Friends with Benefits: An Inside Look of Periodontal Microbes’ Interactions Using Fluorescence In Situ Hybridization—Scoping Review

Fluorescence in situ hybridization (FISH) has proven to be particularly useful to describe the microbial composition and spatial organization of mixed microbial infections, as it happens in periodontitis. This scoping review aims to identify and map all the documented interactions between microbes in periodontal pockets by the FISH technique. Three electronic sources of evidence were consulted in search of suitable articles up to 7 November 2020: MEDLINE (via PubMed), Scopus (Elsevier: Amsterdam, The Netherlands), and Web of Science (Clarivate Analytics: Philadelphia, PA, USA) online databases. Studies that showed ex vivo and in situ interactions between, at least, two microorganisms were found eligible. Ten papers were included. Layered or radially ordered multiple-taxon structures are the most common form of consortium. Strict or facultative anaerobic microorganisms are mostly found in the interior and the deepest portions of the structures, while aerobic microorganisms are mostly found on the periphery. We present a model of the microbial spatial organization in sub- and supragingival biofilms, as well as how the documented interactions can shape the biofilm formation. Despite the already acquired knowledge, available evidence regarding the structural composition and interactions of microorganisms within dental biofilms is incomplete and large-scale studies are needed.


Introduction
Periodontitis is a chronic inflammatory disease that results in the loss of the toothsupporting tissues, which can lead to tooth loss when untreated [1]. It is the consequence of the imbalance between the polymicrobial microbiota, that colonizes the tooth surfaces in form of biofilms, the immune and inflammatory host response within the gingival tissues [1][2][3] and the individual variations in the stock of these taxa [4]. This imbalance, in susceptible individuals, results in the loss of clinical attachment, triggering the formation of periodontal pockets [5,6]. As with other chronic diseases, periodontitis requires supportive care to avoid its recurrence [7]. Furthermore, mounting evidence suggests that many chronic disorders, such as diabetes and cardiovascular diseases, are related to periodontitis via systemic inflammation caused by periodontal bacteria [8].
A total of 2074 genomes and 529 taxa of microbes are estimated to inhabit the oral ecosystem [9]. While most of these microorganisms are not directly associated with periodontitis, they may create the necessary conditions (e.g., nutrient supply or oxygen depletion) for other microorganisms to grow and disrupt the periodontal balance between host and microbes. Microbes' spatial organization, the understanding of the interactions between microbes in supra-and subgingival biofilms, and how it influences the establishment, development, and the outcome of the periodontal diseases have been emerging subjects in periodontal microbiology.
The development of culture-independent methods has allowed the identification of periodontitis-associated uncultured and fastidious species, providing a more detailed look at the bacterial communities in periodontal tissues. Fluorescence in situ hybridization (FISH) has proven to be particularly useful to describe the microbial composition and spatial organization of mixed microbial infections, in time and space within their natural context. This molecular technique relies on the hybridization of single-stranded, fluorescently labeled DNA-, RNA-, or nucleic acid mimics-targeted oligonucleotides probes with fluorescent molecules (e.g., fluorochromes) that hybridize to its complementary conserved 16S or 23S rRNA sequences in the microorganism [10][11][12].
Even in simple samples where only two or three fluorochromes are used, spectral overlap and crosstalk seem difficult to eradicate when performing standard FISH in multiplex experiments. Additionally, the use of bandpass filters in fluorescence image acquisition restricts the number of fluorochromes that can be simultaneously distinguished, making it problematic to recognize one signal from another simultaneously with certainty and consequently restricting the microscopic identification of various taxa of microbes in single samples [13,14].
To avoid spectral overlap and crosstalk, Valm, A.M. et al. (2012) [14,15] developed a strategy-the Combinatorial Labeling and Spectral Imaging (CLASI)-and blended it with FISH (CLASI-FISH). This technique relies on the labeling of microbes of interest with two or more combinations of fluorochromes [14], increasing spectral discrimination of the fluorochromes that have a propensity to overlap in excitation and emission spectra. CLASI-FISH is currently capable to distinguish unambiguously 120 differently labeled organisms [16], resulting in exclusive mixed colors that are distinguished by the application of linear unmixing algorithms.
This scoping review has the purpose to identify and map the existing evidence about the in situ and ex vivo interactions between microbes in periodontal pockets, identified with the FISH technique, as well as to identify and analyze any knowledge gaps that can point to further research directions.

