The acquired genetic characteristics of acute lymphoblastic leukemia (ALL) blasts are often used to guide the intensity of therapy, whereas the germline host genetic characteristics of the patient generally have not been considered. Multiple common, functionally important polymorphisms affect genes whose products determine the pharmacokinetics and pharmacodynamics of antileukemic agents. It is not yet known how genetic polymorphisms may interact to affect the outcome of antileukemic therapy. Combining classification and regression tree with failure time analysis, we assessed whether 16 genetic polymorphisms, alone or in combination, predicted relapses in 246 children with ALL, 116 of whom were treated on the lower-risk (LR) and130 on the higher-risk (HR) arms of the St Jude protocol Total XIIIB. Genotyping was performed for the following polymorphic loci: CYP3A4*1B and CYP3A5*3; GSTP1 313A>G, GSTM1 and GSTT1 deletions; MDR1 exon 21 (2677G>T/A) and MDR1 exon 26 (3435C>T); MTHFR 677C>T and MTHFR 1298A>C; NR3C1 1088A>G; SLC19A1 80G>A; TPMT 238G>C, 460G>A and 719A>G; TYMS enhancer repeat; UGT1A1 promoter repeat polymorphism; VDR intron 8 G>A and VDR FokI (start-site) T>C. In all children with available RNA in their diagnostic ALL blasts, gene expression levels of prognostic genotypes were analyzed using the Affymetrix genechip array HG_U95Av2. Among the HR group, the glutathione S-transferase M1 (GSTM1) non-null genotype was associated with the risk of hematological relapse (5-year cumulative incidence, 17.1%±4.5% compared to 5.1%±2.9% for GSTM1 null genotype, p = 0.03), and among the non-null genotypes, the thymidylate synthetase (TYMS) 3/3 genotype was associated with a further increase in hematologic relapse risk (5-year cumulative incidence, 29.2%±9.5% compared to 10.9%±4.7% for TYMS 2/3 or 2/2 genotypes, p = 0.02). Increased expression levels of these two target genes (p < 0.0001 and p = 0.09, respectively) were consistent with resistance to the drugs interacting with these gene products. For central nervous system relapse, among the HR group, the vitamin D receptor (VDR) start site (p = 0.02) and intron 8 genotypes (p = 0.04) predisposed, whereas for LR patients the TYMS 3/3 genotype predisposed (p = 0.04). The genotypes associated with outcome have pharmacologic plausibility: e.g., high GST activity (GSTM1 non-null) could cause anticancer drug resistance; high TYMS activity (TYMS 3/3) would be less inhibited by antifolates. In conclusion, germline polymorphisms influence the outcome of antileukemic therapy, and therefore represent determinants of response that can be used to optimize therapy.
Read full abstract