Abstract

   Extracellular vesicles (EVs), especially those derived from stem cells,
   have emerged as a novel treatment for promoting wound healing in
   regenerative medicine. However, the clinical application of mammalian
   cells-derived EVs is hindered by their high cost and low yields.
   Inspired by the ability of EVs to mediate interkingdom communication,
   we explored the therapeutic potential of EVs released by the probiotic
   strain Lactobacillus rhamnosus GG (LGG) in skin wound healing and
   elucidated the underlying mechanism involved. Using full-thickness skin
   wound-healing mouse models, we found that LGG-EVs accelerated wound
   healing procedures, including increased re-epithelialization and
   promoted angiogenesis. Using in vitro experiments, we further
   demonstrated that LGG-EVs boosted the proliferation and migration
   capacities of both epithelial and endothelial cells, as well as
   promoted endothelial tube formation. miRNA profiling analysis revealed
   that miR-21-5p was highly enriched in LGG-EVs and LGG-EV treatment
   significantly increased miR-21-5p level in recipient cells.
   Mechanically, LGG-EVs induced regulatory effects via miR-21-5p mediated
   metabolic signaling rewiring. Our results suggest that EVs derived from
   LGG could serve as a promising candidate for accelerating wound healing
   and possibly for treating chronic and impaired healing conditions.

Graphical abstract

   [44]graphic file with name 12951_2024_2893_Figa_HTML.jpg

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-024-02893-8.

   Keywords: Wound healing, Extracellular vesicles, Lactobacillus
   rhamnosus GG, miRNA sequence, Bacteria

Introduction

   Skin wound managements pose significant clinical challenges, due to
   frequent trauma, tearing, cuts, or burns [[45]1]. In certain
   pathological conditions, such as diabetes, obesity, and aging, the
   wound healing processes are disrupted, often resulting in pathological
   conditions, such as chronic or non-healing wounds, and even tumor
   formation, thereby threatening the life quality for millions of people
   worldwide [[46]2]. Despite numerous therapeutic strategies being
   reported, inconsistent clinical outcomes and undesirable side effects
   restrict their long-term efficacy [[47]1], underscoring the need to
   develop optimal treatment approaches.

   Wound healing is a complicated process that aims to regain the
   structural and functional integrity of damaged tissues. The healing
   procedure is orchestrated by multiple cells, signaling pathways, and
   bioactive molecules [[48]3, [49]4]. Typically, it involves a series of
   sequential and overlapping phases, including inflammation,
   proliferation (granulation tissue formation and re-epithelization), and
   resolution. Immediately following injury, blood vessels constrict to
   minimize blood loss and immune cells are recruited into the injured
   sites, removing microbes and orchestrating the multiple healing
   processes. Concurrently, activated fibroblasts migrate to the wounds,
   forming the granulation tissue, which contains abundant extra-cellular
   matrix (ECM) and blood vessels. The filled wound space provides a
   scaffold for tissue repair. In this case, endothelial cells develop new
   blood vessels that provide nutrients and oxygen to lesion sites,
   alleviating uncontrolled inflammation. Keratinocytes proliferate and
   migrate across the wound surface to finish re-epithelization to restore
   the barrier function. Finally, these episodes stop and the tissue is
   remodeled in resolution stage. In diabetic wounds, however,
   hyperglycemia and long-term oxygen deprivation lead to continuous
   pro-inflammatory state, disturbed angiogenic response and compromised
   re-epithelialization, which subsequently contribute to wound healing
   failure. Thus, a feasible approach for chronic wound healing would
   simultaneously induce angiogenesis and modulate epidermal cell
   functions.

   Extracellular vehicles (EVs) are membrane-enclosed particles secreted
   by almost all cells, ranging in size from 30 to 150 nm in diameter and
   containing encapsulated bioactive components, such as microRNAs
   (miRNAs), proteins, and mRNAs [[50]5]. Recent studies have highlighted
   their remarkable biomedical potential in various wound healing
   processes, including re-epithelization, immunoregulation, and ECM
   remodeling [[51]6]. EVs derived from mammalian cells, as exemplified by
   mesenchymal stem cells (MSCs) [[52]7, [53]8], macrophages [[54]9], and
   endothelial progenitor cells [[55]10], have been documented to
   accelerate wound closure by presenting pleiotropic effects
   simultaneously. Despite the potential advantages of EVs as therapeutics
   in regenerative medicine, their clinical application still faces many
   limitations, including difficulty in large quantities, high cost of
   cell culture, and ethical issues.

   Interestingly, bacterial EVs (BEVs) have emerged as an alternative with
   the potential to overcome these limitations. BEVs function as mediators
   of interkingdom communication enabling the shuttle of bioactive
   components and regulating various biological processes, making them
   promising candidates for disease treatment [[56]11]. Several BEVs have
   been reported for disease therapy. For example, Akkermansia
   muciniphila-derived EVs improved intestinal barrier function [[57]12].
   EVs from Lactobacillus druckerii displayed great promise in
   anti-fibrosis [[58]13]. Lactobacillus reuteri extracts haven been shown
   to elicit therapeutic effects on injury repair via PI3K-AKT pathway
   [[59]14]. On the other hand, oral antibiotic treatment delays wound
   repair [[60]15], and supplementation of gut microbiome in drinking
   water accelerates wound healing process [[61]16], indicating that gut
   bacteria or their products may participate in skin wound healing.
   However, little is known about the potential role of EVs from
   probiotics in wound healing. Lactobacillus rhamnosus GG (LGG) is a gut
   probiotic that benefits health in many aspects, including enhancing
   intestinal barrier function [[62]17], regulating gut microbiota
   structure [[63]18], and promoting fin regeneration of zebrafish
   [[64]19]. Particularly, LGG lysate can increase the
   re-epithelialization of keratinocytes [[65]20], which endows them with
   the possibility of pro-healing action. However, the therapeutic effects
   and possible mechanisms of LGG-EVs on cutaneous wound healing remain to
   be determined.

