Short-term tacrolimus-induced modest versus antibiotics-induced strong changes in the gut microbiome
We first investigated the short-term responses (experimental design in Fig. 1A) to tacrolimus and antibiotics on the gut microbiome and metabolism. C57BL/6 mice were treated with antibiotics for 6 days only, tacrolimus for 2 days only, the combination of antibiotics followed by tacrolimus, or untreated no drug control. Whole community metagenomic sequencing of colon intraluminal fecal contents yielded 39.4 ± 8.0 (mean ± s.d.) million reads per sample after quality assessment (Supplemental Table 1A). Intraluminal fecal content from the jejunum showed that the majority (> 98%) of the reads were from the host (Supplemental Table 1B). Thus, the intraluminal content of the colon was used in subsequent analyses. Taxonomic composition using the comprehensive mouse gut metagenome catalog (CMGM) [32] showed 222 taxonomic groups at species level (Supplemental Table 1C) in 64 genera (Supplemental Table 1D). Functional characterization using HUMAnN2 (Human Microbiome Project Unified Metabolic Analysis Network) (v0.11.2) [33] to determine the prevalence and abundance of metabolic functional units is presented in Supplemental Table 1E.
The effects of antibiotics and tacrolimus on gut microbiota were distinct. Antibiotic treatment alone or in combination with tacrolimus significantly reduced gut microbiota diversity (Fig. 1B) and altered the taxonomic composition and structure (Fig. 1C) as well as the functional makeup (Fig. 1D). Compared to tacrolimus alone or the untreated control, the antibiotic effect was much stronger with phylogenetic collateral sensitivity, as taxa from the same phylogenetic groups were simultaneously affected (Supplemental Fig. 1A). The most striking changes were in observed in Firmicutes (aka. Bacillota) with members from the families of Lachnospiraceae (Firmicutes), Oscillospiraceae (Firmicutes), Ruminococceae (Firmicutes), and Muribaculaceae (aka. S24-7, Bacteroidota) being depleted (Supplemental Fig. 1B-E). Antibiotics overpowered the effects of tacrolimus on the gut microbiome when used in combination (Fig. 1E). The most differentially abundant group was Enterobacteriaceae (Enterobacter and Klebsiella pneumoniae) for tacrolimus plus antibiotics, while the antibiotics-only treatment group had a few low abundant groups in Firmicutes and Burkholderiales (Clostridium and Paeniclostridium sordellii) (Supplemental Fig. 2). Otherwise, the antibiotics treated groups with and without tacrolimus were highly similar.
Unlike antibiotics, tacrolimus alone induced only modest changes in gut microbiota, presenting high similarities to the control in community diversity, taxonomic composition and structure, and functional makeup (Fig. 1D-E). The most affected taxa were in low abundance and sporadically distributed in different taxonomic groups without relation to the phylogenetic range (Supplemental Fig. 3). Akkermansia muciniphila (Verrucomicrobiae) was more abundant in the tacrolimus group, whereas a few low abundant Clostridia taxa were more abundant in the control group (Supplemental Fig. 2). Overall, antibiotic treatment, with and without tacrolimus, strongly affected the gut microbiome, and this wide spectrum impact was related to the phylogenetic range. The 2-day tacrolimus treatment had modest effects on the gut microbiome, which was not related to the microbial phylogenetic range.
Short-term tacrolimus treatment induced profound changes in metabolic activities in both gut lumen and serum
To investigate gut metabolism, we profiled the metabolome of paired intraluminal stool and serum samples using capillary electrophoresis-mass spectrometry (CE/MS). After quality assessment, 247 luminal metabolites were included, out of which 233 were annotated by at least one reference from PubChem [34], Kyoto Encyclopedia of Genes and Genomes [KEGG, [35]], or the Human Metabolome Database [HMDB, [36]] (Supplemental Table 2A). According to the KEGG BRITE hierarchical classification system, the most prevalent class of luminal metabolites was amino acid metabolism, comprising 40.2% of all annotated metabolites (Supplemental Table 2C). These metabolites belong to pathways in arginine and proline metabolism, arginine biosynthesis, cysteine and methionine metabolism, histidine metabolism, tryptophan metabolism, and glycine, serine and threonine metabolism. Together with other amino acid metabolites (i.e., β-alanine metabolism, glutathione metabolism), the amino acid metabolism-related metabolites comprised 48.7% of the total luminal metabolome (Supplemental Table 2D). Other prevalent classes included carbohydrate metabolism (14.6%), nucleotide metabolism (11.0%), lipid metabolism (9.0%), metabolism of cofactors and vitamins (8.5%), and the biosynthesis of other secondary metabolites (3.7%). Individual metabolites were characterized in 111 functional modules such as polyamine biosynthesis (arginine = > agmatine = > putrescine = > spermidine) to indicate key metabolic processes (Supplemental Table 2E).
