Gill Junction Injury and Microbial Disorders Induced by Microcystin-Leucine Arginine in Lithobates catesbeianus Tadpoles

Microcystin-LR (MC-LR) is widely present in waters around the world, but its potential toxic effects and mechanisms on amphibian gills remain unknown. In the present study, tadpoles (Lithobates catesbeianus) were exposed to environmentally realistic concentrations of 0.5, 2 μg/L MC-LR, and 0 μg/L MC-LR (Control) for 30 days with the objective to unveil the impairment of gill health. The lysozyme was downregulated, while pattern recognition receptors and complement and adaptive immune processes were upregulated and the ability of gill supernatant to inhibit pathogenic bacteria decreased in the 0.5 and 2 μg/L MC-LR groups. The transcriptions of epithelial barrier components (e.g., CLDN1) were significantly decreased in MC-LR-exposed gills, while the gill content of lipopolysaccharide (LPS) endotoxins and the transcriptions of downstream responsive genes (e.g., TLR4 and NF-κB) were concurrently increased. In addition, the number of eosinophils and the expression of pro-inflammatory cytokines (e.g., IL-1β and TNF-α) were increased. These results imply that exposure of tadpoles to low environmentally concentrations of MC-LR leads to inflammation, increased permeability, and a reduced ability to inhibit pathogenic bacteria. The epithelial cells of inner gill filaments increased and transcriptions of hypoxic stress genes (e.g., HIF-1α, FLT1, and SERPINE1) were upregulated within the exposed group. As a consequence, exposure to MC-LR may lead to hypoxic stress. MC-LR exposure also drove gill microbiota to a dysbiosis. The relative abundance of Elizabethkingia was positively correlated with content of LPS and transcriptions of NF-κB and TNF-α. Overall, this study presents the first evidence about the pronounced impacts of MC-LR exposure on gills of amphibians, highlighting the susceptibility of early developing tadpoles to the environmental risks of MC-LR.


Introduction
In early developing tadpoles and bony fish, the gills have various functions such as gas exchange, immune defense, acid-base balance, and ion regulation [1][2][3]. Gills are usually sensitive to chemicals in water because they are directly and constantly exposed to chemicals dissolved in the water [4]. Chlorpyrifos has been reported to induce damage to the physical and immune barriers of Cyprinus carpio Linnaeus gills, while copper oxide

Transcriptome Analysis
A total of 406,733,698 clean reads and 22,224 genes from 9 RNA libraries were obtained (Table S1). Based on the quantified data for all transcriptions, principal component analysis (PCA) indicated that three groups exposed to environmentally realistic concentrations of MC-LR had changes on the transcription levels in gill tissues ( Figure S2). For the 0.5MCG and 2MCG, 706 and 666 DEGs were identified, respectively (Tables S2 and  S3). In the 0.5MCG, 388 GO terms were markedly enriched (p < 0.05), including oxygen transport, innate immune response, complement activation, removal of superoxide radicals, and apoptotic processes (Tables S4 and Figure S3A). Twelve KEGG pathways were significantly enriched (p < 0.05), including phagosomes, herpes simplex infection, and intestinal immune networks for IgA production (Tables S5 and Figure S3B). In the 2MCG, 281 GO terms were markedly enriched (p < 0.05), including innate immune response, positive regulation of B cell activation, inflammatory response, oxygen transport, and tolllike receptor signaling pathways (Tables S6 and Figure S3C). Eighteen KEGG pathways (p < 0.05), including tight junctions, necroptosis, and complement and coagulation cascades were identified (Tables S7 and Figure S3D).

Transcriptome Analysis
A total of 406,733,698 clean reads and 22,224 genes from 9 RNA libraries were obtained (Table S1). Based on the quantified data for all transcriptions, principal component analysis (PCA) indicated that three groups exposed to environmentally realistic concentrations of MC-LR had changes on the transcription levels in gill tissues ( Figure S2). For the 0.5MCG and 2MCG, 706 and 666 DEGs were identified, respectively (Tables S2 and S3). In the 0.5MCG, 388 GO terms were markedly enriched (p < 0.05), including oxygen transport, innate immune response, complement activation, removal of superoxide radicals, and apoptotic processes (Table S4 and Figure S3A). Twelve KEGG pathways were significantly enriched (p < 0.05), including phagosomes, herpes simplex infection, and intestinal immune networks for IgA production (Table S5 and Figure S3B). In the 2MCG, 281 GO terms were markedly enriched (p < 0.05), including innate immune response, positive regulation of B cell activation, inflammatory response, oxygen transport, and toll-like receptor signaling pathways (Table S6 and Figure S3C). Eighteen KEGG pathways (p < 0.05), including tight junctions, necroptosis, and complement and coagulation cascades were identified (Table S7 and Figure S3D).

DEGs Associated with Hypoxic Stress and Oxygen Transport
The HIF-1α signaling pathway is the main pathway of hypoxia stress. Three DEGs (FLT1, SERPINE1, and HIF-1α) in the HIF-1α signaling pathway were upregulated in the 0.5MCG. Four DEGs (FLT1, SERPINE1, HMOX1, and HIF-1α) in the HIF-1α signaling pathway were also upregulated in the 2MCG. This indicated that MC-LR pollution resulted in hypoxia stress in the gills. Therefore, the oxygen transport capacity was further analyzed. In the 0.5MCG, there was a significant downregulation of HBA5 related to oxygen transport, and 2 DEGs (HBB2 and HBA3) were downregulated in the 2MCG ( Figure  4A,B). This suggested that MC-LR led to a reduced ability of respiratory capacity in the gill.

DEGs Associated with Hypoxic Stress and Oxygen Transport
The HIF-1α signaling pathway is the main pathway of hypoxia stress. Three DEGs (FLT1, SERPINE1, and HIF-1α) in the HIF-1α signaling pathway were upregulated in the 0.5MCG. Four DEGs (FLT1, SERPINE1, HMOX1, and HIF-1α) in the HIF-1α signaling pathway were also upregulated in the 2MCG. This indicated that MC-LR pollution resulted in hypoxia stress in the gills. Therefore, the oxygen transport capacity was further analyzed. In the 0.5MCG, there was a significant downregulation of HBA5 related to oxygen transport, and 2 DEGs (HBB2 and HBA3) were downregulated in the 2MCG ( Figure 4A,B). This suggested that MC-LR led to a reduced ability of respiratory capacity in the gill.

Biochemical Indicators of Gills and Antibacterial Activity
Inspired by the analysis of the transcriptomic data, some biochemical parameters were further analyzed, including gill immune function, anti-pathogenic activity, oxidative stress, apoptosis, endotoxin, and inflammatory factors. Compared with the control, the concentrations of immunoglobulins (IgA, IgG, and IgM) were significantly upregulated and the catalytic activity of LZM was significantly downregulated in the 0.5MCG and2MCG, which was consistent with the transcriptomic data ( Figure 5). In the disc diffusion assay, the area of inhibition of pathogenic bacteria (Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus aureus) decreased with increasing concentration in the MC-LR exposure groups, suggesting that MC-LR exposure caused a reduction in the ability of gills to inhibit the growth of pathogenic bacteria ( Figure 5). The MC-LR groups significantly (p < 0.05) reduced antioxidant enzyme activities, i.e., SOD and GSH levels, while increasing ROS and caspase 3 activity in gill tissue in a concentration-dependent manner relative to the control group ( Figure 6). Furthermore, MC-LR significantly increased the levels of LPS and IL-1β in gills relative to the control group ( Figure 6).

Correlation between Gill Microbiota and Inflammatory Events Triggered by LPS
To assess the potential relationship between inflammatory events triggered by LPS and gill microbiota genera, Spearman analysis was performed to explore any correlation. As shown in Figure 8, relative abundances of Elizabethkingia were significantly and positively correlated with inflammatory events triggered by LPS, including the concentrations of LPS and IL-1β, and the transcript levels of IRAK4, IKKA, NF-κB, and TNF-α. However, there were significantly negative correlations of Elizabethkingia with the transcript level of CLDN1, which represents the gill epithelial barrier component.

