Autoantibodies directed against the islet cells of. Rapid Publications Assay for Islet Cell Antibodies Protein A-Monoclonal Antibody Method

Transcription

1 Rapid Publications Assay for Islet Cell Antibodies Protein A-Monoclonal Antibody Method S. SRIKANTA, A. RABIZADEH, M. A. K. OMAR, AND G. S. EISENBARTH SUMMARY Assays for islet cell antibodies (ICA) are finding increasing application in clinical diabetology. We have developed a new islet cell antibody assay system (ICA-pA), whose salient features include: (1) utilization of fluorescein-conjugated staphylococcal protein A as a standard second-step reagent, the advantages of this approach being improved "signal" (islet)/"noise" (acini) ratio due to reduction of interfering background acinar pancreatic staining, and facilitation of assay standardization provided by the use of a chemically pure conjugated protein A reagent; (2) monoclonal antibody counterstaining with rhodamine-conjugated BISL-32 for the rapid identification of islets in pancreatic sections; and (3) quantitation of circulating serum ICA by microimmunofluorometric techniques. DIABETES 1985; 34: Autoantibodies directed against the islet cells of the pancreas are an important hallmark of type I (insulin-dependent) diabetes mellitus. Besides, significance in research studies probing the mechanism(s) of autoimmune beta cell destruction, islet cell antibody assays are currently finding increasing application in clinical diabetology. In particular, recent reports indicate that islet cell antibodies are an important predictive marker of ongoing preclinical beta cell destruction. 14 In combination with serial assessments of sensitive indices of beta cell mass/function, like the early-phase insulin release to i.v. glucose, islet cell antibody assays are crucial for monitoring the natural history and chronology of the beta cell destructive From the Joslin Diabetes Center, Brigham and Women's Hospital, and Harvard Medical School, Boston, Massachusetts. Address reprint requests to G. S. Eisenbarth, M.D., Ph.D., or S. Srikanta, M.D., Joslin Diabetes Center, Research Division, One Joslin Place, Boston, Massachusetts Received for publication 24 December process. In addition, in subjects with a history of gestational diabetes mellitus 356 and in patients with non-insulin-dependent diabetes mellitus, 78 presence of islet cell antibodies portend the subsequent development of insulin dependence, thus providing valuable information for clinical management. Recent observations suggest that islet cell antibody assays may also prove to be useful in the field of pancreas/islet transplantation, particularly in the differential diagnosis of recurrent autoimmune insulitis 9 versus rejection and other causes of graft dysfunction and failure. Despite its original discovery, in 1974, 1011 and numerous attempts at improvement and modification, 1217 conventional, indirect immunofluorescence and other immunocytochemical assays for islet cell antibodies are beset with numerous methodologic problems, and these have precluded satisfactory standardization of this useful assay procedure. At the current state of knowledge regarding the biochemistry and molecular characteristics of the relevant islet cell "target" autoantigens, the indirect immunofluorescence assay continues to be the best available and clinically relevant measurement for circulating islet cell antibodies. The inevitable need for a fresh frozen, unfixed human pancreatic substrate (from surgical specimens or cadaveric transplant donors) has been amply emphasized in the literature. However, other impeding factors include the high background acinar pancreatic staining and the gross batch-to-batch variability in the quality of the second-step-conjugated antibody (antihuman immunoglobulin) reagent used in the assay. We have observed that use of an optimal dilution of fluoresceinconjugated protein A as a "nonantibody," chemical probe to detect human immunoglobulin binding to the pancreatic islet eliminates many of the problems encountered in the assay and facilitates standardization. The advantages provided by this technique have led us to completely switch from the conventional second-step antibody method to the ICA by the protein A system (ICA-pA) as described in this article. In addition, microphotometric techniques have been used to quantitate circulating serum islet cell antibodies using this assay system. 300 DIABETES, VOL. 34, MARCH 1985

