To analyse the prognostic value of vascular endothelial growth factor (VEGF) in men with clinically localized prostate cancer.
To analyse the prognostic value of vascular endothelial growth factor (VEGF) in men with clinically localized prostate cancer.
Paraffin wax-embedded sections from the radical prostatectomy (RP) specimens of 40 men operated for clinically localized prostate cancer were used to build tissue microarrays. Of these patients, 17 had cancer progression and bone metastases after RP (group 1), and 23 remained free-of-tumour recurrence after RP (group 2). VEGF-A expression was examined in the RP specimens using immunohistochemistry.
The groups had similar tumour characteristics in terms of prostate-specific antigen level, Gleason score, and pathological stage. VEGF-A expression was significantly higher in group 1 than in group 2 (P = 0.046). In logistic regression analysis, VEGF-A expression was the most significant predictive factor of cancer progression after RP.
VEGF-A expression in prostate cancer tissue is associated with the risk of cancer progression after RP. These results suggest that VEGF-A expression has a prognostic impact in clinically localized prostate cancer.
vascular endothelial growth factor
Angiogenesis plays a crucial role in prostate cancer. Vascular endothelial growth factor (VEGF) is a glycoprotein acting as a specific endothelial mitogen. The VEGF gene family encodes five polypeptide growth factors, i.e. VEGF-A, -B, -C, -D and -E. VEGF-A induces mitogenesis of the endothelial cells of blood vessels, and is involved in many variants of tumours, including prostate cancer . VEGF expression in prostatic tissue is associated with microvessel density and with the presence of cancer . Some authors showed a correlation between circulating VEGF levels and the aggressiveness of prostate cancer, e.g. Jones et al. found, in a series of 78 men (16 with localized prostate cancer, 32 with metastatic cancer, nine with BPH, and 21 healthy controls), that the mean serum VEGF-A level was significantly higher in men with hormone-refractory cancer than in all the other groups. Similarly, Duque et al. found that men with advanced prostate cancer had higher plasma levels of VEGF-A than men with localized disease or than healthy controls. Shariat et al. reported, in a series of 215 men undergoing radical prostatectomy (RP), that plasma levels of VEGF-A were associated with the pathological stage of cancer and with the risk of biochemical recurrence after RP.
Although these studies strongly suggest that the VEGF concentration in plasma might be of prognostic value, the clinical impact of VEGF expression in prostatic tissue is unknown. In the present study, we compared VEGF expression in RP specimens of men who had cancer progression after RP, with a control group who remained free of recurrence.
This study included 40 men who had a RP for clinically localized prostate cancer between 1994 and 2004. All the men had bone scintigraphy and pelvic CT before RP. At the time of RP, no patient had clinical or radiological evidence of lymph node involvement or bone metastases, and none had had hormone therapy or radiation therapy. Standard lymphadenectomy at RP showed no lymph node involvement in all cases. After RP, 17 men (group 1) had biological recurrence and developed bone metastases. Of these, 10 received salvage radiotherapy and seven received hormone therapy at the time of biological recurrence. Bone metastases were confirmed by scintigraphy in all cases. The median (range) interval between RP and the occurrence of bone metastases was 48 (16–73) months. The other 23 patients (group 2) had no biological recurrence after RP, none received adjuvant therapy, and all had undetectable PSA levels at the time of analysis, with a median follow-up of 106 (80–131) months.
Paraffin wax-embedded sections from the RP specimens of the 40 patients were reviewed and mapped. Tissue microarrays were built using a manual tissue arrayer (Beecher Instruments, Alphelys, Plaisir, France). Areas representative of the tumour with the highest Gleason score were circled. For each patient, four cores of 0.6 mm from the circled areas of tumour were transferred to a recipient paraffin-wax block. Controls were obtained from non-malignant prostatic tissue. The tissue array set consisted of three blocks, and included the coordinates of each core and case of origin.
The sections were mounted on poly-l- lysine-coated glass slides and stained immunochemically. Before incubation with primary antibodies, deparaffinized and rehydrated sections were incubated with avidin/biotin blocker (Vector Laboratories, Burlingame, CA, USA) and Fc receptor blocked by human serum (5%). Antigen was retrieved by heating the slides at 97 °C in 0.01 m citrate buffer (pH 6) for 15 min. An anti-VEGF A-20 monoclonal antibody (clone sc-152; Santa Cruz Biotechnology, CA, USA) was used at a dilution of 1 : 200 for 1 h. After rinsing in PBS, the biotinylated secondary antibody was applied for 30 min. To visualize the reaction, sections were incubated with an AEC substrate chromogen (Dakocytomation, Copenhagen, Denmark) for 15–20 min at room temperature. Slides were mounted with Glycergel (Dakocytomation) mounting medium and evaluated under conventional light microscopy. Breast carcinoma tissues were used as positive controls.
