Prognostic impact of angiopoietin-2 in multiple myeloma

Abstract

Purpose

Angiogenesis is an essential process for the expansion of multiple myeloma (MM). Angiopoietin-2 (Ang-2), Ang-1 and their receptor possess important roles in this procedure. The aim of the study was to measure serum levels of Ang-2 along with known markers of angiogenesis and to estimate their prognostic impact on the survival.

Methods

Bone marrow microvascular density (MVD), estimated by CD31, and circulating levels of known angiogenic factors Ang-2, interleukin-6, soluble CD105 and platelet-derived growth factor-AB, measured by ELISA, were measured in 77 newly diagnosed patients with active MM and in 57 of them who responded to chemotherapy.

Results

All measured parameters were increased in MM patients, were also increasing in advanced disease and decreased after effective treatment. Ang-2 correlated positively with the other angiogenic factors and MVD. Moreover, Ang-2 values above the median were accompanied by worse survival.

Conclusion

Ang-2 correlates strongly with the angiogenic process and its serum levels are importantly prognostic for survival, highlighting the role of angiopoietins pathway in the biology of MM.

Introduction

Angiogenesis depicts the formation of new vessels and contributes to the biology and progress of multiple myeloma (MM). It is a complex multi-step process, involving several angiogenic regulators. Measurement of microvascular density (MVD) by counting the mean number of immunohistochemically stained microvessels in paraffin-embedded bone marrow (BM) sections can evaluate the angiogenic process (Sezer et al. 2000).

In MM, the mechanisms of angiogenesis include the production of cytokines by both MM and stromal cells. There is a sensitive balance between angiogenic stimulators and inhibitors binding to receptors of endothelial cells. Angiopoietins (Ang’s) participate actively in angiogenic process and are compromised by 4 structurally related proteins, termed Ang-1, 2, 3 and 4. All of them are ligands for the vascular-specific tyrosine kinase Tie-2 (tyrosine kinase with immunoglobulin-like and EGF-like domains) receptor. Ang-1 and Ang-4 behave as activating ligands for Tie-2, whereas Ang-2 and Ang-3 function as competitive antagonists. The role of Ang-1 and Ang-2 seems to be crucial in angiogenesis alongside VEGF. When bound to Tie-2, Ang-1 contributes to maturation and stabilization of the vascular wall (Yancopoulos et al. 2000; Thurston 2002). In contrast, Ang-2 antagonizes Tie-2 binding and induces vessel destabilization, leading to the angiogenic sprouting (Scharpfenecker et al. 2005). The expression of Ang’s has been investigated in human malignancies, such as breast, liver, lung, thyroid, colon and stomach cancers, where it seems that Ang-2, along with vascular endothelial growth factor (VEGF), possesses an important role in the initiation of tumoral angiogenesis (Papetti and Herman 2001). Expression of Ang-2 has also been observed in isolated BM endothelial cells of patients with MM (Giuliani et al. 2003).

Interleukin-6 (IL-6) is the major growth factor for myeloma cells, possessing significant angiogenic activity (Mitsiades et al. 2006). Platelet-derived growth factor (PDGF) is another stimulator of angiogenesis, interacting with 2 tyrosine kinase receptors (PDGFR-alpha and PDGFR-beta). PDGF is expressed, among other, in endothelial cells, whereas its receptors control functions in various cells, such as vascular smooth muscle cells. PDGF also acts as a mitogen for connective tissue and has been implicated in enhanced proliferation and migration of pericytes. In MM, it seems that PDGF possesses pro-angiogenic activity, since its circulating levels have been correlated with BM MVD and various angiogenic cytokines, such as basic-fibroblast growth factor and angiogenin (ANG). Furthermore, PDGF levels have been correlated with disease burden (estimated by Durie-Salmon staging system and levels of IL-6 and beta 2 microglobulin) (Dong et al. 2007; Tsirakis et al. 2012b). Endoglin (CD105) is a membrane protein predominantly and increasingly expressed on endothelial cells under active angiogenesis. Soluble endoglin (sCD105) is the circulating form of endoglin, after being proteolytically cleaved. In MM, serum levels of sCD105 have been correlated with both BM MVD and circulating levels of various angiogenic cytokines, such as ANG, VEGF and interleukin-18. They also have been correlated with disease activity, as estimated by international staging system (ISS) and BM plasma cells’ Ki-67 proliferation index and infiltration (Duff et al. 2003; Tsirakis et al. 2012a; Pappa et al. 2013b).

