Simultaneous Occurrence of Autoimmune Enteropathy and Recurrent Deep Venous Thrombosis

Abstract

Autoimmune enteropathy (AIE) is a immunologically mediated form of intractable diarrhea that is usually present after the first month of life (1,2). This pediatric intestinal disorder is characterized by protracted diarrhea, crypt hyperplastic villous atrophy and specific enterocyte autoantibodies. In these patients organ-and non–organ-specific autoantibodies may be detected (3–6). The autoantibodies that are generally of IgG type are directed against gut epithelial components and are detectable in the serum (1,7,8). Typical reactivities of these autoantibodies include the cytoplasm and/or the brush border of the enterocyte (1,2,7). In some patients anti-goblet cell antibodies are also present (6). It is, however, not clear whether these antibodies are causative or represent only markers of epithelial cell destruction. The pathogenesis of autoimmune enteropathy is incompletely understood, but in contrast to early reports, there are clear links to immunodeficiency (5,9–13). The course of AIE may be severe and refractory to therapy, and the affected patients may require lifelong immunosuppression therapy (14–17). Autoimmune enteropathy may be associated with several autoimmune disorders, such as hemolytic anemia, type I diabetes, systemic vasculitis, eczema, atrophic gastritis, hypothyroidism, common variable hypogammaglobulinemia, interstitial nephritis, and colitis. Young infants are most frequently affected, but the occurrence of AIE has also been reported in adults (18,19). In this article, we describe a teenaged patient, with disease that fulfilled all the criteria for the diagnosis of AIE and who had simultaneous development of recurrent episodes of deep venous thrombosis of the leg. CASE REPORT A 16-year-old boy, blood group O, had suddenly increasing diarrhea with no other major accompanying gastroenterologic symptoms. Minor mesogastric pain, fatigue, and anorexia were also present. There was no history of alcohol, drug, cigarette, or other substance abuse, and no gastroenterologic diseases were reported in the family history. Before the beginning of the disease the patient's body weight was 75 to 80 kg (90–97th percentile) with a height of 180 cm (60th percentile) according to Reinken (20). During the disease, stool frequency increased to up to 8 to 10 times a day. The stool had a pulpy-to-watery consistency without blood deposits. Within a short time a severe malabsorption syndrome developed with hypoproteinemia and hypoalbuminemia (albumin 2.40 g/dl). A small bowel biopsy specimen showed a subtotal atrophy of the villi with flat ulcerations and erosions of the mucosa along with crypt hyperplasia. In serology, a high titer of anti-gliadin antibodies both of IgA class (100 U/l) and IgG class (161 U/l) was found (normal level, <25 U/l); however, neither anti-endomysium nor anti-reticulin antibodies were detectable. At the same time an acute pain developed in the right lower extremity. By means of sonography a deep venous thrombosis of the vena iliaca and vena femoralis was diagnosed. A thrombectomy was performed, and anticoagulant therapy was initiated. The HLA typing indicated one DQ-2/DR-7–positive locus, with DR-5/DR-3 negative. The concentrations in the serum of calcitonin, vasoactive intestinal peptide, serotonin, parathormone, gastrin, glucagon, and complement factors (C3 and C4) were normal. The hepatitis and human immunodeficiency virus serology, antinuclear antibodies, and dsDNA-antibodies were negative. Microbiologic and virologic investigation of stool did not show parasites, viruses, or bacteria. Colonoscopy revealed no abnormalities. An abdominal computed tomographic scan showed normal structure of the abdominal organs and the pancreas. Under the suspicion of gluten-sensitive enteropathy and because of a deterioration of the general condition, a gluten-, egg-, cow's milk–, and lactose-free diet combined with short-term systemic steroids was begun. After this therapy, the severity of the diarrhea decreased, and there was no evidence of persisting thrombosis, so that the boy was temporarily sent home in satisfactory general condition. However, steatorrhea (25.4 g of fat per day) persisted. After 8 months of a strict exclusion diet, he had a clinical relapse with acute watery stool, electrolyte dysfunction (hyponatremia: sodium 115 mEq/l), dehydration, extreme acidosis, protein-loosing enteropathy, and additional weight loss. This time, the weight loss was 39% of his initial weight, the total weight being below the 3rd percentile. Sodium concentrations of more than 100 mEq/l were found in the stool, indicating secretory