• No results found

The association of uterine cervical microbiota with an increased risk for cervical intraepithelial neoplasia in Korea

N/A
N/A
Protected

Academic year: 2021

Share "The association of uterine cervical microbiota with an increased risk for cervical intraepithelial neoplasia in Korea"

Copied!
9
0
0

Loading.... (view fulltext now)

Full text

(1)

The association of uterine cervical microbiota with an increased risk for

cervical intraepithelial neoplasia in Korea

H. Y. Oh1, B.-S. Kim2,3, S.-S. Seo4, J.-S. Kong1, J.-K. Lee5, S.-Y. Park4, K.-M. Hong6, H.-K. Kim6and M. K. Kim1 1) Division of Cancer Epidemiology and Management, National Cancer Centre, Ilsandong-gu, Goyang, 2) Chunlab Inc., Seoul National University,

Seoul, 3) Department of Life Science, Hallym University, Chuncheon, 4) Centre for Uterine Cancer, National Cancer Centre, Ilsandong-gu, Goyang, 5) Department of Obstetrics and Gynaecology, Korea University College of Medicine, Seoul and 6) Division of Cancer Biology, National Cancer Centre, Goyang-si

Abstract

Recent studies have suggested potential roles of the microbiome in cervicovaginal diseases. However, there has been no report on the cervical microbiome in cervical intraepithelial neoplasia (CIN). We aimed to identify the cervical microbiota of Korean women and assess the association between the cervical microbiota and CIN, and to determine the combined effect of the microbiota and human papillomavirus (HPV) on the risk of CIN. The cervical microbiota of 70 women with CIN and 50 control women was analysed using pyrosequencing based on the 16S rRNA gene. The associations between specific microbial patterns or abundance of specific microbiota and CIN risk were assessed using multivariate logistic regression, and the relative excess risk due to interaction (RERI) and the synergy index (S) were calculated. The phyla Firmicutes, Actinobacteria, Bacteroidetes, Proteobacteria, Tenericutes, Fusobacteria and TM7 were predominant in the microbiota and four distinct community types were observed in all women. A high score of the pattern characterized by predominance of Atopobium vaginae, Gardnerella vaginalis and Lactobacillus iners with a minority of Lactobacillus crispatus had a higher CIN risk (OR 5.80, 95% CI 1.73‒19.4) and abundance of A. vaginae had a higher CIN risk (OR 6.63, 95% CI 1.61‒27.2). The synergistic effect of a high score of this microbial pattern and oncogenic HPV was observed (OR 34.1, 95% CI 4.95‒284.5; RERI/S, 15.9/1.93). A predominance of A. vaginae, G. vaginalis and L. iners with a concomitant paucity of L. crispatus in the cervical microbiota was associated with CIN risk, suggesting that bacterial dysbiosis and its combination with oncogenic HPV may be a risk factor for cervical neoplasia. Clinical Microbiology and Infection © 2015 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

Keywords: Atopobium vaginae, cervical intraepithelial neoplasia, human papillomavirus, Lactobacillus crispatus, Lactobacillus iners, microbiome, pyrosequencing

Original Submission: 22 September 2014; Revised Submission: 4 January 2015; Accepted: 26 February 2015 Editor: D. Raoult

Article published online: 6 March 2015

Corresponding author: M.-K. Kim, Translational Epidemiology Branch, Division of Cancer Epidemiology and Prevention, National Cancer Centre, 111, Jungbalsan-ro, Madu-dong, Ilsandong-gu, Goyang-si, Gyeonggi-do 411-769, Korea

E-mail:alrud@ncc.re.kr

These authors (Hea Young Oh and Bong-Soo Kim) contributed equally to this study

Introduction

The importance of the microbiota in health maintenance or disease susceptibility is well known; the gut microbiota is known to control the absorption of nutrients[1]. Imbalance in the gut is associated with Crohn’s disease, type 2 diabetes and chronic allergies [2], for which dysbiosis, not a single pathogen change but shifts in the relative abundance of mi-crobes, has long been suggested. Furthermore, the association of microbial dysbiosis with several cancer types has been suggested in areas surrounded by mucosal membranes where

Clin Microbiol Infect 2015; 21: 674.e1–674.e9

(2)

bacteria live densely [3]. Although only modest effects of dysbiosis on cancer have been presented so far, the long duration of dysbiosis and possible combined effects with other risk factors suggest greater clinical implications [4]. Although human papillomavirus (HPV) has been the major risk for cervical precancerous lesions or cancer[5]; recently, the potential role of the cervicovaginal microbiome in cervical cancer through the elevation of pH has been reported[6]. In addition, the role of the cervicovaginal microbiome in HPV infection has also been reported[7], suggesting a possible role in cervical cancer through potentiation of HPV infection. Although there have been reports on the microbiome in

premenopausal or postmenopausal women [8], and in

bac-terial vaginosis (BV)[9], no study on the microbiome related to cervical precancerous lesions or cancer has been pre-sented. The purpose of this study was to assess the associa-tion between cervical intraepithelial neoplasia (CIN) the and cervical microbiota identified using pyrosequencing.

