Ecological and toxicological assessments of anthropogenic contaminants based on environmental metabolomics.

There has long been a great concern with growing anthropogenic contaminants and their ecological and toxicological effects on living organisms and the surrounding environment for decades. Metabolomics, a functional readout of cellular activity, can capture organismal responses to various contaminant-related stressors, acquiring direct signatures to illustrate the environmental behaviours of anthropogenic contaminants better. This review entails the application of metabolomics to profile metabolic responses of environmental organisms, e.g. animals (rodents, fish, crustacean and earthworms) and microorganisms (bacteria, yeast and microalgae) to different anthropogenic contaminants, including heavy metals, nanomaterials, pesticides, pharmaceutical and personal products, persistent organic pollutants, and assesses their ecotoxicological impacts with regard to literature published in the recent five years. Contaminant-induced metabolism alteration and up/down-regulation of metabolic pathways are revealed in typical organisms. The obtained insights of variations in global metabolism provide a distinct understanding of how anthropogenic contaminants exert influences on specific metabolic pathways on living organisms. Thus with a novel ecotechnique of environmental metabolomics, risk assessments of anthropogenic contaminants are profoundly demonstrated.

Introduction

Metabolomics, one of the latest ‘omics’ technology, aims to characterize a wide variety of small molecules (known as metabolites) in a living organism, subject to external stressors [1]. Development and application of metabolomics grew fast in the last two decades, as reflected by the ever-increasing publications in a variety of research fields (Fig. 1). The biological impacts of metabolomics have been expanded beyond routine identification of simple biomarkers, towards the exploration of molecular mechanisms [2]. With the benefit of probing into organism-environment interactions at the molecular level, metabolomics has shown its capabilities in environmental science, which is termed environmental metabolomics.

Fig. 1

The number of publications and representative research areas during the year 1999–2019 as searched in Web of Science using the theme ‘metabolomics’.

Environmental metabolomics was firstly defined as ‘the metabolomics that investigates both free-living organisms from the nature and lab-reared organisms under nature-simulated conditions’ [3]. In general, it is the practical application of metabolomics that focuses on interpreting the mutual influence between organisms and their surroundings [4]. In comparison to traditional targeted endpoints/assays, environmental metabolomics allows detection and identification of various metabolites, which helps to discover direct alternation of specific metabolic pathways in response to environmental stress [4]. Furthermore, metabolite variations are capable of explaining general observations, such as cortical metabolite changes that lead to abnormal neuron activities in female rat offspring exposed to radioactive iodine during pregnancy [5]. As far-reaching as metabolite profiling, quantitative metabolomics can develop metabolic signatures as biomarkers, to speculate the possible types and contents of pollutants in the environment, which impact on individual organisms. With these advantages, environmental metabolomics may primarily contribute to ecotoxicology through risk assessments of widely-distributed pollutants to the ecosystem.

In this review, we primarily aim to give an overview of the general procedures applied in environmental metabolomics. Then we focus on the representative applications of environmental metabolomics during the years 2015–2020 to profile metabolic responses of a variety of organisms exposed to different anthropogenic contaminants, and further explore their toxicological impacts.

Environmental metabolomics

Environmentally relevant organisms and anthropogenic contaminants

In environmental metabolomics studies, organisms and their living environment are two pivotal factors. Typical applications of environmental metabolomics are the investigations of organismal responses to various environmental biotic and abiotic stressors. Biotic stressors refer to stressors derived from other species or organisms occupying a similar ecological niche, such as predators or male competitors in the same group. Abiotic stressors vary from natural factors (e.g. light, temperature, salinity and desiccation) to anthropogenic factors (e.g. environmental pollutants). Considering the expanding impacts of environmental pollution caused by human activities, recent environmental metabolomics mainly focuses on the toxic effects of anthropogenic contaminants on the ecosystem. These studies can be generally classified into the field of environmental toxicology.

The first step in environmental metabolomics experiments is to select a model organism (either in the field or from the laboratory) and a type of environmental contaminant. Recent environmentally relevant organisms used for environmental metabolomics mainly include animals (rodents, fish, crustacean and earthworms) and microorganisms (bacteria, yeast and microalgae). As shown in Table 1, these studies mainly investigated the ecotoxicity of anthropogenic factors, including heavy metals (HMs), nanomaterials, pesticides, pharmaceuticals and personal care products (PPCPs), persistent organic pollutants (POPs) and other environmental pollutants. The following experimental step is to set the mode of contaminant exposure. The exposure scenario of contaminants usually occurs in laboratory and mimics those under natural conditions, but there is a growing interest in applying environmental metabolomics to complex ecosystems.

Table 1

Representative studies on environmental metabolomics under the stress of anthropogenic contaminants based on animals in 2015–2020.

