What's known on the subject? and What does the study add?
The current basis for diagnosis and prognosis in urinary bladder cancer is based on the pathologists' assessment of a biopsy of the tumour. Urinary biomarkers are preferable as they can be non-invasively sampled. Urinary cytology is the only test with widespread use but is hampered by poor reproducibility and low sensitivity.
By studying the protein expression in bladder tumour tissue samples of proteins previously found in elevated levels in the urine of patients with bladder cancer, we have been able to show that these proteins originate from the tumour. The immunoreactivity of three of the investigated proteins increased with higher stage. Also a serine peptidase inhibitor was found to be predictive of progression from non-muscle-invasive to muscle-invasive tumours.
To analyse the expression of five bladder cancer-associated urinary proteins and investigate if expression is related to the malignant phenotype of the tumour.
To explore the possible prognostic value of these proteins.
Patients and Methods
Urine samples, 16 from patients with bladder cancer and 26 from controls, were used in Western Blotting experiments.
Tissue microarrays with bladder tissue from 344 patients diagnosed with bladder cancer between 1984 and 2005 was used in immunohistochemistry experiments.
The proteins apolipoprotein E (APOE), fibrinogen β chain precursor (FGB), leucine-rich α2-glycoprotein (LRG1), polymerase (RNA) I polypeptide E (POLR1E), α1-antitrypsin (SERPINA1) and topoisomerase 2A (TOP2A) were probed with antibodies validated by the Human Protein Atlas.
Increased expressions of APOE, FGB and POLR1E were correlated with increased tumour stage (P < 0.001).
Expression of SERPINA1 in Ta and T1 tumours was found to increase the risk of tumour progression (hazard ratio 2.57, 95% confidence interval 1.13–5.87; P = 0.025)
All proteins previously detected in urine from patients with bladder cancer were also expressed in bladder cancer tissue.
The expression of APOE, FGB and POLR1E increased with stage and they are potential diagnostic markers.
SERPINA1 was identified as a prognostic marker candidate.
Every year >380 000 people are diagnosed with bladder cancer and 150 000 die from this disease . Many people would therefore benefit from improved diagnostic tools. Bladder tumours are most often discovered when patients seek medical attention for haematuria. Diagnosis of bladder tumours is performed by cystoscopy, which is an unpleasant and invasive procedure. Different treatments, e.g. chemotherapy, radiation and surgery, are applied depending on what stage of bladder cancer is diagnosed. Patients diagnosed with non-muscle-invasive tumours do not need surgery to remove the whole bladder, but require monitoring with cystoscopy for tumour recurrence on a regular basis. These patients would benefit most from a non-invasive urine-based diagnostic and prognostic tool and our research aims to discover new urinary marker proteins for bladder cancer. Urinary biomarkers must not only have the potential to replace the current invasive monitoring procedures but also to reduce the associated costs of monitoring and enable screening of patients for early detection of disease.
In our previous study, using unbiased mass spectrometric (MS) abundance measurement of proteins, apolipoprotein E (APOE), fibrinogen β chain precursor (FGB), leucine-rich α2-glycoprotein (LRG1) and α1-antitrypsin (SERPINA1) were present at higher concentrations in urine from patients with bladder cancer compared with urine from controls . These findings were further confirmed by complementary Western blotting and therefore, these proteins are interesting urinary biomarker candidates for bladder cancer. Also, topoisomerase 2A (TOP2A) and polymerase (RNA) I polypeptide E (POLR1E), were detected by MS experiments and classified as possible tumour-relevant proteins in urine but have not yet been quantified in urine with another method. These proteins can be classified into two groups, acute-phase proteins (APOE, SERPINA1, FGB and LRG1) and proteins involved in RNA transcription (POLR1E and TOP2A). All proteins identified have been reported as cancer relevant, i.e. by affecting the potential of applied treatment (APOE) [3, 4], by enhancing the anti-apoptotic characteristics of cancer cells (SERPINA1) , by being involved in growth, proliferation and metastatic potentials in cell lines (FGB, POLR1E and LRG1) [6-8] or by enhancing the migration and invasiveness of cell lines (LRG1) . These reports refer to cancer cells of other origin than bladder. TOP2A is the only protein previously reported to be involved in bladder cancer tumorigenesis and has been investigated by immunohistochemistry (IHC) . Bladder tumours with high expression of TOP2A correlate with increased risk of recurrence after treatment and increased risk of death . Further, Simon et al.  reported that the TOP2A and human epidermal growth factor receptor 2 (HER2) genes, are often co-amplified and especially in high-grade bladder tumours, and that TOP2A is over-expressed in late stage compared with early stage bladder tumours.
