In situ effector mechanisms in rat kidney allograft rejection. I. Characterization of the host cellular infiltrate in rejecting allograft parenchyma.

The infiltrating inflammatory cells were recovered with collagenase and DNase from rejecting rat kidney allografts and autografts in conditions where the enzyme treatment did not affect the expression of subclass-specific surface markers. As the differential distribution of the inflammatory cells in the dispersate was similar to the distribution of inflammatory cells in tissue imprints, and as any major blood contamination was excluded, we consider the results representative of the composition of the in situ infiltrate. At the peak of rejection on Day 6 after the transplantation, approximately 30% monocytes, 17% macrophages, 31% lymphocytes, 6% (T) lymphoblasts, and 10% (B) plasmablasts and plasma cells were present in the graft. The blast cell response, pathognomonic to immune activation, was less prominent in the recipient spleen, blood, and lymph nodes. Twenty-three percent of the infiltrating lymphocytes expressed the (T-cell-specific) Pta.A.1 surface antigen(s) and 14% were surface Ig positive. The remaining lymphocytes were double-negative “null cells.” In preparative cell electrophoresis most of the allograft-infiltrating lymphocytes carried the low electrophoretic mobility, characteristic to resting B cells. Approximately 70% of allograft-infiltrating macrophages and 50% of infiltrating monocytes but only 30% of the monocytes present in the recipient spleen expressed the Fc receptor to IgG, suggesting an activation (or increase in avidity) of this receptor during the influx of mononuclear cells into the site of inflammation and during maturation of monocytes into tissue macrophages. There was a strong in situ proliferative activity, far stronger than in the central lymphatic system of the recipient rat. After 1 hr in vivo pulse labeling with [3H]thymidine 24% of the infiltrating inflammatory cells carried the label. Most of the labeled cells were blasts or lymphocytes, but a small albeit distinct number of labeled monocytes were also present in situ. In contrast to the recipient spleen, where most of the labeled lymphoid cells had a high electrophoretic mobility of resting T cells, in the infiltrate most of the labeled lymphoid cells had a slow mobility of resting B cells.     

Effector mechanisms in allograft rejection I. Assembly of “sponge matrix” allografts

This communication describes a model that makes it possible to quantitatively recover allograft-infiltrating cells in a functionally viable state. The model is based on the use of an inert spongious matrix tissue into which fibroblasts of strain “A” are grown. Upon transplantation to an allogeneic “B” strain host, graft-directed killer cells infiltrate the sponge. Mere physical compression of the sponge releases virtually all infiltrating cells. More than 90% of the viable cells infiltrating the sponge graft are thus recovered. About 15% of the infiltrating cells are blasts, about 35% are lymphocytes, and about 30% are monocytes and macrophages. The rest are predominantly granulocytes. The allograft-infiltrating cells display an immunologically specific cytolytic response to relevant ⁵¹Cr-labeled target cells in vitro. The infiltrating cells are much more efficient killer cells than spleen or draining lymph node cells. The allograft-infiltrating cells are thus functionally intact and are recovered without enzymatic treatment by mere mechanical means. We therefore consider the sponge matrix model a suitable and reproducible method for studying allograft-infiltrating cells.     

Lymphoid cell subclasses in rejecting renal allograft in the rat

We have quantitated the frequency of lymphoid cell subsets in rejecting renal allografts and in the spleen of the allograft recipient during drug-unmodified rejection in the rat. The number of inflammatory (white) cells in the graft was approximately similar to the number of white cells responding to the allograft in the recipient spleen. The inflammatory population of the graft consisted of lymphoid cells and mononuclear phagocytes, with increasing numbers of macrophages toward the end of rejection. Analysis of allograft cellular dispersates with monoclonal antibodies directed to the lymphoid cell subsets demonstrated that although the majority of allograft-infiltrating lymphocytes were T cells, a sizable B-cell proliferation and immunoglobulin synthesis was associated with the inflammatory response of rejection. Within the T-cell subset, the T suppressor/killer cells predominated in the graft whereas the predominant lymphoid cell subset responding to the allograft in the recipient spleen was the T helper cell.     

