Abstract

Background

   Diabetic foot ulcers (DFU) are severe complications of diabetes, posing
   significant health and societal challenges. Accumulation of reactive
   oxygen species (ROS) and elevated glucose levels are primary factors
   affecting diabetic wound healing. Achieving effective treatment by
   reducing ROS alone is challenging, as high glucose levels continuously
   drive ROS production. The excellent glucose-consuming capacity of
   lactobacilli and the antioxidant function of hydrogen undoubtedly
   provide good therapeutic ideas. Herein, we combined probiotic
   Lactobacillus reuteri with acid-responsive hydrogen-producing
   nanoparticles to construct probiotic active gel LR&AB@CAH to enable a
   cascade of glucose consumption and hydrogen production. Lactobacillus
   reuteri consumed overproduced glucose and thereby released lactic acid
   to activate nanoparticle for hydrogen production, which could
   neutralize excess ROS and promote wound healing.

Results

   In vitro experiments demonstrate that LR&AB@CAH has good
   biocompatibility, antioxidant capacity. LR&AB@CAH reduces excess ROS,
   decreases oxidative substances, and boosts antioxidant enzyme activity.
   In a diabetic wound mouse model, it functions as a glucose scavenger
   and antioxidant, reducing ROS and supporting wound healing.

Conclusion

   LR&AB@CAH offers a novel strategy for the comprehensive treatment of
   DFU. This study provides an artificial-natural composite hydrogel for
   cascade therapy on diabetic wound healing, and suggests a complete
   management approach for diabetic oxidative stress.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-025-03115-5.

   Keywords: Diabetic wound, Hyperglycemia, Oxidative stress,
   Glucose-consuming and hydrogen release, Probiotic active gel

Graphical Abstract

   [48]graphic file with name 12951_2025_3115_Figa_HTML.jpg

   BRIEFS: In the treatment of diabetic ulcers, Lactobacillus reuteri
   encapsulated in LR&AB@CAH plays a cascade with AB@MSN. Lactobacillus
   reuteri depletes glucose from the wound to produce lactate, which
   reduces ROS production at the source. At the same time, the lactic acid
   produced can reduce the pH value of the wound, thereby stimulating the
   acid to produce hydrogen in response to hydrogen-producing particles,
   further removing ROS from the wound and promoting wound healing

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-025-03115-5.

Introduction

   Diabetes mellitus (DM) is a chronic metabolic disease with an extremely
   high prevalence, and associated disability and mortality worldwide
   [[49]1, [50]2]. Diabetic foot ulcer (DFU) is one of the most serious
   chronic complications of diabetes mellitus [[51]3], with up to 34% of
   diabetic patients developing foot ulcers [[52]4, [53]5] and being at
   risk of amputation [[54]6, [55]7]. With the prevalence of diabetes
   mellitus increasing every year, DFU has become a major threat to human
   life, health and well-being.

   The healing process of diabetic wounds is hampered by multiple factors
   such as persistent oxidative stress in the wound, high glucose levels,
   pathogenic bacterial infection [[56]8, [57]9] and inflammation. Current
   clinical standards of care for DFU include pressure off-loading, sharp
   debridement, and wound moisture balance, along with infection control
   and management of peripheral arterial disease [[58]10–[59]12]. However,
   these treatments are often fail to achieve satisfactory outcomes and
   associated with adverse effects [[60]13, [61]14]. Therefore, efforts to
   develop effective therapies for such complicated and chronic wounds are
   required.

   Normal wound healing goes through four main stages: haemostasis,
   inflammation, proliferation and maturation [[62]15]. Chronic wounds
   such as diabetic ulcers are typically trapped in the inflammatory phase
   of the healing process, with overexpression and accumulation of
   reactive oxygen species (ROS) at the wound site [[63]16]. Reactive
   oxygen species (ROS) are considered a double-edged sword. Normal ROS
   levels aid in wound healing by inhibiting microbial infection [[64]17,
   [65]18]. However, high ROS level may lead to damage normal cell
   function through interactions with biologically active molecules
   [[66]19, [67]20]. This in turn leads to the wound being in a state of
   persistent oxidative stress, resulting in a vicious cycle that prevents
   the wound from progressing towards healing [[68]21]. Notably, diabetic
   ulcer wounds exist in a localised hyperglycaemic environment, and high
   glucose further induces excessive intracellular ROS production and
   accumulation [[69]22], further elevating the level of oxidative stress
   at wound site [[70]23]. Conventional antioxidant approaches only focus
   on reducing ROS levels and are not effective in inhibiting the source
   for promoting ROS production. Therefore, combination therapy that
   targets both oxidative stress and high-glucose environments is
   necessary to improving diabetic wound healing.

   Synergistic antioxidant and glucose-lowering therapies offer a good way
   to treat diabetic wounds. Studies have been conducted to develop
   specific systems for hydrogen production from glucose consumption
   [[71]24]. This type of approach combines the advantages of antioxidant
   and hypoglycaemic drugs to compensate for the deficiencies of each in
   the clinical treatment of DFU, highlighting the potential and
   superiority of synergistic treatment and providing a promising strategy
   for the treatment of DFU with good prospects for clinical application.

   Probiotics have the potential to promote skin wound healing [[72]25].
   Chronic DFU wounds have a alkaline pH of 7–9, which is prone to
   bacterial infection [[73]26, [74]27]. Diverging from opportunistic
   pathogens that trigger immune responses resulting in tissue damage
   during infection, these probiotics exhibit some anti-inflammatory
   potential [[75]28, [76]29]. Lactobacillus reuteri, as a type of
   probiotic lactobacilli, exhibits the beneficial effects of decomposing
   glucose into lactic acid into for energy supply [[77]30, [78]31].
   China’s Ministry of Health approved Lactobacillus reuteri as a
   probiotic for health food in 2003. Studies have fully affirmed the role
   of Lactobacillus reuteri in promoting wound healing [[79]28]. When
   Lactobacillus reuteri colonizes DFU wounds, it is able to consume
   glucose in the wound to decrease local blood glucose level, and produce
   the metabolite lactic acid to lower wound pH value. The low pH
   microenvironment created by Lactobacillus reuteri can effectively
   inhibit the growth of harmful bacteria and maintain the wound
   micro-ecological balance.

   Hydrogen can exert selective antioxidant and anti-inflammatory
   functions by scavenging the most toxic hydroxyl and nitrite radicals of
   ROS species [[80]32, [81]33], which is remarkable for treating a wide
   range of inflammatory skin diseases, including improving chronic skin
   trauma, slowing skin aging [[82]34, [83]35]. However, simple H[2] or
   hydrogen enriched water under physiological conditions suffers from low
   water solubility, short residence time and complicated application,
   making it difficult to achieve ideal therapeutic effects [[84]36]. The
   use of solid hydrogen-producing materials for sustained H[2] therapy is
   a promising strategy for the treatment of diabetic wounds. AB@MSN
   nanoparticles, formed by the reaction of hydrogen precursor drug
   aminoborane (AB) with mesoporous silica (MSN), have an ultra-high
   hydrogen storage capacity and acid-responsive H[2] release behaviour
   [[85]37, [86]38]. This excellent property makes AB@MSN can be used as a
   hydrogen-producing antioxidant material, which can be used in
   combination with Lactobacillus reuteri in diabetic wounds to maximize
   the antioxidant effect of H[2].

   Combined polypharmacy has become a major trend in response to the
   pathology of diabetic foot ulcer wounds that are so complex and
   individually influenced [[87]39, [88]40]. The development of composite
   hydrogel dressings ensures a more optimized wound healing environment
   for effective combination therapy [[89]41–[90]44]. They not only
   provide a moist therapeutic environment [[91]45], which facilitates
   rapid wound healing, but also bring innovations to the treatment of
   diabetic ulcers by enhancing mechanical strength and stabilizing drug
   release mechanisms [[92]46, [93]47]. In this study, in order to ensure
   that Lactobacillus reuteri can target the wound and better produce
   acid-responsive cascade hydrogen production with AB@MSN, we selected
   calcium alginate hydrogel with excellent biocompatibility as the wound
   dressing and encapsulated both in the hydrogel. This method effectively
   combines the unique advantages of AB@MSN and Lactobacillus reuteri to
   obtain the artificial live bacteria gel LR&AB@CAH that integrates blood
   sugar reduction and ROS removal (Scheme [94]1). The LR&AB@CAH
   synthesized in this study aims to reduce the wound pH by consuming
   glucose through the growth of Lactobacillus reuteri on the DFU wound
   surface, producing lactic acid, and stimulating AB@MSN to produce H[2].
   Starting from both the source and product of ROS, it reduces the level
   of oxidative stress on the wound surface and promotes DFU healing.

