key: cord-0064133-yw9rgqml
authors: Murray, Benjamin O.; Flores, Carlos; Williams, Corin; Flusberg, Deborah A.; Marr, Elizabeth E.; Kwiatkowska, Karolina M.; Charest, Joseph L.; Isenberg, Brett C.; Rohn, Jennifer L.
title: Recurrent Urinary Tract Infection: A Mystery in Search of Better Model Systems
date: 2021-05-26
journal: Front Cell Infect Microbiol
DOI: 10.3389/fcimb.2021.691210
sha: 56f0cbd1b2faf86f7a950b136765acf6325cb916
doc_id: 64133
cord_uid: yw9rgqml

Urinary tract infections (UTIs) are among the most common infectious diseases worldwide but are significantly understudied. Uropathogenic E. coli (UPEC) accounts for a significant proportion of UTI, but a large number of other species can infect the urinary tract, each of which will have unique host-pathogen interactions with the bladder environment. Given the substantial economic burden of UTI and its increasing antibiotic resistance, there is an urgent need to better understand UTI pathophysiology – especially its tendency to relapse and recur. Most models developed to date use murine infection; few human-relevant models exist. Of these, the majority of in vitro UTI models have utilized cells in static culture, but UTI needs to be studied in the context of the unique aspects of the bladder’s biophysical environment (e.g., tissue architecture, urine, fluid flow, and stretch). In this review, we summarize the complexities of recurrent UTI, critically assess current infection models and discuss potential improvements. More advanced human cell-based in vitro models have the potential to enable a better understanding of the etiology of UTI disease and to provide a complementary platform alongside animals for drug screening and the search for better treatments.

UTIs are among the most common infectious diseases worldwide, causing approximately 150 million cases per annum (Stamm and Norrby, 2001) , but are significantly understudied (Losada et al., 2016) . Although much more common in women, UTI can affect men and children (Foxman, 2010) and after respiratory infection, it is the most common infectious disease of our ageing population (Rowe and Juthani-Mehta, 2013) . In care homes, UTI is the most common infectious disease of all (Rowe and Juthani-Mehta, 2013) . What's more, UTI is one of the most frequent healthcare-acquired infections, and it is particularly problematic for people with multiple sclerosis, spinal injuries, pregnant women and patients requiring urinary catheters (Foxman, 2003) . As a result, UTI imposes a substantial economic and healthcare burden (Foxman, 2010) . In addition, because of the vast number of individual treatments required, the World Health Organization has described antibiotic resistance in uropathogens as a key pressure point in the growing global antimicrobial resistance crisis (WHO, 2014) . As a number of the bacteria involved also infect different bodily niches, resistance can affect treatment of other diseases, including those requiring surgery. There is therefore an urgent need to better understand UTI pathophysiology, so that alternatives to antibiotics can be developed.

Uropathogenic E. coli (UPEC) accounts for about 80% of community-acquired UTI in otherwise healthy people (Foxman, 2003) , but a large number of other bacterial species can cause a UTI (Foxman, 2010) . One of the biggest concerns about UTI is its tendency to recur. While there have been advances in new treatments and vaccines (reviewed in O'Brien et al., 2016) , antibiotics remain the mainstay of therapy. Even despite treatment, up to 25% of women experience a relapse within six months (Foxman et al., 2000) ; in one large study, 2% had six or more episodes within a two-year period (Laupland et al., 2007) . Some patients experience recurrent UTI for years, necessitating prophylactic antibiotics that only increase the risk of antimicrobial resistance (Selekman et al., 2018) . Although the pathophysiology of UTIprimarily of UPEC-associated diseasehas been extensively researched in animal models (Hunstad and Justice, 2010) , relatively little is understood about the infection cycle in humans. In particular, the behavior of E. coli pathogens in the human bladder lumen is understudied, and the situation for other bacterial species is even hazier.

To develop new treatment strategies for UTI, it is imperative that researchers understand the pathophysiology. Although careful studies in UTI patients have been greatly illustrative, as with most diseases, there are limits to what can be understood in this context. The study of UTI with small animal models and cell culture systems, while also incredibly valuable, still have limitations. Further progress in improving the lot of UTI patients will require advances in human-based model systems.

In this review, we will discuss the proposed mechanisms of UTI pathogenesis and recurrence, and highlight the unknowns that remain. Next, we will discuss the animal models used to study UTI, followed by the main in vitro human-cell-based models, examining their strengths and limitations. Then we will review the key features of human bladder physiology that ideally would be present in an improved in vitro model system. Finally, we will discuss state-of-the-art platforms and solutions to these challenges, and conclude with our perspectives on where the field is headed.

Recurrent urinary tract infection (rUTI) represents a massive burden for the economy, for healthcare systems and for the patients who suffer from them. Several hypotheses for the mechanism of rUTI have emerged over the past decadeswhich are not necessarily mutually exclusive. Indeed, differentand likely multiplescenarios will exist within different patients and across different species of uropathogen, as shown recently (Thänert et al., 2019) . We argue that better human-based infection models will be required to elucidate them fully.

One of the most long-standing theories for recurrence of UTI is that the gastrointestinal tract functions as a reservoir for uropathogens, which are repeatedly reintroduced into the urinary tract via contamination of the periurethral surface and subsequent retrograde ascension. Indeed, common uropathogens are abundant as gut commensals, and there are many papers supporting this mechanism (reviewed recently in (Jones-Freeman et al., 2021) . In agreement, recent studies with patients experiencing rUTI revealed that re-infection was usually preceded by a bloom of uropathogens in the intestine (Thänert et al., 2019) , and that gut abundance of Escherichia and Enterococcus served as an independent risk factor for late problematic urine colonization in kidney transplant recipients, with genomic strain analysis supporting the association (Magruder et al., 2019) . Also in support of this hypothesis, molecules designed to selectively deplete uropathogenic E. coli in the gut also decrease the incidence of UTIs in mice [e.g. (Spaulding et al., 2017) ].

The gut is not the only proximal niche that might harbor uropathogens, however. A crosstalk between the vaginal and urinary microbiomes has been reported based on both clinical correlations and experimental models, and the data strongly suggest that this interconnection affects the recurrence of UTIs (Komesu et al., 2020; Lewis and Gilbert, 2020) . Moreover, the now well-established observation that healthy urinary tracts are not sterile and that their microbiota is distinct from those of chronic UTI patients has highlighted the potential relationship between bladder dysbiosis and rUTI (Khasriya et al., 2013; Hilt et al., 2014; Neugent et al., 2020) . Dysbiosis, defined as an imbalance in the natural microbial community, may even affect several host protective mechanisms, such as those mediated by commensals. Also, alteration of the urinary microbiota through the introduction of native microbes from either the gut or the vagina may potentiate the recurrence of UTI. For example, the vaginal commensal Gardnerella vaginalis was reported in the urinary tract of both human males and females (Fairley and Birch, 1983; Lam et al., 1988) . More recently, the inoculation of Gardnerella into the bladders of mice that had recovered from UPEC infection caused urothelium exfoliation and facilitated the exposure of intracellular UPEC reservoirs in the bladder (Gilbert et al., 2017) . In addition, this exposure increased the severity of infection compared with control mice. Of note, a small clinical trial showed that the introduction of a probiotic Lactobacillus species capable of outcompeting uropathogens into the vagina of women with rUTI was able to modestly reduce the incidence of UTI (Stapleton et al., 2011) . Reciprocally, commensal Lactobacillus in the vagina establish a low pH that is non-permissive to uropathogenic species such as E. coli; if these commensals are outcompeted, uropathogens can become the dominant vaginal species, which would facilitate their eventual transit to the urethra (Brannon et al., 2020; Lewis and Gilbert, 2020) .

Antibiotics are another important factor affecting rUTI, not only those taken to treat UTI, but also other indications. Antibiotics favor the development and proliferation of multidrug-resistant organisms (either in the bladder or in the gut/vagina), as well as increase the availability of niches that are no longer inhabited by commensals. As a result, these niches can become dysbiotic and more easily colonized by more virulent and persistent bacteria (Chen et al., 2013) .

Taken together, these various strands of evidence have boosted the search for probiotic approaches and other therapies focused on the modulation of microbiota in the UTI context. However, a human cell-based model that could accurately mimic the human urinary microbiota and its interactions with other microbiota would be a very helpful tool in this endeavor.

Many strain-specific bacterial virulence factors may contribute to the recurrence of UTI, such as flagella/pili, adhesins, extracellular polysaccharides, lipopolysaccharides, toxins, ureases, proteases and iron-scavenging siderophores. These factors allow uropathogens to survive during long periods in a nutrientlimited habitat, helping them to adhere, colonize, damage and invade host cells, as well as to evade host defenses, ultimately increasing their persistence in the urinary tract. For instance, a common uropathogenic strategy is the formation of biofilms, either directly on the urothelial surface or on indwelling devices such as catheters (Jansen et al., 2004; Jacobsen and Shirtliff, 2011) . This behavior has been reported in Klebsiella pneumoniae (Reid et al., 1992; Stahlhut et al., 2012) , Pseudomonas aeruginosa (Ivanova et al., 2015; Saini et al., 2015) , UPEC (Ferrières et al., 2007) and Enterococci (Sillanpää et al., 2010) , and may even be facilitated by polymicrobial interactions during infection (Gaston et al., 2020) . The biofilm provides an ideal physicochemical barrier that protects uropathogens from antimicrobial agents, host immunity and other stresses, allowing them to persist and reinfect the urinary tract. Furthermore, a subpopulation of bacterial cells in biofilms, the so-called persistors, are known to reversibly reduce their metabolic activity, adopting a dormant state that can evade host defenses as well as treatments that target active metabolic pathways or activities such as cell division (Wood et al., 2013) . Another strategy is the adhesion of bacteria either to the host cell surface and/or the extracellular matrix (ECM), through the use of a vast arsenal of virulence factors. Among the most common are curli proteins (similar to amyloid fibers), which are major facilitators of UPEC colonization (Luna-Pineda et al., 2019) . These proteins allow bacterial binding to ECM and serum proteins such as fibronectin, laminin and plasminogen, and are highly expressed among Enterobacteriaceae. P-type pilus and the FimH fimbrial adhesin (part of the Type-1 pilus) have also received attention due to their high binding affinity with urothelial receptors, being important players for UPEC colonization and/or invasion in several UTI models (Mulvey et al., 1998; Zhou et al., 2001; Bishop et al., 2007; Feenstra et al., 2017) . Therefore, a number of therapeutic approaches against rUTI target specific virulence factors, mainly blocking adhesion and/or biofilm formation (Liao et al., 2012; Flores-Mireles et al., 2014; Mike et al., 2016) .

Another bacterial mechanism implicated in rUTI is the ability of bacteria to invade host urothelial cells and establish intracellular bacterial communities (IBCs). This was first discovered in the early 2000s in a UPEC mouse model of induced UTI, in which the bacteria subverted host defenses, invaded urothelial cells and formed IBCs that could later erupt and re-establish UTI (Anderson et al., 2003; Justice et al., 2004) . In this exciting model, after invasion, bacteria rapidly multiply in the cytoplasm of superficial bladder epithelial cells, where they form "pods" that can expand, blister-like, into the lumen of the bladder as the community grows (Anderson et al., 2003) . These communities are embedded in a biofilm-like matrix, which may confer a protective effect similar to that seen in more conventional biofilms. In other cases, UPEC may even invade deeper layers of the urothelium, remaining in a membrane-bound compartment with little to no metabolic activity. These "quiescent intracellular reservoirs" (QIR) can be latent for months in mice, and can be re-activated after the exfoliation of the upper urothelium cell layers (Mulvey et al., 2001; Mysorekar and Hultgren, 2006) .

While IBCs have also been experimentally demonstrated in human cancer cell lines (e.g. (Bishop et al., 2007; Gonzaĺez et al., 2020) , nearly two decades on from the original IBC discovery in mice, surprisingly few papers have reported the existence of UPEC IBCs in patients. A handful show IBC using exfoliated urothelial cells from the urine of human UTI patients (Rosen et al., 2007; Robino et al., 2013; Robino et al., 2014; Cheng et al., 2016) , although only some (Robino et al., 2013; Robino et al., 2014) used imaging resolution with sufficient discriminatory power to distinguish intracellular bacteria from those on the surface of these notoriously flat cells. To our knowledge, only one group has studied human biopsies and reported the existence of IBC and QIR (De Nisco et al., 2019) . In contrast, in a recent study with a porcine model, which has a more similar urogenital anatomy, physiology and immune response to human compared with small-animal models, IBCs were not observed in the bladder after UPEC infection, although high loads of bacteria were detected after prolonged infection, some forming biofilm-like extracellular aggregates (Nielsen et al., 2019) . Therefore, more studies on the role of IBC and QIR in rUTI in humans or human bladder model systems would be welcome.

Although intracellular lifestyle stages have been observed mainly in UPEC infection, other uropathogens, such as Klebsiella pneumoniae , Staphylococcus saprophyticus (Szabados et al., 2008) and Salmonella enterica (Bishop et al., 2007) might also display them, at least in murine models and/or cell lines. More recently, Enterococcus faecalis was shown to reside inside urothelial cells shed from patients with chronic UTI (Horsley et al., 2013) , and to invade a human organoid model (Horsley et al., 2018) . It remains to be seen whether other uropathogens invade human cells, and how widespread the phenomenon is in patients.

Recurrent infection may also be facilitated by the ability of some uropathogens to alter their morphology to evade the host immune system and recolonize naïve regions in the urinary tract more easily. One well-studied example is UPEC cell filamentation, a process which occurs when bacterial cells emerge from IBCs (Justice et al., 2006) . This filamentous form confers resistance to phagocytic engulfment (Justice et al., 2006) and may provide enhanced adhesive properties (Andersen et al., 2012) . Another example is Proteus mirabilis, which can increase flagellar density (Armbruster and Mobley, 2012) . Recently, Mickiewicz et al. reported that UPEC may also acquire a reversible "L-form", a cell-wall deficient phase that can evade antibiotic treatments targeting the bacterial cell wall. In both urine and a zebrafish model, UPEC could rapidly switch into this form; the L-form was also the most prevalent in urine from older rUTI patients (Mickiewicz et al., 2019) .

Antibiotic resistance is a key issue in recurrent UTI, and is the most well-studied mode of treatment failure generally (Ho et al., 2019; Yelin et al., 2019; Mattoo and Asmar, 2020) . The overuse of antibiotics has positively selected for strains with specific genetic traits allowing them to survive and proliferate in the presence of a single or even a class of antimicrobial compounds. In addition, most of these bacteria have the necessary machinery for intraand inter-species transmission (e.g. through horizontal gene transfer mechanisms), which can spread the selected traits rapidly and boost the generation of multidrug-resistant uropathogens (Ho et al., 2019; Mattoo and Asmar, 2020) .

