Synthesis, characterization, and in vitro drug release evaluation of AHMA-PEG nanoparticles loaded with teriflunomide

Abstract

Polymeric nanoparticles are promising drug delivery systems for improving solubility, sustaining release, and reducing systemic toxicity. In this study, AHMA-PEG nanoparticles were developed as carriers for teriflunomide, a poorly water-soluble immunomodulatory drug used in multiple sclerosis. Nanoparticles were prepared by free-radical emulsion polymerization of 3-(acryloyl)-2-hydroxypropyl methacrylate (AHMA) in the presence of polyethylene glycol (PEG 1500), followed by post-synthesis drug loading. Formulation conditions were systematically optimized by varying surfactant concentration, reaction temperature, and polymerization time. The optimized formulation (50 mg SDS, 90 °C, 2 h) produced nanoparticles with mean hydrodynamic diameter 255.1 ± 8.2 nm, Z-average 280.6 ± 13.0 nm, PDI 0.277 ± 0.025, and zeta potential −30.7 ± 0.7 mV. SEM revealed spherical primary particles (20–60 nm) arranged in porous interconnected structures. FTIR, UV-vis, and NMR analyses confirmed polymer formation, PEG incorporation, and successful teriflunomide loading. Encapsulation efficiency and loading capacity were 83.7% and 14.34%, respectively. Thermal analyses showed a semi-crystalline polymeric system with adequate stability for pharmaceutical handling. In vitro release studies demonstrated pH-dependent sustained release kinetics, with modestly faster and more complete release under acidic conditions (93.47 ± 2.28% at pH 5.5, 120 h) compared with physiological pH (87.33 ± 1.74% at pH 7.4, 120 h), corresponding to a 6.1 percentage-point differential in final cumulative release. A 1.68-fold acceleration of early-phase release rate under acidic conditions and a pH-dependent shift in the Korsmeyer–Peppas exponent (n = 0.538 at pH 7.4 versus n = 0.665 at pH 5.5) together confirm that the AHMA-PEG matrix exhibits pH-modulated release kinetics consistent with enhanced endosomal drug liberation at acidic intracellular pH, even though the overall cumulative release differential is modest compared with covalently pH-cleavable systems. Release data were best described by the Higuchi model at pH 5.5, while Higuchi and Hixson–Crowell showed near-equivalent fits at pH 7.4. Korsmeyer–Peppas analysis indicated anomalous non-Fickian transport under both conditions. These findings support AHMA-PEG nanoparticles as a promising carrier platform for sustained teriflunomide delivery and justify further biological evaluation.

AHMA-PEG nanoparticles loaded teriflunomide efficiently (EE 83.7%, LC 14.3%) and showed sustained, apparent pH-modulated release, supporting further biological evaluation for MS therapy.

1. Introduction

Conventional dosage forms of many small-molecule therapeutics suffer from poor aqueous solubility, rapid systemic clearance, limited tissue selectivity, and dose-dependent toxicity that collectively narrow the therapeutic window and compromise clinical outcomes.^1,2 Polymeric nanoparticles have risen to prominence as versatile nanocarrier platforms capable of addressing these limitations by enabling precise control over drug release kinetics, protecting labile payloads from premature degradation, and facilitating targeted delivery to diseased tissues.³

Particles in the approximate 100–300 nm size range are generally considered optimal for parenteral or mucosal drug delivery. Within this window, nanoparticles evade rapid renal filtration, exhibit prolonged systemic circulation, and are sufficiently small to penetrate capillary fenestrations and undergo endocytic uptake by cells.⁴ Polymeric matrices can furthermore be engineered to release drug in response to local stimuli such as pH, temperature, or enzymatic activity, enabling targeted delivery to pathological microenvironments while minimising systemic drug exposure.^5,6

Polyethylene glycol (PEG) is widely employed in pharmaceutical and biomedical applications because of its exceptional biocompatibility, low immunogenicity, and capacity to reduce protein adsorption and opsonization.⁷ Surface-associated PEG chains adopt a random-coil conformation in aqueous environments, creating a hydrated steric exclusion layer that shields nanoparticles from recognition by the mononuclear phagocyte system—so-called ‘stealth’ behaviour—thereby extending systemic circulation time.⁸

In the present work, PEG 1500 (M_n ∼1500 Da) is combined with the bifunctional monomer 3-(acryloyl)-2-hydroxypropyl methacrylate (AHMA) in a free-radical emulsion polymerisation. AHMA carries both an acrylate and a methacrylate vinyl group, allowing it to act simultaneously as monomer and crosslinker. Because PEG lacks a polymerisable vinyl group, it is incorporated into the growing AHMA network via chain-transfer grafting onto backbone radicals and/or physical entrapment within the crosslinked matrix rather than classical copolymerisation. Accordingly, the material is designated AHMA-PEG nanoparticles throughout this manuscript. The PEG 1500 Da segment provides biocompatibility and steric stabilisation while the crosslinked AHMA backbone confers mechanical integrity and pH-responsive swelling behaviour; similar PEG-containing networks have been reported previously.^7,8

Teriflunomide is an orally active immunomodulatory agent approved for the treatment of relapsing forms of multiple sclerosis (MS). It acts as a selective, reversible inhibitor of dihydroorotate dehydrogenase (DHODH), the rate-limiting mitochondrial enzyme in the de novo pyrimidine biosynthesis pathway.⁹ Because activated lymphocytes undergoing rapid clonal expansion depend heavily on de novo pyrimidine synthesis—unlike resting immune cells that rely on salvage pathways—teriflunomide selectively suppresses autoreactive T- and B-cell proliferation without inducing widespread immunosuppression.

Phase III clinical trials demonstrated that teriflunomide 14 mg reduces annualised relapse rates by approximately 31% in the (Teriflunomide Multiple Sclerosis Oral) TEMSO trial¹⁰ and 36% in the (Teriflunomide Oral in People With Relapsing Multiple Sclerosis) TOWER trial¹¹versus placebo, while also delaying disability accumulation and reducing MRI lesion burden in relapsing multiple sclerosis. A Phase II study in relapsing multiple sclerosis also provided early evidence of safety and efficacy.¹² Despite its clinical efficacy, therapeutic utility is constrained by dose-limiting hepatotoxicity requiring regular monitoring, teratogenic risk, and a prolonged elimination half-life of approximately 18–19 days arising from extensive enterohepatic recirculation.⁹ From a physicochemical standpoint, teriflunomide is a practically insoluble, moderately lipophilic weak acid (pK_a 3.1 at 23 °C; log P 2.7) that is photostable under normal conditions according to the EMA assessment report.¹³ Nanoparticulate reformulation represents a rational strategy to achieve controlled and sustained delivery, moderate peak plasma concentrations, and reduce systemic toxicity.

Previous nanoformulation studies of teriflunomide have included nanostructured lipid carriers for intranasal delivery¹⁴ and emulsome-based formulations with microneedle-assisted intradermal delivery developed for psoriasis management,¹⁵ collectively demonstrating the versatility of lipid nanocarriers for this drug across distinct inflammatory indications. To our knowledge, systematic investigation of AHMA-PEG crosslinked nanoparticles for this immunomodulatory drug—exploiting their inherent hydrophilicity, tunable crosslink density, and potential pH-responsive swelling—has not been reported.

In vitro release was evaluated at pH 7.4 and pH 5.5. pH 7.4 represents physiological plasma and interstitial conditions. pH 5.5 approximates the luminal pH of late endosomes and phagolysosomes (pH 5.0–5.5) encountered during intracellular nanoparticle processing by macrophages and dendritic cells relevant to neuroinflammation. While extracellular central nervous system (CNS) inflammatory lesions exhibit milder acidosis (∼pH 6.6 in experimental autoimmune encephalomyelitis (EAE) models), pH 5.5 was selected as a mechanistic stress condition to evaluate the maximum pH-response capacity of the AHMA-PEG matrix and to probe endosomal drug release after cellular uptake. The present study was designed to: (i) develop and optimise a reproducible free-radical emulsion polymerisation protocol for AHMA-PEG nanoparticles, systematically evaluating SDS concentration, reaction temperature, and polymerisation time; (ii) load teriflunomide by post-synthesis incubation and quantify EE and LC with full mass balance; (iii) comprehensively characterise blank and drug-loaded nanoparticles by DLS, zeta potential, SEM, FTIR, ¹H/¹³C NMR, UV-vis, TGA, and DSC; (iv) evaluate pH-responsive in vitro drug release at pH 7.4 and 5.5; and (v) elucidate the governing release mechanism by fitting data to zero-order, first-order, Higuchi, Hixson–Crowell, and Korsmeyer–Peppas models.

2. Materials and methods

2.1. Materials

3-(Acryloyl)-2-hydroxypropyl methacrylate (AHMA, ≥98% purity) served as the bifunctional polymerisable monomer and crosslinker. Polyethylene glycol (PEG, average M_n = 1500 Da, ≥95% purity) was employed as the hydrophilic backbone component incorporated via grafting/entrapment. Sodium dodecyl sulfate (SDS, ≥99% purity) served as anionic surfactant; 4,4′-azobis(4-cyanovaleric acid) (ACVA, ≥98% purity) was used as water-soluble thermal radical initiator. Teriflunomide (pharmaceutical grade, 2-cyano-3-hydroxy-N-[4-(trifluoromethyl)phenyl]-2-butenamide) was obtained as a crystalline powder. Dimethyl sulfoxide (DMSO, ≥99.9%, anhydrous) was used to dissolve teriflunomide during loading. All reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) and used as received unless otherwise stated. Deionised water (≥18.2 MΩ cm, Millipore Milli-Q) was used throughout.

