Scope and relevance. This study addresses a central problem in non-viral mRNA delivery: systemic LNP administration commonly favors the liver, while productive expression in extrahepatic organs requires coordinated control of organ distribution, cellular trafficking, endosomal escape, and tolerability. The authors developed SG5-based LNPs, optimized the final formulation as fSG5, and compared it primarily with a lung-tropic C12-200 formulation containing 50% DOTAP, referred to as C12D. Because ALC-0315 was not among the ionizable lipids listed in the study, the findings inform design questions relevant to ALC-0315 and related LNP systems but do not establish the performance, biodistribution, or safety of ALC-0315 itself. Lim et al., Advanced Science, 2026.
Research Question
The study asked whether an ionizable lipid with two separable protonation transitions could improve systemic lung-selective mRNA expression while avoiding some tolerability liabilities associated with permanently cationic lung-targeting lipids. The authors hypothesized that sequential protonation across the physiological-to-endosomal pH gradient might provide distinct membrane-interaction states and promote productive cytosolic delivery.
The question was therefore broader than whether LNPs could enter cells. It tested whether formulation chemistry could link organ-level functional expression to intracellular release of mRNA. Reporter mRNAs were used as delivery readouts: luciferase measured translated protein output, while Cre mRNA was evaluated through Cre-mediated tdTomato reporter activation. The study then extended the delivery platform to Nrf2 mRNA in an acute lung injury model. Source.
Study Design and Methods
Formulation development
The authors synthesized SG2–SG6 ionizable lipids with varying tail lengths and initially formulated them using a conventional ionizable-lipid:helper-lipid:cholesterol:PEG-lipid molar ratio of 50:10:38.5:1.5. The screening formulation used a total lipid:mRNA weight ratio of 40:1. SG5 produced the strongest luciferase signal among the candidates tested in HeLa cells and mouse lungs and was selected for further optimization.
Helper-lipid screening compared DSPC, DOPC, DOPE, and POPE. DOPE gave the highest in vitro luciferase signal with SG5. Two sequential Taguchi orthogonal-array screens were then used to optimize the lipid ratio without testing all 81 possible combinations. The composition identified as 2T7 was designated fSG5. In the reported fabrication method, the ethanol and citrate-buffer phases were rapidly mixed at a 3:1 aqueous-to-ethanol volume ratio, followed by dilution, concentration, and resuspension in DPBS. Unless otherwise stated, the SG-series formulations used an 80:1 total lipid:mRNA weight ratio, whereas C12-200 formulations used 40:1; the authors also reported a comparison at the same 40:1 ratio.
PEG-lipid was part of the conventional formulation framework, but PEG-lipid composition was not the principal variable tested in this study. The supplied text does not report a direct mRNA encapsulation-efficiency measurement, so expression results cannot be normalized to a reported fraction of encapsulated payload. Lim et al..
Expression, organ targeting, and trafficking assays
Luciferase mRNA was used to screen formulations in HeLa cells and after intravenous administration to mice. Whole-body and ex vivo organ bioluminescence measured luciferase activity, which is a readout of translated protein output rather than direct nanoparticle accumulation or total mRNA uptake. In ROSA26 reporter mice, Cre mRNA delivery was assessed indirectly: translated Cre recombinase removed a loxP-flanked STOP sequence, allowing tdTomato protein expression. Thus, the tdTomato signal represents a downstream consequence of Cre translation and recombination, not a fluorescent label attached to the delivered mRNA.
To examine intracellular trafficking, the authors used Cy5-labeled mRNA and confocal microscopy with EEA1 and LAMP1 staining. These experiments tracked the location of labeled-mRNA signal relative to early endosomes and lysosomes. They did not directly measure translation from the Cy5-labeled cargo or prove that a particular labeled-mRNA molecule had reached the cytosol.
Mechanistic and tolerability studies
The authors characterized SG5 and fSG5 using intrinsic and apparent pKa measurements, zeta-potential titration, TNS fluorescence, Laurdan membrane-packing analysis, and a pH-dependent red-blood-cell hemolysis assay. Molecular-dynamics simulations compared protonation states of SG5 and C12-200 interacting with an endosome-mimicking membrane containing BMP, DOPE, and DOPC. Separate bilayer simulations examined tail splay, membrane area, and a radius-of-gyration-based critical packing parameter.
