RSM optimization of Tetracycline-loaded PLGA nanoparticles in chitosan/HPMC hydrogels for sustained antibacterial delivery Authors: Hasti Salehi, Fatemeh Davodabadi, Seyedeh Sara Esnaashari, Masoomeh Amini Published: 2026 ABSTRACT AbstractThe rising prevalence of antimicrobial resistance and biofilm infections requires innovative drug delivery systems for sustained local antibiotic release with minimal systemic exposure. This study reports the development and characterization of novel nanocomposite hydrogels combining chitosan (CS) and hydroxypropyl methylcellulose (HPMC) matrices embedded with optimized tetracycline (TTC)-loaded poly(lactic-co-glycolic acid) (PLGA) nanoparticles (NPs) for sustained antibacterial delivery. Using a double emulsion solvent evaporation technique, TTC-loaded PLGA NPs were fabricated, and their physicochemical properties—including particle size, encapsulation efficiency % (EE), and drug loading % (DL)—were optimized via Response Surface Methodology (RSM) based on a Box-Behnken design. Optimized NPs exhibited a size of ~ 229 nm, EE of 57.2%, and DL of 0.72%. Fourier transform infrared (FTIR) spectroscopy and scanning electron microscopy (SEM) confirmed successful encapsulation and spherical morphology of the NPs. In vitro release profiles revealed biphasic TTC release from free drug-loaded CS/HPMC hydrogels (95% within 6 h; ~98% by 24 h), whereas NP-loaded nanocomposites exhibited sustained kinetics (initial burst of ~ 48% in 6 h; ~60% at 24 h). Disk diffusion assays confirmed potent antibacterial effects of nanocomposite hydrogels against both E. coli and S. aureus, superior to free drug formulations. Incorporation of optimized NPs into a CS/HPMC hydrogel matrix yielded nanocomposite hydrogels with enhanced mechanical integrity, controlled swelling behavior, and reduced degradation rates compared to hydrogels with free drug or without NPs. This platform leverages hydrogel-NP synergy for prolonged TTC delivery in localized infection management and resistance mitigation.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-55938-3.Keywords: PLGA, Tetracycline, Hydrogel, Nanocomposite, Chitosan, HPMCSubject terms: Biotechnology, Chemistry, Materials science, Nanoscience and technology Introduction The emergence of antimicrobial resistance and biofilm-forming bacteria demands advanced drug delivery systems that maintain therapeutic antibiotic concentrations at infection sites while minimizing systemic exposure. Conventional administration often leads to rapid clearance, frequent dosing, and resistance risks. Nanoparticle-infused hydrogels offer controlled release, enhanced bioavailability, and targeted delivery. These hybrids combine hydrogels’ water-retaining 3D networks with nanoparticles’ efficient drug-carrying capacity1. Among natural polymers, chitosan (CS) and hydroxypropyl methylcellulose (HPMC) have garnered significant attention as hydrogel components due to their biodegradability, biocompatibility, low cost, and versatile structural properties2. CS, a polycationic biopolymer derived from the partial deacetylation of chitin, possesses hydroxyl and amino functional groups that facilitate electrostatic interactions and contribute to its antimicrobial properties2. Studies have demonstrated that formulations with high ratios of chitosan or HPMC exhibit prolonged drug release profiles, with less than 70% of drug released over 24 h, and effective inhibition of microbial growth, including biofilm formation3. The combination of these polymers creates a synergistic matrix with tunable mechanical and biological properties suitable for sustained drug delivery applications. Poly(lactide-co-glycolide) (PLGA) nanoparticles (NPs) represent an established drug delivery system capable of encapsulating various therapeutic agents, including antibiotics4. The biodegradable and biocompatible nature of PLGA5,6, coupled with its ability to protect drugs from degradation, makes it an ideal carrier for antibacterial compounds7–9. Research has demonstrated that the properties of PLGA NPs, including their drug release behavior, can be tailored by modifying parameters such as the lactide/glycolide ratio and molecular weight10. Recent studies have shown that increasing the size of antibiotic-loaded PLGA NPs can significantly enhance antimicrobial efficacy against intracellular infections, suggesting that particle engineering plays a crucial role in optimizing therapeutic outcomes11. Tetracycline (TTC), a broad-spectrum antibiotic effective against various bacterial pathogens, has been successfully incorporated into PLGA nanofibers, demonstrating preserved antimicrobial activity and sustained release profiles10. When loaded into PLGA matrices, TTC maintains its structural stability and antibacterial efficacy, as confirmed by bacterial inhibition tests against organisms such as Staphylococcus aureus10. The entrapment efficiency and release kinetics of TTC from PLGA systems can be modulated by adjusting the LA/GA ratios, molecular weights, and drug loading concentrations, allowing for customized release patterns tailored to specific clinical applications10. Despite the individual advantages of CS/HPMC hydrogels and TTC-loaded PLGA NPs, their integration into a nanocomposite drug delivery system remains unexplored. Prior studies have explored antibiotic or drug-loaded PLGA NPs surface-modified or embedded