Association between Lactobacillus species and bacterial vaginosis-related bacteria, and bacterial vaginosis scores in pregnant Japanese women.

BACKGROUND: Bacterial vaginosis (BV), the etiology of which is still uncertain, increases the risk of preterm birth. Recent PCR-based studies suggested that BV is associated with complex vaginal bacterial communities, including many newly recognized bacterial species in non-pregnant women.
METHODS: To examine whether these bacteria are also involved in BV in pregnant Japanese women, vaginal fluid samples were taken from 132 women, classified as normal (n = 98), intermediate (n = 21), or BV (n = 13) using the Nugent gram stain criteria, and studied. DNA extracted from these samples was analyzed for bacterial sequences of any Lactobacillus, four Lactobacillus species, and four BV-related bacteria by PCR with primers for 16S ribosomal DNA including a universal Lactobacillus primer, Lactobacillus species-specific primers for L. crispatus, L. jensenii, L. gasseri, and L. iners, and BV-related bacterium-specific primers for BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium.
RESULTS: The prevalences of L. crispatus, L. jensenii, and L. gasseri were significantly higher, while those of BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium were significantly lower in the normal group than in the BV group. Unlike other Lactobacillus species, the prevalence of L. iners did not differ between the three groups and women with L. iners were significantly more likely to have BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium. Linear regression analysis revealed associations of BVAB2 and Megasphaera with Nugent score, and multivariate regression analyses suggested a close relationship between Eggerthella-like bacterium and BV.
CONCLUSION: The BV-related bacteria, including BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium, are common in the vagina of pregnant Japanese women with BV. The presence of L. iners may be correlated with vaginal colonization by these BV-related bacteria.

Background

Bacterial vaginosis (BV) is the disturbed vaginal flora, in which normal lactobacilli are replaced by an overgrowth of various anaerobic bacteria [1]. This condition is common in women of reproductive age [1,2] and may cause malodorous vaginal discharge, although in many women it is asymptomatic [3]. In pregnant women, bacterial vaginosis has been suggested to be a risk factor of perinatal complications, including preterm birth [1,4-12] and chorioamnionitis [4,13]. These complications are closely associated with neonatal morbidity and mortality worldwide.

Bacteria detected in BV flora include Gardnerella vaginalis, Mycoplasma hominis, Mobiluncus species (sp.), and other anaerobic bacteria, i.e., Peptostreptococcus sp., Prevotella sp., and Bacteroides sp. [1,14-16]. Recently, bacteria such as Atopobium vaginae, Megasphaera sp., Leptotrichia sp., and Eggerthella-like bacterium have been reported as microorganisms related to this condition by molecular analyses [17-19]. Fredricks et al. identified three phylogenetically distinct bacterial DNA sequences in human vaginal samples highly specific for this condition and designated them BV-associated bacteria (BVAB) 1~3 [19]. They showed that BVABs, Megasphaera, Leptotrichia, and Eggerthella-like bacterium are more specific for BV than Gardnerella and Atopobium [19]. Among BVABs, BVAB2 was shown to be more sensitive for BV than BVAB1 and BVAB3, while the specificities of three BVABs were similar [19]. We use the term "BV-related bacteria" for bacteria including BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium in this manuscript. However, it is important to note that these organisms have not been proven to be causative agents of BV.

The healthy human vaginal flora in reproductive age is usually predominated by Lactobacillus species. Their metabolic products, such as hydrogen peroxide (H2O2), lactic acid, and bacteriocin are believed to play an important role in maintenance of the normal vaginal flora by inhibiting colonization by other pathogens [20-23]. The predominant Lactobacillus species in the normal lactobacillary flora were shown by molecular biological analyses to be L. crispatus, L. gasseri, and L. jensenii [23-28]. In recent studies, L. iners described as L. 1086V by Antonio et al. [24] was identified as one of the common Lactobacillus species colonizing the human vagina [18,28-31]. Only 9% of the strains of this species produce H2O2, whereas almost all strains of L. crispatus and L. jensenii produce H2O2 [24].

