Purpose: To determine the in vivo and in vitro antiangiogenic power of lenalidomide, a “lead compound” of IMiD immunomodulatory drugs in bone marrow (BM) endothelial cells (EC) of patients with multiple myeloma (MM) in active phase (MMEC).
Experimental Design: The antiangiogenic effect in vivo was studied using the chorioallantoic membrane (CAM) assay. Functional studies in vitro (angiogenesis, “wound” healing and chemotaxis, cell viability, adhesion, and apoptosis) were conducted in both primary MMECs and ECs of patients with monoclonal gammopathies (MGUS) of undetermined significance (MGEC) or healthy human umbilical vein endothelial cells (HUVEC). Real-time reverse transcriptase PCR, Western blotting, and differential proteomic analysis were used to correlate morphologic and biological EC features with the lenalidomide effects at the gene and protein levels.
Results: Lenalidomide exerted a relevant antiangiogenic effect in vivo at 1.75 μmol/L, a dose reached in interstitial fluids of patients treated with 25 mg/d. In vitro, lenalidomide inhibited angiogenesis and migration of MMECs, but not of MGECs or control HUVECs, and had no effect on MMEC viability, apoptosis, or fibronectin- and vitronectin-mediated adhesion. Lenalidomide-treated MMECs showed changes in VEGF/VEGFR2 signaling pathway and several proteins controlling EC motility, cytoskeleton remodeling, and energy metabolism pathways.
Conclusions: This study provides information on the molecular mechanisms associated with the antimigratory and antiangiogenic effects of lenalidomide in primary MMECs, thus giving new avenues for effective endothelium-targeted therapies in MM. Clin Cancer Res; 17(7); 1935–46. ©2011 AACR.
This article is featured in Highlights of This Issue, p. 1643
Bone marrow (BM) angiogenesis is an attractive target for the treatment of multiple myeloma (MM). Here, we investigate the antiangiogenic power of lenalidomide, a molecule structurally similar to but less toxic than thalidomide. Lenalidomide impacts angiogenesis in vivo and in vitro and selectively blocks migration of BM endothelial cells (EC) of patients with active disease (MMEC). It downregulates key angiogenic genes and VEGF/VEGFR2-mediated downstream signaling pathways involved in MMEC migration and the NF-κB pathway. Proteomics analysis shows that lenalidomide-treated MMECs specifically modulate the expression levels of angiogenesis-related molecules governing MMEC migration, cell shape and cytoskeletal remodeling, energy metabolism, and protein clearance. Altogether, these data imply the intrinsic complexity of the signaling pathways modulated by lenalidomide at the cellular and molecular levels. Data also suggest that lenalidomide, by targeting angiogenesis, may exert an indirect anti-MM effect. The identified genes and proteins offer candidate targets in MMECs for the development of targeted antiangiogenic therapies.
Lenalidomide (CC-5013, Revlimid; Celgene Corporation) is an oral immunomodulatory drug (IMiD) with shown efficacy and tolerability in patients with multiple myeloma (MM; refs. 1, 2). It is an analogue of thalidomide but has greater potency and a better toxicity profile (3). We have previously shown that thalidomide is antiangiogenic by downregulating key genes involved in autocrine and paracrine angiogenic loops of MM patients (4).
Earlier studies have shown that in the bone marrow (BM) microenvironment, TNF-α increases cell–cell adhesion by inducing NF-κB–dependent upregulation of adhesion molecules on both MM and stromal cells. Although TNF-α only modestly triggers the MM plasma cell proliferation, the subsequent activation of NF-κB stimulates interleukin (IL)-6—another key survival signal—in stromal cells, and both factors are targeted by IMiDs (5). Lenalidomide shows in vivo antitumor properties owing to stimulation of patients' T cells and IL-2/IFN-γ production and inhibition of TNF-α (6). In MM cells, it upregulates apoptotic genes, thus blocking growth (3, 7). It also inhibits MM cell adhesion to stromal cells via downregulation of intercellular adhesion molecule-1 and vascular cell adhesion molecule-1. In peripheral blood mononuclear cells, it reduces the expression of angiogenic factors, such as VEGF and basic fibroblast growth factor (bFGF; ref. 3), TNF-α, and IL-6 (8). However, its precise mechanisms of action and molecular targets in a well-defined subpopulation of BM stromal cells, such as MM endothelial cells (MMEC), are still elusive.
Lenalidomide (Celgene Corporation) was dissolved in dimethyl sulfoxide (Sigma-Aldrich) as a stock solution of 10 mmol/L and stepwise diluted in culture medium before use: the 1.75 μmol/L dose is the interstitial fluid concentration following the therapeutic oral dose of 25 mg/d for an adult patient weighing 70 kg. Recombinant human VEGF165 and TNF-α were purchased from Sigma-Aldrich; heat-inactivated FBS, Dulbecco's modied Eagle's medium (DMEM), and RPMI 1640, antibiotic/antimycotic, glutamine, trypsin/EDTA, and PBS without Ca2+ and Mg2+ were from Euroclone.
