GGT-triggered self-immolative dexamethasone-oligopeptide prodrug nanoparticles ameliorate hepatic fibrosis via dual anti-inflammatory/immunomodulatory pathways with minimized systemic toxicity

Abstract

Liver fibrosis remains a critical unmet medical challenge, where glucocorticoids’ therapeutic potential is constrained by systemic toxicity. We innovatively address this through an amphiphilic prodrug (DEI) engineered by Michael addition-conjugating dexamethasone with I-C-F-6, which self-assembles into γ-glutamyltranspeptidase (GGT)-responsive nanoparticles for liver-targeted delivery. This nanoplatform achieves spatial precision through enzyme-triggered drug self-immolative release exclusively in fibrotic livers, minimizes systemic toxicity, and modulates fibrosis pathogenesis via dual pathways. DEI-NPs could suppress hepatic stellate cell activation via TGF-β/Smad3 and TH17 pathway blockade to halt collagen I/IV deposition, and reprogram intrahepatic immunity through retinol metabolism-mediated TH17/Treg rebalancing (Cyp2a1↓/STAT1↑). Our work pioneers a paradigm-shifting strategy in glucocorticoid therapy that merges GGT-activated spatial control with immunometabolic reprogramming, effectively overcoming critical bottlenecks in fibrosis therapy.

Graphical Abstract

Schematic illustration of DEI NPs and their treatment effects on liver fibrosis in mice. (1) Fibrotic liver-specific activation via GGT-mediated cleavage, enabling targeted DEX release; (2) Dual-pathway inhibition of fibrogenesis—blocking HSCs activation through TGF-β/Smad3 suppression and TH17 pathway interruption to reduce collagen I/IV deposition; (3) Immunometabolic reprogramming via retinol metabolism modulation (Cyp2a1↓/STAT1↑) to restore TH17/Treg equilibrium

Introduction

Liver fibrosis, characterized by pathological deposition of extracellular matrix (ECM) components, predominantly collagens type I/III/IV—represents the common end-stage pathway of chronic liver diseases, with advanced stages progressing to cirrhosis, hepatic decompensation, and portal hypertension, ultimately requiring liver transplantation in > 30% of cases [1]. Its pathogenesis is primarily driven by hepatocyte-initiated inflammatory responses that recruit and polarize monocytes/macrophages, thereby inducing hepatic stellate cells (HSCs) trans-differentiation from quiescent [2, 3], vitamin A-storing phenotypes into activated, α-smooth muscle actin (α-SMA)-positive myofibroblasts [4–6]—the dominant ECM-producing cells [7]. And as liver fibrosis progresses, the balance between TH17 and Treg cells is crucial for immune responses, particularly in injured liver tissue and the restoration of immune homeostasis [8–10]. Research has shown that the development of TH17 cell promotes pro-inflammatory responses in almost all tissues [11, 12]. Retinoic acid acts as a key regulator of TGF-β dependent immune responses and also serves as an auxiliary factor in the generation of Treg cells [13]. It can inhibit the pro-inflammatory differentiation of TH17 cells induced by IL-6 and promote the differentiation of anti-inflammatory Treg cells [9]. Therefore, the key to treating liver fibrosis lies in suppressing the inflammation-induced activation of hepatic stellate cells and preventing further disruption of hepatic immune homeostasis through the restoration of immune balance.

Dexamethasone (DEX) has been well-established as a potent glucocorticoid with significant anti-inflammatory properties and demonstrated efficacy in combating hepatic fibrosis [14–20]. DEX will suppress the activation and proliferation of HSCs by inhibiting the stimulatory effects of infiltrating inflammatory cells such as monocyte-derived macrophages (MDMs), natural killer (NK) cells, and T lymphocytes, and attenuating the pro-inflammatory signaling cascades initiated by liver-resident Kupffer cells [19]. Despite its therapeutic benefits in modulating the fibrotic micro environment, the clinical translation of DEX is significantly limited by dose-dependent systemic adverse effects—such as osteoporosis, immunosuppression and hyperglycemia. Furthermore, its inherently low aqueous solubility considerably compromises bio-availability and hinders targeted delivery to hepatic tissues. These pharmacological constraints underscore the urgent need for novel drug delivery strategies that enhance targeting efficiency and retain the anti-fibrotic efficacy of DEX.

Peptide-conjugated prodrug nanoassemblies represent a transformative strategy for anti-hepatic fibrosis therapy by synergistically combining enzyme-responsive drug activation through liver-specific protease cleavage, and spontaneous self-assembly into nanostructures exhibiting > 90% drug loading efficiency and exceptional colloidal stability [21–24]. γ-glutamyltransferase (GGT) can specifically recognize and catalyze substrates containing gamma-glutamyl groups, and its active sites are mainly exposed to the extracellular environment [25–28]. It is widely used as a target for drug-specific binding. It is highly distributed in the liver and kidney, and GGT in the blood is mainly derived from the liver. When the liver is damaged (e.g. due to inflammation or cancer), the amount of GGT secreted by the liver can be significantly increased compared to normal conditions [29–32].

