Circulating angiopoietin-1 to angiopoietin-2 ratio is an independent prognostic factor for survival in newly diagnosed patients with multiple myeloma who received therapy with novel antimyeloma agents



The circulating levels of several angiogenic cytokines [angiopoietin-1 (Ang-1), angiopoietin-2 (Ang-2), vascular endothelial growth factor (VEGF), angiogenin and basic fibroblast growth factor (bFGF)] were evaluated in 174 consecutive patients with newly diagnosed, symptomatic, multiple myeloma (MM). Circulating levels of Ang-1/Ang-2 were reduced in myeloma patients compared to controls, whereas VEGF and angiogenin levels were increased. Reduced angiopoietin-1/angiopoietin-2 ratio correlated with advanced disease features including international staging system (ISS)-3 stage, renal impairment and extensive bone disease. Based on immunohistochemical results in 20 patients (10 with the higher and 10 with the lower values of circulating angiopoietin-2) we found that angiopoietin-2 is expressed by myeloma cells and correlates with increased microvessel density in subsets of patients. Furthermore, Ang-1/Ang-2 ratio correlated with survival. Patients with circulating Ang-1/Ang-2 below or equal to the median value (6.03) had a median survival of 26.3 months compared to 53 months of all others (p = 0.002). Interestingly, this was mainly observed in patients who received first-line therapy with novel agent-based regimens (65% of our patients). Furthermore, a subset of ISS-3 patients with serum Ang-1/Ang-2 above the median value had favourable prognosis (median survival: 45 months versus 17 months of all others; p = 0.0001). The multivariate analysis revealed that low Ang-1/Ang-2 ratio could independently predict for inferior survival in our cohort of patients (relative risk (RR) 2.07, 95% CI 1.50–2.42; p < 0.001). These results highlight the role of angiopoietins pathway in the biology of MM and reveal novel targets for the development of antimyeloma agents.

Angiogenesis represents an essential step of disease progression in several hematological malignancies, including multiple myeloma (MM).1, 2 Complex interactions between myeloma, stromal and endothelial cells that are mediated by various cytokines with angiogenic properties promote the myeloma cell survival and proliferation in the marrow microenvironment.3 These cytokines include mainly the vascular endothelial growth factor (VEGF), the angiogenin and the basic fibroblast growth factor (bFGF).3 However, several other cytokines are involved in the increased angiogenesis of MM and are under investigation. Angiopoietin-1 (Ang-1) and its natural antagonist angiopoietin-2 (Ang-2) are essential cytokines for normal angiogenesis process.4 Angiopoietins are possibly expressed and secreted by myeloma cells, although reports are contradictory.5, 6 The ratio of circulating levels of Ang-1/Ang-2 was reduced in patients with relapsed/refractory myeloma.7 However, there is limited information on the value of this ratio in patients with newly diagnosed disease who are treated with novel antimyeloma agents, such as thalidomide, bortezomib and lenalidomide that are very effective8–10 and seem to exert antiangiogenic properties.11, 12

The aim of this study was to evaluate the circulating levels of Ang-1, Ang-2, VEGF, VEGF-A, angiogenin and bFGF in a large number of newly diagnosed, previously untreated, symptomatic myeloma patients who received therapy with novel antimyeloma agents and to explore possible correlations with clinical and laboratory data, including survival.

Patients and Methods


We evaluated 174 consecutive, newly diagnosed, symptomatic myeloma patients who received therapy with novel agents (thalidomide, bortezomib or lenalidomide), between June 2005 and July 2008 within the Greek Myeloma Study Group (GMSG). Patients' characteristics are depicted in Table 1.

Table 1. Patients' characteristics
inline image

The study was approved by the local Ethical Committees in accordance with the Declaration of Helsinki and the Good Clinical Practice. All patients provided informed consent for serum sample collection and analysis of the data.

