The gut microbiota as a booster for radiotherapy: novel insights into radio-protection and radiation injury

Elegantly-designed experimental studies establish the strong linkage between the gut microbiota and radiation injury. Resistance to lethal dose total body radiation (TBI) in germ-free mice compared to conventionally-raised mice has been reported since 1950s [4]. In line with this finding, pre-treatment with antibiotics to clear the gut microbiota also significantly improves survival in mice exposed to sublethal dose TBI [5] and accelerates recovery of oral mucositis after snout radiation [6]. Moreover, Fan et al.conducted a series of works by introducing a third factor which interacts with the gut microbiota and thus affects radio-sensitivity of normal tissues to reveal the bridging role of gut microbiota in radiation injury. In their reports, circadian rhythm disturbances [7] and food additive polysorbate-80 [8] disturbing the gut microbiota lead to radiation injury deterioration. Whereas modified diets [9, 10], anti-cholesterol agent simvastatin [9] and anti-oxidant molecules [11‐14] mitigate radiation injury in a microbiota-dependent manner. Noteworthy, a series of gender-specific phenotypes in relation with gut microbiota interplaying with radiation injury is repeatedly reported in Fan’s animal models [5, 9, 10]whose underlying mechanism is unclear. Whether similar manifestations exist in human as a possible explanation for the sex differences in treatment efficacy and toxicity [15] remain to be investigated. Nevertheless, such a phenomenon reminds us that gender is of concern in unravelling the role of gut microbiota in radiation injury. We summarize the animal models used in experimental studies (Table 2), noticing that most studies chose male CB57BL/J6 (B6) mice in this field.

In addition to associations from deprivation and interventional assays, selective microbiota supplementation assays provide more robust evidence. For example, mono-association experiments revealed that Fusobacterium sp., Streptococcus faecalis, Escherichia coli and Pseudomonas sp. respectively contributes to radiation injury augmentation, whereas Clostridium sp., Bifidobacterium pseudolongum and Lactobacillus acidophilus contributes to injury amelioration [16]. Furthermore, repeated transplantations of feaces from gender-matched healthy mice after radiation efficiently mitigate intestinal and hematopoietic toxicity by enhancing angiogenesis and maintaining gut barrier integrity of the radiated mice [5]. On the other hand, transplantation of dysbiosis feaces from radiated mice [17] or from polysorbate-80-treated mice [8] before radiation transmits susceptibility to radiation injury to the recipients.

Compared to cross-sectional studies, repeated sampling in a longitudinal design has higher statistical power, and some studies associate the microbial dynamics throughout radiotherapy with radio-dosimetry [18‐22] and radiation injury [17, 23‐25]. Reportedly, abdominal dose of  > 1.5 Gy is a risk factor of RE and the dose threshold for gastrointestinal lining loss is 6 Gy [26]. The threshold for hematopoietic dysfunction is 0.5 Gy and the exquisitely radio-sensitive lymphocytes are maximally depressed when exposed to radiation of ~ 2 Gy, which is the standard fractionated regimen adopted in the clinic [27]. According to historic reports and our clinical experience, patients typically present gastrointestinal symptoms since the first or second week of radiotherapy [26] and may develop progressive myelosuppression almost immediately within the first week of radiotherapy [27]. Therefore, timely identification of high-risk population (> 2 Gy absorbed dose) in need of medical intervention is critical both for ensuring safety of radiotherapy in clinical settings and for appropriate allocation of medical resources after accidental nuclear exposure. To this end, Lam et al. [19] developed two indices respectively named as “acute ratio”, which climbs within the first week after radiation, and “chronic ratio” which increases from the second till the third week, based on murine faecal microbes whose abundances are significantly and persistently altered after TBI. The two indices effectively determine exposure to single-fraction of 10 Gy radiation and multiple-fractions of 18 Gy/6Fx radiation. Interested in the feasibility of microbiota as radio-dosimetry, Tian’s group [20] using rodent models reported linear correlations between radiation dose and microbial abundances measured at day 3.5 following radiation, including negative association with Eubacterium xylanophilum and positive association with Escherichia Shigella (p-value = 0.046 and 0.026, respectively). These microbial abundances also parallel with the post-radiation lethality and intestinal histopathologic features. Likewise, Carbonero et al. [21] found in minipig and macaque models significant correlations between radiation dose and microbial abundances measured at day 3 following TBI, including the positive association with butyrate-producing Clostridium XIVa as well as the negative association with Prevotella, and the abundance of Prevotella is suggested as a potential positive prognostic biomarker of post-radiation survival.

