Antibody-dependent cell-mediated cytotoxicity in serum samples from patients with ulcerative colitis. Relationship to disease activity and response to total colectomy

Antibody to tropomyosin isoform 5 and complement induce the lysis of colonocytes in ulcerative colitis

OBJECTIVES

Tropomyosins (TMs) are cytoskeletal microfilament proteins present in all eukaryotic cells. Human TM isoform 5 (hTM5) is the predominant isoform in colonic epithelial cells. Antibodies against hTM5 are found both in the sera and in the mucosa of patients with ulcerative colitis (UC) but not Crohn's disease (CD). We investigated whether anti-hTM5 autoantibodies are pathogenic.

METHODS

Normal-appearing colonic mucosal biopsy specimens were incubated with autologous serum. After 45 min, deposition of the complement component C3b was identified by indirect immunofluorescence assay (IFA). Additional specimens were incubated with autologous serum fixed in formalin, and their architecture was examined by hematoxylin and eosin (H&E) staining.

RESULTS

For 79% of UC patients, autologous serum caused C3b staining along the colonic epithelium. Recombinant hTM5 or anti-hTM5 monoclonal antibody blocked serum-induced C3b deposition. Immunoglobulin G (IgG) antibody and affinity-purified anti-hTM5 IgG antibody from UC sera with complement caused C3b deposition, indicating specificity of hTM5 as an autoantigen. When analyzed by H&E staining, sera obtained from 71% of UC patients caused a significant loss of epithelium. This process was inhibited by Fc fragments, indicating that it is complement mediated. With medium, normal, or CD serum, there was no C3b deposition or morphological changes of the colonic epithelium, indicating disease specificity. The ileal mucosa was not affected by UC sera, suggesting specificity for the colon. In UC mucosa, expression of hTM5 increased.

CONCLUSIONS

hTM5 acts as an autoantigen in UC. hTM5-specific IgG autoantibody in sera from UC patients induces C3b deposition and destruction of colonic epithelial cells, suggesting a direct pathogenic effect. If used as a diagnostic test to distinguish UC from CD, IFA would have 79% sensitivity and 100% specificity. Development of blocking antibodies may lead to novel therapies.

Autoantibodies against human tropomyosin isoform 5 in ulcerative colitis destroys colonic epithelial cells through antibody and complement-mediated lysis

INTRODUCTION

Patients with ulcerative colitis (UC) have IgG1 antibodies in serum and colon against human tropomyosin isoform 5 (hTM5), a cytoskeletal microfilament protein found intracellularly and on the surface of colonic epithelial cells (EC). These antibodies may be pathogenic in UC.

METHODS

Sera from patients with UC (n=110) or Crohn's disease (CD) (n=50) and from healthy individuals (Hl) (n=30) were preincubated with recombinant hTM5 or bovine serum albumin (BSA), then cultured for 4h with (51)Cr-labelled colonic adenocarcinoma cells (LS180). Cytotoxicity was determined by (51)Cr release assay.

RESULTS

All serum samples lysed up to 36% of LS180 cells regardless of the source of the serum. However, adding hTM5 to UC, but not to CD or HI, sera reduced cytotoxicity by up to 75%. This hTM5-induced inhibition of cytotoxicity was found especially with sera from UC patients with active disease, and was found even after total colectomy. The hTM5-induced inhibition was mediated by purified IgG from UC sera. Complement was involved since hTM5-induced inhibition of cytotoxicity declined with either heat inactivation of the sera or premixing sera with Fc fragments.

CONCLUSIONS

This study shows that hTM5-specific IgG autoantibodies present in UC sera destroy LS180 cells by antibody and complement-mediated lysis. Such a phenomenon was not seen in CD or HI. This suggests an autoantigenic role of hTM5 and anti-hTM5 antibodies in the pathogenesis of UC. This observation may lead to novel diagnostic and therapeutic possibilities.

