Tifab Deficiency Protects Against Stress-Induced Bone Marrow Failure at the Expense of Pre-Leukemic Selection

Interstitial deletion of a single copy of chromosome 5q is the most frequent cytogenetic alteration in Myelodysplastic Syndromes (MDS), which results in reduced dosage of numerous genes. TRAF-interacting protein with forkhead-associated domain B (TIFAB) resides within the proximal commonly-deleted region on band 5q31.1, and belongs to a family of forkhead-associated domain proteins. TIFAB is deleted in nearly all reported cases of del(5q) MDS and AML. As expected, TIFAB expression is significantly lower in CD34+ and BM mononuclear cells isolated from MDS patients with del(5q) as compared with cells from MDS patients diploid at chr 5q. Recently we have shown that hematopoietic-specific deletion of Tifab in mice results in progressive BM and blood defects, including aberrant HSPC proportions, altered myeloid differentiation, and progressive cytopenia (Varney and Niederkorn et al., JEM 2015). Gene expression analysis of Tifab KO lineage-Sca1+cKit+ (LSK) cells identified dysregulation of immune-related signatures, hypersensitivity to Toll-like receptor stimulation, and enrichment of a p53 regulatory gene-set.

Importantly, p53 is implicated in the pathogenesis of del(5q) MDS: 1) BM cells from murine models and BM from del(5q) patients exhibit increased p53 activity, which is thought to contribute to ineffective hematopoiesis and anemia; and 2) del(5q) MDS patients often acquire concurrent TP53 mutations that result in rapid transformation to AML and poor treatment response. To examine the effects of TIFAB on p53 function, we examined p53 target genes in TIFAB-overexpressing and -deficient cells. Gene expression profiling and qRT-PCR analysis of Tifab KOHSPCs revealed significant upregulation of p53 regulatory genes. In contrast, overexpression of TIFAB in a p53-competent cell line reduced the expression of p53 target gene, p21 . We also observed increased p53 protein in Tifab KO total BM, lineage- BM, and spleen. At steady-state and after in vitro treatment with 2 Gy irradiation, Tifab KO lineage- BM cells exhibit increased p53 protein compared to WT cells.

To test whether this increased p53 stability renders Tifab -deficient HSPCs more sensitive to stress, we treated Tifab KO HSPCs both in vivo and ex vivo using irradiation. When treated with low dose irradiation ex vivo, Tifab KO HSPCs initially exhibit significantly reduced colony forming ability; however, upon secondary re-plating, Tifab KO HSPCs strikingly form more colonies than WT HSPC, resembling a clonal selection event. To model this effect in vivo, we transplanted WT or Tifab KO BM into lethally-irradiated recipient mice. After 4 weeks of hematopoietic reconstitution, recipient mice were treated with 6.5 Gy sublethal irradiation. We find that Tifab KO recipient mice are less likely to succumb to IR-induced bone marrow failure as compared to mice transplanted with WT BM cells. Instead, after 3+ months, the remaining Tifab KO recipient mice have an increased tendency to develop hematologic malignancies, evident by myeloid blasts. Collectively, these data suggest that TIFAB-deficiency protects against cellular stress by eliminating "less fit" HSPCs, at the expense of selecting for pre-leukemic populations, paralleling observations in patients with transformed MDS.

To investigate the mechanism of this phenomenon, we performed a tandem-affinity tag purification and mass-spectrometry analysis of TIFAB complexes in a del(5q) AML cell line (HL60), and identified unique TIFAB-interacting proteins. The top interacting candidate was an ubiquitin-specific peptidase (USP), USP15. USPs play a major role in ubiquitin-dependent processes including DNA damage response signaling, protein degradation, and kinase activation. Specifically, USP15 has been shown to promote p53 degradation via deubiquitination and stabilization of its major negative regulator, MDM2. Through biochemical assays and a series of deletion mutants, we confirmed that TIFAB interacts with USP15 and enhances the deubiquitination of biological USP15 substrates, including MDM2. Thus, our data implicate TIFAB as an important negative regulator of p53 and modulator of USP15 DUB activity.

Altogether, our findings identify a novel role for TIFAB as a modifier of p53 function, potentially via a USP15-MDM2 axis. These findings have important implications in the potential role of TIFAB and p53 signaling in the pathogenesis of del(5q) MDS and transformation to AML.