Aug. 1982

bithionol and halogenated salicylanilide photosensit ivity. Br J Dermatol 82:230-242, 1970

22. de Sousa MAB, Panott DMV: Induction and recall in contact sensitivi ty changes in skin and draining lymph nodes of intact a nd thymectomized mice. J Exp Med 130:671-690, 1969

23. Landsteiner K , Chase MW: Experiments on transfe r of cutaneous sensit ivi ty to simple compound . Proc Soc Exp Bioi Med 49:688-690, 1942

24. Bloom BR, Chase MW: Tra nsfer of delayed-type hypersensitivity. A crit ical review a nd experimenta l study in the guinea p ig. Prog Allerg 10:151-255, 1967

25. Rosenstreich DL, B lake JT, Rosenthal AS: The peritoneal exudate lymphocyte. I. Differences in antigen re ponsiveness between peritoneal exudate and lymph node lymphocytes from immunized guinea pigs. J Exp Med 134:1170-1186, 1971

26. van Boxel JA, Rosenstreich DL: Binding of aggregated y-globulin to activa ted T lymphocytes in the guinea pig. J Exp Med 139:1002-1012, 1974

27. Asherson GL, Zembala M: Contact sensit ivity in the mouse. IV. The role of lymphocytes a nd macrophages in passive transfer and the mechanism of their in teraction. J E xp Med 132: ] - ]5, 1970

28. Miller JFAP, Vadas MA, Whitelaw A, Gamble J: Role of major histocompatibility complex gene products in delayed-type hyp er­sensitivity. P roc Natl Acad Sci 73:2486-2490, 1976

29. Benacerraf B, Germa in RN: The immune response genes of the major histocompatibility complex. Immunol Rev 38:70-117, 1978

30. Zinkernagel RM, Doherty PC: MHC-restricted cytotoxic T cells: Studies on the biological role of polymorphic major tra nsplan­tation antigens determining T -cell restriction-specificity, fu nc­tion, and responsiveness. Adv Immunol 27:51- ] 77, 1979

31. Weinberger JZ, Greene MI, Benacerra f B: Hapten-specific T -cell responses to 4-hydroxy-3-nitrophenyl acetyl. I. Genetic control of delayed-type hypersensitivity by V" and I -A-region genes. J

0022-202X/82/7902-0J 15$02.00/ 0 TH E J OURNAL 0 1' I NVEST IGATIVE DERMATOLOG Y. 79:115-11 8, 1982 Copyright © 1982 by The Williams & Wilkins Co.

CONTACT PHOTOSENSITIVITY TO TCSA IN MICE 115

Exp Med 149:1336-1348, 1979 32. Sunday ME, Benacerraf B, Dorf ME: Hapten-specific T cell re­

sponses to 4-hydroxy-3-nitrophenyl acetyl. VI. Evidence for dif­ferent T cell receptors in cells that mediate H-2-1 restricted and H-2D-restricted cutaneous sensit ivity responses. J Exp Med 152:1554-1562, 1980

33. Asherson GL, Mayhew B, Perera MACC: The production of contact sensitivity by the injection into the foot pad of recipients of lymph node cells from mice. 1 day after painting the skin with contact sensitizer agent: Requirement for matching at the major histocompatibility complex between donor and recipient mice. Immunology 37:241-245, 1979

34 . Thomas WR, Smith FI, Walker ID, Miller JFAP: Contact sensitiv­ity to azobenzenearsonate a nd its inhibition after interaction of sensitized cells with antigen-conjugated cells. J Exp Med 153:1124-1137, 1981

35. Zinkernagel RM: H-2 restriction of virus-specific T -cell-mediated affector functions in vivo. II. Adoptive transfer of delayed-type hypersensitivity to murine lymphocytic choriomeningitis virus is restricted by the K and D region of H-2. J Exp Med 144:776-787, 1976

36. Kochevar IE, Harber LC: P hotoreactions of 3,3' ,4 ',5-tetrachloro­salicylanilide with proteins. J Invest Dermatol 68: 151-156, 1977

