Non-Small Cell Lung Cancer Prognosis Based in a Cut-Off Value for Plasma Basic Fibroblast Growth Factor Expression

Background: Angiogenesis is regulated by angiogenic factors such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) that may be deregulated in lung cancer. The aim of this study was to find out a pattern of VEGF and bFGF protein expression in exhaled breath condensate (EBC) and serum of non-small cell lung cancer (NSCLC) patients and healthy volunteers (smokers and nonsmokers) to obtain early diagnostic values to discriminate initial stages of disease.


INTRODUCTION
Within the European Union, lung cancer accounted for 272 thousand people deaths in 2014 [1]. The prognosis is very poor and the 5-year survival rate decreases with higher stages of disease development [2]. Non-small cell lung cancer parts of the body being often hard to treat, the disease is already in a metastatic state, chemotherapy or surgery has a minimal impact on long-term survival and prognosis remains poor for these patients [2]. The early diagnosis of NSCLC is a key factor to significantly improve overall survival in high risk populations of asymptomatic patients [3] but requires the incorporation of novel techniques with high specificity, which are non-invasive, safe, inexpensive and simple to perform.
Changes in the expression of a biomarker are usually related with the susceptibility to have a disease, the risk of progression of the condition or even with the response to a given treatment. Biomarkers may be detected in peripheral blood, urine or tissue. For respiratory diseases, there can also be used bronchoalveolar lavage [4], sputum [5], exhaled gases [6] or exhaled breath condensate (EBC) [3,7]. Invasive tests like lung biopsies were the only way to investigate the lungs and lower airways, but breath analysis is promising for biomarker detection.
Tumour angiogenesis pathways have been identified as important therapeutic targets for many cancers, including NSCLC, due to angiogenesis is essential in the process of primary tumour growth, proliferation and metastasis [8,9]. Vascular endothelial growth factor (VEGF) seems to have an important role and it has become a major target of antiangiogenic cancer therapy [10]. The literature suggests that the free plasma VEGF concentration is elevated several fold in cancer patients compared to healthy subjects [11].
Besides, the fibroblast growth factor (FGF) family represents a group of heparin-binding, multifunctional polypeptides which also are commonly found dysregulated in malignant tumours [12]. VEGF and FGF have demonstrated co-expression in NSCLC suggesting a synergistic roles in angiogenesis [10].
VEGF initiates the angiogenesis switch, but is not of first importance to the whole process. When the tumours have reached a certain size, other factors such as basic FGF (bFGF), can substitute adequately VEGF. This could explain why in some studies VEGF failed to be a useful marker of long-term survival. Specially in squamous NSCLC, circulating bFGF levels could have a more significant prognostic influence [13,14].
The aim of this prospective study was to find out a pattern of VEGF and bFGF expression in EBC and serum of NSCLC patients and healthy volunteers (smokers and non-smokers) to obtain early diagnostic values that could discriminate initial stages of disease. Ideally, from the data obtained for each biomarker, a cut-off value could be stablished for prognostic purposes.

Study site, patients and controls
The study was approved by the Research Ethics Committee of Galicia (Code 2009/283) and a written informed consent was obtained from all subjects. Participants were recruited from the Pneumology Division at University Hospital Complex of A Coruña, Spain and allocated in experimental or control groups. Experimental group included patients with newly diagnosed NSCLC immediately before histological diagnosis.
None of them had received any form of anti-cancer therapy, invasive diagnostic procedure or primary lung surgery. Control group included subjects without lung cancer and no history of chronic obstructive pulmonary disease (COPD) or other respiratory conditions. This group comprised smokers and exsmokers defined as not having smoked for at least 1 year.

Sample collection
EBC samples were obtained using an EcoScreen condenser (Jaeger, Würzburg, Germany) following current recommendations [15], and rapidly frozen in dry ice. Plasma and serum samples were obtained from peripheral blood.
Samples were stored at -80°C for subsequent analysis.

Enzyme linked immunosorbent assay for the determination of biomarkers
EBC and serum samples were analysed using a sandwich enzyme linked immunosorbent assay (ELISA), Quantikine Human VEGF (R&D Systems Europe, Abingdon UK). This assay is specific for VEGF and does not detect related molecules such as platelet derived growth factor or placental growth factor. The lower limit of detection for VEGF was 9 pg/ml. bFGF was detected in plasma samples using a specific ELISA kit, Quantikine Human bFGF (R&D Systems Europe, Abingdon UK). The limit of detection for bFGF was 3 pg/ml. All the assays were performed in duplicate as previously described [16]. for ROCs were obtained by plotting sensitivity against the false-positive rate (1-specificity). The Youden index [17] gives the value with the higher score of sensitivity and specificity, it was calculated as Youden=sensitivity + (specificity-1) and used to determine optimal cut-off values for VEGF and bFGF for the identification of the better patients prognosis. Survival differences were assessed by the Log-Rank Test. A p-value lower than 0.05 was considered statistically significant.
Controls included 32 healthy smokers and 38 healthy nonsmokers (Table 1). Patients underwent EBC and/or wholeblood collection at enrolment. VEGF and bFGF were detected in serum, plasma and EBC samples. VEGF expression was measured in serum and EBC samples. bFGF expression was measured in plasma samples. VEGF expression was higher in serum compared to EBC in the three groups analysed: cancer, smokers and non-smokers (Table 2). VEGF expression was significantly higher in cancer samples when compared to controls, not only in serum samples (Table 2 and Figure 1A) but also in EBC samples (Table 2 and Figure 1B). Box and Whisker plots identified three outliers in the measures for cancer VEGF EBC, that were not considered for further analysis.   Figure 1C).
There was a positive correlation between serum VEGF and plasma bFGF levels ( Figure 1D  Legend: bFGF, basic fibroblast growth factor; ROC, receiver operating characteristic curve; VEGF, vascular endothelial growth factor.

Figure 1: Box and Whisker plots for VEGF and bFGF levels in serum, EBC and plasma (pg/ml) (A) VEGF concentration in serum samples (B) VEGF concentration in EBC samples (C) bFGF concentration in plasma samples (D) Correlation between serum VEGF and plasma bFGF.
The horizontal lines of the boxes represent 25 %, 50 % (median) and 75 % percentiles (from bottom to top). Whiskers represent minimum and maximum values, and circles are outliers.
The Exp (B) or hazard ratio was 6.837: a NSCLC patient with bFGF expression ≥ cut-off value, the odds of dying are 6.837 times larger than the odds for a patient with bFGF expression < cut-off value.

DISCUSSION
This study demonstrated a positive correlation between VEGF and bFGF measured in blood from NSCLC patients. Moreover, a cut-off value for bFGF expression could be established and used to differentiate those patients with better prognosis and higher survival rates.
The growth of a tumour beyond a certain size requires angiogenesis to supply the extra input of nutrients, and several growth factors such as VEGF and bFGF participate in this process [8,9]. These markers can be detected by immunohistochemistry in lung biopsies [18] or by ELISA in body fluids, mainly blood [19] but also urine [20], bronchoalveolar lavage [21] or pleural effusions [22]. Levels of biomarkers in plasma samples are the reflection of a systemic health situation, but not organ-specific. Blood VEGF varies along the distinct periods of the female life span [23], being more accurate to measure this marker in the EBC for the study of respiratory diseases [16,24]. Nevertheless, some difficulties for EBC analysis have been reported along with some considerations for storage and manipulation of the samples [16]. This study shows that VEGF expression can be detected in EBC and blood, being the concentration higher in serum samples as reported in the literature [16,25]. No statistically significant differences could be observed between cancer and control patients. This result might be triggered by the inherent difficulties of processing and conservation of EBC samples with different storage periods [16]. Anyhow, the possibility of detecting biomarkers in EBC makes it a promising sample for lung determinations. VEGF and bFGF levels were found to be higher in NSCLC blood samples compared to healthy subjects, pointing to their possible role in the disease development.
A positive correlation between serum VEGF and plasma bFGF was found, suggesting that both markers cooperate during lung disease to promote the necessary angiogenesis for the tumour growth [10]. Considering the above, cut-off values were stablished for VEGF and bFGF and used to obtain survival plots. Clear differences in survival were observed when NSCLC patients were classified according to VEGF or bFGF values higher or lower than the cut-off points. Nevertheless, these differences were only significant for bFGF.
The most common histologic subtype of NSCLC is adenocarcinoma (40%), followed by SCC (25% to 30%), and large cell carcinomas (5% to 10%) [26]. NSCLC anti-angiogenic therapies are mainly aimed at patients with adenocarcinomas, where VEGF seems to perform a crucial role [27]. Bevacizumab is a recombinant humanized monoclonal antibody that Besides, active smoking was not found as a prognosis factor in men [35]. Previous publications found that smokers had a higher survival rate compared to non-smokers in NSCLC [36] but also that the impact of smoking status was only important when the status disease was diagnosed in an early stage [37]. Lung cancer in non-smokers is often diagnosed at a late stage, being first attributed to a respiratory infection or even allergies.
Considering that the majority of individuals included in our study were diagnosed of NSCLC in later stages (n=18) versus early stages (n=7), the lack of effect of smoking status might be a biased interpretation. This study has some limitations related to the small sample size that could be biasing the results due to the unproportioned subgroups of NSCLC type. Besides, due to volume sample limitation, bFGF could not be determined in EBC samples so new patient recruitment would be necessary to confirm if the results observed in plasma are replicable in this sample.

CONCLUSION
This study demonstrates a positive correlation between angiogenesis biomarkers (VEGF and bFGF) in blood and also that VEGF could be detected in EBC. A cut-off value for bFGF could be stablished to identify patients with higher overall mortality and poor prognosis, independently of age, sex or smoking habit. Finally, bFGF expression seems to be more important in SSC, suggesting that anti-angiogenic therapy should be selected depending on the histology of the disease.

CONFLICT OF INTEREST
The authors report no conflicts of interest.