mRNA therapy restores euglycemia and prevents liver tumors in murine model of glycogen storage disease

Glycogen storage disease type Ia (GSDIa) results from deficiency of the endoplasmic reticulum enzyme glucose-6-phosphatase-α (G6Pase-α; G6PC), which catalyzes the final step of glycogenolysis and gluconeogenesis. Loss of hepatic G6Pase-α impairs glucose export during fasting, causing severe hypoglycemia and accumulation of glucose-6-phosphate that drives lactic acidemia, hyperlipidemia, hyperuricemia, hypercholesterolemia, steatosis, and organomegaly. Standard care relies on strict dietary regimens (frequent cornstarch feeding and nocturnal gastric glucose infusion) that are burdensome and do not prevent long-term hepatic complications such as hepatocellular adenomas (HCAs) and progression to hepatocellular carcinoma (HCC). Alternatives under investigation include liver stem cell infusion, viral vector-based gene therapies, and enzyme replacement, but each faces obstacles including transient effects, declining transgene expression, potential genotoxicity, preexisting anti-vector immunity, and difficulties purifying and delivering the multispanning ER membrane protein G6Pase-α. Messenger RNA (mRNA) therapy delivered via lipid nanoparticles (LNPs) offers transient, non-integrating, redosable restoration of protein function and leverages cellular machinery for proper protein localization. The study tests whether engineered, codon-optimized mRNA encoding human G6Pase-α can safely restore euglycemia and mitigate hepatic tumor risk in a liver-specific G6pc knockout mouse model that recapitulates human GSDIa.

Prior approaches for GSDIa include: (1) dietary therapy with uncooked/modified cornstarch and nocturnal feeds, which maintains fasting glucose only partially and does not prevent hepatic adenomas/HCC; (2) liver stem cell infusion yielding transient metabolic correction; (3) somatic gene therapy using various viral vectors that correct hypoglycemia and reduce adenomas in animal models and are in early clinical trials, but are limited by dilution of episomal vectors during hepatocyte proliferation, low transduction efficiency, potential genotoxicity, and preexisting neutralizing antibodies that hinder redosing; (4) enzyme replacement therapy is impractical due to the hydrophobic, ER-resident nature of G6Pase-α and delivery challenges. Advances in mRNA chemistry (e.g., modified nucleotides) and LNP delivery have enabled safe, efficient hepatic delivery and repeat dosing in preclinical models of liver metabolic diseases (MMA, AIP, Fabry, others), suggesting potential to restore intracellular/transmembrane proteins considered undruggable by ERT.

Engineered mRNA and LNP formulation: Fully N1-methylpseudouridine-substituted, capped mRNAs encoding human G6Pase-α were synthesized in vitro with defined 5′/3′ UTRs and poly(A) tail, purified, diluted in citrate buffer, and stored frozen. Protein engineering used consensus residue analysis across >100 mammalian orthologs (BLASTP hits realigned with MAFFT; conservation visualized by WebLogo) to identify beneficial substitutions; the S298C variant was selected for improved expression/activity. Codon optimization further enhanced expression. mRNAs were encapsulated in LNPs via ethanol-drop nanoprecipitation using an ionizable lipid (heptadecan-9-yl 8-((2-hydroxyethyl)(8-(nonyloxy)-8-oxooctyl)amino)octanoate), DSPC, cholesterol, and DMG-PEG2000. Formulations were dialyzed into PBS, concentrated, 0.22 µm filtered, QC’d (size, encapsulation, endotoxin), and stored at 4 °C. In vitro studies: HeLa and Hep3B cells were transfected with mRNA (Lipofectamine 2000 or Messenger MAX). Outcomes included G6Pase-α protein (immunoblot normalized to ERP72), enzymatic activity (Taussky–Shorr assay measuring Pi release from G6P), and subcellular localization (confocal immunocytochemistry co-staining with ER marker calnexin and mitochondrial marker TOM20; Mander’s colocalization coefficient). Wild-type mouse pharmacology: Male CD-1 mice received single i.v. doses (1.0 mg/kg) of eGFP, hG6PC-WT, or codon-optimized hG6PC-S298C mRNA-LNP. Livers were harvested at 6, 24, 72, 168, and 336 h for hepatic mRNA (RT-qPCR), protein (immunoblot), and activity (microsomal assay). Half-lives were estimated by non-compartmental analysis. Disease model: Liver-specific G6Pase-α knockout mice (L.G6pc-/-) were generated by albumin-CreERT2–mediated excision of G6pc exon 3 in C57BL/6J background following tamoxifen induction. Male mice were used after at least 4 weeks post-induction. In vivo efficacy in L.G6pc-/- mice:

• Dose-ranging, single dose: L.G6pc-/- mice received i.v. hG6PC-S298C mRNA-LNP at 0.2, 0.5, or 1.0 mg/kg; controls received eGFP mRNA (1.0 mg/kg) or PBS (WT). Fasting began immediately post-dose. Blood glucose was measured at 0, 2.5, 6, and 24 h. At 24 h, livers were collected for morphology, liver weight, hepatic G6P, glycogen, triglycerides, G6Pase-α protein and activity, and serum triglycerides.
• Duration-of-action, single dose: L.G6pc-/- mice received 0.5 or 1.0 mg/kg hG6PC-S298C mRNA (or eGFP). Blood glucose was measured at 2.5 and 6 h fasts on days 0, 2, 4, 7, 10, and 14.
• Repeat-dose efficacy: L.G6pc-/- mice received five i.v. doses of hG6PC-S298C mRNA-LNP at 0.25 mg/kg every 10–14 days over ~8 weeks; controls received eGFP. Fasting (2.5 h) blood glucose was monitored on days 1, 4, 7, and 10 after each dose. Safety assessments: Serum proinflammatory cytokines (IFNγ, IL-1β, TNFα, IL-6) were measured 24 h after single doses (0.2–1.0 mg/kg) and after five repeat doses (0.5 mg/kg). Plasma ALT was assessed. Anti-drug antibodies against G6Pase-α were quantified by ELISA after repeat dosing. Clinical observations included body weight and signs of hypersensitivity. Hepatic tumor prevention study: To accelerate tumorigenesis, WT and L.G6pc-/- male mice were fed a high fat/high sucrose diet starting 3 months before treatment initiation and throughout the study. Mice then received 8–10 i.v. doses of hG6PC-S298C mRNA-LNP (0.25–0.5 mg/kg) or controls (PBS or eGFP mRNA) every 7–14 days. Eight days after the last dose, livers were evaluated for macroscopic lesions (incidence, number per mouse, summed tumor area), histology, and expression of HCA/HCC-associated protein biomarkers (PKM2, β-catenin, p62) and gene markers (Tgfb1, Glul, Ctnnb1); serum AFP was also assessed. Statistics: Data were generally log2-transformed and analyzed by one-way ANOVA with Dunnett’s post hoc test versus eGFP. Duration-of-action and repeat-dose glucose data used two-sample, two-sided t-tests with Bonferroni correction. Sample sizes for each experiment are specified in the figure legends.

• Protein engineering and codon optimization: Among 20 variants screened, hG6Pase-α S298C showed >2-fold higher expression and activity than WT in HeLa cells while retaining proper ER localization. Codon optimization further increased expression and activity in vitro (Hep3B) and in vivo (CD-1 mice), with the combined S298C_CO yielding the greatest hepatic protein and activity.
• Pharmacokinetics in WT mice: Hepatic hG6PC S298C mRNA half-life ~20 h; G6Pase-α S298C protein T1/2 ~79 h; enzymatic activity T1/2 ~74 h. Protein/activity peaked at 24 h and were detectable up to 7 days.
• Single-dose efficacy in L.G6pc-/- mice: hG6PC S298C mRNA (0.2, 0.5, 1.0 mg/kg) significantly improved fasting glycemia at 2.5, 6, and 24 h versus eGFP, maintaining glucose above the therapeutic threshold (60 mg/dL). Liver morphology normalized, liver weight decreased, hepatic G6P, glycogen, and triglycerides were significantly reduced, and serum triglycerides decreased, consistent with dose-dependent increases in hepatic G6Pase-α protein and activity.
• Duration of action: Single doses (≥0.5 mg/kg) improved fasting glucose through days 0, 2, 4, with partial effects through days 7–10; no difference by day 14.
• Repeat dosing: Five i.v. doses of 0.25 mg/kg every 10–14 days sustained significant improvements in fasting glycemia across the 8-week period without loss of effect upon redosing.
• Safety: No increases in serum IFNγ, IL-1β, TNFα, or IL-6 after single or repeat dosing; a trend toward improved ALT; no appreciable anti-G6Pase-α antibody response after five doses; no hypersensitivity signs or adverse effects on body weight.
• Tumor prevention: Under HF/HS diet, macroscopic tumor incidence in L.G6pc-/- controls was 16/26 (~58%) versus 8/34 (~23%) with hG6PC S298C mRNA. Tumors per mouse and total tumor burden (summed lesion area) were significantly reduced by treatment. HCA/HCC biomarkers (PKM2, β-catenin, p62) and related gene expression changes were partially reversed, and serum AFP decreased. Concurrently, fasting glycemia control was maintained during chronic dosing.

The study addresses the central therapeutic need in GSDIa—restoring hepatic G6Pase-α function to prevent life-threatening fasting hypoglycemia and mitigate long-term hepatic neoplasia. Engineered, codon-optimized hG6PC mRNA delivered by LNPs achieved robust ER-localized G6Pase-α expression and activity with favorable hepatic persistence (protein/activity half-lives of ~3 days), translating into correction of fasting glycemia and normalization of key hepatic and serum biomarkers in a liver-specific G6pc knockout model. Importantly, redosability enabled sustained metabolic control without detectable innate cytokine induction or anti-drug antibodies, supporting chronic administration. Beyond acute metabolic correction, repeated treatment significantly lowered the incidence, multiplicity, and burden of hepatic adenomas/carcinomas in a diet-accelerated tumor model, aligning with the mechanistic premise that restoring G6Pase-α reduces tumorigenic drive. Compared to viral gene therapy, mRNA therapy provides transient, non-integrating expression, avoids vector immunity issues, and supports redosing, potentially overcoming limitations such as dilution of episomal vectors in proliferating hepatocytes and transcriptional dysregulation in adenomas reported for AAV approaches. These findings substantiate mRNA therapy as a viable modality for GSDIa with potential disease-modifying effects.

Engineered and codon-optimized hG6PC mRNA delivered via LNPs restored hepatic G6Pase-α function, normalized fasting glycemia and metabolic biomarkers, and reduced hepatic tumor risk in a murine GSDIa model. The therapy was well-tolerated with no evidence of innate cytokine induction or anti-drug antibody formation upon repeat dosing. These results support continued development of mRNA therapy for GSDIa as a chronic, redosable alternative that may address both acute metabolic control and long-term hepatic complications. Future work should include larger-animal safety studies, dose optimization and scheduling for humans, and evaluation of effects on preexisting hepatic adenomas.

• Preclinical evidence is limited to murine models; translatability of efficacy, dosing frequency, and long-term safety to humans remains to be established.
• Dosing frequency inferred for humans relies on allometric scaling; actual human pharmacokinetics/pharmacodynamics may differ.
• Larger-animal (rat, nonhuman primate) safety and immunogenicity studies are warranted but not reported here.
• The tumor study focused on prevention/attenuation under diet-accelerated conditions; effects on established adenomas/HCCs were not directly tested and remain speculative.
• Long-term immunogenicity and safety with chronic lifetime redosing require further evaluation.