Background
Anastomotic failure is one of the most serious complications of intestinal surgery [
24]. An anastomotic leak can lead to high morbidity, increased mortality, and considerable added hospital costs [
18]. Even with extensive research and surgical technique improvements, small intestinal and colorectal anastomosis leakage can occur in 0.05–30% of operations [
18,
24,
30,
31,
43]. Anastomotic leakage can vary in its onset of occurrence. An early leak within the first to second postoperative days occurs most often because of technical reasons, whereas a latent leak occurring by the end of the first postoperative week is most often attributed to a failure in the normal healing mechanisms [
43].
Proper intestinal healing requires a coordinated crosstalk between intestinal epithelial cells (IECs), immune cells, and microbiome. The multifaceted toll-like receptor (TLR) signaling pathways serve as an interface for concerted crosstalk between the intestinal epithelial barrier, microbiota, and immune system [
14]. The TLR signaling pathways help to maintain a constant homeostatic regulation of microbial load and the immune response. There are 10 total TLRs, and they are expressed on macrophages, dendritic cells (DCs), T lymphocytes, and IECs [
14]. TLR4 is expressed in the apical region of the terminal ileum of mice [
1]. TLR4 signaling is required for induction of mucosal healing pathways, the defense against gram-negative bacteria, and maintaining the tolerance to commensal bacteria. During mucosal injury, TLR4 activation on IECs leads to the release of inflammatory cytokines (i.e., IL-8) and chemokines that recruit innate and adaptive immune cells to limit bacterial invasion. TLR4 absence can lead to severe mucosal damage from a loss in epithelial proliferation, impaired inflammatory response, and bacterial translocation. However, TLR4 signaling needs to be tightly regulated as prolonged activation can result in detrimental inflammation that inhibits mucosal repair through decreased enterocyte proliferation and migration [
14].
Many immune cell types function alongside gut microbiota to promote mucosal healing in the steady state and during injury. These include CD103
+ conventional DCs (cDCs), Th17 cells, T regulatory (Treg) cells, iNKT, and iNKTh17 cells [
24]. Typically, the gut microbiota induces differentiation of these immune cells. Treg and iNKT cells will undergo differentiation and secrete the anti-inflammatory cytokine, IL-10, through interactions with
Clostridia [
25]. The promotion of CD4
+ T cells to Th17 cells can be induced by segmented filamentous bacteria (SFB) and Th17 and iNKTh17 cells will produce IL-17a, IL-17f, and IL-22, which is a key cytokine that acts on IL-22R only on epithelial cells in order to promote healing [
25]. However, there is a dichotomy to this system, if SFB colonization it too expressive, systemic Th17 activation will result in increased inflammation and mucosal injury [
25,
56]. Therefore, the IL-10 signaling of CD103
+ DCs and Tregs are key to helping to regulate the Th17 activation [
40]. Transforming growth factor-β drives the formation of Th17 cells from naïve T cells [
34]. Th17 cells are known to produce cytokines IL-17 and IL-22 which have several roles in the gastrointestinal tract, such as cell proliferation, tissue regeneration, pathogen defense, and intestinal barrier maintenance and protection [
22,
34]. IL-22 promotes production of innate antimicrobial molecules, like defensins, Reg family molecules, and S100 proteins by IECs [
34]. Additionally, the intestinal epithelial barrier is maintained via a cross talk between IL-22, innate lymphoid cells type 3, and microbiota.
iNKT cells are subgroup of unconventional, CD1d-restricted T cells that have TCRs and are able to recognize endogenous and exogenous lipids, such as αGalCer, which are presented by the surface molecule CD1d in mice (CD1 in humans) [
44]. After iNKT cell activation, copious amounts of cytokines such as IL-4, IL-10, and IL-22 are secreted to regulate the downstream activation of DCs, NK cells, B cells, or conventional T cells. Recently, there has been growing evidence that iNKT cells play a central role in governing the bidirectional interactions of the host cells and the commensal microbiota, which is key to intestinal homeostasis and preventing inflammation [
13]. The modulation of the mucosal immunity and regulation of bacterial colonization is through iNKT cells recognizing the presentation of commensal-derived lipids by CD1d that is expressed on B cells, DCs, macrophages, IECs, and innate lymphoid cells [
4,
13,
32,
33,
42,
44,
47,
53]. Despite all of the research that has been done to demonstrate intestinal iNKT cell function in intestinal homeostasis, there is still much that remains to be determined as to how iNKT cells populations adjust after intestinal surgery and how these changes correlate to changes in the intestinal commensal bacteria.
CD103
+ DCs are a heterogeneous population and can be categorized into two subsets based on their expression of CD11b. Both populations, CD103
+ CD11b
+ and CD103
+ CD11b
− are found in the small intestinal lamina propria and intestinal lymph, but CD103
+ CD11b
+ make up the majority of the population found in the lamina propria [
8,
15,
20]. The CD103
+ DCs are tolerogenic and are able to induce the differentiation of naïve T cells to FoxP3
+ Treg cells, while CD103
− DCs are not toleragenic [
8,
11,
49]. The intestinal CD103
+ DC induction of Treg differentiation is dependent on retinoic acid and active transforming growth factor (TGF) b [
8,
11,
49].
In this study, we performed two types of anastomosis surgeries on C57BL/6 wildtype mice, Roux-en-Y and end-to-end anastomoses. The focus of this manuscript is to discern the correlative relationship of the microbiota and immune system early after an anastomosis operation. We chose to evaluate these parameters at 3 days post-operative as this is the time when a latent leak could become clinically apparent [
43]. Dissecting the regulatory mechanisms of mucosal healing could impact therapeutic treatments given after an anastomotic operation.
Discussion
This current study further identifies the immune cell and microbiome shifts that occur after small intestinal anastomosis, in comparison to the homeostatic non-surgical control. We did not perform a sham-operated control as we felt the non-surgical control offered a better perspective on the presurgical normal microbiome and immune cell population environment to make comparisons to, however, we do recognize that a sham-operated control could have added more rigor to these studies, and we will consider doing them in our future work. Overall, we observed a decrease in microbiome diversity which included decreases in the major phyla
Firmicutes, Bacteroidetes, and
Saccharibacteria but then an increase in
Proteobacteria after anastomosis. We found that there were increases in Th17 and iNKTh17 cell populations and that these Th17 and iNKTh17 cells had a higher expression of TCRβ, while the increased Treg populations had homeostatic expression levels of TCRβ. Additionally, these Treg populations also had an increase in IL-10 and IL-22, which are necessary for mucosal wounding healing [
25]. Evaluating the DC populations after anastomosis revealed an increase in a subpopulation that was CD11b
hi CD103
mid, while the conventional CD11b
+ CD103
+ DCs decreased. The CD11b
hi CD103
mid DCs were expressing IL-10. The increases in iNKTs, DCs, Th17, and Treg cells correlated to decreases in
Firmicutes and
Bacteroidetes phyla and an increase in the
Proteobacteria phylum within the end-to-end and Roux-en-Y anastomoses surgery groups.
Studies involving germ-free mice have demonstrated that there is an impaired rate of intestinal epithelial cell migration, indicating the necessity of commensal bacteria for effective migration during wound healing [
2]. The pattern-recognition receptor, TLR4 specifically detects the conserved molecular product lipopolysaccharide (LPS) on microorganisms [
37]. The key functions of the TLR family are to act as sensors of microbial infection, initiate inflammatory and immune responses, and to aid in mucosal epithelial healing and homeostasis [
37]. We specifically observed that the expression of TLR4 decreased after anastomotic surgery (Fig.
2). Interestingly, if there is not proper activation of TLRs on epithelial cells, there will be abnormalities in the intestinal epithelial homeostasis and production of cytokines and heat-shock proteins. If there is a defect in the assembly and release of these factors after intestinal injury, the detection and protection of commensals is impeded [
37]. Therefore, this leads to a dysregulated interaction between commensals and TLRs resulting in chronic inflammation and tissue damage. So, a critical balance between the activation of TLRs by commensals and the protective action induced by TLRs will impact whether proper homeostasis can be maintained, or mucosal injury (including by surgery) can be properly healed. How TLR involvement and inflammation affect the microbiome environment could be further explored by comparing the effects of intraperitoneal administration of TLR4 agonist LPS from
Escherichia coli versus administration of TLR4 antagonist LPS from cyanobacteria
Rhodobacter sphaeroides after anastomotic surgery [
51]. However, this exceeds the scope of the current study.
IBD patients are associated with a lower abundance of
Firmicutes and higher abundance of
Proteobacteria and
Bacteroidetes [
2,
28]. Mice that have undergone ileocecal resection (ICR) plus a single antibiotic injection at surgery, have decreased phylogenetic diversity by 7 days post-operative. Prior to ICR,
Firmicutes and
Bacteroidetes were the predominant strains in the jejunum, but then only
Firmicutes remained the dominant population 28 days after ICR, whereas
Bacteroidetes was 0.01% of abundance and there was an increase in
Proteobacteria [
12]. Another study conducted 75% small bowel resection (SBR) in piglets and showed an increase in
Firmicutes and a decrease in
Bacteroidetes and
Proteobacteria 2 and 6 weeks post-operative [
23]. Another study performed 50% proximal SBR in mice and found that there was little to no difference in microbiota diversity between the SBR mice and sham mice that had a transection and anastomosis 7 days post-operative [
48]. They did not however compare the sham mice to non-operated mice. In our anastomotic study, we observed significant decrease in microbiome diversity and in major phyla
Firmicutes, Bacteroidetes, and
Saccharibacteria but an increase in
Proteobacteria in both of the end-to-end and Roux-en-Y anastomosis surgery groups compared to non-operative control. The differences we observed in the microbiome diversity in comparison to the forementioned studies could be due to timing of post-operative fecal collection (our study collected at day 3 vs. others at days 7, 14, and 42), type of surgery performed, and the species. Therefore, our study evaluates an acute time period that is necessary for the establishment of proper healing or by which the timing when most anastomoses start to fail [
43]. Moreover, we correlated these diversity changes to the response of the immune cells at day 3. We correlated the decrease in
Firmicutes after anastomosis surgery to an increase in Th17 IL-22
+ cells, with the greatest increases seen in the proximal segments A and B and a less of an increase in the distal segment C ileal region (Fig.
10D). This contrasts to how segmented filamentous bacteria from the
Firmicutes phylum, are able to bind to the epithelial surface of the ileum and penetrate the mucus layer [
25]. They then can induce the differentiation of CD4
+ T cells to Th17 cells [
19,
25]. Previous studies have also shown that 17 strains of
Clostridia and
B. fragilis support the induction and proliferation of Tregs that produce IL-10 [
9,
41]. Interestingly, we observed a contrasting correlation between the
Clostridia and
Bacteroidia classes, in that after anastomotic surgery they decreased while IL-10
+ Treg numbers increased. These differences again could be due to the timing of post-operative collection and what type of surgery was performed.
Post-operatively, we observed that the Th17, and iNKTh17 populations had a higher expression of TCRβ than control non-surgical mice. TCR upregulation can occur when the cells are preparing for mitosis or even independent of cell division [
27,
46]. It has been proposed that enhancement of TCR expression level overall is directly linked to T cell activation, this is so that the T cell sensitivity is elevated to peptide antigen [
45]. Lastly, ceramide is known to induce TCR upregulation [
45]. After mesenteric ischemic/reperfusion injury, ceramide concentration levels were increased in intestinal vasculature and coating the membranes of bacteria [
17]. Ceramide is known to be involved in the host response to bacterial infections [
17,
54,
55]. Therefore, the increased TCR expression that we observed after anastomotic surgery on Th17 and iNKTh17 cell populations could be due to long stimulation and a change in bacterial populations and ceramide levels, but this awaits further exploration.
Characterizing conventional DCs (cDCs) has led to intense debates and broad investigations. The issue is that several of the markers used to define DCs are also expressed by macrophages, including CD45, CD11c, CD11b, and MHC class II (IA–E). However, macrophages do not express the integrin α
E chain CD103 [
5,
35,
38,
50]. CD103
+ DCs are typical migratory cells in the lamina propria that have mainly been divided into two populations, CD11b
+ CD103
+ and CD11b
− CD103
+ cDCs [
35]. For this reason, it was unexpected to us to find an increase in a new cDC population, CD11b
hi CD103
mid, within all of the intestinal segments after anastomosis surgery. This novel population appeared to be a shift away from the CD11b
+ CD103
+ cDC population as these cells decreased significantly after surgery. Now it is unknown whether there are distinct precursors for the CD11b
+ CD103
+ and CD11b
− CD103
+ cDCs or even what signals they obtain from the intestinal environment that leads to their final differentiation state [
35]. Therefore, our data indicates that a post-operative environment leads to a new CD11b
hi CD103
mid DC differentiated population. Additionally, of the cDC populations, the CD11b
hi CD103
mid DCs were expressing IL-10, an anti-inflammatory cytokine. It is known that activated CD11b
+ DCs in the Peyer’s patches produce higher levels of IL-10 than splenic DCs and are better able to activate CD4
+ T cells to produce higher levels of IL-4 and IL-10 [
10,
11]. CD103
+ DCs, but not CD103
− DCs, from the lamina propria are also known for being able to induce FoxP3 in naïve T cells better than splenic DCs in the presence of exogenous TGFβ [
10,
11]. Our post-operative results showed an increase in IL-10
+ CD11b
hi CD103
mid DCs, which then correlated with an increase in TCRβ
hi CD4
+ T cells and Treg cells. These Treg cells had increased expression of IL-10 and IL-22, which are important for healing pathways [
25]. Along with the IL-10
+ Treg population increase, there was an increase in iNKT IL-10
+ cells post-operatively. CD11c
+ DCs can regulate iNKT cell homeostasis and activation via CD1d-dependent presentation of intestinal lipids [
42]. This DC-iNKT cell crosstalk is important for controlling the bacteria and immune cell populations, such as Tregs within the intestinal compartment as mice with a CD1d conditional deletion have dysbiosis and altered immune homeostasis [
42]. Importantly, Cre
+ CD1d
fl/fl CD11c
Cre mice orally administered the lipid iNKT activator αGalCer had failed to increase Treg populations in the intestinal mesenteric lymph node whereas WT mice had increased Treg populations [
42].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.