Microorganism microneedle micro-engine depth drug delivery – Nature Communications

Microorganisms have the potential to serve as a source of aerodynamic force

Enterobacter Aerogenes (E.A.) can ferment various sugars such as glucose, lactose, and sucrose, producing gas. It primarily inhabits the human and animal intestines, serving as a normal flora in the human body. E.A. exhibits a broad temperature adaptability range, thriving between 15–40 °C. In terms of safety and applicable conditions, E.A. holds potential for applications in the treatment of human diseases. Genetic evidence suggests that E.A. contains hydrogenase proteins (Hyds) that reversibly catalyze H2 release or absorption, including ferredoxin pyruvate redox reductase (PFOR) and NADH ferredoxin redox reductase (NFOR). The main hydrogen production route is to catalyze H2 production using NADH as an electron donor during glycolysis using NADH ferredoxin oxidoreductase (NFOR) (Fig. 2a)20. However, there is limited research attempting to harness the gas-producing capabilities of E.A. To ascertain whether E.A. has the potential to serve as a gas source for microorganism micro-engines, comprehensive research on its gas production components and capabilities is needed. Initially, gas samples produced by E.A. were collected and subjected to gas chromatography analysis. Chromatographic results (Fig. 2a, b, and Supplementary Fig. S1) revealed the useful components to be H2 (6.15%). To further understand E.A.‘s gas production mechanism, gas production variations were evaluated under different inoculum sizes, substrate concentrations, and cultivation times (Fig. 2c–e). The H2 content in the gas was detected, and the methylene blue was decolorized through an oxidation-reduction reaction, indirectly reflecting the differences in gas production. The outcomes of our investigation reveal a direct correlation between substrate concentration and cultivation duration with the observed increase in gas production volume, while the influence of inoculum levels on gas production appeared comparatively nominal. Therefore, when using E.A. as a gas-producing microorganism micro-engine, we primarily control the gas production quantity and rate through the substrate concentration, specifically the concentration of glucose.

Fig. 2: Microorganisms have the potential to serve as a source of aerodynamic force.
figure 2

a E.A. decomposes substrate glucose to produce H2. Created with BioRender.com/l25i234. b The results of gas chromatographic detection of the mixed gas produced by the E.A.. ce Changes in H2 production by the microorganism under different inoculation levels (c), substrate concentrations (d), and cultivation times (e). Mean ± SD of n = 3 independent samples. Two-sided Student’s t test. f The experimental diagram of different masses of agar blocks being pushed by the gas produced by E.A.. g Effect of E.A. produced gas on agar displacement distance of different masses. Mean ± SD of n = 3 independent samples. Two-sided Student’s t test. h Changes in gas production over time under different substrate concentrations. i Changes in gas production rate over time under different substrate concentrations. j Changes in maximum gas production under different substrate concentrations. k Changes in maximum gas production rate under different substrate concentrations. Mean ± SD of n = 20 independent samples. Two-sided Student’s t test. l The time to reach the maximum gas production and gas production rate. Mean ± SD of n = 3 independent samples.

H2 is a unique component in the gas produced by E.A., not only distinguishable from other air components but also present in a proportion (around 6%) sufficient to serve as an indicator for detection. Consequently, in subsequent experiments, H2 will be used as an indicator to infer the overall gas production situation. To quantitatively analyze the factors influencing gas production and identify the relatively optimal conditions for gas generation by E.A., we continuously measured gas production and relative maximum gas production rates over time using a hydrogen gas electrode under different substrate concentrations (Fig. 2h–l and Supplementary Fig. S2). The results demonstrate that the highest gas production volume and gas production rate were observed when the substrate glucose concentration was at 20 mg/mL. However, as the substrate concentration continued to increase, there was no statistically significant enhancement in gas production volume and rate; instead, a decline was noted. This phenomenon is postulated to arise from the inhibitory effect of a high glucose environment on the proliferation of E.A., consequently influencing both gas production volume and rate. In a culture medium containing 30 mL of glucose substrate at a concentration of 20 mg/mL, E.A. could generate up to 28.7 mL of gas within a 12 h timeframe, with a maximum gas production rate reaching 1.65 mL/s. Moreover, the attainment of the peak gas production rate and volume occurred approximately within the initial 5 to 6 hours following inoculation. Thus, it can be deduced that the optimal conditions for E.A. gas production are a substrate concentration of 20 mg/mL, with the maximum gas production rate achieved around 5 to 6 h post-inoculation.

