Abstract

Background

Previous evidence suggests that 25-hydroxyvitamin D 3 [25(OH)D 3 ] protects against several cancers. However, little is known regarding urothelial bladder cancer (UBC). We analyzed the association between plasma 25(OH)D 3 and overall risk of UBC, as well as according to stage and FGFR3 molecular subphenotypes.

Methods

Plasma concentrations of 25(OH)D 3 in 1125 cases with UBC and 1028 control subjects were determined by a chemiluminescence immunoassay. FGFR3 mutational status and expression in tumor tissue were assessed. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by logistic regression adjusting for potential confounders. Analyses were further stratified by tumor invasiveness and grade, FGFR3 expression, and smoking status. Cell proliferation was measured in human UBC cell lines cultured with 1α,25-dihydroxyvitamin D 3 .

Results

A statistically significantly increased risk of UBC was observed among subjects presenting the lowest concentrations of 25(OH)D 3 (OR adj = 1.83; 95% CI = 1.19 to 2.82; P = .006), showing a dose–response effect ( Ptrend = .004). The association was stronger for patients with muscle-invasive tumors, especially among low-FGFR3 expressers (OR adj = 5.94; 95% CI = 1.72 to 20.45; P = .005). The biological plausibility of these associations is supported by the fact that, in vitro, 1α,25-dihydroxyvitamin D 3 upregulates FGFR3 expression in UBC cell lines with low levels of wild-type FGFR3.

Conclusion

These findings support a role of vitamin D in the pathogenesis of UBC and show that 25(OH)D 3 levels are associated with FGFR3 expression in the tumor. Because FGFR3 mutation and overexpression are markers of better outcome, our findings suggest that individuals with low levels of plasma 25(OH)D 3 may be at high risk of more aggressive forms of UBC.

Vitamin D, or cholecalciferol, is a prohormone involved in bone biology that may also protect against a variety of cancers ( 1–3 ). The most active vitamin D metabolite, 1α,25dihydroxyvitamin D 3 [1α,25(OH) 2 D 3 , calcitriol], can regulate proliferation, apoptosis, and cell adhesion at the tumor cell level and it can also affect tumor interaction with the microenvironment through modulation of angiogenesis, invasion, and metastasis [reviewed in ( 1 )]. In addition, it decreases oxidative DNA damage ( 4 ). Epidemiologic evidence shows that vitamin D insufficiency, as defined by low levels of 25-hydroxyvitamin D 3 [25(OH)D 3 , calcidiol], which is the major circulating and most stable form of vitamin D, is associated with an increased risk of colorectal and breast cancers ( 5 , 6 ). As for other cancers, including pancreatic and prostate cancers, evidence of association with 25(OH)D 3 is null or controversial ( 7–10 ). A recent paper on prostate cancer reported a statistically significant association with lethal disease only ( 11 ).

Urothelial bladder cancer (UBC) is an important public health issue because of its high incidence in most developed countries and the high costs to society. Spain presents one of the highest UBC incidence rates worldwide, with a male-to-female ratio of seven ( 12 , 13 ). The main established risk factors for UBC are smoking, occupational exposure to aromatic amines, and high levels of arsenic intake ( 12 ). Smoking accounts for a large proportion of the etiology of UBC, whereas the other factors contribute only to specific risk groups; yet, an important fraction of the disease remains unexplained. Little is known about the contribution of vitamin D to UBC, and only two studies have examined the association between 25(OH)D 3 plasma levels and the risk of this disease. Whereas, in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention–nested case–control study of male smokers, low levels of 25(OH)D 3 were associated with increased risk of UBC ( 14 ), in the PLCO nested case–control study, no association was found ( 15 ). 1α,25(OH) 2 D 3 has been shown to inhibit proliferation and induce apoptosis of human bladder tumor cells in vitro and to reduce tumorigenesis in an N-methylnitrosourea-induced model of bladder cancer in rats ( 16 ). Expression of vitamin D receptor has also been detected in human urothelium ( 17 ).

UBC is a heterogeneous disease at the clinical, pathological, and genetic levels and at least two major progression pathways have been identified: papillary, low-grade, non-muscle-invasive bladder cancer (NMIBC) harbors FGFR3 mutations in approximately 60% of cases and displays low levels of genomic instability; high-grade NMIBC and muscle-invasive bladder cancers (MIBC) display a low prevalence of FGFR3 mutations and frequent alterations in the p53 and Rb pathways ( 18 ). Overall, FGFR3 mutations and FGFR3 protein overexpression characterize a large subgroup of NMIBC with good prognosis ( 19 ), but the molecular mechanisms underlying this association are not well established.

We aimed to assess the association between plasma 25(OH)D 3 levels and risk of UBC in the Spanish Bladder Cancer/EPICURO Study (SBC/EPICURO Study) and explore the molecular mechanisms involved therein. We found that treatment with 1α,25(OH) 2 D 3 leads to the upregulation of FGFR3 messenger RNA and protein in cultured UBC cells with low basal expression levels; these observations led to an analysis of the association of plasma 25(OH)D 3 levels with UBC subphenotypes defined according to tumor invasiveness and grade and FGFR3 mutational and expression status.

