Abstract
Introduction
Vitamin D deficiency is common in the general population. Approximately 25-50% of adult patients seen at routine visits in the United States are found to have an insufficient vitamin D level. Vitamin D has been shown to be prognostic in several types of cancers including breast, prostate and colon cancer. Vitamin D activates a nuclear transcription factor that regulates the expression of almost 200 genes which modulate a variety of cellular processes including angiogenesis, differentiation, proliferation, and apoptosis. Recent research has shown that vitamin D levels may have a prognostic effect in patients with chronic lymphocytic leukemia (CLL), where 25-OH vitamin D insufficiency was associated with shortened time to treatment and poorer overall survival. A centrally important unanswered question relates to causation: does vitamin D insufficiency yield more aggressive cancer disease biology, or do intrinsically progressive cases of CLL cause vitamin D insufficiency? We hypothesized that vitamin D insufficiency alters CLL cell biology and favors a more aggressive disease phenotype.
Methods
Untreated patients within 12 months of initial diagnosis of CLL from Duke University Hospital and the Durham VA were studied. Serum samples from 185 patients were assayed for the 25-OH vitamin D level (immunochemiluminometric assay). A multivariate analysis was performed using: age, race, gender, Rai stage, CD38, Zap70, hierarchical FISH, IGHV, and season of diagnosis to determine whether vitamin D levels are a significant predictor of OS and TTT in this group. Global mRNA expression from 23 patients was analyzed using Affymetrix U133 Plus 2.0 arrays as a function of vitamin D level and gene list generated for those with p values < 10-5. rtPCR was performed on samples from an additional 50 patients to validate the findings from the mRNA expression analysis. Linear regression analysis was conducted to evaluate for significant associations between genes and 25-OH vitamin D levels. An in vitro assessment of 1,25-di-OH vitamin D effects on CLL cell viability in serum free media was evaluated using an MTS assay.
Results
The mean vitamin D level amongst the group of 185 patients was 25.6± 9.7 ng/mL. Eighty-nine patients had a vitamin D level less than 25 and 96 had a level above 25, which we used as our cutoff, as prior reports have used this level to define insufficiency in CLL. Thirty-one of 95 (33%) of the sufficient vitamin D group were treated versus 39 of 89 (44%) of the insufficient vitamin D group (p=0.12). Among those requiring treatment, the mean TTT was approximately the same between the two groups: 4.7±0.3 yrs for the higher vitamin D group vs. 4.6±0.4 yrs for the insufficient group (p=0.126). OS for the higher vitamin D group was 8.3±0.3 vs. 7.0±0.2 years for the lower vitamin D group (p=0.935). Multivariable analysis showed that IGHV mutation (HR = 0.386; p=0.0159) and Rai stage 0 or 1 (HR = 0.174; p=0.0002) predicted TTT, while age and race influenced OS, with age>62 conferring greater risk of death (p=0.0191) and African Americans having decreased survival (p=0.0110). Preliminary studies of gene expression data identified eight probes that were differentially expressed as a function of vitamin D level. rtPCR was then performed on GPR82, MPZL3, FBXW4, ROR1, and CXCL11 to validate these results. Linear regression confirmed that ROR1 and FBXW4 gene expression correlated with vitamin D level (p=0.0065; r2=0.144 and p=0.0185; r2=0.110, respectively). High levels of ROR1 are observed in B-CLL. FBXW4 has been shown to be mutated or under-expressed in a variety of human cancer cell lines. Early in vitro cytotoxicity of 1,25 di-OH vitamin D in CLL (n = 5 patient derived samples) showed an IC50 = 334 nM.
Discussion
Our results show that the basal level of vitamin D is not significantly correlated with either OS or TTT in CLL in contrast to previous studies. No interaction between vitamin D levels and race, age, gender, Rai stage, IGHV mutation, season of diagnosis was observed. However, ongoing in vitro experiments show that 1,25 vitamin D is cytotoxic to CLL, raising the intriguing possibility that intermittent bolus dosing could potentially be used therapeutically. Further, we have identified specific genes where quantitative gene expression is correlated with basal vitamin D levels. These findings expand our understanding of the interaction between vitamin D and B cell malignancies.
Disclosures:
No relevant conflicts of interest to declare.
Vitamin D insufficiency and prognosis in chronic lymphocytic leukemia