Myeloid Suppressor Cells (MSCs) Inhibit NK Cell Cytotoxicity Through Membrane-bound TGF-beta

The immune system is a remarkably versatile defense system comprising innate immunity and adaptive immunity. These two elements are able to employ an enormous variety of cells and molecules acting together in a dynamic network to protect hosts from invading pathogenic microorganisms and cancers. Antigen-presenting cells such as dendritic cells and macrophages, immune effector cells such as natural killer (NK) cells and cytotoxic T-lymphocytes (CTL), can recognize and kill or eliminate the invading pathogenic microorganisms and tumour cells though complex mechanisms. It is obvious that there are lots of immune defence responses against tumour cells in patients and tumor-bearing mice. Simutaneously, tumor induces immune tolerance to escape these immune attacks through secretion of cytokines and chemokines. Among which the generation and/or the expansion of inhibitory or regulatory cell populations (such as regulatory T cells, Treg) that negatively regulate immune cell functions have attracted much attention. In tumor-bearing mice, tumor progression is often associated with increased numbers of cells holding a myeloid morphology and sharing the Gr-1 and CD11b markers in thespleen, bone marrow, lymph node, tumor site, and tumor-infiltrating tissues. The vast majority of studies on these cells provide powerful evidence that they can be involved in the mechanisms for the negative regulation of immune responses in the tumor-bearing hosts.In healthy mice, sizeable number of CD11b⁺Gr-1⁺cells can be only found in the bone marrow, while, they (<4%) can be detected in small numbers in the blood and spleen. The CD11b⁺Gr-1⁺ myeloid cells, now designated as myeloid suppressor cells (MSCs), were first described in the early 1980s, which were previously referred to as immature myeloid cells (IMCs) or natural suppressor cells (NSCs) that constituted a cell population of undefined phenotype. MSCs represent a heterogeneous population comprising different myeloid cells at early stage of differentiation, including granulocytes, monocytes and a pool of immature cells of myelomonocytic lineage that retain the ability to form colonies on agar wich indicates that some of these immature myelomonocytic cells are myeloid precursors. Besides CD11b, a specific marker for myeloid cells of the macrophage lineage, and a marker for granulocytes Gr-1, the immature cells express varying amounts of CD31, ER-MP54, and ER-MP58, antigens that are characteristic of myeloid-cell differentiation. By contrast, they express abnormally low levels of MHC class II molecules and low/undetectable levels of costimulatory molecules such as CD86 and CD40. The counterparts of MSCs in human are mostly CD34⁺, which described by Pak et al for the first time in patients with head and neck cancer (HNC). Subsequently, human MSCs were characterized in the peripheral blood of patients with squamous cell carcinoma, HNC, non-small celllung cancer, and breast cancer. Human MSCs prevalently show a CD34⁺CD33⁺CD13⁺CD15^- immature phenotype and were described as two subsets according to the expression of HLA DR and CD11c molecules: immature monocyte/DCs and earlier myeloid differentiation stages. MSCs have both considerable heterogeneity and a high level of plasticity, confirmed by a line of in vitro and in vivo experiments. When cultured in vitro without cytokines, or after immortalization with retroviruses, these cells mainly give rise to CD11b⁺Gr1^-F4/80⁺ adherent, macrophage-like cells that retain their inhibitory features. They can also be induced to differentiate in vitro into mature granulocytes, macrophages and DCs by means of various cytokine combinations or followings in vivo adoptive transfer. The suppressive phenotype of MSCs extracted from tumor-bearing mice was enhanced by the addition to the culture of Th2 cytokines (such as IL-4 or IL-10). Conversely, MSCs co-cultured with Th1 cytokines (such as IFN-γ) could lose their inhibitory function but enhance antigen-specific T-cell cytotoxicity.The exact mechanism for the accumulation of MSCs in tumor-bearing mice and cancer patients is still unclear. However, large body of evidence indicates that disturbances