1α,25(OH)2D3 reverses exhaustion and enhances antitumor immunity of human cytotoxic T cells

Background Epidemiological surveys have revealed that low serum vitamin D level was correlated with increased risk of tumors. Dysfunctional T cells in patients with tumor are characterized as exhausted with high levels of immune checkpoint receptors (ICRs). However, whether the reduced level of vitamin D in patients with cancer correlates with cytotoxic T-cell exhaustion is unknown. Methods Periphery blood samples from 172 patients with non-small cell lung cancer (NSCLC) were prospectively collected. Patients with NSCLC received one course of intravenous docetaxel (75 mg/m2) followed by treatment with or without rocaltrol at a dose of 0.5–2.0 µg/day for total of 3 weeks. We performed phenotypical and functional analysis of T-cell through flow cytometry. Vitamin D receptor (VDR) knockout and overexpression CD8+ and Vδ2+ T cells were constructed using Cas9-gRNA targeted and overexpressing approaches to identify 1α,25(OH)2D3/VDR-mediated transcription regulation for ICRs or antitumor activity in T cells. Results We show that serum level of vitamin D is negatively correlated with expression of programmed cell death-1 (PD-1), T-cell immunoreceptor with Ig and ITIM domains (TIGIT), and T-cell immunoglobulin and mucin-domain containing-3 (Tim-3), but positively correlated with CD28 expression on CD8+ and Vγ9Vδ2+ T cells in patients with NSCLC. 1α,25(OH)2D3, the active form of vitamin D, promotes the nuclear translocation of VDR, which binds to the promoter region of Pdcd1, Tim3, and Tigit genes and inhibits their expression. Besides, 1α,25(OH)2D3 pretreatment also promotes the methylation of CpG island in the promoter region of the Pdcd1 gene and increases H3K27 acetylation at the promoter region of the Cd28 gene, which leads to surface PD-1 downregulation and CD28 upregulation, respectively. We further reveal that VDR-mediated Ca2+ influx enhanced expression of Th1 cytokines via T-cell receptor activation. Functionally, 1α,25(OH)2D3 pretreated CD8+ T cells or Vγ9Vδ2+ T cells showed increased Th1 cytokine production and enhanced antitumor immunity. Finally, oral 1α,25(OH)2D3 could also decrease expression of PD-1, Tim-3, TIGIT and increase expression of CD28, resulting in cytokine production (associated with antitumor immunity) by cytotoxic T cells of patients with NSCLC. Conclusions Our findings uncover the pleiotropic effects of 1α,25(OH)2D3 in rescuing the exhausted phenotype of human cytotoxic T cells in patients with tumor and in promoting their antitumor immunity. Trial registration number ChiCTR2100051135.


INTRODUCTION
Cancer is a global threat to human health. There is an estimated 19.3 million new cancer cases and 10 million cancer deaths in 2020. 1 Tumor tissues are infiltrated with multiple types of immune cells which contain cytotoxic T cells such as γδ T cells 2 and CD8 T cells 3 and provide one of the major protections in antitumor immunity. γδ T cells recognize danger signals associated with tumorigenesis and are major players in tumor immune surveillance. 2 4 5 Besides, γδ T cells provide the early source of interferon gamma (IFN-γ) in tumor immunity and have potent cytotoxic property. 6 These cells recognize antigen in a major histocompatibility complex (MHC)independent manner, 7 and clinical trials revealed promising effects in autogenic and allogeneic γδ T-cell therapy. [8][9][10] CD8 T cells are antigen-specific tumor killers and are the most important cytotoxic T cells in many cancers. 3 Currently, most tumor immune therapy strategies, such as chimeric antigen receptor T-cell therapy (CAR-T), are based mainly on αβ T cells (CD4 and CD8) and have achieved exciting success in clinic. 3 11 However, tumor cells adopt many approaches to inhibit antitumor immunity mediated by cytotoxic T cells. [12][13][14] Cytokines, such as IFN-γ, stimulate expression of programmed cell death 1-ligand 1 (PD-L1) on tumor cells, which binds to programmed cell death-1 (PD-1) on cytotoxic cells and reduce their effector function. 15 The binding of T-cell immunoglobulin and mucindomain containing-3 (Tim-3) with galectin-9 (Gal-9) or carcinoembryonic antigen cell adhesion molecule 1 (CEACAM-1) in tumor microenvironment negatively regulates antitumor response. 16 17 Tumor or dendritic cells express poliovirus receptor, which binds to T-cell immunoreceptor with Ig and ITIM domains (TIGIT) on CD8 T cells and suppresses their tumor killing ability. 18 Besides, the B7/CD28 costimulatory Open access signaling pathway is also a necessity for antitumor immune response, but tumor burden reduces CD28 expression on cytotoxic T cells. 19 CTLA4/LAG3/FAS-mediated inhibitory signals also hinder anti-tumor responses. 12 13 These inhibitory effects converged to induce exhaustion of cytotoxic T cells and dampen antitumor immunity. Animal models and clinical studies have shown that Tim-3, TIGIT, LAG-3, PD-1, and PD-L1 targeted therapies have promising effects in enhancing antitumor immunity in a wide variety of tumor types.
Vitamin D is most widely known for its beneficial effects on calcium homeostasis and bone mineralization. 20 After being taken up from diet or synthesized by the skin, vitamin D is hydroxylated at C25 by CYP2R1 in the liver to 25(OH) D 3 ; the latter is further hydroxylated at C1 by CYP27B1 into 1α,25(OH) 2 D 3 (calcitriol), the bioactive form. 21 1α,25(OH) 2 D 3 binds to vitamin D receptor (VDR), which functions as a transcription factor to regulate various biological processes. 22 Recent studies revealed that vitamin D also has important regulatory functions on the immune system: vitamin D not only enhances pathogen clearance mediated by innate cells but also exerts immunosuppressive effects to prevent the detrimental responses of prolonged inflammation. 23 Epidemiological surveys have revealed an association between low serum 25(OH)D 3 levels and increased risk of prostate, lung, coloretcal, breast, and hepatic cancers. [24][25][26] The antiproliferative effects of vitamin D on tumor cells were also demonstrated, 27 but whether vitamin D regulates antitumor immunity mediated by cytotoxic T cells is barely known.
Here, we showed that the expressions of PD-1, Tim-3 and TIGIT on Vγ9Vδ2 + T cells and CD8 + T cells are regulated by the 1α,25(OH) 2 D 3 /VDR signaling pathway, which contributes to T-cell cytotoxicity under the condition of tumorantigen and T-cell receptor (TCR) activation. 1α,25(OH) 2 D 3 also modulates the epigenetic programs of Pdcd1 and Cd28 genes, which results in subsequent reduced expression of PD-1 but upregulation of CD28. Besides, we preliminarily reveal that the production of Th1 cytokines in T cells is dependent on VDR and Ca 2+ influx. Mouse models reveal that Vγ9Vδ2 + T cells pretreated with 1α,25(OH) 2 D 3 exert impressive antitumor activity in vivo. In a clinical trial, patients with cancer with NSCLC who received oral calcitriol (1α,25(OH) 2 D 3 ) showed decreased cell-surface level of PD-1, Tim-3 and TIGIT and increased CD28 expression on Vγ9Vδ2 + and CD8 + T cells. Peripheral blood T lymphocytes isolated from 1α,25(OH) 2 D 3 -treated patients with NSCLC showed increased expression of cytokines associated with antitumor immunity. These results converged to strongly support a promotive effect of 1α,25(OH) 2 D 3 /VDR on T-cell cytotoxicity and reveal potential targets for immunotherapy.

