Complete Loss of HLA Class I Heterozygosity in a Patient with Acute Myeloid Leukemia

Loss of heterozygosity (LOH) for certain HLA class I alleles provides tumor cells with an efficient mechanism to escape the T cell immune response. Here we report a case of complete class I LOH in a patient suffering from acute myeloid leukemia, which was associated with an intriguing HLA-typing procedure. The patient was diagnosed with AML FAB M1 and further analysis revealed a nucleophosmin1 (NPM1) mutation and normal male karyotype 46XY.

In February 2013 two independently collected EDTA blood samples (S1 + S2) arrived at our lab for HLA testing. Primary HLA class I and II typing was done for S1 with Sanger sequence based typing (SBT) and Luminex^TM sequence-specific oligonucleotide primed PCR (PCR-SSO). Both methods showed a homozygous phenotype for HLA class I loci (A, B, C) and heterozygosity for HLA class II alleles (DRB1,DQB1). After consolidation, increasing minimal residual disease (MRD) measured by a close NPM1, monitoring was observed in September 2013 and the patient was scheduled for an allogeneic stem cell transplantation. No blasts were present in the peripheral blood of the patient at the time point of molecular relapse. To initiate search for an unrelated donor a confirmatory patient HLA-typing was performed. Due to poor S2 DNA quality a new blood sample (S3) was ordered and used for secondary typing. Whereas HLA class II results were consistent, SBT and PCR-SSO showed discrepant class I typing results when compared to S1, with heterozygosity for all three loci. Consultation of the transplantation centre revealed that S1 + S2 were taken while the patient had 84% blasts in the peripheral blood. Heterozygous patient HLA status was finally confirmed by typing of a saliva sample (S4) with SBT and SSO-PCR, respectively. According to the information about S1 (primary diagnosis, peripheral blasts), S3 (increasing MRD, no peripheral blasts) and S4 (blast free material) we assumed that in S1 cells with somatic HLA class I phenotype were highly outnumbered by blasts bearing a complete class I LOH.

In regard to this case we strongly recommend confirmatory typing to be carried out on blast free samples only, or by using a highly sensitive PCR-SSP in case of a substantial amount of blasts in the blood. This should help to prevent the otherwise potentially fatal selection of a highly mismatched donor. As a consequence we established a routinely feedback on the peripheral blast status of the sample on our test request form.