High Resolution Genomic Mapping of Breast Cancer Derived T-MDS/TAML

Abstract

Therapy-related myelodysplastic syndromes (t-MDS) and therapy-related acute myeloid leukaemia (ICD-O code 9920/3) (t-AML) represent a serious long term complication in patients successfully treated for primary malignancies (PM) using chemotherapy, radiation therapy or a combination of these therapies. Factors that contribute to the incidence of t-MDS/t-AML include,the type and duration of the therapeutic regime used to treat the PM andhost predisposition.Conventional techniques to identify genomic aberrations in t-MDS/t-AML, suffer from low throughput and poor resolution. Recent developments in single-nucleotide polymorphisms (SNP) microarray technology have enabled the exploitation of SNP arrays for high-resolution genome-wide genotyping. We investigated the frequency of chromosomal aberrations and polymorphisms in the GSTM1, GSTT1, GSTP1 and NAT2 enzymes in patients developing t-MDS/t-AML following adjuvant therapy for breast cancer. Using 250K SNP arrays we have analysed two cohorts of similarly treated patients (median age 56 years, range 26–75 years), one group developing t-MDS/t-AML (+t-MDS) and the other remaining disease free (control) within the same follow up time. In all cases constitutive DNA was also simultaneously analysed. We detected multiple chromosomal aberrations including copy number aberrations (CNA) and UPD summarised in Table 1.Table 1AmplificationsDeletionsUPDNo.of pts (%)Median (range) MbNo. of pts (%)Median (range) MbNo. of pts (%)Median (range) Mb+ t-MDS (n=17)10 (48)1.4(0.2–105)13 (62)8.8(0.04–125)9 (43)4.4(2.1–35.4)Control (n=30)6 (20)0.4(0.2–2.5)9 (30)0.3(0.2–0.7)15 (50)3(2–14)There is a distinct difference in CNA between both groups of patients. Interestingly, we identified CNA in the peripheral blood of patients (67%) not developing t-MDS/t- AML which was not observed in any of our normal control samples with no history of chemotherapy. The frequency of both deletions and amplifications was almost twice in + tMDS group in comparison to control samples. Furthermore, the frequency (43% vs. 50%) and median size of UPD (4.4Mb vs. 3Mb) was similar between both groups respectively. SNP array data was also used to test for an association with t-MDS/t-AML. Relative risk (RR) was estimated to determine the effect of each SNP on the risk of t-MDS/t-AML development, with patients not developing t-MDS/t-AML. SNP's were included in the analysis if they physically mapped onto a gene, tagging SNP's and met a significance criteria (p<0.001 and 3<RR<0.33). Our results show tagging SNP's on XRCC4 (p=0.000027, RR=0.020, RUNX1 (p=0.000074, RR=7.8), L3MBTL3 (p=0.0000093, RR=9.6), SUFU (p=0.000087, RR=5.42), ANXA11 (p=0.0002, RR=13.99) and ETFDH (p=0.000865, RR=8.01) to be significantly associated in breast cancer patients developing t-MDS/t- AML. There was no significant difference in the frequency of the GSTM1 homozygous deletion, however, the GSTT1 homozygous deletion was significantly present (p<0.05) in the + t-MDS group. In addition, the double polymorphic deletion (GSTM1 & GSTT1) was significantly present (p<0.01) in patients developing t-MDS/t-AML in comparison to de-novo AML and MDS patients, agreeing with previously published data. There was no significant difference in the incidence of GSTP1 exon 5/6 mutations or NAT2 mutations in both the groups. Our data clearly identifies the utility of using high resolution SNP arrays to identify significant SNP's in genes which may contribute to the pathogenesis of t-MDS/t-AML as well as suggesting that the double deficiency of GSTT1/GSTM1 or single GSTT1 deletion marks a risk for developing haematological malignancies after adjuvant therapy for breast cancer.