Leading article

RELEVANCE OF THE ORAL AND GUT Therapeutic potential of microbial MICROBIOMES IN PANCREATIC CANCER Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from modulation in pancreatic cancer Several recent studies have analysed the oral and gut microbiomes associated Vidhi Chandra ‍ ‍ ,1 Florencia McAllister ‍ ‍ 1,2,3 with pancreatic cancer risk and progres- sion.18 20–25 Periodontal disease has been recognised as a risk factor for pancre- INTRODUCTION the α-diversity of the gut microbiota is atic cancer and may be an initiator of 16–18 Pancreatic cancer, particularly pancreatic lower than that in healthy controls. oral microbial dysbiosis.25 A prospective ductal adenocarcinoma (PDAC), is an With respect to clinical outcomes, the population-­based nested case–control 1 aggressive disease with a poor prognosis. gut microbial α-diversity is higher in study demonstrated that the presence Surgery is the only potential curative treat- patients with melanoma that responded of Porphyromonas gingivalis or Aggrega- ment, but it is only possible in patients to immunotherapy than in those unre- tibacter actinomycetemcomitans in the with early-­stage disease, and most patients sponsive.19 In this review, we highlight oral cavity was indicative of increasing present with non-­resectable disease.1 Even 20 recent studies analysing microbes present the risk of pancreatic cancer. Notably, in patients who undergo resection, the in various compartments (oral cavity, gut, high levels of plasma antibodies reactive recurrence rate is very high due to early against P. gingivalis corresponded with 2 cysts and pancreatic tumours) in patients systemic dissemination. Combinatorial reduced risk of pancreatic cancer, poten- with pancreatic cancer, as well as murine chemotherapy remains the standard of tially due to systemic immunity against care for patients with advanced disease, studies. We also discuss various strategies cancer-­associated oral pathogens.24 but responses to it are heterogeneous, and for microbial modulation. A summary of Another study demonstrated that the oral 3–5 its toxicity limits treatment duration. the published data on enriched microbes microbial composition differs in healthy Results from clinical trials of single-­agent at different sites in clinical studies and controls and patients with established immunotherapy have not proven its effi- murine models of pancreatic cancer is pancreatic cancer, and the investigators cacy in patients with pancreatic cancer, included in table 1. proposed that detection of two microbes not even when used with chemotherapy or other immunotherapeutic agents.6–8 A highly immunosuppressive tumour micro- environment (TME) has been postulated Table 1 The enriched microbes in the oral cavity, gut and tumours in murine models and as one of the main reasons for the lack of human samples efficacy of immunotherapy for pancreatic cancer.9–11 Therefore, novel strategies Host Study group Enriched bacteria Reference Oral microbiota

that modulate the suppressive TME are http://gut.bmj.com/ urgently needed. Human Prediagnostic samples, PDAC Porphyromonas gingivalis, Fan et al20 The GI tract is the largest reservoir of versus matched controls Aggregatibacter actinomycetemcomitans 21 microbes, which play important roles in Human PDAC versus healthy controls Neisseria elongate, Streptococcus mitis Farrell et al modulating metabolism and immunity Gut microbiota 18 through interaction with host cells.12 Human PDAC versus healthy controls Prevotella, Veillonella, Klebsiella, Ren et al Selenomonas, Hallella, Enterobacter, The discovery of the association between Cronobacter

infection with the bacterium Helicobacter Human PDAC versus healthy controls Bacteroidetes, Veillonellaceae, Half et al22 on September 28, 2021 by guest. Protected copyright. pylori and the incidence of gastritis and Akkermansia, Odoribacter peptic ulcer disease earned Barry Marshall Human PDAC versus healthy controls Proteobacteria, Synergistetes, Pushalkar et al23 and Robin Warren the Nobel Prize in Euryarchaeota Physiology or Medicine in 2005, under- Mouse PDAC spontaneous model Actinobacteria, Deferribacteres, scoring the pivotal role of the microbiota Bifidobacterium pseudolongum Mouse PDAC subcutaneous model Bacteroidetes, Firmicutes Sethi et al48 in influencing inflammatory conditions 26 which may predispose to cancer.13 The Mouse PDAC spontaneous model Bacteroides, Alphaproteobacteria Mendez et al microbial -diversity is a metric that Tumour microbiota α 37 quantifies the number of different species Human PDAC versus healthy controls Gammaproteobacteria Geller et al 23 within a defined sample.14 15 In patients Human PDAC tumour versus gut Proteobacteria (Pseudomonas, Pushalkar et al Elizabethkingia) with many malignancies, including Human PDAC Acinetobacter, Afipia, Corynebacterium, Thomas et al27 colorectal, breast and pancreatic cancer, Escherichia, Propionibacterium Human PDAC Gammaproteobacteria, Bacilli, Riquelme et al40 Actinobacteria 1Department of Clinical Cancer Prevention, Houston, Texas, USA Human PDAC long-­term survivors Pseudoxanthomonas, Streptomyces, 2Department of Gastrointestinal Medical Oncology, versus short-­term survivors Saccharopolyspora, Bacillus clausii University of Texas MD Anderson Cancer Center, Cyst microbiota Houston, Texas, USA Human IPMN, MCN, SCA, pseudocysts Bacteroides, Escherichia/Shigella, Li et al35 3 Department of Immunology, University of Texas MD Fusobacterium, Acidaminococcus, Anderson Cancer Center, Houston, Texas, USA Sphingomonas, Bifidobacterium 36 Correspondence to Dr Florencia McAllister, Human Cancerous versus non-­ Fusobacterium nucleatum, Granulicatella Gaiser et al Department of Clinical Cancer Prevention, University of cancerous PCNs adiacens Texas MD Anderson Cancer Center, Houston, TX 77030, IPMN, intraductal papillary mucinous neoplasm; MCN, mucinous cystic neoplasm; PCN, pancreatic cystic neoplasias; PDAC, USA; fmcallister@​ ​mdanderson.org​ pancreatic ductal adenocarcinoma; SCA, serous cystadenoma.

Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 1 Leading article in the oral compartment (Neisseria elon- of Toll-like­ receptors (TLRs) on immuno- cases.29 The role of bacteria-derived­ poly- gata and Streptococcus mitis) can differen- suppressive monocytic cells. Antibiotics-­ amines in pancreatic cancer development tiate patients with pancreatic cancer from based microbial ablation increased CD4+ has yet to be established. Also, because Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from healthy individuals.21 T-cell­ polarisation towards a Th1 pheno- of their wide repertoire of microbial The gut microbiota associated with type and increased cytotoxic CD8+ T enzymes, microbes have great potential to pancreatic cancer development has been cells based on upregulation of T-bet,­ influence host metabolism downstream of assessed in a few clinical studies.18 22 A tumour necrosis factor-α (TNF-α), inter- dietary influences, including metabolism study in China examined a population of feron gamma (IFN-γ), and CD38. Both of xenobiotics by altering their toxicity.30 patients with PDAC and matched healthy CD4+ and CD8+ T cells had overexpres- In summary, the oral and gut micro- controls.18 Besides a lower α-diversity in sion of programmed cell death protein 1 biomes may play an important role in the gut microbiome of patients with PDAC (PD-1), and CD44 and CD4+ T cells had pancreatic cyst biology and cancer initia- versus healthy controls, a unique gut micro- overexpression of inducible costimulator tion and progression through modulation bial profile was detected in patients with (ICOS) and LFA-1 on microbial ablation. of immune and metabolic pathways. Func- PDAC, including increased abundance of They also found that faecal microbial tional mechanisms implicated in these Veillonella, Klebsiella and Selenomonas transplants (FMT) from the spontaneous processes must be further delineated. species and lipopolysaccharide-producing­ PDAC mouse model KPC (LSL-­KrasG12D, Nevertheless, sequential analysis of oral bacteria (Prevotella, Hallella and Entero- LSL-­Trp53R172H and Pdx1Cre) into either and/or gut bacteria, as well as their serum-­ bacter species) but decreased abundance germ-free­ mice or antibiotic-treated­ KC associated metabolites, could emerge as of Bifidobacterium species and some mice significantly accelerated tumour an inexpensive, non-­invasive strategy butyrate-­producing bacteria (eg, Copro- growth in both settings. Bacterial ablation for early detection of pancreatic cancer. coccus, Clostridium, Blautia, Flavon- on recipient mice prior to FMT improves Furthermore, targeting microbial popu- ifractor and Anaerostipes species). Of colonisation with donor’s stools. This lations associated with increased risk of note, patients who presented with biliary study further showed that infiltration of PDAC may also represent a novel cancer-­ obstruction had a unique microbiome, immunosuppressive myeloid cells and preventive methodology. suggesting that biliary fluid stasis has a role macrophages into the TME decreased in the microbial changes found in these with antibiotic-mediated­ bacterial abla- patients.18 Another study that compared tion, which was reversed with murine RELEVANCE OF THE PANCREATIC gut microbes in patients with pancreatic FMT from mice with PDAC in an orthot- INTRATUMORAL MICROBIOME cancer and healthy controls found an opic PDAC murine model. The presence of micro-organisms­ in increase in Bacteroidetes and a reduction A second mechanism postulated for tumours has been described multiple times in Firmicutes abundance in patients with the microbial effect in tumourigenesis over the past 130 years.31–34 Early pancre- pancreatic cancer in two independent implicates metabolites, which are well-­ atic cystic lesions have been described to cohorts.22 Analysis of the gut microbiotas known host modulators.26 Mendez et al26 possess bacterial DNA.35 36 In examina- tion of different types of pancreatic cystic of patients with PDAC in a third study profiled the gut microbial communities http://gut.bmj.com/ revealed enrichment in Proteobacteria, in mice throughout tumourigenesis and lesions, such as intraductal papillary muci- Synergistetes and Euryarchaeota species identified metabolic pathways associated nous neoplasms (IPMNs), mucinous cystic than in matched healthy controls.23 with microbial changes via metagenomics. neoplasms (MCN), serous cystadenomas The functional role of the gut micro- They used the genetically engineered (SCA) and pseudocysts, most samples biome in pancreatic cancer develop- spontaneous PDAC mouse model KPC to had enrichment in Bacteroides spp, Esch- ment has been examined in murine examine microbial and metabolic changes erichia/Shigella species, Fusobacterium 23 26 27 26 studies. Transgenic mice engineered over time. They found that the main acidaminococcus, Sphingomonas species on September 28, 2021 by guest. Protected copyright. to develop the full spectrum of pancreatic metabolic pathways enriched in KPC mice and Bifidobacterium spp.35 Another study premalignant to malignant lesions exhib- were related to the biosynthesis of pyrim- demonstrated that the microbiotas of high-­ ited marked temporal changes in their idines and polyamines, in particular, those grade IPMNs were commonly found as gut microbiota composition as well as involving putrescine, spermidine and sper- part of oral microbiome, like F. nucleatum bacteria-related­ metabolites throughout mine. Polyamine levels were measured in and Granulicatella adiacens; this poten- tumourigenesis.23 26 To examine progres- the serum of KPC mice and were found tial translocation was attributed in part to sive changes in the gut microbiota during to be increased at 4 months, the time at invasive endoscopic procedures.36 pancreatic tumourigenesis, Pushalkar et which most of the KPC mice had advanced In 2017, Geller et al37 reported for al23 longitudinally sampled faeces from pancreatic intraepithelial neoplasia the first time the presence of bacteria the premalignant PDAC mouse model (PanIN).26 To validate this finding, poly- in human PDACs while demonstrating KC (Ptf1aCre, LSL-Kras­ G12D). Actinobac- amine levels were analysed in serum that intratumoral Gammaproteobac- teria and Bifidobacterium pseudolongum samples and were found to be higher in the teria, among the most common bacteria were abundant in KC mice with advanced serum of patients with PDAC than in the detected in human pancreatic tumours, stages of disease progression.23 More- serum of healthy controls.26 Polyamines, reduce the efficacy of chemotherapeutic over, a clear delay in tumourigenesis was which are mainly produced by intestinal drugs like gemcitabine. The mechanism seen in KC mice located in a germ-free­ microbiota, are known to induce cellular for this involves the bacterial production environment but not in their littermates proliferation by contributing to purine/ of cytidine deaminase, which is capable raised in a regular mouse housing facility, pyrimidine cellular biosynthesis.28 One of of metabolising gemcitabine into its inac- highlighting the functional importance of the microbes associated with polyamine tive form.37 Other comprehensive studies the gut microbiota in pancreatic cancer metabolism is Lactobacillus. Authors have have revealed the presence of microbiotas development.23 The mechanisms impli- reported that Lactobacillus rhamnosus is in several other tumour types, such as cated in tumourigenic induction by the capable of affecting polyamine metabo- melanoma and breast, lung, ovarian, bone dysbiotic microbiota included activation lism and tumour growth in gastric cancer and brain tumours.38 Tumour microbial

