Sensitivity to Targeted UBA1 Inhibition in a Myeloid Cell Line Model of Vexas Syndrome

Somatic UBA1 mutations in hematopoietic cells are a hallmark of VEXAS (Vacuoles, E1 enzyme, X-linked, Autoinflammatory, Somatic) syndrome, which is a late-onset, male-predominant myeloid autoinflammatory disease associated with bone marrow failure and high mortality. UBA1 resides on the X chromosome and encodes the E1 enzyme isoforms UBA1a and UBA1b, which are essential for nuclear and cytosolic protein ubiquitination, respectively. The majority of UBA1 mutations in VEXAS syndrome comprise hemizygous missense mutations affecting methionine-41 (M41), leading to loss of UBA1b, gain of a truncated protein isoform (UBA1c),and globally reduced polyubiquitination. The relationship between impaired polyubiquitination, autoinflammation, and dyspoiesis in VEXAS syndrome is poorly understood, highlighting the need for experimental models that reproduce the genetic, biochemical, morphological, and inflammatory features associated with somatic UBA1 ^M41 missense mutations.

To investigate the effects of VEXAS syndrome-associated UBA1 mutations, we used CRISPR-Cas9-based homology directed repair (HDR) in the male myeloid cell line 32D to generate isogenic single cell clones expressing Uba1 ^WT or Uba1 ^M41L from the endogenous mouse Uba1 locus. Sanger sequencing analysis of genomic and complementary DNA indicated the presence of a single Uba1 allele in Uba1 ^WT and Uba1 ^M41L single cell clones. Consistent with prior analyses of VEXAS syndrome patient samples, UBA1 and ubiquitin immunoblots showed that hemizygous Uba1 ^M41L expression was associated with loss of the UBA1b protein isoform, gain of the UBA1c protein isoform and reduced polyubiquitination. Uba1 ^M41L cells also contained abnormal cytoplasmic vacuoles, which were largely absent from Uba1 ^WT cells. Electron microscopy revealed that vacuoles in Uba1 ^M41L cells contained a variety of endolysosomal membranes, including small vesicles, multivesicular bodies, and multilamellar lysosomes. Similar to cytokine profiles in VEXAS syndrome patient samples, concentrations of IL1β and CXCL10 were significantly increased in supernatants from Uba1 ^M41L cells compared to Uba1 ^WT cells; Uba1 ^M41L cell supernatants also contained higher concentrations of CCL5, CCL17, CXCL9, and CXCL12. Thus, hemizygous Uba1 ^M41L expression in a myeloid cell line recapitulates key biochemical, morphological, and inflammatory features of VEXAS syndrome.

The UBA1c protein isoform generated via UBA1 ^M41 missense mutations has reduced capacity to form the UBA1-ubiquitin thioester bond necessary for ubiquitin transfer to E2 enzymes. We hypothesized that decreased UBA1 activity would render Uba1 ^M41L cells vulnerable to the UBA1 inhibitor TAK243. In cell proliferation assays, Uba1 ^M41L cells were more sensitive to TAK243 compared to Uba1 ^WT cells (IC ₅₀ mean±s.d., 1.4±0.6nM vs. 63±28nM, respectively). TAK243 promoted apoptosis in Uba1 ^M41L cells, as indicated by increased annexin-V/propidium iodide staining, H2AX, H2AX pS139, and PARP1 cleavage. TAK243 also led to preferential loss of Uba1 ^M41L cells in competition assays with Uba1 ^WT cells. To determine if on-target UBA1 inhibition was responsible for the increased TAK243 sensitivity of Uba1 ^M41L cells, we used CRISPR-HDR to generate a TAK243 drug-binding mutant, Uba1 ^A580S. Knock-in of Uba1 ^A580S rendered Uba1 ^M41L cells resistant to TAK243; the TAK243 IC ₅₀ was significantly higher in Uba1 ^M41L;A580S cells compared to Uba1 ^M41L cells (mean±s.d., 73±16nM vs. 3.8±1.2nM, respectively). Together, these data demonstrate that hemizygous expression of Uba1 ^M41L in 32D cells confers TAK243 sensitivity via on-target inhibition of UBA1.

Notably, increased TAK243 sensitivity and reduced polyubiquitination in Uba1 ^M41L cells were rescued by overexpression of UBA1b, but not UBA1b/C632A (catalytically inactive) or luciferase, indicating that loss of UBA1b enzymatic function contributes to the phenotype conferred by Uba1 ^M41L.

In summary, we present a myeloid cell line with hemizygous expression of Uba1 ^M41L that exhibits the biochemical, morphological, and inflammatory features of VEXAS syndrome. We find that loss of UBA1b is a cause of reduced polyubiquitination and renders Uba1 ^M41L cells vulnerable to targeted UBA1 inhibition by TAK243. Our Uba1 ^M41L knock-in cell line is a faithful model of VEXAS syndrome that will aid in the study of disease pathogenesis and the development of effective therapies.