in CS8,12,13 or HPMC matrices14, primarily for controlled release, mucoadhesion, and targeted delivery, but none combine TTC hydrochloride with a ternary PLGA-CS/HPMC composite. CS/HPMC hydrogel composites with NPs have been reported for other drugs: CS NPs crosslinked with TPP cast into HPMC films for general delivery15, HPMC/PLGA NPs for mucoadhesive sitagliptin release in gastrointestinal tract (GIT) (higher retention vs. CS alone)16, and PLGA NP-laden HPMC hydrogels for prolonged antimicrobial activity against Pseudomonas aeruginosa14. Other antibiotic systems exist, like fosfomycin in CS hydrogel with PLA microparticles (not NPs) for osteomyelitis17. No studies were found using TTC in PLGA NPs or embedding them in CS/HPMC hydrogels. Besides, the incorporation of NPs into hydrogel matrices can enhance mechanical strength and allow for the development of stimuli-responsive systems, enabling on-demand drug release triggered by external factors such as pH, temperature, or enzymatic activity1. Furthermore, the presence of PLGA NPs within CS hydrogels has been shown to influence thermal stability and structural properties of the composite material18, suggesting complex interactions that may affect drug release and therapeutic efficacy. In this study, we develop and characterize novel nanocomposite hydrogels comprising chitosan and HPMC embedded with TTC-loaded PLGA NPs for sustained antibacterial drug delivery. The Response Surface Methodology (RSM) was used to find the optimized TTC loaded PLGA NPs. Then, we systematically investigate the physicochemical properties and in vitro release kinetics of the TTC loaded PLGA NPs. The optimized NPs were embedded in a HPMC/chitosan hydrogel matrix to evaluate their mechanical stability, structure, release profile, and antimicrobial activity (shown schematically in Fig. 1). Our approach aims to combine the advantages of both hydrogel matrices and nanoparticulate drug carriers to achieve prolonged antibiotic release. This innovative platform holds significant potential for the management of localized bacterial infections, particularly those involving biofilm formation and antibiotic resistance, with applications spanning from wound management to orthopedic implants and chronic infection treatment. Steps of hydrogel nanocomposite production. Materials and methods Materials TTC hydrochloride (≥ 96.0% purity by HPLC) and PLGA (lactide: glycolide 50:50, mol wt 30,000–60,000 Da) were purchased from Sigma-Aldrich (Germany). Polyvinyl alcohol (PVA; fully hydrolyzed, MW 72,000 Da) was obtained from Sigma-Aldrich (Germany). Hydroxypropyl methylcellulose (HPMC; hydroxypropoxyl content ~ 9%, viscosity 80–120 cP, 2% in water at 20 °C) was supplied by Sigma-Aldrich (Germany). Chitosan (CS; for hydrogel preparation, medium molecular weight ~ 190–310 kDa, 75–85% deacetylation) was purchased from Sigma-Aldrich (Germany). Dichloromethane (DCM; analytical grade, ≥ 99.8%) and glacial acetic acid (analytical grade, ≥ 99.7%) were obtained from Dr. Mojallali Chemical Laboratories (Tehran, Iran). NPs preparation TTC-loaded PLGA NPs were prepared using a water-in-oil-in-water (w/o/w) double emulsion method19–21. Briefly, PLGA (50 mg) was dissolved in 4.5 mL DCM as the organic phase (O). TTC (4 mg) was dissolved in 1.5 mL water as the internal aqueous phase (W1), then slowly added to the polymer solution and homogenized for 3 min using a 70-W sonicator to form the primary emulsion. This emulsion was poured into 0.55% (w/v) PVA solution (external aqueous phase, W2) and sonicated again for 3 min. DCM was evaporated under stirring at 25 °C. NPs were collected by centrifugation (10,000 rpm, 30 min), washed three times with distilled water to remove residual PVA and free drug, lyophilized, and stored at 4 °C. Experimental design of RSM RSM was used to predict the effect of independent factors on size, encapsulation efficiency (EE) and drug loading percentages (DL) of NPs. A Box-Behnken design was used to study the effects of formulation variables on NPs properties. Statistical analysis and design were performed using Design-Expert software (version 12, Stat-Ease Inc., USA). The investigated parameters in this research were: polymer to drug mass ratio (X1), PVA concentration (X2), water to oil phase (W1/O1) volume ratio (X3) and which were classified as low, medium and high based on preliminary studies. Table 1 shows the values for selected variables. A number of responses were evaluated, including size (nm), EE%, and DL%. Table 1Box–Behnken experimental design.FactorsLevelsLow(-1)Medium (0)High (+ 1)X1Polymer/drug512.520X2PVA concentration0.1%0.55%1%X3W1/O1 volume ratio23.55 Box–Behnken experimental design. These variables were chosen from established studies showing their critical roles in emulsion stability, particle morphology, and encapsulation outcomes during double emulsion solvent evaporation methods. Specifically, polymer to drug ratio governs DL and size distribution; W1/O1 volume ratio impacts emulsion droplet formation; and PVA concentration controls surface stabilization and burst release22–26. Table 2 shows the Box-Behnken design and observational data. A Box-Behnken design comprising 17 experimental runs (12 factorial points and 5 center points) was employed for the three factors. Table 2Experimental parameters and responses for a three-level three-factorial Box-Behnken design.RunFactor 