Focused Question
This scoping review was conducted following the guidelines of the Transparent Reporting of Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA-ScR) [17], to answer the focused (PCC-Population, Concept, Context) question: "What are the identified interactions between microbes in periodontal pockets as evaluated by the FISH technique?"

Search Strategy and Information Sources
Three electronic sources of evidence were consulted in search of suitable articles that matched the aim of this review: MEDLINE (via PubMed), Scopus (Elsevier: Amsterdam, The Netherlands), and Web of Science (Clarivate Analytics: Philadelphia, PA, USA) online databases. The databases were consulted up to the 7 November 2020.
The electronic database search was supplemented with a hand search across the references of all included papers. The authors of the included papers were contacted to find additional unpublished images that could fulfill our inclusion criteria and were asked permission to reproduce images in this scoping review. When required, additional permission of all reproduced images was granted by the Copyright Clearance Center (Danvers, MA, USA).

Eligibility Criteria
All studies that described ex vivo and in situ interactions between, at least, two microorganisms were found eligible. We only incorporated articles applied in humans.
The exclusion criteria detached experiments using other molecular cytogenetic techniques rather than the FISH technique, articles with no images, or experiments that used manufactured bovine enamel/dentin slabs, acrylic, or epoxy resin appliances to extract dental biofilm. Restrictions were also made to article type excluding reviews, case reports, or letters.

Screening and Selection of Sources of Evidence
Two independent reviewers (G.M.E., L.M.) selected papers by evaluating their titles and their abstracts information. Any disagreements in the acquired results were resolved upon discussion with a third reviewer (A.S.A.). Furthermore, the selected articles were then read in full and were not included if did not fulfill the inclusion criteria or if any of the exclusion criteria was detected. Using the Cohen's Kappa method and IBM SPSS (Version 26) program, the interrater reliability (IRR) was calculated.

Data Extraction and Analysis
Data from the included articles were processed for analysis. Information regarding the year of publication, study design, FISH conditions, probes' names and sequences, images, microorganisms found, their location, and their interactions within the biofilm were collected in parallel by G.M.E. and L.M. The data's interpretation and analysis were debated until a consensus was reached.

Synthesis of Results
The studies were categorized based on the collected data. Evidence is reported in a table and a visual representation, that incorporates all the images in the finest resolution acquired from the FISH experiments performed on the included papers.

Selection of Sources of Evidence
The initial electronic search resulted in 2090 studies, of which 832 located in PubMed, 625 in Web of Science (Clarivate Analytics: Philadelphia, PA, USA), and 633 in Scopus (Elsevier: Amsterdam, The Netherlands). After removing 1144 duplicated studies, 905 studies were rejected after screening articles by title and abstract. The remaining 41 studies were obtained and analyzed. After full-text reading, 8 studies met the inclusion criteria. Additional hand searching of the reference lists of the selected papers retrieved 10 additional studies for full-text reading, of which 2 papers met the inclusion criteria. As such, 10 papers [3,[18][19][20][21][22][23][24][25][26] were included in the present scoping review.
The Cohen's Kappa method was used to calculate the interrater reliability (IRR) in the selection process by titles and abstracts, which yielded a value of 0.965. The PRISMA flow diagram ( Figure 1) demonstrates the selection process.

Characteristics of Sources of Evidence
The general characteristics of the six case series, the three case-control studies, and the one cross-sectional study are presented in this section (Table 1). We divided certain images into three panels to make the analysis of the gathered images simpler (Figures 2-4). Due to the well-defined approach, we paid special attention to the results of one particular study (Mark Welch, J.L., et al.

Characteristics of Sources of Evidence
The general characteristics of the six case series, the three case-control studies, and the one cross-sectional study are presented in this section (Table 1). We divided certain images into three panels to make the analysis of the gathered images simpler (Figures 2-4). Due to the well-defined approach, we paid special attention to the results of one particular study (Mark Welch, J.L., et al.     [18] only 13 out of 57 genera tested had at least 3% mean abundance and were also prevalent, being identified in more than 90% of supragingival specimens (Corynebacterium sp., Capnocytophaga sp., Fusobacterium sp., Leptotrichia sp., Actinomyces sp., Streptococcus sp., Neisseria sp., Haemophilus/Aggregatibacter sp., Porphyromonas sp., Rothia sp., Lautropia sp., Veilonella sp., and Prevotella sp.).
Corynebacterium sp. was remarkably specific to supragingival (12%) and subgingival (8%) plaque. By contrast, genera such as Streptococcus sp., Veillonella sp., and Haemophilus sp. occupied a wide range of substrates in oral ecosystems.
A complex microbial consortium in a hedgehog-shaped structure was observed, showing the spatial organization of the plaque microbiota. In short, these hedgehog structures were radially organized, with a multi-taxa consortium composed of a skeleton mainly of Corynebacterium sp. with Streptococcus sp. cells arranged around the distal tips, a multigenus filament-rich halo composed of Fusobacterium sp., Leptotrichia sp., and Capnocytophaga sp. cells, and a periphery of corncobs structures composed by a filamentous core bordered primarily with Streptococcus sp. cells, Porphyromonas sp. and Haemophilus/Aggregatibacter sp. ( Figure 5).   µm. All images were reprinted and adapted with the publisher's permission.
Corynebacterium sp. was remarkably specific to supragingival (12%) and subgingival (8%) plaque. By contrast, genera such as Streptococcus sp., Veillonella sp., and Haemophilus sp. occupied a wide range of substrates in oral ecosystems.
A complex microbial consortium in a hedgehog-shaped structure was observed, showing the spatial organization of the plaque microbiota. In short, these hedgehog structures were radially organized, with a multi-taxa consortium composed of a skeleton mainly of Corynebacterium sp. with Streptococcus sp. cells arranged around the distal tips, a multi-genus filament-rich halo composed of Fusobacterium sp., Leptotrichia sp., and Capnocytophaga sp. cells, and a periphery of corncobs structures composed by a filamentous core bordered primarily with Streptococcus sp. cells, Porphyromonas sp. and Haemophilus/Aggregatibacter sp. (Figure 5). The corncobs at the periphery showed that "kernels" (coccoid cells) were composed of different taxonomic types and could be either single or double layered ( Figure 6). Singlelayer corncobs had coccoid cells of both Streptococcus sp. or Porphyromonas sp. (in some cases Porphyromonas sp. kernels coexisted with Streptococcus sp. around the same filament), whereas double-layer kernels consisted of a combination of Streptococcus sp. in the inner layer and Haemophilus/Aggregatibacter sp. in the outer layer.
The most common type of kernels visualized were the ones that had a single layer of Streptococcus sp. cells surrounded by a partial or complete layer of Haemophilus/Aggregatibacter species. Porphyromonas sp. cells were only observed organized in single-layer corncobs. On contrary, cells of the genus Haemophilus/Aggregatibacter sp. were only found forming double-layer structures, exclusively with Streptococcus sp. cells, demonstrating an undoubtedly specific relationship. Competitive, exploitative, or mutualistic interactions between these taxa were detected in this specific biofilm architecture between single-and double-layers corncobs.
composed of a filamentous core (sometimes visualized as Corynebacterium filaments but often not stained) bordered primarily by Streptococcus sp. cells but also by Porphyromonas sp. and Haemophilus/Aggregatibacter sp., both in proximal contact with Streptococcus sp. cells. (G) On the periphery of these corncob structures, Corynebacterium sp. filaments pass through the halo that is highly densely colonized with elongated rods of Fusobacterium sp., Leptotrichia sp., and The corncobs at the periphery showed that "kernels" (coccoid cells) were composed of different taxonomic types and could be either single or double layered ( Figure 6). Single-layer corncobs had coccoid cells of both Streptococcus sp. or Porphyromonas sp. (in some cases Porphyromonas sp. kernels coexisted with Streptococcus sp. around the same filament), whereas double-layer kernels consisted of a combination of Streptococcus sp. in the inner layer and Haemophilus/Aggregatibacter sp. in the outer layer.
The most common type of kernels visualized were the ones that had a single layer of Streptococcus sp. cells surrounded by a partial or complete layer of Haemophilus/Aggregatibacter species. Porphyromonas sp. cells were only observed organized in single-layer corncobs. On contrary, cells of the genus Haemophilus/Aggregatibacter sp. were only found forming double-layer structures, exclusively with Streptococcus sp. cells, demonstrating an undoubtedly specific relationship. Competitive, exploitative, or mutualistic interactions between these taxa were detected in this specific biofilm architecture between single-and double-layers corncobs.  Within hedgehogs, bacteria do not form broad single-taxon groups; rather, cells were observed intermingling with at least four different taxa. The authors only did not find this type of interaction between Actinomyces sp. and Corynebacterium sp. cells, where were visualized irregular clumps in the base of the hedgehogs or nearby the hedgehogs, rather than intermixed within this structure.
Hedgehogs were the most observed type of structure. Some samples had multiple hedgehogs' structures adjacent to each other ( Figure S1), especially on the tooth surface on the buccal side and plaque from the gingival margin. Another kind of consortia was also found: a cauliflower structure in plaque ( Figure S2) constituted by Lautropia sp., forming the center of the structure, surrounded by Streptococcus sp., Haemophilus/Aggregatibacter sp., and Veilonella species. Dispersed cells of Prevotella sp., Rothia sp., and Capnocytophaga sp. were also visible.
The findings collected from the included articles in this scoping review are summarized in Table 2. We also included a supplemental table, with details on all the oligonucleotide probes and the FISH conditions used in each microorganisms' analysis (Table S2). Table 2. A summary in terms of the location where the dental biofilm was collected, the microorganisms detected, their shape and spatial organization within the biofilm, the most relevant interactions detected, and the microscopic technique used in each study. The abbreviations can be found at the bottom of the table.   Rothia sp.