Measurement of circulating angiogenic cytokines

Circulating levels of Ang-1, Ang-2, VEGF, VEGF-A (the major angiogenesis component of VEGF), angiogenin and bFGF were evaluated in all patients, using ELISA methodology (R&D Systems, Minneapolis, MN, for all, except VEGF-A: Diaclone, Bensancon, France), as previously described.13 All measurements were performed in a central laboratory of the GMSG.

Collection of serum samples for the evaluation of the above molecules was performed within 1 week before the initiation of any antimyeloma treatment, and the serum samples were stored at −70°C until the day of measurement.

These cytokines were also measured in 34 healthy individuals of similar gender and age (19 M/15 F; median age: 66 years, range: 39–82 years), who served as controls. Their medical history was reviewed to ensure that there was no disease that could alter angiogenesis at the time of sampling (cardiovascular disorder, inflammatory disease, renal impairment, infection, etc.).

Angiopoietins were also measured in 25 patients (15 M/10 F; median age: 67 years, range: 48–72 years) with end-stage chronic kidney disease (ESCKD), on dialysis. The serum was collected on the day of dialysis, just before the start of the procedure.

Measurement of angiopoietin-2 and microvessel density in trephine biopsies

In 20 patients, 10 with the higher values of circulating Ang-2 (median 14.8 ng/ml, range: 12.1–17.9 ng/ml) and 10 with the lower values of circulating Ang-2 (median: 0.52 ng/ml, range: 0.1–0.92 ng/ml) we also performed immunohistochemical evaluation of their bone marrow trephine biopsies for the expression of Ang-2. Biopsies had been obtained at the time of diagnosis. The expression of Ang-2 was assessed using an anti-angiopoietin-2 monoclonal antibody by Santa Cruz Biotechnology, Santa Cruz, CA. The immunoreactivity of Ang-2 was examined on the basis of positive plasma cells with a cutoff value of >20% positive cells to be defined as positive expression.

Immunohistochemical identification of microvascular endothelial cells was also performed in the same trephine biopsies using a human monoclonal antibody against CD34 (DAKO A/S, Glostrup, Denmark) by two experienced pathologists at ×400 magnification. In each biopsy sample, microvessels were counted in at least three independent hot spots per section, and the microvessel density (MVD) of a bone marrow specimen was calculated as the mean value of all independent readings and recorded as the number of microvessels per ×400 field. When the microvessel count was 1–2, angiogenesis was characterized as low grade, while intermediate grade angiogenesis was defined by the presence of a microvessel count of 3–6 and high-grade angiogenesis by the presence of microvessel count ≥7.

Statistical analysis

Descriptive statistics were used to characterize baseline variables. Mann–Whitney test was used to evaluate differences between patients and controls, whereas the Spearman Rank correlation test determined correlations between clinical and laboratory parameters. One-way analysis of variance (ANOVA) was used to test for differences between the groups, if there were more than two groups to compare. Survival probabilities were calculated by the Kaplan–Meier method and comparisons made using the log-rank test to identify potential prognostic factors. Cox regression models were used to assess hazard ratios for survival. Potential correlations between baseline characteristics and survival were investigated in multivariate regression models. A reduced model was generated by backward elimination of the variable with the highest p value from each iteration of the multivariate analysis until only significant variables (0.05 level) remained.


Circulating levels of angiogenic cytokines in patients and controls

Myeloma patients before treatment had elevated serum levels of Ang-2 (p = 0.002; n = 174, Fig. 1a), angiogenin (p = 0.001; n = 174, Fig. 1b), VEGF (p = 0.002; n = 174) and VEGF-A (p = 0.002; n = 174) compared with healthy controls (Table 2). Ang-1 levels were reduced in MM patients compared to controls (p = 0.05; n = 174); thus the ratio of Ang-1/Ang-2 was reduced in MM (p < 0.001). Ang-1/Ang-2 ratio correlated with beta2-microglobulin (r = −0.24, p = 0.002; n = 171), age (r = −0.227, p = 0.003; n = 174), hemoglobin (r = 0.256, p = 0.001; n = 165), platelet counts (r = 0.222, p = 0.004; n = 169) and creatinine (r = −0.23, p = 0.002; n = 173). Patients with ISS-3 had lower Ang-1/Ang-2 ratio (mean value 7.76, median 5.17, range: 0.1–55.9; n = 58) compared with ISS-2 (mean 10.6, median 5.9, range: 1.23–55.8; p = 0.031; n = 61) and ISS-1 patients (mean 14.3, median 7.2, range: 1.26–121.44; p = 0.032; n = 55; Fig. 1c). Patients with lytic bone disease (n = 144) had increased circulating Ang-2 (mean 4.5 ng/ml; median 4.17 ng/ml, range: 2.81–17.9 ng/ml) compared with those who had no lytic lesions in the skeletal survey (mean 3.89 ng/ml; median 3.12 ng/ml, range: 0.1–8.35 ng/ml; p = 0.048). Patients with MM who had increased serum creatinine (>2 mg/dl; n = 24) had no different values of Ang-1, Ang-2 and Ang-1/Ang-2 ratio compared with all others (p = 0.257, p = 0.216, p = 0.181, respectively; see also Table 2). Patients with ESCKD had higher levels of Ang-2 and lower levels of Ang-1 and Ang-1/Ang-2 ratio compared to healthy controls (p < 0.001 for all comparisons; see also Table 2). Patients with MM had increased levels of Ang-1 (p = 0.041) and Ang-1/Ang-2 ratio (p < 0.001) but reduced levels of Ang-2 (p < 0.001) compared to ESCKD patients. When we compared MM patients with increased serum creatinine (>2 mg/dl) with ESCKD patients, we found that MM patients had higher Ang-1/Ang-2 ratio (p = 0.001) due to the lower levels of Ang-2 (p < 0.001) and no differences in Ang-1 levels (p = 0.456).

Figure 1.

Patients with newly-diagnosed, symptomatic multiple myeloma have increased circulating levels of Ang-2 (a) and angiogenin (b) that correlate with ISS stage: p ANOVA = 0.032 for Ang-2 (c) and p ANOVA = 0.004 for angiogenin (d). Ang-2 positive expression by malignant plasma cells in the patient with the highest Ang-2 circulating levels (17.9 ng/ml; magnification ×400, (e) and Ang-2 negative expression in the patient with the lowest circulating Ang-2 value (magnification 250×, f).

Table 2. Circulating levels of angiogenic cytokines (median, range, mean ± SD) in patients and controls
inline image

Angiogenin showed strong correlations with beta2-microglobulin (r = 0.249, p = 0.001; n = 171), serum creatinine (r = 0.24, p = 0.001; n = 173) and hemoglobin (r = −0.265, p = 0.001; n = 165). Furthermore, patients with ISS-3 had increased levels of angiogenin compared to patients with ISS-2 and ISS-1 (p ANOVA = 0.004; Fig. 1d). VEGF circulating levels correlated with Ang-2 (r = 0.317, p < 0.001; n = 174), bFGF (r = 0.348, p < 0.001; n = 174) and lactate dehydrogenase (LDH) (r = 0.188, p = 0.02; n = 171) levels.

Angiopoietin-2 expression by myeloma cells

All ten patients with the higher levels of circulating Ang-2 had positive expression of Ang-2 by myeloma cells (range: 60–100% of the total myeloma cell population; Fig. 1e). All these patients had high grade angiogenesis, as assessed by MVD evaluation (median: 13, range 7–21). On the contrary, in 7/10 patients with the lower circulating levels of Ang-2, myeloma cells were negative for Ang-2 expression (Fig. 1f). These patients had either low or intermediate grade angiogenesis based on MVD (median: 3, range 1–6).

Angiopoietins and survival

In this study, 82 patients (47%) received thalidomide-based regimens and 31 (18%) received bortezomib-based regimens as first-line therapy, whereas 61 (35%) received conventional chemotherapy (CC: VAD or MP) and 29 patients (17%) received high dose melphalan with autologous stem-cell support. No patient received lenalidomide as first-line therapy. All patients received regimens containing bortezomib or an immunomodulatory drug at some point during the course of their disease.

The median survival of the patients was 47 months. The median follow-up period was 26 months. The univariate analysis revealed that ISS stage (p < 0.001), serum LDH (p = 0.003), age (p = 0.02), bone disease status (p < 0.001), serum creatinine (p = 0.017), and the ratio of Ang-1/Ang-2 predicted for overall survival. Patients with serum Ang-1/Ang-2 below or equal to the median value of 6.03 had a median survival of 26.3 months, whereas patients with Ang-1/Ang-2 values above this median value had a median survival of 53 months (p = 0.002; Fig. 2a). Interestingly, this survival difference was mainly observed in patients who received novel agent (thalidomide or bortezomib)-based regimens as first-line therapy (Fig. 2b) but not in those who had been given conventional chemotherapy (CC) (with or without high dose therapy) in the frontline setting (data not shown).

Figure 2.

Myeloma patients who had low levels of Ang-1/Ang-2 ratio had shorter overall survival (a). Those who received thalidomide- or bortezomib- based regimens as frontline therapy and had a low Ang-1/Ang-2 ratio had a median survival of only 22 months (b). Ang-1/Ang-2 increased ratio defines a population in ISS-3 with very poor prognosis (c): these patients had a median age of 70 years (range 50-94 years, n = 29), which was slightly higher than that of the others (n = 29, median age 67, range:44–81 years). There were no significant differences among these two groups of patients of ISS-3 (with high or low Ang-1/Ang-2 ratio) regarding bone disease status, serum creatinine, LDH, hemoglobin, platelet counts, albumin, beta2-microglobulin, VEGF, VEGF-A, bFGF, and angiogenin.

Furthermore, the ratio of Ang-1/Ang-2 could identify a subset of ISS-3 patients with very poor prognosis: patients with serum Ang-1/Ang-2 below or equal to the median value of 5.17 (for ISS-3 patients) had a median survival of 18 months, while patients with Ang-1/Ang-2 values above 5.17 had a median survival of 45 months (p = 0.0001; Fig. 2c). We performed the same analysis for ISS-1 and ISS-2 patients, where Ang-1/Ang-2 either as a median or a mean cutoff value could not identify a poor risk subgroup (data not shown).

The multivariate analysis included all factors that had prognostic significance in the univariate analysis and revealed that only high LDH (RR 1.96, 95% CI 1.18–2.40; p = 0.002), low Ang-1/Ang-2 ratio (RR 2.07, 95% CI 1.50–2.42; p < 0.001) and extensive bone disease (RR 1.28, 95% CI 1.08–1.51; p < 0.001) could independently predict for inferior survival in our cohort of patients. When we included in the analysis, the type of first-line therapy (CC vs. novel agent-based therapies), which had a borderline prognostic significance (p = 0.057) in the univariate analysis, then the same parameters continued to have independent prognostic significance, i.e., high LDH, low Ang-1/Ang-2 ratio and extensive bone disease.


The angiopoietins represent a major family of angiogenic factors. The two best characterized angiopoietins, Ang-1 and Ang-2, are ligands for the Tie-2 receptor, which is present on endothelial cells and endothelial progenitor cells. Ang-2 functions to block the Ang-1/Tie-2 signaling, which serves to inhibit endothelial cell activation. Thus, Ang-2 facilitates endothelial activation in response to inducers of angiogenesis, such as VEGF.4 In our study, we found that circulating Ang-2 is elevated in newly diagnosed, symptomatic myeloma patients, whereas Ang-1 is reduced; thus, the ratio of Ang-1/Ang-2 is reduced and correlates with adverse disease features, including ISS-stage, extensive bone disease and renal impairment. There are few reports on the serum levels of Ang-1 and Ang-2 in plasma cell dyscrasias. Circulating soluble Ang-2 was increased in 24 MM patients at different phases of the disease,14 while our group has reported reduced Ang-1/Ang-2 ratio in patients with relapsed/refractory myeloma,7, 15 with Waldenström's Macroglobulinemia13 and with systemic light chain amyloidosis.16 This study is the largest that evaluated these angiogenic cytokines in newly diagnosed, previously untreated, myeloma patients. As in relapsed/refractory myeloma, we found that the angiopoietin balance was disrupted in favor of Ang-2 in newly diagnosed patients, possibly leading to vessel destabilization and to angiogenic sprouting. The expression of Ang-2 by myeloma cells in patients with high circulating levels and the increased MVD in their marrow biopsies further supports this notion. The reversal of the normal angiopoietin balance in favor of Ang-2 has been also reported in other active hematologic malignancies, including acute myeloblastic leukemia (AML).17 Furthermore, elevated circulating Ang-2 predicted for poor survival in AML patients.18 In our study, reduced level of Ang-1/Ang-2 ratio independently predicted for short survival in myeloma patients. Increased angiogenesis, assessed by MVD in the bone marrow trephine biopsies, correlated with poor survival in myeloma patients treated with CC.1, 19 In our study, increased MVD was present in patients with very high Ang-2 circulating levels. This is the first study, which shows that the Ang-1/Ang-2 ratio is an independent prognostic factor for survival in myeloma patients who received novel agent-based therapies. Furthermore, in the univariate analysis, the effect of Ang-1/Ang-2 ratio was pronounced in patients who were treated upfront with novel agent-based therapies. This is important as bortezomib or thalidomide exert antiangiogenic properties8–10 and it was expected that they could override the adverse prognosis that angiogenic cytokines may have. A possible explanation is that these cytokines may reflect tumor burden. This hypothesis is further supported by the observation that Ang-1/Ang-2 ratio excluded ISS myeloma stage from the multivariate model for survival. Two studies have explored the production of angiopoietins by myeloma cells with conflicting results. In the first study, Ang-1, but not Ang-2, messenger ribonucleic acid (mRNA) and protein were expressed by purified MM cells from approximately 47% of the studied myeloma patients, as well as by myeloma cells of several myeloma cell lines.5 On the contrary, in the second study, expression of Ang-2 by malignant plasma cells was observed in 27/36 (75%) myeloma samples, whereas there was no expression of Ang-1 by myeloma cells. Furthermore, in 18 samples a coexpression of VEGF and Ang-2 was detected.6 In our study, we found Ang-2 expressed by myeloma cells in a subset of myeloma patients with very high-circulating Ang-2 levels. Furthermore, a strong correlation between VEGF and Ang-2 circulating levels was observed. Since VEGF and angiopoietin-2 were reported to work coordinately in angiogenesis,4 it is possible that these molecules contribute in a synergistic manner to the enhancement of angiogenesis in the bone marrow of myeloma patients.

Ang-2 was positively correlated with creatinine in our MM patients. Patients with ESCKD had also elevated circulating Ang-2 even compared to MM patients with high creatinine (>2 mg/dl). This correlation may be another reason for the independent prognostic value of Ang-1/Ang-2 ratio as it may reflect both tumor burden and renal impairment. An important issue is the alterations of Ang-1/Ang-2 levels post therapies with novel agents in newly diagnosed patients. We have shown that the ratio is increased postbortezomib therapy in relapsed/refractory myeloma and this correlates with better survival.7 The effects of other novel agents on Ang-1/Ang-2 ratio, mainly in the frontline setting, are eagerly anticipated.

Increased circulating levels of VEGF and of angiogenin have been previously described in patients with MM20–22 and our results confirm these observations. The effect of circulating VEGF on survival of myeloma patients who received CC was previously reported but the results were conflicting.20, 23 In our study, circulating VEGF had no prognostic value; this may be due to the reduction of VEGF by novel agent-based therapies,11, 22, 24, 25 which may override a possible adverse effect of VEGF on prognosis.

We conclude that circulating Ang-1/Ang-2 is reduced in patients with newly diagnosed, symptomatic myeloma and correlates with advanced disease features and poor survival. These results highlight the role of angiopoietins pathway in the biology of MM and reveal Ang-1/Ang-2/Tie-2 system as a possible target for the development of novel antimyeloma agents.