Int. J. Mol. Sci. 2012, 13, 1444-1460; doi:10.3390/ijms13021444 OPEN ACCESS International Journal of Molecular Sciences ISSN 1422-0067 www.mdpi.com/journal/ijms Article Elevated Peripheral Blood Plasma Concentrations of Tie-2 and Angiopoietin 2 in Patients with Neuroendocrine Tumors Gabriela Melen-Mucha 1,†, Agata Niedziela 1,†, Slawomir Mucha 2, Ewelina Motylewska 1, Hanna Lawnicka 1, Jan Komorowski 2 and Henryk Stepien 1,* 1 Department of Immunoendocrinology, Chair of Endocrinology, Medical University of Lodz, Dr Sterling Str. No. 3, 91-425 Lodz, Poland; E-Mails: [email protected] (G.M-M.); [email protected] (A.N.); [email protected] (E.M.); [email protected] (H.L.) 2 Department of Clinical Endocrinology, Chair of Endocrinology, Medical University of Lodz, Dr Sterling Str. No. 3, 91-425 Lodz, Poland; E-Mails: [email protected] (S.M.); [email protected] (J.K.) † These authors contributed equally to this work. * Author to whom correspondence should be addressed: E-Mail: [email protected]; Tel./Fax: +48-426324854. Received: 9 January 2012; in revised form: 19 January 2012 / Accepted: 19 January 2012 / Published: 31 January 2012 Abstract: Background: Gastro-entero-pancreatic/neuroendocrine (NET) tumors are highly vascularized neoplasms. However, our knowledge concerning circulating levels of the angiogenic factors in NET patients still remains insufficient. Methods: The aim of this study was to measure plasma concentrations of VEGF, angiopoietin 1 (Ang-1), angiopoietin 2 (Ang-2), soluble Tie-2, endostatin, osteopontin (OPN) and chromogranin A (CgA) in 36 NET patients and 16 controls. Results: Only the plasma concentrations of Tie-2 and CgA were higher in NET patients as compared to controls. These levels were within the reference range in controls; however one control demonstrated slightly elevated Tie-2 and 4 elevated CgA. Similarly, in the subgroup of patients with carcinoid syndrome, only Tie-2 and CgA concentrations were higher than those in patients with non-functioning NETs. In turn, in the subgroup of metastatic patients, only Ang-2 levels were higher than in those with localized disease. A positive correlation was found between Ang-2 and Tie-2 levels in metastatic patients and between Ang-1 and Tie-2 in localized NETs. Conclusions: The plasma concentration of Tie-2 is proposed as an additional marker for NET patients Int. J. Mol. Sci. 2012, 13 1445 and seems to be similarly effective as the currently used CgA level. Moreover, higher plasma levels of Ang-2 together with the positive correlation between Ang-2 and Tie-2 levels in metastatic subjects, implies that cases with a Tie-2 level above the upper limits, together with higher level of Ang-2 seem to be highly predictive of metastases. Keywords: angiogenic factors; NET patients; Ang-1; Ang-2, Tie-2 1. Introduction Tumor angiogenesis is a highly complicated process consisting of several steps, known as the angiogenesis cascade, regulated by many pro- and antiangiogenic factors, both of endogenous and exogenous origin [1,2]. The group of endogenous angiogenic factors comprise, among others, the family of vascular endothelial growth factors (VEGFs) and their receptors (VEGF-A, VEGF-B, VEGF-C, VEGF-D and VGFR-2, VEGFR-3), the family of angiopoietins (Ang) and their receptors (Ang-1, Ang-2, Ang3/4 and Tie-1, Tie-2), other growth factor families, as well as endostatin and osteopontin (OPN). Some of these angiogenic factors are responsible for early activation of the angiogenic cascade, in which the VEGF signaling pathway acts as the most important early regulator of angiogenesis. The angiopoietin signaling pathway controls the later stages of the angiogenic cascade related to vessel maturation and quiescence. Gastro-entero-pancreatic/neuroendocrine (NET) tumors are known to be highly vascularized neoplasms. However, our knowledge concerning circulating levels of the angiogenic factors in NET patients is rather weak and inconsistent. Since the incidence rate of these tumors has grown rapidly in the last two decades, this information seems to be very important not only for diagnostic purposes, but also for changes in the therapeutic options. The high vascularization of NETs suggests that the stromal cells play a role in tumor development. Angiopoietin 1 (Ang-1) is a member of the angiopoietin family of growth factors. It plays a crucial role in mediating the reciprocal interactions between the endothelium, surrounding matrix and mesenchyme. It mediates blood vessel maturation and stability. Ang-1 binds and activates the Tie-2 receptor tyrosine kinase, which is expressed on the endothelial and lymphatic endothelial cells and participates in the regulation of angiogenesis and lymphangiogenesis procesess. Angiopoietin 2 (Ang-2) competes with this binding, thus blocking receptor phosphorylation and counteracting the blood vessel maturation and stability mediated by Ang-1 [3]. Its function, however, may be context-dependent. In the absence of angiogenic inducers such as VEGF, Ang-2-mediated loosening of cell-matrix contacts may induce endothelial cell apoptosis with consequent vascular regression. In concert with VEGF, Ang-2 may facilitate endothelial cell migration and proliferation, thus serving as a proangiogenic signal. It is known that Ang-2 regulates tumor angiogenesis in cooperation with VEGF, as well as with Ang-1, through Tie-2-dependent and independent pathways [4]. Endostatin, a 20 kDa fragment of type XVIII collagen inhibits angiogenesis and the growth of many tumors [5]. It has been demonstrated that endostatin prevents the activation of the angiogenic process and can also induce the regression of some established tumors [6]. In contrast to VEGF, circulating levels of endostatin correlate inversely with cancer angiogenesis and relapse-free survival time [7]. Int. J. Mol. Sci. 2012, 13 1446 Although some studies indicate that the antitumor effect of endostatin can be diminished by VEGF [8], contradictory results showing either a positive correlation between these factors [9] or no correlation [10] have also been published. Osteopontin belongs to the small integrin-binding ligand N-linked glycoprotein SIBLING family [11]. Its putative functions include roles in bone metabolism, immune response regulation, wound healing and tumor progression [12]. Growing evidence supports its role as a potential prognostic factor for various human cancers. OPN is expressed on human carcinomas, such as breast cancer, where it may have adhesion/migration functions in promoting invasion and metastasis [13]. Moreover, elevated OPN levels have been detected in cancer patients, including those with malignant mesothelioma [14]. A few studies have attempted to assess the blood concentrations of angiogenic factors in NET patients, but none provide information concerning such a wide panel of factors as that examined in our study. The data from a German study collected from 90 NET patients revealed increased Ang-2 serum concentration in examined NET subjects as compared to healthy controls. Higher Ang-2 levels were observed in metastatic as compared to local stages, which implies that Ang-2 can be applied as a prognostic marker [15]. Spanish data obtained from 42 NET patients showed enhanced serum concentrations of Ang-1, Ang-2 and soluble Tie-2 as compared to a 27-person control group, and revealed that Ang-2 and Tie-2 levels were higher in patients with metastases compared to those without [16]. In turn, data from a UK study on 47 NET patients revealed that serum Ang-2 was elevated in NET subjects, but not Ang-1. The Ang-2 level was higher in patients with metastases as compared to those with local disease [17]. Russian data revealed lower endostatin levels in benign NETs than in malignant cases [18]. The aim of the study was to search for new auxiliary diagnostic markers for patients with neuroendocrine tumors based on the analysis of six circulating angiogenic factors in plasma. 2. Results and Discussion 2.1. Patients The characteristics of 36 patients with NETs including data concerning age, sex, metastases, hormonal function, localization of primary tumor and somatostatin analog therapy is presented in detail in Table 1. Table 1. The characteristics of patients enrolled to the study. Patient Characteristics Age—median (range) 61 years (31–80) Male—number of patients (median of age; range) 17 (57 years; 32–80) Female—number of patients (median of age; range) 19 (66 years;31–76) Patients with metastatic disease—number of patients (%) 27 (75%) Patients without metastases—number of patients (%) 9 (25%) Patients with carcinoid syndrome—number of patients (%) 7 (19%) Patients with non-functioning NETs—number of patients (%) 29 (81%) Int. J. Mol. Sci. 2012, 13 1447 Table 1. Cont. Patient characteristics Type of NETs—number of patients (%) foregut 11 (31%) midgut 12 (33%) hindgut 6 (17%) unknown or others 7 (19%) Localization of the primary tumor—number of patients (%) lung 6 (17%) thymus 1 (3%) pancreas 2 (6%) stomach 2 (6%) small intestine and appendix 12 (33%) rectum 6 (17%) unknown or others 7 (19%) Concurrent SST analogs—number of patients (%) 14 (39%) 2.1.1. Angiogenic Factors Levels in Patients with NETs and Healthy Controls Patients with NETs had significantly higher plasma levels of Tie-2 (p < 0.001) and CgA (p < 0.001) as compared to healthy controls (Figure 1). The median of the Tie-2 levels was more than 40% higher than that of the controls while CgA was more than 2 times higher. The levels of these factors in the 16 subjects within the control group were within reference range, except for 1 subject with slightly elevated Tie-2, and 4 subjects with elevated CgA levels. In turn, the levels of Tie-2 and CgA remained within the reference range in 10 and 9 NET patients respectively (only one patient had these two factors within normal range). Hence, Tie-2 level would seem to be an additional diagnostic marker for NET patients and seem to have a similar effectiveness as the currently-used CgA levels. In addition, in NET patients, a few positive correlations were found between Tie-2 and Ang-2 (p < 0.01), Ang-2 and CgA (p < 0.05), as well as endostatin and OPN (p < 0.001). Figure 1 shows scatter plots of the angiogenic factor levels with their boundaries of reference range and medians in the NET group and healthy controls. Moreover, the medians of Ang-1 and Ang-2 levels were also higher in the NET group; in turn, the medians of endostatin and OPN were lower than that of the controls, however, the differences did not reach the level of statistical significance. The main reason why the markedly different median values were not statistically significant is the great diversity of the obtained results, especially in the NET group. For example, the range of values in the NET group for Ang-1 levels was 836 pg/mL (min) and 74573 pg/mL (max) vs. 2192 pg/mL (min) and 45396 pg/mL (max) in controls; and the range for CgA levels in the NET group was 2.4 U/L (min) and 809.4 U/L (max), and 4.4 U/L (min) and 109.6 U/L (max) in controls. Int. J. Mol. Sci. 2012, 13 1448 Figure 1. Plasma angiogenic factors (Ang-1, Ang-2, Tie-2, endostatin, osteopontin (OPN), VEGF) and CgA levels in NETs patients and controls (16 volunteers without VEGF measurements) assessed with ELISA. Solid horizontal lines correspond to the medians of plasma levels; in turn, two dashed lines (upper and lower) represent the limits of reference values for these factors in EDTA plasma. The reference range for plasma Ang-1 levels is not given by kit producer (R&D Systems). Only the plasma levels of Tie-2 (p < 0.001) and CgA (p < 0.001) were higher in NET patients as compared to controls (Mann-Whitney U test). Int. J. Mol. Sci. 2012, 13 1449 In controls, a negative correlation was observed between age and CgA (p < 0.05), and between age and Tie-2 levels (p < 0.01), but a positive correlation between age and endostatin concentrations (p < 0.05). In controls, the level of CgA positively and strongly correlated with OPN (p < 0.001) and weakly with Tie-2 (p < 0.05). Additionally, of all the examined factors in our control group, only the levels of Ang-2 and Tie-2 (except for one case) did not go beyond the boundaries of reference standards given in R&D Systems materials. OPN levels were below the lower limits in 3 out of 16 healthy volunteers, whereas endostatin levels exceeded the upper limits in 13 out of 16 healthy controls, which suggest the need to reassess and re-establish new standards for plasma concentrations of endostatin (Figure 1). 2.1.2. Angiogenic Factors Levels in Patients with Carcinoid Syndrome in Comparison to Non-Functioning NETs NET patients with carcinoid syndrome (7 pts) had significantly higher levels of Tie-2 (more than 40%, p < 0.05) and CgA (more than 20 times, p < 0.001) as compared to patients with non-functioning tumors (29 pts) (Figure 2). Among the patients with carcinoid syndrome none had Tie-2 or CgA levels within the reference range. The median levels of factors other than endostatin, OPN and Ang-1 were also higher; the median Ang-2 level, for example, was almost 2 times higher, but the differences were insignificant. Moreover, a positive correlation was found between endostatin and OPN (p < 0.05 in NETs with carcinoid syndrome, p < 0.01 in non-functioning NETs) in both subgroups, and also between Ang-2 and Tie-2 in non-functioning tumors (p < 0.05). Figure 2. Plasma angiogenic factors (Ang-1, Ang-2, Tie-2, endostatin, OPN, VEGF) and CgA levels in patients with carcinoid syndrome (7 pts), non-functioning NETs (29 pts) and controls (16 volunteers without VEGF measurments) assessed with ELISA. Solid horizontal lines correspond to the medians of plasma levels; in turn, two dashed lines (upper and lower) represent the limits of reference values for these factors in EDTA plasma. The reference range for plasma Ang-1 levels is not given by kit producer (R&D Systems). Only the plasma levels of Tie-2 (p < 0.05) and CgA (p < 0.001) were higher in the subgroup of patients with carcinoid syndrome than those in patients with non-functioning NETs (post hoc Kruskal-Wallis test). Int. J. Mol. Sci. 2012, 13 1450 Figure 2. Cont. 2.1.3. Angiogenic Factors Levels and Stage of Disease In patients with metastatic disease (27 pts), only Ang-2 levels were significantly higher than in NET patients with localized disease (9 pts) (p < 0.05): the median value being more than 45% higher (Figure 3). However, there was no correlation between the presence of metastases and Ang-2 levels. In metastatic NETs, a positive correlation was found between Ang-2 and Tie-2 levels (p < 0.05), whereas in localized NETs a positive correlation was observed between Ang-1 and Tie-2 (p < 0.01). In both subgroups, a positive correlation was seen between endostatin and OPN (p < 0.01). The median levels of other factors, except for OPN, were also higher in advanced disease but the differences were statistically insignificant. Int. J. Mol. Sci. 2012, 13 1451 Figure 3. Plasma angiogenic factors (Ang-1, Ang-2, Tie-2, endostatin, OPN, VEGF) and CgA levels in metastatic NETs patients (27 pts), NET patients with localized disease (9 pts) and controls (16 volunteers without VEGF measurments) assessed with ELISA. Solid horizontal lines correspond to the medians of plasma levels; in turn, two dashed lines (upper and lower) represent the limits of reference values for these factors in EDTA plasma. The reference range for plasma Ang-1 levels is not given by kit producer (R&D Systems). In the subgroup of metastatic patients, only Ang-2 (p < 0.05) levels were higher than in those with localized disease (post hoc Kruskal-Wallis test). Int. J. Mol. Sci. 2012, 13 1452 Figure 3. Cont. 2.1.4. Angiogenic Factor Level and NET Types There were no statistically significant differences between medians of the examined levels of factors in patients with foregut, midgut and hindgut NETs, however some correlations were observed in each subgroup. In foregut patients, a negative correlation was seen between age and OPN (p < 0.05) and between Ang-1 and endostatin (p < 0.05) while a positive correlation was seen between CgA and Ang-1 (p < 0.01), endostatin and OPN (p < 0.01) and Ang-2 and Tie-2 (p < 0.05); in midgut patients, only a positive correlation was observed between Ang-1 level and VEGF to endostatin ratio (p < 0.001); finally, in hindgut patients, a negative correlation was found between Ang-1 and Tie-2 (p < 0.05) and three positive correlations were found (p < 0.05) between endostatin and Ang-2 to Tie-2 ratio, OPN and Ang-2 to Tie-2 ratio, and between Tie-2 level and Ang-2 to Ang-1 ratio. Ordering the median value of these factors from largest to smallest, a certain regularity can be observed. For Tie-2, the highest levels were in the midgut, then the foregut and then the hindgut (53 vs. 40 vs. 38 ng/mL respectively). For Ang-2, the highest levels were noticed in the midgut, lower in the hindgut and the lowest in the foregut (3555 vs. 2823 vs. 2352 pg/mL respectively). The medians for Ang-1 in foregut and midgut NETs (14,373 and 14,651 pg/mL respectively) were very close and higher than in the hindgut (12,191 pg/mL).The highest median of VEGF levels were observed in foregut NETs, lower in the midgut and the lowest in the hindgut (163 vs. 132 vs. 61 pg/mL respectively). For endostatin and OPN, the highest to the lowest medians occurred in the opposite order to that of VEGF; both factors had the highest median in the hindgut and the lowest in the foregut (endostatin: 240 vs. 196 vs. 168 ng/mL; OPN 63 vs. 36 vs. 19 ng/mL respectively). 2.1.5. Changes in Angiogenic Factors Levels in Selected NET Patients, Who Had Two Blood Samples Taken Double blood sampling in 4 out of 36 patients allows the changes in concentrations of examined angiogenic factors to be analyzed with passing time and, in 3 patients, after the introduction of somatostatin analog therapy. The characteristics of these 4 patients, including data concerning sex, age, metastases, hormonal function, the type of NET, time of the introduction of somatostatin analog therapy, their dosages and time of treatment, together with the levels of examined angiogenic factors and the changes in their levels expressed in percent, is presented in detail in Table 2. Int. J. Mol. Sci. 2012, 13 1453 Table 2. Changes in angiogenic factors levels in four NET patients, who had two blood samples taken. Time of Two Blood CgA Tie-2 Ang-2 Ang-1 End OPN VEGF Patient Samplings & Changes (U/L) ng/mL (pg/mL) (pg/mL) ng/mL ng/mL (pg/mL) ♀ 1/before the first injection of 69-year-old 20 mg octreotide LAR 34.8 64 16921 13571 208 123 ND metastatic non- (Novartis) functioning 2/one month later, ND 43 3024 8228 160 123 ND midgut NET before the second injection (small intestine) changes ↓30% ↓80% ↓40% ↓20% 0% 1/before the first injection of ♂ 90 mg lanreotide autogel 492 37 2062 13671 296 65 ND 57-year-old (Ipsen) metastatic midgut 2/after 7 months of therapy NET with (90 mg every 4 weeks) with 678.5 42 4086 5231 232 52.5 ND carcinoid well-controlled symptoms syndrome changes ↑40% ↑10% ↑100% ↓60% ↓20% ↓20% 1/good clinical status, 2 months before the first dose 75.7 38 3708 27102 92 14 370.6 of 20 mg ♂ octreotide LAR (Novartis) 47-year-old 2/poor clinical status, just metastatic before death, after 1 year and non-functioning 1 month of octreotide therapy NET of the left 60.8 25 3376 28362 100 13 314.3 (20 mg every 4 weeks) and kidney after many resections of metastatic foci changes ↓19% ↓34% ↓9% ↑5% ↑8% ↓7% ↓13% 1/just before GIST removal on 11.1 48 1406 2217 144 6 ND octreotide therapy 2/after almost 2 years passed, ♂ without new GIST foci and 57-year-old with partial regression of NET metastatic non- metastatic foci, continuously functioning rectal 13.2 43.5 2823 7701 240 63 ND treated with octreotide LAR for NET and gastric 3 years and 6 months GIST (Novartis) 20 mg i.m. every 4 weeks ↑ ↓ ↑ ↑ ↑ ↑ changes ND 19% 10% 100% 350% 160% 1000% CgA: chromogranin A; Tie-2: soluble form Tie-2; Ang-1, -2: angiopoetin 1, 2; End: endostatin, OPN: osteopontin, VEGF: vascular endothelial growth factor; ↓: decrease; ↑: increase; ND: not determined. In one of these patients (69-year-old woman, Table 2), a dramatic decrease in the levels of all examined factors except OPN was observed 1 month after the first injection of octreotide LAR, at almost unchanged clinical status. In the next one (57-year-old man, Table 2) with significant improvement in clinical status observed after 7 months of lanreotide autogel therapy (despite of an