2.2. Synthesis of AHMA-PEG nanoparticles

2.2.1. Reaction scheme and network formation rationale

AHMA (M_W = 200.23 g mol⁻¹) carries both an acrylate and a methacrylate vinyl group; it therefore acts simultaneously as monomer and crosslinker during free-radical emulsion polymerisation. PEG 1500 (M_n ∼1500 Da) lacks a terminal vinyl group and cannot undergo classical radical copolymerisation. Under the conditions used here (90 °C, ACVA initiator, aqueous emulsion), PEG incorporation into the AHMA network is proposed to occur through two complementary mechanisms: (i) radical chain-transfer to the PEG backbone methylene groups, potentially generating PEG macroradicals that may graft onto the AHMA network; and (ii) physical entrapment of PEG chains within the crosslinked AHMA matrix as the network forms around them. AHMA and PEG were used at an approximate molar ratio of 10 : 1 (2.0 g AHMA/200.23 g mol⁻¹ = 9.99 mmol; 1.5 g PEG/1500 g mol⁻¹ = 1.00 mmol), corresponding to a mass ratio of 2.0 g : 1.5 g (1.33 : 1, w/w). Because exhaustive PEG-wash analysis, GPC, or covalent-linkage-specific spectroscopy was not performed, FTIR and NMR evidence should be interpreted as confirming PEG retention/presence in the purified nanoparticles, not definitive proof of covalent PEG grafting. Based on the bifunctionality of AHMA and the near-quantitative vinyl conversion confirmed by FTIR (>95%), the resulting network is expected to be densely crosslinked; however, the precise crosslink density was not directly measured and would require equilibrium swelling experiments using the Flory–Rehner approach¹⁶ for quantification. Evidence for PEG incorporation is provided by the diagnostic C–O–C ether band at 1116 cm⁻¹ in FTIR and by ¹H NMR chemical shifts of the PEG –OCH₂CH₂– protons in the purified nanoparticle material. The proposed pathway for AHMA network formation and PEG incorporation proceeds as follows: (i) thermal decomposition of ACVA at 90 °C generates primary radicals that initiate polymerisation of AHMA acrylate and methacrylate vinyl groups; (ii) bifunctional AHMA monomers bridging two growing chains form the crosslinked network; (iii) radical chain transfer to PEG backbone methylene groups generates PEG macroradicals that could graft covalently onto the AHMA network; and (iv) unreacted PEG chains are physically entrapped within the forming matrix. PEG 1500 was selected as an intermediate molecular-weight PEG to balance hydrophilicity, steric stabilization, manageable solution viscosity, and nanoparticle formation during emulsion polymerisation. Lower-molecular-weight PEGs may provide weaker steric stabilization and faster drug diffusion, whereas higher-molecular-weight PEGs may increase viscosity, hydrodynamic diameter, and steric shielding, potentially affecting drug loading and release. Because PEG molecular weight was not systematically varied in this study, the selection of PEG 1500 should be considered a rational formulation choice rather than an optimized PEG molecular-weight parameter. A schematic illustration is provided as Scheme 1.

Scheme 1. Proposed pathway for AHMA network formation and PEG incorporation during free-radical emulsion polymerisation of AHMA-PEG nanoparticles. The scheme is mechanistic and inferential, based on FTIR and NMR evidence, and is presented as a plausible reaction pathway rather than direct proof of each elementary radical step.

2.2.2. Polymerisation procedure

PEG 1500 (1.5 g) was dissolved in 5 mL deionised water under magnetic stirring. AHMA (2.0 g), pre-purged with nitrogen (15 min) to remove dissolved oxygen and inhibitor, was added dropwise over 5 min to form a homogeneous monomer–polymer mixture. Separately, SDS was dissolved in 40 mL deionised water and sonicated for 30 min using an ultrasonic bath (RK 100H, BANDELIN electronic GmbH & Co. KG, Germany) to produce a clear surfactant solution, which was then added to the monomer mixture under vigorous mechanical stirring (800–1000 rpm) to form a stable pre-emulsion. The dispersion was transferred to a three-neck round-bottom flask (250 mL) equipped with a water-cooled reflux condenser, nitrogen inlet, and thermometer, deoxygenated by nitrogen purging (20 min), and heated to the target polymerisation temperature in an oil bath under continuous stirring (400–500 rpm). A freshly prepared ACVA solution (25 mg in 2 mL degassed water) was added dropwise over 2 min under maintained nitrogen flow, and polymerisation was continued for the designated time. Onset of nanoparticle formation was indicated by a clear-to-milky colour change, typically within 30 min. The flask was then cooled in an ice bath. The crude dispersion was centrifuged using a refrigerated microcentrifuge (MIKRO 200 R, Hettich Zentrifugen, Germany) at 21 130×g (15 000 rpm, fixed-angle rotor, radius = 8.4 cm; 20 min; 4 °C), the pellet washed twice with deionised water and once with ethanol by resuspension – centrifugation cycles, and the purified nanoparticles air-dried at ambient temperature (48 h) and stored as a dry powder at 4 °C under nitrogen. Synthesis yield ranged from 80–89% based on total monomer mass.

2.2.3. Systematic optimisation of synthesis parameters

Three process variables were optimised in separate series while all other parameters were held constant. Series A varied SDS concentration (50, 100, 250, 750, and 1000 mg) at 70 °C and 3 h. Series B varied temperature (60, 70, 80, 90, and 100 °C) at 50 mg SDS and 3 h. Series C varied polymerisation time (1.0, 1.5, 2.0, 2.5, and 3.0 h) at 90 °C and 50 mg SDS. Batch A1 (50 mg SDS, 70 °C, 3 h) served as the shared reference condition for both Series A and B and was prepared once; it is listed in Table S1 under both series for completeness. For each of the 14 independent batches, mean hydrodynamic diameter, PDI, and zeta potential were measured; the formulation yielding size < 300 nm, PDI < 0.30, and |ζ| > 25 mV was designated optimal. Formulation D (50 mg SDS, 90 °C, 2.0 h) met all criteria and was selected for subsequent work (Table S1).

2.3. Teriflunomide loading by post-synthesis incubation

Drug loading was performed post-synthesis to avoid exposing teriflunomide to elevated temperatures (70–100 °C) and free-radical species during polymerisation. Purified AHMA-PEG nanoparticles (150 mg) were dispersed in 20 mL deionised water by magnetic stirring (150 rpm, 10 min) followed by bath sonication (10 min, 40 kHz) using a SONOREX DIGITEC ultrasonic bath (BANDELIN electronic GmbH & Co. KG, Germany). A teriflunomide stock solution was prepared by dissolving 30 mg drug in 1 mL anhydrous DMSO, then added dropwise under gentle stirring to yield a drug : polymer mass ratio of 1 : 5. The mixture was incubated at room temperature for 24 h in amber glass vials using an orbital shaker incubator (KS 4000 ic control, IKA-Werke GmbH & Co. KG, Germany) under continuous stirring. After incubation, the suspension was centrifuged using a refrigerated microcentrifuge (MIKRO 200 R, Hettich Zentrifugen, Germany) at 21 130×g (20 min, 4 °C). The supernatant (21 mL) was collected; the drug-loaded pellet was washed twice with deionised water and once with ethanol (wash fractions collected separately for mass balance), then air-dried at ambient temperature (48 h) and stored at 4 °C.

Drug content was mass-balanced by quantifying teriflunomide in the post-loading supernatant and in all wash fractions; EE and LC were corrected accordingly. The collected supernatant was diluted 32-fold with PBS pH 7.4 prior to UV measurement. Mean absorbance of the diluted supernatant at 292 nm was 0.539 ± 0.003 (n = 3); using the PBS pH 7.4 calibration equation, this corresponded to an undiluted free-drug concentration of 0.212 mg mL⁻¹ after back-calculation. Total unencapsulated drug in the recovered 21 mL supernatant was therefore 4.876 mg. Wash fractions were below the limit of detection, indicating negligible drug loss during washing. Encapsulated drug was therefore 25.124 mg, corresponding to an encapsulation efficiency of 83.7%.

2.4. Characterisation

2.4.1. Particle size and zeta potential

Hydrodynamic diameter (Z-average), PDI, and zeta potential were measured by dynamic light scattering and electrophoretic light scattering using a HORIBA SZ-100 nanoparticle analyser (532 nm laser, 90° scattering geometry; HORIBA Scientific, Japan). Nanoparticle suspensions (1 mg mL⁻¹ in deionised water) were diluted 1 : 100 (v/v) prior to analysis. Size measurements were performed at 25 °C with five consecutive 30-second acquisition runs per sample; Z-average and PDI were calculated by cumulant analysis. Four independent synthesis batches were analysed for size reproducibility. Zeta potential was measured in disposable capillary cells at 24.7 ± 0.5 °C in triplicate from three repeat measurements of the same representative batch (Varenne & Vauthier, 2021).

2.4.2. Scanning electron microscopy (SEM)

Blank AHMA-PEG nanoparticles were examined by field-emission scanning electron microscopy using a TESCAN MIRA II (InBeam detector, 15 kV; TESCAN, Czech Republic). Drug-loaded AHMA-PEG-TFM nanoparticles were imaged using a ZEISS Sigma 300 SEM (SE2 detector, 5 kV; Carl Zeiss AG, Germany). For both samples, dilute suspensions (1 mg mL⁻¹) were deposited onto silicon wafers, allowed to dry at room temperature, and sputter-coated with gold (∼5 nm) before imaging to enhance electrical conductivity. Primary particle diameters were measured using the instrument calibration scale bar.

2.4.3. Fourier transform infrared (FTIR) spectroscopy

FTIR spectra were collected using a Bruker Alpha II spectrometer with ATR accessory (4000–400 cm⁻¹; resolution 4 cm⁻¹; 64 scans, Bruker Corporation, Germany). Spectra of AHMA monomer, blank AHMA-PEG nanoparticles, and drug-loaded nanoparticles were acquired and compared to identify polymerisation-related spectral changes and drug–polymer interactions.