The serum metabolome was estimated to be approximately 80% similar to the paired luminal metabolome based on KEGG functional modules. A total of 262 serum metabolites were included after quality assessment, of which 233 were annotated (Supplemental Table 2B). The most prevalent class was amino acids metabolism (43.2%), an even higher proportion than the luminal metabolome. Lipid metabolism (12.5%) and xenobiotics biodegradation and metabolism (3.4%) were also higher in the serum. Conversely, carbohydrate (12.5%) and nucleotide metabolism (8%) were higher in the lumen. Metabolic pathways were also similar in the lumen and serum. The main pathways for which the lumen metabolome had more coverage included protein digestion and absorption, biosynthesis of cofactors, taurine and hypotaurine metabolism, glutathione metabolism, neuroactive ligand-receptor interaction, cysteine and methionine metabolism, purine metabolism, and the cAMP signaling pathway. Conversely, the serum metabolome had higher coverage of lysine degradation, phenylalanine metabolism, tryptophan metabolism, fatty acid biosynthesis, tyrosine metabolism, and glycine, serine, and threonine metabolism. Approximately 80% of the serum functional modules shared key metabolic processes with the lumen (Supplemental Table 4G). The rest were either present in serum or gut. For instance, ornithine biosynthesis (glutamate = > ornithine) was present in lumen but not in serum.
Tacrolimus elicited distinct and strong metabolic changes within 2 days of treatment. Sparse Partial Least-Squares Discriminant Analysis (sPLS-DA) was employed to analyze the large dimensional datasets that had more variables (metabolites) than samples (p > > n) to produce robust and easy-to-interpret models [37]. Distinct metabolic profiles after 2-day tacrolimus treatment were observed in both lumen (Fig. 2A) and serum (Fig. 2B), indicating the significant impact of tacrolimus on the metabolism of both the circulation and gut lumen. The variables that contributed to the separation of treatment groups are shown in Supplemental Fig. 4. An overview of the most differentially abundant compounds among the treatment groups is shown in hierarchical clustering heatmaps (Supplemental Fig. 5). Tacrolimus elicited stronger metabolic changes than antibiotics in terms of the number of metabolites and pathways that were affected. Comparison with antibiotics revealed 10 times more significantly induced luminal metabolites and 4 times more serum metabolites elicited by tacrolimus (Fig. 2C, 2D). Comparison with the no treatment control revealed > 5 times more significantly increased luminal metabolites and 4 times more serum metabolites in tacrolimus than in antibiotics (Supplemental Table 3A, 3B). Pathway analysis also revealed that more pathways were significantly affected by tacrolimus in both the lumen and serum, which was evaluated from the dimensions of pathway topology (i.e., more hits observed in the pathway or more influential “hub” hits) and pathway significance (i.e., more compounds with statistical significance) (Supplemental Table 3C, 3D). Compared to the no treatment control, tacrolimus significantly induced luminal pathways in vitamin B6 metabolism, arginine and proline metabolism, histidine metabolism, glyoxylate and dicarboxylate metabolism, and nicotinate and nicotinamide metabolism. Compared to antibiotics, tacrolimus additionally induced luminal pathways in butanoate metabolism, alanine, aspartate and glutamate metabolism, cysteine and methionine metabolism, pantothenate and CoA biosynthesis, β-alanine metabolism, arginine biosynthesis, and starch and sucrose metabolism. Compared to the no treatment control, tacrolimus significantly increased serum metabolic pathways in alanine, aspartate and glutamate metabolism, glyoxylate and dicarboxylate metabolism, tryptophan indole pathway, primary BA, taurine and hypotaurine, TCA, and alanine, aspartate and glutamate metabolism. Additional serum pathways induced by tacrolimus compared to antibiotics included nicotinate and nicotinamide, tryptophan metabolism of serotonin and L-kynurenine, D-glutamine and D-glutamate, and thiamine metabolism. Overall, tacrolimus exerted a stronger effect on both luminal and circulating metabolism of a selected pool of essential amino acid and carbohydrate metabolism pathways.
