Immortalized normal human bladder epithelial cells, TRT-HU1, were maintained as described previously^[17]. The TRT-HU1 cell line was constructed and extensively characterized in previously published papers. The passage number of each cell line was below 10, and mycoplasma contamination was tested for monthly via PCR analysis. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM), containing 10% fetal bovine serum (FBS, Invitrogen), 1% penicillin/streptomycin, and 1% L-glutamine (Sigma-Aldrich Corp., St. Louis, MO, USA) in a 37 °C humidified incubator with 5% CO₂. To test cell growth during exposure to caffeine, TRT-HU1 cells were seeded onto 6-well plates with a density of 5 × 10⁴ cells/well. Cells were then incubated with standard growth medium with varying doses of caffeine at 0, 0.005 mM, 0.05 mM, or 0.1 mM (C6035, Sigma-Aldrich Corp.) or vehicle for 24, 48, or 72 hrs. Cell proliferation was measured by manually counting cells using a hemocytometer. The averages of each count were used as the total density of the well after each time point. For crystal violet staining, the culture medium was removed, the cells were fixed with 4% paraformaldehyde at room temperature for 5 min and then stained with 0.05% crystal violet for 15 min. The cells were then washed with tap water, after which the water was removed and the cells were dried out on filter paper, and the cell plates were scanned and quantified as described previously^[18]. All experiments were run in triplicate for each cell line, and the data are representative of three independent trials.

Antibodies against various proteins were obtained from the following sources: ACTG2 (ab189385, Abcam), ACTA2 (ab5694, Abcam), MYH2 (ab124937, Abcam), MYH7B (ab172967, Abcam), HISTH2B (ab52599, Abcam), HIST1H2BM (SAB1301739, Sigma), and β-actin (A1978, Sigma). Commercially available horseradish peroxidase (HRP)-conjugated secondary antibodies (7074, 7076) were obtained from Cell Signaling Technology. All other chemical reagents were procured from Sigma Chemical Corp.

Tandem mass tagging (TMT)-based quantitative proteomics was performed as described^[19]. Briefly, cellular protein was extracted from caffeine-treated and control cells using 4% SDS-containing buffer. The protein concentration was measured using the Pierce 660nm Assay Kit. From each sample, 60 μg of protein was digested with trypsin using filter-aided sample preparation (FASP) and labeled with TMT6plex reagents in parallel. After TMT labeling, the peptides were merged, desalted with C18 spin columns (Thermo Scientific), and fractionated via high-pH reversed phase liquid chromatography (RPLC) using an Ultimate 3000 XRS System (Thermo Scientific). For high-pH RPLC, about 50 μg TMT-labeled peptides were loaded onto a 100-mm Hypersil GOLD C₁₈ column (2.1 mm inner diameter, 3 μm particle size, 175 Å pore size) (Thermo Scientific), flushed for 3 min with solvent A (10 mM ammonium formate, pH 10), and then separated with a 7-min linear gradient of 0–40% solvent B (10 mM ammonium formate, 95% acetonitrile, pH 10.0). A total of 24 fractions were collected, concentrated into 12 fractions, and dried down in a SpeedVac (Thermo Scientific). Peptides in each fraction were redissolved with 0.2% formic acid and analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) using an EASY-nLC 1000 connected to an LTQ Orbitrap Elite Mass Spectrometer (Thermo Scientific). Briefly, peptides were loaded onto a 2-cm trap column (PepMap 100 C18, 75 μm inner diameter, 3 μm particles, 100 Å pore size) and separated by a 50-cm EASY-Spray column (PepMap RSLC C₁₈, 75 μm inner diameter, 2 μm particles, 100 Å pore size) heated to 55°C. For low-pH RPLC separation, the mobile phase consisted of 0.1% formic acid in water (phase A) or acetonitrile (phase B). The LC gradient was 4–24% B over 200 min, 24–50% B over 20 min, and 50–100% B over 5 min at a flow rate of 150 nL/min, followed by 100% B over 15 min at a flow rate of 300 nL/min.

Mass spectra were acquired in a data-dependent manner, selecting up to the 15 most abundant precursor ions for higher-energy collisional dissociation. The mass resolution for precursor and fragment ions was set to 120,000 and 30,000, respectively. The isolation width was set as 1.5 and the normalized collision energy was set as 40. Database searching and protein quantification was performed by Proteome Discoverer (v2.1), using the SEQUEST algorithm. The acquired raw data were searched against the human Uniprot Protein Sequence Database (released on 01/22/2016, containing 20,985 protein sequences). Searching parameters were set as follows: trypsin, up to two missed cleavage; precursor ion tolerance of 10 ppm, fragment ion tolerance of 0.02 Da; carbamidomethylation of cysteins and TMT6plex modification of lysines and peptide N-term as fixed modifications; acetylation of protein N-term, oxidation of methionine and deamidation of asparagines and glutamines as variable modifications. A standard false discovery rate of 1% was applied to filter peptide-spectrum matches (PSMs), peptide identifications, and protein identifications.

For protein quantification, peptides with >30% precursor ion interference were excluded to minimize erroneous quantification caused by precursor ion interference. PSM level information was extracted using the Proteome Discoverer^[20]. After quantifying each PSM intensity, the peptide intensities were summarized, as described by Niu et al. ^[21]. In brief, this was done in 4 steps: 1) the normalized log2 intensity of the PSMs matched to each peptide by the substrate mean PSM from each reporter intensity was centered, 2) outliers were detected using Dixon’s Q-test and generalized electrostatic discharge (ESD) test, 3) the mean intensity without outliers was taken, and 4) the grand-mean intensities of the three highest abundant PSMs were added before being mean-centered. Bipartite graphs of peptides and protein groups were generated according to the information of the aligned peptides. Among the proteins in the protein group, we defined the representative protein that had the largest number of peptides or unique peptide^[22]. When there were more than proteins with the same number of peptides in the same protein group, we selected the protein that had the higher sequence coverage. We then computed the relative intensity of the protein group using a linear-programming formulation, as previously described^[23].

To identify the DEPs, we selected proteins that have more than two non-redundant peptides. For the selected proteins, we then performed an one sample t-test using log₂ fold-changes to compute the significance of t value. For this statistical test, an empirical distribution of the null hypothesis, a protein is not differentially expressed, was estimated by following steps: 1) 100,000 random permutations were applied to the samples, 2) t-values were computed using the log₂ fold-changes of the randomly permutated samples, and 3) the Gaussian kernel density estimation method was applied to t-values. The FDRs of each protein from the one sample t-tests were then calculated using Storey’s method^[24]. The DEPs were identified as those with a FDR < 0.05 and absolute log₂ fold change > 0.58 (1.5 fold). Functional enrichment analysis of DEPs was performed using DAVID software (Ver. 6.8)^[25]. Significantly enriched cellular processes were selected for if they had a p-value < 0.05.

Cells were seeded onto 10 cm plates and exposed to 0.05 mM concentrations of caffeine for 72 hrs. Cells were lysed with RIPA buffer (20 mM Tris, 150 mM NaCl, 1% Nonidet, P-40, 0.1 mM EDTA) (Pierce, ThermoFisher) supplemented with a phosphatase inhibitor cocktail (ThermoFisher). The protein concentration of each sample was measured with the Bradford Protein Kit Assay, according to the manufacturer’s instructions (Pierce, ThermoFisher). Equal amounts of protein extract were separated by SDS-PAGE and transferred onto a PVDF membrane. The membranes were then blocked with 5% bovine serum albumin (BSA) or 5% nonfat milk in Tris-buffered saline with 0.1% Tween 20 (TBST [2.42 g/L Tris–HCl, 8 g/L NaCl, and 1 mL/L Tween 20 (pH 7.6)]) and incubated overnight at 4 °C with specific primary antibodies in TBST. The membranes were then incubated with secondary antibodies conjugated with horseradish peroxidase, as described previously. β-actin was used as an internal control. All western blot experiments were run in at least triplicates for each antibody.

TRT-HU1 cells were seeded onto a 24-well Seahorse culture plate at a density of 50,000 cells per well 24 hrs before the Seahorse assay. Media was then changed to XF Base Medium (pH 7.4) supplemented with 10 mM glucose, 1mM sodium pyruvate, and 1mM sodium glutamine. Cells were equilibrated for 1 hr in a non-CO₂ incubator at 37 °C before the assay began. Cells were treated with caffeine at 0.05 mM, or 0.5 mM for experiments. Chemical reagents (Sigma) were used at final concentrations, as follows: 1 μM oligomycin—an ATP synthase inhibitor, 1 μM (FCCP) carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone—an uncoupling agent, and a mixture of 0.5 μM antimycin A—a cytochrome C reductase inhibitor and 0.5 μM rotenone—a complex I inhibitor. Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) was monitored during the duration of the assay run. OCR and ECAR were monitored before and after addition of caffeine. Results were normalized to protein concentrations determined by BCA assay (Thermo Scientific).

Student’s T-test was performed to confirm differential expression of the proteins between two groups. Variables with normal distribution were presented as mean ± SD. All reported p-values are two-tailed, with p-values < 0.05 indicating statistical significance.

Due to the lack of knowledge regarding the effects of caffeine on biological and proteomic perturbations in the normal bladder epithelium, this study aimed to examine the whole proteome alterations caused by caffeine consumption. To gain insight on the underlying mechanism of caffeine on the bladder epithelium, we treated normal bladder epithelial cells with caffeine and performed TMT-based quantitative proteomic analysis, as outlined in Figure 1A. Based on previous literature, we opted to use caffeine concentrations within normal physiological consumption^{[12, 26]}. Whole-cell lysates in biological duplicates were digested with trypsin. Using LC-MS/MS and followed by bioinformatic analyses, we identified 44,597 peptides corresponding 5,832 proteins with more than 2 peptides in more than two sets of pooled lysates from at least three biological samples per condition. The caffeine-treated group had 5,821 identified proteins with high expression, while the control group had 5,808. (Figure 1B). We then performed functional categorization of the proteins to check whether our quantitative proteomic analysis was biased by any cellular compartments or biological process-related proteins using Panther software^[27]. All the identified proteins can be categorized into 14 cellular processes, including cellular process, metabolic process, cellular component organization of biogenesis, localization, biological regulation, response to stimulus, developmental process, multicellular organismal process, biological adhesion, immune system process, locomotion, reproduction, growth, and cell killing, indicating that this proteome reveals major biological functions of bladder epithelial cells in normal physiology (Figure 1C). In addition to this, we examined the cellular localization of the detected proteins, which showed the significant enrichment of 8 cellular compartments including cell part, organelle, macromolecular complex, membrane, extracellular region, cell junction, synapse, and extracellular matrix. (Supplementary Figure 1). Collectively, our quantitative proteomic analysis identified a comprehensive list of proteins in all cellular components and illustrated the cellular functions of highly expressed proteins in normal bladder epithelial cells.

We determined which proteins were differentially expressed after caffeine treatment. DEPs were selected for if they had an absolute log₂ fold change greater than 0.58 and p-value less than 0.05. In total, we identified 32 upregulated and 25 downregulated proteins between the control vs. caffeine groups (Table 1, 2, and Figure 2A). As shown in the volcano plot, some DEPs, including PSMC6 (proteasome 26S subunit ATPase 6), RUFY1 (RUN and FYVE domain containing 1), IARS (isoleucyl-tRNA synthetase), MCU (mitochondrial calcium uniporter), NAV1 (neuron navigator 1), RARG (retinoic acid receptor, gamma), and etc, were significantly increased with caffeine treatment, while ST13P4 (ST13, Hsp70 interacting protein pseudogene 4), EIF5AL1 (eukaryotic translation initiation factor 5A-like 1), WASH2P (WAS protein family homolog 2 pseudogene), RBMS3 (RNA binding motif single stranded interacting protein 3), and etc decreased (Figure 1C).

Next, to understand the function of the perturbed proteins, we performed gene set enrichment analysis (GSEA) using the hallmark gene sets from the Molecular Signature Database (MSigDB)^[28]. As a result, we found that “glycolysis” and “PI3K/AKT/MTOR” signaling gene sets were significantly enriched by differential expression due to caffeine treatment (Figure 2C). We also checked the enrichment of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in the same context and found that the “neurotrophin signaling pathway” and “purine metabolism” were identified to be significantly enriched in the caffeine-treated condition (Figure 2D). This result suggests that caffeine may stimulate bladder epithelial by altering the activation of the PI3K/AKT/MTOR pathway, neurotrophin signaling, and purine metabolism.

For this, we conducted separate functional enrichment analyses on each of the 32 up- and 25 downregulated proteins using DAVID^[25]. The results suggest that upregulated DEPs were significantly enriched for “actin-myosin filament sliding”, “muscle contraction”, “purine nucleotide metabolic process”, “ATP metabolic process”, “autophagy”, “response to external biotic stimulus”, and “response to other organism”. Downregulated DEPs were significantly enriched for “chromatin assembly”, “DNA packaging”, and “cellular macromolecular complex assembly” (Figure 3A). We extracted a list of the most significant upregulated DEPs belonging to “purine nucleotide metabolic process”, “autophagy”, “muscle contraction”, “response to external biotic stimulus”, and “chromatin assembly” downregulated DEPs “chromatin assembly”, respectively (Figure 3B and 3C). While “purine nucleotide metabolic process”, “autophagy”, “muscle contraction”, “response to external biotic stimulus” were enriched for by the upregulated DEPs, “chromatin assembly” was enriched for by downregulated DEPs.

To evaluate the direct effects of caffeine on bladder epithelial cells in vitro, we used TRT-HU1 cells in our study as described in Methods section. After incubating the TRT-HU1 cells in varying concentrations of caffeine (0.005, 0.05, and 0.1 mM) for 24, 48, or 72 hrs, we sought to determine if caffeine affects or controls cell proliferation and metabolism. We found that the cell proliferation rate of caffeine-treated TRT-HU1 cells was not significantly altered, compared to controls (Figure 4A and 4B).

To examine the effects of caffeine on cell metabolism, we performed Seahorse Respirometry Analysis. Caffeine treatment over a 12-hr period did not alter the oxygen consumption rate (OCR, mitochondrial respiration) or extracellular acidification rate (ECAR, glycolysis) in TRT-HU1 cells (Figure 4C). Briefly, basal OCR and ECAR measurements (first three time points) were monitored before caffeine (0.05 mM) or vehicle (methanol) was administered. OCR and ECAR continued to be monitored up to 12 hrs post treatment. Non-treated cells (media administration only) were included to control for any effects of the cells being in the analyzer for an extended period (Figure 4C). We could not detect any changes in both the OCR and ECAR, even after a 12-hr treatment with 0.05 mM caffeine. Additional western blot analysis demonstrated that expression of ACTG2 and HIST1H2BM were reduced in the 0.05 mM caffeine-treated cells, while expression of ACTA2, MYH2 and MYH7B were upregulated. The expression of histone H2B and ACTB remained unchanged (Figure 4D).