1Factor 2Factor 3Response 1Response 2Response 3A: polymer/drug ratioB: PVA concentrationC: W1/O1 volume ratioSize nmEE%DL %150.13.5162300.182200.55520870.40.16312.51242118.50.1342013.531646.40.15512.50.553.521950.70.99612.51518865.30.39712.50.553.520957.20.72812.50.553.526055.60.8912.50.1516567.30.3510200.13.519046.20.1611513.5346380.191250.55516846.80.191350.552408140.1814200.552249160.131512.50.553.523989.70.521612.50.12268140.231712.50.553.522969.60.54 Experimental parameters and responses for a three-level three-factorial Box-Behnken design. In design, a quadratic nonlinear model is generated as follows: The measured response is Y for each factor level combination; A0 is an intercept; A1–A9 are the regression coefficients; X1, X2 and X3 are the studied factors; X21, X 22, X23 are quadratic effects, X1X2 + X2X3 + X1X3 are interaction between variables and E is the error term27. Particle size and zeta potential Particle size of TTC-loaded PLGA NPs was measured after appropriate dilution across 17 runs, according to the variables determined by RSM, using dynamic light scattering (DLS; K-ONE, Korea) at 25 ± 0.5 °C14. Zeta potential was determined for four samples (lyophilized PLGA NPs without drug, lyophilized TTC-loaded PLGA NPs, suspension of TTC-loaded PLGA NPs, and suspension of PLGA NPs without drug) using a zeta sizer (SZ-100, Horiba, Japan). Fourier transform infrared spectroscopy (FTIR) FTIR (Bruker; Equinox55, Germany) was used to characterize active functional groups. PVA, PLGA, TTC drug, lyophilized TTC-loaded PLGA NPs were analyzed. Scanning electron microscopy (SEM) SEM (Mira Tescan, Czech Republic) was used to assess NPs surface morphology. NPs were spread on aluminum foil and dried at 25 °C. As part of the SEM analysis, the samples were mounted on a metal substrate and gold coated under argon atmosphere with 20 V to ensure proper surface conductivity and efficient electron flow. Determination of EE and DL EE% and DL% were determined using UV-visible spectrophotometry, with concentrations measured against a standard curve at 359 nm. EE% was calculated indirectly by quantifying the non-encapsulated drug in the supernatant after centrifugation of NP samples at 10,000 rpm for 30 min at 4 °C: Where X1 is the initial amount of the drug is added and X2 is the amount of free drug in the supernatant. DL% was determined by dissolving a known amount of lyophilized TTC-loaded PLGA NPs in dichloromethane with sonication, followed by centrifugation at 10,000 rpm for 30 min at 4 °C, and measuring the drug concentration in the supernatant: Where X1 is the amount of drug measured in the supernatant and W is weight of lyophilized NPs. Drug release kinetics analysis The in vitro release profile of TTC from optimized PLGA NPs and free drug controls was evaluated using a dialysis method. Lyophilized TTC-loaded PLGA NPs were dispersed in 1 mL of phosphate-buffered saline (PBS, pH 7.4) and placed in a 12 kDa molecular weight cutoff dialysis bag. This system was immersed in 50 mL PBS maintained at 37 °C under continuous agitation (100 rpm) using a shaker incubator (Noor Sanat Ferdous, Iran). Sampling occurred at hourly intervals for the first 6 h, followed by measurements at 24, 48, and 72 h. At each timepoint, 2–3 mL of release medium was withdrawn and immediately replaced with fresh PBS to maintain sink conditions. Drug quantification was performed via UV-Vis spectrophotometry at λ = 359 nm. All experiments were conducted in triplicate to ensure reproducibility. Hydrogel’s preparation Hydrogel film preparation Hydrogel films were fabricated using the solvent casting technique28. CS and HPMC, each at 2 wt%, were dissolved in 1 v/v% aqueous acetic acid. The CS and HPMC solutions were combined at weight ratios of 90:10, 50:50, and 10:90, stirred at ambient temperature for uniform mixing, and poured (10 mL) into 5.5 cm diameter Petri dishes. Films were dried sequentially: 25 °C for 5 h, 40 °C overnight, then 25 °C for 5 h to re-equilibrate and ensure slow solvent evaporation without bubbles (Fan Azma Gostar, Iran). Hydrogel nanocomposite preparation CS (2 wt%) was dissolved in 1 v/v% acetic acid with stirring for 12 h at 25 °C. Separately, HPMC (2 wt%) and 50 mg lyophilized TTC-loaded PLGA NPs were dispersed in distilled water. The solutions were mixed at a 90:10 (CS: HPMC) ratio, homogenized by stirring, and cast (10 mL) into 5.5 cm Petri dishes. Solvent evaporation followed the same drying regimen as above. The resulting films were frozen at − 80 °C and lyophilized. SEM analysis of hydrogel and nanocomposite SEM imaging of the plain hydrogel and nanocomposite was performed using a scanning electron microscope (Mira, Tescan, Czech Republic). The dried specimens were fractured to expose the cross-sectional morphology and were mounted on aluminum stubs using double-sided conductive carbon tape and sputter-coated with gold under argon atmosphere with 20 V to ensure proper surface conductivity and efficient electron flow. Swelling measurement Swelling behavior of three hydrogel types—nanocomposite (0.02 g), drug-loaded (0.02 g), and plain (0.02 g)—was evaluated in 10 mL phosphate-buffered saline (PBS, pH 7.4) at 37 °C. Samples were removed hourly over 6 h, blotted dry, and weighed to record swollen mass. Equilibrium swelling percentage (ES%) was calculated as29: where W1 is the initial dry weight and W2 is the