The bioactive bandage is designed to provide sustained oxygen generation and robust photobiomodulation, promoting keratinocyte migration, fibroblast proliferation, and angiogenesis. This addresses the hypoxic conditions and enhances bioenergetic supply to accelerate the healing process of diabetic wounds.1 The onset of these wounds often leads to significant mental health concerns and a reduced quality of life for patients. If left unmanaged, chronic diabetic wounds can progress to disability and even necessitate amputation. Although preventive measures and treatments such as wound dressings, antibiotics, and hyperbaric oxygen therapy (HBOT) have improved the prognosis of chronic diabetic wounds, they come with notable side effects and limitations. Effective and convenient treatment modalities against chronic diabetic wounds are highly appealing and of substantial significance. Low-level light therapy (LLLT), particularly low-level red and near-infrared light therapy (600 to 950 nm, less than $50\mathrm{mW}/\mathrm{c}{\mathrm{m}}^2$),2 has emerged as a promising approach for treating chronic wounds. The biological effects of LLLT are primarily mediated by photobiomodulation, where photons penetrate the tissue and interact with mitochondria, leading to increased cellular metabolism, reduced pain, and accelerated wound healing. LLLT offers several advantages, including safety, painlessness, and ease of use, making it an attractive option for patients with diabetic chronic wounds.3 However, the bulkiness of traditional light sources and the extended time required for in-clinic treatments have limited its widespread adoption. A wearable, flexible, and patient-friendly photonic device would significantly enhance the convenience and effectiveness of wound management.

LLLT has been extensively investigated as a promising noninvasive modality for the diagnosis and treatment of wound healing (Table S1 in the Supplementary Material). Despite the broad applications of light therapy, especially in treating various skin diseases, the ischemic and hypoxic conditions of diabetic chronic wounds pose significant challenges to achieving satisfactory healing outcomes. HBOT can temporarily improve tissue oxygenation and enhance wound healing, but it requires cumbersome equipment, frequent medical visits, and substantial treatment costs.4 Cyanobacteria, which produce oxygen through photosynthesis by absorbing red lights, offer a promising solution for hypoxic alleviation in diabetic wounds, cancers, and other ischemic diseases. Recently, Gao et al. developed microalgae-loaded microneedles for the treatment of diabetic bacterial infections, combining the continuous oxygen supply from Chlorella with the sterilization activities of poly (ionic liquid).5 Kang et al.6 reported a programmed treatment strategy using modulated light intensity to enable live Haematococcus to perform multiple functions, such as oxygen generation, reactive oxygen species (ROS) scavenging, and immune regulation for chronic diabetic wounds. Microalgae have also demonstrated excellent biosafety and biocompatibility in various applications, including gastrointestinal tract drug delivery, radioprotection, and the treatment of hypoxic cancers.7^,8 Therefore, microalgae represent a promising method for continuous oxygen generation under appropriate light exposure.

In the present work, we developed a wearable and flexible light-emitting bioactive fabric for diabetic wound care (Fig. 1). To fabricate this flexible light-emitting bandage (LEB), optical fibers were micro-punched on the fiber cladding to achieve side emission of illumination and were woven into conventional textile yarns to conform to the complex morphology of the human body. The fabricated LEB was immersed in a cyanobacterial dispersion for cyanobacteria impregnation, creating the light-emitting bioactive bandage (LEB@Cyan). The LEB@Cyan offers continuous oxygen supply under the optimized side emission for synergistic LLLT in chronic diabetic wound healing, presenting the photobiomodulation effects in promoting keratinocyte migration, fibroblast proliferation, and angiogenesis. This wearable light-emitting bioactive bandage represents a convenient, cost-effective, and efficacious strategy for treating chronic diabetic wounds, with significant potential for clinical and in-home commercial translation.

Different concentrations of cyanobacterial suspension (PBS, pH = 7.4) were placed into 35 mm dishes, which were then illuminated at different wavelengths (620, 640, 660, and 680 nm) and various power densities (0 to $10\mathrm{mW}/{\mathrm{cm}}^2$) using an LED optical device. At the same time, the dissolved oxygen concentration was recorded by an oxygen electrode (Unisense, Denmark). In addition, the cycles of a light-dark process were carried out in 15-min shifts.

HaCaTs, HUVECs, and HSFs were seeded in the cell climbing slices of the lower chamber of transwell 24-well plate at $5\times {10}^4\text{cells}$ per well overnight. A 33 mmol glucose and 3% O₂ were used to mimic the hyperglycemic and hypoxic conditions of the diabetic wound microenvironment. Then, the red light (RL) group was irradiated with RL for 30 min, whereas the RL + Cyan group was further treated with cyanobacteria ($1\times {10}^8\text{cells}/\mathrm{mL}$) in the upper chamber ($0.4\text{-}\mu \mathrm{m}$ pore-sized filters). All the treatment groups were pre-treated at 33 mmol of glucose and 3% ${\mathrm{O}}_2$.

The cell climbing slices were collected and fixed with 4% paraformaldehyde for 15 min and permeabilized with 0.3% Triton X-100/PBS for 30 min, and then, cells were blocked in 5% BSA/PBS for 30 min. These slices were further incubated with HIF-1α (CST, #36169), followed by the co-incubation of anti-rabbit IgG (CST, #4412S) for 30 min at room temperature. Finally, immunofluorescence images were obtained using fluorescence microscopy and immunofluorescence signals were quantified using ImageJ software.

HaCaTs, HUVECs, and HSFs were seeded into the lower chamber at a density of $1\times {10}^6\text{cells}$ per well. A 33-mmol glucose and 3% ${\mathrm{O}}_2$ were used to mimic the hyperglycemic and hypoxic conditions of the diabetic wound microenvironment. Then, the RL group was irradiated with RL for 30 min, whereas the RL + Cyan group was further treated with cyanobacteria ($1\times {10}^8\text{cells}/\mathrm{mL}$) in the upper chamber ($0.4\text{-}\mu \mathrm{m}$ pore-sized filters). Then, the intracellular adenosine triphosphate (ATP) content was tested using ATP Assay Kits (Beyotime, China).

The flexible LEB is constituted of side luminous fiber fabric and light-emitting diode (LED), whose wavelength is 660 nm. The tube was affixed to the lamp beads to facilitate the coupling with the side luminous fiber fabric. The intensity of the LEB irradiation can be adjusted between 0 and $3\mathrm{mW}/\mathrm{c}{\mathrm{m}}^2$.

To acquire the optimal efficiency of loading cyanobacteria, the swatches of side luminous fiber fabric ($1\mathrm{cm}\times 1\mathrm{cm}$) were immersed into cyanobacterial suspension ($1\times {10}^8$ or $5\times {10}^8\text{cells}/\mathrm{mL}$) for 10, 30, 60, 120, and 240 min, respectively. These swatches of side luminous fiber fabric were collected and immersed into methanol. The extracted chlorophylls from cyanobacteria were assayed using UV-Vis absorbance spectra.

To establish a standard curve of cyanobacterial concentration based on UV-Vis absorbance, a series of cyanobacterial suspensions with varying concentrations was initially prepared using a blood counting chamber. Then, the chlorophylls of these cyanobacterial suspensions were extracted and detected using UV-Vis absorbance spectra. The standard curve for cyanobacterial concentration was constructed by measuring the absorbance of the samples at 660 nm.

All animal studies conform to the guidelines of the Animal Care Ethics Commission of Shanghai Tenth People’s Hospital under the license of SHDSYY-2024-P0052. Goto-Kakizaki (GK) rat, developed from repeated inbreeding of glucose-intolerant Wistar rats, has been widely used to explore the development of diabetes. To evaluate the efficacy of LEB@Cyan for the therapy of chronic wounds, fifteen GK rats (male, 14 weeks old) were divided randomly into three groups ($N=5$) as a spontaneous type-2 diabetes mellitus model. Five Wistar rats (male, 14 weeks old) were selected as normal control groups. The full-thickness wounds (10 mm in diameter) were created on the dorsal skin after the removal of the hair and anesthetization. The rats from the LEB group and LEB@Cyan group wore LEB and LEB@Cyan for wound care, respectively. Specifically, wound treatment was conducted on days 1, 3, 5, and 7 of the creating wound (parameter for each phototherapy: 660 nm, $2\mathrm{mW}/{\mathrm{cm}}^2$, and 35 min). The photos of wounds were taken on days 1, 3, 7, 11, and 15, and wound size was analyzed using ImageJ software. At the endpoint of the evaluation period, the rats were euthanized and wound tissues were collected. The whole blood was collected at predetermined time points for systematic blood routine examination analysis. One faction of wound tissues and major organs (heart, liver, spleen, lung, and kidney) was fixed with 4% paraformaldehyde overnight for histological and immunofluorescence analysis. The other faction was stored at $-80°\mathrm{C}$ for mRNA sequence.

