Abstract

   Pulmonary fibrosis (PF) is a life-threatening interstitial lung
   disease, characterized by excessive fibroblast activation and collagen
   deposition, leading to progressive pulmonary function decline and
   limited therapeutic efficacy. Here, the inhalable,
   myofibroblast-targeted, and pH-responsive liposomes (FL-NI) were
   developed for effective codelivery of nintedanib, a mainstream
   antifibrotic drug in clinic, and siIL11, a small interfering RNA that
   silences the key profibrosis cytokine IL-11. Notably, FL-NI achieved a
   117.8% increase in pulmonary drug delivery by noninvasive inhalation
   and a 71.5% increase in delivery specifically to fibroblast activation
   protein–positive myofibroblasts while reducing nonspecific immune cell
   and epithelial uptake by 29.8 and 55.8%, respectively. The accurate
   inhalation codelivery of nintedanib and siIL11 into myofibroblasts
   achieved synergistic effects, effectively enhanced myofibroblast
   deactivation, reduced pathological collagen deposition by 50.8%, and
   promoted epithelial tissue repair. FL-NI remodeled the aberrant immune
   microenvironment without inducing systemic toxicities. Therefore, this
   work demonstrated the notable potential for this pluripotent strategy
   for improving PF outcomes and its promising clinical translation.
     __________________________________________________________________

   Inhaled and pH-responsive nanoparticles coloaded siIL11 and nintedanib
   with myofibroblast targeting ability for PF treatment.

INTRODUCTION

   Pulmonary fibrosis (PF) is a chronic interstitial lung disease
   characterized by respiratory challenges and progressive deterioration
   of lung function, often culminating in an extremely poor prognosis
   ([56]1). The median survival rate for untreated patients’ postdiagnosis
   ranges between 3 and 5 years ([57]2). Although the precise etiology of
   PF remains unclear, a range of occupational and external factors,
   including tobacco smoking ([58]3), severe infections ([59]4), genetic
   predisposition ([60]5), and environmental exposures ([61]6), has been
   implicated in its pathogenesis. In addition, PF has emerged as a
   substantial complication of the COVID-19 pandemic, as evidenced by
   multiple clinical reports ([62]7). Nearly all patients with severe
   COVID-19 exhibit some degree of PF, with a notable incidence rate
   ([63]8). In light of the morbidity and mortality, PF has become a
   respiratory disease that poses a serious threat to human health.

   Now, various inhibitors including nintedanib (NIN), a tyrosine kinase
   inhibitor ([64]9), and pirfenitone, an inhibitor of various cytokines
   such as tumor necrosis factor–α (TNF-α), transforming growth factor–β
   (TGF-β), and platelet-derived growth factor-β (PDGF-β) have been widely
   used for PF treatment in the clinic ([65]10). Although these have shown
   some efficiency in alleviating the physiological progression of PF and
   recovering the pulmonary function, their oral administration often
   results in gastrointestinal and liver adverse reactions ([66]11,
   [67]12). Moreover, the lack of organ and cell specificity of these
   drugs significantly limits their therapeutic effectiveness
   ([68]12–[69]14). Immunomodulatory treatments, such as injections of
   N-acetyl-L-cysteine and azathioprine, along with corticosteroids,
   exhibit unsatisfactory efficiency for PF treatment in the clinic
   ([70]13, [71]15). Therefore, developing effective therapeutic
   strategies for PF is of utmost importance.

   In the pathological process of PF, damage to alveolar epithelial type
   II cells initiates pathological epithelial remodeling ([72]16), leading
   to the recruitment and activation of fibroblasts and abnormal
   deposition of extracellular matrix (ECM), ultimately resulting in the
   formation of fibrotic scars and the destruction of lung tissue
   structure ([73]17). Given the pivotal role of fibroblasts in the
   progression of PF, inhibiting fibroblast activation and promoting
   epithelial cell repair are considered as promising therapeutic
   strategies. Interleukin-11 (IL-11) has been identified as a factor in
   this process, as it is specifically secreted by fibroblasts ([74]18).
   In patients with PF, IL-11 levels are elevated in the lungs and
   fibroblasts, and IL-11 will bind to its heterodimeric receptor subunit
   IL-11 receptor subunit alpha (IL-11RA) and glycoprotein 130 (gp130)
   ([75]19), leading to fibroblast activation and increased ECM production
   by regulating the extracellular signal–regulated kinase (ERK) and AKT
   signaling pathways ([76]20). In addition, IL-11 can enhance the
   formation of a profibrotic immune microenvironment through paracrine
   signaling ([77]21). Therefore, silencing IL-11 expression represents a
   potential therapeutic strategy for PF.

   Small interfering RNA (siRNA) therapies have emerged as a safe and
   powerful therapeutic approach in various clinical fields, and
   exogenously synthesized IL-11 siRNA (siIL11) has the ability to silence
   genes expression in a sequence-specific manner ([78]22, [79]23).
   However, because of its susceptibility to degradation, the efficacy of
   siRNA-mediated therapy highly depends on optimized delivery vectors and
   appropriate administration routes ([80]24). Liposomes are considered as
   a mature nonviral nanocarrier for gene therapy, demonstrating
   significant advancements in clinical application, such as patisiran, a
   lipid-based siRNA delivery vector successfully developed and approved
   by Food and Drug Administration ([81]25). To enhance the efficiency of
   drug delivery to the lungs, noninvasive inhalation of nanoparticles
   allows for preferential accumulation and prolonged residence in the
   pulmonary environment, making it an ideal strategy for the treatment of
   PF ([82]26, [83]27). However, the lung microenvironment consists of
   various cell types, and nanocarriers are often nonspecifically engulfed
   by epithelial cells, endothelial cells, immune cells ([84]27, [85]28),
   which seriously hinders the antifibrotic effects of inhaled drugs.
   Therefore, targeting activated fibroblasts for drug delivery could
   notably improve both the efficiency of drug transport and the
   therapeutic efficacy against PF.

   Recent research has demonstrated that fibroblast surface markers can
   serve as effective targets for the specific recognition and uptake of
   nanoparticles by fibroblasts ([86]29, [87]30), offering a promising
   approach for precise cell-targeted therapy. Fibroblast activation
   protein (FAP) is a transmembrane protein specifically expressed in
   activated fibroblasts and serves as an important biomarker for the
   diagnosis and treatment of PF ([88]31). Inspired by this, we developed
   a novel class of liposomes containing FAP-targeted peptides and siIL11
   and NIN for noninvasive inhalation therapy. The FAP-targeting peptide
   enables specific binding to the FAP on activated fibroblasts, thus
   directing the liposomes with selective targeting ability. Moreover, the
   polymer poly(2-ethyl-2-oxazoline) (PEOz) was incorporated to facilitate
   the degradation of liposomes in the acidic conditions of the endosome
   ([89]32), enhancing their ability to escape endosomal compartments.
   Once delivered into activated fibroblasts, NIN effectively inhibits
   signaling pathways mediated by platelet-derived growth factor receptor
   (PDGFR)/fibroblast growth factor receptor (FGFR), while siIL11 reduces
   the expression of IL-11. This down-regulation attenuates the
   IL-11RA/gp130 signaling pathway and decreases the activation of ERK and
   AKT ([90]Fig. 1). Subsequent studies revealed that intratracheal
   delivery of NIN and siIL11 coloaded liposomes significantly promoted
   myofibroblast deactivation, reduced ECM deposition 50.8%, and improved
   the Ashcroft score from 7.6 ± 0.55 to 2.8 ± 0.45 in a mouse model of
   PF. Furthermore, the reduction in IL-11 secretion diminishes the
   profibrotic immune response mediated by its paracrine effects. Thus,
   the noninvasive inhalable FL-NI presents a promising candidate for
   clinical translation in the treatment of PF, owing to its remarkable
   ability to inhibit fibrosis, alleviate chronic inflammation, and
   remodel the immune microenvironment.

