Enhancing the anti-angiogenic action of histone deacetylase inhibitors

Background Histone deacetylase inhibitors (HDACIs) have many effects on cancer cells, such as growth inhibition, induction of cell death, differentiation, and anti-angiogenesis, all with a wide therapeutic index. However, clinical trials demonstrate that HDACIs are more likely to be effective when used in combination with other anticancer agents. Moreover, the molecular basis for the anti-cancer action of HDACIs is still unknown. In this study, we compared different combinations of HDACIs and anti-cancer agents with anti-angiogenic effects, and analysed their mechanism of action. Results Trichostatin A (TSA) and α-interferon (IFNα) were the most effective combination across a range of different cancer cell lines, while normal non-malignant cells did not respond in the same manner to the combination therapy. There was a close correlation between absence of basal p21WAF1 expression and response to TSA and IFNα treatment. Moreover, inhibition of p21WAF1 expression in a p21WAF1-expressing breast cancer cell line by a specific siRNA increased the cytotoxic effects of TSA and IFNα. In vitro assays of endothelial cell function showed that TSA and IFNα decreased endothelial cell migration, invasion, and capillary tubule formation, without affecting endothelial cell viability. TSA and IFNα co-operatively inhibited gene expression of some pro-angiogenic factors: vascular endothelial growth factor, hypoxia-inducible factor 1α and matrix metalloproteinase 9, in neuroblastoma cells under hypoxic conditions. Combination TSA and IFNα therapy markedly reduced tumour angiogenesis in neuroblastoma-bearing transgenic mice. Conclusion Our results indicate that combination TSA and IFNα therapy has potent co-operative cytotoxic and anti-angiogenic activity. High basal p21WAF1 expression appears to be acting as a resistance factor to the combination therapy.


Background
Acetylation and deacetylation of histones by histone acetyltransferases and histone deacetylases (HDACs) alter chromatin structure and modulate transcriptional regulation (reviewed in [1][2][3]. Inhibitors of HDACs (HDACIs) are emerging as a new class of anticancer agents. HDACIs induce cancer cell differentiation, growth arrest, programmed cell death, and inhibit tumour-driven angiogenesis [1,3]. Clinical trials with HDACIs in cancer patients demonstrate that HDACI treatment leads to tumour regression and symptomatic improvement in some heavily pre-treated and multiply relapsed patients, with a surprisingly low side-effect profile [1,4]. However, a large proportion of the patients are not sensitive to the treatment, demonstrating the need to examine the effectiveness of HDACIs in combination with other anti-cancer agents.
Angiogenesis is vital for tumor progression and metastasis [5,6]. As anti-angiogenic therapy is generally less toxic and better tolerated than conventional cytotoxic chemotherapy, strategies which combine anti-angiogenic agents with other anti-cancer drugs have been the focus of current clinical trials to widen the therapeutic index. The interferons (IFNs) are a family of naturally occurring cytokines with anti-proliferative and anti-angiogenic effects [7,8]. Through inhibiting pro-angiogenic gene expression and acting directly on endothelial cells, α-interferon (IFNα) suppresses angiogenesis and tumour growth in vitro and in vivo [7,9]. Rapamycin and its derivatives also inhibit tumour cell proliferation and angiogenesis by acting on the mammalian target of rapamycin and suppressing the transcriptional activity of pro-angiogenic hypoxia-inducible factor 1α (HIF1α), (reviewed in [10]). While clinical trials with IFNα, rapamycin and its derivatives used as single agents have shown some effects, none of the drugs are effective alone in the majority of patients.
It has been reported that a combination therapy with the HDACI, valproate (VPA), and IFNα exerts synergistic anticancer effects in neuroblastoma BE(2)-C cells both in vitro and in vivo [11,12]. Here we evaluated the anticancer actions of combination therapy with HDACIs (Trichostatin A [TSA] or VPA) and anti-cancer agents with anti-angiogenic function (IFNα, rapamycin), and, sought to determine their mechanism of action.

TSA and IFNα exerted co-operative cytotoxic effects in cancer cell lines from a range of different tissue origins
The combination of the HDACI, VPA, and IFNα demonstrated synergistic combinational anti-cancer effects in neuroblastoma BE(2)-C cells both in vitro and in vivo [11,12]. We investigated the synergistic anti-cancer effect of IFNα combined with other HDACIs, and, in cancer cell lines of other tissue origins. We treated breast, lung, colon and prostate cancer cells and MRC-5 normal non-malignant fibroblasts with control, 0.02 μM TSA and/or 500 IU/ml IFNα, and, then assessed for cell viability. As shown in Figure 1A, all of the cancer cell lines tested were sensitive to the cytotoxic effects of the combination, and there was a significant cooperative effect of TSA and IFNα in eight of the nine cell lines tested, with MDA-MB-468 as the only exception. MCF-7, Calu-6, H460, LNCaP, DU-145, HT-29, Caco-2 and BE(2)-C cells were all sensitive to TSA, generally less sensitive to IFNα, and significantly more sensitive to TSA and IFNα combined. MDA-MB-468 breast cancer cells were sensitive to IFNα but resistant to TSA, and no more sensitive to the combination than IFNα alone. When cell sensitivity to the combination treatment was calculated as a percentage of TSA alone (or IFNα alone in case of MDA-MB-468), BE(2)-C, HT-29 and Calu-6 were found to be the most sensitive ( Figure 1A). Importantly, the normal non-malignant MRC-5 fibroblasts were resistant to the treatment of TSA alone, IFNα alone and TSA plus IFNα combination therapy ( Figure  1B). Immunoblot analysis of acetylated histone H3 revealed that treatment with TSA alone or TSA plus IFNα for 6 hours induced drastic histone acetylation in the MRC-5 cells ( Figure 1B).

SAHA and IFNα exerted co-operative cytotoxic effects in cancer cell lines, but not in normal cells
The HDACI, SAHA (vorinostat), is in clinical use for the treatment of cutaneous T-cell lymphoma. We, therefore, tested whether a combination of SAHA and INFα exerted co-operative anti-cancer effects. Neuroblastoma BE(2)-C, breast cancer MCF-7, and normal lung fibroblast MRC-5 cells were treated with control, 0.5 μM SAHA, 500 IU/ml INFα or SAHA plus INFα for 3 days. Alamar blue assays revealed that SAHA and INFα co-operatively reduced the viability of BE(2)-C and MCF-7 cells, although the magnitude was smaller than TSA and INFα. A combination of SAHA and INFα did not co-operatively reduce the viability of MRC-5 cells ( Figure 2).

The effects of other HDACIs and anti-cancer agents used in combination
We compared the cytotoxicity of the TSA and IFNα combination ( Figure 1A) with combinations of another HDACI VPA and IFNα ( Figure 3A). The effect of VPA and IFNα combination therapy on cell viability was similar to TSA and IFNα for the BE(2)-C neuroblastoma cells. However, TSA and IFNα were more effective in MCF-7 and Calu-6 cells than VPA and IFNα. Similar to TSA and IFNα, VPA and IFNα did not show any co-operative cytotoxic effects on normal MRC-5 fibroblasts ( Figure 3C). We next compared the cytoxicity of the TSA and IFNα combination ( Figure 1A), the VPA and IFNα combination ( Figure  3A) with VPA combined with another emerging anticancer agent with both cytotoxic and anti-angiogenic actions, rapamycin [10] in BE(2)-C and MCF-7 cells ( Figure 3B). The VPA and rapamycin treatment had significant cytotoxic effects compared with VPA alone, but the magnitude of these effects was much smaller than VPA and IFNα or TSA and IFNα.
TSA and IFNα exerted co-operative cytotoxic effects in cancer cell lines from a range of different tissue origins, but not in nor-mal non-malignant cells Figure 1 TSA and IFNα exerted co-operative cytotoxic effects in cancer cell lines from a range of different tissue origins, but not in normal non-malignant cells. A. Neuroblastoma [BE(2)-C], breast (MCF-7 and MDA-MB-468), lung (H460 and Calu-6), prostate (DU-145 and LNCaP), and colon (HT-29 and Caco-2) cancer cells were treated with control (Cont), 0.02 μM TSA and/or 500 IU/ml IFNα for 72 hours. Cell viability was examined using the Alamar blue assay, measured as optical density (OD) units of absorbance, and expressed as the absorbance of treated over control samples (ie., % viable cells). ** p < 0.01, *** p < 0.001. B. MRC-5 cells were treated with control, 0.02 μM TSA and/or 500 IU/ml IFNα for 72 hours, and cell viability was assessed as above. Moreover, histone protein was extracted and subject to immunoblot analysis with anti-acetylated histone H3 antibody, after 6 hour exposure to control, TSA and/or IFNα.

Absence of p21 WAF1 expression correlated with sensitivity to TSA and IFNα combination therapy
Up-regulation of p21 WAF1 expression, and p21 WAF1induced cell cycle arrest, have been regarded as one of the main mechanisms through which HDACIs exert their anti-cancer effects [13]. We examined the role of p21 WAF1 in cancer cell sensitivity to the combination therapy. Immunoblot analysis of p21 WAF1 expression was carried out with protein extracted from the eight cell lines of breast, lung, prostate and colon origins ( Figure 4A). p21 WAF1 was basally expressed in untreated H460, DU-145, LNCaP, MCF-7 and MDA-MB-468 cells, but not expressed in Calu-6, HT-29 and Caco-2 cells. Compared with control, TSA induced p21 WAF1 expression in H460, MCF-7 and LNCaP cells. IFNα up-regulated p21 WAF1 only in DU-145 cells, and, combination therapy increased p21 WAF1 in all four cell lines. p21 WAF1 protein was expressed but not altered by treatment with TSA and/or IFNα in MDA-MB-468 cells. Thus, the cancer cell lines which did not have basal expression of p21 WAF1 were generally more sensitive to the combination therapy than those cancer cells expressing p21 WAF1 . This correlation suggested expression of p21 WAF1 might render cancer cells insensitive to the combination therapy.
To determine the role of p21 WAF1 expression in cancer cell sensitivity to TSA and IFNα combination therapy, MCF-7 cells were transfected with control, scrambled siRNA or siRNA specifically targeting p21 WAF1 , and, then treated with control, TSA and/or IFNα. RT-PCR and immunoblot analysis revealed that p21 WAF1 mRNA and protein were knocked down by approximately 75% by the p21 WAF1 siRNA, compared with scrambled control ( Figure 4B). The p21 WAF1 siRNA significantly increased the sensitivity of MCF-7 cells to TSA and IFNα alone, and, in combination, as measured by cell viability assays (p < 0.01) ( Figure 4C).

HDACI and IFNα co-operatively inhibit endothelial cell functions and pro-angiogenic gene expression in cancer cells in vitro
Since HDACIs [2,14] and IFNα [8,9] are known to suppress angiogenesis and tumour growth by acting directly on endothelial cells, we further investigated whether the combination of TSA and IFNα could inhibit endothelial cell function. To exclude the possibility that co-operative anti-angiogenic effects by TSA and IFNα were due to cytotoxicity, we first determined the optimal dosages of TSA and IFNα with Alamar blue cell viability assays. After treatment for 18 hours under normoxic or hypoxic conditions (1% O 2 ), a combination of 0.1 μM TSA and 500 IU/ ml IFNα was found to have no cytotoxicity on endothelial cells within 18 hours after treatment ( Figure 5A). These doses were, therefore, used in all endothelial cell function studies. Surprisingly, TSA or IFNα alone stimulated endothelial cell migration toward the chemoattractant, vascular endothelial growth factor (VEGF) ( Figure 5B). In contrast, the combination of TSA and IFNα suppressed endothelial cell migration under both hypoxic ( Figure 5B) or normoxic conditions (data not shown). Compared with control, IFNα or TSA alone reduced endothelial cell invasion through Matrigel by 35% and 60%, respectively, whereas the combination of TSA and IFNα decreased cell invasion by 80%, under normoxic (data not shown) or hypoxic conditions ( Figure 5C). Under normoxic conditions, compared with control, IFNα or TSA alone decreased the number of complete branches per branching point by 30% and 50%, respectively, while TSA and IFNα did not further decrease complete branches per branching point (data not shown). In contrast, under hypoxic conditions, the combination of TSA and IFNα decreased complete branches per branching point by 50%, while TSA or IFNα alone reduced the average numbers of complete branches from a branching point by only 25% ( Figure 5D).
We next evaluated whether the combination of TSA [15,16] and IFNα [7,17] represses pro-angiogenic gene expression, as measured by RT-PCR, in neuroblastoma BE(2)-C cells. Compared with treatment with TSA or IFNα alone, the combination therapy significantly down-regulated gene expression of HIF1α, VEGF and MMP-9 under normoxic conditions at 72 hours after treatment, while no co-operative effects were observed on the expression of MMP-2, activin A, thrombospondin-1, von Hippel-Lindau protein and bFGF (data not shown). Suppression of HIF1α, VEGF and MMP-9 gene expression by TSA and IFNα was more significant, when compared with TSA or IFNα alone, under hypoxic conditions ( Figure 5E). In the case of HIF1α and VEGF, IFNα alone repressed gene SAHA and IFNα exerted co-operative cytotoxic effects in cancer cell lines, but not in normal non-malignant cells Figure 2 SAHA and IFNα exerted co-operative cytotoxic effects in cancer cell lines, but not in normal non-malignant cells. Neuroblastoma BE(2)-C, breast cancer MCF-7, and normal nonmalignant lung MRC-5 fibroblasts were treated with control, 0.5 μM SAHA and/or 500 IU/ml IFNα for 72 hours. Cell viability was examined using the Alamar blue assay, measured as optical density (OD) units of absorbance, and expressed as the absorbance of treated over control samples (ie., % viable cells). * p < 0.05, ** p < 0.01, *** p < 0.001. expression, however, the combination still had a more significant repressive effect, compared with IFNα alone (p < 0.05). Although MMP-9 gene expression was stimulated by IFNα and TSA alone, the combination suppressed its expression, when compared with control-treated samples (p < 0.05).

TSA and IFNα co-operatively suppress tumour-driven angiogenesis in neuroblastoma-bearingN-Myc transgenic mice
Lastly, we tested whether the combination of TSA and IFNα could co-operatively inhibit tumor-driven angiogenesis in vivo. Abdominal neuroblastoma first became palpable in 100% of homozygote N-Myc transgenic mice at 4 weeks of age [18]. Cohorts of five homozygous MYCN The cytotoxic effects of other HDACI combination therapies Figure 3 The cytotoxic effects of other HDACI combination therapies. A. Neuroblastoma [BE(2)-C], breast (MCF-7), and lung (Calu-6) cancer cell lines were treated with either control, 500 IU/ml IFNα and/or various dosages of VPA for 72 hours B. In separate experiments, BE(2)-C and MCF-7 cells were treated with control, 1 mM VPA and/or various dosages of rapamycin (RAP) for 72 hours. C. Non-malignant lung fibroblast (MRC-5) cells were treated with a range of VPA doses alone, or in combination with 500 IU/ml IFNα. Cell viability was examined by the Alamar blue assay, measured as optical density (OD) units of absorbance, and expressed as a percentage of absorbance for treated samples, over that for control samples (ie., % viable cells). ** p < 0.01, *** p < 0.001.  with Cell Tracker Green CMFDA, migrated through chamber filters toward the chemo-attractant VEGF, and then quantified and expressed as optical density (OD) absorbance units. C. HUVECs were plated into BD BioCoat growth factor-reduced matrigel invasion chambers and treated with control, 0.1 μM TSA and/or 500 IU/ml IFNα for 18 hours. Cells which invaded through the Matrigel were fixed, stained with a Diff Quick staining kit, photographed and then quantified. D. HUVECs were plated onto growth factor-reduced Matrigel in 24 well plates and treated with control, 0.1 μM TSA and/or 500 IU/ml IFNα for 18 hours. Vascular sprouting was quantified by counting the numbers of complete branches per branching point. E. Neuroblastoma BE(2)-C cells were treated with control, 0.02 μM TSA and/or 500 IU/ml IFNα for 72 hours under hypoxic (1% O 2 ) conditions. RNA was extracted and subjected to independent semi-competitive RT-PCR analyses using trans-intron PCR primers, together with primers for the house-keeping gene β-2 microglobulin (β2M). Representative gels for each gene at the 72 hour time point were shown, and fold induction of a target gene by treatment was calculated by ascribing the ratio between the level of expression of a target gene and that of β2M as 1.0 for control treated samples. * p < 0.05, ** p < 0.01, *** p < 0.001. transgenic mice at four weeks of age, were treated with control, IFNα, TSA, or TSA and IFNα for one week after abdominal tumors were first palpable. After mice were sacrificed, tumour volume was measured, and microvasculature assessed by immunohistochemical staining for platelet endothelial cell adhesion molecule 1 (PECAM-1) expression ( Figure 6). When tumour volume was analysed, TSA alone suppressed tumour progression by 87%, while IFNα alone reduced tumour volume by about 36%, compared with control treated mice. The combination of TSA and IFNα reduced tumour volume by more than 92%, although this was not statistically significant compared with TSA treatment alone. When tumour micro-vasculature was assessed by PECAM-1 staining, the use of TSA or IFNα alone, decreased micro-vasculature formation by 32% and 53%, respectively. However, the combination of TSA and IFNα exerted co-operative anti-angiogenic effects, reducing micro-vasculature by almost 90% ( Figure  6)

Discussion
HDACIs have shown great promise in clinical trials in cancer patients. However, a majority of patients have been insensitive to the treatment. In this study, we found that the combination of IFNα with the HDACI TSA induced co-operative cytotoxic effects in almost all cancer cell lines of diverse tissue types, and demonstrated little cytoxicity in normal non-malignant cells. The combination of IFNα with the HDACI SAHA, already in clinical use, also exerted co-operative anti-cancer effects, with little effect on normal cells. The combination of IFNα with another HDACI, VPA, was less effective than IFNα and TSA, but more effective than VPA and rapamycin. These results suggest that HDACI and IFNα combination therapy may be an effective anti-cancer strategy for future clinical trials.
Our data identified p21 WAF1 expression as a key factor responsible for cancer cell resistance to the cytotoxic effects of combination HDACI and IFNα therapy. While IFNα can both induce or suppress p21 WAF1 gene transcription in different cells [19], it is the most common transcriptional target of HDACIs (reviewed in [2]). Previous literature suggested that up-regulation of p21 WAF1 by HDACIs may mediate HDACI-induced cell cycle arrest and growth inhibition [13]. However, recent publications have cast doubt on the role of p21 WAF1 in the action of HDACIs, and, conversely demonstrated that inducible p21 WAF1 reduced HDACI-induced cell death [20][21][22][23][24]. Our TSA and IFNα co-operatively suppress tumour-driven angiogenesis in neuroblastoma bearing transgenic MYCN mice Control data suggests p21 WAF1 expression in some cancer cells acts as a resistance factor for the cytotoxic effects of TSA and IFNα combination therapy.
The individual effects of HDACIs and IFNα on angiogenesis predict a co-operative therapeutic role in blocking tumour angiogenesis. Expression of HDACs is often upregulated under angiogenic stimuli such as hypoxia in cancer cells, and HDACIs can suppress HIF1α expression and its down-stream targets, including VEGF [25]. HDA-CIs have been recently demonstrated to inhibit endothelial cell migration, invasion, vascular sprouting in vitro, and vasculature formation in animal models of cancer [14,16,26]. IFNα can repress VEGF and MMP-9 gene expression, endothelial cell functions, and, inhibit tumour-driven angiogenesis in vivo [9,27]. In our endothelial cell migration experiments, we found in contrast, that either TSA or IFNα alone stimulated migration. We cannot fully explain the discrepancy between our data and previously published migration assays [14], however, this may be due to different characteristics of the migration chamber used. Importantly, the combination of HDACI and IFNα suppressed all endothelial cell functions, indicating a possible role for this drug combination as a therapy for cancer patients at the point of minimal residual disease.

Conclusion
In summary, we have found that the combination of HDACIs, TSA, SAHA and VPA, with IFNα have significant cytotoxic effects on a wide variety of cancer cells, with little toxicity to normal non-malignant cells. Inhibition of p21 WAF1 expression sensitizes p21 WAF1 -expressing cancer cells to the combination therapy. Furthermore, HDACI and IFNα co-operatively suppress pro-angiogenic gene expression in cancer cells, multiple endothelial cell functions in vitro, and tumour-driven vasculature formation in vivo. Our results provide a basis for further in vivo studies and eventual clinical trials using the combination of HDACIs and IFNα.

Methods
Cell culture and reagentsThe neuroblastoma cell line, BE (2)

Endothelial cell culture
Human umbilical vein endothelial cells (HUVECs) were a gift from Dr K MacKenzie (Children's Cancer Institute Australia, Sydney, Australia). HUVECs were maintained in 0.1% gelatin coated tissue culture flasks or wells with medium 199 (Invitrogen, Carlsbad, CA, USA) supplemented with 20% fetal bovine serum, 5% human serum (Sigma), 10 U/ml heparin (Pharmacia & Upjohn, Peapack, NJ, USA), 5 ng/ml basic fibroblast growth factor (bFGF) (Sigma) and 20 ug/ml endothelial growth factor (Roche, Mannheim, Germany). Only passages 5 and 6 were used in the experiments. Hypoxic conditions were maintained in a chamber filled with 1% oxygen.

Alamar blue cell viability assay
After plating in 96 well plates, cells were allowed to attach for 24 hours, followed by treatment with various drugs for 72 hours. Before the end of treatment, cells were incubated with Alamar blue (Invitrogen) for 5 hours, and plates were then read on a micro-plate reader at 570/595 nm. Relative cell viability was calculated according to the readings and expressed as optical density (OD) absorbance units.

Immunoblot analysis
Twenty four hours after treatment with control, TSA and/ or IFNα, protein was extracted from whole cells, separated by electrophoresis, and transferred onto nitrocellulose membrane. Membranes were incubated with mouse antihuman p21 WAF1 antibodies (Santa Cruz Biotechnologies, Santa Cruz, CA, USA) (1:1000), followed by goat antimouse antibody (1:2000) conjugated with horseradish peroxidase. Chemiluminescent detection was performed using SuperSignal reagents (Pierce). Membranes were then re-probed with an anti-β-actin antibody (Pierce), as a loading control.
siRNA transfectionMCF7 cells were transfected with a validated scrambled siRNA or siRNA specifically targeting p21 WAF1 (SmartPool siRNA CDKN1A, Dharmacon Research, Lafayete, CO) with Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer's recommendation. Cells were lysed, and RNA or protein extracted 24 hours later for Reverse Transcriptionpolymerase Chain Reaction (RT-PCR) or immunoblot analysis of siRNA transfection efficacy.