Tissue-derived extracellular vesicles–mediated delivery of a hepatic enzyme living panorama (HELP) for treating multifactorial liver diseases

Abstract

Background

Dysregulation of hepatic drug-metabolizing enzymes is a hallmark of liver injury and contributes to treatment failure. Current strategies often overlook the complexity of the integrated enzyme network. This study reports the use of tissue-derived extracellular vesicles (TEV) as a novel bioinspired delivery platform for a functional hepatic enzyme living panorama (HELP)—a synergistic ensemble of key drug-metabolizing enzymes including cytochrome P450 (CYP450), UDP-glucuronosyltransferase (UGT), sulfotransferase (SULT), and glutathione S-transferase (GST) families.

Results

We isolated TEV from liver tissue and demonstrated their enrichment in functional cytochrome P450s, UGTs, SULTs, and GSTs via proteomics. TEV exhibited robust metabolic activity in vitro and ex vivo. In murine models of acute liver failure, alcoholic liver disease, and metabolic dysfunction-associated steatotic liver disease, TEV administration improved survival, attenuated liver injury, and restored enzyme function without altering host gene expression—confirming direct enzyme delivery. TEV also synergized with conventional therapeutics (regorafenib, silymarin), enhancing efficacy while reducing toxicity. Notably, TEV conferred protection in vulnerable populations with inherent enzymatic deficiencies (neonates, aged mice).

Conclusions

TEV serve as an effective platform for functional enzyme delivery, restoring hepatic metabolic capacity across diverse injury models and special populations. This study introduces HELP as a novel therapeutic strategy for liver diseases characterized by enzymatic dysfunction.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04198-4.

Introduction

Drug metabolism is largely carried out by the cytochrome p450 (CYP) enzyme family, which metabolizes more than 75% of marketed drugs [1, 2]. According to the conventional perspective, the overall activity of the drug-metabolizing system can be represented as the simple sum of the contributions of individual p450 enzymes [3, 4]. However, accumulating evidence suggests that this view represents an oversimplification. Interindividual variation in drug metabolism is substantial and cannot be fully accounted for by genetic polymorphisms or differences in enzyme expression alone [1, 5]. Instead, inter p450 protein–protein interactions are increasingly recognized as critical determinants of metabolic function. Numerous studies have shown that physical interactions among p450s—including the formation of hetero-oligomeric complexes—significantly alter their catalytic behavior [6, 7]. For example, CYP2E1 can activate CYP3A4 and CYP2C19, often at the expense of its own activity, which may be suppressed or unchanged [8]. Moreover, p450 enzymes engage in physical interactions with other key components of the drug disposition cascade, such as microsomal epoxide hydrolase and UDP-glucuronosyltransferases (UGTs) [9, 10]. In light of these complex interactions, we propose the concept of the “hepatic enzyme living panorama " (HELP). HELP refers to a functional and co-enriched ensemble of hepatic metabolic enzymes—predominantly from the CYP450, UGT, sulfotransferase (SULT), and glutathione S-transferase (GST) families—that operates as a highly integrated, non-additive system.

The hepatic metabolic system relies critically on specialized enzyme families, notably p450, UGT, SULT and GST, which collectively govern drug metabolism, detoxification processes, and metabolic homeostasis [11, 12]. Dysregulation of these enzymes is a hallmark of various liver pathologies, such as metabolic dysfunction-associated steatotic liver disease (MASLD) and drug-induced liver injury (DILI) [13, 14]. Despite advances in pharmacotherapy, many patients—especially those with compromised metabolic capacity, such as pediatric or geriatric populations—remain vulnerable to toxicity and poor treatment outcomes [15–17]. In this context, extracellular vesicles (EVs) have recently emerged as promising natural nanocarriers for therapeutic delivery because of their biocompatibility, tissue tropism, and ability to shuttle functional biomolecules [18, 19]. However, their potential as carriers of active metabolic enzymes has remained largely unexplored.

Extracellular vesicles have emerged as promising therapeutic nanocarriers due to their biocompatibility and inherent tropism [20, 21]. However, their application has largely been confined to delivering nucleic acids or signaling molecules; the efficient packaging and targeted delivery of multiple, functional metabolic enzymes—a prerequisite for treating complex hepatic disorders—remains a significant, unmet challenge [22]. This limitation mirrors the broader shortfall of conventional approaches: enzyme replacement therapies (ERTs) are intrinsically single-agent solutions ill-suited for diseases with network-wide metabolic dysfunction [23], while gene therapies struggle with precise, multi-gene orchestration and immunogenic risks [24]. Critically, both fail to recapitulate the native interactome and stoichiometry of hepatic enzymes [5]. We hypothesize that tissue-derived EVs (TEV), which naturally encapsulate a broad spectrum of cognate enzymes from their cells of origin, could overcome these hurdles [25]. By delivering this intact functional ensemble—termed the HELP—TEVs offer a bioinspired strategy to simultaneously restore multiple pathways, presenting a holistic therapeutic paradigm for multifactorial liver diseases.

Here, we demonstrate that TEV is enriched in HELP and display robust metabolic activity in vitro and in vivo. TEV restores liver enzyme function, increases survival, and synergizes with standard therapies, boosting efficacy and reducing toxicity across multiple murine models. TEV also benefit vulnerable groups—neonates and aged mice—by reducing injury and enhancing metabolic resilience without changing gene expression, indicating direct enzyme delivery. This work introduces HELP as a new strategy for recovering hepatic metabolic function and proposes novel therapeutic opportunities in hepatology.

