English

• The remarkable effectiveness of mRNA vaccines during the coronavirus disease 2019 (COVID-19) pandemic not only saved countless lives but also paved a new direction for disease prevention [1–3]. Such vaccines stimulate the body's immune response by utilizing mRNA [4–7]. Compared to traditional vaccines, mRNA vaccines have several advantages, such as avoiding transport into the cell nucleus and integrate with the host genome, eliminating the risk of causing insertional mutations, and offering higher safety. They also enable a shorter research and development cycle and activate a long-lasting immune response [8–10]. Besides mRNA vaccines approved by Food and Drug Administration (FDA) [11–14], numerous mRNA vaccines are under development worldwide, targeting diseases such as respiratory diseases, influenza, cancers, rabies and monkeypox virus [15–20].

  Despite the great potential of mRNA vaccines, they still face some challenges [21]. Because of the negative charge, hydrophilicity, and instability of mRNA, it is readily degraded by nucleases after entering the human body [22–25]. Therefore, developing an efficient mRNA delivery system to ensure that mRNA enters cells and functions effectively is crucial. Lipid-based nanoparticle materials have been widely used in drug delivery due to their advantages such as safety, modifiability and targeting ability [26–29]. Lipid nanoparticle (LNP) technology is now the mainstream vehicle for mRNA delivery among them. LNPs are usually composed of four parts: ionizable lipids/cationic lipids for condensing nucleic acids, cholesterol for enhancing stability and promoting endosome membrane fusion, phospholipids for adjusting fluidity and helping membrane fusion, and PEGylated lipids for adjusting particle size and improving stability [30]. Ionizable lipids, as the key component of LNPs, contain an ionizable amine head group, hydrophobic tails, and a linker. However, even the most effective LNPs currently deliver only ~4% of mRNA into the cytoplasm, which significantly impedes the development of mRNA therapeutics [31,32]. How to further improve the delivery efficiency of mRNA is an urgent problem. Many researchers have optimized the structure of ionizable lipids by combining different multi-branched tails and amine head groups [33–37]. Chen et al. [38] introduced hydroxyl groups on the branches of multi-branched lipids, enhancing the hydrogen bond interactions between the lipids and mRNA. Zhang et al. [39] designed a multi-branched ionizable lipid that is rich in primary amines that can specifically deliver mRNA to the mouse spleen as a single component. Han et al. [40] developed a new combinatorial approach to construct a series of degradable four-branched (DB) "lipidoids", and selected the ionizable lipids showing enhanced therapeutic effects in gene editing and protein supplementation treatments. However, the structure-activity relationship among the tetra-branched hydrophobic tails and amine heads, as well as their impact on delivery efficiency, remains unclear and warrants further exploration to better guide the design of ionizable lipids and provide a deeper understanding of their mechanisms of action.

  Here, a library of ionizable lipids with tetra-branched hydrophobic tails has been constructed using the Michael addition reaction and the highly promising 10A lipid was selected for its improved efficiency of mRNA delivery. Further formulation screening resulted in the development of 10A LNPs, which demonstrated excellent performance in both in vitro and in vivo experiments. By conducting an in-depth analysis of the structure-activity relationship it was found that the efficiency of mRNA delivery is closely linked to the tertiary amine head groups with hydroxyl groups. Compared to di-branched hydrophobic tails, tetra-branched hydrophobic tails significantly enhance the stability of LNPs and improve the efficiency of endocytosis and endosomal escape, thereby effectively enhancing the delivery efficiency of mRNA (Fig. 1A). These findings help develop next-generation ionizable lipids for mRNA delivery. The developed 10A LNP demonstrates a strong potential for clinical translation for disease prevention.

  Figure 1

  

  Figure 1.  Schematic diagram of tetra-branched ionizable lipids for mRNA delivery and construction of the ionizable lipid library. (A) Tetra-branched ionizable lipids form stable LNPs, facilitate efficient cellular uptake and endosomal escape, express high levels of proteins, and ultimately achieve efficient delivery of mRNA. (B) Synthesis scheme for the lipids. Acryloyl chlorides react with fatty alcohol tail to form acrylates, which then further react with amines. (C) Different fatty alcohol tails and amine heads used in the synthesis of ionizable lipids (Fig. 1A was created with BioRender.com).

