Albumin-binding dendrimer-conjugated siRNA enables safe and effective gene silencing throughout the central nervous system

Hassan H Fakih; Masahiro Ohara; Ashley Summers; Samantha L Sarli; Karen Kelly; Rosemary Gagnon; Bruktawit Maru; Brianna Bramato; Anastasia Khvorova; Jonathan K Watts

doi:10.1093/nar/gkag926

. 2026 Sep 30;54(18):gkag926. doi: 10.1093/nar/gkag926

Albumin-binding dendrimer-conjugated siRNA enables safe and effective gene silencing throughout the central nervous system

Hassan H Fakih ^1,^c, Masahiro Ohara ^2,^c, Ashley Summers ³, Samantha L Sarli ⁴, Karen Kelly ⁵, Rosemary Gagnon ⁶, Bruktawit Maru ⁷, Brianna Bramato ⁸, Anastasia Khvorova ^9,^✉, Jonathan K Watts ^10,^✉

¹ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

² RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

³ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

⁴ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

⁵ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

⁶ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

⁷ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

⁸ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

⁹ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

¹⁰ RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States

^✉

To whom correspondence should be addressed. Email: Jonathan.Watts@umassmed.edu

^✉

Correspondence may also be addressed to Anastasia Khvorova. Email: Anastasia.Khvorova@umassmed.edu

Hassan H Fakih and Masahiro Ohara the first two authors should be regarded as Joint First Authors.

Roles

Hassan H Fakih: Conceptualization, Data curation, Formal analysis, Investigation, Project administration, Validation, Visualization, Writing - original draft, Writing - review & editing

Masahiro Ohara: Conceptualization, Data curation, Formal analysis, Investigation, Validation, Visualization, Writing - original draft, Writing - review & editing

Ashley Summers: Investigation

Samantha L Sarli: Investigation

Karen Kelly: Investigation

Rosemary Gagnon: Investigation

Bruktawit Maru: Investigation

Brianna Bramato: Resources

Anastasia Khvorova: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Supervision, Writing - original draft, Writing - review & editing

Jonathan K Watts: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Supervision, Writing - original draft, Writing - review & editing

PMCID: PMC13624663 PMID: 42813369

Abstract

Improving siRNA delivery to the central nervous system (CNS) is a major focus for treating the numerous debilitating neurological conditions that have a genetic basis. Here, we present an albumin-binding siRNA based on an amphiphilic dendrimer conjugate (D-siRNA). We demonstrate that D-siRNA achieves effective and homogeneous delivery throughout the CNS following administration into the cerebrospinal fluid (CSF). In mice, a single CSF administration of D-siRNA resulted in potent and durable gene silencing across various brain regions, with effects lasting 6 months without detectable toxicity. We validate its utility in larger rodents (rats) using intrathecal administration—a clinically relevant route—showing effective and broad delivery and robust silencing. Benchmarking against other clinically relevant siRNA delivery scaffolds revealed that D-siRNA provides comparable delivery and efficacy, with more efficient conversion of gross uptake to functional uptake. These findings support the use of albumin-binding conjugates for brain delivery, and position D-siRNA as a safe, effective, and durable platform for gene silencing in the CNS.

Graphical Abstract

For image description, please refer to the figure legend and surrounding text.

Introduction

Oligonucleotide-based therapeutics, such as small interfering RNA (siRNA), antisense oligonucleotides (ASO), and splice-switching oligonucleotides, have been established as a productive therapeutic modality [1–3]. Of particular interest for gene silencing are siRNAs, which offer long-lasting durability, with some single doses providing an effect for up to a year, leading to eight approved drugs. However, currently approved siRNA therapies are limited to the liver, enabled by robust hepatocyte delivery via the N-acetylgalactosamine (GalNAc) conjugate [2]. Current research efforts are intensely focused on expanding the therapeutic potential of oligotherapeutics to extrahepatic tissues, such as muscle, heart, and the central nervous system (CNS) [4]. The CNS has drawn sharp focus because many severe neurological diseases have well-characterized pathology stemming from identifiable target genes that oligotherapeutics are ideally positioned to address [5, 6]. The CNS can be reached by intrathecal administration of oligonucleotides into the cerebrospinal fluid (CSF), which allows for broad distribution throughout brain regions as is necessary for many neurological and neurodegenerative diseases [6].

When ASOs and siRNAs are administered into the CSF, they are quickly cleared due to their small size and hydrophilicity, limiting their cellular uptake and deep brain region penetration [7–12]. Strategies to reduce clearance and improve wide distribution and uptake have mainly focused on increasing lipophilicity or cell-surface binding through chemical modification of the siRNA or lipophilic conjugation [7, 8, 13, 14]. However, lipophilic conjugation is a double-edged sword: while it can minimize clearance and improve cellular uptake [7], highly hydrophobic conjugates can limit diffusion from the injection site and most importantly can induce toxicity [15, 16].

Recently, we have reported on the development of an amphiphilic dendrimer conjugate for extrahepatic siRNA delivery (D-siRNA) [15, 17–20]. The conjugate’s chemical structure—featuring lipid moieties separated by phosphate groups—imparts its essential amphiphilic nature [17, 19]. This design successfully balances the need for improved lipophilicity to enhance siRNA circulation and cellular uptake with the need to avoid the toxicity frequently observed with highly lipophilic conjugates, driving effective siRNA activity in vivo [15, 19, 20]. Beyond its amphiphilic nature, the delivery benefits of D-siRNA may stem from its capacity for avid and selective albumin binding [19]. This feature is particularly interesting for CNS distribution, as albumin has been shown to distribute within the brain via the glymphatic system [21, 22], which has been shown to contribute to the movement of oligonucleotide therapeutics throughout the brain and spinal cord [11, 23]. We hypothesized that D-siRNA would achieve effective CNS distribution and transport following CSF administration, with improved distribution to various brain regions and low toxicity. Encouragingly, an independent study by Sorets et al. also recently reported that another albumin-binding ligand gave siRNAs a CNS distribution profile consistent with glymphatic system transport [24].

In this work, we explore and characterize the distribution, efficacy, and safety of the D-siRNA conjugate for CNS delivery. Specifically, we assessed D-siRNA following both intracerebroventricular (ICV) injection and the more clinically relevant intrathecal (IT) route. When benchmarked against the clinically relevant divalent siRNA technology (Di-siRNA) [8, 25, 26], D-siRNA demonstrated comparable distribution, efficacy, and safety. Following ICV injection, D-siRNA activity was long-lasting (up to 6 months) and showed potency against multiple targets. In rats, IT injection resulted in wide and effective distribution throughout the CNS, exhibiting no observed toxicity and efficacy lasting at least 1 month, with trends showing improved potency (lower IC₅₀) compared to Di-siRNA. Fluorescent imaging confirmed colocalization with blood vessels, supportive of perivascular transport as expected for glymphatic distribution. This work establishes the amphiphilic-conjugated D-siRNA platform that enables effective and safe oligotherapeutic delivery throughout the CNS.

Materials and methods

Synthesis and purification of oligonucleotides

Compounds were synthesized using a MerMade 12 (LGC Biosearch Technologies) synthesizer following standard protocols at 10–20 μmol scales. Standard RNA 2′-O-methyl, 2′-fluoro modifications were used to enhance siRNA stability (ChemGenes and Hongene Biotech). Di-branched sense strands were synthesized on custom-synthesized branchpoint-functionalized controlled pore glass (CPG) as previously described [8], and unconjugated and dendrimer sense strands were synthesized on CPG functionalized with UnyLinker (ChemGenes) as previously reported [19]. For the dendrimer sense strand, commercially available amidites (i.e. C6, C12, and symmetrical branching amidite from ChemGenes and Glen Research) were used to build the dendritic moiety on the 5′-end as previously described. All sense strands had a dT2 spacer between the oligonucleotide and the conjugate. Antisense strands were synthesized on CPG functionalized with a Unylinker (ChemGenes), and custom 5′-(E)-vinylphosphonate 2′-OMe-uridine CED phosphoramidite (Hongene) was applied to introduce 5′-(E)-vinylphosphonate.

All strands (antisense and sense) were deprotected and purified using 28% aqueous ammonia solution supplemented with 3% diethylamine at 50°C for 16–20 h, followed by drying under vacuum at 40°C, and resuspension in Millipore H₂O. Oligonucleotides were purified using an Agilent 1290 Infinity II HPLC (Agilent Technologies) on a C18 column for lipid-conjugated (dendrimer) sense strands and an ion-exchange column for the other strands. Purified oligonucleotides were desalted by size-exclusion chromatography and characterized by LC-MS analysis on an Agilent 6530 accurate-mass quadrupole time-of-flight (Q-TOF) LC/MS (Agilent Technologies). The sequences and modifications of the oligonucleotides are shown in Supplementary Table S1. HTT and APP sequences were acquired from Alterman et al. [8] and Sarli et al. [15], JAK1 from Tang et al. [27], and MECP2 from Hariharan et al. [28].

Duplex formation

Equimolar amounts of antisense and sense strands were prepared in water and incubated at 95°C for 5 min, cooled to room temperature, and stored at 4°C if not used immediately. The duplexes were then dried and resuspended in artificial CSF (aCSF, 137 mM NaCl, 5 mM KCl, 20 mM glucose, 8 mM HEPES, 2.3 mM CaCl₂, 1.3 mM MgCl₂, pH 7.4—for mice aCSF was supplemented up to 14 mM CaCl₂, 2 mM MgCl₂ [29]). Duplexes were resuspended to a concentration of 10 nmol/5 μl, 10 nmol/10 μl, or 80 nmol/70 μl, depending on the individual experiment. To validate efficient duplex formation, 20 pmol of duplex was loaded onto a non-denaturing 20% tris–borate–EDTA (TBE) gel (Invitrogen # EC63155BOX) and run at constant 180 V for 1 h. The gel was then washed with deionized water for 10 min, stained with 1× SYBR Gold Nucleic Acid Gel Stain (Invitrogen # S11494) for 10 min or 1× SYBR Safe Nucleic Acid Gel Stain (Invitrogen # S33102) for 15 min, and washed again. Bands were visualized on the ChemiDoc MP Imaging system (Bio-Rad # 17001402).