Methods

Biospecimen collection

Liver biopsy samples were collected from West China Hospital of Sichuan University, Sichuan Province, China. Each liver tissue sample was carefully preserved in a 1.5 mL tube and stored at − 80 °C. Adhering strictly to the ethical guidelines outlined in the Declaration of Helsinki, this study received full endorsement from the Ethics Committee of West China Hospital of Sichuan University (Approval No. WCH20180037). Furthermore, to ensure transparency and respect for participant rights, written informed consent was obtained from all individuals involved in the study.

EVs extraction from liver tissue and hepatocytes

Liver-derived EVs were obtained from adult male mice. Specifically, adult mouses were euthanized separately, and fresh liver tissues were harvested and cut into small pieces (1–2 mm) in a sterile low-temperature environment. The tissue fragments were placed in DMEM containing collagenase II (1 mg/mL) and incubated at 37 °C for 2 h. After the tissues were dissociated, the solution was passed through a 70 μm filter to remove large tissue fragments. Cell-derived EVs were obtained from AML-12. The resulting filtrate was centrifuged at 300 × g for 10 min and then at 2000 × g for 20 min to remove cells and cell debris. The supernatant was further centrifuged at 118,000 × g for 2.5 h to obtain EVs. All centrifugation steps were performed at 4 °C.

Enzyme kinetic assay

Activities of alcohol metabolic enzymes in ethanol treated EVs were measured using ADH specific commercial assay kits. EVs were incubated with isopropanol, which results in conversion of NAD⁺ to NADH to generate a colorimetric (450 nm) product proportional to ADH activity (Sigma, MAK053). The metabolism of specific probe by EVs were performed using IPHASE CYP450 Metabolic Phenotyping Kit (Cat#: 0114A101) from IPHASE (Suzhou, China) according to the instructions. All incubations were performed at 37 °C in a shaking water bath. An incubation system in a total volume of 200 µL containing EVs (0.5 mg/mL), NADP⁺ (1.3 mM), MgCl2 (3.3 mM), glucose 6-phosphate (3.3 mM), glucose 6-phosphate dehydrogenase (0.4 U/mL) and sodium citrate (0.05 mM) was prepared in PBS buffer (pH = 7.4) and preincubated at 37 °C for 10 min. Each experiment was carried out in 6 replicates (n = 6).

Precision-cut liver slices (PCLS)

Preparation and incubation of the mouse PCLS were modified from the previously described protocols [26, 27]. In brief, 2 years aged C57BL/6 mice were anaesthetized using 2% isoflurane/1.5% oxygen inhalation, and livers were dissected in Krebs-Henseleit buffer (KHB, pH 7.4, Sigma, K3753). Fresh liver was cored using an 8-mm biopsy punch (Acuderm). Cores were embedded in microwave-preheated 4% low-melting agarose (Lonza, 50101) in KHB once the agarose cooled to 37 °C. Once solidified, the agarose-embedded liver tissues were superglued to the specimen plate, submersed in the buffer tray containing ice-cold KHB, and cut into slices (250 μm of thickness) using a VT 1200s vibratome (Leica) at the speed of 0.08 mm s − 1 and amplitude of 2 mm. Groups of two PCLS were transferred onto the 0.4 μm pore polycarbonate membrane cell culture insert (Sigma, Z353086) in 6-well plate with 1.5 ml of medium (William’s E Medium (Thermo Fisher Scientific, 32551020) containing 2% EVs-free FBS, 1× penicillin/streptomycin, 1× l-glutamine and 2 g l^{− 1} glucose (Gibco, A24940-01)) per well, and cultured in a humidified 37 °C incubator with 5% CO2 with agitation (50 rpm).

To measure drug metabolism after EVs administration, the PCLS from mice were cultured overnight, followed by treatment with PBS or 10 µg ml^− 1 of EVs in the presence of inhibitors. Then, the PCLS were cultured in media containing substrates for the core Cyp enzymes in the presence of PBS or 10 µg ml^− 1 of EVs and co-treated with inhibitors for another 24 h. One millilitre of media was collected to measure drug-metabolizing activity using LC–MS/MS, and PCLS were weighed for normalization, and then cultured in the media containing substrates for the core Cyp enzymes for 24 h. Then, 1 ml of media was collected for drug-metabolizing activity assay using LC–MS/MS and PCLS were weighed for normalization.

Animal experiments

The animal experimental procedures were approved by the Institutional Animal Care and Use Committee of Sichuan University (Approval No. 2018010 A). All animals received humane care according to the Institutional Animal Care and Use Committee (CN) guidelines. We complied with the ARRIVE reporting guidelines. C57BL/6J mice were purchased from DASHUO Animal Company (SCXK (CHUAN) 2015–30, Chengdu, China). After a two-week adaptive period, the mouses were treated according to the experimental protocol.

Statistical analysis

The data are presented as the means ± SEMs and were analyzed via GraphPad Prism 9.0 (San Diego, CA, USA). For the normally distributed data, a significant difference was determined via Student’s t test for comparing 2 groups and 1-way ANOVA with post hoc Tukey’s test for comparing multiple groups under the condition that F was p < 0.05 and that there was no significant variance inhomogeneity. For the nonnormally distributed data, a significant difference was determined via the Mann‒Whitney U test for comparing 2 groups and the Kruskal‒Wallis H test with the post hoc Steel–Dwass test for comparing multiple groups. Statistical significance was defined as p < 0.05.

Other materials and methods are described in the Supplemental Methods.

Results

Significant disorders of hepatic enzyme expression in various liver diseases

To better understand the landscape of hepatic drug enzymes, we analyzed the expression of the enzymes involved in a variety of liver diseases by proteomics. The results indicate that the levels of these proteins, such as p450, alcohol dehydrogenase (ADH) and UGTs, are disrupted in both patients and animal models (Fig. 1a and b), underscoring the critical role of hepatic drug enzymes imbalances in disease progression.

