LNP maintains comparative properties before and after nebulization
An inhalable delivery platform that can withstand the shear forces generated during nebulization is essential for effective lung delivery. Building on our previously developed AA3-DLin ionizable lipid, we formulated LNPs for the co-delivery of siRNA and mRNA, using poloxamer 188 as an excipient to reduce viscosity and surface tension, thus improving nebulization efficiency. The size distribution, hydrodynamic size, and polydispersity index (PDI) of mscFv/siPD-L1@LNP remained nearly unchanged after nebulization (Fig. 2a-c). mscFv/siPD-L1@LNP exhibited a slightly negative zeta potential of approximately -5.5 mV both before and after nebulization, which may help reduce electrostatic interactions with the negatively charged pulmonary mucus layer and minimize entrapment following inhalation (Supplementary Fig. 1). Both mscFv/siPD-L1@LNP and nebulized mscFv/siPD-L1@LNP exhibited similar encapsulation efficiencies of around 86% (Fig. 2d). The morphology of nanoparticles remained spherical and uniform, with consistent size before and after nebulization (Fig. 2e).
Next, we compared the intracellular behavior of LNP encapsulating mRNA and siRNA before and after nebulization in three different cancer cell lines. To investigate cellular uptake, mscFv/siPD-L1@LNP was labeled with Cy5 by substituting 18% of its cholesterol content with cholesterol-Cy5 during preparation. The results showed that mscFv/siPD-L1@LNP achieved nearly 100% cellular uptake both before and after nebulization, with similar amounts of internalized particles observed across all cell lines after four hours of incubation (Supplementary Fig. 2a, b and Fig. 2f). Endocytosis mechanism analysis revealed that LNP were internalized via multiple pathways, including caveolae-mediated endocytosis, macropinocytosis, and clathrin-mediated endocytosis, with clathrin-mediated endocytosis being the dominant route across all cancer cell types (Supplementary Fig. 3). Efficient lysosomal escape following cellular internalization is essential for the activity of mRNA or siRNA. We therefore assessed the lysosomal escape behavior of mscFv/siPD-L1@LNP using fluorescence imaging, which revealed that the LNP rapidly escaped from lysosomes (Supplementary Fig. 4). Using EGFP as a reporter gene, we evaluated the transfection efficiency and gene expression mediated by LNP and nebulized LNP. Both formulations effectively transfected the three cancer cell lines, with no statistically significant differences observed between them (Fig. 2g, h). By encapsulating each of the four siPD-L1 candidates in LNP and transfecting LLC cells, we screened for the candidate with the best knockdown performance (siPD-L1_3), which was then selected for use in both in vitro and in vivo experiments (Supplementary Fig. 5). In all tested cancer cells, mscFv/siPD-L1@LNP significantly downregulated PD-L1 expression, with nebulized mscFv/siPD-L1@LNP achieving a similar knockdown effect (Fig. 2i, j). To enhance the clinical relevance of the developed strategy, we also performed EGFP expression and PD-L1 knockdown assays in three human-derived lung cancer cell lines using both mscFv/siPD-L1@LNP and its nebulized form, which achieved transfection efficiencies exceeding 80% and significant PD-L1 silencing, with no statistically significant differences observed before and after nebulization (Supplementary Fig. 6). These results demonstrate that nebulization does not alter the physicochemical properties of LNP or its ability to mediate gene expression and silencing.
Elevated DDR1 levels in various cancers promote tumor progression by binding to collagen fibers and driving immune exclusion. We hypothesize that blocking the binding of DDR1 ECD to collagen could disrupt fiber alignment and enhance immune infiltration, making DDR1 a promising target for cancer treatment. We first investigated the relationship between DDR1 expression and prognosis in lung cancer cohorts using the Kaplan-Meier plotter platform. Our analysis revealed that DDR1 overexpression was significantly associated with poorer survival outcomes (P = 0.0055) (Supplementary Fig. 7). Additionally, DDR1 expression at both the mRNA and protein levels was notably elevated in a murine lung cancer cell line and other tumor cell lines (Supplementary Fig. 8). These findings suggested the potential of targeting DDR1 signaling as a therapeutic strategy for lung cancer. Therefore, we designed an mRNA encoding anti-DDR1 scFv, which includes a signal sequence for secretion, a heavy chain variable region (V), a light chain variable region (V), a poly-glycine-serine (G4S) linker connecting V and V, a FLAG tag for detection, and a 6×His-tag for purification (Supplementary Fig. 9a). After transfecting HEK293T cells with mscFv, we obtained anti-DDR1 scFv from the cell culture supernatant. Coomassie blue staining results and Western blot analysis confirmed the successful expression and purification of anti-DDR1 scFv (Supplementary Fig. 9b, c). The purified anti-DDR1 scFv was tested for binding affinity to DDR1 ECD using an ELISA assay, revealing an EC of 102.3 nM (Supplementary Fig. 10). mscFv was then encapsulated in LNP along with siPD-L1 and transfected into LLC, B16F10, and 4T1 cells to validate its expression (Fig. 2k). Anti-DDR1 scFv was detected in cell lysis 8 h post-transfection and persisted for over 72 h, though its levels were significantly reduced by the 72-h mark. In contrast, the level of secreted anti-DDR1 scFv in the supernatant continued to accumulate over the 72-h duration. These results confirm the effective expression of mscFv in cancer cells, with the in vitro expression kinetics offering guidance for scheduling administration in in vivo studies. Considering that normal lung cells, including epithelial cells and fibroblasts, also express DDR1, we further investigated whether mscFv affects their viability using the CCK-8 assay. BEAS-2B cells and primary mouse lung fibroblasts (MLF) were used as model cell types. Following treatment with mscFv@LNP at varying mRNA concentrations for 24 h, all groups showed comparable cell viability to the untreated control (Supplementary Fig. 11), indicating that the anti-DDR1 scFv encoded by mscFv was non-toxic to normal lung cells.
The biodistribution of LNP following a single inhalation was examined. To assess in vivo deposition and mRNA translation, mLuc@LNP was labeled with Cy5 within the nanoparticle backbone and administered to C57BL/6 mice via inhalation at a dose of 0.25 mg/kg. Six hours after inhalation, major organs were collected and analyzed for Cy5 and luciferase signals, revealing that both LNP accumulation and luciferase expression were restricted to the lungs (Fig. 2l, m and Supplementary Fig. 12). To identify the cell subtypes internalizing LNP, mice bearing orthotopic LLC lung tumors were treated with mscFv/siPD-L1@LNP. Six hours later, the tumor-bearing lungs were digested and analyzed via flow cytometry. Key lung cell types, including cancer cells, epithelial cells, endothelial cells, and immune cells, were evaluated, showing that 13.8, 51.7, 1.8, and 8.0% of these respective cells internalized LNP (Fig. 2n). Cancer cells and immune cells were the primary cell types taking up LNP, internalizing 31.6 and 41.3% of the total mscFv/siPD-L1@LNP, respectively (Fig. 2o). Functional mRNA expression in vivo was then assessed. Mice were administered mscFv/siPD-L1@LNP by inhalation, and anti-DDR1 scFv levels were measured at various time points. As shown in Fig. 2p, q, anti-DDR1 scFv was detectable in both lung homogenates and bronchoalveolar lavage fluid (BALF), with sustained in vivo expression for at least 120 h. Anti-DDR1 scFv concentrations in lung homogenates peaked at 48 h post-inhalation, while in BALF, peak levels were reached at 72 h, likely due to delayed secretion. These results suggested that a dosing frequency of every 2 days in in vivo studies could sustain high levels of anti-DDR1 scFv in the lung tumor microenvironment.
