An injectable micelle-hydrogel hybrid for localized and prolonged drug delivery in the management of renal fibrosis

Abstract

Localized delivery, comparing to systemic drug administration, offers a unique alternative to enhance efficacy, lower dosage, and minimize systemic tissue toxicity by releasing therapeutics locally and specifically to the site of interests. Herein, a localized drug delivery platform (“plum‒pudding” structure) with controlled release and long-acting features is developed through an injectable hydrogel (“pudding”) crosslinked via self-assembled triblock polymeric micelles (“plum”) to help reduce renal interstitial fibrosis. This strategy achieves controlled and prolonged release of model therapeutics in the kidney for up to three weeks in mice. Following a single injection, local treatments containing either anti-inflammatory small molecule celastrol or anti-TGFβ antibody effectively minimize inflammation while alleviating fibrosis via inhibiting NF-κB signaling pathway or neutralizing TGF-β1 locally. Importantly, the micelle-hydrogel hybrid based localized therapy shows enhanced efficacy without local or systemic toxicity, which may represent a clinically relevant delivery platform in the management of renal interstitial fibrosis.

Graphical abstract

Schematic representation of the “plum‒pudding” system consisting of hyaluronic acid (HA)-derived hydrogel crosslinked via self-assembled triblock micelles, which can serve as a reservoir for hydrophobic solutes to achieve controlled and prolonged release.

1. Introduction

Systemic drug administration often faces premature release and nonspecific distributions resulting in lower efficacy and off-target toxicity. Localized delivery offers an alternative to enhance efficacy, lower dosage, and minimize systemic tissue toxicity by locally releasing therapeutics specifically to the site of interests. Langer and colleagues1, 2, 3 pioneered the technology of polymer-based drug delivery depots to directly deliver therapeutics to the disease site in a controlled manner. However, some require pre-formulation under harsh chemical conditions and complex surgical interventions to place the implants^4,5. As an alternative, injectable hydrogels are gaining attentions over a broad range of applications such as cellular scaffolds, controlled biomolecule delivery, soft tissue replacements, and wound dressings^6,7. Specifically, hydrogel precursors and therapeutics are injected in the solution form, which then undergo in situ gelation via either physical or chemical crosslinking. However, due to their loose structure and hydrophilic nature, localized hydrogel systems disfavor hydrophobic drugs and usually incorporate small molecules via covalent conjugation which relies on heavy chemistry resulting in undesired toxicity^8,9.

Inspired by advances in biomedical engineering and nanotechnology, we proposed an injectable micelle-hydrogel hybrid for noncovalently encapsulating hydrophobic drugs to achieve localized controlled drug release. In this system, terminal functionalized triblock polymeric micelles not only carry on hydrophobic small molecules, but function as hydrogel crosslinkers, displaying a novel “plum‒pudding” structure (Scheme 1). Such a material system is characterized by randomly dispersed micellar nanoparticles (“plums”, content: 50%) in a conventional hydrogel network (“pudding”). With this type of design, poorly water-soluble molecules can be easily encapsulated in the hydrophobic micellar core via hydrophobic interaction. Together, the micelle-crosslinked hydrogel hybrid can serve as a reservoir for hydrophobic solutes to achieve controlled and prolonged release. Additionally, the hydrogel network maintains the nature of rationally modulated mesh sizes which provides proper control over the mass transport of biomolecules and the potential of cell therapy.

Scheme 1.

Schematic representation of the “plum‒pudding” system consisting of hyaluronic acid (HA) derived hydrogel crosslinked via self-assembled triblock micelles.

In recent years, anti-fibrosis drug research mainly focuses on the development of small molecules and antibodies macromolecules. At present, more than 30 small molecule drugs targeting different molecular targets are in the pre-clinical and clinical research stage, involving molecular targets including kinases, growth factors, chemokines, lipid regulation, proteases, ion channels, nuclear factor receptors, etc.^10,11. For highly specific kinase inhibitors, treatment concentrations need to be carefully selected to avoid or reduce adverse effects due to "off-target" distribution. There are also problems with inhibitors of single growth factors. For example, the tyrosine kinase inhibitor imatinib works by inhibiting the platelet growth factor (PDGF) pathway, but its phase II clinical trials for idiopathic pulmonary fibrosis (IPF) and systemic sclerosis (SSc) have failed¹². Currently, the first-line treatment for renal interstitial fibrosis (RIF) mainly focuses on the renin-angiotensin-aldosterone system, including angiotensin-converting enzyme inhibitors, angiotensin II receptor 1 antagonists, and direct renin blockers, and it is limited to reducing symptoms and delaying the progression of RIF13, 14, 15, 16.