diarrhea. The anti-gliadin IgA class antibodies increased to 340 U/l, the IgG class to 249 U/l. Because of the acute deterioration of the clinical picture, total parenteral nutrition, along with metronidazole (25 mg/kg per day) and systemic steroids (60 mg/kg per day, Decortin H; Merck, Germany), was instituted. A new endoscopic examination of the upper gastrointestinal tract was performed before total parenteral nutrition was initiated, to exclude other differential diagnoses (infections, lymphoma, or tumors amyloid or collagenous deposit). The small bowel biopsy showed total villous atrophy and crypt hyperplasia. The intraepithelial lymphocytes (IELs) that were CD3+ and CD8+ by immunohistology were not significantly increased. The surface epithelium showed optically clear, period acid–Schiff (PAS)-negative vacuoles. A dense infiltrate composed of lymphocytes, some plasma cells, and a few granulocytes was found in the lamina propria (Fig. 1A). Immunohistochemistry revealed that the lymphoid cells in the lamina propria were mainly CD3+ T cells (Fig. 1B) with some aggregates of B cells (CD20+). A polymerase chain reaction (PCR) analysis (performed as recently described by Dippel et al. [21], and Marafioti et al. [22]) for the rearrangement of the genes of the heavy immunoglobulin chain and the T-cell receptor γ chain did not disclose a clonal T- or B-cell population. By means of an indirect immunofluorescent technique, according to the protocol previously described (1,6), anti-enterocyte antibodies were detected in the serum up to a dilution of 1:320. These autoantibodies were of IgG class, and directed against the cytoplasm and the apical border of the enterocytes (Fig. 2). Anti-goblet cell antibodies IgG class were also detectable with dilutions of up to 1:256. These findings were confirmed by another immunologic laboratory (unpublished findings, R. Mirakian et al., London Hospital Medical College, London, UK).FIG. 1.: (A) Histology of the small bowel specimen shows complete villous atrophy, vacuolization of the surface epithelium, mixed inflammatory infiltrate in the lamina propria with predominance of lymphoid cells, and low numbers of intraepithelial lymphocytes in the surface epithelium (hematoxylin and eosin staining; magnification, ×100). (B–C) Immunohistochemistry discloses an increased number of CD3+ T-lymphocytes in the lamina propria, many of which express CD8+ phenotype (magnification, ×160).Figure 1: ContinuedFigure 1: ContinuedFIG. 2.: Enterocyte antibodies. Cryostat section of human jejunum obtained before therapy and stained by immunofluorescence with the patient's serum and fluorescein-isothiocyanate–conjugated rabbit anti-human IgG. The staining was localized and was more prominent on the luminal apical border and within the enterocyte cytoplasm (indirect immunofluorescence, original magnification ×100).Figure 2: ContinuedOther tested autoantibodies (anti-pancreatic islet cells, anti-parietal cells, anti-mitochondrial, anti-smooth muscle, anti-kidney tubule, and anti-kidney basal membrane) were not detected. After the beginning of the therapy the serum level of the total immunoglobulin IgG increased slightly; the levels of IgA, IgM, and IgE were normal. During the course of the disease, slight hypogammaglobulinemia (IgG: 526 mg/dl) occurred. The level of the secretory IgA in the saliva was normal. An examination of the systemic immune status showed a normal distribution of the cellular populations, and the surface antigen expression of peripheral blood lymphocytes (CD3+ 85%; CD4+ 40%; CD8+ 34%, and CD56+ 7%) was normal. The in vitro T-cell proliferation after stimulation with phytohemaglutinin (2.4 mg/ml) and anti-CD3 was also normal. CD25 expression was 6%, and HLA-DR expression on T cells was slightly increased at 12%. In view of the clinical relapse and the new serologic and histologic findings, we thought that this was a case of AIE and administered oral immunosuppressive therapy with cyclosporin-A (Cy-A) for 26 months. During this period, the Cy-A blood level was between 150 and 200 ng/ml. The patient experienced repeated episodes of pain in the calf of the right leg. Both ascending contrast venography and color flow duplex sonography confirmed the presence of venous thrombosis. Contrast venography showed in addition that there were two venae poplitea on the right side and disclosed postthrombotic changes in these veins. Genetic analysis, as has been described in detail (23), revealed a heterozygous mutation of the factor V gene with an increased resistance to the activated protein C (ProC Global; Dade Behring Marburg GmbH, Marburg, Germany). By means of an indirect noncompetitive immunoassay for the semiquantitative