In humans, the urinary tract is the second most common site of infection, most frequently by uropathogenic Eschericia coli (UPEC) (Foxman, 2014). Approximately 1% of bladder infections (cystitis) progress to more serious kidney infections (pyelonephritis), and some of these become life-threatening systemic infections (Ki et al., 2004; Jolley et al., 2012). Recurrent urinary tract infections (rUTIs) are very common: 24% of women with an initial UTI will have rUTI within 6 months, and up to 70% will have rUTI within 1 year (Foxman et al., 2000; Foxman, 2014). Approximately 1% of all women (~35 million worldwide) have six or more rUTIs each year (Foxman et al., 2000; Foxman et al., 2002; Foxman, 2014). These women are often given prophylactic antibiotics, contributing to the problem of antibiotic resistance.

It is now appreciated that UPEC has an intracellular niche within the bladder. Multiple studies of UTI and rUTI in adults and children have identified intracellular UPEC, both in bladder epithelial cells shed in urine and in bladder biopsies (Elliott et al., 1985; Rosen et al., 2007; Robino et al., 2014; Liu et al., 2016; De Nisco et al., 2019). One study using confocal microscopy detected intracellular bacteria in 36.8% of samples from children with acute UTI. Notably, a medical chart review demonstrated that the children with intracellular bacteria were significantly more likely to have a history of rUTI (OR, 8.0; 95% CI, 2.3-27.4) (Robino et al., 2014). In C57BL/6 mice, UPEC can persist within bladder epithelial cells for months after inoculation, without bacteriuria (bacteria in urine) and despite antibiotic treatment (Mulvey et al., 2001; Schilling et al., 2002; Eto et al., 2006; Mysorekar and Hultgren, 2006). In patients, most rUTIs (as many as 82%) are caused by a UPEC strain identical to that of the previous infection (Ejrnaes et al., 2006; Czaja et al., 2009; Luo et al., 2012; Skjot-Rasmussen et al., 2013; Koljalg et al., 2014), even when appropriate antibiotic therapy is given. This could reflect reintroduction of UPEC to the bladder from the gut or vaginal reservoir, but is also consistent with the concept that rUTI can be caused by emergence of the initially infecting UPEC strain from an intracellular bladder reservoir (Mulvey et al., 2001; Hickling and Nitti, 2013; Glover et al., 2014). The potential for bladder reservoirs to seed rUTI has prompted preclinical investigations into strategies to eliminate bladder reservoirs as a means of rUTI. Gardnerella strain JCP8151B causes epithelial exfoliation in the vagina in a mouse model (Gilbert et al., 2013). We hypothesized that Gardnerella could likewise cause bladder epithelial (urothelial) exfoliation and thus be a trigger of UPEC rUTI from bladder reservoirs in women. This hypothesis was further supported by three clinical observations potentially linking Gardnerella with UTI. First, women with bacterial vaginosis, a vaginal dysbiosis in which the vagina is overpopulated by a polymicrobial mixture of bacteria including Gardnerella, have an up to 13-fold higher likelihood of UTI than those with Lactobacillus-dominated vaginal microbiotas (OR 2.21-13.75) (Harmanli et al., 2000; Hillebrand et al., 2002; Sharami et al., 2007; Sumati and Saritha, 2009; Amatya et al., 2013). Second, women with rUTI who received vaginal interventions that influence the vaginal microbiota (e.g., vaginal estrogen, probiotic intravaginal Lactobacillus crispatus) experienced fewer rUTIs than those who received placebo (Stapleton et al., 2011; Rahn et al., 2014; Sadahira et al., 2021). Third, multiple urinary microbiome studies have detected Gardnerella in urine collected directly from the bladder (via suprapubic aspiration or catheterization), suggesting that Gardnerella routinely gains access to the bladder in some women (Wolfe et al., 2012; Pearce et al., 2014; Pearce et al., 2015; Thomas-White et al., 2016; Gottschick et al., 2017). Intriguingly, Gardnerella can co-aggregate with UPEC in an in vitro biofilm assay (Castro et al., 2016), suggesting possible synergy between these bacterial species.