2 S. SRIKANTA. A. RABIZADEH. M. A. K. OMAR, AND G. S. EISENBARTH MATERIALS AND METHODS HUMAN PANCREATA These were obtained from cadaveric kidney transplant donors and surgical pancreatectomy specimens. Half-centimeter blocks were snap frozen in precooled isopentane on dry ice and stored at -70 C. Thin cryostat sections (4-5 fjim) were cut onto gelatin dichromate-coated slides before use. (Such slides have been found to withstand storage at -20 C for several months.) HUMAN SERA Sera from type I diabetic subjects, first-degree relatives of type I diabetic subjects, and controls were stored at -20 C before use. SECOND-STEP REAGENTS FITC-conjugated antihuman immunoglobulin. The following immunoglobulins were used: (1) FITC-conjugated IgG fraction goat antihuman immunoglobulin IgG, IgA, IgM, (#16789 Cappel Laboratories, Cochranville, Pennsylvania); (2) FITC-conjugated goat antihuman IgG, heavy and light chain, (65-207S694, Miles Scientific, Naperville, Illinois); (3) FITC-conjugated goat antihuman immunoglobulin, polyvalent, affinity isolated, (# , Tago, Inc., Burlingame, California); and (4) FITC-conjugated goat antihuman IgM, mu-chain specific, affinity isolated (#2192, Tago, Inc.). FITC-conjugated protein A. Protein A used in these experiments was prepared from Staphylococcus aureus (Cowan strain) using ion exchange and gel filtration chromatography, 18 and then labeled with fluorescein isothiocyanate. 19 The final preparation had an immunoglobulin binding capacity of 9.7 mg human IgG/mg protein and contained 70.5 jig FITC/mg protein (Sigma, St. Louis, Missouri, #P-5145). INDIRECT IMMUNOFLUORESCENCE Human pancreatic sections were incubated for 30 min at room temperature with undiluted human sera. The sections were washed twice with phosphate-buffered saline containing 0.05% sodium azide (PBS), and incubated for 30 min at room temperature with suitable dilutions of the second-step reagents (FITC-conjugated antihuman immunoglobulins 1:20 or FITC-conjugated protein A 1:1000, both diluted in PBS with 1% bovine serum albumin). The sections were washed twice in PBS, mounted in PBS with 30% glycerol, and examined under a Leitz Dialux or Diavert fluorescence microscope (Rockleigh, New Jersey). The relative intensity of islet immunofluorescence was graded on an arbitrary visual scale of 0-4, based on the subjective assessment of differential staining between the islet and the background acinar tissue. MICROIMMUNOFLUOROMETRY Quantitation of specific islet cell immunofluorescence was performed using a Leitz MPV compact photometer. The diameter of the photometer-measuring diaphragm was preset (to encircle the cross section of few islet cells in the field of interest to be measured) and held constant for all measurements (Figure 1F). During the actual quantitation of islet immunofluorescence, the pancreatic sections were systematically scanned from one edge to the other (Figure 2A). Coupled measurements of fluorescent intensity were obtained on the islet and on the adjacent acinar pancreas, and the difference between the two readings provided a measurement of specific islet cell immunofluorescence. A minimum of 10 sets of readings was obtained for every serum sample tested, and the results were automatically recorded and statistically analyzed by a Hewlett-Packard (Andover, Massachusetts) 97S I/O calculator connected to the photometer. MONOCLONAL ANTIBODY COUNTERSTAINING Monoclonal islet cell antibody BISL-32 was purified and directly conjugated with the alternative fluorochrome tetramethyl rhodamine B isothiocyanate (TRITC). During the final step in the immunofluorescence procedure, a 1:1000 dilution of the reagent (immunoglobulin concentration 10 mg/ ml) was added onto the pancreatic sections. RESULTS In marked contrast to the polyclonal antihuman immunoglobulin reagents derived from various immunized animal sera, the use of an FITC-conjugated protein A as a standard, "nonantibody," second-step reagent eliminated much of the background exocrine pancreatic staining (Figure 1, A-E). This rendered the visual assessment of positive or negative islet staining easier and more convenient. By the FITC protein A (ICA-pA) technique the islets stood out clearly, brightly stained against a dark background (Figure 1, E and F), whereas, with the use of the anti-immunoglobulin antibody reagents, varying degrees of background staining of the acinar pancreatic cells (Figure 1, A-D) interfered with the visual assessment of the islet cell antibody positivity or negativity, as read by two independent observers. Table 1 illustrates the results of the ICA measurements by our new assay system (ICA-pA) on serum samples previously characterized in our laboratory by the conventional immunofluorescence assays using polyclonal anti-immunoglobulin antibody conjugates. In particular, there were no false negatives in the ICA-pA assay as compared with the results from the conventional assay. At the same time, the assessment of borderline and weak positive readings was made much easier by the new technique. TABLE 1 Results of islet cell antibodies by the protein A system (pilot study)* Study groups ICA-pA positive ICA-pA negative Total Clinical type I diabetes Preclinical type I diabetes Firstdegree relatives Discordant monozygotic twins Unrelated controls *AII these serum samples have been analyzed by both the conventional ICA assay (using polyclonal anti-immunoglobulin conjugates) and the new ICA-pA technique. No false negatives have been observed with the use of ICA-pA system. Following this pilot study, we have completely switched to this new technique for all ICA screening assays DIABETES, VOL. 34, MARCH

3 PROTEIN A-MONOCLONAL ANTIBODY METHOD FIGURE 1. Indirect immunofluorescence assay for islet cell antibodies. (A-E) Comparison of methods using different fluorescein-conjugated, second-step reagents. (A) Goat antihuman immunoglobulin: IgG, IgA, and IgM. (B) Goat antihuman IgG. (C) Goat antihuman immunoglobulin (polyvalent). (D) Goat antihuman IgM (mu-chain specific). (E) Chromatographically purified protein A. These experiments, including photomicrography, were performed under identical conditions using the same Joslin reference ICA positive serum (i.e., #12184). (F) Quantitative microimmunofluorometry. Staining of a human pancreatic section by an ICA positive serum (ICA-pA method). The circles denote the projection of the photometer-measuring diaphragm on the microscopic image of the islet and acinar tissue, respectively. Fluorescence intensity emanating from these two representative areas is quantitated separately by the photometer, the data analyzed, and the difference (islet - acini) automatically expressed as a measure of specific islet cell immunofluorescence. 302 DIABETES, VOL. 34, MARCH 1985

4 5RIKANTA. A. RABIZADEH. M A. K OMAR. AND G. S. EISENBARTH FIGURE 1. (G-J) Utility of monoclonal antibody counterstaining. (G, I) Islet cell antibody negative serum. The islet was first identified under rhodamine filter (G) through BISL-32 monoclonal antibody counterstaining. The same islet immediately viewed under the fluorescein filter (I) illustrates the absence of islet cell staining. (H, J) Islet cell antibody positive serum. Showing islet identification by monoclonal antibody BISL-32 under rhodamine filter (H), followed by visualization of serum islet cell antibody staining under fluorescein filter (J). Figure 2 illustrates the data on the microimmunofluorometric quantitation of islet cell antibodies. Use of an optimal dilution of the fluorescein-conjugated protein A stock solution (Figure 2B) is crucial to obtain the highest signal (islet)/noise (acinar background) ratio and, thus, the maximum sensitivity. Standard curves were generated using dilutions of reference islet cell antibody positive sera (Figure 2C). Repeated exposure of the same field must be avoided during the photometric quantitation to eliminate the effects of rapid fluorescent photobleaching (Figure 2E). Selective islet cell staining by monoclonal antibody BISL32 enabled the rapid identification of the islets using the rhodamine filter (Figure 1, G and H). Switching to the fluorescence filter permitted an immediate assessment and scoring of the islet cell immunofluorescent staining (Figure 1, land J). DISCUSSION Protein A is derived from the cell wall of a bacterium, Staphyloccus aureus, and has the property of attaching to the Fc portion of immunoglobulin molecules of several species.20 In man, protein A reactivity is confined to the lgg1, lgg2, and lgg4 subclasses. Human IgA and IgM proteins are generally negative, though proteins of the human lga2 subclass DIABETES, VOL. 34, MARCH 1985 react with staphylococci and some IgM proteins (provisionally designated lgm2) are also reactive. IgD and IgE immunoglobulin classes are also unreactive. This selective immunoglobulin binding property of protein A has found increasing application in numerous immunochemical techniques, including its recent use in immunocytochemistry.21 The use of protein A conjugated to the fluorochrome FITC as a probe to identify the binding of human anti-islet cell antibodies specifically to the endocrine pancreatic islet