Inhibition of COX-2 signaling favors E. coli during urinary tract infection

Background To avoid the overuse of antibiotics, non-steroidal anti-inflammatory drugs (NSAIDs), acting via cyclooxygenase (COX) inhibition, have been used to reduce pain and as an alternative treatment for uncomplicated urinary tract infections (UTIs). However, clinical studies evaluating NSAIDs versus antibiotics have reported an increased risk of acute pyelonephritis. Therefore, we hypothesized that COX inhibition could compromise the innate immune response and contribute to complications in patients with uncomplicated UTI. Results We here demonstrate that in particular COX-2 inhibition led to decreased expression of the antimicrobial peptides psoriasin and human β-defensin-2 in human uroepithelial cells. Psoriasin expression was altered in neutrophils and macrophages. COX-2 inhibition also had impact on the inflammasome mediated IL-1β expression in response to uroepithelial E. coli infection. Further, COX-2 inhibition downregulated free radicals and the epithelial barrier protein claudin 1, favoring infectivity. In addition, conditioned media from COX-2 inhibited uroepithelial cells infected with E. coli failed to activate macrophages. Conclusions Taken together, our data suggests an adverse innate immune effect of COX-2 inhibition on uroepithelial cells during UTI. Supplementary Information The online version contains supplementary material available at 10.1186/s12950-023-00356-9.


Introduction
Urinary tract infections (UTIs), often caused by E. coli [1], are one of the most common bacterial infections, second only to respiratory infections with respect to antibiotic prescription [2].During UTI, pain associated with voiding is a major symptom and considered to be influenced by an inflammatory reaction.As a result, non-steroidal anti-inflammatory drugs (NSAIDs) are used to dampen the release of pro-inflammatory cytokines and chemokines as well as the pain triggered by elevated prostaglandin E2 levels [3,4].Since 50-60% of women experience at least one UTI in their lifetime and 25-30% of these have recurrent infections, the disease has great impact on the patient's quality of life [5].Although antibiotics are the drug of choice for treating UTI, the increasing antibiotic resistance amongst isolates of E. coli causing UTI [6] demands alternative treatment strategies.
For many patients, UTI is a self-limiting disease [7].Therefore, in an attempt to avoid unneccesary antibiotic use, NSAIDs and similar drugs have been tried as an alternate regimen for mild cystitis.An earlier clinical trial supported this, by suggesting non-inferiority of ibuprofen compared to antibiotics for the treatment of symptomatic uncomplicated UTI [8].Such change of treatment would decrease the antibiotic consumption and reduce the selective pressure for bacteria to develop resistance.In contrast, recent clinical studies have instead shown that NSAIDs were less effective than antibiotics in treating UTI and were associated with more symptoms and with a longer disease duration.Specifically, there was an increased risk of developing acute pyelonephritis in NSAIDs treated patients [9][10][11][12].
Antimicrobial peptides (AMPs) are part of the innate immune response and amongst the body's first line of defense towards invading pathogens.AMPs include cathelicidin [13], β defensin-1 [14], the RNase superfamily [15,16] and psoriasin [17], residing in epithelial and immune cells.NSAIDs mainly act by inhibiting cyclooxygenase (COX)-1 and 2 signalling [18].Although the use of NSAIDs over antibiotics remains alluring, empirical evidence regarding the possible influence of NSAIDs on the innate immune defense during E. coli infection has not been elucidated.We hypothesize that NSAIDs will have immunocompromised effect on the innate immune response during acute infection that in turn will affect the rate of E. coli infection.Therefore, in this study, we aim to investigate the impact of COX-2 inhibitor and underlying relevant mechanisms on the possibility of bacterial infection in the urinary tract.

COX inhibitors
Specific COX-1 inhibitor, SC-560 (1µM) and COX-2 inhibitor, SC-236 (5µM) (Sigma) were dissolved in DMSO and used in all the experiments.Equivalent volume of DMSO alone served as the vehicle control.Above used concentrations were not cytotoxic for human uroepithelial cells, 5637 and T-24 as confirmed by XTT cell viability assay [19] (Supplementary Fig. 1).