Fig. 1. Inhalation of activated fibroblast-targeted liposomes (FL-NI) for
precise codelivery of NIN and siIL11 in PF treatment.

   [91]Fig. 1.
   [92]Open in a new tab

   (A) Components and construction of the FL-NI. (B) Proposed mechanism of
   FL-NI for PF treatment: 1) Noninvasive inhalation enhances local drug
   deposition in the lungs; 2) FAP-targeting peptide modification on
   liposomes improves targeted delivery efficiency to myofibroblasts; 3)
   pH-responsive PEOz promotes drug release and facilitates endosomal
   escape; 4) FL-NI codelivers NIN and siIL11 to myofibroblasts, achieving
   synergistic effects that inhibit fibroblast activation and promote
   collagen degradation; 5) FL-NI remodels the immune microenvironment of
   PF. AEC I, type I alveolar epithelial cells; AEC II, type II alveolar
   epithelial cells; RISC, RNA-induced silencing complex. Created in
   BioRender. Chen, Q. (2025) [93]https://BioRender.com/qbup1xu.

RESULTS

Synthesis and characterization of FL-NI

   To synthesize efficient nanoparticles with selective delivery and
   endosomal escape, blank cationic liposomes were first prepared using
   the thin-film dispersion technique ([94]33). Given the specific
   expression of FAP in activated fibroblasts and the pH-sensitive PEOz,
   we synthesized the FAP-targeting peptide (DRGETGPAC) (fig. S1) and
   1,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE)–PEOz–maleimide
   (MAL) to create DSPE-PEOz-FAP for liposome modification. The
   conjugation of FAP-targeting peptide to DSPE-PEOz-FAP was achieved
   through thiol-MAL chemistry and confirmed by the ^1H nuclear magnetic
   resonance spectrum (fig. S2) and high-performance liquid chromatography
   (fig. S3). Next, liposomes were modified with different molar ratios of
   FAP-targeting peptide (peptides/liposomes) to determine the optimal
   ratio for efficient siRNA delivery. Specifically, activated mouse
   embryonic fibroblasts (MEFs) were coincubated with cationic liposomes
   with different peptide ratios and fluorescein amidite (FAM)–labeled
   siRNA (FAM-siRNA), followed by flow cytometry. It was observed that
   liposomes with peptide ratio above 20% exhibited efficient delivery to
   MEFs ([95]Fig. 2, A and B). Therefore, 20% FAP-targeting peptides were
   selected for the synthesis of FL-NI.

Fig. 2. Characterization of FL-NI.

   [96]Fig. 2.
   [97]Open in a new tab

   (A and B) Fluorescence intensity of MEFs incubated with preparations of
   different peptides/liposomes ratios. (C) Average particle sizes and (D)
   zeta potential of different liposomes determined by DLS. (E)
   Ultraviolet-visible spectra of NIN, siIL11, blank liposome, L-NI, and
   FL-NI. (F) Drug encapsulation efficiency of NIN and siIL11. (G)
   Particle size distribution of FL-NI. (H) Representative transmission
   electron microscopy (TEM) micrograph of FL-NI. Scale bar, 200 nm. (I)
   Particle size of FL-NI in Dulbecco’s modified Eagle’s medium (DMEM)
   with 10% fetal bovine serum (FBS). (J) Particle size of FL-NI at
   different pH values. (K) Stability of naked siRNA and siRNA loaded in
   FL-NI against RNase. All data are presented as means ± SD (n = 3).
   Statistical significance was calculated via one-way analysis of
   variance (ANOVA) (B) in GraphPad Prism. ****P < 0.0001; not significant
   (ns), P > 0.05.

   The cationic liposomes containing NIN and siIL11 (FL-NI) were
   subsequently prepared by mixing cationic liposomes [composed of
   cholesterol, DOTAP (1,2-dioleoyl-3-trimethylammonium-propane), and NIN]
   with a hydrophilic siRNA core consisting of hyaluronic acid, siIL11,
   and protamine. The mixture was then fused with the DSPE-PEOz-FAP at a
   high temperature of 50°C using the postinsertion method. Next, we
   successfully synthesized L-N (loaded with NIN), L-I (loaded with
   siIL11), L-NI (loaded with both NIN and siIL11), and FL-NI (loaded with
   both NIN and siIL11 and surface-modified with FAP-targeting peptide).
   Dynamic light scattering (DLS) measurements indicated that the particle
   sizes for L-N, L-I, L-NI, and FL-NI were 181.3 ± 2.1 nm, 175.1 ± 2.0
   nm, 176.6 ± 2.3 nm, and 182.0 ± 2.2 nm, respectively ([98]Fig. 2C). The
   zeta potentials for these formulations were 39.4 ± 0.3, 38.7 ± 2.8,
   37.8 ± 1.5, and 29.2 ± 2.1 mV, respectively ([99]Fig. 2D).
   Ultraviolet-visible absorption spectra of the dual drugs (NIN and
   siIL11), blank liposomes, and drug-loaded liposomes (L-NI and FL-NI)
   displayed characteristic absorption peaks at 388 nm (NIN) and 260 nm
   (siIL11), confirming the encapsulation of both drugs within the
   liposomes ([100]Fig. 2E). The encapsulation efficiencies of NIN and
   siIL11 in FL-NI were determined to be 92.1 ± 0.2% and 92.9 ± 1.0%,
   respectively ([101]Fig. 2F and fig. S4).

   The polydiseperse index ([102]Fig. 2G) and transmission electron
   microscopy ([103]Fig. 2H) of FL-NI demonstrated a concentrated
   distribution and uniform spherical shape, both of which are favorable
   for stability ([104]34). The stability of FL-NI in Dulbecco’s modified
   Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS, Cellmax,
   SA112.02, Beijing) was assessed over 7 days. Daily measurements
   confirmed excellent particle size stability ([105]Fig. 2I). In
   addition, DLS results revealed nearly no changes in particle size of
   FL-NI after incubation in DMEM (C3103-0500, VivaCell, Shanghai, China)
   containing 10% FBS at both pH 7.4 and 6.8 for 24 hours, indicating that
   FL-NI maintains its structural integrity under mildly acidic
   conditions, typical of extracellular pH in PF lung tissue ([106]Fig.
   2J). Meanwhile, FL-NI was incubated in bronchoalveolar lavage fluid
   collected from mice with PF for 7 days, showing minimal changes in
   particle size and demonstrating excellent stability within the PF
   microenvironment (fig. S5). To assess the protective effect of FL-NI on
   siRNA, naked siRNA and siRNA encapsulated in FL-NI were incubated with
   ribonuclease (RNase) for various durations (0, 15, 30, 60, 120, and
   240 min). While naked siRNA rapidly degraded upon exposure to RNase for
   up to 4 hours, FL-NI–encapsulated siRNA remained intact, demonstrating
   notable protection ([107]Fig. 2K).