   Here, we isolated EVs from LGG and investigated their effects on wound
   healing both in vivo and in vitro. Our results demonstrated that the
   local injection of LGG-EVs promoted mouse wound closure, accompanied by
   accelerated re-epithelialization and increased angiogenesis. Moreover,
   LGG-EVs were found to enhance the proliferation and migration of
   keratinocytes and endothelial cells, as well as the proangiogenic
   ability of endothelial cells. Using miRNA sequencing, we identified
   that LGG-EVs contained a high amount of miR-21-5p and mechanistic
   investigation revealed that this miRNA is crucial for LGG-EVs-mediated
   activation of endothelial cells and keratinocytes that was related to
   AKT and HIF1α signaling pathways. Collectively, our study
   comprehensively illustrates the therapeutic benefits and mechanisms of
   action for LGG-EVs in wound healing and open innovative avenues for
   probiotic bacteria EVs in regenerative medicine.

Materials and methods

LGG culture

   LGG strain was obtained from ATCC 53103 and cultivated in MRS medium
   (Solarbio Science & Technology) [[66]21]. Briefly, after twelve hours
   of culture at 37 °C and 220 rpm (primary culture), LGG was cultivated
   at 37 °C and 220 rpm for 8 h (secondary culture). Subsequently, the
   secondary fermentation broth (1 mL) was supplemented with MRS (50 mL)
   and cultivated at 37 °C and 220 rpm for 24 h. The obtained fermentation
   supernatant was used for EV isolation.

Isolation and identification of LGG-EVs

   The isolation of EVs from LGG was conducted as previously reported
   [[67]21]. Firstly, the fermentation supernatant was centrifugated at
   10,000 g for 15 min to wipe off bacteria and other debris, followed by
   filtration with filters (0.22 μm) to remove large particles. After
   ultra-centrifugation of the solution at 150,000 g for 2 h, the pellet
   was sequentially resuspended with PBS, concentrated through an
   ultrafiltration membrane (50–100 kDa), and subjected to a second step
   ultracentrifugation at 150,000 g for 2 h. The pure EV pellets were
   resuspended with PBS and filtered with 0.22 μm sterile filters. During
   EV preparation, all processes were carried out at 4 °C. The putative
   concentration of LGG-EVs was measured with a BCA assay kit (Beyotime,
   P0010). The particle size distribution of LGG-EVs was determined by
   Nanoparticle tracking analysis (NTA; ZetaView). The morphology of
   LGG-EVs was detected by Transmission electron microscopy (TEM;
   Hitachi). The zeta potential of LGG-EVs was examined by dynamic light
   scattering (Zetasizer Nano ZS90, Malvern).

Mouse skin wound models and LGG-EV treatment

   Eight-week-old male C57BL/6J mice (Vital River Laboratory Animal
   Technology) were used in present study. After anesthetization and hair
   removal, four full-thickness incisions (6 mm) were made on both sides
   of midline of mouse back [[68]22]. Twenty mice were randomly separated
   into two groups Control vehicle (PBS treatment) group and the LGG-EVs
   group (n = 5 mice per group at each time point). The mice were
   subcutaneously injected with either 100 µL PBS (Control) or LGG-EVs
   (50 µg LGG-EVs in 100 µL PBS) around the wounds once a day,
   respectively. On day 0, 4, and 7 post-wounding, a digital photograph of
   the wounds was taken. Wound areas were measured by referring to a
   circle slide with 6 mm in diameter. The wounds and slides were
   photographed, and the wound area was calculated using ImageJ. On day 4
   and 7, mice were sacrificed. The wounds were picked for H&E,
   immunohistochemical and qRT-PCR analysis. All experimental procedures
   were approved by the Institutional Animal Care and Use Committee of
   Shanghai University.

Cell culture

   The human skin keratinocytes (HaCaT cells) were obtained from the Cell
   Bank of the Chinese Academy of Sciences (Shanghai, China), and the
   human umbilical vein endothelial cell line (HUVEC) was purchased from
   OriCell (Guangzhou, China). HaCaT was cultured in DMEM containing
   10%FBS and 1% penicillin-streptomycin. HUVEC was cultured in
   endothelial cell medium (ECM #1001; ScienCell) as previously described
   [[69]23]. HUVEC of passage 3 or 4 was used for experiments in our tests
   and HaCaT of passage less than 7. The culture medium was changed every
   other day and the cells were passaged when they are confluent.

Histological and immunofluorescent analysis

   After fixed with 4% paraformaldehyde (PFA), the wound tissues were
   dehydrated and embedded in paraffin. Then, 5 μm-thickness sections were
   created. For evaluation of the degree of granulation tissue formation,
   the sections were subjected to Masson’s trichrome staining. For
   detection of epidermal regeneration, the sections were subjected to H&E
   staining as standard protocols.

   For immunofluorescent analysis, frozen sections (7 μm) were exposed to
   primary antibodies against CD31 (BD, 557355, 1:1000), F4/80 (Abcam,
   ab6640,1:500) or Ki67 (Servicebio, [70]GB111499, 1:400), followed by
   the secondary antibody treatment (Alexa 594-conjugated antibody, Thermo
   Fisher, A-11007/ A-11037, 1:500; Alexa 488-conjugated secondary
   antibody, Thermo Fisher, A-11029/ A-11006, 1:500). DAPI staining was
   conducted. ImageJ software was used to quantify the staining. At least
   three sections were used for the quantification.

