Cervical Gardnerella vaginalis in women with preterm prelabor rupture of membranes

Objective To determine the association between microbial invasion of the amniotic cavity (MIAC) and/or intra-amniotic inflammation (IAI) and the cervical prevalence of Gardnerella vaginalis DNA in pregnancies with preterm prelabor rupture of membrane (PPROM). Method In total, 405 women with singleton pregnancies complicated with PPROM were included. Cervical fluid and amniotic fluid samples were collected at the time of admission. Bacterial and G. vaginalis DNA were assessed in the cervical fluid samples using quantitative PCR technique. Concentrations of interleukin-6 and MIAC were evaluated in the amniotic fluid samples. Loads of G. vaginalis DNA ≥ 1% of the total cervical bacterial DNA were used to define the cervical prevalence of G. vaginalis as abundant. Based on the MIAC and IAI, women were categorized into four groups: with intra-amniotic infection (both MIAC and IAI), with sterile IAI (IAI without MIAC), with MIAC without IAI, and without either MIAC or IAI. Results The presence of the abundant cervical G. vaginalis was related to MIAC (with: 65% vs. without: 44%; p = 0.0004) but not IAI (with: 52% vs. without: 48%; p = 0.70). Women with MIAC without IAI had the highest load of the cervical G. vaginalis DNA (median 2.0 × 104 copies DNA/mL) and the highest presence of abundant cervical G. vaginalis (73%). Conclusions In women with PPROM, the presence of cervical G. vaginalis was associated with MIAC, mainly without the concurrent presence of IAI.

Introduction Hospital Hradec Kralove in the Czech Republic, was conducted between August 2015 and January 2020. The inclusion criteria were: i) singleton pregnancies complicated by the presence of PPROM and ii) age � 18 years. The exclusion criteria were as follows: i) fetal growth restriction; ii) congenital or chromosomal fetal abnormalities; iii) gestational or pregestational diabetes; iv) gestational hypertension; v) preeclampsia; vi) signs of fetal hypoxia; and vii) significant vaginal bleeding.
Gestational age was established using first-trimester fetal biometry. PPROM was diagnosed by the confirmation of the pooling of amniotic fluid in the vagina during the examination with a sterile speculum. If clinical doubts about PPROM remained after this examination, amniotic fluid leakage was confirmed by the presence of insulin-like growth factor binding proteins (Actim PROM test; Medix Biochemica, Kauniainen, Finland) in the vaginal fluid. Treatment of PPROM was described in our previous studies [14,22].
In total, 311 women from this cohort were used in our previous study based on the presence of L. crispatus-and L. iners-dominated cervical microbiota in women with PPROM [14].
Women's selection and sampling procedures were performed in accordance with the Declaration of Helsinki and applicable local regulatory requirements after approval from the Ethics Committee of the University Hospital Hradec Kralove (July 2014; No 201408 S07P). Written informed consents were obtained from all participants. All participants in the study were Caucasian.

Cervical sampling
Cervical fluid samples were collected prior to administration of corticosteroids, antibiotics, or tocolytics using a Dacron polyester swab placed in the cervical canal for 20 s to achieve saturation. Upon collection, the Dacron polyester swab was inserted into a polypropylene tube containing 1.5 mL of phosphate-buffered saline. Each tube was shaken for 20 min, followed by centrifugation at 300 × g for 15 min at room temperature. The supernatant and pellet were aliquoted and stored at -80˚C until further analyses. The pellet was used to assess the presence of bacterial and G. vaginalis DNA.

Amniotic fluid sampling
Ultrasonography-guided transabdominal amniocentesis was performed prior to administering corticosteroids, antibiotics, or tocolytics. Details about amniotic fluid sampling have been described in our previous publication [14].