Patients and EC cultures
Patients fulfilling the International Myeloma Working Group diagnostic criteria (9) for active MM (n = 44) and monoclonal gammopathy of undetermined significance (MGUS; n = 31) were studied. MM patients had a diagnosis of symptomatic disease and an increase in M-component level in the 3 months before analysis (n = 28) or were in relapse (n = 12) or in leukemic phase (n = 4). There were 24 men and 20 women, aged 44 to 75 (median = 61.5) years, staged as IIA (n = 6), IIB (n = 8), IIIA (n = 23), and IIIB (n = 7), and were not under treatment with lenalidomide. MGUS patients were 18 men and 13 women, aged 43 to 78 (median = 65.7) years. The study was approved by the Ethics Committee of the University of Bari Medical School, and all patients provided their informed consent in accordance with the Declaration of Helsinki.
BM primary ECs from MM (MMEC) and MGUS (MGEC) patients were obtained by centrifugation on Ficoll gradient of heparinized BM aspirates, followed by incubation with magnetic microbeads coated with Ulex europaeus agglutinin-1 lectin (10). Control (healthy) HUVECs were purchased from American Type Culture Collection) and cultured in endothelial growth media-2 (Lonza) with 10% FBS.
Plasma cell conditioned media (CM) were prepared as previously described (10).
Chorioallantoic membrane assay
Fertilized white Leghorn chicken eggs were incubated at 37°C at constant humidity (11). On day 3, the shell was opened and 2 to 3 mL of albumen was removed to detach the chorioallantoic membrane (CAM). On day 8, the CAMs were implanted with 1-mm3 sterilized gelatin sponges (Gelfoam Upjohn) loaded with serum-free medium (SFM) alone (negative control) or CM of MM plasma cells alone (positive control) or with 1.75 μmol/L lenalidomide. The angiogenic response was evaluated on day 12 as the number of vessels converging toward the sponge at 50× and photographed in ovo (Olympus stereomicroscope).
Cell viability and apoptosis assays.
Cell viability was determined by MTT assay (12): cells (5 × 103/100 μL per well) were plated in triplicate in 96-well plates in serum-free DMEM (negative control), complete DMEM alone (positive control), or the lenalidomide doses for 72 hours and treated with MTT for the last 4 hours. The absorbance was measured at 570 nm, with 655 nm as a reference wavelength. Crystal violet assay was carried out as previously described (10). For apoptosis, 5.0 × 105 cells were washed with ice-cold PBS without Ca2+ and Mg2+, incubated with 7-AAD (7-aminoactinomycin D) and PE (phycoerythrin) Annexin V (BD Biosciences), and analyzed by flow cytometry (FACSCantoII; Becton Dickinson).
Chemotaxis was done in triplicate by the Boyden microchamber technique (13) toward SFM alone (negative control) or admixed with the CM of MM plasma cells alone (positive control) or with the lenalidomide doses. Cells were counted on 400× oil-immersion field/membrane.
ECs were grown to confluence on fibronectin-coated (10 μg/mL) 6-cm2 dishes. A “wound” was made by scraping the cell monolayer with a P200 pipette tip. Cells were exposed for 24 hours to SFM alone (negative control) or admixed with the CM of MM plasma cells alone (positive control) or with the lenalidomide doses, fixed, and quantified for motility by counting the cells migrating into the total wound area of each 10× field with an EVOS digital inverted microscope (Euroclone). At least 3 different fields were randomly chosen across the wound length (12).
ECs were plated (2 × 103 cells per well) in triplicate in 96-well fibronectin–coated (10 μg/mL), vitronectin–coated (10 μg/mL), and uncoated plates in SFM alone (positive control) or with the lenalidomide doses, then fixed at 30 minutes with 2.5% glutaraldehyde, stained with crystal violet methanol, and counted as in the cell viability assay, reading absorbance at 595 nm (12).
ECs were plated (1 × 105 cells per well) in duplicate in Matrigel-coated (Becton Dickinson) 24-well plates in SFM alone (positive control) or with the lenalidomide doses. After 18 hours, the skeletonization of the mesh, acquired with the EVOS microscope, was followed by measurement of its topological parameters (“mesh areas,” “length,” and “branching points”) with a computed image analyzer (14).
Total EC protein lysates were subjected to immunoblot with primary and secondary antibodies to the following: VEGFR-2, phospho (p)-extracellular signal regulated kinase (p-Erk)-1/2, Erk-1/2, VE-cadherin, Src, p-NF-κB p65, NF-κB p65, CXCL12/SDF-1, CCL2/MCP-1, myosin light chains (MLC), and p-MLCs (Cell Signaling Technology); p-VEGFR-2 (Y1054) and IEX-1 peptide (Abcam); p-VE-cadherin and p-Src (Invitrogen Corporation); BNIP3 (GeneTex, Inc.); VEGF, bFGF, and SEPW1 (Santa Cruz Biotechnology, Inc.); β-actin (Sigma-Aldrich); and mouse and rabbit horseradish peroxidase–conjugated IgG (Bio-Rad). Immunoreactive bands were visualized by enhanced chemiluminescence (LiteAblot extend substrate; Euroclone) and the Gel Logic 1500 Imaging System (Eastman Kodak Co.), quantified with the Kodak Molecular Imaging Software, and expressed as arbitrary optical density (OD).
Two-dimensional gel electrophoresis, imaging, and quantification.
Total lysates of MMECs and MGECs were prepared as previously described (12). Two-dimensional gel electrophoresis (2-DE) was done in duplicate by isoelectric focusing (IEF) on an IPG-phor system (GE Healthcare Bio-Sciences) with precast gel strips (pH 3–10), followed by SDS-PAGE (Hoefer SE 600 Ruby; GE Healthcare Bio-Sciences), digitized by an ImageScanner (Amersham Biosciences), and analyzed by the ImageMaster 2D Platinum v.5.0 software (GE Healthcare Bio-Sciences) for spot detection and gels matching, as previously described (15). Spots exhibiting an intensity difference between untreated and lenalidomide-treated samples with a P value of 0.05 by the 2-tailored Student's t test for equal or unequal variance (depending on the calculated variance of spots) were considered to be differentially expressed.
In-gel trypsin digestion and analysis by mass spectrometry of selected protein spots.
A preparative gel was obtained by loading whole MMEC lysates. Selected differential protein spots were excised out of the gel, destained with 15 mmol/L K3Fe(CN)6 and 50 mmol/L Na2S2O3, washed with sequential incubation in 25 mmol/L NH4HCO3 in 50% ACN, and dried with 100% ACN. The gel proteins were subjected to cystein reduction and alkylation with 10 mmol/L DTT (dithiothreitol) in 100 mmol/L ambic (ammonium bicarbonate) and 55 mmol/L iodacetamide in 100 mmol/L ambic, respectively, and subsequently reswollen with 10 ng/μL trypsin in 50 mmol/L ambic (10 μg/mL) and digested overnight at 37°C. Two microlitres of 1% trifluoroacetic acid (TFA) was added to stop the enzymatic reaction. Peptides were purified and concentrated by solid-phase extraction (SPE) in ZipTip C18 pipette tips (Millipore) and spotted directly onto a matrix-assisted laser desorption/ionization (MALDI) ground still target (Bruker Daltonics Bremen) upon elution with CHCA matrix (5 mg/mL in 50% ACN and 0.1% TFA). MALDI-mass spectrometry (MS) and MALDI-MS–MS were done with the UltraFlex III MALDI-TOF (time-of-flight)/TOF mass spectrometer (Bruker Daltonics). MS data were acquired in the positive reflectron mode. All acquisitions were done in a mass range of 700 to 3,500 Thomson [relative intensity (m/z)] with voltages of 25 and 21.7 kV for the first and second ion extraction stages, 9 kV for the lens, and 26.3 and 13.8 kV for reflectors 1 and 2, respectively. Quadratic external calibration of TOF was done with monoisotopic mass of bradykinin (clip 1–7) [M + H]+, angiotensin II [M + H]+, angiotensin I [M + H]+, substance P [M + H]+, bombesin [M + H]+, adrenocorticorticotropic hormone (ACTH; clip 1–17) [M + H]+, ACTH (clip 18–39) [M + H]+, and somatostatin [M + H]+. MS spectra were analyzed by the Bruker Flex Analysis 3.0 software. MALDI-TOF/MS–MS analysis was carried out in LIFT mode. Precursor ions were selected manually. MS–MS spectra were acquired with 4,000 to 8,000 (maximum/maximum) laser shots by using the instrument calibration file. Peptide mass fingerprint obtained from MS analysis was used for protein identification in the Swiss-Prot database (SwissProt 2010_04.fasta; 516 081 sequences; 181 677 051 residues) by using the peptide search routine MASCOT 2.3 algorithm (http://www.matrixscience.com) restricted to Homo sapiens taxonomy with carbamidomethyl of cysteines as fixed modification and oxidation of methionines as variable modification. All peptide mass values were considered monoisotopic and mass tolerance was set at 50 ppm. MASCOT scores greater than 56 were considered significant (P ≤ 0.05). A database search for MS–MS data was carried out using the same setting of MS analysis, with a fragment tolerance of 0.4 Da. The main cellular location, function, and biological relationships of identified proteins were searched and determined on PubMed with EndNote software v.5.0.
Real-time reverse transcriptase PCR
Total EC RNA was isolated using the RNeasy Mini kit (Qiagen) and reverse transcribed into total cDNA with the iScript cDNA Synthesis Kit (Bio-Rad). Real-time PCR was carried out using the “StepOne real-time RT-PCR system” (Applied Biosystems). The reaction mixture contained SYBR Green PCR master mix (Applied Biosystems), cDNA, and the following specific (forward/reverse) primers (Invitrogen): VEGF (5′-AAGGAGGAGGGCAGAATCAT-3′/5′-CCAGGCCCTCGTCATTG-3′), bFGF (5′-CCCGACGGCCGAGTTGAC-3′/5′-CACATTTAGAAGCCAGTAATCT-3′), CCL2 (5′-CAAGCAGAAGTGGGTTCAGGAT-3′/5′-TCTTCGGAGTTTGGGTTTGC-3′), CXCL12 (5′-TCAATTGCATCTCCCAGATAATGT-3′/5′-CACGCTGCGTATAGGAATTGG-3′), COL4A1 (5′-CGTAACTAACACACCCTGCTTCAT-3′/5′-CACTATTGAAAGCTTATCGCTGTCTT-3′), CRYAB (5′-AGGTGCATGGAAAACATGAAGA-3′/5′-GGAACTCCCTGGAGATGAAACC-3′), Notch1 (5′-CATGGTACCAATCATGAATCTTTGTT-3′/5′-CTGGAGGGACCAAGAACTTGTATAA-3′), BNIP3 (5′-ATATTCCCCCCAAGGAGTTCCT-3′/5′-ACGCTCGTGTTCCTCATGCT-3′), IER3 (5′-CCGCACTCCCCAAAAAGA-3′/5′-GCTCTCGCGCACCAGGTA-3′), SEPW1 (5′-CGGCCGCCTGGACAT-3′/5′-AATCAACTTCCCGGCTACCA-3′), and control glyceraldehyde 3 phosphate dehydrogenase (GAPDH; 5′-GAAGGTGAAGGTCGGAGT-3′/5′-CATGGGTGGAATCATATTGGAA-3′). Reaction mixture and amplification conditions were maintained according to the manufacturer's instructions. Each RNA was tested in triplicate, and the threshold cycle values were normalized to GAPDH and averaged ±SD. The relative gene expression (fold change) between untreated and lenalidomide-treated ECs was calculated with the 2−ΔΔCT method (16).
MMECs plated on chamber slides (5 × 103 cells per chamber, LabTek; Nalge Nunc International) were treated with VEGF (10 ng/mL) alone or with 1.75 μmol/L lenalidomide, fixed (4% paraformaldehyde), permeabilized (0.2% Triton X-100), blocked (3% BSA/PBS 1X), and then stained with a murine monoclonal antibody to p120 catenin (p120ctn, pp120 Src substrate; BD Transduction Laboratories) and fluorescein (FITC)-conjugated goat anti-mouse IgG (Sigma-Aldrich). Images were obtained using an Axioplan-2 microscope/63X oil-immersion objective (Carl Zeiss).
Lenalidomide suppresses in vivo and in vitro MM-related angiogenesis
To investigate whether lenalidomide could directly inhibit the MM-related angiogenesis, we first examined the in vivo CAM assay. CAMs implanted with a gelatin sponge soaked with the CM of MM patients' plasma cells were surrounded by more newly formed capillaries converging radially toward the sponge in a “spoked-wheel” pattern (mean number of vessels 24 ± 7; Fig. 1A, middle) than with SFM (11 ± 4; Fig. 1A, left). When lenalidomide 1.75 μmol/L was added, a significant inhibition in MM-induced angiogenesis was seen (5 ± 2, P < 0.01; Fig. 1A, right).
Next, we examined whether lenalidomide could impact MMECs in vitro angiogenesis by the Matrigel assay. After 18 hours of incubation, MMECs spread throughout the Matrigel surface and aligned to form branching and anastomosing thick tubes with multicentric junctions, which gave rise to a closely knit network of capillary-like structures (Fig. 1B, left). Lenalidomide dose dependently inhibited this network, which progressively became a poorly organized plexus with few and no anastomosed tubes with scarce junctions. The 0.25 and 0.50 μmol/L concentrations significantly lowered the mesh areas (−34% and −48%, respectively), the vessel length (−45% and −58%), and the vessel branching points (−48% and −55%, P < 0.05; Fig. 1B and C). The MM therapeutic dose of 1.75 μmol/L almost completely inhibited these angiogenic topological parameters (mesh area −67%, vessel length −76%, branching point −80%, P < 0.001; Fig. 1B and C). The 2.5 μmol/L dose fully inhibited MMEC angiogenesis: cells were either isolated or aggregated in small clumps without junctions (mesh area −90%, vessel length and branching points −95%; Fig. 1B and C). Similar, though fainter, effects were observed in the ECs of MGUS patients (MGECs; Supplementary Fig. S1A), whereas there were no significant effects in HUVECs (data not shown).
Lenalidomide affects MMEC functions involved in angiogenesis
MMEC migratory activity was significantly inhibited in a dose-dependent manner, as evidenced both in the “wound” healing assay (Fig. 2A) and in the chemotaxis assay (Fig. 2B). MMEC migration was stimulated by CM of MM plasma cells and reduced in the “wound” by 56% and 75% at 0.5 and 1.75 μmol/L, respectively (number of migrated cells in the “wound” area: 42 ± 3 and 24 ± 2; P < 0.03 and P < 0.01), and in the Boyden microchamber, 0.5 and 1.75 μmol/L inhibited by 27% (P < 0.05) and 55% (P < 0.01). No significant effect on MGECs chemotaxis (Supplementary Fig. S1B) and no effect on HUVECs were observed (data not shown).
Of note, independently of the dose (0.1–10 μmol/L), lenalidomide did not induce cell death of MMECs (neither MGECs nor HUVECs), as assessed by both MTT and crystal violet assays, and did not significantly interfere with their capacity to attach to and spread on fibronectin and vitronectin surfaces and on uncoated dishes. Similar results were obtained with MGECs (Supplementary Fig. S1C) and HUVECs (data not shown). Apoptosis was also not induced (Fig. 2C).
Lenalidomide downregulates key genes involved in the MMEC angiogenic profile
To check possible targets of the lenalidomide antiangiogenic activity, we investigated whether it could modulate the expression profile of genes that are primarily involved in MM angiogenesis (VEGF and bFGF), MMEC migration (CCL2 and CXCL12), and several critical biological functions closely related to angiogenesis (BNIP3, IER3, SEPW1, COL4A1, CRYAB, and Notch1; ref. 17). A significant downregulation was observed only in active MMECs at the 1.75-μmol/L therapeutic dose (Fig. 3). VEGF was 40% less expressed on average than untreated MMECs (P < 0.03); bFGF 50% (P < 0.001); CCL2 and CXCL12 approximately 45% (P < 0.001); BNIP3 40% (P < 0.03); and IER3 and SEPW1 approximately 35% (P < 0.03), whereas COL4A1, CRYAB, and Notch1 were not affected (Fig. 3). Immunoblotting analysis confirmed the relative protein inhibition (Fig. 3). In contrast, lenalidomide produced no or negligible gene/protein regulation in MGECs (Supplementary Fig. S1D) or HUVECs (data not shown).
Lenalidomide abolishes MMEC migration and angiogenesis by affecting VEGF/VEGFR2-mediated downstream signaling pathways
Lenalidomide was shown to exert an inhibitory effect on MMEC angiogenesis by targeting chemotaxis. We have previously found that MMEC functions, including chemotaxis, are mediated by a VEGF/VEGFR2 autocrine loop (12, 18). To assess whether lenalidomide could affect VEGF165-triggered MMEC angiogenesis, serum-starved MMECs were exposed to VEGF165 alone or with added lenalidomide. The VEGF165-triggered VEGFR2 tyrosine phosphorylation lowered significantly and dose dependently with the drug (−25% and −48% with 0.5 and 1.75 μmol/L, P < 0.03 and P < 0.01, respectively; Fig. 4A).
Since the activation of the mitogen-activated protein kinases Erk-1/2 is closely involved in the migration and tubular-like formation of ECs (19, 20), we investigated whether lenalidomide could inhibit Erk-1/2 in MMECs and found that it sizably affected, in a dose-dependent fashion, the VEGF165-induced phosphorylation of Erk-1/2 (−34% and −80% at 0.5 and 1.75 μmol/L, P < 0.003 and P < 0.001, respectively, Fig. 4A). In addition, Erk-1/2 influences the cell motility machinery by phosphoactivating MLC kinase activity. This leads to phophoactivation of MLCs which promotes cytoskeletal myosin-actin contraction needed for cell movement (21). Here, we found that lenalidomide also markedly reduces the VEGF165-induced phosphorylation of MLCs in MMECs (−60% at 1.75 μmol/L, P < 0.01; Supplementary Fig. S2).
It is well documented that VEGFR2 is the principal receptor that mediates the mitogenic and chemotactic effects of VEGF on ECs (22) and that the VEGF/VEGFR2 loop stimulates the tyrosine phosphorylation of VE-cadherin, which is critical for the sprouting of capillary networks during angiogenesis and the metastatic spread of tumor cells (23, 24). Thus, we tested whether lenalidomide could impact VEGF165-mediated VE-cadherin phosphorylation in MMECs. The VEGF165 treatment significantly increased tyrosine phosphorylation of VE-cadherin (on Y658), which was effectively diminished by 0.5 and 1.75 μmol/L lenalidomide (−30% and −66%, respectively, P < 0.01; Fig. 4B). This phosphorylation is mediated by Src tyrosine kinase activity and the activation of the VE-cadherin–Src complex is increased in angiogenic tissues or VEGF-induced ECs, being a necessary step for the cell switch from the quiescent to the angiogenic phenotype (24, 25). We found that lenalidomide consistently reduced VEGF165-induced Src phosphorylation (on Y418) in a dose-dependent manner (−20% and −53% at 0.5 and 1.75 μmol/L, respectively, P < 0.03 and P < 0.01; Fig. 4B). Overall, data suggest that lenalidomide inhibition of VEGF165-triggered angiogenesis in MMECs is mediated by disruption of both Src kinase and subsequent VE-cadherin tyrosine phosphorylation.
Since p120ctn is a VE-cadherin–binding protein in the adherent junctions and affects cell–cell contacts and motility (26), we wondered whether it was modified by lenalidomide. Immunofluorescence studies showed that in VEGF165-stimulated confluent MMECs, the p120ctn-specific fluorescence is diffusely distributed through the cytoplasm, probably because p120ctn cytosolic localization mediates the VEGF-triggered MMEC migration. In contrast, 1.75 μmol/L lenalidomide elicited a massive sequestration of p120ctn at the cell–cell junctions (Supplementary Fig. S3), which may contribute to the inhibition of MMEC motility.
Furthermore, since NF-κB regulates critical genes associated with inflammation and cancer metastases (27), we investigated whether lenalidomide could also modulate the NF-κB–mediated pathway. NF-κB transcriptional activation occurs via heterodimerization of its subunits p50 and p65. Lenalidomide inhibited, in a dose-dependent manner, TNF-α–triggered phosphorylation of the p65 subunit (−65% and −86% at 0.5 and 1.75 μmol/L, respectively, P < 0.003 and P < 0.001; Fig. 4C), without affecting basal p65. However, lenalidomide exerts comparable effects, albeit not significant, in MGECs but not in HUVECs (Supplementary Fig. S4).
Lenalidomide-induced changes in the MMEC proteome
To further dissect the inhibitory effects of lenalidomide at the molecular level, we pursued comparative proteomics analysis of MMECs of 2 patients (1 at diagnosis and 1 at relapse) cultured without (Fig. 5A) or with (Fig. 5B) lenalidomide 1.75 μmol/L for 72 hours, and the same analysis was in parallel carried out for MGECs, as control BM-derived ECs (Fig. 5C and D). At least two 2-DE gels were run per sample, followed by computer-assisted spot matching and silver staining imaging to enable localization of variation spots. More than 500 protein spots were visualized per gel; 10 proteins that were differentially expressed by the treatment (2-fold change) in MMECs were identified by peptide sequencing on the basis of tandem mass spectrometry (MS–MS) and database searching (Fig. 6 and Supplementary Table S1). Among these proteins, 5 were downregulated, the cell membrane–embedded LBP/p40 (RSSA), cytoplasmatic vimentin and keratin2C1, mitochondrial DLDH, and glutamate dehydrogenase (DHE3), whereas 5 were upregulated, cytoplasmatic septin2 (SEPT2), γ-actin (ACTG), PNPH, the glycolytic enzyme ENOA, and PSB4. As shown in Figure 6, all the identified proteins were not significantly affected by lenalidomide treatment in MGECs. Interestingly, basal expression levels of vimentin, keratin2C1, mitochondrial DLDH, DHE3, and ENOA were increased in untreated MMECs versus their normal counterpart (Fig. 6A, for both MMECs and MGECs).
In accordance with the aforementioned antimigratory and antiangiogenic effects of lenalidomide, most of the differently expressed MMEC proteins in response to lenalidomide are angiogenesis-related agents controlling cell shape and invasiveness (LBP/p40; ref. 28), cytoskeletal dynamic remodeling (vimentin, keratin2C1, septin-2, and γ-actin; refs. 29–32), and coordinated events at the interface between cytoskeleton and membrane biology (polymerization of F-actin microfilaments and transport of organelle or ribosomal subunits during protein biogenesis). Alterations of enzymatic activities (DLDH, DHE3, PNPH, and ENOA; refs. 33, 34) are seemingly specific indicators of changes in energy metabolism pathways and oxidative stress responses, even in the absence of a significant apoptosis induction in lenalidomide-treated MMECs. Upregulated levels of PSB4 (35) could be part of a compensatory response to changes in cellular bioenergetics due to altered protein clearance and turnover.
The clinical benefit of lenalidomide in treating myelodysplastic syndromes, MM, chronic lymphocytic leukemia, and non–Hodgkin's lymphoma has drawn heightened attention to its mechanism of action, because specific targets of IMiD compounds remain undefined. In MM plasma cells, lenalidomide inhibits cell growth and promotes apoptosis and blocks their adhesion to stromal cells in the BM milieu (3, 7). In stromal cells, lenalidomide reduces the expression of angiogenic factors and several additional factors that support the plasma cell growth. In addition, lenalidomide stimulates T cells and natural killer cells (6). IMiDs, including lenalidomide, also have antiangiogenic properties that are independent of their immunomodulatory effects (36). We have already shown that thalidomide exerts a direct antiangiogenic activity on MMECs owing to downregulation of key genes mandatory for autocrine and paracrine loops of neovascularization, even though at approximately 10× higher doses than those of lenalidomide used here (4). However, although the IMiDs share many of biological properties of the thalidomide, the relative potency and range of these effects vary substantially from each other and from thalidomide. Thus, each molecule cannot be assumed to have the same overall biological effects or therapeutic properties as thalidomide or other IMiDs. In fact, as shown here, lenalidomide is antiangiogenic in vivo (CAM assay) and in vitro, hence overlapping thalidomide (37, 38). However, lenalidomide halts only MMEC chemotaxis but not proliferation whereas thalidomide halts both EC functions (38) and MMEC proliferation (10). Tentatively, we suggest that the differential effects of the drug may derive from distinct mechanisms of action at the molecular level: thalidomide potently inhibits EC functions, possibly through preventing Erk-1/2 nuclear translocation and/or inhibition of Akt/PKB phophorylation, thus halting mainly VEGF-mediated survival/proliferation (38) whereas lenalidomide inhibits markedly MLC phosphorylation in the cytoplasm (Supplementary Fig. S2), thus acting much more on cell migration. These observations further underscore striking pharmacodynamic differences between lenalidomide and thalidomide.
Accordingly, in this study, we first pointed out the strong potential of lenalidomide in inhibiting in vivo angiogenesis induced by the CM of MM plasma cells (CAM assay) at 1.75 μmol/L, which corresponds to the interstitial fluid level in patients treated with 25 mg/d per os. Then, we examined the activity of lenalidomide in targeting functions associated with in vitro angiogenesis and found that it impairs both MMEC angiogenesis and their migration activity, bringing on the maximum inhibition at the same dose effective for antiangiogenic activity in vivo. Of note, the angiogenic and chemotactic cell functions were significantly hampered only in MMECs derived from patients with active disease whereas the drug exerted only negligible or no effect on ECs from nonactive MM patients or MGECs, or on HUVECs. Previous studies (39) showed that lenalidomide inhibits VEGF-, bFGF-, and TNF-α–mediated migration of HUVECs only at doses 10 to 100 times higher than those used here and inhibits angiogenesis in human umbilical artery explants, with approximately 50% inhibition at 1 μmol/L (40). It is possible that the proangiogenic factors produced by the umbilical artery are either similar to those factors present in the MMEC culture or sufficient to reveal the antiangiogenic activity of lenalidomide.
Although lenalidomide produces in vitro growth arrest and apoptosis of MM plasma cells and a variety of solid tumor cell types (3), the drug does not affect viability and apoptosis of MMECs (neither MGECs nor HUVECs), thus indicating that the MMEC antiangiogenic effect does not correlates with those cell functions. Also, lenalidomide does not interfere with the MMEC fibronectin- and vitronectin-mediated adhesion. Data indicate that lenalidomide may inhibit MM angiogenesis because of an antimigratory rather than cytotoxic or apoptogenic mechanism on MMECs.
Accordingly, we investigated the molecular mechanisms involved in the inhibition of MM angiogenesis by lenalidomide. The drug induces inhibition of genes directly involved in the angiogenic cascade, such as VEGF and bFGF, and of the chemokines CCL2 and CXCL12 that are relevant for migration and homing of MMECs (41, 42). Lenalidomide displayed a selective effect on other angiogenesis-related genes shown to be upregulated in active MMECs (17): inhibition of BNIP3, IER3, and SEPW1 but no effect on COL4A1, CRYAB, and Notch1. BNIP3 is overexpressed in many malignancies due to hypoxia (17, 43). Since hypoxia favors the malignant progression, the BNIP3 overexpression indicates high metastatic ability and aggressive phenotype and implies a poor prognosis. Perhaps, the downregulation of BNIP3 in active MMECs by lenalidomide can be explained as a way to overtake the hypoxia-mediated invasive potential, hence affecting their overangiogenic phenotype. IER3 is a member of the “stress-associated early response gene” family involved both in a wide range of cell functions under stress conditions and in oncogenesis (44). Tentatively, IER3 downregulation by lenalidomide in active MMECs may account for an effective way to block their overangiogenic phenotype, hence the MM plasma cell spreading and invasion. SEPW1 has antioxidant function that protects cells from oxidative stress (45). Since SEPW1-mediated homeostasis in MMECs may be crucial for their life and survival, the lenalidomide-induced SEPW1 downregulation may represent a way to overcome protection of the cells from oxidative stress, thus halting MM angiogenesis. Among the genes not regulated by lenalidomide, CRYAB is antiapoptotic (46) whereas Notch1 drives proliferation of primary MM cells (47). Given their roles in proliferation and apoptosis, they are unlikely to be affected by lenalidomide, which, indeed, has no effects on these MMEC functions as shown here. Although COL4A1 is a major component of the vascular basement membrane, involved in the migration (sprouting) of MM microvessels (10), the lenalidomide-mediated downregulation of other major genes governing the sprouting such as VEGF, bFGF, CCL2, CXCL12, BNIP3, IER3, and SEPW1 can overcome the role of COL4A1.
We found that MMEC angiogenesis is sustained by the VEGF/VEGFR2 autocrine loop (18). VEGF strongly activates, via VEGFR2, the MAP kinase Erk-1/2, which is the last signaling mediator to the nucleus involved in cell proliferation and chemotaxis (19, 20). As a consequence, Erk-1/2-specific inhibitors reduce endothelial tubulogenesis in vitro (20). The block by lenalidomide in active MMECs of the VEGF165-induced phosphorylation of both VEGFR2 and Erk-1/2 further confirms its antiangiogenic power in the MM BM milieu in which VEGF is highly represented (10).
VEGF increases tyrosine phosphorylation of the adherens junction components, such as VE-cadherin, β-catenin, plakoglobin, p120 catenin, and the cell–cell adhesion molecule PECAM-1 in cultured ECs (23), and the VEGF-mediated VE-cadherin phosphorylation causes the loosening of cell–cell contacts in established vessels to allow EC migration and vessel sprouting during angiogenesis (24). Src kinase is enriched at cell–cell contacts and implicated in regulating the phosphorylation of cadherins and catenins; the phosphorylation of VEGF-induced VE-cadherin tyrosine (on Y685) is mediated by Src kinase and is critical for the VEGF-triggered EC migration (48). Here, we show that lenalidomide inhibits VEGF-induced angiogenesis through blocking of Src kinase phosphoactivation and, consequently, tyrosine phosphorylation of VE-cadherin in MMECs, which underlies its antiangiogenic activity in MM. Overall, these data indicate that lenalidomide impacts on MMEC migration in active MM, thereby inhibiting angiogenesis. This inhibition is associated with, and partly mediated by, modifications of intracellular signals triggered by the VEGF-regulated VE-cadherin. Furthermore, we conducted immunolocalization studies on p120ctn, which binds the cytoplasmic tail of cadherin and localizes into adherens junctions, thus playing a key role in cell–cell adhesion and motility (26). In epithelial cells, p120ctn shows a cytoplasmic diffuse distribution in sparse cultures that is indicative of the cell migratory state, whereas in confluent cells, the formation of cadherin-mediated cell–cell contacts is accompanied by sequestering p120ctn to the junction regions, thus abolishing its effects on cell migration (26). Our immunolocalization results suggest that in active MMECs, lenalidomide strongly overcomes VEGF-mediated localization of p120ctn in the cytoplasm by sequestering the protein at the cell–cell junctions. Perhaps, the VEGF-triggered cytoplasmic distribution of p120ctn could impair VE-cadherin–p120ctn binding, thus making ECs free to migrate, whereas lenalidomide-induced sequestration at cell–cell contacts may stabilize the p120ctn–VE-cadherin interaction, thus blocking the MMEC motility.
NF-κB regulates several genes involved in angiogenesis including the chemokine genes governing EC migration (49). It stimulates angiogenesis by inducing IL-8 and VEGF expression (50). By demonstrating here that lenalidomide inhibits the activity of the p65 subunit of NF-κB in active MMECs, we suggest its role in inhibiting cell migration and angiogenesis in MM.
We further extended these findings through a comparative proteomics analysis of the lenalidomide-treated MMECs versus untreated cells. The functional characterization of the identified proteins strengthens the inhibitory effects of lenalidomide on MMEC motility and angiogenic potential associated with changes in cytoarchitecture and mitochondrial energy metabolism. Some of the identified proteins (i.e., LBP/p40, vimentin, keratin2C1, septin2, and γ-actin) are closely correlated with tumor growth, invasion, and metastasis (28–31). How MM signaling networks could affect turnover and the function of such proteins in MMECs remains a major challenge for future investigations. However, all these proteins could represent attractive biomarkers for monitoring resistance to antiangiogenic and cytoskeleton-disrupting drugs (32) in lenalidomide-treated MM patients.
In conclusion, our findings emphasize the intracellular mechanisms whereby lenalidomide exerts its antiangiogenic potential in active MM patients and indicate that this drug may be of improved therapeutic value and could enhance responses to chemotherapy and radiotherapy in MM.
Disclosure of Potential Conflicts of Interest
The authors declare no competing financial interests. The sponsors of this study are public or nonprofit organizations that support science in general. They had no role in gathering, analyzing, or interpreting the data. The authors are fully responsible for the content and editorial decisions for the manuscript.
This work was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC), Investigator Grant, and Special Program Molecular Clinical Oncology 5 per mille n. 9965, Milan, and the Ministry of Health, Progetto Oncologia 2006, Humanitas Mirasole S.p.A., Rome, Italy. A. Basile is supported by a studentship from Fondazione Italiana per la Ricerca sul Cancro (FIRC). The authors received editorial support from Excerpta Medica in the preparation of the manuscript, funded by Celgene Corporation.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).
- Received September 3, 2010.
- Revision received December 14, 2010.
- Accepted January 5, 2011.
- ©2011 American Association for Cancer Research.