Advanced siRNA Designs Further Improve In Vivo Performance of GalNAc-siRNA Conjugates

Abstract

Significant progress has been made in the advancement of RNAi therapeutics by combining a synthetic triantennary N-acetylgalactosamine ligand targeting the asialoglycoprotein receptor with chemically modified small interfering RNA (siRNA) designs, including the recently described Enhanced Stabilization Chemistry. This strategy has demonstrated robust RNAi-mediated gene silencing in liver after subcutaneous administration across species, including human. Here we demonstrate that substantial efficacy improvements can be achieved through further refinement of siRNA chemistry, optimizing the positioning of 2'-deoxy-2'-fluoro and 2'-O-methyl ribosugar modifications across both strands of the double-stranded siRNA duplex to enhance stability without compromising intrinsic RNAi activity. To achieve this, we employed an iterative screening approach across multiple siRNAs to arrive at advanced designs with low 2'-deoxy-2'-fluoro content that yield significantly improved potency and duration in preclinical species, including non-human primate. Liver exposure data indicate that the improvement in potency is predominantly due to increased metabolic stability of the siRNA conjugates.

Keywords: GalNAc conjugates; RNAi; siRNA.

Figure 1

Relative Impact of 2′-F at Each Position in the Antisense and the Sense Strands Based on a Multiple Linear Regression Model The y axis represents model-adjusted mean difference in target silencing, in natural log, of 2′-F from 2′-OMe-containing duplexes at a given position. (A and B) The x axis represents nucleotide position in the duplex, relative to the 5′ antisense (A) or 5′ sense strand (B). Negative numbers indicate activity improvement with inclusion of 2′-F relative to 2′-OMe at that position, positive numbers reflect decreased activity. Asterisks (*) indicate significant differences between 2′-F and 2′-OMe at the noted positions (p < 0.05). (C and D) Confirmation of effect in vitro across four siRNAs targeting the mouse transthyretin (Ttr) gene is depicted as fold-change relative to parent at either 10 nM (A) or 0.1 nM (B). 2′-F and 2′-OMe modifications are depicted as green and black squares, respectively.

Figure 2

Sense and Antisense Strand Optimization (A and B) Antisense (A) and sense (B) strand designs utilizing four siRNAs targeting the mouse transthyretin (Ttr) gene (Table S1) and the combination of best designs (C) evaluated across all 10 siRNAs (Table S1). Impact relative to parent is depicted as the model-adjusted mean difference in activity of design variant (DV) compared to parent, and it is presented in natural log. 2′-F and 2′-OMe modifications are depicted as green and black squares, respectively.

Figure 3

Application of Advanced Designs DV 18 and DV 22 to Multiple Sequences In vivo activity and duration of optimized DVs compared to parent designs post single 3 mg/kg s.c. administration of conjugates in mice (n = 3 per group and time point). Ttr expression was assessed in liver at 7 and 22 days post-dose. Data are presented as average percent of PBS-treated animals. Data were analyzed by two-way ANOVA, followed by Dunnett’s post-test for multiple comparisons. Results of post-test indicated as a single asterisk for significant compared to parent at day 7 and double asterisk for significant compared to parent at day 22. Error is represented as the SD.

Figure 4

Application of New Design DV 22 to siRNA-Targeting AT (A) Mice (n = 3 per group) were treated with a single dose (1 mg/kg) of parent or the DV 22-based derivative. Serum AT levels were assessed by ELISA. (B) Cynomolgus monkeys (n = 3 per group) were treated with a single dose (1 mg/kg) of parent or its DV 22-based derivative, and plasma AT levels were assessed by AT activity assay. Data were analyzed by two-way ANOVA, followed by Dunnett’s (A) or Sidak’s (B) post-test. Time points at which a significant difference was observed between parent and DV 22 are noted with an asterisk. Error is represented as the SD.

Figure 5

In Vivo Evaluation of Parent, DV 18, and Fully 2′-OMe Conjugates Targeting Ttr (A) Time course analysis of antisense strand liver levels (μg/g, microgram antisense strand per gram of liver) from mouse cohorts (n = 3/group/time) dosed with Parent (green line, 2.5 mg/kg), DV 18 (blue line, 0.75 mg/kg), or fully 2′-OMe (black dashed line, 0.75 mg/kg) conjugates targeting Ttr. (B–D) Time course analyses of Ttr mRNA knockdown (dashed black line) and sense (green line) and antisense (blue line) Ago2 loading (ng/g, nanogram sense or antisense strand per gram liver) for parent (B), DV 18 (C), and fully 2′-OMe (D) conjugates. Ttr mRNA levels were normalized to Gapdh mRNA for each animal and then to PBS control animals. (E) Comparison plot of Ago2-loaded antisense with Ttr mRNA knockdown for parent (green circles) and DV 18 (blue squares) conjugates. Semi-log lines of best fit are overlaid. Error is represented as the SD.