Cellular tolerability was assessed by crystal-violet staining and CCK-8 viability assays in HeLa cells. In mice, a single intravenous dose of 1.0 mg mRNA/kg was evaluated using Evans blue lung-permeability measurements, serum ALT and AST, and histopathology. Additional fSG5 dosing at 0.30 mg mRNA/kg was assessed by serum TNF-α and IL-6 and pulmonary MPO and iNOS measurements.
Therapeutic mRNA experiment
For disease-relevant testing, mice received intranasal LPS to induce acute lung injury. Twenty-four hours later, they received a single intravenous dose of Nrf2 mRNA in either C12D or fSG5 at 0.30 mg mRNA/kg. The investigators assessed Nrf2 protein, downstream HO-1, GCLM, and SOD2, lung appearance, edema index, histology, and Ly6G-positive neutrophil-associated inflammation. These experiments moved beyond reporter output to test whether delivery of a functional mRNA cargo produced pathway engagement and tissue-level benefit.
Main Findings
Formulation and organ-level functional expression
SG5 was selected because it gave the highest luciferase activity among the SG-series candidates in the reported in vitro and lung experiments. DOPE was the preferred helper lipid in the initial screen, and the 2T7 composition was selected as fSG5 after in vivo comparison of the three best Taguchi conditions.
In vitro, fSG5 and C12D produced comparable luciferase expression. In vivo, fSG5 produced stronger lung luciferase output than C12D, including when the formulations were compared at the same total lipid:mRNA weight ratio of 40:1. In ROSA26 reporter mice, fSG5 also generated higher tdTomato output after Cre mRNA administration, with lung cryosections supporting the organ-level fluorescence result. These data support greater functional mRNA expression by fSG5 under the tested conditions. They do not by themselves demonstrate that fSG5 particles accumulated in the lung in greater physical quantity, because the principal organ readout was reporter activity after translation.
The study also linked structure to organ preference within its own lipid series. SG5, which has 16-carbon tails, produced lung-directed expression across the tested molar ratios, whereas the single-amine isoSG5 and the longer-spacer proSG5 analogues showed predominantly splenic or weak expression. The analogue experiments support a relationship between the precise amine arrangement of SG5 and the observed expression pattern, but they do not establish that tail length or pKa alone determines organ targeting.
Dual apparent pKa behavior and endosomal escape
SG5 showed two intrinsic protonation events, and fSG5 displayed two apparent pKa transitions. The TNS assay gave apparent pKa values of 7.71 and 5.15, while zeta-potential titration gave 7.64 and 4.71. The authors interpret these transitions as enabling neutral, monovalent, and divalent charge states along the delivery pathway.
Several experiments supported, but did not independently prove, improved endosomal escape. In the Laurdan assay, fSG5 and C12D differed in membrane-packing behavior, and only fSG5 showed an increase in generalized polarization at pH 5.5. In the hemolysis surrogate, fSG5 produced 83.1% hemolysis at pH 5.5 compared with 37.5% for C12D; at pH 7.4, the values were 27.7% and 34.7%, respectively. This pH-dependent membrane-disruption profile is consistent with greater acid-triggered membrane activity, although red-blood-cell hemolysis is not a direct measurement of cytosolic mRNA release in cells.
Confocal analysis of Cy5-mRNA showed similar EEA1 colocalization for C12D and fSG5, with Pearson coefficients of 0.32 and 0.36, respectively. LAMP1 colocalization was lower for fSG5 than C12D, 0.15 versus 0.51. The paper interprets the lower lysosomal overlap as evidence that more mRNA signal escaped during endosomal maturation. A careful distinction is important: these measurements describe the intracellular distribution of labeled-RNA signal, whereas the luciferase, Cre/tdTomato, and Nrf2 assays measure functional consequences of translated mRNA.
Simulation-based mechanistic interpretation
Molecular-dynamics simulations suggested that SG5 protonation states engage endosome-like membranes differently. The mono-protonated state showed greater bilayer penetration, whereas the di-protonated state remained more closely associated with phosphate groups at the membrane interface. The di-protonated SG5 condition had 98.07% phosphate-proximal contact occupancy within the reported 0.35 nm cutoff, compared with 12.17% for the mono-protonated state. By contrast, C12-200 showed a progressive increase in phosphate-proximal contact with protonation.
Interpretation: the simulations provide a plausible molecular explanation for pH-programmed membrane engagement, in which different SG5 charge states may support different interactions during endosomal maturation. They are mechanistic models rather than direct observations of escape in living cells, and the authors appropriately describe the proposed pathway as a basis for further investigation.
Tolerability and therapeutic mRNA activity
In HeLa cells, fSG5 generally preserved higher viability than C12D at the highest tested doses in crystal-violet and CCK-8 assays. In mice given 1.0 mg mRNA/kg intravenously, fSG5 produced less Evans blue leakage from lung tissue than C12D, with leakage similar to untreated controls under the reported conditions. ALT and AST did not differ from untreated controls after fSG5, whereas C12D showed higher values, particularly for AST. Histopathology did not reveal major lesions in the examined organs. The authors also reported no marked increases in serum TNF-α or IL-6 after single or repeated fSG5 administration at 0.30 mg mRNA/kg and no marked pulmonary increases in MPO or iNOS in that assessment.
In the LPS-induced acute lung injury model, both LNPs increased Nrf2-pathway activity, but fSG5 produced stronger Nrf2, HO-1, GCLM, and SOD2 signals than C12D. The edema index was reported as 0.80% for fSG5, 1.05% for C12D, 1.20% for injured mice, and 0.70% for naïve mice. Histology, gross lung appearance, and Ly6G staining were consistent with greater attenuation of inflammatory injury in the fSG5 group. These results support therapeutic activity of Nrf2 mRNA delivered by fSG5 in this acute mouse model; they are not clinical efficacy evidence.
Limitations
- Direct ALC-0315 evidence is absent. The study evaluates SG-series lipids, SG5 analogues, C12-200, DOTAP, and associated helper and PEG-lipid formulations. It does not compare ALC-0315 or establish equivalence to an ALC-0315 LNP.
- Comparator composition is not isolated to one variable. fSG5 and C12D differ in ionizable-lipid chemistry, helper-lipid context, and use of permanently cationic DOTAP. Therefore, the results show superiority of the tested formulations, not a proof that dual pKa behavior alone caused every performance difference.
- Reporter output is not the same as biodistribution. Luciferase and tdTomato primarily report translated-protein or downstream reporter activity. The study does not provide a direct quantitative map of intact particle accumulation, total tissue mRNA, or the fraction of internalized mRNA that reaches the cytosol.
- Cy5-mRNA trafficking has a limited interpretation. The Cy5 signal distinguishes labeled-RNA localization from protein expression, but colocalization with EEA1 or LAMP1 does not directly quantify cytosolic release or translation from the labeled molecules.
- Formulation variables remain incompletely resolved. PEG-lipid was not the main optimization variable, the exact final 2T7 molar composition is not provided in the supplied main text, and encapsulation efficiency is not reported in the supplied methods.
- Simulation models simplify the formulated particle. The SG-series bilayer simulations omitted PEG-lipid, and the membrane-interaction simulations used an endosome-mimicking lipid bilayer rather than a complete LNP and cellular endosomal system.
- Translational scope is limited. The experiments used HeLa cells and mouse models, with short observation windows and limited dose ranges. The authors identify clearance, metabolism, tissue residence, broader dose-response relationships, species generalizability, and more severe or chronic disease models as areas requiring further study.
- Stability was not established. The reported experiments do not provide storage-stability, mRNA-integrity, or long-term particle-stability data, so delivery performance cannot be separated from those practical formulation requirements.
Source
Lim SG, Lee M, Wang Y, et al. “Dual pKa Lipid Nanoparticles for Lung-tropic mRNA Delivery and pH-Programmed Endosomal Escape.” Advanced Science. 2026;0:e77387. https://doi.org/10.1002/advs.77387