Unlike tacrolimus, antibiotics induced only modest changes in the metabolome. The most elevated luminal compounds were primary BAs (Supplemental Fig. 6A). Serum levels of a few compounds were elevated, including the antibiotic itself (i.e., metronidazole) (Supplemental Fig. 6B). Serum pathways of alanine, aspartate, and glutamate metabolism, arginine and proline metabolism, arginine biosynthesis, and valine, leucine and isoleucine biosynthesis were increased in antibiotics. However, most of these pathways decreased in the lumen after antibiotic treatment. Tacrolimus plus antibiotics did not induce additional luminal metabolic pathways compared to antibiotics alone but induced serum taurine and hypotaurine metabolism, primary BA biosynthesis, and histidine metabolism.
Tacrolimus exerted additional effects on gut microbiome and metabolism after prolonged administration
Seven-day tacrolimus treatment was investigated to characterize the impact of prolonged tacrolimus use (experimental design in Fig. 3A). As expected, the gut microbiota of the 2- and 7-day untreated controls were highly similar and clustered together (Supplemental Table 4A), distinct from the other treatment groups, as shown in the heatmap (Fig. 3B) and principal component analysis based on taxonomic composition and structure (Fig. 3C). Seven-day tacrolimus treatment was more effective in terms of the number of differentially abundant taxa (Supplemental Fig. 7A) and significantly reduced diversity of the gut microbiota (Fig. 3D). Compared with the 2-day tacrolimus treatment, the most significantly altered taxonomic groups after 7-day treatment were distributed in a wider phylogenetic range, including Clostridiales, Verrucomincrobiota, and Saccharimonadales (Supplemental Fig. 7B). Overall, these results indicate that tacrolimus reduced commensals and overall diversity and this effect was more pronounced with longer administration.
Metabolome analyses of intraluminal stool revealed 213 metabolites after 2- and 7-day treatments (Supplemental Table 4B, 4C), comprising 88.8% of the 2-day metabolome and 70.3% of the 7-day metabolome. The remaining 11.2% and 29.7% were detected only in the 2- and 7-day treatments, respectively. A distinct set of significantly increased metabolites was observed after 7 days of tacrolimus treatment, including sets of amino acids (Phe, Leu, Trp, Tyr, Gln, Met, Arg, Asn) and dipeptides (His-Glu, Tyr-Glu) (Fig. 2E). Multiple metabolites that were significantly reduced after seven days but increased in the 2-day treatment group included isovalerylalanine, putrescine, λ-Glu-Asp (L-Glutamyl-L-aspartic acid), trimethylamine (TMA), succinic acid, histamine (histidine pathway), and threonic acid. Taking together, these results show that the effects of tacrolimus on the gut microbiota and metabolome are not immediate and accrue over time.
Modularity of gut microbiota and metabolome indicates network effects due to drug treatment
High modularity was observed in the gut microbiota and the luminal and serum metabolome. The concept of modularity was used to reflect the degree of node connectivity to which a network can be divided into subgroups or modules to understand the organization and functional relationships within a complex system. Highly connected components often have similar functions or are part of the same biological process in response to different stimuli [38]. The gut microbiota was de novo clustered into three distinct groups (Fig. 4A): group 1 was enriched in antibiotics only; group 2 was elevated in antibiotics only or with tacrolimus that contained taxa such as Prevotella, Bacteroides, Muribaculum, and Bifidobacterium, which included a large number of taxa enriched in either tacrolimus or no drug control, such as Roseburia, Oscillibactera, CAG-81, Acetatifactor, Lawsonibacter, and Schaedlerella. Taxa within group 2 and 3 were positively correlated, while taxa between groups 2 and 3 were negatively correlated (Supplemental Fig. 8A), suggesting concerted changes among subsets of the gut microbial community in response to different treatments.
Luminal metabolites also formed networks that contained compounds that were either positively or negatively correlated, suggesting concerted responses to different treatments. Four amino acid and derivative networks were observed and were positively correlated within-network (Fig. 4B, Supplemental Fig. 8B): i) Arg, Trp, Val, Ile, Phe, Tyr, Leu, Lys, Tyr-Glu, His-Glu, N6-acetyllysine, and N6, N6, N6-trimethyllysine; ii) His, Thr, and Pro; iii) γ-glutamyl dipeptides γ-Glu-Gln, γ-Glu-Trp, γ-Glu-Met and γ-Glu-Ala; and iv) Ser-Glu, Glu-Glu, Ile-Pro-Pro, and Val-Pro-Pro, which were negatively correlated with amino acid derivatives of carnitine, creatine, and betaine. In addition to amino acids, other networks were found to be functionally related to carbohydrates and nucleosides and nucleotides metabolism.