Fig. 1.

Dysregulation of hepatic drug-metabolizing enzymes expression in liver diseases. (a) Heatmap analysis of hepatic drug enzymes in the patients with NAFLD (n = 15), ALD (n = 15) and ALF (n = 5) derived from proteomics. (b) Heatmap analysis of hepatic drug enzymes in the mice model of Fibrosis (n = 6), ALD (n = 6) and ALF(n = 6) derived from proteomics. (c and d) Immunostaining for CYP2E1, CYP3A4, UGT1A1, SULT1A1 and GSTA1 in human livers. Scale bar, 50 μm. (e and f) Representative western blots of CYP2E1, CYP3A4 and UGT1A1 in ALF and fibrosis mice, and β-actin was used as a protein loading control. Quantification of expression of typical markers in livers. For this figure, the data were analyzed 6 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

To determine the clinical relevance of our findings, we collected liver tissues from patients clinically diagnosed with or without nonalcoholic fatty liver disease (NAFLD) and alcoholic liver disease (ALD), and the related pathological characteristics were analyzed (Fig. S1A and B; Tables 1 and 2 in Supplemental Methods). As shown by the immunohistochemistry (IHC), notably increased protein (CYP2E1, CYP3A4, UGTIA1 and SULT1A1) and decreased hepatic enzyme (GSTA1) levels were observed in NAFLD and ALD patient livers (Fig. 1c and d). In addition, typical acute liver failure (ALF) and fibrosis (Fib) models were induced in the mice (Fig. S1C and D). The protein level of CYP2E1 was elevated, and that of UGT1A1 was decreased in ALF model (Fig. 1e). The protein levels of CYP3A4 and UGT1A1 decreased in Fib mice (Fig. 1f). Taken together, these results highlighted targeting hepatic enzymes dysregulation as a potential clinical therapy.

HELP is active in TEV

To address this clinical dilemma, tissue-derived extracellular vesicles (TEV) and hepatocyte-derived EVs (HEV) were obtained via various steps and differential centrifugation (Fig. 2a). As shown by electron microscopy imaging, the sizes of the HEV and TEV were approximately 158 nm and 197 nm, respectively, and they presented a spherical morphology with a uniform size (Fig. 2b and c). A series of EVs markers, including CD63, TSG101 and Alix, were detected in both types of EVs (Fig. S2A). These results indicated that HEV and TEV were well separated and purified.

Fig. 2.

Characterizations of‌ TEV. (a) Schematic diagram of the preparation of EVs. A certain number of liver tissue and hepatocytes were isolated from the digestive tract. EVs were obtained by differential centrifugation. (b) Representative ultrastructural images of HEV and TEV via TEM. Scale bar, 200 nm. (c) Size distributions of HEV and TEV measured via NTA. (d) KEGG analysis of drug metabolism from the HEV and TEV. (e) Schematic illustration of the animal experiments. (f) Ex vivo fluorescence imaging of major organs, including the heart, liver, spleen, lung, kidney and pancreas, in C57BL/6 mice after various treatments as indicated at different time points. (g) Quantification of the fluorescence signal in the liver after administration at different time points (n = 5). (h) Curves of the fluorescence intensity of TEV over time (n = 5). For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

To explore the differences between HEV and TEV, we further analyzed the obtained EVs. Proteomic analysis revealed that HEV and TEV contained various specific proteins, and the GO term analysis of the proteins revealed that the p450 pathway, alcohol metabolic process and NADPH regeneration were more markedly enriched in TEV than HEV (Fig. S2B-E). Furthermore, the enrichment analyses of the KEGG terms revealed that hepatic enzymes were associated with increased drug metabolism in the TEV. As shown in the heatmap of key differentially expressed genes (DEGs), the levels of proteins related to drug decomposition were greater in the TEV (Fig. 2d and S2F). Similarly, various proteins, such as CYP3A4, SULT1A1and GSTA1, accumulated in TEV (Fig. S2A).

Next, to evaluate the biodistribution of EVs, HEV and TEV were labeled with HPK67 dye [28]. These EVs were administered by i.v. injection into C57BL/C mice via the tail vein, and mice that received the same amounts of free HPK67 dye were used as controls (Fig. 2e). After administration, the mice were sacrificed, respectively, and the organs were excised for IVIS imaging. HEV and TEV accumulated mainly in the liver, pancreas and kidney, while there were no obvious signals in the organs of the free dye group (Fig. 2f). The fluorescence intensity of TEV in those organs was not obviously different than that in the HEV group after injection (Fig. 2g). Notably, TEV displayed the most intense fluorescence in the liver at various time points (Fig. 2h).

More importantly, we next examined the drug-metabolizing capacity of EVs. The metabolic rate curves are shown in Fig. 3a and b. The absence of NAD⁺ did not exhibit activity, and the TEV showed significant kinetics. Other enzymes, like CYP3A4, UGT1A1 and SULT1A1, confirmed considerable metabolic activity also (Fig. 3c, d, e).

Fig. 3.

Hepatic drug-metabolizing enzymes are active in TEV. (a, b) The enzyme kinetic curves and activity of ADH. (c, d, e) The enzyme activity of CYP3A4, UGT1A1and SULT1A1. (f) Schematic illustration of the procedure of drug-metabolizing activity analysis of the precision-cut liver slices (PCLS). Substrates of p450 enzymes (including phenacetin, bupropion, tolbutamide, dextromethorphan and midazolam) were added to the media, and their corresponding metabolites (including acetaminophen, hydroxybupropion, 4-hydroxytolbutamide, dextrorphan and 1-hydroxymidazolam) were analyzed. (g, h) Drug-metabolizing activity of the p450 enzymes in the PCLS sectioned from old mice, treated with HEV or TEV and mathoxsalen (5 µM) or phenobarbital (0.1 mM) for 24 h. n = 5 mice per group. (i) Drug-metabolizing activity of Cyp2c38, Cyp2d9/ Cyp2d10 and Cyp3a11 in the PCLS. For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

Next, we demonstrated that the metabolic capacity of core P450 enzymes for specific substrates was indeed reduced in liver slices obtained from two-year-old mice, as assessed by measuring the levels of their metabolic products (Fig. 3f). Using methoxsalen (a CYP1A2 inhibitor) and phenobarbital (a CYP2B10 inducer) as negative and positive controls, respectively, we further established that ex vivo TEV treatment effectively restored drug metabolism in murine liver slices. This metabolic activity was inhibited by methoxsalen (Fig. 3g) and enhanced by phenobarbital (Fig. 3h). Notably, TEV treatment rescued P450 drug-metabolizing enzyme activity in aged mice—a effect not observed with HEV (Fig. 3i). These results showed that HELP in TEV remain functional.

HELP improves ALF by regulating hepatic enzymes in mice

Based on our findings, we hypothesized that EVs might have beneficial effects on liver injury. We therefore established an acetaminophen (APAP)-induced ALF model exhibiting characteristic hepatic drug enzyme disturbances as described previously [29], followed by tail vein injection treatment with 5 mg/kg EVs daily for 7 days (Fig. 4a). The survival rate of the mice significantly decreased following APAP treatment, which was diminished by EVs, TEV demonstrated superior efficacy (Fig. 4b). Concordant with mouse studies, APAP-induced ALF, as evidenced by increased levels of total serum bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), uric acid (UA) and blood urea nitrogen (BUN) (Fig. 4c-g), and expanded hepatocellular cell death and spotty areas (Fig. 4i and j), while they were effectively improved by EVs. Moreover, APAP boosted the release of interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α), which was remarkable ameliorated by EVs, with TEV exhibiting a more favorable effect than HEV (Fig. 4h). Compared with that in the Ctrl group, the TUNEL results of the mouse liver samples revealed that APAP led to increased cell death in the ALF group, whereas the level of apoptosis was effectively suppressed in the two experimental groups (Fig. 4i and k; Fig. S3A). More importantly, the level of Ki67, a frequently used biomarker of cell proliferation, was significantly lower in ALF mice than in control mice. Conversely, substantial upregulation was observed following TEV or HEV treatment, with the former demonstrating superior outcomes (Fig. 4i and l; Fig. S3B). These data indicated that EVs ameliorate APAP-induced hepatocyte injury and maintain liver function and mouse survival.

Fig. 4.

TEV improves ALF by regulating hepatic enzymes in mice. (a) Overview of the animal experimental procedure. Mice were fasted for 12 h before intraperitoneal injection with 300 mg/kg of APAP to induce ALF, followed by treatment with 5 mg/kg EVs every day (n = 5). (b) Survival rate of APAP and/or EVs–treated mice. (c - g) Serum total bilirubin, AST ALT, UA and Bun levels of APAP and/or EVs–treated mice. (h) Serum IL-1β and TNF-α levels. (i) Representative images of H&E-stained liver tissues. Dotted lines in H&E images outline necroptotic areas. Scale bar, 100 μm. Liver tissues labeled by immunofluorescence with anti-Ki67 (red) and TUNEL (green) were used to detect DNA fragmentation, and DAPI (blue) was used as a counterstain. Scale bar, 50 μm. (j and k) Quantification of ki67- and TUNEL-positive areas. (l) Necroptotic areas. (m and n) Representative western blots and quantification of hepatic drug enzymes, including CYP2E1, UGT1A1, SULT1A1 and GSTA1, were detected in livers, and β-actin was used as a protein loading control. (o - r) AML-12 cells were exposed to APAP plus HEV and TEV and/or siRNAs. The cell viability was detected by CCK8.For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

Upon examining whether TEV was beneficial for liver dysfunction via HELP replenishment, our results displayed the upregulation of CYP2E1 and the downregulation of UGT1A1, SULT1A1 and GSTA1 in the livers of the APAP group, which were significantly replenished after TEV treatment (Fig. 4m and n). To explore the regulatory role of these enzymes in vitro, we performed siRNA-mediated knockdown of Sult1a1, Ugt1a1 and Gsta1 in AML-12 cells. Indeed, after its expression was downregulated (Fig. S3C), cell viability was markedly reduced compared with that following APAP treatment alone (Fig. 4o). Moreover, in gene-silenced AML-12 cells, restoring the TEV significantly saved cell viability, whereas these increases were abrogated by the pan-SULT1A1/UGT1A1/GSTA1-blocking antibody. HEV had no such effect (Fig. 4p-r). Taken together, these data exhibited that TEV prominently complemented active hepatic drug-metabolizing enzymes, further alleviating ALF.

