Abstract

   Radiation proctopathy (RP) is a common complication of radiotherapy for
   pelvic malignancies with high incidence. RP accompanies by microbial
   dysbiosis. However, how the gut microbiota affects the disease remains
   unclear. Here, metabolomics reveals that the fecal and serous
   concentrations of microbiota‐derived 3‐hydroxybutyrate (3HB) are
   significantly reduced in RP mice and radiotherapeutic patients.
   Moreover, the concentration of 3HB is negatively associated with the
   expression of proinflammatory IL6 that is increased along with the
   severity of radiation damage. 3HB treatment significantly downregulates
   IL6 expression and alleviates IL6‐mediated radiation damage. Irradiated
   cell‐fecal microbiota co‐culture experiments and in vivo assays show
   that such a radioprotection of 3HB is mediated by GPR43. Microbiome
   analysis reveals that radiation leads to a distinct bacterial community
   compared to untreated controls, in which Akkermansia muciniphila is
   significantly reduced in RP mice and radiotherapeutic patients and is
   associated with lower 3HB concentration. Gavage of A. muciniphila
   significantly increases 3HB concentration, downregulates GPR43 and IL6
   expression, and ameliorates radiation damage. Collectively, these
   results demonstrate that the gut microbiota, including A. muciniphila,
   induce higher concentrations of 3HB to block GPR43‐mediated IL6
   signaling, thereby conferring radioprotection. The findings reveal a
   novel implication of the gut‐immune axis in radiation pathophysiology,
   with potential therapeutic applications.

   Keywords: 3HB, akkermansia muciniphila, GPR43, IL6, radiation
   proctopathy, radioprotection
     __________________________________________________________________

   Radiation reduces A. muciniphila‐derived 3HB concentration in the gut
   and serum, thereby elevating the expression of proinflammatory IL6 via
   GPR43‐mediated pathway, resulting in the exacerbation of tissue damage.
   This study demonstrate that intervention of gut dysbiosis and abnormal
   metabolomic network by gavage of probiotic A. muciniphila or 3HB
   ameliorates significantly post‐radiation tissue damage, which has
   potential therapeutic applications.

   graphic file with name ADVS-11-2306217-g001.jpg

1. Introduction

   For decades, radiotherapy has been an important component of both
   curative and palliative care for cancer patients, but it is also
   associated with serious unfavorable side effects.^[ [50]^1 ^] Radiation
   proctopathy (RP), a frequent side effect of radiation therapy for
   pelvic malignancies (such as tumors of the bladder, testes, prostate,
   rectum, cervix, and uterus) with high incidence (50–75% of patients
   receiving pelvic radiotherapy experience RP symptoms), is characterized
   by inflammation of colon tissue; however, there is no effective
   treatment available.^[ [51]^2 ^] Moreover, the fundamental mechanism
   behind RP development and advancement has only been partially
   revealed,^[ [52]^2 , [53]^3 ^] demanding additional research on RP
   pathogenesis, progression, and treatment. Latest studies have shown
   that the gut microbiota and its metabolites play a crucial role in
   radiation enteropathy,^[ [54]^4 , [55]^5 , [56]^6 , [57]^7 ^] however,
   more evidence at the molecular level is needed to provide a deeper
   understanding of the mechanisms, which will be instrumental to
   successfully develop helpful intervention strategies in patients for
   radiation‐induced intestinal side effects.

   Although we now know that epithelial injury, the immune system, and the
   gut microbiota are involved in the pathogenesis of radiation
   enteropathy,^[ [58]^8 , [59]^9 , [60]^10 , [61]^11 ^] their interaction
   needs to be investigated further. Growing evidence shows that radiation
   leads to inflammation and gut microbiota dysbiosis. For example, tissue
   damage brought on by radiation is linked to higher levels of cytokine
   expression (e.g., interleukin (IL)1β, IL6) in both human and mouse
   models.^[ [62]^9 , [63]^12 ^] A large clinical study on the association
   of microbiota with radiation enteropathy has shown that an altered
   microbiota is associated with the expression of intestinal mucosal
   cytokines.^[ [64]^4 ^] These studies suggest the existence of an
   immunity‐microbiome axis in radiation enteropathy. As gut microbial
   dysbiosis, which lead to aberrant immune responses, has been associated
   to the development of many diseases,^[ [65]^13 , [66]^14 ^] it's
   conceivable that re‐establishing the balance between microbiota and
   host immunity is just as critical for resolving radiation enteropathy
   symptoms. The clinical application of microorganisms in reducing the
   toxicological effects of radiation seems promising, although it is
   still at an early stage.^[ [67]^1 , [68]^8 ^]

   The physiological functions of gut microbiota and metabolic systems
   (including host and microbial metabolism) are critical in regulating
   human health and diseases.^[ [69]^15 , [70]^16 ^] Metabolites produced
   by commensal bacteria play an important role in the host‐microbiota
   cross talk and affect host health.^[ [71]^17 ^] Growing evidence has
   emphasized the relevance of gut microbiota‐derived metabolites in
   physiology and immunological homeostasis,^[ [72]^18 ^] however, the
   characterization of metabolites and regulatory networks involved in
   host‐microbiota interactions in RP have not yet been elucidated.

   To better understand the interactions between the radiation‐induced
   damage and the metabolites derived from gut microbiome, we aimed to use
   a mouse RP model and multi‐omics strategies to characterize the
   dynamics of the gut microbiota‐derived metabolites and their interplay
   with the host immune system, to identify and uncover the underlying
   mechanisms of the potential metabolites and their producing bacterium
   implicating in radioprotection. Our findings shed new light on the
   causal mechanism of the interplay between the RP damage, the gut
   microbiota‐derived metabolites, and host immunity, which might
   potentially help prevention and treatment strategies of RP in the
   future.

2. Results

2.1. The Concentrations of Gut Co‐Metabolite 3HB are Significantly Reduced in
Feces and Serum of Irradiated Mice and Radiotherapeutic Patients

   To evaluate the impact of radiation treatment on the metabolite
   profiles and identify the metabolites related to tissue damage in RP,
   we took advantage of our previously developed RP mouse model using
   localized rectal radiation,^[ [73]^19 ^] which replicates the
   characteristics of pathological RP in humans. Radiation‐induced damage
   was evaluated using Hematoxylin and eosin (H&E) and Masson staining of
   representative lesions and radiation injury score (RIS) (Figure [74]
   1A–C; Table [75]S1, Supporting Information). We then evaluated whether
   such radiation damage may correlate with alternations of metabolite
   profiles. To address this, LC‐MS analysis of metabolites in sera
   isolated from unirradiated (UR) and RP mice was performed. Rarefaction
   analysis comparing metabolite diversity within individual subject
   revealed that RP mice harbor a clearly distinct metabolite profiles
   compared to UR mice (Figure [76]1D). Volcano plot analysis for all
   identified metabolites found 24 metabolites with significant changes (P
   < 0.05) and a confirmed variable importance (VIP score > 1) in RP mice
   (Figure [77]1E; Figure [78]S1A; Table [79]S2, Supporting Information).
   Furthermore, the differential metabolites between RP and UR mice were
   enriched in different metabolomic signaling pathways using KEGG
   enrichment analysis. Among these pathways, synthesis and degradation of
   ketone bodies was a top one altered in mice (Figure [80]1F).
   Importantly, the concentrations of 3‐hydroxybutyrate (3HB), a
   representative member in this pathway, were significant lower in serum
   of RP mice than UR mice (Figure [81]1G; Figure [82]S1B, Supporting
   Information). However, no significant difference of 3HB abundance was
   found in liver between RP and UR mice (Figure [83]S1C, Supporting
   Information), suggesting that the difference of 3HB concentration in
   serum is not regulated by liver. Notably, the cluster analysis of
   differentially expressed metabolites showed that most of the organic
   acids and their derivatives, including 3HB, are co‐metabolites of gut
   microbiota and host (Figure [84]1H).

Figure 1.

   Figure 1
   [85]Open in a new tab

   Gut co‐metabolite 3HB is downregulated in RP mice and patients. A)
   Experimental design schematic of localized internal rectal radiation in
   C57BL/6J mice. (RP, mice exposed to irradiation; UR, mice without
   irradiation.) B) Representative images of H&E and Masson immunostaining
   of the distal rectums. Insets are demonstrated in higher magnification
   on the right. C) Histopathological change evaluated by calculating the
   radiation injury score (RIS). D) PCA plot of serum metabolites from UR
   and RP mice. E) Volcano plot of all metabolites found in serum samples.
   Red points indicate metabolites with a VIP (variable importance in
   projection) score > 1 and an adjusted P < 0.05 and log2(RP/UR) > 1;
   green points indicate metabolites with a VIP score > 1 and an adjusted
   P < 0.05 and log2(RP/UR) < −1. F) KEGG pathway enrichment analysis of
   differentially enriched metabolites in serum between RP and UR mice. G)
   3HB concentrations in serum between RP and UR mice. H) Cluster analysis
   of differentially enriched metabolites in serum. I) PCA plot of fecal
   metabolites from UR and RP mice. J) Volcano plot of all metabolites
   found in fecal samples. Red points indicate metabolites with a VIP
   (variable importance in projection) score > 1 and an adjusted P < 0.05
   and log2(RP/UR) > 1; green points indicate metabolites with a VIP score
   > 1 and an adjusted P < 0.05 and log2(RP/UR) < −1. K) Cluster analysis
   of differentially enriched metabolites in fecal samples. L) 3HB
   concentrations in fecal samples between RP and UR mice. M) Correlation
   of 3HB between serum and fecal samples measured in RP mice. (N‐O)
   Detection of 3HB concentrations in fecal N) and serum O) samples
   derived from oncology patients before (BT) and after (RT) radiotherapy
   treatment. Data are representative of at least two biological
   replicates. Data are presented as the mean ± SEM. Animal samples were
   collected post 6 weeks irradiation. Patient samples were collected one
   day prior to radiotherapy (BT) and one day after the completion of
   radiotherapy (RT), respectively. N = 8 per group in mouse model. N = 20
   for oncology patients. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P <
   0.0001 determined by the Student's t‐test [(C), (G), and (L)], Spearman
   correlation (M) and paired exact Wilcoxon test, two tailed [(N), and
   (O)].