Bacterial strain
Uropathogenic E. coli strain CFT073 [20] was used for in vitro and in vivo experiments.Bacteria were grown over night on blood agar plates at 37 ° C, resuspended in 1X phosphate-buffered saline (PBS), measured spectrophotometrically and confirmed by viable count.

Mouse model of UTI
Mice experiments were approved by the Northern Stockholm Animal Ethics Committee, and experiments were carried out according to the guidelines of the Federation of Laboratory Animal Science Association and in compliance with the Committee's requirements.Female C57BL/6j mice were obtained from Janvier Laboratories.Animals were housed following standard procedures.10 mg/kg of SC-236 were administered once daily via oral gavage at least 7 days prior to the infection.Mice were anaesthetized using isoflurane and transurethrally infected with 0.5 × 10 8 colony-forming units (CFU) of E. coli CFT073 in 50 µl of PBS.After 24 h infection, mice were sacrificed and respective urinary bladders were aseptically removed, cut open and washed with PBS to remove urine and non-adherent bacteria.To determine the total bacterial load, adhered and intracellular bacteria, bladders were homogenized in 1 ml of PBS, serially diluted and plated on blood agar plates.

Primary neutrophil isolation
Human neutrophils were prepared from buffy coats obtained at the Karolinska University Hospital blood central.According to Swedish law, the use of anonymized buffy coats does not require specific ethical approval.Neutrophils were isolated using a hyperosmotic Percoll (GE Healthcare) solution and subsequent centrifuged (580 x g, 15 min) using standard protocol [22].Blood separated out in 6 distinct bands.Neutrophil layers were carefully collected, and10 ml HBSS without Ca 2+ and Mg 2+ were mixed by inverting the tube and centrifuged at 350 x g for 10 min. 2 ml RBC lysis buffer was added to the pellet, vortexed slowly and centrifuged 250 x g for 5 min.The above described lysis step was repeated to remove the RBCs.Further, cell pellet was washed with HBSS without Ca 2+ and Mg 2+ , centrifuged 250 x g for 5 min to obtain the neutrophils.

Cell infection assays
Cell experiments were carried out in 24 well cell culture plates.TERT-NHUC cells were cultured in Primaria (BD).All other cell lines T-24, 5637, differentiated HL-60 (dHL-60) and differentiated THP-1 (dTHP-1) were cultured in Costar plates.Cells were grown in the presence of 5µM SC-236 for 24 h prior to the infection.After 24 h, old media was removed and replaced with fresh media supplemented with 5µM of SC-236, without antibiotics and serum.10 6 CFU/ml (MOI 5) or 2 × 10 6 CFU/ml (MOI 10) of E. coli CFT073 were added to nearly confluent pre-treated cells and incubated in 37 °C at 5% CO 2 and 80% humidity.At pre-set time points like 15, 30, 60 and 120 min, cells were washed once with PBS and harvested for further analysis.

Bacterial survival assay
5637 and dTHP-1 cells were infected with MOI 5 E. coli in 100 µl of PBS per well, centrifuged for 1 min at 350 g.For assessment of bacterial adhesion, cells were infected for 30 min and washed with PBS.In survival assays, 5637 cells were washed with PBS after 2 h infection to remove non-adherent bacteria and supplemented with fresh medium with or without 5µM of SC-236 for another 2 h.However, for dTHP-1, 100 µg/ml of gentamicin was used for 30 min to kill the extracellular bacteria.In COX-2 inhibited cells, medium was supplemented with 5µM of SC-236 throughout the entire experiment.At indicated time points, cells were lysed with 200µL of ice-cold 0.1% Triton-X-100 in PBS (pH-7.4) and scraped.Cell lysates were serially diluted in PBS and plated on blood agar plates.Cells treated with 5µM of SC-236 were compared to DMSO treated cells.Rate of adhesion and survival were calculated by number of adhered or intracellular bacteria in relation to the total number of bacteria from the same experiment.