Cells of at least four different taxa interact with one another at a micron scale
Lautropia sp.

Summary of Evidence
Our findings from the obtained images indicate a paucity of research focusing specifically on the study of the spatial organization of periodontal pathogens within oral biofilms from the Bacteria domain and its etiologic significance.
In our results, the biofilm's fluorescent intensity and variety of labeled microorganisms increased as images drifted from the tooth to the epithelium side, and from the pocket's depth to the coronal surface, indicating differences in the physiological activity of the cells. Subsequently, we present three possible explanations: (i) the cell structures and morphologies of supragingival biofilms are much more diverse than those of subgingival biofilms; (ii) the unidentified microorganisms may belong to species for which there are no probes available (iii) prior stages of the biofilm that have been shielded from nutrients, comprising dead/inactive cells with low fluorescence activity, are located in the basal layers [26].
In healthy people, the fungal load is thought to be lower than the bacterial load. However, the size and morphology of fungal cells, as well as their synergistic interactions with bacteria, indicate that these species play an important role in the formation of dental biofilms [26,31,32].
The adsorbing of salivary proteins and glycoproteins to the tooth's surface triggers the initial plaque establishment, forming the acquired enamel pellicle (AEP)-a conditioning layer that covers all teeth present in the oral cavity and acts as a substrate for bacterial attachment [33,34]. The AEP is formed during the early stages of teeth eruption when saliva contacts the tooth's surface [34] and is never fully removed even if professional dental hygiene is performed.
Planktonic cells, aggregates of cells, and early colonizers (e.g., Streptococcus sp., Lactobacillus sp., Actinomyces sp., and Candida sp.) adhere to this pellicle via specialized adhesins on the bacterial cell surface [35,36]. These species do not promiscuously bind to any filament available, but rather engage in a highly specific interaction with already adhered cells, such as Corynebacterium sp., Actinomyces sp., [18], or yeast/hyphae cells forming the first layer of supragingival biofilm [26,31].
The role of Actinomyces sp. as a primary colonizer has already been proved due to its significant role in gingivitis [37]. Many Actinomyces species, such as A. naeslundii, A. oris, and A. johnsonii [38] have been found in supra-and subgingival biofilms' specimens from diseased patients, suggesting that these are the most significant biofilm initial formers among the Actinomyces genus. Actinomyces sp. can store intracellular glycogen or search for biofilm material such as extracellular polymeric substances and compounds from dead bacterial cells [26,39], which can create an important advantage in surviving in the deepest layers of the biofilm community. Actinomyces sp. were found in the base of the hedgehogs' structures [18], which also suggests that Corynebacterium sp. cells do not directly colonize the tooth surface but on a previous and established biofilm containing Actinomyces species.
Possibly, these already adhered microorganisms may create a microenvironment favorable to other colonizers to grow. Therefore, biofilm maturation occurs by the coaggregation of planktonic bacteria to an already adhered biofilm [40].
Early colonizers of supragingival dental surfaces are usually facultative anaerobic bacteria, such as Streptococcus sp. These species produce carbon dioxide (CO 2 ), lactate, and acetate, containing hydrogen peroxide (H 2 O 2 ) by consuming oxygen (O 2 ). The redox potential is lowered, allowing strict anaerobes (e.g., Fusobacterium sp., Leptotrichia sp., and Capnocytophaga sp.) to settle and multiply in the biofilm.
The presence of Streptococcus sp. and bacteria from the Cytophaga-Flavobacterium-Bacteroides cluster (CFB-cluster) in the second layer of supragingival biofilm [26] could indicate a critical transition of a supragingival biofilm made of predominantly Grampositive saccharolytic bacteria to a Gram-negative proteolytic subgingival biofilm, which could be caused by nutrient availability (e.g., dietary sugars in the supragingival biofilm, or proteins from saliva and GCF in the subgingival biofilm). Additionally, the presence of Streptococcus sp. in the basal and second layers of biofilm also allows us to infer that these organisms can adapt to a wide variety of environments, settling first as early colonizers but with the ability to expand and prosper. Our hypothesis is summarized in Figure 7. inhibiting the growth of obligate anaerobes [42]. Therefore, the conventional periodontal treatment approaches that involve mechanical removal of biofilm, physical disruption of biofilm structure, or antibiotic therapies could be enhanced.

Figure 7.
Summary hypothesis for the interpretation of the gathered evidence. Microorganisms that thrive in the subgingival environment, in contrast to bacteria in supragingival biofilm, are cut off from high oxygen stress, salivary, and dietary nutrients, depending on gingival crevicular fluid (GCF) for nutrition. The adsorbing of salivary proteins and glycoproteins to the tooth's surface triggers the initial plaque establishment, forming the acquired enamel pellicle (AEP)a conditioning layer for bacterial attachment. In the supragingival biofilm, Corynebacterium sp. filaments bind to an existing biofilm containing Streptococcus sp., Actinomyces sp., yeast cells, and Lactobacillus sp. These species produce carbon dioxide (CO2), lactate, and acetate, containing hydrogen peroxide (H2O2) by consuming oxygen (O2). The redox potential is lowered, allowing strict anaerobes (such as Fusobacterium sp., Leptotrichia sp., and Capnocytophaga sp.) to settle and multiply in the CO2-requiring "halo" section. This environment is also favorable to the growth and development of microorganisms from the Cytophaga-Flavobacterium-Bacteroides cluster (CFB-cluster), especially, Prevotella sp. In the periphery, cells shape  The presence and abundance of Corynebacterium sp. in hedgehog structures in a bushlike skeleton, anchored from the presumed tooth surface [18,41] may also indicate that this genus plays a key role in the biofilm community, while its biofilm specificity indicates that it occupies a niche that is influenced by tooth surface and/or GCF properties.
By reducing the flow of GCF with the use of anti-inflammatory agents the outgrowth of proteolytic bacteria would be prevented. Another alternative would be to use oxygenating or redox agents to make the gingival crevice less anaerobic, selectively inhibiting the growth of obligate anaerobes [42]. Therefore, the conventional periodontal treatment approaches that involve mechanical removal of biofilm, physical disruption of biofilm structure, or antibiotic therapies could be enhanced.