Materials and methods

Study participants and design

The Korean HPV cohort study, which was established to identify epidemic, genetic, viral and ecological factors associ-ated with the development of cervical neoplasia in Korean women, recruited 1096 women, aged 18–65 years, from March 2006 to the present. The study participants were randomly selected from the gynaecological oncology clinics of six tertiary medical centres in Korea. Detailed information regarding the inclusion/exclusion criteria and the design of the baseline measurements for the HPV cohort were presented in a previous study[10]. All study participants provided written informed consent in accordance with good clinical practices. Cervical swabs were collected for a Papanicoaou smear test. Immediately after sampling, each cervix brush (Rovers Medical Devices, Oss, the Netherlands) was rinsed in a vial of Pre-servCyt solution (Cytyc Co., Marlborough, MA, USA), and the vial was placed in a Thin Prep Processor (Cytyc Co.). A second swab was collected for high-risk (HR)-HPV DNA detection and microbiota analysis using a Cervical Sampler Brush (Digene Co. Gaithersburg, MD, USA). Half of the second swab was used immediately for HR-HPV DNA detection and the other half

was stored at −80°C for subsequent DNA extraction, for

2 months up to a maximum of 5 years. A total of 125 women were randomly selected from the 1096 enrolled women, and baseline samples obtained from 120 women (50 controls, 70 with CIN) were analysed; forfive of the samples, metagenomic DNA extraction failed. This study was approved by the Insti-tutional Review Board of the Korean National Cancer Centre

(NCCNCS-06-062) and by the ethics committees of the Korean National Cancer Centre and Korea University Guro Hospital.

Cytological screening and HR-HPV DNA detection The cervical cytologicalfindings were classified according to the

Bethesda system [11]. HPV DNA detection was performed

using the commercially available Digene Hybrid capture II DNA Test (Qiagen, Gaithersburg, MD, USA). The results of a chemiluminescent HPV DNA test were measured in relative light units (RLU) with a probe designed to detect 13 types of HR-HPV. The test results were read as positive at concentra-tions of 1 pg/mL or greater than the RLU/cut-off ratio (RLU of specimen/mean RLU of two positive controls).

DNA extraction and pyrosequencing

Metagenomic DNA samples were extracted using the Fast DNA SPIN extraction kits (MP Biomedicals, Santa Ana, CA, USA). Target fragments of the 16S rRNA gene corresponding to the V1–V3 regions were amplified using bar-coded primers. Amplifications were performed in a final volume of 50 μL containing 10 × Taq buffer, a dNTP mixture (Takara, Shiga, Japan), 10μM of the bar-coded fusion primers, and 2 U of Taq polymerase (ExTaq, Takara). The amplification conditions and pyrosequencing procedure have been previously described

[12]. The beads recovered following emulsion PCR were

deposited on a 454 Pico Titer Plate, and sequencing was per-formed using a Roche/454 GS Junior system (Roche, Branford, CT, USA).

Pyrosequencing data analysis

The raw sequences were sorted using their unique barcodes, and low-quality reads (with an average quality score <25 or a read length <300 bp) were removed [12]. The primer se-quences were trimmed using a pairwise sequence alignment, and sequences were clustered to correct for sequencing er-rors. Representative sequences in each cluster were identified using EzTaxon-e, a public database that contains sequences of the 16S rRNA genes of species type strains, and the taxo-nomic positions of relevant uncultured sequences were identified[13]using the highest pairwise similarity among the BLASTN results. Chimeric sequences were detected and removed by the UCHIME algorithm, which detects chimeric sequences by alignment of a query sequence with two parent

sequences in a reference database [14], and the

diversity indices were calculated using the MOTHURprogram. The heat-map analysis was performed using the MULTI-E XPERI-MENT VIEWER. Pyrosequencing reads are available in the

EMBL SRA database (http://www.ebi.ac.uk/ena/data/view/

(3)

Quantitative real-time PCR assays

Genomic DNA was extracted with 25 KU/mL mutanolysin (Sigma-Aldrich, St Louis, MO, USA) using a QIAamp DNA mini kit (Qiagen, Valencia, CA, USA), as described previously[15]. The DNA targets, primers, probe sequences, oligonucleotides used for PCR, and the amplification conditions are shown in the Supporting information (Table S1) [16].

