The Use of Proteolysis with Ficin, for Immunostaining of Paraffin Sections

Transcription

1 The Use of Proteolysis with Ficin, for Immunostaining of Paraffin Sections A Study of Lymphoid, Mesenchymal, and Epithelial Determinants in Human Tissues RAFAEL E. ANDRADE, M.D., KIMBERLY A. HAGEN, B.Sc, PAUL E. SWANSON, M.D., AND MARK R. WICK, M.D. In order to maximize staining, modifications of immunostaining methods have included proteolytic enzyme digestion of tissue. The authors performed a study of the effect of ficin in 1 paraffinized specimens, including tonsil, lymph nodes, benign vascular and nerve sheath tumors, and various carcinomas and sarcomas. This agent was compared with pepsin and bromelain, as alternative proteases. A panel of monoclonal and polyclonal antibodies was used, with and without previous digestion by ficin, pepsin, and bromelain. A score was assigned to each stain, based on the number and intensity of reactive cells. Ficin enhanced staining markedly in immunostains with antibodies to keratin and Factor VIH-related antigen (F8RAG). Conversely, it abolished staining for LN-2 (a lymphoid marker) and weakened reactivity for S-0 in nerve sheath tumors. Bromelain produced similar results, except that it enhanced S-0. Pepsin was comparatively less active than ficin and bromelain overall but did produce the greatest amplification of vimentin staining in sarcomas. Digestion with any of the three enzymes failed to influence reactivities of leukocyte common antigen, UCHL-1 (a lymphoid marker), alpha-1-antichymotrypsin, carcinoembryonic antigen, epithelial membrane antigen, and blood group isoantigens. These results may reflect a dissimilar recognition of peptide targets in some antigenic proteins, by ficin, bromelain, and pepsin. Hence, one enzymatic agent is unlikely to produce optimal staining for all determinants. With this proviso, however, ficin appeared to be the best general enhancer for antigens known to require vigorous digestion (e.g., keratin; F8RAG) for optimal reactivity in paraffin sections. (Key words: Proteolytic enzymes; Avidin-biotin-peroxidase complex; Intermediate filament proteins; Hematologic antigens) Am J Clin Pathol 1988;90:-9 THE IMMUNOPEROXIDASE TECHNIQUE is a sensitive method for the detection of antigenic substances in either formalin-fixed, paraffin-embedded specimens or frozen tissues." However, a disadvantage of this procedure is its variability, due to the effects of tissue fixa- Received August 1, 1987; received revised manuscript and accepted for publication January, Presented at the 76th Annual Meeting of the International Academy of Pathology, Chicago, March Dr. Wick is the recipient of a Career Development Award in Oncology from the American Cancer Society. Address reprint requests to Dr. Wick: Box 76UMHC, University of Minnesota, 20 Delaware Street, SE, Minneapolis, Minnesota. Division of Surgical Pathology, Department of Laboratory Medicine and Pathology, University of Minnesota School of Medicine, Minneapolis, Minnesota tives; some mordants may cause the destruction of antigenic determinants or masking of antigenic sites.,6 '' 0 ' 12 ' 16 These problems rarely occur in snap-frozen specimens; nevertheless, fresh material is not always available for analysis, especially when dealing with small biopsies and cases submitted for consultation. Several modifications of immunohistochemical technique have been developed, in an attempt to avoid false negative reactions. These include the use of combinations of different immunoperoxidase methods, 919,2-2 diverse chromogenic substrates, 1 and alternative fixatives. 2 Also, protease digestion, as initially introduced in the blood bank to enhance erythrocyte agglutination studies, 1 ' 17 ' 21 has been employed subsequently in the immunohistologic study of paraffin sections. 2 ' ' 6 ' 712 The applicability of proteolytic enzymes to the latter field of investigation is now well recognized; ',20 however, a "universal" protease that would enhance reactivity for all immunostains has remained elusive. Trypsin, pepsin, and pronase are the most common reagents employed toward this end. In general, immunostains for keratins and Factor VIH-related antigen (F8RAG) appear to demonstrate the greatest amplification through the use of such proteases. 2,,16 In an effort to determine the potential enhancing effects of ficin, a plant enzyme (EC..22.),' on a variety of tissue determinants, a systematic study of this reagent was undertaken in our laboratory. Immunostains for 11 antigens were performed in 1 paraffinized specimens, including tonsils, lymph nodes, benign vascular and nerve sheath tumors, and various carcinomas and sarcomas, by use of ficin and other proteases. The results of these analyses constitute the basis of this report. Downloaded from

2 ANDRADE ET AL. A.J.CP.'July 1988 Table 1. Tissues Subjected to Immunostaining with Enzymatic Pretreatment Specimens Tissues Lymphoid tissues Tonsils Lymph nodes Mesenchymal neoplasms Benign vascular tumors Neurofibromas Neurilemmomas Osteosarcomas Malignantfibroushistiocytomas Epithelial neoplasms Breast carcinomas Prostate carcinomas Lung carcinomas Bladder carcinomas Colonic carcinomas Materials and Methods No. All tissue samples used in this study were from human patients and were retrieved from the files of the division of surgical pathology at the University of Minnesota. Table 2. Immunohistochemical Reagents Used in Protease-enhanced Immunostains of Paraffin Sections LN-2* UCHL-l*t Reagent Source Dilution Anti-epithelial membrane antigen* Antivimentin* Anticytokeratins* AE1/AE CAM.2 MAK-6 Anti-leukocyte common antigen* Anti-alpha-1- antichymotrypsin (AACT) Anti-factor VHI-related antigen (F8RAG)* Anti-S-0 protein Anticarcinoembryonic antigen* Anti-blood group isoantigens A,B,H* Techniclone, Inc., Santa Ana, CA Dr. Peter Isaacson, London, England, Santa Barbara, CA ICN Biomedical Co., Lisle, IL Hybritech, Inc., San Diego, CA Becton-Dickinson Immunologicals, Mountain View, CA Triton Biosciences, Alameda, CA Hybritech, Inc. * Murine monoclonal antibodies. t Commercialy available from Prediluted 1:20 1:160 1:800 1:0 1:0 1:7 1:0 1:1,600 1:0 1:1,000 1:1,600 1:80 Table. Enzymes Used in Comparative Proteolysis of Paraffin Sections for Immunohistochemistry Reagent Ficin (EC..22.) Pepsin (porcine) (EC..2.1) Bromelain (EC..22.) * Catalogue no. F12. t Catalogue no. P6887. t Catalogue no. B222. Source Sigma Biochemicals* St. Louis, MOf Sigma Biochemicals Inc.. Sigma Biochemicals^ Concentration (see text for diluents) 0.6% (w/v) 0.% (w/v) 0.02% (w/v) Formalin-fixed, paraffin-embedded specimens were used exclusively. These were all accessioned during the years , during which period fixation times and histologic processing procedures were constant at our hospital. The numbers and types of cases selected are summarized in Table 1. Hematoxylin and eosin slides were prepared in each case to confirm diagnostic classifications and to assess the adequacy of the specimens. Antibodies The reagents employed in this study are summarized in Table 2. These included monoclonal antibodies to leukocyte common antigen, the LN-2 determinant, and the UCHL-1 antigen (used to stain tonsil and lymph nodes), vimentin (malignant fibrous histiocytomas and osteosarcomas), keratin, carcinoembryonic antigen, epithelial membrane antigen, and blood group isoantigens (carcinomas), and F8RAG (hemangiomas). Polyclonal antisera to S-0 protein and alpha-1-antichymotrypsin (AACT) were used in immunostains of peripheral nerve sheath tumors and malignant fibrous histiocytomas, respectively. Proteolytic Enzymes The commercial sources and concentrations of the specified proteases are listed in Table. The characteristics of these agents have been described previously. 1,21 ' 22 Procedure Paraffin sections were cut at - ^m, mounted on glass slides coated with chrome-alum gel, and allowed to stand overnight at 7 C. They were deparaffinized in Americlear and absolute ethanol, and endogenous peroxidase was blocked in absolute methanol-hydrogen peroxidase (0.6% [w/v]) for 0 minutes. Rehydration was accomplished in graded alcohols, distilled water, Downloaded from

3 Vol. 90 No. I FICIN IN IMMUNOHISTOCHEMISTRY and phosphate-buffered saline (PBS) (ph 7.). Sections were wiped free of buffer with a gauze pad, after the last of these steps. In each case, one section was stained without prior protease treatment, and at least one slide each was subjected to digestion with ficin, pepsin, and bromelain, for all antigens being evaluated. Pepsin was prepared fresh at a concentration of 0. mg/dl in 0.01 mol/l NaOH-0.1 mg/dl calcium chloride (ph 7.8); bromelain and ficin were used at concentrations of 0.02% (w/v) and 0.6% (w/v), respectively, in Na 2 HP0-KH 2 P0 buffer (ph 7.). Sections were digested with pepsin and bromelain for 1 minutes at room temperature in a moisture chamber, followed by 1 minutes at 7 C. With ficin, they were digested for 7 minutes at each temperature. After these steps, the sections were rinsed in distilled water (two changes each for minutes) and placed in PBS for an additional minutes. Approximately 2 fil of diluted primary antibody (Table 2) was applied thereafter, and each slide was incubated overnight at C. The sections were developed by use of the avidin-biotin-peroxidase complex (ABC) technique, as previously described," rinsed in phosphate buffer (ph 7.), and immersed in diaminobenzidine solution (0.2 mg/ml) with 0.00% (w/v) hydrogen peroxidase for minutes. Chromogenic development was monitored with light microscopy and quenched with tap water. Staining was enhanced by one dip in osmium-tetroxide solution (0.12% [w/v]), and the sections were then counterstained in Harris hematoxylin. Interpretation of Slides A semiquantitative scale was developed for the scoring of immunostaining intensity. This ranged from 0 to, according to both the number of positive cells and chromogenic density, and is presented in Table. Results The results of staining with each antibody, with and without enzymatic digestion, are summarized in Table. Leukocyte Common Antigen (LCA) Sixteen of the 20 lymphoid tissues ( tonsils and lymph nodes) studied with this antibody without digestion had a staining score greater than or equal to. Only one case had a score as low as 2, and the average score was.9. Similar results were obtained with pepsin digestion (average score.0) and with bromelain (average score.2); the latter enzyme enhanced the reaction in one case from a score of 2 before digestion to a score of Table. Scoring System for Comparative Enzymatic Enhancement of Immunostaining Score Description of Staining Virtually all (>9%) cells stain intensely 7-9% cells stain intensely 0-7% cells stain irregularly but intensely, or All cells stain with moderate intensity 2 2-9% cells stain irregularly but intensely, or All cells stain with weak to moderate intensity 1 1-2% cells stain irregularly but intensely, or All cells stain weakly 0 No staining after digestion. Ficin augmented the reactivity slightly in most cases; 18 had a score of after treatment with this protease, and the average score was.. LN-2 The average score in lymphoid tissues labeled by LN-2 without digestion was.0, and with pepsin digestion it was.. In contrast, very low scores were associated with the use of other enzymes: 1.2 for bromelain, and 0.8 for ficin. Differences between scores obtained in pepsin-treated specimens and undigested tissues were not significant, but pepsin was superior to bromelain and ficin for LN-2 immunostains (Fig. 1). UCHL-1 Scores for UCHL-1 stains were similar for those tissues not subjected to digestion (ND) and those treated with pepsin (average,.9 for both). Slightly increased reactivity was observed with bromelain, but this en- Antigen* LCA LN-2 UCHL-1 S-0 Vimentin F8RAG AACT EMA CEA CK BGI Table. Summary of Effects of Proteolysis on Lymphoid, Mesenchymal, and Epithelial Determinants in Paraffin Sections No Digestion t Ficin Proteases Pepsin Bromelain Abbreviations:LCA = leukocyte common antigen; S-0 = S-0 protein; F8RAG = Factor VIH-related antigen; AACT alpha-l-antichymotrypsin;ema = epithelial membrane antigen;cea = carcinoembryonic antigen;ck = cytokeratin; BGI - blood group isoantigens. t Scores defined in Table. Downloaded from

4 6 ANDRADE ET AL. AJ.C.P.-July 1988 hancement was not significant (average score :1). Ficin-augmented reactivity averaged.2; this difference was similarly only of borderline significance in comparison with undigested tissues (ND vs. ficin), but ficin did increase reactivity in cases with very low ND scores. In four examples, the scores increased from 2 to or. S-0 Ten benign peripheral neural tumors were evaluated for S-0 content, including five neurilemmomas (NLs) and five neurofibromas (NFs). Bromelain was superior overall in enhancing detection of this determinant; average scores were.6 for NFs and. for NLs (Fig. 2). Undigested tissues had scores of for NMs and. for NFs; they were 2.6 for NMs and 2.8 for NFs in cases treated with pepsin. Notably, staining for S-0 was completely abolished by ficin pretreatment in all cases. Alpha-1-Antichymotrypsin (AACT) Differences in reactivity produced by all proteases were not significant in ten cases of MFH stained for AACT. Low scores were seen in all instances (ND: 1.; pepsin, 1.2; bromelain, 1.7; and ficin, 2.). Epithelial Membrane Antigen (EMA) Fifty cases were stained for EMA, including lung carcinomas, prostate carcinomas, ductal breast carcinomas, colonic adenocarcinomas, and transitional cell carcinomas. Breast and colonic cancers generally displayed better reactivity without any enzymatic digestion. Pepsin slightly enhanced EMA staining in lung carcinomas, but without a significant difference relative to ND sections; prostate tumors and transitional cell carcinomas were only moderately reactive under both sets of conditions. Vimentin Twenty cases were stained for vimentin, including osteosarcomas (OSs) and malignant fibrous histiocytomas (MFHs). Undigested tissues exhibited nearly optimal reactivity (score.1 for OSs and.9 for MFHs), although it was somewhat enhanced by pepsin (score.9 for OSs and.77 for MFHs). Bromelain produced no significant changes. A decrease in staining intensity was produced by digestion with ficin (average scores,. for OSs and.1 for MFHs). Scores obtained in pepsintreated sections were significantly different from those seen with ficin (Fig. ). Factor VIII-Related Antigen Significant augmentations of staining for F8RAG were seen when tissues were digested with ficin and bromelain. In ten benign vascular tumors, the average scores were.9 with ficin and. with bromelain; in contrast, lower values were recorded in undigested tissues, and those treated with pepsin (. and.9, respectively). Carcinoembryonic Antigen (CEA) Positivity for CEA was similar in all carcinomas tested. Prostatic tumors had negative results with and without enzymatic digestion. An increase in CEA reactivity was found in lung carcinomas digested with ficin, but this effect was not marked. Cytokeratin (CK) There was a substantial augmentation in CK staining in all carcinomas digested with proteases. This effect was most striking in sections treated with ficin and bromelain; results were statistically significant and constant in all 0 cases in this category. In all cases, scores with ficin were greater than. and those with bromelain were over.1; undigested tissues manifested relatively low scores, averaging. (Fig. ). Blood Group Isoantigens (BGIs) In general, low immunohistochemical scores were obtained in immunostains for BGIs, with only modest FIG. 1 (upper, left). LN-2, immunostain of human tonsillar tissue, after pepsin predigestion. Virtually all follicular lymphocytes are strongly stained for this determinant. Anti-LN-2/pepsin predigestion; avidin-biotin-peroxidase complex immunostain with hematoxylin counterstain (X160). FlG. 2 (upper, right). S-0 protein immunostain of human neurofibroma, after bromelain digestion. Tumoral Schwann's cells are intensely labeled, with a nucleocytoplasmic pattern of staining. Anti-S-0/bromelain predigestion; avidin-biotin-peroxidase complex immunostain with hematoxylin counterstain (X00). FIG. (lower, left). Vimentin immunostain of human osteosarcoma, after pepsin predigestion. Neoplastic osteoblasts are diffusely and intensely reactive for this intermediatefilamentprotein. Antivimentin/pepsin predigestion; avidin-biotin-peroxidase complex immunostain with hematoxylin counterstain (X00). FlG. (lower, right). Cytokeratin immunostain of poorly differentiated human prostatic adenocarcinoma. Essentially all tumor cells are intensely reactive, after predigestion withficin.anticytokeratin (AEl/AE)/ficin predigestion; avidin-biotin-peroxidase complex immunostain with hematoxylin counterstain (XS0). Downloaded from

5 Vol. 90 No. I \ FICIN IN IMMUNOHISTOCHEMISTRY * ir jf: i Downloaded from 7

6 8 ANDRADE ET AL. AJ.CP.'July 1988 staining augmentation in colonic carcinomas that were predigested with ficin (average scores, 1.9 ND; 2.1 pepsin; 1.9 bromelain; and 2.8 ficin). Discussion Several studies have previously reported the usefulness of protease digestion in optimizing immunoreactivity of intermediate filament proteins and other cellular antigens in erythrocytes and formalin-fixed and paraffin-embedded tissues. 1 " 6 ' 121 " 18 ' 2 To our knowledge, however, this is only the second assessment of the effects of ficin in this context. The latter enzyme is a cysteine proteinase derived from figs (Ficus species) and displays a broad substrate specificity. Taschini and MacDonald have previously described the utility of ficin in unmasking keratin proteins in routine specimens that have been fixed in formalin for more than four to six hours. 2 Moreover, these authors suggested that it was comparable in effectiveness to trypsin, a widely implemented protease in diagnostic immunohistochemistry. Perhaps because of the chrome-alum mounting medium used in our laboratory, we have had consistent difficulty in the past in maintaining adherence of tissue sections to glass slides in trypsin-enhanced immunostaining procedures. For this reason we were motivated to examine the attributes of alternative proteolytic agents, particularly with respect to ficin. Our results indicate that the latter reagent is indeed highly efficacious in optimizing reactivity for keratin in paraffin sections. Moreover, stains for Factor VHI-related antigen also were greatly improved by prior ficin digestion. Immunohistochemical reactions for these two determinants that were performed after treatment with pepsin and bromelain demonstrated a general inferiority to others done with ficin. Nevertheless, comment should be made on some exceptions to the statements just made. Probably because of enzymatic cleavage of immunoreactive epitopes, ficin diminished the reactivity of LN-2 (a B-cell and histiocyte marker), 8 vimentin, and epithelial membrane antigen, and abolished staining for S-0 protein. If these antigens are of diagnostic interest in any given case, it should be obvious that an alternative protease must be employed to optimize their exposure to primary antibodies. For example, bromelain appeared to have the most beneficial effect on S-0 immunostains, and vimentin reactivity was best potentiated by pepsin in this study. Therefore, one cannot expect uniform results from the use of a single proteolytic reagent, and enzyme enhancement methods must be tailored to fit the immunostain in question. This process can only be mastered through systematic intrainstitutional evaluation, because of the biochemical variability of commercial enzymes and antibodies that are currently in use. Several other antigens in this series were affected minimally by proteolysis with any of the three enzymes we used. These included UCHL-1 (a pan-t-cell marker), 1 LCA, AACT, CEA, and BGIs. Hence, it is unlikely that a lack of specimen immunoreactivity for any of these determinants can be attributed to aldehyde-induced epitope masking, under routine fixation conditions. In summary, it would seem that ficin is the favored reagent for exposure of antigens such as keratin and F8RAG. These determinants are among those most easily masked by routine tissue processing, and their reactivity is greatly enhanced after ficin pretreatment. Based on comparisons with previous data, 2 ' ' 6 ',12 ' 16 ficin appears to be comparable to trypsin in overall efficacy. However, the latter enzyme must be prepared fresh daily and differs substantially in activity from lot to lot. On the other hand, ficin is stable in diluted form as long as 72 hours after preparation and shows minimal commercial variability (authors' unpublished results). Finally, when it is used as outlined in this report, this protease causes little adverse effect on tissue adhesion to glass slides, making it readily utilizable by most surgical pathology laboratories. References 1. Babin R, Cassaigne R, Christie R, et al: Pharmaceutical enzymes: properties and assay methods. In: Ruyssen R, Lauwers A, eds. Pharmaceutical enzymes. Gent, Belgium: Scientific Publishing, 1978:-0, Battifora H, Kopinski M: The influence of protease digestion and duration offixationon the immunostaining of keratins. A comparison of formalin and ethanolfixation.j Histochem Cytcchem 1986;: Brantzaeg P, Rognom TO: Evaluation of tissue preparation methods and paired immunofluorescence staining for immunocytochemistry of lymphomas. Histochem J 198;1: Brozman M: Immunohistochemical analysis of formaldehydefixed and trypsin or pepsin treated material. Acta Histochem 1978;6S: Corson J, Weiss L, Banks-Schlegel S, Pinkus G: Keratin proteins and carcinoembryonic antigen in synovial sarcomas: an immunohistochemical study of 2 cases. Hum Pathol 198; 1: Curran RC, Gregory J: The unmasking of antigens in paraffin sections of tissue by trypsin. Experientia 1977;: Denk H, Radaszkiewicz T, Weirich E: Pronase pretreatment of tissue sections enhances sensitivity of the unlabelled antibodyenzyme (PAP) technique. J Immunol Methods 1972; 1: Epstein A, Marder R, Winter J, Fox R: Two monoclonal antibodies (LN-1, LN-2) reactive in B and formalin-fixed, paraffiriembedded tissues with follicular center and mantle zone human B lymphocytes and derived tumors. J Immunol 198;1: Hacker GW, Springall DR, Van Noorden S, et al: The irrimunogold-silver staining method: a powerful tool in histopathology. Virchows Arch [Pathol Anat] 198;06: Hautzer N, Wittkuhn J, McCaughey WTE: Trypsin digestion in immunoperoxidase staining. J Histochem Cytochem 1980;28: Hsu S-M, Raine L, Fanger H: The use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a compari- Downloaded from

7 Vol. 90 No. I FICIN IN IMMUNOHISTOCHEMISTRY 9 son between ABC and unlabelled antibody (PAP) procedures. J Histochem Cytochem 1981;29: Huang S, Minnasian H, More J: Application of immunofluorescent staining on paraffin sections improved by trypsin digestion. Lab Invest 1976;: Neumann MP, Solas I, Parkin J, et al: Monoclonal antibody study of Philadelphia chromosome positive blastic leukemias using the alkaline phosphatase anti-alkaline phosphatase (APAAP) technic. Am J Clin Pathol 1986;8: Norton AJ, Ramsay A, Smith S, Beverley P, Issacson P: Monoclonal antibody (UCHL-1) that recognizes normal and neoplastic T cells in routinely fixed tissues. J Clin Pathol 1986;9: Ota S, Muta E, Katakhira Y, Okamoto Y: Reinvestigation of fractionation and some properties of the proteolytically active components of stem and fruit bromelains. J Biochem 198;98: Pinkus G, O'Connor E, Etheridge C, Corson J: Optimal immunoreactivity of keratin proteins in formalin-fixed, paraffin-embedded tissue requires preliminary trypsinization. J Histochem Cytochem 198;: Pirofsky B, Mangum M: Use of bromelain to demonstrate erythrocyte antibodies. Proc Soc Exp Biol Med 199;1: Poston RN, Sidhu YS, Makin CS, Mason DY: Specific enzyme treatment is required for individual monoclonal antibodies in immunohistochemistry with formalinfixedsections. Prot Biol Fluids 198;2: Said JW, Vimadalal S, Nash G, et al: Immunoreactive neuron specific enolase, bombesin, and chromogranin as markers for neuroendocrine lung tumors. Hum Pathol 198;16: Shishi J, Ghazizadeh M, Oguro T, Aihara K, Araki T: Immunohistochemical localization of CA 12 antigen in formalin-fixed paraffin sections of ovarian tumors with the use of pronase. Am J Clin Pathol 1986;8: Simmons A: Methods of detecting antigens and antibodies. In: Simmons A, ed. Technical hematology. rd ed. Philadelphia: JB Lippincott, 1980: Stryer L: Zymogen activation: digestive enzymes and clotting factors. In: Stryer L. ed. Biochemistry. 2nd ed. San Francisco: WH Freeman, 1981: Swanson PE, Hagen KA, Wick MR: Avidin-biotin-peroxidaseantiperoxidase (ABPAP) complex: an immunocytochemical method with enhanced sensitivity. Am J Clin Pathol 1987;88: Taschini PA, MacDonald DM: Protease digestion step in immunohistochemical procedures:ficinas a substitute for trypsin. Laboratory Medicine 1987;18: Walts AE, Said JW, Shintaku IP, Lloyd RV: Chromogranin as a marker of neuroendocrine cells in cytologic material an immunocytochemical study. Am J Clin Pathol 198;8: Downloaded from