J Clin Periodontol 2009; 36: 406–410 doi: 10.1111/j.1600-051X.2009.01393.x

David Polak1, Asaf Wilensky1, Mouse model of experimental Lior Shapira1, Amal Halabi1, Dita Goldstein2, Ervin I. Weiss2 and Yael Houri-Haddad2 periodontitis induced by Departments of1Periodontology and 2Prosthodontics, Faculty of Dental Medicine, The Hebrew University-Hadassah, Medical / Center, Jerusalem, Israel Fusobacteriumnucleatuminfection: bone loss and host response

Polak D, Wilensky A, Shapira L, Halabi A, Goldstein D, Weiss EI, Houri-Haddad Y. Mouse model of experimental periodontitis induced by Porphyromonas gingivalis/ nucleatum infection: bone loss and host response. J Clin Periodontol 2009; 36: 406–410. doi: 10.1111/j.1600-051X.2009.01393.x.

Abstract Aim: To compare the effect of oral infection with Porphyromonas gingivalis or Fusobacterium nucleatum versus infection with both on mouse periodontal tissues, and to characterize the inflammatory response. Materials and Methods: Mice were orally infected with P. gingivalis, F. nucleatum or both. At 42 days post-infection, alveolar bone loss was quantified using micro-computerized tomography. Tumour necrosis factor-a (TNF-a) and interleukin (IL)-1b levels induced by the infection were quantified using the subcutaneous chamber model. Results: Mice orally infected with F. nucleatum/P. gingivalis exhibited significantly more bone loss compared with that of mono-infected and sham-infected mice. F. nucleatum/P. gingivalis infection also increased the levels of TNF-a and IL1b compared with the levels found in the mono-infected groups. Conclusions: Polymicrobial infection with P. gingivalis/F. nucleatum aggravates alveolar bone loss and induces a stronger inflammatory response compared with that Key words: co-aggregation; F. nucleatum; inflammatory response; mixed infection; observed upon infection with either bacterium alone. The results suggest that oral P. gingivalis infection of mice with a mixture of P. gingivalis and F. nucleatum may be superior to mono-infection models of experimental periodontitis. Accepted for publication 4 February 2009

Conflict of interest and sources of is characterized by tory studies (Holt & Bramanti 1991, funding statement an inflammatory process culminating in Socransky & Haffajee 1992, Okuda The authors declare that they have no the destruction of the dental attachment et al. 1994, Haffajee & Socransky conflict of interests. apparatus. A growing body of evidence 2005, Holt & Ebersole 2005). This bac- The present work is part of a PhD thesis by supports the notion that interactions, terium has many virulence factors that David Polak. The research was carried such as co-aggregation between differ- enable it to manipulate the host-inflam- out at the MIS-supported Laboratory for ent bacterial species, are important fac- matory response (Cutler et al. 1995). Periodontal Research and The Ronald E. tors in periodontal disease pathogenesis Fusobacterium nucleatum is also com- Goldstein Center for Esthetic Dentistry (Holt & Bramanti 1991, Haffajee & monly cultivated from the subgingival and Dental Materials Research, The Hebrew University-Hadassah Faculty of Socransky 1994). plaque of inflamed and Dental Medicine, Jerusalem, Israel. The Porphyromonas gingivalis is a key from periodontal pockets (Haffajee & study was supported by a grant from the microorganism in periodontitis, and its Socransky 1994), and participates in Chief Scientist of the Israel Ministry of damage to the has been both adhesion and co-aggregation reac- Health. documented in many clinical and labora- tions with other microbial pathogens, 406 r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S Mixed infection augments periodontitis in mice 407 such as P. gingivalis (Bradshaw et al. brander & Andersen 1989, Genco et al. Data analysis. The animal studies were 1998, Weiss et al. 2000). 1991). carried out with at least six mice per Many investigations have been treatment group. The data were analysed designed to study the virulence of sev- using a statistical software package (Sig- eral oral pathogens and their role in the Oral infection (Baker et al. 1994). Four- maStat, Jandel Scientific, San Rafael, aetiology of periodontal disease. Most to five-week-old Balb/C mice were super- CA, USA). One-way repeated measure of them involved mono-infection animal infected three times at 2-day intervals analysis of variance (RM ANOVA) was models (Kesavalu et al. 1992, Genco with P. gingivalis ( 400 mlofa109 bac- used to test the significance of the dif- et al. 1998, Baker et al. 2000). Several teria/ml suspension for each mouse), F. ferences between the treated groups. If studies have shown that co-infection nucleatum ( 400 mlofa109 bacteria/ml the results were significant, inter-group with different pathogens modulates the suspension for each mouse) or a mixture differences were tested for significance immune response and affects the clinical of both ( 400 mlofa109 bacteria/ml using Student’s t-test and the Bonferroni outcome of the infection in various suspension made up of equal volumes correction for multiple testing. diseases (Vitovec et al. 2001, Stoicov of each bacterium), and 42 days after the et al. 2004). Using the mouse skin last gavage the maxillary jaws were abscess model, Feuille et al. (1996) harvested. demonstrated a synergistic effect on Results soft tissue destruction by mixed infec- P. gingivalis and F. nucleatum tion with P. gingivalis and F. nuclea- Micro-computerized tomography (CT) co-aggregate in vitro and in vivo tum. Destruction of the soft tissue analysis. Maxillary hemi-jaws were continued long after the bacterial infec- analysed by compact fan-beam-type The synergistic effect between two bac- tion with P. gingivalis and F. nucleatum CT (mCT 40, Scanco Medical, Bassers- teria may depend on their ability to was eradicated (Ebersole et al. 1997). dorf, Switzerland) as described pre- co-aggregate in vivo. First we examined More recently, a study on rats, using the viously (Wilensky et al. 2005). the ability of P. gingivalis and F. nucle- oral infection model, has shown that The subcutaneous chamber model was atum to co-aggregate in vitro. The level oral infection with P. gingivalis, Trepo- used to evaluate the local inflammatory of co-aggregation in vitro between nema denticola, response to bacterial challenge as P. gingivalis and F. nucleatum was and F. nucleatum together induced described previously (Houri-Haddad 2–3 on a scale of 0–4. To ascertain that robust alveolar bone loss compared et al. 2000). Four- to five-week-old the two bacteria can also co-aggregate in vivo, a mixture of both bacteria with that obtained upon infection by Balb/C mice were challenged by an 8 either bacterium alone (Kesavalu et al. injection of 100 ml of PBS containing (10 bacteria) was injected into subcuta- 2007). The bone loss did not differ when P. gingivalis (106 bacteria) or F. nucle- neous chambers in mice (Houri-Haddad F. nucleatum was excluded from the atum (106 bacteria) or both (a total of et al. 2000). The chambers were sampled infection. 106 bacteria containing 5 Â 105 of each 2 and 24 h post-infection, and examined We hypothesize that the polymicro- bacterium). Chamber exudates were under phase microscopy (Â 1000). bial infection enhances virulence, both harvested at baseline (before bacterial Co-aggregation of P. gingivalis and accelerating and aggravating bone loss, injection), and 2 and 24 h post-challenge F. nucleatum was evident at the two compared with that occurring upon for cytokine analysis. tested time intervals (Fig. 1). infection with one bacterium. Our objectives were to determine the effect of oral infection with P. gingivalis, Enzyme-linked immunosorbant assay P. gingivalis/F. nucleatum infection induces alveolar bone loss F. nucleatum or both on the mouse (ELISA) of cytokines. The secreted periodontal tissues, at the molecular forms of mouse tumour necrosis fac- We tested the ability of the two bacteria and clinical levels. tor-a (TNF-a) and interleukin (IL)-1b to induce experimental periodontitis in were quantified using two-site-ELISA, mice. Mice were orally infected with the method based on commercially P. gingivalis, F. nucleatum or both, Materials and Methods available antibody pairs (Pharmingen, while control mice received vehicle San Diego, CA, USA), as described (PBS) only. Under the present experi- All experiments were performed in the previously (Houri-Haddad et al. 2000). mental conditions, mice infected with SPF unit of The Hebrew University- The range of detection for each specific P. gingivalis or F. nucleatum alone did Hadassah Medical Center, and approved cytokine was 25–2000 pg/ml. not exhibit significant alveolar bone by the university’s Animal Care and Use loss, which was essentially similar to Committee. that in the non-infected animals (Fig. 2). Co-aggregation assay. A suspension of However, mixed infection with P. gin- Bacterial cultivation. P. gingivalis strain F. nucleatum and P. gingivalis (50ml givalis/F. nucleatum induced significant ATCC 33277 and F. nucleatum strain PK each) was mixed vigorously for 10 s alveolar bone loss, compared with that 1594 were grown in Wilkins broth and a visual rating scale of 0–4 was found in the mono-infected groups and (Chalgren broth, Oxoid Ltd, Cambridge, used to grade the reaction. A score of 0 the sham-infected control group. The UK), in an anaerobic chamber with 85% signified an evenly turbid suspension residual bone volume (as measured by N2,5%H2 and 10% CO2, followed by with no visible aggregates, i.e., no co- volumetric micro-CT analysis) in the three washes in phosphate-buffered saline aggregation, and a score of 4 signified mixed-infection group was  70% of (PBS). The bacterial concentration was maximal clumping, leaving a clear the bone level found in the other three measured spectrophotometrically (Kolen- supernatant (Weiss et al. 2000). groups (Fig. 2). r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S 408 Polak et al.