   We next explored whether 3HB is a co‐metabolite of gut bacteria and
   contribute significantly to the concentration of 3HB in the sera. To
   address this, mice were treated with broad‐spectrum antibiotics for
   removing the bacteria from the intestine (Figure [86]S2, Supporting
   Information). Compared to mice drinking water only, antibiotic‐treated
   mice with depleted intestinal microbiomes exhibited distinct metabolite
   composition and reduced abundance of metabolites in feces (Figure
   [87]S3A–C, Supporting Information). Interestingly, organic acid and its
   derivatives, especially 3HB, were significantly reduced in feces of gut
   microbiota deficient mice compared to control (Figure [88]S3D,F,
   Supporting Information). These results reveal that 3HB is a gut
   co‐metabolite derived from gut microbiota. We then analyzed the
   metabolites in mice sera to validate whether the differences of 3HB are
   due to the alterations of gut microbiota composition. The results
   showed that gut microbiota deficient mice harbor distinct metabolite
   profiles compared with control mice (Figure [89]S4A–C, Supporting
   Information). Remarkably, the reduced concentration of 3HB in serum was
   consistent with the reduction of 3HB observed in feces of gut
   microbiota deficient mice, and was significantly associated with the
   abundance of certain gut microbiota (Figure [90]S4D–H, Supporting
   Information). Importantly, no significant difference of concentration
   of 3HB was found in the liver of gut microbiota deficient and control
   mice during a longitudinal follow‐up of the concentration of 3HB by
   location (Figure [91]S4I; Figure [92]S5, Supporting Information),
   further confirming that the altered 3HB in serum is gut derived.
   Altogether, these results suggest that gut microbiota‐derived 3HB is an
   important contributor to the concentrations of 3HB in serum.

   We subsequently assessed the fecal metabolite profiles of RP mice to
   analyze the abundance of 3HB. PCA plots revealed that metabolite
   profiles in feces are also very distinct between RP and UR
   (Figure [93]1I), and 147 metabolites were identified with significant
   changes and VIP scores, in which 3HB was indeed significantly decreased
   in RP mice (Figure [94]1J–L; Figure [95]S6A,B; Table [96]S3, Supporting
   Information). Importantly, there was a statistically significant
   positive correlation between the abundance of 3HB in fecal and serum
   samples derived from RP mice (Figure [97]1M; Figure [98]S6C; Table
   [99]S4, Supporting Information). A post‐radiation longitudinal
   follow‐up analysis confirmed the decrease of fecal and serous 3HB
   concentrations (Figure [100]S7A,B, Supporting Information). These
   results collectively suggested that such a reduction of 3HB in serum is
   caused by the decrease of 3HB in feces of RP mice. Moreover, a clinical
   detection of samples from patients uncovered that, compared to those
   before radiotherapy (BT), human subjects after radiotherapy (RT)
   exhibit a significantly decreased concentrations of 3HB in both fecal
   and serum samples (Figure [101]1N,O) and is not related to age and
   medicine used (Figure [102]S7C–F, Supporting Information), showing an
   association of decreased 3HB concentration with the exposure of
   radiation, both in mice and human subjects.

   Taken together, these data demonstrate that radiation leads to the
   reduction of levels of gut co‐metabolite 3HB, and suggest that 3HB may
   have radioprotective benefits.

2.2. Oral Administration of 3HB Ameliorates Radiation‐Induced Damage and
Attenuates Inflammation in RP Mice

   Subsequently, we tested whether the increased 3HB concentration can
   ameliorate radiation‐induced damage in RP mice. Compared to mice
   treated with saline, gavage of 3HB significantly increased the 3HB
   concentration in the feces and serum of RP mice (Figure [103] 2A–C).
   The effects of 3HB were assessed by histological analysis. The results
   showed that the clinical score (encompassing the body parameters listed
   in Table [104]S5 (Supporting Information) for evaluating
   radiation‐induced damage, which have been proved to be proportional to
   disease severity^[ [105]^7 ^]) of RP mice with 3HB treatment is
   significantly lower than that of those with saline treatment
   (Figure [106]2D). In addition, 3HB‐treated mice had longer colon length
   and decreased rectum weight. Thus, 3HB remarkably improved the
   pathological damage of the colon and rectum caused by radiation,
   inducing a protective effect against inflammation (Figure [107]2E–H).
   Furthermore, histological analysis revealed that 3HB treatment also
   reduces crypt damage, mucosal ulceration, immune cell infiltration, and
   interstitial edema (Figure [108]2I), which was manifested by a marked
   decrease in the RIS score (Figure [109]2J). Furthermore, 3HB treatment
   ameliorated radiation‐induced damage and did not induce radioprotection
   of the tumor in colorectal cancer (CRC) mice post‐radiation (Figure
   [110]S8, Supporting Information).

Figure 2.

   Figure 2
   [111]Open in a new tab

   Oral administration of 3HB ameliorates radiation‐induced damage. A)
   Experimental diagram for determining the role of 3HB in RP mice.
   Construction of RP mouse model by radiation treatment. RP mice were
   pre‐treated with antibiotics for one week and then orally administrated
   with 3HB (150 mg k^−1g body weight) or saline 3 times per week for 5
   weeks. B,C) Concentrations of 3HB in fecal (B) and serum (C) samples.
   D) Clinical scores of the mice in each group. E,F) Representative
   images of the colon (E) and colon length statistics (F). Boxed regions
   are showed at a higher magnification in G. G,H) Representative images
   of the rectum (G) and rectum weight statistics (H). I) Representative
   images of H&E and Masson immunostaining of the distal rectum. Insets
   are showed at a higher magnification on the right. J) Histopathological
   changes evaluated by calculating the RIS score. Data are representative
   of at least two biological replicates. Data are presented as the
   mean ± SEM. N = 8 per group. Samples were collected after 5 weeks of
   3HB treatment. *P < 0.05, **P < 0.01 and ***P < 0.001 determined by the
   Student's t‐test [(B), (C), (D), (F), (H), and (J)].

   Altogether, these data demonstrate that 3HB effectively ameliorates
   radiation‐induced damage and benefits radioprotection of mice,
   suggesting that 3HB is a mediator of radioprotection.

2.3. Reduced Concentration of 3HB is Negatively Associated with the
Expression of Proinflammatory IL6 that Increased along with Disease Severity

   Since tissue damage of RP is associated with immune system and 3HB can
   significantly improve the feature of colorectal inflammatory
   (Figure [112]2E–H), we then further analyzed whether RP‐induced injury
   may also correlate with the expression of certain cytokines. To this
   end, the tissues were collected from RP and UR mice to analyze the
   colonic and rectal tissue injury and inflammatory cytokine expression.
   The results showed that the clinical score proportional to disease
   severity is significantly higher in RP mice compared to UR mice
   (Figure [113] 3A). The colon length was shorter, and the rectum was
   thicker with increased rectal weight in RP mice, compared to UR
   (Figure [114]3B–E), indicating a severe inflammation.

Figure 3.

   Figure 3
   [115]Open in a new tab

   The concentration of 3HB is negatively associated with the expression
   of proinflammatory IL6 that increased along with the severity of
   radiation damage. A) Clinical score of RP and UR mice. B,C)
   Representative images of the colon (B) and colon length statistics (C).
   Boxed regions shown at higher magnification in D. D,E) Representative
   images of the rectum (D) and rectum weight statistics (E). F) Pooled
   bar graph data show the expression levels of IFNγ, IL10, IL12p70, IL1β,
   IL2, IL4, IL5, IL6, KC/GRO, and TNFα in the animal sera. G) The 6
   differential cytokines involved in the disease severity in RP mice. H)
   Western blot for phosphorylated and total STAT3 in rectal tissue.
   Glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) served as a loading
   control. I) Quantitative immunoblot analysis of pSTAT3 expression
   (compared to STAT3 and GAPDH) calculated by ImageJ. J) Correlation
   between 3HB concentration and IL6 expression in sera of RP mice. K)
   Heatmap showing correlations between 6 differential cytokines and 3HB
   in RP mice. L) Detection of IL6 expression in serum samples derived
   from oncology patients before (BT) and after (RT) radiotherapy
   treatment. Data are representative of at least two biological
   replicates. Data are presented as the mean ± SEM. N = 8 per group in
   mouse model. N = 20 for oncology patients. *P < 0.05, **P < 0.01, ***P
   < 0.001, and ****P < 0.0001 determined by the Student's t‐test [(C),
   (E), (F), (G), and (I)], Mann–Whitney U test (A), Spearman correlation
   [(J), and (K)] and paired exact Wilcoxon test, two tailed (L).

