We performed transcriptomic profiling of a Gene Expression Omnibus (GEO) dataset (GSE32894) composing of 308 UBUC patients [21]. All probes without filtering or preselection were analyzed, and the raw data were imported into the Nexus Expression 3.0 (BioDiscovery, El Segundo, CA, USA) to compute gene abundances, as described in our previous study [22]. A comparative analysis (noninvasive vs. invasive UC) was performed to examine the DEGs related to calcium ion binding (GO:0005509). Those with a significant log${}_2$-transformed expression fold change –0.15 (p 0.01) were selected for further analysis.

A total of 340 UTUC and 295 UBUC patients who underwent curative surgery between 1998 and 2004 were enrolled in the study, and all specimens were procured from our biobank after obtaining informed consent. The patients’ clinical demographic characteristics, pathological features, and survival outcomes were retrospectively reviewed from their medical charts. None of the patients had undergone preoperative radiotherapy or chemotherapy. Postoperative platinum -based adjuvant chemotherapy was administered to patients with nodal involvement or pT3–pT4 diseases. The pathological grade and tumor stage were determined based on the WHO classification criteria and the 7th edition of the american joint committee on cancer (AJCC) staging system, respectively. The Institutional Review Board of Chi Mei Medical Center approved this study (10501005).

Paraffin-embedded tissue blocks were sliced (4 µm) and placed on silane-coated slides, as previously described [22]. Deparaffinization, rehydration, and antigen retrieval were performed according to the standard procedures. The endogenous peroxidase activity was blocked using 3% H${}_2$O${}_2$ (ab64218, Abcam, Cambridge, England). The slides were then incubated with primary antibodies against CRTAC1 (1:100, ab254691, Abcam, Cambridge, England), matrix metallopeptidase 2 (MMP2) (1:100, ab86607, Abcam), or CD31 (1:100, ab28364, Abcam) for one hour. The target proteins were detected using a ChemMate™ EnVision™ Kit (Dako, Carpinteria, CA, USA). Positive CRTAC1 expression was characterized by membranous and/or cytoplasmic staining in UC cells. Sections processed without the primary anti-CRTAC1 antibody were used as negative controls. Two independent pathologists estimated the cancer cell distribution and the intensity of the immunohistochemical (IHC) staining to generate an H-score using the formula $\mathrm{\Sigma }$Pi(i + 1), where Pi is the percentage of stained cancer cells (0%–100%) and i is the intensity of the stained cancer cells (0–3+). The immunoreactivity of CRTAC1 was described as low or high levels of expression according to the median H-score,. As previously described, we identified the CD31-labeled vessels and used ImageJ software to calculate the tumoral microvessel density (MVD) [23].

Five human UC-derived cell lines, namely TCCSUP (American Type Tissue Culture Collection [ATCC], VA), J82 (ATCC, VA), T24 (ATCC, VA), BFTC905 (Food Industry Research and Development Institute [FIRDI], Taiwan), and BFTC909 (FIRDI, Taiwan), were screened for CRTAC1 expression. A non-tumoral uroepithelial cell line, SV-HUC-1 (ATCC, VA), was used as a control. BFTC909 and T24 cells, which exhibit relatively low levels of endogenous CRTAC1, were selected for this study. These cells were cultured as previously described [24]. All cells were incubated at 37 °C in a humidified incubator containing 5% CO${}_2$. The cell lines were regularly tested for mycoplasma and were authenticated by short tandem repeat genotyping.

The Phoenix-Amphotropic (AMPHO) cell line (ATCC, VA) was used to produce lentiviral particles containing CRTAC1. Briefly, the expression plasmids for CRTAC1 and the control vector (pLenti-GIII-CMV) were purchased from Applied Biological Materials, Inc (Richmond, BC, Canada). Following a series of transfections, as described previously [24], T24 and BFTC909 were incubated with culture medium containing viral supernatant and 10 µg/mL polybrene for 24 hours. Afterward, medium containing viral solution was replaced with fresh medium. UC cell lines overexpressing CRTAC1 were obtained as stable clones after puromycin selection (2 µg/mL).

Total RNA was extracted from the UC cells using a Quick-RNA™ Miniprep Kit (Zymo Research, Beijing, China), and was reverse-transcribed using a Maxima First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA). The cDNA was mixed with the corresponding TaqMan assay probes (CRTAC1, Hs00907892_m1; MMP2, Hs01548727_m1; MMP9, Hs00957562_m1; POLR2A, Hs01108291_m1; Applied Biosystems) using TaqMan™ Fast Advanced Master Mix (Thermo Scientific). We then performed quantitative RT-PCR for the mRNA level by the 2${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{CT}}$ method using a StepOnePlus™ System (Applied Biosystems, Waltham, MA, USA). The relative expression levels of the target mRNAs were normalized to those of POLR2A RNA.

We used PRO-PREP™ Protein Extraction Solution (iNtRON Biotechnology, Seongnam-Si, Gyeonggi-do, Republic of Korea) to extract total cellular proteins. The protein concentration was determined by a BCA assay kit (Thermo Fisher Scientific). Thirty micrograms of protein were loaded on an mPAGE Bis-Tris Precast Gel (Merck Millipore, Burlington, MA, USA) and transferred onto an Immobilon®-P PVDF membrane (Merck Millipore). The PVDF membranes were blocked with 5% skim milk (Millipore Sigma, Hong Kong, China) incubated overnight at 4 °C with the following primary antibodies: anti-CRTAC1 (ab254691, Abcam), anti-MMP2 (ab86607, Abcam), or anti-GAPDH (ab181602, Abcam). GAPDH served as an internal control. After washing, the membrane was incubated with a diluted secondary antibody (horseradish peroxidase [HRP] donkey anti-rabbit immunoglobulin G, BioLegend) for one hour. CRTAC1, MMP2, and GAPDH were visualized by enhanced chemiluminescence (Thermo Scientific).

Cell proliferation was evaluated using a Cell Proliferation Assay Kit (Fluorometric; BioVision, Hong Kong, China). Briefly, 1000 cells were pleated in 96-well microplates and incubated for 24, 48, and 72 h at 37 °C. Subsequently, 25 µL of the reaction mixture, including 1$\times$ nuclear dye/cell lysis buffer solution and 1$\times$ nuclear dye, was added to each well. After 15 min of incubation, the fluorescence intensity was determined using a standard microplate reader (excitation/emission = 480/538 nm). The assays were repeated at least three times.

The Boyden chamber technique (Transwell® analysis, Thermo Scientific, Hong Kong, China) was used to determine the cell migration and invasion abilities [25]. Cell migration and invasion assays were performed using Falcon HTS FluoroBlok 24-well inserts (BD Biosciences, Franklin Lakes, NJ, USA) and a QCM™ Collagen Cell Invasion Assay (Millipore), respectively. Each insert was rehydrated with serum-free medium and 1 $\times$ 10${}^6$ UC cells suspended in serum-free medium were plated in the upper chamber and incubated with medium containing 10% FBS in the lower chamber. After 12–24 h of incubation, migrating and invading cells passing through the inserts were detached, stained with the provided dye and transferred to 96-well plates for colorimetric analysis at 560 nm. The assays were repeated at least three times.

Human umbilical vein endothelial cells (HUVECs) were used to investigate the effects of CRTAC1 on UC-induced angiogenesis. The Matrigel® Matrix (Corning, Tewksbury, MA, USA) was used to precoat each inner well of the µ-Slide Angiogenesis (Ibidi, Gräfelfing, Germany) for 30 min at 37 °C. Then, 50 µL of cell suspension containing 7 $\times$ 10${}^3$ HUVECs in 25 µL conditioned medium and 25 µL endothelial cell medium with 2% FBS was seeded on top of the Matrigel®. After incubation for 5 h at 37 °C, the capillary-like tube structures were evaluated and counted under a phase contrast microscope. The assays were repeated at least three times.

The vector pcDNA3-CRTAC1 was constructed as previously described [26]. MMP2 and MMP9 promoter fragments were cloned by PCR amplification and inserted into a luciferase reporter gene plasmid vector (Promega, Fitchburg, WI, USA). The MMP2 and MMP9 promoter reporter constructs and pcDNA3-CRTAC1 construct or their matching empty constructs were co-transfected in cells using PolyJet${}^{\text{TM}}$ transfection reagent (SignaGen Laboratories, Shandong, China) for 48 h. We then used a Dual-Glo® Luciferase Reporter Assay System (Promega) to measure the luciferase activity following the manufacturer’s protocol.

Associations between CRTAC1 expression and different variables were evaluated using the chi-square test. For survival analyses, the disease-specific survival (DSS) and metastasis-free survival (MFS) were plotted using Kaplan–Meier curves and estimated using the log-rank test at the univariate level. The independent prognostic factors were estimated using a multivariate Cox proportional hazards model. All cellular functional studies were done with three replicates. Student’s t-test was used to analyze differences in cell proliferation, migration, invasion, HUVEC tube formation, and luciferase activity. Statistical analyses were performed using SPSS 19 software (SPSS Inc., Chicago, IL, USA). In all analyses, a p-value of 0.05 was considered to reflect statistical significance.

We identified 34 probes covering 22 transcripts associated with calcium ion binding (GO:0005509) in UBUC invasion using a published transcriptome dataset (GSE32894). These genes were significantly downregulated at the high tumor stage (Fig. 1 and Table 1). CRTAC1 was selected for further evaluation because it was the most downregulated gene. Using the Gene Expression Profiling Interactive Analysis database, CRTAC1 was found to be significantly decreased in UBUC (n = 404) compared to in adjacent normal tissues (n = 19) (p 0.001). Intriguingly, violin plots showed that CRTAC1 mRNA levels significantly decreased as the UBUC stage increased (from stage II to IV). Moreover, UBUC patients with high CRTAC1 mRNA expression had better overall survival than those with low CRTAC1 mRNA expression (p = 0.0007) (Supplementary Fig. 1). These results prompted us to assess the prognostic value and clinical relevance of CRTAC1 in a large cohort of UC patients.

Fig. 1.

Data mining. Expression profiles of genes associated with the calcium ion binding (GO:0005509) extracted from a published transcriptome of UC (GSE32894) in the Gene Expression Omnibus database. CRTAC1 was the most downregulated gene associated with UC progression. UC, urothelial carcinoma.

In total, 635 patients (UTUC, 340; UBUC, 295) with a mean age of 65.8 years were included in this study (Table 2). In the UTUC cohort, the majority of patients (n = 284, 83.5%) had a high histological grade, 159 patients (46.8%) had advanced UTUC (pT2-pT4), and 28 patients (8.2%) had metastatic nodal disease at initial diagnosis. Forty-nine patients (14.4%) developed concurrent renal pelvic and ureteral tumors, while 62 (18.2%) developed multiple tumors. In addition, perineural invasion was detected in 19 patients (5.9%), while vascular invasion was detected in 106 patients (31.2%). The UBUC cohort comprised 239 (81.0%) patients with high histological grade tumors, 123 (41.7%) with advanced-stage disease (pT2–pT4), and 29 (7.8%) with metastatic lymph nodes. A total of 156 tumors (52.9%) exhibited high mitotic activity. Perineural invasion and vascular invasion were detected in 49 (16.6%) and 20 (6.8%) patients, respectively.

Immunostaining was performed to evaluate CRTAC1 expression levels in the surgical tissue and revealed that invasive UC had lower CRTAC1 immunoreactivity than noninvasive UC (Fig. 2). The clinical significance of CRTAC1 expression in UC was also examined (Table 2). In the UTUC cohort, CRTAC1 expression significantly correlated with the tumor stage (p 0.001), histological grade (p 0.001), lymph nodal status (p 0.001), vascular invasion (p 0.001), perineural invasion (p = 0.002), and mitotic activity (p 0.001). In the UBUC cohort, low CRTAC1 immunoexpression was associated with pT2–pT4 stage (p 0.001), high histological grade (p 0.001), nodal metastasis (p 0.001), perineural invasion (p = 0.005), vascular invasion (p 0.001), and a high mitotic rate (p = 0.004).

Fig. 2.

Immunohistochemistry. Lower expression of CRTAC1 is associated with higher MMP2 expression, higher CD31-labeled microvascular density, and higher tumor stage in UC. (Scale bar = 200 µm) (CRTAC1, cartilage acidic protein 1; MMP2, matrix metallopeptidase 2; NMIBC, non-muscle invasive bladder cancer; MIBC, muscle invasive bladder cancer).

Within a median follow-up of 31.7 months, 70 UTUC and 76 UBUC patients developed tumor metastasis, while 61 UTUC and 52 UBUC patients died of UC.

Univariate and multivariate analyses were conducted to assess whether CRTAC1 expression affected cancer metastasis and death. With regard to UTUC (Table 3), 55 patients (29.1%) with low CRTAC1-expressing cancers died, while 59 patients (34.7%) subsequently developed cancer metastasis. Only 11 patients (6.5%) with high CRTAC1-expressing cancers developed metastatic cancers, while 9 patients (6.2%) died of UTUC. In particular, low CRTAC1-expressing tumors predicted worse DSS (Fig. 3A; p 0.0001) and MFS (Fig. 3B; p 0.0001) based on the results of the Kaplan-Meier survival analysis. In addition to CRTAC1 immunostaining status, multifocal tumors, advanced tumor stage, high histological tumor grade, nodal metastasis, and vascular and perineural invasion were significantly associated with inferior MFS and DSS. In the multivariate analysis, high CRTAC1 immunoactivity was clearly associated with better MFS and DSS (DSS: hazard ratio [HR]: 0.188, 95% confidence interval [CI]: 0.073–0.481, p 0.001; MFS: HR: 0.255, 95% CI: 0.152–0.426, p 0.001).

Fig. 3.

Kaplan-Meier survival curves. Low CRTAC1 expression is associated with a significant prognostic impact on disease-specific survival and metastasis-free survival of patients with UTUC (A and B, respectively) and UBUC (C and D, respectively). UTUC, upper urinary tract urothelial carcinoma; UBUC, urinary bladder urothelial carcinoma.

In UBUC (Table 4), low CRTAC1 expression levels were associated with higher rates of postoperative cancer metastasis (31.2% vs. 10.0%) and cancer-related deaths (27.1% vs. 8.8%) than high CRTAC1 expression levels. In Kaplan–Meier survival analysis, CRTAC1 immunoexpression (Fig. 3C,D), pT stage, tumor grade, perineural invasion, vascular invasion, lymph node status, and mitotic rate were significantly correlated with worse DSS and MFS. In addition, multivariate Cox regression analysis showed that CRTAC1 immunoexpression status was an independent prognosticator of cancer-related death (HR: 0.220, 95% CI: 0.102–0.474, p 0.001) and metastasis occurrence (HR: 3.374, 95% CI: 0.218–0.642, p 0.001).

To understand the biological function of CRTAC1, endogenous CRTAC1 expression in UC cell lines was determined. Compared to normal urothelial primary cells (SV-HUC-1), all five UC-derived cell lines had lower CRTAC1 mRNA and protein expression (Fig. 4A). Of these, BFTC909 and T24 cells exhibited the lowest levels of CRTAC1 expression; therefore, CRTAC1 overexpression was induced in these two cell lines (Fig. 4B). The overexpression of CRTAC1 in BFTC909 and T24 cells significantly attenuated cell proliferation (Fig. 4C). Matrigel® invasion assays indicated that CRTAC1 overexpression also significantly decreased the number of invading tumor cells, thus indicating its ability to inhibit metastasis (Fig. 4D). Moreover, conditioned medium from CRTAC1-overexpressing BFTC909 and T24 cells markedly inhibited HUVEC tube formation compared to that in the mock group (Fig. 5A). In addition to in vitro studies, we studied the association between CRTAC1 and MVD in our UC specimens. Notably, high CRTAC1 expression was significantly correlated with less CD31-labeled MVD in UTUC and UBUC. (Fig. 5B) To identify the potential cellular pathways that are involved in the regulation of UC invasiveness by CRTAC1, MMP2 was selected for further studies. Initially, IHC staining showed that high CRTAC1 expression negatively correlated with low MMP2 expression in UTUC and UBUC (Table 2; Fig. 2). qRT-PCR and immunoblotting showed that exogenous CRTAC1 expression markedly suppressed MMP2 mRNA and protein expression in BFTC909 and T24 cells (Fig. 6A). Finally, analysis of luciferase activity driven by the MMP2 promoter showed that MMP2 transactivation was negatively associated with CRTAC1 expression in UC cells. These results confirm the role of MMP2 in CRTAC1-driven UC aggressiveness (Fig. 6B).

Fig. 4.

CRTAC1 expression inhibits growth and invasion of UC cells in vitro. (A) Compared to SV-HUC-1 cells, endogenous CRTAC1 mRNA and protein expression is lower in T24 and BFTC909 cell lines. (B) CRTAC1 overexpression was induced in these two cell lines. Compared with lentiviral infection with the mock sequence, lentiviral infection with CRTAC1 significantly increased the mRNA and protein levels of CRTAC1 in BFTC909 and T24 cells. (C) The overexpression of CRTAC1 in BFTC909 and T24 cells significantly attenuates the cellular proliferation. (D) Using Transwell® migration and invasion assays, cell invasion is significantly reduced in CRTAC1-transfected T24 and BFTC909 cell lines, compared to that in the corresponding empty controls. (*, p 0.05).

Fig. 5.

CRTAC1 expression inhibits angiogenesis of UC. (A) Tube formation is markedly decreased when the HUVECs are incubated with conditioned medium from the CRTAC1-overexpressing T24 and BFTC909 cells than that from the mock groups. (B) High CRTAC1 expression is significantly correlated with less CD31-labeled microvascular density in UTUC and UBUC. (*, p 0.05).

Fig. 6.

CRTAC1 inhibits UC invasion by transcriptional repression of MMP2. (A) Exogenous CRTAC1 expression significantly downregulated the MMP2 mRNA and protein levels in BFTC909 and T24 cells, using qRT-PCR and western blotting. (B) The luciferase activity of MMP2 promoter construct was significantly lower in the T24 and BFTC909 cells transfected with the CRTAC1-expressing vector than that in the empty controls. (*, p 0.05).