The aim of the study was to measure serum levels of Ang-2, along with other angiogenic factors, such as IL-6, sCD105 and PDGF-AB, in newly diagnosed patients with active MM, as well in responders to conventional chemotherapy, in order to correlate them with the angiogenic process, as estimated by BM MVD, and to estimate their prognostic impact in the survival.

Materials and methods

Patients

Seventy-seven patients with newly diagnosed MM (38 males and 39 females, median age 68 years, range 46–81 years) were enrolled in the study. Forty-five patients had monoclonal immunoglobulin IgG, 24 IgA and 8 light chain disease. According to the ISS, 22 of them had stage I, 28 stage II and 27 stage III disease. Patients with acute or chronic infection, inflammatory processes and liver or kidney diseases were excluded from this study. Their primal examination was performed before initiation of any myeloma-related therapy. Fifty-seven of them, who responded to chemotherapy, being in complete remission (CR) (12 patients), very good partial remission (VGPR) (18 patients) or in partial remission (PR) (27 patients), were also studied. Of them, 5 had received VAD (vincristine, liposomal doxorubicin, dexamethasone), 5 PAD (bortezomib, liposomal doxorubicin, dexamethasone), 20 VCD (bortezomib, cyclophosphamide, dexamethasone), 16 VMP (bortezomib, melphalan, methyl-prednisolone) and 11 MP (melphalan, methyl-prednisolone) regimens. The median duration of follow-up was 56 (range 6–125) months. Twenty-six age and sex-matched healthy volunteers were used as controls.

Serum samples

Serum, collected from both patients and controls, stored at −70 °C and examined at the end of the study, in order to avoid inter-assay variability. Circulating levels of IL-6, sCD105, PDGF-AB and Ang-2 were measured with a solid-phase sandwich enzyme-linked immunosorbent assay (ELISA), using monoclonal human antibodies against IL-6, sCD105, PDGF and Ang-2 from commercially available test kits (Quantikine^®, R&D Systems Inc. Minneapolis MN, USA), according to manufacturer’s instructions.

BM samples

Both MM patients and controls underwent transiliac trephine BM biopsy. The pattern of plasma cells infiltration was highlighted by their immunostaining with a monoclonal antibody to CD38. Blood vessels were highlighted by immunostaining endothelial cells with a monoclonal antibody to CD31, as has been described in the past (Tsirakis et al. 2012a). The calculation of MVD was based on the mean microvascular count of three hot spots and expressed as vessels/0.0625 mm².

Statistical analyses

All measured parameters are expressed as mean ± SD. Statistical comparison between MM and control group was made using the nonparametric Mann–Whitney test. The nonparametric Kruskal–Wallis test and the one-way analysis of variance (ANOVA) evaluated differences between disease stage groups. The Student–Newman–Keuls test was used for pairwise comparison of values before and after treatment. Correlation between the measured parameters was studied with Spearman’s correlation. Kaplan–Meier curves estimated overall survival of the patients. A 5 % significance level was applied.

Results

Mean (±SD) values of each measured parameter, both in pre-treatment patients and controls, are shown in Table 1. The levels of all angiogenic factors, as well BM MVD, were higher in MM patients compared to controls (p < 0.001 for all cases). Table 2 shows their values in ISS, where all parameters were increasing in parallel with the disease stage (p < 0.001 for all cases). Furthermore, their values in responders to therapy were significantly decreased compared to pre-treatment ones (p < 0.001 for all cases, Table 3). Interestingly, for the case of Ang-2, its post-treatment values were significantly lower in each ISS stage separately, as well (p < 0.001 for all cases).

Table 1.

Mean ± SD serum levels of platelet-derived growth factor-AB (PDGF-AB), interleukin-6 (IL-6) soluble CD105 (sCD105) angiopoietin-2 (ANGIOP-2) and bone marrow microvascular density (MVD) in multiple myeloma patients and in healthy controls

	PDGF-AB (pg/ml)	IL-6 (pg/ml)	sCD105 (pg/ml)	ANGIOP-2 (pg/ml)	MVD
Controls	225.5 ± 146.3	1.2 ± 0.6	7.7 ± 2.9	290.1 ± 56.8	2.0 ± 0.6
Patients	2,114.6 ± 1,608.7	7.0 ± 5.6	11.9 ± 5.9	689.7 ± 474.4	8.5 ± 4.5
p value	<0.001	<0.001	<0.001	<0.001	<0.001

Table 2.

Mean ± SD serum levels of platelet-derived growth factor-AB (PDGF-AB), interleukin-6 (IL-6) soluble CD105 (sCD105) angiopoietin-2 (ANGIOP-2) and bone marrow microvascular density (MVD) between international staging system stages in multiple myeloma patients

	PDGF-AB (pg/ml)	IL-6 (pg/ml)	sCD105 (pg/ml)	ANGIOP-2 (pg/ml)	MVD
Stage 1	1,162.1 ± 648.0	3.2 ± 2.8	8.0 ± 2.6	301.6 ± 145.8	6.4 ± 2.9
Stage 2	2,104.5 ± 1,609.2	6.5 ± 3.6	10.5 ± 4.6	665.8 ± 432.9	8.4 ± 5.2
Stage 3	2,901.0 ± 1,768.1	10.6 ± 6.8	16.7 ± 6.1	1,030.7 ± 445.0	10.4 ± 4.0
p value	<0.001	<0.001	<0.001	<0.001	<0.001

Table 3.

Mean ± SD serum levels of platelet-derived growth factor-AB (PDGF-AB), interleukin-6 (IL-6) soluble CD105 (sCD105) angiopoietin-2 (ANGIOP-2) and bone marrow microvascular density (MVD) in multiple myeloma patients before and after effective treatment

	PDGF-AB (pg/ml)	IL-6 (pg/ml)	sCD105 (pg/ml)	ANGIOP-2 (pg/ml)	MVD
Before treatment	2,133.28 ± 1,561.34	6.88 ± 5.31	11.75 ± 5.79	635.43 ± 414.02	8.51 ± 4.75
After treatment	620.75 ± 338.63	2.57 ± 1.31	8.13 ± 3.41	345.79 ± 160.61	3.00 ± 1.54
p value	<0.001	<0.001	<0.001	<0.001	<0.001

In the entire group of pre-treatment patients, circulating levels of Ang-2 correlated with IL-6 (r = 0.638) (Fig. 1a), PDGF-AB (r = 0.602) (Fig. 1b), sCD105 (r = 0.547) (not shown) and MVD (r = 0.572) (Fig. 1c) (p < 0.0001 for all cases). Furthermore, in each ISS stage separately, serum levels of Ang-2 correlated with MVD as well (r = 0.615 p < 0.002 for stage I, r = 0.441 p < 0.02 for stage II and r = 0.384 p < 0.05 for stage III).

Fig. 1.

Positive correlations between serum levels of angiopoietin-2 (ANGIOP-2) with interleukin-6 (IL-6) (a), platelet-derived growth factor-AB (PDGF-AB) (b) and bone marrow microvascular density (MVD) (c) in multiple myeloma patients

Survival analysis showed that when pre-treatment levels of Ang-2 were higher than the median (589.1 pg/ml), they were accompanied by inferior survival compared to those with lower values (62 vs. 36 months) (p < 0.04).

Discussion

The system of angiopoietins is critically involved in the physiological angiogenic process (Papetti and Herman 2001). Ang-1 is constitutively expressed in pericytes and smooth muscle cells, acting in a paracrine agonistic manner, by increasing tyrosine phosphorylation of Tie-2 receptor. On the other hand, Ang-2 acts primarily as an autocrine functional antagonist of Ang-1/Tie-2, a vessel-destabilizing cytokine, with an essential role in vascular remodeling. Furthermore, Ang-2 may possess context-specific role on Tie-2-expressing cells; agonistic effects on angiogenically activated endothelium and antagonistic effects on resting cells (Folkman and Klagsbrun 1987; Loughna and Sato 2001; Maisonpierre et al. 1997). Ang’s are possibly secreted by myeloma cells, although reports are contradictory (Giuliani et al. 2003; Uneda et al. 2003).

