Macrophages are classified as proinflammatory M1 (classically activated) or anti-inflammatory
M2 (alternatively activated),
1
but binary classification fails to distinguish macrophage expression profiles.
2
Tissue resident muscularis macrophages (MMs) in the gut muscularis externa modulate
gut homeostasis including neuromuscular function.3, 4 Through neuroimmune cross-talk,
MMs may sense cues in the microenvironment from the microbiota or extrinsic sympathetic
neurons and relay this information to enteric neurons.3, 4 We showed that aging shifts
the phenotype of MMs from an M2 to an M1 polarization state, which is associated with
chronic, low-grade inflammation in the enteric nervous system and delayed intestinal
transit.
5
The objective of this study was to evaluate whether alterations in gut microbiota
contribute to age-related changes in MMs and gastrointestinal dysmotility.
To determine the effect of microbiota on MM phenotype, germ-free (GF) mice were colonized
by co-housing with young (age, 2 mo) or old (age, 24 mo) mice (Figure 1
A).
6
After 4 weeks, MMs were sorted from the colonized GF mice, termed GFyoung or GFold,
and analyzed for RNA expression. MMs from GFold mice showed decreased expression of
anti-inflammatory M2 markers, particularly found in inflammatory zone 1, C-type lectin
domain family 10 member A, and interleukin 10 (Figure 1
B), and increased expression of the proinflammatory M1 marker tumor necrosis factor
α (TNFα) (Figure 1
C). Increased TNFα also was detected in protein extract of muscularis externa (Figure 1
D), which mirrors recent studies showing higher TNFα levels in plasma
6
and in ileum
7
of GF mice colonized with old mice microbiota. We also found increased leukocyte infiltration
in the muscularis of GFold mice (Figure 1
E), suggesting that factors in the old microbiota drives inflammation.
Figure 1
Microbiota from old mice alters macrophage phenotype and causes delayed intestinal
transit. (A) GF mice were co-housed with either young or old mice for 1 month before
expression analysis was performed on sorted MMs. MMs from GFold mice showed reduced
expression of anti-inflammatory M2 markers, particularly found in inflammatory zone
1 (FIZZ1), C-type lectin domain family 10 member A (CLEC10A), and interleukin 10 (IL10),
(B) which were statistically significant, and (C) statistically significantly increased
expression of TNFα compared with GFyoung. This change in phenotype corresponded to
(D) increased levels of TNFα and (E) increased infiltration of CD45+ leukocytes in
muscularis. Data are representative of 3 independent experiments. (F) After antibiotic
(Abx) treatment, FMT was performed on old mice with young or old stool. Whole-gut
transit times (WGTTs) were measured at baseline (pre-Abx), before FMT (post-Abx),
and after FMT. (G) A statistically significant reduction in transit times was observed
in mice after FMT with young stool (FTyoung) compared with before and after Abx treatment,
which was not observed in mice transplanted with old stool (FTold). Data are combined
from 2 independent experiments. n ≥ 5 for each group. *P < .05, **P < .01, ***P <
.005 by t test or 1-way analysis of variance, with the Bonferroni multiple comparisons
test. Data are represented as means ± SD.
Next, we tested whether age-dependent differences in the microbiota influence gastrointestinal
(GI) motility. Old mice were pretreated with an antibiotic cocktail and then fecal
microbiota transplantation (FMT) was performed by gavage of young or old mice stool.
Whole-gut transit times were measured at baseline (pre-antibiotic treatment), before
FMT (after antibiotic treatment), and after FMT (Figure 1
F). FMT with young but not old stool resulted in a significant reduction in whole-gut
transit times (Figure 1
G). This suggests that in addition to altering MM phenotype, age-dependent factors
in the microbiota affect gut motility.
Microbiota-mediated loss of barrier function is implicated in age-associated intestinal
inflammation in Drosophila and mice,6, 8, 9 and appears to be TNFα-dependent.
6
This suggests that microbial factors trigger an inflammatory response that presages
disruption of barrier function. Bile acids (BAs) are a logical candidate because they
are metabolized by the microbiota and known to modulate the immune response by signaling
through host receptors.
10
By using targeted metabolomics, we found increased deconjugated BAs in fecal pellets
from old mice (Supplementary Figure 1
A). Deconjugated primary BAs were increased in particular, showing a 20-fold increase
compared with stool from young mice (Supplementary Figure 1
B). Similarly, stool from GFold mice contained increased deconjugated primary BAs,
suggesting that the MM phenotype in our colonized GF mice is mediated through BAs
(Supplementary Figure 1
C).
Takeda G-protein coupled receptor 5 (TGR5), also known as G-protein–coupled bile acid
receptor 1, has been detected on human and mouse small intestinal macrophages3, 11
by RNA sequencing and can modulate the macrophage immune response.
12
Tgr5 is differentially activated by various bile acids,12, 13 with conjugated BAs
leading to greater affinity and activation than deconjugated equivalents.12, 14 We
speculated that increased conjugated BAs in stool of young mice maintain an anti-inflammatory
phenotype in MMs through TGR5 signaling. To test this, we performed immunophenotyping
of MMs from TGR5 knockout mice (tgr5-ko) and wild-type littermates (tgr5-wt) by flow
cytometry. MMs from tgr5-ko mice showed decreased surface expression of the M2 marker
CD206 (Figure 2
A), suggesting that TGR5 deficiency results in loss of M2 phenotype.
Figure 2
TGR5 deficiency in MMs causes loss of M2 phenotype and delayed intestinal transit.
(A) MMs from tgr5-ko mice showed statistically significantly decreased surface expression
of M2 markers CD206 compared with tgr5-wt based on the mean fluorescence intensity
(MFI) normalized to isotype controls. (B) Bone marrow chimeras were generated by injecting
bone marrow from tgr5-ko or tgr5-wt mice into lethally irradiated tgr5-wt mice. Data
are combined from 2 independent experiments. (C) Mice transplanted with bone marrow
from tgr5-ko mice (KO→WT) showed statistically significantly delayed intestinal transit
compared with those transplanted with wild-type bone marrow (WT→WT). Data are representative
of 2 independent whole-gut transit time experiments. (D) Proposed model by which age-dependent
changes in microbiota cause a shift from conjugated to deconjugated BAs, resulting
in a loss of M2 phenotype in MMs and delayed intestinal transit. n ≥ 9 for each group.
*P < .05 by t test.
Tgr5-ko mice have been reported to show delayed intestinal transit.
15
Because TGR5 is expressed in multiple cell types,
15
we asked whether altered MM phenotype caused by TGR5 deficiency contributes to disrupted
intestinal motility. We generated bone marrow chimeras by transplanting bone marrow
from tgr5-ko or tgr5-wt mice into lethally irradiated tgr5-wt mice (Figure 2
B). KO→WT chimeric mice showed delayed intestinal transit compared with WT→WT (Figure 2
C), suggesting that decreased BA signaling through TGR5 in MMs contributes to age-related
disruption in motility. Based on our results, we propose the following model for how
aging affects GI motility (Figure 2
D). Age-dependent alterations in gut microbiota cause a shift from conjugated to deconjugated
BAs. This change in the conjugation status results in diminished TGR5 signaling on
MMs, contributing to loss of anti-inflammatory phenotype and delayed intestinal transit.
Our findings also offer several mechanistic and potential therapeutic interventions
that might be used in treating age-related GI motility disorder. Further studies in
human beings will be necessary to determine the clinical relevance.