Statistical analysis

Factor analysis was performed on the 18 most predominant species using STATA 12.0 (Stata Co., College Station, TX, USA). Among the 18 microbial patterns identified, ten had an eigenvalue >1. The pattern matrix (i.e. the correlation co-efficients between the species and the patterns) is provided in the Supporting information (Table S2). The Wilcoxon

rank-sum and Spearman’s rank correlation tests were performed

to measure differences between two groups and the

dependence between two variables, respectively. Logistic regression analysis was performed after adjustment for age, marital status, menopausal status, oral contraceptive use, smoking habit and HR-HPV infection. Risk estimates are pre-sented as OR with 95% CI. To determine biological in-teractions, ORs for additive scale and multiplicative terms were estimated, and the relative excess risk due to interaction (RERI) and synergy index (S) were calculated as described previously [17]. Values for RERI >0 and S >1 indicate the presence of a synergistic effect.

Results

General characteristics of study participants

Women with CIN 1 (n = 55) and CIN 2 or 3 (n = 15) were accrued, along with control women with normal cytology

FIG. 1.Comparison of the cervical microbiota of women at the species level using heat-map analysis. The relative abundance of each species was represented by a colour; red indicates a high proportion and blue indicates a low abundance. Hierarchical clustering of the cervical microbiota was performed using the Spearman’s correlation.

(4)

(n = 25) or atypical squamous cells of undetermined signifi-cance (ASCUS; n = 25) (seeSupporting information, Table S2). No significant difference was observed between control women and women with CIN except HR-HPV infection (p 0.002).

Cervical microbiota among participants

In total, 1 118 398 high-quality reads were analysed and an average of 74.1 operational taxonomic units (OTUs) per sample was observed. Seven phyla—Firmicutes, Actinobacteria, Bac-teroidetes, Proteobacteria, Tenericutes, Fusobacteria, and

candidate division TM7—were dominant (see Supporting

information, Fig. S1). Although Firmicutes was the predomi-nant phylum in most women (n = 75), Bacteroidetes, Actino-bacteria, Tenericutes and Proteobacteria were also prevalent (n = 45). The composition of the dominant phyla was not clearly distributed by HPV infection status or histological grades. However, the abundances of Bacteroidetes, Actino-bacteria, Tenericutes and Proteobacteria were higher in HPV-positive women (n = 34) than in HPV-negative women (14). Actinobacteria was observed more predominantly in women with CIN (n = 5) than in controls (6).

The OTU number was higher in women with CIN than controls in postmenopausal women, but lower in women with

CIN than controls in premenopausal women (see Supporting

information, Fig. S2a). The OTU number was higher in HPV-negative than HPV-positive women irrespective of meno-pausal status (seeSupporting information, Fig. S2b). The OTU

number in control women and women with CIN was similar in HPV-negative women, but it was higher in controls than in

women with CIN in HPV-positive women (see Supporting

information, Fig. S2c).

Cervical microbiota was distributed into four clusters using heat-map analysis at the species level (Fig. 1). Lactobacillus iners was predominant in cluster I; Atopobium vaginae, Prevotella bivia, Lactobacillus fornicalis, Pseudomonas poae and Gardnerella vaginalis were dominant members in cluster II; L. iners and Lactobacillus crispatus were dominant in cluster III; and L. iners, A. vaginae, P. bivia, Prevotella amnii, and uncultured bacteria ADGP (Meg-asphaera genomosp. type_1 str. 28L) were abundant in cluster IV. Although the clusters were not clearly separated by HPV infection or histology grade, most HPV-negative women carried L. iners and L. crispatus as the dominant members. There were more HPV-positive women in clusters II and IV, and A. vaginae, P. bivia, L. fornicalis, Pseudomonas poae and G. vaginalis were higher in cluster II than in cluster IV.

Risky microbial pattern associated with CIN risk The individual scores for the ten microbial patterns each ob-tained from factor analysis were used for the following step. Only the third pattern showed a significant difference in its score between controls and women with CIN (p 0.007) (Table 1and seeSupporting information, Table S3). The OR of a high tertile of this pattern score was 5.80 (95% CI 1.73–19.4) compared with a low tertile. This risky microbial pattern was characterized by a predominance of A. vaginae, L. iners and

TABLE 1.Odds ratios for cervical intraepithelial neoplasia (CIN) according to the scores of the risky microbial pattern and the

relative abundance of four species

OR and 95% CI for CINsc

Groups Median pa T1 (40)b T2 (40) T3 (40) p for trendd

Score of the risky microbial pattern Controls 0.03

CIN 0.33 0.007 1 (ref.) 1.85 (0.65–5.31) 5.80 (1.73–19.4) 0.004 Groups Median p T1 (67) T2 (27) T3 (26) p for trend Relative abundance of A. vaginae Controls 0.00

