Characterization of urinary extracellular vesicle proteins in muscle-invasive bladder cancer

Abstract

The mechanisms of bladder cancer progression are unknown, and new treatments and biomarkers are needed. Patient urinary extracellular vesicles (EVs) derive in part from bladder cancer cells and contain a specific protein cargo which may provide information about the disease. We conducted a proteomics study comparing EVs from the muscle-invasive bladder cancer (MIBC) cell line TCCSUP to EVs from normal urothelial line SVHUC. GO term analysis showed that TCCSUP EVs are enriched in proteins associated with the cell membrane, extracellular matrix, and inflammation and angiogenesis signaling pathways. Proteins characteristic of cancer EVs were further screened at the mRNA level in bladder cancer cell lines. In Western blots, three of six proteins examined showed greater than fifteenfold enrichment in patient urinary EVs compared to healthy volunteers (n = 6). Finally, we performed immunohistochemical staining of bladder tissue microarrays for three proteins of interest. One of them, transaldolase (TALDO1), is a nearly ubiquitous enzyme and normally thought to reside in the cytoplasm. To our surprise, nuclei were stained for transaldolase in 94% of MIBC tissue samples (n = 51). While cytoplasmic transaldolase was found in 89-90% of both normal urothelium (n = 79) and non-muscle-invasive samples (n = 71), the rate falls to 39% in MIBC samples (P < 0.001), and negative cytoplasmic staining was correlated with worse cancer-specific survival in MIBC patients (P = 0.008). The differential EV proteomics strategy reported here successfully identified a number of proteins associated with bladder cancer and points the way to future investigation.

Keywords: bladder cancer; exosomes; extracellular vesicles; proteomic; transaldolase.

Figure 1. Isolation and proteomic analysis of bladder cancer EVs

(A) Transmission electron micrograph of bladder cancer patient urinary EVs collected at the time of cystectomy and isolated by ultracentrifugation. Bar = 100 nm. (B) Nanoparticle tracking analysis histogram of particle size in a sample of TCCSUP EVs. Particle concentration is corrected to represent the original EV isolate. (C) Proteomic analysis of differential protein inclusion in TCCSUP EVs vs. SVHUC EVs. At the upper left, the Venn diagram represents 719 individual protein-coding genes identified in the PANTHER database. Annotation classes with > 1.3-fold over-representation in TCCSUP EVs are presented in bar graphs. Classes under-represented in TCCSUP appear at the bottom of each category.

Figure 2. Examination of candidate EV proteins in patient and cell line arrays

(A) A bladder cancer gene expression microarray dataset published by Dyrskjøt et al. was used to screen possible candidates. Six genes were found to be over-expressed in one or more cancer types. Significance of over-expression compared to normal tissue samples is represented by asterisks (* is P < 0.05, ** is P < 0.01, *** is P < 0.001). (B) Expression heat map showing qPCR measurement of seven candidate mRNAs plus EDIL3 in a panel of eight bladder cancer cell lines and two normal urothelial lines (SVHUC and TERT-NHUC). Red = high expression. (C) Western blot analysis of EVs derived from four bladder cancer cell lines. HEXB bands are predicted at two distinct protein sizes. EV marker Alix was used as a loading control.

Figure 3. Transaldolase in clinical samples

(A) Western blot analysis of urinary EVs collected from six healthy volunteers and six bladder cancer patients (stage pT1-pT3). HEXB bands are predicted at two distinct protein sizes. EV marker Alix was used as a loading control. (B) Densitometry measurements of the results in (A) normalized by the Alix signal. Bars represent the mean normalized values. HEXB is presented as the sum of the two observed bands. (C and D) Immunohistochemical staining of transaldolase in a bladder TMA (200X) showing a strong cytoplasmic signal in normal urothelium (C) and a strong nuclear signal, with no cytoplasmic staining, in an MIBC sample (D). (E) Pie graphs showing the number of samples scored positive or negative for transaldolase in cytoplasm (top) and nuclei (bottom). (F) Kaplan-Meier curves showing cancer-specific survival in patients with MIBC scoring positive or negative for cytoplasmic transaldolase.