2 Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 Leading article Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from

Figure 1 Overview of clinical FMT trials by status (A), clinical phase (B), delivery mode (C) and cancer (D) reported in ClinicalTrials.gov as of September 2020. autoimm, autoimmune; FMT, faecal microbial transplant; inflamm, inflammatory. signatures can also be extracted from study, tumour immunological profiling amphotericin B reduced tumour growth in publicly available data sets in The Cancer revealed enhanced immunoactivation of orthotopic and autochthonous pancreatic Genome Atlas (TCGA) by filtering out the TME in long-term­ survivors, which cancer models. Repopulation of Malassezia non-­human sequences from tumour and positively correlated with the survivor-­ but not other genera, such as Candida, blood samples and can be used, through enriched microbes, demonstrating a Saccharomyces and Aspergillus, promoted artificial intelligence driven predictive potential role for tumour microbes in tumour growth in orthotopic pancreatic models, to distinguish between cancer altering immune-cell­ function. Two of the tumour-­bearing mice given pretreatment and healthy signatures as well as different most enriched bacterial species identified with antifungals.46 Signalling through the cancer types.39 in the tumours of long-­term survivors mannose-­binding lectin (MBL) pathway Intratumoral bacteria have also been have documented immunomodulatory enabled tumour-associated­ fungal popula- 23 27

identified in murine PDAC models. functions. First, Saccharopolyspora spp tions to activate the complement cascade http://gut.bmj.com/ Of note, though, is that Thomas et al27 induce hypersensitivity pneumonitis41 42 via the C3 molecule.46 reported that genetically engineered mice and promote a proinflammatory response Further work is needed to determine (KrasG12D/+, PTENlox/+ and Pdx1-­Cre) in lung epithelial cells via activation of the route of colonisation of intestinal had tumourigenesis delay when raised protein kinase D1 through the innate bacteria in pancreatic tumours, whether in a germ-­free environment. However, immune signal transduction adaptor through reflux from the duodenum or intratumoral microbiotas did not suffer MyD88.42 Second, B. clausii can mediate via circulation through the bloodstream 47 major changes after they were transferred immune-­cell production of nitrous oxide or lymphatic system. Also, it would be on September 28, 2021 by guest. Protected copyright. from germ-­free to specific pathogen-­ and IFN-γ, along with increased CD4+ important to gain a better understanding free housing conditions, suggesting that T-cell­ proliferation in vitro.43 B. clausii of the local conditions that favour a niche the local tumour microbial communities spores are very popular probiotics in for bacteria colonisation in tumours. in this model may not be as relevant as Europe, branded as Enterogermina.43 In systemic/gut communities. vitro studies have shown that B. clausii More recently, Riquelme et al40 profiled can protect enterocytes from rotavirus ANTIMICROBIALS AS BACKBONE intratumoral bacteria from patients with infection by improving epithelial barrier THERAPY FOR PANCREATIC CANCER resected PDAC and compared short-­term function and reducing the production Disruption of a dysbiotic microbiota has and long-­term survivors in two geograph- of reactive oxygen species and cytokines been associated with beneficial effects and ically distant cohorts. Long-­term survi- like interleukin-8 and interferon-β.44 B. decreased tumour growth in numerous vors had greater intratumoral microbial clausii can also decreased the toxicity of pancreatic cancer murine models.23 27 48 48 α-diversity than did those who died of the pathogens like Clostridium difficile and Sethi et al demonstrated that oral treat- disease within 5 years after resection.40 Bacillis cereus through secretion of a ment with an antibiotic cocktail composed Overall tumour microbial characterisation serine protease.45 No clinical or preclin- of vancomycin, neomycin, metronidazole, revealed a microbial composition similar ical studies have assessed a potential role ampicillin and amphotericin B delays the to the one in human PDAC previously for B. clausii in the context of cancer. growth of subcutaneous murine PDAC described by Geller et al,37 but unique Another notable microbial population implants by increasing the number of enrichment in the following microbes in pancreatic tumours is that of the fungal IFN-γ-producing cytotoxic T cells and was found in tumours from long-­term mycobiome, which was recently reported inhibiting the number of interleukin (IL)-­ survivors in both independent cohorts to be important for pathogenesis.46 Aykut 17A and IL-10 producing protumourigenic studied by Riquelme and colleagues: Pseu- et al46 showed that the fungal genus T cells. Also, Pushalkar et al23 reported doxanthomonas, Streptomyces and Sacch- Malassezia is abundant in murine and that depletion of the gut microbiota aropolyspora and Bacillus clausii. In this human PDACs and that its depletion with with oral antimicrobials led to decreased

Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 3 Leading article Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from

Figure 2 Diagram showing different approaches to microbiota modulation for cancer treatment. The overall goal is to shift the tumour microenvironment from an immunosuppressive to an immune-activ­ ated state. FMT, faecal microbial transplant; MDSC, myeloid-derived­ suppressor cell; TH1, type I helper; Treg, regulatory T cell. tumour size and changes in the immune drugs.51 Increased levels of apoptosis as as treatment with an antibiotic regimen landscape of the TME in orthotopic PDAC measured using cleaved caspase 3 staining consisting of vancomycin, imipenem and mouse models by eliciting both innate and were observed in tumours only when the neomycin resulted in decreased efficacy of adaptive responses to reduce myeloid-­ implanted microdevice released a combi- CpG oligonucleotide immunotherapy and derived suppressor cells (MDSCs) and nation of gemcitabine and antibiotics, platinum-­based chemotherapy. Similarly, increase M1-­type macrophages tumour thus highlighting the specific role of in a subcutaneous sarcoma mouse model infiltration with suppression of TLR 2/5 intratumoral bacteria in opposing chemo- (MCA-205), Vétizou et al59 observed 37

signalling while enhancing the number of therapy activity. Analogous to this, decreased potency of inhibition of the http://gut.bmj.com/ Th1-­type CD4+ and cytotoxic CD8+ T in a colon carcinoma xenograft mouse activity of the immune checkpoint mole- cells. In addition, Thomas et al27 found model, abrogation of the protumourigenic cule cytotoxic T-­lymphocyte-­associated that murine gut microbiota ablation with Fusobacterium load via treatment with protein 4 (CTLA-4) on ablation of the antibiotics dampened tumour growth in metronidazole slowed tumour growth as commensal microbiota with the use of genetic PDAC mouse models (KrasG12D/+, well.52 Localised delivery of antibiotics antibiotics such as ampicillin, colistin and PTENlox/+ and Pdx1-Cre).­ These three using local release devices would be less streptomycin. Using the same antibiotic 60 independent studies using different PDAC toxic and more effective for systemic use, cocktail, Routy et al observed reduced on September 28, 2021 by guest. Protected copyright. models showed that broad microbial abla- potentially representing a promising local efficacy of blocking another immune tion is an effective approach to affect therapeutic strategy to be combined with checkpoint target, PD-1 blockade, in progression of premalignant pancreatic systemic antitumoral therapies.37 53 reducing tumour growth in subcutaneous lesions to cancer and PDAC growth.23 27 48 Despite the promising data on antimi- mouse models of sarcoma (MCA-205) and We envision a clinical trial in which chemo- crobial use for PDAC described previously, melanoma (RET) when combined with therapy is combined with systemic antibi- antibiotics have been associated with antibiotics. otics for treatment of PDAC. However, worse outcomes for other cancer types. From the information in the previous two known toxic effects associated with long-­ Antibiotic use prior to immune check- paragraphs, antibiotics may have different term use of broad-­spectrum antibiotics as point inhibitors treatment has been asso- impacts depending on the tumour type, the well as rise of multidrug resistant bacteria ciated with worse survival in melanoma, concomitant therapies and the presence or would certainly limit the enthusiasm for lung and other cancers.54–56 Meta-­analyses absence of specific bacteria in the cancer such studies.49 50 of published clinical data have also shown cells as in the example of patients with Elimination of bacteria with targeted that antibiotic usage in patients with colorectal cancer, in whom antibiotics only local delivery of antibiotics such as ampi- cancer (melanoma, lung, renal and head play an antitumoral effect when tumour cillin and chloramphenicol resulted in and neck carcinomas) receiving immune cells contain Fusobacterium.52 Therefore, reduced tumour growth in a subcutaneous checkpoint inhibitors leads to poorer identifying the microbial profiles present colon carcinoma mouse model (MC26) survival.57 Several preclinical studies across tumours in different patient popu- as shown by Geller et al.37 Local intratu- have also reported on the negative effects lations is important to elucidate their moral release of antibiotics was achieved of concomitant use of antibiotics and differential effects. If antibiotic treatment through the use of an implantable immunotherapy. Iida et al58 reported a in patients with cancers like sarcomas and microdevice with a standard biopsy needle detrimental effect of depletion of the melanoma results in increased tumour that can release microdoses of single commensal microbiota in a subcutaneous growth, microbial populations in these agents or combinations of therapeutic colorectal cancer mouse model (MC38), settings may be beneficial for the host and

4 Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 Leading article may be needed for therapeutic responses microbiota and outcomes has been to establish the safety and efficacy of of tumours.58–60 In patients with pancre- explored by several groups using preclin- FMT in modulating the human pancreatic atic cancer, however, because most studies ical models.23 40 46 48 tumour microbiome and enhancing the Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from have reported antitumoral effect on anti- To approximate the human scenario, immune response to ultimately improve microbial treatment, microbial enrich- Riquelme et al40 performed human faecal survival. ment associated with this cancer type may microbial transplantation (hFMT) in mice contribute to its pathogenesis, explaining that later received orthotopic tumour SINGLE OR CONSORTIA why its elimination may play a beneficial cell implantation to generate a PDAC BACTERIOTHERAPY role.23 27 48 Future studies must evaluate humanised microbial mouse model. This Bacteriotherapy, consisting of oral admin- factors shaping the overall effect of the study demonstrated that about a quarter istration of either a single or a consortium tumour microbiota, including variables of the human PDAC microbial composi- of bacterial species, is a potential strategy like diversity, location, the microenviron- tion but not the adjacent normal tissue for more targeted manipulation of the ment and host factors. overlapped with the human donors' gut microbiome than FMT. Early studies in Additionally, certain microbes can microbiota, suggesting the existence of a animals looked at the toxicity of spores generate cross-­reactive T cells owing to microbial cross-­talk between the gut and derived from a non-­pathogenic strain molecular mimicry between tumour and pancreatic tumours.40 The same study also of C. novyi, a bacterium that can germi- microbial antigens which may trigger demonstrated that the human donors’ nate in necrotic and hypoxic regions of antitumor immunity through antigen gut bacteria were efficiently transferred tumours.72 73 Other studies evaluated presentation by major histocompatibility to the gut in these mice.40 Moreover, the the roles of specific gut bacterial species complex class I (MHC-I).­ 61 However, tumour microbial composition in human- in augmenting the efficacy of immuno- a recent study showed that pancreatic ised microbial mice was also differentially therapy for tumours.59 74 Bacteroides cancer cells have downregulation of modulated by transplants from different spp, particularly Bacteroides fragilis and MHC-I­ molecules on their surfaces for donors.40 In accordance with previous B. thetaiotaomicron, identified in the evasion of immunosurveillance through murine studies, hFMT using samples gut microbiota in a fibrosarcoma mouse upregulation of autophagy, and inhibition obtained from healthy controls resulted in model, are capable of enhancing CTLA-4-­ of MHC I improves therapeutic outcomes slower tumour growth and a more modest based immunotherapy.59 Bifidobacterium by synergising with immune checkpoint reversal of TME immunosuppression spp, identified in the gut microbiota of a blockade.62 This highlights the importance than did hFMT with samples from short-­ melanoma murine model, induced anti- of ascertaining the balance between elimi- term survivors, whereas the most potent tumourigenic immune responses alone nating harmful microbiotas and enriching reversal of TME immunosuppression and and, more potently, in combination with beneficial microbiotas, which can improve tumour growth was induced by hFMT another immune checkpoint molecule, antitumor immunity. using samples from long-­term survivors of programmed cell death-ligand­ 1 (PD-­ Another emerging strategy for depleting PDAC with no evidence of disease.40 Of L1).74 Engineered non-­pathogenic bacteria

pathogenic intratumoral bacteria is the note, this effect was lost when antibiotics http://gut.bmj.com/ also have been studied for targeted use of a special class of highly selective were given after the transplant or when delivery of therapeutic agents like anti- viruses known as lytic bacteriophages or cytotoxic T cells were depleted.40 This bodies and chemopreventive metabolites phages, which can selectively infect and lyse demonstrated the ability of hFMT to posi- to determine their ability to be enriched certain bacteria and have been tested for tively affect PDAC tumours by modulating in tumours.75–77 Nonetheless, no bacterial their capacity to shift the gut microbiota in the gut/tumour microbial axis along with species have emerged as being therapeutic murine models.53 63 Phage display libraries the immune system. All of this evidence for pancreatic cancer, and bacteriotherapy have been shown to target specific organs in demonstrates that FMT can modulate the on September 28, 2021 by guest. Protected copyright. for this cancer remains largely unexplored. vivo, and they can produce bystander cyto- gut and tumour microbiotas, the immune However, being mindful of the scenarios in toxicity in tumour cells.64–66 In preliminary activation status and outcomes, paving the which addition of a single bacterium may studies, researchers explored the utility of way for FMT to be used in combination result in decreased gut diversity, which phages for delivery of anticancer drugs to with other treatment modalities, such is usually associated with poor cancer pancreatic cancers and demonstrated early as chemotherapy and immunotherapy, outcomes, is important. Thus, supplemen- viability of successful targeting of pancreatic for pancreatic cancer and perhaps other tation of single-­species bacteriotherapy tumour cells.67–69 Whereas substantial data tumour types. with FMT may be needed to provide the demonstrate the effect of phages on regu- FMT is effectively used in clinical advantages of both methodologies. lation of microbial population dynamics, practice for treatment of conditions like evidence of their direct effects on human recurrent C. difficile infection, and several health is limited. Technical limitations ongoing clinical trials at different phases CONCLUSIONS in therapeutically using phages, such as of development are using FMT for treat- Studies that aim to target the PDAC-­ dosing, route of administration, tolerance ment of autoimmune, inflammatory and associated gut and tumour microbiotas by the host immune system, and specificity metabolic conditions (figure 1A–C and may employ three strategies: (1) elimi- to pancreatic cancer microbes and tumour online supplemental table S1).70 71 The use nation of protumourigenic bacteria from antigens, remain to be addressed before this of FMT as a treatment option for cancer the host through the use of antibiotics, strategy can be widely adopted. is gradually attracting more attention. (2) enhancement of immunoactivation Several early-­phase clinical trials of hFMT by delivery of single or multiple micro- FECAL MICROBIAL TRANSPLANTS AS are currently testing its role in cancer treat- bial species and (3) ‘normalisation’ of the BACKBONE THERAPY FOR PANCREATIC ment responses and to reduce treatment-­ dysbiotic gut and tumour environment CANCER associated toxic effects (figure 1D and in patients with PDAC using whole FMT The potential of the whole gut micro- online supplemental table S1). Ongoing (figure 2). Ultimately, all of these strate- biota to modulate the pancreatic tumour and future clinical studies will be required gies have the common goal of shifting the

Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 5 Leading article balance from an immunosuppressive TME 17 Ahn J, Sinha R, Pei Z, et al. Human gut microbiome to an immunoactivated one. Further- and risk for colorectal cancer. J Natl Cancer Inst