MATERIALS AND METHODS Mice
Wild-type (WT) mice and NOD.CB17-Prkdc scid /NcrCrl (NOD-SCID) mice were purchased from Beijing Vital River Laboratory Animal Technology Co. All animals were maintained in a specific pathogen-free facility for use according to the guidelines for experimental animals at Jinan University (Guangzhou, China). Mice were used between 5 weeks and 8 weeks of age.
Expansion of Vγ9Vδ2 T cells Blood from healthy male or female donors was obtained from the Guangzhou Blood Center. Human peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque-based (GE Healthcare, 17-1140-02) density gradient centrifugation. Zoledronate (ZOL) (Sigma, SML0223) expanded Vγ9Vδ2 T (Vδ2 T) cells were generated as described following. 28 Briefly, human PBMCs (2-3×10 6 /mL) were cultured in RPMI 1640 (Gibco, 11875093) medium supplemented with 10% Fetal Bovine Serum (FBS, Gibco, 16140071). Recombinant human interleukin (IL)-2 (Peprotech, 200-02-50) 400 U/mL and ZOL (50 µM) was added at day 0. Recombinant human IL-2 was added to a final concentration of 100 U/mL once every 2 days from day 3. T cells were maintained at a cell density of 1-2×10 6 /mL until the percentage of Vγ9Vδ2/ CD3 T is >90%. The purity of Vγ9Vδ2 T cells was determined by flow cytometry (BD VERSE). After 10-12 days of culture, cells were used in experiments. In some experiments, the Vγ9Vδ2 T cells were further purified by negative selection with EasySep Human Gamma/Delta T Cell Isolation Kit (STEM CELL, 19255). Unless mentioned otherwise, Vγ9Vδ2 T cells used in all the experiments were pretreated with 1α,25(OH) 2 D 3 (50 nM) or vehicle two times at 1-day intervals.
The clinical trial was registered in Chinese Clinical Trial Registry.

Human subjects
A total of 172 patients with NSCLC were recruited from the Department of Oncology, First Affiliated Hospital of Jinan University. A total of 47 healthy donors were recruited from the staff at the Jinan University. Information of patients with cancer and healthy donors is described in online supplemental tables 1 and 2. PBMCs were isolated by Ficoll-Paque-based density gradient centrifugation or stored at −80℃. Cell-surface levels of exhaustion markers PD-1, Tim-3, TIGIT, and costimulatory molecule CD28 on CD4 + , CD8 + , and Vγ9Vδ2 + T cells were determined by flow cytometry.

In vitro treatment of T cells
To measure cell-surface levels of PD-1, Tim-3, TIGIT, and CD28 in human CD8 + and Vγ9Vδ2 + T cells, PBMCs were isolated from patients with NSCLC by gradient centrifugation. PBMCs were then stimulated with 1α,25(OH) 2 D 3 (50 nM) three times at 2-day intervals in vitro. In some experiments, PBMCs were isolated from healthy donors by gradient centrifugation. Purified CD8 + and Vγ9Vδ2 + T cells were treated with 1α,25(OH) 2 D 3 (50 nM) three times at 2-day intervals and then were harvested for further analysis.

Study design
After obtaining informed consent, therapy cycles were scheduled according to the clinical trial standard, approved drug interval and health condition of each patient. For rocaltrol treatment, all patients with NSCLC received one course of intravenous docetaxel (75 mg/m 2 ) followed by rocaltrol at a dose of 0.5-2.0 µg/day for a total of 3 weeks. For the control group, patients with NSCLC received one course of intravenous docetaxel (75 mg/m 2 ) without rocaltrol for 3 weeks. Briefly, blood samples of patients were harvested at baseline (before therapy initiation). Patients with NSCLC were administrated docetaxel (75 mg/m 2 ) or docetaxel combined with rocaltrol (0.5-2.0 µg/day, Roche) for a total of 3 weeks (online supplemental table 2). Blood samples were collected from the patients after 3 weeks of therapy. Fresh trial serum and PBMCs were detected by ELISA and flow cytometry.