I-C-F-6 is an active oligopeptide originally identified from the traditional Chinese medicine turtle shell (Fig. S1, S2). We found that it significantly reduced the CYP2a1 gene expression and inhibited the metabolism of retinoic acid, thereby improving liver fibrosis and the balance of TH17/Treg cells. Related studies have also shown that it has good anti-hepatic fibrosis activity [33]. It can inhibit the activation of Wnt/β-catenin and NF-κB signaling pathways by reducing the expression of β-catenin and p65 proteins, thereby inhibiting the activation of HSCs. Moreover, I-C-F-6 could significantly inhibit the proliferation of hepatic stellate cells and promote their late apoptosis in vitro. Simultaneously, it has a significant therapeutic effect on alcoholic liver injury or acute liver toxicity caused by CCl₄ [34].

To overcome the limitations of DEX, including systemic adverse effects and low aqueous solubility, we engineered an amphiphilic prodrug (DEI) via Michael addition coupling DEX with I-C-F-6. The resulting prodrug self-assembles into water-soluble nanoparticles that via GGT-triggered self-immolation, thereby achieving targeted drug delivery to fibrotic livers and spatially confined release to mitigate systemic toxicity. Mechanistically, DEI potently attenuated the progression of CCl₄-induced hepatic fibrosis in mice by dual actions: inhibiting hepatic stellate cell activation by upstream inflammatory cells and immunometabolic reprogramming via retinol metabolism modulation (Cyp2a1↓/STAT1↑) to restore TH17/Treg equilibrium.

Results and discussion

Synthesis and characterization of GGT-triggered self-immolative prodrug DEI

γ-Glutamyltranspeptidase (GGT) exhibits substrate-specific hydrolysis activity toward γ-glutamylamides, with enzymatic efficiency significantly enhanced by the presence of an α-substituent on the amide moiety [35]. Capitalizing on this biochemical property, we designed and synthesized the amphiphilic prodrug DEI to construct GGT-responsive self-immolative nanoparticles (Fig. 1A). The synthesis involved liquid-phase preparation of DEXA (dexamethasone derivative) and solid-phase synthesis of ICNA (I-C-F-6 derivative), followed by covalent conjugation via Michael addition between DEXA’s vinyl group and ICNA’s cysteine thiol (Fig. S3-S5).

Fig. 1.

Synthesis and characterization of GGT-triggered self-immolative prodrug DEI. (A) Synthetic scheme of DEI. (B) FT-IR spectra of DEX, ICN and DEI. (C) HRMS spectrum of DEI. The molecular weight of DEI (m/z, [M + H]⁺) is 1230.5164, which agreed with the calculated value (m/z, [M + H]⁺, 1230.5153). (D) ¹H-NMR spectra of DEX, ICN and DEI (600 MHz, DMSO-d₆ as solvent)

In addition, comprehensive structural characterization confirmed DEI’s formation (Fig. S6, S7). FT-IR spectra showed the disappearance of the major hydroxyl peak (3464 cm⁻¹) of DEX (Fig. 1B, S5). HRMS results (Fig. 1C) confirmed the molecular weight of DEI as 1230.5164 (m/z, [M + H]⁺), closely matching the calculated value of 1230.5147 (m/z, [M + H]⁺). ¹H NMR analysis (Fig. 1D) revealed the disappearance of the 21-hydroxyl signal in DEX, along with a shift in the 21-methylene chemical signal from 4.05 to 4.53 ppm to 4.79–5.09 ppm, which can be attributed to the effects of ester bond formation.

Self-assembly and molecular dynamics simulation of DEI NPs

Amphiphilic DEI molecules spontaneously self-assemble into nanoparticles (NPs) in aqueous solution using the nanoprecipitation method [36, 37]. The yield of DEI NPs was 90.97%, nearly all initial raw materials were incorporated into the final structure, resulting in minimal waste during preparation. Cold field-emission SEM imaging demonstrated spherical nanoparticles with an average diameter of approximately 170 nm (Fig. 2A).

Fig. 2.

Self-assembly and Molecular Dynamics Simulation of DEI NPs. (A) SEM images of DEI NPs dispersed in DI water. (B) DLS and Zeta potential of DEI NPs dispersed in DI water. (C) Molecular Dynamics Simulation of DEI NPs. (D) Intermolecular interaction patterns of DEI NPs. (E) Electrostatic potential map of DEI NPs

Then we monitored the diameter and zeta potential of DEI NPs using Dynamic Light Scattering (DLS) over a week peroid. The results demonstrated that DEI NPs exhibited good stability in 7 days at 4℃. DLS data revealed an average hydrodynamic diameter of 156.62 ± 2.62 nm with a low polydispersity index (PDI = 0.211 ± 0.003), indicating uniform size distribution (Fig. 2B). Additionally, a zeta potential of +31.28 ± 1.50 mV confirmed favorable colloidal stability and minimized aggregation risks, which was consistent with the DLS data (Fig. 2B).

To understand DEI NPs’ surface electrostatic potential, we used GAUSSIAN 09 to assess individual molecules’ potentials [38]. Optimized with PM3, Fig. 2D showed the poly-peptide contributes to positive charges (red) and polar groups. DEX was nearly neutral and acted as the main non-polar component. These results suggest positive charges are on the cluster surface, giving NPs an overall positive potential. Within the cluster, polar moieties formed stable hydrogen bonds with water, while non-polar components curved inward, forming the core and enhancing particle-solution interface stability.

We conducted molecular dynamics (MD) simulations using GROMACS 2021.6, with systems parameterized by GAFF. The simulations showed that DEI molecules aggregated into stable clusters in water [39]. Figure 2E showed a defined aggregation pattern after 100 ns. The self-assembly process was tracked at 20 ns intervals. System stability was confirmed by an RMSD value of 5.68829 ± 1.176 Ų (Fig. S8A), and SASA measurements showed structural compaction. The average SASA value reduced significantly from 119.42 ± 0.57 nm² to 64.35 ± 0.54 nm² (Fig. S8B), stabilizing after 60 ns, indicating a compact structure. These results indicated DEI’s ability to self-assemble into defined nanostructures in water.

DEI molecules arranged orderly, with hydrophobic groups (blue) in the core and hydrophilic groups (red) on the surface (Fig. 2C). MD simulations revealed hydrogen bonding networks among poly-peptide groups as the main driving force for self-assembly (Fig. 2D, S8C), explaining the nanoparticles stability and organization [40]. Coul-SR interactions averaged −24074.34 ± 114.73 kJ·mol⁻¹, LJ-SR interactions averaged −3056.95 ± 70.25 (Fig. S8D), indicating strong electrostatic stabilization.

GGT-triggered self-immolation of DEI NPs

To validate the GGT-responsive drug release mechanism, DEI NPs (1 mg·ml^− 1) were incubated with GGT (0–10 U·mL⁻¹) in PBS (37 °C) and analyzed via UPLC-MS at timed intervals (0–24 h). The hydrolysis kinetics demonstrated strict enzyme-dependence: complete drug release occurred within 2 h at 10 U·mL⁻¹ GGT, which showed that DEI had good responsiveness to GGT. At lower concentrations, 0.5 U·mL⁻¹ GGT resulted in a gradual release of 58.71% at 8 h, simulation a liver fibrosis microenvironment with elevated GGT expression; while 0.05 U·mL⁻¹ GGT led to a small release of 17.55% over r the same period, mimicking a non-hepatic fibrosis environment with minimal GGT activity. In the absence of GGT (0 U·mL⁻¹ GGT) showed negligible hydrolysis (< 5% at 8 h), confirming enzymatic specificity (Fig. 3A). In the plasma stability test, given the presence of trace levels of GGT in plasma, DEI NPs were slowly released, with less than 20% released at 8 h and 35% released within 72 h (Fig. 3B). To examine the in vivo metabolic profile of DEI NPs, Wistar rats (control vs. model) were administered DEI NPs via intraperitoneal injection at a dose of 11.08 mg·kg⁻¹. Blood samples were serially collected via the orbital venous plexus at predetermined time intervals (0–72 h), and the plasma concentration of DEI was dynamically monitored using HPLC. As shown in Fig. 3B and C, the metabolic rate of DEI NPs in the model group (t_1/2 = 2.21 h) was significantly higher than that in the normal group (t_1/2 = 6.26 h). The results indicated that DEI was completely cleared from plasma within 8 h in the model group (Fig. 3D), which was consistent with in vitro release behavior. Compared to the normal group, the lower C_max, AUC_(0−t) in the model group resulted from accelerated metabolism rather than absorption limitation (Fig. S14). The drug release profile exhibited the highest correlation with the first-order model (R² = 0.9866), suggesting itsrelease behavior follows first-order kinetics. The above data indicated that DEI NPs are stable during circulation and undergo rapid, site‑specific activation only after extravasation into the GGT‑rich fibrotic liver tissue, thereby avoiding off-target systemic effects.

Fig. 3.

GGT-triggered self-immolation of DEI NPs. (A) Drug hydrolysis rate of DEI NPs at different GGT enzyme concentrations in PBS (37 °C). (B) Drug release rate of DEI NPs in plasma at 37℃. (C) The drug time curve of DEI NPs in vivo. (D) Metabolism of DEI NPs and their hydrolysis products ICN and DEX in vivo in the model group. (E) HRMS spectrum of GGT-triggered self-immolative fragments of DEI. (F) Structure of DEI and its reaction with GGT

HRMS analysis after 1 h incubation identified key intermediates and products, the γ-glutamyl cleavage intermediate (observed m/z 1101.4729 [M + H]⁺ vs. calculated 1101.4721), liberated DEX (observed m/z 393.2072 vs. theoretical 393.1999), and ICN (observed m/z 727.2839 vs. expected 727.2828) (Fig. 3E), unequivocally validating the self-immolative cascade depicted in Fig. 3F. According to the cleavage mechanism, glutamic acid was expected as the hydrolysis byproduct. As afundamental amino acid in biological metabolism, it has demonstrated excellent safety [41, 42].

DEI NPs mitigate TGF-β1-induced liver fibrosis in vitro

The anti-fibrotic activity of DEI was validated at the cellular level using a TGF-β1-stimulated JS-1 cell model. Firstly, we assessed the cytotoxicity of I-C-F-6, ICN and DEI NPs in JS-1 cell. The IC₅₀ values of I-C-F-6, ICN, and DEI NPs were approximately 171, 185, and 160 µM, respectively (Fig. 4A), showing they all exhibit good safety in vitro. At a concentration of 50 µM, all drugs exhibited no cytotoxicity to resting JS-1 cells but significantly inhibited activated JS-1 cells.

Fig. 4.

DEI NPs mitigate TGF-β1-induced liver fibrosis in vitro. (A) The CCK8 assay curves of I-C-F-6, ICN, and DEI NPs against JS-1 cells at different concentrations. (B) The cytotoxicity of DEX, I-C-F-6, ICN, and DEI NPs at a concentration of 50 µM on resting (left) and activated (right) JS-1 cells, respectively. (C) The mRNA expression of fibrotic genes COL I and α-SMA in different groups. (D) Protein expression of α-SMA and COL I in different groups. (E) The mRNA expression levels of Cyp2a1, RORγt, RORα, IL17ra, IL17rc, and STAT1 genes. ^###p < 0.001, ^nsp > 0.05 compared to control group. *p < 0.05, *p < 0.05, **p < 0.01, ***p < 0.001 compared to model group. ^☆p < 0.05, ^☆☆p < 0.01, compared to the DEI NPs group

Therefore, the efficacy of DEI NPs was systematically assessed and compared against that of DEX, I-C-F-6, ICN and the mixture (DEX + I-C-F-6). Quantitative analysis through qPCR and western blotting further confirmed DEI’s potent antifibrotic effects through the marked reduction in the expression levels of HSCs activation markers α-SMA and COL I. Notably, DEI NPs outperformed both DEX and I-C-F-6 alone, as well as their physical mixture, in attenuating fibrotic activation, demonstrating the advantages of the nano-prodrug in targeting capability, aqueous solubility, and synergistic efficacy (Fig. 4C and D). Additionally, mRNA expression analysis further showed that DEI NPs downregulated key fibro-inflammatory mediators (Cyp2a1, RORγt, RORα, IL17ra, and IL17rc) while upregulating STAT1 expression (Fig. 4E). Meanwhile, all data revealed that ICN, which is an I-C-F-6 derivative released from GGT-triggered self-immolative of DEI, exhibited similar cytotoxicity and anti-fibrotic effects to I-C-F-6.

Based on these findings, we propose that DEI suppressed the activation and proliferation of HSCs through two primary mechanisms: firstly, by inhibiting the stimulatory and pro-inflammatory signaling effects of infiltrating inflammatory cells; secondly, by modulating retinoic acid metabolism to exert therapeutic effects on ameliorating liver fibrosis and modulating immune responses.

DEI NPs alleviates the CCl₄-induced liver fibrosis and injury in vivo

To systematically evaluate the anti-fibrotic efficacy and mechanisms of DEI nanoparticles (DEI NPs), we established a CCl₄-induced murine liver fibrosis model via 8-week oral administration CCl₄ (n = 8/group), with therapeutic interventions administered from weeks 5 to 8 (Fig. 5A). The DEI NPs groups were given intraperitoneal injections of DEI NPs at three doses (4, 8, and 16 mg·kg⁻¹). The positive drug control groups received intraperitoneal injections of DEX (2.55 mg·kg⁻¹) and I-C-F-6 (3.58 mg·kg⁻¹), equivalent to a molar dose of 8 mg·kg⁻¹ of DEI NPs. Model group livers exhibited gross pathological hallmarks of fibrosis, including surface granularity, edge blunting, and dark red discoloration (Fig. 5B), accompanied by elevated liver weight index (Fig. 5C) and serum ALT/AST/LDH levels (Fig. 5D), confirming hepatic injury. Histopathological analysis via HE staining (Fig. 5E) demonstrated typical fibrotic features in model animals, including hepatocyte vacuolization, inflammatory infiltration. All DEI NPs treatment groups showed dose-dependent attenuation of fibrotic damage, with superior efficacy versus equimolar DEX and I-C-F-6.

Fig. 5.

DEI NPs alleviates the liver damage in CCl₄-induced fibrosis in mice. (A) Schedule of DEI NPs treatment in fibrotic mice induced by CCl₄. (B) Mouse liver appearance. (C) Ratio of liver and body weight. (D) Serum levels of ALT, AST, LDH. (E) Representative images and quantitative statistics of HE-stained liver sections. ^###p < 0.001 compared to control group. *p < 0.05, *p < 0.05, **p < 0.01, ***p < 0.001 compared to model group. ^☆p < 0.05, ^☆☆p < 0.01, compared to the DEI NPs group

The Masson staining results clearly demonstrated that DEI NPs effectively mitigated CCl₄-induced fibrotic responses, as evidenced by a significant reduction in collagen deposition within liver tissue. The results of collagen deposition quantitative analysis using Image J showed that DEI NPs group was significantly better than the model group, DEX group and I-C-F-6 group (Fig. 6A). Quantitative analysis through qPCR and western blot further confirmed DEI’s potent antifibrotic effects, showing marked decreases in the expression levels of HSCs activation markers α-SMA and COL I. In Fig. 6B, qPCR data indicated that DEI (8 mg·kg⁻¹) decreased the mRNA expression of α-SMA to 46.6 ± 4.7% of that in the model group, outperforming the DEX group (63.6 ± 5.3%) and the I-C-F-6 group (66.5 ± 4.1%). Meanwhile, the mRNA expression of COL I was reduced to 36.6 ± 4.7% of that in the model group, compared with the DEX group (51.4 ± 5.3%) and the I-C-F-6 group (55.3 ± 3.4%). Western blot quantification further confirmed that 8 mg·kg⁻¹ of DEI treatment reduced COL I/GADPH protein levels to 1.85 (vs. 4.73 in the model group) and α - SMA/GADPH protein levels to 1.86 (vs. 3.62 in the model group). All differences were statistically significant (p < 0.01), and the results were also superior to those of the DEX group and the I-C-F-6 group (Fig. 6C).

Fig. 6.

DEI NPs alleviates the liver damage in CCl₄-induced fibrosis in mice (n = 8). (A) Representative images and quantitative statistics of liver sections stained with Masson. (B) Hepatic mRNA expression of fibrotic genes COL I and α-SMA. (C) Protein expression of α-SMA and COL I in liver tissues. ^###p < 0.001 compared to control group. *p < 0.05, *p < 0.05, **p < 0.01, ***p < 0.001 compared to model group. ^☆p < 0.05, ^☆☆p < 0.01, compared to the DEI NPs group

Immunofluorescence co-localization analysis yielded results consistent with the aforementioned findings. Quantification of the COL I/DAPI ratio revealed a marked increase in the model group (86.94%) compared to the control group (6.05%). Treatment with DEX reduced this ratio to 39.11%, while DEI treatment at doses of 4, 8, and 16 mg·kg⁻¹ led to further reductions to 40.56%, 32.73%, and 12.59%, respectively. A similar trend was observed for the α-SMA/DAPI ratio, which was significantly elevated in the model group (7.77%) versus the control (1.98%). DEX treatment lowered the ratio to 3.18%, and DEI at 4, 8, and 16 mg·kg⁻¹ progressively reduced it to 3.05%, 2.25%, and 2.05%, respectively (Fig. 7A and B).

Fig. 7.

DEI NPs alleviates the liver damage in CCl₄-induced fibrosis in mice. (A) Immunofluorescence staining of COL I (yellow) and α-SMA (red), nuclei counterstained with DAPI (blue) in liver. (B) The immunofluorescence quantification of COL I/DAPI and α-SMA/DAPI protein levels. (C) Thighbone bone mineral density changes of mice (n = 8). ^###p < 0.001 compared to control group. ^☆p < 0.05, ^☆☆p < 0.01, compared to the DEX group

To compare bone mass across experimental groups, bilateral femora and tibiae were harvested from the mice, and bone mineral density (BMD) was measured using the Kubtec Parameter 3D imaging system. Compared with the control group, the DEX group exhibited a significant reduction in bone mineral density (BMD) by 15.68%. In contrast, the decreases observed in the low-, medium- and high-dose, DEI groups were markedly less severe decreases of only 2.86%, 5.86%, and 7.77%, respectively (Fig 0.7 C).

These data collectively demonstrated that at equimolar doses (8 mg·kg⁻¹ DEI vs. 2.55 mg·kg⁻¹ DEX), DEI NPs exhibited approximately 2.36–2.86 multiple greater anti-fibrotic efficacy compared to conventional DEX therapy. Crucially, DEI NPs circumvented DEX-associated osteoporosis and showed no significant organ toxicity (Fig. S9–S13), demonstrating both therapeutic superiority and safety advantages attributable to targeted delivery.

DEI NPs exhibited enhanced hepatic targeting in fibrotic mice

Given the potent in vivo anti-fibrotic efficacy of the NPs, we next explored their biodistribution using DiR-labeled DEI NPs. Vascular remodeling occurs in the inflammatory tissue areas of liver fibrosis, with the enhanced permeability and retention (EPR) effect [43–46] similar to the solid tumor. The fluorescence imaging signals revealed distinct targeting capability and pharmacokinetic profiles in both healthy and fibrotic mice.

Quantitative fluorescence imaging demonstrated that DiR-DEI NPs achieved hepatic accumulation at 0.5 h post - injection, reaching a peak value at 8 h, with significantly greater intensity in fibrotic livers (1.13 × 10⁷ ± 1.67 × 10⁶ AU) compared to healthy controls (7.61 × 10⁶ ± 2.45 × 10⁵ AU, p < 0.01). This represented a 48.5% enhancement in hepatic targeting efficiency in disease models. In stark contrast, free DiR showed minimal liver retention (fibrotic: 5.93 × 10⁶ ± 5.56 × следующих 10⁵ AU; healthy: 5.42 × 10⁶ ± 5.56 × 10⁵ AU), with DEI NPs exhibiting 2.08-fold higher hepatic accumulation in fibrotic mice (p < 0.001) (Fig. 8A and B). As shown in Fig. 8C and D, the fibrotic model group exhibited significantly higher hepatic accumulation of DEI NPs (8.09 × 10⁶ ± 2.45 × 10⁵ AU) compared to healthy controls (4.05 × 10⁶ ± 4.87 × 10⁵ AU, p < 0.001), while free DiR showed minimal liver retention in both groups (fibrotic: 2.28 × 10⁶ ± 2.33 × 10⁵ AU; healthy: 2.58 × 10⁶ ± 2.43 × 10⁵ AU). Pulmonary distribution of DEI NPs was moderately lower in fibrotic models (7.46 × 10⁵ ± 7.80 × 10⁴ AU) compared to healthy mice (1.14 × 10⁶ ± 1.39 × 10⁵AU, p < 0.05), contrasting with free DiR’s stable lung deposition (1.65 × 10⁶ ± 1.47 × 10⁵ AU).

Fig. 8.

