The last 2 authors contributed equally to this article.
A multiplexed, particle-based flow cytometric assay identified plasma matrix metalloproteinase-7 to be associated with cancer-related death among patients with bladder cancer
Version of Record online: 19 AUG 2010
Copyright © 2010 American Cancer Society
Volume 116, Issue 19, pages 4513–4519, 1 October 2010
How to Cite
Svatek, R. S., Shah, J. B., Xing, J., Chang, D., Lin, J., McConkey, D. J., Wu, X. and Dinney, C. P. (2010), A multiplexed, particle-based flow cytometric assay identified plasma matrix metalloproteinase-7 to be associated with cancer-related death among patients with bladder cancer. Cancer, 116: 4513–4519. doi: 10.1002/cncr.25401
- Issue online: 22 JUN 2010
- Version of Record online: 19 AUG 2010
- Manuscript Accepted: 23 MAR 2010
- Manuscript Revised: 26 FEB 2010
- Manuscript Received: 15 DEC 2009
- matrix metalloproteinases;
- bladder cancer;
The current study was conducted to demonstrate the utility of a multiplexed, particle-based flow cytometric assay for the simultaneous analysis of a panel of matrix metalloproteinases (MMPs) using small volumes of plasma samples from patients with bladder cancer. In addition, the authors attempted to test the hypothesis that plasma levels of MMPs are associated with time to cancer-related death.
Plasma MMP concentrations (MMP-1, -2, -3, -7, -8, -9, and -12) in 135 patients presenting with high-grade ≥T1 bladder cancer were measured. Data regarding clinical and pathologic features was ascertained in a retrospective fashion.
The median duration of follow-up was 30.4 months. At the time of analysis, 61 patients had died, including 45 (33.3%) who died of bladder cancer. Plasma MMP-12 was not measurable. For all other MMPs, the intra-assay coefficient of variation varied from 6.12% to 9.82%. MMP-1, -2, -3, -8, and -9 were not found to be significantly associated with time to cancer-related death. Plasma MMP-7 levels were significantly associated with time to cancer-related death after adjustment for competing clinical and pathologic features (hazard ratio [HR], 2.2; 95% confidence interval [95% CI], 1.1-4.5 [P = .022]). The 5-year median cancer-specific survival rates for those patients with MMP-7 levels above and below the median value (300 pg/mL) were 73.6% (95% CI, 60.0-83.2%) and 48.0% (95% CI, 32.5-61.9%), respectively (P = .01).
Multiplexed, particle-based flow cytometric assay allows for the high-throughput measurement of multiple plasma or serum proteins simultaneously. By using this new technology in a cohort of patients with bladder cancer, plasma levels of MMP-7 were identified as being significantly associated with time to cancer-related death Cancer 2010. © 2010 American Cancer Society.
Urothelial carcinoma of the bladder (UCB) is the fourth most common cancer in men and the fifth most common overall, with an estimated 68,800 new cases of bladder cancer and 14,100 deaths from the disease expected in the United States in 2008.1 Although the majority of patients with invasive bladder cancer present without radiographic or clinical evidence of disease beyond the bladder, eventual metastasis is reported in up to 30% to 56% of patients, most often the result of occult metastasis not detected by current staging modalities.2
Prognostic tools based on standard clinical and pathologic variables can quantify the risk of death from bladder cancer, but their accuracy is imperfect because of the heterogeneous biologic behavior of tumors and/or patients with similar clinical and pathologic features.3-5 The use of biomarkers specific to the tumor and/or patient can provide prognostic utility over that available from routine clinical features.6-8 Previously, the identification of other potentially informative biomarkers to routine clinical and pathologic features has been limited to evaluating 1 or 2 markers at a time. However, recent advances in high-throughput technology allow for the rapid evaluation of multiple potential biomarkers simultaneously.