CIN 0.05 0.003 1 (ref.) 3.85 (1.22–12.2) 6.63 (1.61–27.2) 0.003 Groups Median p T1 (68) T2 (26) T3 (26) p for trend Relative abundance of G. vaginalis Controls 0.00

CIN 0.01 0.044 1 (ref.) 1.25 (0.41–3.76) 3.21 (0.95–10.8) 0.067 Groups Median p T1 (40) T2 (40) T3 (40) p for trend Relative abundance of L. iners Controls 1.08

CIN 3.36 0.188 1 (ref.) 3.16 (0.99–10.0) 2.49 (0.88–7.08) 0.089 Groups Median p T3 (33) T2 (33) T1 (54) p for trend Relative abundance of L. crispatus Controls 0.02

CIN 0.01 0.270 1 (ref.) 3.18 (0.96–10.5) 1.88 (0.65–5.38) 0.339

ap value was from Wilcoxon rank-sum test and p < 0.05 was regarded as significant. b

The scores of the risky microbial pattern and the relative abundances (%) of Atopobium vaginae, Gardnerella vaginalis, Lactobacillus iners and Lactobacillus crispatus were divided into three equal-sized subsets; T1 (tertile 1, low), T2 (tertile 2, medium) and T3 (tertile 3, high). The numbers of women included in each tertile are provided in brackets. p is for trend of OR of the tertiles and p < 0.05 was regarded as significant.

cOdds ratios and 95% CI of all variables were estimated by T1 as a reference category except those of the relative abundance of L. crispatus estimated by T3. All variables in this table

were adjusted for age, marital status, menopausal status, oral contraceptive use and smoking habit as categorical types.

(5)

G. vaginalis and a concomitant paucity of L. crispatus (see Supporting information, Table S4). The women with a high tertile of this pattern score had a high proportion of A. vaginae (mean 28.8%) and G. vaginalis (5.6%), and a low proportion of L. crispatus (0.95%) (Fig. 2a). Lactobacillus iners predominated in women with the medium to high scores (21.2%) (Fig. 2b). Association of A. vaginae, G. vaginalis, L. iners and L. crispatus with CIN risk

There was a significant median difference in relative abundance of A. vaginae between controls and women with CIN (p 0.003) (Table 1). The OR for CIN risk of a high tertile of A. vaginae relative abundance was 6.63 (95% CI 1.61‒27.2) compared with the low tertile. The abundances of L. iners and L. crispatus were not different between controls and women with CIN, and high tertiles of those species were not associated with CIN. How-ever, some women with CIN had a high proportion of L. iners (Fig. 1) and some had complete absence of or low L. crispatus.

Synergistic effect of the risky microbial pattern and HR-HPV on CIN risk

The OR for CIN risk of HR-HPV-positive women with a high score of the risky microbial pattern was 34.1 (95% CI 4.95–234.5), compared with HR-HPV-negative women with a low score. The RERI and S (15.9 and 1.93), and p value for the multiplicative term (p 0.018) also showed a synergistic effect

between the two (Table 2). The OR of HR-HPV-positive

women with A. vaginae in the cervix was 29.9 (95% CI 5.49–163.0), compared with HR-HPV-negative women without A. vaginae (RERI and S, 15.4 and 2.14; p 0.002). The OR of HR-HPV-positive women with L. iners in the cervix was 10.8 (95% CI 1.65–70.9) (RERI and S, 7.39 and 4.06; p < 0.001). Quantitative real-time PCR for A. vaginae and L. crispatus

To compare the results obtained from the quantitative real-time PCR assay and the pyroseqencing for A. vaginae and

FIG. 2.The distribution of four species (Gardnerella vaginalis, Atopobium vaginae, Lactobacillus iners and Lactobacillus crispatus). (a) Area distribution for the relative abundance of four species according to the score of the risky microbial pattern (red, G. vaginalis; orange, A. vaginae; yellow, L. iners; green, L. crispatus). (b) Box plots show a significant difference in the relative abundance of each species between the tertiles of the risky pattern score; (a) p < 0.001, (b) p < 0.001, (c) p = 0.0361, (d) p < 0.001 in the Kruskal–Wallis rank test.

(6)

L. crispatus, the ratio of the two species (log-transformed) was used (Fig. 3). The statistical dependence between the two ratios was significantly high (Spearman’s correlation coefficient 0.868; p 0.001). There was a significant difference between controls and women with CIN in both the ratios obtained from pyro-sequencing (p 0.016) and real-time PCR (p 0.003).

