Abstract 2438

Therapy-related leukemia (t-MDS/AML) is a leading cause of non-relapse mortality in patients treated for cancers, such as Hodgkin lymphoma (HL) or non-Hodgkin lymphoma (NHL). Although the association between therapeutic exposures (alkylators, topoisomerase II inhibitors) and t-MDS/AML is clearly defined, inter-individual variability does exist, suggesting the role for genetic factors. Understanding the full spectrum of genetic susceptibility to t-MDS/AML should help identify those at highest risk, setting the stage for targeted surveillance and/or pharmacological intervention. Using a matched case-control design (clinical characteristics – Table 1), we previously identified an association between t-MDS/AML and GSTM1 deletion, demonstrated as differences in both genotype frequency and gene expression level, suggesting that drug-metabolizing enzymes contribute to individual susceptibility. In general, genotyping study analyzes germ line variant alleles that potentially affect function or production level of genes; while expression analysis investigates differences in transcription level resulting from germ line or somatic variations caused by genetic or epigenetic changes, or treatment exposures. In the previous study, after adjusting for treatment exposures, we had detected association of t-MDS/AML with GSTM1 deletion (P = 0.057), and significantly lower expression of GSTM1 (measured by 2 probesets on Affymetrix HG U133 Plus 2 Array) in cases compared with controls (p = 0.0058). Interestingly, several cases with at least one intact allele of GSTM1 expressed low levels of GSTM1; levels that were comparable to individuals who were homozygous for the null allele. Therefore we hypothesized that additional loci control expression levels of this gene and may also predict susceptibility to t-MDS/AML. These expression quantitative trait loci (eQTL) represent genomic regions for the control of quantitative variation in gene expression, typically through modulation of activity of cis-regulatory elements. To test this hypothesis, we identified one SNP, rs11101992, as a genetic marker to a cis-acting eQTL for GSTM1 in lymphoid cell lines (p = 1.6 × 10−10) based on data from a previous study (Dixon et al. Nat Genet 39, 1202–1207) and confirmed its effect in CD34+ peripheral blood stem cells from 32 subjects of Caucasian ancestry (p = 0.078). We then genotyped this SNP in the 49 pairs of matched cases/controls and identified a much stronger association of t-MDS/AML with this SNP (P = 0.0026). The association persisted after adjusting for GSTM1 deletion (P = 0.0076). This SNP was also correlated with expression level of other GSTM gene family members located within the same chromosomal region in lymphoid cell lines (p = 3.4 × 10−10 with GSTM2, p = 3.8 × 10−7 with GSTM3, and p = 4.1 × 10−7 with GSTM4), possibly due to cross hybridization among probe sets or co-expression control of these homologous genes. Our results suggest that additional germ line variation other than GSTM1 deletion regulates expression level of this or other members of the GSTM family and contributes to the risk of t-MDS/AML. We are currently evaluating genetic effect of individual GSTM genes to t-MDS/AML. To our knowledge, this is the first report of the role played by eQTL in defining the genetic susceptibility of t-MDS/AML and provides a novel strategy to uncover new risk loci for this lethal complication resulting from cancer treatment.

Table 1:

Clinical Characteristics

CharacteristicsGenotyping StudyExpressionStudy
Cases (n=49)Controls (n=49)Cases (n=12)Controls (n=22)
Primary Diagnosis (%)  
    HL 29 24 
    NHL 71 76 100 100  
Age at Primary Diagnosis (Median, Range) 49.1 (13–73) 47.9 (26–75) 48.9 (26–67) 48.2 (26–64) 
Race/ethnicity (%)  
    Caucasians 77.6 81.6 58.3 86.4 
    Hispanics 18.4 14.3 33.3 9.1 
    African-Americans 2.0 2.0 8.3 4.6 
    Asians/others 2.0 2.0 0.0 0.0 
Gender (%)  
    Male 71 59 50 77 
    Female 29 41 50 23 
Type of Treatment (%)  
    Autologous HCT 71 45 100 100 
    Conventional 29 55 
Latency or follow-up from Primary Diagnosis (Year, Median, Range) 4.9 (0.6–29) 8.1 (2.4–30) 3.6 (1.2–9.2) 8.2 (4.8–21) 
Latency or Follow-up from aHCT (Year, Median, Range) 3.1 (0.4–17) 3.9 (0.6–10) 2.8 (0.5–3.7) 7.2 (0.5–8.4) 
Cytogenetic Abnormality (%)  
    5q- or 7q- 57  33  
    11q23 6.5 NA 25 NA 
    Others 30  25  
    Normal 6.5  17  
Treatment Exposure  
    Radiation (%) 69 65 83 68 
    Alkylating Agent Score 3 (1–6) 3 (0–6) 2 (1–4) 3 (1–5) 
    Topoisomerase II Inhibitor Score 3 (0–4) 3 (0–4) 3 (2–4) 3 (2–4) 
CharacteristicsGenotyping StudyExpressionStudy
Cases (n=49)Controls (n=49)Cases (n=12)Controls (n=22)
Primary Diagnosis (%)  
    HL 29 24 
    NHL 71 76 100 100  
Age at Primary Diagnosis (Median, Range) 49.1 (13–73) 47.9 (26–75) 48.9 (26–67) 48.2 (26–64) 
Race/ethnicity (%)  
    Caucasians 77.6 81.6 58.3 86.4 
    Hispanics 18.4 14.3 33.3 9.1 
    African-Americans 2.0 2.0 8.3 4.6 
    Asians/others 2.0 2.0 0.0 0.0 
Gender (%)  
    Male 71 59 50 77 
    Female 29 41 50 23 
Type of Treatment (%)  
    Autologous HCT 71 45 100 100 
    Conventional 29 55 
Latency or follow-up from Primary Diagnosis (Year, Median, Range) 4.9 (0.6–29) 8.1 (2.4–30) 3.6 (1.2–9.2) 8.2 (4.8–21) 
Latency or Follow-up from aHCT (Year, Median, Range) 3.1 (0.4–17) 3.9 (0.6–10) 2.8 (0.5–3.7) 7.2 (0.5–8.4) 
Cytogenetic Abnormality (%)  
    5q- or 7q- 57  33  
    11q23 6.5 NA 25 NA 
    Others 30  25  
    Normal 6.5  17  
Treatment Exposure  
    Radiation (%) 69 65 83 68 
    Alkylating Agent Score 3 (1–6) 3 (0–6) 2 (1–4) 3 (1–5) 
    Topoisomerase II Inhibitor Score 3 (0–4) 3 (0–4) 3 (2–4) 3 (2–4) 

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution