The Aplidin analogs PM01215 and PM02781 inhibit angiogenesis in vitro and in vivo

Novel synthesized analogs of Aplidin, PM01215 and PM02781, were tested for antiangiogenic effects on primary human endothelial cells in vitro and for inhibition of angiogenesis and tumor growth in vivo. Antiangiogenic activity of both derivatives was evaluated by real-time cell proliferation, capillary tube formation and vascular endothelial growth factor (VEGF)-induced spheroid sprouting assays. Distribution of endothelial cells in the different phases of the cell cycle was analyzed by flow cytometry. Aplidin analogs were testedin vivoin chicken chorioallantoic membrane (CAM) assays. Both derivatives inhibited angiogenic capacities of human endothelial cells (HUVECs) in vitro at low nanomolar concentrations. Antiangiogenic effects of both analogs were observed in the CAM. In addition, growth of human multiple myeloma xenograftsin vivoin CAM was significantly reduced after application of both analogs. On the molecular level, both derivatives induced cell cycle arrest in G1 phase. This growth arrest of endothelial cells correlated with induction of the cell cycle inhibitor p16INK4A and increased senescence-associated beta galactosidase activity. In addition, Aplidin analogs induced oxidative stress and decreased production of the vascular maturation factors Vasohibin-1 and Dickkopf-3. From these findings we conclude that both analogs are promising agents for the development of antiangiogenic drugs acting independent on classical inhibition of VEGF signaling.


Background
Growing tumors undergo an angiogenic switch, i. e. tumor cells start to produce angiogenic growth factors that cause destabilization of existing blood vessels, angiogenic sprouting and generation of new immature blood vessels [1]. Normally, endothelial cells are growth-arrested in the human vascular system and stabilized by mural cell coverage. Upon hypoxia or wound healing, factors like vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (FGF2) induce vascular basement membrane degradation, invasion, migration and proliferation of endothelial cells [2]. After capillary tube formation, endothelial cells recruit new mural cells to cover and stabilize newly formed blood vessels [3]. Growing tumors make use of these mechanisms under hypoxic conditions and generate new blood vessels to enlarge and metastasize [4].
Antiangiogenic therapies in cancer medicine make use of drugs that inhibit proliferation of endothelial cells and induce stabilization and maturation of blood vessels [5]. Due to the fact that tumor blood vessels are leaky and immature, they affect blood flow and interstitial blood pressure [6]. Stabilization of blood vessels ensures better delivery of chemotherapeutic drugs to the tumor and enables interstitial blood pressure to be lowered. Thus, cancer medicine uses antiangiogenic drugs like neutralizing antibodies against VEGF-A or small molecules that inhibit the tyrosine-kinase activity of VEGF receptors [7]. Both attempts lead to inhibition of VEGF signaling, but after prolonged treatment alternative pathways cause resistances and further angiogenic processes and tumor progression to develop [8].
This study analyzed substances that are able to inhibit proliferation of human endothelial cells at low non-toxic nanomolar concentrations, thereby inducing growth arrest in tumor endothelial cells. Optimal antiangiogenic compounds should inhibit the proliferation of tumor endothelial cells, but not induce apoptosis in growtharrested endothelial cells, such as normal endothelial cells in the vascular system. Both drugs, bortezomib and Aplidin, have been shown to exert potent anti-myeloma activities by inducing apoptosis in multiple myeloma cell lines [9][10][11][12][13]. Apart from this anti-myeloma activity both display antiangiogenic activity in vitro and in vivo in different tumor models independent of inhibition of VEGF signaling [9,11,[13][14][15]. More than 200 different Aplidin analogs were synthesized and screened for cytotoxic activities against cancer cell lines (WO 02002596). Here, we identified and characterized two novel analogs with reduced in vitro cytotoxicity on human primary cells and more easy chemical synthesis than Aplidin™ and tested them in comparison to the established drugs bortezomib and Aplidin TM for their antiangiogenic effects.

Substances
Bortezomib was purchased from Selleckchem and dissolved in DMSO (SIGMA Biochemicals) to a stock solution of 250 mM. Aplidin™ and Aplidin analogs PM01215 and PM02781 were synthesized in Pharmamar and dissolved in DMSO to stock solutions of 250 mM and stored in aliquots at −80°C. All stocks were further diluted with DMSO to working concentrations of 1 mM and stored at −20°C. Nacetyl cysteine (NAC, Sigma Biochemicals) was dissolved in distilled sterile water, and 30 % H 2 O 2 was purchased from Merck. Thapsigargin was purchased from Life Technologies and dissolved in DMSO (SIGMA Biochemicals) to a stock solution of 1 mM.
Human PBMNCs from healthy donors (n = 3,) were prepared as described elsewhere [16]. In brief, blood samples from blood donors were collected in anticoagulant (EDTA) tubes and transferred to Leucosep® tubes (Greiner Bio-One) containing Ficoll (LSM1077 Lymphocyte separation medium, PAA) for density gradient centrifugation. Thereafter, mononuclear cell faction was washed in PBS, characterized by flow cytometry (size, granularity, CD45+ expression) and used for experiments.
All primary cells were characterized by flow cytometry using a panel of cell type-specific markers (Additional file 1: Table S1) and were tested for the absence of HIV1/2, HBV, HCV and mycoplasma. Only cells of low passages were used for experiments. OPM-2 multiple myeloma cells (AC55) were purchased 2012 directly from DSMZ (Germany), authenticated by us (STR-profiling, flow cytometry: CD138+/CD38+) and cultivated in RPMI1640 medium (Sigma Biochemicals) with 10 % bovine calf serum (Hyclone) and 100 IU/mL penicillin, 100 μg/mL streptomycin and 2 mM glutamine (all PAA Laboratories GmbH) on uncoated plastic material. OPM-2 cells were lentivirally transfected to express eGFP and propagated in the presence of blasticidin (2.5 μg/mL, Invitrogen) before usage for in vivo experiments.

DKK-3 ELISA
Cells were treated for 72 h with 10 nM solution of the respective compounds. For quantitative measurement of DKK-3 in supernatants a commercially available ELISA (human DKK-3 DuoSet; DY1118, R & D Systems) was used according to the manufacturer's guidelines.

Immunofluorescence and confocal microscopy
Cells were seeded on collagen-coated eight-well culture slides (Falcon BD Labware) and incubated with 10 nM of Aplidin, PM01215 and PM02781 for 5 h. Living cells were stained with CellRox®Green reagent to monitor intracellular oxidative stress, and nuclei were stained with NucBlue (Molecular Probes, Life Technologies) according to the manufacturer's protocol. Confocal microscopy was performed with a spinning disc confocal microscopic system (Ultra VIEW VoX; Perkin Elmer, Waltham, MA, USA) that was connected to a Zeiss AxioObserver Z1 inverted microscope (Zeiss). Images were acquired with Velocity software (Perkin Elmer) using a 63x oil immersion objective with a numerical aperture of 1.42.

Real time cell proliferation assays
Real time cell proliferation experiments were performed using the RTCA DP instrument (Roche Diagnostics GmbH), which was placed in a humidified incubator maintained at 5 % CO 2 and 37°C. For proliferation assays, cells were seeded in complete medium in 16-well plates (E-plate 16, Roche Diagnostics GmbH) at a density of 2000 cells/well after coating with 10 μg/mm 2 fibronectin (Sigma Biochemicals). The plate containing gold microelectrodes on its bottom was monitored every 10 min for 4 h (adhesion process), then once every 30 min, until the end of experiment, for a total of 72 h. Data analysis was performed using RTCA software 1.2 supplied with the instrument.

Capillary tube formation and angiogenic sprouting assays
Cells were incubated for 12 h with 10 nM of the respective compounds. To analyze tube formation, 24-well plates were coated with 200 μL growth factor-reduced matrigel (BD Biosciences). HUVECs were resuspended in 200 μL EGM-2 medium (1 × 10 5 cells) containing 10 nM of the respective compound and placed on top of the polymerized matrix; tube formation was observed after 6 hours. Tubes were viewed under an inverted transmission microscope (Zeiss Axiovert 200 M) and documented with a digital imaging system (Axiovision Software, Zeiss).
For sprouting assays HUVEC spheroids were generated overnight in hanging-drop culture consisting of 400 cells in EBM-2 medium, 2 % FCS and 20 % methylcellulose (Sigma Biochemicals). Spheroids were embedded in collagen type I from rat tail (Becton Dickinson) and stimulated with 50 ng/ml VEGF (Sigma Biochemicals) in the presence or absence of compounds or control substances (DMSO, bortezomib). Sprouts were also analyzed by inverted transmission-microscopy (Zeiss Axiovert 200 M) and documented by a digital imaging (Axiovision Software, Zeiss). The cumulative sprout length (CSL) was analyzed after printing of high quality pictures and counting by two independent blinded observers.

Chicken chorioallantoic membrane (CAM)
Fertilized chicken eggs (Gallus domesticus, Charles River) were placed in a 75-80 % humidified 37°C incubator (Grumbach) to allow normal embryo development. On day three eggs were opened, egg shells removed and embryos were placed in a sterile Petri dish in an egg incubator to induce CAM development. On day 8, when CAM and its vasculature were well developed, all experiments were performed. Subsequently, two rings per chicken were grafted on the CAM. Drugs (10 nmol/ring) with VEGF (1 μg/ring) or drugs alone were applied every second day at the center of Permanox™ rings.
On day 6 post-grafting chicken embryos were sacrificed by hypothermia, blood vessels in the ring area were photographed by stereo microscope (Olympus SZW 10) and vessel density was determined by counting with Photoshop CS4 (Adobe).

Human tumor xenograft model in the CAM
OPM-2 eGFP multiple myeloma cells (2.5 × 10 5 ) were mixed with rat-tail collagen and human mesenchymal stromal cells (0.5 × 10 5 ) and the 1 nmol of the respective compounds. Collagen drops (30 μl) were placed on parafilm for 30 min to allow polymerization of the extracellular matrix at 37°C. Then onplants were transferred to the CAM of 7-day-old chick embryos. After 5 days of in vivo growth, onplants were documented by the Olympus SZX10 stereomicroscope (Olympus) equipped with an Olympus DFPL 2-4x objective lens connected with a digital camera (Olympus E410) and flexible cold light (KL200; Olympus). Excised xenografts were transferred into 0.5 ml RIPA Buffer (Sigma Aldrich, Linz, Austria) and homogenized with an Ultra Turrax homogenizer three times for 5 s on ice. Thereafter, homogenate underwent three freezing/thawing-cycles in liquid nitrogen and 37°C water bath. After centrifugation, supernatants were diluted in assay buffer. GFP levels were measured by Cell Biolabs' GFP ELISA Kit (San Diego, CA, USA), using biotinylated anti-GFP antibodies, according to the manufacturer's protocol.

Statistical analysis
Statistical analyses were performed with the GraphPad Prism™ software for Windows. Unpaired t-test was used to study differences between the means of one treatment group and control. The average scores across treated groups were not compared. Statistical analyses of quantitative PCR data were performed according to the delta Ct method described by Pfaffl et al. [17].

PM01215 and PM02781 inhibit cell proliferation and induce cell cycle arrest in human endothelial cells
Testing more than 200 different analogs by Pharmamar in direct comparison to the original compound on tumor cell lines (Patent WO 2002002596) revealed two compounds with similar in vitro activity and more easy chemical synthesis than Aplidin™ (Fig. 1).

PM01215 and PM02781 induce oxidative stress and terminal growth arrest
Aplidin™ has been reported to induce cell death by oxidative stress [10]. Therefore, we tested both Aplidin analogs with CellRox®Green, a fluorogenic probe for measuring oxidative stress in living cells. The cellpermeant dye is weakly fluorescent while in a reduced state and exhibits bright green photostable fluorescence upon oxidation with reactive oxygen species (ROS). Aplidin derivatives induced ROS already 5 h after incubation. ROS was effectively blocked by adding 25 μM Nacetyl-cysteine (NAC) as antioxidant to the culture medium (Fig. 3a). Furthermore, we analyzed them in a real-time proliferation system in direct comparison to H 2 O 2 , a known inductor of ROS, and attempted to rescue cells by adding NAC to culture medium (Fig. 3b). Growth of endothelial cells was inhibited by 20-30 μM H 2 O 2 . Proliferation of H 2 O 2 -treated cells was significantly increased by adding NAC (25 μM) to the culture supernatant. Aplidin analog-treated cells (10 nM each) could not be stimulated for proliferation, even after incubation with the antioxidant NAC, indicating a terminal growth arrest.

PM01215 and PM02781 increase phosphorylation of c-Jun N-terminal Kinase (JNK) after mitotic stress
Aplidin analogs were tested for direct effects on phosphorylation of stress (JNK) and mitogenic (ERK) and survival kinases (AKT). In comparison to DMSOtreated endothelial cells, Aplidin analog-treated cells (each 10 nM, n = 3) showed no significantly altered phosphorylation of prosurvival kinases upon stimulation with EGM2 medium containing mitogenic growth factors such as VEGF and bFGF. Aplidin analogs did not increase phosphorylation of ERK 20 min after mitogenic stimulation (Fig. 3c). The pro-survival AKT (Protein Kinase-B) was not affected after treatment of endothelial cells with PM01215 and PM02781. In comparison to DMSO treated cells, PM01215 and PM02781 were significantly increasing JNK-phosphorylation in endothelial cells 5 min after mitogenic stimulation (Fig. 3c).
PM01215 and PM02781 induce premature senescence and expression of the cell cycle inhibitor p16 INK4A Aplidin analogs induced growth arrest in endothelial cells by induction of p16 INK4A gene expression (Fig. 4a) already 24 h after incubation. Bortezomib (5 nM) did not affect p16 INK4A gene levels. Western Blot analysis after 72 h confirmed that p16 protein was induced by Aplidin and analogs, whereas bortezomib-treated cells were already apoptotic and showed degradation of cellular protein (Fig. 4b). Bortezomib significantly elevated p27 Kip1 gene expression, but had no effect on TP53 gene expression 24 h after stimulation (Fig. 4a). Aplidin analogs did not affect TP53 or p27 Kip1 gene expression (Fig. 4a). As expected from its proteasome inhibitory property, bortezomib (5 nM) increased p53 and p21 protein in endothelial cells after 24 h of incubation (Fig. 4c). Aplidin analogs did not affect p53, p21 or p27 protein levels 24 h after incubation (Fig. 4c) or 72 h after stimulation (data not shown).
Since oxidative stress brings on premature senescence in primary cells, we analyzed the effects of PM01215 and PM01215 PM02781 Fig. 1 Chemical structure of Aplidin™ and the two novel Aplidin derivatives with modified side-chains (boxes). PM01215 and PM02781 are analogs of Aplidin™, in which the pyruvyl-proline side-chain was replaced with a urea derivative based on phenylisocyanate. Additionally, in PM02781 the α − (α − hydroxyisovaleryl) propionyl (Hip) group that is present in the depsipeptide cycle and connected to an isostatine unit by an ester bond was replaced with L-valine. Note: Both analogs are easier to synthesize than Aplidin™ PM02781 for induction of growth arrest by staining for senescence-associated β-galactosidase activity (SA-β-gal) after three days of incubation. In comparison to untreated control cells, Aplidin analogs (10 nM each) significantly increased the number of SA-β -gal positive cells (Fig. 4d).

Aplidin and analogs do not induce canonical unfolded protein response
Beside the analysis of oxidative stress and senescence we wanted to analyze effects of Aplidin and its analogs on induction of endoplasmatic reticulum (ER) stress and activation of unfolded protein response (UPR). Therefore we compared UPR responses in direct comparison to 10 nM thapsigargin, a strong inducer of UPR on human endothelial cells. Three different branches of UPR, the master regulator GRP78/HSP5A, the IRE1α/XPB1 s and PERK/CHOP/DDIT3 pathway were analyzed on gene expression level. Bortezomib, Aplidin and analogs did not induce HSP5A gene expression and splicing of XPB1 within 24 h (Fig. 5a/b). Only thapsigargin was able to  (Fig. 5c). DDIT3/CHOP was significantly elevated on gene expression in bortezomib, Aplidin and analog-treated cells, but not in such a massive response as in thapsigargin-treated control cells (Fig. 5d).

PM01215 and PM02781 inhibit capillary tube formation and angiogenic sprouting of human endothelial cells
Antiangiogenic effects of the Aplidin derivatives were tested in vitro in functional 3D assays inducing capillary tube formation and generation of sprouts from endothelial spheroids. Tube formation of HUVECs (n = 3) in matrigel was inhibited by Aplidin™ (5 nM) and Aplidin derivatives (10 nM each, Fig. 6a). DMSO (0.1 %) served as negative control and bortezomib (5 nM) as positive control for inhibition of capillary tube formation.
In addition, HUVECs (n = 3) were cultured as spheroids in hanging drops for 24 h. Thereafter, spheroids were seeded into methylcellulose/collagen type I matrix together with 100 ng/mL VEGF and drugs (Fig. 6b). In comparison to 0.1 % DMSO (control) Aplidin™ and analogs significantly inhibited angiogenic sprouting at concentrations of 1 and 10 nM.   Three different branches of UPR, the master regulator GRP78/HSP5A, the IRE1a/XPB-1 s and PERK/CHOP/DDIT3 pathway were analyzed. In contrast to thapsigargin (10 nM), bortezomib (5 nM), Aplidin (5 nM) and analogs (10 nM) did not induce HSP5A gene expression (a) and splicing XPB1 within 24 h (b). c Only thapsigargin was able to increase GRP78 and XPB1 s protein as analyzed after 72 h by Western Blot. d DDIT3/CHOP was significant elevated on gene expression after 24 h in bortezomib, Aplidin and analog-treated cells. Aplidin and derivates did not display such an inductive response as observed for thapsigargin. e For analysis of vascular maturation factors HUVECs were incubated with bortezomib (5nM), Aplidin™ (5nM) or Aplidin analogs (each 10 nM) Western Blot analysis of VASH1 protein in cytosolic extracts of bortezomib, Aplidin™ and Aplidin analog-treated endothelial cells. Aplidin™ and analogs downregulated VASH1 and KDR protein levels after 24 h. f DKK-3 release from endothelial cells was reduced upon treatment with bortezomib, Aplidin™ or PM01215 and PM02781 after 72 h, as determined by a sandwich ELISA specific for human DKK-3. The means of one treatment group were compared to untreated control. The level of significance for the analysis was set at p < 0.05 Fig. 7b). Moreover, we observed significant inhibition of spontaneous neovascularization in CAM (Fig. 7b). These data indicate that both analogs have antiangiogenic activities at sublethal concentrations in chicken embryos.
PM01215 and PM02781 inhibit growth and vascularization of human multiple myeloma grafts in the chicken CAM In addition, PM01215 and PM02781 were tested in a human multiple myeloma xenograft model in the chicken. The human myeloma cells OPM-2 eGFP were grafted together with human mesenchymal cells and collagen-type-I matrix on the CAM of chicken embryos.
Both, PM01215 and PM02781, significantly inhibited blood vessel formation adjacent to tumor grafts 5 days after incubation (Fig. 8a). Tumor cell mass was quantified by measuring the transgene GFP in myeloma cells in an ELISA after homogenization of tumors with adjacent host tissue. In comparison to control tumors, PM01215 and PM02781 treated grafts displayed significantly less tumor mass and cell growth (Fig. 8b). This strong reduction of tumor growth was also due to apoptosis of OPM-2 cells in the graft at these high concentrations (1 nmol). OPM-2 myeloma cells undergo apoptosis at high concentrations of Aplidin and analogs (Additional file 2: Table S2).

Discussion
Tumor development and progression strongly depend on angiogenesis [2,3]. Thus, inhibition of angiogenesis by "antiangiogenic drugs" represents an important tool for holding tumors in a small avascular state and inhibiting their growth and metastasis [5,7]. Despite extensive research only few drugs primarily targeting "VEGF signaling" have reached clinical practice and currently face new challenges such as the development of resistances [7,8,22]. Therefore, there is an urgent need for novel compounds that act "antiangiogenically" by stopping endothelial cell proliferation without inducing apoptosis in the vascular network of the body and/or affecting coagulation processes. Next to the proteasome inhibitor bortezomib [13], the cyclodepsipeptide Aplidin™ originally isolated from the Mediterranean tunicate Aplidium albicans, has been demonstrated to exert antiangiogenic effects in vitro and in vivo [12]. This study identified two more easy to synthesize Aplidin analogs as potent antiangiogenic drugs, which in the low nM range induced cell cycle arrest in mitotic endothelial cells. Both analogs were less effective in the induction of apoptosis than the original Aplidin when used at same low nM concentrations. The lower toxicity might result by diminished uptake into human cells due to modification of side chains. Both analogs induced cell cycle arrest in G1 phase and induced expression of the cyclin-dependent kinase inhibitor p16 INK4A . Induction of p16 INK4A and senescenceassociated beta galactosidase is one of the hallmarks of premature senescence and terminal growth arrest. Indeed, it was recently demonstrated by Jenkins et al. that oxidative stress, in particular radical oxygen species, induce p16 INK4A and arrest cells in G1 [23]. Nevertheless, we cannot provide the proof that induction of p16 INK4A is the main trigger for the observed terminal growth arrest or if there are still other mechanisms.
Further analyses revealed that Aplidin analogs induced oxidative stress in endothelial cells. Induction of oxidative stress has already been observed in breast and ovarian cancer cell lines after treatment with Aplidin™ [10,24]. In comparison to these studies performed on tumor cells with high concentrations of Aplidin™ (400 nM), we were not able to rescue cells after adding antioxidants like Nacetyl-cysteine, although we used by far lower nM concentrations of Aplidin analogs. Primary endothelial cells remained in terminal growth arrest and could not be rescued by mitogenic growth medium for further proliferation. Our observations indicate that the cellular senescence program is activated by both Aplidin analogs PM01215 and PM02781.
With regard to changes in endothelial cells after treatment with PM01215 and PM02781 we observed alterations in vascular maturation factors. Release of the Dickkopf Homolog 3 (DKK-3) was downregulated after treatment with Aplidin analogs. DKK-3 has been shown to act on endothelial cells as a differentiation factor [20,21] by inhibiting TGF-beta/Smad signaling [25] and supporting or regulating Wnt/beta-catenin activity [26]. Furthermore, we observed downregulation of the VEGF target gene VASH1. The encoded vasohibin protein has been shown to induce vascular maturation by supporting coverage of blood vessels with smooth muscle cells and pericytes [18,19]. Noteworthy, it was recently shown that downregulation of vasohibin induces oxidative stress and premature senescence in human endothelial cells [27]. Thus, Aplidin analogs could enhance oxidative stress and senescence processes by downregulating vasohibin.
In particular, in our chicken multiple myeloma xenograft models we were able to demonstrate potent antiangiogenic and antimyeloma activities of both Aplidin analogs in sublethal concentrations. Our results of the novel Aplidin analogs are in line with the data of Cers et al. demonstrating the antiangiogenic and anti-myeloma activities of the original Aplidin™ in the 5TMM syngeneic model of multiple myeloma [9].

Conclusion
Our data give evidence that both novel Aplidin analogs show potent antiangiogenic activities in vitro and in vivo assays at low nanomolar concentrations. Therefore, PM01215 and PM02781 are attractive candidates for the development of new antiangiogenic cancer drugs and warrant further analysis in mouse tumor models to study effects on tumor growth and blood vessel formation.

Ethical standards
According to the Austrian law no local ethical approval is required for commercially available human primary cells. According to the Office of Laboratory Animal Welfare of the US public health service avian embryos are not considered live vertebrate animals until hatching. The NIH Office of Laboratory Animal Welfare has provided written guidance in this area (http://www.grants. nih.gov/grants/olaw/references/ilar91.htm and NIH Publication No.: 06-4515). Residual blood samples from volunteers were used after obtaining written informed consent of healthy donors for scientific research projects. From the local ethic commission of the Medical University of Innsbruck (UN4012) we have a permission to use anonymized, voluntarily donated peripheral blood as controls. This is now stated in the ethics paragraph on page 18, line 458.

Additional files
Additional file 1: Table S1. Characterization of human primary cells.

(DOCX 14 kb)
Financial support This work was supported by the EU FP7 project Optatio (No: 278570) and was performed in the framework of BB's PhD program in Molecular Cell Biology and Oncology (MCBO) supported by the Austrian Science Fund (FWF Grant No. W1101).
Author details