Quantitative real-time polymerase chain reaction(qRT-PCR)

   Toal RNA extraction was conducted using TRIzol [[71]24]. For mRNA
   detection, RNA reverse-transcription was conducted using a reverse
   transcription kit (Vazyme, R323-01) and qRT-PCR was conducted using
   SYBR Green (Vazyme, Q711-02). For miRNA detection, RNA was
   reverse-transcribed and quantified using miRNA SYBR qPCR Master Mix
   (Vazyme, MQ101-02). Relative mRNA or miRNA expression was normalized
   with Gapdh (mRNA) or U6 (miRNA), respectively. The relative expressions
   were calculated using the 2^−∆∆Ct method. Primer sequences used were as
   follows: Hif1α_F, ACCTTCATCGGAAACTCCAAAG, Hif1α_R,
   CTGTTAGGCTGGGAAAAGTTAGG; Vegfα_F, TAGAGTACATCTTCAAGCCG, Vegfα_R,
   TCTTTCTTTGGTCTGCATTC; Gapdh_F, CATGTTTGTGATGGGTGTGA, Gapdh_R,
   AATGCCAAAGTTGTCATGGA; miR-21-5p_F: GCAGTAGCTTATCAGACTGATG, miR-21-5p_R,
   GGTCCAGTTTTTTTTTTTTTTCAAC; U6_F, TGGAACGCTTCACGAATTTGCG, U6_R,
   GGAACGATACAGAGAAGATTAGC.

EV internalization assay

   The internalization of LGG-EVs was investigated in HUVEC and HaCaT
   cells. Briefly, Cy5 with red fluorescence was used to label LGG-EVs as
   previous protocol instructed [[72]25]. Then, the labelled EVs were
   administrated to cells. After 6 h of incubation, cells were treated
   with DAPI, and visualized with a confocal microscope.

Cell proliferation assay

   Ki67 staining was used to investigate the cell proliferation. Cells
   (5 × 10^4 cells/well for HUVEC and 8 × 10^4 cells/well for HaCaT) were
   seeded on the glass coverslips located in 12-well plates. After
   transfection with or without miR-21-5p inhibitor (miRi), cells were
   treated with a medium containing LGG-EVs (50 ng/µL) in 1% FBS for 24 h.
   Subsequently, the cells were sequentially fixed in 4% PFA,
   permeabilized with Immunostaining Permeabilization Buffer (Invitrogen,
   00-8333-56) and incubated with 100% goat serum (BOSTER, AR0009) to
   block reactive sites. Subsequently, cells were treated with primary
   antibody against Ki67 (Servicebio, [73]GB111499, 1:400) at 4°C. After
   12 h of incubation, Alexa 488-conjugated antibody (ThermoFisher,
   A-11008, 1:500) was added to cells at a temperature for 1 h. DAPI
   staining was conducted. The images were observed by fluorescence
   microscope (KEYENCE, BZ-X810). ImageJ software was used to quantify the
   staining. Three independent replicates were performed to examine the
   Ki67-positive cells.

Scratch wound-healing assay

   In 48-well plates, cells (4 × 10^4 cells/well) were plated. After
   serum-free media starvation of 24 h, cells were scratched using a 200
   µL pipette tip and washed with PBS to remove the dead cells. HUVEC and
   HaCaT cells were treated with a medium containing LGG-EVs (50 ng/µL) in
   1% FBS for 24 h or 48 h, respectively, and then photographed. The
   wounded area was measured with ImageJ. Each was conducted in
   triplicate. The migration ratio was calculated as migration area (%) =
   (Initial wound area-remaining area at observation point)/ Initial wound
   area× 100.

Transwell assay

   HUVEC was suspended in starvation medium and plated into the apical
   chamber of 24-well plates (8.0 μm pore filters, Corning, 3464) at a
   density of 4 × 10^4 cells/well. Meanwhile, 500 µL culture medium
   containing 10%FBS with or without LGG-EVs (50 ng/µL) was added to the
   basolateral chamber. After 24 h, cells on the lower side of the filter
   membrane were fixed, and stained with crystal violet (Sangon Biotech,
   E607309). The migrating cells were observed under an optical microscope
   (Nikon, Japan) and analyzed with ImageJ.

Tube formation assay

   In 96-well plates, Matrigel (50 µL per well, Corning, 354230) was
   added. After polymerization, HUVECs were overlaid on Martigel-coated
   wells at a density of 4 × 10^4 cells/well, followed by exposure to
   LGG-EVs (50 ng/µL) or PBS for 6 h. Tube formation was visualized with
   microscopy (Nikon, Eclipse Ti2E). ImageJ software was used to quantify
   the tube meshes.

miRNA sequencing and analysis

   The high-throughput sequencing and analysis of miRNAs in LGG-EVs was
   conducted as previously described [[74]25]. TruSeq Small RNA Sample
   Prep Kits (Illumina, San Diego, USA) were used to develop the miRNA
   sequencing library. Kyoto Encyclopedia of Genes and Genomes (KEGG) and
   Gene Ontology (GO) analysis were conducted using R package.

Inhibitor transfection

   miR-21-5p inhibitor and corresponding control were purchased from
   RiboBio (Guangzhou, China). Cells were transfected with miR-21-5p
   inhibitor (100 nmol/L) or the relevant negative control. The
   transfection was conducted in HUVEC and HaCaT cells using Lipofectamine
   3000 (Invitrogen, L3000008) as per the recommendations. The downstream
   experiments were performed after twenty-four hours of transfection.

Western blot

   Proteins were prepared with radioimmunoprecipitation (RIPA) buffer
   (Beyotime, P0013K) containing PMSF (Beyotime, ST506) and phosphatase
   inhibitor cocktails. BCA Kits (Beyotime, P0010) were used to quantify
   proteins. Western blot was performed as described previously [[75]26].
   ImageJ was employed to analyze the signal intensity (n = 3 independent
   replicas). The protein expressions were normalized to the GAPDH. The
   primary antibodies included p-AKT (CST, 9271, 1:1000), AKT (CST,
   4691,1:1000), HIF1α (NOVUS, NB100-449, 1:2000), and GAPDH (Proteintech,
   60004,1:50000).