HLA class-I-restricted and colon-specific cytotoxic T cells from lamina propria lymphocytes of patients with ulcerative colitis

We established cytotoxic T-lymphocyte (CTL) lines against colonic epithelial cell line from LPLs of patients with ulcerative colitis by continuous stimulation with human lymphocyte antigen A (HLA-A) matched colonic epithelial cell lines and clones from LPLs using polyclonal stimulation with phytohemagglutinin. The established CTL lines and clones showed cytotoxicity against HLA-A-matched colonic epithelial cell line but not against HLA-mismatched colonic epithelial cell lines, and HLA-A-matched lung and esophagus cell lines. The CTL response was HLA class I-restricted and mediated by CD8-positive T-lymphocytes. Moreover, the CTL line showed cytotoxicity against autologous B-LCLs pulsed with peptides extracted from HLA-A-matched colonic epithelial cell line but not against other organ-derived peptides pulsed and unpulsed autologous B-LCLs. CTL lines and clones established from LPLs of patients with ulcerative colitis showed colon-specific and HLA class I-restricted killing of human colonic epithelial cell line, suggesting that these CTLs may play a role in colonic epithelial cell damage in some patients with ulcerative colitis.

Generation of MHC class I-restricted cytotoxic T cell lines and clones against colonic epithelial cells from ulcerative colitis

We established CTL lines and clones against colonic epithelial cells from PBLs of patients with ulcerative colitis by continuous stimulation with HLA-A locus-matched colonic epithelial cell lines. We developed a nonradioactive europium release cytotoxicity assay to detect CTLs. PBLs from 3 of 12 patients but not from any of 14 normal controls who shared at least one haplotype of HLA-A locus with two colonic epithelial cell lines, CW2 and ACM, showed increased cytotoxicity against these lines. Three CTL lines established from the PBLs of patients showed increased cytotoxicity against HLA-A locus-matched CW2 or ACM but not against matched lung or esophagus cell lines. The phenotypes of CTL lines were alpha beta-TCR+ CD3+ CD8+ CD16-. The CTL line MS showed increased cytotoxicity against freshly isolated colonic epithelial cells but not against cells with a different HLA-A locus. Two CTL clones were generated from MS and clone 3-2, expressing CD3+ CD8+ CD4- CD56-, showed high MHC class I-restricted cytotoxicity against the colonic epithelial cells. These results indicated that CTLs against colonic epithelial cells may contribute to epithelial cell damage in ulcerative colitis.

Tumor necrosis factor and interferon-gamma augment anticolon antibody- dependent cellular cytotoxicity in ulcerative colitis

The effect of tumor necrosis factor (TNF) and interferon (IFN)-gamma on antibody-dependent cellular cytotoxicity (ADCC) in patients with ulcerative colitis (UC) was investigated. ADCC activity was measured by the 51Cr release assay, using peripheral blood mononuclear cells of healthy subjects as effector cells and RPMI 4788 cells derived from human colon cancer as target cells. ADCC activity under sera from healty subjects remained low whether or not the effector cells were pretreated with TNF (100 U/ml, 16h). Under sera from UC patients, ADCC activity of 13.9%, compared to 9.6% when pretreatment was deleted. The effect of IFN pretreatment (100 U/ml, 16h) was also examined under sera from UC patients; in that experiment activity rose to 26.8%, in comparison to a 10.7% when IFN-gamma pretreatment was deleted. Finally, when the effector cells were pretreated with both TNF and IFN-gamma (100 U/ml of each, 16h) the ADCC activity under sera from UC patients was higher than when either TNF or IFN-gamma were used alone. These results suggest that TNF and IFN-gamma, by increasing ADCC activity in UC lesions, are involved in cell injury in the colonic epithelium. IFN-gamma appears to increase ADCC activity by increasing the number of high affinity monocyte Fc gamma RI receptors, while TNF increases ADCC activity by a different mechanism.

ADCC-mediating capacity in children with cow's milk protein intolerance in relation to IgG subclass profile of serum antibodies to beta-lactoglobulin

In a previous study sera from children with cow's milk protein intolerance (CMPI) exhibiting gastrointestinal symptoms were found to efficiently induce antibody-dependent cell-mediated cytotoxicity (ADCC) to beta-lactoglobulin-coated cells. In contrast, sera from children with coeliac disease showed a low ADCC-mediating capacity, despite high levels of IgG anti-beta-lactoglobulin antibodies. The study described here was undertaken to evaluate whether differences in IgG subclass profile of anti-beta-lactoglobulin antibodies could explain the observed variations in the ADCC-mediating capacity. Forty-eight sera from the following groups of children were investigated: CMPI with predominantly gastrointestinal symptoms, CMPI with skin symptoms of immediate-onset, children with untreated coeliac disease and healthy references. Absorption experiments indicated that primarily IgG1 antibodies were responsible for the ADCC-mediating capacity of the sera. Accordingly, the ADCC reactivity of individual sera correlated with their IgG1 antibody levels. Sera from CMPI children with gastrointestinal symptoms, most of which had a high ADCC reactivity, also demonstrated a distinctive subclass pattern of their anti-beta-lactoglobulin antibodies with higher relative proportions of IgG1 (ratios: IgG1/IgG, IgG1/IgG3 and IgG1/IgG4) than those from the other diagnostic groups. Using logistic regression analysis, the diagnostic potential of ADCC as well as of different IgG subclass variables for the recognition of gastrointestinal symptoms caused by CMPI was evaluated. The ADCC reactivity of sera was found to be the best predictor in this model.

Anti-neutrophil cytoplasmic antibodies in chronic inflammatory bowel disease. Prevalence and diagnostic role

BACKGROUND

Anti-neutrophil cytoplasmic antibodies (ANCA), originally found to be associated with vasculitis, have been reported to be present in chronic inflammatory bowel disease. Most often the ANCA staining pattern is of the perinuclear type (p-ANCA), although nuclear and cytoplasmic stainings are seen. Single studies have shown some of the antibodies to react with lactoferrin or cathepsin G; however, most studies have not been able to determine a main antigenic specificity. We studied the prevalence of ANCA in sera from 155 patients with ulcerative colitis, 128 patients with Crohn's disease, and 51 patients with coeliac disease. The presence of ANCA was correlated to disease activity, extent, and age of onset of the diseases. Furthermore, we tried to characterize the antigen specificity by enzyme-linked immunosorbent assay (ELISA), using elastase, lactoferrin, myeloperoxidase, proteinase 3, and cathepsin G as antigens.

METHODS

The sera were screened for ANCA by indirect immunofluorescence. Anti-nuclear antibodies (ANA) were analysed on HEp2 cells, and ELISA for specific ANCA was performed using the antigens mentioned.

RESULTS

Most of the sera with positive immunofluorescence had the p-ANCA type of pattern. Seventy-eight of 155 (50.3%) of the patients with ulcerative colitis were ANCA-positive, compared with 31 of 128 (24.2%) of patients with Crohn's disease (p < 0.001). However, in the subgroup with Crohn's colitis, 16 of 44 (36.4%) were ANCA-positive. Only 4 of 51 patients (7.7%) with coeliac disease showed positive immunofluorescence (p < 0.001 compared with ulcerative colitis). Less than 10% of the samples were positive in the specific ELISA assays; thus other than the most well known granule proteins can be the target for ANCA in ulcerative colitis.

CONCLUSION

ANCA occur significantly more often in ulcerative colitis than in Crohn's disease. However, the prevalence of ANCA is rather high in Crohn's colitis. ANCA are thus of limited value in differentiating Crohn's colitis from ulcerative colitis. ANCA found in inflammatory bowel disease are different from those associated with vasculitis. The antigen(s) responsible remain to be determined.

Expression of a unique protein on colon cancer cells that reacts with a novel monoclonal antibody and ulcerative colitis serum