  DownLoad: Full-Size Img PowerPoint

  To address the low delivery efficiency of ionizable lipids, the Michael addition reaction was employed to synthesize 98 structurally diverse ionizable lipids (Figs. 1B and C) by combining fourteen different amine heads and seven different fatty alcohol tails. The structure and purity of the synthesized compounds were accurately confirmed and verified by nuclear magnetic resonance (NMR) and high-performance liquid chromatography with an evaporative light scattering detector (HPLC-ELSD) (Figs. S1 and S2 in Supporting information).

  To evaluate the delivery efficiency of ionizable lipids, LNPs were prepared according to the molar ratio of the commonly employed formula (ionizable lipid: 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC): cholesterol: 1,2-dimyristoyl-propyl-3-methoxy polyethylene glycol-2000 (DMG-PEG 2000) = 50:10:38.5:1.5). Their transfection ability in vitro in baby hamster Syrian kidney (BHK) cells was tested. As shown in Fig. 2A, in addition to the amine heads used for Michael addition, additional tertiary amines in the ionizable heads were beneficial for the mRNA delivery. Moreover, it was found that the hydroxyl groups in the heads significantly enhanced the mRNA delivery. The pK_a values of the most efficient LNPs, with 10A LNP at 6.90, 10B LNP at 6.62, and 9A LNP at 6.45 were measured. Previous studies have shown that the pK_a in the range of 6–7 is most conducive to LNP transfection [5], which may explain why ionizable lipids with hydroxyl groups in the head provide better transfection (Fig. S3 in Supporting information). Transfection with modified enhanced green fluorescent protein (eGFP) mRNA in BHK cells also demonstrated the superior mRNA delivery of this portion of ionizable lipids (Fig. S4 in Supporting information).

  Figure 2

  

  Figure 2.  Structural screening, and structure-activity analysis of tetra-branched ionizable lipids. (A) In vitro luciferase (Luc) expression of the ionizable lipid library (n = 3) with 1 µg Luc mRNA. The 10A lipid exhibited the best in vitro Luc expression. RLU, relative light units. (B) In vitro Luc expression of ionizable lipids with hydrophobic tails of varying lengths with 1 µg Luc mRNA (vertical bar indicates average value). (C) Particle size, surface potential, PDI, and in vitro Luc expression of mRNA LNPs formed by di-branched and tetra-branched ionizable lipids. (D) Structure of 10A lipid. ^****P < 0.0001. Data are presented as mean ± standard deviation (SD) (n = 3).

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 2B, the impact of the hydrophobic branched tails on the mRNA delivery was further analyzed. It was found that as the length of the carbon chain increased, the mRNA delivery was impacted, which is attributed to the increase in the overall hydrophobicity of the LNPs. This indicated that for optimal mRNA delivery, LNPs require a certain ratio of hydrophobic to hydrophilic segments. In the comparative study, special attention was paid to the 10A LNP, which displayed the best mRNA delivery efficiency in vitro, and the 10A₁ LNP and 10A₂ LNP formed by ionizable lipids with the same head structure but with di-branched tails. The results showed that the 10A LNP with tetra-branched tails had lower particle size, polydispersity index (PDI), and surface potential, which helped to improve its stability and safety, and achieved nearly a hundredfold increase in delivery efficiency compared to di-branched lipid LNPs (Fig. 2C). This enhancement might be due to the fact that the ionizable lipids with tetra-branched hydrophobic tails have a smaller head and a larger, conically shaped hydrophobic tail that disrupts the local membrane phospholipid organization over a larger area when it incorporates into the membrane, so causing a larger entropic stimulus for interaction. This structure is conducive to promoting the transformation of the membrane to a hexagonal crystal phase during the escape from the endosome, destroying the endosome membrane, and thereby improving the efficiency of mRNA delivery [41,42]. The structures of ionizable lipids obtained through Michael addition with hydroxyl-containing groups, hydrophobic tetra-branched tails that have the ideal transfection efficiency are summarized (Fig. 2D). The 10A lipid was used for subsequent experiments.

  The ratio of components in LNPs significantly affects mRNA delivery [43]. To obtain the best LNPs, 16 different formulations were designed through orthogonal experiments based on the four components (Fig. S6A in Supporting information). Through a comprehensive evaluation of the physical characterization and mRNA delivery efficiency of each formulation, it was found that the size of all formulations ranged from 90.6 nm to 234.2 nm, and the PDI values were between 0.084 and 0.360 (Fig. S6B in Supporting information). Most formulations showed a high level of mRNA delivery in vitro. In particular, formula 13^th showed the best in vitro mRNA expression, significantly better than the LNPs based on commercial lipid SM-102 and commercial formula. The particle size of 10A LNP was 212.9 nm, and the PDI was 0.129. As shown in Fig. S6D (Supporting information), transmission electron microscopy showed that the particles were spherical and evenly distributed and the size of the particles matched the results obtained by dynamic light scattering (DLS). Flow cytometry analysis and cell transfection images both indicated that the selected formulation had superior mRNA delivery efficiency in BHK and HeLa cells compared to SM-102 LNP (Figs. S6C and S7 in Supporting information).

  Further evaluation of the physicochemical properties and in vitro mRNA expression of 10A LNP at different quality ratios showed that an effective encapsulation of mRNA could be achieved at a mass ratio of 10:1, and the best delivery efficiency was achieved at a ratio of 40:1 (Fig. S9 in Supporting information). Additionally, the 10A LNP also demonstrated relatively better stability compared to the SM-102 LNP, exhibiting lower PDI and higher mRNA expression after being stored for 7 days (Fig. S10 in Supporting information). The 10A LNP was identified as having the best delivery efficiency through formulation screening and mass ratio selection. 10A LNP was then used for subsequent experiments.

  Next, a series of cell experiments were designed to explore the endocytosis and endosome escape ability of 10A LNP and compare it with di-branched ionizable lipids and SM-102. As shown in Figs. 3A and B, LNPs were used to encapsulate Cy5 mRNA and observed the endocytosis efficiency of Cy5 mRNA at 2 and 4 h. The results showed that the mRNA encapsulated by 10A LNP could be most effectively internalized by the cells, and a significant Cy5 fluorescence signal was observed. At the same time, the co-location with lysosomes at 4 h was relatively low, indicating that Cy5 mRNA had achieved endosome escape. As shown in Fig. 3C, the results of flow cytometry further confirmed that during the 0.5–4 h of co-cultivation, the endocytosis rate of Cy5 mRNA approached 100%, and the main fluorescence intensity (MFI) of Cy5 in the 10A LNP group was significantly enhanced (Figs. S11 and S12 in Supporting information). As shown in Fig. 3D, the 10A LNP showed great eGFP protein expressions at each time point, forming a sharp contrast with the weak mRNA expression of the 10A₁ LNP group at 4 h. These results indicate that the presence of the hydrophobic tetra-branched tails not only improved the stability of LNPs but also promoted a high level of endocytosis and endosome escape, significantly enhancing the mRNA delivery. In addition, the experiment also supported our hypothesis that the tetra-branched tails had a stronger conical shape in the spatial structure, which helped to improve the efficiency of endosome escape and mRNA expression.

  Figure 3

  

  Figure 3.  10A LNP-mediated intracellular delivery. (A) Confocal microscopy images of BHK cells co-cultured for 2 h with different LNPs encapsulating 1 µg Cy5 mRNA. (B) Confocal microscopy images of BHK cells co-cultured for 4 h with different LNPs encapsulating 1 µg Cy5 mRNA. Scale bar: 20 µm. DAPI, 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride. (C) Flow cytometry analysis of Cy5 expression and mean fluorescence intensity in BHK cells co-cultured for 0–4 h with different LNPs encapsulating 1 µg Cy5 mRNA. (D) Flow cytometry analysis of eGFP expression and mean fluorescence intensity in BHK cells co-cultured for 0–4 h with different LNPs encapsulating 1 µg eGFP mRNA. (E) The effect of different inhibitors on BHK cell uptake of LNPs encapsulating 1 µg Cy5 mRNA. Data are presented as mean ± SD (n = 3).

