Multimeric Conjugates Using Engineered Peptide Scaffolds for Efficient siRNA Delivery

Abstract

Oligonucleotide therapeutics (ONT) traditionally involve a single targeting moiety per oligonucleotide when conjugated for organ delivery. Multimerization represents a novel approach by connecting multiple ONTs to a single scaffold, thereby influencing the drug's activity and biophysical properties in vivo. Recently, others have demonstrated the efficacy of this strategy, showing enhanced tissue retention and extended silencing with the capability to target multiple genes simultaneously. The investigation of diverse multimeric designs is thus an exciting opportunity to explore the delivery of the ONT. In this study, we engineered a versatile peptide branching unit able to link up to four small interfering RNAs together. We conjugated a GalNAc targeting moiety to these scaffolds for liver hepatocyte delivery and assessed their silencing activity. Our approach was further expanded to explore different peptide architectures (linear versus cyclized) and additional functionalities, including endosomal escape domains and dual target silencing. We then evaluated the constructs via subcutaneous and intravenous (i.v.) administration in mice. Notably, the intravenous administration of multimeric siRNA GalNAc demonstrated potent silencing in the liver and significantly affected liver-to-kidney biodistribution. Our findings suggest that peptides as branching units offer a promising pathway for ONT multimerization, advancing the challenges of drug delivery.

1

General synthetic strategy for the siRNA multimers with the dimer-siRNA illustrated as an example. (A) One-pot reaction undergoes both amide coupling between the ligand and the terminal lysine, and the SPAAC reactions between the ON and the azido-lysines core. (B) After isolation of the intermediate using HPLC, the guide strands are annealed to generate the target dimer (C).

2

Dose–response curves generated from the knockdown assays in (A) HEK293-ASGPR1 (48 h incubation) and (B) primary human hepatocytes (72 h incubation). (C) Summary of IC₅₀ and maximum knockdown at 10 μM. Data points represent mean and SD of n = 3.

3

(A) Sketch of linker cleavability on siRNA activity. (B) HEK293-ASGPR1 knockdown experiment of dimeric PPIB constructs, 48 h incubation. Comparison between different cleavable designs. Maximum knockdown was reached at a concentration of 10 μM. All curves are available in Figure S16.

4

(A) Sketch of the chimeric PPIB-SOD1 siRNA scaffold. (B) PPIB dose–response curve knockdown. (C) SOD1 dose–response curve knockdown. Each data point represents mean and SD of n = 3.

5

Top: Sketches of advanced multimer designs. PPIB knockdown of advanced multimer designs, in primary human hepatocytes, 72 h incubation. (A) Cyclic peptide designs. (B) Endosomal escape domains added to the tetrameric constructs. Data points represent mean and SD of n = 3.

6

Dose–response binding curves of multimeric constructs in mouse plasma. All construct concentrations were normalized based on the total amount of siRNA present in each construct.

7

(A) Mice study design and end points, n = 4 biological replicates per time point (total of 8 mice per arm). (B) PPIB antisense strand accumulation in liver and (C) kidney mice homogenates at 48 and 336 h. (D) Liver silencing. Statistics are one-way ANOVA performed with Tukey’s multiple comparisons. All mean KD are available in the Supporting Information. One outlier mouse was excluded from the i.v. administration group, PBS control, t = 48 h. ns: nonsignificant, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. Each data point represents the mean and SD of n = 4 mice. All silencing values can be found in Table S5.