Microbial transitions from health to disease

Recent advances in our understanding of the microbial populations that colonize the human mouth, their acquisition, interdependency, and coevolution with the host, bring a different perspective to the mechanisms underpinning the maintenance of periodontal health and the development of disease. In this work we suggest that our knowledge map of the etiology of periodontal health and disease can be viewed as a broad, highly connected, and integrated system that spans the entire spectrum of microbe/host/clinical interactions. The overall concept of present Periodontology 2000, that the microbial biofilm can be considered a human tissue of bacteriological origin, is entirely consistent with this integrated system view. The health-associated community structure of microbial biofilms can be considered a system that is normally resilient to perturbation. Equally, there is evidence to suggest that the dysbiotic community structure in disease may share similar resilience properties. In both instances, the resilience may be governed by the precise makeup of the acquired microbiome and by the genetics of the host. Understanding the mechanisms that enable the resistance to change of healthy and dysbiotic microbial populations may be important in the development of approaches to prevent the progression of disease and to restore health in diseased individuals.


| INTRODUC TI ON
ultimately lowering of the apical surface of the bone. As before, the model indicates that the interface between these two host compartments is strongly governed by both environment and host genetics. Finally, the clinical signs compartment represents the accumulation of all the outcomes of the aforementioned interactions, which will range from health with low inflammation, and gingivitis with elevated inflammation, to periodontitis with associated inflammation and net tissue loss. Importantly, Page and Kornman 1 also reflected that the clinical signs will impact upon the nature of the microbial challenge to acknowledge the importance of the local environment to the composition and activity of the subgingival microbial communities. 2 Thus, the model describes the potential for a repetitive cycle of increasing microbially driven inflammation leading to increasingly adverse clinical signs and a more aggressive and virulent microbial challenge.

| DE VELOPMENTS IN THE E TIOPATHOG ENE S IS OF PERIODONTAL DISE A SE
There have been a number of subsequent iterations of the original model but the overall concept of a linear interplay between microbial biofilm, inflammatory response, and connective tissue/bone destruction surrounded by environmental and genetic risk factors has remained the core theme. 3 Over the intervening 2 decades, the periodontal research community has made significant progress in describing the fine details within each of the four compartments described above. A summary of this progress is beyond the scope of this review. However, the highlights include: an exponential increase in understanding of the nature of the oral microbiome, its component members and community organization 4-6 ; significant advances in the microbial pattern recognition systems of the host, the associated signaling and cellular responses of innate immunity, and the mechanisms of establishment of the immune systems in the mouth 7,8 ; an appreciation of the control of the inflammatory response and the molecular mechanisms of resolution of inflammation 9 ; enhanced comprehension of the cellular control of tissue turnover in health and the mechanisms of disruption in destructive disease 10 ; and application of these advances to clinical studies of oral health and disease in both animal models and human subjects. 11 The provision of this finer level of detail within each of the four compartments has enabled a more detailed description of the molecular and cellular actions at the interfaces. In so doing, it is now becoming possible to view our knowledge map of the etiology of periodontal health and disease as a broad, highly connected, and integrated system that spans the entire spectrum of microbe/host/ clinical interactions, rather than as a strictly compartmentalized system of four discrete boxes. The overall concept of this issue of Periodontology 2000, that the microbial biofilm can be considered a human tissue of bacteriological origin, is entirely consistent with this integrated system view.
Although researchers in the field have made undoubted progress in filling some of the blank spaces in the original concept design presented by Page and Kornman, 1 there remain significant areas for development. These are especially evident in the identities and roles of the acquired, environmental, and genetic factors, which almost certainly govern the transition points of the system and that ultimately will define susceptibility to the disease. Of course, the range of acknowledged environmental risks for periodontal disease includes modifiable factors, notably tobacco usage and alcohol consumption, as well as a growing list of diseases or conditions including diabetes, obesity, metabolic syndrome, osteoporosis, chronic kidney disease, and low dietary calcium and vitamin D. 12,13 Management of these lifestyle factors and diseases is now firmly embedded in periodontal care. However, with relatively few exceptions, 14 a clear understanding of the mechanism through which these factors may influence disease susceptibility has remained elusive. Similarly, despite some early promise (eg, genetic polymorphisms in the interleukin 6 gene), except for rare, aggressive forms of the disease, our comprehension of the genetic factors that may explain the significant hereditary component of periodontitis is still limited. 12 It has been suggested that the reasons why few true genetic associations for periodontal disease have been identified are two-fold. First, inappropriate sample recruitment: lack of a universally accepted diagnostic criterion for periodontitis; the frequently small sample size studied; and failure to take account of ethnicity differences and environmental effects. Second, inadequate study designs, including, for example, candidate gene association studies with a low sample size, no adjustment for covariates, no study of rare variants, and finally no replication of the results in an independent sample. 15 While this rather harsh critique may have some basis, it may also be the case that hereditary factors other than germ line genetics deserve some consideration.
In this paper, we summarize some recent findings to suggest a more central role for the subgingival microbiome as a key risk factor for susceptibility to periodontal disease that may, in part, explain some of the hereditable components of the disease and the outcome of the interactions at the interface between the microbial challenge and the host response. We draw upon two areas of research investigation to support the proposal: first, increased understanding of the acquisition of the oral microbiome and the role of genetics in this process, and, second, a growing understanding of the stability of microbial populations in both health and disease.

| ACQU IS ITI ON OF THE OR AL MICROB IOME
As pointed out by Wade 16  suggesting that the maternal oral microbiome plays the major role in introducing microbial species to the child. This early microbiome appeared to form the foundation upon which newer microbial communities develop as more colonization niches emerge. The expansion of biodiversity (eg, following tooth eruption) may be attributable not simply to the introduction of exogenous, new species, but to an increase in the abundance of the pre-dentate organisms. In this regard it is significant that two-thirds of the species found in the pre-dentate mucosal microbiome were also seen in the subgingival microbial community of the parent, demonstrating the potentially influential role that the predentate microbiome derived from the mother plays in the development of the periodontal microbial community. A similar mother and infant study by Drell et al 19 demonstrated that the infant gut microbiota harbors a distinctive microbial community that exhibits low similarity with the microbiota that colonizes the mother's gut. By contrast, infants' oral microbiome, as well as mothers' breast milk microbiota, mammary areola microbiota, and oral microbiota, exhibited a high similarity to each other. Interestingly, the mechanism that underlies this maternal-infant transfer may not solely be caused by early exposure following birth. studies are required to confirm the role that vertical microbial transfer from mother to infant plays in the establishment of the oral and subgingival microbiome of the adult. However, it is already clear that this process of establishment of the microbiome in the infant is not simply governed by maternal exposure, and that the genetic landscape of the recipient is also a key determinant.
The influence of the host genetic background on the acquisition of the oral microbiome was examined recently in an elegant investigation involving a genome-wide association study on longitudinal data collected from 752 twin pairs. 21 The study demonstrated that the microbial population diversity of the oral microbiome in monozygotic twins was significantly lower than that for dizygotic or unrelated individuals. This was independent of whether the twins lived together or separately. Furthermore, modeling of these data showed that a number of microbiome phenotypes were more than 50% heritable, consistent with the hypothesis that human genes influence microbial populations: two loci on chromosomes 7 and 12 appeared to be most heritable in the acquisition of the oral microbiome. 21 In summary, there is accumulating evidence to indicate that

| S TAB ILIT Y OF THE OR AL MI CROB I OME IN HE ALTH
Stability is a key feature of the human oral microbiome. 24,25 It is known to be less susceptible to major changes or disruptions by external environmental factors at the individual level. When two individuals were sampled over the course of an entire year, 95% of the operational taxonomic units of the oral bacterial population were found to be stable over the course of the study, while only the minor components of the microbiome were found to be involved in fluctuations. 26 This was in complete contrast to the gut microbiome, which was significantly perturbed by dietary influences, antibiotic usage, and other lifestyle factors. In another study comparing 22 different body sites in 236 healthy adults as part of the Human Microbiome Project, the oral microbiome was found to be the most temporally stable microbial community in the body. 27 This stability, at least in health, may be a function of the dominant and continuous influence of saliva in the nutrition of oral bacteria and only a minimal impact of the diet-except in the case of overwhelming quantities of readily metabolized fermentable dietary sugars-as described by Belstrøm et al. 28 The intraindividual stability of the oral microbiota has been seen to be consistent, despite the interindividual variations in a population. A core healthy human oral microbiome consisting largely of genera such as Streptococcus, Fusobacterium, Prevotella, Rothia, and Neisseria among others has been identified, 29,30 and this has been observed to be stable within individuals for up to a year or longer when tested longitudinally. 24 To date, little information is available on the stability of the subgingival microbiome during episodes of disease although there has been recent progress using animal models of disease, which we will return to later.

| ECOLOG I C AL CON S IDER ATI ON S IN THE COMP OS ITI ON OF THE MI CROB I OTA
The oral cavity is a diverse environment with many different surfaces, topographies, and local environmental conditions. Consequently, the microbial population structure at each ecological site will vary depending on nutrition, pH, host defense factors, and other variables. Hence, when performing a comparison of the microbiology at different locations it is essential to take into account the local environmental conditions. Importantly, these environmental variables may differ independent of the presence or absence of disease. For example, a comparison of the microbial populations in the supragingival vs subgingival tooth surface does not necessarily infer an association with disease. The differences correspond equally to the changes anticipated when comparing ecological sites with different nutritional sources (saliva vs gingival crevicular fluid), different oxygen tensions, differences in pH, and so on. 31 Indeed, even within individual subgingival pockets, the anaerobic microbial population can vary depending on the influence of variations in oxygen sensitivity. 32 The subgingival site, depending upon its depth, may clearly reflect previous disease experience, but the composition of the microbiota at that site does not necessarily correspond to the disease-associated microbial community.
Ximenez-Fyvie et al 33  More studies involving longitudinal sampling within the same individual are needed to better understand the outcome of the disease. Few studies have looked at the subgingival microbiome composition and its variation before and after periodontal therapy. [37][38][39] Although technically challenging in study design, longitudinal sampling coupled with an understanding of the disease experience would represent the gold standard.
Despite these limitations of experimental design that influence much of the microbiological literature in the field, a general consensus is emerging that, during the progression of periodontal disease, the oral microbiota undergoes a major transition, during which the microbial community structure is shifted to an increase in total bacterial diversity. This is accompanied by an increase in the total number of disease-associated bacteria that start dominating the population, which otherwise are present in low numbers in a state of health. 40

| FAC TOR S THAT DRIVE OR AL MICROB IAL TR ANS ITIONS
Three key drivers may be considered important to the shift in microbial populations during periodontal disease. First, certain groups of organisms that subvert the inflammatory response are known to be responsible for influencing a community-wide change on the overall bacterial population. An example is Porphyromonas gingivalis, an organism long associated with the development of periodontal disease. This bacterium has been suggested to exert a "keystone" effect on the oral microbial population during periodontal disease by triggering a state of dysbiosis and inflammation. 43 Porphyromonas gingivalis is involved in both immune subversion and maintaining inflammation in the host tissues by facilitating communication between the C5aR arm of the complement system and toll-like receptor 2 molecules. 44 Studies in mice have also shown that P. gingivalis is not just the sole orchestrator of this shift but is also greatly assisted by the involved activity of the commensal bacterial population. This was demonstrated in germ-free mice, where the absence of the commensal microbiota failed to initiate periodontal disease and alveolar bone loss. 45 More recently, using a combination of metagenomic and meta-transcriptomic approaches in human oral samples, elevated activity of the commensal microbial population, which is not traditionally associated with disease, was observed in the expression of putative virulence factors associated with activities such as stress tolerance and adhesion. 46 This further supports the hypothesis that the entire community acts as a collective pathogenic unit. 47,48 Another potential driver for disease in the oral cavity is the largely inflammophilic nature of the oral microbial population. 49 As stated previously, "periodontal pathogens" are present in oral ecological sites, even in states of health at very low abundance, and these could be responsible for triggering these persistent (albeit low) baseline levels of inflammation, even during healthy conditions. It can be argued that provoking the inflammatory response provides two benefits to an inflammophilic organism: first, through the initiation of tissue destruction, a protected site for colonization is produced, which may enable the organism to outcompete other less inflammophilic organisms; and second, the accumulation of nutrients such as haemin-containing compounds and proteins from tissue exudates/plasma will facilitate the survival of specific types of anaerobic bacteria, thus generating a competitive survival advantage in the ecosystem. Therefore, the inflammophilic nature of the oral microbiome drives a "self-feeding" vicious cycle of tissue damage and bacterial survival and growth. 43 The third driver for the microbial role in periodontal disease is the ability of the oral microbial population to form biofilms. Since these organisms have the propensity to form plaque-like ecosystems, this enables multiple species to coexist in the form of tight complexes with mutually dependent nutritional and survival characteristics.
One of the earliest studies to identify such clustering patterns used a combination of genomic DNA probes and DNA-DNA hybridization methodology to identify five microbial complexes, named red, orange, green, yellow, and purple, with varying patterns of association with health and disease. 50 These microbial clusters tend to operate in a highly mutualistic manner and one member of the cluster is often able to provide protection to all the other members from the inflammatory response of the host. Red complex bacteria have been reported as being involved in consortium in the disruption of homeostasis in the host through activities such as inhibition of interleukin 8 and toll-like receptor 4 signal regulation. 51,52 Some oral bacteria such as Streptococcus mutans are also known for production of specific bacteriocins that directly target and inhibit competitive species that enable its dominance in the plaque 53 and it is evident that similar strategies are at play in the subgingival microbiota. 54,55 The early stages of plaque formation also involve another feature of oral bacteria-coaggregation--with high levels of specificity between mixed species such as Streptococcus and Veillonella, or Streptococcus and Actinomyces. 52,56 Kirst et al 57 used multiple datasets of 16S rRNA gene sequencing data of oral microbial samples and identified two distinct clusters (or "periodontotypes") in the subgingival microbiome, both based on population and functional profiles, a healthy cluster predominated by Rothia and Streptococcus species and another cluster more associated with chronic periodontitis consisting of Fusobacterium and Porphyromonas species. 57 Using cluster analysis on 16S rRNA gene sequencing data of the oral samples of 85 individuals, Boutin et al 58 were able to identify two different microbial "ecotypes" in the oral cavity, one associated with health and/ or mild disease and a second ecotype associated with illness, which could be divided into three subgroups based on the degree of disease progression.

