PD-1 signaling negatively regulates the common cytokine receptor γ chain via MARCH5-mediated ubiquitination and degradation to suppress anti-tumor immunity

Combination therapy with PD-1 blockade and IL-2 substantially improves anti-tumor efficacy comparing to monotherapy. The underlying mechanisms responsible for the synergistic effects of the combination therapy remain enigmatic. Here we show that PD-1 ligation results in BATF-dependent transcriptional induction of the membrane-associated E3 ubiquitin ligase MARCH5, which mediates K27-linked polyubiquitination and lysosomal degradation of the common cytokine receptor γ chain (γc). PD-1 ligation also activates SHP2, which dephosphorylates γcY357, leading to impairment of γc family cytokine-triggered signaling. Conversely, PD-1 blockade restores γc level and activity, thereby sensitizing CD8+ T cells to IL-2. We also identified Pitavastatin Calcium as an inhibitor of MARCH5, which combined with PD-1 blockade and IL-2 significantly improves the efficacy of anti-tumor immunotherapy in mice. Our findings uncover the mechanisms by which PD-1 signaling antagonizes γc family cytokine-triggered immune activation and demonstrate that the underlying mechanisms can be exploited for increased efficacy of combination immunotherapy of cancer.


INTRODUCTION
The advances of cancer immunotherapy have revolutionized the treatment of cancer.[3][4][5][6][7][8][9][10] At the forefront of clinical cancer immunotherapy is immune checkpoint blockade (ICB).Immune checkpoint molecules such as PD-1 of the coinhibitory signaling pathways are evolved to control the magnitude and duration of T cell responses to limit tissue damage and maintain self-tolerance under physiological and pathological conditions. 11However, tumor cells can hijack these inhibitory pathways to escape host immune surveillance such as by overexpressing the PD-1 ligand PD-L1. 11,12This provides the rationale for clinical application of immune checkpoint inhibitors in cancer immunotherapy.Clinically, antibodies blocking PD-1/PD-L1 axis reinvigorate the exhausted T cells in tumor microenvironment (TME) and show remarkable objective response and durable remission with acceptable toxicity profile in large numbers of tumors such as lymphoma and melanoma. 13,14However, not all patients respond to PD-1 monotherapy and there is considerable interest in developing combination therapy to improve the overall response rate and trigger more complete and durable response in patients with cancer.
6][17] γ c is constitutively expressed on various populations of immune cells; mutation of the gene encoding γ c (IL2RG) results in X-linked severe combined immunodeficiency. 16,18,19Cytokines of the γ c family exhibit pleiotropic functions in both innate and adaptive immune responses, contribute to development of T, B, NK and innate lymphoid cells (ILCs), promote either survival or death of immune cells depending on the context, and modulate differentiation of precursor immune cells into more terminally differentiated cells. 15,20,21Because of their important roles in regulating activity of T, NK and other immune cells, some of the γ c family cytokines, such as IL-2, IL-9, IL-15 and IL-21, have shown strong anti-tumor effects. 15,20,21Studies of the γ c family cytokines have allowed remarkable translational advances for autoimmune diseases as well as cancer. 224][25] The immune checkpoint PD-1 blockade plus the γ c family cytokine IL-2 is a promising combination of cancer immunotherapy with several clinical trials ongoing. 15,21,26L-2 had been regarded as a therapeutic agent for cancer due to its powerful ability to stimulate proliferation of cytotoxic T lymphocytes and NK cells. 21,27However, only high-dose of IL-2 has shown therapeutic effects in certain cancer patients, and its widespread utilization is also limited by systemic toxicity, [27][28][29] whereas combination therapy with PD-1 blockade and IL-2 is highly effective in cancer patients. 21,26Understanding the underlying mechanisms responsible for the synergistic effects of this combination is important to design better strategies for cancer immunotherapy.
In this study, we found that PD-1 signaling transcriptionally induced the E3 ubiquitin ligase MARCH5, which targeted γ c for K27-linked polyubiquitination and lysosomal degradation.Additionally, PD-1 signaling also activated SHP2, which mediated dephosphorylation of γ c and inhibition of signaling triggered by the γ c family cytokines.These results suggest that PD-1 signaling antagonizes immune activation triggered by the γ c family of cytokines through two distinct mechanisms.Consistently, PD-1 blockade restored the response of tumor-infiltrating CD8 + T cells to γ c family cytokines.In addition, a MARCH5 inhibitor potently improved the efficacy of immunotherapy triggered by PD-1 blockade and IL-2 in mouse tumor models.Our results reveal the mechanisms by which PD-1 signaling inhibits γ c family cytokinetriggered activation of CD8 + T cells and provide potential targets for increased efficacy of combination immunotherapy of cancer.

PD-1 signaling down-regulates γ c level
To explore whether there is a regulatory relationship between PD-1 signaling and the γ c family cytokines, we firstly analyzed the correlation between PD-L1 and γ c level in TME.Immunohistochemistry (IHC) staining showed that γ c level was negatively correlated with PD-L1 level in human non-small cell lung cancer (NSCLC) biopsies (Fig. 1a).Next, we used a PD-1 antibody to disrupt the PD-1/PD-L1 signaling axis in mouse B16F10 and CT26 tumor models and then analyzed γ c level in the tumor tissues.The results showed that PD-1 blockade up-regulated γ c level in tumor tissues (Fig. 1b).Tumor-infiltrating CD8 + T cells is the main killer cells and rely on the γ c family cytokines for proliferation and activation.Therefore, we analyzed γ c level in tumor-infiltrating CD8 + T cells.The results indicated that the level of γ c was increased in these cells after PD-1 blockade treatment (Fig. 1c), suggesting that PD-1 signaling negatively regulates γ c level in tumor-infiltrating CD8 + T cells.
We next investigated whether PD-1 signaling regulates γ c level in human T cells in vitro.1][32][33] Flow cytometry analysis indicated that γ c level was not markedly changed after TCR activation by anti-CD3 and anti-CD28 antibodies, but was significantly down-regulated after PD-1 ligation by PD-L1 in both primary human CD8 + T cells and Jurkat T cells ectopically expressing PD-1 (Jurkat-PD-1) (Fig. 1d).Immunoblotting analysis further confirmed that PD-L1 treatment caused down-regulation of γ c level in primary human CD8 + T cells (Fig. 1e).Jurkat cells do not express PD-1 and PD-L1 stimulation of Jurkat cells activated by anti-CD3 and anti-CD28 failed to induce γ c down-regulation (Fig. 1f).However, PD-L1 stimulation caused down-regulation of γ c in Jurkat-PD-1 cells (in which PD-1 is ectopically expressed) activated by anti-CD3 and anti-CD28.In these experiments, PD-L1 did not down-regulate γ c in Jurkat-PD-1 cells not pretreated with anti-CD3 and anti-CD28, suggesting that TCR activation is important for PD-L1-triggered down-regulation of γ c (Fig. 1f).Consistently, PD-L1 also triggered down-regulation of γ c in Jurkat-PD-1 cells pretreated with phytohaemagglutinin (PHA) (Fig. 1f).These results further support that PD-1 ligation by PD-L1 causes down-regulation of γ c in activated T cells.Quantitative real-time PCR (qPCR) experiments indicated that PD-L1 co-stimulation did not affect the mRNA level of γ c in human primary CD8 + T cells or Jurkat-PD1 cells (Fig. 1g), suggesting that PD-1 signaling regulates γ c at the protein but not mRNA level.Together, these results suggest that PD-1 signaling down-regulates γ c level in human T cells.

