Normal Inflammatory Bowel Disease Mucosa ... - Semantic Scholar

Normal Inflammatory Bowel Disease Mucosa Conceals Alterations in Natural Killer Cell Activity E. A. F. G A N TOL, H. w. VERSPAGET, A. s. PEfiA & c . B. H. w . LAMERS p e p t . of Gastroenterology and Hepatology, University Hospital, Leideaand Dept. of Gastroenterology and Hepatology, Free University Hospital, Amsterdam, Netherlands

&

3

Scand J Gastroenterol Downloaded from informahealthcare.com by Vrije Universiteit Amsterdam on 03/23/12 For personal use only.

van To1 EAF, Verspaget HW, Peria AS, Lamers CBHW. Normal inflammatory bowel disease mucosa conceals alterations in natural killer cell activity. Scand J Gastroenterol 1992;27:999-1005. fion-major histocompatibility complex-restricted cytotoxicity or natural killer (NK) activity could be detected in all intestinal lamina propria mononuclear cell preparations of histologically normal mucosa from 57 patients with gastrointestinal disease. Similar levels of NK activity were detected among the different disease groups. Within the inflammatorybowel disease patient group, however, Crohn’s disease patients showed a threefold higher level of NK activity than detected in ulcerative colitis patients. Cytotoxicity levels in Crohn’s disease patients were also higher than in the control carcinoma patients, whereas ulcerative colitis patients had considerably lower cytotoxicitylevels than the carcinoma patients. Thus, unaffected normal inflammatory bowel disease mucosa conceals alterations in NK activity which might occur before the inflammation. The colon adenocarcinoma cell line Caco-2 was found to be a representative target for detecting individual differences in NK activity of lamina propria mononuclear cells compared with standard K-562 targets. The latter can he of relevance when studying mucosal immunoregulatory mechanisms in intestinal disease.

--I

Key words: Crohn’s disease; inflammatory bowel disease; natural killer activity; normal mucosa;

ulcerative colitis Hein W. Verspaget, M.Sc., Ph. D . , Dept. of Gastroenterology and Hepatology, University Hospital, Building 1 , C4-PO12, P.O. Box 9600, 2300 RC Leiden, The Netherlands

Current trends in inflammatory bowel disease (IBD) research do not seem to cover the natural killer (NK) cell in the gut as a participant in the etiopathology of this disease. Nowadays, with the availability of more specific monoclonal antibodies aa,ainst NK-cell surface antigens and using adapted assay conditions for measuring cytotoxicity, these cells are found to be commonly present in the lamina propria of normal gut mucosa (1-3). Their immunohistochemical detection, in both tissue sections and isolated mononuclear cell preparations, has shown that NK cells comprise only 1% to 3% of the total intestinal lamina propria mono~iuclearcell (LPMC) population depending on the monoclonal antibodies used (3-6). Intriguingly, the distribution of cells with NK activity in the lamina propria of the gut is proportionally low compared with peripheral blood, and they are of a different phenotype than their peripheral blood counterparts (3,643). Our group has previously described LPMC exerting spontaneous o r non-major histocompatibility complex (MHC) restricted cytotoxicity in both short and prolonged chromium-release assays to b e NK cells, whereas T cells did not participate in this type of cellular cytotoxicity (3). NK cells have important implications in disease with regard to the killing of microorganisms and virus-infected cells, modulation of the immunoglobulin secretion, and cytotoxicity against malignant cells o r epithelial cells that have undergone changes of surface antigen expression (9-11). In general, studying non-MHC-restricted cytotoxicity or NK activity in LPMC preparations from IBD patients has

