Abstract

   Contact lenses (CLs) are prone to adhesion and invasion by pollutants
   and pathogenic bacteria, leading to infection and inflammatory
   diseases. However, the functionalization of CL (biological functions
   such as anti-fouling, antibacterial, and anti-inflammatory) and
   maintaining its transparency still face great challenges. In this work,
   as a member of the MXenes family, vanadium carbide (V[2]C) is modified
   onto CL via a water transfer printing method after the formation of a
   tightly arranged uniform film at the water surface under the action of
   the Marangoni effect. The coating interface is stable owing to the
   electrostatic forces. The V[2]C-modified CL (V[2]C@CL) maintains
   optical clarity while providing good biocompatibility, strong
   antioxidant properties, and anti-inflammatory activities. In vitro
   antibacterial experiments indicate that V[2]C@CL shows excellent
   performance in bacterial anti-adhesion, sterilization, and anti-biofilm
   formation. Last, V[2]C@CL displays notable advantages of bacteria
   elimination and inflammation removal in infectious keratitis treatment.
     __________________________________________________________________

   Water transfer printing of 2D MXene onto contact achieved high
   transmittance, anti-inflammatory, and antibacterial properties.

INTRODUCTION

   Contact lenses (CLs) and orthokeratology (OK) lenses have played an
   important role in refractive correction for myopia patients. On the
   basis of material composition, CL can be classified as either rigid
   gas-permeable lenses fabricated from silicone and fluorine polymers or
   soft hydrogel and silicone hydrogel lenses. Approximately 140 million
   patients worldwide choose to wear CL instead of frame glasses each year
   because of their advantages, such as aesthetics, ease of movement, and
   low cost ([42]1). In the context of the increasingly widespread use of
   electronic products, myopia in adolescents has shown a high incidence
   trend in recent years ([43]2). According to a survey by the National
   Health Commission, the overall myopia rate among children and
   adolescents in China is 52.7%, with 14.3% for 6-year-old children,
   35.6% for primary school students, 71.1% for middle school students,
   and 80.5% for high school students, respectively ([44]3). OK lenses are
   made of hard corneal CL materials with high oxygen permeability
   ([45]4). Wearing OK lenses at night flattens the central part of the
   cornea, which can shorten the eye axis and gradually flatten the
   cornea’s curvature. Thus, patients can achieve clear vision during the
   day without wearing glasses. Notably, as a physical therapy method for
   myopia correction, OK lenses have become one of the most common methods
   for myopia correction ([46]5, [47]6). In addition, cosmetic CLs, also
   known as colored CLs, are popular lenses that can change the color and
   size of the pupils, helping to enhance the appearance of the wearer,
   and are loved by young consumers ([48]7, [49]8). However, prolonged CL
   wear and unhealthy ocular habits greatly increase the risk of bacterial
   adhesion to CL, predisposing to biofilm formation ([50]9). The bacteria
   in the biofilms avoid the killing of lysozyme that initially existed in
   tears. Compared to free bacteria, bacteria within the biofilms have
   increased antibiotic resistance by more than 500 to 1000 times ([51]10,
   [52]11). The proliferation and invasion of bacteria will further lead
   to the occurrence of infectious keratitis ([53]12, [54]13).

   To mitigate surface contamination issues with CL, coating modifications
   have emerged as a principal technology for enhancing anti-fouling
   performance and extending service life. For example, polyethylene
   glycol, heparin, zwitterion, and other hydrophilic coatings were
   constructed by chemical reactions on CL surfaces through “grafting to”
   and “grafting from” technologies ([55]14). The monolayer or
   layer-by-layer coating with interface effect can also be prepared onto
   CL surfaces by supermolecule self-assembly technology ([56]15, [57]16).
   Another type of surface modification technology is the construction of
   biomimetic coatings onto CL surfaces, including self-polymerization
   coatings of dopamine ([58]17) and derivatives resembling mussels
   ([59]18, [60]19), nacre-inspired mineralized films ([61]20), and
   barnacle-like coatings ([62]21). Among them, our team combined super
   spreading and biomineralization to construct a transparent and
   mechanically robust underwater superoleophobic film on the surface of
   optical materials ([63]20). Similarly, we constructed a transparent and
   biocompatible amyloid-like nanofilm on the CL surface. The nanofilm was
   used to encapsulate and release cyclosporin A for dry eye syndrome
   treatment, which displayed excellent therapeutic efficacy with an 82%
   increase in drug bioavailability ([64]21). However, the process of
   chemical modification is relatively cumbersome, which can easily lead
   to a decrease in the transparency of the CL. The preparation process of
   biomimetic-modified coatings is simple with high light transmittance,
   but they rarely have biological functions such as sterilization and
   anti-inflammatory properties.

   It is still a problem to be solved how to modify the surface of the CL
   with transparency and endow it with anti-fouling, bactericidal, and
   anti-inflammatory properties. Our previous review mentioned that
   traditional antibacterial interfaces mainly include three types of
   surfaces: anti-fouling surface, contact-killing surface, and
   bactericide release surface ([65]11). On this basis, composite
   functional antibacterial surfaces and intelligent antibacterial
   surfaces responsive to bacterial infection microenvironments have also
   been derived ([66]22, [67]23). For example, we have constructed an
   antibiotic-loaded coating on the CL surface by surface-initiated
   reversible addition-fragmentation chain transfer polymerization
   ([68]24). The reversible conversion of pH-responsive dynamic chemical
   bonds in the coating realizes multiple loading and intelligent release
   of antibiotics. In addition, researchers have reported the
   antibacterial functionalization modification of biomaterial interfaces
   by drug sustained release coatings responsive to highly expressed
   enzymes in the infected microenvironment ([69]25). In physiological
   environments such as body fluids or blood, bacterial adhesion is
   regarded as the critical initial step in biofilm formation ([70]26,
   [71]27). Once bacteria achieve a transition from reversible adhesion to
   irreversible adhesion, it will inevitably lead to biofilm formation and
   the development of bacterial resistance. Therefore, an antibacterial
   interface with both anti-adhesion and bactericidal effects is very
   beneficial to the safe use of CLs.

   In addition, when infection occurs, it inevitably leads to inflammatory
   reactions and tissue damage. Once the cornea is damaged or its
   transparency decreases, it cannot return to the normal morphological
   structure ([72]28, [73]29). Oxidative stress plays an important immune
   regulatory role in the inflammatory response process related to
   infection, such as bacteria-killing by oxidative free radicals (ROS,
   reactive oxygen species). However, oxidative stress is also essential
   in exacerbating inflammatory reactions, which also leads to corneal
   tissue damage. It has been proved that eliminating oxidative stress
   while killing bacteria to reduce inflammation can promote the repair of
   infected wounds ([74]30, [75]31).

   Among the newly reported antibacterial materials, two-dimensional (2D)
   materials including graphene and its derivatives, black scales, layered
   double hydroxides (LDHs), transition metal chalcogenide (TMDs), and 2D
   transition metal carbides (MXenes) ([76]32) have attracted tremendous
   attention in the field of material science. Two-dimensional materials
   used for antibiosis have advantages such as good biocompatibility, high
   sterilization efficiency, and no drug resistance development owing to
   multiple sterilization mechanisms including physical/mechanical damage,
   lipid extraction, oxidative stress, and photothermal/photodynamic
   effects ([77]33). For example, self-activating implants modified with
   hydroxyapatite/MoS[2] coating present a high antimicrobial efficacy
   against both Staphylococcus aureus and Escherichia coli because of
   bacterial respiration-activated metabolic pathway changes ([78]34).
   Two-dimensional MXenes, such as vanadium carbide (V[2]C) and niobium
   carbide, exhibit excellent reducibility and catalase (CAT)-like and
   superoxide dismutase (SOD)–like activities ([79]35, [80]36). A
   face-to-face water transfer printing technology is so far the only
   known way to accomplish both the growth and transfer steps of 2D
   single-layer film. In addition, this surface modification technique is
   compatible with arbitrary sizes and shapes of the substrate. For
   example, face-to-face water transfer printing technology has been
   applied to prepare a large-scale uniform graphene film on a flexible
   polyimide substrate ([81]37).

   In this work, V[2]C 2D MXenes were modified onto the CL surfaces
   through face-to-face water transfer printing technology ([82]Fig. 1A).
   Briefly, V[2]C MXene nanosheets (NSs) were obtained through
   hydrofluoric acid (HF) etching of bulk MAX-phase V[2]AlC ceramics and
   tetrapropylammonium hydroxide (TPAOH) intercalation. The CL was endowed
   with positive charge properties by polyethyleneimine (PEI) amination
   pretreatment to combine with unilaminar V[2]C through electrostatic
   interactions (V[2]C@CL). The uniform V[2]C coating displayed strong
   anti-adhesion, bactericidal, and anti-biofilm effects against both
   Gram-positive bacteria methicillin-resistant S. aureus (MRSA) and
   Gram-negative bacteria E. coli. The V[2]C@CL also showed excellent
   anti-inflammatory activity through the elimination of ROS based on its
   CAT-like, SOD-like, and glutathione peroxidase (GPx)–like activities.
   Notably, in the infectious keratitis treatment animal model, V[2]C@CL
   with both bactericidal and anti-inflammatory functions showed obvious
   advantages in infectious wound healing and corneal tissue protection
   ([83]Fig. 1B).

