No evidence for association between TGFB1 promoter SNPs and the risk of childhood pre-B acute lymphoblastic leukemia among French Canadians

TGFB1 is a potent growth inhibitor to a wide variety of cell types including hematopoietic cells and deregulation of the TGFB1 signaling pathway has been implicated in the development of several cancers.1 Functional regulatory polymorphisms of the TGFB1 gene have been directly associated with inter-individual variability in TGFB1 plasma levels and modified risk of breast,2 lung,3 colorectal,4 and prostate5 cancers. No studies, to date, have examined the association between genetic polymorphisms in TGFB1 and childhood leukemia. In this study we examined the role of TGFB1 promoter SNPs (pSNPs), -1886G>A, -1571A>G, -1550DEL/AGG, and -509C>T, as genetic modulators of childhood pre-B ALL susceptibility among the French Canadian population.

In a case-control study, we investigated genotypic, haplotypic, as well as multi-SNP combination associations with childhood pre-B ALL. The study population and inclusion criteria were described previously.6 Childhood pre-B ALL cases (n=321) consisted of 189 boys and 132 girls with a median age of 4.7 years. Parental DNA was available for 203 of these children. Healthy controls (n=329) were recruited at the Sainte-Justine Hospital. Study individuals were all French Canadian from the province of Quebec, Canada. Candidate pSNPs were previously identified7 and were selected based on their frequency in the European population. A PCR-based allele-specific oligonucleotide hybridization approach was used to genotype samples, described as previously8.

Hardy-Weinberg equilibrium was tested using the χ goodness of fit test and PedCheck (Version 1.1) was used to identify genotype incompatibilities using the familial data.9 Pearson’s χ test or Fisher’s exact test, as appropriate, was used to compare allele/genotype/haplotype carriership in patients and controls. Crude odds ratios (ORs) were measured using logistic regression in STATA (Release 9.2) and are given with 95% confidence intervals (CIs). Haplotype frequencies were estimated using the FAMHAP Software (Version 16) using parental data when available.10 A likelihood ratio test implemented in FAMHAP was used to examine global haplotype associations with disease status. Multimarker combinations were tested for association with disease using the method proposed by Becker and Knapp implemented in FAMHAP.11 Multiple testing was controlled for using the false discovery rate12 with a type I error rate of 10%.

Frequencies of pSNPs were in agreement with those previously reported in other populations of European descent and all distributions were in Hardy-Weinberg equilibrium. TGFB1 pSNP distributions did not differ significantly between cases and controls (Table 1). Suggestive evidence of a reduced risk of childhood pre-B ALL was demonstrated when carriers of at least one variant -1886A allele were compared to homozygous individuals for the ancestral -1886G allele (OR=0.58, 95% CI: 0.33–1.01, p=0.042). However these results did not sustain multiple testing corrections. A total of 12 promoter haplotypes (pHaps) were inferred but only haplotypes TGH*1, *2 and *3 had frequencies ≥ 0.05 and represented 98% of the observed haplotypes in the population tested (Table 2). The remaining 2% of the chromosomes carried nine minor haplotypes that were grouped under TGH*. We found no significant difference in the overall distribution of the 12 TGFB1-derived pHaps between cases and controls (Global c2= 10.85, 11 degrees of freedom, p=0.46) and found no evidence of association between individual haplotypes and the risk of pre-B ALL. Investigating multi-SNP combinations, the best result was obtained for marker combination -1886A/-1550AGG/-509C, which was associated with a decreased risk of pre-B ALL (OR= 0.86) however the global test statistic failed to reach significance (p=0.15).

Table 1.Allele and genotype frequencies of pSNPs in TGFB1 in childhood pre-B ALL patients and controls.

Though the expected variability of TGFB1 expression levels due to promoter SNPs could indeed contribute to leukemogenesis, our data did not support a role for TGFB1 promoter variants -1886G>A, -1571A>G, -1550DEL/AGG or -509C>T in the etiology of childhood pre-B ALL, at least among the French Canadian population. Having screened 2kb upstream of the transcription start site in 40 individuals,7 a strength of this study was our ability to capture and survey most common genetic variation within the region under investigation with reasonable statistical power. Despite the relatively small size of this dataset, we had 80% power at the 5% level to detect a minimum OR of 1.5 with a minor allele frequency ≥ 15%. However it remains possible that other variants within the TGFB1 gene and surrounding regulatory sequences that were not assessed within the scope of this study could modify disease susceptibility. Additional analyses in independent datasets are required to further support the lack of association between these TGFB1 variants and childhood pre-B ALL since our study did have limited statistical power to detect associations involving weak effects or rarer variants.

Table 2.Distribution of TGFB1 promoter haplotypes in pre-B ALL patients and controls.

Footnotes

• Funding: this study was supported by research funds provided by the Leukemia and Lymphoma Society of Canada, as well as Genome Quebec and Genome Canada. JH is the recipient of a NSERC Canada Graduate’s scholarship. DS holds the François-Karl Viau Chair in Pediatric Oncogenomics and is a scholar of the Fonds de la Recherche en Santé du Québec.