and qualitative determination of cardiolipin antibodies (ACA EIA-Kit; Pharmacia & Upjohn Diagnostics GmbH, Freiburg, Germany), anti-cardiolipin IgG-class antibodies in a concentration of 27 U/ml (normal, <12 U/ml) were detected in the serum. The anti-cardiolipin antibodies type IgM were negative (<6.0 U/ml), and lupus anticoagulant antibodies were not detected (Bioscientia-Labor, Ingelheim, Germany). Lifelong anticoagulant therapy with cumarin (Marcumar, Roche, Germany) was introduced. After the sixth month of Cy-A therapy, stool frequency and the patient's general condition improved. A control biopsy of the small bowel mucosa showed partial villous atrophy (degree IV-III) and decreased inflammation (Fig. 3). The titer of the antigliadin antibodies had also decreased (IgA, 36 U/l; IgG 96, U/l). We recommended strict gluten free-diet, because the consumption of gluten led to an increase of the titer of the antigliadin antibodies (IgA 81 U/l, IgG 99 U/l).FIG. 3.: Second intestinal biopsy after 6 months of therapy with cyclosporin A. Partial villous atrophy with a focal inflammatory infiltration in the lamina propria and low intraepithelial leukocyte counts are observed (hematoxylin and eosin stain; magnification, ×60).After 10 months of Cy-A administration, gluten-free diet, and steroid and anticoagulant therapy, a new biopsy sample showed deterioration of the histologic picture. The small intestinal mucosa was completely remodeled. There was also an increase of the inflammatory infiltrates in the lamina propria and crypt hyperplasia with no goblet cells visible. Immunostaining showed dense T-cell infiltrates present in the lamina propria with a mixture of CD4+ and CD8+ T cells. After 13 to 15 months of Cy-A therapy a new tissue sample was taken. The villi were still completely atrophic (Fig. 4). A reduction of the goblet cell content and persistence of the dense inflammatory infiltrates in the lamina propria with an almost equal number of CD4+ and CD8+ T-lymphocytes were seen. An increase of the IELs within the superficial proportion of the crypts was noted. These lymphocytes expressed CD8 and the TIA1 molecule. Because of increased serum concentration of creatinine (1.8–2.86 mg/ml), we discontinued Cy-A. This led to an immediate relapse of diarrhea and deterioration of the patient's general state. Therapy with azathioprine (2 mg/kg per day) and intermittent steroid administration were instituted. The patient then felt well, and his weight stabilized (74–76 kg). The stool normalized. The last histologic examination of the upper gastrointestinal tract revealed a slight increase of CD3+ CD8+ IELs in the mucosa of the stomach (antrum and corpus) and persistence of the histologic changes including total villous atrophy and absence of goblet cells in the small intestinal specimens. Results of analysis of the biopsy specimens of the colon and rectum were normal.FIG. 4.: (A) Villi are completely atrophic with persistence of the dense inflammatory infiltrates in the lamina propria, cryptic hyperplasia, absence of goblet cells, and increased intraepithelial leukocytes. (B) Immunohistochemistry discloses an increased expression of TIA-1-molecules in CD8+ lymphocytes, more in the crypt, and also in the superficial villus (arrows) (alkaline phosphatase-antialkaline phosphatase [APAAP] technique; magnification, ×160).Figure 4: ContinuedDISCUSSION Autoimmune enteropathy is a generalized intestinal disorder belonging to the group of diseases that cause intractable diarrhea. The hallmarks are protracted diarrhea, severe intestinal crypt hyperplastic villous atrophy, generally with low numbers of IELs, enterocyte autoantibodies, and various associated autoimmune conditions (24–26) We report a teenaged patient in whom we diagnosed AIE after a clinical, serologic, and histologic investigation. He had sudden development of watery stool and a simultaneous thrombosis of the leg. Histologically and serologically we detected complete villous atrophy of the small bowel mucosa, anti-gliadin antibodies, anti-enterocyte antibodies, and anti-goblet cell antibodies. The precise cause of the venous thrombosis in this patient is not clear, because he had two conditions associated with an increased risk for the development of thrombosis: a heterozygous mutation of the coagulation factor V gene with an activated protein C resistance and high titer of IgG anti-cardiolipin antibodies (phospholipid autoantibodies). Mutation of the factor V gene is the most common genetic cause of thrombophilia (27–31). Anti-cardiolipin antibodies can be found in 50% of all the patients with systemic lupus erythematosus (SLE), and also in rheumatic disorders (32–34), in which they may cause venous and arterial thromboses. Our patient did not have SLE, but he definitely had an autoimmune process that affects the small intestinal mucosa and the coagulation system. HLA typing disclosed the presence of DQ2 and DR7, which are typically found in patients with AIE (7,19). At present, it is not known whether the mutation of the coagulant factor V gene is associated with the presence of HLA class II DQ2. We also do not know whether there is an association between AIE and mutations of the coagulant factor V gene. According to the literature, AIE may be associated with other autoimmune diseases, such as diabetes type I, hemolytic anemia, nephropathy, thrombocytopenia, and autoimmune vasculitis (35–37). Therefore, the presence of anti-cardiolipin antibodies and consecutive thrombosis could also be associated with AIE. Recently, venous thrombosis of the leg associated with an anti-cardiolipin antibody and with diarrhea, was reported in a female patient (phospholipid syndrome), but unfortunately in this case no investigations for the presence of anti-enterocyte antibodies were performed (38). Most of the patients with celiac disease involving gluten-sensitive enteropathy also have this heterodimer HLA-DQ (a 1*0501, b 1*0201) (39) and show a similar clinical, histologic, and serologic picture in earlier stages. Although our patient had high-titer IgA and IgG antigliadin antibodies, his illness did not meet the criteria for celiac disease, because his serum was negative for anti-endomysium antibodies, he did not respond to a gluten-free diet, and IELs were not increased in the small bowel specimens initially. In contrast to the two adult patients reported by Corazza et al., (19) our patient had a high titer of antigliadin antibodies at the onset of the disease. These antibodies disappeared after combined therapy consisting of gluten-free diet and steroids with Cy-A. When gluten was reintroduced, the serum concentration of antibodies increased immediately, and watery stools reappeared as well. It is remarkable that after the administration of Cy-A it took approximately 6 to 8 months until these antibodies disappeared. In contrast, the anti-enterocyte, anti-goblet cell, and anti-cardioplin antibodies remained elevated. In our patient, the enterocyte antibodies were of IgG class, and the intensity of the staining was more prominent in the microvillar luminal border. The titer and the pattern of the indirect immunofluorescence (IFL) reaction was not significantly modified, despite immunosuppressive therapy. The anti-enterocyte antibodies probably correlate best with the histologic damage including villous atrophy. The persistence of these autoantibodies is associated with more severe enteropathy. These patients can be refractory to therapy, whereas in patients with a low titer, a milder course may be observed (1). Anti-enterocyte antibodies appear to be reliable serologic markers for AIE, because it is rare that patients with other gastroenterologic diseases display these antibodies (7). Nevertheless, a high incidence of enterocyte autoantibodies was found in adults with human immunodeficiency virus infection (13). Histologically, we detected in our patient a completely remodeled small intestinal mucosa along with almost complete persistence of villous atrophy and crypt hyperplasia. Initially, the IELs in the surface epithelium were not increased. Later during the course of the disease, an increase in IELs was noted in the superficial proportion of the crypts of the small bowel and in the stomach. These IELs were CD3+ and CD8+ and expressed the TIA-1 molecule that is the usual phenotype of IELs of the intestine. The TIA-1 molecule is a cytotoxic granule-associated protein that has been shown to define a subpopulation of CD8+ T-lymphocytes and to be natural killer possessing cytotoxic potential. It is thought to be involved in inducing programmed cell death in target cells (40,41). Goblet cells were absent from the small intestinal mucosa in our patient as has been described by Moore et al. (6) in another boy with autoimmune enteropathy and anti-goblet cell antibodies. In conclusion, we describe a teenaged patient with late-onset AIE and recurrent venous thrombosis. Similar to others, we recommend that patients with unresponsive intestinal villous atrophy be routinely investigated for enterocyte autoantibodies, regardless of the patient's age. Further studies will show whether the association of AIE with thrombosis was merely a coincidence or whether it reflects a true association between these two diseases. Acknowledgments: The authors thank Dr. Rita Mirakian for her support in the diagnosis and Samuel Mucher for help with the manuscript.