Statistical analysis

   All experimental data were expressed as mean ± s.d. from at least three
   independent replicas. We used Graph Pad Prism software for data
   analysis. Statistical Significance was determined using one-way ANOVA
   or Student’s t-test. P values of less than 0.05 were considered
   significant differences (*P < 0.05; **P < 0.01; ***P < 0.001; #
   P > 0.05).

Results

LGG-EV isolation and characterization

   LGG-EVs were isolated from the culture supernatants of LGG
   (Fig. [76]1A). TEM analysis revealed that LGG-EVs exhibited a typical
   sphere-shaped feature (Fig. [77]1B), consistent with previously
   described EVs from mammalian cells [[78]27]. NTA analysis showed that
   the particle size of LGG-EVs ranged from 30 to 200 nm in diameter, with
   the main peak observed around 180 nm (Fig. [79]1C). However, unlike
   mammalian-derived EVs, specific marker for BEV characterization is
   still lacking [[80]5, [81]28]. Zeta potential analysis is a common
   method to evaluate the stability of EVs, and the zeta potential value
   around or less than − 20 mV indicates good stability [[82]29]. As shown
   in Fig. [83]1D, the zeta potential of LGG-EVs was about − 19.4 mV,
   indicative of their high stability. These results collectively indicate
   that LGG-EVs possess a similar structure and size to EVs from other
   mammalian sources.

Fig. 1.

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

   LGG-EV isolation and characterization. A Diagram describing the
   isolation of LGG-EVs. B Representative TEM images of LGG-EVs. Scale bar
   = 100 nm. C NTA to monitor the size of LGG-EVs. D Quantification of the
   zeta potential of LGG-EVs (n = 3). TEM, transmission electron
   microscopy; NTA, nanoparticle tracking analysis

In vivo safety of LGG-EVs

   We then assessed the in vivo safety of LGG-EVs through subcutaneous
   injection into mice once a day for 14 days. H&E staining results
   revealed that compared with Control, there were no obvious pathological
   changes in major organs, including heart, kidney, liver, lung, and
   spleen, after LGG-EV treatment (Supplementary Fig. [86]1), supporting
   the notion that LGG-EVs are well-tolerated. This observation aligns
   with reports of favorable safety profiles for LGG-EVs and EVs from
   other bacteria [[87]25, [88]30]. These findings provide both a
   theoretical and experimental foundation for the clinical application of
   LGG-EVs.

LGG-EV application accelerates cutaneous wound healing

   The retention and internalization are crucial for EV functions. To
   determine the contributions of LGG-EVs to wound healing, we firstly
   examined the retention and internalization of LGG-EVs in skin tissues
   after wound administration. LGG-EVs were labeled with Cy5, a red
   fluorescent dye [[89]25], and topically injected into the tissues
   around the wounds. Fluorescence microscopy analysis revealed that
   abundant Cy5-labeled LGG-EVs were detected and maintained in the
   perinuclear region of both skin epidermal keratinocytes and dermal
   cells at 24 h and 48 h after injection (Supplementary Fig. [90]2A). We
   further detected the capability of key dermal cell types to internalize
   LGG-EVs. To this end, we evaluated the co-localization of endothelial
   cells marked by CD31 and macrophages marked by F4/80 with Cy5 using
   immunostaining. We found that numerous CD31^+ or F4/80^ + cells were
   also with Cy5 signals, indicating that both cell types can internalize
   LGG-EVs (Supplementary Fig. [91]2B). These results confirmed the
   internalization of the LGG-EVs by resident wound cells and their
   retention in vivo.

   Subsequently, we investigated the impact of LGG-EVs on wound healing.
   Full-thickness skin wounds (6 mm punch) were inflicted in the back skin
   of C57BL/6 wildtype mice. LGG-EVs or Control vehicle (PBS) were
   subcutaneously injected into the edges of wounds once a day
   (Fig. [92]2A). The wound-healing kinetics were visualized and
   quantified at corresponding day 4 and day 7 post-injury. Macroscopical
   analysis of digital photographs showed that LGG-EVs treated wounds were
   smaller than in the Control group at day 7 (Fig. [93]2B-D). Wound
   tissues were collected at day 4 and 7 post-injury, which represents the
   inflammatory or proliferative phases of wound healing [[94]31].
   H&E-stained analysis revealed shorter wound widths at days 4 and 7
   post-wounding after LGG-EV treatment (Fig. [95]2E and F). As expected,
   the LGG-EV treatment group showed a longer hyperproliferative
   epithelial tongue on day 4 (Fig. [96]2E and G), indicative of
   accelerated re-epithelialization. Given the nearly enclosed epidermis
   in the LGG-EV group, we did not count the epithelial tongue on day 7.
   In addition to epidermal re-epithelialization, granulation tissue
   formation and ECM deposition are necessary for timely wound closure
   [[97]32]. LGG-EVs treated wounds exhibited thicker granulation tissue
   (gt) on day 4 and day 7 (Fig. [98]2E and H). Collagen synthesis, a
   hallmark of granulation tissue formation, increases to fill the wound
   gap during the early phase of healing [[99]32]. Therefore, Masson
   staining was performed to test the effect of LGG-EVs on collagen
   content. Compared with Control wounds, LGG-EVs treated wounds had more
   collagen content on day 7 (Fig. [100]2I and Supplementary Fig. [101]3),
   indicative of the enhanced granulation tissue formation. Altogether,
   these data suggest that LGG-EVs could accelerate the wound healing
   process in vivo.

Fig. 2.