cells has two important advantages: (1) In comparison with the standard techniques that employ polyclonal anti-immunoglobulin sera derived from various animals, the ICA-pA technique offers a reduction in the interfering background acinar pancreatic staining, thereby enhancing the specificity of the islet cell immunofluorescence (i.e., increased signal/noise ratio). The most likely mechanism for the improvement observed in the ICA-pA assay system seems to be the class/ subclass distribution of the anti-islet antibody reactivity in human sera from type I diabetic subjects. Apparently, most, if not all, of the anti-islet cell specific reactivity resides in the protein A binding immunoglobulin class/subclasses, whereas the interfering background acinar cell reactivity resides in the non-protein A binding fraction (i.e., IgM). (Preliminary studies using immunoglobulin subclass-specific 303

5 PROTEIN A-MONOCLONAL ANTIBODY METHOD -* - ISLET (TOTAL) ACINI (BACK6R0UN0 1/25 1/125 1/685 1/3125 1/ /78125 I/39O625 1/ FITC-PROTEIN A (DILUTION) ICA ICA - 1/4 1/8 1/16 1/32 1/64 SERUM DILUTION ICA-pA VISUAL SCORING \ «20 \ NUMBER OF RE-EXPOSURES OF THE SAME FIELD (ISLET AND ACINI ) FIGURE 2. Quantitative microimmunofluorometric assay for islet cell antibodies (fluorescence expressed as arbitrary units): (A) Scanning protocol for quantitation of islet cell antibodies: immunostained pancreatic sections were systematically scanned from one edge to the other, making coupled measurements of fluorescent intensity on the islet and on the adjacent acinar pancreas. The difference between the two readings (islet - acini) provides a measure of specific islet cell fluorescence. (B) Effect of varying dilutions of the fluorescein-conjugated protein A (stock) on immunofluorescent staining of islets ( ), acinar background ( ), and the difference between the two, i.e., specific islet cell immunofluorescence (A A). The optimal working dilution of the protein A-second-step reagent was 1/625 to 1/3125. (C) Standard curves for specific islet cell fluorescence, generated using dilutions of the Joslin reference ICA positive serum, i.e., # (D) Correlation between the visual assessment and the microimmunofluorometric system for the quantitation of circulating islet cell antibodies (number of sera tested = 11). (E) Fluorescence photobleaching: effect of repeated exposure of the same field (islet and acini) during the quantitation procedure. monoclonal antibodies tend to confirm these observations and indicate a paucity of IgM-ICA response even during the preclinical phase of type I diabetes [M. A. K. Omar, S. Srikanta, and G. S. Eisenbarth, submitted for publication].) Elimination of the anti-acinar cell reactivity of the irrelevant antibody components present in the immunized animal sera (anti-immunoglobulin conjugates) is an additional advantage of the ICA-pA system. (2) The second important advantage of the ICA-pA technique is the chemically pure nature of the FITC protein A. This reagent can be easily prepared in large quantities with a uniform immunoglobulin binding capacity and FITC content. This should allow the use of an identical second-step reagent (free of the vagaries of animal sera such as nonspecific reactions, instability, loss of immunoreactivity, etc.) in all ICA assays, thus facilitating standardization of the immunofluorescence assay not only within a laboratory but between laboratories. With the exception of an identical human pancreatic substrate, every other component in our assay system (protein A fluorescein conjugate, rhodamine-conjugated monoclonal antibody BISL-32) is amenable to uniform quality and infinite availability. The main advantage provided by the immediate pre-identification of islets with monoclonal antibody counterstaining seems to be the time saved during the actual visual assessment and scoring of ICA positivity or negativity. By simply alternating between the two filter systems rhodamine and fluorescein one can rapidly scan a sample of welldefined islets and quickly score their ICA reaction. Precise quantitation of circulating levels of islet cell antibodies can be useful for monitoring the course of islet beta 304 DIABETES, VOL. 34, MARCH 1985

