Immunomodulatory metabolites are a key feature of the tumor microenvironment (TME), but with a few exceptions, their identities remain largely unknown. Here, we analyzed tumors and T cells from the tumors and ascites of patients with high-grade serous carcinoma (HGSC) to reveal the metabolome of these different TME compartments. Ascites and tumor cells have extensive metabolite differences. Compared with ascites, the tumor-infiltrating T cells are significantly enriched in 1-methylnicotinamide (MNA). Although the level of MNA in T cells is elevated, the expression of nicotinamide N-methyltransferase (an enzyme that catalyzes the transfer of methyl groups from S-adenosylmethionine to nicotinamide) is limited to fibroblasts and tumor cells. Functionally, MNA induces T cells to secrete the tumor-promoting cytokine tumor necrosis factor alpha. Therefore, TME-derived MNA contributes to the immune regulation of T cells and represents a potential immunotherapy target for the treatment of human cancer.
Tumor-derived metabolites can have a profound inhibitory effect on anti-tumor immunity, and more and more evidence shows that they can also serve as a key driving force for disease progression (1). In addition to the Warburg effect, recent work has begun to characterize the metabolic state of tumor cells and its relationship with the immune state of the tumor microenvironment (TME). Studies on mouse models and human T cells have shown that glutamine metabolism (2), oxidative metabolism (3) and glucose metabolism (4) can act independently on various immune cell subgroups. Several metabolites in these pathways inhibit the anti-tumor function of T cells. It has been proven that the blockade of the coenzyme tetrahydrobiopterin (BH4) can damage the proliferation of T cells, and the increase of BH4 in the body can enhance the anti-tumor immune response mediated by CD4 and CD8. In addition, the immunosuppressive effect of kynurenine can be rescued by the administration of BH4 (5). In isocitrate dehydrogenase (IDH) mutant glioblastoma, the secretion of enantiometabolic (R)-2-hydroxyglutarate (R-2-HG) inhibits T cell activation, proliferation and cytolysis Activity (6). Recently, it has been shown that methylglyoxal, a by-product of glycolysis, is produced by suppressor cells of myeloid origin, and T cell transfer of methylglyoxal can inhibit effector T cell function. In treatment, the neutralization of methylglyoxal can overcome the activity of myeloid-derived suppressor cells (MDSC) and synergistically enhance checkpoint blockade therapy in mouse models (7). These studies collectively emphasize the key role of TME-derived metabolites in regulating T cell function and activity.
T cell dysfunction has been widely reported in ovarian cancer (8). This is partly due to the metabolic characteristics inherent in hypoxia and abnormal tumor vasculature (9), which results in the conversion of glucose and tryptophan into by-products such as lactic acid and kynurenine. Excessive extracellular lactate reduces the production of interferon-γ (IFN-γ) and drives the differentiation of myelosuppressive subgroups (10, 11). The consumption of tryptophan directly inhibits T cell proliferation and inhibits T cell receptor signaling (12-14). Despite these observations, a lot of work surrounding immune metabolism was carried out in in vitro T cell culture using optimized media, or limited to homologous mouse models in vivo, neither of which fully reflects the heterogeneity of human cancers and Physiological macro and micro environment.
A common feature of ovarian cancer is peritoneal spread and the appearance of ascites. The accumulation of cell fluid in ascites is associated with advanced disease and poor prognosis (15). According to reports, this unique compartment is hypoxic, has high levels of vascular endothelial growth factor (VEGF) and indoleamine 2,3-dioxygenase (IDO), and is infiltrated by T regulatory cells and myeloid inhibitory cells (15-18). The metabolic environment of ascites may be different from that of the tumor itself, so the reprogramming of T cells in the peritoneal space is unclear. In addition, the key differences and heterogeneity between ascites and metabolites present in the tumor environment may hinder the infiltration of immune cells and their function on tumors, and further research is needed.
In order to solve these problems, we designed a sensitive cell separation and liquid chromatography tandem mass spectrometry (LC-MS/MS) method to study different cell types (including CD4 + and CD8 + T cells) as well as within and between tumors Its metabolites span cells in the same ascites and tumor environment of the patient. We use this method in conjunction with high-dimensional flow cytometry and single-cell RNA sequencing (scRNA-seq) to provide a highly resolved portrait of the metabolic status of these key populations. This method revealed a significant increase in the level of 1-methylnicotinamide (MNA) in tumor T cells, and in vitro experiments showed that the immunomodulatory effect of MNA on T cell function was previously unknown. In general, this method reveals the mutual metabolic interactions between tumors and immune cells, and provides unique insights into immune regulation metabolites, which may be useful for the treatment of T cell-based ovarian cancer immunotherapy Treatment opportunities.
We used high-dimensional flow cytometry to simultaneously quantify glucose uptake [2-(N-(7-nitrophenyl-2-oxa-1,3-diaza-4-yl)amino)-2-deoxyglucose (2-NBDG) and mitochondrial activity [MitoTracker Deep Red (MT DR)] (7, 19, 20) are side by side typical markers that distinguish immune cells and tumor cell populations (Table S2 and Figure S1A). This analysis showed that compared with T cells, ascites and tumor cells have higher glucose uptake levels, but have smaller differences in mitochondrial activity. The average glucose uptake of tumor cells [CD45-EpCAM (EpCAM)+] is three to four times that of T cells, and the average glucose uptake of CD4 + T cells is 1.2 times that of CD8 + T cells, which indicates Tumor infiltrating lymphocytes (TIL) have different metabolic requirements even in the same TME (Figure 1A). In contrast, the mitochondrial activity in tumor cells is similar to that of CD4 + T cells, and the mitochondrial activity of both cell types is higher than that of CD8 + T cells (Figure 1B). In general, these results reveal the metabolic level. The metabolic activity of tumor cells is higher than that of CD4 + T cells, and the metabolic activity of CD4 + T cells is higher than that of CD8 + T cells. Despite these effects across cell types, there is no consistent difference in the metabolic status of CD4 + and CD8 + T cells or their relative proportions in ascites compared with tumors (Figure 1C). In contrast, in the CD45-cell fraction, the proportion of EpCAM+ cells in the tumor increased compared with ascites (Figure 1D). We also observed a clear metabolic difference between EpCAM+ and EpCAM- cell components. EpCAM+ (tumor) cells have higher glucose uptake and mitochondrial activity than EpCAM- cells, which is much higher than the metabolic activity of fibroblasts in tumor cells in TME (Figure 1, E and F).
