Introduction
Neuroendocrine neoplasms/tumours (NEN/NET) are heterogeneous tumours arising from a diffuse neuroendocrine cell system, with a broad range of grade, pace of disease, functional status, and primary sites [1]. Their incidence in recent decades is rising and ranges between 1.33 and 2.33/100,000 population in Europe and up to 3.56/100,000 population in the USA [according to the Surveillance, Epidemiology, and End Results (SEER) database] [2], presumably because of improved diagnostic procedures and imaging techniques [1]. The majority of NET encompass well-differentiated tumours with a low proliferation rate (low Ki-67, except NET G3 with Ki-67 above 20%).
The systemic therapy of patients with NEN includes, inter alia, 1st generation somatostatin analogues (SSA) (lanreotide, octreotide) [3–5], both in functional and non-functional NEN. For functional NEN, they reduce production of hormones and secretion of biologically active substances, and control clinical symptoms, but for non-functional, well-differentiated NEN SSA also have an antiproliferative effect, which has been confirmed in 2 randomised studies: PROMID and CLARINET [2].
One of the targets for antineoplastic therapy is inhibition of angiogenesis [6], also in NEN patients. Angiogenesis involves the development of new blood vessels on the basis of already existing previous ones, which may lead to tumour growth and the dissemination of metastasis [7]. As a consequence of hypoxia [8–12], the neoplastic cells secrete vascular endothelial growth factor (VEGF), which stimulates migration and endothelial cell splitting [13], thus inducing angiogenesis of the neoplasms, and it plays important role in metastatic spread. VEGF binds to one of the three tyrosine kinase family receptors: VEGF R1, VEGF R2, and VEGF R3. VEGF has a highest affinity for binding to VEGF R1, but via VEGF R2 it strongly induces endothelial cell proliferation, mainly of blood vessels.
The discovery of antiangiogenic treatment has reduced the mortality rate in neoplasms [14, 15]. Also, in NEN patients, various strategies have been employed therapeutically to antagonize VEGF-mediated tumour angiogenesis. Lyons et al. have proven that VEGF does not stimulate neovascularization in malignant tumour fragments [16]. NEN have strong vascularization, both at the primary site and metastases, so an antiangiogenic treatment by inhibition of angiogenesis is one of the therapy lines in these patients. Moreover, on the basis of immunohistochemistry, high levels of VEGF expression were confirmed on the NEN cells.
The anti-angiogenic effects of SSA were investigated according to the presence of somatostatin receptors (SSTR) on NEN cells and the proliferating vascular endothelium. SSA may suppress angiogenesis directly through SSTR present on endothelial cells and indirectly through the inhibition of growth factor secretion, i.a. VEGF [17, 18]. For the first time, in 1986 O’Dorisio showed the inhibition effect of somatostatin analogues on angiogenesis in vitro models, and then in 1988 with Fassler et al., confirmed also the antiangiogenic effects of octreotide [19, 20]. It comprised preliminary data supporting the antiangiogenic effects of octreotide acetate in a few chicken eggs using the chicken chorioallantoic membrane model. They demonstrated that octreotide acetate could inhibit blood vessel growth. In the next study, Barrie et al. found that the angiogenesis inhibitory ability varied greatly and depended on the structure of the analogue and its amino acid sequence, implying that certain analogues bind to specific SSTR subtypes with varying degrees of affinity [21].
Our study shows the serum VEGF and VEGF R1 before and after treatment with long-acting SSA (lanreotide, octreotide) in NEN patients. Its aim was to determine whether these serum angiogenesis factors can be helpful in assessing the effectiveness of this therapy, thus selecting the appropriate group of NEN patients in whom this therapy gives the greatest benefit. We wanted to see if these tests were warranted both in the decision to start treatment with SSA and in follow-up of the response to this treatment. On the basis of recommendations of the Polish Network of Neuroendocrine Tumours experts (2017), as well as the European Neuroendocrine Tumor Society (ENETS) guidelines (2016), our NEN patients were treated with long-acting octreotide LAR (30 mg i.m. every 4 weeks), and lanreotide Autogel (120 mg s.c. every 4–6 weeks).
Material and methods
Patients
The study enrolled 55 healthy volunteers and 56 NEN patients before (Group 0) and after SSA treatment in various periods (months): 1–12 (Group 1–12, n = 54), 13-24 (Group 13–24, n = 46), 25–36 (Group 25–36, n = 35), 37–60 (Group 37–60, n = 26), and over 60 months (Group > 60, n = 22). The examinations were performed at the Department of Endocrinology and Neuroendocrine Tumours, ENETS Centre of Excellence, and at the Endocrinology Specialist Outpatient Clinic in Katowice. An analysis of medical records and VEGF and VEGF R1 level measurements of NEN patients, who were treated with SSA, were used to examine.
Diagnostic and analytical methods
The serum samples for VEGF and VEGF R1 measurement, both before and after SSA treatment, were collected. After centrifugation at 3000 rpm for 10 minutes, the serum was stored at a temperature of –80°C. Thereafter, serum VEGF and VEGF R1 were determined using Quantikine Human Immunoassay provided by R&D Systems (Minneapolis, MN, USA) according to the manufacturer’s protocol. The results of VEGF and VEGF R1 concentrations were presented in pg/mL.
VEGF metrix: Sensitivity of the method was 9 pg/mL, and intra-assay precision and inter-assay precision was 4.4–6.7% and 6.2–8.8%, respectively.
VEGF R1 metrix: Sensitivity of the method was 3.5 pg/mL, and intra-assay precision and inter-assay precision was 2.6–3.8% and 5.5–9.8%, respectively.
VEGF and VEGF R1 values are expected to be 62–707 pg/mL and 75–179 pg/mL, respectively.
Statistical analysis
Statistical analyses were performed using STATISTICA 13.0 (StatSoft Inc., Tulsa, OK, USA). Concentrations of angiogenesis factors (VEGF and VEGF R1) were expressed as mean values ± standard deviation (median). The comparison between the 2 independent groups (NEN patients and controls) was made using the Mann-Whitney U-test. To investigate the diagnostic capacity of VEGF and VEGF R1 in detecting NEN patients, receiver operating characteristic (ROC) curves were plotted, and the area under the curve (AUC), sensitivity, and specificity were calculated. Intergroup analyses of SSA-treated NEN patients were undertaken using a 2-tailed nonparametric chi-square (Kruskal-Wallis) test and additionally by NIR Fisher’s and Duncan’s test. Test results were considered significant at p < 0.05.
Ethical issues
The study was approved by the Ethics Committee of Medical University of Silesia, Poland (KNW/0022/KB1/130/I/15 and PCN/0022/KB1/97/I/19/20). Informed written permission from all patients and healthy individuals was obtained.
Results
Patients’ and controls’ characteristics are presented in Table 1.
Variable |
NEN |
Controls |
Age [years] Mean (range) |
58 (27–80) |
54 (34–77) |
Gender Male Female |
30 26 |
16 39 |
Grade G1 G2 |
38 18 |
N/A |
Stage I II III IV |
11 11 10 24 |
N/A |
Disease extent — metastases Yes No |
35 21 |
N/A |
Functionality status NF-NEN F-NEN: CS Glucagonoma |
45 11 10 1 |
N/A |
Kind of treatment |
||
SSA Yes No |
56 0 |
N/A |
Surgery Yes No |
29 27 |
N/A |
PRRT Yes No |
0 56
|
N/A |
VEGF
VEGF in all NEN and controls — comparison of these groups
VEGF measurements were elevated in the NEN cohort compared to controls (Tab. 2, Fig. 1A).
Variable |
NEN (n = 56) Mean ± SD (Median) |
Controls (n = 55) Mean ± SD (Median) |
Significance of the difference (Mann-Whitney test) p |
VEGF [pg/mL] |
367.46 ± 277.04 (303.35) |
263.55 ± 173.87 (205.30) |
0.005 |
VEGF R1 [pg/mL] |
365.13 ± 86.99 (345.75) |
96.68 ± 20.53 (92) |
< 0.001 |
AUC for VEGF levels in NEN and controls
AUC analysis could differentiate NEN from controls. Although significant, it should be noted that with an AUC of 0.62, it would be considered a poor biomarker (Fig. 1C). The sensitivity and specificity for the cut-off value were calculated as 74 and 51%, respectively (Tab. 3).
Variable |
AUC |
SE |
95% CI |
p |
Cut-off value |
Sensitivity |
Specificity |
Accuracy |
VEGF |
0.62 |
0.04 |
0.54–0.71 |
0.005 |
206 pg/mL |
74% |
51% |
70% |
VEGF R1 |
1 |
0 |
1–1 |
0 |
190.3 pg/mL |
100% |
100% |
100% |
VEGF in NEN patients according to treatment time groups
According to the Kruskal-Wallis test, we confirmed only 2 significant differences between SSA-treated NEN patients subgroups: Group 1–12 vs. Group 37–60 and Group 1–12 vs. Group > 60 (Fig. 2). During the SSA treatment time, a decrease of the VEGF concentration was observed, i.e. the highest VEGF level was in patients before starting SSA treatment and the lowest was in patients treated for over 60 months. On the other hand, on the basis of NIR Fisher’s and Duncan’s test, we found that the 3 relationships between SSA-treated subgroups were significantly different: Group 0, Group 1–12, and Group 13–24 vs. Group > 60 (p = 0.019, p = 0.034, and p = 0.049, respectively).
VEGF R1
VEGF R1 in all NEN and controls — comparison of these groups
Serum VEGF R1 levels were significantly elevated in the NEN cohort compared to controls (Tab. 2, Fig. 1B).
AUC for VEGF R1 levels in NEN and controls
The AUROC (blue line) for differentiating NEN patients from controls was 1 (95% CI: 1–1, p < 0.001). A maximum AUC = 1 identifies an ideal (perfect) differentiation between these groups. The diagonal red line (AUC = 0.5) in the chart corresponds to chance discrimination. VEGF R1 AUC = 1 (blue line) indicates that it is an excellent biomarker for NEN (Fig. 1D). Both the sensitivity and specificity for the cut-off value were calculated as 100% (Tab. 3).
VEGF R1 in NEN patients according to treatment-time groups
VEGF R1 levels were not significantly different between SSA-treatment NEN patient subgroups (Tab. 4). Increasing VEGF R1 concentration during SSA treatment has been noted — the lowest VEGF R1 level was observed in NEN patients before SSA treatment and the highest in the longest treated patients — for over 60 months.
Factor |
Group 0 |
Group 1–12 |
Group 13–24 |
Group 25–36 |
Group 37–60 |
Group > 60 |
Kruskal-Wallis Test |
VEGF [pg/mL] Mean ± SD (Median) |
410.01 ± 366.46 (321.30) |
395.77 ± 165.87 (327.40) |
388.19 ± 348.32 (267.75) |
375.89 ± 244.11 (334.30) |
268.59 ± 178.18 (264.00) |
247.88 ± 128.76 (236.80) |
c2 = 15.027 p = 0.010 |
VEGF R1 [pg/mL] Mean ± SD (Median) |
359.05 ± 76.03 (342.70) |
359.38 ± 89.12 (344.10) |
362.77 ± 71.23 (361.85) |
375.49 ± 123.02 (361.85) |
365.23 ± 60.42 (347.40) |
383.34 ± 100.54 (360.30) |
c2 = 1.776 p = 0.879 |
Discussion
Numerous studies show that high serum and tumour tissue VEGF concentration indicates intensive development of cancer and is a poor prognostic factor. However, there have also been clinical observations that deny the importance of VEGF in neoplasms, especially its role in the progression of certain neoplasms [22, 23]. Controversy is also raised by the meaning of VEGF activity testing in the clinical evaluation of patients and in making decisions about their treatment [24–28].
Therefore, the aim of this study was to evaluate the relationship between serum VEGF and VEGF R1 and treatment with SSA treatment in NEN patients.
Treatment with SSA is the therapy of choice, both in patients with functional and non-functional NEN, in disease stabilization or progression phase, preferably in well-differentiated NEN (patients with low Ki-67 proliferation index) [5]. Some researchers hypothesized that SSA antitumour effect was i.a. the result of inhibition of angiogenesis [16]. Garcia de la Torre et al. showed that after SSA administration the synthesis and expression of VEGF in colon and rectum tumours were inhibited and serum VEGF levels were decreased [18].
Our results showed significant differences in serum VEGF and VEGF R1 levels between NEN patients and the control group. The mean VEGF concentration was higher in NEN patients than in the control group (367.46 pg/mL vs. 263.55 pg/mL). We also found that the serum VEGF and VEGF R1 level changes reflect the effect of SSA treatment in NEN patients. On the other hand, some treated groups did not reveal such significant VEGF level changes before and after SSA treatment.
We observed a decrease in the VEGF concentration during the time of SSA treatment. We noted the highest concentration in group 0 (before SSA treatment) and the lowest in patients treated for over 60 months (410.01 pg/mL vs. 247.88 pg/mL, respectively).
Perhaps SSA treatment leads to transient responses and further tumour progression because angiogenesis is regulated by various multiple pathways that are able to compensate for each other when a single pathway (VEGF/VEGF R1) is inhibited [29].
The impact of serum VEGF/VEGF R1 on oncogenesis has been a subject of research for many years. Some studies have questioned the accuracy of using serum VEGF as a marker, with the observation that VEGF is released from platelets during venipuncture [30].
Villaume et al. studied the regulation of VEGF production in gastro-entero-pancreatic NEN and the impact of drugs used in NEN therapy on VEGF secretion [31]. The study pointed out that the secretion of VEGF by 3 different endocrine cell lines is significantly decreased by octreotide. Another study [32] analysed in vitro antiangiogenic properties of octreotide. The authors showed that octreotide is able to antagonize the effects of VEGF on endothelial cell proliferation [33] but not on endothelial cell sprouting, and they concluded that the in vitro antiangiogenic effects of SSA are efficiently counterbalanced in the tumour microenvironment by the concomitant release of proangiogenic factors like VEGF. The main mechanism of angiogenesis suppression can be inhibition of endothelial nitric oxide release [34], but inhibition of circulating VEGF also plays a role in the suppression of peritumoral vessel growth [35–36].
Recently, Karpuz et al. evaluated serum VEGF levels as prognostic factors in patients with metastatic colorectal cancer [37]. The analysis included patients before and after treatment with first-line bevacizumab plus chemotherapy. There was no significant correlation between the survival and pre-treatment VEGF level.
In our study mean concentration of serum VEGF R1 was significantly higher in NEN patients than in the control group (365.13 pg/mL vs. 96.68 pg/mL). We also noted increasing concentration during SSA treatment — the lowest level was observed in group 0 (before SSA treatment) and the highest in patients treated for over 60 months (359.06 vs. 383.34 pg/mL, respectively).
A study by Koukorakis et al. analysing serum VEGF levels and tissue activation of VEGF R2 in patients with breast and gynaecological cancer showed significantly higher serum VEGF levels in patients with breast, endometrial, and ovarian cancer compared to healthy controls and patients with benign breast/gynaecological disease in the respective organs [38]. What is more, the expression of phosphorylated VEGF R2 was higher in breast, endometrial, and ovarian cancer in patients with high VEGF serum levels; however, statistical significance was reached when all malignancies were combined.
A recent study by Behelgardi et al. analysed potential effect of new targeted drugs in the treatment of breast cancer [39]. The paper confirmed that simultaneous blockage of VEGF R1 and VEGF R2 inactivates a wider range of signalling pathways of VEGF than blockade of VEGF R1 or VEGF R2 alone, thereby more effectively suppressing tumour growth and metastasis.
Liu et al. investigated the involvement of VEGF R1 in ocular melanoma in animal models. VEGF R1 was responsible for vasculogenic mimicry network formation and was required for efficient choroidal melanoma tumour growth. The study showed VEGF R1 as a potential treatment target [40]. In a study by Enjoji et al., before surgical treatment in patients with biliary carcinoma, VEGF per platelet and VEGF R1 levels were elevated with the lapse of time [41]. Levels of both markers clearly declined as a result of surgical treatment.
A study by Sato et al. suggested that VEGF/VEGF R1 expressions could be associated with cavernous sinus invasion in pituitary neuroendocrine tumours and should be considered as a new direction for targeted therapy [42].
In summary, the studies descriptions indicate that these serum angiogenesis factors can be useful markers for gauging the clinical effect of various treatments on neoplasm patients. Some authors confirmed that in NEN patients with hypervascular tumours, immunohistochemical VEGF expression in NEN cells and serum VEGF are quantitatively correlated [43]. This discovery supports the hypothesis that VEGF production and neovascularization are required for tumour survival. In the available literature, inhibition of VEGF/VEGF R1 pathways seems to be good target for treatment of several neoplasia [44]. The antiproliferative mechanism of SSA in NEN treatment is not fully identified. SSA might suppress angiogenesis; therefore, we looked into the impact of SSA on the concentration of serum angiogenesis factors.
Conclusions
Serum VEGF and VEGF R1 levels seem to have limited usage in the assessment of SSA treatment effectiveness in NEN. Based on our observations, we can only confirm that in NEN patients, some time after treatment, the levels of VEGF increased and VEGF R1 decreased. However, serum VEGF R1 could be a potential marker for distinguishing NET patients from healthy controls.
Limitations
The main limitation of this study is the heterogeneity and different numbers of the NEN patient group. What is more, our analysis was performed on patients treated both with octreotide and lanreotide (in non-equal proportions).
Author contributions
Study conception and design: V.R. Data extraction: V.R., K.J. Analysis of data: V.R. Literature search: V.R., K.J. Writing of the manuscript: V.R., K.J. Responsibility for the paper as a whole: V.R.
Author’s statement
V.R. is the first author.
Conflict of interest
The authors declare no conflict of interest.
Funding
This study was financed by the Medical University of Silesia in Katowice: grant number: KNW-1-131/N/5/0 and KNW-1-110/N/9/K.