37. J enkins FP, Welt i D, Baines D: Photochemical reactions of tetrach­lorosalicylanilide. Nature 201:827-828, 1964.

38. Shearer GM, Schmitt-Verhulst A-M: Major histocompatibility complex restricted cell-mediated immunity. Adv Immunol 25:55-91, 1977

39. Elliott BE, Haskill JS, Axelrad MA: Rosette forming ability of thymus-derived lymphocytes in cell-mediated immunity. I. De­layed hypersensitivity a nd in vitro cytotoxicity. J Exp Med 141:584-599, 1975

40. Ison A, Blank H: Testing drug phototoxicity in mice. J Invest Dermatol 49:508-511, 1967

Vol. 79. No.2 Prin.ted in U.S.A .

Antibodies to Human Epidermal Cytoplasmic Antigens: Incidence, Patterns, and Titers

EDWARD P. PALUCH, PH.D. , AND KURT J. BLOCH, M .D.

The Department of Medicine, Harvard Med,:cal School and the Clinical Im.nl.ll.nology and A llergy Units of the Medical Services Massachu.setts General Hospital. Boston, Massachu.setts, U.S.A .

Serum or plasma specimens were assayed in indirect immunfluorescence tests on cryostat sections of normal human skin for the presence and titer of antibodies reactive with human epidermal cytoplasmic antigens. A polyva lent fluorescein-labeled goat anti-human immu­noglobulin antiserum was used in all tests. Three distinct s taining patterns w ere noted: upper epidermal cytoplas­mic fluorescence, U-CYT, produced by antibodies reac-

Manuscript received July 22, 1980; accepted for publication J a nu ary 22, 1982.

This work was supported by grants from the National Institutes of Health (T32-AI-07040 a nd T32-AM-07258) , The Arthritis Foundation and the L.H. Bendit Foundation.

Reprint requests to: Dr. Edward P . Pa luch, Clinical Immunology & Allergy Units, Massachusetts General Hospital Boston, Massachusetts 02114.

Abbreviations: BCL: basel cell cytoplasmic flu orescence ECA: epidermal cytoplasmic ant igens FITC: flu orescein isothiocyanate G-CYT: general cytoplasmic fluorescence PBS: phospha te-bu ffered saline U-CYT: upper epiderma l cytoplasmic flu orescence

tive with antigen present in cells of the upper and middle layers of the epidermis; general cytoplasmic fluores­cence, G-CYT, produced by antibodies reactive with an­tigens present in cells throughout the epidermis; and basal cell cytoplasmic fluorescence, BCL, produced by antibodies reactive with components present only in basal cells. Sera from 8% of 52 normal blood dOllors produced the U-CYT pattern at dilutions greater than 1:10. The incidence of antibodies reactive with epidermal cytoplasmic antigens in patients with a clinical history of not more than 2 basal cell carcinomas of the skin was 5%, compared to an incidence of 89% in those individuals with 3 or more separate instances of skin neoplasms. There was no difference in the frequency with which cryosurgery was used in the treatment of skin neoplasms in either of these 2 groups. Antibodies to epidermal cytoplasmic antigens were also detected in 10% of pa­tients with nondermatologic, nonpulmonary neoplasms, in 43% of patients with pulmonary neoplasms a nd in 10f 11 patients with nonneoplastic diseases. Positive sera yielded titers ranging from 1:16 to 1:1024. The most common staining patterns noted in all of these cases were the U-CYT and G-CYT patterns; the BCL staining pattern was noted in only one instance.

116 PALUCH AND BLOCH

Ant ibodies which are reactive with several different compo­nents of normal human skin, including t he intercellulru" sub­stance antigen [1,2], basement membrane zone antigens [3], and stratum corneum antigens [4], have been described. In addition, antibodies reactive with epidermal cytoplasmic anti­gens (ECA) of normal human skin have also been identified. These antibodies have been detected in normal individuals [5-7], patients with various malignancies [5,7], burn patients who had rejected a skin graft [6], bone marrow-transplanted patients [8,9] and some individuals with drug reactions [10).

