A High-Fish Oil Diet Can Signicantly Reshape The Gut Microbiota In Mice

Background: Gut microbiota plays an essential role for human health and recent evidence has revealed the benecial effects of sh oil supplements on the gut microbiota. The present study was to investigate the inuence of sh oil on diet-based gut microbiota changes and colitis in mice and whether pyroptosis plays a role in this process. Results: A high-sh oil diet alleviated colitis, resulted in less weight loss and improved pathological scores. Caspase-1, activated in the dextran sulphate sodium (DSS) group, was suppressed by a high-sh oil diet. AIN-93M signicantly decreased the gut microbial diversity of mice, increasing the abundances of Bacteroides and Parabacteroides and decreasing the abundance of Odoribacter. In contrast, gut microbial diversity was maintained in mice fed a high-sh oil diet; the Firmicutes: Bacteroidetes ratio was increased, the abundance of Parabacteroides was increased, and that the abundance of Odoribacter was decreased. Conclusion: AIN-93M can decrease gut microbiota diversity, which may be associated with a potential proinammatory effect. Fish oil has anti-inammatory effects. It can also restore and maintain microbial diversity and suppress pyroptosis activation.


Background
The gut microbiota plays important roles in metabolism and immunity. 1 This microbial community is shaped from infancy and can be greatly in uenced by several factors 2 , among which diet is one of the most important. Many studies have shown that changes in macronutrients can lead to dramatic changes in the gut microbiota. 3 Fish oil is regarded as a bene cial food supplement, with documented bene ts on cardiac vascular health. The primary active ingredient of sh oil is ω-3 polyunsaturated fatty acids (PUFAs), including eicosapentaenoic acid [EPA, 20:5 (n = 3)] and docosahexaenoic acid [DHA, 22:6 (n- 3)]. Studies have also demonstrated the bene cial immunomodulatory effects of ω-3 PUFAs, and the ratio of ω-3 and ω-6 PUFAs was shown to play an important role in in ammation by in uencing eicosanoid metabolism. 4

ω-6
PUFAs are metabolized into several proin ammatory mediators, while ω-3 PUFAs function as competitive inhibitors of enzymes and can be metabolized into several anti-in ammatory mediators. 4 Few studies have investigated the in uence of sh oil on the gut microbiota, and the results are controversial, especially the observed changes in the abundance of different bacteria. Multiple studies have demonstrated that ω-3 PUFAs can increase the abundances of Lactobacillus and Bi dobacteria, [5][6][7][8][9] which are thought to have a bene cial in uence on health. Although some studies have shown that ω-3 PUFAs can increase the abundance of Bacteroidetes and decrease the Firmicutes: Bacteroidetes ratio, 8,10,11 the results of another study supported the opposite conclusion. 12 Pyroptosis, a form of programmed necrosis, has emerged as a general innate immune effector mechanism. Studies [13][14][15][16] have found that pyroptosis is essential for the development of in ammatory bowel disease (IBD). Two pathways were found to mediate pyroptosis. Procaspase-1 is activated by various in ammasome complexes, and Procaspase-4/5/11 is also activated by cytosolic lipopolysaccharide. Activated Caspase-1 and Caspase-4/5/11 cleave Gasdermin D, which can form pores on the plasma membrane, inducing pyroptosis. IL-1β, an in ammatory mediator, was also activated by Caspase-1. 17 Studies 18 found crosstalk between pyroptosis and the gut microbiota, suggesting that pyroptosis plays an essential role in maintaining gut homeostasis. However, few studies have investigated the relationship between sh oil, pyroptosis and the microbiota.
In the present study, we administered dietary interventions to mice to assess the in uence of sh oil on the gut microbiota and colitis and its modulation of pyroptosis.

Fish oil alleviates DSS-induced colitis
Mice fed with DSS developed colitis with diarrhea, bloody stool, weight loss and decreased activity from the 3 rd day. Death started to occur from the 4 th day. Mice from the CD and FD groups showed signi cant more weight loss than the CC and FC groups. (CD vs CC, P<0.05 day 4-7; FD vs FC, P<0.05 day 5). The CD group showed signi cantly more weight loss than the FD group. (CD vs FD, P<0.05 day 4-5) (Fig. 1B).
There were no deaths in the CC and FC groups, and the death rates of the CD and FD groups were 50% (6/12) and 33.3% (2/6), respectively (Fig. 1A). Colons from animals in the CC and FC groups were signi cantly longer than those from animals in the CD and FD groups (CC vs CD, P=0.133; FC vs FD, P=0.013). There was no signi cant difference in colon length between the CD and FD groups (P=0.451) (Fig. 1C). Regarding pathology, the colons from the FC and CC groups had normal mucosal structures with intact crypts. Neither mucosal damage nor in ltration of in ammatory cells was found. For Colon from the DSS groups, colitis of various degrees was observed. In most severe cases, the normal mucosal structure of the anal side disappeared, the crypt structure was completely destroyed, and only some epithelial cells were retained. In ammation involves the mucosal and submucosal layers, with extensive in ltration of in ammation cells. (Fig. 2) Both the CD and FD groups had signi cantly higher pathological scores than the CC and FC groups (CD vs CC, P 0.001; FD vs FC, P=0.008). Moreover, the pathological score of the FD group was signi cantly lower than that of the CD group (P<0.01). (Fig. 1D) Fecal microbiota changes associated with diet.
A comparison of the whole microbiome using Analysis of similarities (ANOSIM) revealed no signi cant differences between the sh oil and control groups before intervention (R=0.1426, P=0.081, FB vs CB). Both the high-sh oil diet and AIN-93M had a signi cant in uence on the microbiota (R=1, P=0.003, CA vs CB; R=0.9185, P=0.001, FA vs FB). Whereas, here was a signi cant difference between the sh oil and control groups after intervention (R=0.8926, P=0.002, FA vs CA) (Fig. 3A).
The alpha diversity tests using the Shannon index with the Wilcoxon signed-rank test showed no difference in microbial diversity between the sh oil and control groups before intervention (P=0.1989, FB vs CB). Compared with baseline, the AIN-93M intervention signi cantly decreased the microbial diversity of the fecal microbiota (P=0.0082, CA vs CB). The high-sh oil diet, on the other hand, had no signi cant in uence on microbial diversity (P=0.7024, FA vs FB) (Fig. 3B).
The MetaStat analysis results showed no signi cant differences in the abundances of Firmicutes and Bacteroidetes, which were the most abundant phyla (abundance>0.1), in the control group. Compared with baseline, the high-sh oil diet intervention signi cantly reduced the abundance of Bacteroidetes and increased Firmicutes (Fig. 3C 3E). Bacteroides, Parabacteroides, Alistipes and Odoribacter were the most abundant genera (abundance>0.01). AIN-93M signi cantly increased the abundances of Bacteroides and Parabacteroides and decreased that of Odoribacter. The high-sh oil diet, on the other hand, increased the abundance of Parabacteroides and decreased that of Odoribacter but had no in uence on the abundance of Bacteroides (Fig. 3D) Pyroptosis Caspase-1, IL-1β and Gasdermin D expression levels were measured using western blotting. No activation of Caspase-1, IL-1β or Gasdermin D was found in the CC, FC and FD groups. In the CD group, no activated forms of Caspase-1 and IL-1β were detected. However, Gasdermin D was activated in the CD group (p=0.0159).

Discussion
A Western-style diet may contribute to the pathogenesis of chronic intestinal diseases, such as in ammatory bowel disease, but the associated mechanisms remain unelucidated. Diet may in uence host health by directly affecting nutrient digestion and absorption or by in uencing the gut microbiota. The relative de ciency of ω-3 PUFAs in the Western diet has been hypothesized to be one of the potential contributors to its adverse effects. 19 With increasing evidence showing that the gut microbiota has an important impact on human health, 1 it is possible that the disruption of the gut microbiota may promote nutrient de ciency with respect to host health. In this study, we observed that diet plays an essential role in host in ammation and gut microbiota, having a large in uence on the structure and diversity of the gut microbial community. By supplementing the diets of mice with ω-3 PUFAs, in ammation was alleviated, the diversity of the microbiota was restored, and its structure was shifted toward an increased Firmicutes: Bacteroidetes ratio.
The results of our study showed that compared to a standard diet, AIN-93M can greatly in uence the gut microbiota. AIN-93 was developed by the American Institute of Nutrition Rodent Diets in 1993 20 . In contrast to the natural diets of mice, the composition of AIN-93 is precise, making it a good tool to study the effects of speci c diet compositions and their in uence on the gut microbiota. However, the results of the present study showed that AIN-93 signi cantly in uences the gut microbiota of mice by reducing species diversity. Feeding AIN-93M mice can increase the abundance of Bacteroides, which was shown to be signi cantly more abundant in IBD patients than in healthy controls. 21 Thus, a better chow with an exact composition and little in uence on the normal microbiota of mice is needed for studies on the interaction between speci c diet ingredients and the microbiota.
As previously reported, the in uence of sh oil on the gut microbiota has been debated and requires further investigation. In our study, mice fed with a high-sh oil diet were observed to have a gut microbiota with greater species diversity, an increased Firmicutes abundance and decreased Bacteroidetes abundance, thereby increasing the Firmicutes: Bacteroidetes ratio. Several studies 5-9, 22,23 have shown that ω-3 PUFAs can in uence speci c probiotic bacteria and pathogens, thereby affecting in ammation of the gut. In our study, we showed that ω-3 PUFAs can reverse the increase in the abundance of Bacteroides, which is thought to be a proin ammatory genus. Mice fed with a diet rich in ω-3 PUFAs showed reduced in ammation, indicating that ω-3 PUFAs may function as an antiin ammatory component by modifying the structure of the gut microbiota.
Pyroptosis plays an important role in innate immunity. Cell experiments 24 have found that ω-3 PUFAs suppress pyroptosis by suppressing NLPR3 and activating Caspase-1. In our study, however, we found that Gasdermin D was activated, but not Caspase-1 and IL-1β. Research 25 found that DSS failed to induce colitis in germ-free mice, hinting at the importance of microbiota in the development of colitis. According to the results of our study, DSS may promote pyroptosis by activating Caspase-4/5/11, which are activated by LPS. ω-3 PUFA intervention resulted in an increased Firmicutes: Bacteroidetes ratio and decreased LPS, thus suppressing pyroptosis, which may be a possible mechanism of its antiin ammatory function.
There are several limitations of this study. The sample size was limited, and additional mice in each group may provide a better understanding of the differences between the groups. Furthermore, we analyzed the fecal microbiota of mice, which can be obtained easily and repeatedly but may not be representative of the microbiota in the host. 26 Further study to assess the microbiota directly from the intestine may allow for a better understanding of the interaction between the host and microbiota. Our study focused on the in uence of sh oil on the microbiota in mice, and additional studies focusing on the associated metabolic and in ammatory pathways would be of great interest.

Conclusions
In summary, in the present study, we showed that AIN-93M had a large in uence on the fecal microbiota of mice, which is believed to be the result of proin ammatory effects. A high-sh oil diet could reverse this effect, maintain a healthy microbiota and alleviate in ammation of the colon. To better understand speci c components of diet and its in uence on the gut microbiota, a standard diet with de ned components and little in uence on the gut microbiota is needed for further studies. The results indicated that a high-sh oil diet would be a better choice for further study than AIN-93M. Furthermore, sh oil has the potential to be a prebiotic, maintaining microbial diversity and inhibiting proin ammatory microbiota.

Study objectives
Thirty male C57BL/6N mice (eight weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd. The mice were maintained in the Institute of Laboratory Animal Science, Chinese  Table 1. Experimental design Male C57BL/6N mice were randomly assigned to two groups using a computer based random order generator: the control (C,n = 18) and sh oil (F, n = 12) groups. All mice were fed with regular diets (provided by Beijing Vital River Laboratory Animal Technology Co., Ltd.) and puri ed water after birth until the dietary intervention. The control group was fed the AIN-93M diet and puri ed water, and the sh oil group was fed the customized diet and puri ed water. The dietary intervention duration was 3 weeks. Feces were collected from each mouse (except 6 from the DSS control group to achieve equal sample size for microbiota analysis) before and at the end of the intervention (Fig. 5). Feces were collected immediately after they were produced and stored at -80°C.
After two weeks of dietary intervention, the mice were then randomly assigned into four subgroups using the same random order generator: control group without DSS (CC n = 6), control group with DSS (CD n = 12), sh oil group without DSS (FC n = 6) and sh oil group with DSS (FD n = 6). For groups with DSS, 2.5% dextran sodium sulfate (DSS, supplied from MP Biomedicals, LLC) was added to puri ed water for 5 days to induce acute colitis. The weight of each mouse was measured every day. Seven days after DSS initiation, the mice were anesthetized with 4% chloral hydrate (10 µL/g, through intraperitoneal injection) and sacri ced by breaking the neck. The whole colon was collected for further study.
Due to different appearance of feed and water, the experiment could neither be blinded to whether the animal was fed AIN-93M or customized diet, nor to whether the animal was fed puri ed water or puri ed water dissolved with DSS.

Mouse colon in ammation assessment
The length of the whole colon of each mouse was measured. The colon was then cut in half along the long axis. Half of the sample was taken to the Peking Union Medical College Hospital Pathology Department for specimen preparation and HE staining. 27 The pathological score 28 was calculated by a pathological specialist who was not a member of our research group and was blinded to the intervention of mice.

Measurement of pyroptosis activation in colonic tissue
The other half of the colon was washed with cold phosphate-buffered saline and stored at -80°C until use. Colonic tissue was homogenized using RIPA lysis buffer (Applygen Technologies Inc.). Tissue homogenates were transferred to microcentrifuge tubes, vortexed, and centrifuged at 4°C for 10 min at 12,000×g. Two microliters of supernatant from each sample was withdrawn to assess the protein concentration using a Coomassie brilliant blue kit (Applygen Technologies Inc.) according to the manufacturers' protocol. Protein samples were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The activation of Caspase-1, IL-1β and Gasdermin D was assessed using antibodies from Abcam (ab179515, ab9722 and ab209845).

Microbiota diversity analysis
Total genomic DNA was extracted from each fecal sample using the sodium dodecyl sulfate (SDS) method and 1% agarose gels. The V4 region of the 16S rRNA gene was PCR ampli ed using the 515F (GTGCCAGCMGCCGCGGTAA) and 806R (GGACTACHVGGGTWTCTAAT) primers with 15 µL of Phusion® High-Fidelity PCR Master Mix (New England Biolabs), 0.2 µM primers and 10 ng of template DNA. The thermal cycling conditions were and initial denaturation step at 98℃ for 1 min; 30 cycles of denaturation at 98℃ for 10 s, annealing at 50℃ for 30 s and elongation at 72℃ for 30 s; and nally 72℃ for 5 min. Finally, sequencing libraries were generated using a TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA) following the manufacturer's recommendations, and index codes were added. The library quality was assessed using a Qubit@ 2.0 Fluorometer (Thermo Scienti c) and an Agilent Bioanalyzer 2100 system, and then, the samples were sequenced on an Illumina NovaSeq platform to generate 250 bp paired-end reads.

Bioinformatics analysis
Paired-end reads were merged using FLASH (V 1.2.7). The data were then ltered to obtain clean, highquality tags according to the QIIME (V 1.9.1) quality-control process. The tags were compared with the reference database (Silva Database) using the UCHIME algorithm to remove chimeric sequences. Availability of data and materials The datasets generated and/or analysed during the current study are available in the repository of Novogene Co. Ltd. And the project number is P101SC18051695-02-F002-B1-41.

Competing interests
The authors declare that they have no competing interests
Authors' contributions HS performed the animal experiment and was one of the major contributors in writing the manuscript. DC and BT helped modi ed the manuscript. All authors read and approved the nal manuscript.    Flow chart of the experiment. A total of 30 male C57/BL6N mice were randomly divided into a control group (N=18) fed AIN-93M and a sh oil group (N=12) fed a high-sh oil diet. Feces were collected from each mouse on day 0 and day 14 except for those in the CD group, in which 6 out of 12 mice were randomly selected for fecal collection for statistical analysis reasons. The mice were then subdivided into CC, CD, FC and FD groups. For the CD and FD groups, DSS was administered for 5 days to induce colitis. On day 21, the mice were sacri ced for further analysis.