Veterinary Immunology and Immunopathology, 26 (1990) 105-113
105
Elsevier Science Publishers B.V., Amsterdam
Immunoperoxidase staining for the detection of autoantibodies in canine autoimmune skin disease; comparison to immunofluorescence results Gregory A. Bradley* and M.B. Calderwood Mays** Department of Comparative and Experimental Pathology, Collegeof VeterinaryMedicine, University of Florida, Gainesville,FL 32610, U.S.A. (Accepted 15 February 1990)
ABSTRACT Bradley, G.A. and Calderwood Mays, M.B., 1990. Immunoperoxidase staining for the detection of autoimmune skin disease; comparison to immunofluorescence results. Vet. Immunol. Immunopathol., 26:105-113. Skin sections from 22 dogs with autoimmune skin disease were stained with anti-canine IgG, IgM and IgA using an immunobridge immunoperoxidase method. Eight cases of lupus erythematosus, three cases of pemphigns vulgaris, and 11 cases of pemphigus foliaceus were included. Results of previously performed, direct immunofluorescence tests for the detection of canine immunoglobulin on skin were available on 17/22 cases. The immunoperoxidase method yielded an overall positive result in 59% (5/8 lupus erythematosus, 2/3 pemphigus vulgaris and 6/11 pemphigns foliaceus) versus an overall positive result of 47% for direct immunofluorescence (3/5 lupus erythematosus, 2/ 2 pemphigus vulgaris and 2 / 10 pemphigns foliaceus). The immunobridge immunoperoxidase method compared favorably to direct immunofluorescenee testing of canine skin for autoantibody in cases of lupus erythematosis and pemphigus vulgaris, and was superior in cases of pemphigns foliaceus. This method should prove useful as an aid in the diagnosis of canine autoimmune skin disease.
INTRODUCTION
The diagnosis of canine autoimmune skin disease has traditionally depended upon clinical evaluation and assessment of histopathology and direct immunofluorescence (IF) testing (Werner et al., 1983). There are several disadvantages to using direct immunofluorescence. Separate specimens must *Present address: Department of Pathobiological Sciences, School of Veterinary Medicine, University of Wisconsin, 2015 Linden Dr. West, Madison, Wisconsin 53706, U.S.A. **Reprint requests: Dr. M.B. Calderwood Mays, Box J-103 JHMHC, Gainesville, FL 32610, U.S.A.
0165-2427/90/$03.50
© 1990 - - Elsevier Science Publishers B.V.
106
G.A. BRADLEY AND M.B. CALDERWOOD MAYS
be used for histopathologic and IF examinations, which can result in discrepancies between the two samples (Suter et al., 1984). There have also been discrepancies between laboratories (Griffin and Rosenkrantz, 1985). A lesion present on the histologic section may be absent from IF specimens. Morphologic detail is poorly preserved with currently used frozen sections for IF testing. Immunofluorescence staining is labile, and therefore, a permanent record cannot be maintained, making retrospective examination impossible (Haines et al., 1987 ). Occasionally, skin specimens preserved in Michel's solution for IF assay are not received with the formalin fixed skin biopsy specimens for histologic examination. After histologic examination is performed and autoimmune skin disease is considered a likely differential diagnosis, confirmation is desired. To c o n f r m the diagnosis, the clinician must submit an additional biopsy specimen in Michel's solution for IF testing (Muller et al., 1983 ) thus resulting in delay of treatment. Immunoperoxidase (IP) techniques have been applied to suspected cases of canine autoimmune skin disease in an effort to resolve these difficulties (Suter et al., 1984; Haines et al., 1987; Seiler, 1983; Moore et al., 1987). The IP techniques have the advantage of using serial sections of formalin fixed tissue for both histologic examination and immunostaining. Although IP would appear to give similar results to IF, published reports either did not compare results of IP stains to results of IF testing on the same cases (Haines et al., 1987; Moore et al., 1987), used special fixatives (4% formalin saline) and paraffin embedding for IF testing (Suter et al., 1984), or performed IP stains on frozen sections (Seiler, 1983). The purpose of our study was to compare the results of IF testing performed on traditional frozen sections with IP staining techniques performed on formalin-fixed, paraffin-embedded tissues from cases ofautoimmune skin disease, and to evaluate the usefulness of IP staining for retrospective examination of canine autoimmune skin diseases. MATERIALS AND METHODS
Records of the University of Florida Veterinary Medical Teaching Hospital, Surgical Pathology service from January 1986 through March 1989, were searched for cases of possible canine autoimmune skin disease. Forty-five cases were identified. The original hematoxylin and eosin ( HE )-stained glass slides were examined, independently, by the authors. Histologic features were considered together with clinical history, signalment and, when available, results of IF tests for canine immunoglobulin on frozen sections of skin from the affected animals. From the original 45 cases, 22 were considered to exhibit canine autoimmune skin disease using established criteria (Muller et al., 1983; Jubb et al., 1985 ). Five-micron-thick paraffin sections were cut and glued to glass slides with Elmer's glue diluted to spreadable consistency with tap water.
