Abstract

   Acne is an inflammatory skin disease mainly caused by Propionibacterium
   acnes, which can cause local inflammatory reactions and develop into
   chronic inflammatory diseases in severe cases. To avoid the use of
   antibiotics and to effectively treat the site of acne, we report a
   sodium hyaluronate microneedle patch that mediates the transdermal
   delivery of ultrasound-responsive nanoparticles for the effective
   treatment of acne. The patch contains nanoparticles formed by zinc
   porphyrin–based metal-organic framework and zinc oxide (ZnTCPP@ZnO). We
   demonstrated activated oxygen-mediated killing of P. acnes with an
   antibacterial efficiency of 99.73% under 15 min of ultrasound
   irradiation, resulting in a decrease in levels of acne-related factors,
   including tumor necrosis factor–α, interleukins, and matrix
   metalloproteinases. The zinc ions up-regulated DNA replication–related
   genes, promoting the proliferation of fibroblasts and, consequently,
   skin repair. This research leads to a highly effective strategy for
   acne treatment through the interface engineering of ultrasound
   response.
     __________________________________________________________________

   Ultrasound-responsive microneedle patches offer a highly effective
   nondrug therapy for acne treatment.

INTRODUCTION

   Acne is a skin disease that ranks as the most common skin disorder
   worldwide. Up to 80% of teenagers (and even adults) are affected by
   acne, which can leave permanent scarring and even affect teenager’s
   psychological development ([44]1–[45]3). Anaerobic bacteria
   Propionibacterium acnes infection is the main cause of acne. Excessive
   lipid secretion of the skin causes blockage of pores, which leads to a
   hypoxic microenvironment of the skin, resulting in the proliferation of
   P. acnes in the skin microbiome and damaging normal tissues ([46]4). P.
   acnes also stimulates skin cells to secrete proinflammatory factors,
   such as tumor necrosis factor–α (TNF-α), interleukins (ILs), and matrix
   metalloproteinases (MMPs) ([47]2, [48]4).

   Now, antibiotics are primarily used to fight skin bacterial infections.
   The most common way to treat acne is by taking antibiotics orally or
   using topical antibiotic creams ([49]5–[50]7). Because of the barrier
   effect of the skin, most drugs cannot be delivered through the skin to
   achieve a therapeutic effect ([51]8–[52]10). Furthermore, because acne
   occurs so frequently, the regular use of oral antibiotics may damage
   normal gut microbes and cause other side effects ([53]11). Notably, P.
   acnes can secrete extracellular polysaccharides to form biofilms,
   making the bacteria resistant to general antibacterial agents and the
   inflammatory cells of the host ([54]3). Therefore, the in situ use of
   effective antibacterial agents is the most effective way to treat acne.

   In recent years, numerous researchers have studied microneedles (MNs)
   for drug loading ([55]12–[56]14). Because of their special morphology,
   MNs can easily penetrate the stratum corneum and enter the viable
   epidermis ([57]15). Therefore, the principal benefit of using MNs is
   the promise of pain-free delivery and the delivery of active
   ingredients into the dermis to treat acne ([58]16).

   In view of the characteristics of P. acnes and the prevention of
   drug-resistant bacteria caused by drug abuse, we used MNs loaded with
   ultrasound response nanocomposite materials to achieve the treatment
   effect of acne infection ([59]17). Ultrasound treatment can effectively
   penetrate the human skin and connective tissue to more than 5 cm under
   the skin ([60]18). Ultrasound is also a safe inspection imaging
   technology that has been widely used in the biomedical field.
   Ultrasound responsive materials can produce reactive oxygen species
   (ROS) under ultrasound excitation, which can kill bacteria ([61]19,
   [62]20). ROS can oxidize key cellular macromolecules through an energy
   transfer mechanism, thereby leading to bacterial killing, regardless of
   whether they are Gram-positive or Gram-negative bacteria
   ([63]21–[64]23). Metal-organic frameworks (MOFs) have been widely used
   in the field of biomedicine because of their special structure and
   photoresponsiveness ([65]21). However, a single MOF has only a weak
   ultrasound response because of the rapid recombination of electron
   holes between its structures. The ultrasound response of MOFs can be
   effectively improved by introducing another semiconductor structure to
   form an interface effect between the two structures and to guide the
   transfer of electrons and the change of energy ([66]24).

   Here, the composite structure of zinc porphyrin–based MOF and zinc
   oxide (ZnTCPP@ZnO) was synthesized and loaded into sodium
   hyaluronate–synthesized MNs. After being pressed on the skin, the
   sodium hyaluronate MNs can dissolve into dermis and release the
   nanoparticles. The interface effect between the composite materials can
   greatly improve the sonocatalytic performance and effectively reduce
   the energy required for oxygen activation. Under ultrasound, oxygen
   changes from a ground state to an excited state by gaining more
   electrons through high interfacial charge transfer, thus rapidly
   producing a large number of ROS that kill P. acnes and hence treat
   acne. Furthermore, zinc ions released from the nanomaterial up-regulate
   metallothioneins (Mt) 1 and Mt2 in skin fibroblasts. Mts can maintain
   zinc homeostasis in cells, and their up-regulation accelerates a
   variety of biochemical and molecular events of skin repair ([67]25).
   Here, we demonstrate that the presence of zinc ions also up-regulates
   DNA replication–related genes replication factor C (Rfc) 3, Rfc4, DNA
   ligase I (Lig1), and mini-chromosome maintenance complex component 6
   (Mcm6) in fibroblasts up-regulated, which further demonstrates the
   function of zinc ions in promoting fibroblast proliferation and
   supporting acne recovery ([68]Fig. 1) ([69]26–[70]28).

Fig. 1. Schematic diagram of MN therapy.

   [71]Fig. 1.
   [72]Open in a new tab

   Sonocatalytic mechanism and the treatment of acne through efficient
   sonodynamic ion therapy–based MN patch. US, ultrasound.

