Sertoli cell oxidative stress as a convergent hazard endpoint for environmental toxicants: A systematic review, meta-analysis, and AOP-based assessment

Abstract

Environmental chemical exposures are primary drivers of declining male reproductive health. Sertoli cells (SCs), indispensable for spermatogenesis and blood-testis barrier (BTB) integrity, are preferential targets for toxicant-induced oxidative injury across chemically diverse classes. We conducted a systematic review and meta-analysis organised around an a priori Adverse Outcome Pathway (AOP): toxicant interaction with SC mitochondria or NADPH oxidases (Molecular Initiating Event) → ROS elevation (Key Event 1, KE1) → SC cellular dysfunction including apoptosis, ferroptosis, and BTB disruption (Key Event 2, KE2) → impaired spermatogenesis (Adverse Outcome). Searching PubMed, Embase, Web of Science, and Scopus from inception to January 2026, we identified 1458 records, of which 96 met inclusion criteria for qualitative synthesis and 29 provided extractable data for meta-analysis. All six chemical classes examined — air pollutants (PM₂.₅), endocrine-disrupting chemicals, pesticides, heavy metals, mycotoxins, and emerging contaminants — significantly elevated SC oxidative stress relative to controls. Metals produced the largest pooled standardised mean difference (SMD > 2.0) and air pollutants the second largest (SMD ∼1.8). The direction of effect was uniformly positive across all 29 studies. Cross-study ecological regression supported the AOP linkage between ROS elevation (KE1) and SC apoptosis (KE2; R² = 0.77, p < 0.001). One Health subgroup analysis showed directionally consistent effect sizes across rodent, livestock, and in vitro evidence streams. These findings position SC oxidative stress as a convergent, chemically non-specific, and phylogenetically conserved hazard endpoint warranting formalisation within tiered reproductive toxicity test strategies.

Graphical abstract

Diverse environmental toxicants air pollutants, EDCs, pesticides, metals, mycotoxins, and emerging contaminants reach the testis and accumulate in SCs at the BTB. These agents stimulate mitochondrial and NOX-derived ROS, overwhelming local antioxidant defences. Excess ROS disrupt tight junctions, reprogram SC glycolysis and lipid metabolism, and activate apoptosis, defective autophagy, ferroptosis, and pyroptosis. SC loss and dysfunction cause BTB leakage, germ cell DNA damage and apoptosis, and reduced sperm count and quality, resulting in male subfertility. Antioxidants, MSC-derived exosomes, and photobiomodulation represent emerging SC-targeted mitigation strategies.

Highlights

• •Six chemical classes uniformly elevate Sertoli cell ROS across species.
• •Metals show the largest oxidative effect (SMD > 2.0); air pollutants rank second.
• •Cross-study AOP regression links ROS burden to Sertoli cell apoptosis (R² = 0.77).
• •SC oxidative stress is a conserved One Health reproductive hazard endpoint.
• •AOP framework anchors SC ROS as a mechanistic endpoint for tiered risk assessment.

1. Introduction

The global decline in male reproductive health is a pressing environmental health concern. Sperm concentration, motility, and morphology have measurably fallen over five decades across multiple populations [[1], [2], [3]]. Environmental and occupational chemical exposures contribute substantially to this trend [[4], [5], [6]]. Multi-centre cohorts and meta-analyses link ambient fine particulate matter (PM₂.₅) to reductions in sperm concentration, progressive motility, and normal morphology, and to elevated sperm DNA fragmentation, at exposure levels commonly encountered in urban populations [7]. Systematic reviews of phthalate biomonitoring data indicate that urinary concentrations of key metabolites within the current human exposure range are associated with impaired semen quality and altered reproductive hormones [8]. Similar associations extend to pesticides, metals, and other occupational exposures, particularly among agricultural workers and men in highly polluted environments [9]. These findings establish that structurally diverse environmental chemicals threaten male reproductive function, yet no study has quantitatively synthesised whether they share a common cellular mechanism of action.

Within the testis, Sertoli cells (SCs) occupy a central position at the interface between the systemic circulation, local somatic cells, and developing germ cells. SCs form the blood-testis barrier (BTB), establish distinct basal and adluminal compartments, and provide the physical scaffolding, metabolic substrates, and paracrine cues that guide germ cells through mitosis, meiosis, and spermiogenesis [[10], [11], [12]]. Because virtually every developing germ cell is anchored to or enveloped by SC processes at some point in its life cycle, toxicant-induced perturbations of SC function propagate directly to spermatogenesis. Lipophilic and transport-mediated toxicants accumulate preferentially in SCs at the BTB interface, making SCs the primary intratesticular target for a broad spectrum of environmental chemicals.

Reactive oxygen species (ROS) are unavoidable by-products of cellular metabolism. At physiological concentrations, ROS serve as second messengers regulating sperm capacitation, the acrosome reaction, and sperm-oocyte fusion [[13], [14], [15]]. However, the testicular microenvironment is highly susceptible to oxidative imbalance: intense mitochondrial activity, rapid cell division, and enrichment of cell membranes in polyunsaturated fatty acids (PUFAs) create a constitutively pro-oxidant baseline. When ROS generation outstrips antioxidant capacity, oxidative stress causes lipid peroxidation, protein carbonylation, and nucleic acid damage in both SCs and germ cells. For many toxicants, ROS generation in SCs is not merely a downstream correlate but a causal mechanism: agents differing widely in chemical structure and primary molecular targets all converge on SC oxidative stress, suggesting a chemically non-specific ‘final common pathway’ from environmental exposure to impaired spermatogenesis [16]. Before triggering terminal events like apoptosis, excess ROS in SCs first cripples their primary nursing function by severely altering cellular metabolism. Specifically, oxidative stress reprograms glucose and lipid metabolism in SCs, disrupting glycolysis and halting the production of lactate—the indispensable energy substrate required for the survival and maturation of developing germ cells. This metabolic shift leads to immediate germ cell starvation. Concurrently, excess ROS disrupt BTB junctional complexes and activate multiple programmed cell death (PCD) modes, ultimately compromising the entire SC niche that supports germ cell development [17,18].

This mechanistic convergence carries direct implications for reproductive hazard assessment. Current OECD reproductive toxicity test guidelines (TG 416, TG 443) rely primarily on apical endpoints — organ weights, sperm parameters, fertility rates — that register SC injury only after substantial somatic cell loss and provide no mechanistic comparability across chemical classes. Formalising SC oxidative stress as a Key Event (KE) within an Adverse Outcome Pathway (AOP) would enable its use as an early, mechanistically grounded biomarker in tiered hazard characterisation. The One Health framework complements this by evaluating whether experimental evidence from animal and in vitro models translates across species to inform human risk assessment.