To date, there have been few studies regarding the frequencies of the BV-related bacteria described above and Lactobacillus species in healthy and abnormal vaginal flora in pregnant women. The present study was performed to evaluate the prevalence of the BV-related bacteria and the common Lactobacillus species in normal and BV flora in pregnant Japanese women. We used a specific PCR method targeting the bacterial 16S ribosomal DNA (rDNA) region for this purpose.

Methods

Patients

A total of 163 pregnant Japanese women were enrolled in this study during routine prenatal visits at Hokkaido University Hospital from May 2005 to February 2006. Informed consent was obtained from all participants in verbal form. Vaginal fluid samples were collected at a mean of 23 weeks of gestation. Estimated date of delivery was determined from the last menstrual period and early gestational fetal ultrasonographic measurements.

Sample collection, Nugent's scoring, and bacterial culture

A sterile speculum was inserted into the vagina and a specimen of vaginal fluid was obtained by brushing the posterior vaginal fornix with a swab. A vaginal smear was prepared by rolling a swab onto a glass slide, which was then air-dried, heat-fixed, and Gram-stained. The smears were then assessed according to Nugent criteria [32]. The other swab was spread onto Columbia blood agar plates, and incubated at 35°C under aerobic conditions in 5% CO2 and anaerobic conditions for 48 h. Lactobacilli were identified to the genus level by Gram staining of colonies and from colony morphology on blood agar plates.

DNA extraction and PCR

Another swab was placed in 1 ml of PBS with subsequent vigorous vortexing to dislodge cells. The cells were centrifuged at 14,000 rpm for 5 min. The pellet was digested with proteinase K at 56°C for 30–60 min and the DNA was extracted and purified with a QIAmp DNA Mini Kit (Qiagen, Germantown, MD) in accordance with the manufacturer's instructions, resulting in 200 μl of DNA solution. PCR mixtures consisted of PCR buffer with 1.5 mM of MgCl2, 10 pmol of each primer, 2.0 μM of each deoxyribonucleoside triphosphate, 0.1 μl of Taq DNA polymerase, and 1.5 μl of template DNA solution in a final volume of 15 μl.

Sequences and annealing temperatures for the various primer sets are listed in Table 1[19,33]. All primers were located in the 16S rDNA region. PCR was carried out for 40 cycles. For the Lactobacillus genus and its four species, the denaturation was performed at 95°C for 15 s followed by a 1-min annealing and extension step. For four BV-related bacteria, the denaturation step was set at 94°C for 30 s, followed by the annealing step for 30 s for BVAB2, Megasphaera, and Leptotrichia and for 40 s for Eggerthella-like bacterium, with extension at 72°C for 1 min for all reactions. A final extension step at 72°C for 7 min was added for all reactions. Aliquots of 7 μl of the PCR products were electrophoresed in agarose gels and visualized by ultraviolet transillumination after ethidium bromide staining.

Table 1

PCR primers

Name	Sequence(5'-3')	Target	Annealing temperature (°C)	Reference
LactoFLactoR	TGGAAACAGRTGCTAATACCGGTCCATTGTGGAAGATTCCC	Lactobacillus	62	[33]
LcrisFLcrisR	AGCGAGCGGAACTAACAGATTTACAGCTGATCATGCGATCTGCTT	L. crispatus	65	[33]
LjensFLjensR	AAGTCGAGCGAGCTTGCCTATAGACTTCTTTCATGCGAAAGTAGC	L. jensenii	60	the present study
LgassFLgassR	AGCGAGCTTGCCTAGATGAATTTGTCTTTTAAACTCTAGACATGCGTC	L. gasseri	63	the present study
LinersFLinersR	CTCTGCCTTGAAGATCGGAGTGCACAGTTGATAGGCATCATCTG	L. iners	65	the present study
Uncxb2-619FUncxb2-1024R	TTAACCTTGGGGTTCATTACAAAATTCAGTCTCCTGAATCGTCAGA	BVAB2	55	[19]
Egger-621FEgger-859R	AACCTCGAGCCGGGTTCCTCGGCACGGAAGATGTAATCT	Eggerthella-like bacterium	58	[19]
Lepto-395FLepto-646R	CAATTCTGTGTGTGTGAAGAAGACAGTTTTGTAGGCAAGCCTAT	Leptotrichia	55	[19]
MegaE-456FMegaE-667R	GATGCCAACAGTATCCGTCCGCCTCTCCGACACTCAAGTTCGA	Megasphaera	55	[19]

