P. intermedia and the related P. nigrescens are known as periodontopathic bacteria and frequently isolated not only from various types of periodontal diseases [21–23] but infections in infants  and sinusitis lesions of dental origin . Previously, we reported that clinical strains of P. nigrescens  and P. intermedia  isolated from chronic periodontitis lesions produce viscous materials under planktonic growth conditions. These materials form dense meshwork structures on cell surfaces as illustrated by SEM. Chemical analyses of the viscous materials isolated from their culture supernatants revealed that they consisted of mostly neutral sugars with mannose constituting more than 80% of the polysaccharide. Purified EPS itself did not exhibit any pathogenicity or immunogenicity. However, whole cells of clinical strains capable of producing EPS, in comparison to EPS non-producing mutants, show a strong ability to induce abscess in mice [7, 8]. Interestingly, the virulence of the mutants to induce abscess in mice can be restored by co-application of the variants with the purified EPS . Therefore, EPS could represent a key component contributing to the virulence of P. intermedia and P. nigrescens. Our finding is consistent with the findings of P. aeruginosa as reported previously . In this regard, Deighton et al.  compared the virulence of slime-positive Staphylococcus epidermidis with that of a slime-negative strain in a mouse model of subcutaneous infection and showed that biofilm-positive strains produce significantly more abscesses and persisted longer in the infection.
To further evaluate the level of pathogenicity on the clinical strains of P. intermedia with EPS productivity, this study compared the ability of EPS-producing P. intermedia and several laboratory reference strains of periodontopathic bacteria (Pi25611, Pg33277, Pg381 and PgW83) to induce abscess formation in mice. EPS-producing Pi17 and PiOD1-16 induced abscess lesions in mice at 107 CFU, but other test periodontopathic bacteria did not when tested at this cell concentration. Pi25611 and P. gingivalis strains used in this study induced detectable abscess formation in mice when the infectious dose was 109 CFU and higher. The abscess model in mice with inoculum sizes of 109-1010 CFU has been used to demonstrate the biological activities of P. gingivalis [27, 28]. Accordingly, the pathogenicity of Pi17 and PiOD1-16 appeared to be stronger than those of the P. gingivalis strains as well as the ATCC strain of P. intermedia used in this study.
A wide range of microorganisms is known to produce EPS as a main constituent of the biofilm extracellular matrix, and recent investigations have revealed that each biofilm-forming bacterium produces distinctive EPS components . In oral microbiota, Capnocytophaga ochracea found in the human oral cavity has been shown to produce mannose-rich EPS that can suppress murine lymphocyte mitogen responses and activate human complement response [30–32]. Kapran et al.  reported that Aggregatibacter actinomycetemcomitans has a gene cluster which is homologous to E. coli pgaABCD and encodes the production of poly-ß-1,6-GlcNAc (PGA) . Rothia mucilaginosa DY-18  and Escherichia hermannii YS-11  isolated from persistent apical periodontitis lesions produced EPS and exhibited cell surface meshwork structures. The meshwork structures of E. hermannii YS-11 disappeared when wzt, one of the ABC-transporter genes, was disrupted by transposon random insertion mutagenesis. Complementation of this gene to the transposant restored and dramatically augmented the formation of meshwork structures. Our studies using an abscess model in mice indicated that this EPS phenotype might be involved in the pathogenicity of this organism . Likewise, as described above, EPS productivity could be associated with P. intermedia and P. nigrescens pathogenicity [7, 8].
In our experience, more than 20% of clinically isolated P. intermedia strains showed viscous material productivity under planktonic growth conditions . This ability was lost in the course of sequential in vitro passage. As a result, less than 2% of clinical isolates remained as viscous material-producing strains (data not shown). Therefore, it is important to note that laboratory reference strains do not always represent the original virulence properties as Fux et al.  previously pointed out. Early studies have pointed out the relation between bacterial pathogenicity and polysaccharide productivity on the reference strains used in this study. Okuda et al.  reported that Pi25611, Pg381 and Pg33277 had capsular structures around the cells and that the capsular polysaccharides extracted from Pg381 contained galactose and glucose as their major constituents. PgW83 is known to produce capsular polysaccharides, and the genetic locus for capsule biosynthesis has been identified [39, 40]. As discussed in these earlier studies [38, 41–44], cell surface-associated polysaccharides could represent another set of virulence factors in addition to production of various proteolytic enzymes, contributing to the pathogenicity of P. gingivalis strains. In this study, we did not detect the presence of capsular polysaccharide or production of EPS in P. gingivalis strains. One possibility is that the tested P. gingivalis strains had lost their ability to produce capsular polysaccharides or EPS because of multiple in vitro passages of the organisms in the laboratory. Further, none of these strains regained capsular polysaccharides or EPS productivity through repetitive animal passages (data not shown). This could explain the less virulent characteristics displayed by the tested P. gingivalis strains shown in our animal virulence experiments.
As our [7, 8] and other earlier studies [26, 45] indicated, it is plausible that the antiphagocytic effect of EPS confers the ability to P. intermedia to induce abscess. It has been demonstrated that the slime or components of slime from S. epidermidis cultures could contribute to the delay of clearance of this organism from host tissues. Similarly, in the murine model of systemic infection, the deletion of ica locus necessary for the biosynthesis of surface polysaccharide of Staphylococcus aureus significantly reduces its virulence . As described above, a study in the early 1970s clearly showed that the addition of slime from P. aeruginosa cultures to E. coli or S. aureus dramatically inhibited phagocytosis by leukocytes . In this study, EPS-producing Pi17 and PiOD1-16 cells were rarely internalized by leukocytes both in vitro and in vivo. Many of these bacteria were seen localized to the cell surface of PMNLs but failed to be ingested. In such a situation, PMNLs are known to produce a harmful effect to the surrounding tissue by elaborating a variety of degradative enzymes and oxygen radicals in an attempt to clear the invaded pathogens [46, 47]. In contrast, the test laboratory reference strains of periodontopathic bacteria, lacking EPS production, were readily engulfed and digested in phagosomes of phagocytes.
Most pathogenic P. gingivalis strains exhibit a higher resistance to phagocytosis than less pathogenic strains do . In our study, the P. gingivalis strains were readily phagocytosed. It has been documented that freshly isolated pathogenic strains of P. gingivalis could lose invasiveness as a result of repeated subcultures . Therefore, it is reasonable to speculate that pathogenic P. gingivalis strains become less pathogenic and more susceptible to phagocytosis by PMNLs when they lose the ability to produce capsular or extracellular polysaccharides though we have to carefully investigate the possibility that multiple factors exist in the observed incapability to induce abscesses in mice. We have not been able to restore any of our working strains of P. gingivalis' ability to express cell surface-associated meshwork structures or the ability to spontaneously produce viscous material in spent culture media. It is still unclear whether P. gingivalis with capsule formation or EPS productivity exhibits similar or higher pathogenicity to those of P. intermedia strains with meshwork structures.