In situ effector pathways of allograft destruction. 1. Generation of the "cellular" effector response in the graft and the graft recipient

Inflammatory leukocytes of DA-to-WF rat renal allografts displayed significant cytolytic activity to natural killer (NK) target cells on Day 2 after transplantation. The NK activity, which was associated with large granular lymphocytes in discontinuous Percoll gradients, peaked on Day 4 and disappeared rapidly thereafter. Coincident with the presence of NK activity in the graft, a decrease in NK activity in the recipient spleen was observed. Low NK activity was also recorded in WF-to-WF autografts. The cells displaying direct cytotoxic activity to donor (but not to recipient) strain peritoneal exudate target cells (PEC) were associated with the T suppressor/killer lymphocytes in affinity chromatography. They appeared in the graft between Days 2 and 4, peaked between Days 6 and 8 and disappeared slowly thereafter. In the spleen the cytotoxic T lymphocyte (CTL) activity appeared later and it reached a maximum between Days 16 and 20 before decreasing. In the blood distinct CTL activity was seen only from Days 16-20 onwards, after the graft had been rejected. No CTL activity was recorded in the graft, blood, or spleen of an autograft recipient. Addition of donor-directed post-transplantation antibody (antibody-dependent cellular cytotoxicity, ADCC) had a slight enhancing effect on the cytotoxic activity of inflammatory leukocytes up to Day 5. After this time, added antibody had a blocking effect on direct CTL activity. No ADCC activity was recorded in the inflammatory population of an autograft. On the contrary, high levels of ADCC activity to donor strain PEC were recorded in the spleens of both autograft and allograft recipients throughout the period of follow-up. The results demonstrate that at least three cellular effector pathways exist in an allograft: a strong natural killer cell component, a strong cytotoxic T lymphocyte component, and (possibly) a weak cell component participating in an ADCC type of cytotoxicity.     

Effector mechanisms in allograft rejection. II. Density, electrophoresis, and size fractionation of allograft-infiltrating cells demonstrating several classes of killer cells.

An analysis of killer cells infiltrating “sponge-matrix” allografts during rejection has been performed by preparative fractionation by density centrifugation, velocity sedimentation, and free flow cell electrophoresis and by the use of heterologous anti-T-cell sera. At the peak of rejection, 7 to 8 days after transplantation, the allograft is infiltrated by several classes of killer cells, most notably by non-T lymphocytes, monocytes/macrophages, and T lymphocytes. The predominant cell types capable of performing in vitro lysis of relevant target cells appeared to be monocytes and non-T lymphocytes. T lymphocytes formed only a minority of the killer cells at this stage of the response. In contrast, as also documented earlier, the predominant killer cells in the regional lymph nodes and the spleen of the graft recipient mice were T lymphocytes (blasts).     

In situ effector pathways of allograft destruction. 2. Generation of the "humoral" response in the graft and the graft recipient

The frequency of both immunoglobulin (Ig)-synthesizing and Ig-secreting B cells have been analyzed in DA-to-WF rat renal allografts (and in control WF-to-WF autografts). We have correlated the in situ B-cell responses with corresponding events in the central lymphatic system of the recipient. Intracellular IgM- and IgG-containing plasma cells appeared in an allograft (but not in an autograft) very shortly after the transplantation. The numbers of both cell types in situ was approximately equal, the highest numbers of each being found on Day 4 after transplantation. A similar early response was observed in the recipient's spleen, however, very few Ig-synthesizing cells were seen in the blood. Only a fraction of the Ig-synthesizing cells in the allograft were involved in immunoglobulin secretion. Thus, the recovery of IgG- and IgM-secreting cells from an allograft was 10 and 2% of intracellular IgG- and IgM-containing cells, respectively. It appears, therefore, that allograft-infiltrating Ig-synthesizing B cells either die or migrate elsewhere before secreting immunoglobulin. The B-cell response in the graft occurs very early and is disproportionally high when the very low frequency of B lymphocytes in the allograft is considered. The data provide no evidence for inflammatory B cells being an integral part of graft rejection. Indeed, the possibility remains that the inflammatory B-cell response observed during the rejection process represents a meaningless byproduct of the inflammatory response.     

Characterization of natural cytotoxic effector cells isolated from rat liver

Nonparenchymal liver cells (NPC) from normal untreated female Wistar/Furth rats were tested for natural cytotoxicity in a 4-hour 51Cr release assay against the murine lymphoma YAC-1, the murine mastocytoma P815, and the syngeneic rat mammary carcinoma TMT-081 tumor cell lines. NPC exerted strong cytotoxicity against all three target cells. In contrast, fresh spleen cells displayed cytotoxicity only against YAC-1, although after culture for 24 h at 37 degrees C cytotoxicity was displayed against all three target cells. Fresh spleen cells contained 2-15% large granular lymphocytes (LGL) as assessed by Giemsa staining whereas NPC contained 10-23% LGL and 10-25% Kupffer cells. Centrifugal elutriation produced fractions that were increased in one or the other of the cell types. More cytotoxic activity was observed in the fraction containing more LGL. The cytolytic activity of fresh spleen cells could be eliminated by either in vivo or in vitro treatment with anti-asialo-GM1 antiserum. On the other hand, the cytolytic activity of NPC was resistant to in vivo treatment, but was partially sensitive to in vitro treatment. Furthermore, the activity of cultured spleen cells was also partially sensitive to in vitro treatment. NPC and cultured spleen cells also were more resistant to suppression by prostaglandin E2 and nordihydroguaiaretic acid than fresh spleen cells. We conclude that LGL is mainly responsible for natural cytotoxicity of NPC and that some effector cells in NPC may be highly activated.     

In situ effector pathways of allograft destruction. 3. Plasminogen activator activity in rat renal allografts

The question of which cell components in a rejecting rat renal allograft secrete plasminogen activator (PA) has been analyzed. Although normal renal parenchymal cells also secreted PA, most of the PA in a renal allograft (and to a lesser extent also in an autograft) was produced by the inflammatory leukocytes. Fractionation at 1 g demonstrated that the inflammatory cell population responsible for the PA production in the allograft sedimented together with the large mononuclear phagocytes (macrophages). Fractions purified for small blast cells and large lymphocytes did not contain any PA activity but they were able to induce resting peritoneal macrophages to produce PA when cocultured in vitro. The results demonstrate that the allograft-infiltrating mononuclear phagocytes are "activated" in the sense that they secrete PA and that the activation of mononuclear phagocytes at the site of inflammation may be partially regulated by the inflammatory lymphoid cells.     

Immunologic responses to a murine mammary adenocarcinoma: in vitro production of specific killer cells is dependent on active T lymphocytes.

The mode of production of specifically armed monocytic killer cells was investigated with the T1699 mammary adenocarcinoma in syngeneic DBA/2 mice. After overnight in vitro incubation of cells from the spleen but not from the lymph nodes, blood, or from the peritoneal cavity produced specific killer cells. The activation of spleen cells was inhibited by pretreatment with anti-theta serum and C; however, already activated specific killer cells were not sensitive to the same treatment. Removal of phagocytic cells did not significantly affect the cytotoxicity of the splenic killer cells whereas removal of rayon-wool adherent cells greatly reduced both the total cytotoxicity, and to a lesser extent, the cytotoxicity indices. Overnight co-cultivation of normal peritoneal-exudate cells with the lymph node cells from tumor-bearers, although neither class of cells alone was cytotoxic to T1699 cells in vitro, produced specific monocytic killer cells, through steps dependent on active T lymphocyte function. Culture spupernatants of tumor-bearer's spleen cells also contained factor(s) which induced cytotoxicity mediated by normal peritoneal-exudate cells against T1699 cells in vitro; and the production of the factor(s) was also inhibited by pretreatment of the spleen cells with anti-theta serum but not by anti-mouse IgG or anti-mouse whole gamma-globulins serum and C.     

Prolongation of murine skin allograft survival by immunologic manipulation with anti-interleukin 2 receptor antibody

Administration of a rat monoclonal antibody (M7/20) directed against the murine interleukin 2 (IL 2) receptor in combination with sublethal x-irradiation of the recipient significantly enhanced the survival of skin allografts, both when the grafts were MHC disparate from the hosts and when only minor histocompatibility differences were present compared with untreated controls. No prolongation in graft survival was seen with either treatment alone at the dose employed. M7/20 and x-ray-treated allograft recipients also displayed significantly decreased alloantigen-specific reactivity against donor-strain spleen cells in both delayed-type hypersensitivity and cytotoxicity assays. Thus, such combination treatment reduces expression of host immune reactivity against graft determinants by several criteria. This work provides additional evidence that monoclonal antibodies directed against the IL 2 receptor may be useful in clinical transplantation.     

Differential sensitivity to natural cell-mediated cytotoxicity of two rat colon adenocarcinoma variants differing in their tumorigenicity: identification of the effector cells as natural killer cells

Summary DHD/K12 TRb (PROb) and DHD/K12 TSb (REGb) are two cancer cell variants originating from the same rat colon adenocarcinoma. They differ in their tumorigenicity: when inoculated into syngeneic BDIX rats, PROb cells induce progressive tumors whereas REGb cells induce tumors which always regress. As previously described, there is an inverse relation between their tumorigenicity and their susceptibility to NCMC mediated by syngeneic spleen or peripheral blood lymphocytes: PROb cells are significantly less sensitive to NCMC than REGb cells. This suggests a role for NCMC in the regression of REGb tumors. In this work the BDIX NCMC effector cells active in vitro against REGb cells were identified as NK cells according to four criteria: (1) efficacy in a 4-h ⁵¹Cr release assay, (2) sensitivity to anti-asGM1 antibody plus complement, (3) LGL morphology, and (4) ability to bind with the same affinity REGb and YAC-1 cells. In spleen, these NK cells were heterogeneous with respect to their asGM1 surface density and their morphology. PROb cells were not lysed by these NK cells in a short-term cytotoxicity assay, but only in a 16-h assay. It was shown that PROb and REGb cells were bound with the same affinity by NK cells, thus they certainly differ in their ability to resist to NK lytic mechanisms. This difference could play a role in the different tumorigenicity of the two variants. Abbreviations used: NK, natural killer; NC, natural cytotoxic; NCMC, natural cell-mediated cytotoxicity; asGM1, asialo GM1; LL, large lymphocytes; LGL, large grnular lymphocytes; LAL, large agranular lymphocytes; PBMNC, peripheral blood mononuclear cells; E:T, effector to target cell ratio; C:H, cold to hot cell ratio; FBS, fetal bovine serum     

Thymus-dependent and thymus-independent effector functions of mouse lymphoid cells. Comparison of cytotoxicity and primary antibody formation in vitro

We have compared two effector functions, antibody formation and cytotoxic capacity in vitro, of mouse cells of various origin with special reference to the T lymphocyte dependence of these processes. We have used addition of PHA and coating of target chicken erythrocytes (CRBC) with antibody as the two means of inducing cytotoxicity. Antibody formation in vitro has been studied both against thymus-dependent sheep erythrocytes (SRBC) and thymus-independent (E. coli) antigens. Spleen cells from thymectomized, lethally irradiated bone marrow-, or fetal liver-repopulated mice were deprived of phagocytic cells by uptake of colloidal iron. They did perform better than normal spleen cells in the antibody-induced cytotoxicity and were also induced to cytotoxicity by PHA. PHA did not induce increased DNA synthesis in these T cell-deprived spleen cell preparations, which could not make primary antibodies to SRBC but were able to do so against E. coli antigens. Fresh bone marrow and fetal liver cells, deprived of phagocytic cells, were also induced into a highly efficient cytotoxicity by anti-CRBC as well as by PHA. Pretreatment of spleen cells with an alloantiserum (θ) against T lymphocytes reduced but did not abolish the PHA-induced cytotoxicity. In contrast, it did not affect the antibody-induced cytotoxicity. Such treated cells could not make antibodies to SRBC but could do so against E. coli. Pretreatment of spleen cells with a heteroantiserum (MBLA) against mouse B lymphocytes completely abolished all cytotoxic- and antibody-forming abilities of the cells, although experiments with combinations of θ-treated and MBLA-treated cells suggested that the MBLA treatment had left behind a significant portion of helper T cells needed for the in vitro antibody response. From these data we conclude, as have others, that the antibody-induced cytotoxicity is independent of T lymphocytes. It can be induced in immature precursor cells from fetal liver or bone marrow, and these cells may also become cytotoxic on interaction with PHA. However, in normal spleen cells, at least part of the PHA-induced cytotoxicity is T cell dependent. Some preliminary data suggest that this PHA-induced cytotoxicity of normal spleen cells may be a joint process between T lymphocytes and other cells.     

Random recirculation of small T lymphocytes from thoracic duct lymph in the mouse

The migration of ⁵¹Cr-labeled nylon-wool separated mouse thoracic duct T cells has been followed in order to determine whether there is a circulation of small (nondividing) T cells through the small intestine. Approximately 6% of the injected dose of T-TDL localized in the small intestine (minus Peyer's patches). Experiments revealed that this gut-localizing cell population consisted almost entirely, if not exclusively, of lymphoblasts present in mouse T-TDL. When lymphoblasts and small lymphocytes from mouse T-TDL were separated by velocity sedimentation, and the migration of separated fractions was studied, we found large cells (66% blasts) migrated well to the gut but poorly to the lymph nodes, whereas small cells (2% blasts) showed minimal migration to the gut but localized randomly in lymph nodes and spleen. The in vivo distribution of small cells from T-TDL was similar to that of T-PLN. Furthermore, the recirculatory patterns of both ⁵¹Cr-labeled T-TDL and T-PLN were found to be identical as accessed by their rate of recovery in the thoracic duct lymph of recipient mice. These results support the notion that the vast majority of T-TDL and T-PLN are part of a common pool of recirculating T cells which recirculate randomly through lymph nodes and spleen and not the small intestine.     

Enhancement by interferon of natural killer cell activity in mice.

Injection of mice with several interferon inducers, Newcastle Disease virus, polyinosinic-polycytidylic acid and tilorone resulted in an increase in spleen cell cytotoxicity for ⁵¹chromium-labeled mouse YAC tumor target cells in 4-hr in vitro assays. This increase in spleen cell cytotoxicity was abrogated by injection of mice with potent anti-mouse interferon globulin. Inoculation of mice with mouse interferon (but not human leucocyte or mock interferon preparations) also resulted in a marked enhancement of spleen cell cytotoxicity. The extent of enhancement of spleen cell cytotoxicity was directly proportional to the amount of interferon injected and a significant increase was observed after inoculation of as little as 10³ to 10⁴ units of interferon. An effect could be detected as soon as 1 hr after injection of interferon. The increase of spleen cell cytotoxicity after inoculation of an interferon inducer was not due to a localization and accumulation of cytotoxic cells in the spleen but reflected a general increase in cytotoxic cell activity in various lymphoid tissues (except the thymus). The splenic cytotoxic cells from interferon or interferon-inducer-injected mice had the characteristics of natural killer (NK) cells since (i) interferon enhanced spleen cell cytotoxicity in athymic (nu/nu) nude mice, (ii) classical spleen cell fractionation procedures by nylon wool columns, anti-Thy 1.2 serum plus complement, anti-Ig columns, and depletion of FcR+ rosette-forming cells, failed to remove the effector cells generated in vivo or in vitro. Therefore like NK cells, interferon-induced cytotoxic cells lack the surface markers of mature T and B lymphocytes, are not adherent, and are devoid of avid Fc receptors. Furthermore like NK cells, the spleen cells from interferon-treated mice lysed various target cells (known for their sensitivity to NK cells) without H-2 or species restriction. Incubation in vitro of normal spleen cells with interferon also resulted in an increase in cytotoxicity for YAC tumor cells. We conclude that interferon acts directly on NK cells and enhances the inherent cytotoxic activity of these cells.     

Sensitivity of rat heart endothelial and myocardial cells to alloimmune lymphocytes and to alloantibody-dependent cellular cytotoxicity.

We have attempted to define the structural target components for the two cytotoxic mechanisms in rejecting rat heart allograft: antibody-independent (direct) and antibody-dependent (ADCC) cellular cytotoxicity. Immune spleen cells and alloantibody were obtained at 1 and 2 weeks after heart allotransplantation, respectively. The cytotoxicity of immune spleen cells, with and without alloantibody, was tested against endothelial- and myocardial-enriched heart cell populations. We found endothelial cells to be sensitive to direct cellular cytotoxicity, most likely mediated by T lymphocytes, while myocardial cells were sensitive to ADCC. Both reactions were shown to be immunologically specific.     

Redistribution of rat renal allograft-responding leukocytes during rejection: 1. Model

We describe a model that enables us to trace the traffic of allograft-responding inflammatory leukocytes to and from the graft without handling of these cells in vitro. At different times after transplantation, the kidney transplant pedicle—including the artery, vein, and draining lymphatics—is clamped. The allograft-responding leukocytes are labeled by a [³H]thymidine pulse either in situ or in the systemic lymphoid organs of the recipient. Fifteen minutes later the pulse is chased with a 1000-fold excess of cold thymidine, and the clamp is opened. The animals are sacrificed 18 hr later, when a balance between the synthesis of new labeled leukocytes from the originally labeled ones and dilution of intracellular label has been achieved. This model was used to analyze the allograft-responding inflammatory cell traffic to and from a renal transplant performed across the major histocompatibility complex in the rat. A sizable traffic was observed to both directions: After systemic injection of label only 0.008 × 10⁶ labeled cells × hr^?1 were found to emigrate into a kidney allograft (control). Already on the third day after transplantation—when the in situ inflammatory response is still at its beginning—more than 0.3 × 10⁶ labeled cells × hr^?1 migrated from the host to the allograft and 1.6 × 10⁶ labeled cells × hr^?1 left the allograft to the recipient spleen. The first figure is several-fold higher than any previous estimate. The findings emphasize the systemic nature of the antiallograft inflammatory response.     

Comparison of the tube leukocyte adherence inhibition test with the lymphocyte stimulation assay in a rat transplantation model.

The tube leukocyte adherence inhibition test (LAI) was analyzed in a heterotopic heart transplantation model in BN and Wag/Rij rats. Seven, fourteen, and twenty-two days after transplantation spleen cells of the recipient rat were tested for recognition of histocompatibility antigens in crude membrane extracts prepared from normal rat tissue. It was found that addition of fetal calf serum affected the cell adherence considerably and that there was a large difference in the percentage of adherent cells from different animals. In a comparative study the lymphocyte stimulation assay proved to be more sensitive and reproducible than the LAI in this rat system.     

Cytological and functional analysis of inflammatory infiltrates in human malignant tumors: III. Further functional investigations using cultured autochthonous tumor cell lines and freeze-thawed infiltrating inflammatory cells

Short-term monolayer cultures, dominated by cells with malignant characteristics, were established from human tumors displaying an unusually strong host-inflammatory response. Upon repeated testings in the ⁵¹Cr release cytotoxicity assay, blood leukocytes were frequently cytotoxic (a) to autologous and allogeneic tumor cells, without any apparent restriction as to tumor origin or HLA type, and (b) to the so-called natural killer (NK) target cells. The anti-tumor cytotoxicity disappeared with time. The in situ inflammatory cells, freshly isolated or recovered from the deep freeze, did not display any type of cytotoxic activity. Nor were they notably suppressive to either type of blood leukocyte cytotoxic activity in the ⁵¹Cr release assay. Cytological analysis demonstrated that the “large granular lymphocytes” (LGL), known to be largely responsible for the NK activity in man, were prominent in the blood but not in the inflammatory infiltrate. These preliminary observations suggest that lack of cytotoxic activity in situ correlates with the absence of effector cells in the inflammatory infiltrate.     

Differential effect of hybrid resistance on the localization of virus-immune effector T cells to spleen and brain

The hybrid resistance (Hr) effect operates in the lymphocytic choriomeningitis (LCM) in vivo transfer model to inhibit both the level of cytotoxicity T lymphocyte (CTL) generation in spleen and the induction of inflammation in cerebrospinal fluid (CSF). The effect is seen when LCM virus-immune T cells that are homozygous for H-2D ^b are injected into virus-infected, immunosuppressed recipients that are heterozygous for this allele, or into radiation chimeras that express an appropriate F₁ phenotype. Evidence that Hr to T -cell transfer is cell-dose-dependent and tends to diminish with age was found in both chimeric and normal F₁ mice. Inhibition of the capacity of injected T cells to cause meningitis is a more sensitive measure of Hr than is the further stimulation of CTL effectors in recipient lymphoid tissue. The injection of large numbers of H-2^b virus-immune T cells into (H-2 ^k X H-2 ^bF₁H-2 ^k) virus-infected recipients did not induce any cellular extravasation into CSF, though potent H-2^b-restricted CTL effectors were generated in recipient spleen. Evidence of minimal inflammatory process was found in one experiment where these chimeras were given a comparable dose of (H-2 ^b X H-2 ^d)F₁ immune spleen cells. Development of this Tcell-mediated immunopathological process depends essentially on the expression of the appropriate H-2 restriction element on radiation-resistant host cells which, in this case, presumably constitute part of the physiological barrier between blood and CSF.     

Identification of effector cells for TNFalpha-mediated cytotoxicity against WEHI164S cells

WEHI164S cells were found to be very sensitive targets for in vitro killing in a 6-h culture when liver or splenic lymphocytes were used as effector cells in mice. Of particular interest, a limiting cell-dilution analysis showed that effector cells were present in the liver with a high frequency (1/4,300). In contrast to YAC-1 cells as NK targets, perforin-based cytotoxicity was not highly associated with WEHI164S killing. The major killer mechanism for WEHI164S targets was TNFalpha-mediated cytotoxicity. By cell sorting experiments, both NK cells and intermediate T cells (i.e., TCR(int) cells) were found to contain effector cells against WEHI164S cells. However, the killer mechanisms underlying these effector cells were different. Namely, NK cells killed WEHI164S cells by perforin-based cytotoxicity, TNFalpha-mediated cytotoxicity, Fas ligand cytotoxicity, and other mechanisms, whereas intermediate T cells did so mainly by TNFalpha-mediated cytotoxicity. These results suggest that TNFalpha-mediated cytotoxicity mediated by so-called natural cytotoxic (NC) cells comprised events which were performed by both NK and intermediate T cells using somewhat different killer mechanisms. Intermediate T cells which were present in the liver were able to produce TNFalpha if there was appropriate stimulation.