Recent reviews have focused on a few aspects of colonization resistance. Protective microorganisms can antagonize pathogens through contact-dependent inhibition or the release of antimicrobial molecules. They can also outcompete pathogens for limited resources such as trace metals, nutrients, receptor donors, or adhesion sites, or construct environments hostile to pathogens, for instance, by lowering environmental pH (see summary of recent findings in Table 1). A recent finding, for instance, is that bacteria of the Clostridiales order outcompete Listeria monocytogenes for nutrition in the small intestine and likely antagonize the pathogen in other ways in the large intestine, providing the host a first line of defense from systemic infection (50).

The concept of the ecological niche is central to colonization resistance. Protective microorganisms are thought to “defend” their nutritional niche by killing incoming pathogens or outcompeting them (74). Pathogens, on the other hand, are thought to overcome colonization resistance by creating or exploiting new spatial-temporal niches (19). However, a niche is not just a nutritional environment, but any environmental feature relevant for colony survival and growth. Protective microbes can also inhibit pathogenic establishment and growth by disrupting pathogenic biofilms, bacterial collectives that undergo regular developmental “life cycles” and protected by synthesized extracellular polysaccharides matrices. Proteases released by Staphylococcus epidermidis can degrade the matrices of pathogenic Staphylococcus aureus biofilms (75, 76). Lactic acid bacteria can disrupt matrix synthesis by interfering with pathogen virulence genes, for instance, by decreasing the expression of Streptococcus mutans genes involved in matrix glucan production (77) or modulating S. aureus gene expressions related to the production of intercellular adhesion polysaccharides (78).

Protective microorganisms can also influence quorum sensing, the cell-to-cell communication system that allows bacteria to perceive information about bacterial population density and to regulate collectively virulence factor production and biofilm development (79, 80). Soluble molecules released by probiotics can interfere with the S. aureus accessory gene regulator (agr) quorum sensing system, which regulates the switch between biofilm and free-floating lifestyles (81). Subtilosin, a protein secreted by Bacillus subtilis, also interferes with the quorum sensing of Gardnerella vaginalis, preventing biofilm formation (82). Biosurfactants are well-known anti-adhesion, anti-biofilm agents that can also disrupt cell-to-cell signaling (83, 84). Cell-bound biosurfactants of lactic acid bacteria disrupt the biofilms of multi-drug-resistant Acinetobacter baumannii, Escherichia coli, and S. aureus (85) as well as Serratia marcescens strains (86).

Colonization resistance is not the only way microorganisms can protect the host from pathogens, as not all types of protection act by inhibiting pathogenic colonies establishment and growth. The role of resident microorganisms in barring entry and dispersal of alien populations is oftentimes ignored (87). Furthermore, inhibition of quorum sensing can affect pathogenicity, the effects of pathogens, without inhibiting bacterial viability, which, as a therapeutic option can avoid the selection of drug-resistant bacteria (88). We thus propose looking beyond colonization resistance to adopt a community ecology point of view that better understands the multiple ways resident communities can prevent or disrupt pathogen invasion.

In invasion ecology, residents can disrupt alien invasion by intervening with alien entry and diversification, by inhibiting its establishment or growth, by mitigating its negative impacts on local communities, or by preventing alien dispersal into other vulnerable environments (89–91) (see summary of stages of invasion in Table 2). A more comprehensive picture of resistance against pathogens must take into consideration the multiple stages of invasion and go beyond competition as the only way residents can fight against invaders.

We thus distinguish three major microbial obstacles against invading pathogens (Table 2; Figure 2): colonization resistance exclusively refers to the prevention of pathogen establishment of a persistent colony and population growth, containment is the prevention of pathogen spread into another body site, sometimes leading to systemic infection, and finally, disease tolerance is the limitation of pathogenic impact on host tissues without killing the pathogens.

Containment is the controlled localization of microbes and their effects within a particular location inside the host body. One example is the intestinal lumen. The epithelium barrier and gut mucosal immune system prevents the translocation of microorganisms from the lumen into the host. Mucosal-associated microorganisms have long been postulated as reinforcers of gut barriers, both directly and indirectly through the host (94–97). For instance, in various organisms, bacteriophages that adhere to mucus glycoproteins can prevent translocation of bacteria, thus providing a host-independent protection of mucosal surfaces against bacterial infections (98). Saccharomyces boulardii also prevents Salmonella liver translocation by directly binding to the pathogens (99).

Secreted factors can interfere with the translocation of gut pathogens into and beyond the epithelia. Supernatants of Enterococcus mundtii and Lactobacillus plantarum inhibit the invasion of L. monocytogenes into epithelial cells (100). A secreted, non-bacteriocin component from Escherichia coli Nissle 1917 also reduces the efficiency of Salmonella enterica serovar Typhimurium epithelial invasion and blocks invasion by many pathogens without eliminating them (101). Pathogen translocation can also be prevented by interference with pathogen adhesion to the epithelium (102, 103). It is well known that resident commensals, especially lactic acid bacteria, compete with pathogens for adhesion sites, for instance, the competition between Lactobacillus reuteri and enteropathogenic E. coli (104). Short chain fatty acids are products of anaerobic intestinal microbiota fermentation of dietary fibers. Butyrate inhibits the virulence genes that enable intestinal S. enterica subspecies to enter the epithelium, specifically pathogenicity island 1 (105). Acetate from Bifidobacteria can enhance epithelial cellular defense functions, inhibiting the translocation of Enterohemorrhagic E. coli (EHEC) shiga toxin into the bloodstream (106).

Disease tolerance, on the other hand, is the limitation of pathogen-induced damage of host tissues without elimination of the pathogen. Various types of processes are associated with disease tolerance: the neutralization or inhibition of pathogenic toxins, the healing process of the host, and the regulation of destructive inflammation and maintenance of metabolic homeostasis (see Box 3). While the latter two are mediated through the host, the first can be host-independent.

An example of host-independent disease tolerance is the ability of resident microbiota to inhibit gene expressions of toxins. Soluble molecules from Bacteroides thetaiotaomicron and the bacteria-produced short chain fatty acid butyrate, for instance, can inhibit EHEC shiga toxin gene expressions and pathways (111, 112). The quorum sensing molecule AI-2 of Ruminococcus obeum also downregulates the Vibrio cholera toxin operon (113). Secreted organic acids can mitigate pathogen toxicity in various ways: lactic acid from Streptococcus thermophilus inhibits Clostridium difficile toxin A gene expression (114) while acetic acid secreted by a Bifidobacteria strain inhibits EHEC shiga toxin production in part by lowering environmental pH (115). Remarkably, the growth of shiga-producing E. coli was not inhibited by the B. breve strain even though toxin production was halted.

The damaging effects of pathogens can also be diminished by the degradation of released toxins. Proteases secreted by a probiotic Bacillus clausii strain acts against toxins from C. difficile and Bacillus cereus (116). Another example is a Saccharomyces boulardii protease that inhibits C. difficile toxin A and B (117). Bifidobacteria and Lactobacilli strains mitigate C. difficile cytotoxic effects by inactivating their secreted toxins (118, 119).

In sum, colonization resistance, containment, and disease tolerance are diverse effects of protective microbes against the spread, establishment and growth, and impact of pathogen invasion. Broadening the concept of protection to include the effects of disease tolerance and containment allows us to consider pathogenicity in a wider context, taking into account the overall ecological process of pathogen invasion (Figure 3). The detailed mechanisms underlying these three types of effects—colonization resistance, containment, and disease tolerance—may overlap. For instance, competition over adherence sites prevents the growth of biofilms that enable the establishment, growth, and spread of many pathogens (Figure 4).

The examples reviewed so far are direct microbe-to-microbe effects. We now turn to indirect forms of protection that involve host systems.

Examples of host-mediated protection, especially colonization resistance, by gut microbiota have already been extensively reviewed (see Box 1). Indirect colonization resistance occurs when resident microorganisms induce host reactions that act against pathogens.

Resident microbiota sampled by pattern recognition receptors can activate downstream secretion of antimicrobial peptides that protect the host by killing or inactivating bacteria, fungi, and viruses, primarily by destroying cellular membranes (124, 125). These include C-type lectin Reg3-γ (38), α and β defensins (35, 126), and the ribonuclease angiogenin (127). The effects are often pathogen- and host-specific (128–130).

Gut microbes induce the development of lymphoid tissues that sample and secrete Ig A into the gut lumen as well as regulate secretory IgA (SIgA) secretion levels (131). SIgA can help trap and exclude pathogens (“immune exclusion”), but when bound to certain microorganisms, can also selectively promote commensal biofilms that confer colonization resistance against pathogens (“immune inclusion”) (132–136). Specificity to microorganisms could be enhanced through a positive feedback loop mediated by SIgA retrotransportation (137–139).

Microbes can also induce protective pro- or anti-inflammatory immune activities by altering the balance between host T-cell subsets (2, 140–142). By tilting the balance toward pro-inflammatory pathways, commensals help generate host-mediated attacks against pathogens. Segmented filamentous bacteria, for example, induce inflammatory Th17 immunity (143), leading to a generally protective, although context-dependent immune state (144). Protozoan Tritrichomonas musculis protects from gut S. Typhimurium infection through promotion of inflammatory Th1 and Th17 type immunity (145).

In the skin, microbiota could be considered an “endogenous adjuvant” of the skin immune system, exerting its influence via the release of products (such as antimicrobial peptides or metabolites) and/or via the modulation of innate and adaptive immunity without invoking inflammatory responses (146–148).

In the respiratory tract, local manipulation of the microorganism composition may have profound consequences on the capacity of the host to mount protective responses (149). For instance, intranasal inoculation of mice with live or heat-inactivated Lactobacillus spp. and non-pathogenic Listeria spp. protects against secondary lethal infection with the virulent pneumonia virus by suppressing virus-induced pro-inflammatory cytokines and viral load (150).

Microbes can also indirectly protect the host by containing pathogens within the gut and lung lumen, as breaching of the epithelial layer triggers systemic inflammation. New developments shed light on the intimate relation between microorganisms and barrier functions. The gut epithelium barrier is fortified by maternal microorganisms before birth (151). After exposure to environmental microbes, microorganisms strengthen and protect the barrier (152). In vitro treatment of Saccharomyces boulardii, for instance, prevents Bacillus anthracis toxins from destroying the integrity of epithelium cells and the tight junctions between them (153). In antibiotics-treated mice, translocation of bacteria and an increase in inflammatory responses can be observed (154, 155).

The separation of host and microorganism with a mucus layer is in part regulated by resident microorganisms. The mucus layer effectively separates luminal components, especially microorganisms, from the epithelium. The constant outgrowth of mucus pushes microorganisms out and provides a medium that concentrates antimicrobial elements such as antimicrobial peptides and SIgA. In the small intestine, Muc2 mucin, the backbone of mucus in mice, requires bacterial cleavage of a proteolytic enzyme to detach mucin from secreting goblet cells; in the colon, “sentinel” goblet cells activated by bacterial components quickly respond by extruding and releasing an explosion of Muc2 (156, 157). Helminthes such as Trichuris muris can restore the Nod2-deficient abnormal goblet cells of mice through T_H2-mediated immunity, thereby inhibiting the colonization of pro-inflammatory Bacteroides species (158).