  DownLoad: Full-Size Img PowerPoint

  Additionally, to explore the endocytic mechanisms of LNPs, experiments using various endocytic inhibitors were conducted (Fig. 3E). At 4 ℃, the rate of endocytosis was significantly reduced, indicating that the process is energy-dependent. Concurrently, under the influence of the inhibitor wortmannin, endocytosis was also notably decreased, suggesting that 10A LNPs are primarily internalized through macropinocytosis and the associated clathrin pathway. Through a hemolysis assay, we further confirmed that 10A LNPs possess greater membrane-disrupting properties under endosomal pH conditions, a characteristic that aids in their effective escape from the endosome (Fig. S14 in Supporting information).

  Next, the in vivo transfection performance of the LNPs was evaluated. Compliance with ethical standards was ensured as all animal-based experiments adhered to the protocols established by Zhejiang University for the utilization of laboratory animals. The research was granted consent by the Zhejiang University Animal Ethics Committee, ensuring the study's alignment with ethical research practices. As shown in Figs. S15A and B (Supporting information), LNPs (containing 2 µg Luc mRNA) were injected into BALB/c mice subcutaneously and measured the protein expression 6 h later by the in vivo imaging system (IVIS). The experimental results were similar to the conclusions of in vitro transfection experiments, and the optimized 10A LNPs showed the highest in vivo transfection effect. Compared with the 10A₁ LNPs made of di-branched ionizable lipids, the transfection effect of 10A LNPs increased by nearly a hundred-fold, and it was also significantly better than the common SM-102 LNPs. As shown in Figs. S15C and D (Supporting information), LNPs (containing 2 µg Luc mRNA) were also injected into BALB/c mice subcutaneously and the protein expression was observed within 0–24 h after injection. It was found that at all time points, the transfection effect of 10A LNPs was better than that of SM-102. These experimental results proved that 10A LNPs have the same high efficiency of mRNA delivery in vivo as in vitro, so they can be further applied to tumor immunotherapy and other in vivo application experiments.

  Given the excellent performance of 10A LNP both in vitro and in vivo, its potential application in tumor immune therapy was further explored. Ovalbumin (OVA) mRNA (mOVA) can promote the proliferation and maturation of antigen-presenting cells in vivo, thereby enhancing the response of CD4⁺ and CD8⁺ T cells and improving the effect of tumor prevention [43–45].

  DC2.4 cells were treated with 10A LNP (mOVA), SM-102 (mOVA), and 10A LNP (luciferase mRNA (mLuc)). After 24 h of co-cultivation, the proportion of CD80⁺CD86⁺ cells was measured by flow cytometry (Fig. 4A). The results showed that the proportion of CD80⁺CD86⁺ cells in DC2.4 cells treated with 10A LNP (mOVA) was higher, which can effectively promote the maturation of dendritic cells (DCs). As shown in Fig. 4B, bone marrow-derived dendritic cells (BMDCs) were further explored with the same LNPs and provided the same conclusion. This indicated that 10A LNPs could efficiently deliver mOVA into DCs, effectively express the OVA protein, stimulate the maturation of antigen-presenting cells, and thereby enhance the immune response.

  Figure 4

  

  Figure 4.  In vivo immune therapeutic effect of mRNA vaccines. (A) Flow cytometry analysis of the proportion of CD80⁺CD86⁺ cells in DC2.4 cells after co-culture. (B) Flow cytometry analysis of the proportion of CD11c⁺CD80⁺ cells (left), CD11c⁺CD86⁺ cells (middle), and CD11c⁺CD80⁺CD86⁺ cells (right) in BMDCs after co-culture. (C, D) Lymph nodes were harvested for flow cytometry analysis after subcutaneous immunization. After three immunizations, the proportion of CD11c⁺CD80⁺ DCs (left), CD11c⁺CD86⁺ cells (middle), and CD11c⁺CD80⁺CD86⁺ cells (right) in the lymph nodes were determined by flow cytometry. (E, F) Spleens were harvested for flow cytometry analysis after subcutaneous immunization. After three immunizations, the proportion of CD3⁺CD4⁺ cells (left) and CD3⁺CD8⁺ cells (right) in the spleen was determined by flow cytometry. Data are presented as mean ± SD (n = 3). (G–I) In vivo immunotherapy study protocol in the B16-OVA model. As shown in schematic (G), after three immunizations, changes in tumor volume (H) and survival curves of mice (I) were determined. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. Data are presented as mean ± SD (n = 5).