On the other hand, a much less well-understood phenomenon involves resilience behaviors that provide temporary antibiotic evasion in a sub-population of bacterial cells. The emergence of resilient bacteria in an overall susceptible bacterial population might therefore play an important role in the selection of fitter uropathogens and subsequent recurrence. Several genetic and non-genetic mechanisms might be involved in the development of resilient profiles, which can be generally categorized as tolerance, persistence and heteroresistance phenotypes (Carvalho et al., 2019) .

One of the best studied bacterial resilient mechanisms involved in rUTI are biofilms, which provide tolerance to external stresses, such as antibiotic treatments and host defenses (Olsen, 2015) . Bacteria in polymicrobial UTI also have the ability to protect one other from clinically relevant antibiotics through the increase of tolerant/resilient phenotypes in the bacterial community (de Vos et al., 2017) . In contrast, the mechanisms of persistence and heteroresistance are still largely unexplored in the UTI context, as well as their impact in clinical settings. However, the development of bacterial subpopulations that are dormant/less metabolically active and/or display a heterogenous (more resistant) phenotype might be crucial for the success of a pathogenic community facing inconsistent and unexpected environmental challenges (Kussell et al., 2005; Carvalho et al., 2019) . Although much remains to be learned about this interesting class of treatment failure, bacterial resilience may have a role in the development of chronic UTIs, while resistance may play a more prominent role in recurrent acute infections (Olsen, 2015) . It is important to note that the host context is likely to influence bacterial resilience phenotypes, so studying this phenomenon in a human-cell (and one day, patient-specific) environment will be important.

Bacteria are not the only players to alter their properties during the host/pathogen interaction. Host urothelial cells can also change upon first infection, a phenomenon observed in experimental mouse models of chronic UPEC infection and in cell lines. These changes influence host cell gene expression, shape, size, growth and proliferation, and depending upon what is changed, may make the urothelium more resilient, or more susceptible, to re-infection (Mysorekar et al., 2009; Shin et al., 2011; .

Other host factors, mainly immune-related, also play a prominent role in the chronicity of rUTI, as the recurrence of infection is frequently associated with a disturbed innate immune response and/or insufficient adaptive immunity (recently reviewed in (Lacerda Mariano and Ingersoll, 2020) . The reasons behind these events are still unclear and largely unexplored due to the lack of knowledge about how the immune response is triggered and develops in human UTIs, which can differ significantly from the murine context. Indeed, up until recently it was assumed that adaptive immunity was not involved at all, due to early cumulative evidence from immune-deficient mouse models lacking interleukins, immunoglobulins and/or T lymphocytes, that could nevertheless resolve UTI as well as their wild-type counterparts or even be resistant to infection Ragnarsdottir et al., 2008) . A revision of this view has paved the way for the development of vaccines against UTI, some of which are showing promising results (Prattley et al., 2020) .

It has been increasingly reported, mainly in children, that a genetic component might be involved, as polymorphisms and mutations in innate immunity-related genes can increase susceptibility to UTIs (Karoly et al., 2007; Tabel et al., 2007) . Additionally, an impaired immune response as well as an exacerbated pro-inflammatory response to a primary infection (Hannan et al., 2014) , might favor recurrence of infection. In fact, the activation of host immunity frequently leads to severe injuries in the urinary mucosa/urothelium, which become more prone to subsequent infections. In one way or another, this common suboptimal and non-sterilizing immune response may create the perfect environment for the occurrence of rUTIs in an experienced host.

Animals of many species were essential for biological scientific experiments as early as the late 19 th century, furthering our understanding of physiology and human diseases. The earliest mention of establishing infection in animal bladders was in 1873, when Fels and Ritter inoculated canine bladders to induce cystitis using urethral ligation (Keyes, 1894) . Many tried to induce infection in animals using pure cultures of microorganisms, but ligation or a wound to the bladder was always necessary to establish cystitis. In 1890, Schnitzler managed to induce cystitis in rabbits without ligation using 'Urobacillus liquefaciens septicus' [later defined as a member of the Bacillus cloacae group (Archer, 1931) ]. Later, it was reported that 'coliform bacillus' could produce cystitis in rabbit kidneys and an inflammatory response in the bladder mucosa (Lepper, 1921) . Animals continued to be important for UTI research, mainly rats and rabbits, until the first murine model of UTI was described in 1967, as an experimental infection model for pyelonephritis (Keane and Freedman, 1967) .

Murine models eventually became established as a valuable tool to study UTI, a trend which continues to this day. Mice were considered superior to rats because they were slightly more relevant to humans, with a greater abundance of globoseries glycolipid receptors on urothelial cells for attachment . Similar to humans, mice do not have a natural vesicoureteral reflux, unlike other rodent models, so would be a more physiologically accurate model for pyelonephritis (Hvidberg et al., 2000) . Mouse and human bladders also have highly conserved uroplakins, which aid type 1 fimbriae adherence and invasion by UPEC (Zhou et al., 2001) . Recently, female C57BL/6 mice urothelial cells were analyzed and classified into eight clusters dependent on expression of cell-specific markers. A novel urothelial cell type expressing Plxna4 was discovered in the mouse bladder which is conserved in humans and may play a role in host immune response (Wen et al., 2010) .

As mice do not naturally develop urinary infections, instillation techniques were developed. Hagberg et al. described the ascending, unobstructed UTI instillation in female CBA mice with E. coli which has since been adapted for other uropathogens such as Proteus mirabilis and Enterococcus faecalis (Jones et al., 1990; Shankar et al., 2001) . This instillation method helped replicate the attachment of human-derived UTI isolates to murine urothelial cells which was considered essential for understanding human infection. CBA mice were preferred due to better bacterial attachment and colonization; female CBA were favored as the anatomy of male mice presented challenges for urethral inoculation . Although UTI is more prevalent in females than males (Foxman, 2010) , little is known about sex differences in UTI pathophysiology. To address this, a protocol for inducing UTI in male mice using transurethral catheterization allowed direct comparison between male and female C57B1/6 mice and their host response (Zychlinsky Scharff et al., 2017) . However, all instillations via catheterization bypass the biology of ascending UTI via the urethra (Barber et al., 2016) .

Many different mouse models have been used for UTI ( Table 1 ). The mouse model offers systemic context and a diverse range of genetic variability, allowing researchers to test specific host factors and immune responses in transgenic and knockdown mutants (Barber et al., 2016) . 'Germ free' murine models are also potentially an attractive option for modeling UTI and the gut microbiota/UTI axis, allowing exclusive colonization of certain pathogens or purposefully introduced commensal species. For immunological purposes, naturalizing mouse models can also be used to help translate research to humans (Graham, 2021) . The immune profiles of laboratory mice are different from that of human as they have a low density of mature T cells (Beura et al., 2016) and lower LPS sensitivity which could lead to differences in pathogenesis and treatment responses in humans (Graham, 2021) . However, germ-free mice are more expensive and need specialized equipment and training (Kennedy et al., 2018) . To our knowledge, however, there have been no reports of naturalized mice used for UTI research.

Although murine models remain incredibly valuable for UTI, they are expensive and labor-intensive to maintain and breed. This precludes their use for high-throughput analysis, particularly for drug screening and testing. Ethical guidelines and applications have also slowed down animal research, although scientists can navigate these obstacles with a little patience.

Aside from these logistical issues, there is a deeper concern that animal models do not always recapitulate the human environment well enough to predict how disease physiology works, nor how prospective treatments might behave in human patients (Herati and Wherry, 2018) , especially as mice do not naturally acquire UTI. Accordingly, the use and accuracy of animal models is a frequent discussion point among UTI researchers, with Barber et al. reviewing their strengths and limitations (Barber et al., 2016) . Since then, further questions about the accuracy of animal models, especially mice, have arisen.

Murine and human bladders differ in certain anatomical features and expression of biomarkers on the epithelial surface ( Figure 1 ). The mouse urothelium is typically 3-4 cells thick, making it thinner compared with human, which is 5-7 cells thick. The difference is due to the number of intermediate cells , which could create differences in physiology and microenvironment between the upper and lower intermediate cells. In addition, the basal cells of the human urothelium have a higher expression of cytokeratin (CK) 5, 13, 14 & 17 compared with mice, whereas human umbrella cells have a higher expression of CK7, 8, 18 & 20 (Laguna et al., 2006) . 8-and 12-weeks post infection with Enterococcus faecalis, C57BL/6 mice had a lower incidence of infected kidneys than SWR mice but a higher proportion of infected kidneys by week 8 compared with mouse strain A (Guze et al., 1985) Due to their popularity not all C57BL/6 strains are the same with differences between C57BL/6J and C57BL/6N strains (Simon et al., 2013; Kelmenson, 2016) Demonstrated extremely low inflammation of the kidneys throughout the 14 day infection with E. coli 1677 (Hopkins et al., 1998) Infected with either E. coli WT NU14 or its fimH − analogue. 6h post-infection, some WT E. coli formed microcolonies in the bladder, while some adhered individually to the surface; fimH − bacteria did not attach (Mulvey et al., 1998) 75% of infected bladders remained colonized with a fimH isogenic mutant of E. coli 2 weeks post-infection (Mulvey et al., 2001) Infected with E. coli UTI89 and CFT073. Intracellular bacterial communities (IBCs) present 6h post infection. IBCs in C57BL/6J occasionally had more loosely defined edges (Garofalo et al., 2007) Infected with E. coli asymptomatic bacteriuria strain VR50; after 18h, VR50 could colonize the urothelium but not mutants for afimbrial adhesin (Beatson et al., 2015) Infected with E. coli asymptomatic bacteriuria strain 83972 which ameliorated the effects of a 24h challenge by a UPEC strain (Rudick et al., 2014) CBA

Large kidneys with proneness to tubulointerstitial lesions (Rudofsky, 1978) Evidence of virulence factor type 1 fimbriae in E. coli (Gunther et al., 2001) Greater abundance of globoseries glycolipid receptors on urothelial cells for attachment of UTI pathogens compared to humans Infected with E. coli UTI89 and CFT073. IBCs present 6h post infection. Similar IBC morphology compared with other mouse strains (Garofalo et al., 2007) More susceptible to UTI infection than BALB/c , C57 and C3H/HeN mice 40 CBA/J mice were infected with E. coli CFT073. 313 bacterial genes were upregulated and 207 downregulated. Chemotaxis-and flagella motility-associated genes were downregulated in vivo compared to in vitro (Snyder et al., 2004) CF1 Albino white mice 11 different E. coli strains were analyzed for infectivity. Presence of C175-94 strain in bladder was found in 4/6 mice 4 weeks post-infection (Hvidberg et al., 2000) Outbred to Charles River Laboratories in 1974 (Chia et al., 2005) Have been used for EGFR gene knockout studies so can be genetically manipulated (Threadgill et al., 1995) FVB/NJ Nephrotic syndrome-like characteristics: albuminuria, cholesterolemia, predisposition to increased T2 cell response; no autoimmune origin (Maier et al., 2007) Infected with E. coli UTI89 and CFT073. IBCs present 6 hours post infection. Similar IBC morphology with other mouse strains (Garofalo et al., 2007) Ccr5 P185L mutation results in resistance to paracetamol (Timmermans et al., 2017) Mx1 s1 results in susceptibility to myxoviruses such as Influenza (Grimm et al., 2007) High susceptibility to glomerular disease (Uchio-Yamada et al., 2016) Used in autoimmunity research due to proneness to develop asthma-like disease with high levels of IgE (Zhu and Gilmour, 2009) C3H/HeJ IL1a Y118_T119del results in resistance to IL1a-mediated inflammation (Timmermans et al., 2017) Clears LPS-containing Gram-negative bacteria at a slower rate compared with C3H/HeN mice. There was no difference for clearing Gram-positive bacteria such as Staphylococcus saprophyticus (Continued) (Timmermans et al., 2017) When infected with E. coli 1677 (O6, type 1), mice had increased bacterial growth in the bladder after 3 days. Inflammation response to LPS was poor and only slightly increased from Day 3 to day 14 post infection (Hopkins et al., 1996) LPS Resistant (HeJ -) (Tötemeyer et al., 2003) Infected with E. coli CI5 to develop prolonged UTI in the bladder. Assessed Forskolin as a potential treatment for UTI (increases cAMP-driven exocytosis). Multiple intraperitoneal injections of 10mg/kg Forskolin at 6, 24 and 48 hours post infection reduced colonization of intracellular E. coli (Bishop et al., 2007) C3H mice have mutation in Nramp1 G169D that makes them susceptible to intracellular pathogens like Mycobacterium, Salmonella & Leishmania (Vidal et al., 1996) Infected with E. coli UTI89 and CFT073. IBCs present 6h post infection. Similar IBC morphology compared with other mouse strains (Garofalo et al., 2007) Filamentation of E. coli UTI89 in C3H/HeJ mice is observed only in the later stages of infection. Could be due to lack of functional TLR4 and resulting lack of inflammatory response (Justice et al., 2006) C3H/OuJ C3H mice have mutation in Nramp1