Methods

Study Participants

Subjects came from the SBC/EPICURO Study, a hospital-based, case–control study conducted in 18 hospitals from five areas in Spain ( 20 ). Briefly, cases were patients newly diagnosed with histologically confirmed UBC in 1998 to 2001. A panel of expert pathologists classified homogeneously all cases according to the invasiveness and grade ( 21 , 22 ). Control subjects were selected from patients admitted to participating hospitals for diagnoses believed to be unrelated to the exposures of interest and were individually matched to the cases on age, sex, ethnic origin, and region. Written informed consent was obtained from all subjects, and the study was approved by the local institutional review boards and the US National Cancer Institute. Information on known or potential cancer risk factors and blood samples were obtained during the inpatient hospital stay for both cases and control subjects. A total of 1219 cases (84% eligible) and 1271 control subjects (88% eligible) agreed to participate in the study and were interviewed. Plasma from blood samples collected at diagnostic time was available from 1130 cases and 1038 control subjects.

Experimental Procedures

Details on the quantification of 25(OH)D 3 , cell proliferation, and FGFR3 expression assays using UBC cell lines and FGFR3 expression and mutational status analyses in tumoral tissue are specified in the Supplementary Methods (available online).

Statistical Analysis

Student’s t test was applied to analyze the effects of vitamin D on proliferation and expression of p21, p27, and FGFR3 in UBC cells. Mann–Whitney U test was used to assess differences between cases and control subjects regarding median plasma concentrations of 25(OH)D 3 . For the analysis of association between 25(OH)D 3 and UBC, logistic regression was applied to estimate odds ratios (ORs) and their 95% confidence intervals (CIs), comparing each category of low plasma 25(OH)D 3 concentration (20–29.99, 15–19.99, 10–14.99, and <10ng/mL) with the reference category (≥30ng/mL). A basic model was adjusted for age at interview, sex, region, smoking status, and season of blood draw. Further adjustments were made for body mass index, alcohol and calcium intake as previously associated with vitamin D ( 23–26 ), and occupational exposure to aromatic amines and toenail arsenic. Tests for linear trend were computed with the median of each category of the plasma 25(OH)D 3 concentration treated as a continuous variable. Because almost a quarter of the control subjects were admitted to the hospital with bone fractures, a sensitivity analysis was carried out excluding these control subjects. The risk of UBC was further evaluated according to smoking status, with the “ever smoker” category created by collapsing the categories of occasional, former, and current smokers. Statistical interaction between smoking and 25(OH)D 3 was assessed by including an interaction term as the product of the median of each category of plasma 25(OH)D 3 and the never smoker or ever smoker categories. This was repeated by adjusting for duration of cigarette smoking, cigarettes per day, and pack-years among the ever smokers. The association with 25(OH)D 3 was also examined by type of tobacco (blond or black) among the ever smokers. The analysis was also stratified by season of blood draw, and the interaction between season and plasma 25(OH)D 3 was assessed.

Polytomous logistic regression models were applied to analyze the association between 25(OH)D 3 and risk of low-grade NMIBC, high-grade NMIBC, and MIBC. Adjustment was performed for the same variables included in the logistic regression models. Difference in odds ratios between case groups was tested using a likelihood ratio test comparing models with and without the odds ratio constrained to be equal for the corresponding case groups. The polytomous logistic regression models were also applied to the stratified analysis for low and high FGFR3 expression and the presence or absence of mutations in FGFR3 in tumor tissue, adjusting for the same variables as before. All statistical tests were two-sided, and results were considered significant when P was less than or equal to .05. Statistical analyses were performed using STATA/SE version 10.1. This study conforms to the guidelines of Strengthening the Reporting of Observational Studies in Epidemiology Statement for observational studies.

Results

Effects of 1α,25(OH) 2 D 3 on Cultured UBC

For this study, four UBC cell lines with distinct features were selected. Under basal conditions, two of them show an epithelial adhesive phenotype and form compact colonies: RT112 displays high levels of constitutively active wild-type FGFR3, whereas MGH-U3 harbors constitutively active mutant FGFR3. In contrast, the other two lines show a less epithelial phenotype: J82 cells have a mutated FGFR3 but lack expression both at the RNA and protein levels, whereas MGH-U4 cells express low levels of wild-type FGFR3 ( 27 ).

Treatment with 1α,25(OH) 2 D 3 induced a more epithelioid phenotype and formation of more compact colonies, consistent with findings reported in other cell types ( Figure 1 , A ). In the four UBC cell lines studied, 1α,25(OH) 2 D 3 (10–100nM) induced a growth arrest ( Figure 1 , B ). This was associated with the upregulation of the CDK inhibitors p21 CIP1 and p27 KIP1 at the protein level ( Figure 1 , C ). As shown in Figure 1 , D and E , 1α,25(OH) 2 D 3 also induced an upregulation of FGFR3 at the messenger RNA level in MGH-U4 and RT112 cells and at the protein level in MGH-U4 cells. Altogether, these findings indicate that vitamin D treatment is associated with phenotypic changes and growth inhibition; in addition, it leads to higher FGFR3 levels in cells with low basal expression.

Figure 1.