Matrix metalloproteinases (MMPs) are a family of zinc-dependent proteinases that are involved in the breakdown and remodeling of the extracellular matrix. They are overexpressed in many cancer tissues and are believed to play an important role in cancer invasion and metastasis in a broad range of tumor types.9-11 In addition, MMPs have been associated with stage and grade in bladder cancer.11-15
In the current study, we sought to demonstrate the utility of a multiplexed, particle-based flow cytometric assay for the simultaneous analysis of a panel of MMPs using small volumes of plasma samples from patients diagnosed with bladder cancer. In addition, we attempted to test the hypothesis that plasma levels of MMPs are associated with time to death from bladder cancer.
MATERIALS AND METHODS
All studies were performed after approval by the local Institutional Review Board. Patients diagnosed with transitional cell UCB were recruited from The University of Texas M. D. Anderson Cancer Center between March 2001 and August 2005. Informed consent was obtained from each patient at the time of plasma withdrawal. Clinical and pathologic characteristics for each patient were collected from our bladder cancer database, which is maintained prospectively. Chart review was used to collect data missing from the database and to determine the patient's status at the time of censor. This was verified from the date of the last follow-up visit, physician correspondence, death record, and/or patient correspondence letters. When patients died, the cause of death was determined by the treating physicians, ascertained by chart review corroborated by death certificates, or by death certificates alone. Most patients who were identified as having died of bladder cancer had progressive, widely disseminated, and often highly symptomatic metastases at the time of death.
The study was comprised of 135 patients diagnosed with ≥T1 UCB, including 107 patients who underwent radical cystectomy. The indications for radical cystectomy were tumor invasion into the muscularis propria or T1 disease on repeat transurethral resection. No patients had evidence of disease metastasis at the time of presentation. All histologic slides were reviewed without knowledge of plasma MMP data. Clinical tumor stage was determined by the highest tumor stage noted on the transurethral resection specimen before cystectomy or assigned by the operative surgeon according to the 2002 American Joint Committee on Cancer (AJCC) TNM classification system.16 Pathologic stage was assigned according to the 2002 AJCC TNM staging system. Pathologic grade was classified according to the 1998 World Health Organization/International Society of Urological Pathology classification system.17
Plasma MMP Measurements
After the patient interview was conducted and the patients provided informed consent, a 40-mL blood sample was drawn into coded heparinized tubes for laboratory analyses. Blood samples were collected into tubes, coded for deidentification, transported immediately to the laboratory, and stored at −4°C. The blood was centrifuged at 500 × gravity for 10 minutes to separate plasma. The plasma was aliquoted and stored at −80°C until batch analysis.
Plasma MMP concentrations (MMP-1, -2, -3, -7, -8, -9, and -12) were measured using a Fluorokine Multianalyte Profiling kit (R&D Systems, Minneapolis, MN) and a Luminex Bioanalyzer (Luminex Corporation, Austin, TX). Measurement of plasma MMP levels were performed according to the package insert of the kit. The Fluorokine Multianalyte Profiling kits contained groups of microspheres. A small volume (15 μL) of plasma was required to run the assay in duplicate. Plasma samples were incubated with antibody-coated microspheres, which bind to specific MMPs in the plasma. After washing, the samples were then incubated with biotinylated antibodies, which bind to the MMPs present on the microsphere-MMP complex. Next, these complexes were incubated with streptavidin-phycoerthrin, which attached to the biotinylated MMP antibodies present on the microspheres. The Luminex Bioanalyzer then quantified the phycoerythrin fluorescence present for each individual microsphere group. The median fluorescence intensity was used to calculate the estimated concentration. Every sample was run in duplicate and the mean value was used for data analyses. Unique microsphere groups were reported by the manufacturer to exhibit <0.5% cross-reactivity and interference with other MMPs analyzed.