Targeted delivery of FL-NI

   Given that FAP serves as a specific recognition target for activated
   fibroblasts ([108]30), we further investigated the enhanced delivery
   capacity of liposomes mediated by FAP-targeting peptide to activated
   MEFs. We synthesized liposomes encapsulating FAM-siRNA, creating two
   variants: FL-siRNA, which contains FAP-targeting peptide, and L-siRNA,
   which does not. Normal and activated MEFs were respectively treated
   with L-siRNA, FL-siRNA, or FL-siRNA plus treatment of anti-FAP
   antibodies with equivalent concentration of siRNA. After 2 hours of
   coincubation, flow cytometry was used to quantify the percentage of
   FAM-positive MEFs ([109]Fig. 3A). The results demonstrated that the
   FL-siRNA group exhibited significantly higher delivery efficiency
   compared to L-siRNA, as evidenced by an increased proportion of
   FAM-positive MEFs. Blocking FAP with antibodies subsequently reduced
   the siRNA delivery efficiency of FL-siRNA. Conversely, when the same
   procedure was applied to normal MEFs, the FL-siRNA group showed only a
   slight increase in FAM-positive MEFs, likely due to minor activation of
   normal MEFs from unavoidable mechanical stimulation during the
   experiment. Moreover, no significant changes were observed in normal
   MEFs after blocking the FAP protein ([110]Fig. 3B). These findings
   confirmed that the FAP-targeting peptide significantly enhances the
   targeting ability of FAM-siRNA to activated fibroblasts.

Fig. 3. Targeting and endosomal escape ability of FL-NI in vitro.

   [111]Fig. 3.
   [112]Open in a new tab

   (A and B) Fluorescence intensity of normal (A) and activated (B) MEFs
   incubated with L-siRNA (pink line), FL-siRNA (orange line), or FL-siRNA
   with pretreatment with anti-FAP antibodies (1:500 concentration, 2
   hours, 37°C) (green line). Bar graphs indicate the percentage of cells
   exhibiting fluorescence offset. (C) Fluorescence microscopic images of
   normal and activated MEFs treated with L-siRNA and FL-siRNA for 2 and 4
   hours at 37°C. Scale bars, 20 and 5 μm (low- and high-magnification
   images, respectively). (D) Confocal laser scanning microscopy images of
   MEFs treated with L[PEG]-siRNA or L-siRNA for 2 and 4 hours at 37°C.
   Endosomes/lysosomes were labeled by LysoTracker Red (red), and nuclei
   were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Scale
   bars, 20 and 5 μm (low- and high-magnification images, respectively).
   (E) Magnified images of the region of interest with colocalization
   analysis. All data are presented as means ± SD (n = 3). Statistical
   significance was calculated via one-way ANOVA [(A) and (B)] in GraphPad
   Prism. **P < 0.01; ****P < 0.0001; and not significant, P > 0.05. h,
   hours.

   Simultaneously, activated and normal MEFs were treated with either
   L-siRNA or FL-siRNA for 2 or 4 hours, followed by analysis using
   confocal microscopy. Consistent with the flow cytometry results,
   FL-siRNA exhibited higher FAM fluorescence intensity in activated MEFs
   compared to L-siRNA. In contrast, normal MEFs showed no notable
   increase in fluorescence when treated with FL-siRNA ([113]Fig. 3C). In
   addition, coumarin 6, a lipophilic fluorescent probe, was encapsulated
   in liposomes to create FL-C6 and L-C6 for modeling the intracellular
   distribution of the hydrophobic drug NIN. It was found that coumarin 6
   fluorescence intensity in the FL-C6 group was higher in activated MEFs
   (fig. S6). Collectively, these results demonstrate the specific
   targeting and delivery capacity of FL-siRNA to activated fibroblasts,
   facilitated by the FAP-targeting peptide.

   Once internalized by cells, liposomes were typically sequestered in
   endocytic vesicles through an energy-dependent endocytic mechanism
   ([114]35), and failure to escape the endosome will hinder drug release
   and therapeutic effects ([115]36). We then investigate the endosomal
   escape capacity of liposomes mediated by PEOz, which reported to
   intensify the “proton sponge effect” ([116]37). Therefore, we
   synthesized L[PEG]-siRNA using DSPE-PEG[2k] as a control for L-siRNA.
   MEFs were treated with either L[PEG]-siRNA or L-siRNA for varying
   durations, and endosomes/lysosomes were labeled with LysoTracker Red.
   Confocal microscopy was then used to monitor the distribution of
   FAM-siRNA within endosome/lysosome ([117]Fig. 3D). It was found that
   FAM-siRNA exhibited lower colocalization with endosomes/lysosomes in
   cells treated with L-siRNA compared to those treated with L[PEG]-siRNA,
   as evidenced by fewer colocalized puncta indicated by orange-colored
   fluorescence, suggesting enhanced endosomal escape ([118]Fig. 3E). In
   addition, with extended incubation time, there was a gradual decrease
   in the colocalization of FAM-siRNA within endosomes/lysosomes, further
   indicating that L-siRNA has an enhanced ability to escape endosomal
   compartments.

Antifibrotic effect of FL-NI in vitro

   After confirming that FL-NI facilitates drug delivery into activated
   fibroblasts via receptor-mediated endocytosis through the FAP-targeting
   peptide, we processed to evaluate its antifibrotic effects in vitro.
   First, the cell viability assay was performed, and the results
   indicated that the cell viability remained 80% after MEFs were
   incubated with different concentrations of FL-NI or L-NI for 24 hours
   (fig. S7). Subsequently, the activated MEFs after stimulating by TGF-β1
   were coincubated with L-I, L-N, L-NI, and FL@NI with equal amounts of
   NIN (1 μmol ml^−1) and siIL11 (40 pmol ml^−1) for 24 hours. Then, we
   used immunofluorescence staining to examine the expression of α–smooth
   muscle actin (α-SMA) and collagen I (COL1A1) as they are key markers of
   PF and indicators of fibroblast activation ([119]10). As expected, MEFs
   exhibited strong red fluorescence of α-SMA and COL1A1 when treated with
   TGF-β1. However, fluorescence associated with both α-SMA ([120]Fig. 4,
   A and B) and COL1A1 ([121]Fig. 4, C and D) decreased to varying degrees
   after treatment with different formulations. Notably, only slight
   fluorescence was detected in the cells treated with dual-drug
   combination therapy, such as L-NI or FL-NI, with the FL-NI group
   showing fluorescence levels reduced to near normal. Slight variation in
   COL1A1 fluorescence was observed between the L-NI and FL-NI treatments,
   which may be attributed to the diminished impact of endocytosis
   differences due to prolonged incubation time. In addition, both Western
   blot (WB) ([122]Fig. 4E) and quantitative polymerase chain reaction
   (qPCR) ([123]Fig. 4, F to H) results confirmed varying degrees of
   reduction in protein and mRNA expression levels of α-SMA and COL1A1
   among various treatments, with MEFs treated with FL-NI exhibiting the
   lowest expression levels.

Fig. 4. The activation and migration of fibroblast inhibited by FL-NI in
vitro.

