Loss of Negative Metabolic Feedback Drives TET2 Mutant Inflammation

TET2 mutations are clinically significant mutations in clonal hematopoiesis of indeterminate potential (CHIP) and myeloid malignancies driving inflammation and clonal progression, but the underlying mechanism of how TET2 loss leads to these phenotypes is not understood. Here, we used in vivo genome-wide genetic perturbations in primary WT or Tet2 KO Cas9+ hematopoietic stem and progenitor cells (HSPCs) in a model of zymosan serositis to decipher the mechanisms of Tet2 KO inflammation.

Previous experimental approaches using CRISPR/Cas9 to study factors governing hematopoiesis in vivo have been hampered by heterogeneity in cellular expansion during hematopoietic reconstitution. To overcome this limitation, we developed an in vivo hematopoietic stem and progenitor (HSPC)-based CRISPR/Cas9 screening system with high-diversity (> 10 million) unique molecular identifiers (UMIs). After transplanting Cas9-expressing Lin^-Sca1⁺c-Kit⁺ (LSK) cells transduced with UMI-labeled guide-RNAs (gRNAs), we compared neutrophils and monocytes recruited during zymosan stimulus compared to their counterparts in the bone marrow. This genome-wide CRISPR knockout screen revealed that Tet2 KO myeloid cells were dependent on lipid metabolic genes for inflammatory recruitment but not WT myeloid cells. In agreement, Tet2 KO neutrophils and macrophages accumulated neutral lipids as confirmed on BODIPY 493/503 neutral lipid staining and lipidomic analysis.

The top dependency for Tet2 KO inflammation was OGT, O-linked N-acetylglucosamine transferase, responsible for attaching O-GlcNAc modifications onto proteins. To investigate whether OGT activity is changed in Tet2 KO, we performed nutrient tracing studies using uniformly labeled C13 glucose and IP-MS for O-GlcNAc in WT and Tet2 KO bone-marrow derived macrophages (BMDMs), which revealed increased utilization of UDP-GlcNAc, the substrate for OGT. To determine whether OGT controls neutral lipid accumulation, we knocked down OGT or used (Z)-PUGNAc, an OGA inhibitor to mimic increased OGT activity, which both showed that OGT controls neutral lipid accumulation.

As OGT is among the top binding partners for TET2, we investigated whether OGT binding to TET2 is critical for inflammation. Mapping studies show that OGT binds to the disordered region on TET2 and when this interaction is disrupted, TET2 mutants have increased inflammation. Conversely, the OGT-binding TET2 peptide is sufficient to reduce inflammation in Tet2 KO whereas the mutant TET2 peptide that does not bind OGT had little effect, suggesting that TET2 restrains OGT to control inflammation.

On chromatin, TET2 negatively regulates OGT and unrestrained OGT leads to increased OGT and O-GlcNAcylation signal on ChIP-Seq in BMDMs. These regions overlap with SREBP1/LXR binding sites, driving inflammatory lipid metabolism. A therapeutic target of OGT hyperactivation in TET2 mutants is ACLY, serving as an important node for lipid accumulation and inflammation in Tet2 KO.

In summary, through a novel in vivo CRISPR screening system combined with metabolomics, proteomics and epigenetic studies, we reveal that TET2 is a negative regulator of a master nutrient sensor OGT. In response to nutrient cues, TET2 physiologically binds OGT on chromatin thus restraining its activity. Upon TET2 loss, OGT binding to genomic loci is increased causing metabolic hyperactivity and lipid accumulation, ultimately driving inflammation.