In the present study we have validated with IHC on tissue microarrays (TMA) the tissue expression of the bladder cancer urinary protein candidate biomarkers, APOE, FGB, LRG1, POLR1E and SERPINA1, identified in our initial study . TOP2A was excluded as it, as described above, has already been investigated with this approach . The aims of the validation were: (i) to validate the presence of TOP2A and POLR1E in bladder cancer urine, (ii) to analyse the tissue immunoreactivity of urinary proteins associated with urinary bladder cancer, (iii) to investigate if the expression is related to malignant phenotype of the tumour and iv) to explore the possible prognostic value of these proteins.
Patients and Methods
The 344 Uppsala University Hospital patients included in this IHC study were diagnosed with bladder cancer between 1984 and 2005. Clinical and histopathological data is presented in Table 1. The follow-up data also included progression-free survival (PFS), overall survival (OS) and disease-specific survival (DSS). The endpoints PFS, OS and DSS were calculated from the date of surgery to date of event or last follow-up. At follow-up the patients with non-muscle-invasive disease were categorised as having none, few (1–2 recurrent tumours ≤18 months) or frequent recurrences (≥3 recurrences within the same time period). Progression was defined as shift of the tumour into a higher stage. The mean (sd, range) time to progression was 51.5 (17.0, 2.0–60.0) months. Follow-up times for non-recurrent and non-progressing cases were ≥4 and ≥5 years, respectively.
Table 1. Clinico-histopathological characteristics of the patients in the TMA study.
In all, 16 urine samples from patients with bladder cancer and 26 from controls, were used in the western blotting experiments .
Western blotting was used to measure the abundance of TOP2A and POLR1E in urine samples from patients with bladder cancer and controls using anti-POLR1E (HPA022527, Atlas Antibodies AB) and anti-TOP2A (HPA006458, Atlas Antibodies AB) as primary antibodies. The procedure was performed on the same sample set and evaluated as described by Lindén et al. . The urine sample collection was approved by the Regional Ethical Review Board of Uppsala (reference numbers 2008:252 and 2007:060).
In the TMA construction , tissue from an independent set of patients (see above), compared with the initial urine study , was used. The use of these patient samples for protein profiling was approved by the Regional Ethical Review Board of Uppsala (reference number 2005:339). Two pathologists identified and categorised the tumour areas used for TMA production, according to the WHO grading system of 2004 . The TMA (URO Fas A) comprised of 120 of each non-muscle-invasive (Ta and T1 stages) tumours and 120 muscle-invasive (T2–T4 stages) tumours. All 360 tissue samples were represented in duplicate tissue cores (1 mm diameter).
IHC was performed according to standard protocol , using the primary antibodies anti-APOE (M068-3, MBL International, Woburn, MA, USA), anti-FGB (HPA001900, Atlas Antibodies AB), anti-LRG1 (HPA001888, Atlas Antibodies AB), anti-POLR1E (HPA022527, Atlas Antibodies AB), anti-SERPINA1 (HPA001292, Atlas Antibodies AB). An immunostaining where the primary antibody had been excluded was used as a negative control.
Scoring of the immunostaining was performed according to a previously published scoring system . Intensities were scored as weak (score 1), medium (score 2) or high (score 3) and the fractions of stained cells were scored as <25% (score 1), 25–75% (score 2), >75% (score 3). The final score was calculated as the product of the intensity and the fraction scores. The following score categories were applied: Score category 0, negative (score 0); score category 1, weak expression (score 1–3); score category 2, moderate/strong expression (score 4–9). All scores were assessed in consensus by two observers, following an independent evaluation.
The mean age for the samples scored on each TMA was calculated. The Mann–Whitney U-test was used to analyse differences between patients and controls in urinary protein Western blotting experiments. The correlation of protein expression with clinical variables, age, gender, stage, grade and recurrence, was assessed using the Spearman's or Pearson's chi-square two-sided tests, when appropriate. Log-rank tests were used to analyse PFS, OS and DSS. Multivariate Cox regression was used to correct significant log-rank P values for additional clinical parameters. In the statistical analysis comparison between the following score categories were performed: (0 vs 1 vs 2), (0 vs 1, 2; negative vs positive) or (0, 1 vs 2). A P < 0.05 was considered to indicate statistical significance.
Our previous results consistently showed, by both MS and Western blotting experiments, that FGB, APOE, LRG1 and SERPINA1 proteins were more abundant in the urine of patients with bladder cancer compared with the urine from controls . POLR1E and TOP2A could not be quantified in a validated manner from the MS results. Only one peptide per protein was detected, which is not statistically reliable according to common practice. Instead, POLR1E and TOP2A were, in the present study, confirmed to be more abundant (P < 0.05) in urine of patients with bladder cancer than in the urine from controls using the complementary Western blotting technique (Fig. 1).