   [102]Fig. 2
   [103]Open in a new tab

   LGG-EV application accelerates cutaneous wound healing. A Schematic of
   full-thickness wounds in mouse skin (n = 5 mice per time point). Either
   LGG-EVs or PBS were subcutaneously injected into skin around wounds
   once a day. On day 4 and day 7 post-wounding, the wounds were
   collected. B Representative time-lapse photographs of wounds treated
   with PBS or LGG-EVs at indicated day post-wounding. Scale bar = 2 mm. C
   Representative wound edge traces were established on each condition
   according to wound photographs. D Quantification of the wound size at
   indicated days (n = 5). E H & E images of the wounds on day 4 and day 7
   (n = 5). The black dotted line indicates epithelial tongue. Scale bar
   = 500 μm. F-H Quantification of the scar widths (F) and granulation
   tissue (gt) thickness (H) on day 4 and day 7, as well as the length of
   the epithelial tongue (G) on day 4 (n = 5). I Representative Masson’s
   trichrome staining of wounds on day 7 (n = 5). Scale bar = 100 μm. he,
   hyperproliferative epithelium

LGG-EV application enhances wound cell proliferation and angiogenesis

   In light of the enhanced re-epithelialization and vascularized
   granulation tissue formation after LGG-EV treatment, we reasoned that
   LGG-EVs may play a role in wound cell proliferation and angiogenesis
   processes. To test this hypothesis, we performed immunofluorescence
   staining for Ki67, a marker of proliferative cells, as well as CD31 in
   wounds. Compared to Control wounds on day 7, LGG-EV treatment
   significantly increased Ki67^+ cells within epithelium as well as
   dermal compartments (Fig. [104]3A-C), which may explain the increased
   re-epithelialization rate and granulation tissue content
   (Fig. [105]2E-I). Meanwhile, LGG-EVs treated wounds displayed a much
   higher proportion of CD31 positive tissue, indicative of more blood
   vessels within the granulation tissue compared with the Control wounds
   on day 7 (Fig. [106]3D and E). Furthermore, a proportion of CD31
   positive cells from Control wounds displayed as singlets, indicative of
   the initial stage of angiogenesis. VEGF-A is one of the most important
   proangiogenic factors that promote the formation of new blood vessels
   [[107]33]. Therefore, we further determined the gene level of Vegfa in
   wound tissues on day 7 post-injury. As expected, Vegfa expression was
   remarkably increased in the LGG-EVs group (Fig. [108]3F).
   Hypoxia-inducible factor 1α (HIF1α), a transcription factor, plays a
   crucial role in VEGF expression and angiogenesis during wound healing
   [[109]34]. Given the enhanced angiogenesis and Vegfa expression, we
   next investigated the influence of LGG-EVs on the expression of HIF1α.
   qRT-PCR results showed that the LGG-EVs group exhibited markedly higher
   Hif1a expression than the Control group (Fig. [110]3F). Taken together,
   our findings suggest that LGG-EVs accelerated the wound healing
   program, both in terms of cell proliferation and angiogenesis.

Fig. 3.

   [111]Fig. 3
   [112]Open in a new tab

   LGG-EV application enhances wound cell proliferation and angiogenesis.
   A Immunofluorescent images of Ki67 (red staining) and DAPI (blue) in
   the wounds on day 7 (n = 5). Scale bar = 100 μm. B, C Quantification of
   Ki67 positive cells in the basal layer of epidermis (B), and dermis
   (C). D Immunofluorescent images of CD31 (red) and DAPI (blue) in the
   wounds on day 7 (n = 5). Scale bar = 100 μm. E Quantification of CD31
   positive area in the wounds. F RT-qPCR analysis of Hif1α and Vegfa in
   the wounds on day 7(n = 5). gt, granulation tissue

LGG-EVs enhance the proliferation and migration of endothelial cells and
keratinocytes

   Wound healing is an intricate process, which relates to the interplay
   of different cells, such as endothelial cells and keratinocytes
   [[113]32]. Given that endothelial cells and keratinocytes are key to
   angiogenesis and re-epithelization, respectively, we tested the effect
   of LGG-EVs on these cells. To this end, we first asked whether LGG-EVs
   could be taken by endothelial cells (HUVEC) and keratinocytes (HaCaT).
   LGG-EVs were labeled with Cy5, and subsequently added to both cells.
   The internalization of LGG-EVs was observed 6 h after treatment.
   Fluorescence microscopy analysis showed that red fluorescence-labeled
   EVs had been largely internalized by HUVEC (Fig. [114]4A) and HaCaT
   (Fig. [115]4B).

Fig. 4.

   [116]Fig. 4
   [117]Open in a new tab

   LGG-EVs enhance the proliferation and migration of endothelial cells
   and keratinocytes. A, B Internalization of LGG-EVs labeled with Cy5
   (red) by HUVEC (A) and HaCaT (B). Scale bar = 25 μm. C, D
   Representative images of Ki67 (green) immunofluorescence staining in
   HUVEC (C) and HaCaT (D) with or without LGG-EVs treatment for 24 h
   (n = 3). Scale bar = 50 μm. E, F Scratch wound-healing assay examining
   the migratory abilities of HUVEC (E) and HaCaT (F) after exposure to
   LGG-EVs or Control vehicle in starvation medium for 24 h and 48 h,
   respectively (n = 3). Scale bar = 200 μm. G Transwell assay examining
   the migratory abilities of HUVEC after exposure to LGG-EVs or control
   vehicle for 24 h. Quantification of the corresponding migration cells
   (right, n = 3). Scale bar = 200 μm. H Tube formation of HUVEC after
   exposure to LGG-EVs or control vehicle for 6 h (n = 3). Scale bar
   = 200 μm. Quantitative analysis of the number of meshes (right) using
   Image J