Gel-based albumin-binding assay

To evaluate albumin-binding properties of siRNA scaffolds, Cy3-labeled siRNAs were incubated with either ultrapure bovine serum albumin (BSA) (Invitrogen # AM2616) or human CSF (Innovative Research # IRHUCSF1ML Lot # 59979) before gel electrophoresis. siRNAs (10 pmol) were mixed with 10 μl of nuclease-free water, 0.3 mg/ml BSA or human CSF, and incubated at 37°C for 1 h. Then, 2 μl each of aCSF (formulation described above) and 30% sucrose were added to each sample, and the mixtures were loaded onto a 4%–20% gradient gel (Bio-Rad, Cat. #4561094). The final BSA concentration in the BSA/siRNA mixture was 0.2 mg/ml, and the final albumin concentration in the CSF/siRNA mixture was estimated to be ∼0.2 mg/ml based on a reported CSF albumin concentration of ~0.3 mg/ml [30]. Electrophoresis was performed in Tris–glycine running buffer (25 mM Tris and 192 mM glycine) at 100 V for 60 min at 4°C. The gel was visualized using a ChemiDoc imaging system (Bio-Rad) with a 2-s exposure time.

Animal experiments

All animal procedures were conducted according to the Institutional Animal Care and Use Committee (IACUC) protocols of the University of Massachusetts Chan Medical School (IACUC protocols 202000010 and 202100196) and in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Animals were housed and maintained in pathogen-free animal facilities at UMass Chan Medical School with a 12-h light/12-h dark cycle at a controlled temperature (23 ± 1°C), standard humidity (50% ± 20%), and free access to food and water.

Stereotactic intracerebroventricular injections in mice

ICV injections were performed as previously described [31]. Briefly, C57BL/6NJ male mice (Jackson Laboratory, Strain # 005304) aged 10–12 weeks were anesthetized with 2.5% isoflurane in oxygen (for Figs 1C, 3 and Supplementary Fig. S5) or by intraperitoneal injection of fentanyl/midazolam/dexmedetomidine (0.1, 5, and 0.25 mg/kg, respectively) (for Fig. 2 and Supplementary Figs S2–S4). Hair at surgical sites was shaved with an electric razor and a sterile surgical field was set up around the animal’s head. A burr was used to drill a small hole at the following coordinates relative to bregma: −0.2 mm posterior, 1.0 mm mediolaterally. The needle was then placed at a depth of −2.5 mm ventrally. A volume of 5 μl was injected per ventricle at a rate of 750 nl/min. Following injection, mice were monitored until sternal. All doses are based on molar amounts of active siRNA (antisense strand).

For image description, please refer to the Fig. legend and surrounding text. — Dendrimer-conjugated siRNA binds albumin in the CSF and shows high accumulation in the CNS. (A) Schematic of various siRNA scaffolds (unconjugated, divalent, and dendritic) assessed for brain distribution, with the chemical modifications used to stabilize the siRNA. (B) Gel-based binding assay of BSA and human CSF with various siRNA scaffolds. (C) The siRNA scaffolds were injected unilaterally in the lateral ventricle, followed by collection of various CNS tissues and siRNA quantification. All study groups are n = 3 mice/group. Statistical analysis was performed in GraphPad Prism: two-way ANOVA analysis followed by Tukey’s multiple comparisons across all groups (ns = non-significant, *P < .05, **P < .01, ***P < .001). Data represented as mean ± S.D.; tissue accumulation measured by PNA assay; drawing elements were created in BioRender. Fakih, H. (2026) https://BioRender.com/cmdmpre.

Evaluation of acute behavioral toxicity after ICV injections

Acute motor phenotypes after ICV injection of siRNAs were assessed using the Evaluation of Acute Drug-Induced NeuroToxicity (EvADINT) scoring system as previously described [32]. Mice were monitored by a blinded investigator over the first 24 h after injection. Scores were assigned based on seizure-like activity, hyperactivity or other atypical motor behaviors, and the time required to resume normal posture and behaviors, including sternal posture, unstimulated movement, movement without ataxia, and grooming/eating/nesting. Higher scores indicate more severe acute neurotoxicity.

Intrathecal delivery of oligonucleotides to rats

Twelve-week-old Sprague–Dawley rats (Charles River Laboratories, Strain # 001) were used for the experiments. Intrathecal administration was performed with slight modifications from previous reports [11]. Briefly, after isoflurane anesthesia, buprenorphine (0.1 mg/kg) was administered subcutaneously. Using a 50-ml conical tube, the rat was placed in the prone position with the spine flexed, and a skin incision was made. An incision was then made through the visible muscle layer, and the L6 lumbar vertebra was identified. A guide cannula (SAI Infusion Technologies) with an inserted 23-gauge needle (Becton Dickinson) was advanced anterior to the L6 vertebra and positioned in place. Subsequently, a catheter–wire assembly (SAI Infusion Technologies) was inserted through the guide cannula into the spinal column. The guide cannula and stylet wire were removed sequentially, leaving only the catheter in the spinal column. The catheter was gently pulled back until the black mark located 2 cm from the tip aligned with the spinal column. Thereafter, 70 μl of aCSF or siRNA was administered through the catheter tip over a period of >30 s. All surgical procedures were performed under sterile conditions. All doses are based on molar amounts of active siRNA (antisense strand). At the time of tissue harvest, rats in which the catheter was not present within the spinal column were considered to have received a misinjection and were excluded from the analysis.

Limb paralysis scoring after IT injection

Functional abnormalities following IT administration have been reported with oligonucleotide administration [33]. We assessed limb paralysis scoring 24 h (Supplementary Fig. S6) or 24 h and 6 days (Supplementary Fig. S7) after intrathecal injection of the Htt-siRNAs using a previously reported scoring system to evaluate potential motor phenotypes [34]. The examiner was blinded to the treatment allocation. In these studies, only scores 5–8 were observed: score 8, normal gait and range of motion in all limbs; score 7, paresis in one limb; score 6, paralysis in one limb or paresis in two limbs; and score 5, paralysis in one limb and paresis in one additional limb or paresis in three limbs.

mRNA quantification and tissue processing

mRNA expression was quantified with the QuantiGene Singleplex Assay Kit (Invitrogen # QS0016), a hybridization-based assay, as previously described [35, 36]. Cultured cells were lysed in a total volume of 250 μl diluted lysis mixture consisting of one part lysis mixture, two parts deionized H₂O and 0.167 mg/l proteinase K (Invitrogen # AM2548). Tissue lysates were thoroughly mixed by pipetting up and down 15–20 times. Probe set validation was performed for each tissue to determine an appropriate lysate volume that would allow unambiguous and linear mRNA detection. The corresponding volumes were transferred to the capture plate as indicated in the manufacturer’s protocol (Invitrogen # MAN0018628), and the protocol was followed further. Luminescence detection was performed using the SpectroMax M5 Microplate Reader (Avantor).

For mouse tissue: At sacrifice, the brain was harvested and placed in a brain matrix (Braintree Scientific). A 1-mm slice was then taken from the regions of interest and 1.5 × 1.5-mm punches were taken and placed in RNALater [935 ml MilliQ water, 700 g ammonium sulfate, 25 ml of 1 M sodium citrate, 40 ml of 0.5 M ethylenediaminetetraacetic acid (EDTA), pH 5.2, using H₂SO₄] for 24 h at 4°C, and then frozen at −80°C until processing for subsequent experiments. Punch biopsies were dried gently, weighed, and added to 500 μl of homogenizing solution (Invitrogen; QS0517) containing 0.2 mg/ml proteinase K (Invitrogen, #AM3546), in a QIAGEN collection microtube holding a 3-mm tungsten bead. The tissues were then homogenized for 10 min under 25 Hz of frequency using a QIAGEN TissueLyser II, followed by incubating at 55°C for 30 min. Lysate was then used as described above for mRNA quantification, or for siRNA accumulation (described below) or stored at −20°C. The following QuantiGene Singleplex RNA probes were used from Thermo Fisher Scientific: Htt (Murine, SB-14150), Hprt (Murine, SB-15463), Jak1 (Murine, SB-3029714), Mecp2 (Murine, SB-11697) Gfap (Murine, SB-14051), Iba-1 (Murine, SB-3027744).

For rat tissue: Rats were euthanized at the specified time point post-injection by CO₂ inhalation; cardiac perfusion with ice-cold phosphate buffered saline (PBS) was performed, followed by removal of the brain and spinal cord. Brains were oriented in an adjustable brain matrix (Braintree Scientific), and 2-mm sagittal sections were collected through the regions of interest; 3-mm biopsy punches were collected from the specific brain areas. Tissue punches were stored in RNAlater overnight at 4°C. For the evaluation of gene silencing, tissue punches were suspended in Affymetrix homogenizing solution containing proteinase K and mechanically dissociated using a Qiagen Tissuelyser with a 2-mm tungsten carbide bead. The tubes were then incubated in a water bath at 65°C until all tubes appeared transparent. Tubes were centrifuged (16 000 × g, 15 min) and supernatant transferred to a 96 well plate for storage at −80°C. The following QuantiGene Singleplex RNA probes were used from Thermo Fisher Scientific: Htt (Rat, SC-34753), Hprt (Rat, SC-14898), App (Rat, SC-15981).

