Human renal fibroblasts are strong immunomobilizers during a urinary tract infection mediated by uropathogenic Escherichia coli

To prevent the onset of urosepsis and reduce mortality, a better understanding of how uropathogenic Escherichia coli (UPEC) manages to infiltrate the bloodstream through the kidneys is needed. The present study elucidates if human renal interstitial fibroblasts are part of the immune response limiting a UPEC infection, or if UPEC has the ability to modulate the fibroblasts for their own gain. Microarray results showed that upregulated genes were associated with an activated immune response. We also found that chemokines released from renal fibroblasts upon a UPEC infection could be mediated by LPS and triacylated lipoproteins activating the TLR2/1, TLR4, MAPK, NF-κB and PKC signaling pathways. Furthermore, UPEC was also shown to be able to adhere and invade renal fibroblasts, mediated by the P-fimbriae. Furthermore, it was found that renal fibroblasts were more immunoreactive than renal epithelial cells upon a UPEC infection. However, both renal fibroblasts and epithelial cells were equally efficient at inducing neutrophil migration. In conclusion, we have found that human renal fibroblasts can sense UPEC and mobilize a host response with neutrophil migration. This suggests that renal fibroblasts are not only structural cells that produce and regulate the extracellular matrix, but also highly immunoreactive cells.

modulate the fibroblast responses to persist and spread to the bloodstream? Hence the need of understanding the interaction between UPEC and renal fibroblasts. We and others have shown that UPEC has the ability to modulate the immune response in the urinary tract via various virulence factors such as type-1 fimbriae, P-fimbriae, α-hemolysin, IrmA and TcpC to colonize the urinary tract [11][12][13][14] . However, which virulence factors UPEC utilizes in the interaction with renal fibroblasts is unknown. Our aim was to elucidate if human renal fibroblasts are a part of the immune response limiting the UPEC infection, or if UPEC has the ability to modulate the fibroblasts for its persistence and spreading.

Results
Gene expression alterations in UPEC infected renal fibroblasts. A microarray analysis was performed on total RNA isolated from primary human renal fibroblasts infected with the UPEC strain CFT073. In total 1196 gene entities were upregulated and 509 gene entities (Supplementary Table S1) were downregulated (corrected p < 0.05) with at least a ≥2 fold change compared to unstimulated renal fibroblasts after 6 hours. The thirty highest upregulated and downregulated gene entities are presented in Tables 1 and 2, respectively. In order to validate the microarray results, real time qPCR was conducted on four significantly upregulated (IL-1β, NOD2, CXCL10 and CXCL9) and four significantly downregulated (PTCH1, TET1, PLCB2 and CPEB1) genes compared to unstimulated renal fibroblasts. In agreement with the microarray analysis, a significant upregulation of IL-1β, CXCL10 and CXCL9 was observed. We also found that PTCH1, TET1, PLCB2 and CPEB1 were significantly downregulated compared to unstimulated renal fibroblasts (Table 3).
Gene ontology and KEGG pathway analysis. Gene ontology analysis was conducted on significantly altered gene entities compared to unstimulated renal fibroblasts. In total, 674 upregulated (Supplementary  Table S2) and 42 downregulated (Supplementary Table S3) gene ontologies were enriched. The top ten upregulated gene ontologies were defense response, immune response, immune system process, response to cytokine, innate immune response, response to other organism, cytokine-mediated signaling pathway, defense response to virus, response to biotic stimulus and response to virus. The top ten downregulated gene ontologies were transcription DNA-templated, regulation of transcription DNA-templated, RNA biosynthetic process, regulation of cellular macromolecule biosynthetic process, regulation of macromolecule biosynthetic process, nucleobase-containing compound biosynthetic process, heterocycle biosynthetic process, regulation of cellular biosynthetic process, aromatic compound biosynthetic process and regulation of biosynthetic process (Fig. 1A). Furthermore, we also found that the upregulated gene entities were significantly (corrected p < 0.05) enriched in 53 KEGG pathways (Supplementary Table S4) and the downregulated gene entities in 1 KEGG pathway (Supplementary Table S5). The top ten upregulated KEGG pathways were TNF signaling pathway, influenza A, NF-kappa B signaling pathway, Herpes simplex infection, cytokine-cytokine receptor interaction, measles, apoptosis, NOD-like receptor signaling pathway, Toll-like receptor signaling pathway and cytosolic DNA-sensing pathway. The downregulated KEGG pathway was transcriptional misregulation in cancer (Fig. 1B).

IL-8 release from renal fibroblasts.
As the microarray analysis showed that several chemokines were highly induced by CFT073, experiments were performed to examine the involvement of various UPEC virulence factors on IL-8 release using P-fimbriae (pap), type-1 fimbriae (fimH), α-hemolysin (hlyA), TIR homologous protein (TcpC), flagellin (fliC) CFT073 deletion mutants. These experiments showed that neither the P-fimbriae, type-1 fimbriae, α-hemolysin, TcpC nor flagellin were responsible for the CFT073 induced IL-8 release from renal fibroblasts after 6 hours. No significant IL-8 release was observed after 3 hours of bacterial stimulation (data not shown). In addition, no difference in IL-8 release was observed between CFT073, UTI89 or MG1655 ( Fig. 2A). Using the same bacterial setup, we also found that neither the P-fimbriae, type-1 fimbriae, α-hemolysin, TcpC nor flagellin were responsible for the CFT073 induced CXCL10 release from renal fibroblasts after 6 hours. However, we did show that MG1655, but not UTI89 induced a significantly higher CXCL10 release compared to CFT073 ( Supplementary Fig. S1). Microarray analysis showed that neither the P-fimbriae, type-1 fimbriae nor α-hemolysin were responsible for any of the significantly altered genes induced by CFT073 compared to unstimulated fibroblasts (data not shown). We continued with evaluating which signaling pathways were involved in the CFT073-induced IL-8 release from renal fibroblasts. We found that inhibition of p38, ERK1/2, NF-κB and PKC resulted in significantly lower IL-8 release from renal fibroblasts compared to cells infected with CFT073 in the presence of the vehicle DMSO after 6 hours (Fig. 2B). In addition, we also showed that LPS and Pam3CSK4, but not ATP, Adenosine or flagellin could induce an increased IL-8 release from renal fibroblasts compared to unstimulated cells after 6 hours (Fig. 2C). Taken together, these results suggest that UPEC-induce chemokine release from renal fibroblasts could be mediated by several different signaling pathways.
CFT073 adheres to and invades renal fibroblasts. Further, evaluation of whether CFT073 could adhere and invade renal fibroblasts was performed. We found that CFT073 had the ability to adhere to renal fibroblasts (Fig. 3A,B) and that this adhesion was mediated by the P-fimbriae and not type-1 fimbriae (Fig. 3A,B). In addition, we also showed that the UPEC strain CFT073 was able to invade renal fibroblasts and that this invasion was partially mediated by the P-fimbriae and not type-1 fimbriae (Fig. 4A-E). We found that the maximum adhesion and invasion occurred after 3 hours. Together, our findings show that UPEC bacteria can adhere and invade renal fibroblast and this may protect the bacteria from the host immune response.
Renal fibroblasts are more immunoreactive to CFT073 then renal epithelial cells. We continued with comparing the immunoreactivity of renal fibroblasts and renal epithelial cells to a CFT073 infection with an Olink inflammation multiplex immunoassay (Supplementary Table S6). We found that 21 of 92 inflammatory-related proteins were significantly upregulated in renal fibroblasts after a CFT073 infection compared to unstimulated renal fibroblasts (Fig. 5). However, only one protein, TGF-alpha, was found to be significantly upregulated in renal epithelial cells after a CFT073 infection compared to unstimulated renal epithelial cells (Fig. 5). These findings show that renal fibroblasts are more immunoreactive than renal epithelial cells against a UPEC infection.

Neutrophil migration is mediated by both renal fibroblasts and renal epithelial cells.
We proceeded with investigating the ability of renal fibroblasts and renal epithelial cells to induce neutrophil migration in response to CFT073. We found that both renal epithelial cells and renal fibroblasts induced a significantly increased neutrophil migration compared to unstimulated cells after 3 hours (Fig. 6). However, we did not find any differences in neutrophil migration between the renal epithelial cells and renal fibroblasts (Fig. 6). These findings suggest that renal fibroblasts have the ability to recruit neutrophils in response to a UPEC infection.

Discussion
To gain a better understanding on how UPEC interacts and manipulates the immune response in the kidneys, is important for preventing the bacteria from reaching the bloodstream. The vast majority of the research has so far been focused on studying the interaction between UPEC, renal epithelial cells and infiltrating leukocytes [15][16][17][18] . However, what role interstitial renal fibroblasts play in the progression of the infection has not been studied to our knowledge. Our aim for this study was to elucidate if human renal fibroblasts are a part of the immune response limiting the infection, or if UPEC has the ability to modulate the fibroblasts for its persistence and spreading. We started by evaluating the global gene expression of primary human renal fibroblasts infected with the UPEC wild type strain CFT073. We found that CFT073 induced significant transcriptional alterations of a large number of genes in renal fibroblasts; 1196 upregulated and 509 downregulated gene entities. These findings were validated with real time RT-PCR for eight of the significantly altered genes (IL-1β, NOD2, CXCL10, PTCH1, TET1, PLCB2, CPEB1 and CXCL9). To the best of our knowledge, this is the first study that has investigated changes in the global gene expression profile of primary human renal fibroblasts infected with UPEC. Importantly, these data help us to understand the specific expression profile of renal fibroblasts evoked by UPEC and how this expression may differ from the profile of renal epithelial cells 19 and infiltrated leukocytes 20,21 .
Gene ontology and KEGG pathway analysis showed that the upregulated genes were associated with an activated immune response. Key immunological pathways such as TNF signaling, NF-kappa B signaling, apoptosis, NOD-like receptor-and Toll-like receptor signaling pathway etc. were all enriched in the upregulated group. All mentioned pathways are known to be associated with pathogen recognition and the mobilization of the innate and adaptive immune response [22][23][24] . These findings are supported by the observed upregulation of several cytokines (e.g. IL-1, -6, -7, -15, -17C, -33) and chemokines (e.g. CXCL1, -2, -3, -4, -6, -8, -9, -10, -11 and CCL1, -2, -3, -5, -7, -8, - 11, -13, -19, 20) in response to UPEC. Furthermore, we also found that the enriched gene ontologies and pathways for the downregulated genes were associated with transcription. A previous study has shown that E. coli is able to control host gene expression through modulation of RNA polymerase II 19 . However, further evaluation is needed to clarify if our observations are pointing towards the same conclusion. Taken together, these results suggests that renal fibroblasts are not only structural cells that produce and regulate the extracellular matrix, but also highly immunoreactive cells that can recognize UPEC and initiate an immune response. It is well known that UPEC utilizes several different virulence factors like the P-fimbriae, type-1 fimbriae, α-hemolysin, TcpC and flagellin to colonize and modulate the immune response in the urinary tract. These virulence factors are known to be able to regulate the release of pro-inflammatory cytokines and chemokines, which we found to be highly induced by CFT073 [25][26][27] . However, we found that neither the P-fimbriae, type-1 fimbriae, α-hemolysin, TcpC nor flagellin were responsible for the CFT073 induced IL-8 or CXCL10 release from renal fibroblasts. Additionally, this was validated by microarray, which showed that neither the P-fimbriae, type-1 fimbriae nor α-hemolysin were responsible for any of the significantly altered genes induced by CFT073 (data not shown). This suggest that neither the P-fimbriae, type-1 fimbriae, α-hemolysin, TcpC nor flagellin are important for the mobilized immune response induced by human renal fibroblasts. Furthermore, we also showed that the UPEC strain UTI89 and the non-pathogenic MG1655 strain could induce IL-8 and CXCL10 release from renal fibroblasts. Suggesting that this induction is not CFT073 or UPEC specific. To further explore the mechanisms behind the IL-8 release, we evaluated which signaling pathways CFT073 utilizes for the cytokine induction. We found that p38, ERK1/2, NF-κB and PKC were all involved in regulating the CFT073-induced IL-8 release from renal fibroblasts. The MAPK-pathway associated p38 and ERK1/2, NF-κB and PKC have all been shown to be important for the activation of the immune response in the kidneys 14,28,29 . Furthermore, we also showed that the TLR2/1 agonist Pam3CSK4 and the TLR4 agonist LPS induced an increased IL-8 release from renal fibroblasts, whereas the TLR5 agonist flagellin, the purinoceptor 1 agonist adenosine or the purinoceptor 2 agonist ATP did not. TLR2/1, TLR4, TLR5, purinoceptor 1 and purinoceptor 2 [30][31][32] have all been shown to be important in the host response against UPEC. Our results suggest that UPEC LPS and triacylated lipoproteins (TLR2/1 agonists)  may be the pathogen associated molecular patterns (PAMPs) activating the renal fibroblasts. Taken together, these results suggest that UPEC-induced chemokine release from renal fibroblasts could be mediated by LPS and triacylated lipoproteins activating the TLR2/1, TLR4, MAPK, NF-κB and PKC signaling pathways. Adhesion to epithelial cells is crucial for bacterial colonization of the urinary tract. UPEC is known to use the type-1 fimbriae for colonizing the bladder and the P-fimbriae for colonizing the kidneys 25 . In agreement with previous findings regarding the P-fimbriae and renal epithelial cells 33,34 , we found that UPEC utilizes the P-fimbriae, and not the type-1 fimbriae, for the adhesion/invasion of renal fibroblasts. The ability of UPEC to adhere and invade renal fibroblasts can be an evasion strategy utilized by UPEC to protect itself from infiltrated neutrophils and antibiotic treatment. Previous studies have associated intracellular UPEC with host evasion, antibiotic failure and recurrent UTI [35][36][37][38] . After the resolution of the immune response, UPEC may emerge out from the intracellular niche and continue the infection. This ability may be a piece of a puzzle explaining how UPEC, after breaching the renal epithelial layer, can persist in the renal interstitium and eventually reach the bloodstream.
By comparing the immune response of renal fibroblasts and renal epithelial cells upon a UPEC infection, a better understanding regarding what role the respective cell type has in clearing the infection might be reached. We found that renal fibroblasts are more immunoreactive than renal epithelial cells upon a UPEC infection, by using the Olink inflammatory multiplex panel. Out of 92 analyzed inflammatory-related proteins, 21 were significantly upregulated in fibroblasts and 1 was significantly upregulated in epithelial cells during a CFT073 infection     [39][40][41] . Our findings suggests that renal fibroblasts may have the ability to attract several different leukocytes during a UPEC infection. It is well known that neutrophils are the main effector cells of the immune response contributing to bacterial clearance during a UTI 42 . However, migrated monocytes and differentiated macrophages have also been shown to influence the inflammatory response during a UTI 43,44 . NK-cells 45 , DCs 46 and T-cells 47,48 have all lately been implicated to be part of the host response during a UTI, but their specific contribution to the host response remains to be determined. With these results in mind, we proceeded with evaluating the ability of renal fibroblasts and renal epithelial cells to induce neutrophil migration in response to CFT073. We found that both fibroblasts and epithelial cells induced a significantly increased neutrophil migration in response to CFT073. However, no difference between the cell types were found. There may be additional neutrophil chemotactic factors released from renal epithelial cells in response to CFT073, e.g. CXCL2, CXCL3, CXCL5 and leukotriene B4 41 that could explain the lack of difference in migration. Taken together, our findings show that renal fibroblasts are strong immunoreactive cells that have the same capacity as renal epithelial cells to recruit neutrophils in response to a UPEC infection. This study provides novel insight into the role renal fibroblasts have during a UPEC infection. The ability to sense UPEC and mobilize a host response with neutrophil migration suggests that renal fibroblasts are not only structural cells that produce and regulate the extracellular matrix, but also highly immunoreactive cells. The ability of UPEC to invade renal fibroblasts may be one of the strategies utilized by the bacteria to subvert the host   responses on its path to the bloodstream. Understanding how UPEC modulate our immune system in the kidneys may help us develop novel treatment strategies to prevent the bacteria from reaching the bloodstream.

Methods
Cell and bacterial culture. Primary human renal fibroblasts (Pelobiotech GmbH, Planegg, Germany) were derived from a healthy woman and the human renal epithelial cell line A498 (HTB-44, ATCC) was derived from a kidney carcinoma. Both cell lines were cultured in Dulbecco's modified eagle medium (DMEM, Lonza, Basel, Switzerland) containing 10% foetal bovine serum (FBS), 2 mM L-glutamine, 1 mM non-essential amino  acids, 50 U/mL penicillin and 50 ml/mL streptomycin (all from Invitrogen Ltd., Paisley, UK) at 37 °C in a 5% CO 2 atmosphere. The cells were serum starved overnight in DMEM containing 2 mM L-glutamine, 1 mM non-essential amino acids and 50 µg/ml Gentamicin (Sigma-Aldrich, St. Louis, MO, USA). During experiments, the medium was replaced with DMEM containing 1% FBS, 1 mM non-essential amino acids and 2 mM L-glutamine. The UPEC strain CFT073 is isolated from a patient with pyelonephritis and urosepsis 49 . The UPEC strain UTI89 is isolated from a patient with an acute bladder infection 50  and green were scored as extracellular adherent bacteria. Intracellular bacteria were only stained green. Images were obtained using an Olympus BX53 fluorescence microscope equipped with an Olympus DP74 camera.
OLINK multiplex protein assay. Renal fibroblasts and renal epithelial cells (A498) were infected with CFT073 at MOI 10 and incubated at 37 °C with 5% CO 2 for 6 hours. A panel of 92 inflammatory proteins were analysed simultaneously in the supernatants using the proximity extension assay on the Proseek Multiplex inflammation panel (Olink Bioscience, Uppsala, Sweden). The protein data is reported as normalized protein expression units where a high value corresponds to high protein concentration.
Neutrophil isolation and migration assay. Human neutrophils were isolated from healthy blood donors by density gradient centrifugation of polymorphprep and lymphoprep reagents (AXIS-SHIELD PoC AS, Oslo, Norway) according to the manufacturer's. An ethical approval has been granted by the regional ethics review board in Uppsala, Sweden (Dnr 2015/437), to isolate blood from healthy individuals after informed consent. Blood from healthy donors were collected according to the ethical guidelines of both the Declaration of Helsinki and the Swedish national board of health and welfare. The viability of the neutrophils was >90% as determined by the trypan blue exclusion test. Renal fibroblasts and renal epithelial cells (A498) were infected with CFT073 at MOI 10 and incubated at 37 °C with 5% CO 2 for 6 h. Supernatants were collected and centrifuged for 5 minutes at 5000 × g to get rid of the bacteria. The bacteria free supernatants were added to the bottom well of a 3 μm pore size transwell system and 1·10 6 neutrophils were added to the top well. Neutrophils were collected from the bottom well after 3 h of incubation and counted in a Bürker chamber.
Statistical analysis and microarray data processing. The differences between groups assay were analysed with one-way ANOVA followed by Bonferroni test. Differences were considered statistically significant when p < 0.05. Data are presented as mean ± SEM, n = number of independent experiments. Microarray analysis was performed using Gene Spring GX version 12.0 (Agilent Technologies) after per chip and gene 75 th percentile shift normalization of samples. Different gene expression between groups was analysed with one-way ANOVA.
Significantly expressed genes (p < 0.05) was obtained by Tukey HSD post-hoc test followed by Bonferroni multiple testing correction with a fold change set at ≥2. GO enrichment and KEGG pathway analysis were conducted with STRING (version 10.5) and the significance was set at a p-value < 0.05.