We earlier developed a MoAb, 7E12H12 (IgM isotype), against a protein present in normal colonic epithelial cells. To examine if 7E12H12-reactive protein is expressed in colon cancer cells and is recognized by ulcerative colitis (UC)-associated autoantibody, we investigated several colon cancer cell lines. 7E12H12 reactivity against the cells was examined by indirect immunofluorescence assay and whole cell ELISA against six colon cancer cell lines HT-29, LoVo, COLO 205, DLD-1, LS 180 and SW 1116. A competitive ELISA was developed using 7E12H12 MoAb and patients' serum to examine the cross-reactive antibodies in the serum. Among the six colon cancer cell lines only LS 180, DLD-1 and SW 1116 reacted with 7E12H12 MoAb, while others did not. The mean (+/- s.e.m.) inhibition of the binding of 7E12H12 MoAb to LS 180 cells by UC serum (n = 51) was 42 +/- 2.1%, whereas in normal subjects (n = 17) it was 14 +/- 2.6%, in Crohn's disease (n = 19) it was 15.3 +/- 2.5%, in infectious diarrhoea (n = 10) it was 11% +/- 3%, and in systemic lupus erythematosus (n = 10) it was 2% +/- 0.6%. The inhibition by the UC group was significantly (P < 0.001 - < 0.0001) higher than any of the non-UC groups, and this inhibition was mainly by IgG1 antibody. The protein in the specific colon cancer cells recognized by the 7E12H12 MoAb cross-reacts with UC-IgG1 antibody and may provide an in vitro system to examine the autoimmune mechanisms in UC.

Immunoglobulin G subclass distribution of human anticolon antibodies in ulcerative colitis

Immunoglobulin G (IgG) subclasses of anticolon antibodies were studied in patients with ulcerative colitis (UC) using enzyme-linked immunosorbent assay (ELISA). The concentrations of total serum IgG subclasses were also measured by ELISA. The values for total serum IgG subclasses in patients with UC were not significantly different from those in normal controls, while the ratio of IgG1 to IgG2 in the patients was significantly higher than that in normal controls. All four IgG subclasses of autoantibodies were demonstrated in the sera of the patients. IgG4 anticolon antibodies were detected most frequently (15 out of 18 patients, 83%). IgG2 was the next most prevalent (9 of 18 patients, 50%). The activity of anticolon antibodies in each subclass did not correlate with the concentration of the corresponding serum IgG subclass. Seven cell lines producing anticolon antibodies were obtained from the colonic mucosa of the patients by Epstein-Barr virus (EBV) transformation. IgG subclasses of anticolon antibodies secreted by these cell lines were also varied. IgG4 subclass was secreted by three EBV transformed cell lines, all of which produced IgG4 anticolon antibodies. These results suggest that all four different IgG subclasses could respond to the colon antigens and that various antigens in colonic mucosa or lumen may contribute to the induction of those autoantibodies. In addition, the prominence of IgG4 anticolon antibodies may support the pathogenic role of this subclass in UC as in other autoimmune diseases.

Interleukin 2 and interferon-gamma augment anticolon antibody dependent cellular cytotoxicity in ulcerative colitis

In vitro effects of cytokines and therapeutic drugs on antibody dependent cellular cytotoxicity (ADCC) mediated by anticolon antibody were investigated in serum samples from patients with ulcerative colitis. A 51Cr release assay was used to examine ADCC activity with the colon cancer cell line, colo 205, as the target and peripheral blood mononuclear cells as the effector. High ADCC activity was shown in 13 of 32 (41%) patients with ulcerative colitis. This ADCC activity was inhibited by protein A treatment of the serum samples. Interleukin 2 (IL2) activated effector cells could enhance ADCC activity, but interferon-gamma (IFN-gamma) or tumour necrosis factor-alpha (TNF-alpha) had no effect on the cytotoxic activity of effector cells. Treatment of target cells with IFN-gamma increased the vulnerability of these cells to ADCC with a large increase of intercellular adhesion molecule-1 (ICAM-1) expression on their surface. Monoclonal antibodies to ICAM-1 inhibited this IFN-gamma enhanced ADCC activity. Interestingly, prednisolone (PSL) reduced ADCC activity, but sulphasalazine (SASP) or 5-aminosalicylic acid (5-ASA) did not. These results suggest that IL2 and IFN-gamma could enhance colonic epithelial cell injury mediated by the ADCC mechanism in ulcerative colitis and that ADCC enhanced by cytokines is restored by PSL treatment.