Free radical formation assay
5637 and dTHP-1 cells were infected as described earlier after incubation with 5µM of SC-236 for 24 h.Supernatants were collected and mixed with equal volumes of Griess reagent (Invitrogen) based on the manufacturer's protocol.Optical density was measured at 550 nm, and free nitrite was evaluated and normalized to DMSO treated control cells.For total ROS analysis, 10 µM DCFH-DA (Sigma) was added to the cells and were incubated at 37° C and 5% CO 2 for another 2 h.Fluorescence intensity was measured at excitation 485 nm and emission 530 nm (Fluostar Omega).Similarly, for mitochondrial ROS analysis after 1 h of E. coli infection, cells were first washed with 1× Hanks' Balanced Salt Solution (HBSS), then 5 µM Mitosox (Life Technologies) was added and left for 30 min, then live cell imaging was done using Zeiss LSM 700 to measure the mitochondrial ROS.Fluorescence intensity was quantified using ImageJ software.

Conditioned media-activation assay
T-24 cells were pretreated with 5µM of SC-236 for 36 h, followed by 2 h infection with E. coli CFT073.Conditioned media was harvested and used for stimulation of dTHP-1, dHL-60 and NKL cells with 1:2 ratio.dHL-60 and NKL cells were harvested after 6 h activation whereas dTHP-1 cells were harvested after 2 h.Early activation marker CD11b (BD Biosciences), CD63 (BD Biosciences) and psoriasin (Santacruz) were used to stain dHL-60 cells whereas M1 activation marker CD86 (BD Biosciences) was used for dTHP-1 cells using standard recommendation of the manufacturer.For NKL cells, Golgi plug/ Brefeldin (final concentration 5 µg/ml, for intracellular proteins) was dispensed after 1 h of stimulation and continued for a total of 6 h activation.Thereafter, NKL cells were washed with PBS + 1% fetal calf serum, live/dead marker (Thermo Scientific) was added and incubated for 30 min at 4 °C.Cells were washed and thereafter fixed and permeabilized with 1X fixation/permeabilization solution (BD Biosciences), Cells were thereafter washed in perm/wash buffer and stained with anti-Granzyme B (BD Biosciences, diluted in perm/wash buffer, for 30 min at 4 °C in the dark.Finally, cells were washed in perm/ was buffer, dissolved in PBS and data acquired at BD FACS Fortessa and analyzed in FlowJo software.

Total RNA isolation and quantitative real-time PCR (qPCR)
Total RNA was extracted using the RNeasy Mini kit (Qiagen) according to the manufacturer's protocol.The concentration and purity of RNA was determined with nanodrop, and up to 0.5 µg of RNA was transcribed to cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems).Expression of target genes was analyzed using SYBR Green reagent in Rotor-Gene PCR cycler (Corbett Life Science) with gene specific primers (Supplementary Table 1).Relative expressions of target genes were presented as 2 −∆CT and fold change as 2 −∆∆CT compared to uninfected or nontreated control.

Western blot
T-24 cells were pretreated with 5µM of SC-236 for 24 h, followed by E. coli CFT073 infection for 3 h with MOI 10.After the infection time, cells were washed with 1 x PBS, 100 µl of RIPA buffer with 1% protease inhibitor cocktail (Sigma) were added and cells are scraped on ice.The cells were homogenized using a syringe and needle.The protein concentration of the samples was measured with the DC protein assay (Bio-Rad Laboratories).Equal amounts of sample were mixed with Laemmli buffer and boiled for 5 min in 95 °C.The samples (10 µg) were separated by 4-20% SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride (PVDF) membrane (Bio-Rad Laboratories).The PVDF membrane was blocked with 3% BSA for 1 h.Caspase 1 protein was detected using a mouse monoclonal (AdipoGen Life Sciences) against human caspase 1. NLRP3 was detected using a mouse monoclonal (Abnova) against human NLRP3.GAPDH was detected with a rabbit polyclonal antibody (Santa Cruz Biotechnology Inc).All primary antibodies were incubated overnight at 4 °C.As secondary antibodies, goat anti rabbit IgG (horseradish peroxidase, HRP) (Abcam) and goat anti mouse IgG (HRP) (Abcam) were used and incubated for 1 h at room temperature.The blots were developed using Luminata Forte Western HRP Substrate (Merck Millipore).

Caspase 1 activity assay
Bladder epithelial cells, T-24 and 5637 were grown in respective media to reach 80% confluence in a 96-well plate.Cells were pretreated with 5µM of SC-236 for 24 h and pre-incubated with 50 µM caspase 1 substrate Ac-YVAD-AMC (Enzo Life Sciences) for 1 h at 37 ˚C 5% CO 2 prior to the E. coli infection.Thereafter, the cells were infected with E. coli at MOI 10.Samples were analyzed after 6 h of infection with a fluorescent plate reader (Cytation 3) at excitation/emission settings of 340/440 nm.Substrate with only medium was used as control to subtract the basal fluorescence later.