  DownLoad: Full-Size Img PowerPoint

  Furthermore, a series of in vivo immune experiments to verify the effect of LNPs in enhancing immune responses was also conducted (Figs. 4C and E). In the experiment, 12 female C57BL/6 mice aged 6 weeks were randomly divided into four groups: phosphate buffer saline (PBS), 10A LNP (mLuc), 10A LNP (mOVA), and SM-102 (mOVA). On days 0, 7, and 14, the mice in each group were treated via subcutaneous injection. By day 21, the mice were euthanized and the lymph nodes and spleens were collected to assess the maturation of DCs and the activation state of T cells. Compared with the SM-102 (mOVA), 10A LNP (mOVA) showed a higher proportion of mature DC cells (CD80⁺CD86⁺) in the lymph nodes (Fig. 4D), and a significantly increased proportion of CD4⁺ T cells in the spleen (Fig. 4F). These results indicated that 10A LNP could effectively promote the maturation of antigen-presenting cells and T cells response in vivo, demonstrating the potential of nano-vaccines in immunotherapy.

  To further evaluate the in vivo prophylactic effect of 10A LNPs as a nano-vaccine, experiments were conducted using the B16-OVA melanoma model. Twenty female C57BL/6 mice aged 6 weeks were randomly divided into four groups: PBS, 10A LNP (mLuc), 10A LNP (mOVA), and SM-102 (mOVA). As shown in Fig. 4G, on days −21, −14, and −7, the mice in each group were treated via subcutaneous injection. On day 0, the mice were subcutaneously injected with B16-OVA tumor cells in the abdomen, and the tumor growth was continuously monitored. According to the ethics of animal experiments, when the tumor volume exceeded 1500 mm³, the mice were euthanized. The results showed that on the 18^th day after the injection of tumor cells, the average tumor volume of the PBS control group was greater than 1000 mm³, while the average tumor volume of the 10A LNP (mOVA) was only about 400 mm³, significantly lower than the PBS and the 10A LNP (mLuc) (Fig. 4H and Fig. S16 in Supporting information). Compared with SM-102 (mOVA), 10A LNP (mOVA) could also effectively prevent tumor growth, confirming that 10A LNP can effectively deliver OVA mRNA, activate the immune response in vivo, and exhibit significant tumor prevention effects in mice. Moreover, the 10A LNP (mOVA) had a longer survival period for the mice (Fig. 4I), and the body weight of the mice in each group continued to increase during the experiment, with no abnormalities observed (Fig. S17 in Supporting information), indicating the safety of 10A LNP. At the end of the experiment, the sectioning of the main organs of the mice showed no obvious pathological changes (Fig. S18 in Supporting information). Hematological, hepatic, and renal function analyses and liver and kidney function tests also showed no significant differences in various physiological indicators between the 10A LNPs group and the PBS control group, further confirming the biosafety of 10A LNPs (Fig. S19 in Supporting information).

  In summary, through the Michael addition reaction, a library of tetra-branched ionizable lipids was synthesized simply and conveniently. After careful analysis of the structure-activity relationship it was found that the presence of hydroxyl groups at the head helps to balance the hydrophilicity/hydrophobicity of LNPs, thereby better exploiting the delivery effect of the lipids. In addition, the length of hydrophobic carbon chains also has a significant impact on delivery efficiency. The tetra-branched tails enhanced the stability of LNPs by adjusting their hydrophilicity/hydrophobicity, effectively delivering mRNA by enhancing the endocytosis and endosome escape during the mRNA process, and ultimately significantly improving the delivery efficiency of mRNA. After optimization of the formula, the selected 10A LNP showed efficient mRNA delivery capabilities in both in vitro and in vivo experiments. In the B16-OVA melanoma model, 10A LNP could effectively prevent tumor growth while showing good biosafety. The conclusions of the structure-activity relationship analysis provide guidance for the development of efficient ionizable lipids and the selection of the 10A LNP as a safe and efficient mRNA delivery carrier with broad clinical application prospects in the future.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Haoran Xu: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Project administration, Investigation, Data curation. Jiaqi Fan: Validation, Methodology. Jiwei Liu: Validation. Qi Wei: Formal analysis. Ruoshui Li: Conceptualization. Pengcheng Yuan: Conceptualization. Bing Xiao: Conceptualization. Ying Piao: Software. Wenjing Sun: Visualization, Validation. Jiajia Xiang: Visualization, Validation. Shiqun Shao: Visualization, Validation. Zhuxian Zhou: Visualization, Validation. Youqing Shen: Writing – review & editing, Funding acquisition, Formal analysis. Nigel K.H. Slater: Writing – review & editing. Jianbin Tang: Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Data curation.

