Vitamin D3 enhances the response to cisplatin in bladder cancer through VDR and TAp73 signaling crosstalk

Abstract Background Vitamin D3 (VitD) deficiency is linked to increased incidence and worse survival in bladder cancer (BCa). In addition to cystectomy, patients are treated with cisplatin‐based chemotherapy, however 30%‐50% of patients do not benefit from this treatment. The effects of VitD deficiency on response to chemotherapy remain unknown. Methods To test effects of VitD supplementation on the response to cisplatin we analyzed patient serum VitD levels and correlated that with survival. In vivo, VitD deficient mice were treated with cisplatin, with or without pretreatment with the active VitD metabolite, 1,25 dihydroxyvitamin D3 (1,25D3). Lastly, using BCa cell lines, T24 and RT‐112, the mechanism of action of 1,25D3 and cisplatin combination treatment was determined by apoptosis assays, as well as western blot and RT‐PCR. Results In this study, we determined that low serum 25 hydroxyvitamin D3 (25D3) levels was significantly associated with worse response to cisplatin. Pretreating deficient mice with 1,25D3, reduced tumor volume compared to cisplatin monotherapy. In vitro, 1,25D3 pretreatment increased the apoptotic response to cisplatin. 1,25D3 pretreatment increased expression of TAp73 and its pro‐apoptotic targets, in a VDR dependent manner. VDR and its transcriptional targets were induced after 1,25D3 treatment and further increased after the combination of 1,25D3 and cisplatin in a TAp73 dependent manner. Conclusions Our data suggest that VitD deficiency could be a biomarker for poor response to cisplatin, and pretreating with VitD can increase the apoptotic response to cisplatin through VDR and TAp73 signaling crosstalk.

chemotherapy are delaying cystectomy in patients with poor response to cisplatin (~30%-50%), and the lack of available biomarkers to identify patients unlikely to respond to cisplatin. 3 Two years after cystectomy 50% of patients relapse with metastatic disease, suggesting a high degree of initial micro-metastasis. 4 Metastatic MIBC has poor survival (5-year survival ~15%) and is treated by systemic chemotherapy. 2,5 Improving the effectiveness of cisplatin could improve overall survival.
p53 is mutated in ~50%-60% of MIBC cases. 14 In a p53-independent mechanism, p73 is activated after cisplatin-induced DNA damage. p73, a p53 family member rarely mutated in cancer, has both multiple isoforms. 15 The anti-apoptotic isoform, ΔNp73, is transcribed from an intrinsic promoter site and is transcriptionally inactive. 16,17 TAp73, the full length pro-apoptotic isoform, transcribes similar targets as p53 such as BAX and NOXA. 18,19 The ratio of TA/ΔN isoforms determines the response to chemotherapy. [20][21][22] Clinical response to cisplatin in BCa does not correlate with p53 mutation status. 14,23 In addition, cisplatin-resistant BCa cell lines lose the ability to induce p73. 24 These findings suggest an important role of p73 in the apoptotic response to cisplatin. 1,25D 3 increases the antitumor response to cisplatin in a variety of preclinical settings. 25,26 However, the importance of VitD sufficiency in the response to cisplatin has not been investigated in BCa. In the current study, we investigated how VitD deficiency altered response to cisplatin in vivo and in patients. We then investigated the mechanism of action of 1,25D 3 and cisplatin combination therapy. Our data provides rationale to determine the serum 25D 3 level in patients with MIBC prior to receiving neoadjuvant chemotherapy and consider VitD supplementation as a pretreatment strategy.

| Cell lines
Human BCa cell lines T24, RT-112, and 253J were purchases from ATCC (Manassas, VA). T24 cells were cultured in McCoys 5A media with l-glutamine supplemented with 10% FBS and 1% penicillin/streptomycin sulfate and used within 6 months of purchase. RT-112 and 253J cells were cultured in RPMI media with L-glutamine supplemented with 10% FBS and 1% penicillin/streptomycin sulfate. All cells were tested for mycoplasma within the past 4 years (PCR Mycoplasma Detection Kit, MD Bioproducts). Cell lines were not authenticated after purchase.

| Tissue microarray
The BCa tissue microarray at Roswell Park was stained with anti-VDR. Nuclear protein expression was digitally scanned using Aperio Scanscope (Aperio Technologies, Inc., Vista, CA) with 20× bright-field microscopy. VDR expression was summarized by patient characteristics using the mean and standard deviation, and the median and IQR (25th-75th percentiles). The association between patient characteristics and expression were evaluated using parametric t-tests or one-way ANOVA; and the non-parametric Mann-Whitney U or Kruskal-Wallis tests. The survival outcomes (overall and disease-specific) were summarized by expression level using standard Kaplan-Meier methods. Estimates of the median time and 3/5-year rates were obtained with 95% confidence intervals. Comparisons were made using the log-rank test.

| In vivo deficiency vs sufficiency
In vivo protocols were approved by the Institutional Animal Care and Use Committee at Roswell Park Cancer Institute and performed in compliance with institutional guidelines and regulations. Ten female nude mice, 4 weeks old, were placed on a VitD deficient diet (Research Diets, 25 IU) and ten female nude mice, 4 weeks old, were placed on a VitD sufficient diet (Research Diets, 1000 IU) for the course of the experiment. After 6 weeks on the diets, mice were inoculated with 3 × 10 6 T24 cells in Matrigel:HBSS (10:1) in the right flank. When tumors reached approximately 300 mm 3 , mice were randomized (n = 5 per group) and given IP injections of 5 mg kg −1 of c (100 μL volume) or saline once a week for 3 weeks. Tumor volume was measured with calipers (calculated by (length × width 2 )/2) and mouse weights were recorded every other day. Fractional tumor volume was calculated using the start of treatment as day 0. Serum 25D 3 levels were analyzed by Heartland Assays using a radioimmunoassay.

1,25D 3 and cisplatin
Forty female nude mice, 4 weeks old, were placed on the VitD deficient diet for 6 weeks. Mice were inoculated with T24 tumors as described. When tumors reached approximately 200 mm 3 , mice were randomized and treated with either saline, 1,25D 3, cDDP, or 1,25D 3 and cDDP (n = 10 per group). 0.625 μg of 1,25D 3 was given IP on Monday, Wednesday, and Friday, and 5 mg kg −1 of cDDP was given on Friday (100 μL volume). Serum and kidneys were harvested from 3 mice per treatment group 8 hours after the first week of treatment. Tumor volume was measured with calipers (calculated by (length × width 2 )/2) and weights were recorded. Fractional tumor volume was calculated using the start of treatment as day 0. An F-test about the treatment-time interaction was used to evaluate differences in tumor growth rates between treatment groups. Post-hoc comparisons were made using Holm-Bonferroni adjusted F-tests about the appropriate linear contrast of model estimates. Serum metabolites of VitD were analyzed by Heartland Assays using a radioiummunoassay.

| Patient serum analysis and survival
BCa patient serum samples were obtained from Roswell Park Cancer Institute Data Bank and Biorepository. Serum 25D 3 levels were analyzed by Heartland Assays using LC/MS/MS. Patients were grouped as low (25D 3 <20 ng mL −1 ) or high (25D 3 ≥20 ng mL −1 ). Survival outcomes were summarized using Kaplan-Meier methods: the 3/5 year and median survival were obtained with 95% confidence intervals, comparisons made using the log-rank test.

| MTT assay
Cells were pretreated with EtOH or increasing concentrations of 1,25D 3 (10 nmol L −1 -1 μmol L −1 ) for 24 hours and followed by increasing concentrations of cDDP (0.1 μg mL −1 -10 μg mL −1 ) for 48 hours. Cell growth was assessed by MTT assay as previously described. 27 Fraction affected was calculated using the equation (1-OD value of treatment cells/OD value of control ethanol treated cells). Combination indexes (CI) were determined using CalcuSyn software. A CI <1 denotes synergy.

| Clonogenic assay
Cells were pretreated with EtOH or 100 nmol L −1 1,25D 3 for 24 hours and then treated with 0.1 μg mL −1 cDDP for 48 hours or left untreated. Cells were trypsinized using 0.25% trypsin, counted and replated in triplicate at a density of 150 cells per well in a 6-well plate. Colonies formed over 10 days and were fixed and stained with 40% methanol crystal violet.