G169D that makes them susceptible to intracellular pathogens like Mycobacterium, Salmonella & Leishmania (Vidal et al., 1996) When infected with E. coli 1677 (O6, type 1), infection became more severe with increased bacterial growth. Unlike other strains in which inflammation gradually decreased, C3H/OuJ had increased inflammation in both, kidney and bladder (Hopkins et al., 1998) LPS sensitive (Hopkins et al., 1996) C3H/HeN IL1a Y118_T119del results in resistance to IL1a-mediated inflammation (Timmermans et al., 2017) Cleared infection of LPS-containing Gram-negative bacteria faster than C3H/HeJ mice but not Gram-positive bacteria Staphylococcus saprophyticus C3H mice have mutation in Nramp1 G169D that makes them susceptible to intracellular pathogens like Mycobacterium, Salmonella & Leishmania (Vidal et al., 1996) Infected with E. coli 1677 (O6, type 1). Demonstrated a high level of infection in the bladder over 14 days compared with other mouse strains (Hopkins et al., 1998) Had a significantly reduced bacterial load when infected with E. coli 1677 than C3H/HeJ and C3H/OuJ over 14 days (Hopkins et al., 1996) Infected with E. coli UTI89 and CFT073. IBCs present 6h post infection. Similar IBC morphology compared with other mouse strains (Garofalo et al., 2007) Infected with E. coli UTI89 or K. pneumoniae TOP52 FimH strain, WT or genetically engineered fimH knockouts. WT-infected bladders had significantly higher CFU counts in E. coli at 6h, 24h and 336h post-infection. In contrast, WT K. pneumoniae showed no significant difference in bacterial count at 6h and 24h post-infection and became significantly lower at 336h post-infection. K. pneumoniae fimH was important for intracellular bacterial colonization Infected with E. coli UTI89 using the technique described in (Olson et al., 2016) . Mice were treated with ceftriaxone, but a few mice were left with residual UPEC. When infection cleared, kidneys of treated mice revealed to have multiple scars containing cellular infiltrate. Males, but not females, can develop 100% penetrant robust ascending UTI presenting with renal abscesses and fibrosis, and fail to resolve it (Olson et al., 2017) DBA/1 & DBA/2 DBA/1 are susceptible to immune mediated nephritis (Xie et al., 2004) Infected with E. coli 1677 (O6, type 1), DBA/1 and DBA/2 strains could resolve the initially high level of infection in bladder and kidney (Hopkins et al., 1998) DBA/2 lack surface expression of CD94/ NKG2A on NK cells known to be expressed on most fetal NK cells (Vance et al., 2002) DBA/1 mice were infected with Pseudomonas aeruginosa for phage therapy applications and found there was more enhanced killing of intracellular bacteria (reviewed in (Ciesĺik et al., 2021) BALB/c Naturally resistant to prolonged UTI (Bishop et al., 2007) 8-and 12-weeks post infection with S. faecalis GK (ATCC 23241), BALB/c mice had a significantly lower incidence of infected kidneys compared with SWR mice (Guze et al., 1985) Used in cancer research and when older can develop renal tumors in males (Noronha, 1977) Infected with E. coli 1677 (O6, type 1), BALB/c mice had highest bacterial presence in the bladder and second highest in the kidneys after 24 hours. After 14 days, infection decreased in both bladder and kidney (Hopkins et al., 1998) Infected with E. coli CI5 to determine the efficacy of Forskolin as a treatment for UTI (increases cAMP-driven exocytosis). Intravesical 100µM Forskolin reduced colonization of intracellular E. coli by more than 75% compared to saline controls (Bishop et al., 2007) (Continued) human urothelium, with CD271 in basal cells and CD227 in umbrella cells; however, there was no specific CD marker for intermediate cells (Liu et al., 2012) . The analogous situation in murine urothelium has not been explored. Another species difference involves the interstitial cells (IC), which are recently described specialized cells residing in the lamina propria and detrusor muscle associated with nerves (McCloskey, 2010). Yu et al. found a high similarity in singlecell types between human and mice bladders, but single-cell transcriptomic mapping highlighted two novel IC expression types specific to human bladders, namely ADRA2A + and HRH2 + . These ICs may play a role in allergic reactions and nerve conduction . Gavaert et al. also reported a difference between mouse and human IC, with human ICs manifesting increased contractile microfilaments, versus a fibroblast phenotype in murine ICs (Gevaert et al., 2017) . The role of ICs in the host response during UTI is unknown but could inform new therapeutic targets.

Urodynamics and voiding patterns also are different between humans and mice. Both sexes of 10-week-old mice void more frequently than humans,~10 times a day vs~6 times a day (Chung and van Mastrigt, 2009; Aizawa et al., 2013) , and also more frequently at night (Ito et al., 2017) . Unlike animals over 3 kg, including humans, which possess a scalable urinary capacity and similar voiding durations, Yang et al. reported that mice scarcely store urine at all, which would affect the shape and stretching parameters of their bladders (Yang et al., 2014) . Contraction of the detrusor muscle in mammalian bladders usually depends on activation of muscarinic-3 (M3) receptors, whereas mice use activation of muscarinic-2 (M2) receptors via an indirect mechanism (Ehlert et al., 2005; Zhou et al., 2010) , which could be important for accurate disease modeling. Age may also play a role; e.g. differences were reported in voiding and storage with 12 week "mature" mice compared to "aged" 27-30 month C57BL/6 mice, with more severe bladder dysfunction in aged male mice compared with female. M3 receptor expression was downregulated in aged male mice, whereas b2-adrenoceptor was downregulated in aged females (Kamei et al., 2018) . As aging is a risk factor in developing UTIs (Foxman, 2010) yet younger mice are mostly selected for UTI research, it is possible that youthful mouse models are inaccurate for modeling UTI in our aging population. After infection with S. faecalis GK (ATCC 23241), AKR mice had established pyelonephritis after 8 weeks and had the second highest bacterial load after 12 weeks in kidneys (Guze et al., 1985) Infected with E. coli 1677 (O6, type 1), AKR mice could resolve the initially high level of infection in the kidneys after 5-7 days but not in the bladder (Hopkins et al., 1998) SJL Susceptible to experimental autoimmune encephalomyelitis useful for multiple sclerosis research (reviewed in (Croxford et al., 2011) SJL mice had a significantly higher proportion of infected kidneys than Strain A at week 8 and similar to other mice (Guze et al., 1985) Elevated T-cell level (Jain et al., 2015) Infected with E. coli 1677 (O6, type 1), SJL strains showed a very low but constant level of infection in bladder and kidney (Hopkins et al., 1998) SWR Develop very severe polydipsia and pyuria as a consequence of nephrogenic diabetes insipidus (Kutscher and Miller, 1974) When infected with S. faecalis GK (ATCC 23241), SWR mice had increased ascending UTI in the kidneys compared to Strain A mice. No histological changes to the kidney were seen and the infection was cleared effectively.

When infected with E. coli strain Yale, SWR mice had a significantly increased bacterial load in the kidneys with more histopathological changes than Strain A mice over 12 weeks (Guze et al., 1987 ) Kidneys unresponsive to vasopressin (Kutscher et al., 1975) 24h post infection with S. faecalis GK (ATCC 23241), SWR mice had increased bacterial growth compared to strain A mice. Cortical abscesses and scars at week 1 and 4 respectively were significantly greater in SWR mice than strain A mice. (Guze et al., 1985) SWR/J are resistant to obesity on high fat diets (Leibowitz et al., 2005) Infected with E. coli 1677 (O6, type 1), SWR mice showed very low but constant level of infection (Hopkins et al., 1998) A Suffer from specific complement deficiency (Timmermans et al., 2017) When infected with S. faecalis GK (ATCC 23241), Strain A mice exhibited lower kidney infection with no significant histopathological changes compared with SWR mice. Infection was cleared effectively for both.

Within 12 weeks post infection with E. coli strain Yale, Strain A mice had a significantly decreased bacterial count in the kidney and no histopathological changes. (Guze et al., 1987) Macrophage defects (Timmermans et al., 2017) When infected with S. faecalis GK (ATCC 23241) cortical abscesses and scars were significantly reduced after 1 week in strain A mice compared to SWR mice. Strain A had a significantly lower proportion of infected kidneys than all other tested mouse strains. (Guze et al., 1985) Host Pathogen Interactions and Immunological Differences

Humans and mice also differ in some innate immunological parameters. In UTI, the innate immune system is particularly important for host recognition of pathogen-associated molecular patterns (PAMPs) via pathogen recognition receptors (PRRs). PRRs recognize pathogens and help regulate a quick proinflammatory immune response, the mechanisms of which have been extensively reviewed (Takeuchi and Akira, 2010; Abraham and Miao, 2015) . The predominant group of PRRs are Toll-like receptors (TLRs). Differences in TLR structure and function between humans and mice could influence how the innate immune system responds to a particular UTI pathogen, which could influence model accuracy.

There are 13 members of TLRs and between humans and mice; TLR1-9 are present in both, TLR10 is found in only in humans and TLR11-13, only in mice (Behzadi and Behzadi, 2016) . In the urothelium, TLR2, 4, 5 and 11 (in mice) are the most competent TLRs against UTIs (Song and Abraham, 2008; Behzadi and Behzadi, 2016) .

TLR4 is the most well-studied of the TLRs. TLR4 forms complexes with CD14 and MD-2 which help bind bacterial lipopolysaccharide (LPS) to initiate signaling for innate immune response (Vaure and Liu, 2014) . LPS is displayed abundantly on the outside of Gram-negative bacteria such as UPEC, highlighting the importance of bacterial TLR4 interactions in the urinary tract. Despite TLR4 being present in both species, noticeable differences exist (reviewed by Vaure and Liu, 2014) , with only a 57% similarity between their MD-2 proteins; this makes human MD-2 more cationic than mouse MD-2 which further affects the activation of TLR4 signaling (Vasl et al., 2009) . The hypervariable region of TLR4, where MD-2 complexes with TLR4, has only 48% similarity between humans and mice, which again might lead to differences in how the host responds to infections. To date, the impact of these differences in TLR4 in human and mouse UTI has not been investigated.

TLR2, expressed in both species, has been shown to be a signaling receptor for bacterial peptidoglycan, important for recognition of Gram-positive bacteria (Takeuchi et al., 1999) , and its ability to form heterodimer complexes with TLR1 and 6 broadens its spectrum to recognize other PAMPs (reviewed in Behzadi and Behzadi, 2016) . TLR5 and 11 are similar in that they recognize UPEC PAMP flagellin (Hatai et al., 2016) . As TLR11 is only expressed in mice, mice might have a higher capacity to recognize flagellin from UPEC, which should be kept in mind when studying this bacterial parameter.

Bacterial phenotypes differ between mouse and human infections, likely influenced by differences in environment, bacterial development and growth. For example, the importance of some bacterial virulence factors seems to be completely different between human cell lines and animal models (Alamuri et al., 2010) . A recent paper explored the gene expression of three UPEC strains isolated from cases of uncomplicated UTI; gene expression between human and mouse FIGURE 1 | Comparison between human versus mouse urothelium. In human bladder, basal cells have higher expression of CK5, 13, 14 and 17, intermediate cells are stacked in 5-7 layers (vs 3-4 in mice), and umbrella cells have higher expression of CK8, 7, 18 and 20, as well as a more cationic MD-2 protein associated with the TLR4, while mice have a high expression of TLR11 (which is not present in humans). Urine is less concentrated in humans, which also have higher storage capacity and lower urination frequency compared with mice. UTI was highly correlated, but 5.4% of analyzed genes were differentially regulated with 30 genes upregulated and 145 downregulated in human UTI (Frick-Cheng et al., 2020) . Most of the downregulated genes were involved in anaerobic metabolism, so the authors hypothesized that the human bladder is more oxygenated than the murine. The presence and full effect of this physiological difference on bacterial phenotypes has not been explored.

Species differences in urine composition might also contribute to differences in bacterial behavior. Mouse urine has been shown to change the gene expression of UPEC when compared with human urine (Hagan et al., 2010) . This study found that type 1 fimbrial genesessential for bacterial adherence in murine modelswere not expressed in 6 of 8 clinical isolates in human urine, suggesting that murine models impart a different expression phenotype. Mouse urine is more concentrated than that of larger mammals, which can affect UPEC biofilm production and cell morphology. FliC, a virulence factor encoding UPEC flagellin, was shown to be downregulated in mouse urine and upregulated in human urine (Snyder et al., 2004; Berry et al., 2009 ). On the other hand, filamentation, which is important for UPEC virulence, occurred in both highly concentrated human and mouse urine (Klein et al., 2015) .

Murine models have been used to study antibiotic efficacy. Chockalingham et al. infected immunocompetent Balb/C mice with UPEC strain CFT073 to investigate resistance development against ampicillin, ciprofloxacin and fosfomycin, concluding that the mouse model was not suitable for studying resistance patterns, but could be useful for studying persistence (Chockalingam et al., 2019) .

Porcine models have recently been mooted as an attractive species to model pathogenesis of UTI. They have more conserved homology and structural motifs to humans compared with mice, which may make it a more accurate model particularly for immunological studies (Dawson et al., 2017) . Thus far pigs have been mainly used for research into pyelonephritis and renal damage, especially vesicoureteric reflux and renal scarring in infants and establishment of upper UTI (Coulthard et al., 2002) . However, until recently pigs had not been used to model cystitis. In 2019, Nielsen et al. reported a UPEC model of infection using a clinical isolate UTI89 to sustain an infection for up to 23 days in female pigs (Nielsen et al., 2019) . Interestingly, as mentioned above, no intracellular UPEC were seen as has been widely reported for UTI89 in mice. The porcine model, as a large mammal, shares similarities in urine density and anatomy with humans (Nielsen et al., 2019) . However, large animal models are particularly expensive. What's more, the scope for genetic manipulation in pigs has yet to match that of mice (Barber et al., 2016) .

The nematode Caenorhabditis elegans has been used for in vivo infection studies that include various UTI pathogens (Lavigne et al., 2008; Szabados et al., 2013; Engelsöy et al., 2019) . Very recently, Hashimoto et al. demonstrated that UPEC mutants with defective iron acquisition-related virulence factors were a significant factor in survival of C. elegans (Hashimoto et al., 2021) . However, as a model, C. elegans has only been useful for survival assays and cannot be explored as a physiological model as it lacks a urinary system. Zebrafish, a popular model for real-time visualization of infections, also lack a urinary system, but has recently been used to determine in vivo L-form switching of UPEC and its role in recurrent UTI (Mickiewicz et al., 2019) .

Non-human primate (NHP) models benefit from their high similarity with humans. To our knowledge there are no papers describing an NHP model for UTI , but they have been used in the past for UTI vaccine studies (reviewed in and viral studies (reviewed in (Estes et al., 2018) . However, their expense, ethical issues and specialized facilities pose significant drawbacks.

In summary, animal models have enabled crucial advances, but relevant human-related models are still needed to fully understand UTI in patients. These models should provide complementary valuable insights into the host-pathogen interactions and allow the development of more humanrelevant therapeutic approaches.

The in vitro recapitulation of the human bladder environment and the reconstitution of the urothelium (or even the organ) is highly desirable because of the advantages such models present compared with animal models (see Section 3) and their potential for studying a wide range of conditions, from UTIs to bladder cancer, tissue regeneration/transplant and the effect of drugs and xenobiotics that contact the urinary tract. Three major strategies have been employed: i) cell monolayers, using commercially available cell lines, cells recovered from urine or derived from biopsies/explant cultures; ii) tissue/organ cultures from patients; and iii) 3D structures/organoids that mimic the urothelium after induced cell stratification and differentiation. An overview of the different bladder models used in UTI and non-UTI related research is presented in Tables 2, 3, respectively.

As human bladder cell lines are easy to maintain in 2D and to manipulate using standard techniques, cell monolayer culture remains one of the most attractive methodologies. There are also an increasing variety of well-established human bladder cell lines that can be purchased; e.g. the American Type Culture Collection currently provides 14 certified human bladder cell lines (4 normal and 10 cancer). Those from cancer sources are among the most popular, mainly HTB-9 and HTB-4, due to their fast growth and easy manipulation. Apart from their obvious relevance for cancer studies, the data obtained with these cells in UTI research should be carefully interpreted due to the differences between normal and cancer cells. Moreover, these cells cannot acquire a native threedimensional urothelial architecture. On the other hand, normal cells are either immortalized, which may compromise cell differentiation (Georgopoulos et al., 2011) , or are derived from human samples (urine or biopsies), maintaining their ability to stratify and differentiate. In the latter case, cells can only be used during a limited number of passages so their routine use requires a large amount of material. Also, donor variation may affect reproducibility. In general, the previously mentioned models generate one of two extremes, either highly differentiated primary cultures with lower serial growth potential, or undifferentiated cultures with higher serial growth potential. Importantly, regardless of the cell source, cells cultured in monolayers do not display growth, architecture and physiology comparable to the human urothelium, limitations that are particularly relevant to UTI (Smith et al., 2011) .

In contrast, a more biologically relevant model is organ/ explant culture, where intact specimens of tissues are cultured for a time. While these may maintain an accurate 3D cell architecture, stratification and differentiation status, the method is time-consuming, requires a ready availability of fresh tissue, and again, may exhibit inter-sample variability. Also, there is an increasing probability that stromal and urothelial cells will undergo unwanted mixing over culturing time, with the overgrowth of stromal cells relative to urothelial (Baker et al., 2014) . Moreover, the lumen of these models is commonly exposed to air, which does not happen in the bladder in situ, and may affect the maintenance of cell differentiation and/or polarization (Visňjar and Kreft, 2013) .

Due to the drawbacks of these strategies, other 3D experimental bladder models are emerging which attempt to biomimic the human urothelium using stem-or stem-like cells subjected to phases of propagation/expansion and stratification/ differentiation. Some of them were even patented (Cross, 2003) . Recently, Kim et al. reported the most complex 3D bladder multilayered model, mimicking an organized architecture of a mature organ through the assembly of a urothelium surrounding stroma and an outer muscle layer (Kim et al., 2020) .

Interestingly, apart from the 3D organoid described by Horsley et al. in 2018, all the aforementioned in vitro models were not exposed to urine, or if so, only for a few hours, up to a dayalthough a few studies report the use of urine components (e.g. ammonia, urea or water). This is important, since the exposure of the apical urothelial surface to urine seems to be essential for proper differentiation and adequate mucopolysaccharide production (Horsley et al., 2018) . Also, bacterial behavior and gene expression is affected by the presence of urine (Hagan et al., 2010; Reitzer and Zimmern, 2019) ; as urine is the natural context in which these microorganisms operate, models that do not include urine will be incomplete.

Notwithstanding the lack of a cellular immune response and the systemic host background that only in vivo studies can provide, it is expected that human-based in vitro biomimetic urothelium constructs with more physiological aspects will accelerate drug discovery for bladder diseases and even reduce the use of animal models (Baker et al., 2014) . In a UTI context, promising platforms already exist for the study of host-pathogen interactions, but they lack key parameters which have a profound impact on normal bladder physiology and may dramatically influence the experimental results obtained. Below we review these key aspects and why they should be included in the next generation of models. 

Advanced in vitro models more closely approximating the human urothelium would be valuable for assessing the cellular changes that occur with infection. Such models would ideally incorporate at least four unique structural and biophysical aspects theorized to be important for host-pathogen interactions including: (1) tissue architecture; (2) apical urine exposure;

(3) dynamic fluid flow; and (4) urothelial stretch. These key features are described in further detail below, in the context of unique bladder physiology, relevance to UTI, and current models or platforms of relevance.

As described above, the human bladder is lined by a unique transitional urothelium formed of approximately 5-7 layers of cells ( Figure 1 ) (reviewed in Khandelwal et al., 2009 ). Terminally 

Xenobiotic studies (Flieger et al., 2008) Tissue specimens Exfoliated cells Human urine (Belik et al., 2008; Dörrenhaus et al., 2000) Finite normal human urothelial (NHU) cells (potential for in vitro stratification and differentiation)

Biopsies/explants of histologically-normal bladder, renal pelvis and ureter, obtained at surgery from adult and/or pediatric patients Development of models/ characterization of bladder epithelium (Rheinwald and O'Connell, 1985; Hutton et al., 1993; Southgate et al., 1994; Varley et al., 2005; Cross et al., 2005; Southgate et al., 2007; Rubenwolf et al., 2009; Rubenwolf et al., 2012; Wezel et al., 2013) Urothelial cells Biopsies of dysfunctional bladder (Southgate et al., 2007) Urine of newborn children (Sutherland and Bain, 1972) From patients with interstitial cystitis and healthy controls (stretch experiments at 20% strain) (Sun et al., 2001; SUN and CHAI, 2004) Urine-derived stem cells Urine from healthy adult men (Lang et al., 2013) Human induced pluripotent stem cells (hiPSCs) derived from mature bladder urothelium hiPSC line FF-PB-3AB4 established from a healthy donor's peripheral blood mononuclear cells (PBMCs) (Suzuki et al., 2019) Normal human transitional cells

Ureter and embryonic bladder explants (Reznikoff et al., 1983) In vitro stratified cell layers Biopsy of lower urinary tract from adult patients undergoing open tumor surgery Bladder irrigation fluids (with exfoliated cells) Anatomically normal bladder and ureteral mucosa of children undergoing open kidney or bladder surgery (Sugasi et al., 2000) UROtsa cell monolayers (with some cytodifferentiation), available from ThermoFisher (Petzoldt et al., 1994; Petzoldt et al., 1995; Rossi et al., 2001) Organ/Tissue culture Bladder, renal pelvis and ureter of patients undergoing urological surgery with benign or no condition (Knowles et al., 1983; Scriven et al., 1997) Bladder from cadavers and deceased heart-beating brain-stem dead donors (Garthwaite et al., 2014) 3D multilayered urothelium (with cell differentiation), dynamic cultures mimicking the urine flow Human urine-derived stem cells, urothelial cells and smooth muscle cells (Wan et al., 2018) Organoids and spheroids for cancer studies Reviewed in (Wang et al., 2017; Vasyutin et al., 2019) differentiated umbrella or facet cells top the intermediate cell layer and face outward into the bladder lumen in contact with urine; these unique and specialized cells are formed by fusion of intermediate layers below, are very large but flattened when stretched, and form a barrier to urine despite their highly dynamic environment. Turnover is slow under healthy conditions (on the order of weeks), but when an umbrella cell is exfoliated, the urothelium ensures replacement through the fusing and differentiation of intermediate cells, which can be replaced in turn via the basal progenitor layer. Umbrella cells deploy a protective apical array of asymmetric unit membrane plaques comprised of uroplakin proteins (Wu et al., 2009 ). The urothelium also elaborates a mucopolysaccharide-rich layer of glycosaminoglycans (GAG) which offers additional protection, enhancing the urothelial barrier against urine (Janssen et al., 2013) . Indeed, the urothelium is one of the strongest epithelial barriers in vivo (Eaton et al., 2019) with native tissue transepithelial electrical resistance (TEER) values exceeding 3000 W*cm 2 (Cross et al., 2005) . The urothelium is not a passive barrier but is responsive to biomechanical cues, serving as a "mechanosensory conductor" as distention of the bladder stretches the urothelium during filling (Winder et al., 2014) . Given that the urothelium exhibits differential expression of specific markers throughout its layers, recapitulating the stratified human urothelium with appropriate markers and functions is critical for in vitro models of human bladder health and disease, and has particular relevance for UTI. The hypothesis that deep "quiescent intracellular reservoirs" in the intermediate cell layer may act as long-term reservoirs for recurrence requires a thicker multi-layer structure for studying this phenomenon. Moreover, as infection causes copious apical shedding, more than three layers are useful for retaining urothelial structure during bacterial insult. At the same time, studying adhesion properties of uropathogens, as well as innate epithelial response, ideally requires a human umbrella cellbacteria interface.

Much of our understanding of stratified urothelial architecture comes from animal models and human bladder biopsies (Jost et al., 1989) , but recreating this architecture in vitro is not trivial. The following trends emerged as we surveyed the literature to elucidate what factors are critical for achieving multiple cell layers alongside umbrella cell differentiation in humancell models.

Three-dimensional, multilayer urothelial organoid development in vitro has been reported using various cell sources: patientderived cells from biopsies and surgical specimens, either by enzymatic dissociation or explant (e.g., Southgate et al., 1994; Daher et al., 2004; Cross et al., 2005; Zhang and Atala, 2013) immortalized cell lines (e.g., UROtsa used by Rossi et al., 2001) ; commercial primary cells or spontaneously immortalized cells (HBEP and HBLAK used by Horsley et al., 2018) ; human urine stem cells (Lang et al., 2013) ; and iPS cells (Suzuki et al., 2019) . However, not all recapitulate the proper stratification and necessary hallmarks of the apical-most layer of umbrella cells, nor were used for infection. Most models report at least some important features, such as uroplakins, cytokeratins, and/or junctions, and include some form of visual evidence of a multilayer structure (e.g., electron or confocal microscopy). It remains unclear whether one cell source is superior to another, since there is no standard approach to urothelial differentiation either in terms of method or assessment. The ideal cell source would be readily available (either commercially and/or through well-described isolation techniques) and lead to reproducible differentiation in culture under defined conditions.

Although there is some confusion about the terms, organoids are usually defined as in vitro culture models that replicate features of native tissue organization, whereas spheroids are a specific, spherical form of organoid (Fang and Eglen, 2017; Vasyutin et al., 2019) . While each approach has its advantages and limitations, a flattened organoid approach may be more desirable for UTI models, as the lumen of spheroid cultures would be difficult to access for inducing infection and readouts of interest beyond imaging (e.g., secreted factors, barrier function assays). Flattened organoids, on the other hand, can establish 3D tissue architecture on a 2D surface allowing for ease of infection, imaging, sampling of media, barrier function (permeability assays or TEER), and the addition of other biophysical cues such as flow and/or stretch as described below.

It is well-established that "high" calcium (2-4 mM) is a major factor for in vitro stratification and differentiation of human urothelial models (e.g., (Southgate et al., 1994; Rossi et al., 2001; Daher et al., 2004) . Other additives have been used, such as FGF-7 (Tash et al., 2001) , FGF-10, PPAR-g agonists, and/or EGFR inhibitors (Suzuki et al., 2019) . Urine exposure may also be required for umbrella cell differentiation (Horsley et al., 2018) . Additional systematic studies of necessary factors may facilitate convergence towards a standardized media for urothelial models.

Beyond the liquid environment, the solid substrate on which cells are cultured can greatly influence 3D tissue architecture. Substrate mechanical properties have been shown to strongly affect cell structure and function, which may influence the optimal choice of materials . The most prevalent platforms have been tissue culture plastic (e.g., well plates) and Transwell ® inserts or similar porous membrane culture systems. The pore size typically used, when reported, is 0.4-0.45 µm diameter. Pore size and pore density may be critical factors in urothelial stratification and differentiation, but to our knowledge have not been systematically studied. Stratification is enhanced on porous membrane substrates, leading to increased urothelial tissue thickness and more uniform stratification (Cross et al., 2005; Suzuki et al., 2019) compared with well plates. In addition to providing nutrients to both sides of the tissue which could allow thicker layer formation, polarization can be achieved by providing differential cues on each side of the microporous membrane. Thus, a microporous substrate may be a requirement for advanced 3D urothelial organoid models. However, Transwells and similar platforms are static cultures not readily adaptable to dynamic biomechanical cues such as fluid flow and stretch. These advanced platforms will be described later.

The extracellular matrix (ECM) can influence cell adhesion, proliferation, differentiation and function (Hynes, 2009; Yue, 2014; Chaudhuri et al., 2020) . Collagen IV is the ECM coating of choice across multiple models, and bestowed accelerated outgrowth of urothelial explants compared with laminin or fibronectin (Daher et al., 2004) . Beyond substrate coatings, more complex ECM scaffolds may be beneficial for 3D urothelial model development. The Atala lab has performed extensive studies of decellularized scaffolds, including the bladder, and has described several approaches to 3D urothelial culture on ECM scaffolds (Zhang and Atala, 2013) . Such scaffolds are naturally porous and provide more complex environmental cues such as multiple ECM components and the topography present in native tissue. Whether decellularized bladder or similar scaffolds enhance urothelial stratification and differentiation compared with other culture substrates remains to be determined.

While many urothelial models are derived from a single cell source, some studies have co-cultured other cell types. The most common addition is fibroblasts (Vasyutin et al., 2019) , although endothelial cells (Sharma et al., 2021) and smooth muscle cells have also been used (Zhang and Atala, 2013) . Additional cell types found in the underlying stroma may provide important cues that influence urothelial proliferation, architecture and differentiation (Southgate et al., 1994) . A challenge with coculture is the inevitable increase in biological variability; alternative approaches such as defined soluble factors may be preferred when cross-talk between cell types is not critical to the study.

The urine microenvironment is an important consideration for UTI models from both the host and pathogen side. To our knowledge only two studies report the use of long-term urine exposure (>24 h) in urothelial models. One model implemented commercially available cells (primary HBEC and spontaneously immortalized HBLAK), differentiation media, and pooled urine in Millicell Transwell inserts (Horsley et al., 2018) . Urine was introduced into the insert (apical side of the cell culture) 24 hr after initiating differentiation of the confluent monolayer and maintained for several weeks. The authors reported that urine was necessary for stratification, differentiation of the umbrella cell layer and GAG elaboration. This model was also used to study infection with Enterococcus faecalis, resulting in urothelial sloughing and formation of intracellular colonies previously observed in human patient cells. Second, a recent pre-print described a bladder-on-chip system with human bladder epithelial cells, bladder microvascular cells and neutrophils in the Emulate microfluidic platform with flow and mechanical stretch (Sharma et al., 2021) . The bladder epithelium was exposed to diluted urine in the co-culture, including during infection. However, the effect of urine on model development and infection was not explicitly studied. In another recent example, a microfluidic platform with the human bladder epithelial cell line PD07i was used in experiments to infect UPEC with short-term (20 h) urine exposure under flow (Iosifidis and Duggin, 2020) .

Given that few human urothelial models use urine, much remains to be learned about this key variable. In vivo, the umbrella cells are exposed to urine and form a tight barrier against it, while being nourished by the underlying vasculature and other cells/tissue in close proximity. Although urine may be necessary for umbrella cell differentiation in vitro, it is a harsh environment and cannot be the sole fluid used, even short term. For an ideal in vitro model, the apical surface of the urothelium would be exposed to urine while the basolateral compartment would be supported by an appropriate differentiation media. This is why platforms incorporating a microporous substrate with apical and basal compartmentalization will be more suitable for advanced urothelial models. The simplest implementation would be a Transwell ® or similar permeable membrane systems, such as described by Horsley et al., 2018 . Other parameters to consider are the timing of urine introduction, acclimation to increasing concentrations of urine over time, and donor characteristics (sex, age, any diseases or conditions that may alter urine composition). More advanced models may implement platforms with fluid flow, such as systems described in the following sections.

Although existing static organoid models can capture some of human urothelial physiology, many biological questions probably can be answered only with a model incorporating additional biophysical aspects, including fluid flow and stretch, which are present in the human bladder and likely affect both normal physiology and infection dynamics. In vitro models that incorporate flow and/or stretch have been developed, but very few for the human bladder. We describe below how flow and stretch are important for both normal bladder function and UTI, and how incorporating these biomechanical cues into an in vitro model would be a valuable addition.

Given the shape and compliance of the bladder, fluid flow during voiding is extremely complex and non-uniform. Computational fluid dynamic (CFD) simulations of flow in the bladder-urethra system revealed that the peak shear stress experienced by cells is approximately 3 dyn/cm 2 (Jin et al., 2010)a moderate level compared with that found in the vascular system, for example. However, this level is only encountered by cells in the urethra, whereas the bladder umbrella cells experience extremely low shear, perhaps well below the levels required to activate shearinduced bacterial adhesion mechanisms. While this does not discount the hypothesis that flow-induced bacterial adhesion plays a role in bladder UTI, it does suggest that umbrella cells are perhaps less affected by this process in vivo, although it may be a likely mechanism in the urethra. A CFD-based study by (Atesçi et al., 2014) produced similar results but in terms of velocity profiles within the bladder and urethra. Although this simulation data cannot be directly extrapolated to assign shear stress values experienced at the bladder wall by umbrella cells, it indicates that it is lower than in the urethra.

Aside from fluid flow possibly affecting uropathogenic bacterial adherence, it may also influence other virulence behaviors including the propensity of UPEC to take on a filamentous form that increases virulence, adhesion, invasion and escape from immune surveillance (Justice et al., 2006; Andersen et al., 2012) . This may be particularly relevant for biofilm formation (Weaver et al., 2012) . In addition, a number of studies have shown that microfluidic flow provides physiological cues that guide tissue architecture; as one example, in an organotypic kidney model, fluid flow altered cell shape, protein expression and transport to better approximate in vivo organization (Jang et al., 2013) . Although currently understudied in the context of bladder cell culture models, flow even with low shear stress could be an important cue to enhance urothelial organoid development in vitro as well as subsequent studies of infection. Models of the urethra with higher shear could provide insights into mechanisms of bacterial adhesion and how uropathogens migrate into the bladder. Examples of models incorporating fluid flow are provided below.

Fluid flow has been incorporated into numerous in vitro model organ systems as a means for fluid and nutrient exchange, waste removal, to maintain cellular survival and tissue architecture, or to induce cellular differentiation (Baudoin et al., 2007; Carraro et al., 2008; Fritsche et al., 2009; Douville et al., 2011; Cheng et al., 2012; Park et al., 2012; Agarwal et al., 2013; Kim and Ingber, 2013; Torisawa et al., 2014; Zhang et al., 2014; Henry et al., 2017; Kasendra et al., 2018; Shin et al., 2019; Sidar et al., 2019; Azizipour et al., 2020) . Many liver and kidney models use fluid flow to maintain tissue structure, deliver drugs, measure reabsorption, or carry out metabolites for ADME-tox studies (Sin et al., 2004; Sivaraman et al., 2005; Kane et al., 2006; Lee et al., 2007; Chao et al., 2009; Mahler et al., 2009; Toh et al., 2009; Jang and Suh, 2010; Novik et al., 2010; Snouber et al., 2012; Legendre et al., 2013; Esch et al., 2015; Lee et al., 2017; Maschmeyer et al., 2015; Vedula et al., 2017; Zhang et al., 2017; Bale et al., 2019; Homan et al., 2019; Jalili-Firoozinezhad et al., 2019; Tan et al., 2020) . Disrupted flow patterns have been applied to induce damage that mimics tissue injury in lung and heart models (Huh et al., 2007; Giridharan et al., 2010; Huh et al., 2010; Khanal et al., 2011; Tavana et al., 2011; Nguyen et al., 2015) , and microfluidic flow systems have been created to study muscle contractility (Grosberg et al., 2012) , endothelial vascularization (Shin et al., 2004) , pulmonary thrombosis (Jain et al., 2018) , tumor-vascular invasion (Nguyen et al., 2019) , blood-brain barrier (Booth and Kim, 2012) , nerve injury (Park et al., 2013; Tsantoulas et al., 2013) and neuronal networks Xiao et al., 2013; van de Wijdeven et al., 2018) .

Microfluidic lung and gut models have also been used to study infection by bacteria and viruses. For example, fluid flow, either alone or together with cyclical strain, enabled not only proper tissue differentiation but also enhanced invasion by the infecting organism Kang et al., 2015; Barr et al., 2015; Benam et al., 2016; Villenave et al., 2017; Ortega-Prieto et al., 2018; Sunuwar et al., 2020; Tang et al., 2020; Thacker et al., 2020; Baddal and Marrazzo, 2021) . Fluid flow led to upregulated invasion of colonic epithelium by Shigella bacteria (in the presence of peristalsis-like mechanical strain) (Grassart et al., 2019) , increased epithelial barrier function in a lung epithelium/ macrophage influenza virus co-infection model (Deinhardt-Emmer et al., 2020) , and enhanced infection of alveolar or airway epithelial cells by SARS-CoV-2 Zhang et al., 2021) . Finally, Kim et al. used a microfluidic gut chip to study the contributions of the microbiome to intestinal pathophysiology and showed that flow plus peristalsis-like mechanical deformations were necessary to protect against bacterial overgrowth reminiscent of inflammatory bowel disease ; other studies have used microfluidic systems to investigate host-microbe interactions Shah et al., 2016) and microbial diversity (Tovaglieri et al., 2019; Jalili-Firoozinezhad et al., 2019) .

Despite its potential relevance, few in vitro bladder models have incorporated flow. One study used a Cellix chip to show that flow of UPEC in cell culture media enabled their adhesion to vascular endothelial cells (mimicking blood-borne dissemination) but had minimal effect on binding to bladder epithelial cells compared with static conditions (Feenstra et al., 2017) . However, this study was performed in 2D culture in the absence of urine, and the primary goal was to investigate bloodborne dissemination of UPEC rather than invasion into bladder cells per se. Another study demonstrated filamentous rod formation and secondary infection of bladder cells by UPEC under conditions of urine flow using a flow-cell chamber (Andersen et al., 2012) . The degree of filamentation was dependent on urine concentration, supporting the idea that flow can affect the infection behavior of UPEC in human cells. However, infection was tested in bladder cell cultures, not a fully differentiated urothelium. As mentioned above, a microfluidic platform with urine flow was used to study UPEC behavior in a human bladder epithelial cell line (Iosifidis and Duggin, 2020) . In a recent pre-print, Sharma et al., 2021 reported a bladder-onchip incorporating fluid flow, mechanical strain and urine exposure applied to differentiated bladder cell monolayers cocultured with microvascular cells (see Section 5.4.3) (Sharma et al., 2021) . Finally, attachment of E. coli to bladder cell line monolayers was found to be highest at low shear stress, allowing for maximal initial attachment of bacterial Fim H receptors via a slip-bond mechanism (Zalewska-Piątek et al., 2020) . Thus, how the low level of shear experienced by umbrella cells in vivo (Jin et al., 2010; Atesçi et al., 2014) affects their physiology and interaction with invading pathogens remains unclear. The above studies suggest the potential for uncovering further benefits of flow in UTI bladder models, including (but not limited to) maintenance of bladder epithelial differentiation and promotion of bacterial-epithelial interactions in a more physiological context.

To study the role of fluid flow in UTI, the ideal platform would provide user control over flow (for example, to mimic urination patterns, to provide perfusion or nutrient replenishment); have the potential to include stretch or other biophysical and biomechanical cues (e.g., ECM); maintain two fluid compartments to aid polarization of the urothelial organoid; incorporate capabilities to interrogate the tissue using various assays; and ideally scale to a throughput allowing experimental replicates while also accommodating multiple variables such bacterial strains and therapeutic dosing. Microfluidic-based approaches such as "microphysiological systems" or "organon-chip" platforms have been deployed extensively for other tissue models (recent reviews include Low et al., 2020; Peterson et al., 2020) and could be an equally promising strategy for UTI models. Bacterial biofilms and antibiotic therapies have been studied in cell-free systems with flow, where flow within a microfluidic configuration allowed analysis of biofilm growth and infection potential (Terry and Neethirajan, 2014; Wright et al., 2015) . As described above, bacterial infection of human cells has also responded to flow in an in vitro system capable of stretch (Grassart et al., 2019) . Such systems demonstrate the utility and impact of flow on bacterial infection in vitro, but have yet to be merged with throughput. Given the timeframe of urothelial organoid differentiation in vitro (2-3 weeks), throughput will be essential for advanced models seeking to study multiple variables or for drug discovery efforts. Flow has been employed in high throughput, although not in an infection context, and the flow mechanism in those studies afforded limited control of flow parameters (Vormann et al., 2021) . Pump-controlled flow has been deployed in high throughput in a microfluidic system with two fluid compartments for a liver model (Tan et al., 2020 ), but the system lacked a stretch component. A logical next step would be to incorporate flow into a high throughput system, while retaining the ability to actuate stretch for a two-fluid compartment system for advanced UTI models.

Many tissues in the human body undergo modest dynamic mechanical deformation, but the bladder expands drastically during the micturition cycle. Unlike the heart and blood vessels which are constantly under pressure from the blood, the bladder experiences the extremes of being completely empty and completely full, with maximum volume capacity of the human bladder ranging from 300-500 ml. Pressures experienced within the bladder range from 0-34 cmH 2 O (0-25 mmHg) while filling and can be as high as 200 cmH 2 O (100-150 mmHg) when full (Wolfe et al., 2015) . In vivo, pressure and stretch are coupled by the physiological processes of gradual urine filling, storage and sudden emptying, but these biomechanical cues can be studied independently ex vivo or in vitro.

Given the ability of the urothelium to increase its surface area in response to pressure or stretch, proscribing a mechanical strain experienced by the tissue is not a simple calculation, and data are difficult to obtain for humans. The urothelium can increase in apical surface area by~50% within hours of experiencing stretch (Wang et al., 2003) . Bladder tissue capacitance, a measure of umbrella cell surface area, increases over approximately 5 hr regardless of the filling rate at a pressure differential of 8 cmH 2 O (Truschel et al., 2002; Carattino et al., 2013) , which represents the pressure during the extended storage phase in rabbits (Levin and Wein, 1982) . Studies of rat bladders suggest than under normal filling loads, the bladder wall itself passively stretches 10-20% (Gloeckner et al., 2002) . Calculations based on responses of rabbit bladder tissue to pressure suggest that umbrella cells can respond to strains in this same range (Carattino et al., 2013) .

Cells in other mechanically active tissues have well-developed mechano-sensing properties; the bladder urothelium is no exception. Several signaling pathways and urothelial receptors contribute to bladder sensing of stretch, and are described in detail in (Janssen et al., 2017) . Biomechanical stretch is known to affect a variety of cellular behaviors, such as proliferation, differentiation, maturation, and tissue-specific functions (Zimmermann, 2013; Yu et al., 2016) . Limited published data exist regarding the effects of stretch on human urothelial cell behavior, although a few are highlighted below. As such, most of our understanding is derived from animal studies, typically rodent and rabbit, where species differences likely exist.

A primary urothelial function is to maintain the urine-blood barrier (Kreft et al., 2010) , no small biological feat during the dramatic and dynamic mechanical fluctuations that occur with bladder filling and voiding. It has been reported that TEER initially dips and then increases during prolonged stretch (Truschel et al., 2002) . Although TEER may drop with stretch, tight junctions are nevertheless maintained, as determined by immunofluorescent staining for claudins and ZO-1 and the limited permeability of tracer molecules (Carattino et al., 2013) . Another group studied human urothelial cell response to various strains and noted increased proliferation under 5% strain (Gao et al., 2018) .

During bladder filling and extended storage phases, umbrella cells undergo remarkable changes in surface area mediated by endo-and exocytosis (Truschel et al., 2002) . Changes in these rates are triggered by mechanical stretch, but not pressure, and begin within seconds of inducing stretch at pressures as low as 2 cmH 2 O ). This behavior was demonstrated ex vivo using rabbit bladder tissue by measured alterations in tissue capacitance, indicating changes in surface area, which increased up to 50% after 5 hr of prolonged stretch. Endocytosis was also remarkably upregulated with stretch as assayed by biotin-labeled membrane internalization and imaging of intracellular FITClabeled wheat germ agglutinin. Prolonged, rather than transient or short-term, stretch is required for observed changes in surface area. However, increased endocytosis occurs in as little as 5 min after initiation of stretched conditions (Truschel et al., 2002) .

Another study showed that mechanical stretch associated with filling and voiding increased the endo/exocytosis behavior of bladder epithelial cells, which use fusiform endocytic vesicle transport to increase and decrease surface membrane area during bladder expansion and contraction, respectively. This vesicle trafficking is of considerable interest because it is hypothesized to be a mechanism hijacked by certain bacterial strains that form intracellular colonies (Bishop et al., 2007) .

Despite the critical physiological relevance, an integrated stretch platform to support complex differentiation of urothelial organoids along with fluidics and polarized urine exposure has only begun to be explored. But at a more basic level, researchers have used multiple strategies to study stretch in urothelial cells, ranging from whole bladder distension to urothelial monolayer culture on commercially available stretch platforms. Most such studies have focused on the impact of mechanical stretch on ATP release, though few have evaluated human cells. ATP activity is linked to exocytosis and endocytosis of the mucosal surface through binding of umbrella P2 receptors; thus, the impact of stretch on ATP secretion inevitably affects membrane transport mechanisms (Wang et al., 2005) which, as pointed out previously, is potentially important for uropathogen invasion. Tanaka et al. used an organ bath to demonstrate the impact of stretch on ATP and prostaglandin E 2 release, which was dependent on volume-based rat bladder distension (Tanaka et al., 2011) . To interrogate the impact of stretch on ATP exocytosis, Mochizuki et al. developed elastic silicone chambers mounted on glass coverslips seeded with primary mouse urothelial cells for simultaneous stretch and in situ Ca2+ imaging. They demonstrated that the TRPV4 cation channel mediates the release of ATP and the influx of Ca2+ via a stretch-dependent mechanism (Mochizuki et al., 2009 ). This stretch methodology was also subsequently used to analyze mechanosensation in the bladder through piezo channels (Miyamoto et al., 2014) .

The importance of stretch on barrier function and tight junction proteins was demonstrated by mounting circular excised sections of rabbit urothelia in modified Ussing chambers and utilizing hydrostatic pressure-induced stretch (Truschel et al., 2002; Carattino et al., 2013) . Multiple studies used a similar system to study the polarization of stretch-induced ATP secretion, demonstrating ten times more concentrated ATP supernatant on the mucosal surface versus the basal surface in rabbits (Yu, 2015) . Although these models are fascinating, all results discussed thus far have been in non-human urothelial cell models and tissues.

A small subset of studies focusing on patient samples has demonstrated that mechanical stretch also affects primary human urothelial cells cultured in vitro. Using commercial stretch platforms, primary human cells isolated from patients with interstitial cystitis (IC), a chronic disease of unknown etiology, were shown to have significantly higher supernatant levels of ATP under stretch compared with stretched healthy control samples (Sun et al., 2001) . Similarly, stretched IC patient cells demonstrated higher levels of purinergic receptor subtype P2X3, which plays a role in transmitting pain signals to the central nervous system (SUN and CHAI, 2004) . This finding supports the hypothesis that urothelial cells can phenotypically mimic sensory neurons, with this phenotype driven by the presence of mechanical stretch. Therefore, primary human urothelial cells in stretch models can address crucial questions in urothelial biology.

In addition to the urothelium, the impact of mechanical stretch on other cell types such as smooth muscle cells (SMCs) isolated from the bladder has been explored. Bu et al. demonstrated that stretch-induced proliferation of SMCs correlated with the upregulation of metalloproteinases (MMPs) MMP-1, 2, 3, and 7 under 10% and 15% stretch conditions in a stretch-dependent manner (Bu et al., 2014) . Other studies have explored the impact of stretch on SMCs, demonstrating the importance of this mechanical stimulus for aspects of the bladder beyond the urothelium (McDermott et al., 2013; Zheng et al., 2016; Pingyu et al., 2019) .

Multiple commercial platforms provide the capability to study various in vitro tissues under mechanical stretch conditions, though certain parameters discussed in this review have yet to be incorporated into a platform capable of supporting a complete bladder infection model. Ideally, an in vitro system capable of uniform, biaxial strain within the range of 10-50% strain is desirable. To support a pragmatic and robust workflow, the ability to experiment with multiple replicates and image experimental samples in situ is vital; however, frequently used stretch platforms in their current forms are not likely to support the complexity of a fully differentiated bladder UTI model.

Existing commercial options achieve multiple-device throughput by implementing stretch on silicone-based substrates such as the uniaxial MCFX (CellScale) and STB-1400 (Strex Inc.) systems, achieving maximum strains of 12.5% and 20%, respectively. These companies also offer models that achieve biaxial stretch up to 20% in the XY plane, but most commercial biaxial models are only offered as single-well systems, such as the MCB1 (CellScale), or the STB-190-XY (Strex Inc.), which sacrifices throughput for the additional strain dimension. The most impressive system to date able to implement biaxial strain in relatively high throughput, the HT BioFlex ® (FlexCell), can actuate 24 wells at once, though the increased throughput sacrifices the maximum strain that the system can achieve, topping out at 8%. The company offers a lower throughput BioFlex ® model that achieves biaxial 20% strain in 6 devices simultaneously (FlexCell).

The ability to provide multi-faceted stimuli for cell-type differentiation is favorable; the C-Stretch system (IonOptix LLC) incorporates electrical stimulation alongside stretch to assist differentiation of naïve cultured cardiomyocytes. The ability to incorporate both fluid flow and stretch in these platforms would be widely beneficial for bladder and other in vitro models. Unfortunately, none of these systems incorporate them simultaneously, and all of the options discussed do not implement permeable substrates compatible with urine exposure differentiation strategies. Figure 2 depicts examples of various platforms used for in vitro bladder studies. To date, the closest relevant platform capable of integrating multiple mechanical stimuli is the Chip S-1 ® system (Emulate Inc.), which utilizes microfluidic flow paths and flexible side walls exposed to vacuum to deliver simultaneous flow up to 0.3 dynes/cm 2 , and demonstrated uniaxial strain of 10%. In a recent pre-print, this platform modeled infection in differentiated HTB9 bladder cancer cell monolayers over a 6hour voiding cycle, incorporating both flow and mechanical stretch with a porous membrane allowing for fluid transport between channels; neutrophils were recruited to sites of infection, and UPEC IBCs formed and persisted in the presence of antibiotics, supporting the hypothesis that IBCs play a role in recurrent infection (Sharma et al., 2021) . Despite these strengths, areas for improvement include urothelial stratification, throughput capacity, and membrane properties, specifically pore size, pore density and drug sorption issues in PDMS (Shirure and George, 2017) . To our knowledge the field still lacks a combined stretch-and flow-based platform with some degree of throughput to support a truly differentiated and stratified human urothelial organoid, fulfilling all the criteria necessary for exploring the complex mechanisms involved in human bladder UTIs (Figure 3 ).

Many mysteries still remain regarding the host/pathogen interactions of UPEC, to say nothing of the wide array of other understudied species that infect the urinary tract. On the horizon, technological advancements promise an exciting array of improvements to in vitro human models both generally, and for studying the infection biology of UTI in particular. The use of human cells also opens up the possibility of precision medicine by allowing the interrogation of UTI in the context of individual patients.

It is important to be aware of the limitations of in vitro systems. For instance, since most human cell-based models are closed structures, they still lack tissue-tissue interfaces, vascularization and circulation of immune cells. Therefore, beyond making the necessary improvements described in this review to isolated urothelial systems, we expect that linking up the bladder environment to its adjacent niches such as gut and kidney using "body-on-chip" platforms will add further nuance. Similarly, coculturing human immune cells and/or underlying human muscle cells, fibroblasts and stroma could breathe further life into existing models. Currently, however, the recapitulation of some crucial physiology is difficult, such as the transport and/or absorption of drugs, nutrients, oxygen, inflammatory molecules and the immune cells themselves. However, significant breakthroughs have occurred recently (reviewed in Ingber, 2020) . Once advanced strategies can be introduced into human-based models, such as the integration of self-organized capillary networks, a complex surrounding ECM, sensors for real-time functional measurements, and the long-term maintenance of co-cultures with different tissue types and their microbiota, the study of human pathophysiological events may well be more accurate and controllable than in animal models.

While animals will continue to reveal crucial insights about human infection biology, we must acknowledge their own limitations openly. We support a strategy whereby human cell model systems are used alongside animals to provide complementary information. It is no secret that paper referees and grant reviewers are quick to demand animal studies or to discount meticulous results gathered using even advanced human in vitro systems, and we quite agree with those (Ingber, 2020) who argue that the scientific community should consider whether such dampening activities are in the best spirit of scientific inquiry.

BM, CF, JR, CW and DF conceived the overall conception and design of the review. All authors contributed to writing various sections of the manuscript. All authors contributed to the article and approved the submitted version. 

We thank Dr. Jeffrey Borenstein (Draper) for his critical review of the manuscript as well as members of the Rohn lab (University College London) for their scientific insight and discussions. Content contained in Figure 2 previously appeared online and are used with permission. 

The Nature of Immune Responses to Urinary Tract Infections

Microfluidic Heart on a Chip for Higher Throughput Pharmacological Studies

Influence of High Fat Diet Feeding for 20 Weeks on Lower Urinary Tract Function in Mice

Adhesion, Invasion, and Agglutination Mediated by Two Trimeric Autotransporters in the Human Uropathogen Proteus Mirabilis

Escherichia Coli Uropathogenesis In Vitro: Invasion, Cellular Escape, and Secondary Infection Analyzed in a Human Bladder Cell Infection Model

Intracellular Bacterial Biofilm-Like Pods in Urinary Tract Infections'. Science

Notes on the Proteus Group of Organisms With Special Reference to a Case of Renal Infection by a Member of This Group

Merging Mythology and Morphology: The Multifaceted Lifestyle of Proteus Mirabilis

Erratum to "Does Urinary Bladder Shape Affect Urinary Flow Rate in Men With Lower Urinary Tract Symptoms?

Evolution of Biochip Technology: A Review From Lab

Refining Host-Pathogen Interactions: Organ-on-Chip Side of the Coin

Biomimetic Urothelial Tissue Models for the in Vitro Evaluation of Barrier Physiology and Bladder Drug Efficacy

A Thermoplastic Microfluidic Microphysiological System to Recapitulate Hepatic Function and Multicellular Interactions

Strengths and Limitations of Model Systems for the Study of Urinary Tract Infections and Related Pathologies. Microbiol

Subdiffusive Motion of Bacteriophage in Mucosal Surfaces Increases the Frequency of Bacterial Encounters

Development of a Renal Microchip for In Vitro Distal Tubule Models

Molecular Analysis of Asymptomatic Bacteriuria Escherichia Coli Strain VR50 Reveals Adaptation to the Urinary Tract by Gene Acquisition

Ribonucleases 6 and 7 Have Antimicrobial Function in the Human and Murine Urinary Tract

Improvements in Culturing Exfoliated Urothelial Cells In Vitro from Human Urine

The Role of Toll-Like Receptors (Tlrs) in Urinary Tract Infections (Utis). Cent. Eur

Small Airway-on-a-Chip Enables Analysis of Human Lung Inflammation and Drug Responses In Vitro

Urothelial Cultures Support Intracellular Bacterial Community Formation by Uropathogenic Escherichia Coli

Normalizing the Environment Recapitulates Adult Human Immune Traits in Laboratory Mice

Modulation of Host Innate Immune Response in the Bladder by Uropathogenic Escherichia Coli

Molecular Basis of Uropathogenic ≪Em<Escherichia Coli≪/Em< Evasion of the Innate Immune Response in the Bladder

Cyclic AMP-regulated Exocytosis of Escherichia Coli From Infected Bladder Epithelial Cells

Characterization of a Microfluidic In Vitro Model of the Blood-Brain Barrier (mbbb)

Invasion of Vaginal Epithelial Cells by Uropathogenic Escherichia Coli

Simulated Physiological Stretch-Induced Proliferation of Human Bladder Smooth Muscle Cells is Regulated by Mmps

Bladder Filling and Voiding Affect Umbrella Cell Tight Junction Organization and Function

In Vitro Analysis of a Hepatic Device With Intrinsic Microvascular-Based Channels

Antibiotic Resilience: A Necessary Concept to Complement Antibiotic Resistance?

Evaluation of a Microfluidic Based Cell Culture Platform With Primary Human Hepatocytes for the Prediction of Hepatic Clearance in Human

Effects of Extracellular Matrix Viscoelasticity on Cellular Behaviour'

Detection of Intracellular Bacteria in Exfoliated Urothelial Cells From Women With Urge Incontinence

Zonation Related Function and Ubiquitination Regulation in Human Hepatocellular Carcinoma Cells in Dynamic vs. Static Culture Conditions

Emerging Resistance Problems and Future Perspectives in Pharmacotherapy for Complicated Urinary Tract Infections

Nonhuman Primate Species as Models of Human Bacterial Sepsis

The Origins and Uses of Mouse Outbred Stocks

Evaluation of Immunocompetent Urinary Tract Infected Balb/C Mouse Model for the Study of Antibiotic Resistance Development Using Escherichia Coli CFT073 Infection

Age and Volume Dependent Normal Frequency Volume Charts for Healthy Males

Animal Models in the Evaluation of the Effectiveness of Phage Therapy for Infections Caused by Gram-Negative Bacteria From the ESKAPE Group and the Reliability of Its Use in Humans

Tamm-Horsfall Protein Protects the Urinary Tract Against Candida Albicans

Renal Scarring Caused by Vesicoureteric Reflux and Urinary Infection: A Study in Pigs

A Biomimetic Tissue From Cultured Normal Human Urothelial Cells: Analysis of Physiological Function

Biomimetic Urothelium

Mouse Models for Multiple Sclerosis: Historical Facts and Future Implications

Growth, Differentiation and Senescence of Normal Human Urothelium in an Organ-Like Culture

An in-Depth Comparison of the Porcine, Murine and Human Inflammasomes; Lessons From the Porcine Genome and Transcriptome. Veterinary Microbiol

Co-Infection With Staphylococcus Aureus After Primary Influenza Virus Infection Leads to Damage of the Endothelium in a

Direct Detection of Tissue-Resident Bacteria and Chronic Inflammation in the Bladder Wall of Postmenopausal Women With Recurrent Urinary Tract Infection

Interaction Networks, Ecological Stability, and Collective Antibiotic Tolerance in Polymicrobial Infections

Cultures of exfoliated epithelial cells from different locations of the human urinary tract and the renal tubular system

Combination of Fluid and Solid Mechanical Stresses Contribute to Cell Death and Detachment in a Microfluidic Alveolar Model

Expansion and Contraction of the Umbrella Cell Apical Junctional Ring in Response to Bladder Filling and Voiding

The M2 Muscarinic Receptor Mediates Contraction Through Indirect Mechanisms in Mouse Urinary Bladder

Impact of Proinflammatory Cytokines on the Virulence of Uropathogenic Escherichia Coli

Multi-Cellular 3D Human Primary Liver Cell Culture Elevates Metabolic Activity Under Fluidic Flow

Nonhuman Primate Models of Human Viral Infections

Actin-Gated Intracellular Growth and Resurgence of Uropathogenic Escherichia Coli

Unconventional Bacteria in Urinary Tract Disease: Gardnerella Vaginalis

Three-Dimensional Cell Cultures in Drug Discovery and Development

Adhesion of Escherichia Coli Under Flow Conditions Reveals Potential Novel Effects of FimH Mutations

Immunoreactivity of p63 in Monolayered and In Vitro Stratified Human Urothelial Cell Cultures Compared With Native Urothelial Tissue

Specific Selection for Virulent Urinary Tract Infectious Escherichia Coli Strains During Catheter-Associated Biofilm Formation

Primary Cultures of Human Urothelial Cells for Genotoxicity Testing

Ebpa Vaccine Antibodies Block Binding of Enterococcus Faecalis to Fibrinogen to Prevent Catheter-Associated Bladder Infection in Mice

Epidemiology of Urinary Tract Infections: Incidence, Morbidity, and Economic Costs

The Epidemiology of Urinary Tract Infection

Risk Factors for Second Urinary Tract Infection Among College Women

The Gene Expression Profile of Uropathogenic Escherichia Coli in Women With Uncomplicated Urinary Tract Infections Is Recapitulated in the Mouse Model

Pulmonary Tissue Engineering Using Dual-Compartment Polymer Scaffolds With Integrated Vascular Tree

Physiological Stretch Induced Proliferation of Human Urothelial Cells Via Integrin a6-FAK Signaling Pathway

Escherichia Coli From Urine of Female Patients With Urinary Tract Infections is Competent for Intracellular Bacterial Community Formation

Use of Donor Bladder Tissues for In Vitro Research

Enterococcus Faecalis Polymicrobial Interactions Facilitate Biofilm Formation, Antibiotic Recalcitrance, and Persistent Colonization of the Catheterized Urinary Tract

Immortalisation of Normal Human Urothelial Cells Compromises Differentiation Capacity

Comparative Study of the Organisation and Phenotypes of Bladder Interstitial Cells in Human, Mouse and Rat

Transient Microbiota Exposures Activate Dormant Escherichia Coli Infection in the Bladder and Drive Severe Outcomes of Recurrent Disease

Microfluidic Cardiac Cell Culture Model (mcccm)

Passive Biaxial Mechanical Properties of the Rat Bladder Wall After Spinal Cord Injury

Selection of Effective Antibiotics for Uropathogenic Escherichia Coli Intracellular Bacteria Reduction

Naturalizing Mouse Models for Immunology

Bioengineered Human Organ-on-Chip Reveals Intestinal Microenvironment and Mechanical Forces Impacting Shigella Infection

Replication Fitness Determines High Virulence of Influenza A Virus in Mice Carrying Functional Mx1 Resistance Gene

Muscle on a Chip: In Vitro Contractility Assays for Smooth and Striated Muscle

In Vivo Dynamics of Type 1 Fimbria Regulation in Uropathogenic Escherichia Coli During Experimental Urinary Tract Infection

Strain-Dependent Difference in Susceptibility of Mice to Experimental Ascending Pyelonephritis

Strain-Dependent Differences in Susceptibility of Mice to Experimental Pyelonephritis

Escherichia Coli Global Gene Expression in Urine From Women With Urinary Tract Infection

Ascending, Unobstructed Urinary Tract Infection in Mice Caused by Pyelonephritogenic Escherichia Coli of Human Origin

Contribution of Adhesion to Bacterial Persistence in the Mouse Urinary Tract

Difference in Susceptibility to Gram-Negative Urinary Tract Infection Between C3H/HeJ and C3H/HeN Mice. Infection Immun

Inhibition of Cyclooxygenase-2 Prevents Chronic and Recurrent Cystitis

Iron Acquisition of Urinary Tract Infection Escherichia Coli Involves Pathogenicity in Caenorhabditis Elegans

Toll-Like Receptor 11 (Tlr11) Interacts With Flagellin and Profilin Through Disparate Mechanisms

Organs-on-Chips With Integrated Electrodes for Trans-Epithelial Electrical Resistance (TEER) Measurements of Human Epithelial Barrier Function

What Is the Predictive Value of Animal Models for Vaccine Efficacy in Humans? Consideration of Strategies to Improve the Value of Animal Models

Urine Is Not Sterile: Use of Enhanced Urine Culture Techniques To Detect Resident Bacterial Flora in the Adult Female Bladder

Flow-Enhanced Vascularization and Maturation of Kidney Organoids In Vitro

Time Course and Host Responses to Escherichia Coli Urinary Tract Infection in Genetically Distinct Mouse Strains

Lipopolysaccharide-Responder and Nonresponder C3H Mouse Strains are Equally Susceptible to an Induced Escherichia Coli Urinary Tract Infection

A Urine-Dependent Human Urothelial Organoid Offers a Potential Alternative to Rodent Models of

Enterococcus Faecalis Subverts and Invades the Host Urothelium in Patients With Chronic Urinary Tract Infection

Interaction Between Antibiotic Resistance, Resistance Genes, and Treatment Response for Urinary Tract Infections in Primary Care

Acoustically Detectable Cellular-Level Lung Injury Induced by Fluid Mechanical Stresses in Microfluidic Airway Systems

Reconstituting Organ-Level Lung Functions on a Chip

Intracellular Lifestyles and Immune Evasion Strategies of Uropathogenic Escherichia Coli

Urothelial Tissue Culture for Bladder Reconstruction: An Experimental Study

Development of a Long-Term Ascending Urinary Tract Infection Mouse Model for Antibiotic Treatment Studies

The Extracellular Matrix: Not Just Pretty Fibrils

Is it Time for Reviewer 3 to Request Human Organ Chip Experiments Instead of Animal Validation Studies?

Distinct Morphological Fates of Uropathogenic Escherichia Coli Intracellular Bacterial Communities: Dependency on Urine Composition and pH

Muro-Neuro-Urodynamics ; a Review of the Functional Assessment of Mouse Lower Urinary Tract Function

Enzyme Multilayer Coatings Inhibit Pseudomonas Aeruginosa Biofilm Formation on Urinary Catheters

Proteus Mirabilis Biofilms and Catheter-Associated Urinary Tract Infections. Virulence

Primary Human Lung Alveolus-on-a-chip Model of Intravascular Thrombosis for Assessment of Therapeutics

Il-21-Driven Neoplasms in SJL Mice Mimic Some Key Features of Human Angioimmunoblastic T-Cell Lymphoma

A Complex Human Gut Microbiome Cultured in an Anaerobic Intestine-on-a-Chip

Human Kidney Proximal Tubule-on-a-Chip for Drug Transport and Nephrotoxicity Assessment

A Multi-Layer Microfluidic Device for Efficient Culture and Analysis of Renal Tubular Cells

Stiffness Sensing by Cells

Mannose-Resistant Proteus-Like Fimbriae Are Produced by Most Proteus Mirabilis Strains Infecting the Urinary Tract, Dictate the In Vivo Localization of Bacteria, and Contribute to Biofilm Formation

Urothelium Update: How the Bladder Mucosa Measures Bladder Filling

The Distribution and Function of Chondroitin Sulfate and Other Sulfated Glycosaminoglycans in the Human Bladder and Their Contribution to the Protective Bladder Barrier

Dynamics Analysis of Bladder-Urethra System Based on

The Microbiome and Host Mucosal Interactions in Urinary Tract Diseases

Construction of a Urease-Negative Mutant of Proteus Mirabilis: Analysis of Virulence in a Mouse Model of Ascending Urinary Tract Infection

The Morphology of Normal Human Bladder Urothelium

Differentiation and Developmental Pathways of Uropathogenic Escherichia Coli in Urinary Tract Pathogenesis

Filamentation by Escherichia Coli Subverts Innate Defenses During Urinary Tract Infection

Age-Related Changes in Function and Gene Expression of the Male and

Liver-Specific Functional Studies in a Microfluidic Array of Primary Mammalian Hepatocytes

Liver Sinusoid on a Chip: Long-term Layered Co-Culture of Primary Rat Hepatocytes and Endothelial Cells in Microfluidic Platforms

Heat Shock Protein 72 (Hspa1b) Gene Polymorphism and Toll-Like Receptor (Tlr) 4 Mutation are Associated With Increased Risk of Urinary Tract Infection in Children

Development of a Primary Human Small Intestine-on-a-Chip Using Biopsy-Derived Organoids

Experimental Pyelonephritis. XIV. Pyelonephritis in Normal Mice Produced by Inoculation of E. Coli Into the Bladder Lumen During Water Diuresis

There IS No SUCH Thing as A C57bl/6 Mouse!

Mouse Microbiota Models: Comparing Germ-Free Mice and Antibiotics Treatment as Tools for Modifying Gut Bacteria

Report on a Series of Experiments to Determine the Value of Certain Solutions as Germicides, Made for Dr

Ischemia/ Reperfusion Injury of Primary Porcine Cardiomyocytes in a Low-Shear Microfluidic Culture and Analysis Device

Cell Biology and Physiology of the Uroepithelium

Spectrum of Bacterial Colonization Associated With Urothelial Cells From Patients With Chronic Lower Urinary Tract Symptoms

Creation of Bladder Assembloids Mimicking Tissue Regeneration and Cancer

Human Gut-on-a-Chip Inhabited by Microbial Flora That Experiences Intestinal Peristalsis-Like Motions and Flow

Gut-on-a-Chip Microenvironment Induces Human Intestinal Cells to Undergo Villus Differentiation

Contributions of Microbiome and Mechanical Deformation to Intestinal Bacterial Overgrowth and Inflammation in a Human Gut-on-a-Chip

Quantification of Filamentation by Uropathogenic Escherichia Coli During Experimental Bladder Cell Infection by Using Semi-Automated Image Analysis

Uropathogenic Escherichia Coli Induces Extrinsic and Intrinsic Cascades To Initiate Urothelial Apoptosis

Uropathogenic Escherichia Coli Potentiates Type 1 Pilus-Induced Apoptosis by Suppressing NF-Kappab

Long-Term Organ Culture of Normal Human Bladder

Defining the Relationship Between Vaginal and Urinary Microbiomes

Formation and Maintenance of Blood-Urine Barrier in Urothelium

Bacterial Persistence: A Model of Survival in Changing Environments

Age-Dependent Polydipsia in the SWR/J Mouse

Renal Deficiency Associated With Diabetes Insipidus in the SWR/J Mouse

The Immune Response to Infection in the Bladder

Keratin Expression Profiling of Transitional Epithelium in the Painful Bladder Syndrome/Interstitial Cystitis

Prevalence of Gardnerella Vaginalis in the Urinary Tract

Prevention of Mucosal Escherichia Coli Infection by FimH-Adhesin-Based Systemic Vaccination

Self-Renewal and Differentiation Capacity of Urine-Derived Stem Cells After Urine Preservation for 24 Hours

Community-Onset Urinary Tract Infections: A Population-Based Assessment

Virulent Synergistic Effect Between Enterococcus Faecalis and Escherichia Coli Assayed by Using the Caenorhabditis Elegans Model

3D Gut-Liver Chip With a PK Model for Prediction of First-Pass Metabolism

An Artificial Liver Sinusoid With a Microfluidic Endothelial-Like Barrier for Primary Hepatocyte Culture

Metabolic Characterization of Primary Rat Hepatocytes Cultivated in Parallel Microfluidic Biochips

Phenotypic Profile of SWR/J and A/J Mice Compared to Control Strains: Possible Mechanisms Underlying Resistance to Obesity on a High-Fat Diet

The Production of Coliform Infection in the Urinary Tract of Rabbits

Response of the In Vitro Whole Bladder (Rabbit) Preparation to Autonomic Agonists

Roles of the Vagina and the Vaginal Microbiota in Urinary Tract Infection: Evidence From Clinical Correlations and Experimental Models

Bacteriophages are Synergistic With Bacterial Interference for the Prevention of Pseudomonas Aeruginosa Biofilm Formation on Urinary Catheters

Single-Cell Transcriptomes of Mouse Bladder Urothelium Uncover Novel Cell Type Markers and Urothelial Differentiation Characteristics

Compounds Targeting YadC of Uropathogenic Escherichia Coli and its Host Receptor Annexin A2 Decrease Bacterial Colonization in Bladder

Bladder Expression of CD Cell Surface Antigens and Cell-Type-Specific Transcriptomes

Expert Panel Recommendations on Lower Urinary Tract Health of Women Across Their Life Span

Organs-on-Chips: Into the Next Decade

Curli of Uropathogenic Escherichia Coli Enhance Urinary Tract Colonization as a Fitness Factor

Gut Uropathogen Abundance is a Risk Factor for Development of Bacteriuria and Urinary Tract Infection

Characterization of a Gastrointestinal Tract Microscale Cell Culture Analog Used to Predict Drug Toxicity

Proteinuria of Nonautoimmune Origin in Wild-type Fvb/Nj Mice

Observation of Intracellular Bacterial Communities in Urinary Sediment Using Brightfield Microscopy; a Case Report

A Four-Organ-Chip for Interconnected Long-Term Co-Culture of Human Intestine, Liver, Skin and Kidney Equivalents

Annotations on Emerging Concerns About Antibiotic-Resistant Urinary Tract Infection

Interstitial Cells in the Urinary Bladder-Localization and Function

Alterations in Acetylcholine, PGE2 and IL6 Release From Urothelial Cells Following Treatment With Pyocyanin and Lipopolysaccharide

Collaboration Between Distinct Rab Small Gtpase Trafficking Circuits Mediates Bacterial Clearance From the Bladder Epithelium

A TRP Channel Senses Lysosome Neutralization by Pathogens to Trigger Their Expulsion

Possible Role of L-form Switching in Recurrent Urinary Tract Infection

Siderophore Vaccine Conjugates Protect Against Uropathogenic Escherichia Coli Urinary Tract Infection

Functional Role for Piezo1 in Stretch-evoked Ca2+ Influx and ATP Release in Urothelial Cell Cultures*

The TRPV4 Cation Channel Mediates Stretch-evoked Ca2+ Influx and ATP Release in Primary Urothelial Cell Cultures

Induction and Evasion of Host Defenses by Type 1-Piliated Uropathogenic Escherichia Coli

Establishment of a Persistent Escherichia Coli Reservoir During the Acute Phase of a Bladder Infection

Mechanisms of Uropathogenic Escherichia Coli Persistence and Eradication From the Urinary Tract

Bone Morphogenetic Protein 4 Signaling Regulates Epithelial Renewal in the Urinary Tract in Response to Uropathogenic Infection

In Vitro Investigations of Tissue-Engineered Multilayered Urothelium Established From Bladder Washings

Advances in Understanding the Human Urinary Microbiome and Its Potential Role in Urinary Tract Infection

A Biomimetic Pancreatic Cancer on-Chip Reveals Endothelial Ablation Via ALK7 Signaling

Effects of Physiologic Mechanical Stimulation on Embryonic Chick Cardiomyocytes Using a Microfluidic Cardiac Cell Culture Model

A Porcine Model for Urinary Tract Infection

Decreased Incidence of Renal Tumors in DMN Treated Balb/C Mice After Orchiectomy

A Microfluidic Hepatic Coculture Platform for Cell-Based Drug Metabolism Studies

Drug and Vaccine Development for the Treatment and Prevention of Urinary Tract Infections

A Mucosal Imprint Left by Prior Escherichia Coli Bladder Infection Sensitizes to Recurrent Disease

Biofilm-Specific Antibiotic Tolerance and Resistance

Androgens Enhance Male Urinary Tract Infection Severity in a New Model

Renal Scar Formation and Kidney Function Following Antibiotic-Treated Murine Pyelonephritis

3D Microfluidic Liver Cultures as a Physiological Preclinical Tool for Hepatitis B Virus Infection

Neuromuscular Junction in a Microfluidic Device

Chip-Based Comparison of the Osteogenesis of Human Bone Marrowand Adipose Tissue-Derived Mesenchymal Stem Cells Under Mechanical Stimulation

Host Cathelicidin Exacerbates Group B Streptococcus Urinary Tract Infection

Application of Microphysiological Systems in Biopharmaceutical Research and Development

Culture and Characterisation of Human Urothelium In Vivo and In Vitro

Immortalisation of Human Urothelial Cells

Cyclic Stretch Promotes Proliferation and Contraction of Human Bladder Smooth Muscle Cells by Cajal-Mediated C-Kit Expression in Interstitial Cells

Role of Vaccines for Recurrent Urinary Tract Infections: A Systematic Review

TLR-and CXCR1-dependent Innate Immunity: Insights Into the Genetics of Urinary Tract Infections

Bacterial Biofilm Formation in the Urinary Bladder of Spinal Cord Injured Patients

Rapid Growth and Metabolism of Uropathogenic Escherichia Coli in Relation to Urine Composition

Growth and Characterization of Normal Human Urothelium In Vitro

Intermediate Filament Proteins as Distinguishing Markers of Cell Type and Differentiated State in Cultured Human Urinary Tract Epithelia

Intracellular Bacteria in the Pathogenesis of Escherichia Coli Urinary Tract Infection in Children

Detection of Intracellular Bacterial Communities in a Child With Escherichia Coli Recurrent Urinary Tract Infections

Detection of Intracellular Bacterial Communities in Human Urinary Tract Infection

Utilization of an Intracellular Bacterial Community Pathway in Klebsiella Pneumoniae Urinary Tract Infection and the Effects of FimK on Type 1 Pilus Expression

Molecular Variations in Klebsiella Pneumoniae and Escherichia Coli FimH Affect Function and Pathogenesis in the Urinary Tract

The Immortalized UROtsa Cell Line as a Potential Cell Culture Model of Human Urothelium

Urinary Tract Infection in Older Adults

Expression and Localisation of Aquaporin Water Channels in Human Urothelium In Situ and In Vitro

Aquaporin Expression Contributes to Human Transurothelial Permeability in Vitro and Is Modulated by Nacl

Asymptomatic Bacteriuria Escherichia Coli Are Live Biotherapeutics for UTI

Renal Tubulointerstitial Lesions in CBA/J Mice

Azithromycin and Ciprofloxacin: A Possible Synergistic Combination Against Pseudomonas Aeruginosa Biofilm-Associated Urinary Tract Infections

Reconstitution of Human Urothelium From Monolayer Cultures

Uropathogen Resistance and Antibiotic Prophylaxis: A Meta-Analysis

A Microfluidics-Based In Vitro Model of the Gastrointestinal Human-Microbe Interface

Role of Enterococcus Faecalis Surface Protein Esp in the Pathogenesis of Ascending Urinary Tract Infection

Dynamic Persistence of Intracellular Bacterial Communities of Uropathogenic Escherichia Coli in a Human Bladder-Chip Model of Urinary Tract Infections

Glia Co-Culture With Neurons in Microfluidic Platforms Promotes the Formation and Stabilization of Synaptic Contacts

Human Intestinal Morphogenesis Controlled by Transepithelial Morphogen Gradient and Flow-Dependent Physical Cues in a Microengineered Gut

Hedgehog/Wnt Feedback Supports Regenerative Proliferation of Epithelial Stem Cells in Bladder

Endothelialized Networks With a Vascular Geometry in Microfabricated Poly(Dimethyl Siloxane)

Design Considerations to Minimize the Impact of Drug Absorption in Polymer-Based Organ-on-a-Chip Platforms

Human Organ Chip-Enabled Pipeline to Rapidly Repurpose Therapeutics During Viral Pandemics

Long-Term Flow Through Human Intestinal Organoids With the Gut Organoid Flow Chip (Goflowchip)

Characterization of the Ebpfm Pilus-Encoding Operon of Enterococcus Faecium and its Role in Biofilm Formation and Virulence in a Murine Model of Urinary Tract Infection

A Comparative Phenotypic and Genomic Analysis of C57BL/6J and C57BL/6N Mouse Strains

The Design and Fabrication of Three-Chamber Microscale Cell Culture Analog Devices With Integrated Dissolved Oxygen Sensors

A Microscale in Vitro Physiological Model of the Liver: Predictive Screens for Drug Metabolism and Enzyme Induction'

Novel Three-Dimensional Organoid Model for Evaluation of the Interaction of Uropathogenic Escherichia Coli With Terminally Differentiated Human Urothelial Cells

Toll-Like Receptor Responses of Normal Human Urothelial Cells to Bacterial Flagellin and Lipopolysaccharide

Analysis of Transcriptomic and Proteomic Profiles Demonstrates Improved Madin-Darby Canine Kidney Cell Function in a Renal Microfluidic Biochip

Transcriptome of Uropathogenic Escherichia Coli During Urinary Tract Infection

TLR-Mediated Immune Responses in the Urinary Tract

Normal Human Urothelial Cells In Vitro

Differentiation Potential of Urothelium From Patients With Benign Bladder Dysfunction

Selective Depletion of Uropathogenic E. Coli From the Gut by a FimH Antagonist

Biofilm Formation of Klebsiella Pneumoniae on Urethral Catheters Requires Either Type 1 or Type 3 Fimbriae

Urinary Tract Infections: Disease Panorama and Challenges

Randomized, Placebo-Controlled Phase 2 Trial of a Lactobacillus Crispatus Probiotic Given Intravaginally for Prevention of Recurrent Urinary Tract Infection

In VITRO Engineering OF Human STRATIFIED Urothelium: ANALYSIS of ITS Morphology AND Function

Up-Regulation of P2X3 Receptor During Stretch of Bladder Urothelial Cells From Patients With Interstitial Cystitis

Augmented STRETCH Activated ADENOSINE Triphosphate RELEASE From BLADDER Uroepithelial CELLS in PATIENTS With INTERSTITIAL Cystitis

Mechanical Stimuli Affect Escherichia Coli Heat-Stable Enterotoxin-Cyclic GMP Signaling in a Human Enteroid Intestine-Chip Model

Culture of Cells From the Urine of Newborn Children

Directed Differentiation of Human Induced Pluripotent Stem Cells Into Mature Stratified Bladder Urothelium

Genetic Factors in Host Resistance to

Variable ADHERENCE to NORMAL Human URINARY-TRACT

Staphylococcus Saprophyticus ATCC 15305 is Internalized Into Human Urinary Bladder Carcinoma Cell Line 5637

Staphylococcus Saprophyticus Surface-Associated Protein (Ssp) is Associated With Lifespan Reduction in Caenorhabditis Elegans

Association of TLR2 Gene Arg753Gln Polymorphism With Urinary Tract Infection in Children

Pattern Recognition Receptors and Inflammation

Differential Roles of TLR2 and TLR4 in Recognition of Gram-Negative and Gram-Positive Bacterial Cell Wall Components

Modulation of Stretch Evoked Adenosine Triphosphate Release From Bladder Epithelium by Prostaglandin E2

Mapping of Cystitis Due to Streptococcus Agalactiae and Escherichia Coli in Mice Identifies a Unique Bladder Transcriptome That Signifies Pathogen-Specific Antimicrobial Defense Against Urinary Tract Infection

Correction: A High-Throughput Microfluidic Microphysiological System (PREDICT-96) to Recapitulate Hepatocyte Function in Dynamic, Re-Circulating Flow Conditions

Human Organs-on-Chips for Virology

Fibroblast GROWTH Factor-7 REGULATES Stratification OF the BLADDER Urothelium

Epithelium Damage and Protection During Reopening of Occluded Airways in a Physiologic Microfluidic Pulmonary Airway Model

A Novel Microfluidic Wound Model for Testing Antimicrobial Agents Against Staphylococcus Pseudintermedius Biofilms

A Lung-on-Chip Model of Early Mycobacterium Tuberculosis Infection Reveals an Essential Role for Alveolar Epithelial Cells in Controlling Bacterial Growth

Comparative Genomics of Antibiotic-Resistant Uropathogens Implicates Three Routes for Recurrence of Urinary Tract Infections

Targeted Disruption of Mouse EGF Receptor: Effect of Genetic Background on Mutant Phenotype

Bacteria-Induced Uroplakin Signaling Mediates Bladder Response to Infection

Antigen-Specific Responses Accelerate Bacterial Clearance in the Bladder

Complete Overview of Protein-Inactivating Sequence Variations in 36 Sequenced Mouse Inbred Strains

A Microfluidic 3D Hepatocyte Chip for Drug Toxicity Testing

Bone Marrow-on-a-Chip Replicates Hematopoietic Niche Physiology In Vitro

Toll-Like Receptor Expression in C3H/HeN and C3H/HeJ Mice During Salmonella Enterica Serovar Typhimurium Infection

Species-Specific Enhancement of Enterohemorrhagic E. Coli Pathogenesis Mediated by Microbiome Metabolites

Stretch-Regulated Exocytosis/Endocytosis in Bladder Umbrella Cells

Probing Functional Properties of Nociceptive Axons Using a Microfluidic Culture System

Tensin2-deficient Mice on FVB/N Background Develop Severe Glomerular Disease

Group B Streptococcus (GBS) Urinary Tract Infection Involves Binding of GBS to Bladder Uroepithelium and Potent but GBS-Specific Induction of Interleukin 1a

Implications of CD94 Deficiency and Monoallelic NKG2A Expression for Natural Killer Cell Development and Repertoire Formation

Structuring a Multi-Nodal Neural Network In Vitro Within a Novel Design Microfluidic Chip

Autocrine Regulation of Human Urothelial Cell Proliferation and Migration During Regenerative Responses In Vitro

Novel Roles of Lysines 122, 125, and 58 in Functional Differences Between Human and Murine MD-2

Bladder Organoids and Spheroids: Potential Tools for Normal and Diseased Tissue Modelling

A Comparative Review of Toll-Like Receptor 4 Expression and Functionality in Different Animal Species

A Microfluidic Renal Proximal Tubule With Active Reabsorptive Function

Natural Resistance to Intracellular Infections: Nramp1 Encodes a Membrane Phosphoglycoprotein Absent in Macrophages From Susceptible (Nramp1 D169) Mouse Strains

Human Gut-On-A-Chip Supports Polarized Infection of Coxsackie B1 Virus In Vitro

Air-Liquid and Liquid-Liquid Interfaces Influence the Formation of the Urothelial Permeability Barrier In Vitro

Implementation of a Human Renal Proximal Tubule on a Chip for Nephrotoxicity and Drug Interaction Studies

A non-Canonical Autophagy-Dependent Role of the ATG16L1(T300A) Variant in Urothelial Vesicular Trafficking and Uropathogenic Escherichia Coli Persistence

The Potential of Organoids in Urological Cancer Research

ATP and Purinergic Receptor-Dependent Membrane Traffic in Bladder Umbrella Cells

Analysis of Hydrostatic Pressure-Induced Changes in Umbrella Cell Surface Area

Urothelium With Barrier Function Differentiated From Human Urine-Derived Stem Cells for Potential Use in Urinary Tract Reconstruction

Fluid Flow Induces Biofilm Formation in Staphylococcus Epidermidis Polysaccharide Intracellular Adhesin-Positive Clinical Isolates

Plexin-A4-semaphorin 3A Signaling is Required for Toll-like Receptor-and Sepsis-Induced Cytokine Storm

Plasticity of In Vitro-Generated Urothelial Cells for Functional Tissue Formation

Signalling Molecules in the Urothelium

11.9 Pressures in the Body," in College Physics for AP ® Courses

Bacterial Persister Cell Formation and Dormancy

Microfluidic Wound Model for Studying the Behaviors of Pseudomonas Aeruginosa in Polymicrobial Biofilms

Uroplakins in Urothelial Biology, Function, and Disease

In Vitro Binding of Type 1-Fimbriated Escherichia Coli to Uroplakins Ia and Ib: Relation to Urinary Tract Infections

Simultaneous Generation of Gradients With Gradually Changed Slope in a Microfluidic Device for Quantifying Axon Response

Strain Distribution Pattern of Susceptibility to Immune-Mediated Nephritis

Duration of Urination Does Not Change With Body Size

Personal Clinical History Predicts Antibiotic Resistance of Urinary Tract Infections

Polarized ATP Distribution in Urothelial Mucosal and Serosal Space is Differentially Regulated by Stretch and Ectonucleotidases

Biology of the Extracellular Matrix: An Overview

Distinct Apical and Basolateral Membrane Requirements for Stretch-Induced Membrane Traffic At the Apical Surface of Bladder Umbrella Cells

Impact of Mechanical Stretch on the Cell Behaviors of Bone and Surrounding Tissues

Single-Cell Transcriptomic Map of the Human and Mouse Bladders

A Shear Stress Micromodel of Urinary Tract Infection by the Escherichia Coli Producing Dr Adhesin

Multisensor-Integrated Organs-on-Chips Platform for Automated and Continual in Situ Monitoring of Organoid Behaviors

Urothelial Cell Culture: Stratified Urothelial Sheet and Three-Dimensional Growth of Urothelial Structure

Advances in Organ-ona-Chip Engineering

Patient-Specific 3d Microfluidic Tissue Model for Multiple Myeloma

Biomimetic Human Disease Model of SARS-CoV-2-Induced Lung Injury and Immune Responses on

Nitric Oxide Mediates Stretch-Induced Ca2+oscillation in Smooth Muscle

Pharmacologic Characteristics of Bladder Micturition Function in Anesthetized Mice

Uroplakin Ia is the Urothelial Receptor for Uropathogenic Escherichia Coli: Evidence From In Vitro FimH Binding

Comparison of Allergic Lung Disease in Three Mouse Strains After Systemic or Mucosal Sensitization With Ovalbumin Antigen

Infection on a Chip: A Microscale Platform for Simple and Sensitive Cell-Based Virus Assays

Biomechanical Regulation of In Vitro Cardiogenesis for Tissue-Engineered Heart Repair

Urinary Tract Infection in a Small Animal Model: Transurethral Catheterization of Male and Female Mice

Conflict of Interest: Authors CW, DF, EM, BC, and JC were or are currently employed by The Charles Stark Draper Laboratory, Inc., a not-for-profit research and development organization that develops hardware for advanced biological models. JR and BM have received research funding from AtoCap Ltd., a University College London spinoff company, to develop novel cures for urinary tract infection and bladder cancer, and JR has share options in the company. JR and CF have also received basic research funding from Pfizer.The remaining 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.Copyright © 2021 Murray, Flores, Williams, Flusberg, Marr, Kwiatkowska, Charest, Isenberg and Rohn. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.