Anti-lactoferrin antibodies and other types of ANCA in ulcerative colitis, primary sclerosing cholangitis, and Crohn's disease

Fifty two serum samples from patients with Crohn's disease, 24 from patients with ulcerative colitis, and 12 from patients with primary sclerosing cholangitis were analysed for the presence of anti-neutrophil cytoplasm antibodies (ANCA) of IgG and IgA class by means of enzyme linked immunosorbent assays using lactoferrin, myeloperoxidase, and antigen extracted from azurophil granules, 'alpha antigen' (that is, an antigen preparation containing proteinase 3) as substrates. A high frequency of positive tests for IgG anti-lactoferrin antibodies was found in sera from patients with ulcerative colitis (50%) and primary sclerosing cholangitis (50%). In Crohn's disease only 4 of 52 (8%) sera had anti-lactoferrin antibodies--in all four instances the sera belonged to patients with disease involving the colon. All patients with sclerosing cholangitis and positive tests for anti-lactoferrin had ulcerative colitis. Low levels of IgG antibodies against myeloperoxidase or alpha antigen were also found occasionally in sera from patients with ulcerative colitis and primary sclerosing cholangitis. IgA antibodies directed against lactoferrin and alpha antigen (but not myeloperoxidase) were seen in all three conditions.

Expression of the 40 kD protein in DLD-1 colon cancer cells and the effect of cytokines

We recently reported the presence of an organ-specific 40 kD colonic protein which acts as an autoantigen(s) in patients with ulcerative colitis. Using a specific monoclonal antibody directed against 40 kD protein (7E12H12, IgM isotype), in conjunction with immunocytochemistry and flow cytometry, we examined the presence of the 40 kD protein on human colon cancer cells, DLD-1, and also characterized the ability of cytokines, IFN-gamma and tumour necrosis factor, to modulate the expression of this protein on these tumour cells. The presence of the 40 kD protein was localized to the plasma membrane; less was present within the cytoplasm. Following exposure to IFN-gamma (10-1000 U/ml), DLD-1 colon tumour cells showed a dose- and time-dependent increase in 7E12H12 antibody associated immunofluorescence, with the maximum 7E12H12 antibody binding observed with 100 U/ml IFN-gamma at 48 h. In contrast, tumour necrosis factor did not alter the levels of anti-40 kD antibody binding over that of control cells. Since IFN-gamma is also known to induce class II major histocompatibility antigens, we examined the possibility of cross-reactivity of HLA class II antigens and Mr 40 kD epitope. Neither pre-incubation of DLD-1 colon tumour cells with anti-HLA class II antibodies followed by 7E12H12 nor co-incubation of both antibodies altered the amount of 7E12H12 antibody binding. Using a direct ELISA, a highly enriched preparation of Mr 40 kD protein reactive to anti-40 kD antibody did not react with HLA class II antibodies. The present results suggest that 40 kD protein is present on DLD-1 human colon tumour cells and that although the 40 kD protein epitope expression is increased by the lymphocyte-derived cytokine, IFN-gamma, the epitope is separate and distinct from the class II HLA antigens. Further studies on the 40 KD protein may elucidate its autoantigenic role in the pathogenesis of inflammatory bowel disease.

Antibody-dependent cell-mediated cytotoxicity to beta-lactoglobulin-coated cells with sera from children with intolerance of cow's milk protein

The capacity of serum antibodies against beta-lactoglobulin to mediate antibody-dependent cell-mediated cytotoxicity (ADCC) was analysed in sera from children with cow's milk protein intolerance (CMPI). The children with CMPI were divided into three groups according to clinical features: delayed-onset CMPI with gastrointestinal symptoms (n = 8); immediate-onset CMPI with gastrointestinal and skin symptoms (n = 8); and immediate-onset CMPI with skin symptoms only (n = 8). The CMPI groups were compared with children with untreated (n = 9) or treated (n = 8) coeliac disease and a control group (n = 22). Sera from the children were examined for cytotoxic effects using lymphocytes from healthy adults as effector cells and radiolabelled beta-lactoglobulin-coated erythrocytes from the same donor as target cells. In addition, IgG and IgA serum antibodies against beta-lactoglobulin were determined with ELISA. Sera from children with CMPI and gastrointestinal symptomatology showed a significantly increased capacity to induce ADCC reactivity as compared with controls. This increased capacity was seen in sera from those with immediate as well as delayed onset of the gastrointestinal symptoms. In contrast, sera from children who had an immediate-onset CMPI with only skin symptoms mediated no such increase in ADCC reactivity. Moreover, children with coeliac disease with a few exceptions, demonstrated low ADCC reactivity, despite the fact that they had high levels of antibodies against beta-lactoglobulin. ADCC may be an immunopathogenic mechanism in certain cases of CMPI with gastrointestinal symptoms.

