Fas-ligand (CD95-L; Fas-L) Fas (CD95) Fas (apoptosis) 年 了 不 度 Fas Fas-L 力 不 Fas/Fas-L T IL-10Fas/Fas-L 不 年 Fas signal-mediated apoptosis 度降 不 不 力 U-118, HeLa, A549, Huh-7 MCF-7, HepG2. PI3K/Akt FasPI3K/Akt Fas/FasL mediated apoptosis is one of the strategies for immune escape. However, we showed previously that tumor infiltrating T cells expressed IL-10 in contact with tumor tissue. In this study, we evaluated the Fas-trigged apoptosis in T cells in contact with tumor cells mimicking the cell-to-cell interactions at regions of tumor. When Jurkat cells were cocultured with U-118, HeLa, A549, and Huh-7, they were less susceptible to the Fas-mediated apoptosis as indicated by amelioration of loss of mitochondria membrane potential, decrease in caspase-8, -3 activation and DFF-45/ICADpoly cleavage, and reduction in breakage of DNA when compared to Jurkat cells cultured alone. This event was independent of the expression of Fas-L and de novo transcription. However, tumor cell lines MCF-7 and HepG2 did not confer Jurkat cells resistance against Fas-mediated apoptosis. Protein kinase inhibitors for ERK, p38, JNK, PKC, PKA did not affect on this protection. PI 3 K inhibitor, LY294002, completely abolished the protection provided by cell-to-cell contact. Accordingly, phosphorylation of AKT is enhanced. Together, our results indicate that the stimulation of β integrin signal of T cells by contact with tumor cells may trigger a novel protective signaling through PI3K/AKT pathway in T cells against Fas-mediated apoptosis. keywordsfas, apoptosis, tumor tissue matrix, PI3K/Akt pathway Introduction: Fas receptor (Fas, CD95, APO-1)-mediated apoptosis plays an important role in controlling cell numbers during embryo development and immune homeostasis. Upon trimerization by its ligand (Fas-L), Fas signal is propagated through a caspase cascade by the activation of caspase-8 which results in apoptosis of Fas+ cells. CD3+ TILs infiltrating in the regions of tumor become Fas-sensitive when they are isolated and challenged with Fas activating agents in vitro. However, there are abundant TILs accumulating around the tumor tissues, such as nasopharyngeal carcinoma, cervical cancer, liver cancer, or ovarian cancer. TILs may play an adverse role during tumor formation, as to express vascular endothelial growth factor (VEGF), which is correlate with the stage of tumor invasion. Recently, we have demonstrated that coculture with glioma cells 1
will induced transcription of the IL-10 gene in peripheral T cells and Jurkat T cells through Fas signaling. These findings raise the questions whether tumors high-jack immune cells to create a tumor growth-favoring environment, eg. evading immune surveillance by elevating IL-10 and promoting angiogenesis to bring in nutrition. Recent study underscores the importance of extracellular matrix (ECM) in informing life and death decisions. Integrins are α/β heterodimeric membranes receptor proteins server the external sensor by interacting with appropriate ECM ligands, through which cells respond to the dynamic changes of distinctive tissue environment constituted by a specific set of ECM. By inflammation and tumorigenesis, the local status of ECM may be profoundly perturbed. Tumors can secret various ECM components and metalloproteases in abundance, thereby altering the makeup of the cells immediate environment. The elevated amount of extracellular proteases of tumor masses is positively correlated to the high aggressiveness. Tethering of epithelial cells to proper ECM enables these cells to assemble cytoskeletons, to gain apical-basal polarity, and to resist all manner of proapoptotihc stimuli including Fas, TRAIL, TNF, and the chemotherapeutic agents. In this paper, we investigated the mechanism for accumulation of immune cells in tumor site. Fas-induced apoptosis in T cells was examined in the context of contact with tumor cells, mimicking integrin-ecm interactions occurred at the tumor site. We provide evidence showing that tumor contact initiate cross-talk of death and survival signals in T cells using an in vitro coculture system. Our data demonstrate that interaction of integrin and cell matrix plays a role in regulating apoptosis in T cells. Results and discussion Coculture with tumor cells prevented Fas-mediated apoptosis in T cells. Despite being susceptible to Fas-mediated apoptosis, tumor infiltrating T cells accumulate in the environment capable of activating Fas signal at the regions of tumors. We attribute the accumulation of T cells in tumor site to a delayed apoptosis by survival signal activated by tumor tissue components. By using a coculture system allowing T cells in contact with tumor cell layer mimicking the scenario at tumor regions, we demonstrated that cell-cell contact with tumor delayed the Fas signal-induced apoptosis in Jurkat cells (Fig.1A). In parallel with the reduction in DNA fragmentation, we found that coculture with glioma cells also suppressed the activation of caspase cascade in Jurkat cells (Fig 2). Caspase-8 is the initial caspase in Fas signaling and consequently turns on the caspase cascade. Then, caspase-3 will be cleaved to activate its protease activity and digests its down stream substrate DFF-45/ICAD. We also showed that this protective ability was tumor cell line-specific. Even cells from same tumor type, eg. hepatoma Huh-7 and HepG2, could have different protection ability; the former was protective, but the latter was not. The physiological relevance of the results was confirmed in activated primary human peripheral T cells (Fig. 1b). Our findings may provide a mechanistic rational to explain the accumulation of T cells only within some tumor tissues. It is to note that most tumor cells used in this study, namely U118MG, HeLa, Huh-7, A549, and HepG2 express Fas receptor and its ligand. However, they were not sensitive to Fas antibody, nor induced apoptosis in Fas-sensitive 2
Jurkat T cells in a 24 h-coculture. Particularly, down regulating FasL of glioma cells did not affect apoptosis in Jurkat cells in coculture. Our results differ from those deduced from FasL overexpression systems, in which FasL directly caused apoptosis in infiltrating T cells. A possible explanation for this discrepancy is that FasL in overexpression systems or on colon cancer is very strong and thus initiate death program of infiltrating T cells. When the FasL gene is under the control of an intrinsic promoter, the amount of FasL on tumor cells is not sufficient to induce drastic apoptosis. Fig. 1. Coculture with tumor cells prevented Fas-mediated apoptosis in T cells. Cells of different origins, including U-118, HeLa cells, NIH3T3, MCF-7, and HepG2, were grown to 90% confluence in 6-well plate. About 3 x 10 5 fresh Jurkat T cells were added to each tumor cultures and incubated for 1 h allowing cell contact. Apoptosis was then induced by treatment with 10 ng/ml anti-fas CH-11 antibody for 24 h. Apoptotic cells were stained by PI showing as subg 0/1 population in flow cytometric analysis (A). Values shown are means with standard deviations (mean ± SD) obtained from three independent experiments. Activation-induced cell death in peripheral T cells was also analyzed (B). Peripheral T cells were isolated from peripheral venous blood and maintained in RPMI 1640 medium supplemented with 1 ng/ml IL-2 for 5 days (NT groups). For activation, cells were treated with 2.5 µg PHA for 16 h and then maintained with 1 ng/ml IL-2 for 5 days (AT groups). T cells were then cocultured with U-118 glioma for 1 h followed by treatment with CH11 (500 ng/ml) for 24 h. Apoptotic cells were stained by PI. Values shown are the mean SD of duplicates. Fig. 1 (A) (B) Fig. 2. The activation of caspase-8, -3 and the cleavage of PARP, DFF-45/35 during Fas-mediated apoptosis were down-regulated by Glioma. Jurkat cells were cocultured with glioma and MCF-7 in presence of 10 ng/ml CH-11 anti-fas antibody for 12 h. Jurkat cells were harvested and total protein extracts were subjected to Western blot analysis detecting the 56-KDa procaspase-8 and its 42/44-KDa cleavage product, 35-KDa procaspase-3 and its 15-17-KDa cleavage product, native 116-KDa PARP and its 85-KDa cleavage product, and the cleavage of DFF-45/ICAD. β-actin served as a control for protein loading. Arrows indicate the cleavage products of caspases. Cocultured with glioma cell did not affect surface Fas on Jurkat cells and No new proteins 3
were required for this inhibition. Glioma cells secrete MMPs, thus will potentially digest the surface Fas on Jurkat cells in coculture. To verify this possibility, we have measured the surface Fas on Jurkat cells being cultured in glioma cells-conditioned media or allowing direct contact with glioma cells. Neither glioma cells-conditioned media nor direct contact with glioma cells caused decrease in surface Fas on Jurkat cells. In addition, we investigated whether new protein synthesis was essential for reducing CH-11-induced cell death of Jurkat T cell, by using ActD to inhibit mrna transcription in Jurkat T cells before cocultured with glioma cells. Cell death by combined treatment with CH-11 and ActD was reduced to 17.2% under cocultured with glioma suggesting that no any new proteins were produced and participated in this protective mechanism. Integrins involved in suppression of the Fas-mediated apoptosis in Jurkat cells. We have excluded the possible involvement of soluble factors secreted by tumor cells using Transwell unit (Fig. 3). Only direct cell-to-cell contact conferred Jurkat cells resistance against Fas activation. Some tumors, eg. MCF-7 and HepG2, assuming with similar ECM constitutes, did not provide protection suggesting additional negative signal existed for those cells. Interestingly, the putative protective ECM molecules were not necessary syngenic, as mouse fibroblast NIH3T3 exhibited similar protective effect for human Jurkat cells. It seems that the ECM molecules for survival signal are conserved in evolution. Furthermore, preventing engagement of tumor ECM and integrins of Jurkat cells by blocking antibodies for β1 and β2 resulted in complete restoring the Fas sensitivity of Jurkat cells. Jurkat cells expressed both β1 and β2 integrins. The glioma cell contact-associated protection was diminished in the presence of blocking antibodies for β1 and β2 integrins (Fig.4). Together, we conclude that β-integrin-matrix interaction plays an essential role. Fig 3. Fig. 4 Fig. 3. Direct cell-cell contact is required to suppress the Fas-mediated apoptosis in Jurkat cells. U118(V) or U118(R), U118MG-derived glioma cells, were cultured in low wells of a transwell plate for 48 h. Jurkat cells were added to both upper and low wells and incubated for 1 h. Only Jurkat cells in low well got contact with glioma cells. Apoptosis was induced by CH-11 (10 ng/ml) for 24 h and detected by PI stain/flow cytometric analysis. Values shown are the mean SD of three independent experiments. Fig. 4. β1 and β2 integrins involved in suppression of the CH-11-induced apoptosis in Jurkat cells. 4
To block the integrin activation, Jurkat cells were pretreated with 10 mg/ml anti-β1 or anti-β2 blocking antibody for 3 h. After remove of the excess antibody, Jurkat cells were cocultured with glioma for 1 h. Apoptosis was induced by CH-11 (10 ng/ml) for 24 h and detected by stain with annexin V-conjugated-FITC antibody/flow cytometric analysis. Values shown are the mean of two independent experiments Integrin-down stream signals implicated in cell viability include PI3K and MAPK pathways. Allan et al. showed that activation of ERK MAPK pathway inhibited caspase-9 activity by direct phosphorylation. These reports suggested that the ERK/MAPK and p38 MAPK pathway might be critical for some apoptotic processes mediated through caspases. However, multiple lines of evidence indicates that MAPK pathway does not contribute to survival of Jurkat cells in response to cell matrix interaction. On one hand, after inhibiting MAPK activities by inhibitors specific for ERK, JNK, or p38, the tumor contact-associated protection against Fas signaling remained in action (Fig. 5). On the other hand, contact with tumor cells, regardless their protective abilities, equally enhanced the ERK phosphorylation of Jurkat cells. For instance, MCF-7 breast cancer cells stimulated Jurkat cells fast and strong phosphorylation of ERK. Yet, these Jurkat cells were still very sensitive to Fas signal-triggered apoptosis. Besides, Fas signaling also directly stimulated the phosphorylation of ERK and p38 to some extent in Jurkat cells when the apoptosis was on going. Phosphorylation of ERK1/2 was increased further in the presence of LY294002, which effectively abolished the tumor contact-associated protection. Fig.5. Tumor cell contact modulated MAPK and PI3kinases activities of Jurkat cells. Jurkat T cells were cocultured with glioma as described in Fig. 1. Apoptosis was induced by 10 ng/ml CH-11 antibody for 12 h. Phosphorylation of MAPK and PI 3 kinases was detected by Western blot analyses (A). To reduce the background activity of PI 3 kinase, Jurkat cells were pretreated with 25 µm LY294002 for 1 h and then cocultured with U-118V in the presence of 10 ng/ml CH-11 antibody for 12 h (B) (A) (B) PI 3 kinase/akt activation mediated the matrix-associated survival signal Coculture with glioma cells significantly elevated the phosphorylation of Akt of Jurkat cells. 5
Moreover, CH-11-trigged reduction in total Akt protein was subdued. LY294002 effectively altered the glioma cell contact-associated protection against Fas activation as compared to the matching control group (Fig. 6). Fig.6. PI 3 kinase involved in tumor cell contact-associated protection for Jurkat cells. Jurkat T cells were pretreated with MEK inhibitor PD98059 (25 µm), MAPK inhibitor U0126 (25µM ), JNK inhibitor SP600125 (25µM ), p38 inhibitor SB202190 (25µM), PKC inhibitor GF190203X (10µM), or PI 3 K inhibitor LY294002 (25µM) for 1 h. Jurkat cells were then cocultured with glioma for 1 h before apoptosis induction by CH-11 (10 ng/ml) for 24 h. Apoptosis was detected by PI stain/flow cytometric analysis. Values shown are the mean SD of three independent experiments. Together, our findings high-light a role of tumor matrix in providing survival signals for T cells. The PI3K/AKT pathway regulating the Fas-mediated apoptosis in T cells entering tumor tissue is illustrated in Fig. 10. Previously, we demonstrated that tumor might highjack immune cells to create an immune suppressive environment by producing IL-10. The data present here suggest that cell contact-pi3k/akt play a key role in delaying the apoptotic program in T cells providing a frame work for further analysis on tumor matrix and immune regulation. Autoimmunity is another pathological relevance of the findings above. Local integrin-matrix interactions in the inflamed tissues provide a survival signal lead to expansion of potentially autoreactive T cells present in the immune repertoire. If it holds truth, the integrin-matrix /PI3K/AKT pathway can be a potential target for disease management. 6