Scheme 1.

   [95]Scheme 1
   [96]Open in a new tab

   Schematic representation of the preparation and DFU wound healing
   effect of artificial live bacterial gel LR&AB@CAH. A AB@MSN
   nanoparticles were successfully prepared by capillary pore adsorption
   of MSN towards AB, and then mixed with cultured Lactobacillus reuteri
   in calcium alginate gel to prepare an artificial live bacteria gel
   (LR&AB@CAH). B LR&AB@CAH exerted a cascade glucose-lowering and
   hydrogen-producing function in wound tissue. Once upon contacting wound
   tissue, Lactobacillus reuteri consumed high level of glucose and
   produced lactic acid to lower the wound pH value. Subsequently, the
   wound acid microenvironment activated AB@MSN to produce hydrogen, which
   decreased the ROS content in wound tissue. Finally, this
   comprehensively facilitated the healing of diabetic ulcer wounds
   through glucose-consuming and ROS clearance

Results and discussion

Morphology, size, and composition characterization of LR&AB@CAH

   MSN and AB@MSN nanoparticles were first prepared successfully using
   sol-gel and capillary adsorption methods, and their morphology, size
   and mesopore structure were characterised (Fig. [97]1). TEM and DLS
   data all indicate that the synthesized MSN and AB@MSN have considerably
   uniform size and good monodispersivity. Both nanoparticles are
   spherical in shape under TEM microscopy (Fig. [98]1A and Supplementary
   Fig. [99]1A), with the particle sizes of MSN being about 100 nm and
   AB@MSN about 125 nm, while the hydration diameters of MSN and AB@MSN
   are slightly larger (122 nm and 164 nm, respectively, Fig. [100]1D)
   owing to surface negative charge and hydrogen bonding between silanol
   (MSN IR peak at 966 cm^− 1 and AB@MSN IR peak at 927.6 cm^− 1 in
   Fig. [101]1E) and water. Zeta potential evaluates the physical
   stability of the particulate dispersion system, showing that both MSN
   and AB@MSN are negatively charged (Supplementary Fig. [102]1B).
   Compared with MSN, the absolute value of Zeta potential of AB@MSN is
   higher, indicating greater electrostatic repulsion between its
   particles and better physical stability after loading AB.

Fig. 1.

   [103]Fig. 1
   [104]Open in a new tab

   Characterization of probiotic active gel LR&AB@CAH. A Transmission
   electron microscopy image of AB@MSN. B Scanning electron microscopy
   image of Lactobacillus reuteri. C Morphology and scanning electron
   microscopy image of LR&AB@CAH. D DLS pattern of MSN and AB@MSN. E FTIR
   characterization of AB, MSN and AB@MSN. F Nitrogen
   adsorption–desorption isotherms of MSN and AB@MSN

   FTIR spectroscopy between 4000 cm^− 1 and 500 cm^− 1 shows that the
   characteristic peaks of AB and MSN overlap, indicating that AB is well
   encapsulated into MSN (Fig. [105]1E). The shoulder peak of AB@MSN at
   1070 cm^− 1 reflects the presence of MSN at a small percentage. Most
   importantly, all the characteristic peaks of AB are well maintained
   after loading by MSN, indicating that AB is stabilized within MSN in
   the absence of decomposition.

   Combining TEM images of MSN and AB@MSN reveals that both nanoparticles
   have a highly mesoporous structure due to the presence of MSN. This
   structure in MSN facilitates high drug loading and sustained release,
   and after encapsulating the hydrogen gas prodrug AB to construct the
   AB@MSN nanomedicine, the hydrogen generated is still able to be
   released in a stable and sustained manner. The BET results shows that
   the pore volume of AB@MSN become smaller compared to MSN, i.e., the
   mesopores of MSN have been blocked after AB loading, indicating that AB
   has been encapsulated successfully into the mesoporous channels of MSN
   (Fig. [106]1F and Supplementary Fig. [107]1C). Meanwhile, the nitrogen
   adsorption-desorption curves show that the lines after adsorption and
   desorption almost completely overlap for both MSN and AB@MSN particles,
   indicating that the nanoparticles have a good pore volume release
   effect, which is conducive to sustained drug release (Fig. [108]1F).

   As a representative strain of probiotic Lactobacillus, Lactobacillus
   reuteri is microscopically short rod-shaped, displaying slightly
   irregular, rounded-ended campylobacteria (Fig. [109]1B). LR&AB@CAH is
   prepared by cross-linking Lactobacillus reuteri (1 × 10^5 CFU/mL) with
   AB@MSN (2.5 mg/mL) at concentrations screened in subsequent
   experiments. Naked eye observation reveals that LR&AB@CAH is in the
   form of a flake gel (lower left corner of Fig. [110]1C). SEM microscopy
   observes a dense pore-like structure on the surface of LR&AB@CAH
   (Fig. [111]1C), while the presence of pores on the surface is found
   under high-resolution microscopy (Supplementary Fig. [112]2). The
   mechanical deformability of LR&AB@CAH is determined by a physical
   stretching method [[113]48]. As illustrated in Supplementary
   Fig. [114]3, the length of the gel prior to and subsequent to the
   stretching process exhibit a change from 35 ± 2 mm to 47 ± 2 mm, and
   its apparent morphology undergo a notable elongation, thereby
   indicating that LR&AB@CAH exhibit a certain degree of deformability.

Glucose-consuming acid-producing properties of Lactobacillus reuteri

   Probiotic lactic acid bacteria are named for their ability to ferment
   carbohydrates into lactic acid. As a type of Lactobacillus,
   Lactobacillus reuteri also has excellent glucose-consuming and
   acid-producing properties. The glucose and lactic acid content of the
   culture medium are quantified before and after the growth of
   Lactobacillus reuteri (Fig. [115]2A). The results of Fig. [116]2B
   reveals a significant reduction in glucose content in the medium after
   24 h with a concentration correlation between 1 × 10^3-1 × 10^5 CFU/mL.
   It is hypothesized that the concentration of Lactobacillus reuteri at
   1 × 10^6 CFU/mL and later is higher, resulting in a small “growth
   competition” between microflora, resulting in a slight decrease in
   glucose consumption. Following a period of 24 h, the overall
   fluctuation of glucose content between 24 and 48 h is not statistically
   significant. This is due to the logarithmic growth phase of
   Lactobacillus reuteri, which resulted in a reduction in glucose content
   due to the depletion of nutrients in the culture medium.

Fig. 2.

   [117]Fig. 2
   [118]Open in a new tab

   Glucose consumption and lactic acid production performance
   of Lactobacillus reuteri. A Schematic diagram of Lactobacillus reuteri
   cultivation, glucose consumption and lactic acid production. In culture
   medium, the amount of glucose (B), lactic acid (C), and pH (D) were
   detected after incubation with different concentrations of
   Lactobacillus reuteri for 24 and 48 h. Furthermore, LR&AB@CAH and
   LR@CAH were placed on a solid medium, and the amount of glucose (E) and
   lactic acid (F) were also measured after incubation for 24 and 48 h.
   **P < 0.01, ***P < 0.001, ****P < 0.0001, indicating statistical
   significance compared with control group

   Figure [119]2C illustrates a notable increase in lactic acid content
   over the 24-hour observation period. This finding is supported by the
   pH results (Fig. [120]2D), showing a significant decline due to lactic
   acid accumulation. Consequently, the solution became acidic. It is
   noteworthy that there is a general decline in lactate content at 48 h
   in comparison to that observed at 24 h. The underlying cause of this
   phenomenon is yet to be determined. One hypothesis suggests that the
   logarithmic growth phase of Lactobacillus reuteri, which occurs after
   24 h, coincides with a reduction in the nutrient content required for
   its growth. This could result in the emergence of alternative metabolic
   pathways for lactate. However, this hypothesis remains to be validated.
   Nevertheless, the reduction of lactic acid has a minimal impact on the
   solution acidity, and there is a negligible difference between the pH
   values at the two time points of 24 h and 48 h. The pH values at the
   two time points of 24 h and 48 h exhibit minimal variation. Upon
   analysis of the aforementioned experimental results, Lactobacillus
   reuteri demonstrates its exceptional capacity to consume glucose and
   produce acid. In light of the glucose and lactate content, pH changes,
   lactate metabolism, and logarithmic growth period of the bacterial
   colony in the microenvironment, it can be concluded that 1 × 10^5
   CFU/mL is the optimal concentration for bacterial encapsulation in
   subsequent experiments. This concentration can be applied to diabetic
   ulcer wounds in conjunction with the acid-responsive hydrogen-producing
   material AB@MSN.

The substance exchange capacity of LR&AB@CAH

   Calcium alginate hydrogels have the capacity to exchange substances.
   After encapsulation in calcium alginate hydrogel, Lactobacillus reuteri
   is still able to absorb glucose from the environment and convert it to
   lactic acid, exerting its ability to consume glucose and produce acid.
   As illustrated in Fig. [121]2E, the glucose concentration within the
   medium at 24 h and 48 h is reduced in comparison to that of the Control
   group (0 h solid medium). The concentration exhibits a gradual decline
   over time. The glucose content in LR@CAH and LR&AB@CAH exhibits a trend
   from absence to presence, followed by a gradual decrease, which
   reflected the process of glucose absorption and utilisation by
   Lactobacillus reuteri after it passed through the hydrogel membrane.
   Figure [122]2F shows that the lactic acid content in the medium is 0.06
   µmol/mL at 0 h. Over time, the lactic acid content in the medium with
   LR@CAH and LR&AB@CAH increases gradually as a result of the consumption
   of glucose by the growth of Lactobacillus reuteri. Due to the
   growth-promoting effect of hydrogen on Lactobacillus reuteri, the
   lactic acid yield of LR&AB@CAH was relatively high compared to LR@CAH.
   The findings demonstrate that Lactobacillus reuteri encapsulated in the
   gel retains the capacity to consume glucose and produce acid, and that
   small nutrients, such as lactic acid and glucose, can be exchanged with
   the external environment through the gel, which provides a theoretical
   basis for probiotic colonisation of wounds and reduction of
   microenvironmental pH.

The acid-responsive hydrogen release behaviour of AB@MSN

   The capacity of AB (Ammonia borane), a hydrogen precursor drug, to
   undergo decomposition in acidic environments, resulting in the
   generation of hydrogen, enables the implementation of
   probiotic-nanoparticle cascade applications. AB molecules have C3v
   symmetry, combining electron-rich NH[3] with electron-poor BH[3]
   [[123]38]. In the presence of acidic conditions, the B-N bond in the AB
   molecule undergoes a cleavage process, resulting in the formation of
   the intermediate BH[3] [[124]49]. The hydrogen atom in BH[3] becomes
   negatively charged, and the subsequent hydrolysis of BH[3] produces
   BO^2−, accompanied by the release of H[2]. The reaction formula for the
   release of hydrogen from AB in an acidic environment is as follows
   [[125]37]: NH[3]·BH[3] + H^+ + 3H[2]O = NH[4]^+ + B(OH)[3] + 3H[2].
   Upon loading AB onto MSN, the synthesised AB@MSN exhibited excellent
   acid-responsive hydrogen release (Fig. [126]3A). The real-time hydrogen
   release curve (Fig. [127]3B) illustrates that when the pH is 7.4, the
   hydrogen concentration per unit time increases gradually with time,
   although the overall rise is not significant, with a peak occurring
   within 6 h. The hydrogen release rate of the solution at pH 7.4
   increases with time (within 6 h), and the hydrogen release rate
   increases with time (within 6 h). Furthermore, when the solution pH is
   acidic, the hydrogen release rate increases with time (within 6 h),
   with a greater rise observed at lower pH values. It is notable that the
   hydrogen release rate of AB@MSN at different pH levels exhibits a
   gradual decline following the peak observed at 6 h. The lower the pH of
   the solution, the more pronounced the decline in the hydrogen release
   rate, which can be attributed to the stabilising effect of the hydrogen
   bond between MSN and AB, which enables a slow release. This slow
   hydrogen splitting behaviour facilitates the sustained release of H[2]
   and is more conducive to long-term hydrogen therapy. As shown in
   Fig. [128]3C, the total hydrogen release at pH = 5.1 is demonstrably
   higher than that observed at other pH values. These findings indicate
   that as the solution pH decreases, the rate of hydrogen release
   increases and the amount of hydrogen produced by AB@MSN at a given time
   also rises. It can be demonstrated that the quantity of hydrogen
   released from AB@MSN over time is proportional to the acidity of the
   environment.

Fig. 3.

   [129]Fig. 3
   [130]Open in a new tab

   The acid-reactive hydrogen release behavior of AB@MSN, and the in
   vitro antioxidant capacity of AB@MSN before and after loading calcium
   alginate. (A) Schematic representation of AB@MSN synthesis and
   acid-responsive hydrogen production. Hydrogen release rate (B) and
   total hydrogen release concentration curves (C) of AB@MSN incubated in
   different pH solutions for 48 h. In vitro scavenging of ABTS^+ by
   AB@MSN (D) as well as LR&AB@CAH (E) in acidic solutions (pH = 5.1). F
   Effect of AB@MSN with different encapsulation concentrations in
   LR&AB@CAH on the proliferation capacity of Lactobacillus reuteri.
   **P < 0.01, indicating statistical significance compared with control
   group

Antioxidant capacity of AB@MSN and LR&AB@CAH

   Hydrogen has been demonstrated to possess selective antioxidant
   capacity, and the excellent acid-responsive hydrogen release capacity
   of AB@MSN nanoparticles has been shown to confer potential antioxidant
   properties. To verify this, an ABTS radical scavenging experiment was
   conducted on AB@MSN nanoparticles in an acidic environment in vitro
   [[131]50]. As illustrated in Fig. [132]3D, AB@MSN nanoparticles exhibit
   a certain degree of free radical scavenging ability. In the
   experimental concentration range, an increase in the concentration of
   AB@MSN is accompanied by a corresponding enhancement in its scavenging
   ability of ABTS^+, thereby demonstrating a dose-dependent scavenging
   ability of ABTS^+.

   Upon encapsulation of AB@MSN in LR&AB@CAH, the latter is similarly
   endowed with antioxidant capacity and demonstrated the ability to
   scavenge ABTS^+ in vitro (Fig. [133]3E). However, in contrast to the
   AB@MSN nanoparticles, both LR&AB@CAH demonstrate a notable capacity for
   free radical scavenging within the experimental concentration range.
   The gel-like LR&AB@CAH has a certain “storage” and “slow release”
   effect on the production and release of H[2], which is evident from the
   scavenging ability of LR&AB@CAH on ABTS^+. Furthermore, the degree of
   scavenging does not change with the change of concentration.

   AB@MSN shows excellent hydrogen production and antioxidant capacity
   under the above different conditions, which provides experimental basis
   for the later treatment of diabetic ulcer combined with Lactobacillus
   reuteri.

In vitro biosafety and stability assessment of LR&AB@CAH

   Hydrogel materials are capable of performing certain physiological
   functions in the context of the treatment of skin wounds [[134]51]. In
   order to ascertain that the successfully prepared LR&AB@CAH can exert
   its therapeutic effect in vivo, its in vitro stability is initially
   determined. The specimens are immersed in PBS phosphate buffer for
   48 h, after which they are photographed at 24-hour intervals to observe
   the morphological changes of LR&AB@CAH. As shown in Supplementary
   Fig. [135]4A, LR&AB@CAH demonstrates the capacity to maintain its
   original morphology following immersion in PBS for 24 and 48 h,
   respectively, without any discernible morphological alterations. This
   indicates that LR&AB@CAH is capable of stabilising in an aqueous
   solution for a prolonged period. In order to guarantee the stability of
   Lactobacillus reuteri within the gel matrix, the culture is continued
   by the addition of LR&AB@CAH to MRS solid medium. The growth of
   Lactobacillus reuteri is observable to the unaided eye after 24 and
   48 h of incubation, respectively (Supplementary Fig. [136]4B). At 24 h,
   milky white starbursts are observed in LR&AB@CAH, with the starbursts
   increasing and enlarging significantly at 48 h. No milky white colonies
   are generated on the medium surrounding LR&AB@CAH, indicating that
   Lactobacillus reuteri is able to grow steadily and without escaping
   from LR&AB@CAH.

   Considering that Lactobacillus reuteri co-survives with calcium
   alginate gel as well as AB@MSN, in order to avoid toxicity of the
   materials on the growth of Lactobacillus reuteri, we test the
   bacteriostatic properties of Lactobacillus reuteri using calcium
   alginate gel loaded with different concentrations of AB@MSN. The
   results in Fig. [137]3F show that the growth rate of the strains is
   over 100% after different concentrations of the materials have been
   applied to Lactobacillus reuteri. This indicates that the materials are
   not toxic to Lactobacillus reuteri and do not inhibit the growth of the
   bacterial population within the experimental concentration range.

   In particular, the growth rate of Lactobacillus reuteri reached 109.09%
   at 2.5 mg/mL, and some growth-promoting effect appeared at this
   concentration, which differed significantly from the control
   (**P < 0.01). In addition, the activity and distribution of
   Lactobacillus reuteri in LR&AB@CAH is assessed using bacterial
   staining. Supplementary Fig. [138]5 visually confirms that
   Lactobacillus reuteri is uniformly dispersed within the hydrogel,
   indicating successful encapsulation and distribution.

Biosafety assessment of LR&AB@CAH at the cellular level

   The wound healing process involves the proliferative activity and
   function of various cell types such as epidermal cells, fibroblasts and
   endothelial cells. As normal subcutaneous tissue cells, HaCaT (human
   immortalised epidermal forming cells) and L929 (mouse fibroblasts) are
   often investigated in diabetic ulcers, so we chose them to carry out
   relevant experiments [[139]52]. Cytotoxicity is one of the key
   indicators of the clinical applicability of a given substance. Given
   that the gel material was too large to be used in 96-well plates, the
   cytotoxicity was determined using the extract. In this study, the
   effects of calcium alginate gels encapsulated with different
   concentrations of AB@MSN (AB@CAH) and Lactobacillus reuteri
   supernatants with different colony-forming units on the viability of
   L929 cells and HaCaT cells, respectively, were evaluated under
   controlled experimental conditions and a defined cell growth
   environment (Fig. [140]4). The analysis of the experimental results
   indicates that the growth rate of both cells is greater than 80%
   following the action of AB@CAH on HaCaT cells (Fig. [141]4A) and L929
   cells (Fig. [142]4B) within the range of experimental concentrations,
   and no obvious cytotoxic effect is observed. The secretions of
   Lactobacillus reuteri do not exhibit cytotoxicity against HaCaT cells
   and L929 cells, as demonstrated in Fig. [143]4C and D. At a
   concentration of 1 × 10^5 CFU/mL, the cell viability of HaCaT cells is
   104.48% at 24 h and 100.69% at 48 h. The cell viability of L929 cells
   is 103.82% at 24 h and 104.41% at 48 h. The cell survival of both cell
   types in the presence of the secretion produced by this colony-forming
   unit is greater than 100%. It is noteworthy that the cell growth rate
   at 48 h is slightly lower than that at 24 h. This may be attributed to
   the stimulation of cells by LR&AB@CAH with the extension of time.
   However, in general, LR&AB@CAH has a good biosafety profile.

Fig. 4.

   [144]Fig. 4
   [145]Open in a new tab

   Cytotoxicity assay of LR&AB@CAH. Effect of AB@CAH on the viability of
   HaCaT cells (A) and L929 cells (B). Effect of Lactobacillus reuteri
   secretion on the survival of HaCaT cells (C) and L929 cells (D)

   Furthermore, to investigate the cellular behaviour of wound healing,
   scratch experiments were conducted utilising fibroblasts to ascertain
   the impact of LR&AB@CAH on cell migration. A comparison of the relative
   migration areas of the various groups revealed that LR&AB@CAH exerted a
   pro-migratory effect on L929 cells (see Supplementary Fig. [146]6 and
   Supplementary Fig. [147]7). At 24 h, 0.62 mg/mL was significantly
   higher than the control group; at 72 h, 0.62 mg/mL was significantly
   higher than the control group; It is proposed that LR&AB@CAH may
   facilitate the migratory capacity of L929 cells within a specific
   concentration range.

   AIE-Mito-R01 is a tetraphenylethylene derivative with a positively
   charged pyridine salt structure, which has the cellular transmembrane
   capacity and good aggregation-induced luminescence properties, and is
   able to specifically label the mitochondrial structure of a variety of
   cells, and the intensity of its fluorescence produces a highly
   significant change, while the fluorescent probe that is not bound to
   the mitochondrion basically does not emit fluorescent signals. In this
   experiment, AIE-Mito-R01 fluorescent probe is employed to label the
   mitochondria of HaCaT cells and L929 cells following co-incubation with
   LR&AB@CAH. Figure [148]5A and Fig. [149]5B illustrate the microscopic
   fluorescence imaging of HaCaT cells and L929 cells following
   co-incubation, respectively. The fluorescence imaging data indicates
   that the mitochondria of both cell types exhibit fluorescence at each
   concentration, with the fluorescence intensity being comparable to that
   of the Control group. This suggests that the cells are in a healthy
   state of growth. Quantitative analysis of fluorescence in Fig. [150]5C
   and D reveales that HaCaT cells show stronger fluorescence at 2.5 mg/mL
   (*P < 0.05), and the intensity of intracellular fluorescence of HaCaT
   cells and L929 cells at the remaining concentrations is not
   significantly different from that of the Control group (P > 0.05),
   further demonstrating the biosafety of the materials.

Fig. 5.

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

   LR&AB@CAH ameliorated intracellular ROS generation in vitro.
   Fluorescence imaging on the mitochondrial of HaCaT cells (A) and L929
   cells (B) after treatment with different concentrations of LR&AB@CAH,
   scale bar is 50 μm. Quantification on the mitochondrial fluorescence
   intensity of HaCaT cells (C) and L929 cells (D). Intracellular ROS
   levels of LPS-induced inflammatory HaCaT cells (E) and L929 cells (F)
   after treatment with different concentrations of LR&AB@CAH. *P < 0.05,
   **P < 0.01, ***P < 0.001, ****P < 0.0001, indicating statistical
   significance compared with control group; ^#P < 0.05, ^##P < 0.01,
   ^###P < 0.001, ^####P < 0.0001, indicating statistical significance
   compared with normal group

Intracellular ROS scavenging capacity of LR&AB@CAH

   Changes in ROS levels in HaCaT cells (Fig. [153]5E) and L929 cells
   (Fig. [154]5F) are detected using the DCFH-DA fluorescent probe to
   evaluate the repairing effect of LR&AB@CAH on cellular oxidative
   damage. The DCFH-DA molecule itself is non-fluorescent and freely
   crosses the cell membrane. Once inside the cell, DCFH-DA is hydrolysed
   to DCFH by intracellular esterases. DCFH is highly hydrophilic and
   cannot penetrate the cell membrane, so it accumulates in large amounts
   inside the cell. Intracellular ROS can oxidise non-fluorescent DCFH to
   fluorescent DCF, which can be detected under a fluorescent enzyme
   marker, so the fluorescence intensity of DCF is proportional to the
   intracellular ROS level. After H[2]O[2] treatment, intracellular ROS
   levels are significantly increased in both cells, and fluorescence
   green under the microscope (Supplementary Fig. [155]8). The
   fluorescence intensity of HaCaT cells increased from 30.47 to 54.91 and
   that of L929 cells increased from 44.36 to 68.67, demonstrating that
   the two cell types were able to produce excess ROS under the
   stimulation of H[2]O[2]. Intracellular ROS levels in both cells
   decreased after treatment. The intracellular ROS levels of HaCaT cells
   remain high when treated with low concentrations. However, at
   2.5 mg/mL, the ROS level of HaCaT cells decreases to 34.66
   (**P < 0.01); and at 5 mg/mL, the ROS level of HaCaT cells decreases to
   38.12 (*P < 0.05), both of which are significantly different from the
   Control group and not significantly different from the Normal group
   (P > 0.05). Due to the differences between the different cell types,
   the intracellular ROS levels of L929 cells were all significantly
   reduced after treatment (****P < 0.0001). Therefore, at a certain
   concentration, LR&AB@CAH has a strong scavenging ability for ROS and a
   certain protective effect against ROS damage, with the concentrations
   of 2.5 mg/mL and 5 mg/mL being the best. Considering cost and
   therapeutic effect, LR&AB@CAH with AB@MSN at a concentration of
   2.5 mg/ml is selected for subsequent animal experiments.

Cascade glucose-consuming hydrogen-producing in LR&AB@CAH promote wound
healing in vivo

   The in vivo pro-healing therapeutic efficacy of LR&AB@CAH was evaluated
   on a full-thickness diabetic wound model (Fig. [156]6A). We randomly
   divided the mice into five groups: model control, CAH (calcium alginate
   hydrogel), LR@CAH (calcium alginate hydrogel containing Lactobacillus
   reuteri), AB@CAH (calcium alginate hydrogel containing AB@MSN) and
   LR&AB@CAH (calcium alginate hydrogel containing AB@MSN and
   Lactobacillus reuteri), and the wound-free diabetic mice served as the
   normal control group. Considering that the logarithmic growth phase of
   Lactobacillus reuteri is 24 h, in order to maintain the vitality of the
   bacteria, prevent bacterial overflow, and balance the growth of
   Lactobacillus reuteri in LR&AB@CAH and the hydrogen release rate of
   AB@MSN nanoparticles, we treated diabetic wounds with a frequency of
   changing LR&AB@CAH every two days. Except for the normal control group
   and the model control group, which were not treated, the remaining
   groups were treated by administering drugs to the wounds at a frequency
   of 2 days/times.

Fig. 6.

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

   In vivo skin diabetic wound repairing ability of LR&AB@CAH. A Schematic
   illustration on the establishment of diabetic wound mouse and treatment
   timeline. B Representative photos of the chronic wound in different
   treated groups (scar bar: 5 mm). C Wound contraction rates in the
   Control group, CAH group, LR@CAH group, AB@CAH group and LR&AB@CAH
   group (n = 3). D Wound healing traces in different groups. H&E (E) and
   Masson’s trichrome (F) staining of wound tissues (scar bar: 500 μm,
   40×). Note: C: inflammatory cells; E: epidermis; H: hair follicle; G:
   skin appendages. *P < 0.05, **P < 0.01, ***P < 0.001, indicating
   statistical significance compared with control group

   During the course of wound treatment, wound healing of the same mouse
   in different groups was observed, photographed and recorded on days 0,
   3, 7, 14, 18 and 24. Figure [159]6B shows a photograph of the wound
   site reflecting the macroscopic changes in the wounds of each group
   over the course of treatment. Figure [160]6C shows the rate of wound
   healing on different days. Figure [161]6D quantitatively reflects the
   process of wound closure on different days. Combining the
   representative pictures of the wounds in Fig. [162]6B and C, it can be
   seen that the wound closure rate of the LR&AB@CAH group was
   significantly higher than that of the other groups, indicating that
   LR&AB@CAH-mediated acid production by Lactobacillus reuteri glucose
   depletion can stimulate hydrogen production in response to the AB@MSN
   acid response to reduce the accumulation and production of ROS at the
   wound site, which plays an important role in accelerating the healing
   of diabetic wounds. In addition, the LR@CAH group also reflects a good
   wound closure rate, which is closely related to the continuous glucose
   depletion of Lactobacillus reuteri at the wound site. In contrast, the
   contribution of the CAH and AB@CAH groups is more limited. In the
   absence of lactobacilli to lower the pH of the microenvironment, the
   diabetic wound itself is in a slightly alkaline environment, limiting
   the play of the acid-responsive hydrogen-producing material AB@MSN, and
   thus the AB@CAH group plays a weak role in accelerating wound healing.

   Pathological tissue preparation and staining techniques are an
   important part of pathology, and the method of section staining can
   show different cell and tissue morphology, as well as certain chemical
   components within the cells and tissues, which in turn can be observed
   in different tissues, the morphology and structure of the cells, as
   well as changes in the levels of certain chemicals within the cells
   [[163]52]. Furthermore, the evolution of the microscopic structure of
   new skins during wound healing was investigated by using the
   hematoxylin-eosin (H&E) and Masson’s trichrome staining methods.
   Figure [164]6E shows H&E stained sections of tissue at the wound site
   on day 14 and day 24 after completion of wound modelling. Inflammatory
   infiltration in the peri-wound tissue is shown microscopically as
   blue-purple stained macrophages and neutrophils. Observation of H&E
   sections on day 14 shows incomplete tissue at the wound site in all
   groups, with poorly defined epidermal borders and a basic absence of
   skin at the wound site, particularly in the control group. In contrast,
   the CAH, LR@CAH and AB@CAH groups have strong skin tissue integrity
   compared to the control group, but show strong purple staining,
   inflammatory cell aggregation and severe inflammation. In the LR&AB@CAH
   group, although the epidermal borders are still indistinct, the
   inflammatory state is less severe and the dermal tissue has recovered
   to some extent. With continuous administration of the drug, the skin
   wounds in the mice essentially healed by day 24, while the control
   group still had some problems, such as the skin’s dermal tissue not
   being restored and hair follicles not being produced. It is worth
   noting that the epidermal tissue in the LR&AB@CAH group was relatively
   thinner and there was already a significant amount of hair follicle
   tissue present, and the epidermis had become intact and smooth compared
   to the other groups.

   Figure [165]6F shows the results of Masson’s staining of tissue
   sections at the wound site on day 14 and day 24. During the experiment,
   muscle fibres can be stained red with magenta and Reichhorn red, and
   collagen fibres can be stained green or blue, we can observe the degree
   of collagen densification of the subepidermal layer by the collagen
   tissue stained blue. On day 14, collagen tissue production was observed
   in all groups except the control group. The subcutaneous collagen
   tissue is more homogeneous and dense in the LR&AB@CAH group, with a
   large number of collagen fibres produced in the tissue and loosely
   distributed under the connective tissue of the wound crust, and
   relatively few collagen fibres in the other groups. On day 24, the
   epidermal structure of the control group is disorganised, with loose
   and disorganised neocollagenous structures arranged beneath the
   epithelial tissue. Compared to the CAH, LR@CAH and AB@CAH groups, the
   LR&AB@CAH group has more skin appendages, a clear hierarchy of
   epidermal structures and a more orderly and regular arrangement of
   collagen fibres under the epithelial tissue. Throughout the wound
   healing process it can be seen that the LR&AB@CAH group showed better
   healing results.

LR&AB@CAH in diabetic wounds depleted of glucose

   The level of glucose consumption in the wound bed of a diabetic ulcer
   is directly related to the degree to which the wound heals faster or
   slower, and LR&AB@CAH has demonstrated excellent glucose consumption
   capacity in the wound. Glucose levels in the traumatised skin are
   measured in three randomly selected mice from each group during the
   wound healing period. As shown in Fig. [166]7A, the glucose content of
   the control group is 48.1 µmol/g, and the glucose content of the
   remaining groups is significantly different from that of the control
   group (****P < 0.0001), where the CAH group and the AB@CAH group are
   due to the fact that the gel itself is hygroscopic and was able to
   absorb wound exudate. The two groups containing Lactobacillus reuteri,
   on the other hand, have lower glucose levels, indicating that
   Lactobacillus reuteri is able to grow stably in the trauma gel and
   significantly reduce glucose levels, greatly reducing the synthesis of
   ROS and providing a new therapeutic idea for the application of live
   bacteria to skin trauma.

Fig. 7.

   [167]Fig. 7
   [168]Open in a new tab

   Glucose levels and antioxidant indexes detected in diabetic wounds from
   treated mice. Glucose content (A), GSH-Px enzyme activity (B), SOD
   enzyme activity (C), CAT enzyme activity (D), MDA content (E), and MPO
   content (F) in the skin of each group after days 3 and 7 of treatment.
   *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, indicating
   statistical significance compared with control group; ^#P < 0.05,
   ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001, indicating statistical
   significance compared with normal group

Antioxidant index

   Superoxide dismutase (SOD), glutathione peroxidase (GSH-Px) and
   catalase (CAT) are the three main classes of antioxidant enzymes that
   are part of the body’s natural defence system against oxidative damage.
   When ROS-mediated injury or oxidative stress occurs, the defence system
   is activated and the three enzymes are strongly activated to defend
   against ROS.

   GSH-Px is an important peroxidation-degrading enzyme that is widely
   distributed in biological organisms, eliminating H[2]O[2] and organic
   hydroperoxides produced by the organism and reducing the destructive
   effects of ROS on the organism. The results showed that the LR&AB@CAH
   group significantly increased GSH-Px enzyme activity in the wounds of
   diabetic mice. On day 3, the enzyme activity in the LR&AB@CAH group is
   135.95 U/mg, significantly higher than the Control group’s activity of
   79.28 U/mg. On day 7, the enzyme activity in the LR&AB@CAH group is
   136.33 U/mg, significantly higher than the Control group’s activity of
   92.97 U/mg (Fig. [169]7B). The main function of SOD is to scavenge
   intracellular O[2]^•− and produce non-toxic O[2] or less toxic
   H[2]O[2].The experimental results showed that the LR&AB@CAH group
   significantly increased SOD enzyme activity in the wounds of diabetic
   mice.On day 3, the enzyme activity in the LR&AB@CAH group is 18.72
   U/mg, significantly higher than the Control group’s activity of 12.23
   U/mg. On day 7, the enzyme activity in the LR&AB@CAH group is 37.22
   U/mg, significantly higher than the Control group’s activity of 19.99
   U/mg (Fig. [170]7C). This may be due to the ability of LR&AB@CAH to
   effectively eliminate excess ROS and inhibit high glucose-stimulated
   ROS production during the early stages of wound healing. As a result,
   SODase degradation was reduced, resulting in higher SODase activity in
   the tissue.

   The main role of CAT is to catalyse the decomposition of H[2]O[2] into
   H[2]O and O[2], thereby scavenging H[2]O[2] from the body, protecting
   cells from the toxicity of H[2]O[2] and providing an antioxidant
   defence mechanism for the organism. On day 3, the enzyme activity in
   the LR&AB@CAH group is 4.54 U/mg, significantly higher than the Control
   group’s activity of 2.55 U/mg. On day 7, the enzyme activity in the
   LR&AB@CAH group is 5.09 U/mg, significantly higher than the Control
   group’s activity of 3.66 U/mg (Fig. [171]7D). The LR&AB@CAH group
   exhibited sustained and elevated CATase activity, indicating that the
   low H[2]O[2] levels at the wound site resulted in minimal CATase
   depletion and therefore CATase in the LR&AB@CAH group exhibited high
   enzyme activity.

   As intracellular SOD levels decrease, there is a corresponding increase
   in O[2]^•− levels and rapid accumulation of the lipid peroxidation
   product malondialdehyde (MDA). MDA levels can reflect the potential
   antioxidant capacity of the body, the rate and intensity of lipid
   peroxidation in the body, and also indirectly the degree of
   peroxidative damage to tissues. The rate of MDA accumulation can
   represent the total free radical scavenging capacity in the tissues,
   and the LR&AB@CAH group had lower MDA levels, suggesting that it
   prevents lipid peroxidation by attenuating oxidative stress at the
   wound site, resulting in lower MDA levels (Fig. [172]7E).

   Myeloperoxidase (MPO), also known as peroxidase, is found in the
   aniline blue granules of myeloid cells, mainly neutrophils and
   monocytes, and is a specific marker for myeloid cells.The primary
   function of MPO is to kill microorganisms in phagocytes, produce
   hypochlorite using hydrogen peroxide and chloride ions, and form
   oxidising free radicals. As the LR&AB@CAH group was able to reduce
   wound oxidative stress, resulting in reduced neutrophil aggregation and
   MPO content, this was consistent with the histological results of H&E
   staining, confirming that LR&AB@CAH reduced inflammatory infiltration
   at the wound site (Fig. [173]7F).

   Based on the above experimental results, it was found that LR&AB@CAH
   could significantly reduce ROS at the wound site, reduce the
   consumption of antioxidant enzymes, increase enzyme activity, and
   reduce the content of inflammatory markers, so as to promote wound
   healing.

Mechanism study of LR&AB@CAH in the treatment of diabetic ulcers

   Having established that LR&AB@CAH plays a role in high glucose and
   oxidative stress environments through the above experiments, to further
   explore its mechanism of action, we extracted mouse ulcerated skin
   tissue for transcriptomic sequencing to analyse the core genes and
   signalling pathways involved.

   Three samples are collected from the normal, control and LR&AB@CAH
   groups for transcriptome sequencing (Fig. [174]8A). Subsequently, Venn
   diagram of differentially expressed genes for two-by-two comparison of
   the three groups are obtained (Fig. [175]8B). A total of 3661
   differentially expressed genes are identified between the control and
   normal groups (Supplementary Fig. [176]9), with 1725 genes exhibiting
   increased expression and 1936 genes displaying decreased expression;
   and a total of 375 genes are differentially expressed between the
   LR&AB@CAH group and the control group (Fig. [177]8C), of which 291 are
   upregulated and 84 are downregulated. The 4036 differentially expressed
   genes resulting from these two-by-two comparisons are subjected to Venn
   diagram analysis. As a result, as shown in Fig. [178]8D, among the
   downregulated differentially expressed genes in the control group,
   LR&AB@CAH is able to upregulate 66 of these differentially expressed
   genes; Among the differentially regulated genes that are upregulated in
   the control group, LR&AB@CAH is able to downregulate 46 of these
   differentially regulated genes. This suggests that these 112
   differentially expressed genes are the core genes for the effect of
   LR&AB@CAH. LR&AB@CAH can regulate these 112 core genes to normalise the
   wounds of diabetic ulcer mice, leading to gradual healing.

Fig. 8.

   [179]Fig. 8
   [180]Open in a new tab

   Transcriptomics analysis of diabetic ulcerated skin tissues from
   treated mice. A: Schematic illustration of transcriptome sequencing; B:
   Venn diagram of differentially expressed genes; C: volcano diagram of
   differentially expressed genes (LR&AB@CAH vs. Control); D: Venn diagram
   analysis of transcriptomes; E: heat map of core gene clustering; F: GO
   analysis of core genes; G: KEGG pathway of core genes. In the figure,
   [181]S1, [182]S2 and [183]S3 are the LR&AB@CAH group; N1, N2 and N3 are
   the normal group; C1, C2 and C3 are the control group

   Clustering and enrichment analyses are also performed on each of the
   112 core genes to explore the targets and pathways involved in the role
   of LR&AB@CAH in the treatment of diabetic ulcers. The clustering
   heatmap results show that the core genes clustered well among the three
   groups (Fig. [184]8E). A subsequent GO functional enrichment analysis
   of the 112 core genes (Fig. [185]8F) reveals that the core genes are
   mainly enriched in biological processes, including cellular
   transmembrane transport and inflammatory response, as well as in
   cellular components such as cell membrane, extracellular matrix, and
   myofibrils. Additionally, the analysis identified molecular functions,
   including arachidonate-CoA ligase activity, cAMP response element
   binding, and other molecular functions. Based on the GO annotation
   classification, KEGG pathway enrichment analysis was performed, and
   Fig. [186]8G shows that the core gene functions were mainly enriched in
   oxidase-inflammation-related pathways, such as peroxisome
   proliferator-activated receptor (PPAR) pathway [[187]53], tumour
   necrosis factor (TNF) pathway [[188]54], and adipocytokine pathway
   [[189]55]. Upon investigation, it is easy to see that these biological
   functions and pathways are inextricably linked to inflammatory
   responses, oxidative stress, etc. and thus have a range of effects on
   skin repair. The above experimental results show that LR&AB@CAH can
   exert its healing role by regulating related signaling pathways and
   core genes, and its specific mechanism of action is left to be studied
   in the future. Generally, transcriptome sequencing has initially
   clarified some of the genes and signalling pathways involved in
   diabetic ulcer healing, and has also provided a certain theoretical and
   practical basis for other researchers to explore the mechanism of
   diabetic ulcer healing in depth in subsequent studies.

Conclusions

   Aiming at the pathological characteristics of diabetic ulcer wounds,
   this study successfully prepared an artificial live bacterial hydrogels
   LR&AB@CAH with cascading glucose-consuming hydrogen-producing capacity
   for targeted oxidative stress mitigation and microenvironmental
   improvement by combining Lactobacillus reuteri and acid-responsive
   hydrogen-producing particles AB@MSN, starting from the two aspects of
   the high-glucose microenvironment and oxidative stress. The core
   concept is to cut off the source of ROS production while consuming
   them, effectively reducing oxidative stress and promoting wound
   healing. The obtained results demonstrate that LR&AB@CAH, has a good
   biocompatibility, antioxidant capacity, and effectively reduces the
   excess ROS, decreases the production of oxidative substances, and
   increases the activity of antioxidant enzymes. In vivo experiments
   shows that LR&AB@CAH reduces the glucose concentration and consumes the
   ROS accumulation in diabetic wounds, thus promoting wound healing in
   diabetic mice by treating both symptom and pathogenesis. Research
   mechanism in the treatment of diabetic ulcers needs to be further
   investigated.

   Overall, novel gel materials offer improved drug efficacy and safety,
   as well as new precision therapeutic interventions for patients with
   diabetic foot ulcers. However, a number of technical and mechanistic
   challenges need to be overcome before widespread clinical application
   can be achieved. In the future, intensive research and technological
   innovation are expected to make this novel gel material an effective
   treatment for diabetic ulcers.

Experimental

Materials

   The strain of Lactobacillus reuteri was purchased from Henan
   Engineering Research Center of Industrial Microbiology (BeNa Culture
   Collection, BNCC192190). Sodium alginate (SA) and anhydrous calcium
   chloride (CaCl[2]) are obtained from Sinopharm Chemical Reagent Co.,
   Ltd. And other analytical grade chemicals from Sigma-Aldrich. In the
   experiments described in this article, the water used was purified
   using a Milli-Q water purification system (Millipore). The kits for
   quantiting antioxidant markers (SOD, CAT, MDA and GSH-PX, MPO) in
   tissue homogenates were purchased from Beijing Solarbio Science &
   Technology Co., Ltd (Beijing, China) and Nanjing Jiancheng
   Bioengineering Institute (Nanjing, China), respectively.

