Peripheral, mucosal, and tumour-infiltrating components of cellular immunity in cancer of the large bowel

Morphology of sporadic colorectal cancer with DNA replication errors

BACKGROUND

Up to 15% of colorectal cancers are characterised by DNA microsatellite instability (MIN), shown by the presence of DNA replication errors (RERs).

AIMS

To identify pathological features that are discriminating for colorectal cancer (CRC) showing extensive MIN.

SUBJECTS

A prospective series of 303 patients with CRC and no family history of either familial adenomatous polyposis or hereditary non-polyposis colorectal cancer.

METHODS

DNA was extracted from fresh tissue samples and the presence of MIN was studied at nine loci that included TGF beta RII, IGFIIR, and BAX. The 61 cases showing RERs were compared with 63 RER negative cases with respect to a comprehensive set of clinical and pathological variables. Predictive utility of the variables was tested by decision tree analysis.

RESULTS

Twenty seven patients with CRC showed extensive RERs (three loci or more) (RER+) and 34 had limited RERs only (28 = one locus; 6 = two loci) (RER+/-), yielding a bimodal distribution. RER+ cancers differed from RER- and RER+/-) cases. Tumour type (adenocarcinoma, mucinous carcinoma, and undifferentiated carcinoma) (p = 0.001), tumour infiltrating lymphocytes (p = 0.001), and anatomical site (p = 0.001) were the most significant of the discriminating variables. Algorithms developed by decision tree analysis allowed cases to be assigned to RER+ versus RER- and +/- status with a global sensitivity of 81.5%, specificity of 96%, and overall accuracy of 93%.

CONCLUSIONS

Pathological examination of CRC allows assignment of RER+ status; assignment is specific and relatively sensitive. Conversely RER- and RER+/- CRC are indistinguishable.

Infiltration of mononuclear inflammatory cells into primary colorectal carcinomas: an immunohistological analysis

Local immunoregulation mediated by mononuclear tumour-infiltrating cells is considered of importance for tumour progression of colorectal cancer, although the balance between immunosuppressor and cytotoxic activities is unclear. Colorectal cancers from 26 patients were investigated using a panel of monoclonal antibodies in order to identify subsets of mononuclear inflammatory cells and to study their pattern of distribution in relation to tumour stage and cytotoxic immune reactivity against the tumour. In all but five tumours, mononuclear cells, lymphocytes or monocytes were present in fairly large numbers, particularly in the stroma. The infiltration of CD4+ mononuclear cells predominated over the CD8+ subset. Infiltration near the tumour cells was found in four cancers only. Stromal infiltration of CD11c+ macrophages was found in all but eight tumours. Small regressive areas, in which the histological architecture of the tumours was broken down, were found in 17 tumours with intense or moderate infiltration by CD4+ lymphocytes or CD11c+ macrophages. Probably this destruction of tumour tissue was caused by cytotoxic activity of the tumour-infiltrating mononuclear cells. In Dukes' class A and B tumours, CD4+ lymphocytes predominated over CD4+ cells with macrophage morphology, but the latter were increasingly found in Dukes' class C and D disease. The occurrence of MHC II-positive macrophages and lymphocytes in different Dukes' classes was similar to that of CD4+ cells. In contrast to this, CD11c+ and CD11a+ cells were more frequent in Dukes' A and B class tumours compared with Dukes' C and D. Four out of nine tumours of the latter stages showed a poor inflammatory reaction. The interpretation of our results is that the subsets of tumour-infiltrating mononuclear cells change with advancing Dukes' class and that the local immune control is gradually broken down in progressive tumour growth, even if some cytotoxic activity is still present.

