Fecal microbiota transplantation reconstructs the gut microbiota of septic mice and protects the intestinal mucosal barrier

The gastrointestinal (GI) tract has long been hypothesized to play an integral role in the pathophysiology of sepsis, and gut microbiota (GM) dysbiosis may be the key factor. Previous studies has confirmed that microbiome is markedly altered in critical illness. We aimed to confirm the existence of gut microbiota imbalance in the early stage of sepsis, observe the effect of fecal microbiota transplantation (FMT) on sepsis, and explore whether FMT can reconstruct the GM of septic mice and restore its protective function on the intestinal mucosal barrier. Through the study of flora, mucus layer, tight junction, immune barrier, and short-chain fatty acid changes in septic mice and fecal microbiota transplanted mice, we found that GM imbalance exists early in sepsis. FMT can improve morbidity and effectively reduce mortality in septic mice. After the fecal bacteria were transplanted, the abundance and diversity of the gut flora were restored, and the microbial characteristics of the donors changed. FMT can effectively reduce epithelial cell apoptosis, improve the composition of the mucus layer, upregulate the expression of tight junction proteins, and reduce intestinal permeability and the inflammatory response, thus protecting the intestinal barrier function. After FMT, Lachnospiraceae contributes the most to intestinal protection through enhancement of the L-lysine fermentation pathway, resulting in the production of acetate and butanoate, and may be the key bacteria for short-chain fatty acid metabolism and FMT success.


Introduction
Sepsis continues to be the leading cause of mortality in the intensive care unit 1, 2 . There were more than 1 million sepsis-related deaths in 2015 in China 3 . In 2017, an estimated 48.9 million incident cases of sepsis were recorded worldwide, with 11.0 million sepsis-related deaths reported, comprising 19.7% of global deaths 4

. The World Health
Organization has recognized sepsis as a global health priority 5 . Despite significant advancement in our understanding of the pathophysiology of sepsis, treatment of sepsis is limited to antibiotics, aggressive fluid resuscitation, vasopressor administration, and supportive care, and no targeted therapeutics for sepsis are approved for usage in patients 6 .
Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection [5][6][7] . The syndrome can be induced by a wide variety of microbes and, by definition, involves a maladaptive response to a pathogen 5 . The gastrointestinal (GI) tract has long been hypothesized to play an integral role in the pathophysiology of sepsis, by both driving and perpetuating multiple organ dysfunction 8,9 . The original concept of gut-derived sepsis proposed that the altered inflammatory milieu induced by overwhelming infection leads to intestinal hyperpermeability, allowing luminal contents, including intact microbes, and microbial products, to escape their natural environment where they can cause either local or distant injury 1,5,6,10 .
The gut microbiome plays a crucial role not only in GI health but also in overall immune development and host health. The gut microbiome contains 40 trillion microorganisms, a similar number of cells as found in the human host. This includes up to 1000 different microbial species and 100 times more bacterial genes than human genes 6,[10][11][12][13] . The symbiotic relationship between microbiota and the host is mutually beneficial 14 . The host provides an important habitat and nutrients for the microbiome, and the gut microbiota (GM) support the development of the metabolic system and the maturation of the intestinal immune system by providing beneficial nutrients, for example, by the synthesis of vitamins and short-chain fatty acids (SCFAs) 15 . The destruction of the balance between the gut microbiome and the host is called dysbiosis. The microbiome is markedly altered in critical illness. Microbial diversity is diminished within 6 h of admission to the intensive care unit, and this lack of diversity has been associated with poor outcomes in critically ill patients 5 .
Considering GM dysbiosis is one of the most important factors that can lead to pathological bacterial translocation and systemic infection, it may be feasible to develop novel therapeutic strategies against gut-derived sepsis by modulating the microbiota 16 . More than 90% of the commensal organisms may be lost during the early stage of critical illness, making it nearly impossible that a single or several probiotic species would be able to completely replenish the diversity of the GM without intervention. Transfer of healthy donor feces containing thousands of microbial species, termed FMT, facilitates replenishment of diminished commensal bacteria and may guide the patient's microbiota toward a healthy state 16 . Fecal microbiota transplantation has been demonstrated to be remarkably successful in the treatment of recurrent Clostridium difficile infection, with a 92% response rate to treatment, and is also increasingly being used for dysbiosis caused by other intestinal pathologies, such as inflammatory bowel disease 12,17 . Yet, FMT is scarcely used in the treatment of septic patients, as antibiotic therapy is frequently used in these cases and continuation of FMT therapy could adversely influence the remodeling of intestinal microbiota 16 . Recently, the use of FMT has been reported in septic patients with MODS and non-C.difficile diarrhea, presenting as refractory to standard medical management. At 2-3 weeks post-FMT, the patients showed resolution of their diarrhea and significant decreases in blood levels of inflammatory mediators, such as TNF-α, interleukin (IL)-1β, IL-6, and C-reactive protein 11,16 . Following FMT, stool microbiota in these patients showed marked alterations resembling the microbiota composition of the donors, with an increase in Firmicutes and a reduction in Proteobacteria. Even though the evidence is limited to a series of case reports, the improved clinical outcomes in these patients following FMT are promising 16,18 .
However, during sepsis, the exact mechanism of action for the use of FMT on the intestine is still unknown 19 .