We conducted a systematic review and meta-analysis to test whether SC oxidative stress constitutes a convergent, cross-chemical hazard endpoint, organised around an a priori AOP: toxicant interaction with SC mitochondria or NADPH oxidases (NOX; MIE), ROS elevation (KE1), SC cellular dysfunction—initially manifesting as the failure to provide metabolic and nutritional sustenance to germ cells, followed by BTB disruption and programmed cell death (KE2), impaired spermatogenesis (AO). A One Health subgroup analysis compared effect magnitudes across human, rodent, livestock, and in vitro evidence streams. To our knowledge, this is the first quantitative synthesis applying both AOP architecture and a One Health framework to SC oxidative stress across six chemical classes, providing a mechanistic foundation for integrating this endpoint into tiered reproductive hazard assessment.

2. Materials and methods

2.1. Study design and conceptual framing

This study is a systematic review and meta-analysis designed a priori within a One Health perspective and an AOP logic model. The primary objective was to quantify the association between environmental or occupational chemical exposures and SC oxidative stress, measured as ROS, MDA/TBARS, 4-HNE, F2-isoprostanes, 8-OHdG, protein carbonyls, oxidation-reduction potential, and antioxidant defences (TAC, SOD, CAT, GPx, GSH). Secondary objectives were to synthesise mechanistic evidence linking exposures to SC oxidative stress and downstream AOP key events (BTB disruption, metabolic dysfunction, impaired autophagic flux, apoptosis, ferroptosis, and pyroptosis), and to map this evidence onto AOP elements.

2.2. Protocol and reporting

The pre-specified protocol is provided in the Supplementary Materials, including eligibility criteria, the data extraction codebook, and the planned synthesis strategy. Reporting follows PRISMA 2020 for systematic reviews and PRISMA-S for literature searches.

The scope and eligibility criteria were structured using the Population-Exposure-Comparator-Outcome (PECO) framework (Table S1; operational decision rules in Table S2). Eligible studies included: (i) human epidemiological studies of quantified environmental or occupational chemical exposures with SC-relevant oxidative stress biomarkers; (ii) controlled in vivo animal studies reporting SC oxidative stress following defined chemical exposures, with a vehicle or sham control; and (iii) in vitro or ex vivo mechanistic studies in SC or testicular tubule models reporting oxidative stress endpoints or downstream AOP-relevant key events. Reviews, editorials, and primary studies lacking extractable quantitative data were excluded. Antioxidant markers (TAC, SOD, GPx, GSH) were sign-inverted so that positive SMDs consistently indicate higher oxidative stress burden.

2.3. Search strategy development, peer review, and validation

MEDLINE, Embase, Web of Science Core Collection, and Scopus were searched from inception to 11 January 2026. Two complementary search blocks combined Sertoli-centric outcome terms with environmental or occupational chemical exposure terms, and SC oxidative stress terms with AOP-relevant mechanistic endpoints, using MeSH controlled vocabulary and Boolean operators (Table S3). Backward and forward citation chasing was performed for all included studies and a pre-specified benchmark set. Grey literature was searched in targeted environmental health repositories. The search strategy was peer-reviewed against the PRESS 2015 checklist by an information specialist, revised accordingly, and validated against the benchmark set before the final run.

Records were imported into EndNote for deduplication, then into Covidence for two-stage screening. Two reviewers independently screened titles and abstracts, then full texts, against the prespecified eligibility criteria. A calibration exercise on 29 records preceded formal screening. Disagreements were resolved by consensus with third-reviewer arbitration. Reasons for full-text exclusion are reported in the PRISMA flow diagram (Fig. 1).

Fig. 1.

PRISMA 2020 flow diagram of the study selection process. The systematic search identified 1458 records from PubMed, Embase, Web of Science, and Scopus. After deduplication and title-abstract screening, 166 full-text articles were assessed for eligibility. A total of 96 studies met inclusion criteria for qualitative synthesis; of these, 29 provided sufficient extractable quantitative data and were included in the meta-analysis and AOP verification.

2.4. Data extraction and coding strategy

Data were extracted using a piloted codebook aligned with the AOP framework, capturing study design, population or model descriptors, exposure characterisation (chemical class, matrix, route, dose or concentration, duration, and window), outcome measurement details, and extractable quantitative results. Each mechanistic endpoint was coded to the corresponding AOP element. Missing dispersion metrics were derived from standard errors, confidence intervals, or p-values using standard conversions; graphical data were digitised. Authors were contacted for essential missing information.

Risk of bias was assessed independently by two reviewers using design-appropriate tools: ROBINS-E for human observational studies; the SYRCLE Risk of Bias Tool for in vivo animal studies; and the OHAT Risk of Bias Rating Tool for in vitro and ex vivo studies. Disagreements were resolved by consensus with third-reviewer arbitration. Risk-of-bias judgements informed sensitivity analyses and certainty-of-evidence grading.

2.5. Effect size calculation and meta-analysis

For controlled in vitro and in vivo studies with continuous outcomes, Hedges' g standardised mean differences (SMDs) with small-sample correction were calculated between exposed and control groups. For human studies, adjusted effect estimates were prioritised; where pooling was feasible, estimates were harmonised to a common exposure contrast. Random-effects models were used throughout. Heterogeneity was quantified with tau-squared (τ²) and I-squared (I²) and explored through prespecified subgroup analyses by chemical class, evidence stream, species, exposure window, and biomarker type. Sensitivity analyses included exclusion of high-risk-of-bias studies, alternative exposure contrasts, and leave-one-out influence analyses. Egger's regression was applied for chemical classes with ten or more studies; funnel plots were presented descriptively for smaller evidence bodies. Analyses were conducted in R using the ‘meta’ and ‘metafor’ packages.

2.6. AOP- and One Health-informed evidence integration

Mechanistic findings were mapped onto the prespecified AOP structure. Coherence across evidence streams was evaluated by concordance in directionality, biological plausibility, and dose-response trends. Where sufficient studies reported paired mechanistic endpoints in comparable experimental contexts, a cross-study ecological regression related KE1 magnitude to KE2 response. We explicitly caution that this regression remains an inter-study associative analysis rather than a demonstration of within-cell mechanistic proof. Its interpretation is inherently limited by potential ecological confounding, as unmeasured study-level covariates—such as species-specific toxicodynamic responses, varying exposure windows, and divergent dose metrics—may independently influence both KE1 and KE2 trajectories.

3. Sertoli cell biology and oxidative vulnerability

3.1. Morphology, cytoskeleton, and the BTB as a toxicant-accumulation interface

SCs are tall columnar cells extending from the seminiferous tubule basement membrane to the lumen, where their lateral and apical cytoplasmic processes envelop developing germ cells at successive differentiation stages (Fig. 3A) [[19], [20], [21]]. A tightly organised cytoskeleton of microtubules, intermediate filaments, and actin provides the scaffold for germ cell migration and the mechanical anchoring for the BTB junctional complex. Adjacent SCs form the BTB through tight junctions (TJs; claudin-11, occludin, junctional adhesion molecule A, ZO-1), adherens junctions (N-cadherin, β-catenin), and gap junctions (connexin 43), all anchored to actin bundles and apical ectoplasmic specialisations (Fig. 3B) [22,23]. Dynamic junctional remodelling accommodates the transit of preleptotene spermatocytes from the basal to the adluminal compartment without compromising barrier integrity. The BTB limits passive entry of hydrophilic xenobiotics; however, lipophilic compounds and substrates of active transporters accumulate preferentially in the SC cytoplasm at the barrier interface, making SCs the primary intratesticular target for a broad spectrum of environmental toxicants [24] (see Fig. 4).

Fig. 3.

Sertoli cells at the centre of spermatogenesis and reproductive toxicology. (A) Schematic cross-section of a seminiferous tubule showing SCs extending from the basement membrane to the lumen and enveloping developing germ cells. (B) Higher magnification of adjacent SCs forming the BTB through tight junctions (claudin-11, occludin, ZO-1), adherens junctions (N-cadherin/β-catenin), and gap junctions (connexin 43), anchored to actin bundles and apical ectoplasmic specialisations. (C) Metabolic and paracrine functions of SCs, including high-rate glycolysis and lactate export to germ cells, fatty acid β-oxidation, cholesterol efflux, and secretion of GDNF, SCF, and retinoic acid.

Fig. 4.

ROS generation and antioxidant defences in testis and semen. (A) Dual roles of ROS: low-to-moderate levels mediate physiological sperm capacitation and the acrosome reaction, whereas excessive ROS drive lipid peroxidation, protein oxidation, DNA damage, and mitochondrial dysfunction. (B) Major intracellular ROS sources in SCs, including NADPH oxidases (NOX1/2) and mitochondrial ETC leakage at Complexes I and III. (C) Antioxidant defences comprising enzymatic systems (SOD, CAT, GPx, glutathione reductase) and non-enzymatic scavengers (GSH, vitamins E and C, selenium, zinc); toxicants amplify both ROS sources while depleting antioxidant capacity.

3.2. Metabolic specialisation and germ cell dependency

SCs operate a distinctive metabolic programme shaped by germ cell dependency. High glycolytic flux converts glucose preferentially to lactate, exported via monocarboxylate transporters as the principal energy substrate for differentiating germ cells, which have limited glycolytic capacity of their own [12,25]. Fatty acid β-oxidation supplements SC bioenergetics, and ABC-transporter-mediated cholesterol efflux maintains seminiferous lipid homeostasis [26,27]. The SC paracrine secretome coordinates germ cell development: GDNF maintains spermatogonial stem cell self-renewal, SCF drives spermatogonial differentiation, and retinoic acid initiates meiosis [[28], [29], [30]]. SCs also interact with Leydig cells, peritubular myoid cells, and immune cells through paracrine and juxtacrine signalling [31,32]. Several toxicants, including fumonisin B₁, atrazine, and certain EDCs, impair SC glycolysis and lactate production, coupling metabolic and oxidative injury in a single mechanistic event [33,34]. Specifically, this oxidative injury translates into metabolic failure through the disruption of redox-sensitive regulatory mechanisms. Excessive ROS accumulation interferes with key metabolic signaling cascades, such as the PI3K/AKT and AMPK pathways, and compromises the stability of hypoxia-inducible factor 1-α. This redox-driven transcriptional and translational dysregulation profoundly alters transporter expression, leading to the downregulation and impaired membrane localization of glucose transporters required for glucose uptake, as well as monocarboxylate transporters essential for lactate export [[35], [36], [37]]. By simultaneously suppressing the catalytic machinery and crippling the nutrient transporter network, toxicant-induced oxidative stress fundamentally dismantles the metabolic supply line critical for germ cell survival.

3.3. ROS sources and the antioxidant network

Two intracellular ROS sources set the SC oxidative baseline. Electron leakage from mitochondrial ETC Complexes I and III generates superoxide as an unavoidable by-product of the intense oxidative phosphorylation required to sustain germ cell support [38]. NOX family enzymes — predominantly NOX1 and NOX2 at the SC plasma membrane — provide a second, stimulus-responsive superoxide source activated by pro-inflammatory cytokines and xenobiotic challenge [39]. Cytosolic systems contribute additively under inflammatory or ischaemia-reperfusion conditions [15,40]. Xenobiotics act as force multipliers on this balance: they amplify both ROS sources through direct redox cycling or indirect NOX and ETC activation, while depleting GSH and inhibiting SOD and CAT [41]. This dual-source amplification, recurring across chemical classes, is the mechanistic basis for the directional universality quantified in Section 6.

The antioxidant network neutralising constitutive SC ROS comprises enzymatic components (SOD1/2/3, CAT, GPx isoforms, glutathione reductase) and non-enzymatic scavengers (GSH, vitamins E and C, selenium, zinc) operating in a coordinated cascade [15,42,43]. Repeated or chronic toxicant exposure depletes GSH and inhibits key enzymes, dismantling the SC's capacity to neutralise the amplified oxidative burden [41].

4. Oxidative stress-induced germ cell damage

Although SCs are the primary focus of this review, a mechanistically complete account must include the direct germ cell oxidative damage that co-occurs with and is amplified by SC dysfunction. Sperm plasma membranes are enriched in PUFAs that, while maintaining fluidity, are inherently susceptible to peroxidative attack; toxicant-induced ROS initiate lipid peroxidation, producing reactive aldehydes (MDA, 4-HNE) that crosslink membrane lipids and proteins and impair motility, capacitation, and fertilising capacity [44]. ROS also generate oxidative DNA lesions — 8-OHdG adducts, strand breaks, and chromosomal aberrations — in spermatogonia and spermatocytes, and environmental toxicants increase sperm DNA fragmentation in both experimental animals and exposed human populations [44,45]; because mature spermatozoa have limited repair capacity, such lesions can persist into the zygote and affect offspring health [46]. Finally, many toxicants target sperm midpiece mitochondria, causing mtROS production, membrane depolarisation, and ATP depletion that, combined with SC dysfunction and BTB disruption, drive asthenozoospermia and teratozoospermia [47,48].

5. Systematic review of chemical class-specific sertoli cell oxidative injury

The mechanistic synthesis below draws on the same systematically screened corpus defined in Section 2; studies are organised by chemical class to expose the class-specific upstream mechanisms that feed the shared downstream signature. Crucially, before examining these chemical classes, it is necessary to establish how the specific vulnerability of SCs fundamentally shifts across developmental windows. SC number is strictly determined by a fetal and early postnatal proliferative phase that ends at puberty; thereafter, this fixed, post-mitotic population sets the absolute ceiling for lifetime sperm production. Consequently, undifferentiated and differentiated SCs exhibit fundamentally distinct mechanistic responses to toxicant-induced oxidative stress [49].

In undifferentiated, highly proliferative SCs (typically modeled by in utero/neonatal exposures), oxidative injury primarily disrupts cell cycle progression, upregulates pro-apoptotic cascades, and permanently limits the final SC population. Because the BTB has not yet formed during this window, toxicants easily access the seminiferous cords, making these early-stage cells disproportionately vulnerable to irreversible developmental reprogramming and apoptosis. In contrast, in mature, differentiated SCs (often modeled by TM4 cells or adult in vivo exposures), the mitotic window is closed. Here, the oxidative burden shifts toward uncoupling structural and supportive functions—specifically driving the disassembly of tight junction proteins, crippling glucose and lipid metabolism, and inducing altered autophagic flux. Toxicant-induced impairment during the undifferentiated window carries outsized consequences that manifest as irreversible adult infertility, whereas injury to mature SCs often precipitates transient or progressive spermatogenic decline. Therefore, the generic dual-source ROS amplification described in Section 3.3 must be interpreted through this developmental lens. The following subsections detail how distinct chemical classes exploit these vulnerabilities.

5.1. Fine particulate matter (PM₂.₅) and air pollutants

PM₂.₅ is a complex mixture of transition metals, polycyclic aromatic hydrocarbons (PAHs), and persistent organic radicals. In rodent inhalation models and SC-line exposure studies, PM₂.₅ induces marked ROS generation in testicular tissue and SCs, accompanied by BTB disruption, germ cell apoptosis, and reduced sperm parameters [17,47]. Its PAHs and transition metals simultaneously activate NOX1/2 and impair ETC Complexes I and III [41]. Transcriptomics and metabolomics profiling of SC lines exposed to exhaust-derived PM₂.₅ reveals coordinate dysregulation of oxidative phosphorylation subunits, ferroptosis-related gene networks, and junctional molecules including claudin-11 and ZO-1, consistent with BTB failure [50]. In primary SCs, urban PM₂.₅ triggers DRP1-mediated mitochondrial fission, dissipates the inner mitochondrial membrane potential, disorganises actin bundles, and downregulates TJ proteins, increasing BTB permeability [51].

Epidemiological evidence substantiates these mechanisms at population-relevant doses. Multi-centre cohorts link interquartile range increases in long-term ambient PM₂.₅ (typically 5–15 μg/m³) to clinically significant decrements in sperm concentration, progressive motility, and normal morphology, with elevated DNA fragmentation [7,52,53]. Modeled testicular particle burdens in chronic animal inhalation studies overlap substantially with those expected under real-world urban exposure, supporting translational extrapolation from SC models to population-level risk [53].

5.2. Endocrine-disrupting chemicals: phthalates, alkylphenols, and bisphenols

EDCs exert reproductive toxicity through both hormonal receptor modulation and direct oxidative injury, with a substantial fraction of SC-targeted effects ROS-mediated [[53], [54], [55]]. Di(2-ethylhexyl) phthalate (DEHP) and its metabolite mono(2-ethylhexyl) phthalate (MEHP) are the most extensively characterised. In utero and peripubertal DEHP exposure in rodents (≥300 mg/kg/day) produces multinucleated germ cells, tubule atrophy, and reduced sperm counts. In SC models, MEHP at 10–100 μM stimulates ROS, induces autophagosome accumulation, and impairs autophagosome-lysosome fusion, producing blocked autophagic flux, SC apoptosis, and BTB breakdown [56]; both the Fas/FasL death-receptor pathway and the intrinsic mitochondrial apoptotic cascade (cytochrome c → caspase-9/-3) are activated downstream of DEHP-induced ROS [57].

4-Nonylphenol, a degradation product of alkylphenol ethoxylates, induces apoptosis and autophagy in SC lines through AMPK and JNK activation and AKT/mTOR inhibition, with ROS as the upstream signal [58]. BPA and its analogues (BPS, BPF) impair mitochondrial function and generate oxidative stress in testicular preparations, altering TJ protein expression and producing germ cell loss [59]. The dose-response behaviour of these EDCs is complicated by non-monotonic relationships; low-dose exposures during sensitive windows may produce effects not predicted from high-dose studies. Consequently, applying traditional meta-analytic pooling that assumes strict monotonicity or log-linear relationships to these data often proves problematic. Such assumptions can dilute significant low-dose effects, artificially inflate inter-study heterogeneity due to differing exposure ranges across primary studies, and ultimately underestimate the true reproductive risk. To overcome these methodological limitations, recent epidemiological syntheses and mixture-modelling approaches eschew strict linear assumptions. For instance, dose-response meta-analyses increasingly utilize restricted cubic splines to capture non-linear trends, while mixture models—such as weighted quantile sum regression, quantile g-computation, and Bayesian kernel machine regression (BKMR)—allow for flexible, non-monotonic fitting. These approaches demonstrate that combined DEHP and dibutyl phthalate metabolite exposure is negatively associated with sperm concentration, motility, and morphology even at individual concentrations at or below regulatory reference doses [60].

Human biomonitoring places these findings in context. Urinary phthalate and bisphenol metabolites within the current range are associated with poorer semen quality across multiple systematic reviews [8]. Mixture-modelling approaches WQS regression, quantile g-computation, and BKMR show that combined DEHP and dibutyl phthalate metabolite exposure is negatively associated with sperm concentration, motility, and morphology even at individual concentrations at or below regulatory reference doses [60]. Late puberty and early adulthood represent windows of heightened EDC vulnerability, consistent with the role of SC ROS in disrupting BTB maturation and autophagic flux [61].

5.3. Pesticides and herbicides

Organophosphate and carbamate insecticides generate ROS in SCs and germ cells, impair ETC function, and activate intrinsic and extrinsic apoptotic cascades [62]. Glyphosate-based herbicides disrupt the BTB and downregulate TJ proteins in rodent testes via NOX1-dependent ROS and MAPK activation [62]. Atrazine inhibits the glycolytic enzymes HK2 and LDHA in SCs, suppressing lactate export to germ cells; this metabolic insult is accompanied by increased ROS and LPO, indicating that glycolytic and mitochondrial injury co-occur and amplify one another [63]. Studies in agricultural workers, greenhouse applicators, and men in high-use regions consistently show reduced sperm counts and motility, elevated DNA fragmentation, and altered serum FSH and Inhibin B in proportion to biomarkers of organophosphate and pyrethroid exposure [64].

5.4. Heavy metals and metalloids

Metals produced the largest pooled effect sizes (SMD > 2.0) in the meta-analysis, consistent with their dual capacity to generate ROS through Fenton and Haber-Weiss chemistry and indirectly through ETC disruption and antioxidant enzyme inhibition. Cadmium (Cd) accumulates preferentially in SCs and BTB-adjacent endothelial cells, stimulating autophagosome formation but blocking autophagosome-lysosome fusion, producing defective autophagic flux, ROS accumulation, and apoptotic activation including Bax translocation and caspase-3 cleavage [65,66]. Manganese (Mn) activates ferroptosis in SC lines by upregulating lipid ROS and suppressing GPX4 and SLC7A11 [67], identifying ferroptosis — alongside apoptosis and defective autophagy — as a metal-driven SC death modality. Lead and arsenic similarly disrupt mitochondrial function, deplete GSH, and elevate LPO biomarkers. Occupational exposures in battery manufacturing, smelting, welding, and mining generate metal burdens approaching or exceeding the biologically active concentrations identified in SC experimental models, supporting translational relevance for occupational reproductive hazard assessment.

5.5. Mycotoxins

Mycotoxins are widespread contaminants of cereal crops and animal feed with documented reproductive toxicity. Zearalenone combines oestrogenic receptor signalling with direct oxidative action: in goat and rodent SC models it activates the p53/Bax apoptotic axis, induces mitochondrial dysfunction and LPO, and drives SC and germ cell apoptosis via the intrinsic caspase cascade [68]. T-2 toxin produces marked ROS generation, mitochondrial swelling, and ultrastructural disorganisation in SCs, causing severe seminiferous epithelial degeneration at dietary-range doses [69]. Fumonisin B₁ couples metabolic and oxidative injury by inhibiting HK2 and LDHA while inducing mitochondrial ROS and fragmentation; mitochondria-targeted quercetin rescues fumonisin B₁-exposed SCs by reducing mitochondrial ROS, restoring HK2/LDHA activity, and preventing apoptosis, mechanistically confirming the mitochondrial origin of this injury [33,70].

5.6. Emerging contaminants: microplastics, nanoparticles, flame retardants, and therapeutics

Emerging contaminants are increasingly detected in human reproductive tissues. Microplastics and nanoplastics, recovered from human testicular tissue and seminal plasma in recent biomonitoring, produce oxidative stress and BTB disruption in rodent models [71]; in vitro, micro- and nanoplastic particles and associated additives, including organophosphate flame retardants, induce ROS and mitochondrial dysfunction in SC preparations [72]. Metal-based nanoparticles (silver, titanium dioxide, zinc oxide) penetrate biological barriers, accumulate in testis, and generate sufficient oxidative stress to impair spermatogenesis in rodents. Therapeutic cytotoxic agents (cyclophosphamide, busulfan) cause severe SC oxidative injury and DNA alkylation, deplete SC number, and compromise the post-treatment spermatogenic niche [73] (see Table 1).

Table 1.

Physiological versus pathological ROS in the male reproductive system.

Context	ROS level	Main source	Impact on male reproduction
Physiological capacitation	Low–moderate	Sperm NADPH oxidase, mitochondria	Promotes capacitation, acrosome reaction, fusion
Acute, compensated stress	Moderate	SC mitochondria, leukocytes	Adaptive antioxidant upregulation, reversible changes
Chronic toxicant/heat exposure	High	NADPH oxidase + mtROS	BTB disruption, SC apoptosis, germ cell loss
Severe pathology (SCOS, I/R)	Very high, focal	mtROS, iron-driven lipid ROS	Ferroptosis/pyroptosis, irreversible spermatogenic failure

Across all six classes, the upstream ROS-generation mechanism varies by toxicant — NOX activation, ETC disruption, direct radical generation, autophagy blockade, or GPX4 suppression — while the downstream cellular signature of BTB disruption, metabolic reprogramming, and PCD activation is shared. This structural convergence underpins the quantitative meta-analytic signal in Section 6. A class-by-class summary of representative stressors, models, ROS readouts, and key pathways is provided in Table 2.

Table 2.

Environmental and chemical stressors that induce oxidative stress and dysfunction in Sertoli cells.

Pollutant	Species/cell model	Typical dose & duration	Main ROS readouts	Key pathways/targets	SC outcomes	Reproductive consequences
PM2.5	Mice (C57BL/6); TM4 cells	In vivo: 4.8, 43.2 mg/kg; In vitro: 25–100 μg/mL (24 h)	N/A	Inhibin B hypermethylation; p21/Caspase-3	Apoptosis; vacuolization	↓ Sperm motility/offspring; ↓ Testosterone; Spermatogenesis inhibition [74]
DEHP/MEHP	Mice/TM3 cells	300 mg/kg (PND21-49); 300 μM (72 h)	↑ROS (NAC rescue)	STAT5B; ↓p-ERK1/2; ↓mitoSTAT3	Leydig cell senescence	↓Testosterone; Premature testicular senescence [75]
	Rat fetal testis (explant)	10−5 M (1-3 days)	N/A	Leydig cell mislocation; ↓Testosterone; ↓AMH; ↓Gonocyte proliferation; ↑Apoptosis	Unaffected number; Vacuolated; ↓AMH	Testis morphology/function disrupted; ↓Testosterone [76]
Atrazine (ATZ)	Mice/TM4 cells	0.3–3000 μg/L (24 h)	No change	↓LDH; Impaired glycolysis	↓Metabolic activity; ↓Glycolysis; ↓Oxidoreductases	Impaired nutritional support; Male infertility risk [63]
Nonylphenol (NP)	GC-1 spg cells	Dose-dependent (N/A)	↑ROS,	ROS/JNK; c-Jun; ↑JNK/MKK4/p53/p38; ↓ERK1/2/MEK1/2	↓Viability; ↑Apoptosis	Testicular disruption; ↓Testosterone (implied) [77]
4-Nonylphenol (NP)	Rat SCs	20-100 mg/kg/2d (IP, 20d)	N/A	AMPK↑; TSC2-mTOR-p70S6K/4EBP1↓; Beclin-1/LC3-II/Atg3/5/7/12↑	↑Autophagy (autophagosomes)	Testicular impairment; ↓Spermatogenesis [78]
Cadmium	Mice/Sertoli cells (Atg5−/− & WT)	2.0 mg/kg (0-24 h)	N/A	↑SC autophagy (Atg5/Atg7/LC3B); ↓Caspase3/PARP	↑Autophagy (LC3 puncta)	↓Germ cell apoptosis (protective) [79]
	SCs	N/A	↑ROS	Autophagic flux blockade (↑LC3-II; ↓Lysosome function/fusion)	↑Apoptosis	Male infertility (implied) [65]
Zearalenone, mycotoxins	Dairy goat SCs	In vitro: 25–100 μM (24 h)	↑ROS, mitochondrial damage	ROS; Caspase; LC3-I/II	↓Viability/MMP; ↑Apoptosis/Autophagy	Male reproductive damage [68]
Heat stress/hyperthermia	Goat SCs (in vivo & primary)	N/A	↑ROS	ROS-DRP1 fission axis; ↑FIS1/DRP1; ↓MFN1/MFN2/OPA1; Mitophagy blockade	↑Apoptosis; ↓ATP/MMP; Excessive fission; Mitophagy flux blocked (Melatonin rescues)	Testicular injury; Abnormal spermatogenesis [80]
Bacteria, Leukocytospermia	Human sperm, Leukocytes	Cut-off: >1 × 106/mL WBC; >103 cfu/mL	↑ROS	Inflammation; Cytokines; Spermatotoxic substances	N/A	Sperm dysfunction; ↓Fertilization/Pregnancy [40]
Mn	Mice/TM3 & TM4 cells	LC50: 230 μM (TM3), 170 μM (TM4)	↑ROS (TM4)	Apoptosis & Ferroptosis; SLC7A15/HMOX1; ↑Caspase3/Casp8ap2/GPX4; ↓SLC7A15	Apoptosis (Z-VAD-FMK rescue) & Ferroptosis (Fer-1 rescue) in TM4; ↓Viability	Spermatogenic damage; ↓Testosterone (implied) [67]

6. Quantitative evidence synthesis

6.1. Study selection and risk of bias

The systematic search retrieved 1458 records from PubMed, Embase, Web of Science, and Scopus. After deduplication and two-stage screening, 166 full texts were assessed for eligibility (Fig. 1). Of these, 96 studies met inclusion criteria for qualitative synthesis and 29 provided extractable quantitative data for meta-analysis. Studies excluded from meta-analysis (n = 67) most commonly lacked variance data, presented only relative change without control group statistics, or used non-comparable exposure contrasts. The 29 meta-analysis studies spanned six chemical classes and three evidence streams: in vivo rodent models (n = 18), in vitro or ex vivo SC preparations (n = 8), and in vivo livestock models (n = 3). Human epidemiological studies contributed to the qualitative synthesis but did not meet quantitative data requirements for meta-analysis. The most frequent risk-of-bias concerns were incomplete confounding control in human studies and limited reporting of randomisation and blinding in experimental studies; overall risk of bias was low to moderate for most studies, and excluding high-risk studies did not materially change the direction of pooled effects.

6.2. Convergent oxidative stress across chemical classes

All six toxicant classes significantly elevated SC oxidative stress relative to controls under random-effects modelling (Fig. 2A; all p < 0.05). Metals produced the largest pooled SMD (>2.0), followed by air pollutants (SMD ∼1.8), mycotoxins, pesticides, EDCs, and emerging contaminants. Crucially, the direction of effect was uniformly positive across all 29 individual studies: no study reported a decrease in oxidative markers or a compensatory increase in antioxidant capacity relative to vehicle controls. This directional universality is the most robust finding of the meta-analysis, independent of heterogeneity. Between-study heterogeneity was high for all classes (I² > 50%), reflecting expected variation in species, dose, exposure duration, biomarker choice, and assay methodology; τ² values are reported in Supplementary Table S4. Egger's regression for the two classes with ten or more studies did not detect systematic small-study effects.

Fig. 2.

Quantitative evidence synthesis based on AOP and One Health frameworks. (A) Forest plot of pooled standardised mean differences (Hedges' g) for SC oxidative stress markers (ROS, MDA) across 29 included studies, stratified by toxicant class; all classes show a significant increase (SMD > 0, all p < 0.05). (B) AOP verification plot showing the cross-study ecological linear regression between relative ROS levels (KE1) and SC apoptosis rates (KE2; R² = 0.77, p < 0.001); this is an associative inter-study correlation and does not establish within-cell causality. (C) One Health subgroup analysis comparing pooled effect sizes of SC oxidative injury across human, rodent, livestock, and in vitro evidence streams, demonstrating cross-species directional conservation.

6.3. AOP verification: cross-study ecological regression of KE1 and KE2

To evaluate the associative coherence of the KE1-to-KE2 linkage, we performed a cross-study ecological regression across the 29 meta-analysis studies, relating KE1 to KE2. A strong positive linear association was observed (R² = 0.77, p < 0.001; Fig. 2B): studies reporting the largest toxicant-induced ROS burdens also showed the highest SC apoptosis rates, and this relationship held across chemical classes. This correlation supports the biological coherence of the AOP linkage. It is, however, an inter-study ecological association subject to confounding by chemical class, dose, species, and design, and should not be interpreted as within-cell mechanistic proof; direct validation requires controlled modulation of KE1 magnitude with proportional tracking of KE2. The association was retained in sensitivity analysis excluding high-risk-of-bias studies.

6.4. One Health subgroup analysis: cross-species conservation

Subgroup analysis by evidence stream revealed broadly comparable pooled SMDs across rodent in vivo models, livestock (porcine and caprine) in vivo models, and in vitro SC cell-line preparations (Fig. 2C). In vitro models showed modestly higher effect sizes, consistent with direct chemical exposure without pharmacokinetic buffering by systemic absorption, distribution, and metabolism. The directional consistency across these phylogenetically distinct systems supports the One Health inference that toxicant-induced SC oxidative stress is a conserved hazard mechanism that can support cross-species extrapolation in reproductive risk assessment (see Fig. 5).

Fig. 5.

Environmental and occupational toxicants converging on SC oxidative stress. Six toxicant classes, air pollutants (PM₂.₅), EDCs, pesticides/herbicides, metals/metalloids, mycotoxins, and emerging contaminants and the class-specific upstream mechanisms by which each activates NOX enzymes, disrupts the mitochondrial ETC, inhibits glycolytic enzymes, or blocks autophagic flux, converging on a shared downstream signature of elevated ROS, BTB disruption, metabolic reprogramming, and PCD.

7. ROS-dependent signalling and programmed cell death in sertoli cells

7.1. Upstream redox-sensing signalling nodes

Toxicant-generated ROS activate several redox-sensitive cascades that converge on PCD execution and BTB disruption (Fig. 6). The MAPKs (JNK, p38, ERK) are activated downstream of PM₂.₅, arachidonic acid metabolites, and EDCs; their effectors promote F-actin disassembly, junctional protein endocytosis, and BTB opening [81]. The AMPK/AKT/mTOR axis integrates redox status with autophagy: ROS activate AMPK and suppress AKT/mTOR, shifting SCs toward autophagy induction, a response exploited therapeutically by curcumin and SC-derived exosomes [81,82]. The Nrf2/Keap1 element induces compensatory SOD, CAT, GPx, and GSH synthesis under acute stress but is overwhelmed by chronic or high-dose exposure [50]. Sustained ROS additionally activate NF-κB and NLRP3 inflammasome assembly, priming SCs for pyroptosis [83] (see Fig. 7).

Fig. 6.

ROS-dependent signalling and programmed cell death pathways in Sertoli cells. Toxicant-induced ROS activate MAPKs (JNK, p38, ERK), the AMPK/AKT/mTOR axis, Nrf2-Keap1 antioxidant responses, and NF-κB/NLRP3 inflammasome signalling, converging on four PCD modes: intrinsic and extrinsic apoptosis (Bax/Bcl-2, Fas/FasL, caspase-9/-3); defective autophagy (autophagosome accumulation, impaired lysosomal fusion, p62 accumulation); ferroptosis (GPX4 inactivation, iron-driven lipid ROS, GSH depletion); and pyroptosis (NLRP3 → caspase-1/4 → gasdermin D → IL-1β/IL-18). The combined outcome is BTB disruption, SC metabolic reprogramming, and secondary germ cell loss.

Fig. 7.

Experimental interventions targeting SC oxidative stress and implications for reproductive toxicology. Overview of antioxidant and nutraceutical strategies (curcumin, quercetin, chlorogenic acid, melatonin, vitamins E and C), MSC-based therapies and derived exosomes, and physical interventions (photobiomodulation, heat preconditioning) that mitigate SC oxidative injury and restore BTB integrity, supporting incorporation of SC-specific oxidative endpoints into reproductive toxicity testing and development of human-relevant in vitro models including BTB-on-chip platforms.

7.2. Bioenergetic crisis, defective autophagic flux, and ferroptosis

Under moderate stress, autophagy and mitophagy are protective, limiting ROS by eliminating dysfunctional mitochondria. However, the immense energetic demands of SCs make them acutely vulnerable to bioenergetic failure. As toxicant-induced ROS progressively damage the mitochondrial electron transport chain, cellular ATP levels plummet, precipitating a profound bioenergetic crisis. This energetic collapse serves as a crucial missing link in the AOP model, fundamentally tethering defective autophagic flux and ferroptosis to terminal cell death. Because late-stage autophagy—specifically autophagosome-lysosome fusion and lysosomal acidification—is highly ATP-dependent, the energetic deficit directly stalls this process. Consequently, toxicants like DEHP, Cd, and heat stress stimulate autophagosome biogenesis but block lysosomal degradation, producing massive vacuole accumulation, lipid accumulation, and unresolvable organellar stress [80,84,85]. Defective autophagic flux thus represents not merely an isolated signaling event, but a direct manifestation of the underlying bioenergetic crisis. Simultaneously, this metabolic crash exacerbates ferroptosis. The depletion of NADPH and glutathione during the energetic crisis cripples GPX4 activity, allowing iron-catalysed lipid peroxides to accumulate and further devastate membrane integrity, particularly in mitochondria. Mn and related metals actively exploit this vulnerability by suppressing SLC7A11, sensitising SCs to ferroptotic execution [67,86].

7.3. Execution of apoptosis as the terminal bioenergetic outcome

When the aforementioned bioenergetic crisis, compounded by defective autophagic clearance and ferroptotic membrane damage, exceeds the point of no return, SCs are forced into terminal apoptosis. This execution is driven predominantly through the intrinsic mitochondrial pathway: unresolvable energetic collapse and ROS accumulation cause mitochondrial outer membrane permeabilisation, cytochrome c release, and caspase-9/-3 cleavage [57]. Furthermore, the extrinsic Fas/FasL pathway can amplify this apoptotic signal from toxicants like DEHP, zearalenone, and chemotherapeutics [57]. Thus, apoptosis in SCs acts as the terminal consequence of a catastrophic bioenergetic failure rather than an independent oxidative outcome.

7.4. Pyroptosis and inflammatory amplification

Pyroptosis, executed by gasdermin D after NLRP3-driven caspase-1/4/5/11 activation, has been implicated in severe SC loss. Elevated caspase-1/4 and IL-1β/IL-18 are documented in testicular tissue from men with Sertoli cell-only syndrome (SCOS), and PM₂.₅ and mycotoxins activate NLRP3 in SC models, amplifying injury through an inflammatory feed-forward loop that extends damage to the surrounding microenvironment [87,88]. These four PCD modalities collectively provide the mechanistic basis for the strong KE1-KE2 association identified in Section 6.3.

8. Critical exposure windows, mixture effects, and translational evidence

8.1. Mixture toxicity and mechanistic convergence

Real-world exposures are inherently multichemical. Because diverse toxicants converge on SC mitochondria and NOX enzymes as shared upstream ROS sources, their combined effects may be additive or supra-additive even when individual concentrations are below single-chemical no-observed-adverse-effect-levels [2,89]. For example, recent in vitro and in vivo models demonstrate that co-exposure to heavy metals, such as cadmium and lead, synergistically exacerbates SC mitochondrial dysfunction, culminating in a supra-additive burst of ROS and accelerated germ cell apoptosis compared to isolated exposures. Similarly, combinations of ubiquitous EDCs, such as BPA and DEHP, have been shown to collaboratively disrupt the BTB; their concurrent presence amplifies intracellular oxidative stress and suppresses tight junction protein expression more profoundly than either chemical alone, driven by this shared ROS-mediated pathway. This 'cocktail effect' is not captured by single-chemical regulatory assessments that assume chemical independence. EDCs and redox-active compounds additionally exhibit non-monotonic dose-response relationships, whereby low-dose exposures during critical windows alter epigenetic programming and SC paracrine signalling through mechanisms diverging from high-dose toxicity [89]. Mixture-aware frameworks are equipped to evaluate cumulative SC oxidative burden. For instance, recent human biomonitoring cohorts applying BKMR have concretely illustrated that concurrent exposure to multiple phthalate metabolites and phenols exerts a cumulative negative impact on semen parameters and SC-specific biomarkers like Inhibin B, underscoring the necessity of these statistical models. Therefore, these advanced frameworks should be routinely applied to reproductive-hazard characterisation in environmentally relevant co-exposure scenarios.

8.2. Translational evidence and non-invasive biomarkers

Direct measurement of SC oxidative stress in living humans requires testicular biopsy and is not feasible in population studies, but substantial indirect evidence is available. Men with high exposure to PM₂.₅, pesticides, metals, or EDCs consistently show reduced sperm counts, motility, and morphology, elevated DNA fragmentation, and altered serum FSH and testosterone, consistent with selective SC targeting [90]. Testicular biopsies from infertile men with SCOS or severe oligozoospermia show elevated MDA, 8-OHdG, activated caspase-3, and BTB disruption markers in SCs, confirming that the experimental oxidative phenotype is present in clinical reproductive failure [87]. Serum Inhibin B and Anti-Müllerian Hormone (AMH) serve as SC-specific functional surrogates: recent studies report negative correlations between environmental phthalate and PM₂.₅ exposures and serum Inhibin B even in men with morphologically normal semen, suggesting SC dysfunction precedes spermatogenic failure detectable by conventional semen analysis [90]. Genetic polymorphisms in SOD2 and GPX1 modify the association between chemical exposures and semen quality, confirming that individual antioxidant capacity shapes susceptibility to SC oxidative hazard [91].

9. Protective interventions and risk assessment implications

9.1. Antioxidant and mitochondria-targeted agents

A convergent body of evidence shows that targeting SC mitochondrial ROS or restoring the antioxidant network rescues SC function under diverse toxicant challenges (Table 3). Curcumin activates the AMPK/SIRT3/SOD2 axis, reducing mtROS, suppressing NLRP3 assembly, and preserving claudin-11 and ZO-1 expression in aged and oxidatively stressed SC models, maintaining BTB integrity [82,92]. Mitochondria-targeted quercetin rescues fumonisin B₁-exposed and heat-stressed SCs by reducing ROS and LPO, restoring HK2/LDHA glycolytic activity, and preventing apoptosis, with superior efficacy to untargeted quercetin [93]. Chlorogenic acid, melatonin, and classical antioxidants attenuate ROS, preserve mitochondrial membrane potential, and maintain TJ protein expression across multiple toxicant models [94]. Beyond classical antioxidant defense, emerging therapeutic strategies increasingly target the bioenergetic crisis by explicitly rescuing the SC glycolytic profile. Because developing germ cells are strictly dependent on SC-derived lactate, pharmacological interventions that directly restore glycolytic flux are indispensable for re-establishing spermatogenesis. For instance, metabolic modulators such as NAD + precursors replenish cellular NAD + pools, activate sirtuin signaling, and de-repress critical glycolytic enzymes, effectively restoring lactate secretion under extreme oxidative stress. Similarly, energy-sensing modulators like metformin and specific PPAR-γ agonists have been demonstrated to robustly upregulate glucose transporters and LDHA in SCs, counteracting toxicant-induced metabolic paralysis. By strategically coupling mitochondrial ROS scavenging with the direct pharmacological reactivation of glycolysis, these combinatorial interventions ensure both the structural preservation of the BTB and the resumption of essential nutritional support for germ cell survival.

Table 3.

Protective compounds and therapies targeting oxidative stress in Sertoli cells.

Intervention	Model	Main mechanisms	Effects on SCs	Effects on spermatogenesis/fertility
Curcumin	Aged/oxidative SC models	AMPK/SIRT3/SOD2 activation, ↓mtROS, ↓NLRP3	Preserves TJ proteins, ↓ apoptosis	BTB integrity improved, spermatogenesis preserved
Quercetin (incl. mito-targeted)	SCs exposed to FB1, heat stress	Antioxidant, mitochondrial protection, glycolytic restoration	↓ ROS/LPO, ↑ HK2/LDHA activity	Germ cell viability & sperm parameters improved
Chlorogenic acid	Heat-stressed porcine SCs	Antioxidant, anti-apoptotic	↓ ROS, ↓ caspase activity	Testicular damage reduced
Melatonin & classic antioxidants	Multiple models	Direct ROS scavenging, anti-apoptotic	Mitochondrial protection, TJ maintenance	Sperm quality and DNA integrity improved
MSCs (amnion/placenta)	Busulfan or chemo-induced toxicity	↓ apoptosis/oxidative stress, ↑ autophagy	Restores SC number & function	Spermatogenesis and fertility restored
BMSC-derived exosomes	Aged testes	AMPK/mTOR modulation, ↓ ROS/NLRP3	Rescues BTB impairment	Spermatogenic microenvironment improved
MgH2 nanomaterial	Irradiated mice	Hydroxyl radical scavenging	↓ SC oxidative damage	Sperm parameters and fertility improved
Photobiomodulation	Hyperthermia models	↓ ROS, ↑ GSH, anti-inflammatory	SC & BTB protection	Spermatogenesis recovery

9.2. Mesenchymal stem cells, exosomes, and photobiomodulation

Stem cell-based approaches address severe SC depletion. Amnion and placenta-derived mesenchymal stem cells transplanted into busulfan-ablated testes reduce oxidative stress and apoptosis, enhance autophagic flux, and partially restore spermatogenesis [73]. Bone marrow MSC-derived exosomes activate AMPK/mTOR signalling, suppress ROS and NLRP3 activation in aging testes, and improve BTB integrity [95]. Photobiomodulation reduces SC ROS, increases GSH, and preserves BTB structure in hyperthermia models [96]. These results confirm that SC oxidative injury is mechanistically reversible and that ROS-targeting strategies at the SC level can restore spermatogenic competence.

9.3. Implications for reproductive hazard assessment

The findings support concrete changes to reproductive toxicity testing strategy. First, SC-specific mechanistic endpoints — mtROS, LPO markers (MDA, 4-HNE), TJ protein integrity (claudin-11, ZO-1, occludin), autophagic flux markers (LC3-II/I, p62), and ferroptosis and pyroptosis indicators (GPX4, gasdermin D) — should be incorporated into tiered in vitro and in vivo test batteries alongside conventional histopathology and sperm parameter assessment. Second, the quantitative AOP developed here, with meta-analytic effect size data supporting both KE1 quantification and the KE1-to-KE2 associative linkage, provides a framework for deriving mechanistic point-of-departure estimates from SC oxidative endpoints. Third, SC lines, testicular organoids, and BTB-on-chip platforms that recapitulate SC junctional and oxidative responses provide validated in vitro alternatives, supporting partial replacement of whole-animal studies. Fourth, the convergent ROS mechanism provides a regulatory rationale for cumulative mixture risk assessment: chemicals acting on the same downstream hazard node should be evaluated for additivity rather than against single-chemical reference doses.

10. Limitations

These mechanistic and quantitative findings must be interpreted in light of crucial methodological limitations. Foremost, the substantial attrition of studies from the qualitative synthesis (n = 96) to the quantitative meta-analysis (n = 29), primarily necessitated by unreported variance data, introduces a potential risk of selection bias. Because studies with non-significant or marginal findings may be less likely to publish complete dispersion metrics, the pooled effect sizes might overestimate the true oxidative burden, thereby limiting the broader generalisability of the meta-analytic estimates across all evaluated chemical classes. Furthermore, the cross-study ecological regression used to substantiate the linkage between KE1 and KE2 is strictly associative. This methodological approach is highly susceptible to the ecological fallacy, as aggregate inter-study correlations do not inevitably mirror true intracellular causal mechanisms. Variations in experimental design, species-specific resilience to oxidative stress, and disparate exposure durations across the included primary studies act as critical confounders that could artificially inflate or obscure the strength of the KE1-to-KE2 associative linkage. Consequently, while the synthesis strongly supports a convergent hazard mechanism, these quantitative relationships should be validated through targeted, multi-endpoint experimental models rather than relying solely on retrospective cross-study pooling.

11. Conclusions

This systematic review and meta-analysis provides the first quantitative, cross-chemical-class evidence that SC oxidative stress is a convergent, chemically non-specific, and phylogenetically conserved endpoint linking diverse environmental hazardous substances to impaired spermatogenesis. Across 29 meta-analysis studies spanning six toxicant classes, three evidence streams, and multiple species, toxicant-induced ROS elevation was the universal directional response in SCs: no study reported a contrary finding. Metals produced the largest effect sizes (SMD > 2.0) and air pollutants the second largest (SMD ∼1.8). Cross-study ecological regression supports the AOP associative linkage from ROS elevation to SC apoptosis and BTB disruption, and the One Health subgroup analysis supports the translational framework linking experimental SC toxicology to human reproductive risk.

These findings position SC ROS measurement as an early, chemically inclusive, and mechanistically grounded biomarker for reproductive hazard characterisation. Priority research directions include: (i) formalisation of this AOP within the OECD AOP-Wiki with explicit key event relationship weighting and Bradford Hill weight-of-evidence scoring; (ii) systematic evaluation of mixture effects on SC ROS at environmentally relevant multichemical concentrations; (iii) development and regulatory validation of standardised SC-based in vitro assays including BTB-on-chip models; and (iv) prospective epidemiological studies correlating serum Inhibin B and AMH with systemic ROS biomarkers in environmentally exposed populations, establishing non-invasive translational biomarker pathways for SC oxidative hazard monitoring at population scale.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the author used Claude (Anthropic) to assist with language editing and manuscript restructuring. After using this tool, the author reviewed and edited the content as needed and takes full responsibility for the content of the published article.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

No data was used for the research described in the article.