Effects of vitamin D on proliferation and expression of p21, p27, and FGFR3 in urothelial bladder cancer (UBC) cells. Vitamin D induces growth arrest and an upregulation of p21, p27, and FGFR3 in UBC cells. A ) Phase contrast microscopy of MGH-U4 and RT112 cells treated with 1α,25-dihydroxyvitamin D 3 [1α,25 (OH) 2 D 3 ] (100nM) or vehicle for 72 hours (scale bar = 100 µm). B ) 1α,25(OH) 2 D 3 treatment inhibits proliferation of UBC cells in vitro. C ) Western blot analysis showing induction of the CDK inhibitors p21 and p27 in MGH-U4 cells treated with 1α,25(OH) 2 D 3 at different time points. Tubulin was used as a loading control. Values indicate fold-change referred to zero time point. D ) Western blot analysis showing changes in FGFR3 in MGH-U4 and RT112 cells treated with 1α,25(OH) 2 D 3 at the indicated time points. E ) Quantitative reverse-transcription polymerase chain reaction analysis showing the upregulation of FGFR3 messenger RNA levels in MGH-U4 and RT112 cells treated with 10nM 1α,25(OH) 2 D 3(gray bars) . Values were normalized to HPRT and referred to expression at time zero ( black bars ). Comparisons with P less than .05 are indicated with an asterisk. Error bars represent standard deviation.

Figure 1.

Effects of vitamin D on proliferation and expression of p21, p27, and FGFR3 in urothelial bladder cancer (UBC) cells. Vitamin D induces growth arrest and an upregulation of p21, p27, and FGFR3 in UBC cells. A ) Phase contrast microscopy of MGH-U4 and RT112 cells treated with 1α,25-dihydroxyvitamin D 3 [1α,25 (OH) 2 D 3 ] (100nM) or vehicle for 72 hours (scale bar = 100 µm). B ) 1α,25(OH) 2 D 3 treatment inhibits proliferation of UBC cells in vitro. C ) Western blot analysis showing induction of the CDK inhibitors p21 and p27 in MGH-U4 cells treated with 1α,25(OH) 2 D 3 at different time points. Tubulin was used as a loading control. Values indicate fold-change referred to zero time point. D ) Western blot analysis showing changes in FGFR3 in MGH-U4 and RT112 cells treated with 1α,25(OH) 2 D 3 at the indicated time points. E ) Quantitative reverse-transcription polymerase chain reaction analysis showing the upregulation of FGFR3 messenger RNA levels in MGH-U4 and RT112 cells treated with 10nM 1α,25(OH) 2 D 3(gray bars) . Values were normalized to HPRT and referred to expression at time zero ( black bars ). Comparisons with P less than .05 are indicated with an asterisk. Error bars represent standard deviation.

Plasma 25(OH)D 3 and Risk of UBC

Based on previous epidemiological evidence and on the above in vitro findings, we analyzed the association between vitamin D levels and UBC risk in the SBC/EPICURO Study, both overall and according to molecular subphenotypes. Cases and control subjects were mostly men, with a high frequency of cigarette smokers ( Table 1 ). Median concentrations of 25(OH)D 3 were lower in cases than in control subjects (13.9 vs 15.0ng/mL; P = .001) ( Table 1 ); 73% of all individuals (75% of cases and 71% of control subjects) had less than 20ng/mL ( Supplementary Table 1 , available online). The distribution of cases by 25(OH)D 3 status was similar in the three tumor subphenotypes examined ( P = .70) ( Supplementary Table 1 , available online).

Table 1.

Characteristics of study participants*

Characteristics Control subjects (N = 1028) Cases (N = 1125) P †  
Age, median (range), y 66 (20–81)  68 (22–81)   1.4×10 −4 
Sex      
    Males 909 88 986 88 .58 
    Females 119 12 139 12  
Region      
    Barcelona 205 20 197 18 .41 
    Valles 159 15 180 16  
    Elche   81   8   88   8  
    Tenerife 153 15 196 17  
    Asturias 430 42 464 41  
Smoking status      
    Never smoker 290 28 157 14  1.9×10 −23 
    Occasional smoker   82   8   45   4  
    Former smoker 383 37 439 39  
    Current smoker 273 27 484 43  
BMI (kg/m 2 )‡       
    <25 415 53 499 58 .13 
    25–26.99 169 22 172 20  
    27–29.99 136 17 123 14  
    >30   65   8   60   7  
Tumor type§      
    Low-grade NMIBC (TaG1/G2) –– –– 579 56  
    High-grade NMIBC (TaG3/T1) –– –– 205 20  
    MIBC (≥T2) –– –– 246 24  
FGFR3 ||  –– ––    
    Wild-type –– –– 496 59  
    Mutated –– –– 340 41  
FGFR3 expression¶ –– ––    
    Low –– –– 396 59  
    High –– –– 271 41  
Plasma 25-hydroxyvitamin D 3 , median (interquartile range), ng/mL  15.0 (10.0–21.2)  13.9 (9.0–19.9)   1.1×10 −3 
Characteristics Control subjects (N = 1028) Cases (N = 1125) P †  
Age, median (range), y 66 (20–81)  68 (22–81)   1.4×10 −4 
Sex      
    Males 909 88 986 88 .58 
    Females 119 12 139 12  
Region      
    Barcelona 205 20 197 18 .41 
    Valles 159 15 180 16  
    Elche   81   8   88   8  
    Tenerife 153 15 196 17  
    Asturias 430 42 464 41  
Smoking status      
    Never smoker 290 28 157 14  1.9×10 −23 
    Occasional smoker   82   8   45   4  
    Former smoker 383 37 439 39  
    Current smoker 273 27 484 43  
BMI (kg/m 2 )‡       
    <25 415 53 499 58 .13 
    25–26.99 169 22 172 20  
    27–29.99 136 17 123 14  
    >30   65   8   60   7  
Tumor type§      
    Low-grade NMIBC (TaG1/G2) –– –– 579 56  
    High-grade NMIBC (TaG3/T1) –– –– 205 20  
    MIBC (≥T2) –– –– 246 24  
FGFR3 ||  –– ––    
    Wild-type –– –– 496 59  
    Mutated –– –– 340 41  
FGFR3 expression¶ –– ––    
    Low –– –– 396 59  
    High –– –– 271 41  
Plasma 25-hydroxyvitamin D 3 , median (interquartile range), ng/mL  15.0 (10.0–21.2)  13.9 (9.0–19.9)   1.1×10 −3 

* BMI = body mass index; MIBC = muscle-invasive bladder cancer; NMIBC = non-muscle-invasive bladder cancer †For age and plasma 25-hydroxyvitamin D 3 , the P value is from the Mann–Whitney U test. For all the other variables, the P value is from the χ 2 test.

‡ Two hundred forty-three control subjects and 271 cases had no information on height or weight or both.

§ Ninety-five cases could not be assigned to any tumor/grade group because the paraffin block could not be retrieved.

|| Two hundred eighty-nine cases did not yield polymerase chain reaction product.

¶ Four hundred fifty-eight cases did not have immunohistochemistry staining for FGFR3.

After adjusting for age, sex, region, smoking status, and season of blood draw, decreasing concentrations of plasma 25(OH)D 3 were found to be associated with increased risk of UBC ( Ptrend = .004) ( Table 2 ). Individuals slightly (OR adj = 1.63; 95% CI = 1.06 to 2.51; P = .03), moderately (OR adj = 1.67; 95% CI = 1.09 to 2.56; P = .02), and severely deficient (OR adj = 1.83; 95% CI = 1.19 to 2.82; P = .006) in vitamin D presented a greater than 50% increased risk of UBC when compared with individuals with sufficient levels. The results were not substantially changed after excluding the control subjects with bone fractures or adjusting for body mass index ( Supplementary Table 2 , available online), intake of alcohol and calcium, occupational exposure to aromatic amines, or toenail arsenic levels (data not shown).

Table 2.

Odds ratios (ORs), 95% confidence intervals (CIs), and P values for the association between plasma 25-hydroxyvitamin D 3 [25(OH)D 3 ] and overall bladder cancer risk, risk of low- and high-grade non-muscle-invasive (NMIBC) and muscle-invasive bladder cancer (MIBC)*

Overall Low-grade NMIBC High-grade NMIBC MIBC 
Vitamin D status  25(OH) D 3 , ng/mL  Controls Cases OR (95% CI)† Cases OR (95% CI)† P Cases OR (95% CI)† P Cases OR (95% CI)† P 
Sufficient ≥30.00 74  51 1.00 (referent)   27 1.00 (referent)  10 1.00 (referent)  1.00 (referent)  
Insufficient 20.00–29.99 227 229 1.40 (0.92 to 2.14) .12 120 1.35 (0.81 to 2.26) .25 38 1.14 (0.53 to 2.45) .73 44 1.44 (0.66 to 3.14) .37 
Slightly deficient 15.00–19.99 212 219 1.63 (1.06 to 2.51) .03 116 1.58 (0.94 to 2.66) .08 46 1.68 (0.79 to 3.57) .18 43 1.83 (0.83 to 4.02) .14 
Moderately deficient 10.00–14.99 255 280 1.67 (1.09 to 2.56) .02 146 1.55 (0.92 to 2.59) .096 49 1.40 (0.66 to 2.99) .38 61 2.13 (0.98 to 4.65) .06 
Severely deficient <10.00 260 346 1.83 (1.19 to 2.82) .006 170 1.64 (0.97 to 2.76) .07 62 1.55 (0.72 to 3.32) .26 89 2.81 (1.29 to 6.13) .009 
Ptrend     .004   .07   .18   .0005 
Overall Low-grade NMIBC High-grade NMIBC MIBC 
Vitamin D status  25(OH) D 3 , ng/mL  Controls Cases OR (95% CI)† Cases OR (95% CI)† P Cases OR (95% CI)† P Cases OR (95% CI)† P 
Sufficient ≥30.00 74  51 1.00 (referent)   27 1.00 (referent)  10 1.00 (referent)  1.00 (referent)  
Insufficient 20.00–29.99 227 229 1.40 (0.92 to 2.14) .12 120 1.35 (0.81 to 2.26) .25 38 1.14 (0.53 to 2.45) .73 44 1.44 (0.66 to 3.14) .37 
Slightly deficient 15.00–19.99 212 219 1.63 (1.06 to 2.51) .03 116 1.58 (0.94 to 2.66) .08 46 1.68 (0.79 to 3.57) .18 43 1.83 (0.83 to 4.02) .14 
Moderately deficient 10.00–14.99 255 280 1.67 (1.09 to 2.56) .02 146 1.55 (0.92 to 2.59) .096 49 1.40 (0.66 to 2.99) .38 61 2.13 (0.98 to 4.65) .06 
Severely deficient <10.00 260 346 1.83 (1.19 to 2.82) .006 170 1.64 (0.97 to 2.76) .07 62 1.55 (0.72 to 3.32) .26 89 2.81 (1.29 to 6.13) .009 
Ptrend     .004   .07   .18   .0005 

* Ninety-five cases could not be assigned to any tumor/grade group because the paraffin block could not be retrieved. Likelihood-ratio test P for the pairwise comparisons between odds ratio of low-grade and high-grade NMIBC and MIBC among the severely deficient (vs sufficient) were .89, low-grade vs high-grade NMIBC; .19, low-grade NMIBC vs MIBC; and .24, high-grade NMIBC vs MIBC.

† Adjusted for age, sex, region, smoking status, and season of blood collection.

The association of plasma 25(OH)D 3 levels and risk of UBC was restricted to smokers, showing a dose–response effect ( P = .003) ( Figure 2 , A ). However, no statistical interaction was observed between tobacco and 25(OH)D 3 levels. Adjusting for duration of cigarette smoking, cigarettes per day, and pack-years among ever smokers did not change these results (data not shown). Stratifying by type of tobacco among ever smokers did not substantially change the results either (data not shown).

Figure 2.

Odds ratios (OR) and 95% confidence intervals for the association between plasma 25-hydroxyvitamin D 3 [25(OH)D 3 ] and bladder cancer risk by smoking status ( A ) and by season of blood collection ( B ). Estimates are adjusted for age, sex, region, smoking status, and season of blood collection, when appropriate.

Figure 2.

Odds ratios (OR) and 95% confidence intervals for the association between plasma 25-hydroxyvitamin D 3 [25(OH)D 3 ] and bladder cancer risk by smoking status ( A ) and by season of blood collection ( B ). Estimates are adjusted for age, sex, region, smoking status, and season of blood collection, when appropriate.

The association of plasma 25(OH)D 3 concentration with risk of UBC was stronger and showed a dose–response pattern among individuals whose blood was drawn in spring and summer seasons, but a statistically significant interaction was not observed ( Figure 2 , B ).

Low concentrations of plasma 25(OH)D 3 were more strongly associated with risk of MIBC. Among individuals severely deficient in vitamin D, the adjusted risk of MIBC (OR adj = 2.81; 95% CI = 1.29 to 6.13; P = .009) was 1.7 times higher than the risk of low-grade NMIBC (OR adj = 1.64; 95% CI = 0.97 to 2.76; P = .07) ( Table 2 ). However, the differences in risk between both tumor types were not statistically significant, possibly due to low sample size.

Plasma 25(OH)D 3 and FGFR3 Mutation and Protein Expression

Tumors with FGFR3 mutations were more likely to show high FGFR3 expression than those without mutations ( P = 2×10 −18 ). Plasma 25(OH)D 3 levels were not associated with somatic FGFR3 mutations ( P = .70) ( Supplementary Table 1 , available online). Risk of UBC among subjects with deficient 25(OH)D 3 concentrations was slightly higher among FGFR3 -mutated than wild-type tumors, although the differences were not statistically significant ( Supplementary Table 3 , available online). The percentage of cases with high FGFR3-expressing tumors was slightly higher in the 25(OH)D 3 -sufficient group than in the 25(OH)D 3 -deficient group ( Supplementary Table 1 , available online), but this difference did not reach statistical significance ( P = .10).

A more detailed analysis revealed that low plasma concentrations of 25(OH)D 3 were associated with an increased risk of developing low FGFR3-expressing UBC but not high FGFR3-expressing UBC. This association was more notable among those severely deficient in vitamin D (OR adj = 3.03; 95% CI = 1.55 to 5.94; P = .001; Ptrend = .0002) ( Supplementary Table 3 , available online). Furthermore, in individuals with low levels of 25(OH)D 3 , the odds of MIBC expressing low levels of FGFR3 were almost 6-fold higher than in those with sufficient levels (OR adj = 5.94; 95% CI = 1.72 to 20.45; P = .005). This association was not statistically significantly different from that of low-grade ( P = .30) and high-grade NMIBC ( P = .10) ( Figure 3 ; Supplementary Table 4 , available online), also possibly due to small sample size in the subgroups.

Figure 3.

Odds ratios (ORs) and 95% confidence intervals (CIs) for the association between plasma 25-hydroxyvitamin D 3 [25(OH)D 3 ] levels and risk of low- and high-grade non-muscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) in relationship with tumor FGFR3 expression levels. Estimates are adjusted for age, sex, region, and smoking status. *Likelihood-ratio test P for the comparison between the odds ratio of MIBC in the low- and high-FGFR3-expression groups = .26.

Figure 3.

Odds ratios (ORs) and 95% confidence intervals (CIs) for the association between plasma 25-hydroxyvitamin D 3 [25(OH)D 3 ] levels and risk of low- and high-grade non-muscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) in relationship with tumor FGFR3 expression levels. Estimates are adjusted for age, sex, region, and smoking status. *Likelihood-ratio test P for the comparison between the odds ratio of MIBC in the low- and high-FGFR3-expression groups = .26.

Discussion

In the present study, we analyzed the association of plasma 25(OH)D 3 with risk of UBC in the largest and most representative patient series tested so far. For the first time, we placed the findings in the context of the molecular taxonomy of this tumor, namely according to alterations in FGFR3 , which is the most commonly mutated oncogene in UBC ( 18 ). We observed an inverse, statistically significant association between plasma 25(OH)D 3 levels and risk of UBC with a dose–response effect: individuals with the lowest concentrations of plasma 25(OH)D 3 presented almost twofold higher odds than individuals with concentrations greater to or equal than 30ng/mL (sufficient status). This risk pattern was mainly observed for MIBC. Furthermore, although the increased risk of low-grade NMIBC was independent of FGFR3 expression in the tumor, that of MIBC was not: individuals with deficient levels of vitamin D showed the highest risk for developing low-FGFR3-expressing MIBC. These findings suggest that vitamin D modulates tumor phenotype in specific tumor subtypes.

Tumors that express low FGFR3 protein levels or that are FGFR3 wild-type are more likely to invade muscle and display an aggressive behavior, whereas tumors that are FGFR3 mutant and express high FGFR3 levels have a lower tendency to progress ( 19 , 28 ). The in vitro data indicate that vitamin D regulates FGFR3 mainly in cells expressing low FGFR3 levels, and a large body of evidence indicates that tumor cells that respond to vitamin D display more differentiated properties and less aggressive in vitro behavior ( 29 , 30 ). Therefore, our findings are consistent with the notion that vitamin D sufficiency may support higher levels of expression of FGFR3, particularly in wild-type tumors, thus favoring a less-aggressive tumor phenotype at presentation. The recent observation that vitamin D can affect gene expression through the epigenetic modulation of histone marks suggests a possible mechanism of this effect ( 31 ). Our results support the notion that in other tumor types, such as colon and breast cancers, similar analyses should be performed to determine whether the effects of vitamin D are restricted to specific tumor subphenotypes and/or molecular pathways.

The study findings build upon data reported by the few previous epidemiologic and molecular studies suggesting that vitamin D may act as a protective factor against UBC ( 14 , 16 ) and extend our understanding by exploring interactions and the association of 25(OH)D 3 plasma levels with UBC FGFR3 subphenotypes. None of the interactions tested were statistically significant. However, the inverse association between 25(OH)D 3 and risk of UBC appeared stronger among ever smokers and among those individuals whose blood was collected during spring or summer months. Although we cannot discard that the relationship with tobacco smoking could be due to chance because of the small sample size of the nonsmoker group, it is in agreement with the findings of Mondul and colleagues, who reported an increased risk of bladder cancer in male smokers associated with low 25(OH)D 3 serum concentrations ( 14 ). The stronger effects found among subjects whose blood was drawn during spring and summer confirm prior findings and suggest that assessing plasma levels of 25(OH)D 3 during sunnier months provides a more sensitive biomarker of a constitutive deficiency of vitamin D and thus better discriminates those individuals with higher susceptibility to UBC ( 14 ).

These results are of relevance given that vitamin D deficiency and insufficiency are highly prevalent in Spain ( 32 , 33 ), where the incidence rates of UBC are among the highest worldwide ( 12 , 13 ), and in many other Western countries. The concentrations of 25(OH)D 3 found in this study were similar to those of same-age individuals from other Southern European countries, although lower than those of individuals from the United States and Sweden. A potential explanation is that, in the latter countries, several food items are fortified with vitamin D ( 34–36 ).

This is the largest study assessing the risk of UBC in relation to 25(OH)D 3 levels and the first one analyzing this association in the context of the molecular features of the tumor and the biological effects of vitamin D, supported by parallel experimental in vitro evidence providing mechanistic explanations for the epidemiological findings. Other relevant strengths of this study are the high participation rates of cases and control subjects as well as their match for area of residence and similar age distribution. In addition, detailed information on several potential confounders (eg, smoking habits, body mass index) were considered. Even though the results of the present study are based on the concentration of plasma 25(OH)D 3 at a single time point, this measurement is considered a good biomarker of long-term vitamin D status because several studies have shown moderate to very high intraclass correlation coefficients (≥0.59), indicating a good concordance in 25(OH)D 3 across time points ( 37–39 ).

However, our study also has some limitations. Despite its large sample size, the assessment of associations in subgroups is limited by the smaller numbers of subjects in each subphenotype, especially when considering the association with tumor FGFR3 mutation and expression. The association between 25(OH)D 3 levels and FGFR3 expression and tumor subtype needs to be confirmed in adequately sized independent series. Temporality should also be taken into account because the study is inherently retrospective and we cannot discard a reverse causality due to the carcinogenesis process. Although we hypothesized that, according to the mechanistic evidence provided here, the protective effect of 25(OH)D 3 should be more pronounced among patients with MIBC low-FGFR3 expressers, we cannot exclude that a cancer diagnosis may lead to a change in diet and outdoor habits, potentially influencing 25(OH)D 3 concentrations. Nevertheless, all present patients with UBC were incident cases, most of whom were in good general health at the time of diagnosis and were not malnourished or cachetic. Also, blood was drawn at time of diagnosis, and plasma levels of 25(OH)D 3 are considered reasonably consistent over time ( 37 , 39 , 40 ). Furthermore, the fact the association is more evident in a group molecularly defined would exclude the possibility of reverse causality. Importantly, our results are in line with previous evidence from both case–control and cohort studies at other cancer sites ( 5 , 6 , 14 ).

In summary, low plasma 25(OH)D 3 concentrations were found associated with an increased risk of UBC, and the effects of vitamin D may be stronger among smokers. Our data suggest that this risk is higher among those individuals with MIBC expressing low FGFR3 levels. The in vitro findings reported here lend support to the biological plausibility of this association. Our results need to be replicated in independent populations, and the benefits of vitamin D intake have to be conclusively assessed through a clinical trial.

Funding

This work was supported by the World Cancer Research Fund (2010/250); Fondo de Investigación Sanitaria, Spain (00/0745, PI051436, PI061614, and G03/174); Red Temática de Investigación Cooperativa en Cáncer, Spain; Fundación Científica de la AECC, Spain; Consolíder ONCOBIO; Comunidad de Madrid, Spain (S2011/BMD-02344, Colomics2); and the Intramural Research Program of the Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health.

We acknowledge the coordinators, field and administrative workers, technicians and patients of the Spanish Bladder Cancer/EPICURO Study, and the Histology and Immunohistochemistry Core Unit of the Spanish National Cancer Research Centre (CNIO).

FX Real and N Malats contributed equally to this work. The authors declare no conflicts of interest.

References

1.
Garland
CF
Gorham
ED
Mohr
SB
et al
Vitamin D for cancer prevention: global perspective
Ann Epidemiol
 
2009
19
(
7
):
468
483
2.
Giovannucci
E.
The epidemiology of vitamin D and cancer incidence and mortality: a review (United States)
Cancer Causes Control
 
2005
16
(
2
):
83
95
3.
Holick
MF.
Vitamin D deficiency
N Engl J Med
 
2007
357
(
3
):
266
281
4.
Fedirko
V
Bostick
RM
Long
Q
et al
Effects of supplemental vitamin D and calcium on oxidative DNA damage marker in normal colorectal mucosa: a randomized clinical trial
Cancer Epidemiol Biomarkers Prev
 
2010
19
(
1
):
280
291
5.
Touvier
M
Chan
DSM
Lau
R
et al
Meta-analyses of vitamin D intake, 25-hydroxyvitamin D status, vitamin D receptor polymorphisms and colorectal cancer risk
Cancer Epidemiol Biomarkers Prev
 
2011
20
(
5
):
1003
1016
6.
Yin
L
Grandi
N
Raum
E
et al
Meta-analysis: serum vitamin D and breast cancer risk
Eur J Cancer
 
2010
46
(
12
):
2196
2205
7.
Helzlsouer
KJ.
Overview of the cohort consortium vitamin D pooling project of rarer cancers
Am J Epidemiol
 
2010
172
(
1
):
4
9
8.
Albanes
D
Mondul
AM
Yu
K
et al
Serum 25-hydroxy vitamin D and prostate cancer risk in a large nested case-control study
Cancer Epidemiol Biomarkers Prev
 
2011
20
(
9
):
1850
1860
9.
Yin
L
Raum
E
Haug
U
et al
Meta-analysis of longitudinal studies: serum vitamin D and prostate cancer risk
Cancer Epidemiol
 
2009
33
(
6
):
435
445
10.
Wolpin
BM
Ng
K
Bao
Y
et al
Plasma 25-hydroxyvitamin D and risk of pancreatic cancer
Cancer Epidemiol Biomarkers Prev
 
2012
21
(
1
):
82
91
11.
Shui
IM
Mucci
LA
Kraft
P
et al
Vitamin D-related genetic variation, plasma vitamin D, and risk of lethal prostate cancer: a prospective nested case-control study
J Natl Cancer Inst
 
2012
104
(
9
):
690
699
12.
Silverman
D
Devesa
S
Moore
L
et al
Bladder cancer
. In:
Schottenfeld
D
Fraumeni
J
Jr
, eds.
Cancer Epidemiology and Prevention
 
3rd ed
New York
Oxford University Press
;
2006
1101
1127
13.
Samanic
C
Kogevinas
M
Dosemeci
M
et al
Smoking and bladder cancer in Spain: effects of tobacco type, timing, environmental tobacco smoke, and gender
Cancer Epidemiol Biomarkers Prev
 
2006
15
(
7
):
1348
1354
14.
Mondul
AM
Weinstein
SJ
Mannisto
S
et al
Serum vitamin D and risk of bladder cancer
Cancer Res
 
2010
70
(
22
):
9218
9223
15.
Mondul
AM
Weinstein
SJ
Horst
RL
et al
Serum vitamin D and risk of bladder cancer in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial
Cancer Epidemiol Biomarkers Prev
 
2012
21
(
7
):
1222
1225
16.
Konety
BR
Lavelle
JP
Pirtskalaishvili
G
et al
Effects of vitamin D (calcitriol) on transitional cell carcinoma of the bladder in vitro and in vivo
J Urol
 
2001
165
(
1
):
253
258
17.
Sahin
MO
Canda
AE
Yorukoglu
K
et al
1,25 Dihydroxyvitamin D(3) receptor expression in superficial transitional cell carcinoma of the bladder: a possible prognostic factor?
Eur Urol
 
2005
47
(
1
):
52
57
18.
Luis
NM
Lopez-Knowles
E
Real
FX.
Molecular biology of bladder cancer
Clin Transl Oncol
 
2007
;9(
1
):
5
12
19.
Hernandez
S
Lopez-Knowles
E
Lloreta
J
et al.  
Prospective study of FGFR3 mutations as a prognostic factor in nonmuscle invasive urothelial bladder carcinomas
J Clin Oncol
 
