Selection strategy of dextran sulfate sodium-induced acute or chronic colitis mouse models based on gut microbial profile

Dextran sulfate sodium (DSS) replicates ulcerative colitis (UC)-like colitis in murine models. However, the microbial characteristics of DSS-triggered colitis require further clarification. To analyze the changes in gut microbiota associated with DSS-induced acute and chronic colitis. Acute colitis was induced in mice by administering 3% DSS for 1 week in the drinking water, and chronic colitis was induced by supplementing drinking water with 2.5% DSS every other week for 5 weeks. Control groups received the same drinking water without DSS supplementation. The histopathological score and length of the colons, and disease activity index (DAI) were evaluated to confirm the presence of experimental colitis. Intestinal microbiota was profiled by 16S rDNA sequencing of cecal content. Mice with both acute and chronic DSS-triggered colitis had significantly higher DAI and colon histopathological scores in contrast to the control groups (P < 0.0001, P < 0.0001), and the colon was remarkably shortened (P < 0.0001, P < 0.0001). The gut microbiota α-diversity was partly downregulated in both acute and chronic colitis groups in contrast to their respective control groups (Pielou index P = 0.0022, P = 0.0649; Shannon index P = 0.0022, P = 0.0931). The reduction in the Pielou and Shannon indices were more obvious in mice with acute colitis (P = 0.0022, P = 0.0043). The relative abundance of Bacteroides and Turicibacter was increased (all P < 0.05), while that of Lachnospiraceae, Ruminococcaceae, Ruminiclostridium, Rikenella, Alistipes, Alloprevotella, and Butyricicoccus was significantly decreased after acute DSS induction (all P < 0.05). The relative abundance of Bacteroides, Akkermansia, Helicobacter, Parabacteroides, Erysipelatoclostridium, Turicibacter and Romboutsia was also markedly increased (all P < 0.05), and that of Lachnospiraceae_NK4A136_group, Alistipes, Enterorhabdus, Prevotellaceae_UCG-001, Butyricicoccus, Ruminiclostridium_6, Muribaculum, Ruminococcaceae_NK4A214_group, Family_XIII_UCG-001 and Flavonifractor was significantly decreased after chronic DSS induction (all P < 0.05). DSS-induced acute and chronic colitis demonstrated similar symptoms and histopathological changes. The changes in the gut microbiota of the acute colitis model were closer to that observed in UC. The acute colitis model had greater abundance of SCFAs-producing bacteria and lower α-diversity compared to the chronic colitis model.


Background
Inflammatory bowel disease (IBD) comprises of Crohn's disease (CD) and ulcerative colitis (UC). The hallmark of CD is chronic colorectal inflammation, which manifests as diarrhea, bloody mucus stool, abdominal pain and other symptoms that severely affect patients' quality of life. The pathogenesis of UC is complex, and involves genetic, immunological and microbiological factors.
The dextran sodium sulfate (DSS)-stimulated colitis model mimics the pathological damage and symptoms of human UC [1], and is routinely used for studying the pathogenesis and pharmacodynamics of UC [2]. DSS is a water-soluble negatively-charged sulfated polysaccharide that can erode the intestinal mucosa [3,4]. DSS increases colon permeability by destroying interepithelial cell tight junctions, reducing mucin levels while simultaneously altering the resident microbiota [5,6]. A compromised intestinal barrier allows penetration of harmful bacterial from the intestinal lumen into the systemic circulation, thereby triggering an inflammatory response. The efficacy of colitis induction depends on the concentration of DSS (typically 1-5%), dosage regimen (long-or short-term), animal strains (BALB/c and C57BL/6 mice are more susceptible), and feeding conditions [7]. With the development of 16S rDNA sequencing, the DSS-induced colitis model been proven to be particularly well-suited for analyzing colitis-associated microbial changes. DSS-induced acute and chronic colitis mimic histopathological damages associated with UC. Intermittent administration of different concentrations of DSS over varying durations result in alternating active and remission phases, which is reflective of disease progression in UC patients. In order to establish acute and chronic models of colitis in mice, either 3% or 2.5% DSS is added to the drinking water for varying durations [8,9].
Short chain fatty acids (SCFAs) are organic fatty acids with aliphatic tails of six or fewer carbons, including acetate, propionate, butyrate, valproate and hexanoate, that are mainly produced by anaerobic microorganisms while fermenting indigestible carbohydrates. SCFAs inhibit the production of inflammatory cytokines in human intestinal epithelial cells in vitro, which promotes intestinal mucosal repair by mitigating inflammation [10]. In vitro studies have shown that SCFAs also exert an anti-tumor effect by reducing the secretion of tumor necrosis factor in intestinal epithelial cells [11]. In addition, SCFAs (especially butyrate) regulate gut microbiota by modulating the intestinal lumen pH to levels that are conducive to the growth of SCFAs-producing bacteria [12]. Acetate, propionate and butyrate are the most abundant SCFAs produced during anaerobic fermentation of dietary fibers in the intestine, and account for more than 85% of all SCFAs. The relative proportion of the three is about 3:1:1 [13], and affected by the host genotype, intestinal microbial composition and diet. Anaerobic bacteria in the colon play an important role in fermentation. The main SCFAs producing-bacteria in the human gut belong to the phylum Firmicutes, in particular Faecalibacterium prausnitzii and Clostridium leptum of family Ruminococcaceae, Eubacterium rectale and Roseburia spp. of Lachnospiraceae, as well as Butyricicoccus, Rikenella and Alistipes [14,15]. In addition, the mucin-degrading bacteria Akkermansia muciniphila (phylum Verrucomicrobia) produces both propionate and acetate [16].
Consistent with our previous studies [17][18][19][20], we found that DSS-triggered acute and chronic colitis were similar in terms of disease activity, colon shortening and histopathological changes in the colon, but differed remarkably in terms of gut microbiota profiles. Given that the initiation and progression of UC centers heavily on gut microbiota proportions, the therapeutic potential of fecal microbiota transplantation is increasingly being assessed in UC animal models. The selection of appropriate colitis mouse models vis-à-vis the gut microflora has become particularly important. In this study, we analyzed the changes in gut microbiota associated with DSS-triggered acute and chronic colitis in a mouse model in order to provide a reference basis for selecting an appropriate model.

Results
Acute and chronic colitis were successfully induced by DSS exposure Mice exposed to DSS of varying concentrations and doses successfully developed acute and chronic colitis [8]. Based on preliminary results, we administered 3% DSS for 1 week to stimulate acute colitis and 2.5% DSS intermittently for 5 weeks to induce chronic colitis. As shown in Fig. 1A and B, the DAI scores of mice increased significantly after 1 week of administering 3% DSS compared to the untreated controls (P < 0.0001). However, the DAI scores fluctuated slightly after 14 days during chronic colitis modeling, corresponding to transient weight loss without change in stool characteristics. model had greater abundance of SCFAs-producing bacteria and lower α-diversity compared to the chronic colitis model.

Keywords:
Dextran sulfate sodium, Acute colitis, Chronic colitis, Gut microbiota This can be attributed to intermittent DSS administration over the 5-week period. Nevertheless, DAI scores were notably raised in the chronic colitis group relative to the control group (P < 0.0001) after 5 weeks. Both acute and chronic colitis groups demonstrated no remarkable differences in terms of DAI scores (P = 0.9198). As shown in Fig. 1C, both forms of colitis were accompanied by significant shortening of the colon compared to their respective controls (P < 0.0001 for both), and the colon lengths did not vary significantly between the acute and chronic colitis models (P = 0.9724). Histopathological examination of the colon tissues of untreated control mice indicated clearly demarcated mucosa, intact epithelium, neatly arranged glands, abundant goblet cells, and absence of inflammatory cell infiltrates in the lamnia propria. In contrast, the DSS-treated mice exhibited considerable structural damage including incomplete glands, loss of epithelial cells, as well as extensive inflammatory cell infiltrates. Interestingly, the acute colitis group demonstrated notably thinned intestinal mucosa in contrast to that seen in the chronic colitis group. Mice with chronic colitis demonstrated locally thickened mucosa, which could be the result of alternating inflammatory and remission phases during the modeling process (Fig. 1E). Nevertheless, the histopathological scores of both groups were similar (P = 0.5113), and markedly higher compared to that of the respective controls (P < 0.0001 for both) (Fig. 1D).

Acute and chronic DSS exposure distinctly altered the diversity of the intestinal microflora
The intestinal microbiota profiles were analyzed by 16S rDNA sequencing. The Pielou index ( Fig. 2A), which is indicative of species evenness, was notably lower in both the acute (P = 0.0022) and chronic (P = 0.0649) colitis groups compared to the respective controls, and did not differ significantly between the forms of colitis (P = 0.0022). The Shannon index (Fig. 2B) is a measure of species richness and evenness, and was decreased in both acute colitis (P = 0.0022) and chronic colitis (P = 0.0931) groups in contrast to the respective controls. Furthermore, acute colitis resulted in significantly lower Shannon index compared to chronic colitis (P = 0.0043). Taken together, DSS-induced colitis lowers mice gut microbiota species diversity, with more drastic changes seen in models of acute colitis.
The β-diversities of the intestinal microbiota across both cohorts were assessed using UPGMA and FastTree (Fig. 2C), NMDS (Fig. 2D), PCA (Fig. S1A), and PCoA (Fig. S1B). The results of UPGMA and FastTree showed that the acute colitis and control groups were further apart compared to the chronic colitis group and its corresponding control, indicating that acute colitis induced more significant changes in the intestinal microflora. The distance between acute and chronic colitis mice models was halfway between each colitis cohort and its respective control. Consistent results were obtained with NMDS, PCA and PCoA (stress< 0.1, PCA1 + PCA2 > 50%, PCoA1 + PCoA2 > 50%). The acute colitis group was the farthest from the rest, indicating that acute exposure to DSS induced significant changes in the gut microbial structure.

Acute and chronic DSS-triggered colitis resulted in a distinct microbiota profile
As shown in Fig At the species level (Fig. 3D), acute colitis increased the relative abundance of Bacte-roides_acidifaciens (31.41% vs. 0.94%, P = 0.004), α-diversity was derived using a combination of the species richness (species situation) and species evenness (distribution) using the Pielou and Shannon indices. C Gut microbiota relationships and distribution are characterized with a UPGMA clustering tree. Samples are represented by end branches. In general, samples belonging to similar groups were clustered into a large branch with different branches representing different groups. Distance is indicated by the vertical axis. Samples with similar structure of bacterial communities are located in the same branch. D NMDS (Non-metric Multi-Dimensional Scaling). NMDS representas a dimension reduction analysis established on distance ranking. It is superior to linear models (including PCA and PCoA) and more accurately translates the nonlinear structure of ecological data. The accuracy of the model is evaluated by the stress value. The smaller the stress value is, the more reliable the model is. Generally, the stress value less than 0.1 is better. **P < 0.01 2.2%, P = 0.522), Lactobacillus_gasseri (0.14% vs. 0.68%, P = 0.037) and Lactobacillus_reuteri (0.34% vs. 0.36%, P = 1.000) was lower compared to the control group.

Acute and chronic DSS-induced colitis led to distinct changes in the intestinal microbiota
Welch's t test was used to analyze the microbial changes induced by acute colitis and chronic colitis. There was a lower abundance of the phyla Firmicutes, Verrucomicrobia, Actinobacteria, Acidobacteria and Planctomycetes in the acute colitis group compared to the control ( Fig. 4A and B). On the other hand, the chronic colitis group, demonstrated increased abundance of Bacteroidetes, Verrucomicrobia and Epsilonbacteraeota, and decreased proportions of Firmicutes and Actinobacteria. As shown in Fig. 4C and D, the relative proportion of the Bacteroidaceae family increased remarkably in the acute colitis group, and that of Lachnospiraceae, Muribaculaceae, Ruminococcaceae, Rikenellaceae, Prevotellaceae, Marinifilaceae and Christensenellaceae decreased As shown in Fig. 5A Fig. 5C and D), acute colitis drastically elevated the relative proportion of Bacteroides_acidifaciens, and lowered that of Ruminococcus_flavefaciens, Lachnospiraceae_bacterium_ COE1 and Clostridium_leptum. Chronic colitis increased the abundance of Bacteroides_acidifaciens, Bacteroides_ massiliensis_dnLKV3, Bacteroides_caecimuris, Akker-mansia_muciniphila and Oscillibacter_sp_1-3. Thus, the proportion of SCFAs-producing bacteria was elevated in the intestines of younger mice exposed to DSS over a longer duration. In contrast, acute colitis increased the abundance of conditional pathogens.

Discussion
DSS and 2,4,6-trinitro-benzenesulfonic acid (TNBS) are established agents for inducing acute and chronic colitis [17][18][19][20]. In this preliminary study, we found that administration of 3 and 2.5% DSS via drinking water can respectively induce acute and chronic colitis in mice while ensuring their survival. The affected mice showed significantly increased DAI scores, shortened colons, and significant damage to the intestinal mucosa. Interestingly, the weight of the control mice also fluctuated during the 5-week long cycle of chronic colitis modeling, which resulted in a transient increase in DAI score. However, the control group of the acute model was not affected due to the short treatment duration. Likewise, the colonic tissue of older mice also displayed limited neutrophil infiltration that did not penetrate beyond the lamina propria. In addition, the alternating active and remission phases during the induction of chronic colitis led to localized thickening of colonic mucosa, in contrast to the uniform thinning observed in acute colitis. Our models of acute and chronic colitis demonstrated good stability and homogeneity with a lack of significant variability in terms of disease features.
However, DSS-induced acute colitis led to notable reduction in the α-diversity of the gut microbiota compared to chronic colitis. Further distinctions between the two groups were highlighted with β-diversity analysis. Table S1 summarizes the changes in microbial structure at the levels of the phylum, family, genus and species. DSS-induced colitis stimulated the growth of Bacteroidetes, Epsilonbacteraeota and Deferribacteres, while lowering that of Firmicutes, Actinobacteria and Patescibacteria. The Firmicutes/Bacteroidetes ratio in the gut is an indicator of intestinal inflammation, and increases during colitis and other inflammatory conditions [21,22]. In addition, the increased abundance of some conditional pathogens is also an important intestinal flora feature of UC. Li et al [23]. collected 56 mucosal microbiome samples from 28 Chinese UC patients and their healthy family partners, and identified four opportunistic "pathogens" (i.e., Clostridium tertium, Odoribacter splanchnicus, Ruminococcus gnavus and Flavonifractor plautii) with potential diagnostic and therapeutic significance for UC, which were inhibited in healthy individuals. Moreover, Huh et al. [24] identified Fusobacterium nucleatum as a marker for early gut microbial dysbiosis in inflammatory bowel diseases.
In our study, DSS-induced acute and chronic colitis simulated UC-associated microbial changes. Acute colitis raised the proportion of the Proteobacteria phylum and lowered that of the Verrucomicrobia phylum, whereas chronic colitis had the opposite effect. Proteobacteria is enriched in UC patients, which correlates with the significant reduction in intestinal SCFAs content (especially in the active phase) [25,26]. In addition, the Verrucomicrobia phylum includes important SCFAs-producing bacteria such as Ruminococcus and Akkermansia. The abundance of the Bacteroidaceae and Helicobacteraceae families had increased, while that of Lachnospiraceae and Ruminococcaceae decreased in mice with acute and chronic colitis. The Helicobacteraceae family consists of conditional pathogens [27,28], while Lachnospiraceae and Ruminococcaceae include SCFAs-producing bacteria that are negatively correlated with colitis [29][30][31]. Enterobacteriaceae and Desulfovibrionaceae were abundant in mice with acute colitis, while Rikenellaceae, Akkermansiaceae and Prevotellaceae showed lower abundance. Excessive proliferation of Enterobacteriaceae has been observed in the gut of UC patients [32,33]. Surprisingly, the SCFAs-producing Rikenellaceae, Akkermansiaceae and Prevotellaceae were raised in mice with chronic colitis, which is contradictory to existing reports. Furthermore, both acute and chronic colitis were associated with a higher abundance of opportunistic pathogens (Bacteroides, Escherichia-Shigella, Helicobacter, Bacte-roides_acidifaciens, Bacteroides_massiliensis_dnLKV3 and Bacteroides_caecimuris), and lower abundance of probiotics/SCFAs-producing bacteria (Lachnospiraceae_ NK4A136_group, Alistipes, Butyricicoccus, Rumino-coccaceae_NK4A214_group, Lactobacillus_gasseri and Lactobacillus_reuteri). Interestingly, the inhibition of SCFAs-producing bacteria was significantly stronger in mice with acute colitis compared to the chronic colitis model, especially for Rikenellaceae, Ruminococcaceae, Ruminiclostridium, Alistipes and Clostridium_leptum. These differences can be attributed to the longer modeling duration of chronic colitis as well as the higher concentration of DSS used to simulate acute colitis. High levels of DSS decreased food intake and resulted in more severe diarrhea and bloody stool symptoms, which lowered the impact of diet and fecal preference on the gut microbiota. Furthermore, the local thickening of the colonic mucosa during chronic colitis likely increased the proliferation of Akkermansia_Muciniphila, a myxotrophin bacterium related to mucous production and intestinal metabolism [34]. This organism is linked to a number of chronic conditions including diabetes and non-alcoholic steatohepatitis, and may be responsible for the aberrant mucin secretion during chronic colitis. Finally, Rikenellaceae, Ruminococcaceae, Ruminiclostridium, Alistipes and Clostridium_leptum thrive in an anaerobic environment, which is possibly disrupted during the induction of acute colitis, resulting in lower proliferation of these bacteria.
However, several limitations of the present study should be noted. First, the analysis of the acute colitis models can be performed within 1-2 weeks, whereas that of the chronic models may take 2-4 months [9]. In the present study, in order to ensure the survival rate of mice and effects of prolonged feeding on gut microbiota, we modeled colitis over a shorter period. Second, we did not assess the impact of different DSS concentrations and administration methods (such as oral gavage) on the intestinal microbiota. Third, we did not compare the effects of acute/chronic DSS exposure in different mouse strains (such as C57BL/6). Finally, the microbiota research method used in this study is relatively simple, and network analysis [23] of differential OTUs may adapt to reveal the complex relationships among microbes.
Unfortunately, due to the small sample size of this study and our limited technical expertise in bioinformatics, we did not perform network analysis. However, these issues will be addressed in our subsequent study.
To summarize, both acute and chronic colitis stimulated changes in the gut microbiota, and the alterations in the acute colitis model are more consistent with that observed in UC patients. Furthermore, acute colitis inhibited the SCFAs-producing bacteria to a significantly greater extent compared to chronic colitis. Finally, younger age and shorter exposure to DSS are more conducive to the stability of the colitis microbiota in the mouse model. Our study provides an experimental basis for the analysis of colitis microbiota, and clearly shows that the acute colitis model simulates UC to a greater extent compared to the chronic colitis model.

Conclusions
DSS-induced acute and chronic colitis demonstrated similar symptoms and histopathological changes. However, the alterations in the acute colitis model are more consistent with that observed in UC patients than chronic colitis models. Acute colitis induced higher abundance of SCFAs-producing bacteria and lower α-diversity compared to chronic colitis.

Establishment of DSS-induced colitis and treatment
This study was carried out in compliance with the ARRIVE guidelines. All animal experiment protocols were approved by the Animal Ethics Committee of Guangzhou First People's Hospital (Approval number: 2017-202). Male BALB/c mice (n = 24, aged 6-8 weeks) were purchased from Guangdong Medical Laboratory Animal Center (Foshan, China, Certificate number SYXK 2013-0002). Animals were reared in a specific pathogen-free (SPF) environment at 24 °C under a 12-h light/dark cycle and relative humidity 50-70%. Mice were allowed free access to food and water. The mice were randomly divided into the acute blank (A_BLK), acute DSS-induced colitis (A_DSS), chronic blank (C_BLK) and chronic DSS-induced colitis (C_DSS) groups (n = 6 each). The control groups received regular drinking water for 1 week and 5 weeks respectively. Mice were stimulated to develop acute colitis through the addition of 3% DSS (MP Biomedicals, USA) into their drinking water for a week. Chronic colitis was induced by supplementing their drinking water with 2.5% DSS every other week for 5 weeks [8,35]. Daily assessments of mice body weights, stool consistency, as well as the presence of blood-stained stool and anus were carried out. Mice were anesthetized with pentobarbital sodium and euthanized via transcardiac perfusion at the end of the experiments, and their colons were harvested, measured, and fixed overnight in 10% neutral buffered formalin. Stool samples were also collected from the cecum.

Extraction of fecal DNA and analysis of gut microbiota
The HiPure Stool DNA Kit (Magen, Guangzhou, China) was used to extract total fecal DNA based on instructions provided by the manufacturer. The following primers were used to amplify the 16S rDNA target regions: 341F: CCT ACG GGNGGC WGC AG; 806R: GGA CTA CHVGGG TAT CTAAT [39]. Amplified products of 400-450 bpm were purified using the Phusion High-Fidelity PCR Master Mix (New England Biolabs, Beverly, MA, USA). Sequencing libraries were generated using the TruSeq DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, CA, USA) based on protocols stipulated by the manufacturer before adding in index codes. The quality of the constructed library was evaluated using the Agilent Bioanalyzer 2100 system and Qubit@ 2.0 Fluorometer (Thermo Scientific, Carlsbad, CA, USA), an the Illumina HiSeq 2500 platform (Tianjin Novogene Bioinformatics Technology Co., Ltd) was used for sequencing [39].

Bioinformatics analyses Quality control and reads assembly
Reads filtering Raw data containing adapters or lowquality reads were further filtered according to the FASTP (version 0.18.0) criteria: 1) containing more than 10% of unknown nucleotides (N) and 2) containing less than 50% of bases with quality (Q-value)>20.
Reads assembly Paired end clean reads were merged as raw tags using FLSAH (version 1.2.11) with a minimum overlap of 10 bp and mismatch error rates of 2%.
Raw tag filtering Noisy sequences of raw tags were filtered by QIIME (version 1.9.1) pipeline under specific filtering conditions [40] to obtain the high-quality clean tags. The filtering conditions are as follows: 1) Break raw tags from the first low quality base site where the number of bases in the continuous low-quality value (the default quality threshold is <=3) reaches the set length (the default length is 3), and 2) filter tags with high-quality base length is less than 75% of the tag length.
In total, we obtained 1,344,698 tags without primers, with 56,029 tags per sample on average. [Link for the NCBI dataset: http:// www. ncbi. nlm. nih. gov/ biopr oject/ 738354]. α-Diversity analysis QIIME (version 1.9.1) was used to derive the α-diversity indices [41]. The ggplot2 package of the R project (version 2.2.1) was used to perform OTU rarefaction and plot rank abundance curves. Various α-indices (Pielou index and Shannon index) between two groups were calculated using the Wilcoxon rank test in the R project Vegan package (version 2.5.3). The Tukey's HSD test was used to evaluate α-indices (Pielou index and Shannon index) among three or more groups.

Operational Taxonomic Units (OTUs) analysis and β-diversity analysis
Effective tags of OTUs ≥ 97% were clustered together with using UPARSE software (version 9.2.64) [42]. Each cluster was represented by the most abundant tag sequence of each cluster. The VennDiagram package [43] of the R project (version 1.6.16) was used to analyze element distribution, while the Vegan package in the R project (version 2.5.3) was utilized for principal component analysis (PCA). Relevant sequences were aligned with the software Muscle [44] (version 3.8.31) and clustering trees were constructed using the FastTree [45] (version 2.1) and UPGMA (Unweighted Pair Group Method with Arithmetic Mean) software. Multivariate statistical techniques including PCoA (principal coordinates analysis) [46] and NMDS (non-metric multi-dimensional scaling) [47] of Bray-Curtis distances were calculated using the Vegan package (version 2.5.3) and plotted using ggplot2 package (version 2.2.1).

Welch's t test
Welch's t test was used to evaluate significant variances in terms of mean abundance of species between two groups. When there are two groups in the comparison group and total repeat samples in each group ≥3, the number of species tags/total tags of at least one sample ≥ 0.1%, and the top 1000 high abundance OTUs are statistically tested with R language vegan data package.

Statistical analysis
Data is depicted in terms of percentage or the mean ± standard deviation (SD). Inter-and intra-group differences were analyzed by the Wilcoxon signed rank test or unpaired t-test, one-way ANOVA and post hoc Tukey's test as appropriate. The SPSS software (version 23.0; IBM Corp.) was utilized for all statistical analyses. Statistical significance was granted when the p values were < 0.05.