Table 1.
Clinical characteristics of patients from the CALGB 80303 and Mayo Clinic studies.
Table 2.
SNPs with concordant effects and p<0.05 in the genome-wide screens of both the CALGB 80303 and Mayo Clinic patients.
Fig 1.
Kaplan-Meier plots of the association between VDR rs2853564 and OS in patients from the CALGB 80303 (A) and Mayo Clinic (B). rs2853564 is an A>G change, and the survival curves for each genotype are reported. Tables indicate the number of patients who were alive and at risk of death at each time point after the diagnosis of pancreatic cancer.
Fig 2.
Kaplan-Meier plots of the association between VDR rs2853564 and OS stratified by the serum baseline levels of 25-hydroxyvitamin D (25(OH)D) (A) and gemcitabine treatment (B). (A) The median (95% CI) OS (months) values stratified by the level of 25(OH)D (high vs. low) and rs2853564 genotypes (AA, AG, GG) are as follows: AA/high (n = 31): 4.3 (95% CI 3.3–5.9), AA/low (n = 33): 6.6 (95% CI 4.2–12.0), AG/high (n = 55): 6.7 (95% CI 4.0–9.6), AG/low (n = 36): 5.3 (95% CI 3.6–8.3), GG/high (n = 16): 11.0 (95% CI 7.1–25.5), GG/low (n = 16): 8.9 (95% CI 5.7–22.0). (B) The median (95% CI) OS (months) values stratified by gemcitabine vs. no chemotherapy and rs2853564 genotypes (AA vs. AG+GG), are as follows: AA/gemcitabine (n = 81): 7.7 (95% CI 6.4–10.2), AA/no chemotherapy (n = 30): 6.3 (95% CI 4.8–8.7), AG+GG/gemcitabine (n = 149): 10.8 (95% CI 9.6–12.5), AG+GG/no chemotherapy (n = 44): 4.4 (95% CI 3.2–7.9). Tables indicate the number of patients who were alive and at risk of death at each time point after the diagnosis of pancreatic cancer.
Fig 3.
Luciferase assays of VDR rs2853564 and rs7979131 in two human cell lines.
Luciferase assays of reporter gene constructs for rs2853564 (A>G) and rs7979131 (T>G) were performed in human pancreatic carcinoma cells (PANC-1) and telomerase-immortalized human microvascular endothelial (TIME) cells. Luciferase activity was determined as a ratio of Firefly to Renilla luciferase activity, normalized to the reference sequence construct (reported as wild-type, WT, in the graph).
Fig 4.
Analysis of the association between VDR rs2853564 and VDR mRNA expression in A) pancreatic adenocarcinomas (n = 66) and B) CCLE pancreatic cancer cell lines (n = 44). A) Genotypes for rs2853564 were determined by a TaqMan SNP genotyping assay. Gene expression data for VDR was derived from Agilent gene expression microarrays normalized for intra-array dye bias. Association between genotype and VDR mRNA expression was assessed using an additive genetic model in a linear regression. The removal of three low samples (2 AA and 1 AG) did not affect the result (p>0.05). B) rs2853559 instead of rs2853564 has been used in this analysis as rs2853559 is in almost complete LD (r2 = 0.9652) with rs2853564. In the figure, we report the genotypes of rs2853564 and their association with VDR expression. Genotype and VDR expression data were obtained from the CCLE, as described in the Supplementary Methods. VDR expression data were normalized using a robust multi-array average. Association between genotype and VDR mRNA expression was assessed using an additive genetic model in a linear regression.