swollen weight at each time point. Experiments were performed in triplicate. Weight loss analysis The degradation rate of the hydrogel was determined by measuring mass loss over time. Initially, each dried hydrogel disk (incubated at 37 °C) was weighed to obtain the initial dry weight (W0). The disks were then immersed in PBS (pH 7.4) at 37 °C for 6 h, removed, re-dried under the same conditions, and weighed to record the remaining dry weight (W1). The weight loss percentage, indicating degradation, was calculated as30: where W0 is the initial dry weight of the hydrogel, and W1 is the dry weight after degradation. Each measurement was performed in triplicate to ensure reproducibility. In vitro drug release studies of nanocomposite hydrogel and hydrogel containing free drug In vitro drug release studies were conducted for both nanocomposite hydrogel and hydrogel containing free drug. A specified amount of each hydrogel, equivalent to 1 mg of TTC, was placed in 50 mL of phosphate buffer (pH 7.4) in a shaking incubator at 37 °C and 100 rpm. At predetermined intervals (1, 2, 3, 4, 5, 6, and 24 h), 5 mL of release medium was withdrawn, replaced with fresh buffer, and analyzed for TTC concentration via UV-Vis spectrophotometry at λ = 359 nm. Assay of antibacterial activity The antibacterial activity of plain hydrogel, hydrogel containing free drug, and nanocomposite hydrogel was evaluated against Staphylococcus aureus (ATCC 25923) and Escherichia coli (ATCC 25922) using the disk diffusion method. Bacterial suspensions were adjusted to 0.5 McFarland standard (~ 1 × 108 CFU/mL), and 100 µL was spread on Mueller-Hinton agar plates, which were allowed to dry for 5 min. Hydrogel samples were cut into 5 mm-diameter disks (each containing 30 µg TTC), placed on the plates, and incubated at 37 °C for 24 h. The zones of inhibition were measured using a digital caliper31. Result and discussion PLGA NPs were developed as a biodegradable drug delivery system for TTC encapsulation. Since particle size, EE, and DL are critical quality attributes affecting stability and performance, response surface methodology (RSM) was applied to optimize the formulation. The effects of polymer/drug ratio (X1), PVA concentration (X2), and W1/O1 volume ratio (X3) were statistically evaluated using polynomial modelling and full ANOVA. The optimized formulation with the highest DL% was subsequently incorporated into a hydrogel matrix. Detailed modelling results for particle size are presented in the following section, supported by the complete ANOVA outputs and the corresponding response surface plots. Statistical modelling and ANOVA analysis of particle size The influence of polymer/drug ratio (X1), PVA concentration (X2), and W1/O1 volume ratio (X3) on particle size was investigated using RSM. Different polynomial models (linear, 2FI, quadratic, and cubic) were evaluated in order to identify the most appropriate regression model for predicting particle size. Model selection was performed based on sequential sum of squares, lack-of-fit tests, and model summary statistics. Among the tested models, the two-factor interaction (2FI) model provided the best predictive performance, with a high coefficient of determination (R² = 0.9179), an adjusted R² of 0.8687, and a predicted R² of 0.7055. In addition, the 2FI model exhibited the lowest PRESS value (30124.40), confirming its robustness and predictive capability. In contrast, the quadratic model showed poor prediction ability (Predicted R² = 0.2988), suggesting overfitting. Therefore, the 2FI model was selected for further analysis. As shown in Table 3, analysis of variance (ANOVA) confirmed that the proposed model was statistically significant (p = 0.0014). Among the main effects, PVA concentration (X2) and W1/O1 volume ratio (X3) exhibited highly significant influences on particle size (p < 0.001). Furthermore, the interaction between polymer/drug ratio and W1/O1 volume ratio (X2X3) was also significant (p = 0.0109), indicating that the effect of polymer/drug ratio on particle size depends on the internal phase volume. The lack-of-fit test was not significant (p = 0.1167), demonstrating good agreement between experimental data and model predictions. Table 3Full ANOVA for particle size response.SourcedfSum of SquaresMean SquareF-valuep-valueSignificance Model 996434.9810715.0012.780.0014 ** X1 (polymer/drug ratio)11830.121830.122.180.1831nsX2 (PVA concentration)129524.5029524.5035.200.0006 *** X3 (W1/O1 volume ratio)147586.1247586.1256.740.0001 *** X1X21841.00841.001.000.3500nsX1X319900.259900.2511.800.0109 * X2X314225.004225.005.040.0597ns (marginal)X1²1423.16423.160.500.5005nsX2²1634.42634.420.760.4133nsX3²11220.421220.421.460.2669nsResidual75870.55838.65———Lack of Fit34333.751444.583.760.1167nsPure Error41536.80384.20———*Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001, ns: not significant (p > 0.05) and marginally significant (0.05 < p < 0.1). Full ANOVA for particle size response. *Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001, ns: not significant (p > 0.05) and marginally significant (0.05 < p < 0.1). The final regression equation in terms of coded variables is expressed as: Where X1 represents polymer/drug ratio, X2 represents PVA concentration, and X3 represents W1/O1 volume ratio. The positive coefficient of factor X2 indicates that increasing PVA concentration leads to an increase in particle size, whereas the negative coefficient of factor X3 demonstrates that higher W1/O1 volume ratios significantly reduce particle size. These findings suggest that surfactant concentration and internal phase volume play dominant roles in controlling droplet stabilization and nanoparticle formation during the emulsification process. The combined effects of formulation variables on particle size are further illustrated by the three-dimensional response surface plots (Fig. 2). These plots highlight the strong influence of PVA concentration and W1/O1 volume ratio, as well as the significant interaction between polymer/drug ratio and internal phase volume, confirming the statistical outcomes obtained from ANOVA. Three-dimensional surface-response diagram showing the effect of different variables on the particle size of PLGA NPs (A) Parameter measurement of W1/O1 volume ratio and PVA concentration (B) Parameter measurement of W1/O1 volume ratio and polymer/drug ratio (C) PVA concentration and polymer/drug ratio. These findings are consistent with reports by Ding et al. (2017), who demonstrated that stabilizer concentration critically affects NP size in double emulsion systems, with higher surfactant levels generally increasing particle size due to changes in interfacial tension and viscosity32. Similarly, the use of double emulsion techniques for protein-loaded PLGA NPs has shown that emulsifier concentration and phase ratios are key determinants of particle size and EE%33. Moreover, the particle size range obtained here is comparable to other studies employing double emulsion solvent evaporation methods, where sizes typically ranged from 150 to 700 nm depending on formulation parameters and processing conditions34. Statistical modelling and ANOVA analysis of EE% A quadratic polynomial model was fitted to evaluate the influence of formulation variables on EE%. The ANOVA results confirmed that the proposed model was statistically significant (p < 0.05), indicating that the selected factors adequately explain the variation in EE%. Table 4 represents ANOVA data for EE response. The EE was adequately described by a quadratic polynomial model. The final coded equation was obtained as: Table 4Full ANOVA for Encapsulation Efficiency response.SourceSum of SquaresdfMean SquareF-valuep-valueSignificance Model 7132.979792.555.500.0175 * X1 (Polymer/drug ratio)315.011315.012.190.1828nsX2 (PVA concentration)14.31114.310.0990.7619nsX3 (W1/O1 volume ratio)4385.1614385.1630.420.0009 *** X1X215.21115.210.1060.7548nsX1X3116.641116.640.8090.3982nsX2X310.56110.560.0730.7944nsX1²878.261878.266.090.0429 * X2²418.321418.322.900.1322nsX3²746.761746.765.180.0570ns (Marginal)Residual1008.967144.14———Lack of Fit24.9938.330.0340.9904nsPure Error983.974245.99———Total8141.9316————*Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001, ns: not significant (p > 0.05) and marginally significant (0.05 < p < 0.1). Full ANOVA for Encapsulation Efficiency response. *Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001, ns: not significant (p > 0.05) and marginally significant (0.05 < p < 0.1). Among the main effects, the W1/O1 volume ratio (X3) exhibited the strongest and most significant influence on EE (p = 0.0009). In contrast, the linear contributions of polymer/drug ratio (X1) and PVA concentration (X2) were not significant within the studied experimental range. Interaction effects between factors (X1X2, X1X3, X2X3) were all non-significant (p > 0.05), suggesting that the factors mainly affected EE independently rather than synergistically. Regarding curvature effects, the quadratic term X1² showed a statistically significant contribution (p = 0.0429), indicating a non-linear relationship between polymer/drug ratio and EE. The lack-of-fit test was not significant (p = 0.9904), confirming the adequacy of the fitted model. The three-dimensional response surface plots (Fig. 3) further illustrate the combined effects of the formulation variables on EE. Three-dimensional surface-response diagram showing the effect of different variables on EE%. (A) Parameter measurement of W1/O1 volume ratio and polymer/drug ratio. (B) Parameter measurement of W1/O1 volume ratio and PVA concentration. (C) PVA concentration and polymer/drug ratio. Achieving high EE for hydrophilic small molecules such as TTC remains a significant challenge in polymeric NP systems, primarily due to their water-soluble nature, which can lead to drug leakage during formulation35. This likely explains the moderate EE observed here. Similar results were reported by Ghasemian et al. for sildenafil-loaded PLGA NPs (EE: 14–62%; DL: 2–6%) prepared via double emulsion22, and by Singh et al. for doxycycline hydrochloride-loaded PLGA NPs (EE: 75.58 ± 1.94%)36. Statistical modelling and ANOVA analysis of DL The quadratic model for DL was statistically significant (F = 4.58, p = 0.0286), with a non-significant lack-of-fit (p = 0.9363), confirming model adequacy. None of the linear or interaction terms showed significant contributions (p > 0.05). However, all quadratic terms (X1², X2², and X3²) were highly significant (p < 0.05), indicating that DL% is predominantly governed by curvature effects and the existence of an optimum region. The three-dimensional response surface plots (Fig. 4) further illustrate these nonlinear relationships, demonstrating that maximum DL occurs at intermediate levels of the formulation variables. ANOVA results for DL were presented in Table 5. The final coded equation was obtained as follow: Three-dimensional surface-response diagram showing the effect of different variables on DL%. (A) Parameter measurement of PVA concentration and polymer/drug ratio. (B) Parameter measurement of W1/O1 volume ratio and polymer/drug ratio. (C) PVA concentration and W1/O1 volume ratio. Table 5Full ANOVA for Drug Loading response.SourceSum of SquaresdfMean SquareF-valuep-valueSignificance Model 0.981390.10904.580.0286*X1 (Polymer/drug ratio)0.002510.00250.10300.7576nsX2 (PVA concentration)0.000410.00040.01890.8945nsX3 (W1/O1 volume ratio)0.022110.02210.92720.3677nsX1X20.000110.00010.00420.9501nsX1X30.000110.00010.00420.9501nsX2X30.004910.00490.20600.6636nsX1²0.450210.450218.930.0034 ** X2²0.198310.19838.340.0234 * X3²0.207510.20758.730.0213 * Residual0.166570.0238———Lack of Fit0.015030.00500.13160.9363nsPure Error0.151540.0379———Total1.1516————*Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001, ns: not significant (p > 0.05) and marginally significant (0.05 < p < 0.1). Full ANOVA for Drug Loading response. *Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001, ns: not significant (p > 0.05) and marginally significant (0.05 < p < 0.1). Numerical optimization Numerical optimization was conducted using the desirability function method in Design-Expert software. The goals were set to minimize nanoparticle size, maximize EE%, and target DL% toward its highest value. The optimal formulation achieved an overall desirability of 0.823. The predicted optimum factor levels were polymer/drug ratio (A) = 12.5, PVA concentration (B) = 0.55, and W1/O1 volume ratio (C) = 3.61. This formulation was predicted to yield a particle size of 225.6 nm, EE% of 66.2%, and DL% of 0.72. A confirmatory experiment was performed under the optimized conditions to validate the numerical optimization. The experimental results were in good agreement with the predicted values, with prediction errors of + 1.55% for particle size, − 13.63% for EE, and + 0.46% for DL (Table 6), confirming the reliability of the optimization model. This formulation was therefore incorporated into the CS/HPMC hydrogel system for further analysis. Table 6Predicted parameters and obtained values.ResponsePredicted ValueExperimental ValueError (%)Size (nm)225.574229+ 1.55%EE (%)66.22357.2−13.63DL (%)0.71670.72+ 0.46 Predicted parameters and obtained values. Characterization of NPs The particle size analysis of TTC-loaded PLGA NPs is presented in Fig. 5. DLS (Fig. 5A) revealed a single, narrow peak with an average hydrodynamic diameter of ~ 229 nm, indicating a monodisperse population in suspension. SEM analysis (Fig. 5B) showed a slightly lower mean diameter of 217 ± 16 nm. (A) DLS and (B) SEM images of TTC-loaded PLGA NPs. The discrepancy between DLS and SEM arises from their principles: DLS measures hydrodynamic diameter (including solvation layer and aggregation), while SEM assesses dried particle physical size. The larger DLS value suggests a hydration shell or minor aggregation, typical for PVA-coated NPs in aqueous media. Both methods confirm well-defined, nanoscale PLGA particles suitable for drug delivery. As shown in Fig. 5B, NPs exhibited spherical morphology, smooth surfaces, and uniform distribution—features enhancing cellular uptake and controlled release. These align with prior reports on double emulsion-synthesized PLGA NPs, which yield discrete, spherical particles with occasional agglomeration based on formulation and drying22. Aggregation of particles is generally minimized when their surface charge, measured as zeta potential, exceeds ± 30 mV, due to the strong electrostatic repulsion between similarly charged particles37. In this study, the zeta potential of lyophilized PLGA NPs was measured at -17.1 mV, which is significantly less negative compared to the − 45.1 mV observed for the PLGA NP suspension. During lyophilization, the structural rearrangement can bury ionizable groups (e.g., terminal carboxylates) within the polymer matrix, reducing their contribution to surface charge. The dehydration process induces polymer chain reorganization, potentially shielding charged moieties from the aqueous interface38 and therefore, reduce the negative charge of the PLGA NPs (Fig. 6). ZP analysis of 4 samples. (A) Lyophilized PLGA NPs without drug (-15.2 mv). (B) Lyophilized TTC-loaded PLGA NPs (-17.1 mv). (C) Suspension of TTC-loaded PLGA NPs (-42.5 mv). (D) Suspension of PLGA NPs without drug (-45.1 mv). The FTIR spectra presented for PVA, TTC, PLGA, and TTC loaded PLGA NPs (as shown in Fig. 7) demonstrate the characteristic functional groups of each component and provide clear evidence of successful drug encapsulation. The TTC spectrum (black) include N-H and O-H stretching around 3200–3400 cm⁻¹. Aliphatic C-H stretching occurs around 2990 cm⁻¹. Carbonyl (C = O) stretching also appears near 1670 cm⁻¹, in agreement with Shao et al.39. The PVA spectrum (red) displays a broad absorption band around 3300 cm⁻¹, corresponding to O–H stretching, and peaks near 2900 cm⁻¹ (C–H stretching), consistent with literature reports40. The PLGA spectrum (blue) is marked by an aliphatic C-H stretching around 2949 cm⁻¹. A strong ester C = O and C–O–C stretching also appear at 1722 cm⁻¹ and 1161 cm⁻¹, which is in agreement with the findings of Makadia & Siegel41. TTC loaded PLGA NPs show FTIR spectra dominated by PLGA ester peaks overlaid with TTC’s N-H and O-H stretching features, confirming drug encapsulation. The strong C = O ester stretching from PLGA remains prominent at 1722 cm⁻¹ and C-O-C band at 1162 cm⁻¹. Aliphatic C-H stretching occurs at 2949 cm⁻¹. TTC contributes N-H/O-H stretching around 3200–3400 cm⁻¹, visible alongside PLGA signals with minor shifts indicating physical entrapment. PVA stabilizer add weak O-H broadening at 3300 cm⁻¹, but these blend into existing features. FTIR spectroscopic analysis of TTC loaded PLGA NPs, PLGA, PVA and TTC. In vitro drug release studies The calibration curve of TTC was constructed using UV–Vis spectrophotometry by plotting absorbance at λmax against standard TTC concentrations (Fig. 8A). The calibration curve showed good linearity over the concentration range of 1.5–12.5 ppm, with a correlation coefficient (R²) of 0.9993. This curve was used to determine the TTC concentration in all subsequent samples. (A) Calibration curve of tetracycline using UV–Vis spectrophotometry (absorbance at λmax: 359 nm vs. concentration) (B) Comparison of TTC release from PLGA NPs and free drug. The in vitro release profile of TTC from PLGA NPs was compared to free TTC to assess the nanoparticulate system’s sustained release capability (Fig. 8B). Free drug showed rapid release, with ~ 100% liberated within 10 h—typical burst behavior for unencapsulated compounds. In contrast, TTC-loaded NPs exhibited a biphasic profile: ~40% release in the first 10 h, increasing gradually to ~ 70% over 72 h.This pattern, matching prior reports, features initial rapid release from NP surface regions via dissolution and diffusion, followed by slower liberation from the polymer matrix due to degradation or both42–44. The sustained profile suggests potential for maintaining therapeutic levels over time, reducing dosing frequency and peak-related side effects. Characterization of hydrogel The ratio of the polymers used (90:10. 50:50 and 10:90) was examined and CS: HPMC = 90:10 proved to form the most promising matrix. Plain hydrogel and nanocomposite hydrogel morphologies were observed using SEM (Fig. 9). Hydrogel has a smooth and porous surface, whereas nanocomposite hydrogel has a non-smooth and uneven surface with NPs distributed throughout18,45. SEM image of (A) plain hydrogel (B) nanocomposite hydrogels. The swelling behavior of three formulations—plain hydrogel, hydrogel with free TTC, and nanocomposite hydrogel (Hydrogel + NP)—was assessed over 6 h (Fig. 10A). All reached peak swelling within 2 h: plain hydrogel at ~ 5200%, free TTC hydrogel at ~ 3200%, and nanocomposite at ~ 3500%. Notably, plain hydrogel and free TTC hydrogel lost structural integrity after 3 and 4 h, respectively, preventing further swelling measurements, while nanocomposite maintained integrity up to 6 h with swelling declining to ~ 1300%. (A) Swelling properties (B) Degradation analysis of free TTC hydrogel, nanocomposite, and plain hydrogel. After 6 h, degradation rates were: plain hydrogel (~ 37%), free TTC hydrogel (~ 10%), and nanocomposite (~ 7%). PLGA NPs likely occupy matrix space, restrict chain mobility, limit water uptake, and slow degradation via hydrophobic reinforcement and gradual ester hydrolysis, contrasting faster breakdown in non-NP formulations. In vitro drug release studies of nanocomposite hydrogel and hydrogel containing free drug Figure 11 displays cumulative TTC release profiles from nanocomposite hydrogel versus hydrogel containing free drug control. Free TTC exhibited rapid burst release, achieving ~ 95% release within 6 h and complete (~ 98%) by 24 h, characteristic of diffusion-dominated kinetics from physically cross-linked matrices. In contrast, the nanocomposite showed moderated release: ~48% in 6 h (reduced burst by 47%) reaching ~ 60% at 24 h. The rapid release from free TTC hydrogel aligns with its high swelling ratio (> 5000%) and weak hydrogen-bonded networks, enabling diffusion of hydrophilic TTC within 24 h. TTC-loaded PLGA NPs integration significantly attenuated burst release (48% vs. 95% at 6 h), attributable to NPs’ hydrophobic PLGA matrix. This pattern indicates that incorporation of tetracycline into PLGA NPs effectively suppresses the burst release typically seen for the free drug and leads to a more sustained release behavior. The in vitro drug release behavior of the hydrogel formulations was evaluated over a 24 h period. This time frame was selected as it is clinically relevant for topical and wound dressing applications and allows meaningful comparison of release profiles among different formulations. The primary objective of the release study was to assess and compare the release behavior rather than to achieve complete drug saturation. The CS/HPMC hydrogel is physically crosslinked via hydrogen bonding and electrostatic interactions. Such physically crosslinked networks may undergo gradual relaxation or erosion upon prolonged exposure to aqueous media. Accordingly, the release study was limited to 24 h to preserve the structural integrity of the hydrogel matrix and ensure reliable and reproducible sampling. In vitro release profile of TTC from free TTC hydrogel and nanocomposite hydrogel at pH 7.4 and 37 °C. These findings are consistent with previous reports on TTC-loaded CS hydrogels46,47 