The specificity of the Lactobacillus species-specific PCR for 14 common intestinal Lactobacillus species was evaluated and confirmed using 106 copies of one Lactobacillus species to each reaction as template DNA (Table 2). The universal Lactobacillus primer amplified all of these Lactobacillus species. The specific primers for L. crispatus, L. jensenii, and L. gasseri only amplified the corresponding species and did not amplify 13 other species (Table 2). They also did not amplify a cloned fragment of 16S rDNA region of L. iners. The specific primers for L. iners did not amplify any of 14 Lactobacillus species (Table 2). We analyzed PCR products from several vaginal samples amplified by the specific primers for L. iners and confirmed that the sequences of PCR products were completely consistent with L. iners (GenBank AY526083).

Table 2

Bacterial strains and the specificity of primers

Species	Strain	LactoF	LcrisF	LjensF	LgassF	LinersF
		LactoR	LcrisR	LjensR	LgassR	LinersR
L. crispatus	ATCC33197	+	+	-	-	-
L. jensenii	ATCC25258	+	-	+	-	-
L. gasseri	ATCC 4963	+	-	-	+	-
L. acidophilus	ATCC 4356	+	-	-	-	-
L. brevis	ATCC 14869	+	-	-	-	-
L. casei	ATCC 334	+	-	-	-	-
L. delbrueckii	ATCC 11842	+	-	-	-	-
L. fermentum	ATCC 14931	+	-	-	-	-
L. johnsonii	ATCC 11506	+	-	-	-	-
L. helveticus	ATCC 521	+	-	-	-	-
L. plantarum	ATCC 8014	+	-	-	-	-
L. reuteri	JCM1112	+	-	-	-	-
L. rhamnosus	ATCC 7469	+	-	-	-	-
L. salvarius	ATCC 11741	+	-	-	-	-

The sensitivities of the species-specific PCR for L. crispatus, L. jensenii, and L. gasseri were measured using serial dilutions of DNA solution of the reference strain. Similarly, for L. iners, serial dilutions of a cloned fragment of 16S rDNA region of L. iners were used instead. The sensitivity of the species-specific PCR for L. crispatus, L. jensenii, and L. gasseri and that of L. iners PCR were 102 to 103 copies and 102 copies per reaction, respectively.

Statistical analysis

Fisher's exact probability test was used for statistical analysis. Multivariate logistic-regression analysis using SPSS™ for Windows was performed to evaluate the independent risk factors, and P < 0.05 was considered statistically significant.

Results

Clinical characteristics of women in three groups divided by Nugent score

A total of 163 samples from 163 pregnant women were obtained during the study period and 31 samples were excluded because of lack of information regarding the gestational week at delivery (n = 21) or Gram staining (n = 10). The remaining 132 samples from 132 women were analyzed and classified according to the Nugent criteria. Ninety-eight women (74.2%) were classified as having normal vaginal flora, 21 (15.9%) were intermediate, and 13 (9.8%) were BV. Samples from these women were divided into normal, intermediate, and BV groups, respectively.

The clinical characteristics of the pregnant women are summarized in Table 3. There were no statistically significant differences in the mean maternal age, number of nulliparous women, gestational week at sampling, gestational week at delivery, or birth weight of the neonate among the three groups. Of the total of 132 women, 35 (26.5%) delivered at <37 weeks, 9 (6.8%) at <33 weeks, and 4 (3.0%) at <30 weeks of gestation.

