Clinical, immunogenetic, and molecular evidence of HLA loss at relapse after MUD HSCT. (A) Clinical course of the patient: histogram bars represent the percentage of leukemic blasts in the bone marrow of the patient at different time points during the posttransplantation follow-up. Boxed in gray is the disease relapse that occurred after the second MUD HSCT, when leukemic blasts were purified by fluorescence-activated cell sorting for further molecular analyses. (B) Genomic HLA typing of the patient before HSCT (and of AML blasts at diagnosis), of the 2 MUDs (donors 1 and 2), of the leukemic blasts purified at relapse after the second MUD HSCT (month 54, boxed in gray in panel A), and of the patient's mother (donor 3). HLA alleles mismatched between donors and patient are shown in bold italics. Homozygosity for HLA-DPB1*04:01 for donor 2 was inferred from univocal genomic typing. (C) Single nucleotide polymorphism (SNP) profile of chromosome 6 from purified AML blasts harvested at diagnosis (top dot plots) and at relapse after the second MUD HSCT (bottom dot plots), analyzed using the Illumina Human660W-Quad BeadChip. Top and bottom plots show the B allele frequency and the LogR ratio, indicating zygosity and gene copy numbers of each SNP, respectively. Note de novo acquired UPD of leukemic blasts at relapse in a 40-Mb region of chromosome 6p encompassing HLA (evidenced in gray).