  Acknowledgments

  This work was financially supported by the National Natural Science Foundation of China (No. T2293753), the National Key R&D Program of China (No. 2021YFA1201200), and the "Pioneer" and "Leading Goose" R&D Program of Zhejiang (No. 022C03022).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.110833.

  
  
• 1. [1]

     B. Chen, M. Liu, C. Huang, J. Pharm. Anal. 11 (2021) 129–137. doi: 10.1016/j.jpha.2020.12.001

  2. [2]

     Z. Zhang, Q. Shen, H. Chang, Front. Immunol. 13 (2022) 843928. doi: 10.3389/fimmu.2022.843928

  3. [3]

     P. Huang, L. Jiang, H. Pan, et al., Adv. Mater. 35 (2023) 2207471. doi: 10.1002/adma.202207471

  4. [4]

     E.J. Sayour, D. Boczkowski, D.A. Mitchell, S.K. Nair, Nat. Rev. Clin. Oncol. 21 (2024) 489–500. doi: 10.1038/s41571-024-00902-1

  5. [5]

     E. Samaridou, J. Heyes, P. Lutwyche, Adv. Drug Deliv. Rev. 154 (2020) 37–63.

  6. [6]

     L. Shi, J. Yang, Y. Nie, Y. Huang, H. Gu, Adv. Healthc. Mater. 12 (2023) 2301261. doi: 10.1002/adhm.202301261

  7. [7]

     W. Xiao, F. Wang, Y. Gu, et al., Chin. Chem. Lett. 35 (2024) 108755. doi: 10.1016/j.cclet.2023.108755

  8. [8]

     X. Cheng, S. Liu, J. Sun, et al., Adv. Mater. 36 (2024) 2310886. doi: 10.1002/adma.202310886

  9. [9]

     N. Pardi, M.J. Hogan, F.W. Porter, D. Weissman, Nat. Rev. Drug. Discov. 17 (2018) 261–279. doi: 10.1038/nrd.2017.243

  10. [10]

      L. Miao, Y. Zhang, L. Huang, Mol. Cancer 20 (2021) 41. doi: 10.1186/s12943-021-01335-5

  11. [11]

      Y.N. Lamb, Drugs 81 (2021) 495–501. doi: 10.1007/s40265-021-01480-7

  12. [12]

      P.S. Arunachalam, M.K.D. Scott, T. Hagan, et al., Nature 596 (2021) 410–416. doi: 10.1038/s41586-021-03791-x

  13. [13]

      J.F. Gutiérrez-Bautista, M. Á. López-Nevot, E. Gómez-Vicente, et al., APMIS 130 (2022) 261–269. doi: 10.1111/apm.13215

  14. [14]

      P. Damani-Yokota, S.T. Yeung, K.M. Khanna, Cell Host Microbe. 29 (2021) 1033–1035. doi: 10.1016/j.chom.2021.06.013

  15. [15]

      W. Xie, B. Chen, J. Wong, Nat. Rev. Drug. Discov. 20 (2021) 735–736. doi: 10.1038/d41573-021-00147-y

  16. [16]

      P. Zamani, M. Mashreghi, M.R. Bazaz, et al., J. Control. Release 360 (2023) 316–334. doi: 10.1016/j.jconrel.2023.06.025

  17. [17]

      T. Carvalho, Nat. Med. 29 (2023) 755–756. doi: 10.1038/d41591-023-00017-7

  18. [18]

      J. Long, C. Yu, Y. Cao, et al., Nano Today 53 (2023) 102038. doi: 10.1016/j.nantod.2023.102038

  19. [19]

      E. Dolgin, Nat. Rev. Drug Discov. 20 (2021) 801–803. doi: 10.1038/d41573-021-00176-7

  20. [20]