(A and B) Median fluorescence intensity (MFI) of glucose uptake (2-NBDG) (A) and mitochondrial activity of CD4 + T cells (MitoTracker dark red) (B) Representative graphs (left) and tabulated data (Right), CD8 + T cells and EpCAM + CD45-tumor cells from ascites and tumor. (C) The ratio of CD4 + and CD8 + cells (of CD3 + T cells) in ascites and tumor. (D) Proportion of EpCAM + tumor cells in ascites and tumor (CD45−). (E and F) EpCAM + CD45-tumor and EpCAM-CD45-matrix glucose uptake (2-NBDG) (E) and mitochondrial activity (MitoTracker dark red) (F) representative graphs (left) and tabulated data ( Right) Ascites and tumor cells. (G) Representative graphs of CD25, CD137 and PD1 expression by flow cytometry. (H and I) CD25, CD137 and PD1 expression on CD4 + T cells (H) and CD8 + T cells (I). (J and K) Naive, central memory (Tcm), effector (Teff) and effector memory (Tem) phenotypes based on the expression of CCR7 and CD45RO. Representative images (left) and tabular data (right) of CD4 + T cells (J) and CD8 + T cells (K) in ascites and tumors. P values ​​determined by paired t-test (*P<0.05, **P<0.01 and ***P<0.001). The line represents matched patients (n = 6). FMO, fluorescence minus one; MFI, median fluorescence intensity.
Further analysis revealed other significant differences between the highly resolved T cell phenotypic status. The activated (Figure 1, G to I) and effector memory (Figure 1, J and K) in tumors are much more frequent than ascites (proportion of CD3 + T cells). Similarly, analyzing the phenotype by the expression of activation markers (CD25 and CD137) and depletion markers [programmed cell death protein 1 (PD1)] showed that although the metabolic characteristics of these populations are different (Figure S1, B to E) , But no significant metabolic differences were consistently observed between naive, effector or memory subsets (Figure S1, F to I). These results were confirmed by using machine learning methods to automatically assign cell phenotypes (21), which further revealed the presence of a large number of bone marrow cells (CD45 + / CD3- / CD4 + / CD45RO +) in the patient’s ascites (Figure S2A ). Among all the identified cell types, this myeloid cell population showed the highest glucose uptake and mitochondrial activity (Figure S2, B to G). These results highlight the strong metabolic differences between the multiple cell types found in ascites and tumors in HGSC patients.
The main challenge in understanding the metabonomic characteristics of TIL is the need to isolate T cell samples of sufficient purity, quality and quantity from tumors. Recent studies have shown that sorting and bead enrichment methods based on flow cytometry may lead to changes in cellular metabolite profiles (22-24). To overcome this problem, we optimized the bead enrichment method to isolate and isolate TIL from surgically resected human ovarian cancer prior to analysis by LC-MS/MS (see Materials and Methods; Figure 2A). In order to assess the overall impact of this protocol on metabolite changes, we compared the metabolite profiles of T cells activated by healthy donors after the above bead separation step with cells that were not bead separated but remained on ice. This quality control analysis found that there is a high correlation between these two conditions (r = 0.77), and the technical repeatability of the group of 86 metabolites has high repeatability (Figure 2B). Therefore, these methods can perform accurate metabolite analysis in cells undergoing cell type enrichment, thus providing the first high-resolution platform for identifying specific metabolites in HGSC, thereby enabling people to gain a deeper understanding of cell specificity Sexual metabolism program.
(A) Schematic diagram of magnetic bead enrichment. Before analysis by LC-MS/MS, the cells will undergo three consecutive rounds of magnetic bead enrichment or remain on ice. (B) The effect of enrichment type on the abundance of metabolites. The average of three measurements for each enrichment type ± SE. The gray line represents a 1:1 relationship. The intra-class correlation (ICC) of repeated measurements shown in the axis label. NAD, nicotinamide adenine dinucleotide. (C) Schematic diagram of the workflow of patient metabolite analysis. Ascites or tumors are collected from patients and cryopreserved. A small portion of each sample was analyzed by flow cytometry, while the remaining samples underwent three rounds of enrichment for CD4+, CD8+ and CD45- cells. These cell fractions were analyzed using LC-MS/MS. (D) Heat map of standardized metabolite abundance. The dendrogram represents Ward’s clustering of Euclidean distances between samples. (E) Principal component analysis (PCA) of sample metabolite map, showing three replicates of each sample, samples from the same patient are connected by a line. (F) The PCA of the metabolite profile of the sample conditioned on the patient (ie, using partial redundancy); the sample type is limited by the convex hull. PC1, main component 1; PC2, main component 2.
Next, we applied this enrichment method to analyze 99 metabolites in the CD4 +, CD8 + and CD45-cell fractions in the primary ascites and tumors of six HGSC patients (Figure 2C, Figure S3A and Table S3 and S4). The population of interest accounts for 2% to 70% of the original large sample of living cells, and the proportion of cells varies greatly between patients. After separating the beads, the enriched fraction of interest (CD4+, CD8+ or CD45-) accounts for more than 85% of all living cells in the sample on average. This enrichment method allows us to analyze cell populations from human tumor tissue metabolism, which is impossible to do from large samples. Using this protocol, we determined that l-kynurenine and adenosine, these two well-characterized immunosuppressive metabolites were elevated in tumor T cells or tumor cells (Figure S3, B and C). Therefore, these results demonstrate the fidelity and ability of our cell separation and mass spectrometry technology to find biologically important metabolites in patient tissues.
Our analysis also revealed a strong metabolic separation of cell types within and between patients (Figure 2D and Figure S4A). In particular, compared with other patients, patient 70 showed different metabolic characteristics (Figure 2E and Figure S4B), indicating that there may be substantial metabolic heterogeneity between patients. It is worth noting that compared with other patients (1.2 to 2 liters; Table S1), the total amount of ascites collected in patient 70 (80 ml) was smaller. The control of inter-patient heterogeneity during principal component analysis (for example, using partial redundancy analysis) shows consistent changes between cell types, and the cell types and/or microenvironment are clearly aggregated according to the metabolite profile (Figure 2F ). The analysis of single metabolites emphasized these effects and revealed significant differences between cell types and microenvironment. It is worth noting that the most extreme difference observed is MNA, which is usually enriched in CD45- cells and in CD4+ and CD8+ cells that infiltrate the tumor (Figure 3A). For CD4 + cells, this effect is most obvious, and the MNA in CD8 + cells also seems to be strongly affected by the environment. However, this is not important, because only three of the six patients can be evaluated for tumor CD8+ scores. In addition to MNA, in different types of cells in ascites and tumors, other metabolites that are poorly characterized in TIL are also differentially rich (Figures S3 and S4). Therefore, these data reveal a promising set of immunomodulatory metabolites for further research.