The optimal light exposure dose for photobiomodulation (PBM) has been extensively documented in prior literature. Sutterby et al.9 examined the effects of three distinct visible light wavelengths at very low intensities: yellow (585 nm, 0.09 mW, $\sim 3.7\mathrm{mJ}/{\mathrm{cm}}^2$), orange (610 nm, 0.8 mW, $\sim 31\mathrm{mJ}/{\mathrm{cm}}^2$), and red (660 nm, 0.8 mW, $\sim 31\mathrm{mJ}/{\mathrm{cm}}^2$). Their findings indicated that all tested wavelengths enhanced proliferation and reduced scratch area in the human keratinocyte cell line (HaCaT), though only red light significantly increased mitochondrial activity. Similarly, Chen et al. investigated the therapeutic potential of LED light (400 to 900 nm) for diabetic wound healing in both in vitro and in vivo models. In vitro results demonstrated that nearly all tested wavelengths (at $10\mathrm{mW}/{\mathrm{cm}}^2$ power density and 0.5 to $4\mathrm{J}/{\mathrm{cm}}^2$ energy density) promoted diabetic wound cell proliferation, with the exception of 510 nm. However, in vivo experiments revealed more nuanced physiological responses: wavelengths of 425 nm (blue), 630 nm (red), 730 nm, and 850 nm (both near-infrared) all accelerated wound healing. Notably, the infrared spectrum (630, 730, and 850 nm) exhibited significantly greater efficacy than blue light (425 nm), likely due to variations in tissue penetration depth.10 Collectively, these studies support the clinical efficacy of low-level red and near-infrared light therapy (600 to 950 nm, $<50\mathrm{mW}/{\mathrm{cm}}^2$) in accelerating wound healing, a modality now widely employed in PBM.

The robust photosynthetic oxygenation effect of cyanobacteria requires optimal parameters for light exposure. Cyanobacteria are capable of absorbing red light (620 to 680 nm) for photosynthetic oxygenation.11 To optimize the irradiation wavelength of LED for LLLT, we initially investigated the oxygenic production of cyanobacteria (${10}^8\text{cells}/\mathrm{mL}$) under LED irradiation at varied wavelengths (620, 640, 660, and 680 nm) and power density (0, 2, 4, 6, 8, and $10\mathrm{mW}/{\mathrm{cm}}^2$) [Fig. 2(a)]. Optimal oxygen production of cyanobacteria could be found at $2\mathrm{mW}/{\mathrm{cm}}^2$ of 660 nm light, $6\mathrm{mW}/{\mathrm{cm}}^2$ of 640 nm light, and $10\mathrm{mW}/{\mathrm{cm}}^2$ of 640 nm light, with up to $125\mu \text{mol}$ of dissolved oxygen generated [Figs. 2(b) and 2(c)]. With the consideration of the energy efficiency, 660 nm and $2\mathrm{mW}/{\mathrm{cm}}^2$ are set as the LED wavelength and power density respectively to irradiate cyanobacterial cells for future evaluations. We also compare the photosynthetic oxygenation performance of varied doses of cyanobacteria under predetermined light irradiation (660 nm, $2\mathrm{mW}/{\mathrm{cm}}^2$) within 15 min. Oxygen production by the cyanobacterial cells was found to be dose-dependent. Specifically, 2 mL of cyanobacteria ($1\times {10}^8\text{cells}/\mathrm{mL}$) generate the highest oxygen up to $126.8\mu \text{mol}$ under such light conditions [Fig. 2(d)]. The stability of cyanobacterial photosynthetic oxygen generation throughout the treatment period is crucial. We evaluated cyanobacterial activity (${10}^8\text{cells}/\mathrm{mL}$) through four consecutive light-dark cycles (15-min red-light irradiation at 660 nm, $2\mathrm{mW}/{\mathrm{cm}}^2$ followed by 15 min dark period) using a microelectrode (Unisense, Denmark) [Fig. 2(e)]. The oxygenation performance is robust and sustained within four lights on and off cycles in 2 h, implicating the excellent bioactivity of the cyanobacterial cells. During the fourth cycle, the dissolved oxygen concentration in the cyanobacterial solution consistently increased by over $120\mu \text{mol}$ under optimal light exposure (15 min), demonstrating that the oxygen generation capacity of cyanobacteria meets the requirements for wound treatment.

Wound healing is a biological process regulated by different types of cells including keratinocytes, fibroblasts, and endothelial cells.12 To investigate the photobiomodulation effects enabled by cyanobacterial cells under red light, the cell biocompatibility of cyanobacteria was initially evaluated on human immortalized keratinocytes (HaCaTs), human umbilical vein endothelial cells (HUVECs), and human skin fibroblasts (HSFs) cell lines respectively using a typical cell counting kit-8 assay. A concentration-dependent decrease in cell viability for both HaCaTs and HSFs with increasing co-incubation concentrations of cyanobacteria ($p<0.05$). Interestingly, HUVECs exhibited a marginal increase in viability at cyanobacteria concentrations of $1\times {10}^8$ and $5\times {10}^8\mathrm{CFU}/\mathrm{mL}$. However, analysis of variance (ANOVA) revealed no statistically significant association between HUVEC viability and cyanobacteria concentration ($p>0.05$). These findings uncover cell-type-dependent heterogeneity in response to cyanobacteria exposure among skin-associated cell lines. All these cell lines remained high relative cell viability (above 80%) when co-incubated with cyanobacterial cells at a dose of $5\times {10}^8\text{cells}/\mathrm{mL}$, indicating good biocompatibility (Fig. S1 in the Supplementary Material). To imitate the hyperglycemic and hypoxic conditions of the microenvironment of a diabetic wound, 33 mmol of glucose and 3% ${\mathrm{O}}_2$ were employed for cell culture and co-incubation [Fig. 3(a)].13 The HaCaT, HSF, and HUVEC were further treated with RL irradiation (RL, 660 nm, $2\mathrm{mW}/{\mathrm{cm}}^2$) or RL + Cyan. As an intracellular hypoxia indicator, hypoxia-inducible factor-$1\alpha$ (HIF-$1\alpha$) is immunostained to the cells under different treatment conditions. These cell lines from the hyperglycemic–hypoxic group with or without the RL irradiation exhibited a significant increase in red fluorescence compared with the cells from the normal group. The red fluorescence signals of the cells with RL + Cyan treatment drastically reduced to the normal level, illustrating that continuous oxygenation of cyanobacteria under RL irradiation can effectively alleviate the over-expression of intracellular HIF-$1\alpha$ [Figs. 3(b)–3(g)].

To investigate the migration effect of HaCaT, the scratch experiment was conducted for 24 h. Compared with the normal culture condition, the migration of HaCaT was substantially suppressed under hyperglycaemic-hypoxic conditions. Moderate alleviation of the migration could be observed for cells treated with RL under hyperglycaemic-hypoxic conditions. Cells from the RL + Cyan group exhibit the highest migration effect under hyperglycaemic-hypoxic conditions, as revealed by the improved percentage of the healing area from 22.7% (hyperglycaemic-hypoxic condition) to 49.3% (hyperglycaemic-hypoxic + RL + Cyan). The presence of RL + Cyan exhibits similar migration regulation effects toward HUVECs and HSFs (Fig. S2 in the Supplementary Material). To evaluate the pro-angiogenesis capability of HUVECs under different treatments, we performed a Matrigel tube formation assay. HUVECs were stained with calcein-AM for confocal observation [Fig. 3(k)]. In the normal group, HUVECs formed a dense network of cord-like structures when cultured on Matrigel. By contrast, HUVECs in the hyperglycemic-hypoxic group showed minimal tube formation. Further treatment with RL or RL + Cyan significantly enhanced tube formation, with the RL + Cyan group demonstrating superior results compared with the normal group. Quantitative analysis of tube formation revealed that the RL + Cyan group had more tube junctions (222.3) than the untreated group (3.3), indicating its effectiveness in improving the hyperglycemic-hypoxic microenvironment by enhanced photobiomodulation effects of cyanobacterial cells under RL. HSF proliferation also plays a crucial role in wound healing. With calcein-AM staining, the proliferation of HSFs in different groups was analyzed using a fluorescence microscope [Figs. 3(m) and 3(n)]. The hyperglycemic–hypoxic group exhibited a lower percentage of green fluorescence signal compared with the normal group. Cells from the RL group showed a moderate increase in green fluorescence signal, while the RL + Cyan group displayed the highest level of green fluorescence signal among all groups. These results suggest that the inhibitory effects of cells under hyperglycemic–hypoxic conditions can be reversed by the continuous oxygen production from cyanobacteria under RL irradiation.

The photobiomodulation effect of RL also lies in the photon-induced bioenergetic regulation of the cells. ATP, the primary intracellular energy currency, plays a crucial role in various physiological and pathophysiological processes associated with diverse cell activities. Elevated ATP levels can effectively ameliorate metabolic-related diseases such as degenerative osteoarthritis and diabetic chronic wounds.14 To validate the photobiomodulation effect of RL promotes the wound healing process through the augmentation of cellular bioenergetics, intracellular ATP concentrations of these cell lines (HaCaT, HUVEC, and HSF) with different treatment conditions were quantified using an ATP assay kit [Figs. 3(o)–3(q)]. RL treatment significantly increased mitochondrial ATP production in HaCaTs, HUVECs, and HSFs under hyperglycemic-hypoxic conditions. This enhancement is likely due to the conversion of photons into biogenic energy through the interaction of red-light photons with the cytochrome C chromophore within mitochondria.15^,16 Collectively, these results demonstrate that RL irradiation of cyanobacterial cells could generate diverse photobiomodulation effects including the promotion of intracellular oxygenation, keratinocyte migration, fibroblast proliferation, and angiogenesis, potentially benefiting the in vivo healing of chronic diabetic wounds with high effectiveness and biocompatibility.

The flexible light-emitting fabric was constructed by weaving methods connected with a rechargeable lithium-ion battery and a 660 nm LED source [Fig. 4(a)]. Under a dark environment, the 660-nm red light-emitted homogeneously from the surface of the flexible fabric, demonstrating that the LEB functions as a nearly planar light source. To evaluate the wearability and safety of the LEB, we placed it on the forearm of human subjects [Fig. 4(b)]. Photographs of the wearable device in actual use showed that the LEB perfectly fits the non-planar structure of the body. Participants wearing the LEB did not experience dermal sensitization or skin irritation, indicating the safety of the LEB for the human skin. The temperature of the LEB during operation was assessed using a thermal camera [Figs. 4(c) and S3 in the Supplementary Material). Thermographic images showed no significant rise in local temperature around the LEB after 1 h of use.

Next, the LEB was directly impregnated with cyanobacterial cells in the solution containing cyanobacterial cells at a concentration of $5\times {10}^8\text{cells}/\mathrm{mL}$ to yield LEB@Cyan. The color of the LEB changed from white to green [Fig. 4(d)]. As shown in Fig. S4 in the Supplementary Material, transmission electron microscope (TEM) images reveal the bacilliform-like morphology of cyanobacteria. From scanning electron microscope (SEM) images, the adhesion of intact rod-shaped cyanobacterial cells onto the fabric could be clearly observed [Fig. 4(e)]. Chlorophyll molecules were extracted from cyanobacteria of different concentrations using methanol, and a strong optical absorption peak near 660 nm was observed, suggesting that 660 nm light exposure effectively activates cyanobacterial photosynthesis [Fig. S5(a) in the Supplementary Material]. The quantitative relationship between the optical absorption at 660 nm and different cyanobacteria concentrations was determined to analyze the impregnation number of cyanobacterial cells of LEB@Cyan [Fig. S5(b) in the Supplementary Material]. The impregnated density of cyanobacteria of LEB@Cyan was evaluated by immersing the LEB in different concentrations of cyanobacteria ($1\times {10}^8$ or $5\times {10}^8\text{cells}/\mathrm{mL}$) at varied impregnation times. The cyanobacterial density of LEB@Cyan shows a positive correlation with both the cyanobacterial concentration of the impregnation solution and the impregnated time (Fig. S6 in the Supplementary Material). Considering the cyanobacterial concentration of the impregnation solution and the time spent for the impregnation process, LEB immersed in $5\times {10}^8\text{cells}/\mathrm{mL}$ cyanobacteria for 30 min was used to acquire LEB@Cyan (cyanobacterial density: $1.6\times {10}^7\text{cells}/\mathrm{mL}$) for further experiments. These results demonstrate that a wearable, flexible, and bioactive LEB@Cyan has been successfully fabricated. To further verify operational stability, we conducted additional measurements during the revision period, monitoring the optical power density of LEB@Cyan at various time intervals (0.5, 1, 3, and 6 h) under continuous operation (Fig. S7 in the Supplementary Material). The results confirmed that the optical power density was consistently maintained at $2\mathrm{mW}/{\mathrm{cm}}^2$ for up to 6 h, demonstrating robust stability in light emission parameters.

The therapeutic efficacy of LEB@Cyan was evaluated on a full-thickness wound on GK rats, a non-obese model of non-insulin-dependent diabetes mellitus (type II diabetes) with metabolic, hormonal, and vascular disorders similar to human diabetes [Fig. 5(a)]. Thirty diabetic rats were randomly divided into three groups: diabetic control, LEB treatment, and LEB@Cyan treatment. Ten Wistar rats served as the normal group. Full-thickness wounds were created on the backs of all rats using a 10-mm biopsy punch. LEB and LEB@Cyan dressings warped the wounds of rats from the LEB group and LEB@Cyan group, respectively [Fig. 5(b)]. Specifically, wound treatment was conducted on days 1, 3, 5, and 7 of the creating wound (dose for each phototherapy: 660 nm, $2\mathrm{mW}/{\mathrm{cm}}^2$, and 35 min). During the therapeutic timeframe, no significant variation in body weight was observed among the experimental groups compared with the normal group (Fig. S8 in the Supplementary Material). The wound areas of rats in different groups were recorded and measured throughout the evaluation period [Figs. 5(c) and 5(d)]. Compared with the normal group, the wound area of diabetic rats reduced much more slowly, confirming the pathology of chronic wounds. Both LEB and LEB@Cyan treatments accelerated wound healing in diabetic rats. On day 3, the wound area ratios of rats in the normal control group, the diabetic control group, the LEB group, and the LEB@Cyan group are 67.03%, 77.38%, 62.94%, and 56.75%, respectively. On day 7, the wound area ratios of rats in the normal control group, the diabetic control group, the LEB group, and the LEB@Cyan group are 17.80%, 41.99%, 36.28%, and 17.54%, respectively. Throughout the treatment period, the LEB@Cyan group exhibited a significantly accelerated wound healing rate compared with both the LEB group and the diabetic control group, with healing kinetics approaching those observed in the normal control group. These results indicate that optimized red light phototherapy effectively promotes diabetic wound healing, whereas the incorporation of cyanobacteria provides a synergistic enhancement to the therapeutic outcome. Notably, the wound area ratio of rats in the LEB@Cyan group (0.26%) was closer to the normal control group (0.74%) on day 15, suggesting that the application of LEB@Cyan effectively promoted chronic diabetic wound healing.

At the end of the therapeutic evaluation, all rats were sacrificed, and the blood was collected from the retro-orbital vein for routine examination and biochemistry. Major organs (heart, liver, spleen, lung, and kidney) were harvested for hematoxylin and eosin (H&E) staining for histopathological study. Blood routine indices showed no noticeable signs of abnormalities (Fig. S9 in the Supplementary Material). In addition, no distinct histological abnormalities were observed in the heart, liver, spleen, lung, and kidney of rats from any of the treatment groups (Fig. S10 in the Supplementary Material), indicating the excellent biocompatibility of LEB@Cyan for in vivo therapeutic applications. At days 3 and 15, wounds were collected and then assessed by immunohistochemistry and immunofluorescence [Figs. 5(e)–5(i)]. In Figs. 5(e)–5(g), the H&E and Masson’s trichrome staining of wound tissue of rats in different groups on day 3 reveal granulation tissue thickness and the percentage of collagen deposition. The granulation tissue thickness and the percentage of collagen deposition of the wound tissue of the diabetic control group were obviously reduced compared with the normal control group, whereas those of the diabetic wound tissue treated with LEB show a moderate increase, whereas those of the diabetic wound tissue treated with LEB@Cyan are close to those of the normal control group, which is due to hypoxic alleviation and photobiomodulation. CD31, a typical endothelial cell marker, was used to detect tissue angiogenesis [Figs. 5(e) and 5(h)]. Similar to the results of in vitro cell tests, LEB@Cyan treatment almost overcame the inhibitory effect on angiogenesis caused by the hypoxia of the diabetic pathophysiology. The fluctuation of HIF-1α immunofluorescence staining in different groups further demonstrated this explanation [Figs. 5(e) and 5(i)]. Cytokeratin 10 (CK10) and cytokeratin 14 (CK14) were utilized as markers for the stratum spinosum and basal keratinocytes, respectively. On day 15, immunofluorescence staining of the wound tissue sections for CK10 and CK14 revealed that the wound tissue in the LEB@Cyan group had nearly completed re-epithelialization, similar to the normal group (Fig. S11 in the Supplementary Material). To evaluate collagen deposition on the wound tissues, Sirius red staining was performed (Fig. S12 in the Supplementary Material). The wound tissue sections of rats from LEB@Cyan groups exhibited a significantly larger area of positive red collagen compared with the diabetic control group, indicating its superior efficacy in promoting collagen deposition and facilitating the healing of chronic diabetic wounds.

To further elucidate the underlying regulatory mechanisms of LEB@Cyan in wound healing, we conducted a whole-transcriptome analysis by sequencing the mRNA from the collected wound tissues of rats from different groups: the normal group, diabetic control group, LEB group, and LEB@Cyan group [Fig. 6(a)]. Among the entire transcriptome, 36,097 differentially expressed genes (DEGs) were identified and dimensionally reduced using unsupervised principal component analysis (PCA). The first two principal components, PC1 (60.45%) and PC2 (16.16%), accounted for the majority of the total variance [Fig. 6(b)]. The close clustering of the samples from the normal and LEB@Cyan groups in the PCA plot suggests that LEB@Cyan treatment can effectively regulate diabetic wound healing, restoring it to a condition closer to normal. The correlation of DEGs between samples from different groups was further analyzed using a heatmap [Fig. 6(c)]. The gene expression profiles of the normal and LEB@Cyan groups were found to be highly similar, indicating that LEB@Cyan treatment can normalize gene expressions in diabetic wounds. In addition, volcano plots were generated to identify DEGs between the normal and diabetic control groups. According to the empirical Bayes method (${\log }_2\text{FoldChange}\ge 1$; padj < 0.05), 2709 genes were upregulated and 2504 genes were downregulated in the diabetic control group compared with the normal group [Fig. 6(d)]. Upon treatment with LEB@Cyan, 7033 significant DEGs were identified in the diabetic wound tissue, with 3445 genes upregulated and 3588 genes downregulated [Fig. 6(e)]. Specifically, among the upregulated genes in the diabetic control group, 2330 were significantly reversed by LEB@Cyan treatment. Similarly, 2077 of the downregulated genes in the diabetic control group were also reversed by LEB@Cyan treatment, accounting for a total reversal percentage of 62.6% [Fig. 6(f)].

Gene Ontology (GO) enrichment analysis revealed that the 2504 downregulated genes in the comparison geneset of diabetic control versus the normal group were enriched in pathways related to epidermis development, skin development, keratinocyte differentiation, and epidermal cell differentiation [Fig. 6(g)]. This suggests that the chronic nature of diabetic wound healing may be attributed to dysregulation in these molecular pathways. Notably, the 3445 upregulated genes in the comparison geneset of LEB@Cyan versus the diabetic control group were enriched in pathways associated with skin development, epidermis development, keratinocyte differentiation, and the establishment of the skin barrier [Fig. 6(h)]. These findings indicate that LEB@Cyan treatment can reverse the inhibitory effects of diabetes on wound healing by modulating key molecular signaling pathways associated with wound healing and tissue regeneration.

The PI3K/AKT signaling pathway has been shown to accelerate wound regeneration by modulating skin-related cell migration and proliferation, collagen deposition, and angiogenesis.17 Meanwhile, the MAPK signaling pathway plays complex roles in different phases of wound healing. It is well-established that excessive activation of c-Jun N-terminal Kinase (JUN) and p38 MAPK induces stress, inflammation, and fibrosis in diabetic wounds.18 We further investigated the therapeutic mechanisms of LEB@Cyan on diabetic wound therapy in rats. The results demonstrate that LEB@Cyan therapy significantly upregulated PI3K/AKT signaling pathway-related genes while downregulating JUN and p38 MAPK signaling pathway-related genes, which reveals that LEB@Cyan promotes diabetic wound healing through coordinated modulation of both PI3K/AKT and MAPK signaling pathways (Fig. S13 in the Supplementary Material).

We have developed a wearable and flexible LEB@Cyan to treat diabetic chronic wounds with high effectiveness and biocompatibility. The optimal irradiation parameters of the RL were initially optimized as 660 nm, $2\mathrm{mW}/\mathrm{c}{\mathrm{m}}^2$ to induce the highest cyanobacterial oxygenation (126.8 mmol per ${10}^8\text{cells}$) during LLLT. Using hyperglycemic-hypoxic culture conditions to imitate the microenvironment of diabetic wounds, accelerated migration of keratinocytes, enhanced endothelial angiogenesis, and improved fibroblast proliferation could be found. Hypoxic alleviation and enhanced bioenergetics were also confirmed as the in vitro photobiomodulation effects of LEB@Cyan. From in vivo therapeutic investigations on chronic wounds in diabetic rat models, sustained oxygen supply and enhanced wound healing performance could be confirmed for the diabetic rats treated with LEB@Cyan dressings. Transcriptional regulations of signaling pathways such as wound healing and tissue regeneration could also be found in rats treated with LEB@Cyan. The present work provides a convenient, effective, and biocompatible wearable dressing to treat ischemic and unhealed skin diseases such as diabetic chronic wounds, offering insights into personal management using self-deployed optical devices for commercial and translational applications.

The statistical significances in this work were analyzed via a two-sided Student’s $t$-test using the software SPSS 20 statistics (version 26.0), n.s. for non-significant; *$P<0.05$, **$P<0.01$, and ***$P<0.001$.

Miss Shenyan Zhang received her master’s degree from Fudan University. Her primary research focuses on the design, fabrication, and application of photomedical devices, as well as the mechanisms of photobiomodulation therapy.

Lingbao Kong is a full professor and the director of the Shanghai Engineering Research Center of Ultra-Precision Optical Manufacturing, School of Information Science and Technology, Fudan University. He received his PhD from The Hong Kong Polytechnic University. His research interests include intelligent optical manufacturing and machine vision measurement, full-process optimization of ultra-precision optical machining, bio-inspired functional structure design and fabrication, multi-spectral and freeform surface metrology, ultra-precision optics and related applications, etc.

Penghao Ji is currently a PhD candidate in basic medical sciences at the School of Medicine, Tongji University. His research focuses on the functional design and fabrication of biomaterials, as well as investigations into their biological effects. His work has been published in prestigious journals including Advanced Materials, Science Advances, Nano Today, Materials Today Bio, and Biomaterials.

Minfeng Huo is a research professor and doctoral supervisor at Tongji University, with a dual appointment as principal investigator at the Tenth People’s Hospital affiliated with Tongji University. He received his PhD from the Shanghai Institute of Ceramics, Chinese Academy of Sciences. His research focuses on the controlled synthesis of nano-biomaterials, their biomedical applications, and investigations into underlying mechanisms.