Dental Infection of Porphyromonas gingivalis Induces Preterm Birth in Mice

Abstract

Background: Epidemiological studies have revealed a link between dental infection and preterm birth or low birth weight (PTB/LBW), however, the underlying mechanisms remain unclear. Progress in understanding the associated mechanisms has been limited in part by lack of an animal model for chronic infection-induced PTB/LBW, mimicking pregnancy under conditions of periodontitis. We aimed to establish a mouse model of chronic periodontitis in order to investigate the link between periodontitis and PTB/LBW.

Methods: To establish chronic inflammation beginning with dental infection, we surgically opened mouse (female, 8 weeks old) 1st molar pulp chambers and directly infected with w83 strain Porphyromonas gingivalis (P.g.), a keystone periodontal pathogen. Mating was initiated at 6 wks post-infection, by which time dental granuloma tissue had developed and live P.g. was cultured from extracted tooth root, which serves as a persistent source of P.g. The gestational day (gd) and birth weight were recorded during for P.g.-infected and control mice, and serum and placental tissues were collected at gd 15 to evaluate the systemic and local conditions during pregnancy.

Results: Dental infection with P.g. significantly increased circulating TNF-α (2.5-fold), IL-17 (2-fold), IL-6 (2-fold) and IL-1β (2-fold). The P.g.-infected group delivered at gd 18.25 vs. gd 20.45 in the non-infected control (NC) group (p < 0.01), and pups exhibited LBW compared to controls (p < 0.01). P.g. was localized to placental tissues by immunohistochemistry and PCR, and defects in placental tissues of P.g. infected mice included premature rupture of membrane, placental detachment, degenerative changes in trophoblasts and endothelial cells, including necrotic areas. P.g. infection caused significantly increased numbers of polymorphonuclear leukocytes (PMNLs) and macrophages in placental tissues, associated with increased local expression of pro-inflammatory mediators including TNF-α and COX-2. Further placental tissue damage was indicated in P.g. infected mice by decreased CD-31 in endothelial cells, increased expression of 8OHdG, an indicator of oxidative DNA damage, and cleaved caspase-3, a marker of apoptosis. In vitro, P.g. lipopolysaccharide significantly increased expression of COX-2, IL-8 and TNF-α, in HTR-8 trophoblasts in an NF-κB-dependent fashion.

Conclusions: Our novel mouse model supports previous epidemiological studies signifying dental infection as predisposing factor for PTB/LBW. We demonstrate PTB and LBW in infected mice, translocation of P.g to placental tissues, increased circulating and local pro-inflammatory markers, and the capability of P.g. LPS to directly induce cytokine production in trophoblasts, in vitro. These findings further underscore the importance of local and systemic infections and inflammation during pregnancy and suggest that prevention and/or elimination of dental infections such as marginal or periapical periodontitis before pregnancy may have a beneficial effect on PTB/LBW.

Fig 1. Dental Infection of P.g. Increases Pro-inflammatory Cytokines in Serum.

ELISA analysis of TNF-α (A), IL-1β (B), IL-6 (C) and IL-17 (D) concentrations in serum (0wk or 6wks post-infection) of non-infected control (NC) and P.g.-infected groups (n = 6 for each). Statistical significance was determined using the Student t-test. **P<0.01.

Fig 2. Dental Infection of P.g. Induces PTB and LBW.

Maternal body weight changes (A) were recorded from P.g. infection (0 wks) to initiation of mating (6 wks post-infection). Gestational day (gd) (B) was confirmed in NC and P.g.-infected groups (n = 20 for each group). Means (±SD) are 20.45 ± 0.56 and 18.25 ± 0.925, respectively. Statistical significance was determined using the unpaired t-test. Birthweight of neonates (C-H) was compared between NC and P.g.-infected groups, considering all neonates or comparing litters with the same number of neonates (to control for litter size effect on birthweight). Statistical significance was determined using the unpaired t-test. SGA (small for gestational age) (I) was defined as pups birth weight less than the 10th percentile in NC group, which is 1.27 g (indicated by dotted line). 13 pups out of 30 pups in P.g.-infected groups were SGA. Statistical analysis for SGA is shown in Table 1. ** p < 0.01. *p < 0.05.

