Comparison of DLin-MC3-DMA and ALC-0315 for siRNA delivery to hepatocytes and hepatic stellate cells

Abstract

Ionizable cationic lipids are essential for efficient in vivo delivery of RNA by lipid nanoparticles (LNPs). DLin-MC3-DMA (MC3), ALC-0315, and SM-102 are the only ionizable cationic lipids currently clinically approved for RNA therapies. ALC-0315 and SM-102 are structurally similar lipids used in SARS-CoV-2 mRNA vaccines, while MC3 is used in a siRNA therapy to knock down transthyretin in hepatocytes. Hepatocytes and hepatic stellate cells (HSCs) are particularly attractive targets for RNA therapy because they synthesize many plasma proteins, including those that influence blood coagulation. While LNPs preferentially accumulate in the liver, evaluating the ability of different ionizable cationic lipids to deliver RNA cargo into distinct cell populations is important for designing RNA-LNP therapies with minimal hepatotoxicity. Here, we directly compared LNPs containing either ALC-0315 or MC3 to knock-down coagulation factor VII (FVII) in hepatocytes and ADAMTS13 in HSCs. At a dose of 1 mg/kg of siRNA in mice, LNPs with ALC-0315 achieved a two-fold and ten-fold greater knockdown of FVII and ADAMTS13, respectively, compared to LNPs with MC3. At a high dose (5 mg/kg), ALC-0315 LNPs increased markers of liver toxicity (ALT and bile acids) while the same dose of MC3 LNPs did not. These results demonstrate that ALC-0315 LNPs achieves potent siRNA-mediated knockdown of target proteins in hepatocytes and HSCs, in mice, though markers of liver toxicity can be observed after a high dose. This study provides an initial comparison that may inform the development of ionizable cationic LNP therapeutics with maximal efficacy and limited toxicity.

Graphical Abstract

Introduction

Ionizable cationic lipids have revolutionized the field of gene therapy by enabling efficient and safe delivery of RNA molecules.¹ Lipid nanoparticles (LNPs) containing ionizable cationic lipids are in three clinically approved RNA therapies, including the two mRNA-based SARS-CoV-2 vaccines. There is potential to extend RNA-LNPs to target or express numerous, if not all, other proteins. Challenges include accessing target proteins in various cell populations, and the potential for hepatotoxicity with high or repeat intravenous dosing.² Insights into the extent that LNPs with different ionizable cationic lipids infiltrate specific hepatic cell populations may aid the design of nanomedicines with decreased risk of hepatotoxicity, and increased efficacy for modulating diseases originating from specific cell types.^3–5

LNPs made with ionizable cationic lipids adopt different charges depending on the pH of the local environment. In circulation and at physiological pH, ionizable cationic lipids adopt a net-neutral surface charge, avoiding the rapid clearance and toxicity associated with permanently cationic LNPs.⁶ At low pH, ionizable lipids adopt a positive charge, enabling efficient cargo loading and endosomal escape.⁶ LNPs can readily access cells in the liver due to the fenestration of the liver vasculature, with different cell populations exhibiting variable LNP uptake.^2,7 LNPs, administered intravenously, associate with Apolipoprotein E (ApoE) in circulation and are endocytosed by hepatic cells via the low-density lipoprotein receptor (LDLR).⁷ Hepatocytes, which are readily transfected with LNPs, express abundant LDLR on the membrane since they play a major role in lipid metabolism and bile production.² Hepatic stellate cells (HSCs) also express LDLR on their cell membrane, however, high LNP transfection efficiency in vivo has not been previously reported.^8–10 Hepatocytes and HSCs are main sites of synthesis for the majority of plasma proteins, including many coagulation, anti-coagulation, fibrinolytic, and anti-fibrinolytic factors. HSCs are a driver of hepatic fibrosis and eventually cirrhosis, an irreversible scarring of the liver that occurs in response to chronic liver injury.^11–13 Liver cirrhosis compromises the ability of hepatocytes to produce proteins that modulate innate immunity in response to pathogenic and inflammatory signals.^3–5,14

To date, three ionizable cationic lipids have been approved for clinical use in RNA-based therapeutics: DLin-MC3-DMA (heptatriaconta-6,9,28,31-tetraen-19-yl-4-(dimethylamino)butanoate),³ ALC-0315 (4-hydroxybutyl) azanediyl)bis (hexane-6,1-diyl)bis(2-hexyldecanoate),¹⁵ and SM-102 (heptadecan-9-yl 8-((2-hydroxyethyl) (6-oxo-6-(undecyloxy) hexyl) amino) octanoate)¹⁶. MC3 was designed for delivery of siRNA to hepatocytes to treat hereditary transthyretin amyloidosis (hATTR).³ ALC-0315 and SM-102 were designed for mRNA delivery, as they are the ionizable cationic lipids in the mRNA-based SARS-CoV-2 vaccines developed by Pfizer/BioNTech/Acuitas and Moderna, respectively.^17,18 All three of these drug formulations use similar “helper” lipids and lipid molar ratios, approximately 50% ionizable cationic lipid, 10% DSPC (1,2-Distearoyl-sn-glycero-3-phosphocholine), 38.5% cholesterol, and 1.5% PEG-DMG (1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000).¹⁹ MC3 contains major structural differences compared to ALC-0315 and SM-102, therefore differences in potency and toxicity may exist. Here, we report a head-to-head comparison of LNPs containing MC3 and ALC-0315, comparing their hepatotoxicity, and ability to deliver siRNA cargo in vivo to hepatocytes and HSCs. Due to the structural similarity to ALC-0315, SM-102 was not investigated in this study; both ionizable lipids exhibit similar branching, and the same functional groups (one hydroxy, one tertiary amine, two esters and only saturated hydrocarbons).²⁰

Experimental Section

siRNA-LNP formulation, and analysis

2’-O-methylated siRNA targeting murine FVII (siFVII), ADAMTS13 (siADAMTS13) and a negative control siRNA targeting Luciferase (siLuc) were obtained commercially (Integrated DNA Technologies (IDT), Coralville, USA). siRNAs were encapsulated in LNPs as previously described.²¹ Briefly, siRNAs were dissolved in sodium acetate (pH 4) and combined with a lipid solution at an amine-to-phosphate (N/P) ratio of 3. The lipid formulations consisted of DSPC, cholesterol and PEG-DMG at a 10:38.5:1.5 % molar ratio, with either 50% DLin-MC3-DMA (synthesized by the lab of Marco Ciufolini, UBC, Vancouver, Canada) or ALC-0315 (DC Chemicals, Shanghai, China). The LNPs were dialysed against phosphate buffered saline (PBS) at pH 7.4 in 500-fold excess. For quality control, cholesterol content was measured using the Cholesterol E Assay Kit (Wako Chemicals, Mountain View, CA). To determine siRNA concentration and entrapment a RiboGreen assay (Quant-IT Ribogreen RNA Assay Kit, ThermoFisher) was used. Encapsulation efficiency was measured by comparing Ribogreen signal in LNP samples with or without Triton X-100 detergent. Particle size was measured using a Malvern-Zetasizer Nano). The resulting LNPs were diluted to a final concentration of 0.1 mg siRNA per mL in PBS prior to intravenous (IV) injection.

Mice

Murine studies were approved by and performed in accordance with the University of British Columbia Animal Care Committee protocol # A20-0142. Isolation of hepatic stellate cells was performed at Michigan State University (MSU) in studies approved by the MSU Institutional Animal Care and Use Committee. C57BL/6J mice (Jackson Labs, Bar Harbor, ME, stock # 000664) age 8–10 weeks were used in all studies.

LNP-siRNA injections

siFVII, siADAMTS13, and siLuc were encapsulated in LNPs containing either ALC-0315 or MC3 as the ionizable cationic lipid. We injected mice with 1 mg siRNA per kg body weight (mg/kg) for knockdown studies, and 5 mg/kg dose for toxicity studies. A dose of 1 mg/kg siRNA in mice is standard for inducing knockdown of mRNA for proteins made in hepatocytes using siRNA-LNPs, whereas 5 mg/kg is a higher dose than the one that would normally be used in mice.³ The recommended dose of ONPATTRO (the clinically approved siRNA for hATTR) is 0.3 mg/kg, which corresponds to a human equivalent dose (HED) of 3.69 mg/kg in mice when using body surface area conversion.^3,22 One week after administration, liver tissue and blood were collected to measure target mRNA and protein levels, respectively, and compared to siLuc-treated mice; half-lives of plasma FVII and ADAMTS13 are 3–6 hours, and 2–3 days, respectively.^23,24 mRNA and protein quantification, and toxicity studies are described further below.

Liver, serum, and plasma extraction from mice

All blood samples were collected via cardiac puncture and liver tissues were surgically excised under isoflurane anesthesia. Plasma samples were separated from whole blood by spinning at 2000 × g for 10 minutes, after blood collection into a syringe containing sodium citrate (0.32 % final). Serum for toxicology analysis was isolated the same way as the plasma, however the blood was collected without sodium citrate and was allowed to clot before centrifuging.

Liver mRNA extraction and quantification

Livers were homogenized in Trizol (ThermoFisher, Waltham, USA) and DNA and RNA were isolated by precipitating them in phenol-chloroform. DNA was digested by incubating samples with TURBO DNase and 10-times TURBO DNase Buffer (ThermoFisher). DNAse was then removed by repeating the precipitation of RNA in Trizol-phenol-chloroform. Reverse transcription was performed using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, USA). Quantitative PCR, was performed using the SYBR Green Master Mix (ThermoFisher), and DNA primers against ADAMTS13 (F:5’-GTGCTCACTAATCTCAATATC-3’, R:5’-AAGGATGAGGTGATGTTG-3’), FVII (F:5’-ACACGCTACAGCCAGATGAG-3’, R:5-’TTTTCAGCACAGCCTGCCAT-3’). PCR was run for 55 cycles, ADAMTS13 and FVII expression was quantified using the ΔΔCt method, relative to the expression of the housekeeping gene Ppia (F:5’-GCGTCTCCTTCGAGCTGTT-3’, R:5’-TGTAAAGTCACCACCCTGGC-3’). All primers were synthesized by IDT. Data was collected and analyzed using the 7500 Software v 2.3.

Analysis of ADAMTS13 activity levels in plasma

Mouse plasma samples (60 μL) were diluted to 100 μL with a buffer (5 mmol/L Bis-Tris, 25 mmol/L CaCl₂, 0.005% Tween-20 at pH 6) on a 96 well-plate. Next, 100 μL of 4 μmol/L of FRETS-VWF73 (Peptide Institute Inc.) was added to the diluted plasma and incubated for 20 min at 37 °C before measuring fluorescence (emission wavelength of 460 nm and excitation of 330 nm) at 30 °C (Tecan microplate reader). Readings were recorded every 20 sec for 40 min. As a negative control, the activity of ADAMTS13 was inhibited by chelating the Ca²⁺ and Zn²⁺ cofactors using EDTA. To account for differences in hemolysis, the baseline value was subtracted from each sample curve, then the data from each plasma sample was normalized by subtracting the EDTA-treated plasma (negative control) curve.

Western blotting

Plasma samples for the blots visualizing FVII were reduced, while ADAMTS13 samples were not reduced. Samples were heated and separated on 4–15 % acrylamide gradient gels (Bio-Rad). Following electrophoresis, the samples were transferred to a PVDF membrane and blocked with protein-free Blot Blocking Buffer (Azure, VWR, Radnor, USA). The membranes were treated with a primary antibody against ADAMTS13 (1:1000; NB110-82382, Novus Biologicals, Littleton, USA), or FVII (1:1000; AF3305, R&D Systems), washed, and treated with HRP-conjugated secondary antibody (Goat Anti-Rabbit: 1:2000, Cell Signaling Inc.; Donkey Anti-Goat: 1/15,000, ab7125, Abcam; or Goat Anti-Mouse IgG: 1:15,000, ab97040, Abcam). Bands were detected using ECL substrates (Bio-Rad) and imaged on a Sapphire Biomolecular Imager (Azure Biosystem). Protein quantification of the bands was performed using ImageJ software. Relative intensity was determined using IgG as the control.

Isolation of HSC from Mouse Liver

Hepatic stellate cells were isolated from mice as described previously with minor modifications.²⁵ The caudal vena cava was cannulated and the livers were perfused with perfusion buffer (50 mM EGTA, 1 M Glucose, and 1% Penicillin-Streptomycin in Hank’s Balanced Salt Solution (HBSS, without calcium and magnesium). Livers were then perfused with HBSS containing 1.0 M calcium chloride, 1.0 M glucose, 1% Penicillin-Streptomycin, and 5.0 mg Pronase (Roche, Indianapolis, IN). Finally, livers were perfused with HBSS containing 1.0 M calcium chloride, 1.0 M glucose, 1% Penicillin-Streptomycin, and 5.0 mg Collagenase (Sigma-Aldrich). The digested livers were transferred to cold media (DMEM, 1% Pen/Strep, 10% heat-inactivated FBS) and mechanically dissociated, and DNAse solution (16 μg/mL DNAse final concentration in sterile saline) was added. The solutions were centrifuged 3 times at 50 × g for 2 min, and the hepatocyte pellet was discarded. The collected supernatants were centrifuged at 1000 × g for 6 min to collect non-parenchymal cells. Hepatic stellate cells were separated from the non-parenchymal cells using an 8% Histodenz (Sigma-Aldrich) gradient, as described previously.²⁵ The hepatic stellate cells were collected from the interface and diluted to 50 mL in HBSS, then centrifuged at 1000 × g for 8 minutes. The resulting hepatic stellate cells were plated in 12-well tissue culture plates (Corning) in DMEM containing 10% FBS and 1% Penicillin-Streptomycin for 2 hours at 37 C. Following incubation, the adherent hepatic stellate cells were gently rinsed in HBSS, then lysed in TRI Reagent and mRNA was isolated using DirectZol mRNA MiniPrep spin columns (Zymo Research, Irvine CA).

Toxicological analysis

Mice were injected IV with either PBS, or with siLuc encapsulated in LNPs with ALC-0315 (siLuc-ALC-0315) or MC3 (siLuc-MC3) at 5 mg/kg (N = 4). While a dose of any LNP at 10 mg/kg usually causes severe toxicity, such as inflammation and liver necrosis, the toxicity after a 5 mg/kg dose depends on the lipid formulation.^26,27 Five hours after the injection, mice were sacrificed, and serum samples were collected as described above. Serum samples were submitted to Idexx BioAnalytics (West Sacramento, USA) for a toxicology panel. Aspartate aminotransferase (AST), alkaline phosphatase (ALP), alanine aminotransferase (ALT), bile acids, total bilirubin (TBIL), blood urea nitrogen (BUN), creatine (CREA), gamma-glutamyl transferase (GGT) levels were analyzed. To note, data regarding bile acid levels in mice treated with PBS and with siLuc-ALC-0315 had N = 3 due to the presence of an outlier in each group (data not shown). The presence of the outliers would have not altered the conclusion, siLuc-ALC-0315 treated mice would have had an even higher bile acid mean and would have been more statistically significant from the PBS-treated mice. Outliers were determined via the ROUT method using GraphPad Prism although limitations such as our small sample size were considered. Bile acid levels commonly range from 0 to 6 μmol/L; however, our results were likely not biologically possible (>130 μmol/L).²⁸