Microbes' Herd Mentality Behavior
Taxa that are present primarily or exclusively in one site may provide clues to the distinctive features of the habitat and the role that those taxa contribute to the site but it's still unclear if pathogens found at the control sites are part of the local/commensal flora or are originated from adjacent periodontal lesion sulcus crevicular fluid [24].
Depending on which partner the microorganism was found coaggregated with, F. alocis shaped several conformations, especially with fusiform bacteria, which formed concentrical radial-orientated structures in mushroom-shaped biofilms or palisades structures when the coaggregation also occurred with eubacterial organisms [3]. When isolated, F. alocis formed test-tube brushes shapes around signal-free channels. This pattern and degree of organization can play a role in the events that occur during biofilm growth and maturation and be strongly linked to biofilm formation.
Our findings propose that other microorganisms adopted different spatial arrangments influenced by their peers in different biofilm areas and tissues. Test-tube brushes shapes were once again found in a complex mixture of cells containing T. forsythia, Campylobacter sp., P. micra, Fusobacterium sp., and Synergistetes group A in the outside layer of subgingival biofilm [26]. On the other hand, Synergistetes group A adopted a wide cigar-like shape in a palisade lining when found isolated, forming aggregates exclusively with themselves. The CFB-cluster also manifested different shapes, especially Prevotella sp. that formed micro-colonies in the top layer of subgingival biofilm but were observed as filamentous, rod-shaped, and coccoid on the intermediate layer.
Depending on which microorganisms they are coaggregated with, some microorganisms appear to "metamorphose". We call this process microbes' "herd mentality behavior" and we believe that might be an important advantage of periodontal pathogens. To verify our hypothesis, future research should focus on the spatial organization of periodontal microbiota. The studies should be performed considering sample collection and preparation that preserves as much spatial organization as possible, such as whole-mount preparations that permit the imaging of 3D structures [18].

Limitations
None of the included articles classified periodontal diseases according to current diagnostic criteria. In June 2018, the American Academy of Periodontology (AAP) released a series of reviews in partnership with the European Federation of Periodontology (EFP) [7,[43][44][45][46] suggesting significant updates in periodontal diagnose, such as combining chronic and aggressive periodontitis in unique-entity periodontitis with various phenotypes. Another major improvement was the association of the periodontal disease's classification with its progression rate and prognostic factors in a more accurate and reliable staging classification [47].
The results demonstrate that the available evidence relies on a limited number of observations. Assuming that these observable interactions are somehow representative of the periodontal microbiota in health and disease, is a leap of faith that needs confirmation in large-scale studies. Some images catch our eye, but in the absence of meticulous cell counting, it may contribute to an overestimation of microbes' population in biofilm specimens, contradicting quantitative findings. This raises the question of whether all FISH studies should involve quantitative cell counting or dynamic time lap observations. We did not include the identification and screening of gray literature in the scoping review process. However, all authors of the included publications in this review were requested to submit unpublished images and/or articles that matched the review's goal.

Conclusions and Future Perspectives
Supragingival biofilm seems to present a radially ordered multiple-taxon structure, backboned by facultative anaerobic long rods, such as Corynebacterium sp. and Lactobacillus sp., bordered by coccoid and filamentous cells. On the other hand, subgingival biofilm seems to present a more layered multiple-taxon structure from the tooth's surface to the pocket's epithelium. Anaerobic microorganisms dominate the subgingival environment, especially in the deepest portions. Outside layers with increased microbial diversity were found in both sub-and supragingival biofilms when compared with the inner layers, and the largest contrast was observed in the subgingival biofilm. Consumers and producers of certain metabolites tend to be spatially related; some microorganisms exhibited the ability to shape various structures influenced by their peers in different biofilm areas. We referred to this last phenomenon as "microbes' herd mentality behavior" and we believe that it may represent an imperative benefit of certain members of the periodontal microbiota.
In vitro studies or studies performed in animals provided the foundations for the molecular pathology reasoning discussed in this scoping review. We suggest that future research should focus on studying species networks in a more realistic and holistic setting, using complex community-like structures as a model.
In conclusion, and in response to the previous query, the FISH technique was able to detect interactions between microorganisms from the three domains of life, enabling the construction of a possible hypothesis for understanding the influence of these interactions in the establishment and development of periodontal diseases. However, the evidence about the structural composition and microorganisms' interactions in supra-and subgingival biofilms is scarce and no definitive conclusion can be drawn. The enhanced FISH-based techniques provide significant information about the spatial organization of a complex natural polymicrobial community, allowing researchers and dental medicine doctors to better understand how the individual taxa interact within a community in periodontal diseases, and how their relations compromise the whole assemblage.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/microorganisms9071504/s1: Figure S1: Multiple hedgehogs' structures adjacent to each other; Figure S2: A cauliflower structure in plaque; Table S1: Database search strategy; Table S2: Oligonucleotide probes' names and sequences, the microorganisms targeted, and the FISH conditions used in the studies included in this scoping review.