(A) Normalized content of MNA in CD4+, CD8+ and CD45- cells from ascites and tumor. The box plot shows the median (line), interquartile range (frame hinge) and data range, up to 1.5 times the interquartile range (frame whisker). As described in Patient Materials and Methods, use the patient’s limma value to determine the P value (*P<0.05 and **P<0.01). (B) Schematic diagram of MNA metabolism (60). Metabolites: S-adenosyl-1-methionine; SAH, S-adenosine-1-homocysteine; NA, nicotinamide; MNA, 1-methylnicotinamide; 2-PY, 1-methyl- 2-pyridone-5-carboxamide; 4-PY, 1-methyl-4-pyridone-5-carboxamide; NR, nicotinamide ribose; NMN, nicotinamide mononucleotide. Enzymes (green): NNMT, nicotinamide N-methyltransferase; SIRT, sirtuins; NAMPT, nicotinamide phosphoribosyl transferase; AOX1, aldehyde oxidase 1; NRK, nicotinamide riboside kinase; NMNAT, nicotinamide mono Nucleotide adenylate transferase; Pnp1, purine nucleoside phosphorylase. (C) t-SNE of scRNA-seq of ascites (grey) and tumor (red; n = 3 patients). (D) NNMT expression in different cell populations identified using scRNA-seq. (E) Expression of NNMT and AOX1 in SK-OV-3, human embryonic kidney (HEK) 293T, T cells and MNA-treated T cells. The folded expression is shown relative to SK-OV-3. The expression pattern with SEM is shown (n = 6 healthy donors). Ct values ​​greater than 35 are considered undetectable (UD). (F) Expression of SLC22A1 and SLC22A2 in SK-OV-3, HEK293T, T cells and T cells treated with 8mM MNA. The folded expression is shown relative to SK-OV-3. The expression pattern with SEM is shown (n = 6 healthy donors). Ct values ​​greater than 35 are considered undetectable (UD). (G) Cell MNA content in activated healthy donor T cells after 72 hours of incubation with MNA. The expression pattern with SEM is shown (n = 4 healthy donors).
MNA is produced by transferring the methyl group from S-adenosyl-1-methionine (SAM) to nicotinamide (NA) by nicotinamide N-methyltransferase (NNMT; Figure 3B). NNMT is overexpressed in a variety of human cancers and is associated with proliferation, invasion and metastasis (25-27). To better understand the source of MNA in T cells in TME, we used scRNA-seq to characterize the expression of NNMT across cell types in the ascites and tumors of three HGSC patients (Table S5). Analysis of approximately 6,500 cells showed that in ascites and tumor environments, NNMT expression was restricted to the presumed fibroblast and tumor cell populations (Figure 3, C and D). It is worth noting that there is no obvious NNMT expression in any population that expresses PTPRC (CD45 +) (Figure 3D and Figure S5A), which indicates that the MNA detected in the metabolite spectrum has been introduced into T cells. The expression of aldehyde oxidase 1 (AOX1) converts MNA into 1-methyl-2-pyridone-5-carboxamide (2-PYR) or 1-methyl-4-pyridone-5-carboxamide (4- PYR); Figure 3B) is also restricted to the population of fibroblasts expressing COL1A1 (Figure S5A), which together indicate that T cells lack the ability of conventional MNA metabolism. The expression pattern of these MNA-related genes was verified using a second independent cell data set from ascites from HGSC patients (Figure S5B; n = 6) (16). In addition, quantitative polymerase chain reaction (qPCR) analysis of healthy donor T cells treated with MNA showed that compared with control SK-OV-3 ovarian tumor cells, NNMT or AOX1 was almost not expressed (Figure 3E). These unexpected results indicate that MNA may be secreted from fibroblasts or tumors into adjacent T cells in TME.
Although candidates include the family of organic cation transporters 1 to 3 (OCT1, OCT2 and OCT3) encoded by the soluble carrier 22 (SLC22) family (SLC22A1, SLC22A2 and SLC22A3), the potential transporters of MNA are still undefined (28) . QPCR of mRNA from healthy donor T cells showed low expression levels of SLC22A1 but undetectable levels of SLC22A2, which confirmed that it had been previously reported in the literature (Figure 3F) (29). In contrast, the SK-OV-3 ovarian tumor cell line expressed high levels of both transporters (Figure 3F).
In order to test the possibility of T cells having the ability to absorb foreign MNA, healthy donor T cells were cultured for 72 hours in the presence of different concentrations of MNA. In the absence of exogenous MNA, the cellular content of MNA cannot be detected (Figure 3G). However, activated T cells treated with exogenous MNA showed a dose-dependent increase in MNA content in the cells, up to 6 mM MNA (Figure 3G). This result indicates that despite the low level of transporter expression and lack of the main enzyme responsible for intracellular MNA metabolism, TIL can still take up MNA.
The spectrum of metabolites in patients’ T cells and in vitro MNA absorption experiments increase the possibility that cancer-associated fibroblasts (CAF) secrete MNA and tumor cells may regulate the phenotype and function of TIL. To determine the effect of MNA on T cells, healthy donor T cells were activated in vitro in the presence or absence of MNA, and their proliferation and cytokine production were evaluated. After 7 days of adding MNA at the highest dose, the population doubling number was moderately reduced, while vigor was maintained at all doses (Figure 4A). In addition, treatment of exogenous MNA resulted in an increase in the proportion of CD4 + and CD8 + T cells expressing tumor necrosis factor-α (TNFα; Figure 4B). In contrast, the intracellular production of IFN-γ was significantly reduced in CD4 + T cells, but not in CD8 + T cells, and there was no significant change in interleukin 2 (IL-2; Figure 4, C and D). Therefore, enzyme-linked immunosorbent assay (ELISA) of supernatants from these MNA-treated T cell cultures showed a significant increase in TNFα, a decrease in IFN-γ, and no change in IL-2 (Figure 4, E to G). . The decrease of IFN-γ indicates that MNA may play a role in inhibiting the anti-tumor activity of T cells. In order to simulate the effect of MNA on T cell-mediated cytotoxicity, chimeric antigen receptor T (FRα-CAR-T) cells targeting folate receptor α and CAR-T (GFP) regulated by green fluorescent protein (GFP) -CAR-T) cells are produced by healthy donor peripheral blood mononuclear cells (PBMC). CAR-T cells were cultured for 24 hours in the presence of MNA, and then co-cultured with human SK-OV-3 ovarian tumor cells expressing folate receptor α at an effector to target ratio of 10:1. MNA treatment resulted in a significant decrease in the killing activity of FRα-CAR-T cells, which was similar to FRα-CAR-T cells treated with adenosine (Figure 4H).
(A) Total viable cell count and population doubling (PD) directly from culture on day 7. The bar graph represents the mean + SEM of six healthy donors. Represents data from at least n = 3 independent experiments. (B to D) CD3/CD28 and IL-2 were used to activate T cells at their respective MNA concentrations for 7 days. Before analysis, cells were stimulated with PMA/ionomycin with GolgiStop for 4 hours. TNFα (B) expression in T cells. Example image (left) and tabular data (right) of TNFα expression in living cells. IFN-γ (C) and IL-2 (D) expression in T cells. The expression of cytokines was measured by flow cytometry. The bar graph represents the mean (n = 6 healthy donors) + SEM. Use one-way analysis of variance and repeated measures (*P<0.05 and **P<0.01) to determine the P value. Represents data from at least n = 3 independent experiments. (E to G) CD3/CD28 and IL-2 were used to activate T cells at their respective MNA concentrations for 7 days. The medium was collected before and after 4 hours of PMA/ionomycin stimulation. The concentrations of TNFα (E), IFN-γ (F) and IL-2 (G) were measured by ELISA. The bar graph represents the mean (n = 5 healthy donors) + SEM. P value determined using one-way analysis of variance and repeated measurements (*P<0.05). The dotted line indicates the detection limit of the detection. (H) Cell lysis assay. FRα-CAR-T or GFP-CAR-T cells were adjusted with adenosine (250μM) or MNA (10 mM) for 24 hours, or left untreated (Ctrl). The percentage killing of SK-OV-3 cells was measured. P value determined by Welch t test (*P<0.5 and **P<0.01).
In order to gain a mechanistic understanding of MNA-dependent TNFα expression regulation, the changes in TNFα mRNA of MNA-treated T cells were evaluated (Figure 5A). Healthy donor T cells treated with MNA showed a two-fold increase in TNFα transcription levels, indicating that MNA is dependent on TNFα transcriptional regulation. To investigate this possible regulatory mechanism, two known transcription factors that regulate TNFα, namely activated T cell nuclear factor (NFAT) and specific protein 1 (Sp1), were evaluated in response to MNA binding to the proximal TNFα promoter ( 30). The TNFα promoter contains 6 identified NFAT binding sites and 2 Sp1 binding sites, overlapping at one site [-55 base pairs (bp) from the 5'cap] (30). Chromatin immunoprecipitation (ChIP) showed that when treated with MNA, the binding of Sp1 to the TNFα promoter increased three-fold. The incorporation of NFAT also increased and approached importance (Figure 5B). These data indicate that MNA regulates the expression of TNFα through Sp1 transcription, and to a lesser extent the expression of NFAT.
(A) Compared with T cells cultured without MNA, the fold change of TNFα expression in T cells treated with MNA. The expression pattern with SEM is shown (n = 5 healthy donors). Represents data from at least n = 3 independent experiments. (B) The TNFα promoter of T cells treated with or without 8 mM MNA after NFAT and Sp1 were combined with (Ctrl) and PMA/ionomycin stimulation for 4 hours. Immunoglobulin G (IgG) and H3 were used as negative and positive controls for immunoprecipitation, respectively. The quantification of ChIP showed that the binding of Sp1 and NFAT to the TNFα promoter in MNA-treated cells increased several times compared with the control. Represents data from at least n = 3 independent experiments. P value determined by multiple t-tests (*** P <0.01). (C) Compared to the ascites of HGSC, T cells (non-cytotoxic) showed increased expression of TNF in the tumor. The colors represent different patients. The displayed cells have been randomly sampled to 300 and jittered to limit overdrawing (** Padj = 0.0076). (D) Proposed model of MNA for ovarian cancer. MNA is produced in tumor cells and fibroblasts in TME and is taken up by T cells. MNA increases the binding of Sp1 to the TNFα promoter, leading to increased TNFα transcription and TNFα cytokine production. MNA also causes a decrease in IFN-γ. Inhibition of T cell function leads to reduced killing ability and accelerated tumor growth.
According to reports, TNFα has front and back-dependent anti-tumor and anti-tumor effects, but it has a well-known role in promoting the growth and metastasis of ovarian cancer (31-33). According to reports, the concentration of TNFα in ascites and tumor tissues in patients with ovarian cancer is higher than that in benign tissues (34-36). In terms of mechanism, TNFα can regulate the activation, function and proliferation of white blood cells, and change the phenotype of cancer cells (37, 38). Consistent with these findings, differential gene expression analysis showed that TNF was significantly up-regulated in T cells in tumor tissues compared with ascites (Figure 5C). The increase in TNF expression was only evident in T cell populations with a non-cytotoxic phenotype (Figure S5A). In summary, these data support the view that MNA has dual immunosuppressive and tumor promoting effects in HGSC.
Fluorescent labeling based on flow cytometry has become the main method for studying TIL metabolism. These studies have shown that compared with peripheral blood lymphocytes or T cells from secondary lymphoid organs, murine and human TIL have a higher tendency to uptake glucose (4, 39) and the gradual loss of mitochondrial function (19, 40). Although we have observed similar results in this study, the key development is to compare the metabolism of tumor cells and TIL from the same resected tumor tissue. Consistent with some of these previous reports, tumor (CD45-EpCAM +) cells from ascites and tumors have higher glucose uptake than CD8 + and CD4 + T cells, supporting that the high glucose uptake of tumor cells can be compared with T cells. The concept of T cell competition. TME. However, the mitochondrial activity of tumor cells is higher than that of CD8 + T cells, but the mitochondrial activity is similar to that of CD4 + T cells. These results reinforce the emerging theme that oxidative metabolism is important for tumor cells (41, 42). They also suggest that CD8 + T cells may be more susceptible to oxidative dysfunction than CD4 + T cells, or that CD4 + T cells may use carbon sources other than glucose to maintain mitochondrial activity (43, 44). It should be noted that we observed no difference in glucose uptake or mitochondrial activity between CD4 + T effectors, T effector memory and T central memory cells in ascites. Similarly, the differentiation state of CD8 + T cells in tumors has nothing to do with changes in glucose uptake, highlighting the significant difference between T cells cultured in vitro and human TIL in vivo (22). These observations were also confirmed by the use of unbiased automatic cell population allocation, which further revealed that CD45 + / CD3- / CD4 + / CD45RO + cells with higher glucose uptake and mitochondrial activity than tumor cells are prevalent but have Metabolic active cell population. This population may represent the putative subpopulation of myeloid suppressor cells or plasmacytoid dendritic cells identified in the scRNA-seq analysis. Although both of these have been reported in human ovarian tumors [45], they still need Further work is to describe this myeloid subpopulation.
Although flow cytometry-based methods can clarify the general differences in glucose and oxidative metabolism between cell types, the precise metabolites produced by glucose or other carbon sources for mitochondrial metabolism in TME have not yet been determined. Assigning the presence or absence of metabolites to a given TIL subset requires purification of the cell population from the excised tissue. Therefore, our cell enrichment method combined with mass spectrometry can provide insights into the metabolites that are differentially enriched in T cells and tumor cell populations in matching patient samples. Although this method has advantages over fluorescence-activated cell sorting, certain metabolite libraries may be affected due to inherent stability and/or rapid turnover rate (22). Nevertheless, our method was able to identify two recognized immunosuppressive metabolites, adenosine and kynurenine, because they vary greatly between sample types.
Our metabonomic analysis of tumors and TIL subtypes provides more insights into the role of metabolites in ovarian TME. First, using flow cytometry, we determined that there was no difference in mitochondrial activity between tumors and CD4 + T cells. However, LC-MS/MS analysis revealed significant changes in the abundance of metabolites among these populations, indicating that conclusions about TIL metabolism and its overall metabolic activity require careful interpretation. Secondly, MNA is the metabolite with the greatest difference between CD45-cells and T cells in ascites, not tumors. Therefore, compartmentalization and tumor location may have different effects on TIL metabolism, which highlights the possible heterogeneity in a given microenvironment. Third, the expression of the MNA-producing enzyme NNMT is mainly limited to CAF, which is tumor cells to a lesser extent, but detectable MNA levels are observed in tumor-derived T cells. The overexpression of NNMT in ovarian CAF has a known cancer-promoting effect, partly due to the promotion of CAF metabolism, tumor invasion and metastasis (27). Although the overall level of TIL is moderate, the expression of NNMT in CAF is closely related to the Cancer Genome Atlas (TCGA) mesenchymal subtype, which is associated with poor prognosis (27, 46, 47). Finally, the expression of the enzyme AOX1 responsible for MNA degradation is also limited to the CAF population, which indicates that T cells lack the ability to metabolize MNA. These results support the idea that although further work is needed to verify this finding, high levels of MNA in T cells may indicate the presence of an immunosuppressive CAF microenvironment.
Given the low expression level of MNA transporters and the undetectable levels of key proteins involved in MNA metabolism, the presence of MNA in T cells is unexpected. Neither NNMT nor AOX1 could be detected by scRNA-seq analysis and targeted qPCR of two independent cohorts. These results indicate that MNA is not synthesized by T cells, but absorbed from surrounding TME. In vitro experiments show that T cells tend to accumulate exogenous MNA.
Our in vitro studies have shown that exogenous MNA induces the expression of TNFα in T cells and enhances the binding of Sp1 to the TNFα promoter. Although TNFα has both anti-tumor and anti-tumor functions, in ovarian cancer, TNFα can promote the growth of ovarian cancer (31-33). The neutralization of TNFα in ovarian tumor cell culture or the elimination of TNFα signal in mouse models can improve TNFα-mediated inflammatory cytokine production and inhibit tumor growth (32, 35). Therefore, in this case, TME-derived MNA can act as a pro-inflammatory metabolite through a TNFα-dependent mechanism through the autocrine loop, thereby promoting the occurrence and spread of ovarian cancer (31). Based on this possibility, TNFα blockade is being studied as a potential therapeutic agent for ovarian cancer (37, 48, 49). In addition, MNA impairs the cytotoxicity of CAR-T cells to ovarian tumor cells, providing further evidence for MNA-mediated immune suppression. Collectively, these results suggest a model in which tumors and CAF cells secrete MNA into extracellular TME. Through (i) TNF-induced ovarian cancer growth stimulation and (ii) MNA-induced T cell cytotoxic activity inhibition, this may have a dual tumor effect (Figure 5D).
In conclusion, by applying a combination of rapid cell enrichment, single-cell sequencing and metabolic profiling, this study revealed the huge immunometabolomic differences between tumors and ascites cells in HGSC patients. This comprehensive analysis showed that there are differences in glucose uptake and mitochondrial activity between T cells, and identified MNA as a non-cell autonomous immune regulatory metabolite. These data have an impact on how TME affects T cell metabolism in human cancers. Although the direct competition for nutrients between T cells and cancer cells has been reported, metabolites can also act as indirect regulators to promote tumor progression and possibly suppress endogenous immune responses. The further description of the functional role of these regulatory metabolites may open up alternative strategies for enhancing the anti-tumor immune response.
Patient specimens and clinical data were obtained through the BC cancer tumor tissue repository certified by the Canadian Tissue Repository Network. According to the protocol approved by the BC Cancer Research Ethics Committee and the University of British Columbia (H07-00463), all patients’ specimens and clinical data obtained informed written consent or formally waived their consent. The samples are stored in the certified BioBank (BRC-00290). Detailed patient characteristics are shown in Tables S1 and S5. For cryopreservation, a scalpel is used to mechanically decompose the patient’s tumor sample and then push it through a 100-micron filter to obtain a single cell suspension. The patient’s ascites was centrifuged at 1500 rpm for 10 minutes at 4°C to pellet the cells and remove the supernatant. Cells obtained from tumor and ascites were cryopreserved in 50% heat-inactivated human AB serum (Sigma-Aldrich), 40% RPMI-1640 (Thermo Fisher Scientific) and 10% dimethyl sulfoxide. These preserved single cell suspensions were thawed and used for metabolomics and metabolite determination described below.
The complete medium consists of 0.22 μm filtered 50:50 supplemented RPMI 1640: AimV. RPMI 1640 + 2.05 mM l-glutamine (Thermo Fisher Scientific) supplemented with 10% heat-inactivated human AB serum (Sigma-Aldrich), 12.5 mM Hepes (Thermo Fisher Scientific), 2 mM l-glutamine (Thermo Fisher Scientific) Fisher Scientific), 1 x Penicillin Streptomycin (PenStrep) solution (Thermo Fisher Scientific) and 50 μMB-mercaptoethanol. AimV (Invitrogen) is supplemented with 20 mM Hepes (Thermo Fisher Scientific) and 2 mM l-glutamine (Thermo Fisher Scientific). The flow cytometer staining buffer consisted of 0.22μm filtered phosphate buffered saline (PBS; Invitrogen) supplemented with 3% heat-inactivated AB human serum (Sigma). The cell enrichment buffer is composed of 0.22μm filtered PBS and supplemented with 0.5% heat-inactivated human AB serum (Sigma-Aldrich).
In 37°C complete medium, cells were stained with 10 nM MT DR and 100 μM 2-NBDG for 30 minutes. Next, the cells were stained with viability dye eF506 at 4°C for 15 minutes. Resuspend the cells in FC Block (eBioscience) and Brilliant Stain Buffer (BD Biosciences), dilute in flow cytometer staining buffer (according to the manufacturer’s instructions), and incubate for 10 minutes at room temperature. Stain the cells with a set of antibodies (Table S2) in flow cytometry staining buffer at 4°C for 20 minutes. Resuspend the cells in flow cytometry staining buffer (Cytek Aurora; 3L-16V-14B-8R configuration) before analysis. Use SpectroFlo and FlowJo V10 to analyze cell count data, and use GraphPad Prism 8 to create the data. The median fluorescence intensity (MFI) of 2-NBDG and MT DR was log-normalized, and then a paired t test was used for statistical analysis to account for matched patients. Remove all populations with fewer than 40 events from the analysis; enter an MFI value of 1 for any negative values ​​before performing statistical analysis and data visualization.
In order to supplement the manual gating strategy of the above process panel, we used the full annotation by the shape restriction tree (FAUST) (21) to automatically assign cells to the population after eliminating dead cells in FlowJo. We manually manage the output to merge populations that seem to be misallocated (combining PD1+ with PD1-tumor cells) and retained populations. Each sample contains an average of more than 2% cells, for a total of 11 populations.
Ficoll gradient density centrifugation was used to separate PBMC from leukocyte separation products (STEMCELL Technologies). CD8 + T cells were isolated from PBMC using CD8 MicroBeads (Miltenyi) and expanded in complete medium using TransAct (Miltenyi) for 2 weeks according to the manufacturer’s instructions. The cells were allowed to stand for 5 days in complete medium containing IL-7 (10 ng/ml; PeproTech), and then re-stimulated with TransAct. On day 7, according to the manufacturer’s instructions, human CD45 MicroBeads (Miltenyi) were used to enrich cells in three consecutive rounds. The cells were aliquoted for flow cytometry analysis (as described above), and one million cells were aliquoted three times for LC-MS/MS analysis. The samples were processed by LC-MS/MS as described below. We estimated the missing metabolite value with an ion number of 1,000. Each sample is normalized by the total ion number (TIC), logarithmically converted and automatically normalized in MetaboAnalystR before analysis.
The single cell suspension of each patient was thawed and filtered through a 40 μm filter into complete medium (as described above). According to the manufacturer’s protocol, three consecutive rounds of positive selection by magnetic bead separation using MicroBeads (Miltenyi) were used to enrich the samples for CD8+, CD4+ and CD45- cells (on ice). In short, the cells are resuspended in cell enrichment buffer (as described above) and counted. The cells were incubated with human CD8 beads, human CD4 beads or human CD45 beads (Miltenyi) at 4°C for 15 minutes, and then washed with cell enrichment buffer. The sample is passed through the LS column (Miltenyi), and the positive and negative fractions are collected. In order to reduce the duration and maximize the cell recovery step, the CD8-fraction is then used for the second round of CD4+ enrichment, and the CD4-fraction is used for the subsequent CD45-enrichment. Keep the solution on ice throughout the separation process.
To prepare samples for metabolite analysis, the cells were washed once with ice-cold salt solution, and 1 ml of 80% methanol was added to each sample, then vortexed and snap frozen in liquid nitrogen. The samples were subjected to three freeze-thaw cycles and centrifuged at 14,000 rpm for 15 minutes at 4°C. The supernatant containing the metabolites is evaporated until dry. The metabolites were re-dissolved in 50 μl of 0.03% formic acid, vortexed to mix, and then centrifuged to remove debris.
Extract metabolites as described above. Transfer the supernatant to a high performance liquid chromatography bottle for metabolomics research. Use a random treatment protocol to treat each sample with a similar number of cells to prevent batch effects. We performed a qualitative assessment of global metabolites previously published on the AB SCIEX QTRAP 5500 Triple Quadrupole Mass Spectrometer (50). Chromatographic analysis and peak area integration were performed using MultiQuant version 2.1 software (Applied Biosystems SCIEX).
An ion count of 1000 was used to estimate the missing metabolite value, and the TIC of each sample was used to calculate the normalized peak area of ​​each detected metabolite to correct for changes introduced by the instrumental analysis from sample processing. After TIC is normalized, MetaboAnalystR(51) (default parameter) is used for logarithmic conversion and automatic norm line scaling. We used PCA with vegan R package to perform exploratory analysis of metabolome differences between sample types, and used partial redundancy analysis to analyze patients. Use Ward method to construct a heat map dendrogram to cluster the Euclidean distance between samples. We used limma (52) on standardized metabolite abundance to identify differentially abundant metabolites across the entire cell type and microenvironment. In order to simplify the explanation, we use the group mean parameter to specify the model, and consider the cell types in the microenvironment as each group (n = 6 groups); for the significance test, we performed three repeated measurements for each metabolite In order to avoid false replication, the patient was included as an obstacle in the limma design. In order to check the differences in metabolites between different patients, we adjusted the limma model including patients in a fixed way. We report the significance of the pre-specified contrast between the cell type and the microenvironment of Padj <0.05 (Benjamini-Hochberg correction).
After vigor enrichment using Miltenyi Dead Cell Removal Kit (>80% viability), single-cell transcriptome sequencing was performed on the total live frozen ascites and tumor samples using a 10x 5′gene expression protocol. Five cases with matching tumors and ascites were analyzed, although the low viability from one tumor sample prevented its inclusion. In order to achieve multiple selections of patients, we combined the samples of each patient in the lanes of the 10x chromium controller, and analyzed the ascites and tumor sites separately. After sequencing [Illumina HiSeq 4000 28×98 bp paired end (PE), Quebec genome; an average of 73,488 and 41,378 reads per cell for tumor and ascites respectively]], we used CellSNP and Vireo (53) (based on CellSNP as The common human SNP (VCF) provided by GRCh38 is assigned a donor identity. We use SNPRelate to infer the closest identity (IBS) of the patient’s genotype status (IBS), excluding unassigned cells and cells identified as duplexes and matching donors between ascites and tumor samples (54). On the basis of this task, we retained three cases with abundant cell representation in the tumor and ascites for downstream analysis. After performing a mass filtration step in the scater (55) and scran (56) BioConductor packaging, this yielded 6975 cells (2792 and 4183 cells from tumor and ascites, respectively) for analysis. We use igraph’s (57) Louvain clustering of shared nearest neighbor network (SNN) based on Jaccard distance to cluster cells by expression. The clusters were manually annotated to putative cell types based on marker gene expression and visualized with t-SNE. Cytotoxic T cells are defined by the expression of CD8A and GZMA, excluding subclusters with low ribosomal protein expression. We accessed the published data of Izar et al. (16), including their t-SNE embedding, can control the expression overlap between immune cell markers and NNMT expression.
PBMC were separated from leukocyte separation products (STEMCELL Technologies) by Ficoll gradient density centrifugation. CD3 + cells were isolated from PBMC using CD3 beads (Miltenyi). In the presence or absence of MNA, CD3+ cells were activated with plate-bound CD3 (5μg/ml), soluble CD28 (3μg/ml) and IL-2 (300 U/ml; Proleukin). On the last day of expansion, the viability (Fixable Viability Dye eFluor450, eBioscience) and proliferation (123count eBeads, Thermo Fisher Scientific) were evaluated by flow cytometry. Evaluate effector function by stimulating cells with PMA (20 ng/ml) and ionomycin (1μg/ml) with GolgiStop for 4 hours, and monitor CD8-PerCP (RPA-T8, BioLegend), CD4-AF700 (RPA-T4) , BioLegend) and TNFα-fluorescein isothiocyanate (FITC) (MAb11, BD). Stimulate qPCR and ChIP cells with PMA (20 ng/ml) and ionomycin (1μg/ml) for 4 hours. The ELISA supernatant was collected before and after stimulation with PMA (20 ng/ml) and ionomycin (1 μg/ml) for 4 hours.
Follow the manufacturer’s protocol to isolate RNA using RNeasy Plus Mini Kit (QIAGEN). Use QIAshredder (QIAGEN) to homogenize the sample. Use high-capacity RNA to cDNA kit (Thermo Fisher Scientific) to synthesize complementary DNA (cDNA). Use TaqMan Rapid Advanced Master Mix (Thermo Fisher Scientific) to quantify gene expression (according to the manufacturer’s protocol) with the following probes: Hs00196287_m1 (NNMT), Hs00154079_m1 (AOX1), Hs00427552_m1 (SLC22A1), Hs02786624_g1 [glyceraldehyde-3-phosphate off Hydrogen (GAPDH)] and Hs01010726_m1 (SLC22A2). The samples were run on the StepOnePlus real-time PCR system (Applied Biosystems) (Applied Biosystems) in the MicroAmp fast optical 96-well reaction plate (Applied Biosystems) with MicroAmp optical film. Any Ct value that exceeds 35 is considered to be above the detection threshold and is marked as undetectable.
Perform ChIP as previously described (58). In short, the cells were treated with formaldehyde (final concentration 1.42%) and incubated at room temperature for 10 minutes. Use supplemented swelling buffer (25 mM Hepes, 1.5 mM MgCl2, 10 mM KCl and 0.1% NP-40) on ice for 10 minutes, then resuspend in immunoprecipitation buffer as described (58 ). The sample was then sonicated with the following cycles: 10 cycles (20 1-second pulses) and a static time of 40 seconds. Incubate ChIP-grade immunoglobulin G (Cell Signaling Technology; 1μl), histone H3 (Cell Signaling Technology; 3μl), NFAT (Invitrogen; 3μl) and SP1 (Cell Signaling Technology; 3μl) antibodies with the sample at 4°C C shake overnight. Incubate protein A beads (Thermo Fisher Scientific) with the sample at 4°C with gentle shaking for 1 hour, then use chelex beads (Bio-Rad) to enrich the DNA, and use proteinase K (Thermo Fisher) for protein digestion. TNFα promoter was detected by PCR: forward, GGG TAT CCT TGA TGC TTG TGT; on the contrary, GTG CCA ACA ACT GCC TTT ATA TG (207-bp product). The images were produced by Image Lab (Bio-Rad) and quantified using ImageJ software.
The cell culture supernatant was collected as described above. The determination was carried out according to the manufacturer’s procedures of human TNFα ELISA kit (Invitrogen), human IL-2 ELISA kit (Invitrogen) and human IFN-γ ELISA kit (Abcam). According to the manufacturer’s protocol, the supernatant was diluted 1:100 to detect TNFα and IL-2, and 1:3 to detect IFN-γ. Use EnVision 2104 Multilabel Reader (PerkinElmer) to measure absorbance at 450 nm.
PBMC were separated from leukocyte separation products (STEMCELL Technologies) by Ficoll gradient density centrifugation. CD3 + cells were isolated from PBMC using CD3 beads (Miltenyi). In the presence or absence of MNA, CD3+ cells were activated with plate-bound CD3 (5μg/ml), soluble CD28 (3μg/ml) and IL-2 (300 U/ml; Proleukin) for 3 days. After 3 days, the cells were collected and washed with 0.9% saline, and the pellet was snap frozen. Cell count was performed by flow cytometry (Cytek Aurora; 3L-16V-14B-8R configuration) using 123count eBeads.
