Microsatellite Instability Assessment by Immunohistochemistry in Acute Myeloid Leukemia: A Reappraisal and Review of the Literature

Background: Microsatellite instability (MSI) is caused by defects in DNA mismatch repair (MMR) components. Inactivation of any MMR gene(s), including hMLH1, hMSH2, hMSH6, and hPMS2, can result in MSI. Immunohistochemistry (IHC) is a sensitive and specific screening tool for MSI that can detect loss of expression of one or more MMR components. Of the four MMR markers, hMLH1 and hMSH2 are considered most informative of MSI status. There has been renewed interest in MSI status in view of its favorable association with response to immune checkpoint inhibitors in some cancers. MMR expression patterns in acute myeloid leukemia (AML) have not been evaluated systematically. Methods: We used clinically-validated IHC assays to assess the expression of hMLH1, hMSH2, hMSH6, and/or hPMS2 in formalin-fixed paraffin-embedded tissue sections of bone marrow core biopsies from patients diagnosed with AML. Mutation profiling was performed using next-generation sequencing to assess for mutations in MMR genes. Results: The study group included 236 patients with AML, including a cohort treated on a clinical trial of azacitidine and nivolumab (NCT02397720). In addition, hMSH6, and/or hPMS2 expression was assessed in 99 AML patients with diploid karyotype. All patients, except two, had retained expression of all MMR markers assessed: One patient from the azacytidine+nivolumab group had zonal patchy loss of staining of hMLH1 and, to a lesser extent, a similar staining pattern of hMSH2; and one patient from the AML with diploid karyotype group had loss of hMSH2 but retained expression of hMLH1, hMSH6 and hPMS2. In addition, a retrospective analysis on a separate cohort of 139 patients with primary AML, on which next generation sequencing profiling was performed, identified 14 cases with alterations in MMR genes. Conclusion and remarks: MMR loss is a rare event in AML, thus does not appear to underlie response patterns to anti-PD1 therapy.