Two pathologists (P.C., C.B.), unaware of the clinical data, independently scored the slides, with the intensity of staining scored as: 0, no detectable signal; 1, weak staining; 2, moderate staining; 3, strong staining (Fig. 1). For each patient, four cores were obtained. The final intensity score was the total of the intensity score of the four cores.
For statistical analysis, the intensity of VEGF-A expression was considered as a non-continuous quantitative variable. A Mann–Whitney test was used to compare the intensity of VEGF-A expression between groups 1 and 2. Student’s t-test was used to compare quantitative variables (patient age and PSA values), and the chi-square test to compare qualitative variables (Gleason score and pathological stage). A multivariate logistic regression analysis was used to determine the predictive factors of prostate cancer progression after RP.
Table 1 summarizes the patients’ characteristics before and after RP. The age of patients, the PSA level before RP, the Gleason score of RP specimens, the pathological stage of tumours, and the rate of positive surgical margins were similar in both groups.
|Variable||Group 1||Group 2||P|
|Age, years||66.3 (6.4)||64.5 (5.7)||0.4|
|PSA level before RP, ng/mL||18.3 (18.9)||13.3 (6.9)||0.2|
|Gleason score, n|
|Pathological stage, n|
|Rate of PSM, %||17.6||13||0.7|
Table 2 shows the repartitioning of the intensity scores of VEGF-A expression. VEGF-A expression in the RP specimens was higher in group 1 than in group 2 (P = 0.046). The percentage of patients with an intensity score of VEGF-A expression >8 was 47% (eight of 17) in group 1, vs 17% (four of 23) in group 2 (P = 0.047).
|Intensity score||Group 1, n (%)||Group 2, n (%)|
|4–5||2 (12)||7 (30)|
|6–7||2 (12)||3 (13)|
|8||5 (29)||9 (39)|
|9–10||5 (29)||3 (13)|
|11–12||3 (18)||1 (4)|
|Total||17 (100)||23 (100)|
In the logistic regression analysis (including VEGF-A expression, the PSA level before RP, Gleason score, and pathological stage of cancer) VEGF-A expression was the most significant predictive factor of cancer progression after RP (Table 3).
|Variable||Odds ratio (95% CI)||P|
|VEGF-A expression||1.383 (0.99–1.94)||0.061|
|PSA level before RP||1.035 (0.97–1.10)||0.280|
|Gleason score||0.912 (0.45–1.85)||0.798|
|Pathological stage||0.997 (0.21–4.70)||0.997|
In the present study, the expression of VEGF-A in prostate cancer was higher in men who developed bone metastases than in men who remained free of recurrence after RP. The groups were similar in terms of preoperative PSA level, Gleason score, pathological stage of tumours and surgical margin status. These findings suggest that VEGF-A expression was the sole factor associated with cancer progression after RP.
There is little information about the clinical impact of VEGF expression in prostate cancer. Several studies analysed the prognostic value of VEGF-C expression; VEGF-C causes proliferation of lymphatic endothelial cells and has a role in tumour lymphangiogenesis, whereas VEGF-A stimulates tumour growth at metastatic sites through angiogenesis. Li et al. compared the expression of VEGF-C receptor (VEGFR-3) in BPH and in prostate cancer; VEGFR-3 was up-regulated in cancer, and there was a correlation between VEGFR-3 expression and PSA level, Gleason score, and lymph node metastases. In another study, Jennbacken et al. reported a higher expression of VEGF-C in patients with lymph node metastases than in those with localized prostate cancer. This was confirmed by Trojan et al. who reported more extensive staining of VEGF-A in lymphatic vessels of prostate cancers with lymph node metastases than those of localized tumours or BPH.
Only two previous studies analysed the clinical impact of VEGF-A expression in prostate cancer. Strohmeyer et al. measured VEGF-A expression in 55 cases, and found a correlation between VEGF-A expression and tumour stage and grade; the expression of VEGF-A was higher in nine men who had cancer progression than in the 46 who had no progression during the follow-up. Similarly, West et al. investigated VEGF-A expression in 67 cases of prostate cancer, and found that cases of VEGF immunoreactivity in the stroma were associated with lower survival than in those with negative staining. Both of these studies suggested that VEGF expression could have an impact on clinical outcome, but they included few patients with cancer progression after treatment. Furthermore, the inclusion criteria were not homogeneous. Indeed, some men with advanced prostate cancer at diagnosis were included, and all did not receive the same treatment. It is therefore difficult to assess the real prognostic impact of VEGF expression from these studies. In the present report, only men with clinically localized prostate cancer were included; all received the same treatment, and their preoperative characteristics were homogeneous. To our knowledge, the present report is the first to analyse the prognostic value of VEGF-A expression in a selected population of patients with clinically localized prostate cancer.
