10.1245/s10434-006-9042-8
Annals of Surgical Oncology 13:1357-1363 (2006)
© 2006 Society of Surgical Oncology
Comparison of Angiogenic Factor Levels in Tumor Drainage and Peripheral Venous Blood From Colorectal Cancer Patients
Yu-Wen Tien, MD, PhD1,
King-Jen Chang, MD, PhD1,
Yen-Feng Chiu, PhD2,
Kai-Wen Huang, MD1 and
Po-Huang Lee, MD, PhD1
1 Department of Surgery and Angiogenesis Center, National Taiwan University Hospital and National Taiwan University College of Medicine, 7 Chung-Shan South Rd., Taipei 1002, Taiwan, Republic of China
2 Division of Biostatistics and Bioinformatics, National Health Research Institute, 35, Keyan Rd., Zhunan Town, Miaoli County 350, Taiwan, Republic of China
Correspondence: Address correspondence and reprint requests to: Yu-Wen Tien, MD, PhD; E-mail: ywt5106{at}ha.mc.ntu.edu.tw
 |
ABSTRACT
|
|---|
Background: The main objective of this study was to determine whether there was a correlation in the levels of various angiogenesis-related factors between the tumor drainage and peripheral venous blood and whether appraisal of angiogenic factor levels in the tumor drainage venous blood could provide better prognostic information for patients with colorectal cancer than assessment of the peripheral venous blood.
Methods: Plasma levels of vascular endothelial growth factor (VEGF), basic fibroblast growth factor, and endostatin were measured and compared in both tumor drainage and peripheral venous blood from 52 patients with colorectal cancer. Plasma levels of angiogenesis-related factors were also correlated with tumor stage and clinical outcomes.
Results: The plasma endostatin level was significantly higher in peripheral blood than in tumor drainage venous blood (P < .001). The plasma VEGF level was significantly correlated with plasma endostatin levels (P = .028 in tumor drainage venous blood and P = .002 in peripheral venous blood). In both tumor drainage and peripheral venous blood, the VEGF level (but not the basic fibroblast growth factor or endostatin level) was significantly correlated with tumor stage and disease recurrence. However, in multivariate analysis, only plasma VEGF level in tumor drainage venous blood remained an independent predictor of disease recurrence.
Conclusions: The plasma VEGF level in tumor drainage venous blood provided better prognostic information than that in peripheral venous blood. The plasma endostatin level was paradoxically significantly higher in peripheral than in tumor drainage blood, and this strongly suggests additional sources of endostatin in peripheral blood.
Key Words: Vascular endothelial cell growth factor • Fibroblast growth factor • Endostatin • Metastasis • Colorectal cancer
|
INTRODUCTION
|
|---|
Angiogenesis is necessary for tumor growth and metastasis, and the quantitation of angiogenesis seems to be a more sensitive prognostic indicator than the evaluation of conventional pathologic parameters.1,2 Angiogenesis has been studied primarily by directly examining new vessel growth within tumor specimens. Microvessel density has been associated with clinical and pathologic features of biologically aggressive colorectal cancer, as well as disease progression and metastasis formation.3–5 However, measurement of microvessel density requires tumor tissue specimens, and this greatly limits its clinical application. An indirect way to measure angiogenic activity in cancers is to evaluate the expression of angiogenic factors in tumor tissue or the quantity of angiogenic proteins secreted in body fluids. Many studies have demonstrated a significant prognostic value of circulating angiogenic factor level in patients with colorectal cancer,6–13 but others did not find such a correlation.14–17
Tumor cells produce or induce angiogenic molecules that act specifically on vascular endothelial cells.18 However, circulating angiogenesis inhibitors have also been documented that, if produced by the primary tumor and secreted in the blood, can suppress angiogenesis and the formation of distant metastasis.19 It is becoming clear that angiogenesis is the result of a net balance between these positive and negative regulators of neovascularization.20 Presumably, angiogenesis in primary tumors is triggered by higher effective local concentrations of angiogenic stimulators than inhibitors.19,21 Similarly, growth inhibition of metastases by some primary tumors is attributed to higher local concentrations of angiogenic inhibitors.19 These contrasting effects imply that tumors produce opposing factors, achieving different outcomes locally and remotely. In theory, measuring the local angiogenic activity rather than the remote antiangiogenic activity would more faithfully reflect the degree of angiogenesis in the primary tumor and could provide better prognostic information. According to the route of venous drainage, angiogenesis-related factors produced by primary colorectal cancers must pass through the capillary beds of the liver, lungs, and body before they can reach the peripheral veins. Some of these angiogenesis-related factors may be metabolized or decayed in these organs and may be detected in the peripheral blood. Thus, the levels of angiogenesis-related factors in the drainage venous blood of a tumor-bearing segment should more faithfully reflect the local angiogenic activity of primary colorectal cancers and should provide better prognostic information than the level of these factors in peripheral venous blood. To test this hypothesis, we measured and compared the plasma level of most commonly studied angiogenesis-related factors (vascular endothelial growth factor [VEGF], basic fibroblast growth factor [bFGF], and endostatin) in both tumor drainage and peripheral venous blood of 52 patients with colorectal cancer. The plasma levels of these angiogenesis-related factors in both tumor drainage and peripheral venous blood were also correlated with clinicopathologic characteristics and clinical outcome.
|
PATIENTS AND METHODS
|
|---|
Study Population
From July 2000 to December 2002, 58 consecutive patients with histologically confirmed colorectal adenocarcinoma operated on by the same surgeon (Y.-W.T.) at our institution were included in this study. Because one of the purposes of this study was to compare the plasma level of angiogenesis-related factors between paired tumor drainage and peripheral venous blood, six patients with synchronous metastatic diseases were excluded (because angiogenesis-related factors produced by the metastatic lesions could confound the analysis). Thus, 52 patients were included in this study. No patient had an active infection or inflammatory bowel disease at the time of operation or had received a blood transfusion, radiotherapy, or chemotherapy before the study. All 52 patients had radical surgical resection. Patients were considered to have undergone radical resections when there were no metastases detected by preoperative radiological scans (ultrasound scan, chest x-ray, magnetic resonance imaging, or computed tomographic scan) and intraoperative assessment, including intraoperative ultrasound scan of the liver. Patients with stage B2, C1, and C2 colon carcinoma were scheduled to receive adjuvant chemotherapy of 5-fluorouracil and folinic acid, and patients with rectal carcinoma were to receive adjuvant radiotherapy in addition to chemotherapy. Tumor stage and grade were classified according to the Dukes’ system. Clinical characteristics for the 52 patients are summarized in Table 1
.
Patients were followed up every 3 months during year 1, every 4 months during year 2, and ever 6 months thereafter in the outpatient clinic. History and physical examinations were performed at each follow-up visit. Additionally, hemography, liver panel, carcinoembryonic antigen level, chest radiographs, and abdominal ultrasound scans were obtained. Colonoscopy was performed at 24 and 48 weeks and then annually. Abdominal computerized tomography was performed in those patients with an increased carcinoembryonic antigen level or abnormal abdominal ultrasound findings. Patients who failed to return to the clinic as scheduled were contacted telephonically. In the event of death, the date and cause of death were recorded. Disease-free survival time was calculated from the date of operation (in patients having complete resection) to the date of first local or distant recurrence of the disease. The control group consisted of 41 age- and sex-matched healthy volunteers (median age, 64 years; range, 48–76 years; 23 men and 18 women) in whom the absence of neoplastic disease was established by clinical history, physical examination, and routine laboratory testing (including liver and renal function tests). The study protocol was approved by the ethics committee of the National Taiwan University, and informed consent was obtained from all patients and control subjects.
Blood sample collection
Immediately after the peritoneal cavity was entered and before mobilization of the tumor-bearing segment of colon, 10 mL of blood was sampled from the drainage vein of the tumor-bearing segment (tumor drainage venous blood sample). At the time of portal venous blood sampling, 10 mL of blood was collected from the peripheral vein (peripheral venous blood sample). Peripheral venous blood samples from 41 healthy volunteers were also analyzed.
Measurement of Angiogenesis-Related Factors in Plasma
Plasma levels of soluble bFGF, VEGF, and endostatin were measured by using solid-phase enzyme-linked immunosorbent assay (R&D Systems, Inc., Minneapolis, MN). Briefly, microtiter plates with immobilized murine monoclonal antibodies specific to human bFGF, VEGF, or endostatin were used. A second alkaline phosphatase–conjugated polyclonal antibody allowed for the quantitation of bound bFGF, VEGF, or endostatin by measuring absorbance values against a standard curve by using a spectrophotometer set at 450 nm. The range of detection for bFGF, VEGF, and endostatin was 0 to 640 pg/mL, 0 to 2000 pg/mL, and 0 to 500 ng/mL, respectively.
Statistical Methods
Data of angiogenesis-related factors levels are presented as the mean ± SD. Comparisons of angiogenesis-related factors levels between paired tumor drainage and peripheral venous blood samples were performed by using the Wilcoxon signed rank test. Comparison of angiogenesis-related factors levels between healthy and pathologic peripheral blood were performed by using the Mann-Whitney test, with no assumption regarding the distribution of the data. Correlations in the individual angiogenesis-related factor levels between tumor drainage and in peripheral venous blood were performed by using the Spearman rank correlation. The Mann-Whitney U-test and the Kruskal-Wallis test were used to test for the differences in angiogenesis-related factor levels between clinical and pathologic features. The Cox proportional hazards model was used to evaluate the association between pathologic features, plasma angiogenesis-related factor levels, and time to disease recurrence. P values <.05 were considered statistically significant.
|
RESULTS
|
|---|
Patient Outcome
The median follow-up term for the surviving patients was 41 months (range, 38–61 months). Twenty-three patients developed recurrent disease (15 distant metastases and 8 local recurrences). Among these 23 patients who had recurrent disease, 5 were stage B and 18 were stage C. The overall probability of being disease free at 24 months after operation was estimated at .81 ± .07. As expected, patients with advanced Dukes’ stages tended to have worse outcomes, and this provided an internal control for the relatively small cohort recruited for this study.
Plasma Levels of bFGF, VEGF, and Endostatin
The mean bFGF plasma level was 1.6 ± 1.58 pg/mL in healthy peripheral blood samples (H-bFGF, cytokine level in healthy peripheral blood samples), 8.04 ± 7.03 pg/mL in tumor drainage venous blood samples (T-bFGF, cytokine level in pathologic tumor drainage), and 3.46 ± 4.78 pg/mL in pathologic peripheral blood samples (P-bFGF, cytokine levels in pathologic peripheral blood samples; Table 2). The difference between mean levels of H-bFGF and P-bFGF was significant (P = .028; Table 2). The difference between mean levels of T-bFGF and P-bFGF was also significant (P < .0001; Table 2). In the same group of patients, the mean VEGF level in tumor drainage (T-VEGF) and peripheral (P-VEGF) venous blood samples was 43.91 ± 41.37 pg/mL and 37.07 ± 40.84 pg/mL, respectively. The healthy peripheral blood mean VEGF level (H-VEGF) was 11.5 ± 15.65 pg/mL. The difference between H-VEGF and P-VEGF was significant (P < .0001; Table 2), but the difference between T-VEGF and P-VEGF was not significant (P = .10; Table 2). The median plasma endostatin level was 2.71 ± 2.72 ng/mL in tumor drainage venous blood (T-endostatin) and was 5.86 ± 6.19 ng/mL in pathologic peripheral blood (P-endostatin). The mean healthy peripheral blood endostatin content (H-endostatin) was 2.11 ± 2.86 ng/mL. The difference between T-endostatin and P-endostatin was significant (P < .0001; Table 2). The difference between P-endostatin and H-endostatin was also significant (P < .0001; Table 2).
View this table:
[in this window]
[in a new window]
|
TABLE 2. Comparisons of angiogenesis factors levels in healthy peripheral venous blood, tumor drainage venous blood, and peripheral venous blood
|
|
Correlation Between Angiogenesis-Related Factor Levels
No significant correlation was noted between bFGF and VEGF levels or between bFGF and endostatin levels in either tumor drainage or pathologic peripheral venous blood samples (Table 3). However, VEGF levels were significantly correlated with endostatin levels (P = .028 in tumor drainage venous blood samples and P = .0018 in pathologic peripheral venous blood samples; Table 3).
View this table:
[in this window]
[in a new window]
|
TABLE 3. Correlation between paired angiogenesis-related factors levels in peripheral and tumor drainage venous blood
|
|
Correlation in the Angiogenesis-Related Factor Levels Between Tumor Drainage and Peripheral Venous Blood Paired Samples
Both bFGF and VEGF levels in tumor drainage venous blood samples were significantly correlated with those in paired pathologic peripheral venous blood samples (P = .0002 for bFGF; P < .0001 for VEGF; Table 3
). The endostatin level in tumor drainage venous blood samples was not significantly correlated with that in paired peripheral venous blood samples (P = .37; Table 3).
Association Between VEGF, bFGF, and Endostatin Levels and Clinical and Pathologic Characteristics
Table 4 shows the association between plasma levels of VEGF, bFGF, and endostatin and the patients’ clinicopathologic characteristics. In both tumor drainage and peripheral venous blood, levels of VEGF were increased in patients with metastases to regional lymph nodes, tumors of T3 classes, tumors with vascular invasion, and tumors of advanced Dukes’ stage. However, levels of both bFGF and endostatin (either in tumor drainage or peripheral venous blood samples) were not significantly correlated with clinicopathologic characteristics of colorectal cancers, such as sex, depth of tumor invasion, presence of metastases to regional lymph nodes, vascular invasion, lymphatic invasion, or Dukes’ stage (Table 5).
View this table:
[in this window]
[in a new window]
|
TABLE 4. Association of pathologic features and plasma angiogenesis-related factor levels in peripheral venous blood samples with the time to develop recurrences determined by the Cox proportional hazards model
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 5. Association of plasma levels of VEGF, bFGF, and endostatin in venous blood samples with clinical and pathologic characteristics of colorectal cancer
|
|
Prognostic Significance of VEGF, bFGF, and Endostatin Levels
Fifty-two patients were classified into one of two groups according to a high or low level of each angiogenesis-related factor in tumor drainage or peripheral venous blood. The cutoff level corresponded to the median value of the entire population for this classification scheme. In univariate analyses, the presence of vascular invasion, an advanced Dukes’ stage, and a high VEGF level in both tumor drainage and peripheral venous blood all significantly correlated with a high risk of recurrence (Tables 4 and 6). However, the levels of some angiogenesis-related factors in tumor drainage and peripheral venous blood were highly correlated (Table 3) and provided similar information (Tables 4 and 6); thus, the effect of these factors was separated in the multivariate model to better estimate the influence of each factor. Thus, in the multivariate analysis, we first evaluated the risk of recurrence of high levels of VEGF, bFGF, or endostatin in tumor drainage venous blood (Table 4) and then evaluated the prognostic significance of high levels of VEGF, bFGF, or endostatin in peripheral venous blood (Table 6). In this multivariate model, only a high level of VEGF in tumor drainage venous blood remained a significant predictor of disease recurrence independently from Dukes’ stage. VEGF levels in tumor drainage venous blood provided better prognostic information than those in peripheral venous blood.
View this table:
[in this window]
[in a new window]
|
TABLE 6. Association of pathologic features and plasma angiogenesis-related factor levels in tumor drainage venous blood with the time to disease recurrence by the Cox proportional hazards model
|
|
|
DISCUSSION
|
|---|
If the primary colorectal tumor is the exclusive or major source of angiogenesis-related factors, then the levels of plasma angiogenesis-related factors should be higher in tumor drainage venous blood than in peripheral venous blood, and there should be a good correlation between paired tumor drainage and peripheral venous blood for every angiogenesis-related factor level. To our surprise, only plasma bFGF levels were significantly higher in tumor drainage venous blood than in peripheral venous blood. The plasma endostatin level was paradoxically higher in peripheral venous blood than in tumor drainage venous blood. Besides, significant correlations between paired tumor drainage and peripheral venous blood levels were noted for VEGF and bFGF but not for endostatin. The facts that the endostatin level was significantly higher in peripheral venous blood than in tumor drainage venous blood and that no correlation in endostatin level was noted between peripheral and tumor drainage venous blood strongly suggest additional sources of endostatin in peripheral venous blood other than the primary tumor. A significant correlation was noted between VEGF and endostatin levels in both tumor drainage and peripheral venous blood. Therefore, we postulated that the production of VEGF by the tumor induces endostatin production in some organs located between the tumor drainage vein and the peripheral vein. If the assumption were true, then endostatin (which has a longer half-life than angiogenic factors) produced outside the tumor could contribute to growth inhibition of distant metastases.
In both peripheral and tumor drainage venous blood samples, only levels of VEGF (but not bFGF or endostatin) were significantly correlated with patients’ clinicopathologic characteristics. In many aspects, the information obtained from the tumor drainage and peripheral venous blood VEGF measurements was similar. Both of them were significantly correlated with the depth of tumor invasion, the presence of metastases to lymph nodes, the presence of vascular invasion, and Dukes’ stage (Table 5). Additionally, the results indicated that high VEGF levels in tumor drainage and/or peripheral venous blood were predictors of disease recurrence in these patients. Univariate analyses revealed that high VEGF levels in the tumor drainage venous blood were a better predictor. Moreover, in the multivariate analyses, high VEGF levels in tumor drainage venous blood independently predicted a reduced disease-free survival, whereas high VEGF levels in peripheral venous blood did not. Thus, VEGF levels in tumor drainage venous blood provided better prognostic information than those in peripheral venous blood.
Despite the presence of so many angiogenic and antiangiogenic factors, the serum or plasma VEGF level has often been reported to be an independently valuable prognostic factor in patients with colorectal cancers. Possible explanations for this include the following. (1) The angiogenic effect of VEGF outmatches the effect of other angiogenesis-related factors. (2) Angiogenesis-related factor levels in tumor drainage venous blood provide better prognostic information, and VEGF levels in paired peripheral and tumor drainage venous blood correlate better than for any other angiogenesis-related factor. Our results showed that both VEGF and bFGF had levels in tumor drainage and peripheral venous blood samples that correlated significantly. Besides, even in tumor drainage venous blood, VEGF levels provided better prognostic information than bFGF or endostatin levels. Therefore, VEGF might indeed play a pivotal role in the angiogenic process.
In conclusion, we measured the plasma levels of most commonly studied angiogenesis-related factors (VEGF, bFGF, and endostatin) in both tumor drainage and peripheral venous blood samples from 52 patients with colorectal cancer and correlated these parameters with tumor stage and clinical outcomes. In both tumor drainage and peripheral venous blood samples, only VEGF, but not bFGF or endostatin, showed a significant correlation with tumor stage and disease-free survival. However, in multivariate analysis, only VEGF levels in tumor drainage venous blood remained an independent predictor of disease recurrence. Our results showed that VEGF levels in tumor drainage venous blood provided better prognostic information than those in peripheral venous blood.
|
ACKNOWLEDGMENTS
|
|---|
This study was supported by grant from Department of Industrial Technology, Minister of Economic Affairs, Taiwan, Republic of China.
Received for publication October 27, 2005.
Accepted for publication April 5, 2006.
|
REFERENCES
|
|---|
- Folkman J. What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst 1990; 82:4–6.[Free Full Text]
- Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis—correlation in invasive breast carcinoma. N Engl J Med 1991; 324:1–8.[Abstract]
- Takahashi Y, Tucker SL, Kitadai Y, et al. Vessel counts and expression of vascular endothelial growth factor as prognostic factors in node-negative colon cancer. Arch Surg 1997; 132:541–6.[Abstract]
- Choi HJ, Hyun MS, Jung GJ, Kim SS. Tumor angiogenesis as a prognostic predictor in colorectal carcinoma with special reference to mode of metastasis and recurrence. Oncology 1998; 55:575–81.[CrossRef][Medline]
- Tanigawa N, Amaya H, Matsumura M, et al. Tumor angiogenesis and mode of metastasis in patients with colorectal cancer. Cancer Res 1997; 57:1043–6.[Abstract/Free Full Text]
- Dirix LY, Vermeulen PB, Hubens G, et al. Serum basic fibroblast growth factor and vascular endothelial growth factor and tumour growth kinetics in advanced colorectal cancer. Ann Oncol 1996; 7:843–8.[Abstract/Free Full Text]
- Kumar H, Heer K, Lee PW, et al. Preoperative serum vascular endothelial growth factor can predict stage in colorectal cancer. Clin Cancer Res 1998; 4:1279–85.[Abstract]
- Fujisaki K, Mitsuyama K, Toyonaga A, Matsuo K, Tanikawa K. Circulating vascular endothelial growth factor in patients with colorectal cancer. Am J Gastroenterol 1998; 93:249–52.[CrossRef][Medline]
- Hyodo I, Doi T, Endo H, et al. Clinical significance of plasma vascular endothelial growth factor in gastrointestinal cancer. Eur J Cancer 1998; 34:2041–5.[CrossRef][Medline]
- Takeda A, Shimada H, Imaseki H, et al. Clinical significance of serum vascular endothelial growth factor in colorectal cancer patients: correlation with clinicopathologic tumor markers. Oncol Rep 2000; 7:333–8.[Medline]
- Chin KF, Greenman J, Reusch P, Gardiner E, Marme D, Monson JR. Vascular endothelial growth factor and soluble Tie-2 receptor in colorectal cancer: associations with disease recurrence. Eur J Surg Oncol 2003; 29:497–505.[CrossRef][Medline]
- Cubo T, Padilla D, de la Osa G, et al. Serum vascular endothelial growth factor levels in patients with colorectal cancer and its prognostic significance. Med Clin 2004; 122:201–4.
- Feldman AL, Alexander HR Jr, Bartlett DL, et al. A prospective analysis of plasma endostatin levels in colorectal cancer patients with liver metastases. Ann Surg Oncol 2001; 8:741–5.[Abstract/Free Full Text]
- Landriscina M, Cassano A, Ratto C, et al. Quantitative analysis of basic fibroblast growth factor and vascular endothelial growth factor in human colorectal cancer. Br J Cancer 1998; 78:765–70.[Medline]
- Davies MM, Jonas SK, Kaur S, Allen-Mersh TG. Plasma vascular endothelial but not fibroblast growth factor levels correlate with colorectal liver metastasis vascularity and volume. Br J Cancer 2000; 82:1004–8.[CrossRef][Medline]
- Karayiannakis AJ, Syrigos KN, Zbar A, et al. Clinical significance of preoperative serum vascular endothelial growth factor levels in patients with colorectal cancer and the effect of tumor surgery. Surgery 2002; 131:548–55.[CrossRef][Medline]
- Akbulut H, Altuntas F, Akbulut KG, et al. Prognostic role of serum vascular endothelial growth factor, basic fibroblast growth factor and nitric oxide in patients with colorectal carcinoma. Cytokine 2002; 20:184–90.[CrossRef][Medline]
- Desai SB, Libutti SK. Tumor angiogenesis and endothelial cell modulatory factors. J Immunother 1999; 22:186–211.[Medline]
- O’Reilly MS, Holmgren L, Shing Y, et al. Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 1994; 79:315–28.[CrossRef][Medline]
- Morelli D, Lazzerini D, Cazzaniga S, et al. Evaluation of the balance between angiogenic and antiangiogenic circulating factors in patients with breast and gastrointestinal cancers. Clin Cancer Res 1998; 4:1221–5.[Abstract]
- Sugarbaker EV, Thornthwaite J, Ketcham A. (1977) Inhibitory effect of a primary tumor on metastasis. In: Day SB, Myers WPL, Stansly P, Garrattini S, Lewis MG, eds. Progress in Cancer Research and Therapy. New York: Raven Press, pp 227–40.
TOP
ABSTRACT
INTRODUCTION
