1. Field of the Invention
The present invention generally relates to methods for detecting the presence or concentration of an analyte in a sample utilizing an immobilized first binding reagent and a second binding reagent which comprises a fusion protein having a reporter domain and a binding domain. The methods may be used to detect rapidly and with high sensitivity, the presence or progress of, e.g., infectious diseases, inflammatory diseases, autoimmune diseases and cancer.
2. Description of the Related Art
Assays detecting human antigen-specific antibodies are medically useful. However, the usefulness of existing simple immunoassay formats is limited by technical considerations such as sera antibodies to contaminants in insufficiently pure antigen, a problem likely exacerbated when antigen panels are screened to obtain clinically useful data.
Most immunoassays use bacterial-expressed proteins for detecting antigen-specific antibodies in human sera [2]. However, since such antigens do not carry post-translational modifications or may fold incorrectly, some immunoassays employ antigens produced in either yeast or insect cells. While these antigens may fold correctly and carry post-translational modifications, they will not carry either mammalian- or disease-specific posttranslational modifications. Tests employing bacterial-produced proteins can produce high backgrounds because it is difficult to completely eliminate or block serum antibodies reactive with trace amounts of bacterial contaminants present in most antigen preparations, even in pharmaceutical grade preparations [3]. The present invention overcomes these problems.
In one aspect, the present invention provides a method for detecting the presence or concentration of an analyte in a sample, said method comprising:
In another aspect, the present invention provides a method for monitoring the course of a disease in a patient having need of such monitoring, said method comprising:
In yet another aspect, the present invention provides a method for detecting the presence or concentration of an analyte in a sample, said method comprising, in any order:
In yet another aspect, the present invention provides a kit for detecting the presence or concentration of an analyte in a sample, said kit comprising:
The present invention provides a rapid, simple and highly sensitive assay to detect the presence or quantify the concentration of an analyte in a sample. Many analytes are within the scope of the present invention. For example, the present assays may be used to detect the presence or quantify the concentration of analytes of medical or biochemical interest, such as antibodies, proteins, antigens, carbohydrates, lipids, etc. The sample containing the analyte to be detected may be any body fluid where the analyte may be found, including but not limited to blood, saliva, ascites, urine, cerebrospinal fluid, amniotic fluid, sputum and gastric fluid.
The first binding reagent of the present invention should be capable of binding to the analyte of interest if present in the sample. As used herein, the term “binding” is intended to mean any interaction or association between the first binding reagent and the analyte that will ultimately permit the analyte to be immobilized via the immobilized first binding reagent, and ultimately to immobilize the second binding reagent. Preferably the binding interaction will have a Kd of about 10−6, more preferably about 10−7, even more preferably about 10−8, and as high as about 10−14. In a preferred embodiment, such binding will be in the nature of a protein/protein or antigen/antibody interaction.
The first binding reagent should also be immobilized. Many immobilization schemes are well known to one of skill in the art, and include covalent immobilization on a solid support such as plastics, magnetic beads, nylon, carbohydrate-based supports, etc. The first binding reagent may be immobilized at any time during the process. For example, it may be immobilized before contact with the fluid sample suspected of containing the analyte. In another embodiment, the first binding reagent may be immobilized after it is contacted with the analyte and/or the second binding reagent. In the latter embodiment, the binding reaction(s) is carried out in solution, then the first binding reagent is subsequently (or simultaneously) immobilized by methods well-known to one of ordinary skill. For example, the first binding reagent in solution may be contacted with a solid medium having an affinity for the first binding reagent, for example a bead coated with an antibody that binds to the first binding reagent.
The first binding reagent may comprise any moiety that is capable of binding to the analyte. Preferred moieties include proteins and antibodies. When the analyte is an antibody, the first binding reagent preferably comprises a protein known to bind to the class of such antibody, such as protein A or protein G.
The second binding reagent of the present invention comprises a fusion protein having a reporter domain and a binding domain. The fusion protein may be made by conventional cloning techniques. The fusion protein may be expressed in a wide range of cells, including mammalian, yeast and plant cells. In a preferred embodiment, the fusion protein is expressed in mammalian cells or cell extracts, such as Cos cells, HeLa, Vero, CHO, NIH 3T3, 293, etc. The use of mammalian cells is particularly preferred for the following reasons. Most immunoassays use bacterial-expressed proteins for detecting antigen-specific antibodies in human sera [2]. However, since such antigens do not carry post-translational modifications or may fold incorrectly, some immunoassays employ antigens produced in either yeast or insect cells. While these antigens may fold correctly and carry post-translational modifications, they will not carry either mammalian- or disease-specific posttranslational modifications. Tests employing bacterial-expressed proteins can produce high backgrounds because it is difficult to completely eliminate or block serum antibodies reactive with trace amounts of bacterial contaminants present in most antigen preparations, even in pharmaceutical grade preparations [3]. Therefore, the use of mammalian-produced fusions can overcome those problems. In a preferred embodiment, such fusions will contain a post-translational modification, such as glycosylation, acetylation, lipidation (e.g., palmitoylation), phosphorylation, citrullination, etc.).
The binding domain of the second binding reagent is capable of binding the analyte if present in the sample, with “binding” being used in the same sense as above. Further, the first and second binding reagents should be capable of binding the analyte simultaneously if present in the sample. When that occurs, it will be apparent that the second binding reagent becomes immobilized through the analyte and the first binding reagent. The binding domain of the second binding reagent may comprise a full-length protein, or a portion of a full-length protein sufficient to bind to the analyte.
The reporter domain of the second binding reagent comprises a detectable moiety that may be used to detect the presence of the second binding reagent. The detectable moiety may be any polypeptide or protein that is capable of detection, either directly or indirectly. Many such moieties are known. In a preferred embodiment, the detectable moiety is a detectable enzyme, such as luciferase, horseradish peroxidase, alkaline phosphatase, etc. Renilla luciferase (abbreviated herein as “Ruc”) is particularly preferred. In another embodiment, the detectable moiety comprises multiple copies of a detectable enzyme. Such may be accomplished by, for example, the use of a cloning vector coding for multiple copies of the enzyme, which may be linked in tandem, or located on either side of the binding domain. Other detectable moieties include, for example, fluorescent proteins such as green fluorescent protein, and various other colored proteins sold by, e.g., Clontech in their Living Colors™ product line.
The methods of the present invention may be used to detect a wide range of analytes, including, but not limited to, proteins, antibodies, carbohydrates, lipids, etc. It will be apparent that the analyte to be detected should be capable of binding simultaneously to the first and second binding reagents. In a preferred embodiment, the analyte is an antibody (e.g., an IgA, IgE, IgG, IgM, etc.), and the first binding reagent and the binding domain of the second binding reagent are both antigens. If the analyte is other than an antibody, then the first binding reagent and the binding domain of the second binding reagent may be antibodies that bind to the analyte.
The analyte to be detected may be indicative of the presence or progress of a disease state. For example, the present invention may be used detect the presence of antibodies generated in response to the presence of pathogens such as viruses, bacteria, fungi, parasites, etc. Any pathogen that generates a humoral response may be detected according to the present invention. A non-exhaustive list of pathogens is available on the American Biological Safety Association website (www.absa.org/resriskgroup.html). Particularly preferred viruses including their subtypes are HIV, CMV, Hepatitis B, Hepatitis C, West Nile virus, HPV, RSV, herpes, HTLV-1, SARS, etc. Particularly preferred bacteria include M. Tuberculosis, H. pylori, anthrax, F. tularensis, streptococcus, pneumococcus, E. coli, Clostridia, staphylococcus, meningococcus, the causative agents of various sexually-transmitted diseases such as syphilis and gonorrhea, the causative agents of Legionnaire's and Lyme disease, etc. Particularly preferred fungi/yeast include Pneumocystis, Candida, Saccharomyces, Histoplasma, Cryptococcus, and Aspergillis. Particularly preferred parasites/protozoa include Plasmodia, Schistosoma, Cryptosporidium, and Toxoplasma.
The present invention may also be used to detect antibodies generated in response to autoimmune diseases. Many such diseases are known, and include, for example, Alopecia greata, Antiphospholipid syndrome, Addison's disease, Arthritis, Ankylosing spondylitis, Dermatomyositis, Fibromyalgia-Fibromyositis, Juvenile arthritis, Polymyalgia Rheumatica, Polymyositis, Raynaud's phenomenon, Reiter's syndrome, Rheumatoid arthritis, Scleroderma, Sjogren's syndrome, Arteritis, Polyarteritis nodosa, Takayasu Arteritis, Temporal arteritis/Giant Cell arteritis, Autoimmune hemolytic anemia, Autoimmune hepatitis, Behcet's disease, Cardiomyopathy, Celiac Sprue, Celiac Sprue-dermatitis, Chronic Fatigue Immune Dysfunction Syndrome, Chronic Inflammatory Demyelinating Polyneuropathy, Churg-Strauss Syndrome, CREST syndrome, Cold Agglutinin Disease, Crohn's disease, Type 1 diabetes, Essential Mixed Cryoglobulinemia, Glomerulonephritis, Graves' disease, Guillain-Barre syndrome, Miller-Fisher syndrome, Idiopathic Pulmonary Fibrosis, Idiopathic Thrombocytopenia Purpura, IgA Nephropathy (Berger's disease), Inflammatory bowel disease, Lichen Planus, Lupus, Lupus nephritis, Systemic lupus erythematosis, Meniere's disease, Mixed connective tissue disease, Multiple sclerosis, Myasthenia gravis, Myocarditis, Pemphigus/pemphigoid, Bullous pemphigoid, Cicatricial pemphigoid, Pemphigus vulgaris, Pernicious anemia, Polychondritis, Polyglandular syndromes, Primary Agamma-globulinemia, Primary biliary cirrhosis, Psoriasis, Retinitis, Rheumatic fever, Sarcoidosis, Stiff-Man syndrome, Thyroiditis, Ulcerative colitis, Uveitis, Vasculitis, Vitiligo, and Wegener's granulomatosis.
The present invention may also be used to detect antibodies generated in response to chronic inflammatory diseases. Many such diseases are known, and include, for example, rheumatoid arthritis, osteoarthritis, chronic obstructive pulmonary disease, etc. The present invention may also be used to detect allergic reactions.
The present invention may also be used to detect antibodies generated in response to the presence of known tumor-associated proteins, such as p53, K-Ras, c-Myc, β-catenin, Smad4, PSA, etc., and hence may be used to detect or follow the course of various cancers, such as colon, breast, prostate, head and neck, etc. In that embodiment, the binding domain of the second binding reagent will comprise the tumor-associated protein, or at least a portion thereof sufficient to bind to the antibodies to be detected, e.g., the antigenic portion of the protein. Any cancer is contemplated to be within the scope of the present invention; a non-exhaustive list may be found at http://www.cancer.gov/cancertopics/alphalist, incorporated herein by reference. The present invention may also be used to detect or quantify the presence of proteins associated with a particular disease state, such as the tumor-associated proteins mentioned above.
Because the present invention is capable of quantifying an analyte, it will be readily apparent that the present invention may be used not only to detect the presence of a disease state, but also to monitor the progress of a disease or condition, and to monitor the progress of treatment of a disease or condition. In that instance, the present assay will be repeated over time to determine the change, if any, of the concentration of the particular analyte over time. Thus, the effectiveness of, e.g., a course of cancer therapy (chemotherapy, radiation, etc.) or infectious disease drug therapy may be determined readily.
The assay of the present invention may be performed by contacting the first binding reagent, the sample and the second binding reagent. If the analyte is present, a complex is formed among the two binding reagents and the analyte. Because the first binding reagent is immobilized, the second binding reagent likewise becomes immobilized if the analyte is present. The immobilized complex is separated from the reaction mixture, for example by washing, and the presence of the second binding reagent in the complex is detected via the reporter domain. The presence (or concentration) of the second binding reagent in the immobilized complex is indicative of the presence (or concentration) of the analyte. The order of addition of the reagents and analyte is not critical. Thus, the analyte may be mixed with the first binding reagent, then the second binding reagent may be added to the mixture. Alternatively, the steps could be reversed, i.e., the analyte may be mixed with the second binding reagent, then the first binding reagent may be added to the mixture. Finally, the present invention contemplates that the first and second binding reagents and the analyte could be mixed together at the same time, or the first and second binding reagents are pre-mixed and the analyte added to the mixture.
It will be apparent that the assay of the present invention may be carried out so as to detect a single analyte, e.g., by testing a sample with a single pair of binding reagents designed to detect a single analyte of interest. It is also contemplated that the present invention may be used to detect multiple analytes in a single sample. That may be done by utilizing multiple pairs of binding reagents designed to detect multiple analytes of interest. Such may be easily achieved by the use of, for example, multiple well plates, wherein the multiple first binding reagents are immobilized in discrete wells in the plate. Alternatively, multiple binding domains, each of which binds the different analytes, may be incorporated with the record binding reagent, for example in tandem.
The present invention also contemplates that the reagents for use in the assays described herein be contained in the form of a kit. Such kits would include, contained within suitable packaging material, an immobilized first binding reagent and a second binding reagent, both as described above. The kits may optionally further contain other components, such as instructional material for use of the kit, reagents for detecting the reporter domain of the second binding reagent (e.g., buffers, compounds that produce light when contacted with the detectable enzyme, etc.), a positive control for comparative or calibration purposes, etc.
The present invention will now be illustrated with reference to the non-limiting examples described below.
We used Ruc-tagged antigen-fusion proteins to develop an immunoprecipitation assay that can quantitatively measure serum antibody reactivity with protein antigens. In brief, crude extract containing the Ruc-antigen fusions, sera and protein A/G beads are mixed together and incubated, during which the antigen fusions become immobilized; antigen-specific antibody is then quantitated by washing the beads and adding the colenterazine substrate. In these assays the amount of light produced is proportional to the amount of soluble fusion protein captured by the antibody-bound beads. It should be noted that the binding capacity of the protein A/G beads (Pierce Biochemical) used to capture either purified monoclonal antibodies or immunoglobulins from crude human or animal antisera is high (24 μg of immunoglobulins/μl of packed beads).
The Immunoprecipitation Assay Shows a Linear Range of Detection with Commercial Antibodies
To illustrate this technology we generated Ruc fusion protein constructs for p53, K-Ras, c-Myc, β-catenin and Smad4 by fusing cDNAs encoding these proteins (in frame) to DNA encoding the C-terminus of Ruc in a mammalian expression vector, pREN2, which also encodes a FLAG epitope tag at the N-terminus of Ruc. Transfections into Cos1 cells of these different constructs yielded crude extracts with 3-10×108 Ruc light units per 100 mm2 plate. We developed a standard assay format by using a commercial anti-FLAG monoclonal antibody and Cos1 cell extracts containing Ruc-p53. When crude extract, antisera and protein A/G beads were incubated together in a single tube, the amount of immunoprecipitated Ruc-p53 was directly proportional to the amount of anti-FLAG antibody over a 1000-fold range of concentrations, with a lower limit of detection of less than 5 picrograms (
Human Cancer Patient Sera Contain Antigen-Specific Antibodies
Since commercial antibodies can immunoprecipitate Ruc-antigen fusions from crude Cos1 extracts, we tested whether our simple assay format could also detect antigen-specific antibodies in clinical sera samples. Our motive for developing this technique was to have an improved method for detecting cancer patient antibody responses to tumor-associated proteins. Thus, we initially tested the assay with a small number of clinical sera samples taken from patients having three types of cancers: breast, colon and head and neck. In order to maximize our chances of detecting positive responses with these clinical sera samples we chose to use p53 and four other tumor-associated proteins (K-Ras, c-Myc, β-catenin and Smad4) that are either frequently mutated and/or overexpressed in various tumors. Wild-type proteins were used as antigens because several studies show that cancer patient sera humoral immune responses are not restricted to or even preferential for the epitopes that usually contain the altered amino acids [16, 18-21]. Cos1 extracts containing Ruc-antigen fusions were used to test a total of 36 sera, comprised of 10 controls and 26 cancer patients (Table 1). Negative and positive controls consisting of protein A/G beads alone and 0.1 μg of anti-FLAG monoclonal antibody with protein A/G beads, respectively, were used for each experiment. As expected, all sera had low reactivity with the non-specific binding control protein, Ruc-alone (Table 1). The positive control, anti-FLAG antibody, immunoprecipitated significant amounts of each of the Ruc-antigen fusions. However, the fraction of the total Ruc activity that could be captured varied amongst the different Ruc-antigen fusions, possibly reflecting reduced accessibility to the N-terminal FLAG epitope in some constructs (data not shown). At least one cancer patient sera had statistically significant antibody responses to each of the five Ruc fusions, where significance is defined as a response greater than the average plus three standard deviations of the 10 control sera (Table 1). Two of 10 head and neck, five of 10 breast, five of six colon cancer sera, but none of 10 healthy control sera gave positive responses. Six of the 12 positive tests were clustered in the six colon cancer patient sera and two antigens, p53 and K-Ras (Table 1). The significance of the relative response rates between different cancer-type sera cannot be calculated because the sample sizes are small and because no effort was made to match the control and patient sera by any criteria. Similarly, we cannot conclude that either K-Ras and/or p53 may be more antigenic in colon cancers than either β-catenin or c-Myc. Interestingly, the only multiple sample from any of the patients, head and neck samples 11 and 12, are sequential samples of which only the more recent sample showed significant levels of anti-p53 antibodies. Since the proteins used to test for antibodies in these 26 cancer patient sera are often mutated and/or overexpressed in the three types of cancer, our results are consistent with studies indicating that these categories of proteins are often antigenic in cancer patients [2]. Our results with colon cancer patient sera also indicate that humoral immune responses to panels of tumor-associated antigens may be clinically useful when single antibody responses are not informative [22, 23].
To determine whether patient antibody responses behave in the same linear manner as the commercial antibodies, we used the most reactive combination of patient sera and fusion antigen in our sample set, colon cancer sera 34 and the Ruc-p53 fusion. Although the amount of Ruc-p53 captured by this serum is roughly linear with incubation time in the presence of protein A/G beads, reaching a plateau by 30-60 minutes (
Competition Experiments with Unmodified Proteins
While human humoral immune responses to post-translational modifications are often ignored and/or undetectable with existing technologies, recent studies demonstrate that disease-related antibody responses can occur to post-translational protein modifications [25]. In at least one case, rheumatoid arthritis, antibody responses to a post-translational modification, citrullination, is now being intensely investigated as a potentially reliable disease indicator [26,27]. In light of these observations, we asked whether each positive sera response seen in Table 1 could include antibodies that were directed toward post-translational modifications by doing competition experiments with unmodified E. coli-produced antigens. These competition experiments (Table 2) show that 0-100% of the immunoprecipitated Ruc-antigen fusions were blocked by preincubating sera with 5 μg of the corresponding E. coli-produced antigens fused to maltose binding protein (MBP). These differences occur even between sera containing antibodies that recognize the same antigen (e.g. p53 or K-Ras), proteins known to contain post-translational modifications. These differences could mean that some tumors tend to produce proteins having more post-translational modifications or that some cancer patients' immune systems tend to produce significantly more antibodies that recognize post-translational modifications. However, this data does not exclude the possibility that some or all of each positive antibody response detected is not even specific for the antigen listed, since the apparent anti-p53 or anti-K-Ras antibodies could be directed toward proteins that are in complexes with these tumor antigens. If the tumor antigens in these complexes were easily replaced by the MBP fusions, one would see higher competition values than if they were inefficiently replaced. Quantitative evaluation of different competition results requires, at a minimum, equal amount of reactive antibodies in each sera, a condition unlikely to be satisfied here, especially for the p53-reactive sera. In addition, when we compared the dose-response competition curves of sera 34 and the commercial polyclonal anti-p53, adjusted to similar capacities for immunoprecipitating Ruc-p53, we found a greater difference than indicated by the end-point values alone (
These data show that our approach of making antigen-enzyme fusions and producing these fusions in mammalian cells is superior to conventional ELISA assays for detecting antigen-specific antibody responses in human sera. Specifically, we have tested the six colon cancer patient sera used here in a standard sandwich type ELISA where the antigen were fused to E. Coli MBP and immobilized on ELISA plates with a monoclonal anti-MBP antibody. In these ELISA tests only two of the six colon cancer sera gave positive responses with any of the five tumor-associated proteins listed in Table 1 (data not shown). In any case, the immunoprecipitation assay described here offers a practical approach for identifying post-translational modification-specific antibody responses and studying their medical relevance.
Quantitative results were obtained by using easily prepared crude cell extracts containing post-translationally modified antigens fused to a light-producing enzyme reporter. While the immunodetection of antigen-enzymes is not new [28, 29], by combining a robust reporter, such as Ruc with the production of recombinant enzyme-antigen fusions in mammalian cells, we have created a highly sensitive, user friendly assay. This assay requires fewer manipulations for reagent preparation and less time than other immunoprecipitation methods including avoiding having to purify and then radiolabel the purified proteins or having to perform additional analysis such as Western blotting after the immunoprecipitations [30]. Thus, it is within the scope of the present invention to utilize the second binding reagent in less than completely pure form, i.e., as a component of a crude extract. Producing the target antigens in mammalian cells offers several potential advantages, including having mammalian-specific and/or disease-specific post-translational modifications added to these antigens. Thus, this immunoprecipitation assay provides a simple, accessible, reliable and reproducible tool for investigations aimed at documenting the role of post-translational modification in disease. Although altered post-translationally modified proteins occur in cancer [31, 32], future studies are needed to explore whether there are detectable cancer patient-specific antibodies to post-translationally-modified tumor proteins. The levels and kinds of post-translational modifications on the Ruc-antigen fusions can be manipulated by exploiting mutant proteins, unique human cell lines (e.g. cell lines overexpressing tyrosine kinases) and various culture conditions. Mammalian-produced antigens have additional advantages over bacterial-expressed antigens including facilitating the study of antibody responses to very large proteins (>100 kDa) that are difficult or impossible to produce as intact proteins in E. coli. Our assay also avoids false positives caused by variable amounts of anti-E. coli antibodies present in patient sera that react with the minor amounts of E. coli proteins that co-purify with bacterial recombinant proteins; such contaminants are even present in some pharmaceutical-grade recombinant protein preparations [3]. These advantages, along with the possibility of improving the assay format, make this assay worthwhile to reevaluate the frequency with which known tumor-associated proteins are detectably antigenic in cancer patients. It is encouraging, although of limited significance, that the frequencies of significant antibody responses for two of the cancers are roughly comparable to reports in the literature. Thus, in colon cancer patients we detected statistically significant antibody responses to Ras and p53 in 50% and 33% of the sera, respectively, compared to published reports of 33% for Ras [1,6] and 26% for p53 [33]. In contrast, we did not find any statistically significant antibody responses to p53 in breast cancer sera, which have been reported to occur with 9% of patient sera [34].
This assay format and high-throughput modifications (e.g. magnetic A/G beads in a microtiter plate format) are obviously directly applicable to detecting human sera antibodies specific for any protein antigen of interest and is also useful for non-human sera, such as sera obtained from animal models of disease, as well as for antibodies in other bodily fluids including from urine, cerebrospinal fluid, amniotic fluid, gastic fluid, ascites and saliva. Variations of this immunoprecipitation assay format might also be useful for studying other types of protein-protein interactions.
Methods
Biochemical Reagents and Antibodies
Ultralink™ immobilized protein A/G beads were obtained from Pierce Biotechnology Inc. Commercially available antibodies were: mouse monoclonal anti-FLAGT™ M2 from Sigma; rabbit anti-acetylated p53 from Upstate Biochemicals and polyclonal rabbit anti-p53, polyclonal rabbit phosphoserine p53 and polyclonal anti-WASP from Santa Cruz Biotechnology.
Patient Sera
The breast and colon cancer patient sera were obtained from the University of Wisconsin collection, now kept at Georgetown University Medical Center. Sera samples from head and neck cancer patients and control sera were collected by Dr. Radoslav Goldman at Georgetown University Medical Center (Washington, D.C.). The sex, age and disease stages of these samples were not examined until after the reactivities for all antigens were measured.
Generation of Constructs Encoding Ruc Fused to Tumor-Associated Antigens
pREN2, a FLAG-epitope-tagged mammalian expression vector, similar to the previously described pREN1 [4], was used to generate all plasmids encoding Ruc fusions. The tumor antigens are at the C-terminus and a single FLAG tag is at the N-terminus of Ruc. A map of pREN2 is shown in
For Ruc alone, a separate construct was prepared containing a stop codon at the end of the luciferase coding sequence in place of the polylinker present in pREN2.
Immunoprecipitation Assays with Ruc Fusion Proteins
Forty-eight hours after Fugene-6 transfection, Cos1 cells in 100 mm2 plates were washed twice with PBS, scraped with 1.0 ml of Buffer A (20 mM Tris, pH 7.5, 150 mM NaCl, 5 mM MgCl2, 1% Triton X-100) plus 50% glycerol and protease inhibitors (10 μg/mL each of leupeptin, aprotinin and pepstatin), sonicated, centrifuged at 13,000×g for 4 min, supernatants collected and used immediately or stored at −20° C. Total luciferase activity in 1 μl of each crude extract was measured by adding it to 100 μl of assay buffer and substrate mixture (Renilla Luciferase Reagent Kit, Promega) in a 12×75-mm glass tube, vortexing and immediately measuring light-forming units with a luminometer (GeneProbe) for 10 sec. Lysate prepared from each 100 mm2 plate of transfected Cos1 cells typically provides enough extract for 60-200 assays. These crude Cos1 extracts containing these Ruc fusions were stable for at least a few weeks when stored in 50% glycerol at −20° C.
Immunoprecipitation assays were performed in 100 μl volumes containing 6 μl of a 30% suspension of protein A/G beads (in PBS), 1-10 μl sera (undiluted or diluted in Buffer A plus 100 μg/ml BSA), sufficient Cos1 cell extract to generate 1-5 million light units (usually 5 μl to 10 μl) and Buffer A and incubated at 4° C. with tumbling for 5-120 minutes, washed 4-5 times with 1.2 ml of cold Buffer A and once with 1.0 ml of PBS. After the final wash, the beads, in a final volume of about 10 μl, were added to the Ruc substrate and light units measured as described above. Since the capacity of these protein A/G beads is 24-32 mg/ml of packed beads, 2 μl of packed beads should be sufficient to immobilize most or all of the IgG in 1 μl of undiluted sera (assumed to be 10 mg/ml IgG). The amount of IgG in 2 μl of each sera that actually bound to protein A/G beads was estimated by measuring the amount of bead-bound sera released by a low pH glycine elution buffer and measured using the BCA Protein Assay kit (Pierce Biotechnology Inc.). The protein values varied from 2.0 to 7.3 μg/μl of patient sera.
Competition experiments were performed using MBP-fusion proteins. Bacterial expression vectors were constructed by subcloning cDNA fragments into the pMAL-c2 vector (New England Biolabs). Recombinant MBP fusion proteins were produced in bacteria, purified by amylose-agarose affinity and eluted with maltose as described by the manufacturer and stored frozen or in 50% glycerol at −20° C. An MBP fusion containing the SPEC2 cDNA [35] was produced and used as a non-specific inhibitor. The integrity of the proteins was confirmed by SDS-PAGE electrophoresis and protein concentration determined. Diluted patient sera (10 μl used of sera diluted 1:10 in buffer A containing 100 μg/ml BSA) were used in the competition experiments described in Table 2, while only 5 μl of 1:10 diluted colon patient sera 34 was used in the experiments described in
aSera, FLAG-Ruc-fusion extracts, protein A/G beads and buffer were mixed together, incubated for 60 minutes and processed. The data, light units, is the average of two experiments and is corrected for background (beads plus extract, but no sera). The standard
bValues of the averages of the 10 control sera plus 3 standard deviations.
cNumbers in bold are statistically significant: greater than the average plus 3 standard deviations of the 10 control sera.
aSera (1 μl), buffer and 5 μg competitor were incubated together for 60 min before adding the fusion
Results
Rapid and Accurate Identification of Human Sera Containing Anti-HIV Antibodies Using an Antigen-Reporter Fusion Protein as the Antigen
We reported the successful use of antigen-Renilla luciferase (Ruc) fusions, produced in Cos1 monkey cells, in a simple immunoprecipitation assay, to quantitatively measure human serum antibody responses to tumor-associated proteins (38). Here, we tested whether a minor modification of this technology could be used to successfully predict the infection status of blinded serum samples, some of which were from patients exposed to infectious agents.
Serum antibody responses to HIV were measured by using a single protein antigen, a fusion between a major HIV core protein, p24, and Renilla luciferase. Transient transfections of Cos1 cells with an expression vector for this fusion yielded crude extracts able to generate 2-10×108 Ruc light units per 100 mm2 plate, similar to the amounts of Ruc light units obtained with other human Ruc-antigen fusions produced under similar conditions (38). Using one such crude extract as the antigen, we profiled 28 sera, consisting of equal numbers of previously characterized HIV positive and negative sera, although we did not know at the time of testing that this set contained equal numbers of positive and negative sera. The methods used to classify these sera as positive or negative (Western Blotting and/or PCR) were also unknown to us (at Georgetown University) at the time of our tests. We tested these sera at two different times, about 6 days apart, using two different crude extracts, both of which were stored at −20° C. before use. For this test we slightly modified our previously described assay format (38). Instead of combining sera, antigen and protein A/G beads all together and then incubating, we first combined and incubated only the sera and antigen and then subsequently added protein A/G beads, under the theory that antigen-antibody complex formation is more likely to be limiting than antibody immobilization onto the beads. That is, antigen-antibody complexes will form more rapidly in solution than after antibody immobilization onto the beads and that quantitative immobilization of antibodies onto the beads is likely to be independent of whether or not they have already formed antibody-antigen complexes. In any case, this modified assay format yielded quantitative values, whose averages span almost four-logs and which probably reflect relative antibody titers (Table 3). When tabulated from lowest to highest, it was believed that that this blinded serum set contained 14 HIV positive and 14 HIV negative sera, which turned out to be correct. The 14 lowest values were less than 1000, with an average of 309 and a relatively small standard deviation of 126. The average of the highest 14 values is 595,300 with a standard deviation of 1,091,000. Clearly, the highest 14 values vary considerably more than the lowest 14, a result expected if our classification was correct and if the 14 highest values are indeed positive sera and are from individuals in different states of disease progression (a speculation also confirmed after the blinded code was broken, see below). Breaking the blinded code confirmed both of our predictions. Thus, this initial blinded assay for HIV yielded 100% sensitivity and 100% specificity.
Additional clinical information about these sera revealed that four of the weakest, but still positive responders, with average titers of 2,716, 4,785, 8,507 and 9,317 light units, were obtained from chronically HIV-infected individuals. These four low anti-p24 antibody titers, in sera from chronically infected individuals, are consistent with several studies showing that anti-p24 HIV antibody responses are lost during disease progression (36; 47) or following successfully antiviral therapy (42). It should be noted that this successful classification was accomplished with crude extracts, was done without optimizing assay conditions and did not involve the use of a training sera set to determine cut-off values. In addition, each set of 28 assays required relatively short hands-on time and less than 15 hours from mixing sera with antigen to data collection.
Comparisons Between the IP and ELISA Assays for Detecting Antibodies Against Pneumocystis jiroveci in Sera from Non-Infected and HIV-Infected Individuals
We next determined whether our immunoprecipitation assay could accurately reproduce the relative antibody titers to Pneumocystisjiroveci, a pathogen often associated with Pneumocystis pneumonia (PCP) in immunocompromised individuals. All of the current ELISA assays used to detect antibodies to Pneumocystisjiroveci employ only fragments of MSG-14, a large major surface glycoprotein of Pneumocystisj. (37; 39) and these fragments are produced in bacteria. Using a crude Cos-1 Ruc-MSG-14Ca fusion extract as the antigen in our immunoprecipitation assay format, we blindly profiled the set of 28 sera already analyzed for anti-p24 HIV antibodies, described above. As shown in Table 3, the antibody responses to this antigen, measured in our immunoprecipitation assay, ranged from 4,318 to 1,260,205 light units. Following unmasking of the blinded samples, we compared the immunoprecipitation and ELISA assay values of the anti-MSG-14 antibody titers and found a high level of correlation (R=0.86), but only if sample #20 was omitted from the analysis (
Compare Methods for Monitoring P. jiroveci-Specific Antibodies in Serial Serum Samples
Antibody titer profiles generated from serial samples often reflect the clinical course of infection. To test whether the present assay is as good as an ELISA for detecting patient antibody responses to P. jiroveci in longitudinal sera samples, we analyzed, in a blinded fashion, a set of 24 samples collected from 6 patients at four time points over about a year spanning their P. jiroveci infections. A comparison of the antibody titer values obtained using the immunoprecipitation assay and a standard ELISA is shown in
The Immunoprecipitation Assays can Detect Antibodies to Multiple Components of Hepatitis C Virus (HCV) in Individual Human Sera
We investigated whether the present immunoprecipitation assay could detect antibody responses to multiple different proteins of HCV in single serum samples. A new, blinded set of 33 clinical sera samples, comprising unknown numbers of non-infected individuals and individuals infected with one or combinations of HBV, HCV and HIV were evaluated for antibody titers against Ruc-core, Ruc-NS3 fragment (C33) and Ruc-NS5A of HCV. The results of these HCV tests are given in Table 5. Based on the relatively sera titers and range of titers for each of the antigens, we predicted that there were 13 positives and 20 negatives for the HCV core, 10 positives and 23 negatives for NS5A and 10 positive and 20 negative for the NS3 fragment (C33). By combining the results for the three independent HCV assays we stipulate that a positive in any of the three HCV tests signals a positive sera test. If so, there are 13 positives and 20 negatives. Our classifications of the HCV status of these 33 samples completely agreed with their clinical status (Table 5). Thus, this initial blinded assay for HCV yielded 100% sensitivity and 100% specificity. Further comparisons show that the HCV core test correctly scored 13/13 positives (100%), the C33 test detected 12/13 positives (92.5%) and the NS5A test detected 10/13 positives (77%). An appealing aspect of this immunoprecipitation test is the large gap between the highest negative test value and the lowest positive test value for many of the antigens tested. Based on these promising results, it is apparent that determining the breadth, strength and kinetics of IgG antibody responses against the entire and/or partial proteome of HCV and other infectious agents is highly feasible and may have significant diagnostic and prognostic value.
Generation of Ruc-Antigen Fusion Constructs
pREN2, a FLAG-epitope-tagged mammalian expression vector was used to generate all plasmids. DNA templates for various pathogens were obtained from Dr. J. Casey for HBV, Dr. C. Rhodes for HIV-1, and Dr. Pad Padmanabhan for the NS5A protein of HCV. The MSG-14 can clone of Pneumocystis j. was previously described (ref). Full-length coding sequences (excluding the initial methionine) were used for the tumor antigens, with the exception of the MSG-14 which encoded amino acids 2-277. In every case a stop codon was included after the C-terminal coding sequences of the tumor antigens. PCR specific linker-primer adapters were used for amplification. The primer adapter sequences used for cloning each antigen are as follows:
Following construction, the different mammalian cell pREN2 expression vectors for the different antigens were purified using Qiagen Midi preparation kits.
In the light of the ability of certain viruses to cause cancer, we tested whether the immunoprecipitation technology employing human papilloma virus (HPV) antigens, could identify blinded sera samples from previously evaluated HPV-DNA positive head and neck tumors. As an initial screening strategy in this pilot study we examined 127 blinded sera samples for high level antibodies against the E7 oncogene of HPV. Sera were evaluated for antibody titers against a Ruc-E7 fusion protein of HPV. High positives were culled, repeated and then this subset was examined for anti-E6 antibodies using a Ruc-E6 fusion protein. Using this less than perfect screening strategy, nine sera were identified by the immunoprecipitation assay as containing high titer antibodies against HPV in this large pilot experiment (Table 6). Following breaking of the blinded code, no HPV-antibody positive calls were made among any of the approximately 60 non-cancer controls. Eight of the nine HPV antibody positives sera were among the head and neck cancer sera samples that were known to have HPV DNA-positive tumors, while three tumors containing high copy number HPV infection were not detected (Table 6). One additionally highly positive HPV immunoreactive sera was also detected, which was not detected as DNA positive and may have represented a head and neck tumor initiated by HPV infection but now resolved or an HPV infection at a different tumor site (e.g. cervical cancer). This pilot study is highly encouraging, in light of additional obvious improvements including using additional HPV gene products to improve sensitivity. It should again be noted that this successful classification was accomplished with crude extracts, was done without optimizing assay conditions and did not involve the use of a training sera set to determine cut-off values. These results with the immunoprecipitation technology profiling exposure to infectious agents should have wide diagnostic applications for HPV-associated cancers and other infection-related cancers.
This application claims the benefit of U.S. Provisional Application No. 60/638,811, filed Dec. 23, 2004. The entire teachings of the referenced application are incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
60638811 | Dec 2004 | US |