Concentration- and schedule-dependent effects of chemotherapy on the angiogenic potential and drug sensitivity of vascular endothelial cells
- First Online:
- Cite this article as:
- Pasquier, E., Tuset, M., Street, J. et al. Angiogenesis (2013) 16: 373. doi:10.1007/s10456-012-9321-x
The anti-angiogenic activity of chemotherapy is both dose- and schedule-dependent. While conventional maximum tolerated dose (MTD) chemotherapy exerts only mild and reversible anti-angiogenic effects, low-dose metronomic (LDM) chemotherapy was developed to specifically target tumour angiogenesis. However, the long-term effects of either MTD or LDM chemotherapy on vascular endothelial cells have never been investigated. Here, we demonstrated that repeated exposure to MTD and LDM chemotherapy differentially impact on the angiogenic potential and chemosensitivity of immortalized endothelial cells. Repeated MTD vinblastine treatment of vascular endothelial cells led to an increased proliferation rate and resistance to paclitaxel. In contrast, repeated LDM treatment with vinblastine or etoposide impaired the angiogenic potential of endothelial cells and increased their chemosensitivity. This effect was associated with a significant decrease in βII- and βIII-tubulin expression. Functional analysis using siRNA showed that silencing the expression of βIII-tubulin in endothelial cells significantly decreased their capacity to form vascular structures and increased their sensitivity to the anti-angiogenic and vascular-disrupting effects of chemotherapy, whereas silencing βII-tubulin expression had no effect. Collectively our results show that LDM chemotherapy impairs the angiogenic potential of endothelial cells while increasing their chemosensitivity—an effect at least in part mediated by the down-regulation of βIII-tubulin expression. Furthermore, our study suggests that βIII-tubulin represents an attractive therapeutic target to increase the anti-angiogenic effects of chemotherapy and overall anti-tumour efficacy.
KeywordsCancer Angiogenesis Metronomic chemotherapy Vascular endothelial cells ABC transporters βIII-tubulin
Bone marrow derived endothelial cells
Fetal calf serum
Fibroblast growth factor β
Human microvascular endothelial cell line 1
Maximum tolerated dose
Reverse-transcriptase polymerase chain reaction
Small interfering RNA
The discovery of the central role played by tumour angiogenesis in cancer progression and metastasis has not only led to the development of novel anti-cancer drugs specifically targeting this key biological process but also instigated the re-examination of conventional chemotherapy agents. A number of studies thus revealed that many of the chemotherapeutic drugs commonly used in the clinic are effective at treating cancer because they can affect both the cancer cells and the tumour vasculature [1–3]. This led to the development of novel treatment modalities aimed at promoting the anti-angiogenic activity of chemotherapy. These new treatment schedules, called metronomic chemotherapy or metronomic scheduling of anti-cancer treatment (MSAT) [4–6], are based on more frequent administration of chemotherapeutic drugs at relatively low dose, as compared to conventional chemotherapy, and with no prolonged drug-free interval. Following initial demonstration of efficacy in pre-clinical models [1, 7], metronomic chemotherapy has shown promising results in a number of clinical applications (reviewed in [5, 6]), such as advanced breast cancer in adults [8, 9] or recurrent medulloblastoma in children . Furthermore, the efficacy of metronomic chemotherapy is currently investigated in several phase II and III clinical trials worldwide in a variety of human malignancies (http//:clinicaltrials.gov).
Despite accumulating evidence of clinical benefits, the complex and multi-faceted mechanism of action of metronomic chemotherapy has only been partially unveiled and warrants further investigation. One of the initial rationales for the development of anti-angiogenic therapies in general, and metronomic chemotherapy in particular, was based on the idea that, unlike cancer cells, vascular endothelial cells are genetically stable and thus less likely to acquire drug resistance [1, 11, 12]. However, accumulating evidence suggests that this hypothesis is, at least in part, incorrect. Endothelial cells isolated from tumour blood vessels were found to display tumour-specific genetic abnormalities [13–15]. In addition, studies have shown that tumour endothelial cells can be intrinsically resistant to chemotherapeutic drugs, such as adriamycin and temozolomide [16, 17], and even acquire drug resistance through VEGF-induced up-regulation of MDR1 expression . The long-term effects of chemotherapy on vascular endothelial cells remains however unexplored.