The biodistribution of DEI NPs in mice (n = 3). (A) In vivo fluorescence imaging of free-DiR and DiR-DEI NPs in control and model group mice. (B) Quantitative analysis of hepatic zones fluorescence intensity. ^#p < 0.05 ^##p < 0.01 compared to free DiR in control group. *p < 0.05, **p < 0.01, ***p < 0.001, compared to free DiR in model group. (C) Representative ex vivo organ fluorescence distribution images at 8 h post-injection. (D) Quantitative analysis of representative ex vivo organ fluorescence distribution at 8 h post-injection. *p < 0.05, ***p < 0.001, compared to free DiR in control group

Together, these data confirmed the excellent fibrotic liver-targeting and retention properties of DEI NPs, which delivered most of their payload to the hepatic fibrosis site within 0.5 h. This targeted enrichment maximizes therapeutic potential at the site of action and concurrently limits systemic Dex exposure, providing a dual benefit of enhanced efficacy and reduced toxicity.

DEI NPs attenuated liver fibrosis progression through concurrent modulation of the TGF-β/Smad signaling axis, retinol metabolism-mediated immunoregulation, and Th17/Treg differentiation balance

To further investigate the potential molecular mechanisms the effects of DEI NPs, RNA sequencing analysis was performed on liver tissue. A total of 56,980 genes were matched from Ensemble: GRCm39, a Mus musculus genome. Differential gene analysis between groups was performed using DESeq2 with a screening threshold of p < 0.05, fold change > 1.2. As shown in Fig. 9A, there were 2,134 differentially expressed genes (DEGs) between the control and model groups, with 1,176 genes upregulated and 958 genes downregulated. Compared to the model group, the DEI NPs group showed 1,315 DEGs, with 781 genes upregulated and 534 genes downregulated. The Principal Component Analysis (PCA) followed by ANOSIM analysis confirmed significant differences in gene expression levels among the three groups (Fig. 9B). The Venn diagram showed that 75 genes are shared by the control, model, and DEI NPs groups (Fig. 9C). The overlapping genes may represent key regulators in the fibrosis process that are restored by DEI NPs treatment.

Fig. 9.

RNA expression profile of livers in different groups of mice. (A) Volcano plots of DEGs between the control vs. model and model vs. DEI groups, respectively. (B) Principal component analysis of the three groups. (C) The Venn diagram of the three groups. (D) KEGG analysis of gene enrichment in the signaling pathway. (E) Heatmap of the expression levels in different samples

To identify the signaling pathways affected by DEI NPs, the KEGG analysis was conducted. The differentially expressed genes were mainly enriched in the retinol metabolism pathway, Th1 and Th2 cell differentiation pathway, and Th17 cell differentiation pathway (Fig. 9D). Hierarchical clustering analysis showed that there were significant differences in gene expression profiles between DEI NPs group and model group, but its profile was closer to control group (Fig. 9E).

To further verify the dual-pathway mechanism of DEI NPs in inhibiting fibrogenesis: first, they block HSCs activation by suppressing the TGF - β/Smad3 pathway; second, they interrupt the TH17 pathway to reduce collagen I/IV deposition. Meanwhile, DEI NPs achieve immunometabolic reprogramming via retinol metabolism modulation (evidenced by Cyp2a1↓/STAT1↑), which helps restore the TH17/Treg equilibrium. We then evaluated hepatic inflammatory responses. Levels of pro-inflammatory cytokines (TNF-α, IL-6, and IL-17) were markedly increased in the fibrosis model groups compared with normal controls (Fig. 10A). Therapeutic intervention with DEI NPs demonstrated superior anti-inflammatory efficacy versus DEX or I-C-F-6 monotherapy, as evidenced by marked reductions in these cytokines. And mRNA expression analysis further showed that DEI NPs downregulated key fibro-inflammatory mediators (Cyp2a1, RORγt, RORα, IL17ra, and IL17rc) while upregulating STAT1 expression (Fig. 10B).

Fig. 10.

DEI NPs attenuate hepatic fibrosis via cytokine suppression and immunometabolic reprogramming. (A) Expression levels of TNF-α, IL-6, and IL-17 in mouse liver. (B) mRNA expression levels of Cyp2a1, RORγt, RORα, IL17ra, IL17rc, and STAT1 genes in mouse liver. (C) Immunofluorescence staining of TH17 (red), Treg (green), M1 (fuchsia), and nuclei counterstained with DAPI (blue) in liver. ^###p < 0.001 compared to control group. *p < 0.05, *p < 0.05, **p < 0.01, ***p < 0.001 compared to model group. ^☆p < 0.05, ^☆☆p < 0.01 compared to the DEI NPs group

To further demonstrate the role of DEI in regulating the immune micro-environment, we employed immunofluorescence staining to investigate the infiltration of immune cell populations, including Treg, Th17, and M1-macrophage, in hepatic fibrosis lesions of mice. TH17/Treg equilibrium. The model group showed increased Th17/M1-macrophage and fewer Treg cells, an imbalance that was reversed by DEI NPs treatment (Fig. 10C). This restoration of the TH17/Treg ratio corroborates the qPCR results and underscores DEI’s role in modulating the fibrotic immune milieu. In summary, DEI NPs attenuate hepatic fibrosis via cytokine suppression and immunometabolic reprogramming.

Conclusion

In this study, we developed a GGT-responsive self-immolative prodrug DEI for targeted liver fibrosis therapy, featuring an amphiphilic structure that self-assembles into stable nanospheres. The response of GGT enables the synchronous release of DEX and ICN, allowing both to exert their pharmacological effects concurrently and achieve pharmacological synergy. DEI NPs exhibit the EPR effect, which can prolong local retention time. Simultaneously, their positive zeta potential further enhances interactions with the negatively charged cell membrane, which further facilitates an “anchoring” effect that promotes cellular binding and uptake. These integrated properties—tissue-specific targeting, enzyme-triggered drug release, prolonged retention, and enhanced membrane interactio—collectively demonstrate that the system exhibits superior efficacy in combating hepatic fibrosis at the tissue and cellular levels and overcoming DEX’s limitations of poor targeting and systemic toxicity.

