From Foodborne Pollutant Carrier to Gastrointestinal Trojan Horse: Simulating the Bioaccessibility of Antibiotics Loaded on Aged Polylactic Acid Microplastics in Human Digestive System

Abstract

Foodborne microplastics (MPs) are suspected carriers of co-ingested food contaminants, yet their digestive fate remains poorly characterized. This study simulates the role of environmentally aged polylactic acid (PLA) MPs—a common food-contact material—in transporting the antibiotic tetracycline (TC) through the human gastrointestinal tract. K₂S₂O₈-induced aging significantly increased PLA surface porosity, oxygen-containing groups, and hydrophilicity, elevating TC adsorption capacity from 0.54 to 0.95 mg/g. While adsorption kinetics were consistent with pseudo-second-order behavior, mechanistic analysis indicates that aging promotes interactions dominated by hydrogen bonding and electrostatic forces, rather than purely physical deposition. Critically, in vitro digestion models revealed that simulated intestinal fluid significantly enhances TC release (up to 62.7% of adsorbed load) compared to gastric conditions. Sequential gastrointestinal simulation yielded a bioaccessibility of 32.6%, indicating substantial digestive mobilization of MP-bound antibiotics. These findings underscore the potential of aged PLA MPs to act as digestive-stage “Trojan horses” for foodborne antibiotics. Our integrated approach—combining controlled aging, adsorption thermodynamics, and physiologically relevant digestion models—provides a mechanistic screening framework for assessing the bioaccessibility and exposure potential of microplastic-vectored contaminants in food safety contexts.

1. Introduction

Microplastics (MPs) are pervasive environmental contaminants that have been consistently detected in food products and beverages, constituting a direct ingestion pathway for humans [1,2]. Their high surface-area-to-volume ratio and environmental persistence enable MPs to adsorb diverse chemical contaminants, including pharmaceuticals and heavy metals, thereby acting as potential carriers into the gastrointestinal tract [3]. This co-transport could modulate contaminant fate, as the unique physicochemical milieu of the gut—marked by shifting pH, digestive enzymes, and surfactants—may influence pollutant release from MPs [4]. However, the behavior of MPs within this complex biological environment, particularly after they undergo environmental weathering, remains inadequately characterized.

A critical factor often overlooked in dietary exposure assessments is the aging of MPs. In food-relevant environments (e.g., during packaging degradation, food processing, or environmental exposure prior to ingestion), plastics undergo photo-oxidation, hydrolysis, and mechanical abrasion [5]. These processes alter surface properties, such as increasing roughness and introducing oxygen-containing functional groups, which can enhance their affinity for co-occurring pollutants [6]. Most studies to date have focused on pristine MPs or employed simplified fluid systems, creating a disconnect between environmentally weathered particles and their subsequent behavior in physiologically relevant digestive conditions.

This knowledge gap is particularly salient for biodegradable plastics, such as polylactic acid (PLA). PLA accounts for a substantial share of bioplastic production and is widely used in food packaging, utensils, and coatings [7,8,9]. Although promoted as sustainable, PLA is susceptible to fragmentation into micro- and nano-scale particles. More importantly, its surface chemistry and degradation behavior—potentially accelerated in the gastrointestinal environment—may significantly affect its role as a contaminant vector [7]. Despite its prevalence in food-contact applications, the combined effect of environmental aging and digestive transit on the vector capacity of PLA MPs is poorly understood, representing a critical uncertainty in food safety.

Tetracycline (TC), a widely used antibiotic frequently detected in environmental and food matrices, serves as a relevant model contaminant for this study due to its complex, pH-dependent speciation and propensity for interfacial interactions [8,9,10]. The co-ingestion of TC with MPs via contaminated food or water raises concerns beyond direct toxicity, including potential impacts on gut microbial communities.

Current literature lacks an integrated, mechanistic investigation that bridges the following key aspects: (1) the controlled simulation of environmentally relevant aging of food-contact PLA MPs; (2) the quantification of contaminant adsorption onto such aged MPs; and (3) the time-resolved release (bioaccessibility) of the contaminant under sequential simulated gastrointestinal conditions that mimic human digestion (SGF followed by SIF). Prior studies have typically examined adsorption or desorption in isolation, often using pristine MPs or single digestive fluids, thereby failing to capture the continuum from environmental transformation to digestive release [11,12,13,14].

This study was designed to address these gaps by systematically evaluating the “Trojan horse” potential of aged PLA MPs for TC. We hypothesize that (a) persulfate-induced oxidative aging—serving as a controlled proxy for long-term environmental weathering—will enhance the adsorption capacity of PLA MPs for TC by modifying surface chemistry and morphology; (b) simulated intestinal fluid (SIF) will be more effective than simulated gastric fluid (SGF) in desorbing TC from aged MPs, leading to significant gastrointestinal bioaccessibility; and (c) aging will shift adsorption mechanisms toward interactions more resistant to simple desorption in aqueous media but susceptible to solubilization by gut surfactants. To test these hypotheses, we employed a three-phase approach: (i) accelerating PLA aging using K₂S₂O₄ to simulate oxidative weathering; (ii) characterizing the adsorption of TC onto aged MPs under varying environmental conditions; and (iii) quantifying TC release using a time-resolved, two-stage simulated gastrointestinal model. This integrated methodology allows for a mechanistic assessment of how aging transforms the vector function of a major food-contact bioplastic within a physiologically relevant context. Our findings aim to provide a quantitative framework for screening the digestive bioaccessibility of MP-vectored contaminants, thereby informing more robust risk evaluation in food safety.

2. Materials and Methods

2.1. Materials and Reagents

TC (purity > 98%) was purchased from Aladdin Industrial Corporation (Shanghai, China). A stock solution (200 mg/L) was prepared and stored at 4 °C in darkness to prevent photodegradation. Polylactic acid (PLA, 1.26 g/cm³) MPs (Average size: 100 μm; D10: 72 ± 8 μm, D50: 101 ± 5 μm, D90: 138 ± 10 μm; irregular shape; No additives; density: 1.26 g/cm³) were supplied by Dongguan Zhangmukou Plastic Co., Ltd. (Dongguan, China). Pepsin (35 kDa), bovine serum albumin (BSA, ≥98%, 66 kDa), and sodium taurocholate (NaT, ≥97%) were also obtained from Aladdin Co. (Shanghai, China). Formic acid (HPLC, ≥99.9%), NaOH, K₂S₂O₈ and HCl were also purchased from Macklin Co. (Shanghai, China). The other reagents were purchased from Guangzhou Chemical Reagent Factory (Guangzhou, China). All solutions were prepared using ultrapure water (18.2 MΩ·cm).

2.2. Artificial Aging of PLA MPs and Physicochemical Characterization

An accelerated oxidative aging protocol was employed to simulate the long-term environmental weathering of PLA MPs, focusing on the pervasive process of oxidation [15]. For each aging time point (0, 60, 120, 180, 240 h), independent batches of PLA MPs (5 g) were treated with 80 mL of 100 mmol/L K₂S₂O₈ solution (pH 7.0) at 70 °C under continuous magnetic stirring at 200 rpm. The oxidant solution was replenished every 12 h, and MPs were rinsed with deionized water every 24 h. Residual oxidant was thoroughly removed via five repeated washes, confirmed by a negative KI-starch test. The aged MPs were then neutralized by deionized water until the pH was 6.8 ± 0.3, dried at 60 °C, and stored in a desiccator. Mass recovery exceeded 98% for all batches.

Surface morphology was examined using scanning electron microscopy (SEM, Thermo Fisher AxiaChemiSEM, Waltham, MA, USA, 3.0 kV) after sputter-coating with gold [16]. Images were acquired at various magnifications (500~5000×) across multiple fields (>20) to ensure representative sampling. Chemical surface modifications were analyzed by attenuated total reflectance Fourier-transform infrared spectroscopy (FT-IR, Thermo Fisher Nicolet iS50, Waltham, MA, USA). Spectra (400~4000 cm⁻¹) were baseline-corrected and normalized to the C–H stretching band (~2950 cm⁻¹) for comparative analysis [17]. Surface charge (zeta potential) was measured with a Litesizer 500 (Anton Paar, Graz, Austria) on fine particulate fractions suspended in 1 mM KCl [18]. Surface hydrophilicity was assessed via static water contact angle measurements (OCA15, DataPhysics, Stuttgart, Germany) on solvent-cast PLA films using the sessile drop method [19].

2.3. TC Adsorption Experiments

Briefly, 20 mg of MPs were incubated with 20 mL of TC solution (20 mg/L for kinetics; 5~25 mg/L for isotherms) in foil-wrapped centrifuge tubes at 25 °C and 150 rpm. Control tubes containing TC solution without MPs were incubated in parallel. HPLC analysis confirmed no significant degradation over 96 h. Samples were collected at predetermined intervals (0~96 h), immediately filtered (0.22 μm), and the TC concentration was quantified. While initial screening used UV-Vis spectrophotometry at 357 nm (BioTek Synergy H1, Winooski, VT, USA), all final adsorption isotherm data and key kinetic points were confirmed by HPLC-UV (see Section 2.4.2 for method). The effects of pH (1~11, adjusted with HCl/NaOH), ionic strength (0~30‰ NaCl), and humic acid (0~20 mg/L, Aladdin standard) were evaluated. For experiments involving humic acid, matrix-matched calibration curves were used. Kinetic data were fitted to pseudo-first-order (PFO) and pseudo-second-order (PSO) models. Isotherm data were fitted to Langmuir and Freundlich models.

2.4. Simulating Human Digestive Transit

2.4.1. Simulated Gastrointestinal Fluids

SGF contained 3.2 g/L pepsin in a 100 mM NaCl matrix, acidified to pH 2.0 with HCl. Simulated SIF contained 5.0 g/L BSA and 6.0 g/L NaT in 100 mM NaCl and 50 mM phosphate buffer, pH 7.0. Fluids were prepared fresh and used immediately. While simplified, this model targets key digestive drivers: acidic proteolysis (SGF) and solubilization by bile salts and proteins (SIF) [20].

2.4.2. Desorption Kinetics and Solubilization Assays

For desorption kinetics, 20 mg of TC-laden MPs were incubated in 20 mL of pre-warmed SGF or SIF at 37 °C with orbital shaking at 150 rpm to simulate basic gastrointestinal peristalsis. Aliquots were taken over 96 h. For SIF samples, proteins were precipitated with acetonitrile (1:1 v/v) prior to centrifugation (10,000 rpm, 10 min) and filtration (0.22 μm). TC concentration was quantified by HPLC-UV using an Agilent 1260 system equipped with a C18 column (Eclipse Plus, 4.6 × 150 mm, 5 μm, Santa Clara, CA, USA). The isocratic mobile phase was 0.1% formic acid/acetonitrile (75:25, v/v) at 1.0 mL/min, with detection at 357 nm. The individual solubilizing effects of pepsin (0.8~4.0 g/L), BSA (1~5 g/L), and NaT (1~6 g/L) on TC were assessed in separate dose–response experiments.

2.4.3. Two-Stage Bioaccessibility Assessment

A sequential two-phase digestion model was employed. TC-laden MPs were first incubated in SGF for 2 h (gastric phase), followed by transfer (via centrifugation and supernatant removal) into fresh SIF for a further 4 h (intestinal phase). All incubations were at 37 °C with 150 rpm shaking. TC released in each phase was quantified by HPLC. Bioaccessibility was calculated as the cumulative percentage of the initially adsorbed TC mass released into the combined digestive fluids.

2.5. Kinetic and Isothermal Modeling for Adsorption Mechanism Deciphering

The adsorption capacity of TC onto MPs at time t (Q_t, mg/g) was calculated using the mass-balance equation (Equation (1)):

$$Q_t={\displaystyle \frac{V}{m}}(C_0-C_t)$$

(1)

where m (g) is the mass of MPs, V (L) is the solution volume, and C₀ and C_t are the initial and time-dependent TC concentrations (mg/L), respectively.

To analyze the adsorption kinetics, the temporal data were fitted by the PFO (Equation (2)) and PSO (Equation (3)) models [21]:

$$Q_t=Q_e(1-e^{-k_{1}t})$$

(2)

$$Q_t={\displaystyle \frac{Q_e^2{k}_{2}t}{1+Qe{k}_{2}t}}$$

(3)

where k₁ (h⁻¹) and k₂ (g/(mg h)) represent the rate constants of the two models; t (h) is the adsorption time, and Q_e (mg/g) is the equilibrium adsorption capacity. The PFO model typically describes diffusion-limited processes, while a superior fit to the PSO model is often associated with adsorption processes where the rate may be influenced by adsorbent–adsorbate affinity and site availability.

Furthermore, adsorption isotherms were modeled using both the Langmuir model (Equation (4)) and Freundlich model (Equation (5)) to identify binding affinity and surface heterogeneity [22]:

$$Q_e={\displaystyle \frac{Q_m{K}_L{C}_e}{1+K_L{C}_e}}$$

(4)

$$Q_e=K_F{C}_e^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.}}$$

(5)

where Q_m (mg/g) signifies the maximum monolayer adsorption capacity; C_e (mg/L) is the equilibrium concentration of TC; K_L (L/mg) and K_F (L/mg) are the Langmuir and Freundlich affinity constants, respectively; n is the heterogeneity factor indicative of adsorption intensity and surface diversity. The Langmuir model assumes homogeneous monolayer adsorption, whereas the Freundlich model is empirical and often applicable to multilayer adsorption on heterogeneous surfaces.

Nonlinear regression was used for all fittings. The goodness of fit was evaluated primarily using the coefficient of determination (R²) and the consistency between fitted and experimental Q_e values. The standard error for each fitted parameter was also calculated and is reported alongside the parameter estimates in the results tables. This integrated modeling approach provides quantitative insights into the adsorption dynamics and equilibrium, aiding in the assessment of the vector potential of antibiotic-carrying MPs.