TEV-delivered HELP rebalances hepatic metabolism and alleviates ALD

ALD is a more complex liver disease associated with elevated activity of enzymes, such as p450 and alcohol-metabolizing families, which contributes to cellular injury [30]. Thus, we established a typical ALD model as described previously, followed by tail vein injection treatment with 5 mg/kg EVs daily for 10 days (Fig. 5a). The liver index (liver weight/body weight) is an important indicator for assessing liver pathology in ALD mice. Compared with the Ctrl group, the ALD group presented an elevated liver index, which was diminished after EVs injection (Fig. 5b). Similarly, AST and ALT levels substantially increased in the ALD group, indicating hepatocyte injury but were effectively eliminated by treatment with TEV. While trends of elevation were observed, triglyceride (TG) and total cholesterol (TC) did not reach statistical significance in this model (Fig. 5c-f). The mice subsequently exhibited the typical histopathological characteristics of ALD in the liver, including noticeable steatosis, lobular inflammation and fibrosis. In contrast, the characteristics of ALD were markedly ameliorated after HEV or TEV injection (Fig. 5h). In addition, the ALD scores were the similar alterations (Fig. 5g). Moreover, BODIPY and Oil Red O staining revealed a significant increase in the number and size of lipid droplets in the ALD group, which decreased with increasing HEV or TEV (Fig. 5h and Fig. S4A). In summary, these data revealed that treatment with TEV more effectively alleviated the progression of ALD.

Fig. 5.

TEV repairs ALD by modulating hepatic drug enzyme profiles. a) Overview of the animal experimental procedure. The NIAAA model was induced as described, followed by tail vein injection treatment with 5 mg/kg EVs every day from days 5 to 28 (n = 5). (b) Liver index. (c - f) Serum AST, ALT, TG and TC levels in ALD- and/or HEV- or TEV-treated mice. (g) ALD activity score. (h) Liver histopathology, including H&E staining, Sirius red staining and Oil Red O staining. Scale bar, 100 μm. BODIPY. Scale bar, 50 μm. (i) PCA of liver samples from the experimental mice. (j) KEGG analysis of samples from the ALD and ALD + TEV group. (k) Heatmap analysis of hepatic drug enzymes in livers. (l and m) Representative western blots and quantification of ethanol and lipid metabolism, including ALDH2, ADH1B, SREBP1c and CPT1, were detected in livers, and β-actin was used as a protein loading control. (n and o) The mRNA expressions of hepatic drug enzymes were detected, and β-actin was used as control. For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

To evaluate the mechanisms of TEV for the treatment of ALD in mice, we collected data from liver samples and used a fully consistent mass spectrometry method. We detected 4879 qualitative proteins via proteome analysis. The visible differences between the four groups were detected via principal component analysis (PCA). Moreover, there were also noticeable distinctions between the ALD + TEV group and the ALD + HEV group (Fig. 5i). These results indicated that the mechanism of treatment for ALD in the TEV group may be distinct from that employed in the HEV group.

To explore the differences after HEV or TEV treatment in ALD mice, we further analyzed the data via size-heatmap analysis of liver samples, which were classified into three distinct clusters (Fig. S4B). The enrichment analyses of the KEGG terms in the two intervention groups revealed significant differences in drug metabolism in the liver, such as p450- and alcohol-metabolizing families (Fig. 5j). The heatmap shows that the KEGG terms associated with hepatic enzymes were significantly different after treatment with the two EVs (Fig. 5k and Fig. S4C). Taken together, these results confirmed that TEV may replenish HELP and further alleviate ALD.

To further verify our data, ethanol and lipid metabolism were detected. Compared with those in the Ctrl group, the levels of aldehyde dehydrogenase 2 (ALDH2), the most commonly utilized biomarker of ethanol metabolism, were markedly elevated in the ALD group. TEV intervention significantly increased ALDH2 protein levels compared to the model group. However, no pronounced difference was observed at the mRNA level, and HEV did not have an effect (Fig. 5l, m and n). More importantly, the mRNA levels of other hepatic enzymes did not significantly differ between the ALD group and the ALD + TEV group (Fig. 5o; Fig. S4D and E). In addition, TEV ameliorate lipid metabolism disorders, as evidenced by western blot results for sterol regulatory element-binding protein 1c (SREBP1c), carnitine palmitoyl-transferase 1 (CPT1) and peroxisome proliferator-activated receptor α (PPARα) (Fig. 5l and m). These results suggested that TEV may provide ALD mice with active enzymes.

To validate this conclusion, compounds that directly disrupt the structure of the enzyme were utilized in a palmitic acid (PA) stimulation-induced lipotoxic cell model [31]. Fenofibrate (Fen), salicylamide (Sal), GSTO-IN-2 (GO2), and nialamide (Nia) are inhibitors of the P450, UGT, GST, and monoamine oxidase (MAO) families, respectively. The cell viability was reduced, which was effectively ameliorated following EVs intervention (Fig. S4F). These compounds could abolish the therapeutic effect of TEV but not HEV in lipotoxic cells and were not very toxic to cells at the concentrations used (Fig. S4G and H). Specifically, there was a clear trend toward the downregulation of related genes, including Cyp2c9, Cyp3a4, Ugt1a1 and Ugt2a1, in PA-treated cells. TEV treatment had no effect on the mRNA levels; however, the expression of the molecules involved was substantially upregulated by the combination of Fen or Sal (Fig. S5A and B). The results of the western blotting analysis demonstrated that PA stimulation restricted the levels of CYP3A4, UGT1A1 and UGT2A1. However, TEV intervention was able to restore their levels. In contrast, inhibitor treatment did not affect the expression of the relevant enzymes, with the exception of CYP2D6 (Fig. S5C and D). The overall trends of the GST and MAO families were similar to previous results (Fig. S5E-H). Taken together, these results indicated that TEV could significantly replenish active hepatic enzymes and further alleviate ALD.

HELP mitigates acute and chronic liver injury

There is still no effective therapy for patients with liver injury who have low levels of hepatic enzymes [32]. Combined with our existing results, we further explored the role of HELP in chronic liver diseases. Thus, in mice pretreated with 10 ml/kg of intraperitoneal carbon tetrachloride (CCl₄) for 6 weeks, the addition of 5 mg/kg TEV daily for five consecutive days alleviated liver injury induced by 100 mg/kg APAP daily for 3 days (Fig. 6a). The survival rate of the mice significantly dropped following APAP treatment, which was diminished by TEV (Fig. 6b). In parallel, APAP induced ALF, as evidenced by raised levels of total serum bilirubin, AST, ALT, UA and BUN (Fig. 6c-g), and increased hepatocellular cell death and spotty areas (Fig. 6h and i), while these effects were effectively improved by TEV. Moreover, APAP advanced the release of IL-1β, which was markedly ameliorated by TEV (Fig. 6j). These data revealed‌ that TEV has a positive effect on the amelioration of APAP-induced injury in conditions associated with underlying liver disease.

Fig. 6.

TEV improves drug metabolism in chronic liver diseases. (a) Overview of the animal experimental procedure. Mice were pre-treated with 10 ml/kg of 20% CCl4 every 3 days for 6 weeks, followed by injection with 100 mg/kg APAP daily for 3 days and 5 mg/kg EVs daily for five consecutive days (n = 5). (b) Survival rate. (c - g) Serum total bilirubin, AST ALT, UA and Bun levels. (h and i) Representative images of H&E and Sirius red-stained liver tissues. Dotted lines in H&E images outline necroptotic areas. Scale bar, 100 μm. (j) Serum IL-1β levels in mice. For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

HELP enhances drug efficacy in NAFLD via metabolic synergy

Next, we established a typical NAFLD model to explore the effects of TEV on drug metabolism in complex situations. The model was developed with a high-fat, high-fructose and high-cholesterol diet (FFC) for 16 weeks, followed by injection of 100 mg/kg silymarin and 5 mg/kg EVs daily for four weeks (Fig. 7a). Compared with the Ctrl group, the NAFLD group presented an elevated liver index, which was diminished after TEV and/or silymarin intervention (Fig. 7b). Similarly, AST, ALT, TG and TC substantially increased in the NAFLD group but were effectively eliminated by treatment with TEV and/or silymarin (Fig. 7c and d). Compared with the other groups, the TEV + silymarin group presented a more favorable effect. The mice subsequently exhibited the typical histopathological characteristics of NAFLD in the liver, including noticeable steatosis, lobular inflammation and fibrosis. In contrast, the characteristics of NAFLD were markedly ameliorated after treatment (Fig. 7e). Moreover, Oil Red O staining revealed a significant increase in the number and size of lipid droplets in the NAFLD group, which decreased with TEV and/or silymarin (Fig. 7e and f). In addition, the NAFLD scores showed comparable improvement patterns (Fig. 7g). TEV combined with silymarin confirmed superior efficacy. In summary, these data clearly demonstrated that TEV could improve the efficacy of drugs.

Fig. 7.

HELP enhances drug efficacy in NAFLD. (a) Overview of the animal experimental procedure. Mice were pre-treated with FFC diet for 12 weeks, followed by oral with 100 mg/kg silymarin and 5 mg/kg EVs daily for four weeks by injection (n = 5). (b) Liver index. (c and d) Serum AST, ALT, TG and TC levels in mice. (e) Liver histopathology, including H&E staining and Masson staining. Scale bar, 100 μm. Oil Red O staining. Scale bar, 50 μm. (f) Quantification of Oil Red O-stained areas. (g) ALD activity score. For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

HELP-Mediated host liver metabolic support enhances cancer therapy

In recent years, liver cancer incidence and mortality rates have remained persistently high globally. Regorafenib is an oral multi-kinase inhibitor used to treat advanced hepatocellular carcinoma (HCC), but its efficacy is markedly compromised in patients with hepatic impairment [33]. Consequently, we established a primary HCC model with liver dysfunction via neonatal mice, which were injected with 200 µg of streptozotocin (STZ) per mouse and induced with an FFC diet for 16 weeks, followed by a 14-day intervention involving the oral administration of 5 mg/kg regorafenib and intravenous injection of 5 mg/kg TEV daily (Fig. 8a and Fig. S6A). H&E staining and quantitative morphometric assessment of the tumor region revealed dysplastic nodules in some tumors (Fig. 8b). The number of tumors further increased, and the maximal tumor volume increased in HCC mice but was effectively eliminated by treatment with TEV + regorafenib or regorafenib alone. The former had a more favorable impact than the latter did, and TEV had no significant therapeutic effects (Fig. 8c and Fig. S6B). Furthermore, in the tumor region, the intensity of Ki67, a marker of malignant proliferative cells, and AFP, a marker of primary carcinoma of the liver, was significantly greater in HCC model mice than in control mice on IHC. Compared with all other treatment regimens, the TEV plus regorafenib combination consistently indicate superior therapeutic efficacy (Fig. 8b, d and e).

Fig. 8.

TEV makes anti-tumor drugs more effective. (a) Overview of the HCC model procedure. In sucking mice, 100 ug/per streptozotocin (STZ) by a single intraperitoneal injection at 2 days after birth served as an accelerator with a FFC diet for 16 weeks, followed by oral with 5 mg/kg Regorafenib and 5 mg/kg EVs daily for two weeks by injection (n = 5). (b) Liver histopathology, including H&E staining and Sirius red staining. Scale bar, 100 μm. Immunostaining for AFP and ki67 in livers. Scale bar, 50 μm. (c) Tumor numbers per mouse in liver. (d and e) Quantification of AFP-positive tissue area and the number of ki67-positive hepatocytes per HPF. (f) Overview of the ICC model procedure. Mice fed with a FFC diet and 500 mg/L thioacetamide in water for 24 weeks, followed by injection with GP group every 3 days and 5 mg/kg EVs daily for three weeks. (g) Liver histopathology, including H&E staining and Sirius red staining. Scale bar, 100 μm. Immunostaining for CK19 and ki67 in livers. Scale bar, 50 μm. (h) ALT/AST ratio in mice. (i and j) Quantification of CK19-positive tissue area and the number of ki67-positive hepatocytes per HPF. For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

The incidence of intrahepatic cholangiocarcinoma (ICC) has increased by 140% over the past decade, and it is now the second most common primary malignant tumor of the liver [34]. Gemcitabine plus cisplatin (GP) remains the standard first-line regimen [35]. However, this combination chemotherapy has significant toxicity, including renal and hepatic impairment, as well as thrombocytopenia. Therefore, we established a representative ICC model by feeding mice an FFC diet with 500 mg/L thioacetamide administered in the drinking water for 24 weeks (Fig. S6C). This was followed by intervention with the standard GP chemotherapy regimen alongside TEV treatment for 3 weeks (Fig. 8f). According to quantitative morphometry, the number of tumors and maximal tumor volume were greater in ICC mice than in control mice but were effectively reduced by treatment with TEV and/or GP alone, and TEV had no significant therapeutic effects (Fig. S6C and D). As shown in Fig. 8g, severe ductular reaction (DR) and excessive collagen deposition, the typical pathological features of ICC, were obvious in ICC mice. However, marked attenuation of these manifestations was observed following GP and/or TEV interventions, with combination therapy demonstrating superior therapeutic efficacy. Notably, hepatotoxicity was significantly lower in the ICC + TEV plus GP group than in the ICC + GP group, as evidenced by a substantial decrease in the ALT/AST ratio (Fig. 8h).

To more precisely evaluate DR, IHC for CK-19, a marker for cholangiocytes, was performed. As expected, there was a marked increase in CK-19 intensity in the ICC groups compared with the Ctrl group, while less CK-19 staining was observed after GP and/or TEV intervention, and TEV combined with GP demonstrated superior therapeutic efficacy (Fig. 7g and i). Furthermore, the numbers of Ki67- and AFP-positive hepatocytes were similar (Fig. 8g and j; Fig. S6E and F). Taken together, these data demonstrated that TEV effectively attenuate chemotherapy-induced hepatotoxicity while increasing the antitumor efficacy of therapeutic agents.

HELP aids in the metabolism of drugs in special populations

Building on our previous exploration of HELP in diverse liver injury models, we note that intrinsically impaired enzymatic activity exists in specific vulnerable cohorts—particularly pediatric populations with immature enzymes and geriatric individuals exhibiting an age-associated decline in dynamics [36, 37]. Thus, we investigated the therapeutic potential of TEV in these specific populations. For 10-day-old suckling mice, subtoxic doses of APAP (50 mg/kg) were administered for three consecutive days to induce hepatotoxicity. Intervention with 5 mg/kg TEV plus either 2 mg/kg Fen or GO2 was subsequently initiated daily (Fig. 9a). The survival rate of the mice significantly decreased following APAP treatment, which was diminished by TEV, and Fen or GO2 inhibited its therapeutic effect (Fig. 9b). Compared with that in the Ctrl group, APAP increased hepatocellular cell death and spotty areas in the H&E and TUNEL results of the mouse liver samples, whereas the level of apoptosis was effectively suppressed in the APAP + TEV groups. However, pharmacological inhibition of enzymatic activity by Fen or GO2 markedly attenuated the hepatoprotective efficacy of TEV (Fig. 9c-e and Fig. S7A). These data showed that TEV could ameliorate APAP-induced hepatocyte injury in neonatal mice.

Fig. 9.

TEV helps the metabolism of drugs in special populations. (a) Overview of the animal experimental procedure. Mice were injected with 50 mg/kg of APAP, followed by treatment with 5 mg/kg EVs and 2 mg/kg drug daily (n = 5). (b) Survival rate. (c and d) Representative images of H&E-stained liver tissues. Dotted lines in H&E images outline necroptotic areas. Scale bar, 100 μm. Liver tissues labeled by immunofluorescence with TUNEL (green) were detected, and DAPI (blue) was used as a counterstain. Scale bar, 50 μm. (e) Quantification of TUNEL-positive areas. (f) Overview of the animal experimental procedure. The old mice were injected with HRZ diet every three days for 10 weeks, followed by treatment with 5 mg/kg EVs and 10 mg/kg inhibitors daily for 3 weeks. (g) Survival rate. (h - l) Serum total bilirubin, AST ALT, CREA and Bun levels of HRZ diet and/or EVs and inhibitors–treated mice. (m and n) Representative images of H&E-stained liver tissues. Dotted lines in H&E images outline necroptotic areas. Scale bar, 100 μm. Liver tissues labeled by immunofluorescence with TUNEL (green) and DAPI (blue) was used as a counterstain. Scale bar, 50 μm. (o) Quantification of TUNEL-positive areas. (p and q) Serum IL-1β and TNF-α levels. (r) The viability was assessed using the CCK8 assay in primary mouse hepatocytes. (s) The IL-1β and TNF-α levels in supernatant. (t) The mRNA expressions of Sod2, Gpx4 and Nrf2 were detected, and β-actin was used as control. For this figure, the data were analyzed 5 times for biologically independent tests, and the results are presented as the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