To test our hypothesis, we developed a mouse model to determine whether Gardnerella could cause exfoliation in the bladder and UPEC emergence from bladder reservoirs (Gilbert et al., 2017; O’Brien et al., 2020). We found that in mice with intracellular UPEC reservoirs in the bladder, two transurethral Gardnerella inoculations triggered UPEC emergence from reservoirs, leading to rUTI. Gardnerella exposure caused urothelial apoptosis and exfoliation, which likely enabled UPEC to emerge from reservoirs. Notably, Gardnerella did not stably colonize the bladder, demonstrating that a brief exposure was sufficient to elicit UTI pathogenesis. Here we sought to further characterize the effect of Gardnerella bladder exposures in order to identify host processes involved in UPEC rUTI that could serve as potential biomarkers or therapeutic targets. We performed RNA sequencing (RNA-seq) on bladders harboring intracellular UPEC reservoirs, with and without exposure to Gardnerella. We found that host gene expression changes following Gardnerella exposure were modest, in keeping with a bacterial exposure that does not result in stable colonization. However, gene set enrichment analyses revealed upregulation of many host pathways. Some of the pathways, such as apoptosis and inflammatory cytokines, corroborated the phenotypic results we previously reported in this model, and other pathways were generally related to UTI. Five genes in the Immediate Early response pathway were significantly upregulated, including Nur77 (aka Nr4a1). Notably, Nur77^-/- mice were protected from Gardnerella-induced rUTI, confirming its role in this model.

Millions of women suffer from recurrent UPEC UTI (Foxman et al., 2000; Foxman, 2014), most commonly caused by the same strain that caused the initial infection (Ejrnaes et al., 2006; Czaja et al., 2009; Skjot-Rasmussen et al., 2013; Koljalg et al., 2014). At least some instances of same-strain rUTI are likely due to UPEC emergence from protected bladder intracellular reservoirs (Mulvey et al., 2001; Hickling and Nitti, 2013; Glover et al., 2014). We and others have demonstrated that UPEC can persist intracellularly in the mouse bladder for weeks or months, and can emerge and elicit rUTI after bladder exfoliation is experimentally induced (Mulvey et al., 2001; Schilling et al., 2002; Eto et al., 2006; Mysorekar and Hultgren, 2006). In our model, we use two Gardnerella exposures as the trigger for bladder exfoliation and rUTI (Gilbert et al., 2017; O’Brien et al., 2020). In the present study we demonstrate that Gardnerella exposures in mice containing UPEC reservoirs affected the expression of host pathways related to urothelial exfoliation and regeneration, mucosal inflammation and immunity, and other processes. Immediate Early (IE) genes were upregulated following one Gardnerella exposure. The orphan nuclear receptor Nur77 is a key IE gene (Yoon et al., 2011; Gao et al., 2016), and we found that whole-body Nur77 knockout mice were protected from recurrent UPEC UTI following Gardnerella exposure – that is to say, the rate of rUTI was not increased in Nur77 ^-/- mice exposed to Gardnerella compared to those mice exposed to PBS. Thus, the IE response may contribute to Gardnerella-induced recurrent UPEC UTI.

Gardnerella has frequently been isolated as the most abundant organism in “urinary microbiome” studies, including those using expanded culture techniques to detect live bacteria (Wolfe et al., 2012; Pearce et al., 2014; Hilt et al., 2014; Pearce et al., 2015; Karstens et al., 2016; Wu et al., 2017). Importantly, these studies collected urine by catheterization or suprapubic needle aspiration to limit vaginal contamination of urine specimens, supporting the conclusion that Gardnerella can indeed be found within the bladder. In a recent longitudinal culturomics study of healthy women, Gardnerella was often present at high relative abundance in urine sample pairs collected from the same woman at different time points (Ksiezarek et al., 2021). Another large study found hospital inpatients with G. vaginalis in their urine were more likely to have a history of rUTI than patients in whom G. vaginalis was not detected (Josephson et al., 1988). Although these clinical studies make it clear that Gardnerella can be found in the bladder and urine, it is not yet known whether Gardnerella stably colonizes the urinary tract, vs. being repeatedly re-introduced into the bladder by hygienic and/or sexual behaviors. We have shown that Gardnerella does not stably colonize the mouse bladder and is cleared within 12 h of exposure, but nonetheless triggers exfoliation and UPEC rUTI (Gilbert et al., 2017). Here we detected significant changes in expression of several pathways even relatively early after a single Gardnerella exposure; more substantial changes occurred after two Gardnerella exposures as indicated by significant changes in GO terms and KEGG pathways. These findings further support a model in which repeated introduction of Gardnerella into the bladder (which may occur in women after sexual activity) can drive host responses and rUTI phenotypes. Thus, our study provides further support for “covert pathogenesis,” the notion that a brief exposure to a microbe (here, Gardnerella) can affect the host enough to drive disease attributed to another microbe (here, E. coli) (Gilbert and Lewis, 2019).

Several pathways identified in this RNA-seq analysis are corroborated by phenotypic observations that we previously reported in this mouse exposure model. As would be expected following bacterial exposure, here we observed changes in many host pathways and processes related to the mucosal immune response. The conclusion that Gardnerella exposures trigger an inflammatory response in bladders harboring UPEC reservoirs is further supported by our previous finding of higher levels of IL-12, IFN-g, and RANTES in bladder homogenates (Gilbert et al., 2017). Interestingly, pathways related to T and B cells were upregulated after the second Gardnerella exposure. This observation could implicate adaptive immunity generated by the initial UPEC infection as a contributor to UPEC reservoir emergence. A previous study showed that an adaptive immune response is necessary to clear a challenge UPEC infection that was introduced by a second UPEC inoculation two weeks after a primary UPEC infection in C57BL/6 mice (Mora-Bau et al., 2015). Other studies have demonstrated that bladder inflammation differs between first and second UPEC exposures and that severe bladder inflammation can impact UTI outcomes (O’Brien et al., 2018; Yu et al., 2019). More work is needed to understand bladder immune responses to Gardnerella exposure and UPEC reservoir emergence. The RNA-seq pathway analysis also implicated urothelial integrity and turnover, as well as apoptosis. Our previous studies corroborate these findings, as we have shown that our model strain, Gardnerella JCP8151B (also used in the present study), elicits both vaginal (Gilbert et al., 2013; Gilbert et al., 2019) and urothelial exfoliation (Gilbert et al., 2017). In the same urinary tract exposure model used here, mice exposed to Gardnerella displayed evidence of apoptosis in the urothelium. Compared to mice exposed to PBS, those exposed to Gardnerella had increased urothelial TUNEL staining and cleaved Caspase-3 12 h after two Gardnerella exposures (Gilbert et al., 2017). Thus, the RNA-seq findings of pathways related to inflammation and urothelial integrity echo the biological phenotypes we previously reported in the mouse model and the findings from our whole organ analysis support the relevance of RNA-seq methods for identifying biologically relevant host responses to Gardnerella bladder exposures.

The model and data presented here have some limitations. Although whole bladder RNA-seq is an established method for investigating host response during UTI (O’Brien et al., 2016; Yu et al., 2019), transcriptional changes cannot be attributed to specific cell types, but rather reflect changes that occurred at the organ level. As well, using the whole organ for bulk RNA-seq may mask gene expression changes that only occur in a small subset of cells, which could be one explanation for the relatively modest number of genes for which expression changes after Gardnerella exposure were statistically significant following FDR correction. Nonetheless, the RNA-seq data corroborated previously reported phenotypic findings in this model (apoptosis, inflammatory cytokines) and identified a host gene (Nur77) that we subsequently found to be necessary for Gardnerella-induced rUTI. These findings support the utility of RNA-seq methods for assessing host responses to even transient bladder microbial exposures and warrant future single-cell transcriptomic studies to specifically examine urothelial, immune, and other bladder cell responses.

The model presented here involves multiple exposures to two different microbes that could independently or synergistically stimulate host responses over time. While this likely reflects the situation occurring in women, it does present some challenges for interpretation. A caveat of the current dataset is that it cannot distinguish changes that required a second Gardnerella exposure from changes that resulted from the first exposure, but required additional time to become apparent. However, since we previously reported that a single Gardnerella exposure was rapidly cleared and did not cause urothelial exfoliation or UPEC rUTI, it is very likely that most of the pathway upregulation observed here was due to the second exposure, but this should be determined in future studies. Our finding that host gene expression pathways were primarily downregulated after one Gardnerella exposure and upregulated after two exposures could suggest that the second time point reflects a host genetic signature of UPEC reservoir reactivation. Additionally, since all of the mice used in the present study harbored UPEC reservoirs (because of our interest in rUTI), we cannot distinguish changes that were due specifically and solely to Gardnerella from changes that resulted from “re-awakening” of UPEC reservoirs. Relatively little is understood with respect to how UPEC emerge from quiescent intracellular reservoirs and what host processes are involved. Furthermore, UPEC infection has a lasting effect on the bladder mucosa that is likely to also impact how the bladder responds to subsequent exposures. Consistent with this idea, we previously reported that the bladder cytokine changes caused by Gardnerella were distinct between naive mice and those with UPEC reservoirs (Gilbert et al., 2017). Future studies aimed at distinguishing between these various possibilities could examine the bladder transcriptome following Gardnerella exposure in naive mice and following induction of UPEC rUTI from reservoirs by other means (e.g. protamine sulfate (Mysorekar and Hultgren, 2006) or chitosan (Blango et al., 2014)).

Our RNA-seq data implicated the nuclear receptor Nur77 (aka Nr4a1) as an early responder to Gardnerella exposure. Nur77 regulates myriad cellular processes, that intersect with the UPEC UTI pathogenic cascade and could influence rUTI outcomes. For example, Nur77 regulates apoptosis in multiple tissue types (Rajpal et al., 2003; Herring et al., 2019), and we previously found that Gardnerella induced bladder exfoliation via apoptosis (Gilbert et al., 2017). Whether or not Nur77 drives exfoliation in the Gardnerella-exposed urothelium remains to be determined. Nur77 also modulates inflammation (Rodriguez-Calvo et al., 2017) and has been specifically implicated in T cell responses (Liebmann et al., 2018) and Ly6C- monocytes (Hanna et al., 2011). Nur77 modulated inflammatory responses to E. coli in the lung during pneumonia in mice (Cui et al., 2019), but the role of Nur77 in mediating bladder responses to UPEC has not been examined. In the present study we did not conduct an in-depth characterization of acute UPEC UTI in Nur77 ^-/- mice. However, the first phase of our rUTI model entails UPEC infection and monitoring of urine bacterial titers over time during the initial infection phase, prior to secondary exposures. We observed no differences between wild type and Nur77 ^-/- mice in the initial UPEC infection phase of the model. As well, we found that Nur77 ^-/- mice did harbor stable bladder UPEC reservoirs. Thus, Nur77 may be dispensable for initial UPEC bladder infection in C57BL/6 mice, though we did not assess chronic bladder or kidney infection or other mouse UTI phenotypes. Strikingly, we found that Nur77 ^-/- mice were protected from Gardnerella-induced recurrent UPEC UTI, as evidenced by no difference in the incidence of post-exposure bacteriuria or remaining bladder reservoir titers between Gardnerella-exposed vs. PBS-exposed Nur77 ^-/- mice. In this pilot study, the wild type and Nur77 ^-/- mice were not littermates. Future studies will directly compare Nur77 ^-/- mice to Nur77 ^+/- and wild type littermate controls to validate our rUTI findings. Nur77 is a druggable target, with several ligands being explored for treatment of diseases such as cancer, metabolic disorders, hyperinflammatory responses and endometriosis (Wu and Chen, 2018; Mohankumar et al., 2020). If our findings translate to rUTI in humans, this could open up a therapeutic avenue that is much needed in the current climate of increasing antibiotic resistance (Tamadonfar et al., 2019). Future studies will more closely examine the mechanism(s) for Nur77-mediated response to Gardnerella in the bladder, will investigate whether IE responses impact UTI phenotypes in other mouse models, and will test whether Nur77 could also play a role in host response to Gardnerella in the vagina.

The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE186800 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE186800).

The animal study was reviewed and approved by The Institutional Animal Care and Use Committee (IACUC) of Washington University School of Medicine

NG and VO’B performed experiments. NG, VO’B, and AL analyzed the data. NG and VO’B drafted the manuscript. All authors provided funding to support the research. All authors contributed to the article and approved the submitted version.

This work was supported by the National Institutes of Health NIAID [R01 AI114635 to AL and R21 AI152049 to AL and NG] and NIDDK [R21 DK092586 to AL and K01 DK110225 to NG], by the National Science Foundation [Graduate Research Fellowship to VO’B #DGE–1143954], by the American Heart Association [Postdoctoral Fellowship to NMG] and by the Center for Women’s Infectious Disease Research at Washington University School of Medicine in St. Louis [Pilot Research Grant to NG]. Some of the animal studies were performed in a facility supported by the NCRR [C06 RR015502]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.