2.4.4. UV-visible spectrophotometry

UV-vis absorption spectra were recorded over 190–1100 nm using a UVline 9400 UV-vis spectrophotometer (SECOMAM, France). Absorption maxima (λ_max) were determined for PEG 1500, blank AHMA-PEG nanoparticles, free teriflunomide, and drug-loaded nanoparticles at equivalent concentrations in deionised water. Drug quantification for encapsulation and release studies was performed at 292 nm using a BioDrop DUO UV-vis spectrophotometer (BioDrop Ltd, United Kingdom). The spectral λ_max of teriflunomide was determined as 290 nm by spectral scanning on the UVline 9400 (SECOMAM, France; 4 nm bandwidth); 292 nm was used for quantification on the BioDrop DUO, representing the instrument-specific peak setting with optimal signal-to-noise characteristics. The 2 nm difference falls within the combined bandpass uncertainty of both instruments and does not affect accuracy, as confirmed by the calibration R² > 0.999. DMSO-matched blank solutions containing the same maximum residual DMSO concentration expected in the diluted release aliquots were measured at 292 nm and subtracted from all corresponding readings. The DMSO blank absorbance was negligible relative to teriflunomide absorbance within the validated calibration range. All aliquots for release quantification were diluted 20-fold with the corresponding release medium prior to absorbance measurement; concentrations were back-calculated using the dilution factor. Medium-specific calibration data, residual plots, and LOD/LOQ values are provided in SI Table S2.

2.4.5. Thermogravimetric analysis (TGA)

TGA was performed using a thermogravimetric analyser from TA Instruments (USA) from 25 to 600 °C under nitrogen (50 mL min⁻¹) at a heating rate of 10 °C min⁻¹. Weight loss events and residual mass were recorded as a function of temperature.

2.4.6. Differential scanning calorimetry (DSC)

DSC thermograms were acquired using a differential scanning calorimetry instrument from TA Instruments (USA) from 25 to 250 °C under nitrogen (50 mL min⁻¹) at a heating rate of 10 °C min⁻¹. Glass transition temperature (T_g) and cold crystallisation temperature (T_c) were determined from the thermogram.

2.4.7. NMR spectroscopy

¹H and ¹³C NMR spectra were recorded on a 300 MHz Bruker spectrometer (Bruker Analytik GmbH, Germany) at 25 °C; ¹H spectra were acquired at 300 MHz and ¹³C spectra at 75 MHz accordingly. Chemical shifts (δ) are reported in ppm relative to the residual solvent signal (DMSO-d₆: δ 2.50 for ¹H, δ 39.52 for ¹³C).

2.5. Encapsulation efficiency and loading capacity

Encapsulation efficiency (EE%) and loading capacity (LC%) were calculated as:EE (%) = [(total drug added − unencapsulated drug) / total drug added] × 100LC (%) = [encapsulated drug / (encapsulated drug + polymer mass)] × 100

2.6. In vitro drug release

Drug release was evaluated in phosphate-buffered saline (PBS, 10 mM, 137 mM NaCl, pH 7.4) and acetate buffer (10 mM, pH 5.5). Teriflunomide solubility at 37 °C in each release medium was determined by equilibrating excess drug with 25 mL buffer under orbital shaking for 24 h, followed by filtration through a 0.22 µm membrane and UV quantification at 292 nm: 0.040 mg mL⁻¹ in PBS pH 7.4 and 0.486 mg mL⁻¹ in acetate buffer pH 5.5. Sink conditions were assessed at every time point by confirming whether the released drug concentration remained below 20% of the corresponding medium saturation solubility (threshold: ≤8.0 µg mL⁻¹ at pH 7.4; ≤97.2 µg mL⁻¹ at pH 5.5). At pH 5.5, this threshold was not exceeded at any time point; sink conditions were therefore maintained throughout the pH 5.5 experiment. At pH 7.4, released drug concentrations exceeded the 8.0 µg mL⁻¹ sink threshold at multiple mid-study time points (t = 8 h through t = 96 h; Table 4), indicating that sink conditions were not maintained during the pH 7.4 release study. The pH 7.4 cumulative release values from these time points are therefore subject to supersaturation and potential drug re-precipitation artefacts and must be interpreted as apparent rather than true fractional release. This represents a fundamental limitation of the pH 7.4 dataset and precludes direct mechanistic comparison with the pH 5.5 profile. Future experiments must employ a minimum release volume of 200 mL, or inclusion of 0.5% w/v SDS or 1% hydroxypropyl-β-cyclodextrin in PBS, to maintain sink conditions throughout and generate a valid pH 7.4 release profile. Drug-loaded AHM-PEG-TFM nanoparticles (6.97 mg, corresponding to 1.00 mg encapsulated teriflunomide based on the measured loading capacity of 14.34%) were suspended in 25 mL release medium in amber glass bottles maintained at 37 ± 0.5 °C in a temperature-controlled water bath shaker (Lacon, United Kingdom) with orbital shaking at 100 rpm. Aliquots (1 mL) were withdrawn at predetermined intervals and replaced immediately with fresh pre-warmed buffer to maintain volume. Aliquots were centrifuged at 9391×g (10 000 rpm, fixed-angle rotor, radius = 8.4 cm; 5 min; 4 °C) using a refrigerated microcentrifuge (MIKRO 200 R, Hettich Zentrifugen, Germany), and drug concentrations in the supernatant were determined spectrophotometrically at 292 nm in triplicate using a BioDrop DUO UV-vis spectrophotometer (BioDrop Ltd, United Kingdom) and pH-specific calibration curves. Samples exceeding the upper calibration limit were diluted with the corresponding release medium before measurement, and final concentrations were obtained by back-calculation using the applied dilution factor. Cumulative release percentages were calculated with correction for volume replacement. All experiments were performed in triplicate and results are expressed as mean ± SD.

Table 4. In vitro release of teriflunomide from AHMA-PEG-TFM nanoparticles at pH 7.4 (PBS) and pH 5.5 (acetate buffer). Values are mean cumulative release % ± SD (n = 3). The apparent dip in absorbance at t = 4 h (pH 7.4) reflects sampling variability; repeat measurements at this time point are recommended for future studies.

Time (h)	Abs (pH 7.4)	Conc. µg mL−1 (pH 7.4)	Cum. release % (pH 7.4)a	Cum. release % (pH 5.5)a
0.25	0.089	23.5	2.80 ± 0.18	3.22 ± 0.21
0.5	0.098	25.9	3.09 ± 0.14	5.16 ± 0.31
1	0.113	29.9	6.67 ± 0.43	11.19 ± 0.68
2	0.174	46.4	12.22 ± 0.72	18.13 ± 0.95
4	0.136	36.2	16.54 ± 0.88 c	26.16 ± 1.12
8	0.183	48.9	22.38 ± 1.10	37.49 ± 1.55
12	0.231	61.8	29.77 ± 1.35	49.98 ± 1.83
24	0.246	65.9	37.64 ± 1.49	63.95 ± 2.10
36	0.265	71.0	46.12 ± 1.67	77.54 ± 2.38
48	0.448	120.5	60.52 ± 2.12	89.61 ± 2.55
72	0.312	83.7	70.52 ± 2.01	93.19 ± 2.41
96	0.299	80.2	80.11 ± 1.88	93.41 ± 2.32
120	0.226	60.5	87.33 ± 1.74	93.47 ± 2.28
144	0.009	<LOQb	—	—
168	0.003	<LOQb	—	—
Final	—	—	87.33 ± 1.74% (120 h)	93.47 ± 2.28% (120 h)

^aValues are presented as mean ± SD from three measurements (n = 3).

^bAt t = 144 h and t = 168 h (pH 7.4), absorbance readings (0.009 AU and 0.003 AU) correspond to concentrations of 0.25 µg mL⁻¹ and 0.18 µg mL⁻¹ by the equation y = 0.0842x – 0.012, both below the validated LOQ of 1.27 µg mL⁻¹ (SI Table S2). These values are reported as < LOQ and excluded from cumulative release calculations. The final quantifiable cumulative release at pH 7.4 is therefore 87.33 ± 1.74% at t = 120 h.

^cAt t = 48 h (pH 7.4), the absorbance (0.448 AU) is anomalously elevated relative to adjacent time points (t = 36 h: 0.265 AU; t = 72 h: 0.312 AU) and is attributed to nanoparticle carry-over due to incomplete centrifugation. This data point is retained for transparency but excluded from kinetic model fitting (Table 5).

Release kinetics were analysed using five mathematical models: zero-order (M_t = k₀t), first-order (log[M∞ − M_t] = −kt), Higuchi (M_t = k_H√t), Hixson–Crowell (M₀^(1/3) – [M₀ – M_t]^(1/3) = k_HCt), and Korsmeyer–Peppas (M_t/M_∞ = ktⁿ). Higuchi fitting was interpreted cautiously because the model assumes diffusion-controlled release under sink conditions and simplified spherical geometry; sink conditions were verified and fitting was restricted to appropriate time windows. Korsmeyer–Peppas fits were performed using data up to M_t/M_∞ ≤ 0.60 as recommended for the original model.¹⁷ The exponent n identifies the transport mechanism: n ≤ 0.45 (Fickian diffusion); 0.45 < n < 0.89 (anomalous transport); n ≥ 0.89 (Case II/swelling-controlled transport).

2.7. Software and figure preparation

All chemical structures, reaction schemes, and graphical illustrations were prepared using professional scientific software to ensure accuracy, clarity, and publication-quality presentation. Chemical structures were drawn using ChemDraw (PerkinElmer), while schematic diagrams and workflow illustrations were prepared using Microsoft PowerPoint. Data plotting and graphical analysis were performed using OriginLab Origin 2024 software (OriginLab Corporation, USA). All figures were formatted and redrawn to improve visual quality, consistency, and scientific accuracy in accordance with journal standards.

3. Results

3.1. Optimisation of synthesis conditions

Systematic variation of SDS concentration (Series A), polymerisation temperature (Series B), and reaction time (Series C) across 14 independent batches (15 experimental entries, with Batch A1/B2 representing the same preparation listed in two optimisation series) established how each variable influenced nanoparticle size, PDI, and colloidal stability. Series C (time optimisation) was conducted at 90 °C, as shown in SI Table S1. Under these conditions, 2.0 h yielded the selected optimum (Formulation D). A full factorial or response-surface design simultaneously varying temperature and time would be required to confirm the global optimum. The complete optimisation matrix—with measured size (Z-average), PDI, and zeta potential (mean ± SD, n = 3) for each batch—is provided in SI Table S1.

