Ureaplasma parvum infection induces inflammatory changes in vaginal epithelial cells independent of sialidase

Ureaplasma, a genus of the order Mycoplasmatales and commonly grouped with Mycoplasma as genital mycoplasma is one of the most common microbes isolated from women with infection/inflammation-associated preterm labor (PTL). Mycoplasma spp. produce sialidase that cleaves sialic acid from glycans of vaginal mucous membranes and facilitates adherence and invasion of the epithelium by pathobionts, and dysregulated immune response. However, whether Ureaplasma species can induce the production of sialidase is yet to be demonstrated. We examined U. parvum-infected vaginal epithelial cells (VECs) for the production of sialidase and pro-inflammatory cytokines. Immortalized VECs were cultured in appropriate media and treated with U. parvum in a concentration of 1 × 105 DNA copies/ml. After 24 h of treatment, cells and media were harvested. To confirm infection and cell uptake, immunocytochemistry for multi-banded antigen (MBA) was performed. Pro-inflammatory cytokine production and protein analysis for sialidase confirmed pro-labor pathways. Infection of VECs was confirmed by the presence of intracellular MBA. Western blot analysis showed no significant increase in sialidase expression from U. parvum-treated VECs compared to uninfected cells. However, U. parvum infection induced 2-3-fold increased production of GM-CSF (p = 0.03), IL-6 (p = 0.01), and IL-8 (p = 0.01) in VECs compared to controls. U. parvum infection of VECs induced inflammatory imbalance associated with vaginal dysbiosis but did not alter sialidase expression at the cellular level. These data suggest that U. parvum’s pathogenic effect could be propagated by locally produced pro-inflammatory cytokines and, unlike other genital mycoplasmas, may be independent of sialidase.


Introduction
Changes in the microbial composition of the vagina, such that there is a decline in the health-promoting Lactobacillus species and an overgrowth of pathobionts including Gardnerella, Mycoplasma, Ureaplasma, Mobiluncus, Bacteroides, Prevotella, etc., can lead to dysbiosis and infection, commonest of which is bacterial vaginosis (BV). BV is the most common vaginal infection of reproductive-age women worldwide (prevalence rates range from 5 to 70%) and continues to be a major health burden, especially in the pregnant population [1][2][3]. BV is a major risk factor for spontaneous PTB (sPTB) as it facilitates ascending intrauterine infection and inflammation leading to preterm prelabour rupture of membranes (PPROM) and preterm labor (PTL) [2].
Ureaplasma, a genus of the order Mycoplasmatales commonly grouped with Mycoplasma as genital mycoplasma, is a common inhabitant of the vagina of reproductive-age women [4]. Like other potentially pathogenic anaerobes, its pathogenicity is inhibited by the acidity created by lactic acid and other antimicrobial substances produced by Lactobacillus species [2]. Mycoplasmas have been identified as markers or symbionts of BV microbiota [4] and independent predictors of PTB in women with PTL and/or short cervix [5]. Vaginal colonization with U. parvum, in particular, is associated with BV [4], acute chorioamnionitis [6], and late abortion or early preterm birth (PTB) [7]. U. parvum and U. urealyticum are two of the most common microbial species isolated from the amniotic cavity of women with PTL and intact membranes [8][9][10][11][12][13][14]. However, the mechanism of U. parvum ascension into the amniotic cavity is still unknown.
BV-associated bacteria can evade the host immune clearance and sometimes antibiotic therapy. They do this by employing several virulence factors, particularly sialidase, whose pathogenic effects are supported by other hydrolytic enzymes, haemolytic toxins and immunomodulatory metabolites [3]. Sialidase is a hydrolytic enzyme that cleaves sialic acid from glycans in vaginal mucous membranes [15,16]. Sialic acid present in the female genital tract plays a part in preventing adherence of pathobionts such as Mycoplasma spp., Ureaplasma spp., Gardnerella vaginalis etc. to the underlying epithelium [3]. Sialic acid blocks adherence of U. urealyticum to epithelial cells, perhaps through its recognition of sialic acid-containing receptors [17]. However, this action may be species-or serotype-specific [17] and has not been demonstrated with U. parvum. Therefore, cleavage of sialic acid by sialidase facilitates bacterial adhesion, invasion of vaginal epithelial cells (VECs), formation of biofilms and inflammation [18][19][20]. Biofilms create antibiotic resistance and facilitate relapse and recurrence of BV [18, 21-23] that can be injurious to pregnancy leading to sPTB [1,[24][25][26][27][28]. Although immunoglobulin A (IgA) protease, urease and phospholipase A1, A2 and C activities have been reported in ureaplasmas [29][30][31][32][33], whether Ureaplasma species can induce the production of sialidase is yet to be demonstrated.
Bacterial-induced production of cellular sialidase is hypothesized to promote ascending infection across the cervicovaginal tissue in multiple ways: (1) breakdown of the vaginal epithelial barrier, (2) local production of inflammation, and (3) migration of vaginal epithelial cells that can propagate inflammation. More so, whether sialidase is involved in Ureaplasma-induced infection and inflammation that lead to sPTB has not been proven. Therefore, to answer these questions, we investigated the production of sialidase and pro-inflammatory cytokines by U. parvuminfected vaginal epithelial cells (VECs).

