A Leukemic Progression Model of Severe Congenital Neutropenia Uncovers a Novel Mechanism of AML Development Involving Elevated Inflammatory Responses, Mutation of CXXC4 and Decreased TET2 Levels

Introduction: Severe congenital neutropenia (SCN) patients receive life-long treatment with CSF3/G-CSF to alleviate neutropenia and have a high risk to develop MDS or AML. The appearance of hematopoietic clones with CSF3 receptor (CSF3R) mutations represents a first step in MDS/AML progression, which is followed by mutations in RUNX1 shortly before MDS/AML becomes clinically overt. How intracellular signaling pathways are affected by the combination of CSF3 therapy, CSF3R and RUNX1 mutations, and how deregulation of these pathways contributes to MDS/AML development is unknown. Also, it has remained unclear how defects in epigenetic regulators such as ASXL1 or SUZ12, which are recurrently but less frequently mutated in SCN/AML, contribute to full malignant transformation.

Aims: (i) To elucidate how CSF3R and RUNX1 mutations in conjunction with CSF3 treatment contribute to leukemic progression in an in vivo serial transplantation mouse model (ii) To determine whether additional mutations, specifically in epigenetic regulators, are needed for full malignant transformation and (iii) To validate the relevance of the findings for clinical SCN and AML.

Mouse model and patient samples: Bone marrow (BM) cells from Csf3r-d715 mice, copying the most frequent CSF3R truncation mutant in SCN patients, were transduced with RUNX1 mutant D171N or empty vector control lentivirus and serially transplanted. Recipient mice were treated 3x a week with CSF3 or PBS (solvent control). Transcriptome analysis (RNA-sequencing + GSEA) and whole exome sequencing (WES) on FACS purified Lineage- cKit+ (LK) populations were done to identify molecular pathways associated with leukemic progression. Sequential CD34+ cell samples from a SCN/AML patient with identical CSF3R and RUNX1 mutations (Beekman, Blood 2012) and whole genome sequencing data from diagnostic AML samples were used for clinical comparisons.

Results: CSF3 treatment of primary recipients of Csf3r-RUNX1 mutant BM cells resulted in sustained (30+ weeks) presence of LK cells (16.5% ± 7), which we morphologically and functionally identified as myeloblasts, in the peripheral blood (PB). None of the mice succumbed to symptoms of AML, suggesting that the elevated PB myeloblast counts reflected a pre-leukemic state. Upon transplantation in secondary and tertiary recipients, mice developed a Csf3r-RUNX1 mutant AML that no longer depended on CSF3 administration. Transcriptome profiles of purified progenitor cells at the sequential steps of transformation in the mouse model and an SCN/AML patient with identical mutations shared striking similarities in deregulation of signaling mechanisms, particularly showing progressive upregulation of TNFα-, interferon- and interleukin-6-driven inflammatory responses during leukemic progression.

WES on the myeloblasts from these stages revealed a single additional clonal mutation in Cxxc4, an Internal tandem duplication (ITD) resulting in the in-frame insertion of two glycines at a.a. position 157 of the protein. This heterozygous mutation appeared in a subclone in the primary recipient (VAF: 0.27) and expanded in secondary and tertiary recipients, in which all AML cells harbored the Csf3r, RUNX1 and Cxxc4 mutations (VAF 0.52±0.06). The mutation resulted in a 10.8 fold (± 2.3) higher expression of a CXXC4 isoform. CXXC4 was shown to inhibit TET2 protein levels (Ko, Nature 2013). In agreement with this, TET2 levels were strongly (6.03 fold ± 0.78) reduced in the CXXC4 overexpressing leukemic samples.

Intriguingly, CXXC4 mutations have also been detected in human AML cases (n=13), including the homologous ITD mutations identified in our mouse model. We are currently investigating the frequencies and co-occurrence of CXXC4 mutations in RUNX1 mutant and other molecular subtypes of AML. Also, studies are in progress to determine how the CXXC4-ITD alters CXXC4 and TET2 protein expression.

Conclusion: By studying the leukemic progression of SCN, driven by CSF3 treatment and mutations in CSF3R and RUNX1, we have identified a mechanism of AML development that involves activation of multiple inflammatory pathways, mutation of CXXC4 and reduction of TET2 expression. These findings corroborate a recent study showing that inflammatory responses drive pre-leukemic myeloproliferation in Tet2 deficient mice (Meisel, Nature 2018).