Value of Immunohistochemistry-Based Direct Visualization for Localization, Lineage Determination and Monitoring of IDH1 p.R132H Mutant Clones in AML

Background

Isocitrate dehydrogenase 1 (IDH1) and IDH2 mutations are important prognostic biomarkers in acute myeloid leukemia (AML). Although the clinicopathologic correlates of IDH mutations have been extensively studied, the distribution of abnormal myeloid cells carrying these mutations has not been studied. Specific localization of cells carrying IDH mutations will be useful in further understanding the pathophysiology and post-treatment biology of IDH mutant cases of AML. This characterization is becoming particularly relevant for identification of minimal residual disease, especially for patients treated with novel IDHinhibitors. In this study, we characterized IDH1 p.R132H clones in bone marrow specimens involved by AML using a mutation specific antibody.

Materials and Methods

Bone marrow tissue sections (biopsy or clot specimens) from 32 AML cases with IDH1 p.R132H mutation were stained with IDH1 p.R132H-mutation specific antibody. These cases include 20 de novoAML and 12 cases of AML with myelodysplasia-related changes (AML-MRC). We also included 10 AML cases with wild-type IDH1 as a control. After confirmation of the positive IDH1 immunohistochemical (IHC) signal in the primary specimens, follow up bone marrow specimens (n=67) including (a) persistent disease, (b) minimal residual disease by flow cytometry, (3) complete remission by morphology and flow cytometry, but, positive for mutation by PCR, as well as (4) relapsed cases after complete remission were included in the study (in progress). We also included pre- and post-treatment (unresponsive with increasing blast counts, stable disease, persistent disease with decreasing blast counts, complete remission, and relapse) bone marrow specimens (n=72) from 16 patients treated with IDH inhibitors (in progress).

Results

All the IDH1 wild type AML cases were negative for IDH1 IHC stain showing 100% specificity. Positive signal was detected in all de novo AML and AML-MRC (allelic frequency ranges from 1.8% to 47% by PCR) except one AML case with 8.9% allele burden which was a limited sample; overall sensitivity was 96%. The IHC signal was detected in the cytoplasm of myelomonocytic cells, their precursors, and megakaryocytes. Erythroid precursors, lymphoid cells, endothelial cells, and osteoblasts were consistently negative. The signal intensity ranged from weak (n=10) to moderate (n=9), to strong (n=13). The positive cells predominantly showed an interstitial distribution in the bone marrow. In the de novo AML group, only the immature cells were positive in 100% of pre-treatment AML cases. However, both mature and immature cells were positive in 7/13 (54%) post-treatment AML cases (6 cases treated with hypomethylating agents). One case was transformed from MPN which also showed positivity in mature and immature cells. In two cases with complete morphologic remission and one case with minimal residual disease detected by flow cytometry, IHC signal was detected in both mature and immature cells; both patients relapsed in 8 and 11 months. In the AML-MRC group, both immature and mature cells were positive in 11/12 (92%) cases of which 2 were not previously treated indicating the possibility that IDH1 mutation is an early event. Since the remaining 9 patients were treated with hypomethylating agents, the positivity of both mature and immature cells as a result of maturation effect versus an early event cannot be assessed. Additional studies for follow-up AML cases, including cases on an IDH inhibitor clinical trial are in progress.

Conclusions

Our preliminary data indicate that IDH1 IHC is a highly specific and sensitive tool to detect IDH1 R132H mutated cases and can be used as a primary method to localize the population of mutation-bearing cells in the bone marrow. IHC also allows determination of whether the IDH1 mutation in the post-treatment setting is arising from immature or mature cells. IHC provides an opportunity to understand the difference between these two populations and, based on characterization of cell type and distribution, may be helpful to predict whether the risk of relapse is high.