   Thus, we subsequently assessed the expression of 10 pro‐ and
   anti‐inflammatory cytokines in the sera of UR and RP mice. These
   cytokines were selected for analyses based on their well‐known
   significance in modulating inflammation response and immune system
   regulation.^[ [116]^20 ^] The expression of pro‐inflammatory cytokines
   IL1β, IL2, IL6, and TNFα was significantly increased in RP mice,
   whereas that of anti‐inflammatory cytokines IL4 and IL5 was
   significantly decreased (Figure [117]3F). However, only IL6 expression
   level paralleled with the severity of radiation‐induced damage
   (Figure [118]3G), suggesting its importance in response to radiation.
   Additionally, compared with UR mice, RP counterparts showed higher
   expression levels of phosphorylated (p)‐STAT3, the intracellular
   signaling cascade pathway activated by IL6 expression
   (Figure [119]3H,I), further confirming that the IL6 signaling pathway
   is activated and enhanced. We then examined whether the abundance of
   selected differential metabolites is correlated to abnormal
   inflammation status in RP mice. Interestingly, although the altered
   expression of six cytokines were observed in RP mice, only IL6
   exhibited a stronger negative correlation to the concentration of 3HB
   (Figure [120]3J,K), suggesting its implication in mediating the
   radioprotection of 3HB. Importantly, the concentration of IL6 was also
   significantly higher in serum samples from subjects post treatment (RT)
   than those before radiotherapy treatment (BT) in clinical oncology
   patients (Figure [121]3L).

   Collectively, these data suggest that IL6 plays a significant role in
   RP development, and may involve in the modulation of radioprotection of
   3HB.

2.4. Radioprotection of 3HB is Correlated with IL6 Signaling

   To investigate whether IL6 plays a causative role or a protective
   compensatory response in RP development, we tested whether inhibition
   of IL6 could ameliorated or worsen the effects of radiation‐induced
   damage. To this end, mice were injected with anti‐IL6 mAb or treated
   with IgG antibody as control (Figure [122] 4A). The results showed that
   RP mice treated with anti‐IL6 MAb have significantly lower clinical
   scores than those injected with saline or IgG (Figure [123]4B).
   Consistently, the protein abundance of pSTAT3 was also decreased in
   mice treated with anti‐IL6 MAb (Figure [124]4C), indicating that
   activated IL6 signaling is weakened. In addition, treatment of IL6
   receptor antagonist significantly improved the colonic length and
   rectal weight of RP mice (Figure [125]4D–G), indicating a reduced
   inflammation symptom. Moreover, the assessment of radiation‐induced
   tissue damage using H&E and Masson immunostaining showed that mice
   receiving anti‐IL6 mAb treatment display fewer lesions and a more
   intact mucosa, and ameliorated RIS scores (Figure [126]4H,I). Thus,
   these data reveal that the blockade of IL6 signaling attenuates the RP
   damage in mice, indicating that IL6 is a major mediator of
   radiation‐induced RP damage.

Figure 4.

   Figure 4
   [127]Open in a new tab

   3HB can downregulate the radiation‐induced expression of IL6, a key
   mediator involved in RP damage. A) Experimental diagram for determining
   whether the blockage of IL6 signaling has protective effect in the RP
   mice. Construction of RP mouse model by radiation treatment. RP mice
   were injected with 5 mg k^−1g anti‐IL6 MAb, or IgG antibody as control,
   or vehicle and injections were given every two days for 6 weeks. B)
   Clinical score for the mice in each group. C) Western blot for
   phosphorylated and total STAT3 in rectal tissue. GAPDH served as a
   loading control. D,E) Representative images of the colon (D) and colon
   length statistics (E). Boxed regions shown at higher magnification in
   F. F,G) Representative images of the rectum (F) and rectum weight
   statistics (G). H) Representative images of H&E and Masson
   immunostaining of the distal rectums. Insets are demonstrated in higher
   magnification at right. I) Histopathological change evaluated by
   calculating RIS score. J) Expression level of IL6 in the animal sera of
   the same mice as shown in Figure [128]2. K) Expression level of IL6 in
   irradiated or unirradiated HIEC‐6 epithelial cells (IECs). L)
   Expression level of IL6 in irradiated IECs treated with 3HB (10 mM) or
   saline. Data are representative of at least two biological replicates.
   Data are presented as the mean ± SEM. Samples were collected after 6
   weeks of anti‐IL6 MAb treatment. N = 8 for saline and anti‐IL6 MAb
   treated groups; N = 7 for IgG treated group. *P < 0.05, **P < 0.01,
   ***P < 0.001, and ****P < 0.001 determined by the one‐way ANOVA with
   Tukey's multiple comparison test [(B), (E), (G), and (I)] and Student's
   t‐test [(J), (K), and (L)].

   Given the fact that blocking IL6 signaling or supplementing 3HB are
   both beneficial for RP mice and that 3HB has signaling roles, we next
   investigated whether the radioprotective effects of 3HB in RP mice is
   related to the inhibition of IL6 signaling. Notably, the
   radioprotection effects showed no significant difference between RP
   mice treated with 3HB alone and 3HB plus the blockade of IL6 signaling
   (Figure [129]S9, Supporting Information). Indeed, 3HB treatment could
   significantly downregulate the expression of IL6 (Figure [130]4J) and
   the protein levels of pSTAT3 of rectal tissue in RP mice (Figure
   [131]S10A,B, Supporting Information). 3HB treatment also significantly
   downregulated IL6 expression in irradiated HIEC‐6 epithelial cells
   compared to those treated with saline (Figure [132]4K,L). These results
   demonstrate that 3HB exerts radioprotection against RP damage at least
   in part mediated by IL6 signaling.

   Collectively, these results indicate that IL6 is a major mediator of RP
   damage and provide evidence that the radioprotective effect of 3HB
   against RP damage is partially ascribed to the regulation of IL6
   expression.

2.5. 3HB Exerts Radioprotective Effect against IL6‐Mediated RP Damage via
GPR43

   We further explored how 3HB regulates IL6 expression. We therefore
   performed KEGG enrichment analysis of the differentially expressed
   metabolites in sera and feces between RP and UR mice, respectively. The
   results showed that these metabolites are clustered in various
   pathways, in which the synthesis and degradation of ketone bodies and
   cAMP signaling pathways are enriched in both serum and fecal samples
   (Figure [133] 5A). On the other hand, as 3HB can act as a signaling
   molecule via G protein‐coupled receptors (GPRs) to initiate additional
   signaling cascades,^[ [134]^21 ^] including the cAMP signaling pathway
   (Figure [135]5A), we next investigated whether and which GPRs mediate
   the radioprotective role of 3HB. GPR43, GPR41, GPR109A, GPR81, GPR35,
   and GPR40 were selected for analysis based on their well‐established
   importance in mediating immune and metabolic functions.^[ [136]^22 ^]
   We found that, compared with UR mice, the expression of GPR43, but not
   other GPRs, in RP mice is significantly increased in colorectal tissues
   (Figure [137]5B,C). Furthermore, the expression of GPR43 in both
   colonic and rectal samples from RP mice treated with 3HB was
   significantly reduced, compared to that in untreated controls
   (Figure [138]5D,E). These results indicate that amelioration of
   radiation‐induced damage by 3HB is associated with downregulated
   expression of GPR43.

Figure 5.

   Figure 5
   [139]Open in a new tab

   The radioprotective effect of 3HB is mediated by GPR43. A) Metabolic
   network integrated biochemical pathways and chemical relationships of
   3HB derived from RP and UR mice. B,C) The expression of GPR receptors
   in the rectal (B) and colon samples (C) derived from UR and RP mice.
   D,E) Expression of GPR43 in the rectal (D) and colon samples (E)
   derived from RP mice treated with 3HB or saline. F) Expression level of
   IL6 in irradiated IECs treated with 3HB (10 mM) and GPR43 agonist
   (4CMTB, 10 µM). G) Expression level of IL6 in gpr43^−/− and control
   cells treated with 3HB (10 mM). H) Experimental diagram for determining
   whether the activation of GPR43 blocks the protective effect in the RP
   mice. Construction of RP mouse model by radiation treatment. RP mice
   were pre‐treated with antibiotics before orally administrated with 3HB
   (150 mg k^−1g body weight) or saline 3 times per week for 5 weeks. Mice
   were injected with GPR43 agonist (4CMTB, 10 mg k^−1g body weight) or
   vehicle 3 times per week at the same time. I) Concentration of IL6 in
   the animal sera. J) Clinical score of the mice in each group. K,L)
   Representative images of the colon (K) and colon length statistics (L).
   Boxed regions are showed at higher magnification in M. M,N)
   Representative images of the rectum (M) and rectum weight statistics
   (N). O) Representative images of H&E and Masson immunostaining of the
   distal rectums. Insets are demonstrated in higher magnification at
   right. P) Histopathological change evaluated by calculating RIS score.
   Data are representative of at least two biological replicates. Data are
   presented as the mean ± SEM. Samples were collected after GPR43 agonist
   treatment for 5 weeks. N = 8 for the vehicle injected group; N = 7 for
   the GPR43 agonist injected group. *P < 0.05, **P < 0.01, and ***P <
   0.001 determined by the Student's t‐test [(B), (C), (D), (E), and (G)],
   one‐way ANOVA with Tukey's multiple comparison test [(F), (I), (J),
   (L), (N), and (P)].

   Given that 3HB decreases the expression level of IL6, we wanted to
   identify the transcriptional regulator of IL6 regulated by 3HB. To
   address this, we initially assessed the expression of il6 mRNA in
   colorectal tissue of RP and UR mice. As expected, the expression levels
   of IL6 were significantly higher in colorectal tissues from RP mice
   than those from UR mice (Figure [140]S11A, Supporting Information). A
   number of known transcription factors that have been shown to regulate
   IL6 gene transcription were selected for analysis.^[ [141]^23 ^] The
   results showed that compared with UR mice, the expression of SP1, which
   is involved in regulating the chronic intestinal inflammation and
   apoptosis of colorectal cell lines,^[ [142]^24 , [143]^25 ^] but not
   other regulators, is both significantly increased in colonic and rectal
   tissues in RP mice (Figure [144]S11B, Supporting Information). Indeed,
   RP mice treated with 3HB significantly reduced the expression of sp1
   and il6 in both colonic and rectal samples, compared to that in
   untreated controls (Figure [145]S11C,D, Supporting Information). These
   results indicate that 3HB inhibits radiation‐induced IL6 expression via
   transcriptional regulator SP1.

   We subsequently tested whether GPR43 is involved in the regulation of
   SP1 using the irradiated cell model and co‐culture experiment. The
   results showed that SP1 inhibitor significantly downregulates the
   expression of sp1 and il6, but not gpr43 in irradiated cells (Figure
   [146]S11E–G, Supporting Information). Remarkably, with or without 3HB
   co‐treatment, GPR43‐synthetic agonist increased the expression of il6
   and sp1, compared to 3HB treatment alone, in irradiated cells
   (Figure [147]5F; Figure [148]S11H, Supporting Information). These
   results not only confirmed a direct role of SP1 in regulation of
   radiation‐induced IL6 but also indicated the control of GPR43 on SP1.
   Moreover, the downregulation of IL6 by 3HB was blocked by GPR43 agonist
   or the knockout of gpr43 (Figure [149]5F,G; Figure [150]S12, Supporting
   Information). Notably, these results are in consistent with those in
   vivo analyses in mice, indicating that the inhibition of 3HB on
   radiation‐induced IL6 expression is mediated by GPR43.

   We next investigated whether GPR43 agonist impairs the radioprotective
   effect of 3HB in RP mice. RP mice pre‐treated with antibiotics were
   treated with or without 3HB and GPR43 agonist (Figure [151]5H). The
   results showed that, compared with those of controls, GPR43 agonist
   treatment significantly suppressed the effects of 3HB in downregulating
   the expression of IL6, reducing pSTAT3 levels, and ameliorating
   clinical scores (Figure [152]5I,J; Figure [153]S10C,D, Supporting
   Information). Consistently, compared with controls, treatment of GPR43
   agonist prevented the protective effects of 3HB on radiation‐induced
   pathological damage of the colonic and rectal tissues
   (Figure [154]5K–N). In addition, histological analysis revealed that
   the GPR43 agonist impairs the effects of 3HB in reducing damage of
   crypts, mucosal ulceration, immune cell infiltration, and interstitial
   edema, resulting in a high RIS score with no statistical significance
   compared to the controls (Figure [155]5O,P). Thus, blockage of the
   radioprotection of 3HB by activating GPR43 indicates GPR43 as a major
   mediator of 3HB against RP damage in mice.

   Taken together, these results demonstrate that GPR43 mediates the
   radioprotection of 3HB by downregulating the expression of IL6 during
   radiation‐induced RP progression.

2.6. Gut Bacterium Akkermansia muciniphila promotes 3HB Concentration and
Downregulates the Radiation‐Induced IL6 Expression

   Since gut but not liver involved in the alteration of the serous 3HB
   concentration in RP mice, we hypothesized that the gut microbiota may
   contribute to control the level of 3HB concentration. To address this,
   we performed 16S rDNA sequencing of fecal samples to analyze the
   composition of the gut bacteria in UR and RP mice. Principal
   co‐ordinates analysis (PCoA) revealed that RP mice harbor a distinct
   bacterial community relative to UR mice (Figure [156] 6A). Compared to
   UR, the abundance of Firmicutes and Bacteroidetes were reduced and the
   abundance of Proteobacteria was increased in RP mice. In particular,
   Verrucomicrobia, and a representative species of Verrucomicrobia,
   Akkermansia muciniphila, were significantly reduced in RP mice
   (Figure [157]6B,C). LEfSe was utilized to identify the bacterial taxa
   linked to radiation‐induced tissue injury. A total of 14 dominant
   bacteria showed a significant change in abundance in RP mice compared
   with UR mice (LDA score > 3, p < 0.05), among which A. muciniphila was
   the most significantly reduced species post radiation (Figure [158]6D).
   Additionally, random forest analysis showed that A. muciniphila
   displays a high Gini score, further confirming its role in RP
   development (Figure [159]6E). Notably, a decrease in the abundance of
   A. muciniphila in the features of radiation‐reshaped gut microbiota was
   consistent with that reported in previous study.^[ [160]^7 ^] We then
   evaluated the association between 3HB concentration and the
   significantly changed bacteria in RP mice to find dominant bacteria
   that can enrich 3HB concentration. We found that the reduction of 3HB
   concentration induced by radiation is significantly correlated with the
   decrease of the abundance of A. muciniphila in RP mice
   (Figure [161]6F,G; Table [162]S6, Supporting Information). Meanwhile,
   mice with depleted intestinal microbiomes were used to find the
   bacteria that play dominant effects on the changes of 3HB concentration
   in feces under normal conditions. The results revealed significant
   correlations between the dominant bacteria with reduced abundance and
   six gut co‐metabolites with decreased concentration and uncovered that
   A. muciniphila abundance is positively correlated with 3HB
   concentration (Figure [163]S13, Supporting Information). Therefore,
   these results collectively suggest that a radiation‐induced decrease of
   the abundance of A. muciniphila led to the reduction of 3HB
   concentration.

Figure 6.

   Figure 6
   [164]Open in a new tab

   The significant reduced A. muciniphila in radiation‐reshaped gut
   bacteria is positively associated with lower 3HB concentration, can
   produce 3HB and downregulate IL6 expression via GPR43. A) PCoA plot
   (based on weighted UniFrac distances). B) The relative abundance of gut
   bacteria at phylum level in the fecal samples. C) Volcano plot of all
   bacterial species found in fecal samples. Red points indicate species
   with an adjusted P < 0.05 and log2(UR/RP) > 1; green points indicate
   species with an adjusted P < 0.05 and log2(UR/RP) < −1. A. muciniphila
   (black box) is a more abundant species enriched in UR (i.e.,
   significantly reduced in RP mice). D) Histogram of the linear
   discriminant analysis (LDA) coupled with effect size measurements
   (LEfSe) [LDA significant threshold (log10) > ±3] identified taxonomic
   biomarkers at species level between UR and RP. E) Predictive importance
   of selected species enriched by LEfSe analysis was assessed by random
   forest analysis. F) Heatmap showing correlations between 3HB
   concentration and the abundance of LEfSe‐enriched species in RP mice.
   G) Correlation between 3HB concentration and A. muciniphila abundance
   in fecal samples of RP mice. H) Growth curve of A. muciniphila cultured
   in BHI under anaerobic conditions. I) Measurement of 3HB concentration
   in cultured A. muciniphila supernatant at the indicated time points. J)
   Analysis of A. muciniphila abundance in stool samples from RP mice
   after 5 weeks of A. muciniphila administration. K,L) Gavage of A.
   muciniphila increased the concentration of 3HB in fecal (K) and serum
   (L) samples. M) IL6 expression in irradiated IECs treated with A.
   muciniphila, inactivated A. muciniphila and saline, respectively. N)
   IL6 expression in irradiated IECs co‐cultured with fecal bacterial
   suspensions obtained from RP mice 6 weeks post‐radiation or from
   age‐matched UR mice. Viable or non‐viable A. muciniphila was added in
   the indicated group. O) IL6 expression level in irradiated IECs treated
   with A. muciniphila and GPR43 agonist (4CMTB, 10 µM). P) Expression
   level of IL6 in gpr43^−/− and control cells treated with A.
   muciniphila. Q) Detection of A. muciniphila abundance in fecal samples
   derived from oncology patients before (BT) and after (RT) radiotherapy
   treatment. Data are representative of at least two biological
   replicates. Data are presented as the mean ± SEM. N = 8 per group in
   mouse model. N = 20 for oncology patients. *P < 0.05, **P < 0.01, and
   ***P < 0.001 determined by the Student's t‐test [(I), (J), (K), (L),
   (M), and (P)], permutation multivariate analysis of variance
   (PERMANOVA) test (A), one‐way ANOVA with Tukey's multiple comparison
   test [(N) and (O)], Spearman correlation [(F), and (G)] and paired
   exact Wilcoxon test, two tailed (Q).