RESULTS

Fabrication and characterization of the MN patch

   Sodium hyaluronate was selected as the sole raw material for MN patch.
   Hyaluronate is a polysaccharide that naturally occurs in corneal skin
   and can absorb up to 1000 times its weight of water, retaining skin
   moisture for transdermal absorption. It has excellent biocompatibility
   and solubility and can be used with other materials to provide the
   ideal effect of promoting absorption ([73]29). Here, a simple method
   was used to prepare MNs. In fig. S1A, sodium hyaluronate was dissolved
   in water, and the dissolved gel-like substance was applied all over the
   surface of the polydimethylsiloxane molds. The molds were vacuum-dried
   in an oven for 5 min. After the molded material was removed, the mold
   surface was coated again with sodium hyaluronate aqueous solution. The
   setup was dried at room temperature overnight, and the blank MN patch
   was removed from the mold. To load the MNs with ZnTCPP@ZnO, the
   prepared nanoparticles were mixed into the dissolved sodium
   hyaluronate, in the first step of patch fabrication, before repeating
   the subsequent steps described above. As shown in fig. S1 (B and C),
   the size of the blank MN patch was 10 mm by 10 mm, the number of MNs on
   the patch was 16 × 16 = 256, and the MNs were evenly distributed on the
   patch. On the basis of the fluorescence microscopy results in
   [74]Fig. 2A, the blank MNs fluoresced green under 490-nm blue light
   irradiation, with the height of individual MNs reported as 600 μm under
   high-magnification field of view. This height is high enough to
   penetrate the cuticle and epidermis (~100 μm) of the skin to reach the
   dermis (~600 μm) ([75]30, [76]31). Regarding the ZnTCPP@ZnO MNs, as
   shown in fig. S1C, loading the ZnTCPP@ZnO nanoparticles caused a color
   darkening, and the patch emitted yellow fluorescence under blue light
   ([77]Fig. 2B), given that the ZnTCPP@ZnO in the MNs would typically
   emit red fluorescence under blue light (fig. S1D). The field-emission
   scanning electron microscopy (FESEM) images (fig. S1, B1 and C1) of
   blank MNs and ZnTCPP@ZnO MNs evidenced that the dimension of the
   ZnTCPP@ZnO MNs remained unchanged after modification. The separation
   between the MNs is about 400 μm, and the tip diameter of a single MN is
   ~50 μm only. It seems that the nanoparticles do not alter the MN
   morphology during the synthesis. Although this small size ensures
   sufficient pressure to penetrate the skin, sufficient mechanical
   strength is also important for MN penetration of the skin. In fig. S1E,
   the stress-strain curves of the MNs were obtained by measuring the
   displacement and force changes. The MN patch was pressed vertically at
   a speed of 0.1 mm/min at the test station with an initial force of 0.01
   N. The curves of the two different samples showed a sudden drop in
   force, which was due to the bending of the MN, a phenomenon that
   persisted until the compression was complete ([78]32). The strain
   distance between the beginning of the MN bending and the final complete
   compression is about 600 μm, consistent with the height of the MN. At
   the same time, although the strength of the MN was weakened after the
   addition of the nanoparticles, it was still higher than the minimum
   penetration strength of the skin ([79]33). To verify the penetration
   effect and skin solubility of the MN patch, a penetration experiment
   was conducted using pig skin. In fig. S1F, the MN patch was pressed on
   the pig skin for 2 min and removed. The MNs on the surface of the patch
   were completely dissolved, leaving uniform pinholes on the surface of
   the pig skin. To verify the yield of the MN patch, the number of MNs on
   the patch was counted. The mold is designed with 256 MNs on one single
   patch, and the height of each MN measures at 600 μm. The results
   highlighted that the yield of the MNs and ZnTCPP@ZnO MN patches
   prepared by this method is 96.7% (table S1). In addition, we identified
   that the height of MNs can be maintained at 600 μm, when vacuum
   treatment is applied during fabrication. Together, these results show
   the successful fabrication of the MN patch and its loading with
   ZnTCPP@ZnO nanoparticles.

Fig. 2. Characterization of MN patch and ZnTCPP@ZnO nanoparticles.

   [80]Fig. 2.
   [81]Open in a new tab

   (A) Fluorescence images of blank MNs. (B) Fluorescence images of
   ZnTCPP@ZnO MNs. Scale bars, 100 μm (left) and 200 μm (right). (C) HRTEM
   image of ZnTCPP@ZnO. (C1) Magnified HRTEM image of ZnTCPP in
   ZnTCPP@ZnO. (C2) Magnified HRTEM image of ZnO in ZnTCPP@ZnO. (D) X-ray
   diffraction (XRD) patterns of ZnO, ZnTCPP, and ZnTCPP@ZnO. a.u.,
   arbitrary units. (E) X-ray photoelectron spectroscopy (XPS) survey scan
   of ZnO, ZnTCPP, and ZnTCPP@ZnO. XPS high-resolution spectra of (F) Zn
   2p peaks and (G) O 1s peaks obtained from ZnO, ZnTCPP, and ZnTCPP@ZnO.
   (H) XPS high-resolution spectra of N 1s peaks obtained from ZnTCPP and
   ZnTCPP@ZnO.

Characterization of ZnTCPP@ZnO nanoparticles

   FESEM and high-resolution transmission electron microscopy (HRTEM) were
   used to observe the microstructure of ZnTCPP@ZnO nanoparticles
   ([82]Fig. 2C). The nanoparticles were mainly composed of regular
   four-sided rod-shaped ZnTCPP MOF and ZnO nanoparticles (fig. S2A), and
   the element mapping of HRTEM showed a uniform distribution of elements
   (C, O, N, and Zn) (fig. S2B). The HRTEM image in ([83]Fig. 2C1) shows
   that the lattice fringe spacing of the rod-like structure is 1.63 nm,
   which is consistent with the computer-simulated crystal plane spacing
   of ZnTCPP ([84]34). In the x-ray diffraction (XRD) patterns
   ([85]Fig. 2D), the peak at 7.1° corresponds to the (100) crystal plane
   and appeared in both the black (pure ZnTCPP) and red (ZnTCPP@ZnO
   nanoparticles) curves, which is consistent with the literature
   ([86]35). Furthermore, as shown in [87]Fig. 2C2, the ZnO structure with
   0.26-nm lattice fringes spacing can be seen, and the characteristic
   peaks of ZnO appear in both blue curve (pure ZnO) and red curve
   (ZnTCPP@ZnO nanoparticles) ([88]Fig. 2D), which proves the presence of
   zinc oxide in ZnTCPP@ZnO nanoparticles, especially the peak located at
   34.41°, which corresponds to the crystal plane (002) and consistent
   with ([89]Fig. 2C2) ([90]36, [91]37).

   To further verify the structure of ZnTCPP@ZnO nanoparticles and the
   interaction between ZnO and ZnTCPP, x-ray photoelectron spectroscopy
   (XPS) tests were performed. [92]Figure 2E shows C, O, and Zn in all
   three curves, and the carbon peak in ZnO may be due to the adsorption
   of CO[2] on the surface. The difference was that both the red and black
   curves had N, which came from the porphyrin ring in the ZnTCPP
   structure ([93]38). To further confirmed the chemical bonding and
   interfacial interaction, the elemental peaks were analyzed and
   compared. As shown in [94]Fig. 2F, the subpeaks at 1045.2 and 1022.1 eV
   were assigned to Zn 2p[3/2] and Zn 2p[1/2] in ZnO, and Zn 2p[3/2] and
   Zn 2p[1/2] in ZnTCPP were located at 1044.2 and 1021.1 eV ([95]38,
   [96]39). In the composite structure of ZnTCPP@ZnO nanoparticles, the
   two characteristic peaks of Zn were Zn 2p[3/2] (1044.9 eV) and Zn
   2p[1/2] (1021.8 eV), which were exactly located between pure ZnO and
   ZnTCPP. These changes in binding energy indicated that the zinc atom
   accepted the electron density of the porphyrins in the nanoparticles.
   Electrons of nitrogen in the porphyrin ring were transferred to an
   unoccupied metal orbital, forming a porphyrin-metal bond ([97]40). When
   the density of the electron cloud increased, the shielding effect of
   the metal atom on the internal electrons also increased. Therefore, the
   binding energy of the inner shell of the electron was reduced ([98]40,
   [99]41). In [100]Fig. 2G, the O 1s region of the ZnO contained Zn─O
   (531.0 eV) and C─O (532.5 eV); the C─O group may also be caused by the
   adsorption of CO[2] on the surface. The O 1s region of the ZnTCPP
   contained C─O (532.6 eV) and C═O (531.5 eV), whereas the ZnTCPP@ZnO
   included C─O (531.9 eV), C═O (531.4 eV), and Zn═O (530.7 eV). Compared
   with ZnO, the Zn─O group in ZnTCPP@ZnO had a lower binding energy. The
   C─O group in ZnTCPP shifted to lower binding energy from 532.6 to 531.9
   eV in ZnTCPP@ZnO ([101]23). In [102]Fig. 2H, the N 1s signals were
   located at the C─N (399.0 eV), C═N (397.5 eV), and pyrrolic N (401.1
   eV) in ZnTCPP, whereas the pyrrolic N group in ZnTCPP@ZnO shifted to
   lower binding energy of 398.1 eV ([103]24, [104]42). The peak shifts of
   Zn 2p, O 1s, and N 1s indicated the formation of a tight interface and
   electron interactions between ZnO and ZnTCPP ([105]24, [106]43). In
   summary, these findings clarify the details of the structure and
   chemical bonds of ZnTCPP@ZnO nanoparticles.

Sonocatalytic activity of ZnTCPP@ZnO and theoretical calculations

   Similar to the principle of photocatalysis, the efficiency of
   sonocatalysis is determined by the efficiency of electrons and hole
   separation and recombination under ultrasound irradiation ([107]18,
   [108]44). Here, photoluminescence (PL) spectra were used to investigate
   the recombination efficiency of electron-hole pairs. In the PL results,
   the PL signal of ZnO was much higher than that of ZnTCPP and
   ZnTCPP@ZnO, which proves that the electron-hole pairs of ZnO recombine
   quickly (fig. S2C). The PL signal of ZnTCPP@ZnO was lower than that of
   ZnTCPP, indicating that the recombination of electron-hole pairs in the
   ZnTCPP@ZnO composite was inhibited (fig. S2D) ([109]45, [110]46).
   Notably, in the ZnTCPP@ZnO curve, the peak at about 450 nm had an
   obvious red shift, indicating the electron state hybridization between
   ZnTCPP and ZnO ([111]20, [112]47). The heterointerface between ZnTCPP
   and ZnO led to a longer lifetime of charge carriers crossing the
   interface, which reduced the recombination efficiency of electron-hole
   pairs and thus improved the sonocatalytic efficiency.

   To further investigate the electron generation of materials under
   ultrasound irradiation, a sonoelectrochemical method was used to detect
   the generation of a sonocurrent. As shown in [113]Fig. 3A, the
   ultrasound response current of ZnTCPP@ZnO was the most obvious in the
   continuous eight-cycle repeat of the on/off ultrasound. This result was
   consistent with the PL result, suggesting that inhibiting the
   recombination of electron-hole pairs can promote sonocurrent
   generation. Furthermore, the electron spin resonance (ESR) was used to
   detect ROS production in the materials during ultrasound irradiation.
   In [114]Fig. 3B, the ^1O[2] was tested by
   2,2,6,6-tetramethylpiperidine, while the
   5,5-dimethyl-1-pyrroline-N-oxide was used as the •OH probe ([115]22,
   [116]48). The characteristic signal (1:1:1) of ^1O[2] showed that,
   after 5 min of ultrasound irradiation, the signal intensity of
   ZnTCPP@ZnO was higher than that of ZnO and ZnTCPP ([117]43, [118]48).
   The difference was that the characteristic signal (1:2:2:1) of •OH in
   ZnTCPP@ZnO was almost identical with that of ZnO and ZnTCPP ([119]49).
   The ^1O[2] detection of ZnTCPP@ZnO was tested by
   1,3-diphenylisobenzofuran (DPBF); the yield of ^1O[2] increased with
   the extension of ultrasound time ([120]Fig. 3C). However, pure DPBF did
   not change with an increase in ultrasound time (fig. S3A).

Fig. 3. Sonocatalytic activity and theoretical calculations of ZnTCPP@ZnO.

   [121]Fig. 3.
   [122]Open in a new tab

   (A) Ultrasound current responses of ZnO, ZnTCPP, and ZnTCPP@ZnO. (B)
   ESR measurements of ^1O[2] and •OH of ZnO, ZnTCPP, and ZnTCPP@ZnO after
   ultrasound irradiation for 5 min. (C) ^1O[2] detection of ZnTCPP@ZnO
   through DPBF. (D) ZnTCPP model structure from the top view and front
   view. (E) Charge density difference of ZnTCPP adsorbed O[2] and Bader
   charge. (F) O[2] activation energy of ZnTCPP. (G) Varied density of
   states of ZnTCPP. (H) ZnTCPP@ZnO model structure from the top view and
   front view. (I) Charge density difference of ZnTCPP@ZnO adsorbed O[2]
   and Bader charge. (J) O[2] activation energy of ZnTCPP@ZnO. (K) Varied
   density of states of ZnTCPP@ZnO.

   To explore the mechanism of sonocatalytic, the Vienna Ab initio
   Simulation Package (VASP) was used to perform all the density
   functional theory (DFT) calculations ([123]50, [124]51). Because the
   main species of ROS was ^1O[2], the main purpose of calculation was to
   explore the oxygen activation and electron transfer of ZnTCPP@ZnO. The
   conversion of O[2] to ^1O[2] was a process of energy transformation and
   electron transfer. The ground state of O[2] (^3O[2]) had two unpaired
   electrons at the individual antibonding π orbital (π[x]* and π[y]*),
   while the excited state of O[2] (^1O[2]) had paired electrons at the
   antibonding π orbital (π[x]*) ([125]52, [126]53). Here, the electron
   transfer and the energy evolution produced by the material with
   ultrasound irradiation could help ^3O[2] accomplished this
   transformation. Therefore, the site of oxygen adsorption on ZnTCPP@ZnO
   and the electrons transfer between two interfaces in ZnTCPP@ZnO were
   investigated by DFT calculations.

   Before this, the structure of a single element of the ZnTCPP and
   ZnTCPP@ZnO was simulated in [127]Fig. 3 (D and H). After oxygen
   adsorption, the results of differential charge density and Bader charge
   showed that there was an obvious charge transfer between the oxygen
   adsorption site and O[2] ([128]Fig. 3, E and I). Specifically, in the
   ZnTCPP model, O[2] can obtain electrons from ZnTCPP by forming Zn─O
   bonds, and the number of transferred electrons was calculated as 0.130
   electron charge. By contrast, in the ZnTCPP@ZnO model, because of the
   interface effect between ZnO and ZnTCPP, electron transfer near O[2]
   increased, resulting in an increased number of electrons oxygen
   obtained; the calculated transferred electron charge was 0.288 electron
   charge. In addition to the enhanced amount of electron transfer in
   ZnTCPP@ZnO model, the calculation also showed that the oxygen
   activation process in ZnTCPP@ZnO model required less energy. As shown
   in [129]Fig. 3F, according to the calculation of the ZnTCPP model, the
   activation energy required for O[2] to change from a ground state
   (^3O[2]) to an excited state (^1O[2]) was 6.16 eV. In the ZnTCPP@ZnO
   model, the activation energy of O[2] was only 3.49 ([130]Fig. 3J), much
   lower than that in the ZnTCPP model, proving that O[2] was more readily
   activated in the ZnTCPP@ZnO model ([131]49).

   Furthermore, the density of electronic states (DOS) calculation was
   conducted using the above two models. In [132]Fig. 3 (G and K), the DOS
   of the ZnTCPP@ZnO indicated more hybridization at the Fermi level than
   ZnTCPP, which was also attributed to the interface effect between ZnO
   and ZnTCPP in the ZnTCPP@ZnO model ([133]22). In summary, through the
   simulation calculation of electron transfer and energy evolution, the
   sonocatalytic mechanism of ZnTCPP@ZnO was elaborated in detail.
   Specifically, O[2] was adsorbed at the interface of ZnTCPP@ZnO due to
   the interface effect between ZnO and ZnTCPP in ZnTCPP@ZnO, and O[2]
   gains more electrons and energy. The energy required for O[2]
   activation in this structure was greatly reduced, which facilitated the
   transition of O[2] from the ground state to the excited state. These
   results indicated that the ZnTCPP@ZnO composite nanoparticles had
   superior sonocatalytic effects and can only produce one type of ROS
   (^1O[2]) to provide support for antibacterial activities.

Antibacterial activity of the MN patch

   P. acnes is the main bacterium that causes acne. To investigate the
   antibacterial effect of MNs against P. acnes, a spread plate method was
   used, as shown in fig. S4A. P. acnes with a concentration of 10^5
   colony-forming units (CFUs)/ml was chosen as the initial bacterial
   density. We observed only a slight change in the number of bacteria in
   the control group [phosphate-buffered saline (PBS)] with or without
   ultrasound. Similarly, for the blank MNs, ultrasound did not affect the
   growth of bacteria, and their numbers did not change too much. By
   contrast, after 15 min of ultrasound irradiation, the number of
   bacteria in ZnTCPP@ZnO MN group was reduced because of the substantial
   number of ROS produced by the ZnTCPP@ZnO nanoparticles in the MNs under
   ultrasound irradiation. Compared with the control group and blank MN
   group, the number of bacteria in ZnTCPP@ZnO MN group also decreased
   slightly without ultrasound irradiation, which may be due to the
   certain inhibitory effect of the zinc ions released from ZnTCPP@ZnO
   nanoparticles in MNs on bacteria (fig. S3B) ([134]44, [135]54,
   [136]55).

   To reflect the antibacterial effect more accurately, we counted the
   average CFU/ml of bacteria in each group. In the ultrasound group, the
   average CFU/ml of P. acnes in the control, blank MN, and ZnTCPP@ZnO MN
   group were 1.5 × 10^8, 1.2 × 10^8, and 8.8 × 10^7, respectively. In the
   ultrasound^+ group, the average CFUs/ml of P. acnes in the control,
   blank MN, and ZnTCPP@ZnO MN group were 1.5 × 10^8, 1.1 × 10^8, and
   4.7 × 10^5 ([137]Fig. 4A). According to the following formula:
   antibacterial efficiency (%) = (number of CFUs in the control group −
   number of CFUs in the experimental group)/(number of CFUs in the
   control group) × 100%, the antibacterial rate of ZnTCPP@ZnO MN group
   was calculated to be high as 99.73%.

Fig. 4. Antibacterial activity and cytocompatibility of the MNs.

   [138]Fig. 4.
   [139]Open in a new tab

   (A) Number of P. acne colonies after treatments with different samples.
   (B) SEM images of P. acne colonies after treatments with different
   samples. Scale bar, 1 μm. (C) Protein leakage concentration for P.
   acne. (D) Fluorescent images of NIH3T3 cultured with various samples
   for 1 day. Actin is stained with fluorescein isothiocyanate (green),
   and nuclei is stained with 4′,6-diamidino-2-phenylindole (blue) at a
   cell density of 4 × 10^5 cells/ml. Scale bar, 100 μm. (E) MTT assay of
   cell viability. Cells were cultured on different samples for 1, 3, and
   7 days. (F) Scratch assay images of NIH3T3 cultured with various
   samples for 1 day. (G) Cell migration quantification of NIH3T3. n = 3
   independent experiments per group. Statistical differences were
   analyzed by one-way analysis of variance (ANOVA) and post hoc Tukey’s
   test (**P < 0.01, ***P < 0.001, and ****P < 0.001).

   The P. acnes morphology of different samples was observed by FESEM. As
   shown in [140]Fig. 4B, except for the morphological folds (marked by
   red arrows) observed in the ZnTCPP@ZnO MNs under ultrasound
   irradiation, the bacteria in the other groups had complete cell
   morphology and regular rod-shaped shapes. In [141]Fig. 4C, after
   ultrasound, the amount of protein leakage of P. acnes increased, and
   the amount of protein leakage in the ZnTCPP@ZnO MN group was the
   largest. This result was consistent with the previous spread plate
   results. The main reason for this difference was that, under the
   irradiation of ultrasound, ZnTCPP@ZnO produced a large number of ROS
   that destroyed the normal structure of the bacteria, leading to the
   contraction or even rupture of the bacterial cell membrane ([142]44).

   To further verify the antibacterial effect of MNs against
   drug-resistant bacteria, methicillin-resistant Staphylococcus aureus
   (MRSA) was selected to test the antibacterial effect of MNs. The
   results showed that the antibacterial trend of different groups against
   MRSA was consistent with the findings shown in [143]Fig. 4A, and the
   ZnTCPP@ZnO MN group had the best antibacterial effect (fig. S4B). On
   the basis of the calculation of the CFU, we can conclude that the
   antibacterial efficiency of ZnTCPP@ZnO MN group reached 99.64% (fig.
   S5A). Similarly, the cell membrane of MRSA was severely damaged by the
   ROS produced in the ZnTCPP@ZnO MN group under ultrasound irradiation
   (fig. S5, B and C), indicating that ROS produced by ZnTCPP@ZnO also had
   an excellent antibacterial effect on drug-resistant bacteria.

In vitro cytotoxicity studies and related mechanisms of cell regulation

   The cell (NIH3T3, mouse embryonic fibroblast cell line) viability in
   different samples was evaluated using a
   3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT)
   assay and cell staining. In [144]Fig. 4D, after 1 day of culture, in
   the control group and blank MN group, the large number of cells and the
   spread of the cellular matrix (green) around the nucleus (blue)
   confirmed that the cells were well spread and were highly active after
   1 day. As for ZnTCPP@ZnO MN group, because ZnTCPP@ZnO had red
   fluorescence itself (fig. S1D), the red color showed up, but we could
   still observe many cells coexisting with the red powder coexistence,
   which indicates that dissolution of the MN into powder posed no
   cytotoxicity. We further acquired more quantitative results using the
   MTT results ([145]Fig. 4E). After the first day, the MN group showed
   some effect on cell growth, but the number of cells increased gradually
   by the third and seventh days, indicating that the MNs themselves did
   not have an obvious toxic effect on cells. Especially in group
   ZnTCPP@ZnO MNs, the number of cells was higher from days 3 to 7 than in
   the other two groups. To verify the cellular toxicity of the MNs upon
   ultrasound stimulation, we supplemented an MTT assay under ultrasound
   irradiation for 15 min. The results shown in fig. S6 demonstrated that
   the transient ultrasound stimulation does not alter the cell viability
   as compared with the control, implying that the ROS induced by
   ZnTCPP@ZnO microstructures will unlikely cause the issue of
   cytotoxicity. Thus, ZnTCPP@ZnO MN group had a certain promoting effect
   on cell proliferation.

   To further confirm the skin repair effect of the ZnTCPP@ZnO MNs on skin
   infection, a cell migration assay was also performed. In [146]Fig. 4F,
   the amount of cell migration was higher in ZnTCPP@ZnO MN group, twice
   as much as in the other two groups ([147]Fig. 4E). This phenomenon may
   be due to the regulation of cells by zinc ions released from the
   ZnTCPP@ZnO MN group. To ascertain that ZnTCPP@ZnO could promote cell
   proliferation, we conducted transcriptome sequencing on cells cotreated
   with different materials. As shown in [148]Fig. 5A, based on the
   expression of genes in different groups, the correlation behavior
   analysis between the groups tested whether the variation between
   biological repeats met the expectations of the experimental design. The
   left and top sides represent the clustering situation of each group.
   Squares of different colors represent the correlation levels of the two
   groups. The larger the value, the greater the correlation between the
   two groups, and the more similar it was. The correlation between the
   control group and the blank MN group was good. However, the gene
   expression of the ZnTCPP@ZnO MN group was different from that of the
   other two groups. As shown in [149]Fig. 5B, compared with the control
   group, the ZnTCPP@ZnO MN group had 4948 differentially expressed genes
   (2920 down-regulated and 2028 up-regulated). In [150]Fig. 5C, compared
   with the blank MN group, the ZnTCPP@ZnO MN group still had 3994
   differentially expressed genes (2547 down-regulated and 1447
   up-regulated). In [151]Fig. 5D, between the control group and the blank
   MN group, there were only 627 differentially expressed genes (476
   down-regulated and 151 up-regulated). These results suggest that the
   ZnTCPP@ZnO component in ZnTCPP@ZnO MN group was the main factor driving
   gene differential expression.

Fig. 5. Analysis of gene expression correlation and differential gene
expression between different samples.

   [152]Fig. 5.
   [153]Open in a new tab

   (A) Correlation heatmap analysis between different groups. Each group
   has three parallel samples. (B to D) Volcano plots of differentially
   expressed genes between different groups. The red dots represent genes
   that are up-regulated, the blue dots represent genes that are
   down-regulated, and the gray dots represent genes that are not
   different. (E) Clustering heatmaps of the top 25 up-regulated and 25
   down-regulated differential genes. Red represents high expression of
   the gene in the sample, and blue represents low expression.

   To further analyze the function of the differential genes, we extracted
   the top 50 differentially expressed genes (25 down-regulated and 25
   up-regulated from [154]Fig. 5, B and C). The heatmap in [155]Fig. 4E
   showed that these 50 target genes were consistently up-regulated or
   down-regulated in all the samples from the ZnTCPP@ZnO MN group compared
   to the other two groups.

   Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG)
   enrichment analyses were performed to study the function of these 50
   target genes. In [156]Fig. 6 (A and B), the horizontal axis represents
   the rich factor—the larger the rich factor, the greater the degree of
   enrichment. The size of the bubble represents the number of genes, and
   the color of the bubble corresponds to different P[adjust] (adjusted P
   value) ranges. In the GO enrichment analysis, the target genes were
   mainly enriched in DNA replication, cellular zinc ion homeostasis, and
   pyrimidine deoxyribonucleotide binding. In KEGG enrichment analysis,
   target genes were mainly enriched in DNA replication, mismatch repair,
   mineral absorption, and the pyrimidine metabolism pathway.

Fig. 6. Functional analysis of target genes and qPCR in different samples.

   [157]Fig. 6.
   [158]Open in a new tab

   (A) GO enrichment analysis of biological processes of the top 25
   up-regulated and 25 down-regulated differential genes. (B) KEGG pathway
   enrichment analysis of the top 25 up-regulated and 25 down-regulated
   differential genes. SNARE, soluble N-ethylmaleimide–sensitive factor
   attachment protein receptor. (C) GO term enrichment of chord diagram.
   (D) KEGG pathway enrichment of chord diagram. (E) Protein-protein
   interaction (PPI) network analysis on target genes. The more connected
   lines, the higher the interaction correlation. (F) Expression of target
   genes in different treated groups. n = 3 independent experiments per
   group. Statistical differences were analyzed by one-way ANOVA and post
   hoc Tukey’s test (*P < 0.05, **P < 0.01, and ***P < 0.001).

   To identify the genes that regulate these functions more intuitively,
   an enrichment chord diagram was constructed for analysis, as shown in
   [159]Fig. 5 (C and D). The main functions of the target genes in the
   two differential enrichment analyses focused on DNA replication and
   metal ion homeostasis. The target genes were Rfc4, Rfc3, Lig1, Mcm6,
   Mt1, Mt2, Dut, and Dctcpp1. The genes involved in DNA replication—Rfc4,
   Rfc3, Lig1, and Mcm6—were up-regulated in the ZnTCPP@ZnO MN group
   ([160]26–[161]28). Dut and Dctcpp1 genes, which were highly expressed
   in the proliferation cells, were also up-regulated in the ZnTCPP@ZnO MN
   group ([162]56, [163]57). These were consistent with MTT data, proving
   that cell proliferation in ZnTCPP@ZnO MN group was faster than that in
   the other two groups. To explore the mutual use of proteins
   corresponding to these genes, these genes were submitted to the STRING
   database to construct a protein-protein interaction (PPI) network, as
   shown in [164]Fig. 6E. The network revealed that the proteins of
   differential genes associated with DNA replication and cell
   proliferation were highly correlated, especially Rfc4. Furthermore, the
   expression levels of Mt1 and Mt2 genes, which are involved in the
   process of intracellular metal ion adsorption and mitosis promotion in
   cell behavior, were highly related to their PPI-suggested involvement
   with zinc ions ([165]25, [166]44). To verify the expression of target
   genes in different groups, fibroblasts cocultured with different groups
   were subjected to RNA extraction and quantitative real-time polymerase
   chain reaction (qRT-PCR) testing, as shown in [167]Fig. 6F. All eight
   genes were more highly expressed in the MN group with ZnTCPP@ZnO
   nanoparticles.

   We then analyzed the secretion of fibroblast-related cytokines by
   enzyme-linked immunosorbent assay (ELISA). The results are shown in
   fig. S7, which is in good agreement with the above results. Stimulated
   by the material, the cells secreted more cytokines related to cell
   proliferation in ZnTCPP@ZnO MNs group. Combined with the previous
   results, these findings indicate that zinc ions released by ZnTCPP@ZnO
   promoted the proliferation of fibroblasts, which would support skin
   repair.

In vivo anti-acne performance of the MN patch

   Before we proceeded to the animal study, a hemolysis test on MNs had
   been conducted. The results suggested that both of the blank MNs and
   ZnTCPP@ZnO MNs do not induce hemolysis (fig. S8). To evaluate the
   effect of a MN patch on P. acnes infection, we created an acne model by
   intradermal injection of P. acnes into the back of Balb/c mice. All
   acne model mice were then randomly divided into six groups (n = 5). At
   day 0, different groups of mice were treated with the different
   materials, and acne was observed and analyzed at 1, 3, and 7 days
   ([168]Fig. 7A). Group 1 (G1) was the control group, with no treatment.
   G2 was treated with an acne drug, Mentholatum Acnes Point Clear, three
   times on day 0. G3 was treated with blank MNs by pressing the blank MN
   patch on the acne for 3 min, after which the needles dissolved in the
   dermis, and the patch was removed before a 10-min ultrasound
   irradiation. G4 was treated with ZnTCPP@ZnO patch (without needles on
   the patch) by pressing the patch on the acne with 10-min ultrasound
   irradiation. Note that the purpose of setting group G4 was to verify
   whether the patch alone can release material into the acne lesions
   under the ultrasound without needle. G5 was treated with ZnTCPP@ZnO MNs
   by pressing the ZnTCPP@ZnO MN patch on the acne for 3 min, after the
   needles were dissolved in the dermis, the patch was removed and
   followed by 10-min ultrasound irradiation. G6 received the same
   treatment as G5 but without ultrasound. Microscopic pores were observed
   on the skin surface of mouse model with acne infection after MN patch
   was applied ([169]Fig. 7A). When examined under SEM, almost all the MNs
   disappeared after the application (fig. S9, A and B). In addition, the
   result of hematoxylin and eosin (H&E) staining suggested that the MNs
   were completely dissolved in the epidermal layer of mouse model after 1
   day of application (fig. S9C).

Fig. 7. Mitigation of acne via different MN patch.

   [170]Fig. 7.
   [171]Open in a new tab

   (A) Schematic illustration of the MN patch for the treatment of acne.
   (B) Photographs of mice treated with various treatments. Scale bar, 5
   mm. (C) Quantitative analysis of thickness of skin from mice in each
   group. (D) Two acne models on the back of the same mouse. (The acne on
   the left was the same treatment as that in G5, while the acne on the
   right was untreated.) (E) Hematoxylin and eosin (H&E) staining of the
   skin tissue sections from the mice after the different treatments
   (magnification, ×5 and ×40; scale bars, 200 and 20 μm, respectively).
   (F) Quantitative analysis of infiltrated inflammation cells in each
   group. n = 3 independent experiments per group. Statistical differences
   were analyzed by one-way ANOVA and post hoc Tukey’s test (*P < 0.05,
   *P < 0.01, ***P < 0.001, and ****P < 0.0001).

   As shown in [172]Fig. 7B, at day 0 (2 days after intradermal injection
   of P. acnes), acne approximately 1 mm thick formed at the injection
   site due to bacterial infection and proliferation, demonstrating the
   successful construction of the P. acnes infection acne model ([173]58).
   After the MNs were pressed on the surface of acne, pinholes with
   uniform distribution and size could be seen on the surface, which
   proved that the MNs could effectively penetrate the surface of acne
   ([174]Fig. 7A). In G1, the size of the untreated acne increased over
   time due to infection. The same trend also appeared in G4, indicating
   that, in the absence of needles on the patch, a single patch could not
   effectively deliver nanoparticles to the infected site. Despite
   ultrasound irradiation, the infected site was still not effectively
   treated due to dermis blocking. To ensure consistency in G2, which used
   the acne drug, the drug was applied three times on day 0. We observed a
   certain degree of inhibition on day 1 with drug use, but, on days 3 and
   7, the acne returned. This may be due to the fact that it was difficult
   to penetrate the dermis to kill bacteria effectively with a single-day
   drug treatment.

   As for G3 blank MNs, because they were not loaded ZnTCPP@ZnO
   nanoparticles, although the needles dissolved in the dermis, they did
   not have sufficient antibacterial effect under ultrasound irradiation;
   only some physical piercing effect was observed on the acne. A similar
   piercing effect also appeared in G6. Although G6 was treated with
   loaded ZnTCPP@ZnO nanoparticles, ultrasound was not used, resulting in
   no ROS generation, and the inhibiting effect on acne only came from
   zinc ions released by the nanoparticles (fig. S3B). Therefore, the
   effect of G6 was slightly better than that of G3 ([175]Fig. 7C). By
   contrast, the treatment effect of G5 was better than that of the other
   groups, and acne was almost completely treated by day 7 after
   treatment. This was because the MNs help the ZnTCPP@ZnO nanoparticles
   enter the infected dermis area, and, when stimulated by ultrasound, the
   nanoparticles in the dermis release a large number of ROS, killing P.
   acnes, and the zinc ions helped the skin repair, allowing acne to be
   effectively treated.

   To further demonstrate the therapeutic effect of G5, two acne models
   were created on the back of the same mouse. The acne on the left was
   the same treatment as that in G5, while the acne on the right was
   untreated. The size of the treated acne after 7 days was smaller than
   that of the untreated acne on the right ([176]Fig. 7D). To verify the
   bacterial inhibition effect, skin tissue from two acne sites was taken
   out and ground up, and the number of bacteria inside the acne was
   observed using the spread plate method ([177]Fig. 7D). The number of
   bacteria in treated acne was much lower than in untreated acne. To
   quantify the treatment effect of acne, the thickness of acne was
   measured during the treatment ([178]Fig. 7C), revealing a change in
   acne thickness that was consistent with the treatment effects observed
   in [179]Fig. 7B.

   To further evaluate the effect of MNs on acne, mice were euthanized on
   day 7, and skin tissue from the acne site was excised for histological
   and immunohistochemical analysis. H&E staining (5×; [180]Fig. 7E)
   revealed that the thickness of the skin section in G5 was thinner than
   in the other groups, which was consistent with the result presented in
   [181]Fig. 7C. As shown in the ×40 H&E staining images, the skin tissue
   in G1 and G4 had an abundance of infiltrated inflammatory cells (black
   arrows). Red blood cells (RBCs) that flowed out from ruptured vessels
   were observed at some sites (red arrows). Similarly, although the
   number of inflammatory cells in G2, G3, and G6 was small, a certain
   degree of infiltrated inflammatory cells was observed. By contrast, a
   clear and intact reticular layer of dermis and hair follicle structures
   can be observed in G5’s skin tissue due to its excellent anti-infection
   properties. Quantitative analysis also proved that G5 had the lowest
   number of inflammatory cells and the best therapeutic effect
   ([182]Fig. 7F).

   P. acnes is an anaerobic Gram-positive bacterium. Its overproliferation
   encourages the synthesis of lipid droplets and triacylglycerol by mouse
   skin sebum cells, causing sebum accumulation to clog hair follicles,
   further causing hypoxia in the skin and promoting bacterial growth
   ([183]2). P. acnes induces and exacerbates inflammation, inducing
   keratinocytes to produce IL-8, TNF-α, and MMP-2 ([184]59).
   Immunohistochemical results showed the expression of these three
   proteins in the acne area in detail. In [185]Fig. 8, except for G5,
   cells with intense staining for the IL-8 protein can be observed in
   other groups. The results of the optical density analysis intuitively
   showed that the expression of inflammatory factors in G5 was the
   lowest. The same trend can be observed in the results of TNF-α and
   MMP-2. It was proved again that the excellent bacterial effect of G5
   reduced the expression of these inflammatory factors and proteins and
   had the best therapeutic effect.

Fig. 8. Immunohistochemical analysis of acne after different treatments.

   [186]Fig. 8.
   [187]Open in a new tab

   (A) Representative images and optical density analysis of IL-8 protein
   expression in the tissue from acne area. (B) Representative images and
   optical density analysis of TNF-α protein expression in the tissue from
   acne area. (C) Representative images and optical density analysis of
   MMP-2 protein expression in the tissue from acne area. Scale bar (A-C),
   20 μm. IOD, integrated optical density. (D) Routine blood analysis from
   the day 7 mice. The range within the dotted line represents the normal
   norm. n = 3 independent experiments per group. Statistical differences
   were analyzed by one-way ANOVA and post hoc Tukey’s test (*P < 0.05,
   **P < 0.01, ***P < 0.001, and ****P < 0.001). WBC, white blood cell
   count; Mon, monocyte; Gran, granulocyte; PLT, platelet count; Lymph,
   lymphocyte; RBC, red blood cell; HCT, hematocrit; HGB, hemoglobin.

   To further confirm the biosafety of the ZnTCPP@ZnO MNs, the blood from
   the mice on day 7 was extracted for routine blood analysis. As shown in
   [188]Fig. 8D, all data were within the corresponding normal scope
   (marked by dotted line), including white blood cell count, monocyte,
   granulocyte, platelet count, lymphocyte, RBC count, hematocrit, and
   hemoglobin. In addition, the corresponding histological analyses of the
   major organs (heart, liver, spleen, lung, and kidney) based on H&E
   staining (fig. S10) showed no signs of organ damage. Together, these
   results further confirmed that the ZnTCPP@ZnO MNs had excellent
   biosafety and immense potential for clinical treatment for bacterial
   infection acne.

DISCUSSION

   The conventional acne treatments mainly include the epidermal
   administration of antibiotics or the physical puncture to drain out the
   pus from lesion. However, repeated applications of antibiotics may
   reduce the sensitivity of bacteria to drugs ([189]2, [190]6, [191]9).
   In addition, the treatment efficiency can be potentially reduced due to
   the permeability of epidermal barrier. Physical and psychological shock
   to the patients had been documented, when long-term and repeated onset
   of acne infection happened. Hence, oral medication and psychological
   intervention are advised to those severe cases ([192]59). We believe
   that the treatment via ROS-enabled MNs, regarding as a minimal invasive
   and transdermal approach, can effectively relieve the acne infection
   ([193]8). MNs serving as the delivery vehicle of antibiotics or
   nonantibiotic substances to address bacterial infection have been
   studied ([194]4, [195]9, [196]15). In the current design, we introduce
   ultrasound-responsive antibacterial nanomaterials to the MN patch that
   not only responds to bacterial infection quickly and efficiently but
   also are able to regulate the epidermal repair driven by fibroblasts.

   In summary, we have developed a simple and effective approach to treat
   acne infection through the adoption of sonocatalytic antibacterial
   nanomaterials in MNs. The MNs made of biocompatible sodium hyaluronate
   can easily deliver ZnTCPP@ZnO to the acne lesion. The interfacial
   effect between ZnTCPP and ZnO within the composite structure has
   improved the electron transfer efficiency generated by ultrasound,
   thereby increasing the interfacial charge transfer to the adsorbed
   O[2]. In addition, the composite material reduces the activation energy
   of O[2], and hence, a substantial amount of ^1O[2] can be produced
   under ultrasound irradiation that warrants the sonocatalytic
   antibacterial ability of the composite. In addition to the elimination
   of P. acnes, ZnTCPP@ZnO is able to intercept the infection induced by
   MRSA. The up-regulation of IL-8, TNF-α, and MMP-2 proteins expressed by
   the fibroblasts demonstrated that the system can effectively convince
   the skin tissue repair, when zinc ion release concurrently happens. The
   cytomic data suggested that two up-regulated genes, namely, Mt1 and
   Mt2, were highly relevant to the absorption of zinc ions in the cells.
   Furthermore, the following up-regulated genes, i.e., Rfc4, Rfc3, Lig1,
   Mcm6, Dut, and Dctcpp1, were evidenced in the process of DNA
   replication and the promotion of fibroblastic proliferation, suggesting
   that these outcomes favored skin tissue repair. Last, we believe that
   the design of ultrasound driven MN patch not only can apply to treat
   acne infection efficiently but also may expand to other skin wound
   infections.

MATERIALS AND METHODS

Preparation of ZnTCPP@ZnO and MNs

   To prepare the ZnTCPP@ZnO, 10.0 mg of
   4,4,4,4-(porphine-5,10,15,20-tetrayl) tetrakis (benzoic acid) (TCPP)
   was dissolved in 50 of ml N,N′-dimethylformamide to form a homogeneous
   solution. After the addition of 10 ml of zinc nitrate (2.5 mg ml^−1;
   aqueous solution) and 10 mg of benzoic acid, the mixture was heated to
   150°C and reacted for 80 min under stirring. Then, the ZnTCPP@ZnO was
   obtained by centrifugation and washed with ethanol several times. We
   used a freeze dryer to get ZnTCPP@ZnO powder for later use.

   For the MNs, we dissolved 600 mg of sodium hyaluronate in 4 ml of
   deionized water and stirred it until completely dissolved. We then
   dropped 150 μl of solution onto the MN mold. After filling the mold, it
   was kept in a vacuum drying oven for 5 min. Bubbles in the mold were
   removed, and the mold was dried at room temperature for 12 hours to
   obtain the MN. To load the MNs with ZnTCPP@ZnO, the prepared
   nanoparticles were mixed into the dissolved sodium hyaluronate in the
   first step of patch fabrication before repeating the subsequent steps
   described above.

Characterization of nanoparticles and MNs

   The microstructure of all samples was observed by FESEM (JSM6510LV and
   Zeiss Sigma 500), inverted fluorescence microscope (IFM; Olympus,
   IX73), and HRTEM (Tecnai G2 20 U-Twin and Titan G2 60-300). The
   mechanical properties of MNs were evaluated by a universal mechanical
   testing machine (Instron 5982). The surface chemical composition of the
   samples was analyzed using XPS (ESCALAB 250Xi, Thermo Fisher
   Scientific, USA). To determine the phase structure of the hydrogels,
   the samples were assessed by XRD (D8A25, Bruker, Germany) with a 2θ
   range from 20° to 80° and a step size of 0.02°. The PL spectra of the
   samples were measured using a fluorescence spectrophotometer
   (PerkinElmer LS55, USA). The sonocurrent and electrochemical impedance
   results were detected at an electrochemical workstation (CHI660E,
   China). Measurements of ROS were performed on ESR (Bruker EMXplus). The
   release of Zn^2+ was recorded via inductively coupled plasma atomic
   emission spectrometry (Optimal 8000, PerkinElmer).

Theoretical calculations

   We used the VASP to perform all the DFT calculations within the
   generalized gradient approximation using the Perdew–Burke–Ernzerhof
   formulation. The projected augmented wave potentials were chosen to
   describe the ionic cores and to take valence electrons into account
   using a plane wave basis set with a kinetic energy cutoff of 400 eV
   ([197]50, [198]51). Partial occupancies of the Kohn-Sham orbitals were
   allowed using the Gaussian smearing method and a width of 0.05 eV
   ([199]50, [200]51). Geometry optimization was considered convergent
   when the force change was smaller than 0.02 eV/Å. Grimme’s DFT-D3
   methodology was used to describe the dispersion interactions. In model
   1, an MOF unit resides in a periodic box with side lengths of 25 Å by
   25 Å by 20 Å. In model 2, one Zn (001) slab with two Zn─O layers (81 Zn
   and 81 O atoms) was added to model 1. During structural optimizations,
   the Γ point in the Brillouin zone was used for k-point sampling, and
   all atoms were allowed to relax. Last, the transition states for the
   elementary reaction steps were determined by a combination of the
   nudged elastic band (NEB) method and the dimer method. In the NEB
   method, the path between the reactant and the product was discretized
   into a series of structural images ([201]50, [202]51). The image
   closest to a likely transition state structure was then selected as an
   initial guess structure for the dimer method.

In vitro antibacterial activity assay

   The spread plate method was used to analyze the antibacterial rate of
   all samples. P. acnes (CCTCC AB 2011019) suspensions (10^5 CFUs/ml)
   were treated by diverse groups, including the control. The ultrasound
   parameter was set as 1.0 MHz, 1.5 W cm^−2, and 50% duty cycle for 15
   min. After treatments, the diluted bacterial suspensions (20 μl) were
   collected and spread on Luria-Bertani agar plates and placed in a
   vacuum environment at 37°C for 2 days. Similarly, the MRSA (CCTCC AB
   2015108) strain (10^7 CFU/ml) was subjected to the same operation,
   except that MRSA only needed to be incubated at 37°C for 1 day. The
   morphologies of the treated bacteria were observed using FESEM. Further
   details of the procedure for the spread plate method and the formula
   for antibacterial efficacy can be found in our previous work ([203]44).

In vitro cytotoxicity studies and cell-related experiments

   NIH3T3 cells (RRID: CVCL_M025) were cultured in a medium supplemented
   with 10% fetal bovine serum and 1% penicillin-streptomycin solution
   (HyClone) in a humidified atmosphere of 5% CO[2] at 37°C, and the
   medium was changed every 2 days. Cell cytotoxicity was measured using
   the MTT assay. Before the MTT assay, all samples were prepared in
   48-well plates. The cells were seeded onto the MNs in the 48-well
   plates. After culturing for 1, 3, and 7 days, 400 μl of MTT solution (5
   μg/ml) was added to each well, and the cells were incubated at 37°C for
   4 hours. After the medium was removed, 400 μl of dimethyl sulfoxide was
   added to each well, and the plates were shaken for 15 min on a shaking
   table. The supernatant fluid was assessed using a SpectraMax i3
   platform (Molecular Devices, CA, USA). To align with the antimicrobial
   experimental protocol, all the MN samples used in the in vitro cell
   experiments were also exposed to ultrasound stimulation (1.0 MHz, 1.5
   W cm^−2, and 50% duty cycle) for 15 min on the day 1 of cell culture
   and the culture continued until day 7. The cell viability was assessed
   on days 1, 3, and 7 by MTT assay. At day 1, the cells were washed
   thrice with PBS (pH 7.4), fixed in 4% formaldehyde solution for 10 min
   at room temperature, and rinsed thoroughly with PBS. Afterward, the
   samples were stained with fluorescein isothiocyanate–phalloidin (YiSen,
   Shanghai) at room temperature in darkness for 30 min and further
   stained with 4′,6-diamidino-2-phenylindole (YiSen, Shanghai) for 30 s
   in darkness. The cell morphology of different samples was examined
   using an IFM (Olympus, IX73). For the transcriptome sequencing test and
   qRT-PCR analysis, 2 ml of NIH3T3 cell suspensions (1 × 10^5 cells/ml)
   was cultured in six-well plates with different samples for 72 hours.
   The cell lysates were then stored at −80°C before sequencing using a
   triazolo reagent (Beyotime Biotechnology). RNA sequencing was performed
   using Illumina HiSeq X10 (Illumina, USA). The value of the gene
   expression was converted to log[10] (transcript per million readings +
   1). The data were analyzed by the online Majorbio cloud platform. The
   qRT-PCR primer sequences are listed in table S2. For the ELISA assay,
   the concentrations of cytokines in cell growth medium and cells were
   measured by an ELISA assay kit (Shanghai Jianglai Biotechnology Co.
   Ltd., Shanghai Enzyme-linked Biotechnology Co. Ltd.).

In vivo anti-acne experiments

   The animal experimental process was approved by the Center for Food and
   Drug Safety Evaluation, Hubei Center for Disease Control and
   Prevention, Hubei Academy of Preventive Medicine. The animal experiment
   ethics approval number was 202110166. All animals were maintained and
   used in accordance with the Animal Management Rules of the Ministry of
   Health of the People’s Republic of China and the Guide for the Care and
   Use of Laboratory Animals—Chinese Version. In addition, all animal
   experiments were conducted with the assistance of the Wuhan Pinuofei
   Biological Technology Co. Ltd.

   Hemolysis assay test was conducted before the in vivo experiment. Fresh
   RBCs (4 ml) were separated by centrifuge at 3000 rpm for 10 min at 4°C,
   washed with saline for three times, and diluted to final concentration
   [5% (v/v)]. Different samples were separately added to 500 μl of RBCs
   and incubated at 37°C for 4 hours. At the end of incubation, the
   solution was centrifuged at 3000 rpm for 10 min at 4°C, and the
   hemolytic activity was determined at 570 nm with a microplate reader:
   Rate of hemolysis (%) = [(OD[sample] − OD[negative])/(OD[positive] −
   OD[negative])] × 100%.

   The mice (Balb/c, female, 8 to 10 weeks, and 25 to 30 g) were weighed
   and anesthetized with the anesthetic isoflurane, the hair on the back
   of the mice was removed in a sterile biosafety cabinet using a shaving
   machine and a hair removal cream, and 100 μl of P. acnes (2 × 10^5
   CFUs/ml) was intradermally injected into the skin of the back of the
   mice after disinfection of the skin with 75% alcohol cotton balls to
   establish an animal model of common acne. There were five mice per
   group. The mice were treated with different materials for 7 days. MNs
   were pressed into the skin by hand against the acne site and held for
   3 min. The blank control group, medical acne ointment, patch, blank MN
   patch, MN patch, and MN patch were treated for 10 min with ultrasound
   (medical ultrasound probe, 1 cm in diameter). During the treatment
   period, the swelling volume of acne was measured daily with
   microcalipers. The changes in the acne on the back of mice were
   photographed and recorded at 1, 3, and 7 days after the experiment, and
   after euthanasia treatment, and the skin was cut along 2 mm of the
   wound edge to complete the tissue acquisition. The expression of
   factors related to wounds, underlying muscle repair, and inflammatory
   indexes were analyzed. For skin tissue acquisition and embedding tissue
   fixation, tissue fixation, tissue dehydration, tissue permeabilization,
   tissue wax immersion, tissue embedding, tissue sectioning were
   performed.

Statistical analysis

   All data in this work are presented as mean values ± SD. A one-way
   analysis of variance (ANOVA) and Student’s t test were used for
   significance analysis. *P < 0.05, **P < 0.01, ***P < 0.001, and
   ****P < 0.0001 were considered statistically significant. All
   experiments were repeated independently more than three times.

Acknowledgments