been discouraging. Cytotoxicity levels were found to be absent or low, and this finding was accompanied by the inability to detect cells with NK-cell markers immunohistochemically o r only at very low percentages (12-15). Studies of mucosal NK activity in IBD usually involved cells isolated from inflamed specimens, whereas control LPMC were isolated from normal mucosa of patients with carcinoma. To define the appropriate control mucosal specimens to study the role of NK cells in IBD seems to be of major relevance. Presumably, LPMC isolated from normal mucosa adjacent t o the inflamed region represent an appropriate control population by eliminating differences inherent in individuals, disease, o r tissue characteristics. In the present study the N K activity of LPMC isolated from histologically normal mucosa was assessed in particular with regard to inflammatory bowel disease. Moreover, the sensitivity of a colonic target cell line for NK activity of LPMC was studied and compared with the standard NK-sensitive K-562 cell line. Immunomodulation of NK activity by inflammatory mediators o r gastrointestinal peptides may also require the use of colonic cell lines as exemplary targets for mucosal NK activity. MATERIALS A N D METHODS Specimens studied In this study intestinal specimens were obtained from 57 patients, comprising 28 women and 29 men, ranging in age

E. A. F. van To1 Y / al

1000

from 21 to 89 years, with a mean of 59 2 2 years, with various diagnoses who underwent a bowel resection in our hospital. The population consisted of 28 patients with colorectal carcinoma (mean age, 67 ? 3 years), 17 individuals with IBD, divided into 9 Crohn's disease (CD) (mean age, 48 t 4 years) and 8 ulcerative colitis (UC) (mean age, 45 2 3 years) patients, and 12 patients with miscellaneous diseases (mean age, 57 4 years)-that is, diverticular disease ( n = S ) , sigmoid volvulus ( n = 2), familial adenomatous polyposis coli ( n = 2), Hirschsprung's disease ( n = l), rectopexy ( n = l), and colon interposition ( n = 1). Diagnosis was based on a combination of established clinical, radiologic, endoscopic, and histologic criteria. In the IBD group 3 patients had no medication (2 CD, 1 UC), and 14 received 5-acetylsalicylic acid (5-ASA)-containing drugs, steroid therapy, or a combination of these drugs (7 CD, 7 UC). Macroscopically normal mucosal specimens were taken distant from the tumour or inflamed or otherwise abnormal region. Immediately after resection and dissection by the pathologist the tissue was transported to the laboratory and processed for isolation of LPMC. All mucosal tissue samples were retrospectively judged as being normal according to histologic evaluation by the pathologist. In this study the origin of the intestinal specimens obtained from the various patient


*

groups was divided as follows: all specimens in the carcinoma patients group comprised colonic tissue; in the CD group we obtained seven ileal and two colonic tissues; in the UC group seven colonic and one ileal specimen; and in the miscellaneous diseases group all samples were colonic tissue except for one ileal specimen.

Isolation of effector cells LPMC were isolated by the method described by Bull & Bookman (16) and Fiocchi et al. (17) with slight modifications. In brief, the mucosa was carefully dissected from the underlying submucosa, cut into small pieces of 0.5 to 1 cm2, and incubated in 1mM dithiothreitol (DTT) (Sigma, St. Louis, Mo., USA) in calcium- and magnesium-free Hanks' balanced salt solution (CMF-HBSS) to remove mucus. Then the tissue was collected, washed extensively and rapidly cut into small fragments. Four to five sequential incubations in ethylenediaminetetraacetic acid (EDTA) (0.75 mM; Merck, Darmstadt, Germany) in CMF-HBSS were performed at 37°C for 1-h periods to remove epithelial cells, followed by overnight incubation at 37°C in medium supplemented with 20 U/ml collagenase CLSPA (Worthington, Freehold, Mass., USA) and 278 U/ml DNase (Worthington). The next morning the cell suspension was

A

B

.. -

% CYTOTOXlClTY

% CYTOTOXlClTY

K-562 4 h

0K-562

0K-562 18h

18h

a K-562 4h

451

45

T

T

n I

30

30

T 15

15

0

0

TOTAL

CARC.

IBD

MISC.