2006
24
(
22
):
3664
3671
20.
Garcia-Closas
M
Malats
N
Silverman
D
et al.  
NAT2 slow acetylation, GSTM1 null genotype, and risk of bladder cancer: results from the Spanish Bladder Cancer Study and meta-analyses
Lancet
 
2005
366
(
9486
):
649
659
21.
Epstein
JI
Amin
MB
Reuter
VR
et al
The World Health Organization/International Society of Urological Pathology consensus classification of urothelial (transitional cell) neoplasms of the urinary bladder. Bladder Consensus Conference Committee
Am J Surg Pathol
 
1998
22
(
12
):
1435
1448
22.
Mostofi
F
Davis
C
Sesterhenn
I.
Histological typing of urinary bladder tumours
 
Berlin
Springer
;
1999
23.
van Dam
RM
Snijder
MB
Dekker
JM
et al
Potentially modifiable determinants of vitamin D status in an older population in the Netherlands: the Hoorn Study
Am J Clin Nutr
 
2007
85
(
3
):
755
761
24.
Jaaskelainen
T
Knekt
P
Marniemi
J
et al
Vitamin D status is associated with sociodemographic factors, lifestyle and metabolic health
[published online ahead of print April 27, 2012].
Eur J Nutr
 
2012
. doi:10.1007/s00394-012-0354-0.
25.
Kinyamu
HK
Gallagher
JC
Rafferty
KA
et al
Dietary calcium and vitamin D intake in elderly women: effect on serum parathyroid hormone and vitamin D metabolites
Am J Clin Nutr
 
1998
67
(
2
):
342
348
26.
Looker
AC
Pfeiffer
CM
Lacher
DA
et al
Serum 25-hydroxyvitamin D status of the US population: 1988–1994 compared with 2000–2004
Am J Clin Nutr
 
2008
88
(
6
):
1519
1527
27.
Jebar
AH
Hurst
CD
Tomlinson
DC
et al
FGFR3 and Ras gene mutations are mutually exclusive genetic events in urothelial cell carcinoma
Oncogene
 
2005
24
(
33
):
5218
5225
28.
Tomlinson
DC
Baldo
O
Harnden
P
et al
FGFR3 protein expression and its relationship to mutation status and prognostic variables in bladder cancer
J Pathol
 
2007
213
(
1
):
91
98
29.
Palmer
HG
Larriba
MJ
Garcia
JM
et al
The transcription factor SNAIL represses vitamin D receptor expression and responsiveness in human colon cancer
Nat Med
 
2004
10
(
9
):
917
919
30.
Palmer
HG
Gonzalez-Sancho
JM
Espada
J
et al
Vitamin D(3) promotes the differentiation of colon carcinoma cells by the induction of E-cadherin and the inhibition of beta-catenin signaling
J Cell Biol
 
2001
154
(
2
):
369
387
31.
Pereira
F
Barbachano
A
Silva
J
et al
KDM6B/JMJD3 histone demethylase is induced by vitamin D and modulates its effects in colon cancer cells
Hum Mol Genet
 
2011
20
(
23
):
4655
4665
32.
Gonzalez-Padilla
E
Soria Lopez
A
Gonzalez-Rodriguez
E
et al
High prevalence of hypovitaminosis D in medical students in Gran Canaria. Canary Islands (Spain)
Endocrinol Nutr
 
2011
58
(
6
):
267
273
33.
Gonzalez-Molero
I
Morcillo
S
Valdes
S
et al
Vitamin D deficiency in Spain: a population-based cohort study
Eur J Clin Nutr
 
2011
65
(
3
):
321
328
34.
Lips
P.
Vitamin D deficiency and secondary hyperparathyroidism in the elderly: consequences for bone loss and fractures and therapeutic implications
Endocr Rev
 
2001
22
(
4
):
477
501
35.
Melhus
H
Snellman
G
Gedeborg
R
et al
Plasma 25-hydroxyvitamin D levels and fracture risk in a community-based cohort of elderly men in Sweden
J Clin Endocrinol Metab
 
2010
95
(
6
):
2637
2645
36.
Looker
AC
Johnson
CL
Lacher
DA
et al
Vitamin D status: United States, 2001–2006
NCHS Data Brief
 
2011
;
59
1
8
37.
Hofmann
JN
Yu
K
Horst
RL
et al
Long-term variation in serum 25-hydroxyvitamin D concentration among participants in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial
Cancer Epidemiol Biomarkers Prev
 
2010
19
(
4
):
927
931
38.
Zeleniuch-Jacquotte
A
Gallicchio
L
Hartmuller
V
et al
Circulating 25-hydroxyvitamin D and risk of endometrial cancer: Cohort Consortium Vitamin D Pooling Project of Rarer Cancers
Am J Epidemiol
 
2010
172
(
1
):
36
46
39.
Meng
JE
Hovey
KM
Wactawski-Wende
J
et al
Intraindividual variation in plasma 25-hydroxyvitamin D measures 5 years apart among postmenopausal women
Cancer Epidemiol Biomarkers Prev
 
2012
21
(
6
):
916
924
40.
Saliba
W
Barnett
O
Stein
N
et al
The longitudinal variability of serum 25(OH)D levels
Eur J Intern Med
 
2012
23
(
4
):
e106
e111