MARCH5 mediates K27-linked polyubiquitination and degradation of γ c
We next investigated the mechanisms responsible for PD-1 signaling-triggered down-regulation of γ c .The protein synthesis inhibitor cycloheximide (CHX) treatment showed that γ c had a shorter half-life after PD-L1 stimulation in Jurkat-PD-1 cells (Fig. 2a), suggesting that PD-1 ligation promotes γ c degradation.γ c degradation after termination of protein synthesis by CHX was markedly inhibited by the lysosomal inhibitor ammonium chloride (NH 4 Cl) but not the proteasome inhibitor MG132 or autophagosome inhibitor 3-methyladenine (3-MA) in Jurkat cells (Fig. 2b).PD-L1-induced degradation of γ c was also inhibited by NH 4 Cl but not MG132 or 3-MA in Jurkat-PD-1 cells (Fig. 2c).Together, these results suggest that PD-1 signaling results in γ c degradation through the lysosomal route.
It has been demonstrated that polyubiquitination of proteins target them for degradation by various routes including the lysosomal pathway. 34Consistently, we found that PD-L1 induced polyubiquitination and down-regulation of γ c in Jurkat-PD1 cells, which was most dramatic at 2 days after PD-L1 stimulation (Fig. 2d).We next attempted to identify the E3 ubiquitin ligases that catalyze degradation of γ c .We screened 196 ubiquitin-related proteins for their abilities to regulate γ c level by co-transfection experiments in HEK293 cells.These efforts led to the identification of MARCH5, which caused down-regulation of γ c in a dosedependent manner (Fig. 2e) but had no marked effects on the γ c family cytokine receptor IL2RB, IL4R or IL7R (Supplementary information, Fig. S1a).There are 11 members in the MARCH E3 ligase family. 35Overexpression of MARCH5 but not other 10 MARCH family members markedly down-regulated the level of γ c (Supplementary information, Fig. S1b).These results suggest that MARCH5 specifically down-regulates γ c .
Previous studies showed that MARCH5 is a membraneembedded E3 ligase that is known to be present on the mitochondrial outer membrane or peroxisomal membrane. 36,37ince γ c is a plasma membrane receptor which is synthesized and translocated from the ER-Golgi-endosome system, 38 we next investigated whether MARCH5 can colocalize with γ c .Confocal microscopy experiments showed that γ c was partially colocalized with MARCH5 at the plasma membrane, endosomes and lysosomes but not at the mitochondria, ER, Golgi and peroxisomes (Supplementary information, Fig. S1c, d).Consistent with the previous reports, 36,37 MARCH5 was also partially located at the mitochondria and peroxisomes (Supplementary information, Fig. S1d).In addition, endogenous co-immunoprecipitation experiments indicated that γ c was basally associated with MARCH5 and PD-L1 stimulation promoted their association in Jurkat-PD-1 cells (Fig. 2f).These results support the conclusion that γ c is polyubiquitinated by MARCH5 and degraded by the lysosomal pathway.Interestingly, these experiments also indicated that MARCH5 was induced after PD-L1 stimulation, which reached to the highest level 3 days after PD-L1 stimulation, and the induction of MARCH5 was correlated with down-regulation of γ c (Fig. 2f).
We generated MARCH5-deficient Jurkat cells using the CRISPR-Cas9 method and found that MARCH5-deficiency up-regulated the protein level of γ c but not IL2Rβ (Fig. 2g).MARCH5-deficiency did not affect the mRNA level of γ c in Jurkat cells (Fig. 2h).In addition, MARCH5-deficiency increased γ c level in human T lymphoid leukemia HPB-ALL cells (Supplementary information, Fig. S2a) and murine T lymphocyte CTLL2 cells (Supplementary information, Fig. S2b).Furthermore, we generated March5 T cell-specific knockout mice (March5 f/f:CD4-Cre ) and isolated the thymocytes from March5 f/f or March5 f/f:CD4-Cre mice.The results showed that MARCH5-deficiency increased γ c level in thymocytes (Supplementary information, Fig. S2c).Previously, it has been demonstrated that phosphorylation of STAT5 Y694/Y699 is a hallmark of γ c family cytokine-triggered signaling activation. 15Consistently, MARCH5-deficiency increased the γ c family cytokine IL-7-and IL-9-induced phosphorylation of STAT5 Y694/Y699 in HPB-ALL cells, which has been shown to be responsive to the two cytokines (Supplementary information, Fig. S2a), 39 whereas MARCH5-deficiency increased IL-2induced phosphorylation of STAT5 Y694/Y699 in CTLL2 cells which was responsive to IL-2 (Supplementary information, Fig. S2b). 40In addition, Il-7-induced phosphorylation of STAT5 Y694/Y699 was increased in March5-deficient thymocytes in comparison to their wild-type counterparts (Supplementary information, Fig. S2c).These experiments also showed that γ c levels were not markedly changed following stimulation by the γ c family cytokines IL-2, IL-7 and IL-9 (Supplementary information, Fig. S2a-c).In addition, NH 4 Cl treatment inhibited MARCH5-mediated degradation of γ c in HEK293 cells (Supplementary information, Fig. S2d).Moreover, MARCH5deficiency impaired PD-L1-induced γ c degradation in Jurkat-PD-1 cells (Fig. 2i) and mouse CD8 + T cells (Fig. 2j).Taken together, these results suggest that MARCH5 mediates PD-L1-induced lysosomal degradation of γ c and inhibition of the γ c family cytokine-triggered signaling in various cells.
We next investigated the molecular mechanisms of MARCH5mediated polyubiquitination of γ c .Overexpression of MARCH5 but not its inactive mutant MARCH5 H43W promoted polyubiquitination T cells or PD-1-expressing Jurkat (Jurkat-PD-1) cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days.The cells were stained with the indicated antibodies and analyzed by flow cytometry.Graph shows mean ± SEM, n = 3 technical repeats.Data were analyzed using two-way ANOVA with GraphPad Prism 8. e The level of γ c is down-regulated after PD-1 ligation.Human CD8 + T cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before immunoblotting analysis with the indicated antibodies.f PD-1 ligation down-regulates γ c level.Left panels: Jurkat cells transduced with empty vector (Jurkat-EV) or Jurkat-PD-1 cells were analyzed by immunoblots with the indicated antibodies.Middle panels: Jurkat-EV and Jurkat-PD-1 cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) as indicated for 3 days before immunoblotting analysis with the indicated antibodies.Right panels: Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for the indicated times before immunoblotting analysis with the indicated antibodies.g Effects of PD-1 ligation on transcription of γ c .Human CD8 + T cells or Jurkat-PD-1 cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before qPCR analysis of mRNA levels of the indicated genes.Graph shows mean ± SEM, n = 3 independent samples from one representative experiment.Data were analyzed using two-way ANOVA with GraphPad Prism 8.The experiments in c-g were repeated for at least two times with similar results. of γ c in HEK293 cells (Fig. 2k).Utilizing ubiquitin mutants in which one or six lysine residues are replaced with arginine (R), we found that MARCH5 increased K27-but not other lysine residue-linked polyubiquitination of γ c (Supplementary information, Fig. S3a).Endogenous ubiquitination assays indicated that PD-L1 stimulation increased K27-linked polyubiquitination and degradation of γ c , and these effects were impaired in MARCH5-deficient Jurkat-PD-1 cells (Fig. 2l).Collectively, these results suggest that MARCH5 mediates K27-linked polyubiquitination and degradation of γ c following PD-1 ligation.
To further identify the residues in γ c that are conjugated with K27-linked polyubiquitin chains by MARCH5, we individually mutated each of the 4 lysine residues within the intracellular region (284-369 aa) of γ c , K294, K315, K338 and K363, to arginine and examined whether these mutants could be modified by K27linked polyubiquitination.The results indicated that mutation of Fig. 2 MARCH5 mediates K27-linked polyubiquitination and lysosomal degradation of γ c . a PD-1 ligation promotes γ c degradation.Jurkat-PD-1 cells were pre-stimulated with PHA (50 ng/mL) for 36 h and then treated with CHX (0.1 mM) for the indicated times before immunoblotting analysis with the indicated antibodies.The γ c band intensities relative to the corresponding β-actin bands were shown in the histograph.b NH 4 Cl inhibits γ c degradation.Jurkat cells were pre-treated with MG132 (100 μM), NH 4 Cl (25 mM) or 3-MA (500 ng/mL) for 4 h, and then treated with CHX (0.1 mM) for 2 h before immunoblotting analysis with the indicated antibodies.c NH 4 Cl inhibits PD-1 ligationinduced degradation of γ c .Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 36 h, and then treated with MG132 (100 μM), NH 4 Cl (25 mM) or 3-MA (500 ng/mL) for 12 h before immunoblotting analysis with the indicated antibodies.d PD-1 ligation induces polyubiquitination of γ c .Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.e MARCH5 promotes degradation of γ c in a dose-dependent manner.HEK293 cells were transfected with the indicated plasmids for 24 h before immunoblot analysis with the indicated antibodies.f Association of γ c with MARCH5.Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.g MARCH5-deficiency up-regulates the level of γ c but not IL2Rβ.Control or MARCH5-deficient Jurkat cells were collected for immunoblotting analysis with the indicated antibodies.h Effects of MARCH5-deficiency on transcription of γ c .Control or MARCH5-deficient Jurkat cells were analyzed by qPCR for mRNA levels of the indicated genes.Graph shows mean ± SEM, n = 3 independent samples from one representative experiment.Data were analyzed using two-way ANOVA with GraphPad Prism 8. i MARCH5-deficiency impairs PD-1 ligation-induced degradation of γ c .Control or MARCH5-deficient Jurkat-PD-1 cell were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before immunoblotting analysis with the indicated antibodies.j PD-1 ligation-induced degradation of γ c is impaired in MARCH5-deficient cells.Mouse CD8 + T cells from sex-and age-matched March5 f/f or March5 f/f:CD4-Cre mice were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound mPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before immunoblotting analysis with the indicated antibodies.k MARCH5 promotes polyubiquitination of γ c .HEK293 cells were transfected with the indicated plasmids for 24 h before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.l MARCH5-deficiency impairs PD-L1-induced K27linked polyubiquitination of γ c .Control or MARCH5-deficient PD-1-expressing Jurkat cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.m MARCH5 increases K27-linked polyubiquitination of wild-type γ c and γ c K294R but not γ c K315R .HEK293 cells were transfected with the indicated plasmids for 24 h before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.n PD-1 induces down-regulation of wild-type γ c but not γ c K315R .Jurkat-PD-1 cells were expressed with Flag-tagged wild-type γ c or γ c K315R mutant and then stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before immunoblotting analysis with the indicated antibodies.All the experiments were repeated for at least two times with similar results.
K315 but not the other 3 lysine residues in γ c to arginine dramatically reduced its K27-linked polyubiquitination, and MARCH5 increased K27-linked polyubiquitination and downregulation of wild-type γ c and γ c K294R but not γ c K315R (Fig. 2m).Consistently, PD-L1 stimulation induced down-regulation of wildtype γ c but not γ c K315R in Jurkat-PD-1 cells (Fig. 2n).Reconstitution of γ c K315R in γ c -deficient HPB-ALL cells increased IL-7-and IL-9induced phosphorylation of STAT5 Y694/Y699 in comparison to cells reconstituted with wild-type γ c (Supplementary information, Fig. S3b).In these reconstitution experiments, the mRNA levels of γ c K315R and wild-type γ c were comparable, but the protein level of γ c K315R was dramatically up-regulated compared to its wild-type counterparts (Supplementary information, Fig. S3b).Taken together, these results suggest that MARCH5 targets γ c K315 for its K27-linked polyubiquitination and degradation.

USP5 antagonizes MARCH5-mediated γ c polyubiquitination and degradation
We next attempted to identify deubiquitinate enzymes that are responsible for removing K27-linked polyubiquitin moieties conjugated to γ c .γ c -bound proteins were immunoprecipitated with anti-γ c and analyzed by mass spectrometry.Among the 184 proteins identified, 7 are deubiquitinate enzymes (Supplementary information, Table S1).Co-transfection experiments indicated that only USP5 but not the other 6 enzymes removed K27-linked polyubiquitin moieties from γ c in HEK293 cells (Fig. 3a).Endogenous co-immunoprecipitation experiments indicated that γ c was constitutively associated with USP5 in Jurkat-PD-1 cells (Fig. 3b).USP5 but not its enzymatic inactive mutant USP5 C335A removed K27-linked polyubiquitin moieties from γ c and up-regulated the level of γ c in HEK293 cells (Fig. 3c).In addition, USP5 removed K27linked polyubiquitination of γ c catalyzed by MARCH5 (Fig. 3d).In these experiments, it is notable that overexpression of USP5 upregulated whereas MARCH5 down-regulated the levels of γ c (Fig. 3c, d).Endogenous ubiquitination assays indicated that PD-L1 stimulation induced K27-linked polyubiquitination of γ c , which was increased in USP5-deficient Jurkat-PD-1 cells (Fig. 3e).In these experiments, the protein level of γ c in USP5-deficient cells was also dramatically down-regulated compared with those in wild-type cells (Fig. 3e).USP5-deficiency also down-regulated the level of γ c in HPB-ALL and CTLL2 cells, and inhibited IL-2-, IL-7-and IL-9induced phosphorylation of STAT5 at Y694/Y699 in these cells (Supplementary information, Fig. S4).These data suggest that USP5 positively regulates γ c level as well as the γ c family cytokinetriggered signaling in various cells.

PD-1 signaling promotes γ c degradation by inducing MARCH5 transcription
In our experiments, we found that PD-1 ligation had no marked effects on the protein level of USP5 as well as its association with γ c (Fig. 3b).USP5-deficiency down-regulated the basal level of γ c but did not affect PD-L1-induced γ c degradation (Fig. 3e).These results suggest that USP5 is not required for PD-1 ligation-induced γ c degradation.On the other hand, we found that the protein level of MARCH5 was not markedly changed after TCR activation, but was up-regulated following PD-L1 stimulation in human CD8 + T cells or Jurkat-PD-1 cells (Figs. 4a and 2f, i, j, l).PD-L1 stimulation did not affect the half-life of MARCH5 in CHX-treated Jurkat-PD-1 cells (Fig. 4b), suggesting that PD-1 ligation does not affect its stability.MARCH5-deficiency caused up-regulation of γ c protein level, which was reversed by reconstitution with wild-type MARCH5 but not MARCH5 H43W in Jurkat-PD-1 cells.Additionally, PD-1 ligation downregulated γ c level in wild-type but not MARCH5-deficient cells ectopically reconstituted with wild-type MARCH5 in Jurkat-PD-1 cells (Fig. 4c).qPCR analysis indicated that PD-L1 treatment upregulated MARCH5 mRNA level in human CD8 + T and Jurkat-PD-1 cells (Fig. 4d).These results suggest that PD-1 ligation regulates MARCH5 level at mRNA but not protein level.
It has been demonstrated that PD-1 ligation induces expression of certain transcription factors including the AP-1 family member BATF, which inhibits T cell function. 13,41Analysis of the Gene Transcription Regulation Database (GTRD, https://gtrd.biouml.org/)indicates that BATF can bind to the promoter region of MARCH5 gene.Consistently, chromatin immunoprecipitation (ChIP) experiments confirmed that BATF bound to the promoter region of MARCH5 gene in Jurkat-PD-1 cells (Fig. 4e).BATF-deficiency caused down-regulation and impairment of PD-L1-induced up-regulation of MARCH5 mRNA level in Jurkat-PD-1 cells (Fig. 4f).PD-1 ligation up-regulated the level of BATF, and BATF-deficiency up-regulated γ c level and inhibited PD-L1-induced down-regulation of γ c level in Jurkat-PD-1 cells (Fig. 4g).These results suggest that the basal level of γ c is constitutively regulated by basal BATF and PD-1 signaling transcriptionally induces MARCH5 in a BATF-dependent manner.

PD-1-triggered SHP2 activation induces dephosphorylation of γ c Y357
It has been shown that the allosteric dephosphorylating enzyme SHP2 is recruited to and activated in the complex of PD-1 upon PD-1 ligation, and the activated SHP2 mediates dephosphorylation of the TCR and CD28 proximal signaling molecules which leads to suppression of T cells. 13,31,33Our biochemical purification and mass spectrometry analysis indicated that SHP2 was a potential γ c -bound proteins (Supplementary information, Table S1).Endogenous co-immunoprecipitation experiments indicated that γ c was weakly associated with SHP2 and PD-L1 stimulation promoted their association at 5 min post stimulation in Jurkat-PD-1 cells (Fig. 5a).These results suggest that PD-1 ligation promotes a physical association of SHP2 with γ c .
We next investigated whether SHP2 mediates dephosphorylation of γ c .It has been demonstrated that the tyrosine kinase JAK3 mediates phosphorylation of γ c upon stimulation of the γ c family cytokines.Overexpression of the SHP2 gain-of-function mutant D61A or E76K but not the wild-type SHP2 or dominant negative SHP2 mutant C459S abolished JAK3-mediated phosphorylation of γ c (Fig. 5b).In similar experiments, SHP2 was not associated with JAK3 and overexpression of the SHP2 gain-of-function mutant D61A or E76K did not abolish the phosphorylation of JAK3 (Supplementary information, Fig. S5a), suggesting that SHP2 does not target JAK3.The level of JAK3 was not markedly changed in SHP2-deficient HPB-ALL cells comparing to their wild-type counterparts (Supplementary information, Fig. S5b), There are 4 tyrosine residues in the intracellular region (284-369 aa) of γ c (Supplementary information, Fig. S5c).Mutagenesis indicated that mutation of Y357 but not the other 3 tyrosine residues to phenylalanine impaired JAK3-mediated tyrosine phosphorylation of γ c (Supplementary information, Fig. S5d).Sequence analysis indicated that γ c Y357 was conserved in various vertebrate species (Supplementary information, Fig. S5c).To determine whether SHP2 dephosphorylates γ c Y357 , we generated a rabbit polyclonal antibody specific for Y357-phosphorylated γ c (p-γ c Y357 ).Immunoblotting analysis confirmed that γ c Y357 was phosphorylated following co-expression of JAK3, which was reversed by the gain-of-function mutant SHP2 D61G in HEK293 cells (Fig. 5c).Endogenous γ c Y357 was phosphorylated following IL-7 stimulation, which was not seen in un-stimulated or γ c -deficient HPB-ALL cells (Fig. 5d).Since Jurkat is a γ c -dependent cell line in which γ c -mediated signaling is constitutively activated, γ c Y357 was constitutively phosphorylated and PD-L1 stimulation caused dephosphorylation of γ c Y357 as early as 5 min post stimulation, while SHP2-deficiency abolished PD-L1-induced dephosphorylation of γ c Y357 in Jurkat-PD-1 cells (Fig. 5e).SHP2-deficiency also enhanced IL-7-induced phosphorylation of γ c Y357 (Fig. 5f) as well as IL-2-, IL-7-and IL-9-induced phosphorylation of STAT5 Y694/Y699 in HPB-ALL and CTLL2 cells, respectively (Fig. 5f, g), suggesting that SHP2 negatively regulates γ c family cytokine-triggered signaling.In addition, the level of γ c was not markedly changed in SHP2-deficient HPB-ALL or Jurkat-PD-1 cells comparing to their wild-type counterparts (Fig. 5e, f), suggesting that JAK3-mediated phosphorylation and SHP2-mediated dephosphorylation of γ c is not involved in MARCH5-mediated down-regulation of γ c .Reconstitution of wild-type γ c but not γ c Y357F in γ c -deficient HPB-ALL cells restored IL-7-and IL-9-induced phosphorylation of STAT5 Y694/Y699 (Fig. 5h).These results suggest that the phosphorylation of γ c Y357 is required for γ c family cytokine-triggered signaling activation, while PD-1 ligation triggers SHP2-mediated dephosphorylation and inactivation of γ c .

