Carcinoembryonic antigen isomer

Abstract

The discovery of an isomeric species of carcinoembryonic antigen and methods of isolation, identification and utilization of the same as an aid in the diagnosis of adenocarcinomas of the gastrointenstinal tract are disclosed.

Description

RELATED APPLICATION

Application Ser. No. 682,028 filed concurrently herewith on Apr. 30, 1976, DIAGNOSTIC METHOD AND REAGENT, Thomas S. Edgington and Edward F. Plow.

INCORPORATION BY REFERENCE

The aforesaid patent application Ser. No. 682,028, filed concurrently herewith on Apr. 30, 1976, and the publications of the inventors, ISOLATION AND CHARACTERIZATION OF A HOMOGENEOUS SPECIES OF CARCINOEMBRYONIC ANTIGEN: CEA-S, Int. J. Cancer: 15, 748-761, 1975, and ASSOCIATION OF AN ISOMERIC SPECIES OF CARCINOEMBRYONIC ANTIGEN WITH NEOPLASIA OF THE GASTROINTESTINAL TRACT, (with Robert W. Astarita), New Eng. J. of Med., 293:103-107 (July 17) 1975, are incorporated herein by reference as fully as though reproduced as part of this specification.

FIELD OF THE INVENTION

This invention relates to the isolation and identification of human tumor-associated antigens, specifically to a species of carcinoembryonic antigen designated CEA-S.sub.1.

BACKGROUND OF THE INVENTION

Carcinoembryonic antigen (CEA) has been defined as glycoprotein present during fetal and embryonic development of the human, present at very low concentration in the mucosal epithelium of the adult human gastrointestinal tract, and produced in high concentration by a variety of tumors of the gastrointestinal tract (Gold, P. & Freedman, S. O.: J. Exp. Med. 122: 439-462, 467-481, 1965).

CEA has been isolated from gastrointestinal tumors and characterized as an approximately 200,000 molecular weight glycoprotein by Gold and Freedman (J. Exp. Med. 128: 387-398, 1968) and by Coligan, J. E., et al (Immunochem. 10: 591-599, 1973).

CEA has been utilized in radioimmunoassays of serums for the detection of cancer in humans, first by Thompson, et al. (Proc. Natl. Acad. Sci. 64: 161-167, 1969) and subsequently by Lo Gerfo, et al. (New Engl. J. Med. 285: 138-141, 1971), Hansen, H. J., U.S. Pat. Nos. 3,697,638 (Oct. 10, 1972) and 3,867,363 (Feb. 18, 1975), Freedman, S. O., Gold, P., and Krupey, J. H., U.S. Pat. No. 3,663,684 (May 16, 1972) and others. This has not provided an adequately precise or specific test for the clinical diagnosis and management of cancer to date since a high incidence of false positive results (3-19% of normal controls and as high as 67% of patients with inflammatory bowel disease) are observed (Hansen, H. J.: Human Pathology 5: 139-147, 1974).

We have isolated what appears to be a tumor-dominant isomer, subspecies or variant of CEA (CEA-S.sub.1). CEA-S is a glycoprotein of approximately 181,000, is devoid of blood group antigens, shares antigenic determinants with CEA but is also immunochemically distinguishable from CEA.

Brief Summary

We have developed and proved the application of CEA-S.sub.1 to an immunodiagnostic assay for the detection of tumors of the gastrointestinal tract (Edgington, T. S., Astarita, R. W. and Plow, E. F.: New Engl. J. Med. 293: 103-107, 1975). The use of this defined isomeric species of carcinoembryonic antigen, isolated from adenocarcinomas of the colon, in a radioimmunoassay of body fluids (i.e., serum) has provided a laboratory assay to distinguish between normal individuals and most of those with tumors of the gastrointestinal tract. The CEA-S.sub.1 assay has been used to evaluate clinical application to the detection and management of human cancer of the gastrointestinal tract. 80.4% of patients with neoplasia of the gastrointestinal tract were positive whereas only 9.3% of other tumors, predominantly of lung and breast origin, were detectable. Serum levels were elevated in only 0.23% of random patients and 0.41% of 725 patients without neoplasia but including those with diseases frequently associated with false positive results by other CEA assays. The results of these studies suggest that CEA-S.sub.1 may represent a tumor dominant isomeric species of carcinoembryonic antigen. Assays of this technique offer a new and an improved approach for diagnosis and management of cancer of the gastrointestinal tract.

DRAWING

FIG. 1 is a graphic depiction of CEA-S.sub.1 and CEA antigen binding data.

GENERAL DESCRIPTION

1. Isolation of CEA-S. (Plow, E. F. and Edgington, T. S.: Int. J. Cancer 15: 748-761, 1975, and unpublished studies) CEA-S.sub.1 is isolated from human adenocarcinomas of the gastrointestinal tract. Isolation involves (a) extraction of glycoprotein; (b) isolation of glycoproteins with a pI of approximately 4.4-4.6; (c) isolation of glycoproteins of approximately 200,000 by molecular exclusion chromatography; (d) recovery by ion exchange chromatography; (e) removal of other proteins, glycoproteins, and blood group antigen related molecules with immunoabsorbants; and (f) isolation of glycoprotein with a buoyant density of approximately 1.41.

This isolation involves: (1) tumor tissue is homogenized in approximately 1.5 volumes cold H.sub.2 O or saline screened through gauze and #25 mesh screening with washes of cold H.sub.2 O. The tumor homogenate is suspended at approximately three times the volume of the starting tumor weight at 4.degree. C. and clarified by ultracentrifugation. The supernatant is brought at 0.degree. C. to 0.9 M with perchloric acid (PCA). The supernatant is recovered following centrifugation and immediately brought to neutrality and dialyzed exhaustively. (2) The PCA soluble glycoprotein fraction is brought to 1% ampholyte concentration, pH 3-6. Precipitate is removed by centrifugation. An additional 1% ampholyte is added, and the sample is electrofocused. The major CEA antigen-containing peak of pI 4.4-4.6 is recovered. (3) This fraction is concentrated and separated by one or more molecular exclusion chromatography steps, typically utilizing first A-1.5 fine agarose (Biorad) then Sephadex G-200 and the peak of approximately 200,000 molecular weight with a Kd of 0.17 (G- 200) is collected and retained as the CEA-S.sub.1 fraction. (4) This glycoprotein fraction is brought to 0.05 M Na.sub.2 HPO.sub.4, pH 8.0, and passed through DEAE-cellulose column at 20.degree.-24.degree. C. equilibrated with the same buffer to remove more highly charged species of glycoprotein. The first major peak is collected and retained. (5) The sample is then passed through immunoadsorbents containing antisera to blood group antigens, human serum proteins, human serum glycoproteins, and antisera to tissue glycoproteins other than CEA or CEA-S.sub.1, i.e., NCA. (6) Following removal of these contaminants, the material that passes through this column is then subjected to equilibrium density gradient ultracentrifugation in cesium chloride. The CEA-S (major band) with a density of approximately 1.41 gm/ml is recovered and utilized as CEA-S.sub.1. The recovery of CEA-S varies between different tumors and represent 1-10% of the original tumor glycoprotein present in the PCA soluble tumor extract.

2. Identification and characterization of CEA-S.sub.1.

(a) Sedimentation velocity. CEA-S.sub.1 has a S.sub.20,.sub.w .degree. of 6.6 Svedberg units when analyzed by linear sucrose density gradient ultracentrifugation (Plow, E. F. & Edgington, T. S.: Int. J. Cancer 15: 748-761, 1975). This contrasts to a S.sub.20,.sub.w .degree. of 6.8 for CEA when analyzed simultaneously.

(b) Immunochemical characteristics. CEA-S.sub.1 precipitates with anti-CEA or anti-CEA-S.sub.1 in gel double diffusion to give an arc of apparent identity with CEA (Plow, E. F. & Edgington, T. S.: Int. J. Cancer 15: 748-761, 1975). Utilizing a variety of equilibrium competitive inhibition radioimmunoassays for CEA-S.sub.1 it is observed that preparation of CEA have 8-33% of the antigen activity of CEA-S.sub.1 and show reactions of only partial quantitative identity with CEA-S.sub.1. CEA-S.sub.1 is preferentially bound by anti-CEA-S.sub.1 antisera as compared to CEA. CEA-S.sub.1 is not bound by antiserum to blood group A or B antigens, Rh antigens, Kell, Duffy or Lewis a antigens, nor by antisera to normal tissue glycoprotein NCA.sub.1 (von Kleist, S., Chavenel, G. & Burtin, P.: Proc. Natl. Acad. Sci. 69: 2492-2494, 1972). Some preparations of CEA are bound by antibodies to blood group A or B antigens and most are bound by antibodies to Lewis a. (Holburn, A. M., Mach, J. P., MacDonald, D. and Newlands, M., Immunology 26: 831-843, 1974.)

(c) Isoelectric point. CEA-S.sub.1 has a single discrete isoelectric point of pI = 4.5 .+-. 0.1 using either isoelectric columns with ampholytes (LKB) with a pH range of 3-6 or thin layer pH 3-9 acrylamide gels containing ampholytes. By contrast, CEA has a variable distribution over a range of pH 3-5.2.

(d) Molecular exclusion chromatography. Characterization of CEA-S.sub.1 by Sephadex G-200 gel exclusion chromatography gives values indistinguishable from CEA with: Kd = 0.17; D.sub.20,.sub.w .degree. = 3.05 .times. 10.sup.-4 cm.sup.2 /sec; mean estimated stokes radius = 65 A (Plow, E. F. & Edgington, T. S.: Int. J. Cancer 15: 748-761, 1975 and unpublished data).

(e) Buoyant density. CEA-S.sub.1 has a buoyant density, when isolated from typical colonic adenocarcinomas, of 1.41 .+-. 0.02 gm/ml determined by isopycnic equilibrium density gradient ultracentrifugation in cesium chloride. By contrast, CEA gives a major peak of approximately 1.36 gm/ml with a wide range of density distributions various subfractions (Plow, E. F. & Edgington, T. S.: Int. J. Cancer 15: 748-761, 1975, and unpublished data).

(f) Estimated molecular weight. Using the Svedberg equation and estimated molecular weight of 181,000 daltons has been determined for CEA-S as contrasted to 201,000 daltons for CEA.

(g) Amino acid composition. The amino acid composition of the single polypeptide chain of CEA-S.sub.1 is similar to CEA.

______________________________________Amino Acid Comparison of CEA and CEA-S.sub.1Isolated from the Same TumorCEA.sub.0.1.3 CEA-S.sub.1.1.3Amino (Rel. % (Rel. %Acid (M/10.sup.5 /cm*) of AA) (M/10.sup.5 /gm*) of AA)______________________________________Lys 27.0 3.0 30.0 3.3His 16.4 1.8 15.9 1.8Arg 32.0 3.5 33.3 3.6Asp 131.1 14.4 127.6 14.0Thr 80.5 8.8 76.2 8.4Ser 92.9 10.2 90.4 9.9Glu 98.8 10.8 97.6 10.7Pro 67.9 7.4 69.3 7.6Gly 58.2 6.4 56.2 6.2Ala 53.3 5.8 52.7 5.8Val 65.2 7.4 64.5 7.1Met 1.9 0.2 1.9 0.2Ileu 44.3 4.9 42.3 4.6Leu 78.9 8.6 75.8 8.3Tyr 32.9 3.6 31.2 3.4Phe 23.3 2.6 21.6 2.4Try 9.9 1.1 12.6 1.4Cys (1/2) 12.2 1.3 14.5 1.6______________________________________ *Based on 33.9% protein as determined from E.sub.280.sup.1% = 5.44 and amino acid yield on a Beckman 121 M amino acid analyser.

(h) Acrylamide gel electrophoresis in 1% sodium dodecyl sulfate at pH 7.0. (method of Weber, K. & Osborn, M.: J. Biol. Chem. 244: 4400-4412, 1969). CEA-S.sub.1 gives a single discrete band with a relative mobility (Rm) of 0.143. It exhibits a slightly lower mobility and is considerably more discrete than CEA.

(i) Radioimmunoassay. CEA-S.sub.1 has been used in radioimmunoassays for the detection, evaluation and management of cancer of the gastrointestinal tract (Edgington, T. S., Astarita, R. W., and Plow, E. F.: New Engl. J. Med. 293: 103-107, 1975). Double blind clinical studies have been conducted to compare CEA-S.sub.1 assays with CEA assays of the same sera. CEA-S.sub.1 is also distinguishable from CEA by: (a) quantitative antigen binding assays with anti-CEA-S.sub.1 and various anti-CEA sera and by (b) competitive inhibition assays using either .sup.125 I CEA-S.sub.1 or .sup.125 I CEA as ligands and either anti-CEA-S.sub.1 or anti-CEA as antisera. In clinical studies CEA-S.sub.1 assay of serum has given less than 1% false positive results (0.23 and 0.41%) as compared to 3-19% false positive for CEA assay (Hansen, H. J.: Human Pathology 5: 139-147, 1974). Detection of cancer of the gastrointestinal tract was 80.4% as compared to 72% with CEA assay and detection of non-gastrointestinal tumors has been very low as compared to equivalent detection by CEA assay.

3. CEA-S.sub.1 radioimmunoassays.

The CEA-S.sub.1 assay employs three features:

(i) .sup.125 I CEA-S.sub.1 of colonic adenocarcinoma origin;

(ii) anti-CEA-S.sub.1 antiserum or selected anti-CEA antiserum exhibiting 2 to 3 fold preferential binding of .sup.125 I CEA-S.sub.1 ;

(iii) High ionic strength buffers such as 0.14 M NaCl, 0.10 M sodium borate buffer, pH 8.2.

In the CEA-S.sub.1 radioimmunoassay, serum and other body fluids are directly assayed by equilibrium competitive inhibition radioimmunoassay of double antibody type (Plow, E. F. & Edgington, T. S.: Int. J. Cancer 15: 748-761, 1975; Edgington, T. S., Astarita, R. W., & Plow, E. F.: New Engl. J. Med. 293: 103-107, 1975). Test samples are clarified (usually at 7,000 .times. g for 10 minutes) and 0.33 ml are transferred with 0.67 ml. of buffer (0.14 M sodium chloride in 0.10 M sodium borate, pH 8.2) to reaction tubes (usually 12 .times. 75 mm). To this is added: (a) 0.25 ml. of .sup.125 I CEA-S.sub.1 (approximately 150-200 pg at 100.mu. Ci .sup.125 I/.mu.g) diluted in 1:33 normal rabbit serum in borate buffer and (b) 0.25 ml. of anti-CEA antiserum (usually rabbit) diluted to give approximately 50-55% binding of the .sup.125 I CEA-S.sub.1. After incubation (approximately 18 hours) at 4.degree. C., 0.5 ml. of hyperimmune second antisera (typically goat anti-rabbit IgG) in mild antibody excess is added. Following approximately 5 hour incubation at 4.degree. C., the tubes are centrifuged and 1.0 ml. of supernatant is automatically transferred to fresh tubes and counted for .sup.125 I. The bound-free ratio is calculated and plotted against arbitrary or absolute standards of tumor CEA-S.sub.1. Unknown samples are quantitated by reference to the standard curve. Alternative radioimmunoassay techniques can be utilized and include sequential (non-equilibrium) assays, the use of counting of the precipitates, variations of time of incubation, and various solid-phase assay techniques provided that the critical elements essential to the assay are maintained.

Isolation and Characterization of CEA-S.sub.1

General Plan

A single homogeneous isomeric species of carcinoembryonic antigen is isolated by reference to solubility in 0.9 M perchloric acid, isoelectric focusing, molecular exclusion chromatography, ion exchange chromatography, passage through immunoabsorbants, and isopyknic density gradient ultracentrifugation. The final product, representing approximately 1.8%-10% of the perchloric acid soluble glycoprotein of the tumor, is homogeneous and devoid of other proteins by polyacrylamide gel electrophoresis. This single species of carcinoembryonic antigen, CEA-S.sub.1, has a sedimentation velocity of 6.6, a diffusion constant of 3.05 .times. 10.sup.-7 cm.sup.2 /sec, a mean Stokes radius of 65 A, a density of 1.41 ml/g in cesium chloride and an estimated molecular weight of 181,000, and it is devoid of detectable A, B, Rh, Kell, Duffy or Lewis a (Le-a) blood-group antigens. Immunochemical studies demonstrate qualitative similarities between CEA-S.sub.1 and conventional carcinoembryonic antigens; however, competitive inhibition analyses demonstrate significant quantitative immunochemical differences between CEA-S.sub.1 and preparations of carcinoembryonic antigen. These results are consistent with the concept that CEA-S.sub.1 is an immunochemical isomer of carcinoembryonic antigen.

Exemplary Procedure

The following procedure is an example of the best mode of practicing the invention and does not limit the scope of the invention.

Source and extraction of CEA-S.sub.1. CEA-S is isolated from 300-600 g of tumor, representing hepatic metastases of primary adenocarcinomas of the colon. Normal tissue is removed by dissection and the tumor is minced, homogenized with approximately 1.5 vol cold H.sub.2 O or saline in a Polytron homogenizer, and passed through gauze or a 25-mesh screen with washes of cold H.sub.2 O or saline. The tumor is suspended in a total of 2 vol H.sub.2 O and stirred for 2-4 h at 4.degree. C. The homogenate is clarified by ultracentrifugation at 22,000 .times. g for 16-18 h at 4.degree. C. in a Spinco 15 rotor. The sediment is homogenized with 1 vol 1 M NaCl for 4-6 h at 4.degree. C. and again clarified by ultracentrifugation. Supernatants from the first and second extractions are combined. An equal volume of cold 1.8 M perchloric acid (PCA) is slowly added dropwise to the tumor extract at 4.degree. C., and after continuous stirring for 1 h the precipitate is removed by centrifugation at 1,800 .times. g for 1 h at 4.degree. C. The supernatant containing the glycoprotein is brought to neutrality, then dialyzed exhaustively against distilled water, the final dialysis also containing mixed bed ion exchange resin (Biorad). (See Freedman, et al, U.S. Pat. No. 3,663,684). To facilitate detection of protein during subsequent isolation, the PCA-soluble glycoprotein fraction may be trace:labelled with .sup.125 I or may be followed by absorbance at 280 nm (E.sub.280.sup.1% = 5.44).

Isoelectric focusing. Isoelectric focusing of the PCA-soluble glycoprotein fraction is performed in a 440 ml LKB electrofocusing column using a 5 to 30% sucrose gradient containing 1% ampholytes, pH 3-6. Electrofocusing is conducted for 60 h at 3.degree. C. maintaining maximum power of approximately 6 watts and increasing to a maximum potential difference at 700 V. The column is eluted and 5 ml fractions are collected.

Ion exchange chromatography. Ion exchange chromatography is performed on a 2.5 .times. 30 cm column of DEAE-cellulose (Schleicher and Schuell, Inc., Keene, N. H., USA, 0.71 meq/g) at 20.degree.-24.degree. C. equilibrated with 0.05 M Na.sub.2 HOP.sub.4 starting buffer. The sample is applied following dialysis against the starting buffer, and a flow rate of 30 ml h is maintained until all of the initial non-bound fraction is eluted. The bound fraction is eluted with 0.50 M NaH.sub.2 PO.sub.4.

Molecular exclusion chromatography. Preparative molecular exclusion chromatography is performed first on Biorad A 1.5 fine agarose and then on Sephadex G-200 (2.5 .times. 80 cm column) both equilibrated with 0.14 M NaCl, 0.01 M sodium phosphate, pH 6.0. A flow rate of 5-6 ml/cm.sup.2 cross-sectional area/hr for A 1.5 and 3.5-4.0 ml/cm.sup.2 cross sectional-area/h for Sephadex G-200 is used. The Kd of the Sephadex G-200 is calculated using a dextran blue determined Vo of 160 ml and a Na.sup.51 Cr determined Vi of 234 ml (Gelotte, 1960). The diffusion constant (D.sub.20,w) and Stokes radius of CEA-S.sub.1 are estimated graphically from the Kd, utilizing proteins with known physical constants for calibration (Siegel and Monty, 1966). Standards included human figrinogen, catalase and albumin.

Immunoabsorbants. The gamma globulin fraction of 5 ml each of human antiserum specific for blood-group A and for B antigens (Spectra Biologicals, Oxnard, Calif., U.S.A.) and 10 ml each anti-human serum (Hyland) are prepared by precipitation at 50% saturation with ammonium sulfate at 4.degree. C. The gamma globulin fraction is dialyzed against 0.1 M sodium borate buffer pH 8.2, and covalently coupled to 200 ml of Sepharose 2B (Pharmacia Fine Chemicals, Uppsala, Sweden) activated at pH 11.0 in 0.10 M sodium phosphate with cyanogen bromide (Cuatrecases et al., 1968). The immunoabsorbant containing 4-6 mg. protein/ml agarose beads is retained in a 2.5 .times. 60 cm column. A separate immunoabsorbant containing only the gamma globulin from 10 ml of high titer anti-A antiserum is coupled to 50 ml of Sepharose 2B and is used in a 15 .times. 30 cm column. The CEA-S.sub.1 preparations are passed over these immunoabsorbants at a rate of 4-6 ml/h at 4.degree. C.

Isopyknic density gradient ultracentrifugation. Preparations are thoroughly dialyzed to water and adjusted to a density of 1.40 with CsCl. Gradients are generated and brought to equilibrium at 76,000 .times. g (mean) for 66 h at 20.degree. C. in a Spinco SW-41 rotor. Gradients are eluted in 0.2 ml fractions from the top using an ISCO fractionator. Density is determined in a Bausch and Lomb refractometer and calculated from the Critical Tables. The concentration of CEA-S.sub.1 has been estimated from an extinction coefficient (E.sub.280.sup.1% = 5.44) determined gravimetrically from partially purified CEA-S.

Sedimentation velocity. The sedimentation velocity of radiolabelled purified CEA-S.sub.1 is estimated in linear 5 to 25% sucrose gradients (in PBS) at 20.degree. C. .sup.125 I CEA-S.sub.1 is mixed with protein standards in 0.2 ml. layered onto 5 ml gradients and centrifuged at 78,000 .times. g for 40 h at 20.degree. C. in a Spinco SW-50.1 rotor. The gradients are fractionated from the top in 0.2 ml aliquots. Standards consist of purified fibrinogen, .sup.131 I F(ab').sub.2, and fresh serum is used as a source of IgG. The position of fibrinogen and IgG is determined by radial immunodiffusion. CEA-S and F(ab').sub.2 are determined by reference to specific radioactivity.

Estimation of molecular weight. Molecular weight is estimated from the Svedberg equation ##EQU1## where s equals the sedimentation coefficient (S.sub.20,w .degree.), D equals diffusion constant (D.sub.20,w .degree.), v equals partial specific volume, .rho. equals density of the solution for sedimentation coefficient analysis but has been corrected to 1.00 in the determination of S.sub.20,w .degree., R equals the gas constant and T equals absolute temperature. Partial specific volume is estimated as the reciprocal of the observed density of the .sup.125 I labelled molecule upon isopyknic density gradient ultracentrifugation in CsCl.

Preparation of CEA. Conventional CEA is prepared from the PCA-soluble glycoprotein fraction of metastatic adenocarcinoma of the colon by molecular exclusion chromatography on Sepharose 6B (Biorad) followed by Sephadex G-200. One ml samples are fractionated at 8 ml/cm.sup.2 cross-sectional area/h on a 2.5 .times. 80 cm column of Sepharose 6B (Biorad). The effluent is monitored at 280 nm for protein and collected in 5 ml aliquots. Two proteins peaks exhibiting CEA or related antigenic activity are regularly observed as described by Coligan et al. (1972). The fraction, with a S.sub.20,w .degree. of 6.8, identical with the value described by Coligan et al. (1972), is further reisolated by Kd 0.17 from a similar column of Sephadex G-200 and is used as conventional CEA. Preparations of purified CEA were also kindly provided from the laboratory of Dr. C. Todd (City of Hope, Calif., U.S.A.) and the British Medical Research Council, National Institute for Biological Standards and Control, (Lot 73/601), Holly Hill, London.

Polyacrylamide gel electrophoresis. Analytical polyacrylamide gel electrophoresis is performed on 5 .times. 75 mm cylindrical gels utilizing a Bio-Rad Model 150 A apparatus. Discontinuous polyacrylamide gel electrophoresis is performed utilizing both sample and spacer gels under either (1) alkaline conditions in tris-glycine, at pH 8.4 or (2) acid conditions in B-alanine-acetic acid, pH 4.5 as described by Ornstein (1964) using 6% polyacrylamide sample gels. Electrophoresis in the presence of 1% sodium dodecyl-sulfate is performed on 4% polyacrylamide gels according to the method of Weber and Osborn (1969). In each case the gels are sliced into 2 mm sections and the mobility and distribution of .sup.125 I CEA-S.sub.1 are determined by reference to radioactivity.

Assay for Blood Groups Antigens. The presence of blood-group A and B antigens in preparations is determined by the hemagglutination-inhibition assay using an ultramicrohemagglutination Coombs technique (Linder and Edgington, 1971). The presence of these blood group antigens as well as Rh, Kell, Duffy and Lewis a antigens has been determined by the ability of specific antisera to bind .sup.125 I CEA-S.sub.1, .sup.125 I CEA or equivalent preparations.

Radioiodination of CEA-S.sub.1. CEA-S is radiolabelled with .sup.125 I by a modified Chloramine-T micromethod (Chisari et al., 1974). As little as 180 ng of CEA-S in 1 .mu.l is effectively labelled to a specific activity of 80-100.mu. Ci/.mu.g. The radioiodinated product retains a density of 1.41 in an isopyknic CsCl density gradient; and more than 97% of the radioactivity is precipitable by 15% trichloracetic acid at 4.degree. C. and by specific antibody. Precipitability does not deteriorate on storage for at least two months.

Antisera. Antiserum to CEA is prepared in rabbits by primary immunization with a) 600 .mu.g of the PCA-soluble glycoprotein fraction or (b) 50 .mu.g purified CEA emulsified in complete Freuend's adjuvant which is injected in the rear footpads. Subsequent booster immunizations of the same dose are given in incomplete adjuvant in multiple subcutaneous sites at 3- to 4-weekly intervals. Antiserum for this study has been collected at 2-3 months and later in the response and rendered specific by exhaustive absorption with normal human serum, normal erythrocyte stroma, porcine and equine soluble blood-group A and B substances, (Dade, Miami, Fla., U.S.A.) and lyophilized homogenates of normal adult colon, lung and liver. This anti-serum, anti-CEA, is monospecific by Ouchterlony gel diffusion against the crude saline tumor extract, the PCA-soluble glycoprotein fraction, conventional CEA or partially purified CEA-S.sub.1 recovered from the immunoabsorbant column. This antiserum forms a line of identity with that (anti-CEA.sub.M) kindly provided by Dr. J. P. Mach (University of Lausanne, Switzerland), an antiserum which was compared with the original anti-CEA antiserum of Gold, et al. The above antiserum binds radiolabelled CEA-S.sub.1, conventional and purified CEA.sub.(Be) from the laboratory of Dr. C. Todd (City of Hope. California, U.S.A.).

Hyperimmune antiserum to rabbit IgG is produced in goats immunized subcutaneously on a monthly schedule with 100 .mu.g of rabbit IgG in incomplete adjuvant. The antiserum is adjusted to an antigen binding capacity at equivalence of .about.0.6 mg IgG/ml for use as the second antibody in the double antibody radioimmunoassay. Antigen binding radioimmunoassays. The capacity of antisera to bind CEA-S.sub.1 is determined by radioimmunoassay consisting of: (1) 0.25 ml of .sup.125 I CEA-S.sub.1 (0.18 ng) in 1:33 normal rabbit serum, 0.14 M NaCl, 0.1 M sodium borate pH 8.2 (RSB): (2) 0.25 ml of serial dilutions from 1:33 of the anti-CEA antiserum in RSB: and (3) 1.0 ml 0.14 M NaCl, 0.1 M sodium borate, pH 8.2 (BB) in a 12 .times. 75 mm tube. Following 18-20 h incubation at 4.degree. C., (4) 0.50 ml of goat anti-rabbit IgG second antibody diluted in BB is added and incubated 5 h at 4.degree. C. Phase separation is accomplished by centrifugation at 1,500 .times. g for 20 min, and 1.0 ml of supernatant is sampled and counted to determine free .sup.125 I CEA-S.sub.1. The bound/free ratio is also calculated.

Competitive inhibition assay of CEA-S.sub.1. CEA-S.sub.1 is assayed by equilibrium competitive inhibition radioimmunoassay of double antibody type similar to that previously described (Plow and Edgington, 1973a), modified with respect to compartment volume and reagent concentrations. Test samples are used directly except for serum which is first clarified at 7,000 .times. g for 10 min. The assay is performed in duplicate and consists of: (1) 0.33 ml of sample which is transferred with 0.67 ml of BB to 12 .times. 75 mm tubes; (2) 0.25 ml of .sup.125 I CEA-S.sub.1 (0.18 ng) in RSB; and (3) 0.25 ml of anti-CEA antiserum diluted in RSB to give 50-60% binding of .sup.125 I CEA-S.sub.1. Following 18 h incubation at 4.degree.-6.degree. C., (4) 0.50 ml of anti-rabbit IgG second antibody is added. After an additional 5 h incubation at 4.degree.-6.degree. C. the tubes are centrifuged at 1,500 .times. g for 20 min. One ml of supernatant is transferred to fresh tubes and counted for .sup.125 I. Controls include trichloric acid precipitable control for maximum precipitation, and no anti-CEA and no precipitating antibody controls for minimum Precipitation. Data are processed and reduced to .sup.125 I CEA-S.sub.1 bound/free ratio, which in turn is plotted against an arbitrary CEA-S.sub.1 standard. This standard consists of an ultracentrifugally clarified saline extract of a lyophilized homogenate of an adenocarcinoma of the colon. It is given a value of 33,300 units/ml. The inhibition profile is linear over the range .about. 6 to .about. 90 units/ml when plotted against the log of competing antigen concentration. The coefficient of variation of the assay, at 13.9 units/ml, is 6.5%.

Soluble tissue constituents were extracted from tumors by homogenization, incubation in 1M NaCl, which are equivalent, and and clarification by ultracentrifugation.

Twelve preparations from primary colonic adenocarcinomas reacted with anti-CEA by gel diffusion and gave concentrations of CEA-S.sub.1 varying from 2,900 to 59,840 units/g and a mean concentration of 30,633 units/g tumor. Metastatic adenocarcinomas contained an average of 6,700 units CEA-S.sub.1 /g tumor. Subsequent addition of perchloric acid at concentrations from 0 to 2.0 M demonstrated no significant loss of CEA-S.sub.1 from the dialyzed supernatant. A concentration of 0.90 M perchloric acid as described by Krupey et al (1968) was accepted and utilized in this study. Subsequent studies have employed 0.60 M PCA.

When such glycoprotein preparations were fractionated on isoelectric focusing columns, marked heterogeneity of net molecular charge was observed for the prototype preparation (Plow and Edgington, Int. J. Cancer: 15, at p. 753). Four major CEA related glycoprotein fractions were recovered in pH 3-6 gradients. These had isoelectric points of approximately 3.8, 4.1, 4.5 and 5.4 with the prototype preparation; the pI 4.5 peak appeared to consist of three components recognized by monitoring of trechloracetic acid precipitate or by optical density at 280 nm. Each of these species of CEA exhibited apparent gross immunochemical identity by gel diffusion against anti-CEA. Significant differences have been observed between protein peaks detected at 280 nm and those detected by reference to .sup.125 I, indicating that different species of glycoproteins are radioiodinated to a different extent. In this respect, a minor pI 5.4 CEA species was virtually unlabelled. CEA-related material isolated from the gradient in the vicinity of pH 4.5 has been utilized for isolation of CEA-S.sub.1.

The pH 4.25-4.80 fraction of preparation has been fractionated into several sizes of glycoproteins by molecular exclusion chromatography on A 1.5 fine then on Sephadex G-200. The second peak of CEA from A 1.5 fine (sedimentation velocity approximately 6.8) has been utilized for subsequent purification, though the first peak is also reactive with anti-CEA anti-sera in gel diffusion assay and radioimmunoassay for CEA. The second and major peak with a Kd of 0.17 from Sephadex G-200 has been retained for subsequent isolation. This fraction has been resolved into two major fractions as well as minor fractions by ion-exchange chromatography on DEAE-cellulose using 0.05 M Na.sub.2 HPO.sub.4 as the equilibration and eluting buffer, as exemplified by the prototype preparation. The initial fraction of the applied glycoprotein was retarded only slightly by the resin, and two minor fractions as well as a major component were recovered by further elution with 0.50 M NaH.sub.2 PO.sub.4.

The partially purified CEA-S.sub.1 has been subjected to two passages through immunoabsorbant columns containing antiserum to human serum and to blood-group antigens A and B. Preparations with high concentrations of A-like antigen have required additional passages through separate anti-A immunoabsorbant columns to achieve complete removal of all detectable A-like antigen. Table 1 summarizes the sequential removal of A-like antigen from prototype CEA-S.sub.1. Material bound by the immunoabsorbant and subsequently eluted with 3 M potassium iodide, pH 7.4, did not possess significant CEA-S.sub.1 by radioimmunoassay though the recovered CEA-S.sub.1 fraction was highly reactive by both radioimmunoassay and gel diffusion using anti-CEA antiserum.

Table 1______________________________________Removal of Blood-Group A RelatedAntigen From Prototype CEA-S.sub.1By Immunoabsorbants A antigen.sup.1 (titer)Passage Immunoabsorbant of preparation______________________________________None None 1:1281 I Anti-A, anti-B, antiserum 1:322 I Anti-A, anti-B, antiserum 1:163 II Anti-A 1:44 II Anti-A Neg______________________________________ .sup.1 Titer of sample (5 .mu.l) required to neutralize anti-A antiserum (1 .mu.l) with a titer of 1:512.

Final purification of CEA-S.sub.1 recovered from the immunoabsorbants has been achieved by isopyknic density gradient ultracentrifugations in CsCl. After 66 h of centrifugation at 76,000 .times. g in a Spinco SW-41 rotor, three components of mean density <1.30, 1.41 and 1.52 were resolved for the prototype preparation. The relative distribution of the three components has varied with preparations from different tumors due to the degree of glycosylation of different species of CEA. The peak extending from 1.37 to 1.45 g/ml with a mean density of 1.41 g/ml is retained as purified CEA-S.sub.1. The peak of higher density, although reactive with anti-CEA, was relatively less reactive in assay of CEA-S.sub.1. The approximate recovery of glycoprotein in CEA-S.sub.1 fractions from each purification relative to the PCA extract is summarized in Table II. The final recovery of CEA-S.sub.1 from this tumor has represented less than 2.0% of the radioiodinated glycoprotein present in the PCA-soluble tumor extract.

Table II______________________________________Relative Recovery of CEA-S.sub.1 Containing Fractionsfrom PCA-Soluble Tumor Glycoprotein (Tumor #1) Recovery.sup.1Procedures (%)______________________________________(1) Isolectric focusing 55(2) Molecular exclusion 43(3) Ion exchange 52(4) Immunoabsorbant 58(5) Isopyknic density gradient 25 Final recovery 1.8______________________________________ .sup.1 Recovery is calculated at the % of the applied .sup.125 I glycoprotein radioactivity recovery in the CEA-S.sub.1 pool from each purification procedure. Recovery differs with different tumors and is comparable when analyzed by absorbance of 280 nm for protein.

Certain features of CEA-S.sub.1 have been established by reference to behavior during isolation, i.e., CEA-S.sub.1 has a pI of approximately 4.5, a buoyant density of 1.41 g/ml and is not associated with blood-group antigens A or B, Rh, Kell, Duffy or Lewis (Le -a). The homogeneity of CEA-S.sub.1 in acid, alkaline and SDS polyacrylamide gels was demonstrated by polyacrylamide gel electrophoresis of purified .sup.125 I CEA-S.sub.1 in 5 .times. 75 mm cylindrical gels using a 4% gel in the presence of 1% sodium dodecyl sulfate, a 6% discontinuous gel in B-alanine acetic acid pH 4.5, and a 6% discontinuous gel in tris-glycine, pH 8.4. (Plow and Edgington, Int. J. Cancer: 15, at p. 756) Under each analytical system CEA-S.sub.1 migrated as a single discrete band, and more than 85% of the applied radioactivity was recovered in the single symmetrical band. In sucrose density gradients, CEA-S.sub.1 sedimented as a discrete and homogeneous species. On the basis of its behavior relative to proteins with known sedimentation coefficients, a S.sub.20,w of 6.6 was estimated for CEA-S.sub.1 using sedimentation velocity ultracentrifugation of .sup.125 I CEA-S.sub.1 in a 5 to 25% sucrose gradient. The ultracentrifugation was performed in a SW-50.1 Spinco rotor at 78,000 .times. g at 20.degree. C. for 40 h. Protein standards utilized to estimate the S.sub.20,w of CEA-S.sub.1 were included in the same tube and were fibrinogen IgG and F(ab').sub.2 fragment of IgG. The diffusion constant and mean Stokes radius of CEA-S.sub.1 were analyzed from molecular exclusion behavior on Sephadex G-200 (Siegel and Monty, 1966). By means of such calibrated columns a Kd of 0.17 was determined for CEA-S.sub.1 graphically. The Kd of CEA-S.sub.1 and standard proteins-fibrinogen, catalase and albumin were determined by molecular exclusion chromatography in Sephadex G-200. The Kd or derivative (Kd).sup.1/3 was plotted against established diffusion constants and Stokes radii of the standard proteins. CEA-S.sub.1 has an estimated diffusion constant (D.sub.20,w) of 3.05 .times. 10.sup.-7 cm.sup.2 /sec and mean Stokes radius of approximately 65 A. The molecular parameters of CEA-S.sub.1 determined in this study are summarized in Table III.

Table III______________________________________Summary of The Molecular Parametersof CEA-S.sub.1Physical constant Value______________________________________pI 4.5S.sub.20,w.sup.o 6.6Kd (G-200) 0.17D.sub.20,w.sup.o 3.05 .times. 10.sup.-1 cm.sup.2 /secStokes radius 65 A.rho. 1.41 g/ml- v 0.709 ml/gM 181,000______________________________________

Immunochemical characterization of CEA-S.sub.1 was performed in double antibody radioimmunoassays utilizing .sup.125 I-CEA-S.sub.1 as the ligand and anti-CEA. The ligand was completely bound by antiserum at a dilution of 1:50 of anti-CEA antiserum and on serial dilution of the antiserum a typical semilogarithmic binding profile with a single component was generated, suggesting the presence of only one antigenic species in the ligand. In competitive inhibition immunoassays binding of the ligand was completely inhibited by: (1) standard tumor extracts; (2) serum from a patient with adenocarcinoma of the colon; (3) conventional CEA; and (4) purified CEA-S.sub.1. Slight differences in competitive inhibition slopes of the competing antigens were within the precision of the assay and suggest that the qualitative antigenic expression of CEA-S.sub.1 in each of these materials is similar if not identical. (Plow and Edgington, Int. J. Cancer: 15, at p. 758) Significant quantitative differences were observed in the immunochemical expression of conventional CEA as compared to CEA-S.sub.1. Identical results were obtained also with different preparations of conventional CEA including preparations Be from the laboratory of Dr. C. Todd and standard CEA prepared by the British Medical Research Council. As compared to the arbitrary CEA-S.sub.1 standard, a purified preparation of CEA-S.sub.1, standardized by reference to absorbance at 280 nm, gave a value of 0.64 ng/unit CEA-S whereas several preparations of CEA required an approximately 5.26 times greater concentration to produce equivalent 50% competitive inhibition. This was equivalent to 3.37 ng CEA/unit of CEA-S.sub.1.

Discussion

Characterization of CEA by Krupey et al. (1968) suggested that such preparations contained a single homogeneous glycoprotein. Homogeneity of conventional CEA preparations was suggested by the studies of Terry et al. (1972) which demonstrated that the major protein constituent of five preparations from different colonic tumors possessed identical sequences through the first 24 amino acid residues. Subsequent studies have demonstrated that preparations of CEA exhibit heterogeneity with respect to a number of analytical parameters. Upon isoelectric focusing (Rule and Goleski-Reilly, 1974) and ion exchange chromatography (Coligan et al., 1973; Eveleigh, 1974), preparations of purified CEA have been resolved into multiple species; and even in the analysis of a preparation of CEA homogeneous with respect to isoelectric point, Turner et al. (1972) observed three discrete components differing in density by isopyknic density gradient ultracentrifugation. Heterogeneity of CEA has also been demonstrated with respect to the expression of blood-group-related antigens; some preparations of CEA appear to be essentially devoid of A-like blood-group determinants, while in other studies blood-group antigens appear to be present on the same molecule expressing CEA determinants (Gold and Gold, 1973; Simmons and Perlmann, 1973). It has been suggested that CEA also possesses blood group Lewis a (Le-a) antigen. (Holburn, A.M. et al, Immunology 26: 831-834, 1974.) Furthermore, since CEA is defined immunochemically, normal serum and tissue constituents which share antigenic determinants and exhibit varying degrees of immunochemical cross reaction with CEA (Mach and Pusztaszeri, 1972; von Kleist et al., 1972) might also be considered as potential heterogeneous species of CEA. The molecular basis and significance of the observed heterogeneity have not been established. Although variation in content of N-acetyl neuraminic acid does influence behavior in charge-related analytical systems, following complete removal of N-acetyl neuraminic acid some residual heterogeneity can still be demonstrated (Coligan et al., 1973). Further, removal of N-acetyl neuraminic acid from CEA does not render it equivalent to CEA-S.sub.1 in immunochemical reactivity. (Edgington, T.S. and Plow, E.F., unpublished).

The previously described heterogeneity of CEA-related glycoproteins with respect to isoelectric point, behavior in ion exchange and density has been observed in the isolation of CEA-S.sub.1 and additional degrees of heterogeneity have been established. The demonstration that CEA-related glycoproteins with a specific isoelectric point can be further fractionated by ion exchange chromatography indicates that heterogeneity of these molecules must reflect differences between species in local charge density as well as net molecular charge. The resolution of CEA into three components by isopyknic density gradient ultracentrifugation following the removal of blood-group A-like activity via immunoabsorption indicates that the presence of a species with a density of 1.52 g/ml is not dependent only upon blood-group activity as previously suggested by Turner et al. (1972). The recovery of CEA-S.sub.1 from anti-A immunoabsorbant columns establishes that CEA and blood-group A determinants can reside on independent molecules. Although some previous preparations of CEA have been reported to be devoid of blood-group activity, the presence of blood-group antigens may reflect only the blood type of the donor rather than any effective or specific removal of these blood-group-related molecules during purification.

A molecular weight of 181,000 has been calculated for CEA-S.sub.1 according to the Svedberg equation. To obtain this value, it has been necessary to assume that the buoyant density of CEA-S.sub.1 in cesium chloride reflects the true partial specific volume of the molecule, that the diffusion constant estimated from the Kd of CEA-S.sub.1 on Sephadex G-200 was not altered by interactions with the gel matrix, and that the sedimentation of CEA-S.sub.1 in sucrose accurately reflects the sedimentation coefficient of the molecule. In simultaneous parallel analyses in sucrose gradient, a sedimentation coefficient of 6.8 S was determined for conventional CEA, a value identical to that reported by Coligan et al. (1972) which supports the validity of the 6.6 S value of CEA-S.sub.1. The v of 0.709 ml/g determined by ultracentrifugation in cesium chloride and provides the only currently available estimate for CEA-S.sub.1. Molecular weights of conventional CEA of approximately 200,000 to 300,000 daltons have been suggested (Coligan et al., 1972; Pusztaszeri and Mach, 1973); however, these estimates are based solely on molecular exclusion properties, a particularly unreliable approach with glycoproteins (Winzor, 1969). The similarity in sedimentation and gel filtration properties on G-200 for CEA and CEA-S.sub.1 suggests that these previous estimates may be high. Although CEA-S.sub.1 may be smaller than conventional preparations of CEA based on the slight differences in sedimentation coefficients, it is still significantly larger than described molecules that cross-react with CEA (Mach and Pusztaszeri, 1972; von Kleist et al., 1972; Burtin et al., 1973; Newman et al., 1974).

Immunochemical analyses utilizing the .sup.125 I-CEA-S.sub.1 ligand (Edgington and Plow, U.S. patent application Ser. No. 682,028 Filed Apr. 30, 1976, concurrently herewith, and Edgington, Astarita and Plow, New Eng. J. Med., 293: 103-107 (July 17, 1975) provide further basic information concerning the isolated molecule and establish its relationship to conventional CEA. First, these analyses indicate that CEA-S.sub.1 as isolated is antigenically comparable to those molecules present in tumors and the serum of certain cancer patients recognized by anti-CEA. The capacity of equilibrium competitive inhibition radioimmunoassays to discriminate minor structural differences has been well documented (Plow and Edgington, 1972, 197 3b; Edgington and Plow, 1975); and the similarity in competitive inhibition slopes of these samples indicates that, despite the extensive handling of CEA-S.sub.1 in isolation, the molecule must remain quite similar to molecules present in vivo. Second, the capacity of conventional CEA and CEA-S.sub.1 to yield complete competitive inhibition and generate curves with similar slopes strongly supports the concept that CEA-S.sub.1 is a species of CEA. Third, the 5.26-fold difference in the quantitative expression of CEA and CEA-S.sub.1 indicates that this isolated and characterized species of CEA is not immunochemically identical to the CEA glycoprotein set but must represent a specific isomer, a concept similar to that developed in studies of enzyme sets and their constituent isoenzymes. By inference other isomers of CEA must also be immunochemically distinct from CEA-S.sub.1.

The immunochemical differences between CEA and CEA-S.sub.1 may provide an explanation for the improved specificity of radioimmunoassays employing CEA-S.sub.1 as the radiolabelled ligand in recognition of patients with neoplasia (Edgington, 1975). In such assays only rare false positives have been observed; three of 662 patients without neoplasia but with various diseases including inflammatory bowel disease and advanced liver disease gave values greater than the selected diagnostic threshold value of 14 units CEA-S.sub.1 /ml serum. Although approximately 82% of patients with colonic adenocarcinomas and 92% of patients with pancreatic carcinomas were positive, only 14% of lung carcinoma and 15% of breast carcinoma patients had elevated serum CEA-S.sub.1. The low incidence of elevated CEA-S.sub.1 in association with these latter tumors and a 3.5% positive rate among other tumors is in marked distinction to other studies utilizing conventional CEA (Hansen, 1974; Lo Gerfo et al., 1971; Pusztaszeri and Mach, 1973). The applied diagnostic significance has been further supported by current double blind studies of 392 sera that indicate a highly significant discorrelation between levels of CEA and of CEA-S.sub.1 assayed in the same sera.

In one aspect the invention is an isolation and identification of a new carcinoembryonic antigen species, here denominated CEA-S.sub.1 and for future reference to distinguish this from all other possible species denominated CEA-S.sub.1 isolated from glycoprotein extract of human adenocarcinomas of the intestinal tract which is useful in the production of highly specific antibodies which are potentially useful in preparing immunotherapeutic reagents and is useful in preparing reagents for various diagnostic procedures. In the diagnosis of cancer of the human gastrointestinal tract the concentrate of isolated CEA-S.sub.1 may be tagged for detection, by radiolabelling with .sup.125 I or .sup.133 I, for example, or by fluorescein labelling for fluorescence detection, labelling with enzymes such as peroxidase or alkaline phosphatase, or may be used directly in visible agglutination detected diagnostic tests, all according to known and published techniques. It may be used to produce anti-CEA-S.sub.1 antisera for these immunodiagnostic assays or for immunofluorescent assays of tissue to demonstrate the presence of CEA-S.sub.1 or autogenically related species. Similarly, it may be employed to bind and purify antibodies of CEA-S.sub.1 by subsequent elution for CEA-S.sub.1. The antibodies can be labelled with isotopes, fluorescent compounds or enzymes or other antigens and employed in radioimmunoassays, in research for demonstration of CEA-S.sub.1 in cells, serum, body fluids and in experimental immunotherapy. At present, use of the new composition, isolated CEA-S.sub.1 as a radioimmunoassay reagent is most advantageous using the equipment available to the inventors, however, the form in which the new composition is used as a diagnostic research or therapeutic reagent depends upon the equipment available to a particular worker and upon the personal preferences and experience of the worker.

The new composition, isolated CEA-S.sub.1, is characterized by a number of physicochemical constants and by immunochemical reactions. Among the distinquishing physiochemical constants are the following.

Sedimentation velocity of CEA-S.sub.1 as measured by the standard sucrose gradient test, using linear 5 to 25% sucrose gradients in PBS at 20.degree. C., S.sub.20,w .degree., is 6.6 Svedberg units compared with 6.8 for CEA. While the absolute value in Svedberg units could vary between laboratories, apparently as a result of minor differences in technique and instrumentation, CEA-S.sub.1 has to date demonstrated a sedimentation velocity about 2 to 4 percent lower than the sedimentation velocity of CEA measured comparatively, using the same technique and instrumentation.

Isoelectric point is another distinguishing feature of CEA-S.sub.1. First, CEA-S.sub.1 is distinct from CEA in that CEA-S.sub.1 is demonstrated by isoelectric point to be a single distinct species producing a single distinct isoelectric point, pI = 4.5 .+-. 0.1, whereas CEA produces, in comparative tests using the same technique and instrumentation, a variable distribution from about pI = 3 to about pI = 4.2. pI for CEA-S.sub.1 was confirmed using two methods and while minor variations according to technique and instrumental errors may result, CEA-S.sub.1 will always produce a single discrete isoelectric point having a value of from about pI = 4.4 to about pI = 4.6.

Buoyant density also serves to distinguish CEA-S.sub.1 from CEA. Turner et al (J. Immunol., 108, 1328-1339, 1972) reported three bands of density in CEA, 1.47, 1.42 and 1.28, but until the present work of isolation no single species of CEA has been heretofore characterized and no isolation CEA-S.sub.1 composition has been prepared. CEA-S.sub.1 has been found consistently to have a buoyant density of about 1.41 .+-. 0.02 gm/ml in cesium chloride as distinguished in comparative measurements with CEA which demonstrated a wide range of density distribution having a broad major peak usually at about 1.36 gm/ml.

It is not possible to unequivocally determine an absolute value for the molecular weight of either CEA or CEA-S.sub.1 but comparative measurements of CEA and CEA-S.sub.1 using the same procedure and instrumentation and the same basic assumptions form a valid basis for distinquishing CEA-S.sub.1 from CEA. Using the Svedberg equation and the premises set forth hereinbefore, and comparative measurements, CEA-S.sub.1 can be stated to be a distinct species of CEA having an estimated molecular weight about 10 percent lower than the estimated molecular weight of CEA; i.e., 181,000 daltons for CEA-S.sub.1 as distinguished from 201,000 daltons CEA.

The relative mobility, Rm, of CEA-S.sub.1 is distinct from that of CEA in comparative measurements. Using the standard acrylamide gel electrophoresis in 1% sodium dodecyl sulfate at pH 7.0 (Weber et al, J. Biol. Chem. 244: 4400-4412, 1969) in comparative tests with bromophenol blue as the marker dye, CEA-S.sub.1 is distinct in that it demonstrates a single narrow, discrete band of about 4% lower mobility, Rm = 0.143, for CEA-S.sub.1 then the wide diffuse band at Rm .perspectiveto. 0.182 demonstrated by CEA.

As demonstrated, CEA-S.sub.1 is strikingly, and unexpectedly, distinct immunochemically from CEA, although it is inherently more difficult to quantify immunochemical differences than phyisicochemical differences. Immunochemically CEA-S.sub.1 is distinct from CEA in a number of ways, the following distinguishing features being selected as being somewhat more subject to relative quantification or definition than other features.

A comparison of antigen binding characteristics, namely the quantitative binding of either .sup.125 I CEA-S.sub.1 or .sup.125 I CEA by antisera to CEA or CEA-S.sub.1, provides one point of identification of CEA-S.sub.1 as compared with CEA. While absolute values vary between different antisera to CEA or CEA-S.sub.1, the antigen binding capacity (ABC) for CEA-S.sub.1 of anti-CEA antisera and anti-CEA-S.sub.1 antisera is significantly greater than for CEA. Antigen binding capacities of about 10% or greater for CEA-S.sub.1 than for CEA are significant and, as shown in Table IV, antigen binding capacities of anti-CEA antisera range typically from about 15% to about 50% greater for CEA-S.sub.1 than for CEA in parallel assays. There is approximately a three-fold preferential binding of CEA-S.sub.1 as compared to CEA by anti-CEA-S.sub.1 antisera in high ionic strength (higher than physiologic), e.g., 0.14 M NaCl, 0.10 M Na.sub.2 B.sub.4 O.sub.7, pH 8.2.

Table IV__________________________________________________________________________Comparative Antigen Binding Characteristicsof Antisera to CEA-S.sub.1 and CEA* Antigen Binding Characteristics CEA CEA-S.sub.1 Sample ABC Slope ABC SlopeAntisera No. (ug/ml) -(.times. 10.sup.3) (ug/ml) -(.times. 10.sup.3)__________________________________________________________________________Anti-CEA A 5.14 25.7 6.42 14.9 B 0.21 25.3 0.38 18.7 C 1.93 23.4 2.73 13.7 D 2.32 24.7 2.74 13.6 E 5.14 24.0 6.42 15.5 F 5.34 26.1 6.31 15.0 mean 24.9.+-.1.0 mean 15.2.+-.1.9Anti-CEA-S.sub.1 0.11 34.1 0.33 23.8__________________________________________________________________________ *See e.g.,Edgington, T.S. and Plow E.F., Functional Molecular Anatomy of Fibrinogen: Antibodies as Biological Probes of Structure, in CONTEMPORARY TOPICS IN MOLECULAR IMMUNOLOGY 2: 237-271, 1973; Plow, E.F. and Edgington T.S., Discriminating Neoantigenic Differences Between Fibrinogen and Fibrin Derivatives, Proc. Nat. Acad. Sci. USA 70: 1169-1173, 1973, for a comprehensive discussion of the validity of the approach exemplified in the data of this table.

The linear regression slope of the % bound of CEA and CEA-S.sub.1 vs. concentration of the antiserum also provides an important distinguishing characteristic of CEA-S.sub.1. The slope of the curves for two identical molecules will be equal, whereas the slope for CEA-S.sub.1 is significantly different from comparable curve for CEA, as demonstrated in Table IV, wherein the slope of the CEA-S.sub.1 curve has averaged about three-fifths the slope of the CEA curve. In general, the slope of the CEA-S.sub.1 curve will be from about one-half to about four-fifths the slope of the CEA curve in comparable tests. This has been found to be true of seven of nine anti-CEA and anti-CEA-S.sub.1 antisera which have been used.

Characteristic equilibrium competitive inhibition radioimmunoassay data for purified CEA-S.sub.1, conventional CEA, standard tumor extract, normal serum (N serum) and serum from a patient with adenocarcinoma of the colon (CA serum) are presented graphically in FIG. 1 of the drawing, and demonstrate the difference between the immunochemical expression in terms of slope and value of conventional CEA as compared with CEA-S.sub.1. The data presented in Table V further distinquish CEA-S.sub.1 from CEA by the quantity of antigen, CEA-S.sub.1 or CEA, (ng/m.) required for 50% inhibition of selected antisera and the slope of the competitive inhibition curve of CEA-S.sub.1 and CEA measured against these antisera.

Table V__________________________________________________________________________Non-Identity of CEA-S.sub.1 and CEA Demonstrated by EquilibriumCompetitive Inhibition Immunochemical AnalysisSystem 50% Inhibition Inhibition Slope* CEAAssay Antiserum Ligand CEA-S.sub.1 CEA CEA-S.sub.I CEA-S.sub.1 CEA__________________________________________________________________________ (ng/ml) -(.times. 10.sup.- 3)A Anti-CEA.sub.sc CEA-S.sub.1 25.0 267 10.7 28.2 31.6B Anti-CEA.sub.sc CEA 56.5 170 3.0 19.7 19.4C Anti-CEA-S.sub.1.1 CEA-S.sub.1 52.7 654 12.4 23.9 31.4 1.1__________________________________________________________________________ *Linear regression analysis. y = a + bx where .times. = slope and y = log.sub.10 competing antigen concentration.

It is to be clearly understood that the invention contemplates one composition, concentrated, isolated CEA-S.sub.1 (including the various forms such as radiolabelled or fluorescein labelled CEA-S) and that while great effort and care has been exercised in specifying the various distinguishing physicochemical characteristics, the invention is not a particular characteristic or a particular set characteristic -- the invention is CEA-S.sub.1, its method of isolation and concentration and its use. Consequently, while in the following claims the invention is defined in terms of physicochemical and immunochemical characteristics and related procedural steps, the substance of the invention is whether or not one has prepared or used CEA-S.sub.1 and not whether a particular characteristic is precisely accurate or whether a composition appears to exhibit some variation in some characteristic or set of characteristics by reason of differing techniques, skills, equipment, instruments or the like. Accordingly, recognizing limitations of the language, a full comprehension of the claimed invention requires recognition of the substance of the invention as hereinbefore stated and described.

Background References

Bjorklund, K. B. U.S. Pat. No. 3,823,126, July 9, 1974. Burtin, P., Chavanel, G., and Hirsch-Marie, H., Characterization of a second normal antigen that cross-reacts with CEA. J. Immunol., 111, 1926-1928 (1973). Chisari, F. V., Gerin, J. L., and Edgington, T. S., Immunochemistry of the hepatitis B virus. .sup.125 I HB Ag ligand. J. Immunol., 113, 543-553 (1974). Coligan, J. E., Henkart, P. A., Todd, C. W., and Terry, W. D., Heterogeneity of the carcinoembryonic antigen. Immunochemistry, 9, 377-387 (1972). Cuatrecasas, P., Wilchak, M., and Anfinsen, C. B., Selective enzyme purification by affinity chromatography. Biochemistry, 61, 636-643 (1968). Edgington, T. S., CEA-S: A distinctive isomeric variant of carcinoembryonic antigen. Fed. Proc., 34, 845 (1975). Edgington, T. S., and Plow, E. F., Conformational and structural modulation of the N-terminal regions of fibrinogen and fibrin associated with plasmin cleavage. J. Biol. Chem., in press (1975). Eveleigh, J. W., Heterogeneity of carcinoembryonic antigen. Cancer Res., 34, 2122-2124 (1974). Freedman, S. O., Gold, P., Krupey, J. H., U.S. Pat. No. 3,663,684, May 16, 1972. Gelotte, B. J., Studies on gel filtration. Sorption properties of the bed material sephadex. J. Chromatog., 3, 330-342 (1960). Gold, P., and Freedman, S. O., Specific carcinoembryonic antigens of the human digestive system. J. exp. Med., 122, 467-481 (1965). Gold, J. M., and Gold, P., The blood group A-like site on the carcinoembryonic antigen. Cancer Res., 33, 2821-2824 (1973). Goldenberg, D. M., U.S. Pat. No. 3,865,684, Feb. 11, 1975. Hansen, H. J., Carcinoembryonic antigen (CEA) assay. A laboratory adjunct of the diagnosis and management of cancer. Hum. Path., 5, 139-147 (1974). Hansen, H. J., U.S. Pat. Nos. 3,697,638, Oct. 10, 1972 and 3,867,363, Feb. 18, 1975. Krupey, J., Gold, P., and Freedman, S., Physicochemical studies of the carcinoembryonic antigens of the human digestive system. J. exp. Med., 128, 387-398 (1968). Linder, E. J., and Edgington, T. S., Ultramicro assay of anti-erythrocyte antibodies and erythrocyte antigens. Vox Sang. 21, 222-232 (1971). Lo Gerfo, P., Krupey, J., and Hansen, H. J., Demonstration of an antigen common to several varieties of neoplasia. New Engl. J. Med., 285, 138-141 (1971). Mach, J. P. and Pusztaszeri, G., Carcinoembryonic antigen (CEA): Demonstration of a partial identity between CEA and a normal glycoprotein. Immunochemistry, 9, 1031-1034 (1972). Moore, T. L., Kupchik, H. Z., Marcon, N., and Zamcheck, N., Carcinoembryonic antigen assay in cancer of the colon and pancreas and other digestive tract disorders. Amer. J. dig. Dis., 16, 1-7 (1971). Newman, E. S., Petras, S. E., Georgiadis, A., and Hanse, H. J., Interrelationship of carcinoembryonic antigen and colon carcinoma antigen III. Cancer Res., 34, 2125-2130 (1974). Ornstein, L., Disc electrophoresis. Ann. N.Y. Acad. Sci., 121, 321-349 (1964). Plow, E., and Edgington, T. S., Molecular events responsible for modulation of neoantigenic expression: The cleavage-associated neoantigen of fibrinogen. Proc. nat. Acad. Sci. (Wash.), 69, 208-212 (1972). Plow, E. F., and Edgington, T. S., Immunobiology of fibrinogen. Emergence of neoantigenic expressions during physiologic cleavage in vitro and in vivo. J. clin. Invest., 52, 273-282 (1973a). Plow, E. F. and Edgington, T. S., Discriminating neoantigenic differences between fibrinogen and fibrin derivatives. Proc. nat. Acad. Sci. (Wash.), 70, 1169-1173 (1973b). Pusztaszeri, G., and Mach, J. P., Carcinoembryonic antigen (CEA) in non-digestive cancerous and normal tissues. Immunochemistry, 10, 197-204 (1973). Rule, A. H., and Goleski-Reilly, C., Phase-specific oncocolon antigens: A theoretical framework for carcinoembryonic antigen specificites. Cancer Res., 34, 2083-2087 (1974). Siegel, L. M., and Monty, K. J., Determination of molecular weights and frictional ratios of proteins in impure systems by use of gel filtration and density gradient centrifugation. Application to crude preparations of sulfite and hydroxylamine reductases. Biochem. Biophys. Acta, 112, 346-362 (1966). Simmons, A. R., and Perlman, P., Carcinoembryonic antigen and blood group substances, Cancer Res., 33, 313-322 (1973). Terry, W. D., Henkart, P., Coligan, J. E., and Todd, C. W., Structural studies of the major glycoprotein preparations with carcinoembryonic antigen activity. J. exp. Med., 136, 200-204 (1972). Terry, W. D. Henkart, P. A., Coligan, J. E., and Todd, C. W., Carcinoembryonic antigen: Characterization and clinical applications. Transplant. Rev., 20, 100-129 (1974). Thompson, D. M. P., Krupey, J., Freedman, S. O., and Gold. P., The radioimmunoassay of circulating carcinoembryonic antigen of the human digestive system. Proc. Nat. Acad. Sci. (Wash.), 64, 161-167 (1969). Turner, M. D., Olivares, T. A., Harwell, L., and Kleinman, M. S., Further purification of perchlorate-soluble antigens from human colonic carcinomata. J. Immunol., 108, 1328-1339 (1972). von Kleist, S., Chavanel, G., and Burtin, P., Identification of an antigen from normal human tissue that crossreacts with the carcinoembryonic antigen, Proc. Nat. Acad. Sci. (Wash.), 69, 2492-2494 (1972). Weber, K., and Osborn, M., The reliability of molecular weight determinations by dodecyl sulfate-polyacrylamide gel electrophoresis. J. Biol. Chem., 244, 4406-4412 (1969). Winzor, D. J., Analytical gel filtration, In: S. J. Leach (ed.), Physical principles and techniques of protein chemistry, part A, pp. 451-494, Academic Press, New York (1969). Edgington, T. S. and Plow, E. F., Functional Molecular Anatomy of Fibrinogen: Antibodies as Biological Probes of Structure, CONTEMPORARY TOPICS IN MOLECULAR IMMONOLOGY 2: 237-271, 1973. Plow, E. F. and Edgington, T. S., Discriminating Neoantigenic Differences Between Fibrinogen and Fibrin Derivatives, Proc. Nat. Acad. Sci., USA 70: 1169-1173, 1973.

Claims

1. The isolated species of carcinoembryonic antigen denominated CEA-S.sub.1, said species being characterized by such physicochemical and immunochemical characteristics as:

(a) having a single, discrete isoelectric point of about 4.5, as measured by isoelectric columns with ampholytes having a pH range of 3 to 6 or by thin layer pH 3 to 9 acrylamide gels containing ampholytes;

(b) having a buoyant density of about 1.41, as measured by isopyknic equilibrium density gradient ultracentrifugation in cesium chloride solution;

(c) having a molecular weight of about 181,000 daltons as calculated from the Svedberg equation;

(d) precipitating with anti-carcinoembryonic antigen, precipitating preferentially with anti-CEA-S.sub.1, but not precipitating with antisera to blood group A or B or antisera to normal tissue glycoprotein;

(e) having a percent bound antigen vs. antiserum concentration linear regression slope of from about one half to about four fifths the slope of a like curve for carcinoembryonic antigen comparatively tested against identical anti-carcinoembryonic antigen in high ionic strength buffer.

2. The isolated tumor-associated antigen denominated CEA-S.sub.1 suitable for use as an aid in the diagnosis of cancer of the human gastrointestinal tract, consisting essentially of the isomer of carcinoembryonic antigen isolated and concentrated from glycoprotein extract of human adenocarcinomas of the intestinal tract being characterized by such physicochemical and immunochemical characteristics as:

having a sedimentation velocity of S.sub.20,w .degree. = about 6.6 Svedberg units when analyzed by linear sucrose density ultracentrifugation as distinguished from carcinoembryonic antigen having a sedimentation velocity of about 6.8 as determined by the same method;

precipitating with antiserum to carcinoembryonic antigen and antiserum to carcinoembryonic antigen isomer in gel double diffusion to give an arc of apparent identity with carcinoembryonic antigen, being preferentially bound by antiserum carcinoembryonic antigen isomer, but not being bound by antisera to blood group A or B or antisera to normal tissue glycoprotein;

having a single discrete isoelectric point of about 4.5 measured by isoelectric columns with ampholytes having a pH range of 3 to 6 or by thin layer pH 3 to 9 acrylamide gels containing ampholytes as distinguished from carcinoembryonic antigen isomer which exhibits a distribution of isoelectric points over the range of from about 3 to about 5.2;

being indistinguishable from carcinoembryonic antigen when characterized by Sephadex G-200 gel exclusion chromatography;

having a buoyant density when isolated from typical colonic adenocarcinomas of about 1.41 when measured by isopyknic equilibrium density gradient ultracentrifugation in cesium chloride solution; and

an estimated molecular weight of about 181,000 daltons as calculated from the Svedberg equation ##EQU2## wherein s is the sedimentation coefficient measured by linear sucrose density ultracentrifugation, D is the diffusion constant as measured by molecular exclusion chromatography on Sephadex G-200 equilibrated with 0.14 M. NaCl, 0.01 sodium phosphate at pH 6.0, v is the partial specific volume estimated as the reciprocal of buoyant density by isopyknic density gradient ultracentrifugation in CsCl, .rho. is the density of solution for sedimentation coefficient analysis corrected to 1.00, R is the gas constant and T is the absolute temperature, as distinguished from carcinoembryonic antigen having an estimated molecular weight of about 201,000 measured in the same manner.

3. The method of isolating the species of carcinoembryonic antigen denominated CEA-S.sub.1, comprising the steps of:

(a) extracting from adenocarcinomas of the human gastrointestinal tract glycoproteins including carcinoembryonic antigen;

(b) isolating from said glycoproteins the carcinoembryonic antigen; and

(c) isolating from the carcinoembryonic antigen that species which is characterized by having an estimated molecular weight of about 181,000 daltons as calculated from the Svedberg equation ##EQU3## wherein s is the sedimentation coefficient measured by linear sucrose density ultracentrifugation, D is the diffusion constant as measured by molecular exclusion chromatography on Sephadex G-200 equilibrated with 0.14 M. NaCl, 0.01 sodium phosphate at pH 6.0, v is the partial specific volume estimated as the reciprocal of buoyant density is isopyknic density gradient ultracentrifugation in CsCl, .rho. is the density of solution for sedimentation coefficient analysis corrected to 1.00, R is the gas constant and T is the absolute temperature, as distinguished from carcinoembryonic antigen having an estimated molecular weight of about 201,000 measured in the same manner, a single discrete isoelectric point of about 4.5 plus or minus 0.1, a sedimentation velocity of about 6.6 Svedberg units and the immunochemical characteristic of precipitating with anti-carcinoembryonic antigen, precipitating preferentially with anti-carcinoembryonic antigen species, but not precipitating with antisera to blood group A or B or antisera to normal tissue glycoprotein.

4. The method of isolating a species of carcinoembryonic antigen comprising the steps of:

(a) extracting glycoprotein from adenocarcinomas of the human gastrointestinal tract;

(b) isolating extracted glycoproteins having a PI of about 4.5 plus or minus 0.1 from the glycoproteins extracted in step (a);

(c) isolating from the product of step (b) the glycoproteins having an average molecular weight of about 200,000 daltons as measured by molecular exclusion chromatography on Sephadex G-200 equilibrated with 0.14 M. NaCl, 0.01 sodium phosphate, pH 6.0;

(d) fractionating by charge density relative to binding to the ion exchange resin DEAE-cellulose with recovery of the first major peak during elution in 0.05 M Na.sub.2 HPO.sub.4, pH 8.

(e) removing from the product of step (c) other proteins, glycoproteins and blood group antigens related molecules by immunoabsorption; and

(f) isolating from the product of step (d) the glycoprotein having a buoyant density of about 1.41glml.

5. The method of isolating carcinoembryonic antigen species comprising the steps of:

(a) extracting glycoproteins from adenocarcinomas of the human gastrointestinal tract;

(b) isolating by isoelectric focusing of extracted glycoproteins that glycoprotein fraction having an isoelectric point of about 4.5;

(c) isolating by molecular exclusion chromatography that fraction of glycoprotein having an isoelectric point of about 4.5 and an average molecular weight of about 200,000 daltons as measured by molecular exclusion chromatography on Sephadex G-200 equilibrated with 0.14 M. NaCl, 0.01 sodium phosphate at pH 6.0;

(d) fractioning by charge density relative to binding to the ion exchange resin DEAE-cellulose with recovery of the first major peak during elution in 0.05 M Na.sub.2 HPO.sub.4, pH 8;

(e) removing from the fraction of glycoprotein having an isoelectric point of about 4.5 and a molecular weight of about 200,000 daltons by immunoabsorption of proteins, glycoproteins other than carcinoembryonic antigen species, and blood group antigen related molecules, and

(f) isolating by isopyknic density gradient ultracentrifugation that fraction of glycoproteins which are carcinoembryonic antigen species having an estimated molecular weight of about 181,000 daltons as calculated from the Svedberg equation ##EQU4## wherein s is the sedimentation coefficient measured by linear sucrose density ultracentrifugation, D is the diffusion constant as measured by molecular exclusion chromatography on Sephadex G-200 equilibrated with 0.14 M. NaCl, 0.01 sodium phosphate at pH 6.0, v is the partial specific volume estimated as the reciprocal of buoyant density by isopyknic density gradient ultracentrifugation in CsCl, .rho. is the density of solution for sedimentation coefficient analysis corrected to 1.00, R is the gas constant and T is the absolute temperature, as distinguished from carcinoembryonic antigen having an estimated molecular weight of about 201,000 measured in the same manner; and an isoelectric point of about 4.5 and being free of other glycoproteins and proteins and blood group antigen related molecules.

6. The product of the process of claim 5.

7. The isolated species of carcinoembryonic antigen denominated CEA-S.sub.1, said species being characterized by:

having an antigen binding capacity of anti-CEA for CEA-S.sub.1 significantly greater by at least about 10 percent than the antigen binding capacity for CEA for the same anti-CEA in comparative tests; and

having a percent bound antigen vs. antiserum concentration linear regression slope substantially less, from about one half to about four fifths the slope of a like curve for CEA comparatively tested against identical anti-CEA in high ionic strength buffer.

8. The method of isolating a species of carcinoembryonic antigen denominated CEA-S.sub.1, comprising the steps of:

(a) extracting from adenocarcinomas of the human gastrointestinal tract glycoproteins including carcinoembryonic antigen;

(b) isolating from said glycoproteins the carcinoembryonic antigen; and

(c) isolating from the carcinoembryonic antigen that species which is characterized by having an antigen binding capacity of anti-CEA for CEA-S.sub.1 significantly greater by at least about 10 percent that the antigen binding capacity for CEA for the same anti-CEA in comparative tests, and by having a percent bound antigen vs. antiserum concentration linear regression slope substantially less, from about one half to about four fifths the slope of a like curve for CEA comparatively tested against identical anti-CEA in high ionic strength buffer.