In the present study, we applied the unique hydrogel‒micelle hybrid system to locally deliver therapeutics for the management of RIF in mice. Celastrol (CLT), a natural compound extracted from Tripterygium wilfordii, was selected as a model therapeutic for the treatment of RIF due to its excellent anti-inflammatory and antiproliferation activities17, 18, 19. Our laboratory previously reported an albumin nanoparticle with a well-defined size achieving site-specific delivery of CLT to mesangial cells for the treatment of mesangioproliferative glomerulonephritis²⁰. Without a proper carrier system, CLT following intravenous administration often resulted in dose-related systemic toxicities possibly via interaction with multiple cellular targets^21,22. In this work, we developed a hyaluronic acid-derived hydrogel hybrid crosslinked by terminally functionalized Pluronic micelles to achieve localized and prolonged release of CLT in the kidney. In addition, due to the critical role of TGF-β during fibrosis²³, anti-TGFβ antibody was selected as a model biomolecule to further demonstrate the sustained therapeutic effect of the localized delivery strategy against RIF using a unilateral ureteral obstruction (UUO) mice model.

2. Materials and methods

2.1. Materials

Pluronic® F127 (molecular weight 12,600 Da) was purchased from BASF (Tarrytown, USA). Sodium hyaluronate (HA, molecular weight 77 kDa) was provided by Bloomage Biotechnology Corporation Limited (Ji′nan, China). N-Hydroxysuccinimide (NHS), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDCI), and cysteamine hydrochloride (CHS·HCl) were obtained from Sigma‒Aldrich (USA). Methacrylic anhydride (MA) was purchased from Best-Reagent (Chengdu, China). Bovine serum albumin (BSA) was provided by Biosharp (Hefei, China). Cyanine 5.5-N-hydroxysuccinimide (Cy5.5-NHS) was purchased from Hypercyte (Beijing, China). Celastrol (CLT) was obtained from Must Biotechnology (Chengdu, China), and murine monoclonal anti-TGFβ1 antibody was provided by R&D (Minneapolis, MN, USA).

2.2. Synthesis of hydrogel precursors

Bis-F127-methacrylate (bis-F127-MA) was synthesized according to Supporting Information Scheme S1. Briefly, F127 (5 g, 0.4 mmol) was dissolved in 100 mL of anhydrous dichloromethane in a round bottom dry flask followed by dropwise addition of 2 mL of triethylamine and methacrylic anhydride (4 mL, 0.026 mol) under ice bath. After stirring for 24 h at room temperature, the reaction mixture was concentrated to 20 mL via rotary evaporation, precipitated in cold diethyl ether, and then dried in vacuum. The obtained bis-F127-MA (yield: 80.2%) was validated by ¹H NMR (400 MHz, CDCl₃, δ): 6.13 (s, 2H), 5.58 (s, 2H), 3.57 (bs, PEG), 1.13 (bs, 6H) (Supporting Information Fig. S1).

Thiolated hyaluronic acid (HASH) was obtained following a pre-established method (Supporting Information Scheme S2)²⁴. The obtained product (yield: 89.9%) was validated by ¹H NMR (400 MHz, D₂O, δ): 4.38 (bs, 2H), 3.0‒3.8 (bs, 5H), 2.54 (m, 2H), 2.30 (m, 2H), 1.87 (bs, 3H) (Supporting Information Fig. S2).

2.3. Gelation and characterizations of physicochemical properties

2.3.1. Preparation of hydrogel

F127 micelle crosslinked hydrogel (F127-HA) was synthesized via Michael addition between thiol residuals on the HA backbone and the terminal methacrylate on bis-F127-MA. Briefly, bis-F127-MA micelles were produced by a thin film-dispersion method. First, 100 mg of bis-F127-MA was dissolved in 10 mL of acetonitrile in a 50 mL round bottom flask, evaporated via rota-vap at 45 °C to form a thin film, then hydrated in 2 mL of HEPES to afford a micelle dispersion (pH 8.0), which was then mixed with HASH in ddH₂O of various concentrations, e.g., 1%, 5%, and 10% (w/v). Pregel solutions were quickly mixed at equal volume and allowed to gel at 37 °C, and gelation was assessed by the standard vial tilting method²⁵.

To load therapeutic components, 8 mg of CLT and 100 mg of bis-F127-MA were dissolved in the organic solvent such as acetonitrile prior to thin film formation, which was then subjected to hydration in 2 mL of HEPES to afford CLT loaded 5% (w/v) bis-F127-MA micelles, while anti-TGFβ1 antibody was dissolved in 5% (w/v) HASH solutions. The final concentrations of CLT and anti-TGFβ1 antibody were 4 and 6.7 mg/mL, respectively. After two parts were fully mixed in equal volume, the gel precursor solution was then ready for injection in mice.