Localized protection is also achieved by disease tolerance. In the gut, Bacteroides fragilis affects systemic T-cell responses through the action of the bacterium-derived polysaccharide A, which protects against pathobiont H. hepaticus colitis via the production of anti-inflammatory IL-10 by CD4+ T cells and the promotion of regulatory T cells (159). Moreover, this protection has no effect on pathogen fitness or on population load in the intestine. SIgA can neutralize intracellular toxins and viruses, as well as directly suppress bacterial virulence mechanisms (133, 135).

In sum, resident microbiota can protect the host by locally inhibiting the establishment and growth of pathogens, mitigating pathogen-induced damage to the host, and preventing the spread of microbes to other sites. Local protection is ensured by mechanisms that are directly targeted against microorganisms or indirectly through the host immune system.

Microbiota at one location can induce systemic colonization resistance in the host. Composition of the gut microbiota modulates the severity of malaria, where a fecal transplant from resistant mice transfers the resistance to germ-free mice by elevating humoral immune responses. The resistant gut microbiota are characterized by an increased abundance of Lactobacillus and Bifidobacterium (173). In humans, gut microbe E. coli O86:B7 induces antibodies that target plasmodium sporozoites, conferring a protective cross-response against malaria transmission (174).

Systemic innate and adaptive immune responses to Toxoplasma gondii infection in humans rely on the indirect stimulation of dendritic cells by normal gut microflora (175). Several studies have shown that gut microbiota also have the potential to reduce systemic viral infection and disease (176). A clinical trial observed a positive effect of a Lactobacillus strain as an immune adjuvant for live-attenuated H3N2 influenza vaccine in healthy adults, with higher seroconversion rates in patients treated by probiotics (177), suggesting a non-specific immune response activation. In HIV infection, higher abundance of the gut Lactobacillales order in patients is positively associated with CD4+ T cell count and negatively associated with viral load, indicating that bacteria from Lactobacillales could in some direct or indirect way modulate the infectivity or pathology of HIV infection (178).

Finally, as a basal stimulant, bacteria can prime activation threshold of antiviral innate immunity against systemic viruses (179). Antibiotic-treated mice before lymphocytic choriomeningitis (LCMV) infection elicited an impaired innate and adaptive immune response to LCMV infection, and an increased mortality after influenza infection (179).

Microbiota or its components that are normally distributed across the host system can also induce system-wide colonization resistance in the host. Recently, the idea that viruses can be mutualistic, and in particular can offer protection against pathogens, has gained popularity (180–182). A healthy human is infected by more than 10 permanent chronic systemic viruses, and this number may in fact be far higher (182). Two cases show that viruses induce in their hosts a higher basal immunity that explains normal immunity as well as responses to novel microbes. First, baseline activities can be upregulated and maintained by viruses in latent states, conferring protection against subsequent infections (176, 183). In mice, gammaherpesvirus 68 or murine cytomegalovirus (CMV) (which are models for the human pathogens Epstein–Barr virus and CMV, respectively) in dormancy confers symbiotic protection against bacterial infection to L. monocytogenes and Y. pestis in an antigen-independent way, involving interferon production and macrophages systemic activation (184, 185).

In a clinical study, HIV patients co-infected with non-pathogen GB flavivirus showed reduced mortality rate, suggesting the possibility of an inhibition of HIV replication due to this flavivirus (186). HIV replication was reproducibly inhibited in cultures of peripheral-blood mononuclear cells by GBV-C coinfection (187). Patients with GBV-C infection present an increased proportion of naive T cells and a reduction in T-cell activation and proliferation that could increase disease tolerance and finally the survival among HIV infected individuals co-infected with GBV-C (188).

It is generally assumed that microorganisms are pathogenic when spread throughout the system while non-pathogens are contained within the gut and lung lumens and outside of skin surfaces. Indeed, bacteremia and endotoxemia, i.e., the spreading of bacteria or bacterial structural components through blood circulation, respectively, are usually pathological. For instance, intra-uterine infection during pregnancy in mice can be caused by pathogens that do not belong to vaginal microflora, but to oral microorganisms via hematogenous transmission (189).

However, non-pathological, physiological blood circulation of microorganisms or microbial fragments occurs in healthy subjects (190–192). Substantial numbers of live gut species have also recently been found living in the mesenteric lymph nodes and systemic lymphoid organs under normal situations (164). In this section, we thus consider the possibility that circulation of bacteria or their metabolic compounds can constitute a long-reaching route of protection. Even though a long-reaching seeding through the bloodstream and lymphatic systems has never been demonstrated, it is a possible route protective microbes could take.

The involvement of circulating long-reaching mechanisms in host protection is only a hypothesis at this stage, but investigating this potential long-reaching influence undoubtedly constitutes a very promising avenue for future research, paving the way to targeted experiments.

To the best of our knowledge, the only explicit long-reaching protective effects from one local site to another described in the literature are the immune-mediated relation between the gut microbiota and lung infections (203, 206). Klebsiella pneumoniae lung infection has been extensively studied as a model of pulmonary infection. Germ-free mice infected with K. pneumoniae are drastically susceptible to bacterial infection in an IL-10-dependent manner, with an increase of bacterial growth and dissemination. Interestingly, activating microbial pattern recognition receptors helps fight against K. pneumoniae infection in the lungs (207, 208). Another example is tuberculosis. Helicobacter pylori infection is suspected to modify the clinical outcomes of M. tuberculosis infection, with the presence of H. pylori associated with a lower rate of tuberculosis infection (209), but increased tuberculosis severity (210). In this case, a microbe at one location (H. pylori in the stomach) could modulate long-distance immunity at yet another location (in the lungs) in response to infection.

Antibiotic-treated mice exhibit impaired innate and adaptive mucosal immune responses to influenza infection, with increased damages and host mortality (179). A decrease in the number of gut commensals via treatment with a broad-spectrum antibiotic resulted in blunted T-cell and B-cell responses to an intranasal infection with the A/PR8 strain of influenza (211). Activation of Toll-like receptors restored the immune response in antibiotic-treated mice through inflammasome-mediated induction of interleukins such as IL-1β and IL-18. The authors suggest that either the microbial products can diffuse systemically, or activation of the inflammasome does not need to occur at the site of infection.

In an E. coli pneumonia model, antibiotic depletion of commensals in mice also causes a drastic bacterial burden both in lungs and blood, with a significant increase of mortality. Lipopolysaccharides supplementation during antibiotherapy reversed these effects, suggesting a distal action of commensal microbes through toll-like receptors (212). Deleterious effects of antibiotic depletion have been shown in a mouse model of Streptococcus pneumoniae infection, reversed by fecal microbiota transplantations that enhance macrophage functions in the primary alveolar (213).

Gut microbiota also play a role in expanding and maintaining viral-specific memory T-cell populations in the lungs. In a mouse model of MCMV-associated lung disease, the frequency of virus-specific CD8⁺ T cells in the MCMV-infected lungs of germ-free mice was restored by fecal bacteria from specific pathogen-free mice, likely through cross-activity between gut microbiota peptides and epitopes of MCMV-specific memory T cells (214).

Furthermore, mice challenged with pulmonary staphylococcal infection but lacking segmented filamentous bacteria in their gut microbiota showed more severe infection with higher bacterial load and mortality, associated with diminished lung concentration of Th17 immune effector cells. Reintroduction of segmented filamentous bacteria was sufficient to restore protective effect (215). Antibiotics also significantly decreased lung Th17 cell numbers during pulmonary acute fungal infection, restored by segmented filamentous bacteria colonization (216).

Even though direct and host-mediated causes of microbial protection are oftentimes difficult to distinguish and intertwined at a local scale, the direct versus indirect distinction can help identify ecological versus physiological routes of influence at a global scale. Current evidence indicates that long-distance protection is mostly mediated by the host, but ecological routes are also available when we consider the way fluids and air flow through the host. Many organs are interconnected through airways and liquid channels. We hypothesize microbe-induced protection of downstream conditions in the respiratory tract and the gastrointestinal tract.