Cell culture
Immortalized vaginal epithelial cells (VECs) [34,35] stored in liquid nitrogen were thawed in a water bath for 3 min and incubated for 48 h in 5% CO 2 at 37 o C in a T25 flask containing 5 ml of a mixture of Keratinocyte Growth Medium-2 (KGM-2) Bulletkit (CC-3107, Lonza) and Gib-coTM Keratinocyte Serum-Free Growth Medium (KSFM, 1X) (37010022, Fisher) complete media in equal amounts (1:1) [35]. KSFM complete media was made by adding 1 ml of 22.25 mg/ml stock of sterilized CaCl 2 , 1 ml of primocin (Invitrogen, Waltham, MA, USA, Cat. No: cat-pm-1, 50 mg/mL, Carlsbad, CA, US) and supplemented with bovine pituitary extract and human recombinant epidermal growth factor (rEGF) [35]. The media was changed every 48 h and after 7 days the cells were transferred from T25 to T75. The cells were split into three T75 flasks (i.e., N1-1 control, 1 treatment and 1 stock) at 1:3 ratio after attaining > 80% confluency. The treatment sample consisted of VECs infected with U. parvum, while the Control sample consisted of VECs without infection. The stock VECs were preserved for subsequent experiments.

Infection of VECs with U. parvum
After 14 h of incubation, 1 × 10 5 DNA copies/ml of U. parvum was added to VEC media. U. parvum in media was then transferred into T75 containing VECs, swirled gently over the cells and incubated separately in 5% CO 2 at 37 o C for 24 h. The negative control (VECs not infected with U. parvum) was also incubated under the same condition but in a different incubator to avoid cross-contamination (N = 4).
Approximately 30,000 VECs were cultured on 8-well coverslips (CELLTREAT Scientific Products, Pepperell, MA, USA, Cat. No: 229168) for immunocytochemistry in a total N = 4 for each group. U. parvum at 1 × 10 5 DNA copies/ mL concentration was added the next day in the treatment wells, and cell culture media only was used in the negative control wells. Treated cells were incubated at 37 °C, 5% CO 2 and 95% air humidity for 24 h.

Cell lysis and protein extraction
The cell morphology was confirmed under the microscope and the media was collected from the treated cells with a pipette into a 50 ml tube and stored immediately at -80 0 C. The cells in the flask were washed with 2 ml phosphatebuffered saline (PBS, Corning, Cat. No: 21-040) and 700 µl of radioimmunoprecipitation assay (RIPA) buffer from a stock solution of 1 ml RIPA, supplemented with 10 µl phosphatase inhibitor cocktail, 10 µl protease inhibitor cocktail, and 10 µl phenylmethylsulfonyl fluoride (PMSF) and incubated on ice for 10 min. After incubation, the cells were scraped from the surface of the flask into solution using a scrapper. The cells in the buffer were then collected into an eppendorf tube and vortexed for 30 s, sonicated for another 30 s and vortexed briefly again. This was followed by incubation at RT for another 10 min, brief vortexing and storage at -80 o C. The media from the control sample was also collected and stored at -80 o C for comparison with the treated sample. Four biological replicates (N1-N4) were produced using the same experimental (infection) protocol.

Multiple-banded antigen
After 24 h of treatment, the media was removed and cells were fixed. After fixing, the cells were blocked with 3% BSA/PBS for 30 min at RT, and then incubated with primary antibody for multi-banded antigen (MBA) -U. parvum (Invitrogen, Cat. No: MA5-17010, DF: 1:200) overnight at 4 o C. The next day, the cells were washed three times for 10 min with 1 x PBS and incubated with Alexa Fluor secondary antibody (Alexa Fluor® 594, Invitrogen, Cat. No: ab150080) diluted in PBS for 1 h at RT. After that, the cells were stained with DAPI (4′,6-diamidino-2-phenylindole, Invitrogen), washed three times for 10 min and then mounted using Mowiol mounting medium (475904-100GM-M; Sigma-Aldrich).

Microscopy and image analysis
Bright-field microscopy images were captured using a Nikon Eclipse TS100 microscope (×4, ×10, ×20) (Nikon). Three regions of interest per condition were used to determine the overall cell morphology. A Keyence All-in-one Fluorescence BZ-X810 microscope, ×4, ×10, and ×40 magnification) was used to determine MBA expression. IMARIS (Version 9.7.2, Bitplane) 3-dimensional imaging software was used to localize red MBA staining within the VECs.

Western blot
Frozen culture media from the control and infected VECs were thawed and centrifuged at 9600 x g for 20 min at 4 o C. The supernatant was aspirated into a new tube and placed on ice ready for total protein quantification by BCA (Bicinchoninic acid) protein assay using the Pierce BCA protein assay kit (23225, Thermoscientific).

Gel electrophoresis
10 µg of total protein was combined with 10X RIPA lysis buffer and sample buffer containing beta-mercaptoethanol (reducing agent) and 4x Laemmli sample buffer (1610747, Bio-Rad) to give a total volume of 50 µl. The samples were heated at 101 o C for 5 min and then allowed to cool on ice before loading on to the gel. The samples were run on the gel for 40 min at 120 V along with a molecular weight marker using a 10x Tris/Glycine/SDS running buffer. After electrophoresis, the protein bands were transferred (blotted) to the PVDF membrane according to the instructions for using the Bio-Rad Trans-Blot Turbo Transfer System RTA transfer kits.

Antibody incubation
Non-specific binding was blocked by placing the membrane in 5% milk in TBST (tris buffered saline containing 0.1%Tween 20) on a shaker for 1-2 h. Primary antibodies anti-Neuraminidase antibody [EPR15712] (ab197020, abcam) (1:2000) and anti-actin antibody (ab8226) (1:15,000) diluted in 5% milk/TBST was added and incubated on the shaker at 4°C overnight. After incubation, the membrane was washed 3 times with TBST for 10 min per wash while shaking. A secondary antibody diluted in 5% milk/TBST (1:10,000) was added and incubated for 1 h at RT while shaking. Again, the membrane was washed 3 times with TBST for 10 min per wash while shaking.

Visualization of U. parvum in VECs
To determine the downstream effects of U. parvum infection in VECs, we first confirmed the localization and infection of VECs in culture. VECs were grown to confluence under standard cell culture conditions and then treated with 1 × 10 5 DNA copies/ml of U. parvum for 24 h. To confirm intracellular infection, immunocytochemistry of VECs with MBA antibody, a specific marker for U. parvum, was performed and counterstained with DAPI. The presence of MBA within VECs cytoplasm was visualized by fluorescence microscopy ( Fig. 1 A). Using Z-stack images processed with IMARIS 3D software, MBA was visualized in a 3D space (Fig. 1B) and localized to the cytoplasm and adjacent to the nucleus ( Fig. 1 C). The identification of MBA within VECs confirmed intracellular infection after 24 h of treatment.

U. parvum infection did not induce sialidase expression in VECs
After confirmation of infection, downstream pro-labor signalling pathways were analysed in U. parvum-treated VECs. Western blot analysis of cell lysate collected from both infected and uninfected cells produced low levels of sialidase (Fig. 2 A). U. parvum-treated VECs did not show a significant increase in sialidase expression compared to uninfected cells (Fig. 2B).

U. parvum infection induces pro-inflammatory cytokine production by VECs
Changes in the microbial composition of the vagina or dysregulation of pro-inflammatory cytokines can cause immune dysfunction within the vaginal tract, which is a risk factor for sPTB. To understand if U. parvum plays a role in the induction of these pro-labor pathways, pro-and anti-inflammatory cytokines were measured in supernatants from treated or untreated VECs. Multiplex cytokine assays showed U. parvum infection induced the production of GM-CSF (p = 0.03), IL-6 (p = 0.01), and IL-8 (p = 0.01) in VECs compared to controls (Fig. 3). IL-1β was expressed at low levels under both conditions but did not change after treatment (Fig. 3). Anti-inflammatory cytokine IL-10 and pro-inflammatory cytokine Tumor Necrosis Factor Alpha (TNF-α) were also measured though not detected in cell supernatants (data not shown). These data suggest an association between U. parvum infection and pro-inflammatory response in VECs.

Detection/visualization
The protein bands were visualized on the Bio-Rad Chemi-Doc™ imaging system after Clarity Western ECL substrate or Clarity Max Western ECL substrate was added to the blots on the membrane. Following the instructions on the BioRad imager, we obtained and merged colorimetric images of the protein of interest and the membrane to get the reference molecular weight marker included in the image. After initial visualization, the blot stripping buffer was added for 15 min and washed three times with TBST for 10 min per wash while shaking. Non-specific binding was blocked again by placing the membrane in 5% milk in TBST on a shaker for 1-2 h. Bands were analysed using ImageLab software.

Cytokine analysis
Media collected from VEC culture flasks after treatment (control or U. parvum) were used to measure the following inflammatory markers: granulocyte-macrophage colonystimulating factor (GM-CSF), interleukin 1-β (IL-1β), IL-6, IL-10, Tumor Necrosis Factor alpha (TNF-α), and IL-8. Milliplex MAP panel and detection kit (EMD Millipore Corporation, Billerica, MA, USA) were used according to the kit protocol. Briefly, the plate was washed by shaking in a wash buffer for 10 min and then decanted. Standards, controls, samples, assay buffer, and magnetic beads were added to the plate. The plate was sealed and incubated with agitation on a plate shaker overnight at 4 o C in the dark. The plate contents were removed and washed three times with a wash buffer. 1X Milliplex MAP detection antibody was added to the plate and incubated on a shaker for 1 h at room temperature (20-25 o C) in the dark; 1X streptavidin-phycoerythrin was added and incubated on a plate shaker for 30 min in the dark. After aspirating the plate content and washing for three times, an amplification buffer was then added. The plate was run on a Luminex 200 (LX200-XPON-IVD, Luminex Corporation, Austin, TX, USA) apparatus. Collected data were analysed.

Statistical analyses
All data were analysed using GraphPad Prism 9.3.1 (350) software (GraphPad Software, La Jolla, CA, USA). Student's t-test was used to compare results with two means. Ordinary one-way analysis of variance followed by the Tukey's multiple comparison test was used to compare normally distributed data with at least three means. The Kruskal-Wallis test with Dunn's multiple comparison test were used for data that were not normally distributed. Asterisks denote p values as follows: * p < 0.05; ** p < 0.01; *** p < 0.001, *** p < 0.0001. imbalance in the lower genital tract that could lead to poor reproductive outcomes.
Increased sialidase expression has been shown to facilitate bacterial adhesion, invasion, formation of biofilms, and dysregulated immune response [3]. Sialidase is central to the dysregulated immune response observed in BV which appear to underpin the strong association between BV and spontaneous preterm birth (sPTB) [3]. Along with other virulence factors, sialidase is produced by several BV-associated bacteria including Mycoplasma spp. [3,20]. Though sialidase-mediated pathogenic pathways have been documented for other colonizers of the vagina [3,20], it is

Discussion
In this in vitro study, we demonstrated infection of VECs by U. parvum confirmed by intracytoplasmic localization of U. parvum's MBA after 24 h of treatment compared to uninfected VECs which did not show any trace of the antigen. The infected VECs responded by producing significantly higher amounts of pro-inflammatory cytokines, including IL-6, IL-8, and GM-CSF. However, the relatively low levels of sialidase produced by the infected and uninfected cells did not differ significantly between both groups. These data suggest that U. parvum could induce inflammatory   [30,39,40]. Subsequent experiments are required to confirm the concentration of sialidase that can weaken the VEC barrier function and whether sialidase is propagated through EVs.
Furthermore, because sialidase is implicated in the immunosuppression that underpins the absence of overt inflammation in BV [3], and the sialidase level in the current study was not significantly different in the infected and uninfected cells, it could be that the pro-inflammatory state we observed was induced by MBA (as we detected) and other virulence factors [30,40] of U. parvum. MBA is a surface-exposed virulence (immunodominant) lipoprotein of Ureaplasma spp. [40][41][42][43][44], the sizes of which modulate host immune response to the bacterial infection [45]. The MBA activates NF-κB and the production of cytokines by Toll-like receptors (TLRs) 1, 2, and 6 signalling [40,[42][43][44], while the whole bacterium is required to activate TLR9 [43].
As shown in a mouse model [40], U. parvum also expresses a diacylated lipopeptide (UPM-1) that can initiate inflammation by upregulating IL-1β, IL-6, IL-12p35, TNF-α, macrophage inflammatory protein 2 (MIP2), LPSinduced CXC chemokine (LIX), and inducible nitric oxide synthase (iNOS) and cause PTB at high concentrations [40]. Another mouse model indicated chemical damage to the cervical epithelium facilitated intrauterine U. parvum infection, upregulated pro-inflammatory cytokines, and increased PTB rates by more than 2-fold [46]. The above reports give strong evidence that Ureaplasma alone can cause PTB in a controlled experimental fashion. Though we did not study unknown if U. parvum could also play a role in the induction of sialidase-mediated inflammation and pro-labor signalling. Data from this study demonstrated: (1) U. parvum can infect VEC in culture, (2) sialidase expression does not change under U. parvum treatment, and (3) U. parvum infection can cause immune status dysregulation by creating a pro-inflammatory vaginal environment. It can be speculated that U. parvum in the vaginal cavity could induce a localized epithelial cell response to promote ascending infection and associated inflammatory changes.
It has been shown that sialidase can be transported and delivered to host cells by cervicovaginal mucus-derived extracellular vesicles (EVs, 100-200 nm in diameter) [37,38]. The implication of these plausible EV-mediated intercellular communication and involvement of common vaginal anaerobes such as U. parvum to the incidence of infection/inflammation-associated sPTB is still unclear. Here, we did not observe a significant difference in sialidase level between U. parvum-infected vs. uninfected VECs. That is, sialidase did not change in VECs colonized by U. parvum. This is suggesting that sialidase is unlikely a mediator of U. parvum's pathogenicity unlike other BV pathogens. However, whether sialidase is packaged in EVs to the same degree in both infected and uninfected VECs is still to be elucidated. Quantitative measurement of exosomal sialidase needs to be conducted as our preliminary western blot analysis showed sialidase in control and infected VECs. Another possibility is that, the pathogenicity of sialidase could be potentiated by the pro-inflammatory role of MBA as in our model of U. parvum infection and/ of the current study. Subsequent experiments can also test whether U. parvum, via sialidase activity, can induce epithelial exfoliation and epithelial-mesenchymal transformation and promote ascending infection and inflammation similar to Group B Streptococcus [52,53].

Conclusion
In conclusion, experimental U. parvum infection of VECs induced an increased production of IL-6, IL-8, and GM-CSF. This pro-inflammatory state was associated with the presence of multi-banded antigen, a virulence factor specific to U. parvum. However, there was no corresponding increase in sialidase due to the infection, indicating that U. parvum may not be an important pathogen in the cervicovaginal environment. There is need for more in vitro and in vivo studies to clarify the role of vaginal U. parvum infection during pregnancy.

Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s11033-022-08183-6. the expression of UPM-1, these reports support the 2-3-fold increased production of IL-6, IL-8, and GM-CSF in the infected VECs we observed, which are pro-inflammatory mediators that can be transported by EVs [47].
Our current in vitro model somewhat substantiates the findings of our recent systematic review on genital mycoplasma [48] where we showed that U. parvum is not a great pathogen in the cervicovaginal environment unlike other virulent species (e.g., Gardnerella vaginalis). And unless the reproductive tract is damaged as in Pavlidis et al. [46], the impact of this bacteria is going to be minimal and may not be more than a localized inflammatory response. Furthermore, vaginal U. parvum and U. urealyticum infection were not associated with BV symptoms/signs in non-pregnant women [49]. Another recent study concluded that cervicovaginal Mycoplasma infection (including Mycoplasma hominis and Ureaplasma spp.) was not a risk factor for sPTB, even in the presence of other vaginal pathogens [50]. The relationship of genital mycoplasmas and adverse pregnancy and birth outcomes in the presence or absence of BV or location of infection in the genital tract is still debatable [48,51] and warrants further investigation. However, we have also previously shown that U. urealyticum infection of HeLa cells induced greater cytotoxicity and release of higher amounts of ammonia and IL-8 within 24 h compared to L. crispatus [39]. Release of ammonia through urea hydrolysis is another virulence factor of Ureaplasma spp. [30,39]. Therefore, we propose that U. parvum can cause cellular irritation, localized inflammation but is incapable of producing classic BV like antigenic factors such as sialidase and prolidase required for ascending infection.
The current study is perhaps the first to determine sialidase expression in U. parvum-infected VECs and has uncovered several areas for further study. The sample sizes, especially for the quantification of sialidase levels, were relatively low. The intracellular MBA localization was not quantitative, hence, we could not ascertained if the multiplicity of infection (MOI) was sufficient to determine whether U. parvum infection induced sialidase upregulation. We were also unable to determine whether the cytokine response was directly related to cells with intracellular MBA or from bystander cells (i.e. without intracellular MBA). However, we previously demonstrated significant increase in IL-8 production by U. urealyticum infected HeLa cells after 24 h with an MOI of 3640 CFU/ml [39]. There is also the need to isolate and characterize EVs, confirm the presence of tetraspanin markers to determine the purity of exosomes from uninfected and infected VECs; as well as the presence of sialidase and other virulence factors in EVs derived from VECs infected with U. parvum. Future experiments on a larger sample size with appropriate MOI should validate and give more credence to the findings guests? Glycobiology 31 (6)