2013;105:1907–11. Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from more, identification and better under- To cite Chandra V, McAllister F. Gut Epub ahead of 18 Ren Z, Jiang J, Xie H, et al. Gut microbial profile standing of the mechanisms employed print: [please include Day Month Year]. doi:10.1136/ analysis by MiSeq sequencing of pancreatic carcinoma by downstream effector molecules (eg, gutjnl-2019-319807 patients in China. Oncotarget 2017;8:95176–91. host or microbial metabolites) mediating Received 13 April 2020 19 Gopalakrishnan V, Spencer CN, Nezi L, et al. Gut microbiome modulates response to anti-­PD-1 microbial responses as biomarkers and Revised 16 March 2021 Accepted 12 April 2021 immunotherapy in melanoma patients. Science potentially as therapeutic targets may also 2018;359:97–103. be very important. In summary, modula- Gut 2021;0:1–7. 20 Fan X, Alekseyenko AV, Wu J, et al. Human oral tion of the microbiome may emerge as a doi:10.1136/gutjnl-2019-319807 microbiome and prospective risk for pancreatic cancer: a population-­based nested case-control­ study. Gut supplement for existing cancer therapies ORCID iDs 2018;67:120–7. with the main goal of increasing their Vidhi Chandra http://orcid.​ ​org/0000-​ ​0003-3642-​ ​2144 21 Farrell JJ, Zhang L, Zhou H, et al. Variations of oral efficacy by reversing immunosuppression. Florencia McAllister http://orcid.​ ​org/0000-​ ​0002-9915-​ ​ microbiota are associated with pancreatic diseases Because most functional studies reported 0943 including pancreatic cancer. Gut 2012;61:582–8. 22 Half E, Keren N, Reshef L, et al. Fecal microbiome thus far have been conducted using animal signatures of pancreatic cancer patients. Sci Rep models, we hope to get more information REFERENCES 2019;9:16801. from upcoming clinical trials targeting the 1 Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. 23 Pushalkar S, Hundeyin M, Daley D, et al. The pancreatic CA Cancer J Clin 2020;70:7–30. microbiome. cancer microbiome promotes oncogenesis by induction 2 Rhim AD, Mirek ET, Aiello NM, et al. Emt and of innate and adaptive immune suppression. Cancer dissemination precede pancreatic tumor formation. Discov 2018;8:403–16. Contributors FM designed, wrote and reviewed the Cell 2012;148:349–61. 24 Michaud DS, Izard J, Wilhelm-Benartzi­ CS, et al. Plasma article. VC wrote and reviewed the article. 3 Conroy T, Desseigne F, Ychou M, et al. Folfirinox versus antibodies to oral bacteria and risk of pancreatic Funding VC is supported by NCI T32 in Cancer gemcitabine for metastatic pancreatic cancer. N Engl J cancer in a large European prospective cohort study. Biology, MD Anderson Cancer Centre. FM is supported Med 2011;364:1817–25. Gut 2013;62:1764–70. by NCI 1R37CA237384-­01A1, CPRIT RP200173, V 4 Von Hoff DD, Ervin T, Arena FP, et al. Increased 25 Michaud DS, Joshipura K, Giovannucci E, et al. Foundation Translational Award, Andrew Sabin Family survival in pancreatic cancer with nab-­paclitaxel plus A prospective study of periodontal disease and Fellowship and Philantropic Moonshot Programmes gemcitabine. N Engl J Med 2013;369:1691–703. pancreatic cancer in US male health professionals. J (MDACC). We also acknowledge the MD Anderson 5 Neoptolemos JP, Stocken DD, Bassi C, et al. Natl Cancer Inst 2007;99:171–5. Research Medical Library. Adjuvant chemotherapy with fluorouracil plus folinic 26 Mendez R, Kesh K, Arora N, et al. Microbial dysbiosis acid vs gemcitabine following pancreatic cancer and polyamine metabolism as predictive markers for Competing interests None declared. resection: a randomized controlled trial. JAMA early detection of pancreatic cancer. Carcinogenesis Patient consent for publication Not required. 2010;304:1073–81. 2020;41:561–70. doi:10.1093/carcin/bgz116 6 O’Reilly EM. Durvalumab with or without 27 Thomas RM, Gharaibeh RZ, Gauthier J, et al. Intestinal Provenance and peer review Commissioned; tremelimumab for patients with metastatic pancreatic microbiota enhances pancreatic carcinogenesis in externally peer reviewed. ductal adenocarcinoma: a phase 2 randomized clinical preclinical models. Carcinogenesis 2018;39:1068–78. Supplemental material This content has been trial. JAMA Oncol 2019;5:1431–8. doi:10.1001/ 28 Tofalo R, Cocchi S, Suzzi G. Polyamines and gut supplied by the author(s). It has not been vetted by jamaoncol.2019.1588 microbiota. Front Nutr 2019;6:16. http://gut.bmj.com/ BMJ Publishing Group Limited (BMJ) and may not have 7 Kamath SD, Kalyan A, Kircher S, et al. Ipilimumab and 29 Russo F, Orlando A, Linsalata M, et al. Effects of been peer-­reviewed. Any opinions or recommendations gemcitabine for advanced pancreatic cancer: a phase Lactobacillus rhamnosus GG on the cell growth and discussed are solely those of the author(s) and are Ib study. Oncologist 2020;25:e808–15. polyamine metabolism in HGC-27 human gastric not endorsed by BMJ. BMJ disclaims all liability and 8 Royal RE, Levy C, Turner K, et al. Phase 2 trial of single cancer cells. Nutr Cancer 2007;59:106–14. responsibility arising from any reliance placed on the agent ipilimumab (anti-­CTLA-4) for locally advanced or 30 Koppel N, Maini Rekdal V, Balskus EP. Chemical transformation of xenobiotics by the human gut content. Where the content includes any translated metastatic pancreatic adenocarcinoma. J Immunother microbiota. Science 2017;356. doi:10.1126/science. material, BMJ does not warrant the accuracy and 2010;33:828–33. aag2770. [Epub ahead of print: 23 Jun 2017]. reliability of the translations (including but not limited 9 Feig C, Gopinathan A, Neesse A, et al. The pancreas 31 Russell W. An address on a characteristic organism of on September 28, 2021 by guest. Protected copyright. cancer microenvironment. Clin Cancer Res to local regulations, clinical guidelines, terminology, cancer. Br Med J 1890;2:1356–60. 2012;18:4266–76. drug names and drug dosages), and is not responsible 32 Rous P. A sarcoma of the fowl transmissible by an 10 von Bernstorff W, Voss M, Freichel S, et al. Systemic for any error and/or omissions arising from translation agent separable from the tumor cells. J Exp Med and local immunosuppression in pancreatic cancer and adaptation or otherwise. 1911;13:397–411. patients. Clin Cancer Res 2001;7:925s–32. 33 Wuerthelecaspe V, Alexanderjackson E, Anderson 11 Liyanage UK, Moore TT, Joo H-­G, et al. Prevalence JA, et al. Cultural properties and pathogenicity of regulatory T cells is increased in peripheral blood of certain microorganisms obtained from various and tumor microenvironment of patients with proliferative and neoplastic diseases. Am J Med Sci pancreas or breast adenocarcinoma. J Immunol 1950;220:638–48. 2002;169:2756–61. 34 Aries V, Crowther JS, Drasar BS, et al. Bacteria and 12 Durack J, Lynch SV. The gut microbiome: relationships the aetiology of cancer of the large bowel. Gut Open access This is an open access article with disease and opportunities for therapy. J Exp Med 1969;10:334–5. distributed in accordance with the Creative Commons 2019;216:20–40. 35 Li S, Fuhler GM, Bn N, et al. Pancreatic cyst Attribution Non Commercial (CC BY-NC­ 4.0) license, 13 Mégraud F. A humble bacterium sweeps this year’s fluid harbors a unique microbiome. Microbiome which permits others to distribute, remix, adapt, build Nobel Prize. Cell 2005;123:975–6. 2017;5:147. upon this work non-commercially,­ and license their 14 Human Microbiome Project Consortium. Structure, 36 Gaiser RA, Halimi A, Alkharaan H, et al. Enrichment of derivative works on different terms, provided the function and diversity of the healthy human oral microbiota in early cystic precursors to invasive original work is properly cited, appropriate credit is microbiome. Nature 2012;486:207–14. doi:10.1038/ pancreatic cancer. Gut 2019;68:2186–94. given, any changes made indicated, and the use is nature11234 37 Geller LT, Barzily-Rokni­ M, Danino T, et al. Potential role non-­commercial. See: http://​creativecommons.​org/​ 15 Arthur JC, Perez-­Chanona E, Mühlbauer M, et al. of intratumor bacteria in mediating tumor resistance licenses/​by-​nc/​4.​0/. Intestinal inflammation targets cancer-­inducing activity to the chemotherapeutic drug gemcitabine. Science of the microbiota. Science 2012;338:120–3. © Author(s) (or their employer(s)) 2021. Re-­use 2017;357:1156–60. 16 Goedert JJ, Jones G, Hua X, et al. Investigation of the permitted under CC BY-­NC. No commercial re-use­ . See 38 Nejman D, Livyatan I, Fuks G, et al. The human tumor association between the fecal microbiota and breast rights and permissions. Published by BMJ. microbiome is composed of tumor type-­specific cancer in postmenopausal women: a population-­based intracellular bacteria. Science 2020;368:973–80. ►► Additional supplemental material is published case-­control pilot study. J Natl Cancer Inst 2015;107. 39 Poore GD, Kopylova E, Zhu Q, et al. Microbiome online only. To view, please visit the journal online doi:10.1093/jnci/djv147. [Epub ahead of print: 01 Jun analyses of blood and tissues suggest cancer (http://dx.​ ​doi.org/​ ​10.1136/​ ​gutjnl-2019-​ ​319807). 2015]. diagnostic approach. Nature 2020;579:567–74.

6 Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 Leading article

40 Riquelme E, Zhang Y, Zhang L, et al. Tumor microbiome 53 Duerkop BA, Clements CV, Rollins D, et al. A composite 65 Trepel M, Stoneham CA, Eleftherohorinou H, et al. A diversity and composition influence pancreatic cancer bacteriophage alters colonization by an intestinal heterotypic bystander effect for tumor cell killing after

outcomes. Cell 2019;178:795–806. commensal bacterium. Proc Natl Acad Sci U S A adeno-associated­ virus/phage-mediated,­ vascular-­ Gut: first published as 10.1136/gutjnl-2019-319807 on 27 April 2021. Downloaded from 41 Marx JJ, Emanuel DA, Dovenbarger WV, et al. Farmer’s 2012;109:17621–6. targeted suicide gene transfer. Mol Cancer Ther lung disease among farmers with precipitating 54 Elkrief A, El Raichani L, Richard C, et al. Antibiotics 2009;8:2383–91. antibodies to the thermophilic actinomycetes: a are associated with decreased progression-­free 66 Pasqualini R, Ruoslahti E. Organ targeting in vivo clinical and immunologic study. J Allergy Clin Immunol survival of advanced melanoma patients treated with using phage display peptide libraries. Nature 1978;62:185–9. immune checkpoint inhibitors. Oncoimmunology 1996;380:364–6. 42 Kim Y-­I, Park J-E,­ Brand DD, et al. Protein kinase D1 is 2019;8:e1568812. 67 Asar MC, Franco A, Soendergaard M. Phage display essential for the proinflammatory response induced 55 Huemer F, Rinnerthaler G, Westphal T, et al. selection, identification, and characterization of novel by hypersensitivity pneumonitis-­causing thermophilic Impact of antibiotic treatment on immune-­ pancreatic cancer targeting peptides. Biomolecules actinomycetes Saccharopolyspora rectivirgula. J checkpoint blockade efficacy in advanced non-­ 2020;10. doi:10.3390/biom10050714. [Epub ahead Immunol 2010;184:3145–56. squamous non-­small cell lung cancer. Oncotarget of print: 05 May 2020]. 43 Urdaci MC, Bressollier P, Pinchuk I. Bacillus clausii 2018;9:16512–20. 68 Bedi D, Gillespie JW, Petrenko VA. Selection of probiotic strains: antimicrobial and immunomodulatory 56 Pinato DJ, Howlett S, Ottaviani D, et al. Association of pancreatic cancer cell-­binding landscape phages and activities. J Clin Gastroenterol 2004;38:S86–90. prior antibiotic treatment with survival and response to their use in development of anticancer nanomedicines. 44 Paparo L, Tripodi L, Bruno C, et al. Protective action of immune checkpoint inhibitor therapy in patients with Protein Eng Des Sel 2014;27:235–43. Bacillus clausii probiotic strains in an in vitro model of cancer. JAMA Oncol 2019;5:1774–8. 69 Wang T, Narayanaswamy R, Ren H, et al. Phage-­ rotavirus infection. Sci Rep 2020;10:12636. 57 Huang X-­Z, Gao P, Song Y-­X, et al. Antibiotic use and derived protein-­mediated targeted chemotherapy of 45 Ripert G, Racedo SM, Elie A-­M, et al. Secreted the efficacy of immune checkpoint inhibitors in cancer pancreatic cancer. J Drug Target 2018;26:505–15. compounds of the probiotic Bacillus clausii strain O/C patients: a pooled analysis of 2740 cancer patients. 70 Bakken JS, Borody T, Brandt LJ, et al. Treating inhibit the cytotoxic effects induced by Clostridium Oncoimmunology 2019;8:e1665973. Clostridium difficile infection with fecal microbiota difficile and Bacillus cereus toxins. Antimicrob Agents 58 Iida N, Dzutsev A, Stewart CA, et al. Commensal transplantation. Clin Gastroenterol Hepatol Chemother 2016;60:3445–54. bacteria control cancer response to therapy by 2011;9:1044–9. 46 Aykut B, Pushalkar S, Chen R, et al. The fungal modulating the tumor microenvironment. Science 71 Vrieze A, Van Nood E, Holleman F, et al. Transfer of mycobiome promotes pancreatic oncogenesis via 2013;342:967–70. intestinal microbiota from lean donors increases activation of MBL. Nature 2019;574:264–7. 59 Vétizou M, Pitt JM, Daillère R, et al. Anticancer insulin sensitivity in individuals with metabolic 47 Thomas RM, Jobin C. Microbiota in pancreatic health immunotherapy by CTLA-4 blockade relies on the gut syndrome. Gastroenterology 2012;143:913–6. and disease: the next frontier in microbiome research. microbiota. Science 2015;350:1079–84. 72 Diaz LA, Cheong I, Foss CA, et al. Pharmacologic and Nat Rev Gastroenterol Hepatol 2020;17:53–64. 60 Routy B, Le Chatelier E, Derosa L, et al. Gut toxicologic evaluation of C. novyi-­NT spores. Toxicol Sci 48 Sethi V, Kurtom S, Tarique M, et al. Gut microbiota microbiome influences efficacy of PD-1-based­ 2005;88:562–75. promotes tumor growth in mice by modulating immunotherapy against epithelial tumors. Science 73 Krick EL, Sorenmo KU, Rankin SC, et al. Evaluation of immune response. Gastroenterology 2018;155:33–7. 2018;359:91–7. Clostridium novyi-­NT spores in dogs with naturally 49 Barbosa TM, Levy SB. The impact of antibiotic use on 61 Fluckiger A, Daillère R, Sassi M, et al. Cross-­reactivity occurring tumors. Am J Vet Res 2012;73:112–8. resistance development and persistence. Drug Resist between tumor MHC class I-­restricted antigens 74 Sivan A, Corrales L, Hubert N, et al. Commensal Updat 2000;3:303–11. and an enterococcal bacteriophage. Science Bifidobacterium promotes antitumor immunity 50 Bonomo RA. Multiple antibiotic-­resistant bacteria 2020;369:936–42. and facilitates anti-­PD-­L1 efficacy. Science in long-­term-­care facilities: an emerging problem 62 Yamamoto K, Venida A, Yano J, et al. Autophagy 2015;350:1084–9. in the practice of infectious diseases. Clin Infect Dis promotes immune evasion of pancreatic cancer by 75 Chowdhury S, Castro S, Coker C, et al. Programmable 2000;31:1414–22. degrading MHC-­I. Nature 2020;581:100–5. bacteria induce durable tumor regression 51 Jonas O, Landry HM, Fuller JE, et al. An implantable 63 Hsu BB, Gibson TE, Yeliseyev V, et al. Dynamic and systemic antitumor immunity. Nat Med microdevice to perform high-­throughput in vivo modulation of the gut microbiota and metabolome by 2019;25:1057–63. http://gut.bmj.com/ drug sensitivity testing in tumors. Sci Transl Med bacteriophages in a mouse model. Cell Host Microbe 76 Forbes NS. Engineering the perfect (bacterial) cancer 2015;7:284ra57. 2019;25:803–14. therapy. Nat Rev Cancer 2010;10:785–94. 52 Bullman S, Pedamallu CS, Sicinska E, et al. Analysis of 64 Landon LA, Deutscher SL. Combinatorial discovery of 77 Ho CL, Tan HQ, Chua KJ, et al. Engineered commensal Fusobacterium persistence and antibiotic response in tumor targeting peptides using phage display. J Cell microbes for diet-­mediated colorectal-­cancer colorectal cancer. Science 2017;358:1443–8. Biochem 2003;90:509–17. chemoprevention. Nat Biomed Eng 2018;2:27–37. on September 28, 2021 by guest. Protected copyright.

Chandra V, McAllister F. Gut Month 2021 Vol 0 No 0 7