Fig. 1. Schematic illustration of V[2]C@CL preparation and application.

   [84]Fig. 1.
   [85]Open in a new tab

   (A) The process of preparing V[2]C@CL using the water transfer printing
   method. (B) Multifunctional properties of V[2]C@CL in prevention and
   treatment of infectious keratitis. PEI, polyethyleneimine; IL-1β,
   interleukin-1β; NF-κB, nuclear factor κB; TNF-α, tumor necrosis
   factor–α.

RESULTS

Preparation and characterization of V[2]C@CL

   As shown in [86]Fig. 2A, Al atoms were etched from the V[2]AlC MAX
   phase in HF solution to obtain multilayer V[2]C (mV[2]C). The obtained
   blocky precipitates were added to TPAOH aqueous solution for
   intercalation, and the supernatant was collected to obtain V[2]C MXene
   NSs aqueous solution. Compared to the block structure of V[2]AlC (fig.
   S1A) and the multilayer structure of mV[2]C (fig. S1B), the
   transmission electron microscopy (TEM) image of fully peeled V[2]C
   MXene NSs showed single-layer and 2D sheet-like structure ([87]Fig.
   2B). Selected area electron diffraction (SAED) patterns further showed
   the well-defined polycrystalline structure of V[2]C MXene NSs ([88]Fig.
   2C). The crystal forms of V[2]AlC and V[2]C were measured through x-ray
   diffraction (XRD) spectroscopy analysis ([89]Fig. 2D). XRD patterns
   revealed characteristic peaks at 13.70° and 41.46° in V[2]AlC,
   corresponding to the (002) and (103) crystal forms, respectively.
   Following etching, these signature V[2]AlC peaks were absent in the
   diffractogram of the resulting V[2]C material. A previously
   unidentified peak emerged at 7.54° in the V[2]C diffractogram,
   indicating the (002) crystalline phase. In the final dispersion of
   V[2]C MXene NSs, the predominant fraction of V[2]C NSs exhibited
   successful layering, with virtually no unreacted MAX phase impurities
   remaining. In addition, high-resolution TEM (HRTEM) unveiled the
   micro-lattice structure of V[2]C MXene NSs ([90]Fig. 2E) with a lattice
   spacing of 0.24 nm, thereby confirming the successful synthesis of
   V[2]C MXene NSs. The V[2]C MXene NSs exhibited a negative charge, as
   evidenced by a zeta potential of −38 ± 20 mV ([91]Fig. 2F), which was
   attributed to the presence of =O and -OH functional groups on the MXene
   surface. The inherent negative charges on the V[2]C MXene surface
   facilitated its interaction with positively charged CL, which had been
   pretreated through PEI amination.

Fig. 2. Preparation of V[2]C MXene NSs and characterization of V[2]C MXene
NSs and V[2]C@CL.

   [92]Fig. 2.
   [93]Open in a new tab

   (A) Illustration of the preparation of V[2]C MXene NSs. (B)
   Transmission electron microscopy (TEM) image, (C) selected area
   electron diffraction image, (D) x-ray diffraction patterns, (E)
   high-resolution TEM image, and (F) zeta potential of V[2]C MXene NSs.
   (G) Scanning electron microscopy (SEM) images of CL, V[2]AlC@CL,
   mV[2]C@CL, V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL. (H) Transmittance
   was assessed by an ultraviolet-visible (UV-Vis) spectrometer. (I and J)
   Typical x-ray photoelectron spectroscopy full-spectrum and V2p spectrum
   of V[2]C-1@CL. (K) SEM elemental mapping of V[2]C-1@CL.

   Subsequently, we used the face-to-face water transfer printing method
   to modify the positively charged CL surface with one to three layers of
   V[2]C to obtain V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL, respectively.
   Through the same steps, the V[2]AlC MAX phase without HF etching and
   mV[2]C without TPAOH intercalation was modified on the CL surface to
   obtain V[2]AlC@CL and mV[2]C@CL to be used as control. In the
   face-to-face water transfer printing process, the V[2]C MXene NS
   ethanol dispersion was slowly dripped into a glass dish containing
   deionized water (0.1 ml/s). Under the influence of the Marangoni
   effect, V[2]C MXene NSs rapidly diffused and formed a continuous thin
   film at the gas-liquid interface. Tweezers were used to clamp the edge
   of CL horizontally to slowly press down on the water and quickly lift
   it up to transfer the film onto the outer surface of CL. The CL was
   then folded to make the inner surface protrude, and V2C MXene was
   transferred onto the inner CL surface in the same way. The surface
   moisture was blown dry with nitrogen gas to obtain a single-layer V[2]C
   MXene NS–coated CL. This process was iterated to produce CLs coated
   with two and three layers of V[2]C MXene. As depicted in [94]Fig. 2G,
   scanning electron microscopy (SEM) imaging of CLs subjected to
   different coatings revealed that the surfaces of V[2]AlC@CL and
   mV[2]C@CL exhibited rough and uneven textures, while V[2]C-1@CL,
   V[2]C-2@CL, and V[2]C-3@CL essentially maintained their surface
   flatness.

   It is well known that the smoothness and flatness of the coating
   notably influences the transparency of CLs. To assess this, CLs
   modified by different kinds of coatings were placed on color printing
   paper, black and white printing paper, and gauze to evaluate the
   clarity of text and texture. A notable reduction in light transmittance
   was observed in V[2]AlC@CL and mV[2]C@CL groups due to the uneven
   coatings of V[2]AlC and mV[2]C. In contrast, V[2]C-1@CL, V[2]C-2@CL,
   and V[2]C-3@CL consistently maintained satisfactory transparency of
   text, color, and texture, akin to unmodified CL (fig. S2). This
   observation was further substantiated by ultraviolet-visible (UV-Vis)
   spectroscopy. Compared to V[2]AlC@CL and mV[2]C@CLwith transmittance
   below 60%, the transmittance of V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL
   all exceeded 80% ([95]Fig. 2H). To ensure adequate optical performance
   of CL, three-layer V[2]C was considered to be the limited number of
   layers for CL modification. Ellipsometry tests showed that the coating
   thicknesses of one-, two-, and three-layer V[2]C were 9.16 ± 0.65 nm,
   16.42 ± 0.23 nm, and 30.09 ± 0.75 nm, respectively (fig. S3). The
   electrokinetic analyzer for solid surface analysis demonstrated that
   the zeta potential of the CL surface increased from −31.25 ± 0.39 mV to
   35.52 ± 1.31 mV after PEI modification. In addition, the zeta
   potentials were measured as 4.33 ± 1.06 mV, −13.11 ± 1.69 mV, and
   −20.09 ± 1.22 mV for the V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL
   surfaces, respectively (fig. S4). Furthermore, after modification, the
   water contact angle of V[2]C-1@CL decreased from 42.35° ± 0.75° to
   15.5° ± 0.1° (fig. S5). Therefore, V[2]C@CL exhibited good
   hydrophilicity, which would provide improved wearing comfort. To
   simulate the shearing force of eyelids on CL, V[2]C-1@CL was rinsed
   with a water flow (10 ml/s). After 12 hours, V[2]C-1@CL retained
   favorable hydrophilicity with a mean contact angle of 24.97° (fig. S6).
   To further assess coating stability, V[2]C-1@CL was immersed in pH 4
   (HCl) and pH 10 (NaOH) solutions for 12 hours. Subsequent contact angle
   measurements confirmed the retained hydrophilicity. Thus, coating
   persistence under both acidic and basic conditions demonstrated the
   stability of the V[2]C MXene coating (fig. S7).

   Then, x-ray photoelectron spectroscopy (XPS) was used to analyze the
   types and valence states of elements on the surface of V[2]C@CL
   ([96]Fig. 2I). The XPS analysis showed binding energy absorption peaks
   at 516, 400, 153, and 102 eV, attributed to V2p, N1s, Si2s, and Si2p,
   respectively. These peaks originated from the V[2]C coating, pretreated
   PEI, and the intrinsic silicon hydrogel composition of CLs. In
   addition, peak fitting of the V2p region revealed that most V elements
   existed in the V^5+ oxidation state with binding energies at 516.3 and
   523.7 eV ([97]Fig. 2J). A minor fraction of V^4+ was also detected as
   indicated by the two fitting peaks at 513.3 and 520.6 eV, respectively.
   The high valence state of V in V[2]C@CL was due to the presence of
   oxygen-containing groups and oxidized monolayers on the surface of
   V[2]C MXene. SEM element mapping and corresponding elemental analysis
   were conducted to explore the types and distribution of elements on the
   V[2]C@CL surface ([98]Fig. 2K and fig. S8). The results showed that
   V[2]C and PEI were uniformly distributed on the CL surface with V and N
   as the characteristic elements, respectively. This observation
   confirmed that the uniform and high-density modification of CL with
   high transparency was successfully achieved through PEI pretreatment
   and face-to-face water transfer printing method.

Biomimetic enzyme activity of V[2]C@CL

   In this part, the biological enzyme-like activities of V[2]C@CL were
   explored including CAT-like, SOD-like, and GPx-like activities, which
   could eliminate overexpressed ROS in the microenvironment when
   infection and inflammation occurred. Among them, hydrogen peroxide
   (H[2]O[2]) was the main component of ROS and the substrate for further
   formation of free radical components. As illustrated in [99]Fig. 3A,
   oxygen (O[2]) and water (H[2]O) were the products during the catalytic
   decomposition of H[2]O[2] by V[2]C. To indirectly gauge the catalytic
   activity of each sample group, the amount of generated O[2] was
   quantified using a dissolved O[2] analyzer. As demonstrated in
   [100]Fig. 3G, CL and V[2]AlC@CL exhibited no CAT-like activity, and
   mV[2]C@CL only displayed minimal CAT-like activity. In contrast,
   V[2]C@CL exhibited robust CAT-like activity with O[2] production in a
   time-dependent pattern. Notably, as the number of V[2]C layers
   increased, the catalytic performance was strongly enhanced. Likewise,
   the CAT-like activity of V[2]C@CL was assessed through quantification
   of residual H[2]O[2] in the system using the H[2]O[2] assay kit
   ([101]Fig. 3D and fig. S9). V[2]C@CL exhibited excellent CAT-like
   activity compared to V[2]AlC@CL and mV[2]C@ CL. Moreover, V[2]C-2@CL
   and V[2]C-3@CL also exhibited significantly enhanced activity compared
   to V[2]C-1@CL. However, considering the transparency of CL and the
   already obtained high enzymatic catalytic activity, the number of V[2]C
   layers modified onto CL was not further increased.

Fig. 3. Biomimetic enzyme activity of V[2]C@CL.

   [102]Fig. 3.
   [103]Open in a new tab

   (A to C) Schematic illustration of CAT-like activity, SOD-like
   activity, and GPx-like activity of V[2]C@CL. (D to F) CAT-like
   activity, SOD-like activity, and GPx-like activity of different CL. (G)
   The time-dependent evolution of O[2] generation resulting from H[2]O[2]
   decomposition under different treatments. (H) total antioxidant
   capacity (T-AOC) of different CL. (I) Ultraviolet radiation-inhibition
   activity of different CL. Data are presented as means ± SD, n = 3,
   biological replicate, and significances are determined by one-way
   analysis of variance (ANOVA) with Tukey’s correction. *P < 0.05, **P <
   0.01, ***P < 0.001, and ****P < 0.0001; ns, no significance.

   Furthermore, superoxide radicals (•O[2]^−) are persistent ROS produced
   during the normal physiological activities of life metabolism, and
   their expression level significantly increases during infection and
   inflammatory reactions. As a key antioxidant enzyme against ROS in
   cells, SOD catalyzes the disproportionation of •O[2]^− to generate O[2]
   and H[2]O[2] to reduce oxidative stress ([104]Fig. 3B). Therefore, SOD
   mimics are used as potential therapeutic agents to combat inflammatory
   diseases caused by oxidative stress ([105]38). The test results showed
   that the SOD-like activity of V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL
   was 34.34, 45.94, and 61.13%, respectively, demonstrating much higher
   biomimetic enzyme activity than that in the control groups ([106]Fig.
   3E and fig. S10). In addition, GPx used intracellular tripeptide
   glutathione (GSH) as a reducing agent to catalyze the decomposition of
   H[2]O[2] into H[2]O, while oxidizing GSH to glutathione (GSSG). GSSG is
   reduced to GSH with the assistance of glutathione reductase and
   coenzyme nicotinamide adenine dinucleotide phosphate. Therefore, GPx
   played a crucial role in maintaining H[2]O[2] levels in the body
   ([107]Fig. 3C). The GPx kit was used to test the GPx-like enzyme
   activity of CLs in each group. The test results demonstrated that the
   GPx activity of V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL was 51.88,
   62.99, and 76.95%, respectively ([108]Fig. 3F and fig. S11).

   The total antioxidant capacity (T-AOC) of V[2]C@CL was measured using
   the Total Antioxidant Capacity Assay kit with ABTS as an indicator. As
   displayed in [109]Fig. 3H, the T-AOC of V[2]C-1@CL, V[2]C-2@CL, and
   V[2]C-3@CL reached 44.07, 52.5, and 64%, respectively. UV radiation was
   also used to stimulate the generation of ROS, including hydroxyl
   radical (•OH) and H[2]O[2]. The CLs in each group were submerged in
   phosphate-buffered saline (PBS) and exposed to UV irradiation for 30
   min to evaluate the UV protective capabilities of V[2]C@CL. As
   illustrated in [110]Fig. 3I, the quantification of produced ROS within
   each system was determined using 3, 3′, 5, 5′-tetramethylbenzidine
   (TMB) as an indicator. The result showed that V[2]C@CL significantly
   reduced the generation of ROS under UV irradiation. In comparison, the
   enzyme activities were negligible in the CL, V[2]AlC@CL, and mV[2]C@CL
   control groups. Therefore, it indicated that V[2]C could only exhibit
   various biomimetic enzyme activities in a monolayer state rather than
   in a stacked state. A uniform large-scale single-layer V[2]C film was
   obtained through the face-to-face water transfer printing method onto
   CL surfaces, providing sufficient specific surface area for substrate
   catalysis.

In vitro cytocompatibility, antioxidant, and anti-inflammatory properties

   After incubation of human corneal epithelial cells (HCECs) with
   different types of CL for 24 hours, both cell live/dead staining and
   Cell Counting Kit-8 (CCK-8) methods were used to conduct the
   biocompatibility testing. The live/dead staining images showed that
   most of the cells in each group exhibited green fluorescence
   representing live cells, while the dead cells were minimal with red
   fluorescence ([111]Fig. 4A). The CCK-8 results showed that the survival
   rate of cells cultured in each group was greater than 90% compared with
   the original CL, displaying no significant cytotoxicity in various
   groups ([112]Fig. 4B). Therefore, different types of CL sample all
   showed good cell compatibility. Then, the intracellular oxidative
   stress of HCECs induced by H[2]O[2] was used to detect the ROS
   scavenging effect of various coatings using 2′,7′-dichlorofluorescin
   diacetate (DCFH-DA) as a fluorescent probe. As indicated in [113]Fig.
   4C, clear green fluorescence appeared in CL-, V[2]AlC@CL-, and
   mV[2]C@CL-treated groups, indicating intracellular produced ROS in
   HCECs stimulated by H[2]O[2]. However, in the V[2]C-1@CL–, V[2]C-2@CL–,
   and V[2]C-3@CL–treated groups, there was almost no green fluorescence,
   indicating the elimination of intracellular ROS by V[2]C with various
   enzyme-like activities. In addition, immunofluorescence staining of the
   DNA damage marker γ-H2AX was conducted after H[2]O[2] induction of
   HCECs. As shown in fig. S12, the green fluorescence intensity of HCECs
   in the V[2]C-1@CL–, V[2]C-2@CL–, and V[2]C-3@CL–treated groups was
   significantly lower than that in the control groups, which demonstrated
   the ROS scavenging ability of V[2]C@CL in DNA protection based on the
   efficient enzyme-like activity.

Fig. 4. In vitro cytocompatibility, antioxidant, and anti-inflammatory
properties of V[2]C@CL.

   [114]Fig. 4.
   [115]Open in a new tab

   (A) Live/dead cell staining of HCECs after 24-hour incubation with
   different CL. Green indicates live cells, and red indicates dead cells.
   (B) The cytocompatibility of different CL toward HCECs through CCK-8
   examination. (C) Fluorescence micrographs of 2′,7′-dichlorofluorescin
   diacetate–stained HCECs after different treatments. (D) Fluorescence
   micrographs of IL-1β immunofluorescence staining of differently treated
   RAW 264.7 cells. (E to G) Enzyme-linked immunosorbent assay (ELISA)
   quantification of IL-1β, IL-6, and TNF-α in the cell culture medium of
   RAW 264.7 cells after different treatments. Data are presented as means
   ± SD, n = 3, biological replicate, and significances are determined by
   one-way ANOVA with Tukey’s correction. *P < 0.05, ***P < 0.001, and
   ****P < 0.0001.

   After infection, inflammatory reactions and oxidative stress in the
   immune system coexist and promote each other. Bacterial toxins
   stimulate lesional tissue and antigen-presenting cells to produce
   inflammatory cytokines and chemokines, which induces the recruitment of
   cytotoxic immune cells, including granulocytes, T cells, B cells,
   natural killer cells, and macrophages, to the lesion site. Immune cells
   kill pathogenic bacteria through the secretion of antibodies, oxidative
   stress, and phagocytosis. However, the simultaneous production of
   excessive ROS leads to oxidative stress in tissues, releasing cytokines
   and further exacerbating inflammatory reactions. In particular, if
   corneal tissue is damaged, then it is likely to cause a series of
   problems such as corneal defects, neovascularization, and scarring,
   ultimately leading to decreased corneal transparency. Therefore, if
   material-based killing effects replace or synergize immune killing
   effects while simultaneously reducing oxidative stress and inflammatory
   damage, then it may be a better treatment strategy to protect the
   integrity and transparency of the cornea. To verify this concept, RAW
   264.7 cells were stimulated by lipopolysaccharide (LPS) and cultured
   with different groups of CLs. The expression of various intracellular
   inflammatory factors including interleukin-1β (IL-1β), IL-6, and tumor
   necrosis factor–α (TNF-α) was observed after immunofluorescence
   staining ([116]Fig. 4D and figs. S13 and S14). It was found that the
   cells in the CL-, V[2]AlC@CL-, and mV[2]C@CL-treated groups showed
   obvious green fluorescence, indicating the massive expressions of
   IL-1β, IL-6, and TNF-α inflammatory factors. When infection occurred,
   it mainly stimulated nuclear factor κB (NF-κB) inflammatory pathways to
   mediate the expression of various inflammatory factors. However, in the
   V[2]C-1@CL–, V[2]C-2@CL–, and V[2]C-3@CL–treated groups, the green
   fluorescence was significantly reduced, indicating the highly reduced
   expressions of IL-1β, IL-6, and TNF-α, which certified the
   anti-inflammatory effect through eliminating oxidative stress.

   To further validate the anti-inflammatory effect, the culture media of
   RAW 264.7 cells were collected after different treatments, and
   inflammatory factors were quantitatively analyzed using an
   enzyme-linked immunosorbent assay (ELISA) kit ([117]Fig. 4, E to G).
   The significantly reduced expressions of IL-1β, IL-6, and TNF-α
   inflammatory factors were found in V[2]C-1@CL–, V[2]C-2@CL–, and
   V[2]C-3@CL–treated groups. In addition, compared to the expressions of
   inflammatory factors in CL and V[2]AlC@CL groups, only slightly reduced
   expressions were observed in the mV[2]C@CL group. This result was well
   consistent with the above immunofluorescence tests. In summary, on the
   basis of the CAT-like, SOD-like, and GPx-like activities, V[2]C@CL
   could catalyze ROS decomposition in cells stimulated by LPS to reduce
   the inflammatory response. The expressions of inflammatory mediators
   further decreased with the increase of the V[2]C layer number.

Anti-adhesion, bactericidal, and anti-biofilm properties

   In complex pathophysiological environments, biomaterials are
   susceptible to direct protein contact, further mediating bacterial
   adhesion. Once the bacteria colonize the material surface, they
   complete quantitative proliferation and extracellular matrix (ECM)
   secretion forming the initial state of the biofilms. Then,
   mushroom-like bacterial colonies continue to grow and interact with
   each other, which are regulated by the quorum sensing system to form a
   mature biofilm growth ecosystem ([118]27). Therefore, bacterial
   adhesion to the material interface is considered the initial and most
   important step in biofilm formation. To test the anti-adhesion ability,
   the CLs in each group were incubated with MRSA and E. coli solutions,
   respectively. After washing off the planktonic bacteria, the attached
   bacteria were further dissociated from the CL through ultrasound to
   form a bacterial suspension. The content of bacteria was measured by
   plate counting method to determine the anti-adhesion ability of CLs in
   each group. As shown in [119]Fig. 5, A and B, both initial CL and
   V[2]AlC@CL basically showed no anti-adhesion effect to MRSA and E.
   coli. Meanwhile, mV[2]C@CL also showed a limited anti-adhesion effect
   with a large amount of bacterial adhesion on CL. However, V[2]C-1@CL,
   V[2]C-2@CL, and V[2]C-3@CL all exhibited excellent anti-adhesion
   effects with adhesion rates of 8.98, 1.81, and 0.9% to MRSA and 22.11,
   5.47, and 0.95% to E. coli, respectively.

Fig. 5. Anti-adhesive properties of V[2]C@CL.

   [120]Fig. 5.
   [121]Open in a new tab

   (A and B) Images of bacterial colonies on agar plates and quantitative
   analysis of adhesion rate of bacteria onto different CL. (C) Images of
   crystal violet staining of different CL after incubation with bacteria.
   (D) SEM images of bacteria adhering on different CL surfaces. Data are
   presented as means ± SD, n = 3, biological replicate, and significances
   are determined by one-way ANOVA with Tukey’s correction. ***P < 0.001
   and ****P < 0.0001.

   To further determine the anti-adhesion ability of the coatings, the CLs
   were cultured with bacteria solutions and stained with crystal violet.
   Then, the CL surface was dyed blue-purple to quantify the adhered
   bacteria and secreted ECM. As shown in [122]Fig. 5C, the purple color
   on the surface of V[2]AlC@CL was even darker than that on the initial
   CL. The quantification of crystal violet staining showed that MRSA and
   E. coli adhesion on V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL surfaces
   accounted for 22.01, 15.07, 5.56% and 20.96, 12.79, 7.67%,
   respectively, of that adhered on pristine CL surfaces (fig. S15).
   Because of the rough surface fit for bacterial adhesion, the depth of
   purple on mV[2]C@CL and V[2]AlC@CL surfaces only negligibly decreased
   compared to that of pristine CLs, indicating the absence of bacterial
   anti-adhesion ability. On the contrary, there was basically no purple
   color on the surfaces of V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL,
   indicating that bacteria could not adhere to the single-layer 2D V[2]C
   surface. A similar result was also found in live/dead staining of MRSA
   and E. coli on CL surfaces (fig. S16). There was a large amount of
   green fluorescence with very little red fluorescence on initial CL,
   V[2]AlC@CL, and mV[2]C@CL surfaces indicating the live state of
   bacteria. In contrast, a remarkable increase in red fluorescence and
   substantially decreased green fluorescence on V[2]C@CL surfaces showed
   significant anti-adhesion and bactericidal effects. Fluorescence
   quantification analysis revealed green fluorescence was reduced by
   71.15, 80.02, and 94.02% for MRSA and 78.29, 88.18, and 96.17% for E.
   coli on V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL surfaces compared to
   that on initial CL surface, respectively. In addition, the bacterial
   adhesion on various CL surfaces was visually displayed through SEM
   testing ([123]Fig. 5D). As shown in the SEM images, a large number of
   bacteria accumulated on pristine CL, V[2]AlC@CL, and mV[2]C@CL
   surfaces. Furthermore, there was only a small number of bacteria on
   V[2]C-1@CL and V[2]C-2@CL surfaces, while there was basically no
   bacteria adhering to the surface of V[2]C-3@CL. On the one hand, the
   uniform coverage of V[2]C onto the CL surface was achieved through the
   face-to-face water transfer printing method. On the other hand, the
   single-layer V[2]C was not conducive to bacterial adhesion. In
   addition, the CLs in each group were incubated with MRSA and E. coli
   separately, and the bacterial solutions were plated to enumerate viable
   cells (fig. S17). The results showed that V[2]C-1@CL, V[2]C-2@CL, and
   V[2]C-3@CL exhibited killing efficacy of 51.20, 70.34, and 84.82%
   against MRSA and 59.44, 76.86, and 87.21% against E. coli,
   respectively, demonstrating potent bactericidal capabilities of V[2]C
   MXene coating. Further exploration of anti-adhesion and bactericidal
   mechanisms was conducted through RNA sequencing (RNA-seq) methods.

   Furthermore, biofilm formation and the survival of bacteria on CL
   surfaces were identified using a live/dead staining method after
   long-term contact culture of modified CL with bacterial solutions (72
   hours). Laser scanning confocal microscopy was used to observe the 3D
   biofilm morphology on CL surfaces ([124]Fig. 6A) and the ratio of red
   and green fluorescence intensity ([125]Fig. 6, B and C). The 3D
   structure observation showed that the formed biofilms on the initial CL
   surface showed the largest thickness with mainly green fluorescence,
   indicating the survival of internal bacteria. Similarly, uniform and
   distinct viable bacterial biofilms formed on V[2]AlC@CL and mV[2]C@CL
   surfaces with predominating green fluorescence. These two modification
   methods could not resist bacterial adhesion or effectively kill
   bacteria within the biofilms. In addition, in the V[2]C-1@CL–treated
   group, it was found that 50.88% of MRSA and 36.13% of E. coli were
   stained with red fluorescence by fluorescence intensity statistical
   analysis. This indicated that V[2]C-1@CL played a good bactericidal
   role in the contact with bacteria. As the increase of V[2]C layer
   number onto the CL surface, the biofilms content on the CL surface
   further decreased, accompanied by a significant increase in the
   proportion of red fluorescence. The ratio of red fluorescence intensity
   on V[2]C-2@CL and V[2]C-3@CL surfaces were 57.44 and 78.91% for MRSA
   and 61 and 77% for E. coli, respectively. The green fluorescence
   intensities on V[2]C-2@CL and V[2]C-3@CL surfaces decreased by 79.13
   and 92.94% for MRSA and by 83.56 and 96.35% for E. coli compared to
   initial CL, which indicated enhanced anti-adhesion and bactericidal
   effect as the increase of V[2]C layers. In particular, V[2]C-3@CL
   basically completely resisted the formation of E. coli biofilms and had
   the best bactericidal effect ([126]Fig. 6D).

Fig. 6. Anti-biofilm properties of V[2]C@CL and MRSA RNA-seq analysis.

   [127]Fig. 6.
   [128]Open in a new tab

   (A) Confocal 3D images of biofilms on different CL surfaces with live
   and dead bacteria represented by green and red colors, respectively. (B
   and C) Relative fluorescence intensity of MRSA and E. coli biofilms on
   different CL surfaces. (D) Schematic illustration of the antibacterial
   mechanism for V[2]C@CL. (E) Volcano plots from RNA-seq analysis
   depicting differentially expressed genes between the CL and V[2]C-3@CL
   groups. (F) Heatmap depicting differentially expressed genes related to
   bacterial biofilm formation, virulence, proliferation, and drug
   resistance. (G and H) KEGG pathways analysis and GO analysis of
   differentially expressed genes between the CL and V[2]C-3@CL groups.

   As shown in [129]Fig. 5, the surface of V[2]C@CL could resist most of
   bacterial adhesion in the short term (for several hours). However,
   long-term CL wearing inevitably resulted in the adhesion of pollutants.
   Fortunately, V[2]C@CL could effectively kill the adhered bacteria on
   the surface to resist biofilm formation. The bactericidal effect and
   mechanism of 2D materials such as graphene and its derivatives have
   been reported in previous works ([130]39, [131]40). However, the
   bactericidal mechanism has not been clarified for 2D transition metal
   materials modified at the surface of biomaterials. Therefore, MRSA and
   E. coli were respectively cultivated with CL and V[2]C@CL, followed by
   RNA-seq of bacteria to explore the antibacterial mechanisms. As shown
   in [132]Fig. 6E and fig. S18, differential gene expression analysis
   revealed that after treatment by V[2]C-3@CL, 139 genes were
   up-regulated and 230 genes were down-regulated in MRSA. While 29 genes
   were up-regulated and 30 genes were down-regulated in E. coli. In
   addition, the analysis showed that the down-regulated genes in MRSA
   were associated with biofilm formation, virulence, proliferation, and
   drug resistance, while the down-regulated genes of E. coli were related
   to biofilm formation, virulence, and proliferation ([133]Fig. 6F and
   fig. S19). The V[2]C modification significantly inhibited bacterial
   biofilm formation and decreased bacterial virulence, proliferation, and
   drug resistance.

   Furthermore, we investigated the differentially expressed genes in the
   Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO)
   enrichment analysis. KEGG is a database that contains information on
   genes and proteins, biochemical pathways, and regulatory pathways,
   which can be applied to check which of the differentially expressed
   genes belong to known KEGG pathways. If certain pathways are notably
   enriched, then it can reveal the biological processes or pathways that
   are most prominent and potentially important in the gene set. The
   enrichment of differentially expressed genes in 10 pathways was
   observed in KEGG pathway analysis ([134]Fig. 6G and fig. S20). The top
   three affected pathways in MRSA were glycolipid metabolism, limonene
   and pinene degradation, and ABC transporters. As for E. coli, the top
   three affected pathways were cationic antimicrobial peptide resistance,
   amino sugar and nucleotide sugar metabolism, and LPS biosynthesis. In
   addition, GO is a structured vocabulary to describe the biological
   processes, molecular functions, and cellular components associated with
   gene products. GO enrichment was determined by a similar analysis with
   GO terms instead of KEGG pathways. GO analysis of differentially
   expressed genes in MRSA showed that the three most relevant pathways
   were the plasma membrane, cell periphery, and nitrate metabolic process
   ([135]Fig. 6H). In E. coli, the top three relevant pathways were the
   biosynthetic process, plasmid maintenance, and protein glycosylation
   (fig. S21). On the basis of the above analysis, V[2]C@CL mainly damaged
   the cell membrane and other peripheral structures of bacteria, and
   acted on some metabolic pathways, thereby playing a significant role in
   bactericidal and anti-adhesion effects ([136]Fig. 6D).

In vivo biocompatibility and antibacterial activity evaluation

   For animal experiments, V[2]C-3@CL was selected owing to its excellent
   biomimetic enzyme activity, anti-adhesion, bactericidal effects, and
   acceptable light transmittance (table S1). Before the animal model
   intervention experiment, the V[2]C-3–coated CL was worn on the eyes of
   Sprague-Dawley (SD) rats for biocompatibility evaluation by slit lamp,
   hematoxylin and eosin staining (H&E), and tonometer. As shown in
   [137]Fig. 7A, the blood vessel patterns of the cornea were clear after
   incubation with V[2]C-3@CL and initial CL. There was no significant
   transparency reduction compared with normal cornea, as well as no
   obvious edema, redness, and inflammatory reactions. It revealed that no
   punctate fluorescence formed on the cornea after sodium fluorescein
   staining (fig. S22). Tear film breakup time showed no significant
   difference among different groups (fig. S23). Further H&E staining
   observation showed that the corneal texture of the V[2]C-3@CL–treated
   group was clear compared with the normal cornea ([138]Fig. 7B). There
   was no significant increase in thickness and inflammatory infiltration
   cells, indicating the absence of significant inflammatory reaction in
   the cornea. The tear secretion test strip showed that there was no
   significant difference in the amount of tear secretion among the
   groups, indicating that CL wearing did not affect the stability of the
   tear film and tear secretion (fig. S24). In addition, there was no
   significant change in intraocular pressure (IOP) during material
   contact with the eye (fig. S25). Therefore, the above animal
   experiments verified good biocompatibility of V[2]C-3@CL to the eyes
   without significant biological irritation and toxicity.

Fig. 7. In vivo biocompatibility evaluation of V[2]C@CL and therapeutic
efficacy assessment of V[2]C@CL in a keratitis model using SD rats.

   [139]Fig. 7.
   [140]Open in a new tab

   (A) Slit lamp images of rat eyes on days 0, 1, 3, 5, 7, and 10
   following different treatments. (B) Images of H&E-stained corneas from
   rats on day 10 after different treatments. (C) Schematic illustration
   depicting the in vivo assessment of therapeutic efficacy in an
   MRSA-infected keratitis model. (D and E) Images of bacterial colonies
   cultured from corneal homogenates on agar plates and quantification of
   viable bacteria remaining in corneas following different treatments. (F
   to H) Slit-lamp images, quantified corneal clinical scores (scales 0 to
   20), and central corneal thickness from bacterially infected rats
   following different treatments on days 0, 1, 3, and 5. Data are
   presented as means ± SD, n ≥ 3, biological replicate, and significances
   are determined by one-way ANOVA with Tukey’s correction. *P < 0.05, **P
   < 0.01, and ***P < 0.001.

   In the animal experiment of infectious keratitis treatment, infectious
   corneal wounds were established the day before treatment ([141]Fig.
   7C). Then, the initial CL, vancomycin, and V[2]C-3@CL were applied for
   5 days of intervention, and the treatment process was tracked with the
   slit lamp, central corneal thickness, and inflammation score
   measurements. After treatment, electroretinography was used to observe
   the physiological function of the retina. Furthermore, after
   euthanasia, the cornea was fixed, sliced, and subjected to inflammatory
   factor expression and H&E staining tests.

   As shown in [142]Fig. 7F, there was no significant improvement in
   corneal transparency in the control group without intervention and the
   initial CL–treated group with the corneas remained in a state of edema,
   redness, and inflammatory response. In the vancomycin-treated group,
   the corneal keratitis state was significantly reduced on day 3, but
   there was still a large area of white lesions in the central area. Only
   in the V[2]C-3@CL–treated group did the inflammatory response of the
   cornea obviously reduce 1 day after treatment, and the lesion basically
   disappeared on day 5. The tracked clinical inflammation score and
   corneal center thickness showed that the corneas in V[2]C-3@CL–treated
   group displayed rapidly decreased clinical corneal score and
   inflammatory reaction ([143]Fig. 7, G and H). On day 5 after treatment,
   the inflammation score of corneas in the V[2]C-3@CL–treated group was
   around 2.5, indicating the basic disappearance of inflammatory
   response. The decreased corneal thickness illustrated the reduction of
   corneal edema symptoms. Similarly, it was found that the corneal
   thickness in the V[2]C-3@CL–treated group quickly decreased and had
   already returned to normal corneal thickness level on day 5 (below 200
   μm). After treatment, corneal tissues were homogenized and diluted, and
   the bacterial was content quantified by plate counting. As depicted in
   [144]Fig. 7, D and E, the vancomycin-treated group displayed reduced
   bacterial content with a 61.93% decrease compared to the control group.
   In contrast, the V[2]C-3@CL–treated group exhibited an 86.40% decrease
   in bacterial content compared to the control group, resulting in the
   lowest bacterial level.

   Furthermore, electroretinography testing post-treatment revealed
   significantly reduced a-wave and b-wave amplitudes in the control,
   initial CL–, and vancomycin-treated groups, indicating substantial
   visual impairment. In contrast, the electroretinography spectrum of the
   V[2]C-3@CL–treated group retained high a-wave and b-wave amplitudes,
   suggesting maintenance of good visual function ([145]Fig. 8, A to C).
   Furthermore, the expressions of the classic inflammatory factors
   including IL-1β, IL-6, and TNF-α related to infection were analyzed
   using immunofluorescence ([146]Fig. 8D). Compared to a large amount of
   red fluorescence in the corneal region in the control, initial CL–, and
   vancomycin-treated groups, only the corneas in the V[2]C-3@CL–treated
   group showed almost no red fluorescence indicating the subsided
   inflammatory response. In addition, the real-time quantitative
   polymerase chain reaction (qRT-PCR) method was used to analyze the
   expression of mRNA related to inflammatory factors ([147]Fig. 8F).
   Compared to the control group, the mRNA expressions related to IL-1β,
   IL-6, and TNF-α in the V[2]C-3@CL–treated group significantly decreased
   by 79.33, 65.72, and 82.7%, respectively. Therefore, after V[2]C-3@CL
   treatment, effective bacterial infection elimination significantly
   reduced the expression of inflammatory factors. Then, H&E staining of
   corneal slices was also performed to observe the morphology and
   structure of the cornea, as well as the inflammatory infiltration into
   the cornea ([148]Fig. 8E). The H&E-stained corneal images showed
   significant corneal thickening with a large number of inflammatory cell
   infiltration and fiber structure disorder in the control and initial CL
   groups. In contrast, the corneal structure and thickness in the
   V[2]C-3@CL–treated group had returned to a healthy state. There was
   almost no inflammatory cell infiltration with a clear structural
   hierarchy of the cornea.

Fig. 8. Electroretinography and histological assessment to evaluate
therapeutic efficacy of V[2]C-3@CL.

   [149]Fig. 8.
   [150]Open in a new tab

   (A to C) Electroretinography recordings and quantitative analysis of
   a-wave and b-wave amplitudes of SD rats with corneal infections after
   different treatments. (D) Immunofluorescence images of IL-1β, IL-6, and
   TNF-α in SD rat corneal tissue. (E) Images of H&E-stained corneas. (F)
   qRT-PCR analysis of IL-1β, IL-6, and TNF-α mRNA levels in rat corneal
   tissues. Data are presented as means ± SD, n ≥ 3, biological replicate,
   and significances are determined by one-way ANOVA with Tukey’s
   correction. *P < 0.05, **P < 0.01, and ***P < 0.001.

DISCUSSION

   The increasingly widely used CL is vulnerable to the adhesion of
   pollutants and pathogens in the tears, leading to bacterial keratitis
   and other problems. In this work, a face-to-face water transfer
   printing method was developed to modify V[2]C MXenes onto CL to obtain
   a transparent coating with multifunctional properties. It was certified
   in XPS, SEM images, and UV-Vis spectra that uniformly distributed V[2]C
   coatings formed on CL surfaces pretreated with PEI amination. Water
   contact angle measurements and water flushing experiments demonstrated
   that the obtained V[2]C@CL exhibited favorable hydrophilicity and
   coating stability.

   In addition, V[2]C@CL showed significant CAT-like, SOD-like, and
   GPx-like activities, which greatly reduced intracellular ROS in HCECs
   and the LPS-stimulated inflammatory response in RAW 264.7 cells.
   Furthermore, V[2]C@CL showed excellent bacteria anti-adhesion and
   bactericidal effects in resisting biofilm formation. In RNA-seq, the
   gene expressions highly decreased in bacteria related to biofilm
   formation, virulence, proliferation, and drug resistance. It was also
   found that V[2]C@CL mainly damaged the cell membrane and other
   peripheral structures of bacteria and acted on some metabolic pathways
   for bactericidal and anti-adhesion properties. Compared with the
   reported drug-loaded antibacterial or anti-inflammatory CLs ([151]41,
   [152]42), the V[2]C@CL designed in this work shows the obvious
   advantage of not changing the body structure of the CLs with
   well-maintained light transmittance. The convenient modification
   process and simple composition of V[2]C@CL without additional drug
   loading highly facilitate its clinical application.

   In infectious keratitis treatment, V[2]C@CL exhibited much higher
   bactericidal and anti-inflammatory effects compared to clinical
   vancomycin eye drops, which effectively protected the structure of
   corneal tissue. Therefore, this facile material surface modification
   technology that is independent of the substrate can be widely applied
   in optical materials and implant surface modification for multiple
   biological functions.

MATERIALS AND METHODS

Materials

   V[2]AlC was purchased from Forsman Scientific Co., Inc. (Beijing,
   China). TPAOH and HF were purchased from Aladdin (Shanghai, China). PEI
   was purchased from Macklin (Shanghai, China). CL was purchased from
   Bausch & Lomb (Shanghai, China). HCECs and RAW 264.7 cells were
   purchased from the American Type Culture Collection (USA). CCK-8 was
   purchased from MedChemExpress (USA). Total Antioxidant Capacity Assay
   Kit, Hydrogen Peroxide Assay Kit, Total Superoxide Dismutase Assay Kit,
   DNA Damage Assay Kit, and Calcein/PI Cell Viability/Cytotoxicity Assay
   Kit were purchased from Beyotime (Jiangsu, China). Glutathione
   Peroxidase Assay Kit was purchased from Leagene (Beijing, China). Mouse
   IL-1β ELISA Kit, Mouse IL-6 ELISA Kit, and Mouse TNF-α ELISA Kit were
   purchased from Elabscience (Wuhan, China). E. coli (ATCC 25922) and
   MRSA (ATCC 43300) strains were obtained from Shanghai Luwei Microbial
   Science and Technology (Shanghai, China).

Characterizations

   TEM images, HRTEM images, and SAED patterns were acquired using a
   JEM-2100F instrument (JEOL, Japan). The zeta potential of V[2]C NSs was
   measured by dynamic light scattering using a Malvern Zetasizer Nano
   ZS90 (Malvern Instruments Limited, UK). XRD patterns were obtained with
   a Rigaku SmartLab diffractometer (Rigaku, Japan). XPS was performed on
   an Escalab 250Xi spectrometer (Thermo Fisher Scientific). The water
   contact angle was measured by OCA15EC (Dataphysics, Germany). UV-Vis
   spectrophotometry for transmittance measurements was conducted using a
   UV-1780 instrument (SHIMADZU, Japan). Coating thickness was tested
   using a UVISEL PLUS ellipsometer (Horiba, France). The zeta potential
   of CLs was measured by a SurPASS 3 electrokinetic analyzer (Anton Paar,
   Austria). Because of the requirement of the instrument on sample size
   for measuring (2 × 4 cm^2), glass slides with similar surface
   properties to CLs were used for solid surface zeta potential testing.
   SEM images and element mapping were obtained with a JSM-7500F
   microscope (JEOL, Japan). Bacterial and cell fluorescence staining
   images were acquired using a Leica DMi8 fluorescence microscope (Leica,
   Germany). Fluorescence imaging of cornea sections for H&E and
   immunofluorescence staining was conducted using a Leica DM4B
   fluorescence microscope (Leica, Germany).

Preparation of V[2]C MXene NSs

   V[2]AlC powder (1 g) was added to 30 ml of 40 wt% HF solution for
   etching at room temperature with magnetic stirring at 50°C for 72
   hours. The resulting multilayered V[2]C suspension was centrifuged at
   2300g for 10 min, and the supernatant was repeatedly washed with
   deionized water until the supernatant pH was above 5.0. The obtained
   lumpy precipitate was added to 40 ml of TPAOH solution and magnetically
   stirred for 24 hours at room temperature. The residual TPAOH was
   removed by rinsing three times with deionized water. After centrifuging
   at 1500g for 50 min, the supernatant was collected to obtain an aqueous
   solution of V[2]C MXene NSs.

Preparation of V[2]C@CL

   The silicone hydrogel CLs were removed from their packaging, washed
   with pure water to remove the commercial solution, and then immersed in
   a 5% PEI solution for 12 hours. Then, 1 ml of V[2]C MXene solution (5
   mg/ml) was centrifuged at 10,000 rpm for 1 hour. The resulting
   precipitate was resuspended in 5 ml of anhydrous ethanol and slowly
   dripped (0.1 ml/s) into a glass dish containing deionized water.
   Through the Marangoni effect, the V[2]C MXene NSs rapidly diffused and
   formed a continuous film at the gas-liquid interface. The PEI-modified
   CLs were touched to the water surface to transfer the film to the outer
   CL surfaces and then immediately lifted with V[2]C MXene bonded by
   electrostatic forces. Then, the CL was folded to protrude from the
   inner surface, and the V[2]C MXene layer was transferred onto the inner
   CL surface in the same way. A single-layer V[2]C MXene–coated CL was
   obtained after drying the surface moisture by nitrogen gas. This
   process was repeated to obtain double and triple V[2]C MXene layers
   coated CL, namely, V[2]C-1@CL, V[2]C-2@CL, and V[2]C-3@CL. To compare
   the performance of V[2]C MXene, MAX phase V[2]AlC, and mV[2]C, the
   above procedure was repeated using 1 ml of V[2]AlC suspension (5 mg/ml)
   and 1 ml of mV[2]C suspension (5 mg/ml) to prepare V[2]AlC@CL and
   mV[2]C@CL, respectively. After preparation, the samples were immersed
   in sterile deoxygenated water, sealed in containers, and stored at 4°C.
   To test the stability of the coating, V[2]C-1@CL was rinsed with a
   water flow (10 ml/s) for 0, 3, 6, and 12 hours. The water contact angle
   of the samples was measured after each rinse duration to verify the
   integrity of the coating. Furthermore, V[2]C-1@CL was immersed in HCl
   (pH 4) and NaOH (pH 10) solutions for 12 hours. Water contact angle
   measurements were then conducted to verify coating stability.

CAT-like, SOD-like, and GPx-like activities of V[2]C@CL

   CAT-like activity was measured using the Hydrogen Peroxide Assay Kit.
   First, Fe^2+ was oxidized to Fe^3+ by H[2]O[2] and reacted with
   dimethylphenol orange to produce a purple product. The amount of purple
   product was determined by reading absorbance at 560 nm using a
   microplate reader to indirectly calculate CAT-like activity. Briefly,
   circular 5-mm-diameter sections were cut from different kinds of CLs.
   Then, CL samples, H[2]O[2] solution, and H[2]O[2] detection reagent
   were incubated at 37°C for 60 min. The absorbance at 560 nm was then
   calculated using a microplate reader (SpectraMax 190, Molecular
   Devices) to explore the CAT-like activity. CAT-like activity was
   further assessed by measuring O[2] produced in the presence of 1 mM
   H[2]O[2]. The O[2] produced upon the addition of each CL group was
   quantified using the JPB-607A portable dissolved oxygen meter.

   The SOD-like activity was measured using the Total SOD Activity Test
   Kit. Xanthine oxidase in the kit reacted with xanthine to generate a
   superoxide anion, which reduced nitroblue tetrazolium (NBT) to
   formazan, a blue-colored product with an absorption peak at 560 nm.
   Briefly, circular 5-mm-diameter sections were cut from different kinds
   of CLs. Then, CL samples were incubated with NBT/enzyme working
   solution and reaction initiation solution at 37°C for 30 min. The
   absorbance at 560 nm was then read using a microplate reader to
   calculate SOD-like activity.

   The GPx-like activity was tested using the GPx assay kit. GPx catalyzed
   the conversion of reduced GSH and H[2]O[2] to GSSG and H[2]O,
   respectively. The remaining GSH reacted with
   5,5′-dithiobis(2-nitrobenzoic acid) to produce the yellow
   5-thio-2-nitrobenzoic acid (TNB) product. The amount of produced TNB
   was measured by reading the absorbance at 422 nm using a microplate
   reader to determine the reduction of GSH. This allowed the indirect
   calculation of the GPx-like activity of the samples. Briefly, circular
   5-mm-diameter sections were cut from different kinds of CLs. Then, CL
   samples were incubated with GSH working solution at 37°C for 5 min.
   Oxidizing working solution was then added and further incubated at 37°C
   for 30 min. The mixture was centrifuged at 3500g for 10 min after
   adding 2 ml of acidic precipitant. The supernatant was incubated with
   GPx assay buffer and benzoic acid color-developing solution for 1 min
   at room temperature before calculating GPx-like activity by reading
   absorbance at 422 nm using a microplate reader.

Ultraviolet radiation inhibition and T-AOC of V[2]C@CL

   CL samples were placed in 1 ml of PBS in microcentrifuge tubes and
   irradiated under a 365-nm UV lamp for 30 min at a dose of 20 J/cm^2. In
   the following exposure, 0.5 mM TMB substrate was added to each tube.
   The absorbance at 652 nm was subsequently measured using a microplate
   reader.

   T-AOC was measured using the Total Antioxidant Capacity Assay Kit. ABTS
   could be oxidized to green ABTS+ by an oxidant, which could be
   inhibited by antioxidants. The amount of produced ABTS+ was measured by
   a microplate reader with absorbance at 414 nm to calculate T-AOC.
   Briefly, circular 5-mm-diameter sections were cut from different kinds
   of CLs. Then, the CL samples, peroxidase working solution, and ABTS
   working solution were incubated together at 37°C for 60 min. The CL
   samples were then removed and T-AOC was calculated by reading
   absorbance at 414 nm using a microplate reader.

Cytocompatibility of V[2]C@CL

   Both cell viability assay kit (CCK-8) and live/dead cell staining
   assays were used to evaluate the cytocompatibility of V[2]C@CL.
   Briefly, HCECs were inoculated into 96-well plates at 10,000 cells per
   well. The plates were incubated in a cell culture incubator (37°C, 5%
   CO[2]) for 12 hours. Sterile 5-mm-diameter circular sections were cut
   from the different kinds of CLs and placed into wells for 24-hour
   incubation. For the positive control, 10 μl of dimethyl sulfoxide was
   added to the wells. Then, 100 μl of CCK-8 solution was added to each
   well and incubated for 1 to 4 hours. A microplate reader was used to
   measure the produced crystal at 450 nm. Furthermore, the
   cytocompatibility of the samples was evaluated using a live/dead
   staining method. HCECs were seeded in 96-well plates at 10,000 cells
   per well and incubated for 12 hours. CL samples were then added to the
   wells for 24-hour incubation. Last, the cells were stained with
   calcein/PI and imaged under a fluorescence microscope (DMi8, Leica) to
   observe cell viability.

In vitro antioxidant and anti-inflammatory properties of V[2]C@CL

   HCECs were seeded in 96-well plates at 10,000 cells per well and
   incubated for 12 hours. The medium was removed and replaced with 90 μl
   of fresh medium and 10 μl of 10 mM H[2]O[2] solution per well. The
   5-mm-diameter circular CL samples were added to the wells and incubated
   with HCECs for 12 hours. Then, the cells were stained with DCFH-DA and
   visualized under an inverted fluorescence microscope (DMI8, Leica). In
   addition, the antioxidant properties of the samples were further
   evaluated using a DNA damage detection kit. After incubation of cells
   with different kinds of CLs and H[2]O[2] as described,
   immunofluorescent staining for the DNA damage marker γ-H2AX was
   performed using the kit. The cells were harvested and subsequently
   imaged under an inverted fluorescence microscope (DMI8, Leica).

   In vitro anti-inflammatory capacity was determined by
   immunofluorescence staining of inflammatory factors and ELISA. For
   immunofluorescence staining, the RAW 264.7 cells were seeded in 24-well
   plates at 50,000 cells per well and incubated for 12 hours. The CL
   samples were added to the wells and the cells were stimulated with LPS
   (1 μg/ml) for 4 hours. After removing the CLs, immunofluorescent
   staining of inflammatory cytokines (IL-1β, IL-6, and TNF-α) was
   performed to acquire images using a fluorescence microscope (DM4B,
   Leica). Antibodies used for immunofluorescence are listed in table S3.
   For the ELISA assay, the RAW 264.7 cells were seeded at 50,000 cells
   per well and incubated for 12 hours before adding CL samples. The CL
   samples were added to the wells and the cells were stimulated with LPS
   (1 μg/ml) for 4 hours. For the negative control group, cells were
   incubated for 4 hours without any treatment. After the incubation, the
   cell culture medium was collected from each group. Levels of IL-1β,
   IL-6, and TNF-α were measured by ELISA kit.

Anti-adhesive, bactericidal, and anti-biofilm properties of V[2]C@CL

   The anti-adhesive of V[2]C@CL was determined using a spread plate
   method. Logarithmic growth phase (log-phase) MRSA and E. coli were
   diluted to 1 × 10^6 colony-forming units (CFU)/ml. CL samples were cut
   into 5-mm-diameter discs and placed into 96-well plates with 200 μl of
   bacterial suspension per well. Plates were incubated at 37°C for 24
   hours. The CL samples were carefully recovered and washed three times
   with PBS to remove unbound bacteria, and then transferred to sterile
   centrifuge tubes with 1 ml of PBS. Tubes were sonicated for 10 min and
   vortexed for 2 min to detach surface-adhered bacteria. MRSA suspensions
   and E. coli suspensions were 1000-fold and 10,000-fold diluted,
   respectively, followed by plating to enumerate CFU.

   The anti-adhesive ability was further examined by crystal violet
   staining, live/dead bacterial staining, and SEM observation. CL samples
   in each group were incubated with log-phase MRSA and E. coli at 37°C
   for 24 hours. After incubation, planktonic bacteria were gently washed
   away with PBS. After staining of CLs with crystal violet, the stain was
   solubilized with 30% acetic acid to evaluate the remained bacteria at
   the absorbance at 590 nm. In the live/dead bacterial staining test, the
   incubated CLs were gently washed with PBS and then stained with a SYTO
   9/PI dye mixture for visualization and imaging using TCS-SP8 confocal
   laser scanning microscopy (Leica, Germany). As for SEM observation, the
   incubated CLs were fixed in 2.5% glutaraldehyde at 4°C for 12 hours.
   Samples underwent stepwise dehydration in ethanol solutions (50, 70,
   90, 95, and 100%) for 15 min each, followed by drying in a vacuum oven
   for 24 hours. Dried samples were visualized and imaged using SEM.

   The bactericidal property of V[2]C@CL was determined using a spread
   plate method. The log-phase MRSA and E. coli were diluted to 1 × 10^3
   CFU/ml. The CL samples were incubated with the bacterial solution for
   12 hours, and the number of surviving bacteria was enumerated by plate
   counting.

   The anti-biofilm property of V[2]C@CL was determined by live/dead
   bacterial staining methods. Log-phase MRSA and E. coli were diluted to
   1 × 10^6 CFU/ml. CL samples were cut into 5-mm-diameter discs and
   placed in 96-well plates. Two hundred microliters of bacterial
   suspension was added to each well and incubated at 37°C for 72 hours to
   form biofilms with renewal of the bacterial solution every 24 hours.
   Samples were gently washed twice with PBS and stained with SYTO 9/PI
   dyes for visualization and capturing of 3D biofilm images using TCS-SP8
   confocal laser scanning microscopy (Leica, Germany).

RNA-seq analysis

   Log-phase MRSA and E. coli were diluted to 1 × 10^6 CFU/ml. Then, CL
   and V[2]C-3@CL were separately incubated with 5-ml aliquots of the
   bacterial solution at 37°C for 12 hours. The bacterial solutions were
   harvested to extract the total RNA. The concentration and purity of the
   extracted RNA were assessed using an Agilent 2100 Bioanalyzer. The
   complementary DNA (cDNA) was synthesized by reverse transcription and
   specific libraries and enriched by polymerase chain reaction (PCR).
   Sequencing was performed on Illumina NovaSeq 6000 platform by Shanghai
   Personal Biotechnology Co. Ltd. The raw data underwent quality
   assessment and filtration using fastp (version 0.22.0), to eliminate
   reads with adapters, ploy-N, and low quality. Subsequent analyses
   relied on the resultant high-quality clean data. The reference genome
   index was built with Bowtie2 and the filtered reads were aligned to the
   reference genome. The gene read count values were determined by HTSeq
   (version 0.9.1), representing the initial gene expression levels.
   Normalization of gene expression across various genes and samples was
   used by FPKM (fragments per kilobase of exon per million fragments
   mapped). DESeq (version 1.38.3) discerned differentially expressed
   mRNAs, with transcripts meeting |log[2]FoldChange| > 1 and P value <
   0.05 deemed significant. GO enrichment analysis was performed using the
   topGO package to determine biological functions enriched among
   differentially expressed genes. GO terms with P values under 0.05 were
   considered statistically significantly enriched. The ClusterProfiler
   package (version 4.6.0) was used for KEGG pathway enrichment analysis
   of differentially expressed genes to identify signaling pathways
   showing significant enrichment (P < 0.05).

In vivo biocompatibility of V[2]C@CL

   Six-week-old male SD rats weighing approximately 150 g were purchased
   from the Animal Administration Center of Wenzhou Medical University
   (Wenzhou, China). The rats were single-housed in an environment
   maintained at 25° ± 2°C temperature, 50 ± 5% relative humidity, and a
   12-hour light/dark cycle. Standard rodent chow and filtered water were
   provided ad libitum. All animal experiments were approved by the
   Institutional Animal Care and Use Committee of Wenzhou Medical
   University under protocol number wydw2021-0484. The procedures were
   carried out in strict accordance with the ARRIVE 2.0 guidelines. Nine
   rats in good health with similar weight were selected to ensure that
   they were free of ocular diseases and were randomly divided into three
   groups for PBS, initial CL, and V[2]C@CL daily treatment. Specifically,
   5 μl of PBS was dropped on the ocular surface of rats twice a day in
   the control group. The rats were fitted with initial CL and V[2]C-3@CL
   for 12 hours daily for a period of 10 days. IOP, tear film breakup
   time, and slit lamp examinations were conducted before treatment as
   baseline and on days 1, 3, 5, 7, and 10 after different treatments.
   After completion of the examination on day 10, tear secretion was
   examined using the Schirmer test in each group. The rat in each group
   was randomly selected for cervical dislocation and execution. Corneal
   tissues were fixed in 4% paraformaldehyde (PFA), embedded in optimal
   cutting temperature compound, sectioned, and subsequently stained with
   H&E.

Infectious keratitis treatment

   SD rats of similar weight with healthy corneas were selected. A
   3-mm-diameter region of the central corneal epithelium was scraped off.
   Subsequently, 10 μl of log-phase MRSA bacteria at 1 × 10^7 CFU/ml was
   applied to the corneal defects, and the eyelids were closed. After 24
   hours of incubation, a slit lamp examination was performed to observe
   the rat corneas to obtain keratitis eyes with uniform infection degree.
   Rats that had successfully developed keratitis were then randomly
   divided into four treatment groups using a random number generator,
   with five rats in each group, for the different interventions. The four
   groups included the control group (no treatment), initial CL group
   (fitted 12 hours/day), vancomycin group (5 μl of 5% vancomycin topical
   administration twice daily), V[2]C-3@CL group (fitted V[2]C-3@CL, 12
   hours/day). The total treatment duration was 5 days. During the
   experiment, any rats that died or sustained corneal ruptures were
   excluded. On days 1, 3, and 5, slit lamp examination was performed to
   observe the corneal transparency. The corneal thickness was also
   measured using a SP-3000 pachymeter (Tomey, Japan). Corneal
   inflammation was scored on the basis of seven criteria as previously
   reported (table S2) ([153]43). On day 5, electroretinography assays
   were done with a RETI-Port21 system (Roland, Germany). Rats were
   euthanized by cervical dislocation for eyeball removal. To quantify the
   viable bacteria remaining in corneas following various treatments, the
   corneal tissues were homogenized in 0.5 ml of sterile solution and
   serially diluted 100-fold. Aliquots of 100 μl from the dilutions were
   plated on agar and incubated at 37°C for 12 hours. Bacterial colonies
   were subsequently enumerated and photographs were acquired.

Histopathological evaluation and qRT-PCR analysis

   The corneal inflammation levels were evaluated via H&E-stained
   sections, immunofluorescence detection, and qRT-PCR quantification of
   IL-1β, IL-6, and TNF-α expression measurement. The eye samples were
   fixed in 4% PFA at 4°C for 1 hour. The corneas were excised along the
   corneoscleral margins using microscopic scissors. The obtained corneal
   samples underwent gradient dehydration at 4°C for 2 hours in 10, 20,
   and 30% sucrose solutions. The corneal tissues were embedded in an
   optimal cutting temperature compound. The tissues were sectioned into
   10-μm slices using a frozen sectioning machine and transferred onto
   slides. The sections were stained with H&E and immunofluorescence.
   Antibodies used for immunofluorescence are listed in table S3. The
   tissue morphology was photographed and observed under a fluorescence
   microscope (DM4B, Leica). For qRT-PCR measurement, corneal tissues were
   homogenized and total corneal RNA was extracted for reverse
   transcription to obtain cDNA. Then, qRT-PCR experiments were performed
   using the Taq Pro Universal SYBR Green qPCR Master Mix kit (Q712-02,
   Vazyme, Nanjing, China) to detect IL-1β, IL-6, and TNF-α gene
   expressions. The sequences of forward and reverse primers for cytokines
   testing are listed in table S4.

Statistical analysis

   All data were analyzed statistically using GraphPad Prism software
   (version 9.3.0). One-way analysis of variance (ANOVA) was used for
   comparisons between multiple groups. Statistical significance was
   defined as P < 0.05, with varying levels indicated as follows: *P <
   0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Acknowledgments