Animal model
In breast cancer model, NOD-SCID mice were inoculated subcutaneously with 5×10 6 MCF-7 cells. Ten days later, when MCF-7 tumors had reached a volume of 50 mm 3 , the mice were randomized into three groups and treated respectively with Vγ9Vδ2 T cells (10×10 6 ), 1α,25(OH) 2 D 3 pretreated Vγ9Vδ2 T cells (10×10 6 ), 100 µL Phosphate Buffer Solution (PBS) as control. All animals were transferred intravenously with Vγ9Vδ2 T cells or PBS once a week.
In a B16-F0 tumor model, age-matched and sex-matched B6 WT mice (age of 6-8 weeks) were inoculated subcutaneously with 4.0×10 5 B16-F0 cells. On day 0, the mice were randomized into two groups, and PBS (200 µL) or 1α,25(OH) 2 D 3 (0.03 mg/kg, 200 µL), were intravenously injected into the mice every 2 days, and eight times in total. Tumor size and mice survival were recorded every 2 days from day 6. Tumor volume (TV) was recorded and calculated using the following equation: where W=width, L=length. Mice bearing a tumor with size larger than 15 mm in any direction were euthanized.

Tumor-infiltrating lymphocyte (TIL) isolation
In order to analyze the cytokine production and expression of surface markers on tumor-infiltrating T cells (TILs), mice were euthanized on day 18. Tumor tissues were cut into pieces and suspended with 10 mL tumor digestion buffer (5% FBS, 1.5 mg/mL collagenase IV, and 10 µg/mL DNase I). After rotation for 1 hour at 37℃, tumor tissues were digested. The cell suspension was filtered with a 70 µm filter to harvest single-cell suspension. Leukocytes were isolated by density-gradient centrifugation using 40% and 70% Percoll (GE, 17089102). Then, tumor-infiltrating leukocytes were labeled with specific antibodies for different markers.

Flow cytometry
For cell-surface staining, approximately 10 5 -10 6 human Vγ9Vδ2 + T or CD8 + T cells were incubated with specific antibodies for 15  Subsequently, cells were cultured in normal medium for an additional 2 days. After puromycin selection two times, T cells were harvested and detected by western blot and quantitative real-time PCR. Furthermore, transduced T cells were restimulated with 1α,25(OH) 2 D 3 (50 nM) two times at 1-day intervals and determined by flow cytometry. For overexpression VDR, CD8 + T cells were transfected with PTSB-CMV-GFP-VDR (The VDR overexpression gene was inserted into a CMV expression vector with a GFP gene) and control vector PTSB-CMV-GFP (TransSheep Biotechnology), respectively, at 500×g for 90 min at 4℃. Cells were then restimulated with 1α,25(OH) 2 D 3 (50 nM) two times at 1-day intervals. The cellsurface levels of immune checkpoint receptors (ICRs) were detected by flow cytometry. The decline ratio of ICRs was calculated according to following formula: .
RNA extraction, cDNA synthesis, and quantitative real-time PCR Cells were harvested in 1 mL TRIzol (Invitrogen, Thermo Fisher Scientific), and total RNA was isolated with RNA simple Total RNA kit (Tiangen, DP419). RNA from each sample (200-500 ng) was reversely transcribed with the PrimeScript RT reagent kit in accordance with the manufacturer's instructions (TaKaRa, RRO37A). The cDNA was diluted at 1:5 or 1:10 in RNase/DNAse-free water for further analysis of quantitative real-time PCR (qPCR). Expression of target genes was determined by CFX Connect System (Bio-Rad) with 2×SYBR Green PCR Master Mix (Bimake, B21203) according to the manufacturer's instructions, and each qPCR reaction had a final volume of 20 µL. β-actin or GAPDH was used as the internal control. Each sample was examined at least in triplicate. PCR product specificity was confirmed by a melting-curve analysis. The relative mRNA expression was calculated using 2 −ΔΔCt method. Hot map was described as fold change according to the following formula: Primer pairs for quantitative real-time PCR were purchased from Sangon Biotech. The list of primer sequences is provided in online supplemental table S4.

Pyrosequencing
The PyroMark Assay Design V.2.0 software was used to design primers for the analysis of iNOS methylation.

Open access
Briefly, the following primer sequences were used: 5′-(Btn)-−3′ (5-biotin labeled) and 5′--3′, sequencing primer: bisulfite conversion was carried out using the EpiTect Plus DNA Bisulfite Kit (QIAGEN). We carried out the bisulfite conversion following the manufacturer's guidelines (QIAGEN). The PCR amplification was conducted in 50 µL reactions containing 2 µL of DNA (bisulfite conversion), 1 µL PCR primers, 2 µL dNTP mix, 10 µL 5×PCR GC buffer, and 0.2 µL 5 U Taq DNA polymerase (KAPA Biosystems). The PCR conditions were as following: 95°C for 3 min, 45 cycles of (94°C for 30 s, 50°C for 30 s and 72°C for 1 min) followed by 7 min hold at 72°C and then 4°C. The PCR products were then captured using streptavidin coated agarose beads under denaturing conditions to obtain single-stranded DNA. The pyrosequencing reaction was then carried out using the PyroMark Q96 machine (QIAGEN) and PyroMark Gold-Q96 Reagents kit (QIAGEN), using the following sequencing primer: 5′--3. The methylation status was analyzed using TPyro Q-CpG software (QIAGEN). The frequency of DNA methylation at the CpG islands in Pdcd-1 promoter regions were detected by pyrosequencing (Beijing Genomics Institute BGI). The primer sequences are provided in online supplemental table 5.