in cytokine homeostasis resulted from intense immune activation due to the tumor growth or development can alter the equilibrium of this population leading to its accumulation in lymphoid organs and blood and ultimately influencing their maturation toward a suppressive phenotype. Several tumor-derived factors (TDFs) such as GM-CSF, IL-6, IL-10, and TGF-6, alone or in combination, promote not only MSCs recruitment but also their maturation toward an immature and fullyimmunosuppressive phenotype. Studies both in animal model and cancer patients, in vitro or in vivo, indicated that normal bone marrow cells cultured in conditioned media from tumor cells lines could give rise to immunosuppressive elements and the number of MSCs can dramatically decrease after removal of the primary tumor.The accumulation of MSCs has been observed in both cancer patients and tumor bearing animals in many different types of cancers and is associated with increased tumor burden and predicts poor survival rates. It has been demonstrated that these immune suppressive cells have capability to inhibit the T cell proliferative response induced by alloantigens, CD3 ligation, or various mitogens, and can also inhibit IL-2 utilization. The CD3ζ expression on T cells can be down-regulated by co-culturing with MSCs derived from tumor-bearing mice and can be restored by MSCs depletion from the cell suspension. Furthermore, the CD11b⁺Gr-1⁺ immune suppressive cell population was shown to be involved in inhibition of CD8⁺ generation and activity in the tumor host through apoptosis and a contact-dependent mechanism. Moreover, treatment of mice with the monoclonal anti-Gr-1 antibody was shown to reduce tumor progression and be able to induce a CTL response resulting in tumor eradication. On the other hand, adoptive transfer of MSCs from moribund tumor-bearing animals into mice bearing 5-day-old tumor caused rapid tumor progression and a relatively weak CTL response. Various mechanisms for the inhibitory function by MSCs have been proposed, such as arginine exhausted, release of NO, TGF-6 and so on. Whereas numerous findings correlate hyporesponse of T cells with systemic MSCs accumulation in tumor-bearing hosts, up to now, there is no report about their role inimpaired NK cell function.NK cells make up 5-10% of all circulating lymphocytes and are also found in peripheral tissues including the liver, peritoneal cavity and placenta. These cells described for their nonspecific cytotoxicity to pathogens and tumors, are the first line of defense against tumors. NK cells attack target cells by several mechanisms. The principal way NK cells used to eliminate tumor cells is by the release of cytoplasmic granules, kinds of complex organelles that combine specialized storage and secretory functions with the generic degradative functions of lysosomes. These granules derived from NK cells contain a lot of proteins, such as perform and granzymes, which lyse target cells. Large body of studies in perforin-deficient mice indicated that this protein is required for most NK cell cytotoxicity. NK cells also kill target cells by the expression of tumour necrosis factor (TNF)-family members, or through nitric oxide signalling. In some circumstances, Fc receptors on NK cells may mediate antibody-dependent cell-mediated cytotoxicity (ADCC) via binding to antibody-coated tumor cells. NK cells also produce an array of important cytokines, including IFN-γ, TNF, granulocyte-macrophage colony stimulating factor (GM-CSF), IL-5, IL-10 and IL-13, which contribute vital roles in immune regulation and influence both innate and adaptive immunity. Given no antigen-specific receptors on NK cells, the mechanisms by which NK cells recognize altered self-cells and distinguish them from normal cells through employing two different categories of receptors, one that binds class I major histocompatibility complex molecules and transmits inhibitory signals to NK cells, and another that couples ligands on tumours and delivers activation signals. As therecognition of tumor cells by NK cells is not MHC restricted, the activity of these cells is not compromised by the decreased MHC expression on some tumor cells. It is now clear that there are many different cell-surface receptors for activation signals and a number of different kinds for inhibitory ones. It is the balance between activating signals and inhibitory ones that is believed to enable NK cells to distinguish healthy cells from cancerous ones. Targeted killing by NK cells depends on the engagement of activating receptors on NK cells, among which the NKG2D appears to play a major role in NK cell-mediated cytotoxicity. NKG2D, one of NK cell activating receptors, is a type II transmembrane-anchored glycoprotein expressed as a disulfide-linked homodimer on the surface of all mouse and human NK cells. As most activating receptors, NKG2D is a multi-subunit receptor complex, whereby NKG2D signaling is mediated by specialized signaling adaptors. Mouse NKG2D can associate with two distinct adaptors DAP-10 and DAP-12/KARAP, while human NKG2D exclusively uses DAP-10. In murine NK cells, it is thus possible that one NKG2D homodimer simultaneously associates with one DAP-10 and one DAP-12 homodimer. Irrespectively, the regulated association of mouse NKG2D with two distinct signaling adaptors provides a potential for diversity of NKG2D-dependent cellular responses. NKG2D may recognize cell surface glycoproteins structurally related to MHC class I expressed by some cancer cells. It is important to be aware of that NK activity is stimulated by soluble factors such as IFN-a, IFN-β, IL-12 and IL-15 through delivery of additional NK-activating signals. However, TGF-6 derived from tumor can inhibit the lysis of NK cells by decreasing their NGG2D expression. Downmodulation ofNKG2D, which is associated with elevated levels of TGF-6, could be also observed in cancer patients and tumor-bearing mice.Alhough both accumulations of MSCs and impaired function of NK cells have been demonstrated in cancer patients and tumor-bearing animal models. It is unclear whether these two phenomenons exist in cause-and-effect. To clarify the possible relationship between the increasing MSCs and impaired NK cell function in tumors, we prepared the Hepatic tumor and 3LL experimental metastases models and investigated the changes of MSCs, NK cell cytotoxicity and NKG2D expression. With the development of Hepatic tumor progression, the property of MSC cells in splenocytes increased from 4-6% to 20-40%, while the cytotoxicity of NK cells against Yac-1 was down-regulated dramatically. To further identify the factors associated with NK cells function, we analyzed the expression of a line of activating receptors (Ly49D, CD94NKG2C and NKG2D) and inhibitory receptors (Ly49A, Ly49C and CD94NKG2A) together with a set of molecules that involved in the cytotoxicity of NK cells towards target cells, such as FasL, perform, granzyme B and Trail, in the different stages of tumor development. The down-regulation of NKG2D may contribute to the inhibition of NK cell cytotoxicity as the number of NKG2D⁺NK cells and the mean fluorescence intensity (MFI) of NKG2D declined gradually concomitant with the tumor progression. Nevertheless, there was no remarkable change being observed in the expression of FasL, perform, granzyme B, and Trail. Similar results have been obtained in the 3LL tumor-bearing mice at different stages. To confirm the critical role of NKD2D in NK cell mediated tumor cytotoxicity, weadded the specific monoclonal antibody against NKD2D in the coculture system of NK cells and Yac-1 cells. The anti-NKG2D mAb blocked the cytotoxicity of NK cells effectively. These data suggested that the increased frequency of MSCs and down-regulated NKG2D expression may be linked to the impaired NK cell function in tumor-bearing mice.To investigate if elevated MSC population in tumor-bearing mice was responsible for the decreased NK cell cytotoxicity, we incubated freshly isolated NK cells from normal mice with MSCs from tumor-bearing mice at a 1:1 ratio. Co-culture of NK cells with MSCs resulted suppression of NK cell cytotoxicity at 3h and a significant suppression of NK cell cytotoxicity at 12h. Simultaneously, we examined the change of NKG2D expression when the above-mentioned two kinds of cells were cultured together for 3h, 6h and 12h. And a similar tendency was observed. Hence, these data suggest that MSCs can inhibit NK cell cytotoxicity through down-regulation of NKG2D expression.We further investigated the underlying mechanisms for the down-regulation of NKG2D expression induced by MSCs. To identify whether soluble molecules and/or cell-cell interaction were involved in this process, we isolated NK cells and MSCs with 0.4μM transwell or incubated NK cells with fixed MSCs. MSCs were deprived completely of the capacity to suppress NKG2D expression when separated them with transwell, whereas fixed MSCs could not block this process. That was to say, membrane-bound factors rather than soluble ones play a critical role. As TGF-β is an important inhibitory cytokine produced by MSCs and it also is the critical regulatoryfactor for NKG2D expression, we next added anti-TGF-β1 mAb into the coculture system. Interestingly, neutralization of TGF-61 restored the level of NKG2D, indicating that TGF-61 was responsible for the reduction in NKG2D expression by MSCs. This postulation was further testified that MSCs could not down-regulate NKG2D expression of NK cells isolated from Smad3(ex8/ex8) mice which have impaired TGF-β1 signal result from mutant on gene encoding Smad. However, co-culturing NK cells in the presence of the recombinant TGF-β1 had no influence in NKG2D expression. Furthermore, anti-TGF-β1 mAb blocked the down-modulation of MSCs towards NK cells, and the function of NK cells purified from Smad3^ex8/ex8 mice was nonresponsible for incubation with MSCs. We next investigated whether MSCs expressed membrane-bound TGF-β1. FACS-Purified MSCs were treble-stained with anti-Gr-1-APC, anti-TGF-Bl- FITC and DAPI. Confocal microscopy analysis was performed with a Leica TCS SP2 confocal laser microscope. The results showed that membrane-bound TGF-β1 was expressed on the surface of MSCs. Taken together, our data suggested that MSCs inhibit NK cell cytotoxcity through membrane-bound TGF-61.The property of MSCs was as little as 2-5% in spleen and mesenteric lymph nodes (MLN) of normal mice that lived in pathogen-free condition. If the NK cell cytotoxicity was normal, about 30-60% of these cells are NKG2D positive. To identify whether MSCs down-regulated NK cell function in vivo, MSCs were purified by sorting (purity >97%) and were intraperitoneally (i.p.) injected to normal C57BL/6 mice (3×l0⁶ per mouse). After adoptive transfer of MSCs, splenocytes and MLN cellswere examined at 12h, 18h and 24h respectively. We found that in vivo injection of MSCs significantly reduced NK cell function and NKG2D expression within 24h. We next investigated whether depletion of MSCs in vivo could influence NK cell function. Considering that high frequency of MSCs and low NK cell function were observable in the tumor-bearing mice (21days), we depleted MSCs by injection of anti-Gr-1 mAb into those mice bearing tumors for 21 days and then examined NK cell cytotoxicity and NKG2D expression 24h later. Very similar data were obtained that depletion of MSCs in vivo restored NK cell cytotoxicity and NKG2D expression.It is well known that CD4⁺CD25⁺ regulatory T cells (Treg), a critical player in the maintenance of peripheral tolerance, can inhibit NK cell function in cancer patients and tumor-bearing mice. To further identify the significance of the inhibitory function of MSCs towards NK cells, we studied the differences of MSCs frequency, NK cell function and NKD2D expression in liver in both normal mice and tumor-bearing mice. As compared to that in normal mice, there were dramatically elevated MSCs (from 4-7% to 40-50%) and evidently decreased NKG2D expression (from 30-50% to 4-9%) in tumor-bearing mice. However, there were less CD4⁺Foxp3⁺CD25⁺ Treg being detected in liver than that in spleen, moreover, the frequency of hepatic Treg had little change being observed in neither normal mice nor tumor-bearing mice. Hence, we presumed that the increased MSCs may be one of main contributors to downregulate NKG2D expression. It confirmed our postulation that transfer MSCs to normal mice decreased NKG2D expression and depletion of MSCs in tumor-bearing mice restored NKG2D expression. Collectively, our data suggest that MSCs are the criticalnegative-regulator for NK cell cytotoxicity in tumor-bearing host. Our results outline new mechanistic explanation for tumor immune escape.