A comprehensive multi-institutional study on postoperative adjuvant immunotherapy with oral streptococcal preparation OK-432 for patients after gastric cancer surgery. Kyoto Research Group for Digestive Organ Surgery

The current study was designed to compare the effects of oral administration of the streptococcal preparation, OK-432, as an adjuvant immunotherapy versus those of intradermal administration of OK-432 on the survival of patients after surgery for gastric cancer. The patients were stratified into two groups after surgery: a curative surgery stratum and a palliative surgery stratum. Then the patients in each stratum were randomly assigned into three groups: an oral placebo group, an oral OK-432 group, and an intradermal OK-432 group. All of the patients were given fluoropyrimidines orally in combination with OK-432 or placebo for 2 years after surgery. A total of 1011 patients were registered between 1982 and 1985, and 970 patients were eligible for statistical analysis. The survival rate of the oral OK-432 group was significantly higher than those of the other two groups after curative surgery. There were no significant difference in the survival rates between the three groups after palliative surgery, however. The effect of oral OK-432 was quite pronounced in patients after curative surgery for stage II to IV gastric cancer, especially in those patients with regional node involvement. Furthermore, it was found that the spleen is necessary for effective immunotherapy with oral OK-432, because the survival rate of the oral OK-432 group was significantly improved in patients whose spleens were preserved, when compared with splenectomized patients. These results demonstrate that oral adjuvant immunotherapy with OK-432 is beneficial after curative surgery for gastric cancer.

Anti-CD3- or PHA-induced cytotoxicity in human intraepithelial and lamina propria lymphocytes

Morphological data in humans and rodents and functional data of intraepithelial lymphocytes of mice support the idea that cytotoxic cells are a significant population of the human mucosa. Previously it was shown that human IEL have no spontaneous cell-mediated cytotoxicity and that human LPL cells have anti-CD3-mediated cytotoxicity. We confirm that most individuals have anti-CD3- or PHA-mediated cytotoxicity of LPL. In IEL we do not find cytotoxic function in short-term assays. There is no difference between patients with colon cancer and inflammatory bowel disease.

Immune-competent cells of regional lymph nodes in colorectal cancer patients: II. Immunohistochemical analysis of Leu-7+ cells

Distribution of Leu-7+ cells in cancer tissues and regional lymph nodes was immunohistochemically examined to estimate the role of NK cells in colorectal cancer patients. Leu-7+ cells were rarely observed both in the primary and distant metastatic lesions, but the number of these cells was large in the germinal center of the lymph nodes. Studies of Leu-7+ cell population in the lymph node indicated that intermediate nodes in n0 and n1 groups showed significantly higher values than those in the control group. Comparative study of the population in identical patients revealed a significantly higher rate in the intermediate nodes than in the paracolic nodes both in n0 and n1 groups. These data suggest that Leu-7+ cells might be related to the defense mechanism of regional lymph nodes against tumor.

Orally administered streptococcal preparation, OK-432 augments the antitumor immunity of patients with gastric or colorectal cancer

A streptococcal preparation, OK-432, was orally administered at a dose of 5 KE to patients with gastric or colorectal cancer for 7-14 days before their operations, and its immunomodulatory effects on peripheral blood lymphocytes (PBL), regional node lymphocytes (RNL) and tumor infiltrating lymphocytes (TIL) were assessed. The group treated with OK-432 included 8 gastric and 6 colorectal cancer patients, and the control group included 8 gastric and 8 colorectal cancer patients. The NK cell activity of PBL was significantly augmented by the oral administration of OK-432, and the proportions of Leu 7+ and Leu 11+ cells in PBL also increased. The responses of PBL and TIL to autologous tumor extracts in the presence of interleukin-2 were enhanced after the oral administration of OK-432. The proportion of OKT8+ cells in PBL increased after treatment with oral OK-432, whereas the proportion in RNL significantly decreased. These results indicate that oral OK-432 affects NK and T cells and may augment the antitumor immunity of patients with gastrointestinal cancer.