   In response to injury, endothelial cells and keratinocytes undergo a
   series of cellular activities, including proliferation and migration to
   support healing processes [[118]32]. Hence, we next investigated the
   effect of LGG-EVs on the functions of endothelial cells and
   keratinocytes via proliferation assay and mobilization assay. Ki67
   staining assay demonstrated that the proliferation of HUVEC
   (Fig. [119]4C) and HaCaT (Fig. [120]4D) was significantly increased
   after exposure to LGG-EVs. Analysis of the mobilization ability of
   cells using scratch assays and transwell assays displayed that compared
   with that of Control, LGG-EVs enhanced the closure of cell-free gaps of
   HUVEC (Fig. [121]4E) and HaCaT (Fig. [122]4F), and migration of HUVEC
   (Fig. [123]4G). Furthermore, we also examined the vessel-forming
   ability of HUVECs using tube formation assay. As shown in Fig. [124]4H,
   compared to the Control, LGG-EVs exhibited a significantly increased
   number of meshes, suggesting that LGG-EVs promoted angiogenesis. In
   conclusion, these results indicate that LGG-EVs can endow keratinocytes
   and endothelial cells with enhanced capabilities, including
   proliferation, migration, as well as vessel formation for endothelial
   cells.

Profiling of LGG-EV miRNAs

   Although EVs derived from other bacteria have been demonstrated to
   regulate the wound healing process [[125]13], their effective contents
   remain to be elucidated. The EV cargos consist of various miRNAs,
   proteins, lipids, and other nucleic acid compositions. Among them,
   miRNA is one of the most abundant, and many of the biological
   activities of EVs have been considered to be related to miRNAs
   [[126]35]. To investigate the key miRNAs in LGG-EVs contributing to
   wound healing, miRNA sequencing was conducted. miRNA identification was
   performed based on miRBase and the genome and EST database. A total of
   494 miRNAs have been identified, including 193 known miRNAs and 301
   novel miRNAs. Heatmap showed that many miRNAs related to cell
   proliferation, migration, and angiogenesis (miR-21-5P, miR-25-3p,
   miR-92a-3p etc.) were highly expressed in LGG-EVs (Fig. [127]5A). In
   addition, KEGG pathway enrichment analysis and GO analysis of these
   miRNAs were performed to verify the pathways and biological functions
   potentially involved. KEGG analysis displayed that miRNAs were engaged
   in signal transduction of PI3K-AKT, metabolic, HIF-1, and VEGF pathways
   (Fig. [128]5B), all of which were reported to be associated with wound
   closure [[129]36–[130]38]. In terms of biological process (BP), miRNAs
   were mainly involved in cell proliferation, cycle and migration,
   angiogenesis, cell differentiation, and extracellular matrix
   organization (Fig. [131]5C). As for cellular component (CC), miRNAs
   participated in regulating membrane composition, cytosol, nucleus, and
   cell junction (Fig. [132]5D). In molecular function (MF) analysis, they
   were associated with the binding of proteins, nucleotide, ATP, and
   cadherin, as well as kinase activity (Fig. [133]5E). These results were
   consistent with the pro-healing effect of LGG-EVs observed in vivo and
   in vitro.

Fig. 5.

   [134]Fig. 5
   [135]Open in a new tab

   Profiling of LGG-EV miRNAs. A Heatmap showing the top miRNAs in
   LGG-EVs. n = 3. B KEGG pathway enrichment analysis of the top miRNAs in
   LGG-EVs. C-E GO analysis of the top miRNAs of LGG-EVs in BP, CC, MF. F
   The expression of five key miRNAs in LGG-EVs. G The predicted key
   target gene network of the five key miRNAs in LGG-EVs. KEGG: Kyoto
   Encyclopedia of Genes and Genomes; BP, biological process; CC, cellular
   component; MF, molecular function

   Of importance, there are multiple miRNAs involved in cell
   proliferation, migration, and angiogenesis among the top 20 miRNAs,
   such as let-7i-5p [[136]39, [137]40], miR-21-5p [[138]41–[139]43],
   miR92a-3p [[140]44, [141]45], miR-122-5p [[142]46, [143]47], and
   miR182-5p [[144]48, [145]49] (Fig. [146]5F). Cytoscape was used to
   visualize the target genes and networks of these five miRNAs
   (Fig. [147]5G and Supplementary Fig. [148]4A). Of importance, the
   network revealed multiple target genes associated with energy
   metabolism, such as PDK4, SLC2A3, LAMP2 and CDK13 [[149]50–[150]53],
   indicative the relationship between LGG-EV targets with metabolism.
   Furthermore, GO and KEGG analysis of these five miRNAs were also
   performed. KEGG analysis also showed that these five miRNAs were
   enriched in PI3K-AKT, HIF-1, MAPK, and metabolic signaling pathways
   (Supplementary Fig. [151]4B), which is consistent with cytoscape
   analysis. GO analysis suggested that these five miRNAs contributed to
   positive regulation of transcription, protein and cell differentiation
   (Supplementary Fig. [152]4C), nucleus, an integral component of
   membrane, cytosol, and Golgi apparatus in the category of CC
   (Supplementary Fig. [153]4D), and the binding of protein, nucleotide,
   and ATP, as well as hydrolase activity and DNA transcription factor
   (Supplementary Fig. [154]4E). Collectively, the results revealed that
   the miRNA cargo in LGG-EVs has the potential to benefit wound healing.
   Of note, the miRNA database associated with BEVs is currently
   deficient, and deserves further investigation.

LGG-EVs shuttle miR-21-5p into endothelial cells and keratinocytes

   Having demonstrated the miRNA contents of LGG-EVs, we next investigated
   the functional miRNAs that mediate the pro-wound healing effects. Since
   miR-21-5p displayed the second highest expression level and had
   positive roles in regulating cell functions and angiogenesis
   [[155]41–[156]43], all of which give rise to the pro-healing
   properties, we thus selected miR-21-5p as a possible candidate miRNA.
   We first confirmed the transfer of miR-21-5p into target cells. HUVEC
   and HaCaT were administered with LGG-EVs for 6 h followed by qRT-PCR
   analysis. In both cells, the expression of miR-21-5p was markedly
   increased after exposure to LGG-EVs (Fig. [157]6A and B). These data
   demonstrated that miR-21-5p can be shuttled by LGG-EVs to recipient
   cells to regulate gene expression.

Fig. 6.

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

   LGG-EVs activate the functions of endothelial cells and keratinocytes
   via transferring miR-21-5p. A, B The expression of miR-21-5p in HUVEC
   (A) and HaCaT (B) after treatment with or without LGG-EVs for 6 h
   (n = 3). C, D Representative images of Ki67 (green) and DAPI (blue)
   immunofluorescence staining. HUVEC (C) and HaCaT (D) were
   pre-transfected with miRi for 24 h, followed by treatment with LGG-EVs
   or PBS (Control) for 24 h (n = 3). Scale bar = 50 μm. E-H Scratch
   wound-healing assay examining the migration. HUVEC (E) or HaCaT (G) was
   transfected with miRi for 24 h, followed by exposure to LGG-EVs or
   vehicle control for 24 h or 48 h, respectively (n = 3). Scale bar
   = 200 μm. F, H Quantification of cell migration ratio in E and G,
   respectively. miRi: miR-21-5p inhibitor

LGG-EVs activate the functions of endothelial cells and keratinocytes via
transferring miR-21-5p

   To investigate if the pro-healing effects of LGG-EVs were mediated by
   miR-21-5p, we transfected a miRi that specifically reduces miR-21-5p
   before LGG-EVs treatment in HUVEC and HaCaT. LGG-EVs promoted cell
   proliferation and this effect was abolished by miRi as judged by Ki67
   staining (Fig. [160]6C and D, Supplementary Fig. [161]5). Furthermore,
   after miR-21-5p inhibition, the pro-migratory effects of LGG-EVs were
   attenuated, though not completely abolished in both cells
   (Fig. [162]6E-H). Intriguingly, LGG-EVs supplemented with miRi still
   displayed a higher migration ratio than control group (Fig. [163]6E-H),
   indicating that EVs have complex contents and other cargos in EVs may
   also contribute to cell migration. These findings showed that the
   effects of LGG-EVs on endothelial cells and keratinocytes were, at
   least in part, mediated through miR-21-5p.

The activation of the AKT/HIF1α pathways is involved in response to LGG-EV
stimulation

   We next asked which signaling pathway was affected by LGG-EVs during
   wound healing. The wound healing process is associated with the
   activation of various signaling pathways, such as PI3K-AKT and HIF1α
   [[164]36, [165]37]. PI3K-AKT and HIF1α are the key pathways involved in
   keratinocyte proliferation and migration, as well as angiogenesis.
   Previous reports also suggested that miR-21-5p might modulate migration
   through AKT signaling by targeting PTEN, a known negative regulator of
   PI3K-AKT and HIF1α signaling pathways [[166]54, [167]55]. Combining
   with above KEGG analysis (Fig. [168]5B), we hence investigated the
   influence of LGG-EVs on these two pathways. As shown in Fig. [169]7,
   LGG-EVs remarkably increased the levels of phosphorylation of AKT
   (p-AKT), an indicator of PI3K-AKT activation, and HIF1α in HUVEC
   (Fig. [170]7A-C) and HaCaT (Fig. [171]7D-F). We also found that the
   miRi abolished the activation induced by LGG-EVs (Fig. [172]7A-F).
   Therefore, LGG-EVs enhanced the function of endothelial cells and
   keratinocytes by miR-21-5p mediated PI3K-AKT/HIF1α signaling
   activation.

Fig. 7.

   [173]Fig. 7
   [174]Open in a new tab

   The activation of the AKT/HIF1α pathways is involved in response to
   LGG-EV stimulation. A-F Western Blot analysis of HIF1α, AKT, and p-AKT
   levels. HUVEC (A-C) and HaCaT (D-F) were transfected with miRi for
   24 h, followed by exposure to LGG-EVs or PBS (Control) for 24 h
   (n = 3). miRi, miR-21-5p inhibitor

Discussion

   Despite considerable research efforts, timely wound healing or
   regeneration remains an area of highly unmet medical need that requires
   novel therapeutics. In this study, we investigated the therapeutic
   effects and potential mechanisms of EVs released by probiotic LGG on
   cutaneous wound healing. We provided evidence for the first time that
   LGG-EVs facilitated wound closure, improved pro-healing-associated
   biomarkers, including enhanced angiogenesis and re-epithelialization
   capacities. Furthermore, we also revealed that LGG-EVs can be
   internalized by endothelial cells and keratinocytes, activate the
   PI3K-AKT and HIF1α signaling pathways, and enhance cell proliferation
   and migration as well as angiogenesis properties by shuttling
   miR-21-5p.

   In recent years, the crosstalk between gut microbiota and skin has
   received significant attention from scientists and clinicians [[175]56,
   [176]57]. Particularly, the gut microbiome participates in regulating
   homeostasis and pathogenesis via “gut-skin-axis” [[177]57]. Clinical
   trials have shown promising results with improved skin conditions
   following oral and topical probiotic interventions [[178]57, [179]58],
   indicating the potential of probiotics in dermatology. EVs are lipid
   bilayer particles loaded with bioactive molecules that transfer
   biological signals between local or distant organs. Interestingly,
   emerging evidence supports that EVs from gut bacteria function as
   mediators, facilitating the transport of bioactive components into the
   skin, regulating various biological processes, and exhibiting
   therapeutic effects in the disease therapy field [[180]59]. For
   instance, Wu et al. showed that Lactobacillus druckerii-derived EVs can
   accelerate wound closure as well as inhibit scar formation [[181]13],
   conferring regenerative capability. Fuhrmann et al. demonstrated that
   EVs from Lactobacilli increased cell migration and inhibited
   inflammation in vitro, thus improving cutaneous wound healing
   [[182]60]. Lactobacillus plantarum-derived EVs protect against ischemic
   brain injury [[183]61]. Interestingly, in a very recent study, Kuhn et
   al.. embedded EVs from Lactobacillus plantarum and Lactobacillus casei,
   respectively, into hydrogel and furtherly demonstrated their
   anti-inflammatory effects in vitro and capability to accelerate wound
   healing in vivo [[184]62]. However, to our knowledge, no study has
   reported the effects and underlying mechanisms of LGG-EVs concerning
   wound healing. In our study, we found that LGG-EVs can promote
   cutaneous wound healing, as evidenced by faster wound closure, higher
   rates of re-epithelialization, and more vessel formation. Therefore, we
   propose that gut probiotics-derived EVs may be an innovative
   therapeutic strategy in regenerative medicine, deserving further
   investigation. Notably, several EVs have entered into clinical trials
   (ClinicalTrials.gov.), underscoring their significant potential.
   However, these EVs are typically derived from mammalian cells, which
   face challenges such as poor and slow yield, high economic investment,
   and medical ethics concerns. BEVs can overcome these limitations as
   they are easily accessible and produced by fed-batch fermentation
   protocols. Importantly, manipulation of bacterial EVs is feasible, and
   this work opens new possible areas of study, such as engineered LGG-EVs
   to enhance therapeutic effects.

   BEVs also encapsulate novel bioactive molecules to regulate recipient
   cells [[185]11]. Among the most abundant loads of EVs, miRNAs play an
   important role in regulating biological activities by transferring them
   into recipient cells. Our miRNA sequencing revealed that miRNAs
   associated with cell proliferation, migration, and angiogenesis were
   enriched in LGG-EVs (Fig. [186]5). In line with previously reported
   engineered LGG-EVs [[187]25], we observed that miR-21-5p was also
   enriched in naïve LGG-EVs, with the second-highest expression level in
   our assay. This supports the notion that physicochemical modifications
   of EVs did not significantly change their components [[188]63]. In
   addition, while this study predominantly concentrates on miRNAs, we
   recognize the potential impact of other LGG-EV cargos. Future
   investigations are needed to provide a comprehensive and in-depth
   analysis of the bioactive components encapsulated in LGG-EVs and their
   respective functions.

   miR-21-5p may be the key regulator in LGG-EVs associated with the
   improved functions of keratinocytes and endothelial cells
   [[189]41–[190]43]. LGG-EVs contained abundant miR-21-5p. After
   incubation of LGG-EVs for a short time (6 h), the expression of
   miR-21-5p was significantly increased in those two cells, indicating
   that miR-21-5p can be shuttled from LGG-EVs to recipient cells. By
   using specific inhibitor of miR-21-5p, we were able to identify the
   therapeutic effects, as observed that inhibition of miR-21-5p
   attenuated LGG-EVs effects on endothelial cells and keratinocytes,
   thereby highlighting the key role of miR-21-5p in LGG-EVs. miR-21-5p is
   a vascular-specific member of the family of miRNA precursors that
   mediates angiogenesis, invasion, and metastasis of tumors, including
   skin tumors [[191]64]. Intriguingly, miR-21 expression is increased
   after skin injury [[192]65], while the expression of miR-21-5p is lower
   in diabetic mouse skin wounds [[193]66], suggesting its potential role
   in chronic wounds. Therefore, further study is needed to investigate
   the role of miR-21-5p and LGG-EVs in diabetic wounds and impaired
   healing conditions. In addition, multiple studies also have
   investigated the beneficial effects of miR-21-5p on the functional
   properties of endothelial cells, fibroblasts, and tumor cells
   [[194]41–[195]43]. For example, miR-21-5p has been reported to promote
   the proliferation and migration of endothelial cells and fibroblasts in
   vitro and angiogenesis in tumor microenvironment [[196]41–[197]43]. In
   normal and keloid keratinocytes, miR-21-5p overexpression induced the
   functional activity of these keratinocytes [[198]54]. However, few
   studies have illustrated the functions of miR-21-5p in wound
   re-epithelialization and revascularization. Our findings further
   consolidated that miR-21-5p up-regulation may promote wound healing.

   Multiple signaling pathways have been involved in wound healing
   processes. KEGG analysis of key pathways associated with miRNAs in
   LGG-EVs demonstrated the enrichment in PI3K-AKT and HIF1α pathways,
   both of which are required for timely wound healing. Consistent with
   the finding, LGG-EVs upregulated HIF1α and p-AKT, indicating the
   activated PI3K-AKT and HIF1α signals and miR-21-5p inhibition
   attenuated the activation effect, which means that LGG-EVs regulated
   the functional properties, at least in part, by inducing PI3K-AKT and
   HIF1α activity via miR-21-5p.

Conclusion

   In summary, we demonstrate that probiotic LGG-derived EVs are enriched
   with miR-21-5p, which enhance the functions of endothelial cells and
   keratinocytes by metabolic signaling rewiring, and subsequently promote
   re-epithelization and angiogenesis, thereby accelerating cutaneous
   wound healing. Our findings support the intriguing hypothesis that
   probiotic EVs-mediated signals can promote and facilitate the
   restoration of tissue integrity after injury in various aspects. Future
   experiments will determine whether these EVs can facilitate wound
   healing in pathological conditions and ultimately expand the
   therapeutic benefits, potentially leading to their application in
   clinical settings. Moreover, further pre-clinical in vivo studies are
   necessary to elucidate the innovative mechanisms of the signal exchange
   between probiotics and skin, specifically focusing on gut
   microbiota-derived EVs.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [199]Supplementary Material 1^ (3.3MB, docx)

Acknowledgements