Prime Editor Gene Therapy and TREX1 Mosaicism in Retinal Vasculopathy with Cerebral Leukoencephalopathy

Abstract

TREX1 mutations underlie a variety of human diseases, including retinal vasculopathy with cerebral leukoencephalopathy (RVCL or RVCL-S), a catastrophic adult-onset vasculopathy that is often confused with multiple sclerosis, systemic vasculitis, or systemic lupus erythematosus. Patients with RVCL develop brain, retinal, liver, and kidney disease around age 35-55, leading to premature death in 100% of patients expressing an autosomal dominant C-terminally truncated form of TREX1. We previously demonstrated that RVCL is characterized by high levels of DNA damage, premature cellular senescence, and risk of early-onset breast cancer before age 45. Here, we report human TREX1 mosaicism causing organ-limited RVCL in the retina, as well as a gene therapy to synthetically create TREX1 mosaicism as a potential treatment for RVCL. In our patient with organ-limited disease, the mosaic TREX1 mutant allele underwent germline transmission to 3 children, who developed severe multi-organ disease at ~ age 40, unlike their mosaic parent, who has organ-limited disease at age 74. Additionally, we describe our TREX1 prime editor gene therapy that corrects the most common RVCL-causing TREX1 variant in cell culture and in mice. Thus, TREX1 mosaicism causes organ-limited RVCL with a normal lifespan, suggesting that a gene therapy to create TREX1 mosaicism in adults may someday become useful as a treatment for patients with RVCL.

Keywords: CRV; Cerebroretinal vasculopathy; Dementia; Gene therapy; HERNS; Prime editor; RVCL; RVCL-S; Retinal vasculopathy with cerebral leukoencephalopathy and systemic manifestations; Small vessel disease; TREX1; Vasculopathy.

Fig. 1

Organ-limited RVCL due to parental TREX1 mosaicism with classic multi-organ RVCL in her progeny. (A) Pedigree of the family with inheritance of a mosaic disease-causing TREX1 mutation. Shaded boxes represent patients with germline TREX1 E266X mutations and the patterned shading indicates mosaicism in the mother. (B) MRI of patient II.2 from (A) with inherited RVCL. (C) Key clinical features in the affected family. (D) Schematic of the TREX1 protein with known disease-causing mutations. (E) Representative Western blot of TREX1 E266X expression in 293T cells. Data in (E) are representative of n = 3 independent experiments. (F) Frequency of TREX1 protein variants found in our cohort of 120 patients with RVCL. (G) Representative confocal immunofluorescence images of 293T cells expressing HA-tagged WT TREX1 or TREX1 E266X with immunostaining for TREX1 (HA). Scale bar = 15 μm. Data are representative of all cells observed in 2 independent experiments

Fig. 2

Prime-editing gene therapy converts the RVCL-causing TREX1 mutant protein to wild-type TREX1 in cultured cells and in mice. (A) Schematic of prime-editing gene therapy for TREX1 V235Gfs, Created in BioRender. Miner, J. (2024) https://BioRender.com/g08d573 (B) Western blot of TREX1 in 293T cells expressing TREX1 V235Gfs and transfected with the RVCL gene editor. After SDS-PAGE, membranes were probed with antibodies against the N-terminal HA tag (left), or an antibody against the C-terminus of TREX1 (middle), or GAPDH (right). On the right, we include diagram indicating the regions of TREX1 detected by antibodies to the N- and C-termini, created in BioRender. Miner, J. (2024) https://BioRender.com/t88w064. Data are representative of three independent experiments. (C) Prime editing efficiency in 293T cells (left) and livers of mice treated intravenously with AAV encoding the RVCL prime editor (right). For statistical analysis, the data represent the mean ± SEM of n = 3 samples pooled from three independent experiments and were analyzed by Student’s t test (^**P < 0.01; ^***P < 0.001). (D) Representative epifluorescence images of liver sections from unedited and AAV prime-edited LSL CAG-Cre TREX1 mice expressing HA-tagged TREX1 V235Gfs and WT TREX1 (control), expression was induced with daily tamoxifen injection for three or four days. Scale bar = 20 μm. Data in (D) are representative of two independent experiments