Recent studies revealed that chemotherapy dosing schedule strongly influences drug resistance development in cancer cells. Weekly docetaxel treatment was found to induce the up-regulation of a number of genes involved in drug resistance, including TUBB3 (i.e. the gene encoding β-III tubulin) and ATP-binding cassette (ABC) transporters ABCB1 and ABCC10, in xenografted ovarian tumours whereas continuous docetaxel treatment did not . Furthermore, Kerbel and colleagues demonstrated that extended low-dose metronomic (LDM) cyclophosphamide therapy could lead to drug resistance in prostate cancer cells in vivo but these cells retained sensitivity to maximum-tolerated dose (MTD) cyclophosphamide . This strongly suggests that resistance to LDM chemotherapy in cancer cells imparts distinct mechanisms of resistance to MTD chemotherapy. Here, we hypothesized that, similarly to what was reported in tumour cells, repeated exposure to either LDM or MTD chemotherapy would differentially impact on the angiogenic potential and chemosensitivity of endothelial cells, through modulation of the expression of genes involved in drug resistance, such as ABC transporters and β-tubulin isotypes.
Materials and methods
BMH29L cells are bone marrow-derived endothelial cells (BMECs) that were immortalized by ectopic expression of human telomerase reverse transcriptase . They were grown in Medium 199 (Invitrogen) containing 10 % heat-inactivated Fetal Calf Serum (FCS), 5 % male human serum, AB only (Sigma-Aldrich, Castle Hill, Australia), 1 % penicillin and streptomycin, 1 % heparin, 5 ng/mL recombinant human FGFβ (fibroblast growth factor β; Sigma-Aldrich) and 20 μg/mL Endothelial Cell Growth Factor (Roche, Dee Why, Australia). HMEC-1 (Human Microvascular Endothelial Cell line 1) cells were originally developed by Prof Ades  and obtained from the Cell Culture Laboratory in the Hôpital de la Conception (Assistance Publique Hôpitaux de Marseille, Marseille, France). They were grown in MCDB-131 medium (Invitrogen, Mount Waverley, Australia) containing 10 % heat-inactivated FCS, 2 mM l-glutamine, 1 % penicillin and streptomycin, 1 μg/mL hydrocortisone and 10 ng/mL epithelial growth factor (BioScientific, Gymea, Australia). Both cell lines were routinely maintained in culture on 0.1 % gelatin-coated flasks at 37 °C and 5 % CO2. Cell lines were regularly screened and are free from mycoplasma contamination.
Long-term drug treatment
BMH29L cells were incubated with vinblastine (VLB) or etoposide (VP16) following a maximum-tolerated dose (MTD) conventional schedule or a low-dose metronomic (LDM) schedule for a total duration of 100 days (Online Resource 1). In the MTD schedule, BMH29L cells were treated every 2 weeks for a total of 7 courses with drug-free medium (Ctrl-MTD) or containing VLB or VP16 at the IC80 for cell proliferation (i.e. 20 nM and 50 μM of VLB and VP16, respectively). In the LDM schedule, BMH29L cells were treated 5 days a week for a total of 14 courses with drug-free medium (Ctrl-LDM) or containing VLB or VP16 at the highest non-toxic concentration (i.e. 1 nM and 0.5 μM of VLB and VP16, respectively). This led to the establishment of 6 new BMH29L subclones: Ctrl-LDM, Ctrl-MTD, VLB-LDM, VLB-MTD, VP16-LDM and VP16-MTD. These subclones were cryopreserved at the end of the long-term treatment and not kept in culture for more than 1 month for all subsequent experiments.
Doubling time and growth inhibition assay
Following long-term drug treatment, BMH29L subclones were seeded onto 6-well plates at a cell density of 20,000 cells/well and counted every 24 h for 4 days using the trypan blue exclusion method. Doubling times were determined by mathematical regression using GraphPad Prism 4 software (GraphPad Software Inc, La Jolla, CA). Growth inhibition assays were performed as previously described . Briefly, cells were seeded at 1,500 cells/well (BMH29L) or 3,750 cells/well (HMEC-1) in 96-well gelatin-coated plates. After 24 h, cells were treated with a range of concentrations of chemotherapeutic drugs and after 72 h drug incubation, metabolic activity was detected by addition of Alamar blue and spectrophotometric analysis. Cell proliferation was determined and expressed as a percentage of untreated control cells. The determination of IC50 and IC80 values was performed by point-to-point fit spline analysis using GraphPad Prism 4 software.
The expression of ABC transporters ABCB1, ABCC1, ABCC2 and ABCC10 and β-tubulin genes TUBB, TUBB2A and TUBB3 was examined in BMH29L subclones using real-time quantitative RT-PCR, as previously described [24, 25]. Total RNA was extracted and DNAse treated using the Qiagen Mini RNeasy kit according to the manufacturer instructions (Qiagen, Doncaster, Australia). cDNA synthesis was performed using High capacity cDNA reverse transcription kit with RNAse inhibitor according to the manufacturer instructions (Applied Biosystem, Melbourne, Australia). Real time PCR was run on 7900HT Fast Real-Time PCR system using either Taqman® gene expression assays (Applied Biosystems) for ABCB1 (Hs00184500), ABCC1 (Hs01561503), ABCC10 (Hs00375716) and endogenous control HPRT1 (4326321E) or Power SYBR® green (Applied Biosystems) for TUBB (QT00089775), TUBB2A (QT01677326), TUBB3 (QT00083713) and endogenous control GAPDH (QT01192646). ABCC2 forward and reverse primer sequences were 5′-AGAGAACAGCTTTCGTCGAACAC-3′ and 5′-CATTCCGAGTTTTCAAGGAGTTTC-3′, respectively. ABCC2 probe sequence was ACCTAGAACTGCGGCTA. Gene expression levels were determined using the ΔΔC t method, normalized to the HPRT1 control for ABC transporters and the GAPDH control for β-tubulin genes, and expressed relative to a calibrator .
Radiolabelled drug accumulation assay
For drug accumulation studies, BMH29L subclones, seeded in 12-well plates, were incubated for 4 h at 37 °C with [3H]-vincristine (15.8 Ci/mmol; final concentration 50 nM) in presence or absence of 10 μM verapamil. Cells were then washed thrice with ice-cold PBS to eliminate the extracellular tritiated drug and lysed in 0.5 M NaOH. Intracellular [3H]-vincristine concentration was determined by β-scintillation counting and normalized to protein content, as determined by BCA assay .
In vitro Matrigel™ assay
Matrigel™ (BD Biosciences, North Ryde, Australia) assay was used to determine the effects of repeated exposure to chemotherapy and βII and βIII tubulin knockdown on the angiogenic potential and chemosensitivity of endothelial cells, as previously described . For the anti-angiogenesis analysis, cells were treated with different drug solutions 20 min after seeding on Matrigel and photographs were taken after 8 h drug incubation using the 5X objective of an Axiovert 200 M fluorescent microscope coupled to an AxioCamMR3 camera driven by the AxioVision 4.7 software (Carl Zeiss, North Ryde, Australia). For the vascular-disruption analysis, cells were allowed to undergo morphogenesis and form capillary-like structures for 6 h before drug treatment was initiated. Photographs were then taken using the same microscope device after 2 h drug incubation. The anti-angiogenic and vascular-disrupting effects were then quantitatively evaluated by measuring the total surface area of capillary tubes formed in at least 10 view fields per well using AxioVision 4.7 software.
Gene silencing by small interfering RNA
βII- and βIII-tubulin gene expression were silenced in endothelial cells using siRNA sequences whose potency and specificity have been validated previously [28, 29] and obtained from Dharmacon (Thermo Fisher Scientific, Scoresby, Australia) and Qiagen (Qiagen), respectively. The optimum amount of siRNA was determined to be 200 pmol (data not shown) and was used in all subsequent experiments. A non-silencing control siRNA, which has no sequence homology to any known human gene sequence, was used as a negative control in all experiments (Qiagen). Cells were transfected using the Nucleofector® II device (Lonza, Mount Waverley, Australia) as previously described . Briefly, HMEC-1 and BMH29L cells were resuspended in nucleofector® solution R and V, respectively, and transfected with siRNA using specific nucleofector® programs (T-016 and S-003 for HMEC-1 and BMH29L, respectively). In all subsequent experiments, drug treatment was initiated 72 h after siRNA transfection—when the knockdown of the targeted gene was the most effective.