The gas production rate and volume of E.A. were found to be considerable. To demonstrate their potential as a microorganism micro-engine, we inoculated the bacteria at the bottom of agar-filled test tubes and sealed them for cultivation. We observed variations in agar displacement distances caused by gas propulsion of different masses (Fig. 2f, g, and Supplementary Fig. S4). The findings reveal that under conditions of substrate concentration at 20 mg/mL and a temperature of 37 °C, after a 12 h cultivation period, E.A. is capable of propelling agar with a maximum mass of 4 g, resulting in a maximum displacement distance of 12 mm. By means of a comprehensive analysis of force and motion, the calculated maximum propulsive force is approximately 0.04 N (where g = 9.8 m/s²). In conclusion, these data demonstrate that E.A., as a microorganism capable of producing gas through substrate decomposition, can potentially serve as a microorganism micro-engine for dynamic gas therapy.

Preparation and bacterial activity validation of microorganism micro-engine microneedles (MM-MNs)

The method of transdermal therapy using microneedles loaded with small-molecule drugs or biologics has become quite common. However, encapsulating active microorganisms into microneedles and attempting to harness the continuous production of gas from these active microorganisms21 to sustain their basic physiological processes poses significant technical challenges. To overcome this hurdle, we conducted a series of characterizations on the matrix composition selection for microneedles, the preparation process, and the validation of microbial activity. The ability of E.A. to serve as a source of gas propulsion has been verified. We mixed the bacterial suspension with the substrate glucose and the matrix material poly (ethylene glycol) diacrylate (PEGDA), which is used for fabricating the microneedles, and placed it into a mold. It was solidified using vacuum suction and ultraviolet (UV) irradiation (Fig. 3a), thus encapsulating the microorganisms within the microneedles.

Fig. 3: Preparation and bacterial activity validation of microorganism micro-engine microneedles (MM-MNs).
figure 3

aThe preparation process of MM-MNs. Created with BioRender.com/e35z291. b The effect of ultraviolet irradiation time on the activity of E.A.. Mean ± SD of n = 3 independent samples. c Digital microscopy characterization of microneedle morphology. Scale bar, 3 mm (Left); 2 mm (Middle); 500 μm (Right). d Mechanical force for the fracture of the microneedle tips. e Scanning electron microscopy characterization of microneedle morphology.Scale bar, 500 μm (Left); 300 μm (Right). f Fluorescence microscopy results of E.A. activity encapsulated in microneedles. Scale bar, 800 μm (Up); 200 μm (Down). g, h Confocal microscopy results of E.A. activity encapsulated in microneedles. Scale bar, 500 μm (g); Scale bar, 400 μm (h). A representative image of four biologically independent samples from each group is shown in c, e, f, g, and h.

To demonstrate the ability of E.A. to maintain viability and provide gas propulsion while mixing with the matrix material PEGDA, vacuum suction, and UV curing, we verified the viability of the microorganisms encapsulated in the microneedles. Firstly, we measured the effect of UV irradiation time on bacterial viability (Fig. 3b). The results showed that although bacterial viability gradually decreased with increasing UV irradiation time, the synthesis process of microorganism micro-engine microneedles (MM-MNs) only required ~5 s of UV irradiation, which had a negligible effect on bacterial viability.

By comparing the digital microscope images (Fig. 3c) and scanning electron microscope images (Fig. 3e) of the microorganism micro-engine microneedles (MM-MNs) and empty microneedles (MNs), it can be observed that the surface morphology of the microneedles remains unchanged before and after microbial encapsulation. The puncture force of the microneedle tips was measured using a universal mechanical tester (Fig. 3d). The results showed that the encapsulation of microorganisms did not cause significant changes in the puncture force, which reached approximately 0.6 N, well exceeding the force required for penetrating human skin. To validate the viability of microorganisms after encapsulation in the microneedles, we stained the MM-MNs with fluorescent dyes of Dry methylaluminoxane (DMAO) and Ethidium Homodimer III (EthD-III). The results of fluorescence microscopy (Fig. 3f and Supplementary Fig. S8) demonstrate that nearly all microorganisms maintained a high level of activity, with minimal occurrence of diseased cells. In addition, we tracked the activity of E.A. in microneedles over 48 h under room temperature (25 °C) and storage at 4 °C (Supplementary Figs. S9S11). The results showed that low temperatures could effectively extend the activity duration of E.A. in microneedles. However, even at room temperature (25 °C), E.A. maintained a considerable level of activity within the first 24 h.

We conducted layer-by-layer scanning of the loaded microneedles using confocal microscopy. In comparison to the positive control group subjected to sterilization treatment (Fig. 3g), the loaded microneedles exhibited a higher level of activity. In summary, the above data indicate that E.A. maintains a high activity during the preparation of microneedles. This provides the conditions for continuous gas production by active microorganisms in microneedles, forming the foundation for pneumatic motor microneedles.

Characterization of gas penetration depth of microorganism micro-engine microneedles (MM-MNs)

E.A. has been verified to maintain its activity within microneedles. However, for it to truly function as a pneumatic motor, the smooth release of gas from the microneedles is crucial. To quantitatively assess the gas production of MM-MNs and verify whether the duration and total amount of gas production can be controlled by adjusting the concentration of the loaded glucose substrate, we prepared microbial microneedles loaded with different concentrations of glucose: MM-MNs with 30 mg/mL glucose (MM-30G-MNs), MM-MNs with 20 mg/mL glucose (MM-20G-MNs), MM-MNs with 10 mg/mL glucose (MM-10G-MNs) and MM-0G-MNs. We selected one patch of microneedles and immersed them in simulated body fluid. The gas production was measured using an H2 electrode probe (Fig. 4a). The results showed that the microbial microneedles loaded with 30 mg/mL of glucose had the most significant gas production duration and amount (Fig. 4b). In subsequent experiments when investigating the drug delivery performance and mechanism of the microbial microneedles, we defaulted to using 30 mg/mL as the substrate concentration when not otherwise specified (MM-MNs).

Fig. 4: Characterization of gas penetration depth of microorganism micro-engine microneedles (MM-MNs).
figure 4

a Experimental diagram illustrating the detection of gas production from MM-MNs in simulated body fluids using a gas microelectrode system. Created with BioRender.com/e13d275. b Gas release from single microneedles loaded with different concentrations of glucose when immersed in simulated body fluid. c Test diagram of the penetration of MM-MNs through different skin thicknesses by using a transdermal test instrument. Created with BioRender.com/e13d275. d Trends in hydrogen gas concentration in simulated body fluid over time after penetrating different thicknesses of pig skin. Mean ± SD of n = 3 independent samples. e Experimental diagram of microelectrode real-time monitoring of subcutaneous gas concentrations at different depths. Created with BioRender.com/e13d275. f, g Trends in hydrogen gas concentration (f) and gas production rate (g) within the skin over time.

This indicates that E.A. not only maintains a high activity after encapsulation in the microneedles, but can still use the substrate retained in the microneedles and the surrounding to simulate glucose in body fluids to produce gas. To investigate the ability of gas produced by MM-MNs to penetrate the human epidermis effectively, we used H2 as a detection indicator. We explored the penetration depth of gas generated by MM-MNs in human tissues.

First, we inserted MM-MNs into pigskin of varying thicknesses. We used a transdermal testing apparatus to measure the gas concentration that passed through the skin in simulated body fluid, reflecting the gas permeability of MM-MNs (Fig. 4c). The results revealed a gradual reduction in the quantity of penetrated gas with increasing thickness of the porcine skin (Fig. 4d). Even within porcine skin of 5 mm thickness, commendable penetration capability persisted, as the increase in skin thickness did not lead to a substantial decrease in gas permeation. Subsequently, we selected porcine skin with a thickness of 5 mm and introduced MM-MNs into it to simulate the process of gas generation within the skin. Employing H2 microelectrodes, we monitored the temporal changes in H2 concentrations at different depths beneath the skin tissue surface (ranging from 100 μm to 1000 μm) and the corresponding alterations in gas production rates (Fig. 4e–g). The dissolved H2 concentration within the tissue indirectly reflected the overall mixed gas concentration within the tissue. The above data show that the gas generated by the loaded E.A. as a gas power source in the microneedles can penetrate the dermis to the depth of the skin tissue, and the deepest penetration can be more than 5 mm of the skin, providing a basis for its use as a power source to push drugs into the tissue.

Characterization of drug delivery performance of microorganism micro-engine microneedles (MM-MNs)

After verifying the performance of microbial pneumatic power, further validation is needed to determine whether double-layered microneedles can effectively propel the drug for enhanced release. This is crucial for assessing whether the drug’s penetration ability into deeper tissues has been improved22. Next, we proceeded to validate the functionality of MM-MNs in drug release. To begin with, we substituted the pharmaceutical agent with rhodamine dye and simulated the drug release process in vitro. The preparation protocol entailed blending rhodamine with the soluble microneedle matrix material, polyvinyl alcohol (PVA), and subsequently casting it into molds for solidification. Following this, insoluble matrix materials, poly (ethylene glycol) diacrylate (PEGDA), with and without microbial inclusion, were introduced and solidified as a bilayer core-shell structured microneedle (Fig. 5a, b). To accomplish this, two distinct variations of microneedles were formulated: rhodamine-loaded MM-MNs (RMM-MNs) and rhodamine-loaded MNs (R-MNs). Initially, these two types of microneedles were separately introduced into the simulated physiological fluid at 37 °C. Over time, the absorbance of the released rhodamine dye into the simulated physiological fluid environment was quantified using ultraviolet spectrophotometry (Fig. 5c). It is evident that microbial aero-dynamic forces significantly contribute to drug release, facilitating the expedited dissolution of the outer PVA shell and the accelerated release of the pharmaceutical agent.

Fig. 5: Characterization of drug delivery performance of microorganism micro-engine microneedles (MM-MNs).
figure 5

a Double-layer microneedles with an outer layer loaded with Rhodamine dye and an inner layer containing Poly(ethylene glycol) diacrylate (PEGDA) loaded with E.A.. Created with BioRender.com/e13d275. b Digital microscope image of RMM-MNS. Scale bar, 3 mm (Left); 500 μm (Right). c The release of rhodamine over time in simulated body fluid. Mean ± SD of n = 3 independent samples. d Confocal microscopy imaging of the penetration depth of Rhodamine dye released by microneedles loaded with different concentrations of glucose into the skin. Scale bar, 500 μm. e Fluorescence intensity of Rhodamine dye diffusion at different depths beneath the skin released by microneedles loaded with varying concentrations of glucose. Mean ± SD of n = 3 independent samples. Two-sided Student’s t test. f Double-layer microneedles with an outer layer loaded with Calcipotriol and an inner layer containing PEGDA loaded with E.A. Created with BioRender.com/e13d275. g Determination of the concentration of Carpotriol in the skin’s surface and subcutaneous tissue by high-performance liquid Chromatography (HPLC). Mean ± SD of n = 6 independent samples. Two-sided Student’s t test. h Changes in caprotriol concentration in skin surface and subcutaneous tissue over time as determined by HPLC. Mean ± SD of n = 6 independent samples. Two-sided Student’s t test. A representative image of four biologically independent samples from each group is shown in (b) and (d).

In order to verify whether the gas production and rate can be controlled indirectly by adjusting the concentration of glucose encapsulated in the microneedles to control the depth of the gas to push the dye through the skin, we conducted the following verification experiments. Four distinct variations of microneedles were formulated: rhodamine-loaded MM-MNs with 30 mg/mL glucose (RMM-30G-MNs), rhodamine-loaded MM-MNs with 20 mg/mL glucose (RMM-20G-MNs), rhodamine-loaded MM-MNs with 10 mg/mL glucose (RMM-10G-MNs) and rhodamine-loaded MNs (R-MNs). Subsequently, these four microneedle variants were inserted into equivalent porcine skin tissue. Following a 12 h incubation period, confocal microscopy was employed to observe the depth of dye penetration within the tissue (Fig. 5d, e). The results elucidate a notably enhanced depth of penetration of dye in the RMM-30G-MNs group, reaching up to about 900 μm beneath the skin, where gas-producing bacteria participated in the diffusion process. The aforementioned experiments employed dye to simulate drug release ex vivo. Quantitative analysis of fluorescence intensity at different subcutaneous depths in each group concluded that the impact of varying glucose substrate concentrations on drug diffusion depth and range is minimal within the 0-400 μm range (Fig. 5e). Significant differences only appear at depths greater than 500 μm, confirming that our designed microbial gas-engine-powered microneedles are controllable and can significantly increase drug delivery depth and range.

Moving forward, carpool (a clinical pharmaceutical agent used for psoriasis treatment) was employed in lieu of rhodamine dye within the microneedles (Fig. 5f). Consequently, two other sets of microneedles were generated: carpool-loaded MM-MNs (CMM-MNs) and carpool-loaded MNs (C-MNs). These two microneedle groups were inserted into the depilated dorsal region of BALB/c mice, and following a 12-hour period, the microneedles were retrieved, accompanied by the collection of dorsal skin and subcutaneous tissue samples. Subsequently, liquid chromatography was employed to ascertain the concentration of carpool within these samples (Fig. 5g, h and Supplementary Fig. S15). The outcomes distinctly demonstrate a heightened drug concentration within the subcutaneous tissue of the CMM-MNs group, thereby substantiating the capacity of these biomimetic microbial-aerodynamic microneedles to facilitate deep-tissue drug permeation within animal tissue. We also collected skin samples from mice in the CMM-MNs group at different time points to measure the drug retention in the epidermis and subcutaneous tissue. This verified that, over time, the drug gradually penetrated from the epidermis into the subcutaneous tissue along with the gas production process. It can also be concluded that the microbial microneedles we designed exhibit excellent transdermal efficiency and drug retention rates in vitro. The increase in penetration depth of microneedles drug delivery ensures that the drug is more effectively absorbed by the skin. This is particularly important for drugs requiring higher bioavailability, as it can reduce the required dosage and alleviate potential side effects. It provides a faster route for drug delivery, accelerating the onset of treatment, which is particularly beneficial in situations where rapid symptom relief is crucial.

The diffusion of drugs under the skin is controlled by the gas-producing behavior of microbial engines

Microbial microneedles powered by E.A. as the gas-producing engine have been proven to significantly improve transdermal drug delivery efficiency and increase the range and depth of drug diffusion. However, further verification is needed to understand how the micro-engine precisely controls the drug release behavior. To more accurately investigate the diffusion process and range of the drug within the skin, we set up an observation period of 24 h. During this period, we observed the time-dependent release of a dye simulating the drug subcutaneously and conducted a quantitative evaluation of the fluorescence intensity (Fig. 6a–c and Supplementary Fig. S16). Firstly, it is evident that in the shallower epidermal layers (0–400 μm), the drug can diffuse extensively within 1-2 h. This can be attributed to the larger volume at the base of the conical microneedles, which contain more gas-producing microorganisms. After 4 h, the fluorescence range in the deeper skin tissues (500–600 μm) begins to expand. Between 6 and 8 h, the drug can diffuse into deeper tissues (>700 μm). As the observation period extends to 24 h, the diffusion depth of the drug can further increase, likely due to the ongoing molecular diffusion and thermal motion of the drug or dye molecules. To verify that this drug diffusion behavior is closely related to E.A. gas production, we compared the time-dependent gas production curve of the microneedles with the diffusion curves of the dye at various subcutaneous depths (Fig. 6d). It is clear that as the total amount of gas increases during the gas production process, the drug begins to diffuse in the superficial layers of the skin at the same time. The diffusion of the drug in the deeper layers shows a delay of about 1-2 h as it gradually increases with subcutaneous depth, indicating that the gas-driven drug diffusion takes time. However, it can still be concluded that the gas production process of E.A. is highly correlated with drug diffusion.

Fig. 6: The diffusion of drugs under the skin is controlled by the gas-producing behavior of microbial engines.
figure 6

a Confocal microscopy imaging of the release and diffusion over time of dye-simulated drugs at different depths beneath the skin. Scale bar, 500 μm. b, c The average fluorescence intensity of dye-simulated drugs diffusing over time at different depths beneath the skin. d The curves show the average fluorescence intensity of drug diffusion over time at different depths, and the curves depict gas production by microneedles over time. e A schematic diagram illustrating the gradual dissolution and diffusion of dye-simulated drugs under the skin due to gas production by microneedles, penetrating into deeper tissues. Created with BioRender.com/u33l993. A representative image of four biologically independent samples from each group is shown in (a).

The microneedle matrix encapsulating E.A. is made of cured PEGDA, while the outer layer encapsulating the drug/dye is made of PVA. These two media have different densities. When the gas released by E.A. reaches the interface between the two media, it inevitably causes a change in pressure. Generally speaking, gas pressure is produced by the frequent and continuous collisions of a large number of molecules undergoing random thermal motion against the walls of the container. The impact of a single molecule colliding with the wall is brief, but the frequent collisions of many molecules with the wall create a continuous and uniform pressure. Therefore, from the perspective of molecular kinetic theory, gas pressure is the average force exerted by a large number of gas molecules on the unit area of the container wall.

Treatment of psoriasis using microorganism micro-engine microneedles (MM-MNs)

Among various clinical applications, we tested microbial pneumatics microneedles for transdermal delivery of calcipotriol to treat psoriasis. Psoriasis is a genetically and immunologically mediated inflammatory skin disease whose treatment primarily relies on symptom control23. Current treatment methods include local therapy and systemic therapy. Systemic therapy has many limitations, such as hepatic first-pass metabolism, hepatotoxicity, gastrointestinal disturbances, needle pain, and phobia, and it requires administration by healthcare professionals. On the other hand, local treatment for mild to moderate psoriasis faces challenges such as poor skin permeability, short drug retention time, greasy topical vehicles, and lack of controlled release.

Microneedles, as a novel transdermal delivery method, bypass the skin barrier and enhance direct drug penetration into the skin, addressing the limitations of traditional transdermal drug delivery systems. However, due to the limited length of the needle shaft, they cannot penetrate the thickened epidermis, thus reducing the transdermal diffusion efficiency of drugs. Therefore, we chose psoriasis as the disease model to evaluate the effectiveness of microbial biomimetic pneumatics microneedles in clinical therapy. We used the imiquimod (IMQ)-induced psoriasis mice model, which can reproduce typical clinical symptoms and pathological features of human psoriasis, making it a feasible animal model for studying psoriasis (Fig. 7a).

Fig. 7: The therapeutic efficacy of microorganism micro-engine microneedles (MM-MNs) for psoriasis treatment.
figure 7

a Establishment of the animal model and treatment procedure. Created with BioRender.com/u33l993. b Changes in PASI scores during the treatment process. Mean ± SD of n = 3 independent samples. ch ELISA experiment to measure the levels of psoriasis-related inflammatory factors. (c): IL-17A; (d): IL-17F; (e): IL-10; (f): IL-22; (g): IL-23; (h): CXCL1; (ch) mean ± SD of n = 3 independent samples. Two-sided Student’s t test. i Changes in the condition of the back of mice during the treatment process. A representative image of three biologically independent samples from each group is shown in (i).

We evaluated the progression of psoriatic lesions on the mice’s back, including modeling efficacy and treatment efficacy, using the Psoriasis Area and Severity Index (PASI) score based on four dimensions: erythema, scaling, thickness, and skin lesion severity. We established five experimental groups: the modeled untreated group (Control), the calcipotriol treatment group (Calcipotriol), the calcipotriol-loaded microneedles group (C-MNs), the calcipotriol-loaded MM-MNs group (CMM-MNs), and the MM-MNs group (MM-MNs).

Throughout the modeling and treatment process, we recorded the PASI scores of the mice back (Fig. 7b and Supplementary Figs. S17, S18) and observed the skin appearance of the back (Fig. 7i and Supplementary Fig. S19). We found that the CMM-MNs group had the best treatment efficacy and the fastest improvement in disease progression. In addition, we conducted tests on psoriasis-related inflammatory factors in each group, and the results (Fig. 7c–h) confirmed the relevant data regarding treatment efficacy, with the CMM-MNs group having the lowest levels of inflammatory factors. By observing skin tissue sections stained with hematoxylin and eosin (H&E) (Supplementary Fig. S20a), we observed a significant improvement in inflammatory cell infiltration and a reduction in skin epidermal thickening in the CMM-MNs group. Furthermore, the results of H&E-stained liver and kidney tissue sections, inflammatory factor data in the control group, and cytotoxicity tests showed that microbial pneumatics micro-needles did not exhibit significant hepatotoxicity, nephrotoxicity, or cytotoxicity, indicating their potential for clinical application (Supplementary Fig. S20b–d).

However, the IMQ-induced psoriasis animal model is an acute induction model characterized by acute inflammatory responses, and thus cannot fully replicate the chronic inflammatory state of psoriasis. In pharmacological studies, cessation of IMQ can alleviate symptoms in psoriatic mice. To better simulate the chronic inflammatory state of psoriasis, we also employed a second modeling method, namely therapeutic intervention under continuous IMQ induction conditions (Fig. 8a). In the continuous IMQ-induced mouse model, the disease progression continues to worsen, but significant improvements in the scaling, redness, and inflammation factor levels, as well as a notable reduction in skin thickness, were observed in the CMM-MNs group (Fig. 8b–k). In conclusion, microbial pneumatics microneedles demonstrated effective treatment of psoriasis through transdermal delivery of calcipotriol, offering potential clinical value in improving the condition and reducing inflammation response. At the same time, we observed an interesting phenomenon: in the MM-MNs group without drug loading, a relatively noticeable therapeutic effect was observed in the first non-sustained modeling experiment. The amount of skin scaling on the backs of the mice was significantly reduced, and there was also a significant difference in the reduction of inflammatory factors compared to the control group (Fig. 7c–i). This suggests that in the non-sustained psoriasis model, the gas produced by the active microorganisms in the microneedles also has a clear therapeutic effect. This observation led us to hypothesize that certain components of this gas might also help alleviate the inflammation associated with psoriasis, which shifted our focus to H2. Hydrogen (H2), as a molecule with strong antioxidant and anti-inflammatory properties, has shown potential therapeutic effects in various diseases in recent years24. For immune-mediated diseases like psoriasis, oxidative stress, and inflammatory responses are key factors in the pathological process. We speculate that H2 could alleviate psoriasis symptoms by reducing oxidative stress, thereby inhibiting the overactivation of inflammatory signaling pathways19. However, in the group subjected to continuous IMQ modeling, the therapeutic effect was significantly weakened and was limited to the skin area treated with microneedles. We believe this can be explained by the fact that psoriasis is an immune-mediated skin disease, with its pathological characteristics mainly reflected in the dysregulation of immune cells within the dermis. The activity of these immune cells in the dermis leads to local inflammation, accelerated proliferation of keratinocytes, and the formation of typical psoriatic plaques. The application of H2 alone only alleviated inflammation in the epidermis and did not impact the immune cells in the dermis. Therefore, in the continuous modeling animal model, where the disease continues to worsen, systemic therapeutic effects could not be achieved. However, for the CMM-MNs group, H2 can assist in deep drug delivery treatment, achieving a combined therapeutic effect. This unexpected coincidence greatly enhances the application efficacy of microbial microengineered microneedles.

Fig. 8: The therapeutic efficacy of microorganism micro-engine microneedles (MM-MNs) for chronic psoriasis treatment.
figure 8

a Establishment of the animal model and treatment procedure. Created with BioRender.com/u33l993. b Changes in PASI scores during the treatment process. Mean ± SD of n = 3 independent samples. c H&E staining results of the back skin in each group. Scale bar, 300 μm. d Changes in the condition of the back of mice during the treatment process. ej ELISA experiment to measure the levels of psoriasis-related inflammatory factors. (e): IL-17A; (f): IL-17F; (g): IL-10; (h): IL-22; (i): IL-23; (j): CXCL1; ej mean ± SD of n = 3 independent samples. Two-sided Student’s t test. A representative image of three biologically independent samples from each group is shown in (c) and (d).