δ T-cell receptor in dermatitis herpetiformis

GASTROENTEROLOGY

1992;102:1499-1505

Increased Jejunal Intraepithelial Lymphocytes Bearing y/6 T-Cell Receptor in Dermatitis Herpetiformis MAURIZIO AMILCARE Department

VECCHI, LUCIA CROSTI, EMIL10 BERTI, CERRI, and ROBERTO DE FRANCHIS

of Internal Medicine

and First Department

T-cell receptor 1 (r/S) expression was studied in 19 jejunal or duodenal specimens from patients with dermatitis herpetiformis and in 16 jejunal or duodenal specimens showing normal histology. In normal specimens, r/6+ cells represented 10.6% of intraepithelial CD3+ lymphocytes. Around 50% of these cells were recognized by the Al3 monoclonal antibody, which detects products of the Vyl/ViSl gene rearrangement and the non-disulfide-linked form of T-cell receptor 1. The remaining 50% reacted with the BB3 monoclonal antibody, which recognizes products of the Vy9/ViQ rearrangement and the disulfide-linked form of receptor. Very few y/6+ cells were observed in the lamina propria. In jejunal specimens from patients with dermatitis herpetiformis, a significant increase in the prevalence of y/6+ intraepithelial lymphocytes was observed (P < 0.001). This finding was largely accounted for by an increase in those cells recognized by the Al3 monoclonal antibody, thus possibly expressing the Vrl/VGl rearrangement and the nondisulfide-linked form of receptor. These data suggest that similar pathogenetic mechanisms may be active in determining the jejunal damage in celiac disease and dermatitis herpetiformis.

M

ost peripheral blood CD3+ lymphocytes in humans and animals express the T-cell receptor 2 or TcR2 composed of the two disulfide-linked subunits a and p.‘*” Recently, a second CDS-associated T-cell receptor termed y/6 or TcRl has been identified in a small proportion of peripheral T cells.3*4 In humans, most of these circulating TcRl+ lymphocytes are CD8- CD4-. In the mouse, y/6 T lymphocytes represent around 50% of gut intraepithelial lymphocytes and are CD8+.5*6 These findings have prompted a suggestion of a possible specific role for TCRl+ cells in mucosal immunity.7*8 Studies on normal human jejunum, however, have shown that only a small proportion (2%-8%) of human intestinal intraepithelial T cells are TcRl+; most of them are CDs- CD4- (double

of Dermatology,

DANIELA

University

AGAPE,

of Milan, Italy

negatives),g-” whereas virtually no TcRl+ lymphocytes are observed in the lamina propria. A significant increase in such jejunal intraepithelial lymphorecently observed in celiac cytes has been disease,12s13 suggesting that TcRl+ lymphocytes, although not a prominent population in normal jejunum, may play a pathogenetic role in some human intestinal diseases. Dermatitis herpetiformis is a chronic, pruritic, blistering skin disease of unknown origin. The disease is characterized by the presence of granular deposition of immunoglobulin A (&A), fibrin deposition, and neutrophil aggregates in the subepidermal papillae. The jejunal mucosa of patients with dermatitis herpetiformis shows lesions almost indistinguishable from those observed in celiac disease and consisting of patchy villous atrophy, crypt hyperplasia, and dense mononuclear cell infiltrates in the lamina propria and in the epithelial layer.14 Further similarities between the two diseases include the presence of circulating antibodies to gliadin and to endomysium as well as the response of jejunal lesions to a gluten-free diet.15*‘” To further evaluate the relationships existing between these two diseases and the role of TcRl+ intraepithelial lymphocytes in human intestinal diseases, we studied, by means of a double-step immunofluorescence method, the phenotype and distribution of intraepithelial lymphocytes expressing the TcRl in the jejunal mucosa of patients with dermatitis herpetiformis. Materials and Methods Patients Nineteen patients with dermatitis herpetiformis undergoing duodenal endoscopic and/or jejunal suction biopsy in the work-up of their skin disease were enrolled in the study. The diagnosis of dermatitis herpetiformis was made in all patients by showing typical IgA granular de0 1992by the American Gastroenterological Association oolfi-5995/92/$3.99

1500

VECCHI ET AL.

GASTROENTEROLOGY

posits in the subepidermal papillae by means of immunofluorescence on perilesional skin.“*” Skin specimens from 9 of these patients were also available for the study of lymphocyte subpopulations on lesional and perilesional skin. Sixteen patients undergoing diagnostic upper gastrointestinal endoscopy for dyspepsia (6 subjects) or jejunal biopsy for the work-up of chronic diarrhea (10 subjects) were used as controls. All 6 patients undergoing upper gastrointestinal endoscopy for dyspepsia gave informed consent to have two biopsy specimens taken from the descending duodenum. In the 10 subjects studied for chronic diarrhea and used as controls, a diagnosis of celiac disease was excluded on the basis of clinical features, results of absorption studies, and jejunal morphology. Tissue Specimens At least two duodenal mucosal specimens were obtained from the descending duodenum in the course of upper gastrointestinal endoscopy. Jejunal mucosal specimens were obtained by means of suction biopsy at the ligament of Treitz using a directable catheter under fluoroscopic control. With this method, two mucosal specimens can usually be obtained. One of the two specimens obtained by means of endoscopy or suction biopsy was fixed in 10% buffered formalin and used for routine histology; the other specimen was immediately embedded in OCT compound (Miles, Elkhart, IN) and frozen in liquid nitrogen. The specimens were kept at -2O’C until the time of the experiments for the study of lymphocyte subpopulations. The same procedure was used for skin biopsy specimens. Antibodies The Ig isotypes, the antigen specificities, and the lymphocyte populations recognized by the mouse monoclonal antibodies used are reported in Table 1. The monoclonal antibody TCRB-1 was purchased from T Cell Science (Cambridge, MA). Monoclonal antibodies Al3 and BB3”‘*lgwere a kind gift of Drs. L. Moretta and S. Ferrini from the Istituto Nazionale dei Tumori, Genoa, Italy. The monoclonal antibody TiyAzOwas a kind gift of Dr. T. Hercend from the Institute Gustave-Roussy, VilleTable

1. Characteristics

Monoclonal antibody

of the MonocJonaJ

Antibodies

Vol. 102, No. 5

juif, France. The monoclonal antibody Ber-ActEt2*was kindly provided by Dr. H. Stein from the Pathology Institute, University of Berlin, Germany. The monoclonal antibody to cluster differentiation antigen 3 (CD3, clone MEM 57) was obtained as undiluted ascites from the IV International Workshop on Human Leukocyte Differentiation Antigens. ‘* The monoclonal antibody to CD8 (OKT8) was obtained from Ortho (Raritan, NJ). The monoclonal antibody anti-human collagen IV was purchased from ICN Immunobiologicals (Lisle, Israel). Secondary antibodies to mouse IgG subclasses were purchased from Southern Biotechnology (Birmingham, AL) and adsorbed by mixing them with human AB serum before each experiment. lmmunofluorescence Five-micrometer-thick cryostat sections were obtained from OCT-embedded jejunal mucosa or skin specimens and dried at room temperature overnight. The sections were fixed in acetone for 10 minutes, dried for 30 minutes at room temperature, washed three times in Trisbuffered saline (TBS), pH 7.6, and incubated in diluted normal goat and human serum to reduce unspecific binding of the reagents. Sections were then exposed to the primary mouse monoclonal antibody (TCRd-1, BB3, A13, Ber-Act8, or Ti-yA, IgGl subclass) for 60 minutes at room temperature in a humid chamber. To obtain a good recognition of the epithelial cell layers, an anti-human collagen IV mouse monoclonal antibody (IgGl subclass) was coincubated together with the specific anti-TcR antibody. After three washes in TBS, the sections were incubated with fluorescein-conjugated anti-mouse IgGl goat IgG fraction for 60 minutes at room temperature in a humid chamber and washed three times in TBS. Sections were then incubated with anti-CD3 or anti-CD8 monoclonal antibody (IgG2a subclass) for 60 minutes at room temperature, washed in TBS, and incubated in rhodamine-conjugated anti-mouse IgGPa goat IgG fraction for 60 minutes. Sections were then washed in TBS, mounted in glycerin, and coded. The sections were examined blindly for immunoreactivity under a Zeiss epifluorescence microscope (Jena, Germany). The expression of TcR y/b-related antigens was looked for in each CD3+ cell. The localization (intraepithelial vs. lamina propria, epidermal or dermal) of T lympho-

Used

Mouse Ig isotype

Antigen specificity

TCR-6-l A13

IgG1 IgGl

6 Chain Undetermined

BB3

IgGl

V62

TiyA

IgGl

VP

Ber-Act8

I@

Undetermined

MEM 57 OKT8

IgG2a IgG2a

CD3 CD8

Lymphocyte population recognized All y/s lymphocytes VylVSl non-disulfide-linked (Cy2) TcR1 + lymphocytes VflV62 disulfide-linked (Cyl) TcR1 + lymphocytes Vy~V62 disulfide-linked (Cyl) TcRl + lymphocytes Intestinal intraepithelialactivated CD8 + lymphocytes All mature T lymphocytes All CD8 + lymphocytes

May 1992

JEJUNAL y/6 LYMPHOCYTES IN DERMATITIS HERPETIFORMIS

cytes observed was made possible by the presence of collagen IV reactivity. Statistical Analysis To compare the results obtained in patients with dermatitis herpetiformis and in controls, a Wilcoxon’s two-tailed test was used. Results

The vast majority of CD3+ lymphocytes reactive to the monoclonal antibody TcR-6-l was seen among intraepithelial lymphocytes (Figure 1). The same finding was observed when monoclonal antibodies Al3 and BB3 were tested. Only scattered (O%-3%) CD3+ lymphocytes in the lamina propria showed reactivity with TcR-S-1, Al3, or BB3 monoclonal antibodies, both in dermatitis herpetiformis and in controls. In jejunal specimens from patients with dermatitis herpetiformis, TcR-6-l+ cells ranged from 16% to 91% of CD3+ intraepithelial lymphocytes, with a mean of 40.1% (Figure 2; Table 2). This figure is significantly different from that observed in control specimens (P < 0.001) (Figure 3). Reactivity to Al3 monoclonal antibody was observed in 23.4% (range, 4.1%-47.4%) of CD3+ intraepithelial lymphocytes in the specimens from patients with dermatitis herpetiformis compared with 7.2% (range, l.lOb-27%) in controls (P < 0.001). The corresponding numbers observed with the BB3 monoclonal antibody were 10.4% (range, 1.9%-40%)

Figure 1, Jejune1 specimen from a patient with dermatitis herpetiformis with partial villous atrophy (immunofiuorescence; original magnification X40).Several intraepithelial lymphocytea reactive to the TcR-S-1 monoclone1 antibody are seen. The reactivity of anti-collagen IV monoclonal antibody identifies the basal membrane.

1501

and 7.9% (range, 2.3%-29.5Oh), respectively (NS). The results of these experiments are summarized in Table 2. When patients with dermatitis herpetiformis were divided according to the degree of jejunal damage, the increase in TcR-6-l+ and Al3+ intraepithelial lymphocytes did not significantly differ in patients with subtotal villous atrophy compared with patients with partial villous atrophy or patients on gluten-free diet with normal villi (Table 2). The sections showing a high number of lymphocytes reactive to TcR-6-l monoclonal antibody were also analyzed to assess the number of y/6+ cells that were also CD8+. CD8 reactivity was observed in 10.9% (range, O%-25%) of TcRl+ lymphocytes in dermatitis herpetiformis (Figure 4) and in 7.9% (range, 0%-27%) of TcRl+ lymphocytes in controls. These values were not statistically different. CDs+ intraepithelial lymphocytes were significantly increased in jejunal specimens from patients with dermatitis herpetiformis (40.0 -t 11.7/mm mucosa) compared with controls (13.5 -t 11.5/mm mucosa) (P < 0.01). Intraepithelial lymphocytes reactive to Ber-Act8 monoclonal antibody were significantly more frequent in the jejunal mucosa of patients with dermatitis herpetiformis (35.5 f 7.5/mm mucosa) than in controls (10.0 + 4.9/mm mucosa) (P < 0.001). In the skin, no TcR-6-l+ lymphocytes were observed in the epidermal layers. Only scattered peri-

1502 VECCHI ET AL.

Figure 3. Jejunal specimen from a patient with dermatitis herpetiformis with partial villous atrophy (double immunofluorescence; original magnification X60). CD3+ TcR-S-l+ lymphocytes are seen in yellow. CD3+ TcR-S-l- lymphocytes are seen in red. TcR-S-l+ cells represent a major population of CD3+ intraepithelial lymphocytes.

vascular lymphocytes bearing the TcRl were observed in perilesional and lesional skin. Discussion

On the basis of their tissue distribution and functional characterization in the mouse,5’6 y/6

GASTROENTEROLOGY

Vol. 102, No. 5

Figure 3. Jejunal specimen from a control subject (immunofluorescence; original magnification X46). One TcR-S-l+ lymphocyte is observed in the intraepithelial compartment.

(TcRl)+ T cells have been proposed to play an important role in intestinal mucosal immunity.7g* Several studies on human tissues, however, ha ve shown that TcR1-t lymphocytes represent only ’ a minor proportion of intraepithelial lymphocytes in

Figure 4. Jejunal specimen from a patient with dermatitis herpetiformis with partial villous atrophy (double immunofluorescence; original magniftcation X46). CD6+ lymphocytes are seen in green and TcR-S-l+ lymphocytes in red. Yellow cells are TcR-&l+ Cd6+.

JEJUNAL y/6 LYMPHOCYTES IN DERMATITIS HERPETIFORMIS

May 1992

1503

of CD3f Intraepithelial Lymphocytes in the Jejunal Mucosa of Patients With Dermatitis Herpetiformis and Controls Tested With Monoclonal Antibodies TCR-61,A13, and BB3

Table 2. Reactivity Reactivity to monoclonal antibodies TCRB-l+ Al3+ BB3+

DH total (n = 19) 40.1+ 17.9O (16.1-91.0) 23.4+ 10.0" (4.1-47.4) 10.4f 9.0" (1.9-40.0)

DH + SVA (n = 7)

DH + PVA (n = 9)

DH normal villi (n = 3)

39.7+ 12.v (23.3-63.3) 27.7f ll.la (14.8-47.4) 10.1+_6.3" (3.2-20.0)

44.8+_22.4' (16.1-91.0) 21.7+ 9.2b (4.1-35.5) 11.8+ 11.7= (2.8-40.0)

26.9+ 6.0 (20.0-31.0) la.1 + a.4 (9.4-26.1) 6.9+ 6.4 (1.9-14.2)

Controls (n = 16)

10.8rtr6.9 (2-O-23.3) 7.2+ 6.9 (1.1-27.0) 7.9+ 6.7 (2.3-29.5)

NOTE. Results are expressed as mean + SD. Ranges are reported in parentheses. DH, Dermatitis herpetiformis; SVA, subtotal villous atrophy; PVA, partial villous atrophy. “P < 0.001; bP < 0.01; “NS, vs. controls.

human intestinal tract, ranging from 2% to 10% of intraepithelial CD3+ cells.9-** In the present study, the prevalence of TcRl+ intraepithelial lymphocytes in control jejunal tissue specimens ranged from 2% to 22% (mean, 10.8%), in agreement with previously reported figures.“” Our data confirm that, despite the high variability observed in the prevalence of TcRl+ cells in normal subjects, lymphocytes expressing the TcRl represent only a minor population of intraepithelial lymphocytes in normal human jejunal mucosa. Our data also confirm that lamina propria lymphocytes very seldom express the TCRl phenotype. To further investigate the phenotype of TcRl+ lymphocytes in intestinal mucosa, jejunal sections were tested by immunohistochemistry using the monoclonal antibodies A13, BB3, and TiyA. The Al3 antibody recognizes a common monomorphic determinant present on the non-disulfide-linked form of y/6 receptor and possibly identifies the TcRl subset expressing the Vrl/V81 gene rearrangement.‘* On the other hand, BB3 is thought to recognize products of the Vy8/V62 gene rearrangement present in a subpopulation characterized by the disulfide-linked form of TcRl.” TiyA monoclonal antibody has been developed against the Vy9 gene product and has been reported to react with most peripheral blood lymphocytes,” with a similar pattern of reactivity as BB3. By means of these monoclonal antibodies, it was possible to show that the intraepithelial population of TcRl+ lymphocytes differs to some extent from the circulating one. In fact, Al3 reactivity, which is detected in around 30% of TcRl+ peripheral blood lymphocytes, was present in 50%-70% of TcRl+ intraepithelial lymphocytes. These numbers are in agreement with the data obtained by using the monoclonal antibody 6-TCS-1,*3which is thought to react with a subpopulation of TcRl+ cells expressing the non-disulfide-linked form of receptor similar to that

recognized by Al3 monoclonal antibody. Furthermore, whereas most circulating TcRl+ cells have been shown to react with TiyA monoclonal antibody,” only a minor proportion of TcRl+ lymphocytes in jejunal epithelium were reactive in our experiments (data not shown). These differences observed by us and others between the intraepithelial and the circulating compartments of TcRl+ lymphocytes might be interpreted on the basis of recent data obtained in mice. These studies suggest in fact that the intestinal epithelium may have the property of exerting a positive selection, possibly by class II major histocompatibility complex mo1ecules,23 of specific TcR gene rearrangements on extrathymically derived lymphocytes such as TcRl+ intraepithelial lymphocytes.24 Our study also confirms that only around 10% of TcRl+ intraepithelial lymphocytes bear the CD8 antigen, this antigen being expressed only by a proportion of A13+ cells.9-” When jejunal specimens from patients with dermatitis herpetiformis were studied, an increase in TcRl+, as expressed by TcR-6-l reactivity, as well as in CD3+, CD8+ lymphocytes was observed. The mean prevalence of TcRl+ intraepithelial lymphocytes was 40% in the jejunum of patients with dermatitis herpetiformis with a highly significant difference (P < 0.001) compared with controls. These findings are in agreement with those reported by Halstensen et al.” and Spencer et a1.13in celiac disease and provide further support to the hypothesis that similar pathogenetic mechanisms play a role in the jejunal lesions of celiac disease and dermatitis herpetiformis. In our patients, the prevalence of TcRl+ intraepithelial lymphocytes did not appear to be related to the degree of jejunal involvement, because similar figures were observed in patients with partial villous atrophy and patients with total villous atrophy. Moreover, three patients who were on a gluten-free

1504

VECCHI ET AL.

diet at the moment of biopsy and whose specimens showed normal jejunal architecture also showed increased intraepithelial TCRl+ lymphocytes. This is in agreement with the previous report of increased TcRl+ lymphocytes in treated celiac patients, even with normal jejunal morphology.25 Interestingly, the analysis of Al3 and BB3 subsets of TcRl+ cells showed that only A13+ lymphocytes were significantly increased in patients with dermatitis herpetiformis (P < 0.001 vs. controls), whereas no significant difference was observed in BB3+ cells. This finding is similar to what has been reported in celiac disease, where non-disulfide-linked forms of TcRl lymphocytes account for the increase in TcRl+ cells.13 The increase in A13+ lymphocytes appeared to be independent from the degree of jejunal involvement, similarly to the observations made by us and others about total TcRl+ cells. Similarly to what observed in normal jejunal mucosa, only few TcRl+ cells expressing the phenotype recognized by TiyA monoclonal antibody were observed in jejunal specimens from patients with dermatitis herpetiformis (data not shown). When TcRl+ lymphocytes were analyzed according to the expression of CD8 molecules, around 90% of them resulted to be CD8-, regardless of patient’s diagnosis. This finding is in agreement with that of Halstensen et al.” Furthermore, a significant increase in CD8+, Ber-Act8+ intraepithelial T lymphocytes was observed in the specimens from patients with dermatitis herpetiformis, very few of which expressed the TcR-6-l phenotype. No cells showing Ber-Act8 reactivity were detected in the lamina propria infiltrate. In conclusion, our data confirm that TcRl+ cells represent a minor population of CD3t lymphocytes in the epithelial layer and, even less, in the lamina propria of normal human jejunal mucosa. A highly significant increase in TcRl+ cell subpopulations, which mostly react with the Al3 monoclonal antibody detecting the non-disulfide-linked form of TcRl and do not express the CD8 antigen, can be observed in the intraepithelial compartment in the course of dermatitis herpetiformis, similarly to what has been reported for celiac disease. These findings suggest that similar pathogenetic mechanisms may play a role in determining the jejunal damage in these two diseases. References 1. Hedrick SM, Germain RN, Bevan MJ, Dorf M, Engel I, Fink P, Gascoigne N, Heber-Katz E, Kapp J, Kaufmann Y, Kaye J, Melchers F, Pierce C, Schwartz RH, Sorensen C, Taniguchi M, Davis MM. Rearrangement and transcription of a T cell receptor beta-chain in different T cell subsets. Proc Nat1 Acad Sci USA 1985;82:531-535.

GASTROENTEROLOGY

Vol. 10.2, No. 5

2. Marrack P, Kappler J. The antigen-specific, major histocompatibility complex-restricted receptor on T cells. Adv lmmunol 1986;38:1-30. 3. Brenner MB, McLean J, Dialynas DP, Strominger JL, Smith JA, Owen FL, Seidman JG, Stephen IP, Rosen F, Krangel MS. Identification of a putative second T-cell receptor. Nature 1986;322:145-149. 4. Brenner MB, McLean J, Scheft H, Riberdy J, Ang SL, Seidman JG, Devlin P, Krangel MS. Two forms of the T-cell receptor protein found on peripheral blood cytotoxic T lymphocytes. Nature 1987;325:689-694. 5. Bonneville M, Janeway CA, Ito K, Haser W, Ishida I, Nakanishi N, Tonegawa S. Intestinal intraepithelial lymphocytes are a distinct set of y/6 T cells. Nature 1988;336:479-481. 6. Goodman T, LeFrancois L. Expression of the y/6 T-cell receptor on intestinal CD8+ intraepithelial lymphocytes. Nature 1988;333:855-858. 7. Janeway CA. Frontiers of the immune system. Nature 1988;333:804-806. 8. Bluestone JA, Matis LA. TCRy/G-cells: minor redundant T cell subset or specialized immune system component? J Immunol 1989;142:1785-1788. 9. Brandtzaeg P, Bosnes V, Halstensen TS, Scott H, Sollid LM, Valnes KT. T lymphocytes in human gut epithelium preferentially express the a/P antigen receptor and are often CD45/ UCHLl positive. Stand J Immunol 1989;30:123-128. 10. Bucy RP, Chen C-LH, Cooper MD. Tissue localization and CD8 accessory molecule expression of Ty/6 cells in humans. J Immunol 1989;142:3045-3049. 11. Groh V, Porcelli S, Fabbi M, Lanier LL, Picker LJ, Anderson T, Warnke RA, Bhan AK, Strominger JL, Brenner MB. Human lymphocytes bearing T cell receptor y/6 are phenotypically diverse and evenly distributed throughout the lymphoid system. J Exp Med 1989;169:1277-1294. 12. Halstensen TS, Scott H, Brandtzaeg P. Intraepithelial T cells of the TcRy/G+ CDs- and V&/J&+ phenotypes are increased in coeliac disease. Stand J Immunol 1989;30:665-672. 13. Spencer J, Isaacson PG, Diss TC, MacDonald TT. Expression of disulfide-linked and non-disulfide-linked forms of the T cell receptor gamma/delta heterodimer in human intestinal intraepithelial lymphocytes. Eur J Immunol1989;19:1335-1338. 14. De Franchis R, Primignani M, Monti M, Cipolla M, Vecchi M, Berti E, Agape D. Small bowel involvement in dermatitis herpetiformis and in linear IgA bullous dermatosis. J Clin Gastroenterol 1983:5:429-436. 15. Hall RP. The pathogenesis of dermatitis herpetiformis: recent advances. J Am Acad Dermatol1987;16:1129-1144. 16. Reunala T, Chorzelski TP, Viander M, Sulej J, Vainio E, Kumar V, Beutner EH. IgA anti-endomysial antibodies in dermatitis herpetiformis: correlation with jejunal morphology, gluten-free diet and anti-ghadin antibodies. Br J Dermatol 1987;117:185-191. 17. Dabrowsky J, Chorlzelsky TP, Jaqblonska S. The ultrastructural localization of IgA in skin of a patient with mixed form of DH and PB. J Invest Dermatol 1978;70:76-79. 18. Ferrini S, Prigione I, Bottino C, Ciccone E, Tambussi G, Mammoliti S, Moretta L, Moretta A. Monoclonal antibodies which react with the T cell receptor y/6 recognize different subsets of CD3+ WT31- T lymphocytes. Eur J Immunol 1989;19:5761. 19. Bottino C, Tambussi G, Ferrini S, Ciccone E, Varese P, Mingardi MC, Moretta L, Moretta A. Two subsets of human T lymphocytes expressing y/6 antigen receptor are identifiable by monoclonal antibodies directed to two distinct molecular forms of the receptor. J Exp Med 1989;168:491-505.

May 1992

JEJUNAL y/6 LYMPHOCYTES IN DERMATITIS HERPETIFORMIS

20. Faure F, Jitsukawa S, Triebel F, Hercend T. CD3/TiyA: a functional r-receptor complex expressed on human peripheral lymphocytes. J 1mmuno11988;140:1372-1379. 21. Schwarting R, Kruschwitz M, Fritzsche G, Stein H. The antigen detected by the new monoclonal antibody Ber-Act 8 defines late activated CD8 positive lymphocytes and is identical to the antigens recognized by monoclonal antibodies HML-1 and B-LY7. Proceedings of the 3rd Meeting of the European Association for Haematopathology. Wurzburg, Germany, October 7-11,1990. 22. Berti E, Cattoretti G, Delia G, Parravicini C, Pinto N, Soligo D. IV International Workshop on human leukocyte differentiation antigens: a report. Vienna, Austria, February 21-25, 1989. Haematologica 1989;74:401-407.

1505

23. LeFrancois

L, LeCorre R, Mayo J, Bluestone JA, Goodman T. Extrathymic selection of TcR -yS+ T cells by class II major histocompatibility complex molecules. Cell 1990;63:333-340. 24. Bandeira A, Itohara S, Bonneville M, Burlen-Defranoux 0, Mota-Santos T, Coutinho A, Tonegawa S. Extrathymic origin of intestinal intraepithelial lymphocytes bearing T-cell antigen receptor $5 Proc Nat1 Acad Sci USA 1991;88:43-47. 25. Viney J, MacDonald TT, Spencer J. Gamma/delta T cells in the gut epithelium. Gut 1990;31:841-844,

Received December 17,199o. Accepted September 18,1991. Address requests for reprints to: Maurizio Vecchi, M.D., Istituto di Medicina Interna, Via Pace 9, 20122 Milan, Italy.