Three different patterns of human epidermal cytoplasmic staining have been recognized by investigators using indu·ect immunofluorescence [6-8,11-14). These 3 patterns have been defmed as follows [7,13]: (1) upper epidermal cytoplasmic flu­orescence, abbreviated U -CYT, produced by antibodies reactive with antigens present in the cytoplasm of the cells in the upper and middle layers of the epidermis, exclusive of the basal cell layer (Figure, C); (2) general cytoplasmic fluorescence (G-CYT) produced by antibodies reactive with cytoplasmic antigens which are present in cells throughout the epidermis (Figure, B); and (3) basal cell cytoplasmic fluorescence (BCL) produced by antibodies reactive with cytoplasmic components present only in the cytoplasm of basal cells of normal human epidermis (Figure D) .

In the present study, serum or plasma samples from normal blood donors, patients with skin neoplasms and patients with various noncutaneous malignancies and nonmalignant condi­tions were examined in indirect immunofluorescence tests for both the presence and titer of antibodies to human ECp

FIG Indirect immunofluorescence tests performed with cryostat sec­tion of normal human skin incubated with (A) normal human serum; (B) serum producing the G-CYT staining pattern; (C) serum producing the U -CYT staining pattern; and (D) serum producing the BCL staining pattern.

Vol. 79, No.2

MATERIALS AND METHODS Serum or Plasma Specimens

Serum or EDTA plasma specimens were collected f.rom four groups of individuals: 52 blood donors; 31 patients with skin neoplasms; 80 pat ients with various noncutaneous malignancies; and 11 patients with ~onneoplastic conditions. Samples were stored at -20°<;: prior to test­mg.

Subslrale

Human skin specimens obtained by punch biopsy for diagnostic purposes and found to be histologically normal a nd free of immuno­globulins or complement, as determined by immunofluorescence, were used as a substrate in indirect immunofluorescence tests. Skin frag­ments which had been kept moist on a saline-soaked sponge for no longer than 20 min were embedded in OTC embedding compound (Lab­T ek Products, Naperville, ILLinois) , frozen at -40°C and stored at -70°C for no longer than 28 days.

Laboralory Methods

Four-micron cryostat sections of normal human skin were cut just prior to use and allowed to air-dry at room temperature for approxi­mately 20 min. Sections were flxed in acetone for 5 min at room temperature, removed from the fixative and allowed to air dry. Dilu­tions (1:10) of normal human sera were prepared, using .01 M phosphate­buffered saline (PBS) , pH 7.5, containing 1% bovine serum albumin, as a diluent. Tissue sections were covered with the diluted serum and incubated at room temperature in a humid chamber for 45 min. The sections were washed twice for 5 min per wash in PBS and covered with a 1:100 dilution of polyvalent fluorescein isothiocyanate (FITC)­labeled goat anti-human immunoglobulin a ntiserum (Atlantic Anti­bodies, Westbook, Maine), which had an FITC to protein ratio of 2.3 molar and a total protein concentration of 16 mg/ ml. The sections were incubated as described above and were then washed twice with PBS as described. Sections were examined by UV microscopy. Those specimens which were found to react positively with ECA were t itered to the endpoint; the highest dilution yielding visible fluorescence was defined as the titer. Titration was begun with samples diluted 1:8.

ContTols

Sera which were known to contain ant ibodies capable of producing the G-CYT, U-CYT, and BCL staining patterns were included as positive controls in each assay. This procedul·e ensured that the ECA of the substrate had not deteriorated upon storage. Blocking experi­ments were performed on ten sera which contained antibodies reactive with human ECA. A 1:10 dilution of each of these sera was incubated for 45 min at room temperature with cryostat sections of norma l human epidermis, washed twice for 5 min per wash in PBS, and incubated for 45 min with a 1:10 dilution of unlabeled goat anti-human immunoglob­ulin antiserum (Hyland Laboratories, Costa Mesa, CAl. The slides were then washed as described and incubated for 45 min with a 1:100 dilu t ion of fluorescein labeled goat a nti-human inlmunoglobulin antiserum . After additional washing, the slides were examined as described.

Statistical Analyses

Analyses were performed by the Chi-sq uru·e method.

RESULTS

Indirect Immunofluorescence Assay

Fifty-two serum specimens obtained from blood donors were assayed in indirect immunofluorescence tests at dilutions of 1:10 on cryostat sections of normal human epidermis. Of these, 18% were found to contain antibodies .producing the U-CYT, 6% the G-CYT and none the BCL staining pattern. However, only 8% of the 52 serum specimens yielded positive staining at dilutions greater than 1:10; in all of these instances, the U-CYT staining pattern was noted (Table).

The incidence of antibodies reactive with ECA in samples from patients with a clinical history of 1 or 2 basal cell carci­nomas of the skin, was approximately 5%, compared to an incidence of 89% in those indivi~uals with 3 or more separate instances of skin neoplasms (usually consisting of only basal cell carcinomas, but in some cases, of both basal and squamous cell carcinomas). This difference was significant (p < 0.005) .

Aug. 1982 ANTIBODIES TO HUMAN EPIDERMAL ANTIGENS 117

TABLE Incidence, patterns and titers of antibodies to hu.man epidermal cytoplasmic antigens in various categories

Category Mean age Cytoplasmic staining paLLerns and titers

No. Pt ·. No. + %+ U-CYT G-CYT BCL

Blood donors 28 52 4 8 16, 16, 32, 32 Skin neoplasms (52 separate instances) 51 22 1 5 64 Skin neoplasms (2:3 sepa1'3te instances) 60 9 8 89 256,64, 32", 256, 16", 16"

64, 64, 512, 32" AdenocaJ'cinoma of the lung 58 10 3 30 512, 128 128 Oat ce ll carcinoma of the lung 54 4 2 50 256 16 Squamous cell caJ'cinoma of the lung 58 4 3 75 1024, 64 128 Large cell carcinoma of the lung 53 5 2 40 256 16 Adenocarcinoma of the colon 62 35 3 9 64 <, 256,512 16" Adenocarcinoma of the rectum 63 9 1 1] 64 LeiomyosaJ'coma 46 1 1 100 128 Mesothelioma 36 1 1 100 512 Gastric ulcer 37 3 1 33 256 Miscellaneous disorders' 51 19 0 0

I Those specimens reacting with ECA at dilutions> 1:10 were considered positive. 2 No reactivity with ECA was noted as dilutions 2: 1:10 with specimens obtained from patients with the following miscellaneous disorders:

metastatic SaJ'coma (1); adenocaJ'cinoma of the common bile duct (J); adenocarcinoma of the endometrium (1); parotid gland tumor (2); adenocarcinoma of the breast (4); pancreatic caJ'cinoma (2); actinic keratosis (5); fibrocystic disease of the breast (1); acute and chronic cystitis (1); cholecystitis and cholelithiasis (1).

". I,. "Superscript letters identify those samples yielding more than one staining patterns.

The most common staining pattern was the U-CYT pattern; positive sera yielded titers ranging from 1:16 to 1:512.

Of 80 patients with nondermatologic malignancies, 16 were found to have antibodies reactive with ECA. Approximately 60% of the positive samples contained antibodies producing the U -CYT and 40% the G-CYT staining pattern; titers ranged from 1:16 to 1:1024. Twenty-three of the 80 patients with nondermatologic malignancies had pulmonary neoplasms; of these 23 patients, samples from 10 were positive for antibody to ECA; 60% yielded the U-CYT and 40% the G-CYT staining pattern. Of the remaining 57 patients in this category, 6 were found to have antibodies to ECA; samples from 50% yielded the U-CYT, 17% the G-CYT, 17%.the BCL and 16% yielded both the U-CYT and G-CYT staining patterns. Of 11 patients with nonneoplastic diseases, one was found to have antibodies pro­ducing the U-CYT staining pattern.

The fluorescent staining patterns observed were considered to reflect specific binding of human immunoglobulins to ECA. Cryostat sections of normal human epidermis were incubated first with sera found to contain antibodies to ECA, followed by an incubation with unlabelled goat anti-human immunoglobu­lin antiserum. The subsequent binding of FITC- labeled goat anti-human immunoglobulin antiserum was blocked by the unlabeled goat anti-human immunoglobulin antiserum.

Therapy

Cryosurgery was used to treat the skin neoplasms of 21 of the 22 patients with one or 2 tumors; surgical excision was used to treat the remaining patient. Of 9 patients with a history of 3 or more skin neoplasms, cryosurgery was used to treat at least 3 skin neoplasms on at least 3 separate occasions in each of the 9 patients in this category. Ofthe 80 patients with noncutaneous malignancies, 15 were receiving chemotherapeutic agents at the time serum or plasma specimens were obtained. Of these 15 patients, 27% were found to possess antibodies reactive with ECA. Of the 65 patients who were not receiving chemoth erapy, 17% were found to possess antibodies reactive with ECA (p>0.75).

DISCUSSION

The existence of at least 3 distinct epidermal cytoplasmic antigens has been demonstrated in most, if not all human skin by several investigators [5,7,12,13]. Low titer antibodies to these antigens were noted in 24% of 52 normal blood donors examined in this study; values of 7% [5] and 34% [7] have been reported by others. The results of the present study suggest that both

the incidence and titer of antibodies reactive with human ECA appear to be greater in patients with repeated episodes of skin neoplasms compared to normal blood donors or those with a history of one or two episodes of skin neoplasms. Serum speci­mens from 89% of patients with repeated episodes of skin neoplasms produced the U -CYT staining pattern; 2 of t he sera also produced the G-CYT staining pattern. In contrast only 1 of 22 samples from patients with 1 or 2 instances of basal cell carcinomas of the skin contained antibodies reacting with ECA. The reason for this disparity is uncertain. Possibly, t he repeated use of cryosUJ-gery in the treatment of skin tumors may have a role in inducing these antibodies. Cryosurgical prostatectomy has been reported to result in the formation of antibodies reactive with cytoplasmic constituents of secretory epithelial cells of human prostate [15,16]. Similarly, Bagley and Faraci observed that mice with 3-methylcholantlu-ene-induced fibro­sarcomas which were treated with either cryosurgery or elec­trocauterization had significantly greater humoral and cell me­diated immunity to t heir tumor, as compared to those animals whose tumors were untreated or were surgically removed [17].

In the present study, cryosurgery was repeatedly used to treat the tumors of the nine patients with a history of 3 or more skin neoplasms. This treatment may have resulted in the release of epidermal antigens and repeated exposure of the host to epidermal antigens may have induced the formation of anti­bodies reactive with ECA. Whether other forms of trauma are responsible for the induction of antibodies to ECA in normal subjects is not known. Saurat et al observed the presence of antibodies to ECA in 88% of 17 bone marrow transplant patients [8,9]. Since skin lesions occur commonly after bone marrow transplantation [18], it was suggested that antibodies to ECA may develop as a result of chronic skin damage accompanied by release of epidermal antigens. Of the 80 patients with non­cutaneous malignancies included in this study, 65 were not receiving chemotherapy at the time serum or plasma specimens were obtained. Antibodies to ECA were detected in 17% of these 65 patients and in 27% of t he 15 patients who were receiving chemotherapy. This difference was not statistically significant (p >.75). Therefore, it seems unlikely that antibodies to ECA observed in patients with non cutaneous malignancies included in this study were drug-induced. Furthermore, chemo­therapy did not appear to interfere with the expression of antibodies to ECA.