A further step made by Epperly et al. [25] is the inclusion of microbiota information into post-radiation survival models which efficiently improves the predictability. They set the repeatedly measured microbiota abundances as time-varying covariates combined with static covariates of several radiation mitigators including granulocyte-colony stimulating factor (G-CSF), programmed cell death inhibitors JP4-039, necrostatin-1 and baicalein in a Cox proportional hazards model. Results showed that both the abundance of Lactobacillus (HR = 0.15, p-value = 0.001) and butyrate-producing Ruminococcus (HR = 0.85, p-value = 0.04) are independently robust predictors of increased survival, whose addition into the model improves the concordance index, even outweighing the significance of therapeutic G-CSF or the triple regimen of JP4-039/baicalein/necrostatin-1 (p-value = 0.08 and 0.09, respectively). Surprisingly, abundance of Akkermansia is a deleterious predictor of survival (HR = 2, p-value = 0.046) in experiments incorporating G-CSF although the prognostic value of Akkermansia is insignificant in experiments using programmed cell death inhibitors only. Given the marginally significantly decrease in Akkermansia in survivors treated with G-CSF, the possibility that the interplay between Akkermansia and G-CSF renders this microbe the unexpected correlation with radiation-induced death needs to be clarified.

To summarize, preclinical results strongly support that the gut microbiota is closely involved into the development of radiation injury. Of note, over 96% of the microbial biomarkers identified in murine faeces have been matched to genera existing in human [19], highlighting the translational value of these laboratory results.

Two small-sample-size studies [28, 29] incorporating mixed cohorts of gynecological, colorectal and anal cancers first contributed to this field by presenting that pre-existing faecal microbiota features help identify patients predisposed to acute radiation-induced diarrhea, and that patients who prone to develop diarrhea have lower Shannon diversity index (SDI) and higher Firmicute/Bacteroides ratio in their baseline faecal specimens compared to those who do not. Two subsequent studies with uniform cohorts of cervical cancer patients confirmed that dysbiosis in the baseline faecal specimens from patients prone to acute RE is featured with significantly reduced α diversity (including SDI) and increased β diversity [30], and that increased SDI is an independent predictor for better near-term gastrointestinal function, when all timepoints data were analyzed together [31]. However, no association between the α diversity and occurrence of chronic RE is found in another mixed cohort of 59 patients [32], which emphasizes further microbiota exploration on the chronic RE in larger-sample-size uniform cohorts. A translational contribution deserving attention is the establishment of the first model for predicting Grade2 acute RE in a cohort of 17 gynecological cancer patients based on microbial features measured after radiotherapy [33], whose area under the receiver-operating characteristics curve (AUC) metric reaches 0.87, indicating robust predictive performance.

In prostate cancer patients, Ferreira et al. [34] also observed a trend towards lower α diversity (Chao index) after radiotherapy in both acute and chronic symptomatic patients compared to their non-symptomatic counterparts despite that SDI was not analyzed. Compositionally, patients with acute and chronic RE symptoms are both rich in short chain fatty acid (SCFA)-producers, including Clostridium IV, Roseburia and Phascolarctobacterium whose abundances all significantly correlate positively with the elevated acute symptoms and all demonstrate more drastic post-radiotherapy decline in patients with persistent symptoms. Overabundances of Roseburia is even an independent predictor of occurrences of severe chronic RE. Moreover, the SCFA-producers negatively correlate with IL-15 and positively correlate with eotaxin in rectal mucosal tissues sampled from chronic RE patients, indicating an inflammation regulatory role for these microbes. Functionally, there also exists an increasing trend of butyrate metabolism and propionate metabolism associated with aggravating acute symptoms, whereas fatty acid metabolism significantly reduces with aggravating chronic symptoms.