To evaluate the in vitro antitumor efficacy of mscFv/siPD-L1@LNP, a co-culture system of tumor cells and T cells was established. LLC, B16F10, and 4T1 cells were pretreated with various formulations and then co-cultured with activated CD8 T cells for 24 h. Tumor cell apoptosis was measured by flow cytometry after staining with Annexin V/propidium iodide (PI) (Fig. 3a-d). CD8 T cell-mediated tumor killing was modest in the control and LNP groups, which could be attributed to the immune escape mechanisms facilitated by PD-1/PD-L1 signaling. Treatment with mscFv alone induced certain cancer cell apoptosis, possibly because the DDR1/collagen binding also activates antiapoptotic signaling pathways in cancer cells. After simultaneously downregulating PD-L1 and blocking DDR1 signaling, the mscFv/siPD-L1@LNP group exhibited the highest levels of apoptotic tumor cells. We further analyzed the cytokine-producing capacity of CD8 T cells in the co-culture system (Fig. 3e-g). Treatment with mscFv/siPD-L1@LNP significantly increased the release of granzyme B, IFN-γ, and TNF-α by CD8 T cells, demonstrating restored T cell function and enhanced cytotoxicity against cancer cells. The levels of cytokines released into the cell culture supernatant were measured by Western blot, revealing increased secretion of pro-inflammatory cytokines, including TNF-α, IL-1β, and IL-6, along with a reduced production of the immunosuppressive cytokine IL-10 following treatment with mscFv/siPD-L1@LNP (Fig. 3h). PD-L1 and DDR1 are known to promote tumor cell migration, so we also investigated the effects of different treatments on this process. Simultaneous blockade of PD-L1 and DDR1 signaling led to the greatest inhibition of cancer cell migration (Supplementary Fig. 13).
We then evaluated the in vivo function of the anti-DDR1 scFv designed in this study following inhalation of mscFv@LNP at low or high doses. An orthotopic lung cancer mouse model was first established by surgically implanting 2 × 10 LLC-Luc cells into the left lung lobe of C57BL/6 mice. The procedure did not cause any complications, as indicated by unchanged pulmonary function before and after surgery (Supplementary Fig. 14). mscFv@LNP was administered every 2 days for a total of six inhalations (Supplementary Fig. 15a). Second harmonic generation (SHG) imaging revealed that, without treatment, the collagen fibers within the tumor were highly aligned (Fig. 4a), forming a barrier that impeded immune cell infiltration. Treatment with mscFv@LNP at doses of 0.5 or 1 mg/kg disrupted fiber alignment to varying degrees, with the 1 mg/kg dose resulting in the fibers becoming loose and disordered. The degree of fiber alignment, quantified by the coefficient of alignment, was significantly reduced to 75.1 and 48.4% of the PBS-treated controls following treatment with mscFv@LNP at doses of 0.5 and 1 mg/kg, respectively (Fig. 4b). In addition, the length of collagen fibers was markedly shortened in the 1 mg/kg group (Fig. 4c). The ability of mscFv@LNP to alter collagen fiber configuration was also verified in a pulmonary metastasis model of breast cancer, established by intravenously injecting 4T1-Luc cells. Similarly, inhalation of 1 mg/kg mscFv@LNP caused the collagen fibers to shift from dense and aligned to loose and disordered, with a significant decrease in both the coefficient of alignment and fiber length compared to the PBS group (Fig. 4d-f). Aligned collagen fibers have been shown to increase tissue matrix stiffness. Therefore, we assessed tumor stiffness using atomic force microscopy. The stiffness of tumors decreased in a dose-dependent manner following mscFv inhalation in both the orthotopic LLC tumor (Fig. 4g, h) and the breast cancer lung metastasis model (Fig. 4i, j), demonstrating that mscFv@LNP significantly altered the mechanical properties of the tumor tissue.
Based on these findings, we next investigated the impact of mscFv@LNP on immune infiltration. A dose gradient ranging from 0.05 to 1.25 mg/kg was administered to mice bearing LLC lung tumors, resulting in a dose-dependent control of tumor growth (Supplementary Fig. 15b). The concentration of anti-DDR1 scFv in the tumor tissue was measured by ELISA, which showed a direct correlation with the inhalation dose of mscFv@LNP (Supplementary Fig. 16). The number of infiltrated immune cells, including total immune cells, CD4 T cells, CD8 T cells, NK cells, and NKT cells within TME was analyzed by flow cytometry (Fig. 4k-o). An increase in the presence of all tested immune cell types was observed with escalating doses of mscFv, indicating a strong correlation between immune cell infiltration and the dose of mscFv. Collectively, these data demonstrate that inhalation of mscFv effectively disrupts aligned collagen fibers within lung tumor tissues and reduces tumor stiffness, leading to the reversal of immune exclusion and enhanced infiltration of various immune effector cells.
We assessed the antitumor effects of the inhaled therapy in a mouse model of orthotopic lung cancer. The mice were administered PBS, empty LNP, mscFv/siNC@LNP, mEGFP/siPD-L1@LNP, or mscFv/siPD-L1@LNP via inhalation on days 6, 8, 10, 12, 14, and 16 (Fig. 5a). Tumor progression during treatment was monitored using IVIS Spectrum imaging, and tumor burden was quantified by measuring the intensity of the luciferase signal (Fig. 5b, c). Tumors grew rapidly in the PBS and LNP groups without drug intervention. Treatment with mscFv/siNC@LNP and mEGFP/siPD-L1@LNP significantly delayed tumor progression, while the inhalation of mscFv/siPD-L1@LNP achieved the most favorable therapeutic outcomes among the five groups. One day after the final treatment, tumor-bearing lungs from each group were collected and examined by H&E staining (Fig. 5d). Compared to the PBS and LNP groups, treatment with either mscFv alone, siPD-L1 alone, or their combination effectively controlled tumor growth, with the mscFv/siPD-L1@LNP treatment showing the most significant tumor regression, as evidenced by the reduced tumor areas. We recorded the body weight of mice from different groups until day 27 post-tumor inoculation. A significant decrease in body weight was observed only in the PBS and empty vehicle groups (Fig. 5e), revealing aggressive tumor progression. In parallel, lung weight measurements revealed a significant decrease in the lungs of the mscFv/siPD-L1@LNP group compared to the other four groups, indicating minimal tumor growth in this group (Fig. 5f). The survival of mice from different groups was monitored until day 60 post-tumor inoculation (Fig. 5g). All mice in the PBS and LNP groups died by day 31, with median survival times of 23 and 25.5 days, respectively. Although treatment with siPD-L1 alone inhibited tumor growth during the treatment cycle, these effects were not durable, as indicated by the moderate impact on extending mouse survival. In the mscFv/siNC@LNP group, median survival was significantly prolonged to 34 days. Treatment with mscFv/siPD-L1@LNP further extended the median survival time to 54 days, with 50% of the mice still alive at the end of the monitoring period. These data suggest that inhalation of mscFv/siPD-L1@LNP can effectively hinder the advancement of orthotopic lung cancer, demonstrating significantly better therapeutic effects compared to mscFv or siPD-L1 therapy alone. The therapeutic efficacy of inhaled mscFv/siPD-L1@LNP was also evaluated in a male mouse model of LLC-induced orthotopic lung cancer, using the same tumor implantation method and administration regimen (Supplementary Fig. 17). Consistently, treatment with mscFv/siPD-L1@LNP significantly inhibited tumor growth and prolonged overall survival, indicating that the therapeutic effectiveness of this strategy is not sex-dependent. Additionally, we performed TUNEL staining on lung tissue sections from mice treated with mscFv/siPD-L1@LNP to assess whether the treatment induced unintended apoptosis in DDR1-expressing normal tissues (Supplementary Fig. 18). While DDR1 expression was detected in both tumor and lung regions, apoptotic signals were predominantly localized to tumor areas, with minimal apoptosis observed in normal lung tissues, supporting that this therapy safely inhibited tumor progression without causing toxicity to normal tissues.
We first examined gene expression and silencing in tumor-bearing mice as a basis for assessing the potential of mscFv/siPD-L1@LNP to reshape TME. The expression of anti-DDR1 scFv was confirmed by immunofluorescent staining, revealing its production after inhalation of mscFv/siNC@LNP or mscFv/siPD-L1@LNP, with no expression observed in the other three groups (Supplementary Fig. 19a). Collagen fiber rearrangement mediated by anti-DDR1 scFv was analyzed using picrosirius red staining. Treatment with mscFv resulted in disordered and shorter collagen fibers compared to the other groups (Fig. 5h, i). Immunofluorescent staining and Western blot analysis of PD-L1 demonstrated PD-L1 downregulation in the treatment groups containing siPD-L1 (Fig. 5j and Supplementary Fig. 20).
Next, we investigated alterations in the TME following inhalation treatment with mscFv/siPD-L1@LNP in an orthotopic LLC model (Fig. 6a). The different immune cell populations within the tumors were characterized and analyzed using flow cytometry. The proportion of CD8 T cells increased following treatment with mscFv/siNC@LNP or mLuc/siPD-L1@LNP compared to the PBS or placebo groups. The synergistic effects of mscFv and siPD-L1 were evident in the significantly higher frequency of CD8 T cells in the mscFv/siPD-L1@LNP group compared to the other groups (Fig. 6b). The blockade of PD-1/PD-L1 signaling is known to alleviate immunosuppression in the TME. Therefore, we examined the levels of immunosuppressive cells in tumors after treatment. The frequency of the primary immunosuppressive cell types, including regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs), was significantly reduced in the treatment regimen that included siPD-L1 (Fig. 6c, d), indicating that downregulation of PD-L1 expression via RNAi effectively reversed immunosuppression. The capacity of CD8 T cells to secrete IFN-γ was evaluated (Fig. 6e). In the mscFv/siNC@LNP and mLuc/siPD-L1@LNP groups, the proportion of CD8IFN-γ T cells increased by 2.3-fold and 2.5-fold, respectively, compared to the PBS group. Notably, a 4-fold increase in the frequency of CD8IFN-γ T cells was observed in the mscFv/siPD-L1@LNP group compared to the PBS group. Meanwhile, the proliferative capacity of CD8 T cells was significantly enhanced, as indicated by the markedly higher proportion of CD8Ki67 T cells in the mscFv/siPD-L1@LNP group compared to the other groups (Fig. 6f). To verify the enhanced immune infiltration in lung tumors achieved by the developed inhaled therapy, we measured the number of total immune cells (CD45 cells), CD4 T cells, and CD8 T cells within the tumors by flow cytometry and immunofluorescent staining (Fig. 6g-i and Supplementary Fig. 9b, c). In the PBS and LNP groups, immune infiltration levels were low but showed a slight increase following treatment with mLuc/siPD-L1@LNP. However, the inclusion of mscFv in the inhaled formulation significantly boosted the number of immune cells in the TME. Specifically, treatment with mscFv/siNC@LNP increased the number of CD45 cells, CD4 T cells, and CD8 T cells by fourfold, 4.2-fold, and 4.9-fold, respectively. The mscFv/siPD-L1@LNP treatment further enhanced these increases by fourfold, 4.3-fold, and 7.9-fold, respectively, compared to the PBS group. Additionally, cytokine levels in the TME were quantified by ELISA. The concentrations of immunostimulatory cytokines, including IFN-γ, IL-2, IL-12p70, and TNF-α, were significantly elevated following mscFv/siPD-L1@LNP treatment compared to the other four groups (Fig. 6j-m). Conversely, mscFv/siPD-L1@LNP treatment substantially reduced the level of IL-10, a potent immunosuppressive cytokine (Fig. 6n and Supplementary Fig. 19d). While enhanced immune infiltration and pro-inflammatory cytokine secretion were noted in the pulmonary TME, serum levels of pro-inflammatory cytokines remained similar across treatment groups (Supplementary Fig. 21). This indicates that pulmonary delivery of mRNA/siRNA-LNPs resulted in minimal systemic cytokine production, thus mitigating potential safety concerns. Collectively, the inhalation of mscFv/siPD-L1@LNP effectively reprogrammed the hostile lung TME by simultaneously boosting immune infiltration and reversing immunosuppression, thereby creating a more favorable environment for a stronger immune response against cancer cells.