   We thus measured the concentration of 3HB in the culture supernatant at
   different time points based on the growth curve of A. muciniphila to
   validate whether A. muciniphila can produce and promote the
   accumulation of 3HB. The results showed that the concentration of 3HB
   increased during the growth phases of A. muciniphila, with much higher
   concentrations during the stationary phase (Figure [165]6H,I).
   Moreover, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis and
   previous studies showed that FabG might be involved in the biosynthesis
   of 3HB,^[ [166]^26 , [167]^27 ^] and the presence of an expressible
   fabG gene in A. muciniphila was confirmed by PCR and qPCR analyses,
   respectively (Figure [168]S14A,B, Supporting Information). When treated
   with epigallocatechin gallate (EGCG), an inhibitor of FabG,^[ [169]^28
   ^] the 3HB concentration in the culture supernatant of the bacteria in
   stationary phase was decreased with a EGCG concentration dependent
   manner, whereas the survival and growth rate of A. muciniphila did not
   affect (Figure [170]S14C,D, Supporting Information), further confirming
   that A. muciniphila can produce and accumulate 3HB. Furthermore, gavage
   of A. muciniphila significantly increased fecal and serous 3HB
   concentrations in the RP mice (Figure [171]6J–L). Thus, these data
   collectively indicate that gut A. muciniphila is an important
   contributor to affect 3HB levels in RP mice.

   Next, we tested whether A. muciniphila can play a role in the
   regulation of radiation‐induced IL6 expression. The irradiated cells
   treated with live but not heat‐inactivated A. muciniphila, resulted in
   significant downregulation of the expression of IL6 (Figure [172]6M).
   IL6 level was higher in the irradiated cells treated with fecal
   suspensions derived from RP mice than those treated with that derived
   from UR mice. The fecal suspension obtained from RP mice showed no
   significant difference compared to the saline group, which may be due
   to the lack of A. mucinipila. Moreover, supplementation of live but not
   inactivated A. muciniphila into the microbiota derived from RP mice
   significantly reduced IL6 expression in co‐cultured irradiated cells
   (Figure [173]6N). We also studied whether GPR43 is involved in the
   downregulation of IL6 expression by A. muciniphila treatment. Compared
   with controls, A. muciniphila treatment significantly downregulated IL6
   expression, whereas A. muciniphila plus the GPR43 agonist blocked the
   decrease of IL6 expression (Figure [174]6O; Figure [175]S15, Supporting
   Information). Importantly, knockout of gpr43 in HIEC‐6 cells impaired
   the ability of A. muciniphila on downregulation of radiation‐induced
   IL6 expression (Figure [176]6P). Therefore, A. muciniphila can directly
   downregulate radiation‐induced expression of IL6 via GPR43‐mediated
   pathway. Like the results from mice, the abundance of fecal A.
   muciniphila was significantly lower in clinical oncology patients after
   radiotherapy treatment (BT) than those before treatment (RT)
   (Figure [177]6Q). These results indicate that decreased A. muciniphila
   levels are related to radiation injury, and suggest that gavage of A.
   mucinihila may provide radiation protection.

   Taken together, these results demonstrate that A. muciniphila, an
   important contributor of 3HB levels, can also downregulate IL6
   expression through GPR43‐mediated pathway, suggesting a potential
   radioprotective benefits for RP.

2.7. Gavage of A. muciniphila Increases 3HB Concentration, reduces IL6
Expression and Ameliorates Radiation‐Induced Damage in RP Mice

   We then investigated whether A. muciniphila can confer radioprotective
   effects against RP in mice. To address this, RP mice pre‐treated with
   antibiotics were orally administrated A. muciniphila or saline
   (Figure [178] 7A). Fluorescence in situ hybridization (FISH) analysis
   by using an A. muciniphila‐specific probe showed that A. muciniphila
   cells colonize the intestinal epithelium of mice fed with A.
   muciniphila, compared with saline‐treated mice (Figure [179]7B).
   Colonization of A. muciniphila in colonic mucosa were also observed by
   transmission electron microscopy (Figure [180]S16A, Supporting
   Information). Furthermore, the increased colonization of A. muciniphila
   was confirmed using A. muciniphila‐specific qPCR (Figure [181]7C).
   These results indicate that A. muciniphila indeed can colonize in
   intestinal compartment upon gavage.

Figure 7.

   Figure 7
   [182]Open in a new tab

   Gavage of A. muciniphila increases 3HB concentration, reduces IL6
   expression and ameliorates radiation‐induced damage in RP mice. A)
   Experimental diagram for determining the role of A. muciniphila in RP
   mice. Construction of RP mouse model by radiation treatment. RP mice
   were pre‐treated with antibiotics for one week before orally
   administrated with A. muciniphila (2 × 10^8) or saline 3 times per week
   for 5 weeks. B) Representative fluorescent in situ hybridization (FISH)
   and confocal microscopic imaging analysis of A. muciniphila (green) in
   the intestinal mucosa using an A. muciniphila‐specific probe. DAPI (4′,
   6‐diamidino‐2‐phenylindole) was used for nuclear staining (blue). Scale
   bars, 10 µm. C) Analysis of A. muciniphila abundance in stool samples
   from mice at days 9 and 43 after antibiotic treatment. (D‐E) 3HB shows
   much higher concentrations in fecal D) and serum E) samples from mice
   with oral administration of A. muciniphila than that treated with
   saline. F,G) Expression of GPR43 was assessed by qPCR‐based analysis
   using mRNA extracted from the colon (F) and rectal samples (G). H) The
   expression of IL6 level in the animal sera. I) Clinical score of the
   mice in each group. J,K) Representative images of the colon (J) and
   colon length statistics (K). Boxed regions shown at higher
   magnification in L. L,M) Representative images of the rectum (L) and
   rectum weight statistics (M). N) Representative images of H&E and
   Masson immunostaining of the distal rectums. Insets are demonstrated in
   higher magnification at right. O) Histopathological change evaluated by
   calculating RIS score. Data are representative of at least two
   biological replicates. Data are presented as the mean ± SEM. Samples
   were collected after A. muciniphila treatment for 5 weeks. N = 8 for
   the saline treated group and N = 7 for the A. muciniphila treated
   group. *P < 0.05, **P < 0.01 and ***P < 0.001 determined by the
   Student's t‐test [(D), (E), (F), (G), (H), (I), (K), (M), and (O)],
   two‐way ANOVA with Tukey's multiple comparison test (C).

   Notably, gavage of A. muciniphila significantly increased the fecal and
   serous 3HB concentrations in the mice (Figure [183]7D,E). The
   expression of GPR43 in both colonic and rectal samples from RP mice
   treated with A. muciniphila was also significantly reduced, compared to
   that in untreated controls (Figure [184]7F,G). Indeed, compared to mice
   fed with saline, mice fed with A. muciniphila resulted in downregulated
   expression of IL6 and pSTAT3 in rectal tissues and lower clinical
   scores (Figure [185]7H,I; Figure [186]S16B,C, Supporting Information).
   As expected, A. muciniphila treatment significantly prevented
   shortening of colonic length and thickening of rectal tissue
   (Figure [187]7J–M). Moreover, pathological and histological analyses
   showed that gavage of A. muciniphila remarkably protects the distal
   colorectal tissue from radiation‐induced damage (Figure [188]7N,O).
   Furthermore, A. muciniphila treatment ameliorated radiation‐induced
   damage and did not induce radioprotection of the tumor in CRC mice
   post‐radiation (Figure [189]S17, Supporting Information). Thus, these
   results reveal that A. muciniphila confers radioprotective effects on
   radiation‐induced RP damage by increasing the fecal and serous 3HB
   concentration, downregulating the expression of GPR43 and IL6.

3. Discussion

   Increasing evidence has highlighted that immune system and the gut
   microbiota are involved in the pathogenesis of RP,^[ [190]^8 , [191]^9
   , [192]^10 , [193]^11 ^] and that metabolites produced by commensal
   bacteria play an important role in the host‐microbiota cross‐talk.
   However, the relevance and characterization of regulation between gut
   microbiota and metabolomic involved in physiology and immunological
   homeostasis of RP have not yet been completely elucidated. This study
   demonstrated that dysbiosis and abnormal status of metabolic systems
   (including host and microbial metabolism) induced by radiation are
   related to the pathogenesis of RP, and that A. muciniphila and 3HB
   contribute substantially to radioprotection in mice, which is
   associated with the expression of proinflammatory cytokine IL6.

   Our study showed that the metabolite profiles of RP mice is
   characterized by significant reduced concentration of 3HB, a gut
   co‐metabolite derived from commensal bacteria, higher levels of IL6,
   and severe tissue damage. We further provided evidence that gut
   microbiota‐derived 3HB plays a radioprotective role in intestinal
   inflammation and tissue damage. Mechanistically, gavage of 3HB
   significantly ameliorates radiation‐induced damage by downregulating
   GPR43‐mediated IL6 expression in RP mice. Moreover, our study showed
   that the gut microbiome pattern of RP mice is characterized by the
   reduction of core species, particularly A. muciniphila which serves as
   an important contributor for 3HB concentration. We confirmed that
   gavage of A. muciniphila increases 3HB concentration of mice and
   contributes substantially to radioprotection. Our findings contribute
   to advance our knowledge of the link between the RP disease mechanism
   and gut microbiome, and provide potential prevention and treatment for
   alleviating clinical radiation‐induced damage.

   As an initial step of this study, we attempted to identify the
   metabolites and inflammatory cytokines linking to the pathogenesis of
   RP. Our results demonstrated that radiation can alter the fecal and
   serous metabolite profiles of RP mice, in which the concentration of
   3HB is significantly reduced. Radiation also increases the expression
   of inflammatory cytokine IL6 that exhibits significant negative
   correlation with 3HB. Previous study^[ [194]^12 ^] and our results show
   that the expression of IL6 is correlated with the severity of
   radiation‐induced damage and therefore is a good marker of RP. Although
   3HB has been shown to have significant therapeutic effects in colitis
   and colorectal cancer,^[ [195]^29 , [196]^30 ^] whether it serves as an
   immune effector remains unknown. We confirmed that 3HB directly
   downregulates radiation‐induced IL6 expression and exerts a
   radioprotective effect. This result is also be supported by a previous
   study demonstrating that treatment with 3HB suppresses the levels of
   inflammatory cytokines IL1β, IL6, and IL8 in human placental tissue
   culture.^[ [197]^31 ^] Another study reported that 3HB reduced NLRP3
   inflammasome‐mediated IL1β and IL18 production in human monocytes.^[
   [198]^32 ^] These findings suggest that 3HB is an immune effector and
   may be a promising mediator of radioprotection. In addition to 3HB, we
   also identified other metabolites whose levels were reduced in RP mice.
   Further studies are required to investigate whether these compounds
   have radioprotective properties.

   The current data using our RP mouse model demonstrated that irradiation
   leads to dysbiosis, which is consistent with previous findings.^[
   [199]^9 , [200]^10 , [201]^11 ^] Alterations in intestinal microbiota
   were observed in patients with RP, but a causal association with
   disease activity remains limited.^[ [202]^1 ^] Our research revealed
   for the first time that radiation treatment impairs colonization and
   abundance of A. muciniphila. In addition, the current study shows that
   the abundance of A. muciniphila is positively correlated with the
   concentration of 3HB in fecal and serous samples from RP mice, and both
   in vitro and in vivo experiments confirmed that A. muciniphila exhibits
   a dominant role in mediating the accumulation of 3HB. Therefore, at
   least one of the radioprotective pathways of A. muciniphila may be
   achieved by mediating the accumulation of 3HB. Moreover, other
   metabolites derived from A. muciniphila are known to attenuate
   proinflammatory cytokine responses.^[ [203]^33 , [204]^34 ^] Therefore,
   identifying more effective metabolites may provide a comprehensive
   understanding of the pleiotropic effects of A. muciniphila on
   radioprotection.

   In vitro findings in the current study demonstrated that A. muciniphila
   has a direct down‐regulation effect on the expression of IL6 in
   epithelial cell line. Thus, A. muciniphila plays a protective role in
   intestinal inflammation and tissue damage, at least in part through the
   increase of 3HB concentration and the reduction of IL6 level. Previous
   studies have shown that lack or decrease in the abundance of A.
   muciniphila is linked to many diseases, including steatosis of the
   liver, inflammation, obesity, diabetes, and the response to cancer
   immunotherapies.^[ [205]^35 ^] Our data added a novel and effective
   application for A. muciniphila in RP treatment.

   Our finding on the function of GPR43 has a potential clinical
   relevance. Previous studies have shown that GPR43 is extensively
   expressed in immune tissues and cells,^[ [206]^36 ^] suggesting its
   important role in immune responses. However, the role of GPR43 in the
   inflammatory response remains elusive. GPR43 may have both pro‐ and
   anti‐inflammatory effects, depending on the disease model employed.^[
   [207]^37 , [208]^38 , [209]^39 , [210]^40 , [211]^41 ^] In addition,
   GPR43 may mediate interactions between the human host and gut
   microbiome.^[ [212]^42 ^] Our results demonstrated that activation of
   GPR43 blocks the downregulation of IL6 by A. muciniphila or A.
   muciniphila‐mediated 3HB, thus prevents the radioprotection. This
   result further points to the unique association of GPR43 with gut
   microbiota and radiation‐induced IL6 expression in the pathogenesis of
   RP. The relationship between GPR43 and IL6 has also been reported in
   other studies. For example, knockout of GPR43 reduces IL6/IL1β/TNFα
   levels in cells transfected with bacterial pathogen Klebsiella
   pneumoniae.^[ [213]^43 ^] Another study found that GPR43 activation
   enhances psoriasis‐like inflammation by upregulating IL6 signaling in
   the epidermis.^[ [214]^44 ^] These findings suggest that GPR43 may also
   play a crucial role in radiation‐induced intestinal injury. Our results
   will be particularly relevant for the clinical application of GPR43
   agonists or antagonists in patients with RP. Therefore, future clinical
   trials are needed to evaluate the inhibition effect of GPR43 signaling
   in patients with RP.

   Gut microbiota may be useful for the prevention and treatment of RP. We
   showed that RP mice has depletion in the Firmicutes and Bacteroidetes
   phyla and an increase in Proteobacteria, similar to what was observed
   in inflammatory bowel disease (IBD).^[ [215]^45 ^] Thus, dysbiosis is a
   common feature in IBD and RP. Moreover, we found that RP and IBD share
   similarities, as both are accompanied by fibrosis, inflammation,
   epithelial barrier breaches, and mucosal immune cell infiltration.
   Modulation of IBD activity by manipulating the microbiome is now
   gaining a lot of attention, thus it seems sense for RP to follow suit.
   Probiotics, and fecal microbial transplants (FMT) have all been
   recommended for the treatment of IBD and may also be effective for the
   management of RP. Attempts at bacteriotherapy in the setting of RP
   appear to be safe, according to several research conducted on both
   rodents and humans.^[ [216]^46 , [217]^47 ^]

   In conclusion, the results of this study demonstrate radiation‐induced
   changes in the microbiome and metabolome and how such an alteration
   impacts RP‐induced inflammation and RP progression. These findings
   suggest the potential of A. muciniphila and 3HB in the treatment of RP,
   and that effective treatment of RP could be achieved by modulating the
   gut microbiota and its interaction with the immune system.
   Collectively, given the fact that current therapeutic methods for
   remission of the adverse side effects of RP are limited and expensive,
   our study provides a promising clinical option for the management of
   the damage associated with RP.

4. Experimental Section

Mice

   Female C57BL/6J mice aged 6–8 weeks or C57BL/6‐APC^min/+ mice aged 8–10
   weeks were kept in a specified pathogen‐free (SPF) environment with
   unrestricted access to drinking water and food under
   temperature‐controlled settings and a 12‐hour light/dark cycle. The
   littermates were used in the experiments. All animal research was
   approved by the Institutional Animal Care and Use Committee of the
   Sixth Affiliated Hospital, Sun Yat‐sen University (IACUC‐20200813).

RP Mouse Model

   The mice were treated with rectal radiation using an RS2000 device (Rad
   Source, USA) as we previously described.^[ [218]^19 ^] The mice were
   subjected to irradiation (25 Gy; 2.14 Gy min^−1) while being protected
   by a 4 mm thick lead shelter that exposed the lower pelvic area (1
   cm^2) including the rectum in the center of the field. After
   irradiation, the mice were kept in an SPF setting with regular feed and
   water. Unless otherwise specified, mice were examined for changes in
   body weight and other bodily parameters 6 weeks following radiation.
   Clinical scores were calculated based on weight reduction, physical
   appearance, posture, mobility, anal hair, and hydration using a minor
   modified cumulative scoring method (Table [219]S5, Supporting
   Information).^[ [220]^7 ^]

Clinical Oncology Patients

   Human fecal and serum samples were obtained from 20 patients with
   rectal cancer. All patients included in this study received the same
   standardized radiation therapy regimen, which was PTV‐GTV 50 Gy/25F and
   PTV‐CTV 45 Gy/25F for 5 weeks. The detailed information and the
   radiotherapy regimen of the oncology patients was also showed in Table
   [221]S7 (Supporting Information). Patients with typical clinical
   symptoms of RP after 5‐weeks radiotherapy were enrolled in the study,
   which was confirmed by endoscopic examinations of patients before
   radiotherapy and after the completion of radiotherapy. Representative
   endoscopy examination showed subsequent changes of the intestinal
   mucosa after pelvic radiotherapy, reflecting a clear phenotype of the
   disease (Figure [222]S18A, centre). After exposure and radiation,
   mucosal edema appeared in the rectum and colon, with discoloration and
   blackening of the mucosa, disappearance of normal blood vessel
   distribution, pallor of the mucosa, disappearance of mucosal folds, and
   irregular superficial ulcers. The same standard was used to collect
   samples from all RP patients, namely, all samples were collected one
   day prior to radiotherapy (BT) and one day after the completion of
   5‐weeks radiotherapy (RT), respectively. Fecal samples were collected
   in stool collection containers and transferred immediately to a
   refrigerator with a temperature of −80 °C until use. After collection,
   blood samples were placed in a refrigerator at 4 °C, centrifuged at
   3000 rpm at 4 °C for 10 minutes, and the upper serum layer was removed
   and stored in a −80 °C refrigerator until use. The experiment was
   approved by the internal review and the ethics boards of Ethics
   Committee of the Sixth Affiliated Hospital of Sun Yat‐sen University
   (2021ZSLYEC‐262) with patient informed consent.