According to a small-sample-size pilot study on rectal cancer [35], the SCFA-producing family Lachnospiraceae belonging to Clostridia class and the lactate-producing Bifidobacterium are overrepresented in patients with milder acute diarrhea compared to the severe counterparts. A subsequent expanded analysis showed that the microbial diversity (including SDI) negatively correlates with the severity of both myelosuppression [36] and chronic gastrointestinal toxicity significantly (data not published), and non-significantly with acute gastrointestinal toxicity [36]. Compositionally, Akkermansia muciniphila, Bifidobacterium adolescentis and Coprococcus eutactus are overrepresented at baseline in patients with milder myelosuppression and in patients with milder acute gastrointestinal toxicity, respectively compared to their corresponding severe groups; whereas β-glucuronidase-producing microbes including Escherichia fergusonii related with reactivation of the irinotecan end-product, SN38G, are overrepresented in patients with severe myelosuppression and in patients with severe acute gastrointestinal toxicity compared to the mild counterparts [36]. Established on the abundances of discriminative microbes measured prior to radiotherapy, toxicity prediction classifiers perform inspiringly with the AUC for predicting Grade2 + myelosuppression reaching 0.89, and that for predicting Grade3 + acute gastrointestinal toxicity reaching 0.86. Another study on anal squamous cell cancer [37] revealed a trend towards higher baseline microbial diversity (including SDI) in patients predisposing to Grade 2 + acute anal dermatitis compared to their Grade 0–1 counterparts, with Pielou evenness reaching significance. There is no significant association between severity of acute RE with microbial diversity. However, Clostridia class and Clostridiales order are enriched at baseline in patients with severe acute RE, which contrasts to findings in rectal cancer. Notably, this study took anal swabs from the tumor site, which is different from the faecal specimens used in other studies. Given the unique biological feature of HPV-driven anal squamous carcinoma, and the limited sequencing volume from anal swabs, as well as only 12 cases for RE analysis and 16 cases for anal dermatitis analysis, findings in this study merit further confirmation.

Collectively, some microbial features have been consistently identified in both animal models and clinical cohorts to be associated with amelioration of radiation injury, including lactate-producers [16, 25, 35, 36], SCFA-producers [16, 25, 29‐31, 34‐36] and Akkermansia [23, 36] (Fig. 1), as well as high α diversity [17, 20, 36] especially SDI [28, 30, 31, 33]. Nonetheless, disagreements also exist on the differential distribution of some microorganisms in different symptom groups, which will be discussed in the next section. Regarding the discrepancy on the statistical significance for α diversity, we deduce that limited sample size, as well as differences in specimen sources, study populations (varied race, gender, cancer types, etc.), methods for toxicity assessment and group stratifications may account. Summary in Table 3 may provide clues for designing confirmative studies in the future. Despite use of baseline specimens to profile the pre-existing unaffected microbiota in most clinical studies, the temporal relationship is insufficient for inferences on causality. Hence, we should interpret a particular microorganism to be beneficial or deleterious cautiously, and unravelling the molecular mechanisms is crucial.

Unravelling the convergent biological functions behind the divergent microbial compositions may be more generalizable given the variability of microbiota profiles between individuals. In post-radiation murine plasma and urine, increased metabolism pathways include primary bile acid (BA) biosynthesis, taurine metabolism and lipid metabolism [39‐41] whereas decreased metabolites include indole-3-propionic acid (IPA), 3-indoxylsulfate and indole lactate [39] which are transformed by the gut microbiota [41] from tryptophan, a type of aromatic amino acid. In irradiated murine feaces, abundances of primary BAs and taurine rise whilst abundance of secondary BAs decline with the elevated radiation dose, while aromatic amino acid metabolism and BA secretion pathways are co-activated [18]. The dramatic decline in murine faecal SCFAs [8, 42] (including acetate, propionate, butyrate and valerate) and IPA [43] after radiation which are reversed by faecal microbiota transplantation (MT) also support radio-protective attributes for these metabolites. Concordantly, in feaces from cervical cancer patients susceptible to acute RE, SCFAs demonstrate a significantly sharper decline while BAs increase after radiotherapy [44]. This finding aligns with Ferreira’s findings [34, 45] on prostate cancer that intestinal SCFA-producers and faecal butyrate level both decline more dramatically in patients with acute RE, and with Shi’s finding [35] on rectal cancer that BA biosynthesis pathway enriched in patients with milder diarrhea. Notably, butyrate alteration is not detected in plasma or urine specimens probably indicating a localized effect of this metabolite [45]. Regarding lipid metabolism, the predictive performance of the lipid-like metabolite ptilosteroid A for acute RE in gynecological cancer patients underscores its significance [33]. To summarize, post-radiation metabolomics from both rodent and human specimens consistently highlight the potential values of SCFAs and aromatic amino acid-derived indole compounds, as well as metabolism dysregulation in BAs and lipids in radiation injury.

Both SCFAs and indole compounds are produced from gut microbiota and regarded as free-radical scavengers. SCFAs are the preferred energy source to colonocytes and serve as histone deacetylation inhibitors and G-protein-coupled-receptor (GPR) ligands. By inactivating NF-κB pathway, SCFAs upregulate anti-inflammatory cytokine IL-10, downregulate pro-inflammatory cytokine IL-6, enhance anti-microbial functions of macrophages and regulate colonic Tregs homeostasis [46]. Whether these anti-inflammatory properties contribute to radio-protection requires investigations. Fan’s group proved that butyrate attenuates post-radiation intestinal injury and inflammation, reverses dysbiosis and enhances gut integrity by activating GPR41/43 pathway [8], and that both valerate [42] and IPA [43] impart hematopoietic and gastrointestinal radio-mitigation by restoring microbial homeostasis. Mechanistically, valerate re-elevates the post-radiation decline of KRT1, a component of cytoskeleton capable of sustaining epithelial integrity against stresses whereas IPA exerts radio-protection by re-activating the down-regulated pregnane X receptor/acyl-CoA-binding protein pathway to enhance gut integrity. In parallel, selective post-radiation enrichment of Lachnospiraceae and Enterococcaceae families renders corresponding mice the “survivor phenotype” represented as attenuated hematopoietic and intestinal impairment after TBI. Butyrate and propionate derived from Lachnospiraceae are partly responsible by reducing DNA damage and reactive oxygen species release, and tryptophan-derived indole compounds (kynurenic acid and indole-3-carboxaldehyde) are also proved to be radio-protective albeit lacking mechanistic explanations in this study [47]. Reportedly, indole compounds as ligands to aryl hydrocarbon receptor-mediated transcriptional activity enhance proliferation of intestinal stem cells (ISCs) [48] and promote anti-oxidant and anti-inflammatory immune responses [49, 50]. One example is 3,3′-diindolylmethane which is hydrolyzed by gastric acid from indole-3-carbinol, imposes radio-protection by promoting post-radiation DNA damage repair, ISCs survival and anti-oxidative reactions as well as reversing radiation-induced gut dysbiosis [11].

BAs participate into absorption and metabolism of dietary lipids and cholesterols, and affect the detoxification of xenobiotics and the regulation of intestinal inflammation as well. Exposure of enterocytes to primary BAs aggravates inflammation in IBD whereas secondary BAs possess potent anti-inflammatory capacities [51]. The increase in primary BAs is generally attributed to post-radiation malabsorption and gut dysbiosis [18]. What deserves attention is that the beneficial Lactobacillus and Ruminococcaceae both participate into production of secondary BAs, whose post-radiation decline may therefore aggravate enteritis [18, 25]. Nonetheless, there is still a lack of direct evidence on how BA metabolism dysregulation induced by dysbiosis affects radiation injury.

Regarding the role of microbiota-regulated lipid metabolism in radiation injury, the elevated energy demand during post-radiation repairment phase may be a breakthrough although supportive evidence is lacking. Fasting-induced adipose factor (Fiaf) is an important lipid metabolism regulator secreted from epithelium which stimulates fat mobilization, triglycerides deposition, endothelial survival and vascular sprouting. According to Crawford and Gordon [4], the existence of gut microbiota suppresses the expression of Fiaf, which then increases apoptosis of endothelial cells and lymphocytes in the intestinal villi after TBI thereby aggravating enteritis and shortening post-radiation mice survival. However, no subsequent research unravels the molecular mechanism for Fiaf’s radio-protection. Another assumption is that undesirable microorganisms may radio-sensitize normal tissues by exacerbating lipid peroxidation, which is a major target for oxidative stress and radiation injury. For example, Pseudomonas aeruginosa, an opportunistic pathogen commonly habituating in hospitalized immunocompromised patients, induces ferroptosis, augments post-radiation gut barrier breach and ensuing lethality through producing 15-lipoxygenase. 15-lipoxygenase also promotes a transition to pro-inflammatory milieu by modulating inflammation-related lipid signaling, immunocyte recruitment and cytokine production in small intestine [52].

Experimental evidence suggests that pre-treatment with Lactobacillus mitigates radiation injury by improving gut dysbiosis, intestinal histopathology and by inhibiting endotoxemia occurrence and bacterial translocation [20, 53]. Radio-protective mechanisms concordantly highlight regulations of Lactobacillus on ISCs, which are radio-resistant and important for post-radiation gut repairment. In this process, mesenchymal cells play indispensable mediating roles by sensing microbial signals and activating downstream pathways [53‐56]. According to Ciorba’s study [53, 54], gut Lactobacillus releases lipoteichoic acid to activate the Toll-like receptor-2 (TLR-2)/MyD88 pathway in intestinal peri-cryptal macrophages, thereby promoting the mesenchymal stem cells to migrate from villi to crypts where ISCs located, in a CXCL12-CXCR4 chemotactic manner. Lipoteichoic acid also activates TLR-2 on the migrated mesenchymal stem cells to promote the production of cyclooxygenase-2 and the ensuing synthesis of prostaglandin E₂ (PGE₂). Locally secreted PGE₂ then binds E prostaglandin recepter-2 on ISCs to execute anti-apoptotic and pro-proliferative functions by activating epidermal growth factor(EGF)-Akt-Bax signaling [57]. Notably, PGE₂ is also radio-protective of hematopoietic stem cells via activating Wnt/β-catenin pathway and of cancer cells via activating EGF-ERK2-MAPK pathway [57]. In this study, however, Lactobacillus renders no radio-protection on hematopoietic stem cells and even delays the growth of subcutaneously implanted colon cancer cells [54]. A possible explanation is that Lactobacillus functions by producing PGE₂, which is unstable and only acts over a short distance. And hematopoietic and cancer cells are spared from exposure to PGE₂ in gavage studies with Lactobacillus. In another study [55], Lactobacillus-derived lactate interacts with GPR81 expressed on Paneth cells and mesenchymal cells thereby increasing the secretion of Wnt3 and Wnt2b. Supply of these Wnt ligands effectively promotes β-catenin nuclear translocation in ISCs [56] and consequently improves the epithelial regeneration capacity and mice survival after exposed to radiation plus methotrexate.

However, Bifidobacterium also producing lipoteichoic acid and lactate did not exhibit radio-protection when administered alone according to Ciorba’s study [53], suggesting that there may exist other mechanisms as replenishment. Reportedly, Lactobacillus ameliorates inflammation via downstream protein p40 to promote immunoglobulin A (IgA) secretion against pathogenic infections, to reduce inflammatory cell infiltration, cytokine secretion and cytokine-induced apoptosis by activating EGF-PI3K-Akt signaling [58, 59]; and triggers anti-oxidant responses [60] by producing superoxide dismutases [61] and PGE₂ [62]. Whether and how such anti-inflammatory and anti-oxidant properties contribute to radio-protection should be elucidated in future research.

Akkermansia muciniphila residing in the mucus layer participates into the production and degradation of mucus, with SCFA as the end products. As a next-generation probiotic candidate it alleviates inflammatory disorders and metabolic syndromes and improves therapeutic responses to immune checkpoint inhibitors [63, 64]. Yet in radiation injury, the specific role of Akkermansia seems paradoxical at the first sight [23‐25]. On the one hand, Epperly et al. reported that oral gavage of Akkermansia strikingly increases murine 30-day-survival rate from 0 to 100% following TBI [23]. Kim et al. [63] further demonstrated that Akkermansia effectively upregulates ISCs proliferation and differentiation towards secretory cells, leading to increased goblet cells and thickened mucus layer. Like lactate [55], pre-treatment with Akkermansia ameliorates intestinal injury in mice and organoids exposed to radiation plus methotrexate. Mechanistically, supplement of Akkermansia dramatically alters murine intestinal microbial composition and increases SDI, which in turn enhances the production of SCFAs. Of these, acetate and propionate effectively promote epithelial proliferation by increasing the Wnt ligands secretion from Paneth cells to activate the β-catenin-RAS-ERK signaling in ISCs. On the other hand, the expansion of Akkermansia is consistently observed in both radiated and colitis models [20, 23, 24], supporting this microbe in association with inflammation. Notably, Gerassy et al. [17] proved that radiation induces a proinflammatory dysbiosis featured by expansion of microbes including Akkermansia. Such dysbiosis is transmissible via faecal MT and enhances host secretion of IL-1β, a responsive mediator to radiation and colitis stimuli which deteriorates intestinal injury. Combining these findings, we deduce that the rise in Akkermansia may represent a repairment response towards the activation of inflammatory regulation, and a compensatory protection towards the shortage of SCFA-producers after radiation stimuli, rather than a deleterious factor in spite of its reported negative prognostic value in mice [25] and minipigs [21]. Given the extensive evidence in other disorders for this microbe improves prognosis by enhancing mucosal barrier and suppressing local inflammation, its role in radiation injury merits in-depth investigations.

Notably, the post-radiation decreasing trend of the order Clostridiales (including the family Clostridiaceae and the genus Clostridium) is concordantly reported in radiated mice [17‐19, 43] and radiated patients [31, 37]. Consistent results that Clostridia [35] and Clostridiales [29‐31, 47] are significantly enriched in hosts with milder post-radiation toxicities also lend strong support to this radio-sensitive microbial cluster as radio-protective, albeit with contradictory finding in one small-sample-size study investigating microbiome sampled from tumor site [37]. In literature, class Clostridia is perceived as beneficial by engaging into the degradation of taurine, which is a sulfur-containing amino acid involved in BA conjugation and possesses anti-oxidation and de-toxication properties [18], and Clostridium sporogenes is the only identified IPA-producing microbe. More importantly, bacteria in class Clostridia are commonly equipped with butyrate production capacity and thus involved into immunomodulation by promoting expansion of intraepithelial lymphocytes, IgA-producing cells and colonic Tregs [65]. However, Clostridiales is composed of mixed species including pathogens like Clostridium perfrigens as well. The specific radio-protective member(s) in Clostridiales and underlying mechanisms requires systemic investigations.

Of interest, Coprococcus belonging to the family Lachnospiraceae and the order Clostridiales, is significantly underrepresented at baseline in both gynecological and rectal cancer patients predisposing to acute post-radiation toxicities [30, 36], supporting this SCFA-producer being radio-protective. In prostate cancer, however, SCFA-producers were found to be overrepresented in both acute and chronic symptomatic patients throughout the treatment [32, 34]. Regarding the paradoxical finding, two interesting explanations proposed by Ferriera et al. merit validation: (1) the overrepresentation of SCFA-producers in feaces results from excessive microbe shedding secondary to the increased competition with pathogens, which is more common in patients vulnerable to radiation injury; (2) patients with pre-existing subclinical intestinal dysfunctions possess a compromised microbiota and may therefore have higher dependence on SCFA supply which is represented by the overrepresentation of SCFA-producers in their feaces, and this group may be predisposed to severe RE.

