rolfe, Abstr. Ann. Meet. Am. Soc. Microbiol., 69:62 (1987).]

[0029] C. difficile has also been reported to produce other toxins such as an enterotoxin different from toxins A and B [Banno et al., Rev. Infect. Dis., 6(Suppl. 1:S 11-S20 (1984)], a low molecular weight toxin [Rihn et al., Biochem. Biophys. Res. Comm., 124:690-695 (1984)], a motility altering factor [Justus et al., Gastroenterol., 83:836-843 (1982)], and perhaps other toxins. Regardless, C. difficile gastrointestinal disease is of primary concern.

[0030] It is significant that due to its resistance to most commonly used antimicrobials, C. difficile is associated with antimicrobial therapy with virtually all antimicrobial agents (although most commonly ampicillin, clindamycin and cephalosporins). It is also associated with disease in patients undergoing chemotherapy with such compounds as methotrexate, 5-fluorouracil, cyclophosphamide, and doxorubicin. [S. M. Finegold et al., Clinical Guide to Anaerobic Infections, pp. 88-89, Star Publishing Co., Belmont, Calif. (1992).]

[0031] Treatment of C. difficile disease is problematic, given the high resistance of the organism. Oral metronidazole, bacitracin and vancomycin have been reported to be effective. (Finegold et al., p. 89.) However there are problems associated with treatment utilizing these compounds. Vancomycin is very expensive, some patients are unable to take oral medication, and the relapse rate is high (20-25%), although it may not occur for several weeks. Id.

[0032] C. difficile disease would be prevented or treated by neutralizing the effects of these toxins in the gastrointestinal tract. Thus, what is needed is an effective therapy against C. difficile toxin that is free of dangerous side effects, is available in large supply at a reasonable price, and can be safely delivered so that prophylactic application to patients at risk of developing pseudomembranous enterocolitis can be effectively treated.

DESCRIPTION OF THE DRAWINGS

[0033] FIG. 1 shows the reactivity of anti-C. botulinum IgY by Western blot.

[0034] FIG. 2 shows the IgY antibody titer to C. botulinum type A toxoid in eggs, measured by ELISA.

[0035] FIG. 3 shows the results of C. difficile toxin A neutralization assays.

[0036] FIG. 4 shows the results of C. difficile toxin B neutralization assays.

[0037] FIG. 5 shows the results of C. difficile toxin B neutralization assays.

[0038] FIG. 6 is a restriction map of C. difficile toxin A gene, showing sequences of primers 1-4 (SEQ ID NOS:1-4).

[0039] FIG. 7 is a Western blot of C. difficile toxin A reactive protein.

[0040] FIG. 8 shows C. difficile toxin A expression constructs.

[0041] FIG. 9 shows C. difficile toxin A expression constructs.

[0042] FIG. 10 shows the purification of recombinant C. difficile toxin A.

[0043] FIG. 11 shows the results of C. difficile toxin A neutralization assays with antibodies reactive to recombinant toxin A.

[0044] FIG. 12 shows the results for a C. difficile toxin A neutralization plate.

[0045] FIG. 13 shows the results for a C. difficile toxin A neutralization plate.

[0046] FIG. 14 shows the results of recombinant C. difficile toxin A neutralization assays.

[0047] FIG. 15 shows C. difficile toxin A expression constructs.

[0048] FIG. 16 shows a chromatograph plotting absorbance at 280 nm against retention time for a pMA1870-680 IgY PEG preparation.

[0049] FIG. 17 shows two recombinant C. difficile toxin B expression constructs.

[0050] FIG. 18 shows C. difficile toxin B expression constructs.

[0051] FIG. 19 shows C. difficile toxin B expression constructs.

[0052] FIG. 20 shows C. difficile toxin B expression constructs.

[0053] FIG. 21 is an SDS-PAGE gel showing the purification of recombinant C difficile toxin B fusion protein.

[0054] FIG. 22 is an SDS-PAGE gel showing the purification of two histidine-tagged recombinant C. difficile toxin B proteins.

[0055] FIG. 23 shows C. difficile toxin B expression constructs.

[0056] FIG. 24 is a Western blot of C. difficile toxin B reactive protein.

DEFINITIONS

[0057] To facilitate understanding of the invention, a number of terms are defined below.

[0058] As used herein, the term “neutralizing” is used in reference to antitoxins, particularly antitoxins comprising antibodies, which have the ability to prevent the pathological actions of the toxin against which the antitoxin is directed.

[0059] As used herein, the term “overproducing” is used in reference to the production of clostridial toxin polypeptides in a host cell and indicates that the host cell is producing more of the clostridial toxin by virture of the introduction of nucleic acid sequences encoding said clostridial toxin polypeptide than would be expressed by said host cell absent the introduction of said nucleic acid sequences. To allow ease of purification of toxin polypeptides produced in a host cell it is preferred that the host cell express or overproduce said toxin polypeptide at a level greater than 1 mg/liter of host cell culture.

[0060] As used herein, the term “fusion protein” refers to a chimeric protein containing the protein of interest (i.e., C. difficile toxin A or B and fragments thereof) joined to an exogenous protein fragment (the fusion partner which consists of a non-toxin protein). The fusion partner may enhance solubility of the C. difficile protein as expressed in a host cell, may provide an affinity tag to allow purification of the recombinant fusion protein from the host cell or culture supernatant, or both. If desired, the fusion protein may be removed from the protein of interest (i.e., toxin protein or fragments thereof) prior to immunization by a variety of enzymatic or chemical means known to the art.

[0061] The term “protein of interest” as used herein refers to the protein whose expression is desired within the fusion protein. In a fusion protein the protein of interest will be joined or fused with another protein or protein domain, the fusion partner, to allow for enhanced stability of the protein of interest and/or ease of purification of the fusion protein.

[0062] As used herein, the term “maltose binding protein” referes to the maltose binding protein of E. coli. A portion of the maltose binding protein may be added to a protein of interest to generate a fusion protein; a portion of the maltose binding protein may merely enhance the solubility of the resulting fusion protein when expressed in a bacterial host. On the other hand, a portion of the maltose binding protein may allow affinity purification of the fusion protein on an amylose resin.

[0063] As used herein, the term “poly-histidine tract” when used in reference to a fusion protein refers to the presence of two to ten histidine residues at either the amino- or carboxy-terminus of a protein of interest. A poly-histidine tract of six to ten residues is preferred. The poly-histidine tract is also defined functionally as being a number of consequtive histidine residues added to the protein of interest which allows the affinity purification of the resulting fusion protein on a nickle-chelate column.

[0064] As used herein, the term “purified” or “to purify” refers to the removal of contaminants from a sample. For example, antitoxins are purified by removal of contaminating non-immunoglobulin proteins; they are also purified by the removal of immunoglobulin that does not bind toxin. The removal of non-immunoglobulin proteins and/or the removal of immunoglobulins that do not bind toxin results in an increase in the percent of toxin-reactive immunoglobulins in the sample. In another example, recombinant toxin polypeptides are expressed in bacterial host cells and the toxin polypeptides are purified by the removal of host cell proteins; the percent of recombinant toxin polypeptides is thereby increased in the sample.

[0065] The term “recombinant DNA molecule” as used herein refers to a DNA molecule which is comprised of segments of DNA joined together by means of molecular biological techniques.

[0066] The term “recombinant protein” or “recombinant polypeptide” as used herein refers to a protein molecule which is expressed from a recombinant DNA molecule.

[0067] The term “native protein” as used herein refers to a protein which is isolated from a natural source as opposed to the production of a protein by recombinant means.

[0068] As used herein the term “portion” when in reference to a protein (as in “a portion of a given protein”) refers to fragments of that protein. The fragments may range in size from four amino acid residues to the entire amino acid sequence minus one amino acid.

[0069] As used herein “soluble” when in reference to a protein produced by recombinant DNA technology in a host cell is a protein which exists in solution in the cytoplasm of the host cell; if the protein contains a signal sequence the soluble protein is exported to the periplasmic space in bacteria hosts and is secreted into the culture medium eucaryotic cells capable of secretion or by bacterial host possessing the appropriate genes (i.e., the kil gene). In contrast, an insoluble protein is one which exists in denatured form inside cytoplasmic granules (called an inclusion bodies) in the host cell. High level expression (i.e., greater than 1 mg recombinant protein/liter of bacterial culture) of recombinant proteins often results in the expressed protein being found in inclusion bodies in the bacterial host cells. A soluble protein is a protein which is not found in an inclusion body inside the host cell or is found both in the cytoplasm and in inclusion bodies and in this case the protein may be present at high or low levels in the cytoplasm.

[0070] A distinction is drawn between a soluble protein (i.e., a protein which when expressed in a host cell is produced in a soluble form) and a “solubilized” protein. An insoluble recombinant protein found inside an inclusion body may be solubilized (i.e., rendered into a soluble form) by treating purified inclusion bodies with denaturants such as guanidine hydrochloride, urea or sodium dodecyl sulfate (SDS). These denaturants must then be removed from the solubilized protein preparation to allow the recovered protein to renature (refold). Not all proteins will refold into an active conformation after solubilization in a denaturant and removal of the denaturant. Many proteins precipitate upon removal of the denaturant. SDS may be used to solubilize inclusion bodies and will maintain the proteins in solution at low concentration. However, dialysis will not always remove all of the SDS (SDS can form micells which do not dialyze out); therefor, SDS-solubilized inclusion body protein is soluble but not refolded.

[0071] As used herein, the term “therapeutic amount” refers to that amount of antitoxin required to neutralize the pathologic effects of Clostridium difficile toxin in a subject.

SUMMARY OF THE INVENTION

[0072] The present invention relates to clostridial antitoxin therapy for humans and other animals. Antitoxins which neutralize the pathologic effects of clostridial toxins are generated by immunization of avian hosts with recombinant toxin fragments. In one embodiment, the present invention contemplates a fusion protein comprising a non-toxin protein sequence and a portion of the Clostridium difficile toxin B sequence of SEQ ID NO:10. It is not intended that the present invention be limited by the nature of the fusion protein. For example, the fusion protein may comprise the portion of the Clostridium difficile toxin B sequence as set forth in SEQ ID NO:20 and the maltose binding protein (or portion thereof). On the other hand, the fusion protein may comprise the Clostridium difficile toxin B sequence as set forth in SEQ ID NO:21 along with a poly-histidine tract.

[0073] In one embodiment, the present invention contemplates a method of generating a neutralizing antitoxin directed against Clostridium difficile toxin B comprising: a) providing in any order: i) a purified fusion protein comprising a non-toxin protein sequence and a portion of the Clostridium difficile toxin B sequence of SEQ ID NO:10, and ii) an avian host; and b) immunizing said host with said purified fusion protein so as to generate an antitoxin capable of neutralizing native Clostridium difficile toxin B. Again by way of illustration only, the fusion protein may comprise the portion of the Clostridium difficile toxin B sequence as set forth in SEQ ID NO:20 and the maltose binding protein (or portion thereof). On the other hand, the fusion protein may comprise the Clostridium difficile toxin B sequence as set forth in SEQ ID NO:21 along with a poly-histidine tract. The method may further comprise a step c) collecting said antitoxin from said host and, even further, a step d) purifying said antitoxin. The present invention contemplates the antibody, as a composition of matter, raised according to the above-described methods.

[0074] The present invention further contemplates a method of treatment comprising: a) providing: i) a subject, and ii) at least one neutralizing antitoxin directed against a fusion protein comprising a non-toxin protein sequence and a portion of Clostridium difficile toxin B sequence of SEQ ID NO:10; and b) administering said antitoxin to said subject. In one embodiment, the present invention contemplates administering by oral administration. The present invention further contemplates that the subject treated may or may not have been exposed to Clostridium difficile and its toxins. That is to say, in one instance, exposure to at Clostridium difficile toxin B may have occurred prior to administration of the antitoxin. In another instance, the subject has not been exposed to at Clostridium difficile toxin B prior to administration of the antitoxin.

[0075] The present invention also contemplates fusion proteins containing toxin A fragments. In one embodiment, the present invention contemplates a fusion protein comprising a non-toxin protein sequence and a portion of the Clostridium difficile toxin A sequence consisting of SEQ ID NO:7. In still another embodiment, the present invention contemplates a fusion protein comprising a non-toxin protein sequence and a portion of the Clostridium difficile toxin A sequence comprising the amino acid sequence of SEQ ID NO:8. In the above embodiments, the non-toxin part of the fusion protein may be selected from a variety of non-toxin protein sequence types. In a preferred embodiment, the non-toxin sequence is the maltose binding protein sequence (or a portion thereof).

[0076] The present invention contemplates a method of generating a neutralizing antitoxin directed against Clostridium difficile toxin A comprising: a) providing in any order: i) a purified fusion protein comprising a non-toxin protein sequence (for example, the maltose binding protein sequence or portion thereof) and a portion of the Clostridium difficile toxin A sequence (for example, the toxin A sequence as set forth in SEQ ID NO:7), and ii) an avian host; and b) immunizing said host with said purified fusion protein so as to generate an antitoxin capable of neutralizing said Clostridium difficile toxin A. The method may further comprise step c) collecting said antitoxin from said host and step d) purifying said antitoxin.

[0077] The present invention also contemplates uses for the toxin fragments in vaccines and diagnostic assays. The fragments may be used separately as purified, soluble antigens or, alternatively, in mixtures or “cocktails.”

DESCRIPTION OF THE INVENTION

[0078] The present invention contemplates treating humans and other animals intoxicated with at least one bacterial toxin. It is contemplated that oral administration of antitoxin will be used to treat patients effected by or at risk of symptoms due to the action of bacterial toxins. The organisms, toxins and individual steps of the present invention are described separately below.

[0079] I. Clostridium Species, Clostridial Diseases and Associated Toxins

[0080] A preferred embodiment of the method of the present invention is directed toward obtaining antibodies against Clostridium species, their toxins, enzymes or other metabolic by-products, cell wall components, or synthetic or recombinant versions of any of these compounds. It is contemplated that these antibodies will be produced by immunization of humans or other animals. It is not intended that the present invention be limited to any particular toxin or any species of organism. In one embodiment,

TABLE 2
Clostridial Toxins
Organism        Toxins and Disease-Associated Antigens
C. botulinum    A, B, C1, C2, D, E, F, G
C. butyricum    Neuraminidase
C. difficile    A, B, Enterotoxin (not A nor B), Motility Altering
                Factor, Low Molecular Weight Toxin, Others
C. perfringens  α, β, ε, ι, γ, δ, ν, θ, κ, λ, μ, υ
C. sordelli/    HT, LT, α, β, γ
C. bifermentans
C. novyi        α, β, γ, δ, ε, ζ, ν, θ
C. septicum     α, β, γ, δ
C. histolyticum α, β, γ, δ, ε plus additional enzymes
C. chauvoei     α, β, γ, δ

[0081] toxins from all Clostridium species are contemplated as immunogens. Examples of these toxins include the neuraminidase toxin of C. butyricum, C. sordellii toxins HT and LT, toxins A, B, C, D, E, F, and G of C. botulinum and the numerous C. perfringens toxins. In one preferred embodiment, toxins A and B of C. difficile are contemplated as immunogens. Table 2 above lists various Clostridium species, their toxins and some antigens associated with disease.

[0082] It is not intended that antibodies produced against one toxin will only be used against that toxin. It is contemplated that antibodies directed against one toxin (e.g., C. perfringens type A enterotoxin) may be used as an effective therapeutic against one or more toxin(s) produced by other members of the genus Clostridium or other toxin producing organisms (e.g. Bacillus cereus, Staphylococcus aureus, Streptococcus mutans, Acinetobacter calcoaceticus, Pseudomonas aeruginosa, other Pseudomonas species, etc.). It is further contemplated that antibodies directed against the portion of the toxin which binds to mammalian membranes (e.g., C. perfringens enterotoxin A) can also be used against other organisms. It is contemplated that these membrane binding domains are produced synthetically and used as immunogens.

[0083] II. Obtaining Antibodies in Non-Mammals

[0084] A preferred embodiment of the method of the present invention for obtaining antibodies involves immunization. However, it is also contemplated that antibodies could be obtained from non-mammals without immunization. In the case where no immunization is contemplated, the present invention may use non-mammals with preexisting antibodies to toxins as well as non-mammals that have antibodies to whole organisms by virtue of reactions with the administered antigen. An example of the latter involves immunization with synthetic peptides or recombinant proteins sharing epitopes with whole organism components.

[0085] In a preferred embodiment, the method of the present invention contemplates immunizing non-mammals with bacterial toxin(s). It is not intended that the present invention be limited to any particular toxin. In one embodiment, toxin from all clostridial bacteria sources (see Table 2) are contemplated as immunogens. Examples of these toxins are C. butyricum neuraminidase toxin, toxins A, B, C, D, E, F, and G from C. botulinum, C. perfringens toxins α, β, ε, and ι, and C. sordellii toxins HT and LT. In a preferred embodiment, C. difficile toxins A and B are contemplated as immunogens.

[0086] A particularly preferred embodiment involves the use of bacterial toxin protein or fragments of toxin proteins produced by molecular biological means (i.e., recombinant toxin proteins). In a preferred embodiment, the immunogen comprises interval 6 of C. difficile toxin A produced by recombinant DNA technology. In yet another preferred embodiment, the immunogen comprises interval 3 of C. difficile toxin B produced by recombinant DNA technology.

[0087] When immunization is used, the preferred non-mammal is from the class Aves. All birds are contemplated (e.g., duck, ostrich, emu, turkey, etc.). A preferred bird is a chicken. Importantly, chicken antibody does not fix mammalian complement. [See H. N. Benson et al., J. Immunol. 87:616 (1961).] Thus, chicken antibody will normally not cause a complement-dependent reaction. [A. A. Benedict and K. Yamaga, “Immunoglobulins and Antibody Production in Avian Species,” in Comparative Immunology (J. J. Marchaloni, ed.), pp. 335-375, Blackwell, Oxford (1966).] Thus, the preferred antitoxins of the present invention will not exhibit complement-related side effects observed with antitoxins known presently.

[0088] When birds are used, it is contemplated that the antibody will be obtained from either the bird serum or the egg. A preferred embodiment involves collection of the antibody from the egg. Laying hens transport immunoglobulin to the egg yolk (“IgY”) in concentrations equal to or exceeding that found in serum. [See R. Patterson et al., J. Immunol. 89:272 (1962); and S. B. Carroll and B. D. Stollar, J. Biol. Chem. 258:24 (1983).] In addition, the large volume of egg yolk produced vastly exceeds the volume of serum that can be safely obtained from the bird over any given time period. Finally, the antibody from eggs is purer and more homogeneous; there is far less non-immunoglobulin protein (as compared to serum) and only one class of immunoglobulin is transported to the yolk.

[0089] When considering immunization with toxins, one may consider modification of the toxins to reduce the toxicity. In this regard, it is not intended that the present invention be limited by immunization with modified toxin. Unmodified (“native”) toxin is also contemplated as an immunogen.

[0090] It is also not intended that the present invention be limited by the type of modification—if modification is used. The present invention contemplates all types of toxin modification, including chemical and heat treatment of the toxin. The preferred modification, however, is formaldehyde treatment.

[0091] It is not intended that the present invention be limited to a particular mode of immunization; the present invention contemplates all modes of immunization, including subcutaneous, intramuscular, intraperitoneal, and intravenous or intravascular injection, as well as per os administration of immunogen.

[0092] The present invention further contemplates immunization with or without adjuvant. (Adjuvant is defined as a substance known to increase the immune response to other antigens when administered with other antigens.) If adjuvant is used, it is not intended that the present invention be limited to any particular type of adjuvant—or that the same adjuvant, once used, be used all the time. While the present invention contemplates all types of adjuvant, whether used separately or in combinations, the preferred use of adjuvant is the use of Complete Freund's Adjuvant followed sometime later with Incomplete Freund's Adjuvant. Another preferred use of adjuvant is the use of Gerbu Adjuvant. The invention also contemplates the use of RIBI fowl adjuvant and Quil A adjuvant.

[0093] When immunization is used, the present invention contemplates a wide variety of immunization schedules. In one embodiment, a chicken is administered toxin(s) on day zero and subsequently receives toxin(s) in intervals thereafter. It is not intended that the present invention be limited by the particular intervals or doses. Similarly, it is not intended that the present invention be limited to any particular schedule for collecting antibody. The preferred collection time is sometime after day 100.

[0094] Where birds are used and collection of antibody is performed by collecting eggs, the eggs may be stored prior to processing for antibody. It is preferred that eggs be stored at 4° C. for less than one year.

[0095] It is contemplated that chicken antibody produced in this manner can be buffer-extracted and used analytically. While unpurified, this preparation can serve as a reference for activity of the antibody prior to further manipulations (e.g., immunoaffinity purification).

[0096] III. Increasing the Effectiveness of Antibodies

[0097] When purification is used, the present invention contemplates purifying to increase the effectiveness of both non-mammalian antitoxins and mammalian antitoxins. Specifically, the present invention contemplates increasing the percent of toxin-reactive immunoglobulin. The preferred purification approach for avian antibody is polyethylene glycol (PEG) separation.

[0098] The present invention contemplates that avian antibody be initially purified using simple, inexpensive procedures. In one embodiment, chicken antibody from eggs is purified by extraction and precipitation with PEG. PEG purification exploits the differential solubility of lipids (which are abundant in egg yolks) and yolk proteins in high concentrations of PEG 8000. [Polson et al., Immunol. Comm. 9:495 (1980).] The technique is rapid, simple, and relatively inexpensive and yields an immunoglobulin fraction that is significantly purer in terms of contaminating non-immunoglobulin proteins than the comparable ammonium sulfate fractions of mammalian sera and horse antibodies. The majority of the PEG is removed from the precipitated chicken immunoglobulin by treatment with ethanol. Indeed, PEG-purified antibody is sufficiently pure that the present invention contemplates the use of PEG-purified antitoxins in the passive immunization of intoxicated humans and animals.

[0099] IV. Treatment

[0100] The present invention contemplates antitoxin therapy for humans and other animals intoxicated by bacterial toxins. A preferred method of treatment is by oral administration of antitoxin.

[0101] A. Dosage of Antitoxin

[0102] It was noted by way of background that a balance must be struck when administering currently available antitoxin which is usually produced in large animals such as horses; sufficient antitoxin must be administered to neutralize the toxin, but not so much antitoxin as to increase the risk of untoward side effects. These side effects are caused by: i) patient sensitivity to foreign (e.g., horse) proteins; ii) anaphylactic or immunogenic properties of non-immunoglobulin proteins; iii) the complement fixing properties of mammalian antibodies; and/or iv) the overall burden of foreign protein administered. It is extremely difficult to strike this balance when, as noted above, the degree of intoxication (and hence the level of antitoxin therapy needed) can only be approximated.

[0103] The present invention contemplates significantly reducing side effects so that this balance is more easily achieved. Treatment according to the present invention contemplates reducing side effects by using PEG-purified antitoxin from birds.

[0104] In one embodiment, the treatment of the present invention contemplates the use of PEG-purified antitoxin from birds. The use of yolk-derived, PEG-purified antibody as antitoxin allows for the administration of: 1) non(mammalian)-complement-fixing, avian antibody; 2) a less heterogeneous mixture of non-immunoglobulin proteins; and 3) less total protein to deliver the equivalent weight of active antibody present in currently available antitoxins. The non-mammalian source of the antitoxin makes it useful for treating patients who are sensitive to horse or other mammalian sera.

[0105] B. Delivery of Antitoxin

[0106] Although it is not intended to limit the route of delivery, the present invention contemplates a method for antitoxin treatment of bacterial intoxication in which delivery of antitoxin is oral. In one embodiment, antitoxin is delivered in a solid form (e.g., tablets). In an alternative embodiment antitoxin is delivered in an aqueous solution. When an aqueous solution is used, the solution has sufficient ionic strength to solubilize antibody protein, yet is made palatable for oral administration. The delivery solution may also be buffered (e.g., carbonate buffer pH 9.5) which can neutralize stomach acids and stabilize the antibodies when the antibodies are administered orally. In one embodiment the delivery solution is an aqueous solution. In another embodiment the delivery solution is a nutritional formula. Preferably, the delivery solution is infant formula. Yet another embodiment contemplates the delivery of lyophilized antibody encapsulated or microencapsulated inside acid-resistant compounds.

[0107] Methods of applying enteric coatings to pharmaceutical compounds are well known to the art [companies specializing in the coating of pharmaceutical compounds are available; for example, The Coating Place (Verona, Wis.) and AAI (Wilmington, N.C.)]. Enteric coatings which are resistant to gastric fliud and whose release (i.e., dissolution of the coating to release the pharmaceutical compound) is pH dependent are cominerically available [for example, the polymethacrylates Eudragit® L and Eudragit® S (Röhm GmbH)]. Eudragit® S is soluble in intestinal fluid from pH 7.0; this coating can be used to microencapsulate lyophilized antitoxin antibodies and the particles are suspended in a solution having a pH above or below pH 7.0 for oral administration. The microparticles will remain intact and undissolved until they reached the intestines where the intestinal pH would cause them to dissolve thereby releasing the antitoxin.

[0108] The invention contemplates a method of treatment which can be administered for treatment of acute intoxication. In one embodiment, antitoxin is administered orally in either a delivery solution or in tablet form, in therapeutic dosage, to a subject intoxicated by the bacterial toxin which served as immunogen for the antitoxin.

[0109] The invention also contemplates a method of treatment which can be administered prophylactically. In one embodiment, antitoxin is administered orally, in a delivery solution, in therapeutic dosage, to a subject, to prevent intoxication of the subject by the bacterial toxin which served as immunogen for the production of antitoxin. In another embodiment, antitoxin is administered orally in solid form such as tablets or as microencapsulated particles. Microencapsulation of lyophilized antibody using compounds such as Eudragit® (Rohm GmbH) or polyetheylene glycol, which dissolve at a wide range of pH units, allows the oral administration of solid antitoxin in a liquid form (i.e., a suspension) to recipients unable to tolerate administration of tablets (e.g., children or patients on feeding tubes). In one preferred embodiment the subject is an child. In another embodiment, antibody raised against whole bacterial organism is administered orally to a subject, in a delivery solution, in therapeutic dosage.

[0110] V. Multivalent Vaccines Against Clostridial Species

[0111] The invention contemplates the generation of multivalent vaccines for the protection of an organism (particularly humans) against several Clostridal species. Of particular interest is a vaccine which stimulates the production of a humoral immune response to C. botulinum, C. tetani and C. difficile in humans. The antigens comprising the vaccine prepartion may be native or recombinantly produced toxin proteins from the clostridial species listed above. When native toxin proteins are used as immunogens they are generally modified to reduce the toxicity. A preferred toxin modification is formaldehyde treatment.

[0112] The invention contemplates that recombinant C. difficile toxin A and toxin B proteins be used in conjunction with either native toxins or toxoids from C. botulinum and C. tetani as antigens in a multivalent vaccine preparation. It is also contemplated that recombinant C. botulium toxin proteins be used in the multivalent vaccine preparation. It is comtemplated that, due to the sturctural similarity of C. botulinum and C. tetani toxin proteins, a vaccine comprising C. difficile and botulinum toxin proteins (native or recombinant or a mixture thereof) be used to stimulate an immune response against C. botulinum, C. tetani and C. difficile.

[0113] VI. Detection of Toxin

[0114] The invention contemplates detecting bacterial toxin in a sample. The term “sample” in the present specification and claims is used in its broadest sense. On the one hand it is meant to include a specimen or culture. On the other hand, it is meant to include both biological and environmental samples.

[0115] Biological samples may be animal, including human, fluid, solid (e.g., stool) or tissue; liquid and solid food products and ingredients such as dairy items, vegetables, meat and meat by-products, and waste. Environmental samples include environmental material such as surface matter, soil, water and industrial samples, as well as samples obtained from food and dairy processing instruments, apparatus, equipment, disposable and non-disposable items. These examples are not to be construed as limiting the sample types applicable to the present invention.

[0116] The invention contemplates detecting bacterial toxin by a competitive immunoassay method that utilizes recombinant toxin A and toxin B proteins, antibodies raised against recombinant bacterial toxin proteins. A fixed amount of the recombinant toxin proteins are immobilized to a solid support (e.g. a microtitier plate) followed by the addition of a biological sample suspected of containing a bacterial toxin. The biological sample is first mixed with affinity-purified or PEG fractionated antibodies directed against the recombinant toxin protein. A reporter reagent is then added which is capable of detecting the presence of antibody bound to the immobilized toxin protein. The reporter substance may comprise an antibody with binding specificity for the antitoxin attached to a molecule which is used to identify the presence of the reporter substance. If toxin is present in the sample, this toxin will compete with the immobilized recombinant toxin protein for binding to the anti-recombinant antibody thereby reducing the signal obtained following the addition of the reporter reagent. A control is employed where the antibody is not mixed with the sample. This gives the highest (or reference) signal.

[0117] The invention also contemplates detecting bacterial toxin by a “sandwich” immunoassay method that utilizes antibodies directed against recombinant bacterial toxin proteins. Affinity-purified antibodies directed against recombinant bacterial toxin proteins are immobilized to a solid support (e.g., microtiter plates). Biological samples suspected of containing bacterial toxins are then added followed by a washing step to remove substantially all unbound antitoxin. The biological sample is next exposed to the reporter substance, which binds to antitoxin and is then washed free of substantially all unbound reporter substance. The reporter substance may comprise an antibody with binding specificity for the antitoxin attached to a molecule which is used to identify the presence of the reporter substance. Identification of the reporter substance in the biological tissue indicates the presence of the bacterial toxin.

[0118] It is also contemplated that bacterial toxin be detected by pouring liquids (e.g. soups and other fluid foods and feeds including nutritional supplements for humans and other animals) over immobilized antibody which is directed against the bacterial toxin. It is contemplated that the immobilized antibody will be present in or on such supports as cartridges, columns, beads, or any other solid support medium. In one embodiment, following the exposure of the liquid to the immobilized antibody, unbound toxin is substantially removed by washing. The exposure of the liquid is then exposed to a reporter substance which detects the presence of bound toxin. In a preferred embodiment the reporter substance is an enzyme, fluorescent dye, or radioactive compound attached to an antibody which is directed against the toxin (i.e., in a “sandwich” immunoassay). It is also contemplated that the detection system will be developed as necessary (e.g., the addition of enzyme substrate in enzyme systems; observation using fluorescent light for fluorescent dye systems; and quantitation of radioactivity for radioactive systems).

[0119] Experimental

[0120] The following examples serve to illustrate certain preferred embodiments and aspects of the present invention and are not to be construed as limiting the scope thereof.

[0121] In the disclosure which follows, the following abbreviations apply: ° C. (degrees Centigrade); rpm (revolutions per minute); BBS-Tween (borate buffered saline containing Tween); BSA (bovine serum albumin); ELISA (enzyme-linked immunosorbent assay); CFA (complete Freund's adjuvant); IFA (incomplete Freund's adjuvant); IgG (immunoglobulin G); IgY (immunoglobulin Y); IM (intramuscular); IP (intraperitoneal); IV (intravenous or intravascular); SC (subcutaneous); H2O (water); HCl (hydrochloric acid); LD100 (lethal dose for 100% of experimental animals); aa (amino acid); HPLC (high performance liquid chromatography); kD (kilodaltons); gm (grams); μg (micrograms); mg (milligrams); ng (nanograms); μl (microliters); ml (milliliters); mm (millimeters); nm (nanometers); μm (micrometer); M (molar); mM (millimolar); MW (molecular weight); sec (seconds); min(s) (minute/minutes); hr(s) (hour/hours); MgCl2 (magnesium chloride); NaCl (sodium chloride); Na2CO3 (sodium carbonate); OD280 (optical density at 280 nm); OD600 (optical density at 600 nm); PAGE (polyacrylamide gel electrophoresis); PBS [phosphate buffered saline (150 mM NaCl, 10 mM sodium phosphate buffer, pH 7.2)]; PEG (polyethylene glycol); PMSF (phenylmethylsulfonyl fluoride); SDS (sodium dodecyl sulfate); Tris (tris(hydroxymethyl)aminomethane); Ensure® (Ensure®, Ross Laboratories, Columbus Ohio); Enfamil® (Enfamil®, Mead Johnson); w/v (weight to volume); v/v (volume to volume); Amicon (Amicon, Inc., Beverly, Mass.); Amresco (Amresco, Inc., Solon, Ohio); ATCC (American Type Culture Collection, Rockville, Md.); BBL (Baltimore Biologics Laboratory, (a division of Becton Dickinson), Cockeysville, Md.); Becton Dickinson (Becton Dickinson Labware, Lincoln Park, N.J.); BioRad (BioRad, Richmond, Calif.); Biotech (C-C Biotech Corp., Poway, Calif.); Charles River (Charles River Laboratories, Wilmington, Mass.); Cocalico (Cocalico Biologicals Inc., Reamstown, Pa.); CytRx (CytRx Corp., Norcross, Ga.); Falcon (e.g. Baxter Healthcare Corp., McGaw Park, Ill. and Becton Dickinson); Fisher Biotech (Fisher Biotech, Springfield, N.J.); GIBCO (Grand Island Biologic Company/BRL, Grand Island, N.Y.); Harlan Sprague Dawley (Harlan Sprague Dawley, Inc., Madison, Wis.); Mallinckrodt (a division of Baxter Healthcare Corp., McGaw Park, Ill.); Millipore (Millipore Corp., Marlborough, Mass.); New England Biolabs (New England Biolabs, Inc., Beverly, Mass.); Novagen (Novagen, Inc., Madison, Wis.); Pharmacia (Pharmacia, Inc., Piscataway, N.J.); Qiagen (Qiagen, Chatsworth, Calif.); Sasco (Sasco, Omaha, NE); Showdex (Showa Denko America, Inc., New York, N.Y.); Sigma (Sigma Chemical Co., St. Louis, Mo.); Sterogene (Sterogene, Inc., Arcadia, Calif.); Tech Lab (Tech Lab, Inc., Blacksburg, Va.); and Vaxcell (Vaxcell, Inc., a subsidiary of CytRX Corp., Norcross, Ga.).

[0122] When a recombinant protein is described in the specification it is refered to in a short-hand manner by the amino acids in the toxin sequence present in the recombinant protein rounded to the nearest 10. For example, the recombinant protein pMB1850-2360 contains amino acids 1852 through 2362 of the C. difficile toxin B protein. The specification gives detailed construction details for all recombinant proteins such that one skilled in the art will know precisely which amino acids are present in a given recombinant protein.

EXAMPLE 1

Production of High-Titer Antibodies to Clostridium difficile Organisms in a Hen

[0123] Antibodies to certain pathogenic organisms have been shown to be effective in treating diseases caused by those organisms. It has not been shown whether antibodies can be raised, against Clostridium difficile, which would be effective in treating infection by this organism. Accordingly, C. difficile was tested as immunogen for production of hen antibodies.

[0124] To determine the best course for raising high-titer egg antibodies against whole C. difficile organisms, different immunizing strains and different immunizing concentrations were examined. The example involved (a) preparation of the bacterial immunogen, (b) immunization, (c) purification of anti-bacterial chicken antibodies, and (d) detection of anti-bacterial antibodies in the purified IgY preparations.

[0125] a) Preparation of Bacterial Immunogen

[0126] C. difficile strains 43594 (serogroup A) and 43596 (serogroup C) were originally obtained from the ATCC. These two strains were selected because they represent two of the most commonly-occurring serogroups isolated from patients with antibiotic-associated pseudomembranous colitis. [Delmee et al., J. Clin. Microbiol., 28(10):2210 (1990).] Additionally, both of these strains have been previously characterized with respect to their virulence in the Syrian hamster model for C. difficile infection. [Delmee et al., J. Med Microbiol., 33:85 (1990).]

[0127] The bacterial strains were separately cultured on brain heart infusion agar for 48 hours at 37° C. in a Gas Pack 100 Jar (BBL, Cockeysville, Md.) equipped with a Gas Pack Plus anaerobic envelope (BBL). Forty-eight hour cultures were used because they produce better growth and the organisms have been found to be more cross-reactive with respect to their surface antigen presentation. The greater the degree of cross-reactivity of our IgY preparations, the better the probability of a broad range of activity against different strains/serogroups. [Toma et al., J. Clin. Microbiol., 26(3):426 (1988).]

[0128] The resulting organisms were removed from the agar surface using a sterile dacron-tip swab, and were suspended in a solution containing 0.4% formaldehyde in PBS, pH 7.2. This concentration of formaldehyde has been reported as producing good results for the purpose of preparing whole-organism immunogen suspensions for the generation of polyclonal anti-C. difficile antisera in rabbits. [Delmee et al., J. Clin. Microbiol., 21:323 (1985); Davies et al., Microbial Path., 9:141 (1990).] In this manner, two separate bacterial suspensions were prepared, one for each strain. The two suspensions were then incubated at 4° C. for 1 hour. Following this period of formalin-treatment, the suspensions were centrifuged at 4,200× g for 20 min., and the resulting pellets were washed twice in normal saline. The washed pellets, which contained formalin-treated whole organisms, were resuspended in fresh normal saline such that the visual turbidity of each suspension corresponded to a #7 McFarland standard. [M. A. C. Edelstein, “Processing Clinical Specimens for Anaerobic Bacteria: Isolation and Identification Procedures,” in S. M. Finegold et al (eds.)., Bailey and Scott's Diagnostic Microbiology, pp. 477-507, C.V. Mosby Co., (1990). The preparation of McFarland nephelometer standards and the corresponding approximate number of organisms for each tube are described in detail at pp. 172-173 of this volume.] Each of the two #7 suspensions was then split into two separate volumes. One volume of each suspension was volumetrically adjusted, by the addition of saline, to correspond to the visual turbidity of a #1 McFarland standard. [Id.] The #1 suspensions contained approximately 3×108 organisms/ml, and the #7 suspensions contained approximately 2×109 organisms/ml. [Id.] The four resulting concentration-adjusted suspensions of formalin-treated C. difficile organisms were considered to be “bacterial immunogen suspensions.” These suspensions were used immediately after preparation for the initial immunization. [See section (b).]

[0129] The formalin-treatment procedure did not result in 100% non-viable bacteria in the immunogen suspensions. In order to increase the level of killing, the formalin concentration and length of treatment were both increased for subsequent immunogen preparations, as described below in Table 3. (Although viability was decreased with the stronger formalin treatment, 100% inviability of the bacterial immunogen suspensions was not reached.) Also, in subsequent immunogen preparations, the formalin solutions were prepared in normal saline instead of PBS. At day 49, the day of the fifth immunization, the excess volumes of the four previous bacterial immunogen suspensions were stored frozen at −70° C. for use during all subsequent immunizations.

[0130] b) Immunization

[0131] For the initial immunization, 1.0 ml volumes of each of the four bacterial immunogen suspensions described above were separately emulsified in 1.2 ml volumes of CFA (GIBCO). For each of the four emulsified immunogen suspensions, two four-month old White Leghorn hens (pre-laying) were immunized. (It is not necessary to use pre-laying hens; actively-laying hens can also be utilized.) Each hen received a total volume of approximately 1.0 ml of a single emulsified immunogen suspension via four injections (two subcutaneous and two intramuscular) of approximately 250 μl per site. In this manner, a total of four different immunization combinations, using two hens per combination, were initiated for the purpose of evaluating both the effect of immunizing concentration on egg yolk antibody (IgY) production, and interstrain cross-reactivity of IgY raised against heterologous strains. The four immunization groups are summarized in Table 3.

[0132] The time point for the first series of immunizations was designated as “day zero.” All subsequent immunizations were performed as described above except that the bacterial immunogen suspensions were emulsified using IFA (GIBCO) instead of

TABLE 3
Immunization Groups
GROUP	IMMUNIZING	APPROXIMATE
DESIGNATION	STRAIN	IMMUNIZING DOSE
CD 43594, #1	C. difficile	1.5 × 108 organisms/hen
	strain 43594
CD 43594, #7	″	1.0 × 109 organisms/hen
CD 43596, #1	C. difficile	1.5 × 108 organisms/hen
	strain 43596
CD 43596, #7	″	1.0 × 109 organisms/hen

[0133] CFA, and for the later time point immunization, the stored frozen suspensions were used instead of freshly-prepared suspensions. The immunization schedule used is listed in Table 4.

[0134] c) Purification of Anti-Bacterial Chicken Antibodies

[0135] Groups of four eggs were collected per immunization group between days 80 and 84 post-initial immunization, and chicken immunoglobulin (IgY) was extracted according to a modification of the procedure of A. Polson et al., Immunol. Comm., 9:495 (1980). A gentle stream of distilled water from a squirt bottle was used to separate the yolks from the whites, and the yolks were broken by dropping them through a funnel into a graduated cylinder. The four individual yolks were pooled for each group. The pooled, broken yolks were blended with 4 volumes of egg extraction buffer to improve antibody yield (egg extraction buffer is 0.01 M sodium phosphate, 0.1 M NaCl, pH 7.5, containing 0.005% thimerosal), and PEG 8000 (Amresco) was

TABLE 4
Immunization Schedule
DAY OF	FORMALIN-	IMMUNOGEN
IMMUNIZATION	TREATMENT	PREPARATION USED
0	1%, 1 hr.	freshly-prepared
14	1%, overnight	″
21	1%, overnight	″
35	1%, 48 hrs.	″
49	1%, 72 hrs.	″
70	″	stored frozen
85	″	″
105	″	″

[0136] added to a concentration of 3.5%. When all the PEG dissolved, the protein precipitates that formed were pelleted by centrifugation at 13,000× g for 10 minutes. The supernatants were decanted and filtered through cheesecloth to remove the lipid layer, and the PEG was added to the supernatants to a final concentration of 12% (the supernatants were assumed to contain 3.5% PEG). After a second centrifugation, the supernatants were discarded and the pellets were centrifuged a final time to extrude the remaining PEG. These crude IgY pellets were then dissolved in the original yolk volume of egg extraction buffer and stored at 4° C. As an additional control, a preimmune IgY solution was prepared as described above, using eggs collected from unimmunized hens.

[0137] d) Detection of Anti-Bacterial Antibodies in the Purified IgY Preparations

[0138] In order to evaluate the relative levels of specific anti-C. difficile activity in the IgY preparations described above, a modified version of the whole-organism ELISA procedure of N. V. Padhye et al., J. Clin. Microbiol. 29:99-103 (1990) was used. Frozen organisms of both C. difficile strains described above were thawed and diluted to a concentration of approximately 1×107 organisms/ml using PBS, pH 7.2. In this way, two separate coating suspensions were prepared, one for each immunizing strain. Into the wells of 96-well microtiter plates (Falcon, Pro-Bind Assay Plates) were placed 100 μl volumes of the coating suspensions. In this manner, each plate well received a total of approximately 1×106 organisms of one strain or the other. The plates were then incubated at 4° C. overnight. The next morning, the coating suspensions were decanted, and all wells were washed three times using PBS. In order to block non-specific binding sites, 100 μl of 0.5% BSA (Sigma) in PBS was then added to each well, and the plates were incubated for 2 hours at room temperature. The blocking solution was decanted, and 100 μl volumes of the IgY preparations described above were initially diluted 1:500 with a solution of 0.1% BSA in PBS, and then serially diluted in 1:5 steps. The following dilutions were placed in the wells: 1:500, 1:2,500, 1:62,5000, 1:312,500, and 1:1,562,500. The plates were again incubated for 2 hours at room temperature. Following this incubation, the IgY-containing solutions were decanted, and the wells were washed three times using BBS-Tween (0.1 M boric acid, 0.025 M sodium borate, 1.0 M NaCl, 0.1% Tween-20), followed by two washes using PBS-Tween (0.1% Tween-20), and finally, two washes using PBS only. To each well, 100 μl of a 1:750 dilution of rabbit anti-chicken IgG (whole-molecule)-alkaline phosphatase conjugate (Sigma) (diluted in 0.1% BSA in PBS) was added. The plates were again incubated for 2 hours at room temperature. The conjugate solutions were decanted and the plates were washed as described above, substituting 50 mM Na2CO3, pH 9.5 for the PBS in the final wash. The plates were developed by the addition of 100 μl of a solution containing 1 mg/ml para-nitrophenyl phosphate (Sigma) dissolved in 50 mM Na2CO3, 10 mM MgCl2, pH 9.5 to each well, and incubating the plates at

TABLE 5
Results Of The Anti-C. difficile Whole-Organism ELISA
IgY	DILUTION OF	43594-COATED	43596-COATED
PREPARATION	IgY PREP	WELLS	WELLS
CD 43594, #1	1:500	1.746	1.801
	1:2,500	1.092	1.670
	1:12,500	0.202	0.812
	1:62,500	0.136	0.179
	1:312,500	0.012	0.080
	1:1,562,500	0.002	0.020
CD 43594, #7	1:500	1.780	1.771
	1:2,500	1.025	1.078
	1:12,500	0.188	0.382
	1:62,500	0.052	0.132
	1:312,500	0.022	0.043
	1:1,562,500	0.005	0.024
CD 43596, #1	1:500	1.526	1.790
	1:2,500	0.832	1.477
	1:12,500	0.247	0.452
	1:62,500	0.050	0.242
	1:312,500	0.010	0.067
	1:1,562,500	0.000	0.036
CD 43596, #7	1:500	1.702	1.505
	1:2,500	0.706	0.866
	1:12,500	0.250	0.282
	1:62,500	0.039	0.078
	1:312,500	0.002	0.017
	1:1,562,500	0.000	0.010
Preimmune IgY	1:500	0.142	0.309
	1:2,500	0.032	0.077
	1:12,500	0.006	0.024
	1:62,500	0.002	0.012
	1:312,500	0.004	0.010
	1:1,562,500	0.002	0.014

[0139] room temperature in the dark for 45 minutes. The absorbance of each well was measured at 410 nm using a Dynatech MR 700 plate reader. In this manner, each of the four IgY preparations described above was tested for reactivity against both of the immunizing C. difficile strains; strain-specific, as well as cross-reactive activity was determined.

[0140] Table 5 shows the results of the whole-organism ELISA. All four IgY preparations demonstrated significant levels of activity, to a dilution of 1:62,500 or greater against both of the immunizing organism strains. Therefore, antibodies raised against one strain were highly cross-reactive with the other strain, and vice versa. The immunizing concentration of organisms did not have a significant effect on organism-specific IgY production, as both concentrations produced approximately equivalent responses. Therefore, the lower immunizing concentration of approximately 1.5×108 organisms/hen is the preferred immunizing concentration of the two tested. The preimmune IgY preparation appeared to possess relatively low levels of C. difficile-reactive activity to a dilution of 1:500, probably due to prior exposure of the animals to environmental clostridia.

[0141] An initial whole-organism ELISA was performed using IgY preparations made from single CD 43594, #1 and CD 43596, #1 eggs collected around day 50 (data not shown). Specific titers were found to be 5 to 10-fold lower than those reported in Table 5. These results demonstrate that it is possible to begin immunizing hens prior to the time that they begin to lay eggs, and to obtain high titer specific IgY from the first eggs that are laid. In other words, it is not necessary to wait for the hens to begin laying before the immunization schedule is started.

EXAMPLE 2

Treatment of C. difficile Infection with Anti-C. difficile Antibody

[0142] In order to determine whether the immune IgY antibodies raised against whole C. difficile organisms were capable of inhibiting the infection of hamsters by C. difficile, hamsters infected by these bacteria were utilized. [Lyerly et al., Infect. Immun., 59:2215-2218 (1991).] This example involved: (a) determination of the lethal dose of C. difficile organisms; and (b) treatment of infected animals with immune antibody or control antibody in nutritional solution.

[0143] a) Determination of the Lethal Dose of C. difficile Organisms

[0144] Determination of the lethal dose of C. difficile organisms was carried out according to the model described by D. M. Lyerly et al., Infect. Immun., 59:2215-2218 (1991). C. difficile strain ATCC 43596 (serogroup C, ATCC) was plated on BHI agar and grown anaerobically (BBL Gas Pak 100 system) at 37° C. for 42 hours. Organisms were removed from the agar surface using a sterile dacron-tip swab and suspended in sterile 0.9% NaCl solution to a density of 108 organisms/ml.

[0145] In order to determine the lethal dose of C. difficile in the presence of control antibody and nutritional formula, non-immune eggs were obtained from unimmunized hens and a 12% PEG preparation made as described in Example 1(c). This preparation was redissolved in one fourth the original yolk volume of vanilla flavor Ensure®.

[0146] Starting on day one, groups of female Golden Syrian hamsters (Harlan Sprague Dawley), 8-9 weeks old and weighing approximately 100 gm, were orally administered 1 ml of the preimmune/Ensure® formula at time zero, 2 hours, 6 hours, and 10 hours. At 1 hour, animals were orally administered 3.0 mg clindamycin HCl (Sigma) in 1 ml of water. This drug predisposes hamsters to C. difficile infection by altering the normal intestinal flora. On day two, the animals were given 1 ml of the preimmune IgY/Ensure® formula at time zero, 2 hours, 6 hours, and 10 hours. At 1 hour on day two, different groups of animals were inoculated orally with saline (control), or 102, 104, 106, or 108 C. difficile organisms in 1 ml of saline. From days 3-12, animals were given 1 ml of the preimmune IgY/Ensure® formula three times daily and observed for the onset of diarrhea and death. Each animal was housed in an individual cage and was offered food and water ad libitum.

[0147] Administration of 106-108 organisms resulted in death in 3-4 days while the lower doses of 102-104 organisms caused death in 5 days. Cecal swabs taken from dead animals indicated the presence of C. difficile. Given the effectiveness of the 102 dose, this number of organisms was chosen for the following experiment to see if hyperimmune anti-C. difficile antibody could block infection.

[0148] b) Treatment of Infected Animals with Immune Antibody or Control Antibody in Nutritional Formula

[0149] The experiment in (a) was repeated using three groups of seven hamsters each. Group A received no clindamycin or C. difficile and was the survival control. Group B received clindamycin, 102 C. difficile organisms and preimmune IgY on the same schedule as the animals in (a) above. Group C received clindamycin, 102 C. difficile organisms, and hyperimmune anti-C. difficile IgY on the same schedule as Group B. The anti-C. difficile IgY was prepared as described in Example 1 except that the 12% PEG preparation was dissolved in one fourth the original yolk volume of Ensure®.

[0150] All animals were observed for the onset of diarrhea or other disease symptoms and death. Each animal was housed in an individual cage and was offered food and water ad libitum. The results are shown in Table 6.

[0151] Hamsters in the control group A did not develop diarrhea and remained healthy during the experimental period. Hamsters in groups B and C developed diarrheal disease. Anti-C. difficile IgY did not protect the animals from diarrhea or death, all animals succumbed in the same time interval as the animals treated with preimmune

TABLE 6
The Effect Of Oral Feeding Of Hyperimmune IgY
Antibody on C. difficile Infection
	TIME TO	TIME TO
ANIMAL GROUP	DIARRHEAa	DEATHa
A pre-immune IgY only	no diarrhea	no deaths
B Clindamycin, C. difficile, preimmune IgY	30 hrs.	49 hrs.
C Clindamycin, C. difficile, immune IgY	33 hrs.	56 hrs.
amean of seven animals.

[0152] IgY. Thus, while immunization with whole organisms apparently can improve sub-lethal symptoms with particular bacteria (see U.S. Pat. No. 5,080,895 to H. Tokoro), such an approach does not prove to be productive to protect against the lethal effects of C. difficile.

EXAMPLE 3

Production of C. botulinum Type A Antitoxin in Hens

[0153] In order to determine whether antibodies could be raised against the toxin produced by clostridial pathogens, which would be effective in treating clostridial diseases, antitoxin to C. botulinum type A toxin was produced. This example involves: (a) toxin modification; (b) immunization; (c) antitoxin collection; (d) antigenicity assessment; and (e) assay of antitoxin titer.

[0154] a) Toxin Modification

[0155] C. botulinum type A toxoid was obtained from B. R. DasGupta. From this, the active type A neurotoxin (M.W. approximately 150 kD) was purified to greater than 99% purity, according to published methods. [B. R. DasGupta & V. Sathyamoorthy, Toxicon, 22:415 (1984).] The neurotoxin was detoxified with formaldehyde according to published methods. [B. R. Singh & B. R. DasGupta, Toxicon, 27:403 (1989).]

[0156] b) Immunization

[0157] C. botulinum toxoid for immunization was dissolved in PBS (1 mg/ml) and was emulsified with an approximately equal volume of CFA (GIBCO) for initial immunization or IFA for booster immunization. On day zero, two white leghorn hens, obtained from local breeders, were each injected at multiple sites (intramuscular and subcutaneous) with 1 ml inactivated toxoid emulsified in 1 ml CFA. Subsequent booster immunizations were made according to the following schedule for day of injection and toxoid amount: days 14 and 21-0.5 mg; day 171-0.75 mg; days 394, 401, 409-0.25 mg. One hen received an additional booster of 0.150 mg on day 544.

[0158] c) Antitoxin Collection

[0159] Total yolk immunoglobulin (IgY) was extracted as described in Example 1(c) and the IgY pellet was dissolved in the original yolk volume of PBS with thimerosal.

[0160] d) Antigenicity Assessment

[0161] Eggs were collected from day 409 through day 423 to assess whether the toxoid was sufficiently immunogenic to raise antibody. Eggs from the two hens were pooled and antibody was collected as described in the standard PEG protocol. [Example 1(c).] Antigenicity of the botulinal toxin was assessed on Western blots. The 150 kD detoxified type A neurotoxin and unmodified, toxic, 300 kD botulinal type A complex (toxin used for intragastric route administration for animal gut neutralization experiments; see Example 6) were separated on a SDS-polyacrylamide reducing gel. The Western blot technique was performed according to the method of Towbin. [H. Towbin et al., Proc. Nat'l Acad. Sci. USA, 76:4350 (1979).] Ten μg samples of C botulinum complex and toxoid were dissolved in SDS reducing sample buffer (1% SDS, 0.5% 2-mercaptoethanol, 50 mM Tris, pH 6.8, 10% glycerol, 0.025% w/v bromphenol blue, 10% β-mercaptoethanol), heated at 95° C. for 10 min and separated on a 1 mm thick 5% SDS-polyacrylamide gel. [K. Weber and M. Osborn, “Proteins and Sodium Dodecyl Sulfate: Molecular Weight Determination on Polyacrylamide Gels and Related Procedures,” in The Proteins, 3d Edition (H. Neurath & R.L. Hill, eds), pp. 179-223, (Academic Press, NY, 1975).] Part of the gel was cut off and the proteins were stained with Coomassie Blue. The proteins in the remainder of the gel were transferred to nitrocellulose using the Milliblot-SDE electro-blotting system (Millipore) according to manufacturer's directions. The nitrocellulose was temporarily stained with 10% Ponceau S [S. B. Carroll and A. Laughon, “Production and Purification of Polyclonal Antibodies to the Foreign Segment of β-galactosidase Fusion Proteins,” in DNA Cloning: A Practical Approach, Vol.III, (D. Glover, ed.), pp. 89-111, IRL Press, Oxford, (1987)] to visualize the lanes, then destained by running a gentle stream of distilled water over the blot for several minutes. The nitrocellulose was immersed in PBS containing 3% BSA overnight at 4° C. to block any remaining protein binding sites.

[0162] The blot was cut into strips and each strip was incubated with the appropriate primary antibody. The avian anti-C. botulinum antibodies [described in (c)] and pre-immune chicken antibody (as control) were diluted 1:125 in PBS containing 1 mg/ml BSA for 2 hours at room temperature. The blots were washed with two changes each of large volumes of PBS, BBS-Tween and PBS, successively (10 min/wash). Goat anti-chicken IgG alkaline phosphatase conjugated secondary antibody (Fisher Biotech) was diluted 1:500 in PBS containing 1 mg/ml BSA and incubated with the blot for 2 hours at room temperature. The blots were washed with two changes each of large volumes of PBS and BBS-Tween, followed by one change of PBS and 0.1 M Tris-HCl, pH 9.5. Blots were developed in freshly prepared alkaline phosphatase substrate buffer (100 μg/ml nitroblue tetrazolium (Sigma), 50 μg/ml 5-bromo-4-chloro-3-indolyl phosphate (Sigma), 5 mM MgCl2 in 50 mM Na2CO3, pH 9.5).

[0163] The Western blots are shown in FIG. 1. The anti-C. botulinum IgY reacted to the toxoid to give a broad immunoreactive band at about 145-150 kD on the reducing gel. This toxoid is refractive to disulfide cleavage by reducing agents due to formalin crosslinking. The immune IgY reacted with the active toxin complex, a 97 kD C. hotulinum type A heavy chain and a 53 kD light chain. The preimmune IgY was unreactive to the C. botulinum complex or toxoid in the Western blot.

[0164] e) Antitoxin Antibody Titer

[0165] The IgY antibody titer to C. botulinum type A toxoid of eggs harvested between day 409 and 423 was evaluated by ELISA, prepared as follows. Ninety-six-well Falcon Pro-bind plates were coated overnight at 4° C. with 100 μl/well toxoid [B. R. Singh & B. R. Das Gupta, Toxicon 27:403 (1989)] at 2.5 μg/ml in PBS, pH 7.5 containing 0.005% thimerosal. The following day the wells were blocked with PBS containing 1% BSA for 1 hour at 37° C. The IgY from immune or preimmune eggs was diluted in PBS containing 1% BSA and 0.05% Tween 20 and the plates were incubated for 1 hour at 37° C. The plates were washed three times with PBS containing 0.05% Tween 20 and three times with PBS alone. Alkaline phosphatase-conjugated goat-anti-chicken IgG (Fisher Biotech) was diluted 1:750 in PBS containing 1% BSA and 0.05% Tween 20, added to the plates, and incubated 1 hour at 37° C. The plates were washed as before, and p-nitrophenyl phosphate (Sigma) at 1 mg/ml in 0.05 M Na2CO3, pH 9.5, 10 mM MgCl2 was added.

[0166] The results are shown in FIG. 2. Chickens immunized with the toxoid generated high titers of antibody to the immunogen. Importantly, eggs from both immunized hens had significant anti-immunogen antibody titers as compared to preimmune control eggs. The anti-C. botulinum IgY possessed significant activity, to a dilution of 1:93,750 or greater.

EXAMPLE 4

Preparation of Avian Egg Yolk Immunoglobulin in an Orally Administrable Form

[0167] In order to administer avian IgY antibodies orally to experimental mice, an effective delivery formula for the IgY had to be determined. The concern was that if the crude IgY was dissolved in PBS, the saline in PBS would dehydrate the mice, which might prove harmful over the duration of the study. Therefore, alternative methods of oral administration of IgY were tested. The example involved: (a) isolation of immune IgY; (b) solubilization of IgY in water or PBS, including subsequent dialysis of the IgY-PBS solution with water to eliminate or reduce the salts (salt and phosphate) in the buffer; and (c) comparison of the quantity and activity of recovered IgY by absorbance at 280 nm and PAGE, and enzyme-linked immunoassay (ELISA).

[0168] a) Isolation of Immune IgY

[0169] In order to investigate the most effective delivery formula for IgY, we used IgY which was raised against Crotalus durissus terrificus venom. Three eggs were collected from hens immunized with the C. durissus terrificus venom and IgY was extracted from the yolks using the modified Polson procedure described by Thalley and Carroll [Bio/Technology, 8:934-938 (1990)] as described in Example 1(c).

[0170] The egg yolks were separated from the whites, pooled, and blended with four volumes of PBS. Powdered PEG 8000 was added to a concentration of 3.5%. The mixture was centrifuged at 10,000 rpm for 10 minutes to pellet the precipitated protein, and the supernatant was filtered through cheesecloth to remove the lipid layer. Powdered PEG 8000 was added to the supernatant to bring the final PEG concentration to 12% (assuming a PEG concentration of 3.5% in the supernatant). The 12% PEG/IgY mixture was divided into two equal volumes and centrifuged to pellet the IgY.

[0171] b) Solubilization of the IgY in Water or PBS

[0172] One pellet was resuspended in ½ the original yolk volume of PBS, and the other pellet was resuspended in ½ the original yolk volume of water. The pellets were then centrifuged to remove any particles or insoluble material. The IgY in PBS solution dissolved readily but the fraction resuspended in water remained cloudy.

[0173] In order to satisfy anticipated sterility requirements for orally administered antibodies, the antibody solution needs to be filter-sterilized (as an alternative to heat sterilization which would destroy the antibodies). The preparation of IgY resuspended in water was too cloudy to pass through either a 0.2 or 0.45 μm membrane filter, so 10 ml of the PBS resuspended fraction was dialyzed overnight at room temperature against 250 ml of water. The following morning the dialysis chamber was emptied and refilled with 250 ml of fresh H2O for a second dialysis. Thereafter, the yields of soluble antibody were determined at OD280 and are compared in Table 7.

TABLE 7
Dependence Of IgY Yield On Solvents
	ABSORBANCE OF 1:10	PERCENT
FRACTION	DILUTION AT 280 nm	RECOVERY
PBS dissolved	1.149	100%
H2O dissolved	0.706	61%
PBS dissolved/H2O dialyzed	0.885	77%

[0174] Resuspending the pellets in PBS followed by dialysis against water recovered more antibody than directly resuspending the pellets in water (77% versus 61%). Equivalent volumes of the IgY preparation in PBS or water were compared by PAGE, and these results were in accordance with the absorbance values (data not shown).

[0175] c) Activity of IgY Prepared with Different Solvents

[0176] An ELISA was performed to compare the binding activity of the IgY extracted by each procedure described above. C. durissus terrificus (C.d.t.) venom at 2.5 μg/ml in PBS was used to coat each well of a 96-well microtiter plate. The remaining protein binding sites were blocked with PBS containing 5 mg/ml BSA. Primary antibody dilutions (in PBS containing 1 mg/ml BSA) were added in duplicate. After 2 hours of incubation at room temperature, the unbound primary antibodies were removed by washing the wells with PBS, BBS-Tween, and PBS. The species specific secondary antibody (goat anti-chicken immunoglobulin alkaline-phosphatase conjugate (Sigma) was diluted 1:750 in PBS containing 1 mg/ml BSA and added to each well of the microtiter plate. After 2 hours of incubation at room temperature, the unbound secondary antibody was removed by washing the plate as before, and freshly prepared alkaline phosphatase substrate (Sigma) at 1 mg/ml in 50 mM Na2CO3, 10 mM MgCl2, pH 9.5 was added to each well. The color development was measured on a Dynatech MR 700 microplate reader using a 412 nm filter. The results are shown in Table 8.

[0177] The binding assay results parallel the recovery values in Table 7, with PBS-dissolved IgY showing slightly more activity than the PBS-dissolved/H2O dialyzed antibody. The water-dissolved antibody had considerably less binding activity than the other preparations.

EXAMPLE 5

Survival of Antibody Activity After Passage Through the Gastrointestinal Tract

[0178] In order to determine the feasibility of oral administration of antibody, it was of interest to determine whether orally administered IgY survived passage through the gastrointestinal tract. The example involved: (a) oral administration of specific immune antibody mixed with a nutritional formula; and (b) assay of antibody activity extracted from feces.

TABLE 8
Antigen-Binding Activity of IgY Prepared with Different Solvents
		PBS	H2O	PBS/
DILUTION	PREIMMUNE	DISSOLVED	DISSOLVED	H2O
1:500	0.005	1.748	1.577	1.742
1:2,500	0.004	0.644	0.349	0.606
1:12,500	0.001	0.144	0.054	0.090
1:62,500	0.001	0.025	0.007	0.016
1:312,500	0.010	0.000	0.000	0.002

[0179] a) Oral Administration of Antibody

[0180] The IgY preparations used in this example are the same PBS-dissolved/H2O dialyzed antivenom materials obtained in Example 4 above, mixed with an equal volume of Enfamil®. Two mice were used in this experiment, each receiving a different diet as follows:

[0181] 1) water and food as usual;

[0182] 2) immune IgY preparation dialyzed against water and mixed 1:1 with Enfamil®. (The mice were given the corresponding mixture as their only source of food and water).

[0183] b) Antibody Activity After Ingestion

[0184] After both mice had ingested their respective fluids, each tube was refilled with approximately 10 ml of the appropriate fluid first thing in the morning. By mid-morning there was about 4 to 5 ml of liquid left in each tube. At this point stool samples were collected from each mouse, weighed, and dissolved in approximately 500 μl PBS per 100 mg stool sample. One hundred and sixty mg of control stools (no antibody) and 99 mg of experimental stools (specific antibody) in 1.5 ml microfuge tubes were dissolved in 800 and 500 μl PBS, respectively. The samples were heated at 37° C. for 10 minutes and vortexed vigorously. The experimental stools were also broken up with a narrow spatula. Each sample was centrifuged for 5 minutes in a microfuge and the supernatants, presumably containing the antibody extracts, were collected. The pellets were saved at 2-8° C. in case future extracts were needed. Because the supernatants were tinted, they were diluted five-fold in PBS containing 1 mg/ml BSA for the initial dilution in the enzyme immunoassay (ELISA). The primary extracts were then diluted five-fold serially from this initial dilution. The volume of primary extract added to each well was 190 μl. The ELISA was performed exactly as described in Example 4.

TABLE 9
Specific Antibody Activity After Passage Through
the Gastrointestinal Tract
		CONTROL
		FECAL	EXP. FECAL
DILUTION	PREIMMUNE IgY	EXTRACT	EXTRACT
1.5	<0	0.000	0.032
1:25	0.016	<0	0.016
1:125	<0	<0	0.009
1:625	<0	0.003	0.001
1:3125	<0	<0	0.000

[0185] There was some active antibody in the fecal extract from the mouse given the specific antibody in Enfamil® formula, but it was present at a very low level. Since the samples were assayed at an initial 1:5 dilution, the binding observed could have been higher with less dilute samples. Consequently, the mice were allowed to continue ingesting either regular food and water or the specific IgY in Enfamil® formula, as appropriate, so the assay could be repeated. Another ELISA plate was coated overnight with 5 μg/ml of C.d.t. venom in PBS.

[0186] The following morning the ELISA plate was blocked with 5 mg/ml BSA, and the fecal samples were extracted as before, except that instead of heating the extracts at 37° C., the samples were kept on ice to limit proteolysis. The samples were assayed undiluted initially, and in 5× serial dilutions thereafter. Otherwise the assay was carried out as before.

TABLE 10
Specific Antibody Survives Passage Through
The Gastrointestinal Tract
		CONTROL
DILUTION	PREIMMUNE IgY	EXTRACT	EXP. EXTRACT
undiluted	0.003	<0	0.379
1:5	<0	<0	0.071
1:25	0.000	<0	0.027
1:125	0.003	<0	0.017
1:625	0.000	<0	0.008
1:3125	0.002	<0	0.002

[0187] The experiment confirmed the previous results, with the antibody activity markedly higher. The control fecal extract showed no anti-C.d.t. activity, even undiluted, while the fecal extract from the anti-C.d.t. IgY/Enfamil®-fed mouse showed considerable anti-C.d.t. activity. This experiment (and the previous experiment) clearly demonstrate that active IgY antibody survives passage through the mouse digestive tract, a finding with favorable implications for the success of IgY antibodies administered orally as a therapeutic or prophylactic.

EXAMPLE 6

In vivo Neutralization of Type C. botulinum Type A Neurotoxin by Avian Antitoxin Antibody

[0188] This example demonstrated the ability of PEG-purified antitoxin, collected as described in Example 3, to neutralize the lethal effect of C. botulinum neurotoxin type A in mice. To determine the oral lethal dose (LD100) of toxin A, groups of BALB/c mice were given different doses of toxin per unit body weight (average body weight of 24 grams). For oral administration, toxin A complex, which contains the neurotoxin associated with other non-toxin proteins was used. This complex is markedly more toxic than purified neurotoxin when given by the oral route. [I. Ohishi et al., Infect. Immun., 16:106 (1977).] C. botulinum toxin type A complex, obtained from Eric Johnson (University Of Wisconsin, Madison) was 250 μg/ml in 50 mM sodium citrate, pH 5.5, specific toxicity 3×107 mouse LD50/mg with parenteral administration. Approximately 40-50 ng/gm body weight was usually fatal within 48 hours in mice maintained on conventional food and water. When mice were given a diet ad libitum of only Enfamil® the concentration needed to produce lethality was approximately 2.5 times higher (125 ng/gm body weight). Botulinal toxin concentrations of approximately 200 ng/gm body weight were fatal in mice fed Enfamil(t containing preimmune IgY (resuspended in Enfamil® at the original yolk volume).

[0189] The oral LD100 of C. botulinum toxin was also determined in mice that received known amounts of a mixture of preimmune IgY-Ensure® delivered orally through feeding needles. Using a 22 gauge feeding needle, mice were given 250 μl each of a preimmune IgY-Ensure® mixture (preimmune IgY dissolved in ¼ original yolk volume) 1 hour before and ½ hour and 5 hours after administering botulinal toxin. Toxin concentrations given orally ranged from approximately 12 to 312 ng/gm body weight (0.3 to 7.5 μg per mouse). Botulinal toxin complex concentration of approximately 40 ng/gm body weight (1 μg per mouse) was lethal in all mice in less than 36 hours.

[0190] Two groups of BALB/c mice, 10 per group, were each given orally a single dose of 1 μg each of botulinal toxin complex in 100 μl of 50 mM sodium citrate pH 5.5. The mice received 250 μl treatments of a mixture of either preimmune or immune IgY in Ensure® (¼ original yolk volume) 1 hour before and ½ hour, 4 hours, and 8 hours after botulinal toxin administration. The mice received three treatments per day for two more days. The mice were observed for 96 hours. The survival and mortality are shown in Table 11.

TABLE 11
Neutralization of Botulinal Toxin A In Vivo
TOXIN DOSE		NUMBER OF	NUMBER OF
ng/gm	ANTIBODY TYPE	MICE ALIVE	MICE DEAD
41.6	non-immune	0	10
41.6	anti-botulinal toxin	10	0

[0191] All mice treated with the preimmune IgY-Ensure® mixture died within 46 hours post-toxin administration. The average time of death in the mice was 32 hours post toxin administration. Treatments of preimmune IgY-Ensure® mixture did not continue beyond 24 hours due to extensive paralysis of the mouth in mice of this group. In contrast, all ten mice treated with the immune anti-botulinal toxin IgY-Ensure® mixture survived past 96 hours. Only 4 mice in this group exhibited symptoms of botulism toxicity (two mice about 2 days after and two mice 4 days after toxin administration). These mice eventually died 5 and 6 days later. Six of the mice in this immune group displayed no adverse effects to the toxin and remained alive and healthy long term. Thus, the avian anti-botulinal toxin antibody demonstrated very good protection from the lethal effects of the toxin in the experimental mice.

EXAMPLE 7

Production of an Avian Antitoxin Against Clostridium difficile Toxin A

[0192] Toxin A is a potent cytotoxin secreted by pathogenic strains of C. difficile, that plays a direct role in damaging gastrointestinal tissues. In more severe cases of C. difficile intoxication, pseudomembranous colitis can develop which may be fatal. This would be prevented by neutralizing the effects of this toxin in the gastrointestinal tract. As a first step, antibodies were produced against a portion of the toxin. The example involved: (a) conjugation of a synthetic peptide of toxin A to bovine serum albumin; (b) immunization of hens with the peptide-BSA conjugate; and (c) detection of antitoxin peptide antibodies by ELISA.

[0193] a) Conjugation of a Synthetic Peptide of Toxin A to Bovine Serum Albumin

[0194] The synthetic peptide CQTIDGKKYYFN-NH2 was prepared commercially (Multiple Peptide Systems, San Diego, Calif.) and validated to be >80% pure by high-pressure liquid chromatography. The eleven amino acids following the cysteine residue represent a consensus sequence of a repeated amino acid sequence found in Toxin A. [Wren et al., Infect. Immun., 59:3151-3155 (1991).] The cysteine was added to facilitate conjugation to carrier protein.

[0195] In order to prepare the carrier for conjugation, BSA (Sigma) was dissolved in 0.01 M NaPO4, pH 7.0 to a final concentration of 20 mg/ml and n-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS; Pierce) was dissolved in N,N-dimethyl formamide to a concentration of 5 mg/ml. MBS solution, 0.51 ml, was added to 3.25 ml of the BSA solution and incubated for 30 minutes at room temperature with stirring every 5 minutes. The MBS-activated BSA was then purified by chromatography on a Bio-Gel P-10 column (Bio-Rad; 40 ml bed volume) equilibrated with 50 mM NaPO4, pH 7.0 buffer. Peak fractions were pooled (6.0 ml).

[0196] Lyophilized toxin A peptide (20 mg) was added to the activated BSA mixture, stirred until the peptide dissolved and incubated 3 hours at room temperature. Within 20 minutes, the reaction mixture became cloudy and precipitates formed. After 3 hours, the reaction mixture was centrifuged at 10,000× g for 10 min and the supernatant analyzed for protein content. No significant protein could be detected at 280 nm. The conjugate precipitate was washed three times with PBS and stored at 4° C. A second conjugation was performed with 15 mg of activated BSA and 5 mg of peptide and the conjugates pooled and suspended at a peptide concentration of 10 mg/ml in 10 mM NaPO4, pH 7.2.

[0197] b) Immunization of Hens with Peptide Conjugate

[0198] Two hens were each initially immunized on day zero by injection into two subcutaneous and two intramuscular sites with 1 mg of peptide conjugate that was emulsified in CFA (GIBCO). The hens were boosted on day 14 and day 21 with 1 mg of peptide conjugate emulsified in IFA (GIBCO).

[0199] c) Detection of Antitoxin Peptide Antibodies by ELISA

[0200] IgY was purified from two eggs obtained before immunization (pre-immune) and two eggs obtained 31 and 32 days after the initial immunization using PEG fractionation as described in Example 1.

[0201] Wells of a 96-well microtiter plate (Falcon Pro-Bind Assay Plate) were coated overnight at 4° C. with 100 μg/ml solution of the toxin A synthetic peptide in PBS, pH 7.2 prepared by dissolving 1 mg of the peptide in 1.0 ml of H2O and dilution of PBS. The pre-immune and immune IgY preparations were diluted in a five-fold series in a buffer containing 1% PEG 8000 and 0.1% Tween-20 (v/v) in PBS, pH 7.2. The wells were blocked for 2 hours at room temperature with 150 μl of a solution containing 5% (v/v) Carnation® nonfat dry milk and 1% PEG 8000 in PBS, pH 7.2. After incubation for 2 hours at room temperature, the wells were washed, secondary rabbit anti-chicken IgG-alkaline phosphatase (1:750) added, the wells washed again and the color development obtained as described in Example 1. The results are shown in Table 12.

TABLE 12
Reactivity Of IgY With Toxin Peptide
	ABSORBANCE AT 410 nm
		IMMUNE ANTI-
DILUTION OF PEG PREP	PREIMMUINE	PEPTIDE
1:100	0.013	0.253
1:500	0.004	0.039
1:2500	0.004	0.005

[0202] Clearly, the immune antibodies contain titers against this repeated epitope of toxin A.

EXAMPLE 8

Production of Avian Antitoxins Against Clostridium difficile Native Toxins A and B

[0203] To determine whether avian antibodies are effective for the neutralization of C. difficile toxins, hens were immunized using native C. difficile toxins A and B. The resulting egg yolk antibodies were then extracted and assessed for their ability to neutralize toxins A and B in vitro. The Example involved (a) preparation of the toxin immunogens, (b) immunization, (c) purification of the antitoxins, and (d) assay of toxin neutralization activity.

[0204] a) Preparation of the Toxin Immunogens

[0205] Both C. difficile native toxins A and B, and C. difficile toxoids, prepared by the treatment of the native toxins with formaldehyde, were employed as immunogens. C. difficile toxoids A and B were prepared by a procedure which was modified from published methods (Ehrich et al., Infect. Immun. 28:1041 (1980). Separate solutions (in PBS) of native C. difficile toxin A and toxin B (Tech Lab) were each adjusted to a concentration of 0.20 mg/ml, and formaldehyde was added to a final concentration of 0.4%. The toxin/formaldehyde solutions were then incubated at 37° C. for 40 hrs. Free formaldehyde was then removed from the resulting toxoid solutions by dialysis against PBS at 4° C. In previously published reports, this dialysis step was not performed. Therefore, free formaldehyde must have been present in their toxoid preparations. The toxoid solutions were concentrated, using a Centriprep concentrator unit (Amicon), to a final toxoid concentration of 4.0 mg/ml. The two resulting preparations were designated as toxoid A and toxoid B.

[0206] C. difficile native toxins were prepared by concentrating stock solutions of toxin A and toxin B (Tech Lab, Inc), using Centriprep concentrator units (Amicon), to a final concentration of 4.0 mg/ml.

[0207] b) Immunization

[0208] The first two immunizations were performed using the toxoid A and toxoid B immunogens described above. A total of 3 different immunization combinations were employed. For the first immunization group, 0.2 ml of toxoid A was emulsified in an equal volume of Titer Max adjuvant (CytRx). Titer Max was used in order to conserve the amount of immunogen used, and to simplify the immunization procedure. This immunization group was designated “CTA.” For the second immunization group, 0.1 ml of toxoid B was emulsified in an equal volume of Titer Max adjuvant. This group was designated “CTB.” For the third immunization group, 0.2 ml of toxoid A was first mixed with 0.2 ml of toxoid B, and the resulting mixture was emulsified in 0.4 ml of Titer Max adjuvant. This group was designated “CTAB.” In this way, three separate immunogen emulsions were prepared, with each emulsion containing a final concentration of 2.0 mg/ml of toxoid A (CTA) or toxoid B (CTB) or a mixture of 2.0 mg/ml toxoid A and 2.0 mg/ml toxoid B (CTAB).

[0209] On day 0, White Leghorn hens, obtained from a local breeder, were immunized as follows: Group CTA. Four hens were immunized, with each hen receiving 200 μg of toxoid A, via two intramuscular (I.M.) injections of 50 μl of CTA emulsion in the breast area. Group CTB. One hen was immunized with 200 μg of toxoid B, via two I.M. injections of 50 μl of CTB emulsion in the breast area. Group CTAB. Four hens were immunized, with each hen receiving a mixture containing 200 μg of toxoid A and 200 μg of toxoid B, via two I.M. injections of 100 μl of CTAB emulsion in the breast area. The second immunization was performed 5 weeks later, on day 35, exactly as described for the first immunization above.

[0210] In order to determine whether hens previously immunized with C. difficile toxoids could tolerate subsequent booster immunizations using native toxins, a single hen from group CTAB was immunized for a third time, this time using a mixture of the native toxin A and native toxin B described in section (a) above (these toxins were not formaldehyde-treated, and were used in their active form). This was done in order to increase the amount (titer) and affinity of specific antitoxin antibody produced by the hen over that achieved by immunizing with toxoids only. On day 62, 0.1 ml of a toxin mixture was prepared which contained 200 μg of native toxin A and 200 μg of native toxin B. This toxin mixture was then emulsified in 0.1 ml of Titer Max adjuvant. A single CTAB hen was then immunized with the resulting immunogen emulsion, via two I.M. injections of 100 μl each, into the breast area. This hen was marked with a wing band, and observed for adverse effects for a period of approximately 1 week, after which time the hen appeared to be in good health.

[0211] Because the CTAB hen described above tolerated the booster immunization with native toxins A and B with no adverse effects, it was decided to boost the remaining hens with native toxin as well. On day 70, booster immunizations were performed as follows: Group CTA. A 0.2 ml volume of the 4 mg/ml native toxin A solution was emulsified in an equal volume of Titer Max adjuvant. Each of the 4 hens was then immunized with 200 μg of native toxin A, as described for the toxoid A immunizations above. Group CTB. A 501 μl volume of the 4 mg/ml native toxin B solution was emulsified in an equal volume of Titer Max adjuvant. The hen was then immunized with 200 μg of native toxin B, as described for the toxoid B immunizations above. Group CTAB. A 0.15 ml volume of the 4 mg/ml native toxin A solution was first mixed with a 0.15 ml volume the 4 mg/ml native toxin B solution. The resulting toxin mixture was then emulsified in 0.3 ml of Titer Max adjuvant. The 3 remaining hens (the hen with the wing band was not immunized this time) were then immunized with 200 μg of native toxin A and 200 μg of native toxin B as described for the toxoid A+ toxoid B immunizations (CTAB) above. On day 85, all hens received a second booster immunization using native toxins, done exactly as described for the first boost with native toxins above.

[0212] All hens tolerated both booster immunizations with native toxins with no adverse effects. As previous literature references describe the use of formaldehyde-treated toxoids, this is apparently the first time that any immunizations have been performed using native C. difficile toxins.

[0213] c) Purification of Antitoxins

[0214] Eggs were collected from the hen in group CTB 10-12 days following the second immunization with toxoid (day 35 immunization described in section (b) above), and from the hens in groups CTA and CTAB 20-21 days following the second immunization with toxoid. To be used as a pre-immune (negative) control, eggs were also collected from unimmunized hens from the same flock. Egg yolk immunoglobulin (IgY) was extracted from the 4 groups of eggs as described in Example 1(c), and the final IgY pellets were solubilized in the original yolk volume of PBS without thimerosal. Importantly, thimerosal was excluded because it would have been toxic to the CHO cells used in the toxin neutralization assays described in section (d) below.

[0215] d) Assay of Toxin Neutralization Activity

[0216] The toxin neutralization activity of the IgY solutions prepared in section (c) above was determined using an assay system that was modified from published methods. [Ehrich et al., Infect. Immun. 28:1041-1043 (1992); and McGee et al. Microb. Path. 12:333-341 (1992).] As additional controls, affinity-purified goat anti-C. difficile toxin A (Tech Lab) and affinity-purified goat anti-C. difficile toxin B (Tech Lab) were also assayed for toxin neutralization activity.

[0217] The IgY solutions and goat antibodies were serially diluted using F 12 medium (GIBCO) which was supplemented with 2% FCS (GIBCO)(this solution will be referred to as “medium” for the remainder of this Example). The resulting antibody solutions were then mixed with a standardized concentration of either native C difficile toxin A (Tech Lab), or native C. difficile toxin B (Tech Lab), at the concentrations indicated below. Following incubation at 37° C. for 60 min., 100 μl volumes of the toxin +antibody mixtures were added to the wells of 96-well microtiter plates (Falcon Microtest III) which contained 2.5×104 Chinese Hamster Ovary (CHO) cells per well (the CHO cells were plated on the previous day to allow them to adhere to the plate wells). The final concentration of toxin, or dilution of antibody indicated below refers to the final test concentration of each reagent present in the respective microtiter plate wells. Toxin reference wells were prepared which contained CHO cells and toxin A or toxin B at the same concentration used for the toxin plus antibody mixtures (these wells contained no antibody). Separate control wells were also prepared which contained CHO cells and medium only. The assay plates were then incubated for 18-24 hrs. in a 37° C., humidified, 5% CO2 incubator. On the following day, the remaining adherent (viable) cells in the plate wells were stained using 0.2% crystal violet (Mallinckrodt) dissolved in 2% ethanol, for 10 min. Excess stain was then removed by rinsing with water, and the stained cells were solubilized by adding 100 μl of 1% SDS (dissolved in water) to each well. The absorbance of each well was then measured at 570 nm, and the percent cytotoxicity of each test sample or mixture was calculated using the following formula: 1$\%\mathrm{CHO}\mathrm{Cell}\mathrm{Cytotoxicity}=\left[1-\left(\frac{\mathrm{Abs}.\mathrm{Sample}}{\mathrm{Abs}.\mathrm{Control}}\right)\right]\times 100$

[0218] Unlike previous reports which quantitate results visually by counting cell rounding by microscopy, this Example utilized spectrophotometric methods to quantitate the C. difficile toxin bioassay. In order to determine the toxin A neutralizing activity of the CTA, CTAB, and pre-immune IgY preparations, as well as the affinity-purified goat antitoxin A control, dilutions of these antibodies were reacted against a 0.1 μg/ml concentration of native toxin A (this is the approx. 50% cytotoxic dose of toxin A in this assay system). The results are shown in FIG. 3.

[0219] Complete neutralization of toxin A occurred with the CTA IgY (antitoxin A, above) at dilutions of 1:80 and lower, while significant neutralization occurred out to the 1:320 dilution. The CTAB IgY (antitoxin A+toxin B, above) demonstrated complete neutralization at the 1:320-1:160 and lower dilutions, and significant neutralization occurred out to the 1:1280 dilution. The commercially available affinity-purified goat antitoxin A did not completely neutralize toxin A at any of the dilutions tested, but demonstrated significant neutralization out to a dilution of 1:1,280. The preimmune IgY did not show any toxin A neutralizing activity at any of the concentrations tested. These results demonstrate that IgY purified from eggs laid by hens immunized with toxin A alone, or simultaneously with toxin A and toxin B, is an effective toxin A antitoxin.

[0220] The toxin B neutralizing activity of the CTAB and pre-immune IgY preparations, and also the affinity-purified goat antitoxin B control was determined by reacting dilutions of these antibodies against a concentration of native toxin B of 0.1 ng/ml (approximately the 50% cytotoxic dose of toxin B in the assay system). The results are shown in FIG. 4.

[0221] Complete neutralization of toxin B occurred with the CTAB IgY (antitoxin A+toxin B, above) at the 1:40 and lower dilutions, and significant neutralization occurred out to the 1:320 dilution. The affinity-purified goat antitoxin B demonstrated complete neutralization at dilutions of 1:640 and lower, and significant neutralization occurred out to a dilution of 1:2,560. The preimmune IgY did not show any toxin B neutralizing activity at any of the concentrations tested. These results demonstrate that IgY purified from eggs laid by hens immunized simultaneously with toxin A and toxin B is an effective toxin B antitoxin.

[0222] In a separate study, the toxin B neutralizing activity of CTB, CTAB, and pre-immune IgY preparations was determined by reacting dilutions of these antibodies against a native toxin B concentration of 0.1 μg/ml (approximately 100% cytotoxic dose of toxin B in this assay system). The results are shown in FIG. 5.

[0223] Significant neutralization of toxin B occurred with the CTB IgY (antitoxin B, above) at dilutions of 1:80 and lower, while the CTAB IgY (antitoxin A+toxin B, above) was found to have significant neutralizing activity at dilutions of 1:40 and lower. The preimmune IgY did not show any toxin B neutralizing activity at any of the concentrations tested. These results demonstrate that IgY purified from eggs laid by hens immunized with toxin B alone, or simultaneously with toxin A and toxin B, is an effective toxin B antitoxin.

EXAMPLE 9

In vivo Protection of Golden Syrian Hamsters from C. difficile Disease by Avian Antitoxins Against C. difficile Toxins A And B

[0224] The most extensively used animal model to study C. difficile disease is the hamster. [Lyerly et al., Infect. Immun. 47:349-352 (1992).] Several other animal models for antibiotic-induced diarrhea exist, but none mimic the human form of the disease as closely as the hamster model. [R. Fekety, “Animal Models of Antibiotic-Induced Colitis,” in O. Zak and M. Sande (eds.), Experimental Models in Antimicrobial Chemotherapy, Vol. 2, pp.61-72, (1986).] In this model, the animals are first predisposed to the disease by the oral administration of an antibiotic, such as clindamycin, which alters the population of normally-occurring gastrointestinal flora (Fekety, at 61-72). Following the oral challenge of these animals with viable C. difficile organisms, the hamsters develop cecitis, and hemorrhage, ulceration, and inflammation are evident in the intestinal mucosa. [Lyerly et al., Infect. Immun. 47:349-352 (1985).] The animals become lethargic, develop severe diarrhea, and a high percentage of them die from the disease. [Lyerly et al., Infect. Immun. 47:349-352 (1985).] This model is therefore ideally suited for the evaluation of therapeutic agents designed for the treatment or prophylaxis of C. difficile disease.

[0225] The ability of the avian C. difficile antitoxins, described in Example 1 above, to protect hamsters from C. difficile disease was evaluated using the Golden Syrian hamster model of C. difficile infection. The Example involved (a) preparation of the avian C. difficile antitoxins, (b) in vivo protection of hamsters from C. difficile disease by treatment with avian antitoxins, and (c) long-term survival of treated hamsters.

[0226] a) Preparation of The Avian C. difficile Antitoxins

[0227] Eggs were collected from hens in groups CTA and CTAB described in Example 1(b) above. To be used as a pre-immune (negative) control, eggs were also purchased from a local supermarket. Egg yolk immunoglobulin (IgY) was extracted from the 3 groups of eggs as described in Example 1(c), and the final IgY pellets were solubilized in one fourth the original yolk volume of Ensure® nutritional formula.

[0228] b) In vivo Protection of Hamsters Against C. difficile Disease by Treatment with Avian Antitoxins

[0229] The avian C. difficile antitoxins prepared in section (a) above were evaluated for their ability to protect hamsters from C. difficile disease using an animal model system which was modified from published procedures. [Fekety, at 61-72; Borriello et al., J. Med. Microbiol., 24:53-64 (1987); Kim et al., Infect. Immun., 55:2984-2992 (1987); Borriello et al., J. Med. Microbiol., 25:191-196 (1988); Delmee and Avesani, J. Med. Microbiol., 33:85-90 (1990); and Lyerly et al., Infect. Immun., 59:2215-2218 (1991).] For the study, three separate experimental groups were used, with each group consisting of 7 female Golden Syrian hamsters (Charles River), approximately 10 weeks old and weighing approximately 100 gms. each. The three groups were designated “CTA,” “CTAB” and “Pre-immune.” These designations corresponded to the antitoxin preparations with which the animals in each group were treated. Each animal was housed in an individual cage, and was offered food and water ad libitum through the entire length of the study. On day 1, each animal was orally administered 1.0 ml of one of the three antitoxin preparations (prepared in section (a) above) at the following timepoints: 0 hrs., 4 hrs., and 8 hrs. On day 2, the day 1 treatment was repeated. On day 3, at the 0 hr. timepoint, each animal was again administered antitoxin, as described above. At 1 hr., each animal was orally administered 3.0 mg of clindamycin-HCl (Sigma) in 1 ml of water. This treatment predisposed the animals to infection with C. difficile. As a control for possible endogenous C. difficile colonization, an additional animal from the same shipment (untreated) was also administered 3.0 mg of clindamycin-HCl in the same manner. This clindamycin control animal was left untreated (and uninfected) for the remainder of the study. At the 4 hr. and 8 hr. timepoints, the animals were administered antitoxin as described above. On day 4, at the 0 hr. timepoint, each animal was again administered antitoxin as described above. At 1 hr., each animal was orally challenged with 1 ml of C. difficile inoculum, which contained approx. 100 C. difficile strain 43596 organisms in sterile saline. C. difficile strain 43596, which is a serogroup C strain, was chosen because it is representative of one of the most frequently-occurring serogroups isolated from patients with antibiotic-associated pseudomembranous colitis. [Delmee et al., J. Clin. Microbiol., 28:2210-2214 (1985).] In addition, this strain has been previously demonstrated to be virulent in the hamster model of infection. [Delmee and Avesani, J. Med. Microbiol., 33:85-90 (1990).] At the 4 hr. and 8 hr. timepoints, the animals were administered antitoxin as described above. On days 5 through 13, the animals were administered antitoxin 3× per day as described for day 1 above, and observed for the onset of diarrhea and death. On the morning of day 14, the final results of the study were tabulated. These results are shown in Table 13.

[0230] Representative animals from those that died in the Pre-Immune and CTA groups were necropsied. Viable C. difficile organisms were cultured from the ceca of these animals, and the gross pathology of the gastrointestinal tracts of these animals was consistent with that expected for C. difficile disease (inflamed, distended, hemorrhagic cecum, filled with watery diarrhea-like material). In addition, the

TABLE 13
Treatment Results
	No. Animals
Treatment Group	Surviving	No. Animals Dead
Pre-Immune	1	6
CTA (Antitoxin A only)	5	2
CTAB (Antitoxin A + Antitoxin B)	7	0

[0231] clindamycin control animal remained healthy throughout the entire study period, therefore indicating that the hamsters used in the study had not previously been colonized with endogenous C. difficile organisms prior to the start of the study. Following the final antitoxin treatment on day 13, a single surviving animal from the CTA group, and also from the CTAB group, was sacrificed and necropsied. No pathology was noted in either animal.

[0232] Treatment of hamsters with orally-administered toxin A and toxin B antitoxin (group CTAB) successfully protected 7 out of 7 (100%) of the animals from C. difficile disease. Treatment of hamsters with orally-administered toxin A antitoxin (group CTA) protected 5 out of 7 (71%) of these animals from C. difficile disease. Treatment using pre-immune IgY was not protective against C. difficile disease, as only 1 out of 7 (14%) of these animals survived. These results demonstrate that the avian toxin A antitoxin and the avian toxin A+toxin B antitoxin effectively protected the hamsters from C. difficile disease. These results also suggest that although the neutralization of toxin A alone confers some degree of protection against C. difficile disease, in order to achieve maximal protection, simultaneous antitoxin A and antitoxin B activity is necessary.

[0233] c) Long-Term Survival of Treated Hamsters

[0234] It has been previously reported in the literature that hamsters treated with orally-administered bovine antitoxin IgG concentrate are protected from C. difficile disease as long as the treatment is continued, but when the treatment is stopped, the animals develop diarrhea and subsequently die within 72 hrs. [Lyerly et a., Infect. Immun., 59(6):2215-2218 (1991).] In order to determine whether treatment of C. difficile disease using avian antitoxins promotes long-term survival following the discontinuation of treatment, the 4 surviving animals in group CTA, and the 6 surviving animals in group CTAB were observed for a period of 11 days (264 hrs.) following the discontinuation of antitoxin treatment described in section (b) above. All hamsters remained healthy through the entire post-treatment period. This result demonstrates that not only does treatment with avian antitoxin protect against the onset of C. difficile disease (i.e., it is effective as a prophylactic), it also promotes long-term survival beyond the treatment period, and thus provides a lasting cure.

EXAMPLE 10

In vivo Treatment of Established C. difficile Infection in Golden Syrian Hamsters with Avian Antitoxins Against C. difficile Toxins A And B

[0235] The ability of the avian C. difficile antitoxins, described in Example 8 above, to treat an established C. difficile infection was evaluated using the Golden Syrian hamster model. The Example involved (a) preparation of the avian C. difficile antitoxins, (b) in vivo treatment of hamsters with established C. difficile infection, and (c) histologic evaluation of cecal tissue.

[0236] a) Preparation of the Avian C. difficile Antitoxins

[0237] Eggs were collected from hens in group CTAB described in Example 8(b) above, which were immunized with C. difficile toxoids and native toxins A and B. Eggs purchased from a local supermarket were used as a pre-immune (negative) control. Egg yolk immunoglobulin (IgY) was extracted from the 2 groups of eggs as described in Example 1(c), and the final IgY pellets were solubilized in one-fourth the original yolk volume of Ensure® nutritional formula.

[0238] b) In vivo Treatment of Hamsters with Established

[0239] C. difficile Infection The avian C. difficile antitoxins prepared in section (a) above were evaluated for the ability to treat established C. difficile infection in hamsters using an animal model system which was modified from the procedure which was described for the hamster protection study in Example 8(b) above.

[0240] For the study, four separate experimental groups were used, with each group consisting of 7 female Golden Syrian hamsters (Charles River), approx. 10 weeks old, weighing approximately 100 gms. each. Each animal was housed separately, and was offered food and water ad libitum through the entire length of the study.

[0241] On day 1 of the study, the animals in all four groups were each predisposed to C. difficile infection by the oral administration of 3.0 mg of clindamycin-HCl (Sigma) in 1 ml of water.

[0242] On day 2, each animal in all four groups was orally challenged with 1 ml of C. difficile inoculum, which contained approximately 100 C. difficile strain 43596 organisms in sterile saline. C. difficile strain 43596 was chosen because it is representative of one of the most frequently-occurring serogroups isolated from patients with antibiotic-associated pseudomembranous colitis. [Delmee et al., J. Clin. Microbiol., 28:2210-2214 (1990).] In addition, as this was the same C. difficile strain used in all of the previous Examples above, it was again used in order to provide experimental continuity.

[0243] On day 3 of the study (24 hrs. post-infection), treatment was started for two of the four groups of animals. Each animal of one group was orally administered 1.0 ml of the CTAB IgY preparation (prepared in section (a) above) at the following timepoints: 0 hrs., 4 hrs., and 8 hrs. The animals in this group were designated “CTAB-24.” The animals in the second group were each orally administered 1.0 ml of the pre-immune IgY preparation (also prepared in section (a) above) at the same timepoints as for the CTAB group. These animals were designated “Pre-24.” Nothing was done to the remaining two groups of animals on day 3.

[0244] On day 4, 48 hrs. post-infection, the treatment described for day 3 above was repeated for the CTAB-24 and Pre-24 groups, and was initiated for the remaining two groups at the same timepoints. The final two groups of animals were designated “CTAB-48” and “Pre-48” respectively.

[0245] On days 5 through 9, the animals in all four groups were administered antitoxin or pre-immune IgY, 3× per day, as described for day 4 above. The four experimental groups are summarized in Table 14.

TABLE 14
Experimental Treatment Groups
Group Designation Experimental Treatment
CTAB-24           Infected, treatment w/antitoxin IgY started @ 24
                  hrs. post-infection.
Pre-24            Infected, treatment w/pre-immune IgY started @ 24
                  hrs. post-infection.
CTAB-48           Infected, treatment w/antitoxin IgY started @ 48
                  hrs. post-infection.
Pre-48            Infected, treatment w/pre-immune IgY started @ 48
                  hrs. post-infection.

[0246] All animals were observed for the onset of diarrhea and death through the conclusion of the study on the morning of day 10. The results of this study are displayed in Table 15.

[0247] Eighty-six percent of the animals which began receiving treatment with antitoxin IgY at 24 hrs. post-infection (CTAB-24 above) survived, while 57% of the animals treated with antitoxin IgY starting 48 hrs. post-infection (CTAB-48 above) survived. In contrast, none of the animals receiving pre-immune IgY starting 24 hrs. post-infection (Pre-24 above) survived, and only 29% of the animals which began receiving treatment with pre-immune IgY at 48 hrs. post-infection (Pre-48 above) survived through the conclusion of the study. These results demonstrate that avian antitoxins raised against C. difficile toxins A and B are capable of successfully treating established C. difficile infections in vivo.

TABLE 15
Experimental Outcome-Day 10
Treatment Group	No. Animals Surviving	No. Animals Dead
CTAB-24	6	1
Pre-24	0	7
CTAB-48	4	3
Pre-48	2	5

[0248] c) Histologic Evaluation of Cecal Tissue

[0249] In order to further evaluate the ability of the IgY preparations tested in this study to treat established C. difficile infection, histologic evaluations were performed on cecal tissue specimens obtained from representative animals from the study described in section (b) above.

[0250] Immediately following death, cecal tissue specimens were removed from animals which died in the Pre-24 and Pre-48 groups. Following the completion of the study, a representative surviving animal was sacrificed and cecal tissue specimens were removed from the CTAB-24 and CTAB-48 groups. A single untreated animal from the same shipment as those used in the study was also sacrificed and a cecal tissue specimen was removed as a normal control. All tissue specimens were fixed overnight at 4° C. in 10% buffered formalin. The fixed tissues were paraffin-embedded, sectioned, and mounted on glass microscope slides. The tissue sections were then stained using hematoxylin and eosin (H and E stain), and were examined by light microscopy.

[0251] Upon examination, the tissues obtained from the CTAB-24 and CTAB-48 animals showed no pathology, and were indistinguishable from the normal control. This observation provides further evidence for the ability of avian antitoxins raised against C. difficile toxins A and B to effectively treat established C. difficile infection, and to prevent the pathologic consequences which normally occur as a result of C. difficile disease.

[0252] In contrast, characteristic substantial mucosal damage and destruction was observed in the tissues of the animals from the Pre-24 and Pre-48 groups which died from C. difficile disease. Normal tissue architecture was obliterated in these two preparations, as most of the mucosal layer was observed to have sloughed away, and there were numerous large hemorrhagic areas containing massive numbers of erythrocytes.

EXAMPLE 11

Cloning and Expression of C. difficile Toxin A Fragments

[0253] The toxin A gene has been cloned and sequenced, and shown to encode a protein of predicted MW of 308 kd. [Dove et al., Infect. Immun., 58:480-488 (1990).] Given the expense and difficulty of isolating native toxin A protein, it would be advantageous to use simple and inexpensive procaryotic expression systems to produce and purify high levels of recombinant toxin A protein for immunization purposes. Ideally, the isolated recombinant protein would be soluble in order to preserve native antigenicity, since solubilized inclusion body proteins often do not fold into native conformations. To allow ease of purification, the recombinant protein should be expressed to levels greater than 1 mg/liter of E. coli culture.

[0254] To determine whether high levels of recombinant toxin A protein can be produced in E. coli, fragments of the toxin A gene were cloned into various prokaryotic expression vectors, and assessed for the ability to express recombinant toxin A protein in E. coli. Three prokaryotic expression systems were utilized. These systems were chosen because they drive expression of either fusion (pMALc and pGEX2T) or native (pET23a-c) protein to high levels in E. coli, and allow affinity purification of the expressed protein on a ligand containing column. Fusion proteins expressed from pGEX vectors bind glutathione agarose beads, and are eluted with reduced glutathione. pMAL fusion proteins bind amylose resin, and are eluted with maltose. A polyhistidine tag is present at either the N-terminal (pET16b) or C-terminal (pET23a-c) end of pET fusion proteins. This sequence specifically binds Ni2+ chelate columns, and is eluted with imidazole salts. Extensive descriptions of these vectors are available (Williams et al. (1994) DNA Cloning: Expression Systems, in press), and will not be discussed in detail here. The Example involved (a) cloning of the toxin A gene, (b) expression of large fragments of toxin A in various prokaryotic expression systems, (c) identification of smaller toxin A gene fragments that express efficiently in E. coli, (d) purification of recombinant toxin A protein by affinity chromatography, and (e) demonstration of functional activity of a recombinant fragment of the toxin A gene.

[0255] a) Cloning of the Toxin A Gene

[0256] A restriction map of the toxin A gene is shown in FIG. 6. The encoded protein contains a carboxy terminal ligand binding region, containing multiple repeats of a carbohydrate binding domain. [von Eichel-Streiber and Sauerborn, Gene 96:107-113 (1990).] The toxin A gene was cloned in three pieces, by using either the polymerase chain reaction (PCR) to amplify specific regions, (regions 1 and 2, FIG. 6) or by screening a constructed genomic library for a specific toxin A gene fragment (region 3, FIG. 6). The sequences of the utilized PCR primers are P1: 5′ GGAAATT TAGCTGCAGCATCTGAC 3′ (SEQ ID NO.:1); P2: 5′ TCTAGCAAATTCGCTTGT GTTGAA 3′ (SEQ ID NO.:2); P3: 5′ CTCGCATATAGCATTAGACC 3′ (SEQ ID NO.:3); and P4: 5′ CTATCTAGGCCTAAAGTAT 3′ (SEQ ID NO.:4). These regions were cloned into prokaryotic expression vectors that express either fusion (pMALc and pGEX2T) or native (pET23a-c) protein to high levels in E. coli, and allow affinity purification of the expressed protein on a ligand containing column.

[0257] Clostridium difficile VPI strain 10463 was obtained from the ATCC (ATCC #43255) and grown under anaerobic conditions in brain-heart infusion medium (BBL). High molecular-weight C. difficile DNA was isolated essentially as described by Wren and Tabaqchali (1987) J. Clin. Microbiol., 25:2402, except proteinase K and sodium dodecyl sulfate (SDS) was used to disrupt the bacteria, and cetyltrimethylammnonium bromide precipitation [as described in Ausubel et al., Current Protocols in Molecular Biology (1989)] was used to remove carbohydrates from the cleared lysate. The integrity and yield of genomic DNA was assessed by comparison with a serial dilution of uncut lambda DNA after electrophoresis on an agarose gel.

[0258] Fragments 1 and 2 were cloned by PCR, utilizing a proofreading thermostable DNA polymerase (native pfu polymerase; Stratagene). The high fidelity of this polymerase reduces the mutation problems associated with amplification by error prone polymerases (e.g., Taq polymerase). PCR amplification was performed using the indicated PCR primers (FIG. 6) in 50 μl reactions containing 10 mM Tris-HCl(8.3), 50 mM KCl, 1.5 mM MgCl2, 200 μM each dNTP, 0.2 μM each primer, and 50 ng C. difficile genomic DNA. Reactions were overlaid with 100 μl mineral oil, heated to 94° C. for 4 min 0.5 μl native pfu polymerase (Stratagene) added, and the reaction cycled 30× at 94° C. for 1 min, 50° C. for 1 min, 72° C. for 4 min, followed by 10 min at 72° C. Duplicate reactions were pooled, chloroform extracted, and ethanol precipitated. After washing in 70% ethanol, the pellets were resuspended in 50 μl TE buffer [10 mM Tris-HCL, 1 mM EDTA pH 8.0]. Aliquots of 10 μl each were restriction digested with either EcoRI/HincII (fragment 1) or EcoRI/PstI (fragment 2), and the appropriate restriction fragments were gel purified using the Prep-A-Gene kit (BioRad), and ligated to either EcoRI/SmaI-restricted pGEX2T (Pharmacia) vector (fragment 1), or the EcoRI/PstI pMAlc (New England Biolabs) vector (fragment 2). Both clones are predicted to produce in-frame fusions with either the glutathione-S-transferase protein (pGEX vector) or the maltose binding protein (pMAL vector). Recombinant clones were isolated, and confirmed by restriction digestion, using standard recombinant molecular biology techniques. [Sambrook et al., Molecular Cloning, A Laboratory Manual (1989), and designated pGA30-660 and pMA660-1100, respectively (see FIG. 6 for description of the clone designations).]

[0259] Fragment 3 was cloned from a genomic library of size selected PstI digested C. difficile genomic DNA, using standard molecular biology techniques (Sambrook et al.). Given that the fragment 3 internal PstI site is protected from cleavage in C. difficile genomic DNA [Price et al., Curr. Microbiol., 16:55-60 (1987)], a 4.7 kb fragment from PstI restricted C. difficile genomic DNA was gel purified, and ligated to PstI restricted, phosphatase treated pUC9 DNA. The resulting genomic library was screened with a oligonucleotide primer specific to fragment 3, and multiple independent clones were isolated. The presence of fragment 3 in several of these clones was confirmed by restriction digestion, and a clone of the indicated orientation (FIG. 6) was restricted with BamHI/HindIII, the released fragment purified by gel electrophoresis, and ligated into similarly restricted pET23c expression vector DNA (Novagen). Recombinant clones were isolated, and confirmed by restriction digestion. This construct is predicted to create both a predicted in frame fusion with the pET protein leader sequence, as well as a predicted C-terminal poly-histidine affinity tag, and is designated pPA1100-2680 (see FIG. 6 for the clone designation).

[0260] b) Expression of Large Fragments of Toxin A In E. coli

[0261] Protein expression from the three expression constructs made in (a) was induced, and analyzed by Western blot analysis with an affinity purified, goat polyclonal antiserum directed against the toxin A toxoid (Tech Lab). The procedures utilized for protein induction, SDS-PAGE, and Western blot analysis are described in detail in Williams et al (1994), supra. In brief, 5 ml 2XYT (16 g tryptone, 10 g yeast extract, 5 g NaCl per liter, pH 7.5+100 μg/ml ampicillin were added to cultures of bacteria (BL21 for pMAl and pGEX plasmids, and BL21(DE3)LysS for pET plasmids) containing the appropriate recombinant clone which were induced to express recombinant protein by addition of IPTG to 1 mM. Cultures were grown at 37° C., and induced when the cell density reached 0.5 OD600. Induced protein was allowed to accumulate for two hrs after induction. Protein samples were prepared by pelleting 1 ml aliquots of bacteria by centrifugation (1 min in a microfuge), and resuspension of the pelleted bacteria in 150 μl of 2× SDS-PAGE sample buffer [Williams et al. (1994), supra]. The samples were heated to 95° C. for 5 min, the cooled and 5 or 10 μl aliquots loaded on 7.5% SDS-PAGE gels. BioRad high molecular weight protein markers were also loaded, to allow estimation of the MW of identified fusion proteins. After electrophoresis, protein was detected either generally by staining gels with Coomassie blue, or specifically, by blotting to nitrocellulose for Western blot detection of specific immunoreactive protein. Western blots, (performed as described in Example 3) which detect toxin A reactive protein in cell lysates of induced protein from the three expression constructs are shown in FIG. 7. In this figure, lanes 1-3 contain cell lysates prepared from E. coli strains containing pPA1100-2860 in B121(DE3)lysE cells; lanes 4-6 contain cell lysates prepared from E. coli strains containing pPA1100-2860 in B121(DE3)lysS cells; lanes 7-9 contain cell lysates prepared from E. coli strains containing pMA30-660; lanes 10-12 contain cell lysates prepared from E. coli strains containing pMA660-1 100. The lanes were probed with an affinity purified goat antitoxin A polyclonal antibody (Tech Lab). Control lysates from uninduced cells (lanes 1, 7, and 10) contain very little immunoreactive material compared to the induced samples in the remaining lanes. The highest molecular weight band observed for each clone is consistent with the predicted size of the full length fusion protein.

[0262] Each construct directs expression of high molecular weight (HMW) protein that is reactive with the toxin A antibody. The size of the largest immunoreactive bands from each sample is consistent with predictions of the estimated MW of the intact fusion proteins. This demonstrates that the three fusions are in-frame, and that none of the clones contain cloning artifacts that disrupt the integrity of the encoded fusion protein. However, the Western blot demonstrates that fusion protein from the two larger constructs (pGA30-660 and pPA1100-2680) are highly degraded. Also, expression levels of toxin A proteins from these two constructs are low, since induced protein bands are not visible by Coomassie staining (not shown). Several other expression constructs that fuse large sub-regions of the toxin A gene to either pMALc or pET23a-c expression vectors, were constructed and tested for protein induction. These constructs were made by mixing gel purified restriction fragments, derived from the expression constructs shown in FIG. 6, with appropriately cleaved expression vectors, ligating, and selecting recombinant clones in which the toxin A restriction fragments had ligated together and into the expression vector as predicted for in-frame fusions. The expressed toxin A interval within these constructs are shown in FIG. 8, as well as the internal restriction sites utilized to make these constructs.

[0263] As used herein, the term “interval” refers to any portion (i.e., any segment of the toxin which is less than the whole toxin molecule) of a clostridial toxin. In a preferred embodiment, “interval” refers to portions of C. difficile toxins such as toxin A or toxin B. It is also contemplated that these intervals will correspond to epitopes of immunologic importance, such as antigens or immunogens against which a neutralizing antibody response is effected. It is not intended that the present invention be limited to the particular intervals or sequences described in these Examples. It is also contemplated that sub-portions of intervals (e.g., an epitope contained within one interval or which bridges multiple intervals) be used as compositions and in the methods of the present invention.

[0264] In all cases, Western blot analysis of each of these constructs with goat antitoxin A antibody (Tech Lab) detected HMW fusion protein of the predicted size (not shown). This confirms that the reading frame of each of these clones is not prematurely terminated, and is fused in the correct frame with the fusion partner. However, the Western blot analysis revealed that in all cases, the induced protein is highly degraded, and, as assessed by the absence of identifiable induced protein bands by Coomassie Blue staining, are expressed only at low levels. These results suggest that expression of high levels of intact toxin A recombinant protein is not possible when large regions of the toxin A gene are expressed in E. coli using these expression vectors.

[0265] c) High Level Expression of Small Toxin A Protein Fusions in E. coli

[0266] Experience indicates that expression difficulties are often encountered when large (greater than 100 kd) fragments are expressed in E. coli. A number of expression constructs containing smaller fragments of the toxin A gene were constructed, to determine if small regions of the gene can be expressed to high levels without extensive protein degradation. A summary of these expression constructs are shown in FIG. 9. All were constructed by in-frame fusions of convenient toxin A restriction fragments to either the pMALc or pET23a-c vectors. Protein preparations from induced cultures of each of these constructs were analyzed by both Coomassie Blue staining and Western analysis as in (b) above. In all cases, higher levels of intact, full length fusion proteins were observed than with the larger recombinants from section (b).

[0267] d) Purification of Recombinant Toxin A Protein

[0268] Large scale (500 ml) cultures of each recombinant from (c) were grown, induced, and soluble and insoluble protein fractions were isolated. The soluble protein extracts were affinity chromatographed to isolate recombinant fusion protein, as described [Williams et al. (1994), supra]. In brief, extracts containing tagged pET fusions were chromatographed on a nickel chelate column, and eluted using imidazole salts as described by the distributor (Novagen). Extracts containing soluble pMAL fusion protein were prepared and chromatogrammed in column buffer (10 mM NaPO4, 0.5M NaCl, 10 mM β-mercaptoethanol, pH 7.2) over an amylose resin column (New England Biolabs), and eluted with column buffer containing 10 mM maltose as described [Williams et al. (1994), supra]. When the expressed protein was found to be predominantly insoluble, insoluble protein extracts were prepared by the method described in Example 17, infra. The results are summarized in Table 16. FIG. 10 shows the sample purifications of recombinant toxin A protein. In this figure, lanes 1 and 2 contain MBP fusion protein purified by affinity purification of soluble protein.

TABLE 16
Purification Of Recombinant Toxin A Protein
		Yield Affinity	% Intact	Yield Intact
		Purified Soluble	Soluble	Insoluble
Clone(a)	Protein Solubility	Protein(b)	Fusion Protein(c)	Fusion Protein
pMA30-270	Soluble	4 mg/500 mls	10%	NA
PMA30-300	Soluble	4 mg/500 mls	5-10%	NA
pMA300-660	Insoluble	—	NA	10 mg/500 ml
pMA660-1100	Soluble	4.5 mg/500 mls	50%	NA
pMA1100-1610	Soluble	18 mg/500 mls	10%	NA
pMA1610-1870	Both	22 mg/500 mls	90%	20 mg/500 ml
pMA1450-1870	Insoluble	—	NA	0.2 mg/500 ml
pPA1100-1450	Soluble	0.1 mg/500 mls	90%	NA
pPA1100-1870	Soluble	0.02 mg/500 mls	90%	NA
pMA1870-2680	Both	12 mg/500 mls	80%	NA
pPa1870-2680	Insoluble	—	NA	10 mg/500 ml
(a)pP = pET23 vector, pM = pMALc vector, A = toxin A.
(b)Based on 1.5 OD280 = 1 mg/ml (extinction coefficient of MBP).
(c)Estimated by Coomassie staining of SDS-PAGE gels.

[0269] Lanes 3 and 4 contain MBP fusion protein purified by solubilization of insoluble inclusion bodies. The purified fusion protein samples are pMA1870-2680 (lane 1), pMA660-1100 (lane 2), pMA300-600 (lane 3) and pMA1450-1870 (lane 4).

[0270] Poor yields of affinity purified protein were obtained when poly-histidine tagged pET vectors were used to drive expression (pPA1100-1450, pP1100-1870). However, significant protein yields were obtained from pMAL expression constructs spanning the entire toxin A gene, and yields of full-length soluble fusion protein ranged from an estimated 200-400 μg/500 ml culture (pMA30-300) to greater than 20 mg/500 ml culture (pMA1610-1870). Only one interval was expressed to high levels as strictly insoluble protein (pMA300-660). Thus, although high level expression was not observed when using large expression constructs from the toxin A gene, usable levels of recombinant protein spanning the entire toxin A gene were obtainable by isolating induced protein from a series of smaller pMAL expression constructs that span the entire toxin A gene. This is the first demonstration of the feasibility of expressing recombinant toxin A protein to high levels in E. coli.

[0271] e) Hemagglutination Assay Using the Toxin A Recombinant Proteins

[0272] The carboxy terminal end consisting of the repeating units contains the hemagglutination activity or binding domain of C. difficile toxin A. To determine whether the expressed toxin A recombinants retain functional activity, hemagglutination assays were performed. Two toxin A recombinant proteins, one containing the binding domain as either soluble affinity purified protein (pMA1870-2680) or SDS solubilized inclusion body protein (pPA1870-2680) and soluble protein from one region outside that domain (pMA1100-1610) were tested using a described procedure. [H. C. Krivan et. al., Infect. Immun., 53:573 (1986).] Citrated rabbit red blood cells (RRBC)(Cocalico) were washed several times with Tris-buffer (0.1M Tris and 50 mM NaCl ) by centrifugation at 450× g for 10 minutes at 4° C. A 1% RRBC suspension was made from the packed cells and resuspended in Tris-buffer. Dilutions of the recombinant proteins and native toxin A (Tech Labs) were made in the Tris-buffer and added in duplicate to a round-bottomed 96-well microtiter plate in a final volume of 100 μl. To each well, 50 μl of the 1% RRBC suspension was added, mixed by gentle tapping, and incubated at 4° C. for 3-4 hours. Significant hemagglutination occurred only in the recombinant proteins containing the binding domain (pMA 1870-2680) and native toxin A. The recombinant protein outside the binding domain (pMA1100-1610) displayed no hemagglutination activity. Using equivalent protein concentrations, the hemagglutination titer for toxin A was 1:256, while titers for the soluble and insoluble recombinant proteins of the binding domain were 1:256 and about 1:5000. Clearly, the recombinant proteins tested retained functional activity and were able to bind RRBC's.

EXAMPLE 12

Functional Activity of IgY Reactive Against Toxin A Recombinants

[0273] The expression of recombinant toxin A protein as multiple fragments in E. coli has demonstrated the feasibility of generating toxin A antigen through use of recombinant methodologies (Example 11). The isolation of these recombinant proteins allows the immunoreactivity of each individual subregion of the toxin A protein to be determined (i.e., in a antibody pool directed against the native toxin A protein). This identifies the regions (if any) for which little or no antibody response is elicited when the whole protein is used as a immunogen. Antibodies directed against specific fragments of the toxin A protein can be purified by affinity chromatography against recombinant toxin A protein, and tested for neutralization ability. This identifies any toxin A subregions that are essential for producing neutralizing antibodies. Comparison with the levels of immune response directed against these intervals when native toxin is used as an immunogen predicts whether potentially higher titers of neutralizing antibodies can be produced by using recombinant protein directed against a individual region, rather than the entire protein. Finally, since it is unknown whether antibodies reactive to the recombinant toxin A proteins produced in Example 11 neutralize toxin A as effectively as antibodies raised against native toxin A (Examples 9 and 10), the protective ability of a pool of antibodies affinity purified against recombinant toxin A fragments was assessed for its ability to neutralize toxin A.

[0274] This Example involved (a) epitope mapping of the toxin A protein to determine the titre of specific antibodies directed against individual subregions of the toxin A protein when native toxin A protein is used as an immunogen, (b) affinity purification of IgY reactive against recombinant proteins spanning the toxin A gene, (c) toxin A neutralization assays with affinity purified IgY reactive to recombinant toxin A protein to identify subregions of the toxin A protein that induce the production of neutralizing antibodies, and determination of whether complete neutralization of toxin A can be elicited with a mixture of antibodies reactive to recombinant toxin A protein.

[0275] a) Epitope Mapping of the Toxin A Gene

[0276] The affinity purification of recombinant toxin A protein specific to defined intervals of the toxin A protein allows epitope mapping of antibody pools directed against native toxin A. This has not previously been possible, since previous expression of toxin A recombinants has been assessed only by Western blot analysis, without knowledge of the expression levels of the protein [e.g., von Eichel-Streiber et al, J. Gen. Microbiol., 135:55-64 (1989)]. Thus, high or low reactivity of recombinant toxin A protein on Western blots may reflect protein expression level differences, not immunoreactivity differences. Given that the purified recombinant protein generated in Example 11 have been quantitated, the issue of relative immunoreactivity of individual regions of the toxin A protein was precisely addressed.

[0277] For the purposes of this Example, the toxin A protein was subdivided into 6 intervals (1-6), numbered from the amino (interval 1) to the carboxyl (interval 6) termini.

[0278] The recombinant proteins corresponding to these intervals were from expression clones (see Example 11(d) for clone designations) pMA30-300 (interval 1), pMA300-660 (interval 2), pMA660-1100 (interval 3), pPA1100-1450 (interval 4), pMA1450-1870 (interval 5) and pMA1870-2680 (interval 6). These 6 clones were selected because they span the entire protein from amino acids numbered 30 through 2680, and subdivide the protein into 6 small intervals. Also, the carbohydrate binding repeat interval is contained specifically in one interval (interval 6), allowing evaluation of the immune response specifically directed against this region. Western blots of 7.5% SDS-PAGE gels, loaded and electrophoresed with defined quantities of each recombinant protein, were probed with either goat antitoxin A polyclonal antibody (Tech Lab) or chicken antitoxin A polyclonal antibody [pCTA IgY, Example 8(c)]. The blots were prepared and developed with alkaline phosphatase as previously described [Williams et al. (1994), supra]. At least 90% of all reactivity, in either goat or chicken antibody pools, was found to be directed against the ligand binding domain (interval 6). The remaining immunoreactivity was directed against all five remaining intervals, and was similar in both antibody pools, except that the chicken antibody showed a much lower reactivity against interval 2 than the goat antibody.

[0279] This clearly demonstrates that when native toxin A is used as an immunogen in goats or chickens, the bulk of the immune response is directed against the ligand binding domain of the protein, with the remaining response distributed throughout the remaining ⅔ of the protein.

[0280] b) Affinity Purification of IgY Reactive Against Recombinant Toxin A Protein

[0281] Affinity columns, containing recombinant toxin A protein from the 6 defined intervals in (a) above, were made and used to (i) affinity purify antibodies reactive to each individual interval from the CTA IgY preparation [Example 8(c)], and (ii) deplete interval specific antibodies from the CTA IgY preparation. Affinity columns were made by coupling 1 ml of PBS-washed Actigel resin (Sterogene) with region specific protein and {fraction (1/10)} final volume of Ald-coupling solution (1M sodium cyanoborohydride). The total region specific protein added to each reaction mixture was 2.7 mg (interval 1), 3 mg (intervals 2 and 3), 0.1 mg (interval 4), 0.2 mg (interval 5) and 4 mg (interval 6). Protein for intervals 1, 3, and 6 was affinity purified pMAl fusion protein in column buffer (see Example 11). Interval 4 was affinity purified poly-histidine containing pET fusion in PBS; intervals 2 and 5 were from inclusion body preparations of insoluble pMAL fusion protein, dialysed extensively in PBS. Aliquots of the supernatants from the coupling reactions, before and after coupling, were assessed by Coomassie staining of 7.5% SDS-PAGE gels. Based on protein band intensities, in all cases greater than 50% coupling efficiencies were estimated. The resins were poured into 5 ml BioRad columns, washed extensively with PBS, and stored at 4° C.

[0282] Aliquots of the CTA IgY polyclonal antibody preparation were depleted for each individual region as described below. A 20 ml sample of the CTA IgY preparation [Example 8(c)] was dialysed extensively against 3 changes of PBS (1 liter for each dialysis), quantitated by absorbance at OD280, and stored at 4° C. Six 1 ml aliquots of the dialysed IgY preparation were removed, and depleted individually for each of the six intervals. Each 1 ml aliquot was passed over the appropriate affinity column, and the eluate twice reapplied to the column. The eluate was collected, and pooled with a 1 ml PBS wash. Bound antibody was eluted from the column by washing with 5 column volumes of 4 M Guanidine-HCl (in 10 mM Tris-HCl, pH 8.0). The column was reequilibrated in PBS, and the depleted antibody stock reapplied as described above. The eluate was collected, pooled with a 1 ml PBS wash, quantitated by absorbance at OD280, and stored at 4° C. In this manner, 6 aliquots of the CTA IgY preparation were individually depleted for each of the 6 toxin A intervals, by two rounds of affinity depletion. The specificity of each depleted stock was tested by Western blot analysis. Multiple 7.5% SDS-PAGE gels were loaded with protein samples corresponding to all 6 toxin A subregions. After electrophoresis, the gels were blotted, and protein transfer confirmed by Ponceau S staining [protocols described in Williams et al. (1994), supra]. After blocking the blots 1 hr at 20° C. in PBS+ 0.1% Tween 20 (PBST) containing 5% milk (as a blocking buffer), 4 ml of either a 1/500 dilution of the dialysed CTA IgY preparation in blocking buffer, or an equivalent amount of the six depleted antibody stocks (using OD280 to standardize antibody concentration) were added and the blots incubated a further 1 hr at room temperature. The blots were washed and developed with alkaline phosphatase (using a rabbit anti-chicken alkaline phosphate conjugate as a secondary antibody) as previously described [Williams et al. (1994), supra]. In all cases, only the target interval was depleted for antibody reactivity, and at least 90% of the reactivity to the target intervals was specifically depleted.

[0283] Region specific antibody pools were isolated by affinity chromatography as described below. Ten mls of the dialysed CTA IgY preparation were applied sequentially to each affinity column, such that a single 10 ml aliquot was used to isolate region specific antibodies specific to each of the six subregions. The columns were sequentially washed with 10 volumes of PBS, 6 volumes of BBS-Tween, 10 volumes of TBS, and eluted with 4 ml Actisep elution media (Sterogene). The eluate was dialysed extensively against several changes of PBS, and the affinity purified antibody collected and stored at 4° C. The volumes of the eluate increased to greater than 10 mls during dialysis in each case, due to the high viscosity of the Actisep elution media. Aliquots of each sample were 20× concentrated using Centricon 30 microconcentrators (Amicon) and stored at 4° C. The specificity of each region specific antibody pool was tested, relative to the dialysed CTA IgY preparation, by Western blot analysis, exactly as described above, except that 4 ml samples of blocking buffer containing 100 μl region specific antibody (unconcentrated) were used instead of the depleted CTA IgY preparations. Each affinity purified antibody preparation was specific to the defined interval, except that samples purified against intervals 1-5 also reacted with interval 6. This may be due to non-specific binding to the interval 6 protein, since this protein contains the repetitive ligand binding domain which has been shown to bind antibodies nonspecifically. [Lyerly et al., Curr. Microbiol., 19:303-306 (1989).]

[0284] The reactivity of each affinity purified antibody preparation to the corresponding proteins was approximately the same as the reactivity of the 1/500 diluted dialysed CTA IgY preparation standard. Given that the specific antibody stocks were diluted 1/40, this would indicate that the unconcentrated affinity purified antibody stocks contain 1/10-1/20 the concentration of specific antibodies relative to the starting CTA IgY preparation.

[0285] c) Toxin A Neutralization Assay Using Antibodies Reactive Toward Recombinant Toxin A Protein

[0286] The CHO toxin neutralization assay [Example 8(d)] was used to assess the ability of the depleted or enriched samples generated in (b) above to neutralize the cytotoxicity of toxin A. The general ability of affinity purified antibodies to neutralize toxin A was assessed by mixing together aliquots of all 6 concentrated stocks of the 6 affinity purified samples generated in (b) above, and testing the ability of this mixture to neutralize a toxin A concentration of 0.1 μg/ml. The results, shown in FIG. 11, demonstrate almost complete neutralization of toxin A using the affinity purified (AP) mix. Some epitopes within the recombinant proteins utilized for affinity purification were probably lost when the proteins were denatured before affinity purification [by Guanidine-HCl treatment in (b) above]. Thus, the neutralization ability of antibodies directed against recombinant protein is probably underestimated using these affinity purified antibody pools. This experiment demonstrates that antibodies reactive to recombinant toxin A can neutralize cytotoxicity, suggesting that neutralizing antibodies may be generated by using recombinant toxin A protein as immunogen.

[0287] In view of the observation that the recombinant expression clones of the toxin A gene divide the protein into 6 subregions, the neutralizing ability of antibodies directed against each individual region was assessed. The neutralizing ability of antibodies directed against the ligand binding domain of toxin A was determined first.

[0288] In the toxin neutralization experiment shown in FIG. 11, interval 6 specific antibodies (interval 6 contains the ligand binding domain) were depleted from the dialysed PEG preparation, and the effect on toxin neutralization assayed. Interval 6 antibodies were depleted either by utilizing the interval 6 depleted CTA IgY preparation from (b) above (“-6 aff. depleted” in FIG. 11), or by addition of interval 6 protein to the CTA IgY preparation (estimated to be a 10 fold molar excess over anti-interval 6 immunoglobulin present in this preparation) to competitively compete for interval 6 protein (“-6 prot depleted” in FIG. 11). In both instances, removal of interval 6 specific antibodies reduces the neutralization efficiency relative to the starting CTA IgY preparation. This demonstrates that antibodies directed against interval 6 contribute to toxin neutralization. Since interval 6 corresponds to the ligand binding domain of the protein, these results demonstrate that antibodies directed against this region in the PEG preparation contribute to the neutralization of toxin A in this assay. However, it is significant that after removal of these antibodies, the PEG preparation retains significant ability to neutralize toxin A (FIG. 11). This neutralization is probably due to the action of antibodies specific to other regions of the toxin A protein, since at least 90% of the ligand binding region reactive antibodies were removed in the depleted sample prepared in (b) above. This conclusion was supported by comparison of the toxin neutralization of the affinity purified (AP) mix compared to affinity purified interval 6 antibody alone. Although some neutralization ability was observed with AP interval 6 antibodies alone, the neutralization was significantly less than that observed with the mixture of all 6 AP antibody stocks (not shown).

[0289] Given that the mix of all six affinity purified samples almost completely neutralized the cytotoxicity of toxin A (FIG. 11), the relative importance of antibodies directed against toxin A intervals 1-5 within the mixture was determined. This was assessed in two ways. First, samples containing affinity purified antibodies representing 5 of the 6 intervals were prepared, such that each individual region was depleted from one sample. FIG. 12 demonstrates a sample neutralization curve, comparing the neutralization ability of affinity purified antibody mixes without interval 4 (−4) or 5 (−5) specific antibodies, relative to the mix of all 6 affinity purified antibody stocks (positive control). While the removal of interval 5 specific antibodies had no effect on toxin neutralization (or intervals 1-3, not shown), the loss of interval 4 specific antibodies significantly reduced toxin neutralization (FIG. 12).

[0290] Similar results were seen in a second experiment, in which affinity purified antibodies, directed against a single region, were added to interval 6 specific antibodies, and the effects on toxin neutralization assessed. Only interval 4 specific antibodies significantly enhanced neutralization when added to interval 6 specific antibodies (FIG. 13). These results demonstrate that antibodies directed against interval 4 (corresponding to clone pPA1100-1450 in FIG. 9) are important for neutralization of cytotoxicity in this assay. Epitope mapping has shown that only low levels of antibodies reactive to this region are generated when native toxin A is used as an immunogen [Example 12(a)]. It is hypothesized that immunization with recombinant protein specific to this interval will elicit higher titers of neutralizing antibodies. In summary, this analysis has identified two critical regions of the toxin A protein against which neutralizing antibodies are produced, as assayed by the CHO neutralization assay.

EXAMPLE 13

Production and Evaluation of Avian Antitoxin Against C. difficile Recombinant Toxin A Polypeptide

[0291] In Example 12, we demonstrated neutralization of toxin A mediated cytotoxicity by affinity purified antibodies reactive to recombinant toxin A protein. To determine whether antibodies raised against a recombinant polypeptide fragment of C. difficile toxin A may be effective in treating clostridial diseases, antibodies to recombinant toxin A protein representing the binding domain were generated. Two toxin A binding domain recombinant polypeptides, expressing the binding domain in either the pMALc (pMA1870-2680) or pET 23(pPA1870-2680) vector, were used as immunogens. The pMAL protein was affinity purified as a soluble product [Example 12(d)] and the pET protein was isolated as insoluble inclusion bodies [Example 12(d)] and solubilized to an immunologically active protein using a proprietary method described in a pending patent application (U.S. patent application Ser. No. 08/129,027). This Example involves (a) immunization, (b) antitoxin collection, (c) determination of antitoxin antibody titer, (d) anti-recombinant toxin A neutralization of toxin A hemagglutination activity in vitro, and (e) assay of in vitro toxin A neutralizing activity.

[0292] a) Immunization

[0293] The soluble and the inclusion body preparations each were used separately to immunize hens. Both purified toxin A polypeptides were diluted in PBS and emulsified with approximately equal volumes of CFA for the initial immunization or IFA for subsequent booster immunizations. On day zero, for each of the recombinant preparations, two egg laying white Leghorn hens (obtained from local breeder) were each injected at multiple sites (intramuscular and subcutaneous) with 1 ml of recombinant adjuvant mixture containing approximately 0.5 to 1.5 mgs of recombinant toxin A. Booster immunizations of 1.0 mg were given on days 14 and day 28.

[0294] b) Antitoxin Collection

[0295] Total yolk immune IgY was extracted as described in the standard PEG protocol (as in Example 1) and the final IgY pellet was dissolved in sterile PBS at the original yolk volume. This material is designated “immune recombinant IgY” or “immune IgY.”

[0296] c) Antitoxin Antibody Titer

[0297] To determine if the recombinant toxin A protein was sufficiently immunogenic to raise antibodies in hens, the antibody titer of a recombinant toxin A polypeptide was determined by ELISA. Eggs from both hens were collected on day 32, the yolks pooled and the antibody was isolated using PEG as described. The immune recombinant IgY antibody titer was determined for the soluble recombinant protein containing the maltose binding protein fusion generated in p-Mal (pMA1870-2680). Ninety-six well Falcon Pro-bind plates were coated overnight at 4° C. with 100 μl /well of toxin A recombinant at 2.5 μg /μl in PBS containing 0.05% thimerosal. Another plate was also coated with maltose binding protein (MBP) at the same concentration, to permit comparison of antibody reactivity to the fusion partner. The next day, the wells were blocked with PBS containing 1% bovine serum albumin (BSA) for 1 hour at 37° C. IgY isolated from immune or preimmune eggs was diluted in antibody diluent (PBS containing 1% BSA and 0.05% Tween-20), and added to the blocked wells and incubated for 1 hour at 37° C. . The plates were washed three times with PBS with 0.05% Tween-20, then three times with PBS. Alkaline phosphatase conjugated rabbit anti-chicken IgG (Sigma) diluted 1:1000 in antibody diluent was added to the plate, and incubated for 1 hour at 37° C. The plates were washed as before and substrate was added, [p-nitrophenyl phosphate (Sigma)] at 1 mg/ml in 0.05M Na2CO3, pH 9.5 and 10 mM MgCl2. The plates were evaluated quantitatively on a Dynatech MR 300 Micro EIA plate reader at 410 nm about 10 minutes after the addition of substrate.

[0298] Based on these ELISA results, high antibody titers were raised in chickens immunized with the toxin A recombinant polypeptide. The recombinant appeared to be highly immunogenic, as it was able to generate high antibody titers relatively quickly with few immunizations. Immune IgY titer directed specifically to the toxin A portion of the recombinant was higher than the immune IgY titer to its fusion partner, the maltose binding protein, and significantly higher than the preimmune IgY. ELISA titers (reciprocal of the highest dilution of IgY generating a signal) in the preimmune IgY to the MBP or the recombinant was <1:30 while the immune IgY titers to MBP and the toxin A recombinant were 1:18750 and >1:93750 respectively. Importantly, the anti-recombinant antibody titers generated in the hens against the recombinant polypeptide is much higher, compared to antibodies to that region raised using native toxin A. The recombinant antibody titer to region 1870-2680 in the CTA antibody preparation is at least five-fold lower compared to the recombinant generated antibodies (1:18750 versus >1:93750). Thus, it appears a better immune response can be generated against a specific recombinant using that recombinant as the immunogen compared to the native toxin A.

[0299] This observation is significant, as it shows that because recombinant portions stimulate the production of antibodies, it is not necessary to use native toxin molecules to produce antitoxin preparations. Thus, the problems associated with the toxicity of the native toxin are avoided and large-scale antitoxin production is facilitated.

[0300] d) Anti-Recombinant Toxin A Neutralization of Toxin A Hemagglutination Activity in vitro

[0301] Toxin A has hemagglutinating activity besides cytotoxic and enterotoxin properties. Specifically, toxin A agglutinates rabbit erythrocytes by binding to a trisaccharide (gal 1-3B1-4GlcNAc) on the cell surface. [H. Krivan et al., Infect. Immun., 53:573-581 (1986).] We examined whether the anti-recombinant toxin A (immune IgY, antibodies raised against the insoluble product expressed in pET) can neutralize the hemagglutination activity of toxin A in vitro. The hemagglutination assay procedure used was described by H. C. Krivan et al. Polyethylene glycol-fractionated immune or preimmune IgY were pre-absorbed with citrated rabbit erythrocytes prior to performing the hemagglutination assay because we have found that IgY alone can agglutinate red blood cells. Citrated rabbit red blood cells (RRBC's)(Cocalico) were washed twice by centrifugation at 450× g with isotonic buffer (0.1 M Tris-HCl, 0.05 M NaCl, pH 7.2). RRBC-reactive antibodies in the IgY were removed by preparing a 10% RRBC suspension (made by adding packed cells to immune or preimmune IgY) and incubating the mixture for 1 hour at 37° C. The RRBCs were then removed by centrifugation. Neutralization of the hemagglutination activity of toxin A by antibody was tested in round-bottomed 96-well microtiter plates. Twenty-five μl of toxin A (36 μg /ml) (Tech Lab) in isotonic buffer was mixed with an equal volume of different dilutions of immune or preimmune IgY in isotonic buffer, and incubated for 15 minutes at room temperature. Then, 50 μl of a 1% RRBC suspension in isotonic buffer was added and the mixture was incubated for 3 hours at 4° C. Positive control wells containing the final concentration of 9 μg/ml of toxin A after dilution without IgY were also included. Hemagglutination activity was assessed visually, with a diffuse matrix of RRBC's coating the bottom of the well representing a positive hemagglutination reaction and a tight button of RRBC's at the bottom of the well representing a negative reaction. The anti-recombinant immune IgY neutralized toxin A hemagglutination activity, giving a neutralization titer of 1:8. However, preimmune IgY was unable to neutralize the hemagglutination ability of toxin A.

[0302] e) Assay of in vitro Toxin A Neutralizing Activity

[0303] The ability of the anti-recombinant toxin A IgY (immune IgY antibodies raised against pMA1870-2680, the soluble recombinant binding domain protein expressed in pMAL, designated as Anti-tox. A-2 in FIG. 14 , and referred to as recombinant region 6) and pre-immune IgY, prepared as described in Example 8(c) above, to neutralize the cytotoxic activity of toxin A was assessed in vitro using the CHO cell cytotoxicity assay, and toxin A (Tech Lab) at a concentration of 0.1 μg/ml, as described in Example 8(d) above. As additional controls, the anti-native toxin A IgY (CTA) and pre-immune IgY preparations described in Example 8(c) above were also tested. The results are shown in FIG. 14.

[0304] The anti-recombinant toxin A IgY demonstrated only partial neutralization of the cytotoxic activity of toxin A, while the pre-immune IgY did not demonstrate any significant neutralizing activity.

EXAMPLE 14

In vivo Neutralization of C. difficile Toxin A

[0305] The ability of avian antibodies (IgY) raised against recombinant toxin A binding domain to neutralize the enterotoxin activity of C. difficile toxin A was evaluated in vivo using Golden Syrian hamsters. The Example involved: (a) preparation of the avian anti-recombinant toxin A IgY for oral administration; (b) in vivo protection of hamsters from C. difficile toxin A enterotoxicity by treatment of toxin A with avian anti-recombinant toxin A IgY; and (c) histologic evaluation of hamster ceca.

[0306] a) Preparation of the Avian Anti-Recombinant Toxin A IgY for Oral Administration

[0307] Eggs were collected from hens which had been immunized with the recombinant C. difficile toxin A fragment pMA1870-2680 (described in Example 13, above). A second group of eggs purchased at a local supermarket was used as a pre-immune (negative) control. Egg yolk immunoglobulin (IgY) was extracted by PEG from the two groups of eggs as described in Example 8(c), and the final IgY pellets were solubilized in one-fourth the original yolk volume using 0.1M carbonate buffer (mixture of NaHCO3 and Na2CO3), pH 9.5. The basic carbonate buffer was used in order to protect the toxin A from the acidic pH of the stomach environment.

[0308] b) In vivo Protection of Hamsters Against C. difficile Toxin A Enterotoxicity by Treatment of Toxin A with Avian Anti-recombinant Toxin A IgY

[0309] In order to assess the ability of the avian anti-recombinant toxin A IgY, prepared in section (a) above to neutralize the in vivo enterotoxin activity of toxin A, an in vivo toxin neutralization model was developed using Golden Syrian hamsters. This model was based on published values for the minimum amount of toxin A required to elicit diarrhea (0.08 mg toxin A/Kg body wt.) and death (0.16 mg toxin A/Kg body wt.) in hamsters when administered orally (Lyerly et al. Infect. Immun., 47:349-352 (1985).

[0310] For the study, four separate experimental groups were used, with each group consisting of 7 female Golden Syrian hamsters (Charles River), approx. three and one-half weeks old, weighing approx. 50 gms each. The animals were housed as groups of 3 and 4, and were offered food and water ad libitum through the entire length of the study.

[0311] For each animal, a mixture containing either 10 μg of toxin A (0.2 mg/Kg) or 30 μg of toxin A (0.6 mg/Kg) (C. difficile toxin A was obtained from Tech Lab and 1 ml of either the anti-recombinant toxin A IgY or pre-immune IgY (from section (a) above) was prepared. These mixtures were incubated at 37° C. for 60 min. and were then administered to the animals by the oral route. The animals were then observed for the onset of diarrhea and death for a period of 24 hrs. following the administration of the toxin A+IgY mixtures, at the end of which time, the following results were tabulated and shown in Table 17:

TABLE 17
Study Outcome At 24 Hours
	Study Outcome at 24 Hours
Experimental Group	Healthy1	Diarrhea2	Dead3
10 μg Toxin A + Antitoxin Against	7	0	0
Interval 6
30 μg Toxin A + Antitoxin Against	7	0	0
Interval 6
10 μg Toxin A + Pre-Immune Serum	0	5	2
30 μg Toxin A + Pre-Immune	0	5	2
1Animals remained healthy through the entire 24 hour study period.
2Animals developed diarrhea, but did not die.
3Animals developed diarrhea, and subsequently died.

[0312] Pretreatment of toxin A at both doses tested, using the anti-recombinant toxin A IgY, prevented all overt symptoms of disease in hamsters. Therefore, pretreatment of C. difficile toxin A, using the anti-recombinant toxin A IgY, neutralized the in vivo enterotoxin activity of the toxin A. In contrast, all animals from the two groups which received toxin A which had been pretreated using pre-immune IgY developed disease symptoms which ranged from diarrhea to death. The diarrhea which developed in the 5 animals which did not die in each of the two pre-immune groups, spontaneously resolved by the end of the 24 hr. study period.

[0313] c) Histologic Evaluation of Hamster Ceca

[0314] In order to further assess the ability of anti-recombinant toxin A IgY to protect hamsters from the enterotoxin activity of toxin A, histologic evaluations were performed on the ceca of hamsters from the study described in section (b) above.

[0315] Three groups of animals were sacrificed in order to prepare histological specimens. The first group consisted of a single representative animal taken from each of the 4 groups of surviving hamsters at the conclusion of the study described in section (b) above. These animals represented the 24 hr. timepoint of the study.

[0316] The second group consisted of two animals which were not part of the study described above, but were separately treated with the same toxin A+pre-immune IgY mixtures as described for the animals in section (b) above. Both of these hamsters developed diarrhea, and were sacrificed 8 hrs. after the time of administration of the toxin A +pre-immune IgY mixtures. At the time of sacrifice, both animals were presenting symptoms of diarrhea. These animals represented the acute phase of the study.

[0317] The final group consisted of a single untreated hamster from the same shipment of animals as those used for the two previous groups. This animal served as the normal control.

[0318] Samples of cecal tissue were removed from the 7 animals described above, and were fixed overnight at 4° C. using 10% buffered formalin. The fixed tissues were paraffin-embedded, sectioned, and mounted on glass microscope slides. The tissue sections were then stained using hematoxylin and eosin (H and E stain), and were examined by light microscopy.

[0319] The tissues obtained from the two 24 hr. animals which received mixtures containing either 10 μg or 30 μg of toxin A and anti-recombinant toxin A IgY were indistinguishable from the normal control, both in terms of gross pathology, as well as at the microscopic level. These observations provide further evidence for the ability of anti-recombinant toxin A IgY to effectively neutralize the in vivo enterotoxin activity of C. difficile toxin A, and thus its ability to prevent acute or lasting toxin A-induced pathology.

[0320] In contrast, the tissues from the two 24 hr. animals which received the toxin A+pre-immune IgY mixtures demonstrated significant pathology. In both of these groups, the mucosal layer was observed to be less organized than in the normal control tissue. The cytoplasm of the epithelial cells had a vacuolated appearance, and gaps were present between the epithelium and the underlying cell layers. The lamina propria was largely absent. Intestinal villi and crypts were significantly diminished, and appeared to have been overgrown by a planar layer of epithelial cells and fibroblasts. Therefore, although these animals overtly appeared to recover from the acute symptoms of toxin A intoxication, lasting pathologic alterations to the cecal mucosa had occurred.

[0321] The tissues obtained from the two acute animals which received mixtures of toxin A and pre-immune IgY demonstrated the most significant pathology. At the gross pathological level, both animals were observed to have severely distended ceca which were filled with watery, diarrhea-like material. At the microscopic level, the animal that was given the mixture containing 10 μg of toxin A and pre-immune IgY was found to have a mucosal layer which had a ragged, damaged appearance, and a disorganized, compacted quality. The crypts were largely absent, and numerous breaks in the epithelium had occurred. There was also an influx of erythrocytes into spaces between the epithelial layer and the underlying tissue. The animal which had received the mixture containing 30 μg of toxin A and pre-immune IgY demonstrated the most severe pathology. The cecal tissue of this animal had an appearance very similar to that observed in animals which had died from C. difficile disease. Widespread destruction of the mucosa was noted, and the epithelial layer had sloughed. Hemmorhagic areas containing large numbers of erythrocytes were very prevalent. All semblance of normal tissue architecture was absent from this specimen. In terms of the presentation of pathologic events, this in vivo hamster model of toxin A-intoxication correlates very closely with the pathologic consequences of C. difficile disease in hamsters. The results presented in this Example demonstrate that while anti-recombinant toxin A (Interval 6) IgY is capable of only partially neutralizing the cytotoxic activity of C. difficile toxin A, the same antibody effectively neutralizes 100% of the in vivo enterotoxin activity of the toxin. While it is not intended that this invention be limited to this mechanism, this may be due to the cytotoxicity and enterotoxicity of C. difficile Toxin A as two separate and distinct biological functions.

EXAMPLE 15

In vivo Neutralization of C. Difficile Toxin A by Antibodies Against Recombinant Toxin A Polypeptides

[0322] The ability of avain antibodies directed against the recombinant C. difficile toxin A fragment 1870-2680 (as expressed by pMA1870-2680; see Example 13) to neutralize the enterotoxic activity of toxin A was demonstrated in Example 14. The ability of avian antibodies (IgYs) directed against other recombinant toxin A epitopes to neutralize native toxin A in vivo was next evaluated. This example involved: (a) the preparation of IgYs against recombinant toxin A polypeptides; (b) in vivo protection of hamsters against toxin A by treatment with anti-recombinant toxin A IgYs and (c) quantification of specific antibody concentration in CTA and Interval 6 IgY PEG preparations.

[0323] The nucleotide sequence of the coding region of the entire toxin A protein is listed in SEQ ID NO:5. The amino acid sequence of the entire toxin A protein is listed in SEQ ID NO:6. The amino acid sequence consisting of amino acid residues 1870 through 2680 of toxin A is listed in SEQ ID NO:7. The amino acid sequence consisting of amino acid residues 1870 through 1960 of toxin A is listed in SEQ ID NO:8.

[0324] a) Preparation of IgY's Against Recombinant Toxin A Polypeptides

[0325] Eggs were collected from Leghorn hens which have been immunized with recombinant C. difficile toxin A polypeptide fragments encompassing the entire toxin A protein. The polypeptide fragments used as immunogens were: 1) pMA 1870-2680 (Interval 6), 2) pPA 1100-1450 (Interval 4), and 3) a mixture of fragments consisting of pMA 30-300 (Interval 1), pMA 300-660 (Interval 2), pMA 660-1100 (Interval 3) and pMA 1450-1870 (Interval 5). This mixture of immunogens is referred to as Interval 1235. The location of each interval within the toxin A molecule is shown in FIG. 15A. In FIG. 15A, the following abbreviations are used: pP refers to the pET23 vector (New England BioLabs); pM refers to the pMAL™-c vector (New England BioLabs); A refers to toxin A; the numbers refer to the amino acid interval expressed in the clone. (For example, the designation pMA30-300 indicates that the recombinant clone encodes amino acids 30-300 of toxin A and the vector used was pMAL-c).

[0326] The recombinant proteins were generated as described in Example 11. The IgYs were extracted and solubilized in 0.1M carbonate buffer pH 9.5 for oral administration as described in Example 14(a). The IgY reactivities against each individual recombinant interval was evaluated by ELISA as described in Example 13(c).

[0327] b) In vivo Protection of Hamsters Against Toxin A by Treatment with Anti-Recombinant Toxin A Antibodies

[0328] The ability of antibodies raised against recombinant toxin A polypeptides to provide in vivo protection against the enterotoxic activity of toxin A was examined in the hamster model system. This assay was performed as described in Example 14(b). Briefly, for each 40-50 gram female Golden Syrian hamster (Charles River), 1 ml of IgY 4× (i.e., resuspended in ¼ of the original yolk volume) PEG prep against Interval 6, Interval 4 or Interval 1235 was mixed with 30 μg (LD100 oral dose) of C. difficile toxin A (Tech Lab). Preimmune IgY mixed with toxin A served as a negative control. Antibodies raised against C. difficile toxoid A (Example 8) mixed with toxin A (CTA) served as a positive control. The mixture was incubated for 1 hour at 37° C. then orally administered to lightly etherized hamsters using an 18G feeding needle. The animals were then observed for the onset of diarrhea and death for a period of approximately 24 hours. The results are shown in Table 18.

TABLE 18
Study Outcome After 24 Hours
	Treatment group	Healthy1	Diarrhea2	Dead3
	Preimmune	0	0	7
	CTA	5	0	0
	Interval 6	6	1	0
	Interval 4	0	1	6
	Interval 1235	0	0	7
	1Animal shows no sign of illness.
	2Animal developed diarrhea, but did not die.
	3Animal developed diarrhea and died.

[0329] Pre-treatment of toxin A with IgYs against Interval 6 prevented diarrhea in 6 of 7 hamsters and completely prevented death in all 7. In contrast, as with preimmune IgY, IgYs against Interval 4 and Interval 1235 had no effect on the onset of diarrhea and death in the hamsters.

[0330] c) Quantification of Specific Antibody Concentration in CTA and Interval 6 IgY PEG Preparations

[0331] To determine the purity of IgY PEG preparations, an aliquot of a pMA1870-2680 (Interval 6) IgY PEG preparation was chromatographed using HPLC and a KW-803 sizing column (Shodex). The resulting profile of absorbance at 280 nm is shown in FIG. 16. The single large peak corresponds to the predicted MW of IgY. Integration of the area under the single large peak showed that greater than 95% of the protein eluted from the column was present in this single peak. This result demonstrated that the majority (>95%) of the material absorbing at 280 nm in the PEG preparation corresponds to IgY. Therefore, absorbance at 280 nm can be used to determine the total antibody concentration in PEG preparations.

[0332] To determine the concentration of Interval 6-specific antibodies (expressed as percent of total antibody) within the CTA and pMA1870-2680 (Interval 6) PEG preparations, defined quantities of these antibody preparations were affinity purified on a pPA1870-2680(H) (shown schematically in FIG. 15B) affinity column and the specific antibodies were quantified. In FIG. 15B the following abbreviations are used: pP refers to the pET23 vector (New England BioLabs); pM refers to the pMAL™-c vector (New England BioLabs); pG refers to the pGEX vector (Pharmacia); pB refers to the PinPoint™ Xa vector (Promega); A refers to toxin A; the numbers refer to the amino acid interval expressed in the clone. The solid black ovals represent the MBP; the hatched ovals represent glutathione S-transferase; the hatched circles represent the biotin tag; and HHH represents the poly-histidine tag.

[0333] An affinity column containing recombinant toxin A repeat protein was made as follows. Four ml of PBS-washed Actigel resin (Sterogene) was coupled with 5-10 mg of pPA1870-2680 inclusion body protein [prepared as described in Example (17) and dialyzed into PBS] in a 15 ml tube (Falcon) containing {fraction (1/10)} final volume Ald-coupling solution (1 M sodium cyanoborohydride). Aliquots of the supernatant from the coupling reactions, before and after coupling, were assessed by Coomassie staining of 7.5% SDS-PAGE gels. Based upon protein band intensities, greater than 6 mg of recombinant protein was coupled to the resin. The resin was poured into a 10 ml column (BioRad), washed extensively with PBS, pre-eluted with 4 M guanidine-HCl (in 10 mM Tris-HCl, pH 8.0; 0.005% thimerisol) and re-equilibrated with PBS. The column was stored at 4° C.

[0334] Aliquots of a pMA1870-2680 (Interval 6) or a CTA IgY polyclonal antibody preparation (PEG prep) were affinity purified on the above affinity column as follows. The column was attached to an UV monitor (ISCO) and washed with PBS. For pMA1870-2680 IgY purification, a 2× PEG prep (filter sterilized using a 0.45μ filter; approximately 500 mg total IgY) was applied. The column was washed with PBS until the baseline was re-established (the column flow-through was saved), washed with BBSTween to elute nonspecifically binding antibodies and re-equilibrated with PBS. Bound antibody was eluted from the column in 4 M guanidine-HCl (in 10 mM Tris-HCl, pH 8.0; 0.005% thimerisol). The entire elution peak was collected in a 15 ml tube (Falcon). The column was re-equilibrated and the column eluate was re-chromatographed as described above. The antibody preparation was quantified by UV absorbance (the elution buffer was used to zero the spectrophotometer). Total purified antibody was approximately 9 mg and 1 mg from the first and second chromotragphy passes, respectively. The low yield from the second pass indicated that most specific antibodies were removed by the first round of chromatography. The estimated percentage of Interval 6 specific antibodies in the pMA1870-2680 PEG prep is approximately 2%.

[0335] The percentage of Interval 6 specific antibodies in the CTA PEG prep was determined (utilizing the same column and methodology described above) to be approximately 0.5% of total IgY.

[0336] A 4× PEG prep contains approximately 20 mg/ml IgY. Thus in b) above, approximately 400 μg specific antibody in the Interval 6 PEG prep neutralized 30 μg toxin A in vivo.

EXAMPLE 16

In vivo Treatment of C. difficile Disease in Hamsters by Recombinant Interval 6 Antibodies

[0337] The ability of antibodies directed against recombinant Interval 6 of toxin A to protect hamsters in vivo from C. difficile disease was examined. This example involved: (a) prophylactic treatment of C. difficile disease and (b) therapeutic treatment of C. difficile disease.

[0338] a) Prophylactic Treatment of C. difficile Disease

[0339] This experiment was performed as described in Example 9(b). Three groups each consisting of 7 female 100 gram Syrian hamsters (Charles River) were prophylactically treated with either preimmune IgYs, IgYs against native toxin A and B [CTAB; see Example 8(a) and (b)] or IgYs against Interval 6. IgYs were prepared as 4× PEG preparations as described in Example 9(a).

[0340] The animals were orally dosed 3 times daily, roughly at 4 hour intervals, for 12 days with 1 ml antibody preparations diluted in Ensure®. Using estimates of specific antibody concentration from Example 15(c), each dose of the Interval 6 antibody prep contained approximately 400 μg of specific antibody. On day 2 each hamster was predisposed to C. difficile infection by the oral administration of 3.0 mg of Clindamycin-HCl (Sigma) in 1 ml of water. On day 3 the hamsters were orally challenged with 1 ml of C. difficile inoculum strain ATCC 43596 in sterile saline containing approximately 100 organisms. The animals were then observed for the onset of diarrhea and subsequent death during the treatment period. The results are shown in Table 19.

TABLE 19
Lethality After 12 Days Of Treatment
Treatment Group	Number Animals Alive	Number Animals Dead
Preimmune	0	7
CTAB	6	1
Interval 6	7	0

[0341] Treatment of hamsters with orally-administered IgYs against Interval 6 successfully protected 7 out of 7 (100%) of the animals from C. difficile disease. One of the hamsters in this group presented with diarrhea which subsequently resolved during the course of treatment. As shown previously in Example 9, antibodies to native toxin A and toxin B were highly protective. In this Example, 6 out of 7 animals survived in the CTAB treatment group. All of the hamsters treated with preimmune sera came down with diarrhea and died. The survivors in both the CTAB and Interval 6 groups remained healthy throughout a 12 day post-treatment period. In particular, 6 out of 7 Interval 6-treated hamsters survived at least 2 weeks after termination of treatment which suggests that these antibodies provide a long-lasting cure. These results represent the first demonstration that antibodies generated against a recombinant region of toxin A can prevent CDAD when administered passively to animals. These results also indicate that antibodies raised against Interval 6 alone may be sufficient to protect animals from C. difficile disease when administered prophylactically.

[0342] Previously others had raised antibodies against toxin A by actively immunizing hamsters against a recombinant polypeptide located within the Interval 6 region [Lyerly, D. M., et al. (1990) Curr. Microbiol. 21:29]. FIG. 17 shows schematically the location of the Lyerly, et al. intra-Interval 6 recombinant protein (cloned into the pUC vector) in comparision with the complete Interval 6 construct (pMA1870-2680) used herein to generate neutralizing antibodies directed against toxin B. In FIG. 17, the solid black oval represents the MBP which is fused to the toxin B Interval 6 in pMA1870-2680.

[0343] The Lyerly, et al. antibodies (intra-Interval 6) were only able to partially protect hamsters against C. difficile infection in terms of survival (4 out of 8 animals survived) and furthermore, these antibodies did not prevent diarrhea in any of the animals. Additionally, animals treated with the intra-Interval 6 antibodies [Lyerly, et al. (1990), supra] died when treatment was removed.

[0344] In contrast, the experiment shown above demonstrates that passive administration of anti-Interval 6 antibodies prevented diarrhea in 6 out of 7 animals and completely prevented death due to CDAD. Furthermore, as discussed above, passive administration of the anti-Interval 6 antibodies provides a long lasting cure (i.e., treatment could be withdrawn without incident).

[0345] b) Therapeutic Treatment of C. difficile Disease: in vivo Treatment of an Established C. difficile Infection in Hamsters with Recombinant Interval 6 Antibodies

[0346] The ability of antibodies against recombinant interval 6 of toxin A to therapeutically treat C. difficile disease was examined. The experiment was performed essentially as described in Example 10(b). Three groups, each containing seven to eight female Golden Syrian hamsters (100 g each; Charles River) were treated with either preimmune IgY, IgYs against native toxin A and toxin B (CTAB) and IgYs against Interval 6. The antibodies were prepared as described above as 4× PEG preparations.

[0347] The hamsters were first predisposed to C. difficile infection with a 3 mg dose of Clindamycin-HCl (Sigma) adminstered orally in 1 ml of water. Approximately 24 hrs later, the animals were orally challenged with 1 ml of C. difficile strain ATCC 43596 in sterile saline conatining approximately 200 organisms. One day after infection, the presence of toxin A and B was determined in the feces of the hamsters using a commerical immunoasssay kit (Cytoclone A+B EIA, Cambridge Biotech) to verify establishment of infection. Four members of each group were randomly selected and tested. Feces from an uninfected hamster was tested as a negative control. All infected animals tested positive for the presence of toxin according to the manufacturer's procedure. The intitiaton of treatment then started approximately 24 hr post-infection.

[0348] The animals were dosed daily at roughly 4 hr intervals with 1 ml antibody preparation diluted in Ensure® (Ross Labs). The amount of specific antibodies given per dose (determined by affinity purification) was estimated to be about 400 μg of anti-Interval 6 IgY (for animals in the Interval 6 group) and 100 μg and 70 μg of anti-toxin A (Interval 6-specific) and anti-toxin B (Interval 3-specific; see Example 19), respectively, for the CTAB preparation. The animals were treated for 9 days and then observed for an additional 4 days for the presence of diarrhea and death. The results indicating the number of survivors and the number of dead 4 days post-infection are shown in Table 20.

TABLE 20
In vivo Therapeutic Treatment With Interval 6 Antibodies
Treatment Group	Number Animals Alive	Number Animals Dead
Preimmune	4	3
CTAB	8	0
Interval 6	8	0

[0349] Antibodies directed against both Interval 6 and CTAB successfully prevented death from C. difficile when therapeutically administered 24 hr after infection. This result is significant since many investigators begin therapeutic treatment of hamsters with existing drugs (e.g., vancomycin, phenelfamycins, tiacumicins, etc.) 8 hr post-infection [Swanson, et al. (1991) Antimicrobial Agents and Chemotherapy 35:1108 and (1989) J. Antibiotics 42:94].

[0350] Forty-two percent of hamsters treated with preimmune IgY died from CDAD. While the anti-Interval 6 antibodies prevented death in the treated hamsters, they did not eliminate all symptoms of CDAD as 3 animals presented with slight diarrhea. In addition, one CTAB-treated and one preimmune-treated animal also had diarrhea 14 days post-infection. These results indicate that anti-Interval 6 antibodies provide an effective means of therapy for CDAD.

EXAMPLE 17

Induction of Toxin A Neutralizing Antibodies Requires Soluble Interval 6 Protein

[0351] As shown in Examples 11(d) and 15, expression of recombinant proteins in E. coli may result in the production of either soluble or insoluble protein. If insoluble protein is produced, the recombinant protein is solubilized prior to immunization of animals. To determine whether, one or both of the soluble or insoluble recombinant proteins could be used to generate neutralizing antibodies to toxin A, the following experiment was performed. This example involved a) expression of the toxin A repeats and subfragments of these repeats in E. coli using a variety of expression vectors; b) identification of recombinant toxin A repeats and sub-regions to which neutralizing antibodies bind; and c) determination of the neutralization ability of antibodies raised against soluble and insoluble toxin A repeat immunogen.

[0352] a) Expression of the Toxin A Repeats and Subfragments of These Repeats in E. coli Using a Variety of Expression Vectors

[0353] The Interval 6 immunogen utilized in Examples 15 and 16 was the pMA1870-2680 protein, in which the toxin A repeats are expressed as a soluble fusion protein with the MBP (described in Example 11). Interestingly, expression of this region (from the SpeI site to the end of the repeats, see FIG. 15B) in three other expression constructs, as either native (pPA1870-2680), poly-His tagged [pPA1870-2680 (H)] or biotin-tagged (pBA1870-2680) proteins resulted in completely insoluble protein upon induction of the bacterial host (see FIG. 15B). The host strain BL21 (Novagen) was used for expression of pBA1870-2680 and host strain BL21(DE3) (Novagen) was used for expression of pPA1870-2680 and pPA1870-2680(H). These insoluble proteins accumulated to high levels in inclusion bodies. Expression of recombinant plasmids in E. coli host cells grown in 2× YT medium was performed as decribed [Williams, et al. (1994), supra].

[0354] As summarized in FIG. 15B, expression of fragments of the toxin A repeats (as either N-terminal SpeI-EcoRI fragments, or C-terminal EcoRI-end fragments) also yielded high levels of insoluble protein using pGEX (pGA1870-2190), PinPoint™-Xa (pBA1870-2190 and pBA2250-2680) and pET expression systems (pPA1870-2190). The pGEX and pET expression systems are described in Example 11. The PinPoin™-Xa expression system drives the expression of fusion proteins in E. coli. Fusion proteins from PinPoint™-Xa vectors contain a biotin tag at the amino-terminal end and can be affinity purified SoftLink™ Soft Release avidin resin (Promega) under mild denaturing conditions (5 mM biotin).

[0355] The solubility of expressed proteins from the pPG1870-2190 and pPA1870-2190 expression constructs was determined after induction of recombinant protein expression under conditions reported to enhance protein solubility [These conditions comprise growth of the host at reduced temperature (30° C.) and the utilization of high (1 mM IPTG) or low (0.1 mM IPTG) concentrations of inducer [Williams et al. (1994), supra]. All expressed recombinant toxin A protein was insoluble under these conditions. Thus, expression of these fragments of the toxin A repeats in pET and pGEX expression vectors results in the production of insoluble recombinant protein even when the host cells are grown at reduced temperature and using lower concentrations of the inducer. Although expression of these fragments in pMal vectors yielded affinity purifiable soluble fusion protein, the protein was either predominantly insoluble (pMA1870-2190) or unstable (pMA2250-2650). Attempts to solubilize expressed protein from the pMA1870-2190 expression construct using reduced temperature or lower inducer concentration (as described above) did not improve fusion protein solubility.

[0356] Collectively, these results demonstrate that expression of the toxin A repeat region in E. coli results in the production of insoluble recombinant protein, when expressed as either large (aa 1870-2680) or small (aa 1870-2190 or aa 2250-2680) fragments, in a variety of expression vectors (native or poly-his tagged pET, pGEX or PinPoint™-Xa vectors), utilizing growth conditions shown to enhance protein solubility. The exception to this rule were fusions with the MBP, which enhanced protein solubility, either partially (pMA1870-2190) or fully (pMA1870-2680).

[0357] b) Identification of Recombinant Toxin A Repeats and Sub-Regions to Which Neutralizing Antibodies Bind

[0358] Toxin A repeat regions to which neutralizing antibodies bind were identified by utilizing recombinant toxin A repeat region proteins expressed as soluble or insoluble proteins to deplete protective antibodies from a polyclonal pool of antibodies against native C. difficile toxin A. An in vivo assay was developed to evaluate proteins for the ability to bind neutralizing antibodies.

[0359] The rational for this assay is as follows. Recombinant proteins were first pre-mixed with antibodies against native toxin A (CTA antibody; generated in Example 8) and allowed to react. Subsequently, C. difficile toxin A was added at a concentration lethal to hamsters and the mixture was adminstered to hamsters via IP injection. If the recombinant protein contains neutralizing epitopes, the CTA antibodies would lose their ability to bind toxin A resulting in diarrhea and/or death of the hamsters.

[0360] The assay was performed as follows. The lethal dose of toxin A when delivered orally to nine 40 to 50 g Golden Syrian hamsters (Sasco) was determined to be 10 to 30 infection g. The PEG-purified CTA antibody preparation was diluted to 0.5× concentration (i.e., the antibodies were diluted at twice the original yolk volume) in 0.1 M carbonate buffer, pH 9.5. The antibodies were diluted in carbonate buffer to protect them from acid degradation in the stomach. The concentration of 0.5× was used because it was found to be the lowest effective concentration against toxin A. The concentration of Interval 6-specific antibodies in the 0.5× CTA prep was estimated to be 10-15 μg/ml (estimated using the method described in Example 15).

[0361] The inclusion body preparation [insoluble Interval 6 protein; pPA1870-2680(H)] and the soluble Interval 6 protein [pMA1870-2680; see FIG. 15] were both compared for their ability to bind to neutralizing antibodies against C. difficile toxin A (CTA). Specifically, 1 to 2 mg of recombinant protein was mixed with 5 ml of a 0.5× CTA antibody prep (estimated to contain 60-70 μg of Interval 6-specific antibody). After incubation for 1 hr at 37° C., CTA (Tech Lab) at a final concentration of 30 μg/ml was added and incubated for another 1 hr at 37° C. One ml of this mixture containing 30 μg of toxin A (and 10-15 μg of Interval 6-specific antibody) was administered orally to 40-50 g Golden Syrian hamsters (Sasco). Recombinant proteins that result in the loss of neutralizing capacity of the CTA antibody would indicate that those proteins contain neutralizing epitopes. Preimmune and CTA antibodies (both at 0.5×) without the addition of any recombinant protein served as negative and positive controls, respectively.

[0362] Two other inclusion body preparations, both expressed as insoluble products in the pET vector, were tested; one containing a different insert (toxin B fragment) other than Interval 6 called pPB1850-2070 (see FIG. 18) which serves as a control for insoluble Interval 6, the other was a truncated version of the Interval 6 region called pPA1870-2190 (see FIG. 15B). The results of this experiment are shown in Table 21.

TABLE 21
Binding Of Neutralizing Antibodies By Soluble Interval 6 Protein
Study Outcome After 24 Hours
Treatment	Group1	Healthy2	Diarrhea3 Dead4
Preimmune Ab	0	3	2
CTA Ab	4	1	0
CTA Ab + Int 6 (soluble)	1	2	2
CTA Ab + Int 6 (insoluble)	5	0	0
CTA Ab + pPB1850-2070	5	0	0
CTA Ab + pPA1870-2190	5	0	0
1C. difficile toxin A (CTA) was added to each group.
2Animals showed no signs of illness.
3Animals developed diahrrhea but did not die.
4Animals developed diarrhea and died.

[0363] Preimmune antibody was ineffective against toxin A, while anti-CTA antibodies at a dilute 0.5× concentration almost completely protected the hamsters against the enterotoxic effects of CTA. The addition of recombinant proteins pPB1850-2070 or pPA1870-2190 to the anti-CTA antibody had no effect upon its protective ability, indicating that these recombinant proteins do not bind to neutralizing antibodies. On the other hand, recombinant Interval 6 protein was able to bind to neutralizing anti-CTA antibodies and neutralized the in vivo protective effect of the anti-CTA antibodies. Four out of five animals in the group treated with anti-CTA antibodies mixed with soluble Interval 6 protein exhibited toxin associated toxicity (diarrhea and death). Moreover, the results showed that Interval 6 protein must be expressed as a soluble product in order for it to bind to neutralizing anti-CTA antibodies since the addition of insoluble Interval 6 protein had no effect on the neutralizing capacity of the CTA antibody prep.

[0364] c) Determination of Neutralization Ability of Antibodies Raised Against Soluble and Insoluble Toxin A Repeat Immunogen

[0365] To determine if neutralizing antibodies are induced against solubilized inclusion bodies, insoluble toxin A repeat protein was solubilized and specific antibodies were raised in chickens. Insoluble pPA1870-2680 protein was solubilized using the method described in Williams et al. (1994), supra. Briefly, induced cultures (500 ml) were pelleted by centrifugation at 3,000× g for 10 min at 4° C. The cell pellets were resuspended thoroughly in 10 ml of inclusion body sonication buffer (25 mM HEPES pH 7.7, 100 mM KCl, 12.5 mM MgCl2, 20% glycerol, 0.1% (v/v) Nonidet P-40, 1 mM DTT). The suspension was transferred to a 30 ml non-glass centrifuge tube. Five hundred μl of 10 mg/ml lysozyme was added and the tubes were incubated on ice for 30 min. The suspension was then frozen at −70° C. for at least 1 hr. The suspension was thawed rapidly in a water bath at room temperature and then placed on ice. The suspension was then sonicated using at least eight 15 sec bursts of the microprobe (Branson Sonicator Model No. 450) with intermittent cooling on ice.

[0366] The sonicated suspension was transferred to a 35 ml Oakridge tube and centrifuged at 6,000× g for 10 min at 4° C. to pellet the inclusion bodies. The pellet was washed 2 times by pipetting or vortexing in fresh, ice-cold RIPA buffer [0.1% SDS, 1% Triton X-100, 1% sodium deoxycholate in TBS (25 mM Tris-Cl pH 7.5, 150 mM NaCl)]. The inclusion bodies were recentrifuged after each wash. The inclusion bodies were dried and transferred using a small metal spatula to a 15 ml tube (Falcon). One ml of 10% SDS was added and the pellet was solubilized by gently pipeting the solution up and down using a 1 ml micropipettor. The solubilization was facilitated by heating the sample to 95° C. when necessary.

[0367] Once the inclusion bodies were in solution, the samples were diluted with 9 volumes of PBS. The protein solutions were dialyzed overnight against a 100-fold volume of PBS containing 0.05% SDS at room temperature. The dialysis buffer was then changed to PBS containing 0.01% SDS and the samples were dialyzed for several hours to overnight at room temperature. The samples were stored at 4° C. until used. Prior to further use, the samples were warmed to room temperature to allow any precipitated SDS to go back into solution.

[0368] The inclusion body preparation was used to immunize hens. The protein was dialysed into PBS and emulsified with approximately equal volumes of CFA for the initial immunization or IFA for subsequent booster immunizations. On day zero, for each of the recombinant recombinant preparations, two egg laying white Leghorn hens were each injected at multiple sites (IM and SC) with 1 ml of recombinant protein-adjuvant mixture containing approximately 0.5-1.5 mg of recombinant protein. Booster immunizations of 1.0 mg were given of days 14 and day 28. Eggs were collected on day 32 and the antibody isolated using PEG as described in Example 14(a). High titers of toxin A specific antibodies were present (as assayed by ELISA, using the method described in Example 13). Titers were determined for both antibodies against recombinant polypeptides pPA1870-2680 and pMA1870-2680 and were found to be comparable at >1:62,500.

[0369] Antibodies against soluble Interval 6 (pMA1870-2680) and insoluble Interval 6 [(inclusion body), pPA1870-2680] were tested for neutralizing ability against toxin A using the in vivo assay described in Example 15(b). Preimmune antibodies and antibodies against toxin A (CTA) served as negative and positive controls, respectively. The results are shown in Table 22.

TABLE 22
Neutralization of Toxin A by Antibodies Against Soluble Interval
6 Protein Study Outcome After 24 Hours
Antibody Treatment Group	Healthy1	Diarrhea2	Dead3
Preimmune	1	0	4
CTA	5	0	0
Interval 6 (Soluble)4	5	0	0
Interval 6 (Insoluble)	0	2	3
1Animals showed no sign of illness.
2Animal developed diarrhea but did not die.
3Animal developed diarrhea and died.
4400 μg/ml.

[0370] Antibodies raised against native toxin A were protective while preimmune antibodies had little effect. As found using the in vitro CHO assay [described in Example 8(d)] where antibodies raised against the soluble Interval 6 could partially neutralize the effects of toxin A, here they were able to completely neutralize toxin A in vivo. In contrast, the antibodies raised against the insoluble Interval 6 was unable to neutralize the effcets of toxin A in vivo as shown above (Table 22) and in vitro as shown in the CHO assay [described in Example 8(d)].

[0371] These results demonstrate that soluble toxin A repeat immunogen is necessary to induce the production of neutralizing antibodies in chickens, and that the generation of such soluble immunogen is obtained only with a specific expression vector (pMal) containing the toxin A region spanning aa 1870-2680. That is to say, insoluble protein that is subsequently solubilized does not result in a toxin A antigen that will elicit a neutralizing antibody.

EXAMPLE 18

Cloning and Expression of the C. difficile Toxin B Gene

[0372] The toxin B gene has been cloned and sequenced; the amino acid sequence decided from the cloned nucleotide sequence predicts a MW of 269.7 kD for toxin B [Baroso et al., Nucl. Acids Res. 18:4004 (1990)]. The nucleotide sequence of the coding region of the entire toxin B gene is listed in SEQ ID NO:9. The amino acid sequence of the entire toxin B protein is listed in SEQ ID NO:10. The amino acid sequence consisting of amino acid residues 1850 through 2360 of toxin B is listed in SEQ ID NO:11. The amino acid sequence consisting of amino acid residues 1750 through 2360 of toxin B is listed in SEQ ID NO:12.

[0373] Given the expense and difficulty of isolating native toxin B protein, it would be advantageous to use simple and inexpensive procaryotic expression systems to produce and purify high levels of recombinant toxin B protein for immunization purposes. Ideally, the isolated recombinant protein would be soluble in order to preserve native antigenicity, since solubilized inclusion body proteins often do not fold into native conformations. Indeed as shown in Example 17, neutralizing antibodies against recombinant toxin A were only obtained when soluble recombinant toxin A polypeptides were used as the immunogen. To allow ease of purification, the recombinant protein should be expressed to levels greater than 1 mg/liter of E. coli culture.

[0374] To determine whether high levels of recombinant toxin B protein could be produced in E. coli, fragments of the toxin B gene were cloned into various prokaryotic expression vectors, and assessed for the ability to express recombinant toxin B protein in E. coli. This Example involved (a) cloning of the toxin B gene and (b) expression of the toxin B gene in E. coli.

[0375] a) Cloning of the Toxin B Gene

[0376] The toxin B gene was cloned using PCR amplification from C. difficile zgenomic DNA. Initially, the gene was cloned in two overlapping fragments, using primer pairs P5/P6 and P7/P8. The location of these primers along the toxin B gene is shown schematically in FIG. 18. The sequence of each of these primers is: P5: 5′ TAGAAAAAATGGCAAATGT 3′ (SEQ ID NO:11); P6: 5′ TTTCATCTTGTA GAGTCAAAG 3′ (SEQ ID NO:12); P7: 5′ GATGCCACAAGATGATTTAGTG 3′ (SEQ ID NO:13); and P8: 5′ CTAATTGAGCTGTATCAGGATC3′ (SEQ ID NO:14).

[0377] FIG. 18 also shows the location of the following primers along the toxin B gene: P9 which consists of the sequence 5′ CGGAATTCCTAGAAAAAATGGCAA ATG 3′ (SEQ ID NO:15); P10 which consists of the sequence 5′ GCTCTAGAATGA CCATAAGCTAGCCA 3′ (SEQ ID NO:16); P11 which consists of the sequence 5′ CGGAATTCGAGTTGGTAGAAAGGTGGA 3′ (SEQ ID NO:17); P13 which consists of the sequence 5′ CGGAATTCGGTTATTATCTTAAGGATG 3′ (SEQ ID NO:18); and P14 which consists of the sequence 5′ CGGAATTCTTGATAACTGGAT TTGTGAC 3′ (SEQ ID NO:19). The amino acid sequence consisting of amino acid residues 1852 through 2362 of toxin B is listed in SEQ ID NO:20. The amino acid sequence consisting of amino acid residues 1755 through 2362 of toxin B is listed in SEQ ID NO:21.

[0378] Clostridium difficile VPI strain 10463 was obtained from the American Type Culture Collection (ATCC 43255) and grown under anaerobic conditions in brain-heart infusion medium (Becton Dickinson). High molecular-weight C. difficile DNA was isolated essentially as described [Wren and Tabaqchali (1987) J. Clin. Microbiol., 25:2402], except 1) 100 μg/ml proteinase K in 0.5% SDS was used to disrupt the bacteria and 2) cetytrimethylammonium bromide (CTAB) precipitation [as described by Ausubel et al., Eds., Current Protocols in Molecular Biology, Vol. 2 (1989) Current Protocols] was used to remove carbohydrates from the cleared lysate. Briefly, after disruption of the bacteria with proteinase K and SDS, the solution is adjusted to approximately 0.7 M NaCl by the addition of a 1/7 volume of 5M NaCl. A 1/10 volume of CTAB/NaCl (10% CTAB in 0.7 M NaCl) solution was added and the solution was mixed thoroughly and incubated 10 min at 65° C. An equal volume of chloroform/isoamyl alcohol (24:1) was added and the phases were thoroughly mixed. The organic and aqueous phases were separated by centrifugation in a microfuge for 5 min. The aqueous supernatant was removed and extracted with phenol/chloroform/ isoamyl alcohol (25:24:1). The phases were separated by centrifugation in a microfuge for 5 min. The supernatant was transferred to a fresh tube and the DNA was precipitated with isopropanol. The DNA precipitate was pelleted by brief centrifugation in a microfuge. The DNA pellet was washed with 70% ethanol to remove residual CTAB. The DNA pellet was then dried and redissolved in TE buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA). The integrity and yield of genomic DNA was assessed by comparison with a serial dilution of uncut lambda DNA after electrophoresis on an agarose gel.

[0379] Toxin B fragments were cloned by PCR utilizing a proofreading thermostable DNA polymerase [native Pfu polymerase (Statagene)]. The high fidelity of this polymerase reduces the mutation problems associated with amplification by error prone polymerases (e.g., Taq polymerase). PCR amplification was performed using the PCR primer pairs P5 (SEQ ID NO:11) with P6 (SEQ ID NO:12) and P7 (SEQ ID NO:13) with P8 (SEQ ID NO:14) in 50 μl reactions containing 10 mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 200 μM of each dNTP, 0.2 μM each primer, and 50 ng C. difficile genomic DNA. Reactions were overlayed with 100 μl mineral oil, heated to 94° C. for 4 min, 0.5 μl native Pfu polymerase (Stratagene) was added, and the reactions were cycled 30 times at 94° C. for 1 min, 50° C. for 1 min, 72° C. (2 min for each kb of sequence to be amplified), followed by 10 min at 72° C. Duplicate reactions were pooled, chloroform extracted, and ethanol precipitated. After washing in 70% ethanol, the pellets were resuspended in 50 μl TE buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA).

[0380] The P5/P6 amplification product was cloned into pUC19 as outlined below. 10 μl aliquots of DNA were gel purified using the Prep-a-Gene kit (BioRad), and ligated to SmaI restricted pUC19 vector. Recombinant clones were isolated and confirmed by restriction digestion using standard recombinant molecular biology techniques (Sambrook et al., 1989). Inserts from two independent isolates were identified in which the toxin B insert was oriented such that the vector BamHI and SacI sites were 5′ and 3′ oriented, respectively (pUCB10-1530). The insert-containing BamHI/SacI fragment was cloned into similarly cut pET23a-c vector DNA, and protein expression was induced in small scale cultures (5 ml) of identified clones.

[0381] Total protein extracts were isolated, resolved on SDS-PAGE gels, and toxin B protein identified by Western analysis utilizing a goat anti-toxin B affinity purified antibody (Tech Lab). Procedures for protein induction, SDS-PAGE, and Western blot analysis were performed as described in Williams et al. (1994), supra. In brief, 5 ml cultures of bacteria grown in 2× YT containing 100 μg/ml ampicillin containing the appropriate recombinant clone were induced to express recombinant protein by addition of IPTG to 1 mM. The E. coli hosts used were: BL21(DE3) or BL21(DE3)LysS (Novagen) for pET plasmids.

[0382] Cultures were induced by the addition of IPTG to a final concentration of 1.0 mM when the cell density reached 0.5 OD600, and induced protein was allowed to accumulate for two hrs after induction. Protein samples were prepared by pelleting 1 ml aliquots of bacteria by centrifugation (1 min in microfuge), and resuspension of the pelleted bacteria in 150 μl of 2× SDS-PAGE sample buffer (0.125 mM Tris-HCl pH 6.8, 2 mM EDTA, 6% SDS, 20% glycerol, 0.025% bromophenol blue; β-mercaptoethanol is added to 5% before use). The samples were heated to 95° C. for 5 min, then cooled and 5 or 10 μs loaded on 7.5% SDS-PAGE gels. High molecular weight protein markers (BioRad) were also loaded, to allow estimation of the MW of identified fusion proteins. After electrophoresis, protein was detected either generally by staining the gels with Coomassie Blue, or specifically, by blotting to nitrocellulose for Western blot detection of specific immunoreactive protein. The MW of induced toxin B reactive protein allowed the integrity of the toxin B reading frame to be determined.

[0383] The pET23b recombinant (pPB10-1530) expressed high MW recombinant toxin B reactive protein, consistent with predicted values. This confirmed that frame terminating errors had not occurred during PCR amplification. A pET23b expression clone containing the 10-1750aa interval of the toxin B gene was constructed, by fusion of the EcoRI-SpeI fragment of the P7/P8 amplification product to the P5-EcoRV interval of the P5/P6 amplification product (see FIG. 18) in pPB10-1530. The integrity of this clone (pPB10-1750) was confirmed by restriction mapping, and Western blot detection of expressed recombinant toxin B protein. Levels of induced protein from both pPB10-1530 and pPB10-1750 were too low to facilitate purification of usable amounts of recombinant toxin B protein. The remaining 1750-2360 aa interval was directly cloned into pMAL or pET expression vectors from the P7/P8 amplification product as described below.

[0384] b) Expression of the Toxin B Gene

[0385] i) Overview of Expression Methodologies

[0386] Fragments of the toxin B gene were expressed as either native or fusion proteins in E. coli. Native proteins were expressed in either the pET23a-c or pET16b expression vectors (Novagen). The pET23 vectors contain an extensive polylinker sequence in all three reading frames (a-c vectors) followed by a C-terminal poly-histidine repeat. The pET 16b vector contains a N-terminal poly-histidine sequence immediately 5′ to a small polylinker. The poly-histidine sequence binds to Ni-Chelate columns and allows affinity purification of tagged target proteins [Williams et al. (1994), supra]. These affinity tags are small (10 aa for pET16b, 6 aa for pET23) allowing the expression and affinity purification of native proteins with only limited amounts of foreign sequences.

[0387] An N-terminal histidine-tagged derivative of pET16b containing an extensive cloning cassette was constructed to facilitate cloning of N-terminal poly-histidine tagged toxin B expressing constructs. This was accomplished by replacement of the promoter region of the pET23a and b vectors with that of the pET16b expression vector. Each vector was restricted with BglII and NdeI, and the reactions resolved on a 1.2% agarose gel. The pET16b promoter region (contained in a 200 bp BglII-NdeI fragment) and the promoter-less pET23 a or b vectors were cut from the gel, mixed and Prep-A-Gene (BioRad) purified. The eluted DNA was ligated, and transformants screened for promoter replacement by NcoI digestion of purified plasmid DNA (the pET16b promoter contains this site, the pET23 promoter does not). These clones (denoted pETHisa or b) contain the pET16b promoter (consisting of a pT7-lac promoter, ribosome binding site and poly-histidine (10aa) sequence) fused at the NdeI site to the extensive pET23 polylinker.

[0388] All MBP fusion proteins were constructed and expressed in the pMAL™-c or pMAL™-p2 vectors (New England Biolabs) in which the protein of interest is expressed as a C-terminal fusion with MBP. All pET plasmids were expressed in either the BL21(DE3) or BL21(DE3)LysS expression hosts, while pMal plasmids were expressed in the BL21 host.

[0389] Large scale (500 mls to 1 liter) cultures of each recombinant were grown in 2× YT broth, induced, and soluble protein fractions were isolated as described [Williams, et al. (1994), supra]. The soluble protein extracts were affinity chromatographed to isolate recombinant fusion protein, as described [Williams et al., (1994) supra]. In brief, extracts containing tagged pET fusions were chromatographed on a nickel chelate column, and eluted using imidazole salts or low pH (pH 4.0) as described by the distributor (Novagen or Qiagen). Extracts containing soluble pMAL fusion protein were prepared and chromatographed in PBS buffer over an amylose resin (New England Biolabs) column, and eluted with PBS containing 10 mM maltose as described [Williams et al. (1994), supra].

[0390] ii) Overview of Toxin B Expression

[0391] In both large expression constructs described in (a) above, only low level (i.e., less than 1 mg/liter of intact or nondegraded recombinant protein) expression of recombinant protein was detected. A number of expression constructs containing smaller fragments of the toxin B gene were then constructed, to determine if small regions of the gene can be expressed to high levels (i.e., greater than 1 mg/liter intact protein) without extensive protein degradation. All were constructed by in frame fusions of convenient toxin B restriction fragments to either the pMAL-c, pET23a-c, pET16b or pETHisa-b expression vectors, or by engineering restriction sites at specific locations using PCR amplification [using the same conditions described in (a) above]. In all cases, clones were verified by restriction mapping, and, where indicated, DNA sequencing.

[0392] Protein preparations from induced cultures of each of these constructs were analyzed, by SDS-PAGE, to estimate protein stability (Coomassie Blue staining) and immunoreactivity against anti-toxin B specific antiserum (Western analysis). Higher levels of intact (i.e., nondegraded), full length fusion proteins were observed with the smaller constructs as compared with the larger recombinants, and a series of expression constructs spanning the entire toxin B gene were constructed (FIGS. 18, 19 and 20 and Table 23).

[0393] Constructs that expressed significant levels of recombinant toxin B protein (greater than 1 mg/liter intact recombinant protein) in E. coli were identified and a series of these clones that spans the toxin B gene are shown in FIG. 19 and summarized in Table 23. These clones were utilized to isolate pure toxin B recombinant protein from the entire toxin B gene. Significant protein yields were obtained from pMAL expression constructs spanning the entire toxin B gene, and yields of full length soluble fusion protein ranged from an estimated 1 mg/liter culture (pMB1100-1530) to greater than 20 mg/liter culture (pMB1750-2360).

[0394] Representative purifications of MBP and poly-histidine-tagged toxin B recombinants are shown in FIGS. 21 and 22. FIG. 21 shows a Commassie Blue stained 7.5% SDS-PAGE gel on which various protein samples extracted from bacteria harboring pMB1850-2360 were electrophoresed. Samples were loaded as follows: Lane 1: protein extracted from uninduced culture; Lane 2: induced culture protein; Lane 3: total protein from induced culture after sonication; Lane 4: soluble protein; and Lane 5: eluted affinity purified protein. FIG. 22 depicts the purification of recombinant proteins expressed in bacteria harboring either pPB1850-2360 (Lanes 1-3) or pPB1750-2360 (Lanes 4-6). Samples were loaded as follows: uninduced total protein (Lanes 1 and 4); induced total protein (Lanes 2 and 5); and eluted affinity purified protein (Lanes 3 and 6). The broad range molecular weight protein markers (BioRad) are shown in Lane 7.

[0395] Thus, although high level expression was not attained using large expression constructs from the toxin B gene, usable levels of recombinant protein were obtained by isolating induced protein from a series of smaller pMAL expression constructs that span the entire toxin B gene.

[0396] These results represent the first demonstration of the feasibility of expressing recombinant toxin B protein to high levels in E. coli. As well, expression of small regions of the putative ligand binding domain (repeat region) of toxin B as native protein yielded insoluble protein, while large constructs, or fusions to MBP were soluble (FIG. 19), demonstrating that specific methodologies are necessary to produce soluble fusion protein from this interval.

[0397] iii) Clone Construction and Expression Details

[0398] A portion of the toxin B gene containing the toxin B repeat region [amino acid residues 1852-2362 of toxin B (SEQ ID NO:20)] was isolated by PCR amplification of this interval of the toxin B gene from C. difficile genomic DNA. The sequence, and location within the toxin B gene, of the two PCR primers [P7 (SEQ ID NO:13) and P8 (SEQ ID NO:14)] used to amplify this region are shown in FIG. 18.

[0399] DNA from the PCR amplification was purified by chloroform extraction and ethanol precipitation as described above. The DNA was restricted with SpeI, and the cleaved DNA was resolved by agarose gel electrophoresis. The restriction digestion with SpeI cleaved the 3.6 kb amplification product into a 1.8 kb doublet band. This doublet band was cut from the gel and mixed with appropriately cut, gel purified pMALc or pET23b vector. These vectors were prepared by digestion with HindIII, filling in the overhanging ends using the Klenow enzyme, and cleaving with XbaI (pMALc) or NheI (pET23b). The gel purified DNA fragments were purified using the Prep-A-Gene kit (BioRad) and the DNA was ligated, transformed and putative recombinant clones analyzed by restriction mapping. pET and pMal clones containing the toxin B repeat insert (aa interval 1750-2360 of toxin B) were verified by restriction mapping, using enzymes that cleaved specific sites within the toxin B region. In both cases fusion of the toxin B SpeI site with either the compatible XbaI site (pMal) or compatible NheI site (pET) is predicted to create an in frame fusion. This was confirmed in the case of the pMB1750-2360 clone by DNA sequencing of the clone junction and 5′ end of the toxin B insert using a MBP specific primer (New England Biolabs). In the case of the pET construct, the fusion of the blunt ended toxin B 3′ end to the filled HindIII site should create an in-frame fusion with the C-terminal poly-histidine sequence in this vector. The pPB1750-2360 clone selected had lost, as predicted, the HindIII site at this clone junction; this eliminated the possibility that an additional adenosine residue was added to the 3′ end of the PCR product by a terminal transferase activity of the Pfu polymerase, since fusion of this adenosine residue to the filled HindIII site would regenerate the restriction site (and was observed in several clones).

[0400] One liter cultures of each expression construct were grown, and fusion protein purified by affinity chromatography on either an amylose resin column (pMAL constructs; resin supplied by New England Biolabs) or Ni-chelate column (pET constructs; resin supplied by Qiagen or Novagen) as described [Williams et al. (1994), supra]. The integrity and purity of the fusion proteins were determined by Coomassie staining of SDS-PAGE protein gels as well as Western blot analysis with either an affinity purified goat polyclonal antiserum (Tech Lab), or a chicken polyclonal PEG prep, raised against the toxin B protein (CTB) as described above in Example 8. In both cases, affinity purification resulted in yields in excess of 20 mg protein per liter culture, of which greater than 90% was estimated to be full-length recombinant protein. It should be noted that the poly-histidine affinity tagged protein was released from the Qiagen Ni-NTA resin at low imidazole concentration (60 mM), necessitating the use of a 40 mM imidazole rather than a 60 mM imidazole wash step during purification.

[0401] A periplasmically secreted version of pMB1750-2360 was constructed by replacement of the promoter and MBP coding region of this construct with that from a related vector (pMAL™-p2; New England Biolabs) in which a signal sequence is present at the N-terminus of the MBP, such that fusion protein is exported. This was accomplished by substituting a BglII-EcoRV promoter fragment from pMAL-p2 into pMB1750-2360. The yields of secreted, affinity purified protein (recovered from osmotic shock extracts as described by Riggs in Current Protocols in Molecular Biology, Vol. 2, Ausubel, et al., Eds. (1989), Current Protocols, pp. 16.6.1-16.6.14] from this vector (pMBp1750-2360) were 6.5 mg/liter culture, of which 50% was estimated to be full-length fusion protein.

[0402] The interval was also expressed in two non-overlapping fragments. pMB1750-1970 was constructed by introduction of a frameshift into pMB1750-2360, by restriction with HindIII, filling in the overhanging ends and religation of the plasmid. Recombinant clones were selected by loss of the HindIII site, and further restriction map analysis. Recombinant protein expression from this vector was more than 20 mg/liter of greater than 90% pure protein.

[0403] The complementary region was expressed in pMB1970-2360. This construct was created by removal of the 1750-1970 interval of pMB1750-2360. This was accomplished by restriction of this plasmid with EcoRI (in the pMalc polylinker 5′ to the insert) and III, filling in the overhanging ends, and religation of the plasmid. The resultant plasmid, pMB1970-2360, was made using both intracellularly and secreted versions of the pMB1750-2360 vector.

[0404] No fusion protein was secreted in the pMBp1970-2360 version, perhaps due to a conformational constraint that prevents export of the fusion protein. However, the intracellularly expressed vector produced greater than 40 mg/liter of greater than 90% full-length fusion protein.

[0405] Constructs to precisely express the toxin B repeats in either pMalc (pMB1850-2360) or pET16b (pPB1850-2360) were constructed as follows. The DNA interval including the toxin B repeats was PCR amplified as described above utilizing PCR primers P14 (SEQ ID NO:19) and P8 (SEQ ID NO:14). Primer P14 adds a EcoRI site immediately flanking the start of the toxin B repeats.

[0406] The amplified fragment was cloned into the pT7 Blue T-vector (Novagen) and recombinant clones in which single copies of the PCR fragment were inserted in either orientation were selected (pT71850-2360) and confirmed by restriction mapping. The insert was excised from two appropriately oriented independently isolated pT71850-2360 plasmids, with EcoRI (5′ end of repeats) and PstI (in the flanking polylinker of the vector), and cloned into EcoRI/PstI cleaved pMalc vector. The resulting construct (pMB1850-2360) was confirmed by restriction analysis, and yielded 20 mg/l of soluble fusion protein [greater than 90% intact (i.e., nondegraded)] after affinity chromatography. Restriction of this plasmid with HindIII and religation of the vector resulted in the removal of the 1970-2360 interval. The resultant construct (pMB1850-1970) expressed greater than 70 mg/liter of 90% full length fusion protein.

[0407] The pPB1850-2360 construct was made by cloning a EcoRI (filled with Klenow)-BamHI fragment from a pT71850-2360 vector (opposite orientation to that used in the pMB1850-2360 construction) into NdeI (filled)/BamHI cleaved pET16b vector. Yields of affinity purified soluble fusion protein were 15 mg/liter, of greater than 90% full length fusion protein.

[0408] Several smaller expression constructs from the 1750-2070 interval were also constructed in His-tagged pET vectors, but expression of these plasmids in the BL21 (DE3) host resulted in the production of high levels of mostly insoluble protein (see Table 23 and FIG. 19). These constructs were made as follows. pPB1850-1970 was constructed by cloning a BgIII-HindIII fragment of pPB1850-2360 into BglII/HindIII cleaved pET23b vector. pPB1850-2070 was constructed by cloning a BglII-PvuII fragment of pPB1850-2360 into BglII/HincII cleaved pET23b vector. pPB1750-1970(c) was constructed by removal of the internal HindIII fragment of a pPB1750-2360 vector in which the vector HindIII site was regenerated during cloning (presumably by the addition of an A residue to the amplified PCR product by terminal transferase activity of Pfu polymerase). The pPB1750-1970(n) construct was made by insertion of the insert containing the NdeI-HindIII fragment of pPB1750-2360 into identically cleaved pETHisb vector. All constructs were confirmed by restriction digestion.

[0409] An expression construct that directs expression of the 10-470 aa interval of toxin B was constructed in the pMalc vector as follows. A NheI (a site 5′ to the insert in the pET23 vector)-AfII (filled) fragment of the toxin B gene from pPB10-1530 was cloned into XbaI (compatible with NheI)/HindIII (filled) pMalc vector. The integrity of the construct (PMB10-470) was verified by restriction mapping and DNA sequencing of the 5′ clone junction using a MBP specific DNA primer (New England Biolabs). However, all expressed protein was degraded to the MBP monomer MW.

[0410] A second construct spanning this interval (aa 10-470) was constructed by cloning the PCR amplification product from a reaction containing the P9 (SEQ ID NO:15) and P10 (SEQ ID NO:16) primers (FIG. 18) into the pETHisa vector. This was accomplished by cloning the PCR product as an EcoRI-blunt fragment into EcoRI-HincII restricted vector DNA; recombinant clones were verified by restriction mapping. Although this construct (pPB10-520) allowed expression and purification (utilizing the N-terminal polyhistidine affinity tag) of intact fusion protein, yields were estimated at less than 500 μg per liter culture.

[0411] Higher yield of recombinant protein from this interval (aa 10-520) were obtained by expression of the interval in two overlapping clones. The 10-330aa interval was cloned in both pETHisa and pMalc vectors, using the BamHI-AflIII (filled) DNA fragment from pPB10-520. This fragment was cloned into BamHI-HindIII (filled) restricted pMalc or BamHI-HincII restricted pETHisa vector. Recombinant clones were verified by restriction mapping. High level expression of either insoluble (pET) or soluble (pMal) fusion protein was obtained. Total yields of fusion protein from the pMB10-330 construct (FIG. 18) were 20 mg/liter culture, of which 10% was estimated to be full-length fusion protein. Although yields of this interval were higher and >90% full-length recombinant protein produced when expressed from the pET construct, the pMal fusion was utilized since the expressed protein was soluble and thus more likely to have the native conformation.

[0412] The pMB260-520 clone was constructed by cloning EcoRI-XbaI cleaved amplified DNA from a PCR reaction containing the P11 (SEQ ID NO:17) and P10 (SEQ ID NO:16) DNA primers (FIG. 18) into similarly restricted pMalc vector. Yields of affinity purified protein were 10 mg/liter, of which approximately 50% was estimated to be full-length recombinant protein.

[0413] The aa510-1110 interval was expressed as described below. This entire interval was expressed as a pMal fusion by cloning the NheI-HindIII fragment of pUCB10-1530 into XbaI-HindIII cleaved pMalc vector. The integrity of the construct (pMB510-1110) was verified by restriction mapping and DNA sequencing of the 5′ clone junction using a MBP specific DNA primer. The yield of affinity purified protein was 25 mg/liter culture, of which 5% was estimated to be full-length fusion protein (1 mg/liter).

[0414] To attempt to obtain higher yields, this region was expressed in two fragments (aa510-820, and 820-1110) in the pMalc vector. The pMB510-820 clone was constructed by insertion of a SacI (in the pMalc polylinker 5′ to the insert)-HpaI DNA fragment from pMB510-1110 into SacI/StuI restricted pMalc vector. The pMB820-1110 vector was constructed by insertion of the HpaI-HindIII fragment of pUCB 10-1530 into BamHI (filled)/HindIII cleaved pMalc vector. The integrity of these constructs were verified by restriction mapping and DNA sequencing of the 5′ clone junction using a MBP specific DNA primer.

[0415] Recombinant protein expressed from the pMB510-820 vector was highly unstable. However, high levels (20 mg/liter) of >90% full-length fusion protein were obtained from the pMB820-1105 construct. The combination of partially degraded pMB510-1110 protein (enriched for the 510-820 interval) with the pMB820-1110 protein provides usable amounts of recombinant antigen from this interval.

[0416] The aa1100-1750 interval was expressed as described below. The entire interval was expressed in the pMalc vector from a construct in which the AccI(filled)-SpeI fragment of pPB10-1750 was inserted into StuI/XbaI (XbaI is compatible with SpeI; StuI and filled AccI sites are both blunt ended) restricted pMalc. The integrity of this construct (pMB1100-1750) was verified by restriction mapping and DNA sequencing of the clone junction using a MBP specific DNA primer. Although 15 mg/liter of affinity purified protein was isolated from cells harboring this construct, the protein was greater than 99% degraded to MBP monomer size.

[0417] A smaller derivative of pMB1100-1750 was constructed by restriction of pMB1100-1750 with AflII and SalI (in the pMalc polylinker 3′ to the insert), filling in the overhanging ends, and religating the plasmid. The resultant clone (verified by restriction digestion and DNA sequencing) has deleted the aa1530-1750 interval, was designated pMB1100-1530. pMB1100-1530 expressed recombinant protein at a yield of greater than 40 mg/liter, of which 5% was estimated to be full-length fusion protein.

[0418] Three constructs were made to express the remaining interval. Initially, a BspHI (filled)-SpeI fragment from pPB10-1750 was cloned into EcoRI(filled)/XbaI cleaved pMalc vector. The integrity of this construct (pMB1570-1750) was verified by restriction mapping and DNA sequencing of the 5′ clone junction using a MBP specific DNA primer. Expression of recombinant protein from this plasmid was very low, approximately 3 mg affinity purified protein per liter, and most was degraded to MBP monomer size. This region was subsequently expressed from a PCR amplified DNA fragment. A PCR reaction utilizing primers P13 [SEQ ID NO:18; P13 was engineered to introduce an EcoRI site 5′ to amplified toxin B sequences] and P8 (SEQ ID NO:14) was performed on C. difficile genomic DNA as described above. The amplified fragment was cleaved with EcoRI and SpeI, and cloned into EcoRI/XbaI cleaved pMalc vector. The resultant clone (pMB1530-1750) was verified by restriction map analysis, and recombinant protein was expressed and purified. The yield was greater than 20 mg protein per liter culture and it was estimated that 25% was full-length fusion protein; this was a significantly higher yield than the original pMB1570-1750 clone. The insert of pMB1530-1750 (in a EcoRI-SalI fragment) was transferred to the pETHisa vector (EcoRI/XhoI cleaved, XhoI and SalI ends are compatible). No detectable fusion protein was purified on Ni-Chelate columns from soluble lysates of cells induced to express fusion protein from this construct.

TABLE 23
Summary Of Toxin B Expression Constructsa
Clone	Affinity Tag	Yield (mg/liter)	% Full Length
pPB10-1750	none	low (estimated from	?
		Western blot hyb.)
pPB10-1530	none	low (as above)	?
pMB10-470	MBP	15	mg/l	0%
pPB10-520	poly-his	0.5	mg/l	20%
pPB10-330	poly-his	>20 mg/l (insoluble)	90%
pMB10-330	MBP	20	mg/l	10%
pMB260-520	MBP	10	mg/l	50%
pMB510-1110	MBP	25	mg/l	5%
pMB510-820	MBP	degraded (by
		Western blot hyb)
pMB820-1110	MBP	20	mg/l	90%
pMB1100-1750	MBP	15	mg/l	0%
pMB1100-1530	MBP	40	mg/l	5%
pMB1570-1750	MBP	3	mg/l	<5%
pPB1530-1750	poly-his	no purified	?
		protein detected
pMB1530-1750	MBP	20	mg/l	25%
pMB1750-2360	MBP	>20	mg/l	>90%
pMBp1750-2360	MBP	6.5 mg/l (secreted)	50%
pPB1750-2360	poly-his	>20	mg/l	>90%
pMB1750-1970	MBP	>20	mg/l	>90%
pMB1970-2360	MBP	40	mg/l	>90%
pMBp1970-2360	MBP	(no secretion)	NA
pMB1850-2360	MBP	20	mg/l	>90%
pPB1850-2360	poly-his	15	mg/l	>90%
pMB1850-1970	MBP	70	mg/l	>90%
pPB1850-1970	poly-his	>10 mg/l (insoluble)	>90%
pPB1850-2070	poly-his	>10 mg/l (insoluble)	>90%
pPB1750-1970(c)	poly-his	>10 mg/l (insoluble)	>90%
pPB1750-1970(n)	poly-his	>10 mg/l (insoluble)	>90%
aClones in italics are clones currently utilized to purify recombinant protein from each selected interval.

EXAMPLE 19

Identification, Purification and Induction of Neutralizing Antibodies Against Recombinant C. difficile Toxin B Protein

[0419] To determine whether recombinant toxin B polypeptide fragments can generate neutralizing antibodies, typically animals would first be immunized with recombinant proteins and anti-recombinant antibodies are generated. These anti-recombinant protein antibodies are then tested for neutralizing ability in vivo or in vitro. Depending on the immunogenic nature of the recombinant polypeptide, the generation of high-titer antibodies against that protein may take several months. To accelerate this process and identify which recombinant polypeptide(s) may be the best candidate to generate neutralizing antibodies, depletion studies were performed. Specifically, recombinant toxin B polypeptide were pre-screened by testing whether they have the ability to bind to protective antibodies from a CTB antibody preparation and hence deplete those antibodies of their neutralizing capacity. Those recombinant polypeptides found to bind CTB, were then utilized to generate neutralizing antibodies. This Example involved: a) identification of recombinant sub-regions within toxin B to which neutralizing antibodies bind; b) identification of toxin B sub-region specific antibodies that neutralize toxin B in vivo; and c) generation and evaluation of antibodies reactive to recombinant toxin B polypeptides.

[0420] a) Identification of Recombinant Sub-Regions Within Toxin B to Which Neutralizing Antibodies Bind

[0421] Sub-regions within toxin B to which neutralizing antibodies bind were identified by utilizing recombinant toxin B proteins to deplete protective antibodies from a polyclonal pool of antibodies against native C. difficile toxin B. An in vivo assay was developed to evaluate protein preparations for the ability to bind neutralizing antibodies. Recombinant proteins were first pre-mixed with antibodies directed against native toxin B (CTB antibody; see Example 8) and allowed to react for one hour at 37° C. Subsequently, C. difficile toxin B (CTB; Tech Lab) was added at a concentration lethal to hamsters and incubated for another hour at 37° C. After incubation this mixture was injected intraperitoneally (IP) into hamsters. If the recombinant polypeptide contains neutralizing epitopes, the CTB antibodies will lose its ability to protect the hamsters against death from CTB. If partial or complete protection occurs with the CTB antibody-recombinant mixture, that recombinant contains only weak or non-neutralizing epitopes of toxin B. This assay was performed as follows.

[0422] Antibodies against CTB were generated in egg laying Leghorn hens as described in Example 8. The lethal dosage (LD100) of C. difficile toxin B when delivered I.P. into 40 g female Golden Syrian hamsters (Charles River) was determined to be 2.5 to 5 μg. Antibodies generated against CTB and purified by PEG precipitation could completely protect the hamsters at the I.P. dosage determined above. The minimal amount of CTB antibody needed to afford good protection against 5 μg of CTB when injected I.P. into hamsters was also determined (1× PEG prep). These experiments defined the parameters needed to test whether a given recombinant protein could deplete protective CTB antibodies.

[0423] The cloned regions tested for neutralizing ability cover the entire toxin B gene and were designated as Intervals (INT) 1 through 5 (see FIG. 19). Approximately equivalent final concentrations of each recombinant polypeptide were tested. The following recombinant polypeptides were used: 1) a mixture of intervals 1 and 2 (INT-1, 2); 2) a mixture of Intervals 4 and 5 (INT-4, 5) and 3) Interval 3 (INT-3). Recombinant proteins (each at about 1 mg total protein) were first preincubated with a final CTB antibody concentration of 1× [i.e., pellet dissolved in original yolk volume as described in Example 1(c)] in a final volume of 5 mls for 1 hour at 37° C. Twenty-five μg of CTB (at a concentration of 5 μg/ml), enough CTB to kill 5 hamsters, was then added and the mixture was then incubated for 1 hour at 37° C. Five, 40 g female hamsters (Charles River) in each treatment group were then each given 1 ml of the mixture I.P. using a tuberculin syringe with a 27 gauge needle. The results of this experiment are shown in Table 24.

TABLE 24
Binding Of Neutralizing Antibodies By INT 3 Protein
	Number Of	Number Of
Treatment Group1	Animals Alive	Animals Dead
CTB antibodies	3	2
CTB antibodies + INT1, 2	3	2
CTB antibodies + INT4, 5	3	2
CTB antibodies + INT 3	0	5
1C. difficile toxin B (CTB) was added to each group.

[0424] As shown in Table 24, the addition of recombinant proteins from INT-1, 2 or INT-4, 5 had no effect on the in vivo protective ability of the CTB antibody preparation compared to the CTB antibody preparation alone. In contrast, INT-3 recombinant polypeptide was able to remove all of the toxin B neutralizing ability of the CTB antibodies as demonstrated by the death of all the hamsters in that group.

[0425] The above experiment was repeated, using two smaller expressed fragments (pMB 1750-1970 and pMB 1970-2360, see FIG. 19) comprising the INT-3 domain to determine if that domain could be further subdivided into smaller neutralizing epitopes. In addition, full-length INT-3 polypeptide expressed as a nickel tagged protein (pPB1750-2360) was tested for neutralizing ability and compared to the original INT-3 expressed MBP fusion (pMB1750-2360). The results are shown in Table 25.

TABLE 25
Removal Of Neutralizing Antibodies By Repeat Containing Proteins
	Number Of	Number Of
Treatment Group1	Animals Alive	Animals Dead
CTB antibodies	5	0
CTB antibodies + pPB1750-2360	0	5
CTB antibodies + pMB1750-2360	0	5
CTB antibodies + pMB1970-2360	3	2
CTB antibodies + pMB1750-1970	2	3
1C. difficile toxin B (CTB) was added to each group.

[0426] The results summarized in Table 25 indicate that the smaller polypeptide fragments within the INT-3 domain, pMB1750-1970 and pMB1970-2360, partially lose the ability to bind to and remove neutralizing antibodies from the CTB antibody pool. These results demonstrate that the full length INT-3 polypeptide is required to completely deplete the CTB antibody pool of neutralizing antibodies. This experiment also shows that the neutralization epitope of INT-3 can be expressed in alternative vector systems and the results are independent of the vector utilized or the accompanying fusion partner.

[0427] Other Interval 3 specific proteins were subsequently tested for the ability to remove neutralizing antibodies within the CTB antibody pool as described above. The Interval 3 specific proteins used in these studies are summarized in FIG. 23. In FIG. 23 the following abbreviations are used: pP refers to the pET23 vector; pM refers to the pMALc vector; B refers to toxin B; the numbers refer to the amino acid interval expressed in the clone. The solid black ovals represent the MBP; and HHH represents the poly-histidine tag.

[0428] Only recombinant proteins comprising the entire toxin B repeat domain (pMB1750-2360, pPB1750-2360 and pPB1850-2360) can bind and completely remove neutralizing antibodies from the CTB antibody pool. Recombinant proteins comprising only a portion of the toxin B repeat domain were not capable of completely removing neutralizing antibodies from the CTB antibody pool (pMB1750-1970 and pMB1970-2360 could partially remove neutralizing antibodies while pMB1850-1970 and pPB1850-2070 failed to remove any neutralizing antibodies from the CTB antibody pool).

[0429] The above results demonstrate that only the complete ligand binding domain (repeat region) of the toxin B gene can bind and completely remove neutralizing antibodies from the CTB antibody pool. These results demonstrate that antibodies directed against the entire toxin B repeat region are necessary for in vivo toxin neutralization (see FIG. 23; only the recombinant proteins expressed by the pMB1750-2360, pPB1750-2360 and pPB1850-2360 vectors are capable of completely removing the neutralizing antibodies from the CTB antibody pool).

[0430] These results represent the first indication that the entire repeat region of toxin B would be necessary for the generation of antibodies capable of neutralizing toxin B, and that sub-regions may not be sufficient to generate maximal titers of neutralizing antibodies.

[0431] b) Identification of Toxin B Sub-Region Specific Antibodies That Neutralize Toxin B in vivo

[0432] To determine if antibodies directed against the toxin B repeat region are sufficient for neutralization, region specific antibodies within the CTB antibody preparation were affinity purified, and tested for in vivo neutralization. Affinity columns containing recombinant toxin B repeat proteins were made as described below. A separate affinity column was prepared using each of the following recombinant toxin B repeat proteins: pPB1750-2360, pPB1850-2360, pMB1750-1970 and pMB1970-2360.

[0433] For each affinity column to be made, four ml of PBS-washed Actigel resin (Sterogene) was coupled overnight at room temperature with 5-10 mg of affinity purified recombinant protein (first extensively dialyzed into PBS) in 15 ml tubes (Falcon) containing a 1/10 final volume Ald-coupling solution (1 M sodium cyanoborohydride). Aliquots of the supernatants from the coupling reactions, before and after coupling, were assessed by Coomassie staining of 7.5% SDS-PAGE gels. Based on protein band intensities, in all cases greater than 30% coupling efficiencies were estimated. The resins were poured into 10 ml columns (BioRad), washed extensively with PBS, pre-eluted with 4M guanidine-HCl (in 10 mM Tris-HCl, pH 8.0) and reequilibrated in PBS. The columns were stored at 4° C.

[0434] Aliquots of a CTB IgY polyclonal antibody preparation (PEG prep) were affinity purified on each of the four columns as described below. The columns were hooked to a UV monitor (ISCO), washed with PBS and 40 ml aliquots of a 2× PEG prep (filter sterilized using a 0.45μ filter) were applied. The columns were washed with PBS until the baseline value was re-established. The columns were then washed with BBStween to elute nonspecifically binding antibodies, and reequilibrated with PBS. Bound antibody was eluted from the column in 4M guanidine-HCl (in 10 mM Tris-HCl, pH 8.0). The eluted antibody was immediately dialyzed against a 100-fold excess of PBS at 4° C. for 2 hrs. The samples were then dialyzed extensively against at least 2 changes of PBS, and affinity purified antibody was collected and stored at 4° C. The antibody preparations were quantified by UV absorbance. The elution volumes were in the range of 4-8 ml. All affinity purified stocks contained similar total antibody concentrations, ranging from 0.25-0.35% of the total protein applied to the columns.

[0435] The ability of the affinity purified antibody preparations to neutralize toxin B in vivo was determined using the assay outlined in a) above. Affinity purified antibody was diluted 1:1 in PBS before testing. The results are shown in Table 26.

TABLE 26
Neutralization Of Toxin B By Affinity Purified Antibodies
	Number
	Animals	Number Animals
Treatment groupa	Aliveb	Deadb
Preimmune1	0	5
CTB1 ; 400 μg	5	0
CTB (AP on pPB1750-2360);2 875 μg	5	0
CTB (AP on pMB1750-1970);2 875 μg	5	0
CTB (AP on pMB1970-2360);2 500 μg	5	0
aC. difficile toxin B (CTB) (Tech Lab; at 5 μg/ml, 25 μg total) at lethal concentration to hamsters is added to antibody and incubated for one hour at 37° C. After incubation this mixture is injected intraperitoneally (IP) into hamsters. Each treatment group received toxin premixed with antibody raised against the indicated protein, as either: 14X antibody PEG prep or 2affinity purified (AP) antibody (from CTB PEG prep, on
# indicated columns). The amount of specific antibody in each prep is indicated; the amount is directly determined for affinity purified preps and is estimated for the 4X CTB as described in Example 15.
bThe numbers in each group represent numbers of hamsters dead or alive, 2 hr post IP administration of toxin/antibody mixture.

[0436] In all cases similar levels of toxin neutralization was observed, such that lethality was delayed in all groups relative to preimmune controls. This result demonstrates that antibodies reactive to the repeat region of the toxin B gene are sufficient to neutralize toxin B in vivo. The hamsters will eventually die in all groups, but this death is maximally delayed with the CTB PEG prep antibodies. Thus neutralization with the affinity purified (AP) antibodies is not as complete as that observed with the CTB prep before affinity chromatography. This result may be due to loss of activity during guanidine denaturation (during the elution of the antibodies from the affinity column) or the presence of antibodies specific to other regions of the toxin B gene that can contribute to toxin neutralization (present in the CTB PEG prep).

[0437] The observation that antibodies affinity purified against the non-overlapping pMB1750-1970 and pMB1970-2360 proteins neutralized toxin B raised the possibility that either 1) antibodies specific to repeat sub-regions are sufficient to neutralize toxin B or 2) sub-region specific proteins can bind most or all repeat specific antibodies present in the CTB polyclonal pool. This would likely be due to conformational similarities between repeats, since homology in the primary amino acid sequences between different repeats is in the range of only 25-75% [Eichel-Streiber, et al. (1992) Molec. Gen. Genetics 233:260]. These possibilities were tested by affinity chromatography.

[0438] The CTB PEG prep was sequentially depleted 2× on the pMB1750-1970 column; only a small elution peak was observed after the second chromatography, indicating that most reactive antibodies were removed. This interval depleted CTB preparation was then chromatographed on the pPB1850-2360 column; no antibody bound to the column. The reactivity of the CTB and CTB (pMB1750-1970 depleted) preps to pPB1750-2360, pPB1850-2360, pMB1750-1970 and pMB1970-2360 proteins was then determined by ELISA using the protocol described in Example 13(c). Briefly, 96-well microtiter plates (Falcon, Pro-Bind Assay Plates) were coated with recombinant protein by adding 100 μl volumes of protein at 1-2 μg/ml in PBS containing 0.005% thimerisol to each well and incubating overnight at 4° C. The next morning, the coating suspensions were decanted and the wells were washed three times using PBS. In order to block non-specific binding sites, 100 μl of 1.0% BSA (Sigma) in PBS (blocking solution) was then added to each well, and the plates were incubated for 1 hr. at 37° C. The blocking solution was decanted and duplicate samples of 150 μl of diluted antibody was added to the first well of a dilution series. The initial testing serum dilution was (1/200 for CTB prep, (the concentration of depleted CTB was standardized by OD280) in blocking solution containing 0.5% Tween 20, followed by 5-fold serial dilutions into this solution. This was accomplished by serially transferring 30 μl aliquots to 120 μl buffer, mixing, and repeating the dilution into a fresh well. After the final dilution, 30 μl was removed from the well such that all wells contained 120 μl final volume. A total of 5 such dilutions were performed (4 wells total). The plates were incubated for 1 hr at 37° C. Following this incubation, the serially diluted samples were decanted and the wells were washed three times using PBS containing 0.5% Tween 20 (PBST), followed by two 5 min washes using BBS-Tween and a final three washes using PBST. To each well, 100 μl of 1/1000 diluted secondary antibody [rabbit anti-chicken IgG alkaline phosphatase (Sigma) diluted in blocking solution containing 0.5% Tween 20] was added, and the plate was incubated 1 hr at 37° C. The conjugate solutions were decanted and the plates were washed 6 times in PBST, then once in 50 mM Na2CO3, 10 mM MgCl2, pH 9.5. The plates were developed by the addition of 100 μl of a solution containing 1 mg/ml para-nitro phenyl phosphate (Sigma) dissolved in 50 mM Na2CO3, 10 mM MgCl2, pH 9.5 to each well. The plates were then incubated at room temperature in the dark for 5-45 min. The absorbency of each well was measured at 410 nm using a Dynatech MR 700 plate reader.

[0439] As predicted by the affinity chromatography results, depletion of the CTB prep on the pMB1750-1970 column removed all detectable reactivity to the pMB1970-2360 protein. The reciprocal purification of a CTB prep that was depleted on the pMB1970-2360 column yielded no bound antibody when chromatographed on the pMB1750-1970 column. These results demonstrate that all repeat reactive antibodies in the CTB polyclonal pool recognize a conserved structure that is present in non-overlapping repeats. Although it is possible that this conserved structure represents rare conserved linear epitopes, it appears more likely that the neutralizing antibodies recognize a specific protein conformation. This conclusion was also suggested by the results of Western blot hybridization analysis of CTB reactivity to these recombinant proteins.

[0440] Western blots of 7.5% SDS-PAGE gels, loaded and electrophoresed with defined quantities of each recombinant protein, were probed with the CTB polyclonal antibody preparation. The blots were prepared and developed with alkaline phosphatase as described in Example 3. The results are shown in FIG. 24.

[0441] FIG. 24 depicts a comparison of immunoreactivity of IgY antibody raised against either native or recombinant toxin B antigen. Equal amounts of pMB1750-1970 (lane 1), pMB1970-2360 (lane 2), pPB1850-2360 (lane 3) as well as a serial dilution of pPB1750-2360 (lanes 4-6 comprising 1×, 1/10× and 1/100× amounts, respectively) proteins were loaded in duplicate and resolved on a 7.5% SDS-PAGE gel. The gel was blotted and each half was hybridized with PEG prep IgY antibodies from chickens immunized with either native CTB or pPB1750-2360 protein. Note that the full-length pMB1750-1970 protein was identified only by antibodies reactive to the recombinant protein (arrows).

[0442] Although the CTB prep reacts with the pPB1750-2360, pPB1850-2360, and pMB1970-2360 proteins, no reactivity to the pMB1750-1970 protein was observed (FIG. 24). Given that all repeat reactive antibodies can be bound by this protein during affinity chromatography, this result indicates that the protein cannot fold properly on Western blots. Since this eliminates all antibody reactivity, it is unlikely that the repeat reactive antibodies in the CTB prep recognize linear epitopes. This may indicate that in order to induce protective antibodies, recombinant toxin B protein will need to be properly folded.

[0443] c) Generation and Evaluation of Antibodies Reactive to Recombinant Toxin B Polypeptides

[0444] i) Generation of Antibodies Reactive to Recombinant Toxin B Proteins

[0445] Antibodies against recombinant proteins were generated in egg laying Leghorn hens as described in Example 13. Antibodies were raised [using Freunds adjuvant (Gibco) unless otherwise indicated] against the following recombinant proteins: 1) a mixture of Interval 1+2 proteins (see FIG. 18); 2) a mixture of interval 4 and 5 proteins (see FIG. 18); 3) pMB1970-2360 protein; 4) pPB1750-2360 protein; 5) pMB1750-2360; 6) pMB1750-2360 [Titermax adjuvant (Vaxcell)]; 7) pMB1750-2360 [Gerbu adjuvant (Biotech)]; 8) pMBp1750-2360 protein; 9) pPB1850-2360; and 10) pMB1850-2360.

[0446] Chickens were boosted at least 3 times with recombinant protein until ELISA reactivity [using the protocol described in b) above with the exception that the plates were coated with pPB1750-2360 protein] of polyclonal PEG preps was at least equal to that of the CTB polyclonal antibody PEG prep. ELISA titers were determined for the PEG preps from all of the above immunogens and were found to be comparable ranging from 1:12500 to 1:62500. High titers were achieved in all cases except in 6) pMB1750-2360 in which strong titers were not observed using the Titermax adjuvant, and this preparation was not tested further.

[0447] ii) Evaluation of Antibodies Reactive to Recombinant Proteins by Western Blot Hybridization

[0448] Western blots of 7.5% SDS-PAGE gels, loaded and electrophoresed with defined quantities of recombinant protein (pMB1750-1970, pPB1850-2360, and pMB1970-2360 proteins and a serial dilution of the pPB1750-2360 to allow quantification of reactivity), were probed with the CTB, pPB1750-2360, pMB1750-2360 and pMB1970-2360 polyclonal antibody preparations (from chickens immunized using Freunds adjuvant). The blots were prepared and developed with alkaline phosphatase as described above in b).

[0449] As shown in FIG. 24, the CTB and pMB1970-2360 preps reacted strongly with the pPB1750-2360, pPB1850-2360, and pMB1970-2360 proteins while the pPB1750-2360 and pMB1970-2360 (Gerbu) preparations reacted strongly with all four proteins. The Western blot reactivity of the pPB1750-2360 and pMB1970-2360 (Gerbu) preparations were equivalent to that of the CTB preparation, while reactivity of the pMB1970-2360 preparation was <10% that of the CTB prep. Despite equivalent ELISA reactivities only weak reactivity (approximately 1%) to the recombinant proteins were observed in PEG preps from two independent groups immunized with the pMB1750-2360 protein and one group immunized with the pMB1750-2360 preparation using Freunds adjuvant.

[0450] Affinity purification was utilized to determine if this difference in immunoreactivity by Western blot analysis reflects differing antibody titers. Fifty ml 2× PEG preparations from chickens immunized with either pMB1750-2360 or pMB1970-2360 protein were chromatographed on the pPB1750-2360 affinity column from b) above, as described. The yield of affinity purified antibody (% total protein in preparation) was equivalent to the yield obtained from a CTB PEG preparation in b) above. Thus, differences in Western reactivity reflect a qualitative difference in the antibody pools, rather than quantitative differences., These results demonstrate that certain recombinant proteins are more effective at generating high affinity antibodies (as assayed by Western blot hybridization).

[0451] iii) In vivo Neutralization of Toxin B Using Antibodies Reactive to Recombinant Protein

[0452] The in vivo hamster model [described in Examples 9 and 14(b)] was utilized to assess the neutralizing ability of antibodies raised against recombinant toxin B proteins. The results from three experiments are shown below in Tables 27-29.

TABLE 27
In Vivo Neutralization Of Toxin B
Treatment Groupa	Number Animals Aliveb	Number Animals Deadb
Preimmune	0	5
CTB	5	0
INT 1 + 2	0	5
INT 4 + 5	0	5
pMB1750-2360	0	5
pMB1970-2360	0	5
pPB1750-2360	5	0
aC. difficile toxin B (CTB) (at 5 μg/ml; 25 μg total; Tech Lab) at lethal concentration to hamsters is added to antibody and incubated for one hour at 37° C. After incubation this mixture is injected intraperitoneally (IP) into hamsters. Each treatment group received toxin premixed with antibody raised against the indicated protein, as a 4X antibody PEG prep.
bThe numbers in each group represent numbers of hamsters dead or alive, 2 hours post IP administration of toxin/antibody mixture.

[0453]

TABLE 28
In Vivo Neutralization Of Toxin B Using Affinity Purified Antibodies
	Number Animals	Number Animals
Treatment Groupa	Aliveb	Deadb
Preimmune(1)	0	5
CTB(1)	5	0
pPB1750-2360(1)	5	0
1.5 mg anti-pMB1750-2360(2)	1	4
1.5 mg anti-pMB1970-2360(2)	0	5
300 μg anti-CTB(2)	5	0
a C. difficile toxin B (CTB) (at 5 μg/ml; 25 μg total; Tech Lab) at lethal concentration to hamsters is added to antibody and incubated for one hour at 37° C. After incubation, 1 ml of this mixture is injected intraperitoneally (IP) into hamsters. Each treatment group received toxin premixed with antibody raised against the indicated protein, as either (1)4X antibody PEG prep or (2)affinity purified antibody (on a pPB1750-2360 resin), either 1.5 mg/group (anti-pMB1750-2360 and
# anti-pMB1970-2360; used undiluted affinity purified antibody) or 350 μg/group (anti-CTB, repeat specific; used 1/5 diluted anti-CTB antibody).
b The numbers in each group represent numbers of hamsters dead or alive, 2 hr post-IP administration of toxin/antibody mixture.

[0454]

TABLE 29
Generation Of Neutralizing Antibodies Utilizing The Gerbu Adjuvant
	Number Animals	Number Animals
Treatment Groupa	Aliveb	Deadb
Preimmune	0	5
CTB	5	0
pMB1970-2360	0	5
pMB1850-2360	0	5
pPB1850-2360	0	5
pMB1750-2360 (Gerbu adj)	5	0
a C. difficile toxin B (CTB) (Tech Lab) at lethal concentration to hamsters is added to antibody and incubated for one hour at 37° C. After incubation this mixture is injected intraperitoneally (IP) into hamsters. Each treatment group received toxin premixed with antibody raised against the indicated protein, as a 4X antibody PEG prep.
b The numbers in each group represent numbers of hamsters dead or alive, 2 hrs post IP administration of toxin/antibody mixture.

[0455] The ability of each immunogen to neutralize toxin B in vivo has been compiled and is shown in Table 30. As predicted from the recombinant protein-CTB premix studies (Table 24) only antibodies to Interval 3 (1750-2366) and not the other regions of toxin B (i.e., intervals 1-5) are protective. Unexpectedly, antibodies generated to INT-3 region expressed in pMAL vector (pMB1750-2360 and pMpB1750-2360) using Freunds adjuvant were non-neutralizing. This observation is reproducible, since no neutralization was observed in two independent immunizations with pMB1750-2360 and one immunization with pMpB1750-2360. The fact that 5× quantities of affinity purified toxin B repeat specific antibodies from pMB1750-2360 PEG preps cannot neutralize toxin B while 1× quantities of affinity purified anti-CTB antibodies can (Table 28) demonstrates that the differential ability of CTB antibodies to neutralize toxin B is due to qualitative rather than quantitative differences in these antibody preparations. Only when this region was expressed in an alternative vector (pPB1750-2360) or using an alternative adjuvant with the pMB1750-2360 protein were neutralizing antibodies generated. Importantly, antibodies raised using Freunds adjuvant to pPB1850-2360, which contains a fragment that is only 100 amino acids smaller than recombinant pPB1750-2360, are unable to neutralize toxin B in vivo (Table 27); note also that the same vector is used for both pPB1850-2360 and pPB1750-2360.

[0456] The pPB1750-2360 antibody pool confers significant in vivo protection, equivalent to that obtained with the affinity purified CTB antibodies. This correlates with the observed high affinity of this antibody pool (relative to the pMB1750-2360 or pMB1970-2360 pools) as assayed by Western blot analysis (FIG. 24). These results provide the first demonstration that in vivo neutralizing antibodies can be induced using recombinant toxin B protein as immunogen.

[0457] The failure of high concentrations of antibodies raised against the pMB1750-2360 protein (using Freunds adjuvant) to neutralize, while the use of Gerbu adjuvant and pMB1750-2360 protein generates a neutralizing response, demonstrates that conformation or presentation of this protein is essential for the induction of neutralizing antibodies. These results are consistent with the observation that the neutralizing antibodies produced when native CTB is used as an immunogen appear to recognize conformational epitopes [see section b) above]. This is the first demonstratration that the conformation or presentation of recombinant toxin B protein is essential to generate high titers of neutralizing antibodies.

TABLE 30
In Vivo Neutralization Of Toxin B
			Antigen	In vivo
		Tested	Utilized	Neutrali-
Immunogen	Adjuvant	Preparationa	For AP	zationb
Preimmune	NA1	PEG	NA	no
CTB (native)	Titermax	PEG	NA	yes
CTB (native)	Titermax	AP	pPB1750-2360	yes
CTB (native)	Titermax	AP	pPB1850-2360	yes
CTB (native)	Titermax	AP	pPB1750-1970	yes
CTB (native)	Titermax	AP	pPB1970-2360	yes
pMB1750-2360	Freunds	PEG	NA	no
pMB1750-2360	Freunds	AP	pPB1750-2360	no
pMB1750-2360	Gerbu	PEG	NA	yes
pMB1970-2360	Freunds	PEG	NA	no
pMB1970-2360	Freunds	AP	pPB1750-2360	no
pPB1750-2360	Freunds	PEG	NA	yes
pPB1850-2360	Freunds	PEG	NA	no
pMB1850-2360	Freunds	PEG	NA	no
INT 1 + 2	Freunds	PEG	NA	no
INT 4 + 5	Freunds	PEG	NA	no
aEither PEG preparation (PEG) or affinity purified antibodies (AP).
b‘Yes’ denotes complete neutralization (0/5 dead) while ‘no’ denotes no neutralization (5/5 dead) of toxin B, 2 hours post-administration of mixture.
1‘NA’ denotes not applicable.

EXAMPLE 20

Determination of Quantitative and Qualitative Differences Between pMB1750-2360, pMB1750-2360 (Gerbu) or pPB1750-2360 IgY Polyclonal Antibody Preparations

[0458] In Example 19, it was demonstrated that toxin B neutralizing antibodies could be generated using specific recombinant toxin B proteins (pPB1750-2360) or specific adjuvants. Antibodies raised against pMB1750-2360 were capable of neutralizing the enterotoxin effect of toxin B when the recombinant protein was used to immunize hens in conjunction with the Gerbu adjuvant, but not when Freunds adjuvant was used. To determine the basis for these antigen and adjuvant restrictions, toxin B-specific antibodies present in the neutralizing and non-neutralizing PEG preparations were isolated by affinity chromatography and tested for qualitative or quantitative differences. The example involved a) purification of anti-toxin B specific antibodies from pMB1750-2360 and pPB1750-2360 PEG preparations and b) in vivo neutralization of toxin B using the affinity purified antibody.

[0459] a) Purification of specific Antibodies from pMB1750-2360 and pPB1750-2360 PEG Preparations

[0460] To purify and determine the concentration of specific antibodies (expressed as the percent of total antibody) within the pPB1750-2360 (Freunds and Gerbu) and pPB1750-2360 PEG preparations, defined quantities of these antibody preparations were chromatographed on an affinity column containing the entire toxin B repeat region (pPB1750-2360). The amount of affinity purified antibody was then quantified.

[0461] An affinity column containing the recombinant toxin B repeat protein, pPB1750-2360, was made as follows. Four ml of PBS-washed Actigel resin (Sterogene) was coupled with 5 mg of pPB1750-2360 affinity purified protein (dialyzed into PBS; estimated to be greater than 95% full length fusion protein) in a 15 ml tube (Falcon) containing 1/10 final volume Ald-coupling solution (1M sodium cyanoborohydride). Aliquots of the supernatant from the coupling reactions, before and after coupling, were assessed by Coomassie staining of 7.5% SDS-PAGE gels. Based on protein band intensities, greater than 95% (approximately 5 mg) of recombinant protein was coupled to the resin. The coupled resin was poured into a 10 ml column (BioRad), washed extensively with PBS, pre-eluted with 4M guanidine-HCl (in 10 mM Tris-HCl, pH 8.0; 0.005% thimerosal) and re-equilibrated in PBS and stored at 4° C.

[0462] Aliquots of pMB1750-2360, pMB1750-2360 (Gerbu) or pPB1750-2360 IgY polyclonal antibody preparations (PEG preps) were affinity purified on the above column as follows. The column was attached to an UV monitor (ISCO), and washed with PBS. Forty ml aliquots of 2× PEG preps (filter sterilized using a 0.45μ filter and quantified by OD280 before chromatography) was applied. The column was washed with PBS until the baseline was re-established (the column flow-through was saved), washed with BBSTween to elute nonspecifically binding antibodies and re-equilibrated with PBS. Bound antibody was eluted from the column in 4M guanidine-HCl (in 10 mM Tris-HCL, pH 8.0, 0.005% thimerosal) and the entire elution peak collected in a 15 ml tube (Falcon). The column was re-equilibrated, and the column eluate re-chromatographed as described above. The antibody preparations were quantified by UV absorbance (the elution buffer was used to zero the spectrophotometer). Approximately 10 fold higher concentrations of total purified antibody was obtained upon elution of the first chromatography pass relative to the second pass. The low yield from the second chromatography pass indicated that most of the specific antibodies were removed by the first round of chromatography.

[0463] Pools of affinity purified specific antibodies were prepared by dialysis of the column elutes after the first column chromatography pass for the pMB1750-2360, pMB1750-2360 (Gerbu) or pPB1750-2360 IgY polyclonal antibody preparations. The elutes were collected on ice and immediately dialyzed against a 100-fold volume of PBS at 4° C. for 2 hrs. The samples were then dialyzed against 3 changes of a 65-fold volume of PBS at 4° C. Dialysis was performed for a minimum of 8 hrs per change of PBS. The dialyzed samples were collected, centrifuged to remove insoluble debris, quantified by OD280, and stored at 4° C.

[0464] The percentage of toxin B repeat-specific antibodies present in each preparation was determined using the quantifications of antibody yields from the first column pass (amount of specific antibody recovered after first pass/total protein loaded). The yield of repeat-specific affinity purified antibody (expressed as the percent of total protein in the preparation) in: 1) the pMB1750-2360 PEG prep was approximately 0.5%, 2) the pMB1750-2360 (Gerbu) prep was approximately 2.3%, and 3) the pPB1750-2360 prep was approximately 0.4%. Purification of a CTB IgY polyclonal antibody preparation on the same column demonstrated that the concentration of toxin B repeat specific antibodies in the CTB preparation was 0.35%.

[0465] These results demonstrate that 1) the use of Gerbu adjuvant enhanced the titer of specific antibody produced against the pMB1750-2360 protein 5-fold relative to immunization using Freunds adjuvant, and 2) the differences seen in the in vivo neutralization ability of the pMB1750-2360 (not neutralizing) and pPB1750-2360 (neutralizing) and CTB (neutralizing) PEG preps seen in Example 19 was not due to differences in the titers of repeat-specific antibodies in the three preparations because the titer of repeat-specific antibody was similar for all three preps; therefore the differing ability of the three antibody preparations to neutralize toxin B must reflect qualitative differences in the induced toxin B repeat-specific antibodies. To confirm that qualitative differences exist between antibodies raised in hens immunized with different recombinant proteins and/or different adjuvants, the same amount of affinity purified anti-toxin B repeat (aa 1870-2360 of toxin B) antibodies from the different preparations was administered to hamsters using the in vivo hamster model as described below.

[0466] b) In vivo Neutralization of Toxin B Using Affinity Purified Antibody

[0467] The in vivo hamster model was utilized to assess the neutralizing ability of the affinity purified antibodies raised against recombinant toxin B proteins purified in (a) above. As well, a 4× IgY PEG preparation from a second independent immunization utilizing the pPB1750-2360 antigen with Freunds adjuvant was tested for in vivo neutralization. The results are shown in Table 31.

[0468] The results shown in Table 31 demonstrate that:

[0469] 1) as shown in Example 19 and reproduced here, 1.5 mg of affinity purified antibody from pMB1750-2360 immunized hens using Freunds adjuvant does not neutralize toxin B in vivo. However, 300 μg of affinity purified antibody from similarly immunized hens utilizing Gerbu adjuvant demonstrated complete neutralization of toxin B in vivo. This demonstrates that Gerbu adjuvant, in addition to enhancing the titer of antibodies reactive to the pMB1750-2360 antigen relative to Freunds adjuvant (demonstrated in (a) above), also enhances the yield of neutralizing antibodies to this antigen, greater than 5 fold.

[0470] 2) Complete in vivo neutralization of toxin B was observed with 1.5 mg of affinity purified antibody from hens immunized with pPB1750-2360 antigen, but not with pMB1750-2360 antigen, when Freunds adjuvant was used. This demonstrates, using standardized toxin B repeat-specific antibody concentrations, that neutralizing antibodies were induced when pPB1750-2360 but not pMB1750-2360 was used as the antigen with Freunds adjuvant.

[0471] 3) Complete in vivo neutralization was observed with 300 μg of pMB1750-2360 (Gerbu) antibody, but not with 300 μg of pPB1750-2360 (Freunds) antibody. Thus the pMB1750-2360 (Gerbu) antibody has a higher titer of neutralizing antibodies than the pPB1750-2360 (Freunds) antibody.

TABLE 31
In vivo Neutralization Of Toxin B Using Affinity Purified Antibodies
	Number Animals	Number Animals
Treatment Groupa	Aliveb	Deadb
Preimmune1	0	5
CTB (300 μg)2	5	0
CTB (100 μg)2	1	4
pMB1750-2360 (G) (5 mg)2	5	0
pMB1750-2360 (G) (1.5 mg)2	5	0
pMB1750-2360 (G) (300 μg)2	5	0
pMB1750-2360 (F) (1.5 mg)2	0	5
pPB1750-2360 (F) (1.5 mg)2	5	0
pPB1750-2360 (F) (300 μg)2	1	4
CTB (100 μg)3	2	3
pPB1750-2360 (F) (500 μg)1	5	0
aC. difficile toxin B (CTB) (Tech Lab) at lethal concentration to hamsters (25 μg) was added to the antibody (amount of specific antibody is indicated) and incubated for one hour at 37° C. After incubation, this mixture was injected IP into hamsters (1/5 total mix injected per hamster). Each treatment group received toxin premixed with antibody raised against the indicated protein (G = gerbu adjuvant, F = Freunds adjuvant).
1indicates the antibody was a 4X IgY PEG prep;
2indicates the antibody was affinity purified on a pPB1850-2360 resin; and
3indicates that the antibody was a 1X IgY PEG prep.
bThe numbers in each group represent numbers of hamsters dead or alive, 2 hrs post IP administration of toxin/antibody mixture.

[0472] 4) Complete neutralization of toxin B was observed using 300 μg of CTB antibody [affinity purified (AP)] but not 100 μg CTB antibody (AP or PEG prep). This demonstrates that greater than 100 μg of toxin B repeat-specific antibody (anti-CTB) is necessary to neutralize 25 μg toxin B in vivo in this assay, and that affinity purified antibodies specific to the toxin B repeat interval neutralize toxin B as effectively as the PEP prep of IgY raised against the entire CTB protein (shown in this assay).

[0473] 5) As was observed with the initial pPB1750-2360 (IgY) PEG preparation (Example 19), complete neutralization was observed with a IgY PEG preparation isolated from a second independent group of pPB1750-2360 (Freunds) immunized hens. This demonstrates that neutralizing antibodies are reproducibly produced when hens are immunized with pPB1750-2360 protein utilizing Freunds adjuvant.

EXAMPLE 21

Diagnostic Enzyme Immunoassays for C. difficile Toxins A and B

[0474] The ability of the recombinant toxin proteins and antibodies raised against these recombinant proteins (described in the above examples) to form the basis of diagnostic assays for the detection of clostridial toxin in a sample was examined. Two immunoassay formats were tested to quantitatively detect C. difficile toxin A and toxin B from a biological specimen. The first format involved a competitive assay in which a fixed amount of recombinant toxin A or B was immobilized on a solid support (e.g., microtiter plate wells) followed by the addition of a toxin-containing biological specimen mixed with affinity-purified or PEG fractionated antibodies against recombinant toxin A or B. If toxin is present in a specimen, this toxin will compete with the immobilized recombinant toxin protein for binding to the anti-recombinant antibody thereby reducing the signal obtained following the addition of a reporter reagent. The reporter reagent detects the presence of antibody bound to the immobilized toxin protein.

[0475] In the second format, a sandwich immunoassay was developed using affinity-purified antibodies to recombinant toxin A and B. The affinity-purified antibodies to recombinant toxin A and B were used to coat microtiter wells instead of the recombinant polypeptides (as was done in the competitive assay format). Biological samples containing toxin A or B were then added to the wells followed by the addition of a reporter reagent to detect the presence of bound toxin in the well.

[0476] a) Competitive Immunoassay for the Detection of C. difficile Toxin

[0477] Recombinant toxin A or B was attached to a solid support by coating 96 well microtiter plates with the toxin protein at a concentration of 1 μg/ml in PBS. The plates were incubated overnight at 2-8° C. The following morning, the coating solutions were removed and the remaining protein binding sites on the wells were blocked by filling each well with a PBS solution containing 0.5% BSA and 0.05% Tween-20. Native C. difficile toxin A or B (Tech Lab) was diluted to 4 μg/ml in stool extracts from healthy Syrian hamsters (Sasco). The stool extracts were made by placing fecal pellets in a 15 ml centrifuge tube; PBS was added at 2 ml/pellet and the tube was vortexed to create a uniform suspension. The tube was then centrifuged at 2000 rpm for 5 min at room temperature. The supernantant was removed; this comprisies the stool extract. Fifty μl of the hamster stool extract was pipetted into each well of the microtiter plates to serve as the diluent for serial dilutions of the 4 μg/ml toxin samples. One hundred μl of the toxin samples at 4 μg/ml was pipetted into the first row of wells in the microtiter plate, and 50 μl aliquots were removed and diluted serially down the plate in duplicate. An equal volume of affinity purified anti-recombinant toxin antibodies [1 ng/well of anti-pMA1870-2680 antibody was used for the detection of toxin A; 0.5 ng/well of anti-pMB1750-2360(Gerbu) was used for the detection of toxin B] were added to appropriate wells, and the plates were incubated at room temperature for 2 hours with gentle agitation. Wells serving as negative control contained antibody but no native toxin to compete for binding.

[0478] Unbound toxin and antibody were removed by washing the plates 3 to 5 times with PBS containing 0.05% Tween-20. Following the wash step, 100 μl of rabbit anti-chicken IgG antibody conjugated to alkaline phosphatase (Sigma) was added to each well and the plates were incubated for 2 hours at room temperature. The plates were then washed as before to remove unbound secondary antibody. Freshly prepared alkaline phosphatase substrate (1 mg/ml p-nitrophenyl phosphate (Sigma) in 50 mM Na2CO3, pH 9.5; 10 mM MgCl2) was added to each well. Once sufficient color developed, the plates were read on a Dynatech MR700 microtiter plate reader using a 410 nm filter.

[0479] The results are summarized in Tables 32 and 33. For the results shown in Table 32, the wells were coated with recombinant toxin A protein (pMA1870-2680). The amount of native toxin A added (present as an addition to solubilized hamster stool) to a given well is indicated (0 to 200 ng). Antibody raised against the recombinant toxin A protein, pMA1870-2680, was affinity purified on the an affinity column containing pPA1870-2680 (described in Example 20). As shown in Table 32, the recombinant toxin A protein and affinity-purified antitoxin can be used for the basis of a competitive immunoassay for the detection of toxin A in biological samples.

[0480] Similar results were obtained using the recombinant toxin B, pPB1750-2360, and antibodies raised against pMB1750-2360(Gerbu). For the results shown in Table 33, the wells were coated with recombinant toxin B protein (pPB1750-2360). The amount of native toxin B added (present as an addition to solubilized hamster stool) to a given well is indicated (0 to 200 ng). Antibody raised against the recombinant toxin B protein, pMB1750-2360(Gerbu), was affinity purified on the an affinity column containing pPB1850-2360 (described in Example 20). As shown in Table 33, the recombinant toxin B protein and affinity-purified antitoxin can be used for the basis of a competitive immunoassay for the detection of toxin B in biological samples.

[0481] In this competition assay, the reduction is considered significant over the background levels at all points; therefore the assay can be used to detect samples containing less than 12.5 ng toxin A/well and as little as 50-100 ng toxin B/well.

TABLE 32
Competitive Inhibition Of Anti-C. difficile Toxin A By Native Toxin A
ng Toxin A/Well OD410 Readout
200             0.176
100             0.253
50              0.240
25              0.259
12.5            0.309
6.25            0.367
3.125           0.417
0               0.590

[0482]

TABLE 33
Competitive Inhibition Of Anti-C. difficile Toxin B By Native Toxin B
ug Toxin B/Well OD410 Readout
200             0.392
100             0.566
50              0.607
25              0.778
12.5            0.970
6.25            0.902
3.125           1.040
0               1.055

[0483] These competitive inhibition assays demonstrate that native C. difficile toxins and recombinant C. difficile toxin proteins can compete for binding to antibodies raised against recombinant C. difficile toxins demonstrating that these anti-recombinant toxin antibodies provide effective diagnostic reagents.

[0484] b) Sandwich Immunoassay for the Detection of C. difficile Toxin

[0485] Affinity-purified antibodies against recombinant toxin A or toxin B were immobilized to 96 well microtiter plates as follows. The wells were passively coated overnight at 4° C. with affinity purified antibodies raised against either pMA1870-2680 (toxin A) or pMB1750-2360(Gerbu) (toxin B). The antibodies were affinity purified as described in Example 20. The antibodies were used at a concentration of 1 μg/ml and 100 μl was added to each microtiter well. The wells were then blocked with 200 μl of 0.5% BSA in PBS for 2 hours at room temperature and the blocking solution was then decanted. Stool samples from healthy Syrian hamsters were resuspended in PBS, pH 7.4 (2 ml PBS/stool pellet was used to resuspend the pellets and the sample was centrifuged as described above). The stool suspension was then spiked with native C. difficile toxin A or B (Tech Lab) at 4 μg/ml. The stool suspensions containing toxin (either toxin A or toxin B) were then serially diluted two-fold in stool suspension without toxin and 50 μl was added in duplicate to the coated microtiter wells. Wells containing stool suspension without toxin served as the negative control.

[0486] The plates were incubated for 2 hours at room temperature and then were washed three times with PBS. One hundred μl of either goat anti-native toxin A or goat anti-native toxin B (Tech Lab) diluted 1:1000 in PBS containing 1% BSA and 0.05% Tween 20 was added to each well . The plates were incubated for another 2 hours at room temperature. The plates were then washed as before and 100 μl of alkaline phosphatase-conjugated rabbit anti-goat IgG (Cappel, Durham, N.C.) was added at a dilution of 1:1000. The plates were incubated for another 2 hours at room temperature. The plates were washed as before then developed by the addition of 100 μl/well of a substrate solution containing 1 mg/ml p-nitrophenyl phosphate (Sigma) in 50 mM Na2CO3, pH 9.5; 10 mM MgCl2. The absorbance of each well was measured using a plate reader (Dynatech) at 410 nm. The assay results are shown in Tables 34 and 35.

TABLE 34
C. difficile Toxin A Detection In Stool Using Affinity-Purified Antibodies
Against Toxin A
ng Toxin A/Well OD410 Readout
200             0.9
100             0.8
50              0.73
25              0.71
12.5            0.59
6.25            0.421
0               0

[0487]

TABLE 35
C. difficile                   Toxin B Detection In Stool Using Affinity-Purified Antibodies
Against Toxin B
ng Toxin B/Well OD410 Readout
200             1.2
100             0.973
50              0.887
25              0.846
12.5            0.651
6.25            0.431
0               0.004

[0488] The results shown in Tables 34 and 35 show that antibodies raised against recombinant toxin A and toxin B fragments can be used to detect the presence of C. difficile toxin in stool samples. These antibodies form the basis for a sensitive sandwich immunoassay which is capable of detecting as little as 6.25 ng of either toxin A or B in a 50 μl stool sample. As shown above in Tables 34 and 35, the background for this sandwich immunoassay is extremely low; therefore, the sensitivity of this assay is much lower than 6.25 ng toxin/well. It is likely that toxin levels of 0.5 to 1.0 μg/well could be detected by this assay.

[0489] The results shown above in Tables 32-35 demonstrate clear utility of the recombinant reagents in C. difficile toxin detection systems.

[0490] From the above it is clear that the present invention provides compositions and methods for effective therapy against clostridial toxin disease therapy. It is also contemplated that these antitoxins and recombinant toxin proteins be used for diagnostic purposes.

Claims

1. A composition comprising nucleic acid encoding a soluble fusion protein, said fusion protein comprising a non-toxin protein sequence and a portion of the Clostridium difficile toxin B sequence of SEQ ID NO:10, wherein said portion contains at least amino acids 1750-2360 of said Clostridium difficile toxin B sequence.

2. The composition of claim 1, wherein said nucleic acid is in a vector.

3. The composition of claim 2, wherein said vector is in a host cell.

4. The composition of claim 3, wherein said host cell is Escherichia coli.

5. A composition comprising a recombinant fusion protein in soluble form wherein said recombinant fusion protein comprises a non-toxin protein sequence and a portion of the Clostridium difficile toxin B sequence of SEQ ID NO:10, and wherein said portion contains at least amino acids 1750-2360 of said Clostridium difficile toxin B sequence.

6. The composition of claim 5, wherein said protein is soluble at a level greater than one milligram per liter of host cell culture.

7. The composition of claim 5, wherein said protein is soluble at a level of at least 6.5 milligrams per liter of host cell culture.

8. The composition of claim 5, wherein said protein is soluble at a level of at least 20 milligrams per liter of host cell culture.