Mice
The study included 6-8-week-old male clean-grade C57BL/6 mice weighing between 20-25g (Beijing Vital River Laboratory Animal Technology Co.). The variation in intestinal microbiota composition between inbred C57BL/6 mice derived from one commercial vendor is known to be highly limited 20 . Breeding conditions included a 12/12-hour light/dark cycle at a room temperature of 20-25℃. All mice were given free access to food and water. This experiment was approved by the Committee on the Ethics of Animal Experiments of the Hebei General Hospital.

Sepsis model
Sepsis was induced by cecal ligation and puncture (CLP) as a previously described 21,22 . Mice were anesthetized using an intraperitoneal injection at 50 mg/kg of 2% sodium pentobarbital. Briefly, a 1.5 cm midline laparotomy was performed under aseptic conditions to expose the cecum.
A single through-and-through puncture was performed by a 22-gauge needle between the ligation site and the end of the cecum, and then a small amount of fecal material was extruded through the puncture. The cecum was repositioned into the peritoneal cavity, and the peritoneum and skin were closed in layers. Sham mice were treated identically except the cecum was neither ligated nor punctured. All mice received 1.0 ml normal saline subcutaneously after the surgery to compensate for fluid loss. Mice were euthanized at 12, 24, and 48 h following CLP for acute studies (Sham: n = 6 per group; CLP: n = 10/11 per group; FMT: n = 10/11 per group).

Fecal microbiota transplantation
Fresh feces was collected from 10 healthy C57BL/6 mice, that were from the same strain of the recipient, homogenized in 10 ml of sterile phosphate-buffered saline (PBS) and centrifuged for 30 sec at 3,000 rpm, 4 ℃, to pellet the particulate matter. The optical density (OD) value of the supernatant slurry was checked to calculate the concentration of total bacteria (OD = 0.5 represents 10 8 cells ). For each mouse, 1×10 8 bacterial cells (1×10 9.8 bacterial cells represents the sum of the total bacterial population within 2 g of cecal contents) were centrifuged for 5 min at 12,000 rpm, at 4 ℃, and them bacterial pellets were resuspended in 0.2 ml PBS and gavaged into each mouse one day at a time 23,24 . For acute studies, mice in the FMT group received a single dose of fecal microbiota just prior to cecal ligation and puncture for the first day 22 . The CLP and Sham group mice were gavaged with 0.2 ml PBS one day at a time. All mice were treated at the same time. For the survival studies, mice in the FMT group were treated with fecal microbiota once a day for the first 3 days, and the other mice were treated with PBS daily for 3 days.

Sample processing for animal experiments
After the mice were sacrificed, their GI tracts were quickly removed. The colons were gently separated, by cutting at the cecum-colon junction and rectum, and immediately preserved in Carnoy's fixative (dry methanol: chloroform: glacial acetic acid in the ratio 60:30:10) 25,26 . The Carnoy's fixative was made fresh with anhydrous methanol, chloroform and glacial acetic acid. The colons were fixed in Carnoy's solution for 3 h followed by transfer to fresh Carnoy's solution for 2-3 h. The colons were then washed in dry methanol for 2 h, placed in cassettes and stored in fresh dry methanol at 4℃ until further use. Cecal contents from each animal were divided into replicates, and each distal end was instantly flash-frozen in liquid nitrogen and then stored at -80℃ until further use.

Morphological examination and histological analysis
The colonic sections were stained with hematoxylin/eosin (HE) and examined under a phase-contrast microscope for morphological characteristics. The histological damage was scored using previously published criteria 23 , including extent of destruction of normal epithelial architecture, presence and degree of inflammatory infiltration, presence of edema, extent of vascular dilatation and congestion, and presence or absence of goblet cell depletion, presence or absence of crypt abscesses.
To observe the thickness of the mucus layer, we stained with Alcian, and observed under a DP80 microscope (Olympus, Japan).

Immunohistochemistry and immunofluorescence
Samples were fixed in Carnoy's fixative, embedded in paraffin, and cut into sections (5m thickness). Sections were mounted onto polylysine-coated slides, deparaffinized, rehydrated and placed in a 3% citrate buffer to repair antigens. After being pretreated with 3% H2O2 for Samples were cut into 60-80 nm sections, and stained with uranyl acetate and lead citrate. The sections were analyzed by electronic microscopy (HT7700 TEM; Hitachi Inc., Tokyo, Japan ).
Selected blots were quantified with ImageJ software.

Real-time PCR
Total RNA was isolated from colonic tissue (that had been snap-frozen in

Statistical analysis
The Kaplan-Meier estimator was used to draw the survival curve of the mice, and the log-rank method was used to compare the survival rates between different groups. The measurement data had a normal distribution, the variance was uniform, and the one-way ANOVA and LSD tests were used for comparison between multiple groups. The Student's T-test was used to compare the two independent groups, and the measurement result was expressed as the mean ± standard deviation ( mean ± sd. ). The Kruskal-Wallis H test was used for data with non-normal distribution and/or uneven variance. The results were expressed in medians (interquartile range). SPSS 21.0, GraphPad Prism 8.0, Photoshop CS5, Image-Pro Plus, and ImageJ were used for data analysis, and P < 0.05 was considered statistically significant.

Mortality among three groups
The survival rates of the Sham group, CLP group, and FMT group were compared 7 days following sepsis modeling. The CLP group had a mortality rate of 30% at 24 h and a 50% survival rate at 7 days. There were no deaths at 24 h in the FMT group, and a 90% survival at 7 days, and no deaths in the Sham group. Compared with the Sham group and the FMT group, the mortality of the CLP group was significantly higher, P < 0.05. (Figure 1. I).

Colonic pathology score and apoptosis
Histological scores were performed by professionals who were unfamiliar with our experiment according to the standards proposed by Li et al 23 .
The colonic pathology score of the CLP group was significantly higher than the scores of both the Sham and FMT groups at 12, 24, or 48 h, P < 0.0001. The colonic pathology score of the FMT group was higher than that of the Sham group at 12 h, P < 0.004, but there was no significant difference between the Sham and FMT groups at 24 and 48 h ( Figure   1.II). Examination of colonic pathology showed the damage in the FMT group was -significantly less than that in the CLP group, so we next  (IV) Relative expression of caspase 3 compared with GAPDH between the CLP and FMT groups at 24 and 48 h. * P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns -no statistical difference, respectively.

Mouse serum inflammatory factors TNF-α, IL-6, and IL-10.
The concentration in pg/mL of serum IL-6 (91.62 ± 25.53) and IL-10 There was no significant difference among the three groups in the IL-10 level at 48 h.
All the results are shown in Table 1, Table 2, and Table 3.   The expression of IL-6 in the CLP group was the highest at 24 h after modeling, and then decreased, but it was still higher than that in the Sham group at 48 h. Expression of IL-6 in the FMT group was slightly higher than that in the CLP group at 12 h, but there was no significant difference, to the contrary, IL-6 levels were markedly lower than the CLP group at 24 and 48 h. The TNF-α level in the CLP group continued to increase, and it was the highest among the three groups at 48 h, however, the IL-10 level was lowest at 24 h after modeling ( Figure 2 ).

The thickness of the mucus layer (nm) and MUC2 expression
The AB-PAS method was used to detect the colonic mucus layer thickness at 12, 24, and 48 h after sepsis modeling in the three groups.  Figure 3. II).     and NF-κB levels in the colon. *P < 0.05, **P < 0.01.

16SrRNA sequence analysis
Firmicutes and Bacteroidetes were the main bacteria, accounting for 55% and 33% of the abundance found in the normal mice, respectively.  IV). The core of the KEGG database is a biological metabolic pathway analysis database, in which metabolic pathways are classified into six categories, including metabolism, genetic information processing, environmental information processing, cellular processes, organismal systems, and human diseases. We found a higher relative abundance in infectious diseases. (Figure 7. V). Therefore, we further analyzed the species composition of the differential pathways, and found that the Lachnospiraceae contributed the most to L-lysine fermentation to acetate and butanoate (Figure 7. VI).

Analysis of short-chain fatty acids (SCFAs) by LC-MS
We detected the content of short-chain fatty acids, including acetate, propionate, isobutyrate, butyrate, valerate, isovalerate, caproate, and heptanoic acid in the CLP group (CD) and the FMT group (TD) at 48 h.
The first four acids were the most important SCFAs, therefore, we used these four major short-chain fatty acids for PCA analysis. (n = 7 per group) (Figure 8.). That ZO-1 outperformed other TJ markers may reflect the concomitant organ epithelial injury which occurs in MODS 36 . Therefore, we observed the expression of the above two proteins at 24 and 48 h in the three groups, and the results showed that occludin and ZO-1 in the CLP group were significantly lower than those in the Sham group and the FMT group. Levels of the two indicators in the FMT group were higher than those in the CLP group, or they were similar to those in the Sham group.
The intestinal mucus layer is divided into two layers, a dense inner layer and a loose outer layer 26,43 . The mucous gel is formed by high molecular-weight mucins secreted by the goblet cells and the outer mucus layer serves as a habitat for the resident microflora by providing a beneficial microenvironment 43,44 . The amount and composition of the mucus layer reflect the balance between mucus secretion, and its erosion and degradation by bacteria 45 . Research has shown that the thickness of the mucus layer is dependent on commensal bacteria 12 . We performed blinded thickness measurements of the colonic mucus layer in each mouse using Alcian blue-stained sections. We further validated the thickness of the mucus layer by immunofluorescence staining of the MUC2 mucins using a-MUC2 antibody 25 , and observed that the thickness of the mucus layer in septic mice was significantly reduced, and the thickness in the FMT group was significantly increased. This change was consistent with the changes in the flora of the two groups.
Although commensal bacteria reside innocuously in the gut mucosa, they share common microbe-associated molecular patterns (MAMPs) with the pathogenic bacteria that invade through the intestinal epithelium 46,47 . Commensal bacteria, therefore, have the potential to activate immune responses through pattern recognition receptors such as toll-like receptors (TLRs) and nucleotide-oligomerization domain (NOD)-like receptors (NLRs) 47 . Mild or physiologically acceptable levels of stimulatory signals provided by commensal flora are essential for the development and maintenance of an appropriate mucosal immune system, which provides the first line of defense together with the epithelial barrier system against undesired, foreign antigens 44 .
TLRs are type I transmembrane glycoproteins and serve as signal transduction receptors for innate immunity and inflammation 15 . TLR4 is the best-characterized pathogen-recognition receptor. Its downstream effects are varied, and the TLR/MyD88/p38 MAPK/NF-κB pathway is popularly believed to play a critical role in the inflammatory response 33,48 . A major pathophysiological mechanism of sepsis involves recruitment of inflammatory cells and generation of an overwhelming pro-inflammatory response. Cell-wall components from Gram-negative and Gram-positive bacteria activate PRRs such as TLR4 and TLR2, respectively, resulting in a "cytokine storm" of pro-inflammatory mediators generated mainly via the mitogen-activated protein kinase and NF-κB pathways 19 . In this study, the TLR4/MyD88/NF-κB pathway in septic mice increased significantly at both the protein level and the gene level, and the inflammatory factors TNF-α and IL-6 increased at different levels at 24 and 48 h after modeling, while IL-10 decreased at 24 and 48 h after modeling. In contrast, the FMT group showed an increase in inflammatory response inhibition. This is consistent with the pathological score of bowel injury in the three groups. It was observed that the amount of apoptosis protein caspase 3 was different between the CLP group and the FMT group, and the amount of apoptosis in the CLP group was significantly higher than that of the FMT group.
The GM support the development of the metabolic system and the maturation of the intestinal immune system by providing beneficial nutrients, by synthesizing vitamins and short-chain fatty acids (SCFAs) 15 .
Acetate, especially, is often considered to play a critical role in protection against pathogenic infection 16,49 . Studies have shown that SCFAs have anti-inflammatory and immunoregulatory activities and may reduce butyrate-producing bacteria such as Ruminococcaceae, Faecalibacterium, and Roseburia 13 . Therefore, we compared the relative abundance of Ruminococcaceae in each group, and the results suggest that fecal bacteria were significantly reduced in the sepsis model group, while in the FMT group, bacterial counts were similar to or even slightly higher than those found in normal mice. (In our study, the relative abundance of normal mice was 0.07%, 0.0025% in the CLP group and 0.067% in the FMT group at 48 h). It appears that Ruminococcaceae plays a role in the inflammatory response, and fecal microbiota transplantation can effectively improve the intestinal bacteria composition, thereby improving the inflammatory state. We found through functional prediction that the relative abundance of intestinal flora was significantly different in infectious diseases. Further analysis of the species composition of the different pathways revealed that Lachnospiraceae contributed the most to L-lysine fermentation to acetate and butanoate, consistent with previous studies 50 . This indicated that Lachnospiraceae may be the key bacteria for the effectiveness of FMT.
We further explored the changes of SCFAs within 48 h in the CLP group and the FMT group. We mainly studied acetate, propionate, isobutyrate and butyrate. After drawing the PCA chart, we found that there may be a different trend between the two groups, but further studies are needed.
This study has certain limitations. First, we selected the C57BL/6 mouse as our animal model, so it may be difficult to generalize our conclusions to humans. Second, we administrated FMT once a day, and the 7-day mortality observation group received FMT once a day for the first three days. We did not attempt to discern the complete reconstruction time of intestinal flora after initiating FMT, nor did we compare the frequency and number of fecal bacteria transplantation, but it appears that early application of FMT has a protective effect on the intestinal function of septic mice. Thirdly, this experimental study does not involve acquired immunity, a topic demanding further experimentation and verification. Fourth, the transmission electron microscope observed that the effect of FMT is not limited to the intercellular connection, but also involves mitigation of organelle damage. We did not explore the reason for this difference. Finally, the issue of metabolics needs deeper study.