Detection of cervical G. vaginalis DNA
Nucleic acid was isolated from the pellets using the tissue protocol with the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) [14]. In-house real-time PCR was developed to detect G. vaginalis. Primers and hydrolysis probes for G. vaginalis were designed from 16S rRNA region to amplify a 164 bp amplicon (forward primer GV F578 GTG AAA GCC CAT CGC TTA AC, reverse primer GV R741 TTC GCT TCT CAG CGT CAG TA and probe FAM-BHQ1 AAT TCT CGG TGT AAC GGT GG) and real-time PCR was performed on a Rotor-Gene Q instrument (Qiagen) in 25 μL reaction volumes, containing universal 2x gb IPC PCR Master Mix (Generi Biotech, Hradec Kralove, Czech Republic) with an internal positive control, primers at concentrations of 400 nM each, and dual-labeled hydrolysis probes (FAM-BHQ1) at a concentration of 200 nM [14]. Primers and probes were synthesized by Generi Biotech. Amplification parameters were as follows: 95˚C for 5 min, followed by 45 cycles of 95˚C for 15 s and 60˚C for 30 s [14]. PCR detection was performed by absolute quantification and a standard curve was generated from serial 10-fold dilutions of linearized and normalized plasmids containing the cloned target sequences for G. vaginalis at concentrations of 10 7 copies/μL (Generi Biotech, Hradec Kralove, Czech Republic).
To identify the relative abundance of G. vaginalis, total bacterial DNA detection was performed using quantitative RT-PCR-BactQuant [23]. To quantify the bacterial load, we used the forward primer CCT ACG GGD GGC WGC A, reverse primer GGA CTA CHV GGG TMT CTA ATC, and hydrolysis probe FAM-BHQ1 CAG CCG CGG TA, and a calibration curve was generated using 10-fold dilutions of linearized and normalized plasmid containing the cloned target sequence of the 466-bp region in the V3-V4 domain of 16S rRNA at a concentration of 10 7 copies/μL (Generi Biotech) [14]. RT-PCR for G.vaginalis detection was validated with a commercially available positive control Amplirun Gardnerella vaginalis DNA Control (Vircell Microbiologists, Spain).
Relative abundance of G. vaginalis in cervical microbiota was expressed as the ratio of the DNA quantity of G. vaginalis to the total cervical bacterial DNA load. The PCR conditions used in the BactQuant assay were the same as those used for G. vaginalis.

Detection of Ureaplasma spp., Mycoplasma hominis, and Chlamydia trachomatis and others bacteria in amniotic fluid
DNA was isolated from the amniotic fluid using a QIAamp DNA mini kit according to the manufacturer's instructions.
RT-PCR was performed on a Rotor-Gene 6000 instrument using the commercial Ampli-Sens 1 C. trachomatis/Ureaplasma/M. hominis-FRT kit (Federal State Institution of Science, Central Research Institute of Epidemiology, Moscow, Russia) to detect the DNA from Ureaplasma spp., M. hominis, and C. trachomatis in the same PCR tube. As a control, we amplified beta-actin, a housekeeping gene, to exclude the presence of PCR inhibitors.
Bacterial DNA was identified by PCR targeting 16S rRNA using the following primers: 5 0 -CCAGACTCCTACGGGAGGCAG-3 0 (V3 region) and 5 0 -ACATTTCACAACACGAGCTGAC GA-3 0 (V6 region) [24]. The products were visualized on an agarose gel. Positive reactions yielded 950 bp products that were subsequently analyzed using sequencing. The 16S PCR products were purified and sequenced using PCR with the above primers and BigDye Terminator kit v3.1 (Thermo Fisher Scientific, Waltham, MA). The bacteria were then typed using the sequences obtained from BLAST 1 and SepsiTest TM BLAST.

Aerobic and anaerobic cultures of the amniotic fluid
The amniotic fluid samples were cultured in Columbia agar with sheep's blood, G. vaginalis selective medium, MacConkey agar, Neisseria-selective medium (modified Thayer-Martin medium), Sabouraud agar, and Schaedler anaerobe agar. The plates were cultured for 6 days and checked daily. The species were identified by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry using MALDI Biotyper software (Bruker Daltonics, Billerica, MA).
Concentration of IL-6 in the amniotic fluid samples taken between December 2018 and January 2020 were evaluated using automated electrochemiluminescence immunoassay method with immuno-analyzer Cobas e602, which is a part of the Cobas 8000 platform (Roche Diagnostics, Basel, Switzerland). The measurement range was 1.5-5,000 pg/mL, which could be extended to 50,000 pg/mL with a 10-fold dilution of the sample. The coefficients of variation for inter-and intra-assay precisions were < 10% [22].