      Y. Tian, M. Li, Y. Yang, et al., Chin. Chem. Lett. 35 (2024) 109270. doi: 10.1016/j.cclet.2023.109270

  21. [21]

      J. Chen, J. Chen, Q. Xu, Annu. Rev. Biomed. Eng. 24 (2022) 85–109. doi: 10.1146/annurev-bioeng-110220-031722

  22. [22]

      Y. Xiao, Z. Tang, X. Huang, et al., Chem. Soc. Rev. 51 (2022) 3828–3845. doi: 10.1039/d1cs00617g

  23. [23]

      C. Bai, C. Wang, Y. Lu, Small 19 (2023) 2303713. doi: 10.1002/smll.202303713

  24. [24]

      X. Hou, T. Zaks, R. Langer, Y. Dong, Nat. Rev. Mater. 6 (2021) 1078–1094. doi: 10.1038/s41578-021-00358-0

  25. [25]

      S. Chen, X. Huang, Y. Xue, et al., Nat. Rev. Methods Primers 3 (2023) 63. doi: 10.1038/s43586-023-00246-7

  26. [26]

      R. Liao, Z. Sun, L. Wang, et al., Bioact. Mater. 43 (2025) 406–422.

  27. [27]

      Z. Sun, R. Liao, C. Xian, J. Control. Release 375 (2024) 300–315. doi: 10.1016/j.jconrel.2024.09.013

  28. [28]

      Z. Fang, X. Zhang, H. Huang, J. Wu, Chin. Chem. Lett. 33 (2022) 1693–1704. doi: 10.1016/j.cclet.2021.11.050

  29. [29]

      J. Liu, Y. Bai, Y. Li, X. Li, K. Luo, EBioMedicine 107 (2024) 105301. doi: 10.1016/j.ebiom.2024.105301

  30. [30]

      J.A. Kulkarni, M.M. Darjuan, J.E. Mercer, et al., ACS Nano 12 (2018) 4787–4795. doi: 10.1021/acsnano.8b01516

  31. [31]

      S. Liu, Q. Cheng, T. Wei, et al., Nat. Mater. 20 (2021) 701–710. doi: 10.1038/s41563-020-00886-0

  32. [32]

      J. Gilleron, W. Querbes, A. Zeigerer, et al., Nat. Biotechnol. 31 (2013) 638–646. doi: 10.1038/nbt.2612

  33. [33]

      D. Zhang, E.N. Atochina-Vasserman, J. Lu, et al., J. Am. Chem. Soc. 144 (2022) 4746–4753. doi: 10.1021/jacs.2c00273

  34. [34]

      Z. He, Z. Le, Y. Shi, et al., Angew. Chem. Int. Ed. 62 (2023) e202310401. doi: 10.1002/anie.202310401

  35. [35]

      A.Y. Jiang, J. Witten, I.O. Raji, et al., Nat. Nanotechnol. 19 (2023) 364–375.

  36. [36]

      R. Wang, J. Li, J.J. Chen, et al., New J. Chem. 48 (2024) 4175–4183. doi: 10.1039/d3nj05949a

  37. [37]

      Y. Yan, X. Liu, L. Wang, et al., Biomaterials 301 (2023) 122279. doi: 10.1016/j.biomaterials.2023.122279

  38. [38]

      Q. Chen, X. Wang, Y. Zhang, et al., Natl. Sci. Rev. 11 (2024) nwae135. doi: 10.1093/nsr/nwae135

  39. [39]

      X. Zhang, K. Su, S. Wu, et al., Angew. Chem. Int. Ed. 63 (2024) e202405444. doi: 10.1002/anie.202405444

  40. [40]

      X. Han, J. Xu, Y. Xu, et al., Nat. Commun. 15 (2024) 1762. doi: 10.1038/s41467-024-45537-z

  41. [41]

      M.F.W. Trollmann, R.A. Böckmann, Biophys. J. 121 (2022) 3927–3939. doi: 10.1016/j.bpj.2022.08.037

  42. [42]

      S.C. Semple, A. Akinc, J. Chen, et al., Nat. Biotechnol. 28 (2010) 172–176. doi: 10.1038/nbt.1602

  43. [43]