Impaired local natural killer cell activity in human colorectal carcinomas

The present study was undertaken to study natural killer (NK) cell activity in patients with colorectal cancer at peripheral and local levels. Mononuclear cells were isolated from uninvolved colorectal mucosa, tumor tissue and peripheral blood, and tested against the colon carcinoma cell line CaCo-2 and the erythroleukemia cell line K-562. Peripheral blood NK cell activity from the patients showed similar levels compared with healthy controls, whereas, mononuclear cells of tumor tissue were found to have a significantly decreased NK cell activity compared to the normal intestinal mucosa (P less than 0.01). No relation was found between the NK cell activity and the advancement of the disease according to the Duke's stage. Interferon-gamma (IFN-gamma) stimulated the NK cell activity of the mononuclear cells from blood, mucosa and tumor. However, the increase of NK cell activity after IFN-gamma stimulation was lower in the tumor compared to the mucosa (P less than 0.02). The lectin, phytohaemagglutinin, increased the cytotoxicity of mononuclear cells from blood, mucosa and tumor to a similar level. These results suggest that patients with colorectal tumors exhibit a normal NK cell activity in peripheral blood and intestinal mucosa; however, a diminished NK cell activity exists at the tumor level. Although mononuclear cells isolated from the tumor have a normal response to lectin stimulation they show hyporesponsiveness to IFN-gamma stimulation with regard to their NK cell activity.

Characterization of activated lymphocytes in colon cancer

The percentages of activated lymphocytes in colon cancers were compared to clinical features of the tumors and to functional characteristics of the tumor-infiltrating lymphocytes (TIL) in order to evaluate the role of activated TIL in controlling tumor growth. The TIL isolated from colon cancers contained 80 +/- 6% T cells [cluster designation (CD) 2+], 42 +/- 9% CD4+ cells, 27 +/- 9% CD8+ cells, 17 +/- 5% cells (surface immunoglobulin or SIg+), 6 +/- 2% null cells (CD2-, SIg-), 6 +/- 4% Leu 7+ cells, 1 +/- 0% macrophages (CDw 14+), and no plasma cells (PCA+). The lamina propria lymphocytes (LPL) isolated from adjacent colonic mucosa contained a similar distribution but with fewer CD8+ (10 +/- 5%) and Leu 7+ (0.6 +/- 0.3%) lymphocytes (P less than 0.05). More TIL than LPL expressed the interleukin-2 (IL-2) receptor (9 +/- 8% vs 1 +/- 0.8% CD25+ cells; P less than 0.05) and the transferrin receptor (4 +/- 5% vs 0.3 +/- 0.6% T9+ cells; P less than 0.05). Those TIL expressing activation antigens were CD2+, SIg-. The percentage of TIL that were CD25+ did not correlate with the extent of tumor spread, with the degree of tumor differentiation, nor with the percentage of HLA-DR+ tumor cells. However, there were significantly more CD25+ TIL from tumors located in the left colon, particularly small lesions, than from tumors located in the right colon (P less than 0.05). The proliferation of TIL preparations cultured with medium alone, with mitogens, or with IL-2 did not vary according to the percentage of CD25+ lymphocytes in the TIL. T cells may be activated by foreign surface determinants on tumor cells, particularly in small, left-sided lesions.

Human mucosal T-cell cytotoxicity

Non-major histocompatibility complex-restricted cytotoxicity triggered by antibodies to the CD3 component of the human T-cell receptor complex is thought to be an indirect measure of in vivo primed cytotoxic T-cell activity. We have used this technique to examine the lytic activity of freshly isolated T cells from noninflamed human colonic mucosa. Anti-CD3-triggered T-cell (anti-CD3-T) cytotoxicity was found in all mucosal specimens studied. The mucosal anti-CD3-T effectors do not have Fc receptors for immunoglobulin G, and are therefore distinct from T gamma cells, which mediate antibody-dependent cellular cytotoxicity. The surface antigen phenotype of mucosal anti-CD3-Ts is CD2+, CD3+, CD8+, CD4-, CD16-, and Leu7-. In contrast, peripheral blood anti-CD3-T effectors are Leu7+. Although non-major histocompatibility complex-restricted, mucosal anti-CD3-T cytotoxicity has considerable target specificity, which differs from that of natural killer and lymphokine-activated killer cells. The profile of target cell susceptibility and the inhibitory effects of anti-CD45 antibody suggest that the CD45 molecule on the effector cell may be an important determinant of anti-CD3-T sensitivity. As anti-CD3-triggered lysis may be a marker of in vivo primed mucosal T cells of undetermined antigen specificity, this technique might have important implications in inflammatory bowel disease, where the antigen(s) inciting the mucosal immune reactivity is not certain.

Human mucosal cytotoxic effector cells

Human intestinal lamina propria mononuclear cells have been shown to mediate mitogen-induced cellular cytotoxicity, antibody-dependent cellular cytotoxicity, and lymphokine activated killer cell function. However, although natural killer cells have been demonstrated in the gut mucosa of rodents, recent reports found little or no spontaneous cytotoxic activity in the lamina propria of the human gut. Using the natural killer cell-related monoclonal antibody NKH-1, which has not previously been applied to studies of mucosal killer cell function, we have shown by immunofluorescence that 2%-3% of enzymatically dispersed lamina propria lymphocytes are NKH-1+. A "panning" technique was then used to enrich for the NKH-1+ cells. Panned cells were consistently greater than or equal to 80% NKH-1+ by indirect immunofluorescence. Unlike their counterparts in the peripheral blood, the mucosal NKH-1+ cells were Leu-11-. Although unseparated lamina propria lymphocytes failed to exhibit natural killer activity against K562 targets in 4-h chromium release assays at effector to target ratios of up to 100:1, the NKH-1+ cells were cytolytically active at ratios of less than 5:1. Mucosal lymphocytes depleted of natural killer cells (NKH-1-) exhibited cytotoxic activity when cultured for 72 h with interleukin-2. The precursors of the lymphokine culture activated phenomenon were NKH-1-, Leu-11-, T4-, T3-, T11+, and T8+. Although lamina propria T3+ cells did not exhibit spontaneous or culture activated cytotoxicity, they were shown to exhibit nonspecific anti-CD3 (anti-T3)-induced T-cell cytotoxicity. In conclusion, functional natural killer and lymphokine activated killer cells are both present in the human gut mucosa and represent distinct populations of cytotoxic cells. In addition, anti-CD3-induced cytotoxicity is a feature of mucosal T cells. These mucosal killer cell subsets differ phenotypically from those previously described in the peripheral blood.

Definite spontaneous cell-mediated cytotoxicity and HNK-1 cells in the human large intestine

Spontaneous cell-mediated cytotoxicity (SCMC) and the marker of natural killer (NK) cells mediating SCMC of the human large intestine were studied. Lamina proprial lymphoid cells (LPL) were isolated by sequential dithiothreitol-EDTA-collagenase treatment of the gut specimen. SCMC was measured by the chromium release method. Target cells included P4788 in monolayer, a cell line derived from colon cancer, Chang cells in monolayer, and K562 in suspension. Target cells in monolayer including colon cancer cell line were chosen because they were thought to be more appropriate to assess SCMC for lymphoid cells in the solid organ. While lower compared to cytotoxicities (CT) by peripheral blood lymphoid cells (PBL), define CT were observed in LPL against all three targets. NK cells marker was studied both on LPL by an indirect fluorescent antibody method and on the gut tissue by indirect immunoperoxidase staining using anti HNK-1 monoclonal antibody which defines virtually all NK cells. HNK-1 positive (HNK-1 +) cells were identified in both methods. HNK-1 + cells were observed in the epithelium, lamina propria, and lymph follicle with or without germinal centers. These results clearly demonstrated the presence of SCMC and HNK-1 + cells in the human large bowel.

Examination of the low proliferative capacity of human jejunal intraepithelial lymphocytes

The proliferation of human jejunal intraepithelial lymphocytes (IEL) was examined to determine how it differed from that of peripheral blood (PB) T lymphocytes. The IEL were mainly T lymphocytes of the cytotoxic-suppressor (T8+) phenotype. They demonstrated lower proliferative responses to various stimuli (2,501 +/- 565 ct/min with phytohaemagglutinin; PHA) compared to unseparated PB T lymphocytes (73,678 +/- 2,495) or the T8+ subset (68,939 +/- 10,053 ct/min) (P less than 0.001). This low proliferative response was also a characteristic of the T8+ T lymphocytes in the lamina propria (4,606 +/- 1,226 ct/min) but not the T4+ subset (43,447 +/- 10,188 ct/min) (P less than 0.05). These findings were not due to isolation techniques or to differences in kinetics. Mixing experiments revealed that the IEL did not contain cells which suppressed proliferation. In addition, the IEL could be stimulated by mitogens, as they produced the same amount of interleukin 2 (IL-2) and IL-2 receptors as did PB T lymphocytes. Although the lectin-induced proliferative response of IEL was unaltered by the addition of autologous macrophages and minimally increased by IL-2, it was markedly enhanced by the addition of sheep red blood cells (SRBC). The enhancing effect of SRBC was not due to T cell recognition of xenogenic antigens on the erythrocytes since neither allogeneic non-T lymphocytes nor other xenogenic erythrocytes produced the same effect. Both intact SRBC and membrane fragments from osmotically lysed cells augmented lymphocyte proliferation. Thus, jejunal IEL could be activated by mitogen and proliferated as much as PB T lymphocytes if exposed to a membrane component found on SRBC.

Local immune mechanisms in inflammatory bowel disease and colorectal carcinoma. Natural killer cells and their activity

Mononuclear cell (MNC) populations isolated from intestinal mucosa, mesenteric lymph nodes, and peripheral blood have been assessed for their natural killer (NK) (Leu-7+) cell proportions and NK cell activity against K-562 erythroleukemic target cells. In peripheral blood, normal proportions of Leu-7+ cells were found in patients with Crohn's disease or ulcerative colitis, whereas increased proportions in colorectal carcinoma may have been related to the higher mean age of these patients. Low proportions of Leu-7+ cells (less than 3%) were present in intestinal MNCs in Crohn's disease, ulcerative colitis, colon cancer, and miscellaneous intestinal diseases. All groups of patients had diminished NK activity of peripheral blood MNCs compared with a group of healthy controls. Intestinal NK cell activity from histologically normal mucosa correlated with autologous peripheral blood NK cell activity (p less than 0.001) but no such correlation was seen for patients with inflammatory bowel disease. Mucosal or nodal NK cell activity showed a wide range of activity but did not relate to the underlying disease, mucosal histopathology, drug therapy, or, in patients with cancer, Dukes' grading. Intestinal MNCs from all patient groups responded to stimulation with lymphoblastoid interferon, except in a small number of patients whose unstimulated activity was not detectable. In conclusion, the NK cell on intestinal mucosa behaves similarly in various intestinal diseases. However, the disparity between NK activity of autologous peripheral blood and intestinal MNCs in inflammatory bowel disease highlights the difficulty in extrapolating peripheral blood findings to mucosal immune events.

In vitro cellular cytotoxicity for a human colon cancer cell line by mucosal mononuclear cells of patients with colon cancer and other disorders

In vitro cellular cytotoxicity of mononuclear cells of intestinal mucosa and peripheral blood for a colon cancer cell target was measured in patients with colon cancer and other disorders requiring resection. Four- and 24-hour cytotoxicity assays were conducted using selenium 75 (75Se)-labeled RPMI-4788 human colon cancer target cells grown in culture. In the cancer group mean cytotoxicity was 30.4% at 24 hours for peripheral blood effectors and 8.0% for effectors from normal mucosa. Values in patients with Crohn's disease were 10.4% for blood and 17.2% for effectors from normal mucosa, and 13.6% and 18.5%, respectively, for blood and abnormal mucosa. Values in patients with other diseases were 25% for blood and 14.7% for mucosa. Mean cytotoxicity at 4 hours did not exceed 6.4% for any group, and assays in autologous serum gave results similar to tests in calf serum. In additional studies, K 562 chronic leukemia cells were somewhat more sensitive to lysis than RPMI-4788 by blood mononuclear cells, but there was no lysis of K 562 by mucosal populations that were cytotoxic for RPMI-4788. There was no competitive inhibition by either target cell for the other. It was concluded that 75Se RPMI-4788 colon cancer cells are suitable targets for evaluating in vitro cytotoxicity by intestinal mucosal cells and that mucosal cytotoxicity in patients with colon cancer is depressed compared to cytotoxicity by peripheral blood effectors.