Enzyme-linked immunosorbent assay
T-24, dTHP-1, NKL and dHL-60 cells were treated with 5µM of SC-236 for 36 h, followed by 2 h infection.Supernatants were collected and centrifuged at 350 g for 10 min and stored at -80 °C until needed.IL-1β (R&D Biosystems) ELISA was analyzed according to the manufacturer's recommendations.Cells treated with DMSO were served as control.

Statistical analysis
All statistical tests were performed in Graph pad Prism version 5. Data were obtained from Students unpaired t-test, non-parametric test using Mann Whitney u test, paired t-test and non-parametric one-way Anova, Bonferonni's multiple comparison, Dunett's one-way Anova test as appropriate.Differences with p values below 0.05 were considered statistically significant.

COX-2 inhibition specifically decreases antimicrobial peptide expression
COX enzymes are responsible for the formation of prostaglandin E synthase 2 [23].Therefore, we first investigated the effect of both COX-1 and 2 inhibitions in 5637 uroepithelial cells on the mRNA expression of prostaglandin E synthase 2, PGES2.A significant inhibition was demonstrated in the presence of COX-2 inhibitor, SC-236, whereas no difference was observed with COX-1 inhibitor, SC-560 (Fig. 1A).Likewise, treatment with the COX-2 inhibitor downregulated the AMP psoriasin, encoded by S100A7, in human uroepithelial cells, 5637 (Fig. 1B) and TERT-NHUC (Supplementary Fig. 2A) during non-infected state and after E. coli infection on both mRNA (Fig. 1C) and protein levels (Fig. 1D).Similar results were also observed for another AMP, human β defensin-2 (hBD2) in 5637 cells (Supplementary Fig. 2B-D).Therefore, we continued further with COX-2 inhibition only.
During acute infection, neutrophils contribute to the antibacterial activity by producing and releasing antimicrobial peptides [24].Interestingly, SC-236 treatment reduced the S100A7 mRNA levels in human primary neutrophils isolated from blood (Supplementary Fig. 2E) and DMSO differentiated promyeloblasts (dHL-60) (Supplementary Fig. 2F).Recruitment of neutrophils is dependent on the release of inflammogenic host factors from infected uroepithelial cells [24].Therefore, we investigated the effect of conditioned media from E. coli infected uroepithelial T-24 cells, on the expression of psoriasin in dHL-60 cells.Stimulation of dHL-60 cells with conditioned media resulted in decreased psoriasin on the mRNA (Supplementary Fig. 2G), and increased protein levels (Supplementary Fig. 2H) compared with unstimulated cells.Macrophages, in addition to neutrophils, also play an important role in the response to E. coli infection [25].In contrast to dHL-60 cells, conditioned media supplemented to THP-1 differentiated macrophages, dTHP-1, did not affect the expression of psoriasin.However, addition of SC-236 followed by E. coli infection in dTHP-1 macrophages showed a trend of reduced psoriasin at mRNA (Fig. 1E) but a significant reduction in protein levels (Fig. 1F).

