Construction of a self-amplifying mRNA-LNP complex for phenylalanine hydroxylase and its efficacy evaluation in gene therapy for phenylketonuria Authors: Ziqi Wen, Yiling Long, Wenjing Shi, Wenjing Song, Rongjun Xu, Xingri Zhan, Xiuyuan Wang, Rui Su, Yu Wang, Jia Fei Published: 2026 ABSTRACT SummaryPhenylketonuria (PKU) is an autosomal recessive disorder caused by deficient phenylalanine hydroxylase (PAH) activity, with current pharmacotherapies showing limited response rates and age-dependent restrictions. We compared two liver-targeted lipid nanoparticle formulations delivering PAH-encoding mRNA—a self-amplifying construct (saRNA) and a conventional non-replicating counterpart (nr-mRNA)—in Pah-knockout mice. Both constructs reduced circulating phenylalanine (Phe) after a single intravenous injection. The saRNA formulation achieved comparable efficacy at one-sixth the nr-mRNA dose and maintained physiological Phe levels for 14 days, whereas nr-mRNA produced only transient reductions. Repeated saRNA administration was well tolerated, with no detectable immunogenicity or hepatic toxicity. These results establish that saRNA-mediated PAH replacement provides sustained activity with substantial dose reduction, overcoming the short-lived effect of non-replicating mRNA. This platform may inform broader strategies for durable enzyme therapy in inherited metabolic diseases requiring liver-restored function. Introduction Phenylketonuria (PKU) is an inherited metabolic disorder caused by mutations in the hepatic PAH gene, leading to impaired conversion of phenylalanine (Phe) to tyrosine (Tyr).1 The resulting accumulation of Phe in blood and brain causes severe clinical manifestations, including irreversible intellectual disability, autism spectrum disorders, motor dysfunction, hypopigmentation, and characteristic musty odor of urine.2,3 PKU exhibits marked genetic heterogeneity, with an estimated incidence of approximately 1:10,000–15,000 live births in Caucasian and East Asian populations.4 Early diagnosis through newborn screening has significantly improved disease management; however, current therapeutic options remain limited.2 Lifelong dietary restriction of Phe is the mainstay treatment but is difficult to maintain and often associated with poor compliance.5 Pharmacological interventions, including tetrahydrobiopterin supplementation and pegylated Phe ammonia lyase, benefit only a subset of patients or are restricted by safety concerns such as immunogenicity. These limitations highlight the urgent need for more effective and broadly applicable therapeutic strategies.6,7 The remarkable allelic diversity of the PAH gene, which spans approximately 90 kb and contains 13 exons, poses a major challenge for mutation-specific therapies, such as gene editing.8 Although genome editing technologies have shown promise, their clinical application remains constrained by the need for individualized design and concerns regarding safety and off-target effects. In contrast, mRNA-based protein replacement therapy offers a mutation-independent approach, enabling transient, non-integrating expression of functional PAH protein and thereby providing a safer and more versatile therapeutic alternative.9 In 2021, the FDA approval of BioNTech’s BNT162b2 mRNA vaccine for COVID-19 prevention in individuals aged ≥16 years marked a milestone in the clinical translation of mRNA technology,10 accelerating its expansion from infectious diseases into oncology and metabolic disorders. In this context, mRNA-based therapeutics offer a reversible, non-integrating alternative with a favorable safety profile. Recent advances have further introduced saRNA platforms, which have been widely investigated in vaccine development due to their ability to achieve robust and sustained protein expression at lower doses. Notably, the innovation of the present study lies in the effective application of saRNA not as a prophylactic vaccine but as a therapeutic modality for protein replacement, prolonging PAH protein expression and reducing dosing frequency, thereby providing a more sustainable and scalable treatment paradigm for genetically heterogeneous metabolic disorders such as PKU. Structurally, saRNAs (9–12 kb) are more complex than conventional non-replicating mRNAs (nr-mRNAs), containing core mRNA elements (5′cap, 5′UTR, 3′UTR, and poly(A) tail) plus a large open reading frame (ORF). The ORF encodes Venezuelan equine encephalomyelitis virus (VEEV)-derived nonstructural proteins (nsP1–4), a subgenomic promoter, and the target gene (replacing viral structural genes).11 Lack of structural proteins prevents infectious virus formation while preserving self-amplification capacity.12 SaRNA translation initiates with nsP2-mediated cleavage of the nsP1–4 polyprotein into nsP1–3 and nsP4, forming the early RNA-dependent RNA polymerase (RDRP) for negative-strand RNA synthesis. In late replication, full hydrolysis of nsP1–4 yields nsP1–4, which assemble into a positive-strand-specific RDRP complex. This complex uses negative-strand templates to synthesize full-length genomic RNA and abundant subgenomic RNAs (sgRNAs) via the subgenomic promoter, driving high-level target protein expression (Figure 1A). SaRNAs offer advantages including low production costs, high efficacy, facile synthesis, and scalability. Low-dose administration prolongs protein expression—addressing the transience of nr-mRNAs13,14—while mitigating immunogenicity.Figure 1Comparison of PAH sa-mRNA-LNP and PAH nr-mRNA-LNP at the cellular level(A) Production and localization of PAH in HEK 293T and HepG2 cells. HEK 293T and HepG2 cells were transfected with PAH sa-mRNA-LNP and PAH nr-mRNA-LNP after 24 h. PAH protein stained with dye-conjugated (A594) anti-PAH antibody was visualized by EVOS FL Auto, which appears red. The cell nucleus is stained with DAPI and appears blue. Scale bars, 50 μm.(B and C) Changes in the levels of PAH mRNA and protein expression in HEK293T cells and HepG2 cells. (B) PAH sa-mRNA-LNP and PAH nr-mRNA-LNP were transfected into HEK 293T cells and HepG2 cells, and the expression level of PAH mRNA was detected after 6, 24, 48, 72, and 96 h by RT-qPCR. Data are presented as mean ± SEM; n = 3. (C) PAH sa-mRNA-LNP and PAH nr-mRNA-LNP were transfected into HEK293T cells and HepG2 cells, and changes in PAH protein expression levels were detected after 6, 24, 48, 72, and 96 h by western blot. Lanes 1, 3, 5, 7, and 9 were the PAH nr-mRNA-LNP transfection group; lanes 2, 4, 6, 8, and 10 were the PAH sa-mRNA-LNP transfection group; and lane 11 was the BK group.(D) Comparison of PAH enzyme activity produced by PAH sa-mRNA LNP and PAH nr-mRNA LNP in vitro. Two types of RNA were transfected into HEK293T cells and HepG2 cells. After 1–5 days, the cell lysate was collected to detect the activity of Phe decomposition, and an untreated control group was set up. No significant differences, ns; p > 0.5; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; PAH sa-mRNA-LNP group compared with PAH nr-mRNA-LNP group; mean ± SEM; n = 3. Comparison of PAH sa-mRNA-LNP and PAH nr-mRNA-LNP at the cellular level (A) Production and localization of PAH in HEK 293T and HepG2 cells. HEK 293T and HepG2 cells were transfected with PAH sa-mRNA-LNP and PAH nr-mRNA-LNP after 24 h. PAH protein stained with dye-conjugated (A594) anti-PAH antibody was visualized by EVOS FL Auto, which appears red. The cell nucleus is stained with DAPI and appears blue. Scale bars, 50 μm. (B and C) Changes in the levels of PAH mRNA and protein expression in HEK293T cells and HepG2 cells. (B) PAH sa-mRNA-LNP and PAH nr-mRNA-LNP were transfected into HEK 293T cells and HepG2 cells, and the expression level of PAH mRNA was detected after 6, 24, 48, 72, and 96 h by RT-qPCR. Data are presented as mean ± SEM; n = 3. (C) PAH sa-mRNA-LNP and PAH nr-mRNA-LNP were transfected into HEK293T cells and HepG2 cells, and changes in PAH protein expression levels were detected after 6, 24, 48, 72, and 96 h by western blot. Lanes 1, 3, 5, 7, and 9 were the PAH nr-mRNA-LNP transfection group; lanes 2, 4, 6, 8, and 10 were the PAH sa-mRNA-LNP transfection group; and lane 11 was the BK group. (D) Comparison of PAH enzyme activity produced by PAH sa-mRNA LNP and PAH nr-mRNA LNP in vitro. Two types of RNA were transfected into HEK293T cells and HepG2 cells. After 1–5 days, the cell lysate was collected to detect the activity of Phe decomposition, and an untreated control group was set up. No significant differences, ns; p > 0.5; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; PAH sa-mRNA-LNP group compared with PAH nr-mRNA-LNP group; mean ± SEM; n = 3. NrRNAs have shown efficacy in treating COVID-19, but they require high doses. For example, Moderna’s mRNA-1273 is administered at 100 μg,15 even though nucleoside modifications have been used to enhance translatability and reduce inflammation. High doses not only increase costs but also raise the risk of adverse events. In contrast, saRNAs—a newer mRNA platform—exhibit superior therapeutic potential in protein replacement, enabling low-dose, sustained efficacy with reduced toxicity. Naked RNA is vulnerable to RNase degradation, necessitating delivery systems. Lipid nanoparticles (LNPs)—bilayer-based nanocarriers—efficiently and safely deliver therapeutic RNAs to target cells, comprising four key components: ionizable cationic lipids, helper lipids, cholesterol, and PEGylated lipids. Compared to viral vectors, LNPs offer multi-faceted advantages: improved safety (such as no genomic integration risk), biodegradable lipids, and low toxicity.16 LNPs have a highly efficient encapsulation rate of nucleic acids as well as good protective power, and more than 90% nucleic acid encapsulation rate can be achieved by using ionizable lipid SM-102, which encapsulates nucleic acids within a hydrophobic core and effectively avoids the degradation of RNase.17 Additionally, the lipid ratio, surface modification, and particle size of LNP can be flexibly regulated to load different RNAs, such as siRNAs,18 microRNAs,19 and CRISPR/Cas9 complexes.20 It can also be made to target different target organs, such as the liver, lung,21 and pancreas.22 Currently, LNPs have a well-established, specialized production pipeline. They demonstrate optimal delivery efficiency in the liver, a key advantage for PKU gene therapy. Since hepatocytes are the primary site where natural PAH is expressed and functions, LNPs emerge as an ideal vehicle for delivering PAH saRNA to these cells. Using a Pah-knockout (KO) mouse model (exhibiting classical PKU phenotypes: stunted growth, hypopigmented coat, and hyperphenylalaninemia), we evaluated two complexes: PAH sa-mRNA-LNP and N1-methylpseudouridine-modified (m1Ψ) PAH nr-mRNA-LNP. Single tail vein injections of both high/low doses normalized serum Phe in Pah-KO mice, but nr-mRNA-LNP showed Phe rebound by day 2, whereas sa-mRNA-LNP maintained Phe < 360 μmol/L. Assessments included body weight/coat color monitoring, serial blood sampling, and post-study liver analyses. Repeated low-dose sa-mRNA-LNP injections further validated efficacy and safety. These data will provide support for further saRNA in PKU gene therapy clinical trials. Results Comparative evaluation of in vitro translatability of PAH sa-mRNA-LNP and PAH nr-mRNA-LNP HEK293T and HepG2 cells were transfected with PAH sa-mRNA-LNP and PAH nr-mRNA-LNP, respectively. HEK293T cells, known for high transfection efficiency, served as a protein expression validation model, while HepG2 cells, exhibiting hepatocyte-like properties, mimicked liver cell transfection. At 24 h post-transfection, cellular imaging revealed 4′,6-diamidino-2-phenylindole (DAPI)-stained nuclei in blank control (BK) cells with negligible PAH channel fluorescence, confirming basal PAH expression. In contrast, cells transfected with either mRNA-LNP complex showed robust cytoplasmic PAH fluorescence colocalized with nuclear DAPI staining (Figure 1A), indicating successful protein expression and cytoplasmic localization. Both constructs demonstrated translational competence in HEK293T and HepG2 cells, though cell-type- and mRNA-specific expression variances were observed, warranting further investigation in subsequent experiments. To compare the expression kinetics of PAH sa-mRNA-LNP and PAH nr-mRNA-LNP, HEK 293T and HepG2 cells were separately transfected with these constructs, alongside BK control groups. Cells were harvested at 6, 24, 48, 72, and 96 h post-transfection for qPCR analysis of PAH mRNA levels and Western blotting of protein expression. In both HEK293T and HepG2 cells, PAH mRNA expression exhibited a transient peak followed by a decline, with notable differences between delivery systems. Relative to basal levels in BK controls, transfection with PAH sa-mRNA-LNP or PAH nr-mRNA-LNP significantly upregulated PAH mRNA. In HEK293T cells, the nr-mRNA-LNP group showed a rapid 6 h peak (3.55 × 105-fold vs. BK), whereas the sa-mRNA-LNP group reached a higher peak (4.10 × 105-fold) at 48 h, maintaining elevated expression through 96 h HepG2 cells mirrored this pattern: nr-mRNA-LNP peaked at 6 h (1.40 × 105-fold) before declining, while sa-mRNA-LNP peaked at 24 h (2.13 × 105-fold) with sustained expression (Figure 1B). Western blot analysis correlated with mRNA kinetics: neither construct induced significant protein expression within 6 h. The nr-mRNA-LNP group showed a transient protein peak at 24 h (1.45-fold vs. BK), whereas the sa-mRNA-LNP group exhibited delayed yet sustained expression, peaking at 72 h (2.7-fold vs. nr-mRNA-LNP) and remaining elevated throughout the time course (Figure 1B). Western blot analysis validated sustained PAH protein expression mediated by sa-mRNA-LNP across HEK293T and HepG2 cell lines. Neither constructs induced significant protein changes within 6 h post-treatment. In HEK293T cells, PAH nr-mRNA-LNP peaked at 24 h (2.78-fold vs. BK), declining thereafter, whereas PAH sa-mRNA-LNP reached a higher peak at 72 h (4.7-fold vs. BK) and maintained elevated expression at 72–96 h (2.09- and 1.92-fold vs. nr-mRNA-LNP, respectively). HepG2 cells mirrored this pattern: nr-mRNA-LNP peaked at day 1 (1.45-fold vs. BK), indistinguishable from sa-mRNA-LNP at this time point, while sa-mRNA-LNP exhibited a 72-h peak (2.7-fold vs. nr-mRNA-LNP) with sustained expression throughout the time course (Figure 1C). These results underscore the sa-mRNA-LNP system’s superiority in achieving prolonged, high-level protein production across cell types. To detect the enzymatic activity of PAH produced in vitro, we verified it by the change in the content of the substrate Phe. Cell lysates were extracted for assay on days 1–5 post-transfection, respectively. Both mRNA-LNP constructs significantly reduced Phe levels by day 1 in HEK 293T; however, the PAH sa-mRNA-LNP group maintained sustained Phe depletion through day 5, peaking at day 3 with a 90% reduction relative to baseline. Conversely, the PAH nr-mRNA-LNP group exhibited a transient effect, with Phe levels rebounding by day 3, indicating rapid degradation of the encoded enzyme (Figure 1D). The changes in the levels of Phe were slightly different for different cell types, but the trends were similar, indicating that different cell types may have an effect on the effectiveness of the two RNA-LNPs in reducing Phe; however, this effect was significantly smaller than the difference between the two RNA types. A single intravenous injection of PAH sa/nr-mRNA-LNP showed a therapeutic effect in Pah-KO mice Figure 2A depicts the schematic illustration of the study design. Pah-KO mice exhibit classic PKU phenotypes, including reduced body weight and hypopigmented fur (Figures 2B and S4). Six-week-old homozygous (HO) Pah-KO mice (both female and male) were randomly allocated to five groups: PBS control, PAH sa-mRNA-LNP (10 μg/animal, medium dose; 60 μg/animal, high dose), and PAH nr-mRNA-LNP (10 μg/animal, medium dose; 60 μg/animal, high dose). An additional heterozygous (HE) blank control group was set up. After 6 weeks, both dose groups of PAH sa-mRNA-LNP successfully converted Phe to Tyr, resulting in fur color darkening and normalization. Notably, no significant differences were observed between the nr-mRNA-LNP groups and the HO-PBS control (Figures 2C and S5).Figure 2Single tail vein injection of Pah-KO mice(A) Experimental protocol. Serum samples were collected 5 days before treatment and on days 1, 2, 4, 7, 14, and 42 after treatment for Phe detection.(B) Comparison of coat color between homozygous (HO) and heterozygous (HE) Pah-KO mice with the PKU phenotype.(C) Changes in coat color in PAH sa-mRNA-LNP mice in different treatment groups 6 weeks after a single tail-vein injection. Regardless of the dose, 60 μg or 10 μg, the coat color of the mice darkened and tended to normalize.(D) Changes in serum Phe levels in each group. Data are presented as mean ± SEM; n = 4.(E) Changes in body weight in the PAH sa-mRNA-LNP treatment group 6 weeks after a single tail-vein injection. Data are presented as mean ± SEM; n = 4.(F) Immunohistochemical staining of the liver. Immunohistochemical staining of PAH protein in liver tissue was performed. Dark brown represents positive staining areas of PAH protein, and light blue represents counterstaining of cell nuclei by hematoxylin (10-fold). Scale bars, 200 μm.(G) H&E staining of the liver. Six weeks after treatment, Pah-KO mice were dissected and livers were removed. Paraffin-embedded sections were stained with H&E (10x). Scale bars, 200 μm.(H) Serum ALT levels in mice remained stable at 6 h after treatment, with no significant differences (ns, p > 0.5; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; comparison between PAH sa-mRNA-LNP group and PAH nr-mRNA-LNP group; mean ± SEM; n = 4). Single tail vein injection of Pah-KO mice (A) Experimental protocol. Serum samples were collected 5 days before treatment and on days 1, 2, 4, 7, 14, and 42 after treatment for Phe detection. (B) Comparison of coat color between homozygous (HO) and heterozygous (HE) Pah-KO mice with the PKU phenotype. (C) Changes in coat color in PAH sa-mRNA-LNP mice in different treatment groups 6 weeks after a single tail-vein injection. Regardless of the dose, 60 μg or 10 μg, the coat color of the mice darkened and tended to normalize. (D) Changes in serum Phe levels in each group. Data are presented as mean ± SEM; n = 4. (E) Changes in body weight in the PAH sa-mRNA-LNP treatment group 6 weeks after a single tail-vein injection. Data are presented as mean ± SEM; n = 4. (F) Immunohistochemical staining of the liver. Immunohistochemical staining of PAH protein in liver tissue was performed. Dark brown represents positive staining areas of PAH protein, and light blue represents counterstaining of cell nuclei by hematoxylin (10-fold). Scale bars, 200 μm. (G) H&E staining of the liver. Six weeks after treatment, Pah-KO mice were dissected and livers were removed. Paraffin-embedded sections were stained with H&E (10x). Scale bars, 200 μm. (H) Serum ALT levels in mice remained stable at 6 h after treatment, with no significant differences (ns, p > 0.5; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; comparison between PAH sa-mRNA-LNP group and PAH nr-mRNA-LNP group; mean ± SEM; n = 4). Serum Phe baseline levels were measured 5 days prior to injection, confirming that HO Pah-KO mice exhibited characteristically high Phe levels, while the HE-BK group showed normal levels. Phe dynamics were assessed on days 1, 2, 4, 7, 14, and 42 post-injection (Figure 2D). As expected, the HO-PBS control group maintained high Phe levels with only slight fluctuations, validating its role as a negative control. In the high-dose (60 μg) groups, both PAH sa-mRNA-LNP and PAH nr-mRNA-LNP significantly reduced Phe levels (to 580 μmol/L) on day 1. Phe levels in the PAH sa-mRNA-LNP group remained low until day 14, while the nr-mRNA-LNP group rebounded on day 2 and recovered to Pah-KO baseline levels by day 4. In the low-dose (10 μg) group, PAH nr-mRNA-LNP reduced Phe levels to 1,000 μmol/L on day 1, followed by a gradual increase. The PAH sa-mRNA-LNP group also reached 1,000 μmol/L on day 1, with a smaller initial decrease than the high-dose group, but remained at a relatively low level thereafter. By day 2, Phe levels in all treatment groups were close to normal and remained comparable to the hyperfunctional hepatitis (HE) control group 6 weeks post-injection, confirming the sustained efficacy of PAH sa-mRNA-LNP at different doses. Further evaluation is needed to confirm its Phe conversion capacity. Body weight was continuously monitored during the 6-week treatment period; no significant differences were found between groups, and no dose-response relationship was observed, but a trend toward post-treatment weight gain was observed (Figures 2E and S6). After 6 weeks of treatment, mouse liver tissues from each group were subjected to immunohistochemical (IHC) and hematoxylin-eosin (H&E) staining to assess PAH expression and histopathological changes, respectively. IHC analysis showed robust PAH-positive staining in HE mice, whereas, compared to the HO-PBS control group, both the HO 60 μg and HO 10 μg PAH sa-mRNA-LNP groups exhibited distinct brown staining with increased intensity, thereby confirming PAH expression in the liver and validating LNP-mediated hepatic targeting, this was further corroborated by qPCR analysis of PAH expression in the liver (Figures 2F and S11). The high-dose (60 μg) group had larger, darker brown areas (21.02% positive area), indicating higher PAH expression; the 10 μg group (13.97% positive area) showed less extensive and lighter staining than the high-dose group but significantly stronger signals than the HO-PBS control (8.04%) and intensity comparable to the HE-BK group (14%), confirming effective PAH expression even at lower doses (Figure S7). In contrast, after a single intravenous injection of PAH nr-mRNA-LNP for 6 weeks, liver IHC staining results were not different from those of blank HO mice (Figure S8). Concurrently, H&E staining revealed that hepatocytes in the HE-BK, HO-PBS, PAH sa/nr-mRNA-LNP HO-60 μg and 10 μg groups maintained regular morphology with distinct nuclei, without obvious pathological alterations (e.g., cellular swelling, necrosis, or inflammatory cell infiltration) (Figures 2G and S6). Hepatic lobules remained structurally intact, with no abnormalities in central veins or portal areas, and hepatic sinusoidal endothelial cells were neatly arranged. These findings indicate that 10 μg or 60 μg PAH sa-mRNA-LNP treatment induces effective hepatic PAH expression without adverse effects on liver tissue structure or cellular morphology, supporting a favorable safety profile of PAH sa-mRNA-LNP in PKU treatment within a certain dose range. Researchers collected blood samples at 6-h time points on day 5 before treatment and days 7 and 14 after treatment to measure ALT and AST levels (Figures 2H and S10). As key indicators of hepatocellular injury, ALT and AST levels remained relatively stable in all groups of mice, with no significant increases observed. Anti-PAH antibody detection showed an increase by day 7, though this appeared to be dose-dependent (Figure S12). These results indicate that high/medium doses of PAH sa-mRNA-LNP or PAH nr-mRNA-LNP, and the extended observation period of up to 14 days, did not induce significant hepatocellular injury in mice. Repeated intravenous injection of PAH sa/nr-mRNA-LNP showed a therapeutic effect in Pah-KO mice To investigate the impact of PAH sa/nr-mRNA-LNP on immune responses, a 5-day treatment cycle was adopted (Figure 3A), involving three tail vein injections of 10 μg PAH sa-mRNA-LNP and PAH nr-mRNA-LNP. On day 11 of the experiment, mice in the repeated injection groups were observed: compared with the HO-PBS group, both the PAH sa-mRNA-LNP and HO-PAH nr-mRNA-LNP groups exhibited deepened and normalized fur color, with the effect being more pronounced in the PAH sa-mRNA-LNP group (Figure 3B). Serum samples were collected 5 days before the experiment and 24 h after each injection to measure Phe levels (Figure 3C). Results showed that serum Phe was significantly reduced 24 h after each injection, returning to levels comparable to those in non-PKU mice. These findings indicate that repeated low-dose (10 μg) injections of PAH sa-mRNA-LNP effectively lowered serum Phe in Pah-KO mice, gradually restoring it to near-normal levels. In contrast, 10 μg PAH nr-mRNA-LNP only reduced Phe to approximately 1,000 μmol/L. Additionally, no significant differences in body weight were observed among the treatment groups throughout the 11-day period (Figure 3D).Figure 3Repeated tail-vein injections in Pah-KO mice(A) Experimental schedule.(B) A comparison of fur color was conducted among HO-PBS group mice (with PKU phenotype), mice receiving repeated PAH sa-mRNA-LNP injections, PAH nr-mRNA-LNP group mice, and HE-BK group mice.(C) Changes in serum Phe levels in each group. Serum was taken for Phe 5 days before treatment and the day after each injection. Data are presented as mean ± SEM; n = 3.(D) Body weight changes in different treatment groups after repeated tail vein injections. Data are presented as mean ± SEM; n = 3.(E and F) Liver was taken after 11 days for IHC staining and H&E staining. Scale bars, 200 μm.(G) IL-6 cytokine levels. Blood samples were collected at 6 and 24 h following three repeated intravenous injections to measure serum IL-6 levels in mice from each group (ns, p > 0.5, ∗∗∗∗p < 0.0001; PAH sa-mRNA-LNP group compared with PAH nr-mRNA-LNP group; mean ± SEM; n = 3). Repeated tail-vein injections in Pah-KO mice (A) Experimental schedule. (B) A comparison of fur color was conducted among HO-PBS group mice (with PKU phenotype), mice receiving repeated PAH sa-mRNA-LNP injections, PAH nr-mRNA-LNP group mice, and HE-BK group mice. (C) Changes in serum Phe levels in each group. Serum was taken for Phe 5 days before treatment and the day after each injection. Data are presented as mean ± SEM; n = 3. (D) Body weight changes in different treatment groups after repeated tail vein injections. Data are presented as mean ± SEM; n = 3. (E and F) Liver was taken after 11 days for IHC staining and H&E staining. Scale bars, 200 μm. (G) IL-6 cytokine levels. Blood samples were collected at 6 and 24 h following three repeated intravenous injections to measure serum IL-6 levels in mice from each group (ns, p > 0.5, ∗∗∗∗p < 0.0001; PAH sa-mRNA-LNP group compared with PAH nr-mRNA-LNP group; mean ± SEM; n = 3). On day 11 following repeated injections, liver IHC staining results showed that repeated injections resulted in the production of large amounts of PAH in the liver (Figure 3E). H&E staining of mouse livers (Figure 3F) revealed regular hepatocyte morphology with clear nuclei in the HE-BK, HO-PBS, HO-PAH sa-mRNA-LNP, and HO-PAH nr-mRNA-LNP groups, with no obvious pathological changes such as cellular swelling, necrosis, or inflammatory cell infiltration. These observations suggest that PBS injection, low-dose PAH sa-mRNA-LNP, and PAH nr-mRNA-LNP had minimal impact on liver tissue structure or cellular morphology, and repeated injections were well tolerated. Following three repeated injections, serum levels of the cytokine IL-6 were measured at 6 and 24 h, respectively (Figure 3G). The results indicated that, although the treatment groups exhibited significant differences in IL-6 levels at the 6-h mark, these levels had returned to the normal range by 24 h. Discussion This study successfully constructed and evaluated two liver-targeted LNP formulations for gene therapy in PKU: a sa-mRNA encoding PAH and a traditional nr-mRNA counterpart. Both formulations demonstrated significant efficacy in reducing Phe levels in Pah-KO murine models, with the PAH sa-mRNA-LNP exhibiting superior dose-sparing effects and sustained therapeutic outcomes. Specifically, the PAH sa-mRNA-LNP achieved equivalent Phe reduction at one-sixth the dosage (10 μg vs. 60 μg) compared to the nr-mRNA-LNP, maintaining physiological Phe concentrations for up to 42 days post-administration. Repeated dosing of PAH sa-mRNA-LNP further validated its robust tolerability and lack of immunogenicity or toxicity. These findings underscore the potential of saRNA-based therapies as a transformative approach for PKU and other inborn errors of metabolism. The PAH sa-mRNA-LNP demonstrated distinct advantages over its nr-mRNA-LNP counterpart across multiple dimensions. In terms of efficacy, sa-mRNA-LNP showed prolonged and enhanced protein expression: in vitro, it exhibited delayed but sustained PAH mRNA and protein peaks, correlating with prolonged enzymatic activity and Phe reduction; in vivo, a single dose maintained near-normal Phe levels for 42 days, whereas nr-mRNA-LNP-treated mice showed rapid rebound by day 2, attributed to saRNA’s ability to amplify intracellular RNA copies. Furthermore, LNPs with neutral surface charge evade RNase-mediated degradation and macrophage phagocytosis in vivo, enter the cytoplasm via endocytosis, and release encapsulated RNA through the proton sponge effect—facilitated by lysosomal acidification-induced membrane disruption—enabling more translation of saRNAs in the cytoplasm of the cell (Figure 4). The dose-sparing capability is another critical advancement, as 10 μg of sa-mRNA-LNP achieved efficacy comparable to 60 μg of nr-mRNA-LNP, reducing production costs and potential toxicity risks. Additionally, both formulations showed excellent safety profiles, with stable serum ALT/AST levels, no liver damammation in histological analyses, andflammation in histological analyses, and repeated sa-mRNA-LNP dosing not elevating IL-6, confirming minimal immunogenicity. Moreover, the SM-102 LNP system effectively delivered PAH mRNA to hepatocytes, as evidenced by hepatic PAH expression in immunohistochemistry, highlighting its organ-specific targeting, crucial for PKU.Figure 4Schematic diagram of LNP-encapsulated saRNA functioning in micesaRNA contains a replicase region encoding non-structural proteins (nsP1–nsP4), which form an RNA-dependent RNA polymerase (RDRP), and a subgenomic promoter driving the expression of the target ORF. After entering the cytoplasm, the genomic (+) RNA is first translated to produce the replicase complex. The RDRP then synthesizes complementary negative-strand RNA, which serves as a template for amplification of genomic RNA and generation of subgenomic RNA. The subgenomic RNA is subsequently translated into large amounts of target protein, enabling signal amplification.For in vivo delivery, saRNA is encapsulated into LNP, which protect it from RNase degradation in circulation and facilitate systemic administration. After reaching target tissues, LNP are internalized by cells via endocytosis. Following endosomal acidification, LNPs destabilize and release saRNA into the cytoplasm, where translation occurs. Schematic diagram of LNP-encapsulated saRNA functioning in mice saRNA contains a replicase region encoding non-structural proteins (nsP1–nsP4), which form an RNA-dependent RNA polymerase (RDRP), and a subgenomic promoter driving the expression of the target ORF. After entering the cytoplasm, the genomic (+) RNA is first translated to produce the replicase complex. The RDRP then synthesizes complementary negative-strand RNA, which serves as a template for amplification of genomic RNA and generation of subgenomic RNA. The subgenomic RNA is subsequently translated into large amounts of target protein, enabling signal amplification. For in vivo delivery, saRNA is encapsulated into LNP, which protect it from RNase degradation in circulation and facilitate systemic administration. After reaching target tissues, LNP are internalized by cells via endocytosis. Following endosomal acidification, LNPs destabilize and release saRNA into the cytoplasm, where translation occurs. The findings of this study have important implications and point to several future directions. Clinically, the study provides a robust preclinical foundation for saRNA-based PKU therapy, with future work focusing on phase I trials to evaluate safety and pharmacokinetics in humans, leveraging the dose-sparing advantage to minimize side effects. In terms of broader applications, the success of sa-mRNA-LNP in PKU suggests its applicability to other monogenic metabolic disorders (e.g., tyrosinemia23 and urea cycle defects24), and the platform’s modularity allows for rapid adaptation of target genes, potentially accelerating therapy development. Mechanistically, further research should explore saRNA design modifications (e.g., UTR optimization and nsP1-4 engineering) to enhance translational efficiency and reduce innate immune activation, and combining saRNA with novel LNP formulations (e.g., targeting extrahepatic tissues) could expand the therapeutic scope. Additionally, future studies must incorporate longer follow-ups to assess sustained efficacy, immune tolerance, and off-target effects, with companion biomarkers (e.g., anti-drug antibodies and cytokine profiles) being critical for clinical monitoring. Overall, this study establishes PAH sa-mRNA-LNP as a promising therapeutic candidate for PKU. Limitations of the study This study has several limitations that need to be acknowledged. Firstly, the observation window for single-dose effects (42 days) and repeated dosing (11 days) is relatively short, providing limited insights into long-term safety and efficacy, which is essential for chronic PKU management to assess durability, potential antibody formation, and cumulative toxicity. Secondly, the current study was limited by the small sample size of mice used in dosage and administration frequency evaluations. This constraint may restrict the robustness of conclusions regarding the optimal therapeutic dosage and administration interval that balances efficacy and safety. A larger sample size would enhance statistical power, enabling more precise characterization of dose-response relationships and the identification of administration regimens that maximize therapeutic outcomes while minimizing potential risks. Resource availability Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jia Fei (tfeijia@jnu.edu.cn). Materials availability No unique materials were generated in this study. Data and code availability All data associated with this study are presented in the paper or the supplemental information. No datasets were deposited in public repositories. Materials that support the findings of this study are available from the lead contact upon reasonable request. Acknowledgments We thank Yaohai Bio-pharma (Jiangsu, China) for providing LNP service and the Medical Experimental Center, School of Medicine, Jinan University (Guangzhou, China), for providing the Instrument platform. This work was supported by grants from the Shenzhen Science and Technology Major Project 2024 (KJZD20240903100304007); Science and Technology Program of Guangzhou City (grant no. 202206010063); and the 2025 faculty-level Postgraduate Research Project of the School of Basic Medicine and Public Health, Jinan University (“Strengthening Talents” project). Author contributions All named authors have read and approved the manuscript, and no individuals meeting authorship criteria have been omitted. The listed author order has also been approved by all contributors. Specifically, J.F. and Y.W. conceptualized the study and designed the experiments, while Z.W., Y.L., W. Shi, and W. Song conducted the experimental work. Data analysis was performed by Z.W., Y.L., W.S., X.Z., R.X., and X.W. J.F. and Z.W. provided materials, reagents, and analytical tools, with Z.W., R.S., and J.F. leading manuscript drafting. Y.W. supervised the study, critically reviewed and revised the manuscript, and approved the final version for submission. All data were generated in-house, and no paper mill services were utilized. All authors accept accountability for the integrity and accuracy of all aspects of the work. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesPAH Polyclonal antibodyProteintechCat No. 16347-1-AP; RRID:AB_2878247Mouse anti-GAPDH antibodyAffinityT0004; RRID:AB_2833041Goat anti-rabbit IgG (H + L)-HRP Fluor594-conjugatedAffinityS0002; RRID:AB_2839430Chemicals, peptides, and recombinant proteinsPAH-sa-mRNA-LNPYaohai Bio/Yaohai Bio-PharmaN/APAH-nr-mRNA-LNPYaohai Bio/Yaohai Bio-PharmaN/ACritical commercial assaysECL Chemiluminescence KitAffinityN/AMouse Interleukin 6,IL-6 ELISA KITWUHAN HUAMEI BIOTECH CO.,LTD.CSB-E04639mPhenylalanine Assay KitSigma-Aldrich CorporationMAK484-1 KTDeposited dataDeposited dataThis studyNo data were deposited in public repositoriesExperimental models: Cell linesHEK293T cellsThis laboratoryN/AHepG2 cellsThis laboratoryN/AExperimental models: Organisms/strainsPah knockout miceShanghai Model Organisms CenterNM-KO-2110565OligonucleotidesPAH forward primerBeijing Tsingke Biotech Co., Ltd.N/APAH reverse primerBeijing Tsingke Biotech Co., Ltd.N/AGAPDH forward primerBeijing Tsingke Biotech Co., Ltd.N/AGAPDH reverse primerBeijing Tsingke Biotech Co., Ltd.N/ARecombinant DNAPAH sa-mRNA template plasmidThis studyN/APAH nr-mRNA template plasmidThis studyN/ASoftware and algorithmsGraphPad Prism 9.0GraphPad SoftwareN/AGraphPad SoftwareAdobeN/A REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesPAH Polyclonal antibodyProteintechCat No. 16347-1-AP; RRID:AB_2878247Mouse anti-GAPDH antibodyAffinityT0004; RRID:AB_2833041Goat anti-rabbit IgG (H + L)-HRP Fluor594-conjugatedAffinityS0002; RRID:AB_2839430Chemicals, peptides, and recombinant proteinsPAH-sa-mRNA-LNPYaohai Bio/Yaohai Bio-PharmaN/APAH-nr-mRNA-LNPYaohai Bio/Yaohai Bio-PharmaN/ACritical commercial assaysECL Chemiluminescence KitAffinityN/AMouse Interleukin 6,IL-6 ELISA KITWUHAN HUAMEI BIOTECH CO.,LTD.CSB-E04639mPhenylalanine Assay KitSigma-Aldrich CorporationMAK484-1 KTDeposited dataDeposited dataThis studyNo data were deposited in public repositoriesExperimental models: Cell linesHEK293T cellsThis laboratoryN/AHepG2 cellsThis laboratoryN/AExperimental models: Organisms/strainsPah knockout miceShanghai Model Organisms CenterNM-KO-2110565OligonucleotidesPAH forward primerBeijing Tsingke Biotech Co., Ltd.N/APAH reverse primerBeijing Tsingke Biotech Co., Ltd.N/AGAPDH forward primerBeijing Tsingke Biotech Co., Ltd.N/AGAPDH reverse primerBeijing Tsingke Biotech Co., Ltd.N/ARecombinant DNAPAH sa-mRNA template plasmidThis studyN/APAH nr-mRNA template plasmidThis studyN/ASoftware and algorithmsGraphPad Prism 9.0GraphPad SoftwareN/AGraphPad SoftwareAdobeN/A Experimental model and study participant details Animal model Male and female Pah-KO heterozygous mice with a C57BL/6J genetic background were procured from Shanghai Model Organisms Center, Inc. A Pah-KO genotype mouse model was constructed via CRISPR/Cas9 technology. Pah homozygous knockout mice exhibited phenotypes characteristic of classical PKU disease, including abnormal growth and development, altered coat color, and elevated blood phenylalanine levels. These mice were continuously propagated and bred at the Laboratory Animal Management Center of Jinan University (IACUC-20240531–01, No. self81662) to screen the pure heterozygous mice used for animal experiments. All animal experiments(IACUC-20241223–30) were approved by and conformed to the relevant regulatory standards of the Institutional Animal Care and Use Committee at the Institute of Laboratory Animal Science, Jinan University (Guangzhou, China). Cell lines HEK 293T cells and HepG2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, USA) with high glucose supplemented with 10% fetal bovine serum (FBS; Gibco) and 1% penicillin-streptomycin (P/S; Pricella, China). Cell culture dishes (NEST, China) were placed in an incubator at 37°C for culture, and HEK 293T and HepG2 cells in logarithmic phase were selected for the experiments. All cell lines were routinely tested for mycoplasma contamination and confirmed to be negative. Method details Study design This study aimed to assess the therapeutic efficacy in PKU of two different RNA-LNP complexes, both of which encode the PAH protein involved in Phe and Tyr metabolism. We used PAH nr-mRNAs modified N1-methyluracil (m1Ψ). In addition, we designed a sa-mRNA encoding the same PAH. And encapsulated them using SM-102 cationic lipid LNP, which is hepatic-targeted and suitable for encapsulating mRNAs, to form RNA-LNP complexes. The immunogenicity and efficacy of PAH sa-mRNA-LNP and PAH nr-mRNA-LNP complexes were compared in selected studies. Codon optimization and UTR design PAH codon usage was optimized using algorithms developed at GenSmart™ Codon optimization. The optimized sequences were subjected to GC content analysis (GC content calculator | Yunzhou Bio) and codon adaptation index (MERRYBIO). Optimization of codons in the PAH gene based on human codon usage preferences without changing the amino acid sequence to improve the translation efficiency of both RNAs. The 5′ and 3' UTRs of PAH sa-mRNA retain the original sequence of VEEV. The UTRs found at the 5′ and 3' ends of VEEV contain different core promoter elements for antisense and sense-stranded RNA synthesis, which have important functions in viral replication. With important functions. The 5′ and 3' UTRs of PAH nr-mRNA were selected from the same UTRs of the Pfizer BNT162b2 mRNA vaccine. The 5′ UTR was selected from human alpha bead protein, and the 3' UTR consisted of two fragments derived from human mitochondrial 12S rRNA (mtRNR1) and the human AES/TLE5 gene. mRNA production In this project, the plasmid was constructed based on VEEV viral vector with optimized PAH gene instead of GOI sequence, After linearization of the plasmid, the plasmid was transcribed in vitro by using T7 polymerase, and the cap was added by using the cowpox virus capping enzyme (Vaccinia Capping Enzyme, VCE) method as well as one-step tailing method. The following structures were obtained for PAH sa-mRNA 5′ to 3': cap structure, 5′UTR, nsP1-4 nonstructural protein, 26S subgenic promoter, Pah coding region of the target gene, 3′UTR, and Poly(A). The PAH nr-mRNA 5′ to 3' structure was as follows: hat structure, optimized 5′UTR, target gene Pah coding region, optimized 3′UTR, and Poly(A) with m1Ψ modification. LNP formulation of mRNAs For the formulation of LNP, the well-established SM-102 LNP was selected, with the lipid components (SM-102: DSPC: cholesterol: PEG 2000 = 50: 10: 38.5: 1.5) being rapidly mixed with an aqueous buffer solution containing PAH sa-mRNA or nr-mRNA by microfluidics, resulting in homogeneous lipid complexes and high efficiency of encapsulation, then the feed solution was concentrated to 0.1 mg/mL or 0.3 mg/mL using ultrafiltration and the buffer was changed to a neutral buffer solution to remove unencapsulated mRNA, excess lipids and acetic acid. The particle size and potential of LNP were detected using a Malvern instrument, as well as the RiboGreen assay to determine the encapsulation efficiency. PAH sa/nr-mRNA-LNP using a microfluidic procedure was prepared by Yaohai Bio-pharma (Jiangsu, China). The PAH sa-mRNA-LNP and PAH nr-mRNA-LNP complexes were then stored at −80°C. Cell culture and transfection HEK 293T and HepG2 cells were collected separately and inoculated at 4×105 cells/well in 6-well plates overnight. On the next day, 2.5 μg of PAH sa/nr-mRNA-LNP (0.1 mg/mL) was added to Opti-MEM (Gibco) to formulate 1 mL of the total system. The original medium was aspirated, PBS was washed once, and the transfection system was gently mixed and added to the wall of a 6-well plate. 6 h later, 1 mL of DMEM medium containing 20% FBS was added to restore the 10% FBS culture system. Continue incubation for 6 h, 1–4 days. Quantitative real-time PCR To monitor the temporal changes in the relative amount of PAH mRNA in HEK 293T and HepG2 cells post-transfection, the relative expression levels of mRNA were determined by the 2−ΔΔCt method, with Glyceraldehyde phosphate dehydrogenase (GAPDH), which is a housekeeping gene commonly utilized as a loading control. The primer sequences are as follows: GAPDH F-5′-CAATGACCCCTTCATTGACC-3′, R-5′-GACAAGCTTCCCGTTCTCAG-3';PAH F-5′-GCAGACTTGCACTGGTTTCC-3′, R-5′-ATGGCAGATGTCAGGTTCGG-3'. Western blotting analyses Whole-cell lysates from PAH sa/nr-mRNA-LNP-transfected cells were tested for the presence of the PAH protein at the desired timepoints. An immunoblot was performed on polyvinyl difluoride membranes. PAH was detected using goat anti-PAH polyclonal antibody (Proteintech, China) and Goat anti-rabbit immunoglobulin G (H + L) HRP (Affinity, USA) as the primary and secondary antibodies. Mouse anti-GAPDH antibody (Affinity) was employed as the primary antibody to detect GAPDH. PAH enzymatic assay To assess the in vitro enzymatic activity of PAH, three experimental groups were established: the BK group, the PAH sa-mRNA-LNP group, and the PAH nr-mRNA-LNP group. Each group was transfected for 1 to 5 days, with three replicate wells per group. Cells from each group in 24-well plates were harvested by centrifugation at 1000 rpm for 3 min, and the supernatant was discarded. Subsequently, cells were lysed with 200 μL of RIPA (Beyotime, China) lysis buffer to extract the transfected PAH protein. After a 30 s incubation at room temperature, the lysates were placed on ice. For the enzymatic reaction, 200 μL of cell lysate was incubated at room temperature for 5 min with 80 μL of HEPES buffer (30 mM, Macklin, China), 35 μL of bovine liver catalase(20 mg/mL, Aladdin, China), and 35 μL of L-Phe (10 mM, Solarbio, China). Then, 70 μL of ammonium iron(II) sulfate hexahydrate (100 μM, Sigma, USA) and 250 μL of HEPES (30 mM) were added, followed by a 1 min incubation at room temperature. Finally, 26.4 μL of BH4 (5 mg/mL, KKL) dissolved in 2 mM DTT (Aladdin) was introduced into the reaction mixture, which was stirred at 300 rpm for 3 h at 30°C. The final concentration of the reaction system contained 1 mg/mL catalase, 10 μm iron(II) ammonium sulfate hexahydrate, 0.5 mM L-Phe, and 600 μm BH 4. Then collected and stored at - 80°C for subsequent analysis.25 The relative Phe content in each sample was determined using a Phe Assay kit (Sigma) to evaluate the PAH protein activity. Single-dose treatment The experiments were conducted in the experimental area of the Animal Experiment Management Center of Jinan University using Pah-KO mice aged 6 to 8 weeks. Mice of both genders were randomly divided into four groups, with four mice in each group, as follows: 1) Heterozygous Blank group; 2) Homozygous PBS Injection group; 3) 60 μg-dose PAH sa/nr-mRNA-LNP injection group; 4) 10 μg-dose PAH sa/nr-mRNA-LNP Injection group. For the 60 μg-dose injection, mice were given a single injection of 200 μL of PAH sa/nr-mRNA-LNP (0.3 mg/mL) via the caudal vein, and for the 10 μg-dose injection, 200 μL of PAH sa/nr-mRNA-LNP (0.05 mg/mL) was administered in the same way. The control group received a 200 μL injection of PBS via the caudal vein. Blood samples were collected four days before the injection to determine the pre-treatment serum Phe levels. Serum samples were obtained on days 1, 2, 4, and 7 after the single injection, and the Phe content in the serum was detected using a Phe Assay kit (Sigma). Subsequently, the detection was carried out every 1 to 2 weeks. The body weight of the mice and any changes in their fur color were recorded. Repeated-dose treatment Mice of both genders were randomly divided into three groups, with three mice in each group, as follows: 1) Heterozygous Blank group; 2) Homozygous PBS injection group; 3) 10 μg PAH sa/nr-mRNA-LNP Repeated Injection group. Blood samples were collected five days before the treatment, 24 h after each injection, and at the time of euthanasia. Additionally, the livers were collected for further analysis. The body weight and fur color of the mice were determined before the first injection, throughout the entire experimental process, and at the time of euthanasia (on the 11th day). Histochemical analysis and immunohistochemistry Formalin-fixed, paraffin-embedded liver tissues were cut into 5-μm-thick sections using a sectioning machine. Subsequently, the tissue sections were stained with hematoxylin and eosin (H&E). Immunohistochemistry (IHC) was performed using goat anti-PAH polyclonal antibody (Affinity). ALT and AST measurement Serum samples were collected from each group of mice 7 days prior to the experimental commencement and subsequently at 7-day intervals following the injection. Each serum sample volume was maintained at a minimum of 50 μL. The levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in the collected serum samples were quantified using a fully automated multifunctional animal biochemical analyzer (Indiko Plus, Thermo Fisher, USA). Quantification and statistical analysis Statistical analysis Statistical analysis was performed using GraphPad Prism software(version 9.0). Data were expressed as means ± standard error of the mean (SEM). Comparisons between individual and corresponding control group were analyzed by unpaired, two tailed Student’s test Comparisons of multiple groups were performed using oneway or two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons tests. Degrees of significance were defined as p < 0.05 (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001). Figures were prepared using Adobe Illustrator 2023 (Adobe Systems). Additional resources No additional resources were generated in this study.