2.3.2. Characterization of rheological properties

The gelation kinetics was monitored by rheometric study. The following rheological measurements were carried out at 37 °C on a Haake Mars Modular Advanced Rheometer (Thermo Fisher Scientific, USA). A time sweep mode was adopted (frequency, 1 Hz; strain, 5%) to determine the sol‒gel transition point, using a parallel plate setting (diameter, 40 mm; gap, 1000 μm). Next, at a fixed strain of 5%, a frequency sweep test was performed from 0.1 to 10 Hz to measure the storage modulus (G′), and loss modulus (G″). The frequency range was ensured to stay in the linear viscoelastic region of each hydrogel sample as determined by amplitude sweeps (data not shown). To determine the injectability of such hydrogel hybrids, the viscosity of hydrogel precursor solutions was determined using a standard viscometer (Brookfield, HBDV-II+PRO, USA) as previously described²⁶.

2.3.3. Characterization of hydrogel microstructure

The hydrogel samples were frozen in liquid nitrogen and then lyophilized. The cross-sections of lyophilized hydrogels were sputter-coated with gold, and subjected to observation under scanning electron microscopy (SEM, JSM-7500F, JEOL, Japan) as previously described²⁷.

2.3.4. In vitro swelling

To perform swelling studies, hydrogel discs were fabricated, freeze-dried and weighed (W_d). Next, 3 mL of 1× PBS (10 mmol/L, pH 7.4) was added to the lyophilized hydrogel to allow swelling at 37 °C. At each given time point, hydrogel disks were removed, the surface water was carefully blotted with filter paper, and the wet weight (W_t) of each hydrogel disk was recorded. The real-time swelling ratio (RSR) has been calculated according to the following Eq. (1):

$$\text{RSR}(\text{\%})=\frac{W_t-W_{\text{d}}}{W_{\text{d}}}\times 100$$ (1)

2.3.5. In vitro degradation

To explore hydrogel degradation profiles in vitro, fully cured F127-HA hydrogels were lyophilized, and the lyophilized samples were placed in 10 mL of 1× PBS (10 mmol/L, pH 7.4) to allow swelling and degradation at 37 °C. The weight of lyophilized hydrogel was recorded as W₀, which was allowed to swell in the buffer solution over time. At each given time point, the hydrogel sample was collected, freeze-dried and the gel weight at time t was recorded as W_t. The degradation profile of F127‒HA hydrogel hybrid was presented as the percentage of weight loss (W_loss,%) calculated by the following Eq. (2):

$$W_{\text{loss}}(\text{\%})=\frac{W_t-W_0}{W_0}\times 100$$ (2)

2.4. In vitro release

As mentioned in Section 2.3, 1 mg of DiD and 100 mg of bis-F127-MA were dissolved in acetonitrile (8 mL), evaporated using a rota-vap, and rehydrated in 2 mL of HEPES to afford DiD loaded bis-F127-MA micelle solution (5%, w/v). An equal volume of 5% DiD loaded micelle solution was mixed with 5% HASH solution to afford DiD loaded F127-HA hydrogel. To load Cy5.5-BSA, 1 mg of Cy5.5-BSA was dissolved in 5% (w/v) HASH solution and then mixed with bis-F127-MA micelle to obtain Cy5.5-BSA loaded hydrogel.

The release study was then performed to determine the cumulative release percentages of DiD and Cy5.5-BSA over time. Briefly, 500 μL of the hydrogel precursor solution was placed into the mold and allowed to gel. The fully cured hydrogel samples were then removed and placed in vials containing 2 mL of 1× PBS (10 mmol/L, pH 7.4) at 37 °C under shaking (100 rpm). At predetermined time points, release media were collected and replaced with fresh media. The concentrations of DiD and Cy5.5-BSA in the release samples were then analyzed by fluorescence spectrometry using a Varioskan flash multimode microplate reader (Thermo Fisher Scientific, USA). The cumulative release percentage was then calculated, and the release profiles of DiD and Cy5.5-BSA were analyzed using mathematical models as previously described²⁸.

2.5. In vivo release of DiD and Cy5.5-BSA

Wild-type male BALB/c mice (6–8 weeks old, 18–22 g) were provided by the Laboratory Animal Center of Sichuan University (Chengdu, China). All animal protocols and studies were approved by the Ethics Committee of Sichuan University and performed per university guidelines. To conduct local injection, 15 μL of hydrogel precursor solution was injected into the renal capsule of the left kidney of each mouse. Following administration, mice were imaged using IVIS Spectrum Imaging System (PerkinElmer, USA) on Days 1, 3, 7 and then weekly until Day 21. Fluorescence intensity (Cy5.5-BSA: λ_ex/λ_em = 620/710 nm; DiD: λ_ex/λ_em = 644/665 nm) was semi-quantitatively analyzed and normalized to the peak intensity of each mouse as previously reported²⁹.

2.6. Therapeutic efficacy in unilateral ureteral obstruction (UUO) mice

Wild-type male BALB/c mice were used to establish mouse UUO model by ligation of the left ureter as previously reported³⁰. Immediately after ligation, mice in six groups were treated accordingly: i) sham operation group (sham); ii) UUO group with no treatment (UUO); iii) UUO group followed by intravenous injection of CLT solution (UUO + CLTs, 3 mg/kg × 4 times); iv) UUO followed by a single injection of F127-HA hydrogel into the renal capsule (15 μL blank hydrogel); v) UUO followed by a single injection of F127-HA hydrogel containing CLT (3 mg/kg CLT, 15 μL); vi) UUO followed by a single injection of F127-HA hydrogel containing anti-TGF-β1 antibody (100 μg). To evaluate the therapeutic efficacy, healthy mice without any operation or treatment served as negative control. Above hydrogel precursor solutions containing different treatments were locally injected at the renal capsule of the affected kidney immediately after the initial UUO procedure^31,32. Mice were sacrificed on Days 7, 14, and 21 following the initial UUO operation. All animals were handled in compliance with internal guidelines, and above animal protocols and operations were approved by the Ethics Committee of Sichuan University, China.

2.7. Immunofluorescence staining of macrophages

Paraffin-embedded kidney sections were dewaxed with a gradient ethanol solution, then antigen-repaired with EDTA antigen repair solution, and blocked with BSA for 30 min. Next, the primary antibodies were incubated in a wet box at 4 °C overnight followed by secondary antibody incubation at room temperature for about 50 min in the dark. DAPI was used to counterstain nuclei for 10 min. The following primary antibodies were used: rat polyclonal F4/80 antibody (1:100 dilution, Abcam, USA), rabbit polyclonal CD206 antibody (1:100 dilution, Abcam, USA) and rabbit polyclonal iNOS antibody (1:100 dilution, Abcam, USA) to label anti-inflammatory (M2) and pro-inflammatory (M1) macrophages, respectively. The secondary antibodies included: Cy3 conjugated goat anti-rabbit IgG (catalog number GB21303, 1:300 dilution, Servicebio, China), and Alexa Fluor® 488-conjugated AffiniPure goat anti-rabbit IgG (catalog number GB25303, 1:400 dilution, Servicebio, China). Fluorescence images were taken on a confocal microscope (LSM 800, Carl Zeiss, Jena, Germany) obtaining 18 images per section. As previously mentioned³³, the average fluorescence intensity was quantitated by ImageJ (NIH, Bethesda, USA). The mean gray value (MOD) of positive staining in the visual field of each slide was calculated respectively.

2.8. Immunostaining and histological analysis

To evaluate the in vivo efficacy and systemic toxicity, standard hematoxylin and eosin (H&E) staining was conducted on kidney and vital organ sections. To evaluate cell apoptosis, in situ cell death detection kit-POD (Roche, Switzerland) was used per manufacturer's instruction, and the cell nuclei were counterstained by hematoxylin. Next, fibrosis was evaluated following Masson's trichrome staining (Servicebio, China). Apoptotic cells were analyzed at 40× magnification, and the fibrotic area was semi-quantitatively analyzed for trichrome staining using Image-Pro Plus software (Media Cybernetics, USA).

To further characterize fibrosis and inflammatory cytokine levels of the affected kidney, immunostaining was performed on kidney sections using primary antibodies for α-smooth muscle actin (α-SMA, 1:100 dilution, Abcam, USA), cluster of differentiation 68 (CD68, 1:80 dilution, Abcam, USA), Collagen I (1:100 dilution, Abcam, USA), tumor necrosis factor α (TNF-α, 1:120 dilution, Abcam, USA), interleukin 6 (IL-6, 1:100 dilution, Abcam, USA), and interleukin 1β (IL-1β, 1:80 dilution, Affinity, USA) with biotinylated goat anti-rabbit IgG (Zhongshan Jinqiao Biotechnology, Beijing, China) as the secondary antibody. Representative images were obtained using an inverted microscope (Leica, Germany). The levels of collagen deposition and inflammatory cytokines such as TNF-α, IL-6, and IL-1β were semi-quantitatively analyzed at 40× magnification using Image-Pro Plus software (Media Cybernetics, USA).

2.9. Statistics

Data were presented as mean ± standard deviation (SD) of minimum three replicates unless otherwise indicated. Comparisons among multiple groups were conducted by one-way ANOVA followed by a Tukey post hoc analysis using GraphPad Prism 5.0 (La Jolla, CA, USA). P-values of less than 0.05 were considered statistically different.

3. Results and discussions

3.1. Synthesis of hydrogel precursors

Triblock Pluronics® (PEO-PPO-PEO) are commercially available amphiphiles, which undergo self-assembly into nanoscale micelles thus being a good pharmaceutical excipient for solublizing hydrophobic drugs. Among commercially available Pluronics, F127 (Mw 12,600 Da) was selected due to its suitable hydrophilic-lipophilic balance value, which has previously been proven to assemble into micelles with an average size of around 30 nm³⁴. Thiolated HA (HASH) and bismethacrylate Pluronic F127 (bis-F127-MA) were synthesized using established methods (Schemes S1 and S2)35, 36, 37. The chemical structures of HASH and Bis-F127-MA were validated by magnetic resonance spectroscopy (Figs. S1 and S2). In the following study, we would explore a simple approach utilizing a triblock Pluronic-based micelle with HA hydrogel for the controlled and prolonged delivery of both small molecule and macromolecule drugs. Using shell functionalized micellar particles as the hydrogel crosslinkers, efficient in situ gelation are expected to be achieved under physiological conditions which then act as a drug-containing depot to provide long-term release in vivo.