   [124]Fig. 4.
   [125]Open in a new tab

   (A and B) Immunofluorescence analysis of α-SMA expression in
   fibroblasts. Scale bars, 20 μm. Data are presented as means ± SD
   (n = 5). (C and D) Immunofluorescence analysis of COL1A1 expression in
   fibroblasts. Scale bars, 20 μm. Data are presented as means ± SD
   (n = 5). (E) WB analysis of COL1A1 and α-SMA in fibroblasts treated by
   PBS, L-N, L-I, L-NI, and FL-NI. (F to H) qPCR analysis of IL-11,
   COL1A1, and α-SMA in fibroblasts treated by PBS, L-N, L-I, L-NI, and
   FL-NI. Data are presented as means ± SD (n = 3). (I) Representative
   images and (J) statistical analysis of migration cell counts from the
   transwell migration assay in fibroblasts with different treatments.
   Scale bar, 50 μm. Data are presented as means ± SD (n = 4). (K)
   Representative images and (L) statistical analysis on wound healing
   area of fibroblast with different treatments. Scale bars, 200 μm. Data
   are presented as means ± SD (n = 3). Statistical significance was
   calculated via one-way ANOVA in GraphPad Prism. **P < 0.01;
   ***P < 0.001; ****P < 0.0001; not significant, P > 0.05. GAPDH,
   glyceraldehyde phosphate dehydrogenase.

   Given the pivotal role of fibroblast migration into fibroblastic foci
   during PF development ([126]38), we further assessed the inhibitory
   effect of FL-NI on this process using transwell migration and wound
   healing assays. MEFs after the same treatment procedure as above were
   seeds to wells for the migration experiment (fig. S8), and the
   migration capability of MEFs was quantified by counting migrated cells.
   The migrated number of MEFs after activated by TGF-β1 was 160 ± 11,
   while the number decreased to 36 ± 7 after treated with the FL-NI,
   significantly lower than 66 ± 11 of those with L-NI, indicating the
   superior efficacy of FL-NI in attenuating TGF-β1–induced activation and
   migration of MEFs ([127]Fig. 4, I and J). For the wound healing
   experiments, MEFs subjected to same treatments with the above and the
   images were captured at 0 and 24 hours post-administration ([128]Fig.
   4K), and the relative wound closure area at 24 hours compared to 0
   hours was analyzed to evaluate the migration ability ([129]Fig. 4L).
   When MEFs were activated with TGF-β1, the wound healing area was
   66.02 ± 5.69%. However, following incubation with FL-NI, the wound
   healing area significantly decreased to 17.97 ± 3.95%, which was
   notably lower than 29.82 ± 2.50% observed in the L-NI treatment group.
   These findings suggest that FL-NI, which codelivers NIN and siIL11 to
   myofibroblasts, effectively induces their deactivation and inhibits
   their migration in the context of PF. This outcome demonstrates
   superior antifibrotic efficacy compared to L-NI, likely due to the
   enhanced internalization facilitated by the FAP-targeting peptide.

Targeted delivery of FL-NI in fibrotic lungs

   To investigate the targeted delivery of FAP-targeting
   peptide–conjugated liposomes in vivo, the naked siRNA, L-siRNA, and
   FL-siRNA containing equivalent cyanine 5 (Cy5) siRNA were inhaled into
   the lungs of mice by a pulmonary spray needle. After 24 hours, major
   organs were harvested, and the fluorescence distribution and intensity
   of Cy5 were visualized using an in vivo imaging systems (IVIS) imaging
   system ([130]Fig. 5A and fig. S9). The results demonstrated a uniform
   and robust Cy5 fluorescence signal in the lungs of the FL-siRNA group,
   indicating the superior lung deposition of FL-NI through enhanced
   uptake by myofibroblasts, compared to the L-siRNA group. In contrast,
   naked siRNA, without protective encapsulation, showed significant
   degradation and elimination, as evidenced by reduced Cy5 fluorescence
   signals. Moreover, these data highlight the advantages of inhalation
   administration, as weak Cy5 fluorescence signals were observed in the
   liver, kidney, heart, and spleen across all three groups ([131]Fig.
   5A). Quantitative analysis of fluorescent signals in these organs
   ([132]Fig. 5B) and calculation of the lung/liver ratio ([133]Fig. 5C)
   further supported the preferential lung deposition of FL-NI compared to
   L-NI, which was also corroborated with the immunofluorescence
   fluorescence imaging of frozen lung sections, where the naked siRNA
   group exhibited minimal fluorescence signals due to the degradation by
   nucleases ([134]Fig. 5D). To further explore the reasons behind the
   superior pulmonary deposition of FL-siRNA, frozen lung sections from
   the L-siRNA and FL-siRNA groups were subjected for FAP
   immunofluorescence staining. Confocal microscopy images revealed a
   higher degree of colocalization of FAP-positive fibroblasts (labeled by
   AF488) and Cy5-siRNA in the FL-siRNA group, manifesting enhanced
   targeting specificity of myofibroblasts in vivo compared to the L-siRNA
   group ([135]Fig. 5E).

Fig. 5. Targeted delivery of FL-NI in vivo.

   [136]Fig. 5.
   [137]Open in a new tab

   (A) IVIS images of the heart, liver, lungs, spleen, and kidneys, 24
   hours after inhalation of naked Cy5-siRNA, L-siRNA, and FL-siRNA. (B)
   Fluorescence intensity of the heart, liver, lung, spleen, and kidney,
   expressed as average radiant efficiency units. (C) Fluorescence
   intensity ratio of the lung to liver. (D) Confocal microscopy image of
   Cy5-siRNA distribution in the lung sections in naked siRNA, L-siRNA,
   and FL-siRNA groups. Scale bars, 100 μm. (E) Immunofluorescence
   distribution analysis of FAP^+ myofibroblasts and Cy5-siRNA in the lung
   sections in L-siRNA and FL-siRNA groups. Scale bars, 20 μm. (F)
   Representative flow cytometry diagram showed the distribution of
   Cy5-siRNA in the lungs after inhalation of the different preparations.
   (G to J) Statistical data showed the proportion of Cy5-positive cells
   in (G) myofibroblasts, (H) endothelial cells, (I) immune cells, and (J)
   epithelial cells after inhalation of the different preparations. All
   data are presented as means ± SD (n = 3). Significant difference was
   calculated via one-way ANOVA [(B), (C), and (G) to (J)] in GraphPad
   Prism. AF488, Alexa Fluor 488; *P < 0.05; **P < 0.01; ***P < 0.001; not
   significant, P > 0.05.

   To precisely determine the distribution of FL-siRNA among major cell
   subtypes in the PF microenvironment, we digested the lungs and prepared
   into single-cell suspensions for flow cytometry analysis. This analysis
   quantified the proportion of Cy5-positive cells among epithelial cells
   (CD326^+), endothelial cells (CD31^+), immune cells (CD45^+), and
   activated fibroblasts (CD45^− and FAP^+) ([138]Fig. 5F), with the flow
   cytometry gating strategy shown in fig. S10. The results indicated that
   the FL-siRNA group exhibited a significantly higher proportion of
   Cy5-positive cells in fibroblasts, accounting for 21.3%, compared to
   12.2% in the L-siRNA group ([139]Fig. 5G). Conversely, negligible
   difference was observed in the proportion of Cy5-positive cells between
   the FL-siRNA and L-siRNA groups among endothelial cells ([140]Fig. 5H).
   The proportion of Cy5-positive cells was significantly reduced in both
   immune cells ([141]Fig. 5I) and epithelial cells ([142]Fig. 5J) in the
   FL-siRNA group. These findings suggest that following inhalation,
   FL-siRNA exhibits a higher propensity for uptake by FAP-positive
   fibroblasts while demonstrating reduced nonspecific entry into
   epithelial cells and immune cells. This indicates that FAP peptides
   facilitate targeted delivery in vivo, enhancing the therapeutic
   potential of FL-siRNA for treating PF.

Antifibrotic effects of FL-NI in vivo

   To evaluate the antifibrotic efficacy of FL-NI in vivo, a PF mouse
   model was established using a single intratracheal injection of
   bleomycin (BLM) at a dose of 2 USP kg^−1, followed by a 7-day
   progression period. Then, BLM-induced C57BL/6 mice were randomly
   divided into the following groups: BLM, L-N, L-I, L-NI, and FL-NI, with
   normal C57BL/6 mice serving as the normal control group. Starting on
   day 1 postmodeling and continuing until day 14, each group received
   corresponding formulation with NIN and siIL11 at doses of 1 mg kg^−1
   each time via intratracheal instillation every other day ([143]Fig.
   6A). Throughout the experiment, the weights of mice were monitored
   every 2 days as an indicator of disease burden. Normal mice exhibited
   gradual weight gain, whereas all mice treated with BLM experienced
   weight loss. Notably, compared to the BLM group, mice receiving
   different formulations displayed varying degrees of weight loss
   attenuation. Among these, the L-NI group was more effective than the
   L-N and L-I groups in slowing weight loss. Notably, FL-NI–treated mice
   demonstrated significantly greater alleviation of weight loss compared
   to all other groups, likely due to the enhanced in vivo targeting
   capability of FAP ([144]Fig. 6B).

Fig. 6. FL-NI alleviates BLM-induced PF in vivo.

   [145]Fig. 6.
   [146]Open in a new tab

   (A) Timeline for modeling PF and inhalation treatment. (B) Body weights
   of mice in different groups. G1: normal; G2, BLM; G3, L-N; G4, L-I; G5,
   L-NI; G6, FL-NI. (C) Morphologic images of lungs after mice were
   euthanized at the end of treatment period. (D) Quantitative analysis of
   hydroxyproline content in mouse lung tissues after different
   treatments. (E) Ashcroft scores of fibrotic lung sections. (F) H&E
   staining of lung sections from different groups. The collagen fibers
   are stained blue in (G) Masson staining and red in (H) Van Gieson
   staining of lung sections from different groups. (I to K)
   Immunohistochemistry staining of COL1A1 (I), α-SMA (J), and FN1 (K) of
   lung sections from different groups. (L) Immunofluorescence analysis of
   SFTPC of lung sections from different groups. (M) WB analysis of IL-11,
   COL1A1, and α-SMA expression of lung sections from different groups.
   (N) qPCR analysis of α-SMA, COL1A1, FN1, and SFTPC expression of lung
   sections from different groups. All data are means ± SD (n = 5).
   Significant difference was calculated via Student’s t test (B) and
   one-way ANOVA [(D), (E), and (N)] in GraphPad Prism. *P < 0.05;
   **P < 0.01; ***P < 0.001; ****P < 0.0001; not significant, P > 0.05.

   At the end of treatment, the mice were euthanized for different
   analysis. As exhibited in [147]Fig. 6C, mice treated with BLM alone
   exhibited hemorrhagic necrosis in the lung, while the FL-NI–treated
   group demonstrated substantial restoration of lung morphology.
   Quantitative analysis of hydroxyproline levels ([148]Fig. 6D) revealed
   that the FL-NI group exhibited the most effective remodeling of
   pathological collagen accumulation in the lungs. To visualize alveolar
   structure damage caused by PF, we performed hematoxylin and eosin (H&E)
   staining of lung sections and Ashcroft scoring. The Ashcroft score of
   BLM-induced fibrosis was 7.6 ± 0.6, but it was significantly reduced to
   2.8 ± 0.5 following FL-NI inhalation, which was significantly lower
   than that in the L-NI group ([149]Fig. 6E). H&E staining showed severe
   disruption of alveolar structure in BLM-induced fibrosis model,
   characterized by extensive fibrotic foci, thickened alveolar walls, and
   infiltration of inflammatory cells. The FL-NI–treated group exhibited
   marked reduction in necrotic areas, with near-complete restoration to
   normal alveolar architecture ([150]Fig. 6F and fig. S11). Given that
   excessive pathological collagen accumulates in the pulmonary
   interstitium of fibrotic tissue, leading to parenchymal defects and
   respiratory impairment, Masson’s trichrome staining was used to
   differentiate collagen distribution in the lung ([151]Fig. 6G). The
   FL-NI group showed greater reduction in interstitial collagen
   deposition compared to the L-NI group, indicating that FL-NI has an
   optimal antifibrotic effect. Similarly, Van Gieson staining confirmed
   the same results ([152]Fig. 6H). Collectively, these findings
   demonstrated the potent efficacy of FL-NI in reducing collagen protein
   and fiber deposition in BLM-induced PF.

   In PF, the up-regulation of α-SMA and excessive pathological collagen
   secretion are key contributors to disease progression ([153]10).
   Immunohistochemical staining was carried out to evaluate the expression
   levels of COL1A1, α-SMA, and fibronectin 1 (FN1) in fibrotic lungs. It
   was found that FL-NI treatment substantially normalized the expression
   of COL1A1 ([154]Fig. 6I), α-SMA ([155]Fig. 6J), and FN1 ([156]Fig. 6K),
   indicating effective inhibition of activated fibroblasts in the lung
   interstitium and elimination of the ECM. Given the strong association
   between PF and epithelial injury ([157]39), we further assessed the
   repair and functional remodeling of the epithelium by
   immunofluorescence staining of surfactant protein C (SFTPC). Obviously,
   the increased expression of SFTPC was observed in the FL-NI treatment
   group, suggesting its reparative effect on the epithelium. These
   findings were corroborated by WB, which showed significant
   up-regulation of IL-11, COL1A1, and α-SMA in the BLM group, with
   varying degrees of down-regulation following FL-NI treatment, neatly
   restoring them to normal levels ([158]Fig. 6M). The qPCR results
   further confirmed the decreased expression of COL1A1, α-SMA, and FN1,
   alongside increased SFTPC levels after FL-NI treatment ([159]Fig. 6N).
   Considering that fibrosis in young mice usually spontaneously reverse 4
   weeks after a single intratracheal instillation of BLM, despite this
   model being the most widely used for studying the pathogenesis and
   therapeutic effects of PF, we used both the repetitive BLM-induced PF
   model and the aged mice PF model to access the antifibrosis effects of
   FL-NI. In the repetitive BLM model, the treatment process followed the
   experimental timeline outlined in the flow chart (fig. S12A). It was
   found that FL-NI treatment extended survival (fig. S12B) and
   significantly alleviated fibrosis progression in mice subjected to
   repetitive BLM treatment (fig. S12C). In addition, H&E and Masson
   staining revealed substantial restoration of alveolar structure
   disruption and a marked reduction in collagen deposition in the
   FL-NI–treated mice (fig. S12D). Notably, the Ashcroft score for the
   FL-NI group was reduced to 1.8 ± 1.5 (fig. S12E), in contrast to the
   higher score observed in the BLM group. Consistently, in the aged mice,
   following the administration protocol outlined in fig. S13A, mice
   inhaled with FL-NI exhibited significantly greater alleviation of
   weight loss (fig. S13B), hemorrhagic necrosis (fig. S13C), alveolar
   structure disruption, and collagen deposition (fig. S13D) compared to
   the BLM group. The Ashcroft score in the FL-NI–treated group was
   notably lower, at 2.6 ± 1.1 (fig. S13E), compared to the BLM group.
   These results collectively highlight the exceptional antifibrotic
   efficacy of FL-NI. Moreover, FL-NI and other formulations showed no
   additional toxicity to major organs such as the liver and kidneys.
   Quantification of alanine aminotransferase and aspartate
   aminotransferase levels in mouse serum indicated negligible
   hepatotoxicity (fig. S14). Similarly, H&E staining of the heart, liver,
   spleen, and kidneys revealed nearly no histopathological changes in
   these organs (fig. S15). Therefore, FL-NI effectively inhibits
   excessive fibroblast activation and ECM deposition while promoting
   epithelial repair in vivo, without causing side effects to other major
   organs.