In this study, we showed that Ang-2 levels are significantly higher in myeloma patients compared to healthy population. Higher levels of Ang-2 have been already measured in MM patients when compared to controls (Terpos et al. 2012; Bhaskar et al. 2013a; Pappa et al. 2013a). We also found that Ang-2 levels differed significantly according to ISS disease stage, with higher levels in advanced disease. These findings confirm previous data describing increased expression of VEGF and Ang-2 in myeloma cells, and lower values of Ang-2 in cases of monoclonal gammopathy of undetermined significance (MGUS) and reactive plasmacytosis (Dong et al. 2007). Furthermore, its increased serum levels have already been correlated, with disease stage and have shown correlations with levels of beta-2 microglobulin, LDH, CRP, IL-6 and grade of bone involvement (Pappa et al. 2013a; Quartarone et al. 2006). We also found significant decrease in Ang-2 levels in responders to conventional chemotherapy, when compared to pre-treatment values, in the entire group of patients, as well in each stage, confirming previous measurements (Terpos et al. 2012; Pappa et al. 2013a; Anargyrou et al. 2008). This notion suggests that circulating levels of Ang-2 may be correlated to disease burden. Nevertheless, its expression from myeloma plasma cells remains contradictory (Giuliani et al. 2003; Dong et al. 2007; Terpos et al. 2012; Agnelli et al. 2005). On the other hand, its production by, both mature and progenitor, endothelial cells is well-established (Holash et al. 1999; Plank et al. 2004). We suggest that serum levels of Ang-2 may be derived from both plasma and endothelial vascular cells.

Another important point in this study is the correlation of Ang-2 with the angiogenic process. In the entire group of untreated MM patients, as well in each ISS stage, its circulating levels correlated significantly with MVD. Furthermore, Ang-2 showed strong correlations with circulating levels of well-established angiogenic cytokines IL-6, sCD105 and PDGF-AB, in the entire group of untreated MM patients. These observations imply the important role of Ang-2 in the angiogenic process per se. Since Ang-2 acts in endothelial cells by an autocrine, as well a paracrine manner (by other cells, possibly myeloma ones), it seems that there is an interactive procedure in the activity of Ang-2. Therefore, the correlation of Ang-2 with the angiogenic process has to do by both the facts that Ang-2 may enhance angiogenesis, as well that its increased serum levels are also the result of increased angiogenesis.

Finally, another important observation from our study was that patients with serum levels of Ang-2 above the median value had significant lower survival. It has already been reported that reduced Ang-1/Ang-2 ratio correlated with features of advanced disease, including ISS, renal impairment and bone disease (Duff et al. 2003) and moreover was an independent predictor for response to therapy (Terpos et al. 2012; Bhaskar et al. 2013a; Anargyrou et al. 2008). Patients with decreased Ang-1/Ang-2 ratio had inferior median survival, even when receiving first-line therapy with novel agent-based regimens. On the other hand, patients with increased serum Ang-1/Ang-2 ratio had favourable prognosis even in advanced stages (Terpos et al. 2012). Recently, it was reported that increased serum levels of VEGF when accompanied with increased mRNA expression of Ang-2 were predictive of inferior progression free survival, further suggesting that VEGF and Ang-2 act in synergy (Bhaskar et al. 2013b). Therefore, Ang-2 seems to play a crucial role in the angiogenic process in MM patients, with a very important impact on their prognosis.

Conclusion

Our study shows that Ang-2 correlates strongly with the angiogenic process. Furthermore, its serum levels, alone or in ratio with Ang-1, as has been reported in the past, are importantly prognostic for response to therapy and mainly for survival. These results highlight the role of angiopoietins pathway in the biology of MM and reveal novel targets for the development of anti-myeloma agents.