(556g) Invited Talk: Next Generation siRNA-Lipid Conjugates for Enhanced Gene Targeting in the Brain after Central Administration

Neurodegenerative diseases are mostly incurable. The lack of available therapies for these diseases can be attributed in part to the challenge of delivering compounds across the blood-brain barrier (BBB), which effectively prevents all large biological drugs from entering the brain when its properties are intact.To circumvent the BBB, drugs can be injected into the cerebrospinal fluid (CSF), leading to distribution in brain regions connected to bulk CSF circulation. This is the most commonly used delivery route for oligonucleotide drugs, such as antisense oligonucleotides (ASOs) and short interfering RNAs (siRNAs), which are extremely promising agents for personalized medicine owing to their ability to silence disease-relevant genes. A major success story in this space is Nusinersen (Spinraza), an ASO approved to treat type 1 spinal muscular atrophy in infants. However, ASOs have thus far been ineffective against age-related neurodegenerative diseases, as exemplified by recent phase 3 clinical trial failures of ASOs targeting the HTT gene for treatment of Huntington’s disease. Interestingly, ASOs and siRNAs are commonly functionalized to contain phosphorothioate nucleotide linkages (in place of phosphodiester bonds), which replaces the hydroxyl group at the 2’ position of sugars with synthetic moieties such as 2′-O-methyl and 2′-deoxy-2′-fluoro; these modifications prevent nuclease degradation and increase tissue penetration. Yet, Tominersen—Roche’s ASO targeting HTT—had these modifications and still fell short of its clinical endpoints, likely in part from insufficient delivery into deep brain structures. Such failures highlight the need for innovation and novel strategies for oligonucleotide delivery to enable treatment of neurodegenerative diseases.

Prior work has shown that conjugation of oligonucleotide drugs—particularly siRNA—to hydrophobic lipid moieties can enhance biodistribution after intracerebroventricular (ICV) or intrathecal injection into CSF. The most well-studied conjugates are natural lipid moieties such as cholesterol and fatty acids, which have been explored for both systemic and CSF delivery approaches; however, lipid structures can be highly toxic in vivo, which ultimately limits clinical translation. We recently developed a novel diacyl lipid structure with 18-carbon stearyls and 18 EG repeats between the lipid and the branch point (termed EG18) that significantly increased serum half-life by binding to albumin, the most abundant protein in serum. Since albumin is also the most abundant protein in CSF and has been shown to traffic through perivascular spaces—which provides access to deep brain structures—we hypothesized that conjugation of siRNA to EG18 could improve biodistribution in brain and enhance gene silencing activity.

For experiments, we carried out ICV injections in wild-type C57BL/6 mice using free siRNA, siRNA conjugated to cholesterol (siRNA-chol), or siRNA conjugated to EG18 (siRNA-EG18). After 7 days, we assessed distribution patterns, regional gene knockdown, cell-specific uptake and gene knockdown, and toxicity using standard techniques and assays (fluorescence microscopy, Quantigene, and flow cytometry). Using confocal and light sheet microscopy, we confirmed that siRNA-EG18 likely traffics through perivascular spaces, whereas siRNA-chol is transported by bulk diffusion. We further determined that siRNA-chol and siRNA-EG18 were significantly more effective at silencing genes relative to free siRNA and a non-targeting control. However, siRNA-chol caused reactive microgliosis and focal disruption of the BBB, as indicated by extensive fibrinogen deposition in brain parenchyma even at a modest dose, whereas siRNA-EG18 exhibited no negative outcomes. At the cell level, siRNA-chol was taken up more effectively by different cells, but in certain cases, siRNA-EG18 was more effective at gene silencing. For example, in microglia (an attractive target for neuroinflammatory conditions), siRNA-chol had significantly more uptake than siRNA-EG18, yet siRNA-EG18 exhibited significantly better gene knockdown (>2-fold relative to siRNA-chol). These outcomes highlight the safety and potency of the siRNA-EG18 conjugate upon central administration.

Moving forward, we are currently assessing gene knockdown at later time points and quantifying conjugate uptake and knockdown potency across all cell types. We are also moving into transgenic models of Alzheimer's disease to assess the efficacy of gene knockdown in a more therapeutically relevant context. Overall, our results demonstrate the promise of siRNA-EG18 for brain-wide gene targeting and motivate its further exploration as a candidate for treating neurodegenerative diseases.