To validate the expression of urinary bladder cancer-marker proteins in tissue, the five proteins FGB, APOE, LRG1, POLR1E and SERPINA1 were stained with corresponding antibodies in IHC experiments on a TMA of bladder cancer samples. As many as 120 samples of three different stage groups, in total 360 samples, were included on the TMA in duplicate cores. After immunostaining and scoring of the immunoreactivity statistical analysis was applied. The total number of scores for a specific protein was, however, <360 in the statistical analyses due to occasionally missing cores. A maximum of 2.2% of the cores were missing for a specific staining. The interobserver κ-values were as follows: FGB (0.69), APOE (0.74), LRG1 (0.71), POLR1E (0.70), SERPINA1 (0.75) and were considered as good agreement (κ-value >0.6) between observers. All studied proteins were shown to be expressed in urinary bladder cancer tissue to a lesser or greater extent. For POLR1E and LRG1, the main part of the tumours in each stage (>80%) were stained, while for SERPINA1 the corresponding number was <20%. Further, the immunoreactivity for the different proteins had different characteristics (Fig. 2). The APOE, FGB, POLR1E and SERPINA1 proteins were evenly distributed throughout the cytoplasm and POLR1E was also expressed in the nuclei. Interestingly, at lower cancer stages, immunoreactivity of LRG1 was detected in a small area close to the nuclei. However, in higher cancer stages the localisation of LRG1 shifted into a more general distribution in the cytoplasm (Fig. 2C).
Next, we investigated the high expression of proteins correlated with increasing tumour stage and for APOE, FGB and POLR1E a correlation was observed (P < 0.001). Representative immunostainings are presented in Fig. 2 and the scores of the immunostaining for different stages are presented in Fig. 3. The LRG1 and SERPINA1 staining did not correlate to stage. Further, clinical parameters were investigated, and Cox regression analysis of SERPINA1 expression in Ta and T1 tumours showed a correlation (P < 0.025) with progression. Figure 4 shows that PFS was less likely in patients diagnosed with tumours of Ta or T1stages expressing SERPINA1 compared with patients that lacked SERPINA1 expression. When correcting for age, gender, grade and stage, SERPINA1 had a hazard ratio (HR) of 2.57, 95% CI 1.13–5.87 (P = 0.025). None of the other investigated proteins correlated with progression. Neither did age, gender, grade, recurrence, OS or DSS correlate to expression of any of the proteins investigated.
The IHC experiments showed that the investigated proteins associated with urinary bladder cancer due to elevated levels in urine were also expressed in bladder cancer tissue. This is an important confirmation, as the tumour then is a likely source of these proteins in urine. Today, although it has weaknesses, IHC on TMA is the high-throughput method of choice for investigating expression of proteins in tissue. First, the reliability of this method is dependent on the selection of primary antibodies. Often many antibodies against one protein are available and in the present study the antibodies chosen have been assessed by the Human Protein Atlas (HPA) to have high specificity. Second, the results are dependent on the sample selection. Even though as many as 360 different tumour samples were investigated, the findings need to be further verified in an independent tissue material to explore the potential of these proteins as well-defined biomarkers. Furthermore, the IHC scoring is subjective. Although the scoring was done by experienced personnel variations occur and it would be more reliable if the evaluation could be standardised in the future by using computerised image analysis.
The proteins have characteristic immunoreactivities within the cellular compartments and expression of some proteins also correlate to malignant phenotype and clinical parameters. Below the results of the TMA experiments are discussed, protein by protein.
The acute-phase protein APOE is a regulating part of the immune system involved in destroying tumour cells . This protein is also important for the CD1-dependent presentation of lipid antigens  and patients with bladder cancer treated with intravesical BCG vaccine show lower recurrence frequency, if the tumour cells have high expression of antigen-presenting CD1 molecules, major histocompatability complex-I, class I and chemokines . APOE has also been implicated in ovarian cancer, where cancer cells need APOE for proliferation and survival . The present study showed that high expression of APOE was associated with advanced stages of bladder tumours (Fig. 3A). In the future, a study including patients treated with BCG would be of great interest.
The expression of fibrinogen, of which the FGB is one part, is shown to increase the metastatic potential of lung cancer and melanoma cell lines . Also, endogenously synthesised fibrinogen has been shown to enhance growth of lung and prostate cancer cell lines . The FGB protein was, in the present study, confirmed as interesting also for bladder cancer, as its expression correlated with higher tumour stage (Fig. 3B). Previously, Martinez-Pineiro et al.  showed that fibrinogen degradation products could be used for monitoring recurrence in patients with bladder cancer. In the present study, the elevated levels of fibrinogen-related compounds are exclusively related to FGB . The effect of BCG treatment against bladder cancer has been shown to be enhanced if the BCG is injected s.c. together with fibrinogen . Further investigation is needed to determine if BCG treatment is more successful when treating tumours expressing FGB.
For the acute-phase protein, LRG1 immunoreactivity did not correlate with stage (Fig. 3C) or any other clinical parameter. Interestingly, we observed changed intracellular protein localisation (Fig. 2C) with stage but the implications of this must be further investigated. Low levels of LRG1 have been reported to decrease motility in neuroblastoma cells and the activity of the protein might depend on the cellular localisation. In experiments with neuroblastoma cell lines it has been shown that presence of a microRNA targeting the LRG1 mRNA reduces cell migration and invasiveness . LRG1 contains a leucin-zipper that implies that the protein is involved in protein–DNA and protein–protein interactions, and has also been related to liver cancer. In two hepatoma cell lines, one sensitive and one resistant to TGF-β1, LRG1 is co-expressed with the TGF-βI type II receptor in the sensitive cell line, making it sensitive to TGF-β1 growth inhibitory signals . However, at present LRG1 is not of further interest in bladder cancer.
The present results confirm that POLR1E is a urinary protein marker for bladder cancer. Further we describe a correlation between POLR1E immunoreactivity and an increased stage of bladder cancer (Fig. 3D). This RNA-transcription involved protein is needed by proliferating cells and it initiates transcription of rRNA from rDNA in the nucleus. The initiation occurs when POLR1E interacts with the upstream binding factor and Pol I . Moreover, the present results showed that POLR1E was expressed in high-stage bladder cancer cells and located in their nucleoli (Fig. 2D), where rRNA transcription occurs. The protein has also been reported to be located in cell junctions of e.g. the epidermoid carcinoma cell line A-431 . Compared with proliferating cells, confluent cells have 38% less expression of POLR1E . When POLR1E is blocked, specific transcription is inhibited. This effect is due to the prevention of the initiation complex formation , again showing its involvement in cell growth. In mice, POLR1E binds strongly to the Pol I complex and is a true Pol I subunit . These properties of the POLR1E protein make it a drug target candidate for treatment of bladder cancer.
In the present study, the expression of the acute-phase protein SERPINA1 was not related to tumour stage (Fig. 3E) but, more importantly, SERPINA1 expression was a potential predictor of progression in Ta and T1 tumours (Fig. 4). SERPINA1 has an anti-apoptotic effect mediated by its inhibitory effect on caspase-3  and has previously been shown to predict progression in cutaneous squamous cell carcinoma . SERPINA1 also contributes to progression towards metastatic characteristics in epithelial ovarian cancer cells . High expression of SERPINB5, another serine peptidase inhibitor, is correlated to higher risk of progression in bladder cancer . The expression of SERPINA1 in the initially resected tumour tissue thus has potential as a progression-related biomarker for determining if radical surgery would be beneficial or not.
High expression of TOP2A has previously been shown to correlate with recurrence and increased risk of death , and with late-stage bladder cancer tumours . In the present study, we confirmed that TOP2A was a protein with higher abundance in urine from patients with bladder cancer compared with controls. TOP2A is therefore classified as a urinary protein marker candidate of bladder cancer.
In conclusion, the present study showed that five selected urinary proteins, with higher abundance in samples from patients with bladder cancer compared with controls, also were expressed in bladder tumours investigated in a Swedish cohort. All proteins had their major immunoreactivity in the cytoplasm, but for POLR1E nucleolic immunoreactivity and for LRG1 a possible immunoreactivity of the Golgi apparatus was observed. The tissue expression for APOE, FGB and POLR1E correlated with higher tumour stage (P < 0.001) and could be helpful in the future to characterise and diagnose tumours. Interestingly, immunoreactivity of SERPINA1 in Ta and T1 tumours correlated with PFS, opening up the possibility for a new prognostic marker.
Funding was obtained from Swedish Cancer Society (P-U.M., S.B.L.), European Union grant to UROMOL consortium no. 201663, The Kjell and Märta Beijer Foundation, The Foundation in Memory of Johanna Hagstrand and Sigfrid Linners (Stiftelsen Johanna Hagstrand och Sigfrid Linnérs Minne), P.O. Zetterling Foundation and Lions Cancer Foundation (Uppsala). The HPA project is accessible at http://www.proteinatlas.org and is funded by the Knut & Alice Wallenberg Foundation (Stockholm, Sweden). The Atlas is part of the Human Proteome Organisation (HUPO) Human Antibody Initiative (Montreal, Quebec, Canada).
The authors wish to thank Lisa Wernroth for the input on statistical analyses and Evelina Sjöstedt for assistance in immunoreactivity scoring.