CD

uc

Fig. 1A. Cytotoxicity of freshly isolated lamina propria mononuclear cells (LPMC) from histologically normal mucosa of patients with intestinal disease against K-562 target cells in both 4-h and 18-h "Crrelease assays. Mean cytotoxicity levels of the entire patient group studied (total), carcinoma patients (carc.), inflammatory bowel disease patients (IBD), and patients with miscellaneous disease (rnisc.) are shown. Number of individuals studied in each group as indicated within bars. 1B. Cytotoxic activity of freshly isolated LPMC from normal mucosa of patients with inflammatory bowel disease (CD = Crohn's disease, UC = ulcerative colitis) against K-562 target cells in both 4-h and 18-h W r release. Significant difference between IBD subgroups: * p = 0.005, * * p = 0.001. Number of individuals as indicated within bars.

Normal IBD Mucosa and NK Activity


filtered through a 50-pm mesh, washed in medium, counted, and analyzed for viability. This heterogeneous cell population was further enriched for mononuclear cells by density gradient centrifugation using Ficoll-Isopaque (1.077 g/ml). Viability of this cell population was approximately 90%, whereas contamination with epithelial cells was rarely found.

Target cells K-562 erythroleukemia cells and Caco-2 colon adenocarcinoma cells were used as targets in the chromium-release assay. The cells were cultured at 37°C in a 5% C02/95% air incubator and kept in the growing phase by transferring approximately 1:5 twice a week. Caco-2 cells were used in the cytotoxicity assay when they had reached confluency in culture. Cytotoxicity assay Both 4-h and 18-h assays were performed using freshly isolated LPMC and target cells labeled with sodium chromate-51 (5’Cr; 100 pCi/5 x lo6 cells, Amersham Laboratories, U.K.) for 1 h at 37°C. LPMC were tested at a 500: 1 ratio, thus correcting for the relatively low number of cells with NK activity in the gut mucosa according to Gibson et al. (1). This high E/T ratio in mucosal cytotoxicity experiments is used to study differences in activity levels between patient groups and enables the use of less susceptible target cells, such as Caco-2. Incubations for spontaneous and maximum release of chromium by the target cells were performed in sextuplicate, whereas the incubations with effector cells were done in triplicate. Specific cytotoxicity was calculated by the following formula: Cytotoxicity (%) -

Experimental release - Spontaneous release x loo. Maximal release - Spontaneous release

The intra-assay coefficient of variation of our cytotoxicity assay detecting NK activity by LPMC was 5.2 2 0.4%.

Analyses of data Statistical analyses were performed using SPSS/PC+ (Microsoft Corp., Redmond, Wash., USA). Data are expressed as the mean ? SEM unless stated otherwise. Student’s t test or separate variance analysis was used to calculate the significance of the differences between groups. Correlations were described using linear regression analysis. RESULTS The mean cytotoxic activity of LPMC suspensions against K-562 target cells was 13 2 2% in the 4-h assay ( n = 5 5 ) and 28 2 3% in the 18-h assay ( n = 54). The cytotoxic activity of LPMC against Caco-2 target cells showed relatively greater differences between the two incubation periods than found using K-562 targets: 4 h, 4 L 1% ( n = 27), and 18 h, 23 t 4% ( n = 26). NK activity of LPMC against K-562 within the