For the evaluation of Iba1 and Gfap mRNA levels in rats shown in Fig. 5 and Supplementary Fig. S7, and in mice shown in Supplementary Fig. S3, tissue samples were homogenized in TRIzol reagent (Invitrogen) using a Qiagen Tissuelyser with a 3-mm tungsten carbide bead. RNA was extracted using RNA Clean & Concentrator Kit (Zymo Research, Cat. #R1016) and quantified on a NanoDrop (Thermo Fisher Scientific). For reverse transcription, 1 µg of RNA was used with the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Cat. #4368813) per the manufacturer’s protocol. Quantitative real-time PCR (qRT-PCR) was carried out in technical duplicates using iTaq Universal SYBR Green Supermix (Bio-Rad, Cat. #1725122) on Bio-Rad CFX-96 real time machine using gene-specific primers from Integrated DNA Technologies (IDT, Coralville, Iowa, USA): rat Gfap forward primer: 5′ GTTAAGCTAGCCCTGGACATC 3′; rat Gfap reverse primer: GATCTGGAGGTTGGAGAAAGTC; rat Iba1 forward primer: 5′ TCCGAGGAGACGTTCAGTTA 3′; rat Iba1 reverse primer: 5′ GTTGGCTTCTGGTGTTCTTTG 3′; rat Gapdh forward primer: 5′ ACAAGATGGTGAAGGTCGGTG 3′; rat Gapdh reverse primer: 5′ ACCATGTAGTTGAGGTCAATGAAGG 3′; mouse Gfap forward primer: 5′ CACCTACAGGAAATTGCTGGAGG 3′; mouse Gfap reverse primer: CCACGATGTTCCTCTTGAGGTG; mouse Iba1 forward primer: 5′ TCTGCCGTCCAAACTTGAAGCC 3′; mouse Iba1 reverse primer: 5′ CTCTTCAGCTCTAGGTGGGTCT 3′. Product assay ID for mouse Hprt is Mm.PT.39a.22214828.

siRNA tissue accumulation quantification

To quantify siRNA accumulated in tissues following administration, PNA hybridization assay was used on the lysate prepared for mRNA quantification, as previously described with slight modification [35–37]. All samples that were quantified were weighed to calculate siRNA accumulation per milligram of tissue. Briefly, the accumulation was quantified using custom Alexa488-labeled fluorescent PNA oligonucleotide probes that are fully complementary to the antisense strand of interest (PNABio). The probes used are Jak1 (Alexa488-OO- CATCAGCTACAAGCGAT), Mecp2 (Alexa488-OO-ATCTGACAAAGCTTCCCGATA), and App (Alexa488-OO-ATCAATTACCAAGAATTCTCA). To hybridize to the antisense strand of target siRNAs, lysates were annealed with the corresponding PNA probe (95°C for 5 min, followed by 5 min at 55°C and 5 mins at 4°C). Then, lysates were cleaned by precipitating out sodium dodecyl sulfate by adding 50 μl of 3 M potassium chloride followed by centrifugation for 15 min at 5000 × g. The clear supernatant containing the bound antisense strand was then collected. Anion-exchange chromatography was used to analyze the sample mixtures on an Agilent 1260 Infinity quad-pump HPLC with a 1260 FLD fluorescent detector, and quantification was based on comparison to standard curves of spike-in siRNA in tissue lysate as previously described [4].

Rat protein quantification with western blots

Rat brain samples were homogenized and lysed in ice-cold RIPA buffer (Boston BioProducts, Cat. #BP-115S) supplemented with Halt protease and phosphatase inhibitor cocktail (Life Technologies, Cat. #78441). Lysates were centrifuged at 12 000 rpm for 10 min at 4°C, and the supernatant was collected into fresh microcentrifuge tubes, leaving the pellet behind. Protein concentrations were determined using a BCA protein assay (Thermo Fisher Scientific, Cat. #23225). Equal amounts of protein (30 µg per sample) were mixed with 4× Laemmli SDS sample buffer (Thermo Scientific Chemicals, Cat. #J60015.AD) and boiled at 95°C for 5 min. Samples were vortexed and briefly centrifuged before electrophoresis. Proteins were separated on 4%–20% Tris–glycine SDS–PAGE gels (Bio-Rad, Cat. #4561 094) in running buffer (Bio-Rad, Cat. #1610732). Electrophoresis was performed at 50 V for 5 min to allow stacking, followed by 100 V for 1 h. Proteins were transferred to nitrocellulose membranes (Bio-Rad, Cat. #162-0115) using a semi-dry transfer apparatus. Transfer was carried out at 20 V for 1 h at room temperature. Then, membranes were blocked with Intercept blocking buffer (LI-COR, Cat. #927-60001) at room temperature for 1 h with agitation, followed by overnight incubation at 4°C with primary antibodies diluted in blocking buffer [1:1000 for APP (Cell Signaling, Cat. #2452S); 1:2000 for GAPDH (Sigma-Aldrich, Cat. #MAB374)]. After three washes with 1× TBST (10 min each), membranes were incubated with IRDye 680 anti-rabbit antibody for APP (1:5 000; LICORbio, Cat. #925-68073) and IRDye 800 anti-mouse antibody for GAPDH (1:5000; LICORbio, Cat. #92632210) for 1 h at room temperature. Membranes were then washed three additional times with 1× TBST. The bands were visualized using the LI-COR imaging system according to the manufacturer’s instructions.

Histology

Mice or rats were perfused with ice-cold PBS followed by ice-cold 4% paraformaldehyde (PFA) (Thermo Scientific Chemicals, Cat. #J19943-K2). Brains and spinal cords were post-fixed in 4% PFA at 4°C overnight, and subsequently cryoprotected in 30% sucrose at 4°C overnight. For experiments examining brain distribution in rats, brain hemispheres were initially preserved in RNAlater and stored at −80°C, followed by processing using the same protocol described above. Tissues were embedded in OCT compound (Sakura Finetek USA), and frozen sections were cut at a thickness of 20 µm using a cryostat (Thermo Fisher). Sections were mounted onto glass slides (MATSUNAMI, Cat. # SUMGP11). In experiments to assess siRNA distribution, tissue sections were air-dried at room temperature, washed three times with PBST, and coverslipped using a mounting medium containing DAPI (VECTOR laboratories, Cat. #H-1200-10). For immunofluorescence, sections were washed three times with 1× TBST and incubated in 5% goat serum in TBST for 60 min at room temperature to block nonspecific binding. Samples were then incubated with anti-CD31 antibody (1:200; Cell Signaling Technology, Cat. #15585T) or anti-GFAP antibody (1:200; Cell Signaling Technology, Cat. #80788S) at 4°C overnight, followed by three washes with TBST. Subsequently, sections were incubated with Alexa Fluor 488-conjugated anti-rabbit IgG secondary antibody (1:500 for CD31 staining and 1:1000 for GFAP staining; Cell Signaling Technology, Cat. #4412S) diluted in 5% goat serum/TBST for 1.5 h at room temperature. After three additional washes with TBST, sections were mounted using a DAPI-containing mounting medium (VECTOR laboratories, Cat. #H-1200-10). For the quantitative evaluation of GFAP-stained area, we used the Analyze Particle tool in ImageJ as previously described [38]. Regions of interest (333 × 333 µm) were placed in the left and right anterior horns of the lumbar spinal cord, and the stained areas were compared.

Spatial distribution of siRNA in the brain with line-profile analysis

Quantitative line-profile analysis for brain images after ICV injection of 10 nmol siRNAs was performed using original fluorescence images. Images were analyzed in Fiji/ImageJ using the siRNA fluorescence channel. For each brain image, a 200-µm-wide line ROI was manually drawn from the lateral ventricle/injection site toward the cortex over a distance of 1.0 mm, or from the lateral ventricle/injection site toward the cerebellum over a distance of 2.5 mm. Fluorescence intensity along each ROI was measured using the Plot Profile function in Fiji. The resulting intensity profiles were exported and plotted as fluorescence intensity versus distance from the lateral ventricle.

Blood diagnostics

Blood sampling for blood diagnostics was performed when mice were terminated at 24 h post injection, via cheek bleed. In total, 200 μl of blood was collected in a lithium heparin-coated BD Microtainer tube (BD, #365965) for blood chemistry test. Hundred microliters of blood was collected in a K2 EDTA-coated BD Microtainer tube (BD, #365974) for a complete blood count (CBC) test. Blood chemistry and CBC diagnostics were conducted by the Diagnostics Laboratory in the Department of Animal Medicine at UMass Chan Medical School.

Graphs and statistical analyses

Data were analyzed using GraphPad Prism 10.1.2 software for Windows (GraphPad Software, Inc., San Diego, CA). For each independent experiment, the levels of mRNA silencing were normalized to the mean of the control NTC (non-targeting control) group. Data were analyzed using a one-way or two-way ANOVA with post-hoc multiple comparisons as specified in the figure captions. To assess differences in overall target mRNA knockdown across CNS regions among groups, data were analyzed by two-way ANOVA, followed by Tukey’s multiple comparisons test. Within each CNS region, treatment groups were compared with the control group using Dunnett’s multiple comparisons test. Statistical significance for specific pairwise comparisons between groups was determined using adjusted P-values. IC₅₀ and IC₇₀ values were obtained from four-parameter logistic fits (nonlinear regression), with bottom and top constrained to 0 and 100, respectively, and the “Find ECanything” function was used to determine the concentrations corresponding to 50% and 70% knockdown on the fitted curves. Asterisks (*) denote gene expression significance (*P < .05, **P < .01, ***P < .001, ^****P < .0001). Graphs are plotted as mean ± standard deviation.