Western blotting analysis
Total cellular proteins from endothelial cells (10–15 μg) were resolved on 12 % SDS-PAGE before electrotransfer onto nitrocellulose membrane. Immunoblotting was done using antibodies directed against βI-tubulin (clone SAP 4G5; Abcam), βII-tubulin (clone 7B9; Chemicon), βIII-tubulin (clone TUJ1; Chemicon), βIV-tubulin (clone ONS 1A6; Sigma-Aldrich), total β-tubulin (clone TUB 2.1; Sigma-Aldrich) and glyceraldehyde-3-phosphate dehydrogenase (GADPH; Abcam). The membranes were then incubated with horseradish peroxidase–conjugated IgG secondary antibodies, and protein was detected with ECL Plus (GE Healthcare Life Sciences). The blots were scanned and densitometric analysis performed as previously described .
All experiments were performed at least in triplicate. Statistical significance was determined using two-sided student’s t-test in the GraphPad Prism 4 software (GraphPad Software, Inc).
Conventional and metronomic chemotherapy differentially impact on the angiogenic potential of vascular endothelial cells
Matrigel assay was then used to analyse the effects of long-term chemotherapy treatment on the capacity of endothelial cells to undergo morphological differentiation into vascular structures in vitro (Fig. 1b). Repeated exposure to etoposide resulted in impaired angiogenic properties, irrespective of the treatment schedule (−33 and −39 % vessel formation for VP16-LDM and VP16-MTD cells as compared to control cells, respectively; p < 0.05). In contrast, the impact of vinblastine on the angiogenic potential of endothelial cells was schedule-dependent. VLB-LDM cells displayed impaired angiogenic properties (−20 % as compared to control cells; p < 0.05) whereas VLB-MTD cells were able to form vascular structures to the same extent as control untreated cells.
Conventional and metronomic chemotherapy differentially impact on the chemosensitivity of vascular endothelial cells
Angiogenesis assay further confirmed these results. Repeated exposure to etoposide resulted in increased sensitivity to the anti-angiogenic effects of vinblastine (Fig. 2d; p < 0.01), etoposide (Fig. 2e; p < 0.05) and paclitaxel (Fig. 2f; p < 0.05), irrespective of the treatment schedule. Long-term effects of vinblastine however were schedule-dependent. Repeated exposure to vinblastine at MTD did not affect the sensitivity of endothelial cells to the anti-angiogenic activity of chemotherapeutic drugs whereas repeated exposure to vinblastine at low concentration significantly increased sensitivity (Fig. 2d–f; p < 0.05).
Repeated exposure to chemotherapy modulates the expression of ABC transporter without dramatically affecting intracellular drug accumulation
Radiolabelled drug uptake studies were then performed to determine whether the changes in ABCB1 and ABCC2 gene expression were functional and associated with altered intracellular accumulation of the P-gp and ABCC2 substrate, vincristine. Our results showed that repeated exposure to chemotherapy only marginally altered the intracellular accumulation of tritiated vincristine ([3H]-VCR) in endothelial cells (Fig. 3b). Repeated treatment with etoposide thus led to a slight decrease in [3H]-VCR uptake (−11 and −9 % for VP16-LDM and VP16-MTD as compared to control cells; p < 0.06) whereas repeated exposure to vinblastine did not. Furthermore, the addition of P-gp inhibitor, verapamil, did not significantly increase [3H]-VCR uptake in the BMH29L subclones, with the exception of a 7 % increase observed in VP16-MTD cells (p < 0.05). This is in marked contrast with the 234 % increase in [3H]-VCR uptake induced by verapamil in P-gp expressing neuroblastoma cells . Our results thus demonstrated that the increase in ABCB1 gene expression observed following repeated exposure to chemotherapy did not translate into an increase in P-gp activity.
Repeated exposure to chemotherapy induces significant changes in β-tubulin isotype expression
βIII-tubulin expression mediates the sensitivity of endothelial cells to the anti-angiogenic effects of chemotherapy
Inhibition of tumour angiogenesis is a recognized mechanism of action of chemotherapy [2, 3, 34]. However, the long-term effects of chemotherapeutic drugs on vascular endothelial cells have not been explored. In this study, we used bone-marrow derived endothelial cells immortalized by ectopic expression of human telomerase reverse transcriptase  to analyse the impact of repeated exposure to chemotherapy on the angiogenic potential and chemosensitivity of vascular endothelial cells. We found that the long-term effects of chemotherapy on endothelial cells were drug-, concentration- and schedule-dependent.
First, we demonstrated that repeated exposure to vinblastine at the maximum-tolerated dose (MTD) increased the angiogenic potential of vascular endothelial cells by promoting their proliferation without altering their capacity to form vascular structures on Matrigel. Furthermore, this treatment regimen led to a 1.8-fold increase in paclitaxel resistance in endothelial cells. Recently, VEGF was found to increase resistance to paclitaxel in tumour endothelial cells by inducing ABCB1 expression . Previous studies have also reported that endothelial cells isolated from tumour tissues, such as hepatocellular carcinoma  and malignant glioma , displayed increased angiogenic potential and intrinsic drug resistance, which may limit the effectiveness of chemotherapy. Here, we provide the first evidence that endothelial cells can acquire some level of drug resistance as a result of repeated exposure to chemotherapy in vitro and may contribute to drug resistance acquisition in tumours in vivo following MTD chemotherapy. Further investigations are currently underway to validate our findings in vivo.
The increase in angiogenic potential and drug resistance induced by MTD chemotherapy was in marked contrast with the long-term effects of low-dose metronomic (LDM) chemotherapy. Indeed, our results demonstrated that repeated exposure to LDM chemotherapy led to sustained impairment of the angiogenic potential of endothelial cells and significantly increased their chemosensitivity. Cells treated with LDM chemotherapy thus exhibited a slower proliferation rate, decreased capacity to form vascular structures on Matrigel and increased sensitivity to the anti-proliferative and anti-angiogenic effects of chemotherapy agents.
Our findings are in accordance with previous studies showing that the presence and length of treatment-free intervals are major determinants in the development of drug resistance in tumour cells, and that drug resistance acquisition can be hindered, prevented and even circumvented simply by modifying the dose and schedule of chemotherapy administration [1, 19, 20]. In the seminal study that led to the development of metronomic chemotherapy, Browder et al. demonstrated that using more frequent cyclophosphamide administration could re-sensitize Lewis lung carcinoma tumours that were made resistant to the same drug . More recently, cyclophosphamide was found to generate dramatically different resistance phenotypes in prostate and breast cancer cells in vivo, depending on the dose and schedule of treatment . While cancer cells made resistant to MTD cyclophosphamide in vivo showed stable drug resistance in vitro and in vivo, cells made resistant to LDM cyclophosphamide in vivo retained sensitivity to chemotherapy agents in vitro and to MTD cyclophosphamide in vivo. Here, we report differential effects of vinblastine in endothelial cells when administered at the MTD or following a LDM schedule, similar to that observed with taxanes and cyclophosphamide in tumour cells. In contrast, the effects of etoposide appeared to be less schedule-dependent in endothelial cells and repeated exposure led to impaired angiogenic potential and increased chemosensitivity, irrespective of the treatment schedule.
Since intracellular drug accumulation is the basic requirement for chemotherapy agents to exert their effects, upregulation of ABC transporters is of crucial importance in drug resistance acquisition . Here we found that repeated exposure to chemotherapy induced a significant upregulation of ABCB1 gene expression, irrespective of the treatment schedule and the type of chemotherapy agent used. In contrast, the effects on ABCC2 gene expression were schedule-dependent, with an increase observed in VLB-MTD cells and a decrease in VLB-LDM and VP16-LDM cells. However, the changes in ABC transporter gene expression did not correlate with major changes in intracellular drug accumulation. All 6 endothelial subclones were found to accumulate very high levels of tritiated vincristine (~45 pmol per mg of protein), which was two- to six-fold higher than in neuroblastoma cells. This high vincristine uptake was unaltered by co-administration of verapamil, suggesting that the increase in ABCB1 gene expression did not translate into increased drug efflux. Interestingly, increased uptake of paclitaxel in endothelial cells was previously reported and proposed as a potential mechanism for the anti-angiogenic activity of paclitaxel . Here we demonstrated that endothelial uptake of Vinca alkaloids is also high and may contribute to their potent anti-vascular properties .
The lack of correlation between changes in ABC transporter expression and intracellular drug accumulation strongly suggests that the effects of repeated exposure to chemotherapy on the drug sensitivity of endothelial cells are not mediated by ABC transporters. We therefore sought to investigate other candidate genes that may be involved. The expression level of specific β-tubulin isotypes has been shown to influence drug sensitivity in a range of cancer cell lines and malignancies, including lung, ovarian, breast and prostate cancers (reviewed in ). Here we found that changes in chemosensitivity of endothelial cells induced by repeated exposure to chemotherapy were associated with significant changes in β-tubulin isotype expression. In particular, all 3 BMH29L subclones displaying increased drug sensitivity (VLB-LDM, VP16-LDM and VP16-MTD) were found to express lower protein levels of βII- and βIII-tubulin. This is the first demonstration that repeated exposure to microtubule-targeting agents, such as vinblastine, and other classes of chemotherapy agent, such as the topoisomerase inhibitor etoposide, can result in significant changes in tubulin isotype expression in non-cancer cells, most likely through transcriptional repression.
Using a series of elegant experiments, Piquette-Miller and colleagues demonstrated that continuous administration of taxanes (i.e. paclitaxel and docetaxel) was considerably more efficacious than intermittent taxane therapy (i.e. once or thrice weekly) against ovarian cancer cells in vitro and in vivo [19, 38, 39]. Interestingly, intermittent docetaxel treatment of ovarian tumours led to significant upregulation of ABCB1, ABCC10, bcl2 and TUBB3 among other genes involved in drug resistance whereas continuous docetaxel did not induce upregulation of any analysed genes but instead led to a downregulation of ABCC10, TUBB3 and stathmin . Our results are in accordance with these studies and confirm the influence of treatment schedule on the modulation of the expression of genes involved in drug resistance not only in tumour cells but also in endothelial cells.
We previously reported relatively high expression of βII- and βIII-tubulin in vascular endothelial cells  and showed that both these tubulin isotypes strongly influence the response of non-small cell lung carcinoma cells (NSCLC) to chemotherapeutic drugs [25, 28, 29, 41]. In the present study, functional analysis using gene silencing technology showed that βIII-tubulin, but not βII-tubulin, plays a crucial role in the anti-angiogenic activity of chemotherapy. Knocking down βIII-tubulin expression thus resulted in a small, but significant, decrease in the angiogenic potential of endothelial cells and increased their sensitivity to the anti-angiogenic and vascular-disrupting activity of chemotherapeutic drugs. This finding provides further evidence of the key role played by βIII-tubulin in tumour progression, angiogenesis and drug resistance. Recently, βIII-tubulin was identified as a marker of angiogenic perivascular cells in rat mesenteric tissues during active capillary sprouting . Furthermore, a large number of studies have linked abnormal or high levels of βIII-tubulin expression with more aggressive and drug-resistant phenotypes in a range of tumour types, including lung, ovarian, breast and gastric cancers (reviewed in [33, 43]). In NSCLC for instance, expression of βIII-tubulin is associated with poorly differentiated tumour tissue, high-grade malignancy and increased metastatic potential , as well as lower response rate to paclitaxel/vinorelbine-based chemotherapy and shorter progression-free and overall survival [44, 45]. Collectively the results of our current study together with our previous work [25, 28, 29, 41] and accumulating clinical data strongly suggest that βIII-tubulin represents an attractive therapeutic target to increase the anti-angiogenic effects of chemotherapy and overall anti-tumour efficacy.
This work was supported by the Children’s Cancer Institute Australia for Medical Research, which is affiliated with the University of New South Wales and Sydney Children’s Hospital, and by grants from the Cancer Institute New South Wales (EP), the Balnaves Foundation (EP) and the National Health and Medical Research Council (MK). EP was supported by a Cancer Institute New South Wales “Early Career Development” Fellowship and MK is supported by a National Health and Medical Research Council Senior Research Fellowship. The authors would like to thank Amanda Russell and Claudia Flemming (Children’s Cancer Institute Australia) for their assistance with the ABC transporter gene expression analysis and the radiolabelled drug accumulation assay, respectively.
Conflict of interest
The authors declare that they have no conflict of interest.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.