Enhanced protection against hypoxia/reoxygenation-induced apoptosis in H9c2 cells by puerarin-loaded liposomes modified with matrix metalloproteinases-targeting peptide and triphenylphosphonium

Abstract Based on the inhibition of mitochondrial permeability transition pore (mPTP) opening, puerarin (PUE) has a good potential to reduce myocardial ischemia/reperfusion injury (MI/RI). However, the lack of targeting of free PUE makes it difficult to reach the mitochondria. In this paper, we constructed matrix metalloproteinase-targeting peptide (MMP-TP) and triphenylphosphonium (TPP) cation co-modified liposomes loaded with PUE (PUE@T/M-L) for mitochondria-targeted drug delivery. PUE@T/M-L had a favorable particle size of 144.9 ± 0.8 nm, an encapsulation efficiency of 78.9 ± 0.6%, and a sustained-release behavior. The results of cytofluorimetric experiments showed that MMP-TP and TPP double-modified liposomes (T/M-L) enhanced intracellular uptake, escaped lysosomal capture, and promoted drug targeting into mitochondria. In addition, PUE@T/M-L enhanced the viability of hypoxia-reoxygenation (H/R) injured H9c2 cells by inhibiting mPTP opening and reactive oxygen species (ROS) production, reducing Bax expression and increasing Bcl-2 expression. It was inferred that PUE@T/M-L delivered PUE into the mitochondria of H/R injured H9c2 cells, resulting in a significant increase in cellular potency. Based on the ability of MMP-TP to bind the elevated expression of matrix metalloproteinases (MMPs), T/M-L had excellent tropism for Lipopolysaccharide (LPS) -stimulated macrophages and can significantly reduce TNF-α and ROS levels, thus allowing both drug accumulation in ischemic cardiomyocytes and reducing inflammatory stimulation during MI/RI. Fluorescence imaging results of the targeting effect using a DiR probe also indicated that DiR@T/M-L could accumulate and retain in the ischemic myocardium. Taken together, these results demonstrated the promising application of PUE@T/M-L for mitochondria-targeted drug delivery to achieve maximum therapeutic efficacy of PUE.


Introduction
Acute myocardial infarction (AMI) is one of the common causes of mortality and disability worldwide (Johansson et al. 2017, Wang et al. 2021).Critically ill patients are usually treated with early reperfusion using percutaneous coronary intervention to restore the blood oxygen supply of blocked vessels in time (Fordyce et al. 2015, Baeza-Herrera et al. 2020).However, the process of coronary reperfusion may paradoxically induce myocardial ischemia-reperfusion injury (MI/RI) (Chouchani et al. 2014, Maneechote et al. 2017, Mui and Zhang 2021).Currently, there are still no specific therapies to prevent MI/RI (Cheng et al. 2019).Therefore, there is a pressing need to develop effective drugs and technologies to address this issue.
Although the exact mechanism of MI/RI has not been fully clarified, most researchers have indicated that these facts of the continuous explosion of reactive oxygen species (ROS) (Jiang et al. 2021), overload of mitochondrial calcium, and continuous opening of mitochondrial permeability transition pore (mPTP) (Miao et al. 2019, Li et al. 2019b) could induce mitochondrial dysfunction.Mitochondrial dysfunction allows cytochrome C to be released from the mitochondria into the cytosol, triggering a cascade effect of Caspases and inducing apoptosis in cardiomyocytes (Zhou et al. 2021).Therefore, mitochondria happen to be the key regulators of cardiomyocyte survival and mitochondria-targeted therapeutic strategies could be promising for preventing MI/RI.
Puerarin (PUE) is the main active ingredient extracted from Pueraria lobata (Willd.)Ohwi, which has good efficacy in reducing MI/RI (Xu et al. 2019).There is increasing evidence that the decreasing ischemia myocardium of PUE is mediated through regulating mitochondrial function (Gao et al. 2005(Gao et al. , 2006)).PUE has been shown to inhibit the opening of mPTP and reduce the production of ROS and inflammatory cytokines, consequently inhibiting the mitochondria-dependent apoptosis pathway (Gao et al. 2006, Wang et al. 2018).Therefore, efficient targeted delivery of PUE to the mitochondria of ischemic cardiomyocytes is expected to achieve better therapeutic effects.
In recent years, triphenylphosphonium (TPP) has been widely used as a mitochondrial targeting carrier to improve therapeutic effects (Ye et al. 2022).The cationic TPP group could impart a delocalized charge and lipophilic character to a compound that is favorable for mitochondrial targeting (Biswas et al. 2012).In recent years, based on the mitochondrial targeting of TPP, series of drug delivery systems such as Tan-TPP and ANP/TPP-BN-LPNs have been developed, which can deliver drugs to mitochondria and significantly improve the effectiveness of drugs in reducing MI/RI.(Zhao et al. 2019, Wang et al. 2021).However, effective accumulation of nanocarriers in the heart, which is a prerequisite for achieving mitochondrial targeting of drugs, remains a major challenge.This is partially due to the relatively low permeability of coronary blood vessels (Horwitz et al. 1994) and the extremely rapid blood flow during reperfusion (Wexler et al. 1968), leading to the heart's poor retention of therapeutic drugs for a prolonged time.Ischemic myocardial targeted drug delivery systems have the potential to overcome these barriers.
The available literature indicates that matrix metalloproteinases (MMPs), especially MMP-2, are significantly elevated after myocardial ischemia (MI) (Nguyen et al. 2015, Fan et al. 2019).When cardiomyocytes are damaged or necrotic, the cells release large amounts of inflammatory cytokines to recruit macrophages.Macrophages then synthesize large amounts of MMPs that infiltrate into the site of necrotic ischemic myocardium, resulting in a significant elevation of MMPs in the ischemic myocardium.(Nguyen et al. 2015).Thus, the overexpression of MMPs is positively correlated with MI severity, which is considered a prognostic marker of MI (Halade et al. 2013).Nguyen et al designed a micellar vehicle containing a matrix metalloproteinases-targeting peptide (MMP-TP, GGGGCTTHWGFTLC).Their research result showed that the MMP-TP micelles could preferentially target the infarcted myocardium of the heart in an MMP-dependent manner (Nguyen et al. 2015).The MMP-TP could guide the lipid micelles to accumulate in the ischemic myocardial areas rich in matrix metalloproteinases and infiltrating macrophages (Spinale 2007).It is hypothesized that MMP-TP can guide drug carriers into the inflammatory sites of ischemic myocardium.
Based on the ischemic myocardial targeting of MMP-TP and the mitochondrial targeting function of TPP, we construct a stepwise targeted delivery system consisting of MMP-TP and TPP, anticipating that the nanoparticles would first aggregate around macrophages in the ischemic myocardium and then allow the therapeutic drug to enter the mitochondria to modulate mitochondrial function (Scheme 1).Invitrogen (Carlsbad, USA).Rabbit anti-mouse MMP-2 antibody was obtained from ABclonal Technology.The ELISA kits for TNF-a was purchased from Cusabio (Wuhan, Hubei, China).H9c2 cells derived from rat myocardium and murine Raw 264.7 macrophage cells were purchased from the American Type Culture Collection (Manassas, VA).Cell cultures were maintained in a humidified atmosphere of 5% CO 2 at 37 C.All other chemicals and reagents were analytical grade or chromatography grade.

Synthesis of functional conjugates
The synthesis of MMP-TP-PEG-PE conjugate was performed following the previously published procedure (Nguyen et al. 2015).Briefly, 100 mg of DSPE-PEG-NHS and 50 mg of MMP-TP were dissolved in 10 ml of anhydrous N-methyl-2-pyrrolidone (NMP), and then 50 mL of triethylamine was added.The mixture was stirred at room temperature for 24 h.The reaction solution was transferred into a dialysis bag (4KD) and dialyzed in pure water for 24 h to remove the unreacted materials and byproducts.The dialysate was freeze-dried to obtain the crude product of MMP-TP-PEG-PE.Then MMP-TP-PEG-PE was further purified by preparative liquid phase using a 70 min gradient from 30 to 100% acetonitrile in the presence of 0.1% trifluoroacetic acid (TFA).The structures of the synthesized MMP-TP-PEG-PE polymer were confirmed by 1 H NMR and MALDI-TOF mass spectrometry (MALDI-TOF MS).TPP-PEG-PE was prepared by the method reported in our previous literature (Li et al. 2019a).Briefly, DSPE-PEG-NH 2 was reacted with CTPP (1:1.5 molar ratio) in chloroform containing EDC and NHS in the dark.The mixture was gently stirred overnight at room temperature under argon.The reactants were then dialyzed in a dialyzed bag to remove the unreacted materials and byproducts.Finally, the solution was freeze-dried and stored at À20 C. The structures of the synthesized TPP-PEG-PE polymer were confirmed by 1 H NMR spectroscopy.

Preparation and physico-chemical characterization of liposomes
PUE-loaded liposomes co-modified by TPP and MMP-TP (PUE@T/M-L), PUE-loaded liposomes modified by TPP (PUE@T-L), PUE-loaded liposomes modified by MMP-TP (PUE@M-L), PUE-loaded liposomes modified by none (PUE@L), all of which were obtained by thin film hydration.In brief, soybean lecithin (SLC), cholesterol (Chol), TPP-PEG-PE, and MMP-TP-PEG-PE were mixed in a molar ratio of 52:26:11:11.The mixture and PUE were well dissolved in chromatographic methanol and then evaporated to form a thin film at the bottom of the vials.The film was further dried under a high vacuum overnight to remove the traces of solvents.Then the dried lipid film was hydrated with distilled water or PBS and incubated in the water bath at 37 C for 30 min.The mixture was ultrasonicated for 5 min to ensure proper resuspension of the film.The final PUE@T/M-L was then extruded 20 times by passage through a 100 nm pore-sized polycarbonate membrane using Avanti Mini-Extruder.PUE@T-L, PUE@M-L, and PUE@L were prepared following the above procedures.The corresponding lipid composition of all PUE-loaded liposomes were shown in Table 1.The morphology of PUE@T/M-L was observed by the JEM 2100 transmission electron microscope (TEM, JEOL Ltd., Japan).The nanoparticle size (Diameter, nm) and surface charge (Zeta potential, mV) of liposomes were measured by the dynamic light scattering (DLS) using Malvern Zetasizer Nano ZS (Malvern Instruments Ltd, UK) at 25 C.The size distribution of liposomes was analyzed using a NanoSight NS300 (Malvern Instruments Ltd, UK).To study the cellular uptake of drug-loaded liposomes by H9c2 cells, the coumarin-6 (fluorescent material) loaded unmodified (C6@L) or modified liposomes (C6@T-L, C6@M-L, C6@T/M-L) was prepared by the similar process as that for PUE-loaded liposomes.

Fourier transform infra-red spectrometer (FT-IR) analysis
A Thermo Fisher Nicolet iS10 Fourier transforms infra-red spectrometer (FT-IR) was used to characterize the status of PUE in PUE@T/M-L.Free PUE, blank liposomes (T/M-L) and PUE@T/M-L were all dispersed on KBr sheets, and FT-IR spectra of these samples were recorded from 400 to 4000 cm-1 by the KBr spheroid method on an FT-IR spectrometer at room temperature.

Entrapment efficiency (EE) and drug loading (DL)
HPLC was used to analyze the drug loading efficiency (DL%) and entrapment efficiency (EE%) for PUE in PUE@T/M-L.The drug DL% and EE% of liposomes were calculated by the following equations.The content assay of the drugs was determined by the HPLC system using acetonitrile: water (18:82) as the mobile phase at the flow rate of 1 ml/min with the detection at 250 nm.

Stability analysis
The PUE@T/M-L was stored at 4 C for 60 days in a dark environment.The stability of the PUE@T/M-L was evaluated by observing the physical characteristics (precipitation or crystal growth, particle size, and zeta potential) of the systems during the storage time.The drug loading and encapsulation rate of PUE@T/M-L before and after storage were determined by HPLC.Moreover, the serum stability of PUE@T/M-L was incubated with 10% fetal bovine serum (FBS).The changes of PUE@T/M-L in particle size and zeta potential at 0, 4, 8, 12, and 24 h were monitored by DLS.

In vitro drug release
The in vitro release study of PUE from drug-loaded liposomes was investigated using dialysis technology.Briefly, free PUE, PUE@L, PUE@T-L, PUE@M-L, and PUE@T/M-L at the same amount of 230 lg PUE was sealed in a dialysis bag (MW.Cut-off: 3 kDa).The bag was then directly submerged in 50 ml of PBS (pH 7.4) while being shaken at 100 rpm.The temperature was maintained at 37 C during the experiment.At predetermined time points, 150 lL aliquots were withdrawn from the release medium and replaced with the same volume of fresh PBS.The concentration of PUE in samples was then analyzed by the HPLC method as described in the above section.The accumulated release was calculated using the following equation: where M PUE represents the amount of PUE in the liposomes, V t represents the whole volume of the release media (V t ¼ 50 ml), V 0 is the volume of the replaced media, and C n represents the concentration of PUE in the n-th sampling point.

Hemolysis assay
A hemolysis experiment was used to reflect the in vivo safety of PUE@T/M-L after intravenous injection.Red blood cells (RBCs) of the rat were repeatedly washed with PBS until the supernatant became clear.The purified RBCs were then mixed with free PUE, PUE@L, PUE@M-L, PUE@T-L and PUE@T/M-L at a concentration of PUE ranging from 18.75 to 300 lg/mL.Distilled water and PBS were used as the positive and negative control, respectively.The mixture was incubated at 37 C for 2 h and then centrifuged at 2000 rpm for 5 min.The absorbance values (A) of the supernatants at 540 nm were measured using a Multiskan MK3 microplate reader.The hemolysis ratio (HR %) was calculated as the following equation: Hypoxia/reoxygenation injured H9c2 cells model

Cellular uptake
Considering the non-fluorescent property of PUE, C6 was used as the classic fluorescent labeling dye incorporated into the liposomes for observation of cellular uptake.H9c2 cells were seeded into 6-well plates (1 Â 10 5 cells/well).After hypoxia for 8 h, cells were exposed to the DMEM with free C6, C6@L, C6@T-L, C6@M-L or C6@T/M-L (C6-equivalent dose: 50 ng/mL) under the normal cell culture conditions.After 1 h and 4 h incubation, cells were washed three times with PBS (pH 7.4) to eliminate residual C6 outside the cells.The nuclei of the cells were then labeled with DAPI staining for 10 min, followed by an OLYMPUS IX73 inverted microscope (Olympus, JAPAN) to obtain fluorescent images of the cells.To investigate quantitatively cellular uptake, the intracellular C6 fluorescence intensity of each H9c2 cell after incubation with the drug was measured using flow cytometry (BD LSR II, Biosciences).
The research methods reported in the literature were adopted to investigate the endocytic pathway of C6@T/M-L (Zhang et al. 2019).Sucrose (inhibitor of clathrin-mediated uptake, 150 mg/mL), 2-Deoxy-D-glucose-sucrose (ATP depletion, 1 mg/mL), colchicine (inhibitor of macropinocytosis, 0.8 mg/mL), methyl-b-cyclodextrin (inhibitor of caveolae-mediated uptake, 0.005 mg/mL) were respectively added to H/R injured H9c2 cells, and the cells were incubated for 2 h at 37 C in hypoxic culture medium.Then, the cells were cultured with fresh DMEM containing C6@T/M-L and incubated for 4 h.The cellular uptake of C6 in each H9c2 cell was quantitatively analyzed by flow cytometry following the above method.

Intracellular distribution and co-location analysis
Coumarin 6 labeled C6@T/M-L was used to investigate the subcellular distribution of PUE delivered by PUE@T/M-L.H9c2 cells were seeded at 12-well plates (1 Â 10 5 cells/well) and cultured for 12 h.After hypoxia for 8 h, the cells were cultured with fresh DMEM containing free C6, C6@L, C6@T-L, C6@M-L or C6@T/M-L for 1 h and 4 h.Then, the cells were washed with PBS three times, fixed with 4% paraformaldehyde for 20 min and stained with Mitotracker Red (200 nM) or Lysotracker Red (60 nM).Finally, the cells were incubated with DAPI solution (5 lg/mL) or Hoechst 33342 (5 lg/mL) for 10 min.The distribution of C6 labeled liposomes in the cells was observed using an OLYMPUS IX73 inverted microscope (Olympus, JAPAN) and the colocation analysis was finished by the Image pro-plus 6.0 Image J software.

Determination of cell viability and cell apoptosis
The cell viability was determined by CCK-8 assay and apoptosis was analyzed by Annexin V-FITC/PI staining assay.For cell viability, H9c2 cells were incubated in hypoxia for 8 h and then incubated with the drug in reoxygenation.The cells were treated with fresh DMEM containing free PUE, PUE@L, PUE@T-L, PUE@M-L or PUE@T/M-L (PUE equivalent dose: 20 lM) for 12 h.Then the H9c2 cells were added to 10% CCK-8 solution and incubated for 2 h at 37 C. Optical density (OD) at 450 nm was detected by a microplate reader (Thermo Fisher Scientific, Inc.).Cell viability was detected using the CCK-8 assay kit (Meilunbio, DaLian) according to the manufacturer's protocol.For cell apoptosis, H9c2 cells were exposed to H/R and treated with free PUE, PUE@L, PUE@T-L, PUE@M-L, or PUE@T/M-L as described above.Then the cells were collected and washed with PBS.Following this, the cells were treated with Annexin V-FITC/PI and analyzed by flow cytometry.

Determination of mPTP opening
Calcein AM and CoCl 2 were used to analyze the opening of mitochondrial permeability transition pores (mPTP).H9c2 cells were planted and cultured as described above.After hypoxia for 8 h, the cells were cultured with fresh DMEM containing free PUE, PUE@L, PUE@T-L, PUE@M-L or PUE@T/M-L (PUE-equivalent dose: 20 lM) for 12 h, and then the cells were incubated with 1.0 mM of calcein-acetomethoxy ester (Molecular Probes) and 1.0 mM CoCl 2 in HBSS for 20 min at 37 C. Finally, the fluorescence intensity of H9c2 cells was then measured by flow cytometry.

ROS scavenging capacity
2 0 ,7 0 -dichlorodihydrofluorescein diacetate (DCF-DA) was used to monitor intracellular ROS levels.Briefly, H9c2 cells were exposed to hypoxia for 8 h and treated with free PUE, PUE@L, PUE@T-L, PUE@M-L or PUE@T/M-L as described above.Then 10 lM DCF-DA was added and incubated with H9c2 cells for 30 min.The cells were washed thrice with PBS and the 2 0 ,7 0 -dichlorofluorescein (DCF) fluorescence was observed via fluorescence microscopy.

Western blot analysis
Apoptosis-related proteins were analyzed using a western blot.H9c2 cells were exposed to hypoxia for 8 h and treated with PUE@L, PUE@T-L or PUE@T/M-L at an equal concentration of 20 lM PUE for 12 h.The total protein was extracted from each group.Denatured proteins were separated by dodecyl sulfate sodium-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to polyvinylidene difluoride (PVDF) membranes.Membranes were blocked with 5% milk solution and then incubated with primary antibodies of Bcl-2 and Bax overnight at 4 C.After incubation with the corresponding secondary antibodies, specific protein bands were detected using ChemiDoc XRS with Quantity One software (Bio-Rad, Hercules, CA, USA).
Colocalization of MMP-TP modified liposomes with MMP-2 RAW 264.7 cells were treated with 100 ng/mL LPS for 24 h to release inflammatory cytokines and induce the cells to increase their MMP secretion.The binding of MMP-TP modified liposomes to MMP-2 was analyzed using immunofluorescence analysis.Briefly, LPS-activated RAW 264.7 cells were incubated with free C6, C6@L, C6@T-L, C6@M-L, and C6@T/M-L (C6-equivalent dose: 50 ng/mL) for 2 h, then incubated with Rabbit anti-mouse MMP-2 antibody, followed by the fluorescent secondary antibody.MMP-2 was labeled with the MMP-2 antibody and the liposomes were labeled with C6.The binding of green fluorescent liposomes to red fluorescent MMP-2 was observed via Immunofluorescence microscopy and the colocation analysis was finished by the Image pro-plus 6.0 Image J software.
Binding of MMP-TP modified liposomes to LPS-Activated RAW 264.7 cells RAW 264.7 cells were incubated with media containing 100 ng/mL LPS for 24 h, after which the medium was replaced with fresh media containing free C6 and four types of C6-labeled liposomes (C6@L, C6@T-L, C6@M-L and C6@T/M-L).The cells were further incubated for 1-4 h and then washed three times.To verify whether MMP-TP modified liposomes fused with LPS-activated RAW 264.7 Cells, we labeled LPS-activated RAW 264.7 cells with Dil (cell membrane-specific fluorescent probe).Similarly, we used DAPI-labeled nuclei to study the cellular uptake of MMP-TP modified liposomes by RAW 264.7 cells treated with LPS.
Flow cytometry was used to quantify the binding and internalization of free C6, C6@L, C6@T-L, C6@M-L and C6@T/M-L to LPS-Activated RAW 264.7 Cells.The LPS-induced RAW 264.7 cells inflammation model was the same as above, and then the cells were co-incubated with liposomes for 2 h, followed by thorough washing at least three times.Finally, the intracellular C6 fluorescent intensity in each RAW 264.7 cell was quantitatively analyzed by flow cytometer.

Attenuation effects of MMP-TP modified liposomes on inflammation in vitro
Stimulation of RAW264.7 cells with 100 ng/mL LPS for 24 h induced inflammation in the cells, resulting in increased intracellular levels of TNF-a.The cells were then co-incubated with free PUE, PUE@L, PUE@T-L, PUE@M-L and PUE@T/M-L (PUE-equivalent dose: 20 lM) for 12 h.The generation of intracellular TNF-a was measured via ELISA assay according to the manufacturer's instructions.Furthermore, LPS is a typical ROS inducer, and we use LPS to stimulate macrophages to produce large amounts of ROS according to the previously reported methods (Gao et al. 2020).Briefly, RAW 264.7 cells were pre-incubated with free PUE, PUE@L, PUE@T-L, PUE@M-L and PUE@T/M-L (PUE-equivalent dose: 20 lM) for 2 h before treated with 400 ng/mL LPS for 6 h, resulting in increased intracellular levels of ROS.Referring to the above method for the determination of ROS level in H9c2 cells, the ROS level of each group was determined via fluorescence microscopy using DCF-DA as the sensor.

Ex vivo biodistribution determined by fluorescence imaging
The animal experiment was approved by the Animal Ethics Committee at the Second Xiangya Hospital of Central South University and performed in conformed with the guidelines of the National Act on the Use of Experimental Animals (China).male BALB/c mice were purchased from the Hunan Slack Scene of Laboratory Animal Co., Ltd.According to the literature and our previously reported methods (Cheng et al. 2015, Li et al. 2019a), the induction of myocardial ischemia in mice was performed by subcutaneous injection of isoprenaline (ISO) (100 mg/kg) at an interval of 24 h for 2 days.It has been reported and our previous study also confirmed that ISO has deleterious cardiac effects in mice, including mitochondrial alterations, oxidative damage, Macrophage infiltration, and morphological changes of cardiomyocytes, which is similar to that in the infarcted human heart (Nichtova et al. 2012, Li et al. 2019a).DiR, a near-infra-red fluorescent probe, was encapsulated into normal liposomes (DiR@L) and TPP and MMP-TP co-modified liposomes (DiR@T/M-L) according to the similar method described above.Free DiR solution, DiR@L and DiR@T/M-L at DiR-eq dose were injected into the myocardial ischemia mice tail veins.Then, the mice were scanned at 6, 12, 18 and 24-h post-injection using an AniView 100 Fluorescence imaging device at appropriate wavelengths (DiR, EX: 748 nm, EM: 780 nm) and then were sacrificed.Their major organs (heart, liver, spleen, lung, and kidney) were harvested for ex vivo imaging.

Histological evaluation of liver samples
Myocardial ischemia was induced in mice by subcutaneous injection of ISO as described above.Free PUE, PUE@L and PUE@T/M-L were then administered to myocardial ischemic mice in equal doses by tail vein injection, and the control group was injected with equal volumes of saline.Liver tissues were taken and immediately fixed in 4% paraformaldehyde, embedded in paraffin wax and cut into 5 lm thick sections, then stained with hematoxylin-eosin, and finally the liver tissue sections of each group were observed by light microscopy.

Statistical analysis
The results were reported as the mean ± standard deviation (SD) from triplicate determinations.Comparisons among multiple groups were assessed for significance using a one-way analysis of variance (ANOVA).The criterion for statistical significance was at Ã p < 0.05 and ÃÃ p < 0.01.

Synthesis and identification of MMP-TP-PEG-PE and TPP-PEG-PE copolymer
The characteristic signals of aromatic protons of matrix metalloproteinases-targeting peptide (MMP-TP) at 7.0-8.7 ppm appeared in the 1 H NMR spectrum of MMP-TP-PEG-PE, indicating that the MMP-TP was successfully linked to DSPE-PEG-NHS to form MMP-TP-PEG-PE (Figure S1a).MALDI-TOF mass spectrum was also carried out to characterize and validate the synthetic product.MMP-TP-PEG-PE (molecular weight 4400 Da) was successfully synthesized using DSPE-PEG-NHS (molecular weight 3004 Da) and MMP-TP (molecular weight 1396 Da), by an acylation reaction between succinimide and amino group (Figure S1b).Peaks of parental DSPE-PEG-NHS completely vanished and shifted to the right side, the increased molecular weight was approximately the molecular weight of the MMP-TP.The results indicated that the DSPE-PEG-NHS could be quantitatively reacted with MMP-TP according to the equal molar ratio.
TPP was conjugated to DSPE-PEG-NH 2 to construct the targeting copolymer, TPP-PEG-PE.As shown in Figure S1c, the 1 H NMR spectrum of polymer synthetic product showed the characteristic groups of the -CH 2 -CH 2 -Ogroup of PEG (3.65 ppm) and aromatic rings group of TPP (7.75 ppm), indicating the successful synthesis of TPP-PEG-PE.

Characterization of liposomes
The lipid compositions, particle size and zeta potential of the resulting PUE@T/M-L, PUE@T-L, PUE@M-L and PUE@L were shown in Table 1.Functionalized liposomes and PUE-loaded liposomes exhibited particle sizes in the range of 117.0-144.9nm.Among them, DLS analysis revealed that the mean diameters of PUE@T/M-L were approximately 20 nm higher than that of other liposomes due to the surface modified with TPP-PEG-PE and MMP-TP-PEG-PE.In addition, NanoSight was also used to investigate the distribution profile of PUE@T/M-L and revealed a peak of 122 nm (Figure 1(b)).As shown in the TEM images, the morphology of PUE@T/M-L exhibited a roughly spherical structure (Figure 1(a)).
The size of liposomes observed in the TEM images was slightly smaller than the diameter measured by DLS.The reason is that DLS gives the hydrodynamic diameter of liposomes whereas TEM results in the size of liposomes in a dried state.(Figure 1(c)).Overall, they were small enough (less than 200 nm) to avoid detection and elimination by the reticuloendothelial system (RES) in the blood circulation (Raza et al. 2016, Li et al. 2018).This facilitates nanoparticle aggregation in ischemic myocardium via the enhanced permeability and retention (EPR) effect (Dong et al. 2017, Li et al. 2019a).Even though there was not much difference in the mean diameters of the liposomes, the zeta potential of the liposomes changed significantly upon the addition of TPP-PEG-PE.The mean zeta potential of PUE@L was À 40.2 ± 0.6 mV, while the PUE@T-L and PUE@T/M-L were À18.1 ± 0.1 mV and À19.4 ± 0.5 mV, respectively (Table 1, Figure 1(d)).The increased zeta potential was possibly due to the presence of cationic TPP on the surface of PUE@T/M-L or PUE@T-L (Li et al. 2019a).The drug-loading content (LC) and encapsulation efficiency (EE) of PUE@T/M-L were 6.2 ± 0.1% and 78.9 ± 0.6%, respectively.
The drug entrapment into the lipophilic core of PUE@T/M-L was confirmed by the analysis of FT-IR. Figure S2 shows the FT-IR spectra of PUE, blank T/M-L and PUE@T/M-L.The FT-IR spectra of PUE@T/M-L were similar to that of the blank T/M-L.In contrast, the characteristic peaks of PUE at 1515.34, 1397.40, 836.49, and797.96cm À1 were hardly seen in PUE@T/M-L.This result indicated that PUE was successfully entrapped into the lipophilic core of PUE@T/M-L.

Stability assay
Stability is an important property of liposomes, especially for functionally modified liposomes.The stability of PUE@T/M-L after storage at 4 C for 60 d or in the presence of 10% FBS was evaluated.As shown in Figure S3, the size and zeta potential of PUE@ T/M-L remained almost unchanged for two months at 4 C, without any signs of precipitation or crystal growth, indicating that PUE@T/M-L had considerably high thermodynamic stability (Wu et al. 2013).The drug loading and encapsulation rate of PUE@T/M-L were also basically unchanged before and after storage, thus indicating that PUE@T/M-L was relatively stable during storage.Moreover, the PUE@T/M-L also exhibited excellent stability when incubated with 10% FBS at 4 C for 24 h, with no significant difference in size or zeta potential.

In vitro drug release
As shown in Figure S4a, it can be seen that the free PUE solution presented a rapid release with more than 90% of the drug released from the dialysis bag within the first 24 h.while the release of PUE-loaded liposomes (PUE@L, PUE@T-L, PUE@M-L and PUE@T/M-L) exhibited a burst release in the early stage, followed by a slow and sustained release of PUE throughout almost 72 h.These sustained behaviors mechanistically may be attributed to the slow degradation of polymeric materials and the release of PUE from the lipid layer by diffusion (Wang et al. 2021).PUE@T-L, PUE@M-L, and PUE@T/M-L showed similar sustained behaviors, while PUE@L without modification exhibited a relatively faster release profile.Similar reports in some literature revealed that surface modification of nano-carriers could delay the drug release and bring the systems more sustained release patterns (Zhang et al. 2018, Shao et al. 2019).The longer release time of PUE@T/M-L could protect the drugs for a relatively long time from being degraded in the circulation system.PUE@T/ M-L had sustained release properties in vitro, and it was presumed that effective drug concentrations and long-lasting therapeutic effects may be maintained at lesion sites in vivo.(Li et al. 2019a).

Hemolysis assay
Hemocompatibility is a key point for in vivo injectable therapeutics of liposomes.As shown in Figure S4b, the PUE exhibited slight concentration-dependent hemolysis, while the hemolysis percentage of PUE-loaded liposomes was lower than that of the same concentration of PUE.Although the positively charged TPP was reported to have a risk of hemolysis (Kuznetsova et al. 2019), the functionally modified liposomes such as PUE@T/M-L had superiority in hemocompatibility and could be regarded as suitable for intravenous administration.It was speculated that the effect of steric shielding or coating of liposomes may significantly reduce the association of erythrocytes with TPP.In addition, the strongly negatively charged liposomes may neutralize the positively charged TPP, resulting in PUE@T/M-L remaining negatively charged.In a physiological environment, both the RBCs and plasma protein were negatively charged, which may reduce the non-specific interaction with the negatively charged PUE@T/M-L.

Cellular uptake study
To trace the cellular uptake of the liposomes, a hypoxia/reoxygenation (H/R) model of H9c2 Cells was established.The cellular uptake study of C6@T/M-L was evaluated by fluorescence microscopy and flow cytometry.After 1 h of incubation, the fluorescence intensity of C6 in both C6@T-L and C6@T/M-L treated cells was significantly brighter compared to C6@L and clustered around the nucleus.As the incubation time was prolonged from 1 h to 4 h, the cellular uptake of C6@T-L or C6@T/M-L by H/R injured H9c2 cells was significantly increased as compared with C6@L (Figure 2(a)).It revealed that lipophilic TPP cation on the surface of C6@T-L or C6@T/M-L could effectively increase the cellular uptake of liposomes in H/R injured H9c2 cells.
The cellular uptake of liposomes was also quantitatively analyzed by flow cytometry.With the increase in incubation time, the uptake of green fluorescent nanoparticles in H/R injured H9c2 cells displayed a time-dependent manner.The fluorescence intensity of C6@T/M-L in 4 h increased by 2.2fold as compared with that incubation for 1 h.After 4 h incubation, the greatest fluorescence intensity was also found in the H9c2 cells exposed to C6@T/M-L, which was almost 2.7fold and 1.6-fold higher than cells exposed to free C6 and C6@L, respectively (Figure 2(b)).The results further indicated that the TPP cation on the surface of C6@T/M-L could facilitate their interaction with H9c2 cells, leading to preferential accumulation of C6@T/M-L at the H/R injured H9c2 cells.This might result from the fact that cationic TPP could be convenient for the liposomes to interact with the negatively Figure 2. Cellular Uptake study.(a) Fluorescent images for qualitative cellular uptake of free C6, C6@L, C6@M-L, C6@T-L and C6@T/M-L incubation in H9c2 cells at 1 and 4 h.(b) H9c2 cells were incubated with free C6, C6@L, C6@M-L, C6@T-L and C6@T/M-L and the intracellular fluorescence intensity was measured by flow cytometry.(c) Relative uptake efficiency of C6@T/M-L on H/R injured H9c2 cells in the presence of various endocytosis inhibitors ( ÃÃ p < 0.01, Ã p < 0.05).Scale bar:100 lm.
charged cell membrane, leading to absorptive endocytosis (Biswas et al. 2012).Figure 2(c) shows that sucrose and colchicine significantly inhibited the cellular uptake of C6@T/M-L in H/R injured H9c2 cells.It implied that C6@T/M-L was internalized by H/R injured H9c2 cells mainly through clathrin-mediated endocytosis and micropinocytosis pathways (Zhang et al. 2019).

Intracellular distribution and co-location analysis
We next examined the intracellular localization of C6@T/M-L in H9c2 cells by fluorescence microscopy.The lysosomes of H9c2 cells were stained with Lysotracker red dye.As shown in Figure 3, a multitude of free C6 and C6@L were observed to be trapped in lysosomes, with the Pearson's correlation coefficients (R) value of 0.81 ± 0.01 and 0.74 ± 0.05 for C6 and C6@L (Figure S5a), respectively.However, there was little overlap of the fluorescence between green (C6@T-L and C6@T/M-L) and red (Lysotracker) signals after 4 h incubation.A significantly reduced R value of lysosomes with TPP modified liposomes was observed for C6@T-L (0.51 ± 0.10) and C6@T/M (0.48 ± 0.01).The results suggested that C6 @T-L or C6@T/M can partially escape from the lysosomes (Figure S5b).It was speculated that the 'proton sponges' effect of cationic TPP may lead to swelling and rupture of the lysosome (Biswas et al. 2012).Thus, degradation of the TPPmodified liposomes (C6@T-L and C6@T/M) in lysosomes could be reduced as much as possible for efficient mitochondria-targeting delivery.
We used Mitotracker red dye to label mitochondria and observed the overlap between green fluorescent liposomes and red mitochondria.As shown in Figure 4.The overlapping regions appeared in liposomes with TPP function (C6@T/M-L and C6@T-L), while there are fewer overlapping regions for free C6, thus indicating that free C6 rarely enters mitochondria.In particular, after 4 h, the R values of mitochondria with TPP-functional liposomes (C6@T/M-L and C6@T-L) were 0.77 ± 0.02 and 0.75 ± 0.01, (Figure S5C) respectively, which were significantly higher than those of C6@L (0.62 ± 0.08) and free C6 (0.60 ± 0.05) (Figure S5d).These results showed that the TPP functionalized liposomes had excellent mitochondrial targeting ability mainly depends on the electrostatic absorption (Lu et al. 2022), resulting in a specific accumulation of liposomes at the mitochondrial region (Biswas et al. 2012).The reason may be that TPP had a highly lipophilic ligand with three phenyl groups and a positive charge on phosphorus, which enhanced its cell association and mitochondrial targeting ability (Li et al. 2019a).

Effects of PUE@T/M-L on the mitochondrial function of H/R injured H9c2 cells
Mitochondrial dysfunction plays a crucial role in H/R injured H9c2 cells, especially apoptosis (Zhou et al. 2021).The PUE@T/M-L in the present study were used to increase the mitochondrial targeting of PUE and improve the protection of H9c2 cardiomyocytes against H/R-induced apoptosis, which was detected by Annexin V-FITC/PI double staining in H9c2 cells.The results demonstrated that the apoptosis rate was significantly higher in the H/R group compared with the control group, while the administration groups showed a decreasing trend in the apoptosis rate of H9c2 cells.Among them, the lowest percentage of apoptosis was observed after The rapid increase of ROS levels in mitochondria is the main cause of mitochondrial dysfunction (Zhang et al.  2021a).Excessive ROS generation is also associated with cardiomyocyte apoptosis (Liu et al. 2021).Thus, ROS clearance can benefit the viability to rescue the H/R injured H9c2 cells.The effect of PUE@T/M-L on mitochondrial ROS was shown in Figure 5(c), the fluorescence intensity was significantly increased in H9c2 cells with H/R injury compared to the control group, suggesting that H/R stimulation significantly increased ROS production.After treatment with PUE and PUE-loaded liposomes, the fluorescence intensity of H/R injured H9c2 cells showed different degrees of decrease.Compared with PUE@L, both PUE@T/M-L and PUE@T-L showed stronger inhibition of mitochondrial ROS production.The effect of PUE@T/M-L on mPTP opening in H/R injured H9c2 cells was assessed by using mitochondrial Calcein-AM labeling as previously described (Zhang et al. 2019).As shown in Figure 5(d), It was interesting to note that both PUE@T-L and PUE@T/M-L exerted stronger inhibition of mPTP opening in H/R injured H9c2 cells as compared with PUE@L.Moreover, we further investigated the effects of PUE@T/M-L on the expression of apoptosis-related proteins.As shown in Figure 5(e), Hypoxic reoxygenation (H/R) significantly increased Bax expression and decreased Bcl-2 expression in H9c2 cells, and this result was significantly reversed by the administration of the drugs.Compared with the PUE@L, both PUE@T/M-L and PUE@T-L exhibited a stronger effect on the expression of apoptosis-related proteins.Taken together, these results reconfirmed the advantages of TPP, which could effectively improve the mitochondrial targeting of PUE, with a large amount of PUE accumulating in the mitochondria, inhibiting the opening of mPTP, decreasing the production of ROS, and regulating the expression of apoptosis-related proteins, thus enhancing the drug protective effect on H/R-induced apoptosis.

Colocalization of MMP-TP modified liposomes with MMP-2
During the MI/RI, damaged cardiomyocytes release inflammatory cytokines and chemokines, leading to the recruitment of macrophages (Nguyen et al. 2015).Macrophage infiltration is a hallmark pathological change observed in the early stage of MI/RI.Macrophages can secrete matrix metalloproteinases in large quantities in response to inflammatory stimuli, thus causing a significant increase in matrix metalloproteinase expression around ischemic cardiomyocytes (Spinale 2007, Halade et al. 2013).According to the methods reported in the literature, we used LPS stimulation of macrophages to simulate the inflammatory induction process during MI/RI (Shen and He 2021).An inflammatory response was induced in RAW 264.7 cells by LPS treatment, leading to overexpressed MMPs.Immunofluorescence analysis showed green fluorescently labeled MMP-TP modified liposomes (C6@M-L and C6@T/M-L) could bind well to red fluorescently labeled MMP-2 (Figure 6(a)).The co-localization effect of C6@M-L and C6@T/M-L was significantly stronger than the other groups, indicating that MMP-TP modified liposomes were able to bind to MMP-2 in vitro (Figure 6(b)).The result was exactly in agreement with the previous literature (Nguyen et al. 2015).
Binding of MMP-TP modified liposomes to LPS-Activated RAW 264.7 cells Macrophages are the main source of MMPs after myocardial infarction and a portion of the secreted MMPs can be bound to the cell membrane.This results in elevated expression of MMPs on macrophage membranes (Nguyen et al. 2015).As depicted in Figure S6, fluorescence microscopy observed less fusion of both free C6 and C6@L with Dil-labeled cell membrane within 4 h.By contrast, In the C6@M-L and C6@T/ M-L treated groups, the fusion of green and red fluorescence on the cell surface appeared yellow within 2 h, and the yellow color became apparent with incubation time (until 4 h), demonstrating the targeting ability of MMP-TP modified liposomes to LPS-activated RAW 264.7 cells.Nguyen et al have also reported high binding of MMP-TP to other MMPexpressing cell lines, such as U937 cells (Nguyen et al. 2015).As the liposomes fused with the macrophage membrane, some of the liposomes were taken up by the cells and accumulated around the blue cell nucleus.As shown in Figure S7, strong green fluorescence was observed in RAW 264.7 cells incubated with C6@M-L and C6@T/M-L, indicating that MMP-TP modified liposomes were significantly taken up by macrophages.In contrast, almost no fluorescence was observed in RAW 264.7 cells treated with free C6.Of note, C6@T-L exhibited a moderately higher accumulation around the nuclei than C6@L, suggesting that TPP cation promote the cellular uptake of liposomes.The strongest fluorescence intensity of C6@T/M-L among all treatment groups was likely attributed to the dual effect of pro-cellular uptake of TPP cation and MMPs targeting of MMP-TP.The quantitative cellular uptake results determined by flow cytometry were also essentially similar to the above qualitative fluorescence assay.The quantitative assay showed that the fluorescence intensity of C6@T/M-L in RAW 264.7 cells was 6.6 and 3.5 times higher than that of free C6 and C6@L, respectively, and also significantly stronger than that of C6@M-L and C6@T-L (Figure 7).These data indicated that C6@T/M-L had a significantly stronger binding capacity to LPS-activated RAW 264.7 cells than other groups, which may be explained by the combined effect of specific binding of MMP-TP and pro-cellular uptake of TPP cation.

Attenuation effects of MMP-TP modified liposomes on inflammation in vitro
Macrophage infiltration plays an important role in inducing inflammation during MI/RI.Macrophages release large amounts of inflammatory cytokines that exacerbate the oxidative stress response in ischemic cardiomyocytes, leading to massive apoptosis of cardiomyocytes (Zhang et al. 2021b).Meanwhile, excessive ROS can also cause cardiac inflammation, which in turn aggravates myocardial damage (Zhao et al. 2022).Therefore, alleviating the production of ROS and inflammatory cytokines produced by macrophages is considered an effective way to alleviate MI/RI.Compared with other groups, PUE@T/M-L significantly reduced TNF-a in LPS-stimulated RAW 264.7 cells (Figure S8a), while decreasing ROS levels (Figure S8b), curbing the vicious cycle of ROS and inflammation interacting with each other.This is attributed to the dual effect of specific binding of MMP-TP and pro-cellular uptake of TPP cation, which effectively deliver large amounts of drugs into cells for better anti-inflammatory effects.

Ex vivo biodistribution determined by Fluorescence imaging
As clearly shown in Figure S9, the intensity of cardiac fluorescence in the DiR@T/M-L group was significantly stronger than that in the other groups, especially at 24 h of administration, the cardiac fluorescence in the DiR@T/M-L group was still maintained, while the cardiac fluorescence in the other two groups showed almost disappearance.These findings suggested that T/M-L had better targeting and retention effects in the ischemic myocardium.This may be due to the synergistic effect of the ischemic myocardial targeting of MMP-TP and the pro-cellular uptake of TPP cation.On the one hand, MMP-TP could guide the DiR@T/M-L to accumulate at the ischemic myocardium by binding to MMPs secreted by macrophages.On the other hand, TPP cation can also accumulate several hundred-fold within mitochondria due to the highly negative mitochondrial membrane (Ye et al. 2022).It has been reported in the literature that 18 F-FPTP, as mitochondrial voltage sensors in PET myocardial imaging, demonstrated higher accumulation and retention in cardiac cells owing to the higher mitochondrial membrane potential and a huge bulk of mitochondria in cardiomyocytes, which could enhance their electrostatic attraction to the TPP moiety of 18 F-FPTP (Kim and Min 2015).

Histological assessment of liver samples
As observed from HE staining imaging, the liver tissue of PUE@T/M-L was intact with clear liver lobules and regular arrangement of hepatocytes, and no inflammatory cell infiltration or hepatocyte swelling was observed (Figure S10).This indicated that PUE@T/M-L was almost non-toxic to the liver.

Conclusion
In the present study, we successfully developed PUE@T/M-L co-modified with MMP-TP and TPP, which exhibited good lysosomal escape ability and mitochondrial targeting, largely enhancing the drug's effect in reducing H/R-induced H9c2 cell apoptosis.In addition, PUE@T/M-L facilitated drug entry into LPS-stimulated RAW 264.7 cells, inhibited the release of inflammatory factors, and contributed to the improvement of the inflammatory microenvironment at ischemic myocardial sites in the future.
Scheme 1. Schematic illustrations for PUE@T/M-L stepwise targeting process to the mitochondria in ischemic cardiomyocytes.

Table 1 .
The lipid composition, size and zeta potential of liposomes.