BAALC Is a Key Mediator of Leukemia Development in Congenital Neutropenia

We recently reported high frequency of cooperating acquired CSF3R and RUNX1 mutations in severe congenital neutropenia (CN) patients that developed AML or MDS. CN is a pre-leukemic bone marrow failure syndrome (BMFS) and longitudinal retrospective analysis of CN bone marrow samples allows identification of the pathological intracellular events that lead to leukemogenic transformation.

In the present study, we report the identification of a key mechanism of leukemogenesis in CN. We made use of the induced pluripotent stem cell (iPSC) model and embryoid body (EB)-based in vitro hematopoietic differentiation of iPSCs generated from a CN patient harboring ELANE mutation before (CN-iPSCs) and after (CN/AML-iPSCs) overt AML. CN/AML-iPSCs clones revealed CSF3R and RUNX1 mutations and trisomy 21. EB-based hematopoietic differentiation revealed that CN-iPSCs produced similar amounts of CD34⁺ cells, but approximately 2-fold reduction of mature granulocytes and CFU-G colonies in CFU-assay, as compared to healthy donor-derived iPSCs. CN/AML-iPSCs produced CD34⁺ cells but did not differentiate into mature granulocytes at all.

Analysis of the gene expression profiles of five CN/AML patients including our index patient, revealed that genes upregulated in early hematopoietic stem/progenitor cells (HSPCs), such as Brain And Acute Leukemia, Cytoplasmic (BAALC), Hematopoietic Prostaglandin D Synthase (HPGDS), Natriuretic Peptide Receptor 3 (NPR3) and CD109 were markedly increased in CN/AML blasts harboring both RUNX1 and CSF3R mutations, as compared to hematopoietic cells prior to leukemia. This genetic signature suggests transformation of hematopoietic cells into more primitive hematopoietic progenitors after acquisition of CSF3R and RUNX1 mutations. We were able to confirm strong increase of BAALC, HPGDS, NPR3 and CD109 mRNA levels in CD34⁺ cells derived from CN/AML-iPSCs, as compared to CN-iPSCs, suggesting that our iPSC-model is reliable to study leukemogenesis.

In order to evaluate which of these candidate genes (BAALC, HPGDS, NPR3 and CD109)is responsible for the leukemogenic transformation in CN, we generated CN/AML-iPSCs clones deficient for one selected genes. We established CRISPR-Cas9-mediated gene knockout using transfection of CN/AML-iPSCs with pSpCas9(BB)-2A-GFP vector expressing Cas9 and sgRNA, that were designed to specifically target first exon of each candidate gene. CN/AML-iPSCs transfected with specific sgRNAs were sorted and single cell clones were analyzed further.

Knockout of the selected genes did not affect survival and pluripotency of CN/AML-iPSCs and we were able to study EB-based hematopoietic differentiation of gene-edited CN/AML-iPSCs clones. Interestingly, knockout of HPGDS, NPR3 or CD109 did not correct defective granulocytic differentiation of CN/AML-iPSCs, whereas knockout of BAALC in CN/AML-iPSCs resulted in the tremendous increase of the granulocytic differentiation, as assessed by FACS, CFU assay, morphological analysis of cytospin preparations and nitroblue tetrazolium (NBT) assay. Granulocytic differentiation of BAALC -deficient CN/AML-iPSCs was comparable to CN-iPSCs, but was still reduced in comparison to healthy donor-derived iPSCs.

These data clearly showed that knockout of BAALC in CN/AML-iPSCs overcomes differentiation block at the early hematopoietic progenitor cell stage, demonstrating a fundamental role of BAALC in leukemia development in CN. BAALC, HPGDS and CD109 are known to be markedly up-regulated in RUNX1 -mutated de novo AML and qRT-PCR of CD34⁺ cells generated from BAALC -knockout CN/AML-iPSCs revealed a decrease of mRNA expression levels of BAALC, HPGDS and CD109 . At the same time, knockout of RUNX1 in CN/AML-iPSCs using CRISPR/Cas9-editing revealed even higher decrease of mRNA levels of BAALC, NPR3 and CD109 . These data suggest that HPGDS and CD109 up-regulation is downstream of RUNX1 -mediated BAALC induction in AML.

Taken together, we established a reliable in vitro model to study leukemogenesis in the pre-leukemic BMFS, using CN as an example. With the help of this model, we identified BAALC up-regulation as a key "leukemogenic" factor downstream of RUNX1 and CSF3R mutations. Based on these observations, we conclude that inhibition of BAALC in hematopoietic cells of CN patients may prevent leukemogenic transformation of hematopoiesis or eliminate AML blasts in CN-AML.