Stretch-Responsive Adhesive Microcapsules for Strain-Regulated Antibiotic Release from Fabric Wound Dressings

Abstract

Bacterial infection of a wound is a major complication that can significantly delay proper healing and even necessitate surgical debridement and other complications. Conventional non-woven fabric dressings, including gauzes, bandages and cotton wools, often fail in treating wound infections in a timely manner due to their passive release mechanism of antibiotics. Here, we propose adhesive mechanically-activated microcapsules (MAMCs) capable of strongly adhering to a fibrous matrix to achieve a self-regulated release of antibiotics upon uniaxial stretching of non-woven fabric dressings. To achieve this, a uniform population of polydopamine (PDA)-coated MAMCs (PDA-MAMCs) are prepared using a microfluidics technique and subsequent oxidative dopamine polymerization. The PDA-MAMC allows for robust mechano-activation within the fibrous network through high retention and effective transmission of mechanical force under stretching. By validating the potential of a PDA-MAMCs-laden gauze to release antibiotics in a tensile strain-dependent manner, we demonstrate that PDA-MAMCs can be successfully incorporated into a woven material and create a smart wound dressing for control of bacterial infections. This new mechano-activatable delivery approach will open up a new avenue for a stretch-triggered, on-demand release of therapeutic cargos in skin-mountable or wearable biomedical devices.

Introduction

Extensive full-thickness wounds, including acute postsurgical incisions, transcutaneous prosthesis, burns, and non-healing ulcers, are susceptible to undesirable bacterial infection due to their hypoxic and protein-rich environments that are ideal for bacterial growth.^1–3 In particular, the formation of blisters due to various mechanical forces in the wound area, especially over joints that are constantly under motion (e.g., knee, elbow and finger knuckle), increases the risk of developing severe infection.^4–7 Improper treatment of the infection potentially gives rise to the development of biofilms which cause delayed wound healing and sometimes necessitate surgical debridement and in some extreme cases amputation.^1,8 For clinical wound management, non-woven fabric dressings, such as gauzes, bandages and cotton wools, are the long-standing used materials and most common options for protecting the wound bed from mechanical trauma, dehydration and infections.^9–11 However, the passive release of antibiotics from these conventional fabric-based dressings falls short of providing a reliable and timely treatment of infections in dynamic wound environments.^12,13 Even though several active releasing systems have been developed to treat infected skin wounds,^14–16 most of them are based on polymeric hydrogels to fill and repair tissue defects with their own confined matrix. Therefore, it is highly desirable to design smart, non-woven fabric dressings for programmable treatment of bacterial infections while maintaining their intrinsic breathability.

Microcapsules are capable of protecting encapsulated therapeutics within a solid shell against degradation and environmental factors and controlling their release based on the characteristics of the shell.^17–19 Various types of stimuli-responsive microcapsules that can release their contents in response to heat, chemical, and light activation have been successfully demonstrated. One endogenous stimulus that has not been taken advantage of to enable self-activation of drug release from microcapsules, despite their importance of mechanical forces in human physiology, is the mechanical deformation and mechanical loading that accompanies a variety of body motions and functions. A few recent studies have shown that microcapsules encapsulating active agents suspended in solution or embedded in hydrogels can be induced to undergo rupture and release their contents under mechanical forces and that such mechanical activation of microcapsules can be tailored by changing their dimensions.^20–24 To realize effective strain-triggered release of therapeutics from mechanically activated microcapsules (MAMCs)^23,24 within fibrous matrices such as those used in wound dressings, MAMCs need to have strong interactions with the fibrous substrate of the dressing, especially in a physiologically hydrated condition.¹⁷ Conventional methods for binding microcapsules onto fabrics, however, often result in considerable toxicity, low adhesion, and binder-induced film formation which could hinder the release of encapsulated drugs.^20,25–27

In this work, we develop a stretch-responsive delivery system by imparting strong adhesion between MAMCs and a fibrous matrix via a mussel-inspired coating to enable mechanically activated release of antibiotics from non-woven fabric dressings in response to tensile strains. Polydopamine (PDA) has emerged as one of the simplest and most versatile strategies to functionalize the surface of a wide variety of materials. The hydrated adhesive properties as well as biocompatibility and biodegradability make PDA particularly suitable for biomedical applications.^28–30 We hypothesize that PDA coating will markedly enhance the adhesion of MAMCs onto fibrous substrates in hydrated environments, and that the release of antibiotics from the adhesive MAMCs can be triggered by stretching of the dressing (Fig. 1A). We prepare a uniform population of the PDA-coated MAMCs (PDA-MAMCs) using a microfluidic technique^23,24 followed by oxidative dopamine polymerization (Fig. 1B),²⁸ and evaluate their adhesive properties and stretch-responsive mechano-activation functionality. The tensile strain-dependent release profile and antibacterial performance of ciprofloxacin (CIF)-loaded PDA-MAMCs embedded in an ordinary gauze are also assessed.

Fig. 1.

Fabrication and characterization of PDA-MAMCs. Schematic illustration of (A) the stretch-triggered antibiotics release from fabric wound dressing and (B) the fabrication of the PDA-MAMCs. (C) Microscopic images of a) the emulsification process within a capillary microfluidic device and b) the generated water-in-oil-in-water (W/O/W) double emulsions. (D) Fluorescent images of the MAMCs with labeled shells (red) and PDA-MAMCs with labeled PDA coating layer (green). (E) UV-Vis spectroscopy of the MAMCs, the PDA-MAMCs and a PDA-coated transparent polystyrene surface. Blue regions indicate the characteristic absorbance peaks for dopaminochrome at 388 nm and dimers of dopaminochrome and 5,6-dihyrixindole at 400-450 nm on the PDA-MAMC surface.

Results and discussion

Monodisperse poly(D,L-lactide-co-glycolide) (PLGA)-based MAMCs with a diameter of ~ 56 μm and a shell thickness of ~0.95 μm are fabricated by generating water-in-oil-in-water (W/O/W) double emulsions using a glass capillary microfluidic device, followed by solvent removal (Fig. 1C and Fig. S1, ESI^†).^23,24 PLGA, a bioresorbable polymer approved by the US Food and Drug Administration (FDA), is used as the shell material of MAMCs to make them suitable for various biomedical applications.³¹ Nile Red is added to the middle phase to fluorescently label the MAMC shell, facilitating their visualization. Incubation of the resulting MAMCs in an alkaline dopamine solution (1 mg/ml dopamine-hydrochloride; pH 8.5) leads to the formation of PDA coating on the MAMC surface,²⁵ as determined by fluorescence imaging (Fig. 1D). The successful coating is also confirmed by a distinct color change of the PDA-MAMC surface to light brown due to the oxidation of the dopamine monomers (Fig. S2, ESI^†). Consistent with this result from UV-Vis spectroscopy, we observe the characteristic absorbance peaks for dopaminochrome at 388 nm and dimers of dopaminochrome and 5,6-dihyrxyindole at 400-450 nm on the PDA-MAMC surface (Fig. 1E).^28,32From the ellipsometry result, the thickness of PDA-coated PLGA film is equivalent to the sum of thicknesses of PDA coating layer and PLGA film (Fig. S3A, ESI^†), suggesting that PDA likely is not infiltrating the PLGA film. Similarly, the PDA coating layer is located along the outer outline of PLGA shell without penetrating into the core of MAMC, which is confirmed by confocal microscopy (Fig. S3B, ESI^†).

The adhesive property of the PDA-MAMCs is evaluated by testing their ability to adhere to either the surface of porcine skin or a plastic surface under a hydrated condition (Fig. 2A). As-prepared MAMCs without PDA coating in Tris-HCl buffer are used as a negative control. The majority of the PDA-MAMCs remain adhered on both surfaces after rinsing with distilled water (DW), whereas the majority of the unmodified MAMCs are washed away from the surface, indicating superior adhesiveness of the PDA-MAMCs on wet substrates independent of the nature of the surface.

Fig. 2.

Adhesiveness of PDA-MAMCs. (A) Images showing adhesion of microcapsules to porcine skin and a plastic surface. The right-most images present confocal enlargements of the retained microcapsules on the plastic surface after washing. (B, C) Microcapsule detachment profiles from (B) porcine skin and (C) a plastic surface as a function of centrifugal force (n ≥ 4 specimens, ***p<0.005, **p<0.01, *p<0.05; Kolmogorov-Smirnov test).

The adhesion strength of the PDA-MAMC is quantitatively characterized by a centrifugation approach. In this method, a custom-designed substrate-holding block is used to maintain the substrate parallel to the centrifugal axis, allowing microcapsules on the substrate to experience only the centrifugal force (Fig. S4, ESI^†). Cationic poly-L-lysine (PLL) is coated onto the MAMC surface (PLL-MAMC) via simple agitation of the MAMCs in a PLL solution to serve as a positive control and to compare the adhesion strength to that of PDA-MAMCs.^33,34 The fraction of MAMCs that are removed from the surfaces of the two substrates as a function of centrifugal force clearly show that the PDA-MAMC has the strongest adhesion to both the porcine skin and the plastic surfaces. The adhesion force of PDA-MAMCs onto the plastic surface is 5 times higher compared to unmodified MAMCs and 2 times higher than PLL-MAMCs (Table 1 and Fig. 2B,C). All three types of MAMCs adhere more avidly to the porcine skin than to the plastic surface, which is likely due to the relatively higher surface roughness of the porcine skin. Again, PDA-MAMCs show the greatest increase in the adhesion strength on the porcine skin; the enhanced adhesion of PDA-MAMCs on the porcine skin could be due to the formation of covalent bonds between the catechol groups and the amine or thiol functional groups on the surface of porcine skin via Michael addition or Schiff base reactions.^28,35 In addition, an increase in the concentration of PDA coating (from 1 mg/ml to 10 mg/ml) on MAMCs does not result in an increase in adhesion strength on both the porcine skin and the plastic surfaces (Fig. S5, ESI^†), probably due to the full coverage of PDA coating layer on the surface of MAMCs (Fig. S3B, ESI^†) and the saturation of adhesion strength when coated with 1 mg/ml PDA.

Table 1.

Adhesion strength of PDA-MAMCs to porcine skin, a plastic surface and gauze (n ≥ 4).

Substrate	MAMC	PDA-MAMC	PLL-MAMC
Porcine skin	112 ± 29 g	633 ± 106 g	321 ± 77 g
Plastic	94 ± 19 g	510 ± 77 g	243 ± 58 g
Gauze	372 ± 63 g	1,092 ± 94 g	610 ± 104 g

To test PDA-MAMCs as strain-responsive drug carriers for wound dressing, we assess their release properties in a three-dimensional (3D) fibrous matrix of a commercial non-woven gauze under physiologically relevant strains. MAMCs are loaded into the gauze by placing a drop of MAMC suspension on top of the gauze and subsequently rinsing with DW. PDA-MAMCs exhibit a higher level of retention in the stretchable fabric gauze under a wet condition, compared to the unmodified MAMCs, suggesting the strong adhesiveness of the PDA-MAMC to the fibrous matrix (Fig. 3A and Fig. S6, ESI^†). We hypothesize that PDA-MAMCs that have strong adhesion to the fibers will experience compression and shear forces during stretching of the gauze, resulting in the mechano-activation and rupture of the PDA-MAMCs (Fig. 3B). To test this hypothesis, a PDA-MAMCs-laden gauze is subjected to stepwise grip-to-grip strains in uniaxial tension using a custom-designed micromechanical device (Fig. 3C and Fig. S7, ESI^†).³⁶ The shell and core of MAMCs are loaded with Nile Red and fluorescently labeled bovine serum albumin (BSA), respectively, to facilitate the characterization of mechano-activation and release of the cargo. MAMCs are used after confirming that the percentage of full microcapsules (% Full) is over 95 (Fig. S3C, ESI^†). When 20% tensile strain (~245 kPa) is applied, ~85% of the unmodified MAMCs detach from the fibrous network, while the majority of the PDA-MAMCs remain adhered on the gauze fibers (Fig. 3D,E). Upon stretching to 50% strain (~610 kPa), approximately 80% of the PDA-MAMCs remain within the gauze due to their strong adhesion to the fibers, whereas very few unmodified MAMCs remain under the same conditions. Since the gauze shows a plastic deformation and a breakage when tensile strain exceeds 50% (Fig. S8, ESI^†), strains of up to 50% are applied to the PDA-MAMCs-laden gauze. Importantly, more than half of the remaining PDA-MAMCs exhibit clear deformations after 20% strain of the gauze, with many losing their structural integrity. Loss of structural integrity is evidenced by the loss of fluorescent signal from the core of the PDA-MAMCs in ~97% of microcapsules, as shown in the 3D volume reconstruction of confocal z-stack images, indicating the release of their content (BSA) (Fig. 3D,F). The application of 50% tensile strain, physiologically relevant in the bending of knees, finger joints and wrists,^37,38 results in the mechano-activation of almost all of the PDA-MAMCs within the fibrous network. The shape of ruptured MAMCs show anisotropy, indicating that they experience a compression induced by the geometric change of the fibrous network upon uniaxial stretching (Fig. S9, ESI^†). In contrast, most of the unmodified MAMCs retain their spherical shape without detectable release of encapsulated fluorescent BSA, even after stretching the gauze to 50% strain. These unmodified MAMCs are likely dislodged from the fiber substrate during the gauze deformation, and so mechanical force is not transmitted to them. Collectively, these findings suggest that the adhesive interaction between the PDA-MAMCs and the underlying fiber substrate is essential for an effective mechano-activation of the PDA-MAMCs with uniaxial stretching of the gauze and that this enables a self-regulated release of the therapeutic cargos in a highly spatiotemporally controlled manner.

Fig. 3.

Stretch-induced mechano-activation of PDA-MAMCs in a fibrous matrix. (A) Pseudo-colored SEM images of gauze embedded with MAMCs and PDA-MAMCs. The red spheres indicate the microcapsules. (B) Schematic illustration of the stretch-induced drug release from the adhesive PDA-MAMCs in a fibrous matrix. (C) Schematic illustration showing the stepwise increments in strain as a function of time (left), with strain levels for investigation highlighted in red. Photographs of the PDA-MAMCs-laden gauze before (top) and after (bottom) application of tensile strain (50%). (D) Confocal microscopy images showing changes in the shape of the MAMCs with increasing strain. Fluorescent BSA is released from the aqueous core upon fracture of the MAMCs. (E, F) Percentage of (E) retained microcapsules (%) and (F) retained microcapsules that are full (%) as a function of tensile strain application. Data represent the mean ± standard deviation with statistical significance as indicated (n ≥ 500 microcapsules/loading regimen/type, 4 specimens/loading regimen/type, ***p<0.005; one-way ANOVA test with Tukey’s post-hoc test).

To evaluate the potential of the PDA-MAMCs to promote the release of antibiotics from fibrous matrices in response to a tensile strain, PDA-MAMCs containing CIF are prepared and loaded into the gauze. To support their utilization in this manner, we chose the potent gram-negative antibiotic CIF that is used clinically to treat a number of bacterial infections, including bone, joint, and skin infections.³⁹ The CIF-loaded PDA-MAMCs (CIF@PDA-MAMCs) are mechano-activated by stretching the gauze to 20% and 50% strains and the solutions containing released CIF are collected (Fig. 4A). A plain unaltered gauze (a negative control), an intact CIF@PDA-MAMCs-laden gauze (a non-activated control), and a gauze treated with the equivalent amount of free CIF (a positive control) are used for comparison. Similar to the results from confocal microscopy, stretching of the gauze induces the release of CIF from the CIF@PDA-MAMCs, with an increasing amount of CIF released with strain; conversely, CIF-loaded unmodified MAMCs (CIF@MAMCs) do not show any significant release of CIF with stretching (p<0.05, Fig. 4B). In addition, the amounts of released CIF from the CIF@PDA-MAMCs measured immediately after and at 24 h after stretching are not different, which suggests a burst release of CIF from the PDA-MAMCs upon rupture (Fig. S10, ESI^†).

Fig. 4.

Stretch-responsive antibiotics delivery from the CIF@PDA-MAMCs-laden fabric dressings. (A) A schematic representation of antibacterial assay using groups including no treatment (negative control), intact CIF@PDA-MAMCs, mechano-activated CIF@PDA-MAMCs and free CIF. The CIF@PDA-MAMCs are prepared with CIF in their inner core. Green color indicates presence of CIF in MAMC and in the medium. (B) In vitro release of CIF from the CIF@PDA-MAMCs-laden gauze as a function of tensile strain levels (n ≥ 500 microcapsules/loading regimen/type, 4 specimens/loading regimen/type). (C) Images and (D) area of the inhibition zones formed around the gauzes (n ≥ 3 specimens/type). (E) Viability of E. coli on CIF@PDA-MAMCs-laden gauze after 1 day of culture (n ≥ 3 specimens/type). Data represent the mean ± standard deviation with statistical significance as indicated (***p<0.005; one-way ANOVA test with Tukey’s post-hoc test).

The antibacterial performance of the mechano-activated CIF@PDA-MAMCs is assessed by the disk diffusion test using Escherichia coli (Fig. 4C). Intact CIF@PDA-MAMCs without mechanical activation exhibit negligible inhibition of bacterial growth, indicating that the CIF remains encapsulated in the PDA-MAMCs during the period of incubation. In contrast, significant inhibition zones are observed around CIF@PDA-MAMCs-laden gauze activated by 20% strain of stretching. The diameter of inhibition zones increases with an increase in the strain of mechano-activation, indicating a larger amount of CIF released from the CIF@PDA-MAMCs (Fig. 4D). The mechano-activation of the CIF@PDA-MAMCs at 50% strain achieves an inhibitory zone with a diameter comparable with that of administration of free CIF. The viability of E. coli on CIF@PDA-MAMCs-laden gauze in liquid media is also investigated after 1 day of incubation (Fig. 4E). Mechano-activated CIF@PDA-MAMCs-laden gauze significantly reduces the viability of E. coli at the end of the incubation period compared to non-activated CIF@PDA-MAMCs-laden gauze, consistent with the disk diffusion results. The efficacy of CIF@PDA-MAMCs-laden gauze on the inhibition of E. coli growth also increases with an increase in applied strain. Notably, incubation of bacterial cells on CIF@PDA-MAMCs-laden gauze that was stretched to 50% strain leads to more than 95% loss of bacterial viability. These results clearly demonstrate that stretching with different strain levels enables self-regulated release of antibiotics. We believe the PDA-MAMC technology has a great potential for the release various therapeutic agents such as pain relievers and anti-inflammatory agents or growth factors to promote tissue repair involving dermal and wound healing applications. Our future studies will focus on testing these possibilities via animal studies.

Conclusions

In conclusion, we developed stretch-responsive therapeutic MAMCs to achieve self-regulated release of an antibiotic from fabric wound dressings. Inspired by the adhesion capability of bivalve molluscs and exploiting a novel microfluidic encapsulation technique, we achieve strong adhesive interactions between drug-delivering PDA-MAMCs and the surrounding fibrous matrix of gauze under hydrated conditions, as well as robust mechano-activation of the PDA-MAMCs in response to uniaxial stretching of the gauze. Our findings highlight that the strain-dependent release of bioactive contents from PDA-MAMCs can enable a versatile approach to modulate the administration timing of therapeutics in response to body motions, such as bending of joints or exogenously applied pressure. The potential of utilizing PDA-MAMCs for controlled release of antibiotics to inhibit bacterial growth is validated, suggesting avenues for programmable antibacterial performance of fabric wound dressings. Our data strongly supports that the PDA-MAMCs with antibiotic payloads can be successfully employed to control bacterial infections. This has wide-ranging clinical application to the treatment of wounds, injuries or surgical sites, ranging from skin (e.g., forearm skin under strains of up to 25% in daily life⁴⁰) to internal tissues such as lung (over 20% strain during mechanical ventilation⁴¹) or aortic valve (40% strain in the radical direction⁴²) where notable strains are inherent or applied. The clinical applications and variety of therapeutic molecules that can be delivered using this MAMC delivery platform are far-reaching.

Experimental

Preparation of PDA-MAMCs

The MAMCs are fabricated using a glass capillary microfluidic device to generate W/O/W double emulsions as previously described.43 An aqueous solution of 1 mg/mL BSA (Sigma-Aldrich, St Louis, MO, USA) as a model drug or 5% (w/v) of the antibiotic CIF (Sigma-Aldrich) for antibacterial activity analyses is loaded in the inner core of the MAMCs. The middle phase consists of 85:15 PLGA (0.55-0.75 dL/g, ester-terminated; Lactel, Birmingham, AL, USA) dissolved in chloroform with the addition of 100 μg/mL Nile Red (Sigma-Aldrich) to fluorescently label the shell. The outer aqueous phase contains 2% (w/v) poly(vinyl alcohol) (PVA; Sigma-Aldrich) for stabilizing the outer oil/water interface, thus preventing coalescence of droplets during generation of double emulsions. The generated double emulsions are left in a large reservoir with the collecting solution of 0.1% (w/v) BSA in phosphate buffered saline (PBS, pH 12; Sigma-Aldrich) for 72 h to allow evaporation of chloroform from the middle phase and hardening of the shell wall.

After solidification, MAMCs are collected and washed with 10 mM Tris-HCl (pH 8.5; Sigma-Aldrich) solution. Subsequently, MAMCs are coated with PDA by immersing them in 10 mM Tris-HCl solution containing 1 mg/ml dopamine-hydrochloride (Sigma-Aldrich) with constant shaking at 100 rpm under ambient conditions for 12 h.⁴⁴ As positive controls for the adhesion study, MAMCs are alternatively treated with 10 mg/ml dopamine-hydrochloride or 1 mg/ml PLL (Sigma-Aldrich) dissolved in PBS.

For visualization of the PDA layer, fluorescein isothiocyanate (FITC) isomer 1 (Sigma-Aldrich) is added to the reacting solution. After rinsing with fresh Tris-HCl solution three times to remove residual dopamine monomers, PDA-MAMCs are collected in Tris-HCl and transferred to PBS (pH 7.4) just prior to use. The prepared PDA-MAMCs are observed by a confocal microscope (20× magnification, mid-plane imaging; Fluoview FV 1000; Olympus, Shinjuku, Tokyo, Japan) and the average outer diameter is measured using the ImageJ software (v.1.52; National Institutes of Health, Bethesda, MD, USA). The presence of PDA coating is also confirmed by monitoring the color change of MAMC suspensions and measuring the absorbance with a comparison with a PDA-coated transparent polystyrene surface using an UV-Vis spectrometer (Infinite M200; TECAN, Zürich, Switzerland). The concentration of PDA-MAMCs is defined as the number of microcapsules/ml (n ≥ 3 aliquots per fabrication batch) as measured under an optical microscope (Eclipse TE200; Nikon, Minato, Tokyo, Japan) within the first 3 days.

To investigate the possibility of PDA infiltration into the PLGA shell, the thickness change of a PLGA film after PDA coating is measured using an ellipsometer (M-2000 V and Alpha-SE; J. A. Woollam, Lincoln, NE, USA). PLGA is deposited onto a silicon wafer at the same concentration used to form the MAMCs. PDA is coated onto the PLGA film as above-mentioned.

Adhesion Analyses

To characterize the adhesion properties of PDA-MAMCs on a biologically relevant matrix, a porcine skin sheet (Stellen Medical, Saint Paul, MN, USA), a plastic surface (Thermo Fisher Scientific, Waltham, MA, USA) and a gauze pad (CVS Pharmacy, Inc., Woonsocket, RI, USA) are used as a model fabric wound dressing. For a qualitative analysis, ~1,000 microcapsules are placed onto the substrates under hydrated conditions and washed with DW after 30 min. The adhesion strength of the microcapsules is quantitatively measured using a custom centrifugal method that we adopted.^45,46 A polyurethane-based substrate-holding block is custom-designed and 3D-printed to maintain the substrate parallel to the centrifugal axis during centrifugation (Fig. S3, ESI^†). Microcapsules are placed onto the substrates (10 mm × 10 mm) and left for 30 min to allow for surface adhesion. The microcapsules-adhered substrate is firmly fixed to the holding block and loaded into a centrifuge tube, which is filled with deionized water. After centrifugation, adhesion strength (as defined by the force necessary to detach one half of the microcapsules, n ≥ 4 specimens) is measured by counting the number of remaining microcapsules using a digital microscope (5-MP; Celestron, Torrance, CA, USA) and converting the centrifugal force to a normal force according to the equipment manual of the centrifuge (Allegra X-12; Beckman coulter, Brea, CA, USA).

Stretch-Responsive Mechano-Activation Analyses

To monitor mechano-activation in a 3D fibrous matrix, the microcapsules are embedded in a commercial grade gauze (a piece of 10 mm × 50 mm; CVS Pharmacy) in the same way they were adhered onto the model surfaces. The morphological analysis of microcapsules-laden gauzes is performed using a scanning electron microscope (SEM; Quanta 600 FEG ESEM; FEI, Hillsboro, OR, USA). For uniaxial stretching of the microcapsules-laden gauze, a custom micromechanical test device is used.³² Samples are kept hydrated in PBS throughout testing. Then, grip-to-grip tensile strains of 20% or 50% are applied in 10% stepwise increments at a strain rate of 1%/s (Fig. 2C). After application of the tensile deformation, the detached microcapsules are collected and counted to determine the percentage of microcapsules retained in the gauze (n ≥ 500 microcapsules/loading regimen/type, 4 specimens/loading regimen/type). The stretched gauzes are incubated in PBS for 24 h under ambient condition to allow complete diffusion of inner contents post-rupture. The percentage of full (intact) microcapsules after load are measured using the maximum intensity projections obtained from the confocal z-stacks (10× magnification, n ≥ 500 microcapsules/loading regimen/type, 4 specimens/loading regimen/type).

The tensile strength of a gauze upon stretching is measured via uniaxial tensile tests using a universal testing machine (Instron, Norwood, MA, USA) equipped with a 10 N load cell. The gauze is immersed in PBS at room temperature prior to test, and grip-to-grip tensile strains are applied until complete separation in 10% stepwise increments at a strain rate of 1%/s.

Antibiotic Delivery and Analyses

To evaluate the efficacy of delivering antibiotics in response to tensile strain-induced mechano-activation, 5% (w/v) CIF is encapsulated in inner core of the microcapsules. After stretching to strains of 20% or 50%, 50 μL of fresh PBS solution is dropped onto the microcapsules-laden gauze to collect the released CIF. The amount of released CIF is quantified by measuring the absorbance at 277 nm using the UV-Vis spectrometer (n ≥ 500 microcapsules/loading regimen/type, 4 specimens/loading regimen/type).

The impact of CIF delivery from mechano-activated PDA-MAMCs is also assessed by analyzing antibacterial performances of the microcapsule-loaded gauzes against E. coli JM109 strain using the Kirby-Bauer method.⁴⁷ Gauze specimens of 8 mm diameter are placed on Luria-Bertani (LB) agar plates (Sigma-Aldrich) seeded with 100 μL of log-phase bacterial cells, and are incubated at 37 °C for 24 h. The area of inhibition zone is measured using Image J (n ≥ 3 specimens/type). The colony forming units (CFU) are also counted after incubation of the microcapsule-loaded gauzes at 37 °C for 24 h in 500 μL of LB medium seeded with bacterial cells in a log-phase at 1:100 (v/v) in a tissue culture plate. Relative bacterial cell viability is determined by dividing the CFU count in culture broth with the gauze specimens or soluble CIF by the CFU count without the gauze (n ≥ 3 specimens/type).

Statistical Analyses

All data are obtained from independent experiments carried out at least in triplicate (n ≥ 3). After Shapiro-Wilk test to evaluate the normality of data distribution, the significance of the data obtained from each group is statistically analyzed via one-way ANOVA test with Tukey’s post-hoc test (for a normal distribution) or Kruskal-Wallis test with Dunn’s post-hoc test (for a non-normal distribution). Cumulative distribution plots for the detachment of microcapsules in centrifuge-based adhesion analyses are analyzed via Kolmogorov-Smirnov test. The data represent the mean ± standard deviation with statistical significance as indicated (*p < 0.05, **p < 0.01, and ***p < 0.005). All data are processed using the R software (v.3.2.1; R Development Core Team, Vienna, Austria).