Clinical significance of vascular endothelial growth factor C and vascular endothelial growth factor receptor 3 in patients with nonsmall cell lung carcinoma

Authors


Abstract

BACKGROUND

Vascular endothelial growth factor C (VEGF-C) plays an important role in lymphangiogenesis and activates VEGF receptor 3 (VEGFR-3). By contrast, lymphatic spread is an important prognostic factor in patients with nonsmall cell lung carcinoma (NSCLC). The objective of the current study was to determine whether the expression of VEGF-C and VEGFR-3 correlates with clinicopathologic factors and prognosis in patients with primary NSCLC.

METHODS

The authors conducted a retrospective review of 180 consecutive patients who underwent complete resection for NSCLC and who did not receive any chemotherapy or radiotherapy prior to surgery. Immunohistochemical staining for VEGF-C and VEGFR-3 was performed. The clinicopathologic implications of VEGF-C and VEGFR-3 expression were analyzed statistically.

RESULTS

Of 180 patients with NSCLC, 137 patients (76.1%) were positive for VEGF-C, and 40 patients (22.2%) were positive for VEGFR-3. VEGF-C expression was observed frequently in patients with adenocarcinoma (P = 0.026). For VEGFR-3 expression, significant correlations were demonstrated with age (P = 0.02), gender (P = 0.008), and histologic differentiation in patients with squamous cell carcinoma (P = 0.03). Patients who had positive staining for VEGF-C showed significantly less favorable survival rates compared with patients who had negative staining for VEGF-C (P = 0.003). The survival rates of patients who had positive staining for VEGFR-3 also were significantly lower compared with patients who had negative staining for VEGFR-3 (P < 0.001). Patients who had positive staining for both VEGF-C and VEGFR-3 exhibited the most unfavorable prognoses. Univariate analysis revealed the following prognostic factors: gender (P = 0.03), tumor status (T1,T2 vs. T3; P < 0.01), lymph node status (negative vs. positive; P < 0.01), tumor size (≤ 35 mm vs. > 35 mm; P < 0.01), disease stage (Stage I vs. Stages II and III; P < 0.01), VEGF-C expression (negative vs. positive; P < 0.01), VEGFR-3 expression (negative vs. positive; P < 0.01) and combined VEGF-C and/or VEGFR-3 expression (both positive vs. VEGF-C or VEGFR-3 positive; P < 0.01). Multivariate analysis demonstrated that VEGFR-3 expression was the only independent negative prognostic factor (P < 0.01).

CONCLUSIONS

VEGF-C and VEGFR-3 expression may be indicative of survival rates for patients with NSCLC. Cancer 2003;97:457–64. © 2003 American Cancer Society.

DOI 10.1002/cncr.11073

Nonsmall cell lung carcinoma (NSCLC) is the most common cause of disease-related death worldwide, and the outcome of patients with NSCLC remains poor. Current studies have focused on patient prognosis by identifying biologic markers in NSCLC. Disease stages are used widely as prognostic factors after patients undergo surgery.1 Lymph node metastasis specifically is recognized as an important prognostic factor.2 Lymphatic invasion and blood vessel invasion are additional indicators of disease recurrence or poor overall survival.3, 4 However, the mechanism of lymphatic spread remains unclear.

Vascular endothelial growth factor C (VEGF-C) recently has been identified as a new member of the VEGF family.5, 6 The VEGF-C gene is located on chromosome 4q347 and produces a propeptide that, unlike other members of the VEGF family, is cleaved proteolytically. VEGF-C, which is believed to be the only lymphangiogenic factor in the VEGF family, activates both vascular endothelial growth factor receptor 2 (VEGFR-2) and VEGFR-3.8 VEGF-C stimulates the migration and proliferation of endothelial cells in addition to increasing vascular permeability.9 In normal adult human tissue, VEGF-C expression is highest in the heart, placenta, ovary, small intestine, and thyroid gland, suggesting that VEGF-C also plays a role in the paracrine maintenance of differentiated endothelial cell function in the lymphatic vessels.5 VEGF-C induces capillary endothelial cell migration and proliferation in culture10, 11 and stimulates angiogenesis in the cornea and in ischemic muscle.5, 9, 12 VEGF-C induces the proliferation of lymphatic vessels in the stroma of primary gastric carcinoma by activating VEGFR-3 in lymphatic endothelial cells.13 The major receptor for fully processed VEGF-C appears to be the tyrosine kinase flt-4/VEGFR-3, which has two isoforms that differ in their signaling properties;14 however, VEGF-C also binds to VEGFR-2. It has been demonstrated that VEGFR-3 protein plays a role in lymphangiomatosis.15 During development, the expression of VEGFR-3 becomes restricted to lymphatic endothelia.16, 17 Immunohistochemical evaluation has demonstrated that VEGFR-3 was present in large numbers of vascular tumors.18

Current studies suggest a clinicopathologic role for VEGF-C and VEGFR-3 in various malignancies. In patients with lung carcinoma, a positive association has been reported between VEGF-C with lymph node metastasis and lymphatic invasion.19 However, little is known about the expression of VEGF-C and VEGFR-3 in patients with NSCLC. Therefore, in the current study, we used immunohistochemistry to examine the correlation between the expression of VEGF-C or the expression of VEGFR-3 and the clinicopathologic implications.

MATERIALS AND METHODS

Patients

Using a data base of consecutive patients who underwent either lobectomy or pneumonectomy for primary NSCLC at Oita Medical University between 1990 and 1996, we retrospectively reviewed 180 patients who underwent complete resection with systematic lymphadenectomy who did not undergo chemotherapy or radiotherapy before surgery. The patients included 133 males and 47 females, and they ranged in age from 35 years to 84 years (average age, 65 years). The lesions included 101 adenocarcinomas, 65 squamous cell carcinomas, 6 large cell carcinomas, 5 adenosquamous carcinomas, and 3 other lesion types. Grading of postoperative lung carcinoma was undertaken using the TNM classification system published by the International Union Against Cancer in 1997.20 Fifty-two patients had Stage IA disease, 49 patients had Stage IB disease, 9 patients had Stage IIA disease, 29 patients had Stage IIB disease, 40 patients had Stage IIIA disease, and 1 patient had Stage IIIB disease. The median follow-up was 1639 days (range, 43–4038 days).

Immunohistochemistry

Formalin fixed and paraffin embedded sections of tumor tissue from resected lung were cut to a thickness of 4 μm and placed on silane coated slides. After deparaffinization in xylene and rehydration, endogeneous peroxidase activity was blocked by incubation with 3% hydrogen peroxidase for 20 minutes. Tissue sections were then autoclaved at 121 °C in 10 mM citrate buffer, pH 6.0, for 10 minutes for antigen retrieval and were cooled at room temperature for 30 minutes. Sections were immersed in normal goat serum for 15 minutes at room temperature. Immunohistochemical staining for VEGF-C and VEGFR-3 was performed using the streptavidin-biotin-peroxidase complex method (Histofine SAB-PO® kit; Nichirei, Tokyo, Japan). In brief, sections were incubated for 12 hours at 4 °C with a 1:40 dilution of anti-VEGF-C rabbit polyclonal antibody (Immuno-Biological Laboratories Company Ltd., Gunma, Japan) and with a 1:100 dilution of anti-VEGFR-3 rabbit polyclonal antibody (Santa Cruz Biotechnology, Inc, Santa Cruz, CA). Bound peroxidase was visualized using a solution of diaminobenzidine as the chromogen, and nuclei were counterstained with hematoxylin. All sections were examined by two independent investigators who were blinded to the clinical data. Positive staining was defined as the presence of VEGF-C immunoreactivity in at least 30% of tumor cells.21 The immunoreactivity levels of VEGF-C and VEGFR-3 were graded as negative, weak, positive, or strongly positive according to the staining intensity in the cytoplasm of the tumor cells.

Statistical Analysis

The frequency distributions between two groups were tested by using a Fisher exact probability test. A chi-square test was used for histologic type, and the Mann–Whitney U test was used for tumor differentiation. Survival rates were calculated using the Kaplan–Meier method, and significant differences in survival were determined by both log-rank test and the generalized Wilcoxon test. Patients who died in hospital within 30 postoperative days were excluded from this study. For the multivariate analysis, the Cox proportional hazards model was used to identify variables that were associated significantly with survival. All reported P values are two-sided, and P values < 0.05 were considered statistically significant.

RESULTS

VEGF-C and VEGFR-3 Expression in Lung Carcinoma Tissue

VEGF-C and VEGFR-3 were observed almost exclusively in the cytoplasm of lung tumor cells and macrophages (Fig. 1). No staining was seen in normal lung tissue. Of 180 patients with NSCLC, 137 patients (76.1%) were positive for VEGF-C, and 40 patients (22.2%) were positive for VEGFR-3.

Figure 1.

Vascular endothelial growth factor C (VEGF-C) was observed in the cytoplasm of patients with squamous cell carcinoma (a) and adenocarcinoma of the lung (b). VEGF receptor 3 (VEGFR-3) also was observed in the cytoplasm of patients with squamous cell carcinoma (c) and adenocarcinoma of the lung (d).

Correlation between VEGF-C and VEGFR-3 Expression and Clinicopathologic Factors

Table 1 shows that VEGF-C expression was observed frequently in patients with adenocarcinoma (P = 0.026); however, no significant correlations with age, gender, tumor size, tumor (T) status, lymph node (N) status, disease stage, or histologic differentiation were observed. For VEGFR-3 expression, significant correlations were demonstrated with age (P = 0.02), gender (P = 0.008), and histologic differentiation of squamous cell carcinoma (P = 0.03). A positive association of VEGF-C expression and VEGFR-3 expression (P = 0.02) was demonstrated (Table 2). Thirty-six patients showed both VEGF-C expression and VEGFR-3 expression, 105 patients showed expression of either VEGF-C or VEGFR-3, and 39 patients showed expression of neither. No significant correlation between clinicopathologic features and VEGF-C or VEGFR-3 expression was demonstrated after a combined analysis (Table 3).

Table 1. Correlation of Clinicopathologic Features and Expression of Vascular Endothelial Growth Factor C and Vascular Endothelial Growth Factor Receptor 3 in Patients with Nonsmall Cell Lung Carcinoma
CharacteristicVEGF-C expression (%)P valueVEGFR-3 expression (%)P value
Negative/weakPositiveNegative/weakPositive
  • VEGF: vascular endothelial growth factor; VEGFR: VEGF receptor; ca: carcinoma.

  • a

    Fisher exact probability test.

  • b

    Statistically significant.

  • c

    Chi-square test.

  • d

    Mann–Whitney U test.

Age (yrs)      
 ≤ 6519 (10.6)52 (28.9)0.48a49 (27.2)22 (12.2)0.02ab
 > 6524 (13.3)85 (47.2)91 (50.6)18 (10.0)
Gender      
 Male31 (17.2)102 (56.7)0.84a97 (53.9)36 (20.0)0.008ab
 Female12 (6.7)35 (19.4)43 (23.9)4 (2.2)
Tumor size (mm)      
 ≤ 3525 (13.9)88 (48.9)0.48a87 (48.3)26 (14.5)0.85a
 > 3518 (10.0)49 (27.2)53 (29.4)14 (7.8)
T status      
 T1–T240 (22.2)123 (68.3)0.77a128 (71.1)35 (19.4)0.54a
 T3–T43 (1.7)14 (7.8)12 (6.7)5 (2.8)
N status      
 N027 (15.0)84 (46.7)> 0.99a90 (50.0)21 (11.7)0.20a
 N1–N216 (8.9)53 (29.4)50 (27.8)19 (10.5)
Stage      
 I–II35 (19.5)104 (57.8)0.54a110 (61.1)29 (16.1)0.40a
 III8 (4.4)33 (18.3)30 (16.7)11 (6.1)
Histologic type      
 Squamous cell ca24 (13.3)41 (22.8)0.026bc55 (30.6)10 (5.6)0.25c
 Adenocarcinoma19 (10.5)82 (45.6)75 (41.7)26 (14.4)
 Other type0 (0.0)14 (7.8)10 (15.6)4 (2.2)
Differentiation      
 Squamous cell ca      
  Well5 (7.7)4 (6.2)0.30d8 (12.3)1 (1.6)0.03bd
  Moderate15 (23.0)27 (41.5)39 (60.0)3 (4.6)
  Poorly4 (6.2)10 (15.4)8 (12.3)6 (9.2)
 Adenocarcinoma      
  Well6 (5.9)35 (34.7)32 (31.7)9 (8.9)
  Moderate9 (8.9)22 (21.8)0.87d24 (23.8)7 (6.9)0.30d
  Poorly4 (4.0)25 (24.7)19 (18.8)10 (9.9)
Total4313714040
Table 2. Relation between Vascular Endothelial Growth Factor C Expression and Vascular Endothelial Growth Factor Receptor 3 Expression in Patients with Nonsmall Cell Lung Carcinoma
VEGFR-3 expressionVEGF-C expression (%)P value
Negative/weakPositive
  • VEGF: vascular endothelial growth factor; VEGFR: VEGF receptor.

  • a

    Statistically significant (Fisher exact probability test).

Negative39 (21.7)101 (56.1)0.02a
Positive4 (2.2)36 (20.0)
Total43137
Table 3. Relation between Clinicopathologic Features and Expression of Vascular Endothelial Growth Factor C and Vascular Endothelial Growth Factor Receptor 3 in Patients with Nonsmall Cell Lung Carcinoma
CharacteristicVEGF-C/VEGFR-3 expression (%)P valuea
Positive/positivePositive/negative or negative/ positiveNegative/negative
  • VEGF: vascular endothelial growth factor; VEGFR: VEGF receptor.