Fig 3. Dental Infection of P.g. Induces Defects in Placental Tissues in Mice.

Representative histological findings in gd 15-placental tissue from the NC group (A-D) and P.g.-infected group (E-H) by H&E staining (n = 6 for each group). (A, E) Overview of mouse placenta: La, Labyrinth; Sp, Spongiotrophoblasts; De, Decidua; Ut, Uterus wall. (B, F) Epithelium of amnion: in the P.g.-infected group, amnion epithelium (black arrow) is degenerative and detached from chorionic plate (white arrow). (C, G) Labyrins and Spongiotrophoblasts layers: Trophoblasts and endothelial cells are necrotic (*) in P.g.-infected group. (D, H) Decidua and uterus wall: Placental abruption is evident in the P.g.-infected group. ** shows the separation between placenta and uterus at maternal-fetus junction. (I) Immunolocalization of P.g. (brown pigments) (n = 6 for each group); 1000x-magnification by oil-immersed microscopy. (J) Gene expression for the mgl-gene of P.g-W83 strain by nested PCR; representative results from 6 mice for each group. PC, positive control. Scale bars, 100μm. Olympus BH2 microscope and Nikon digital sight DS-L2 camera were used for capturing images.

Fig 4. Increased Inflammatory Cell Infiltration and Inflammatory Mediators in Placental Tissues Following P.g. Infection.

Immunohistochemical staining of gd 15 placental tissues of NC and P.g.-infected groups (n = 6 for each group). (A,B) PMNLs, (C,D) F4/80 positive macrophages, (E,F) TNF-α positive macrophages, (G, H) COX-2 positive staining, (I, J) CD-31 stained endothelial cells, (K,L) 8OHdG (oxidative DNA stress marker), (M,N) cleaved-caspase 3 (apoptosis marker), and (O,P) histomorphometric analysis of PMNLs and F4/80 (macrophages). Scale bars, 100μm. Statistical significance was determined using the Student t-test. *p < 0.05, **p < 0.01. Olympus BH2 microscope and Nikon digital sight DS-L2 camera were used for capturing images. (Q) Representative results of mRNA expression of COX-2, Gal-3, IL-8 and TNF-α in placental tissues in NC group and P.g.-infected group (n = 3 for each), GAPDH was used as internal control. Molecular weights were labeled to the right of each band. Yellow dashed square shows that PMNLs are restricted to the blood vessels.

Fig 5. P.g. LPS Up-regulates Expression of Inflammatory Mediators via NF-κB Signaling in Trophoblasts.

(A) HTR-8 trophoblasts were seeded (5x10⁵ cells/well) in 6-well culture plates and culture media were changed once before stimulation. Cells were treated by P.g.-LPS (1 μg/ml unless otherwise noted) and both culture medium and cells were collected. mRNA expressions of COX-2, IL-8 and TNF-α were analyzed from cell pellets. (B) HTR-8 cells were stimulated with P.g.-LPS, Li-P.g. (live-P.g.) or D-P.g. (dead-P.g.) and cell lysates were examined by immunoblotting analysis using p-p65, p65. Molecular weight is labeled to the right of each band. (C,D) HTR-8 cells were pretreated with or without CAPE for 4 hrs and then stimulated with or without P.g.-LPS, Li-P.g. (live-P.g.) or D-P.g. (dead-P.g.) for 24 hrs. mRNA expressions of COX-2, IL-8 and TNF-α (C); protein secretion of TNF-α were examined (D). TNF-α amount in negative controls were subtracted as basal levels when calculating the percentage of down-regulation. GAPDH or β-actin was used as internal control. MOI, multiplicity of infection. *p < 0.05, **p < 0.01. Experiments were performed at least three times with similar results.