Integrating the microbiome into the barrier theory of cancer

The barrier theory of cancer integrates environmental, genetic, and infectious contributions to oncogenesis into a single framework. The full spectrum of symbiotic influences on oncogenesis, however, still needs to be brought into this framework, particularly for symbionts that are classified as commensals or mutualists. This paper contributes to this goal by presenting evidence that these symbionts may improve the effectiveness of immunological defenses against cancer and by drawing attention to the interactions between parasitic and mutualistic microbes in relation to oncogenesis. Although integration of protective symbionts with the barrier theory encompasses a variety of interactions, we highlight one particular aspect: effects of symbionts on the host’s ability to attack tumor cells through the relaxation of immune checkpoints.


| Intro duc tion: The Barrier Theor y of C ancer an d the Sy mbiosis Continuum
The barrier theory of oncogenesis integrates cellular protections against oncogenesis with infectious and environmental factors that abrogate these protections (Ewald & Swain Ewald, 2013). It organizes the vast array of information on oncogenesis within a causal framework that distinguishes the small number of barriers to oncogenesis from the many restraints on oncogenesis. Barriers are defined as processes that block the process of oncogenesis and that therefore must be broken or inactive in a particular cell for oncogenesis to occur. The five noted barriers to oncogenesis are cell cycle arrest, apoptosis, regulation of telomerase, cell adhesion, and asymmetric cell division. Abrogations of these barriers are considered essential causes of oncogenesis, which are distinguished from exacerbating causes. The latter contributes to oncogenesis by relaxing restraints on oncogenesis, which are defined as defenses against cancer that suppress but do not block oncogenesis. The barrier theory is diagrammed in Figure 1. From a practical perspective, the barrier theory draws attention to interventions that preserve barriers and therefore can most directly prevent or cure cancer.
Abrogation of barriers could arise by mutation, epigenetic modification, and direct interference from pathogens. The most powerful contributions of pathogens to oncogenesis arise from mechanisms that evolved to promote productive pathogen persistence within hosts because barriers to cancer are often barriers to this persistence (Ewald & Swain Ewald, 2012). Pathogens that evolve increased persistence by compromising these barriers thus contribute directly to the process of oncogenesis. Less direct effects of pathogens on oncogenesis arise from haphazard effects of pathogens, such as stimulation of inflammatory responses that may increase the probability of mutations or enhance cellular proliferation (Garrett, 2015).
The barrier theory is evolutionary because it integrates the evolution of protective mechanisms against cancer as well as the evolution of symbiont characteristics that compromise or enhance these protective adaptations. It draws on several conceptual frameworks for understanding cancer, including the clonal evolution theory (Nowell, 1976), the stem cell theory (Clarke et al., 2006), the Hallmarks of Cancer (Hanahan & Weinberg, 2000, and the concept of tumor virology (Moore & Chang, 2010). The barrier theory, however, is distinct from each of these frameworks because it integrates the known processes of oncogenesis into a causal evolutionary framework based on first principles. It differs, for example from the clonal theory by integrating mechanisms beyond mutation through which oncogenic viruses contribute to oncogenesis. It differs from the stem cell theory by building from principles of somatic evolutionary selection on cells rather than the metaphor of organogenesis and thus proposes how tumor cells may have stem-cell-like characteristics without suggesting that oncogenesis reflects the process of organogenesis.

Barriers to oncogenesis
Mutualistic microbiota may protect against oncogenic factors that abrogate barriers of tumor virology in a broader framework of oncogenesis that can encompass pathogens other than viruses and differentiates oncogenic characteristics of pathogens that are adaptations from side effects of infection.
Although the barrier theory incorporates insights from these other conceptual frameworks, its value derives from its differences from them. Some of the Hallmarks of Cancer, for example, involve processes that block cancer; others do not. Enhancement of angiogenesis and immune evasion are hallmarks of cancer but may not be necessary for oncogenesis to proceed. According to the barrier theory, these effects involve relaxation of restraints and thus play an exacerbating rather than an essential role in oncogenesis. Also, in contrast with these other frameworks, the barrier theory builds on the foundation of evolutionary selection-differential survival and reproduction-to characterize barriers and restraints ( Figure 1).
But the barrier theory applies this foundation to a subcategory of somatic selection, termed oncogenic selection, to identify the barriers that must be inactive or inactivated for oncogenesis to occur ( Figure 1; Ewald & Swain Ewald, 2013).
The goal of this paper is to broaden the scope of the barrier theory to include protection against oncogenesis arising from the microbiome. We consider direct protective effects of microbiota well as the interplay between protective microbiota and parasites that contribute directly or indirectly to oncogenesis. This analysis complements recent papers that have considered the joint contributions of parasitic organisms to oncogenesis (Dheilly, Ewald, Brindley, Fichorova, & Thomas, 2019;Ewald & Swain Ewald, 2014) and protection against oncogenesis that arises from immune responses to pathogens that are not oncogenic (Cramer & Finn, 2011;Iheagwara et al., 2014;Jacqueline & Finn, 2020;Oh et al., 2019).

| Illus tr ative O ver view O f Microbiome Composition A n d Protec tion A gains t C ancer
Mutualistic microbiota can function as a symbiotic complement to  Similarly, a mouse model of colitis-associated colorectal cancer, using a carcinogen plus an inflammatory agent, revealed that the process of tumorigenesis shifted the microbiome (Zackular et al., 2013).
Over the course of tumor development, step-wise alterations from the baseline microbiome resulted in less overall microbial diversity and major changes in microbial composition. When this altered microbiome was established in germ-free mice prior to initiating tumorigenesis, more and larger tumors developed relative to germ-free mice preestablished with a healthy microbiota. Importantly, repeated exposure to the inflammatory agent alone had an initial and sustained impact on the microbiome but did not reduce overall diversity. At a mechanistic level, the balance maintained by these checkpoints is influenced by microbiome components. At an evolutionary level these interactions must surely be molded by natural selection because a rogue immunological response would lower the organism's fitness by destroying valuable cells, but over suppression of these processes would make organisms vulnerable to infection and cancer.

| Ef fec t s of Sy mbiont s on Immune
The current spectrum of cancer treatments includes inhibition of these checkpoints, which is often associated with improved control of the cancer (Routy, Gopalakrishnan, et al., 2018). Current evidence indicates that symbionts help maintain this responsiveness.
One line of evidence has been generated from antibiotic treatment of cancer patients. Administration of broad-spectrum antibiotics was associated with decreased effectiveness of checkpoint inhibition through the PD-1 mechanism in lung, renal, and bladder cancer (Routy, Le Chatelier, et al., 2018). This finding implicates microbes in a process that permits escalated immunological destruction of cancer cells.
A complementary line of evidence involves effects of microbial supplementation on checkpoint inhibition. In particular, experimental transfers of bacteria have been associated with responsiveness to checkpoint inhibition. Transfer of stool samples from patients who responded to immune checkpoint inhibition ("responders") enhanced the ability of mice to respond to this inhibition and better control of cancer (Routy, Gopalakrishnan, et al., 2018). This favorable response did not occur in mice receiving stool samples from nonresponder patients, but could be generated subsequently in these mice by se- The bacterium Akkermansia muciniphila was isolated more frequently from feces of cancer patients who responded strongly to checkpoint inhibition of PD-1 than from patients who showed a weaker checkpoint blockade response (Matson et al., 2018;Routy, Le Chatelier, et al., 2018). A. muciniphila restored the response to checkpoint inhibition in mice that had been treated with broad-spectrum antibiotics or had received fecal transplants from nonresponders.
The mechanism by which A. muciniphila improves response to checkpoint inhibition may involve effects on T-cell function.
Interferon associated memory T cell responses were stronger in responders than in nonresponders when the T cells were cultured in Oncogenic: pertaining to the development of cancer; oncogenic pathogens to refer to pathogens that contribute to the process of oncogenesis.
The effects of A. muciniphila in mice depended on interleukin-12, which favors the development of cell-killing functions of T cells, and were correlated with reductions in regulatory T cells, which enforce cellular checkpoints (Routy, Le Chatelier, et al., 2018). In sterile mice, A. muciniphila led to the presence in mesenteric lesions of a helper T cell subset that is associated with regression of the lesions (Routy, Le Chatelier, et al., 2018).
Associations of cancer with different abundances of gut microbiome species and outcomes of bacterial supplementation suggest the important role of mutualist bacteria in cancer prevention as well as treatment efficacy and tolerability (Routy, Gopalakrishnan, et al., 2018;Wang, Yin, Chen, & Davis, 2018). Using Bifidobacterium species as an example, studies have found an association between lower levels of these bacteria in colorectal cancer patients relative to healthy controls, higher levels in patients that responded to immune checkpoint inhibition therapy, and slower progression to this cancer in a mouse model according to the relative presence of Bifidobacterium. Supplementation with Bifidobacterium species was associated with a better response to immune checkpoint inhibition therapy and a reduction in therapy-associated colitis in mice, as well as better recovery after colon cancer surgery in humans (Routy, Gopalakrishnan, et al., 2018;Wang et al., 2018). and anti-PD-1 antibodies, the abundance of Faecalibacterium prausnitzii was elevated in baseline stool samples from responders relative to nonresponders (Frankel et al., 2017).
The responders also had a higher incidence of immune-related colitis, which is a common side effect of checkpoint inhibition therapy (Garrett, 2015) and probably reflects a tradeoff associated with immune checkpoints, namely that relaxation of the constraints on immunological attacks increases the risk of friendly fire from the immune system. The association between mutualistic symbionts and checkpoint responsiveness suggests that the actions of both mutualists and parasites must be integrated into our understanding of the physiological and evolutionary maintenance of checkpoints. The tradeoff between the positive and negative effects of checkpoint inhibition also raises the possibility that the symbionts that increase immune checkpoint responsiveness could be ambisymbionts that have an overall negative effect on host survival if the negative effects of checkpoint relaxation (e.g., due to autoimmune damage) outweigh the positive effects of escalated destruction of dangerous cells.

| Inter ac tions A mong Par asites an d Mutualis t s
In the last few years, cancer research has increasingly recognized the need to understand the joint contribution of parasites to oncogenesis. Knowledge in this area has been integrated with the barrier theory with many illustrative examples (Ewald andSwain Ewald, 2014, Dheilly et al., 2019). A major challenge will be to broaden this understanding to determine the ways in which mutualists interact with parasites to inhibit oncogenesis. These interactions include protective effects of mutualists against oncogenic pathogens.
Bifidobacterium adolescentis, for example, induces a protein with activity against hepatitis B virus and is associated with a decline in viral titer (Lee, Kang, Shin, Park, & Ha, 2013). This effect may protect against hepatobiliary cancers caused by hepatitis B virus.
Interactions between mutualists and parasites may involve chains of effects. Mutualistic microbes may, for example, inhibit pathogens that indirectly exacerbate cancers for which other pathogens play a more direct oncogenic role. Using malaria as an example, recent research suggests that specific groups of gut microbiota may be protective against Plasmodium infection (Ippolito, Denny, Langelier, Sears, & Schmidt, 2018). This appears to be true for multiple Plasmodium species that are found across separate as well as overlapping geographic regions (Ippolito et al., 2018). By helping to control plasmodia, these gut microbiota may help curb the development of Burkitt's lymphoma, which is caused by the Epstein Barr virus infection acting through the abrogation of barriers to cancer but exacerbated by Plasmodium facliparum (Ewald & Swain Ewald, 2012. Interactions between mutualists and parasites may sometimes involve exacerbation of oncogenic infections by organisms that are generally beneficial. If such interactions cause a symbiont that is normally mutualistic to have a net negative effect on the host in the presence of an oncogenic pathogen, the symbiont would be best considered an ambisymbiont (see Box 1) rather than a mutualist.
In a mouse model of colorectal cancer, for example, A. muciniphila contributed to tumorigenesis, but suppressed it in the presence of Helicobacter typhlonius (Dingemanse et al., 2015). Similarly, A. muciniphila can worsen Salmonella enterica infection (Routy, Gopalakrishnan, et al., 2018), which in turn has been associated with hepatobiliary carcinomas (Samaras, Rafailidis, Mourtzoukou, Peppas, & Falagas, 2010  Aside from the direct contribution of oncogenic pathogens to cancer, explorations of the possible role of the microbiota in the etiology, progression, inhibition, and prevention of cancer are in early stages. The human microbiome and host have been shaped by coevolutionary forces. In this dynamic setting, species outcompete others, provide resources for other microbiota and the host, impact metabolism, exploit and or damage the host, and interact with the immune system. When considering cancer, it may be necessary to look both at the level of function (e.g., provisioning of short-chain fatty acids, competitive colonization of pathogen binding sites) as well as the interactions among symbionts across the symbiosis continuum.

| Conclusion
Symbionts are influenced by their immediate environment and the environments of their hosts, including aspects such as host diets and transmission between hosts (Swain Ewald and Ewald, 2018). Their capabilities for causing or protecting against cancer depend on their evolutionary adaptations to abrogate barriers or enhance host protective systems. A thoroughly integrated barrier theory of cancer will need to incorporate these environmental influences along with an understanding of evolutionary adaptations of host and microbe that result in vulnerability to or protection from oncogenesis.

ACK N OWLED G M ENTS
The manuscript was improved by suggestions from two anonymous reviews.

DATA AVA I L A B I L I T Y S TAT E M E N T
All information used for this study are available in the cited references.