Results

D-siRNA, but not Unc- and Di-siRNA, binds albumin in CSF

We previously showed that dendrimer-siRNA (D-siRNA) predominantly binds albumin in plasma [19]. While CSF also contains albumin, its concentration is typically ∼100-fold lower in CSF than in plasma [30]. To test whether albumin binding to D-siRNA is still relevant under these conditions, we assessed albumin binding via a gel-based binding assay using Cy3 labeled siRNAs, comparing D-siRNA performance to that of standard unconjugated siRNAs (Unc-siRNA) and clinical-stage divalent siRNA (Di-siRNA) (Fig. 1A and B). It is worth noting that all the tested scaffolds contain a 5-nucleotide single-stranded, fully phosphorothioate-modified tail (or two tails in the Di-siRNA compound), which are known to contribute significantly to protein binding independently from any conjugate [8, 39].

In the study, we used human CSF (estimated final albumin concentration 0.2 mg/ml), as well as an additional control system consisting of a solution of 0.2 mg/ml BSA in buffer. This corresponds to ~45 pmol of albumin in a 15 μl reaction, to which we added 10 pmol of siRNA. In the siRNA-only lanes, each siRNA migrated as a free oligonucleotide band. After incubation with BSA, Unc-siRNA and Di-siRNA did not show a detectable shifted Cy3-positive band, whereas D-siRNA produced an additional band, consistent with the formation of an albumin-associated complex. Similarly, after incubation with human CSF, a Cy3-positive band with similar mobility to the BSA/D-siRNA band was observed only for D-siRNA. These results indicate that D-siRNA associates with albumin and suggest that this interaction is relevant at the concentrations of albumin observed in healthy human CSF.

D-siRNA achieves brain-wide distribution in mice following administration into the CSF

To evaluate the CNS distribution of different siRNA chemistries, mice received a unilateral ICV injection of 10 nmol of Cy3-labeled, fully chemically modified Unc-siRNA, Di-siRNA, or D-siRNA (equivalent to 135 µg of Cy3-labeled siRNA) (Fig. 1A). This unilateral injection allowed for the assessment of distribution to both hemispheres and the broader CNS parenchyma (Fig. 1C). Twenty-four hours post-injection, the accumulation of siRNA was quantified in the ipsilateral and contralateral hemispheres, cerebellum, and spinal cord. Both Di- and D-siRNA significantly enhanced siRNA distribution compared to Unc-siRNA across all collected tissues (***P < .001), with the exception of the spinal cord, where the enhancement did not reach statistical significance. Distribution levels of Di- and D-siRNA were generally comparable throughout the CNS, though D-siRNA showed an improvement in the contralateral hemisphere and cerebellum compared to Di-siRNA (400 versus 200 fmol/mg *P < .05 and 400 versus 200 fmol/mg **P < .01, respectively). We also assessed whether such administration has any impact on CBC or blood chemistry as readouts of any potential systemic toxicity and found no alterations to any of the markers measured compared to controls (Supplementary Fig. S1).

Fluorescence imaging of brain sections further confirmed the dendrimer conjugate’s ability to modify both the distribution and retention of the siRNA cargo relative to Unc-siRNA and Di-siRNA (Fig. 2A and Supplementary Fig. S2). When co-staining for CD31 (blood vessels), a correlation was observed for all siRNA scaffolds; however, the extent of parenchymal signal differed among the constructs, with Di-siRNA showing the strongest parenchymal signal, followed by D-siRNA and then Unc-siRNA (Fig. 2B and C). We next evaluated the spatial distribution of the signal within the brain (Fig. 2C). Di-siRNA showed prominent periventricular accumulation, with signal intensity gradually decreasing farther from the ventricle. In contrast, D-siRNA showed a more uniform signal distribution across the analyzed brain regions. Collectively, these data demonstrate that the albumin-binding D-siRNA achieves wide CNS distribution, comparable to that of clinical-stage divalent siRNA chemistry and representing a significant improvement relative to unconjugated siRNAs.

D-siRNA supports safe, potent, and durable activity in the CNS

Following the distribution assessment, we evaluated the efficacy and safety of D-siRNA in the mouse CNS. Mice were bilaterally injected (ICV) with a total dose of 10 nmol (5 nmol/ventricle) of an NTC (equivalent to 129 µg of siRNA) or previously developed Huntingtin (Htt)-targeting sequence [8, 15] (equivalent to 128 µg of siRNA), comparing Unc-, Di-, or D-siRNA chemistries (Fig. 3A). Efficacy was assessed by quantifying Htt mRNA expression, while toxicity was monitored using Iba1 and Gfap mRNA expression [8]. This was done both at 2 months (Fig. 3B) and 6 months post injection (Fig. 3C) to evaluate durability.

At 2 months post-injection (Fig. 3B), region-by-region analyses showed that all chemistries achieved similarly robust silencing of Htt mRNA (∼50%–60% silencing) compared to the NTC control in each CNS region evaluated (two-way ANOVA; Fig. 3B, left panel). In a separate analysis comparing the overall silencing effects of three siRNA chemistries across all CNS regions, Di-siRNA showed significantly stronger silencing compared to Unc- and D-siRNA (two-way ANOVA; ***P < .001, **P < .01, respectively), whereas no significant difference was observed between Unc-siRNA and D-siRNA (two-way ANOVA; P = .93). Analysis of toxicity indicators Iba1 and Gfap in the thalamus and hippocampus—regions closest to the injection site with the highest siRNA accumulation and activity [8]— showed no observed changes, supporting the safety of D-siRNA (Fig. 3B, right panel).

By 6 months post-injection, region-by-region analyses showed that Unc-siRNA lost its activity in all brain regions, whereas both Di- and D-siRNA maintained significant silencing in most regions (two-way ANOVA; Fig. 3C). In the striatum, medial cortex, and brainstem, only the D-siRNA maintained significant Htt mRNA silencing at 6 months (∼25%, **P < .01 in the striatum and *P < .05 in the medial cortex and brainstem compared to NTC) (Fig. 3C, left panel). When comparing the overall silencing effects across all CNS regions, both D-siRNA and Di-siRNA were significantly more potent than Unc-siRNA (two-way ANOVA; ^****P < .0001 for both comparisons). No significant difference was observed between D- and Di-siRNA (two-way ANOVA; P > .99). Toxicity markers remained unchanged across all the groups, further supporting the long-term safety of D-siRNA with a profile similar to that of Di-siRNA (Fig. 3C, right panel).

To further evaluate the safety and tolerability of D-siRNA, mice received a high dose of 30 nmol by bilateral ICV injection (equivalent to 405 µg of siRNA), followed by acute toxicity assessment and evaluation of neuroinflammation markers (Supplementary Fig. S3). We also did not observe any significant changes for the D-siRNA, further supporting its safety profile even at exaggerated doses (Supplementary Figs S3 and S4). Overall, the data demonstrated comparable silencing durability and safety of D-siRNA to that of Di-siRNA, with durability enhancement compared to Unc-siRNA, which lost efficacy at 6 months post injection.

To confirm the broad applicability of the D-siRNA platform, the dendrimer was conjugated to siRNAs targeting Jak1 (Fig. 3D) and Mecp2 (Supplementary Fig. S5), both using previously identified cross-species-reactive sequences [27, 28]. Following a bilateral ICV injection (10 nmol total dose) (equivalent to 136 µg of Jak1-targeting siRNA), the activity of D-siRNA was compared to Di-siRNA 1-month post-injection (Fig. 3D, right panel). Both constructs demonstrated robust silencing efficacy, achieving ~75% Jak1 mRNA silencing across all regions (Fig. 3D, right panel).

Assessment of JAK1-siRNA accumulation showed no significant difference in accumulation between D-siRNA and Di-siRNA across various brain regions at 1 month post-injection (Fig. 3E, left panel). While at early timepoints (Fig. 1C; 24 h) D-siRNA showed increased accumulation in some brain regions, this difference disappears at the longer 1-month timepoint (Fig. 3E; 1 month). The high accumulation at early timepoints (24 h) may reflect siRNA that is albumin-bound but not yet fully internalized; once this is cleared, there is no accumulation advantage of the D-siRNA over Di-siRNA. This is consistent with the idea that the association with albumin may drive an extended CSF retention [24].

Plotting the Jak1 mRNA silencing versus siRNA accumulation allowed for a direct comparison of the PK/PD relationship between the two chemistries (Fig. 3E, right panel). Both constructs showed a strong PK/PD correlation, but D-siRNA achieved a significantly lower IC₇₀ (135.7 fmol/mg with 95% CI 101–176.1) compared to Di-siRNA (354.6 fmol/mg with 95% CI 235.3–609.0), indicating enhanced delivery efficiency (i.e. conversion of gross uptake to functional uptake). Similar results were obtained with Mecp2-targeting siRNAs where both chemistries achieved potent efficacy of ∼75% silencing across all regions (130 µg of Mecp2-targeting siRNA) (Supplementary Fig. S5).

D-siRNA enables safe and brain-wide distribution via intrathecal administration in rats

Based on the promising ICV results in mice, we next evaluated intrathecal (IT) administration, a delivery method more widely used in the clinical setting for oligonucleotides in the CNS [2]. Although ICV dosing can achieve efficacy in deep brain regions due to these areas being close to the injection site, IT delivery has been reported to have limited access to deep structures [12, 40]. Therefore, we assessed whether D-siRNA would retain distribution and efficacy across the CNS following IT administration. To first examine CNS distribution, we administered Cy3-labeled siRNAs intrathecally to rats (50 nmol, Unc-, Di- and D-siRNAs) (equivalent to 677 µg of siRNA) and evaluated their distribution in the brain and spinal cord 48 h post injection (Fig. 4A).

In the brain, the Unc-siRNA showed weaker signal than either Di- or D-siRNA. In the representative image, the Di-siRNA showed signal in the brain parenchyma, in addition to signal along the pia mater and perivascular structures (Fig. 4B). The D-siRNA appeared to show relatively stronger signal along perivascular structures (Fig. 4B), a pattern consistent with previous reports of albumin-related transport of appropriately conjugated siRNAs through perivascular pathways [24]. A similar pattern was observed in the lumbar spinal cord: D-siRNA produced signal levels comparable to Di-siRNA and stronger than Unc-siRNA (Supplementary Fig. S6). Quantification of siRNA accumulation demonstrated that D-siRNA accumulated most strongly in the lumbar spinal cord and, overall, tended to accumulate similarly to Di-siRNA (Fig. 4C).

For toxicity assessment, we evaluated GFAP immunofluorescence in the lumbar spinal cord, the region with the highest D-siRNA accumulation, as it has been reported to transiently increase with oligonucleotide administration [8]. No obvious differences in GFAP staining area were observed between groups (Supplementary Fig. S6B and C). Furthermore, motor scoring at 24 h post-injection revealed no detectable functional abnormalities (Supplementary Fig. S6D).

Robust knockdown of App with intrathecally administered D-siRNA

Based on the favorable safety profile and distribution of D-siRNA following IT administration, we then assessed the silencing efficacy using siRNAs targeting a disease-relevant gene. We compared the pharmacodynamics and pharmacokinetics of siRNAs targeting amyloid beta (Aβ) precursor protein gene (App), a target for which siRNAs are already in clinical development as a treatment for Alzheimer’s disease [7, 41]. App-targeting Unc-, Di-, and D-siRNAs were IT administered at 80 nmol dose of each compound into rats and evaluated readouts at 1 month post injection (equivalent to 1030 µg of NTC-siRNA, and 1035 µg of App-targeting siRNA) (Fig. 5A). D-siRNA showed robust target gene silencing of ~50%–70% across almost all brain regions (Fig. 5B). Notably, significant knockdown was observed even in deep brain regions in D-siRNA-treated rats, whereas Unc-siRNA showed limited silencing (two-way ANOVA; Fig. 5B). When comparing the overall silencing, D-siRNA demonstrated a statistically higher knockdown compared to Unc-siRNA (two-way ANOVA; ^****P < .0001), whereas comparable to Di-siRNA (two-way ANOVA; P = .99). At the protein level, D-siRNA achieved a significant reduction of APP throughout the cortex, hippocampus, and brainstem (Fig. 5C).

In terms of siRNA accumulation, both Di-siRNA and D-siRNA showed an overall higher accumulation than Unc-siRNA (two-way ANOVA across all CNS regions; ^****P < .0001 for Di- and D-siRNA versus Unc-siRNA), though the differences were statistically significant only in some individual brain regions (two-way ANOVA; Fig. 5D). There was no significant difference in siRNA accumulation across all CNS regions between Di- and D-siRNA (two-way ANOVA; P = .95). Interestingly, when PK/PD relationships were evaluated—excluding the cerebellum, which shows low silencing with high accumulation as previously reported [42, 43]—both D-siRNA and Di-siRNA showed substantially better silencing per tissue accumulation than Unc-siRNA. D-siRNA showed the highest efficiency of converting gross uptake to functional silencing, with an in-tissue IC₅₀ of 50.8 fmol/mg tissue; 95% CI: 6.6 – 94.1 (Fig. 5E). This result is in broad concordance with the data observed by ICV injection in mouse (Fig. 3). No evidence of neuroinflammation was observed based on (Iba1 and Gfap) mRNA levels (Fig. 5F).

The broad applicability of the D-siRNA was further confirmed by silencing Htt as an additional target. Cy3-labeled, Htt-targeting siRNAs administered intrathecally to rats at a relatively low dose (50 nmol) (equivalent to 675 µg of NTC-siRNA, and 677 µg of Htt-targeting siRNA) showed moderate but overall significant efficacy at 1 month post injection (Supplementary Fig. S7A and B—two-way ANOVA, ^****P < .0001 for Di-siRNA and D-siRNA relative to NTC-siRNA). Importantly, these compounds did not induce any obvious motor phenotypes at 24 h or 6 days after administration (Supplementary Fig. S7C). Furthermore, neuroinflammation markers (Iba1 and Gfap) showed no changes between groups, even in the lumbar spinal cord where accumulation was the highest (Supplementary Fig. S7D), at 1-month post injection.

Discussion

Oligonucleotide therapeutics show promise to treat neurological diseases in the CNS. However, the CNS is a delicate and sensitive system, and some modifications and conjugates that work for systemic administration show toxicity when delivered to the CNS [29, 32, 44]. Lipophilic conjugation has been shown to improve PK/PD of oligotherapeutics [7, 14, 45], but highly lipophilic conjugates tend to increase toxicity (seizure-like phenotypes) and limit brain distribution when administered in the brain parenchyma [15, 16]. Lipophilic conjugates with more moderate hydrophobicity appear to have an improved safety profile and wider brain distribution [7, 15]. Here, we showed that an amphiphilic conjugate that has previously been demonstrated to selectively and avidly bind albumin in plasma [19], also binds albumin in CSF, exhibits favorable PK/PD properties for CNS delivery, and is well-tolerated in rodents. Another albumin-binding ligand was also recently shown to have promising delivery properties in the CNS [24].

D-siRNAs demonstrated a promising therapeutic profile in the CNS, comparable to the divalent siRNA chemistry, which has already received IND approval [26] and is characterized by broad distribution and extended durability without compromising safety [8]. The ability of D-siRNA to achieve significant contralateral accumulation following unilateral ICV administration suggests that the conjugate overcomes the diffusion limitations often observed with highly lipophilic modifications [16]. This improved parenchymal permeation may result in part from the albumin-binding capacity of the dendrimer, which may exploit glymphatic transport for wider convective distribution, a mechanism consistent with other albumin-binding constructs [24] and supported in our work by the observed strong colocalization with blood vessels. This is consistent with the high CNS concentration of D-siRNA at the 24 h timepoint in whole hemispheres relative to the other modifications (Fig. 1C). However, this likely reflects siRNA that is not fully internalized. Thus, although the albumin concentration in the CSF is significantly lower than that in plasma, our data, together with the observations of others [24], suggest that albumin binding may influence siRNA distribution in the CNS.

Looking across tissue sections at longer timepoints, the Di-siRNA shows the highest uptake in cells closest to CSF, while the D-siRNA seems to show lower but more even distribution moving away from ventricles (Fig. 2C). While dendrimer and divalent chemistries exhibited comparable accumulation with a trend of slightly lower accumulation with dendrimer, the level of knockdown was comparable between the two. This suggests that D-siRNA more efficiently converts gross tissue uptake to functional uptake, as reflected in the lower in-tissue IC₅₀ and IC₇₀. At much later timepoints, the technologies appear comparable (Fig. 3C), consistent with the idea that over a timescale of months, gross uptake can be converted to productive (functional) uptake and that this efficiency advantage becomes less important.

We don’t know whether the efficient delivery observed here is related to the albumin-binding ability of the dendrimer conjugate or to other productive interactions with cell-surface or endosomal biomolecules/structures, though the known ability of albumin to be rescued from endosomal degradation may be relevant [46]. Others have reported on albumin-binding conjugates that enhanced delivery and knockdown in the brain following CSF administration, further supporting the potential utility of such chemical scaffolds for CNS delivery of oligotherapeutics [24]. The mechanistic details of uptake enhancement by albumin binders are worth pursuing in future studies.

The translational potential of this platform is further supported by intrathecal (IT) administration in rats, more closely mirroring clinical protocols. The transition to a larger rodent species (∼five-fold brain volume increase) did not diminish efficacy; D-siRNA maintained potent silencing of the Alzheimer’s target App across the spinal cord and brain regions. Consistent with ICV findings, IT administration yielded a lower in-tissue IC₅₀ compared to divalent siRNAs, reinforcing the hypothesis that the amphiphilic dendrimer improves the efficiency of conversion of gross uptake to functional uptake. Crucially, the absence of neurotoxicity markers (neuroinflammation or motor deficits) across both routes confirms that this enhanced potency does not come at the cost of the safety profile.

As the landscape of RNA-based neurotherapeutics evolves, there is a distinct drive toward modalities that enable effective brain delivery via systemic administration. The most advanced of these technologies are transferrin receptor-targeting antibody-oligonucleotide conjugates, which facilitate transport of oligonucleotides across the blood–brain barrier [47, 48]. While these systemic platforms offer a clear administration advantage over lipophilic conjugates and other distribution-enhancing agents requiring direct CNS access, they will also come with new risks, and intrathecal administration remains a clinically viable and highly validated route. When paired with wide-distributing chemistries—such as the D-siRNA discussed here—intrathecal delivery provides a proven path to the clinic that maximizes CNS exposure while minimizing systemic off-target tissue risks.

In conclusion, this work showcases the applicability of the albumin-binding dendrimer conjugate as a safe and effective delivery tool for siRNAs in the CNS. Collectively, the data support safe, wide distribution and efficacious delivery throughout the brain and spinal cord, offering an alternative delivery technology for gene silencing in the brain for the treatment of neurological diseases.