The mechanism of FL-NI for antifibrotic treatment

   To elucidate the underlying mechanism by which FL-NI ameliorates PF, we
   performed bulk RNA sequencing (RNA-seq) analysis on lung tissues from
   the normal, BLM, and FL-NI groups (every group contains five samples)
   ([160]Fig. 7A). Transcriptome sequencing revealed significant gene
   expression changes: 413 genes were down-regulated, and 1715 mRNAs were
   up-regulated in the BLM group compared to the normal group, while 608
   genes were down-regulated and 840 genes were up-regulated in the FL-NI
   group compared to the BLM group ([161]Fig. 7B). Notably, FL-NI
   treatment effectively restored the expression of fibrosis-related
   molecules. Specifically, it down-regulated the expression of ECM
   components, growth factors and their receptors, chemokines and their
   receptors, matrix metalloproteinases, and other fibrosis-associated
   regulatory proteins while up-regulating surfactant proteins ([162]Fig.
   7C). To further identify the pathways modulated by FL-NI, we
   intersected the mRNAs up-regulated in the BLM group and down-regulated
   in the FL-NI group, identifying 451 differentially expressed mRNAs for
   functional enrichment analysis ([163]Fig. 7D). The Kyoto Encyclopedia
   of Genes and Genomes (KEGG) enrichment analysis indicated that FL-NI
   treatment significantly inhibited several vital signaling pathways,
   including cytokine-cytokine receptor interaction, chemokine, Janus
   kinase/signal transducer and activator of the transcription (JAK-STAT),
   phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PI3K-AKT),
   nuclear factor κB, and Ras signaling pathways (highlighted in orange)
   ([164]Fig. 7E). In addition, Gene Ontology (GO) analysis revealed that
   FL-NI regulated the process related to cell chemotaxis, immune
   response, and inflammatory response (highlighted in green). Notably,
   the inflammatory responses regulated by FL-NI appear to be associated
   with cytokine signaling pathways involving TNF, IL-6, and IL-1, as
   visualized in an enriched chord diagram (highlighted in blue)
   ([165]Fig. 7, F and G). To validate these findings, we measured the
   levels of TNF-α ([166]Fig. 7H), IL-6 ([167]Fig. 7I), and IL-1β
   ([168]Fig. 7J) in lung tissue homogenate supernatants using
   enzyme-linked immunosorbent assay (ELISA). The results showed that
   cytokine levels were significantly reduced in the FL-NI group.
   Therefore, siIL11 and NIN inhibit the activation of IL-11RA and
   PDGFR/FGFR signaling pathways, respectively, which, in turn, suppress
   downstream PI3K-AKT and Ras signaling pathways to alleviate fibrosis.
   In addition, reducing IL-11 and associated chemokine levels helps
   attenuate the recruitment of inflammatory cells and decreases the
   expression of TNF-α, IL-1β, and IL-6, thereby suppressing the
   inflammatory response. As a result, FL-NI demonstrates great potential
   for effectively regulating the pathological microenvironment in the
   mesenchyme and remolding the parenchyma in PF treatment ([169]Fig. 7K).

Fig. 7. Antifibrotic mechanism of FL-NI.

   [170]Fig. 7.
   [171]Open in a new tab

   (A) RNA-seq analysis of the lung tissue of mice from different groups.
   (B) Volcano plots of differentially expressed genes comparing normal
   versus BLM and BLM versus FL-NI (the latter versus the former). (C)
   Heatmap illustrating the average expression of fibrosis-related genes
   in lung tissue of mice from different groups. (D) Venn diagram showing
   the intersection of genes regulated by BLM and FL-NI. (E) KEGG
   enrichment analysis of 451 mRNAs that are simultaneously down-regulated
   by FL-NI and up-regulated by BLM. (F) GO enrichment analysis revealing
   biological pathways enriched for 451 mRNAs. (G) Chord diagram showing
   the enrichment of KEGG pathways and the related genes. (H to J) The
   levels of TNF-α (H), IL-6 (I), and IL-1β (J) in lung tissue homogenates
   measured by ELISA. Data are means ± SD (n = 5). (K) Diagram
   illustrating the dual-drug synergistic mechanism for antifibrotic
   treatment. Significant difference was calculated via one-way ANOVA [(H)
   to (J)] in GraphPad Prism. **P < 0.01; ***P < 0.001; ****P < 0.0001;
   not significant, P > 0.05. FDR, false discovery rate; FC, fold change;
   NF-κB, nuclear factor κB; MAPK, mitogen-activated protein kinase.

   To further investigate the role of FL-NI in immune response and cell
   chemotaxis in PF, immune-related signaling pathway enrichment analysis
   was conducted using the RNA-seq data. The analysis revealed that FL-NI
   treatment significantly influenced neutrophil, monocyte, and macrophage
   chemotaxis ([172]Fig. 8A). Furthermore, using the ImmuCellAI-mouse
   (Immune Cell Abundance Identifier for mouse) ([173]40), we evaluated
   immune cell infiltration in PF lung tissues. The results showed a
   significant increase in the abundance of infiltrated immune cells
   including macrophages, neutrophils, and monocytes in the BLM group,
   which was restored to normal levels following FL-NI treatment
   ([174]Fig. 8B and fig. S16). To validate these findings, flow cytometry
   analysis was performed. C57/BL6 mice were treated according to the
   protocol described in [175]Fig. 6A, and single-cell suspensions were
   prepared from lung lobes for analysis ([176]Fig. 8C). The results
   demonstrated that FL-NI treatment significantly reduced the percentage
   of monocytes in fibrotic lungs ([177]Fig. 8D). Similarly, the
   proportion of neutrophils or granulocytic myeloid-derived suppressor
   cells (MDSCs) (CD11b^+Ly6G^+Ly6C^−) was also obviously decreased in the
   FL-NI group, while no significant changes were observed in the
   proportion of monocytic MDSCs (CD11b^+Ly6G^−Ly6C^+) ([178]Fig. 8E and
   fig. S17). Moreover, macrophages, known to induce fibroblast
   differentiation into myofibroblasts to promote PF progression
   ([179]41), were significantly reduced following FL-NI treatment
   ([180]Fig. 8F). Further analysis divided pulmonary macrophages into
   CD11c^+ [resident alveolar macrophages (AM)] and CD11b^+ (interstitial
   macrophages or recruited AM). CD11b^+ macrophages, which are implicated
   in fibrosis development and the promotion of a profibrotic phenotype
   ([181]42), were significantly elevated in PF lungs. However, their
   levels returned to normal following FL-NI treatment ([182]Fig. 8G).
   Conversely, there was a notable trend in CD11c^+ resident AM (fig.
   S17). Furthermore, the M2 macrophages in CD11b^+ macrophages are
   reported to regulate tissue damage repair by secreting profibrotic
   cytokines ([183]41). Consistently, the results showed that the
   proportion of M2 macrophages significantly increased in the BLM group
   and decreased in all treatment groups, especially returning to normal
   in the FL-NI group ([184]Fig. 8H). Collectively, these findings suggest
   that inhalation therapy with FL-NI can effectively remodel the normal
   immune microenvironment and protect against BLM-induced PF.

Fig. 8. Immune microenvironment of PF remodeled by FL-NI.

   [185]Fig. 8.
   [186]Open in a new tab

   (A) GO enrichment analysis results for immunomodulatory biological
   pathways. (B) Analysis of immune cell infiltration in different groups
   of mice. (C) Schematic representation of immuno-evaluation of lung
   tissues of mice from different groups. (D to H) Representative flow
   cytometry plots and percentages of CD14^+ cells in CD45^+ cells (D),
   Ly6G^+Ly6C^− and Ly6C^+Ly6G^− cells in CD11b^+ cells (E), CD11b^+ cells
   in F4/80^+ cells (F), CD11b^+CD11c^− and CD11c^+CD11b^− cells in
   F4/80^+ cells (G), and CD206^+CD80^− cells in F4/80^+CD11b^+ cells (H)
   in mouse lungs after various treatments. Data are means ± SD (n = 5).
   Significant difference was calculated via one-way ANOVA [(H) to (J)] in
   GraphPad Prism. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001;
   not significant, P > 0.05.

DISCUSSION

   In this work, we embarked on an innovative “one-stone-three-birds”
   strategy by developing a multifunctional nanoplatform (FL-NI) designed
   to achieve efficient and step-by-step drug delivery across three key
   dimensions: organ deposition, cellular endocytosis, and intracellular
   release. This strategy was tailored to enhance the targeted delivery
   and release of a combination of siIL11 and NIN, ultimately facilitating
   a synergistic therapeutic effect to improve treatment efficacy.
   Numerous therapeutic strategies have been proposed for PF; however,
   drug delivery remains inefficient due to the reliance on either
   intravenous or inhalation administration without cellular targeting
   ([187]43–[188]45). Nanoparticles often undergo nonspecific uptake by
   normal cells, seriously limiting the efficacy of antifibrotic
   treatments following inhalation. Therefore, combining inhaled delivery
   with myofibroblast targeting offers a promising strategy for enhanced
   drug delivery and improved therapeutic outcomes. Specifically, our
   findings demonstrated that inhalation administration of FL-NI
   significantly enhanced lung deposition, achieving a 117.8% increase
   compared to the deposition observed with L-NI. This was followed by a
   precise interaction between FAP-targeting peptides and FAP on the
   surface of activated fibroblasts, resulting in a 71.5% increase in
   myofibroblast targeting compared to L-NI. This was accompanied by 29.8
   and 55.8% reduction in nonspecific immune cell and epithelial uptake,
   respectively. Furthermore, FL-NI effectively circumvented the acidic
   lysosomal environment through its pH-responsive PEOz components,
   enabling controlled drug release. Subsequent in vivo assessments
   revealed that FL-NI inhalation exhibited strong antifibrosis and
   immune-regulatory efficacy. Specifically, it significantly reduced
   collagen deposition by 50.8%, bringing it back to normal levels, and
   reduced the Ashcroft score from 7.6 ± 0.55 to 2.8 ± 0.45. In addition,
   FL-NI treatment inhibited fibroblast activation and structural damage
   associated with PF. Besides the single BLM instillation model, in
   progressive PF models that closer to PF clinical pathology, such as
   repetitive instillation BLM and aged mouse model ([189]46, [190]47),
   FL-NI still demonstrated excellent antifibrotic capacity, indicating a
   strong clinical value. FL-NI inhalation also normalized the profibrotic
   microenvironment, showing significant reductions in inflammatory
   factors, such as IL-1β, IL-6, and TNF-α, as well as in immune cell
   infiltration, including M2 macrophages and granulocytic MDSCs,
   indicating remarkable restoration of tissue homeostasis. Mechanistic
   investigations, including RNA-seq, revealed that the enhanced
   antifibrotic efficacy of FL-NI was derived from the synergistic effects
   of siIL11 and NIN. These components effectively inhibited key signaling
   pathways involved in fibrosis, including JAK-STAT, PI3K-AKT, and ERK
   pathways. Moreover, evaluations of physicochemical properties and
   safety of FL-NI demonstrated its long-term colloidal stability and high
   biocompatibility after inhalation in mice, without any safety concerns.
   In conclusion, the multifunctional nanoplatform developed in this work
   represents a significant advancement in drug delivery precision,
   providing a highly effective therapeutic strategy for PF, and it not
   only demonstrates substantial therapeutic potential but also offers
   insights for the future development of noninvasive precision therapy.

MATERIALS AND METHODS

Materials

   The compounds DOTAP and cholesterol were acquired from Weihua
   Biotechnology Co. Ltd. (Guangzhou, China). The liposome size extruder
   was obtained from Avanti Polar Lipid Inc. (Alabaster, AL, USA), while
   polycarbonate membranes were sourced from Merck Millipore (Billerica,
   MA, USA). The NIN, MTT
   [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide],
   protamine sulfate, and hyaluronic acid were acquired from Sigma-Aldrich
   (St. Louis, MO). The FAP-targeting peptide (DRGETGPAC) was obtained
   from Jiangsu Jitai Technology Co. Ltd. DSPE-PEOz[2k]-MAL was provided
   by Weihua Biotechnology Co. Ltd. (Guangzhou, China), and DSPE-PEOz-FAP
   was synthesized by combining the FAP-targeting peptide with
   DSPE-PEOz[2k]-MAL based on the previous established protocol
   ([191]33).The serum-free cell cryopreservation solution (FL10020L) and
   trypsin-EDTA digestion solution (FL20011) were purchased from Bioland
   Biotechnology Co. Ltd. (Hangzhou, China).

Synthesis of FL-NI

   FL-NI were prepared according to a stepwise self-assembly process
   described in the previous established protocols ([192]33). First, using
   a hydration-extrusion method, the DOTAP/cholesterol/NIN lipid film was
   hydrated with diethyl pyrocarbonate–treated water and followed by
   extrusion through 200- and 400-nm polycarbonate membranes (Millipore,
   Billerica, MA) to form cationic liposomes. For the siIL11 core
   formation, solution A (containing siRNA and hemagglutinin) and solution
   B (containing protamine) were allowed for 10 min. The resulting siIL11
   core solution was then incubated with the prepared liposomes for 10 min
   to encapsulate siIL11. Following this, DSPE-PEOz-FAP or DSPE-PEOz[2k]
   was coincubated with the mixture for 1 hour at 50°C to form liposomes
   modified with or without the FAP targeting peptide. Free NIN and
   unencapsulated siIL11 were removed by ultrafiltration centrifugation
   using a 100-kDa MWCO filter (Millipore, USA). For comparison, other
   liposomal formulations containing NIN (L-N), siIL11 (L-I), or both
   (L-NI) were synthesized following the same procedure.

Characterization of FL-NI

   The particle size, zeta potential, and the morphology of the
   nanoparticles were determined using Zetasizer Nano ZS (Malvern
   Instrument) and transmission electron microscopy (FEI Tecnai 20). The
   free NIN and siRNA unencapsulated in liposomes were removed by
   ultrafiltration centrifugation. The microplate reader (Bio-Rad) was
   used to detect the free and total amount of NIN and siRNA to calculate
   the encapsulation efficiency according to the following equations;
   encapsulation efficiency (EE) % = (W[Total] − W[Free])/W[Total] × 100%.
   Furthermore, the stability of the nanoparticles was assessed by
   measuring particle size at predetermined time points in DMEM containing
   10% FBS.

