Aerobic vaginitis is associated with carbonic anhydrase IX in cervical intraepithelial neoplasia

The aim of this study was to analyze the association between vaginal microbiota, carbonic anhydrase IX (CAIX) and histological findings of cervical intraepithelial neoplasia (CIN). The study included 132 females, among them 66 were diagnosed with high-grade intraepithelial lesion (CIN2, CIN3, and cancer), 14 with low-grade disease, and 52 assigned to the control group. An interview focused on the behavior risk factors, together with vaginal fluid pH measurement, wet mount microscopy, detection of Chlamydia trachomatis, and Trichomonas vaginalis were performed. After colposcopy, high-grade abnormalities were detected via direct biopsies and treated with conization procedure. Conuses were immuno-stained with CAIX antibody. The histological findings were CIN1 (n = 14), and CIN2+ (included CIN2 (n = 10), CIN3 (n = 49), and cancer (n = 7; squamous cell carcinomas)). Prevalence of bacterial vaginosis (BV) was similar between the groups. Moderate or severe aerobic vaginitis (msAV) was diagnosed more often among CIN2+ (53.0%) than CIN1 (21.4%). Moderate or strong immunostaining of CAIX (msCAIX) was not detected among CIN1 cases. Thus, msAV was prevalent in CAIX non-stained group (p = 0.049) among CIN2 patients. Co-location of msAV and msCAIX was found in CIN3. Regression model revealed that msAV associated with high-grade cervical intraepithelial neoplasia independently from smoking and the number of partners.


Participants
Women (n = 52) with results being negative for intraepithelial lesion or malignancy (NILM) from triennial Papanicolaou test (PAP smear) for National cervical cancer screening program were enrolled into the control group at Onos Gurevičienės Family Clinic in Marijampole city in the same period.Fourteen women with detected hr-HPV and low-grade intraepithelial neoplasia (LSIL) had confirmative colposcopies with punch biopsies at main university affiliated tertial oncogynecology centers (National Cancer Institute and Kaunas Clinics, Lithuanian University of Health Sciences) and were included in the CIN1 group.
Respondents (n = 66; 23-60-year-old) with detected high grade intraepithelial neoplasm on PAP smears, confirmed in colposcopies with punch biopsies were referred for conization to main university affiliated tertial oncogynecology centers (National Cancer Institute and Kaunas Clinics, Lithuanian University of Health Sciences).
Inclusion criteria for both groups were: reproductive age, ability to provide informed written consent, free from hr-HPV and cervical pathology in control group and confirmed cervical pathology and/or hr-HPV in CIN1, CIN2+ group accordingly.
Exclusion criteria were lack of wet mount microscopy data, pregnancy, earlier diagnosed oncological pathology, endometrial hyperplasia, and human immunodeficiency virus (HIV).

Wet mount smear evaluation
During gynecological exam, unmoistened speculum was used.The vaginal sample was taken with plastic spatula from lower vaginal wall, spread on glass slide through pH strip (range 3.1-7.0;Macherey-Nagel Inc.) 44 .The sample for wet mount microscopy was air dried and transported to the laboratory (Vilnius University Life Sciences Center) 45 .The sample was rehydrated 45 and evaluated under microscopy.Vaginal pH value was considered normal if it was lower than 4.5.
• Inflammation (count of white blood cells and leucocytes with active vacuoles (so called "toxic") per high power field/epithelial cell.• Miscellaneous (erythrocytes, cream, sperm) 46 .
Lactobacilli grade was classified according to Donder's modified classification described by Schroder 47,48  • LGB I-dominance of lactobacilli morphotypes.
Full-blown bacterial vaginosis (BV) was identified if granular microflora dominated and > 20% of clue cells were observed per field.Smears with mixed areas of this type of flora with sporadic "clue cells" and other types of bacteria were classified as partial BV 49 .Aerobic vaginitis (AV) score was calculated following the definition by Donders et al. 14 , considering moderate or severe (msAV) disease ≥ 5.

Colposcopic evaluation
All patients underwent PAP test and hr-HPV evaluation.After virus detection, colposcopy was performed for CIN detection.Following the standard procedure, the colposcopy-directed biopsies were performed in the cases of cervical abnormalities 50 .After application of 5% acetic acid the transformation zone was evaluated.The visual suspicion of CIN in cervical epithelium were biopsied with forceps.

Tissue immunohistochemical staining for CAIX
Samples of punch biopsies and conization material were collected into the labeled containers with 10% buffered formalin and sent to the National Center of Pathology, or Department of Pathological Anatomy, Lithuanian University of Health Sciences.Two pathologists (ED and AC) stained the samples with hematoxylin and eosin, evaluated the samples and made a pathological diagnosis as following: CIN1, CIN2, CIN3, CIS (carcinoma in situ), cancer (squamous cell carcinoma), cervicitis, or normal tissue 51 .
Tissue immunohistochemical staining was performed using monoclonal H7 antibodies for CAIX as previously described 52 .Immunohistochemical staining from the conus of the cervical transformation zone was performed using fully automated BenchMark ULTRA IHC/ISH Staining system (Roche, Basel, Switzerland) at the National Center of Pathology and Department of Pathological Anatomy (LUHS).Immunohistochemical staining was performed on both the formalin-fixed and paraffin-embedded tissue samples.EnVision FLEX, High pH kit was used (Dako, Agilent, K8023, Santa Clara, CA, USA) with purified MAb H7 52 (0.02 mg/ml) as well as hematoxylin (Merck, 75290, Burlington, MA, USA) nuclear staining.Specific immunohistochemical staining was defined by the presence of a brown reaction product on cell membrane under × 40 magnification.Absent membrane staining or faint staining of the cytoplasm was considered negative.Pathologists (ED and AC) confirmed the diagnosis and evaluated immunostaining in atypical cells.They reported staining intensity (negative, weak, moderate, and strong), and the proportion of stained cells following Woelberg 24 .Stained tissues were photographed using ScanScope XT system (Aperio, Leica Biosystems Inc., Bualo Grove, IL, USA).

hr-HPV DNA detection by genotyping
During first visit biological material from the cervix was obtained with cervical brush (Rovers medical devices, Oss, Netherlands) into the "CyMol" medium (Copan, Brescia, Italy) for transportation.Clinical samples were centrifuged at 2000×g for 15 min at 4 °C and hr-HPV DNA was extracted using GeneJet Genomic DNA Purification Kit (Thermo Fisher Scientific, Vilnius, Lithuania) according to the manufacturer's protocol.Genomic DNA was eluted with 100 μl of Elution Buffer.The purified DNA was used for PCR analysis or stored at − 20 °C53 .

Chlamydia trachomatis detection
Clinical samples from cervix were using a brush and transformed to a vial of liquid preservative for the determination of C. trachomatis.The DNA extraction was performed using the GeneJET Genomic DNA Purification Kit (Thermo Fisher Scientific, Lithuania) 54 .

Statistical analysis
Data were analyzed using SPSS 28 software.Results were presented in proportions, means, medians and ranges.Mann-Whitney U test (MWU) was used for comparison of independent samples.Asymptotic exact sign or Fisher's exact test p value was interpreted as statistically significant if p < 0.05.The risk of CIN2+ development depending on various risk factors was calculated as odds ratios.1).
The histological findings in study group followed as CIN1 (n = 14), CIN2 (n = 10), CIN3 (n = 49), and cancer (n = 7; all were cervical squamous cell carcinomas).In the cases when several grades of lesion were detected, the most advantageous lesion was included in the analysis.
Table 2 shows the differences in pH value and microbiota findings among respondents with and without intraepithelial lesion.Lactobacilli grade variation was insignificant between the groups.Vaginal pH that increases in the cases of flora alteration was higher in CIN1+ or CIN2+ that in controls.Women with cervical pathology had more leucocytes, and msAV.Prevalence of bacterial vaginosis (both any (p > 0.05) or just full blow vaginosis (p > 0.05) did not differ between the groups.The main reason was the higher prevalence of msAV among CIN2+ (53.0%) than CIN1 (21.4%) (Table 2).BV cases between the groups variated insignificantly both for any (p > 0.05) or just full blow vaginosis (p > 0.05).
The immunohistochemical staining with CAIX antibodies was performed on surgical specimens (Fig. 2).The reaction was negative or weak for three cases without dysplasia, 8 of LSIL, 53 CIN1+ and 45 CIN2+.Moderate or strong immunohistochemical staining was observed among CIN2+ only in 20 cases.However, co-existence of msAV and msCAIX showed different pattern.

Table 1.
Characteristics of study group participants with CIN1 + and controls (no CIN).*Primary, secondary or vocational education.**Two people did not report their education, thus the number of participants in "No CIN" group was 50 and the total was 130. ***Two persons did not report their smoking status, thus the number of participants in CIN1+ group was 78 and the total was 130. $ MWU p value.www.nature.com/scientificreports/Higher number of partners and smoking were more prevalent in women with CIN2+ (OR 0.94 (95% CI 0.89-0.98),p = 0.01 and OR 7.1 (95% CI 2.7-17.4),p < 0.01) compared to women without cervical dysplasia.Microbiota alteration such as msAV (OR 5.3 (95% CI 2.3-12.8),p < 0.01) was more frequent in study group than controls (Table 3).

Characteristics
Multivariate logistic regression covered analysis of number of partners, smoking and msAV.All of them were independent risk factors for CIN2+ increasing the risk of the disease five times if msAV is diagnosed or smoking is present.

Discussion
This study analyzed the association between vaginal dysbiosis, grade of cervical intraepithelial lesion, and CAIX expression.The moderate or severe dysplasia is usually detected in the mild dysplasia area due to progression 55 , so we considered the most advanced grade of lesion trying to understand co-factors of the process.Results showed that microbiota alteration such as msAV was associated with higher grade intraepithelial lesions (CIN2+), when both smoking and number of partners were addressed.Moreover, msAV was detected together with msCAIX only in more advanced (CIN3) cases.This was not a case among patients with CIN2.
The association between intraepithelial lesion and behavioral risk factors such as smoking or the number of partners have been described in other studies [56][57][58][59] as well as the relation of elevated vaginal pH, abnormal vaginal microbiota diagnosed on wet mount microscopy and cervical precancerous lesions 13,60,61 .Our results were in line with these findings.
Studies show that hr-HPV carcinogenesis is associated with local inflammation 62 , increased number of leucocytes in vaginal fluid and concentration of inflammatory interleukins (IL) 63,64 .Inflammatory microenvironment is genotoxic, stimulates epigenetic changes and DNA damage in the cervix.
Lactobacilli spp.increase adenosine and cytosine levels and reduce inflammation 65 , but they are suppressed in msAV and BV cases.Vaginal microbiome called community state type IV with decreased number of Lactobacilli spp. was associated with CIN2+ persistence after 24 months 39 .However, in this study, community state type IV was not differentiated into subtypes I and II in accordance with prevalence of anaerobic (in BV cases) or aerobic (in AV cases) bacteria.Despite lack of leucocytes, modern metabolomic analysis shows that BV maybe also related to proinflammatory dysbiosis with elevated cervicovaginal cytokines and chemokines, such as IL-1β, IL-6, and IL-8 66,67 .
We did not find differences of lactobacilliary grade between control group and patients with cervical anomaly as well as between low and high grade cervical intraepithelial neoplasia.After comparison of two types of dysbiosis (BV and AV), the proportion of respondents with bacterial vaginosis was similar in all groups.Plisko  12 .Moreover, highly increased concentrations of IL-1-β and IL-6, that promoted CAIX expression 29 , were observed in msAV 15 .Enterococcus (part of msAV flora) was associated with elevation of IL-6 and IL-8 also in Moscicki et al. study 41 .Escherichia coli (that is also found in msAV) produced the 2-hydroxyglutarate (oncometabolite) in colorectal carcinogenesis 68 .
Plisco et al. discussed that BV can be seen as an indirect marker of sexual behavior leading to hr-HPV acquisition rather than be directly involved in the pathogenesis of cervical cancer 12 .Another study found that multiple sexual partners were associated with hr-HPV DNR detection 69 , that could contribute to higher CIN1 rate.
We found that BV was not related to msCAIX immunostaining.The msAV was not prevalent in msCAIX stained CIN2 group but was found among patients with CIN3.Thus, we speculate that presence of msAV and msCAIX staining could be associated with CIN2 progression to CIN3.However, we do not know if microbiota alteration is directly related to increased hypoxia or other mechanisms are affecting inflammatory and hypoxia process during progression of the lesion.The cross-sectional study design was sufficient to find an association but is not sufficient for observation of the lesion progression.Retrospective data about sexual habits and risk factors usually limits the possibility to identify association with intensity of exposition.However, this was not our aim.Microbiota was evaluated, but microbiome and metabolomic analysis wasn't performed and could be seen as limitation of the study as well as a possible direction for further research.
The strength of the study was the use of standard procedures for colposcopy and wet mount microscopy, confirmation of severity of cervical lesions by histologic evaluation of biopsies before treatment and in conuses after surgery, as well as automatized standardized CAIX immunostaining.Thus, this enables comparability between the studies for further meta-analysis.
More studies are needed to understand the relation of abnormal vaginal microbiome and the development of cervical cancer and more aggressive disease in hr-HPV-positive women.Further evaluation of importance of msCAIX and msAV interplay and interrelation with behavioral factors is necessary for drawing the whole picture of interrelations among risk factors.It is also important to clarify whether the treatment of abnormal vaginal microbiota could help prevent the development of cervix cancer or could contribute to cancer prognosis.

Conclusions
The alteration of vaginal microbiota, such as msAV, is associated with more advanced cervical intraepithelial lesions even if behavioral risk factors, such as smoking, and the number of partners is considered.The associated inflammation due to msAV and increased CAIX expression in cervical cells should be taken into consideration when studying vaginal microbiota in women with pre-invasive cervical lesions.

Figure 2 .
Figure 2. Immunohistochemical staining with CAIX on surgical specimens.The first sample was immunohistochemically stained with hematoxylin and eosin (A) and CAIX with negative results (B).The second sample was immunohistochemically stained with hematoxylin and eosin (C) and CAIX with positive results (D).

Table 2 .
Wet mount microscopy findings and pH in accordance with CIN severity groups.Numbers in % are shown in the brackets.CIN cervical intraepithelial neoplasia, LGB lactobacilli grade, hpf high power field, BV bacterial vaginosis, AV aerobic vaginitis, msAV moderate/severe aerobic vaginitis, msCAIX moderate/severe carbonic anhydrase immunostaining.

Table 3 .
Result of univariate and multivariate regression between the risk factors such as smoking, number of partners, and moderate or severe aerobic vaginitis (msAV) and CIN2+ condition.msAV moderate/severe aerobic vaginitis.