Sample Collection

   Fecal and serum samples were collected at 6 weeks post‐radiation
   because the clinical scores of RP mice at 6 and 8 weeks showed no
   difference (Figure [223]S18B, Supporting Information). Samples from
   each animal were either preserved at −80 °C for microbiota composition
   analyses, metabolite profiles, or processed for subsequent
   investigations. To do this, the samples were combined by vortex,
   homogenized in saline, then centrifuged at 800 rpm for 5 minutes to
   pellet bacterial cells. Pellets were washed in PBS, reconstituted in an
   equivalent volume of brain‐heart infusion (BHI) liquid media (with 50%
   glycerol), and kept at −80 °C for co‐culture investigations. Mice in
   the radiation, control, and treatment groups were put to death six
   weeks after being exposed to radiation. Blood was collected,
   centrifuged for 10 min at 3000 rpm after being held at ambient
   temperature for 30 min. The supernatant serum was kept at −80 °C for
   metabolite profiles and cytokine assays. Colonic segments and liver
   tissue were promptly frozen in liquid nitrogen and kept at −80 °C. The
   rectal tissue 1 cm above the anus was excised, washed, and sliced into
   two equal length specimens along the craniocaudal axis. For western
   blotting or quantitative PCR, one specimen was preserved in liquid
   nitrogen; for histopathological investigation, the second specimen was
   fixed in 4% formaldehyde.

Histopathological Analysis

   Rectal samples, fixed in 4% formaldehyde, were used for hematoxylin and
   eosin (HE) and Masson's trichrome staining. Sections of the tissues
   that were four microns thick were cut using a rotary microtome (Leica,
   Germany), placed on slides for staining and then examined under a light
   microscope in accordance with established techniques (DM2500, Leica). A
   certified pathologist double‐blindly carried out histological
   examinations. Collagen, cell nuclei, and cytoplasm received blue, dark
   purple, and red/pink, respectively, colors in the Masson's trichrome
   staining. After microscopic analysis of the stained slides, the
   radiation injury score (RIS), which was modified from Langberg
   et al.,^[ [224]^48 ^] was computed to assess the histological
   alterations. Such changes were based on Masson and HE staining, which
   revealed typical histological characteristics of RP lesions, including
   mucosal ulceration, infiltrating inflammatory cells, edema, vessel
   stenosis, and submucosal fibrosis (Table [225]S1, Supporting
   Information).

Cytokines Assay

   The circulating concentrations of IFNγ, IL10, IL12p70, IL1β, IL2, IL4,
   IL5, IL6, KC/GRO, and TNFα were measured using the MSD V‐Plex
   Proinflammatory Panel 1 Mouse Kit (Meso Scale Diagnostics, Cat#
   K15048D). Serum cytokine IL6 level was measured using an ELISA kit
   (Cloud‐Clone Corp, SEA079Mu) as directed by the manufacturer's
   instructions.

Fecal 16S rDNA Gene Sequencing

   Fecal sample preparation and 16S rDNA gene sequencing were carried out
   as it was previously described.^[ [226]^33 , [227]^34 ^] The QIAamp
   Fast DNA stool Mini Kit (Qiagen, Cat# 51604) was used to extract
   bacterial DNA from fecal samples in accordance with the manufacturer's
   instructions. By utilizing barcoded primer pairs that targeted the
   V3‐V4 region of the 16S rDNA gene, fecal DNA samples were amplified by
   polymerase chain reaction (PCR). The Illumina NovaSeq6000 was used to
   sequence the PCR amplicons (Novogene Co., Ltd., China) based on
   standard protocols. QIIME (version 1.9.1) was used to examine the
   resultant bacterial sequence fragments. The 16S rDNA variable region
   primers used to target the region's V3‐V4 were listed in Table [228]S8
   (Supporting Information). The raw sequences were saved in the Sequence
   Read Archive database.

   Principal co‐ordinates analysis (PCoA) plots were used to show the
   generated matrices. Heatmaps among gut microbiota, cytokines and
   metabolite results were created with the statistical computing
   environment R's “heatmap” function. The effect size (LEfSe) of linear
   discriminant analysis (LDA) was utilized to find changes in relative
   abundance. The Spearman's correlation coefficient was used to calculate
   correlations. Random forest analysis was performed using the R Studio
   (v 3.5.0).

Immunoblot Analysis

   Immunoblotting was carried out as it was previously reported.^[
   [229]^33 ^] Briefly, rectal tissues collected from the indicated mice
   were lysed, and total protein was isolated and measured. Extracted
   proteins were electrophoresed in 10% SDS‐PAGE with loading buffer
   before being transferred to PVDF membranes (Bio‐Rad) and then put in 5%
   fat‐free milk to block for one hour. The membranes were treated with
   primary antibodies overnight at 4 °C before being incubated for 1 hour
   at room temperature with a secondary antibody (Cell Signaling). The
   protein bands were visualized using an ECL kit (Millipore). Images were
   captured using a Tanon5200 machine (Tanon, China). ImageJ 1.43 software
   was used to analyze the data.

Antibodies

   Rabbit‐signal transducer and activator of transcription 3 (STAT3)
   (#4904, Cell Signaling), rabbit‐phosphorylated p‐STAT3 (#9145, Cell
   Signaling), rabbit‐glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH)
   (#5174, Cell Signaling), HRP‐linked anti‐rabbit IgG antibody (#7074,
   Cell Signaling), anti‐IL6 MAb Tocilizumab (HY‐P9917, MCE), and isotype
   control IgG (HY‐[230]P70251, MCE) were used.

IL6 Receptor Antagonist Treatment

   After irradiation of RP mice, the animals were injected with 5 mg k^−1g
   anti‐IL6 MAb, IgG antibody control, or saline every two days per week
   for six weeks. The clinical score plots and samples were then
   collected.

Co‐Culture of Epithelial Cells with Microbiota

   Human HIEC‐6 intestinal epithelial cell line was obtained from the
   American Type Culture Collection and cultured in Dulbecco's modified
   Eagle's medium (11965092, Gibco) with 10% fetal bovine serum (12483020,
   Gibco) at 37 °C in 5% CO[2] and 95% air for 24 h. The HIEC‐6 cells were
   irradiated (8 Gy) using an RS2000 device. For the co‐culture
   experiments, the epithelial cells were seeded (5×10^5 cells) into Nunc
   EasYFlask (25 cm^2) (Thermo Scientific), and then the medium was
   removed and the bacterial cell suspensions (MOI = 10^2) obtained from
   irradiated and un‐irradiated mice were added. For irradiated cells
   treated with A. muciniphila, active or pasteurized A. muciniphila (MOI
   = 10^2) were added and co‐culture at 37°C in 5% CO[2] and 95% air for
   24 h. For irradiated cells co‐cultured with fecal suspensions derived
   from RP mice, the feces were collected from RP mice post radiation for
   6 weeks as described in sample collection section, and the bacterial
   cell suspension were prepared and added to irradiated cells at an MOI =
   10^2, then were kept at 37°C in 5% CO[2] and 95% air for 24 h. For
   irradiated cells treated with bacterial cell suspension plus A.
   muciniphila at an MOI = 10^2, in which A. muciniphila was at a ratio of
   2%, then were co‐cultured at 37°C in 5% CO[2] and 95% air for 24 h.
   Cell samples were then collected after the co‐culture treatments.

RNA Extraction and Real‐Time PCR

   TRIzol reagent (Takara, Japan) was used to extract total RNA from
   tissues or cells, and a Nanodrop spectrometer was used to measure total
   RNA concentration. Using a Promega cDNA Synthesis Kit, first‐strand
   cDNA was created by following the manufacturer's instructions. To
   measure the expression of IL6 and GPR43, quantitative real‐time PCR was
   carried out in triplicate using SYBR Green. The endogenous control
   GADPH was utilized to standardize gene expression. The primers used for
   IL6, GPR43, and GAPDH were listed in Table [231]S8 (Supporting
   Information). The data were evaluated using an ABI StepOnePlus
   real‐time PCR system (Applied Biosystems).

Liquid culture of A. muciniphila

   A. muciniphila (ATCC BAA‐835) was cultured in BHI medium at 37 °C under
   anaerobic conditions, as it was previously described.^[ [232]^34 ^] The
   growth curve (OD[600]) of A. muciniphila was established every 12 h
   using a Genesys spectrophotometer (Thermo Scientific). For detecting
   the effects of FabG on 3HB concentration, A. muciniphila cells in
   stationary phase were given escalating doses of epigallocatechin
   gallate (EGCG) for 24 h as previously reported.^[ [233]^28 ^] For gut
   colonization of mice by A. muciniphila, cultures were cleaned and
   concentrated in anaerobic saline containing 25% (v/v) glycerol under
   anaerobic conditions. In addition, an equivalent amount of A.
   muciniphila cultured in the same medium was heat‐inactivated at 70 °C
   for 30 minutes.

Antibiotics Treatment

   For the germ‐free mice model, a mixture of ampicillin (1 mg mL^−1,
   Sigma, CAS# 7177–48‐2), neomycin sulfate (1 mg mL^−1, Sigma, CAS#
   1405–10‐3), vancomycin (0.5 mg mL^−1, Sigma, CAS# 1404–90‐6), and
   metronidazole (0.2 mg mL^−1, Sigma, CAS# 443–48‐1) was added to
   sanitize the mice's drinking water. Three times every week, the
   solutions and bottles were replaced. UR mice were drinked with
   antibiotics mixture or water for 6 weeks. RP mice were drinked with
   antibiotics mixture for 1 week after radiation, and then treated with
   A. muciniphila or 3HB for another 5 weeks. A. muciniphila treated group
   were drinked with normal water and 3HB treated group were drinked with
   antibiotic water in the duration of treatment. Fecal pellets were
   cultivated on blood agar plates, resuspended in BHI + 50% glycerol
   (0.1 g mL^−1), and incubated for 48 hours at 37 °C in both aerobic and
   anaerobic conditions to test antibiotic sensitivity.

A. muciniphila Oral Administration

   For the treatment of mice using A. muciniphila, the cells were grown in
   BHI broth medium at 37 °C under anaerobic conditions, as it was
   previously reported.^[ [234]^33 , [235]^34 ^] Gut colonization of
   antibiotics pre‐treated RP mice by A. muciniphila was conducted by oral
   gavage with 200 µL bacterial suspension containing 2 × 10^8 cells or
   saline 3 times per week. Clinical score plots and samples were
   collected after five weeks of A. muciniphila treatment.

Detection of A. muciniphila

   According to the previous investigation, the abundance of A.
   muciniphila in fecal samples was evaluated using quantitative PCR
   (qPCR).^[ [236]^33 ^] Following the manufacturer's instructions,
   genomic DNA was extracted from tissues or feces using the QIAamp DNA
   Tissue or Stools Mini Kit (Qiagen). SYBR Green‐based targeted qPCR
   devices were used. Primers were utilized as previously described.

Fluorescence In Situ Hybridization (FISH)

   FISH and immunofluorescence of A. muciniphila were performed as it was
   described previously.^[ [237]^33 ^] Briefly, deparaffinization was
   performed on paraformaldehyde‐fixed paraffin‐embedded colon tissue
   slices (5 mm). To identify bacterial colonization, a
   fluorescein‐labeled oligonucleotide probe targeting one area of the A.
   muciniphila 16S rDNA gene was utilized. As a negative control,
   nonspecific hybridization was detected using the non‐EUB probe. In situ
   hybridization was carried out at 50 °C overnight. Slides were coated
   with ProLong Gold with DAPI (Invitrogen), sealed with coverslips, and
   permitted to dry overnight at 4°C in the dark before being photographed
   using a confocal microscope (LSM 880 with Airyscan).

Transmission Electron Microscopy (TEM)

   Transmission electron microscopy (TEM) was used to investigate A.
   muciniphila colonization as it was described previously.^[ [238]^33 ^]
   Briefly, the intestinal tissues of mono‐colonized mice were cut into
   2–3 mm cubes and promptly fixed at 4°C overnight using a 0.1 M sodium
   cacodylate buffer containing 3% glutaraldehyde. The samples were then
   implanted in epoxy resin after being post‐fixed in a 2% osmium
   tetroxide buffered solution. Then, the samples were processed in the
   manner previously described. FEI Tecnai G2 Spirit BioTwin 634 was used
   to create electron micrographs.

3HB Treatment and Determination

   RP mice pretreated with antibiotics were orally gavaged with 200 µL 3HB
   (150 mg k^−1g body weight) or saline 3 times per week. Clinical score
   plots and samples were collected after five weeks of 3HB treatment.
   Irradiated HIEC‐6 cells, with or without 3HB (10 mM), were cultured in
   5% CO[2] and 95% air conditions for 24 h, and then the expression of
   IL6 was assessed. Each gram of feces and liver tissue were homogenized
   with 1 mL saline. Ten minutes of centrifugation at 4 °C of 12,000 rpm
   to obtain the supernatants for further analysis. The concentration of
   3HB in mice sera, feces and liver and supernatant of A. muciniphila was
   quantified using an ELISA kit (Cloud‐Clone Corp, CEB022Ge), as directed
   by the manufacturer's instructions.

Metabolomics Analysis

   The samples stored at −80°C were thawed on ice. To precipitate the
   proteins, 300 µL of methanol were added to 100 µL of each sample,
   vortexed for 3 min, and remained for 10 min at room temperature.
   Following 20 minutes of centrifugation at 4°C of 12,000 rpm, the
   supernatants were analyzed by LC‐MS analysis, following the
   manufacturer's instructions (Metware Biotechnology Co., Ltd. Wuhan).

   The first‐ and second‐order spectra obtained by mass spectrometry were
   qualitatively analyzed using the self‐database, metware database
   (MWDB), and the public database of metabolite information. Multiple
   reaction monitoring (MRM) triple quadrupole mass spectrometry was used
   to quantify the metabolites. The statistical function prcomp inside R
   was used to perform unsupervised principal component analysis (PCA).
   VIP ≥ 1, p‐value < 0.05, and absolute Log2FC ≥ 1 were used to identify
   significantly regulated metabolites across groups. The Mann‐Whitney
   U‐test was used for each metabolite to compute the p‐value for
   significance. VIP values were calculated using the R package
   MetaboAnalystR and retrieved from the OPLS‐DA data, which included
   included score plots and permutation plots. The cor function in R was
   used to generate the Spearman correlation coefficient between samples,
   which was then shown as heatmaps. The KEGG database
   ([239]http://www.kegg.jp/kegg) was used to annotate the identified
   metabolites, and the annotated metabolites were then linked to the KEGG
   pathway database. A hypergeometric p‐value test was used to find
   pathways that had significantly modulated metabolites.

GPR43 Agonist Treatment

   For irradiated HIEC‐6 cells co‐cultured with A. muciniphila or treated
   with 3HB, vehicle [10% DMSO (D2650, Sigma‐Aldrich) + 40% PEG300
   (HY‐Y0873, MCE) + 5% Tween 80 (HY‐Y1891, MCE) + 45% saline] was used to
   dissolve GPR43 agonist to prepare 10 mM storage concentration.
   Irradiated cells were treated with 10 µM GPR43 agonist (4‐CMTB,^[
   [240]^49 ^] HY‐P1125, MCE) or vehicle and then the expression of IL6
   was assessed. To determine the effect of the GPR43 agonist on mice, RP
   mice pretreated with antibiotics were orally gavaged with 200 µL 3HB
   (150 mg k^−1g body weight) or saline and were injected with 200 µL
   4CMTB (10 mg k^−1g body weight) or vehicle. Clinical scores and samples
   were obtained following five weeks of treatment.

Generation of GPR43 KO in Human HIEC6 Cell Line

   The CRISPR–Cas9‐based KO strategy was used for GPR43 knock out in human
   HIEC6 cells. The sgRNAs targeting GPR34 were designed using Benchling.
   CRISPR‐Cas9 sgRNA sequences were listed in Table [241]S8 (Supporting
   Information). CRISPR/Cas9 was performed using LentiCRISPRv2 vector.
   sgRNAs were cloned into lentiCRISPRv2 according to the lentiCRISPRv2
   cloning protocol. Lentiviruses were then generated by transfection of
   HEK293T cells with lentiCRISPRv2‐sgGPR34 and the packaging plasmids
   psPAX2 and pMD2.G using Lipofectamine 3000 according to the
   manufacturer's instructions. Viral particles in the cell culture
   supernatant were filtered with 0.45 um filters to remove cellular
   debris. HIEC6 cells were transduced with lentiviruses in the presence
   of polybrene (10 ug mL^−1), 72 h after infection, cells were selected
   under 2 ug mL^−1 puromycin for 7 days. The cells were cultured for at
   least five passages to establish the stable cell lines.

Data Analysis

   GraphPad Prism 6 (GraphPad Software, USA) was used for statistical
   analysis. At least two biological replicates were conducted in vivo
   animal experiments, and at least three biological replicates were
   conducted in vitro cell, bacterial, and molecular experiments. The
   figure legends indicated the number of animals (n) used in the studies.
   One dot or lane indicated one mouse or sample. All values were shown as
   the mean ± SEM, with *P < 0.05, **P < 0.01, ***P < 0.001, ****P <
   0.0001; ns, not significant. The D'Agostino‐Pearson omnibus test was
   used to determine data normal distributions. If statistical
   significance between two groups was not mentioned in the figure
   legends, it was calculated using an unpaired, two‐tailed Student's t
   test, Mann‐Whitney test, or permutation multivariate analysis of
   variance (PERMANOVA) test, and significance of more than two groups was
   established using one‐way ANOVA or two‐way ANOVA in GraphPad Prism with
   default setting depending on experience.

Conflict of Statement

   The authors declare no conflict of interest.

Author Contributions

   Z.G., H.M.W., H.W., Y.L. planned the experiments. Z.G., C.C., J.C.,
   Z.J., L.C., Y.W., H.C., L.H., Y.Z., X.L., H.Z. performed the
   experiments and analysis. Z.G., Y.L. wrote the draft of the manuscript.
   All authors contributed to the edits of the manuscript.

Supporting information

   Supporting Information
   [242]ADVS-11-2306217-s001.pdf^ (3.7MB, pdf)

Acknowledgements