TTLL4 glutamyltransferase is a therapeutic target for NPM1-mutated Acute Myeloid Leukemia

Abstract

Abstract Mutations in nucleophosmin (NPM1) are identified in 35% of acute myeloid leukemia (AML) patients, but therapeutic strategies targeting this subtype of AML remain limited [PMID:23436734]. The most common mutation in NPM1, termed NPM1c, is a C-terminal frameshift mutation that results in cytoplasmic mis-localization of the mutant protein and aberrant expression of HOXA/B and MEIS1 in myeloid progenitors, which suppresses differentiation and promotes self-renewal. Overexpression of these genes is mediated by direct binding of the NPM1c A2 acidic stretch to chromatin [PMID:36455613]. This region is susceptible to various post translational modifications that regulate its function. We found that TTLL4 catalyzes post-translational glutamylation of both NPM1c and NPM1 specifically at E126 in the A2 stretch. Although TTLL4 expression was equivalent across NPM1-mutant and wildtype AML cell lines, mass spectrometry analysis of mutant and wild type NPM1 purified from AML cells and recombinant NPM1 and NPM1c proteins showed that the mutant protein is the preferred biochemical substrate for TTLL4. Genetic inactivation of TTLL4 byCRISPR-Cas9 or shRNA approaches in NPM1-mutated OCI-AML3 cells reduced NPM1c glutamylation, impaired proliferation and colony formation, and triggered myeloid differentiation, as evidenced by significantly increased CD11b and CD14 expression and acquisition of morphologic changes consistent with neutrophil and monocyte morphology. RNA sequencing performed after TTLL4 shRNA knockdown in OCI-AML3 cells with a degron-tagged NPM1c (PMID: 30205049) revealed activation of myeloid gene expression programs in TTLL4-deficient cells. The effects on gene expression correlated significantly with the effects of NPM1c degradation. Gene ontology showed significant overlap in processes regulated by TTLL4 and NPM1c, including enrichment of myeloid cell differentiation and myeloid leukocyte activation signatures and loss of stem cell-associated gene signatures, suggesting that glutamylation promotes leukemogenesis through an NPM1c-dependent mechanism. We observed significant overlap of TTLL4 knockdownwith inhibition of KMT2A complex interactors, including menin, ENL, KAT6A, and BRD4. The menin-KMT2A complex is a key mediator of NPM1c chromatin binding; thus, our findings suggest that TTLL4-mediated glutamylation has a plausible role in modulating the KMT2A-mediated NPM1c transcriptional response [PMID:36455613]. The TTLL4-knockdown transcriptome also overlapped with that of selinexor inhibition of XPO1 in OCI-AML3 cells, which promotes differentiation by inducing nuclear re-localization of NPM1c. Fractionation immunoblots revealed re-localization of NPM1c from the cytoplasm to the chromatin fraction of TTLL4 knockdown cells, suggesting that glutamylation may be important to maintain NPM1c in the cytoplasm. Studies using NPM1c [PMID:30692594] and NPM1c/NRAS-driven [PMID:28835438] mouse AML models validated TTLL4 as a targetable vulnerability. We found that Ttll4 deletion impaired colony formation of NPM1c+but not of wild type mouse bone marrow, suggesting that TTLL4 is specifically important for Npm1-mutatedcells. Recipients of Ttll4-deleted Npm1cA/+;NrasG12D/+ leukemic bone marrow cells had significantly prolonged survival compared to controls before succumbing to AML (median survival 59d v. 36d, p = 0.008). We observed significant expansion of c-Kit-negative myeloid populations and increased expression of the myeloid markers CD11b, CD14, and CD115 in Ttll4-deleted leukemic bone marrow, consistent with myeloid differentiation of leukemic blasts. Finally, we identified a small molecule, EN7, that selectively inhibits TTLL4 and recapitulates the anti-proliferative and pro-differentiation phenotype in NPM1c AML cell lines. EN7 did not have these effects on NPM1wt cell lines or on healthy CD34+ bone marrow progenitors, suggesting a therapeutic window for targeting TTLL4. These findings identify glutamylation as a new axis of leukemic regulation and highlight TTLL4 as a druggable vulnerability in NPM1c AML.