The respiratory tract ecosystem consists of the upper and lower respiratory tracts, with the latter further divided into trachea, bronchi, and bronchioles. The gastrointestinal tract ecosystem is divided into the stomach, small intestine, and large intestine, with secretions that come from associated organs such as the gall bladder, liver, and pancreas. It is possible that microbial communities in one patch of the lung and gut meta-communities can protect the host by influencing the entry and growth of pathogens in a downstream patch.

Bronchoalveolar lavage fluid samples of the lungs reveal a substantial microbiota community with multi-kingdom interactions due to air-borne fungi (217). The oral microbiota are the physiological source of lower respiratory tract microorganisms, predominately acquired through microaspiration and especially during sleep (218). It is also the source of disease-inducing pathogens in the lungs, especially pneumonia and cystic fibrosis. The sampled microorganisms of a healthy lung consist primarily of transitory populations determined by a steady balance of incoming migration and outgoing expulsion through physiological clearance or immune responses. In disease states, such as cystic fibrosis, regional growth conditions support the settled colonization and expansion of microbial communities (219, 220). The respiratory tract ecosystem is thus an excellent example of dominance in dispersal processes over processes of establishment and growth, similar to the metacommunity principles governing islands (220–224).

Microbial communities in the upper respiratory tract (e.g., oral cavity) are likely to confer resistance against the colonization and growth of pathogens in the lower respiratory tract. There is an association between oral dysbiosis and lower respiratory tract infections such as pneumonia. The absence of normal residents in the upper respiratory tract may thus contribute to the overwhelming migration of residential or opportunistic pathogens to the lower respiratory tract (225). Pathogen “blooms” in the lower respiratory tract generally occur under two conditions: when there is a positive influx of microorganisms over those eliminated or when altered local conditions favor pathogenic growth [see adapted island biogeography model in Ref. (223), and review in Ref. (224)]. We thus suspect that a loss of colonization resistance of upper respiratory tract residential microbiota may be the reason for lower respiratory tract microbial establishment and growth, resulting in disease.

Another possible downstream protective effect is in the gastrointestinal tract. The habitat of the gastrointestinal tract is similar to a river ecosystem (226). The intestine receives inputs from the source (nutrients and acids from the stomach) and associated organs (liver, gall bladder, and pancreas) and the flow slows down at the colon, which is dominated by greater bacterial loads and the accumulated nutrients and metabolites from upstream.

Secondary bile acids inhibit the germination and growth of C. difficile in the colon. Bile acids are released and mostly reabsorbed upstream in the small intestine, decreasing greatly in concentration further downstream (227, 228). However, small intestinal microbiota can inhibit re-absorption by deconjugating bile acids, thus promoting their excretion down and out of the large intestine (229). Downstream secondary bile acids in the colon can thus depend on upstream commensals. Indeed, antibiotic treatment that decreases colonization resistance against C. difficile also decreases bile acid amounts in fecal matter (64), which are both restored by fecal matter transfer (65). Upstream commensals can thus possibly protect the host from downstream pathogenic growth by releasing metabolite inhibitors.

Microbes can protect the host by regulating the migration flows between the lung “island” communities and by influencing “downstream” communities in the gut. The analogy between lung and gut ecosystems to island and river ecosystems, respectively, point to new perspectives. These are clearly local-to-local effects, although not necessarily organ-to-organ effects (they concern “subparts” of organs rather than organs per se). Yet they point to the possibility of organ-to-organ protective effects, and offer insights about how these effects could be investigated in the future. It is worth investigating whether migrations between other sites, for instance, the upper respiratory tract and the stomach (218), or direct dispersal between oral and vaginal microbiota via sexual behavior, also exhibit upstream-to-downstream colonization resistance against infections.