and PLGA-based delivery systems48, where the introduction of a diffusional barrier (like nano/micro particles) lengthens the diffusion path of the drug and slows its transport into the surrounding medium. Such a two-stage release pattern, characterized by a limited initial burst followed by a gradual, diffusion-controlled phase, is particularly desirable for local antibiotic therapy, because it can maintain therapeutically relevant concentrations at the target site for an extended period while potentially reducing dosing frequency and systemic side effects49. Assay of antibacterial activity Table 7 presents zones of inhibition for TTC-loaded hydrogels against S. aureus and E. coli. Plain hydrogel and negative controls showed no zones, confirming no intrinsic activity. Although chitosan is widely reported to possess antibacterial activity, its effectiveness is strongly influenced by experimental conditions, molecular characteristics, and the mode of interaction with bacterial cells. Chitosan’s antimicrobial mechanisms predominantly involve electrostatic attraction between protonated amino groups and negatively charged bacterial membranes, leading to disruption of cell integrity50. However, when chitosan is embedded within a physically crosslinked hydrogel network, its mobility and diffusion into the surrounding agar medium in a disk diffusion assay are limited. Under these conditions, contact-dependent antimicrobial effects may not result in a visible zone of inhibition despite inherent activity, as similarly noted in other studies where chitosan-coated materials did not exhibit significant inhibition in agar diffusion tests51. Hydrogel with free TTC exhibited 22 mm against E. coli and 32 mm against S. aureus, while nanocomposite hydrogel showed 20 mm against E. coli and superior 40 mm against S. aureus. These results confirm successful TTC release from nanocomposite hydrogel, with enhanced S. aureus activity indicating cooperative NP effects for Gram-positive penetration. Findings align with studies on other TTC-loaded structures52,53. Representative images of the disk diffusion assay against S. aureus and E. coli are provided in the Supplementary Information (Supplementary Fig. S1). Table 7Zones of inhibition of TTC-Loaded Hydrogels against bacterial strains.SampleEscherichia coli Zone of inhibition (mm)Staphylococcus aureusZone of inhibition (mm)Plain hydrogelNo zoneNo zoneFree TTC hydrogel2232Nanocomposite hydrogel2040Negative controlNo zoneNo zoneZone diameterinterpretive standards a18–2224–33a Zone diameter interpretive standards (in mm) define antibiotic susceptibility and resistance; zone sizes exceeding the standard indicate microbial sensitivity to the sample. Zones of inhibition of TTC-Loaded Hydrogels against bacterial strains. Zone diameter interpretive standards a a Zone diameter interpretive standards (in mm) define antibiotic susceptibility and resistance; zone sizes exceeding the standard indicate microbial sensitivity to the sample. Conclusion This research successfully developed an optimized nanocomposite hydrogel by embedding TTC-loaded PLGA NPs in a CS/HPMC matrix. Response Surface Methodology enabled precise tuning of NP parameters, yielding desirable size, high encapsulation efficiency, and drug loading for sustained antibacterial release. The hybrid hydrogels demonstrated superior mechanical stability, controlled swelling/degradation profiles, structural integrity, and enhanced antibacterial activity compared to plain or free-drug hydrogels. In vitro assays confirmed prolonged, biphasic TTC release from NP encapsulation and hydrogel entrapment, with disk diffusion results showing nanocomposite hydrogels achieving superior zones of inhibition against S. aureus and E. coli relative to free TTC hydrogel, indicating effective drug release and NP synergy against Gram-positive bacteria. These attributes reduce dosing frequency and systemic side effects while addressing resistant and biofilm infections. Future in vivo studies on efficacy, biocompatibility, and stimuli-responsive mechanisms will enhance clinical translation. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements During the preparation of this work, the authors used AI (perplexity.ai and DeepSeek) in the writing process to improve the readability and language of the manuscript. After using this AI, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. Abbreviations Poly (lactic-co-glycolic acid) Nanoparticles Tetracycline polyvinyl alcohol Drug loading Entrapment efficiency (hydroxypropyl) methyl cellulose Chitosan Food and drug administration Response Surface Methodology Poly dispersity index Fourier transform infrared spectroscopy Dynamic light scattering Scanning electron microscopy Phosphate buffer Author contributions Design of the proposal was performed by S.S. E and H. S. The experimental section was done by H. S. and M. A. The interpretation of data, preparing the manuscript and its editing were performed by F. D., H. S. and S.S. E. All authors reviewed the manuscript. Funding All authors declare that no funds or other support were received during the preparation of this manuscript. Data availability All data supporting the findings of this study are available within the paper. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All data supporting the findings of this study are available within the paper.