DETECTION OF AUTOANTIBODIESIN CANINE AUTOIMMUNESKIN DISEASE
107
Slides were incubated in a 56°C oven for 30 min, then rehydrated. Three percent hydrogen peroxide (H202) in distilled water was applied for 5 min, then slides were rinsed under running tap water for 5 min. Slides were briefly rinsed in phosphate buffered saline (PBS) and treated with 0.25% trypsin (Gibco, Grand Island, NY) in PBS for 45 min at 37°C, then rinsed under running tap water for 5 min, followed by three brief rinses in PBS. Slides were stained by Garvin's modification (Garvin et al., 1974) of Mason's immunobridge technique (Mason et al., 1969). Normal goat serum ( 1 : 50) (Organon Teknika-Cappel, Westchester, PA) was applied to the slides for 20 min. Excess serum was drained off, and the slides were incubated with one of the primary antibodies (rabbit anti-canine IgG, IgM, IgA) ( 1:500 ) (Miles Scientific, Napperville, IL) or with normal rabbit serum (NRS) (1:500) for 2 h in a moist chamber (anti-canine IgM and IgA were heavy chain specific). Slides were washed three times in PBS for 5 rain each. The secondary antibody (goat anti-rabbit immunoglobulin, 1:50) (Organon Teknika-Cappel, Westchester, PA) was applied for 15 min, followed by three 5-rain rinses in PBS. Next, rabbit anti-horseradish peroxidase (1: 100) (Organon-TeknikaCappel, Westchester, PA) was applied for 15 min, followed by three 5-min rinses in PBS. Horseradish peroxidase ( l mg°/0) (Sigma, St. Louis, MO ) was then applied for 15 rain, followed by three 5-min washes in PBS. The slides were placed in diaminobenzidine (DAB) (Sigma, St. Louis, MO) ( 15 mg/ 50 ml PBS) containing three drops of 3% H202 for 3-7 minutes. Slides were washed under running tap water when brown color had darkened sufficiently, and then counterstained with Harris' hematoxylin for 30 to 60 s. Slides were dehydrated and permanently coverslipped. Dermal plasma cells served as positive internal controls. NRS-stained slides served as negative controls. Slides were examined, independently, by the two authors by light microscopy. Criteria for a positive reaction were discrete outlining of intercellular regions of epidermis and/or staining of the basement membrane. Diffuse staining of the cell cytoplasm in the epidermis was not considered a positive reaction. The IF assays were all performed in the immunology laboratory at the University of Florida at the time of the original submission. The results were obtained from the case records. All assays were performed with sheep anti-canine IgG (not rendered heavy chain specific). The technique has been previously described (Werner et al., 1983 ). RESULTS
Evaluation of HE-stained slides, clinical history and IF results on 45 selected cases of canine skin disease yielded 22 for which the likelihood of autoimmune skin disease was considered very high; these included eight cases of systemic or discoid lupus erythematosus (LE), three ofpemphigus vulgaris
108
G.A. BRADLEY
AND M.B. CALDERWOOD
MAYS
(PV), and 11 o f pemphigus foliaceus (PF). IF results were available on 17/ 22 cases. The remaining cases were accepted based on classic histologic lesions, clinical history, and ANA titers where applicable. Bullous pemphigoid, pemphigus erythematosus and pemphigus vegetans were not represented in our case material. IF results were reported as follows for 17 cases: 3/5 cases o f LE in which IF was performed had positive basement membrane fluorescence (one case) or multifocal intercellular fluorescence (two cases ). ANA titers > 1 : 20 were accepted as positive, and 3/5 o f these cases, in which titers were measured, had positive titers. Of the two cases o f PV for which IF results were available, the results were reported as "positive intercellular" in both instances. IF results were "positive intercellular" in 2 / 1 0 cases o f PF. In one additional case, the TABLE
1
Results of immunofluorescence and immunoperoxidase tests Case #
Diagnosis
IF results
Location
I
IP results
Location
IgA
IgM
IgG
-
-
+
BM I and
BM
C88-1413
LE l
ND
C88-1251
LE
.
C87-433
LE
ND
-
-
+
C87-1988
LE
+
IC
-
-
-
C86-1104
LE
+
IC
-
-
+
IC
C86-1162
LE
-
-
-
+
IC
C86-2805
LE
+
-
-
-
C86-882
LE
ND
2(a)
-
+
+
BM
(b)
+
-
+
BM
-
.
.
BM
C88-2420
PV
+
IC
-
-
C87-1606
PV
+
IC
-
-
+
IC
C89-579
PV
ND
-
-
+
BM
C88-1812
PF
-
-
-
+
IC
C87-1634
PF
+
-
-
+
IC
C87-1956
PF
-
-
-
+
IC
C87-2045
PF
.
C86-423
PF
ND
-
-
+
IC
C86-1345
PF
+
-
-
+
IC
C86-1567
PF
.
C86-1800
PF
+
-
-
C86-2283
PF
.
.
.
C86-2553
PF
.
.
.
C86-2635
PF
-
-
+
IC
.
.
IC .
and
IC
and
IC
.
. BM 3
IC l
.
. . . -
IC
)LE=lupus erythematosus, PV=pemphigus vulgaris, PF=pemphigus foliaceus, ND=not done, basement membrane, I C = intercellular. 2 T w o specimens from separate sites. 3Inconsistent with histologic diagnosis.
BM =
DETECTIONOF AUTOANTIBODIESIN CANINEAUTOIMMUNESKINDISEASE
109
Fig. 1. Lupus erythematosus skin stained with anti-canine IgG, demonstrating diffuse staining of the basement membrane (arrow) (C86-882). Bar=25/~m.
Fig. 2. Pemphigus vulgaris mucous membrane stained with anti-canine IgG, demonstrating diffuse intercellular staining (arrow) (C89-579). Bar=25/~m.
pattern of fluorescence followed the basement membrane and was not found intercellularly as in the other cases of PF, and thus was not considered "positive". The results of IP staining are recorded in Table 1. Five cases of LE exhibited either diffuse IgG-positive basement membrane staining (one case, two specimens) (Fig. 1 ), intercellular staining of the epidermis (two cases), or a combination of both patterns (two cases). Skin from two locations was examined in one case. Sections from the nose exhibited IgM positive basement
l 10
G.A. BRADLEY AND M.B. CALDERWOOD MAYS
Fig. 3. Pemphigus foliaceus skin stained with anti-canine IgG, demonstrating intercellular staining, especially in the superficial epidermis (arrow) (C87-1634 ). Bars = 25/~m.
membrane staining in addition to the positive IgG staining, while sections from an unspecified location (but not nose) yielded a positive basement membrane staining for IgA in addition to IgG. Two cases of PV demonstrated positive intercellular staining with IgG (Fig. 2), and one of these also had strong basement membrane staining. All were negative for IgA and IgM deposits. The results of IP staining for IgG on the 11 cases of PF yielded positive, multifocal, intercellular staining which discretely outlined cells of the superficial epidermis in six cases (Fig. 3 ). Staining intensity varied from weak to strong between cases. Stains for IgA and IgM were negative in all cases. Diffuse staining of epithelial cell cytoplasm was occasionally present in areas of inflammation or epithelial erosion and was not considered a positive reaction. Four cases were negative for both IF and IP. In all cases examined by the immunobridge IP technique, background staining was low in IgG-stained sections and almost nonexistent in IgA- and IgM-stained slides. Negative controls were uniformly negative. Dermal plasma cells demonstrated strong cytoplasmic positivity for IgG, and rarely, IgM. In all slides, the hematoxylin counterstain allowed simultaneous evaluation of histologic features and the antibody staining pattem. No fading was detected in slides stained for up to 10 months. Of the eight cases of LE examined, 2/5 were confirmed by IF that were missed by IP. Conversely, 1/5 was confirmed by IP, but missed by IF. 1/2 cases of PV in which IF was done, was missed by IP but confirmed by IF. The other case was positive by both methods. Neither of the two cases of PF definitively confirmed by IF was missed by the IP technique. Conversely, 3/10 cases confirmed by IP were missed by IF.
DETECTION OF AUTOANTIBODIES IN CANINE AUTOIMMUNE SKIN DISEASE
l 11
DISCUSSION
We obtained slightly better overall results with IP compared to IF. Percent positive results for IF in our case material (47%) closely resembles those previously reported for 84 cases of autoimmune skin disease in dogs (52%) (Werner et al., 1983 ). IP staining of formalin-fixed, paraffin-embedded tissue from the same cases yielded a higher percentage of positives (59%) overall. ARhough our number of cases was small and not subjected to statistical analysis, the difference is noteworthy. The difference was marked when considering PF alone. Whereas only 20% of the cases with IF results were positive, 55% of IP stained slides were positive in PF. Anti-IgG proved to be the best reagent in all three diseases. IgG was detected in all positive cases, with only one case (LE) exhibiting IgA and IgM deposition in addition to IgG. Our specimens were not evaluated for the presence of C3; however, the contribution of C3 in canine autoimmune skin disease is reportedly minimal, occurs only in areas of acantholysis, and is always accompanied by positivity for antibody (Scott et al., 1982). False negatives were still a problem with our IP assay, as was found by others using only IF. Four cases diagnosed as PF were negative for antibody deposits by both IF and IP methods. This illustrates the necessity of using clinical signs, history, and histology, as well as immunopathologic tests such as IF or IP to arrive at a diagnosis of autoimmune skin disease, especially of PF. Reliance on immunopathologic methods alone results in a significant number of missed diagnoses (Werner et al., 1983). Small numbers of false negatives for both the IP and IF techniques were detected by a positive reaction on the other test. There are at least two possible explanations for this discrepancy between the techniques. Formalin fixation is known to reduce the antigenicity of tissue via the formation of aldehyde cross-linkages between antigenic sites (Haines et al., 1987; Moore et al., 1987). This effect is especially pronounced in specimens fixed for greater than 24 h. The small number of false negatives for the IP technique may have been due to this cross-linkage effect even though the sections were trypsinized to help restore antigenicity. Secondly, the IF and IP specimens were collected from separate sites on the dogs. Immunoglobulin may be present only transiently in lesions of autoimmune skin disease (Werner et al., 1983). Thus while autoantibody was present in one specimen it may have been absent in the other. Our results indicate that it is better to perform both IF and IP techniques, if possible, when LE or PV are suspected, in order to confirm additional cases missed by one technique alone. Of the seven cases of LE and PV for which IF results were available, application of our IP technique allowed detection of one additional case of LE. Conversely, three cases (two LE and one PV) would have been missed if only the IP technique had been applied. A somewhat
l 12
G.A. BRADLEY AND M.B. CALDERWOOD MAYS
different situation exists for PF. Our IP technique was more sensitive than IF and helped us confirm three additional cases o f PF. N o n e o f the P F cases detected by IF were missed by IP. We chose to utilize an older and somewhat c u m b e r s o m e immunobridge IP technique for examination of our case material. Others have reported good results utilizing the peroxidase-antiperoxidase ( P A P ) (Suter et al., 1984; M o o r e et al., 1987) and avidin-biotin complex ( A B C ) (Haines et al., 1987) IP techniques. In our hands, the i m m u n o b r i d g e IP technique resulted in less background staining than either ABC or PAP. Less background staining made evaluation of slides easier, as positive staining was distinct. In conclusion, our study illustrates that IP staining o f formalin-fixed paraffin-embedded canine skin for antibody deposits compares favorably with the results o f IF testing on the same animal, and m a y confirm additional cases o f a u t o i m m u n e skin disease, especially PF, missed by IF alone. Several other advantages are realized by using an IP method, namely the ability to evaluate the same or serial sections for both morphology and antibody staining, establishment o f a permanent record of stained slides, and the ability to retrospectively examine canine skin for a u t o i m m u n e diseases. ACKNOWLEDGEMENTS The authors wish to thank Betty J. Hall for her technical assistance.
REFERENCES Garvin, A.J., Spicer, S.S., Parmley, R.T. and Munster, A.M., 1974. Immuno-histochemical demonstration of IgG in Reed-Sternberg and other cells in Hodgkins' disease. J. Exp. Med., 139: 1077-1083. Griffin, C.E. and Rosenkrantz, W.R., 1985. Comparison of two laboratories in the diagnosis of immune mediated skin disease. AAVD and ACVD Proceed Annual Meeting I. Haines, D.M., Cooke, E.M. and Clark, E.G., 1987. Avidin-biotin-peroxidase complex immunohistochemistry to detect immunoglobulin in formalin fixed skin biopsies in canine autoimmune skin disease. Can. J. Vet. Res., 51: 104-109. Jubb, K.V.F., Kennedy, P.C. and Palmer, N., 1985. Pathology of Domestic Animals. Academic Press, San Diego, CA, Voi l, 574 pp. Mason, T.E., Phifer, R.F., Spicer, S.S., Swallow, R.A. and Dreskin, R.B., 1969. An immunoglobulin-enzyme bridge method for localizing tissue antigens. J. Histochem. Cytochem., 17" 563. Moore, F.M., White, S.D., Carpenter, J.L. and Torchon, E., 1987. Localization of immunoglobulins and complement by the peroxidase antiperoxidase method in autoimmune and nonautoimmune canine dermatopathies. Vet. Immunol. Immunopathol., 14: 1-9. Muller, G.H., Kirk, R.W. and Scott, D.W., 1983. Small Animal Dermatology. W.B. Saunders, Philadelphia, PA, 889 pp. Scott, D.W., Wolfe, M.J., Smith, C.A. and Lewis, R.M., 1980. The comparative pathology of non-viral bulious skin diseases in domestic animals. Vet. Pathol., 17:257-281.
DETECTIONOFAUTOANTIBODIESIN CANINEAUTOIMMUNESKINDISEASE
113
Scott, D.W., Manning, T.O., Smith, C.A. and Lewis, R.M., 1982. Observations on the immunopathology and therapy of canine pemphigus and pemphigoid. J. Am. Vet. Med. Assoc., 180 1: 48-52. Scott, D.W., Walton, D.K., Manning, T.O., Lewis, R.M. and Smith, C.A., 1983. Pitfalls in immunofluorescence testing in canine dermatology. Cornell Vet, 73:131-136. Seiler, R.J., 1983. Staphylococcal protein a labeled with horseradish peroxidase for the immunohistologic diagnosis of canine autoimmune disease. Am. J. Vet. Res., 44-2:195-200. Suter, M.M., Palmer, D.G., Zindel, S. and Schenk, H., 1984. Pemphigus in the dog: comparison of immunofluorescence and immunoperoxidase method to demonstrate intercellular immunoglobulin in the epidermis. Am. J. Vet. Res., 45-2: 367-369. Vandervelde, M., Husi, E. and Isler, A., 1983. Histological immunoenzyme techniques in canine tissues: evaluation of various methods and modifications. Res. Vet. Sci., 34: 193-198. Werner, L.L., Brown, K.A. and Halliwell, R.E.W., 1983. Diagnosis of autoimmune skin disease in the dog: correlation between histopathologic, direct immunofluorescent and clinical findings. Vet. Immunol. Immunopathol., 5: 47-64.