Molecular dynamics simulations confirmed the nanostructure’s core-shell architecture, with hydrophobic DEX prodrug fragments sequestered internally and hydrophilic I-C-F-6 derivatives exposed for GGT recognition. DEI NPs demonstrated excellent stability in the plasma, upon reaching fibrotic liver regions, it underwent enzymatic cleavage in perivascular areas, releasing active DEX and ICN. At the cellular level or in CCl₄-induced fibrosis models, DEI NPs exhibited dual therapeutic mechanisms. Firstly they inhibited upstream inflammatory cytokines through the classic anti-inflammatory pathway of DEX, thereby reducing the production of TNF-α, IL-6, IL-17 and subsequent HSCs activation. Secondly, DEI NPs suppressd HSCs activation by downregulating Cyp2a1-mediated retinoic acid metabolism. Concurrently they modulated the TH17/Treg balance through coordinated inhibition of RORγt/RORα/IL-17ra/IL-17rc and upregulation of STAT1. This synergistically improved liver function, attenuated fibrogenesis, and avoided DEX-associated osteoporosis/immune dysregulation. The GGT-triggered self-immolative nanoplatform demonstrated superior efficacy and safety over conventional DEX, establishing a paradigm for enzyme-responsive combination therapy in fibrotic diseases.

Materials and methods

Synthesis and Characterization of GGT-triggered self-immolative prodrug DEI

DEX-acryloyl chloride derivatives were synthesized via liquid-phase synthesis and purification. Dexamethasone (DEX, 100 mg, 0.25 mM) and triethylamine (26 mg, 0.25 mM) were dissolved in tetrahydrofuran (THF) under ice-bath conditions. Acryloyl chloride (23 mg, 0.25 mM) was added drop-wise with stirring for 3 min (reaction completion monitored by TLC). The reaction mixture was concentrated in vacuo, redissolved in DCM (40 mL), and transferred to a separatory funnel for sequential washing with deionized water (2X) and saturated brine (1X). The organic layer was dried over anhydrous Na₂SO₄, concentrated, and purified by silica gel chromatography (petroleum ether/ethyl acetate 15:1, v/v) to afford DEXA as a white solid.

I-C-F-6 and its-derivatives ICNA (cysteine-glutamic acidconjugate) were synthesized by GL Biochem (Shanghai) Ltd. via solid-phase synthesis and and subsequently purified by HPLC to yield a 98% product. ICN for cell experiments was obtained by enzymatic hydrolysis of DEI with GGT. DEXA (200 mg, 0.438 mmol in methanol:water=3:1, 500 µL) and ICNA (100 mg in ultrapure water, 500 µL; added pyridine, 500 μL) were combined in a 5 mL flask and stirred for 24 h (TLC-monitored). The mixture was lyophilized after rotary evaporation and resolubilization in water, then purified by preparative HPLC (Fig. S13) to obtain DEI.

FT-IR spectroscopic analysis

The FT-IR spectra of DEX, ICNA, DEXA and DEI were recorded by FT-IR spectrometer (Alpha II, Bruker, Germany) at 4000–400 cm^− 1 with an average scan of 200 times.

Preparation of DEI NPs by anti-solvent method

Briefly, 5 mg of DEI powder was dissolved in a mixture of methanol (MeOH) and tetrahydrofuran (THF) (1:3 ratio, 500 µL). Then, 5 mL of deionized water was added dropwise to the drug conjugate solution while stirring gently for 2 h at 25 °C. Subsequently, the dispersion was centrifuged at 14,000 rpm for 5 min, washed with deionized water, and centrifuged again. This process was repeated two more times to eliminate any remaining organic reagent.

The carrier-free DEI NPs are composed solely of self-assembled DEI single-molecules without an external matrix, making the concepts of EE% and DL% mathematically inapplicable. The yield of DEI NPs serves as the appropriate analogous metric [47, 48].

Yield % = (Mass of lyophilized DEI NPs/Mass of initial DEI) x 100.

SEM analysis of DEI NPs

2.5 µL of DEI NPs aqueous solutions (1 mg·mL^− 1 and 0.1 mg·mL^− 1) were applied onto silicon wafers. The samples were vacuumdried at room temperature for 4 h. Subsequently, the dried samples were gold-sprayed and examined using a scanning electron microscope (SEM; HITACHI-Regulus 8100, Japan) at a working voltage of 3 kV.

Morphological characterization of DEI NPs

Dynamic light scattering (DLS) was used to determine the particle size distribution and charge properties of the NPs. The particle size and zeta potential was measured using Malvern Zetasizer Nano-ZS (Zetasize Nano ZS 90, Malvern Instruments, UK).

Molecular dynamic simulation

The DEI molecule was constructed by Chem3D, and then the 6-polymerization of DEI was performed using PACKMOL 18.169. All-atom molecular dynamics simulation of 6-mer as the initial structure using AMBER 18 [49]. The detailed steps were available in the Supporting Information.

Cell culture

JS-1 cells were obtained from the Chinese Academy of Medical Sciences & Peking Union Medical College and cultured in DMEM (Thermo Fisher Scientific) containing 10% (vol·vol^− 1) dialyzed FBS, 100 U·mL^− 1 penicillin and 100 mg·mL^− 1 streptomycin.

The cytotoxicity of these compounds was tested on JS-1 and in vitro by the CCK-8 assay. In short, 100 µL of cells with a density of 5 × 10⁴ cells·mL^− 1 were added to per well of 96-well plates, and incubated for 24 h at 37℃ with 5% CO₂. Then the cells were exposed to various concentrations of the tested compounds and incubated for 72 h. The CCK-8 solution (10 µL) was added to each well. After 0.5 h of incubation, the absorbance was measured at a wavelength of 450 nm. Wells without reagent were used as the blanks. The Cell Counting Kit-8 was available from MeilunBio. The IC₅₀ values were defined as the concentration of compounds that produced a 50% proliferation inhibition of surviving cells and calculated with the GraphPad Prism 8. The inhibitory rate was calculated in the following equation:

% Inhibition = 100% × [1 - (OD_{Sample group} - OD_{Blank group})/(OD_{Control group} - OD_{Blank group}).

Animal studies

C57BL/6J mice (8 weeks old, male, 18–22 g) were purchased from SIBEIFU Biotechnology Co, Ltd. (Beijing, China). Mice had free access to normal chow diet and sterile water. Mice were randomly divided into six groups (n = 8). They were given olive oil (Control), CCl₄ (Model), CCl₄ + DEX (2.55 mg·kg^− 1), CCl₄ + I-C-F-6 (3.58 mg·kg^− 1), CCl₄ + DEI (4 mg·kg^− 1), CCl₄ + DEI (8 mg·kg^− 1) and CCl₄ + DEI (16 mg·kg^− 1), respectively. CCl₄ was dissolved in olive oil (13% v·v^− 1, 1 mL·kg^− 1). In the CCl₄-induced model, mice received intragastric administration of CCl₄ twice weekly for eight weeks. DEX, I-C-F-6 and DEI NPs were administered daily via intraperitoneal (i.p.) injection from week 4 to week 8. At the conclusion of the experiment, mice were anesthetized using isoflurane and subsequently euthanized for the collection of blood, liver, and other major organs. All animal experiments and procedures were approved by the Institutional Animal Care and Use Committee of Beijing University of Chinese Medicine and conducted in compliance with relevant guidelines and regulations.

qPCR assay

Total RNA of liver and cells was extracted and cDNA was synthesized. The sequence of primers was designed as Table S1. The results were analyzed by a relative quantitative 2^-ΔΔCt method, and then the expression of mRNA in each group was calculated and compared. The amounts of transcript were normalized to those for glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Primer sequences were presented in Table S1.

Western blotting

Cellular and tissue proteins were extracted by RIPA lysis, and protein quantification was performed using the BCA method. Then 10 µg of denatured proteins were loaded into Tris-Gly gels for electrophoresis. The gel was then transferred to a PVDF membrane, closed with 5% BSA, and incubated with primary and secondary antibodies. Finally, the bands were visualized using ECL chemiluminescent reagents. Image Lab software was used for quantitative analysis. Antibodies aganst α-SMA (80008-1-RR), Collagen I (14695-1-AP), and GAPDH (60004-1-lg) were obtain from Proteintech Group, Inc. (China).

DiR-labeled DEI NPs preparation

First, 10.0 mg DEI was weighed and dissolved into 5.0 mL deionized (DI) water. Then 1 mL DiR dichloromethane solution was added (2.0 mg·mL^− 1) dropwise to the DEI solution. Immediately after the addition, the mixture was homogenized for 0.5 h using a high-speed tissue homogenizer until a homogeneous solution was obtained. The resulting product was lyophilized to obtain the DiR-loaded DEI powder. For the subsequent experiment, an appropriate amount of this sample was weighed and dissolved in DI water to prepare a DiR-loaded DEI solution containing 0.1 mg·mL^− 1 DiR. A free DiR aqueous solution at the same concentration (0.1 mg·mL^− 1) was prepared on the spot.

RNA sequencing analysis of liver

Liver tissues of control, model, and DEI groups were collected for RNA-seq and qPCR analysis. The detailed steps were available in the Supporting Information.

Statistical analysis

All data were expressed as mean ± SD and analyzed using the Graphpad Prism (version 9.0). The significance was tested with one-way analysis of variance (ANOVA) among three or more groups. Statistical significance was set at p < 0.05, 0.01 and 0.001.

Author contributions

Z.Z. was responsible for conceptualization, methodology, validation, visualization, and writing original draft. J.A. was responsible for investigation and data curation. ZJ.Z. was responsible for investigation and software. F.G. was responsible for writing – review & editing. DL.L., J.W., XR.H and KX.K. was responsible for investigation. YJ.L. was responsible for methodology. XC.D. was responsible for project administration. B.X. was responsible for project administration, funding acquisition, and supervision. HM.L. was responsible for resources, supervision, and funding acquisition. All authors reviewed the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China (No. 82104365, 82274082, 82403682), Beijing “high-grade, precision and advanced” project, Beijing Key Laboratory for Basic and Development Research on Chinese Medicine (Beijing, 100102).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The animal study proposal was approved by the ethics committee of Beijing University of Chinese Medicine with the permit number 1121043000408. All animal experimental procedures were performed in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals approved by the State Council of People’s Republic of China.

Competing interests

The authors declare no competing interests.