Human intestinal mucosal mononuclear cells exhibit lymphokine-activated killer cell activity

Previous investigations have shown that mononuclear cells present in human intestinal mucosa possess cytotoxic properties that are unique and different from those of cells circulating in the peripheral blood. We have further explored the cytolytic capacity of human intestinal mucosal mononuclear cells using both phenotypic and functional criteria. Using an immunoperoxidase technique on gut frozen sections, we failed to identify cells bearing surface markers displayed by natural killer cells and recognized by the monoclonal antibodies Leu7, Leu11, and Leu15. Freshly isolated human lamina proprial mononuclear cells (LPMC) were unable to lyse K562 target cells, even after using a variety of experimental conditions which included depletion of adherent cells, fractionation by Percoll gradients or panning with monoclonal antibodies, and treatment with a prostaglandin synthetase inhibitor or interferon-gamma. In contrast, when LPMC were cultured in the presence of the lymphokine interleukin 2 (IL 2), they displayed high levels of cytotoxicity against both K562 and Daudi target cells. Such cytolytic activity appeared after 1 or 2 days of culture, was dependent on the amount of IL 2 added to the cultures, was independent of plastic adherent cells, and could be inhibited by agents that block proliferation. Interferon-gamma, when used under experimental conditions identical to those adopted for IL 2, was unable to induce any significant cytotoxicity by LPMC or enhance the level of killing obtained by stimulation with IL 2 alone. This IL 2-induced, nonspecific cytotoxicity of LPMC probably represents a form of lymphokine-activated killer cell function similar to that recently described for human peripheral blood lymphocytes. In view of the absence of morphologic and functional evidence for natural killer cells in human intestinal mucosa, the phenomenon of lymphokine-activated killer cell activity displayed by LPMC may represent an alternate cytotoxic function potentially relevant to intestinal mucosa immunity.

Isolation of intestinal mononuclear cells: factors released which affect lymphocyte viability and function

Whether toxic or immunomodulatory factors are released during the collagenase digestion phase of the isolation of mononuclear cells from human intestinal mucosa was investigated by assessing the effect of the collagenase supernatant on the viability and natural killer activity of normal peripheral blood mononuclear cells. Three hours' incubation in collagenase supernatant suppressed natural killer activity by 25 +/- 4% and decreased the viability of peripheral blood mononuclear cells by 11 +/- 2%. The ability of collagenase supernatants to kill 51Cr-labelled peripheral blood mononuclear cells over four hours was assessed in 16 collagenase supernatants, eight of which produced lysis of 20 +/- 4%. There was no ultrastructural evidence of early degenerative changes in the viable intestinal mononuclear cells fresh from the isolation process or in peripheral blood mononuclear cells incubated in collagenase supernatant. Because prostaglandins are known to inhibit natural killer activity, PGE was measured in 20 collagenase supernatants by radioimmunoassay and found to be high at 27.5 +/- 4.0 ng/ml. Addition of indomethacin to the collagenase medium, however, failed to abolish the inhibitory effect of collagenase supernatant on natural killer activity and did not increase the natural killer activity of isolated intestinal mononuclear cells. The release of cytotoxic and immunomodulatory factors during the isolation of intestinal mononuclear cells indicates the necessity for careful assessment of the potential effects of the isolation process on any function being examined and casts doubt upon the relationship between in vitro findings and in vivo functional capabilities.

Natural killer cells and spontaneous cell-mediated cytotoxicity in the human intestine

Spontaneous cell-mediated cytotoxicity (SCMC) has been investigated in mononuclear cells (MNC) isolated from intestinal mucosa and autologous peripheral blood from human subjects. The proportion of cells with the NK-K phenotype (Leu 7+) were substantially lower in intestinal MNC than in autologous peripheral blood. SCMC of K-562 target cells when tested at an effector to target (E:T) ratio equivalent to that used for peripheral blood MNC was markedly deficient in intestinal MNC. This was not due to the effect of EDTA and collagenase used in the isolation process. However, at high E:T, ratios, significant cytotoxicity was demonstrated for most intestines examined probably reflecting a low proportion of effector cells within the intestinal MNC population. SCMC in both intestinal and autologous peripheral blood MNC were similarly related to the Leu 7+:T ratios used in the assay indirectly suggesting that the Leu 7+ cell may be responsible for the observed cytotoxicity. It is concluded that the apparent functional difference between similar cells derived from different sites may be largely related to differing proportions of effector cells. The findings indicate the need for specific definition of the effector cell and suggest that intestinal SCMC in health and various disease states requires re-appraisal.