Discussion

This study showed that a cervical microbial composition char-acterized by a predominance of A. vaginae, G. vaginalis and L. iners with a concomitant paucity of L. crispatus is associated with a higher risk of CIN. The predominance of A. vaginae was a major contributor to this risk. Furthermore, a synergistic interaction between this microbial pattern and HR-HPV infec-tion that robustly increased the risk was observed.

In most women, L. crispatus and L. iners were the dominant cervical microbiota. These two species were also found to be predominant in previous studies of the vaginal microbiota of

Asian women [8,18]. In contrast to previous reports, A

vaginae, G. vaginalis, P. bivia and Pneumococcus poae were the dominant species in several women in this study. In previous

reports, a higher proportion of Proteobacteria, a lower pro-portion of Firmicutes, and a greater diversity of bacteria were observed in postmenopausal women than in premenopausal women[8]. However, in the present study, various microbial compositions and a higher proportion of Firmicutes were observed in postmenopausal women. In addition, higher numbers of OTU were detected in premenopausal women than in postmenopausal women. This difference may be attributed to differences in the target region of 16S rRNA genes analysed or in sequencing methods[19]. Although these differences could produce disparities among the results of studies of microbiota, the predominant members of the cer-vical microbiota detected here were consistent with those detected in previous studies that used different target regions [19]. The similarity of the results is a result of the relatively low diversity of the cervical microbiota, compared with microbiota from other parts of the human body. Therefore, various other factors, including sample collection techniques, temporal shifts in the microbiota during the menstrual cycle, hygiene practices, glycogen levels and host genetic factors could be associated with individual variation; these variations may underlie the discrepancy observed between the results of this study and previous studies[8,18–21].

TABLE 2.Synergistic joint effect of the risky microbial pattern with high-risk human papillomavirus (HR-HPV) infection on the

increase of cervical intraepithelial neoplasia risk

Joints Symbola N w/wob OR (95% CI)c RERI/Sd

Risky microbial pattern (a) and HR-HPV (b) 0, 0 5/19 1 (ref.) 15.9/1.93 0, 1 42/24 8.32 (2.18–31.7)

1, 0 7/3 10.8 (1.71–68.8) 1, 1 16/4 34.1 (4.95–234.5)

p for trende <0.001

p for interaction (a × b)f 0.018

Atopobium vaginae (a) and HR-HPV (b) 0, 0 4/15 1 (ref.) 15.4/2.14 0, 1 27/21 7.74 (1.70–35.3)

1, 0 8/7 7.76 (1.26–48.0) 1, 1 31/7 29.9 (5.49–163.0)

p for trend <0.001

p for interaction (a × b) 0.002

Gardnerella vaginalis (a) and HR-HPV (b) 0, 0 7/18 1 (ref.) 1.23/1.19 0, 1 27/16 5.17 (1.52–17.7)

1, 0 5/4 3.14 (0.55–18.1) 1, 1 31/12 8.55 (2.34–31.2)

p for trend 0.001

p for interaction (a × b) 0.031

Lactobacillus iners (a) and HR-HPV (b) 0, 0 2/6 1 (ref.) 7.39/4.06 0, 1 12/11 2.87 (0.43–19.1)

1, 0 10/16 1.55 (0.24–10.1) 1, 1 46/17 10.8 (1.65–70.9)

p for trend <0.001

p for interaction (a × b) <0.001

Lactobacillus crispatus (a) and HR-HPV (b) 1, 0 7/14 1 (ref.) −2.28/0.66 1, 1 30/15 6.82 (1.79–26.0)

0, 0 5/8 1.92 (0.40–9.11) 0, 1 28/13 5.45 (1.48–20.1)

p for trend 0.120

p for interaction (a × b) 0.250

aThe symbol‘0’ corresponds to the low score of the risky microbial pattern (low to medium tertile), the absence of indicated microbial species (the relative abundance = 0) and

non-HPV infection. The symbol‘1’ stands for the high risky microbial pattern score (high tertile), the presence of that (the relative abundance > 0) and HPV infection.

bThe number of women with/without cervical intraepithelial neoplasia (CIN) lesions was presented.

cOdds ratio (OR) was estimated after adjustment for age, marital status, menopausal status, oral contraceptive use and smoking history as categorical variables.

dThe biological interaction of the joint based on the additive model was investigated using relative excess risk due to interaction (RERI) and synergy index (S). RERI >0 and S > 1

indicate a synergistic effect.

ep is for the linear trend of odds ratios according to the order of combination provided by the logistic regression analysis. fP is for the interaction provided by the logistic regression for the multiplicative terms (a × b).

(7)

Clusters I and III contained a greater abundance of L. iners and L. crispatus. Clusters II and IV contained various bacteria as the predominant members, but differed in composition. These clusters were similar to previously reported community state types[18]. Clusters I and III were similar to community state types I and III described in a previous report[18]. Cluster II was similar to state type IV_B, and cluster IV was similar to type IV_A. Diverse microbial associations containing high proportions of Atopobium, Prevotella, Sneathia and Gardnerella have been reported to be related to BV [22]. Although the diverse community detected in clusters II and IV could be related to BV, diverse bacterial communities do not always

have persistently high Nugent scores [21]. A high Nugent

score is one of the methods available for diagnosis of BV, but it is not routinely used by physicians because of its inaccuracy. Molecular quantification of A. vaginae and G. vaginalis shows a higher reproducibility for diagnosis of BV than Nugent scores[16].

Although confirmation using other diagnostic tests is required, BV may be involved in the microbial composition associated with a high risk of CIN identified in the study. To date, several studies have explored the role of BV in cervical neoplasia; they have reported conflicting findings. In one pro-spective study, the CIN rate and the quantity of nitrosamines produced in women with BV did not differ from those in women with normalflora[23]. Another longitudinal study re-ported no influence of abnormal flora on the presence of atypical endocervical cells[24]. In contrast, Nam et al. showed a significant correlation between BV and CIN, regardless of severity [25]. Furthermore, a recently published systemic

review and meta-analysis showed a positive correlation

be-tween BV and CIN [26]. However, these ambivalent results

may be attributed to the different tools used in the diagnosis of BV. Of the above-mentioned studies, two used a BV diagnosis with three other measurements, including pH, clue cells after Gram stain, and a positive amine or whiff test[23,25], and the other used the Nugent scale [24]. Although the correlation between BV and CIN remains to be determined, our results are in support of such a correlation.

In this study, the women with the highest risk of CIN showed low levels of L. crispatus. Hydrogen peroxide-producing Lactobacilli are present in 96% of women with a normal vaginal bacterial community [27]. Species in the genus Lactobacillus maintain a low pH by producing lactic acid. Lactobacillus crispatus was more strongly associated with a low vaginal pH than L. iners, suggesting that these two species differ in ecological function [9]. Whereas a high abundance of L. crispatus was detected in the low-risk group in this study, L. iners was abundant in the medium-risk group, showing that this population indicates a susceptibility to CIN. Lactobacillus iners has been reported to become a predominant part of the microbial community when the vaginal microbiota transitions between abnormal and normal states[28]. HPV can alter mucosal metabolism and host immunity, and can induce changes in the vaginal microbiota

[29]. Atopobium vaginae also stimulates an innate immune

response from vaginal epithelial cells, leading to the secretion of interleukin, and the production of defending proteinβ-defensin (BD)[30]. Increased protein and mRNA expression of human BD2 and human BD3 was also observed in the genital condy-lomata acuminata of the vulvovaginal tract, induced by HPV

FIG. 3.Quantitative real-time PCR assay of Atopobium vaginae and Lactobacillus crispatus. Log of the ratio of A. vaginae and L. crispatus provided from pyrosequencing (the relative abundances) (a) and real-time PCR (equivalent/μL) (b). A Spearman’s rank correlation test was performed for measuring the statistical dependence between two variables; rank correlation coefficient = 0.868, p < 0.001. The p value was from the Wilcoxon rank-sum test for controls and women with cervial intraepithelial neoplasia (CIN).

(8)

[31]. Colonization of the vaginal epithelia by A. vaginae and G. vaginalis induced production of interleukin-8 and regulated on activation, normal T-cell expressed and secreted (RANTES), and amplified the pro-inflammatory response to both

mem-brane lipophosphoglycan/ceramide–phosphoinositol–glycan

core and Trichomonas vaginalis [32]. Levels of RANTES and macrophage inflammatory protein-1β were significantly higher in the cervical mucosa of HIV/HPV-co-infected individuals than in HPV-mono-infected individuals[33]. In a recent survey of the interactions between HPV types in healthy individuals, more than 50% of healthy women showed co-infection with two to three HPV types in vaginal tissue [34]. Evidence of complex interactions between viruses and immune regulators, bacterial flora and immune regulators, and co-infection with multiple viruses in the genital tract is accumulating[30–34]. Ourfindings demonstrated an interaction between the bacterial flora and HR-HPV-type infections, and a possible association between this interaction and clinical cervical neoplasia. The synergistic effect that increases the risk of CIN noted between a pre-dominance of BV-related bacteria and a concomitant HPV infection could be caused by a robust response of the same immune regulators.