To determine if the tumor suppression observed with the developed inhalation strategy is limited to orthotopic lung cancers, we established a mouse model of pulmonary metastasis of breast cancer by intravenous injection of 1 × 10 4T1-Luc cells and administered different formulations via inhalation, with a total of six treatments (Fig. 7a). The IVIS images and tumor signal intensity curve revealed substantial tumor growth in the PBS and LNP groups (Fig. 7b, c), highlighting the aggressiveness of 4T1 breast cancer lung metastasis. The tumor inhibition by mEGFP/siPD-L1@LNP treatment was modest, showing some effect during the early phase of tumor development, likely due to the resistance of 4T1 breast cancer to ICB therapy. Inhalation of mscFv/siNC@LNP suppressed tumor growth, and this therapeutic effect was further enhanced in mice treated with mscFv/siPD-L1@LNP, which resulted in the most pronounced tumor inhibition among the treatment groups. To assess tumor burden, tumor-bearing lungs were excised on day 18 post-tumor inoculation for morphological observation and H&E staining, with the number of metastatic foci also quantified (Fig. 7d-f and Supplementary Fig. 22). Consistent with the IVIS results, lungs from the PBS and LNP vehicle groups displayed dense metastatic foci, with a slight reduction observed in the mEGFP/siPD-L1@LNP group. While mscFv alone significantly reduced tumor burden, the combination therapy resulted in the fewest metastatic foci in the lung tissues. Since tumor growth in the lung increased lung weight, the lung weights followed a similar pattern to the number of metastatic foci across the five groups (Fig. 7g). The rapid tumor growth in the PBS, LNP, and mEGFP/siPD-L1@LNP groups resulted in a significant decrease in body weight, whereas the average normalized body weight of mice in the treatment groups containing mscFv remained above 100% at 20 days post-tumor implantation (Supplementary Fig. 23). The rapid tumor progression culminated in the death of all animals within 31 days in the PBS, LNP, and mEGFP/siPD-L1@LNP groups (Fig. 7h). Treatment with mscFv/siNC@LNP extended median survival to 35 days, while mscFv/siPD-L1@LNP recipients showed significantly longer survival compared to other groups, owing to its synergistic therapeutic effects.
We further investigated changes in the TME after treatment. The presence of immunosuppressive cells, including Tregs and MDSCs, was analyzed by flow cytometry, revealing that PD-L1 downregulation via the siRNA strategy effectively alleviated immunosuppression (Fig. 7i, j). The number of total immune cells, CD4 and CD8 T cells, as well as NK and NKT cells that directly kill cancer cells, was assessed (Fig. 7k-o). By disrupting collagen fiber alignment, treatment with mscFv/siNC@LNP or mscFv/siPD-L1@LNP significantly reduced immune exclusion and promoted the infiltration of these immune cells. Overall, mLuc/siPD-L1@LNP achieved similar effects in reversing immunosuppression as mscFv/siPD-L1@LNP, while mscFv/siNC@LNP had comparable effects in relieving immune exclusion. However, the therapeutic efficacy of mscFv/siPD-L1@LNP was significantly superior to both, highlighting that addressing both insufficient immune infiltration and immunosuppression is crucial for effective lung cancer treatment.
To assess in vivo toxicity following inhalation of mscFv/siPD-L1@LNP, we administered six doses to healthy mice. Major organs and blood samples were collected 24 h after the final inhalation for evaluation (Supplementary Fig. 24a). H&E staining of major organs, including the heart, liver, spleen, and kidneys, showed no lesions, histopathological injuries, or abnormalities caused by the inhalation of mscFv/siPD-L1@LNP (Supplementary Fig. 24b). Particular attention was given to lung structure, where LNPs primarily accumulate (Supplementary Fig. 24c). The alveoli, large airways, and small airways remained intact and clear, with no signs of inflammation or damage in any of the groups. The potential side effects of mscFv/siPD-L1@LNP on liver function, heart function, kidney function, and glycometabolism were assessed through serum biochemical analysis, which showed no statistically significant differences among the treatment groups for any of the tested parameters (Supplementary Fig. 25a). Hematological analysis also confirmed no impact on blood components or function (Supplementary Fig. 25b). These findings suggest a favorable biosafety profile for mscFv/siPD-L1@LNP, supporting its potential for further preclinical or clinical investigation.