3.2. Gelation study and characterizations

To prepare Pluronic micelles, a thin film evaporation and dispersion method was adopted, and bis-F127-MA was proven to self-assemble into near spherical particles in the aqueous solution with an average size of about 35.6 nm (PDI = 0.168 ± 0.02), and a zeta potential of −0.13 ± 0.011 mV (Fig. 1A and B), which are consistent with the size and zeta potential of Pluronic F127 micelles without any chemical modification²⁷. The encapsulation efficiency for various hydrophobic small molecules in bis-F127-MA micelles remained above 90% with loading capacities of around 2% (Supporting Information Table S1) indicating the suitability and feasibility of bis-F127-MA micelles for drug encapsulation and solubilization.

Figure 1.

Physicochemical characterizations of bis-F127-MA micelles crosslinked HA hydrogels. Size distribution profile of bis-F127-MA micelles determined by dynamic light scattering (A). Representative TEM images of bis-F127-MA micelles (B), and F127-HA hydrogel (C). Representative SEM image of lyophilized F127-HA hydrogel (D). Scale bars = 100 nm. Rheological characterization of F127-HA hydrogels via dynamic time sweep (E) and frequency sweep (F) tests. The viscosities (G), swelling (H) and degradation (I) over time profiles of F127-HA hydrogels. Data represent mean ± SD (n = 3). ∗P < 0.05, ∗∗P < 0.01 and ∗∗∗P < 0.001.

To perform gelation study, bis-F127-MA micelles were synthesized as aforementioned and dispersed in HEPES (0.5 mmol/L, pH 8.0) at polymer concentrations of 1%, 5% and 10% (w/v). Meanwhile, HASH was dissolved in ddH₂O at concentrations of 1%, 5% and 10% (w/v). Bis-F127-MA micelle dispersions and HASH were then mixed and achieved complete gelation within 5 min at 37 °C (Supporting Information Fig. S4). To examine the micelle structure within the fully cured hydrogel network, equal volumes of bis-F127-MA micelles and HASH solution were well mixed and allowed to form a thin hydrogel layer over the transmission electron microscopy (TEM) grids. The TEM image showed that bis-F127-MA micelles uniformly distributed within the hydrogel network and maintained their near spherical morphology similar to bis-F127-MA micelle suspensions (Fig. 1C). After lyophilization, F127-HA hydrogel network displayed the typical porous microstructure under SEM (Fig. 1D), the pore size of which could be modulated by varying the total polymer concentration or crosslinking density.

Next, the gelation kinetics of the F127-HA hydrogels was characterized by strain oscillatory shear experiments. Per the time sweep test, all three formulations achieved rapid sol‒gel transition within 5 min (Fig. 1E). Also, the gelation time gradually decreased with increasing concentrations of precursor solutions (Fig. 1E). Consistent with previous findings, changing the concentration of hydrogel precursors may help modulate the rate of crosslinking and the gelation time^35,38. Moreover, dynamic modulus measurements were employed to gain insights into the viscoelastic properties of fully cured hydrogels. All three formulations showed G′ (storage modulus)>G″ (loss modulus), indicating an elastic nature of fully cured F127-HA hydrogels. G′ values have also been shown to increase with increasing polymer concentrations (Fig. 1F). To determine the injectability properties, the viscosity of hydrogel precursor solutions was determined using a standard viscometer, and all three formulations displayed viscosity values below 50 cP, which indicates the hydrogel precursor solutions can easily go through 22-gauge needles (Fig. 1G).

Per the dynamic equilibrium swelling study, 5% hydrogel underwent rapid swelling in PBS reaching the equilibrium swelling state in about 20 min (Fig. 1H), and its equilibrium swelling rate (ESR) reached 200%, presumably due to the loose and porous structure of hyaluronan-based networks. Moreover, the 5% hydrogels started to lose their bulk integrity over time as observed by mass loss after 20 days (Fig. 1I).

Another unique feature of the F127-HA hydrogel is its capacity to load poorly water-soluble small molecule drugs and to achieve controlled and prolonged release. Interestingly, without encapsulating in the bis-F127-MA micellar core prior to gelation, physically mixed dexamethasone (DEX) in F127-HA hydrogel displayed an opaque feature which appeared to be suspended in the hydrogel matrix (Supporting Information Fig. S3). In contrast, when DEX was encapsulated in bis-F127-MA micelles prior to forming crosslinked hydrogel, the DEX loaded F127-HA hydrogel displayed a transparent feature (Fig. S3), indicating the extraordinary solubilizing capacity of Pluronic F127 micelles. However, the SEM images did not reveal much variations in the microstructure of these two hydrogel formulations both showing a porous structure after freeze-drying (Fig. S3).

3.3. In vitro and in vivo release studies

According to a preliminary 14 d release study, CLT and Cy5.5-BSA from F127-HA hydrogel hybrid showed a rapid initial burst with over 60% cumulative release within the first 24 h followed by gradually increasing cumulative release for up to 14 days (Supporting Information Fig. S4). The rapid initial burst release in vitro is likely due to the rapid swelling effect of the hydrogel hybrid as demonstrated in the previous section; however, when in the in vivo environment, the hydrogel hybrid may not undergo such rapid swelling due to the limited body fluid around, which may result in reduced initial burst.

To gain connection between the in vitro and in vivo release, a near infrared hydrophobic probe DiD and Cy5.5-BSA were selected as two model solutes in the following release studies. DiD was encapsulated in the hydrophobic core of the bis-F127-MA micelles and further afforded the DiD loaded F127-HA hydrogel. Meanwhile, Cy5.5-BSA was synthesized, mixed with HASH solution, and then crosslinked by bis-F127-MA micelles. Both DiD and Cy5.5-BSA achieved continuous release for up to 7 days in vitro (Fig. 2A), while in vivo, both solutes showed continuous release for up to 21 days with a gradual decay of the fluorescence intensity over time (Fig. 2B‒D). Consistently, Cy5.5-BSA showed more rapid and higher cumulative release than DiD from F127-HA hydrogels at all given time points both in vitro and in vivo (Fig. 2A and B), indicating the release of Cy5.5-BSA was mainly driven by diffusion and polymer relaxation. In contrast, the lower cumulative release of DiD at each given time point from in vitro and in vivo studies was likely due to the encapsulation in the micellar core of the bis-F127-MA micelles thus presenting an additional barrier for solute diffusion other than HA scaffolds. Additionally, the solutions of both DiD and Cy5.5-BSA following local injection in the kidney were shown to retain at the injection site for about three days (Supporting Information Fig. S5), which further supports that localized delivery via hydrogel contributes to the prolonged release in vivo. Per mathematical modelling analysis, the Ritger‒Peppas equation showed the best correlation for both DiD and Cy5.5-BSA from in vivo release profiles, which indicates that the releases of both small molecule drugs and biomacromolecules out of the F127-HA hydrogel hybrid were mostly non-Fickian diffusion driven (Supporting Information Table S2). Also, the release was proportional to time and partially dependent on the swelling and relaxation of polymer chains within the hydrogel.

Figure 2.

Release profiles of F127-HA hydrogels. In vitro release profiles of DiD and Cy5.5-BSA from F127-HA hydrogels (A). Semiquantitative in vivo release profiles of DiD and Cy5.5-BSA from F127-HA hydrogels (B). Representative in vivo optical images of DiD (C) and Cy5.5-BSA (D) loaded hydrogel following localized injection in healthy mice. Data represent mean ± SD (n = 3).

3.4. In vivo gelation and biocompatibility

To explore the injectability and gelation in vivo, DiD-loaded hydrogel precursor solutions were injected subcutaneously in mice. Complete gelation was observed at 5 min post injection (Supporting Information Fig. S6), which suggests that the F127-HA hydrogel be suitable for localized injection in the kidney.

Next, to evaluate the injectability and gelation in kidney, 15 μL of hydrogel precursor solution containing either DiD or Cy5.5-BSA was locally injected in the kidney of healthy mouse. After injection, mice were subjected to in vivo imaging at predetermined time points. As shown in the representative images, the injected hydrogel was well retained at the injection site in the affected kidney for up to three weeks (Fig. 2C and D), which will likely result in localized and prolonged drug release.

Per histological analysis, the injection of the F127-HA hydrogel did not invoke any obvious tissue toxicity or inflammatory cell infiltration in the affected kidney for up to 21 days (Supporting Information Fig. S7). In addition, compared with the normal group, the implantation of the hydrogel containing near infrared dyes DiD or Cy5.5-BSA will not have any influence on the renal function indexes (Supporting Information Fig. S8), indicating remarkable biocompatibility of the F127-HA hydrogel hybrid in vivo.

3.5. Mice UUO model and therapeutic efficacy

After confirming the biocompatibility of F127-HA hydrogels, a UUO mice model was established to evaluate the anti-fibrosis effect of the proposed localized therapy^31,32,39,40. CLT and anti-TGFβ1 antibody were selected as model therapeutics to explore the therapeutic efficacy of localized drug delivery in anti-fibrosis study in vivo. CLT was successfully encapsulated within the core of bis-F127-MA micelles to afford CLT loaded F127-HA hydrogel, while anti-TGFβ1 antibody was uniformly dispersed in the HASH solution and further crosslinked by bis-F127-MA micelles to afford anti-TGFβ1 antibody loaded F127-HA hydrogel. Treatment was initiated immediately post UUO surgery, the renal tissues after treatments were collected on Days 7, 14, and 21 for immunohistochemical analysis (Scheme 2).

Scheme 2.

UUO in BALB/c mice and dosing regimen.

Compared with sham control, the UUO group showed tubular atrophy and extensive inflammatory cell infiltration over time (Fig. 3A). Compared with UUO group, all treatment groups demonstrated positive therapeutic efficacy with decreased inflammatory cell infiltration and reduced morphological changes (Fig. 3A). Also, the systemic toxicity of CLT following intravenous injection was greatly relieved by the localized therapy as indicated by reduced cardiotoxicity in the CLT hydrogel group (Fig. S8), which demonstrates the great potential of delivering anti-inflammatory drugs locally instead of systemically.

Figure 3.

Representative images of the affected kidney sections by H&E (A) and Masson's trichrome staining (B). Scale bars = 100 μm.

To evaluate fibrosis, the Masson's trichrome and α-SMA staining was performed and the results further demonstrated an obvious increase in the fibrosis area of the UUO group over time (Fig. 3B, Supporting Information Figs. S10 and S11). At Day 21, the fibrosis area showed progressive and significant increases for the untreated UUO and the group treated with either F127-HA hydrogel alone or CLT solution alone, indicating CLT solution following systemic administration did not show efficacy against renal fibrosis (Fig. 3B and Fig. S10). As a comparison, F127-HA hydrogel containing either CLT or anti-TGFβ1 antibody showed significantly reduced fibrotic areas compared with untreated UUO at day 21 (Fig. 3B and Fig. S11). Consistently, collagen I deposition was greatly downregulated in F127-HA/CLT or F127-HA/anti-TGFβ1 groups on Day 21 (Supporting Information Fig. S12). On Day 21, macrophage infiltration was significantly reduced in the groups treated with F127-HA hydrogel containing either CLT or anti-TGFβ1 antibody as compared to the F127-HA hydrogel only group (Fig. S10 and Supporting Information Fig. S13).

At present, TGF-β1 is recognized as one of the key regulatory factors contributing to the deposition of extracellular matrix (ECM) in the fibrotic region^41,42. Therapeutics commonly used for the treatment of renal interstitial fibrosis include anti-TGFβ1 antibody43, 44, 45, anti-fibrosis cytokine46, 47, 48, and various kinase inhibitors^{13,15,16,29,49}. As expected, anti-TGFβ1 loaded F127-HA hydrogel showed remarkable efficacy against renal fibrosis (Fig. 3).

3.6. Localized treatments differentially modulate macrophage plasticity

Different subtypes of macrophages have been found to play critical roles in many human diseases or experimental disease models such as cancer, inflammation, and fibrosis50, 51, 52, 53, 54, 55. Pro-inflammatory macrophages (M1-like) produce pro-inflammatory cytokines, e.g., IL-1, TNF-α and macrophage migration inhibitory factor, thus causing acute tissue injury, while anti-inflammatory or pro-fibrotic macrophages (M2-like) promote fibrosis via multiple mechanisms⁵⁶.

To elucidate the role of macrophages during the progression of fibrosis in UUO model, the plasticity patterns of renal macrophages in the affected kidney after unilateral ureteral ligation were elucidated via immunofluorescence staining. Interestingly, the percentage of M1-like macrophage labelled by iNOS was found to be dramatically higher than that of anti-inflammatory M2-like macrophages labelled by CD206 within the first five days post ligation, suggesting acute inflammatory responses in UUO mice (Supporting Information Fig. S14). About 7 days post ligation, increases in the percentage of M2-like macrophages were observed in both localized treatment groups, i.e., CLT or anti-TGFβ hydrogel, indicating that the immunomodulatory effect kicked in by accumulating a large number of M2-like macrophages to suppress inflammation (Fig. 4A and C). Specifically, localized anti-TGFβ hydrogel showed a more profound impact on upregulating the level of M2-like macrophages on day 7 post ligation (Fig. 4A and C). Compared with the 7-day results, M2-like macrophages were significantly upregulated on Day 21 in the UUO group (Fig. 4B and D), suggesting a possible correlation between the massive accumulation of M2 macrophages and the excessive deposition of ECM. This observation is consistent with literature findings that M2-like macrophages in UUO mice show pro-fibrotic effect⁵⁶. In comparison, the localized CLT or anti-TGFβ1 treatment groups showed enhanced M2-like macrophage polarization on Day 7 (Fig. 4), whereas after 21 days, localized CLT group showed significantly reduced M2-like macrophage percentage as compared to that of the UUO group (P < 0.05). However, localized anti-TGFβ1 treatment did not result in an obvious reduction in M2-like macrophages versus UUO group (Fig. 4). Taken together, these results showed correlation with Masson's trichrome staining results (Fig. 3B), which suggests renal fibrosis can be alleviated by localized CLT or anti-TGFβ1 antibody treatment, possibly by inhibiting acute inflammation by upregulating M2-like macrophages in the early stage while reversing the excessive deposition of ECM by downregulating the accumulation of M2-like macrophages in the later stage. CLT and anti-TGFβ antibody work as anti-fibrosis therapeutics via different pathways and mechanisms. Hence, localized delivery of CLT or anti-TGFβ1 antibody appeared to differentially modulate macrophage plasticity at the different stages of UUO mice to minimize matrix deposition.

Figure 4.

Anti-inflammatory (M2-like) macrophage profiles on histological injury and development of renal fibrosis in UUO mice. Representative immunofluorescence staining images of CD206/F4/80 on the affected kidney sections at Day 7 (A) and Day 21 (B). Scale bars = 20 μm. Macrophage distribution was assessed semi-quantitatively by fluorescence counting (C) and (D). Each evaluation represents the mean ± SD (n = 6). ∗∗P < 0.01, ∗∗∗P < 0.001 vs. UUO group. ns, not significant.

3.7. Anti-fibrosis effect via neutralizing TGFβ1 or inhibiting NF-κB locally

Active TGFβ1 plays a critical role in activating Smad2/3 and mediating renal fibrosis, and therefore, fibrotic kidney is often associated with an upregulated level of TGFβ1^57,58. However, systemic administration of anti-TGFβ1 antibody may lead to various issues due to the extensive distribution of TGFβ1 in the body, which is commonly involved in various physiological pathways⁵⁹. Thus, neutralizing TGFβ1 locally at the disease site serves as a potential strategy to overcome fibrosis. To further elucidate the anti-fibrosis mechanism, we analyzed the apoptosis rate of the affected kidney sections via TUNEL staining, which showed downregulated apoptosis levels for localized treatment with either CLT or anti-TGFβ1 as compared to UUO group (Supporting Information Fig. S15). However, systemic administration of CLT showed no improvement in the apoptosis rate compared to UUO group (Fig. S14).

When it comes to inflammation and renal fibrosis, kidney injury induces inflammation, while unresolved inflammation may lead to fibrosis as well as end-stage renal disease⁶⁰. The activation of NF-κB and the infiltration of macrophages and lymphocytes have been identified as the major inflammatory components⁶¹. Thus, administration of inhibitors towards NF-κB signaling pathway may help resolve inflammation and possibly minimize fibrosis. According to our previous works and others’, CLT is a well proven inhibitor for NF-κB pathway thus showing outstanding anti-inflammatory activities in a number of diseases including acute pancreatitis⁶², rheumatoid arthritis⁶³, and glomerulonephritis²⁰. As a result, we analyzed the downstream signaling factors of NF-κB signaling pathway including tumor necrosis factor α (TNF-α), interleukin 6 (IL-6), and interleukin 1β (IL-1β) (Supporting Information Fig. S16‒S18). Compared with sham group, untreated UUO showed significantly upregulated levels of TNF-α, IL-6 and IL-1β (P < 0.05, Fig. 5), whereas localized delivery of CLT via F127-HA hydrogel showed greater downregulations of the inflammatory chemokines and cytokines than systemic administration of CLT repeatedly (Fig. 5). Collectively, these results further support that the sustained release of anti-inflammation therapeutics following localized administration may greatly benefit the treatment of RIF.

Figure 5.

Representative images of affected kidneys by immunostaining of TNF-α, IL-6 and IL-1β at 21 days post unilateral ureteral ligation (A). Semi-quantitative analysis of the levels of TNF-α (B), IL-6 (C) and IL-1β (D) following different treatments. Scale bar = 100 μm. Data are mean ± SD (n = 3). ∗P < 0.05 vs. UUO group.

4. Conclusions

In summary, an injectable triblock copolymeric micelle-crosslinked hyaluronan hydrogel hybrid offers a unique “plum‒pudding” structure to deliver hydrophobic small molecules as well as biomacromolecules in a prolonged and controlled manner. Compared to existing hydrogel implants, the F127-HA micelle hydrogel hybrid was proven to solubilize poorly water-soluble drugs while achieving localized controlled drug release in vivo. Treatment with F127-HA hydrogel hybrid containing anti-inflammatory CLT greatly downregulated proinflammatory cytokine levels thus minimizing local inflammatory responses. The localized anti-inflammation therapy represents a viable strategy to manage chronic kidney diseases, which are proven to impact inflammation, apoptosis, and ECM deposition in UUO mice. Furthermore, delivering anti-TGFβ1 antibody locally via F127-HA hydrogel hybrid demonstrated great potential in the management of UUO renal fibrosis. Overall, this versatile injectable F127-HA hydrogel hybrid represents a neat and clinically relevant platform system to achieve localized therapy, which may further be explored to achieve concurrent loading and prolonged release of both small molecules and macromolecular therapeutics for other disease therapies.

Appendix A. Supplementary data

The following is the Supplementary data to this article: