Combined c-Myc and caveolin-1 expression in human prostate carcinoma predicts prostate carcinoma progression

Authors


Abstract

BACKGROUND

Over-expression of the oncogene c-Myc has been implicated in the development and progression of human prostate carcinoma. However, previous assessments of c-Myc expression have not revealed its potential for predicting prostate carcinoma progression. Caveolin-1 is associated with prostate carcinoma progression and is a downstream target gene of c-Myc. The observation that caveolin-1 can suppress c-Myc-induced apoptosis suggested the potential for cooperation between c-Myc and caveolin-1 in malignant progression. In this study, the authors evaluated the prognostic potential of combined c-Myc and caveolin-1 expression in human prostate carcinoma progression.

METHODS

Immunostaining with c-Myc and caveolin-1-specific antibodies was performed on paraffin sections from 104 radical prostatectomy specimens from men with lymph node negative prostate carcinoma. Combined c-Myc and caveolin-1 immunostaining scores were related with the clinical and pathologic features and the probability of prostate-specific antigen recurrence after surgery.

RESULTS

The combination of c-Myc and caveolin-1 immunopositivity correlated positively with Gleason score (ρ = 0.219; P = 0.0253) and positive surgical margin (ρ = 0.333; P = 0.0006). The combination of positive c-Myc and caveolin-1 in patients with clinically confined prostate carcinoma was a significant prognostic marker for the time to disease progression after surgery in both univariate analysis (P = 0.0039; hazard ratio, 3.035) and multivariate analysis (P = 0.0114; hazard ratio, 2.916).

CONCLUSIONS

The coexpression of c-Myc and caveolin-1 showed potential as a useful prognostic marker for human prostate carcinoma. The current results suggest interactions between c-Myc and caveolin-1 in the progression of human prostate carcinoma. Cancer 2005. © 2005 American Cancer Society.

Although it has been shown that several preoperative clinical and pathologic indices, including prostate-specific antigen (PSA) levels and the grade of the carcinoma obtained from needle biopsy, are important predictive indices of pathologic and clinical outcomes in patients with prostate carcinoma,1, 2 the predictive power of these indices is diminished greatly for those patients with malignancies that fall in the middle range of a given index. The identification and use of molecular biomarkers to supplement clinical information concerning the biologic aggressiveness of a tumor may allow for better differentiation between the patients who would benefit from aggressive treatment and those with more indolent disease and may allow for better selection among different treatment regimens.

The oncogene c-Myc plays a critical role in the development of malignancies in a variety of human organs.3 In the prostate, there is compelling evidence that deregulation of c-myc gene is involved in tumorigenesis. c-myc gene amplification was identified in 30–50% of prostate carcinoma specimens.4–7 c-Myc over-expression at both the mRNA and protein levels are found in as high as 70% of prostate carcinomas, as demonstrated by northern blot analysis, in situ hybridization, and immunohistochemistry.8, 9 It was documented previously that c-myc over-expression produced hyperplastic lesions in vivo in studies using the mouse prostate reconstitution model.10 More recent findings that transgenic expression of human c-myc in mouse prostate leads to lesions similar to human high-grade prostatic intraepithelial neoplasia (PIN) and prostate carcinoma11 provide categorical evidence that c-myc is a critical gene in prostate carcinoma development.

In addition to tumorigenesis, it also has been suggested that c-myc deregulation contributes to the progression of prostate carcinoma.12, 13 c-myc over-expression tends to be associated with tumors that are poorly differentiated,5, 9 have developed lymph node metastases,6, 13 have poor clinical outcomes,13 and have become androgen-independent after hormone treatment.6, 7, 14 In some tumors, it has been found that c-myc over-expression has prognostic value,15–18 whereas, in other tumors, no association between c-myc over-expression and prognosis was observed.19 In human prostate carcinoma, c-myc over-expression shown by fluorescent in situ hybridization (FISH) is associated with both systemic progression and patient death in univariate analysis.13 However, it is our understanding that there are no reports demonstrating that c-myc is an independent prognostic marker for prostate carcinoma.

A unique consideration for the use of c-myc over-expression as a prognostic marker for patients with carcinoma relates to the dual biologic functions of the c-myc gene. It is well known that cell proliferation and tumor growth promoted by c-myc over-expression is antagonized by myc-induced apoptosis; and suppression of this apoptotic pathway in c-myc over-expressing cells promotes disease progression.20 In some c-myc transgenic mouse models, expression of specific growth factors may counteract the apoptotic activities promoted by c-myc over-expression that lead to the development of prostate carcinoma.11, 21 However, in other c-myc transgenic mouse models, hyperplastic lesions do not progress to malignancy.21 These apparent inconsistencies result in part from the levels of c-myc expression generated by the specific transgene and/or the genetic background of the mouse.11, 21

Caveolin-1 is a major structural component of caveolae, specialized plasma membrane invaginations that are involved in diverse cell activities, such as molecular transport, cell adhesion, and signal transduction, in a cell-specific and context-specific fashion.22 It has been found that caveolin-1 is over-expressed in both mouse and human metastatic prostate carcinoma cells.23 Positive correlations between caveolin-1 over-expression and disease progression/metastasis have been found in human prostate carcinomas24–26 and in a number of malignancies, as reviewed recently.27 The metastasis-promoting activity of caveolin-1 likely is related in part to its antiapoptotic function(s) in tumor cells.28, 29 It has been shown that caveolin-1 mediates the inhibition of PP1 and PP2A, which, in turn, can lead to Akt activation and prosurvival effects in cultured prostate carcinoma cells.30 Over-expression of caveolin-1 also increases translocation of phosphorylated androgen receptor to the nucleus significantly30 and potentiates the prosurvival activity of androgen.31 Prostate carcinoma cells with high levels of caveolin-1 are more resistant to castration-induced apoptosis.32

Caveolin-1 may collaborate with c-Myc in facilitating tumor progression. In our previous work, we showed that human prostate carcinoma cell lines with increased caveolin-1 expression are more resistant to myc-induced apoptosis and have increased capacity for growth in soft agar.33 Thus, it is reasonable to postulate that an interaction of c-Myc and caveolin-1 may potentiate tumor progression and that an evaluation of the combined status of c-Myc and caveolin-1 may be informative of prostate carcinoma progression. In the current study, we assessed the prognostic value of combined c-Myc and caveolin-1 expression for predicting carcinoma progression. In this retrospective study, 104 prostate carcinoma specimens that were obtained at radical prostatectomy from lymph node-negative tumors were immunostained with c-Myc-specific and caveolin-1-specific antibodies. Coexpression of c-Myc and caveolin-1 was evaluated in association with preoperative clinical information and with postoperative pathologic diagnosis. Our results indicate that a coexpression of c-Myc and caveolin-1 is a novel, independent prognostic marker for disease progression in patients with prostate carcinoma.

MATERIALS AND METHODS

Patients and Prostate Specimens

In this study, 104 prostate carcinoma radical prostatectomy specimens were obtained from the Baylor Prostate Cancer Specialized Program of Research Excellence (SPORE) tissue bank.34 The specimens were from a series that was selected previously in a consecutive manner by an independent statistician, as described elsewhere,24 and were from patients with prostate carcinoma that had been staged as clinically confined and that was determined to be lymph node negative at surgery. Radical prostatectomy specimens were sliced into 5-mm-thick tissue blocks according a procedure described previously.24 The tissue blocks were then fixed in 10% neutral buffered formalin and embedded in paraffin. From each tissue block, hematoxylin and eosin-stained sections were made for pathologic evaluation. For each patient, 1 tissue block that was identified as containing the index carcinoma was cut into 6-μm sections for immunostaining. The index carcinoma, which was defined as the largest tumor focus and/or the most invasive and/or highest Gleason score, was identified and mapped on wholemount sections. In rare instances, the Gleason score of the index carcinoma focus was lower than other foci that were judged as having a reduced potential for clinical impact. The Gleason score of the index carcinoma was used as the overall surgical Gleason score of the prostatectomy and was used in the analysis correlating tumor histologic differentiation with protein expression. The clinical and pathologic data describing this patient set are presented in Table 1. None of the patients received any disease-specific treatment before the surgery. Preoperatively, the patients had their disease staged according to the TNM staging system.35 Postoperative pathologic findings; extraprostatic extension, seminal vesicle invasion, surgical margin, and Gleason score were recorded by a single pathologist (T.M.W.).

