Engineered ionizable lipid siRNA conjugates enhance endosomal escape but induce toxicity in vivo

Purpose: Lipid-conjugated siRNAs can reach extrahepatic tissues, but silencing efficacy remains lower than in liver largely because only ~1–2% of internalized siRNA escapes endosomes into the cytoplasm. Ionizable lipids enhance endosomal escape in lipid nanoparticles (LNPs), but direct covalent conjugation of an ionizable lipid to siRNA had not been investigated.

Hypothesis: Direct conjugation of an ionizable lipid analog, DLin-MC3-DMA, to siRNA will enhance endosomal escape and improve cytoplasmic availability/silencing, but may also induce toxicity or non-specific gene modulation in vivo, especially in tissues with high siRNA accumulation.

Aims: Develop a synthetic route to covalently attach a DLin-MC3-DMA analog to siRNA. - Characterize the physicochemical properties and RISC activity of DLin-MC3-DMA-conjugated siRNAs in vitro. - Evaluate whether the ionizable lipid conjugate enhances endosomal escape using a Gal8-YFP reporter system. - Assess tissue distribution, retention, target silencing, and toxicity/non-specific gene modulation after systemic administration in mice.

Delivery system: Platform: Covalent ionizable lipid–siRNA conjugate; not an LNP formulation. - Conjugate: DLin-MC3-DMA analog attached to the 3′ end of the siRNA sense strand via a custom functionalized CPG solid support. - Payload: Fully chemically stabilized siRNA targeting huntingtin (Htt) or cyclophilin B (Ppib); non-targeting control (NTC) also used. - Controls: Unconjugated siRNA and cholesterol-conjugated siRNA. - Physicochemical properties: DLin-MC3-DMA-siRNA showed similar hydrophobicity to cholesterol-siRNA by reverse-phase HPLC but formed aggregates/micelles (~8.3 nm by DLS), unlike unconjugated or cholesterol-siRNA (~2 nm). - Route: Subcutaneous injection at 20 mg/kg in mice.

Approach: In vitro: MDA-MB-231 cells stably expressing Gal8-YFP; RNAiMAX transfection; dose–response silencing of HTT; fluorescence microscopy for endosomal escape. - In vivo: 7-week-old female FVB/NJ mice; subcutaneous injection of PBS, NTC, Htt-targeting, or Ppib-targeting siRNAs at 20 mg/kg; n = 6 per group for silencing; n = 3 for distribution. - Tissue collection: 48 h post-injection for siRNA quantification/spatial distribution; 1 week post-injection for mRNA silencing. - Disease context: None; healthy mice only.

Key methods: Reverse-phase HPLC for relative hydrophobicity. - DLS and TEM for aggregation/self-assembly. - QuantiGene bDNA assay for Htt, Ppib, and Hprt mRNA levels. - Gal8-YFP fluorescence microscopy for endosomal disruption/escape. - PNA hybridization assay for siRNA antisense-strand quantification in tissues. - Fluorescence microscopy for intra-tissue spatial distribution. - One-way ANOVA with Dunnett correction for in vivo comparisons.

Key results: DLin-MC3-DMA conjugation retained RISC activity in vitro: IC50 values were 322 pM (unconjugated), 365 pM (cholesterol), and 481 pM (DLin-MC3-DMA). - Endosomal escape was enhanced: DLin-MC3-DMA-siRNA increased Gal8+ large-foci cells to 51% vs 30% for cholesterol and 35% for vehicle; Gal8-YFP intensity was ~130,000 FLU vs ~70,000 FLU for controls. - Tissue retention at 48 h: unconjugated siRNA ~15%; cholesterol-siRNA ~100%; DLin-MC3-DMA-siRNA ~80% of injected dose retained across 15 tissues. - DLin-MC3-DMA-siRNA accumulated in the vascular compartment of kidney and fat, unlike cholesterol-siRNA. - In tissues with >20 pmol/mg siRNA accumulation (liver, kidney, spleen), DLin-MC3-DMA-siRNA caused non-specific gene modulation, including downregulation of Hprt and target genes even with NTC. - In tissues with <20 pmol/mg (heart, fat, muscle), target silencing was observed without Hprt downregulation or NTC effects: ~35% in heart, ~60% in fat, ~30% in muscle. - Despite lower accumulation in heart and muscle than cholesterol-siRNA, DLin-MC3-DMA-siRNA produced better silencing: heart 37 ± 6% vs 10 ± 7%; muscle 30 ± 14% vs 17 ± 13%.

Interpretation: This is the first report of direct ionizable lipid conjugation to siRNA. The DLin-MC3-DMA conjugate enhances endosomal escape and can improve silencing in some extrahepatic tissues, but high tissue accumulation (>20 pmol/mg) causes non-specific gene modulation/toxicity. The authors conclude that engineering the conjugate structure is a promising strategy to alter intra-tissue and intracellular siRNA trafficking, but careful fine-tuning is needed to balance endosomal escape with safety.

Limitations: Toxicity/non-specific gene modulation at high tissue accumulation limits therapeutic potential of this specific conjugate. - No disease model or therapeutic efficacy endpoint; only reporter/target mRNA silencing in healthy mice. - Only subcutaneous administration was tested; no intravenous or other routes. - Only one ionizable lipid analog was evaluated; broader chemical space not explored. - In vitro and in vivo results diverged: endosomal escape/toxicity effects were tolerated in vitro but not in vivo. - No carrier-free internalization in vitro without a transfecting agent. - No detailed mechanism of toxicity or long-term safety data. - No large-animal validation.