Unculturable and culturable periodontal-related bacteria are associated with periodontal inflammation during pregnancy and with preterm low birth weight delivery

Recent studies revealed culturable periodontal keystone pathogens are associated with preterm low birth weight (PLBW). However, the oral microbiome is also comprised of hundreds of ‘culture-difficult’ or ‘not-yet-culturable’ bacterial species. To explore the potential role of unculturable and culturable periodontitis-related bacteria in preterm low birth weight (PLBW) delivery, we recruited 90 pregnant women in this prospective study. Periodontal parameters, including pocket probing depth, bleeding on probing, and clinical attachment level were recorded during the second trimester and following interviews on oral hygiene and lifestyle habits. Saliva and serum samples were also collected. After delivery, birth results were recorded. Real-time PCR analyses were performed to quantify the levels of periodontitis-related unculturable bacteria (Eubacterium saphenum, Fretibacterium sp. human oral taxon(HOT) 360, TM7 sp. HOT 356, and Rothia dentocariosa), and cultivable bacteria (Aggregatibacter actinomycetemcomitans, Porphyromonas gingivalis, Tannerella forsythia, Treponema denticola, Fusobacterium nucleatum and Prevotella intermedia) in saliva samples. In addition, ELISA analyses were used to determine the IgG titres against periodontal pathogens in serum samples. Subjects were categorized into a Healthy group (H, n = 20) and periodontitis/gingivitis group (PG, n = 70) according to their periodontal status. The brushing duration was significantly lower in the PG group compared to the H group. Twenty-two of 90 subjects delivered PLBW infants. There was no significant difference in periodontal parameters and serum IgG levels for periodontal pathogens between PLBW and healthy delivery (HD) groups. However, ordinal logistic regression analysis revealed that a higher abundance of Treponema denticola, Prevotella intermedia, Fretibacterium sp. HOT360 and lower levels of Rothia dentocariosa were significantly associated with the presence of periodontal disease during pregnancy. Moreover, the amount of Eubacterium saphenum in saliva and serum IgG against Aggregatibacter actinomycetemcomitans were negatively correlated with PLBW. Taken together, unculturable periodontitis-associated bacteria may play an important role both in the presence of periodontal inflammation during pregnancy and subsequent PLBW.


Results
Subject characteristics in the healthy and periodontitis/gingivitis groups. The duration of tooth brushing in the Periodontitis/Gingivitis group was significantly shorter than that in the Healthy group (Table 1). There were no significant differences between the two groups in terms of socio-economic characteristics or serum levels of high-sensitivity C-reactive protein (hs-CRP).
There were no significant differences between the two groups in IgG antibodies against periodontal pathogens ( Fig. 1A and Table S1). Fig. 2A and Table S2, the amount of P. gingivalis, A. actinomycetemcomitans, T. forsythia, T. denticola, P. intermedia and Fretibacterium sp. HOT 360 in saliva in the PG group were significantly higher than those in the H group. In contrast, R. dentocariosa was significantly lower in the PG group.

Comparison of bacterial load between healthy and periodontitis/gingivitis groups. As showed in
Correlation of periodontal parameters with bacterial copy number. Figure 3A and Table S3 shows the results of the correlation analysis between periodontal parameters and bacterial copy number (log scale). The number of P. gingivalis and A. actinomycetemcomitans were positively correlated with BOP positive sites (%). Moreover, the bacterial load of P. intermedia, T. denticola and Fretibacterium sp. HOT 360 in saliva were positively correlated with mean PPD, mean CAL, and BOP positive sites (%). On the other hand, R. dentocariosa was negatively correlated with mean PPD, mean CAL, and BOP positive sites (%).
Subject characteristics in healthy delivery and preterm low birth weight groups. Seven of 90 pregnant women delivered PB infants and the other fifteen individuals delivered SGA infants. There was no significant difference in the gestational week at birth (birth week), birth weight, incidence (%) of PB, and SGA between the Healthy and Periodontitis/Gingivitis groups (Table 1). Table 2 shows the subject characteristics of the HD and PLBW groups. There were no significant differences in oral hygiene behaviors, socio-economic characteristics, and serum cytokine levels against periodontal pathogens between the two groups. There were no significant differences between the two groups in IgG antibodies against periodontal pathogens ( Fig. 1B and Table S4).
Comparison of bacterial load between healthy delivery and preterm low birth weight groups. Aa showed in Fig. 2B and Table S5, the amount of E. saphenum was significantly lower in the PLBW group than that in the HD group (p < 0.01). There was no other significant difference in periodontal pathogens between the two groups.
Correlation of birth results with log copy number of bacteria. As showed in Fig. 3B and Table S6, the log copy number of E. saphenum was positively correlated with the gestational week at birth (r = 0.22, p = 0.03). There was no other significant difference between bacteria and birth results.
Potential risk indicators for periodontal disease during pregnancy and PLBW. Ordinal logistic regression analysis revealed that the high levels of T. denticola, P. intermedia, Fretibacterium sp. HOT360 and low amount of R. dentocariosa in saliva were significantly associated with the presence of gingival inflammation during pregnancy (Table 3). Meanwhile, the amount of E. saphenum and serum IgG against A. actinomycetemcomitans showed a negative correlation with the prevalence of PLBW ( www.nature.com/scientificreports/

Discussion
The oral microbiome can be affected by numerous environmental factors including pH, anaerobic conditions, nutrition, and hormone levels [14][15][16] . During pregnancy, periodontal tissues show an enhanced inflammatory response to plaque microbiome, believed to be mediated by female sex hormones 17 . Periodontal pathogens, such as P. intermedia, P. gingivalis, T. forsythia, Campylobacter rectus, F. nucleatum, and A. actinomycetemcomitans are highly abundant in the saliva and subgingival biofilm of pregnancy gingivitis patients [18][19][20] . In addition, P. gingivalis, T. denticola, Fretibacterium sp., and P. intermedia levels in saliva and plaque samples from pregnant women were associated with gingival bleeding 21 . Similarly, in the present study, P. gingivalis, A. actinomycetemcomitans, P. intermedia, T. denticola and Fretibacterium sp. HOT 360 levels were positively correlated with periodontal parameters during pregnancy ( Fig. 3A and Table S3). In the current study, although IgG antibodies against periodontal pathogens were present at higher levels in the PG group, there was no significant difference between the H and PG groups. Fretibacterium sp. HOT 360, which belongs to phylum Synergistetes, is found in high proportions in subgingival plaque or saliva samples of patients with chronic periodontitis 13,[22][23][24][25] . However, the characteristics of this bacterial species remain unknown. The present study findings and that of our previous report 13 , suggest that Fretibacterium sp. HOT 360 may be a periodontal disease-related bacteria, and it was positively correlated with a deterioration of maternal gingival status.
R. dentocariosa is a gram-positive bacterium that is representative of the normal flora in the human oral cavity and pharynx 22,26 . Kumar et.al. 27 also reported that R. dentocariosa exists in high proportions under healthy gingival conditions. In addition, our study demonstrated that there was a negative correlation between R. dentocariosa and periodontal parameters, such as PPD and BOP (Fig. 3A), and a decreased amount of R. dentocariosa was associated with an increased risk for gingivitis in pregnant women (Table 3). This suggests that a decrease in R. dentocariosa may be a potential risk indicator for and play a role in the onset of gingivitis in pregnancy. However, R. dentocariosa has also been isolated from dental caries [28][29][30] . Taken in aggregate, these findings suggest that R. Table 1. Subject characteristics in healthy and periodontitis/gingivitis groups. Parametric continuous variables were tested using an unpaired t-test and are given as means ± standard deviations; nonparametric continuous variables were tested using the Mann-Whitney U-test and are given as medians (quartile). PG, periodontitis/gingivitis; PB, preterm birth; SGA, small for gestational age; PPD, pocket probing depth; CAL, clinical attachment loss; BOP, bleeding on probing; hs-CRP, high-sensitivity C-reactive protein; NS, not significant. *Significantly different from the H group (P < 0.05). www.nature.com/scientificreports/ dentocariosa may play a role in the development of dental caries, but decreasing levels of R. dentocariosa may predispose pregnant women to the development of gingivitis. Eubacterium saphenum is a newly discovered anaerobic gram-positive and rod-shaped bacteria, which was reported as a putative periodontal pathogen, as it was isolated from periodontal pockets in periodontitis patients 10,22,31 . No study has addressed the role of E. saphenum in systemic diseases. In this study, the amount of E. saphenum in saliva did not show any significant differences between the H and PG groups. However, a decreased amount of E. saphenum was correlated with PLBW (Table 4). Furthermore, the amount of E. saphenum was positively correlated with the birth week versus the birth weight ( Fig. 3B and Table S6). This indicated the lacking of E. saphenum might be related to the progression of preterm birth than intrauterine growth restriction. A possible mechanism for this finding may be that there is a loss of a beneficial effect in controlling an oral microbiome dysbiosis caused by a reduction in E. saphenum that subsequently induces an altered maternal and fetal immunological response and systemic inflammation, leading to preterm birth. Further investigation is warranted to reveal the potential mechanism. Also, the lack of IgG against A. actinomycetemcomitans was associated with an increased risk for PLBW. Maternal IgG might protect the fetus from exposure to pathogens 32,33 . Taken together, these data suggest that loss of E. saphenum and loss of IgG response to A. actinomycetemcomitans may be a risk indicator for PLBW.
Although brushing frequency was similar in the Healthy and Periodontitis/Gingivitis group, the shorter brushing time for the PG group might contribute to the development of periodontal inflammation during pregnancy in this group. Hence, proper personal plaque control measures and periodontal therapy should be pursued to reduce periodontal inflammation and increases in periodontitis-associated bacteria during pregnancy.
The present investigation is the first study to explore the relationship between unculturable periodontitisassociated bacteria and PLBW. However, this study has some limitations, including an uneven distribution of subjects in the H and PG groups. A second limitation is its somewhat cross-sectional design. The microbial etiology of maternal periodontitis appears to be complex, and this fact warrants further large-scale longitudinal studies to understand these bacteria and their role in PLBW outcomes. www.nature.com/scientificreports/ In conclusion, the results of our study showed that a greater abundance of T. denticola, P. intermedia, Fretibacterium sp. HOT360 and a lower abundance of R. dentocariosa in saliva may pose an increased risk for gingival/ periodontal inflammation during pregnancy. Moreover, a low amount of E. saphenum in saliva and anti-A. actinomycetemcomitans IgG in serum may increase the risk of PLBW. Study population. The present study was a longitudinal observational study. The optimal sample size was determined using the G*Power software 34 . The tail was set at one, the odds ratio at 2, Pr(Y = 1/X = 1)HO at 0.2, alpha error at 0.05, and power (1 -beta) at 0.8, R 2 other X at 0, X distribution at normal, X parm μ at 0 and X parm σ at 1 for logistic regression, and the minimum sample size was estimated to be 88 participants.

Methods
Ninety Chinese pregnant women aged 21 to 39 who received a periodontal examination during their middle trimester of pregnancy were recruited in the Department of Periodontology, West China Dental Hospital, Sichuan University from May 2015 to May 2018. All subjects had a minimum of 20 teeth and did not receive any periodontal treatment or antibiotic treatment within three months of their participation in the study. Any subjects with systemic disease or multiple gestations were excluded. All patients were divided into a periodontitis/gingivitis group (PG; n = 70) and gingival health group (H; n = 20) according to their periodontal status (described subsequently). After delivery, gestational age at birth, birth weight, infant gender, and delivery mode were recorded. Preterm birth (PB) was defined as a gestational age less than 37 weeks and small for gestational age (SGA) delivery was defined as a birth weight of less than the 10th percentile for gestational age based on www.nature.com/scientificreports/ Chinese norms 35 . Therefore, the PLBW group consisted of PB and/or SGA. The healthy delivery (HD) group was defined as pregnancy's showing a gestational age ≥ 37 weeks and a birth weight of more than the 10th percentile for gestational age.
Interview and periodontal examination. All pregnant women underwent a full-mouth periodontal examination at 26 to 28 weeks of gestation. Before the examination, information about confounding factors, such as oral hygiene behavior (the frequency of tooth brushing, duration of tooth brushing, utilization of special appliances, and utilization of an oral rinse), smoking habits, socioeconomic characteristics (education level, income level, and health insurance status), and an alcohol consumption habits was obtained by structured interviews. Periodontal examinations were performed by one periodontal specialist (C.Y.) with the use of a manual probe (PCP-UNC15, Hu-Friedy, Chicago, USA). The periodontal parameters, including pocket probing depth (PPD), clinical attachment level (CAL), and bleeding on probing (BOP), were recorded at all six sites for each tooth (mesiobuccal, mid-buccal, distobuccal, mesiolingual, mid-lingual and distolingual). Subsequently, subjects were divided into a periodontitis/gingivitis (PG) group (n = 70) and a Healthy (H) group (n = 20) according to periodontal status. Subjects showing BOP in > 25% of sites and/or ≥ 4 teeth with one or more sites of PPD ≥ 4 mm were placed in the PG group. Subjects who did not fit the above criteria were diagnosed as having clinical gingival health and were placed in the H group 36-38 . Sample collection. One mL of unstimulated saliva was obtained at the same time as the periodontal examination and then stored at − 80 °C. Participants were instructed to abstain from food intake and oral hygiene procedures for at least 2 h before saliva collection. Also, at 28 weeks of gestational age, peripheral blood samples were collected and centrifuged at 1500×g for 10 min at 4 °C. Subsequently, serum samples were collected and immediately stored at − 80 °C until further analysis.
Bacterial detection in saliva. Genomic DNA was extracted using the QIAamp DNA Mini kit (QIAGEN., CA, USA). The amount of specific bacterial species in duplicate saliva samples were then quantified by realtime polymerase chain reaction (RT-PCR) using primers and TaqMan probes for E. saphenum, R. dentocariosa, Fretibacterium sp. HOT 360, TM7 sp. HOT 356, A. actinomycetemcomitans, P. gingivalis, T. forsythia, T. denticola, F. nucleatum, and P. intermedia 13,[39][40][41] (Table 5). The PCR reaction was performed in a total volume of 20 μL of reaction mixtures containing 10 μL Premix Ex Taq, 0.4 μL each of the forward and reverse primer (final concentration 10 nM), 0.5 μL TaqMan probe (final concentration 10 nM), 1 μL template DNA solution, and an appropriate volume of sterilized DNase-and RNase-free water. The PCR TaqMan Master Mix without DNA was used as negative control. www.nature.com/scientificreports/ Table 2. Subjects' characteristic in health delivery group and preterm low birth weight group. Parametric continuous variables were tested using an unpaired t-test and are given as means ± standard deviations; nonparametric continuous variables were tested using the Mann-Whitney U-test and are given as medians (quartile). PB, preterm birth; SGA, small for gestational age; PPD, pocket probing depth; CAL, clinical attachment loss; BOP, bleeding on probing; hs-CRP, high-sensitivity C-reactive protein; NS, not significant.. *Significantly different from the HD group (P < 0.05).  www.nature.com/scientificreports/ The optimized thermal cycling conditions for amplification were carried out on a thermal cycler device (Real-Time System II; Takara-bio, Tokyo, Japan) at 95 ºC for 30 s, 95 ºC for 5 s (40 cycles) and 60 ºC for 30 s. Standard curves were obtained by using tenfold serial dilutions of artificial synthetic genes, which contained the amplified region corresponding to 16S sequences of each bacteria (Eurofin genomics, Tokyo, Japan).

Statistical analysis.
Statistical analysis was performed using SPSS 20.0 software (SPSS Inc., Chicago, IL, USA). Initially, the distribution pattern of the data was assessed using the Shapiro-Wilk test. Thereafter, an unpaired t-test was used to analyze the statistical differences in age, brushing frequency, brushing time, weeks at birth, and birth weight between the H and PG groups, or HD and PLBW groups, respectively. In addition, a Mann-Whitney U-test was used to analyze the mean PPD, mean CAL, sites with PPD ≥ 3 mm (%), sites with PPD ≥ 4 mm (%), sites with PPD ≥ 5 mm (%), BOP-positive sites (%), serum CRP concentration, the number of bacteria in saliva, and the serum IgG titer between the H and PG groups, or HD and PLBW groups. Also, a chisquared test was used to analyze the primiparity rate, the utilization rate of dental floss or interproximal brushing, the utilization rate of mouth rinse, the amount showing more than 12-years of education, the amount with more than the local average income, the amount with dental insurance, Caesarean section rate, PB rate and SGA rate between the PG and H groups, or the PLBW and HD groups. Furthermore, Pearson's correlation coefficient test was used to analyze the correlation between periodontal parameters and bacterial load.
To determine potential bacterial and periodontal risk factors for pregnant women, we evaluated correlations between the presence of periodontal disease or PLBW with PPD ≥ 5 mm (%), BOP, the duration of tooth brushing, the amount of P. gingivalis, A. actinomycetemcomitans, T. forsythia, T. denticola, P. intermedia, Fretibacterium sp. HOT 360, R. dentocariosa, E. saphenum in saliva and IgG against A. actinomycetemcomitans using ordinal regression analysis, respectively. Odds ratios and 95% confidence intervals were calculated. A P value less than 0.05 was considered statistically significant. Table 5. PCR primers and probes for periodontal pathogens and "not-yet-culturable" periodontitis-related bacteria. F, forward; R, reverse.  F:TAG CTT GCT AAG GTC GAT GG  TGC GTA ACG CGT ATG CAA CTT GCC  R:CAA GTG TAT GCG GTT TTA GT   A. actinomycetemcomitans  F:TCT TAC CTA CTC TTG ACA TCC GAA  AGA ACT CAG AGA TGG GTT TGT GCC TTAG  R:ATG CAG CAC CTG TCT CAA AGC   T. forsythia  F:CGA CGG AGA GTG AGA GCT TTCT  CGT CTA TGT AGG TGC TGC ATG GTT