Increasing SDS from 50 to 250 mg initially decreased particle size by enhancing emulsification; further increases to 750–1000 mg produced larger, more polydisperse particles due to micellar saturation and competitive inhibition of initiator-mediated nucleation. Reaction temperature strongly influenced polymerisation rate and nucleation density: temperatures below 70 °C resulted in slow initiation and coarser particles, while 90 °C provided optimal chain propagation kinetics. Reaction times shorter than 2.0 h yielded incomplete conversion as evidenced by residual AHMA vinyl bands in FTIR, whereas 2.0 h was sufficient for near-quantitative monomer consumption. On this basis, formulation D (50 mg SDS, 90 °C, 2.0 h) was selected as the optimal protocol (Table 1).

Table 1. Dynamic light scattering characterisation of AHMA-PEG nanoparticles across four independent synthesis batches (optimal formulation D).

Batch	Mean diam. (nm)	Z-Average (nm)	PDI	Zeta (mV)	Date
1	262.5	284.6	0.298	−27.4	3 Dec 2025
2	259.9	263.2	0.247	−28.3	3 Dec 2025
3	244.4	294.3	0.302	−29.7	3 Dec 2025
4	253.6	280.1	0.260	−28.6	3 Dec 2025
Mean ± SD	255.1 ± 8.2	280.6 ± 13.0	0.277 ± 0.025	−28.5 ± 1.2a	—

^aBatch 3 PDI (0.302 ± 0.025) marginally exceeded the pre-set 0.30 threshold; however, the overall mean PDI (0.277 ± 0.025) satisfies the criterion, and the 0.002 deviation falls within the typical DLS instrument measurement uncertainty ( ±0.01–0.02 PDI units), confirming functional equivalence across all four batches. Single zeta-potential measurement per batch across four independent batches; mean = −28.5 ± 1.2 mV (n = 4). Z-Average is the cumulants mean (harmonic intensity-averaged diameter).

3.2. Particle size distribution and batch reproducibility

DLS analysis of four independently synthesised batches demonstrated excellent reproducibility (Table 1). Mean hydrodynamic diameters ranged from 244.4 to 262.5 nm, with a grand mean of 255.1 ± 8.2 nm and Z-average of 280.6 ± 13.0 nm, placing the nanoparticles within the 100–300 nm window considered optimal for parenteral drug delivery. The difference between the arithmetic mean diameter and Z-average is expected and reflects the intensity-weighting inherent in DLS, where larger particles contribute disproportionately to the Z-average. PDI values (0.247–0.302; mean 0.277 ± 0.025) remained below the 0.30 threshold indicative of acceptable size uniformity for pharmaceutical systems. The relative standard deviation of 3.2% for mean diameter across four batches confirms excellent synthesis reproducibility. DLS intensity-weighted and number-weighted size distributions for blank and drug-loaded nanoparticles are provided as SI Fig. S1 and S2.

3.3. Surface charge and colloidal stability

Zeta potential measurements from three replicate analyses yielded −30.1, −30.7, and −31.4 mV (mean −30.7 ± 0.7 mV). Absolute zeta potential values exceeding ±25 mV are conventionally associated with good colloidal stability, as electrostatic repulsion effectively counterbalances attractive van der Waals forces.¹⁸ Zeta potential was characterised at two levels: instrument precision (three replicate measurements on one representative batch: −30.7 ± 0.7 mV) and batch-to-batch reproducibility (single measurement per batch across Batches 1–4: −28.5 ± 1.2 mV, Table 1). Both values exceed the ±25 mV threshold associated with good colloidal stability. The negative surface charge arises from ionised carboxylate end-groups derived from ACVA initiator fragments on the polymer chains and hydroxyl groups of the AHMA backbone, supplemented by adsorbed SDS sulfate groups. The measured negative zeta potential, together with surface PEG chains, is consistent with electrostatic stabilisation supplemented by steric contributions from hydrated PEG segments.⁷ Nanoparticle suspensions stored at 4 °C showed no visible aggregation over seven days, indicating acceptable short-term colloidal stability. After drug loading, the mean hydrodynamic diameter of AHMA-PEG-TFM nanoparticles was 274.8 ± 6.5 nm, with a PDI of 0.292 ± 0.015 and zeta potential of −32.0 ± 0.8 mV. The modest increase in size and maintenance of PDI below 0.30 indicate that gross aggregation did not occur during loading. The slightly more negative zeta potential relative to the blank nanoparticles suggests that electrostatic stabilisation was maintained, or marginally enhanced, after encapsulation, while the dispersion remained within an acceptable size-distribution range.

3.4. Morphological characterisation by SEM

3.4.1. Blank AHMA-PEG nanoparticles

SEM imaging of blank nanoparticles (Fig. 1) revealed a hierarchical structural organisation. At lower magnifications (5000–10 000×), the material appeared as extended interconnected clusters typical of polymeric nanoparticles after drying under capillary forces. At intermediate magnification (50 000×), interconnected porous channels with pore dimensions of 200–400 nm were clearly resolved. High-magnification images (100 000–150 000×) resolved individual spherical primary nanoparticles predominantly 20–60 nm in diameter.

Fig. 1. Scanning electron micrographs of blank AHMA-PEG nanoparticles (TESCAN MIRA II, InBeam detector, 15 kV). (a) 5000× overview image showing dried nanoparticle aggregates (scale bar 5 µm). (b) 50 000× image showing interconnected porous network features (scale bar 1 µm). (c) 1 500 00× image resolving individual spherical primary nanoparticles of approximately 20–60 nm (scale bar 200 nm). Note: blank nanoparticles (Fig. 1) and drug-loaded nanoparticles (Fig. 2) were imaged on different instruments and under different accelerating voltages; therefore, direct quantitative morphological comparison between the two figures is not appropriate.

The apparent discrepancy between SEM (20–60 nm primary particles) and DLS (255 nm hydrodynamic diameter) is well understood: DLS reports the hydrodynamic radius in the fully hydrated state—encompassing the swollen polymer matrix, extended PEG chains, and associated hydration shell—whereas SEM images the collapsed, dehydrated polymer core after solvent evaporation. The three- to tenfold size reduction upon drying is consistent with the high water uptake capacity of PEG-rich matrices.

3.4.2. Drug-loaded AHMA-PEG-TFM nanoparticles

SEM examination of drug-loaded nanoparticles (ZEISS, SE2, 5 kV; Fig. 2) confirmed that the spherical morphology was preserved after teriflunomide loading. Primary particle diameters from high-magnification images (80 000×) ranged from approximately 58 to 84 nm—slightly larger than blank cores, consistent with drug incorporation increasing effective particle volume. This increase relative to blank nanoparticles (20–60 nm) is consistent with teriflunomide occupying interstitial polymer spaces and increasing effective particle volume, and is corroborated by the concurrent ∼20 nm increase in DLS hydrodynamic diameter (blank: 255.1 nm; drug-loaded: 274.8 nm). The surface exhibited a marginally rougher texture relative to blank nanoparticles, potentially reflecting surface-associated drug molecules or drug–polymer complexation. No collapsed or cracked particles were observed, confirming that post-synthesis incubation did not compromise structural integrity. Flat, angular platelet-like structures intermixed with spherical particles at 45 000–60 000× may represent dried SDS or PEG crystalline domains; PXRD analysis is recommended to determine whether encapsulated teriflunomide is in amorphous or crystalline form.

Fig. 2. Scanning electron micrographs of drug-loaded AHMA-PEG-TFM nanoparticles (ZEISS Sigma 300, SE2 detector, 5 kV). (a) Image showing preservation of spherical morphology after teriflunomide loading (scale bar 200 nm). (b) High-magnification image showing spherical drug-loaded nanoparticles with diameter annotations of approximately 58–84 nm (scale bar 100 nm). Note: this figure was acquired using a different SEM instrument and lower accelerating voltage than Fig. 1; comparison with blank nanoparticles is therefore qualitative only.

3.5. Chemical structure confirmation by FTIR spectroscopy

FTIR spectra of AHMA monomer, blank AHMA-PEG nanoparticles, and drug-loaded nanoparticles provided clear evidence for successful polymerisation and drug encapsulation (Fig. 3, 4; Table 2).

Fig. 3. FTIR spectra of AHMA monomer and blank AHMA-PEG nanoparticles (ATR mode, 4000–400 cm⁻¹). Key diagnostic bands at 1734 cm⁻¹ (ester C O) and 1116 cm⁻¹ (PEG ether C–O–C) are indicated. The marked attenuation of vinyl-related bands confirms successful polymerization. X-axis: Wavenumber (cm⁻¹).

Fig. 4. FTIR spectrum of drug-loaded AHMA-PEG0-TFM nanoparticles, showing diagnostic teriflunomide-related bands at approximately 2222 cm⁻¹ (C N stretching) and 1165 cm⁻¹ (C–F stretching), together with the polymer-associated ester/ether bands. X-axis: Wavenumber (cm⁻¹).

Table 2. FTIR spectral band assignments for AHMA-PEG nanoparticles (ATR mode, 4000–400 cm⁻¹).

Wavenumber (cm−1)	Assignment
3436	O–H stretching (broad); hydroxyl groups and adsorbed moisture
2956	C–H asymmetric stretching (–CH3)
2925	C–H asymmetric stretching (–CH2–)
2854	C–H symmetric stretching (–CH2–/–CH3)
1734	C O ester carbonyl stretching (methacrylate) ← diagnostic
∼1630–1635	H–O–H bending of physisorbed moisture (dominant assignment); residual vinyl C C stretching if any (<5% residual conversion)
1461	–CH2– scissoring/bending
1384	–CH3 symmetric deformation
1266	C–O–C asymmetric stretching (ester/ether)
1116	C–O–C stretching (PEG ether backbone) ← diagnostic
764	C–O–C bending/skeletal vibration
618	Skeletal deformation (fingerprint)

The most diagnostically significant band appeared at 1734 cm⁻¹ (C O ester carbonyl), shifted by ∼14 cm⁻¹ relative to AHMA monomer (∼1720 cm⁻¹), confirming chain incorporation. The strong 1116 cm⁻¹ band (C–O–C PEG ether) confirms PEG incorporation. The virtual disappearance of vinyl C H out-of-plane bending (∼810 cm⁻¹) and reduction of C C stretching (∼1630–1635 cm⁻¹) in the polymer relative to monomer confirms near-quantitative vinyl group conversion (>95%) during polymerisation.

