Leveraging adrenergic receptor blockade for enhanced nonalcoholic fatty liver disease treatment via a biomimetic nanoplatform

Abstract

Nonalcoholic fatty liver disease (NAFLD) is characterized by excessive lipid accumulation, steatosis and fibrosis. Sympathetic nerves play a critical role in maintaining hepatic lipid homeostasis and regulating fibrotic progression through adrenergic receptors expressed by hepatocytes and hepatic stellate cells; however, the use of sympathetic nerve-focused strategies for the treatment of NAFLD is still in the infancy. Herein, a biomimetic nanoplatform with ROS-responsive and ROS-scavenging properties was developed for the codelivery of retinoic acid (RA) and the adrenoceptor antagonist labetalol (LA). The nanoplatform exhibited improved accumulation and sufficient drug release in the fibrotic liver, thereby achieving precise codelivery of drugs. Integration of adrenergic blockade effectively interrupted the vicious cycle of sympathetic nerves with hepatic stellate cells (HSCs) and hepatocytes, which not only combined with RA to restore HSCs to a quiescent state but also helped to reduce hepatic lipid accumulation. We demonstrated the excellent ability of the biomimetic nanoplatform to ameliorate liver inflammation, fibrosis and steatosis. Our work highlights the tremendous potential of a sympathetic nerve-focused strategy for the management of NAFLD and provides a promising nanoplatform for the treatment of NAFLD.

Highlights

The potential of adrenoceptor-blockade strategy was explored for reprogromming the activated hepatic stellate cells and reducing hepatic lipid accumulation in the treatment of NAFLD.

A biomimetic nanoplatform with ROS-responsive and ROS-scavenging properties was tailored for achieving precise delivery of retinoic acid and adrenoceptor antagonists as well as consuming ROS to provide anti-inflammatory benefits.

The biomimetic nanoplatform integrating adrenoceptor blockade exhibits excellent ability to reduce liver steatosis and alleviate inflammation and fibrosis, thereby achieving enhanced NAFLD treatment.

Introduction

Nonalcoholic fatty liver disease (NAFLD) has become one of the most common causes of chronic liver disease, and it affects an estimated 5% of population worldwide [1, 2]. NAFLD is a metabolic syndrome characterized by excessive lipid accumulation and liver steatosis inflammation that is typically accompanied by hepatocyte injury, which can progress to liver fibrosis and cirrhosis [3,4,5]. Liver fibrosis is recognized as an inevitable process that progresses from NAFLD to more severe disease and is a major predictor of liver-related mortality [6,7,8]. Currently, NAFLD treatments have focused on the classical pathogenesis of liver lipid deposition, inflammation or fibrosis [3]. Unfortunately, there are currently limited available options for treating NAFLD in the clinic.

The sympathetic nervous system (SNS), an important branch of the autonomic nervous system, is responsible for the physiological regulation of internal organs, including tissue homeostasis, organogenesis, plasticity, and regeneration [9, 10]. Liver tissues are highly infiltrated and innervated by sympathetic nerves [11, 12]. Emerging evidence indicates that sympathetic nerves are activated during liver fibrosis and that sympathetic nerve hyperactivity with elevated levels of norepinephrine is involved in the progression of liver fibrosis and cirrhosis [13, 14]. Hepatic stellate cells (HSCs), as a center of liver fibrosis formation and development, may be modulated by the SNS because HSCs abut and touch sympathetic nerve fibers and express functional adrenoceptors [15, 16]. HSCs can be activated by sympathetic nerve neurotransmitters to a proliferative and myofibroblast phenotype and subsequently produce excessive amounts of collagen. Previous reports have shown that mice with genetic deletion of dopamine-β-hydroxylase and a lack of norepinephrine are resistant to hepatic fibrosis [17]. On the other hand, the sympathetic neurotransmitter plays a key role in the regulation of liver lipid homeostasis through adrenoceptors overexpressed on hepatocytes. Hepatic adrenergic activity increases with lipid accumulation and is considered a vital driver of liver steatosis [18, 19]. Although these findings suggest that sympathetic overactivity promotes the development of NAFLD through the secretion of norepinephrine, the potential of manipulating liver sympathetic nerves to ameliorate liver steatosis and enhance antifibrotic effects remain mostly unexplored.

All-trans retinoic acid (RA), a natural metabolite of vitamin A, has attracted intensive attention for the treatment of NAFLD. RA not only induces the quiescence of HSCs and potently suppresses collagen production but also alleviates obesity-induced liver steatosis by inhibiting lipogenesis and insulin resistance [20,21,22]. However, its clinical application is limited by its poor bioavailability and unsatisfactory efficacy. We hypothesized that blockade of adrenoceptors to interrupt efferent nerve trafficking by precise targeting might have a synergistic or additive effect with RA in ameliorating liver steatosis and fibrosis, ultimately resulting in a superior therapeutic outcome for NAFLD.