1001

different patient groups is shown in Fig. 1A. Considerable NK activity was detected in all patient groups when tested at a 500:l effector to target cell ratio. Interestingly, the NK activity of LPMC from normal mucosa was not found to differ when the various patient groups were compared in both 4-h and 18-h assays; that is, patients with carcinomas, miscellaneous diseases, or IBD had cytotoxicity levels similar to those calculated for the entire population. However, when further differentiating the IBD group in patients with Crohn’s disease and those with ulcerative colitis, considerable differences in cytotoxicity levels were found in both 4-h (21 t 4% versus 7 ? 2%, p = 0.005) and 18-h (49 ? 7% versus 16 t 3%, p = 0.001) assays, hence the level in the Crohn’s disease patient group was threefold the level detected in the ulcerative colitis patient group (Fig. 1B). In addition, the cytotoxicity level in the Crohn’s disease group was higher than the carcinoma patient group (4 h, 21 t 4% versus 13 t 2%, p = 0.059; 18 h, 49 t 7% versus 28 t 4%, p < 0.02) and the miscellaneous diseases group in the prolonged assay (18 h, 49 7% versus 23 ? 6 % , p < 0.01), whereas a lower level of cytotoxic activity was detected in the ulcerative colitis group than in the carcinoma group (4 h, 7 L 2% versus 13 2 2%, p < 0.05; 18 h, 16 t 3% versus 28 t 4%,p< 0.05). Of importance was the observation that two patients with Crohn’s colitis had high levels of NK activity by LPMC isolated from the normal colonic mucosa (mean NK activity in the 4-h assay, 21%; in the 18-h assay, 47%), whereas cells from small-bowel tissue of a single UC patient showed low NK activity (4-h assay, 8%; 18-h assay, 17%). No relation was found between drug therapy and the apparent differences in NK activity as detected in LPMC preparations from CD and U C patients in particular. Furthermore, both UC and CD groups were age-matched, and no relation was found between age and NK activity levels either in the entire study group or in the disease subgroups. Further stratifying the miscellaneous diseases group did not show differences in NK activity between patients with mucosal or physical changes of the intestine. Linear regression analysis was performed to evaluate individual differences in NK activity against Caco-2 target cells compared with the activity measured against the standard NK-sensitive K-562 cells in both 4-h and 18-h assays. Levels of individual cytotoxicity between 4-h and 18-h assays, irrespective of using either K-562 or Caco-2 target cells, were found to correlate remarkably well (Fig. 2). Hence individuals with high or low levels of cytotoxicity in the standard 4-h 51Cr-releaseassay against K-562 were also found to have high or low cytotoxicity levels, respectively, in the K-562 18h, Caco-2 4-h, and Caco-2 18-h assays as well.

*

DISCUSSION Non-MHC-restricted cytotoxicity or NK activity in LPMC preparations from intestinal mucosa could be demonstrated

0

\

m

%.

W

n

0

,\

\

0

X Cytotoxicity Caco-2 t8h, 500:l 0

X Cytotoxicity COCO-2 4h, 500:l

X Cytotoxicity COCO-2 18h, 500:l 0

X Cytotoxicity K-562 18h, 500:l


w

0 w 0


Normal IBD Mucosa and N K Activity

in each individual patient by the use of a high effector to target cell ratio. This was earlier described by Gibson et al. (l), but curiously, it has not yet been introduced as a common procedure to study mucosal NK activity despite immunohistochemical and functional experiments demonstrating NK cells to be present in low proportions. It therefore remains difficult to understand the biologic significance of such a sparsely distributed cell in mucosal immunopathology. So far, most studies have reported difficulties in detecting NK activity in LPMC or LPL preparations, or only very low levels have been found when using regular (50:l) effector to target cell ratios (6, 12, 14, 15). Significant levels of cytotoxicity, however, were obtained not only by introducing high effector to target ratios but also by preincubating the effector cells, prolonging the assay duration (2, 13, 18), and through enriching for NK cells by panning (6). The use of a high E/T ratio enabled us to detect differences in the functional contribution of lamina propria NK cells with regard to inflammatory bowel diseases. Such an observation could have been obscured in previous studies due to the use of low E/T ratios, thus failing to detect significant NK activity levels. Our group recently described the cell responsible for the spontaneous cytotoxic activity by LPMC preparations in both 4-h and extended 18-h assays as having NK-cell characteristics (3). Depletion of cells with specific NK-cell markers from LPMC resulted in abrogation of cytotoxicity in both 4-h and 18-h assays. However, these effector cells are of a somewhat different phenotype-that is, all CD56+ and some CD16+-than the equivalent functional cells in peripheral blood which co-express both CD56 and CD16 differentiation antigens. Moreover, CD3+ T cells were not found to contribute to this type of non-MHC restricted cytotoxicity in the intestinal lamina propria, as demonstrated by functional depletion experiments. Cytotoxicity levels by LPMC were reported to be low in IBD when using low E/T ratios, as shown by Fiocchi et al. ( 5 ) and MacDermott et al. (13), and no differences were reported between Crohn’s disease and ulcerative colitis. By using the low effector to target ratio in a prolonged 24-h assay, Beeken et al. (18) found significant levels of cytotoxicity in LPMC preparations of both normal and inflamed mucosa from Crohn’s disease patients. In addition, when combining the high effector to target cell ratio and a prolonged 24-h assay, Gibson & Jewel1 (4) detected significant levels of NK activity in LPMC preparations, but they were found to be independent of the underlying disease, drug therapy, or anatomic origin of the effector cells. In contrast to our study presented here, they did not describe differences