Stressor	Concentration	Detection method	Organism	Disturbed metabolite/Pathway	Reference
Heavy metals
As	1/10 LD50	UPLC-MS/MS	SD rats	Phenylalanine metabolism; aminoacyl-tRNA, phenylalanine, tyrosine, tryptophan, ubiquinone and terpenoid-quinone biosynthesis	[19]
Cu	25, 62.5, 125 and 187.5 mM	GC-MS	Mussels (Perna canaliculus)	Oxidative stress and apoptosis involved in the transsulfuration pathway; glutathione and taurine metabolism	[20]
Hg	0.77 ± 0.09 mg/kg sediment	1H–13C NMR	Earthworms (Eisenia fetida)	Osmoregulation, amino acid and energy metabolism	[21]
Metal (loid): V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Cd, Sb and Pb	0.01–2745.73 μg/L	NMR	Mosquitofish (Gambusia holbrooki)	Aspartate, histidine, myo-inositol, taurine and choline	[14]
Nanomaterials
AgNPs	0.02 and 1 ppb	GC/LC-Q-TOF-MS	Daphnia (Daphnia similis)	Amino acid (serine, threonine and tyrosine), fatty acid (arachidonic acid) and sugar (d-allose) metabolism; protein digestion and absorption	[22,23]
	8 mg/kga	NMR	BALB/c mice	Antioxidative defence, immunoregulation and detoxification; TCA cycle	[24]
TiO2 NPs	0, 5, 50 and 500 mg/kg	GC-MS	Earthworms (Eisenia fetida)	Glutathione metabolism (glycine and pyroglutamic acid); carbohydrate, protein and lipid synthesis	[25]
GO	0, 0.01, 0.1 and 1.0 mg/L	GC-MS	Zebrafish (Danio rerio)	Amino acid, fatty acid and energy-associated pathways (e.g. TCA cycle)	[26]
Pesticides
Acetamiprid and halosulfuron-methyl	1/10 LC50	GC-MS	Zebrafish (Brachydanio rerio)	Amino acid metabolism; TCA cycle (malic acid and fumaric acid); neurotransmitter balance (glutamic acid, taurine and glycine)	[27]
Imidacloprid	3.36, 1.68, 0.84 and 0.42 ng/cm2 for contact tests; 1.0, 0.75, 0.5, 0.375, 0.25 and 0.05 mg/kg for soil tests	NMR	Earthworms (Eisenia fetida)	Maltose, glucose, glutamate, glutamine, malate, fumarate, ATP, lactate/threonine, myo-inositol, arginine, lysine, tyrosine, leucine, phenylalanine, inosine, isoleucine, betaine, valine, alanine, tryptophan and scyllo-inositol	[28]
Acetochlor, bromoxynil, carbofuran, chlormequat, ethephon, fenpropimorph, glyphosate and imidacloprid	246, 12, 22.5, 35, 22.5, 15.5, 12, 12.5 μg/kg/d	NMR	Wistar Crl:WI(Han) rats	TCA cycle; energy production and storage; lipid, carbohydrate and amino acid metabolism	[29]
Pharmaceuticals and personal care products
Bisphenol A	1 mg/L	NMR	Midges (Chironomus riparius)	Energy metabolism; protein biosynthesis; gluconeogenesis; methionine pathways	[30]
Fluoxetine, N,N-diethyl-meta-toluamide, 17 alpha-ethynylestradiol and diphenhydramine	1 μg/L	GC-MS	Oysters (Crassostrea virginica)	Cellular energetics (i.e., Krebs cycle intermediates), amino acids and fatty acids; anaerobic metabolism; osmotic stress; oxidative stress	[31]
Sulfamethoxazole	10 μg/L	HPLC-HRMS	Mussels (Mytilus galloprovincialis)	Amino acids (aspartate, phenylalanine, valine and tryptophan) involved in osmotic regulation and energy metabolism, some nucleotides (guanosine and inosine) and a carboxylic acid	[32]
Persistent organic pollutants
Benzo [a]pyrene	5 μM and 50 nM	LC-MS/MS	Murine hepatoma Hepa1c1c7 cells	Amino acids, acylcarnitines and glycerophospholipids	[33]
	1 and 10 μg/L	NMR	Oysters (Pinctada martensii)	Energy metabolism; osmotic regulation; immune response	[34,35]
	2.0 mg/kg/d	GC-MS	SD rats	Energy, tyrosine, methionine, cysteine and glutathione metabolism; long-term potentiation	[36,37]
2,2′,4,4′-tetrabromodiphenyl ether	0.1 or 1 mg/kg/d	GC-MS	SD rats	Amino acid (valine, leucine and isoleucine and arginine), lipid, carbohydrate and energy metabolism	[38]
	0, 0.002 and 0.2 mg/kg	NMR	ICR mice	Lipid metabolism	[39]
	1, 10 and 100 mg/kg/d	LC-Orbitrap-MS	C57BL/6 mice	Purine, aspartate, glutamate, phenylalanine, tyrosine, tryptophan and glutathione alanine	[40]
Other contaminants					[41]
PM2.5	3 mg/kg	GC-MS	C57BL/6 mice	Energy, cholesterol, arachidonic acid, inositol phosphate and aspartic acid metabolism	[41]
	40 mg/kg	NMR	BALB/c mice	Amino acid, energy, cofactor and vitamin, lipid and carbohydrate metabolism; protein biosynthesis	[42]
	88.2 μg/m3	GC/LC-MS	C57Bl/6 J mice	Protein digestion and absorption; glycine, serine, threonine, d-Alanine and carbon metabolism; ATP-binding cassette transporters	[43]
Wastewater treatment plant effluent	–	NMR and GC-MS	Zebrafish (Danio rerio)	Energy, amino acid and lipid metabolism	[44]
	–	LC-HRMS	Mussels (Mytilus galloprovincialis)	Amino acid, neurohormone, purine and pyrimidine metabolism; citric acid cycle intermediates; oxidative stress	[45]

Abbreviations: NPs, nanoparticles; GO, graphene oxide; PM, particle matter; UPLC, ultra-performance liquid chromatography; HPLC, high performance liquid chromatography; Q, quadrupole; TOF, time of flight; HRMS, high resolution mass spectrometry; LD50, median lethal dose; SD, Sprague-Dawley; ICR, Institute of Cancer Research.

Body weight.

Identification of metabolites and pathways

The instrumental analysis is the most critical step in metabolite identification and the subsequent pathway analysis. Instruments that are amenable to analyze a large number of metabolites include mass spectrometry (MS) and nuclear magnetic resonance (NMR).

The rapidly developing MS analyzers can qualitatively or even quantitatively profile metabolites from a complex mixture with high selectivity and sensitivity [6], among which time of flight (TOF) and Orbitrap are usually applied in environmental metabolomics [7]. MS techniques can be coupled with gas chromatography (GC), liquid chromatography (LC) and capillary electrophoresis, or be used alone by direct infusion [8]. Combined chromatographic separation and MS detection of metabolites have been widely used to study environmental organisms that respond to environmental stressors. GC-MS is designed for the analysis of volatile and thermally stable compounds after derivatization, whose detection limits can reach nmol/pmol levels with highly analytical reproducibility [9]. It is quite suitable for analyzing small-molecule organic acids, amino acids, fatty acids, sugars and amines [10,11], outperforming LC-MS and NMR in low instrument cost. LC-MS has gained popularity in recent metabolomics studies because it does not require sample derivatization, and is amenable for the analysis of polar, semi-polar and non-polar metabolites. The ultrahigh-pressure liquid chromatography mass spectrometry (UPLC-MS) is increasingly utilized in high-throughput metabolomics studies, owing to the small pore size (sub 2 μm) in the column that significantly decreases the time for analysis, while maintaining high separation efficiency [12]. Currently, relevant technologies of MS are dynamically developed to improve the separation and measurement processes, making it a promising tool in the field of metabolomics, gaining more intensive attention.

NMR is a non-invasive approach that provides rich information on metabolite structures in the solution state. There is no requirement for separation and derivatization of metabolites as in the case of MS-based metabolomics, thereby showing high superiority in the analysis of sugar, amines and volatile ketones. The acquisition time (2–5 min) of a high-field NMR spectrometer is much shorter than that of GC/LC-MS, while the accuracy and precision have shown to be around 1% in a quantitative 1H NMR analysis [13]. These properties make NMR a powerful tool for toxicological assessments of environmental organisms from contaminated fields [14]. However, the disadvantages of NMR involve low sensitivity, non-selectivity and the overlapping resonances [15,16], making it not suitable for targeted metabolomics. The low sensitivity challenge may be overcome by the recent development of higher field magnets (900 MHz) and cryogenically cooled probes [17]. The problem of overlapping resonances can be largely solved by two-dimensional acquisition techniques [18]. By combining MS and NMR, it is feasible to obtain a more comprehensive set of metabolic information.

Environmental metabolomics in animals

Metabolomics shows superiority in underlying contaminant-induced cellular metabolic alternations, even when no significant differences are presented in traditional endpoints such as morphology, mortality, reproduction and weight loss [21,25]. It can also identify specific biomarkers and elucidate the action mode of environmental contaminants, mainly in targeted animals of rodents, earthworms and aquatic organisms.

Heavy metals

Although metals play a role in the biosphere by participating in many biological processes, their excess levels introduced to the environment through anthropogenic activities may adversely impact on free-living organisms. Numerous metabolomics studies have been undertaken, aiming at the responses of various organisms to a wide range of HMs, such as As [46,47], Cd [[48], [49], [50]], Cr [51], Cu [20,52], Hg [21,53], Pb [54,55], Se [56], Zn [57,58] and mixed metal (loid) pollution [[59], [60], [61]]. Many studies on the environmental metabolomics in terrestrial mammals have adopted rodents as target organisms since they are the most established mammalian models [62]. Different rodent organs, including testis [19], brains [63], lungs [64], hippocampus and cerebellum [65], have been targeted to elaborate on the specific function of metabolites, thereafter providing defence mechanisms or strategies for animals in response to HM stress.

Metabolomics results have revealed that HMs induce changes in metabolism mostly associated with oxidative stress and energy metabolism. The protective response and defence mechanism can be unveiled by key metabolites in organisms under HM stress. HMs can induce oxidative stress by raising reactive oxygen species (ROS) levels [20]. Kidneys and livers are prime targets of HMs [66], whose cells are protected from the ROS-induced damage by antioxidant enzyme defence systems. Upon Cd exposure, antioxidant activities of superoxide dismutase (SOD) and catalase (CAT) declined in both rat urine and liver tissues, implying oxidative injury to two organs [48,67]. Energy metabolism perturbation has been observed in various creatures induced by HMs [55]. In order to cope with HM toxicity, organisms tend to change energy demand, which can be satisfied through several approaches like regulating the tricarboxylic acid (TCA) cycle and glycolysis. Aquatic species exposed to As and Zn showed a similar toxic feature in energy metabolism. For examples, stimulated glycolysis and inhibited the TCA cycle were induced by As and Zn in the midge Chironomus tepperi and rockfish Sebastes schlegelii respectively [47,57]. However, Cd exposure led to depleted contents of glucose in the polychaete Perinereis aibuhitensis [49], while low levels of Hg caused an increase in ATP and intermediate levels of the TCA cycle in the earthworm Eisenia fetida [21]. A great many of the ROS, TCA and glucose-related metabolites are potential precursors to neurotoxicity and DNA damage [68], which deserve carefully monitoring in stressed organisms.

The metabolism changes, though quantitatively different between various organisms, can be indicated by a multiple set of metabolites. These metabolites can be identified as biomarkers for biological processes in response to HM toxicity. Melvin et al. [14] suggested that a metabolite suite of aspartate, myo-inositol, histidine, choline and taurine in mosquitofish may act as indicators of metal toxicity in a complex environment. In particular aspartate, a dose-response metabolite exhibited the potential to assign levels of the exposure. Xu et al. [50] picked out the panels of biomarkers in rat urine separately for Cd, chlorpyrifos and their mixture exposure. As exposure was found to cause membrane structural and functional damages in the digestive gland of the clam Scrobicularia plana [46]. As a detoxification mechanism, increased homocysteine and methionine were related to the enhanced transformation of inorganic arsenic to innocuous arsenobetaine, leading to less toxic dimethylarsenic in the methylation cycle. Additionally, 19 proteins and 2 metabolites (tyrosine and allopregnanolone), specifically related to male reproduction, were found to be significantly altered in male rats due to As treatment [19]. These findings suggest that amino acids, which are closely linked with the structure, function and activity of proteins and enzymes [69], might be sensitive indicators with unique fingerprints in mitigating HM toxicity. It is thus possible to assign significantly changed metabolites as common biomarkers in different organisms, under the stress of various HMs.

Nanomaterials

With the steadily developing nanotechnology, engineered nanomaterials have been applied in a wide range of areas [70]. The environmental concentrations of nanoparticles (NPs) were found to be in the range of ng/L (e.g. ZnO NPs of 10–500 ng/L) or even computed to be μg/L (e.g. TiO2 NPs of 0.7–16 μg/L) [71,72]. The unique features of NPs, including extraordinarily tiny size and enormous surface area, determine their behaviours, biological effects and toxicity, which may cause organ injury, neurological damage, inflammation and carcinogenesis [73].

As remarkable antibacterial agents [74], silver nanoparticle (AgNPs) can release Ag+, contributing to their ecotoxicity [75]. Li et al. [76] compared the metabolomics profiles of Daphnia magna treated with AgNPs and AgNO3 (Ag+) and proved that Ag+ takes the main responsibility for acute interference with energy metabolism and oxidative stress. Upon chronic AgNP exposure, arachidonic acid, a key fatty acid determining sex ratios, decreased in Daphnia similis [23]. A decrease in the levels of arachidonic acid was observed in the same daphnia species exposed to low-dose AgNPs, corresponding to the lower reproduction rates (Wang et al., 2018f). These observations suggest that arachidonic acid supplements might be crucial for daphnia species to enhance reproduction under AgNP stress. Besides organism-specific responses, organ-specific responses induced by AgNPs in mice were also verified by Jarak et al. [24]. The NMR metabolomics revealed that antioxidative defence, immunoregulation and detoxification were crucial in alleviating damages to mouse livers, the TCA cycle was intensified in hearts but down regulated in lungs. Metal oxides (e.g. TiO2 NPs), another popular ingredient in nanomaterial manufacture, have also become serious concerns for the potential toxicity to biota. Zhu et al. [25] confirmed the significant changes in glutathione metabolism (glycine and pyroglutamic acid) and biosynthesis of carbohydrates, proteins and lipids in earthworm E. fetida under long-term exposure to TiO2 NPs. The genes regulating ribosome biogenesis, however, were up-regulated with unaffected growth and reproduction of earthworms. These results revealed the global responses that cannot be achieved by conventional endpoints.

Carbon-based nanomaterial has attracted increased attention. However, they may also pose acute cytotoxicity and even genetic damages [77]. Therefore, risk assessment of carbon nanomaterials is imperative, and metabolomics is a useful tool to implement it. Sun et al. [26] demonstrated notable alterations in zebrafish Danio rerio metabolome in response to graphene oxide (GO) treatment, especially biosynthesis of amino/fatty acids and the TCA cycle. Metabolomics has also been applied to investigate the hormesis effects caused by fullerenes at molecular levels, and the observed metabolic changes indicate oxidative stress, hypoxia and energy metabolism disturbance [78,79]. The toxicity of nanopolystyrene [80] and laser printer-emitted NPs [81], originated from other modern techniques, on organisms are also a concerning issue. This will inherently require experimental design and instrumental analysis of greater complexity in environmental metabolomics studies.

Pesticides

Pesticides, including insecticides, herbicides, fungicides, rodenticides, nematicides and other substances [82]. Aim to prevent, destroy, repel or mitigate pests [83]. They possess features like biotoxicity, persistency, enormous input into the environment and biomagnification [[84], [85], [86]], which are closely related to various ecological and health problems.

Aquatic organisms are directly affected by pesticides, mostly through runoff from croplands. Various fish species [[87], [88], [89], [90]], crustaceans [[91], [92], [93], [94]] and amphibians [[95], [96], [97]] have been commonly regarded as targets of aquatic risk assessments. Zebrafish are popular in toxicological experiments because of their physiological and genetic similarity to mammals [[98], [99], [100]]. Specifically, embryonic or larval fish are extremely vulnerable to exotic stressors [101], thus suffering more from the developmental toxicity of pesticides. Developmental malformations (locomotor defect and notochord deformation) were observed in embryonic and larval zebrafish Danio rerio exposed to fipronil and flutolanil, which could be ascribed to the changed pathway of aminoacyl-tRNA biosynthesis [102,103]. Unfortunately, further damages to DNA and RNA were induced by fipronil and flutolanil in embryo/larval zebrafish, accompanied by the disturbance in nucleotide (purine and pyridine) metabolism [102,104]. This is likely to alter genetic expression and transcription, consequently increasing the pesticide resistance in organisms. Therefore, environmental metabolomics would be a promising tool to link metabolic phenotypes with potential biomarkers and dominant genes.

Soils are huge reservoirs retaining a variety of pesticides. Earthworms play an indispensable role in soil ecosystems by participating in soil formation and structure maintenance, aeration, drainage and fertility [105]. Monitoring earthworm metabolome shall indicate the health condition of soils and evaluate the ecotoxicity of pesticides [106]. Various earthworm species have been used in soil exposure experiments, and among them, E. fetida is the most popular, recommended by Organization for Economic Cooperation and Development. The majority of metabolomics studies targeted at whole-worm extracts [28,107,108]. The coelomic fluid, however, is the center of many bioactivities. Its extraction is less invasive and simpler, and bears less loss [109]. In fact, in comparison to whole worms and coelomocytes, coelomic fluid was detected as the most sensitive matrix in chlorothalonil-exposed earthworms [110]. A novel metabolite, malylglutamate, was identified and postulated to serve roles in whole earthworm homeostasis, due to the simpler background produced by coelomic fluid components [111]. Additionally, coelomic fluid metabolomics could phenotypically distinguish between earthworm species by species-specific aromatic metabolites [112]. The expanded biological materials, including coelomic fluid, coelomocytes and tissues, will contribute significantly to the systematic assessments of pesticide toxicity in soil ecosystems.

Mouse blood, urine and organs are also gaining popularity in metabolomics experiments [[113], [114], [115], [116]]. Bonvallot et al. [29] put forward a theory about the action mode of a mixture of eight pesticides on pregnant rats and their offspring with NMR-based metabolomics upon plasma, liver and brain. The results suggested that maternal exposure to the pesticide mixture caused liver dysfunction, altered fatty acid component and lipid content, while brain samples revealed the affected energy metabolism in mother rats and male offspring. Considering that >30% of pesticides are chiral [117] such as metalaxyl, penconazole and profenofos, many studies employed mice to prove and investigate enantioselective metabolic effects and differential ecotoxicity of chiral pesticides [[118], [119], [120]]. A broader scope of environmental metabolomics can be achieved by conducting experiments, involving more organismal targets under pesticide stress.

Pharmaceuticals and personal care products

PPCPs are widely used in a vast amount worldwide [121,122]. They are continually discharged into domestic/industrial sewage systems [123], and their removal efficiency in conventional wastewater treatment plants (WWTPs) is limited [124]. As emerging contaminants, PPCPs have aroused great concern due to their ubiquity, pseudo-persistence and bioactivity in the environment [125,126], as well as their unexpected adverse impacts upon the ecosystem [127].

Aquatic organisms are primary concerns since they are under continual exposure to PPCPs and encountered the most exposure risks [[128], [129], [130], [131], [132], [133], [134], [135]]. The metabolic effects of PPCPs on edible organisms deserve the main focus of concern. Antibiotics earn special attention due to their omnipresent existence in the environment, corresponding induction of antibiotic-resistance genes and potential impacts on human beings [136]. Metabolomics studies have noted significant changes in amino acid metabolism of fish and crustaceans. Sulfamethazine, an antibiotic extensively used to control bacterial infections, was proved to disturb amino acid profiles in marine medaka Oryzias melastigma [137]. Hepatic damages, immune function defect, oxidative stress and consequent defences, energy shift, osmoregulation and nucleotide metabolism were found correlated to changes in amino acid levels under the stress of sulfamethoxazole in marine mussel Mytilus galloprovincialis [32]. As an adverse consequence of disturbed amino acid metabolism, the nutritional value of fish and mussels, for human consumption would be affected by antibiotics, possibly manifested by the varied quality of dietary proteins. Antibiotics may provoke antioxidant stress as well, leading to accelerated degradation of proteins in blue mussel Mytilus edulis [138]. From a perspective on human health, it is of great interest to further evaluate the effects of PPCPs on the biosynthesis of vitamins, lipids and other valuable metabolites due to dietary exposure from animal origins.

Endocrine-disrupting chemicals (EDCs), including hormones [31], plastics and plasticizers (e.g. bisphenols and phthalates) [[139], [140], [141], [142]], can disrupt hormonal systems and homeostasis in organisms. Owing to the extensive use, severe pollution and endocrine-disrupting effects, hormones are progressively gaining attention these years, including steroid estrogens (e.g. natural estrone, 17 beta-estradiol and synthetic 17 alpha-ethynylestradiol), progestagens, androgens and glucocorticoids [143]. Studies provide evidence that these EDCs affect gonad, thyroid, kidney, neuroendocrine and cardiovascular systems in the laboratory in vitro/vivo models [144], whose correlation with metabolite changes have been investigated by metabolomics. For instance, 17-alpha-ethynylestradiol-induced disturbed lipid metabolism, destructing the homeostasis of steroid hormones and causing energy deprivation in germ cells [145]. These were responsible for the adverse reproduction in crucian carp Carassius auratus [145]. In contrast to energy deprivation, oysters might utilize their lipid reserves to satisfied increased energy demands caused by 17 alpha-ethynylestradiol [31], and proteolysis was stimulated in medaka to produce more energy to metabolize and eliminate 2,4-dichlorophenol [146]. EDCs can also perturb reproduction through interference with gamete generation and development, by either inhibiting or promoting spermatogenesis [147]. A transgenerational decline in spermatogenesis was reported, accompanying with the reduced activity of betaine homocysteine S-methyltransferase in the testes of rats exposed to di (n-butyl) phthalate [148]. Moreover, bisphenol A, S, F and AF can trigger other impacts including obesity and DNA/histone methylation [30,149,150]. Despite considerable efforts to investigate metabolic diseases, the existing mechanisms remain unclear in organisms exposed to newly emerging EDCs, and metabolomics may offer a scientific basis to solve the uncertainty.

Persistent organic pollutants

POPs are toxic substances that can be transported globally, persist in the environment and magnify in the food web [151,152]. Their pollution is a global issue, and joint strategies have been developed to control their emission and environmental impacts [153].

Polycyclic aromatic hydrocarbons (PAHs) have been confirmed to be carcinogenic and mutagenic [154]. Many PAHs are inhaled by organisms through contaminated air, thereafter threatening the respiratory system. The damage in airway cellular membrane underlined a possible mechanism [155,156]. To gain more insights into the PAH toxicity, numerous studies have focused on alterations in the metabolome of organismal tissues, including brains, lungs, liver, gills and glands. Disrupted energy metabolism, osmotic regulation and amino acid metabolism were identified as predominant metabolic dysfunction under PAH stress in different animals [[33], [34], [35]]. Benzo [a]pyrene is a typical PAH that targets the developing brain to cause nervous system impairment. The neurotoxicity of benzo [a]pyrene was demonstrated through metabolic changes correlated with cerebellum injury in rats [36]. The exposure to benzo [a]pyrene aggravated energy consumption, while the abnormal glycolysis and interrupted creatine/phosphocreatine kinase system indicated insufficient energy supply. This will cause ROS production and the methionine cycle disorder, leading to further responsible for neurobehavioral impairment. Consequently, amino acid metabolism was accelerated to compensate for energy shortage, particularly valine and leucine metabolism. Similarly, abnormal metabolism of glutathione, tyrosine and glycolysis induced by benzo [a]pyrene was noted in rat hippocampus, which ultimately affected the antioxidant system, neurotransmitter synthesis and energy supply, respectively [37]. These should account for the practical learning and memory deficits of adult rats. Therefore, untargeted metabolomics responds effectively to a request for information on the PAH-specific toxicity on developing organismal systems.

Brominated flame retardants (BFRs) are the most popular group of flame retardants [157,158]. Among the more than 75 BFRs available in the market [159], hexabromocyclododecane is a highly bioaccumulated group, while polybrominated diphenyl ethers and tetrabromobisphenol A are produced in large volumes [160]. BFRs possess neurotoxicity, immunotoxicity and carcinogenicity [40,[161], [162], [163]], which can exert toxic effects on endocrine systems, especially thyroid hormones, reproduction and steroid homeostasis [159,[164], [165], [166]]. Obesity is one of the negative results of BFR exposure. It was demonstrated that perigestational exposure of 2,2′,4,4′-tetrabromodiphenyl ether caused body weight gain in male rat offsprings and dams at low doses, and the maternal serum metabolome change might be a possible contributing factor [38]. These gender-specific effects were attributed to sex steroids and distinct modification of thyroid hormone production. The differential serum metabolites implicated that amino acid metabolism was the most disturbed, among which branched-chain amino acids declined, and hypothetically, they were transferred to offspring, contributing to the risk of obesity. Some studies reported that high-fat diet stimulated obesity and induced more pronounced metabolic changes in male mouse offspring treated with BFRs [39,167]. The possible linkage between POP exposure and other diseases can also be verified in model organisms with metabolic evidence.

Other categories of prevalent POPs, e.g. short-chain chlorinated paraffin, have been proved to significantly affect organismal metabolism in rat livers/kidneys [168,169] and zebrafish embryos/larvae [170]. Dioxins and many halogenated compounds were demonstrated to induce systemic metabolic dysfunctions in zebrafish [171], daphnia [172], nematodes [173], earthworms [174] and rats/mice [[175], [176], [177]].

Other contaminants

In addition to the above categories of contaminants, multitudes of other pollutants generating from anthropogenic activities deserve in-depth exploration of their ecotoxicity. These include flame-retardant phosphate [178,179], microplastics [180], microcystin [181], inorganic chlorine, sulfur and nitrogen [[182], [183], [184]], physical stressors [[185], [186], [187]] and other contaminant mixture.

Since polybrominated diphenyl ethers and polychlorinated biphenyls were listed in the ban list of Stockholm Convention of POPs, organophosphorus flame retardants are used as substitutes [188]. Triphenyl phosphate is one of the most broadly utilized and frequently detected organophosphate flame retardants in the environment [189,190]. Although not as persistent as some BFRs, sex-specific and dose-dependent metabolic disturbances of triphenyl phosphate were reported by neonatal exposure in mice [191], obesity and metabolic syndrome were further investigated at a fetal dose [188]. In our previous studies, degradation metabolites of triphenyl phosphate were studied in earthworm Perionyx excavates, and novel metabolites were putatively identified and categorized as thiol conjugates and glucoside conjugates [192,193]. These novel biotransformation products revealed the metabolism of anthropogenic contaminants, as a key detoxification mechanism in exposed organisms.

Actually, a great many contaminants co-exist in the environment, and living organisms are continually facing a combination of various contaminants under different uncontrolled, varying and irregular conditions. These contaminants interact with each other, and exert synergistic or antagonistic effects on living environments. In a coastal lagoon received large amounts of agricultural and urban pollutants, metabolites of amino acids, osmotic protectants and nucleotides increased after 7-d exposure, and drastically diminished in the digestive glands of clam Ruditapes decussatus at longer exposure duration [194]. Conventional WWTPs are not designed to remove HMs, NPs, pesticides, PPCPs and POPs. The effluent usually contains numerous newly emerging pollutants when discharging into the aquatic environment, though at relatively low levels [195]. Several metabolomics studies regarded wastewater effluent as the pollutant source and elucidated the biological effects of such exposure on either freshwater or marine organisms [196]. It was found that being chronically exposed to low concentrations of 38 pharmaceutically active compounds and 4 pesticides from treated wastewater, metabolite levels were impacted in amino acid, purine, pyrimidine and neurohormone pathways in mussel M. galloprovincialis [45]. Whilst major impacts were exerted on energy, amino acid and lipid metabolism of zebrafish D. rerio in surface water receiving complex pollutant mixtures from WWTPs [44]. Fine particulate matter (e.g. PM2.5), a major air pollutant, is commonly composed of HMs, PAHs, inorganic minerals and even pathogens [197]. It can adversely affect respiratory and cardiovascular systems [198]. Nevertheless, studies on its interference with metabolism pathways are scarce. PM2.5 exposure affected hormone metabolites and circadian rhythm biomarkers like melatonin, retinal and 5-methoxytryptophol [43], or induced neural damages [41]. Distinctly, metabolic changes induced by water-soluble components of PM2.5 could be differentiated from those by water-insoluble components [42]. Overall, it is possible to characterize the specific or systemic biological effects of co-existing contaminants with variability and complexity, owing to the global metabolite profiles of stressed organisms. It was thus concluded that environmental metabolomics could be efficient in evaluating the comprehensive responses of organisms to various contaminants.

Environmental metabolomics in microorganisms

Microorganisms, omnipresent in environment, are becoming popular indicators for cytotoxicity of contaminants and cellular responses to them. Microbial metabolomics is increasingly applied to unmask a series of factors that affect ecosystem health induced by toxic compounds including HMs [199], NPs [[200], [201], [202], [203]], PAHs [204], pesticides [84], ammonia, phosphate and ocean acidification [115,184,205] (Table 2).

Table 2

Environmental metabolomics studies based on microorganisms in 2015–2020.

Stressor	Concentration	Detection method	Organism	Disturbed metabolite/Pathway	Reference
Heavy metals
Cu	1, 3 and 6 mM	LC-MS	Yeast (Saccharomyces cerevisiae)	Oxidative stress and DNA damage	[199]
	0, 0.3, 6 and 10 μg/L	GC-TOF-MS	Diatom Tabellaria flocculosa (Roth) Kützing	Hydroxylamine, unsaturated fatty acids, saturated fatty acids, 2-palmitoylglycerol, glycerol and diterpenoid compounds	[206]
Zn	30, 500 and 1000 μg/L	GC-TOF-MS	Diatom Tabellaria flocculosa (Roth) Kützing	Fatty acids, amino acids, terpenoids, glycerol, phosphate, sucrose and lumichrome; antioxidant systems and extracellular ion chelation (exopolysaccharides, frustulins)	[207]
Nanomaterials
AgNPs	0, 1, 10 and 100 μg/L	GC-Q-TOF-MS	Microalgae (Scenedesmus obliquus)	d-galactose, sucrose, d-fructose and glycine	[201]
	100 mg/kg soil	GC-MS	Soil microorganisms	Sugar, amino acid and fatty acid metabolism	[208]
CeO2 NPs	10 mg/L	FT-ICR	Freshwater alga (Chlamydomonas reinhardtii)	Carbon fixation and photosynthesis associated with energy metabolism	[200]
MoS2 nanosheets	100 and 1000 μg/mL	GC-MS	Escherichia coli	Glycine, serine, threonine and pyruvate metabolism; urea cycle; protein biosynthesis	[209]
TiO2 and ZnO NPs	2.5–10 and 0.025–0.2 mg/L respectively	GC-TOF-MS	Escherichia coli	Arginine, proline, glycine, serine and threonine metabolism	[210]
GO and CNTs	0.01–10 mg/L	GC-MS	Microalgae (Chlorella vulgaris)	Carbohydrate, amino acids, fatty acids and urea	[211]
Pesticides
2,4-D	10 mM	GC-MS	Escherichia coli	Oxidative phosphorylation; ABC transport; peptidoglycan biosynthesis; amino acid, nucleotide and sugar metabolism	[212]
	1 ppm	LC-Q-TOF-MS)	Mice gut microorganisms	Urea degradation, amino acid and carbohydrate metabolism	[114]
Persistent organic pollutants
Phenanthrene, pyrene and benzo(a)pyrene	47.36 ± 0.34, 50.13 ± 1.42 and 9.66 ± 0.77 mg/kg respectively	GC-TOF-MS	Soil microorganisms	Fatty acids, carbohydrates and amino acids	[204]
Other contaminants
Ocean acidification	–	UPLC-HRMS	Brown macroalgae (Lobophora rosacea)	Lobophorenols B and C other oxylipin derivatives	[205]
Triphenyl phosphate	0.01–10 mg/L	GC/LC-QTOF-MS	Microalgae (Chlorella vulgaris and Scenedesmus obliquus)	Respiration; osmoregulation; membrane lipid synthesis and lipolysis	[115]

Abbreviations: CNTs, carbon nanotubes; 2,4-D, 2,4-dichlorophenoxyacetic acid; FT-ICR, Fourier-transform ion cyclotron resonance.

Model microorganisms, e.g. Escherichia coli, have been well characterized for elucidating xenobiotics' cytotoxicity, especially through ‘omics’ technologies [209]. Most environmental contaminants impose oxidative stress, which is mainly confirmed by ROS generation, and trigger the defensive system [213]. Having found that both TiO2 and ZnO NPs elevated ROS in E. coli under solar irradiation [214,215], Pathakoti et al. [210] further figured out the mechanisms of their cellular toxicity by GC-TOF-MS-based metabolomics. Noticeable fluctuations were noted in glycine, serine, threonine, arginine and proline metabolism. Combining NMR-based metabolomics and MS-based proteomics, up-regulated metabolites (putrescine and valine) and proteins (succinate dehydrogenase and aconitate hydratase) correlated to oxidative stress protection were recognized in E. coli [216]. In another study, Bhat et al. [212] posed E. coli to sublethal levels of 2,4-dichlorophenoxyacetic acid (2,4-D) herbicide and coupled its morphological changes to intracellular metabolic alterations. 2,4-D induced oxidative stress in E. coli, which was manifested by the depletion of glutathione metabolites to protect cells from antioxidant species. As a modulator of cellular redox state, glutathione is the second most abundant metabolite, acting as the first line of defence against oxidants [217]. The lack of it is to blame for inhibited cell division and altered peptidoglycan biosynthesis [218], representing as remodeled cell envelop, membrane defects and shifts in surface roughness/charge [212]. Therefore, model species provide shreds of evidence with metabolomics to develop global understanding towards biological responses in microorganisms under environmental exposure. As shown in Fig. 2, there may be a common mechanism for ecotoxicity of most contaminants to different organisms, e.g. a dose-dependent effect of HMs, NPs, pesticides, PPCPs, POPs and other contaminants on either animals or microorganisms, sharing similar changes in amino acid species, with different phenotype alternations.

Fig. 2

A common mechanism for ecotoxicity of different anthropogenic contaminants to animals and microorganisms: Amino acid metabolism as an example.

Gut microbiota is a cluster of microorganisms in the intestine, capable of affecting host metabolism and is associated with several diseases [219]. Gut microbiota can be easily affected by external stressors, presenting as alterations of gut microbiota composition as well as the metabolome, and hence further perturb host metabolic pathways via intimately interacting with its host [220]. Mice gut microbiome is increasingly used as an indicator in environmental metabolomics. Perturbed by chlorpyrifos, clear correlations between certain gut bacteria and altered urine metabolites in mice were manifested, for instance, Bifidobacteriaceae was positively correlated with glycerol, phenylacetylglycine and lithocholate [116]. It was found that occupationally relevant doses of 2,4-D profoundly altered gut-associated metabolic pathways including urea degradation, amino acid and carbohydrate metabolism in mice by fecal metabolomics, in concert with observed modifications in mice gut microbiota composition and diversity [114]. Moreover, oral exposure to TiO2 NPs altered rat gut microbiota and their associated metabolism. Hence the metabolic disorders of gut microbiota, and the subsequent lipopolysaccharides production, led to the inflammatory response and oxidative stress in the intestine [221]. These results inferred that the interactions between contaminants and gut microbiota might have been greatly underestimated. The dysbiosis of gut microbiome is, of necessity, to be studied in association with nutritional balance and related diseases in the host intestine. More bioinformatics data on regulatory metabolites and pathways may provide clues to better maintain the host health.

As the ecosystem is a huge tank containing many contaminants, there is no doubt that these contaminants interfere with different microbial activities. In the soil ecosystem, engineered nanomaterials have been applied in environmental remediation [70]. AgNPs, however, have been found to affect soil microbes of different functions and therefore modify carbon and nitrogen cycles in soil matrix [208]. Oxidative stress induced by AgNPs impacted on microbial cell membranes and thus led to altered profiles of phospholipid fatty acids. In the presence of TiO2 and CuO NPs, phospholipid fatty acids can serve as signs indicating soil microbial community shifts [222]. The evidence supports the conclusion that changes of microbial communities are primarily responsible for metabolome alterations, rendering some sensitive metabolites in certain microbial taxa, e.g. iso-branched fatty acids in Gram-positive bacteria and cyclopropane fatty acids in Gram-negative bacteria as an adaptation response to carbon nanotubes (CNTs) [223]. More importantly, the soil is the cradle of tremendous terrestrial plants and the input of the food chain, which provides space for interactions between plants and soil microbes, which are both exposed to various contaminants [25,203]. As pesticides pass down the food chain through biomagnification, they would finally reach human bodies and cause disorders. In the aquatic ecosystem, a marked increase in antioxidant activity of SOD and CAT were observed to restrain oxidative stress induced by Cu and Zn, which was fueled by higher energy production in freshwater diatom Tabellaria flocculosa [206,207]. The cell division of the microalga Chlorella vulgaris was suppressed after 96-h exposure to GO or CNTs in a dose-dependent manner [211]. GC-MS-based metabolomics proposed the alternation involving carbohydrate, fatty acid and amino acid metabolisms as well as urea cycle as nano-toxicological mechanisms (Fig. 3), demonstrating the threats posing by carbon-based nanomaterials.

Fig. 3

Representative effects of environmental contaminants on the metabolism of microorganisms: Main metabolic pathways of microalgae (Chlorella vulgaris) disturbed by nanomaterials (GO and CNTs). The blue/red arrows represent the up/downregulation of metabolites, and the black solid/dotted arrows indicate direct/indirect reactions in the presence of GO/CNTs, respectively. Reprinted with permission from Hu et al. [211]. Copyright 2015 American Chemical Society.

The microorganism is one of the most basic and crucial components of ecosystems. The existence of various contaminants in the ecosystem would disturb the functions and activities of microbes as well as alter microbial composition in a microcosm. Changes that happen in microcosm can be therefore magnified and consequently affect macrocosm, altering the original ecological state. For this reason, ecological and toxicological assessments of anthropogenic contaminants based on microorganisms merit more careful consideration.

Conclusions

In this review, we emphasize environmental metabolomics as a promising ecotechnique to evaluate the impacts of contaminant-related stressors introduced by anthropogenic activities. The contributions of metabolomics to the environment can be concluded: Firstly, assessing the ecotoxicity of contaminants, especially at environmentally relevant concentrations, on different organisms on a molecular level; secondly, investigating the in-depth mechanisms and action modes of contaminants through detection of metabolite alternations and perturbation of metabolism pathways; thirdly, identifying representative differentiated metabolites as specific biomarkers, which will be promising tools in environmental monitoring; ultimately, providing timely and effective information to protect environmental health from noxious contaminants better, and offering fundamental knowledge for the sake of the establishment of precautions and remedies. Similar amino acid metabolisms are proposed for animals and microorganisms in response to ecotoxicity of different anthropogenic contaminants. With these significant contributions, there will be vaster applications of environmental metabolomics for exploration.

Author contributions

Hu-Chun Tao conceived the project. Li-Juan Zhang and Lu Qian drafted the manuscript. Ling-Yun Ding, Lei Wang, Ming Hung Wong and Hu-Chun Tao edited the manuscript.

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.