Electrophoretic mobility shift assay

   Electrophoretic mobility shift assay was used to evaluate the
   resistance of siRNA encapsulated in FL-NI to nuclease degradation.
   Briefly, naked siRNA and FL-NI (containing an equivalent amount of
   siRNA) were coincubated with RNase (10 μg ml^−1) for 15, 30, 60, 120,
   and 240 min. After incubation, 2% sodium heparin solution was added to
   the FL-NI group to extract siRNA, followed by mixing with 6× loading
   buffer (No. K1161, APExBIO, Houston, USA). Equal volumes of naked siRNA
   and FL-NI samples were then loaded into gel electrophoresis wells, and
   electrophoresis was performed at 80 V for 12 min. The results were
   photographed under ultraviolet irradiation.

Assessment of specific binding ability of FL-siRNA

   MEFs (0.8 × 10^5 cells per well) were seeded onto 24-well plates
   containing coverslips (1170000, SAINING Biotechnology) for overnight at
   37°C. Then, these fibroblasts were activated by incubation with TGF-β1
   (10 ng ml^−1) for 24 hours. The activated and normal MEFs were washed
   with phosphate-buffered saline (PBS) and then treated with L-siRNA and
   FL-siRNA for 2 and 4 hours at 37°C. The cells were then fixed with 4%
   paraformaldehyde for 1 hour, followed by nuclei staining with
   4′,6-diamidino-2-phenylindole. Cells were mounted on coverslips with an
   antifade reagent. FAM fluorescence was subsequently visualized using
   confocal fluorescence microscopy.

   Meanwhile, for flow cytometry analysis, MEFs (1.5 × 10^5 cells per
   well) were seeded into 12-well plates overnight and then activated by
   TGF-β1 (10 ng ml^−1) for 24 hours. Both activated and normal MEFs were
   treated with L-siRNA or FL-siRNA containing an equivalent amount of
   FAM-siRNA for 2 hours at 37°C. To block cell surface ligands targeted
   by FAP-targeting peptides, cells were pretreated with 1:500 anti-FAP
   antibodies for 2 hours before the treatment with FL-siRNA. After
   incubation, cells were harvested by trypsinization, resuspended in
   300 μl of PBS, and analyzed for FAM fluorescence (siRNA
   internalization) using flow cytometry.

Antifibrosis treatment in vitro

   MEFs were seeded in six-well plates until they reached 90% confluence.
   MEFs were incubated with TGF-β1 (10 ng ml^−1) for 24 hours for
   activation. After activation, the following treatments were applied to
   the MEFs for another 24 hours: L-N, L-I, L-NI, FL-NI, or an equivalent
   volume of PBS. Normal MEFs treated with PBS served as the positive
   control. Following incubation, the treated cells were subjected to WB
   with precast gel (GenScript, M00664) and secondary antibodies
   (Signalway Antibody, L3032), qPCR, and wound healing assays for
   assessing the antifibrosis ability of different treatments.

Antifibrotic effects of FL-NI in vivo

   The PF models were established by receiving a single BLM (2 USP kg^−1)
   intratracheal injection for 7 days. BLM-induced C57BL/6 mice were
   randomly divided into the following groups: BLM, L-N, L-I, L-NI, and
   FL-NI, with normal C57BL/6 mice serving as the normal control group
   (n = 5). Each group received a corresponding formulation with the dose
   of NIN at 1 mg kg^−1 and siIL11 at 1 mg kg^−1 using a microsprayer
   aerosolizer every other day for 2 weeks. During the experiment, body
   weight was recorded every 2 days. The mice were euthanized for
   different analysis 2 days after the last treatment. All animal
   experiments were conducted in accordance with the relevant laws and
   guidelines and were approved by the Institutional Animal Care and Use
   Committee of Soochow University (no. SUDA20250304A06).

Content of hydroxyproline and cytokines in lungs

   The content of hydroxyproline in fibrotic lungs was evaluated by a
   hydroxyproline detection kit. The lung homogenized supernatants were
   used to measure IL-6 (Beijing Solarbio Science and Technology Co. Ltd.,
   SEKM-0007), IL-1β (Beijing Solarbio Science and Technology Co. Ltd.,
   SEKM-0002), and TNF-α (Beijing Solarbio Science and Technology Co.
   Ltd., SEKM-0034) by ELISA according to the manufacturer’s instructions.

RNA-seq and data analysis

   To explore the mechanism behind the antifibrotic efficacy of FL-NI,
   lung tissues from normal mice, PF mice, and PF mice treated with FL-NI
   were analyzed by RNA-seq on the Illumina HiSeq platform (GENEWIZ
   Biotech Inc.). Bioinformatic analysis of RNA-seq was performed using
   tools available at [193]https://omicshare.com/tools/. The
   fibrosis-related molecules listed and analyzed are derived from the
   Molecular Signatures Database ([194]48). The abundance of cellular
   infiltration in the lung microenvironment was evaluated on the basis of
   RNA-seq data from the mice with PF. Immune infiltration analysis was
   conducted using ImmuCellAI-mouse ([195]40), a tool designed to estimate
   the abundance of 36 different immune cell types from gene expression
   datasets. These immune cells are categorized into seven major types: B
   cell, natural killer cell, monocytes, macrophages, T cells, dendritic
   cells, and granulocytes.

In vivo immunological evaluation

   To investigate the immune regulation effects of FL-NI in vivo, mice in
   all groups were euthanized 2 weeks after treatment. The lungs were
   euthanized from each mouse to prepare single-cell suspensions. These
   suspensions were treated with red blood cell (RBC) lysis buffer
   (ZUNYAN, NanJing, China, ZYFB006) to eliminate RBC, following by
   washing and blocking with anti-CD16/CD32 (BioLegend, 101303) to prevent
   Fc receptor binding. The cells were then stained with
   fluorophore-labeled antibodies as follows. To analyze monocyte, the
   cells were stained with anti–CD14–fluorescein isothiocyanate (FITC;
   BioLegend, 123308) and anti–CD45–phycoerythrin (PE; BioLegend, 147712).
   To analyze AMs and interstitial macrophages, the cells were stained
   withanti-Percp-F4/80 (BioLegend, 123126), anti–CD11b-FITC (BioLegend,
   101206), and anti–CD11c-allophycocyanin (APC; BioLegend, 117310). To
   analyze M2 macrophages, the cells were stained with anti–Percp-F4/80
   (BioLegend, 123126), anti–CD11b-FITC (BioLegend, 101206), anti–CD80-PE
   (BioLegend, 104708), and anti–CD206-APC (BioLegend, 141708). To
   evaluate monocytic MDSCs and granulocytic MDSCs, the cells were stained
   with anti–Percp-CD45 (BioLegend, 103130), anti–CD11b-FITC (BioLegend,
   101206), anti–Ly6G-PE (BioLegend, 127608), and anti–Ly6C-APC
   (BioLegend, 128016). Subsequently, these cells were analyzed by flow
   cytometry (FongCyte).

Statistics and reproducibility

   Data are presented as means ± SD. Student’s t test or one-way analysis
   of variance (ANOVA) followed by the Tukey test was applied for
   comparison between two groups or among multiple groups using GraphPad
   Prism 9.5.1, respectively. P < 0.05 was considered to be statistically
   significant. Each experiment is repeated at least three times
   independently with similar results, and the representative dataset is
   presented.

Acknowledgments