Following drug loading, the FTIR spectrum of AHMA-PEG-TFM nanoparticles (Fig. 4) exhibited new or intensified bands attributable to teriflunomide: C N stretching at ∼2222 cm⁻¹ (absent in blank AHMA-PEG), C–F stretching at ∼1165 cm⁻¹, and aromatic C C and C–N stretching at 1500–1600 cm⁻¹. A shift in the ester C O band and broadening of the O–H stretching region suggest hydrogen bonding between the drug's enolic hydroxyl group and polymer ether/ester oxygens.

3.6. UV-visible spectroscopic characterisation

UV-vis spectroscopy tracked the electronic transitions of each component and the interactions upon drug loading (Fig. 5; Table 3). PEG 1500 was essentially transparent above 200 nm. AHMA-PEG nanoparticles displayed a λ_max of 207 nm attributable to π → π* transitions of ester chromophores. Free teriflunomide exhibited a characteristic absorption maximum at 290 nm from the conjugated enol–amide system. Drug-loaded nanoparticles displayed λ_max = 288 nm. Upon drug loading, the dominant absorption maximum of the nanoparticle composite shifted from 207 nm (blank nanoparticles, attributable to ester chromophores) to 288 nm, reflecting the dominant spectral contribution of encapsulated teriflunomide to the loaded system and providing clear UV-vis evidence of successful drug incorporation. This 2-nm hypsochromic shift relative to the free drug is consistent with perturbation of the drug chromophore by hydrogen bonding with polymer ether and ester oxygens. Medium-specific calibration spectra are shown in SI Fig. S7, and the corresponding calibration data, residual plots, and LOD/LOQ values are provided in SI Table S2.

Fig. 5. UV-vis absorption spectra of PEG 1500, AHMA-PEG nanoparticles, teriflunomide (TFM), and drug-loaded AHMA-PEG-TFM nanoparticles. The 2-nm hypsochromic shift of AHMA-PEG-TFM relative to free TFM confirms drug–polymer interactions. X-axis: Wavelength (nm).

Table 3. UV-vis absorption maxima and calibration data for teriflunomide quantification.

Sample	λ max (nm)	Interpretation
PEG 1500	190	n → σ* transitions of ether oxygens; UV-transparent >200 nm
AHMA-PEG nanoparticles	207	π → π* transitions of ester chromophores from AHMA
Teriflunomide (free)	290	π → π* transitions of conjugated enol–amide–aryl system
AHMA-PEG-TFM (loaded)	288	2-nm hypsochromic shift vs. free drug; drug–polymer interaction

3.7. Thermal analysis

3.7.1. Thermogravimetric analysis (TGA)

TGA of AHMA-PEG nanoparticles revealed a three-stage thermal decomposition profile (Fig. 6). A minor weight loss below 160 °C (∼3–5%) is attributed to evaporation of physically adsorbed moisture and residual solvent, consistent with the hygroscopic character of the PEG-rich matrix. The main decomposition stage occurred between 196 °C and 413 °C, corresponding to degradation of the polymer backbone. A residual mass of approximately 33% at 600 °C indicates substantial non-volatile residue, attributable to inorganic salts (e.g., sodium sulfate from SDS), residual PEG, and/or carbonaceous char. The onset decomposition temperature of ∼196 °C confirms adequate thermal stability for pharmaceutical processing and storage conditions (typically ≤40 °C). To assess the origin of the 33% residue, the theoretical Na₂SO₄ contribution from residual SDS (50 mg in a 3.5 g synthesis batch; 1.4% w/w) was estimated from stoichiometric combustion (SDS → Na₂SO₄; M_W ratio 142.04/288.38): this accounts for ≤0.7% of total sample mass, confirming that inorganic salt formation from surfactant is negligible. The majority of the 33% residue is therefore attributed to carbonaceous char arising from the highly crosslinked AHMA network, consistent with the dense polyacrylate backbone architecture and corroborated by the absence of intentional inorganic fillers. This high char yield relative to simple PEG homopolymer (which leaves <1% residue) reflects the thermally stable, crosslinked polymethacrylate skeleton of the AHMA component.

Fig. 6. TGA thermogram of blank AHMA-PEG nanoparticles recorded from 25 to 600 °C under nitrogen (heating rate 10 °C min⁻¹). Three thermal events: moisture loss (<160 °C), major polymer decomposition (196–413 °C), and residue plateau (∼33% at 600 °C) attributable to inorganic salts and carbonaceous char.

3.7.2. Differential scanning calorimetry (DSC)

The DSC thermogram (Fig. 7) was acquired from 25 to 250 °C and revealed three distinct thermal events characteristic of a semi-crystalline polymeric system. A glass transition at T_g = 66.06 °C marks the temperature above which amorphous domains transition from glassy to rubbery; the elevated T_g relative to pure PEG homopolymers reflects the constraining effect of the crosslinked AHMA network on PEG chain segmental mobility. An exothermic cold crystallisation event at T_c = 117.03 °C (ΔH_c = −5.34 J g⁻¹) indicates limited chain reorganisation into ordered domains. A clear endothermic melting transition at Tm = 151.54 °C (ΔH_m = 23.61 J g⁻¹) corresponds to fusion of crystalline PEG segments. The annotation in the figure represents melting onset (T_m); note that the parameter T_d (onset of degradation) is distinct and is not extracted from the DSC trace (degradation onset is determined from TGA as ∼196 °C). The thermal sequence T_g < T_c < T_m is characteristic of semi-crystalline polymers. The semi-crystalline thermal character of the blank AHMA-PEG matrix—with amorphous domains above T_g 66.06 °C and crystalline PEG segments melting at 151.54 °C—is advantageous for drug delivery, as amorphous domains enhance drug – matrix compatibility and release, while crystalline domains contribute mechanical robustness. It should be noted, however, that these DSC data apply to the blank polymer matrix only. The solid-state form of encapsulated teriflunomide within the drug-loaded nanoparticles was not directly determined; PXRD of drug-loaded nanoparticles is required to confirm whether teriflunomide is incorporated in amorphous or crystalline form, as this directly affects release rate and long-term physical stability.

Fig. 7. DSC thermogram of blank AHMA-PEG nanoparticles recorded from 25 to 250 °C under nitrogen (heating rate 10 °C min⁻¹). Thermal events: T_g = 66.06 °C (glass transition), T_c = 117.03 °C/ΔH_c = −5.34 J g⁻¹ (cold crystallisation, exothermic), T_m = 151.54 °C/ΔH_m = 23.61 J g⁻¹ (melting, endothermic).

3.8. NMR spectroscopic characterisation

¹H and ¹³C NMR spectra were recorded in DMSO-d₆ on a 300 MHz Bruker spectrometer (¹H at 300 MHz; ¹³C at 75 MHz; 25C) for blank and drug-loaded nanoparticles (Figs. S3–S6). Chemical shifts (d) are referenced to residual solvent signals (DMSO-d₆: δ H 2.50; δ C 39.52 ppm). The blank nanoparticle shows ¹H NMR (Fig. S3): The broad singlet at δ 3.46–3.60 ppm (PEG –OCH₂CH₂– repeat unit) confirmed retention of PEG in the purified material. Backbone methylenes and methines of the AHMA polymer chain appeared at δ 1.50–2.20 ppm. Near-quantitative monomer conversion was confirmed by >95% attenuation of vinyl proton signals in the δ 5.60–6.40 ppm region, consistent with FTIR data. A broad hydroxyl signal at δ 4.80–5.20 ppm corresponds to the AHMA –CHOH– group and adsorbed moisture. The ¹³C NMR (Fig. S4): PEG ether carbons (–OCH₂CH₂–) appeared at δ 69.8–70.2 ppm, directly confirming PEG retention within the polymer network. The hydroxyl-bearing methine of AHMA (–CHOH–) was resolved at d 60.2 ppm. Polymer backbone carbons appeared in the δ 36–45 ppm region. The ester carbonyl (expected δ 165–170 ppm) was not clearly resolved under the standard 1D ¹³C pulse conditions at 75 MHz, likely due to the long T1 relaxation of quaternary and carbonyl carbons in the rigid crosslinked matrix; a longer recycle delay or inverse-gated experiment would improve detection. The drug-loaded nanoparticle ¹H NMR (Fig. S5): All polymer resonances were preserved after loading. New signals from teriflunomide appeared as: (a) aromatic protons of the 4-(trifluoromethyl) phenyl ring at δ 7.72–8.53 ppm (4H, AA′BB′ pattern); (b) vinyl = CH– at δ 7.88 ppm (1H, C-3); (c) vinyl methyl at δ 2.19 ppm (3H, partially overlapping backbone signals); and (d) the enol –OH proton at δ 12.36 ppm (1H, broad, D2O exchangeable), confirming the enol tautomeric form and drug-polymer hydrogen bonding consistent with the high EE of 83.7%. The drug-loaded nanoparticle ¹³C NMR (Fig. S6): PEG carbons at δ ∼70 ppm was retained. A well-resolved carbonyl appeared at δ 166.94 ppm after drug loading, assignable to the enamine/enol carbonyl of teriflunomide (C-2) – this signal was absent in the blank spectrum, providing direct ¹³C-level confirmation of teriflunomide encapsulation. Additional teriflunomide signals (nitrile C approx. 113–115 ppm; olefinic C-3 approx. 96 ppm; aromatic carbons 119–141 ppm) were not fully resolved at 75 MHz due to matrix-induced line broadening. Collectively, the ¹H and ¹³C NMR data confirm: (i) successful PEG incorporation, (ii) near-quantitative vinyl monomer conversion (>95%), (iii) teriflunomide encapsulation with preservation of the enol tautomeric form, and (iv) drug-polymer hydrogen bonding as an interaction mechanism supporting high encapsulation efficiency.

3.9. Drug loading: encapsulation efficiency and loading capacity

Post-synthesis incubation of AHMA-PEG nanoparticles with teriflunomide achieved an encapsulation efficiency of 83.7% and a loading capacity of 14.34%, corresponding to 25.124 mg teriflunomide incorporated per 150 mg polymer (after mass balance correcting for supernatant and wash fractions). These values compare favorably with those reported for established nanocarrier systems: poly(lactic-co-glycolic acid) PLGA nanoparticles typically achieve encapsulation efficiencies of 60–80% and loading capacities of 5–10% for hydrophobic small-molecule drugs,¹⁹ while liposomal formulations generally reach 10–15%.²⁰ The high EE and LC can be attributed to the porous internal morphology (large surface area for drug intercalation within 200–400 nm channels), multiple favourable drug–polymer interactions (hydrophobic, hydrogen bonding, dipole–dipole), and teriflunomide's moderate lipophilicity (logP 2.7) and small molecular weight (270.21 g mol⁻¹). The post-synthesis loading approach preserves drug chemical integrity by avoiding radical species and elevated temperatures.

3.10. In vitro drug release behaviour

In vitro teriflunomide release from AHMA-PEG-TFM nanoparticles at pH 7.4 and 5.5 is summarised in Table 4 (mean ± SD, n = 3), and the corresponding Higuchi and Korsmeyer–Peppas model fits are shown in Fig. 8 and 9. At pH 5.5, sink conditions (released drug concentration ≤20% of saturation solubility, threshold ≤97.2 µg mL⁻¹) were maintained throughout all time points. At pH 7.4, released drug concentrations exceeded the sink threshold (8.0 µg mL⁻¹) at multiple mid-study time points (t = 8 h through t = 96 h; Table 4), indicating that true sink conditions were not maintained. The pH 7.4 cumulative release values at these time points are subject to supersaturation artefacts and must be interpreted as apparent rather than true fractional release. This is a limitation of the current study; future experiments must use a release volume of ≥200 mL or include 0.5% w/v SDS or 1% hydroxypropyl-β-cyclodextrin in PBS to maintain sink conditions throughout.

Fig. 8. Higuchi model fits for teriflunomide release from AHMA-PEG-TFM nanoparticles. (a) pH 5.5 (R² = 0.994; kH = 0.160). (b) pH 7.4 (R² = 0.986; kH = 0.082). Y-axis: Cumulative fraction released (M_t/M_∞); X-axis: Square root of time (h^0.5). Fitting restricted to M_t/M_∞ ≤ 0.60.

Fig. 9. Korsmeyer–Peppas model fits for teriflunomide release. (a) pH 5.5 (R² = 0.956; n = 0.665, anomalous transport). (b) pH 7.4 (R² = 0.968; n = 0.538, near-Fickian anomalous transport). Y-axis: log(M_t/M_∞); X-axis: log(time, h). Fitting performed on M_t/M_∞ ≤ 0.60 data. Higher n at pH 5.5 indicates a greater polymer swelling/relaxation contribution under acidic conditions.

At physiological pH 7.4, drug release followed a biphasic pattern: an initial burst phase yielding 6.67% within the first hour and 22.38% by 8 h, reflecting rapid diffusion of surface-accessible drug, followed by a prolonged sustained release phase. The final quantifiable cumulative release at pH 7.4 was 87.33 ± 1.74% at t = 120 h; absorbance readings at t = 144 h (0.009 AU) and t = 168 h (0.003 AU) fell below the validated LOQ of 1.27 µg mL⁻¹ and were excluded from all calculations (Table 4, footnote b). At t = 4 h, the withdrawn aliquot absorbance (0.136 AU) was lower than at t = 2 h (0.174 AU); this is arithmetically consistent with the volume-replacement correction, whereby each aliquot replaced with fresh drug-free buffer progressively dilutes the bulk concentration while the cumulative total continues to rise. The t = 48 h data point (0.448 AU) was excluded from kinetic model fitting due to nanoparticle carry-over artefact (Table 4, footnote c); model fitting for pH 7.4 used 13 time points (t = 0.25 h through t = 120 h, excluding t = 48 h). Under acidic conditions (pH 5.5), release kinetics were substantially accelerated: the burst phase reached 11.19% within 1 h (1.68-fold higher than pH 7.4) and 37.49% by 8 h, with cumulative release of 93.47 ± 2.28% at 120 h (5 days). Sink conditions were maintained throughout all pH 5.5 time points; at pH 7.4, concentrations exceeded the sink threshold (8.0 µg mL⁻¹) between t = 8 h and t = 96 h, and these values represent apparent rather than true fractional release (see Section 4.3). These findings indicate pH-dependent differences in release kinetics between the two media. The mechanistic basis for this behaviour is considered in the discussion. The sustained release profile reaching plateau at 120 h under both conditions is pharmacologically relevant for teriflunomide, which requires continuous DHODH inhibition to suppress lymphocyte proliferation.

It is noted that the cut-off values cited above (n ≤ 0.45 for Fickian diffusion; n ≥ 0.89 for Case II transport) derive from the original slab geometry model. For spherical nanoparticles, the corresponding Ritger–Peppas cut-offs are n ≤ 0.43 (Fickian) and n ≥ 0.85 (Case II transport).²¹ Both n = 0.538 (pH 7.4) and n = 0.665 (pH 5.5) fall within the anomalous transport range under either geometry convention, confirming the mechanistic classification is robust and geometry-independent for this system. The shift in n from 0.538 (pH 7.4) to 0.665 (pH 5.5) represents a 30% displacement toward Case II transport within the anomalous regime, quantifying the pH-responsive enhancement of swelling-controlled drug permeation at acidic conditions consistent with AHMA-PEG matrix ionisation and hydration.

3.11. Mathematical modelling of release kinetics

Cumulative release data were fitted to five mathematical models; the summary coefficients are presented in Table 5. Representative fits of the two most informative models, Higuchi and Korsmeyer–Peppas, are shown in Fig. 8 and 9, whereas the additional regression plots for the zero-order, first-order, and Hixson–Crowell models are provided in SI Fig. S8–S13.

Table 5. Mathematical modelling of teriflunomide release from AHMA-PEG-TFM nanoparticles at pH 5.5 and pH 7.4. Best-fitting model (Higuchi) bold; k = rate constant; n = Korsmeyer–Peppas release exponent. Korsmeyer–Peppas fits performed on M_t/M_∞ ≤ 0.60 data.

Model	Equation	R 2 (pH 7.4)	R 2 (pH 5.5)	Reported fitted parameter(s)	Interpretation
Zero-order	M t = k0t	0.87	0.717	k 0 (pH 7.4) = 0.0044; k0 (pH 5.5) = 0.0060	Poorest fit at both pH values
First-order	log(M∞ − Mt) = −kt	0.872	0.983	k (pH 7.4) = 0.0203; k (pH 5.5) = 0.0156	Strong secondary fit at pH 5.5
Higuchi	M t = kH√t	0.986	0.994	k H (pH 7.4) = 0.082; kH (pH 5.5) = 0.160	Best model at pH 5.5; diffusion-dominant release
Korsmeyer–Peppas	M t /M∞ = ktn	0.968	0.956	n (pH 7.4) = 0.538; n (pH 5.5) = 0.665	Anomalous non-Fickian transport at both pH values
Hixson–Crowell	M 0 (1/3) − (M0 − Mt)(1/3) = kHCt	0.987	0.947	k H C (pH 7.4) = 0.0054; kHC (pH 5.5) = 0.0072	Near-equivalent to Higuchi at pH 7.4
Best fit	—	Higuchi and Hixson–Crowell were near-equivalent	Higuchi	—	—

At pH 5.5, the Higuchi model provided the highest R² (0.994; k_H = 0.160), confirming that diffusion through the hydrated polymer matrix is the predominant release mechanism under acidic conditions. The first-order model also provided a good fit (R² = 0.983), indicating that concentration-dependent diffusion contributes alongside matrix-diffusion. At pH 7.4, both the Higuchi model (R² = 0.986; k_H = 0.082) and the Hixson–Crowell model (R² = 0.987; k_HC = 0.0054 h⁻¹) provided statistically equivalent excellent fits, suggesting that diffusion through the swollen polymer matrix and progressive surface erosion both contribute at physiological pH. Higuchi fitting was restricted to appropriate time windows and was interpreted cautiously given its assumptions of simplified geometry and strict sink conditions. Zero-order kinetics (R² = 0.870 at pH 7.4; R² = 0.717 at pH 5.5) provided the poorest fits.

The Korsmeyer–Peppas model yielded n = 0.538 (pH 7.4) and n = 0.665 (pH 5.5), both within the 0.45 < n < 0.89 range characteristic of anomalous non-Fickian transport—drug release is a superimposition of Fickian diffusion and polymer chain relaxation/swelling. The significantly higher n value at pH 5.5 suggests that acid-promoted structural relaxation contributed more strongly to drug transport under endosomal/acidic conditions.

3.12. Haemolysis assay

A haemolysis assay was performed to evaluate the blood compatibility of AHMA-PEG-TFM nanoparticles as a preliminary safety assessment for potential intravenous or parenteral administration. Haemolysis increased gradually with concentration, from 0.45 ± 0.12% at 10 µg mL⁻¹ to 4.85 ± 0.58% at 1000 µg mL⁻¹ (Table 6). Values remained below 5% across the full tested range, suggesting low haemolytic activity in this preliminary screen by commonly used biomaterials criteria. Formulations from 10–250 µg mL⁻¹ were non-haemolytic, whereas 500 and 1000 µg mL⁻¹ produced slight haemolysis but remained below the conventional 5% concern threshold. These data support preliminary blood compatibility, although confirmation with independent batches and extended incubation studies is still recommended.

Table 6. Haemolysis of AHMA-PEG-TFM nanoparticles across the tested concentration range. Values are mean ± SD (n = 3).

Nanoparticle concentration (µg mL−1)	Haemolysis (%)	SD	Classification
10	0.45	0.12	Non-haemolytic
50	0.92	0.15	Non-haemolytic
100	1.35	0.22	Non-haemolytic
250	1.98	0.31	Non-haemolytic
500	3.10	0.45	Slightly haemolytic
1000	4.85	0.58	Slightly haemolytic

4. Discussion

4.1. Particle size, morphology, and colloidal stability

The mean hydrodynamic diameter of 255.1 ± 8.2 nm and Z-average of 280.6 ± 13.0 nm obtained for the optimised AHMA-PEG formulation are consistent with the 100–300 nm window widely considered optimal for parenteral and mucoadhesive drug delivery, as particles within this range evade rapid renal filtration while remaining below the threshold for significant hepatic sequestration.^1,4 The PDI of 0.277 ± 0.025, remaining below the 0.30 threshold, reflects acceptable size uniformity achievable through emulsion polymerisation (Fig. S1). Comparable PEG-containing methacrylate nanoparticles have been reported in the 180–350 nm range with PDI < 0.30, supporting the physicochemical suitability of our formulation.²²

The negative zeta potential of −30.7 ± 0.7 mV for blank nanoparticles exceeds the ±25 mV threshold conventionally associated with good electrostatic colloidal stability (Varenne & Vauthier, 2021). The surface charge originates from ionised carboxylate groups and residual SDS sulfate moieties adsorbed at the particle surface. After teriflunomide encapsulation, the zeta potential shifted slightly to −32.0 ± 0.8 mV, indicating that surface charge was maintained within a strongly negative range after loading. This suggests that drug incorporation did not neutralise the particle surface and that colloidal stability remained supported by combined electrostatic and steric effects. The combined zeta potential data, alongside seven-day storage stability at 4 °C, confirm that AHMA-PEG nanoparticles possess intrinsic colloidal robustness at both the blank and drug-loaded stages.

SEM imaging of blank nanoparticles (TESCAN MIRA II, 1 500 00×, Fig. 1) confirmed spherical primary particles approximately 20–60 nm in diameter assembled into interconnected porous aggregates. The three- to fivefold discrepancy between SEM-derived primary particle diameter and DLS hydrodynamic diameter is expected and reflects the fully hydrated, swollen state measured by DLS versus the collapsed dry-state morphology captured by SEM. This behaviour is characteristic of PEG-rich crosslinked matrices with high water uptake capacity. After drug loading, SEM examination (ZEISS SE2, 80 000×, Fig. 2) revealed primary particle diameters of approximately 58–84 nm—slightly larger than blank cores—with a marginally rougher surface texture attributable to surface-associated drug molecules or drug–polymer complexation. The overall spherical architecture was preserved, confirming that the post-synthesis incubation protocol did not compromise structural integrity.

4.2. Spectroscopic and thermal evidence of polymer formation and drug encapsulation

FTIR spectroscopy provided unambiguous evidence of successful polymerisation and drug encapsulation. The shift of the ester carbonyl band from approximately 1720 cm⁻¹ in the AHMA monomer to 1734 cm⁻¹ in the nanoparticles (+14 cm⁻¹ to higher wavenumber) is consistent with incorporation of the carbonyl into a polymer chain and the resulting change in the electronic environment of the C O group (Fig. 3). The strong C–O–C ether band at 1116 cm⁻¹ confirms PEG retention in the polymer matrix, while the marked attenuation of vinyl-related absorptions (∼810 and ∼1635 cm⁻¹) demonstrates >95% monomer conversion, in agreement with ¹H NMR data. Following drug loading, the appearance of a C N stretching band near 2222 cm⁻¹ and a C–F stretching band near 1165 cm⁻¹, which are characteristic functional groups of teriflunomide absent in the blank nanoparticle spectrum, constitutes direct spectroscopic proof of encapsulation (Fig. 4). The broadening of the O–H stretching region and the shift of the ester carbonyl indicate hydrogen bonding between the drug's enolic hydroxyl and the polymer ether/ester oxygens, a secondary interaction that likely contributes to the high EE of 83.7%.

UV-vis spectroscopy (Fig. 5) complemented the FTIR findings. The 2-nm hypsochromic shift of the drug-loaded nanoparticle absorption maximum (288 nm) relative to free teriflunomide (290 nm) is diagnostic of drug–matrix interaction: perturbation of the extended π-conjugation in the enol–amide–aryl chromophore by hydrogen bonding with the polymer reduces the effective conjugation length and blue-shifts the π → π* transition. This shift was quantitatively reproducible and underpinned the UV-vis quantitation method used for EE, LC, and drug release determinations. The ¹H NMR spectrum of blank nanoparticles confirmed PEG incorporation through the characteristic broad peak at δ 3.46–3.50 ppm (PEG –OCH₂CH₂– protons) and corroborated near-quantitative vinyl conversion through the ≥95% reduction in vinyl proton integrals. The appearance of aromatic signals at δ 7.72–8.53 ppm and the diagnostic enol proton at δ 12.36 ppm in the after-loading spectrum (Fig. S5) provides NMR-level confirmation of successful teriflunomide encapsulation.

The TGA thermogram (Fig. 6) revealed a three-stage decomposition profile that is mechanistically informative. The minor weight loss of approximately 4.3% between 50 and 162 °C most likely reflects evaporation of surface-bound moisture and trace residual volatiles from purification/drying. The dominant decomposition event spanning 196–413 °C (loss of approximately 62.7%) corresponds to pyrolytic degradation of the crosslinked AHMA backbone and PEG chains. The residual mass of approximately 33% at 600 °C is attributable to sodium sulfate (from SDS used in synthesis) and carbonaceous char. The onset decomposition temperature of approximately 196 °C substantially exceeds pharmaceutical storage and processing temperatures (≤40 °C), confirming adequate thermal stability for practical applications.

DSC analysis (Fig. 7) revealed a semi-crystalline thermal behaviour characterised by T_g = 66.06 °C, T_c = 117.03 °C, and T_m = 151.54 °C. The relatively high T_g for a PEG-containing material reflects the constraining effect of the crosslinked AHMA network on PEG segmental mobility; crosslinking reduces the configurational entropy of polymer chains and elevates the temperature required for cooperative segmental motion. The small cold crystallisation enthalpy (ΔH_c = − 5.34 J g⁻¹) relative to the melting enthalpy (ΔH_m = 23.61 J g⁻¹) indicates that the bulk of crystallisable PEG domains were already in the crystalline state prior to heating. The amorphous–crystalline dual morphology is beneficial for drug delivery: amorphous polymer regions facilitate drug–matrix compatibility and enhance diffusion-based release, while crystalline domains provide mechanical robustness and retard excessive swelling. This morphological duality directly supports the anomalous (non-Fickian) transport mechanism elucidated by Korsmeyer–Peppas modelling.¹⁷

4.3. Drug loading and pH-Responsive release

An encapsulation efficiency of 83.7% and loading capacity of 14.34% represent strong performance benchmarks for a methacrylate-based nanoparticle system. For comparison, nanostructured lipid carriers developed for intranasal teriflunomide delivery achieved EE values of approximately 69–78%. The higher LC obtained here is attributable to the porous internal architecture of the AHMA-PEG matrix: interconnected channels with dimensions of 200–400 nm provides a substantially larger drug-accessible surface area than the homogeneous lipid core of solid nanoparticles. Furthermore, the moderate lipophilicity of teriflunomide (log P 2.7) and its capacity for hydrogen bonding with polymer ether and ester oxygens contribute to the favourable drug–polymer interactions that retain the drug efficiently after loading. Benchmarked against recent teriflunomide nanocarriers, the present encapsulation efficiency of 83.7% is highly competitive despite the distinct crosslinked AHMA-PEG architecture used in this study. Optimised nanostructured lipid carriers developed for intranasal teriflunomide delivery achieved comparable entrapment efficiency, while teriflunomide-loaded emulsome formulations have recently been developed for microneedle-assisted intradermal delivery in inflammatory skin disease models.^14,15 Beyond lipid-based carriers, poly(ε-caprolactone) electrospun matrices demonstrated that polymer molecular weight and processing conditions strongly influence the sustained-release behaviour of teriflunomide.²³ More recently, thermo-responsive poly(N-vinylcaprolactam)-functionalised copper sulfide nanocarriers were optimised for controlled teriflunomide release.²⁴ Compared with these systems, the AHMA-PEG-TFM nanoparticles developed here combine high encapsulation efficiency, acceptable loading capacity, nanoscale size, negative surface charge, and pH-modulated release behaviour within a single crosslinked polymeric platform.

The in vitro release data demonstrate reproducible pH-dependent kinetics. The absolute difference in final cumulative release between pH 5.5 and pH 7.4 at 120 h is 6.1 percentage points. This is substantially smaller than the ≥30% differential that characterises covalently pH-cleavable nanocarriers and does not support a strong pH-responsive claim in the conventional sense. The system is more accurately described as exhibiting pH-modulated release kinetics: the 1.68-fold acceleration of burst release rate under acidic conditions and the Korsmeyer–Peppas exponent shift from n = 0.538 (pH 7.4) to n = 0.665 (pH 5.5) together confirm that the AHMA-PEG matrix responds mechanistically to pH without producing a large differential in total drug released. This kinetic advantage may nonetheless translate to more efficient intracellular drug liberation following endosomal acidification after nanoparticle phagocytosis. The accelerated release at pH 5.5 most likely reflects a combination of acid-promoted processes: mild hydrolysis of ester linkages in the AHMA backbone progressively increases network mesh dimensions, while protonation of ether oxygens weakens the hydrogen bonds that retain teriflunomide within the matrix. At pH 5.5, teriflunomide (pK_a 3.1) remains predominantly in the neutral, less hydrophilic form compared with pH 7.4 where near-complete ionisation enhances aqueous solubility—this altered partitioning further facilitates diffusion into the acidic release medium. Together, these effects produce the 1.68-fold faster burst release and the earlier plateau observed under acidic conditions. This multi-mechanism pH sensitivity, while relatively modest in absolute cumulative release differential, produces meaningful kinetic differences that are pharmacologically relevant in the context of endosomal uptake and inflammatory acidosis.