X4000

X1000

Fig. 1. Light microscopy of chamber exu- Fig. 3. Inflammatory cytokine response. Mice (n 5 6 in each group) received an intra- dates following mixed infection with Por- chamber challenge of Porphyromonas gingivalis, Fusobacterium nucleatum or a mixture of phyromonas gingivalis and Fusobacterium the two. The control group was challenged with saline alone. Chamber exudates were nucleatum. Mice received an intra-chamber obtained at baseline (before challenge), and at 2 and 24 h post-infection. The levels of each challenge of a mixture of P. gingivalis (108 cytokine in the chamber exudates were determined by an enzyme-linked immunosorbant 8 assay. The results are expressed as the mean Æ standard error. #, a significant difference from bacteria) and F. nucleatum (10 bacteria). n Chamber exudates were harvested at 24 h the control group; , a significant difference from the other groups. po0.05. (a) Tumour post-infection and examined under phase necrosis factor-a levels induced by mono versus mixed infection. (b) Interleukin-1b levels microscopy at  1000 and  4000 magni- induced by sub-clinical mono versus mixed infection. fication.

F. nucleatum or P. gingivalis/F. nucle- periodontal disease, and the possibility atum, compared with almost undetect- of a synergistic effect between the bac- able levels in the control group. The teria, a mouse model of experimental TNF-a levels induced by infection with periodontits using more than one bacter- P. gingivalis alone were not different ium could further our understanding of from that found in the control mice disease pathogenesis. To test this notion, (Fig. 3a, po0.05). Although the levels we focused on a mouse model of mixed of TNF-a were higher in the P. gingi- infection with two prevalent periodontal valis/F. nucleatum group than that in the pathogens: P. gingivalis and F. nucle- F. nucleatum group, the difference was atum. These two bacteria were found to not statistically significant. At 24 h post- co-aggregate in vitro and in vivo. To test infection, there was a decline in TNF-a our hypothesis, we compared the clin- levels in all the tested groups, with no ical outcome of an oral infection with a Fig. 2. Evaluation of residual bone loss difference between the mono-infection mixture of the two bacteria to infection induced by oral infection with Porphyromo- and the mixed-infection groups. with either bacterium alone. Mono- nas gingivalis (P. gingivalis), Fusobacter- infection with P. gingivalis or F. nucle- ium nucleatum (F. nucleatum) or both. Mice (n 5 8 in each group) were challenged orally IL-1b levels atum did not induce significant bone three times at 2-day intervals with an inno- loss compared with that of the control. culum of P. gingivalis, F. nucleatum or a As found for TNF-a, 2 h post-challenge Although P. gingivalis was previously mixture of the two. Sham-infected mice high levels of IL-1b were induced by shown to induce bone loss in a similar served as control. Six weeks later, the jaws infection with F. nucleatum or P. gingi- model (Baker et al. 1994, Wilensky were harvested and the alveolar bone valis/F. nucleatum, compared with et al. 2005), the bacterial concentration volume was measured using micro-compu- almost undetectable levels in the control in the previous studies was 10-fold high- terized tomography. The results are group. The IL-1b level induced by expressed as the mean Æ standard error. n, er than the concentration used in the infection with P. gingivalis alone was present one. In addition, more aggres- a significant difference from the other not different from that found in the groups. po0.05. sive strains of P. gingivalis (ATCC control mice, and was significantly low- 52977, 381 and W50) were used in the er than that in the other two infected previous studies. Another possible groups (Fig. 3b, po0.05). At 24 h post- explanation for the lack of differences Mixed infection augments the pro- infection this trend was stable, although inflammatory cytokine response between the mono-infected and the con- only the difference between the mixed trol groups may be the small sample size infection and the P. gingivalis groups To evaluate the early local cytokine of each group (n 5 8), but this possibi- was statistically significant (p 0.05). response, mice were infected in pre- o lity seems less likely. In our animal implanted subcutaneous chambers with model, although each bacterium alone P. gingivalis, F. nucleatum or both. did not induce alveolar bone loss, the Saline was injected in the control group. Discussion mixed oral infection, using the same Most of the rodent models of experi- concentration as in the mono-infection mental periodontis involve oral infec- groups, resulted in significant bone loss TNF-a levels tion with a single bacterium or ligature in comparison with that observed in the Two hours post-challenge, high levels of placement around a single tooth. mono- and sham-infected groups. This TNF-a were induced by infection with Because of the polymicrobial nature of could be a result of synergism in viru- r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S Mixed infection augments periodontitis in mice 409 lence between the two bacteria leading same model, tissue destruction contin- bial communities during aeration. Infection to significant induction of periodontal ued for 7 days after the bacterial chal- and Immunity 66, 4729–4732. breakdown, or that the bacterial mixture lenge, while the bacterial presence in the Cutler, C. W., Kalmar, J. R. & Genco, C. A. results in greater survival than the infected sites was detectable only up to (1995) Pathogenic strategies of the oral anae- monocultures. A study of oral infection 2 days post-challenge (Ebersole et al. robe, Porphyromonas gingivalis. Trends in in rats showed that mixed infection with 1997). This suggests that tissue destruc- Microbiology 3, 45–51. Ebersole, J. L., Feuille, F., Kesavalu, L. & Holt, P. gingivalis, T. denticola, T. forsythia tion continues long after the periopatho- S. C. (1997) Host modulation of tissue and F. nucleatum induced significant gen infection is eradicated, and that the destruction caused by periodontopathogens: bone loss compared with that following destruction process is mediated by the effects on a mixed microbial infection com- mono-infection with each bacterium host response. It cannot be excluded that posed of Porphyromonas gingivalis and alone (Kesavalu et al. 2007). In their the tested cytokine levels found in the Fusobacterium nucleatum. Microbial Patho- study, the addition of F. nucleatum to P. gingivalis mono-infected group and genesis 23, 23–32. the mixed infection did not change bone the mixed-infection group may also be Feuille, F., Ebersole, J. L., Kesavalu, L., Stepfen, resorption. A model using mixed infec- affected by the presence of P. gingiva- M. J. & Holt, S. C. (1996) Mixed in- tion with two bacteria may result in lis’s proteases. Moreover, it is possible fection with Porphyromonas gingivalis and different bacterial interactions than that the presence of F. nucleatum along Fusobacterium nucleatum in a murine lesion model: potential synergistic effects on viru- mixed infection with four bacteria. with P. gingivalis may interfere to some lence. Infection and Immunity 64, 2094– Furthermore, we previously showed extent with the proteolysis activity of 2100. that the micro-CT technology is more gingipains and may also affect the Genco, C. A., Cutler, C. W., Kapczynski, D., sensitive than the two-dimensional detected cytokine levels. Maloney, K. & Arnold, R. R. (1991) A novel method used by Kesavalu et al. (2007) It is generally accepted that perio- mouse model to study the virulence of and (Wilensky et al. 2005). dontal disease is a result of more than host response to Porphyromonas (Bacter- We hypothesized that the dissimilar- one infective organism, albeit the inter- oides) gingivalis. Infection and Immunity ity in disease expression between mono- action between the 59, 1255–1263. and mixed infection is correlated with is not clear, nor is the impact on disease Genco, C. A., Van Dyke, T. & Amar, S. (1998) differences in host response at the mole- outcome. The results of the present Animal models for Porphyromonas gingiva- cular level. To test this assumption, we study clearly show that the mixed in- lis-mediated periodontal disease. Trends in Microbiology 6, 444–449. used the subcutaneous chamber model, fection was more destructive than Haffajee, A. D. & Socransky, S. S. (1994) which enabled us to characterize the mono-infection in our mouse model. Microbial etiological agents of destructive inflammatory response to a specific bac- The nature of the synergistic destruction periodontal diseases. 2000 terial infection. We are not suggesting of the mixed infection may be at least 5, 78–111. that the chamber model is a model for partly attributable to the inter-genera Haffajee, A. D. & Socransky, S. S. (2005) periodontal infection, but a well-estab- interaction. However, the impact of the Microbiology of periodontal diseases: intro- lished accepted model for quantification interaction between bacteria in a mixed duction. Periodontology 2000 38, 9–12. of the local host response to bacterial infection on host response merits further Holt, S. C. & Bramanti, T. E. (1991) Factors in infection. A comparison of the inflam- investigation. The present study sug- virulence expression and their role in perio- matory response induced by mixed gests that the oral mixed infection model dontal disease pathogenesis. Critical Reviews in Oral Biology and Medicine 2, 177–281. infection versus mono-infection showed with P. gingivalis and F. nucleatum may Holt, S. C. & Ebersole, J. L. (2005) Porphyr- that the levels of IL-1b and TNF-a were serve as a rodent model for perio- omonas gingivalis, denticola, and augmented in the mixed-infection dontitis. Tannerella forsythia: the ‘‘’’, a group. The difference between the prototype polybacterial pathogenic consor- P. gingivalis group and the mixed- tium in periodontitis. Periodontology 2000 infection group was significant, whereas 38, 72–122. Houri-Haddad, Y., Soskolne, W. A., Halabi, A., the difference between the F. nucleatum References group and the mixed-infection group Barak, V. & Shapira, L. (2000) Repeat bac- was not. The lack of a statistically Baker, P. J., Dixon, M., Evans, R. T. & Roope- terial challenge in a subcutaneous chamber significant difference between the nian, D. C. (2000) Heterogeneity of Porphyr- model results in augmented tumour necrosis factor-alpha and interferon-gamma response, F. nucleatum group and the mixed- omonas gingivalis strains in the induction of alveolar bone loss in mice. Oral Microbiol- and suppression of interleukin-10. Immunol- infection group might be due to the ogy 99, 215–220. limited number of animals per group ogy and Immunology 15, 27–32. Baker, P. J., Evans, R. T. & Roopenian, D. C. Kesavalu, L., Ebersole, J. L., Machen, R. L. & (n 5 6) used. (1994) Oral infection with Porphyromonas Holt, S. C. (1992) Porphyromonas gingivalis IL-1b and TNF-a are two cytokines gingivalis and induced alveolar bone loss in virulence in mice: induction of immunity to that are well recognized as tissue- immunocompetent and severe combined bacterial components. Infection and Immunity destructive mediators (Bascones et al. immunodeficient mice. Archives of Oral Biol- 60, 1455–1464. 2005), and their high levels in the mixed ogy 39, 1035–1040. Kesavalu, L., Sathishkumar, S., Bakthavatchalu, infection group may explain the results Bascones, A., Noronha, S., Gomez, M., Mota, V., Matthews, C., Dawson, D., Steffen, M. & of the oral infection. These findings P., Gonzalez Moles, M. A. & Dorrego, M. V. Ebersole, J. L. (2007) Rat model of polymi- concur with the results of a previous (2005) Tissue destruction in periodontitis: crobial infection, immunity, and alveolar bacteria or cytokines fault? Quintessence bone resorption in periodontal disease. Infec- study that showed that mixed infection International 36, 299–306. tion and Immunity 75, 1704–1712. with P. gingivalis and F. nucleatum Bradshaw, D. J., Marsh, P. D., Watson, G. K. & Kolenbrander, P. E. & Andersen, R. N. (1989) induces a synergistic destructive effect Allison, C. (1998) Role of Fusobacterium Inhibition of coaggregation between Fuso- on soft tissue in the murine abscess nucleatum and coaggregation in anaerobe bacterium nucleatum and Porphyromonas model (Feuille et al. 1996). In the survival in planktonic and biofilm oral micro- (Bacteroides) gingivalis by lactose and r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S 410 Polak et al.

