Activation of aryl hydrocarbon receptor by dioxin directly shifts gut microbiota in zebrafish

Environmental Pollution 255 (2019) 113357

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Activation of aryl hydrocarbon receptor by dioxin directly shifts gut microbiota in zebrafish* Yumiao Sun a, 1, Lizhu Tang a, 1, Yang Liu b, Chenyan Hu c, Bingsheng Zhou a, Paul K.S. Lam d, James C.W. Lam d, e, Lianguo Chen a, * a

State Key Laboratory of Freshwater Ecology and Biotechnology, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan 430072, China College of Life Sciences, Henan Normal University, Xinxiang, Henan 453007, China School of Chemistry and Environmental Engineering, Wuhan Institute of Technology, Wuhan 430072, China d State Key Laboratory of Marine Pollution and Department of Chemistry, City University of Hong Kong, Kowloon, Hong Kong, China e Department of Science and Environmental Studies, The Education University of Hong Kong, 10 Lo Ping Road, Tai Po, New Territories, Hong Kong, China b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 July 2019 Received in revised form 12 September 2019 Accepted 5 October 2019 Available online 7 October 2019

Gut microbiota is of critical importance to host health. Aryl hydrocarbon receptor (AhR) is found to be closely involved in the regulation of gut microbial dynamics. However, it is still not clear how AhR signaling shapes the gut microbiota. In the present study, adult zebrafish were acutely exposed to an AhR antagonist (CH223191), an AhR agonist (polychlorinated biphenyl 126; PCB126) or their combination for 7 d. Overall intestinal health and gut microbial community were temporally monitored (1 d, 3 d and 7 d) and inter-compared among different groups. The results showed that single exposure to PCB126 significantly disrupted the overall health of intestines (i.e., neural signaling, inflammation, epithelial barrier integrity, oxidative stress). However, CH223191 failed to inhibit but enhanced the physiological toxicities of PCB126, implying the involvement of extra mechanisms rather than AhR in the regulation of intestinal physiological activities. Dysbiosis of gut microbiota was also caused by PCB126 over time as a function of sex. It is intriguing that CH223191 successfully abolished the holistic effects of dioxin on gut microbiota, which inferred that growth of gut microbes was directly controlled by AhR activation without the involvement of host feedback modulation. When coming to detailed alterations at certain taxon, both antagonistic and synergistic interactions existed between CH223191 and dioxin, depending on fish sex, exposure duration and bacterial species. Correlation analysis found that gut inflammation was positively associated with pathogenic Legionella bacteria, but was negatively associated with epithelial barrier integrity, suggesting that integral intestinal epithelial barrier can prevent the influx of pathogenic bacteria to induce inflammatory response. Overall, this study has deciphered, for the first time, the direct regulative effects of AhR activity on gut microbiota. Future research is warranted to elucidate the specific mechanisms of AhR action on certain bacterial population. © 2019 Elsevier Ltd. All rights reserved.

Keywords: Gut microbiota Aryl hydrocarbon receptor PCB126 CH223191 Zebrafish intestinal health

1. Introduction A large quantity and diversity of microbes commensally reside in the gut of animals, forming a complex ecosystem of microbiota. It is increasingly appreciated that gut microbial community plays important roles in the maintenance of host health (Holmes et al., 2011; Kinross et al., 2011). By posing direct or indirect effects

* This paper has been recommended for acceptance by Dr. Sarah Harmon. * Corresponding author. E-mail address: [email protected] (L. Chen). 1 Y.S. and L.T. contributed equally to this work.

https://doi.org/10.1016/j.envpol.2019.113357 0269-7491/© 2019 Elsevier Ltd. All rights reserved.

through microbiota-gut-brain axis, gut microbes can tightly regulate a battery of physiological activities of hosting organisms, such as neural signaling, energy metabolisms and immune functions (Tremaroli and Backhed, 2012). With regard to the intricate involvement of gut microbiota in host fitness, dysbiosis of gut microbial community composition will dysregulate the metabolic homeostasis, compromise the wellbeing of host animals and eventually cause the onset of multiple diseases in the long run, including obesity, diabetes, inflammatory symptom and neurological disorder (Mathis and Benoist, 2012; Snedeker and Hay, 2012; Tilg and Kaser, 2011). Furthermore, dynamics of gut microbiota are highly sensitive to the challenges of exogenous stressors, among which environmental pollutants are of special concern (Kan