2.6. Statistical Analysis

All experimental data were derived from a minimum of three independent replicates and are presented as mean ± standard deviation. The effects of environmental parameters (e.g., pH) on adsorption were assessed via one-way analysis of variance (ANOVA) with Tukey’s post hoc test, where a p-value of <0.05 was considered statistically significant. All statistical computations and graphical analyses were performed using OriginPro (Version 2021, OriginLab Corporation, Northampton, MA, USA).

3. Results and Discussion

3.1. Development and Characterization of an Artificially Aged Microplastic Model

Surface morphological analysis via SEM revealed substantial alterations in PLA MPs following K₂S₂O₈-induced aging (Figure 1). Pristine MPs displayed relatively smooth surfaces, whereas aged specimens exhibited progressively roughened morphologies with pronounced porosity and microcrack formation. These topographical changes are attributed to radical-induced polymer degradation, wherein sulfate radicals (·SO₄⁻) and secondary oxygen radicals (OH/O₂⁻) attack the aliphatic ester linkages (-COO-), leading to chain scission and surface embrittlement [23]. For food-contact plastics, such weathering could occur during storage, processing, or environmental exposure prior to ingestion. The resulting high-surface-area defects are critical as they can enhance the adherence and accumulation of foodborne contaminants [24].

Figure 1. SEM images of original PLA (a), aging 60 h PLA (b), 120 h PLA (c), 180 h PLA (d) and 240 h PLA (e).

FTIR spectroscopy further elucidated chemical modifications during aging (Figure 2). ATR-FTIR spectra, normalized to the C–H stretching band (~2950 cm⁻¹), showed consistent trends across triplicate measurements. The relative intensity of peaks associated with C–H bending (1371, 1449 cm⁻¹) increased, suggesting chain rearrangement. Enhanced absorption at 1091 cm⁻¹ and 1201 cm⁻¹ indicated ester bond breakdown and C–O formation. Notably, the elevated relative intensity of the carbonyl band (C=O, 1761 cm⁻¹) confirmed surface oxidation and the introduction of oxygen-containing groups, consistent with polymer oxidative degradation [25]. This chemical transformation is pivotal as it can directly alter the interaction potential of MPs with polar food contaminants like antibiotics.

Figure 2. FT-IR of original and different aging time PLAs.

Water contact angle measurements on solvent-cast PLA films demonstrated a significant shift from hydrophobic (101.2°) to hydrophilic (59.6°) wetting behavior with aging (mean ± SD, n = 5 measurements per sample) (Figure 3). This transition, driven by oxidative functionalization [26], can affect the aggregation state and bioavailability of MPs in aqueous food matrices and the gastrointestinal environment [27].

Figure 3. Contact angle of original PLA (a), aging 60 h PLA (b), 120 h PLA (c), 180 h PLA (d) and 240 h PLA (e).

Surface charge analysis via zeta potential (Figure 4) supported the chemical changes, showing increased negative surface potential with aging. Measurements were performed on particle suspensions in 1 mM KCl (pH adjusted with HCl/NaOH, 2 min equilibration) at a concentration of 0.1 mg/mL, following established protocols for colloidal systems [18].

Figure 4. Zeta potential of original and aged microplastics.

Collectively, the persulfate-accelerated aging protocol induced key surface transformations—oxidation, increased roughness, and hydrophilication—that are hallmarks of environmentally weathered MPs, including those derived from food packaging [28]. While this controlled oxidative model effectively simulates a dominant pathway of long-term aging, we acknowledge it represents one specific mechanism under accelerated conditions. Nevertheless, it provides a reproducible and powerful tool for probing the consequences of surface weathering on microplastic-contaminant interactions in contexts relevant to food safety [29].

3.2. Interaction Mechanisms and Environmental Behavior of TC with Aged PLA MPs

3.2.1. Adsorption Kinetics and Proposed Mechanisms

The adsorption kinetics of TC onto both pristine and aged PLA MPs reached equilibrium within ~48 h (Figure 5a). Aged MPs exhibited a substantially enhanced equilibrium adsorption capacity, increasing by 43.8% to 83.7% compared to pristine samples. This augmentation is attributed to aging-induced increases in surface roughness, porosity, and abundance of oxygen-containing functional groups, which collectively create a more fragmented polymeric matrix with diffusion pathways and adsorption site availability for TC molecules [8,30].

Figure 5. Adsorption behavior of PLA on TC: (a) fitting curves of adsorption kinetics models for different aging times; (b) fitting curve of Weber–Morris model; and (c) fitting graphs of isotherm models.

As detailed in Table 1, the adsorption kinetics were best described by the PSO model (R² = 0.987~0.991), which is often indicative of adsorption processes where the rate may be influenced by adsorbent–adsorbate interactions and site availability. The equilibrium adsorption capacity (Q_e) increased markedly from 0.515 ± 0.014 mg/g for pristine PLA to 0.931 ± 0.031 mg/g for the sample aged 240 h [24,31]. Interestingly, the PSO rate constant (k₂) decreased slightly with extended aging, which may reflect changes in the overall adsorption energy landscape or site accessibility as the surface becomes more complex (Figure 5a).

Table 1. Parameters of the kinetics and model for sorption experimental data of TC on PLA MPs.

Kinetics	Parameter	Samples
		PLA-Ori	aPLA-60 h	aPLA-120 h	aPLA-180 h	aPLA-240 h
Pseudo-first order	k1 (h−1)	0.086 ± 0.008	0.135 ± 0.010	0.116 ± 0.015	0.116 ± 0.015	0.122 ± 0.014
	qe (mg/g)	0.515 ± 0.014	0.721 ± 0.017	0.755 ± 0.026	0.843 ± 0.035	0.931 ± 0.031
	R2	0.976	0.988	0.975	0.963	0.967
Pseudo-second order	k2 (g(mg·h)−1)	0.178 ± 0.021	0.177 ± 0.008	0.174 ± 0.016	0.160 ± 0.020	0.156 ± 0.015
	qe (mg/g)	0.586 ± 0.016	0.815 ± 0.009	0.859 ± 0.016	0.944 ± 0.027	1.019 ± 0.021
	R2	0.987	0.998	0.995	0.988	0.991

The Weber–Morris intraparticle diffusion model (Figure 5b, Table 2) revealed a multi-stage process [32]. The multi-linear profiles, characterized by initial boundary layer diffusion, followed by intra-particle diffusion and finally equilibrium, indicate that intraparticle diffusion is a contributing but not the sole rate-limiting step. The intra-particle diffusion rate constants increased significantly with aging, attributable to the formation of porous architectures and cracks (Figure 2), which facilitate deeper penetration of TC molecules.

Table 2. Parameters of the Weber–Morris model for sorption experimental data of TC on PLA MPs.

Model	Parameter	Samples
		PLA-Ori	aPLA-60 h	aPLA-120 h	aPLA-180 h	aPLA-240 h
One step	ks1 (mg/g h−1/2)	0.078 ± 0.013	0.171 ± 0.018	0.222 ± 0.008	0.242 ± 0.008	0.260 ± 0.007
	C	0.015 ± 0.027	−0.058 ± 0.028	−0.101 ± 0.011	−0.087 ± 0.010	−0.098 ± 0.011
	R2	0.913	0.966	0.998	0.997	0.998
Two step	ks2 (mg/g h−1/2)	0.058 ± 0.009	0.111 ± 0.019	0.116 ± 0.011	0.128 ± 0.013	0.139 ± 0.013
	C	0.0105 ± 0.048	0.096 ± 0.077	0.118 ± 0.004	0.124 ± 0.039	0.144 ± 0.038
	R2	0.928	0.946	0.981	0.969	0.975
Third step	ks3 (mg/g h−1/2)	0.007 ± 0.005	0.020 ± 0.009	0.026 ± 0.004	0.027 ± 0.006	0.025 ± 0.008
	C	0.455 ± 0.045	0.0573 ± 0.068	0.560 ± 0.028	0.629 ± 0.047	0.722 ± 0.071
	R2	0.646	0.712	0.931	0.914	0.817

3.2.2. Adsorption Isotherms and Transition in Uptake Mechanisms

The adsorption isotherms (Figure 5c, Table 3) quantitatively confirm the enhanced affinity of aged PLA MPs for TC. The experimental data were fitted to Langmuir, Freundlich, and linear models to elucidate the governing equilibrium mechanisms. Objective model selection, based on the Akaike Information Criterion (AIC) alongside R² values, revealed a distinct trend: adsorption onto pristine and slightly aged (60 h) MPs was best described by the Langmuir model, suggesting a tendency toward monolayer coverage on relatively uniform sites [33]. In contrast, for MPs aged 120 h and beyond, the Freundlich model provided a superior fit, indicating the development of heterogeneous surface sites with a range of binding energies. This transition correlates directly with the observed physical and chemical surface alterations—increased roughness, porosity, and oxygen functionality—which create a more diverse array of interaction sites [24].

Table 3. Parameters of the isotherm models for sorption experimental data of TC on PLA.

Isotherm	Parameter	Samples
		PLA-Ori	aPLA-60 h	aPLA-120 h	aPLA-180 h	aPLA-240 h
Freundlich	Kf ((mg/g)/(L/mg)n)	0.079 ± 0.011	0.096 ± 0.008	0.117 ± 0.008	0.127 ± 0.012	0.171 ± 0.011
	1/n	0.660 ± 0.046	0.645 ± 0.031	0.614 ± 0.025	0.610 ± 0.034	0.536 ± 0.021
	R2	0.987	0.993	0.995	0.990	0.995
Langmuir	Kl (L/mg)	0.035 ± 0.006	0.040 ± 0.002	0.040 ± 0.007	0.050 ± 0.012	0.051 ± 0.006
	Qm (mg/g)	1.386 ± 0.155	1.486 ± 0.049	1.671 ± 0.180	1.586 ± 0.220	1.712 ± 0.096
	R2	0.990	0.999	0.985	0.975	0.991
Linear	Kd (L/g)	0.021 ± 0.002	0.025 ± 0.002	0.025 ± 0.002	0.029 ± 0.001	0.025 ± 0.001
	R2	0.968	0.963	0.977	0.985	0.980

The Freundlich parameters provide further mechanistic insight. The exponent 1/n remained below unity (0.536–0.660) across all samples, confirming favorable and non-linear adsorption. More significantly, the Freundlich constant K_F, representing adsorption capacity, increased steadily from 0.079 ± 0.011 (pristine) to 0.171 ± 0.011 mg/g·(L/mg)ⁿ (240 h-aged). This progression underscores a substantial augmentation in overall adsorption strength with aging, driven by the synergistic effects of newly introduced polar groups (enhancing hydrogen bonding and electrostatic interactions) and increased surface accessibility [34].

3.2.3. Environmental Modulation of Adsorption: Implications for Contaminant Mobility

The interaction between TC and aged PLA MPs is critically modulated by key aqueous environmental parameters, which inform their potential behavior in natural and food-relevant systems (Figure 6). Control experiments verified TC stability across the entire tested pH range over the duration of the adsorption experiments. Adsorption exhibited a pronounced bell-shaped dependence on pH, with maximum uptake observed at neutral pH (~7.0). This profile arises from the interplay between the pH-dependent speciation of TC—a multiprotic molecule—and the surface charge of the MPs [35]. At neutral pH, favorable interactions such as hydrogen bonding and hydrophobic effects are optimized. Under highly acidic or alkaline conditions, electrostatic repulsion between similarly charged species reduces adsorption efficiency [36]. In contrast, variations in ionic strength (0~30‰ NaCl) induced negligible changes in TC adsorption across all aging stages, indicating that the primary adsorption mechanisms are not governed by simple electrostatic screening from monovalent ions [37]. Conversely, the presence of humic acid significantly suppressed TC adsorption in a concentration-dependent manner, primarily through competitive sorption for surface sites.

Figure 6. The effect of factors on the adsorption of TC by PLA and aged PLA: (a) pH; (b) salinity; (c) HA. Data are mean ± standard deviation (SD), n = 3. Different letters show statistically significant differences between different groups (p < 0.05).

Collectively, these findings demonstrate that the adsorption of antibiotics onto aged biodegradable MPs is a process governed predominantly by specific surface chemistry and solution pH, rather than by ionic strength. This understanding is crucial for predicting the fate and vector potential of MP-bound contaminants as they transition from environmental matrices into the pH- and surfactant-rich milieu of the gastrointestinal tract [38].

3.3. Intestinal-Driven Desorption and Bioaccessibility of TC from Aged MPs

3.3.1. Desorption Kinetics in Simulated Digestive Fluid

The desorption kinetics of TC from PLA MPs were investigated in SIF, SGF, and ultrapure water (Figure 7). A key finding was the superior desorption efficiency of SIF across all samples. Pristine PLA MPs released up to 62.7% of their adsorbed TC load in SIF, significantly higher than in SGF (19.4%) or water (14.2%). This trend persisted for aged MPs, although the fractional desorption in SIF exhibited a gradual decline with increasing aging duration (e.g., 54.4% for aPLA-60 h to 44.6% for aPLA-240 h).

Figure 7. (a) Desorption kinetics of original and aged PLA in simulated gastric fluid, simulated intestinal fluid and Milli-Q water, (b) desorption rate of TC in simulated gastric fluid, simulated intestinal fluid and Milli-Q water.

This pronounced desorption in SIF is attributed to its chemical composition (Figure 8). The presence of solubilizing agents—specifically BSA and the bile salt NaT—enhances the apparent solubility of TC in the aqueous phase, thereby shifting the adsorption–desorption equilibrium toward release. In contrast, the lower desorption in water reflects the absence of these solubilizing drivers, while the moderately higher desorption in the acidic SGF (pH 2.0) may be influenced by the pH-dependent speciation of both TC and the MP surface charge, which can alter interfacial interactions.

Figure 8. Bioaccessibility of TC adsorbed on original and aged PLA by using in vitro gastrointestinal model. Data are mean ± standard deviation (SD), n = 3. Different letters show statistically significant differences between different groups (p < 0.05).

3.3.2. Solubilization of TC by Gastrointestinal Components

To mechanistically understand the desorption results, the individual solubilizing capacity of key digestive components was quantified (Figure 9). Experiments were conducted at controlled, physiologically relevant pH (SGF condition: pH 2.0; SIF condition: pH 7.0). Pepsin demonstrated a modest solubilizing effect on TC compared to water at neutral pH. In contrast, both BSA and NaT significantly increased TC solubility in a concentration-dependent manner. The enhancement by BSA is ascribed to ligand-specific binding interactions that increase TC’s apparent solubility in the aqueous phase. The effect of NaT is consistent with micellar solubilization, wherein hydrophobic molecules like TC are incorporated into the hydrophobic core of bile salt micelles. The combined action of these components in SIF creates a solubilizing milieu that effectively competes with the MP surface for TC, facilitating its desorption [39,40].

Figure 9. Solubility of TC in various concentrations of (a) Pepsin; (b) BSA; (c) NaT solutions and (d) solubility of TC in simulated gastric fluid, simulated intestinal fluid and ultra-pure water. Different lowercase letters in the figure indicate the significant differences.

3.3.3. Adsorption Isotherms in Simulated Digestive Fluids

The adsorption behavior of TC onto PLA MPs was also directly assessed in SGF and SIF to understand the equilibrium under digestive conditions (Figure 10, Table 4). The Freundlich model provided the best fit for the isotherm data in both fluids. A critical observation was the notable reduction in the Freundlich adsorption capacity parameter (K_F) in SIF compared to its value in either pure water or SGF. For instance, for 240 h-aged MPs, K_F decreased from ~0.170 (mg/g) (L/mg)ⁿ in water to ~0.044 (mg/g) (L/mg)ⁿ in SIF. This suppression of adsorption affinity in SIF provides direct evidence for the competitive effect of intestinal fluid components. BSA and NaT likely reduce TC uptake by either occupying adsorption sites on the MP surface (competitive adsorption) or by sequestering TC in the solution phase (solubilization), thereby decreasing its effective activity and partitioning onto the solid phase [41].

Figure 10. Isotherm model fitting of TC adsorption by original and aged PLA in simulated gastric fluid (a) and simulated intestinal fluid (b).

Table 4. Adsorption isotherm parameter of TC on PLA MPs in gastrointestinal fluids.

Isotherm		Parameter	Samples
			PLA-Ori	aPLA-60 h	aPLA-120 h	aPLA-180 h	aPLA-240 h
Gastric fluid	Freundlich	Kf ((mg/g)/(L/mg)n)	0.090 ± 0.005	0.116 ± 0.004	0.132 ± 0.006	0.182 ± 0.010	0.205 ± 0.007
		1/n	0.648 ± 0.035	0.606 ± 0.019	0.597 ± 0.028	0.501 ± 0.036	0.491 ± 0.025
		R2	0.998	0.998	0.998	0.995	0.997
	Langmuir	Kl (mg/L)	0.047 ± 0.006	0.050 ± 0.005	0.047 ± 0.004	0.059 ± 0.007	0.097 ± 0.010
		Qm (mg/g)	1.301 ± 0.087	1.447 ± 0.087	1.643 ± 0.070	1.530 ± 0.078	1.359 ± 0.063
		R2	0.995	0.994	0.997	0.991	0.991
	Linear	Kd (L/g)	0.024 ± 0.001	0.025 ± 0.002	0.026 ± 0.001	0.022 ± 0.002	0.029 ± 0.002
		R2	0.987	0.982	0.993	0.972	0.978
Intestinal fluid	Freundlich	Kf ((mg/g)/(L/mg)n)	0.023 ± 0.002	0.035 ± 0.003	0.036 ± 0.002	0.039 ± 0.000	0.044 ± 0.002
		1/n	0.412 ± 0.085	0.294 ± 0.011	0.310 ± 0.057	0.312 ± 0.008	0.294 ± 0.059
		R2	0.981	0.955	0.990	0.999	0.988
	Langmuir	Kl (mg/L)	0.096 ± 0.017	0.178 ± 0.033	0.159 ± 0.031	0.226 ± 0.010	0.240 ± 0.058
		Qm (mg/g)	0.120 ± 0.007	0.109 ± 0.006	0.120 ± 0.005	0.122 ± 0.002	0.127 ± 0.007
		R2	0.973	0.945	0.952	0.996	0.917
	Linear	Kd (L/g)	0.002 ± 0.000	0.001 ± 0.000	0.001 ± 0.000	0.002 ± 0.000	0.002 ± 0.000
		R2	0.969	0.921	0.975	0.997	0.965

3.3.4. Bioaccessibility in a Sequential Gastrointestinal Simulation

The integrated bioaccessibility of TC—defined as the percentage of initially adsorbed TC released during sequential SGF (2 h) and SIF (4 h) incubation—was quantified (Figure 8; data are mean ± SD, n = 4). Bioaccessibility decreased with aging, from 32.6% (pristine) to 18.7% (240 h-aged) in the intestinal phase. This trend aligns with the stronger adsorption interactions (e.g., hydrogen bonding) on the oxidized surfaces of aged MPs. However, when combined with the elevated adsorption capacity of aged MPs, the net mass of TC mobilized remains considerable. These results highlight that digestive conditions, particularly in the intestine, are a major driver for the release of MP-bound antibiotics, defining their bioaccessible pool for potential intestinal uptake [42,43].

3.4. Limitations and Future Perspectives

While providing a mechanistically reproducible assessment of contaminant bioaccessibility from aged food-contact microplastics, this study is necessarily defined by specific methodological boundaries. Acknowledging these limitations is essential for contextualizing the findings and charting a precise course for more physiologically and environmentally relevant food safety research.

Microplastic Model. The experimental framework relied on monodisperse, commercially sourced PLA particles to ensure uniformity and isolate the variable of controlled aging. This choice, however, limits direct extrapolation to the heterogeneous mixtures of particles—varying in size, shape, and polymer crystallinity—that result from the environmental degradation of diverse food packaging materials. Future studies must incorporate MPs sourced directly from consumer products and environmental matrices to better represent the actual physical and chemical diversity ingested via the diet.

Aging Simulation. The accelerated oxidative aging protocol using K₂S₂O₈ effectively generated surface properties characteristic of long-term weathering, such as increased oxygen functionality and porosity. Nonetheless, this single-stressor approach does not replicate the synergistic effects of combined environmental forces (e.g., UV photoaging with mechanical abrasion and biodegradation) that plastics experience in real-world scenarios. Subsequent research should employ sequential or simultaneous multi-stressor aging regimens to more accurately simulate the complex lifecycle of food-packaging debris.

Contaminant Framework. TC served as a robust model contaminant due to its pH-dependent speciation, enabling a detailed mechanistic study of adsorption–desorption kinetics. A recognized limitation is that its behavior may not be representative of other prevalent foodborne contaminants with differing physicochemical properties. Moreover, the concentrations used, while optimal for model fitting, exceed typical environmental or dietary levels. Future investigations should therefore expand to complex, low-dose mixtures of relevant co-contaminants (e.g., other pharmaceuticals, pesticides) to assess competitive interactions and establish more realistic exposure scenarios.

Digestion Model. The application of a standardized static in vitro gastrointestinal model was pivotal for controlled, time-resolved bioaccessibility measurements. Its primary limitation lies in its simplification of the dynamic, shear-rich, and microbiologically active human gut environment. To enhance physiological relevance, the logical progression is toward dynamic digestion systems that incorporate continuous flow, mucosal interfaces, and, critically, the influence of gut microbiota on contaminant release and potential transformation.

Integrated Risk Assessment. The current study provides essential in vitro bioaccessibility coefficients and mechanistic insights. The critical translational step is to bridge these data with in vivo bioavailability and toxicological endpoints. Concurrently, advancing analytical methodologies for detecting and characterizing aged microplastics and the sorbed contaminant loads within complex food matrices is paramount. This integrated, tiered approach will be fundamental for developing robust, evidence-based risk assessment frameworks and informing the design of next-generation food-contact materials.

4. Conclusions

This study elucidates how environmental aging transforms PLA MPs—a polymer widely used in food packaging—into potential carriers of antibiotics such as TC within the human digestive tract. Accelerated aging enhanced TC adsorption on PLA MPs through increased surface oxidation and porosity, while simulated intestinal fluid was shown to significantly promote TC release compared to gastric or aqueous environments. Although the release percentage decreased with aging, the concomitant rise in adsorption capacity indicates that aged MPs may still deliver substantively bioaccessible TC loads in the gut.

These findings directly inform food safety and food packaging material research, as aged PLA fragments may enter the human gastrointestinal system via contaminated food or degraded packaging. The experimental framework established here—integrating aging simulation, adsorption analysis, and physiologically relevant digestion models—provides a methodological basis for screening the vector potential of food-contact microplastics under digestive conditions.

While this study offers mechanistic insight under controlled settings, further research is warranted to assess real-world exposure scenarios, including the effects of complex food matrices, chronic low-dose intake, and combined contaminants. Future work incorporating aged MPs extracted from food-contact materials, along with in vivo and epidemiological studies, will be essential to translate these findings into actionable food safety guidelines and material design innovations.