In vitro anticolon antibody production by mucosal or peripheral blood lymphocytes from patients with ulcerative colitis

Serum anticolon antibody and in vitro anti-colon antibody production by peripheral blood and mucosal lymphocytes was investigated in patients with ulcerative colitis. The frequency of serum anticolon antibody was 71% in 41 patients with ulcerative colitis, estimated by enzyme linked immunosorbent assay (ELISA) using isolated rat colon epithelial cells. This finding confirms our previous report on the frequency of serum anticolon antibody detected by flow cytometry analysis. The estimated frequencies of IgG anticolon antibody secreting cells were 1.5-12.5/10(6) cells in the colonic mucosa and 0.1-0.5/10(6) cells in peripheral blood, from patients with ulcerative colitis when Epstein-Barr virus (EBV) was used as a B cell polyclonal activator. Poisson analysis of limiting dilution culture showed that about one per 140 IgG cells in the colonic mucosa synthesised anticolon antibody. Two monoclonal IgG antibodies were obtained from EBV transformed anticolon antibody secreting cells by limiting dilution method. One reacted with goblet cells in the intestine, and the other reacted mainly with colonic epithelial cells. These results suggest that heterogeneous anticolon antibodies are present in patients with ulcerative colitis and that colonic mucosa may be the main source of anticolon antibody. Local autoimmune reaction might have an important role in causing the inflammation of colonic mucosa in this disease.

Advances in the assessment of disease activity in inflammatory bowel disease

Knowledge of the severity and extent of the inflammation in inflammatory bowel diseases provides a means of determining rational therapeutic strategies in affected patients. During the past 3 decades, several clinical, laboratory, and combined indices have been proposed for the assessment of inflammatory bowel disease; refinements in radiologic methods and the availability of endoscopy and biopsy have facilitated the accurate assessment of the extent and severity of the disease. In relapsing conditions such as inflammatory bowel disease, however, the use of such procedures is limited by the radiation exposure or the relatively invasive nature of the technique. In this article, we review the proposed methods and recent advances in assessment of patients with inflammatory bowel disease; we also discuss possible strategies at the time of diagnosis, during recurrence, and in evaluation of the efficacy of drug or dietic therapy.

Ulcerative colitis: specific antibodies against a colonic epithelial Mr 40,000 protein

A colon tissue-bound disease specific IgG (CCA-IgG) antibody can be eluted from the colonic tissue of patients with ulcerative colitis (UC). CCA-IgG recognizes an Mr 40,000 protein present in both normal and diseased colon. Using an anti-Mr 40,000 murine monoclonal antibody this protein has been localized exclusively in colonic epithelial cells mainly along the baso-lateral domains of plasma membrane. Non-colonic epithelial tissues from 12 different organs including small intestine did not react with the anti-Mr 40,000 monoclonal antibody. The Mr 40,000 protein is present in the RPMI-4788 colon cancer cells but not in HeLa. The antibody dependent cell-mediated cytolysis by UC sera against RPMI-4788 cells was inhibited by anti-Mr 40,000 polyclonal antibody. These findings suggest that CCA-IgG and the Mr 40,000 colonic epithelial protein are involved in an autoimmune reaction and may be important in the pathogenesis of UC.

Ulcerative colitis specific cytotoxic IgG-autoantibodies against colonic epithelial cancer cells

Serum antibodies cytotoxic to the colon cancer cell line RPMI 4788 were studied in 42 patients with ulcerative colitis, 61 patients with Crohn's disease, 27 patients with other inflammatory diseases (disease-controls) and 22 healthy controls. Cytotoxicity of antibodies towards RPMI 4788 was studied by means of a chromium release assay using peripheral blood mononuclear leucocytes of healthy subjects as effector cells. Using a four hour antibody dependent cell mediated cytotoxicity assay sera from 29% of ulcerative colitis patients contained antibodies cytotoxic for the target, while only 3% of the Crohn's patients and 6% of the disease controls and non of the healthy controls were positive. When an 18 hour assay was applied, however, not only 40% of ulcerative colitis patients, but also 14% of Crohn's patients and 21% of disease controls were found positive. The reactive antibody in the four hour assay was mainly of the IgG class, which points at a classical antibody dependent cell mediated cytotoxicity mechanism. In the 18 hour cytotoxic assay IgG and particularly IgM antibodies were found to be reactive. This suggests that in the latter case other cellular cytotoxic mechanism might be involved. There was a significant inverse correlation between the appearance of the ulcerative colitis restricted IgG-anticolon epithelial cell antibodies (four hour assay) and the disease activity (p less than 0.01). Absorption studies showed that the reactive antigen is not specific for ulcerative colitis colonic tissue, but is similarly found in Crohn's bowel tissue, and to a lower extent in normal bowel, liver and kidney. The reactive antigen, however, could not be detected in brain and lymphoblastoid cells.

Immunofluorescence assay using Crohn's disease tissue-injected athymic nude mouse lymph nodes in the diagnosis of inflammatory bowel diseases

Sometimes, even after extensive investigative efforts, the diagnosis of inflammatory bowel disease remains in doubt. The accurate diagnosis is important if appropriate therapy is to be instituted. A simple indirect immunofluorescence assay that tests the patient's serum against lymphoid tissues from athymic nude (nu/nu) mice receiving injections of filtrates of Crohn's disease tissue is proposed. Eighty percent of serum samples from patients with active, symptomatic Crohn's disease give positive results of immunofluorescence when tested with these lymphoid tissues. The false-positive rate has been very low (less than 10 percent). Because this assay is fairly sensitive and least invasive, it was used for the clarification of many puzzling cases that were seen at the Albert Einstein College of Medicine over the past three years. Ten of these cases were selected for illustration and discussion and are presented in this report.

Isolation and characterization of a colonic autoantigen specifically recognized by colon tissue-bound immunoglobulin G from idiopathic ulcerative colitis

Patients with idiopathic ulcerative colitis (UC) have a colonbound antibody (CCA-IgG) that reacts with colon tissue extracts. We have partially characterized a colonic protein that is specifically recognized by CCA-IgG. CCA-IgG was eluted from operative colon specimens from 10 patients with UC. A colon tissue-bound IgG was similarly eluted from six patients with Crohn's colitis, two with ischemic colitis, and one with diverticulitis. Purified serum IgG from patients with Crohn's disease, from normal subjects and a patient with myeloma were also used as additional controls. For detection of antigen(s), tissue extracts were prepared from 26 specimens of colon (UC, 12; Crohn's disease, 6; normal, 4; other controls, 4), 8 specimens of human normal stomach, duodenum, ileum, and liver (2 each). Tissue extracts were also prepared from rats and mice, including germ-free rat colons and rat's fetal colons. Immunorecognition of CCA-IgG to the tissue extracts was examined by affinity-column chromatography and by transblot analysis. Tissue-extracted proteins were electrophoresed in SDS-polyacrylamide gel, transferred to nitrocellulose sheet, and probed with iodinated CCA-IgG, colonic IgG from other inflammatory bowel disease patients, UC serum IgG, and control serum IgG. Although many proteins were present in colon tissue extracts, 9 of 10 CCA-IgG consistently recognized a protein of 40 kD. None of the nine IgG preparations from colon specimens of patients with Crohn's colitis and other colonic inflammatory diseases reacted with the 40-kD protein. Five of six symptomatic UC serum IgG and none of eight control serum IgG reacted with the 40-kD protein. The 40-kD protein was present in all colon specimens and it appeared to be organ specific. It was absent in mouse and rat tissues, including colon. The 40-kD protein is not actin and nor a part of the Ig molecule. These results suggest that the 40-kD protein is a colonic "autoantigen" that may initiate a specific IgG antibody response in UC.