Serum metabolites are also highly modular. Multiple correlation networks were observed (Fig. 4C, Supplemental Fig. 8C): i) purine metabolism; ii) pentose phosphate pathway; iii) amino acid metabolism that includes Arg, Lys, Pro, Met, Thr, Ala, and Ser and Nω-methyarginine, citrulline, and ornithine, many of which were enriched in antibiotics groups; and iv) amino acid derivatives (isovalerylalanine, 1N-acetylleucine, 1H-Imidazole-4-propionic acid, 3-phenylpropionic acid, N-acetylphenylalanine) that were negatively correlated with glycerophospholipid ethanolamine phosphate. Overall, modularity indicates concerted interactions among functionally related components involved in key biological processes that govern the gut ecosystem and systemic metabolism.
Highly correlated gut and systemic metabolism
Sparse partial least squares (or projection to Latent Space, PLS) was employed to represent paired gut microbiota and luminal metabolome of the same mouse in the same latent space to demonstrate their level of agreement [39, 40]. The metabolic phenotype in the tacrolimus or antibiotic groups produced more “homogeneous” sample projections, as depicted by the short average arrow length between the paired gut microbiota and luminal metabolome (Fig. 4D), luminal and serum metabolome (Fig. 4E), gut microbiome, and serum metabolome (Fig. 4F).
The network modules of the gut microbiota and metabolome were correlated. Microbiota group 2 was positively correlated with luminal metabolites belonging to the carbohydrate metabolism pathway (glucaric acid, gluconic acid, 6,8-thioctic acid, quinic acid, gluconolactone, and creatinine) (Supplemental Fig. 8D), as well as with serum amino acids (Val, Tyr, Ala, Leu, Ser, Phe, Lys, Met, Asn, Pro, Thr and Arg) (Fig. 4G), many of which were enriched in antibiotics only or antibiotics with tacrolimus (Supplemental Fig. 5B). Microbiota group 3 was positively correlated with luminal metabolites of S-adenosylmethionine (SAM), GABA, glyceric acid, glycine, symmetric dimethylarginine (SDMA), asymmetric dimethylarginine (ADMA), spermidine, N1-acetylspermidine, citrulline, and O-acetylcarnitine. These metabolites are essential compounds in interconnected pathways of arginine metabolism, polyamine metabolism, nitric oxide regulation, and urea cycle. Microbiota group 3 was also positively correlated with the serum metabolites of serotonin (5-HT), 5-hydroxyindoleacetic acid (5-HIAA), lactate acid, glycerol 3-phosphate, Nω-methylarginine, butyrobetaine, and choline (Fig. 4H), which were within the same network (Supplemental Fig. 8C). Furthermore, the luminal amino acid network correlated with serum metabolites involved in glucose homeostasis, nitric oxide regulation, and BA metabolism (Fig. 4F). These results indicate that the correlations between gut and systemic metabolism occure through interconnected pathways, particularly amino acid metabolism.
Altered amino acid metabolism in an immune suppressed environment
Since tacrolimus elicited distinct metabolic phenotypes, we sought to define the metabolic phenotype, or “metabotype”, which reflects one or a set of compounds that inform about the treatment effect [31]. Metabolites from the functional pathways most induced by tacrolimus were investigated. In particular, we used the ratio of two metabolites that were either directly linked or shared a common precursor in a pathway. The ratio is less subject to individual variations and is more reflective of the dynamic changes in metabolic fluxes or shifts compared to the absolute concentration of a single metabolite [41].
Amino acid metabolic pathways, including histidine, tryptophan and arginine metabolism, were the most prominent among the compounds most significantly affected by tacrolimus. There was increased conversion of histidine to histamine and then to 1-methyl-4-imidazoleacetic acid, instead of conversion to 4-(β-acetylaminoethyl)imidazole (Fig. 5A, 5B), suggesting that the metabolism of histidine in the lumen is upregulated by tacrolimus. Three tryptophan metabolism pathways were observed in the serum, including the kynurenine, indole pyruvate, and serotonin pathways (Fig. 5C). The indole and serotonin pathways were significantly elevated by tacrolimus and/or reduced by antibiotics (Fig. 5D). The ratio of substrates involved in the indole pyruvate pathway, in which tryptophan is converted to indole-3-propionic acid (IPA) or ILA, was highest in the tacrolimus group (Fig. 5E). Serotonin (5-HT, 5-hydroxytryptamine) were also elevated by tacrolimus and/or reduced by antibiotics. Since the indole pathway requires microbial metabolism, while serotonin is primarily produced in the enterochromaffin cells of the gastrointestinal tract and released into the bloodstream, our results indicated that tryptophan metabolism in response to tacrolimus included synergistic reactions by the gut microbiome and intestinal epithelia that together directed the enzymatic reactions in tryptophan metabolism.
We further investigated the hydroxylated form of amino acids; hydroxylation is a post-translational modification that significantly influences protein structure and function, subsequently influencing immune responses[42, 43]. There were 20 hydroxylated serum metabolites and 9 fecal hydroxylated metabolites. The serum hydroxylated metabolites were defined by treatment (Supplemental Fig. 9A), indicating distinct modifications under treatment conditions. Tacrolimus increased 2-hydroxyglutaric acid, an "oncometabolite”, and its accumulation promotes the formation and progression of cancer [44]. In the antibiotic-only group, p-hydroxyphenylpyruvic acid and 2-hydroxybutryic acid were elevated. Antibiotics with tacrolimus distinctively increased the levels of hydroxyproline and decreased the levels of 3-(4-hydroxyphenyl)propionic acid, 5-HIAA (5-hydroxyindoleacetic acid), 5,6-dihydroxyindole, 5-hydroxypentanoic acid, and 3-hydroxybutyric acid (Supplemental Fig. 9B). Metabolites such as 2-hydroxyglutaric acid hydroxyproline and 5-HIAA are known immune regulators, and others such as 2-hydroxybutryic acid and 3-(4-hydroxyphenyl)propionic acid are known metabolites derived from gut microbiota, indicating the impact of treatments on disrupted microbiota, metabolite production, and immune responses.
Tacrolimus induces augmented polyamine metabolism
Tacrolimus induced distinct changes in arginine metabolism in both the lumen and serum (Fig. 6A). In serum, there was increased putrescine accompanied by an increased arginine to putrescine conversion ratio (Fig. 6B). The level of 4-acetamidobutanoate was significantly decreased, accompanied by a significantly increased ratio of spermidine to 4-acetamidobutanoate, indicating the directed metabolism by tacrolimus from arginine to putrescine, followed by spermidine and N1-acetylspermidine in the circulation (Fig. 6D). The luminal arginine metabolic pathway was also active (Fig. 6E). This was supported by the increased putrescine to arginine ratio (Fig. 6F), indicating directed metabolism by tacrolimus from arginine to putrescine, and then to N1-acetylspermidine and N8-acetylspermidine in the gut lumen. Together, these results suggest that active arginine metabolism driven by tacrolimus that is directed towards polyamine metabolism in both lumen and circulation.
Arginase I and nitric oxide (NO) synthase (NOS) compete for arginine to produce either polyamines or NO. Potent NOS inhibitors [45–48], including asymmetric dimethylarginine (ADMA) and its enantiomer symmetric dimethylarginine (SDMA), heightened after tacrolimus administration (Fig. 6C), indicating the inhibition of NOS. S-adenosylmethionine (SAM) is involved in the methylation of arginine to form ADMA, and SAM levels were increased by tacrolimus, supporting the notion that increased SAM levels lead to the formation of more ADMA, which in turn inhibited NOS activity. Together, these results demonstrate the diverted metabolism from arginine towards increased polyamine biosynthesis by tacrolimus, reflecting an increased requirement for cellular growth and proliferation in an immunosuppressed environment.
Altered BA conjugation in gut lumen and circulation
BA conjugation is an essential process that occurs in the liver, where primary BAs, such as cholic acid (CA), are combined with amino acids, such as glycine or taurine, to form conjugated BAs of glycocholic acid (GCA) or taurocholic acid (TCA) (Fig. 7A). BAs are normally reabsorbed in the intestine and recycled back into the liver through enterohepatic circulation. The metabolic phenotypes of the primary BAs under different drug treatments were distinct (Fig. 7B). Elevated serum GCA and TCA levels were observed, accompanied by elevated GCA to glycine and TCA to taurine ratios by tacrolimus (Fig. 7B). Increased luminal CA to GCA and CA to TCA ratios by tacrolimus were also observed, suggesting increased deconjugation of BAs in the lumen, a process known to be driven primarily by gut microbiota via bile salt hydrolase (BSH) enzymatic activities. Antibiotic treatment caused significantly higher luminal GCA and TCA levels (Supplemental Fig. 10), indicating reduced deconjugation in the gut lumen. As shown above (Fig. 1E, Supplemental Fig. 1A), antibiotics affected entire taxonomic groups of Firmicutes and Bacteroidota, including the majority of identified BSH-containing bacteria such as Blautia, Eubacterium, Clostridium, Lactobacillus, and Roseburia [49, 50]. Together, these results indicate that antibiotic and tacrolimus treatments both disrupt BA homeostasis, but likely through different mechanisms.
Microbiome-dependent metabolic activities
In addition to previous knowledge on microbe-derived metabolites [43], we performed in silico modeling to characterize microbial involvement in metabolic processes. To relate actual metabolite measurements to paired microbiome metabolic potentials (CMP), we calculated the set of metabolic reactions that each microbial taxon is predicted to be capable of performing using MIMOSA2 (Model-based Integration of Metabolite Observations and Species Abundances) [51]. The top metabolites correlated with the abundance of CMP of the whole microbial community are shown in Supplemental Fig. 11. The gut microbiome metabolic pathways of arginine and proline metabolism (arginine, hydroxyproline), histidine metabolism (histamine), BAs metabolism (cholic acid, glycine, taurine), alanine, aspartate and glutamate metabolism (fumaric acid), pyruvate metabolism (pyruvate), and purine metabolism (thymidine) are among the ones most correlated with paired serum metabolite concentrations. The lumen metabolite correlation result was highly similar to that of serum, with additional relations to interconnected pathways such as the β-alanine metabolism (pantothenic acid) and polyamines (spermidine). Interestingly, many of these metabolic pathways were also significantly altered by tacrolimus or antibiotics treatment. The individual bacterial species that contain the genetic potentials correlating with metabolite measurements are listed in Supplemental Table 5. For example, species containing the BA hydrolase gene (choloylglycine hydrolase, cbh) include Acutalibacter muris, Bacteroides thetaiotaomicron, Enterobacter cloacae, Clostridium celerecrescens, Lactobacillus johnsonii, Akkermansia muciniphila, suggesting their potential involvement. This analysis was limited to genes with annotated KEGG metabolic reactions, which comprised 44.7% and 43.3% of all serum and fecal metabolites, respectively. Some important metabolites, such as tryptophan indole pathway compound IPA, which were significant in our metabolome analyses, could not be included. Based on in silico modeling and previous knowledge of microbe-derived metabolites [43], the major metabolic pathways attributed to the tacrolimus metabotype are likely microbiome-dependent.
Antibiotics and tacrolimus, alone and synergistically, modulate LN and intestinal immune compartments
Immune system structure was assessed by flow cytometry of important leukocyte subsets in mesenteric LN (mLN) and peripheral LN (pLN), and by immunohistochemistry for these same subsets in LNs and intestine, and for stromal laminins in LNs. Flow cytometry showed no significant differences in the overall cellularity of CD4 + T cells, CD8 + T cells, Foxp3 + Tregs, and B220 + B cells in the mLN or pLN in any of the groups after two days of treatment (Supplemental Fig. 12). Immunohistochemistry (IHC) showed that F4/80 + macrophages (MΦ) were significantly increased in the mLN around the high endothelial venules (HEV) by tacrolimus compared to the other groups (Fig. 8A). CD11c + dendritic cells (DCs) were increased around the HEV and within the cortical ridge (CR) for all treatment groups, especially in the combined treatment group (Fig. 8B). In the pLN, Foxp3 + Tregs decreased in the CR and around the HEV in the tacrolimus treatment groups, both with and without antibiotics, but not with antibiotics alone (Fig. 8B). CD11c + DCs decreased in the pLN CR and around HEV in all treatment groups, but most significantly in the tacrolimus treatment group (Fig. 8B). In the pLN CR, laminin a4:a5 ratios were highly increased by tacrolimus-only treatment and slightly increased by antibiotics-only treatment (Fig. 8B). In the intestine, Foxp3 + Tregs were also slightly increased by tacrolimus compared with antibiotics and antibiotics with tacrolimus (Fig. 8C, Supplemental Fig. 12c). Similarly, F4/80 + MF in the intestine was increased by tacrolimus alone compared to antibiotics, both with and without tacrolimus (Fig. 8C). Measurement of intestinal barrier function showed that antibiotics increased gut permeability compared to tacrolimus alone or both treatments together, suggesting that antibiotic impairment of barrier function was ameliorated by the addition of tacrolimus (Fig. 8D). Overall, tacrolimus displayed rapid anti-inflammatory properties within two days of treatment, and this effect was distinct from that of the other groups.