Common genetic variants are associated with accelerated bone mineral density loss after hematopoietic cell transplantation

Abstract

Background: Bone mineral density (BMD) loss commonly occurs after hematopoietic cell transplantation (HCT). Hypothesizing that genetic variants may influence post-HCT BMD loss, we conducted a prospective study to examine the associations of single nucleotide polymorphisms (SNP) in bone metabolism pathways and acute BMD loss after HCT.

Methods and findings: We genotyped 122 SNPs in 45 genes in bone metabolism pathways among 121 autologous and allogeneic HCT patients. BMD changes from pre-HCT to day +100 post-HCT were analyzed in relation to these SNPs in linear regression models. After controlling for clinical risk factors, we identified 16 SNPs associated with spinal or femoral BMD loss following HCT, three of which have been previously implicated in genome-wide association studies of bone phenotypes, including rs2075555 in COL1A1, rs9594738 in RANKL, and rs4870044 in ESR1. When multiple SNPs were considered simultaneously, they explained 5-35% of the variance in post-HCT BMD loss. There was a significant trend between the number of risk alleles and the magnitude of BMD loss, with patients carrying the most risk alleles having the greatest loss.

Conclusion: Our data provide the first evidence that common genetic variants play an important role in BMD loss among HCT patients similar to age-related BMD loss in the general population. This infers that the mechanism for post-HCT bone loss is a normal aging process that is accelerated during HCT. A limitation of our study comes from its small patient population; hence future larger studies are warranted to validate our findings.

Figure 1. Annualized rate of bone mineral density change after hematopoietic cell transplantation by genotype for SNPs significant in the multivariable models including clinical risk factors.

Mean and standard error of the annualized rate of BMD change is plotted by genotype for significant SNPs in the multivariable models including clinical risk factors. If the frequency of the homozygous variant genotype was ≤5%, it was combined with the heterozygous genotype (includes rs3787557 [CYP24A1], rs6265 [PTH], rs1801131 [MTHFR] and rs2075555 [COL1A1]). Please refer to Table S2 for a description of functions of these genes and SNPs. Abbreviation: BMD, bone mineral density; HCT, hematopoietic cell transplantation; SNP, single nucleotide polymorphism.

Figure 2. Annualized rate of bone mineral density change after hematopoietic cell transplantation by the number of risk alleles of the SNPs significant in the multiple SNP models.

Mean and standard error of annualized rate of BMD change is plotted by the number of risk alleles in significant SNPs from the multiple SNP models as shown in Table 3. The p-values for a trend between the number of risk alleles and BMD change were ≤0.003 for all three outcomes. The proportion of HCT patients carrying the specified number of risk alleles is shown. The multiple SNP analysis was not performed for spinal BMD change after allogeneic HCT since only one SNP was significant in the multiple SNP model. Please refer to Table S2 for a description of these genes and SNPs. Abbreviation: BMD, bone mineral density; HCT, hematopoietic cell transplantation; SNP.