Table 3

Demographic and obstetric characteristics of women in normal, intermediate, and BV groups

	Nugent score
	0–3		4–6		7–10
No. of women	98		21		13
Age (years)	32.6 ± 5.3	(19–44)	32.1 ± 5.8	(20–40)	29.1 ± 5.3	(21–37)
Nulliparity (%)	46.2		52.4		50.0
Gestational week at sampling	22.6 ± 8.6	(5–36)	23.0 ± 9.1	(7–36)	24.2 ± 11.2	(7–37)
Gestational week at delivery	37.1 ± 2.9	(21–41)	36.4 ± 4.7	(18–40)	37.3 ± 5.5	(20–41)
Preterm birth at <37 weeks	26	26.5%	7	33.3%	2	15.4%
Preterm birth at <33 weeks	6	6.1%	2	9.5%	1	7.7%
Preterm birth at <30 weeks	2	2.0%	1	4.8%	1	7.7%
Birth weight (g)	2807 ± 565	(360–3805)	2602 ± 794	(165–3660)	2807 ± 831	(350–3555)

Range is shown in parenthesis.

Detection rate of lactobacilli and BV-related bacteria in three groups by PCR

Genus Lactobacillus (any Lactobacillus) was detected in almost all women irrespective of Nugent score (Table 4). The detection rates of L. crispatus, L. jensenii, and L. gasseri were significantly higher in the normal group than in the BV group, while that of L. iners did not differ between the three groups. In contrast, the detection rates of BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium were significantly lower in the normal group than in the BV group.

Table 4

Distribution of lactobacilli and bacterial vaginosis-related bacteria in women in normal, intermediate, and BV groups determined by PCR

	Nugent score
	0–3	4–6	7–10
No. of women	98	21	13
any Lactobacillus	97(99.0%)	21(100.0%)	12(92.3%)
L. crispatus	60(61.2%)	6(28.6%)§	2(15.4%)§
L. jensenii	29(29.6%)	4(19.0%)	0(0.0%)†
L. gasseri	33(33.7%)	9(42.9%)	0(0.0%)§
L. iners	39(39.8%)	10(47.6%)	6(46.2%)
BVAB2	3(3.1%)	4(19.0%)†	5(38.5%)§
Megasphaera	11(11.2%)	13(61.9%)§	9(69.2%)§
Leptotrichia	14(14.3%)	5(23.8%)	7(53.8%)§
Eggerthella-like bacterium	7(7.1%)	7(33.3%)§	7(53.8%)§

†: P < 0.05, §: P < 0.01, vs group of Nugent score 0–3

Independent risk factors for abnormal Nugent score

Multivariate logistic regression analysis was performed to evaluate the independent contributions of the various bacteria to the abnormal vaginal flora (Tables 5 and 6). Seven bacteria, i.e., L. crispatus,L jensenii,L. gasseri, BVAB2, Megasphaera,Leptotrichia and, Eggerthella-like bacterium, were entered as variates to be analyzed (P < 0.1, Fisher's exact probability test). The absence of L. crispatus and the presence of Megasphaera were selected as two independent risk factors of Nugent score ≥ 4, giving Odds ratios of 0.2 and 13.3, respectively (Table 5). Likewise, the presence of Eggerthella-like bacterium was selected as an independent risk factor of Nugent score ≥ 7, giving an Odds ratio of 6.2 (Table 6). Linear regression analyses revealed that BVAB2 and Megasphaera were associated with Nugent score.

Table 5

Independent risk factors for Nugent score ≥ 4 by multivariate regression analysis

	β	SE	p-value	Odds (95%CI)
Constant	-1.32	0.35
L. crispatus	-1.51	0.52	0.004	0.22 (0.08–0.61)
Megasphaera	2.60	0.51	0.001	13.33 (4.92–36.11)

Table 6

Independent risk factors for Nugent score ≥ 7 by multivariate regression analysis

	β	SE	p-value	Odds (95%CI)
Constant	-2.24	0.47
Eggerthella-like bacterium	1.83	0.65	0.005	6.24 (1.75–22.21)

Coexistence of BV-related bacteria with L. iners

L. iners was detected by PCR in 55 of 132 (41.7%) women, and its prevalence did not differ between the groups classified according to Nugent score (Table 4). However, the presence of L. iners appeared to be positively associated with colonization by BV-related bacteria (Table 7). The detection rates of all BV-related bacteria were significantly higher in samples harboring L. iners. No such association was seen between the presence or absence of L. iners and the detection rate of any other Lactobacillus species.

Table 7

Prevalence of various bacteria according to the presence or absence of L. iners

	L. iners
	present	absent	p-value
Number	55	77
Lactobacillus	55 (100.0%)	75 (97.4%)	0.510
L.crispatus	25 (45.5%)	43(55.8%)	0.290
L. jensenii	13 (23.6%)	20(26.0%)	0.840
L. gasseri	15 (27.3%)	27(35.1%)	0.449
BVAB2	11 (20.0%)	1 (1.3%)	<0.001
Megasphaera	21 (38.2%)	12(15.6%)	0.004
Leptotrichia	19 (34.5%)	7(9.1%)	<0.001
Eggerthella-like bacterium	15 (27.3%)	6 (7.8%)	0.004

Difference in the detection of Lactobacillus species between PCR and culture methods

Lactobacillus was cultured from 91 (92.9%) of 98 samples, 11 (52.4%) of 21 samples, and 2 (15.4%) of 13 samples of the normal, intermediate, and BV groups, respectively (data not shown). These observations conflicted markedly with the results obtained by the PCR method, especially in women with abnormal vaginal flora with respect to detection of Lactobacillus. Of the eleven women with BV from whom Lactobacillus was uncultured but detected by PCR, L. iners was detected in 5 women and L. crispatus was detected in only one woman by PCR, suggesting that L. iners is less likely to be cultured than L. crispatus. To determine which species of Lactobacillus is difficult to culture, the detection rates by PCR of various species of Lactobacillus were compared with those by the conventional culture method (Table 8). Among 130 samples determined to contain any Lactobacillus by the PCR method, 104 (80.0%) were positive for Lactobacillus by the culture method (Table 8). More than 90% of samples determined to contain L. crispatus, L. jensenii, or L. gasseri by the PCR method were determined to have Lactobacillus by the conventional culture method. Among 24 samples in which L. crispatus was the only Lactobacillus species identified by the PCR method, Lactobacillus was cultured from 22 samples (91.7%), while Lactobacillus was cultured only from 47.6% and 27.3% of samples in which L. iners and unspecified Lactobacillus species, respectively, were the only Lactobacillus species identified by the PCR method. Thus, L. iners, and unspecified Lactobacillus species other than L. crispatus, L. jensenii, or L. gasseri appeared to have stringent cultivation requirements.

Table 8

Differences in detection between PCR and cultivation methods

	PCR	CULTURE	CULTURE/PCR(%)
any Lactobacillus	130	104	80.0%
L. crispatus	68	63	92.6%
L. jensenii	33	32	97.0%
L. gasseri	42	39	92.9%
L. iners	55	41	74.5%
L. crispatus only	24	22	91.7%§
L. jensenii only	5	4	80.0%§
L. gasseri only	14	13	92.9%§
L. iners only	21	10	47.6%§
Lactobacillus, unspecified spp.only	11	3	27.3%§

§P < 0.01

Number of samples in which lactobacillus species were detected.

Discussion

In the present study, we confirmed that L. crispatus, L. gasseri, and L. jensenii were common species in pregnant Japanese women with normal vaginal flora by species-specific PCR of the 16S rDNA region. These three species were less prevalent in women with BV. In contrast, four BV-related bacteria, i.e., BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium, were detected at higher prevalence in women with BV. As all these results were in accordance with those of Fredricks et al. [19] who analyzed the vaginal fluid of non-pregnant women with and without BV using the broad-range 16S rDNA PCR and cloning methods, BV is suggested to have remarkably similar microbiological profiles among women with different demographic characteristics, including race and pregnancy, as suggested by the conventional cultivation method.

L. crispatus, L. gasseri, and L. jensenii are common Lactobacillus species found in the vagina [24-28,31,34]. L. iners, described recently as a new Lactobacillus species [29], is one of the common Lactobacillus species of the vaginal microbiota [18,19,23,28,30,31], which was also confirmed in the present study. The results showed that L. iners was present in 40% to 50% of women irrespective of Nugent score, as observed in an earlier study [19]. We examined twelve samples positive for L. iners (6 from normal flora and 6 from BV flora) to determine whether the abundance of L. iners was different in the two groups. The species-specific PCR for L. iners using serial dilutions of each sample revealed that both normal and BV flora contained 103 to 105 copies/μl of L. iners and the median concentration was 104 copies/μl for both.

As the presence of H2O2-producing lactobacilli in the vaginal fluid is associated with a reduced risk of BV [15,24] and because the concentration of H2O2 in the vaginal fluid is low in women with BV as compared with those with normal vaginal flora [21], the H2O2-producing ability of lactobacilli is thought to play a significant role in protecting the vaginal ecosystem from BV infection, although direct evidence to support this notion is lacking. Nearly all strains of L. crispatus and L. jensenii have been reported to produce H2O2, whereas only 9% of the strains of L. iners produce H2O2 [24]. The prevalences of L. crispatus and L. jensenii were significantly higher in the normal group than in the BV group and the detection rates of all BV-related bacteria were significantly higher in women with than in those without L. iners in this study. Although this observation is consistent with the notion that H2O2-producing ability of lactobacilli is important in protecting the vaginal ecosystem from BV infection, it remains to be determined whether these observations resulted from differences in H2O2-producing ability of these lactobacilli.

The newly proposed "BV-related bacteria," including BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium, were all shown to be associated with BV in the present study, confirming the results of a recent study by Fredricks et al. [19]. However, the detection rates of these bacteria in women with BV were lower, while those in women with normal flora were similar to their results [19]. BVAB2 is cultivation-resistant, one of three bacteria (provisionally named BV-associated bacteria: BVAB1, BVAB2, and BVAB3) newly found to be highly specific for BV in the vagina of non-pregnant women [19], and not closely related to other bacteria as shown by comparison of 16S rDNA. In the present study, BVAB2 was present in 38.5% (5/13) and 3.1% (3/98) of women with BV and with normal vaginal flora, respectively, while Fredricks et al. reported these rates to be 88.9% (24/27) and 4.3% (2/46), respectively [19]. Similarly, detection rates of Megasphaera (69.2%),Eggerthella-like bacterium (53.8%), and Leptotrichia (53.8%) in women with BV in the present study were lower than those of 96.3%, 92.6%, and 85.2% reported by Fredricks et al. [19], while detection rates of Megasphaera (11.2%),Eggerthella-like bacterium (7.1%), and Leptotrichia (14.3%) in women with normal vaginal flora were comparable to their values of 8.7%, 8.7%, and 4.3%, respectively [19].

The results of the present study raised the possibility that the four BV-related bacteria were less prevalent in pregnant Japanese women with BV as compared with non-pregnant American women. However, the number of subjects with BV in the present study was too low to draw definitive conclusions about the prevalence of bacteria in different populations. Further studies using different demographic populations are needed to determine the roles of these BV-related bacteria in the pathogenesis of BV.

Twelve (92%) of 13 women with BV were positive for genus Lactobacillus by 16S rDNA PCR using the universal Lactobacillus primer, including 5 women with L. iners, one with both L. iners and L. crispatus, one with L. crispatus, and 5 with unspecified Lactobacillus. Of these 13 women, only one with L. iners was positive for Lactobacillus by general cultivation methods and positive for Gram-positive rods on Gram staining. These results suggested that many women with BV harbor genus Lactobacillus in the vagina and that the number of these lactobacilli colonizing the vagina is small. Further, as L. iners has been reported to require specialized blood agar media for isolation [29], the conventional culture method used in this study may have failed to reveal its colonization in the vagina.

Conclusion

Our results suggested that BV-related bacteria, including BVAB2, Megasphaera, Leptotrichia, and Eggerthella-like bacterium, were associated with BV in pregnant Japanese women. The presence of L. iners, one of the common Lactobacillus species in the vagina, may be correlated with vaginal colonization by these BV-related bacteria. It remains to be determined whether BV-related bacteria cause BV or are common and abundant as a consequence of BV.

Competing interests

The author(s) declare that they have no competing interests.

Authors' contributions

The manuscript was written by RT. RT and TY contributed 16S rDNA-based bacterial identification. TY and IF supervised the microbiology laboratory work. TY and KC performed the statistical analyses. TY and MM provided clinical samples. HY, NS and HM critically reviewed the manuscript. All authors read and approved the final manuscript.

Pre-publication history

The pre-publication history for this paper can be accessed here: