Combination of low vascular endothelial growth factor A (VEGF-A)/VEGF receptor 2 expression and high lymphocyte infiltration is a strong and independent favorable prognostic factor in patients with nonsmall cell lung cancer


  • Tom Donnem MD, PhD,

    Corresponding author
    1. Institute of Clinical Medicine, University of Tromso, Tromso, Norway
    2. Department of Oncology, University Hospital of North Norway, Tromso, Norway
    • Department of Oncology, University Hospital of North Norway, 9038 Tromso, Norway
    Search for more papers by this author
    • Fax: (011) 47 77626779

  • Khalid Al-Shibli MD,

    1. Institute of Medical Biology, University of Tromso, Tromso, Norway
    2. Department of Pathology, Nordland Central Hospital, Bodo, Norway
    Search for more papers by this author
  • Sigve Andersen MD,

    1. Institute of Clinical Medicine, University of Tromso, Tromso, Norway
    2. Department of Oncology, University Hospital of North Norway, Tromso, Norway
    Search for more papers by this author
  • Samer Al-Saad MD,

    1. Institute of Medical Biology, University of Tromso, Tromso, Norway
    2. Department of Pathology, University Hospital of North Norway, Tromso, Norway
    Search for more papers by this author
  • Lill-Tove Busund MD, PhD,

    1. Institute of Medical Biology, University of Tromso, Tromso, Norway
    2. Department of Pathology, University Hospital of North Norway, Tromso, Norway
    Search for more papers by this author
  • Roy M. Bremnes MD, PhD

    1. Institute of Clinical Medicine, University of Tromso, Tromso, Norway
    2. Department of Oncology, University Hospital of North Norway, Tromso, Norway
    Search for more papers by this author



There seems to be a close interplay between angiogenesis and the immune system. The authors of this report investigated the prognostic role of angiogenic markers in coexpression with immune system markers in patients with nonsmall cell lung cancer (NSCLC).


Tumor resection samples from 335 patients with stage I to IIIA NSCLC were obtained, and tissue microarrays were constructed. Immunohistochemistry was used to evaluate the expression of vascular endothelial growth factor (VEGF) A (VEGF-A), VEGF receptor 2 (VEGFR-2), and lymphocytes that were positive for the cluster of differentiation 4 (CD4) and CD8 coreceptors.


In univariate analysis, 5-year survival rates were 87% for the combination of low tumor cell expression of VEGF-A and VEGFR-2 (↓VEGF-A/↓VEGFR-2) and high tumor cell expression of CD4 and CD8 (↑CD4/↑CD8) (n = 19), 58% for mixed combinations (n = 290), and 27% for the ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD8 combination (n = 26). In multivariate analysis, the coexpression of ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD8 was an independent negative prognostic factor (hazard ratio, 9.16; 95% confidence interval, 2.11-39.8; P = .003).


Low tumor cell VEGF-A and VEGFR-2 expression in combination with high adaptive immune cell expression in the tumor-related stroma had a strong and independent favorable prognostic impact in patients with NSCLC. Cancer 2010. © 2010 American Cancer Society.

Lung cancer is the leading cause of cancer-related mortality, and new treatment strategies are warranted.1 Surgery has been the cornerstone in curative NSCLC treatment, whereas chemotherapy has been the main treatment modality in advanced disease.2 However, postoperative radiotherapy as well as antiangiogenic treatment in combination with chemotherapy are treatment options for subgroups of patients with NSCLC.2, 3 Recent data on immunotherapy for NSCLC appear promising, but it remains to be established which patients will benefit from such therapy and whether it should be combined with other targeted therapeutic interventions.4, 5

The immune system can be divided into the innate and adaptive immune systems. The innate immune system includes dendritic cells, natural killer cells, macrophages, granulocytes, and mast cells. The antigen-specific, adaptive immune system response is connected to B-lymphocytes, cluster of differentiation 4 (CD4)-positive helper T lymphocytes and CD8-positive cytotoxic T lymphocytes. In cancer, an abundance of infiltrating, innate immune cells, such as macrophages and mast cells, often correlates with a poor prognosis; whereas, in contrast, an increased number of infiltrating lymphocytes frequently is correlated with a favorable prognosis.6 In short, immunotherapy can either support antitumor adaptive immunity or neutralize cancer-promoting properties of innate immune cells.6

There seems to be a close interplay between the immune system and angiogenesis in lung cancer development. In addition to its well documented role in tumor neovascularization, vascular endothelial growth factor (VEGF) A (VEGF-A) also mediates tumor evasion of normal immune surveillance by inhibiting the development of both dendritic cells and T cells.7 Furthermore, VEGF-A stimulates regulatory T cells (T-regs) and inhibits effector T cells, favoring tumor growth.8

We previously reported the prognostic impact of VEGFs, VEGF receptors (VEGFRs), and CD4-positive/CD8-positive lymphocytes.9, 10 On the basis of the proposed interaction between VEGF-A and the adaptive immune system, we wanted to explore the impact of the coexpression between VEGF-A/VEGFR-2 and CD4/CD8 in patients with resected NSCLC.


Patients and Clinical Samples

Primary tumor tissue samples from anonymized patients who were diagnosed with NSCLC (pathologic stages I-IIIA) at the University Hospital of North Norway and Nordland Central Hospital from 1990 through 2004 were used in this retrospective study. In total, 371 patients were registered from the hospital database. Of these, 36 patients were excluded from the study because 1) they had received radiotherapy or chemotherapy before to surgery (n = 10); 2) they had another malignancy within 5 years before their NSCLC diagnosis (n = 13); or 3) they had inadequate paraffin-embedded, fixed tissue blocks (n = 13). Adjuvant chemotherapy was not introduced in Norway during this period (1990-2004). Thus, 335 patients who had complete medical records and adequate paraffin-embedded tissue blocks were eligible.

This report includes follow-up data as of November 30, 2008. The median follow-up of survivors was 86 months (range, 48-216 months). Complete demographic and clinical data were collected retrospectively. Formalin-fixed, paraffin-embedded tumor specimens were obtained from the archives of the Departments of Pathology at University Hospital of North Norway and Nordland Central Hospital. The tumors were staged according to the International Union Against Cancer TNM classification and histologically subtyped and graded according to World Health Organization guidelines.11 Regarding lymph node (N) status, ipsilateral peribronchial or hilar lymph nodes and intrapulmonary lymph nodes are defined as N1, whereas N2 includes ipsilateral mediastinal or subcarinal lymph nodes. The National Data Inspection Board and the Regional Committee for Research Ethics approved this study.

Microarray Construction

All lung cancer specimens were reviewed histologically by 2 pathologists (S.Al-S. and K.Al-S.), and the most representative areas of viable tumor cells (neoplastic epithelial cells) and central tumor stroma were carefully selected and marked on a hematoxylin and eosin-stained slide and sampled for tissue microarray (TMA) blocks. The TMAs were assembled using a tissue-arraying instrument (Beecher Instruments, Silver Springs, Md). The detailed methodology has been reported previously.9 Briefly, we used a 0.6-mm-diameter stylet, and the study specimens were sampled routinely with 2 replicate core samples (different areas) of neoplastic tissue and 2 areas of tumor stroma. Both normal lung tissues localized distant from the primary tumor and 1 slide with a normal lung tissue sample from 20 patients without a cancer diagnosis were used as controls.

To include all core samples, 8 tissue array blocks were constructed. Multiple 4-μm-sections were cut with a Micron microtome (HM355S) and stained by specific antibodies for immunohistochemical analysis.


The applied antibodies were subjected to in-house validation by the manufacturer for immunohistochemistry (IHC) analysis on paraffin-embedded material. The antibodies that we used in this study were as follows: VEGF-A (rabbit polyclonal; RB-1678; Neomarkers, Fremont, Calif; dilution, 1:10), VEGFR-2 (rabbit polyclonal; RB-9239; Neomarkers; dilution, 1:25), CD4 (1F6; Novocastra Laboratories Ltd., Newcastle upon Tyne, United Kingdom; dilution, 1:5), and CD8 (1A5, Ventana Benchmark XT; Ventana Medical Systems Inc., Tucson, Ariz). The detailed IHC procedures were published previously.9, 10 For each antibody, including negative staining controls, all TMA staining was done in a single experiment.

Scoring of IHC

By using light microscopy, representative and viable tissue sections were scored. Regarding the angiogenic markers VEGF-A and VEGFR-2, the dominant cytoplasmic staining intensity in tumor cells was scored semiquantitatively as 0 (negative), 1 (weak), 2 (intermediate), or 3 (strong) (Fig. 1). The mean score for duplicate cores from each individual was calculated separately. Regarding stromal CD4 and CD8, we calculated the percentage of positive cells compared with the total amount of nucleated cells in the whole surface area of the 2 stromal cores. The same cutoff values that were published previously were used.9, 10 All samples were anonymized and were scored independently by 2 pathologists (S.Al-S. and K.Al-S.). In case of disagreement, the slides were re-examined, and a consensus was reached by the observers. In most tumor cores and in some stromal cores, there was a mixture of stromal cells and tumor cells. However, according to morphologic criteria, we scored the intensity of staining only in tumor cells from tumor cores, and we scored both the intensity and the density of stromal cells in stromal cores. When assessing a variable for a given core, the observers were blinded to the scores of the other variables and to outcome.

Figure 1.

In an immunohistochemical analysis of nonsmall cell lung cancer, different cytoplasmic staining intensity scores for tumor cell and stromal expression are shown for (A) tumor cell vascular endothelial growth factor (VEGF) A (VEGF-A) (score = 3 [strong]), (B) tumor cell VEGF-A (score = 1 [weak]), (C) stromal cluster of differentiation 4 (CD4) (high score), and (D) stromal CD4 (low score).

The interobserver scoring agreement was validated previously in the same TMA blocks.9 After categorizing specimens into a high-expression group and a low-expression group, the percentage discordance among the pathologists was as follows: tumor cell ligand, 8%; stromal ligand, 8%; tumor cell receptor, 2%; and stromal receptor, 4%.

Statistical Analysis

All statistical analyses were done using the statistical package SPSS version 15 (SPSS Inc., Chicago, Ill). The chi-square test and the Fisher exact test were used to examine the associations between molecular marker expression and various clinicopathologic parameters. Univariate analyses were done by using the Kaplan-Meier method, and statistical significance between survival curves was assessed by the log-rank test. Disease-specific survival was determined from the date of surgery to the time of death from lung cancer. To assess the independent value of different pretreatment variables on survival, in the presence of other variables, multivariate analyses were carried out using a Cox proportional hazards model. Only variables that had significant value in the univariate analyses were entered into the Cox regression analysis. The probabilities of stepwise entry and removal were set at .05 and .10, respectively, and the level of significance was defined as P < .05.


Clinicopathologic Variables

Demographic, clinical, and histopathologic variables are listed in Table 1. The median patient age was 67 years (range, 28-85 years), and the majority of patients were men (75%). The NSCLC tumors comprised 191 squamous cell carcinomas, 113 adenocarcinomas, and 31 large cell carcinomas. Because of lymph node metastasis or nonradical surgical margins, 55 patients (16%) received adjuvant radiotherapy.

Table 1. Prognostic Clinicopathologic Variables as Predictors of Disease-Specific Survival in 335 Patients With Nonsmall Cell Lung Cancer: Univariate Analyses (Log-Rank Test)
CharacteristicNo. of Patients (%)Median Survival, mo5-Year Survival, %P
  • NR indicates not reached; ECOG, Eastern Cooperative Oncology Group; SCC; squamous cell carcinoma; LCC, large-cell carcinoma.

  • a

    Wedge (n = 10).

Age, y    
 ≤65156 (47)8355.34
 >65179 (53)NR60 
 Women82 (25)19063.20
 Men253 (75)8356 
Smoking history    
 Never15 (5)1943.23
 Current215 (64)NR60 
 Former105 (31)7154 
ECOG performance status    
 0197 (59)NR63.013
 1120 (36)6452 
 218 (5)2533 
Weight loss    
 <10%303 (90)12758.71
 >10%32 (10)9857 
 SCC191 (57)NR66.08
 Adenocarcinoma113 (34)5445 
 LCC31 (9)9856 
 Poor138 (41)4747<.001
 Moderate144 (43)19064 
 Well53 (16)NR68 
Surgical procedure    
 Lobectomy and wedgea243 (73)19061.004
 Pneumonectomy92 (27)3747 
Pathologic stage    
 I212 (63)19069<.001
 II91 (27)4143 
 IIIA32 (10)1819 
Tumor classification    
 T190 (27)19075.002
 T2218 (65)7453 
 T327 (8)4735 
Lymph node status    
 N0232 (69)19066<.001
 N176 (23)3543 
 N227 (8)1818 
Surgical margins    
 Free307 (92)19058.29
 Not free28 (8)4747 
Vascular infiltration    
 No284 (85)19058<.001
 Yes51 (15)2732 

Correlations Between the Coexpression VEGF-A/VEGFR-2 and CD4/CD8 and Clinicopathologic Factors

The coexpression VEGF-A/VEGFR-2 and CD4/CD8 was not correlated with age, sex, performance status, weight loss, differentiation, or vascular infiltration. The coexpression of high tumor cell levels of VEGF-A and VEGFR-2 (↑VEGF-A/↑VEGFR-2) and low tumor cell levels of CD4 and CD8 (↓CD4/↓CD8) were observed most often in patients who had T3 tumors (T1, 4.4%; T2, 7.5%; T3, 18.5%; P = .017) and in patients with positive lymph node status (N0, 5.2%; N1, 11.8%; N3, 18.5%; P = .04).

Univariate Analysis: Coexpression of VEGF-A/VEGFR-2 and CD4/CD8

Among the clinical variables listed in Table 1, performance status (P = .013), tumor differentiation (P < .001), surgical procedure (P = .004), pathologic stage (P < .001), tumor classification (P = .002), lymph node status (P < .001), and vascular infiltration (P < .001) were significant prognostic indicators for disease-specific survival. We previously published survival outcome data for these markers9, 10; however, because of a survival update in November 2008, there are some minor changes with respect to the survival data. Updated results are provided in Table 2. In addition, Table 2 and Figure 2 indicate that the combination of ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD8 expression was associated with a significantly negative prognostic impact (P < .001). The 5-year survival rates were 87% for patients who had VEGF-A/↓VEGFR-2 and ↑CD4/↑CD8 coexpression (n = 19), 58% for patients who had mixed combinations (n = 290), and 27% for patients who had ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD8 coexpression (n = 26).

Table 2. Cluster of Differentiation 4 (CD4), CD8, and Angiogenic Markers as Predictors of Disease-Specific Survival in 335 Patients With Nonsmall Cell Lung Cancer: Univariate Analysis (Log-Rank Test)
Marker ExpressionNo. of Patients, (%)Median Survival, mo5-Year Survival, %P
  1. VEGF-A indicates vascular endothelial growth factor-A; VEGFR-2, VEGF receptor-2; CD4, cluster of differentiation 4; ↓, low expression; ↑, high expression.

 Tumor   .001
  Low192 (52)19066 
  High142 (43)5448 
  Missing1 (0)   
 Tumor   .005
  Low140 (42)NR66 
  High195 (58)7151 
 Stroma   <.001
  Low82 (25)2542 
  High251 (74)19063 
  Missing2 (1)   
 Stroma   .003
  Low268 (80)7554 
  High61 (18)19076 
  Missing6 (2)   
VEGF-A/VEGFR-2 and CD4/C8   <.001
 ↓VEGF-A/↓VEGFR-2 and ↑CD4/↑CD819 (6)NR87 
 Other combinations290 (86)12758 
 ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD826 (8)1327 
Figure 2.

Disease-specific survival curves are shown according to the coexpression of tumor cell vascular endothelial growth factor (VEGF) A (VEGF-A)/VEGF receptor 2 (VEGFR-2) and positive (+) stromal CD4/CD8 expression in an unselected cohort of 335 patients with resected nonsmall cell lung cancer. Downward arrows indicate low expression; upward arrows, high expression.

Multivariate Cox Proportional Hazards Analysis: Coexpression VEGF-A/VEGFR-2 and CD4/CD8

Results from the multivariate analysis are presented in Table 3. All statistically significant (P < .05) clinicopathologic and molecular variables from the univariate analyses were entered into the multivariate analysis. Tumor (T) classification, lymph node status, performance status, vascular infiltration, and differentiation all were identified independent prognostic factors. The 3-level coexpression of VEGF-A/VEGFR-2 and CD4/CD8 had an independent and clearly greater prognostic impact (P < .001) compared with clinicopathologic variables. For patients who had ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD8 coexpression, the hazard ratio (HR) was 9.16 (95% confidence interval [CI], 2.11-39.8) compared with patients who had ↓VEGF-A/↓VEGFR-2 and ↑CD4/↑CD8 coexpression.

Table 3. Results of Cox Regression Analysis Summarizing Significant Independent Prognostic Factorsa
FactorHR95% CIP
  • HR indicates hazard ratio; 95% CI, 95% confidence interval; ECOG, Eastern Cooperative Oncology Group; VEGF-A, vascular endothelial growth factor-A; VEGFR-2, VEGF receptor-2; CD4, cluster of differentiation 4; ↓, low expression; ↑, high expression.

  • a

    Clinicopathologic variables that were significant in univariate analyses (P < .05) and different levels of VEGF-A/VEGFR2 and CD4/CD8 coexpression were included in the analysis.

  • b

    This P value indicates that the variable had overall significance as a prognostic factor.

Tumor classification  .035b
Lymph node status  <.001b
ECOG performance status  .004b
Vascular infiltration  .003
Differentiation  .007b
VEGF-A/VEGFR-2 and CD4/C8  <.001b
 ↓VEGF-A/↓VEGFR-2 and ↑CD4/↑CD81.00  
 Other combinations3.080.75-12.6.12
 ↑VEGF-A/↑VEGFR-2 and ↓CD4/↓CD89.162.11-39.8.003


In this report, we present a large-scale study of an unselected population of patients with surgically resected NSCLC in which we used high-throughput TMA. We observed that the coexpression of ↑VEGFA/↑VEGFR-2 and ↓CD4/↓CD8 was an independent and strong negative prognostic indicator. In contrast, patients who had ↓VEGF-A/↓VEGFR-2 and ↑CD4/↑CD8 coexpression had a highly beneficial prognosis.

There are several indications that the immune system is of importance in lung cancer development. A detrimental survival effect by blood transfusions to patients undergoing surgical treatment for NSCLC has been observed.12, 13 The immunosuppressive effect of allogenic blood transfusions is as a possible explanation.14, 15 The incidence of NSCLC also is increased after organ transplantation and in patients with acquired immunodeficiency syndrome.16, 17 Tumor-infiltrating lymphocytes are considered an indication of the host-immune reaction to tumor antigens, and we demonstrated previously that high tumor-related stromal CD4 and CD8 expression is strongly associated with a good prognosis.10

To our knowledge, this is the first study to report that the combination of low adaptive stromal immunity and high tumor cell expression of angiogenic markers has a substantial and independent, negative impact on patients with lung cancer. We previously reported that stromal CD4 expression and CD8 expression were individual, independent prognostic factors with an HR of 2.6 and 1.9, respectively.10 In our analysis of the combined coexpression of lymphocytes and angiogenic markers, the observed HR was 9.3. The HR indicates at least an additive effect for which there are several plausible biologic explanations. For instance, VEGF-A–induced defects in T-cell development are manifested in a dramatic induction of thymic atrophy and loss of thymocyte cellularity.7 In a murine study, Ohm et al observed a dramatic reduction in CD4/CD8 thymocytes after exposure to recombinant VEGF-A.18

Furthermore, T-regs seem to be stimulated by VEGF-A.8 T-regs are a small subset (10%) of thymus-derived, CD4-positive T cells, and their suppression of autoreactive T cells helps prevent autoimmune disorders. Conversely, T-regs inhibit the immune response by suppressing the effector functions of cytotoxic T cells and thereby suppress antitumor CD8-positive T-cell activity.6

This is consistent with our finding that ↑CD4/↑CD8 expression indicates a less favorable prognosis when combined with high angiogenic marker expression. Without knowing the T-reg/effector T-cell ratio, we may only speculate that, in patients who have high angiogenic marker expression levels, the numbers of T-regs are probably higher relative to the numbers of T-effector cells.

Like what has been reported in many other human tumor types, patients with lung cancer often have increased numbers of circulating T-regs with a significant accumulation of these cells at the tumor site.19 The extent to which this accumulation of T-regs is induced by VEGF-A in NSCLC is less known. Nevertheless, in an in vitro study by Wada et al, VEGF blockade decreased the Treg/effector T-cell ratio. It is noteworthy that those authors also observed that VEGFR-2 expression was higher in T-regs than in other CD4-positive T cells.8 Adding our findings to these previous results actualizes therapeutic angiogenesis blockade in combination with immunotherapy as an interesting approach in patients with resected NSCLC.

Tumor-infiltrating lymphocytes can be divided into 3 groups: lymphocytes within cancer cell nests (epithelial lymphocytes), lymphocytes in the central cancer stroma (stromal lymphocytes), and lymphocytes present along the invasive margins (peritumoral lymphocytes).9, 20 In the current study, the cores were punched from the central parts of the tumors, and the CD4/CD8 coexpression staining represents stromal lymphocytes. This is important, because different localization of the lymphocytes may indicate a different prognostic impact.9, 20, 21 Furthermore, VEGF-A/VEGFR-2 expression was scored in neoplastic epithelial cells; however, we know that this is only a piece of the complex interplay between tumor cells, endothelium, and stromal cells in angiogenesis. Because little is known about the more precise localization of the interplay between angiogenesis and lymphocytes, this issue should be addressed in future studies.

In conclusion, there is a lack of curative therapy for the vast majority of patients with NSCLC and also for patients with limited disease. This means that patients with NSCLC are prime candidates for the evaluation of new treatment modalities. The relative novelty of immunotherapy for NSCLC and its limited clinical success indicates an increased need for translational research within this field. We observed that the combination of ↓VEGF-A/↓VEGFR-2 and ↑CD4/↑CD8 coexpression had a major, positive, independent impact on survival. This finding may inspire further research regarding immunotherapy strategies designed to increase CD4/CD8 lymphocyte infiltration combined with angiogenic blockade.


The authors made no disclosures.