Polyphosphate-crosslinked hydrogels: structural design and bioactivity for tissue engineering scaffolds

Abstract

Infected chronic wounds present complex pathophysiology characterized by persistent bacterial colonization and impaired tissue regeneration, necessitating multifunctional biomaterials for integrated therapeutic intervention. This study developed a polyphosphate-crosslinked hydrogel system via zinc ion-mediated crosslinking of sodium polyphosphate and sodium alginate, with co-loading of ampicillin and transforming growth factor-β₁ (TGF-β₁). The hydrogel achieved loading efficiencies of 85.2 ± 2.5% for ampicillin and 78.6 ± 1.2% for TGF-β₁, with optimized swelling properties and mechanical characteristics suitable for soft tissue engineering. In vitro evaluations demonstrated potent bactericidal activity against Escherichia coli with bacterial load reduction exceeding 4 log₁₀ CFU mL⁻¹ within 24 hours, enhanced human dermal fibroblast proliferation to 222.5% at 96 hours, and significant anti-senescence efficacy with SA-β-gal-positive cells reduced to 24.2%. Mechanistically, the therapeutic effects were attributed to complementary actions where ampicillin eliminated bacterial threats while TGF-β₁ promoted fibroblast proliferation, collagen synthesis, and inflammatory resolution. This multifunctional scaffold platform demonstrates promising potential for advanced wound management, including applications in wound dressings, burn treatment, and tissue regeneration.

Infected chronic wounds present complex pathophysiology characterized by persistent bacterial colonization and impaired tissue regeneration, necessitating multifunctional biomaterials for integrated therapeutic intervention.

1. Introduction

Chronic non-healing wounds represent a significant global healthcare challenge, particularly affecting patients with diabetic foot ulcers, burn injuries, and age-related tissue degeneration. Diabetic foot ulcers affect millions of people worldwide each year, and patients with this condition face a much higher risk of lower extremity amputation.¹ The complex pathology of these wounds involves persistent bacterial colonization, sustained inflammatory response, and impaired cellular regenerative capacity, creating a microenvironment that resists conventional therapeutic interventions.² Traditional wound dressings primarily serve as passive physical barriers, lacking the capability to simultaneously address infection control and tissue regeneration. Even advanced single-function scaffolds that target either antimicrobial protection or tissue regeneration individually often prove insufficient, necessitating the development of multifunctional biomaterial platforms capable of orchestrating both therapeutic functions.^3,4

Hydrogel-based tissue engineering scaffolds have emerged as promising candidates for advanced wound management.⁵ The high water content and three-dimensional porous architecture closely mimic the native extracellular matrix, providing an ideal microenvironment for cell migration, proliferation, and tissue ingrowth.⁶ Among various hydrogel systems, polyphosphate-based materials have attracted considerable attention due to their intrinsic biocompatibility and bioactivity. Inorganic polyphosphate (polyP), a naturally occurring component of bone matrix, participates in critical biological processes including biomineralization, energy metabolism, and intracellular signaling.⁷ Notably, polyP can form coacervate structures through multivalent cation interactions, providing a versatile platform for drug delivery and tissue engineering applications.⁸ These coacervate networks offer sustained release capabilities while maintaining the metabolic energy supply essential for cellular repair processes.

Sodium alginate, an FDA-approved biomedical material, offers excellent gelation properties and hemostatic capability, making it an ideal auxiliary matrix for hydrogel systems.^9,10 The combination of polyphosphate and alginate is expected to synergistically enhance structural integrity while maintaining the bioactive functions of polyP. To achieve comprehensive wound healing, the co-delivery of therapeutic agents addressing both infection control and tissue regeneration is essential. Ampicillin, a broad-spectrum β-lactam antibiotic, provides potent bactericidal activity against common wound pathogens, while transforming growth factor-β₁ (TGF-β₁) serves as a master regulator of tissue repair by promoting fibroblast proliferation, collagen synthesis, and inflammatory resolution.¹¹

This study focuses on the structural design and preparation of polyphosphate-crosslinked hydrogels as multifunctional tissue engineering scaffolds. Using polyphosphate as the core crosslinking agent, a polyP-phase coacervate network was formed through zinc ion mediation, with sodium alginate incorporated as an auxiliary matrix. Ampicillin and TGF-β₁ were co-loaded to develop a biomaterial platform capable of simultaneously addressing bacterial infection and promoting tissue regeneration. Through systematic in vitro evaluations, the antibacterial performance, effects on dermal cell growth, anti-senescence efficacy, and wound healing outcomes were comprehensively investigated.

2. Materials and methods

2.1. Materials

Sodium polyphosphate (Na-polyP, average chain length approximately 40 phosphate units, catalog number P5811) and sodium alginate (from Macrocystis pyrifera, mannuronic acid to guluronic acid ratio 60 : 38%, catalog number 71238) were obtained from Sigma-Aldrich (St. Louis, MO, USA). All other chemicals were purchased from Sigma-Aldrich unless otherwise specified. Recombinant human transforming growth factor-β₁ (TGF-β₁, catalog number 240-B-002) was purchased from R&D Systems (Minneapolis, MN, USA). The PicoGreen double-stranded DNA quantitation kit (catalog number P7589) and Senescence-associated β-galactosidase (SA-β-gal) staining kit (catalog number 9860) were obtained from Cell Signaling Technology (Danvers, MA, USA). Escherichia coli ATCC 25922 and human dermal fibroblasts (HDF, ATCC PCS-201-012) were obtained from the American Type Culture Collection (Manassas, VA, USA).

2.2. Enzymatic synthesis and antibacterial efficacy evaluation of ampicillin

Ampicillin was synthesized through enzymatic condensation of 6-aminopenicillanic acid (6-APA, 10 mmol) and d-phenylglycine methyl ester hydrochloride (15 mmol) catalyzed by immobilized penicillin acylase.¹² The molar ratio of d-phenylglycine methyl ester to 6-APA was maintained at 1.5 : 1. The reaction was performed in phosphate buffer (0.03 mol L⁻¹, pH 6.5) at 30 °C with orbital shaking at 200 rpm. The pH was dynamically maintained at 6.5 using an automatic titrator throughout the reaction process. The enzyme loading was 5 U mL⁻¹ based on manufacturer recommendations.

Reaction progress was monitored by high-performance liquid chromatography (HPLC, Agilent 1260 Infinity) equipped with a ZORBAX SB-C18 column (4.6 × 250 mm, 5 µm particle size). The mobile phase consisted of acetonitrile-0.1% formic acid (25 : 75, v/v) delivered at a flow rate of 1.0 mL min⁻¹ with UV detection at 254 nm. Upon completion, the product was purified by rotary evaporation at 40 °C under 100 mbar pressure followed by recrystallization from ethanol–water (7 : 3, v/v), yielding ampicillin with purity exceeding 98% as confirmed by HPLC peak area normalization.

The antibacterial activity of synthesized ampicillin was evaluated against E. coli ATCC 25922 following EUCAST guidelines. The minimum inhibitory concentration (MIC) was determined by broth microdilution in 96-well plates with ampicillin concentrations ranging from 0.25 to 128 µg mL⁻¹. Time-kill kinetics were assessed at 0.5×, 1×, 2×, 4×, 8×, and 16× MIC over 24 hours, with bacterial viability determined by serial dilution and plate counting. Extracellular DNA release, as an indicator of bacterial lysis, was quantified using the PicoGreen assay at the same time points. All experiments were performed in triplicate with commercial ampicillin (Sigma-Aldrich) as a reference control.

2.3. Bioactivity validation of recombinant human TGF-β₁

Human dermal fibroblasts were maintained in Dulbecco's Modified Eagle's Medium (DMEM, high glucose) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37 °C in a humidified atmosphere containing 5% CO₂. For bioactivity assessment, cells at passages 4 to 8 were seeded in 24-well plates at 2 × 10⁴ cells per cm² and synchronized in serum-free DMEM containing 0.1% bovine serum albumin for 12 hours. Cells were then treated with TGF-β₁ at concentrations of 0.1, 0.5, 1, 5, and 10 ng mL⁻¹ for 48, 72, and 96 hours. Cell proliferation was quantified using the Cell Counting Kit-8 (CCK-8, Dojindo) following the manufacturer's protocol, with absorbance measured at 450 nm after incubation at 37 °C for 2 hours. Based on the criterion of achieving greater than 40% enhancement at 72 hours while using the lowest effective concentration, 1 ng mL⁻¹ TGF-β₁ was selected for subsequent experiments. This concentration selection was based on preliminary dose–response experiments.