Extract metabolites as described above. The dried extract was reconstituted at a concentration of 4000 cell equivalents/μl. Analyze the sample by reversed-phase chromatography (1290 Infinity II, Agilent Technologies, Santa Clara, CA) and CORTECS T3 column (2.1×150 mm, particle size 1.6-μm, pore size 120-Å; #186008500, Waters). Polar mass spectrometer (6470, Agilent), in which electrospray ionization operates in positive mode. Mobile phase A is 0.1% formic acid (in H2O), mobile phase B is 90% acetonitrile, 0.1% formic acid. The LC gradient is 0 to 2 minutes for 100% A, 2 to 7.1 minutes for 99% B, and 7.1 to 8 minutes for 99% B. Then re-equilibrate the column with mobile phase A at a flow rate of 0.6 ml/min for 3 minutes. . The flow rate is 0.4ml/min, and the column chamber is heated to 50°C. Use MNA’s pure chemical standard (M320995, Toronto Research Chemical Company, North York, Ontario, Canada) to establish retention time (RT) and transformation (RT = 0.882 minutes, transformation 1 = 137→94.1, transformation 2 = 137→92 , Conversion 3 = 137→78). When all three transitions occur at the correct retention time, transition 1 is used for quantification to ensure specificity. The standard curve of MNA (Toronto Research Chemical Company) was generated by six serial dilutions of the stock solution (1 mg/ml) to obtain standards of 0.1, 1.0, 10 and 100 ng/ml and 1.0 and 10μg/ml respectively liquid. The detection limit is 1 ng/ml, and the linear response is between 10 ng/ml and 10μg/ml. Each injection of two microliters of sample and standard is used for LC/MS analysis, and a mixed quality control sample is run every eight injections to ensure the stability of the analysis platform. The MNA responses of all MNA-treated cell samples were within the linear range of the assay. Data analysis was done using MassHunter quantitative analysis software (v9.0, Agilent).
The second generation αFR-CAR construct was taken from Song et al. (59). In short, the construct contains the following contents: CD8a leader sequence, human αFR-specific single-chain variable fragment, CD8a hinge and transmembrane region, CD27 intracellular domain and CD3z intracellular domain. The complete CAR sequence was synthesized by GenScript, and then cloned into the second-generation lentiviral expression vector upstream of the GFP expression cassette used to evaluate the transduction efficiency.
Lentivirus is produced by transfection of HEK293T cells [American Type Culture Collection (ATCC); grown in Dulbecco's modified Eagle medium containing 10% fetal bovine serum (FBS) and 1% PenStrep, and used CAR-GFP vector and The packaging plasmids (psPAX2 and pMD2.G, Addgene) use lipofection amine (Sigma-Aldrich). The virus-containing supernatant was collected 48 and 72 hours after transfection, filtered, and concentrated by ultracentrifugation. Store the concentrated viral supernatant at -80°C until transduction.
PBMC are separated from healthy donor leukocyte separation products (STEMCELL Technologies) by Ficoll gradient density centrifugation. Use positive selection CD8 microbeads (Miltenyi) to isolate CD8+ cells from PBMC. Stimulate T cells with TransAct (Miltenyi) and in TexMACS medium [Miltenyi; supplemented with 3% heat-inactivated human serum, 1% PenStrep and IL-2 (300 U/ml)]. Twenty-four hours after stimulation, T cells were transduced with lentivirus (10 μl concentrated virus supernatant per 106 cells). 1 to 3 days after transduction on Cytek Aurora (on FSC (Forward Scatter)/SSC (Side Scatter), Singlet, GFP+), evaluate the GFP expression of the cells to demonstrate a transduction efficiency of at least 30% .
CAR-T cells were cultured for 24 hours in Immunocult (STEMCELL Technologies; supplemented with 1% PenStrep) under the following conditions: untreated, treated with 250 μM adenosine or 10 mM MNA. After pretreatment, CAR-T cells were washed with PBS and combined with 20,000 SK-OV-3 cells [ATCC; in McCoy 5A medium (Sigma-Aldrich) supplemented with 10% FBS and 1% PenStrep at 10: The effector to target ratio of 1 was amplified in triplicate in supplemented Immunocult medium. SK-OV-3 cells and SK-OV-3 cells lysed with digitalis saponin (0.5mg/ml; Sigma-Aldrich) were used as negative and positive controls, respectively. After 24 hours of co-cultivation, the supernatant was collected and the lactate dehydrogenase (LDH) was measured according to the manufacturer’s instructions (LDH Glo Cytotoxicity Assay Kit, Promega). The LDH supernatant was diluted 1:50 in LDH buffer. The percentage of killing was measured using the following formula: percentage of killing = percentage of correction / maximum killing rate x 100%, where percentage of correction = co-culture-T cells only, and maximum killing rate = positive control-negative control.
As described in the text or materials and methods, use GraphPad Prism 8, Microsoft Excel or R v3.6.0 for statistical analysis. If multiple samples are collected from the same patient (such as ascites and tumor), we use a paired t test or include the patient as a random effect in a linear or generalized model as appropriate. For metabolomics analysis, the importance test is performed in triplicate.
For supplementary materials for this article, please see http://advances.sciencemag.org/cgi/content/full/7/4/eabe1174/DC1
This is an open access article distributed under the terms of the Creative Commons Attribution-Non-Commercial License, which allows the use, distribution and reproduction in any medium, as long as the final use is not for commercial gain and the premise is that the original work is correct. Reference.
Note: We only ask you to provide your email address so that the person you recommend to the page knows that you want them to see the email and that it is not spam. We will not capture any email addresses.
This question is used to test whether you are a visitor and prevent automatic spam submission.
Marisa K. Kilgour (Marisa K. Kilgour), Sarah MacPherson (Sarah MacPherson), Lauren G. Zacharias (Lauren G. Zacharias), Abigail Eli Aris G. Watson (H. Watson), John Stagg (John Stagg), Brad H. Nelson (Brad H. Nelson), Ralph J. De Bellardini (Ralph J. DeBerardinis), Russell G. Jones (Russell G. Jones), Phineas T. Hamilton (Phineas T.
MNA contributes to the immune suppression of T cells and represents a potential immunotherapy target for the treatment of human cancer.
Marisa K. Kilgour (Marisa K. Kilgour), Sarah MacPherson (Sarah MacPherson), Lauren G. Zacharias (Lauren G. Zacharias), Abigail Eli Aris G. Watson (H. Watson), John Stagg (John Stagg), Brad H. Nelson (Brad H. Nelson), Ralph J. De Bellardini (Ralph J. DeBerardinis), Russell G. Jones (Russell G. Jones), Phineas T. Hamilton (Phineas T.
MNA contributes to the immune suppression of T cells and represents a potential immunotherapy target for the treatment of human cancer.
©2021 American Association for the Advancement of Science. all rights reserved. AAAS is a partner of HINARI, AGORA, OARE, CHORUS, CLOCKSS, CrossRef and COUNTER. ScienceAdvances ISSN 2375-2548.