[Ca2+]i Oscillations and IL-6 Release Induced by α-Hemolysin from Escherichia coli Require P2 Receptor Activation in Renal Epithelia*

Background: The virulence factor HlyA elicits [Ca2+]i oscillations in renal epithelial cells. Results: These oscillations and following IL-6 release are reduced by inhibition or lack of P2Y2 receptors. Conclusion: The effects of HlyA in renal epithelial cells are mediated by P2Y2 receptors. Significance: ATP release and P2Y2 receptor activation are essential parts of the early interaction between Escherichia coli and the renal epithelium. Urinary tract infections are commonly caused by α-hemolysin (HlyA)-producing Escherichia coli. In erythrocytes, the cytotoxic effect of HlyA is strongly amplified by P2X receptors, which are activated by extracellular ATP released from the cytosol of the erythrocytes. In renal epithelia, HlyA causes reversible [Ca2+]i oscillations, which trigger interleukin-6 (IL-6) and IL-8 release. We speculate that this effect is caused by HlyA-induced ATP release from the epithelial cells and successive P2 receptor activation. Here, we demonstrate that HlyA-induced [Ca2+]i oscillations in renal epithelia were completely prevented by scavenging extracellular ATP. In accordance, HlyA was unable to inflict any [Ca2+]i oscillations in 132-1N1 cells, which lack P2R completely. After transfecting these cells with the hP2Y2 receptor, HlyA readily triggered [Ca2+]i oscillations, which were abolished by P2 receptor antagonists. Moreover, HlyA-induced [Ca2+]i oscillations were markedly reduced in medullary thick ascending limbs isolated from P2Y2 receptor-deficient mice compared with wild type. Interestingly, the following HlyA-induced IL-6 release was absent in P2Y2 receptor-deficient mice. This suggests that HlyA induces ATP release from renal epithelia, which via P2Y2 receptors is the main mediator of HlyA-induced [Ca2+]i oscillations and IL-6 release. This supports the notion that ATP signaling occurs early during bacterial infection and is a key player in the further inflammatory response.

ulence factors such as the exotoxin ␣-hemolysin (HlyA) 2 (2,3). HlyA belongs to the Repeat-in-Toxin (RTX) family, of which some are able to form pores in cell membranes, rendering the cells permeable to ions and water and ultimately lyse them (for review, see Ref. 4). Recently, we discovered that the HlyA-induced lysis of erythrocytes is amplified by ATP release and subsequent P2X receptor activation (5,6), because either inhibition of the P2X receptors or scavenging of extracellular ATP markedly reduced the HlyA-induced hemolysis (5). This cellular amplification system was found to be equally relevant in hemolysis induced by other cytolysins such as ␣-toxin from Staphylococcus aureus (7), leukotoxin A from Aggregatibacter actinomycetemcomitans (8) and complement activation (9). Interestingly, the size of the created pore by bacterial toxins is critical for whether or not cell lysis is amplified by ATP (10). The authors show that bacterial pores too narrow to allow passage of ATP do not exhibit P2X receptor-dependent amplification of cell lysis (10). In support for this, we have shown that HlyA-induced ATP release is likely to occur directly through the HlyA-pore itself (11). This is exceedingly interesting because it implies that ATP will be released from any cell attacked by HlyA and thus, P2 receptors will be activated in an auto-and paracrine fashion directly subsequent to membrane insertion of HlyA. Moreover, an immediate consequence would be that the effect of HlyA would follow the P2 receptor expression pattern, which is one of the key elements in the current study.
In proximal tubule cells from rats, HlyA causes distinct, reversible oscillations in the intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) (12). In this study, these HlyA-induced [Ca 2ϩ ] i oscillations were suggested to require G-protein activation and to be essential for a successive release of IL-6 and IL-8 from the epithelium (12). We have previously demonstrated that spontaneous [Ca 2ϩ ] i oscillations in renal epithelia result from constitutive ATP release from the cells followed by P2 receptor activation (13). Based on the results in erythrocytes, we specu-lated that the HlyA-induced [Ca 2ϩ ] i oscillations in renal epithelia could result from ATP release, followed by auto-and paracrine signaling. This notion is supported by several studies, which show that extracellular ATP is a keen trigger of IL-6/IL-8 release from other cell types (14 -18).
Here, we show that HlyA induced [Ca 2ϩ ] i oscillations both in renal epithelial cells in culture (Madin-Darby canine kidney (MDCK) cells) and isolated medullary thick ascending limb (mTAL) from mice require P2 receptor activation. In MDCK cells and mTAL isolated from P2Y 2 ϩ/ϩ mice, the HlyA-induced [Ca 2ϩ ] i oscillations were markedly lowered by scavenging of extracellular ATP by apyrase. In murine mTAL, the HlyA-induced [Ca 2ϩ ] i oscillations were significantly less pronounced in P2Y 2 Ϫ/Ϫ mice compared with control. These data were substantiated in biosensor cells that do not express any type of P2 receptors. We demonstrate that HlyA-induced [Ca 2ϩ ] i oscillations could only be inflicted when these cells had been transfected with hP2Y 2 receptors and thus, the HlyA effect completely depends on P2 receptor expression. Moreover, HlyA readily inflicted IL-6 release from the renal inner stripe of outer medulla (ISOM) from P2Y 2 ϩ/ϩ mice, whereas this response was completely absent in P2Y 2 Ϫ/Ϫ mice. These data pinpoint extracellular ATP as an early signaling molecule in the epithelial response to bacterial virulence factors. Thus, ATP-mediated P2 receptor signaling is likely to be a crucial player in the following inflammatory response.

Experimental Procedures
Cells and Tissue-Type 1 MDCK cells (passage 56 -69) from the American Type Culture Collection (LGC Standards, Boras, Sweden) and 132-1N1 astrocytoma cells were grown in a monolayer on 25-mm diameter coverslips (VWR, Denmark) or in 96-well plates in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (Gibco, Invitrogen, Taastrup, Denmark), 2 mM glutamine and 1 unit ml Ϫ1 of penicillin and 100 g ml Ϫ1 of streptomycin but without riboflavin and phenol red as previously described (19). The hP2Y 2 -transfected 132-1N1 astrocytoma cells were kindly provided by Robert Nicholas (Chapel Hill, NC) and grown in DMEM with 440 g ml Ϫ1 of geneticin. mTALs were isolated by manual dissection from ISOM from P2Y 2 ϩ/ϩ and P2Y 2 Ϫ/Ϫ mice (3-6 weeks) of either sex. The mice were on a B6D2/SV129 background, generously provided by Dr. B. H. Koller, University of North Carolina, NC. Animals were kept and handled according to the Danish animal welfare regulation.
HlyA from E. coli-HlyA was purified from the E. coli strain ARD6 (O6:K13:H1) as described (20) and used in a concentration (EC 50 ) that produces 50% hemolysis of human erythrocytes (1.25% v/v, 60 min at 37°C). This was determined as hemoglobin release measured as A 540 nm of the supernatant in PowerWave Microplate Spectrophotometer, Biotek Instruments, Winooski, VT.
Live Cell Imaging-Cell cultures and mTAL were loaded with 5 M fluo 4-AM and mounted in a semi-open perfusion chamber (modified RC-21BRFS, Warner Instruments, Hamden, CT). The cultured cells were imaged at 1 Hz on an inverted microscope (TE-2000, Nikon) with ϫ60/1.4 NA planApo objective (Nikon), a monochromator (488 nm, Visitech International, Sunderland, UK), and emission collected Ͼ520 nm by an intensified SVGA CCD camera and imaging software (Quanticell 2000/Image Pro, VisiTech). The mTALs were imaged at 1 Hz by ϫ40/1.35 NA planApo objective (Olympus), a monochromator (488 nm, Polychrome V, Till-Phototonics, Munich, Germany), and emission was collected by an iCCDcamera (sensicam qe, PCO, Kelheim, Germany) and software Live Acquisition (Till-Photonics, Munich, Germany). All experiments were carried out at 37°C, pH 7.4, and fluorescence intensity was expressed relative to baseline. Regions of interest were placed over each of the cultured cells or side-by-side along the entire mTAL length and analyzed in Igor Pro (Wavemetrics, Lake Oswego, OR).
Cytokine Measurements-ISOM was dissected and incubated for 6 h at 37°C and 5% CO 2 in DMEM in the presence or absence of HlyA (EC 50 ). Samples of the supernatant were frozen immediately (Ϫ20°C, storage Ͻ10 days). Cytokines were measured by a mouse IL-6 solid phase sandwich ELISA kit from BD Bioscience and a mouse keratinocyte chemoattractant (KC) solid phase sandwich ELISA kit from RayBiotech, Inc. (Norcross, GA) according to the manufacturer's protocols. The amount of IL-6 and KC was determined by absorbance at 450 and 570 nm in a plate reader (PowerWave, Biotek Instruments, Winooski, VT).
ATP Measurements-MDCK cells grown in 96-well plates were incubated for 1 h in HEPES-buffered solution at 37°C. ATP release was determined by luciferin/luciferase (Invitrogen) in the presence or absence of HlyA (EC 50 ) with slight modifications to the instruction. All experiments included a standard row and samples and standards were read in a Mithras LB940 Multimode Reader (Berthold Technologies, Bad Wildbad, Germany).
Statistical Analysis-Data are presented as mean Ϯ S.E. mean. The n value indicates number of mice or cell culture preparations. Data were tested for normal distribution by Kolmogorov-Smirnov test. Statistical significance was determined by Student's t test, Mann-Whitney-Wilcoxon test, or Wilcoxon

HlyA Induces [Ca 2ϩ ] i Oscillations in Cell
Culture-HlyA is known to inflict reversible [Ca 2ϩ ] i oscillations in proximal tubule cells from rat (12). Our data shows that HlyA similarly induces reversible [Ca 2ϩ ] i oscillations in MDCK cells, a cell line isolated from canine distal tubules ( fig. 1). Fig. 1a shows the spontaneous [Ca 2ϩ ] i oscillations previously reported for this cell type (13). Five minutes after addition of HlyA (EC 50 ), there was a significant increase in [Ca 2ϩ ] i oscillatory activity ( Fig. 1, b and d). The HlyA-induced [Ca 2ϩ ] i oscillations were reversible and the cells were still able to respond with an increase in [Ca 2ϩ ] i when exposed to ATP after washout of HlyA (Fig. 1, c and e). Please note that none of the cells showed a sudden drop in fluo 4-fluorescence upon exposure to HlyA, and thus do not lyse during the observation period. Scavenging of extracellular ATP with apyrase (10 units ml Ϫ1 , Fig. 2, a-e) abolished the HlyA-induced [Ca 2ϩ ] i oscillations in MDCK cells, suggesting involvement of purinergic signaling. Because scavenging of ATP markedly reduces the cellular response to HlyA, we anticipate that HlyA may trigger ATP release from MDCK cells. Fig.  2f shows that ATP release from MDCK cells increased immediately after addition of HlyA (EC 50 ). However, it does not become statistically significantly different from control before 12 min incubation because of the mechanically induced ATP release that occurs upon addition of HlyA or vehicle. The release of ATP stayed elevated throughout the observation period.
If HlyA-induced [Ca 2ϩ ] i oscillations are a consequence of ATP release and P2 receptor signaling it should depend on P2 receptor expression. Renal epithelial cells generally express a large variety of different P2 receptors (21,22). The unspecific P2 receptor antagonists (PPADS and suramin) completely abolished the HlyA-induced [Ca 2ϩ ] i oscillations in MDCK cells (data not shown). To completely block all P2 receptor subtypes, the concentration of P2 receptor antagonists has to be substantial and thus, potentially cause unspecific effects. Therefore, we tested the effect of HlyA on ATP-biosensor cells. The 132-1N1 astrocytoma cell line does not express any P2 receptor subtypes and are thus, ideal control cells for potential P2 mediated signaling. In native 132-1N1 cells, addition of HlyA only caused discrete increases in [Ca 2ϩ ] i , which could not be detected over baseline ( Fig. 3, a, b, and g). The lacking effect was not a result of the inability of these cells to create receptor-mediated [Ca 2ϩ ] i increases in general, because the muscarine receptor agonist, carbachol (100 M), caused a sharp rise in [Ca 2ϩ ] i in native 132-1N1 cells (Fig. 3, c and j). Note the complete lack of effect of ATP (100 M), which complies with a total absence of P2 receptors in these cells. The 132-1N1 cells, which are stably transfected with the hP2Y 2 receptor, show spontaneous [Ca 2ϩ ] i oscillations (Fig. 3d), which markedly increase after addition of HlyA (Fig. 3, e and h). The HlyA-induced [Ca 2ϩ ] i oscillations in hP2Y 2 -transfected 132-1N1 cells could be inhibited by the nonselective P2 receptor antagonist suramin (Fig. 3i). These results show that [Ca 2ϩ ] i oscillations induced by HlyA require P2 receptor activation in both MDCK and 132-1N1 cells.
The murine mTAL also expresses P2X receptors on the basolateral side of the tubule (23,24). To test if these receptors are involved in the HlyA-induced [Ca 2ϩ ] i oscillatory activity observed in P2Y 2 Ϫ/Ϫ mTAL, we used the irreversible selective P2X receptor antagonist oxidized ATP. Fig. 5b shows that oxidized ATP (50 M) did not reduce [Ca 2ϩ ] i oscillations in P2Y 2 Ϫ/Ϫ mTAL further (p ϭ 0.64) at a concentration that was shown to completely block ATP-induced transport inhibition in mTAL (24). These data indicate that P2X receptors are not the main receptors involved in HlyA-induced [Ca 2ϩ ] i oscillations in murine mTAL.
Because murine mTAL also expresses P2Y 6 (24) and possibly P2Y 1 receptors (23), we tested the effect of YB-074, which is a P2Y 6 and P2X 1-3 receptor antagonist (25) and MRS2179, an agonist with preference for P2Y 1 receptors. YB-074 (100 M) did not reduce the [Ca 2ϩ ] i oscillatory activity induced by HlyA in P2Y 2 Ϫ/Ϫ mTAL (Fig. 5b, p ϭ 0  tor is unlikely to be involved in the generation of these oscillations. HlyA-induced Release of IL-6 and IL-8 -Uhlén et al. (12) has shown that HlyA-induced [Ca 2ϩ ] i oscillations trigger the release of IL-6 and IL-8 from rat proximal tubule cells. Because ATP is involved in the transcription and release of IL-6 from respiratory epithelia (14), we speculated that HlyA-induced ATP release may stimulate release of cytokines from renal epithelia. After incubation for 6 h, the baseline release of IL-6 from ISOM isolated from P2Y 2 ϩ/ϩ mice was 47.0 Ϯ 9.9 pg ml Ϫ1 (mg of tissue) Ϫ1 and HlyA increased the release to 101.4 Ϯ 23.3 pg ml Ϫ1 (mg of tissue) Ϫ1 (Fig. 6a, p ϭ 0.028). In ISOM from P2Y 2 Ϫ/Ϫ mice, the baseline IL-6 release was 31.47 Ϯ 4.0 pg ml Ϫ1 (mg of tissue) Ϫ1 . Addition of HlyA did not result in a statistically significant increase in IL-6 release in ISOM from P2Y 2 Ϫ/Ϫ mice (51.37 Ϯ 9.3 pg ml Ϫ1 (mg of tissue) Ϫ1 , p ϭ 0.132). Moreover, the HlyA-induced IL-6 release was statistically significantly lower in ISOM from P2Y 2 Ϫ/Ϫ compared with P2Y 2 ϩ/ϩ (p ϭ 0.043). These results show that IL-6 is released from murine ISOM in response to HlyA and that the release requires P2Y 2 receptor activation. Mice do not express IL-8. Instead KC is accepted as the mouse homologue to human IL-8 because it binds to the IL-8 type 2 receptor (26 -28). As shown in Fig. 6b, HlyA releases KC in both P2Y 2 ϩ/ϩ and P2Y 2 Ϫ/Ϫ and there is no difference in the released amount (p ϭ 0.45). Surprisingly, we observed a small but statistically significant difference in baseline KC release between P2Y 2 ϩ/ϩ and P2Y 2 Ϫ/Ϫ (Fig. 6b). These results show that HlyA-induced KC release from murine ISOM in contrast to IL-6 does not involve P2Y 2 receptors.

Discussion
Urinary tract infections are frequently inflicted by the Gramnegative bacterium E. coli (29) and strains that produce the virulence factor HlyA are more prone to inflict severe cases such as pyelonephritis (for review see Refs. 1 and 30). Uhlén et al. (12) found that HlyA from E. coli supernatant induces [Ca 2ϩ ] i oscillations in rat proximal tubule cells followed by IL-6 and IL-8 release. This potentially suggests that HlyA enables the renal epithelial cells to initiate an immunological response before the bacteria have crossed the epithelial barrier.
The current study considers the possibility that HlyA-induced [Ca 2ϩ ] i oscillations are mediated by ATP released from the renal epithelial cells followed by P2 receptor activation. This notion is based on the abundance of ATP-sensitive P2 receptors expressed in virtually all types of renal epithelial cells (for overview see Ref. 21) and that these receptors specifically have been shown to be responsible for spontaneous [Ca 2ϩ ] i oscillations observed in renal epithelia (13). Moreover, there is substantial evidence that extracellular ATP induces IL-6 and IL-8 release from other tissues (14 -18). Interestingly, it was previously shown that IL-6 and IL-8 concentrations are higher in urine from children with febrile urinary tract infection compared with asymptomatic bacteriuria and that bacteria isolated from febrile children were more likely to produce virulence factors including HlyA (31). As mentioned, we have previously established that ATP release and P2X receptor activation is important for the cytolytic effect of HlyA in erythrocytes. Recently, we have shown that ATP may pass directly through the HlyA pore (11). This finding was somewhat surprising because previous studies on the biophysical properties of HlyA, as a pore, suggest it to be cation selective (32,33) and not immediately favor passage of negative charged ATP. Nevertheless, we found that HlyA insertion induced non-lytic ATP release from artificial phospholipid vesicles, which is consistent with HlyA either passing through the pore itself or slip through between the membrane and protein at sites of HlyA insertion (11). That said, it is currently not known how HlyA induces ATP release from renal epithelia. One could speculate that ATP leaves through the HlyA pore, but ATP may equally as well be released through one of the established ATP release pathways in renal  epithelia; vesicular release (34) or through connexin hemichannels (35).
Here, we confirm that HlyA induces [Ca 2ϩ ] i oscillations in renal epithelial cells. Similar to the previously reported effect in rat proximal tubular cells (12), HlyA-induced oscillations were also observed in MDCK cells and isolated murine mTAL. These HlyA-induced [Ca 2ϩ ] i oscillations were previously shown to be inhibitable by IP 3 receptor blockage (12) and thus, require release of Ca 2ϩ from intracellular Ca 2ϩ stores. This finding potentially argues for G␣ q activation in response to HlyA, which certainly comply with ATP-mediated P2Y receptor activation. It must, however, be noted that other groups support that the [Ca 2ϩ ] i events are caused by the HlyA-pore itself without any kind of subsequent signaling (36). To test the concept of HlyA-induced P2Y receptor activation, we used the same types of renal epithelia, in which we previously characterized the ATP and P2 receptor dependence of spontaneous [Ca 2ϩ ] i oscillations: MDCK cells and murine mTAL. We show that HlyA (EC 50 ) accentuates the [Ca 2ϩ ] i oscillations observed in cultured MDCK cells. At this concentration, the effect of HlyA did not cause cell lysis in either of the preparations (measured as immediate drop of fluo 4 fluorescence). Moreover, the cells readily showed a brisk increase in [Ca 2ϩ ] i in response to extracellular ATP after washout of HlyA, a response that serves as a further viability control of the cells. The HlyA-induced [Ca 2ϩ ] i oscillations in MDCK cells were significantly reduced by P2 receptor antagonists (data not shown) and scavenging of extracellular ATP. This finding was substantiated in 132-1N1 cells, which is the only cell line presently known to be completely absent of P2 receptors. In these cells, which hardly show any spontaneous [Ca 2ϩ ] i oscillations (13), HlyA only inflicted an occasional few increases in [Ca 2ϩ ] i , which could not be detected over baseline in the summarized data. However, when 132-1N1 cells were transfected with the hP2Y 2 receptor, the baseline oscillatory activity increased significantly as previously demonstrated (13) and now HlyA readily triggered a vivid [Ca 2ϩ ] i oscillatory activity. These data are consistent with the hypothesis that P2 receptor activation is a player in HlyA-induced [Ca 2ϩ ] i oscillations in general.
To test this in intact renal tissue, we used murine mTAL, where the P2 receptor expression has been functionally charac-terized (23). The P2Y 2 receptor is the predominant receptor expressed on the luminal and basolateral membrane in mTAL (23 Ϫ/Ϫ mTAL. In addition, the mTAL has been found to produce mRNA for the P2Y 4 receptor, which may potentially explain the remaining oscillations (37). In this context, it must be noted that the effect of HlyA seems to follow the P2 receptor expression pattern of the given tissue. In TAL there is a clear dependence on the P2Y 2 receptor, whereas P2X receptors apparently are not involved in the HlyA-induced effects. In erythrocytes, however, P2X receptor antagonists completely block HlyA-induced hemolysis, whereas P2Y receptors do not play a role in HlyA-induced effects in these cells (5).
The present study also confirms that HlyA triggers release of IL-6 and IL-8 from renal tissue. In murine ISOM, where the main components are thick ascending limbs, the baseline IL-6 release was similar in tissue isolated from P2Y 2 Ϫ/Ϫ and P2Y 2 ϩ/ϩ mice. Interestingly, the HlyA-induced IL-6 release was completely absent in ISOM isolated from P2Y 2 Ϫ/Ϫ mice. This finding strongly suggests that HlyA-induced ATP release and subsequent P2Y 2 receptor activation are required for the IL-6 secretion in ISOM. Interestingly, ATP is known to induce release of IL-6 in a P2Y 2 receptor-dependent fashion from human airway epithelia (14) and from the human renal epithelial cell line A498 (38). Because release of IL-6 is important for clearance of the bacterial infection and IL-6-deficient mice have higher mortality in response to E. coli-induced pyelonephritis (39), it would be interesting to investigate whether the same is true for P2Y 2 Ϫ/Ϫ mice. We could also show that HlyA readily inflicted release of KC, the murine IL-8 homologue (26 -28). Surprisingly, the HlyA-induced KC release did not require P2Y 2 receptor activation in murine ISOM. This is not immediately consistent with the clear P2Y receptor dependence of the HlyA-induced IL-8 release in cultured human uroepithelia (40). It must, however, be stressed that KC only is a homolog of IL-8 and its release may not be regulated completely similar to the human IL-8. Moreover, there are still remaining [Ca 2ϩ ] i oscillations in P2Y 2 Ϫ/Ϫ mice and if the KC release is slightly more Ca 2ϩ sensitive than the IL-6 release, it could potentially be unaffected by the lack of P2Y 2 receptors.
Our current findings support an important role of ATP-mediated auto-and paracrine signaling in the first encounter of renal epithelial cells with uropathogenic E. coli. One could speculate that extracellular ATP may act as a potential defense mechanism against bacteria. Extracellular ATP is a potent inhibitor of renal epithelial transport (26,41,42) and could thereby increase the washout of intratubular bacteria. Moreover, the current data are consistent with ATP as the mediator of HlyA-induced IL-6 release and thus, potentially essential to immune system-mediated eradication of the infection. On this note, it is interesting to speculate that the P2Y 2 receptor is required for an adequate immune response in renal epithelia.
The critical point is: why is HlyA a virulence factor in terms of ascending urinary tract infections? Generally, hemolysins have been suggested to provide the iron important to sustain bacterial growth, because chelation of iron significantly hampers the growth of E. coli (43). Iron is recognized as an essential nutrient for many of the key steps in the development of any pathogen in its host (44). Hemolysis has been speculated to be an important physiological source of iron for E. coli, because the iron-transport system of the bacterium use heme and hemoglobin and not transferrin and lactoferrin (45). This certainly is one potential way for HlyA to enhance bacterial growth in vivo and thus, be a virulence factor. It does, however, not necessarily explain why this increases the chance of the bacterium to advance into the kidney. Here interaction with the epithelial barrier is likely to take precedence. In this context, it is interesting that both E. coli (46) and aerolysin from Aeromonas hydrophila have been shown to modify tight junctions (47) and several cytokines including IL-6 have indeed been shown to modulate the expression pattern of occludins and claudins as well as modulate tight junctional function in certain epithelia/ endothelia (48,49). Thus, it is likely that HlyA and the following ATP-dependent cytokine release may interfere with the barrier function of the epithelia and thus, its ability to keep the bacteria from invasion.
It is still not known what the functional phenotype of the P2Y 2 receptor is in respect to ascending urinary tract infections. Our presented data points to ATP release and signaling as one of the very early events when an epithelial barrier is exposed to HlyA. Therefore, these data support that interference with purinergic signaling may significantly affect the pathogenesis of urinary tract infections.