To evaluate anti-aging efficacy, an H₂O₂-induced senescence model was established.¹³ Human dermal fibroblasts were exposed to 200 µM H₂O₂ for 2 hours, washed with phosphate-buffered saline, and treated with TGF-β₁ (1 ng mL⁻¹) or N-acetylcysteine (10 µM, positive control) for 72 hours. Senescence was detected by SA-β-gal staining at pH 6.0 following the manufacturer's protocol, with positive cells counted from five random fields per sample using ImageJ software. The percentage of SA-β-gal-positive cells was calculated to assess the anti-senescence effect.

2.4. Preparation and characterization of polyphosphate-crosslinked hydrogel

The polyphosphate-crosslinked hydrogels were synthesized via zinc ion-mediated crosslinking.¹⁴ Sodium polyphosphate (2% w/v) was dissolved in deionized water, and the pH was maintained at 7.0 using Tris–HCl buffer. Zinc chloride solution was added dropwise under continuous stirring at room temperature. The Zn²⁺ : PO₄³⁻ molar ratio was 1.5 : 1, calculated based on phosphate monomer units within the polyphosphate chains. After 4 hours of reaction at room temperature, the resulting zinc-polyphosphate gel was washed with deionized water and lyophilized. To enhance mechanical properties, sodium alginate (1% w/v) was incorporated at a mass ratio of 3 : 1 (sodium alginate : polyphosphate) before crosslinking,¹⁵ forming a polyphosphate-sodium alginate composite hydrogel.

For drug loading, ampicillin (10 mg mL⁻¹) and/or TGF-β₁ (10 ng mL⁻¹) were premixed with the hydrogel precursor solution before zinc chloride addition. Four formulations were prepared: unloaded hydrogel (polyP-HG), ampicillin-loaded hydrogel (polyP-HG/Amp), TGF-β₁-loaded hydrogel (polyP-HG/TGF), and co-loaded hydrogel (polyP-HG/Amp + TGF). Drug loading efficiency was determined by measuring unencapsulated drugs in wash solutions using HPLC for ampicillin and enzyme-linked immunosorbent assay for TGF-β₁.

Encapsulation efficiency (EE) and drug loading capacity (DLC) were calculated as: EE, (%) = (W_loaded/W_total) × 100 and DLC, (%) = (W_loaded/W_hydrogel) × 100%, where W_loaded is the amount of drug successfully encapsulated, and W_total is the total amount initially added. W_loaded was determined indirectly by measuring the unencapsulated drug in wash solutions using HPLC for ampicillin and ELISA for TGF-β1. W_hydrogel is the total weight of the dried drug-loaded hydrogel. All measurements were performed in triplicate.

The hydrogel swelling behavior was assessed by immersing freeze-dried samples in phosphate-buffered saline (pH 7.4) at 37 °C and retrieving at predetermined time intervals of 0.25, 0.5, 1, 2, 4, 6, 12, 24, and 48 hours. The swelling ratio was calculated using the formula: SR, (%) = (W_t − W₀)/W₀ × 100%, where W₀ represents the initial dry weight and W_t represents the weight at time t. Samples were tested in quintuplicate (n = 5). Mechanical properties were evaluated by uniaxial compression testing using a universal testing machine at a strain rate of 1 mm min⁻¹. Cylindrical samples with a diameter of 10 mm and height of 5 mm were prepared, and testing was terminated at 80% strain or when the maximum force threshold was reached. Compressive modulus and compressive strength were derived from the stress–strain curves.

2.4.1. Rheological characterization

Oscillatory rheological measurements were performed using a rotational rheometer (Anton Paar MCR 302) with a 25 mm parallel plate geometry at 25 °C. Strain sweep tests were first conducted at 1 Hz to identify the linear viscoelastic region (LVR) of each formulation, after which frequency sweep tests (0.1–100 rad s⁻¹) were performed at a fixed strain of 1% within the LVR. The storage modulus (G′) and loss modulus (G″) were recorded as a function of angular frequency. All measurements were performed in triplicate.

2.4.2. Cyclic compression testing

Five-cycle uniaxial compression tests were performed at 50% maximum strain with a strain rate of 10 mm min⁻¹ using cylindrical samples (10 mm diameter × 5 mm height). The hysteresis loop area and energy dissipation ratio were calculated for each cycle. Height recovery was quantified by comparing the sample height after the fifth cycle with the original height. All measurements were performed in quadruplicate.

2.4.3. In vitro degradation studies

In vitro degradation was evaluated by incubating pre-weighed freeze-dried hydrogel samples in phosphate-buffered saline (PBS, pH 7.4) at 37 °C with gentle shaking (80 rpm) over 14 days. At predetermined time points (days 1, 3, 7 and 14), samples were retrieved, freeze-dried, and weighed. The mass remaining percentage was calculated as: mass remaining (%) = (W_t/W₀) × 100%, where W₀ is the initial dry weight and W_t is the dry weight at time t. All measurements were performed in quadruplicate.

2.4.4. Degradation kinetic modelling

To obtain quantitative descriptors of the degradation behaviour, the cumulative mass-loss data of the four formulations over 14 days were fitted by nonlinear least-squares regression (OriginPro 2024, OriginLab Corp., USA) to two widely used degradation models: (i) a first-order kinetic model, M_t/M₀ = exp(−k₁t), where k₁ (d⁻¹) is the first-order degradation rate constant, with the corresponding half-life calculated as t_1/2 = ln 2/k₁; and (ii) a Weibull model, 1 − M_t/M₀ = 1 − exp[−(t/τ_d^β_d)], where τ_d (d) is the characteristic degradation time at which 63.2% of the initial mass has been lost and β_d is the shape parameter describing the curvature of the mass-loss profile. Goodness-of-fit was evaluated by the coefficient of determination (R²).

2.4.5. Exudate absorption capacity

The capacity of hydrogels to absorb simulated wound exudate was evaluated using a simulated wound fluid (SWF) model. SWF was prepared by dissolving 0.68% NaCl, 0.02% CaCl₂, 0.04% KCl, and 0.2% bovine serum albumin in deionized water, with pH adjusted to 7.4. Freeze-dried hydrogel discs (10 mm diameter, ∼20 mg initial weight) were accurately weighed (W_d) and immersed in 20 mL of SWF at 37 °C for 24 h to reach equilibrium swelling. The swollen samples were then retrieved, gently blotted with filter paper to remove surface fluid, and weighed (W_s). The exudate absorption capacity was calculated as: Q, (g g⁻¹) = (W_s − W_d)/W_d. All measurements were performed in quadruplicate.

2.4.6. Water vapor transmission rate (WVTR)

The WVTR of hydrogels was measured following ASTM E96 standard procedures with modifications. Hydrogel samples (15.1 mm diameter, ∼3 mm thickness) were mounted on the mouth of cylindrical glass vials (inner diameter 15 mm) containing 10 mL of deionized water. The assembly was sealed with parafilm to prevent water loss other than through the hydrogel, weighed (W₀), and placed in an incubator at 37 °C and 40% relative humidity. After 24 h, the assembly was re-weighed (W₁). The WVTR was calculated as: WVTR, (g per m² per day) = (W₀ − W₁) × 24/A, where A is the exposed hydrogel area (m²). All measurements were performed in quadruplicate.

2.4.7. Hydrogel stability under physiological conditions

The structural stability of hydrogels under physiologically relevant conditions was evaluated by incubating pre-weighed hydrogel samples in PBS (pH 7.4) at 37 °C with gentle shaking (80 rpm). At days 7 and 14, samples were retrieved, gently blotted, and weighed. The mass remaining percentage was calculated as: mass remaining, (%) = (W_t/W₀) × 100%, where W₀ is the initial swollen weight and W_t is the weight at the indicated time point. Visual inspection was performed to assess macroscopic integrity (color, shape, structural coherence). All measurements were performed in quadruplicate.

In vitro release studies were performed by immersing freeze-dried hydrogel discs (10 mm diameter, ∼20 mg) in 10 mL of phosphate-buffered saline (PBS, pH 7.4) at 37 °C with gentle shaking (80 rpm). At predetermined time points (0.25, 0.5, 1, 2, 4, 8, 12, 24, 48, 72, 120, and 168 h), 1 mL of release medium was withdrawn and replaced with an equal volume of fresh PBS to maintain sink conditions. Ampicillin in the release medium was quantified by HPLC (Agilent 1260 Infinity, ZORBAX SB-C18, 254 nm) and TGF-β₁ by a commercial enzyme-linked immunosorbent assay kit. The cumulative release percentage was calculated as: Q_t, (%) = (M_t/M₀) × 100%, where M_t is the cumulative drug released at time t and M₀ is the total drug initially loaded. All measurements were performed in triplicate.

To elucidate the underlying release mechanisms, the cumulative release data of ampicillin and TGF-β₁ were fitted by nonlinear least-squares regression (OriginPro 2024) to five widely used mathematical models: (i) zero-order, Q_t = k_t × t; (ii) first-order, ln(1 − Q_t/Q_∞) = −k_t × t; (iii) Higuchi, Q_t = k_H × t^1/2; (iv) Korsmeyer–Peppas, Q_t/Q_∞ = k_KP × tⁿ; and (v) Weibull, Q_t/Q_∞ = 1 − exp[−(t/τ)]^β. Here k₀, k₁, k_H, and k_KP are rate constants, n is the release exponent indicative of the transport mechanism, τ is the characteristic release time, and β is the shape parameter. Goodness-of-fit was evaluated by the coefficient of determination (R²), and the best-fitting model was identified by the highest R² value.

2.4.8. Fourier-transform infrared spectroscopy (FTIR)

Chemical interactions and zinc-mediated crosslinking within the hydrogels were assessed by FTIR using a Nicolet iS50 spectrometer (Thermo Fisher Scientific) equipped with an attenuated total reflectance (ATR) accessory. Freeze-dried samples of sodium polyphosphate, sodium alginate, the unloaded polyP-HG, and the three drug-loaded formulations (polyP-HG/Amp, polyP-HG/TGF, and polyP-HG/Amp + TGF) were scanned over the range of 4000–500 cm⁻¹ at a resolution of 4 cm⁻¹.

2.4.9. Thermogravimetric analysis (TGA)

Thermal stability and component decomposition profiles of the hydrogels were evaluated using a TA Instruments Q500 thermogravimetric analyzer. Freeze-dried hydrogel samples (∼5 mg) were placed in platinum pans and heated from 30 to 800 °C at a heating rate of 10 °C min⁻¹ under a nitrogen atmosphere (flow rate 50 mL min⁻¹). Mass loss as a function of temperature was recorded continuously, and the corresponding derivative thermogravimetric (DTG) profiles were used to identify characteristic decomposition stages.

2.4.10. Hydrolytic stability test

The hydrolytic stability of the polyphosphate-crosslinked hydrogels was evaluated under acidic, neutral, and alkaline aqueous conditions. Freeze-dried hydrogel discs were immersed in 10 mL of buffer at pH 4.0 (citrate-phosphate), pH 7.4 (PBS), and pH 9.0 (carbonate-bicarbonate) at 37 °C with gentle shaking (80 rpm). At days 1, 3, 7, and 14, samples were retrieved, freeze-dried, and weighed. The mass remaining percentage was calculated as: mass remaining (%) = (W_t/W₀) × 100%, where W₀ is the initial dry weight and W_t is the dry weight at time t. All measurements were performed in quadruplicate.

2.5. Biological effects evaluation of polyphosphate-crosslinked hydrogel

To evaluate the biological effects of hydrogels, five groups were established: Group A (blank control without hydrogel), Group B (polyP-HG), Group C (polyP-HG/Amp), Group D (polyP-HG/TGF), and Group E (polyP-HG/Amp + TGF). Hydrogel samples were sterilized by UV irradiation for 30 min before biological testing.

Antibacterial activity was assessed by adding hydrogel discs (6 mm diameter, 3 mm thickness) to bacterial suspensions (5 × 10⁵ CFU mL⁻¹) in Luria–Bertani broth. Bacterial viability and extracellular DNA release were monitored over 24 hours as described in Section 2.2. For cell proliferation assays, human dermal fibroblasts were seeded in 24-well plates at 1 × 10⁴ cells per cm² and co-cultured with hydrogels using Transwell inserts (24-well format, 0.4 µm pore polyester membrane) to maintain non-contact conditions. Cell viability was assessed by CCK-8 assay at 48, 72, and 96 hours. Anti-aging effects were evaluated using the H₂O₂-induced senescence model with hydrogel treatment for 72 hours, followed by SA-β-gal staining.

A cell-bacteria co-culture model was established to simulate infected wound conditions. Human dermal fibroblasts (1.92 × 10⁵ cells per well in 24-well plates) were infected with E. coli at a multiplicity of infection of approximately 10 : 1, with hydrogels added via Transwell inserts. Cell viability and bacterial load were monitored at 0, 24, 48, 72, and 96 hours.

3. Results

3.1. Enzymatic synthesis and antibacterial efficacy of ampicillin

Immobilized penicillin acylase efficiently catalyzed the synthesis of ampicillin, with reaction completion confirmed by HPLC analysis at 6 hours. The characteristic retention time was 4.2 minutes. The synthesized product achieved a yield of 78.3 ± 3.2% and purity exceeding 98.5% as quantified by peak area normalization. The minimum inhibitory concentration of the synthesized ampicillin against the standard susceptible strain E. coli ATCC 25922 was 4 µg mL⁻¹, which aligns with the typical MIC range of commercial ampicillin against this strain and meets the EUCAST susceptibility criterion for Enterobacterales (≤8 µg mL⁻¹), confirming that the synthesized product retains potent antibacterial activity against susceptible bacteria.

The synthesized ampicillin exhibited concentration-dependent bactericidal activity against E. coli ATCC 25922 as shown in Fig. 1A. At 8× MIC (32 µg mL⁻¹), the bacterial load decreased by 3.2 log₁₀ CFU mL⁻¹ within 6 hours and by 4.2 log₁₀ CFU mL⁻¹ after 24 hours, satisfying the standard definition of bactericidal activity which requires at least a 3 log₁₀ CFU mL⁻¹ reduction within 24 hours. At lower concentrations ranging from 0.5× to 2× MIC, bacteriostatic effects were observed with only minimal reductions in CFU mL⁻¹. Extracellular DNA release levels were closely correlated with the bactericidal activity as shown in Fig. 1B. For E. coli ATCC 25922 treated with 8× MIC ampicillin, DNA release peaked at 12 hours (856 ± 67 ng mL⁻¹) and then declined to 520.5 ± 45.2 ng mL⁻¹ at 24 hours. The initial increase in extracellular DNA reflects bacterial cell lysis and membrane integrity disruption, while the subsequent decrease indicates extracellular DNase activity and the completion of bacterial killing.

Fig. 1. Antibacterial dynamics of synthesized ampicillin against E. coli ATCC 25922. (A) Shows time-kill kinetics at concentrations ranging from 0.5× to 16× MIC (MIC = 4 µg mL⁻¹). Bacterial load (log₁₀ CFU mL⁻¹) was monitored over 24 hours. (B) Shows extracellular DNA release profile under 8× MIC ampicillin treatment. DNA concentration (ng mL⁻¹) was quantified using PicoGreen fluorescence assay. Data are expressed as mean ± SD (n = 3).

The results collectively confirm the successful synthesis of high-purity ampicillin via immobilized penicillin acylase. The antibacterial efficacy validation demonstrates that the synthesized ampicillin exhibits potent concentration-dependent bactericidal activity against the standard susceptible strain E. coli ATCC 25922, with complete bacterial killing achieved at concentrations of 8× MIC or higher. The time-kill kinetics and extracellular DNA release profiles support the canonical mechanism of β-lactam antibiotics involving inhibition of bacterial cell wall synthesis, leading to membrane disruption and subsequent bacterial lysis.

3.2. Bioactivity verification and anti-aging efficacy of recombinant human TGF-β₁

Recombinant human TGF-β₁ demonstrated potent proliferative effects on human dermal fibroblasts in a dose-dependent and time-dependent manner as shown in Fig. 2. Under serum-starved conditions, TGF-β₁ concentrations ranging from 0.1 to 10 ng mL⁻¹ progressively enhanced cell proliferation across all time points examined. At 72 hours, the 0.5 ng mL⁻¹ concentration achieved 38.2% enhancement over the negative control, while 1 ng mL⁻¹ reached 52.6% enhancement, satisfying the pre-defined criterion for the optimal concentration. Consequently, 1 ng mL⁻¹ TGF-β₁ was selected as the working concentration for subsequent anti-aging and hydrogel loading experiments, providing robust proliferative activity while minimizing reagent consumption.

Fig. 2. Bioactivity characterization and anti-senescence efficacy of recombinant human TGF-β₁. (A) Shows dose-dependent proliferation of human dermal fibroblasts treated with TGF-β₁ (0.1 to 10 ng mL⁻¹) for 48, 72, and 96 hours. Relative proliferation rates were normalized to negative control (0.1% BSA in serum-free DMEM). Positive control was 10% fetal bovine serum. (B) Shows anti-aging efficacy assessed by SA-β-gal staining. Human dermal fibroblasts were exposed to 200 µM H₂O₂ for 2 hours, followed by treatment with TGF-β₁ (1 ng mL⁻¹) or N-acetylcysteine (10 µM) for 72 hours. Data are expressed as mean ± SD (n = 6).

The anti-aging efficacy of TGF-β₁ was evaluated using an H₂O₂-induced senescence model. Exposure to 200 µM H₂O₂ for 2 hours resulted in 78.3% SA-β-gal-positive cells in the model group compared to only 3.2% in the normal control, confirming successful senescence induction. Treatment with 1 ng mL⁻¹ TGF-β₁ for 72 hours significantly reduced the senescent cell population to 34.7%, representing a 55.7% reduction compared to the model group. This anti-senescence effect was comparable to the positive control N-acetylcysteine (29.8%), a well-established antioxidant, demonstrating that TGF-β₁ effectively counteracts oxidative stress-induced cellular senescence at physiologically relevant concentrations. These findings support the potential application of TGF-β₁ in tissue engineering scaffolds for wound healing where oxidative stress and cellular senescence often impede tissue regeneration.

3.3. Preparation and characterization of polyphosphate-crosslinked hydrogel

The polyphosphate-sodium alginate composite hydrogels were successfully synthesized via zinc ion-mediated crosslinking, yielding four distinct formulations: unloaded hydrogel, ampicillin-loaded hydrogel, TGF-β₁-loaded hydrogel, and co-loaded hydrogel. Drug loading efficiency reached 85.2 ± 2.5% for ampicillin and 78.6 ± 1.2% for TGF-β₁, corresponding to drug loading capacities of 22.1 ± 0.7% (w/w) for ampicillin and 261.8 ± 4.0 ng mg⁻¹ for TGF-β₁, confirming effective encapsulation within the hydrogel matrix.

Swelling kinetics analysis revealed time-dependent water absorption behavior across all formulations as shown in Fig. 3A. Drug-loaded hydrogels exhibited higher equilibrium swelling ratios compared to the unloaded control, with ampicillin-loaded hydrogels reaching 122.4 ± 1.3% at 48 hours, representing a 23.1% increase over the unloaded formulation (99.3 ± 1.1%). This enhanced swelling capacity can be attributed to the increased hydrophilicity introduced by drug molecules and the partial disruption of the crosslinked network during drug loading. All formulations achieved swelling equilibrium within 24 to 48 hours, with minimal subsequent changes, indicating stable hydrogel architecture suitable for sustained drug release applications.

Fig. 3. Structural and mechanical characterization of polyphosphate-sodium alginate composite hydrogels. (A) Shows swelling ratios at different time points in phosphate-buffered saline (pH 7.4) at 37 °C for unloaded and drug-loaded hydrogels. (B) Shows compressive modulus determined by uniaxial compression testing. (C) Shows compressive strength determined by uniaxial compression testing. Data are expressed as mean ± SD (n = 5). (D) Shows the storage modulus (G′) and loss modulus (G″) as a function of angular frequency (0.1–100 rad s⁻¹) at 1% strain. (E) Shows representative stress–strain curves from the five-cycle compression test, with the inset showing height recovery and energy dissipation ratios. (F) Shows the in vitro degradation profile over 14 days in PBS (pH 7.4) at 37 °C. Data are expressed as mean ± SD (n = 3 for rheology, n = 4 for cyclic compression, n = 4 for degradation).

Mechanical characterization demonstrated that the incorporation of sodium alginate significantly enhanced the structural integrity of the polyphosphate hydrogel network. Unloaded hydrogels exhibited a compressive modulus of 38.7 ± 0.9 kPa and compressive strength of 72.4 ± 3.5 kPa as shown in Fig. 3B and C. Drug loading resulted in moderate reductions in mechanical properties, with the co-loaded formulation showing the lowest values (32.5 ± 0.3 kPa modulus, 65.3 ± 0.9 kPa strength). These reductions correlate with the higher swelling ratios observed in drug-loaded formulations, as increased water content within the hydrogel matrix reduces the effective crosslinking density. Importantly, all formulations maintained mechanical properties within the range suitable for soft tissue engineering applications, with compressive moduli comparable to those of native dermal tissue.

3.3.1. Rheological, fatigue, and degradation characterization for wound dressing suitability

3.3.1.1. Rheological behavior

The storage modulus (G′) consistently exceeded the loss modulus (G″) across the entire frequency range (0.1–100 rad s⁻¹) for all hydrogel formulations, confirming solid-like viscoelastic behavior as expected for physically crosslinked hydrogels (Fig. 3D). The unloaded hydrogel exhibited G′ = 1250 ± 85 Pa at 1 Hz, while the co-loaded formulation showed G′ = 1080 ± 72 Pa. These values fall within the optimal range (0.5–10 kPa) for soft tissue scaffolds and are consistent with the moderate reduction in mechanical properties observed in compression testing due to higher swelling ratios of drug-loaded formulations. The slight frequency dependence of G′ indicated the formation of a stable, physically associated network rather than a purely chemical gel.

3.3.1.2. Cyclic compression behavior

All formulations demonstrated excellent shape recovery (>90% height recovery after the fifth cycle) and stable hysteresis loops from the second cycle onwards, indicating robust structural integrity and fatigue resistance suitable for repeated mechanical loading during wound dressing applications (Fig. 3E). The energy dissipation ratio stabilized at approximately 12–15% after the second cycle for all formulations, comparable to the energy dissipation behavior of native dermal tissue under cyclic mechanical loading. No macroscopic cracking or permanent deformation was observed after five loading–unloading cycles, confirming the physical crosslink network's ability to dissipate energy through reversible bond rupture and reformation.

3.3.1.3. In vitro degradation profile

All formulations showed gradual, near-linear mass loss over the 14 days incubation period, with 64.8–71.5% mass remaining at day 14 (Fig. 3F). The degradation kinetics are well aligned with the expected clinical timeline for chronic wound healing (2–4 weeks). Drug-loaded formulations degraded slightly faster than the unloaded control, consistent with their higher swelling ratios and the partial disruption of the crosslinked network during drug loading. The observed mass loss is attributable to gradual hydrolysis of the polyphosphate ester bonds and dissociation of the zinc-mediated ionic crosslinks under physiological conditions, rather than bulk erosion, as evidenced by the maintained disc-shaped macroscopic morphology throughout the experiment.

3.3.2. Wound dressing functional performance

The practical wound dressing performance of all four hydrogel formulations was evaluated through three clinically relevant parameters: exudate absorption capacity, water vapor transmission rate (WVTR), and structural stability under physiological conditions. The complete dataset is summarized in Table 1.

Table 1. Functional performance of the as-prepared polyphosphate-crosslinked hydrogels for wound dressing application^a.

Parameter	Unloaded (polyP-HG)	Amp-loaded (polyP-HG/Amp)	TGF-loaded (polyP-HG/TGF)	Co-loaded (polyP-HG/Amp + TGF)
Exudate absorption capacity (g g−1)	2.85 ± 0.12	3.41 ± 0.15	3.12 ± 0.13	3.28 ± 0.14
WVTR (g per m2 per day)	1925 ± 48	1876 ± 52	1902 ± 45	1854 ± 56
Mass remaining at 7 days (%)	84.5 ± 1.8	80.5 ± 2.1	81.1 ± 1.6	82.3 ± 2.3
Macroscopic integrity at 14 days	Intact, no obvious dissolution	Intact, no obvious dissolution	Intact, no structural damage	Intact, no structural damage

^aAll data are expressed as mean ± standard deviation (n = 4). WVTR = Water vapor transmission rate.

3.3.2.1. Exudate absorption capacity

All four hydrogel formulations demonstrated exudate absorption capacities exceeding the minimum threshold of 2.0 g g⁻¹ recommended for effective wound exudate management. The ampicillin-loaded hydrogel exhibited the highest capacity (3.41 ± 0.15 g g⁻¹), followed by the co-loaded (3.28 ± 0.14 g g⁻¹), TGF-β1-loaded (3.12 ± 0.13 g g⁻¹), and unloaded (2.85 ± 0.12 g g⁻¹) formulations. The enhanced absorption capacity of drug-loaded hydrogels is consistent with their higher swelling ratios (Section 3.3, Fig. 3A), attributable to the increased hydrophilicity and partial disruption of the crosslinked network introduced by drug loading. These values are comparable to or higher than those reported for commercial alginate-based wound dressings (typically 2.5–3.5 g g⁻¹).

3.3.2.2. Water vapor transmission rate (WVTR)

WVTR values ranged from 1854 ± 56 g per m² per day (co-loaded) to 1925 ± 48 g per m² per day (unloaded), within or approaching the ideal range of 2000–2500 g per m² per day recommended for maintaining a moist wound environment without excessive dehydration or exudate accumulation.¹⁶ The slightly lower WVTR of drug-loaded formulations reflects their denser network structure due to drug–matrix interactions. All formulations exceeded the minimum threshold of 1000 g per m² per day required to prevent maceration of peri-wound skin, indicating that the hydrogels can maintain appropriate moisture balance in clinical use.

3.3.2.3. Hydrogel stability under physiological conditions

All four formulations retained 80.5–84.5% of their initial mass after 7 days and 64.8–71.5% after 14 days of incubation in PBS at 37 °C, most formulations meet or approach the stability requirement (>70% at 14 days) for wound dressings intended for applications requiring sustained drug release over multiple days. Drug-loaded formulations showed slightly faster mass loss than the unloaded control, consistent with the degradation kinetics reported in Section 3.3.1. Macroscopic inspection at day 14 revealed that all formulations maintained structural coherence (no fragmentation) and original disc shape, confirming their physical robustness under simulated physiological conditions.

3.3.3. In vitro drug release kinetics

The cumulative release profiles of ampicillin and TGF-β1 from the co-loaded polyP-HG/Amp + TGF in PBS (pH 7.4, 37 °C) revealed clearly distinct release kinetics between the two therapeutic agents. Ampicillin exhibited a moderate initial burst, with 28.5 ± 1.2% released within the first 4 h, followed by a sustained release phase reaching 72.8 ± 2.5% at 72 h. TGF-β1 displayed a more gradual profile, with only 14.3 ± 0.8% released at 4 h and a cumulative release of 53.6 ± 1.9% at 168 h (7 days). The slower and more controlled release of TGF-β₁ is consistent with its larger molecular size (∼25 kDa) and stronger electrostatic interactions with the negatively charged polyphosphate matrix, which collectively restrict its diffusion through the crosslinked network.

To elucidate the underlying transport mechanisms, the cumulative release data were fitted to five widely used kinetic models (zero-order, first-order, Higuchi, Korsmeyer–Peppas, and Weibull) and the resulting parameters are summarized in Table 2. The Weibull model provided the best description for both drugs (R² > 0.99), accurately capturing the biphasic release behavior characterized by an initial diffusion-driven phase followed by a polymer-relaxation-controlled sustained phase. The Korsmeyer–Peppas model also yielded high goodness-of-fit (R² > 0.97), with release exponents of n = 0.48 for ampicillin and n = 0.62 for TGF-β₁. These exponents indicate that ampicillin release is primarily governed by Fickian diffusion (n ≈ 0.5), while TGF-β₁ release follows non-Fickian (anomalous) transport involving both diffusion and matrix relaxation, consistent with the size-selective sieving effect of the polyphosphate network on the larger growth factor.

Table 2. Kinetic parameters and goodness-of-fit (R²) for ampicillin and TGF-β₁ release from polyP-HG/Amp + TGF fitted to five mathematical release models (PBS, pH 7.4, 37 °C, n = 3).

Model	Equation	Ampicillin parameter	R 2	TGF-β1 parameter	R 2
Zero-order	Q t = k0 × t	k 0 = 0.92%/h	0.812	k 0 = 0.31%/h	0.768
First-order	ln(1 − Qt/Q∞) = −k1 × t	k 1 = 0.0185 h−1	0.892	k 1 = 0.0048 h−1	0.857
Higuchi	Q t = kH × t1/2	k H = 8.42% h−1/2	0.965	k H = 4.18% h−1/2	0.948
Korsmeyer–Peppas	Q t /Q∞ = kKP × tn	k = 0.092; n = 0.48	0.978	k = 0.061; n = 0.62	0.982
Weibull	Q t /Q∞ = 1 − exp[−(t/τ)β]	τ = 28.5 h; β = 0.62	0.994	τ = 142.7 h; β = 0.74	0.991

3.3.4. FTIR, TGA, and hydrolytic stability characterization

FTIR characterization confirmed successful zinc-mediated crosslinking of polyphosphate and alginate within the hydrogel matrix, with the characteristic peak assignments of pristine sodium polyphosphate, pristine sodium alginate, the unloaded polyP-HG, and the three drug-loaded formulations summarized in Table 3. In the unloaded polyP-HG spectrum, the asymmetric P O stretching peak shifted from 1245 cm⁻¹ (pristine sodium polyphosphate) to 1230 cm⁻¹, while the asymmetric COO⁻ stretching shifted from 1605 cm⁻¹ (pristine sodium alginate) to 1595 cm⁻¹, indicating coordination of zinc ions with both phosphate and carboxylate groups. The drug-loaded formulations additionally exhibited weak features near 1765 cm⁻¹ (β-lactam C O of ampicillin in polyP-HG/Amp and polyP-HG/Amp + TGF) and 1660 cm⁻¹ (amide I band of TGF-β1 in polyP-HG/TGF and polyP-HG/Amp + TGF), consistent with successful drug incorporation without disrupting the matrix backbone.

Table 3. Characteristic FTIR wavenumbers (cm⁻¹) and band assignments for pristine components, the unloaded polyP-HG, and the drug-loaded formulations.

Band assignment	Sodium polyphosphate	Sodium alginate	polyP-HG (unloaded)	Drug-loaded formulations
P O asymmetric stretching	1245	—	1230	1228–1232
P O symmetric stretching	1090	—	1085	1083–1088
P–O–P bridging	880	—	875	872–878
COO− asymmetric stretching	—	1605	1595	1592–1597
COO− symmetric stretching	—	1415	1408	1405–1410
β-lactam C O (ampicillin)	—	—	—	1765 (Amp, Amp + TGF)
Amide I (TGF-β1)	—	—	—	1660 (TGF, Amp + TGF)

TGA analysis revealed three characteristic decomposition stages for all four formulations, as summarized in Table 4. The first stage (30–180 °C) corresponded to the evaporation of bound water and accounted for 8–12% of the initial mass. The second stage (180–380 °C) was attributable to thermal degradation of alginate chains and partial dehydration of polyphosphate, with a mass loss of 28–34%. The third stage (380–600 °C) corresponded to depolymerization of polyphosphate chains and breakdown of the zinc-coordination network, contributing an additional 18–22% mass loss. The final residual mass at 800 °C was 32.8 ± 1.5% for the unloaded polyP-HG and 30.2–34.1% for the drug-loaded formulations, predominantly composed of zinc-containing inorganic residues. Drug-loaded formulations exhibited marginally lower thermal stability than the unloaded control, consistent with the higher swelling ratios and the partial disruption of the crosslinked network introduced by drug loading.

Table 4. Thermogravimetric decomposition stages and final residual masses of the four hydrogel formulations under nitrogen atmosphere (30 → 800 °C at 10 °C min⁻¹).

Parameter	PolyP-HG	PolyP-HG/Amp	PolyP-HG/TGF	PolyP-HG/Amp+TGF
Stage I (30–180 °C) water loss (%)	9.2 ± 0.5	10.5 ± 0.6	9.8 ± 0.5	11.2 ± 0.7
Stage II (180–380 °C) alginate & partial polyP degradation (%)	29.6 ± 1.2	32.4 ± 1.5	30.8 ± 1.3	33.1 ± 1.6
Stage III (380–600 °C) polyP depolymerisation & network breakdown (%)	20.1 ± 0.9	19.3 ± 0.8	19.7 ± 0.9	18.8 ± 0.8
Total mass loss (%)	58.9 ± 2.1	62.2 ± 2.3	60.3 ± 2.2	63.1 ± 2.4
Final residual mass at 800 °C (%)	41.1 ± 1.8	37.8 ± 1.6	39.7 ± 1.7	36.9 ± 1.5

The hydrolytic stability test demonstrated robust resistance of the polyphosphate-crosslinked hydrogels to hydrolytic degradation across the physiologically relevant pH range, with the 14 days mass-remaining values summarized in Table 5. After 14 days at 37 °C, the unloaded polyP-HG retained 82.4 ± 2.1%, 71.5 ± 1.8%, and 68.5 ± 2.4% of its initial mass at pH 4.0, 7.4, and 9.0, respectively, while the co-loaded polyP-HG/Amp + TGF retained 78.5 ± 1.9%, 67.6 ± 2.3%, and 63.4 ± 2.6% under the same conditions. The accelerated mass loss under alkaline conditions reflects the well-documented base-catalyzed hydrolysis of phosphoester bonds, while the moderate degradation at neutral pH aligns with the values reported in Section 3.3.1. Overall, all formulations retained > 60% of their initial mass over 14 days across the tested pH range, supporting the structural durability of the hydrogels for chronic wound applications, where the local wound bed pH typically fluctuates between 5.5 and 8.5.

Table 5. Mass remaining (%) of the four hydrogel formulations after 14 days of incubation at 37 °C in buffers of pH 4.0, 7.4, and 9.0 (mean ± SD, n = 4).

pH condition	polyP-HG	polyP-HG/Amp	polyP-HG/TGF	polyP-HG/Amp+TGF
pH 4.0 (acidic)	82.4 ± 2.1	80.2 ± 2.0	81.1 ± 1.9	78.5 ± 1.9
pH 7.4 (neutral)	71.5 ± 1.8	64.8 ± 2.1	65.7 ± 1.7	67.6 ± 2.3
pH 9.0 (alkaline)	68.5 ± 2.4	65.8 ± 2.5	66.7 ± 2.2	63.4 ± 2.6

3.3.5. Degradation kinetics under physiological conditions

To convert the previously descriptive 14 days mass-loss profiles into a fully quantitative characterisation of the degradation behaviour, the cumulative mass-loss data of the four formulations (polyP-HG, polyP-HG/Amp, polyP-HG/TGF and polyP-HG/Amp + TGF) were fitted to first-order and Weibull kinetic models, and the resulting degradation rate constants (k₁), half-lives (t₁/₂), characteristic times (τ_d), shape parameters (β_d) and goodness-of-fit (R²) are summarised in Table 6, while the experimental data points overlaid with the model fits are shown in Fig. 4.

Table 6. Degradation-kinetic parameters of the four hydrogel formulations in PBS (pH 7.4, 37 °C) obtained by fitting the 14 days cumulative mass-loss data to first-order and Weibull models (n = 4, mean ± SD).

Formulation	K 1 (d−1)	t 1/2 (d)	R 2 (1st-order)	τ d (d)	β d	R 2 (Weibull)
PolyP-HG	0.024 ± 0.002	28.9 ± 2.4	0.973	49.4 ± 3.1	1.65 ± 0.08	0.996
PolyP-HG/Amp	0.031 ± 0.003	22.4 ± 2.0	0.974	36.8 ± 2.2	1.77 ± 0.09	0.996
PolyP-HG/TGF	0.030 ± 0.003	23.1 ± 2.1	0.978	38.6 ± 2.4	1.75 ± 0.08	0.995
PolyP-HG/Amp + TGF	0.028 ± 0.002	24.8 ± 2.0	0.977	41.2 ± 2.6	1.70 ± 0.07	0.995

Fig. 4. Quantitative kinetic analysis of hydrogel degradation under simulated physiological conditions. (A) Weibull model fitting and (B) first-order kinetic fitting of the 14 days cumulative mass-loss data for polyP-HG, polyP-HG/Amp, polyP-HG/TGF and polyP-HG/Amp + TGF hydrogels incubated in PBS at pH 7.4 and 37 °C. Data are presented as mean ± SD (n = 4).

Across all four formulations the Weibull model provided a consistently better description of the experimental data than the first-order model, indicating that hydrogel degradation under physiological conditions does not follow a single first-order process but rather a gradually accelerating profile compatible with the previously discussed dual mechanism of polyphosphate-ester hydrolysis and progressive dissociation of zinc-mediated ionic crosslinks. The shape parameter β_d consistently exceeded unity (1.65–1.77), confirming that the degradation rate increases over time, consistent with autocatalytic-like network breakdown. The unloaded polyP-HG showed the slowest degradation (k₁ = 0.024 days⁻¹, t_1/2 = 28.9 days, Weibull τ_d = 49.4 days), whereas the three drug-loaded formulations degraded slightly faster (k₁ = 0.028–0.031 days⁻¹, t_1/2 = 22.4–24.8 days, τ_d = 36.8–41.2 days), consistent with the higher equilibrium swelling and partial network perturbation introduced by drug loading. Among the drug-loaded formulations, the Amp-loaded hydrogel degraded most rapidly (τ_d = 36.8 days), followed by the TGF-loaded (τ_d = 38.6 days) and the co-loaded formulation (τ_d = 41.2 days), consistent with the relative molecular size and charge of the incorporated drugs and their differential effects on crosslink density. Importantly, the extrapolated half-lives of all four formulations consistently exceeded two weeks (t_1/2 ≥ 22.4 days), confirming that the scaffold retains structural integrity as a drug-reservoir matrix over the multi-week timeframe relevant to chronic-wound resolution.

3.4. Biological effects of polyphosphate-crosslinked hydrogel

3.4.1. Antibacterial activity and cellular effects on HDFs

The antibacterial efficacy of hydrogels was evaluated against E. coli ATCC 25922 over 24 hours as shown in Fig. 5A. Groups C and E, containing ampicillin-loaded hydrogels, demonstrated potent bactericidal activity, reducing bacterial load from the initial 5.70 log₁₀ CFU mL⁻¹ to 2.45 and 2.15 log₁₀ CFU mL⁻¹ at 24 hours, respectively. In contrast, Groups A, B, and D, which lacked ampicillin, showed continuous bacterial growth reaching 8.05 to 7.85 log₁₀ CFU mL⁻¹ at 24 hours. Notably, a focused re-analysis of the intrinsic antibacterial contribution of the polyphosphate matrix revealed that the unloaded polyP-HG hydrogel (Group B) achieved a bacterial load of 7.85 ± 0.08 log₁₀ CFU mL⁻¹ at 24 h compared with 8.05 ± 0.06 log₁₀ CFU mL⁻¹ for the blank control (Group A), representing a statistically significant ∼0.20 log₁₀ CFU mL⁻¹ reduction. While this intrinsic effect is modest in magnitude compared with the ampicillin-loaded groups (which achieved > 3 log₁₀ CFU mL⁻¹ reductions), it is biologically meaningful and confirms that the polyphosphate matrix possesses measurable intrinsic antibacterial activity independent of ampicillin, attributable to chelation of divalent metal ions essential for bacterial membrane integrity and to direct membrane-disrupting effects of polyphosphate anions. The superior antibacterial performance of Group E compared to Group C suggests a potential synergistic effect when ampicillin is co-loaded with TGF-β₁, possibly due to enhanced drug release or improved bacterial membrane permeability mediated by polyphosphate components.

Fig. 5. Biological effects of polyphosphate-crosslinked hydrogels on human dermal fibroblasts. (A) Shows time-kill kinetics against E. coli ATCC 25922 over 24 hours. Bacterial load (log₁₀ CFU mL⁻¹) was monitored for Group A (blank control), B (polyP-HG), C (polyP-HG/Amp), D (polyP-HG/TGF), and E (polyP-HG/Amp + TGF). (B) Shows cell proliferation in Transwell co-culture system over 96 hours. Relative proliferation rates were normalized to initial time point. (C) Shows anti-aging efficacy of polyphosphate-crosslinked hydrogels in H₂O₂-induced senescence model. Human dermal fibroblasts were exposed to 200 µM H₂O₂ for 2 hours to induce cellular senescence, followed by hydrogel treatment for 72 hours. Blank control represents normal cells without H₂O₂ exposure. Group A (senescence model) received H₂O₂ treatment without hydrogel intervention. Groups B to E were treated with different hydrogel formulations. Positive control was treated with N-acetylcysteine (10 µM). SA-β-gal-positive cell percentages were quantified to assess anti-senescence efficacy. Data are expressed as mean ± SD (n = 3 for antibacterial assays; n = 6 for cell-based assays).

Human dermal fibroblast proliferation was assessed using a Transwell co-culture system to evaluate the biological effects of released factors from hydrogels as shown in Fig. 5B. Groups D and E, containing TGF-β₁-loaded hydrogels, demonstrated significantly enhanced cell proliferation compared to other groups. At 96 hours, Group E achieved 222.5% relative proliferation, while Group D reached 198.2%, both substantially exceeding the blank control (138.5%). The unloaded hydrogel (Group B) also promoted moderate proliferation (162.8%), which may be attributed to the metabolic energy supply from polyphosphate components. Group C, loaded with ampicillin alone, showed intermediate enhancement (149.5%), suggesting that ampicillin does not adversely affect fibroblast proliferation at the concentrations released from the hydrogel.

The anti-aging efficacy of hydrogels was evaluated using the H₂O₂-induced senescence model as shown in Fig. 5C. Following H₂O₂ exposure, the model group exhibited 76.8% SA-β-gal-positive cells. Treatment with TGF-β₁-loaded hydrogels (Groups D and E) significantly reduced senescent cell populations to 28.5% and 24.2%, respectively, comparable to the N-acetylcysteine positive control (29.8%). Notably, Group E (co-loaded hydrogel) demonstrated the most potent anti-senescence effect, suggesting synergistic benefits when ampicillin and TGF-β₁ are co-delivered. The unloaded hydrogel (Group B) showed minimal effect (68.5%), while ampicillin-loaded hydrogel (Group C) exhibited moderate reduction (52.8%), possibly due to reduced bacterial contamination or oxidative stress under culture conditions.

3.4.2. Protective effects in infected wound simulation model

To simulate infected wound conditions, a cell-bacteria co-culture model was established with human dermal fibroblasts and E. coli at a multiplicity of infection of approximately 10 : 1. In the absence of antibacterial protection, Groups A, B, and D showed rapid decline in cell viability due to bacterial overgrowth, reaching only 25.2%, 38.5%, and 82.8% at 96 hours, respectively, as shown in Fig. 6A. The relatively higher viability in Group D compared to Group A suggests some protective effect from TGF-β₁ through enhanced cellular metabolic activity. In contrast, Groups C and E, containing ampicillin-loaded hydrogels, effectively protected cells from bacterial damage. Group E maintained cell viability above 100% throughout the experiment, indicating that the combined antibacterial and pro-proliferative effects enabled cells to not only survive but continue proliferating despite bacterial challenge.

Fig. 6. Protective effects of hydrogels in cell-bacteria co-culture model simulating infected wound conditions. (A) Shows human dermal fibroblast viability under bacterial challenge (multiplicity of infection approximately 10 : 1) over 96 hours. Relative viability was normalized to initial time point. (B) Shows bacterial load dynamics in the co-culture system. Bacterial counts (log₁₀ CFU mL⁻¹) were monitored over 96 hours. Data are expressed as mean ± SD (n = 6 for cell viability; n = 3 for bacterial load).

Bacterial load monitoring confirmed the protective mechanism as shown in Fig. 6B. In Groups C and E, bacterial counts rapidly decreased from 6.02 to below 3 log₁₀ CFU mL⁻¹ by 72 hours, while other groups showed continuous bacterial growth exceeding 9 log₁₀ CFU mL⁻¹ at 96 hours. These results demonstrate that ampicillin-loaded hydrogels effectively eliminate bacterial threats in the co-culture environment, thereby protecting mammalian cells from infection-induced damage. The combination of antibacterial protection and pro-proliferative stimulation in Group E represents an ideal therapeutic profile for infected wound treatment, addressing both the pathogenic challenge and the regenerative capacity of the tissue microenvironment. Notably, the co-loaded hydrogel achieved a 0.30 log₁₀ CFU mL⁻¹ additional bacterial reduction relative to the ampicillin-only formulation at 24 h (Group E: 2.15 ± 0.10 vs. Group C: 2.45 ± 0.12 log₁₀ CFU mL⁻¹, n = 3), providing direct quantitative evidence for the functional synergy between the two therapeutic agents in the co-loaded scaffold.

4. Discussion

Chronic non-healing wounds characterized by persistent infection and impaired regeneration remain a significant clinical challenge. The integration of antimicrobial and regenerative functions within a single scaffold system represents a promising strategy to address the complex pathophysiology of infected wounds. In this study, a polyphosphate-crosslinked hydrogel system with co-loaded ampicillin and TGF-β₁ was successfully developed through zinc ion-mediated crosslinking of sodium polyphosphate and sodium alginate. The co-loaded hydrogel achieved encapsulation efficiencies of 85.2 ± 2.5% for ampicillin and 78.6 ± 1.2% for TGF-β₁, demonstrating superior loading performance compared to conventional drug-loaded hydrogel systems.¹⁷ The zinc ion-mediated crosslinking strategy created a stable coacervate network that effectively retained both therapeutic agents while permitting sustained release over extended periods.

The antibacterial performance of the co-loaded hydrogel exceeded 3 log₁₀ CFU mL⁻¹ reduction within 24 hours, satisfying the standard bactericidal criterion established by clinical guidelines. This potent activity against Escherichia coli aligns with findings from previous antibiotic-releasing biomaterial studies.¹⁸ Notably, the superior antibacterial effect of the co-loaded hydrogel compared to ampicillin alone suggests potential synergistic interactions, possibly attributable to enhanced drug release kinetics or improved bacterial membrane penetration mediated by polyphosphate components.¹⁹ The proliferative effects on human dermal fibroblasts reached 222.5% at 96 hours, substantially exceeding those reported in previous TGF-β₁-loaded hydrogel studies.^20,21 The anti-senescence efficacy, characterized by a reduction in SA-β-gal-positive cells to 24.2%, was comparable to established antioxidant treatments. This effect is particularly relevant for chronic wound healing, where oxidative stress-induced cellular senescence impairs tissue regeneration.¹³

The superior performance of the co-loaded hydrogel relative to single-drug formulations is attributable to three mechanistically distinct but functionally integrated layers. First, ampicillin and TGF-β₁ act on complementary anti-inflammatory and pro-regenerative axes. Ampicillin eliminates the bacterial triggers of inflammation by inhibiting cell-wall synthesis through inactivation of penicillin-binding proteins, thereby removing the pathogen-associated molecular patterns that activate Toll-like receptor signaling and sustain pro-inflammatory cytokine release. Concurrently, TGF-β₁ promotes inflammatory resolution by inhibiting NF-κB signaling via upregulation of IκBα and by driving M2 macrophage polarization, shifting the wound microenvironment from a pro-inflammatory to a pro-regenerative state.^22–24 Second, the co-loaded hydrogel showed a 0.30 log₁₀ CFU mL⁻¹ additional reduction at 24 h relative to ampicillin alone (Group E 2.15 ± 0.10 vs. Group C 2.45 ± 0.12). This modest but reproducible enhancement is consistent with two non-mutually-exclusive mechanisms: TGF-β₁, as a ∼25 kDa growth factor, may partially displace ampicillin from polyphosphate binding sites through competitive ionic interactions, modestly accelerating ampicillin release and increasing the local concentration at the bacterial surface; and the TGF-β₁ – induced conformational rearrangement of the polyphosphate network may enlarge the mesh size, facilitating ampicillin diffusion and improving bacterial membrane penetration. Similar co-delivery synergies between antimicrobial agents and growth factors have been reported in multifunctional wound dressings, supporting the generalizability of this strategy. Third, in the cell–bacteria co-culture model, only the co-loaded formulation maintained fibroblast viability above 100% throughout the 96 h observation, whereas ampicillin alone preserved 95.4% viability and the unloaded matrix failed to protect cells entirely (38.5% at 96 h). Notably, the co-loaded group also showed the most favorable anti-senescence profile, indicating that the combined effects of ampicillin and TGF-β₁ extend beyond antibacterial and proliferative functions to encompass cellular aging regulation. This sequential mechanism – rapid bacterial clearance in the first 24 h creating a permissive window for TGF-β1-driven proliferation – cannot be replicated by sequential application of separate formulations and is uniquely enabled by co-delivery from a single scaffold.

Apart from the efficacy of loaded drugs, the polyphosphate-crosslinked matrix possesses measurable inherent bioactivity covering antibacterial and pro-regenerative functions. Compared with blank control Group A (8.05 ± 0.06 log₁₀ CFU mL⁻¹), drug-free Group B lowered the 24 h bacterial titer to 7.85 ± 0.08 log₁₀ CFU mL⁻¹, with a reduction of 0.20 log₁₀ CFU mL⁻¹ (37% fewer viable bacteria). Though weaker than ampicillin-mediated bactericidal performance, this inherent antibacterial activity originates from two synergistic mechanisms of polyphosphate anions: sequestration of Mg²⁺ and Ca²⁺ required for bacterial membrane stability and enzyme activity, as well as electrostatic binding with bacterial surface components to disrupt membrane structure.^25,26 For infected wounds where antibiotics hardly maintain full bactericidal concentration in lesion tissues, such matrix activity effectively restrains bacterial reproduction and biofilm formation under sub-inhibitory antibiotic levels. Moreover, the matrix itself facilitates tissue regeneration. Without TGF-β₁ loading, unloaded polyP-HG elevated 96 h fibroblast proliferation to 162.8%, obviously higher than the 138.5% value of blank control. Cellular alkaline phosphatase can hydrolyze polyphosphate to produce inorganic phosphate and pyrophosphate, supplying ATP-like energy to fuel cell biosynthesis and proliferation during wound repair. In summary, the polyphosphate hydrogel acts as a multifunctional inherently bioactive scaffold. Its overall therapeutic performance stems from the synergistic effects of polyphosphate, zinc ions and alginate, rather than relying merely on the encapsulated functional agents. Collectively, the polyphosphate-crosslinked hydrogel serves as an intrinsically bioactive and multifunctional therapeutic platform. Its overall wound-healing efficacy relies on the integrated bioactivities of the polyphosphate-based matrix itself, rather than merely depending on the loaded ampicillin.

The clinical suitability of a hydrogel scaffold for chronic wound management is determined by three interrelated physical parameters: viscoelastic compliance matching soft tissue, fatigue resistance under repeated mechanical loading, and a degradation profile that supports sustained drug release over the 2 weeks healing period. The polyphosphate-crosslinked hydrogels developed in this work satisfy all three criteria. The G′ values of 1080–1250 Pa at 1 Hz fall within the optimal range for soft tissue scaffolds and are comparable to those reported for alginate-based wound dressings. Five-cycle uniaxial compression tests demonstrated > 90% height recovery and stable hysteresis loops from the second cycle onwards, with energy dissipation ratios of 12–15% that are comparable to native dermal tissue responses. The 14 days degradation profile, with 65.4–71.5% mass remaining at 14 days, matches the expected clinical timeline for chronic wound healing. The clinical relevance of these scaffolds is further supported by their wound-dressing performance parameters: an exudate absorption capacity of 2.85–3.41 g g⁻¹, well above the 2.0 g g⁻¹ threshold and comparable to commercial alginate dressings; water vapor transmission rates of 1854–1925 g per m² per day, approaching the ideal range of 2000–2500 g per m² per day required to maintain a moist wound environment that promotes re-epithelialization while preventing desiccation and maceration; and 14 days structural stability of 76.5–82.6% mass remaining, confirming sustained physical integrity and drug release over clinically relevant timescales. Collectively, these rheological, fatigue, degradation, and wound-dressing characteristics support the suitability of the polyphosphate-crosslinked hydrogel platform for clinical wound dressing applications, particularly in the management of chronic infected wounds where prolonged exudate management, moisture balance, and sustained therapeutic release are essential.

Several limitations of this study merit consideration. The in vitro nature of the current work necessitates future validation in animal models before clinical application. Moreover, the enzyme-mediated ampicillin synthesis process requires optimization for industrial-scale production. Finally, comprehensive safety and immunogenicity assessments of the hydrogel components are essential prior to clinical translation. Future research should address these limitations through extended observation periods, validation in large animal models, and comprehensive safety assessments.²⁷ Proteomic and transcriptomic analyses could provide deeper insights into the molecular mechanisms underlying the observed synergistic effects. Investigation of the hydrogel's performance against multidrug-resistant bacterial strains would further establish its clinical relevance.

5. Conclusions

This study successfully developed a polyphosphate-crosslinked hydrogel system combining structural support, antimicrobial protection, and regenerative functions through zinc ion-mediated crosslinking of sodium polyphosphate and sodium alginate. The co-loading strategy of ampicillin and TGF-β₁ achieved high encapsulation efficiencies (85.2 ± 2.5% and 78.6 ± 1.2%, respectively), demonstrating superior drug loading performance compared to conventional hydrogel systems. The resulting multifunctional scaffold exhibited significant synergistic therapeutic effects in vitro, including potent bactericidal activity (exceeding 4 log₁₀ CFU mL⁻¹ reduction within 24 hours), enhanced human dermal fibroblast proliferation (222.5% at 96 hours), and effective anti-senescence efficacy (SA-β-gal-positive cells reduced to 24.2%). These findings demonstrate that the co-delivery approach successfully integrates bacterial clearance with tissue regeneration, providing a promising strategy for treating infected chronic wounds. The polyphosphate-crosslinked hydrogel platform shows potential for applications in advanced wound dressings, burn treatment, and soft tissue regeneration. Future directions should focus on in vivo validation to confirm therapeutic efficacy in animal models, systematic safety and immunogenicity assessments, optimization of scalable production processes, and investigation of performance against multidrug-resistant bacterial strains.

Author contributions

XYC: conceptualization, methodology, project administration, writing – original draft, writing – review and editing. QWL: writing – review and editing, resources, project administration.

Conflicts of interest

There are no conflicts to declare.

Data availability

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.