This study demonstrated that bacterial dysbiosis, character-ized by a predominance of A. vaginae, G. vaginalis and L. iners and a concomitant paucity of L. crispatus, in its combination with oncogenic HPV may be a risk factor for cervical neoplasia. Although further studies using a large number of women with CIN 2 or CIN 3 and considering other latent sexually trans-mitted organisms with no symptoms are required, we suggest that a monitoring programme for the relative distribution of these species in the cervicovaginal system could be helpful for the effective prevention of cervical precancerous lesions.

Transparency declaration

The authors have no conflict of interest.

Acknowledgements

This research was supported by grants from the National Cancer Centre in Korea (1110320, 1310360).

Appendix A. Supplementary data

Supplementary data related to this article can be found athttp:// dx.doi.org/10.1016/j.cmi.2015.02.026

References

[1] Ley RE, Hamady M, Lozupone C, Turnbaugh PJ, Ramey RR, Bircher JS, et al. Evolution of mammals and their gut microbes. Science 2008;320: 1647–51.

[2] Clemente JC, Ursell LK, Parfrey LW, Knight R. The impact of the gut microbiota on human health: an integrative view. Cell 2012;148: 1258–70.

[3] Turner JR. Intestinal mucosal barrier function in health and disease. Nat Rev Imm 2009;9:799–809.

[4] Bultman SJ. Emerging roles of the microbiome in cancer. Carcino-genesis 2014;35:249–55.

[5] Schiffman M, Castle PE, Jeronimo J, Rodriguez AC, Wacholder S. Human papillomavirus and cervical cancer. Lancet 2007;370: 890–907.

[6] Clarke MA, Rodriguez AC, Gage JC, Herrero R, Hildesheim A, Wacholder S, et al. A large, population-based study of age-related associations between vaginal pH and human papillomavirus infection. BMC Infect Dis 2012;12:33.

[7] Taha TE, Hoover DR, Dallabetta GA, Kumwenda NI, Mtimavalye LA, Yang LP, et al. Bacterial vaginosis and disturbances of vaginalflora: association with increased acquisition of HIV. AIDS 1998;12: 1699–706.

[8] Lee JE, Lee S, Lee H, Song YM, Lee K, Han MJ, et al. Association of the vaginal microbiota with human papillomavirus infection in a Korean twin cohort. PloS One 2013;8:e63514.

[9] Hummelen R, Fernandes AD, Macklaim JM, Dickson RJ, Changalucha J, Gloor GB, et al. Deep sequencing of the vaginal microbiota of women with HIV. PloS One 2010;5:e12078.

[10] Hwang JH, Lee JK, Kim TJ, Kim MK. The association between fruit and vegetable consumption and HPV viral load in high-risk HPV-positive women with cervical intraepithelial neoplasia. Cancer Causes Control CCC 2010;21:51–9.

[11] Solomon D, Davey D, Kurman R, Moriarty A, O’Connor D, Prey M, et al. The 2001 Bethesda system: terminology for reporting results of cervical cytology. J Am Med Assoc 2002;287:2114–9.

[12] Jeon YS, Chun J, Kim BS. Identification of household bacterial com-munity and analysis of species shared with human microbiome. Curr Microbiol 2013;67:557–63.

[13] Kim OS, Cho YJ, Lee K, Yoon SH, Kim M, Na H, et al. Introducing eztaxon-e: a prokaryotic 16s rrna gene sequence database with phy-lotypes that represent uncultured species. Int J Syst Evol Microbiol 2012;62:716–21.

[14] Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. Uchime im-proves sensitivity and speed of chimera detection. Bioinformatics 2011;27:2194–200.

[15] Yuan S, Cohen DB, Ravel J, Abdo Z, Forney LJ. Evaluation of methods for the extraction and purification of DNA from the human micro-biome. PloS One 2012;7:e33865.

[16] Menard JP, Fenollar F, Henry M, Bretelle F, Raoult D. Molecular quantification of Gardnerella vaginalis and Atopobium vaginae loads to predict bacterial vaginosis. Clin Infect Dis 2008;47:33–43.

[17] Rothman KJ. Epidemiology. An introduction. New York: Oxford University Press; 2002.

[18] Ravel J, Gajer P, Abdo Z, Schneider GM, Koenig SS, McCulle SL, et al. Vaginal microbiome of reproductive-age women. Proc Natl Acad Sci USA 2011;108(Suppl. 1):4680–7.

[19] Smith BC, McAndrew T, Chen Z, Harari A, Barris DM, Viswanathan S, et al. The cervical microbiome over 7 years and a comparison of methodologies for its characterization. PloS One 2012;7:e40425. [20] Zhou X, Brown CJ, Abdo Z, Davis CC, Hansmann MA, Joyce P, et al.