      R. Zhang, S. Shao, Y. Piao, et al., Adv. Mater. 35 (2023) 2303614. doi: 10.1002/adma.202303614

  44. [44]

      B. Li, A.Y. Jiang, I. Raji, et al., Nat. Biomed. Eng. 9 (2025) 167–184. doi: 10.1038/s41551-023-01082-6

  45. [45]

      H. Zhang, X. You, X. Wang, et al., Proc. Natl. Acad. Sci. U. S. A. 118 (2021) e2005191118. doi: 10.1073/pnas.2005191118

• 

  Figure 1  Schematic diagram of tetra-branched ionizable lipids for mRNA delivery and construction of the ionizable lipid library. (A) Tetra-branched ionizable lipids form stable LNPs, facilitate efficient cellular uptake and endosomal escape, express high levels of proteins, and ultimately achieve efficient delivery of mRNA. (B) Synthesis scheme for the lipids. Acryloyl chlorides react with fatty alcohol tail to form acrylates, which then further react with amines. (C) Different fatty alcohol tails and amine heads used in the synthesis of ionizable lipids (Fig. 1A was created with BioRender.com).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Structural screening, and structure-activity analysis of tetra-branched ionizable lipids. (A) In vitro luciferase (Luc) expression of the ionizable lipid library (n = 3) with 1 µg Luc mRNA. The 10A lipid exhibited the best in vitro Luc expression. RLU, relative light units. (B) In vitro Luc expression of ionizable lipids with hydrophobic tails of varying lengths with 1 µg Luc mRNA (vertical bar indicates average value). (C) Particle size, surface potential, PDI, and in vitro Luc expression of mRNA LNPs formed by di-branched and tetra-branched ionizable lipids. (D) Structure of 10A lipid. ^****P < 0.0001. Data are presented as mean ± standard deviation (SD) (n = 3).

  下载: 全尺寸图片 幻灯片
  

  Figure 3  10A LNP-mediated intracellular delivery. (A) Confocal microscopy images of BHK cells co-cultured for 2 h with different LNPs encapsulating 1 µg Cy5 mRNA. (B) Confocal microscopy images of BHK cells co-cultured for 4 h with different LNPs encapsulating 1 µg Cy5 mRNA. Scale bar: 20 µm. DAPI, 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride. (C) Flow cytometry analysis of Cy5 expression and mean fluorescence intensity in BHK cells co-cultured for 0–4 h with different LNPs encapsulating 1 µg Cy5 mRNA. (D) Flow cytometry analysis of eGFP expression and mean fluorescence intensity in BHK cells co-cultured for 0–4 h with different LNPs encapsulating 1 µg eGFP mRNA. (E) The effect of different inhibitors on BHK cell uptake of LNPs encapsulating 1 µg Cy5 mRNA. Data are presented as mean ± SD (n = 3).

  下载: 全尺寸图片 幻灯片
  

  Figure 4  In vivo immune therapeutic effect of mRNA vaccines. (A) Flow cytometry analysis of the proportion of CD80⁺CD86⁺ cells in DC2.4 cells after co-culture. (B) Flow cytometry analysis of the proportion of CD11c⁺CD80⁺ cells (left), CD11c⁺CD86⁺ cells (middle), and CD11c⁺CD80⁺CD86⁺ cells (right) in BMDCs after co-culture. (C, D) Lymph nodes were harvested for flow cytometry analysis after subcutaneous immunization. After three immunizations, the proportion of CD11c⁺CD80⁺ DCs (left), CD11c⁺CD86⁺ cells (middle), and CD11c⁺CD80⁺CD86⁺ cells (right) in the lymph nodes were determined by flow cytometry. (E, F) Spleens were harvested for flow cytometry analysis after subcutaneous immunization. After three immunizations, the proportion of CD3⁺CD4⁺ cells (left) and CD3⁺CD8⁺ cells (right) in the spleen was determined by flow cytometry. Data are presented as mean ± SD (n = 3). (G–I) In vivo immunotherapy study protocol in the B16-OVA model. As shown in schematic (G), after three immunizations, changes in tumor volume (H) and survival curves of mice (I) were determined. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. Data are presented as mean ± SD (n = 5).

  下载: 全尺寸图片 幻灯片
•