The present results suggest that the VEGF-A expression has a prognostic impact; high expression in prostate tissue might be associated with a less favourable outcome. In clinical practice, this has at least two potential implications. When measured in the RP specimen, the expression of VEGF-A could help to identify patients at high risk of tumour recurrence. Currently the main postoperative factors that predict the risk of recurrence are the Gleason score, the pathological stage, and the surgical margin status [11,12]. VEGF-A expression in the RP specimen could be an additional factor for predicting outcome; it could be considered before indicating an adjuvant treatment. Moreover, when measured in prostate biopsies, the expression of VEGF-A could help to distinguish patients at high risk of extracapsular extension. Despite the use of the clinical stage of cancer, PSA level, Gleason score on biopsies, percentage of positive biopsy cores, and imaging, the pre-therapeutic staging of prostate cancer remains a challenge. More than 20% of clinically localized tumours are staged pT3 on pathological analysis of the RP specimen, and >30% of men with extracapsular tumours have positive surgical margins [13,14]. VEGF-A expression in prostate biopsies could help to identify men at high risk of extracapsular tumour, and spare them from unnecessary intervention. VEGF-A expression is simple to measure in current practice; it is based on standard immunochemical methods, and it is easily reproducible. VEGF-A expression could therefore become an additional prognostic marker of clinically localized prostate cancer.
A limitation of the present study is that there were few patients, because of our inclusion criteria for group 1; we included only men who developed bone metastases, instead of all those who had biological recurrence. Our choice was based on the strong prognostic impact of bone metastases. After biochemical recurrence, most patients receive adjuvant treatment, consisting of radiation or hormone therapy. After such treatment, some remain free of disease progression. Brooks et al. recently reported, in 114 men treated with salvage radiotherapy for biological recurrence after RP, a 14% risk of distant metastases after 6 years. In fact, a great proportion of men with biochemical recurrence after RP will keep a favourable prognosis, whereas the vast majority with metastases will soon die. To better determine the prognostic impact of VEGF-A expression, we included only men with bone metastases. Another limitation is our inclusion criteria for group 2. We could not include all men operated upon during the same period; >400 men had RPs between 1994 and 2004, and most remained free of recurrence after surgery, so the material available for determining VEGF-A expression was limited. We therefore included a similar number of patients in both groups, and selected 23 men whose initial characteristics were similar to those in group 1.
Notably, we included some men who had had RP >10 years earlier, mainly because biochemical recurrence and cancer progression after RP are often delayed. In 131 men undergoing RP for clinically localized prostate cancer, with a minimum follow-up of 22.5 years, Swanson et al. found that the median time to recurrence was 7 years. In the present study, the median interval between RP and the diagnosis of bone metastases was 4 years. A long-term follow-up after RP was therefore needed to evaluate the prognostic impact of VEGF-A expression. Another reason is that the oncological results of RP have improved significantly over the last decade, due to a better selection of patients and to technical advances. In recent reports from some high-volume institutions, the 5-year PSA-recurrence rates are <10%. In the present study, it was therefore necessary to include patients who had RP a long time before, due to the decreasing rate of cancer progression after RP.
The critical issue remains the therapeutic impact of VEGF. Recent advances in the understanding of the VEGF system have led to new therapeutic strategies for advanced prostate cancer. Anti-angiogenic agents include bevacizumab, the humanized murine monoclonal antibody to VEGF. The Cancer and Leukaemia Group B study 90006 treated men with hormone-refractory prostate cancer with a combination of bevacizumab and docetaxel ; the regimen was active, with a 53% partial response rate in measurable disease and a 65% biochemical response rate. Anti-angiogenic therapy is still under investigation, and its clinical benefit in advanced prostate cancer remains to be determined.
In conclusion, the expression of VEGF-A in RP specimens of men who developed bone metastases after RP was higher than in men who remained free of recurrence during the follow-up. These results suggest that VEGF expression in tumour tissue has a prognostic impact in clinically localized prostate cancer. To date, the assessment of VEGF-A expression cannot be recommended in daily practice. Indeed, few data are available on the clinical impact of VEGF in localized prostate cancer, and the present findings need to be confirmed by larger longitudinal studies.