| S TAB ILIT Y AND RE S ILIEN CE OF DYSB IOSIS
As stated previously, very few studies in humans have examined the stability of the dysbiotic microbiota in periodontal disease. However, studies involving animal models of disease suggest that it is not just the healthy oral microbiome that has an inherent stability. Once altered, a newly formed dysbiotic population also exhibits a high degree of stability. Studies in specific pathogen-free mice using an oral gavage model of P. gingivalis have demonstrated longitudinal stability ( Figure 1) as well as the transferability of this dysbiotic, diseased microbial population, both between cohabiting mice in the form of a horizontal transfer, and also from parent to offspring in the form of a vertical transfer (Figure 2). 59 The transmissible dysbiotic community is not just stable at the population level but also results in the same manifestation of destructive periodontal disease as the source, as observed in the mice in the form of alveolar bone loss. Similar examples of the transferability of a diseased microbiota and the disease phenotype have also been observed under other conditions. For example, the obese phenotype was replicated in germ-free mice by the transfer of an "obese microbiota", which was found to be more efficient at energy retrieval from a high-fat diet compared with the recipients of "lean microbiota". 60 One implication of these findings concerns the frequently de- followed by the reacquisition of the original dysbiotic microbial population. We recognize that the stability of a microbial community is not simply maintained by inertia, but by the action of restoring forces within a dynamic system. 61 In the case of the oral microbiome, these may include a complex set of metabolic and functional F I G U R E 2 Resilience of the dysbiotic oral microbiome. (Upper panel) Porphyromonas gingivalis-mediated dysbiotic microbiomes stably transfer horizontally into healthy germ-free mice and lead to periodontal disease in the recipients: bacterial composition of the oral microbiome (left) was determined by culture of control and P. gingivalis-treated C3H/Orl mice and conventionalized germ-free mice of identical genotype co-caged with the respective specific pathogen-free (SPF) mice for 14 days. Alveolar bone levels (right) in control (SPF) mice, P. gingivalis challenged and conventionalized germ-free mice were determined after 16 weeks of co-caging. (Lower panel) P. gingivalismediated dysbiotic microbiomes and disease stably transfer vertically between generations: bacterial composition of the oral microbiome (left) was determined by culture of control and P. gingivalis-treated parents at 16 weeks and litters of controls and P. gingivalis-treated parents at 16 and 28 weeks. Alveolar bone levels (right) were determined at 16 weeks in control and P. gingivalis-treated parents and their respective litters at 16 and 28 weeks. The sizes of the pie-charts are indicative of the variations in the total oral bacterial counts in the different groups. The graphs have been plotted using the observed number of colony-forming units of each microorganism in each group. Bone loss was expressed as negative values relative to the baseline. Each point represents the mean bone level for an individual mouse with horizontal lines representing the mean bone levels per group ± standard deviation. (**P <.05; ***P < .005; ****P < .0005) (Adapted from Payne et al [59]) [Colour figure can be viewed at wileyonlinelibrary.com] interrelationships that develop within dental biofilms and also between biofilms and the host. 62

| SUMMARY
The considerations raised in this work are summarized schematically in Figure 3. After birth, the infant mouth is steadily colonized in an This new system also displays resilience to change, which maintains the stability of the dysbiotic state and leads to microbially driven tissue destruction by the resultant deregulated inflammatory response.
Understanding the nature of the parameters that underpin the resilience of healthy and dysbiotic microbial populations may be important to the development of approaches to prevent the progression of disease and to restore health in diseased individuals.