Fanconi anemia gene variants in therapy-related myeloid neoplasms

Therapy-related myeloid neoplasms (t-MN) include myelodysplastic syndromes (MDS) and acute myeloid leukemias (AML) occurring as a late effect of chemotherapy and/or radiotherapy for a primary malignancy or for autoimmune diseases. 1 – 2 The incidence of this complication has been raising in the past years because of the prolonged survival and the higher number of treated patients. Still, o 5% of patients exposed to cytotoxic drugs and radiotherapy develop a t-MN, suggesting an underlying individual susceptibility. Primary malignancies most frequently associated to t-MN are breast cancer and lymphoproliferative diseases. Other recurrent clinical characteristics are presence of multiple primary neoplasms in the same individual and cancer familiarity. 2 So far, higher frequency of single-nucleotide variants of detoxi ﬁ cation and DNA-repair enzymes, alone or in association, have been reported in t-MN, but none has been validated as signi ﬁ cant risk factor in large patient groups. 3 – 6 The association of breast and other


Fanconi anemia gene variants in therapy-related myeloid neoplasms
Blood Cancer Journal (2015) 5, e323; doi:10.1038/bcj.2015.44; published online 3 July 2015 Therapy-related myeloid neoplasms (t-MN) include myelodysplastic syndromes (MDS) and acute myeloid leukemias (AML) occurring as a late effect of chemotherapy and/or radiotherapy for a primary malignancy or for autoimmune diseases. [1][2] The incidence of this complication has been raising in the past years because of the prolonged survival and the higher number of treated patients. Still, o 5% of patients exposed to cytotoxic drugs and radiotherapy develop a t-MN, suggesting an underlying individual susceptibility. Primary malignancies most frequently associated to t-MN are breast cancer and lymphoproliferative diseases. Other recurrent clinical characteristics are presence of multiple primary neoplasms in the same individual and cancer familiarity. 2 So far, higher frequency of single-nucleotide variants of detoxification and DNA-repair enzymes, alone or in association, have been reported in t-MN, but none has been validated as significant risk factor in large patient groups. [3][4][5][6] The association of breast and other cancers to myeloid neoplasms is frequent in Fanconi anemia (FA), a childhood syndrome characterized by chromosomal instability, developmental abnormalities, aplastic anemia and by predisposition to cancer, particularly gynecological, head-neck and gastrointestinal. [7][8][9] FA is associated to the occurrence of bi-allelic loss-of-function mutations in the family of FANC genes, comprising 16 DNA-repair genes. Mutations in FANCA, FANCC and FANCG are the most common, and account for~85% of the FA patients. Less than 5% of FA patients do not appear to have mutations in these 16 genes, indicating that additional genes may be involved. Carriers of germline homozygous mutations of at least five of the sixteen FA genes, including FANCD1/BRCA2, FANCJ/BRIP1, FANCN/PALB2, FANCO/RAD51C and FANCQ/ERCC4, are at a higher-risk to develop cancer.
So far, mutations in FA genes have been rarely described in hematological malignancies outside the syndromic picture of FA. [10][11][12] We were interested in the prevalence of FA variants in t-MN following cytotoxic treatment for breast cancer and lymphoproliferative diseases, compared with de novo acute myeloid leukemia (dnAML).
We studied 37 patients with a t-MN, diagnosed at our Department of Hematology between November 1995 and January 2012. Table 1 shows patient characteristics. Median age was 63 years (range 30-78). According to the proportion of blasts, there were 19 t-MDS and 18 t-AML. Cytogenetic analysis of bone marrow mononuclear cells obtained at t-MN diagnosis revealed clonal aberrations in 24 out of 32 available samples (75%). The primary malignancy was Hodgkin lymphoma in 7 patients, non-Hodgkin lymphoma in 12 patients and breast cancer in 18 patients (associated with a second malignancy in five patients). As a control, 24 dnAML patients, of a median age of 61 years (range 40-78 years), were studied at the Istituto Seragnoli of the University of Bologna. Patients gave informed consent to the study and the protocol received approval from the local Ethical Committees.
Bone marrow mononuclear cells were separated at the time of initial t-MN diagnosis using Ficoll Gradient centrifugation (Cedarlane, ON, Canada), and DNA was extracted by using the QIAamp DNA Mini Kit (Qiagen AG, Hilden, Germany). DNA extracted from different tissues of the same individuals served as germline controls and included: lymphnode biopsies or breast specimens (neoplastic and non-neoplastic) collected at the time of the primary cancer diagnosis, buccal brush cytology or hair follicles collected at the time of t-MN or dnAML diagnosis, peripheral blood or bone marrow samples collected during follow-up of the primary malignancy or at the time of complete remission.
We   Table 1). Reagents and conditions were as recommended by the manufacturers (PyroMark Q96 ID, Diatech Pharmacogenetics, Jesi, Italy). A BLAST search on NCBI, Cosmic and UCSC databases was performed for all identified variants. The Polyphen-2 tool (genetics.bwh.harvard. edu) was consulted to define the putative effect of FA discovered variants, predicting the possible impact of the amino acid substitutions on the structure and function of the proteins. Odds ratios, with 95% confidence interval were also calculated.
FA gene variants were frequent in our t-MN patients, with 6 out of 37 patients (16%) carriers of at least one genomic variant, with similar prevalence in t-MN secondary to lymphoproliferative diseases versus solid tumors (4 out of 19 vs 2 out of 18, respectively). We found seven heterozygous FANC variants, including two FANCA (L6F and S90T), three FANCD2 (T1376A, P256S and M1023V), one FANCJ (I364V) and one FANCC (L36F) ( Table 2). There were no differences in the time between primary cytotoxic treatment and t-MN diagnosis in carriers of FA variant versus wild-type patients (median, 72 vs 69 months). Six variants were novel, according to the NCBI and Cosmic databases, whereas the FANCA L6F had been previously described in the 1000 Genome project (www.1000genomes.org) and reported in the UCSC Genome Browser in the transcript variant 2 of FANCA mRNA (uc002fow.1). This same variation has been conflictually reported as SNV in the 5ʹ UTR of NCBI Reference Assembly (rs189841793).
To determine whether FANC genetic variants were somatically acquired in t-MN patients, germline material was screened in five patients for whom control tissues were available. The FANC mutation was germline in all cases (5 confirmed in 37 t-MN patients, 13.5%) ( Table 2). As a control, the frequency of FANC variant in t-MN was compared with that of 24 dnAML patients, studied using whole-exome-sequencing. We found that 3 out of 24 patients (12.5%) were carriers of a germline FA variant (1 FANCA, 1 FANCL and 1 FANCJ). This translates into a similar prevalence of FA SNV in t-MN and dnAML. Since SNV in FA genes seem frequent both in dnAML and in t-MN, we suggest that these variants may have a role as risk factors in the development of myeloid neoplasms.
As all FANC variants were novel, further studies will be necessary to clarify the functional impact of these variants on the gene product and DNA damage repair. According to the Polyphen-2 database (genetics.bwh.harvard.edu) five out of eight FA variants were defined probably or possibly damaging DNA repair (Supplementary Table 2). Functional studies using constructs including the mutated sequences may help to answer this question. 13 Our data may indicate that heterozygous carriers of FA variants may have increased susceptibility to environmental carcinogens and to the DNA-damaging action of cytotoxic therapy used to treat primary tumors, leading to de novo or secondary leukemogenesis. In this line, a trend towards higher DEB-induced chromosome breakage and an increased release of fragmented DNA after exposure to X-irradiation have been reported in FA heterozygotes relative to controls. 14 We have previously reported that t-MN secondary to lymphoproliferative diseases were characterized by a very low frequency or absence of epigenetic and spliceosome gene mutations, compared to t-MN following solid tumors. 15 As the frequency of FA germline variants in this study did not appear to be different between lymphoproliferative neoplasms and breast cancer and between de novo and t-MN, the probable functional role of these variants in carcinogenesis does not appear to be restricted to a specific type of primary cancer or its treatment.
In the near future, we might have to revise our concept of genetic susceptibility to AML, taking into account not only recurrent genetic variants, such as polymorphisms, but also rare individual germline variants in critical genes.