Synthesis of AB@MSN and LR&AB@CAH dressing

   MSN was prepared using the sol-gel method. 0.6 g of
   Cetyltrimethylammonium tosylate (CTAT) was dissolved in 40 mL of
   deionized water, and 0.16 g of triethanolamine (TEAH[3]) was added
   under continuous stirring at 80 °C for 30 min, yielding a clear
   solution. Next, 4.0 g of tetraethyl orthosilicate (TEOS) was added
   dropwise to the above solution with vigorous stirring for an additional
   4 h. The acquired white product was collected by centrifugation,
   washed, and refluxed in a solution of 160 mL ethanol containing 9 mL of
   37% HCl. Finally, the synthesized MSN was collected by centrifugation
   and gently washed with deionized water three times.

   AB was dissolved in an ethanol solution and then added to the MSN stock
   solution (2 mg/mL). The resulting mixture was shaken at 500 rpm for
   24 h, and then the product was collected by centrifugation and washed
   to obtain the AB@MSN.

   Calcium alginate gel forms through ion exchange, where Na^+ in sodium
   alginate is replaced by Ca^2+ from calcium chloride. The preparation of
   LR&AB@CAH dressing was achieved by cross-linking sodium alginate with
   calcium chloride through this principle. First, 10% calcium chloride
   solution (CaCl[2]) and 3% sodium alginate solution (SA) were prepared
   using deionized water. With the help of vortex oscillation and pulse
   ultrasound, AB@MSN and Lactobacillus reuteri were evenly dispersed in
   the sodium alginate solution (10 mL), with loading amounts of 2.5 mg/mL
   and 1 × 10^5 CFU/mL, respectively. 2 mL of the mixed solution was
   poured into the mold to form a thin layer, and calcium chloride
   solution was added from the edges to cross-link and form gels over
   1–3 min. As to treatment on the animal model, in order to fit the wound
   shape better, LR&AB@CAH was pressed into a circle with an 8 mm sterile
   biopsy punch, so that it could stay on the wound naturally.

Characterization of AB@MSN and LR&AB@CAH dressing

   MSN and AB@MSN were deposited on a carbon-coated copper grid and dried
   naturally at room temperature, the size and morphology were measured
   with a JEM-2100 F transmission electron microscope (TEM, JEOL, Ltd.,
   Japan). The particle size and zeta potential of the nanoparticles were
   characterized with a Nano-ZS 90 Nanosizer (Malvern Instruments Ltd.,
   Worcestershire, UK). The composition of AB, MSN and AB@MSN were
   characterized by Fourier transform infrared spectroscopy (FTIR, Bruker,
   Germany). FTIR analysis was performed in the 4000 –500 cm^− 1 range to
   scan and determine functional groups in MSN and AB@MSN synthesis. The
   pore size volume of MSN and AB@MSN was determined and calculated by
   nitrogen isothermal adsorption desorption method (Micromeritics ASAP
   2020). Observation of the surface morphology and size of Lactobacillus
   reuteri by scanning electron microscopy (SEM, HITACHI SU8100; energy
   spectrum: AZtecLive Ulive Maxa 100). The surface morphology of
   LR&AB@CAH dressing was characterized by scanning electron microscopy
   (SEM, Zeiss sigma300, energy spectrum: Oxford Xplore30, detected by
   Sci-go Instrument Testing Platform).

Detection of lactic acid generation and glucose depletion

   The L. reuteri strain was cultured in MRS Broth at 37℃ until it reached
   the logarithmic growth phase (24 h). Subsequently, the bacterial
   suspension was adjusted to a concentration of 1 × 10^8 CFU/mL for
   reserve (OD[600 nm]=0.1). According to experimental requirements, the
   bacterial suspension was diluted by a concentration gradient and
   incubated at 37℃ for 24/48 h (1 × 10^8, 1 × 10^7, 1 × 10^6, 1 × 10^5,
   1 × 10^4, and 1 × 10^3 CFU/mL). After centrifugation at 4000 rpm for 5
   min, the supernatant was filtered and analyzed using an lactic
   acid/glucose content detection kit (Abbkine Scientific Co., Ltd.,
   Wuhan, China) at time points of 0 h, 24 h, and 48 h, respectively,
   while the pH value was detected and recorded.

   MRS solid medium was used to simulate diabetic ulcer wounds: LR&AB@CAH
   was placed on MRS solid medium for a period of time, and 1 cm^3 of
   solid medium around it was taken to detect whether there was material
   exchange between LR&AB@CAH and solid medium, and to detect glucose
   depletion and lactic acid production through the kit.

Antioxidant activity of AB@MSN and LR&AB@CAH dressing: ABTS determination

   The ABTS free radical scavenging experiment was employed to determine
   the in vitro antioxidant activity of AB@MSN and LR&AB@CAH dressing. The
   sample to be measured was dissolved and diluted to different
   concentrations, and 40 µL of the sample to be tested was mixed with 160
   µL ABTS working solution and reacted for 10 min at room temperature.
   Absorbance is measured at 734 nm.

   The cleanup capability (%)= [1 − (Ai − Aj) / A0] × 100%.

   Where Ai represents the absorbance value of the solution after adding
   the sample, Aj represents the absorbance value of the sample solution,
   and A0 presents the absorbance value of the ABTS + solution.

Determination of AB@MSN hydrogen release rate

   Aqueous solutions of different pH values were prepared to determine the
   hydrogen release rate of AB@MSN particles under different pH conditions
   by using the microelectrode method: the pH value of the solution was
   maintained at 5.1, 6.2, and 7.4, respectively, by precise measurement
   with a pH meter. A certain concentration of AB@MSN hydrogen-producing
   particles was taken, dissolved thoroughly and homogeneously, and then
   added into the solution to avoid light and stirring, and the hydrogen
   release of AB@MSN was monitored in real time by a hydrogen probe, and
   the rate of hydrogen release was calculated from the hydrogen yield.

Toxicity test of AB@MSN against lactobacillus reuteri

   Different concentrations of AB@MSN solution were prepared and filtered
   through 0.22 μm filter membrane to test whether it has an inhibitory
   effect on Lactobacillus reuteri. Take a 96-well plate, respectively, in
   each well, add the same volume of Lactobacillus reuteri (1 × 10^6
   CFU/mL) and different concentrations of AB@MSN particles, at least five
   replicate wells in each group, and set up a blank control and a
   negative control, mix well and then static cultivation in the bacterial
   37℃ incubator, after 12–16 h of visual observation of the bacterial
   growth, and the enzyme standard instrument 570 nm wavelength detection
   of the absorbance value of each well.

   Bacterial growth rate (%)=(A1-A0)/(A-A0)×100%.

   Where A1 represents the absorbance value after adding AB@MSN particles
   and bacterial solution, A0 represents the absorbance value of blank
   medium, and A represents the absorbance value after adding bacterial
   solution.

Bacterial dispersion and escape assay

   In order to verify whether Lactobacillus reuteri can survive and
   disperse uniformly in calcium alginate gel, the bacteria were stained
   using live and dead bacterial dyes and later mixed in aqueous sodium
   alginate solution to prepare LR&AB@CAH and fluorescence changes were
   observed under inverted fluorescence microscope. If the bacteria
   survived, the presence of green fluorescence was observed and the
   bacteria were evenly dispersed in the field of view. In addition, the
   LR&AB@CAH materials were placed on MRS solid petri dishes and incubated
   at 37 °C in an incubator for 24–48 h. Naked eyes were used to observe
   whether the lactic acid bacteria escaped, and photographs were taken to
   record the results.

Cytotoxicity assay

   Extract preparation method: 2 mL of AB@MSN mother liquor (10 mg/mL) was
   centrifuged (10000 rpm/8 min) and washed 3 times with PBS to fully
   remove the anhydrous ethanol. After that, 200 µL PBS was used for
   resuspension and gradient dilution was carried out sequentially, and
   900 µL sodium alginate solution was added and mixed thoroughly to
   prepare a gel respectively, and the gel was immersed into 1 mL of DMEM
   medium for 24 h to obtain the leachate with different concentration
   gradients.

   The cytotoxicity of the materials was assessed using the MTT cell
   viability assay. HaCaT with L929 cells in good growth condition were
   passaged for three consecutive times, and the cells were counted and
   then inoculated in 96-well plates at a density of 8 × 10^4 cells/mL for
   overnight incubation. To each well, 100 µL of LR&AB@CAH extract
   containing different concentrations of AB@MSN and supernatant of
   different colony forming units were added to treat the cells for 24 h.
   To each well, 10 µL of MTT solution with a concentration of 5 mg/mL was
   added to each well to avoid the light, and the cells were incubated at
   37 °C for 4 h. The supernatant was discarded, and the purple crystals
   on the bottom were completely solved by the addition of 150 µL of DMSO
   to each well and the absorbance value of the wells was detected at
   570 nm. absorbance value of each well.

   Cell viability (%) = (OD experiment-OD blank)/(OD control-OD
   blank)×100%.

Intracellular clearance of ROS levels

   DCFH-DA fluorescent probe was used to detect the changes of
   intracellular ROS levels. HaCaT and L929 cells with good growth status
   were selected and inoculated in 96-well fluorescent plates at a density
   of 8 × 10^4 cells/mL, and wall-affixed by overnight incubation at 37℃.
   After discarding the supernatant, the experimental wells were added
   with 100 µL of the extract to continue the treatment for 24 h. The
   peroxidative environment was set up artificially by adding H[2]O[2] to
   the medium, and each well was induced by adding 100 µL of H[2]O[2] at a
   concentration of 0.5 mM for 30 min, except for the normal group which
   did not undergo any treatment. after the plates were washed with PBS
   for three times, all the wells were loaded with a final concentration
   of 4 µM of the DCFH-DA probe and the experimental procedures were After
   incubation at 37℃ for 20 min, the plates were washed 3 times to fully
   remove the DCFH-DA that had not entered the cells; the parameters of
   the luciferase marker were set to 488 nm for excitation light and
   525 nm for emission light, and the fluorescence intensity was detected
   in each well.

Cellular mitochondrial assay

   AIE-Mito-R01 is a tetraphenylethylene derivative with a positively
   charged pyridine salt structure. It has cellular transmembrane capacity
   and good aggregation-induced luminescence properties, which can
   specifically label the mitochondrial structure of a variety of cells,
   and its fluorescence intensity produces a highly visible change, while
   the fluorescent probe that is not bound to the mitochondrion basically
   does not emit a fluorescent signal, so there is no need for further
   cleaning after the co-incubation, and it can be directly imaged.
   Therefore, no further cleaning is required after co-incubation, and
   imaging can be performed directly. After drug incubation, 100 µL of
   working solution at a final concentration of 4 µM was added to incubate
   the wall-adherent cells for 15–30 min, and the whole process was
   carried out in a light-proof environment. The fluorescence intensity of
   each well was detected by fluorescence microscopy with the excitation
   light of 488 nm and the emission light of 665 nm. And the cell state of
   each well plate was observed under fluorescence microscope and
   photographed and recorded as soon as possible.

Diabetic ulcer mouse model building and treatment

   The protocol was approved by the ethical review of the Medical Animal
   Centre of Shandong Second Medical University (No. 2024SDL332). C57BL/6J
   male mice (7 weeks old, weighing 18–22 g) were purchased from Jinan
   Pengyue Laboratory Animal Breeding Co. Ltd. and were randomly grouped
   and housed in cages with 12 h of alternating light and darkness. The
   temperature and humidity were maintained at 25 ± 2 °C and 70-90%,
   respectively. The mice were given sufficient drinking water and food
   during rearing, and the bedding was kept dry. After one week of
   acclimatisation, STZ 50 mg/kg/d was injected intraperitoneally for 3
   consecutive days. After 1 week of stabilisation, blood was taken by the
   tail-tip blood sampling method, and the fasting blood glucose level of
   the mice was monitored by a blood glucose tester; if the blood glucose
   level was > 16.7 mM, the mouse was confirmed to be a diabetic mouse,
   and the modelling was successful. It is a common practice in the field
   of modelling to select males as subjects, as females have been found to
   have a lower success rate and may also have a higher mortality rate.

   After successful modelling of diabetic mice, the mice were
   anaesthetised using sodium pentobarbital at a concentration of 1%
   (50 mg/kg) by intraperitoneal injection and depilated. A total
   excisional skin wound through the meat disc was created on both sides
   of the exposed dorsal midline of the mice using a sterile 6 mm diameter
   biopsy punch, followed by treatment. The mice were randomly divided
   into 6 groups: Normal group, Control group, CAH group, LR@CAH group,
   AB@CAH group, LR&AB@CAH group, Each group of 20 animals and 5 animals
   per cage were kept. Except for the normal control group and the model
   control group which were left untreated, the remaining groups were
   treated by administering medication to the wounds at a frequency of 2
   days/times. The wound area was photographed using a camera at
   designated time points to record wound healing and changes. During
   treatment, mice in each group were randomly selected for euthanasia on
   days 0, 3, 7, 14, and 24, respectively, after modelling, and the skin
   at the wound site was collected, half of which was fixed in 4%
   paraformaldehyde for paraffin embedding, and subsequently stained with
   H&E and Masson’s trichrome staining and observed under a light
   microscope; and half of which was immediately placed in pre-cooled
   sterile centrifuge tubes, and the tissue homogenization was carried out
   in liquid nitrogen or on ice. Tissue homogenisation was carried out to
   detect the glucose content in the wound skin, as well as the enzymatic
   activity of antioxidant enzymes.

Transcriptomic sequencing

   Diabetic mice are randomly divided into three groups (Normal group,
   Control group, LR&AB@CAH group). After the wound was modeled and
   treated, 1 cm^2 of skin tissue around the wound was placed in a sterile
   enzyme-free cryopreservation tube and frozen in liquid nitrogen, and
   then moved to a cryogenic freezer for cryopreservation. During the
   experiment, three samples were randomly selected from each group for
   transcriptome sequencing. All RNA in the skin tissue was extracted and
   subjected to quality control. From the total RNA, mRNA was sorted and
   purified, mRNA was reverse transcribed to cDNA, an appropriate
   transcriptome library was selected and constructed, and finally the
   samples were put on the machine to sequence the transcriptome. FPKM
   values were obtained for all genes, and differentially expressed genes
   (DEGs) between different groups were identified using the DEGseq R
   package, and differentially expressed genes between the LR&AB@CAH group
   vs. control group and control group vs. normal group were screened
   (P-value or q-value < 0.05, and |Log2 (Fold Change)| > 1).
   Differentially expressed genes with Log2 (Fold Change) > 1 were
   identified as up-regulated genes, and those with Log2 (Fold Change) <
   -1 were identified as down-regulated genes, and the number of
   differentially expressed genes with up- and down-regulation was
   counted. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes
   (KEGG) enrichment analyses of the differentially expressed genes were
   performed using R language clusterProfiler software to evaluate the
   biological functions of the DEGs. In order to alleviate the unnecessary
   effects of different batches of mice on the experiment, external
   factors such as animal species, sex, age, feeding conditions and
   experimental methods during the experiment are guaranteed to be the
   same as those of the previous batch.

Statistical analysis

   Experimental data are expressed as mean ± standard deviation
   (mean ± SD), with each experiment independently repeated three times.
   One-way ANOVA and Student’s t-test were performed using GraphPad Prism
   8.0.2, with a P < 0.05 considered. GraphPad Prism 8.0.2 was used to
   graph the experimental results.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [190]Supplementary Material 1^ (3MB, pdf)

Author contributions

   Conceptualization: Y.S.W.; Formal analysis: L.H.S., J.N.L., F.Y.W.,
   Z.H.Z., Y.X.W., X.Y.D., D.Q.; Funding acquisition: Y.S.W., Y.Y.G.;
   Investigation: Y.S.W., L.H.S.; Methodology: L.H.S., J.N.L., F.Y.W.,
   Z.H.Z., Y.X.W., X.Y.D., D.Q., F.M.C.; Project administration: Y.S.W.,
   T.Y.S., D.S., Y.Y.G.; Supervision: Y.S.W., T.Y.S., C.G., Y.Y.G.;
   Validation: Y.S.W., L.H.S., J.N.L.; Roles/Writing - original draft: all
   the authors; Writing - review & editing: Y.S.W., D.Q., T.Y.S., D.S.,
   Y.Y.G., L.H.S. All the authors contributed to the initial writing and
   have read and approved the final manuscript.

Funding

   This work was supported by the National Natural Science Foundation of
   China (No.82202317), Natural Science Foundation of Shandong Province
   (No.ZR2022QC087 and No.ZR2022MH152) and the Colleges and Universities
   Youth Innovation Team Talent Induction Program of Shandong Province
   “Precision Drug Delivery and Diagnosis and Treatment Application
   Innovation Team”, and Young Scholars Program of kite city.

Data availability

   No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

   All animal experiments protocols were approved by the ethical review of
   the Medical Animal Centre of Shandong Second Medical University (No.
   2024SDL332).

Consent for publication

   All authors read and agreed to submit the manuscript.

Competing interests

   The authors declare no competing interests.

Footnotes

   Publisher’s note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

   Yingshuai Wang, Lihua Shi and Junna Lu contributed equally to this
   work.

Contributor Information

   Yuanyuan Gao, Email: yyg20062006@126.com.

   Cheng Gao, Email: chenggao@szu.edu.cn.

   Tongyi Sun, Email: tysun@sdsmu.edu.cn.

References