Table 1. Clinicopathologic Characteristics of Patients with Lymph Node-Negative Prostate Carcinoma
VariableNo. of patients (%)Mean (range)
  1. PSA: prostate-specific antigen.

Age (yrs)104 (100)64.2 (49.6–78.5)
Preoperative PSA (ng/mL)88 (84.6)12.4 (0.6–88.3)
Tumor classification  
 T1a–c18/99 (18.2) 
 T2a–c74/99 (74.7) 
 T3a7/99 (7.1) 
Gleason grade  
 < 742 (40.4) 
 749 (47.1) 
 > 713 (12.5) 
Positive margins23 (22) 
Extraprostatic extension60 (57.7) 
Seminal vesicle invasion17 (16.3) 
Biochemical recurrence40 (38.5) 
Follow-up (mos)104 (100.0)62.7 (0.5–144.9)

Immunohistochemistry

For c-myc immunostaining, the sections were deparaffinized and dehydrated. They were heated for 20 minutes in a 750-Watt microwave oven in Dako buffer, pH 8.0, for antigen retrieval. Endogenous peroxidase was inactivated with 2% H2O2 for 10 minutes. The sections were then blocked in 3% normal horse serum in 0.2 M phosphate buffered saline (PBS), pH 7.4, followed by incubation with a monoclonal antibody to c-Myc (NCL-cMYC, clone 9E11; Novocastra Laboratories Ltd., Newcastle, United Kingdom). In immunostaining protocols, this antibody detects both cytoplasmic and nuclear c-Myc and has been used successfully in formalin fixed paraffin sections of human carcinoma.16, 36–42 The c-Myc antibody was used at a dilution of 1:100 in PBS with 0.5% normal horse serum. The sections were incubated in the primary antibody for 2 hours at room temperature. They were then processed following a standard avidin-biotin complex (ABC) immunostaining procedure with an ABC kit (Vector Laboratories, Burlingame, CA). Immunoreaction products were visualized in a 3,3′-diaminobenzidine/H2O2 solution. To verify the specificity of the immunoreactions, some sections were incubated with normal mouse serum or with PBS replacing primary antibody. For immunostaining with caveolin-1 antibody, a previously described procedure was followed.24

All immunostained sections were evaluated at a power of × 200 under a Zeiss microscope. All specimens were evaluated without any knowledge of the patients' clinical information. For each specimen, the entire index carcinoma area was scanned, and an average of 22 ± 14 (mean ± standard deviation) microscopic fields were evaluated. Cytoplasmic and nuclear c-Myc staining was scored according to the percentage of tumor cells labeled. Expression of caveolin-1 also was scored as the percentage of tumor cells labeled by the antibody, as opposed to setting a cut-off labeling rate.24 For univariate and multivariate analyses, specimens were considered positive for caveolin-1 or c-Myc if they demonstrated any positively stained tumor cells.

Statistical Analysis

The correlation of c-Myc immunoreactivity alone or combined c-Myc/caveolin-1 status with patients' clinical and pathologic variables was analyzed by the Spearman rank-correlation test. The predictive value of c-Myc and caveolin-1 alone or of c-Myc plus caveolin-1 were tested univariately using a Kaplan–Meier actuarial analysis43 and the log-rank test. In addition, univariate analyses of c-Myc and caveolin-1 alone or of c-Myc plus caveolin-1 immunostaining and multivariate analyses with other clinical and pathologic variables, such as clinical stage, seminal vesicle involvement, extraprostatic extension, positive surgical margins, and Gleason score, were performed using the Cox proportional hazards regression model.44 The hazard ratio (HR) and 95% confidence interval were recorded for each marker. P values < 0.05 were considered statistically significant in all of the analyses. All analyses were performed with the SPSS 11.0 statistical software package (SPSS Inc., Chicago, IL).

RESULTS

c-Myc and Caveolin-1 Expression in Human Prostate Carcinoma

Figure 1 shows that normal glandular epithelium (Fig. 1A) and histologically normal glandular epithelium adjacent to tumor cells (Fig. 1B) showed little c-Myc immunoreactivity, whereas high c-Myc immunoreactivity was present in prostate carcinoma specimens. Some high-grade PIN lesions also exhibited high c-Myc immunoreactivity (Fig. 1C). The c-Myc immunoreactivity was localized predominantly to the neoplastic epithelial cells, but some stromal cells surrounding tumor cells also were labeled weakly. Among the 104 specimens, 53 (51%) demonstrated detectable c-Myc tumor cell staining (Table 2). In these c-Myc-positive tumors, the proportion of positively labeled tumor cells varied greatly among individual specimens, with an average labeling index of 38% (range, 3–95%). Within the same specimen in which there was more than one Gleason pattern, the c-Myc immunoreactivity tended to be present in the region with the higher Gleason grade, i.e., the more poorly differentiated tumor. At the cellular level, the c-Myc immunoreactivity was localized mostly to the cytoplasm of the tumor cells with scattered, positively labeled nuclei also apparent, especially in poorly differentiated tumors (Fig. 1D).

Figure 1.

c-Myc expression pattern in normal prostate tissue (A), prostate carcinoma (B), and prostate carcinoma adjacent to a prostatic intraepithelial neoplasia lesion (C). c-Myc reactivity was seen most commonly in the cytoplasm of tumor cells, but a small fraction of tumor cells also exhibited nuclear labeling (D). Original magnification × 200 (A); × 400 (B–D).

Table 2. c-Myc and Caveolin-1 Expression in Patients with Clinically Confined Prostate Carcinoma
Caveolin-1 expressionc-Myc expression: No. of patients (%)
Totalc-Myc negativec-Myc positive
Total104 (100)51 (49)53 (51)
Caveolin-1 positive21 (20)7 (7)14 (13)
Caveolin-1 negative83 (80)44 (42)39 (38)

Among the tumors examined, 21 tumors were caveolin-1 immunopositive. The labeling index varied from 4% to 85% positive, with a mean of 19%. These results were consistent with previous reports.24, 26 Among the caveolin-1 positive tumors, 67% (14 of 21 tumors) were also c-Myc-positive. This ratio was higher compared with the caveolin-1-negative tumors, in which only 53% (44 of 83 tumors) were c-Myc-positive. Indeed, there was a significant positive correlation (ρ = 0. 215; P = 0.0283) between the c-Myc labeling index and the caveolin-1 labeling index in tumor cells among these specimens (Fig. 2).

Figure 2.

Bivariate scatterplot showing the correlation between c-Myc expression and caveolin-1 expression in prostate carcinoma expressed as labeling indices (% positive).

Correlation of Combined c-Myc and Caveolin-1 Expression with Clinical/Pathologic Features

The association of combined c-Myc and caveolin-1 expression with clinical and pathologic variables is summarized in Table 3. There was a positive correlation between combined c-Myc and caveolin-1 immunopositivity and Gleason score (ρ = 0.219; P = 0.0253). The frequency of combined c-Myc and caveolin-1 expression was higher in the tumors that had positive surgical margins compared with tumors that had negative surgical margins (ρ = 0.333; P = 0.0006). These correlational analyses suggest that patients who have prostate carcinoma that is positive for both c-Myc and caveolin-1 are more likely to have poorly differentiated and more aggressive tumors.

Table 3. Correlation of Combined c-Myc and Caveolin-1 Positivity with Clinical and Pathologic Parameters in Patients with Clinically Confined Prostate Carcinoma
c-Myc and caveolin-1 positiveNo. of patientsRhoaP value
  • UICC: International Union Against Cancer; PSA: prostate-specific antigen.

  • a

    Spearman correlation coefficient.

Gleason score1040.2190.0253
Positive surgical margins1040.3330.0006
Seminal vesicle involvement1040.0540.5847
Extraprostatic extension1040.1100.2678
Clinical stage (UICC)990.0820.4181
Preoperative PSA (ng/mL)880.2390.0250

Combined c-Myc and Caveolin-1 Expression Predicts Biochemical Recurrence

We evaluated the ability of the combined expression of c-Myc and caveolin-1 to predict biochemical recurrence in patients with lymph node-negative prostate carcinoma. Among the 104 patients who had clinically confined prostate carcinoma, 40 patients had a biochemical recurrence, as defined by a serum PSA level ≥ 0.4 ng/mL on 2 successive measurements after surgery. The follow-up for the 104 patients ranged from 0.5 months to 144 months with a median of 67.5 months. The actuarial probability of remaining free of disease progression for these patients during follow-up after the surgery was calculated by using the Kaplan–Meier method. The patients who had tumors that were positive for c-Myc had a significantly worse prognosis, as demonstrated on the Kaplan–Meier plot (Fig. 3A). Consistent with our previous study,24 patients with caveolin-1-positive tumors (Fig. 3B) also demonstrated a shorter recurrence-free survival. It is noteworthy that patients who had tumors that were positive for both c-Myc and caveolin-1 expression also had a worse prognosis (Fig. 3C).

Figure 3.

Recurrence-free probability for c-Myc-positive (A), caveolin-1 (Cav-1)-positive (B), and c-Myc-positive and caveolin-1 positive (C) patients in the cohort of 104 men with clinically confined, lymph node-negative prostate carcinoma.

The prognostic value of combined c-Myc and caveolin-1 was analyzed further using the Cox proportional hazard regression models. Univariate analyses showed that combined c-Myc and caveolin-1 expression is a significant prognostic predictor of a shorter time to progression after surgery (HR, 3.0; P = 0.0039) (Table 4). Univariate analysis also indicated that c-Myc alone (HR, 2.7; P = 0.0032) or caveolin-1 alone (HR, 2.2; P = 0.0220) offered prognostic significance. These results are consistent with our previous study,24 which demonstrated that positive caveolin-1 status was associated with a shorter recurrence-free survival. Higher HRs for the combined c-Myc and caveolin-1 expression than for either c-Myc alone or caveolin-1 alone suggest that it is better at distinguishing between high-risk and low-risk patients.

Table 4. Univariate and Multivariate Cox Proportional Hazards Model for Time to Recurrencea
Prognostic parameter/modelHR95%CIP value
  • HR: hazard ratio; 95%CI: 95% confidence interval.

  • a

    Multivariate models also included clinical stage, Gleason score, seminal vesicle invasion, surgical margins, and extracapsular extension.

c-Myc alone   
 Univariate2.7091.395–5.2590.0032
 Multivariate1.6290.789–3.3230.1799
Caveolin-1 alone   
 Univariate2.2121.121–4.3630.0220
 Multivariate2.7621.28–5.9580.0096
c-Myc and caveolin-1   
 Univariate3.0351.429–6.4460.0039
 Multivariate2.9161.273–6.6760.0114

Multivariate analyses were performed for combined c-Myc and caveolin-1 expression, positive c-Myc alone, or positive caveolin-1 alone using clinical stage and common pathologic parameters, i.e., Gleason score, extraprostatic extension, seminal vesicle involvement, and positive surgical margins (Table 4). The preoperative PSA level was not included in these analyses, because > 15% of patients were missing this variable (Table 1). The results indicated that c-Myc expression alone was not an independent prognostic factor for the time to disease progression in this group of patients (HR, 1.6; P = 0.1799), whereas caveolin-1 expression was a prognostic factor (HR, 2.8; P = 0.0096). However, the combination of positive c-myc and caveolin-1 had significant, independent predictive value of recurrence-free survival in these patients with a higher HR compared with the group who had positive caveolin-1 alone (HR, 2.9; P = 0.0114).

DISCUSSION

In this study, we demonstrate that the positive combination of c-Myc and caveolin-1 in prostate carcinoma is an independent prognostic marker for biochemical recurrence. Although we showed previously that caveolin-1 alone is an independent prognostic marker in patients with organ-confined prostate carcinoma,24, 26 the current results using univariate and multivariate analyses suggest that combining c-Myc with caveolin-1 may be a more powerful indicator.

Another important finding of this study is that prostate carcinomas with high c-Myc levels also have a higher probability of being caveolin-1-positive (Fig. 2). The molecular mechanism for this relation remains to be explored. In Rat-1 fibroblast systems with a regulatable c-Myc (MycER™), caveolin-1 expression is down-regulated by c-Myc in vitro33, 45; however, enforced expression of caveolin-1 with an adenoviral vector suppresses c-Myc-induced apoptosis.33 It has been reported that protein kinase C ϵ, which is an oncogenic protein, is able to enhance coexpression of both caveolin-1 and c-Myc in cultured prostate carcinoma cells and is sufficient to induce the growth of androgen-independent tumors in the CWR22 model of human prostate carcinoma.46, 47 Moreover, it has been reported that caveolin-1 over-expressing cells also exhibited a significant activation of the PI3K/Akt pathway,30 which may antagonize the apoptotic activity induced by c-Myc. In this regard, there is increasing evidence that Akt,48 particularly in the active phosphorylated form,49 has prognostic significance. Thus, it is likely that the association of high c-Myc levels with caveolin-1 expression during disease progression involves complex interactions between multiple oncogenic proteins.

Despite evidence that c-myc frequently is over-expressed or amplified and that its deregulated expression is associated with tumorigenesis and progression, the value of c-myc expression alone as a prognostic biomarker in prostate carcinoma has not been established previously, although the copy number of c-myc, as determined by FISH, is predictive of a poor outcome.13 Our results suggest that immunostaining for c-Myc with the described monoclonal antibody yields significant results. Detecting c-Myc expression by immunostaining is difficult due to the extremely short half-life of the c-Myc protein.50 Previous immunostaining results of analysis of tissues from human malignancies have shown that cytoplasmic accumulation of c-Myc protein is more prevalent than nuclear staining.16, 36, 37, 39 Although the reason(s) for these results are not clear, it seems likely that biologic rather that methodologic reasons are responsible, because multiple investigators have reported similar results, and we also have observed similarly increased cytoplasmic-to-nuclear ratios with other antibodies (unpublished results).

The complexities of the network of genes regulated by c-Myc also may contribute to diagnostic difficulties, because cellular context determines whether a cell will undergo proliferation or apoptosis in response to c-Myc. These opposing pathways in the tumorigenic process may obscure the importance of c-Myc expression in individual tumors. It is noteworthy that, in a previous article, we reported that caveolin-1 antagonizes apoptosis potentiated by c-Myc over-expression33 and that caveolin-1 also may render prosurvival properties to metastatic prostate carcinoma cells through activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway.30 Thus, the evaluation of the combined positivity of c-Myc and caveolin-1 may reveal two of the most critical aspects of prostate carcinoma progression, tumor cell growth, and survival potential.

The current study has important implications with regard to the evaluation of prognostic biomarkers. It emphasizes that a potential biomarker should be evaluated not only individually but also in combination with other potential biomarkers that have the capacity to modify its activity. The biologic basis for this approach when evaluating malignancies is rooted firmly in the pioneering studies of oncogenic collaboration.51 In addition to caveolin-1 up-regulation, a variety of abnormal protein activities have been demonstrated in human prostate carcinoma that also may collaborate with c-Myc in disease development/progression, including up-regulation of antiapoptotic proteins, such as bcl-2 and phosphorylated Akt; up-regulation of cell proliferation-promoting proteins, such cyclin D and cyclin E; and down-regulation of tumor suppressors, such as p27 and p53.49, 52, 53 A comprehensive evaluation of the expression of these proteins (many of which are c-Myc targets) for their diagnostic or prognostic significance, together with c-Myc, could yield valuable information.

In summary, we have shown that combined c-Myc and caveolin-1 immunostaining likely reflects the aggressiveness of prostate carcinoma and may offer prognostic value for this malignancy. Because this is a preliminary retrospective study, additional studies in a larger, randomly selected population in both a retrospective and prospective manner are warranted to establish the value of combined c-Myc and caveolin-1 as a clinical biomarker for prognosis in patients with prostate carcinoma.

Acknowledgements

The authors thank Josephine Addai for technical assistance.

Ancillary