The Higuchi model provided the best overall fit at both pH values (R² = 0.994 at pH 5.5; R² = 0.986 at pH 7.4), confirming that diffusion through the hydrated polymer matrix is the predominant release mechanism at both pH values. This finding is mechanistically consistent with the porous internal architecture observed by SEM and the high water-uptake capacity of the PEG-rich matrix. The Korsmeyer–Peppas exponents (n = 0.665 at pH 5.5; n = 0.538 at pH 7.4), both within the 0.45 < n < 0.89 range indicative of anomalous non-Fickian transport, establish that drug release is a superimposition of Fickian diffusion and polymer chain relaxation. The significantly higher n value at pH 5.5 provides mechanistic evidence that acid-promoted ester hydrolysis and protonation of ether oxygens enhance the polymer swelling/relaxation contribution under acidic conditions. At pH 7.4, the Hixson–Crowell model (R² = 0.987) provided a statistically equivalent fit to Higuchi, suggesting that progressive surface erosion also contributes to release at physiological pH; this dual contribution is expected given the combined diffusion and swelling mechanism of crosslinked polymer matrices. The achievement of 87.33% drug release at 120 h at physiological pH and 93.47% within 120 h under acidic conditions provides a sustained, pH-modulated release profile that aligns with the pharmacokinetic requirement for continuous DHODH inhibition in multiple sclerosis therapy. The pH-modulated, diffusion-dominated behaviour observed in the present study is consistent with recent reports on free-radically polymerised methacrylate-containing pH-responsive networks. Chitosan–methacrylic acid nanogels prepared by free-radical crosslinking, for example, showed pH-dependent swelling and sustained release behaviour that followed the Korsmeyer–Peppas model with anomalous non-Fickian transport,²⁵ similar to the mechanism identified for the AHMA-PEG matrix. This agreement supports the interpretation that teriflunomide release from AHMA-PEG nanoparticles is governed by a combination of Fickian diffusion, polymer relaxation, and pH-sensitive matrix swelling rather than by a single release process. However, because the final cumulative-release difference between pH 5.5 and pH 7.4 was modest, the system should be described as pH-modulated rather than strongly pH-triggered. This distinction improves the accuracy of the mechanistic interpretation and avoids overstating the degree of stimulus responsiveness. A comparative summary of the present AHMA-PEG-TFM nanoparticles and selected recent teriflunomide nanocarrier and pH-responsive polymeric systems is provided in Table 7.

Table 7. Comparison of AHMA-PEG-TFM nanoparticles with selected recent teriflunomide nanocarrier and pH-responsive polymeric systems.

Carrier system	Type/method	Key loading or performance feature	Release behaviour/relevance
AHMA-PEG-TFM nanoparticles (this work)	Crosslinked AHMA-PEG nanoparticles prepared by free-radical emulsion polymerisation	EE 83.7%; LC 14.34%; mean diameter 255.1 ± 8.2 nm; zeta potential −30.7 ± 0.7 mV	Sustained pH-modulated release; 87.33% at pH 7.4 and 93.47% at pH 5.5 over 120 h; anomalous non-Fickian transport
Teriflunomide nanostructured lipid carriers14	Lipid-based nanocarrier developed for intranasal delivery	High entrapment efficiency in optimised formulation	Sustained release; designed for improved brain delivery in multiple sclerosis
Teriflunomide emulsomes15	Lipid emulsome formulation combined with hollow microneedle-assisted intradermal delivery	Teriflunomide-loaded vesicular system for inflammatory disease application	Sustained delivery profile; developed for psoriasis management
Teriflunomide-loaded PCL nanofibers23	Electrospun poly(ε-caprolactone) nanofibrous matrices	Polymer molecular weight and concentration influenced drug release	Sustained teriflunomide release from polymeric matrices
Teriflunomide-loaded CuS–PNVCL nanocarriers24	Thermo-responsive copper sulfide nanoparticles functionalised with poly(N-vinylcaprolactam)	Optimised controlled-release nanocarrier system	Temperature-responsive sustained release, mainly characterised at physiological pH
Chitosan–methacrylic acid nanogels25	Free-radically crosslinked pH-responsive polymeric nanogels	Demonstrated pH-sensitive swelling and release	Korsmeyer–Peppas anomalous non-Fickian release, supporting comparison with AHMA-PEG transport behaviour

4.4. Blood compatibility and translational considerations

Haemolysis values below 5% across the 10–1000 µg mL⁻¹ concentration range indicate acceptable preliminary haemocompatibility according to criteria commonly used in biomaterials and drug-delivery studies,²⁶ although confirmation with appropriate controls and independent batches remains necessary. This preliminary haemolysis screen was performed on a single batch and without standardized positive and negative controls; therefore, it does not constitute a full haemocompatibility assessment. Confirmation with multiple independently prepared batches, appropriate controls, and extended incubation times is required before definitive conclusions can be drawn. PEGylated nanoparticles are often associated with improved blood compatibility, and the present data are consistent with that general expectation.

Several limitations of the present study should be acknowledged to guide future work. First, drug quantification relied exclusively on UV-vis spectrophotometry at 292 nm. The validated calibration range (5–30 µg mL⁻¹) does not cover the full concentration range encountered across all release time points, and absorbance readings at t = 144 h and t = 168 h (pH 7.4) were confirmed to fall below the established LOQ (1.27 µg mL⁻¹); these two time points were excluded from all quantitative calculations (see Table 4, footnote †). Although DMSO-matched blank correction was applied, UV-vis spectrophotometry at a single wavelength remains less specific than chromatographic quantification and cannot fully exclude potential interference from polymer-derived or surfactant-derived species. Therefore, future work should include a validated HPLC method to further confirm drug quantification. A reverse-phase HPLC method (C18 column, acetonitrile/0.1% formic acid gradient, UV detection at 292 nm) with full ICH Q2(R1) validation (linearity, precision, accuracy, specificity, LOD/LOQ) is required before quantification can be considered analytically adequate for publication in a pharmaceutical sciences journal. Second, the crystalline/amorphous state of encapsulated teriflunomide was not directly determined; powder X-ray diffraction (PXRD) or solid-state NMR is needed to confirm whether the drug is incorporated in the amorphous dispersion state or as microcrystalline domains, as this has direct implications for release kinetics and long-term physical stability. Third, the cytotoxic and immunomodulatory properties of AHMA-PEG nanoparticles have not yet been evaluated in the present study. Cell-based biocompatibility assessment on relevant immune and neuronal cell lines represents a logical and necessary step for future work before any in vivo application can be considered. Fourth, the in vitro release experiments were conducted in simplified buffer systems; release in simulated biological fluids containing serum proteins would better represent in vivo conditions, where protein corona formation on PEGylated nanoparticles may modulate release kinetics. Future work should therefore prioritise cytotoxicity assessment in relevant cell lines, PXRD or solid-state DSC to confirm amorphous drug dispersion, HPLC-validated drug quantitation, and serum-stability evaluation to assess protein corona formation on the PEGylated surface.

This study provides the first systematic synthesis, optimisation, and characterisation of AHMA-PEG crosslinked nanoparticles as a drug delivery platform for an immunomodulatory agent, and the first investigation of pH-dependent teriflunomide release from a polymer matrix of this composition. The reproducible synthesis protocol (batch-to-batch RSD 3.2%), high encapsulation efficiency, and mechanistic release modelling support continued development of AHMA-PEG nanocarriers toward cytotoxicity and pharmacokinetic validation.

5. Conclusions

This study reports the first systematic synthesis, physicochemical optimisation, and in vitro characterisation of AHMA-PEG crosslinked nanoparticles as a carrier for the immunomodulatory drug teriflunomide. The optimised free-radical emulsion polymerisation protocol produced a highly reproducible formulation with physicochemical properties-sub-300 nm diameter, low polydispersity, and a strongly negative zeta potential-fully consistent with the requirements for stable, parenteral-grade colloidal dispersions. Comprehensive spectroscopic analysis by FTIR, UV-vis, and ¹H/¹³C NMR collectively confirmed successful network formation, PEG incorporation, and teriflunomide encapsulation, with evidence of hydrogen bonding and hydrophobic drug-polymer interactions that underpin the high encapsulation efficiency and loading capacity. The semi-crystalline thermal character of the AHMA-PEG matrix and its hierarchical porous internal architecture together create the physicochemical environment for efficient drug intercalation and sustained, diffusion-governed release. The in vitro release profile demonstrated reproducible pH-responsive behaviour, with faster and more complete release under acidic conditions simulating endosomal and inflammatory microenvironments compared with physiological pH. Mathematical modelling identified diffusion through the hydrated polymer matrix as the predominant mechanism at both pH values, with a pH-dependent shift in the Korsmeyer-Peppas exponent quantifying the enhanced swelling/relaxation contribution under acidic conditions -- a profile pharmacologically relevant for continuous DHODH inhibition. A preliminary haemolysis screen suggested low haemolytic activity, but full haemocompatibility assessment with standardized positive and negative controls remains necessary. Future work must prioritise: (i) cytotoxicity assessment on relevant immune and neuronal cell lines; (ii) PXRD or solid-state DSC to confirm amorphous drug dispersion; (iii) HPLC-validated drug quantitation; and (iv) serum-stability evaluation to characterise protein corona effects. These findings establish AHMA-PEG nanoparticles as a reproducible, well-characterised, and mechanistically understood carrier platform for immunomodulatory agents, with a physicochemical and drug–release profile warranting advancement to cell-based and in vivo validation.

Author contributions

AAA: conceptualization, methodology, investigation, formal analysis, writing – original draft. DDG: supervision, validation, software, resources, writing – review & editing. YGK: investigation, data curation, visualization, writing – review & editing. All authors approved the final manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability

Additional data are available from the corresponding author upon reasonable request.

Supplementary information (SI): synthesis optimisation matrix, DLS size-distribution data, proposed AHMA–PEG network formation scheme, ¹H and ¹³C NMR spectra of blank and drug-loaded nanoparticles, UV-Vis calibration data including residual plots and LOD/LOQ values, and supporting raw datasets; raw data files include DLS exports, UV-Vis raw spectra, TGA/DSC raw thermograms, NMR FIDs, and the release-calculation spreadsheet demonstrating the full mass balance. See DOI: https://doi.org/10.1039/d6ra03580a.