ChIP-qPCR and ChIP-seq
Chromatin immunoprecipitations were performed according to the protocol of Simple ChIP Enzymatic Chromatin IP Kit (CST, 9003) as previous described. 29 For VDR binding detection, CD8 + T cells (5×10 6 -1×10 7 ) were treated with 1α,25(OH) 2 D 3 (50 nM) for 12 hours followed by chromatin immunoprecipitations. For acetylation detection, Vγ9Vδ2 + T cells (5×10 6 -1×10 7 ) were treated with 1α,25(OH) 2 D 3 (50 nM) or vehicle two times at 1-day intervals, followed by chromatin immunoprecipitations. Briefly, cells were fixed for 12 min at room temperature with 1% formaldehyde, followed by addition of glycine, and incubated for another 5 min at room temperature. Subsequently, cells were lysed; chromatin was harvested and fragmented (150-900 bp) using enzymatic digestion. The chromatin was then subjected to immunoprecipitation with specific antibodies. Rabbit polyclonal antibodies to anti-VDR (CST, 12550), anti-H3K27ac, anti-H3 (CST, 4620), and normal rabbit IgG (CST, 8173) were used. Finally, the immune complexes were washed and eluted. Eluted DNA and 2% input DNA were incubated at 60℃ to reverse the cross-linking and then purified with spin columns. The relative abundance of precipitated DNA fragments was analyzed by qPCR using SYBR Green PCR Master Mix. Primers (forward and reverse) for amplification of the region upstream of Pdcd1, Cd28, Tigit, and Tim-3 are provided in online supplemental table 6. The results are showed as being relative to the total chromatin input or normalized to total histone H3 expressions to account for the nucleosomal occupancy at the promoter regions of genes. In addition, ani-H3K27ac enriched DNA fragments were used for further chromatin immunoprecipitation sequencing (Beijing Genomics Institute BGI).

Statistical analysis
Statistical analysis was performed using GraphPad Prism V.6.0 (GraphPad Software). Data are presented as the mean±SD. Statistical significance was determined as indicated in the figure legends. Significance was set to p<0.05 and represented as *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, not significant. For sample sizes of n=3, a two-tailed unpaired Student's t-test was used when the variance was similar between two groups, and a two-tailed unpaired Student's t-test with Welch's correction was used when variances were different. For sample sizes of n>3, the data distribution was first checked using a Kolmogorov-Smirnov test (GraphPad Software). If the data fitted a normal distribution, a two-tailed unpaired Student's t-test was used when variances were similar, whereas a twotailed unpaired Student's t-test with Welch's correction was used when variances were different. If the data did not fit a normal distribution, a Mann-Whitney U test was used. A two-tailed paired Student's t-test was used when the difference value was Gaussian distributed, and a twotailed paired Student's t-test with Wilcoxon correction was used when data were skew distribution. Correlation analysis was performed using Pearson's test. Animals were randomly allocated to the treatment groups.

RESULTS
1α,25(OH) 2 D 3 decreases expression of multiple ICRs but increases CD28 expression Low-serum vitamin D level was correlated with increased risk of tumors, [24][25][26] and supplementation of vitamin D reduced the risk of cancer death. 30 Tumor burden also induces exhaustion of cytotoxic T cells and hinders antitumor immunity. 13 However, whether reduced vitamin D participates in the etiology of exhaustion of cytotoxic T cells is unknown. Thus, we recruited 53 patients with advanced lung cancer (non-small cell lung cancer (NSCLC)), and 47 healthy volunteers, serum vitamin D (25(OH)D 3 , 1α,25(OH) 2 D 3 ) levels, and baseline expression of ICRs on CD8 + T and Vγ9Vδ2 + T cells from PBMCs were analyzed. A summary of the characteristics of both healthy volunteers and patients with cancer with NSCLC is described in tables 1 and 2. Consistent with previous reports, patients with cancer showed immune exhaustion phenotype as indicated by increased cell-surface expression of PD-1, TIGIT, and Tim-3 but reduced CD28 expression on CD8 + T cells and Vγ9Vδ2 + T cells, while cell surface CTLA-4 was barely detectable (online supplemental figure S1A,B). The serum levels of vitamin 25(OH)D 3 and the bioactive form 1α,25(OH) 2 D 3 in patients with cancer were also significantly lower than those of healthy controls (online supplemental figure  S2A,B), which was in accordance with previous studies. Importantly, we found that the serum levels of 25(OH) D 3 and 1α,25(OH) 2 D 3 in patients with NSCLC were negatively correlated with PD-1, TIGIT, and Tim-3, but positively correlated with the CD28 level on cytotoxic T cells ( figure 1A,B, and online supplemental figure S3A,B). To test whether vitamin D could directly regulate the expression of ICRs on cytotoxic T cells, we isolated PBMC from patients with NSCLC and cultured in the presence or absence of 1α,25(OH) 2 D 3 for 6 days (online supplemental figure S4A). Notably, 1α,25(OH) 2 D 3 treatment significantly reduced PD-1, TIGIT, and Tim-3 but increased CD28 expression on CD8 + T and Vγ9Vδ2 + T cells (online supplemental figure S4B-D). Besides, 1α,25(OH) 2 D 3 also significantly reduced expression of PD-1, TIGIT and Tim-3 (Co-expression of PD-1 with Tim-3 or TIGIT) and increased CD28 expression on T cells (CD8 + , CD4 + , and Vγ9Vδ2 + T cells) from healthy donors during de novo activation (online supplemental figure S5A-E). Furthermore, Vγ9Vδ2 + T cells pretreated with 1α,25(OH) 2 D 3 showed higher levels of Ki-67 (online supplemental figure S5F). These results strongly indicated that vitamin 1α,25(OH) 2 D 3 could directly regulate the expression of ICRs on T cells. We next checked whether vitamin D supplementation could reverse the exhaustion phenotype of cytotoxic T cells in patients with NSCLC. In order to do so, all patients with NSCLC were administered with one course of intravenous docetaxel (75 mg/m 2 ) followed therapy with rocaltrol at a dose of 0.5-2.0 µg/day for total of 3 weeks ( figure 1C). Treatment cycles were scheduled according to the clinical trial standard, approved drug interval and health condition of each patient. Patient information is provided in online supplemental table 2 (for cell-surface markers detection). After therapy initiation, patients with NSCLC showed elevated 1α,25(OH) 2 D 3 levels in serum at the indicated time points ( figure 1D). Delightedly, oral intake of calcitriol reduced PD-1, TIGIT and Tim-3 levels and increased CD28 expression on CD8 + and Vγ9Vδ2 + T cells compared with a single dose of Open access