Another microorganism deserving attention is Sutterella, belonging to the phylum Proteobacteria, overrepresented in patients predisposing to acute RE following cervical or prostate radiotherapy [31, 34]. In rodent models, Sutterella is significantly enriched following rectum radiation and also positively correlated with colonic injury [17]. However, Sutterella abundance correlated inversely with occurrence of chronic RE in a mixed clinical cohort [34]. Reportedly, Sutterella degrades secretory IgA [66] which acts as an important local immunoregulator restricting pathogenic infections especially intracellular pathobionts like Fusobacterium. The Sutterella expansion and ensuing IgA low phenotype increases host’s susceptibility to colitis stimuli [66] and resistance to ulcerative colitis treatment [67], hence may account for the aggravation of acute RE. On the other hand, Sutterella possesses almost negligible capacity of promoting inflammatory cytokines production in IBD [67] which may explain its nonharmful role in the non-inflammatory pathogenesis of chronic RE although more investigations are needed to illustrate how this microbe influences radiation injury development.

Overall, evidence supports that the gut microbiota has magnificent impacts on radiation injury via: (a) regulating cytotoxicity and repairment, and therefore sustaining epithelial and mucus barrier to defend bacterial translocation; (b) modulating host inflammation and immune functions; (c) controlling post-radiation oxidative stresses. These functional pathways are not mutually exclusive and rather occur complementally, and we briefly summarize the mechanisms in Fig. 2 and Table 4.

Faecal MT is established as an effective preventive option for recurrent Clostridium difficile infection [68] and the first product Rebyota [69] has recently been approved by FDA. However, severe complications including deaths used to be reported and data from long-term follow-up is scarce. Protocol standardization is still under cautious investigations and freshly purified materials combining multiple delivery routes are recommended for the sake of efficacy and safety [68]. In radiation injury, two research groups from China reported the feasibility of washed faecal MT in treating chronic RE, which is proven to be safer, more precise and quality-controllable than traditional crude faecal MT [70]. In Ding’s report [71], five cases aged from 45 to 81 with persistent Grade 2–4 chronic RE symptoms (per RTOG criteria) after gynecological cancer radiotherapy had at least one course of single-donor MT through the naso-jejunal transendoscopic enteral tubing. Three cases achieved the primary endpoint of ≥ one-grade reduction in toxicity at eight-week follow-up, with only one adverse event of transient nausea. However, the three cases all relapsed at longer follow-up, indicating short-lasting efficacy of this approach. In this context, strategic modifications merit consideration: (a) commencing MT earlier, which requires in advance prediction of chronic RE since this disorder is progressive; (b) repeating the courses to consolidate colonization; (c) optimizing sources of donor based on knowledge of beneficial microbiota compositions. Supportive evidence comes from Zheng’s study [72] where one 45-year-old case with continuous blood loss due to Grade 2 chronic RE (per RTOG criteria) after cervical cancer radiotherapy underwent multi-donor MT via lower-gut delivery as the first course and mid-gut delivery as the remaining three courses. At six-month follow-up, the patient’s symptoms relieved without any adverse events. This case report demonstrates that repeated transplantations of a more diverse microbiota effectively and safely mitigate hemorrhagic chronic RE, which deserves confirmation in larger cohorts with longer follow-up.

Of interest, oral microbiota is also correlated with aggravation of post-radiation oropharyngeal mucositis in nasopharyngeal cancer [73, 74], and one Chinese group successfully conducted oral MT from healthy mice to radiated mice to alleviate oral mucositis [75]. However, oral MT in colitis colorectal cancer model impaired radiotherapy efficacy and aggravated gastrointestinal toxicity, which was attributed to Fusobacterium, a pathogenic commensal in oral cavity [76]. Therefore, clarification of ideal consortia for mitigating toxicity is a prerequisite for application of MT in radio-protection and radio-mitigation.