MARCH5 knockdown improves anti-tumor immunity and suppresses tumor growth
It has been reported that March5 homozygous deletion is embryonically lethal. 42To investigate the physiological functions of MARCH5, March5 flox/flox mice were generated and crossed with Vav1-Cre mice to obtain March5 hematopoietic-specific knockout strain (March5 f/f:Vav1-Cre ).Unexpectedly, the March5 f/f:Vav1-Cre mice were born normally, but all died during 4-6 weeks after birth.Thus, we used March5 +/f and March5 +/f:Vav1-Cre mice for further investigation.We verified that the mRNA and protein levels of MARCH5 in the bone marrow (BM), spleen and thymus of March5 +/f:Vav1-Cre mice were about half to that in March5 +/f mice (Supplementary information, Fig. S6a, b).Flow cytometry analysis indicated that MARCH5 knockdown up-regulated the level of γ c in CD4 + T, CD8 + T, NK and B cells from the spleen (Fig. 6a).MARCH5 knockdown had no marked effects on the percentages of doublenegative (DN), double-positive (DP), CD4 + single-positive (CD4SP) cells, but increased the percentage of CD8 + single-positive (CD8SP) cells in the thymus (Fig. 6b; Supplementary information, Fig. S6c).March5 +/f:Vav1-Cre mice exhibited significantly higher percentages of CD8 + T and NK cells and slightly lower percentages of B and CD4 + T cells in spleen and the peripheral blood (Fig. 6c; Supplementary information, Fig. S6d).The percentage of CD44 high CD62L high central memory (CM) CD8 + T cells was also increased in total CD8 + T cells of spleen and the peripheral blood of March5 +/f:Vav1-Cre mice (Fig. 6c).The percentage of CD44 high CD62L high central memory CD4 + T cells did not show significant difference in CD4 + T cells of spleen and the peripheral blood of March5 +/f:Vav1-Cre mice (Supplementary information, Fig. S6d).These results suggest that MARCH5 negatively regulates γ c level as well as CD8 + T and NK cell development in mice.
We next used mouse MC38 colorectal carcinoma and B16F10 melanoma models to investigate the biological functions of MARCH5 in anti-tumor immunity.In these models, March5 +/f:Vav1-Cre mice showed slower tumor progression than March5 +/f mice (Fig. 6d; Supplementary information, Fig. S7a).We isolated tumor-infiltrating lymphocytes (TILs) from the tumor tissues of March5 +/f and March5 +/f:Vav1-Cre mice and analyzed the immunological changes in tumors.The results showed that March5 +/f:Vav1-Cre mice exhibited significantly higher percentages of CD8 + T, NK and central memory CD8 + T cells in TILs comparing to March5 +/f mice (Fig. 6e), while the percentages of CD4 + T, Treg and B cells in TILs were similar between March5 +/f and March5 +/f:Vav1-Cre mice (Supplementary information, Fig. S7b, c).To determine whether MARCH5 knockdown affects the activation of tumor-infiltrating CD8 + T cells or the profile of exhausted T cells, we examined the T cell activation maker Granzyme B (GzmB) and the exhausted T cell marker TIM3 on infiltrated CD8 + T cells in the mouse tumor models.The results showed that the percentage of TIM3 + cells in CD8 + T cells from the March5 +/f:Vav1-Cre mice was down-regulated, while the percentage of GzmB + cells in CD8 + T cells did not show marked changes (Supplementary information, Fig. S7b, c), suggesting that knockdown of MARCH5 inhibited exhaustion of CD8 + T cells and did not affect the CD8 + T cell response.Taken together, these data suggest that MARCH5 negatively regulates anti-tumor immunity in mouse models.
Based on their powerful abilities to stimulate proliferation of cytotoxic CD8 + T and NK cells, certain γ c family cytokines such as IL-2 have long been used in clinical cancer immunotherapy. 27Since MARCH5 knockdown leads to higher level of γ c in CD8 + T and NK cells, we reasoned that this would improve the efficacy of the γ c family cytokines in anti-tumor immunotherapy.To test this, we used the B16F10 melanoma and MC38 colon cancer mouse tumor models.Administration of IL-2 in March5 +/f:Vav1-Cre mice showed increased efficacy in suppression of tumor growth and improvement of the overall survival comparing to that in March5 +/f mice (Fig. 6f, g; Supplementary information, Fig. S7d).We further combined IL-2 with PD-1 blocking antibody for tumor immunotherapy in March5 +/f and March5 +/f: Vav1-Cre mice.The results showed that the anti-tumor effects of combined IL-2 with PD-1 blockade in March5 +/f:Vav1-Cre mice were superior to that in March5 +/f mice, which eliminated tumors in 2 out of 8 March5 +/f:Vav1-Cre mice (Fig. 6f, g).In addition, MARCH5-deficiency in B16F10 and MC38 cells had no marked effects on their proliferation (Supplementary information, Fig. S8).Taken together, these results suggest that MARCH5 knockdown sensitizes the anti-tumor effects of IL-2 as well as its combination with PD-1 blocking antibody.
Combination of a MARCH5 inhibitor with IL-2 and PD-1 blockade significantly increases anti-tumor efficacy Our results reveal an important role of MARCH5 in PD-1-triggered immune suppression, which suggest that MARCH5 is a potential target for cancer immunotherapy.Unfortunately, no pharmacological inhibitors of MARCH5 have been reported.We designed a reporter system in which HEK293 cells are co-transfected with two plasmids encoding MARCH5 and γ c -luciferase fusion protein respectively.In this system, inhibition of MARCH5 by an inhibitor would result in increased luciferase activity in reporter assays (Supplementary information, Fig. S9a).Using this system, we screened a collection of Food and Drug Administration-approved drugs and identified 10 compounds that increased luciferase activity to more than 1.5-folds (Supplementary information, Fig. S9b and Table S2).Further confirmative experiments identified Pitavastatin calcium (PC) as the most potent inhibitor that specifically increased the luciferase activity in MARCH5expressing but not control HEK293 cells (Supplementary information, Fig. S9c).Pitavastatin is a unique lipophilic statin and potent inhibitor of HMG-CoA reductase with a strong effect on lowering plasma total cholesterol and triacylglycerol levels.Pitavastatin has also been reported to have pleiotropic beneficial effects such as suppression of inflammation, regulation of angiogenesis and osteogenesis, improvement of endothelial function and arterial stiffness. 43Recent studies have also shown that PC has anti-tumor activity by promoting tumor cell apoptosis. 44,45Consistent with a specific inhibition of MARCH5 activity in the reporter system, PC treatment up-regulated the level of γ c in a dose-dependent manner in all examined cells including primary human and mouse CD8 + T cells, and Jurkat, HPB-ALL and CTLL2 cells (Fig. 7a).It has been shown that PC as well as other statin drugs such as Lovastatin, Simvastatin, Fluvastatin Sodium and Rosuvastatin Fig. 3 USP5 removes K27-linked polyubiquitin moieties from γ c . a USP5 removes K27-linked polyubiquitin moieties from γ c .HEK293 cells were transfected with the indicated plasmids for 24 h before immunoblotting analysis with the indicated antibodies.b Association of γ c with USP5.Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.c USP5 but not its enzymatic inactive mutant USP5 C335A removes K27-linked polyubiquitin moieties from γ c .HEK293 cells were transfected with the indicated plasmids for 24 h before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.d USP5 removes K27-linked polyubiquitin moieties of γ c catalyzed by MARCH5.HEK293 cells were transfected with the indicated plasmids for 24 h before immunoblotting analysis with the indicated antibodies.e USP5-deficiency enhances PD-1 ligation-induced K27-linked polyubiquitination of γ c .Control or USP5-deficient Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.All the experiments were repeated for at least two times with similar results.
Calcium target HMG-CoA reductase. 46Therefore, we examined the effects of other statin drugs on γ c level.The results indicated that only PC but not the other examined statin drugs including Lovastatin, Simvastatin, Fluvastatin Sodium and Rosuvastatin Calcium up-regulated γ c level (Supplementary information, Fig. S9d), suggesting that inhibition of HMG-CoA reductase does not have an effect on γ c level.In MARCH5-deficient cells, the basal level of γ c was increased and PC treatment did not further increase its protein level (Fig. 7b).There results suggest that PC is a specific inhibitor of MARCH5 and capable of inhibiting MARCH5-mediated γ c degradation independent of its inhibition of HMG-CoA reductase.
We further tested the anti-tumor effects of PC in mouse MC38 colorectal carcinoma and B16F10 melanoma models.Administration of PC alone significantly suppressed tumor growth in mice (Fig. 7c).Administration of PC also resulted in higher level of γ c in tumor-infiltrating lymphocytes (Fig. 7d).The percentages of CD8 + T, NK, CD4 + T, Treg and B cells in tumor-infiltrating lymphocytes did not show significant difference after PC administration (Supplementary information, Fig. S10a).PC treatment upregulated the percentage of GzmB + cells and reduced the percentage of TIM3 + cells in infiltrated CD8 + T cells, suggesting that PC enhances infiltration of CD8 + cytotoxic T cells and inhibits exhaustion of CD8 + T cells (Fig. 7e).
We further investigated the anti-tumor effects of PC combined with IL-2 and PD-1 blockade (Supplementary information, Fig. S10b).Consistently, PC significantly increased the efficacies of IL-2 or IL-2 plus PD-1 blocking antibody on tumor suppression (Fig. 7f) and overall survival of mice (Fig. 7g).In addition, we investigated whether the other HMG-CoA reductase inhibitors such as Rosuvastatin Calcium affect the anti-tumor efficacies of IL-2.Administration of PC or Rosuvastatin Calcium alone suppressed tumor growth in mice.However, only PC but not Rosuvastatin Calcium has a synergistic effect with IL-2 on tumor suppression (Supplementary information, Fig. S10c), suggesting that PC promotes anti-tumor immunity via its inhibition of MARCH5 but not HMG-CoA reductase.Taken together, these results suggest that administration of MARCH5 inhibitor leads to increased efficacies of cancer immunotherapy by IL-2 plus PD-1 blockade.

DISCUSSION
Combination therapy with PD-1 blockade and IL-2 substantially improves anti-tumor efficacy compared with PD-1 blockade or IL-2 monotherapy, which has been extensively explored for immunotherapy of various cancers. 15,21,26,47The basic principle of combined immunotherapy with PD-1 blockade and IL-2 is to remove the PD-1 inhibitory brake and in the meantime provide a stimulatory signal for cytotoxic T lymphocytes with IL-2.Recent studies have also shown that combination immunotherapy with PD-1 blockade and IL-2 modifies CD8 + T cell exhaustion program. 47,48Currently, the cellular and molecular basis responsible for the synergistic effects observed between PD-1 blockade and IL-2 remains enigmatic.In this study, we showed that PD-1 signaling negatively regulates the stability and activity of γ c , which impairs the γ c family cytokine-triggered signaling and immune activation of CD8 + T cells in the TME.PD-1 blockade removes the inhibitory effects on γ c , thereby restoring the responses of CD8 + T cells to the γ c family cytokines such as IL-2 and leading to the synergistic effects of combination immunotherapy of PD-1 blockade and IL-2.Most importantly, we uncovered the two distinct molecular mechanisms responsible for PD-1 signaling-triggered inhibition of γ c -mediated immune activation, and validated the membraneassociated E3 ligase MARCH5 as a potential target for combination immunotherapy.
0][51][52][53] To evade host immune surveillance, tumor cells take advantage of this inhibitory pathway by overexpressing PD-L1. 12,54γ c is a component of the receptors for IL-2, IL-4, IL-7, IL-9, IL-15 and IL-21, which is constitutively expressed on various populations of immune cells. 19In our experiments, we found that the level of γ c in TME was negatively correlated with PD-L1.PD-1 blockade in mouse tumor model upregulated the level of γ c in tumor-infiltrating CD8 + T cells.6][57] Our results confirmed that the level of γ c was not markedly changed after T cell activation or γ c family cytokine stimulation.However, PD-1 ligation promoted lysosomal degradation of γ c .
In a screen for potential E3 ubiquitin ligases that mediate γ c degradation, we identified MARCH5, which is a member of the membrane-associated E3 ubiquitin ligase family. 35Previous studies have demonstrated that MARCH5 is involved in regulation of mitochondrial morphology and pexophagy. 36,37,58Our results indicated that MARCH5 was also colocalized with γ c at the plasma membrane and intracellular membrane organelles such as endosome and lysosome.Endogenous coimmunoprecipitation experiments indicated that MARCH5 was associated with γ c , which was increased following PD-L1 stimulation.Overexpression of MARCH5 but not its inactive mutant MARCH5 H43W promoted K27linked polyubiquitination and lysosomal degradation of γ c .Fig. 4 PD-1 signaling promotes γ c degradation by inducing transcription of MARCH5.a MARCH5 is up-regulated after PD-L1 but not PHA stimulation.Human CD8 + T cells or Jurkat-PD-1 cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/ mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before immunoblotting analysis with the indicated antibodies (left panels).Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for the indicated times before immunoblotting analysis with the indicated antibodies (right panels).b Effects of PD-1 ligation on MARCH5 degradation.Jurkat-PD-1 cells were pre-stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 36 h and then treated with CHX (0.1 mM) for the indicated times before immunoblotting analysis with the indicated antibodies.The MARCH5 band intensities relative to the corresponding β-actin bands were shown in the histograph.c MARCH5-deficiency impairs PD-1 ligation-induced degradation of γ c .Control or MARCH5-deficient Jurkat-PD-1 cells were reconstituted with wild-type MARCH5 or MARCH5 H43W mutant and then stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 2 days before immunoblotting analysis with the indicated antibodies.d Effects of PD-1 ligation on MARCH5 mRNA level.Human CD8 + T cells or Jurkat-PD-1 cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before qPCR analysis of mRNA levels of the indicated genes.Graph shows mean ± SEM, n = 3 independent samples from one representative experiment.Data were analyzed using two-way ANOVA with GraphPad Prism 8. e BATF binds to the promoter region of MARCH5 gene.Jurkat-PD-1 cells were analyzed by ChIP with the indicated antibodies, and then de-crosslinked DNA was subjected to qPCR analysis using specific primers.Graph shows mean ± SEM, n = 3 independent samples from one representative experiment.Data were analyzed using a Student's unpaired ttest with GraphPad Prism 8. f BATF-deficiency impairs PD-1 ligation-induced transcription of MARCH5.Control or BATF-deficient Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 2 days before qPCR analysis of mRNA levels of the indicated genes.Graph shows mean ± SEM, n = 3 independent samples from one representative experiment.Data were analyzed using two-way ANOVA with GraphPad Prism 8. g BATF-deficiency impairs PD-1 ligation-induced degradation of γ c .Control or BATF-deficient Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 2 days before immunoblotting analysis with the indicated antibodies.All the experiments were repeated for at least two times with similar results.
MARCH5-deficiency up-regulated the level of γ c , which was reversed by reconstitution with wild-type MARCH5 but not MARCH5 H43W .PD-1 ligation induced K27-linked polyubiquitination and lysosomal degradation of γ c , which were blocked in MARCH5deficient cells.In contrast, PD-1 ligation had no marked effects on the protein level of γ c in MARCH5-deficient cells reconstituted with ectopically-expressed wild-type MARCH5.These results suggest that PD-1 signaling leads to MARCH5-mediated K27-linked polyubiquitination and degradation of γ c , but does not affect MARCH5 activity per se.Mechanistically, we found that PD-1 ligation induced the transcription factor BATF, which transcriptionally induced MARCH5 by binding to the promoter of MARCH5 gene.BATF-deficiency impaired PD-L1-induced up-regulation of MARCH5 mRNA and down-regulation of γ c protein level.Our experiments suggest that PD-1 signaling induces the transcription factor BATF, which in turn induces expression of the E3 ubiquitin ligase MARCH5, leading to K27-linked polyubiquitination and lysosomal degradation of γ c .
Our experiments further identified USP5 as an enzyme that constitutively deubiquitinates γ c .USP5 removed K27-linked polyubiquitin moieties of γ c conjugated by MARCH5.Knockout of USP5 increased PD-1 ligation-induced K27-linked polyubiquitination of γ c and down-regulated its protein level in cells.These results suggest that USP5 acts as a constitutive guard for γ c stability to ensure proper responses of CD8 + T cells to the γ c family cytokines.Consistently, our experiments showed that MARCH5-deficiency potentiated the γ c family cytokine-triggered signaling and immune activation, whereas USP5-deficiency had the opposite effects.Taken together, our experiments suggest that the BATF-MARCH5-γ c axis mediates PD-1-triggered inhibition of γ c family cytokine-triggered signaling and immune activation.
Our experiments also suggest that SHP2 mediates another mechanism responsible for PD-1-triggered inhibition of γ c family cytokine-triggered signaling and immune activation.Previously, it has been shown that PD-1 ligation triggers clustering of PD-1 with TCR, and PD-1 is phosphorylated by TCR proximal Src family kinases. 13,31,33Phosphorylation of PD-1 recruits and activates the allosteric dephosphorylating enzyme SHP2, which mediates dephosphorylation of TCR proximal signaling molecules.The clustering of PD-1 and TCR appears to be required for PD-1 phosphorylation and SHP2 recruitment.However, PD-1 is partially segregated from TCR after clustering, while it interacts with SHP2 constitutively, 30,33,59 suggesting that PD-1-associated SHP2 has the potential to mediate dephosphorylation of molecules that is not part of the TCR signalosome.Consistently, T cell costimulatory receptor CD28, which is not a TCR-associated component, is dephosphorylated by PD-1-SHP2. 30In our experiments, we found that PD-1 ligation promoted the association of γ c with SHP2.Overexpression of gain-of-function SHP2 mutants (D61G or E76K) reduced JAK3-mediated phosphorylation of γ c Y357 .Knockout of SHP2 impaired PD-1 ligation-induced γ c Y357 dephosphorylation.Knockout of SHP2 also increased the γ c family cytokine-induced phosphorylation of γ c Y357 and STAT5 Y694/Y699 .Reconstitution of wild-type γ c but not γ c Y357F in γ c -deficient HPB-ALL cells restored IL-7-and IL-9-induced phosphorylation of STAT5 Y694/Y699 .These results suggest that PD-1 ligation-triggered SHP2 activation induces dephosphorylation of γ c Y357 , resulting in desensitization of γ c -mediated signaling and immune activation.In our study, we found that the protein levels of γ c in MARCH5-deficient cells reconstituted with ectopically-expressed wild-type MARCH5 did not markedly changed after PD-1 ligation, suggesting that SHP2-mediated dephosphorylation of γ c is not involved in MARCH5mediated down-regulation of γ c .
Based on our results, we propose a model on the regulatory mechanisms of γ c stability and activity by PD-1 signaling.In TME with high expression of PD-L1 in tumor cells, PD-1 signaling in immune cells is hijacked and activated.The activated PD-1 recruits and activates SHP2, which subsequently mediates dephosphorylation of γ c Y357 , leading to its inactivation and unresponsiveness to γ c family cytokines.On the other hand, PD-1 signaling induces the transcription factor BATF, which induces expression of the membrane-associated E3 ubiquitin ligase MARCH5.MARCH5 is recruited to γ c and mediates its K27-linked polyubiquitination at K315 and lysosomal degradation.Therefore, PD-1 signaling suppresses the γ c family cytokine-triggered immune activation via two distinct mechanisms.As shown in our experiments, PD-1triggered SHP2 activation and dephosphorylation of γ c occurred in minutes, whereas PD-1-triggered induction of MARCH5 and degradation of γ c was obvious one day after PD-1 ligation; thus, we propose that PD-1 signaling inhibits γ c family cytokine-triggered immune activation via the two mechanisms in a temporal manner.In this context, it has previously been shown that PD-1 signaling causes dephosphorylation of TCR proximal molecules and subsequent transcriptional inhibition. 13,41,59Targeting of components involved in these regulatory mechanisms, such as SHP2, MARCH5 and USP5, may increase the efficacy of cancer immunotherapy by PD-1 blockade and the γ c family cytokines (Fig. 8).Consistently, our experiments indicated that MARCH5-knockdown up-regulated the level of γ c in lymphocytes, increased the percentage of CD8 + singlepositive (CD8SP) cells in thymocytes and enhanced the percentages of CD8 + T and NK cells in the spleen and peripheral blood in mice.MARCH5-knockdown suppressed tumor growth, and sensitized the anti-tumor effects of IL-2 as well as its combination with PD-1 blocking antibody in mice.
In our study, we also identified PC as an inhibitor of MARCH5.Our experiments indicated that PC up-regulated the level of γ c in a MARCH5-dependent manner.In MARCH5-deficient cells, the basal level of γ c was increased and PC treatment did not further increase its protein level.These results suggest that PC inhibits MARCH5 activity.PC is a unique lipophilic statin and potent inhibitor of HMG-CoA reductase.PC but not the other HMG-CoA reductase inhibitors such as Lovastatin, Simvastatin, Fluvastatin Sodium and Rosuvastatin Calcium treatment up-regulated the level of γ c , suggesting that inhibition of HMG-CoA reductase does not have an effect on γ c level and PC-induced downregulation of γ c is independent of its inhibition of HMG-CoA reductase.Consistently, PC treatment up-regulated the level of γ c in tumor-infiltrating lymphocytes, inhibited tumor growth, and sensitized the anti-tumor effects of IL-2 as well as its Fig. 5 PD-1 signaling activates SHP2 to induce dephosphorylation of γ c Y357 .a Association of γ c with SHP2.Jurkat-PD-1 cells were stimulated with PHA (150 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for the indicated times before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.b Effects of SHP2 on dephosphorylation of γ c .HEK293 cells were transfected with the indicated plasmids for 24 h before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.c JAK3 mediates phosphorylation of γ c Y357 .HEK293 cells were transfected with the indicated plasmids for 24 h before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.d IL-7 treatment induces γ c Y357 phosphorylation.Control or γ c -deficient HPB-ALL cells were stimulated with IL-7 (100 ng/mL) for the indicated times before co-immunoprecipitation and immunoblotting analysis with the indicated antibodies.e SHP2-deficiency impairs PD-L1-induced γ c Y357 dephosphorylation.Control or SHP2-deficient Jurkat-PD-1 cells were stimulated with PHA (150 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein for the indicated times before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.f SHP2-deficiency enhances IL-7-induced γ c Y357 phosphorylation.Control or SHP2-deficient HPB-ALL cells were stimulated with IL-7 (100 ng/mL) for the indicated times before coimmunoprecipitation and immunoblotting analysis with the indicated antibodies.g SHP2-deficiency promotes γ c family cytokine-induced phosphorylation of STAT5 Y694/Y699 .Control or SHP2-deficient HPB-ALL cells were stimulated with IL-9 (100 ng/mL) for the indicated times before immunoblotting analysis with the indicated antibodies (left panels).Control or SHP2-deficient CTLL2 cells were stimulated with IL-2 (400 IU/mL) for the indicated times before immunoblotting analysis with the indicated antibodies (right panels).h Effects of γ c Y357F mutant on the γ c family cytokine-induced phosphorylation of STAT5 Y694/Y699 .γ c -deficient HPB-ALL cells were reconstituted with wild-type γ c or γ c Y357F mutant and then stimulated with IL-7 (100 ng/mL) or IL-9 (100 ng/mL) for the indicated time lengths before immunoblotting analysis with the indicated antibodies.All the experiments were repeated for at least two times with similar results.combination with PD-1 blocking antibody.In addition, only PC but not Rosuvastatin Calcium had a synergistic effect with IL-2 on tumor suppression, suggesting that PC promotes anti-tumor immunity via its inhibition of MARCH5 but not HMG-CoA reductase.Previous studies have demonstrated that MARCH5 inhibits mitochondrial fission by promoting Drp1 degradation, 36,60 whereas other studies suggest that MARCH5 is required for Drp1-dependent mitochondrial division. 58,61In single-positive (CD8SP).The CD4 -CD8 -double-negative (DN) cells were analyzed by flow cytometry for the percentage of CD44 + singlepositive (DN1), CD44 + CD25 + double-positive (DN2), CD25 + single-positive (DN3) and CD44 -CD25 -double-negative (DN4).Graph shows mean ± SEM, n = 6 independent samples.Data were analyzed using Student's unpaired t-test with GraphPad Prism 8. c Effects of MARCH5 knockdown on the percentages of CD8 + T and NK cells from spleen and the peripheral blood.Splenocytes or peripheral blood leukocytes from sex-and age-matched March5 +/f or March5 +/f:Vav1-Cre mice were analyzed by flow cytometry for the percentage of CD8 + T (CD3 + CD8 + ), NK (CD3 -NKp46 + ), native CD8 + T (CD44 low CD62L high CD8 + T cells), central memory CD8 + T (CD44 high CD62L high CD8 + T cells, CM) and effector/ effector memory CD8 + T (CD44 high CD62L low CD8 + T cells, Effector/EM) cells.Graph shows mean ± SEM, n = 6 independent samples.Data were analyzed using Student's unpaired t-test with GraphPad Prism 8. d MARCH5 knockdown inhibits tumor growth.Sex-and age-matched March5 +/f or March5 +/f:Vav1-Cre mice were subcutaneously injected with MC38 cells (5 × 10 5 ).On day 3 after tumor cell inoculation, tumor sizes were measured every two days by caliper.Tumor-bearing mice were euthanized on day 13, and then tumor tissues were separated from the mice.Tumor weights were measured by Analytical Balance.Graph shows mean ± SEM, n = 6.Data were analyzed using Student's unpaired ttest with GraphPad Prism 8. e MARCH5 knockdown increases the percentages of CD8 + T and NK cells in TILs.TILs were isolated from the MC38 tumor tissues in d.TILs were stained with the indicated antibodies and analyzed by flow cytometry.Graph shows mean ± SEM, n = 6 independent samples.Data were analyzed using Student's unpaired t-test with GraphPad Prism 8. f Combination of IL-2 and PD-1 blockade has increased anti-tumor efficacy in March5 +/f:Vav1-Cre mice.March5 +/f and March5 +/f:Vav1-Cre mice were subcutaneously injected with MC38 cells (5 × 10 5 ).On day 5 after tumor cell inoculation, mice were intraperitoneally injected with control, IL-2 (50,000 IU per mouse) or anti-PD-1 (100 μg per mouse) (Supplementary information, Fig. S7d).Tumor sizes were measured every two days by caliper from day 5. WT: March5 +/f , CHZ: March5 +/f:Vav1-Cre .Graph shows mean ± SEM, n = 8.Data were analyzed using two-way ANOVA with GraphPad Prism 8. g Combination of IL-2 and PD-1 blockade increases the survival rate in March5 +/f:Vav1-Cre mice.March5 +/f and March5 +/f:Vav1-Cre mice were subcutaneously injected with MC38 cells (5 × 10 5 ).On day 5 after tumor cell inoculation, mice were intraperitoneally injected with control, IL-2 (50,000 IU per mouse) or anti-PD-1 (100 μg per mouse) (Supplementary information, Fig. S7d).Mice were sacrificed when the tumor size was bigger than 15 mm of the mean tumor diameter, tumor volume exceeded 2000 mm 3 , or tumor had ulcers with diameter reached 10 mm.Statistical analysis was performed using the GraphPad Prism 8 software, n = 8.Kaplan-Meier survival curves and corresponding log-rank (Mantel-Cox) tests were used to evaluate the statistical differences between groups in survival studies.There is a significant difference when the P < 0.05.addition, it has also been reported that MARCH5 is partially localized at the peroxisome to regulate pexophagy. 37Whether and how the functions of MARCH5 in these cellular processes other than its down-regulation of γ c are involved in its regulation of anti-tumor immunity needs to be investigated in future studies.Given its important role of in regulating γ c stability, USP5 can be a potential target for cancer immunotherapy.Unfortunately, USP5 is aberrantly expressed in human cancers and promotes tumor growth and metastasis, [62][63][64] which may limit its application as a target for cancer immunotherapy.
In conclusion, our current study suggests that MARCH5mediated degradation and SHP2-mediated inactivation of γ c triggered by PD-1 signaling render CD8 + T cells in TME irresponsive to immune activation by the γ c family of cytokines.PD-1 blockade therapy restores γ c stability and activity, thereby sensitizing CD8 + T cells to IL-2.Our findings define the underlying mechanisms on how IL-2 combined with PD-1 blockade can be synergistic in tumor immunotherapy.As validated in our study, regulation of these mechanisms would provide new strategies for improved efficacy of combination immunotherapy of cancer.).On day 5 after tumor cell implantation, mice were intraperitoneally injected with control, PC (5 mg/kg), IL-2 (50,000 IU per mouse) or anti-PD-1 (100 μg per mouse) (Supplementary information, Fig. S10b).Tumor sizes were measured every two days by caliper from day 5. Graph shows mean ± SEM, n = 8.Data were analyzed using two-way ANOVA with GraphPad Prism 8. g PC promotes the survival rate of mice treated with IL-2 and PD-1 blockade.C57BL/6J mice were subcutaneously injected with MC38 cells (5 × 10 5 ).On day 5 after tumor cell implantation, mice were intraperitoneally injected with control, PC (5 mg/kg), IL-2 (50,000 IU per mouse) or anti-PD-1 (100 μg per mouse) (Supplementary information, Fig. S10b).Mice were sacrificed when the tumor size is bigger than 15 mm of the mean tumor diameter, tumor volume exceeded 2000 mm 3 , or tumor had ulcers with diameter reached 10 mm.Statistical analysis was performed using the GraphPad Prism 8 software, n = 8.Kaplan-Meier survival curves and corresponding log-rank (Mantel-Cox) tests were used to evaluate the statistical differences between groups in survival studies.There is a significant difference when the P < 0.05.

Cells
Jurkat cells were obtained from American Type Culture Collection.HEK293 cells were originally provided by Dr. Gary Johnson (National Jewish Health, Denver, CO).HPB-ALL cells were provided by Dr. Hudan Liu (Wuhan university).CTLL2 cells were obtained from Cell Resource Center (IBMS, CAMS/PUMC).Human CD8 + T cells were obtained from Milestone® Biotechnologies.Mouse CD8 + T cells were isolated from the spleen of 6-8-week-old C57BL/6 mice by negative selection magnetic beads (STEM CELL Technologies, Catalog #19853 A).B16F10 cells, CT26 cells and MC38 cells were provided by Dr. Jinfang Zhang (Wuhan University).

Transfection
HEK293 cells were transfected by standard calcium phosphate precipitation.The empty control plasmid was added to ensure that each transfection receives the same amount of total DNA.

CRISPR-Cas9 knockout
Double-stranded oligonucleotides corresponding to the target sequences were cloned into the Lenti-CRISPR-V2 vector, which were co-transfected with the packaging plasmids into HEK293 cells.Two days after transfection, the viruses were harvested, ultra-filtrated (0.45-μm filter, Millipore) and used to infect cells in the presence of polybrene (8 μg/mL).The infected cells were selected with puromycin (Jurkat: 1 μg/mL, HPB-ALL: 2 μg/mL, CTLL2: 4 μg/mL) for at least 6 days.The information of gRNA sequences is shown in Supplementary information, Table S4.

Co-immunoprecipitation, ubiquitination and immunoblotting analysis
Cells were lysed in 1 mL of NP-40 lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1% Triton X-100, 10 μg/mL aprotinin, 10 μg/mL leupeptin, and 1 mM phenylmethylsulfonyl fluoride, PMSF).For each immunoprecipitation reaction, a 0.4 mL aliquot of lysate was incubated with 0.5-2 μg of the indicated antibody or control IgG and 35 μL of a 1:1 slurry of Protein-G Sepharose at 4 °C for 3 h.The Sepharose beads were washed three times with 1 mL of lysis buffer containing 500 mM NaCl.The precipitates were fractionated by SDS-PAGE, and immunoblotting analysis was performed with the indicated antibodies.For ubiquitination assays, the immunoprecipitants were re-extracted in NP-40 lysis buffer containing 1% SDS and denatured by heating for 10 min.The supernatants were diluted with regular lysis buffer until the concentration of SDS was decreased to 0.1%, following by re-immunoprecipitation with the indicated antibodies.The immunoprecipitants were analyzed by immunoblotting with the ubiquitin antibody.qPCR Total RNA was isolated for qPCR analysis to measure mRNA abundance of the indicated genes.Data shown are the relative abundance of the indicated mRNAs normalized to that of GAPDH.The qPCR data were collected with Bio-Rad CFX96 (Version 3.1) and analyzed with Bio-Rad CFX Manager (Version 3.1).Gene-specific primer sequences are listed in Supplementary information, Table S5.

ChIP
ChIP was performed according to the manufacture's instruction.Ten million cells were fixed with 1% formaldehyde for 10 min, quenched with 0.125 M glycine for 5 min at 37 °C and lysed in SDS Lysis Buffer.Cell lysate was sonicated by Bioruptor Pico Sonifier to shear chromatin DNA to a size range of 200-1000 bp.The supernatant was diluted 10-fold in ChIP Dilution Buffer and precleared with 60 mL agarose beads for 30 min.The supernatant fraction was immunoprecipitated with the indicated antibodies (2 μg) against BATF overnight at 4 °C.The antibody-chromatin complexes were pulled down with protein A agarose/salmon sperm DNA beads (Sigma, Catalog #16-157) for 1 h at 4 °C.The de-crosslinked DNA was subjected to qPCR analysis using specific primers listed in Supplementary information, Table S6.

Statistical analysis
Data were analyzed using Student's t-test, multiple t-test or twoway ANOVA with GraphPad Prism 8.The correlation study was analyzed using a Spearman rank correlation test.The number of asterisks represents the degree of significance with respect to P values, with the latter presented within each figure or figure legend.All the biochemical experiments, particularly immunoblotting analysis, were repeated for at least two times with similar results.The reproducibility of other experiments is described in the respective figure legends.

Fig. 1
Fig. 1 PD-1 signaling is negatively correlated with γ c level. a γ c level is negatively correlated with PD-L1 level in human NSCLC tumor biopsies.Representative images from IHC staining of γ c and PD-L1 in human NSCLC tumor biopsies are shown (left panels).Quantification of PD-L1 and γ c staining intensities was performed by semi-quantitative scoring (right panel).n = 127 independent samples, R = -0.2966.Correlation coefficients were calculated using the Pearson test.Two-sided P-value was given.Scale bars, 100 μm.b PD-1 blockade increases γ c level in tumor tissues.B16F10 cells (5 × 10 5 ) were subcutaneously injected into C57BL/6J mice.CT26 cells (5 × 10 5 ) were subcutaneously injected into Balb/c mice.Mice were intraperitoneally injected with control or anti-PD-1 antibody (100 μg per mouse) every three days (four times in total) five days after inoculation of cells.After 17 days, tumor-bearing mice were euthanized and tumor tissues were analyzed.Representative images from IHC staining of γ c in tumor sections are shown.Scale bars, 100 μm.c PD-1 blockade increases γ c level in tumorinfiltrating CD8 + T cells.TILs were isolated from B16F10 tumor tissues (left panels) or CT26 tumor tissues (right panels), stained with the indicated antibodies and analyzed by flow cytometry.Graph shows mean ± SEM, n = 8 independent samples.Data were analyzed using Student's unpaired t-test with GraphPad Prism 8. MFI, median fluorescence intensities.d PD-1 ligation down-regulates γ c level.Human CD8 +T cells or PD-1-expressing Jurkat (Jurkat-PD-1) cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days.The cells were stained with the indicated antibodies and analyzed by flow cytometry.Graph shows mean ± SEM, n = 3 technical repeats.Data were analyzed using two-way ANOVA with GraphPad Prism 8. e The level of γ c is down-regulated after PD-1 ligation.Human CD8 + T cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before immunoblotting analysis with the indicated antibodies.f PD-1 ligation down-regulates γ c level.Left panels: Jurkat cells transduced with empty vector (Jurkat-EV) or Jurkat-PD-1 cells were analyzed by immunoblots with the indicated antibodies.Middle panels: Jurkat-EV and Jurkat-PD-1 cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) as indicated for 3 days before immunoblotting analysis with the indicated antibodies.Right panels: Jurkat-PD-1 cells were stimulated with PHA (50 ng/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for the indicated times before immunoblotting analysis with the indicated antibodies.g Effects of PD-1 ligation on transcription of γ c .Human CD8 + T cells or Jurkat-PD-1 cells were stimulated with plate-bound anti-CD3 (1 μg/mL) and solubilized anti-CD28 (5 μg/mL) in the presence of plate-bound hPD-L1-Fc fusion protein or control hIgG1 (2 μg/mL) for 3 days before qPCR analysis of mRNA levels of the indicated genes.Graph shows mean ± SEM, n = 3 independent samples from one representative experiment.Data were analyzed using two-way ANOVA with GraphPad Prism 8.The experiments in c-g were repeated for at least two times with similar results.

Fig. 7
Fig.7Pitavastatin calcium potentiates anti-tumor immunity triggered by combination therapy of IL-2 and PD-1 blockade.a PC treatment up-regulates the level of γ c .Human CD8 + T, mouse CD8 + T, Jurkat, HBP-ALL or CTLL2 cells were treated with PC (0, 0.5, 1, 2 μM) for 24 h before immunoblotting analysis with the indicated antibodies.The γ c band intensities relative to the corresponding β-actin bands were shown in the histograph.b MARCH5-deficiency impairs PC-induced up-regulation of γ c .Control or MARCH5-deficient Jurkat cells were treated with PC (0, 0.5, 1 μM) for 24 h before immunoblotting analysis with the indicated antibodies.The immunoblots were repeated for two times with similar results.c PC treatment suppresses tumor growth.C57BL/6J mice were subcutaneously injected with 5 × 10 5 of MC38 or B16F10 cells.On day 3 (MC38) or 5 (B16F10) after tumor cell implantation, mice were intraperitoneally injected with control or PC (5 mg/kg/day).Tumor sizes were measured every two days by caliper.Tumor-bearing mice were euthanized on day 13 (MC38) or day 15 (B16F10).Tumor weights were measured by Analytical Balance.Graph shows mean ± SEM, n = 6.Data were analyzed using Student's unpaired t-test with GraphPad Prism 8. d Effects of PC on the level of γ c in TILs.TILs were isolated from the MC38 tumor tissues in c.TILs were stained with the indicated antibodies and analyzed by flow cytometry.Graph shows mean ± SEM, n = 6 independent samples.Data were analyzed using a Student's unpaired t-test with GraphPad Prism 8. MFI, median fluorescence intensities.e PC treatment increases tumor infiltrating CD8 + cytotoxic T cells.TILs were isolated from the MC38 tumor tissues in c.TILs were stained with the indicated antibodies and analyzed by flow cytometry.Graph shows mean ± SEM, n = 6 independent samples.Data were analyzed using Student's unpaired t-test with GraphPad Prism 8. f PC potentiates the antitumor efficacy of IL-2 and PD-1 combination.C57BL/6J mice were subcutaneously injected with MC38 cells (5 × 10 5 ).On day 5 after tumor cell implantation, mice were intraperitoneally injected with control, PC (5 mg/kg), IL-2 (50,000 IU per mouse) or anti-PD-1 (100 μg per mouse) (Supplementary information, Fig.S10b).Tumor sizes were measured every two days by caliper from day 5. Graph shows mean ± SEM, n = 8.Data were analyzed using two-way ANOVA with GraphPad Prism 8. g PC promotes the survival rate of mice treated with IL-2 and PD-1 blockade.C57BL/6J mice were subcutaneously injected with MC38 cells (5 × 10 5 ).On day 5 after tumor cell implantation, mice were intraperitoneally injected with control, PC (5 mg/kg), IL-2 (50,000 IU per mouse) or anti-PD-1 (100 μg per mouse) (Supplementary information, Fig.S10b).Mice were sacrificed when the tumor size is bigger than 15 mm of the mean tumor diameter, tumor volume exceeded 2000 mm 3 , or tumor had ulcers with diameter reached 10 mm.Statistical analysis was performed using the GraphPad Prism 8 software, n = 8.Kaplan-Meier survival curves and corresponding log-rank (Mantel-Cox) tests were used to evaluate the statistical differences between groups in survival studies.There is a significant difference when the P < 0.05.

Fig. 8 A
Fig.8A model on regulation of γ c stability and activity by PD-1 signaling.In tumor microenvironment, PD-L1/PD-1 signaling results in inhibition of the γ c family cytokine-triggered signaling and immune activation by two mechanisms.Immediately after PD-1 ligation, SHP2 is recruited to PD-1 and activated, which in turn dephosphorylates γ c at Y357, leading to its inactivation and unresponsiveness to γ c family cytokines.More later after PD-1 ligation, the transcription factor BATF is induced, which up-regulates the expression of the membrane-associated E3 ubiquitin ligase MARCH5.MARCH5 is recruited to γ c and mediates its K27-linked polyubiquitination at K315 and lysosomal degradation.Targeting of components involved in these regulatory mechanisms, such as by a combination of PD-1 blocking antibody (1), IL-2 (2) and MARCH5 inhibitor (3) leads to potent anti-tumor immunity.