Introduction
Argininosuccinic aciduria (ASA) is an inherited rare metabolic disorder affecting 1 in 70,000 - 220,000 births worldwide, and is caused by argininosuccinate lyase (ASL) deficiency. The enzyme ASL catalyzes the hydrolytic cleavage of argininosuccinate into arginine and fumarate, an essential step in the liver for detoxifying ammonia via the urea cycle (Figure 1). Mutated ASL leads to hyperammonemia, encephalopathy, and respiratory distress, affecting primarily the liver but also the kidneys and brain. The symptoms of ASA typically occur within the first 2-3 days of life.
As depicted in Figure 1, Daly et al. (2023) have investigated the use of lipid nanoparticle (LNP)-formulated mRNA encoding normal, unmutated ASL as an intravenous injection for the tandf ASA in a mouse model, using improved survival as the principal measure of efficacy. This blog summarizes key aspects of this collaborative research by BioNTech and Genevant, which utilized TriLink mRNA reagents and TriLink mRNA custom synthesis, and included a 2023 Nobel laureate, Katalin Karikó.
FIGURE 1. Graphical abstract from Daly et al. (2023) reproduced under a CC BY license.
Design of ASL mRNA Therapeutic Candidates
To design an optimal in vitro transcribed (IVT) mRNA for the treatment of ASA, Daly et al. engineered four ASL mRNA coding sequences (CDS1-CDS4) wherein diverse parts of the mRNA were varied to increase its efficiency and safety (Figure 2). For efficient translation, each of the four candidate mRNAs was 5’ capped with the frequently used Cap 1 structure. The 5’ and 3’ untranslated regions (UTRs) were comprised of different motifs, sequences, and structural elements that regulate the translational efficiency of mRNA.
FIGURE 2. Schematic of four optimized ASL mRNA constructs CDS1-CDS4. Taken from Daly et al. (2023) and free to use under a CC BY license.
Two different pairs of 5’ UTR and 3’ UTR were studied. The first pair, comprised of the α-globin 5’ UTR and a 3’ UTR referred to as AES and mtRNR1, was previously used in the Pfizer-BioNTech COVID-19 vaccine and are known to contain motifs that increase the translation and stability of mRNA (Sahin et al. 2020). The second pair was a minimally unstructured 5’ UTR and the α-globin 3’ UTR.
The mRNA coding sequence for the wild-type ASL enzyme (CDS1) was compared with three ASL sequences that were codon-optimized. Increasing GC content and replacing rare codons with more frequently occurring codons has been shown to enhance the translational efficiency of the mRNA. Therefore, two of the coding sequences were GC-rich (CDS2 and CDS4) and the third was non-GC rich (CDS3).
To achieve improved safety of therapeutic mRNA, Daly et al. used N1-methylpseudouridine (m1Ψ)- and pseudouridine (Ψ)-modified mRNAs that are known to lower immunogenicity and minimize potential side effects caused by the release of proinflammatory cytokines and chemokines (Karikó 2021). To ensure robust enhancement of translation, polyadenylation of mRNA at its 3’ end is necessary and was therefore incorporated in each design.
Manufacturing of ASL mRNA-LNP Candidates
BioNTech manufacturing of ASL mRNAs CDS1, CDS2, and CDS3 was performed by IVT of linearized plasmid DNA templates wherein uridine 5’-triphosphate (UTP) was replaced with TriLink m1Ψ-5’-triphosphate. Each of these three mRNAs was co-transcriptionally 5’ capped with the Cap1 structure using TriLink CleanCap® Reagent AG, while the poly(A) tail was encoded by the template. ASL mRNAs CDS1-CDS3 were each purified using cellulose chromatography (Baiersdörfer et al. 2019).
Manufacturing of the ASL mRNA CDS4 was performed by TriLink mRNA Synthesis Services with replacement of UTP by TriLink Ψ-5’-triphosphate. ASL mRNA CDS4 was enzymatically capped and polyadenylated post-transcriptionally, then purified by TriLink using HPLC.
Purified ASL mRNA CDS1-CDS4 final products were each QC-checked for length, RNA integrity, capping efficiency, and negligible levels of double-stranded RNA contaminants. Each final product was encapsulated in LNPs by Genevant using a controlled mixing process in which an aqueous solution of mRNA was combined with an ethanolic solution of four lipids in precisely measured ratios, as detailed elsewhere (Lam et al. 2023).
In Vitro Expression
To examine protein expression of the four LNP-formulated ASL mRNAs CDS1-CDS4 in vitro, Daly et al. first studied transfection of Chinese hamster ovary (CHO) and human embryonic kidney (HEK293) cell lines. Cell lysates were harvested on Day 1 and Day 3 after the transfection and Western blot analysis using an ASL antibody showed protein translation from all four candidates, with relatively more expression from CDS1 and CDS2.
This preliminary evaluation was followed by comparing CDS1-CDS4 in primary hepatocytes, which are more relevant to the principal tissue effected by ASA. Transfection at 3 different concentrations and evaluation of relative expression levels at Day 6 showed a dose response for all four candidates, with CDS1 and CDS2 again providing higher levels of expression.
In Vitro Safety Assessments
It is known that mRNA is sensed by pattern-recognition receptors in the cell, and that LNP components are recognized by Toll-Like receptors located on the cell surface of macrophages and other cell types. Resultant induction of cytokine and chemokine secretion, which can lead to mRNA-LNP infusion side-reactions, including fever and chills, was expected to be blunted by using nucleoside-modified mRNA and Genevant’s biodegradable LNP.
To measure levels of immune activation after the application of the designed ASL mRNA-LNPs, Daly et al. transfected human peripheral blood mononuclear cells (PBMCs) in vitro. Cytokines and chemokines were quantified as biomarkers of immunotoxicity 24 h after the application of each of the four candidate therapeutics over a wide range of concentrations.
Insignificant amounts of interferon-ɣ-induced proteins were secreted, independent of the applied dose. The lack of a dose response and nearly background levels of these proteins indicated minimal reactogenicity of human PBMCs to the ASL mRNA-LNPs, thus predicting a high margin of safety.
Dose-Response in Wild-Type Mice
To determine the lead ASL mRNA-LNP therapeutic in vivo, a dose-response study (0.3, 1, and 3 mg/kg) was conducted in wild-type mice. Protein expression was evaluated in the liver of the mice 24 h after dosing using Western blots with an ASL antibody. A dose-dependent increase in ASL expression was observed and, in agreement with the above-mentioned in vitro data, the highest expressing mRNAs were CDS1 and CDS2 across all three dose levels.
To investigate the half-life of the ASL protein, 6 mice were administered the CDS3-encoding mRNA-LNP at 0.5 mg/kg. The livers from 3 mice were harvested 24 h post-dosing and the remaining 3 mice were sacrificed 7 days later. The expression at these two time points was quantified by mass spectrometry, which indicated a ASL protein half-life of 5 days, which guided the selected dose frequencies that were investigated in the following ASL-deficiency (ASLD) disease model.
Repeat-Dose Efficacy in an ASLD Disease Mouse Model
The efficacy of repeat-dose administration of LNP-formulated ASL mRNA CDS2, one of the two lead-candidate therapeutics, was evaluated in an ASLD disease mouse model using ASLNeo/Neo mice. In these mice, a neomycin-resistance cassette (Neo) inserted into the ASL gene disrupts its transcription by >80% compared to wild-type ASL in normal mice.
Groups of ASLNeo/Neo mice were fed a high-protein diet and treated with CDS2 mRNA-LNP on Day 0 at 4 at different dose levels (0.25, 0.5, 1, and 3 mg/kg) in a once-a-week (OW) or twice-a-week (BIW) dosing regimen. The primary readout was daily body weight change, as impaired ureagenesis would lead to rapid body weight loss in untreated animals. Dosing continued for 5 doses and any animal with ≥20% body-weight loss was euthanized. Phosphate-buffered saline (PBS) control-treated animals reached this weight loss limit within 7 days.
An obvious dose-dependent benefit to survival was evident with either dosing regimen, with animals treated BIW showing a better survival profile (Figure 3). Importantly, the 1 and 3 mg/kg treatments showed full survival during the dosing.
FIGURE 3. LNP-ASL CDS2 mRNA treatment leads to survival benefit in the ASLD mouse model. Kaplan–Meier curves for ASLNeo/Neo mice (N = 5/group) treated bi-weekly while challenged with a high-protein diet. Data from 4 dose levels and PBS control animals are shown. Taken from Daly et al. (2023) and free to use under a CC BY license.
Changes in body weights agreed with the survival data. There was a larger increase in bodyweight change for the 3 mg/kg dose level during the BIW dosing, with some animals gaining >50% of the initial starting weight. However, the body weights for the 3 mg/kg group began to decline 7 days after the last mRNA-LNP treatment, underlining the importance of the expression kinetics of the delivered mRNA.
Conclusions
Daly et al. noted that incorporation of m1Ψ into LNP-ASL mRNA CDS1-CDS3 or Ψ into LNP-ASL mRNA CDS4 leading to increased translational capacity and decreased immunogenicity, compared to unmodified mRNA, is consistent with prior findings for m1Ψ and Ψ. They added that differences in UTRs cannot be excluded as contributing to the finding that the GC-rich m1Ψ ASL mRNA CDS2 translated better compared to GC-rich Ψ ASL mRNA CDS4.
It was also concluded that all in vitro experiments using diverse methodologies and various cell types indicated that LNP-ASL mRNAs CDS1 and CDS2 were the best two lead-optimized mRNA candidates for the treatment of ASA. Furthermore, dose-response analysis of all four mRNA optimizations in wild-type mice confirmed that the highest expressing mRNAs across the dose-range were LNP-ASL mRNA CDS1 and CDS2, providing confidence in the reproducibility of the findings.
Importantly, repeated delivery of LNP-ASL mRNA CDS2 was clearly shown to be protective in a relevant in vivo disease model of morbidity/mortality associated with ASA urea-cycle disorder. While all four mRNA-LNPs also showed good tolerability in vitro and in vivo, Daly et al. suggested that, to fully leverage the potential of this ASL mRNA-LNP platform, further optimization may be necessary to refine the therapeutic window and enable full protection of animals at lower doses.
Your comments are welcomed, as usual.
Please feel free to share this blog with your colleagues or on social media.