Herein, a fibrotic liver-targeting biomimetic nanocarrier consisting of a CD44 aptamer-modified macrophage membrane shell and a diselenide-bridged mesoporous silica (MSN) core was developed for the delivery of RA and the adrenoceptor antagonist labetalol (LA). The macrophage membrane coating bestowed the biomimetic nanoplatform with reduced capture of Kupffer cells, and the CD aptamer enhanced its specific binding capability to activated HSCs overexpressing the CD44 receptor. Additionally, the MSN core encapsulating RA and LA could react with ROS to trigger drug release in the pathological microenvironment of the fibrotic liver, which not only improved the therapeutic specificity and decreased the side effects of the drug, but also consumed ROS at injured liver to provide anti-inflammatory benefits. The results from in vitro experiments confirmed that the nanoplatform’s high delivery efficiency to activated HSCs (aHSCs) could effectively restore aHSCs to a quiescent state and interrupt the crosstalk between HSCs and sympathetic nerves, which consolidated and improved the antifibrotic effect. The improved therapeutic outcomes achieved through the blockade of adrenoceptors were further validated in a BDL-induced liver fibrosis animal model. Importantly, a durable reduction in hepatic lipid accumulation and recovery of liver function were achieved by blockade of adrenoceptors in high-fat diet-fed mice. Overall, this study not only reveals the significance of adrenoceptor blockade in alleviating liver fibrositis and lipid deposition but also provides versatile, efficient, and safe carriers for the clinical treatment of NAFLD (See Scheme 1).

Scheme 1

Schematic illustration of nanodrugs that can inhibit the vicious crosstalk of sympathetic nerves with HSCs and hepatocytes through blocking adrenoceptor for alleviating liver fibrositis and lipid deposition during the treatment of NAFLD

Experimental section

Synthesis of MSNs

A modified sol-gel approach was applied to synthesize MSNs using tetraethyl orthosilicate (TEOS, 98%) and bis[3-(triethoxysilyl)propyl] diselenide (BTESePD) as silica sources and cetyltrimethyl ammonium bromide (CTAB). Briefly, 0.2 g of triethanolamine (TEAH₃) was mixed with 0.8 g of CTAB in 60 mL of deionized water at 80 °C with stirring for 20 min. Then, a mixture of 1.0 g of BTESePD and 5.0 g of TEOS was slowly injected into the above-prepared surfactant solution and then stirred at 80 °C for 6 h. The products were collected by centrifugation and refluxed in an ethanol solution of NH₄NO₃ (0.8% w/v) for 6 h to form mesopores. The obtained MSNs were collected, washed, dried and stored at 4 °C for subsequent experiments.

Cyanine 3 amine (Cy3) and cyanine 5.5 amine (Cy5.5)-labeled MSNs were prepared through a modified EDC/NHS reaction. Briefly, 200 µL 3-aminopropyltriethoxysilane (APTES) were dropwise added into 80 ml ethanol solution of MSNs (1.5 mg/mL) and refluxed at 105 °C for 12 h. Then, the products were collected and dispersed in N, N-dimethylformamide (DMF) solution. Next, 100 µL triethylamine (TEA) and 1.5 g succinic anhydride were added 100 mL DMF solution of MSNs (2 mg/mL) was stirred at 80 °C for 24 h to form carboxyl-functionalized MSNs. Subsequently, the 10 mg of carboxyl-functionalized MSNs were mixed with 100 mL of EDC/NHS aqueous solution and sonicated for 30 min. Next, 1 mg Cy3 or Cy5.5 was added into the mixture and stirred at room temperature for 24 h. The Cy3- and Cy5.5-labeled MSNs were collected washed, dried and stored at 4 °C for subsequent experiments.

Drug loading, degradation and release

Two milligrams of RA and 2 mg of LA were dissolved in 5 mL of dimethyl sulfoxide (DMSO). Subsequently, 5 mg of MSNs was added to the mixture and stirred at room temperature for 24 h. Then, we collected the supernatant to quantify LA and RA using HPLC. The drug encapsulation efficiency (EE%) was calculated by the following equation: EE% = mass of drugs in NPs/mass of total drugs. The drug loading content (DL%) was calculated by the following equation: DL% = mass of drugs in NPs/mass of drug-loaded NPs. In vitro degradation and release were evaluated in 100 µM H₂O₂ solution. The samples were collected and observed by TEM after 3 days. For drug release evaluation, MSNs@RA/LA were dispersed in 5 mL of PBS solution containing 0 or 100 µM H₂O₂ on a shaker. Then, the released amounts of RA and LA were detected in the supernatant at predetermined times by HPLC.

Macrophage membrane derivation and biomimetic nanoplatform fabrication

To fabricate cell membrane vesicles, 1 × 10⁷ RAW264.7 cells were lysed in 20 mL of hypotonic lysing buffer and sonicated at 30 kHz for 2 min by a Vibra-cell VCX 400 probe sonicator. Then, the resulting ghosts were resuspended in 2 mL of deionized water and sonicated for 10 min, followed by extrusion through serial polycarbonate membranes (Whatman) to prepare membrane vesicles. The obtained vesicles were mixed with MSNs@RA/LA via sonication and extruded through 200-nm polycarbonate membranes to fabricate cell membrane-cloaked MSNs@RA/LA (MSNs@RA/LA@CM). To further modify the MSNs@RA/LA@CM with the CD44 aptamer, 100 µM DSEP-PEG-CD44 aptamer was added to 10 mL of an aqueous solution of MSNs@RA/LA@CM (1 mg/mL) and stirred at room temperature for 1 h. Then, the CD44 aptamer-modified MSNs@RA/LA@CM (MSNs@RA/LA@CM-AP) were collected by centrifugation and stored at 4 °C for subsequent experiments.

Cell culture

Hepatic embryo cells HL-7702 and human hepatocarcinoma cells HepG2 were cultured in RPMI-1640 medium supplemented with 10% (v/v) fetal bovine serum under 5% CO₂. Human hepatic stellate cells LX-2 and murine macrophages cell line (RAW264.7) were maintained in dulbecco’s modified eagle medium (DMEM)with 10% (v/v) fetal bovine serum (FBS).

Intracellular lipid deposition measurement

HepG2 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin in a 5% CO₂ atmosphere at 37 °C. HepG2 cells were cultured in the bottom chamber, and sympathetic neurons were cultured in the upper chamber in a Transwell model. To establish the in vitro NAFLD model, 2 mM free fatty acid (FFA) (sodium palmitate: sodium oleate = 1:2) supplemented with MSNs@RA/LA@CM-AP or other formulations was added to the medium of HepG2 cells at 70% confluence. After 48 h, the FFAs were removed, and the cells were washed with PBS three times. Then, the cells were stained with Bodipy and DAPI for 30 min. To measure the intracellular lipid levels, these cells were lysed and the supernatant was collected for detection of the intracellular contents of TC and TG using the corresponding assay kits.

Cellular internalization evaluation

To investigate the aHSC targeting property, LX-2 cells were seeded in a 6-well plate and cultured with TGF-β1 at a concentration of 2 ng/mL for 24 h. Then, Cy3-labeled MSNs@RA/LA, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP (12.5 µg/mL) were incubated with LX-2 cells for 6 h. Subsequently, these cells were stained with DAPI for 20 min, washed with PBS and then detected by a laser confocal assay (Olympus FV1000; Olympus, Tokyo, Japan) and flow cytometry (BD Biosciences, Franklin Lakes, NJ, USA). To investigate immune evasion, Cy3-labeled MSNs@RA/LA, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP were incubated with RAW264.7 cells. After 6 h, the cellular internalization of these nanoparticles in RAW264.7 cells were detected by a laser confocal assay and flow cytometry.

Cytotoxicity evaluation

LX-2 cells, HL7702 cells and HUVECs were seeded at a 96-well plate at a density of 5 × 10³ cells/well overnight. Then, MSNs@RA/LA, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP were added into the cells at various concentration. After 24 h of coincubation, the cell viability was detected using a CCK-8 assay kit.

Animal models

ICR male mice and C57BL/6 male mice (6 weeks) were purchased from Jilin University Animal Experimental Center. All the animal experiments followed a protocol in line with the National Institute of Health Guide for the Care and Use of Laboratory Animals and were approved by the ethical committee of Jilin University Animal Experimental Center (SY202403017). Liver fibrosis models were established using the bile duct ligation (BDL) method according to a previous report [23]. Briefly, male ICR mice were anesthetized, and a 3 cm incision was made in the middle of the abdomen. Then, two adjacent positions of the common bile duct 1 cm from the porta hepatis were ligated. The liver fibrosis models were generated 14 days after surgery. To establish the NAFLD model, C57BL/6 mice were continuously fed a high-fat diet for 10 weeks.

Systemic toxicity

Healthy ICR male mice were intravenously injected with MSNs@RA/LA@CM-AP at a dose of 5 mg/kg every day for 21 days. Then, all the mice were sacrificed. The major organs including the heart, liver, spleen, lung and kidney were harvested, fixed and stained with hematoxylin-eosin (H&E). Blood was collected for analyzing serum biochemical indexes including alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), blood urea nitrogen (BUN), creatinine (CR), cholesterol (TC), total bilirubin (TBIL) and triglyceride (TG) using a Coulter LX2D (Beckman, Brea, CA).

Biodistribution evaluation

Liver fibrosis model mice and normal control mice were intravenously injected with Cy 5.5-labeled MSNs@RA/La@CM or MSNs@RA/La@CM-AP at a dose of 5 mg/kg. All the mice were sacrificed 8 h after administration. Major organs, including the liver, spleen, kidney, lung and heart, were harvested for ex vivo imaging. These organs were weighed and homogenized, and then the fluorescence signals were detected to quantify the biodistribution.

Antifibrotic effect in vitro

LX-2 cells were seeded into a 6-well plate at a density of 2 × 10⁵ cells/well and incubated with TGF-β1 (2 ng/mL) for 24 h. Then, these cells were treated with PBS, MSNs, free RA, free LA, free RA plus LA, MSNs@RA@CM-AP, MSNs@ LA@CM-AP and MSNs@RA/LA@CM-AP for 24 h, respectively. Next, the expression of mRNA levels was detected by RT-PCR. Total collagen was measured by a Sirius Red assay.

ROS scavenging in vitro

LX-2 cells were plated into a 6-well plate at a density of 2 × 10⁵ cells/well and incubated with 100 ng/mL lipopolysaccharide (LPS) to increase the intracellular level of ROS. Afterward, the cells were treated with MSNs@CM-AP and MSNs@RA/LA@CM-AP (12.5 µg/mL) for 24 h and subsequently stained with 2′,7′ -dichlorodihydrofluorescein diacetate (DCFH-DA) probe for 20 min. Then, these cells were washed with PBS and detected using a laser confocal microscope and flow cytometry.

Antifibrotic effect in vivo

Liver fibrosis model mice were intravenously injected with saline, MSNs@CM-AP, RA plus LA, MSNs@RA@CM-AP, MSNs@LA@CM-AP or MSNs@RA/LA@CM-AP every day for 21 days. The dose of MSNs@RA/LA@CM-AP administered was 5 mg/kg. All mice were sacrificed after 21 days. Subsequently, the serum was collected, and liver tissues were harvested. The levels of AST and ALT in the blood and the content of hydroxyproline were tested by commercial ELISA kits. The liver tissues were fixed with 4% formalin and sliced for hematoxylin and eosin (H&E), Sirius Red and immunofluorescence staining.

Anti-steatosis effect in NAFLD models

NAFLD model animals were intravenously administered saline, MSNs@RA@CM-AP, MSNs@LA@CM-AP or MSNs@RA/LA@CM-AP every day for 14 days. The body weights of the mice were recorded every three days. All the mice were anesthetized on Day 15, and liver tissues and serum were collected. The liver tissues were weighed, fixed in 4% formalin and stained with oil red O to measure lipid levels. The levels of AST, ALT, TG, TC and low-density lipoprotein cholesterol (LDL-C) in the blood were measured using commercial ELISA kits.

Statistics

Student’s t-test was used to analyze differences between the two groups. One-way ANOVA (Tukey’s multiple comparison test) or two-way ANOVA (Tukey’s and Sidak’s multiple comparisons test) were used to analyze differences between more than two groups. All results are expressed as mean ± SD. Statistical analysis was conducted by GraphPad Prism 8.0.1.

Results

A previously reported sol-gel method was used to synthesize diselenide-bridged mesoporous silica nanoparticles (MSNs). Electron microscopy images (Fig. 1A) clearly revealed that the prepared MSNs had a uniformly spherical morphology with a size of approximately 80 nm. The N₂ adsorption/desorption isotherms indicated a large surface area (487.2 m²/g), pore volume (0.41cm³/g), and average pore diameter (3.8) of the MSNs (Figure S1), which bestowed a high ability to preload RA and LA. The drug loading and cumulative release of RA/LA-loaded MSNs (MSNs@RA/LA) was measured by HPLC. The encapsulation efficiencies (EE%) of RA and LA was calculated to be 32.8 ± 2.0% and 25.1 ± 1.0%, while the drug loading content (DL%) of RA and Labetalol was 11.6 ± 0.64% and 9.1 ± 0.33%, respectively (Table S1). Considering that the fibrotic liver features excessive ROS, we investigated the ROS-responsive degradation and drug release of MSNs@RA/LA in a solution containing 100 × 10^− 6 M H₂O₂. Due to the cleavage of the diselenide bridge, MSNs@RA/LA gradually collapsed into irregular aggregates in the presence of 100 × 10^− 6 M H₂O₂ at 3 days (Fig. 1B and S2). Consistent with the degradation behavior, the cumulative release of RA and LA reached 61.1% and 66.5%, respectively, in H₂O₂-containing solution, whereas 7.9% RA and 11.2% LA were released from MSNs@RA/LA after 72 h in PBS solution (Fig. 1D and E). The ROS-responsive release was conducive to improving the drug release efficiency in fibrotic livers with high levels of ROS while minimizing drug side effects in normal tissues. Additionally, MSNs gradually consumed H₂O₂ as the concentration and time increased (Fig. 1C and S3), suggesting the ROS-scavenging ability of MSNs during the treatment of NAFLD.

Fig. 1

Characterization of MSNs@RA/LA@CM-AP. (A) TEM image of MSNs. (B) Morphology of MSNs 3 days after incubation with 100 µM H₂O₂. (C) H₂O₂ consumption 3 days after incubated with various concentration of MSNs, n = 4. (D-E) The release behavior of RA (D) and LA (F) from MSNs@RA/LA A under PBS or100 µM H₂O₂, n = 3. (H) Zeta potential analysis of MSNs@RA/LA@CM-AP, n = 3. (G) TEM image of MSNs@RA/LA@CM. (H) Zeta potential analysis of MSNs@RA/LA@CM-AP, n = 3. (G) Size distribution of MSNs@RA/LA@CM-AP, n = 3. Data represent mean ± SD

To prolong the blood circulation time and decrease the capture of Kupffer cells, cell membranes were derived from macrophages and coated on the surface of RA/LA coencapsulated MSNs (MSNs@RA/LA to form MSNs@RA/LA@CM. Subsequently, CD44 aptamer-conjugated DSPE-PEG was directly fused with MSNs@RA/LA@CM using a noncovalent method. As shown in Fig. 1F, the formed biomimetic nanocomposites (MSNs@RA/LA@CM-AP) had a spherical structure with an MSN core enclosed by a thin membrane shell. DLS measurements revealed that MSNs@RA/LA@CM-AP had a slightly larger hydrodynamic size than did bare MSNs@RA/LA, and the surface potential decreased from − 6.2 mV for MSNs@RA/LA to − 21.3 mV due to the coating of negatively charged MSC membranes (Fig. 1G and H). These results indicated the successful coating of the cell membrane onto the nanoparticles. Additionally, the characteristic absorption of the CD44 aptamer at 260 nm was detected in the UV‒vis spectrum of MSNs@CM-AP, confirming the successful modification of the CD44 aptamer (Figure S4). Furthermore, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP displayed relatively good stability after 7 days of stockage in the cell medium, whereas aggregations were found on MSNs@RA/LA without cell membrane coating (Figure S5).

To explore their ability to target aHSCs, we investigated the cellular uptake of Cy3-labeled MSNs@RA/LA@CM-AP by TGF-β1-stimulated LX-2 cells via a laser confocal assay and flow cytometry. After treatment with TGF-β1 (2 ng/mL) for 24 h, the expression of fibrogenic markers such as α-SMA, Col1a1, Col3a1 and Col5a1 was significantly elevated (Figure S6), demonstrating the successful induction of aHSCs. Additionally, aHSCs had higher expression of CD44 than quiescent HSCs (qHSCs), as verified by immunofluorescence staining and flow cytometry (Figure S7). Furthermore, our data showed the internalization of MSNs@RA/LA@CM-AP in aHSCs was remarkably greater than that of MSNs@RA/LA@CM and MSNs@RA/LA (Fig. 2A and C). However, the uptake of MSNs@RA/LA@CM-AP was low in qHSCs, which was comparable to that of MSNs@RA/LA@CM (Fig. 2B and D). Moreover, the high cellular internalization of MSNs@RA/LA@CM-AP with aHSCs was markedly attenuated when 1 mM CD44 antibodies was preincubated with aHSCs for 1 h, further indicating CD44-mediated cellular uptake (Figure S8). Additionally, flow cytometry analysis indicated that MSNs@RA/LA@CM-AP could be endocytosed by HepG2 cells and HL-7702 cells, and the cellular internalization efficiency of MSNs@RA/LA@CM-AP was like that of MSNs@RA/LA@CM (Figure S9). These results indicated the aHSC-specific targeting property of MSNs@RA/LA@CM-AP. To investigate the ability of MSNs@RA/LA@CM-AP to escape macrophages, MSNs@RA/LA, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP were incubated with RAW264.7 macrophages for 6 h. As shown in Figure S10, the uptake of MSNs@RA/LA@CM-AP and MSNs@RA/LA@CM was lower than that of MSNs@RA/LA in RAW264.7 cells, suggesting that the cell membrane coating might decrease the capture of Kupffer cells. Additionally, MSNs@RA/LA@CM-AP and MSNs@RA/LA@CM displayed a markedly increased blood retention than MSNs@RA/LA, which was likely due to the good colloidal stability and the immune-evasion propery of the macrophage membrane (Figure S11). Having demonstrated that the MSNs@RA/LA@CM-AP were able to target aHSCs in vitro, we investigated the tissue biodistribution of MSNs@RA/LA@CM-AP in liver-fibrosis mice induced by bile duct ligation (BDL) and in normal mice. We found that MSNs@RA/LA@CM-AP predominantly accumulated in liver tissues 8 h after intravenous administration (Fig. 2E and G). A fluorescent quantitation assay indicated that MSNs@RA/LA@CM-AP showed greater accumulation efficiency than MSNs@RA/LA@CM in the fibrotic livers, whereas there is no obvious difference of accumulation between MSNs@RA/LA@CM-AP and MSNs@RA/LA@CM in the livers of healthy mice (Figure S12), which could be explained by the greater uptake of MSNs@RA/LA@CM-AP by aHSCs in the fibrotic liver due to the modification of the CD44 aptamer. Consistent results were detected in the ICP-OES assays (Fig. 2F and H). Additionally, fluorescence imaging of fibrotic livers stained with α-SMA revealed a higher colocalization of MSNs@RA/LA@CM-AP with aHSCs than that of MSNs@RA/LA@CM with aHSCs (Figure S13A). Moreover, flow cytometry assays of fibrotic livers indicated MSNs@RA/LA@CM-AP showed higher accumulation in aHSCs than MSNs@RA/LA@CM (Figure S13B), further confirming the excellent aHSCs-targeting property of MSNs@RA/LA@CM-AP. Collectively, these results indicated the excellent ability of MSNs@RA/LA@CM-AP to actively target the fibrotic liver.

Fig. 2

Fibrotic liver-targeting property of MSNs@RA/LA@CM-AP. (A-B) Fluorescence microscopy images of activated HSCs (A) and quiescent HSCs (B) after incubation with MSNs@RA/LA@CM-AP, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP for 6 h, the scale bar = 10 μm. (C-D) Quantitative fluorescence intensity of Cy3-labeled MSNs activated HSCs (C) and quiescent HSCs (D) after treated with MSNs@RA/LA@CM-AP, MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP for 6 h, n = 3. (E) Ex vivo fluorescence images of Cy5.5-labeled MSNs in major organs of the liver fibrosis mice 8 h after treatment with MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP. (F) Quantitative ICP-OES analysis of MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP in the major organs of liver fibrosis mice 8 h after intravenous administration, n = 3. (G) Ex vivo fluorescence images of Cy5.5-labeled MSNs in major organs of the healthy mice 8 h after various treatments. (H) Quantitation of the biodistribution of MSNs@RA/LA@CM and MSNs@RA/LA@CM-AP in the healthy mice 8 h after intravenous administration by ICP-OES, n = 3. *p < 0.05, **p < 0.01, ***p < 0.001. Data are presented as the mean ± SD

The biosafety of nanoparticles is an important concern for medical applications. We evaluated the cytotoxicity of these nanoparticles in LX-2 cells, HL7702 cells and HUVECs. Our data showed that the viability of these cells was above than 80% even when the concentration of MSNs and MSNs@RA/LA@CM-AP reached 100 µg/mL (Figure S14A-C), indicating that MSNs@RA/LA@CM-AP had a low effect on cell survival. We subsequently evaluated the systemic toxicity of MSNs@RA/LA@CM-AP in the mice. Encouragingly, repeated administration of MSNs@RA/LA@CM-AP did not induce significant changes in body weight, blood biochemistry, or organ histopathology compared with those of the control group (Figure S14D-F). These results suggested that MSNs@RA/LA@CM-AP were a low-toxicity nanocarrier system.

To investigate the antifibrotic activity of MSNs@RA/LA@CM-AP, we detected the expression of fibrogenic markers in TGF-β1-activated LX-2 cells after treatment with various formulations. QT-PCR revealed that the expression of α-SMA, Col1a1, Col3a1 and Col5a1 was obviously downregulated following treatment with free RA, free LA, RA plus LA MSNs@RA@CM-AP, MSNs@LA@CM-AP or MSNs@RA/LA@CM-AP (Fig. 3A-D). Consistently, these formulations reduced collagen-1a2 gene expression and collagen secretion of aHSCs (Fig. 3E and F). Additionally, MSNs@RA@CM-AP showed better antifibrotic effects than free RA due to the improved cellular internalization of RA by the nanocarrier system. Notably, compared with MSNs@RA@CM-AP, MSNs@RA/LA@CM-AP induced lower expression of fibrogenic markers and less secretion of collagen, which could be explained by the fact that norepinephrine is an autocrine factor for HSCs and that the loaded LA inhibited the self-produced norepinephrine from promoting the proliferation and activation of HSCs through the autocrine closed loop.

Fig. 3

Anti-Fibrotic effect of MSNs@RA/LA@CM-AP in vitro. (A-D) Quantitative analysis of the expression of fibrosis-related genes including α-SMA (A), Col1a1 (B), Col3a1 (C) and Col5a1 (D) in aHSCs after various treatments, n = 3. (E) Relative collagen-1a2 gene expression in aHSCs after various treatments, n = 3. (F) Quantitative analysis of total collagen secreted from aHSCs after treated with various formulations by Sirius Red assay, n = 3. (G) Flow cytometry analysis of the ROS levels in aHSCs after treated with various formulations, n = 3. (H) Fluorescence microscopy images of total ROS in aHSCs after various treatments, the scale bar = 10 μm. (I-J) The mRNA levels of α-SMA (I) and collagen-1a2 (J) in aHSCs, n = 3. (K-M) Secretion of TNF-α, IL-1β and IL-6 levels from aHSCs after various treatments, n = 3. *p < 0.05, **p < 0.01, ***p < 0.001. Data are presented as the mean ± SD

The excessive production of ROS plays a vital role in activating the HSCs to generate abundant collagen during liver injuries. We expected that our MSNs could scavenge excessive ROS to inhibit HSC activation. As shown in Fig. 3G and H, LPS stimulation significantly increased the intracellular ROS level of LX-2 cells. However, such effect was obviously inhibited by the treatments of MSNs@CM-AP and MSNs@RA/LA@CM-AP. Additionally, both MSNs@CM-AP and MSNs@RA/LA@CM-AP suppressed the increasement of α-SMA and collagen-1a2 expression in LPS-activated LX-2 cells (Fig. 3I and J). Consistent results were detected for the secretion of proinflammatory cytokines including TNF-α, IL-1β and IL-6 in LPS-stimulated LX-2 cells (Fig. 3K-M). Additionally, MSNs@CM-AP showed a similar antifibrotic activity to MSNs and MSNs@CM, which suggested that MSNs were the main factor for fibrosis inhibition in LPS-stimulated LX-2 cells through their ROS consumption (Figure S15). Collectively, these results indicated the anti-oxidant and anti-inflammatory capabilities of MSNs@RA/LA@CM-AP, which might benefit the NAFLD treatment.

To investigate the cellular crosstalk between HSCs and sympathetic neurons, we constructed a coculture system of aHSCs and sympathetic neurons in a Transwell model. As illustrated in Fig. 4A, the number of HSCs was significantly greater after coculture with sympathetic neurons for 72 h than after coculture with free DMEM. Furthermore, increased expression of α-SMA, Col1a1, Col3a1, Col5a1 and collagen-1a2 and increased secretion of collagen were detected in HSCs cocultured with sympathetic neurons (Fig. 4B-E and S16). Additionally, the addition of NE to HSCs cultured with free DMEM also induced an increase in cell viability and fibrosis-related gene expression in HSCs. To further explore the bidirectional influence, we detected the production of sympathetic nerve neurotransmitters in the coculture system. The coculture system secreted a surge of norepinephrine, which was markedly greater than that secreted by single-cultured sympathetic neurons (Figure S17). The increased level of norepinephrine was possibly because that aHSCs not only self-produced norepinephrine through the autocrine closed loop, but also enhanced sympathetic nerve activity through releasing catecholamines to induce more production of norepinephrine from sympathetic neurons [13, 15, 16, 24]. These results suggested that sympathetic neurons increased the proliferation and activation of HSCs. We expected that the undesirable crosstalk between HSCs and sympathetic neurons could be inhibited by the nanoplatform to achieve efficient treatment of hepatic fibrosis. Consistent with our expectation, the proliferation and activation of HSCs were significantly inhibited by the administration of MSNs@LA@CM-AP, MSNs@LA@CM-AP and MSNs@RA/LA@CM-AP (Fig. 4A-F and S16). Moreover, MSNs@RA/LA@CM-AP showed a greater deactivation effect than MSNs@RA@CM-AP and MSNs@LA@CM-AP, and strongly suppressed the proliferation and activation of HSCs after the addition of NE. These results indicated that blockade of the adrenoceptor could improve the HSC deactivation effect of the nanoplatform.

Fig. 4

MSNs@RA/LA@CM-AP interrupts vicious interactions of sympathetic neurons. (A) cell viability of HSCs when cocultured with sympathetic neurons after various treatments, n = 4. (B-D) Quantitative RT-PCR analysis of the expression of α-SMA (B), Col1a1 (C) and Col3a1 (D) of HSCs in the coculture system after various treatments, n = 3. (E) Quantitative analysis of collagen-1a2 gene expression of HSCs after various treatments, n = 3. (F) Total collagen content secreted from HSCs in the coculture system, n = 3. (G) TG levels in steatotic HepG2 cells, n = 3. (H) Bodipy staining showing the lipid accumulation in steatotic HepG2 cells after various treatments, the scale bar = 10 μm. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data are presented as the mean ± SD

To explore the lipid-lowering effect of MSNs@RA/LA@CM-AP, TG and TC levels in FFA-treated HepG2 cells were detected after various treatments. FFA treatment indeed elevated the TG and TC contents in HepG2 cells. However, MSNs@RA@CM-AP induced significant decreases in TG and TC contents similar to MSNs@RA/LA@CM-AP, which was likely because that RA inhibited FFA uptake and lipogenesis (Figure S18). To investigate whether MSNs@RA/LA@CM-AP could decrease cellular lipid deposition by interrupting crosstalk between hepatocytes and sympathetic nerves, sympathetic neurons seeded on the upper chamber and FFA-treated HepG2 cells were layered on the bottom chamber with various formulations. Figure 4G and S19 indicates that the TG and TC levels were further increased in the FFA-treated HepG2 cells when cocultured with sympathetic neurons. However, the upregulation of TG and TC induced by sympathetic neurons was eliminated by MSNs@LA@CM-AP. Notably, MSNs@RA/LA@CM-AP led to the greatest reduction in lipid deposition in FFA-treated HepG2 cells. Consistently, Bodipy staining revealed that the lipid deposition was obviously lower in steatotic HepG2 cells after treatment with MSNs@RA/LA@CM-AP than those treatments with MSNs@RA @CM-AP and MSNs@LA@CM-AP (Fig. 4H and S20). These results suggested that MSNs@RA/LA@CM-AP integrating adrenoceptor antagonist had a beneficial lipid-lowering effect.

We investigated the antifibrotic effect of MSNs@RA/LA@CM-AP in vivo using BDL-induced liver fibrosis models. As shown in Fig. 5A-C and S21, MSNs@CM-AP slightly reduced the α-SMA expression and collagen deposition as well as pathological area in the liver tissue, which was possibly due to the ROS-scavenging effect of MSNs. Furthermore, both MSNs@RA@CM-AP and MSNs@LA@CM-AP partly alleviated liver fibrosis. Compared with treatment with MSNs@RA@CM-AP and MSNs@LA@CM-AP, treatment with MSNs@RA/LA@CM-AP led to a greater decrease in α-SMA expression and collagen deposition, indicating the potential benefits of the blockade of adrenoceptors in the nanoplatform on HSC inactivation. Additionally, MSNs@RA@CM-AP, MSNs@LA@CM-AP and MSNs@RA/LA@CM-AP elicited a striking reduction in liver parenchymal necrosis and the liver fibrotic area compared with saline and MSNs@CM-AP. Furthermore, ALT and AST, the main biochemical markers of liver function, as well as hydroxyproline, an important component of the ECM that reflects collagen deposition and the severity of liver fibrosis, significantly decreased in the MSNs@RA@CM-AP, MSNs@LA@CM-AP and MSNs@RA/LA@CM-AP groups (Fig. 5D-F). Notably, the ability of MSNs@RA/LA@CM-AP to reduce liver fibrosis and damage was greater than that of MSNs@RA@CM-AP and MSNs@LA@CM-AP. These results further confirmed that the combination of MSNs@RA/LA@CM-AP with adrenoceptor antagonists could enhance the treatment of liver fibrosis.

Fig. 5

Anti-Fibrotic effect of MSNs@RA/LA@CM-AP in vivo. (A) Immunofluorescence staining of α-SMA in the fibrotic liver of mice after various treatments, the scale bar = 200 μm. (B) Masson staining of collagen in liver sections of mice after various treatments, the scale bar = 200 μm. (C) H&E staining analysis of pathological changes in liver sections, the scale bar = 25 μm. (D-F) The levels of ALT (D) and AST (E) in serum as well as Hyp (F) in liver of mice after treated with various formulations

NAFLD is considered a result of liver steatosis. The activation of liver sympathetic nerves regulates liver steatosis during obesity. To explore whether MSNs@RA/LA@CM-AP combined with adrenoceptor antagonists could decrease obesity-induced liver steatosis to ameliorate NAFLD, NAFLD mouse models were established through long-term high-fat diet (HFD) feeding, and these mice were subsequently administered MSNs@RA/LA@CM-AP or other formulations (Fig. 6A). As anticipated, the body weight and liver weight were greater in the mice after 10 weeks of HFD feeding than those in the mice fed the normal diet (healthy mice) (Fig. 6B and C). Additionally, the sympathetic nerves of the HFD-fed mice exhibited increased activity compared with those of the normal chow-fed mice, as evidenced by the increased production of norepinephrine (Figure S22). These findings further indicated that HFD-induced NAFLD is associated with liver sympathetic nerve overactivity. Notably, all treatments, especially MSNs@RA/LA@CM-AP, significantly decreased the body weight, liver weight and liver/body weight ratio of HFD-fed mice (Fig. 6B-D). Liver lipid deposition is an important issue in NAFLD; therefore, lipid levels were detected to investigate the lipid-lowering efficacy of MSNs@RA/LA@CM-AP. The levels of TC, TG and LDL-C in the liver and serum of mice after treatment with MSNs@RA/LA@CM-AP were lower than those in the liver and serum of mice after treatment with MSNs@RA@CM-AP and MSNs@LA@CM-AP (Fig. 6E-F and Figure S23). Additionally, oil red O staining of neutral lipids revealed that MSNs@RA/LA@CM-AP led to the least lipid droplet accumulation in the cytoplasm of liver tissues, indicating that MSNs@RA/LA@CM-AP exerted a lipid-lowering effect (Fig. 6I). Furthermore, the ALT and AST levels in the blood were the lowest in the MSNs@RA/LA@CM-AP group among other treatment groups in the NAFLD mouse models, which were approximate to those in normal mice (Fig. 6G and H). These results confirmed the excellent ability of MSNs@RA/LA@CM-AP to reduce liver steatosis and damage, suggesting its tremendous potential in NAFLD therapy.

Fig. 6

Therapeutic effect of MSNs@RA/LA@CM-AP in HFD-induced mouse models. (A) Schemic illustration of the treatment timeline for the HFD-induced NAFLD models. (B) Body weight changes during the treatment, n = 5. (C) liver weight changes, n = 5. (D) liver/body weight ratio, n = 5. (E-F) Levels of TG and TC (F) in the liver tissues at the end of various treatments, n = 5. (G-H) Serum ALT (G) and AST (H) were detected at the end of various treatments, n = 5. (I) Oil red O staining analysis of liver sections of mice after various treatments, the scale bar = 50 μm

NAFLD is a chronic disease associated with steatosis, inflammation and fibrosis. Effective treatments of NAFLD remained a great challenge due to the limited available therapeutic options in the clinic. Increasing evidence has revealed that hepatic sympathetic innervation promoted the activation and proliferation of HSC and regulated liver lipid acquisition during obesity through adrenoceptors expressed by HSCs and hepatocytes. Therefore, adrenoceptors might be promising targets for novel approaches to treat NAFLD. In the current study, we hypothesized that blockade of adrenoceptors for breaking the vicious crosstalk of sympathetic nerves in the liver might provide benefits to the treatment of NAFLD. Herein, a biomimetic nanocarrier was tailored to co-deliver RA, a commonly used anti-NAFLD agent and LA, an adrenoceptor antagonist. The biomimetic nanocarrier itself consumed the excessive ROS at the pathologic site, which contributed to decrease the inflammatory response and inhibit the fibrosis progression. Furthermore, the biomimetic nanocarrier possessed aHSC-targeting property and ROS-responsiveness release as well as Kupffer cell-evading ability, which facilitated the drug delivery and improved the therapeutic efficiency of RA for NAFLD.

Norepinephrine, as an autocrine factor produced from HSCs, can also promote the proliferation of HSCs. The MSNs@RA/LA@CM-AP integrating the blockade of adrenoceptors effectively inhibit the stimulatory effects of self-produced norepinephrine on the HSC proliferation. Importantly, we found that our prepared MSNs@RA/LA@CM-AP intercepted cellular crosstalk of sympathetic neurons with HSCs, which effectively inhibited the activated and proliferative effect of sympathetic neurons on HSCs. Therefore, MSNs@RA/LA@CM-AP showed great ability to reduce liver fibrosis and damage in BDL-induced liver fibrosis animal models.

There are no obvious differences in the lipid accumulation of steatotic HepG2 cells after treated with MSNs@RA@CM-AP and MSNs@RA/LA@CM-AP, which was possibly because the lipid-lowering effect was mainly originated from the reduction of FFA uptake and lipogenesis by RA [20]. However, MSNs@RA/LA@CM-AP exhibited better lipid-lowering effect in steatotic HepG2 cocultured with sympathetic neurons and in HFD-induced NAFLD animal models. Norepinephrine can act on the overexpression of adrenoceptors in hepatocytes, resulting in an increase in lipid deposition according to the previous report [25]. Therefore, MSNs@RA/LA@CM-AP integrating adrenoceptor antagonist exhibited excellent ability to reduce liver steatosis, which further indicated that our nanoplatform regulated all stages of NAFLD involving steatosis, inflammation and fibrosis.

Conclusion

In summary, we created a biomimetic fibrotic liver-targeting nanoplatform to simultaneously deliver antifibrotic agents and adrenoceptor antagonists for the treatment of NAFLD. The biomimetic nanoplatform exhibited a prolonged blood circulation time and reduced Kupffer cell capture owing to macrophage membrane cloaking, as well as high accumulation efficiency in the fibrotic liver as a result of the CD44 aptamer modification. Additionally, the biomimetic nanoplatform exhibited ROS-responsive degradation and release properties as well as ROS-scavenging ability, which not only facilitated drug delivery in the microenvironment of the fibrotic liver, but also suppressed oxidative stress of injured liver tissues to enhance therapeutic outcome. Futhermore, the biomimetic nanoplatform with adrenoceptor antagonists effectively interrupted the malignant interactions between HSCs and sympathetic nerves and thereby showed a greater ability to inhibit the proliferation and activation of HSCs than the corresponding nanoplatform without adrenoceptor antagonists. Using a liver fibrosis model induced by BDL, we demonstrated that the biomimetic nanoplatform effectively decreased collagen deposition and fibrogenesis-associated biomarker expression, and restored hepatic function. Additionally, the nanoplatform led to the greatest decrease in LDL-C, TG, and TC levels and liver lipid droplets in HFD-fed mice, confirming its excellent lipid-lowering effect. Overall, the biomimetic nanoplatform integrating the blockade of adrenoceptors had a synergistic effect or an additive effect on ameliorating liver fibrositis and lipid deposition. Our work indicates that adrenoceptors may be potential targets for novel therapeutic approaches in the clinical treatment of NAFLD.