Next, we further investigated the function of HELP in elderly mice. HRZ (isoniazid, rifampicin, pyrazinamide) serves as the core first-line regimen for tuberculosis treatment and requires long-term administration [38]. Therefore, this poses significant hepatotoxicity risks in special populations. To model chronic drug-induced liver injury, we administered standard HRZ therapy to aged mice for 10 weeks, followed by interventions with 5 mg/kg TEV and 10 mg/kg FEN or GO2 starting from the seventh week (Fig. 9f). Compared with the Ctrl group, chronic HRZ administration resulted in a significant reduction in the survival rate of aged mice. While TEV alone improved survival metrics, coadministration with Fen or GO2 diminished these benefits (Fig. 9g). In parallel, HRZ increased the levels of total serum bilirubin, AST, ALT, creatinine (CREA) and BUN (Fig. 9h-l) and increased hepatocellular cell death and spotty areas (Fig. 9m-o and Fig. S7B), which were effectively improved by TEV. Moreover, HRZ increased the release of IL-1β and TNF-α, which was markedly ameliorated by TEV, with the benefits of TEV being suppressed by Fen or GO2 (Fig. 9p and q). These data revealed that TEV could ameliorate HRZ-induced liver injury.

We further investigated whether TEV act through these enzymes in neonatal mouse hepatocytes, which are a natural model with insufficient enzyme activity. The cell viability gradually decreased with increasing APAP concentration, which was effectively ameliorated by TEV but not by HEV (Fig. S7C and D). Fen or GO2 abolished the therapeutic effect of TEV (Fig. 9r). Specifically, the release of IL-1β and IL-18 significantly increased after APAP stimulation but decreased in the APAP + TEV groups, and Fen or GO2 inhibited their therapeutic effects (Fig. 9s). Moreover, there was a clear trend toward the downregulation of antioxidant genes, including Sod2, Gpx4 and Nrf2, in APAP-treated cells. Backfilling TEV significantly increased the expression of these genes, whereas these increases were abrogated by these compounds (Fig. 9t). Taken together, these data indicated that TEV significantly enhanced hepatic enzyme activity and mitigated liver injury in specialized models.

Discussion

This study demonstrated that TEV serve as effective carriers of active hepatic metabolic enzyme systems, showing therapeutic benefits in reducing liver injury across various murine models. Notably, TEV are enriched in HELPs—such as p450s, UGTs, SULTs, and GSTs—and maintain strong drug-metabolizing activity both in vitro and in vivo. The administration of TEV led to significant improvements in survival and liver function, frequently outperforming HEV, thereby establishing TEV as a promising platform for restoring enzymatic homeostasis in impaired livers.

Importantly, TEV exhibited pronounced efficacy in models of chronic and multifactorial liver diseases, such as ALD‌ and NAFLD‌, where enzyme dysfunction both drives and exacerbates disease progression. The restoration of ‌ALDH2, CYP3A4, UGT1A1‌, and other critical enzymes by TEV—without concomitant changes in mRNA levels—suggests a posttranslational mechanism of action, likely mediated through direct enzyme delivery. This hypothesis was corroborated by pharmacological inhibition studies, in which disruption of enzyme activity abolished the therapeutic effects of TEV, confirming their enzyme-dependent mode of action. Beyond acute injury models, we extended TEV applications to vulnerable populations with inherent enzyme deficiencies, including ‌neonates and aged individuals‌, where conventional pharmacotherapy is often limited by toxicity. TEV not only enhanced survival and attenuated injury in these cohorts but also demonstrated synergistic effects with standard therapies, such as ‌regorafenib‌ and ‌silymarin‌, improving both therapeutic efficacy and safety profiles.

A context-dependent rationale is essential for enzyme-based therapy. Our TEV-delivered HELP strategy aims to restore systemic metabolic homeostasis rather than merely supplement individual enzymes. This approach addresses two key pathological scenarios. In states of broad enzymatic deficiency, such as ALF where conjugating enzymes are depleted (Fig. 4m and n), HELP directly reconstitutes lost function, improving outcomes (Fig. 4). In diseases of metabolic imbalance like ALD—where induced enzymes (e.g., CYP2E1) overwhelm detoxification—HELP rebalances the network by supplying cooperating partners (e.g., ALDH2, Fig. 5l and m) to clear toxic intermediates without exacerbating the initial insult (Fig. 5). Thus, HELP transcends single-enzyme replacement by offering a holistic, network-restorative therapy for complex liver pathologies where isolated corrections are inadequate or risky.

In conclusion, our study establishes TEV as a versatile and effective platform for restoring hepatic enzyme function in a wide spectrum of liver diseases. By leveraging the natural cargo and targeting properties of EVs, HELP offers a novel paradigm for the treatment of metabolic liver disorders, with potential applications in precision medicine and combination therapy.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Xin Zeng: Project administration; Investigation; Data curation; Methodology; Formal analysis; Roles/Writing - original draft. Wei Jiang: Project administration; Funding acquisition; Resources; Roles/Writing - original draft. Baohong He, Zhao Li, Lan Li and Fudong Fu: Software; Formal analysis. Tian Wu and Chengshi Wang: Data Curation. Guangneng Liao and Han Yao: Methodology. Dongbo Wu: Funding acquisition; Conceptualization; Supervision; Validation; Visualization; Writing - review & editing.

Funding

This work was supported by the Science and Technology Program of Sichuan [grant numbers: 2024YFFK0214, 2024NSFSC1690 and 2022NSFSC1559]; The Program of National Natural Science Foundation of China [grant numbers: 81801589]; The 1.35 project for disciplines of excellence, West China Hospital, Sichuan University [grant numbers: ZYGD23030].

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval

All experimental procedures were approved by the Institutional Animal Care and Use Committee of Sichuan University (Approval No. 2018010 A). The use of human liver tissues was reviewed and approved by the West China Hospital Ethics Committee (Approval No. WCH20180037). All participants provided written informed consent in accordance with the Declaration of Helsinki.

Competing interests

The authors declare no competing interests.