Bespoke Base and Prime Editing Approaches for STING-Associated Vasculopathy with Onset in Infancy (SAVI)

IntroductionSTING-associated vasculopathy with onset in infancy (SAVI) is a rare, severe type I interferonopathy caused by gain-of-function mutations in STING1, leading to early-onset systemic inflammation, cutaneous vasculopathy, and life-threatening interstitial lung disease. Current treatments attenuate downstream inflammation without fully addressing the pathogenic driver and carry significant adverse effects. Recently, gene editing has emerged as a paradigm-shifting approach for IEI, with substantial potential for gain-of-function disorders that require allele-specific correction. We present a highly efficient and specific base- and prime-editing strategy that supports both ex vivo and in vivo therapeutic applications.ResultsK562 were transduced to generate stable STING1-WT and STING1-V155M lines (K562wt, K562mut). 8 sgRNAs were screened for base editing across PAM-flexible nucleases (SpCas9-NG, SpRY) in K562mut, and sgRNA2 and sgRNA5 were selected based on on-target editing. Given the high bystander activity observed with sgRNA5, we leveraged a previously developed library-derived in silico TadA design tool to identify >15 variants for testing. The TadA variant prioritized after experimental screening (var_17) reduced bystander editing (p<0.05). To overcome bystander constraints, PEGsm was selected for prime editing after comprehensive screening. Top-performing guides and selected editor variants were delivered by electroporation as in vitro–transcribed mRNA, achieving high on-target editing (90.10% ± 2.73%) for sgRNA5 and PEGsm (PE2 and PE3) and 53.48% ± 2.02% for sgRNA2. To functionally assess editing, interferon-stimulated gene transcripts (IFIT1, ISG15, and IFI44L) were quantified by droplet digital PCR (ddPCR) before and after 2′3′-cGAMP stimulation. ISG15 and IFIT1 were significantly reduced in unstimulated K562mut-edited cells (sgRNA2, PEGsm) versus K562mut, while after stimulation, K562wt and K562mut-edited reached the K562mut plateau as expected. In healthy donor hematopoietic stem and progenitor cells (HSPCs), a surrogate sgRNA was used to benchmark editing efficiency within the target window.Surrogate sgRNA achieved a mean bystander editing rate of 75.0% ± 4.4%. Finally, peripheral blood mononuclear cells (PBMCs) were reprogrammed into patient-derived induced pluripotent stem cells (iPSCs) using non-integrating Sendai vectors. iPSCs were prime edited to generate fully corrected isogenic single-cell–derived clones.ConclusionWe developed a comprehensive, bespoke base- and prime-editing platform for SAVI, enabling efficient and specific STING1 variant correction with functional normalization of interferon-stimulated genes (ISGs) expression, paving the way for future therapeutic applications. We also generated multiple fully corrected isogenic iPSC clones to support mechanistic studies of SAVI pathogenesis.Figure 1.Engineering of a STING1-V155M K562 model and functional assessment of bespoke base and prime editing. (A) Bidirectional third-generation lentiviral (LV) vector used to generate stable K562wt and K562mut lines expressing STING1-WT or STING1-V155M together with an EGFRt surface marker. (B) In vitro transcription (IVT) mRNA electroporation of top guide–editor pairs yields high A-to-G editing in K562mut, including base editing with sgRNA2/sgRNA5, a bystander-reducing TadA variant (var17) with sgRNA5, and prime editing with PEGsm in PE2 and PE3 configurations; editing was quantified by Sanger sequencing (EditR). (C) Functional readout by ddPCR showing IFIT1 and ISG15 expression (normalized to HPRT1) in K562mut across editing conditions and after STING inhibition (H-151). Statistics were computed by one-way ANOVA with multiple comparisons correction; significance is indicated, and where not otherwise specified, **** denotes p < 0.0001.