Definition of G. vaginalis as abundant bacteria in cervical microbiota
G. vaginalis was classified as abundant bacteria in cervical microbiota when its relative abundance was > 1% of the total cervical microbiota [26][27][28][29] (cervical loads of G. vaginalis DNA > 1% of the amount of the cervical bacterial DNA).

Diagnosis of MIAC
MIAC was defined as the presence of microorganisms detected by culture and/or the detection of microbial nucleic acids in amniotic fluid (PCR analysis of Ureaplasma spp., M. hominis, C trachomatis and PCR analysis of 16S rRNA gene). Due to heterogeneity of microorganisms present in amniotic fluid obtained from PPROM pregnancies, with Ureaplasma spp. being the most common, [30,31] women with MIAC were further split into three groups: i) with just Ureaplasma spp. in the amniotic fluid; ii) with the bacteria other than Ureaplasma spp. in the amniotic fluid; and iii) with both Ureaplasma spp. and the other bacteria in the amniotic fluid.
Based on the presence of MIAC and/or IAI, the women were categorized into four groups: with intra-amniotic infection (both MIAC and IAI), with sterile IAI (IAI without MIAC), with MIAC without IAI, and without either MIAC or IAI.

Definitions of select aspects of short-term neonatal morbidity
Maternal and perinatal medical records were reviewed by seven investigators (DL, MK, JS, TF, JM, PV, and IM). Data regarding short-term neonatal morbidity were reviewed for all newborns. "Compound neonatal morbidity" was defined in this study as follows: the need for intubation, and/or need for nasal continuous positive airway pressure, and/or respiratory distress syndrome (defined by the presence of two or more of the following criteria: evidence of respiratory compromise, persistent oxygen requirement for more than 24 h, administration of exogenous surfactant, radiographic evidence of hyaline membrane disease); and/or transient tachypnea of newborns (defined as any oxygen supplement requirement during the first 6 h that did not increase during the subsequent 18 h as clinical conditions improved within 3-6 h and chest radiographs was either normal or indicating reduced translucency, infiltrates, and hyperinsufflation of the lungs); and/or bronchopulmonary dysplasia (defined as infant oxygen requirement at 28 days of age); and/or pneumonia (diagnosed by abnormal findings on chest X rays); and/or retinopathy of prematurity (identified using retinoscopy); and/or intraventricular hemorrhage (diagnosed by cranial ultrasound examinations according as described by Papile et al. [34]); and/or necrotizing enterocolitis (defined as radiologic findings of either intramural gas or free intra-abdominal gas); and/or early-(during the first 72 h of life) or lateonset-(between the ages of 4 and 120 days) sepsis (either proven by bacterial culture or clinically highly suspected sepsis); and/or neonatal death before hospital discharge.

Statistical analyses
The normality of the data was tested using the D'Agostino-Pearson omnibus normality test and Shapiro-Wilk test. Continuous variables were compared using nonparametric Mann-Whitney U test or Kruskal-Wallis test, and presented as the median value (interquartile range [IQR]). Categorical variables were compared using Fisher's exact test or chi-square test, and presented as a number (%). A partial correlation was used to adjust the results for the different methods of IL-6 assessment. Differences were considered significant at p < 0.05. All p values were obtained from two-tailed tests, and all statistical analyses were performed using Graph-Pad Prism version 6.0h software for Mac OS X (GraphPad, Inc., San Diego, CA, USA) and SPSS version 19.0 for Mac OS X (SPSS, Inc., Chicago, IL, USA).

Results
A total of 464 women with singleton pregnancies at a gestational age of 24+0 to 36+6 weeks were admitted during the study period. A total of 24 women were excluded because their cervical fluid was not collected (n = 15), amniocentesis failure (n = 3), unconfirmed PPROM (n = 2), and delivery prior to amniocentesis (n = 4). In total, 35 women were subsequently excluded due the following medical reasons: gestational diabetes mellitus (n = 11), fetal growth restriction (n = 8), chronic hypertension (n = 5), pre-gestation diabetes mellitus (n = 4), preeclampsia (n = 4), and pregnancy induced hypertension (n = 3). A total of 405 women were included in the study.
MIAC and IAI were identified in 24% (96/405) and 18% (73/405) of the women, respectively. G. vaginalis DNA in amniotic fluid was detected in 1% (4/405) of the women. The women's demographic and clinical data based on the presence and absence of MIAC and IAI are shown in Table 1. Intra-amniotic infection was observed in 12% (47/405), sterile intra-amniotic inflammation in 6% (26/405), MIAC without IAI in 12% (49/405), and the absence of either MIAC or IAI in 70% (283/405) of the women. Bacterial species identified in the amniotic fluid obtained in women with intra-amniotic infection and with MIAC without IAI are shown in Table 2.
Cervical G. vaginalis DNA was found in 94% (379/405) of the women. Abundant cervical G. vaginalis (cervical loads of G. vaginalis DNA > 1% of the amount of the cervical bacterial DNA) was found in 49% (197/405) of the women. Differences between gestational age at sampling and the rates of the presence of cervical G. vaginalis DNA (p = 0.01, Fig 1A) and abundant cervical G. vaginalis DNA (p = 0.007, Fig 1B) were observed.
Among women with the presence of cervical G. vaginalis DNA, those with MIAC were also split into three subgroups: i) with just Ureaplasma spp. in the amniotic fluid (n = 49); ii) with the bacteria other than Ureaplasma spp. in the amniotic fluid (n = 36); and iii) with both Ureaplasma spp. and the other bacteria in the amniotic fluid (n = 11). No difference in the loads of cervical G. vaginalis DNA (Ureaplasma spp.: median 6.5 × 10 3 copies DNA/mL, IQR 4.3× 10 2 -4.9 × 10 5 vs. other: median 5.6 × 10 4 copies DNA/mL, IQR 6.1 ×

IAI and cervical G. vaginalis
The cervical presence of G. vaginalis DNA was identified in 96% (70/73) and 93% (309/332) of the women with and without IAI, respectively.
No difference in the rates of the presence of the abundant cervical G. vaginalis was identified between the women with and without IAI [with IAI: 52% (38/73) vs. without IAI: 48% (158/332); p = 0.70].

The presence of MIAC without IAI and cervical G. vaginalis DNA
Among women with the presence of cervical G. vaginalis DNA along with a higher cervical load of G. vaginalis DNA, those with MIAC without IAI had a higher load of the cervical G. vaginalis DNA than remaining women in the crude analysis (MIAC without IAI: median 2.0 × 10 4 copies DNA/mL, IQR 1.9 × 10 2 -5.2 × 10 5 vs. remaining women: median 4.0 × 10 3 copies DNA/mL, IQR 2.7 × 10 2 -2.1 × 10 5 ; p = 0.04; Fig 2C) but not after the adjustment for the different methods of IL-6 assessment (p = 0.48).

Short-term neonatal morbidity and cervical G. vaginalis
The presence of the abundant cervical G. vaginalis was not associated with the differences in the short-term neonatal morbidity (Table 3).

Discussion
The principal findings of this study are as follows: i) the cervical G. vaginalis DNA was found in 94% of women with PPROM between gestational ages 24+0 and 36+6 weeks; ii) the abundant cervical G. vaginalis DNA was identified in 49% of women with PPROM between gestational ages 24+0 and 36+6 weeks; iii) women with MIAC had a higher microbial load of cervical G. vaginalis DNA and a higher rate of the abundant cervical G. vaginalis; iv) the presence of MIAC without IAI was associated with a higher microbial load of cervical G. vaginalis DNA and a higher rate of the abundant cervical G. vaginalis; and v) no association between a short-term neonatal morbidity and the abundant cervical G. vaginalis was revealed.
G. vaginalis is one of the essential bacteria associated with bacterial vaginosis [17,35]. Its vaginal presence during pregnancy seems to be common. The prevalence of the vaginal presence of G. vaginalis DNA differs among the studies, from 27% to 94% of the pregnant women [36][37][38]. Since an association between bacterial vaginosis and PPROM is well established [19][20][21]39], a higher cervical presence of G. vaginalis among women with PPROM was expected. In this study, RT-PCR was used to detect G. vaginalis DNA. This method was able identify the cervical G. vaginalis DNA at very low loads owing to a detection limit of this method between 1 and 10 copies of the nucleic acid. Because of this, the presence of cervical G. vaginalis DNA was identified in almost in all women (94%) with PPROM. However, 30% (115/379) of the women with proven cervical G. vaginalis DNA had very low loads (below 1000 copies DNA/ mL) of G. vaginalis DNA. A clinical relevance for such extremely low loads of cervical G. vaginalis is highly unlikely, and the abundant presence (> 1% of all cervical bacterial DNA) of cervical G. vaginalis was also taken into consideration in this study. Given this more clinically relevant and meaningful criterion for the cervical presence of G. vaginalis, the finding regarding the abundant presence of cervical G. vaginalis (49%) was in line with our previous study Table 3. Selected aspects of short-term neonatal morbidity with respect to the presence or absence of the abundant cervical Gardnerella vaginalis in women with preterm prelabor rupture of membranes.

The presence (n = 197)
The absence (n = 208) p-value  [15], where G. vaginalis DNA identified using 16S rRNA gene sequencing by pyrosequencing, was found in 46% (28/61) of the women with PPROM [15]. MIAC in pregnancies in PPROM represents a very heterogeneous condition due to different: i) microorganisms present in the amniotic fluid [25,30,40]; ii) number of microbial species identified in the amniotic fluid [14,30,40]; iii) loads of microorganisms revealed in the amniotic fluid [41,42]; and iv) an intensity of intra-amniotic inflammatory response [25,31,40,41]. Furthermore, MIAC in PPROM has two different clinical phenotypes: i) intra-amniotic infection and ii) the presence of MIAC without IAI [7]. The latter has been shown to have similar intensities of intra-amniotic [7], maternal [8,9], and fetal [43] inflammatory responses to those without MIAC or IAI. Regardless, the fact that a spectrum of the microorganisms revealed in amniotic fluid from women with PPROM with intra-amniotic infection and with MIAC without IAI looks very similar, their microbial loads seem to be lower in the presence of MIAC without IAI, at least in those with Ureaplasma spp. in the amniotic fluid [7]. Moreover, intra-amniotic infection is associated with the most intense intra-amniotic [7], maternal [8,9], and fetal [43] inflammatory responses.
In this study, we found that the presence of MIAC, but not IAI, was associated with higher abundant cervical G. vaginalis DNA presence. Interestingly, when women with PPROM were divided into four subgroups (intra-amniotic infection, sterile IAI, MIAC without IAI, and without either MIAC or IAI), the loads of G. vaginalis DNA and the rate of the abundant cervical presence of G. vaginalis were the highest in women with MIAC without IAI. This clinically interesting and relevant finding suggests that the presence of the cervical G. vaginalis in women with PPROM is associated mainly with the development of the less threatening form of MIAC, which does not provoke an inflammatory response in the maternal, fetal, and intraamniotic compartments.
G. vaginalis is a pathogen that exhibits the following important virulence determinants: i) adhesion to vaginal epithelial cells [44]; ii) production of biofilms that can be detected not only in the vagina [17,44] but also in the endometrium and fallopian tubes [45]; and iii) cytotoxic activity [46,47]. These findings confirm a unique role of G. vaginalis in the pathophysiology of bacterial vaginosis [35]. The ability of G. vaginalis to create a biofilm might explain our observation regarding the association between the cervical presence of G. vaginalis and the presence of MIAC without IAI. It has been shown that some uropathogens such Escherichia coli and Enterococcus faecalis can enter into the G. vaginalis biofilm, survive in this environment, and they can continue to ascend in the upper part of the urogenital system, like the bladder [48,49]. Therefore, we hypothesize that similar mechanism can enhance an ascension of microorganisms with either low virulence potential or with low loads (not enough to trigger an intensive intra-amniotic inflammatory response), from the vagina through the cervix into the amniotic cavity, in pregnancies with PPROM. At this stage we do not have an exact mechanistic explanation of the association between cervical G. vaginalis DNA and the presence of MIAC without IAI. Nevertheless, this observation requires further investigation using larger cohorts of women with very well defined microbial and inflammatory compositions of the cervical and amniotic fluid compartments.
There is evidence that the cervical/vaginal presence of some bacteria might affect some aspects of short-term neonatal morbidity, mainly early-onset sepsis, in newborns from PPROM pregnancies [14,50]. For example, Brown et al. have shown that vaginal microbiota enriched for Catonella spp. and Sneathia spp. increased the risk of early-onset sepsis in the newborn [50]. Recently, our group has revealed a relation between the absence of L. crispatusdominated cervical microbiota and early-onset sepsis of the newborn [14]. Therefore, selected aspects of short-term neonatal morbidity in relation to the abundant presence of the cervical G. vaginalis were investigated in this study. However, no associations between the abundant cervical presence of G. vaginalis and early-onset sepsis and the other selected aspects of shortterm neonatal morbidity were identified. This observation was complementary to our previous finding that L. iners-dominated cervical microbiota did not affect the short-term neonatal morbidity [14].
The main strength of this study was a large cohort of the women with singleton pregnancy complicated by PPROM, for whom the results from amniotic fluid for MIAC and IAI at the time of admission were known. However, this study has some limitations. In the murine model, it has been shown that vaginal infection by G. vaginalis can lead to an ascending intrauterine infection by this bacteria, mainly in the cases with higher vaginal microbial loads of G. vaginalis [51]. Nevertheless, the ability of G. vaginalis to ascend into the uterus has not been demonstrated in the pregnant murine model [52]. Despite the fact that amniotic and cervical fluid samples used in this study were obtained at the time of admission into the labor room, we were unable to use these paired samples for a more comprehensive assessment of the association between the cervical and amniotic fluid presence of G. vaginalis, owing to the different analytical approaches used to identify G. vaginalis in both types of the samples. Second, leaking amniotic fluid through the cervix is an unavoidable confounder in PPROM pregnancies that may have affected the obtained cervical fluid samples in the following manner: i) inhibition of the growth of some bacteria [53,54]; ii) dilution of the cervical fluid samples and a decrease in the load of G. vaginalis DNA; and iii) elevation of the cervical bacterial DNA loads in the samples from the women with proven MIAC. Third, G. vaginalis is a bacteria with known phenotypic and genetic heterogeneity [55]. Based on the 473 G. vaginalis genes, and by sequencing a single 552 bp region of the chaperonin-60 gene, four clades (1-4) and four subgroups (A-D) of G. vaginalis having different pathogenic potentials were revealed [56][57][58][59][60][61]. The design of this study did not give us an opportunity to take heterogeneity of G. vaginalis into consideration. Fourth, data on Nugent score or the presence of Amsel's clinical criteria for bacterial vaginosis is not available for this cohort of the women. Fifth, the absence of the control cohort consisting from women with an uncomplicated singleton pregnancy, matched for gestational age at sampling, prevented us from the evaluation of the association between the presence of cervical G. vaginalis DNA and PPROM. In addition, the population of the women with PPROM used in this study was very homogenous involving just Caucasian women without comorbidities. This fact should be considered as a shortcoming due to limited generalization of these results on populations with broad racial and ethnic diversities. Next, two different methods were used in this study to assess amniotic fluid levels of IL-6 (with different cut-off values for IAI) [22,25]. This fact prevented us from evaluation of the relationship between loads of the cervical G. vaginalis DNA or an abundance of the cervical G. vaginalis and the intensity of intra-amniotic inflammation. In addition, the use of two different methods to assess levels of amniotic fluid IL-6 should be taken as a possible confounder. Therefore, appropriate results were adjusted for the different methods of IL-6 assessment. After the adjustment, differences in the rates of the abundant cervical G. vaginalis among the women with intra-amniotic infection, sterile IAI, MIAC without IAI and without either MIAC or IAI and between the women with MIAC without IAI and remaining women remained statistically significant. However, no difference in the cervical load of G. vaginalis DNA between women with MIAC without IAI and remaining women was observed after the adjustment for the different methods of IL-6 assessment. Lastly, the presence of MIAC and IAI might be associated with the alteration of the microbial loads of the others cervical microorganisms. Therefore, we cannot rule out that our observations were not a transient effect.
In conclusion, in women with PPROM, the presence of cervical G. vaginalis was associated with MIAC, mainly without the concurrent presence of IAI.