6 cell destruction, 9 and studying the influence of potential immunotherapeutic interventions. Pending the precise biochemical characterization of the target islet cell autoantigens and development of sensitive radioassays for ICA, the quantitative microimmunofluorometric system described in this report represents an approach in this direction. ACKNOWLEDGMENTS The secretarial help of P. Cronin and data management by J. Connelly are gratefully acknowledged. We thank Dr. J. P. Caulfield of Harvard Medical School for his help and expert guidance in the performance of microphotometric assays. This work was supported by grants from the National Institutes of Health (AM ) and grants from the Juvenile Diabetes Foundation (045C81, 83F200). REFERENCES 1 Srikanta, S., Ganda, 0. P., Eisenbarth, G. S., and Soeldner, J. S.: Islet cell antibodies and beta cell function in monozygotic triplets and twins initially discordant for Type I diabetes mellitus. N. Engl. J. Med. 1983; 308: Srikanta, S., Ganda, O. P., Jackson, R. A., Gleason, R. E., Kaldany, A., Garovoy, M. R., Milford, E. L, Carpenter, C. B., Soeldner, J. S., and Eisenbarth, G. S.: Type I diabetes mellitus: chronic progressive beta cell dysfunction. Ann. Intern. Med. 1983; 99: Srikanta, S., and Eisenbarth, G. S.: Disappearing anti-islet antibodies? Lancet 1984; 1: Srikanta, S., Ganda, 0. P., Jackson, R. A., Brink, S. J., Fleischnick, E., Yunis, E., Alper, C, Soeldner, J. S., and Eisenbarth, G. S.: Pre-Type I (insulin-dependent) diabetes: common endocrinological course despite immunological and immunogenetic heterogeneity. Diabetologia 1984; 27 (Suppl): Steel, J. M., Irvine, W. J., and Clarke, B. R: The significance of pancreatic islet cell antibody and abnormal glucose tolerance during pregnancy. J. Clin. Lab. Immunol. 1980; 4: Ginsberg-Fellner, F, Mark, E. M., Nechemias, C, Hausknect, R. U., Rubenstein, P., Dobersen, M. J., and Notkins, A. L: Islet cell antibodies in gestational diabetics. Lancet 1980; 2: Irvine, W. J., McCallum, C. J., Gray, R. A., and Duncan, L. J. P.: Clinical and pathogenic significance of pancreatic islet cell antibodies in diabetics treated with oral hypoglycemic agents. Lancet 1977; 1: Irvine, W. J., Sowers, J. S. A., Feek, C. M., Prescott, R. J., and Duncan, L. J. P.: The value of islet cell antibody in predicting secondary failure to oral hypoglycemic agent therapy in diabetes mellitus. J. Clin. Lab. Immunol. 1979; 2: Sutherland, D. E. R., Sibley, R., Xu, X. Z., Michael, A., Srikanta, S., Taub, F, Najarian, J., and Goetz, F. C: Twin to twin pancreas transplantation: reversal and reenactment of the pathogenesis of Type I diabetes. In press. Trans. Assoc. Am. Physicians Bottazzo, G. F, Florin-Christensen, A., and Doniach, D.: Islet cell antibodies in diabetes mellitus with autoimmune polyendocrine deficiencies. Lancet 1974; 2: MacCuish, A. C, Barnes, E. Z., Irvine, W. J., and Duncan, L. J. P.: Antibodies to pancreatic islet cells in insulin-dependent diabetics with coexistent autoimmune disease. Lancet 1974; 2: Dobersen, M. J., Bell, A. M., Jenson, A. B., Notkins, A. L., and Fellner, F. G. Detection of antibodies to islet cells and insulin with paraffin embedded pancreas as antigen. Lancet 1979; 2: Yagihashi, S., Suzuki, H., Dobersen, M. J., Onodera, T., Notkins, A. L, and Fellner, F G.: Autoantibodies to islet cells, comparison of methods. Lancet 1982; 2: Sewell, H., Smith, D. I., Wilcox, A., Barnes, C. A., and Matthews, J. B.: Demonstration of islet cell antibodies. Lancet 1980; Krell, J., and Rabin, B. S.: Comparison of an immunohistochemical and immunofluorescence procedure to detect antibody to pancreatic islets. Diabetes 1984; 33: Marner, B., Lernmark, A., Nerup, J., Molenaar, J. L., Tuk, C. W., and Bruining, G. J.: Analysis of islet cell antibodies on frozen sections of human pancreas. Diabetologia 1983; 25: Veda, Y., Miyake, S., Takahashi, A., Tsujihata, M., and Nagatoki, S.: Incidence of islet cell antibody in insulin-dependent diabetes mellitus by a new method using peroxidase labeled protein A. Abstract. Diabetes 1983; 32 (Suppl. 1):47A. 18 Sjoquist, J., Meloun, B., and Hjelm, H: Protein A isolated from staphyloccus aureus after digestion with lysostaphin. Eur. J. Biochem. 1972; 29: McKinney, R. M., Spillane, J. T., and Pearch, G. W.: A simple method for determining the labeling efficiency of fluorescein isothiocyanate products. Anal. Biochem. 1966; 14: ^Stanworth, D. R., and Turner, M. W.: Immunochemical analysis of immunoglobulins and their subunits. In Handbook of Experimental Immunology, Vol. 1. Weir, D. M., Ed. Oxford, Blackwell Scientific Publications, 1978: Van Noorden, S., and Polak, J. M.: Immunocytochemistry today: techniques and practice. In Immunocytochemistry: Practical Applications in Pathology and Biology. Polak, J. M., and Van Noorden, S., Eds. Bristol, England, Wright PSG, 1983: DIABETES, VOL. 34, MARCH