2

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

et al., 2015; Jin et al., 2017). Albeit with distinct physicochemical properties, a variety of pollutants (e.g., persistent organic pollutants and antibiotics) can significantly shift the composition, abundance and diversity of gut microbial population. Aryl hydrocarbon receptor (AhR) is widely expressed in different cell types throughout the body (e.g., gut), which mediates the cellular response and toxicity to environmental xenobiotics (Mandal, 2005). The canonical acting model of AhR signaling starts from the binding of a ligand, which enables the translocation of AhR from cytoplasm into cellular nucleus. Then, AhR will heterodimerize with its nuclear translocator (ARNT) and bind to the DNA dioxin response element to activate the transcription of xenobiotic metabolizing genes, including Cytochrome P450 1A1 (CYP1a1) and CYP1b1 (Lamas et al., 2018). In addition to the modulation of xenobiotic metabolisms and toxicity, recent studies report the involvement of AhR signaling in the dysbiosis of gut microbiota by environmental pollutants. By comparing with AhR-deficient mice, Zhang et al. (2015) finds that dietary exposure to 2,3,7,8tetrachlorodibenzofuran (TCDF), a dioxin-like pollutant, modifies gut bacterial community composition, triggers intestinal inflammation and disturbs gut microbiota-host metabolic homeostasis in an AhR-dependent manner. Another AhR agonist, polychlorinated biphenyl 126 (PCB126) also abruptly shapes the structure of gut microbial community to an AhR-responsive profile after acute exposure of adult zebrafish, which is primarily contributed by the Aeromonas genus (Chen et al., 2018a). Therefore, the toxicological regulation of AhR signaling could be a novel route of environmental pollutants to shift gut microbiota, necessitating more mechanistic works. Although AhR pathway has been confirmed to modulate the dynamics of intestinal microbes, it is still unknown how AhR shifts gut microbiota and what are the direct targets of AhR among the complex of toxic responses in the intestines. In the present study, adult zebrafish (Danio rerio) were acutely exposed for 7 d to an AhR agonist (PCB126), an AhR antagonist (CH223191) or a combination of PCB126 and CH223191. During the exposure, a time-course observation was performed at 1 d, 3 d and 7 d, respectively. Intercomparison of adverse effects from different groups (agonist or/ and antagonist) can facilitate the filtering of molecular targets under direct regulation of AhR signaling. By comparing the changes at different time points, it is expected to further unveil how AhR signals gradually impact gut health and microbial population over time, thus identifying the earliest responsive events and the most sensitive indicators of AhR-mediated toxicity. 2. Materials and methods 2.1. Chemicals PCB126 (AhR agonist) at a purity of >99.0% was purchased from AccuStandard (New Haven, CT, USA). CH223191 (AhR antagonist) was obtained from Sigma-Aldrich Corp. (St. Louis, MO, USA; purity
98.0%). Stock solutions of PCB126 and CH223191 were prepared in dimethyl sulfoxide (DMSO; purity >99%; Sigma-Aldrich). All other chemicals used in the present study were of analytic grade. 2.2. Fish maintenance and exposure Adult zebrafish (D. rerio) of 4-month age were maintained in a semi-static system containing charcoal-filtered fully-aerated tap water according to previous instructions (Chen et al., 2018a). Ambient temperature was constantly controlled at 28 ± 0.5  C and the photoperiod was set as 14-h light: 10-h dark. The fish were fed twice daily with flake feed and newly-hatched Artemia nauplii until the fish stopped feeding. Approximately equal amount of feeds was

provided to the fish in all tanks. After acclimation for two weeks, adult zebrafish were randomly divided into four groups, including a control group, a CH223191 group (100 nM), a PCB126 group (1 mg/L, being equal to 3 nM) and a coexposure group (100 nM CH223191 and 1 mg/L PCB126). Equal amount of DMSO was received in every tank (v/v <0.001%). Kim et al. (2006) claims that 30 nM CH223191 can potently inhibit the activation of AhR by dioxins, while other in vitro incubations show that CH223191 up to 1 mM can effectively inhibit the activities of AhR (Attignon et al., 2017; Zhao et al., 2010). After full consideration about previous results and chemical availability for current waterborne exposure, we chose a medium concentration of CH223191 (i.e., 100 nM). PCB126 at 1 mg/L has been capable of significantly shifting zebrafish gut microbiota in an acute exposure (Chen et al., 2018a). Each group included three replicate tanks (n ¼ 3). Each tank contained 25-L exposure medium with 30 females and 30 males. The water was renewed daily to maintain appropriate chemical concentrations. After exposure for 1 d, 3 d and 7 d, approximately 9 female and 9 male zebrafish were randomly selected and dissected, respectively, at each time point to obtain intestine tissues, which were snap-frozen in liquid nitrogen and stored at 80  C for following molecular analyses. Zebrafish were not fed that morning before dissection. 2.3. Gene transcriptions Transcriptions of AhR-responsive genes (CYP1a1 and CYP1b1) in intestines were measured using quantitative real-time PCR assays according to previous descriptions (Chen et al., 2012). Three intestines of the same sex (male or female) were pooled together a replicate (n ¼ 3). Briefly, total RNA was extracted using TRIzol Reagent following manufacturer’s manual (Invitrogen, Carlsbad, CA, USA). RNA integrity and purity were examined by electrophoresis on 1% agarose gel and by readings on a Nanodrop 2000 platform (Thermo Scientific, DE, USA), respectively. First-strand cDNA was synthesized using commercial reverse transcription kits (Yeasen Biotech Co., Ltd., Shanghai, China). Gene transcriptions were analyzed using SYBR Green Kits (Yeasen Biotech Co., Ltd) on a CFX 384 Touch Real-Time PCR Detection System (Bio-Rad, Munich, Germany). Primers of CYP1a1 (Forward: AATCCCAGACGGGCTACA; Reverse: CCGGGCCATAGCACTTAC) and CYP1b1 (Forward: GCTCAGCTGGTCCATTGATACC; Reverse: CATCAGCGACAGCAACA€ nsson et al., CAC) were adopted from previous publications (Jo 2007; Blüthgen et al., 2012). The housekeeping gene ribosomal protein L8 (rpl8; Forward: TTGTTGGTGTTGTTGCTGGT; Reverse: GGATGCTCAACAGGGTTCAT) was used as the internal gene, which transcription did not vary during exposure. Transcriptional levels of target genes were normalized against that of rpl8 using the 2eDDT method. 2.4. Physiological measurement of the intestines After exposure at different duration (1 d, 3 d and 7 d), holistic health of intestine tissues was evaluated by a suite of physiological markers, including serotonin for neural signal transduction, tight junction protein 2 (TJP2) for intestinal epithelial barrier permeability, interleukin 1b (IL1b) for inflammatory responses, and reactive oxygen species (ROS) for oxidative stress (Chen et al., 2018b). Three intestines of the same sex were pooled together as a biological replicate (n ¼ 3). The intestines were homogenized in 0.5 mL normal saline (0.9% sodium chloride) on ice using a tissue tearer (BioSpec Products, Bartlesville, OK, USA). After centrifuging at 6000g for 10 min at 4  C, the supernatant was transferred to a new tube for further physiological measurements. Serotonin concentrations, indicative of intestinal neural

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

signaling, were measured using a commercial enzyme-linked immunosorbent assay (ELISA) kit for fish according to the manual (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Protein expressions of TJP2 and IL1b, which indicate intestinal epithelial barrier integrity and inflammatory responses, respectively, were also examined using fish-specific ELISA assay kits (Mybiosource, San Diego, CA). Concentrations of intestinal ROS were measured using 20 ,70 dichlorofluorescein diacetate (DCHF-DA) to indicate oxidative stress (e.g., hydrogen peroxide, peroxynitrite and hydroxyl radicals) based on the manual of a fluorimetric kit (Nanjing Jiancheng Bioengineering Institute). Briefly, 20-mL aliquots of intestinal tissue supernatant were incubated with 180 mL DCHF-DA (100 mM) at room temperature in a 96-well plate. After a 30-min reaction in the dark, the fluorescence intensity was measured using a microplate reader SpectraMax i3x (Molecular Devices, Silicon Valley, CA), with excitation at 502 nm and emission at 530 nm. The ROS levels were expressed in arbitrary units as DCF generations per milligram of protein. 2.5. 16S rRNA amplicon sequencing and bioinformatic analyses Three intestines of the same sex were pooled together as a replicate (n ¼ 3). Genomic DNA of whole intestine tissues was extracted using a DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). Quality of DNA extracts was checked by a NanoDrop 2000 (Thermo Scientific) and by agarose gel electrophoresis. Then, concentrations of genomic DNA were measured using a Qubit fluorometer (Thermo Scientific). An aliquot of 30 ng DNA sample was used for the amplification of 16S rRNA using the primer pair 341F (50 -ACTCCTACGGGAGGCAGCAG-30 ) and 806R (50 -GGACTACNNGGGTATCTAAT-30 ), targeting the V3V4 hypervariable regions. The concentrations of DNA libraries were quantified by Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). DNA libraries were then sequenced on the Illumina HiSeq 2500 platform to generate paired-end raw reads (i.e., 2  150 bp). After removing low-quality sequences, the clean reads were merged by use of FALSH software to obtain clean tags, which were then clustered by UPARSE pipeline to generate Operational Taxonomic Units (OTUs) according to a cutoff of 97% similarity. Chimera sequences were removed by use of UCHIME. Taxonomic annotation of OTU representative sequences was performed using RDP classifier (version 2.2) based on the Green Genes Database. Relative abundances (%) of each bacterial taxon (e.g., phylum or genus) were calculated by dividing respective count of sequences against total count of sequences in each sample. 2.6. Statistical analyses Values in the present study were represented as mean ± SEM of three replicates. Normality of data and homogeneity of variances were evaluated by the ShapiroWilk test and Levene’s test, respectively. Significant difference among groups was determined by one-way analysis of variance (ANOVA), followed by post-hoc LSD test. Data were log-transformed if necessary. Kruskal-Wallis ANOVA test with Dunn-Bonferroni post-hoc comparison would be used alternatively to determine statistical significance when normal distribution or homogenous variance could not be met. All statistical analyses were performed on SPSS software version 22.0 (IBM SPSS Statistics, IBM Corporation, Armonk, New York). A Pvalue <0.05 was set as the criterion of significant difference. Only representative bacterial genera of relative abundance >1% in at least one replicate sample were used in following bioinformatic analyses. Hierarchical clustering analysis of

3

representative genera was performed using Gene Cluster 3.0 software. Spearman Rank Correlation was chosen to calculate the similarity metric among genera with centroid linkage as the clustering method. Principal component analysis (PCA) was conducted with the input of relative abundances of representative genera (>1%) according to the variance-covariance matrix in PAST software (Chen et al., 2018a). Pearson correlation analysis between representative bacterial genera (>1%) and physiological indices was performed on SPSS software version 22.0. A co-occurrence network of representative abundant genera was constructed on the RStudio platform (R 2.13; The R Foundation for Statistical Computing, Vienna, Austria) based on the pearson correlation results (R > 0.5 or < -0.5; P < 0.001). The diagram was then visualized by the interactive platform Gephi (WebAtlas, Paris, France). 3. Results and discussion 3.1. Impairment of zebrafish intestinal health Gut environment is a susceptible target organ of environmental pollutants. Considering the enterohepatic cycle of PCB126, a continual impact on gut health and microbiota is expected (Petriello et al., 2018). In both male and female intestines of zebrafish, PCB126 exposure remarkably and significantly induced the gene transcriptions of CYP1a1 (Fig. 1A and Fig. 2A, respectively) and CYP1b1 to a lesser extent (Figs. 1B and 2B, respectively) from 1 d until the conclusion of exposure. It is well known that PCB126 is a coplanar dioxin-like congener, having a high affinity to bind and activate AhR signaling (Hennig et al., 2002). Expressions of CYP1a1 and CYP1b1 are under target regulation of AhR activity, although CYP1a1 shows a higher inducibility than CYP1b1 (Lin et al., 2003). CH223191 is proposed as a potent and specific antagonist of AhR, which can efficiently prevent dioxin-elicited CYP induction in vitro at 30 nM (Kim et al., 2006). However, in the present study, the presence of 100 nM CH223191 during combined exposure had almost no antagonistic effects on PCB126-induced CYP transcriptions, except a mild significant decrease of 1.4-fold in male CYP1a1 gene at 7 d relative to PCB126 single exposure (Fig. 1A). This difference between in vivo and in vitro outcomes implicates the complexity of in vivo environments, which may require a higher concentration of CH223191 to antagonize AhR activity. In addition, acute exposure to PCB126 led to significant alterations in intestinal health of male (Fig. 1) and female zebrafish (Fig. 2). As early as 1-d exposure, systematic impairment of intestinal health as characterized by various physiological markers has been recorded here, further highlighting intestine as a susceptible organ to environmental pollutants and the sensitivity of physiological indicators to reflect adverse effects (Choi et al., 2010; Chen et al., 2018c; Chi et al., 2019). Serotonin is a key neurotransmitter in gut lumen along microbiota-gut-brain axis, posing critical influences on bowel movement and microbial balance (O’Mahony et al., 2015). Compared to control group, PCB126 exposure significantly increased serotonin concentrations in male intestines at 1 d (Fig. 1C), but significantly decreased serotonin concentrations in female intestines at 7 d (Fig. 2C). Depending on sex, addition of CH223191 in exposure media of the combined group (PCB þ CH) also showed contrasting effects on serotonin compared to dioxin single exposure, which was able to successfully antagonize PCB126 effects in male at 1 d (Fig. 1C), but unexpectedly enhance PCB126 effects in female at 1 d and 3 d (Fig. 2C). This indicates that serotonin is a potential target molecule of AhR activation. Furthermore, sex-specific responses to environmental pollutants are frequently documented, which are presumably attributed to the inherent distinction of each sex on innate hormonal levels and detoxifying capacity (Chen et al., 2018a). Due to the high levels of sex hormones

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

400 300 200

70

c

b

b b

100

c

a 12 8 4 0

a b

a a 1d

a

3d

7d

(C) Serotonin

16

12

8

4

0

2.4

a a

a 1d

3d

b

b b

b

20 10

a a

a a 1d

a a

3d

(D) TJP2

7d

b

250 200 150 100 50 0

b

ab a

a

b

ab a

a 1d

(F) ROS

a

1.2

ab b b

0.8 0.4

1d

3d

7d

80

3d

a

7d

a b

60

b b

b

100

1.6

0.0

30

120

2.0

b

40

300

7d

(E) IL1

b

50

350

b

(B) CYP1b1

60

0

Intestinal levels (pg/mg protein)

Intestinal levels (ng/mg protein)

c

Control CH223191 PCB126 PCB+CH

500

20

Intestinal levels (pg/mg protein)

(A) CYP1a1

Intestinal levels relative to control (%)

Fold change relative to control

700 600

Fold change relative to control

4

abab b

b b

40 20 0

1d

3d

7d

Fig. 1. Alterations in intestinal health of male zebrafish after 1 d, 3 d and 7 d exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126). (A) CYP1a1 gene transcription; (B) CYP1b1 gene transcription; (C) Intestinal serotonin concentrations; (D) Intestinal TJP2 protein expression; (E) Intestinal IL1b levels; (F) Intestinal ROS levels. Values represent the mean ± SEM of three replicates (n ¼ 3). Different characters (a, b, c) indicate significant difference among exposure groups.

in females, crosstalk between AhR and estrogen signals is found to differentially regulate the activation of detoxification system of each sex, thus accounting for sex-specific toxicities (Marques et al., 2013). In male intestines, acute PCB126 exposure significantly increased the protein abundances of TJP2 at 3 d and 7 d (Fig. 1D). After single exposure to CH223191 or PCB126 until 7 d, TJP2 expression was also significantly up-regulated in females (Fig. 2D). Consistent to current results, AhR activation by ligands has been previously shown to increase the expressions of TJPs and improve the epithelial barrier integrity (Bansal et al., 2010; Han et al., 2016). A compensatory mechanism may be involved to repair intestinal injury caused by PCB exposure (Petriello et al., 2018). In addition to the rectification of epithelial barrier permeability, activation of AhR signaling can concurrently reduce the inflammation in intestines (Bansal et al., 2010), which is supported by the decreased levels of IL1b (Figs. 1E and 2E) and ROS (Fig. 1F) in PCB126 exposure group of

this study. Intestinal epithelial cell layer provides a protective interface between external luminal environment and internal host body. The increased integrity of epithelial barrier will decrease permeability and prevent the influx of microbial products into lamina propria, thus reducing the interaction with immune cells and production of pro-inflammatory cytokines (Salim et al., 2014). Although CH223191 is proposed as a potent inhibitor of AhR activity, the unexpected enhancement of dioxin toxicity by CH223191 was observed over time to physiological indices (i.e., TJP2, IL1b and ROS) in both sexes, implying the association of extra signaling pathway with the modulation of these biochemicals rather than AhR. Physiological events related to epithelial integrity, inflammation and oxidative stress were thus eliminated from the direct molecular targets of AhR activity in the complex network of intestinal toxicology. As stated previously, the only exception appeared to be serotonin in male intestines (Fig. 1C), further highlighting serotonin as a direct target of AhR regulation.

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

200

b b

b

100

b a a

1 0

70

Intestinal levels (ng/mg protein)

70

b

a

1d

a

a a

3d

(C) Serotonin

7d

60

a

50 40 30 20 10

a ab

0

2.4

ab

b

1d

ab ab b 3d

b

b b

7d

(E) IL1

2.0

a

0.4

1d

3d

7d

b b

b b

20 10

a a

a a

1d

a a

3d

(D) TJP2

7d

b ab

b

90

b b

ab ab

60

a

a

b

a

30

a 0

120

b b

0.8

0.0

30

1d

3d

7d

(F) ROS

160

1.6 1.2

b

40

200

a

b

50

120

a

(B) CYP1b1

60

0

Intestinal levels (pg/mg protein)

300

Intestinal levels (pg/mg protein)

b

Control CH223191 PCB126 PCB+CH

Fold change relative to control

(A) CYP1a1

Intestinal levels relative to control (%)

Fold change relative to control

400

5

b

ab a

b a

ab

ab a

80

40

0

1d

3d

7d

Fig. 2. Alterations in intestinal health of female zebrafish after 1 d, 3 d and 7 d exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126). (A) CYP1a1 gene transcription; (B) CYP1b1 gene transcription; (C) Intestinal serotonin concentrations; (D) Intestinal TJP2 protein expression; (E) Intestinal IL1b levels; (F) Intestinal ROS levels. Values represent the mean ± SEM of three replicates (n ¼ 3). Different characters (a, b) indicate significant difference among exposure groups.

3.2. Dysbiosis of gut microbiota Temporal changes in gut microbial population of zebrafish were also monitored in this study after 1-d, 3-d and 7-d exposure to CH223191, PCB126 or their combination. Parameters of alphadiversity were calculated to reflect the overall status of bacterial community diversity and richness (Table 1). After exposure till 7 d, the alpha-diversity results showed that PCB126 single exposure significantly disturbed the gut microbial community composition in abundance and diversity. Inverse changing trend of Shannon and Simpson indexes was noted for male and female fish, pointing to sex-specific responses of gut microbiome to AhR activation (Chen et al., 2018a). In contrast to general enhancement of PCB126 toxicities by CH223191 on intestinal physiological health of zebrafish host (e.g., inflammation, epithelial integrity and oxidative stress), comparing the microbial profile of combined PCB þ CH group with that of PCB126 single exposure group found that CH223191 effectively and completely abolished the effects of dioxin on gut

microbiota (Table 1). This distinction in changing trend of host intestinal health and intestinal microbiota indicated that gut microbial population may be directly controlled and directed by AhR signaling without the involvement of host feedback regulation. Previous researches find that dioxin-like pollutants can alter gut microbial population in an AhR-dependent manner (Zhang et al., 2015; Chen et al., 2018a). However, it is still unclear whether gut microbial dysbiosis is directly caused by dioxins or due to indirect regulation of impaired intestinal health (e.g., inflammation and oxidative stress). For the first time, current results deciphered gut microbes as the direct targets of AhR regulation in the complexity of intestinal responses, although the specific roles of AhR on bacterial growth and reproduction remain to be clarified. Taxonomic annotation showed detailed changes in gut bacterial community at different levels (e.g., phylum and genus). Proteobacteria, Actinobacteria, Planctomycetes, Chlamydiae and Fusobacteria were dominant phyla in male and female intestines (Fig. 3A and Fig. 4A). Compared to control group, acute exposure to

171±11 193±1 159±6 161±13 142±9 197±15 187±17 184±34 143±9 203±16 191±18 194±31 c

d

Different characters indicate significant difference among exposure groups. Values represent the mean±SEM of three replicates. Indicative of bacterial community diversity. Values of Simpson are inversely proportional to bacterial diversity. Indicative of bacterial community richness.

a a

b

2.9±0.3 2.2±0.5 3.4±0.4 2.9±0.2 164±11 180±1 142±10 151±16 136±9 190±13 177±15 174±34 207±15 168±35 187±22 164±30

1d

3.0±0.2 2.6±0.5 2.1±0.4 2.1±0.3 206±9 a 182±19 a 108±13 b 223±53 a 223±8 191±33 165±21 190±4

7d 3d 1d

140±19 181±16 176±25 176±24

Control CH223191 PCB126 PCBþCH Male Control CH223191 PCB126 PCBþCH

a,b

b

212±15 180±36 193±22 179±35 ab

a

0.39±0.13 0.23±0.10 0.07±0.01 0.10±0.04 0.15±0.06 0.18±0.09 0.12±0.02 0.12±0.02 0.16±0.05 0.29±0.10 0.08±0.03 0.14±0.02 ab

a

1.8±0.3 2.4±0.3 3.4±0.0 3.2±0.4 2.8±0.5 3.0±0.6 2.7±0.1 2.9±0.2

a a

179±14 200±4 160±8 158±14

a

b

a ab

b b

a

7d

211±15 178±34 193±23 174±30

233±4 204±33 186±16 201±4

3d 7d 3d

233±5 214±40 189±11 211±8 159±15 201±25 193±25 195±20 0.20±0.05 0.22±0.10 0.35±0.23 0.11±0.03 0.11±0.01 0.23±0.08 0.33±0.12 0.24±0.08

3d 7d 3d

2.6±0.3 2.4±0.4 2.0±0.7 3.0±0.2

3.3±0.1 2.5±0.4 0.9±0.4 2.9±0.7

a

1d

0.09±0.01 0.25±0.08 0.69±0.16 0.19±0.12

a

1d

d

Chao1 c

Simpson c

Shannon Observed species Female

Table 1 Alpha diversity in zebrafish intestines after 1 d, 3 d and 7 d exposure to control, CH223191, PCB126 and PCBþCH (a combination of CH223191 and PCB126).

217±11 191±19 124±15 244±46

a

1d

d

ACE

154±10 195±21 190±25 194±18

7d

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

212±8 a 190±19 a 119±7 b 238±48 a

6

single PCB126 significantly increased the relative abundances of Planctomycetes phylum at 3 d (Fig. 3B), but significantly decreased the relative abundances of Fusobacteria at 3 d (Fig. 3C), Actinobacteria at 7 d (Fig. 3D) and Firmicutes at 1 d (Fig. 3E) in male intestines. In female intestines relative to the control, PCB126 exposure significantly decreased the relative abundances of phyla Bacteroidetes at 3 d (Fig. 4C) and Actinobacteria at 7 d (Fig. 4D). Previous time-course study using mice also observes significant time-dependent decreases in abundances of Bacteroidetes and Firmicutes (Petriello et al., 2018). 2013. Bacteroidetes and Firmicutes are well-known to provide key functions during lipid metabolism of hosting animals (Kan et al., 2015). In addition, CH223191 coexposure demonstrated sex-, species- and time-specific effects on gut microbiota compared to PCB126 alone. Both antagonistic and synergistic actions of CH223191 were observed on dioxin toxicities. At the genus level, Hyphomicrobium, Cetobacterium, Clavibacter and Rhodobacter were the abundant genera in most zebrafish intestines (Fig. 5). In male fish (Fig. 5A), PCB126 acute exposure significantly decreased the relative abundances of Propionibacterium at 1 d, Cetobacterium at 3 d, Shewanella at 3 d and Flavobacterium at 7 d, but significantly increased the abundances of genera Clavibacter at 3 d and Hyphomicrobium at 7 d. Belonging to Fusobacteria phylum, the abundance of Cetobacterium genus is relevant to fermentative metabolism of peptides and carbohydrates and production of vitamin B12 (Larsen et al., 2014), while Hyphomicrobium genus bacteria of Proteobacteria phylum perform key functions during denitrification process (Martineau et al., 2015). In female intestines (Fig. 5B), consistent decreases were observed in the relative genera abundances of Propionibacterium, Acinetobacter, Sphingomonas and Paracoccus. Propionibacterium bacteria are named for their unique ability to synthesize short chain fatty acids, in particular propionic acid (Brüggemann et al., 2004). Bacteria of Sphingomonas genus completely lacks lipopolysaccharides in their cell envelopes and are able to degrade polycyclic aromatic hydrocarbons (Balkwill et al., 2003). Similar to phylum level, CH223191 exerted varying interactions with dioxin toxicities depending on exposure duration, genus and sex, showing both antagonistic (e.g., female Propionibacterium, Acinetobacter, Sphingomonas and Paracoccus at 7 d) and synergistic modes (e.g., male Sphingomonas at 3 d and Kaistobacter at 7 d). The exact regulatory mechanisms of AhR pathway on the growth and reproduction of specific bacteria await future research to clarify. Whether intestinal bacteria can express homologue proteins of AhR functions is also an intriguing topic to resolve. With the input of relative abundances of representative genera (>1%), PCA plots were constructed for male and female intestines at various sampling points (1 d, 3 d and 7 d). Based on the distribution of samples on the plot, PCA can differentiate the interactive modes of AhR antagonist and agonist. In male intestines (Fig. 6), PCB126 exposure caused a deviation of gut microbiota from control group at 3 d (Figs. 6B) and 7 d (Fig. 6C), while the addition of CH223191 restored the control profile of gut microbiota in coexposure group at 7 d (Fig. 6C). In female intestines (Fig. 7), similar distributive patterns were also noted at 3 d (Fig. 7B) and d (Fig. 7C). Therefore, PCA plots provided additional evidences about the direct modulation of gut microbiota by AhR signals when considering gut microbiota as a whole. 3.3. Correlation between gut microbiota and fish health Based on relative abundances of representative genera, construction of co-occurrence network showed that gut microbial populations were vigorously interacted with each other only in a positive way, probably indicating a self restructure of microbiota

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

(A)

1d

Phylum relative abundance (%)

100

7

7d

3d

Proteobacteria Actinobacteria Planctomycetes Firmicutes Chlamydiae Fusobacteria Bacteroidetes Cyanobacteria Tenericutes TM6 Acidobacteria Chloroflexi TM7 Others

80

60

40

20

C on tro C l H 22 31 9 PC 1 B1 2 PC 6 B+ C H C on tro C l H 22 31 9 PC 1 B1 2 PC 6 B+ C H C on tro C l H 22 31 9 PC 1 B1 2 PC 6 B+ C H

0

40

b

30

ab

20 10 0

ab

a

a 1d

ab ab

3d

b

7d

(D) Actinobacteria

a 20

ab

15

ab

10

b

5

1d

3d

7d

(C) Fusobacteria

a

60 50 40 30

b

20

b b

10 0

15

25

0

Phylum relative abundance (%)

70

Control CH223191 PCB126 PCB+CH

50

30

Phylum relative abundance (%)

(B) Planctomycetes

Phylum relative abundance (%)

Phylum relative abundance (%)

60

1d

3d

7d

3d

7d

(E) Firmicutes

12

9

6

a 3

0

a

b ab 1d

Fig. 3. Changes in gut microbial phyla of male zebrafish after 1 d, 3 d and 7 d exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126). (A) Composition profile of gut microbial community based on representative phyla (>1% relative abundance); (B) Planctomycetes relative abundance; (C) Fusobacteria relative abundance; (D) Actinobacteria relative abundance; (E) Firmicutes relative abundance. Values represent the mean ± SEM of three replicates (n ¼ 3). Different characters (a, b) indicate significant difference among exposure groups.

upon AhR regulation (Fig. 8A). Correlation analysis found few significant association of gut microbiota with host physiological indices, further implying the limited effects of intestinal health on gut microbes. It is reasonable to speculate that shift in gut microbial

community could be mainly due to the direct modulation of AhR pathway. The only exception was the significant positive correlation between Legionella genus abundance and intestinal IL1b concentration (Fig. 8B). Legionella bacteria include several pathogenic

8

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

(A)

1d

Phylum relative abundance (%)

100

7d

3d

Proteobacteria Actinobacteria Planctomycetes Firmicutes Chlamydiae Fusobacteria Bacteroidetes Cyanobacteria Tenericutes TM6 Acidobacteria Chloroflexi TM7 Others

80

60

40

20

0

l l l 1 1 6 6 H H H 91 26 tro tro tro C C 19 B12 19 B12 n n n + +C + 3 3 31 B1 o o o B B B 2 2 2 C C C 2 2 2 PC PC PC PC PC PC H H H C C C Control CH223191 PCB126 PCB+CH

40 30 20 10

a

a

a ab

a 0

1d

3d

(D) Actinobacteria

b ab

a

b

b

10

5

b

0

1d

3d

6

4

2

0

7

20

7d

(C) Bacteroidetes

8

7d

25

15

Phylum relative abundance (%)

10

50

30

Phylum relative abundance (%)

(B) Chlamydiae b

Phylum relative abundance (%)

Phylum relative abundance (%)

60

ab a ab

a b

1d

ab

b

ab

3d

7d

(E) Firmicutes

6

b

5 4 3

ab

2

a

1

a 0

1d

3d

7d

Fig. 4. Changes in gut microbial phyla of female zebrafish after 1 d, 3 d and 7 d exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126). (A) Composition profile of gut microbial community based on representative phyla (>1% relative abundance); (B) Chlamydiae relative abundance; (C) Bacteroidetes relative abundance; (D) Actinobacteria relative abundance; (E) Firmicutes relative abundance. Values represent the mean ± SEM of three replicates (n ¼ 3). Different characters (a, b) indicate significant difference among exposure groups.

species that can cause severe infection and disease, with destructive inflammation occurring concomitantly (Bhopal, 1995). Because

the intestinal epithelial barrier will prevent the influx of bacterial products and subsequently reduce the production of pro-

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

9

Fig. 5. Hierarchical clustering analysis based on representative genera (>1% relative abundance) in zebrafish intestines (A, male; B, female) after 1 d, 3 d and 7 d exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126). Different characters (a, b, c) indicate significant difference among exposure groups. Red intensity is proportional to genus relative abundance.

inflammatory cytokines (Salim et al., 2014), a significant negative correlation was observed for TJP2 and IL1b (Fig. 8B). Gene transcriptions of CYP1a1 were significantly positively correlated with TJP2 protein expression (Fig. 8C), which is consistent with previous findings about the up-regulation of TJPs by AhR activation (Lamas et al., 2018). 4. Conclusion This study investigated the temporal changes (i.e., 1 d, 3 d and 7 d) in zebrafish intestinal health and microbiota after acute exposure to CH223191 (an AhR antagonist), PCB126 (an AhR

agonist) or their combination. Regarding the overall intestinal health, CH223191 failed to inhibit but enhanced physiological disturbances by dioxin in characteristics of epithelial barrier integrity, inflammation and oxidative stress. The only exception among physiological indices was serotonin, which may be an early target under direct regulation of AhR activity. Dynamics of gut microbial community were also significantly disrupted by PCB126 single exposure compared to the control, while the presence of CH223191 effectively abolished the holistic effects of PCB126 on gut microbiota. This discerned gut microbes as direct targets of AhR signals without the involvement of host feedback influences. When coming to specific phylum or genus, varying interactive modes of

10

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

Fig. 6. Principal component analysis based on the relative abundances (>1%) of genera in male intestines after 1 d (A), 3 d (B) and 7 d (C) exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126).

CH223191 were noted with dioxin, including both antagonistic and synergistic models depending on sex, exposure duration and bacterial species. Overall, the present study deciphered gut microbiota from intestinal complex toxic network and inferred gut microbiota

Fig. 7. Principal component analysis based on the relative abundances (>1%) of genera in female intestines after 1 d (A), 3 d (B) and 7 d (C) exposure to control, CH223191, PCB126 and PCB þ CH (a combination of CH223191 and PCB126).

as a direct target of AhR regulation. Future mechanistic studies are warranted to clarify the specific roles of AhR pathway in microbial growth and reproduction.

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

11

h

Fig. 8. Correlation analysis between gut microbiota and zebrafish physiological indices. (A) Bacterial co-occurrence network based on representative genera (relative abundance >1%) and pearson correlation screening (R > 0.5 or < -0.5; P < 0.001). Circular area represents genus relative abundance and color labeling represents the belonging phylum. Red lines indicate significant positive correlations among genera. (B) Significant correlation of IL1b intestinal levels with TJP2 protein expression (negative) and Legionella abundance (positive). (C) Significant positive correlation between CYP1a1 transcription and TJP2 protein expression in zebrafish intestines.

Declaration of competing interest The authors report no conflicts of interest. Acknowledgements This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB14040103), and State Key Laboratory of Freshwater Ecology and Biotechnology (2019FBZ03). References Attignon, E.A., Distel, E., Le-Grand, B., Leblanc, A.F., Barouki, R., de Oliveira, E., Aggerbeck, M., Blanc, E.B., 2017. Down-regulation of the expression of alcohol dehydrogenase 4 and CYP2E1 by the combination of a-endosulfan and dioxin in HepaRG human cells. Toxicol. In Vitro 45, 309e317. Balkwill, D.L., Fredrickson, J.K., Romine, M.F., 2003. Sphingomonas and related genera. In: Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.H., Stackebrandt, E. (Eds.), The Prokaryotes: A Handbook on the Biology of Bacteria, third ed., vol. 7. Springer, New York, pp. 605e629. Bansal, T., Alaniz, R.C., Wood, T.K., Jayaraman, A., 2010. The bacterial signal indole increases epithelial-cell tight-junction resistance and attenuates indicators of

inflammation. Proc. Natl. Acad. Sci. U.S.A. 107, 228e233. Bhopal, R., 1995. Source of infection for sporadic Legionnaires’ disease: a review. J. Infect. 30, 9e12. Blüthgen, N., Zucchi, S., Fent, K., 2012. Effects of the UV filter benzophenone-3 (oxybenzone) at low concentrations in zebrafish (Danio rerio). Toxicol. Appl. Pharmacol. 263, 184e194. Brüggemann, H., Henne, A., Hoster, F., Liesegang, H., Wiezer, A., Strittmatter, A., Hujer, S., Dürre, P., Gottschalk, G., 2004. The complete genome sequence of Propionibacterium acnes, a commensal of human skin. Science 305, 671e673. Chen, L., Hu, C., Lai, N.L.S., Zhang, W., Hua, J., Lam, P.K.S., Lam, J.C.W., Zhou, B., 2018c. Acute exposure to PBDEs at an environmentally realistic concentration causes abrupt changes in the gut microbiota and host health of zebrafish. Environ. Pollut. 240, 17e26. Chen, L., Lam, J.C.W., Hu, C., Tsui, M.M.P., Wang, Q., Giesy, J.P., Lam, P.K.S., 2018b. Perfluorobutanesulfonate exposure causes durable and transgenerational dysbiosis of gut microbiota in marine medaka. Environ. Sci. Technol. Lett. 5, 731e738. Chen, L., Yu, K., Huang, C., Yu, L., Zhu, B., Lam, P.K.S., Lam, J.C.W., Zhou, B., 2012. Prenatal transfer of polybrominated diphenyl ethers (PBDEs) results in developmental neurotoxicity in zebrafish larvae. Environ. Sci. Technol. 46, 9727e9734. Chen, L., Zhang, W., Hua, J., Hu, C., Lai, N.L.S., Qian, P.Y., Lam, P.K.S., Lam, J.C.W., Zhou, B., 2018a. Dysregulation of intestinal health by environmental pollutants: involvement of the estrogen receptor and aryl hydrocarbon receptor. Environ. Sci. Technol. 52, 2323e2330. Chi, Y., Lin, Y., Lu, Y., Huang, Q., Ye, G., Dong, S., 2019. Gut microbiota dysbiosis

12

Y. Sun et al. / Environmental Pollution 255 (2019) 113357

correlates with a low-dose PCB126-induced dyslipidemia and non-alcoholic fatty liver disease. Sci. Total Environ. 653, 274e282. Choi, Y.J., Seelbach, M.J., Pu, H., Eum, S.Y., Chen, L., Zhang, B., Hennig, B., Toborek, M., 2010. Polychlorinated biphenyls disrupt intestinal integrity via NADPH oxidaseinduced alterations of tight junction protein expression. Environ. Health Perspect. 118, 976e981. Han, B., Sheng, B., Zhang, Z., Pu, A., Yin, J., Wang, Q., Yang, K., Sun, L., Yu, M., Qiu, Y., Xiao, W., Yang, H., 2016. Aryl hydrocarbon receptor activation in intestinal obstruction ameliorates intestinal barrier dysfunction via suppression of MLCKMLC phosphorylation pathway. Shock 46, 319e328. Hennig, B., Meeraranj, P., Slim, R., Toborek, M., Daugherty, A., Silverstone, A.E., Robertson, L.W., 2002. Proinflammatory properties of coplanar PCBs: in vitro and in vivo evidence. Toxicol. Appl. Pharmacol. 181, 174e183. Holmes, E., Li, J.V., Athanasiou, T., Ashrafian, H., Nicholson, J.K., 2011. Understanding the role of gut microbiome-host metabolic signal disruption in health and disease. Trends Microbiol. 19, 349e359. Jin, Y., Wu, S., Zeng, Z., Fu, Z., 2017. Effects of environmental pollutants on gut microbiota. Environ. Pollut. 222, 1e9. €nsson, M.E., Orrego, R., Woodin, B.R., Goldstone, J.V., Stegeman, J.J., 2007. Basal and Jo 3,30 ,4,40 ,5-pentachlorobiphenyl-induced expression of cytochrome P450 1A, 1B and 1C genes in zebrafish. Toxicol. Appl. Pharmacol. 221, 29e41. Kan, H., Zhao, F., Zhang, X.X., Ren, H., Gao, S., 2015. Correlations of gut microbial community shift with hepatic damage and growth inhibition of Carassius auratus induced by pentachlorophenol exposure. Environ. Sci. Technol. 49, 11894e11902. Kim, S.H., Henry, E.C., Kim, D.K., Kim, Y.H., Shin, K.J., Han, M.S., Lee, T.G., Kang, J.K., Gasiewicz, T.A., Ryu, S.H., Suh, P.G., 2006. Novel compound 2-methyl-2H-pyrazole-3-carboxylic acid (2-methyl-4-o-tolylazo-phenyl)-amide (CH-223191) prevents 2,3,7,8-TCDD-induced toxicity by antagonizing the aryl hydrocarbon receptor. Mol. Pharmacol. 69, 1871e1878. Kinross, J.M., Darzi, A.W., Nicholson, J.K., 2011. Gut microbiome-host interactions in health and disease. Genome Med. 3, 14. Lamas, B., Natividad, J.M., Sokol, H., 2018. Aryl hydrocarbon receptor and intestinal immunity. Mucosal Immunol. 11, 1024e1038. Larsen, A.M., Mohammed, H.H., Arias, C.R., 2014. Characterization of the gut microbiota of three commercially valuable warmwater fish species. J. Appl. Microbiol. 116, 1396e1404. Lin, P., Hu, S.W., Chang, T.H., 2003. Correlation between gene expression of aryl

hydrocarbon receptor (AhR), hydrocarbon receptor nuclear translocator (ARNT), cytochromes P4501a1 (CYP1a1) and 1b1 (CYP1b1), and inducibility of CYP1a1 and CYP1b1 in human lymphocytes. Toxicol. Sci. 71, 20e26. Mandal, P.K., 2005. Dioxin: a review of its environmental effects and its aryl hydrocarbon receptor biology. J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 175, 221e230. Marques, M., Laflamme, L., Gaudreau, L., 2013. Estrogen receptor alpha can selectively repress dioxin receptor-mediated gene expression by targeting DNA methylation. Nucleic Acids Res. 41, 8094e8106. Martineau, C., Mauffrey, F., Villemur, R., 2015. Comparative analysis of denitrifying activities of Hyphomicrobium nitrativorans, Hyphomicrobium denitrificans, and Hyphomicrobium zavarzinii. Appl. Environ. Microbiol. 81, 5003e5014. Mathis, D., Benoist, C., 2012. The influence of the microbiota on type-1 diabetes: on the threshold of a leap forward in our understanding. Immunol. Rev. 245, 239e249. O’Mahony, S.M., Clarke, G., Borre, Y.E., Dinan, T.G., Cryan, J.F., 2015. Serotonin, tryptophan metabolism and the brain-gut-microbiome axis. Behav. Brain Res. 277, 32e48. Petriello, M.C., Hoffman, J.B., Vsevolozhskaya, O., Morris, A.J., Hennig, B., 2018. Dioxin-like PCB 126 increases intestinal inflammation and disrupts gut microbiota and metabolic homeostasis. Environ. Pollut. 242, 1022e1032. Salim, S.Y., Kaplan, G.G., Madsen, K.L., 2014. Air pollution effects on the gut microbiota. Gut Microb. 5, 215e219. Snedeker, S.M., Hay, A.G., 2012. Do interactions between gut ecology and environmental chemicals contribute to obesity and diabetes? Environ. Health Perspect. 120, 332e339. Tilg, H., Kaser, A., 2011. Gut microbiome, obesity, and metabolic dysfunction. J. Clin. Investig. 121, 2126e2132. Tremaroli, V., Backhed, F., 2012. Functional interactions between the gut microbiota and host metabolism. Nature 489, 242e249. Zhang, L., Nichols, R.G., Correll, J., Murray, I.A., Tanaka, N., Smith, P.B., Hubbard, T.D., Sebastian, A., Albert, I., Hatzakis, E., Gonzalez, F.J., Perdew, G.H., Patterson, A.D., 2015. Persistent organic pollutants modify gut microbiota-host metabolic homeostasis in mice through aryl hydrocarbon receptor activation. Environ. Health Perspect. 123, 679e688. Zhao, B., DeGroot, D.E., Hayashi, A., He, G., Denison, M.S., 2010. CH223191 is a ligand-selective antagonist of the Ah (dioxin) receptor. Toxicol. Sci. 117, 393e403.