Open access
docetaxel treatment in patients with NSCLC (figure 1E-P and online supplemental figure S6A-D). Collectively, the aforementioned findings clearly revealed the therapeutic potential of 1α,25(OH) 2

D 3 in overcoming cytotoxic T-cell exhaustion and indicated that the decreased vitamin D level in patients with cancer with NSCLC was the etiology of cytotoxic T-cell exhaustion.
Cytokine secretion of T cells in patients with NSCLC treated with rocaltrol was promoted To further explore the role of 1α,25(OH) 2 D 3 in patients with NSCLC, we analyzed samples from patients with advanced lung cancer receiving rocaltrol (1α,25(OH) 2 D 3 ) therapies. In clinical trials of oral rocaltrol, patients with high levels of 1α,25(OH) 2 D 3 may not benefit from our therapy project, so we excluded them from the clinical trial. Baseline levels of 1α,25(OH) 2 D 3 in enrolled patients treated with docetaxel and rocaltrol or docetaxel alone were less than 20 pg/mL. The difference of baseline levels of 1α,25(OH) 2 D 3 between groups of docetaxel or docetaxel combined with rocaltrol was not statistically significant (figure 2A). Patients with NSCLC had received docetaxel or docetaxel combined with rocaltrol treatment at the indicated time points (figure 2B and online supplemental figure S7A). A summary of patients' information is described in online supplemental table 2 (for cytokine detection). CT scans were performed at the baseline and after treatment with docetaxel and rocaltrol. Changes of tumor size based on the CT imaging after 3 weeks were recorded. Obvious difference of tumor size shrinkage was observed under the treatment of docetaxel combined with rocaltrol (figure 2C-E). Next, we analyzed cytokine secretion of peripheral blood CD4 + , CD8 + , and Vγ9Vδ2 + T cells during docetaxel or docetaxel combined with rocaltrol therapy in patients with NSCLC. Administration with docetaxel and rocaltrol increased the percentage of IFN-γ + CD4 + , TNF-α + CD8 + , and TNF-α + Vγ9Vδ2 + T cells and the level of TNF-α in serum compared with that of administration with a single dose of docetaxel (figure 2F-I and online supplemental figure S7B-D) in patients with NSCLC. To assess the percentage of no recurrence patients responding to docetaxel and rocaltrol therapy, we focused our analysis on patients who had at least a 1.2-fold increase in the frequency of IFN-γ + and TNF-α + CD3 T cells. In accordance with our predictions, most of the patients with NSCLC with IFN-γ + and TNF-α + CD3 T cells activated by docetaxel and rocaltrol did not relapse; 1α,25(OH) 2 D 3 level was marginally increased, but 1α,25(OH) 2 D 3 level negatively correlated with PD-1 expression on CD8 + T cells (online supplemental figures S7E-G). These data suggested that 1α,25(OH) 2 D 3 treatment might also be important for activation of T cells in patients with cancer with NSCLC.
VDR is a transcription factor for Pdcd1, Tim3 and Tigit genes The canonical action pathway of vitamin 1α,25(OH) 2 D 3 is through binding to its receptor (VDR), which translocates into nucleus and functions as a transcription factor. 21 We used distinct active forms of vitamin VD 3 , 25(OH)D 3 , and 1α,25(OH) 2 D 3 (low, moderate, and high activities) to stimulate T cells, respectively. The results showed that ICR levels were declined maximum by 1α,25(OH) 2 D 3 , which indicated that the expression of ICRs on T cells was regulated mainly by 1α,25(OH) 2 D 3 ( figure 3A-D and online supplemental figures S8A,B). Indeed, 1α,25(OH) 2 D 3 treatment increased VDR expression and promoted its nuclear translocation in cytotoxic T cells ( figure 3E,F). Thus, we next investigated whether VDR was involved in regulating the expression of ICRs. VDR-KO CD8 + T cells were constructed using CRIS-PR-Cas9 targeted approach and the results demonstrated that knockout of VDR remarkably potentiated the expression of PD-1, TIGIT, and Tim-3, whereas CD28 level was marginally reduced (figure 3G and online supplemental figure S9A-C). Consistently, VDR overexpression in CD8 + T cells were constructed using PTSB-CMV-VDR approach, and the results showed that the expression of PD-1, TIGIT, and Tim-3 were inhibited under stronger 1α,25(OH) 2 D 3 /VDR signaling, whereas CD28 expression was either unchanged or only marginally increased (figure 3H,I, and online supplemental figure S9D). Thus, we hypothesized that VDR might be a transcription factor that inhibits the transcription of Pdcd1, Tigit, and Tim3 genes. Besides, we used the JASPAR website to predict VDR binding sites in the promoter region of Pdcd1, Tigit, Tim3, and Cd28 genes (online supplemental figure S9E), and ChIP-qPCR analysis showed that 1α,25(OH) 2 D 3 administration notably promoted the binding of VDR to these three genes ( figure 3J,K). Collectively, the aforementioned results demonstrated that VDR is an inhibitory transcription factor for Pdcd1, Tigit, and Tim3 in cytotoxic T cells.

Open access
Open access was largely unaffected by either 5-Aza-dC or TSA (online supplemental figure S10E,F). These results suggested that the inhibition of PD-1 by 1α,25(OH) 2 D 3 was dependent on DNA methylation, and promotion of CD28 was dependent on histone acetylation. Indeed, 1α,25(OH) 2 D 3 treatment promoted the methylation of CpG sites of the Pdcd1 gene ( figure 4E,F). Moreover, ChIP-qPCR analysis further confirmed the upregulation of H3K27 acetylation in the Cd28 gene but not in Pdcd1, Tim3, or Tigit ( figure 4G-I). Taken together, these results corroborated that 1α,25(OH) 2 D 3 /VDR signaling adopted epigenetic approaches to decrease the expression of inhibitory checkpoint receptors and to promote costimulatory CD28 expression.
VDR mediates Ca 2+ influx to promote cytokine production of activated T cells From the aforementioned data, we demonstrated that 1α,25(OH) 2 D 3 could downregulate expression of inhibitory checkpoint receptors on T cells, and release T-cell inhibition, so we wondered if 1α,25(OH) 2 D 3 promote their cytokine production. To investigate whether 1α,25(OH) 2 D 3 promoted cytokine production of T cells in response to TCR stimulation in vitro, we cultured Vγ9Vδ2 + T lymphocytes and CD8 + T cells from PBMCs of healthy donors in the presence or absence of 1α,25(OH) 2 D 3 treatment. The expression of cytokine-related genes (figure 5A) in Vγ9Vδ2 + and CD8 + T cells was detected with quantitative real-time PCR, and the intracellular level of cytokines was determined by flow cytometry. The results showed that 1α,25(OH) 2 D 3 treatment enhanced Th1 cytokine production of Vγ9Vδ2 + (figure 5B,C) and CD8 + (figure 5D,E) T cells in response to TCR activation (figure 5F and online supplemental figure S11A) and also promoted CD107a degranulation (online supplemental figure S11B) but had no obvious effects on expression of Fas, perforin, and NKG2D (online supplemental figures S11C-E). Interestingly, production of IFN-γ and TNF-α in response to 1α,25(OH) 2 D 3 treatment was insignificantly altered under PMA and ionomycin stimulation (online supplemental figures S12A,B), which suggested that the enhanced Th1 cytokine production in 1α,25(OH) 2 D 3 pretreated T cells was associated with TCR recognition. To determine whether Ca 2+ ions were involved in regulating IFN-γ and TNF-α production in 1α,25(OH) 2 D 3 pretreated Vγ9Vδ2 T cells, the intracellular Ca 2+ ions were chelated with BAPTA-AM, and the results showed a significant decrease in the percentage of IFN-γ and TNF-α production (online supplemental figure S13). Besides, we found that stimulation of anti-CD3 and anti-CD28 antibodies was incapable of inducing intracellular Ca 2+ ions release in activated T cells ( figure 5G). However, supplementation of Vγ9Vδ2 + T cells with 2 mM Ca 2+ ions resulted in sustainable Ca 2+ ions influx in 1α,25(OH) 2 D 3 pretreated T cells under the brief stimulation of the TCR (figure 5H). Importantly, our results found that stimulation with TCR in vitro promoted VDR migration from cytoplasm and nucleus to the membrane of T cells ( figure 5I). Therefore, we hypothesized that VDR might need to migrate to the cell surface in order to efficiently mediate Ca 2+ ion influx in cytotoxic T cells. We next checked whether VDR was involved in regulating Ca 2+ ion influx-mediated cytokine production in cytotoxic T cells. Hence, we adopted CRISPR-Cas9 technology to knockout VDR in Vγ9Vδ2 + T cells and found that both of VDR and Ca 2+ ions (figure 5J) were necessary for production of IFN-γ and TNF-α (figure 5K,L), and Ca 2+ ions influx is VDR dependent. These data demonstrated that VDR mediates Ca 2+ ion influx to enhance cytokine production in activated T cells.
1α,25(OH) 2 D 3 promotes antitumor immunity of cytotoxic T cells Recent clinical studies also underscore the safety and efficacy of vitamin C mediated allogeneic Vγ9Vδ2 T-cell immunotherapy in patient with late-stage lung or liver cancer. 10 Since 1α,25(OH) 2 D 3 could reverse the exhaustion phenotype of cytotoxic T cells, we assumed that it might also regulate antitumor immunity mediated by cytotoxic T cells. To test this assumption, 1α,25(OH) 2 D 3 treated Vγ9Vδ2 + T cells were cocultured with multiple lines of CFSE-labeled tumor cells for 6 hours and the apoptosis of tumor cells were analyzed. Vγ9Vδ2 + T cells recognize and lead to tumor cell lysis directly ( figure 6A). As expected, 1α,25(OH) 2 D 3 treated Vγ9Vδ2 + T cells showed increased cytotoxicity (figure 6B and online supplemental figure S14A,B). To examine whether 1α,25(OH) 2 D 3 treatment would also be necessary for reinvigoration of antitumor responses of Vγ9Vδ2 + T cells, NOD-SCID mice were inoculated with MCF-7 (online supplemental figure S15A) and U2932 cells (online supplemental figure S15B). We tracked tumor growth by measuring TV and observed that 1α,25(OH) 2 D 3 pretreated Vγ9Vδ2 + T cells could slow tumor growth in MCF-7 tumor model ( figure 6C,D). Besides, it should be noted that antitumor immunity mediated by 1α,25(OH) 2 D 3 pretreated Vγ9Vδ2 + T cells had no side effects on organs, including the liver, kidneys, and lungs, displaying the safety of Vγ9Vδ2 T cells ( figure 6E). We also found that 1α,25(OH) 2 D 3 pretreated Vγ9Vδ2 + T cells showed enhanced antitumor immunity in adoptive therapy, and they even provided better protection than ibrutinib in treating diffuse large B-cell lymphoma (DLBCL) U2932 ( figure 6F,G). However, in tumor-bearing mice models, no obvious difference of tumor size shrinkage was observed under the therapy of 1α,25(OH) 2 D 3 pretreated Vγ9Vδ2 + T cells combined with αPD-L1 (figure 6H-K and online supplemental figure S15C,D). To further investigate the physiological function of 1α,25(OH) 2 D 3 (calcitriol), we next used mouse B16-F0 melanoma models to investigate the effect of calcitriol in antitumor immunity. The results showed that mice treated with calcitriol had slower tumor progression than PBS treated mice ( figure 6L-N). Moreover, administration with calcitriol led to more production of IFN-γ and TNF-α in tumor-infiltrating CD8 + and γδ + T cells ( figure 6O,P). Analysis of CD8 + and γδ + T cells isolated Open access Open access from tumor tissues showed that the expression of PD-1 was obviously reduced in mice treated with calcitriol (online supplemental figure S15E). These results indicated that 1α,25(OH) 2 D 3 treatment reversed exhaustion and enhanced T-cell cytotoxicity both in vitro and in vivo.

DISCUSSION
From the immunological point of view, tumor growth and progression are due to imperfect immune surveillance and failure of cancerous cell eradication. Tumor cells evade immune surveillance through different mechanisms, including activation of different immune checkpoint pathways that suppress antitumor immune responses. ICR-targeted therapies such as blockade of PD-1/PD-L1/ CTLA4 reinvigorate antitumor immune responses and promote immune-mediated elimination of tumor cells. 33 Some of these approaches have been approved for certain cancer treatments, and hundreds more are under clinical trials. However, accumulating evidence showed that only a fraction of patients with cancer benefit from ICR-targeted therapies. This is probably because cancer cells adopt multiple immune suppressing approaches and only a fraction of patients show increased expression of a certain ICR. Thus, combined ICR therapies were developed to increase blockade efficacy. Our study demonstrated that 1α,25(OH) 2 D 3 /VDR signaling suppresses the expression of multiple ICRs including PD-1, TIGIT, and Tim-3 in cytotoxic T cells, and enhanced their antitumor activity. 1α,25(OH) 2 D 3 treatment also reversed the decreased costimulatory CD28 expression on them. Besides, we demonstrated that the mechanism of 1α,25(OH) 2 D 3 / VDR in regulating antitumor immunity of T cells is associated with rescuing their exhausted phenotype and VDR-mediated Ca 2+ ions influx of human cytotoxic T cells. These results suggested that 1α,25(OH) 2 D 3 /VDRtargeted therapy might have broader scope of application.
Another obstruction that hinders the application of ICR-targeted therapy is severe immune-related adverse events, which arise in some patients with cancer. 34 These therapies probably cause imbalance of the normal physiological barriers against autoimmunity and lead to various local and systemic autoimmune responses. In contrast, vitamin D possesses immune-suppressing function in inflammatory responses elicited by innate immune cells or CD4 T cells. 23 Besides, vitamin D also favors the differentiation of Foxp3 + Treg cells, 35 which are potent suppressors in controlling autoimmune responses. These studies suggested that vitamin D treatment might be much safer with low risks of systemic inflammatory storm and autoimmunity. This needs further clinical validation in the future. In addition, vitamin D also directly suppresses the proliferation and induces apoptosis of a variety types of tumors. 27 Epidemiological surveys have revealed that low level of vitamin D is associated with high risks of tumor, and vitamin D supplementation reduces the risk of cancer death. 30 36 37 These observations potentiated the rationale for vitamin D therapy in cancer.
Interestingly, some studies showed the reduction of ICRs by 1α,25(OH) 2 D 3 only occurred on cytotoxic T cells but not on CD4 + T cells, suggesting different ways of action of vitamin D in these cell subtypes. The precise mechanisms underlying these different actions need further investigation. Previous studies also revealed the divergent role of vitamin D on PD-1 expression. Bendix et al found that long-term administration of vitamin D3 on patients with Crohn's disease promoted PD-1 expression on CD4 + CD25 int T cells, 38 and Sheikh et al showed that stimulation of CD4 + T cells with vitamin D3 in vitro enhanced the expression of PD-1. 39 In contrast, Pincikova et al reported that vitamin D treatment on patients with cystic fibrosis was associated with reduced PD-1 expression on both CD4 + and CD8 + T cells. 40 Our results showed that 1α,25(OH) 2 D 3 could reverse exhaustion of CD4 T cells derived from PBMCs of healthy donors in vitro. These findings suggest that the physical condition of patients may influence the action of vitamin D, and the effects of vitamin D in patients with different tumors need to be validated.
Vitamin 1α,25(OH) 2 D 3 binds to VDR, which triggers the nuclear translocation of VDR and regulates the transcription of multiple genes. 21 In this study, we report that VDR activation by 1α,25(OH) 2 D 3 induces transcriptional suppression of Pdcd1, Tim3, and Tigit genes in cytotoxic T cells. However, how VDR mediates the inhibition of transcription of these genes is unknown. Previous studies have revealed that blimp-1 41 and T-bet 42 are inhibitory transcription factors for PD-1. Does VDR activation promote recruitment of these factors to Pdcd1 gene locus? VDR frequently forms a heterodimer with retinoid X receptor for DNA binding. 43 Epigenetic modification such as DNA methylations and histone modifications were also involved in PD-1 regulation. 44 Xu et al reported that DNA modification 5-methylcytosine at CpG islands in gene promoters or at transcriptional enhancers was related with silence of gene expression. 45 Some studies demonstrated that two CpG-rich regions upstream of Pdcd1 transcription start site (TSS) were dynamically methylated in CD8 T cells during infection of acute lymphocytic choriomeningitis virus (LCMV) or chronic HIV. 46 47 On a genomic level, promoters were enriched histones such as H3 lysine 27 acetylation (H3K27ac) and were generally considered as promoted gene expression. 48 When PD-1 expression was enhanced on CD8 T cells in vitro, H3K27ac and H3K9ac were located at CR-C (this element, which are located 100 bp and 1.1 kb upstream of the TSS, contains multiple transcription factor binding sites). 41 In addition, the CR-C region of Pdcd1 was enriched with repressive histone modification H4K20me3, H3K27m3, and H3K9me3 to inhibit PD-1 expression. 49 1α,25(OH) 2 D 3 reduced the level of H3K27 acetylation at the promoter of Pdcd1, indicating that histone modification might be also involved in regulation of PD-1 expression. Our results also suggested that the inhibition of PD-1 expression by 1α,25(OH) 2 D 3 was dependent on DNA methylation, but the function relevance for DNA methyltransferase is unclear yet, and how methylation and acetylation coordinately regulate the expression of ICRs and costimulatory molecules deserved further study. VDR also recruits co-regulatory complexes for chromatin remodeling, histone modification and RNA polymerase II activation. 50 Whether these elements coordinate to mediate transcriptional suppression of Pdcd1, Tim3, and Tigit genes needs further investigation. Therefore, it is important to recognize that 1α,25(OH) 2 D 3 molecule has broad pleiotropic effects, and other pathways are probably implicated.
T cells are central players in immune responses against exogenous pathogens and endogenous cancers. T-cell activation requires two signals: major antigen-induced signal from the TCR and a secondary signal from costimulatory receptors. 51 A previous study has shown that the TCR, Ca 2+ , and CD28 together form a dual-positive-feedback circuit that substantially amplifies T-cell signaling and thus increases antigen sensitivity. 52 Using VDR-knockout Vγ9Vδ2 + T cells, we demonstrated that VDR deficiency in T cells resulted in decreased extracellular Ca 2+ ions influx and reduced IFN-γ and TNF-α production in the context of αCD3/CD28 stimulation. Our studies found that 1α,25(OH) 2 D 3 regulated the cytokine production of T cells through VDR and Ca 2+ influx. Interestingly, we discovered that VDR in T cells could migrate from cytoplasm and nucleus to cell membrane in response to αCD3/CD28 treatment. However, the mechanisms of how VDR migrated to the cell membrane and how they worked there remain to be investigated.
Currently, most immune cell therapy strategies, such as CAR-T, are mainly based on αβ T cells (CD4 and CD8 T) and have achieved ground-breaking success in CD19 + diseases, 11 53 but it is associated with unique acute toxicities, which can be severe or even fatal for patients, the application of such a strategy have shown limited success in treating solid tumors. 54 Notably, the published article has intensively reviewed the clinical potential of allogeneic Vγ9Vδ2 T cells as a new strategy for cancer immunotherapy, showing the safety of Vγ9Vδ2 T cells. 10 55 Previous reports have proven that T-cell exhaustion is a poor responsive status with an upregulated level of ICRs, decreased production of antitumor-related Th1 cytokines, and suppressed antitumor efficacy in patients with cancer. 33 56 Interestingly, recent studies also show that tumor mutation burden, microsatellite instability, and immune microenvironment (PD-1 and PD-L1 expressions) are prognostic factors associated with better patient survival in ICR inhibitor-targeted therapies. [56][57][58][59][60] Vitamin 1α,25(OH) 2 D 3 downregulated ICRs, increased expression of antitumor-related cytokines by activated T cells, and elevated antitumor activity in vitro and in vivo. We found that the proportion of no recurrence patients treated with the combination of docetaxel and rocaltrol was slightly higher than that of patients treated with docetaxel alone. Considering the number of enrolled patients and the lack of observation time for treatment effects, we expect more patients to be enrolled in clinical trials and observed for a longer time to fully evaluate the effect of this therapy. Our study therefore expands on the knowledge of 1α,25(OH) 2 D 3 biology and highlights the clinical potential of the regulation of antitumor immunity through the 1α,25(OH) 2 D 3 /VDR pathway.
In summary, our study demonstrated that 1α,25(OH) 2 D 3 reduced expression of multiple ICRs and increased expression of costimulatory molecule CD28 through modification of their gene promoter regions, enhanced production of Th1 cytokines through VDR-mediated Ca 2+ ion influx signaling pathway, and facilitated antitumor immunity by cytotoxic T cells. Primary clinical evidence also supported our findings. Taken together, 1α,25(OH) 2 D 3 and VDR-targeted therapy might be a safe and economical approach with pleiotropic effects and broad application scope to promote antitumor responses of patients with cancer. 1α,25(OH) 2 D 3 combined with some ICR inhibitors or drugs may have better efficacy.

Open access
Open access This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See https://creativecommons.org/ licenses/by/4.0/.
Author note Zhinan Yin, Ligong Lu and Yunfei Gao jointly supervised this work.