The ENDOMIX perspective: how everyday chemical mixtures impact human health and reproduction by targeting the immune system†.

Hormones are biochemical messengers that bind to specific receptors in responsive cells, thereby modulating important physiological processes in the body. Among these, lipophilic sex hormones are capable of crossing the plasma membrane and interacting with nuclear receptors to initiate slow genomic responses [ 30 ]. The main reproductive hormones are estrogens, testosterone, and progesterone (P4). In females, P4 and estradiol (E2) regulate the menstrual cycle and subsequently maintain the pregnancy [ 31 ]. During the menstrual cycle, ovaries and adipose tissue produce E2, which in turn stimulates ovulation through the release of luteinizing hormone (LH) [ 32 ]. Concurringly, the corpus luteum and the placenta secrete P4 to prepare the uterus for implantation [ 33 ]. In the event of pregnancy, E2 promotes uterine blood flow, myometrial growth, breast development, and cervical softening [ 32 ]. Remarkably, these steroid hormones are not sex-specific. In males, testosterone synthesis in the testis relies on the effect of gonadotropin-releasing hormone, LH, and follicle-stimulating hormone (FSH). This androgen drives the development of male genitals and secondary sexual characteristics and promotes spermatogenesis [ 34 ]. Nevertheless, the role of sex hormones is not limited to their reproductive function; they also modulate a network of body systems: neural, cardiovascular, musculoskeletal, dermatological, adipose, hematopoietic, and immune [ 35 ]. Therefore, disturbances in the hormonal balance caused by environmental factors, such as chemicals, can lead to severe fertility and reproductive issues and underlie other diseases, specially those related to chronic inflammation. EDC action mechanisms are linked to their capacity to bind SHRs and interfere with endocrine signaling and metabolism. Several of these chemicals can mimic hormones by acting as agonists or antagonists of SHRs [ 18 ]. For instance, BPA and DES are estrogenic compounds capable of binding and activating ERs [ 36 , 37 ], whereas phthalates like diethylhexylphthalate (DEHP) and the fungicide vinclozolin are classified as androgen antagonists or antiandrogenic [ 38 , 39 ]. Other identified EDC action mechanisms involve modifying endogenous hormone signaling, concentration, bioavailability, or turnover. Notoriously, atrazine inhibits cyclic adenosine monophosphate (cAMP)-specific phosphodiesterase-4 leading to cAMP intracellular accumulation, a second messenger mediating cellular responses to many hormones [ 40 ]. Likewise, TCDD and BPA target two main metabolic sensors, the aryl hydrocarbon receptor (AhR) and peroxisome proliferator-activated receptors, respectively [ 36 , 41 ]. Mechanistically, the interaction of chemicals with nuclear receptors (e.g., ER, AR, and AhR) can modulate genomic responses and disrupt epigenetic changes, namely DNA methylation and histone modifications during gamete maturation and fetus development [ 42 , 43 ]. Recently, a systemic approach integrating and classifying EDC modes of action into key characteristics has been developed [ 4 ]. This tool demonstrates how mechanistic data might be of relevance for predicting the toxicological hazard of traditional and emerging chemicals in human populations. EDCs’ impact on sex hormone production, transport, and function has been linked to an increased risk of female and male reproductive outcomes. For example, gestational exposure to phenols, phthalates, and metals (e.g., barium and manganese) was correlated with sex-specific variations in steroid hormone concentrations in both mother and newborn [ 44 ]. In this sense, several epidemiological studies have assessed the adverse effects of single EDCs or simple mixtures during windows of susceptibility. Prenatal exposure to BPA was correlated with an increased risk of preeclampsia [ 45 ] and placental abnormalities [ 46 , 47 ] in women, whereas the mixture of BPA/phthalate metabolites induced adverse birth outcomes and newborn DNA methylation signatures, respectively [ 48 ]. Exposure to other simplified EDC mixtures during fetal development, including PFAS/organochlorine pesticides (OCPs) and parabens/triclosan, has been associated with small-for-gestational-age births and neonatal health issues [ 49 , 50 ]. Furthermore, in utero chemical-derived hormonal disruption can impact later life stages, from child development to puberty and reproductive maturity. Robust population-based data have validated the association between prenatal exposure to EDC mixtures and metabolic disorders in children. Notoriously, the multicenter Swedish Environmental Longitudinal Mother and child Asthma and allergy—SELMA study correlated PFAS, triclosan, phthalates, non-phthalate plasticizers, bisphenols, PAHs, pesticides, and polychlorinated biphenyls (PCBs) measured in 1118 mothers with low birth weight, abnormal weight gain trajectories at age 4, and sex-dependent body fat changes at age 7 [ 51 , 52 ]. In 8–14-year-old male teens, phthalate exposure during the third trimester was inversely correlated with pubic hair amount [ 53 ] and sex hormone levels [ 54 ]. Similarly, exposure to phthalates, parabens, and phenols was associated with hormone concentration fluctuations, adolescent breast density, and sexual maturation in females at age 12 [ 55–57 ]. Other chemical-mediated outcomes are sex-dependent, including lipid and overall body fat profiles [ 52 , 58 ]. Nevertheless, epidemiological causality is influenced by confounding factors (e.g., sex, timing of exposure, chemical type, and mixture interactions) and provides limited insights into EDC molecular targets and action mechanisms. The effect of single and mixed EDCs on reproductive health has been extensively verified in vivo, in both male and female rodents. For instance, male rats exposed to malathion for 60 days produced sperm cells with restricted mobility and increased Bax/Bcl-2-mediated apoptosis [ 59 ]. Similarly, the administration of CdCl 2 or sodium arsenite in rodents for 30 days altered their testicular enzymatic activity and semen quality, which was associated with structural changes in the seminiferous tubules, Sertoli cells, and spermatozoa [ 60 , 61 ]. Exposure to phenols and phthalates at low doses was also detrimental to rat sperm quality [ 62 , 63 ]. Mechanistically, BPA disrupts the hypothalamic–pituitary–testicular axis and ERK signal pathway in sperm cells, while phthalates target steroid hormone metabolism [ 62 , 64 , 65 ]. In female rats, BPA, phthalates, DES, and methotrexate interfere with the process of folliculogenesis by disturbing the hypothalamic axis, anti-Müllerian hormone production, and antioxidant enzyme activities (e.g., superoxide dismutase and glutathione peroxidase) [ 66–70 ]. Remarkably, both persistent and non-persistent EDCs have adverse effects on female fertility. Short-term and chronic exposure to BPA, phthalates, and DES at low and high doses decreased fertility rates, which were associated with disruption of LH, FSH, E2, and P4 levels, zygote replication following in vivo fertilization, and ovary structure in rodents [ 71–74 ]. In sum, in vivo studies underline how different EDC doses, routes, and timing of exposure cause structural and physiological alterations in male and female reproductive systems by targeting the endocrine axis. Future animal studies, including those planned within ENDOMIX, should include complex EDC mixtures as a blueprint for real-life exposure scenarios.

ENDOMIX will integrate and harmonize data from multiple European cohorts, in vitro barrier and target organ models, and multispecies animal models to define how specific chemicals of concern trigger or perpetuate disease ( Fig. 2 ). The strong study design of ENDOMIX eases the analysis of interactions between chemical exposure and cofounder factors to recognize pivotal disease determinants. In this sense, the project encompasses life course exposure to EDCs with a particular emphasis on adverse health effects in vulnerable groups, including pregnant women, children, and adults of fertile age. Furthermore, the potential hazard of in utero exposure to chemical mixtures will be assessed by in vitro placenta barrier models and transgenerational mouse studies. Overall, the ambition of ENDOMIX is to close the existing knowledge gap regarding complex EDC mixtures by clearly defining novel risk assessment tools, toxicological endpoints, and immunological biomarkers. The end goal of the consortium involves translating paradigm-changing knowledge about EDC toxicology into mitigation exposure strategies accessible to policymakers, regulatory bodies, and stakeholders, ultimately improving EU public health and well-being. To achieve this endeavor, ENDOMIX will address the following main objectives: Characterize chemical mixtures (real-life EDCs) within the context of social, lifestyle, and environmental factors. Evaluate immunotoxicity as a central mechanism of EDC action by implementing complex bioassays with immune and non-immune cells. Validate novel findings on multiple European cohorts of mother-child pairs with data on chemical exposure and non-communicable diseases across the life course. Determine how relevant chemical mixtures disrupt molecular mechanisms to perpetuate disease by utilizing in vitro placenta barrier and organoids models, as well as, mouse models in compliance with the 3R rules. Integrate human cohort, in vitro and in vivo data based on advanced modeling strategies and artificial intelligence (AI) algorithms. Concept, approach, and methodologies of ENDOMIX. The ENDOMIX consortium will assess the adverse health effects of real-life EDCs grounded on immunotoxicity as a central mechanism. Human populations are in constant contact with consumer products containing EDC mixtures, which can alter development, growth, and metabolism by targeting the endocrine, immune, and reproductive systems. This project will cover the complete chain of causality, from exposure to disease onset, by validating adverse effects of EDC mixtures of concern in European cohorts during windows of susceptibility (e.g., pregnancy and childhood). Subsequently, the impact of these EDC mixtures on the phenotype and function of immune and target cells will be characterized by using in vitro bioassays, barrier and organoid systems, and animal models. Newly-generated data and information available in databases will be employed to feed machine learning algorithms and develop roadmaps for EDC risk assessment (created with BioRender.com ). The first objective involves the identification of complex and representative EDC mixtures through non-targeted screening and data mining of toxicology literature and existing biomonitoring studies. Likewise, we will evaluate the chemical-specific effects and interactions within the tested chemical mixtures, including concentration addition or “something from nothing” phenomena, by using immunokinetic assays and in silico modeling. The relevant contributors to mixture effects will be retrospectively measured in a subgroup of cohort blood samples to quantify their real-life concentration in the European population. Next, we will test the single components and several designed EDC mixtures at the detected concentration by candidate high-throughput (HT) bioassays [ 226 , 227 ]. Based on these results, a selection of ten chemical mixtures will be prepared and tested using immunoassays with primary human cells, barrier systems, and organ and animal models. To investigate the direct cell toxicity and modulatory capacity of prioritized EDC mixtures on innate and adaptive immune cells, we will implement in vitro bioassays with peripheral blood mononuclear cells from female and male donors. Firstly, we will identify changes in immune cell subset frequencies in response to the chemical mixtures by using mass cytometry (CyTOF—cytometry by time of flight) to perform high-dimensional immune profiling at the single-cell level [ 228 ]. Then, we will design proliferation, activation, and cytotoxic immunoassays with both innate and adaptative immune cell subpopulations markedly susceptible to the EDC mixtures. The target primary immune cells will be isolated and treated with the defined single chemicals and mixtures at real-life concentrations (Objective I) in the presence and absence of cell activation agonists, such as LPS, toll-like receptors (TLR)-like molecules, and anti-CD3/28. Well-established maturation and differentiation protocols will be utilized to prepare in vitro differentiated macrophages, DCs, and lymphocytes, and to assess their functional response to chemical stress. The immune-related readouts will include macrophage polarization into proinflammatory M1-like or anti-inflammatory M2-like states and secretion of proinflammatory (TNF-α, IL-1β, IL-12, and IL-23) or regulatory (IL-10 and TGF-β) cytokines in culture supernatants [ 229 ]; DC maturation and activation measured by the expression of the surface markers CD83, CD86, and MHC-II [ 230 ]; and CD4 + T cell differentiation and plasticity into Th1, Th2, Th17 or Treg programs [ 118 ]. We will conduct the immune cell phenotyping with CyTOF or spectral flow cytometry and quantify cytokines using cytometry bead arrays. Altogether, our systemic approach will provide detailed information about EDC effects on immune cell differentiation, maturation, plasticity, and functionality, thereby increasing the understanding of how chemicals drive disease by directly disrupting immune regulation. We will harmonize mother-child pair data from 15 prospective cohorts to address how exposure to EDC mixtures during pregnancy, childhood, and adolescence impacts the incidence of disease outcomes later in life. The epidemiological association studies will take into account interactions with cofounder variables (e.g., lifestyle, socioeconomic state, sex, and immunodeficiency) and between chemicals, and will encompass health endpoints for which insufficient data is available; such as disorders in the development and function of the nervous, cardiovascular, and immune systems, as well as metabolic syndrome (obesity and diabetes), allergies, autoimmune diseases, hormone-dependent cancers, and abnormalities in sexual maturation (e.g., mini puberty, prepuberty, and puberty). To evaluate the immune-mediated effects of chemical exposure and their underlying biological mechanism during different windows of susceptibility, we will carry on HT proteomic screenings using Olink inflammation panels. This technique incorporates >300 biomarkers, including cytokines, chemokines, and chemokines receptors, among other functional proteins. The adverse effect of chemicals on epigenetic footprint, metabolome, and gut microbiome will also be considered. By this means, ENDOMIX will identify EDC mixtures of concern and potential intermediate biomarkers of exposure based on real-life chemicals already measured in the human cohorts: PCBs, OCPs, PFAS compounds, metals, parabens, phthalates, phenols, and organophosphate pesticides. Through this strategy, we will determine the most sensitive windows of susceptibility to specific profiles of EDC exposure and propose measures to prevent undesirable health effects. We will expand the in vitro analysis of the chemical mixtures to placenta co-culture models, organoids, and barrier models that have been developed in our laboratory. Remarkably, dose-relationship assays and single chemical controls will be considered in the experimental design when needed. ENDOMIX will employ an in vitro trophoblast-based barrier model to access EDC effects on placenta functionality. The placenta barrier model will be exposed to EDC mixtures and co-culture with immune cells that can eventually infiltrate the decidua (NK and T cells), in both steady-state and inflammatory conditions. A second phase will involve 3D placenta models, such as spheroids and organoids generated by culturing trophoblast cell lines and primary cells with adequate extracellular matrices as previously published [ 231 , 232 ]. The organoids will be co-cultured with complementary cell types (e.g., fetal endothelial cells) and immune cells to address EDC mixtures’ effect on cell interactions. Afterward, the following endpoints will be analyzed: cellular viability, migration, invasion, gene expression of cell-specific markers, and hormone production. In the third phase, we will cross-scale chemical mixture doses to animal studies. Special emphasis will be placed on transgenerational 3R-conform mouse models to validate the observed toxic effects on human cohorts and placenta barrier and organoid models. In this manner, we can triangulate reproductive outcomes, alterations in the placenta functionality, and modulation of pregnancy-relevant immune cell populations associated with EDC exposure. Upon chemical exposure of female mice during pregnancy and lactational periods, we will determine changes in metabolic and reproductive health (e.g., vaginal opening, estrus cycle, and sperm) [ 233 ] parameters among the female and male offspring. The impact of EDC in utero exposure on the offspring will also be investigated in settings of induced asthma [ 234 , 235 ] and rheumatoid arthritis [ 236 ]. Altogether, this robust design involving physiologically relevant in vitro bioassays will reduce the use of animals for experimentation and enable exploring the underlying mechanisms governing chemical disruption. The last objective involves adopting the weight-of-evidence (WoE) approach to concatenate, weigh the lines of evidence, and draw conclusions about the immunotoxicity of the tested EDC mixtures. The WoE data input will comprise epidemiological, in vitro, target organ/barrier models, and in vivo results from ENDOMIX, as well as existing literature and database-derived information. Moreover, we will employ the automated strategy Network and Dynamical Reasoning Assembler (INDRA) to discover molecular pathways targeted by EDC mixtures of concern. INDRA classifies and sorts out redundant information from published papers and databases and applies predictive AI models (e.g., natural language processing) to generate a framework of mechanistic networks of disease outcomes [ 237 ]. Accordingly, the resulting INDRA risk statements (e.g., BPA alters androgenic activity) can be implemented in rule-based and network models, such as knowledge graphs, and compiled in a comprehensive guide for EDC risk assessment in terms of human health and reproduction. Overall, ENDOMIX aims to provide novel scientific-based evidence concerning the relevance of updating the toxicological risk assessment paradigm, from the traditional chemical-by-chemical approach toward a more realistic mixture risk assessment (MRA) concept, as previously proposed [ 238 ]. We have established a rigorous system biology strategy that integrates epidemiological, in vitro, and in vivo data to validate the toxicological relevance of MRA. In this sense, we expect to uncover chemical interactions targeting the immune system and posing a risk to human health and reproduction, as a complement to new assessment methods developed by other EU-funded initiatives (EuroMix, EDC-MixRisk, EUToxRisk, HBM4EU, and JRC) [ 239 ]. In concrete, to influence policy-making and the development of EU regulations, we will provide research-based recommendations about hazard and risk assessment. These include a guide of co-exposure patterns for specific consumer groups, depending on lifestyle and diet, as well as hazardous and unintentional chemical mixtures derived from key industries, such as pharmaceuticals, medical implants, tobacco, clothing, and food. To encourage the acceptance of innovative base guidelines, the project involves scientific advisors and a dedicated Stakeholder Board that will advocate for future policies. Likewise, citizens will be invited to participate in dissemination and educational activities, promoting awareness of how to mitigate exposure to real-life chemical mixtures that threaten human health and our natural environment. First, ENDOMIX anticipates to uncover the health impact of realistic EDC mixtures by identifying representative mixtures of concern. Nonetheless, the inferred mixtures may overrepresent EDC that are commonly reported in toxicological databases and HBM studies, neglecting chemicals with short half-lives that can often only be detected in repeated biospecimen measurements [ 240 , 241 ]. This exposure assessment bias could affect the selection of individual EDCs and the preparation of the mixtures for the bioassays. Second, we can solely test the effects of the prioritized single chemicals and their mixture in our complex barrier and organ models due to resources and cost restrictions. Although this can limit the validation of key toxic events for some EDC mixtures, we anticipate partially overcoming the gap via in silico models based on HT screening results and human cohort data. Finally, we will screen and prioritize thousands of potential EDCs focused on the mixture concept of concentration addition, which could overlook modulating or antagonistic chemical interactions. Yet, additivity and synergy are the most frequent mixture effects linked to adverse health outcomes [ 13 ].

The ENDOMIX consortium conducts research with human data and animals under the appropriate legislation declared by the Declaration of Helsinki, ICH guideline for Good Clinical Practice, ISO guidelines on Good Clinical Practice ISO 14155:2020, the European Directive 2001/20/EC on Good Clinical Practice, and the EU General Data Protection Regulation (2016/679).

Given the relevance of hormones for sexual development and maternal immunotolerance, EDC exposure endangers the physiological functions governing reproductive and pregnancy outcomes. Although epidemiological-based studies have correlated chemical exposure with preterm pregnancies and spontaneous abortions [ 121–123 ], EDC’s underlying mechanisms remain elusive. As we describe above, most of the literature regarding the impact of EDCs on reproduction has evaluated causality based solely on statistical correlations between single chemicals and specific health endpoints. Additionally, in vitro and vivo studies mainly focus on narrow immunotoxicity responses, namely immunosuppression and immunopotentiation [ 124 ], while neglecting realistic chemical mixtures from HBM reports. In this section, we summarize the literature regarding the immune modulation of innate and adaptative cells by single EDCs and their mixtures ( Fig. 1 ) and propose how these effects on the immunome could trigger or perpetuate health outcomes, particularly reproductive-related disorders ( Table 1 ). A schematic representation of the impact of EDCs on the immune system. EDCs have complex and heterogeneous structures and can interact with hormone receptors in immune cells. This interaction may result in the activation, blockade, or modulation of the receptor function, altering signaling cascades and ultimately the cell phenotype. EDC-induced immunomodulation encompasses changes in surface protein expression, cytokine profile secretion, and cytotoxic activity. These phenotypic changes are not limited to individual immune cells but can translate to cell-to-cell interactions; for instance, the differentiation of dendritic cells toward a pro-inflammatory or tolerogenic subset will determine the cell fate of CD4 + T cells. Therefore, EDCs can potentially disrupt the immune system regulatory network, contributing to the development of allergies, autoimmune diseases, and reproductive outcomes (created with BioRender.com ). Modes of action, immunomodulatory effects, and associated reproductive outcomes of exposure to common EDCs (continued) Continued As stated above, SHR-expressing immune subsets are not only receptive to endogenous hormone signaling but also susceptible to chemical modulation [ 204 ]. Therefore, EDCs can potentially alter the development, phenotype, and functions of immune cell populations, including those driving maternal immunotolerance, allergies, and autoimmune diseases [ 205 , 206 ]. Chemicals with estrogenic activity can influence the cell fate and function of innate immune cells. It has been described that BPA and parabens enhance histamine release in vitro and in vivo [ 132 , 207 ], which can explain associations between exposure to these substances and increased allergy incidence [ 208 , 209 ]. Among other EDCs, the organotin compound tributyltin (TBT) can modulate the cytokine profile of human primary NK cells, by either stimulating or repressing their secretion of proinflammatory cytokines (e.g., TNFα, IL-1β, and IL-6) in response to differential chemical doses [ 166–168 ]. Furthermore, in vitro treatment with phenols and OCPs diminished NK cell surface protein expression (CD16, CD18, and CD56) and lytic function, abrogating their capacity to bind and eliminate tumor cells [ 135 , 136 ]. During early pregnancy, exposure to mixed BPA and BP-3 negatively impacted fetal and placental development, while increasing the uNK cell numbers in dams at gestational day 14 [ 27 ]. The immunotoxic effect of BPA on macrophage phenotype and function is contradictory. BPA impairs macrophages’ in vitro secretion of TNF-α, IL-1β, and IL-8, while it reduces their IL-6 and nitric oxide (NO) production and phagocytic activity through an estrogen-mediated mechanism [ 210 , 211 ]. In contrast, other studies found that macrophages cultured with BPA produced increased levels of proinflammatory cytokines and low IL-10 and TGF-β concentrations [ 152 , 212 ]. This discrepancy is probably related to the usage of different cell lines (human THP-1 vs. murine RAW 264.7), human donor variations, BPA treatment time points (15 vs. 24 h), and type of cytokine agonist (lipopolysaccharide [LPS] vs. none). Nevertheless, an imbalance between proinflammatory M1 and tolerogenic M2 macrophages can lead to negative pregnancy outcomes, such as pre-eclampsia and fetal growth restriction [ 213 ], and favor the accumulation of apoptotic cells in systemic lupus erythematosus (SLE) patients [ 214 ]. It has been shown that phenol-containing compounds can also alter the phenotype and functions of APCs. For instance, BPA appears to impact DC maturation, differentiation, and function. In vitro studies of DCs treated with BPA have generated opposite findings regarding the expression of HLA-DR and costimulatory molecules (CD80 and CD86) but consistently reported an expansion of CD1a + DCs [ 153 , 215 ]. Interestingly, the CD1a + DC subset accumulates in the inflammatory arthritis synovium and can stimulate T cell effector function by overruling Treg suppression [ 216 , 217 ], a hallmark of rheumatoid arthritis physiopathology [ 218 ]. This highlights how chemical-driven dysruption of DC antigen presentation can potentially influence T cell responses. However, EDC-induced immunotoxicity is not restricted to indirect effects on T cell function through APC modulation, since antigen-specific lymphocytes express SHRs that can be directly targeted by chemicals [ 29 ]. Several in vivo studies have shown that prenatal and adult exposure to single phenols or their mixtures dysregulate CD4 + T cell differentiation and function, triggering allergic-like reactions [ 219–222 ] and reduced fetal development [ 27 ]. In this sense, mice exposed to BPA displayed an augmented Th2/Treg ratio in a dose-dependent manner [ 222 ], whereas nonylphenol induced a similar shift in the Th2/Treg programs and diminished the therapeutic effect of ER blocking in mice with allergic rhinitis [ 223 ]. Mechanistically, these compounds increased the production of IL-4 and antigen-specific IgE in sera by activating the Ca 2+ /calcineurin-dependent nuclear factor of activated T cells (NF-AT) pathway [ 219–221 ]. In contrast, one study described that BPA increased Th1 response and IFN-γ production in mouse splenocytes [ 224 ]. This discrepancy regarding the EDC effect on Th polarization possibly reflects protocol variations associated with dose, purity of cells treated in vitro (T cell vs. splenocytes), and in vivo timing of exposure. Yet, the disruption of Th balance and low frequencies of Tregs in peripheral blood have been linked to implantation failure, abortion, preterm labor, and preeclampsia [ 101 ]. In male mice, Treg depletion promotes exacerbated inflammation in the testis and epididymis and the generation of anti-sperm antibodies inducing lower mobility rates and sperm counts [ 225 ].

The exposure to real-life chemicals and their mixtures is variable and dynamic among human populations, making it challenging to establish causality between EDCs and disease outset. Currently, both observational and experimental strategies to investigate EDC exposure effects present limitations. For instance, epidemiological studies have inherent pitfalls associated with reliable exposure assessment, such as difficulty in measuring EDC exposure during critical periods of life (e.g., fetal development and puberty), latency between the exposure to non-persistent EDCs and outcome onset, correlations with cofounding bias, and lack of unexposed control groups [ 19 ]. Experimental studies using in vitro and animal models assess variable chemical exposure and doses but sometimes their design does not account for non-monotonic dose responses (NMDR), low dose effects, or interactions (e.g., additive, synergetic, or antagonist) within chemical mixtures [ 20 ]. NMDR describes the biological effect of a substance that does not consistently increase or decrease with the dose [ 21 ], arguing against risk assessment strategies that predict chemical low dose safety solely based on extrapolations of high-dose adverse effects [ 22 ]. Consequently, the requirement for a modern strategy integrating epidemiological, risk assessment, and immunotoxicology methodologies has been proposed [ 23 ]. To our knowledge, such research initiatives have not been undertaken. In the past decades, component-based methodologies have been successful in addressing EDC mixture effects. In this approach, a restricted number of chemicals (usually <10) with defined concentrations and modes of action (e.g., antiandrogens and xenoestrogens) are mixed and tested [ 24 ]. However, these simplified EDC mixtures do not reflect the toxicological profiles of real-life co-exposures, involving at least thousands of chemicals. Several human biomonitoring (HBM) Europe-wide studies have measured EDCs in blood, plasma, and urine samples showing the vast heterogeneity of the internal exposome. Remarkably, a cross-sectional German study of children and adolescents (3–17 years old) detected 31 chemical substances above the quantification limit in urine; the components and their concentration (median in μg/g creatinine) in the mixture comprised 5 elements (0.05–6.89), 17 phthalates (0.93–21.36) and 3 substitutes (2.03–4.66), 6 polycyclic aromatic hydrocarbons (PAHs; 0.04–3.15), and BPA (1.6). Another HBM report of Spanish workers above age 16 found similar levels of phthalates (0.62–189.47) and substitutes (0.7–0.43) in urine, as well as perfluorononanoic acid and perfluorodecanoic acid (0.95 and 0.37 μg/L) in blood. Altogether, this underscores the complexity of the internal exposome and the challenges it poses for the risk assessment of EDC toxicity. The toxicological profile of a chemical mixture does not equal the sum of its single component’s effects. Substances within the mixture can display additive or synergistic interactions provoking adverse effects through various modes of action, even when each chemical component is below the no-observable-adverse-effect-level (NOAEL) [ 25 ]. Between 2000 and 2020, about half of in vivo studies on EDC exposure at low doses (<NOAEL) found greater size responses when chemicals were mixed, suggesting a concentration additive effect [ 24 ]. Accordingly, a meta-analysis reported synergism as the main antiandrogenic mixture effect among the immunotoxicity endpoints [ 13 ]. For instance, gestational exposure to the herbicide atrazine combined with different types of EDCs (e.g., perfluorooctanoic acid, BPA, and tetrachlorodibenzodioxin [TCDD]) at doses equivalent to the tolerable daily intake (TDI) or < NOAEL values enhanced sex-specific neurotoxicity compared to the single components [ 26 ]. In contrast, we showed that BPA at TDI seems to antagonize benzophenone-3 (BP-3) negative effect on fetal weight when applied at a single skin dose [ 27 ]; this is plausible as BPA and BP-3 display competitive ER binding [ 28 , 29 ]. Overall, antagonistic interactions can mask single chemical exposure effects leading to underestimating toxicity, whereas additive or “something from nothing” effects of low dose co-exposures can modify the mixture toxicity magnitude increasing the risk of adverse outcomes.

In the last decades, the risk assessment of EDC exposure in humans has been addressed using independent and non-standardized epidemiological and experimental designs. Thus far, most of the studies have focused on single chemicals and neglected their combined adverse effects on hormone-regulated organ systems, such as the immune system. In addition, EDC outcomes among vulnerable populations like pregnant women, newborns, and children are barely considered in risk and hazard assessment or studies addressing chemical impact on health. Therefore, the lack of robust scientific evidence regarding the health impact of real-life EDC mixtures has impaired the establishment of EU regulations to mitigate exposure in susceptible populations. In this sense, the ENDOMIX consortium tackles this knowledge gap with an interdisciplinary and systemic approach that brings together renowned toxicologists, epidemiologists, environmental immunologists, clinical and data scientists, and risk assessment experts. ENDOMIX’s main objective is to provide novel insights into the mechanistic relationship between chemical-driven immune modulation and adverse effects on health and male and female reproduction. Our approach involves characterizing EDC mixtures of concern in large human cohorts and assessing their direct toxic effects on immune cells, organoids, physiological barriers, and transgenerational animal models. The project envisions the triangulation of epidemiological, in vitro, and in vivo data by using statistical and computational strategies to identify intermediate biomarkers and toxicological endpoints of exposure during critical life stages. The end goal of ENDOMIX is to deliver pertinent evidence regarding EDC-induced immunotoxicity and science-to-policy translation guidelines to European society, comprising stakeholders, health professionals, the scientific community, and policymakers. Therefore, enabling the construction of a regulatory framework and empowered decision-making to minimize the EDC burdens on public health and the natural environment.

One of the key factors for a successful pregnancy is maternal immune tolerance, which involves a vivid interplay between the mother, placenta, and semi-allogeneic fetus [ 75 ]. The pregnancy tolerogenic state comprises adaptations in the mother’s immune system at tissue, cellular, and molecular levels [ 76 ]. At the feto–maternal interface, villous and extravillous trophoblasts (EVTs) expressing specific major histocompatibility complex (MHC) molecules regulate immune functions. For instance, EVTs can bind the Ig-like transcript 2 (ILT2) on uterine natural killer (uNK) cells and macrophages via HLA-G [ 77–80 ]. Overall, the regulatory microenvironment surrounding the fetus involves uNK cell subsets, such as decidual (d)NK cells, tolerogenic DCs, and alternatively activated M2 macrophages, as well as regulatory T and B lymphocytes [ 32 ]. The function of these innate and adaptative immune cells during human gestation is discussed next. Immunological tolerance at the maternal–fetal interface relies on the crosstalk between antigen-presenting cells (APCs) and lymphocytes, as well as endocrine-driven functions [ 81 , 82 ]. During the first trimester of pregnancy, cytokine-producing CD56 bright CD16 − uNK are the most abundant leukocytes (50%–70%) [ 83 ] and contribute to embryo implantation and trophoblast differentiation and invasion [ 84–87 ]. uNK cell maturation, proliferation, and secretion of fetal growth-promoting factors are modulated by HLA-G/ILT2 [ 88 ] and estrogen receptor (ER)β activation [ 89–91 ], among other factors [ 92 ]. The second largest leukocyte population (20%) within the decidua are tissue-resident macrophages [ 93 ], which support trophoblast implantation, angiogenesis, and endometrial regeneration following childbirth [ 94 ]. In the uterus, HLA-G/ILT2-ILT4 signaling promotes the differentiation of M2 macrophages that display anti-inflammatory properties [ 95 , 96 ] and scavenger receptors (e.g., CD163 and CD206). The clearance of trophoblast debris and apoptotic cells promotes macrophage regulatory responses via IL-10 production [ 93 , 97 ]. Recently, a subset of decidual CD11c + HLA-DR mid macrophages specialized in regulatory T cell (Treg) induction was described [ 98 ]. Lymphocyte subpopulations orchestrate cellular responses against foreign antigens and mediate immune tolerance toward fetal alloantigens. The balance between CD4 + T helper (Th) cell programs, Th1, Th2, Treg, and Th17 [ 99 ], have been proposed as the central paradigm of fetal immunotolerance [ 100 , 101 ]. Gestation drive the decrease of CCR6 + Th1, Th2, and Th17 cells and the expansion of CCR6 − Th1 cells and CD4 + CD25 hi FoxP3 +/hi Tregs in the decidua [ 102–104 ]. Mechanistically, Tregs suppress the cytotoxic and proinflammatory functions of allogenic-specific effector T cells via inhibitory receptor activation and IL-10 signaling [ 105 , 106 ]. In mice, fetal rejection can be prevented by transferring Treg cells from normal pregnant animals that generate a tolerant micromilieu [ 107 , 108 ]. Moreover, a splenic IL-10-producing B lymphocyte subset (B10) that expands during murine pregnancy has been identified [ 109 , 110 ]. B10/Breg cells prevented fetal rejection when transferred to abortion-prone mice, possibly through IL-10 signaling in DCs [ 111 ]. Sex hormone levels fluctuate throughout pregnancy endorsing an anti-inflammatory immune profile, which comprises M2 macrophages, Tregs, and B10 cells [ 112 ]. Immune cells are susceptible to direct endocrine modulation as they express intracellular receptors for steroid hormones, such as E2, estriol, and P4, whose levels increase during gestation and peak by the third trimester [ 113 ]. Indeed, both forms of ERs (α/β), and the progesterone receptor (PR) are differentially expressed in lymphocytes, macrophages, NK cells, and DCs [ 114 , 115 ]. For instance, ERα is upregulated in T cells, while ERβ is prominently expressed in B cells [ 116 ]. Among innate immune cells, ER and PR signaling inhibits NK cell cytotoxic and proliferative capacity and NF-κB-mediated inflammatory response in macrophages [ 117 ]. In antigen-specific lymphocytes, E2, P4, and human chorionic gonadotropin (hCG) diverging effects seem to define the Th balance; ER activation promotes the inflammatory cell programs Th1/Th17, while P4 and hCG enhance Treg responses [ 117 , 118 ]. In contrast, both hormones favor the expansion of follicular Th and B cells [ 119 ], underscoring the complexity of the endocrine-immune axis. Altogether, the interplay between trophoblast and immune cells via cell-to-cell contact, soluble factors, and hormones results in the establishment of an immunosuppressive milieu in proximity to the fetus. Furthermore, immune cells exhibit a distinct phenotype in the uterine environment compared to the periphery and promote pregnancy-related functions in response to endocrine signals, principally embryo implantation, angiogenesis, trophoblast invasion, spiral artery remodeling, decidualization, placentation, uterine contraction, and placenta regeneration [ 120 ].

Modern human populations are constantly co-exposed to several environmental chemicals that can induce transitory or permanent biological deviations from homeostasis, such as acute cellular responses, epigenetic reprogramming, or even mutagenesis [ 1 , 2 ]. A group of heterogenous substances classified as endocrine-disrupting chemicals (EDCs) are of particular concern because they interfere with endocrine functions, which regulate development, growth, metabolism, and reproduction. These substances consist of naturally occurring molecules (e.g., phytoestrogens) and synthetic compounds utilized in several industrial products, including plastics (bisphenol A [BPA]), plasticizers (phthalates), pharmaceutical agents (diethylstilbestrol [DES]), surfactants (per- and polyfluoroalkyl substances [PFAS]), herbicides (atrazine), and metal elements (cadmium), among others [ 3 ]. EDCs modify not only the production, secretion, transport, and degradation of endogenous hormones but also their binding, signaling, and metabolism in hormone-producing and responsive cells, such as mature reproductive cells and immune cells [ 4 ]. Therefore, EDC exposure can impact human and environmental well-being, raising public health concerns that are being addressed with chemical regulatory frameworks in the European Union (EU). EDC exposure has been linked with an increased risk of neurodevelopmental disorders [ 5 ], metabolic syndrome triggering obesity and diabetes [ 6 ], thyroid dysfunction [ 7 ], male and female infertility [ 8 ], and mammal gland and prostate cancers [ 9 ]. Indeed, these chemicals are of special concern for reproduction and fertility since they mainly target steroid hormone receptors (SHRs) in hormone-responsive cells [ 10 ]. Furthermore, due to the role of immune-endocrine cross-talk during pregnancy and the expression of SHRs in immune cells [ 11 ], the immune system has been recognized as a key mediator of EDC effects [ 12 ]. The association between EDC exposure and disease outcome is frequently taken as demonstrated. However, most of the scientific evidence is limited to the scope of single chemicals or simple mixtures of two or three molecular-related substances but it rarely investigates EDC’s mechanistic mode of action [ 13 ], such as immunotoxicity. A systemic approach to studying EDC should integrate data from prospective human cohorts, complex in vitro bioassays, and adequate animal models while considering the exposure to real-life chemical mixtures during highly susceptible life stages. Most epidemiological and experimental studies have focused on correlating single exposures to EDCs with a specific outcome. However, a principle of “something from nothing” has been described, by which EDC mixtures can provoke biologically relevant alterations at levels that single chemicals do not induce measurable effects [ 14 ]. Thus, investigating the combined effect of EDC mixtures on health and reproduction is physiologically relevant and could provide a more accurate estimation of EDC exposure risk. This is in line with the exposome approach to EDC research, which comprises the assessment of life-course environmental exposures from fetal development onwards [ 15 ]. Here, we propose a more practical strategy involving the analysis of EDC-derived biological disturbances during critical stages of life, namely pregnancy, adolescence, and reproductive age [ 16 ]. This effort requires multicenter studies that harmonize epidemiological data and triangulate target organs and cellular mechanistic studies in the context of relevant EDC mixtures at real-life concentrations. EDC exposure is ubiquitous among global populations and poses a substantial health and economic burden, estimated at €163 billion (1.28% of EU Gross Domestic Product) per year in the EU, from which ~5% is attributed to reproductive disorders [ 17 ]. Even though the high costs of chemicals-induced disease and disability have been documented, there is a huge gap between science and policy that hampers the definition and enforcement of regulations focused on minimizing European citizens’ exposure to EDCs [ 18 ]. Within the funding frame of the EU’s Horizon Health Program, the multidisciplinary project “Understanding how endocrine disruptors and chemical mixtures of concern target the immune system to trigger or perpetuate disease” —ENDOMIX aims to determine immune-mediated health and reproductive impacts of EDC mixtures. This review outlines the conceptualization, methodology, and analytical approach behind ENDOMIX with a particular focus on fertility and reproductive outcomes. Firstly, we provide an overview of how EDCs can interfere with hormonal functions during pregnancy and reproductive development. Next, we summarize relevant literature about immune cell interactions susceptible to chemical stress and their potential involvement in the development of fertility and reproductive disorders. Lastly, ENDOMIX objectives and methodological approach are presented, underlining how novel findings regarding EDC toxicology are needed to bridge science-policy gaps.