Overexpression of vascular endothelial growth factor (VEGF) - Nature

1 downloads 66 Views 287KB Size Report
Vascular endothelial growth factor (VEGF) and its cellular receptor VEGFR-2 ... antiangiogenic and antileukemic treatment strategies in AML. Leukemia (2002) .... strated,30,38 the anti-VEGF antibody used in this study is spe- cific for VEGF-A ...
Leukemia (2002) 16, 1302–1310  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu

Overexpression of vascular endothelial growth factor (VEGF) and its cellular receptor KDR (VEGFR-2) in the bone marrow of patients with acute myeloid leukemia T Padro´1, R Bieker1, S Ruiz1, M Steins1, S Retzlaff1, H Bu¨rger2, T Bu¨chner1, T Kessler1, F Herrera1, J Kienast1, C Mu¨ller-Tidow1, H Serve1, WE Berdel1 and RM Mesters1 1

Department of Medicine/Hematology and Oncology, University of Muenster, Muenster, Germany; and 2Gerhard-Domagk Institute of Pathology, University of Muenster, Muenster, Germany

Vascular endothelial growth factor (VEGF) and its cellular receptor VEGFR-2 have been implicated as the main endothelial pathway required for tumor neovascularization. However, the importance of the VEGF/VEGFR-2 system for angiogenesis in hematologic malignancies such as AML remains to be elucidated. In 32 patients with newly diagnosed untreated AML, we observed by immunohistochemical analysis of bone marrow biopsies significantly higher levels of VEGF and VEGFR-2 expression than in 10 control patients (P ⬍0.001). In contrast, VEGFR-1 staining levels in AML patients were in the same range as in the controls. Expression of VEGF and VEGFR2 was significantly higher in patients with a high degree of microvessel density compared to those with a low degree (VEGF: P ⴝ0.024; VEGFR-2: P ⴝ0.040) and correlated well with bone marrow microvessel density (rs ⴝ0.566 and 0.609, respectively; P ⬍0.001). Furthermore, in patients who achieved a complete remission following induction chemotherapy VEGFR-2 staining levels decreased into the normal range. In conclusion, our results provide evidence for increased expression of VEGF/VEGFR-2 of leukemic blasts and correlation with angiogenesis in the bone marrow of AML patients. Thus, VEGF/VEGFR-2 might constitute promising targets for antiangiogenic and antileukemic treatment strategies in AML. Leukemia (2002) 16, 1302–1310. doi:10.1038/sj.leu.2402534 Keywords: vascular endothelial growth factor receptor-2, angiogenesis; acute myeloid leukemia

Introduction Angiogenesis – a complex, multistep process by which new microvessels are formed from the preexisting vasculature – is an absolute requirement for the viability and growth of solid tumors.1 Emerging data suggest an involvement of angiogenesis in the pathophysiology of hematologic malignancies, as well. Recently, we and others have reported increased angiogenesis in the bone marrow of patients with acute myeloid leukemia (AML)2–4 and normalization of bone marrow microvessel density when patients achieved a complete remission (CR) after induction chemotherapy.3 Similarly, increased bone marrow microvessel density was described in adult patients with myelodysplastic syndromes (MDS) and multiple myeloma,4–7 as well as in children with acute lymphoblastic and adults with B cell chronic lymphocytic leukemias.8,9 Thus, angiogenesis may be one potential target for new treatment strategies in AML.10 Tumor angiogenesis depends on the expression of specific mediators that initiate the cascade of events leading to the formation of new microvessels.11 Among these, vascular endothelial growth factor type A (VEGF) plays a pivotal role in the induction of neovascularization in solid tumors.12

Expression of VEGF is associated with angiogenesis, tumor aggressiveness or poor prognosis in several human tumors, including breast, colon, lung, gastric, renal and oropharyngeal cancers.13–20 The two primary signaling receptor tyrosine kinases that mediate the various biologic effects of VEGF are VEGFR-1 (Flt1) and VEGFR-2 (Flk-1/KDR).12 Homozygous gene knockout studies suggest that expression of VEGFR-2 is important for vasculogenesis and angiogenesis whereas VEGFR-1 may be more important for vascular remodeling.21,22 Furthermore, disruption of VEGFR-2 signaling resulted in inhibition of tumor growth and tumor metastasis in murine models23–25 and glioblastoma growth was inhibited in mice dominant-negative for VEGFR-2.26 Recent data suggest an important role for VEGF in hematologic malignancies, as well. It has been demonstrated that several human leukemic cell lines express VEGF.27,28 Isolated AML blasts and B cells in chronic lymphocytic leukemia have been found to produce and secrete VEGF.27,29 A recent report has shown that expression of VEGF is restricted to myeloblasts and immature myeloid elements in various MDS subtypes30 and VEGF has been found to be elevated in patients with idiopathic myelofibrosis.31 Furthermore, in a series of 99 patients with newly diagnosed AML, Aguayo et al32 found that increased cellular VEGF is an independent poor prognostic factor in patients with AML. VEGF type C, another member of the vascular endothelial growth factor family, is expressed by leukemic blasts in a significant proportion of patients with AML.33 VEGFR-2 was thought to be exclusively expressed by adult endothelial cells. However, it has recently been shown that VEGFR-2 is present on a subset of primitive hematopoietic stem cells34 as well as on some leukemic cell lines and isolated blasts from AML patients.27,28,35,36 In contrast, Bellamy et al30 were unable to detect VEGFR-2 expression in AML blasts performing immunohistochemistry on bone marrow clot biopsies. Thus, the role of VEGFR-2 expression in AML remains controversial. The goal of our study was to investigate the expression of VEGF/VEGF-receptors and its relation to angiogenesis in the bone marrow of adult patients with newly diagnosed untreated AML. In addition, we evaluated the expression of VEGF and its two receptors in the bone marrow of AML patients following induction chemotherapy. Materials and Methods

Bone marrow specimens Correspondence: RM Mesters or WE Berdel, Department of Medicine/Hematology and Oncology, University of Muenster, AlbertSchweitzer-Strasse 33, D-48129 Muenster, Germany; Fax: 49/251/83 48745 Received 24 September 2001; accepted 18 February 2002

Thirty-two patients (21–78 years) were randomly chosen from a cohort of 62 patients with newly diagnosed, untreated AML of whom the degree of bone marrow angiogenesis at presentation has previously been reported.3 In order to have a rep-

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

resentative subgroup of the original group of 62 patients, randomization was adjusted to the frequency distribution of the most common AML subtypes (M1, M2, M4, M5) according to the definition of the French–American–British (FAB) Cooperative Group.37 Due to the low frequency in the original cohort, FAB types M0, M3, M6 were excluded from randomization to avoid potential bias. Ten unselected controls were taken from an initial group of 22 patients which have been analyzed for bone marrow microvessel density, as well.3 These controls consist of adult patients (17–77 years) with various diseases but with normal bone marrow morphology as demonstrated by cytological and histological analyses. From all AML and control patients, a bone marrow core biopsy (iliac crest) for histological diagnosis was obtained at presentation (day 0). From the subgroup of eight AML patients, additional biopsies were available on day 16 of induction chemotherapy (hypoplastic bone marrow) and at the time of CR. After every core biopsy, a bone marrow aspiration was obtained through a separate puncture for cytologic analyses.

Immunohistochemical staining Bone marrow specimens were fixed in paraformaldehyde, decalcified with EDTA, and embedded in paraffin. Serial sections (4-␮m-thick) of each sample were processed immunohistochemically for the expression of VEGF and its receptors with rabbit polyclonal anti-human VEGF (sc-152; Santa Cruz Biotechnology, Santa Cruz, CA, USA; working dilution 1:2000) and anti-human VEGFR-1 antibodies (sc-316, Santa Cruz Biotechnology; working dilution 1:400) and with a mouse monoclonal anti-human VEGFR-2 antibody (sc-6251, Santa Cruz Biotechnology; working dilution 1:50). As described by the manufacturers and previously demonstrated,30,38 the anti-VEGF antibody used in this study is specific for VEGF-A and does not cross-react with other known VEGF/placental growth factor family members. The VEGFR-1 and VEGFR-2 antibodies are also specific and do not crossreact with each other or with other receptor tyrosine kinases. Controls for immunostaining using non-immune mouse IgG or rabbit IgG (sc-2025 and sc-2027; respectively; Santa Cruz Biotechnology) in substitution for the specific first antibodies were consistently negative (data not shown). Immunohistochemical staining was performed by the alkaline phosphatase/antialkaline phosphatase double bridge technique (Dako-APAAP Kit; Dako, Glostrup, Denmark). Briefly, tissue sections were deparaffinized in xylene and rehydrated in a graded ethanol series. Samples were pretreated to promote antigen retrieval in a microwave oven at 400 W twice for 7 min in 10 mmol/l sodium citrate, pH 6.0 (Dako). The primary antibodies were applied overnight at 4°C. Subsequent steps were performed according to the manufacturer’s instructions. The fast red substrate (Dako) supplemented with 0.1% (w/v) levamisole was employed for revelation of phosphatase activity (30 min at room temperature). Sections were counterstained with 0.1% (w/v) erythrocin solution.

Evaluation of VEGF, VEGFR-1, and VEGFR-2 expression Immunostaining was simultaneously assessed by two independent experienced investigators using light microscopy. The investigators were blinded to the clinical characteristics of the patients and the bone marrow microvessel counts when per-

forming the growth factor evaluation. Expression of VEGF, VEGFR-1 and VEGFR-2 protein was semi-quantitatively assessed by scoring the proportion and intensity of stained cells according to published methods, with appropriate modifications.17,38–39 At first, the entire bone marrow section was systematically scanned at ⫻100 and ⫻250 magnification and the percentage of positive cells stained with each specific antibody within cellular areas was estimated according to a threegrade scale (1⬅ ⬍10% positive cells, 2⬅ 10–50% positive cells, 3⬅ ⬎50% positive cells). Subsequently, the intensity of positive staining was evaluated in three representative ⫻500 fields (0.126 mm2 field area) selected in each section after the initial screening at ⫻100–⫻250 magnification. The total area of these three ⫻500 fields represented 7.5 to 20% of the total cellular bone marrow cross-sectional area. The degree of cellular staining intensity within these areas was defined in the following manner: 1⬅ faint or negative, 2⬅ ⬍50% cells with moderate staining, 3⬅ ⬎50% cells with moderate staining, 4⬅ ⬍50% cells with intense staining, 5⬅ ⬎50% cells with intense staining. In order to achieve a reasonable assessment of the protein expression in the whole bone marrow section, the mean of cellular staining intensity from three representative ⫻500 fields was subsequently multiplied with the number obtained when estimating the percentage of positive cells according to the three grade scale in the entire bone marrow section at ⫻100 and ⫻250 magnification. The results were expressed as arbitrary units (AU). In each biopsy sample, expression of VEGF, VEGFR-1 and VEGFR-2 was evaluated in two to three sections processed in independent immunostainings and the mean value was calculated. The variability between the investigators for the immunostaining scores was 10.5% (interquartile range, IQR: 4.3–20.2%). Immunohistochemical staining for VEGF, VEGFR-1 and VEGFR-2 yielded reproducible results. Staining scores of three different ⫻500 fields from the same bone marrow section had a median coefficient of variation of 10.0% for VEGF (IQR: 5.0–20.0%), 2.5% for VEGFR-1 (IQR: 0.0–16.1%), and 10.5% for VEGFR-2 (IQR: 7.0–17.7%). Analysis of two to three marrow sections from a single biopsy specimen revealed a median intra-individual variability lower than 30% for each of the antibodies used in the study (median (IQR): VEGF, 18.3% (7.4–33.2%); VEGFR1, 10.1% (0.5–51.4%); VEGFR-2, 28.3% (15.0–44.2%)). To ensure the reliability of the quantification assay, marrow slides have been randomly selected during the study and reanalyzed with excellent agreement. A bone marrow specimen has been used as internal control in the study. Sections of this specimen have been repeatedly immunostained for the different antibodies and quantified by the two investigators at different times. As previously described for bone marrow microvessel countings,3 we evaluated signals only within the cellular areas of the marrow since the non-cellular areas (bone lamellae, fat and connective tissue areas, necrotic foci) are devoid of microvessels, which might hamper comparison between expression of VEGF or its receptors and the degree of angiogenesis in the marrow samples. Areas of staining adjacent to bone or dense connective tissue were also excluded, because vascularization is not representative of angiogenesis in these areas.3

1303

Quantification of leukemic blast infiltration and criteria for response to chemotherapy Quantitative analysis of leukemic blast infiltration was performed in bone marrow aspirates by routine cytologic analysis Leukemia

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

1304

as described by the FAB Cooperative Group.37 A CR was defined as a bone marrow with normal hematopoiesis of all cell lines, less than 5% blast cells, and a peripheral blood cell count with at least 1500 neutrophils/␮l and 100 000 platelets/␮l.40

Statistics Data are presented as individual data plots or as medians, interquartile ranges (low quartile–high quartile (LQ–HQ)), and ranges. Differences in expression of VEGF, VEGFR-1 or VEGFR-2 between AML and control groups as well as between AML patients with high and low microvessel densities were analyzed by the Mann–Whitney rank sum test for independent groups. Statistical significance of overall differences between more than two groups was analyzed by the Kruskal–Wallis one-way analysis of variance. Distribution of disease-free survival between patients with high and low levels of VEGF and its receptors was estimated by the Kaplan– Meier method41 and comparisons were based on the log-rank test.42 The Wilcoxon matched-pair signed rank test was used to compare protein expression in the bone marrow of individual patients at diagnosis and following induction chemotherapy. Correlation between variables or between variables and microvessel counts3 was assessed by the Spearman rank correlation coefficient (rs). Two-sided P values lower than 0.05 were considered significant, corrections for the number of comparisons (Bonferroni test) were performed when required.

staining (⬍10% total cellular area). As shown in Figure 1a, VEGF signals were seen primarily in the cytoplasm of leukemic blasts and as diffuse staining in the intercellular spaces. Capillaries positive for VEGF, distributed in hot-spots, were occasionally found (not shown). In contrast to the AML specimens, a heterogeneous spotted pattern of cell-associated VEGF antigen staining was observed throughout the cellular areas of the control bone marrow samples (Figure 1b). Both in AML and control specimens, megakaryocytes were also stained with anti-VEGF antibodies. However, these were easily recognized by their characteristic size and morphology. VEGF protein staining scores of the 32 bone marrow specimens from AML patients ranged from 1.1 to 9.0 AU with a median value of 4.2 AU (LQ–HQ: 3.1–6.1), whereas control

Results Expression of VEGF and its receptors VEGFR-1 and VEGFR2 was evaluated by immunohistochemical staining in bone marrow core sections of 32 AML patients at presentation (day 0). Patient characteristics are depicted in Table 1. Results from these AML patients were compared with the degree of protein expression in bone marrow biopsies obtained at diagnosis from 10 patients with different diseases but with a normal bone marrow morphology. In the cases of non-Hodgkin’s lymphoma and kidney carcinoma, the bone marrow was histologically not involved by the underlying disease (Table 1). VEGF antigen was consistently detected in every bone marrow specimen from AML as well as control patients. Areas of homogeneous positive staining for VEGF were widely distributed in cellular regions of the bone marrow from AML patients at presentation. Thirty patients exhibited extensive VEGF staining, whereas only two patients had limited areas of VEGF

Table 1

Patient characteristics

Age (years)a Sex (males/females) FAB distributionb Percentage of leukemic blastsa (bone marrow) Disease

a

Median (range). French–American–British classification for AML.37

b

Leukemia

Figure 1 Immunohistochemical staining of bone marrow sections obtained at presentation from AML and control patients for VEGF (a, b), VEGFR-1 (c, d) and VEGFR-2 (e, f). Immunohistochemical localization was performed by the respective specific antibodies and the alkaline phosphatase/anti-alkaline phosphatase technique (DakoAPAAP Kit; Dako). Note the overexpression of VEGF and VEGFR-2 and the homogeneous distribution of signals in the marrow sections of AML patients (a, e) compared to the controls (b, f). In contrast, similar staining patterns and intensity were observed for VEGFR-1 in the bone marrow of AML (c) and control patients (d). Original magnification ⫻500.

AML patients (n ⫽ 32)

Control patients (n ⫽ 10)

64 (21–78) 18/14 5 M1, 13 M2, 6 M4, 8 M5 83% (50–90) —

54 (17–77) 7/3 — — 1 non-Hodgkin’s lymphoma, 1 kidney carcinoma, 8 non-malignant disorders

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

bone marrow samples (n⫽10) showed significantly weaker VEGF protein staining (median, LQ–HQ: 2.9, 2.5–3.3 AU; Mann–Whitney test, P ⫽0.01; Figure 2a). Bone marrow of AML patients stained positive for the cell membrane-associated receptor tyrosine kinases VEGFR-1 and VEGFR-2 (Figure 1c, 1e). However, the intensity of the staining for VEGFR-1 and VEGFR-2 was different. Nine patients (28%) had a very faint or no staining at all for VEGFR-1 (score ⬍1.5 AU; Figure 2b). In contrast, 31 out of 32 patients displayed a strong staining, but only one patient had a weak staining for VEGFR-2 (score 1.4 AU; Figure 2c). Positive staining was cytoplasmatic with membrane accentuation for VEGFR-2. Expression of VEGFR-1 protein was heterogeneously distributed within the cellular areas of the marrow sections with a similar pattern of cellular staining as the controls. A different staining pattern emerged for VEGFR-2. Positive signals were widely and uniformly observed throughout the cellular regions of the AML marrows. Overall, there were more VEGFR-2-positive than VEGFR-1-positive areas. Accordingly, VEGFR-1 protein expression in the bone marrow of AML patients displayed low staining scores (median, (LQ– HQ): 1.8 (1.4–2.7) AU) and did not significantly differ from those found in bone marrow sections of control patients (2.7 (1.6–3.6) AU; Mann–Whitney test, P ⫽0.224; Figures 1c, d and 2b). In contrast, bone marrow sections in AML patients had significantly higher VEGFR-2 scores than the control group (median (LQ–HQ): AML, 3.8 (2.3–6.3) AU; controls, 1.8 (1.3–1.9) AU; Mann-Whitney test, P ⬍0.001; Figures 1e, 1f and 2c). Bone marrow in the AML patients studied was usually highly infiltrated by blast cells. The median (LQ-HQ) percentage of blasts was 83% (50–90). Although VEGF and its receptors were consistently detected in leukemic blasts of the bone marrow, staining scores for VEGF, VEGFR-1 or VEGFR-2 in the biopsies of AML patients did not correlate with the percentage of blasts determined in the marrow aspirates (VEGF: rs ⫽0.083; VEGFR-1: rs ⫽⫺0.241; VEGFR-2: rs ⫽0.284, P ⬎0.05 for each variable). Median (LQ-HQ) staining scores for VEGF, VEGFR-1, and VEGFR-2 in relation to the AML FAB-subtypes are shown in

Table 2. Statistical analyses did not reveal significant differences in VEGF nor in its receptors between the four AML subtypes considered for analysis (Table 2). From each bone marrow specimen, adjacent sections to those processed in the present study have previously been immunostained with specific markers of endothelial cells (anti-human von Willebrand factor and anti-human thrombomodulin antibodies) for the assessment of microvessel density (MVD) as reported.3 The median MVD in the 32 AML patients was 22.4 (LQ–HQ: 19.2–28.8) microvessels/⫻500 field and 12.2 (10.6–14.3) microvessels/⫻500 field in the 10 control patients, respectively. The MVD corresponds to the average number of microvessels counted in the field area at ⫻500 magnification (0.126 mm2) using light microscopy.3 Thus, the MVD values for the randomly chosen subgroups were not significantly different from those previously reported in the entire population of 62 AML and 22 control patients.3 The expression of VEGF and VEGFR-2 in the bone marrow was related to the MVD levels in the group of AML patients as well as in the total study population (32 AML and 10 control patients). Bone marrows of AML patients with strong angiogenesis (76–100% quartile of the population) showed higher median staining scores for VEGF and VEGFR-2 than AML marrows with low MVD (1–25% quartile) (marrows with ⬎28.8 microvessels/⫻500 field vs marrows with ⬍19.2 microvessels/⫻500 field: VEGF, 4.6 AU vs 2.4 AU, P ⫽0. 018; VEGFR-2, 5.4 AU vs 2.6 AU, P ⫽0.036, Mann–Whitney test; Figure 3a, c). In contrast, median scores for VEGFR-1 did not significantly differ between AML patients with high and low bone marrow MVD (2.0 AU vs 1.8 AU, P ⫽0.874 Mann– Whitney test; Figure 3b). Furthermore, there was a positive correlation of VEGF expression in the bone marrow with the degree of MVD (rs ⫽0.566 for 32 AML and 10 control patients, P ⬍0.001; Figure 4a). Microvessel counts also correlated with the staining scores for VEGFR-2 (rs ⫽0.609, P ⬍0.001; Figure 4c), but not with VEGFR-1 (rs ⫽⫺0.094, P ⬎0.05; Figure 4b). From the 32 AML patients investigated, 22 achieved a CR after receiving standard induction chemotherapy according to the AMLCG protocol.43 Expression of VEGF or its receptors VEGFR-1 and VEGFR-2 in the bone marrow of these patients at presentation did not differ significantly from the levels of VEGF, VEGFR-1 and VEGFR-2 in the bone marrow of patients not achieving a CR (median (LQ–HQ): VEGF, 4.7 (3.3–6.5) vs

1305

Table 2 VEGF, VEGFR-1 and VEGFR-2 staining scores according to the AML subtype

VEGF (AU)a

Figure 2 Expression of VEGF and VEGF-Receptors in the bone marrow of AML and control patients. Adjacent sections of bone marrow biopsies from 32 AML patients at presentation and 10 controls were stained in parallel for VEGF (a), VEGFR-1 (b) and VEGFR-2 (c). Staining scores for each protein were calculated as described in Methods. Data are presented as individual values (open circles) and medians (horizontal bars). VEGF and VEGFR-2 levels (arbitrary units, AU) were significantly higher in AML than in control patients (P ⫽0.01 and P ⬍0.001, respectively; Mann–Whitney test). No differences were found between AML patients and controls for VEGFR-1 (P ⫽0.224).

AML subtypeb M1 M2 M4 M5 Control patients

(n ⫽5) (n ⫽13) (n ⫽6) (n ⫽8) (n ⫽10)

2.9 4.0 6.6 4.1 2.9

(2.2–4.4) (3.3–4.9) (5.7–7.0) (3.7–5.6) (2.5–3.3)

VEGFR-1 (AU)a

1.4 2.7 1.9 1.8 2.7

VEGFR-2 (AU)a

(1.0–1.5) 5.3 (4.7–5.9) (1.5–3.5) 4.4 (2.9.–7.0) (1.6–2.1) 2.7 (2.2–3.8) (1.5–2.2) 3.6 (2.3–4.1) (1.6–3.6) 1.6 (1.3–1.9)

a Values represent medians and interquartile ranges of arbitrary units (AU). Analysis of significance for overall differences between AML subtypes was performed by the Kruskal–Wallis test (VEGF: P ⫽0.090; VEGFR-1: P ⫽0.076; VEGFR-2: P ⫽0.129) b AML subtype according to the French–American–British classification for AML.37

Leukemia

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

1306

Figure 3 VEGF, VEGFR-1, and VEGFR-2 expression in the bone marrow of AML patients with low and high microvessel densities. Microvessel density (MVD) in AML bone marrow sections has been quantified in a previously reported study.3 From the total of 32 AML patients, two subgroups with different microvessel densities were selected. Low MVD denotes a group of AML patients (n ⫽8) with low number of microvessel counts in the bone marrow (1–25% quartile of the entire population ⬅ ⬍19.2 microvessels/⫻500 field) and high MVD denotes a group of AML patients (n ⫽8) with microvessel counts within the highest quartile (76–100%) of the total population (⬎28.8 microvessels/⫻500 field). Levels of VEGF (a), VEGFR-1 (b) and VEGFR-2 (c) in these groups are presented as individual values (open circles) and medians (horizontal bars). VEGF and VEGFR-2 expression was significantly elevated in the high MVD group compared to the low MVD group (VEGF: P ⫽0.018; VEGFR-2: P ⫽0.036; Mann–Whitney test). VEGFR-1 staining levels did not differ significantly between the low and high MVD groups (P ⫽0.874).

3.9 (2.2–4.7) AU; VEGFR-1, 1.8 (1.3–1.5) vs 1.5 (1.3–3.3) AU; VEGFR-2, 3.9 (2.5–5.2) vs 2.3 (2.1–4.7) AU; Mann–Whitney test: P ⫽0.07, P ⫽0.48 and P ⫽0.23, respectively). Furthermore, using the median staining scores as cut-off level (VEGF: 4.2 AU; VEGFR-1: 1.8 AU; VEGFR-2: 3.8 AU), neither VEGF nor VEGF receptor expression was associated with diseasefree survival as evaluated by log-rank analysis of Kaplan– Meier plots comparing the patient group with VEGF, VEGFR1, and VEGFR-2 levels below the cut-off with the patient group with protein expression levels above the cut-off (P⫽0.189, P ⫽0.942 and P ⫽0.741, respectively). To study the effect of chemotherapy on the expression of VEGF, VEGFR-1 and VEGFR-2, follow-up biopsies from eight AML patients were available for the studies on day 16 of induction chemotherapy using the TAD-protocol43 (standarddose cytarabine, daunorubicin and 6-thioguanine) and at the time of CR (Table 3). Residual cells in the hypoplastic bone marrows on day 16 of induction chemotherapy displayed a positive staining for VEGF. Intra-individual comparisons between biopsy specimens taken on day 16 and at presentation (day 0) did not reveal any significant difference in the VEGF staining scores (P ⫽0.09, Wilcoxon test; Table 3). VEGF expression was persistently elevated in the bone marrow biopsies obtained from the patients after having achieved a CR (CR vs day 0, P ⫽0.87, Wilcoxon test; Figure 5a). In spite of the persistence of high staining scores, changes in the VEGF distribution pattern were obvious in the latter samples. Bone marrow biopsies taken at CR showed an intense but heterogeneous and cell associated staining for VEGF (compare Figures 6a and 1a). The distribution pattern of VEGFR-1 in bone marrow biopsies of AML patients was unaffected by induction chemotherapy (see Figure 6b). Statistical analysis did not reveal any significant differences in the levels of VEGFR-1 Leukemia

Figure 4 Correlation of VEGF and VEGF receptor expression with microvessel density in the bone marrow of AML and control patients. As described in Figure 3, microvessel densities of bone marrow sections (32 patients with AML and 10 controls) have previously been determined.3 For each sample, microvessel counts were plotted against the immunostaining scores for VEGF (a), VEGFR-1 (b) and VEGFR-2 (c). Significance of the regression analysis was calculated by the Spearman test. Microvessel density significantly correlated with the levels of VEGF (rs ⫽0.566, P ⬍0.001) and VEGFR-2 expression (rs ⫽0.609, P ⬍0.001). No correlation was found between microvessel density and VEGFR-1 expression (rs ⫽⫺0.094).

Table 3 VEGF, VEGFR-1 and VEGFR-2 expression at presentation and following induction chemotherapy

VEGF (AU)a Day 0 Day 16 CR a

4.4 (3.2–4.9) 2.9 (2.6–3.9) 4.4 (3.8–5.1)

VEGFR-1 (AU)a VEGFR-2 (AU)a 1.3 (1.0–1.3) 1.3 (1.2–1.8) 2.0 (1.5–2.7)

4.2 (2.4–6.9) 1.3 (1.2–1.8) 1.3 (1.3–1.3)

Values represent medians and interquartile ranges of arbitrary units (AU). Differences in VEGFR-1 levels between bone marrows at presentation (day 0) and at complete remission (CR) were significant (P ⫽0.014; Wilcoxon test). VEGFR-2 levels were significantly reduced at day 16 and at the time of CR compared with levels at day 0 (P ⫽0.009; Wilcoxon test). Accepted significance level according to Bonferroni’s procedure: P ⭐0.025.

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

1307

Figure 5 Effects of induction chemotherapy on the expression of VEGF, VEGFR-1 and VEGFR-2 in the bone marrow of AML patients. Staining scores are depicted for presentation (day 0), and the time of complete remission (CR). Each pair of dots linked by a line represents the values from an individual patient. (a) VEGF expression did not differ significantly between day 0 and CR groups (P ⫽0.870, Wilcoxon test). (b) VEGFR-1 expression was significantly higher in the bone marrows at the time of CR compared with the values at presentation (P ⫽0.014, Wilcoxon test). (c) VEGFR-2 expression was significantly decreased at the time of CR compared with the values at presentation (P ⫽0.005, Wilcoxon test).

between biopsy specimens taken from the patients at presentation and on day 16 of induction chemotherapy (P ⫽0.100, Wilcoxon test; Table 3). However, VEGFR-1 expression was significantly higher in the bone marrows at CR (P ⫽0.014 for comparison with day 0, Wilcoxon test; Table 3). In contrast to VEGFR-1, levels of VEGFR-2 in the bone marrow significantly decreased from 4.2 AU at presentation to a median of 1.3 AU on day 16 of induction chemotherapy (P ⫽0.012; Table 3). Similarly, bone marrow biopsies obtained from patients after having achieved a CR displayed significantly lower VEGFR-2 levels (median: 1.3 AU) compared with initial presentation (P ⫽0.005; Figure 5b). Bone marrow sections were mainly negative for VEGFR-2 protein. Some weak VEGFR-2-positive cells only were scattered within the cellular areas of the bone marrows of the AML patients after having achieved a CR (compare Figures 6c and 1e). Bone marrows of patients at presentation showed higher VEGFR-2 than VEGFR-1 levels (Table 3, P ⬍0.001; Figure 1c and e). In contrast, bone marrow specimens obtained from the same patients after achieving a CR displayed an opposite pattern, with an excess of VEGFR-1 over VEGFR-2 (Table 3, P ⬍0.001; Figure 6b and c). Discussion The present study was prompted by the recent evidence of increased angiogenic activity in patients with AML.2–4,32 In contrast to solid tumors, in which VEGF and its cellular receptor VEGFR-2 have been implicated as the key endothelial pathway required for tumor neovascularization,26 few data about the importance of these angiogenic growth factor/receptors in the pathogenesis of hematologic malignancies have been reported so far. The current investigation, based on the immunohistological analyses of bone marrow biopsy samples, has clearly documented an elevated expression of VEGF and its receptor VEGFR-2 in leukemic blasts of patients with newly diagnosed, untreated AML. VEGF and VEGFR-2 expression have pre-

Figure 6 Immunohistochemical staining of the bone marrow sections with anti-VEGF, anti-VEGFR-1 and anti-VEGFR-2 antibodies at the time of complete remission (CR). Immunohistochemical localization was performed as described in Figure 1. (a, b and c) show representative areas of staining for VEGF, VEGFR-1 and VEGFR-2 in marrow sections of AML patients who have achieved a CR after induction chemotherapy. Note the strong cell-associated VEGF antigen staining in (a). Cells with intense cytoplasmic staining for VEGFR-1 were scattered throughout the bone marrow section (b). (c) shows the bone marrow of the same patient demonstrating a very low-intensity signal for VEGFR-2. Original magnification ⫻500.

viously been demonstrated in primary isolated blasts from AML patients as well as in AML cell lines.2,27,28,32 However, by the use of a semi-quantitative scoring system, our report is the first that demonstrates a significant increase of VEGF and VEGFR-2 levels in situ in the bone marrow of patients with AML compared with controls. Indeed, the bone marrow of 70% of the AML patients showed a 1.5- to 6.0-fold higher VEGFR-2 expression than the median of the control group. Although interpretation of these data is subject to the limitation of quantitative immunohistochemical methods, our blinded evaluation and several measures which have been taken to ensure the reliability of the quantification assay protect our results from diagnostic bias. Our findings are in line with reports of elevated VEGF protein in supernatants of primary AML blasts compared with normal mononuclear marrow cells27 and of high plasma4 and cellular32 VEGF levels in AML patients compared with control subjects. Furthermore, strong VEGFR-2 protein expression has recently been reported in ectopically implanted leukemias Leukemia

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

1308

Leukemia

(chloromas) in a murine model.36 Studies have demonstrated that VEGF inhibits chemotherapy-induced apoptosis in hematopoietic cells by inducing the antiapoptotic factor MCL144 and that VEGF stimulates proliferation of leukemic cells in vitro,30,36 an effect mediated through VEGFR-2.36 In their studies, Dias et al36 showed that neutralization of human VEGFR2 with specific monoclonal antibodies prolonged survival in mice xenotransplanted with human VEGFR-2-positive leukemic cell lines. Recently, our group has reported a stable remission after administration of the receptor tyrosine kinase inhibitor SU5416, targeting VEGFR-2 and the stem cell factor receptor c-kit, in a patient with AML relapse refractory toward standard chemotherapy regimens.45 However, SU5416 inhibits signalling via VEGFR-1, as well.46 Therefore, we cannot exclude that the inhibition of VEGFR-1 might have contributed to the observed remission. Together, these data suggest that increased VEGF and VEGFR expression in leukemic blasts play an important role in the pathophysiology of AML. VEGF secreted by AML blasts may not only support leukemic cell growth through a paracrine (by increasing the bone marrow endothelial cell mass) but also through autocrine mechanisms (increasing leukemic cell proliferation and promoting leukemic cell survival mediated by VEGFR-2). The observation that VEGF was in part associated with the cytoplasmic membrane of leukemic blasts may suggest binding of VEGF to its receptors, thus underscoring the autocrine VEGF/VEGFR2 pathway. The frequency of AML bone marrows positive for VEGFR2 (⬎95%) in our study clearly exceeded the frequencies of VEGFR-2 expression reported in marrow isolated blasts from AML patients (20–50%).27,36 The reasons for this apparent discrepancy may lie in the patient selection, in the method for detection, or in the different material used for analysis (isolated marrow blasts vs bone marrow core biopsies). Indeed, beside leukemic blasts other cellular components of the stromal bone marrow may account for increased VEGFR2 expression. A recent study based on the immunohistochemical analyses of lung carcinoma biopsies demonstrated that 75% of the stromal fibroblasts and ⬎90% of the stromal endothelial cells were positive for VEGFR-2.38 Bellamy et al30 investigated the expression of VEGFR-2 by immunohistochemistry of bone marrow clot biopsies of leukemic patients. They were unable to detect VEGFR-2 protein in a group of 17 patients with AML. This discrepancy with our study is most likely explained by the fact that those authors did not use antigen retrieval methodologies for immunohistochemical detection. Expression of VEGFR-1 has been documented within the hematopoietic compartment of normal bone marrow, mainly in monocytes–histiocytes, but also in rare myeloid elements.30 We observed strong VEGFR-1-positive cells scattered in cellular areas of normal bone marrows. The intensity and distribution pattern of the staining signals was not different in AML specimens. This observation does not seem to support a major role for the VEGF/VEGFR-1 signalling pathway in AML and is in line with results from a human leukemia model, in which monoclonal antibodies against human or murine VEGFR-1 had no effect on mice survival.47 However, the apparent similarity in VEGFR-1 expression between normal and leukemic bone marrow in our patient population does not exclude a potential role for VEGFR-1 in AML. It is entirely possible that certain AML subtypes or certain cell populations in AML depend on VEGFR-1 signalling. Future studies should address this issue. A compelling observation in the present study was the

association of VEGF and VEGFR-2 expression with the degree of neovascularization in the bone marrow of AML patients. Indeed, biopsy specimens with elevated microvessel counts exhibited significantly higher levels of VEGF and its receptor VEGFR-2 than those AML samples with low vascularization. In contrast, no association between marrow microvessel densities and VEGFR-1 levels was found. Furthermore, we could demonstrate a positive correlation between individual levels of VEGF and VEGFR-2 with the number of microvessels in the bone marrow of the total study population. These results support our hypothesis that VEGF secreted by AML blasts support leukemic cell growth by increasing microvessel density (paracrine mechanism). The increased endothelial cell mass could in turn enhance the proliferation of leukemic blasts by secreting growth factors like granulocyte–macrophage colonystimulating factor.27 The lack of correlation between expression of VEGF or its receptors and the bone marrow blast count at presentation may be due to the low variation of the percentage of blast infiltration in our study population (median: 83%; HQ–LQ: 50–90) or interindividual differences in expression of these factors by leukemic blasts. Several reports have suggested VEGF expression as a marker for poor prognosis in human solid tumors.13–20 Aguayo et al32 reported that increased cellular VEGF is an independent adverse prognostic factor in patients with AML, as well. However, our results suggest that neither bone marrow levels of VEGF nor those of its receptors at presentation predict the clinical outcome in terms of achieving a CR after standard chemotherapy or in terms of disease-free survival. The apparent discrepancy between these two studies may be explained by differences in the procedure for statistical analysis, the smaller sample size in our study, or the method of detection used. Indeed, in their study, Aguayo et al. also reported a lack of association between the cellular VEGF levels and the CR rate or survival when using the median as a cut-off level as we did in our study. Following standard induction chemotherapy, bone marrow of patients with AML had different trends for expression of VEGF and VEGFR-2. On day 16, we found a significant decrease in the content of VEGFR-2, most likely due to chemotherapy-induced clearance of VEGFR-2-positive leukemic blasts. In contrast, levels of VEGF did not differ significantly from those at presentation. The fact that residual cells, in the otherwise hypocellular bone marrows, were frequently positive for VEGF, might be due to VEGF bound to bone marrow stromal cells or to the extracellular matrix. Levels of VEGF in the bone marrow remained elevated when patients achieved a CR after induction chemotherapy. Hematopoietic progenitor cells including erythroblasts that express and release VEGF upon stimulation with cytokines in vitro,48 have been identified as a major source of VEGF in normal marrows.49 These cells might contribute to the elevated levels of VEGF in the regenerating bone marrow. Interestingly, VEGFR-1 was the major receptor in these bone marrows in CR, whereas VEGFR-2 was the most predominant receptor in the AML bone marrows at presentation. From these results, one may speculate that malignant transformation of leukemic cells may result in sustained upregulation of VEGFR-2, which may enhance leukemic cell proliferation through an autocrine loop. On the other hand, it has been demonstrated that cytotoxic drugs directly or indirectly induce endothelial cell apoptosis in animal models.50,51 Thus, increased VEGF/VEGFR-1 expression in CR marrows might not only be important for regeneration of hematopoiesis but also for vas-

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

cular remodeling during the recovery of the bone marrow from intensive chemotherapy. Indeed, VEGFR-1 deficient mice have vascular malformations suggesting that VEGFR-1 has a role in vascular remodeling.22,52,53 The type of cells with increased VEGFR-1 expression in CR remains to be identified in future studies. However, it is well known that VEGFR-1 is expressed by various mature hematopoietic cells, including dendritic and monocytic cells.54,55 Activated VEGFR-1-positive myeloid cells have been shown to release angiogenic factors such as VEGF, platelet-derived growth factor and brainderived neurotrophic factor (BDNF), which enhances vessel formation and stability.56,57 Moreover, recruitment of VEGFR1-positive hematopoietic cells confers stability to the neovessels in a mouse tumor model.58 Together, these data support our hypothesis that the observed VEGF/VEGFR-1 expression in CR marrows occurs in non-malignant hematopoietic cells and contributes not only to regeneration of hematopoiesis but also to vascular remodeling after intensive cytotoxic chemotherapy. In summary, we have demonstrated that acute myeloid leukemia is associated with increased expression of VEGF and its membrane receptor VEGFR-2 in the bone marrow. Furthermore, VEGF and VEGFR-2, but not VEGFR-1 were related to the degree of bone marrow angiogenesis. These results, together with the significant decrease of VEGFR-2 in AML patients who have achieved a CR after induction chemotherapy, strongly support the hypothesis of an important role for VEGF and VEGFR-2 in AML. Furthermore, these findings suggest that therapies using VEGFR-2 as a target could constitute a novel strategy for the treatment of AML.

10 11 12 13

14 15

16

17

18

19

Acknowledgements

20

The authors are indebted to Dr Achim Heinecke (Department of Biostatistics, University of Muenster, Muenster, Germany) for his biostatistical advice and assistance.

21

References

22

1 Folkman J, D’Amore PA. Blood vessel formation: what is its molecular basis? Cell 1996; 87: 1153–1155. 2 Hussong JW, Rodgers GM, Shami PJ. Evidence of increased angiogenesis in patients with acute myeloid leukemia. Blood 2000; 95: 309–313. 3 Padro´ T, Ruiz S, Bieker R, Bu¨rger H, Steins M, Kienast J, Bu¨chner T, Berdel WE, Mesters RM. Increased angiogenesis in the bone marrow of patients with acute myeloid leukemia. Blood 2000; 95: 2637–2644. 4 Aguayo A, O’Brien S, Keating M, Manshouri T, Barlogie B, Koller C, Kantarjian H, Albitar M. Angiogenesis in acute and chronic leukemias and myelodysplastic syndromes. Blood 2000; 96: 2240–2245. 5 Pruneri G, Bertolini F, Soligo D, Carboni N, Cortolezzi A, Ferruci PF, Buffa R, Lambertenghi-Deliliers G, Pezzella F. Angiogenesis in myelodysplatic syndromes. Br J Cancer 1999; 81: 1398–1401. 6 Vacca A, Ribatti D, Roncali L, Ranieri G, Serio G, Silvestris F, Dammacco F. Bone marrow angiogenesis and progression in multiple myeloma. Br J Haematol 1994; 87: 503–508. 7 Rajkumar SV, Fonseca R, Witzig TE, Gertz MA, Greipp PR. Bone marrow angiogenesis in patients achieving complete response after stem cell transplantation for multiple myeloma. Leukemia 1999; 13: 469–472. 8 Perez-Atayde AR, Sallan SE, Tedrow U, Connors S, Allred E, Folkman J. Spectrum of tumor angiogenesis in the bone marrow of children with acute lymphoblastic leukemia. Am J Pathol 1997; 150: 815–821. 9 Kini AR, Kay NE, Peterson LC. Increased bone marrow angiogen-

23

24

25 26 27

28 29

esis in B cell chronic lymphocytic leukemia. Leukemia 2000; 14: 1414–1418. Zwiebel JA. New agents for acute myelogenous leukemia. Leukemia 2000; 14: 488–490. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000; 407: 249–257. Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z. Vascular endothelial growth factor (VEGF) and its receptors. FASEB J 1999; 13: 9–22. Eppenberger U, Kueng W, Schlaeppi JM, Roesel JL, Benz C, Mueller H, Matter A, Zuber M, Luescher K, Litschgi M, Schmitt M, Foekens JA, Eppenberger-Castori S. Markers of tumor angiogenesis and proteolysis independently define high- and low-risk subsets of node-negative breast cancer patients. J Clin Oncol 1998; 16: 3129–3136. Linderholm B, Tavelin B, Grankvist K, Henriksson R. Vascular endothelial growth factor is of high prognostic value in node-negative breast carcinoma. J Clin Oncol 1998; 16: 3121–3128. Linderholm B, Grankvist K, Wilking N, Johansson M, Tavelin B, Henriksson R. Correlation of vascular endothelial growth factor content with recurrences, survival, and first relapse site in primary node-positive breast carcinoma after adjuvant treatment. J Clin Oncol 2000; 18: 1423–1431. Takahashi Y, Kitadai Y, Bucana CD, Cleary KR, Ellis LM. Expression of vascular endothelial growth factor and its receptor, KDR, correlates with vascularity, metastasis and proliferation of human colon cancer. Cancer Res 1995; 55: 3964–3968. Yuan A, Yu CJ, Chen WJ, Lin FY, Kuo SH, Luh KT, Yang PC. Correlation of total VEGF mRNA and protein expression with histologic type, tumor angiogenesis, patient survival and timing of relapse in non-small-cell lung cancer. Int J Cancer 2000; 89: 475–483. Maeda K, Chung YS, Ogawa Y, Takatsuka S, Kang SM, Ogawa M, Sawada T, Sowa M. Prognostic value of vascular endothelial growth factor expression in gastric carcinoma. Cancer 1996; 77: 858–863. Paradis V, Ben Lagha N, Zeimoura L, Blanchet P, Eschwege P, Ba N, Jardin A, Bedossa P. Expression of vascular endothelial growth factor in renal carcinomas. Virchows Arch 2000; 436: 351–356. Smith BD, Smith GL, Carter D, Sasaki CT, Haffy BG. Prognostic significance of vascular endothelial growth factor protein levels in oral and oropharyngeal squamous cell carcinoma. J Clin Oncol 2000; 18: 2046–2052. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breitman ML, Schuh AC. Failure of blood-island formation and vasculogenesis in Flk-1 deficient mice. Nature 1995; 376: 62–66. Fong GH, Rossant J, Gertsenstein M, Breitman ML. Role of Flt1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 1995; 376: 66–70. Shaheen RM, Davis DW, Liu W, Zebrowski BK, Wilson MR, Bucana CD, McConkey DJ, McMahon G, Ellis LM. Antiangiogenic therapy targeting the tyrosine kinase receptor for vascular endothelial growth factor receptor inhibits the growth of colon cancer liver metastasis and induces tumor and endothelial cell apoptosis. Cancer Res 1999; 59: 5412–5416. Prewett M, Huber J, Li Y, Santiago A, O’Connor W, King K, Overholser J, Hooper A, Pytowski B, Witte L, Bohlen P, Hicklin DJ. Antivascular endothelial growth factor receptor (fetal liver kinase 1) monoclonal antibody inhibits tumor angiogenesis and growth on several mouse and human tumors. Cancer Res 1999; 59: 5209–5218. Skobe M, Rockwell P, Goldstein N, Vosseler S, Fusenig NE. Halting angiogenesis suppresses carcinoma cell invasion. Nat Med 1997; 3: 1222–1227. Millauer B, Schawver LK, Plate KH, Risau W, Ullrich A. Gioblastoma growth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature 1994; 367: 576–579. Fiedler W, Graeven U, Ergu¨n S, Verago S, Kilic N, Stockschla¨der M, Hossfeld DK. Vascular endothelial growth factor, a possible paracrine growth factor in human acute myeloid leukemia. Blood 1997; 89: 1870–1875. Bellamy WT, Richter L, Frutiger Y, Grogan TM. Expression of vascular endothelial growth factor and its receptors in hematopoietic malignancies. Cancer Res 1999; 59: 728–733. Chen H, Treweeke AT, West DC, Kathleen JT, Cawley JC, Zuzel M, Toh CH. In vitro and in vivo production of vascular endothelial

1309

Leukemia

VEGF/VEGFR-2 overexpression in AML T Padro´ et al

1310 30

31

32

33 34

35

36

37

38

39

40

41 42

43

44

Leukemia

growth factor by chronic lymphocytic leukemia cells. Blood 2000; 96: 3181–3187. Bellamy WT, Richter L, Sirjani D, Roxas C, Glinsmann-Gibson B, Frutiger Y, Grogan TM, List AF. Vascular endothelial growth factor is an autocrine promoter of abnormal localized immature myeloid precursors and leukemia progenitor formation in myelodysplastic syndromes. Blood 2001; 97: 1427–1434. Di Raimondo F, Azzaro MP, Palumbo GA, Bagnato S, Stagno F, Giustolisi GM, Cacciola E, Sortino G, Guglielmo P, Giustolisi R. Elevated vascular endothelial growth factor (VEGF) serum levels in idiopathic myelofibrosis. Leukemia 2001; 15: 976–980. Aguayo A, Estey E, Kantarjian H, Mansouri, Gidel C, Keating M, Giles F, Estrov Z, Barlogie B, Albitar M. Cellular vascular endothelial growth factor is a predictor of outcome in patients with acute myeloid leukemia. Blood 1999; 94: 3717–3721. Fiedler W, Graeven U, Ergu¨n S, Verago S, Kilic N, Stockschla¨der M, Hossfeld DK. Expression of FLT4 and its ligand VEGF-C in acute myeloid leukemia. Leukemia 1997; 11: 1234–1237. Ziegler BL, Valtieri M, Almeida Porada G, De Maria R, Mu¨ller R, Masella B, Gabbianelli M, Casella I, Pelosi E, Bock T, Zanjani ED, Peschle C. KDR receptor: a key marker defining hematopoietic stem cells. Science 1999; 285: 1553–1558. Katoh O, Tauchi H, Kawaishi K, Kimura A, Satow Y. Expression of the vascular endothelial growth factor (VEGF) receptor gene, KDR, in hematopoietic cells and inhibitory effect of VEGF on apoptotic cell death caused by ionizing radiation. Cancer Res 1995; 55: 5687–5692. Dias S, Hattori K, Zhu Z, Heissig B, Choy M, Lane W, Wu Y, Chadburn A, Hyjek E, Gill M, Hicklin DJ, Witte L, Moore MA, Rafii S. Autocrine stimulation of VEGFR-2 activates human leukemic cell growth and migration. J Clin Invest 2000; 106: 511– 521. Bennet JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnick HR, Sultan C. Proposed revised criteria for the classification of acute myeloid leukemia: A report of the French–American–British Cooperative Group. Ann Intern Med 1985; 103: 620–625. Decaussin M, Sartelet H, Robert C, Moro D, Claraz C, Brambilla C, Brambilla E. Expression of vascular endothelial growth factor (VEGF) and its two receptors (VEGF-R1-Flt1 and VEGF-R2Flk1/KDR) in non-small cell lung carcinomas (NSCLCs): correlation with angiogenesis and survival. J Pathol 1999; 188: 369– 377. Gobbi H, Dupont WD, Simpson W, Plummer WD, Schuyler PA, Olson SJ, Arteaga CL, Page DL. Transforming growth factor-␤ and breast cancer risk in women with mammary epithelial hyperplasia. J Natl Cancer Inst 1999; 91: 2096–2101. Cheson BC, Cassileth PA, Head DR, Schiffer CA, Bennett JM, Bloomfield CD, Brunning R, Gale RP, Grever MR, Keating MJ, Sawitsky A, Stass S, Weinstein H, Woods WG. Report of the National Cancer Institute-sponsored Workshop on definitions of diagnosis and response in acute myeloid leukemia. J Clin Oncol 1990; 8: 813–819. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson U, Peto J, Smith PG. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and example. Br J Cancer 1977; 35: 1–39. Bu¨chner T, Hiddemann W, Wo¨rmann B, Lo¨ffler H, Gassmann W, Haferlach T, Fonatsch C, Haase D, Schoch C, Hossfeld D, Lengfelder E, Aul C, Heyll A, Maschmeyer G, Ludwig WD, Sauerland MC, Heinecke A. Double induction strategy for acute myeloid leukemia: the effect of high-dose cytarabine with mitoxantrone instead of standard-dose cytarabine with daunorubicin and 6-thioguanine: a randomized trial by the German AML Cooperative Group. Blood 1999; 93: 4116–4124. Katoh O, Takahashi T, Oguri T, Kuramoto K, Mihara K, Kobayashi

45

46

47

48

49 50

51

52 53

54

55

56 57

58

M, Hirata S, Watanabe H. Vascular endothelial growth factor inhibits apoptotic death in hematopoietic cells after exposure to chemotherapeutic drugs by inducing MCL1 as an antiapoptotic factor. Cancer Res 1998; 58: 5565–5569. Mesters RM, Padro´ T, Bieker R, Steins M, Kreuter M, Go¨ner M, Kelsey S, Scigalla P, Fiedler W, Bu¨chner T, Berdel WE. Stable remission after administration of the receptor tyrosine kinase inhibitor SU5416 in a patient with refractory acute myeloid leukemia. Blood 2001; 98: 241–243. Mendel DB, Laird AD, Smolich BD, Blake RA, Liang C, Hannah AL, Shaheen RM, Ellis LM, Weitman S, Shawver LK, Cherrington JM. Development of SU5416, a selective small molecule inhibitor of VEGF receptor tyrosine kinase activity, as an antiangiogenesis agent. Anticancer Drug Des 2000; 15: 29–41. Dias S, Hattori K, Heissig B, Zhu Z, Wu Y, Witte L, Hicklin DJ, Tateno M, Bohlen P, Moore MAS, Rafii S. Inhibition of both paracrine and autocrine VEGF/VEGFR-2 signaling pathways is essential to induce long-term remission of xenotransplanted human leukemias. Proc Natl Acad Sci USA 2001; 98: 10857–10862. Bautz F, Rafii S, Kanz L, Mo¨hle R. Expression and secretion of vascular endothelial growth factor-A by cytokine-stimulated hematopoietic progenitor cells: possible role in the hematopoietic microenvironment. Exp Hematol 2000; 28: 700–706. Tordjman R, Delaire S, Plouet J, Ting S, Gaulard P. Fichelson S, Rome´o PH, Lemarchandel V. Erythroblasts are a source of angiogenic factors. Blood 2001; 97: 1968–1974. Klement G, Baruchel S, Rak J, Man S, Clark K, Hicklin DJ, Bohlen P, Kerbel RS. Continuous low-dose therapy with vinblastine and VEGF receptor-2 antibody induces sustained tumor regression without overt toxicity. J Clin Invest 2000; 105: R15–R24. Browder T, Butterfield CE, Kra¨ling BM, Shi B, Marshall B, O’Reilly MS, Folkman J. Antiangiogenic scheduling of chemotherapy improves efficacy against experimental drug-resistant cancer. Cancer Res 2000; 60: 1878–1886. Fong GH, Zhang L, Bryce DM, Peng J. Increased hemangioblast commitment, not vascular disorganization, is the primary defect in flt–1 knock-out mice. Development 1999; 126: 3015–3025. Carmeliet P, Moons L, Luttun A, Vincenti V, Compernolle V, De Mol M, Wu Y, Bono F, Devy L, Beck H, Scholz D, Acker T, DiPalma T, Dewerchin M, Noel A, Stalmans I, Barra A, Blacher S, Vandendriessche T, Ponten A, Eriksson U, Plate KH, Foidart JM, Schaper W, Charnock-Jones DS, Hicklin DJ, Herbert JM, Collen D, Persico MG. Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nat Med 2001; 7: 575–583. Barleon B, Sozzani S, Zhou D, Weich HA, Mantovani A, Marme´ D. Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via VEGF receptor flt-1. Blood 1996; 87: 3336–3343. Clauss M, Weich H, Breier G, Knies U, Rockl W, Waltenberger J, Risau W. The vascular endothelial growth factor receptor Flt-1 mediates biological activities. Implications for a functional role of placenta growth factor in monocyte activation and chemotaxis. J Biol Chem 1996; 271: 17629–17634. Takakura N, Watanabe T, Suenobu S, Yamada Y, Noda T, Ito Y, Satake M, Suda T. A role for hematopoietic stem cells in promoting angiogenesis. Cell 2000; 102: 199–209. Donovan MJ, Lin MI, Wiegn P, Ringstedt T, Kraemer R, Hahn R, Wang S, Ibanez CF, Rafii S, Hempstead BL. Brain derived neurotrophic factor is an endothelial cell survival factor required for intramyocardial vessel stabilization. Development 1999; 124: 4531–4540. Lyden D, Hattori K, Dias S, Costa C, Blaikie P, Butros L, Chadburn A, Heissig B, Marks W, Witte L, Wu Y, Hicklin D, Zhu Z, Hackett NR, Crystal RG, Moore MAS, Hajjar KA, Manova K, Benezra R, Rafii S. Impaired recruitment of bone-marrow-derived endothelial and hematopoietic precursor cells blocks tumor angiogenesis and growth. Nat Med 2001; 7: 1194–1201.