A Novel Multi-Analytic Clinical Diagnostic Assay to Detect HLA Loss of Heterozygosity in Post-Hematopoietic Cell Transplant Relapse Patients

Abstract

Genomic loss of heterozygosity in the HLA region (HLA LOH) can be an underlying driver of disease relapse in haploidentical and HLA mismatched hematopoietic cell transplantation (HCT) in which relapsing cells evade donor T-cell allorecognition by deleting non-shared HLA alleles through an acquired somatic uniparental disomy event. The presence of HLA LOH may indicate a loss of graft vs. leukemia effect, render treatments such as donor leukocyte infusion ineffective, and impact donor selection for subsequent allogeneic HCT. A major challenge in developing a clinical diagnostic assay to detect HLA LOH post-HCT is sensitive detection of a loss of genomic material against a background of multiple genomes. Cell enrichment may address this issue by isolating relapsed cells for downstream molecular analysis. However, detection of HLA LOH is most clinically useful early in relapse when the relapsed cells are likely in low abundance, the relapsed cells may not be a single clonal population, and the immunophenotypic aberrancies may be variable and/or unknown. To address these issues, we have developed a comprehensive multi-analytic approach to detect HLA LOH that includes sensitive flow cytometric enrichment of leukemic cells followed by genomic analysis. First, a "difference from normal” flow cytometry approach (ΔN:™), the approach used for determination of residual disease in Children's Oncology Group AML clinical trials, is utilized to characterize the aberrant cells for each patient and relapse episode to ensure specific enrichment of relapsed cells to high purity. Next, two complementary molecular assays are used, including detection of short tandem repeat (STR) loci throughout the HLA region and next-generation sequencing (NGS) of HLA loci to identify deleted HLA alleles. In combination, the two assays provide a sensitive and spatially resolved assessment of LOH across the entire HLA genomic region, enabling molecular characterization of LOH events ranging from an entire HLA haplotype, partial HLA haplotype, or a single HLA locus. The ability to detect a locus-specific HLA LOH event could be useful in mismatched unrelated HCT donor/recipient pairs which may differ at only one or two HLA loci. Validation of the flow cytometric methods and genomic assays were performed separately. DNA isolated from the flow sorting process was shown to be suitable for both genomic assays. The sensitivity of the molecular assays to detect HLA LOH was defined using artificial mixes of DNA from two individuals with allelic mismatching at various HLA loci. Cell lines with known HLA LOH were also used to confirm the performance of the molecular assays. As of this writing, 16 samples from post-HCT relapse patients were tested with this assay and HLA LOH was detected in 3 cases (19%), consistent with published reports on the frequency of HLA LOH detected in allogeneic HCT patients. Samples with a range of relapsed leukemic cells from <1% to 85% and cell sorting output from 800 to 250,000 cells all yielded valid results. Results were reported within an average of 7 days, allowing adequate time for integration into therapeutic decision making. Challenging and unusual patient samples such as those with extremely low blast counts will be presented. Initial results indicate this clinical diagnostic assay provides reliable and accurate detection of HLA LOH during relapse in post-HCT patients. Detection of HLA LOH using our multi-analytic diagnostic approach may also aid in optimizing therapeutic decisions to improve outcomes in allogeneic HCT patients.