The association between categorical variables was assessed with the Pearson chi-square test. Cox proportional hazards models were used to evaluate the relation between plasma MMP levels and time to death from bladder cancer.18 Histogram plots and the Shapiro-Wilk test for normality indicated that the MMP levels were not normally distributed and therefore they were analyzed as dichotomous variables on the basis of the 50% distribution. We report hazard ratios (HRs) for potential prognostic factors with 95% confidence intervals (95% CIs). The accuracy of multivariable models was tested using the Harrell concordance index, which approximates the area under the curve in censored data.19 Backwards stepwise regression was used to reduce the multivariable model. This eliminated the least statistically significant variables 1 by 1 from the full model until the only variables that remained were statistically significant at an α level of 10%. The Kaplan-Meier method20 was used to estimate the median cancer-specific survival. Survival estimates between groups were compared using the log-rank test statistic. Significance for interaction was tested using the likelihood ratio test. All statistical tests were 2-sided with a type I error rate of 5%. Statistical analyses were performed using STATA software (Version 10.1; StataCorp LP, College Station, TX).
Descriptive characteristics of the study cohort are shown in Table 1. Plasma samples from 135 patients with bladder cancer were analyzed. The median follow-up time was 30.4 months (interquartile range, 15.3-50.0 months). A total of 61 (45.2%) patients died at the time of data acquisition, including 45 (33.3%) patients who died of bladder cancer. The 5-year survival rate for the entire cohort was 51.2% (95% CI, 41.4-60.1%). All patients were found to have high-grade cancer on transurethral resection. Of the 107 patients who underwent radical cystectomy, 44 (41.1%) received perioperative chemotherapy, including 32 (29.9%) patients before surgery, 9 (8.4%) patients after surgery, and 3 (2.8%) patients both before and after surgery.
Descriptive characteristics of the plasma MMPs are shown in Table 2. Plasma MMP-12 levels were not measureable. The intra-assay coefficient of variation for MMPs -1, -2, -3, -7, -8, and -9 ranged from 6.12% to 9.82% (Table 2). The association between plasma MMP levels and clinical stage of disease is shown in Table 3. Higher clinical stage was found to be significantly associated with lower levels of plasma MMP-2 (P = .026) and higher levels of plasma MMP-7 (P = .022). The univariable association between plasma MMP levels and cancer-related mortality is shown in Table 4. Plasma MMP-1, -2, -3, -8, and -9 levels were not found to be significantly associated with bladder cancer-related mortality. However, the level of plasma MMP-7 was found to be significantly associated with bladder cancer-related death on univariable analysis. The 5-year median cancer-specific survival rates for those patients with MMP-7 levels above and below the median value (300 pg/mL) were 73.6% (95% CI, 60.0-83.2%) and 48.0% (95% CI, 32.5-61.9%), respectively (P = .01) (Fig. 1).
|Mean age at diagnosis (range), y||64 (38-88)|
|Radical cystectomy (%)|
|Clinical T classification|
|Neoadjuvant and adjuvant||3 (2.2)|
|All-cause mortality||61 (45.2)|
|Cancer-specific mortality||45 (33.3)|
|Analyate||Concentration (in pg/mL) Cohort Data (n=135)|
|Clinical T Classification||Median Plasma MMP Levels (IQR)|
|≤T2||195 (87-462)||20,772 (16,890-27,701)||2120 (1375-3294)||283 (188-592)||2813 (1524-5989)||4915 (2379-8118)|
|>T2||175 (393-479)||15,696 (18,680-19,863)||2120 (1325-2751)||1045 (353-1419)||4417 (1896-7098)||6741 (3619-9984)|
|Univariable||Multivariable (Base)||Multivariable (Reduced)a|
|HR||95% CI||P||HR||95% CI||P||HR||95% CI||P|
|Base model predictors||Age||1.02||0.99-1.05||.201||1.02||0.99-1.06||.142||—||—||—|
|Clinical T classification|
|Univariable||Multivariable (Base + 1 MMP)|
|Plasma level of MMPs||MMP-1|
To determine whether plasma MMP levels were associated with bladder cancer-related mortality after adjustment for known clinical features, we performed multivariable analyses. Initially, we fitted a base model that was comprised of routine clinical features available at the time of presentation. Of these clinical features, male gender (P = .012), advanced clinical T classification (P = .004), and absence of radical cystectomy (P < .001) were found to be independently associated with cancer-related mortality (Table 4). The predictive accuracy of this model was 70.3%. Of the individual variables, the presence of radical cystectomy represented the most informative multivariable prognostic factor, as evidenced by the decrease in predictive accuracy noted when the variable was removed from the multivariable model (61.8%).
