Microenvironmental Links Between TET2/DNMT3A Mutations and Arginase 1 Overexpression in Human MDS/CMML

Introduction:

Immune dysregulation is a common feature of myeloid cancers including MDS and CMML. Precisely how cells such as macrophages (MΦs) contribute to or are impacted by aberrant innate immunity is not fully understood. Previously, our group demonstrated Arginase 1 (Arg1) overexpression in MDS/CMML patient samples as well as in a murine model of Tet2-deficiency (Cull et al., Blood Abstract 2013: 1578). Arg1, an L-arginine metabolizing enzyme, is highly expressed by alternatively activated (M2) macrophages and myeloid-derived suppressor cells. In the current study, our goals were to (1) establish whether the presence of certain mutations (i.e. in TET2) corresponds with high Arg1 immunohistochemical (IHC) staining, and (2) determine the environmental vs. cell-intrinsic effects of Tet2-deficiency on Arg1 expression in murine peritoneal macrophages (PMΦs).

Methods:

H&E staining and anti-human Arg1 IHC was performed on FFPE BM biopsies obtained from Sunnybrook and Kingston General Hospitals (8 age-matched controls and 30 MDS/CMML cases including 9 RCMD, 5 RAEB-1, 2 RAEB-2, 2 RARS, 1 RA, 1 del(5q) and 10 CMML-1). IHC staining was scored independently by 2 blinded Hematopathologists. Genomic DNA from each patient was subjected to mutation profiling using a 1,552-amplicon Ion Torrent Sequencing Panel. Primary PMΦ were obtained from peritoneal lavages performed on hematopoietic-specific Vav1-Cre-driven Tet2 knockout (Tet2-/-) mice generated from parental floxed (Tet2f/f) and Vav1-Cre mice (JAX) in accordance with Queen's Animal Care protocols.

Results:

Following an initial pilot study, we generated Arg1 IHC data and corresponding genome sequences from 8 age-matched control and 30 MDS/CMML patient FFPE BM biopsies. In keeping with our previous findings, we observed increased BM biopsy Arg1 IHC expression in CMML and low-grade MDS for both hospital cohorts. Taking all patient samples together, we found a significant correlation between high Arg1 IHC status and the presence of either TET2 or DNMT3A mutations (Fisher's exact test, p-value = 0.0173), both of which are highly prevalent in MDS and CMML patients. While this result - that mutations in a DNA methyltransferase and a DNA demethylating protein could lead to increased Arg1 protein expression and ultimately a similar disease phenotype - may appear counterintuitive, a recent publication provided evidence that Tet2 and Dnmt3a act cooperatively at specific sites within the genome (PMID 27428748). In addition, we noted upregulation of Ccr5, Tgfbi, Socs3, Bst2, Ccl6, Ifit2, Ccl2, Cxcl1, Arg1 and Ccl7 in Dnmt3a knockout vs. wild-type mouse PMΦs (PMID 27240213). Intriguingly, our results showed that all of these genes were also upregulated in Tet2-/- knockout PMΦs. Taken together, this suggests a biological basis linking high Arg1 IHC levels with mutations in TET2 and DNMT3A.

In order to study environmental influences that may impact Arg1 expression, we isolated PMΦs from Tet2f/f and Tet2-/- mice and treated these cells with peritoneal lavage fluid derived from both control and Tet2-knockout animals. We observed a 10.2-fold increase in Arg1 mRNA expression in Tet2f/f cells treated with Tet2-/- peritoneal lavage fluid. Moreover, Tet2-/- lavage fluidtreatment of Tet2-/- PMΦs induced a 2.2-fold increase in Arg1 expression (from 5.8- to 12.8-fold compared to untreated Tet2f/f control cells).

Given these findings, we hypothesized that Tet2-/- peritoneal lavage fluid contains a factor(s) which leads to induction of Arg1. In a pilot experiment, Tet2f/f and Tet2-/- lavage fluid (n=3/genotype) samples were sent for 31-plex cytokine/chemokine array (Eve Technologies). Levels of the pro-inflammatory molecules IL-12 (p70) and MIG (CXCL9) were elevated in knockout lavage fluid versus controls, both demonstrating trends towards significance. While Arg1 upregulation may be occurring in response to aberrant inflammation, we have also observed that PMΦs can be cultured ex vivo over a period of at least 72h all the while maintaining elevated Arg1 expression. This suggests a cell-intrinsic mechanism by which these cells maintain high levels of Arg1 transcript independent of the peritoneal microenvironment.

Conclusion:

Arg1 overexpression in both TET2/DNMT3A mutant human MDS/CMML and Tet2-deficient mouse models points to the potential utility of this enzyme as a marker of aberrant inflammation.