Differences in the composition of vaginal microbial communities found in healthy caucasian and black women. ISME J 2007;1:121–33.

(9)

[21] Gajer P, Brotman RM, Bai G, Sakamoto J, Schütte UM, Zhong X, et al. Temporal dynamics of the human vaginal microbiota. Sci Transl Med 2012;4:132ra152.

[22] Fredricks DN, Fiedler TL, Marrazzo JM. Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med 2005;353: 1899–911.

[23] Boyle DC, Barton SE, Uthayakumar S, Hay PE, Pollock JW, Steer PJ, et al. Is bacterial vaginosis associated with cervical intraepithelial neoplasia? Int J Gynecol Cancer 2003;13:159–63.

[24] Ludwin I, Ludwin A, Basta A. Influence of vaginal microflora on the presence of persistent atypical squamous cells and atypical glandular cells in pap smear–a 3-year study. Ginekol Polska 2010;81:364–9. [25] Nam KH, Kim YT, Kim SR, Kim SW, Kim JW, Lee MK, et al.

Asso-ciation between bacterial vaginosis and cervical intraepithelial neoplasia. J Gynecol Oncol 2009;20:39–43.

[26] Gillet E, Meys JF, Verstraelen H, Verhelst R, De Sutter P, Temmerman M, et al. Association between bacterial vaginosis and cervical intraepithelial neoplasia: systematic review and meta-analysis. PloS One 2012;7:e45201.

[27] Nam H, Whang K, Lee Y. Analysis of vaginal lactic acid producing bacteria in healthy women. J Microbiol (Seoul, Korea) 2007;45: 515–20.

[28] Jakobsson T, Forsum U. Lactobacillus iners: a marker of changes in the vaginalflora? J Clin Microbiol 2007;45:3145.

[29] Scott M, Stites DP, Moscicki AB. Th1 cytokine patterns in cervical human papillomavirus infection. Clin Diagn Lab Immunol 1999;6:751–5. [30] Libby EK, Pascal KE, Mordechai E, Adelson ME, Trama JP. Atopobium

vaginae triggers an innate immune response in an in vitro model of bacterial vaginosis. Microbe Infect 2008;10:439–46.

[31] Meyer-Hoffert U, Schwarz T, Schroder JM, Glaser R. Expression of human beta-defensin-2 and -3 in verrucae vulgares and condylomata acuminata. J Eur Acad Dermatol Venereol 2008;22:1050–4. [32] Fichorova RN, Buck OR, Yamamoto HS, Fashemi T, Dawood HY,

Fashemi B, et al. The villain team-up or how trichomonas vaginalis and bacterial vaginosis alter innate immunity in concert. Sex Transm Infect 2013;89:460–6.

[33] Behbahani H, Walther-Jallow L, Klareskog E, Baum L, French AL, Patterson BK, et al. Proinflammatory and type 1 cytokine expression in cervical mucosa during HIV-1 and human papillomavirus infection. J AIDS (1999) 2007;45:9–19.

[34] Ma Y, Madupu R, Karaoz U, Nossa CW, Yang L, Yooseph S, et al. Human papillomavirus community in healthy persons, defined by metagenomics analysis of human microbiome project shotgun sequencing data sets. J Virol 2014;88:4786–97.

References

Related documents

The methodology of this research work consist of 5 main levels containing PDF file to image conversion, Text Extraction ,Preprocessing, Feature extraction and

Concerning a search for plant extracts and compounds derived from plants with efficacy as modulators of inflammation, antiviral and anticancer agents, three extracts from

‘‘Dynamic economic emission dispatch using nondominated sorting genetic algorithm-II’’, Electrical Power and Energy Systems, 30, pp.. ‘‘Combined economic emission dispatch

Earnings Yield (TTM) R ev en ue G ro w th (T TM ) -2 5% 10 0% 10% -15% FA VO RA BL E UN FA VO RA BL E GOOGL GOOGL GOOGL GOOGL GOOGL BIDU BIDU BIDU BIDU BIDU YHOO YHOO YHOO YHOO

A successful web-based decision support is in place in PAU, Ludhiana, which use the risk evaluation model to determine the disease severity of late blight in

Since participants are future PG Students, enriching religious freedom with right knowledge certainly helps to resolve religious conflict in particular and Human

They are COFACE (Confedera on of Family Organisa ons in the EU), EASPD (European Associa on of Service Providers for Persons with Disabili es), EDF (European Disability

Both left-wing and right-wing populism, Mouffe argues, aim to reclaim democracy by re-inscribing the distinction between the establishment and “the people.” The two