In the second step, we fitted multivariable models that were comprised of all variables from the base model and 1 of the plasma MMPs (Table 4). Only plasma MMP-7 achieved independent predictor status. The combined predictive accuracy of the base model with MMP-7 was 72.5%. The accuracy of this full model significantly exceeded that of the base model (+2.2%; P < .001).
We repeated our analysis after including only those patients with clinical T2 and higher disease (n = 126). In this analysis, plasma MMP-7 levels above the median value were also found to be associated with bladder cancer-related mortality on univariable analysis (HR, 2.2; 95% CI, 1.2-4.0) and after adjustment for routine clinical features (HR, 2.1; 95% CI, 1.0-4.3). In addition, the addition of MMP-7 provided a significant improvement in predictive accuracy compared with the base model (73.2% vs 71.3%, respectively; P = .040). The 5-year median cancer-specific survival rates for those patients with MMP-7 levels above and below the median value were 73.1% (95% CI, 58.8-83.1%) and 47.4% (95% CI, 31.5-61.7%), respectively (P = .01).
By using a multiplexed, particle-based flow cytometric assay, we demonstrated that a panel of MMPs can be reliably measured with small volumes of plasma samples from patients diagnosed with bladder cancer. In addition, we tested the ability of several plasma MMP levels to improve prognostication for patients with bladder cancer. We found that plasma levels of MMP-7 significantly improved our prediction of time to death from bladder cancer after adjusting for the effects of routine clinical features. Predictions based on the full model that included the base model plus the plasma MMP-7 level were found to be significantly more accurate than predictions based on the base model alone.
Previous studies have supported an association between MMPs and the presence and stage of bladder cancer. MMP-2 and MMP-9 levels have been found to be associated with bladder cancer stage and grade in previously published small series.21-23 By using enzyme-linked immunoadsorbent assays (ELISAs), Guan et al21 noted that serum MMP-9 levels were significantly higher in 52 patients with bladder cancer compared with 32 healthy controls, and they also observed MMP-9 levels to be significantly associated with tumor stage and grade. Gerhards et al observed urinary levels of MMP-2 and MMP-9 to be highly correlated with tumor stage and grade.23 By using ELISAs, Staack et al24 observed median MMP-2 plasma levels to be elevated in patients with bladder cancer compared with controls. In addition, they found that the combination of MMP-9 and a tissue inhibitor of MMP-1 (TIMP-1) improved the sensitivity of MMP-2 for detecting bladder cancer.24
We identified plasma levels of MMP-7 as being significantly associated with time to death from bladder cancer in our cohort. Furthermore, this association was observed after correction for competing clinical predictors. Patients with a plasma MMP-7 level above the median value of 300 pg/mL experienced a 148% increase in their hazard of death from bladder cancer compared with those patients with an MMP-7 level below the median value. At 5 years, the median survival estimate was approximately 25% higher for those patients with a plasma MMP-7 level above the median value compared with those with a plasma MMP-7 level below the median value.
