Bladder inflammatory transcriptome in response to tachykinins: Neurokinin 1 receptor-dependent genes and transcription regulatory elements

All animal experimentation described here was performed in conformity with the "Guiding Principles for Research Involving Animals and Human Beings (OUHSC Animal Care & Use Committee protocol #00-109 and #00-108). Groups of ten to twelve-week old female mice were used in these experiments. NK₁R^-/- and wild type (WT, C57BL6) littermate control mice were generated by Dr. Norma P. Gerard. The colonies at OUHSC were genotyped as described previously [35].

All mice in this study were sensitized with 1 μg DNP₄-human serum albumin (HSA) in 1 mg alum on days 0, 7, 14, and 21, intraperitoneally (i.p.). In normal mice, this protocol induces sustained levels of IgE antibodies up to 56 days post-sensitization [36]. One week after the last sensitization, cystitis was induced. Briefly, sensitized WT and NK₁R^-/- mice were anesthetized (ketamine 40 mg/kg and xylazine 2.5 mg/kg, i.p.), then transurethrally catheterized (24 Ga.; 3/4 in; Angiocath, Becton Dickson, Sandy, Utah), and the urine was drained by applying slight digital pressure to the lower abdomen. The urinary bladders were instilled with 200 μl of pyrogen-free saline or DNP₄-OVA (1 μg/ml). One, four, and twenty-four hours after instillation, mice were sacrificed with pentobarbital (100 mg/kg, i.p.) and bladders were removed rapidly.

Previous results from our laboratory demonstrated a mandatory role of NK₁R on antigen-induced cystitis [19, 37]. In the present work, we also investigated whether NK₁Rs are important for both SP- and LPS-induced cystitis. For this purpose, an additional group of NK₁R^-/- and wild type (WT, C57BL6) were anesthetized as described above and challenged intravesically with 200 μl of pyrogen-free saline, SP (10 μM), or Escherichia coli LPS strain 055:B5 (Sigma, St. Louis, MO; 100 μg/ml). Twenty-four hours after instillation, mice were euthanized with pentobarbital (200 mg/kg, i.p.), and the bladders were removed rapidly for evaluation of inflammatory cell infiltrates and the presence of interstitial edema. A semi-quantitative score using defined criteria of inflammation severity was used to evaluate cystitis [37]. A cross-section of bladder wall was fixed in formalin, dehydrated in graded alcohol and xylene, embedded in paraffin, and cut serially into four 5-μm sections (8 μm apart) to be stained with hematoxylinand eosin (H&E) and Giemsa. H&E stained sections were visualized under microscope (Eclipse E600, Nikon, Lewisville, TX). All tissues were photographed at room temperature by a digital camera (DXM1200; Nikon). Exposure times were held constant when acquiring images from different groups. Images were analyzed with Image-Pro Analyzer^® (Media Cybernetics Inc.; Silver Spring, MD 20910). The severity of lesions in the urinary bladder was graded as follows: 1+, mild (infiltration of 0–10 neutrophils/cross-section in the lamina propria, and little or no interstitial edema); 2+, moderate(infiltration of 10–20 neutrophils/cross-section in the lamina propria, and moderate interstitial edema); 3+, severe (diffuse infiltration of >20 neutrophils/cross-section in the lamina propria and severe interstitial edema) [19, 37, 38]. Identification of mast cells was performed in Giemsa-stained sections [37].

We used a novel approach [34] to fully annotate and represent NK₁R-dependent genes by using the Ingenuity Pathways Analysis tool [43]. Using Ingenuity knowledge base network, we identified specific and canonical pathways downstream of NK₁R activation.

We employed a bioinformatics approach to hypothesize functionally relevant transcriptional regulatory elements (TREs) of NK₁R-dependent and -independent genes. The regulatory network was determined by a combination of micro array-selected transcripts and PAINT 3.3 [44], available online [45], to query the transcription factor database (TRANSFAC) [46]. PAINT 3.3 was employed to examine 2000 base pairs of regulatory regions upstream of the transcriptional start site of each differentially expressed gene detected with the microarray. PAINT is a suite of bioinformatics and computational tools that integrate functional genomics information, as is the case of our microarray-based gene expression data, with genomic sequence and TRE data to derive hypotheses on the TREs relevant to the biological function under study.

Genbank accession numbers were used as the gene identifiers in PAINT test files. Over-representation of TREs in the matrix was calculated at levels of 0 < p <= 0.01 and 0.01 < p < 0.05 when compared to the reference (TREs regulating all genes in the original array). Employing the microarray accounts for any 'bias' present in the genes on the microarray relative to entire genome, nd guards from incorrectly concluding that certain TREs are relevant to the current experiment. The TRE hypotheses were generated from statistical enrichment analysis and were defined as those TREs that are significantly enriched such as NK₁R-dependent and -independent genes, over random occurrence in the gene groups [44].

Anesthetized C57BL6 female mice were instilled with 200 μl of saline or SP (10 μM) and bladders were removed 2, 6, and 24 hours after instillation. In one additional group (zero hours), the urinary bladders were removed without instillation. Urinary bladders were placed in cold phosphate buffered saline (0°C), containing peptidase inhibitors (aprotinin, pepstatin, leupeptin at 0.01 mg/ml,) and the mucosa was dissected away from the muscle, as described previously [33]. Nuclear proteins were extracted and used for electrophoretic mobility shift assay for Nkx-2.5 and NF-kappaB.

The NF-kappaB probe was constructed by annealing complementary synthetic oligonucleotides (5' GAT CAT GGG GAA TCC CCA 3'). Annealed probes were end-labeled with [α-³²P] ATP (3000 Ci/mmole; GE Healthcare) and T4 polynucleotide kinase (New England Biolabs), and then purified using a G-50 column (GE Healthcare). Nuclear extracts (10 μg) were incubated with 1 ng of [³²P] NF-kappaB double-stranded probe in 20 μl 20 mM HEPES, 70 mM KCl, 2 mM DTT, 0.01% NP-40, 4% Ficoll, 1 mg/ml BSA, and 1.4 μg poly d(I-C). For competition reactions, a 50-fold excess of unlabelled NF-kappaB probe was added to the reaction mixture, and it was incubated at room temperature for 5 minutes before the addition of the [³²P] NF-kappaB double-stranded probe. Reaction mixtures were incubated for 20 minutes at room temperature. DNA-protein complexes were resolved on a non-denaturing 6% polyacrylamide gel at 200 V for 3 hours in 0.5 TBE (45 mM Tris-borate and 1 mM EDTA). Gels were vacuum-dried and visualized on Kodak Biomax MS Film and quantified using ImageJ Software (NIH) and statistical differences were determined using GraphPad Prism Software (GraphPad Software, San Diego, CA).

Electrophoretic mobility shift assays for Nkx-2.5 were performed using a non-radioactive Gel Shift Kit (Panomics, Redwood City, CA) according to the experimental procedures provided by the manufacturer. Briefly, nuclear extracts (5 μg) were incubated with biotinylated Nkx-2.5 probe in Binding Buffer with poly d(I-C). For competition reactions, an excess of unlabelled Nkx-2.5 probe was added to the reaction mixture and it was incubated at room temperature for 5 minutes before the addition of the biotinylated Nkx-2.5 probe. Reaction mixtures were incubated for 30 minutes at 20°C. DNA-protein complexes were resolved on a non-denaturing 6% polyacrylamide gel at 200 V for 2 hours in 0.5× TBE (45 mM Tris-borate and 1 mM EDTA). The gel was then transferred onto a 0.45 μm Nytran SuperCharge membrane (Schleicher & Schuell, Keene, NH) at 300 mA for 30 minutes in 0.5× TBE. The blot was dried for 60 minutes at 80°C and then UV cross-linked for 3 minutes. After blocking, the blot was incubated with Streptavidin-HRP and developed using the substrate solutions included in the kit. Blots were visualized on Pierce CL-XPosure film (Pierce, Rockford, IL). Bands were quantified using ImageJ Software (NIH) and statistical differences were determined using GraphPad Prism Software (GraphPad Software, San Diego, CA). The sequence for the NKx-2.5 motif-containing probe, not provided by the manufacturer, is 5' AAA CAA GTC ATA ATA GGA AGC A 3'.

Previous results from our laboratory demonstrated a mandatory role of NK₁R on antigen-induced cystitis [19, 37]. In the present work we also investigated whether NK₁Rs are important for both SP- and LPS-induced cystitis. In contrast to mice treated with saline (Figure 1A), instillation of LPS into the bladder of wild type mice leads to inflammation characterized by edema (Figure 1B; black line delimits the area of sub-epithelial edema) and infiltration of inflammatory cells (Figure 1D; black arrows indicate PMNs). In contrast, the urinary bladders of NK₁R-knockout mice failed to mount an inflammatory response to LPS (Figures 1C and 1E) despite the capacity of LPS to induce urothelial cell injury, as indicated by intra-cytoplasmatic vacuolization [47] (Figure 1E; black circles). Similar results were obtained with WT mice that mounted an inflammatory reaction to intravesicall instillation of SP as seem in H&E (Figure 2A and 2C) and Giemsa stained (Figure 2E) sections. NK₁R-knockout failed to mount an inflammatory response to SP (Figures 2B, 2D, and 2F) with a reduced number of migrating inflammatory cells but presenting visible resident mast cells (Figure 2F; black line). Figure 3 presents the quantification of the results here described. In conclusion, these results extend our previous observation of a mandatory role for NK₁R in antigen-induced cystitis to other pro-inflammatory mediators.

Two hundred and nine genes fulfilled the established criteria (see material and methods) and were considered NK₁R-dependent. In contrast, 236 genes were found to be up-regulated secondary to antigen challenge in tissues isolated from WT and NK₁R^-/- mice and, therefore, were considered to be NK₁R-independent.

PAINT 3.3 was employed to examine 2000 base pairs of regulatory regions upstream of the transcriptional start site of each differentially expressed gene from the microarray expression data. Genbank accession numbers were used as gene identifiers in PAINT input files. For the NK₁R-dependent genes, out of a total of 209 genes from the expression analysis, only 153 genes had corresponding upstream sequence information owing to the incomplete nature of genomic annotation. A total of 87 TREs were identified on these sequences using MATCH^® tool in the TRANSFAC Professional database. Similarly, for the NK₁R-independent genes, only 188 promoters were retrieved for the 236 genes from expression analysis. A total of 88 TREs were identified on these sequences using MATCH^®. These TREs were examined in the enrichment analysis to derive regulatory network hypotheses. The resulting candidate interaction network can be visualized as an interaction matrix where the individual elements of the matrix are color-coded based on the p-values for statistical enrichment. A p-value threshold of 0.07 was used to filter the enrichment analysis results to derive the regulatory network hypotheses. These results are shown in Figure 4 for NK₁R-dependent genes and in Figure 5 for NK₁R-independent genes.

Figure 6 depicts the hypothesized regulatory network corresponding to the NK₁R-dependent genes whose TRE were found to be over-represented (0.01 < p < 0.05) in this set when compared to a reference (all genes in the array). cRel was the predominant TRE, driving 22% of the NK₁R-dependent genes. The order of predominance for the different TREs was: cRel, v-Myb, CRE-BP1/c-Jun, USF, AP-1_Q2, Pax-6, AP-1_C, NF-kappaB_Q6, Efr-1, Egr-3, and AREB6 (Figure 6 and Table 1 [additional file 1]).

In order to make sense of the vast information generated by cDNA array expression, we used the recently developed Ingenuity Pathways Knowledge Base [34] to design pathways of NK1R-dependent genes and to query canonic pathways regarding the relative importance of each of the transcripts (Figure 7). As a result, genes were localized to different compartments depending on the predominant expression of the protein they encoded (Figure 7). In this way, four compartments are depicted: extracellular space, plasma membrane, cytoplasm, and nucleus. According to their primary function, ingenuity grouped NK₁-dependent genes into 4 different networks: cell morphology, cell cycle, inflammation, and cell death. However, a strong degree of overlapping between cell morphology, cell cycle, cell death, and cancer was observed. In contrast, genes involved in inflammation were also listed to be involved in embryonic development. The NK1R-dependent genes were significantly correlated with the following canonical signaling pathways: p38 MAPK, NF-kappaB, PPAR, IL-6, death receptor, apoptosis, and SAPK/JNK (Table 2, [additional file 2]).

Overall NK₁R-dependent genes were classified following the biological processes with which they are involved (GO anthology): apoptosis (GZMA, TNFRSF1B, TNFRSF1A, TRAF3, NOS2A, and BID); cell adhesion/hyaluronic acid binding (CD44 and AGC1); cell cycle (CCND2 and CCNG1); cell-cell signaling (FGF11 and GJA7); cytokinesis (CDC42 and KIF1B); development (FMR2); extracellular transporters & carriers (APOE); G-protein coupled receptors (GNA13 and PTGIR); growth factors (MXD1); heat shock proteins (PRNP, HSPH1, and HSPD1); immune response (BST-1, CTSW, and IL1R1); interferons (INFGR1) intracellular kinases (WBP6); intracellular transducers (MAP3K7); kinase activators & inhibitors (YWHAH); membrane channels (KCNAB1, KCNJ12, KCNQ1, and SLC30A4); nucleotide metabolism (PCSK1); oncogenes & tumor suppressors (BRCA1, MAP3K8, RET, VIL2, FLI1, MET, NF2, and VEGFR1); receptor mediated endocytosis (DAB2); receptor tyrosine kinase (EPHA2); regulation of transcription (NEUROD6); serotonin biosynthesis (YTPH1); symporters & antiporters (SLC16A1 and SLC1A1); transcription activators & repressors (FOXA1, HSF1, IER2, and NR1H2); and synaptic transmission (GRID1).

The whole set of genes that were up-regulated at least 3-fold during inflammation including NK₁R-dependent and NK₁R-independent genes were analyzed in PAINT. A p-value threshold of 0.05 was used to derive statistically enriched regulatory elements in these two gene groups (Figure 8). The enrichment in this case was obtained by using the interaction matrix for the combined gene list as a reference. This enabled us to contrast the two gene groups relative to each other and characterize those regulatory elements that are specific to one group or the other. We observed differential enrichment of several regulatory elements including over-expression of AP1 specific in the NK₁R-dependent genes, as we previously suggested [41] (Figure 8). However, the most striking finding of this analysis was the revelation that Nkx-2.5, a murine homeo box gene, is a unique discriminator of NK₁R-dependent genes. Nkx-2.5 matrix _01 was found over represented in the set of NK₁R-independent transcripts (colored in red) and under-represented or suppressed in the set of NK1R-dependent TREs (colored in cyan). In contrast, Nkx-2.5 matrix _02 had the completely opposite behavior since it was over-represented in NK1R-dependent and under-represented in the NK₁R-independent set. Table 3 [additional file 3] lists all genes under the control of both Nkx-2.5 matrix _01 and matrix _02. With the exception of A10 (L21027) and EGF (J00380) that were associated with both matrixes, a unique set of genes is correlated with Nkx-2.5 matrix _01 and Nkx-2.5 matrix _02.

To determine whether activation of NK₁R increases the transcriptional activities of NF-kappaB and Nkx-2.5, an EMSA was performed. Results show that instillation of SP into the mouse bladder increased the amount of shifted Nkx-2.5 and NF-kappaB probes that peaked at 24 hours post stimulation (Figure 9 A-D). Because of the high level of expression of both NKx-2.5 and NF-kappaB in the bladders pre-treated with saline, in an additional group, urinary bladders were removed without insertion of the catheter or fluid instillation. Interestingly, in this group (0 hours) almost the same degree of constitutive transcriptional activities was observed.