Class II major histocompatibility complex and organ specific autoimmunity in man

Transcriptional regulation of chemokine genes: a link to pancreatic islet inflammation?

Enhanced expression of chemotactic cytokines (aka chemokines) within pancreatic islets likely contributes to islet inflammation by regulating the recruitment and activation of various leukocyte populations, including macrophages, neutrophils, and T-lymphocytes. Because of the powerful actions of these chemokines, precise transcriptional control is required. In this review, we highlight what is known about the signals and mechanisms that govern the transcription of genes encoding specific chemokine proteins in pancreatic islet β-cells, which include contributions from the NF-κB and STAT1 pathways. We further discuss increased chemokine expression in pancreatic islets during autoimmune-mediated and obesity-related development of diabetes.

Autoimmune disease is not a feature of hepatitis C infection in Ireland

To determine the prevalence of autoimmune disease, autoantibody positivity, or both in Irish persons with hepatitis C, we surveyed 98 such patients (55 recipients of anti-D, 25 intravenous drug abusers, and 18 blood transfusion recipients). We studied them clinically and tested for anti-nuclear, anti-smooth muscle, and anti-mitochondrial, liver-kidney microsomal, thyroid microsomal, thyroid globulin, and gastric parietal antibodies; and also for rheumatoid factor. In the anti-D antibody group (all female), two patients reported generalized musculoskeletal symptoms but had no demonstrable physical signs. We did not find cryoglobulins in any patient. We detected thyroid microsomal antibodies in only 6 of 55 (10.9%) patients. (In two of these, thyroid globulin antibodies were also positive). These patients were all clinically euthyroid, but two had borderline low-normal thyroid function tests. Titers for anti-nuclear antibodies were weakly positive in 5 of 55 (9.1%) patients, and gastric parietal antibodies were positive in 5 of 55 (9.1%) patients. In particular, we noted no antibodies to liver-kidney microsome. Rheumatoid factor was detected in eight patients. Forty-seven of 55 patients were genotype 1b, and 8 of 55 were genotype 3. In the intravenous drug abusers (8 women, 17 men), we detected no autoantibodies. Seven of the 25 genotypes were tested; three were genotype 3 and four were genotype 1b. In the transfusion group (10 women, 8 men), we detected no autoantibodies apart from weak anti-nuclear antibody Titers (1:10), which we found three patients. Five of 10 genotypes tested were of genotype 3 and the other five were of genotype 1b. These findings suggest that in Irish patients with hepatitis C, neither genotype nor source (and dose) of inoculum contributes to the development of autoimmune disease. How hepatitis C virus is associated with autoimmune disease in other studies remains unknown. The answer may, at least in part, be found in genetic; HLA typing studies should provide useful information.

Expression of major histocompatibility complex class II molecules by dermal inflammatory cells, epidermal Langerhans cells and keratinocytes in canine dermatological disease

The distribution of Major Histocompatibility Complex (MHC) Class II molecules in skin biopsies from 42 dogs with primary skin disease and six control dogs was examined by indirect immunohistochemical methods. Canine epidermal Langerhans cells constitutively expressed MHC Class II and the number of such cells was increased in the epidermis of dogs with discoid lupus erythematosus (n = 10), hypersensitivity dermatitis (n = 10) or deep pyoderma (n = 13) but not in eight cases of pemphigus foliaceus. Expression of MHC Class II by epidermal and follicular keratinocytes was observed in a ninth dog with pemphigus foliaceus. Within dermal inflammatory infiltrates, Class II was expressed by macrophages, dendritic cells and fibroblasts but not by neutrophils, eosinophils, mast cells or plasma cells. Only occasional dermal lymphocytes expressed surface MHC Class II, but nodular aggregates of strongly Class II-positive lymphocytes were observed within the dermis of two dogs with discoid lupus erythematosus.

Immunologic features and HLA associations in chronic viral hepatitis

BACKGROUND/AIMS

Chronic viral hepatitis may have immunologic manifestations, and such features may reflect genetic predispositions. The aim of this study was to assess associations between immune manifestations and HLA-DR antigens.

METHODS

Ninety-five patients were evaluated prospectively for immunologic features. A microlymphocytotoxicity technique was used to determine DR3, DR4, and A1-B8-DR3 phenotypes. DR antigens were also determined by restriction fragment length polymorphism in 76 patients with chronic viral hepatitis and 80 normal subjects.

RESULTS

Autoantibodies were found in 59 patients (62%), and concurrent immunologic diseases were found in 22 patients (23%). Patients with antinuclear antibodies had the A1-B8-DR3 phenotype more commonly than seronegative counterparts (26% vs. 6%; P = 0.02) and had DR3 positivity more frequently than normal subjects (41% vs. 18%; P = 0.03). In contrast, patients with concurrent immunologic diseases had DR4 positivity more commonly than patients without these findings (68% vs. 27%; P = 0.001) and normal subjects (68% vs. 30%; P = 0.003).

CONCLUSIONS

Patients with chronic viral hepatitis commonly have autoantibodies and/or concurrent immunologic diseases. The expression of antinuclear antibodies is associated with the A1-B8-DR3 phenotype, and the presence of concurrent immunologic diseases is associated with the DR4 phenotype. In these instances, autoimmune expression may reflect a genetic predisposition that is facilitated by viral infection or is coincidental with it.

Genetic predispositions for the immunological features of chronic active hepatitis

To assess the frequency and genetic predispositions of concurrent immunological diseases and immunoserological markers in autoimmune hepatitis and chronic viral hepatitis, we assessed 185 patients prospectively, including 122 patients with autoimmune hepatitis and 63 patients with viral disease. Human leukocyte antigens were determined in all patients. Sixty patients (32%) had concurrent immunological diseases, and the majority of the diseases (68%) had known human leukocyte antigen associations. Although patients with autoimmune hepatitis had concurrent immunological diseases more commonly than those with viral disease (38% vs. 22%; p = 0.04), the nature of the diseases was similar in both groups, as were the frequencies of human leukocyte antigen-DR4 (42% vs. 39%; p = 0.7). The presence of human leukocyte antigen-DR4 was associated with the concurrence of immunological diseases in both autoimmune (62% vs. 33%; p = 0.01) and viral hepatitis (75% vs. 29%; p = 0.009). In autoimmune hepatitis, human leukocyte antigen-DR4 was also associated with the expression of smooth muscle antibodies and high-titer antinuclear antibodies. We conclude that concurrent immunological diseases and immunoserological markers are common in autoimmune and chronic viral hepatitis. Both conditions have a common genetic predisposition for concurrent immunological disease associated with human leukocyte antigen-DR4. The expression of smooth muscle antibodies and high-titer antinuclear antibodies is associated with human leukocyte antigen-DR4 in autoimmune hepatitis only, suggesting that this response is associated with triggering antigens and immune recognition systems that are different from those in viral disease.

Increased epithelial expression of HLA-DQ and HLA-DP molecules in salivary glands from patients with Sjögren's syndrome compared with obstructive sialadenitis

Salivary gland specimens from 10 patients with primary Sjögren's syndrome (pSS) were examined by two-colour immunofluorescence with various combinations of monoclonal and polyclonal antibody reagents of the following specificities: human leucocyte antigen (HLA) class I and II (DR, DP and DQ), CD3, CD45 (leucocyte common antigen), various cytokeratins, and factor VIII-related antigen. Tissue specimens from 10 normal glands and 10 glands with obstructive sialadenitis (no known autoimmunity) served as controls. Only some intercalated ducts and scattered acini of the normal major glands expressed HLA class II determinants (< 5% of total epithelial area); the relative proportion of positive elements indicated differential expression (DR > DP > DQ). SS glands contained substantial T cell infiltrates and increased numbers of activated (DR+) T cells; adjacent epithelium showed extensive differential expression of HLA class II determinants (DR > DP > DQ). Glands with obstructive sialadenitis showed similarly increased epithelial expression of HLA-DR but with surprisingly small amounts of concomitant HLA-DP and -DQ expression. Epithelial HLA class II expression probably depends on cytokines as an inductive event, which is not unique for SS but particularly prominent in this disorder. Our results suggest that epithelial expression of HLA-DP or -DQ, rather than -DR, might be a prerequisite for the autoimmune process of SS to develop in genetically susceptible individuals.

Evidence for urothelial cell activation in interstitial cystitis

Bladder biopsy samples from 17 interstitial cystitis patients and 20 controls were evaluated for urothelial cell activation using a panel of monoclonal antibodies to HLA-DR, intercellular adhesion molecule 1, interleukin 1 alpha and tumor necrosis factor alpha. Urothelial cells in the majority (13 of 16, 81%) of the biopsies from patients with interstitial cystitis showed increased expression of HLA-DR, while fewer samples were positive for intercellular adhesion molecule 1 (3 of 16, 19%), interleukin 1 alpha (2 of 17, 12%) or tumor necrosis factor alpha (1 of 15, 7%). No urothelial cell expression of intercellular adhesion molecule 1, interleukin 1 alpha or tumor necrosis factor alpha was detected in the controls, and only 1 of 20 control samples contained HLA-DR positive urothelial cells. These results suggest that an unusual type of cellular activation is present in interstitial cystitis. In vitro studies with cultured normal urothelial cells indicated that cells activated with gamma interferon and tumor necrosis factor alpha expressed intercellular adhesion molecule 1 and HLA-DR, although increases in intercellular adhesion molecule 1 expression occurred earlier. Urothelial cells in interstitial cystitis patients may be defective in ability to express intercellular adhesion molecule 1. Alternatively, the differential expression of HLA-DR and intercellular adhesion molecule 1 in interstitial cystitis specimens may represent a functional subset of interstitial cystitis or reflect different stages of the disease. Urothelial cell activation in interstitial cystitis may result in aberrant immune responses and immune activation within the bladder. Because HLA-DR can be detected in paraffin-embedded tissues, evaluation of urothelial cell HLA-DR expression, although not specific for interstitial cystitis, may become a useful tool in the pathological evaluation of biopsy tissues from patients with this disease.

Expression of MHC class II antigens by conjunctival epithelial cells in trachoma: implications concerning the pathogenesis of blinding disease

In an attempt to understand the pathogenesis of corneal changes and conjunctival scarring in trachoma, which are responsible for blindness in some 7 million people worldwide, the surface expression of major histocompatibility complex class II antigens on conjunctival epithelial cells was sought in patients with different clinical features of the disease. Class II expression was significantly associated with active (inflammatory) trachoma, but an independent association with corneal pannus or conjunctival scarring could not be shown. Longitudinal data suggest that class II expression may antedate clinically active disease and persist after it has resolved.

Class II antigen expression in peripheral neuropathies

The expression of class II antigen was studied in sural nerve biopsies from patients with peripheral neuropathies. These included patients with chronic demyelinating polyradiculoneuropathy (CIDP), non-immune mediated neuropathies of diverse etiologies and controls without evidence of neuropathy. The major finding in CIDP was a marked increase in class II expression on Schwann cells. Endoneurial Schwann cell staining to the same degree as in CIDP was seen in diabetic symmetric proximal motor neuropathy, neuropathies associated with monoclonal gammopathies and hereditary sensory and autonomic neuropathy type 1. In the control nerves and the other non-immune mediated neuropathies class II expression was mainly restricted to endothelial and perineurial cells. Increased endoneurial expression of class II antigen was found to correlate with elevated cerebrospinal fluid (CSF) protein levels but not with other clinical variables or demyelination as defined by electrophysiologic criteria or teased fiber analysis. The increased expression of class II antigen on Schwann cells may be indicative of a breakdown in immunological tolerance but should not be used as a diagnostic marker for dysimmune neuropathies due to overlap with non-immune mediated neuropathies.

Lymphoid cell subsets in colonic mucosa and HLA-DR antigens on colonic epithelia in colitis excluding ulcerative colitis and Crohn's disease

Lymphoid cell subsets, including T cells as well as Ig-containing cells in the colonic mucosa and HLA-DR antigens on colonic epithelia, were examined in non-IBD colitis (colitis excluding ulcerative colitis (UC) and Crohn's disease) by the indirect immunoperoxidase staining method. Mouse anti-CD5, CD8, CD4, IgG, IgA1, IgA2, IgM, IgD, IgE, HLA-DR, and NuIa monoclonal antibodies were used as the first antibody. The results were compared to those of the normal controls and UC. T cell subsets in non-IBD colitis were almost similar to those of the controls and UC. The number of Ig-containing cells of all classes, except for IgA, tended to be increased in non-IBD colitis. In particular, both IgG- and IgE-containing cells were significantly increased compared to those in the controls. Compared to UC, IgG-containing cells were decreased in non-IBD colitis. Namely, in non-IBD colitis, as well as in UC, the change of Ig-containing cells (B cell lineage) was more pronounced than that of T cells. The frequency of the expression of HLA-DR antigens on colonic epithelia in non-IBD colitis was 70%, which was significantly higher than that in controls (0%), but significantly lower than that in UC (100%). Whether the differences in the number of IgG-containing cells, and the frequency of epithelial HLA-DR expression between non-IBD colitis and UC was due to the differences of the degree of local inflammation or due to the differences of the nature of the two diseases was not elucidated in this study.

Class II (HLA-DR, -DP, and -DO) antigens on intestinal epithelia in ulcerative colitis, Crohn's disease, colorectal cancer and normal small intestine

Class II antigens in involved colonic epithelia of ulcerative colitis (UC), Crohn's disease (CD), colorectal cancer, and epithelia of normal small intestine were investigated using an immunoperoxidase method. Ten normal colonic mucosa served as normal controls. Ten specimens were studied for each group. Normal colonic epithelia did not express class II antigens. In colonic diseases; (1) with regard to the frequency and extent of class II antigens on epithelia. HLA-DR antigens were the most highly and greatly expressed, followed by HLA-DP and then HLA-DQ antigens. (2) The extent of HLA-DR and -DP, but not HLA-DQ, expression on epithelia of UC and colorectal cancer seemed to be positively correlated with the degree of mononuclear cell infiltration in the lesion. (3) The extent of class II antigen expression on colonic epithelia in CD was not related to the degree of mononuclear cell infiltration in the lesion. (4) The extent of HLA-DR and -DP expression on epithelia of colorectal cancer seemed to be positively correlated with increasing undifferentiation. In the normal small intestine, where HLA-DR antigens were physiologically expressed on epithelia. HLA-DQ antigens were not expressed. The heterogeneity among colonic diseases, and between the physiological and pathological status, may reflect differences in immunoregulation.

HLA-DR antigen expression in macroscopically uninvolved areas of intestinal epithelia in Crohn's disease

HLA-DR antigens in small and large intestinal epithelia were examined in Crohn's disease (CD). Seventy-two biopsy specimens (10 from the duodenum, 18 from the terminal ileum, and 44 from various sites of the large intestine) were obtained from 5 patients with CD chronologically; before treatment, during elemental diet (ED) treatment, and during home elemental enteral hyperalimentation (HEEH). HLA-DR antigens on the intestinal epithelia were identified by the indirect immunoperoxidase staining method using two mouse anti-HLA-DR monoclonal antibodies. In the small intestine, HLA-DR antigens were expressed differently from normal controls in macroscopically or microscopically pathological areas, but not in macro- or microscopically normal areas. In the large intestine, when the disease was in full vigor before treatment, HLA-DR antigens were always aberrantly expressed even in the macroscopically uninvolved areas, as well as in macroscopically involved areas. However, when the disease was suppressed after ED followed by HEEH treatment, the antigens were not always expressed in the macroscopically uninvolved areas, or in the macroscopically involved areas. These findings suggest that HLA-DR antigens are not merely expressed as a non-specific response to inflammation, but may be expressed as a primary change which results in macroscopic lesions. The possible role of HLA-DR antigens in the immunological mechanism in CD is proposed.

Class I and class II HLA antigen expression by transitional cell carcinoma of the bladder: correlation with T-cell infiltration and BCG treatment

HLA class I and II glycoproteins from transitional cell carcinoma (TCC) and from perineoplastic and healthy vesical mucosa were characterized together with infiltrating cells by means of immunochemistry using specific monoclonal antibodies on frozen sections obtained during resection or radical cystectomy. Specimens were taken from 11 patients with TCC and five with healthy bladder mucosa. Four patients with TCC and four with healthy mucosa had been previously treated with a course of intravesical bacillus Calmette-Guerin (BCG). Ten out of 11 TCC samples expressed class I glycoproteins with a membrane pattern (diffuse in seven, focal in three) as normal epithelial cells from either controls or perineoplastic bladder. Interestingly, eight out of 11 TCC samples expressed class II antigens on their membrane that were also present in six cases in the perineoplastic tissue while the epithelial cells from four out of five normal bladders were completely negative. The epithelial display of class II antigens in the non-neoplastic areas and in the normal bladder correlates (p less than 0.001) with the degree of cellular infiltrate while such a relationship was not found between the HLA II expression of neoplastic cells and the infiltrate. BCG treatment was associated with a higher amount of inflammatory cells, prevalently T "activated" cells (CD5+,DR+), with a CD4/CD8 ratio always greater than 1. In the light of the role played by HLA glycoproteins in immune mechanisms, these results could help explain the positive action of BCG and the relative immunosensitivity of TCC.

Immunopathology of the placenta in pemphigoid gestationis and linear IgA disease

We have investigated the immunopathology of the placenta in bullous diseases by studying the deposition of immune complexes and expression of MHC class II subregion products by immunohistological methods. Placentae from seven patients with pemphigoid gestationis (PG) and two patients with linear IgA disease were studied. In PG immune complexes containing IgGI and C3 were identified in six cases. In linear IgA disease IgAI containing immune complexes were found in both cases. Placentae from patients with PG showed aberrant expression of MHC Class II products. This was not seen in the placentae from patients with linear IgA disease. In PG there was incoordinate expression of the subregion products, DP and DR being more extensively and consistently expressed than DQ. These results and previous immunogenetic studies suggest that PG may be unique among organ specific autoimmune disease, the autoantibodies forming during an allogenic response rather than target cells behaving as antigen presenting cells.

Expression of HLA-DQ, -DR, and -DP antigens in normal kidney and glomerulonephritis

Expression of the defined subtypes of HLA-class II antigens DQ, DR, DP, as well as of a putatively new HLA-class II determinant DY was evaluated with specific monoclonal antibodies on frozen sections of 15 normal kidneys, as well as of renal tissue of 65 patients with different forms of glomerulonephritis (GN). In normal kidney HLA-DR and/or -DY versus DQ or DP antigens were shown to be differentially expressed on subpopulations of glomerular and interstitial cells, as well as vascular endothelia. Normal proximal tubular epithelia lacked HLA-DQ and -DP antigens, but carried -DY and variably -DR products constitutively. In comparison, aberrant presence of HLA-DQ and/or -DP antigens was found on proximal tubular cells in the majority of patients with rapidly progressive (RPGN), membranoproliferative GN (MPGN), or focal glomerular sclerosis (FGS), but more rarely observed in other forms of proliferative or non-proliferative GN. In addition all cases with RPGN revealed reduction of HLA-DQ, -DR, -DP or -DY+ glomerular cells. Decline of HLA-DP and/or -DR+ glomerular cells was variably seen in mesangioproliferative glomerulonephritis (MesPGN) and MPGN, whereas in FGS HLA-DQ antigens appeared to be increased in glomeruli. HLA-DQ, -DR, -DY+ interstitial cellular infiltrates were present in RPGN, FGS and MPGN and only occasionally occurred in other forms of GN. Altered renal expression of HLA-class II antigens may indicate specific sites of immunologically-mediated kidney injuries in GN.

Molecular mechanisms of autoimmunity

Autoimmune diseases result from a combination of genetic susceptibility factors and exogenous influences such as infection or chemical (including drug) exposure. Germline DNA variations in genetic type as well as defects in antigen recognition acquired during thymic education of developing T-lymphocytes both contribute to impaired self: nonself discrimination and set the stage for later development of such diseases as myasthenia gravis, polymyositis, or systemic lupus erythematosus. In addition, drugs such as D-penicillamine, hydralazine, procainamide, or quinidine induce T-cell or B-cell changes which precipitate auto-reactivity and cause drug-induced disease. Intervention in autoimmune diseases with prednisone, alkylating agents or the future use of more selective monoclonal antibody reagents may be life-saving in some of these disorders.

Interferon-gamma in vivo. Induction and loss of class II MHC antigens and immature myelomonocytic cells in rat organs

The effects of recombinant rat interferon-gamma on class II major histocompatibility complex antigen expression in vivo were studied by immunohistology in LEW rats after continuous intravenous infusion for three days. Interferon-gamma administration led to a systemic induction of class II molecules in previously negative parenchymal and stromal cells. The induction patterns observed were highly reproducible, but not closely dose dependent within a 25-fold dose difference tested. However, the effect of interferon infusion differed profoundly in individual cell types, and appeared to be related to the differentiation stage of each cell population. Thus, epithelial cells like duct epithelia, urothelium or basal ear skin keratinocytes as well as endothelia in big vessels were strongly and easily induced for class II antigen expression. Parenchymal cells like cardiomyocytes and hepatocytes showed intermediate reactivity, while capillary endothelia, neurons in the brain, straight proximal kidney tubules or endocrine pancreatic islet cells did not express class II antigens. The induced expression was rapidly lost from most cells within one or two days after interferon withdrawal; the only exception occurred in keratinocytes. Long-term alterations were, however, still found 14 days after infusion. Interstitial class II-positive dendritic-shaped cells were increased in the organs and hepatic Kupffer cells carried class II antigens. On conventional histology all organs appeared perfectly normal at this date. After three days of interferon, cells of an immature myelomonocytic phenotype occluded medium-sized and small veins in all organs and occurred in granuloma-like lesions in the liver. Although these cells quickly disappeared after interferon withdrawal they might have been at least partially responsible for single deaths on day three. Our study provides a basis for testing the immunological in vivo function of parenchymal class II antigen expression and its differentiation-specific regulation.

Skeletal muscle expression of class II histocompatibility antigens (HLA-DR) in polymyositis and other muscle disorders with an inflammatory infiltrate

Muscle biopsy specimens with a pronounced inflammatory component were evaluated for myocyte expression of class II histocompatibility antigens (HLA-DR) by immunohistochemical techniques. All 15 cases of polymyositis were positive; six cases of muscular dystrophy (two Duchenne, four facio-scapulo-humeral), and one case of acute denervation (motor neurone disease) were negative, despite having a comparably intense mononuclear infiltrate. Twelve entirely normal biopsy specimens were also negative for HLA-DR expression. Expression of this membrane glycoprotein may have a clinically important aetiological role in polymyositis, and demonstration of its presence may prove useful as a marker for this disorder in selected problematic biopsy specimens.

Microenvironment of thymic myoid cells in myasthenia gravis

The microenvironment of myoid cells (MyCs) was studied in myasthenia gravis (MG) thymitis with lymphoid follicular hyperplasia (LFH) (nine cases) and with diffuse B cell infiltration (one case), and compared with findings in the thymuses of non-myasthenic control subjects (ten cases). Double immunostaining was used to demonstrate MyCs labelled by anti-desmin together with other thymic components such as keratin-positive epithelial cells, Ki-M 1-positive interdigitating reticulum cells (IDCs), Ki-M 4-positive follicular dendritic reticulum cells, Ki-M 6-positive macrophages, CD22-positive B-cells, CD1-positive cells, CD3-positive T-cells or HLA-DR-positive cells. Round or elongated MyCs were confined to the thymic medulla and were surrounded by CD3-positive T-cells and CD22-positive B-cells. In MG thymitis MyCs were localized in the vicinity of, but not inside germinal centres (GCs). MyCs were always HLA-DR-negative, but were invariably embedded in a cellular micromilieu with strong HLA-DR expression. A remarkable feature of MG thymitis was that the great majority of MyCs were in intimate contact with intramedullary IDCs. Morphometric studies confirmed that such contacts were significantly less frequent in thymuses from non-myasthenic subjects. This indicates that an IDC-dependent antigen-presenting process for T-cells may actively involve MyCs in MG thymitis.

Residual insulin positivity and pancreatic atrophy in relation to duration of chronic type 1 (insulin-dependent) diabetes mellitus and microangiopathy

The relationship of residual insulin positivity in chronic Type 1 (insulin-dependent) diabetes and atrophy of the exocrine pancreas to duration of diabetes, age at onset and microangiopathy was studied in 26 patients (disease duration 2 to 54 years, mean 26 years). Islets containing insulin cells were found in 13/26 pancreata. In 5/13 pancreata insulin positive cells were detected in only one lobule, while in 8/13 insulin positivity was multifocal. All patients with diabetes duration less than 11 years had residual insulin cells; whereas, the rate of insulin positivity was near 40% with diabetes duration of more than 11 and 21 years, respectively. Survival of insulin cells was not clearly related to age at onset. HLA-DR expression on insulin cells was seen in one case. Insulitis was lacking. Pancreatic volume determined in 18 patients ranged from 14-110 ml (age adjusted mean 56.3 ml) and was significantly less than that of control subjects (age adjusted, mean 89.9 ml, p less than 0.0001). Computerized morphometry of the exocrine pancreas revealed severe acinar atrophy due to a reduction in size of acinar cells. Acinar atrophy correlated neither with the degree of insulin positivity, disease duration nor severity of microangiopathy. The findings suggest that in about 40% of patients with Type 1 diabetes a small population of insulin cells may escape autoimmune destruction, irrespective of disease duration or age at onset. Though exocrine atrophy and insulin deficiency are associated, the variable extent of pancreatic atrophy could not be related to such factors as amount of surviving insulin cells, duration of diabetes or microangiopathy.