The majority of MMPs are believed to be produced predominately by stromal cells such as fibroblasts, endothelial cells, and macrogphages.25 However, to the best of our knowledge, MMP-7 is 1 of the few MMPs that is produced directly by tumor cells, making it an attractive biomarker for identifying an aggressive phenotype.25 MMP-7 is 1 of the smallest MMPs and it lacks a C-terminal hemopexin domain common to other MMP members.25 Its expression is restricted to carcinoma cells and its levels have been associated with lymph node metastasis in patients with endometrial cancer.11, 26 MMP-7 levels have also been associated with gastric and colorectal cancer and its immunoreactivity has been correlated with vascular invasion and metastasis in these tumors.27, 28 It is frequently overexpressed in cancer tissue and there is considerable evidence that the expression of MMP-7 is associated with advanced cancer stage and prognosis.25
The assay used in the current study has several distinct advantages over conventional ELISA assays to measure plasma proteins. It is a multiplex system that allows for the simultaneous measurement of multiple proteins in a single plasma sample. As a result, the estimates are less influenced by operator error as observed with ELISAs performed on multiple analytes. The assay can measure proteins with a small volume (only 15 μL) of plasma or serum, which is generally much smaller than that required for 1 ELISA assay. In addition, the fluorescence intensity measurement provides a more direct and sensitive readout than the colorimetric methods used for ELISAs.29
The observation of a significant association between male gender and an increased risk of cancer-related death on multivariable analysis was an unexpected finding. Female gender was shown to portend a higher risk for cancer-related mortality in a large multicentered cohort study.30 We speculate that our unique observation may be related to the small sample size and limited number of events in the current study. To this end, we were unable to perform multivariable analysis in only those patients who underwent radical cystectomy. Moreover, we acknowledge the possibility that the lack of an association between MMPs and outcome was a result of insufficient sample size. We also recognize additional limitations to this study. The current study did not measure plasma levels of MMPs in patients without bladder cancer. Certainly, consideration of the levels of MMPs in age-matched controls would strengthen this analysis. It is possible that other factors will be demonstrated to be independent prognostic factors in a larger patient cohort and additional validation studies are needed. Finally, the assessment of multiple biomarkers increases the possibility of a false-positive finding and we did not perform statistical adjustment for the P value based on the examination of several MMPs. Nevertheless, this does not repudiate the significant association observed between MMP-7 and time to cancer-specific mortality.
Despite these limitations, the results of the current study demonstrate the utility and feasibility of a novel multiplex assay to measure plasma MMP levels. In addition, our observations indicate that MMP-7 may be an important prognostic biomarker for patients with bladder cancer, a finding that warrants further investigation. The substantial number of patients noted to develop disease recurrence after radical cystectomy indicates that improved risk stratification is needed in patients with invasive bladder cancer. The identification of patients at highest risk for metastasis and death from disease is important for tailoring treatment and selective clinical trial enrollment. If externally validated as a prognostic marker, the clinical utility of MMP-7 can be assessed in a properly conducted prospective trial. Current prognostic models for patients with bladder cancer incorporate important clinical and pathologic features but fail to capture the biologic heterogeneity of the individual.4, 31 In addition, new therapeutic approaches will be increasingly dependent on a more reliable set of markers to serve as prognosticators, targets, and/or surrogate endpoints of disease progression and response to therapy. Indeed, we are approaching a time when biomarkers will be used routinely to assist with detection, monitoring, and/or managing disease progression. Moreover, because of feasibility, the most attractive biomarkers are those that can be assayed from blood and, as a result, there is much interest in profiling serum and/or plasma proteins.
Multiplexed, particle-based flow cytometric assay allows for the high-throughput measurement of multiple plasma or serum proteins simultaneously. By using this new technology in a cohort of patients with invasive bladder cancer, we identified elevated plasma levels of MMP-7 as being significantly associated with time to cancer-related death.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 16Urinary bladder. In: GreeneFL, PageDL, FlemingID, et al, eds. AJCC Cancer Staging Manual. 6th ed. New York: Springer-Verlag; 2002: 301-323..
- 18Regression models and life tables (with discussion). J R Stat Soc B. 1972; 34: 187-220..
- 30Evidence-based gender related outcomes after radical cystectomy: results of a large multicenter study. Am Urol Assoc. 2009., , , et al.