| shRNA lentiviral transduction
Plasmid DNA was prepared by adding 10 μL of each glycerol stock to 5 mL of 2A-LB broth (low salt) and 100 μg mL −1 of carbenicillin. DNA was isolated using a DNA mini prep kit (Zymo Research). shRNA lentivirus was produced in HEK293T cells using pCMV-dR8.74, VSV-G, and Lipofectamine 2000. T24 and RT-112 cells were transduced with the shRNA lentivirus containing media and 4 μg mL −1 of polybrene. Cells were selected with puromycin (2.5 μg mL −1 ). shRNA knockdown was confirmed at both mRNA and protein levels using qRT-PCR and Western blot analysis. TAp73 shRNA sequence: TGCTGTTGA CAGTGAGCGAGGCCATGCCTGTTTACAAGAATAGT GAAGCCACAGATGTATTCTTGTAAACAGGCATGG CCCTGCCTACTGCCTCGGA.

| siRNA transfection
Cells were plated in antibiotic-free complete media. 5 μmol L −1 siRNA stock solutions were prepared in 1× siRNA buffer. siRNA was diluted in serum-free media. DharmaFECT transfection reagent 1 was diluted in serum-free media. siRNA and DharmaFECT were mixed and incubated for 20 minutes at room temperature. Antibiotic-free complete media was added to a final siRNA concentration of 25 nmol L −1 . Media was removed from cultured cells and replaced with siRNA containing media and incubated at 37°C in 5% CO 2 for 24 hours. siRNA knockdown was confirmed at both mRNA and protein levels using qRT-PCR and Western blot analysis.

| Real-time quantitative RT-PCR
RNA was isolated using Directzol RNA MiniPrep (ZymoResearch) and quantified. cDNA was made using Transcriptor First Strand cDNA Synthesis Kit (Roche) and 1 μg of total RNA. qRT-PCR was performed on an Applied Biosystems 7300 real-time system (Applied Biosystems, Foster City, CA) using iTaq ™ Universal SYBR ® Green Supermix (Bio Rad), 500 nmol L −1 of forward and reverse primers, dH 2 O, and 1 ul of cDNA. Standard thermal cycler conditions were used. Primer pairs were purchases from IDT, sequences can be found in Supplementary Methods.

| Western blot analysis
Cell lysates were prepared using Triton X-100/SDS lysis buffer supplemented with 4% protease inhibitor cocktail and 1% phosphatase inhibitor cocktail. Protein concentrations were determined using Bio-Rad Protein Assay Dye Reagent and made into samples with 30 μg of protein. Samples were run on a 4%-20% SDS-PAGE gel and transferred using a Bio-Rad Semi-dry Transfer Cell. Membranes were blocked with 5% milk and probed with antibodies described in Materials.

| Statistical analysis
All experiments were performed in biological and technical triplicate and graphically represented as the average ± SE. ANOVA followed by post-hoc Bonferroni comparisons were used to determine significance between treatment groups unless stated otherwise. GraphPad Prism was used to calculate significance (P value <0.05). Asterisks alone denote significance compared with control treated samples, asterisks and lines denote significance between treatment groups.

| High VDR expression is associated with improved survival
To evaluate the clinical significance of VDR protein expression in BCa tumor tissues, 359 primary tumor samples collected at Roswell Park Cancer Institute compiled on tissue microarrays (TMAs) were stained with anti-VDR. Patient demographics and clinical characteristics are listed in Tables S1 and S2. H-score was calculated as a combination of intensity and % positive nuclei. Representative images of low (1.5595) and high H-scores (232.2660) can be found in Figure 1A-B. A significant decrease in VDR H-score was observed in patients with muscle invasive disease compared with superficial disease (Table S3, P < 0.001). VDR expression decreased with an increase in grade (P = 0.002), clinical T stage, Path AJCC, Path T and Path N stage (Table S3, P < 0.001). To study survival, VDR H-scores were then classified into quartiles (low: H-score < 45.58, high: H-score > 140.60, and intermediate: 45.58 ≤ H-score ≤ 140.60). A significant association was observed with improved overall survival and disease-specific survival in patients with higher VDR expression ( Figure 1C-D, P = 0.05 and P = 0.008, respectively).

| Vitamin D 3 sufficient mice have an improved response to cisplatin
To investigate the importance of serum VitD levels in the response to cisplatin, we placed mice on a VitD deficient (25 IU, serum 25D 3 4.16 ± 0.97 ng mL −1 ) or sufficient diet (1000 IU, serum 25D 3 17.2 ± 1.12 ng mL −1 ) for 6 weeks to allow serum levels to equilibrate. Mice were inoculated subcutaneously with T24 cells. Tumor growth was not affected by the diets (Figure 2A). Mice were treated with saline or 5 mg kg −1 of cisplatin once a week for 3 weeks ( Figure 2B, day 0, 7, and 14). Five mg kg −1 of cisplatin was determined as the maximum tolerated dose for this schedule (data not shown). After the first dose of cisplatin, all mice had a reduction in tumor volume of ~30% ( Figure 2B, black arrow). However, mice on the deficient diet showed no further reduction in tumor volume despite 2 additional treatments and tumor volume returned to baseline. In contrast, mice on the sufficient diet had a maximum reduction of tumor volume of ~50%, which was stably maintained up to 36 days post treatment (P < 0.05). Tumor volume in the sufficient mice increased slightly 3 weeks after treatment ended, but maintained a 25% reduction. Cisplatin treatment caused a slight drop in mouse weight (5%-7%), which was not altered by the different diets ( Figure S1A). These data suggest that VitD sufficient mice have an improved response to cisplatin compared with deficient mice.

| 1,25D 3 can increase the response to cisplatin in vitamin D 3 deficient mice
After determining that VitD deficient mice have a worse response to cisplatin, we investigated if pretreating deficient mice with 1,25D 3 (0.625 μg mouse −1 , Monday, Wednesday, and Friday (MWF)) could improve response. Mice were placed on a VitD deficient diet (25 IU) for 6 weeks. Mice were inoculated with T24 cells subcutaneously and treated with saline, 1,25D 3 (0.625 μg mouse −1 , MWF), cisplatin (5 mg kg −1 , Friday), or 1,25D 3 followed by cisplatin. After treatment with 1,25D 3 , serum levels of 1,25D 3 increased from approximately 93 to >210 pg mL −1 , the upper limit of the assay. Mice treated with the combination had the largest reduction in tumor volume and a significant reduction in growth rate (P < 0.01 compared with all other groups, Figure 2C, Figure S1D). Mouse weights decreased slightly after cisplatin treatment but rebounded prior to the next round of treatment and were not affected by 1,25D 3 ( Figure S1C). These data indicate that pretreating deficient mice with 1,25D 3 can increase the response to cisplatin in vivo.

| High serum 25D 3 levels associate with improved survival in BCa patients treated with cisplatin
Having determined that VitD sufficient mice have an improved response to cisplatin treatment, we next investigated the effects of serum 25D 3 levels on patients' response to cisplatin. We analyzed serum 25D 3 levels in a cohort of BCa patients who had been treated with cisplatin-based chemotherapy. Out of 71 patients, 26 (36.6%) had low (deficient) 25D 3 (<20 ng mL −1 ) and 45 (63.4%) had high (sufficient or greater) 25D 3 (≥20 ng mL −1 ). 8 Significant associations were found with overall (P = 0.007) and cancer specific survival (P = 0.014) where patients with high 25D 3 had improved outcomes ( Figure 3A-B). This data confirms that VitD deficiency impairs patients' response to cisplatin, and provides rational for serum 25D 3 level to be used as a biomarker for cisplatin response.

| 1,25D 3 and cisplatin combination therapy increases apoptosis
Our lab has shown previously that 1,25D 3 , cisplatin and gemcitabine combination can increase apoptosis. To verify this and investigate other mechanisms where VitD may affect cisplatin response we evaluated multiple aspects of cisplatin activity that may be altered by 1,25D 3 pretreatment. Pretreatment with 1,25D 3 did not significantly alter DNA-platinum adduct formation ( Figure S3A, P < 0.001) or repair ( Figure S3B, P < 0.05), which has been linked to VitD. 29 Nor did pretreatment with 1,25D 3 affect cell cycle ( Figure S3C). However, the percentage of apoptotic cells significantly increased from 5% in control treated cells to 17% with cisplatin monotherapy and to 22% with the combination ( Figure 4E, P < 0.02). These results suggest that 1,25D 3 pretreatment increases the anti-tumor response to cisplatin by increasing apoptosis, consistent with prior data in a squamous cell carcinoma and pancreatic models. 26,30

TAp73 expression
Because the combination treatment decreased the surviving fraction in all 3 cell lines, we hypothesized that p73, not p53,  Figure 5C, *P < 0.05, **P < 0.01). These results suggest that the increase in apoptosis seen after 1,25D 3 pretreatment may be in part due to the increased transcription of TAp73 over ΔNp73. mRNA levels of transcriptional targets of TAp73, BAX and NOXA, and protein levels of BAX were found to increase after the combination treatment ( Figure 5D-F, *P < 0.05, **P < 0.01). TAp73 protein levels increase after the combination treatment in T24 and RT-112 cells ( Figure 5F). These observations confirm an increase in TAp73 expression and transcriptional activity by the combination treatment. Finally, we tested the requirement for VDR in mediating the transcriptional increase in TAp73 and BAX after treatment in T24 cells. VDR knockdown using siRNA ( Figure 6A-mRNA expression, B-protein expression) prevented the induction of TAp73 and BAX mRNA by either monotherapy or the combination ( Figure 6C-D, P < 0.05, normalized to ns-shRNA EtOH treated (no decrease in TAp73 mRNA expression)). These data confirm that the combination treatment increases TAp73 expression and transcriptional activity, and that this induction requires VDR.

| 1,25D 3 and cisplatin treatment requires TAp73
Next, the requirement of TAp73 for the response to the combination treatment was determined. Two different shRNA constructs targeting TAp73 were transduced into T24 and RT-112 cells and stable cell lines were generated (data shown with one per cell line, similar results in each). TAp73 shRNA transduction reduced mRNA and protein levels of TAp73 ( Figure 7A-B). The contribution of TAp73 in the clonogenic capacity was then examined by clonogenic assays. The nonsilencing control cells followed the pattern of the parental T24 and RT-112 cell lines with the greatest reduction in surviving fraction found after the combination treatment ( Figure 7C, P < 0.05). In TAp73 shRNA cells, all effects of treatment were abrogated. Surviving fractions ranged from 86%-96% in T24 cells and 92%-107% in RT-112 cells compared with ethanol treated cells ( Figure 7C, P < 0.05). Representative images from each cell line and treatment group are presented in Figure S4A. The transcriptional induction of BAX after the combination treatment was also abrogated after TAp73 knockdown, confirming that TAp73 contributes to the transcriptional regulation of BAX ( Figure 7D, TAp73 shRNA normalized to EtOH, P < 0.05). These findings indicate that TAp73 is required for the reduction in surviving fraction by the combination treatment.

| 1,25D 3 and cisplatin combination treatment increases the vitamin D 3 response
In order to determine how combination treatment affects the VitD response, T24 and RT-112 cells were treated with ethanol, 1,25D 3 , cisplatin, or 1,25D 3 and cisplatin. Treatment with either 1,25D 3 or cisplatin monotherapy increased VDR mRNA to a similar degree in both T24 and RT-112 cells. The combination treatment significantly increased VDR mRNA expression compared with either monotherapy ( Figure  S5A-B, *P < 0.05, **P < 0.01). VDR protein levels increase with 1,25D 3 and further with 1,25D 3 followed by cisplatin ( Figure 7A). To determine whether increased VDR leads to a functional response, we analyzed two VDR target genes, CYP24A1 and cathelicidin antimicrobial peptide (CAMP). Consistent with the effect on VDR expression, 1,25D 3 alone

| TAp73 is required for the induction of VDR and CYP24A1
Having discovered that TAp73 shRNA reduced the effects of 1,25D 3 monotherapy on surviving fraction ( Figure 6E), we next wanted to determine if TAp73 was required for VitD signaling. We treated T24 and RT-112 cells transduced with TAp73 shRNA as described and analyzed the expression of VDR and CYP24A1. As expected, ns shRNA cells had a significant increase in CYP24A1 mRNA expression and VDR protein expression after 1,25D 3 and an even greater increase after the combination ( Figure 8B-D, P < 0.05). In contrast, in cells transduced with TAp73 shRNA, 1,25D 3 alone or in combination with cisplatin was no longer able to induce VDR protein and CYP24A1 mRNA levels to the same magnitude ( Figure 8B-D, P < 0.05). This result suggests that TAp73 protein is in part required for VitD signaling and may regulate VDR expression.

| DISCUSSION
Treatment with cisplatin in advanced BCa patients has a limited response (5%-10% improved survival over cystectomy). 31,32 Approximately 30%-50% of BCa patients treated with cisplatin-based chemotherapy will not benefit from therapy and 50% of patients treated with cystectomy develop metastatic disease within 2 years. [3][4][5]31,33 New approaches to increase the efficacy of cisplatin are necessary to improve patient outcomes. We determined that VitD deficiency impairs the response to cisplatin, and that supplementation, either through dietary intervention or treatment with 1,25D 3 , can increase the response to cisplatin. It stands to reason that these results could be expanded to other tumor types, specifically those where incidence has been related to vitamin D deficiency such as prostate cancer. The mechanism of action of 1,25D 3 and cisplatin response is through VDR and TAp73 signaling crosstalk to increase the apoptotic response to cisplatin (Figure 9). We identified TAp73 as a key player in the response to 1,25D 3 and cisplatin. This is consistent in the literature where p53 status is not linked to cisplatin response. 14,23 Pretreatment with 1,25D 3 increases the ratio of TAp73/ΔNp73 mRNA, previously determined to be important in the response to chemotherapy. [20][21][22] The expression of the transcriptional targets of TAp73, BAX and NOXA, increases with the combination treatment, in a VDR dependent manner. In addition, treatment with 1,25D 3 and cisplatin appears to increase VitD response through VDR induction. The combination therapy had a greater effect on the expression of VDR and its targets, CYP24A1 and CAMP, than 1,25D 3 alone. It appears that in the presence of ligand, cisplatin can induce VDR protein expression, in a TAp73 dependent manner. This pathway crosstalk creates a positive feedback loop that results in an enhanced apoptotic response. In the literature, there is evidence of crosstalk between VDR and p53 or p73. [34][35][36] Our studies have identified VDR and TAp73 crosstalk in BCa cells as a result of treatment with 1,25D 3 and cisplatin. An important limitation to our studies is the lack of mechanistic data in wild type p53 cell lines. Typically, cell lines with wild type p53 express very little p73, making changes in expression nearly impossible to detect. Due to high structural similarity between p53 and TAp73 functional domains, we hypothesize that crosstalk between VDR and p53 is possible and could explain the response to 1,25D 3 and cisplatin in cell lines with wild type p53. Both VDR and TAp73 have been shown to physically interact with mutant p53 and alter normal function. 37,38 We hypothesize that our treatments may overcome this repression and restore VDR and TAp73 transcriptional regulation.
In summary, this study suggests that VitD plays a role in the response to cisplatin in BCa. VitD sufficiency, through dietary intervention or treatment with 1,25D 3 , can increase the therapeutic response to cisplatin. Patients who received cisplatin treatment with high 25D 3 levels had improved survival. The mechanism of action of 1,25D 3 and cisplatin anti-tumor response is through TAp73 and VDR signaling crosstalk to potentiate apoptosis. Lastly, VitD deficiency may have potential use as biomarkers for response to cisplatin. Our data suggests that patients should be screened for their 25D 3 levels prior to treatment and given intervention if they are found to be VitD deficient.
F I G U R E 9 Diagram of mechanism of action: 1,25D 3 monotherapy activates the transcription factor VDR, which heterodimerizes with retinoid-X receptor (RXR) and transcribes target genes such as CYP24A1. Cisplatin monotherapy activates the transcription factor TAp73 which transcribes pro-apoptotic target genes such as BAX. When we treat with the combination of 1,25D 3 and cisplatin due to signaling crosstalk there is a greater increase in both VDR and TAp73 expression and as a result an increase in transcription regulation of CYP24A1 and BAX, and ultimately the induction of apoptosis