COX-2 inhibition prevents inflammasome mediated IL-1β secretion and activation of immune cells
Upon infection, IL-1β is produced by the inflammasome pathway [26].E. coli infected SC-236 treated T-24 cells showed lower mRNA expression of the inflammasome markers NLRP3 (Fig. 2A), ASC (Fig. 2B) and CASPASE1 (Fig. 2C) compared to vehicle controls.Similarly, after 3 h of E. coli infection, SC-236 treated T-24 cells showed no change in NLRP3 levels or cleavage of pro-caspase 1 (Fig. 2D).Thus, indicating inhibition of the inflammasome pathway in comparison to control infected cells, whereas in uninfected cells, we did not observe any influence of SC-236 on inflammasome markers (Fig. 2D).Further to confirm, caspase 1 activity was measured in uroepithelial cells infected with E. coli.After 6 h of infection, caspase 1 activity was lowered in SC-236 treated cells in both the uroepithelial cells T-24 (Fig. 2E) and 5637 (Supplementary Fig. 3A).
Inhibition of caspase 1 via SC-236 resulted in lower expression of IL-1β on the mRNA (Fig. 2F) and the secreted protein after infection with E. coli (Fig. 2G).Similar to the in vitro results, E. coli infected SC-236 treated mice showed a trend of lower Il1b expression (Fig. 2H).
Immune cells are often triggered by secretory molecules released from infected uroepithelial cells thereby promoting migration to the site of infection.Therefore, activation markers were analyzed in dTHP-1 and NKL cells after treatment with conditioned media from T-24 uroepithelial cells.We observed lower expression of CD86 (Fig. 2I) in dTHP-1 cells treated with conditioned media from SC-236 and E. coli infected uroepithelial cells.In addition, intracellular granzyme B content was higher in SC-236 treated NKL cells in comparison with vehicle treated control cells (Fig. 2J).

Compromised free radical formation is associated with COX-2 inhibition
During infection, proinflammatory cytokines influence the release of free radicals [27].Therefore, we investigated the effect of COX-2 inhibition on free radicals.Interestingly, NOS2 mRNA was downregulated in SC-236 treated 5637 cells (Fig. 3A) and remained reduced after E. coli infection both on mRNA (Fig. 3B) and protein levels (Fig. 3C).In agreement with the mRNA levels, lower levels of free NO were observed in SC-236 treated 5637 cells (Fig. 3D).In addition to the reduction in NO, total intracellular ROS was also downregulated in SC-236 treated 5637 cells (Fig. 3E).Similarly, after E. coli infection and SC-236 treatment of 5637, a significant reduction in mitochondrial ROS was observed (Fig. 3F).To verify whether this observation was limited to uroepithelial cells only, dTHP-1 cells were analyzed showing lower NO (Fig. 3G) and total intracellular ROS (Fig. 3H) and lower mitochondrial ROS (Fig. 3I) in SC-236 treated cells compared with untreated cells.
Antioxidants interact with free radicals and neutralize their effects [28].We observed a lower expression free radical in SC-236 treated uroepithelial cells and macrophages, therefore we expected a strong effect of SC-236 on regulators of cellular resistance to antioxidants, like nuclear factor erythroid 2 related factor 2 (NRF2).Interestingly, we observed increased expression of NRF2 in SC-236 treated human uroepithelial cells, 5637 (Fig. 4A).The effect was even more prominent with nuclear translocation of NRF2 in SC-236 treated E. coli infected cells (Fig. 4A, zoomed images).However, we also observed increased KEAP1 in these conditions.Furthermore, to confirm the effect of NRF2 translocation, NRF2 target genes like GCLC (Fig. 4B) and HMOX1 (Fig. 4C) were investigated and found to have increased mRNA expression in SC-236 treated uroepithelial cells.

COX-2 inhibition compromises epithelial integrity and promotes E. coli infection
Psoriasin is known to induce barrier proteins [29,30], and maintain epithelial integrity.The decreased psoriasin levels seen upon treatment with COX-2 inhibition of uroepithelial cells further prompted us to investigate the effect of SC-236 on tight junction proteins.A significant reduction in the expression of claudin-1 in SC-236 treated 5637 and TERT-NHUC cells, at the mRNA (Fig. 5A and Supplementary Fig. 4A) and protein level (Fig. 5B) was observed.
Loss of epithelial integrity encouraged us to investigate the implication of E. coli infection in uroepithelial cells.E. coli survived better in SC-236 treated 5637 (Fig. 5C) and TERT-NHUC cells (Supplementary Fig. 4B).In line with our in vitro results in uroepithelial cells, bladders from mice treated with SC-236 showed limited to no clearance of E. coli relative to control mice, although it did not reach significance (Fig. 5D).In contrast, a significant difference was observed in dTHP-1 cells (Fig. 5E).Higher bacterial burden could be due to alteration in epithelial cell surface receptors.We therefore evaluated the expression of TLR4, responsible for the interaction with E. coli.Our results showed that TLR4 was upregulated both at mRNA (Fig. 5F) and protein (Fig. 5G) levels in 5637 cells after treatment with SC-236.These results could explain the increased adhesion of E. coli that was observed in SC-236 treated 5637 cells (Supplementary Fig. 4C).

Discussion
We here demonstrate that the inhibition of COX-2 signaling pathway compromised the innate immune response and weakened the epithelial barrier.These findings offer an understanding of the clinical observation that NSAID treatment of patients was associated with increased pyelonephritis episodes [9][10][11][12].
Our results show that inhibition of COX-2 in uroepithelial cells mediates decreased psoriasin and hBD2 in response to E. coli.Inhibition of COX-2 signaling in phagocytic cells like neutrophils and macrophages resulted in compromised expression of psoriasin.These results are in line with a previous report where COX-2 expression was critical for the expression of hBD2 and hBD3 in response to S. aureus infection in keratinocytes [31].
Expression of the COX-2 precursor prostaglandin E2 is known to trigger the expression of antimicrobial peptides [32], it has also been shown that the antimicrobial peptide, LL-37, stimulates the activation of COX-2 signaling in keratinocytes [33].In the context of uroepithelial cells, there are no reports suggesting the influence on antimicrobial peptides.However, we cannot completely rule out the possible interaction of psoriasin, since S100B, a member of the S100 family proteins, to which psoriasin belongs, induces COX-2 signaling in microglial cell lines [34].
The epithelial immune response is critical for effective clearance of bacteria during UTI [35].In an autocrine pathway, IL-1β triggers the expression of COX-2 [36,37] and cytokines like IL-8 [38,39] which are important for the recruitment of neutrophils.Therefore, this pathway is crucial in response to invading pathogens in the urinary tract.We showed that COX-2 inhibition in uroepithelial cells compromised their IL-1β secretion, corresponding to what has been shown in ovarian granulosa cells [40].Moreover, PGE2 is known to regulate the expression and release of cytokines particularly in innate immune cells like macrophages, neutrophils and NK cells [41].Interestingly, when uroepithelial cells failed to release IL-1β in the presence of COX-2 inhibitor and even during E. coli infection, we observed increased expression of IL-1β in macrophages and NK cells which might be assumed to compensate the loss of effect in an in vivo model.Uropathogenic E. coli is known to activate NLRP3 and caspase 1 mediated inflammasome in bladder epithelial cells [42].It is also important to note that, during in vivo infection exfoliation of infected superficial bladder epithelial cells from urinary bladder is dependent on the NLRP3 inflammasome [43].In this process, the human body clears the infected cells.Our result showing lower expression of inflammasome markers in COX-2 inhibitor treated and E. coli infected cells supports impediment of the inflammasome pathway.Secretion of IL-1β takes place with activation of the inflammasome [44].The cleavage of the IL-1β precursor occurs by active caspase 1 [44].In the present study, we observed lower caspase 1 activity in E. coli infected COX-2 inhibited uroepithelial cells, which resulted in compromised IL-1β levels.Altogether, our data suggests negative impact of COX-2 inhibition in uroepithelial cells.This might affect the exfoliation of infected cells in patients treated with NSAIDs, allowing the bacteria to survive longer in the host cells and be a potential threat for recurrent infection.
Activation of macrophages is important for effective phagocytic activity and bacterial clearance [45].In agreement with this, conditioned media of COX-2 inhibited E. coli infected uroepithelial cells failed to fully activate macrophages, as infected uroepithelial cells release innate immune molecules to recruit innate immune cells to the site of infection.Similarly, granzyme B is an important determinant for the cytotoxicity of NKL cells, and NK cell protein levels of granzyme B can increase rapidly upon activation.Interestingly, we observed an increase in intracellular granzyme B in COX-2 inhibited NKL cells.We believe that this at least partly is a response to activating cytokines in the conditioned media.Although target cell interaction is the major trigger for granzyme B release, granzyme B can be non-specifically released after prolonged IL-2 stimulation [46].We therefore speculate that the increased levels of granzyme B in SC-236 treated cells, could also be due to delayed release.This possibility could overall compromise the innate immunity and effective clearance of bacteria in patients treated with COX-2 inhibitors or NSAIDs.
Furthermore, COX-2 inhibited uroepithelial cells and macrophages produced less free radicals.Moreover, NO seems to interfere directly with the activity of COX-2, as production of prostaglandins by the COX-2 pathway modulates the biosynthesis of NO [47].It has also been reported that ROS activates the expression of the COX-2 downstream protein, PGE2 [48].Therefore, decreased levels of free radicals might further influence the COX-2 signaling pathway.
Antimicrobial peptides also largely contribute to the expression of epithelial barrier proteins.Along with psoriasin, several antimicrobial peptides, such as LL-37 and hBD2, are reported to have a beneficial effect on strengthening the epithelial barrier [29,30,49,50].Compromised psoriasin levels after COX-2 inhibition are an important observation in uroepithelial cells.It is known that COX-2 deficiency leads to epithelial barrier dysfunction [51] and application of PGE2 to ischemic-injured ileal mucosa recovered the barrier function [52].As psoriasin is known to upregulate claudin-1 in human keratinocytes [29], we cannot rule out the indirect impact of COX-2 inhibitor, SC-236, on claudin-1 in uroepithelial cells via the psoriasin mediated pathway.Loss of epithelial integrity favors bacterial entry into the host [53].Specifically, it has been shown that loss in claudin-1 levels resulted in increased IL-1β as well as inflammation during staphylococci infection [54].
Apart from compromising epithelial barrier function, COX-2 inhibition increased the expression of TLR4 in uroepithelial cells, which facilitated greater E. coli attachment.This is of high significance as PGE2 is also known to downregulate the expression TLR4 [55].As COX-2 inhibition downregulated PGE2 in uroepithelial cells, the increase in TLR4 levels is expected and as a consequence, resulted in greater attachment of E. coli to uroepithelial cells.
Overall, we conclude that COX-2 inhibition compromises parameters of the innate immune response including the production of antimicrobial peptides, proinflammatory cytokines, epithelial barrier integrity and free radicals.Alterations in the prostaglandin receptor may influence the innate immune response status as inhibition of COX-2 creates a favorable condition for bacterial attachment and intra-cellular survival.Our data suggest that balanced COX-2 signaling is vital for the innate immune response against invading pathogens.Therefore, patients with lower UTI given NSAIDs to help treat inflammation and the pain during voiding, might be at risk of an iatrogenic-derived severe infection.Karolinska Institutet's Research Foundation (SM, HB, and AB).HB is supported by grants from the Swedish Society for Medical Research, the Swedish Cancer Foundation, the Swedish Medical Association, Region Stockholm (clinical research appointment), Hudfonden, Clas Groschinsky, Åke Wiberg, and Magnus Bergvall.Open access funding provided by Karolinska Institute.

Fig. 2
Fig. 2 Inhibition of inflammasome pathway in COX-2 inhibited E. coli infected uroepithelial cells, compromised activation of immune cells.Expression of (A)NLRP3, (B)ASC, (C)CASPASE1 mRNA after 2 h E. coli infection (n = 4) in T-24 cells.(D) Representative western blot analysis of NLRP3 and Pro-caspase 1 compared to housekeeping gene, GAPDH after 3 h E. coli infection (n = 3).(E) Relative caspase 1 activity assay in T-24 cells after 6 h E. coli infection.Expression of the cytokine IL-1β at the (F) mRNA level (n = 3) and (G) secreted level from the supernatants using ELISA (n = 3) after 2 h E. coli infection in T-24 cells.(H) Expression of IL1b mRNA in the urinary bladders of SC-236 treated mice and after 24 h E. coli infection (n = 6/7).(I) Expression of CD86 in dTHP-1 cells after 2 h stimulation with conditioned media obtained from E. coli infected bladder epithelial cells, T-24 (n = 3) using microscopy.Relative densitometry of CD86 is shown.(J) Measurement of intracellular granzyme B levels in NKL cells after 6 h stimulation with conditioned media obtained from E. coli infected bladder epithelial cells, T-24 (n = 8).Mean fluorescence intensity (MFI) was measured and normalized to control cells.In vitro analysis was performed in either duplicate or triplicate.T-24 cells were treated with 5µM SC-236 for 24 h followed by E. coli infection.Conditioned media was obtained from 5µM SC-236 treated and E. coli infected T-24 cells after 2 h.Data are shown as mean + SEM.Results from mice are presented as median.Significance levels mentioned as *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001