Statistical Analysis

To ensure a t-test could be used between two groups, the F-test was performed to confirm the standard deviation (SD) between groups was not statistically different. Comparisons between the mean of two groups were performed with a one-tailed unpaired parametric t-test or with Welch’s t test if the SD between groups was significantly different. For Figure 3C, simple linear regression was used to determine statistical difference in the slopes. All analysis was done using GraphPad Prism (Version 9.2.0). Significance was designated at P values < 0.05.

Figure 3. ALC-0315 causes increased markers of liver toxicity after a 5 mg/kg dose.

Serum was collected five hours after administration of LNPs with MC3 or ALC-0315 at a dose of 5 mg siRNA/kg body weight, or control PBS. Markers of liver toxicity were measured, including ALT (A), AST (B), and bile acids (C). N = 4. *P < 0.05, ***P < 0.001, ns indicates no significant difference. Error bars represent mean±SEM.

Results

To examine the ability of ALC-0315 in siRNA delivery to the liver, by measuring knockdown of two liver derived proteins, by comparing to MC3 LNPs, which were developed specifically for siRNA delivery to the liver. We chose to test siRNA targeting FVII and ADAMTS13 because they are well-characterized in the literature to be synthesized in hepatocytes and hepatic stellate cells, respectively.^29,30,4,5

ALC-0315 achieves more potent siRNA-mediated knockdown in hepatocytes compared to MC3.

Mice treated with 1 mg/kg siFVII encapsulated in LNPs containing either ALC-0315 or MC3 (siFVII-ALC-0315 or siFVII-MC3, respectively) exhibited significant knockdown of FVII mRNA, compared to control mice treated with siLuc-ALC-0315 or siLuc-MC3 (Figure 1A). Mice treated with siFVII-ALC-0315 at the same dose had greater FVII mRNA knockdown (1.6 ± 0.3 % residual mRNA, P = 0.0004), compared to mice treated with siFVII-MC3 (15.3 ± 3% residual mRNA, P = 0.002) (Fig. 1). Plasma proteins levels between siFVII-ALC-0315 (18 ± 8 %, P = 0.003) and siFVII-MC3 (6 ± 2% plasma protein, P = 0.02) treated mice did not significantly differ (Fig. 1B).

Figure 1. ALC-0315 induces greater knockdown of FVII in hepatocytes than MC3.

Mice were injected with a single dose of siRNA FVII (siFVII) or control siLuc, encapsulated in ALC-0315 or MC3 LNPs. Livers and blood were collected one-week post-injection to quantify FVII mRNA and plasma protein levels. A) Hepatic FVII mRNA levels, relative to siLuc-treated. B) Representative western blot of FVII in blood plasma from mice. N = 3. **P < 0.01, ***P < 0.001. Error bars represent mean±SEM.

There were no differences detected between male and female mice. Encapsulation of siRNA within LNPs was quantified, and there was no substantial difference in RNA loading between siFVII-MC3 (88%), siFVII-ALC-0315 (78%), siLuc-MC3 (90%) and siLuc-ALC-0315 (66%).

ALC-0315 achieves siRNA-mediated knockdown in HSCs, while knockdown by MC3 is minimal.

Mice treated with siADAMTS13-ALC-0315 at the same dose had greater FVII mRNA and protein knockdown (31 ± 13% residual mRNA, P = 0.038, and 40 ± 20% plasma protein, P = 0.060), compared to mice treated with siADAMTS13-MC3 (86 ± 18% residual mRNA, P = 0.221, and 75 ± 9.5% plasma protein, P = 0.274) (Fig. 2A–B). Thus, mice treated with siADAMTS13-ALC-0315 resulted in a 69% knock down in ADAMTS13 mRNA expression while mice treated with siADAMTS13-MC3 did not have a statistically significant decrease.³¹ Between siADAMTS13-ALC-0315 and siADAMTS13-MC3, the difference in Adamts13 mRNA knock down was statistically significant (P = 0.0243).

Figure 2. ALC-0315 enables siRNA-mediated knockdown of ADAMTS13 in HSCs.

Mice were injected with a single dose of siADAMTS13 (siA13) or control siLuc, encapsulated in ALC-0315 or MC3 LNPs. Livers and blood were collected one-week post-injection to quantify ADAMTS13 mRNA, protein, and activity levels. A) Hepatic Adamts13 mRNA levels, relative to siLuc-treated. B) Representative western blot of ADAMTS13 in blood plasma from mice. C) Plasma ADAMTS13 activity measured by cleavage of a fluorescent substrate. Representative ADAMTS13 activity in blood plasma from the same mice used in the western blot. D) Mice were injected with a single dose of siADAMTS13 or siLuc encapsulated in ALC-0315 LNPs. One-week post-injection, HSCs were isolated from the liver, and qPCR was performed to quantify ADAMTS13 mRNA. PPIA was used as an internal control. Grey dashed line represents a maximum amplification cycle of 55. N = 4. *P < 0.05; ns indicates no significant difference. Error bars represent mean±SEM.

The enzymatic activity of ADAMTS13 in plasma was determined by measuring the rate of cleavage of a fluorogenic substrate (Figure 2C).³² Samples from mice treated with siLuc-MC3 and siLuc-ALC-0315 exhibited high ADAMTS13 activity (5.2 ± 0.04 and 3.7 ± 0.04 RFU/sec, respectively) that were quenched in the presence of EDTA, an inhibitor of ADAMTS13 activity. Plasma from siADAMTS13-ALC-0315 and siADAMTS13-MC3-treated mice both showed a diminished ADAMTS13 activity (0.42 ± 0.02 RFU/sec and 2.4 ± 0.05 RFU/sec, respectively, both P < 0.05), indicating a significant decrease in activity compared to their respective siLuc-treated groups. Groups were not powered to detect statistical significance of sex-differences, however, the knockdown between males and females appeared to be the same. There was also no substantial difference in RNA loading between siADAMTS13-MC3 (94%), siADAMTS13-ALC-0315 (82%), siLuc-MC3 (94%) and siLuc-ALC-0315 (80%).

To validate that ADAMTS13 was knocked down in HSCs, ADAMTS13 mRNA was measured in HSCs isolated from livers of mice treated with siLuc or siADAMTS13 encapsulated in ALC-0315 LNPs. mRNA encoding ADAMTS13, and housekeeping gene Ppia, were measured via qPCR. Extracted RNA yields were low, corresponding to the small population of cells isolated, but detection (cycle threshold) of Ppia was similar in samples from mice treated with siLuc and siADAMTS13 (32.5 ± 1.19 and 32.1 ± 0.76, respectively); ADAMTS13 mRNA was detected in samples from mice treated with siLuc (44.4 ± 4.6) but was not detected in RNA extracted from HSCs of mice treated with siADAMTS13, up to a maximum of 55 amplification cycles.

A high intravenous dose of ALC-0315 LNPs increases some markers of liver toxicity.

To do an initial assessment of the safety of ALC-0315 and MC3 LNPs, serum from mice treated with a high dose (5 mg/kg) of siLuc-ALC-0315, siLuc-MC3, or PBS as vehicle control were analyzed. Mice treated with siLuc-ALC-0315 had significantly higher serum levels of ALT and bile acids (105.5 ± 11 U/L, P = 0.0003, and 7.6 ± 2 μmol/L, P = 0.0253, respectively) compared to PBS-treated mice (29 ± 3 U/L and 2.3 ± 1 μmol/L, respectively) (Figure 3). Serum AST may have been higher, but not statistically significant. All other markers measured (ALP, TBIL, BUN, CREA, GGT) were not statistically different from control. Mice treated with siLuc-MC3 did not have increased levels of any liver toxicity markers that we measured compared to mice treated with PBS. While the size of LNPs can contribute to the LNP immunogenicity,³³ MC3 and ALC-0315 LNPs were similar in size with a diameter of 45nm.

Discussion

We found differences in gene silencing in two distinct liver cell populations when siRNA was delivered with LNPs containing the ionizable cationic lipids, MC3 or ALC-0315. There was increased silencing of hepatocyte- and HSC-specific mRNA targets when the corresponding siRNA was delivered using ALC-0315 LNPs compared to MC3 LNPs. We speculate that ALC-0315 achieves a higher transfection potency in liver cells broadly, rather than knockdown in stellate cells being enabled through cell selectivity. Increased siRNA delivery may arise from improved endosomal escape by ALC-0315 LNPs. An important characteristic for endosomal escape is for the ionizable cationic lipid to exhibit a cone-shape structure where the area subtended by the headgroup is less than the area subtended by the alkyl chains. This cone shape promotes the formation of structures that can disrupt the endosomal membrane, such as inverted hexagonal phases.⁶ MC3 contains two C18 linoleic acid tails, each of which includes two double bonds that create a subtle cone-shape.³⁴ In contrast, ALC-0315 has four alkyl tails, resulting in a more pronounced cone-shape that could engender more effective endosomal escape.¹⁵ This is consistent with previous studies showing that only 2–5% of siRNA delivered by MC3 LNPs escapes into the cytosol, while the rest were degraded in lysosomes or released in exosomes.^35–37 The quantification of ALC-0315 endosomal release has not been reported, despite an estimated 10-fold higher mRNA silencing efficacy than MC3,^15,19 though other structurally similar ionizable lipids of this new generation exhibit endosomal release of 15% or higher.³⁸ Beyond endosomal escape, lipid chemistry can influence the composition of the bound protein corona, which in turn impacts specific cellular uptake.^39–41 Other factors that affect LNP performance, such as LNP ultrastructure, size and pKa are similar between LNPs containing MC3 or ALC-0315.¹⁵ Though neither FVII nor ADAMTS13 are synthesized in Kupffer cells, Kupffer cells internalize nanoparticles, alter the amount of nanoparticles delivered to other cells and thus may have influenced the extent of knockdown in hepatocytes and HSCs.^42,43

Despite its role in disease processes, there is a lack of specific inhibitors of ADAMTS13 for in vivo research. In this study, we developed the first-reported approach that can achieve potent in vivo knockdown of Adamts13 mRNA in HSCs, resulting in depletion of the protein from circulation. ADAMTS13 is a metalloprotease that cleaves von Willebrand factor (VWF).⁴⁴ VWF is a plasma glycoprotein that plays a key role in hemostasis by adhering and aggregating platelets.⁴⁵ ADAMTS13 has previously been suggested as a therapeutic target to prevent nonsurgical bleeding during left ventricular assist device (LVAD) support,⁴⁴ and increased levels of ADAMTS13 has been linked to diverse pathologies, including acute ischemic brain injury, hepatocellular carcinoma, and severe sepsis.^46–48 Extreme caution should however be applied in the development of ADAMTS13 inhibitors, as patients with a severe ADAMTS13 deficiency (<10% normal ADAMTS13 activity) exhibit a serious condition known as thrombotic thrombocytopenic purpura (TTP).⁴⁹ The dosing of siADAMTS13 can easily be tailored to ensure knockdown of circulating ADAMTS13 remains above this safe target threshold for therapeutic applications. In contrast, siADAMTS13 can also be leveraged to easily create animal models of TTP by ensuring dosing regimen achieves target protein knockdown of <10% of normal. Mice with ADAMTS13 deficiency alone do not exhibit TTP, except when they have a CASA/Rk genetic background.⁵⁰ Other approaches to induce TTP in mice have been developed including high IV injection doses of recombinant VWF in Adamts13^−/− mice.⁵⁰

In this study, the LNP hepatotoxicity was tested using siLuc, the most commonly used siRNA control. In order to minimize any confounding variables that may arise as a result of knocking down hepatic and coagulation proteins, the two inhibitory siRNAs (siFVII or siADAMTS13) were not tested. In this study, we detected a mild hepatotoxic response in mice after administering a high dose of siLuc-ALC-0315, compared to PBS-treated mice. This dose was at least five-times higher than what was required to achieve 98% knockdown of FVII, a hepatocyte-derived protein, and 69% knockdown of ADAMTS13, an HSC-derived protein. Intramuscular administration, as is done for the SARS-CoV-2 mRNA vaccines, limits the amount of lipid content reaching the liver, and would not be expected to show the same elevation of hepatotoxicity markers.⁵¹ We speculate that the increased levels of ALT and bile acids in mice treated intravenously with siLuc-ALC-0315 is due to the increased ability of ALC-0315 LNPs to infiltrate HSCs and hepatocytes compared to MC3. Bile acids are produced in hepatocytes, and are used as indicators of normal liver function, which may be disrupted with excess amounts of LNP uptake.⁵² ALC-0315 uptake may also trigger HSC activation, which is a central driver of liver fibrosis and cirrhosis.¹¹ HSC activation causes release of cytokines and inflammatory responses commonly observed in liver diseases and following liver injury. Mice treated with siLuc-MC3 at the same dose showed no elevation of hepatotoxicity markers compared to PBS-treated mice. A decreased ability of MC3 LNPs to infiltrate hepatocytes and HSCs could explain both the lack of hepatotoxicity, and weaker mRNA silencing efficacies compared to ALC-0315 LNPs. It should also be noted that since ALC-0315 was designed for enhanced degradability, it may cause less toxicity in repeated dosing scenarios, whereas MC3 LNPs with decreased degradability may accumulate in hepatic cells.⁵³ While the results here provide a starting point for understanding the safety of ALC-0315, in order to directly compare the hepatoxicity of ALC-0315 to MC3 more extensive characterization of their therapeutic indexes would be required.

Conclusions

This work represents proof-of-concept that ionizable cationic LNPs can be used to access and knockdown HSCs-specific targets like ADAMTS13. It shows there are differences in efficacy of MC3 and ALC-0315, which are two of the three clinically approved ionizable cationic lipids used in the LNP delivery platform. Insights from this head-to-head comparison may enable the optimization of RNA-based agents to modulate expression of proteins that were previously inaccessible in vivo for research and therapeutic purposes.

Abbreviations

ADAMTS13

a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13

ALC-0315

ionizable cationic lipid (4-hydroxybutyl) azanediyl)bis (hexane-6,1-diyl)bis(2-hexyldecanoate)

ALT

alanine transaminase

APOE

apolipoprotein E

FVII

coagulation factor VII

HSC

hepatic stellate cell

IV

intravenous

LDLR

low-density lipoprotein receptor

LNP

lipid nanoparticle

MC3

ionizable cationic lipid DLin-MC3-DMA (heptatriaconta-6,9,28,31-tetraen-19-yl-4-(dimethylamino)butanoate)

mg/kg

mg siRNA per kg body weight

PBS

phosphate buffered saline

PPP

platelet poor plasma

PRP

platelet rich plasma

qPCR

quantitative polymerase chain reaction

SEM

standard error of the mean

siADAMTS13

siRNA targeting ADAMTS13

siFVII

siRNA targeting coagulation factor VII

siLuc

siRNA targeting luciferase

siRNA

small interfering RNA

siRNA-LNP

lipid nanoparticle containing small interfering RNA

SM-102

ionizable cationic lipid (heptadecan-9-yl 8-((2-hydroxyethyl) (6-oxo-6-(undecyloxy) hexyl) amino) octanoate)