related sugars. Infection and Immunity 57, oplasma gondii infections. Journal of Wilensky, A., Gabet, Y., Yumoto, H., 3204–3209. Immunology 173, 3329–3336. Houri-Haddad, Y. & Shapira, L. (2005) Okuda, K. S., Hirai, T., Harano, K. & Kato, T. K. Vitovec, J., Aldova, E., Vladik, P. & Krovacek, Three-dimensional quantification of alveolar (1994) Role of antibodies in periodontopathic K. (2001) Enteropathogenicity of Plesiomo- bone loss in Porphyromonas gingivalis- bacterial infection. In: Genco, R., Hamada, S., nas shigelloides and Aeromonas spp. in infected mice using micro-computed Lehner, T., McGhee, J. & Mergenhagen, S. experimental mono- and coinfection with tomography. Journal of Periodontology 76, (eds) Molecular Pathogenesis of Periodontal Cryptosporidium parvum in the intestine of 1282–1286. Disease, 1st edition, pp. 257–265, Washing- neonatal BALB/c mice. Comparative Immu- ton DC: ASM press. nology Microbiology and Infectious Diseases Socransky, S. S. & Haffajee, A. D. (1992) The 24, 39–55. bacterial etiology of destructive periodontal Address: Weiss, E. I., Shaniztki, B., Dotan, M., Ganesh- disease: current concepts. Journal of Perio- David Polak. dontology 63, 322–331. kumar, N., Kolenbrander, P. E. & Metzger, Z. Department of Periodontics Stoicov, C., Whary, M., Rogers, A. B., Lee, F. (2000) Attachment of Fusobacterium nucle- School of Dental Medicine S., Klucevsek, K., Li, H., Cai, X., Saffari, R., atum PK1594 to mammalian cells and its The Hebrew University-Hadassah Ge, Z., Khan, I. A., Combe, C., Luster, A., coaggregation with periodontopathogenic PO Box 12272 Fox, J. G. & Houghton, J. (2004) Coinfection bacteria are mediated by the same galac- Jerusalem 91120 modulates inflammatory responses and clin- tose-binding adhesin. Oral Microbiology Israel ical outcome of Helicobacter felis and Tox- and Immunology 15, 371–377. E-mail: [email protected]

Clinical Relevance Principal findings: Polymicrobial experimental periodontitis by P. gin- Scientific rationale for the study:Most infection with P. gingivalis and givalis and F. nucleatum. (2) The rodent models for periodontal disease F. nucleatum in mice resulted in P. gingivalis/F. nucleatum experi- use mono-infection despite the polymi- augmented alveolar bone loss and a mental periodontitis model is suita- crobial nature of the disease. Our aim robust inflammatory response, com- ble for future studies of disease was to establish an animal model of pared with that obtained upon infec- pathogenesis and novel prevention poly-microbial infection for the study tion with one bacterium alone. approaches. of periodontal disease pathogenesis, Practical implication: (1) There is a which would better mimic the disease. synergistic effect in the induction of

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S