105
Elsevier Science Publishers B.V., Amsterdam
Immunoperoxidase staining for the detection of autoantibodies in canine autoimmune skin disease; comparison to immunofluorescence results Gregory A. Bradley* and M.B. Calderwood Mays** Department of Comparative and Experimental Pathology, Collegeof VeterinaryMedicine, University of Florida, Gainesville,FL 32610, U.S.A. (Accepted 15 February 1990)
ABSTRACT Bradley, G.A. and Calderwood Mays, M.B., 1990. Immunoperoxidase staining for the detection of autoimmune skin disease; comparison to immunofluorescence results. Vet. Immunol. Immunopathol., 26:105-113. Skin sections from 22 dogs with autoimmune skin disease were stained with anti-canine IgG, IgM and IgA using an immunobridge immunoperoxidase method. Eight cases of lupus erythematosus, three cases of pemphigns vulgaris, and 11 cases of pemphigus foliaceus were included. Results of previously performed, direct immunofluorescence tests for the detection of canine immunoglobulin on skin were available on 17/22 cases. The immunoperoxidase method yielded an overall positive result in 59% (5/8 lupus erythematosus, 2/3 pemphigus vulgaris and 6/11 pemphigns foliaceus) versus an overall positive result of 47% for direct immunofluorescence (3/5 lupus erythematosus, 2/ 2 pemphigus vulgaris and 2 / 10 pemphigns foliaceus). The immunobridge immunoperoxidase method compared favorably to direct immunofluorescenee testing of canine skin for autoantibody in cases of lupus erythematosis and pemphigus vulgaris, and was superior in cases of pemphigns foliaceus. This method should prove useful as an aid in the diagnosis of canine autoimmune skin disease.
INTRODUCTION
The diagnosis of canine autoimmune skin disease has traditionally depended upon clinical evaluation and assessment of histopathology and direct immunofluorescence (IF) testing (Werner et al., 1983). There are several disadvantages to using direct immunofluorescence. Separate specimens must *Present address: Department of Pathobiological Sciences, School of Veterinary Medicine, University of Wisconsin, 2015 Linden Dr. West, Madison, Wisconsin 53706, U.S.A. **Reprint requests: Dr. M.B. Calderwood Mays, Box J-103 JHMHC, Gainesville, FL 32610, U.S.A.
0165-2427/90/$03.50
© 1990 - - Elsevier Science Publishers B.V.
106
G.A. BRADLEY AND M.B. CALDERWOOD MAYS
be used for histopathologic and IF examinations, which can result in discrepancies between the two samples (Suter et al., 1984). There have also been discrepancies between laboratories (Griffin and Rosenkrantz, 1985). A lesion present on the histologic section may be absent from IF specimens. Morphologic detail is poorly preserved with currently used frozen sections for IF testing. Immunofluorescence staining is labile, and therefore, a permanent record cannot be maintained, making retrospective examination impossible (Haines et al., 1987 ). Occasionally, skin specimens preserved in Michel's solution for IF assay are not received with the formalin fixed skin biopsy specimens for histologic examination. After histologic examination is performed and autoimmune skin disease is considered a likely differential diagnosis, confirmation is desired. To c o n f r m the diagnosis, the clinician must submit an additional biopsy specimen in Michel's solution for IF testing (Muller et al., 1983 ) thus resulting in delay of treatment. Immunoperoxidase (IP) techniques have been applied to suspected cases of canine autoimmune skin disease in an effort to resolve these difficulties (Suter et al., 1984; Haines et al., 1987; Seiler, 1983; Moore et al., 1987). The IP techniques have the advantage of using serial sections of formalin fixed tissue for both histologic examination and immunostaining. Although IP would appear to give similar results to IF, published reports either did not compare results of IP stains to results of IF testing on the same cases (Haines et al., 1987; Moore et al., 1987), used special fixatives (4% formalin saline) and paraffin embedding for IF testing (Suter et al., 1984), or performed IP stains on frozen sections (Seiler, 1983). The purpose of our study was to compare the results of IF testing performed on traditional frozen sections with IP staining techniques performed on formalin-fixed, paraffin-embedded tissues from cases ofautoimmune skin disease, and to evaluate the usefulness of IP staining for retrospective examination of canine autoimmune skin diseases. MATERIALS AND METHODS
Records of the University of Florida Veterinary Medical Teaching Hospital, Surgical Pathology service from January 1986 through March 1989, were searched for cases of possible canine autoimmune skin disease. Forty-five cases were identified. The original hematoxylin and eosin ( HE )-stained glass slides were examined, independently, by the authors. Histologic features were considered together with clinical history, signalment and, when available, results of IF tests for canine immunoglobulin on frozen sections of skin from the affected animals. From the original 45 cases, 22 were considered to exhibit canine autoimmune skin disease using established criteria (Muller et al., 1983; Jubb et al., 1985 ). Five-micron-thick paraffin sections were cut and glued to glass slides with Elmer's glue diluted to spreadable consistency with tap water.
DETECTION OF AUTOANTIBODIESIN CANINE AUTOIMMUNESKIN DISEASE
107
Slides were incubated in a 56°C oven for 30 min, then rehydrated. Three percent hydrogen peroxide (H202) in distilled water was applied for 5 min, then slides were rinsed under running tap water for 5 min. Slides were briefly rinsed in phosphate buffered saline (PBS) and treated with 0.25% trypsin (Gibco, Grand Island, NY) in PBS for 45 min at 37°C, then rinsed under running tap water for 5 min, followed by three brief rinses in PBS. Slides were stained by Garvin's modification (Garvin et al., 1974) of Mason's immunobridge technique (Mason et al., 1969). Normal goat serum ( 1 : 50) (Organon Teknika-Cappel, Westchester, PA) was applied to the slides for 20 min. Excess serum was drained off, and the slides were incubated with one of the primary antibodies (rabbit anti-canine IgG, IgM, IgA) ( 1:500 ) (Miles Scientific, Napperville, IL) or with normal rabbit serum (NRS) (1:500) for 2 h in a moist chamber (anti-canine IgM and IgA were heavy chain specific). Slides were washed three times in PBS for 5 rain each. The secondary antibody (goat anti-rabbit immunoglobulin, 1:50) (Organon Teknika-Cappel, Westchester, PA) was applied for 15 min, followed by three 5-rain rinses in PBS. Next, rabbit anti-horseradish peroxidase (1: 100) (Organon-TeknikaCappel, Westchester, PA) was applied for 15 min, followed by three 5-min rinses in PBS. Horseradish peroxidase ( l mg°/0) (Sigma, St. Louis, MO ) was then applied for 15 rain, followed by three 5-min washes in PBS. The slides were placed in diaminobenzidine (DAB) (Sigma, St. Louis, MO) ( 15 mg/ 50 ml PBS) containing three drops of 3% H202 for 3-7 minutes. Slides were washed under running tap water when brown color had darkened sufficiently, and then counterstained with Harris' hematoxylin for 30 to 60 s. Slides were dehydrated and permanently coverslipped. Dermal plasma cells served as positive internal controls. NRS-stained slides served as negative controls. Slides were examined, independently, by the two authors by light microscopy. Criteria for a positive reaction were discrete outlining of intercellular regions of epidermis and/or staining of the basement membrane. Diffuse staining of the cell cytoplasm in the epidermis was not considered a positive reaction. The IF assays were all performed in the immunology laboratory at the University of Florida at the time of the original submission. The results were obtained from the case records. All assays were performed with sheep anti-canine IgG (not rendered heavy chain specific). The technique has been previously described (Werner et al., 1983 ). RESULTS
Evaluation of HE-stained slides, clinical history and IF results on 45 selected cases of canine skin disease yielded 22 for which the likelihood of autoimmune skin disease was considered very high; these included eight cases of systemic or discoid lupus erythematosus (LE), three ofpemphigus vulgaris
108
G.A. BRADLEY
AND M.B. CALDERWOOD
MAYS
(PV), and 11 o f pemphigus foliaceus (PF). IF results were available on 17/ 22 cases. The remaining cases were accepted based on classic histologic lesions, clinical history, and ANA titers where applicable. Bullous pemphigoid, pemphigus erythematosus and pemphigus vegetans were not represented in our case material. IF results were reported as follows for 17 cases: 3/5 cases o f LE in which IF was performed had positive basement membrane fluorescence (one case) or multifocal intercellular fluorescence (two cases ). ANA titers > 1 : 20 were accepted as positive, and 3/5 o f these cases, in which titers were measured, had positive titers. Of the two cases o f PV for which IF results were available, the results were reported as "positive intercellular" in both instances. IF results were "positive intercellular" in 2 / 1 0 cases o f PF. In one additional case, the TABLE
1
Results of immunofluorescence and immunoperoxidase tests Case #
Diagnosis
IF results
Location
I
IP results
Location
IgA
IgM
IgG
-
-
+
BM I and
BM
C88-1413
LE l
ND
C88-1251
LE
.
C87-433
LE
ND
-
-
+
C87-1988
LE
+
IC
-
-
-
C86-1104
LE
+
IC
-
-
+
IC
C86-1162
LE
-
-
-
+
IC
C86-2805
LE
+
-
-
-
C86-882
LE
ND
2(a)
-
+
+
BM
(b)
+
-
+
BM
-
.
.
BM
C88-2420
PV
+
IC
-
-
C87-1606
PV
+
IC
-
-
+
IC
C89-579
PV
ND
-
-
+
BM
C88-1812
PF
-
-
-
+
IC
C87-1634
PF
+
-
-
+
IC
C87-1956
PF
-
-
-
+
IC
C87-2045
PF
.
C86-423
PF
ND
-
-
+
IC
C86-1345
PF
+
-
-
+
IC
C86-1567
PF
.
C86-1800
PF
+
-
-
C86-2283
PF
.
.
.
C86-2553
PF
.
.
.
C86-2635
PF
-
-
+
IC
.
.
IC .
and
IC
and
IC
.
. BM 3
IC l
.
. . . -
IC
)LE=lupus erythematosus, PV=pemphigus vulgaris, PF=pemphigus foliaceus, ND=not done, basement membrane, I C = intercellular. 2 T w o specimens from separate sites. 3Inconsistent with histologic diagnosis.
BM =
DETECTIONOF AUTOANTIBODIESIN CANINEAUTOIMMUNESKINDISEASE
109
Fig. 1. Lupus erythematosus skin stained with anti-canine IgG, demonstrating diffuse staining of the basement membrane (arrow) (C86-882). Bar=25/~m.
Fig. 2. Pemphigus vulgaris mucous membrane stained with anti-canine IgG, demonstrating diffuse intercellular staining (arrow) (C89-579). Bar=25/~m.
pattern of fluorescence followed the basement membrane and was not found intercellularly as in the other cases of PF, and thus was not considered "positive". The results of IP staining are recorded in Table 1. Five cases of LE exhibited either diffuse IgG-positive basement membrane staining (one case, two specimens) (Fig. 1 ), intercellular staining of the epidermis (two cases), or a combination of both patterns (two cases). Skin from two locations was examined in one case. Sections from the nose exhibited IgM positive basement
l 10
G.A. BRADLEY AND M.B. CALDERWOOD MAYS
Fig. 3. Pemphigus foliaceus skin stained with anti-canine IgG, demonstrating intercellular staining, especially in the superficial epidermis (arrow) (C87-1634 ). Bars = 25/~m.
membrane staining in addition to the positive IgG staining, while sections from an unspecified location (but not nose) yielded a positive basement membrane staining for IgA in addition to IgG. Two cases of PV demonstrated positive intercellular staining with IgG (Fig. 2), and one of these also had strong basement membrane staining. All were negative for IgA and IgM deposits. The results of IP staining for IgG on the 11 cases of PF yielded positive, multifocal, intercellular staining which discretely outlined cells of the superficial epidermis in six cases (Fig. 3 ). Staining intensity varied from weak to strong between cases. Stains for IgA and IgM were negative in all cases. Diffuse staining of epithelial cell cytoplasm was occasionally present in areas of inflammation or epithelial erosion and was not considered a positive reaction. Four cases were negative for both IF and IP. In all cases examined by the immunobridge IP technique, background staining was low in IgG-stained sections and almost nonexistent in IgA- and IgM-stained slides. Negative controls were uniformly negative. Dermal plasma cells demonstrated strong cytoplasmic positivity for IgG, and rarely, IgM. In all slides, the hematoxylin counterstain allowed simultaneous evaluation of histologic features and the antibody staining pattem. No fading was detected in slides stained for up to 10 months. Of the eight cases of LE examined, 2/5 were confirmed by IF that were missed by IP. Conversely, 1/5 was confirmed by IP, but missed by IF. 1/2 cases of PV in which IF was done, was missed by IP but confirmed by IF. The other case was positive by both methods. Neither of the two cases of PF definitively confirmed by IF was missed by the IP technique. Conversely, 3/10 cases confirmed by IP were missed by IF.
DETECTION OF AUTOANTIBODIES IN CANINE AUTOIMMUNE SKIN DISEASE
l 11
DISCUSSION
We obtained slightly better overall results with IP compared to IF. Percent positive results for IF in our case material (47%) closely resembles those previously reported for 84 cases of autoimmune skin disease in dogs (52%) (Werner et al., 1983 ). IP staining of formalin-fixed, paraffin-embedded tissue from the same cases yielded a higher percentage of positives (59%) overall. ARhough our number of cases was small and not subjected to statistical analysis, the difference is noteworthy. The difference was marked when considering PF alone. Whereas only 20% of the cases with IF results were positive, 55% of IP stained slides were positive in PF. Anti-IgG proved to be the best reagent in all three diseases. IgG was detected in all positive cases, with only one case (LE) exhibiting IgA and IgM deposition in addition to IgG. Our specimens were not evaluated for the presence of C3; however, the contribution of C3 in canine autoimmune skin disease is reportedly minimal, occurs only in areas of acantholysis, and is always accompanied by positivity for antibody (Scott et al., 1982). False negatives were still a problem with our IP assay, as was found by others using only IF. Four cases diagnosed as PF were negative for antibody deposits by both IF and IP methods. This illustrates the necessity of using clinical signs, history, and histology, as well as immunopathologic tests such as IF or IP to arrive at a diagnosis of autoimmune skin disease, especially of PF. Reliance on immunopathologic methods alone results in a significant number of missed diagnoses (Werner et al., 1983). Small numbers of false negatives for both the IP and IF techniques were detected by a positive reaction on the other test. There are at least two possible explanations for this discrepancy between the techniques. Formalin fixation is known to reduce the antigenicity of tissue via the formation of aldehyde cross-linkages between antigenic sites (Haines et al., 1987; Moore et al., 1987). This effect is especially pronounced in specimens fixed for greater than 24 h. The small number of false negatives for the IP technique may have been due to this cross-linkage effect even though the sections were trypsinized to help restore antigenicity. Secondly, the IF and IP specimens were collected from separate sites on the dogs. Immunoglobulin may be present only transiently in lesions of autoimmune skin disease (Werner et al., 1983). Thus while autoantibody was present in one specimen it may have been absent in the other. Our results indicate that it is better to perform both IF and IP techniques, if possible, when LE or PV are suspected, in order to confirm additional cases missed by one technique alone. Of the seven cases of LE and PV for which IF results were available, application of our IP technique allowed detection of one additional case of LE. Conversely, three cases (two LE and one PV) would have been missed if only the IP technique had been applied. A somewhat
l 12
G.A. BRADLEY AND M.B. CALDERWOOD MAYS
different situation exists for PF. Our IP technique was more sensitive than IF and helped us confirm three additional cases o f PF. N o n e o f the P F cases detected by IF were missed by IP. We chose to utilize an older and somewhat c u m b e r s o m e immunobridge IP technique for examination of our case material. Others have reported good results utilizing the peroxidase-antiperoxidase ( P A P ) (Suter et al., 1984; M o o r e et al., 1987) and avidin-biotin complex ( A B C ) (Haines et al., 1987) IP techniques. In our hands, the i m m u n o b r i d g e IP technique resulted in less background staining than either ABC or PAP. Less background staining made evaluation of slides easier, as positive staining was distinct. In conclusion, our study illustrates that IP staining o f formalin-fixed paraffin-embedded canine skin for antibody deposits compares favorably with the results o f IF testing on the same animal, and m a y confirm additional cases o f a u t o i m m u n e skin disease, especially PF, missed by IF alone. Several other advantages are realized by using an IP method, namely the ability to evaluate the same or serial sections for both morphology and antibody staining, establishment o f a permanent record of stained slides, and the ability to retrospectively examine canine skin for a u t o i m m u n e diseases. ACKNOWLEDGEMENTS The authors wish to thank Betty J. Hall for her technical assistance.
REFERENCES Garvin, A.J., Spicer, S.S., Parmley, R.T. and Munster, A.M., 1974. Immuno-histochemical demonstration of IgG in Reed-Sternberg and other cells in Hodgkins' disease. J. Exp. Med., 139: 1077-1083. Griffin, C.E. and Rosenkrantz, W.R., 1985. Comparison of two laboratories in the diagnosis of immune mediated skin disease. AAVD and ACVD Proceed Annual Meeting I. Haines, D.M., Cooke, E.M. and Clark, E.G., 1987. Avidin-biotin-peroxidase complex immunohistochemistry to detect immunoglobulin in formalin fixed skin biopsies in canine autoimmune skin disease. Can. J. Vet. Res., 51: 104-109. Jubb, K.V.F., Kennedy, P.C. and Palmer, N., 1985. Pathology of Domestic Animals. Academic Press, San Diego, CA, Voi l, 574 pp. Mason, T.E., Phifer, R.F., Spicer, S.S., Swallow, R.A. and Dreskin, R.B., 1969. An immunoglobulin-enzyme bridge method for localizing tissue antigens. J. Histochem. Cytochem., 17" 563. Moore, F.M., White, S.D., Carpenter, J.L. and Torchon, E., 1987. Localization of immunoglobulins and complement by the peroxidase antiperoxidase method in autoimmune and nonautoimmune canine dermatopathies. Vet. Immunol. Immunopathol., 14: 1-9. Muller, G.H., Kirk, R.W. and Scott, D.W., 1983. Small Animal Dermatology. W.B. Saunders, Philadelphia, PA, 889 pp. Scott, D.W., Wolfe, M.J., Smith, C.A. and Lewis, R.M., 1980. The comparative pathology of non-viral bulious skin diseases in domestic animals. Vet. Pathol., 17:257-281.
DETECTIONOFAUTOANTIBODIESIN CANINEAUTOIMMUNESKINDISEASE
113
Scott, D.W., Manning, T.O., Smith, C.A. and Lewis, R.M., 1982. Observations on the immunopathology and therapy of canine pemphigus and pemphigoid. J. Am. Vet. Med. Assoc., 180 1: 48-52. Scott, D.W., Walton, D.K., Manning, T.O., Lewis, R.M. and Smith, C.A., 1983. Pitfalls in immunofluorescence testing in canine dermatology. Cornell Vet, 73:131-136. Seiler, R.J., 1983. Staphylococcal protein a labeled with horseradish peroxidase for the immunohistologic diagnosis of canine autoimmune disease. Am. J. Vet. Res., 44-2:195-200. Suter, M.M., Palmer, D.G., Zindel, S. and Schenk, H., 1984. Pemphigus in the dog: comparison of immunofluorescence and immunoperoxidase method to demonstrate intercellular immunoglobulin in the epidermis. Am. J. Vet. Res., 45-2: 367-369. Vandervelde, M., Husi, E. and Isler, A., 1983. Histological immunoenzyme techniques in canine tissues: evaluation of various methods and modifications. Res. Vet. Sci., 34: 193-198. Werner, L.L., Brown, K.A. and Halliwell, R.E.W., 1983. Diagnosis of autoimmune skin disease in the dog: correlation between histopathologic, direct immunofluorescent and clinical findings. Vet. Immunol. Immunopathol., 5: 47-64.