Probiotics containing identified live organisms may be preferrable to a mixture of uncertain components in feaces material. Lactate-producing probiotics as anti-diarrhea preventive strategy in radiotherapy patients is recommended in guidelines [77], nonetheless the low quality and heterogeneity of the analyzed clinical trials hamper its widespread application. Reports regarding influences of these probiotics on myelosuppression are scarce and debated [78, 79]. Also due to incomplete data, toxicity profile of probiotics has not been clearly summarized either. Potential risks in vulnerable population like immunocompromised cancer patients include systemic infections, deleterious metabolism, overactivated immunoregulation and gene transfer with hosts [80]. In our literature review, adverse events of probiotics use in radiotherapy patients include neutropenic infection [39], fever [78] and gastrointestinal discomforts [30] but transient and with a low incidence. Complementally, prebiotics and synbiotics also represent as viable options by fostering growth of desired microorganisms. Prebiotics containing inulin and fructo-oligosaccharide, which favor lactate-producers growth, shorten duration with watery stools but confer no significant improvements in radiation-induced symptoms in gynecological cancer patients [81]. However, synbiotics containing inulin, guar gum mixture and Lactobacillus reportedly attenuate acute radiation proctitis in prostate cancer patients [82]. These pilot randomized clinical trials warrant further investigations.

Up to now, some insightful implications have been accrued from probiotics practice: (a) mixture preparations outperform single-strain probiotic[83]; (b) Lactobacillus is preferred to Bifidobacterium [38, 53]; (c) prophylactic use outperforms salvage therapy [38]; (d) more prominent benefits in reducing severe diarrhea (Grade ≥ 2, per CTCAE) [84]. However, the unclear efficacy and safety require optimization of probiotics in dose, delivery and formulation. Massive mechanism-driven works also pinpoint additional radio-protective microbiota components other than Lactobacillus and lactate. We look forward to the development of next-generation probiotics which should be corroborated in well-designed high-quality clinical trials to change practice.

The un-predictable risk with live microorganisms prompts search for pure functional constituents from microbiota as alternatives. The soluble anti-inflammatory protein p40 purified from Lactobacillus by Yan’s group [59] represents an example, but it is still in the laboratory stage, as are most of the keystone metabolites discussed above. Moderate-quality evidence from one small-sample-size trial suggests that sodium butyrate enema probably cannot reduce acute gastrointestinal toxicities in men undergoing prostate radiotherapy [85], which requires further investigations. Of note, lactate, SCFAs and indole compounds all possess anti-oxidant capacities. Future investigations on these metabolites will have to focus on side effects, therapeutic time window and impacts on tumor control since they constitute the major obstacles for application of anti-oxidants in radio-protection, like the case with amifostine [2].

Mechanistic exploration on microbe-host interactive pathways paves way for developing selective agonists or antagonists for radio-protection. One attractive target is TLR signaling with downstream effector PGE₂. The TLR-4 ligand lipopolysaccharide produced by gram negative microbes [86], TLR-5 ligand derived from Salmonella flagellin [87, 88] and TLR-2 ligands like lipoteichoic acid from gram positive microbes [53, 54] as well as lipopeptide from Mycoplasma are outlooking radioprotectors [89]. Especially the engineered de-toxic derivative of Salmonella flagellin, Entolimod (CBLB502) is now under investigation with FDA [90]. Inhibitors of inflammatory cytokines and lipid mediators as well as microbial enzymes also promise to be tailored radio-protective strategies, including the FDA-approved IL-1β antagonist anakinra which can be administered in gut dysbiotic patients [17], and 15-lipoxygenase inhibitor baicalein which can be prescribed in Pseudomonas aeruginosa-infected patients [52].

Taken together, we are still at a very early stage on translating the knowledge of microbiota interacting with radiation injury into effective countermeasures. Procedure standardization and safety monitor for microbiota manipulation strategies require in-depth investigations. Both prophylactic use of probiotics and TLR agonists outperform post-radiation salvage [53, 87, 88], again highlighting the importance of prediction to tailor the radio-protective strategy according to radiation injury risk.