SNP Array Analysis In Unrelated HLA-Matched Transplanted Patients Reveals Loss of Patient- Specific Allele at the Mismatched Locus

Abstract 4692

Allogeneic hematopoietic cell transplantation (HCT) is the most effective curative therapy in patients acute myeloid leukemia (AML) and myelodisplastic syndromes (MDS). Despite improvements in the clinical management, relapse from the original disease is still one of the main causes of treatment failure. There is scarce available data on the influence of genomic changes in relapsed patients after HCT, in particular in HLA identical transplantation. Selective pressure mediated by T cells can result in genomic changes in the HLA region on chromosome 6, as shown by several animal models and in human haploidentical transplantation. In this study we examined genome wide DNA copy number changes and long contiguous stretches of homozygosity (LCSH) in 21 patients with a diagnosis of AML or MDS/AML undergoing HCT.

SNP-chip analysis revealed a total of 162 genomic aberrations (GA) in 21 pairs of samples before allogeneic transplantation and at relapse in patients diagnosed with AML or AML/MDS. Total GA were significantly higher after transplant compared with the pre transplant samples (p<0.05) only in the group of patients with normal cytogenetics. Within this last group of patients, only a small deletion was detected in the pre transplant sample not detected by conventional cytogenetics. In contrast, in the group of patients with abnormal/complex karyotype, 5 additional deletions and 2 duplications were detected in 5 patients before transplantation. The median number of GA per patient pre transplant was 3 (range 0–19), while the median number of GA at relapse post transplantation was 5 (range 1–20).

The median copy number loss per patient was 1.6 before transplant and 2.8 at relapse, while the median number of LCSH at the same time points was 2 and 2.8 respectively. Copy number gains were significantly less common (p<0.05) with a median of 0.5 per patient before transplant and 1 at relapse. Of 21 patients, eleven (52%) shared the same GA before transplant and at relapse indicating common clonality. Conversely 26 novel deletions, 9 LCSH and 6 duplications arose at relapse detected in 17 patients (76%). Of interest 3 out of 9 of these new LCSH involving chromosomes 6p, 20q and 22q, were recurrent GA at relapse detected in 2, 5 and 4 patients respectively. HLA typing of the blasts from patients with LCSH in chromosome 6p revealed a loss of the patient specific band at the mismatched locus leading to homozygosity for the HLA haplotype shared by the patient and the donor. LCSH was the most frequent lesion present in 81% of the patients followed by deletions (57%), trisomies (48%) and duplications (24%). All chromosomes had one or more GA, with significant differences among the 22 autosomes. Chromosome 2 had only 2 GA in the post transplant sample while chromosome 17 had 9 in the same time point. The most frequent lesion was LCSH in chromosome 20 q11.21 in the relapse sample, occurring in 5 of 21 patients (24%). Another recurrent GA was a deletion in chromosome 12 p13 detected in 4 patients (19%) at relapse. To confirm the SNP-chip results detected in allogeneic transplanted patients, extensive validation analysis was performed. FISH signal for probes TP53 (17q13.1), ETV6 (12p13) and ATM (11q 22.3) revealed one signal. These regions also showed hemizygous deletion by SNP-chips analysis. D8Z2 probe revealed three signals in patients with trisomy 8 detected with the array platform. Two patients with LCSH in the short arm of chromosome 9 p24 were homozygous for JAK2 mutation as detected by pyrosequencing. STR analysis using the polymorphic marker YNZ 22 confirm the allelic loss of at least one allele in the 17 p13.3 region. Results of SNP, FISH, pyrosequencing and STR analysis were completely congruent. These validation results suggest that SNP analysis reflected genomic changes. Patients with late onset GvHD (> day 100 post transplantation) had a significant higher number of GA after transplantation (p<0.05). No significant association was found between acquired GA and the rest of the clinical variables analyzed including outcome. Our results indicate that GA in patients undergoing HCT are likely to play a role. LCSH in chromosome 6p can result in immune escape by leukemic cells from graft-versus-leukemia effect that can lead to relapse. In addition homozygosity in mutated genes such as FLT3 (13 q12) and JAK2 (9 p24) due to deletion or LCSH has important implication for selecting a treatment in patients that relapse after HCT.