Fig. 2. Correlations between cytotoxic activity of individual lamina propria mononuclear cell preparations against K-562 and Caco2 target cells in 4-h or 18-h 51Cr-releaseassays. ( n = number of individuals tested; R = correlation coefficient; p = significance of

correlation.)

1003

between Crohn’ disease and ulcerative colitis patients, and almost all tissue samples included in their study were taken from inflamed mucosa. In general, studies on NK activity in IBD comprised inflamed mucosa, and LPMC from unaffected normal mucosa from carcinoma patients or from patients of a miscellaneous group served as controls. In the present study LPMC isolated from normal mucosa distant from the tumor region in carcinoma patients as well as normal mucosa from a miscellaneous disease group were both found to be appropriate control LPMC populations to compare with normal IBD mucosa. We found that even histologically normal IBD mucosa shows alterations in NK activity-that is, high levels in Crohn’s disease versus low levels in patients with ulcerative colitis. These observed abnormalities were strengthened by the findings that NK activity in Crohn’s disease patients or ulcerative colitis patients were also higher or lower, respectively, than in the control carcinoma patients. Beeken et al. (18) did not find differences in cytotoxicity levels between normal and abnormal mucosal tissue in Crohn’s disease, which might favor our observation that even normal mucosa conceals alterations in NK activity in this disease, perhaps preceding inflammation. In a preliminary study we found differences in the proportions of both CD16- and CD56positive cells between Crohn’s disease and ulcerative colitis which were suggestive for the difference in NK activity that we report here. In addition, Hirata et al. (19) found higher proportions of Leu7’ NK cells by immunohistochemical detection in lamina lymphocyte populations from both normal and inflamed mucosa in IBD. There is evidence from peripheral blood studies that differences in antigen expression designate different NK-cell subsets with regard to their lytic efficacy (7, 20, 21). In relation to the immunopathology of IBD it is important to observe that considerable differences in cytotoxic activity as reported here may be associated with selective compartmentalization of NK cells in the intestine, a phenomenon that has been described earlier by Gibson et al. (22). It needs to be recognized, however, that despite interesting differences in NK activity within the IBD patient group, the biologic relevance of such a sparsely distributed cell in the intestinal mucosa remains to be elucidated, in particular with regard to the pathogenesis of IBD. Nevertheless, the present finding of a difference in NK activity between UC and CD patients may support the concept that we are dealing with different disease processes. At this stage it cannot be completely ruled out that the origin of the tissue from which the LPMC have been isolated contributes to the observed differences in NK activity between CD and UC. However, a few important observations prompted us to consider the differences to be disease-related. First, colonic cells from two patients with Crohn’s colitis also showed markedly higher NK activity levels than those from control colonic tissue of the carcinoma or miscellaneous diseases group and much higher activity than those of the UC group. Secondly, cells from small-


1004

E. A . F. van To1 et al.

bowel mucosa of a single UC patient showed much lower NK activity than the LPMC isolated from normal ileum of CD patients. Autologous epithelial cells from the colon have been used as ‘physiologic targets’ to study spontaneous or antibodyarmed cytotoxicity (23,24), but colon target cells are rarely used in peripheral blood or mucosal NK-activity studies. Another colon tumor target cell line, RPMI-4788, was used in studies by MacDermott et al. (13). Despite the low levels of cytotoxicity against K-562 targets, they obtained similar results when using these colon target cells. The latter cell line has been reported to be more sensitive for LPMC cytotoxicity than K-562 in studies performed by Beeken et al. (18). They postulated that RPMI-4788 cells were better targets for evaluation of mucosal cytotoxicity. The lower susceptibility of Caco-2 cells to NK activity in the short 4-h assay of the present study seems to be in accordance with previous findings using peripheral blood mononuclear effector cells (25). Despite the difference in susceptibility of Caco-2 cells to NK activity as reported here, we found nice correlations between the activities against both target cell lines in both 4-h and 18-h assays. Caco-2 cells can therefore be used as target for detecting differences in individual NK activity by LPMC, which may be of interest in mucosal cytotoxicity studies because these cells originate from colonic epithelium, a putative target for local NK activity. Our observation and those reported by Beeken et al. (18) and Taunk et al. (26) indicate that the mucosal NK cell has a different profile of target cell recognition from its peripheral blood counterpart. Furthermore, changes of mucosal nervous inncrvation and peptide concentration have been reported to be associated especially with Crohn’s disease (27,28). Since we found gastrointestinal peptides to be able to modulate NK activity (29,30), it would be interesting to include target cells from intestinal origin as well in studies of mucosal NK activity in IBD. ACKNOWLEDGEMENTS We appreciate the technical assistance of Annie van der Zon and Marij Mieremet-Ooms. REFERENCES 1. Gibson PR, Dow EL, Selby WS, Strickland RG, Jewell DP.

Natural killer cells and spontaneous cell-mediated cytotoxicity in the human intestine. Clin Exp Immunol 1984;56:438-44. 2. Gibson PR, Hermanowicz A, Jewell DP. Factors affecting the spontaneous cell-mediated cytotoxicity of intestinal mononuclear cells. Immunology 1984;53:267-74. 3. van To1 EAF, Verspaget HW, Pena AS, Elzo Kraemer CV, Lamers CBHW. The CD56 adhesion molecule is the major determinant for dctecting non-MHC-restricted cytotoxic mononuclear cells from the intestinal lamina propria. Eur J Immunol 1992;22:23-9. 4. Gibson PR, Jewell DP. Local immune mechanisms in inflammatory bowel disease and colorectal carcinoma. Natural killer cells and their activity. Gastroenterology 1986;90:12-9. 5. Fiocchi C, Youngman KR, Yen-Lieberman B, Tubbs RR.

Modulation of intestinal immune reactivity by interleukin 2. Phenotypic and functional analysis of lymphokine-activated killer cells from human intestinal mucosa. Dig Dis Sci 1988; 33:1305-15. 6 . Shanahan F, Brogan M, Targan S. Human mucosal cytotoxic effector cells. Gastroenterology 1987;92:1951-57. 7. Lanier LL, Le AM, Civin CI, Loken MR, Phillips JH. The relationship of CD16 (Leu-11) and Leu-I9 (NKH-1) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J Immunol 1986;136:4480-86. 8. Hersey P, Bolhuis R. ‘Nonspecific’ MHC-unrestricted killer cells and their receptors. Immunol Today 1987;8:233-9. 9. Whiteside TL, Herberman RB. The role of natural killer cells in human disease. Clin Immun Immunopathol l989;53: 1-23. 10. Robertson MJ, Ritz J. Biology and clinical relevance of human natural killer cells. Blood 1990;76:2421-38. 11. Oldham RK. Natural killer cells: history, relevance, and clinical applications. Nat Immun Cell Growth Regul 1990;9:297-312. 12. MacDermott RP, Franklin GO, Jenkins KM, Kodner IJ, Nash GS, Weinrieb IJ. Human intestinal mononuclear cells. I. Investigation of antibody-dependent, lectin-induced, and spontaneous cell-mediated cytotoxic capabilities. Gastroenterology 1980;78:47-56. 13. MacDermott RP, Bragdon MJ, Kodner IJ, Bertovich MJ. Deficient cell-mediated cytotoxici ty and hyporesponsiveness to interferon and mitogenic lectin activation by inflammatory bowel disease peripheral blood and intestinal mononuclear cells. Gastroenterology 1986;90:6-11. 14. Fiocchi C, Tubbs RR, Youngman KR. Human intestinal mucosal mononuclear cells exhibit lymphokine-activated killer cell activity. Gastroenterology 1985;88:625-37. 15. Hogan PG, Hapel AJ, Doe WF. Lymphokine-activated and natural killer cell activity in human intestinal mucosa. J Immunol 1985;135:1731-8. 16. Bull DM, Bookman MA. Isolation and functional characterization of human intestinal mucosal lymphoid cells. J Clin Invest 1977;59:96674. 17. Fiocchi C, Battisto JR, Farmer RG. Gut mucosal lymphocytes in inflammatory bowel disease. Isolation and preliminary functional characterization. Dig Dis Sci 1979;24:705-17. 18. Beeken WL, Gundel RM, St. Andre-Ukena S, McAuliffe T. In vitro cellular cytotoxicity for a human colon cancer cell line by mucosal mononuclear cells of patients with colon cancer and other disorders. Cancer 1985;55:1024-29. 19. Hirata I, Berrebi G , Austin LL, Keren DF, Dobbins WO. Immunohistological characterization of intraepithelial and lamina propria lymphocytes in control ileum and colon and in inflammatory bowel disease. Dig Dis Sci 1986;31:593-603. 20. Lanier LL, Phillips JH. Evidence for three types of human cytotoxic lymphocyte. Immunol Today 1986;7:132-4. 21. Nagler A, Lanier LL, Cwirla S, Phillips JH. Comparative studies of human FcRIII-positive and negative natural killer cells. J Immunol 1989;153:3183-91. 22. Gibson PR, Verhaar HJJ, Selby WS, Jewell DP. The mononuclear cells of human mesenteric blood, intestinal mucosa and mesenteric lymph nodes: compartmentalization of NK cells. Clin Exp Immunol 1984;56:445-52. 23. Shorter RG, McGill DB, Baker RC. Cytotoxicity of mononuclear cells for autologous colonic epithelial cells in colonic disease. Gastroenterology 1984;86:13-22. 24. Gibson PR, van de Pol E, Pullman W, Doe WF. Lysis of colonic epithelial cells by allogenic mononuclear and lymphokine activated killer cells derived from peripheral blood and intestinal mucosa: evidence against a pathogenic role in inflammatory disease. Gut 1988;29:1076-84. 25. Aparicio-Pages MN, Verspaget HW, Pena AS, et al. In vitro cellular cytotoxicity in Crohn’s disease and ulcerative colitis: relation with disease activity and treatment, and the effect of recombinant gamma-interferon. J Clin Lab Immunol 1989;29: 11%24. 26. Taunk J , Roberts AI, Ebert E. Spontaneous cytotoxicity of human intraepithelial lymphocytes against epithelial cell tumors. Gastroenterology 1992;102:69-75.

Normul I B D Mucosa and N K Activity

27. Sjolund OB, Schaffalitzky de Muckadell OB, Fahrenkrug J , Hlkanson R, Peterson BG, Sundler F. Peptide containing nerves fibers in the gut wall in Crohn’s disease. Gut 1983;24: 724-33. 28. Koch TR, Carney JA, Go VL. Distribution and quantitation of gut neuropeptides in normal intestine and inflammatory bowel disease. Dig Dis Sci 1987;32:369-76. 29. van To1 EAF, Verspaget HW, Lamers CBHW. Neuropeptide


Received 3 March 1992 Accepted 14 July 1992

1005

regulation of cell-mediated cytotoxicity against human tumor cells. Neuropeptides 1990;16:25-32. 30. van To1 EAF, Verspaget HW, Pena AS, Jansen JBMJ, AparicioPages MN, Lamers CBHW. Modulatory effects of VIP and related peptides from the gastrointestinal tract on cell mediated cytotoxicity against tumor cells in vitro. Immunol Invest 1991; 20:257-67.