Discussion
In the present study, exposure to MC-LR resulted in increase in the permeability of the gill epithelium barrier, hyperplasia of inner gill filaments, impaired respiratory function, immune dysregulation, and inhibited antimicrobial capacity. MC-LR subchronic ex- Figure 8. Correlation between gill microbial genera and metabolites. Blue and red colors indicate negative and positive correlations, respectively. Color intensity is proportional to extent of change. Significant correlations are indicated by asterisk (* p < 0.05 and ** p < 0.01; r > 0.5 or < −0.5).

Discussion
In the present study, exposure to MC-LR resulted in increase in the permeability of the gill epithelium barrier, hyperplasia of inner gill filaments, impaired respiratory function, immune dysregulation, and inhibited antimicrobial capacity. MC-LR subchronic exposure also drove the gill microbiome to dysbiosis. In addition, oxidative stress-mediated apoptosis and increased lipopolysaccharide (LPS) endotoxin can lead to the activation of the TLR4/TNF-α signaling pathway, which is an important regulatory mechanism of gill injury after subchronic exposure to MC-LR. The present findings highlighted the susceptibility of early developing tadpoles to MC-LR, which needs an urgent and accurate evaluation.
The gill epithelial barrier separates the organism from its environment and protects it from pathogens and toxins [36]. Tight junctions are a major component of the gill epithelial barrier [37]. The disruption of the tight junctions leads to increased permeability and the penetration of exogenous substances into the body inducing local or systemic inflammatory response [38,39]. After MC-LR exposure, the transcription levels of tight junction genes were downregulated, implying that real environmental concentrations (0.5 µg/L and 2 µg/L) resulted in an impaired gill epithelial barrier. A previous study has shown that higher exposures of 10 µg/L MC-LR impaired the gill tissue barrier in Cyprinus carpio Linnaeus [5]. The implication is that MC-LR exposure can leads the gill tissue barrier of amphibian tadpoles more susceptible to damage than some fish. In addition, eosinophils can promote the progression of inflammation by the release the contents of granules [40]. There was a significant concentration-dependent increase in eosinophils in the gill tissue of the MC-LR exposed group. The transcription of genes associated with inflammation was also upregulated in the MC-LR group, manifesting as an enhanced inflammatory response. These results emphasize is similar to previous observations about MC-LR exposure in L. catesbeianus gut tissue [14]. MC-LR usually induces oxidative stress in different tissues, such as gonad and liver [31,37]. The oxidative stress was also observed in tadpole gills after exposure to low environmental concentrations of MC-LR. Furthermore, oxidative stress can usually activate the caspase cascade reaction and apoptosis [41][42][43]. The caspase 3 activity and other related apoptotic genes transcription were upregulated after exposure to low environmental concentrations of MC-LR. These results emphasize that MC-LR exposure produces oxidative stress to promote apoptosis, which is similar to previous observations about MC-LR exposure in Rana nigromaculata liver and gonad tissue [31,44]. However, no oxidative stress was observed in Xenopus laevis tadpoles [45], suggesting that there may still be differences in MC-LR sensitivity in different amphibians.
The gill mucosa is an important immune barrier, including innate and adaptive immunity [46,47]. Lysozyme is the widely distributed enzyme of the innate immune system that exhibits antibacterial activity [48]. After MC-LR exposure, the expression of lysozyme was reduced in gill tissue, while other nonspecific and specific immunities were upregulated. The disc assay can assess immune substances in tissues on the ability to inhibit bacteria after contaminant exposure [30,49]. The disc diffusion test showed that MC-LR exposure induced a reduction in the ability of the gills to inhibit the growth of pathogenic bacteria. Therefore, altered immune material following MC-LR exposure affects gill ability to inhibit pathogens, further implying that gill tissue is also more susceptible to pathogenic bacteria following MC-LR exposure.
Early developmental tadpoles exchange gas primarily through the gills. After exposure, the epithelial cells of inner gill filaments increased with increasing concentrations of MC-LR. HIF-1α signaling pathway genes associated with hypoxic stress were upregulated, implying the presence of hypoxic stress after MC-LR exposure. Generally, hypoxic stress implies a reduction in oxygen transport capacity [50], and a similar phenomenon of reduced oxygen transport capacity was seen in tadpole gills with MC-LR exposure. Thus, realistic environmental concentrations of MC-LR exposure induce the proliferation of gill filament epithelial cells, which may thicken gill filament, separate the blood from the air, produce hypoxic stress, and reduce the oxygen transport capacity. However, the mechanism of gill filament epithelial cell proliferation is yet to be further studied after environmental pollutants exposure [5].
The gill microbiome, together with the gill, serves as a barrier for the host to adapt to external environments and maintain its health [10,11,51]. LPS constitutes a significant portion of the cell wall of gram-negative bacteria [52]. A significant increase in LPS endotoxin content was found in the gills of tadpoles exposed to MC-LR. The correlation analysis indicated a significant positive correlation between LPS content and relative abundances of Elizabethkingia genus that is Gram-negative. This implies that Elizabethkingia may be an important source of LPS in the gill tissue after MC-LR exposure. LPS infiltration activated TLR4 and induced the key inflammatory transcript factor NF-κB to translocate into the nucleus, leading to the upregulation of a variety of pro-inflammatory cytokines, such as IL-1β and TNF-α [53,54]. Activated TLR4/NF-κB signaling pathways and inflammation were also observed in several animal models and cell lines exposed to microcystins [44,[55][56][57]. In this study, a significant increase in LPS content in the gill was found after MC-LR exposure, and the transcriptions of TLR4 and NF-κB were also upregulated. In addition, expression of the pro-inflammatory cytokines IL-1β and TNF-α was significantly elevated in the gills after MC-LR exposure. Therefore, LPS infiltration activated the TLR4/TNF-α signaling pathway may be an important pathway to cause gill inflammation. MC-LR exposure at environmentally realistic concentrations altered the composition and decline diversity of the gill microbial community. Significantly reducing the diversity of gill microbiota may lead to a reduction in competition or ability to fight pathogens [58]. Acetobacterium, Succinispira, Lachnoclostridium, Eubacterium, and Desulfovibrio bacteria have the function of maintaining the immune barrier [59][60][61][62][63]. The significant reduction of related genera in the gill of MC-LR-exposed tadpoles may mean that MC-LR exposure further interferes with gill mucosal immunity. Norank Clostridia-UCG-014 and Robinsoniella bacteria have potential anti-inflammatory effects [64,65]. The significant reduction of norank Clostridia-UCG-014 and Robinsoniella bacteria after exposure in MC-LR indicates fluctuations in abundances of norank Clostridia-UCG-014 and Robinsoniella may not be conducive to the elimination of inflammation.

Conclusions
In this study, MC-LR subchronic exposure seriously affected L. catesbeianus tadpoles' gill function and microbiome. After MC-LR exposure, the permeability of the gill epithelial barrier was increased, while the ability to inhibit pathogenic bacteria was decreased. The increase of epithelial cells of inner gill filaments leads to hypoxia stress. Elizabethkingia may be an important source of LPS in the gill tissue after MC-LR exposure. LPS infiltration activated the TLR4/TNF-α signaling pathway, which may be an important pathway when it comes to causing gill inflammation after MC-LR exposure. MC-LR exposure disrupted the balance of redox homeostasis and activated the apoptosis. In addition, MC-LR subchronic exposure drove the gill microbiome to dysbiosis. In summary, these findings clue the first evidence that amphibian gills are very sensitive to MC-LR pollution, which provide a new basis for accurately assessing the risks of MC-LR pollution to amphibians.

Ethics Statement
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of Anhui Normal University and renewed on 7 June 2022 (protocol code AHNU-ET2022051).

Rearing of Animals and Sample Collection
Tadpoles (Lithobates catesbeianus) were exposed to environmental reared conditions based on the method of He et al. (2022) [14]. Briefly, tadpoles (G26 stage, 10 days post fertilization) were divided into three exposure groups: negative control, 0.5, and 2 µg/L MC-LR group (MC-LR purity > 95.0%; Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China). Actual concentrations of MC-LR were 0, 0.5, and 1.95 µg/L, as determined by ELISA kits (a minimum detectable concentration limit of 0.1 ng/mL, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China) throughout the exposure period. As the measured concentrations were very close to the nominal concentrations, references to nominal concentrations are made throughout. In the following content, control, 0.5, and 2 µg/L MC-LR groups correspond to 0MCG, 0.5MCG, and 2MCG, respectively. Three replicate tanks of each group were conducted with each tank containing 20 tadpoles in 5 L medium. After 30 days (40 days post-fertilization) of MC-LR exposure, tadpoles from each tank were anesthetized with 0.02% MS-222 and gill samples were collected.

Histopathology of Gill
Gill tissues were fixed in 4% paraformaldehyde for histopathological observation. After paraformaldehyde-fixed gill tissues were dehydrated in a graded ethanol series, cleared in xylene, paraffin-embedded, and cut to 5 µm thickness, then stained with hematoxylin-eosin (HE) for histological observation [33]. Fifteen complete histological sections were randomly selected from each gill tissue (n = 6 gill tissue per group) and photographed using light microscope (Olympus BX61, Tokyo, Japan), and the number of gill filament epithelial cells and eosinophils were counted using Image-Pro Plus 6.0 software.

Biochemical Parameters of Gill
Each biological replicate (n = 3 replicates per group) consisted of three gills from each tank pooled together for biochemical analysis. Homogenized gill tissues were centrifuged (4 • C, 15 min, 6000× g) to collect supernatants, which were transferred to fresh tubes for further analysis. Commercial colorimetric kits were used to detect the levels of glutathione (GSH) and reactive oxygen species (ROS), as well as the activities of caspase 3, superoxide dismutase (SOD), and lysozyme (LZM). Moreover, the enzyme-linked immunosorbent assay (ELISA) kits were conducted to check the contents of immunoglobulins (IgA, IgM, and IgG), interleukin-1β (IL-1β), and lipopolysaccharide (LPS). All kits were used following the manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) with the biochemical measures normalized to the protein content. The optical density measurement was performed using a Tecan Spark 10M microplate reader (Tecan Trading AG, Zurich, Switzerland).

Antibacterial Activity of the Gill
Each biological replicate (n = 3 replicates per group) consisted of three gills from each tank pooled together for antimicrobial assays. Potential changes in gill antibacterial capacity induced by MC-LR exposure were assessed by disc diffusion assays [30,49]. Three pathogenic bacteria (e.g., Escherichia coli, CMCC44102; Pseudomonas aeruginosa, CICC 21625; and Staphylococcus aureus, CMCC26003) were used.

Transcriptome Sequencing and Bioinformatic Analyses
Each biological replicate (n = 3 replicates per group) consisted of three gills from each tank pooled together for transcriptome sequencing. TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract total RNA, while total RNA quality was analyzed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and 9 cDNA libraries were prepared using the TruSeq ® RNA Sample Preparation Kit (Illumina, San Diego, CA, USA). And then performed sequencing on an Illumina Novaseq™ 6000 at the LC Sciences (Hangzhou, China). We first used the Cutadapt software (version 1.18) to remove the joint sequences and low-quality sequences in the original data to obtain clean data. Clean data was compared to the genomic data (L. catesbeianus, assembly RCv2.1) using HISAT2 software (version 2.1.0), and then data were assembled with StringTie software (v2.1.1). Assembly annotations were obtained using gffcompare software (version: gffcompare-0.9.8. Linux-x86-64) by comparing transcripts with the reference genome. The expression levels of all transcripts and genes were determined by calculating the fragments per kilobase per million reads mapped (FPKM) using StringTie and ballgown. Differentially expressed genes (DEGs) were determined with the judgment criteria of a fold change >2 or <0.5 and adjusted p ≤ 0.05 by the edgeR in R package. Then, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichments of the DEGs were analyzed. The bioinformatics analyses were performed as previously published protocols [14].

16S rRNA Amplicon Sequencing and Bioinformatics
Each biological replicate (n = 3 replicates per group) consisted of three gills from each tank pooled together for 16S rRNA gene sequencing analysis. Total genomic DNA was extracted from the soil using the Mobio PowerSoil DNA Isolation Kit according to the manufacturer's instructions (Mobio, Laboratories, Carlsbad, CA, USA). After genomic DNA extraction, the extracted genomic DNA was detected by 1% agarose gel electrophoresis. The 16S rDNA V3-V4 region was amplified with the primer sequences (338F, ACTCC-TACGGGAGGCAGCAG and 806R, GGACTACHVGGGTWTCTAAT). Sequencing of the amplified fragment was performed on -the Illumina MiSeq platform (Illumina, San Diego, CA, USA). The paired-end reads obtained from MiSeq sequencing were spliced based on overlap relationships, while the sequence quality was quality controlled and filtered, then analyzed using QIIME2, and the reads were clustered into operational taxonomic units (OTUs) using Uparse with 97% similarity. The taxonomy of each OTU representative sequence was analyzed by RDP Classifier version 2.1.1 against the 16S rRNA database (Silva v138) using a confidence threshold of 0.7. The alpha diversity indices including Sobs, Chao, Shannon, and Simpson at OTU level were calculated by MOTHUR (v1.31.2). The beta-diversity based on the OTU level was calculated by measuring the Bray-Curtis distance using QIIME (version 1.9.1) and visualized using principal coordinate analysis (PCoA). Significance among samples were determined with analysis of similarity (ANOSIM) algorithm (9999 permutations) implemented using the Vegan package in R.

Statistical Analyses
All experimental data were analyzed statistically using R software version 3.6.3 (R Project for Statistical Computing, Vienna, Austria). Statistical comparisons among multiple groups were carried out by one-way ANOVA followed by the least significant difference (LSD) test (data with normal distribution and homogeneity of variance) or the Kruskal-Wallis test followed by the Dunn's test (data not normally distributed or without homogeneity of variance). The correlation between the abundance of gill microbiota genera and inflammatory events indicators by LPS (TLR4, IRAK4, IKKA, NF-κB, TNF-α, CLDN1 gene expression levels, and LPS, IL-1β content) was determined by Pearson analysis. p-value indicates a significant difference and p < 0.05 indicates statistical significance.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/toxins14070479/s1, Figure S1, the number of internal gill epithelial cells and the average number of eosinophils. Figure Table S1: Summary of generated transcriptome data; Table S2: 706 DEGs between 0.5 µg/L MC-LR group and Control group; Table S3: 666 DEGs between 2 µg/L MC-LR group and Control group; Table S4: 388 GO enrichment between 0.5 µg/L MC-LR group and Control group; Table S5: 12 pathway enrichment between 0.5 µg/L MC-LR group and Control group; Table S6: 281 GO enrichment between 2 µg/L MC-LR group and Control group; Table S7: 18 pathway enrichment between 0.5 µg/L MC-LR group and Control group; Table S8: Alterations in the alpha diversity indices of gill microbiota. Data Availability Statement: In this study, sequencing of 16S rRNA genes from frog tadpole gill microbiota samples has been submitted to the NCBI Sequence Read Archive under accession numbers PRJNA843355. The raw transcriptome data have been submitted to the NCBI. Sequence Read Archive under accession PRJNA843352.

Acknowledgments:
We would like to thank Liang Chen from the Institute for Ecological Research and Pollution Control of Plateau Lakes, School of Ecology and Environmental Science, Yunnan University, China, for his helpful comments.

Conflicts of Interest:
The authors declare no conflict of interest.