  • a

    Mann–Whitney U test.

Tumor size (mm)    
 ≤ 3524 (13.3)66 (36.6)23 (12.8)0.49
 > 3512 (6.7)39 (21.7)16 (8.9)
T status    
 T1–T225 (13.9)83 (46.1)31 (17.2)0.37
 T3–T411 (6.1)22 (12.2)8 (4.4)
N status    
 N019 (10.6)67 (37.2)25 (13.9)0.23
 N1–N217 (9.4)38 (21.1)14 (7.8)
Stage    
 I–II25 (13.9)83 (46.1)31 (17.2)0.31
 III11 (6.1)22 (12.2)8 (4.4)
Total3610539

Prognostic Impact of VEGF-C and VEGFR-3 Expression

The prognostic impact of VEGF-C and VEGFR-3 expression using Kaplan–Meier survival curves are shown in Figure 2. The 5-year survival rates of patients who were positive for VEGF-C and patients who were negative for VEGF-C were 47% and 70%, respectively, and the 10-year survival rates were 39% and 65%, respectively (Fig. 2a). Patients who had positive VEGF-C staining showed a significantly lower survival rate compared with patients who had negative VEGF-C staining (P = 0.003). The 5-year and 10-year survival rates for patients who were positive for VEGFR-3 were 28% and 20%, respectively; and the 5-year and 10-year survival rates for patients who were negative for VEGFR-3 were 59% and 53%, respectively (Fig. 2b). The survival rates of patients who were positive for VEGFR-3 were significantly lower compared with patients who were negative for VEGFR-3 (P < 0.001). Survival rates also were evaluated according to combinations of VEGF-C expression and/or VEGFR-3 expression (Fig. 2c). The differences between three groups were statistically significant (P < 0.001). Patients who were positive for both VEGF-C and VEGFR-3 had unfavorable prognoses, and patients who were negative for both VEGF-C and VEGFR-3 expression had more favorable prognoses.

Figure 2.

Survival curves for patients with nonsmall cell lung carcinoma (NSCLC) were stratified by the expression of vascular endothelial growth factor C (VEGF-C) (a), VEGF receptor 3 (VEGFR-3) (b), and a combination of VEGF-C and VEGFR-3 (c) using the Kaplan–Meier method. (a) Patients who had positive staining for VEGF-C showed significantly less favorable survival rates compared with patients who had negative staining for VEGF-C (P = 0.003). (b) The survival rates of patients who had positive staining for VEGFR-3 were significantly lower compared with patients who had negative staining for VEGFR-3 (P < 0.001). (c) Survival rates also were evaluated according to the combination of VEGF-C expression and VEGFR-3 expression. The group of patients with positive staining for both VEGF-C and VEGFR-3 exhibited the poorest prognosis (statistically significant; P < 0.001).

Univariate and Multivariate Analyses

Univariate analysis revealed the following potential prognostic factors: gender (P = 0.03), T status (T1,T2 vs. T3; P < 0.01), N status (negative vs. positive; P < 0.01), tumor size (≤ 35 mm vs. > 35 mm; P < 0.01), disease stage (Stage I vs. Stage II and III; P < 0.01), VEGF-C expression (negative vs. positive; P < 0.01), VEGFR-3 expression (negative vs. positive; P < 0.01) and VEGF-C expression and/or VEGFR-3 expression (both positive vs. VEGF-C positive or VEGFR-3 positive; P < 0.01). Multivariate analysis was used to examine further these potential prognostic parameters. Multivariate analysis demonstrated that VEGFR-3 was the only independent prognostic factor (P < 0.01). VEGF-C was not an independent prognostic factor (P = 0.056) (Table 4).

Table 4. Multivariate Prognostic Analyses of Various Factors in Patients with Nonsmall Cell Lung Carcinoma
ParameterRelative risk ratio95% Confidence limitP value
  • VEGF: vascular endothelial growth factor; VEGFR: VEGF receptor.

  • a

    Statistically significant.

Age (≤ 65 yrs vs. > 65 yrs)1.02090.8187–1.27860.854
Gender (male vs. female)1.14170.8706–1.53490.347
T status (TI, TII vs. TIII)1.34680.7754–2.10940.264
N status (N0 vs. N1, N2)0.7580.4218–1.41320.367
Stage (I vs. II, III)1.36070.7139–2.68090.349
VEGF-C (negative vs. positive)0.64330.4639–0.85890.056
VEGFR-3 (negative vs. positive)1.26960.9932–1.6195< 0.01a

DISCUSSION

Current studies suggest a clinicopathologic role for VEGF-C in various malignancies. Positive correlations between lymph node metastasis or lymphatic vessel invasion and VEGF-C expression have been reported in patients with carcinoma of the head and neck,22 thyroid,23, 24 esophagus,21, 25 breast,26 stomach,13, 27, 28 large intestine,29 uterus,30, 31 and prostate.32 Furthermore, it has been shown that VEGF-C is an independent prognostic predictor only for patients with gastric carcinoma27, 33 and cervical carcinoma.30

Few studies in patients with lung carcinoma have linked VEGF-C expression to clinicopathologic factors. VEGF-C expression was reported in 38.7% of patients19 to 45.1% of patients34 with malignant tumors. In the current study, the frequency of VEGF-C expression was greater than that reported previously (76.1%). Our cohort included large numbers of patients with adenocarcinoma, which exhibits significantly greater VEGF-C expression compared with other tumor tissue types. The primary antibody used in this study was different from other studies, which also may explain the unexpected results. VEGF-C expression was correlated with VEGFR-3 expression, as previously reported,34 but showed no correlation with lymph node metastasis. Although the use of VEGF-C expression as an independent prognostic factor has been supported by several current reports,19, 34 our multivariate analysis indicated that VEGF-C expression tends to be a poor prognostic factor.

In adjacent normal tissue, we observed an absence or reduction of VEGF-C expression, as reported previously.22, 27, 35, 36 Using real-time reverse transcriptase-polymerase chain reaction analysis, Niki and colleagues observed that a high ratio of VEGF-C/VEGF-D expression was correlated with the frequency of lymphatic invasion.37 The current data are consistent with the hypothesis that VEGF-D plays a role in lymph node metastasis and, thus, has prognostic implications. Furthermore, VEGF-C induced macrophage chemotaxis in melanomas, indicating that VEGF-C can act as a direct immunomodulator. The amount of peritumoral macrophage infiltration was increased in the skin surrounding VEGF-C-transfected melanomas.38 It is known that macrophage migration inhibitory factor (MIF) is a proinflammatory cytokine and a glucocorticoid-induced immunomodulator as well as a regular of tumor growth. Intracellular MIF distribution is indicative of prognosis in patients with adenocarcinoma of the lung.39 This offers an alternative explanation why VEGF-C was not identified as an independent prognostic factor. TNM status is related to lymphatic vessel involvement4 but does not have a direct influence on lymph node status in patients with lung tumors. We did not evaluate lymphatic vessel involvement in the current study, although it may be related to VEGF-C expression.

VEGFR-3 expression in tumor cells has been reported in patients with malignant mesothelioma,35, 40 Kaposi sarcoma,41 vascular tumors,42 nasopharyngeal tumors,43 and adenocarcinoma of the lung.44 These reports evaluated VEGFR-3 expression in vascular endothelial cells but not in tumor cells. To our knowledge, Kajita and colleagues reported VEGFR-3 expression in lung tumor cells for the first time, but its impact on prognosis or its clinicopathologic correlations with VEGFR-3 expression have not been evaluated in patients with NSCLC.19 It is noteworthy that VEGFR-3 expression had a significant prognostic impact as evaluated in stained tumor cells in the current study.

A correlation between the expression of both VEGF-C and VEGFR-3 and lymph node metastasis was reported recently in patients with gastric carcinoma,13 prostatic carcinoma,31 adenocarcinoma of the lung,44 breast carcinoma,45 and colorectal carcinoma.29 In the current study, we found a clear and significant correlation between VEGF-C expression and VEGFR-3 expression (P = 0.02). The combination analysis of VEGF-C and VEGFR-3 expression demonstrated a negative impact on prognosis. No correlation between clinicopathologic factors and both VEGF-C expression and VEGFR-3 expression was observed directly; however, these findings may enable more a accurate assessment of the postoperative prognosis of patients with NSCLC.

In conclusion, this study demonstrated that VEGF-C and VEGFR-3 expression had clinicopathologic implications in patients with NSCLC. These findings suggest that VEGF-C and VEGFR-3 may be ideal targets for diagnosis or therapy to improve the prognosis of patients with this deadly disease. Further investigation is necessary to clarify and understand the roles of VEGF-C and VEGFR-3 in patients with NSCLC.

Acknowledgements

The authors thank Miss Naomi Kawano, Miss Yoko Iwata, and Miss Kaori Soe at the Department of Surgery II, Oita Medical University, for technical assistance with immunohistochemical staining.

Ancillary