SERF1 Is Required for G-CSF Resistance of Start-Codon Mutant ELANE Granulocytic Precursors

Abstract Severe congenital neutropenia (SCN) is frequently associated with mutations in the ELANE gene encoding neutrophil elastase (NE), an azurophilic granule protein which represents a major fraction of all protein synthesized by neutrophilic promyelocytes and myelocytes. The reduction of mutant ELANE translation inversely correlates with neutrophil differentiation. Chronic administration of granulocyte colony-stimulating-factor (G-CSF) has been successfully used to treat some of these patients, especially those ones with mutations in C-terminus translating exons-4 and -5. Patients with proximal mutations of ELANE tend to be resistant to G-CSF therapy. The SCN pathobiology is poorly understood due to the lack of efficient disease animal models, and difficulty in obtaining relevant primary cell populations from patients. We have demonstrated that ELANE exon-3 mutations, sensitive to higher doses of G-CSF administration, cause maturation arrest and apoptosis of promyelocyte and myelocyte stages due to mutant NE inducing unfolded protein response (UPR) and endoplasmic reticulum (ER) stress. We have also shown that SCN patients with mutations in ELANE at its translation initiation codon (ELANE-ATG mut) result in synthesis of alternate truncated neutrophil elastase peptides from downstream in-frame ATG due to the presence of internal ribosomal entry sites. However, the mechanisms of SCN pathogenesis associated with mutations at the translation initiation codon of ELANE gene associated with G-CSF resistance remain unclear. We employed a disease-in-a-dish approach by directed hematopoietic and granulopoietic differentiation of gene edited (CRISPR/Cas9) isogenic patient-derived iPSC lines. We used four iPSC lines (healthy donor 28L and its start codon knock-in isogenic 28L-GTG; and patient-derived ATG mut SCN110 and its isogenic corrected line SCN110C) to evaluate if ELANE start codon mutations are both necessary and sufficient to cause disease pathogenesis, and to unravel the underlying molecular mechanisms. We found that myeloid precursors derived from 28-GTG and SCN110 iPSC lines generate low molecular weight truncated NE that is detected by a C-terminal region (C225-H238) specific anti-NE antibody (Ab). The survival, expansion and differentiation of ELANE start codon mutant hematopoietic progenitors were significantly reduced in cytokine supplemented in vitro myelopoiesis cultures, and the correction of the mutation (SCN110C) restored the granulocytic precursor expansion and differentiation. The reduced expansion in 28L-GTG and SCN110 lines was associated with both increased apoptosis and expression of proapoptotic BH3-only proteins in comparison to their isogenic 28L and SCN110C lines. However, this apoptosis was not associated with UPR/ER-stress in ELANE-ATG mut myeloid precursors. Granulopoietic differentiation of the 28L-GTG and SCN110 iPSC derived hematopoietic progenitors was significantly reduced, while it was rescued in corrected isogenic lines. Correlating with their clinical behavior, high-dose G-CSF in vitro did not restore the granulopoietic differentiation of ELANE ATG mutant hematopoietic progenitors. Mechanistically, 28L-GTG and SCN110 myeloid precursors contained high-molecular weight, sodium dodecyl sulfate (SDS) resistant NE bands that were detected by anti-NE Ab specific for the C-terminal (C225-238), but not the N-terminal region (C19-38), suggesting that truncated mutant NE generates insoluble NE aggregates. The association of truncated NE with classical autophagy marker LC3B strongly suggested the activation of an aggrephagy process in these cells. We found that SERF1 (small EDRK-rich factor), an RNA-chaperoning protein, known to localize in misfolded protein aggregates of neurodegenerative diseases, was highly upregulated in GTG knock-in and SCN patient myeloid precursors. SERF1 was distributed in both nucleus and cytoplasm, co-localized and physically interacted with mutant truncated NE in the cytoplasm of ELANE-ATG mut cells as shown by proximity ligation assays. Silencing of SERF1 made the survival and differentiation of 28L-GTG and SCN110 derived myeloid precursors sensitive to G-CSF. ELANE translation initiation mutant was associated with truncated NE aggregates and induced apoptosis mediated by SERF1. This data opens a path for therapeutic intervention of G-CSF resistant ELANEmut SCN. Disclosures Kalfa: Agios Pharmaceuticals, Inc.: Other: Steering Committee, Research Funding; FORMA Therapeutics, Inc: Research Funding. Lutzko: Aruvant Sciences: Patents & Royalties: preclinical vector development. Cancelas: Hemerus LLC: Research Funding; Vascular Solutions Inc.: Research Funding; Westat Inc: Research Funding; Fresenius-Kabi LLC: Research Funding; Cytosorbents Inc: Research Funding; Hemanext: Membership on an entity's Board of Directors or advisory committees, Research Funding; TerumoBCT: Research Funding; Cerus Co: Research Funding; University of South Florida/MEQU Inc: Research Funding.