Bioactivity Profiling of Chemical Mixtures for Hazard Characterization

Full text

Bioactivity Profiling of Chemical Mixtures for Hazard
Characterization
Xiaojing Li,*
,◆
Jiarui Zhou,
◆
Yaohui Bai, Meng Qiao, Wei Xiong, Tobias Schulze, Martin Krauss,
Timothy D. Williams, Ben Brown, Luisa Orsini, Liang-Hong Guo, and John K. Colbourne
Cite This: https://doi.org/10.1021/acs.est.4c11095
Read Online
ACCESS Metrics & More Article Recommendations *
sı Supporting Information
ABSTRACT: The assessment and regulation of chemical toxicity
to protect human health and the environment are done one
chemical at a time and seldom at environmentally relevant
concentrations. However, chemicals are found in the environment
as mixtures, and their toxicity is largely unknown. Understanding
the hazard posed by chemicals within the mixture is critical to
enforce protective measures. Here, we demonstrate the application
of bioactivity profiling of environmental water samples using the
sentinel and ecotoxicology model species Daphnia to reveal the
biomolecular response induced by exposure to real-world mixtures.
We exposed a Daphnia strain to 30 sampled waters of the Chaobai
River and measured the gene expression response profiles. Using a
multiblock correlation analysis, we establish correlations between
chemical mixtures identified in 30 water samples with gene expression patterns induced by these chemical mixtures. We identified 80
metabolic pathways putatively activated by mixtures of inorganic ions, heavy metals, polycyclic aromatic hydrocarbons, industrial
chemicals, and a set of biocides, pesticides, and pharmacologically active substances. Our data-driven approach discovered both
known bioactivity signatures with previously described modes of action and new pathways linked to undiscovered potential hazards.
This study demonstrates the feasibility of reducing the complexity of real-world mixture toxicity to characterize the biomolecular
eects of a defined number of chemical components based on gene expression monitoring of the sentinel species Daphnia.
KEYWORDS: freshwater, chemical mixtures, transcriptomics, biomolecular eect data, new approach methodologies (NAMs)
■INTRODUCTION
Chemical pollution is a global threat to public health
1
due to
unregulated mixtures from domestic, agricultural and industrial
processes entering the environment. Freshwater ecosystems,
especially rivers, are particularly impacted, as they receive
untreated and treated wastewater from domestic and industrial
euents.
2
This is because conventional wastewater treatment
processes are not designed to eectively remove industrial,
agricultural and domestic pollutants from wastewater.
3
These
chemicals enter the food chain and water supply through
irrigation and aquifer recharges, adversely aecting environ-
mental and human health.
4
Current regulatory frameworks assess chemical safety one
substance at a time, neglecting the cumulative toxicity of
chemical mixtures.
5,6
The total number of chemicals tested by
2022 was 12,714,
7
corresponding to less than 0.2% of the
chemicals potentially present in the environment. The low
percentage of tested chemicals can be explained by the low
throughput of traditional risk assessments, which require animal
testing.
New approach methodologies (NAMs) in regulatory
toxicology include bioactivity measurements (e.g., genes and
metabolites) of changes induced by exposure to chemicals. They
are regarded as high-throughput alternatives to traditional
methods, are suitable for detecting sublethal eects of chemical
and chemical mixtures
8,9
and enable the grouping of chemicals
based on their bioactivity.
10
This helps define points of
departure to adversity
11
and classify substances by their modes
of action (MoAs).
12
However, despite their advantages,
regulatory frameworks have been reluctant to adopt NAMs
due to perceived scientific, technical, regulatory, and economic
challenges.
13
Validating the robustness and advantages of NAMs
in real-world cases can help build confidence in their use and
create harmonized guidelines for their application.
We recently published a conceptual framework that uses gene
expression monitoring to identify correlations between ambient
chemical mixtures and biomolecular responses in the sentinel
Received: October 15, 2024
Revised: December 4, 2024
Accepted: December 5, 2024
Articlepubs.acs.org/est
© XXXX The Authors. Published by
American Chemical Society A
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
This article is licensed under CC-BY 4.0
Downloaded via HELMHOLTZ-ZENTRUM UMWELTFORSCHUNG on January 6, 2025 at 15:57:32 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

species Daphnia.
14
This framework enables Daphnia to be used
as an early warning system of the chemical hazards by profiling
gene expression responses to chemical mixtures and identifying
hazardous chemicals within mixtures. Here, we demonstrate the
use of this framework in a case study by exposing Daphnia magna
to surface river waters from the Chaobai River system in China.
These water samples are chemically complex as the river system
receives chemical pollutants from agricultural runos, domestic
reclaimed waters, and industrial wastewater euents.
15,16
We
characterized the chemical fingerprints of 30 samples collected
along the river, using target and target screening methods. We
exposed a commercial strain ofD. magnato these water samples,
conducting both immobilization assays and genome-wide
transcriptome profiling. We combined gene coexpression
network analysis with multiblock correlation analysis to link
chemicals within the environmental mixtures and the bioactivity
profiles inD. magna, revealing networks of genes dysregulated by
chemicals. By applying pathway overrepresentation analysis, we
identified enriched functional pathways associated with those
chemicals. We benchmarked our findings by identifying known
associations between metabolic pathways and chemicals from
published studies and discovered new toxicity pathways. The
identification of known pathways through this data-driven
approach provides confidence in the newly discovered
associations between chemicals and novel coresponsive path-
ways. This data-driven approach for integrating analytical
chemical analysis and bioactivity profiling enhances our ability
to unravel the hazards of environmental chemical mixtures,
paving the way to more informed and eective environmental
risk assessment.
■MATERIALS AND METHODS
Field Sampling Site and Strategy. The Chaobai River
system, the second longest river in the Beijing area, spans 458
km and covers a drainage area of 13,846 km2
17
across the Hebei-
Beijing-Tianjin region. It is formed by three rivers, namely the
Bai, Chao and Chaobai Rivers. The Bai River and the Chao River
both originate from Yunwu Mountain in Hebei province. They
converge to form the Chaobai River, which flows through
populated towns in northern Beijing and agricultural lands in
Tianjin before reaching the Pacific Ocean. This river system is
impacted by industrial euents from various sectors, including
food, cosmetics, pharmaceuticals, and automotive manufactur-
ing.
18
The downstream (Chaobai River) receives reclaimed
water from the Wenyu River, contributing up to 38 billion cubic
meters to its annual flow.
19
Additionally, wastewater treatment
plants along this river release 9.2 billion cubic meters of treated
euents into the water flow yearly.
20
For this study, we collected water from seven locations along
the Bai River, six locations along the Chao River, and 17 from the
Chaobai River (Figure 1). The pH and total dissolved solids
were measured in situ using a multiparameter water quality
probe (MYRON Co.) (Table S1). From each location, ∼8L of
surface water was collected in Duran amber glass bottles and
transported to the laboratory at room temperature (19−23 °C).
Of the total collected water per sample, 500 mL was used for
immobilization assays with D. magna, and the remaining volume
was used for chemical fingerprinting.
Chemical Fingerprinting of the Chaobai Water
Samples. The chemical fingerprints of the Chaobai water
samples were generated by target and target screening analyses
(Table S2). The target analysis measured inorganic elements
and polycyclic aromatic hydrocarbons (PAHs), while the target
screening analysis quantified polar organic compounds.
Altogether, we measured 609 chemicals, of which 385 were
detected.
Target Chemical Analysis. The target chemical analysis was
used to quantify 19 inorganics (e.g., P, N and heavy metals) and
Figure 1. Workflow of the Chaobai case study. A total of 30 water samples were collected from three rivers: the Chao River (6 sites), the Bai River (7
sites), and the Chaobai River (13 sites). These samples were used for chemical fingerprinting and exposure assays with a commercial strain of Daphnia
magna. Exposed Daphnia that were not immobilized after exposure were used for transcriptome profiling (RNA-Seq). Genome-wide gene expression
was used to generate coexpression gene modules. Multiblock correlation analysis was then used to identify correlations between chemicals within
mixtures and pathways enriched within coexpression modules.
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
B

16 PAHs (Table S2). For this analysis, water samples were
filtered through 0.45 μm glass fiber membrane (GF/F,
Whatman) and stored at 4 °C before analysis. Total nitrogen
(TN), ammonia (NH4+), nitrate (NO3−) and nitrite (NO2−)
were measured with UV spectrophotometry after alkaline
potassium persulfate digestion, while total phosphorus (TP)
and phosphate (PO43−) were quantified using ammonium
molybdate spectrophotometry, following methods described in
Liao et al.
21
Heavy metal concentrations (Cd, Cr, Cu, Ni, Zn, Pb,
and As) were determined using inductively coupled plasma-
mass spectrometry (ICP-MS, 7500a, Plasma Quad 3), while
concentrations of K, Ca, Na, and Mg were assessed via
inductively coupled plasma optical emission spectrometry
(ICP-OES, OPTIMA 2000, PerkinElmer), following methods
described in Xiong et al.
17
Sixteen PAHs were quantified using
an Agilent 7890A gas chromatographer coupled with a 5795C
mass spectrometry detector with electrospray ionization sources
in the selective ion monitoring mode following methods
described in Qiao et al.,
19
with detailed methods outlined in
Supporting Information Section A.
Target Screening Analysis of Polar Organic Substances.
The target screening analysis quantified 574 organic chemicals
(Table S2), including pharmaceuticals, pesticides, biocides,
industrial solvents, etc. For this analysis, two liters of surface
water were filtered through a 0.7 μm glass microfiber membrane
(GF/F, Whatman) and extracted using solid-phase extraction
with HLB cartridges (500 mg, 6 mL, Waters), which were
preconditioned with methanol and deionized water. Each
sample underwent positive and negative ion modes analysis by
liquid chromatography-high resolution mass spectrometry (LC-
HRMS), as also described in Section A. The conversion of raw
mass spectral data (mzML) and centroiding was accomplished
using ProteoWizard, incorporating the instrument’s built-in
library.
22
Data processing, including peak picking, alignment,
gap filling, and peak annotation, was performed using MZmine
2.52 (http://mzmine.github.io). The annotated peak list
underwent further analysis with an in-house R-package,
MZquant (version 0.7.22), for semiautomated quantification.
A final cleanup of the annotated peaks, blank filtering, automatic
internal standard assignment, and peak quantification was
completed using MZquant (version 0.7.22). Compounds with
broad peaks or high background noise, which could not be
quantified through the semiautomatic workflow, were manually
annotated with TraceFinder 4.1 (Thermo Scientific). The raw
peak tables were refined based on final peak annotations
following assessments by MZquant and TraceFinder.
Data Preprocessing and Analysis. Chemicals detected in
target and target screening assays that were below detection
limits or absent in more than 50% of the samples were excluded
from downstream analysis. This was done to meet the minimum
requirements for missing value imputation. The k-nearest
neighbor algorithm (k= 5)
23
was used for imputing missing
values of 215 chemicals retained after the filtering process. The
principal component analysis (PCA) plot was used to reveal the
similarity of chemical fingerprints based on the standardized and
normalized concentration values of 215 selected chemicals.
Bioactivity Profiling of Daphnia Exposed to the
Chaobai Water Samples. We performed immobilization
assays and genome-wide transcriptome profiling to quantify the
Daphnia magna’s biological and biomolecular responses to
chemical mixtures from the Chaobai water samples.
Immobilization Assays. Surface waters collected from 30
sites were filtered using the 0.45 μm glass fiber membrane (GF/
F, Whatman) to remove suspended particles and bacteria. The
filtered water samples were then frozen at −20 °C prior to
exposure to reduce any bacterial or viral bioactivities. The
filtered water samples were acclimated at 20 °C overnight prior
to OECD immobilization assays, where immobilization is
recorded as an apical end point.
24
Three biological replicates
of the D. magna commercial strain IRCHA 5 (Water Research
Centre, Medmenham, U.K.), each containing five 24 h-old
daphnids, were used per sample. Daphnia were exposed to river
waters filtered by a 0.7 μm glass fiber membrane (GF/F,
Whatman). After 48 h of exposure, the number of immobilized
daphnids was counted, whereas the mobile daphnids were
collected, flash-frozen in liquid nitrogen and stored at -80 °C
until transcriptome profiling.
Transcriptome Profiling. Flash-frozen daphnids from each
exposure were homogenized for total RNA extraction using the
Agencourt RNAdvance Tissue Total RNA kit (Beckman
Coulter), following the manufacturer’s instructions. The
concentration and the quality (integrity and purity) of the
extracted RNA were quantified with the Nanodrop 8000
Spectrophotometer (Labtech Ltd., U.K.) and the TapeStation
2200 (Agilent Technologies), respectively. The mRNA was
prepared into cDNA libraries using the NEBNext Ultra
Directional RNA Library Prep Kit (New England Biolab
E7420L) with NEBnext Multiplex Oligos for Illumina Dual
Index Primers (New England Biolabs E7600S) RNA extraction
and library preparation were performed with the Biomek FxP
workstation (Beckman Coulter A31842). The quality of the
constructed libraries was assessed using the Tapestation 2200.
Libraries were pooled at a final molarity of 3 nM and 100-bp
paired-ended sequenced on a HiSeq4000 (Illumina) at the
Beijing Genomics Institute (BGI), aiming for 5 million reads per
sample.
Sequence Preprocessing. RNA sequences were trimmed
with Trimmomatic (version 0.32)
25
to remove adapter
sequences and retain reads with an average Phred score ≥30
with FastQC (version 0.11.9). High-quality reads were mapped
on the D. magna reference genome
26
using STAR (version
2.7.10a), and the expression levels of the mapped genes were
quantified by RSEM (version 1.3.2) using default settings.
27
Gene counts were preprocessed in R (version 4.0.3) with a
customized script to exclude genes with average read counts <5,
total counts <10, or zero counts in more than 60% of the
samples. The count table of filtered genes was normalized using
the conditional quantile normalization approach provided by
the “cqn” package (version 6.0.0).
28
The normalized gene
counts were then used to perform similarity analysis with PCA,
construct coexpression network, and multiblock correlation, as
explained in the following.
Linking the Chemical Fingerprint to the Bioactivity
Profile. Gene coexpression module identification. We identified
coexpression modules (clusters of genes with covarying
expression patterns) with a customized computational script.
We conducted 50 permutations, which included two steps in
each run: first, we calculated Pearson correlations based on
bootstrapped data with noise; second, we applied threshold on
correlation coecients to ensure that the coexpression network
follows a scale-free topology, with a detailed explanation of
diagnostic plots presented in the Supporting Information
section A. After that, we ensembled results from 50 runs to
construct a coexpression network. We then used the “infomap”
algorithm
29
to identify gene clusters with similar expression
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
C

patterns across 30 water samples, hereafter called coexpression
modules.
Multiblock Correlation Analysis. To identify correlations
between chemical fingerprints and bioactivity profiles, we used
the ‘Sparse Generalized Canonical Correlation Analysis’
(SGCCA) algorithm in the RGCCA package (version
2.1.2).
30,31
This approach identifies correlations between
variations of chemical substances within mixtures and variations
of genes within modules on a low-dimensional latent space. In
this latent space, projected gene and chemical vectors, so-called
gene and chemical low-dimension representations (LDRs), are
linearly correlated. The relative contribution of individual
chemicals and genes to the corresponding LDRs is measured
by squared weights (SWs); the sum of SWs of all chemicals or
genes selected per LDR equals 1.
Module Enrichment Analysis. Using a Mann−Whitney U
test, an enrichment analysis was conducted to identify the
coexpression module consisting of genes with significantly larger
SWs than genes not within the module. The resulting P-values
were adjusted for a false discovery rate of 0.05 using the
Benjamini-Hochberg approach. This analysis provided a list of
significantly enriched coexpression modules for each LDR.
Pathway Overrepresentation Analysis of Significantly
Enriched Coexpression Modules. We performed a pathway
overrepresentation analysis on each significantly enriched
coexpression module per LDR, using the KEGG Pathway
database.
32
This enabled the identification of overrepresented
metabolic pathways per module and LDR. Significantly enriched
pathways (adjusted P-value <0.05) were identified following a
chi-square test adjusted for multiple testing using the Benjamini-
Hochberg procedure.
■RESULTS AND DISCUSSION
Diversity and Distribution of Chemical Pollutants
among the Sampling Sites. The chemical fingerprints of
30 river samples detected and quantified 385 chemical
substances. As shown in Figure S1, a large proportion of the
variance in the chemical fingerprints separated the upstream
(Bai River and Chao River) from the downstream sites (Chaobai
River) (31.4% of variance); conversely, 9.8% of the variance
separated sites within the rivers. The upstream sites clustered
closer together, indicating a higher similarity in the chemical
mixtures in upstream than downstream waters past their
confluence point (Figure S1). This could be explained by the
ongoing impacts of domestic pollution and intensive farming
activities across the downstream sites.
16
Of the 385 chemicals detected, 215 were present at more than
15 sites, including 16 inorganic compounds (7 were heavy
metals), 15 polycyclic aromatic hydrocarbons (PAHs), and 184
polar organic compounds (Table S2). Additionally, 44
chemicals were exclusive to a single sampling site, whereas 90
were ubiquitously detected across all 30 sites (Table S2). The
215 chemicals detected at more than 15 sites, could be grouped
in 19 chemical classes, including PAHs, perfluorinated
compounds (PFCs), pharmaceuticals and pesticides (Figure
S2). Fifteen PAHs were detected across the three rivers. Four
PFCs were prevalently detected in downstream river sites,
including perfluorooctanoic acid (PFOA), perfluorooctanesul-
fonic acid (PFOS), perfluoroheptanoic acid, and 6:2 fluo-
rotelomer sulfonic acid; these PFCs are known as “forever
compounds” due to their persistence in the environment, and
they primarily come from industrial discharges, landfill leachate,
and atmospheric deposition.
33
Out of 33 pharmaceuticals
commonly detected across the 30 river sites, 17 were
consistently detected in downstream waters (specifically in the
Chaobai River). Six pharmaceuticals (i.e., n-acetyl-4-amino-
antipyrine, amantadine, budesonide, dimethylamino phenazone,
diazepam, and 4-formyl-antipyrine) were detected in all 30 sites,
likely resulting of untreated domestic wastewater entering the
rivers.
15
It is possible that some of the detected compounds are
transformation products of parent compounds. Of the 51
pesticides detected in more than 15 sites, 20 were common to all
downstream sites of the Chaobai River (Table S2). These
pesticides and their transformation products likely originate
from agricultural runos from farmlands
34
along the Chaobai
River. Many of the pesticides found in our study are also found in
other rivers in China due to their wide application in controlling
weeds in crops.
35,36
Quantifying 385 chemical substances allowed us to generate a
comprehensive chemical fingerprinting of the water samples
using target and target screening approaches. However, this
number is likely an underestimation of the total number of
chemical compounds present in the Chaobai River. Further
refinement of detection methods and broader analytical
coverage are needed to capture the full complexity of the
chemical mixtures in these waters.
Bioactivity Profiles of the Chaobai Water Samples.
Although no acute toxicity (immobilization) was observed in the
Daphnia exposed to the Chaobai water samples, we observed
biomolecular responses through variations in the expression of
19,942 genes. Following preprocessing (filtering) steps
described in the methods, 10,440 genes were retained for
downstream analysis. With these genes, we built a coexpression
network to identify gene coexpression modules. A total of 22
modules (including 7608 genes) were identified through this
approach, comprising between 23 (module 22) and 1026 genes
(module 1) (Figure S3).
The PCA plot based on 10,440 genes showed no clear
partitioning of the data by location and/or river (Figure S4),
which was unexpected given the partitioning of the chemicals by
rivers (Figure S2). This suggests that the variation in the mixture
composition between sites, as revealed by quantifying hundreds
of chemicals, is not reflected in the changes in the expression of
individual genes. This result diers from two previous studies
using Daphnia magna’s transcriptomic profiles to assess
biomolecular responses to surface water samples.
37,38
However,
those previous studies measured physicochemical parameters
(e.g., pH, temperature, suspended solids) and a limited number
of inorganic substances (e.g., TN, TP, As, Cr, etc.). It is plausible
that other environmental factors not considered in our study
(e.g, dissolved oxygen, salinity, humic substances) may have
influenced the bioavailability of the chemicals studied. However,
the absence of distinct gene-specific responses across river sites
is not unexpected. Many of the quantified chemicals are present
at trace levels, which are unlikely to elicit strong, gene-specific
expression changes. Instead, these chemicals are more likely to
induce subtle, coordinated changes in the expression of multiple
genes. These changes, while biologically relevant, may fall below
the statistical detection threshold, potentially explaining the lack
of significant gene-by-site partitioning observed in the PCA plot
(Figure S4). Therefore, we used the coexpression patterns
among genes to discover covariation with ambient chemical
mixtures in the Chaobai River system.
Correlation Analysis via SGCCA. To correlate coex-
pression modules of bioactivity profiles with ambient chemical
mixtures, we used ‘Sparse Generalized Canonical Correlation
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
D

Analysis’ (SGCCA). This correlation analysis identified five low-
dimension representations (LDRs) with correlation coecients
ranging between 0.14 (LDR5) and 0.67 (LDR1) (Figure S5).
Based on the relative contribution (squared weight, SW) of
individual chemical features in each LDR, 44 chemical
substances with SW more than 0.01 contributed to the
SGCCA model, including nine inorganic chemicals in LDR1,
Figure 2. Chemical substance contributions to individual low-dimensional representation (LDR). These chemicals are ranked according to the
squared weights (SW) and color-coded by their chemical classes. The names and chemical classes of these chemical substances are described in Table
S2. Only chemical substances with squared weights greater than 0.01 are shown. The details of SWs are listed in Table S3.
Figure 3. Biomolecular eects associated with individual low-dimensional representation (LDR). The correlation analysis (SGCCA) identifies five
correlations between chemicals and bioactivity profiles. The five LDRs are coded by color. The strength of correlation is determined by chemical
classes’ sums of squared weights (SWs) listed in Table S3.
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
E

six heavy metals in LDR2, nine PAHs in LDR3, ten polar organic
chemicals in LDR4, and a dierent set of ten polar organic
chemicals in LDR5 (Figure 2). The details of the SWs for these
44 chemical substances are listed in Table S3. Notably, these 44
chemical substances together explain 46.4% of the total variance
in transcriptomic data, which was 86.7% of the variance
explained by 215 chemical substances used in the SGCCA
model. The chemical substances selected in LDR4 explained the
highest amount (21.8%) of the total variance of the bioactivity
profiles, followed by LDR 5 (14.3%), LDR3 (4.1%), LDR2
(3.1%) and LDR1 (3.0%) (Figure S5). These patterns suggest
that these 44 chemical substances may be the driving factor of
gene expression responses in exposed D. magna in this study,
with organic chemical substances being the primary contributor.
While the rest may subtly influence the daphnid transcriptome,
it is premature to determine whether such chemicals pose
discernible hazards to exposed daphnids and may future
experimental investigation.
Based on the SW of individual genes, we performed a module
enrichment analysis within LDRs. Twenty-one coexpression
modules showed significant enrichment within one or more
LDRs (adjusted P-value <0.05) (Figure S6). Four coexpression
modules (modules 5, 7, 8, and 10) were shared across the LDRs.
Eight modules were shared between LDR1 and LDR2, and eight
modules were shared among LDR3, LDR4 and LDR5. Modules
11 and 21 were unique to LDR4, whereas module 12 was unique
to LDR5 (Figure S6).
We conducted a pathway overrepresentation analysis to
identify significantly enriched pathways in the modules
contributing significantly to each LDR. Five LDRs were
associated with 80 KEGG pathways, which were categorized
into 28 biomolecular eect categories (Figure 3 and Table S4).
Five categories of biomolecular eects were shared among LDRs
(LDR3-LDR5), including “carbohydrate metabolism”, ‘folding,
sorting, and degradation’, “transport and catabolism”, “energy
metabolism”, and “digestive system”. This pattern suggests that
chemical mixtures may induce multiple biomolecular eects on
distinct biological processes. Furthermore, top-ranked pathways
are distinctive among these three LDRs (Table S5), which
suggests that similar hazards can be associated with distinctive
metabolic pathways. Besides, it is necessary to compare the
bioactivity profiles of individual compounds with the bioactivity
profiles of the chemical mixtures identified by this data-driven
model to further determine whether synergistic interactions
among chemicals within the mixture are occurring.
39,40
Synergistic interactions may induce distinct pathways than
those expected from exposures to individual chemicals.
41,42
Furthermore, we cannot exclude that additional chemicals that
were not quantified within our study from having contributed to
the observed bioactivity profiles.
Figure 4. Pathway-level bioactivity signatures of individual low-dimensional representation (LDR). The pathway overrepresentation analysis evaluates
the significance of enriching a specific pathway in significantly enriched modules (Figure S6) per LDR. These 30 pathways were selected based on their
adjusted P-values. The log10 transformed adjusted P-values (Padj) of the chi-square testing significant correlations after FDR correction are shown.
Asterisks (*) indicate pathways that are significantly enriched in an LDR based on Padj <0.05.
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
F

Bioactivity Signatures of Three Organic Chemical
Mixtures. We benchmarked and evaluated the pathway-level
bioactivity signatures of three organic chemical mixtures from
LDR3−5 using a comprehensive review of published studies. A
total of 78 relevant papers were identified through a systematic
search using keywords including the specific chemicals,
“Daphnia,” “gene/transcript,” and “pathway.” These studies
provided critical insights into the links between the chemicals
and their biomolecular functions in Daphnia. When Daphnia-
specific studies were unavailable, we supplemented our analysis
with studies of other species to elucidate the chemical modes of
action. This approach allowed us to infer potential mechanisms
of toxicity and molecular interactions, providing a broader
context for interpreting the observed pathway-level bioactivity
detected in this study.
In LDR3, we identified 26 enriched pathways (Table S5)
associated with nine PAHs, i.e., Chry, Pyr, Ace, BaA, Acy, Fluo,
Phe, Flua, and BaP. The top-ranked pathways in this LDR
comprise the synaptic vesicle cycle, oxidative phosphorylation,
chemical carcinogenesis (forming DNA adducts), steroid
biosynthesis, metabolism of xenobiotics by cytochrome P450
(CYP), drug metabolism, lysosome, and glutathione metabolism
(Figure 4). PAH in LDR3 correlated with xenobiotic
metabolism, oxidative stress, carcinogenesis, lipid metabolism
and nervous system functions. CYPs are the functional enzymes
in phase I of xenobiotic metabolism. Among these, there are the
CYP monooxygenases involved in the biotransformation of
PAHs in Daphnia.
43
Enzymes observed in LDR3 and linked to PAHs have been
previously associated with redox responses and phase II
xenobiotic metabolism,
44
including D. magna.
45
The pathway
related to the formation of DNA adducts, which we identified in
the top-ranked pathways, has been previously linked to
carcinogenesis.
46
The perturbation in lipid metabolism observed
in our study has been previously associated with chronic PAH
toxicity.
47
Previous studies have shown that BaP exposure may induce
neurotoxicity in aquatic invertebrates
48,49
by inhibiting
acetylcholinesterase (AChE) and choline acetyltransferase
(ChAT) activities.
50
Although the cholinergic synapse pathway,
which includes AChE and ChAT, was not significantly enriched
in LDR3; AChE, a key biomarker in these enriched pathways,
was a member of enriched modules. This finding suggests that
the PAH mixture induces neurotoxicity by aecting AChE
activity, which is crucial for terminating neurotransmission
through the breakdown of acetylcholine released during the
synaptic vesicle cycle.
In this LDR, we also identified novel pathways of toxicity. For
example, a pathway regulating cardiac muscle contraction and
three pathways critical for maintaining normal cardiac muscle
contraction (aspartate β-hydroxylase, cytochrome c oxidase
subunit 6b, and voltage-dependent calcium channel α-2/delta-
1) were among the top-ranked pathways. Perturbation of these
pathways could lead to severe cardiac dysfunctions and suggests
the hazards posed by PAH mixtures can be more severe than
those from single PAH exposures.
51
This finding agrees with a
previous study showing the enhanced eects of PAH mixtures in
D. magna.
52
In LDR4, we identified 35 enriched pathways (Table S5)
correlated with nine organic polar compounds, including PFOA,
pesticides (imidacloprid-urea, 2-hydroxyatrazine), a rubber
additive (2-hydroxybenzothiazole), a food ingredient (har-
mine), three industrial chemicals (2,2,6,6-tetramethyl-4 piper-
idone, n-butylbenzenesulfonamide, 4-tert-butylbenzenesulfona-
mide), a pharmaceutical (diazepam), and a plastic additive
(triethylphosphate). These pathways include steroid biosyn-
thesis, biosynthesis of unsaturated fatty acids, fatty acid
metabolism, mTOR signaling pathway, ABC transporters and
PPAR signaling pathways, which have been previously
associated with PFOA in rodents.
53−55
Specifically, the MoAs
of PFOA have been linked to the deregulating PPARα-
dependent signaling pathway,
53
the increased expression of
mTOR that controls lipid synthesis,
54
and the inhibition of
cholesterol biosynthesis via the downregulation of the 3-
hydroxy-3-methylglutaryl-CoA reductase and acyl CoA choles-
terol acyltransferase.
55
Studies of Daphnia’s responses to PFOA
are limited, and a comprehensive overview of the pathway-level
bioactivity signature of PFOA was not yet possible.
56
While we
recognize that there are major dierences between mammals
and invertebrates,
57
this bioactivity signature identified in our
study indicates the putative biomolecular targets in Daphnia.
Furthermore, we also identified a correlation between
imidacloprid-urea (IDPu), a degradation product of a neuro-
toxic insecticide (imidacloprid) that acts on nicotinic acetylcho-
line receptors, and the synaptic vesicle cycle pathway
58
(Figure
4), suggesting that IDP impairs the Daphnia nervous system.
59
This chloronicotinyl insecticide is consistently detected in all
downstream sites (Chaobai River, Table S2) and is commonly
detected in aquatic environments.
59
Additionally, IDPu has
been found to suppress terpenoid backbone biosynthesis and
fatty acid metabolism,
60
which were also observed in our study.
In LDR5, we identified 38 enriched pathways (Table S5),
including tyrosine metabolism, synaptic vesicle cycle, cholester-
ol metabolism, glutathione metabolism, PI3K-Akt signaling
pathway, and cGMP-PKG signaling pathway (Figure 4). These
pathways were associated with a pesticide metabolite (dimetha-
chlor ESA), a biocide and its metabolite (fipronil and fipronil
sulfide), a food ingredient (harman), and six pharmaceuticals
(amantadine, tramadol, clofibric acid, losartan, n-acetyl-4-
aminoantipyrine, and n-formyl-4-aminoantipyrine) (Figure 4).
Within this bioactive mixture, dimethachlor ESA (a metabolite
of the chloroacetanilide herbicide dimethachlor), fipronil
(phenylpyrazole insecticide), and fipronil sulfide (a trans-
formation product of fipronil) are GABA antagonists and
obstruct the ion channel. Exposure to these pesticides can
reduce GABA release, disrupting intraneural signal transduction
and thereby suppressing neurotransmitter release.
61
Our study
also identified four pathways that were previously linked to
cardiac and muscular dysfunctions. Organochlorine pesticides
62
and fipronil
63
were previously linked to irregular heart rates in
Daphnia.
In addition, clofibric acid, as a metabolite of clofibrate, targets
cholesterol metabolism to promote fatty acid metabolism. This
agrees with the pathways identified in our study, such as
cholesterol metabolism and fat digestion and absorption.
64
Furthermore, the renin-angiotensin system pathway related to
the endocrine system was identified in LDR5. This pathway is
involved in the modulation of angiotensin peptides. Dysregu-
lation of these enzymes can lead to angiotensin II and III
imbalances, potentially causing inflammation or even organ
damage.
65
As losartan is an angiotensin II receptor blocker, it
specifically inhibits the action of angiotensin II by blocking its
binding to the AT1 receptor,
66
which could be involved in the
modulation of the renin-angiotensin system.
Advantages and Limitations of the Data-Driven
Approach. A key advantage of the data-driven approach used
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
G

in this study is its ability to identify hazards induced by real-
world chemical mixtures without bias. Unlike traditional
methods that are hypothesis-driven and test chemicals at
concentrations that the organisms rarely encounter in the
natural environment, our approach examines biomolecular
signatures induced by sublethal concentrations of ambient
chemical mixtures. These bioactivity signatures can serve as
early warning indicators of potential hazards, allowing protective
measures to be taken. The multiblock correlation analysis that
we employed had reduced the chemical complexity of
environmental water samples to five components that have
distinct pathway-level bioactivity signatures. Some signatures
are indicative of known toxicity pathway and known adversities,
yet here are associated with the potential MoAs of a combination
of chemicals. Although it is not possible (not necessary) to
discern how the chemicals interact as a component of the total
mixture to produce their distinct biological eects, this approach
significantly reduces the need for extensive testing across
hundreds of chemicals. This is particularly advantageous since
most chemicals from domestic and industrial sources are
released into the environment as unintended mixtures with
largely unknown toxicities.
67
By linking chemicals to specific
biomolecular signatures, our approach supports high-through-
put screening of potential environmental hazards, identifying
likely toxicity targets for further experimental validation.
14
Specifically, in each LDR, we could associate biomolecular
functions with potential chemical hazards within the ambient
mixtures. Therefore, this methodology serves as an initial,
unbiased screening tool for unintentional (thus unknown)
hazards, potentially improving the eciency and throughput of
environmental risk assessments when routinely applied.
Our approach has its limitations. We selected the 48 h time
point for this study because it aligns with the time frame used to
measure ecological end points commonly adopted by regulatory
agencies for risk assessment.
24
This provides a practical and
relevant foundation for benchmarking our data-driven approach
against established ecological metrics. By targeting this
regulatory benchmark, we aimed to ensure the translational
relevance of our findings, particularly in the context of
environmental decision-making frameworks. Conversely, early
onset molecular events, such as gene expression changes, can
occur before measurable ecological impacts become evident.
These early responses can serve as sensitive indicators of
chemical exposure and potential stressor impacts, oering the
advantage of early warning signals in environmental monitor-
ing.
14
Integrating such early biomolecular data with apical end
points could improve predictive modeling of long-term or
population-level eects, providing a more proactive approach to
risk assessment. Future work could benefit from incorporating
multiple time points to capture the temporal dynamics of
molecular responses
68
and their correlation with ecological end
points. Such an approach would allow us to refine our data-
driven framework, potentially enhancing its sensitivity and
predictive capacity while broadening its applicability across
dierent regulatory contexts.
Another limitation of our study is that identifying correlations
between biomolecular signatures and chemical mixtures does
not necessarily establish causality. The associations discovered
by our approach require future experimental validation using
surrogate animal species to move from correlation to
causation.
69
The validation of these observations are crucial
for confirming the biological relevance and potential environ-
mental impact of the identified correlations.
In addition to these limitations, adopting the novel method-
ologies proposed in this study will require adjustments in
regulatory frameworks, including phases of testing, validation,
and acceptance by regulatory bodies. Despite these challenges,
our data-driven approach has the potential to be transformative
by reducing the need to conduct “forward” testing regimes,
where first bionart, then trinary and higher-order mixtures are
tested to assess the chemical mixture eects. Otherwise, these
combinatorial tests are intractable at the scale of real-world
complex euents composed of hundreds or thousands of
distinct substances. More importantly, it enables the generation
of bioactivity signatures that enhance the mechanistic under-
standing of the hazardous eects of environmental chemical
mixtures and their MoAs.
■ASSOCIATED CONTENT
Data Availability Statement
Chemical fingerprinting data (target and target screening
analysis) are listed in the Supporting Information Table S2.
Raw transcriptomics data are submitted to NCBI
(PRJNA809147).
*
sı Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.est.4c11095.
Section A (Detailed methodology); Section B (Six
supplementary figures: PCA plots for chemical and
bioactivity profiles; Chemical fingerprinting; Diagnostic
plots of gene coexpression network analysis; Visual-
izations of multiblock correlation analysis; Module
enrichment analysis); Section C (Five supplementary
tables: Description of sampling sites; Description of
chemical fingerprinting; Square weight of individual
chemicals in multiblock correlation analysis; Biomolecu-
lar eect profiles; Pathway profiles) (PDF)
Table S1. Overview of the sampling sites along the
Chaobai River Basin; Table S2. Description of the
analytical chemical fingerprints; Table S3. The relative
contribution of individual chemical substance to each
LDR; Table S4. Biomolecular eect of each LDR; Table
S5. Bioactivity signature of each LDR (XLSX)
■AUTHOR INFORMATION
Corresponding Author
Xiaojing Li −Centre for Environmental Research and Justice
(CERJ), School of Biosciences, The University of Birmingham,
Birmingham B15 2TT, U.K.; orcid.org/0000-0002-2796-
5342; Email: x.li.12@bham.ac.uk
Authors
Jiarui Zhou −Centre for Environmental Research and Justice
(CERJ), School of Biosciences, The University of Birmingham,
Birmingham B15 2TT, U.K.; orcid.org/0000-0002-1025-
718X
Yaohui Bai −Research Centre for Eco-Environmental Sciences,
Chinese Academy of Sciences, Beijing 100085, P. R. China;
orcid.org/0000-0002-2086-4477
Meng Qiao −Research Centre for Eco-Environmental Sciences,
Chinese Academy of Sciences, Beijing 100085, P. R. China
Wei Xiong −Key Laboratory of Environmental Biotechnology,
Research Centre for Eco-Environmental Sciences, Chinese
Academy of Sciences, Beijing 100085, P. R. China
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
H

Tobias Schulze −Department Exposure Science, Helmholtz
Centre for Environmental Research −UFZ, 04318 Leipzig,
Germany
Martin Krauss −Department Exposure Science, Helmholtz
Centre for Environmental Research −UFZ, 04318 Leipzig,
Germany; orcid.org/0000-0002-0362-4244
Timothy D. Williams −Centre for Environmental Research and
Justice (CERJ), School of Biosciences, The University of
Birmingham, Birmingham B15 2TT, U.K.
Ben Brown −Environmental Genomics and Systems Biology
Division, Lawrence Berkeley National Laboratory, Berkeley
94720, United States
Luisa Orsini −Centre for Environmental Research and Justice
(CERJ), School of Biosciences, The University of Birmingham,
Birmingham B15 2TT, U.K.; The Alan Turing Institute,
British Library, London NW1 2DB, U.K.
Liang-Hong Guo −Hangzhou Institute for Advanced Study,
UCAS, Hangzhou, Zhejiang 310020, P. R. China;
orcid.org/0000-0003-1399-5716
John K. Colbourne −Centre for Environmental Research and
Justice (CERJ), School of Biosciences, The University of
Birmingham, Birmingham B15 2TT, U.K.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.est.4c11095
Author Contributions
◆
X.L. and J.Z. share the first authorship. L.O., L.-H.G., and
J.K.C. share senior authorship. X.L. and J.Z. analyzed data for the
case study. Y.B., M.Q., W.X., T.S., and M.K. generated the target
and target screening data of inorganic substances, PAHs, and
organic substances of the Chaobai River waters. T.W. and B.B.
supervised the writing. X.L. coordinated data analysis and
interpretation. X.L., L.O., and J.K.C. wrote various drafts of the
manuscript and compiled the final version of the manuscript. All
authors contributed to the manuscript writing.
Notes
The authors declare no competing financial interest.
■ACKNOWLEDGMENTS
This work was part of the China-UK Research of Safeguarding
Natural Water project, funded by the Royal Society Interna-
tional Collaboration Award (grant No IC160121), and of the
PrecisionTox project that has received funding from the
European Union’s Horizon 2020 research and innovation
programme (grant No. 965406). This output reflects only the
authors’ view and the European Union cannot be held
responsible for any used that may be made of the information
contained therein. The field sampling was funded by the
National Natural Science Foundation of China (grant No.
U20A20133). X.L. is funded by the Natural Environmental
Research Council Innovation People (grant No. NE/Y005120/
1). The field sampling, exposure assay with Daphnia magna, and
target chemical analysis were supported by coauthors from the
Research Centre for Eco-Environmental Sciences. We gratefully
acknowledge access to the platform CITEPro (Chemicals in the
Environment Profiler) at Helmholtz Centre for Environmental
Research (UFZ, Germany) funded by the Helmholtz Associa-
tion for the LC-HRMS analyses. We thank William Scavone for
the artwork of the graphical abstract and Andrey Frankfurt for
the artwork of Figure
1
.
■REFERENCES
(1) Landrigan, P. J.; Fuller, R.; Acosta, N. J. R.; Adeyi, O.; Arnold, R.;
Basu, N.; Nil; Baldé, A. B.; Bertollini, R.; Bose-O’Reilly, S.; Boufford, J.
I.; Breysse, P. N.; Chiles, T.; Mahidol, C.; Coll-Seck, A. M.; Cropper, M.
L.; Fobil, J.; Fuster, V.; Greenstone, M.; Haines, A.; Hanrahan, D.;
Hunter, D.; Khare, M.; Krupnick, A.; Lanphear, B.; Lohani, B.; Martin,
K.; Mathiasen, K. V.; McTeer, M. A.; Murray, C. J. L.; Ndahimananjara,
J. D.; Perera, F.; Potocnik, J.; Preker, A. S.; Ramesh, J.; Rockström, J.;
Salinas, C.; Samson, L. D.; Sandilya, K.; Sly, P. D.; Smith, K. R.; Steiner,
A.; Stewart, R. B.; Suk, W. A.; van Schayck, O. C. P.; Yadama, G. N.;
Yumkella, K.; Zhong, M. The Lancet Commission on Pollution and
Health. Lancet 2018,391 (10119), 462−512.
(2) Wang, B.; Dai, G.; Deng, S.; Huang, J.; Wang, Y.; Yu, G. Linking
the environmental loads to the fate of PPCPs in Beijing: Considering
both the treated and untreated wastewater sources. Environ. Pollut.
2015,202, 153−159.
(3) Falås, P.; Wick, A.; Castronovo, S.; Habermacher, J.; Ternes, T. A.;
Joss, A. Tracing the Limits of Organic Micropollutant Removal in
Biological Wastewater Treatment. Water Res. 2016,95, 240−249.
(4) Fuller, R.; Landrigan, P. J.; Balakrishnan, K.; Bathan, G.; Bose-
O’Reilly, S.; Brauer, M.; Caravanos, J.; Chiles, T.; Cohen, A.; Corra, L.;
Cropper, M.; Ferraro, G.; Hanna, J.; Hanrahan, D.; Hu, H.; Hunter, D.;
Janata, G.; Kupka, R.; Lanphear, B.; Lichtveld, M.; Martin, K.;
Mustapha, A.; Sanchez-Triana, E.; Sandilya, K.; Schaefli, L.; Shaw, J.;
Seddon, J.; Suk, W.; Téllez-Rojo, M. M.; Yan, C. Pollution and Health:
A Progress Update. Lancet Planet. Health 2022,6(6), e535−e547.
(5) Hollender, J.; Schymanski, E. L.; Singer, H. P.; Ferguson, P. L.
Nontarget Screening with High Resolution Mass Spectrometry in the
Environment: Ready to Go? Environ. Sci. Technol. 2017,51 (20),
11505−11512.
(6) Kortenkamp, A.; Faust, M.; Backhaus, T.; Altenburger, R.;
Scholze, M.; Muller, C.; Ermler, S.; Posthuma, L.; Brack, W. Mixture
Risks Threaten Water Quality: The European Collaborative Project
SOLUTIONS Recommends Changes to the WFD and Better
Coordination across All Pieces of European Chemicals Legislation to
Improve Protection from Exposure of the Aquatic Environment to
Multiple Pollutants. Environ. Sci. Eur. 2019,31 (1), 69.
(7) Olker, J. H.; Elonen, C. M.; Pilli, A.; Anderson, A.; Kinziger, B.;
Erickson, S.; Skopinski, M.; Pomplun, A.; LaLone, C. A.; Russom, C. L.;
Hoff, D. The ECOTOXicology Knowledgebase: A Curated Database of
Ecologically Relevant Toxicity Tests to Support Environmental
Research and Risk Assessment. Environ. Toxicol. Chem. 2022,41 (6),
1520−1539.
(8) Harrill, J. A.; Everett, L. J.; Haggard, D. E.; Sheffield, T.; Bundy, J.
L.; Willis, C. M.; Thomas, R. S.; Shah, I.; Judson, R. S. High-
Throughput Transcriptomics Platform for Screening Environmental
Chemicals. Toxicol. Sci. 2021,181 (1), 68−89.
(9) Li, Y.; Zhang, Z.; Jiang, S.; Xu, F.; Tulum, L.; Li, K.; Liu, S.; Li, S.;
Chang, L.; Liddell, M.; Tu, F.; Gu, X.; Carmichael, P. L.; White, A.;
Peng, S.; Zhang, Q.; Li, J.; Zuo, T.; Kukic, P.; Xu, P. Using
Transcriptomics, Proteomics and Phosphoproteomics as New
Approach Methodology (NAM) to Define Biological Responses for
Chemical Safety Assessment. Chemosphere 2023,313, No. 137359.
(10) Varsou, D.-D.; Ellis, L.-J. A.; Afantitis, A.; Melagraki, G.; Lynch, I.
Ecotoxicological Read-across Models for Predicting Acute Toxicity of
Freshly Dispersed versus Medium-Aged NMs to Daphnia Magna.
Chemosphere 2021,285, No. 131452.
(11) Martin, R.; Hazemi, M.; Flynn, K.; Villeneuve, D.; Wehmas, L.
Short-Term Transcriptomic Points of Departure Are Consistent with
Chronic Points of Departure for Three Organophosphate Pesticides
across Mouse and Fathead Minnow. Toxics 2023,11 (10), 820.
(12) O’Rourke, A.; Beyhan, S.; Choi, Y.; Morales, P.; Chan, A. P.;
Espinoza, J. L.; Dupont, C. L.; Meyer, K. J.; Spoering, A.; Lewis, K.;
Nierman, W. C.; Nelson, K. E. Mechanism-of-Action Classification of
Antibiotics by Global Transcriptome Profiling. Antimicrob. Agents
Chemother. 2020,64 (3), No. e01207-19.
(13) Westmoreland, C.; Bender, H. J.; Doe, J. E.; Jacobs, M. N.; Kass,
G. E.; Madia, F.; Mahony, C.; Manou, I.; Maxwell, G.; Prieto, P.;
Roggeband, R.; et al. Use of New Approach Methodologies (NAMs) in
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
I

regulatory decisions for chemical safety: Report from an EPAA Deep
Dive Workshop. Regul. Toxicol. Pharmacol. 2022,135, No. 105261.
(14) Abdullahi, M.; Li, X.; Abdallah, M. A.-E.; Stubbings, W.; Yan, N.;
Barnard, M.; Guo, L.-H.; Colbourne, J. K.; Orsini, L. Daphnia as a
Sentinel Species for Environmental Health Protection: A Perspective
on Biomonitoring and Bioremediation of Chemical Pollution. Environ.
Sci. Technol. 2022,56 (20), 14237−14248.
(15) Yang, L.; He, J.; Liu, Y.; Wang, J.; Jiang, L.; Wang, G.
Characteristics of change in water quality along reclaimed water intake
area of the Chaobai River in Beijing, China. J. Environ. Sci. 2016,50,
93−102.
(16) Su, D.; Ben, W.; Strobel, B. W.; Qiang, Z. Occurrence, Source
Estimation and Risk Assessment of Pharmaceuticals in the Chaobai
River Characterized by Adjacent Land Use. Sci. Total Environ. 2020,
712, No. 134525.
(17) Xiong, W.; Ni, P.; Chen, Y.; Gao, Y.; Shan, B.; Zhan, A.
Zooplankton Community Structure along a Pollution Gradient at Fine
Geographical Scales in River Ecosystems: The Importance of Species
Sorting over Dispersal. Mol. Ecol. 2017,26 (16), 4351−4360.
(18) Wang, G.; Xia, J.; Chen, J. Quantification of effects of climate
variations and human activities on runoff by a monthly water balance
model: A case study of the Chaobai River basin in northern China.
Water Resour. Res. 2009,45 (7), No. W00A11.
(19) Qiao, M.; Fu, L.; Li, Z.; Liu, D.; Bai, Y.; Zhao, X. Distribution and
Ecological Risk of Substituted and Parent Polycyclic Aromatic
Hydrocarbons in Surface Waters of the Bai, Chao, and Chaobai Rivers
in Northern China. Environ. Pollut. 2020,257, No. 113600.
(20) He, B.; He, J.; Wang, J.; Li, J.; Wang, F. Characteristics of GHG
Flux from Water-Air Interface along a Reclaimed Water Intake Area of
the Chaobai River in Shunyi, Beijing. Atmos. Environ. 2018,172, 102−
108.
(21) Liao, K.; Bai, Y.; Huo, Y.; Jian, Z.; Hu, W.; Zhao, C.; Qu, J.
Integrating Microbial Biomass, Composition and Function to Discern
the Level of Anthropogenic Activity in a River Ecosystem. Environ. Int.
2018,116, 147−155.
(22) Beckers, L.-M.; Brack, W.; Dann, J. P.; Krauss, M.; Muller, E.;
Schulze, T. Unraveling Longitudinal Pollution Patterns of Organic
Micropollutants in a River by Non-Target Screening and Cluster
Analysis. Sci. Total Environ. 2020,727, No. 138388.
(23) Pujianto, U.; Wibawa, A. P.; Akbar, M. I. K-nearest neighbor (k-
NN) based missing data imputation. In 2019 5th International
Conference on Science in Information Technology (ICSITech); IEEE,
2019; pp 83−88.
(24) OECD. Test No. 202: Daphnia Sp. Acute Immobilisation Test;
OECD Guidelines for the Testing of Chemicals, Section 2; OECD, 2004.
(25) Bolger, A. M.; Lohse, M.; Usadel, B. Trimmomatic: A Flexible
Trimmer for Illumina Sequence Data. Bioinformatics 2014,30 (15),
2114−2120.
(26) Chaturvedi, A.; Li, X.; Dhandapani, V.; Marshall, H.; Kissane, S.;
Cuenca-Cambronero, M.; Asole, G.; Calvet, F.; Ruiz-Romero, M.;
Marangio, P.; Guigó, R.; Rago, D.; Mirbahai, L.; Eastwood, N.;
Colbourne, J. K.; Zhou, J.; Mallon, E.; Orsini, L. The Hologenome of
Daphnia magna Reveals Possible DNA Methylation and Microbiome-
Mediated Evolution of the Host Genome. Nucleic Acids Res. 2023,51
(18), 9785−9803.
(27) Dobin, A.; Gingeras, T. R. Mapping RNA-seq Reads with STAR.
Curr. Protoc. Bioinf. 2015,51 (1), 11−14.
(28) Hansen, K. D.; Irizarry, R. A.; Wu, Z. Removing Technical
Variability in RNA-Seq Data Using Conditional Quantile Normal-
ization. Biostatistics 2012,13 (2), 204−216.
(29) Rosvall, M.; Bergstrom, C. T. Maps of Random Walks on
Complex Networks Reveal Community Structure. Proc. Natl. Acad. Sci.
U.S.A. 2008,105 (4), 1118−1123.
(30) Girka, F.; Camenen, E.; Peltier, C.; Gloaguen, A.; Guillemot, V.;
Le Brusquet, L.; Tenenhaus, A. Multiblock data analysis with the
RGCCA package. J. Stat. Software 2023, 1−36.
(31) Tenenhaus, A.; Philippe, C.; Guillemot, V.; Le Cao, K.-A.; Grill,
J.; Frouin, V. Variable Selection for Generalized Canonical Correlation
Analysis. Biostatistics 2014,15 (3), 569−583.
(32) Kanehisa, M.; Furumichi, M.; Tanabe, M.; Sato, Y.; Morishima,
K. KEGG: New Perspectives on Genomes, Pathways, Diseases and
Drugs. Nucleic Acids Res. 2017,45 (D1), D353−D361.
(33) Kurwadkar, S.; Dane, J.; Kanel, S. R.; Nadagouda, M. N.;
Cawdrey, R. W.; Ambade, B.; Struckhoff, G. C.; Wilkin, R. Per- and
Polyfluoroalkyl Substances in Water and Wastewater: A Critical Review
of Their Global Occurrence and Distribution. Sci. Total Environ. 2022,
809, No. 151003.
(34) Chen, J.; Chen, L.; Zhongfang, Y.; Jiyuan, W. A. N. G. Residues
and characteristics of organochlorine pesticides in the surface water in
the suburb of Beijing. Earth Sci. Front. 2008,15 (5), 242−247.
(35) Sun, J. T.; Pan, L. L.; Zhan, Y.; Tsang, D. C. W.; Zhu, L. Z.; Li, X.
D. Atrazine Contamination in Agricultural Soils from the Yangtze River
Delta of China and Associated Health Risks. Environ. Geochem. Health
2017,39 (2), 369−378.
(36) Zhao, J.-H.; Hu, L.-X.; Xiao, S.; Zhao, J.-L.; Liu, Y.-S.; Yang, B.;
Zhang, Q.-Q.; Ying, G.-G. Screening and Prioritization of Organic
Chemicals in a Large River Basin by Suspect and Non-Target Analysis.
Environ. Pollut. 2023,333, No. 122098.
(37) Talu, M.; Seyoum, A.; Yitayew, B.; AdaneMihret; Aseffa, A.; Jass,
J.; Mamo, G.; Olsson, P.-E. Transcriptional Responses of Daphnia
magna Exposed to Akaki River Water. Environ. Monit. Assess. 2022,194
(5), No. 349.
(38) Kim, B.-M.; Ahn, S. H.; Choi, N. R.; Heo, J.; Kim, H.; Kwon, K.
H.; Lee, Y.-M.; Rhee, J.-S. Transcriptome Profiles of Daphnia magna
across to the Different Water Chemistry of Surface Water of the Korean
Demilitarized Zone. Toxicol. Environ. Health Sci. 2017,9(3), 188−198.
(39) Spurgeon, D. J.; Jones, O. A. H.; Dorne, J.-L. C. M.; Svendsen, C.;
Swain, S.; Sturzenbaum, S. R. Systems Toxicology Approaches for
Understanding the Joint Effects of Environmental Chemical Mixtures.
Sci. Total Environ. 2010,408 (18), 3725−3734.
(40) Altenburger, R.; Scholz, S.; Schmitt-Jansen, M.; Busch, W.;
Escher, B. I. Mixture Toxicity Revisited from a Toxicogenomic
Perspective. Environ. Sci. Technol. 2012,46 (5), 2508−2522.
(41) Hernández, A. F.; Gil, F.; Lacasana, M. Toxicological Interactions
of Pesticide Mixtures: An Update. Arch. Toxicol. 2017,91 (10), 3211−
3223.
(42) Ulrich-Merzenich, G.; Panek, D.; Zeitler, H.; Wagner, H.; Vetter,
H. New Perspectives for Synergy Research with the “Omic”-
Technologies. Phytomedicine 2009,16 (6−7), 495−508.
(43) Akkanen, J.; Kukkonen, J. V. K. Biotransformation and
Bioconcentration of Pyrene in Daphnia magna.Aquat. Toxicol. 2003,
64 (1), 53−61.
(44) Branco, V.; Matos, B.; Mourato, C.; Diniz, M.; Carvalho, C.;
Martins, M. Synthesis of Glutathione as a Central Aspect of PAH
Toxicity in Liver Cells: A Comparison between Phenanthrene,
Benzo[b]Fluoranthene and Their Mixtures. Ecotoxicol. Environ. Saf.
2021,208, No. 111637.
(45) Wan, X.; Cheng, C.; Gu, Y.; Shu, X.; Xie, L.; Zhao, Y. Acute and
Chronic Toxicity of Microcystin-LR and Phenanthrene Alone or in
Combination to the Cladoceran (Daphnia magna). Ecotoxicol. Environ.
Saf. 2021,220, No. 112405.
(46) Ewa, B.; Danuta, M.-S. Polycyclic Aromatic Hydrocarbons and
PAH-Related DNA Adducts. J. Appl. Genet. 2017,58 (3), 321−330.
(47) Li, F.; Xiang, B.; Jin, Y.; Li, C.; Ren, S.; Wu, Y.; Li, J.; Luo, Q.
Hepatotoxic Effects of Inhalation Exposure to Polycyclic Aromatic
Hydrocarbons on Lipid Metabolism of C57BL/6 Mice. Environ. Int.
2020,134, No. 105000.
(48) Gauthier, P. T.; Norwood, W. P.; Prepas, E. E.; Pyle, G. G.
Behavioural Alterations from Exposure to Cu, Phenanthrene, and Cu-
Phenanthrene Mixtures: Linking Behaviour to Acute Toxic Mecha-
nisms in the Aquatic Amphipod, Hyalella azteca.Aquat. Toxicol. 2016,
170, 377−383.
(49) Rodrigues, A. P.; Lehtonen, K. K.; Guilhermino, L.; Guimaraes,
L. Exposure of Carcinus maenas to Waterborne Fluoranthene:
Accumulation and Multibiomarker Responses. Sci. Total Environ.
2013,443, 454−463.
(50) Guo, B.; Feng, D.; Xu, Z.; Qi, P.; Yan, X. Acute Benzo[a]Pyrene
Exposure Induced Oxidative Stress, Neurotoxicity and Epigenetic
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
J

Change in Blood Clam Tegillarca granosa.Sci. Rep. 2021,11 (1),
No. 18744.
(51) Barata, C.; Calbet, A.; Saiz, E.; Ortiz, L.; Bayona, J. M. Predicting
Single and Mixture Toxicity of Petrogenic Polycyclic Aromatic
Hydrocarbons to the Copepod Oithona davisae.Environ. Toxicol.
Chem. 2005,24 (11), 2992−2999.
(52) Vandenbrouck, T.; Jones, O. A. H.; Dom, N.; Griffin, J. L.; De
Coen, W. Mixtures of Similarly Acting Compounds in Daphnia magna:
From Gene to Metabolite and Beyond. Environ. Int. 2010,36 (3), 254−
268.
(53) Varga, T.; Czimmerer, Z.; Nagy, L. PPARs Are a Unique Set of
Fatty Acid Regulated Transcription Factors Controlling Both Lipid
Metabolism and Inflammation. Biochim. Biophys. Acta, Mol. Basis Dis.
2011,1812 (8), 1007−1022.
(54) Zhang, X.; Ren, X.; Sun, W.; Griffin, N.; Wang, L.; Liu, H. PFOA
Exposure Induces Aberrant Glucose and Lipid Metabolism in the Rat
Liver through the AMPK/mTOR Pathway. Toxicology 2023,493,
No. 153551.
(55) Guruge, K. S.; Yeung, L. W. Y.; Yamanaka, N.; Miyazaki, S.; Lam,
P. K. S.; Giesy, J. P.; Jones, P. D.; Yamashita, N. Gene Expression
Profiles in Rat Liver Treated with Perfluorooctanoic Acid (PFOA).
Toxicol. Sci. 2006,89 (1), 93−107.
(56) Yang, H.-B.; Zhao, Y.-Z.; Tang, Y.; Gong, H.-Q.; Guo, F.; Sun,
W.-H.; Liu, S.-S.; Tan, H.; Chen, F. Antioxidant Defence System Is
Responsible for the Toxicological Interactions of Mixtures: A Case
Study on PFOS and PFOA in Daphnia magna.Sci. Total Environ. 2019,
667, 435−443.
(57) Li, K.; Gao, P.; Xiang, P.; Zhang, X.; Cui, X.; Ma, L. Q. Molecular
mechanisms of PFOA-induced toxicity in animals and humans:
Implications for health risks. Environ. Int. 2017,99, 43−54.
(58) Matsuda, K.; Buckingham, S. D.; Kleier, D.; Rauh, J. J.; Grauso,
M.; Sattelle, D. B. Neonicotinoids: Insecticides Acting on Insect
Nicotinic Acetylcholine Receptors. Trends Pharmacol. Sci. 2001,22
(11), 573−580.
(59) Tisler, T.; Jemec, A.; Mozetic, B.; Trebse, P. Hazard
Identification of Imidacloprid to Aquatic Environment. Chemosphere
2009,76 (7), 907−914.
(60) Erban, T.; Sopko, B.; Talacko, P.; Harant, K.; Kadlikova, K.;
Halesova, T.; Riddellova, K.; Pekas, A. Chronic Exposure of
Bumblebees to Neonicotinoid Imidacloprid Suppresses the Entire
Mevalonate Pathway and Fatty Acid Synthesis. J. Proteomics 2019,196,
69−80.
(61) Mahmoud, Y. K.; Ali, A. A.; Abdelrazek, H. M. A.; Aldayel, T. S.;
Abdel-Daim, M. M.; El-Menyawy, M. A. I. Neurotoxic Effect of Fipronil
in Male Wistar Rats: Ameliorative Effect of L-Arginine and L-Carnitine.
Biology 2021,10 (7), 682.
(62) Duong, T.-B.; Odhiambo, B. K.; Oldham, D.; Hoffman, M.;
Frankel, T. E. Acute Exposure to an Organochlorine Pesticide Alters
Locomotor Behavior and Cardiac Function in the Freshwater
Invertebrate Daphnia magna.Mar. Freshwater Behav. Physiol. 2021,
54 (2), 51−64.
(63) Bownik, A.; Szabelak, A. Short-Term Effects of Pesticide Fipronil
on Behavioral and Physiological Endpoints of Daphnia magna.Environ.
Sci. Pollut. Res. 2021,28 (25), 33254−33264.
(64) Weston, A.; Caminada, D.; Galicia, H.; Fent, K. Effects of Lipid-
Lowering Pharmaceuticals Bezafibrate and Clofibric Acid on Lipid
Metabolism in Fathead Minnow (Pimephales promelas). Environ.
Toxicol. Chem. 2009,28 (12), 2648−2655.
(65) Karamyan, V. T.; Speth, R. C. Enzymatic Pathways of the Brain
Renin−Angiotensin System: Unsolved Problems and Continuing
Challenges. Regul. Pept. 2007,143 (1−3), 15−27.
(66) Barreras, A.; Gurk-Turner, C. Angiotensin II Receptor Blockers.
Baylor Univ. Med. Cent. Proc. 2003,16 (1), 123−126.
(67) Kienzler, A.; Bopp, S. K.; van der Linden, S.; Berggren, E.; Worth,
A. Regulatory assessment of chemical mixtures: requirements, current
approaches and future perspectives. Regul. Toxicol. Pharmacol. 2016,80,
321−334.
(68) Bakker, R.; Ellers, J.; Roelofs, D.; Vooijs, R.; Dijkstra, T.; van
Gestel, C. A.; Hoedjes, K. M. Combining time-resolved transcriptomics
and proteomics data for Adverse Outcome Pathway refinement in
ecotoxicology. Sci. Total Environ. 2023,869, No. 161740.
(69) Foreman, A. L.; Warth, B.; Hessel, E. V.; Price, E. J.; Schymanski,
E. L.; Cantelli, G.; Parkinson, H.; Hecht, H.; Klánová, J.; Vlaanderen, J.;
Hilscherova, K.; et al. Adopting Mechanistic Molecular Biology
Approaches in Exposome Research for Causal Understanding. Environ.
Sci. Technol. 2024,58 (17), 7256−7269.
Environmental Science & Technology pubs.acs.org/est Article
https://doi.org/10.1021/acs.est.4c11095
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
K