Although sera from patients with malignant melanoma were not examined in the present study, Abel and Bystryn reported that 34% to 53% of patients with malignant melanoma had circulating antibodies which produced the U-CYT staining

118 PALUCH AND BLOCH

pattern [7]. These antibodies were noted to be presen t in patients with malignant melanoma in a frequency 2.5 times that observed in normal individuals. No correlation was observed between stage of disease and incidence of anti-ECA antibod ies. However, the sera in this study were examined at a 1:5 dilution, with no attempt being made to determine the antibody titers of those sera which were found to react positively in indirect immunofluorescence tests.

The finding of antibodies reactive with ECA is not restricted to patients with skin neoplasms but has also been reported in patients with other malignancies. Bystryn, Abel and Weidman reported the presence of antibodies reactive with ECA in 32% of 34 patients with various malignancies [5]. Similarly, the results of the present study indicate that antibodies reactive with ECA exist in patients with malignancies; these antibodies were found in 19% of patients with noncutaneous malignancies. Of 23 patients with pulmonary neoplasms, 43% were found to have anti-ECA antibodies, compared to an incidence of only 9% in 35 patients with adenocarcinoma of the colon (p < .01). Although sera from patients with other neoplasms were also examined, the small number of cases precludes an accurate estimate of the incidence of anti ECA antibodies.

Antibodies reactive with t he basal cell layer of human epi­dermis were found in the sera of 13% of 24 bum victims [19] and in 16% of 42 patients with drug reactions involving the skin [10]. T hese antibodies have also been noted in patients with pemphigus and bullous pemphigoid; in these patients, titers ranged from 1:40 to 1:1280 [11]. In the current study, one patient with adenocarcinoma of the rectum was found to have circulating antibodies reacting with the cytoplasm of human epidermal basal cells. Sera containing antibodies responsible for the BCL staining pattern continue to produce t his pattern after absorption with A and B blood group substances [6]. Antisera directed against blood group substances A and B do react with human epidermis, but the pattern resembles the staining produced by sera from patients with pemphigus [20]. Finally, basal cells have been shown to lack A and B blood group antigens [21], making it still less likely that th e BCL staining pattern is attributable to these substances.

Epidermal cytoplasmic antigens also do not appear to be HLA antigens, since the latter are distributed on the cell surface rather than intracellularly [22]. Support for the concept that some of the ECA may be differentiation antigens comes from the work of Bystryn and Frances [23]. These investigators noted that sera which produced the U -CYT or BCL staining patterns with normal human skin as substrate did not react with the cytoplasm of all basal or squamous cell carcinomas of the skin. In contrast, sera producing th e G-CYT staining pattern reacted with the cytoplasm of all basal and squamous cell carcinomas of the skin. T hese findings suggest that malignant transfo~mation of cells in the epidermis may, in some cases, result 111 a loss of antigens which are associated with the differentiation of epidermal cells (ie, those antigens responsible for the U-CYT and BCL staining patterns). Additional evidence in support of ECA being differentiation antigens is provided by the work of Saurat et al [9]. Using the technique of indirect immunofluorescence, performed on cryostat sections of rabbit lip, sera containing antibodies reactive with ECA were found not to react identically on mucosal, parakeratotic and ortho­keratotic epithelium. T hese results suggest that ECA are in­volve.d in keratinocyte differentiation.

Vol. 79, No . 2

Although the present study did not determine the immuno­globulin classes of the antibodies reactive with ECA, other investigators h ave reported them to be of the IgG class [8-10]. Whether the antibodies which react with ECA are auto- or isoantibodies is unknown. In one instance, a patient's serum produced the BCL staining pattern in indirect immunofluores­cence tests with allogenic normal human epidermis, but did not react with autologous skin [11].

REFERENCES 1. Beutner EH, Jordon RE: Demonstration of skin antibodies in sera

of pemphigus vu lgaris patients by indirect immunonuorescent staining. Proc Soc Exp BioI Med 117:505-510, 1964

2. Peck SM: Studies in bu llous diseases: Immunonuorescent serologic tests. N Engl J Med 279:951-958, 1968

3. Jordon RE: Basement membrane zone ant ibodies in human pem­phigoid. JAMA 200:751-756, 1967

4. Krough KH: Antibodies in human sera to stratum corneum. lnt Arch Allerg Appl Immunol 36:4 15-427, 1969

5. Bystryn JC, Abel E, Weidman A: Antibodies against the cytoplasm of human epidermal ce lls. Arch Dermatol108:24 1-244 , 1973

6. Ackermann-Schopf C, Ackerman R, T erasaki P I, Levy J : Natural and acquired epidermal autoant ibodies in ma n. J Immunol 112:2063-2067, 1974

7. Abel E, Bystryn JC: Epidermal cytoplasmic antibodies: Incidence and type in normal persons and patients with melanoma. J Invest Dermatol 66:44-48, 1976

8. Saurat JM , Bonnetblanc JM, Gluckman E , Didierjean L, Bussel A, P uissant A: Skin antibodies in bone marrow transplanted pa­tients. Clin Exp Dermatol 1:377-384, 1976

9. Saurat JH, Didierjean L, Galoppin L, Gl uckman E: Antibodies with affinity to the cytoplasm of keratinocytes produced after bone marrow transpla ntation in man. A review. Cut Immunopathol INSERM 80:77-92,1978

10. Van Joost T : Incidence of circulating an tibodies reactive with basal cells of skin in drug reactions. Acta Derm Venereol (Stockh) 54:183-188, 1974

11. Bystryn JC: Clinical significance of basal cell layer antibodies. Arch Dermatol 113:1380-1382, 1977

12. Bystryn JC, Nash M, Robins P: Epidermal cytoplasmic antigens. II. Concurrent presence of ant igens of different specificities in normal human skin. J Invest Dermatol 71:110-113, 1978

13. Burnham TK: Two epidermal cytoplasmic immunoflu orescen t pat­terns detected by indirect immunonuorescence. J Invest Der­matol 63:100-105, 1974

14. Nishikawa T, Kurihara S, Harada T , Hatano H: Binding of bullous pemphigoid antibodies to basal cells. J Invest Dermatol 74:389-391, 1980

15. Ablin RJ , Danaher J , Soanes WA, Gonder MJ : Antibodies reactive with autologous prostatic t issue in adenocarcinoma of th e pros­tate. Urology 3:491-499, 1974

16. Ablin RJ, Gonder MJ , Soioll1es WA: Immunohistologic studies of carcinoma of the prostate. III. E lu tion of interepithelial a nt i­bodies from carcinomatous human prostatic tissue fo llowing cry­oprostatectomy. Oncology 29:329-334, 1974

17. Bagley D, Faraci R: T umor immunity fo llowing cryosW"gery or electrocauterization. Natl Cancer Inst Monogr 49:371-373, 1978

18. Thomas ED: Bone marrow tran. plantation, Clinical Immunobiol­ogy, vol 2. Edited by FH Bach, RA Good. New York, Academic Press, 1974, pp 1-32

19. Thivolet J , Beyvin AJ: Recherche par immuno{luorescence d'autocorps seriques vis-a.-vis constituants de I'epiderme chex les brules. Experientia 24:945-946, 1968

20. Szu LLman EA: T he histological distribution of blood group sub­stances A and B in ma n. J Exp Med 111:785-799, 1960

21. Davidson J, Ni L Y: Immunocytology of cancer. Acta CytoI14:276-282, 1970

22. SaUl'a t JH, Gluckma n E , D idierjean L, Andersen E, Sockeel F , Puissant A: Cytoplasmic and HL-A antigens in human epidermis. Br J Dermatol 96:603-608, 1977

23. Bystryn J C, Frances C: CytoplasmiC differentiation antigens of human epidermal cells. Transplantation 27:392-396, 1979