Delivery of a modified U1 small nuclear RNA alleviates splicing-defective coagulation Factor VII expression in mouse models

Background: The small nuclear RNA U1 (U1snRNA), the component of the spliceosomal U1snRNP, is an attractive therapeutic molecule because of its ability to rescue mRNA splicing impaired by mutations. However, the correction efficacy has been proven only in cellular models. Coagulation factor deficiencies represents ideal models to test the U1snRNA-based strategy since even modest increase of functional protein levels would ameliorate the patients’ clinical phenotype. We previously demonstrated in cellular models that the modified U1snRNA U1+5a restores mRNA splicing and coagulant function of coagulation factor VII (FVII) impaired by the F7 c.840+5G>A splicing mutation, which causes severe FVII deficiency. Aims: To demonstrate the U1+5a-mediated correction of the F7 c.840+5G>A mutation in mouse models. Methods: Mouse models of FVII deficiency caused by splicing mutations are not available. We created a novel mouse model of human FVII (hFVII) deficiency by liver-restricted expression, either transient (by hydrodynamic injection of plasmids) or prolonged (by adeno-associated viral [AAV] vectors), of the hFVII splicing-competent cassette harbouring the FVII c.840+5G>A mutation (FVII+5A) in C57BL/6 mice. The rescue was assessed by co-expression of the U1+5a, under the control of its own promoter. To avoid competition for AAV receptor binding, two AAV serotypes with liver tropism were used (AAV2-FVII+5A and AAV8-U1+5a). hFVII expression was evaluated by human-specific assays. Results: While delivery of plasmid pFVII+5A alone was ineffective, co-delivery of pFVII+5A with a molar excess (1.5X) of pU1+5a resulted in a significant increase of circulating hFVII levels (178±126 ng/ml), with a peak of 367ng/ml corresponding to 17% of pFVII-wt. This finding was corroborated by the appearance of hFVII-positive staining cells and of correctly spliced hFVII transcripts (26±10% of total transcripts) in mouse liver, thus indicating the ability of the U1+5a to efficiently re-direct usage of the mutated hFVII splicing site in vivo. To assess prolonged correction we exploited AAV. Mice were injected with 1.2x 1012 vector genomes (vg)/mouse of AAV2-FVII+5A alone or with the AAV8-U1+5a (1.2x1011 or 6x1011 vg/mouse). Noticeably, circulating hFVII antigen levels were appreciable only in mice treated with the AAV8-U1+5a. The U1+5a-mediated effect was dose-dependent, as measured by hFVII levels of 3.9±0.8 ng/ml (1.2x1011 vg/mouse) or 23.3±5.1 ng/ml (6x1011 vg/mouse) at two weeks post-injection. These findings were corroborated by the appearance of correctly spliced hFVII transcripts (4±0.5% or 16±3% of the total transcripts, respectively) and of hFVII-positive cells in mouse hepatocytes. Worth noting that in our experimental model hFVII expression results from hepatocytes simultaneously transduced by both AAV2-FVII+5A and AAV8-U1+5a. When mice were injected with an increased dose of template AAV2-FVII+5A (6x1012 vg/mouse) and the lowest AAV8-U1+5a dose, the correction effect (6.9±1.7 ng/ml) was more pronounced. This suggests that in FVII deficient patients, expressing the target F7 pre-mRNA in all hepatocytes, the rescue would be likely more robust. Conclusions: Altogether, we propose a novel methodology to model human FVII deficiency caused by splicing defects and to evaluate correction approaches in vivo. Our data provide the first in vivo proof-of-principle of the U1snRNA-mediated rescue of gene expression, and highlight its therapeutic potential in coagulation factor disorders.