Supplementary Material

gkag926_Supplemental_File

gkag926_supplemental_file.pdf^{(8.2MB, pdf)}

Acknowledgements

We thank Khvorova/Watts lab members (past and present) for insightful discussions and support. We thank Hanadi F. Sleiman for the initial collaboration on dendrimer conjugate for siRNA delivery, and the productive insights and discussions. We thank the animal medicine core at UMass Chan Medical School for their help and support on multiple studies. All figures and the Graphical abstract were created with PowerPoint and BioRender; all BioRender figure elements are available at https://BioRender.com/cmdmpre.

Author contributions: H.H.F., M.O., A.K., and J.K.W. conceived the project. H.H.F., M.O., A.K., and J.K.W. contributed to the experimental design. H.H.F., M.O., A.S., S.L.S., K.K., R.G., B.M.B., and B.M. contributed experimentally. H.H.F., R.G., B.M.B., and B.M. synthesized all siRNA compounds. H.H.F., M.O., and J.K.W. co-wrote the manuscript. All authors provided feedback and approved of the manuscript.

Notes

Present address: Samantha L. Sarli, Boston Children’s Hospital, Harvard University, Boston, MA 02115, United States

Contributor Information

Hassan H Fakih, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Masahiro Ohara, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Ashley Summers, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Samantha L Sarli, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Karen Kelly, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Rosemary Gagnon, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Bruktawit Maru, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Brianna Bramato, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Anastasia Khvorova, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Jonathan K Watts, RNA Therapeutics Institute, University of Massachusetts Chan Medical School, Worcester, MA 01605, United States.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

The authors have filed patent applications related to this work. H.H.F. and J.K.W. are on the SAB of M2DS therapeutics.

Funding

The authors acknowledge support from the NIH (R01 NS111990 to J.K.W., R35 GM131839, S10 OD020012, and S10 OD036329 to A.K.). The open access publication charge for this paper has been waived by Oxford University Press – NAR Editorial Board members are entitled to one free paper per year in recognition of their work on behalf of the journal.

Data availability

All data are presented in the main text and the supplementary data, with raw data available from the corresponding authors upon request.

References

1. Belgrad J, Fakih HH, Khvorova A. Nucleic acid therapeutics: successes, milestones, and upcoming innovation. Nucleic Acid Ther. 2024;34:52–72. 10.1089/nat.2023.0068 [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Egli M, Manoharan M. Chemistry, structure and function of approved oligonucleotide therapeutics. Nucleic Acids Res. 2023;51:2529–73. 10.1093/nar/gkad067 [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Watts JK. Nucleic Acid Therapeutics. In: Blackburn GM, Egli M, Gait MJ, Watts JK (eds.), Nucleic Acids in Chemistry and Biology. Cambridge, UK, The Royal Society of Chemistry, 2022, pp.350–402. 10.1039/9781837671328: 10.1039/9781837671328-00350 [DOI] [Google Scholar]
4. Tang Q, Khvorova A. RNAi-based drug design: considerations and future directions. Nat Rev Drug Discov. 2024;23:341–64. 10.1038/s41573-024-00912-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Hu B, Zhong L, Weng Y et al. Therapeutic siRNA: state of the art. Signal Transduct Target Ther. 2020;5:101. 10.1038/s41392-020-0207-x [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Mansour HM, El-Khatib AS. Oligonucleotide-based therapeutics for neurodegenerative disorders: focus on antisense oligonucleotides. Eur J Pharmacol. 2025;998:177529. 10.1016/j.ejphar.2025.177529 [DOI] [PubMed] [Google Scholar]
7. Brown KM, Nair JK, Janas MM et al. Expanding RNAi therapeutics to extrahepatic tissues with lipophilic conjugates. Nat Biotechnol. 2022;40:1500–8. 10.1038/s41587-022-01334-x [DOI] [PubMed] [Google Scholar]
8. Alterman JF, Godinho BMDC, Hassler MR. A divalent siRNA chemical scaffold for potent and sustained modulation of gene expression throughout the central nervous system. Nat Biotechnol. 2019;37:884–94. 10.1038/s41587-019-0205-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Boyd MM, Mazur C, Pribnow K et al. Comparison of MALAT1 antisense oligonucleotide distribution following intracerebroventricular and lumbar intrathecal routes of administration. Mol Ther Nucleic Acids. 2025;36:102742. 10.1016/j.omtn.2025.102742 [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Cook BE, McLaren DG, Sullivan JM et al. Central nervous system biodistribution and pharmacokinetics of radiolabeled tofersen in rodents, nonhuman primates, and humans. J Nucleic Med. 2026;67:139–44. 10.2967/jnumed.125.270731 [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Mazur C, Powers B, Zasadny K. Brain pharmacology of intrathecal antisense oligonucleotides revealed through multimodal imaging. JCI Insight. 2019;4:e129240. 10.1172/jci.insight.129240 [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Jafar-nejad P, Powers B, Soriano A et al. The atlas of RNase H antisense oligonucleotide distribution and activity in the CNS of rodents and non-human primates following central administration. Nucleic Acids Res. 2020;49:657–73. 10.1093/nar/gkaa1235 [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Wolfrum C, Shi S, Jayaprakash KN et al. Mechanisms and optimization of in vivo delivery of lipophilic siRNAs. Nat Biotechnol. 2007;25:1149–57. 10.1038/nbt1339 [DOI] [PubMed] [Google Scholar]
14. Osborn MF, Coles AH, Biscans A et al. Hydrophobicity drives the systemic distribution of lipid-conjugated siRNAs via lipid transport pathways. Nucleic Acids Res. 2019;47:1070–81. 10.1093/nar/gky1232 [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Sarli SL, Fakih HH, Kelly K et al. Quantifying the activity profile of ASO and siRNA conjugates in glioblastoma xenograft tumors in vivo. Nucleic Acids Res. 2024;52:4799–817. 10.1093/nar/gkae260 [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Alterman JF, Hall LM, Coles AH et al. Hydrophobically modified siRNAs silence Huntingtin mRNA in primary neurons and mouse brain. Mol Ther Nucleic Acids. 2015;4:e266. 10.1038/mtna.2015.38 [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Lacroix A, Edwardson TGW, Hancock MA et al. Development of DNA nanostructures for high-affinity binding to human serum albumin. J Am Chem Soc. 2017;139:7355–62. 10.1021/jacs.7b02917 [DOI] [PubMed] [Google Scholar]
18. Lacroix A, Fakih HH, Sleiman HF. Detailed cellular assessment of albumin-bound oligonucleotides: increased stability and lower non-specific cell uptake. J Control Release. 2020;324:34–46. 10.1016/j.jconrel.2020.04.020 [DOI] [PubMed] [Google Scholar]
19. Fakih HH, Tang Q, Summers A et al. Dendritic amphiphilic siRNA: selective albumin binding, in vivo efficacy, and low toxicity. Mol Ther Nucleic Acids. 2023;34:102080. 10.1016/j.omtn.2023.102080 [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Fakih HH, Tang Q, Summers A et al. Albumin-binding dendritic siRNA improves delivery and efficacy to solid tumors in a melanoma model. Mol Ther Nucleic Acids. 2025;36:102579. 10.1016/j.omtn.2025.102579 [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Mestre H, Hablitz LM, Xavier ALR. Aquaporin-4-dependent glymphatic solute transport in the rodent brain. eLife. 2018;7:e40070. 10.7554/eLife.40070 [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Wolf MS, Chen Y, Simon DW et al. Quantitative and qualitative assessment of glymphatic flux using Evans blue albumin. J Neurosci Methods. 2019;311:436–41. 10.1016/j.jneumeth.2018.09.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Ohara M, Hattori T. The glymphatic system in cerebrospinal fluid dynamics: clinical implications, its evaluation, and application to therapeutics. Neurodegener Dis. 2025;25:1–13. 10.1159/000546286 [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Sorets AG, Schwensen KR, Francini N et al. Elucidating brain transport pathways and cell type-dependent gene silencing of a durable lipid–siRNA conjugate administered into cerebrospinal fluid. Nucleic Acids Res. 2025;53:gkaf600. 10.1093/nar/gkaf600 [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Andreone BJ, Lin J, Tocci J. Durable suppression of seizures in a preclinical model of KCNT1 genetic epilepsy with divalent small interfering RNA. Epilepsia. 2025;66:1677–90. 10.1111/epi.18278 [DOI] [PubMed] [Google Scholar]
26. Gentile JE, Corridon TL, Serack FE et al. Divalent siRNA for prion disease. Nucleic Acids Res. 2026; 54:gkag287, 10.1093/nar/gkag287. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Tang Q, Gross KY, Fakih HH et al. Multispecies-targeting siRNAs for the modulation of JAK1 in the skin. Mol Ther Nucleic Acids. 2024;35:102117. 10.1016/j.omtn.2024.102117 [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Hariharan VN, Summers A, Clipperton-Allen AE et al. Single-dose administration of therapeutic divalent siRNA targeting MECP2 prevents lethality in an MECP2 duplication mouse model. Nat Commun. 2026;17:8877. 10.1038/s41467-026-75611-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Miller R, Paquette J, Barker A et al. Preventing acute neurotoxicity of CNS therapeutic oligonucleotides with the addition of Ca²⁺ and Mg²⁺ in the formulation. Mol Ther Nucleic Acids. 2024;35:102359. 10.1016/j.omtn.2024.102359 [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Blennow K, Fredman P, Wallin A et al. Protein analysis in cerebrospinal fluid. II. Reference values derived from healthy individuals 18-88 years of age. Eur Neurol. 1993;33:129–33. [DOI] [PubMed] [Google Scholar]
31. Belgrad J, Tang Q, Hildebrand S et al. A programmable dual-targeting siRNA scaffold supports potent two-gene modulation in the central nervous system. Nucleic Acids Res. 2024;52:6099–113. 10.1093/nar/gkae368 [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Moazami MP, Rembetsy-Brown JM, Sarli SL et al. Quantifying and mitigating motor phenotypes induced by antisense oligonucleotides in the central nervous system. Mol Ther. 2024;32:4401–17. 10.1016/j.ymthe.2024.10.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Berman-Booty LD, Klein SK, Mazur C et al. Toxicologic pathology forum: opinion on interpretive challenges for procedure-related effects associated with direct central nervous system delivery of oligonucleotides to rodents, dogs, and nonhuman primates. Toxicol Pathol. 2023;51:375–89. 10.1177/01926233231218953 [DOI] [PubMed] [Google Scholar]
34. Graber DJ, Hickey WF, Harris BT. Progressive changes in microglia and macrophages in spinal cord and peripheral nerve in the transgenic rat model of amyotrophic lateral sclerosis. J Neuroinflam. 2010;7:8. 10.1186/1742-2094-7-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Tang Q, Fakih HH, Zain Ui Abideen M et al. Rational design of a JAK1-selective siRNA inhibitor for the modulation of autoimmunity in the skin. Nat Commun. 2023;14:7099. 10.1038/s41467-023-42714-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Biscans A, Caiazzi J, McHugh N et al. Docosanoic acid conjugation to siRNA enables functional and safe delivery to skeletal and cardiac muscles. Mol Ther. 2021;29:1382–94. 10.1016/j.ymthe.2020.12.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Godinho B, Gilbert JW, Haraszti RA et al. Pharmacokinetic profiling of conjugated therapeutic oligonucleotides: a high-throughput method based upon serial blood microsampling coupled to peptide nucleic acid hybridization assay. Nucleic Acid Ther. 2017;27:323–34. 10.1089/nat.2017.0690 [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Kahriman A, Bouley J, Smith TW et al. Mouse closed head traumatic brain injury replicates the histological tau pathology pattern of human disease: characterization of a novel model and systematic review of the literature. Acta Neuropathol Commun. 2021;9:118. 10.1186/s40478-021-01220-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Gaus HJ, Gupta R, Chappell AE et al. Characterization of the interactions of chemically-modified therapeutic nucleic acids with plasma proteins using a fluorescence polarization assay. Nucleic Acids Res. 2019;47:1110–22. 10.1093/nar/gky1260 [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Goto A, Yamamoto S, Iwasaki S. Biodistribution and delivery of oligonucleotide therapeutics to the central nervous system: advances, challenges, and future perspectives. Biopharm Drug Dispos. 2023;44:26–47. 10.1002/bdd.2338 [DOI] [PubMed] [Google Scholar]
41. Regan Lindenboom C, Brodsky JF. Alnylam reports additional positive interim phase 1 results for ALN-APP. Development for Alzheimer’s Disease and Cerebral Amyloid Angiopathy. 2023. https://investors.alnylam.com/press-release?id=27761 (28 July 2026, date last accessed)
42. Frei JA, Gentile JE, Lian Y et al. Cell type distribution of intrathecal antisense oligonucleotide activity in deep brain regions of non-human primates. Nucleic Acid Ther. 2025;35:234–40. 10.1177/21593337251371594 [DOI] [PubMed] [Google Scholar]
43. Mortberg MA, Gentile JE, Nadaf NM et al. A single-cell map of antisense oligonucleotide activity in the brain. Nucleic Acids Res. 2023;51:7109–24. 10.1093/nar/gkad371 [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Kuroda T, Yoshioka K, Mon SSL et al. Unraveling and controlling late-onset neurotoxicity of antisense oligonucleotides through strategic chemical modifications. Mol Ther Nucleic Acids. 2025;36:102692. 10.1016/j.omtn.2025.102692 [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Biscans A, Coles A, Haraszti R et al. Diverse lipid conjugates for functional extra-hepatic siRNA delivery in vivo. Nucleic Acids Res. 2019;47:1082–96. 10.1093/nar/gky1239 [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Hoogenboezem EN, Duvall CL. Harnessing albumin as a carrier for cancer therapies. Adv Drug Deliv Rev. 2018;130:73–89. 10.1016/j.addr.2018.07.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Hammond SM, Abendroth F, Goli L et al. Antibody-oligonucleotide conjugate achieves CNS delivery in animal models for spinal muscular atrophy. JCI Insight. 2022;7:e154142. 10.1172/jci.insight.154142 [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Barker SJ, Thayer MB, Kim C et al. Targeting the transferrin receptor to transport antisense oligonucleotides across the mammalian blood-brain barrier. Sci Transl Med. 2024;16:eadi2245. 10.1126/scitranslmed.adi2245 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gkag926_Supplemental_File

gkag926_supplemental_file.pdf^{(8.2MB, pdf)}

Data Availability Statement

All data are presented in the main text and the supplementary data, with raw data available from the corresponding authors upon request.

[B1] 1. Belgrad J, Fakih HH, Khvorova A. Nucleic acid therapeutics: successes, milestones, and upcoming innovation. Nucleic Acid Ther. 2024;34:52–72. 10.1089/nat.2023.0068 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B2] 2. Egli M, Manoharan M. Chemistry, structure and function of approved oligonucleotide therapeutics. Nucleic Acids Res. 2023;51:2529–73. 10.1093/nar/gkad067 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B3] 3. Watts JK. Nucleic Acid Therapeutics. In: Blackburn GM, Egli M, Gait MJ, Watts JK (eds.), Nucleic Acids in Chemistry and Biology. Cambridge, UK, The Royal Society of Chemistry, 2022, pp.350–402. 10.1039/9781837671328: 10.1039/9781837671328-00350 [DOI] [Google Scholar]

[B4] 4. Tang Q, Khvorova A. RNAi-based drug design: considerations and future directions. Nat Rev Drug Discov. 2024;23:341–64. 10.1038/s41573-024-00912-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B5] 5. Hu B, Zhong L, Weng Y et al. Therapeutic siRNA: state of the art. Signal Transduct Target Ther. 2020;5:101. 10.1038/s41392-020-0207-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[B6] 6. Mansour HM, El-Khatib AS. Oligonucleotide-based therapeutics for neurodegenerative disorders: focus on antisense oligonucleotides. Eur J Pharmacol. 2025;998:177529. 10.1016/j.ejphar.2025.177529 [DOI] [PubMed] [Google Scholar]

[B7] 7. Brown KM, Nair JK, Janas MM et al. Expanding RNAi therapeutics to extrahepatic tissues with lipophilic conjugates. Nat Biotechnol. 2022;40:1500–8. 10.1038/s41587-022-01334-x [DOI] [PubMed] [Google Scholar]

[B8] 8. Alterman JF, Godinho BMDC, Hassler MR. A divalent siRNA chemical scaffold for potent and sustained modulation of gene expression throughout the central nervous system. Nat Biotechnol. 2019;37:884–94. 10.1038/s41587-019-0205-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B9] 9. Boyd MM, Mazur C, Pribnow K et al. Comparison of MALAT1 antisense oligonucleotide distribution following intracerebroventricular and lumbar intrathecal routes of administration. Mol Ther Nucleic Acids. 2025;36:102742. 10.1016/j.omtn.2025.102742 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B10] 10. Cook BE, McLaren DG, Sullivan JM et al. Central nervous system biodistribution and pharmacokinetics of radiolabeled tofersen in rodents, nonhuman primates, and humans. J Nucleic Med. 2026;67:139–44. 10.2967/jnumed.125.270731 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B11] 11. Mazur C, Powers B, Zasadny K. Brain pharmacology of intrathecal antisense oligonucleotides revealed through multimodal imaging. JCI Insight. 2019;4:e129240. 10.1172/jci.insight.129240 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B12] 12. Jafar-nejad P, Powers B, Soriano A et al. The atlas of RNase H antisense oligonucleotide distribution and activity in the CNS of rodents and non-human primates following central administration. Nucleic Acids Res. 2020;49:657–73. 10.1093/nar/gkaa1235 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B13] 13. Wolfrum C, Shi S, Jayaprakash KN et al. Mechanisms and optimization of in vivo delivery of lipophilic siRNAs. Nat Biotechnol. 2007;25:1149–57. 10.1038/nbt1339 [DOI] [PubMed] [Google Scholar]

[B14] 14. Osborn MF, Coles AH, Biscans A et al. Hydrophobicity drives the systemic distribution of lipid-conjugated siRNAs via lipid transport pathways. Nucleic Acids Res. 2019;47:1070–81. 10.1093/nar/gky1232 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B15] 15. Sarli SL, Fakih HH, Kelly K et al. Quantifying the activity profile of ASO and siRNA conjugates in glioblastoma xenograft tumors in vivo. Nucleic Acids Res. 2024;52:4799–817. 10.1093/nar/gkae260 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B16] 16. Alterman JF, Hall LM, Coles AH et al. Hydrophobically modified siRNAs silence Huntingtin mRNA in primary neurons and mouse brain. Mol Ther Nucleic Acids. 2015;4:e266. 10.1038/mtna.2015.38 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B17] 17. Lacroix A, Edwardson TGW, Hancock MA et al. Development of DNA nanostructures for high-affinity binding to human serum albumin. J Am Chem Soc. 2017;139:7355–62. 10.1021/jacs.7b02917 [DOI] [PubMed] [Google Scholar]

[B18] 18. Lacroix A, Fakih HH, Sleiman HF. Detailed cellular assessment of albumin-bound oligonucleotides: increased stability and lower non-specific cell uptake. J Control Release. 2020;324:34–46. 10.1016/j.jconrel.2020.04.020 [DOI] [PubMed] [Google Scholar]

[B19] 19. Fakih HH, Tang Q, Summers A et al. Dendritic amphiphilic siRNA: selective albumin binding, in vivo efficacy, and low toxicity. Mol Ther Nucleic Acids. 2023;34:102080. 10.1016/j.omtn.2023.102080 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B20] 20. Fakih HH, Tang Q, Summers A et al. Albumin-binding dendritic siRNA improves delivery and efficacy to solid tumors in a melanoma model. Mol Ther Nucleic Acids. 2025;36:102579. 10.1016/j.omtn.2025.102579 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B21] 21. Mestre H, Hablitz LM, Xavier ALR. Aquaporin-4-dependent glymphatic solute transport in the rodent brain. eLife. 2018;7:e40070. 10.7554/eLife.40070 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B22] 22. Wolf MS, Chen Y, Simon DW et al. Quantitative and qualitative assessment of glymphatic flux using Evans blue albumin. J Neurosci Methods. 2019;311:436–41. 10.1016/j.jneumeth.2018.09.031 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B23] 23. Ohara M, Hattori T. The glymphatic system in cerebrospinal fluid dynamics: clinical implications, its evaluation, and application to therapeutics. Neurodegener Dis. 2025;25:1–13. 10.1159/000546286 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B24] 24. Sorets AG, Schwensen KR, Francini N et al. Elucidating brain transport pathways and cell type-dependent gene silencing of a durable lipid–siRNA conjugate administered into cerebrospinal fluid. Nucleic Acids Res. 2025;53:gkaf600. 10.1093/nar/gkaf600 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B25] 25. Andreone BJ, Lin J, Tocci J. Durable suppression of seizures in a preclinical model of KCNT1 genetic epilepsy with divalent small interfering RNA. Epilepsia. 2025;66:1677–90. 10.1111/epi.18278 [DOI] [PubMed] [Google Scholar]

[B26] 26. Gentile JE, Corridon TL, Serack FE et al. Divalent siRNA for prion disease. Nucleic Acids Res. 2026; 54:gkag287, 10.1093/nar/gkag287. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B27] 27. Tang Q, Gross KY, Fakih HH et al. Multispecies-targeting siRNAs for the modulation of JAK1 in the skin. Mol Ther Nucleic Acids. 2024;35:102117. 10.1016/j.omtn.2024.102117 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B28] 28. Hariharan VN, Summers A, Clipperton-Allen AE et al. Single-dose administration of therapeutic divalent siRNA targeting MECP2 prevents lethality in an MECP2 duplication mouse model. Nat Commun. 2026;17:8877. 10.1038/s41467-026-75611-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B29] 29. Miller R, Paquette J, Barker A et al. Preventing acute neurotoxicity of CNS therapeutic oligonucleotides with the addition of Ca²⁺ and Mg²⁺ in the formulation. Mol Ther Nucleic Acids. 2024;35:102359. 10.1016/j.omtn.2024.102359 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B30] 30. Blennow K, Fredman P, Wallin A et al. Protein analysis in cerebrospinal fluid. II. Reference values derived from healthy individuals 18-88 years of age. Eur Neurol. 1993;33:129–33. [DOI] [PubMed] [Google Scholar]

[B31] 31. Belgrad J, Tang Q, Hildebrand S et al. A programmable dual-targeting siRNA scaffold supports potent two-gene modulation in the central nervous system. Nucleic Acids Res. 2024;52:6099–113. 10.1093/nar/gkae368 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B32] 32. Moazami MP, Rembetsy-Brown JM, Sarli SL et al. Quantifying and mitigating motor phenotypes induced by antisense oligonucleotides in the central nervous system. Mol Ther. 2024;32:4401–17. 10.1016/j.ymthe.2024.10.024 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B33] 33. Berman-Booty LD, Klein SK, Mazur C et al. Toxicologic pathology forum: opinion on interpretive challenges for procedure-related effects associated with direct central nervous system delivery of oligonucleotides to rodents, dogs, and nonhuman primates. Toxicol Pathol. 2023;51:375–89. 10.1177/01926233231218953 [DOI] [PubMed] [Google Scholar]

[B34] 34. Graber DJ, Hickey WF, Harris BT. Progressive changes in microglia and macrophages in spinal cord and peripheral nerve in the transgenic rat model of amyotrophic lateral sclerosis. J Neuroinflam. 2010;7:8. 10.1186/1742-2094-7-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B35] 35. Tang Q, Fakih HH, Zain Ui Abideen M et al. Rational design of a JAK1-selective siRNA inhibitor for the modulation of autoimmunity in the skin. Nat Commun. 2023;14:7099. 10.1038/s41467-023-42714-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B36] 36. Biscans A, Caiazzi J, McHugh N et al. Docosanoic acid conjugation to siRNA enables functional and safe delivery to skeletal and cardiac muscles. Mol Ther. 2021;29:1382–94. 10.1016/j.ymthe.2020.12.023 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B37] 37. Godinho B, Gilbert JW, Haraszti RA et al. Pharmacokinetic profiling of conjugated therapeutic oligonucleotides: a high-throughput method based upon serial blood microsampling coupled to peptide nucleic acid hybridization assay. Nucleic Acid Ther. 2017;27:323–34. 10.1089/nat.2017.0690 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B38] 38. Kahriman A, Bouley J, Smith TW et al. Mouse closed head traumatic brain injury replicates the histological tau pathology pattern of human disease: characterization of a novel model and systematic review of the literature. Acta Neuropathol Commun. 2021;9:118. 10.1186/s40478-021-01220-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B39] 39. Gaus HJ, Gupta R, Chappell AE et al. Characterization of the interactions of chemically-modified therapeutic nucleic acids with plasma proteins using a fluorescence polarization assay. Nucleic Acids Res. 2019;47:1110–22. 10.1093/nar/gky1260 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B40] 40. Goto A, Yamamoto S, Iwasaki S. Biodistribution and delivery of oligonucleotide therapeutics to the central nervous system: advances, challenges, and future perspectives. Biopharm Drug Dispos. 2023;44:26–47. 10.1002/bdd.2338 [DOI] [PubMed] [Google Scholar]

[B41] 41. Regan Lindenboom C, Brodsky JF. Alnylam reports additional positive interim phase 1 results for ALN-APP. Development for Alzheimer’s Disease and Cerebral Amyloid Angiopathy. 2023. https://investors.alnylam.com/press-release?id=27761 (28 July 2026, date last accessed)

[B42] 42. Frei JA, Gentile JE, Lian Y et al. Cell type distribution of intrathecal antisense oligonucleotide activity in deep brain regions of non-human primates. Nucleic Acid Ther. 2025;35:234–40. 10.1177/21593337251371594 [DOI] [PubMed] [Google Scholar]

[B43] 43. Mortberg MA, Gentile JE, Nadaf NM et al. A single-cell map of antisense oligonucleotide activity in the brain. Nucleic Acids Res. 2023;51:7109–24. 10.1093/nar/gkad371 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B44] 44. Kuroda T, Yoshioka K, Mon SSL et al. Unraveling and controlling late-onset neurotoxicity of antisense oligonucleotides through strategic chemical modifications. Mol Ther Nucleic Acids. 2025;36:102692. 10.1016/j.omtn.2025.102692 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B45] 45. Biscans A, Coles A, Haraszti R et al. Diverse lipid conjugates for functional extra-hepatic siRNA delivery in vivo. Nucleic Acids Res. 2019;47:1082–96. 10.1093/nar/gky1239 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B46] 46. Hoogenboezem EN, Duvall CL. Harnessing albumin as a carrier for cancer therapies. Adv Drug Deliv Rev. 2018;130:73–89. 10.1016/j.addr.2018.07.011 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B47] 47. Hammond SM, Abendroth F, Goli L et al. Antibody-oligonucleotide conjugate achieves CNS delivery in animal models for spinal muscular atrophy. JCI Insight. 2022;7:e154142. 10.1172/jci.insight.154142 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B48] 48. Barker SJ, Thayer MB, Kim C et al. Targeting the transferrin receptor to transport antisense oligonucleotides across the mammalian blood-brain barrier. Sci Transl Med. 2024;16:eadi2245. 10.1126/scitranslmed.adi2245 [DOI] [PubMed] [Google Scholar]

PERMALINK

Albumin-binding dendrimer-conjugated siRNA enables safe and effective gene silencing throughout the central nervous system

Hassan H Fakih

Masahiro Ohara

Ashley Summers

Samantha L Sarli

Karen Kelly

Rosemary Gagnon

Bruktawit Maru

Brianna Bramato

Anastasia Khvorova

Jonathan K Watts

Roles

Abstract

Graphical Abstract

Graphical Abstract.

Introduction

Materials and methods

Synthesis and purification of oligonucleotides

Duplex formation

Gel-based albumin-binding assay

Animal experiments

Stereotactic intracerebroventricular injections in mice

Figure 1.

Figure 3.

Figure 2.

Evaluation of acute behavioral toxicity after ICV injections

Intrathecal delivery of oligonucleotides to rats

Limb paralysis scoring after IT injection

mRNA quantification and tissue processing

Figure 5.

siRNA tissue accumulation quantification

Rat protein quantification with western blots

Histology

Spatial distribution of siRNA in the brain with line-profile analysis

Blood diagnostics

Graphs and statistical analyses

Results

D-siRNA, but not Unc- and Di-siRNA, binds albumin in CSF

D-siRNA achieves brain-wide distribution in mice following administration into the CSF

D-siRNA supports safe, potent, and durable activity in the CNS

D-siRNA enables safe and brain-wide distribution via intrathecal administration in rats

Figure 4.

Robust knockdown of App with intrathecally administered D-siRNA

Discussion

Supplementary Material

Acknowledgements

Notes

Contributor Information

Supplementary data

Conflict of interest

Funding

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases