Abstract

   Effective management of postoperative pain and wound healing presents
   significant challenges in clinical settings, driving the need for
   innovative therapeutic approaches. The analgesic and wound healing
   effects of hydrogen (H[2]) have gradually been recognized; however, the
   lack of efficient hydrogen delivery systems remains a major limitation.
   This study introduces a novel transdermal drug delivery system, which
   utilizes sustained‐release photothermal microneedles (MNs) to
   ameliorate incisional pain and accelerate wound healing. Polydopamine
   (PDA)‐modified ZIF‐8@ammonia borane (AB) nanoparticles with
   photothermal conversion properties are designed, along with
   temperature‐responsive QX‐314‐loaded polycaprolactone (PCL)
   microspheres for controlled release, which are delivered in vivo by
   dissolvable MNs. In vitro results showed that PDA@ZIF‐8@AB
   nanoparticles can release H[2] continuously for up to 5 days in an
   acidic microenvironment, while the photothermal properties of PDA
   facilitated controlled release of QX‐314 through 6 cycles of
   near‐infrared (NIR) exposure. In vivo experiments demonstrated that the
   MN system provided sustained analgesia for up to 5 days and promoted
   wound healing in the acidic microenvironment of postoperative
   incisions. Upon NIR exposure, the photothermal conversion of PDA
   activated membrane ion channels and induced thermally triggered
   deformation of PCL@QX‐314 microspheres, allowing for on‐demand release
   of QX‐314 and targeted neuronal uptake, thus offering personalized
   analgesia. In vitro cell experiments and in vivo studies confirmed the
   biocompatibility of the system. This innovative approach not only
   highlights the dual role of H[2] in pain relief and wound healing but
   also provides a new personalized treatment strategy for postoperative
   pain management with promising clinical applications.

   Keywords: hydrogen, microneedle, personalized analgesia, photothermal
   excitation, postoperative pain
     __________________________________________________________________

   This study introduces a novel transdermal delivery system of molecular
   hydrogen for postoperative pain relief and wound healing. This system
   utilizes sustained‐release photothermal microneedles to integrate
   hydrogen‐releasing nanoparticles and temperature‐responsive
   microspheres, achieving acid‐responsive hydrogen release and controlled
   drug release. The innovative approach offers sustained analgesia,
   individualized pain treatment, and promotion of wound healing through
   multi‐faceted therapeutic mechanisms.

   graphic file with name ADVS-12-e03698-g003.jpg

1. Introduction

   Postoperative pain management is a critical issue that needs to be
   addressed in modern anesthesiology and perioperative medicine.^[ [40]^1
   ^] Studies have shown that over 80% of surgical patients experience
   acute pain during the postoperative period, and ≈10% of patients
   develop chronic pain due to inadequate management of acute pain.^[
   [41]^2 ^] Poor pain control not only complicates recovery, potentially
   prolongs hospital stays, but also leads to delayed wound healing,
   resulting in impairments of quality of life.^[ [42]^3 ^] In clinical
   practice, local anesthetics are most commonly used to manage
   postoperative pain, because of their effective analgesic performance
   and minimal side effects.^[ [43]^4 ^] However, the short duration of
   action following a single administration has hindered the application
   of local anesthetics in postoperative pain managements.^[ [44]^5 ^]

   In recent years, researchers have begun to explore effective
   sustained‐release analgesic strategies and innovative drug delivery
   systems. Among these innovations, liposomal formulations^[ [45]^6 ^]
   and hydrogels^[ [46]^7 ^] have been developed to extend the duration of
   pain relief. Our previous study has also investigated a pH‐responsive
   self‐monitoring microneedle (MN), which can provide controlled drug
   release tailored to individual patient needs and effective
   postoperative analgesia for up to 72 h.^[ [47]^8 ^] However, existing
   MN systems still face challenges related to drug loading capacity and
   precision of drug release, which impede the clinical translation.
   Therefore, there is an urgent need to develop new analgesia system to
   postoperative pain managements.

   Recently, hydrogen (H[2]), a colorless, odorless, non‐toxic, renewable
   gas, has garnered attention for its potential medical applications.^[
   [48]^9 ^] For instance, H[2] has demonstrated potential in alleviating
   various types of pain, including postoperative and neuropathic pain,^[
   [49]^10 ^] by targeting oxidative stress and inflammatory pathways,
   offering a promising non‐traditional analgesic option.^[ [50]^11 ^] A
   study by Li et al.^[ [51]^12 ^] confirmed that H[2] could be used as a
   therapeutic agent to alleviate postoperative pain through the
   Trx1/ASK1/MMP9 signaling pathway. Additionally, H[2] has been also
   reported to promote wound healing and tissue regeneration by reducing
   inflammation and oxidative stress.^[ [52]^13 ^] Therefore, we
   hypothesize that H[2] could serve a dual role in postoperative incision
   management, providing both analgesic effects and promoting healing.
   However, delivering H[2] in a controlled and sustained manner still
   presents challenges, particularly in the acidic microenvironment of
   postoperative wounds. Current methods, such as H[2] inhalation, oral
   intake of H[2]‐rich water, or injection of H[2] solutions, often
   require frequent dosing, which can reduce patient compliance and lack
   the precision needed to maintain localized therapeutic concentrations
   at surgical sites over extended periods.^[ [53]^14 ^]

   To tackle these issues, we propose a novel therapeutic strategy: a
   transdermal drug delivery system that utilizes molecular hydrogen
   delivered via a photothermal MN. This approach enables localized H[2]
   delivery to control postoperative incisional pain and tissue healing,
   while also allowing for the on‐demand release of the local analgesic
   (QX‐314) tailored to individual pain thresholds (Figure [54]1 ),
   optimizing postoperative incision management. The system utilizes MNs
   as a drug delivery vehicle, which have emerged as a promising
   transdermal delivery system.^[ [55]^15 ^] There are three steps in the
   development of this H[2]‐based therapeutic system.

Figure 1.

   Figure 1
   [56]Open in a new tab

   Schematic illustration of the H[2]‐releasing and photothermal‐triggered
   MN patch fabrication and its therapeutic mechanism. a) Synthesis and
   preparation process of PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs. ZIF‐8@AB
   was synthesized via a one‐pot reaction, with AB encapsulated for
   acid‐responsive H[2] release, then coated with PDA and incorporated
   into the MN patch. PCL@QX‐314 MSs were prepared by emulsion and
   integrated into the MN system. b) Mechanism of action in the
   postoperative incision environment. In the acidic wound environment,
   PDA@ZIF‐8@AB NPs release H[2] that reduces oxidative stress and
   inflammation, providing sustained analgesic effects while promoting
   angiogenesis and cell migration. NIR irradiation activates PDA,
   inducing a photothermal effect that triggers controlled QX‐314 release
   and opens membrane ion channels, enhancing localized pain relief. NPs:
   nanoparticles; PCL@QX‐314 MSs: PCL microspheres loaded with QX‐314.

   The first step in this analgesic delivery system involves using ammonia
   borane (AB) as an H[2] donor,^[ [57]^16 ^] as it is an efficient
   hydrogen storage material that exhibits acid‐responsive H[2] release
   behavior, ensuring a stable supply of H[2] in the acidic
   microenvironment of postoperative wounds.^[ [58]^17 ^] First of all, AB
   is encapsulated within ZIF‐8 using a one‐step synthesis method,
   resulting in composite ZIF‐8@AB nanoparticles (NPs). Next, this
   composite undergoes a self‐polymerization reaction with dopamine
   hydrochloride to generate PDA@ZIF‐8@AB (Figure [59]1a). The
   acid‐sensitive degradation of PDA@ZIF‐8@AB ensures the gradual release
   of H[2], providing localized analgesic and therapeutic effects around
   the incision site. This H[2]‐release provides effective pain relief up
   to five days while promoting tissue regeneration, thereby facilitating
   a smoother healing process (Figure [60]1b).

   The second step addresses the inherent subjectivity of pain, which is
   influenced by factors such as genetics, psychological state, and prior
   pain experiences.^[ [61]^18 ^] To address the current
   “one‐size‐fits‐all” approach in clinical postoperative pain
   management,^[ [62]^19 ^] we incorporated a personalized pain management
   strategy into the system. In the current study, QX‐314, a quaternary
   ammonium derivative of lidocaine, which can enter primary sensory
   neurons through activated membrane ion channels, resulting in targeted
   regional anesthesia,^[ [63]^1 , [64]^20 ^] is applied to provide
   supplementary on‐demand analgesia. QX‐314‐loaded polycaprolactone (PCL)
   microspheres (MSs) are prepared using a water/oil/water (W/O/W)
   emulsion‐solvent evaporation method (Figure [65]1a). Under
   near‐infrared (NIR) laser irradiation, the photothermal conversion of
   PDA triggers the deformation of PCL MSs, enabling the on‐demand release
   of QX‐314. The activation of ion channels at elevated temperatures
   promotes the entry of QX‐314 into neurons, providing targeted pain
   relief (Figure [66]1b). This mechanism allows for personalized
   treatment tailored to individual patient pain sensitivity, effectively
   bridging the gap between standardized pain management and personalized
   care.

   The final step involves delivering the above‐mentioned formulations
   (PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs) to the surgical incision site
   using hyaluronic acid (HA)‐based dissolvable MNs (Figure [67]1b). We
   validated the effectiveness of this system using an animal model with
   plantar incisions and investigated the analgesic and promoting wound
   healing mechanisms of H[2] through RNA sequencing (RNA‐seq) techniques.

   This system integrates innovative analgesic strategies and regenerative
   therapies to tackle the dual challenges of postoperative pain relief
   and wound healing, in line with the trend of Enhanced Recovery After
   Surgery (ERAS).

2. Results

2.1. Optimization and Characterization of PDA@ZIF‐8@AB NPs

   We optimized the synthesis of PDA@ZIF‐8@AB NPs by assessing several
   critical parameters, including the reaction medium, the method of AB
   addition, the stirring duration for the preparation of ZIF‐8@AB, the
   concentration of dopamine hydrochloride (DA), and the reaction time for
   DA self‐polymerization and modification. This comprehensive analysis
   enabled us to establish the optimal preparation protocol for the NPs
   (Figure [68]S1, Supporting Information). Following PDA modification,
   the nanomaterial transitioned from a milky white to a deep gray color
   (Figure [69]2b). The microstructural characteristics of ZIF‐8,
   ZIF‐8@AB, and PDA@ZIF‐8@AB were examined using transmission electron
   microscopy (TEM). Both ZIF‐8 and ZIF‐8@AB exhibited uniform rhombic
   dodecahedral morphologies, consistent with previously reported
   structures.^[ [70]^21 ^] In contrast, PDA@ZIF‐8@AB displayed a
   morphology that was closer to spherical, featuring a rough surface with
   a granular texture (Figure [71]2a).^[ [72]^22 ^] The particle sizes
   were measured at 103 ± 3 nm for ZIF‐8, 135 ± 6 nm for ZIF‐8@AB, and 177
   ± 5 nm for PDA@ZIF‐8@AB (Figure [73]2c), indicating that the
   modifications led to a significant increase in particle size, likely
   due to the incorporation of PDA and AB into the structure. Following
   PDA modification, the surface charge of the NPs shifted from positive
   to negative, significantly enhancing their retention in biological
   tissues and thereby improving therapeutic efficacy (Figure [74]2d).^[
   [75]^23 ^]

Figure 2.

   Figure 2
   [76]Open in a new tab

   Characterization of PDA@ZIF‐8@AB NPs. a) TEM images of ZIF‐8, ZIF‐8@AB,
   and PDA@ZIF‐8@AB NPs. b) Visual appearance of ZIF‐8@AB and PDA@ZIF‐8@AB
   NPs dispersions. c) Particle size distribution of NPs as determined by
   DLS. d) Zeta potential measurements for NPs. e‐f) XRD patterns (e) and
   FT‐IR spectra (f) of different materials. g XPS spectra before and
   after etching for PDA@ZIF‐8@AB NPs. h) EDS mapping images of ZIF‐8@AB,
   PDA, PDA@ZIF‐8@AB NPs. i) EDS quantification of elemental composition
   in PDA@ZIF‐8@AB NPs. j) TGA curves of PDA@ZIF‐8 and PDA@ZIF‐8@AB NPs.
   k) DLS measurements of PDA@ZIF‐8@AB NPs in various media over 24 h. l)
   Photographic images of ZIF‐8@AB and PDA@ZIF‐8@AB NPs in different media
   at 0, 12, and 24 h.

   Further investigation of the crystalline structure of the NPs was
   conducted using X‐ray diffraction (XRD). The results indicated no
   significant changes in the position and intensity of the diffraction
   peaks for ZIF‐8@AB compared to blank ZIF‐8; however, notable
   alterations were observed in PDA@ZIF‐8@AB. These findings supported the
   TEM results, suggesting that the loading of AB did not disrupt the
   crystalline structure of ZIF‐8, while the surface polymerization of PDA
   did have an effect (Figure [77]2e). Fourier transform infrared
   spectroscopy (FT‐IR) confirmed the structured composition of the NPs.
   The distinct N‐B‐H stretching vibration peak at ≈1100 cm⁻^1 and the B‐H
   stretch around 2500 cm⁻^1 of AB in the ZIF‐8@AB spectrum confirmed the
   successful encapsulation of AB within the ZIF‐8 framework.
   Additionally, the N‐H peak between 1500 and 1600 cm⁻^1 in the ZIF‐8@AB
   spectrum indicated the successful incorporation of AB and its
   interaction with ZIF‐8. No significant shifts were observed in the
   ZIF‐8 and ZIF‐8@AB spectra, indicating that the crystalline structure
   of ZIF‐8 remained intact post‐loading. After PDA modification, the
   slight broadening and shifts in peaks, especially between 1400 and 1600
   cm⁻^1, indicated successful surface polymerization of PDA, introducing
   new functional groups without affecting the ZIF‐8 structure
   (Figure [78]2f). In the ultraviolet‐visible (UV‐vis) absorption
   spectrum, a new absorption peak between 280 and 290 nm was observed in
   the PDA@ZIF‐8@AB sample, which was absent in the ZIF‐8@AB spectrum.
   This peak was attributed to the π‐π* transition of the catechol groups
   present in dopamine, confirming the successful modification with PDA
   (Figure [79]S2c, Supporting Information).

   X‐ray photoelectron spectroscopy (XPS) was conducted to confirm the
   surface composition and structural modifications of NPs. Before
   etching, characteristic peaks corresponding to zinc (Zn 2p), nitrogen
   (N 1s), carbon (C 1s), and oxygen (O 1s) were observed, indicating the
   presence of the ZIF‐8 structure and successful PDA surface
   modification. The prominent O 1s signal confirmed effective PDA
   coating. After etching, the oxygen signal decreased significantly,
   while the Zn 2p peak became more pronounced, indicating the removal of
   the PDA layer and the exposure of the ZIF‐8 core. The appearance of a
   boron (B 1s) peak post‐etching provided evidence for the encapsulation
   of AB within the ZIF‐8 framework (Figure [80]2g). Additional
   confirmation was obtained through TEM mapping (Figure [81]2h) and
   energy‐dispersive X‐ray spectroscopy (EDS) analysis (Figure [82]2i),
   which visualized the distribution of carbon, zinc, and oxygen, further
   affirming the structure of the NPs. Based on thermogravimetric analysis
   (TGA), the drug loading of AB in the PDA@ZIF‐8@AB NPs was calculated to
   be ≈10% (Figure [83]2j), with an encapsulation efficiency of around 17%
   (Figure [84]S2a, Supporting Information).

2.2. Stable Dispersion and Enhanced Biocompatibility after PDA Modification

   To evaluate the impact of PDA modification on the stability of NPs,
   ZIF‐8@AB and PDA@ZIF‐8@AB NPs were placed in equal volumes of various
   solutions, including deionized water, phosphate‐buffered saline (PBS),
   Dulbecco's Modified Eagle Medium (DMEM), and fetal bovine serum (FBS),
   at the same concentration and allowed to stand. As shown in
   Figure [85]2l, both types of particles remained uniformly dispersed in
   all four solutions over a 24‐hour period, with no visible signs of
   sedimentation. Dynamic light scattering (DLS) measurements indicated
   that the particle size of ZIF‐8@AB (Figure [86]S2b, Supporting
   Information) and PDA@ZIF‐8@AB NPs (Figure [87]2k) in each solution did
   not exhibit significant changes during this time. The consistent
   particle size across different media demonstrated that the PDA
   modification did not compromise the stability of the NPs. This stable
   dispersion, without aggregation, suggested that the modified NPs could
   maintain uniformity in diverse biological environments, which was
   beneficial for practical applications.

   ZIF‐8 exhibits dose‐dependent biocompatibility. One study demonstrated
   that at concentrations below 30 µg mL^−1, ZIF‐8 did not exhibit
   significant cytotoxicity. However, at higher concentrations, the
   release of Zn^2⁺ ions induced mitochondrial reactive oxygen species
   (ROS) production, causing irreversible DNA damage and activating
   apoptotic pathways.^[ [88]^17a ^] In our research, we aimed to explore
   whether PDA modification could improve the biocompatibility of ZIF‐8
   NPs. The apoptosis rate of cells treated with ZIF‐8@AB and PDA@ZIF‐8@AB
   NPs was evaluated using an Annexin V‐FITC/PI apoptosis kit. As shown in
   Figure [89]3i, the apoptosis rate increased with the concentration of
   NPs; however, it was lower in the PDA@ZIF‐8@AB group compared to the
   ZIF‐8@AB group across all concentrations. This finding indicated that
   PDA modification significantly improved the biocompatibility of the
   NPs. At the highest concentration (100 µg mL^−1), the apoptosis rate
   for ZIF‐8@AB reached 30 ± 2%, whereas it was only 9.7 ± 1.0% for
   PDA@ZIF‐8@AB (Figure [90]3j). These results suggested that PDA coating
   reduced the cytotoxicity of ZIF‐8@AB NPs, thereby enhancing their
   potential for safe biomedical applications.

Figure 3.

   Figure 3
   [91]Open in a new tab

   Acid‐responsive H[2] release, photothermal properties, and cytotoxicity
   of PDA@ZIF‐8@AB NPs. a) TEM images of PDA@ZIF‐8@AB NPs at different
   time points under pH 7.4 and pH 6.8 conditions. b) Photographic images
   of PDA@ZIF‐8@AB NPs in MB‐Pt probe solution at various time intervals
   over 168 h. c) Thermal images of PDA@ZIF‐8@AB dispersions at different
   concentrations under NIR irradiation (1.0 W cm^−2) for 8 min. d)
   Photothermal conversion efficiency calculation for PDA@ZIF‐8@AB NPs. e)
   In vitro cumulative H[2] release profiles of PDA@ZIF‐8@AB NPs at pH 7.4
   and pH 6.8. f‐g) Temperature elevation of PDA@ZIF‐8@AB NPs at various
   concentrations (1.0 W cm^−2) and NIR power densities (500 µg mL^−1)
   over 8 min. h) Temperature cycling of PDA@ZIF‐8@AB NPs (500 µg mL^−1)
   under repeated NIR irradiation (1.5 W cm^−2) on‐off cycles. i) Flow
   cytometry analysis of apoptosis rates in cells treated with varying
   concentrations of ZIF‐8@AB and PDA@ZIF‐8@AB NPs. j) Quantification of
   apoptosis rates, comparing ZIF‐8@AB and PDA@ZIF‐8@AB NPs treatments at
   different concentrations. Data represented mean ± SEM from n = 3
   independent experiments. P value: ^* p < 0.05 and ^** p < 0.01.

2.3. Acid‐Triggered Degradation and Sustained H[2] Release from PDA@ZIF‐8@AB
NPs

   The acid‐triggered degradation of PDA@ZIF‐8@AB NPs was evaluated under
   different pH conditions (pH 7.4 and pH 6.8). TEM analysis revealed
   distinct degradation behaviors depending on the pH environment
   (Figure [92]3a). At pH 7.4, the NPs largely maintained their structural
   integrity throughout the 7‐day observation period, with only minor
   morphological changes observed. In contrast, at pH 6.8, degradation was
   notably accelerated. By day 5, the particles exhibited significant
   structural disintegration, with their shapes becoming progressively
   indistinct and blurred. By day 7, the NPs had almost entirely
   disintegrated, displaying clear signs of breakdown and degradation.
   These results underscored the acid‐sensitive nature of the PDA@ZIF‐8@AB
   NPs, suggesting their potential for targeted drug release in low‐pH
   environments, such as wound sites or areas of inflammation.

   H[2] was detected using the MB‐Pt probe,^[ [93]^24 ^] clearly
   demonstrating the crucial role of AB in the NP system's ability to
   generate H[2]. As illustrated in the series of photographs, under
   acidic conditions (pH 6.8), the color of the MB‐Pt probe solution
   containing PDA@ZIF‐8@AB NPs gradually faded over time (Figure [94]3b),
   while the solution with PDA@ZIF‐8 NPs exhibited almost no fading during
   the same period (Figure [95]S3c, Supporting Information). These
   observations suggested that AB was essential for the release of H[2] in
   the PDA@ZIF‐8@AB formulation.

   The cumulative release curves further demonstrated that under acidic
   conditions (pH 6.8), PDA@ZIF‐8@AB NPs continuously released H[2] over a
   period of 120 hours, with the release gradually plateauing between 120
   and 168 hours. In contrast, at pH 7.4, the release rate was slower
   (Figure [96]3e). Additionally, a comparison with the release curves of
   free AB indicated that encapsulating AB within the NPs facilitated for
   a more gradual and sustained release, highlighting the advantages of
   this system for applications requiring prolonged H[2] delivery (Figure
   [97]S3b, Supporting Information). These findings underscored the
   efficiency of PDA@ZIF‐8@AB NPs in achieving controlled, pH‐sensitive
   H[2] release, suggesting significant potential for therapeutic
   applications.

2.4. Excellent Photothermal Properties of PDA@ZIF‐8@AB NPs

   The modification of PDA not only reduced the toxicity of ZIF‐8 and
   enhanced its stability but also endowed the NPs with photothermal
   conversion capabilities, allowing for controlled drug release.^[
   [98]^25 ^] The photothermal properties of PDA@ZIF‐8@AB NPs were
   evaluated under 808 nm NIR laser irradiation, and the results
   demonstrated excellent photothermal conversion performance. As shown in
   the infrared thermal images and temperature variation curves
   (Figure [99]3c,f,g), the temperature increased with rising
   concentrations of the materials (0, 250, 500, 800 µg mL^−1) and NIR
   power densities (0.5, 1.0, 1.5 W cm^−2), following a nonlinear upward
   trend until reaching a plateau.

   At a laser power density of 1.0 W cm^−2, the 800 µg mL^−1 PDA@ZIF‐8@AB
   suspension reached a temperature of 59 °C within 8 min. In comparison,
   suspensions with concentrations of 500 µg mL^−1 and 250 µg mL^−1
   reached peak temperatures of 53 °C and 40 °C, respectively. The control
   group (0 µg mL^−1) exhibited negligible temperature changes under the
   same irradiation conditions (Figure [100]3f). When exposed to a power
   density of 1.5 W cm^−2, the 500 µg mL^−1 suspension exceeded 58 °C
   during laser irradiation, while the temperature increase was more
   moderate at 1.0 W cm^−2 and 0.5 W cm^−2, reaching ≈52 °C and 41 °C,
   respectively. Notably, PDA@ZIF‐8@AB NPs displayed excellent
   photothermal stability during cyclic NIR irradiation (Figure [101]3h),
   with consistent temperature elevations observed over five “on/off”
   cycles. The calculated photothermal conversion efficiency (η) for these
   NPs was 41% (Figure [102]3d), highlighting their potential as effective
   agents for photothermal therapy.

2.5. Preparation and Characterization of PCL@QX‐314 MSs

   PCL is a biodegradable and biocompatible polymer widely recognized for
   its excellent properties in biomedical applications, including low
   toxicity, mechanical flexibility, and superior thermal performance.^[
   [103]^26 ^] These characteristics make it particularly suitable for
   controlled‐release systems. Moreover, its ability to undergo
   temperature‐induced shape changes further enhances its potential in
   drug‐controlled release delivery systems.^[ [104]^27 ^] In this study,
   PCL and QX‐314 were selected as the core materials for the preparation
   of MSs following established methods,^[ [105]^28 ^] and the stabilizer
   in the inner aqueous phase was optimized to improve the quality of the
   MSs. Scanning electron microscopy (SEM) images (Figure [106]S4a–c,
   Supporting Information) revealed significant differences in surface
   morphology and size distribution based on the stabilizer used (Blank,
   PEG, and HA). MSs without stabilizers (Blank) exhibited irregular
   shapes and rough surfaces. In contrast, PEG‐stabilized MSs had smoother
   surfaces but larger particle sizes, which may limit their application
   in MN‐based drug delivery systems. However, HA‐stabilized MSs (Figure
   [107]4a,b) demonstrated higher density, smaller, and more uniform
   particle sizes, indicating their suitability for drug delivery
   applications. Cross‐sectional images obtained using ion‐slicing
   technology demonstrated a dense internal structure, suggesting enhanced
   mechanical stability (Figure [108]4b).^[ [109]^29 ^] This compact
   arrangement reflected a robust MS formation process, which could
   improve the overall durability of the MSs during handling and delivery.
   Particle size distribution analysis indicated that the majority of the
   MSs ranged between 3 and 10 µm, with an average diameter of 6.6 ± 1.5
   µm (Figure [110]4c).

Figure 4.

   Figure 4
   [111]Open in a new tab

   Characterization, thermal stability, and release profile of PCL@QX‐314
   MSs. a) SEM image showing the surface morphology of PCL@QX‐314 MSs. b)
   High‐magnification SEM images of individual PCL@QX‐314 MS. c) Size
   distribution histogram of PCL@QX‐314 MSs. d) TGA curves of PCL and
   PCL@QX‐314 MSs. e‐f) XRD patterns (e) and FT‐IR spectra (f) of
   different particles. g) DSC curves of PCL and PCL@QX‐314 MSs. h) SEM
   images showing the stability of PCL@QX‐314 MSs with multiple NIR
   irradiation cycles and without irradiation over 7 days. i) In vitro
   cumulative release profiles of QX‐314 from PCL@QX‐314 MSs under
   different pH conditions with and without NIR irradiation over 7 days.
   Arrows in the graph indicated the time points of NIR irradiation. Data
   represented mean ± SEM from n = 3 independent experiments.

   The XRD analysis demonstrated the crystalline characteristics of both
   PCL and QX‐314 (Figure [112]4e). In the physical mixture of PCL and
   QX‐314, both sets of peaks were observed, indicating no significant
   structural changes in either component. However, in the PCL@QX‐314 MSs,
   the reduction in the intensity of the QX‐314 peaks suggested that the
   drug's crystalline structure had been partially masked during
   encapsulation, indicating successful dispersion and encapsulation of
   QX‐314 within the PCL matrix. The FT‐IR spectrum (Figure [113]4f)
   provided deeper insights into the chemical interactions within the MSs.
   Key characteristic peaks of PCL, such as the C═O stretching vibration
   at 1724 cm⁻^1, were clearly observed. For QX‐314, the spectrum revealed
   characteristic peaks between 1630 and 1680 cm⁻^1 (C═O stretching) and
   3000–3100 cm⁻^1 (aromatic C─H stretching), confirming its chemical
   structure. In the encapsulated PCL@QX‐314 MSs, shifts and reductions in
   the intensity of these peaks, particularly in the C═O stretching
   region, suggested molecular interactions between PCL and QX‐314 within
   the MSs. These spectral changes confirmed successful drug encapsulation
   and provided evidence of the chemical interaction between the polymer
   and the encapsulated drug.

   TGA was employed to evaluate the thermal stability and drug loading
   efficiency of the MSs. The TGA curve (Figure [114]4d) showed a
   characteristic weight loss profile for PCL@QX‐314, featuring two
   primary stages of degradation. The initial weight loss corresponded to
   the degradation of the encapsulated drug, while the second stage was
   associated with the degradation of the PCL matrix. Based on the TGA
   results, the drug loading of the PCL@QX‐314 MSs was calculated to be
   ≈8%, with an encapsulation efficiency of around 45% (Figure [115]S4d,
   Supporting Information). These values confirmed the successful
   incorporation of QX‐314 into the PCL matrix. Differential Scanning
   Calorimetry (DSC) was employed to further evaluate the thermal
   properties of the PCL@QX‐314 MSs (Figure [116]4g). The melting point of
   PCL was observed at around 57.5 °C, which aligned with the established
   thermal properties of PCL.^[ [117]^30 ^] In the PCL@QX‐314 MSs, minor
   shifts in the thermal transition peaks were detected, indicating that
   the encapsulated drug influenced the thermal behavior of the polymer.
   Based on these DSC results and additional thermal testing data, a laser
   irradiation condition of 1.0 W cm^−2 for 5 min was selected for the
   controlled release of QX‐314. The temperature rise induced by this
   irradiation also facilitated the opening of TRPV1 channels, allowing
   QX‐314 to enter cells and exert its analgesic effect.

2.6. Thermally Induced Deformation and Photothermal‐Controlled Release of
PCL@QX‐314 MSs

   The thermally induced deformation and photothermal‐controlled release
   of PCL@QX‐314 MSs were systematically investigated by co‐incubating the
   MSs with PDA@ZIF‐8@AB NPs. Since PCL lacks inherent photothermal
   conversion capabilities, the heating effect under NIR irradiation was
   primarily driven by the PDA component, which efficiently absorbed NIR
   energy and generated localized heat. Morphological analysis using SEM
   (Figure [118]4h) revealed that, in the absence of NIR irradiation, the
   MSs maintained their structural integrity with smooth, spherical
   surfaces over a 7‐day period. However, upon exposure to NIR irradiation
   (1.0 W cm^−2, 5 min), significant structural changes occurred in the
   MSs after multiple irradiation cycles. By the fourth to sixth cycles,
   the MSs exhibited clear deformation, characterized by surface
   irregularities and partial collapse. These results suggested that the
   photothermal effect induced by NIR irradiation caused softening and
   deformation of the PCL matrix, thereby facilitating the release of the
   encapsulated QX‐314.

   The release of QX‐314 from PCL@QX‐314 MSs was evaluated under various
   pH conditions (Figure [119]4i). In the absence of NIR irradiation, the
   release of QX‐314 remained minimal (below 20%) at both pH 7.4 and 6.8
   over a period of 7 days, indicating the strong stability of the MSs
   without external stimuli. However, under NIR irradiation, a significant
   stepwise increase in release was observed with each irradiation cycle,
   resulting in over 90% of QX‐314 being released after six cycles in both
   pH environments. These results suggested that pH had a little effect on
   the release, with the primary trigger being the NIR‐induced
   photothermal effect. The stepwise release pattern highlighted the
   effectiveness of NIR as a precise trigger for controlled drug release,
   demonstrating the potential of PCL@QX‐314 MSs for on‐demand drug
   delivery.

2.7. Morphology, Mechanical Properties, and Skin Penetration Efficiency of
PDA@ZIF‐8@AB and PCL@QX‐314‐Loaded MNs

   PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs were encapsulated in HA‐based
   dissolvable MN patches (HA‐PP MN) using a layer‐by‐layer casting
   technique (Figure [120]5a).^[ [121]^31 ^] Optical images revealed that
   the HA‐PP MNs formed a well‐organized 10 × 10 array on a 7 mm × 7 mm
   patch, featuring sharp and well‐defined needle tips (Figure [122]5b).
   The incorporation of PDA@ZIF‐8@AB NPs imparted a dark coloration to the
   tips of the needles, which were anticipated to facilitate precise drug
   delivery into the skin upon application (Figure [123]5b). Each needle
   exhibited a base diameter of 400 µm and a height of 850 µm.
   High‐magnification SEM images confirmed the structural integrity of the
   needles, displaying uniform conical shapes and well‐preserved surface
   morphology (Figure [124]5c).

Figure 5.

   Figure 5
   [125]Open in a new tab

   Preparation, characterization, mechanical properties, and photothermal
   properties of HA‐PP MN patch. a) Schematic illustration of the
   fabrication process of the HA‐PP MN patch. b) Optical images of the
   fabricated HA‐PP MN patch, with an enlarged view showing the MN array.
   c) SEM images of the HA‐PP MN patch. d) Force‐displacement curve of the
   HA‐PP MN patch (Inset: optical images of MNs before and after the
   mechanical strength test). e) Fluorescence imaging of the HA‐PP MN
   patch showing the distribution of PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs
   within the MNs. f) Optical images of the HA‐PP MN patch applied to
   porcine skin. g) Images of frozen sections of MNs inserted in porcine
   skin. The arrows in the images indicated the PCL@QX‐314 MSs loaded
   within the MNs, which successfully remained in the skin after MN
   insertion. h) Stability test of the HA‐PP MN patch under multiple
   cycles of NIR irradiation (1 W cm^−2). i) Infrared thermographic images
   of different MN patches under NIR irradiation (1 W cm^−2), illustrating
   temperature elevation over time. HA‐PP MN patch: the HA‐based MN patch
   loaded with PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs.

   The drug loading capacity of the patches was assessed, revealing that
   each patch contained ≈285 ± 9 µg of PDA@ZIF‐8@AB NPs and 241 ± 7 µg of
   PCL@QX‐314 MSs, indicating successful loading of both components
   (Figure [126]S5a, Supporting Information). Mechanical strength testing
   (Figure [127]5d) demonstrated that each needle could withstand forces
   of up to 1.5 ± 0.5 N before bending or breaking, significantly
   exceeding the 0.1 N required for effective skin penetration.^[ [128]^32
   ^] This confirmed that the MN patches possessed sufficient mechanical
   robustness for transdermal application without compromising their
   structural integrity.

   Fluorescence visualization was performed using Rhodamine B and
   fluorescein sodium as model drugs for NPs and MSs, respectively
   (Figure [129]5e). Following application to porcine skin, the MNs
   achieved over 99% drug delivery, successfully penetrating the skin
   (Figure [130]5f). Analysis of the frozen sections revealed that
   Rhodamine B remained well encapsulated within the NPs, while
   fluorescein sodium exhibited slight diffusion due to the inherent burst
   release characteristics of water‐soluble drugs in MSs (Figure [131]5g).
   The overlay of fluorescence signals confirmed the successful
   co‐delivery of both components into the skin tissue. These findings
   highlighted the effective design and mechanical stability of the HA‐PP
   MN patches, enabling efficient transdermal delivery of therapeutic
   agents.

2.8. Photothermal Activation and Visualization of Laser‐Controlled Drug
Release in HA‐PP MN Patches

   To investigate the photothermal response and controlled drug release of
   HA‐based MN patches, we prepared three groups of patches: blank HA MN,
   HA‐PZA MN containing only PDA@ZIF‐8@AB (PZA) NPs, and HA‐PP MN. The
   objective was to determine whether the PDA component in the patches
   could generate sufficient heat under NIR irradiation to facilitate
   targeted drug release. Upon exposure to an 808 nm NIR laser (1.0
   W cm^−2), thermal imaging results showed minimal temperature changes in
   the blank HA MN patches. In contrast, both HA‐PZA MN and HA‐PP MN
   patches exhibited rapid temperature increases, with the HA‐PP MN group
   reaching 51 °C after 5 min, thereby meeting the experimental
   requirements for triggering the controlled release of QX‐314 from the
   PCL@QX‐314 MSs (Figure [132]5i). This heating effect confirmed the role
   of PDA@ZIF‐8@AB NPs in converting NIR energy into localized heat within
   the MN patches.

   To evaluate laser‐controlled drug release, HA‐PP MN patches were
   inserted into a skin‐mimicking agarose gel and subjected to repeated
   NIR irradiation cycles (Figure [133]5h). Rhodamine B, encapsulated in
   NPs as a stable model drug, exhibited minimal diffusion into the gel
   throughout the cycles, likely due to the structural stability of the
   NPs under heat. This limited diffusion indicated that repeated NIR
   stimulation did not significantly affect the drug release rate from the
   NPs. In contrast, fluorescein sodium, encapsulated within the PCL MSs,
   demonstrated a different behavior. Designed to respond to the
   photothermal effect, fluorescein sodium exhibited a stepwise increase
   in release with each NIR cycle. The heat generated by PDA in the MN
   patches effectively triggered the diffusion of fluorescein sodium,
   resulting in an intensified and prolonged fluorescence signal in the
   gel with successive irradiations. This controlled, on‐demand release
   pattern highlighted the potential of HA‐PP MN patches for applications
   requiring precise, NIR‐triggered drug delivery.

2.9. PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs Retention and Drug Release in Local
Tissue Microenvironment

   To evaluate the retention and sustained release of PDA@ZIF‐8@AB NPs and
   PCL@QX‐314 MSs in the local tissue microenvironment, fluorescent HA‐PP
   MN patches were applied for in vivo rat paw skin puncture testing.

   Fluorescence images were captured daily to monitor the retention and
   release of NPs and MSs. The results showed that after the application
   of HA‐PP MN patches, the NPs and MSs were effectively delivered to the
   tissue and remained stably localized at the application site.
   PDA@ZIF‐8@AB NPs primarily relied on an acid‐responsive mechanism for
   the gradual release of the fluorescent model drug, with fluorescence
   remaining detectable for up to 5 days. The presence or absence of NIR
   irradiation had minimal impact on their release and degradation (Figure
   [134]S6a, b, Supporting Information). In contrast, PCL@QX‐314 MSs
   showed significant changes under NIR irradiation. Without NIR exposure,
   PCL@QX‐314 MSs remained localized in the tissue with minimal drug
   release (Figure [135]S6a, Supporting Information). However, with daily
   NIR exposure, the fluorescence intensity gradually decreased with each
   light cycle, and by Day 5, the fluorescence had nearly disappeared
   (Figure [136]S6b, Supporting Information).

2.10. Enhanced Cell Migration and Wound Healing Driven by the H[2]‐Releasing
of MN Patches

   Human umbilical vein endothelial cells (HUVECs, Procell Life Science &
   Technology Co., Ltd., China) and mouse embryonic fibroblast cells
   (NIH‐3T3, Procell Life Science & Technology Co., Ltd., China) were
   utilized to assess the cell migration and angiogenesis‐promoting
   abilities of HA‐PP MN patches. In the cell migration assays for HUVECs
   (Figure [137]6a) and NIH‐3T3 cells (Figure [138]S7a, Supporting
   Information), the groups containing PDA@ZIF‐8@AB NPs demonstrated
   significantly enhanced wound healing rates compared to the Control and
   HA‐PZ MN groups. Specifically, at 24 h, the wound healing rate for the
   HA‐PZ MN group was 29 ± 2%, while the healing rates for the HA‐PZA MN,
   HA‐PZA MN + NIR, HA‐PP MN, and HA‐PP MN + NIR groups were ≈40%
   (Figure [139]6c). At 48 h, the wound healing rate for the HA‐PZ MN
   group was 50 ± 3%, while the groups containing PDA@ZIF‐8@AB NPs showed
   healing rates exceeding 70% (Figure [140]6d). NIH‐3T3 cells exhibited a
   similar trend (Figure [141]S7b, c, Supporting Information). These
   results highlighted that the enhanced effect was attributed to the H[2]
   generation capability of PDA@ZIF‐8@AB NPs, which has been shown to
   facilitate cell migration and promote wound healing processes.

Figure 6.

   Figure 6
   [142]Open in a new tab

   Evaluation of wound healing and angiogenic effects of various MN
   patches on HUVEC cells. a) Bright field images showing the scratch
   assay of HUVEC treated with different MN patches at 0, 24, and 48 h. b)
   Bright field and Calcein‐AM staining images of HUVEC cells after
   treatment with different MN patches, demonstrating tube formation and
   angiogenic network structures. c) Quantification of wound healing rate
   at 24 h for each treatment group. d) Quantification of wound healing
   rate at 48 h for each treatment group. e) Number of junctions in HUVEC
   networks formed after treatment with each MN patch. f) Total segment
   length in HUVEC networks for each treatment. Data represented mean ±
   SEM from n = 3 independent experiments. P value: ns means no
   significance, ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, and ^**** p <
   0.0001.

   In the tube formation assay with HUVECs (Figure [143]6b), the HA‐PZA MN
   and HA‐PP MN groups, particularly under NIR irradiation, demonstrated
   more robust and well‐defined tubular structures. Quantitative analysis
   showed that the HA‐PZ MN group had a number of junctions of 12.0 ± 1.6
   and total segment length of 620 ± 7, while the groups containing
   PDA@ZIF‐8@AB NPs had a number of junctions greater than 30
   (Figure [144]6e) and total segment length exceeding 1200
   (Figure [145]6f). This effect was primarily attributed to the H[2]
   generation driven by PDA@ZIF‐8@AB NPs, which stimulated angiogenic
   activity.

2.11. Prolonged and Potent Analgesic Effect of HA‐PP MN Patches with NIR
Activation for Postoperative Pain

   Before initiating the formal analgesic studies, we assessed the
   potential skin effects of the photothermal response from the HA‐PZA MN
   patches, which showed the highest temperature rise. The patches were
   applied to the dorsal (Figure [146]S8a, Supporting Information) and
   plantar skin (Figure [147]S8b, Supporting Information) of rats and
   subjected to six intermittent NIR irradiation cycles. Observations
   recorded during and 48 h post‐irradiation showed no visible signs of
   skin damage, confirming that the photothermal effect of the HA‐PZA MN
   patches did not cause adverse effects on the skin.

   For the analgesic study, rats underwent plantar incision surgery to
   establish a postoperative pain model (Figure [148]7a).^[ [149]^33 ^]
   Mechanical pain sensitivity was evaluated with von Frey filaments
   (Figure [150]7c–e; Figure [151]S9a, b, Supporting Information), showing
   that the S and S + HA‐PZ MN group experienced an immediate reduction in
   paw withdrawal threshold to 21 ± 3% and 18 ± 3% of baseline, indicating
   heightened pain sensitivity. This sensitivity only began to subside
   from day 4 post‐surgery, returning to 88 ± 6% and 91 ± 4% of baseline
   by day 7. In contrast, the HA‐PZA MN, HA‐PZA MN + NIR, and HA‐PP MN
   groups, featuring the release of H[2], demonstrated significant
   analgesic effects, with comparable pain relief lasting up to the fifth
   day post‐surgery, maintaining paw withdrawal thresholds of 68 ± 4%, 73
   ± 11%, and 74 ± 3%, respectively. Notably, the HA‐PP MN + NIR group
   exhibited the most potent analgesic effect, reaching near‐complete pain
   threshold recovery within 24 h post‐surgery and maintaining a high paw
   withdrawal threshold of 100 ± 9% on day 5. This robust analgesic effect
   was attributed to the combination of H[2] release from PDA@ZIF‐8@AB NPs
   and NIR‐triggered release of QX‐314, working synergistically to inhibit
   nociceptive signaling and improve pain outcomes. In comparison, the
   HA‐PP Gel + NIR group showed significantly weaker pain relief,
   indicating lower transdermal delivery efficiency and drug utilization
   compared to the MN formulation. For clinical comparison, the analgesic
   effect of a single subcutaneous morphine injection lasted only 4 h.
   Although the 3 mg kg^−1 morphine dose significantly reduced pain
   sensitivity, the level of pain relief was not as effective as the HA‐PP
   MN + NIR group. With repeated injections, the analgesic effect of
   morphine slightly diminished.

Figure 7.

   Figure 7
   [152]Open in a new tab

   Behavioral tests for pain assessment following plantar incision surgery
   in rats and the effects of various MN treatments. a) Schematic of the
   plantar incision model and timeline of behavioral tests, including the
   von Frey filament and Hot plate assessments. b) Body weight changes of
   rats across 15 days post‐surgery in different treatment groups. Data
   were expressed as mean ± SEM (n = 16). c‐e) Percentage change in
   mechanical paw withdrawal threshold across treatment groups compared to
   baseline. f‐h) Percentage change in thermal withdrawal latency across
   treatment groups compared to baseline. i‐j) Bubble charts illustrating
   the time‐course improvement in mechanical and thermal pain thresholds
   across groups, with intensity changes visualized at each time point.
   All data were expressed as mean ± SEM (n = 8). P value: ns means no
   significance, ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, and ^**** p <
   0.0001.

   Thermal pain sensitivity tests further supported these findings
   (Figure [153]7f–h; Figure [154]S9c, d, Supporting Information). In the
   S group, the thermal pain threshold decreased immediately post‐surgery
   to 36 ± 3%, indicating abnormal thermal nociception. The HA‐PP MN + NIR
   group consistently showed increased paw withdrawal latency on the hot
   plate, starting at 2 h post‐surgery with a threshold of 82 ± 5%, and
   maintaining an elevated threshold above baseline up to day 7 (114 ±
   4%). This sustained reduction in thermal sensitivity demonstrated
   effective pain relief compared to the S group. The movies captured the
   thermal pain testing of the S and HA‐PP MN + NIR groups 24 h
   post‐surgery, showcasing the marked differences in pain sensitivity
   (Movies [155]S1–[156]S6, Supporting Information). In groups with only
   PDA@ZIF‐8@AB NPs or one additional condition (PCL@QX‐314 MSs or NIR
   irradiation), analgesic effects in thermal pain also persisted for over
   five days, though with reduced intensity. This suggested that while
   PDA@ZIF‐8@AB's H[2] generation contributed significantly to pain
   relief, the combination of NIR‐triggered QX‐314 release in the HA‐PP MN
   + NIR group maximized the analgesic effect. The bubble plots
   (Figure [157]7i,j) provided a visual representation of the analgesic
   effects of different treatments for mechanical pain (Figure [158]7i)
   and thermal pain (Figure [159]7j), highlighting the superior pain
   relief in the HA‐PP MN + NIR group.

   Throughout the 15‐day study, the rats’ body weights were monitored to
   assess general health (Figure [160]7b). All groups showed steady weight
   gain, indicating that the treatments, including the NIR‐triggered HA‐PP
   MN patches, did not adversely affect the health or growth of rats.
   These findings collectively underscored the potential of HA‐PP MN
   patches with NIR, to deliver prolonged and controlled postoperative
   pain relief safely.

2.12. Suppression of Pain Pathway Activation by HA‐PP MN Patches with NIR

   Studies confirm that 24 h post‐surgery is a critical phase for pain
   pathway activation and related gene expression.^[ [161]^34 ^] Our
   behavioral findings also showed significant analgesic effects of HA‐PP
   MN patches at this point. Thus, we selected this time to investigate
   the pain‐suppressive mechanisms induced by MN patches.

   In the L4/L5 DRG, we analyzed c‐Fos and TRPV1 as markers associated
   with nociceptive neuron activation (Figure [162]8a,e; Figure [163]S10a,
   Supporting Information). TRPV1, a key receptor in pain transmission, is
   highly expressed in DRG nociceptive sensory neurons and plays an
   essential role in relaying peripheral pain signals to the central
   nervous system.^[ [164]^35 ^] C‐Fos serves as an indicator of neuronal
   activity,^[ [165]^36 ^] with increased c‐Fos^+/TRPV1^+ expression
   denoting heightened sensory neuron response to painful stimuli.
   Compared to the Control group, increased c‐Fos/TRPV1 co‐positive
   neurons in the S group indicated heightened nociceptive activation
   post‐incision (22.4 ± 1.5% vs 41 ± 3%). HA‐PZA MN and HA‐PP MN + NIR
   treatments significantly reduced this activation, with the HA‐PP MN +
   NIR group showing the most pronounced effect (22.2 ± 1.7%), suggesting
   effective inhibition of pain signal transmission.

Figure 8.

   Figure 8
   [166]Open in a new tab

   Immunofluorescence analysis of neuronal and glial activation in DRG and
   spinal cord following MN treatments. a) Representative
   immunofluorescence images showing c‐Fos and TRPV1 expression in DRG
   neurons under different treatment conditions. b) Representative
   immunofluorescence images for c‐Fos and NeuN in the SG area of spinal
   cord dorsal horn under different treatment conditions. c) GFAP
   immunostaining in the spinal cord dorsal horn revealed astrocyte
   activation across groups. d) Iba‐1 immunostaining in the spinal cord
   dorsal horn demonstrated microglial activation across groups. e‐h)
   Quantification of c‐Fos⁺TRPV1⁺/TRPV1⁺ (e), c‐Fos⁺NeuN⁺/NeuN⁺ (f) cell
   percentages, and integrated fluorescence intensities for GFAP (g) and
   Iba‐1 (h) across groups, respectively. All data were expressed as mean
   ± SEM (n = 9 slices from three rats). P value: ^* p < 0.05, ^** p <
   0.01, ^*** p < 0.001, and ^**** p < 0.0001.

   In the substantia gelatinosa (SG) region of the spinal cord, we
   analyzed c‐Fos and NeuN as markers of central pain processing
   (Figure [167]8b,f; Figure [168]S10b, Supporting Information). The SG
   region is pivotal for relaying and modulating nociceptive input from
   the periphery to higher brain centers.^[ [169]^37 ^] Elevated
   co‐expression of these markers in the S group (32 ± 4%) indicated
   central sensitization and enhanced transmission of pain signals. Both
   HA‐PZA MN and HA‐PP MN + NIR groups significantly lowered c‐Fos and
   NeuN expression in the SG, with HA‐PP MN + NIR showing the strongest
   inhibition (12.7 ± 1.4%), suggesting reduced central sensitization.

   Additionally, we assessed glial cell activation by staining GFAP
   (astrocyte marker, Figure [170]8c,g) and Iba‐1 (microglial marker,
   Figure [171]8d,h) in the dorsal horn of the spinal cord, as astrocytes
   and microglia are involved in neuroinflammatory processes that
   exacerbate pain.^[ [172]^38 ^] Elevated GFAP and Iba‐1 expression in
   the S group reflected neuroinflammatory activation consistent with pain
   sensitization. Both HA‐PZA MN and HA‐PP MN + NIR treatments,
   particularly the HA‐PP MN + NIR group, significantly reduced GFAP and
   Iba‐1 levels. The decreased glial activation indicated that HA‐PP MN +
   NIR not only suppressed neuronal activation but also inhibited
   neuroinflammatory pathways involved in pain amplification.

   Together, these findings suggested that HA‐PP MN patches, particularly
   with NIR activation, effectively modulated both peripheral and central
   nociceptive pathways, offering a promising strategy for targeted,
   sustained postoperative pain relief.

2.13. Accelerated Wound Healing Induced by the H[2] Released from MN Patches

   Daily photographic observations of the surgical wounds demonstrated
   that all H[2]‐releasing MN‐treated groups, including HA‐PZA MN, HA‐PZA
   MN + NIR, HA‐PP MN, and HA‐PP MN + NIR, achieved more rapid wound
   closure compared to the S, HA MN, HA‐PZ MN, HA‐PP Gel + NIR, and
   Morphine s.c. groups, which showed slower healing (Figure [173]9a).

Figure 9.

   Figure 9
   [174]Open in a new tab

   Evaluation of wound healing, collagen deposition, and angiogenesis
   following MN treatments. a) Representative images of wound healing
   across different treatment groups, showing wound closure progression.
   b) H&E‐stained tissue sections of the wound area. c) Masson's trichrome
   staining for collagen fibers in wound tissues. d) Immunofluorescence
   staining for CD31 in wound sections to assess angiogenesis. e‐g)
   Quantification of epidermal healing rate (e), CVF (f), and integrated
   fluorescence intensity of CD31 (g) across groups. All data were
   expressed as mean ± SEM (n = 3). P value: ns means no significance, ^*
   p < 0.05, ^** p < 0.01, and ^**** p < 0.0001. CVF: collagen volume
   fraction.

   On day 5, H&E staining highlighted that H[2]‐releasing MN patches
   facilitated more organized re‐epithelialization and tissue recovery at
   the incision site, with reformed epithelial layers and well‐structured
   tissue observed in these groups (Figure [175]9b,e). Notably, this
   effect was absent in the HA‐PZ MN group, underscoring the role of H[2]
   generation in promoting effective wound repair.

   By day 7, Masson's trichrome staining showed robust collagen deposition
   in the H[2]‐generating groups, indicating enhanced tissue regeneration
   (Figure [176]9c). Collagen organization was markedly improved, which
   supported better structural recovery in the wound area
   (Figure [177]9f). This trend suggested that H[2] release aided in the
   collagen synthesis necessary for wound strength and stability. CD31
   immunofluorescence staining further revealed an increase in
   microvascular density in the wounds treated with H[2]‐releasing MNs by
   day 7, pointing to a significant pro‐angiogenic effect that supported
   sustained tissue repair (Figure [178]9d,g). This enhanced
   vascularization was likely to promote nutrient delivery and waste
   removal, key factors in effective wound healing.

2.14. Transcriptome Analysis of DRG with H[2]‐Releasing MN Patches Applied
After Surgery

   To elucidate the potential molecular mechanisms underlying the
   analgesic and wound‐healing effects of H[2] released from MN patches,
   RNA‐seq was performed on L4/5 DRG tissues collected 24 h post‐surgery
   from three groups: Control (C group), Surgery (S group), and Treatment
   (T group, S + HA‐PZA MN). Considering the analgesic properties of
   QX‐314 are well‐established,^[ [179]^20 ^] this transcriptome study
   just focused on the specific role of H[2] in postoperative pain relief
   and tissue repair. Sequencing analysis yielded consistently high
   mapping rates across all groups, with an average exceeding 97%,
   ensuring data robustness and reliable transcript alignment (Table
   [180]S1, Supporting Information). Principal component analysis (PCA)
   demonstrated a clear separation between the Control, Surgery, and
   Treatment groups, indicative of distinct transcriptional shifts
   associated with surgical trauma and the reparative influence of
   H[2]‐releasing MN treatment (Figure [181]10a,b).

Figure 10.

   Figure 10
   [182]Open in a new tab

   Transcriptomic analysis of DRG tissues from Control, Surgery, and
   Treatment groups. a‐b) PCA plots showing the distribution of samples
   among the Control, Surgery, and Treatment groups. c‐d) Volcano plots
   highlighting differentially expressed genes between the Surgery vs
   Control groups (c) and the Treatment vs Surgery groups (d). e) Heatmap
   of gene expression profiles. f) GO enrichment analysis for the Surgery
   vs Control comparison, highlighting up‐regulated pathways. g) GO
   enrichment analysis for the Surgery vs Control comparison, showing
   significant down‐regulated pathways. h) KEGG pathway enrichment
   analysis of differentially expressed genes in the Surgery vs Control
   comparisons (up‐regulated). i) KEGG pathway enrichment analysis of
   differentially expressed genes in the Surgery vs Control comparisons
   (down‐regulated). Data were presented from n = 3 biological replicates
   per group. Treatment group: group with H[2]‐releasing MN patches
   applied after surgery.

   Differential expression analysis revealed significant transcriptomic
   changes across conditions. In the comparison between Surgery and
   Control groups, 192 genes were significantly differentially expressed,
   with 164 down‐regulated and 28 up‐regulated, reflecting substantial
   transcriptomic alterations due to surgical stress (Figure [183]10c).
   When examining differences between the Treatment and Surgery groups,
   112 genes exhibited significant regulation, with 73 up‐regulated and 39
   down‐regulated, emphasizing the impact of H[2] release on the DRG
   transcriptome in response to surgical insult (Figure [184]10d). A
   comprehensive heatmap of gene expression profiles further illustrated
   these differences, showing distinct clustering patterns that
   underscored the unique transcriptional landscapes associated with each
   experimental condition (Figure [185]10e).

   Subsequently, we performed functional annotation of the differentially
   expressed genes (DEGs) across the groups. Gene Ontology (GO), Kyoto
   Encyclopedia of Genes and Genomes (KEGG), and Reactome pathway analyses
   were conducted based on the identified DEGs. The transcriptomic
   comparison between the Surgery and Control groups highlighted key early
   responses to surgical stress. The GO analysis revealed significant
   alterations in pathways associated with inflammation and pain, receptor
   and signal transduction, immune response and cell differentiation,
   metabolic regulation and enzyme activity, and muscle and cytoskeletal
   functions (Figure [186]10f,g). Specifically, pathways such as cellular
   response to histamine, prostaglandin receptor activity, icosanoid
   receptor activity and S‐methyltransferase activity were up‐regulated in
   the Surgery group, suggesting enhanced inflammatory and pain signaling
   (Figure [187]10f). In contrast, MHC class I receptor activity, receptor
   serine/threonine kinase binding, positive regulation of gamma‐delta T
   cell differentiation, and myofilament were down‐regulated
   (Figure [188]10g). KEGG pathway analysis identified the neuroactive
   ligand‐receptor interaction pathway as up‐regulated (Figure [189]10h),
   while the cytokine‐cytokine receptor interaction and hematopoietic cell
   lineage pathways were down‐regulated (Figure [190]10i). Reactome
   analysis further revealed the up‐regulation of signal transduction and
   receptor‐related pathways, such as G alpha q signaling events and
   effects of PIP2 hydrolysis. Additionally, pathways associated with
   oxidative stress, such as biological oxidations, were activated.
   Changes in inflammation regulation were observed, including alterations
   in inflammasome activity and TNF receptor binding, along with
   modifications in muscle function, as evidenced by the modulation of
   muscle contraction (Figure [191]11e). In summary, surgical stress
   significantly impacted tissue inflammation, pain signaling, and cell
   function, thereby influencing the overall healing and recovery process.

Figure 11.

   Figure 11
   [192]Open in a new tab

   Enrichment analyses comparing Surgery vs Control groups and Treatment
   vs Surgery groups. a) GO enrichment analysis for the Treatment vs
   Surgery comparison, highlighting up‐regulated pathways. b) GO
   enrichment analysis for the Treatment vs Surgery comparison, showing
   significant down‐regulated pathways. c) KEGG pathway enrichment
   analysis of differentially expressed genes in the Treatment vs Surgery
   comparisons (up‐regulated). d) KEGG pathway enrichment analysis of
   differentially expressed genes in the Treatment vs Surgery comparisons
   (down‐regulated). e) Reactome enrichment analysis for the Surgery vs
   Control groups. f) Reactome enrichment analysis for the Treatment vs
   Surgery groups. Data were presented from n = 3 biological replicates
   per group.

   To explore how H[2]‐releasing MNs control postoperative pain and
   accelerate wound healing, we analyzed transcriptomic differences
   between the Treatment and Surgery groups. This analysis specifically
   focused on the therapeutic effects of H[2]. The GO analysis revealed
   significant alterations in pathways associated with inflammation, pain,
   receptor binding and signal transduction, immune response, metabolic
   and redox processes, and cell differentiation. Notably, pathways
   involved in pericyte cell differentiation, monoamine transmembrane
   transporter activity, NADPH activity, NAD^+ binding, water channel
   activity, and MHC class II protein complex were up‐regulated in the
   Treatment group (Figure [193]11a). In contrast, pathways such as
   prostaglandin receptor activity, icosanoid receptor activity,
   acetylcholine receptor binding, and MAPK kinase activity were
   down‐regulated, suggesting an inhibition in pain and inflammatory
   signaling (Figure [194]11b). KEGG pathway analysis identified the
   up‐regulation of hematopoietic cell lineage and cytokine‐cytokine
   receptor interaction pathways in the Treatment group, reflecting a
   modulation of immune and inflammatory responses (Figure [195]11c,d).
   Reactome analysis highlighted several pathways related to extracellular
   matrix metabolism and immune signaling. Pathways such as
   glycosaminoglycan metabolism, extracellular matrix degradation, and
   dermatan sulfate biosynthesis were up‐regulated, indicating an enhanced
   wound healing process. Additionally, the up‐regulation of T cell
   receptor (TCR) signaling and passive transport by aquaporins pathways
   was observed, consistent with the GO analysis findings
   (Figure [196]11f). The above enrichment results suggested that
   H[2]‐releasing MNs alleviated postoperative pain and promoted wound
   healing by inhibiting inflammation and oxidative stress, suppressing
   pain signal transduction and nociceptive receptor activity, and
   enhancing tissue regeneration.

   Based on the transcriptomic analysis, we further validated the
   sequencing results using RT‐qPCR for key differentially expressed
   genes. These genes were selected based on their relevance to our study,
   particularly in inflammation, pain signaling, and tissue healing
   processes. RT‐qPCR results confirmed the findings from RNA‐seq and
   showed that the regulatory trend of the selected genes was consistent
   with the sequencing results. For example, Ptgdr, which is involved in
   inflammatory pathways and is known to play a role in pain modulation,
   was significantly up‐regulated in the Surgery group compared to the
   Control group. Treatment with H[2]‐releasing MNs effectively reduced
   the Ptgdr expression, suggesting its potential role in alleviating
   postoperative pain (Figure [197]12a). Similar trends were observed for
   genes related to tissue regeneration. The gene expression levels of
   Eln, Fam180a, and Twist1 were significantly higher in the Treatment
   group, indicating the enhanced tissue repair and regeneration processes
   facilitated by H[2] release (Figure [198]12b,c,f). Notably, the
   down‐regulation of Acta2 and Cldn24 in the Surgery group was reversed
   in the Treatment group, suggesting a restoration of cellular functions,
   particularly in cell‐cell adhesion and tissue integrity, suggesting an
   improvement in tissue homeostasis following H[2] treatment
   (Figure [199]12d,e).

Figure 12.

   Figure 12
   [200]Open in a new tab

   qPCR analysis of key genes associated with pain and wound healing. a‐f)
   Relative mRNA expression of selected genes in Control, Surgery, and
   Treatment groups. Genes analyzed included: Ptgdr (a), Eln (b), Fam180a
   (c), Acta2 (d), Cldn24 (e), and Twist1 (f). All data were presented as
   mean ± SEM (n = 4). P value: ^* p < 0.05, ^** p < 0.01 and ^*** p <
   0.001.

2.15. Evaluation of Biocompatibility and Biosafety of HA‐PP MN Patches

   The biocompatibility and biosafety of MN patches were comprehensively
   assessed through in vitro and in vivo studies. In vitro cytotoxicity
   tests on HUVEC (Figure [201]S12b‐d, Supporting Information) and NIH‐3T3
   (Figure [202]S13b‐d, Supporting Information) cells demonstrated
   excellent cell viability across varying concentrations of the patch
   components (HA, PDA@ZIF‐8@AB NPs, and PCL@QX‐314 MSs). The
   Calcein‐AM/PI staining indicated minimal cell death, with cell
   viability consistently exceeding 90% across all treatment groups,
   indicating that the materials did not adversely affect cellular
   integrity (Figures [203]S12a and [204]S13a, Supporting Information).

   In vivo biocompatibility was evaluated using hemolysis assays and organ
   function markers. Hemolysis rates in all treated groups were well below
   the 5% safety threshold, suggesting minimal red blood cell lysis and
   excellent blood compatibility (Figure [205]S14a, Supporting
   Information).^[ [206]^39 ^] Serum analyses for liver (ALT, AST) and
   kidney (BUN, CR) function revealed no significant differences between
   treatment and control groups, indicating that HA‐PP MN patches with NIR
   irradiation did not compromise major organ functions (Figure
   [207]S14b‐e, Supporting Information). Furthermore, H&E staining of
   internal organs (heart, liver, spleen, lungs, and kidneys) showed no
   signs of inflammation, tissue damage, or morphological abnormalities,
   supporting the biocompatibility of the MN patches in vivo (Figure
   [208]S15c, Supporting Information).

   Upon application of MN patches to rat skin, the puncture sites were
   observed to completely close within 30 min, demonstrating the minimal
   invasiveness of the MNs (Figure [209]S15a, Supporting Information).
   Subsequent H&E staining of the skin at application sites confirmed the
   absence of inflammatory infiltration or structural damage, further
   verifying the safety of repeated application and NIR‐triggered drug
   release (Figure [210]S15b, Supporting Information). Collectively, these
   results underscored the excellent biosafety profile of HA‐PP MN
   patches, affirming their suitability for safe, transdermal therapeutic
   applications.

3. Discussion

   In this study, we introduced a novel H[2]‐based therapeutic system
   delivered via intelligent photothermal MNs for postoperative pain
   management and wound healing. Using a dual‐phase MN system, we
   successfully achieved sustained hydrogen encompassment and controlled
   release of QX‐314 at the surgical sites. Our findings underscore the
   efficacy of H[2] as a non‐traditional analgesic and its potential to
   accelerate wound healing. Besides, the photothermally triggered release
   of QX‐314 enables on‐demand pain relief, optimizing pain management
   based on individual patient pain thresholds.

   Several studies have explored the use of H[2] as a therapeutic agent,
   particularly for its antioxidant^[ [211]^40 ^] and anti‐inflammatory^[
   [212]^41 ^] properties, which play a crucial role in pain management
   and tissue repair. Li et al.^[ [213]^12 ^] demonstrated that H[2]
   reduced neuroinflammation and microglial activation in a mouse model by
   inhibiting the ASK1/JNK/p38/MMP9 signaling pathway, thereby alleviated
   postoperative pain. Similarly, Chen et al.^[ [214]^42 ^] utilized H[2]
   as an antioxidant to alleviate chronic wound inflammation and promote
   healing. Our HA‐PP MN exhibits similar therapeutic effects and can
   provide effective analgesia for up to 5 days.

   Our transcriptomic analysis provided deeper insights into the molecular
   mechanisms underlying the observed analgesic and regenerative effects
   of H[2] therapies. The data revealed distinct biological pathways
   involved in the response to surgical injury and subsequent treatment.
   Surgical trauma led to inflammatory responses and oxidative stress,
   simultaneously upregulating pain‐related signaling pathways and
   receptor activities, such as prostaglandin receptor activity. This
   suggest that the inflammation and other responses triggered by surgical
   stimuli play a critical role in postoperative pain. In contrast,
   H[2]‐releasing MNs led to a shift in gene expression reflecting both
   analgesic and regenerative effects. The up‐regulation of antioxidant
   stress pathways in the Treatment group suggested that H[2] may mitigate
   oxidative damage, promote tissue repair, and accelerate wound healing.
   Additionally, the up‐regulated pathways related to pericyte
   differentiation, NADPH activity, and extracellular matrix regulation
   suggested enhanced tissue repair and resolution of inflammation.
   Modulation of immune‐related pathways further highlighted the role of
   H[2] in regulating inflammation. The down‐regulation of pain‐related
   signaling pathways, such as prostaglandin receptor activity and MAPK
   signaling, reinforced the analgesic potential of H[2]‐releasing MNs.
   Building on previous research, these findings emphasize the therapeutic
   potential of H[2] in analgesia and tissue regeneration by inhibiting
   inflammation and oxidative stress.

   Delivering H[2] in a controlled and sustained manner, especially for
   localized applications^[ [215]^43 ^] are still challenging. In our
   study, we addressed this issue by using ZIF‐8 NPs as a scaffold for AB,
   allowing for controlled and sustained H[2] release within the MN
   system. ZIF‐8 is a structurally stable and biocompatible metal‐organic
   framework (MOF) widely used in drug delivery systems.^[ [216]^44 ^]
   Previous research by Fang et al.^[ [217]^45 ^] showed that modifying
   ZIF‐8 could improve the stability and efficacy of anticancer drugs,
   further supporting its potential in biomedical applications. Notably,
   ZIF‐8's pH‐responsive degradation is particularly advantageous in
   postoperative acidic environments, where it facilitates the release of
   encapsulated drugs.^[ [218]^46 ^] In our system, the encapsulation of
   AB within ZIF‐8 NPs enabled sustained and localized H[2] release around
   the surgical sites, overcoming the typical rapid diffusion associated
   with H[2] therapies.

   On‐demand analgesia strategies have become a focal point in recent pain
   management research. For example, Song et al.^[ [219]^47 ^] introduced
   a nanoparticle platform that provides sustained, on‐demand pain relief
   using moderate‐intensity ultrasound, while Reeder et al.^[ [220]^48 ^]
   developed a soft, biodegradable cooling device that achieves on‐demand
   analgesia by lowering nerve temperature. These systems allow for
   real‐time modulation of pain relief, pushing the boundaries of
   personalized pain management. QX‐314 represents a promising approach
   for on‐demand analgesia due to its permanent positive charge, which
   prevents it from crossing cell membranes.^[ [221]^49 ^] As a result, it
   must be delivered intracellularly, such as through activated ion
   channels, to exert its effects.^[ [222]^50 ^] Based on this property,
   we incorporated QX‐314 into the MN system, utilizing the photothermal
   responsiveness of PDA and the temperature‐sensitive properties of PCL
   MSs. This system enabled controlled, on‐demand release of QX‐314 under
   NIR irradiation, facilitating the entry of QX‐314 into neurons and
   providing localized pain relief at the surgical site. This approach,
   offering precise modulation of QX‐314 release in response to individual
   pain thresholds, represents a significant step forward in personalized
   pain management, offering a more targeted alternative to traditional
   methods.

   MN transdermal delivery systems offer several advantages, including
   minimally invasive administration, high patient compliance, and
   controlled release capabilities.^[ [223]^51 ^] Recent studies have
   explored dual‐release systems to address diverse therapeutic needs. For
   instance, Gu et al.^[ [224]^52 ^] developed a glucagon patch with both
   immediate and biologically responsive release modes for the emergency
   treatment of severe hyperglycemia. Younas et al.^[ [225]^53 ^] created
   a MN‐based dual‐release system for rapid analgesic delivery combined
   with sustained antimicrobial release to promote wound healing. In our
   study, we developed a dual‐phase MN system that integrates both
   sustained and on‐demand release strategies. H[2] was continuously
   released to reduce inflammation, relieve pain and promote tissue
   repair, while QX‐314 was released on‐demand for immediate, localized
   pain relief. This dual‐phase approach not only offers personalized pain
   management but also enhances the precision of postoperative care,
   marking a notable advancement over traditional analgesic strategies and
   facilitating more tailored therapeutic interventions.

   While our system demonstrates promising results, there are still some
   limitations. First, the need for an external NIR device to trigger
   QX‐314 release limits the system's practicality and integration. An
   autonomous, fully integrated system would be more desirable for
   clinical use, and further development is needed to eliminate the
   reliance on external equipment. Additionally, while H[2] delivery via
   ZIF‐8 NPs effectively reduces inflammation and promotes tissue
   regeneration, controlling the release rate over extended periods
   remains a challenge. The long‐term stability of H[2] delivery, as well
   as the potential for local tissue saturation, require further
   optimization and investigation.

4. Conclusions

   In conclusion, the smart photothermal‐responsive MN system developed in
   this study successfully integrated sustained H[2] release and
   photothermal‐triggered QX‐314 delivery, achieving both long‐lasting
   analgesia and accelerated tissue repair. The ZIF‐8‐based hydrogen
   release system provided localized, long‐lasting anti‐inflammatory
   effects, while controlled QX‐314 release allowed for customized pain
   relief, addressing the limitations of conventional pain management
   strategies. These findings underscore the therapeutic potential of this
   dual‐action MN system, offering a personalized approach that meets the
   analgesic and regenerative needs in postoperative care. This technology
   not only holds promise for enhancing recovery protocols but also lays
   the groundwork for the future development of responsive drug delivery
   systems.

5. Experimental Section

Materials

   All reagents used in this study were of analytical grade and purchased
   from commercial suppliers. These included ammonia borane (AB, 97%,
   Aladdin, China), 2‐Methylimidazole (2‐MI, ≥98%, Aladdin, China), zinc
   nitrate hexahydrate (99%, Chronchem, China), dopamine hydrochloride
   (98%, Macklin, China), QX‐314 (≥98%, Sigma Aldrich, USA), hyaluronic
   acid (HA, 40–100 kDa, Macklin, China), polycaprolactone (PCL, Mn 80000,
   Sigma Aldrich, USA), polyvinyl alcohol (PVA, Mw 9000–10000, Sigma
   Aldrich, USA), sodium chloride (99.5%, Macklin, China), HA (10 kDa,
   Rhawn, China), methylene blue (≥98%, Macklin, China), chloroplatinic
   acid hexahydrate (Pt ≥ 37.5%, Aladdin, China), fluorescein sodium salt
   (70%, Aladdin, China), and rhodamine B (95%, Solarbio, China). DMEM and
   FBS were obtained from Procell (China). The Annexin V‐FITC/PI Apoptosis
   Kit was purchased from Elabscience (China), while both Cell Counting
   Kit‐8 (CCK‐8) and Calcein‐AM/PI Cell Viability/Cytotoxicity Assay Kit
   were sourced from Beyotime (China). The antibodies used in this study
   included TRPV1 (Affinity Biosciences, DF10320), NeuN (Abcam, ab177487),
   c‐Fos (Servicebio, GB12069), Iba‐1 (Abcam, ab283319), GFAP (Cell
   Signaling Technology, 3670S), Goat Anti‐Rabbit IgG H&L (FITC
   conjugated, Servicebio, GB22303), Goat Anti‐Mouse IgG H&L (Cy3
   conjugated, Servicebio, GB21301), Goat Anti‐Rabbit IgG H&L (Cy3
   conjugated, Servicebio, GB21303), and Goat Anti‐Mouse IgG H&L (FITC
   conjugated, Servicebio, GB22301).

Synthesis of PDA@ZIF‐8@AB NPs

   ZIF‐8@AB NPs were synthesized using a one‐step method.^[ [226]^54 ^]
   Zinc nitrate hexahydrate (104.125 mg) was dissolved in 700 µL of
   deionized water and sonicated for 5 min at room temperature to ensure
   complete dissolution. Meanwhile, 2‐MI (2.05 g) and AB (50 mg) were
   dissolved in 10 mL of deionized water, and the mixture was stirred at
   500 rpm for 5 min to create a homogeneous solution. The zinc nitrate
   solution was then added dropwise to the 2‐MI‐AB mixture while stirring
   at 1500 rpm at room temperature for 30 min. After this, the resulting
   mixture was centrifuged at 10 000 rpm for 15 min to collect the
   precipitate, which was washed with deionized water, yielding ZIF‐8@AB.
   The precipitate was then resuspended in 10 mL of deionized water, and
   the pH was adjusted to 8.5 using Tris‐HCl buffer. Subsequently, 10 mg
   of dopamine hydrochloride was added, and the solution was stirred at
   1500 rpm for 2 h at room temperature. After 2 h, the resulting product
   was collected by centrifugation at 10 000 rpm for 15 min, washed three
   times with deionized water, and then freeze‐dried under vacuum for 48 h
   to obtain PDA@ZIF‐8@AB NPs. The blank NPs were prepared with the same
   procedures without adding AB to the 2‐MI solution.

Preparation of PCL@QX‐314 MSs

   Drug‐loaded MSs were prepared using an emulsification‐solvent
   evaporation method.^[ [227]^55 ^] In brief, 30 mg of QX‐314 and 20 mg
   of HA (40‐100 kDa) were dissolved in 1 mL of deionized water to form
   the inner aqueous phase (W1). Separately, 250 mg of PCL was dissolved
   in 5 mL of dichloromethane to create the organic phase (O). The two
   solutions were then emulsified for the first time using an ultrasonic
   processor under ice bath conditions, with a power setting of 150 W for
   1 min, resulting in a water‐in‐oil (W/O) primary emulsion. Next, the
   primary emulsion was quickly transferred to 10 mL of a PVA (2%, w/v)
   and sodium chloride (1%, w/v) solution (W2). The mixture was
   hand‐shaken at a speed of 100 shakes min^−1 for 1 min, resulting in a
   water‐in‐oil‐in‐water (W1/O/W2) multiple emulsion. The multiple
   emulsion was then rapidly poured into 20 mL of a PVA (0.2%, w/v)
   solution and stirred at 300 rpm at room temperature for 4 h to allow
   for the evaporation of dichloromethane, leading to the solidification
   of the polymer matrix and the formation of MSs. The products were
   collected by centrifugation at 6000 rpm for 15 min, washed three times
   with deionized water, and then freeze‐dried under vacuum for 48 h to
   obtain PCL@QX‐314 MSs. The preparation steps for the blank MSs were
   identical, except that no QX‐314 was added to the inner aqueous phase
   (W1).

Characterization of PDA@ZIF‐8@AB NPs

   The microstructure of the nanoparticles was characterized using TEM
   (JEOL, JEM‐2100). The particle size, polydispersity index (PDI), and
   zeta potential of the samples were measured using a Malvern Zetasizer
   (Malvern Instruments, Zetasizer Nano ZSP), with samples prepared in
   deionized water at a concentration of 1 mg mL^−1. The crystal structure
   of the nanoparticles was analyzed by XRD (PANalytical, XPert Pro) over
   a 2 ϴ range from 10° to 80°. FT‐IR was conducted using an FT‐IR
   spectrometer (Thermo Fisher Scientific, Nicolet iS50 FTIR). Samples
   were prepared using the KBr pellet method to examine interactions
   between functional groups in the samples. UV‐vis spectra of the
   prepared samples were recorded using a UV‐Vis spectrophotometer
   (Shimadzu, UV‐2600). XPS (Thermo Fisher Scientific, ESCALAB Xi+) was
   utilized to identify the types and chemical states of elements present
   on the materials’ surfaces. XPS measurements were conducted both before
   and after the etching process. Prior to etching, surface elements were
   analyzed to obtain the composition of the original material. The
   etching depth was set to 20 µm, which was determined based on the
   particle sizes of ZIF‐8@AB NPs and PDA@ZIF‐8@AB NPs. After the etching
   process, XPS analysis was repeated to examine the internal elements of
   the NPs. Elemental mapping was performed using a field emission
   transmission electron microscope (JEOL, JEM‐F200). Finally, TGA
   (Mettler‐Toledo, TGA2 / DSC3) was conducted to determine the drug
   loading capacity.

Measurement of Stability and Biocompatibility of PDA@ZIF‐8@AB NPs

   ZIF‐8@AB and PDA@ZIF‐8@AB NPs were separately suspended in four
   different solvents: deionized water, PBS, DMEM, and FBS. The samples
   were allowed to sit at 4 °C, and their dispersibility was observed
   through photographs taken at 12 and 24 h. Additionally, DLS was used to
   measure the particle size, allowing for the analysis of the impact of
   PDA modification on material stability.

   To evaluate the biocompatibility of the nanoparticles, NIH‐3T3 cells
   were utilized. The cells were cultured in DMEM supplemented with 10%
   FBS, penicillin (100 units mL^−1), and streptomycin (100 µg mL^−1).
   Different concentrations of ZIF‐8@AB and PDA@ZIF‐8@AB NPs were
   co‐cultured with the cells for 24 h. Subsequently, the treated cells
   were stained using an Annexin V‐FITC/PI apoptosis kit according to the
   manufacturer's instructions.

Acid‐Responsive Degradation and H[2] Release Behavior of PDA@ZIF‐8@AB NPs In
Vitro

   To investigate the degradation characteristics of the nanoparticles in
   acidic environments, PDA@ZIF‐8@AB NPs were suspended in PBS at pH
   values of 7.4 and 6.8, maintaining equal concentrations. The entire
   system was placed on a shaker (150 rpm) at 37 °C. At specific time
   points, namely days 1, 2, 5, and 7, 100 µL samples were collected from
   the PBS. After drying at room temperature, these samples were analyzed
   using TEM to evaluate the extent of degradation.

   H[2] has the capability to rapidly reduce MB to colorless reductive MB
   under the catalysis of Pt nanocrystals.^[ [228]^56 ^] Therefore, MB‐Pt
   probe was used to indirectly detect H[2] release. A probe solution was
   prepared containing MB at a concentration of 300 µg mL^−1 and Pt
   nanocrystals (3 nm) at a concentration of 13.9 µg mL^−1. This
   concentration ensured that the MB remained within the detection range
   of the UV‐vis spectrophotometer and could fully absorb all H[2]
   released from PDA@ZIF‐8@AB NPs. The absorbance of MB was measured at
   664 nm. 10 mg of PDA@ZIF‐8@AB NPs were dispersed in 4 mL of the MB‐Pt
   probe solution at different pH values (7.4 and 6.8) and incubated on a
   shaker (150 rpm) at 37 °C. The release of hydrogen was quantified by
   monitoring the absorbance changes of the solution at predetermined time
   points (0, 1, 2, 4, 6, 12, 24, 48, 72, 96, 120, 144, and 168 hours),
   while the color changes of the solution were photographed.
   Subsequently, an in vitro cumulative release curve was plotted based on
   the calculated data.

Evaluation of the Photothermal Properties of PDA@ZIF‐8@AB NPs

   To evaluate the photothermal properties of PDA@ZIF‐8@AB NPs, a series
   of experiments were performed. First, PBS solutions containing
   different concentrations of PDA@ZIF‐8@AB NPs (0, 250, 500, and 800
   µg mL^−1) were irradiated with an 808 nm NIR laser at a power density
   of 1.0 W cm^−2 for 8 min. Temperature changes were continuously
   monitored using a thermal infrared imager (FLIR SYSTEMS AB,
   FLIRSC7700M), capturing infrared images at 1‐minute intervals for each
   sample. Furthermore, the 500 µg mL^−1 solution of PDA@ZIF‐8@AB was
   subjected to various laser power densities (0.5, 1.0, and 1.5 W cm^−2)
   for 8 min, with temperature variations similarly monitored. To assess
   the stability of the photothermal response, the PDA@ZIF‐8@AB suspension
   underwent cyclic NIR irradiation, focusing on the temperature
   fluctuations during exposure. The 500 µg mL^−1 solution was irradiated
   with the 808 nm laser at a power density of 1.5 W cm^−2 for 8 min,
   followed by cooling to room temperature with the laser turned off. This
   process was repeated five times to analyze photothermal stability. The
   photothermal conversion efficiency (η) was calculated using established
   methods.^[ [229]^57 ^]

Characterization of PCL@QX‐314 MSs

   The morphology of the MSs was analyzed using SEM, and particle sizes
   were measured using the ImageJ software based on the SEM images.
   Ion‐slicing technology (TESCAN, TESCAN Amber equipped with C‐ToF‐SIMS)
   was employed to obtain cross‐sections of the PCL@QX‐314 MSs for
   observing the internal structure. Structural characterization of the
   MSs was conducted using XRD and FT‐IR, with the detection methods being
   the same as those used for the NPs. The thermal properties of the
   samples were evaluated using TGA and DSC (Mettler‐Toledo, TGA2/DSC3),
   with TGA employed to calculate the drug loading capacity and DSC used
   to evaluate the thermal transitions of the materials.

Thermal Response Characteristics and QX‐314 Release Behavior of PCL@QX‐314
MSs In Vitro

   Equal concentrations (500 µg mL^−1) of PCL@QX‐314 MSs and PDA@ZIF‐8@AB
   NPs were dispersed in PBS solution and incubated on a shaker at 150 rpm
   and 37 °C. For the NIR irradiation group, samples were exposed to NIR
   irradiation (1.0 W cm^−2 for 5 min) at designated time points (24, 48,
   72, 96, 120, and 144 h), while the control group remained unirradiated.
   Samples from the NIR irradiation group were collected after 0, 2, 4,
   and 6 cycles of laser exposure, while samples from the control group
   were collected at 0, 3, 5, and 7 days. The morphological changes of the
   MSs were examined using SEM.

   To investigate the release behavior of QX‐314 at varying pH levels (7.4
   and 6.8), 10 mg of PCL@QX‐314 MSs and an equal amount of PDA@ZIF‐8@AB
   NPs were added to each PBS solution. The mixtures were then incubated
   on a shaker at 150 rpm and 37 °C. For the NIR irradiation group,
   samples were exposed to NIR irradiation (1.0 W cm^−2 for 5 min) at
   specific time points (24, 48, 72, 96, 120, and 144 h). At predetermined
   intervals (0, 1, 2, 4, 6, 12, 24, 48, 72, 96, 120, 144, and 168 h), 200
   µL of supernatant was collected from each system and stored at 4 °C,
   with 200 µL of fresh PBS added to maintain volume. The samples were
   analyzed using a UV‐vis spectrophotometer at 193 nm to calculate the
   cumulative release rate of QX‐314.

Fabrication and Characterization of MN Patches

   PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs (3%, w/v) were dispersed in a
   Tris‐HCl buffer solution (pH 8.0) containing 18% (w/v) HA to obtain the
   matrix solution. This solution was injected into polydimethylsiloxane
   (PDMS, Dow Corning, USA) molds and vacuumed for 2 h, then the molds
   were transferred to a high‐speed centrifuge and centrifuged at 4200 rpm
   for 5 min. After removing the excess polymer solution, additional blank
   HA matrix solution was injected into the molds, which were then
   subjected to vacuum treatment for 4 h and dried in a 40 °C oven for 48
   h. The resulting dual‐phase MN patches (HA‐PP MN) were subsequently
   peeled off from the PDMS molds. HA‐PZ MN loaded only PDA@ZIF‐8 NPs,
   while HA‐PZA MN contained PDA@ZIF‐8@AB NPs without any MSs, and HA MN
   was prepared from a blank matrix solution. Rhodamine B was used as a
   model drug for the NPs, and fluorescein sodium served as a model drug
   for the MSs, allowing for fluorescence visualization of the MN patches.
   Furthermore, we also prepared a gel formulation (HA‐PP Gel) using the
   same concentration of HA (18%, w/v), which loaded equal amounts of
   PDA@ZIF‐8@AB NPs and PCL@QX‐314 MSs. This was done to compare the
   transdermal efficiency and therapeutic effects between the MNs and the
   conventional formulation.

   The morphology of the MN patches was observed using an optical
   microscope (LEICA, TL3000 Ergo) and SEM. Mechanical strength was
   assessed with a force measurement system (Mark‐10, ESM303). A single MN
   was fixed vertically on a steel platform, and pressure was applied at a
   constant speed of 2.1 mm min^−1. The pressure‐displacement curve was
   recorded to evaluate the mechanical response of the MN. To evaluate the
   skin penetration capability of the HA‐PP MNs, isolated porcine skin was
   used in experiments and observed under an optical microscope after
   applying the MNs. Subsequently, the porcine skin was embedded and
   sliced into 12 µm thick sections using a cryostat microtome (Thermo
   Fisher Scientific, HM525 NX) for analysis under an inverted
   fluorescence microscope (Olympus, IX73).

Photothermal Properties and Laser‐Controlled Drug Release of MN Patches

   To assess the NIR irradiation activation of PDA@ZIF‐8@AB NPs in the MN
   patches, the HA MN patches, HA‐PZA MN patches, and HA‐PP MN patches
   were exposed to an 808 nm NIR laser (1.0 W cm^−2) for 5 min. The
   temperature changes and thermal images of the MN patches were collected
   and monitored by a thermal infrared imager in real time.

   To examine the potential influence of PDA's photothermal effect on drug
   release from NPs and MSs, fluorescent HA‐PP MN patches were inserted
   into a skin‐mimicking agarose gel (1.4%, w/v). The patches underwent
   laser irradiation for 5 min, followed by a 60‐minute cooling period
   with the laser off. This “on/off” light cycle was repeated for a total
   of 6 times. The diffusion of the two model drugs within the agarose gel
   was then evaluated using an inverted fluorescence microscope.

   To evaluate the retention and sustained release of NPs and MSs in the
   local tissue microenvironment, fluorescent HA‐PP MN patches were
   applied for in vivo puncture testing. To minimize interference from
   factors like wound exudates on visualization, non‐surgical rat paw skin
   was used. During MN preparation, the matrix pH was adjusted to 6.8,
   creating a local acidic environment upon in vivo dissolution, which
   triggers the degradation and release of NPs. For the NIR irradiation
   group, the samples were exposed daily to an 808 nm NIR laser (1.0
   W cm^−2) for 5 min. Fluorescence images were captured over a 5‐day
   period, and fluorescence analysis was used to assess the duration of
   NPs and MSs retention and release in the local tissue.

Promotion of Migration and Angiogenesis by MN Patches In Vitro

   HUVEC and NIH‐3T3 cells were utilized to evaluate the impact of MN
   patches on cell migration. The culture was maintained in a humidified
   incubator at 37 °C with 5% CO[2]. Cells were seeded in 6‐well plates
   and cultured for 24 h. A 200 µL pipette tip was then used to create a
   uniform scratch in the center of each well. After washing with PBS to
   remove any floating cells, the cells were subjected to different
   conditions, divided into six groups: Control group, HA‐PZ MN group,
   HA‐PZA MN group, HA‐PZA MN with NIR laser group, HA‐PP group, and HA‐PP
   with NIR laser group. The MN patches were initially dissolved in DMEM
   containing 10% FBS. For the groups requiring NIR irradiation, the
   patches were subjected to 5 min of NIR exposure before being added to
   the wells. Cell migration was recorded at 0, 24, and 48 h.

   To evaluate HUVEC tube formation, 100 µL of thawed Matrigel (Abwbio,
   0827065) was added to each well of a 24‐well plate and incubated at 37
   °C for 1 h to solidify. Then, 1.5 × 10^5 HUVECs were seeded onto the
   Matrigel and exposed to various treatments. Tube formation was observed
   using an inverted fluorescence microscope, and live cells were
   visualized with a Calcein‐AM kit. ImageJ software was utilized for
   quantitative analysis of the tube structures.

Development of Plantar Incision Model

   All procedures were approved by the Experimental Animal Welfare Ethics
   Committee, Zhongnan Hospital of Wuhan University (Approval No:
   ZN2024100). Healthy male Sprague‐Dawley rats aged 6–8 weeks, sourced
   from China Three Gorges University, were kept in a controlled
   environment with humidity maintained at 40–60% and a temperature of 22
   ± 2 °C. The animals were provided free access to food and water, kept
   on a 12‐hour light cycle, and their body weights were recorded daily
   during the experiment. All rats were randomly assigned to 10 groups (n
   = 8 per group). Except for the Control group, all underwent plantar
   incision surgery and received various treatments (S, S + HA MN, S +
   HA‐PZ MN, S + HA‐PZA MN, S + HA‐PZA MN + NIR, S + HA‐PP MN, S + HA‐PP
   MN + NIR, S + Morphine s.c., and S + HA‐PP Gel + NIR groups).

   The plantar incision model was established following the previously
   described method.^[ [230]^8 ^] The rats were anesthetized via
   intraperitoneal injection of 2% sodium pentobarbital. After
   disinfection, a longitudinal incision of 1 cm was made on the left hind
   paw, positioned 2 mm from the heel, penetrating to the muscle while
   preserving the integrity of the muscle attachments at both ends. The
   skin was closed with 5‐0 sutures. The rats in the Control group
   received anesthesia, disinfection, and antibiotic treatment without
   undergoing any surgical procedures. After the recovery of the righting
   reflex, the appropriate treatment measures were immediately
   administered to each group. For the MN and gel groups, the treatments
   were applied directly to the incision site as a single treatment
   immediately post‐surgery. The morphine group received subcutaneous
   injections (3 mg kg^−1) once a day around the incision site immediately
   after the surgery and within 3 days after the surgery. The NIR
   irradiation group was exposed to a 5‐minute daily laser treatment at
   the wound site for 6 days post‐surgery.

Behavior Test for Postoperative Pain

   Behavioral assessments for postoperative pain were conducted at
   specific intervals (baseline, 0, 2, 4, 6, 12, 24, 48, 72, 96, 120, 144,
   and 168 h after surgery). Mechanical pain sensitivity in the rats was
   evaluated using von Frey filaments. Prior to each session, the animals
   were allowed a 30‐minute acclimation period in the testing environment.
   A series of calibrated filaments with varying forces (1.0, 1.4, 2.0,
   4.0, 6.0, 8.0, 10.0, and 15.0 g) were applied sequentially to the area
   around the plantar incision. Each filament was inserted into the tissue
   three times, with a 5‐minute interval between applications. The
   pressure was gradually increased until a withdrawal response was
   observed, with the force at which this occurred recorded as the paw
   withdrawal threshold. Additionally, thermal pain sensitivity was
   assessed. For three days prior to testing, rats were acclimatized to
   the hot plate for 30 min each day. During the test, the rats were
   placed on a heated surface set to 52.5 ± 0.5 °C, and the latency to
   withdraw or lick their paw was measured. Each rat underwent three
   trials, spaced 10 min apart, with a cut‐off time of 30 s implemented to
   prevent potential tissue damage.

Immunofluorescence Staining

   The L4/L5 DRG and spinal cord segments from the ipsilateral side (model
   side) of rats were collected 24 h post‐surgery for immunofluorescence
   staining. Samples were prepared as frozen sections, with the spinal
   cord cut to a thickness of 20 µm and the DRG to 10 µm. The expression
   levels of c‐Fos, TRPV1, NeuN, GFAP, and Iba‐1 were evaluated to
   investigate the activation of pain pathways. Positive neurons were
   quantified, and integrated fluorescence signal was calculated using
   ImageJ software.

Assessment of Surgical Wound Healing and Histological Analysis

   For a duration of seven days post‐surgery, photographs were taken daily
   to document the healing process of the surgical incision. On the fifth
   postoperative day, skin samples surrounding the plantar incision were
   collected for H&E staining to evaluate the healing of the damaged skin.
   Similarly, on the seventh postoperative day, samples were obtained for
   Masson's trichrome staining and CD31 immunofluorescence staining, aimed
   at assessing collagen deposition and microvascular density,
   respectively.

RNA Sequencing and Data Analysis

   RNA‐seq technology was applied to investigate the mechanisms by which
   H[2] released from MN patches contributed to postoperative analgesia
   and wound healing. At 24 hours post‐surgery, the L4/5 DRG were
   harvested from the Control group (C group), Surgery group (S group),
   and the S+HA‐PZA MN group (Treatment, T group) for RNA‐seq library
   preparation and analysis. Each group contained three biological
   replicates. A total of 0.5 µg of RNA from each sample was used for
   library preparation, which was performed using the VAHTS mRNA Capture
   Beads 2.0 kit (Vazyme, China). Poly‐T oligo‐attached magnetic beads
   were used to purify mRNA. RNA integrity was assessed using the Agilent
   Bioanalyzer 2100 system (Agilent Technologies, USA). Library
   construction and quality control were carried out following standard
   protocols. Sequencing was performed on the Illumina NovaSeq 6000
   platform (Illumina Inc., USA), with a paired‐end 150 bp (PE150) read
   length.

   Raw data were processed using in‐house Perl scripts for quality
   control, including the removal of adaptor‐contaminated reads,
   low‐quality reads, and reads containing a high proportion of N bases.
   Clean reads were aligned to the mRatBN7.2 reference genome using HISAT2
   v2.1.0,^[ [231]^58 ^] with genome and annotation files downloaded from
   the ENSEMBL database
   ([232]http://ftp.ensembl.org/pub/current_gtf/rattus_norvegicus/Rattus_n
   orvegicus.mRatBN7.2.112.chr.gtf.gz). Gene‐level quantification was
   performed using HTSeq v0.6.0,^[ [233]^59 ^] and gene expression levels
   were normalized and expressed as Fragments Per Kilobase Millon Mapped
   Reads (FPKM).

   Differential gene expression analysis was conducted using DESeq2
   v1.6.3.^[ [234]^60 ^] Genes with a p value < 0.05 and an absolute
   |log2FC| ≥ 1 were considered significantly differentially expressed.
   Functional enrichment analysis, including GO, KEGG, and Reactome
   pathway analysis, was performed. GO and KEGG enrichment were carried
   out using the clusterProfiler R package,^[ [235]^61 ^] while Reactome
   pathway analysis was conducted using the ReactomePA package.^[ [236]^62
   ^] Enrichment was evaluated using a hypergeometric test, with a p value
   < 0.05 considered statistically significant. GO analysis included three
   categories: Biological Process (BP), Molecular Function (MF), and
   Cellular Component (CC). The Reactome pathway analysis was conducted by
   selecting the first 30 items according to the ascending order of p
   values.

RT‐qPCR Analysis of Key Genes

   The RNA concentration was assessed using a NanoDrop spectrophotometer
   (Thermo Scientific, Waltham, MA, United States). 1 µg of total RNA was
   used for complementary DNA synthesis using the HiScript IV RT SuperMix
   for qPCR (+gDNA wiper) Kit (Vazyme cat# R423‐01). Quantitative PCR was
   carried out on a Rotor GeneQ (Qiagen) cycler with ChamQ Universal SYBR
   qPCR Master Mix (Vazyme cat# Q711‐03) using primers for target genes
   and β‐actin as an internal control. Primers were designed for the
   following key genes based on transcriptomic enrichment analysis: Ptgdr,
   Eln, Fam180a, Acta2, Cldn24, and Twist1. All transcript levels were
   normalized to β‐actin mRNA using the 2^−ΔΔCT method and each PCR
   reaction was run in duplicate for each sample and repeated at least
   twice. The primer sequences are shown in Table [237]S2 (Supporting
   Information).

In Vitro Cytotoxicity and In Vivo Biocompatibility Assessments

   Cytotoxicity assays were conducted using NIH‐3T3 and HUVEC cells. The
   viability of cells exposed to various treatments for 24 h was evaluated
   with the Calcein‐AM/PI cell viability/cytotoxicity assay kit.
   Additionally, the proliferation toxicity of PDA@ZIF‐8@AB NPs,
   PCL@QX‐314 MSs, and HA at different concentrations was assessed using
   the CCK‐8 assay. For in vivo biocompatibility evaluation, blood samples
   were collected from the orbital sinus of rats for hemolysis testing,
   and serum was obtained via centrifugation to analyze liver and kidney
   function. MN patches were inserted into the dorsal skin of rats, with
   careful monitoring for any signs of inflammation, redness, or swelling
   at the insertion sites. After a 7‐day period, skin and internal organs
   (heart, liver, spleen, lungs, and kidneys) were harvested and subjected
   to H&E staining to investigate potential inflammatory responses or
   tissue damage.

Statistical Analysis

   Statistical analysis was performed with data presented as mean ±
   standard error of the mean (SEM). All evaluations were conducted using
   GraphPad Prism 9 (GraphPad Software Inc., USA). Differences between two
   groups were analyzed using an unpaired two‐tailed Student's t‐test. For
   comparisons involving multiple groups, one‐way analysis of variance
   (ANOVA) followed by Tukey's post hoc test was employed. Equality of
   variances was assessed using Brown‐Forsythe test, and Welch's
   correction was applied if the assumption of equal variances was
   violated. A p value of less than 0.05 was considered statistically
   significant, with significance levels indicated as follows: ^* p <
   0.05, ^** p < 0.01, ^*** p < 0.001, and ^**** p < 0.0001.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   A.Zhang and X.Jiang. contributed equally to this work and share first
   authorship. W. Li and M. Peng designed and supervised the project. A.
   Zhang and X. Jiang designed the experimental strategies. A. Zhang, X.
   Jiang, B. Xiong, J. Chen, X. Liu, S. Wang and B. Li performed the
   experiments and analyzed the data. A. Zhang, W. Li, and M. Peng wrote
   the manuscript.

Supporting information

   Supporting Information
   [238]ADVS-12-e03698-s004.docx^ (25.2MB, docx)

   Supplemental Movie 1
   [239]Download video file^ (6.3MB, mp4)

   Supplemental Movie 2
   [240]Download video file^ (5.8MB, mp4)

   Supplemental Movie 3
   [241]Download video file^ (6.4MB, mp4)

   Supplemental Movie 4
   [242]Download video file^ (2MB, mp4)

   Supplemental Movie 5
   [243]Download video file^ (2MB, mp4)

   Supplemental Movie 6
   [244]Download video file^ (1.9MB, mp4)

   Supporting Tables
   [245]ADVS-12-e03698-s001.docx^ (17.3KB, docx)

Acknowledgements