Molecular profiling of signalling proteins for effects induced by the anti-cancer compound GSAO with 400 antibodies
- 4.2k Downloads
GSAO (4-[N-[S-glutathionylacetyl]amino] phenylarsenoxide) is a hydrophilic derivative of the protein tyrosine phosphatase inhibitor phenylarsine oxide (PAO). It inhibits angiogenesis and tumour growth in mouse models and may be evaluated in a phase I clinical trial in the near future. Initial experiments have implicated GSAO in perturbing mitochondrial function. Other molecular effects of GSAO in human cells, for example on the phosphorylation of proteins, are still largely unknown.
Peripheral white blood cells (PWBC) from healthy volunteers were isolated and used to profile effects of GSAO vs. a control compound, GSCA. Changes in site-specific phosphorylations, other protein modifications and expression levels of many signalling proteins were analysed using more than 400 different antibodies in Western blots.
PWBC were initially cultured in low serum conditions, with the aim to reduce basal protein phosphorylation and to increase detection sensitivity. Under these conditions pleiotropic intracellular signalling protein changes were induced by GSAO. Subsequently, PWBC were cultured in 100% donor serum to reflect more closely in vivo conditions. This eliminated detectable GSAO effects on most, but not all signalling proteins analysed. Activation of the MAP kinase Erk2 was still observed and the paxillin homologue Hic-5 still displayed a major shift in protein mobility upon GSAO-treatment. A GSAO induced change in Hic-5 mobility was also found in endothelial cells, which are thought to be the primary target of GSAO in vivo.
Serum conditions greatly influence the molecular activity profile of GSAO in vitro. Low serum culture, which is typically used in experiments analysing protein phosphorylation, is not suitable to study GSAO activity in cells. The signalling proteins affected by GSAO under high serum conditions are candidate surrogate markers for GSAO bioactivity in vivo and can be analysed in future clinical trials. GSAO effects on Hic-5 in endothelial cells may point to a new intracellular GSAO target.
KeywordsHuman Umbilical Vein Endothelial Cell Signalling Protein Adenine Nucleotide Translocator Peripheral White Blood Cell Serum Culture
The term 'cancer' encompasses a wide variety of distinct, multigenic diseases. Even within a specific tumour type, a remarkable degree of heterogeneity on the level of DNA lesions and affected signalling pathways is apparent. Many cancer relevant signalling molecules, but also many molecular targets of anti-cancer drugs, therefore remain unknown. Prominent examples of signalling protein classes long known to be involved in generating cancer pathologies include GTPases, protein kinases and transcription factors. By contrast, protein phosphatases have only recently entered the stage as known players in cancer development. At least 30 protein phosphatases are now implicated in cancer development and other human diseases [1, 2, 3]. In some of these cases, mutational inactivation of a protein phosphatase appears to mimic the constitutive activation of its target kinase(s) . In other cases, hyperactivation or deregulation of a phosphatase may contribute to kinase activation. For example, overexpression of the Cdc25 family phosphatases, which control cell cycle progression, is well documented for a variety of cancers, making the Cdc25 proteins interesting potential targets for anti-cancer therapies [4, 5, 6, 7] and references therein).
The modulation of specific cellular signalling pathways to treat human cancers has only recently developed into an area of intense clinical research activity. A large number of clinical trials for novel signal transduction modulator (STM) drugs are currently planned or under way. STM drugs often have relatively low toxicity, so determination of the maximum tolerated dose (MTD) may not be a prime goal for phase I clinical trials. Instead, identification of an optimal biologically active dose (OBD) is essential . Rapid determination of the OBD requires that in vivo markers of drug activity are available before or very early during the clinical trial.
GSAO has anti-angiogenic activity in vitro and in vivo . Mitochondria and in particular the adenine nucleotide translocator (ANT) have been described as one target of GSAO. However, mitochondria are present in virtually all living cells. Therefore, inhibition of ANT does not per se explain the low toxicity and anti-angiogenic activity of GSAO.
Recent studies have revealed that differences in multidrug resistance proteins (MRP1 and 2) and cellular glutathione levels  appear to contribute to the preferential in vivo activity of GSAO in endothelial cells. Nevertheless, it is well known that practically all drugs in clinical use have multiple molecular in vivo targets. They depend on the affected cell type and applied doses. Some of these targets are intimately linked to the adverse effects of the drug. In cancer cells, with their heterogeneous genetic lesions, drug targets can even be present in different combinations or at different expression and activity levels when comparing individual patients. Identification and monitoring of novel molecular targets could thus help to understand why a drug is more effective or toxic in some patients compared to others. PWBC from donated blood were chosen for the analyses, since they are easily obtainable from healthy volunteers prior to clinical trial and later from the patients enrolled in the trials. A main focus of the experiments was to detect changes in protein phosphorylation induced by GSAO treatment, because GSAO is a hydrophilic derivative of phenylarsine oxide (PAO), a well known protein phosphatase inhibitor [12, 13]. In contrast to GSAO, PAO is unsuitable for in vivo applications and toxic for cells at nanomolar concentrations . It penetrates very rapidly into cell membranes and is hence not equally distributed throughout the vascular system of an organism. As GSAO is expected to enter a phase I clinical trial in the near future a molecular profiling of its effects on signalling proteins prior to the trial was performed. The aim was to provide data for trial-supporting pharmacodynamic molecular assays, trial design and possibly identification of additional mechanisms of GSAO action.
Isolation and culture of PWBC from healthy human volunteers
PWBC were obtained from 14 healthy volunteers not taking any medication (8 male, 6 female; aged between 23 and 60 years) with informed consent. Ethical approval was granted by the Oxfordshire Ethics Committee (study number 05/Q1605/59). Cell preparations from more than 40 different blood donations were used. All blood donations took place between 9 and 11 a.m. to minimise potential circadian changes of signalling proteins. Whole blood (54 ml) was collected from a peripheral arm vein into 6 ml of 46.7% (v/v) trisodium citrate (final concentration 4.67%) and syringes immediately placed on a nutator at room temperature for at least 5 min to ensure complete mixing. Batches of 20 ml citrated blood were then carefully layered onto 20 ml Polymorphprep™ density gradient media (Axis Shield PoC AS, Oslo, Norway) and centrifuged at 500 × g for 30 min at 20°C to remove erythrocytes. PWBC were collected from the medium/serum interface, transferred into a 50 ml centrifuge tube and gently resuspended in sterile PBS. PWBC were then centrifuged for 20 min at 600 × g at 20°C, PBS was aspirated and the PWBC pellet stored in ice. PWBC were subsequently cultured in either 100% donor serum or RPMI1640 supplemented with 0.5% foetal bovine serum (FBS), 100 μg/ml penicillin-streptomycin and 2 mM L-glutamate. Donor serum was simultaneously prepared freshly from peripheral blood. For this, blood was collected without anticoagulant and centrifuged for 30 min at 3700 × g at 20°C. The supernatant was transferred into a fresh centrifuge tube and incubated for 2 h at room temperature with nutation to allow coagulation, before centrifugation at 3700 × g for 30 min at 20°C to obtain serum. Pelleted PWBC were resuspended in either growth medium or donor serum and equilibrated for 2 h at 37°C in a humidified cell culture incubator with 5% CO2, before treatment with the compound of interest. Cells were treated with either GSAO, the active drug, or 4-N-(S-glutathionyl acetyl) amino) benzoic acid (GSCA) in which the trivalent arsenic of GSAO has been replaced by a carboxylic acid, as a control, for the concentrations and durations indicated. Stock solutions of both GSCA and GSAO were made at a concentration of 3 mM in RPMI 1640 and, since GSAO is prone to oxidation, stored in aliquots at -80°C unless used immediately. PAO was freshly dissolved in cell culture grade DMSO at a concentration of 10 mM and diluted to 1 μM with medium.
Preparation and analyses of protein lysates
Treated or mock-incubated PWBC were scraped from tissue culture flasks, transferred to a 50 ml centrifuge tube filled with PBS and centrifuged at 3700 × g for 20 min at 20°C to pellet the cells. The PBS was completely aspirated from the PWBC pellet and the pellet lysed directly in boiling SDS-PAGE sample buffer (70 mM TrisHCl pH6.8, 5% (v/v) β-mercaptoethanol, 40% (v/v) glycerol, 3%(w/v) SDS, 0.05% (w/v) bromophenol blue). The samples were then boiled for a further 5 min, cooled to room temperature and sonicated for 30 seconds to fragment DNA contained in the cell lysate. Proteins in the lysate were resolved by SDS polyacrylamide gel electrophoresis (SDS-PAGE) and electrotransferred to PVDF membranes (Hybond P, Amersham Biosciences). The membranes were blocked and probed overnight at 4°C with different phospho-epitope specific or other antibodies according to the antibody manufacturers instructions. Bound primary antibodies were detected using horseradish peroxidase-conjugated anti-mouse or anti-rabbit immunoglobulins (Jackson ImmunoResearch Inc.) and ECL (Amersham Biosciences). Equal loading of different protein extract samples was routinely confirmed by Coomassie-blue staining of an SDS-PAGE-separated aliquot prior to blot analyses. Antibodies used in this study are compiled in the additional files 8 and 9. The BD PowerBlot Western Array Screen Service (BD Biosciences) was utilised to analyse expression levels of 240 individually selected signalling proteins using monoclonal antibodies.
Culture and treatment of HUVEC
Human umbilical vein endothelial cells (HUVEC) were purchased from TCS Cell Works, (Botolph Claydon, UK) and cultured according to the manufacturers conditions with recommended growth media, cell attachment factor and passage conditions. Cells were cultured in 10 cm plates until 80% confluent before treatment with GSAO or GSCA in growth media with high serum (15% FBS) for 24 hours at 37°C with 5% CO2. Cells were lysed with boiling sample buffer. Protein analysis, SDS PAGE and Western blots were carried out as described for the PWBC.
Pleiotropic effects of GSAO on PWBC cultured in low serum conditions
In initial experiments it was determined how rapidly protein phosphorylation changes were detectable upon GSAO treatment of freshly isolated PWBC from healthy donors under low serum culture conditions. For this, PWBC were incubated with GSAO or its more hydrophobic relative PAO. As previously reported in the literature for various cell types, increased tyrosine phosphorylation was rapidly observed with PAO. Incubation of PWBC with 1 μM PAO for 1 h resulted in a hyper-phosphorylation of protein-tyrosyl residues that was comparable to an incubation of PWBC with 50 μM GSAO for 8 h (additional file 1). This was not unexpected, since GSAO is specifically modified to greatly slow down its uptake into cells. Incubation of PWBC with a single dose of GSAO for 24 h instead of 8 h greatly enhanced the hyperphosphorylation of proteins on tyrosine residues (figure 1C). Treatment with 50 μM of the control compound GSCA (see figure 1B for structure), which lacks the crucial arsenic moiety, did not detectably affect basal PWBC protein-tyrosyl phosphorylation (figure 1C). Subsequently, titration experiments showed that incubation of PWBC for 24 h with 1.5 μM GSAO readily induced hyper-phosphorylation (figure 1D). Similar results were obtained for threonine phosphorylation of proteins upon GSAO treatment of PWBC (additional file 2). Nine different commercial anti-pSer antibodies were also tested, but all were found not to be suitable for our purposes (listed in additional file 9; data not shown).
In a subsequent set of experiments, the effects of GSAO vs. GSCA on specific signalling proteins in PWBC were investigated. The results are described and shown in additional text file and additional files 1 to 11.
From this large data set we conclude that GSAO has rather pleiotropic effects on the cell signalling protein network of PWBC cultured in low serum conditions. Since it is very unlikely that a drug which affects numerous key signalling proteins would show the low toxicity profile seen in vivo , we modified the PWBC culture system.
GSAO affects only a small subset of signalling proteins when PWBC are cultured in high serum
In summary, the screen with a large number of epitope-specific antibodies has resulted in the detection of three candidate surrogate markers for GSAO activity in PWBC under high serum conditions. To search for further potential markers, antibodies targeted against phosphorylated substrate epitopes or docking motifs of a specific kinase or kinase family were used on GSAO-, GSCA- or mock-treated PWBC lysates.
Detection of further proteins affected by GSAO in PWBC high serum culture using kinase substrate-specific antibodies
GSAO affects signalling proteins in HUVEC endothelial cells
STM drugs are playing an increasingly important role in cancer therapy. Many of these drugs appear to have low toxicity profiles, so biological responses are expected below drug concentrations, which lead to massive side effects. Based on the results from previous animal studies [10, 18], one can speculate that this may also be the case for novel anti-cancer compound GSAO. Surrogate markers that can be monitored to detect the biological activity of a STM drug in patients shortly after drug application may therefore be of great value for its optimal clinical development.
GSAO is thought to primarily interfere with proliferating endothelial cells , hence disrupting the further growth of tumour vasculature. GSAO activity is therefore dependent on systemic distribution through the blood vessels and capillaries but probably does not need to penetrate deeply into tissues to exert key effects. As endothelial cells and PWBC are both key components of the blood vessel system and will be exposed to circulating GSAO to a similar degree, PWBC should be well suited for the analysis of surrogate markers for GSAO activity in vivo.
Our experiments clearly show that classical low serum culture, often used to reduce background signals prior to addition of a drug or factor, is not ideal to find potential markers for GSAO bioactivity in vivo. Furthermore, this study was only possible through the multitude of commercial signalling protein antibodies that have become available in the last few years. These antibodies, directed against specific phosphorylated or nonphosphorylated protein epitopes in signalling proteins, allow the rapid analysis of a significant part of the signalling phosphoproteome. In order to obtain reliable data with these antibodies, their specificity must be carefully evaluated. In our western blot experiments, we frequently found multiple bands resulting from a single antibody analysed. Signals appeared often at unexpected sizes, especially for polyclonal anti sera. Until now, high quality monoclonal antibodies are only available for a small number of phosphoepitopes. We feel that this would make, at present, results from high throughput (HTP) array-format assays or FACS-based HTP analyses like multiplex-bead assays difficult to properly evaluate.
Monitoring of drug effects on total protein expression levels, is simpler, but appears to be overall less informative as many phospho-changes are not mirrored by changes in overall protein expression. Nevertheless, it can give some initial leads and many good quality monoclonal antibodies are readily available. For optimal retrieval of information from western blot type assays it is necessary to perform non-automated visual re-inspections to detect all changes. The GSAO-induced Hic-5 mobility shift was found through this approach and would have been missed using a spot array or another technique not separating proteins by apparent molecular weight.
Hic-5 is an interesting candidate for further analyses of GSAO actions in cells. It is a member of the paxillin family [19, 20] and has been implicated in a variety of biological processes including those that regulate cell motility, invasion, survival and proliferation (summarised in ). Moreover, recent publications have highlighted additional roles for Hic-5, for example, the suppression of Lef/Tcf-driven transcription during vertebrate development . Hic-5 was also reported to undergo redox-sensitive nuclear-cytoplasmic shuttling [22, 23] and to interact with Traf4, which is implicated in the oxidative regulation of endothelial cell migration . GSAO uptake and action has also been shown to depend on redox parameters like cellular glutathione levels , but it remains to be studied if GSAO directly influences theredox regulation of cells or cellular subcompartments, for example by interfering with redox-regulated phosphatases like PTP-PEST  or other signalling proteins.
The consistent activation of the MAP kinase Erk2 by GSAO under high and low serum conditions is also intriguing, since no activation of its regulator kinases MEK1/2 could be observed (data not shown). Apart from a direct effect on Erk-inactivating phosphatases, like MKPs, other more indirect changes in PWBC signalling may lead to Erk2 activation. Since the intracellular GSAO uptake is very slow and effects are usually seen after many hours, a plethora of possible indirect mechanisms, including changes in the transcription of certain genes may underlie the observed Erk2 activity change. Erk2 activation is thought to trigger its dimerisation and nuclear translocation , possibly leading to the phosphorylation of nuclear substrate proteins. Analyses of selected Erk substrates like p90Rsk, and Elk-1 with phospho-epitope specific antibodies (results summarised in additional file 9) did not pinpoint an affected Erk2 substrate so far and an initial screen for substrates of proline-directed kinases (like Erks and Cdks) with an antibody which recognises an xxx-pThr-Pro-xxx motif did also not detect candidate substrates in GSAO-treated PWBC culture in high serum (not shown). Since more than 70 Erk substrates have been reported so far, a large number of potential substrates remain to be analysed. Beyond this, experiments with substrate epitope-directed phospho-antibodies specific for targets of other kinases clearly indicate that additional GSAO-affected proteins, which could be useful bioactivity markers, remain to be discovered.
Changes in one of the three candidate surrogate markers for GSAO activity found in PWBC is detectable in endothelial cells, but additional changes in tyrosine phosphorylation were observed in a multitude of proteins at 15 μM GSAO treatment in HUVEC. This suggests that overlapping, but distinct groups of signalling pathways are affected by GSAO in the different cell types. Further studies into the biological mechanisms of GSAO interference with endothelial cell proliferation are needed to elucidate the in vivo activity of GSAO in this cell type.
This study has shown that GSAO effects on the PWBC phosphoprotein signalling network are quite selective, provided the appropriate experimental conditions are used. The results also give suggestions for the development of new reagents, which may be useful to support future clinical studies and to possibly gain further insight into the cellular targets of GSAO. Analyses with substrate epitope-specific antibodies show that further GSAO-affected proteins remain to be identified. Variation in phosphorylation patterns between PWBC and HUVEC suggest distinct pathway patterns become activated by GSAO in different cell types. As PWBC are easily obtainable they are more probably suitable for measuring surrogate markers in future clinical studies, but endothelial cells are clearly also an important cell type that needs to be analysed further with molecular and functional assays.
We are grateful to the volunteers for donating blood and to Cancer Research UK for funding our work.
- 3.Wang Z, Shen D, Parsons DW, Bardelli A, Sager J, Szabo S, Ptak J, Silliman N, Peters BA, van der Heijden MS, Parmigiani G, Yan H, Wang TL, Riggins G, Powell SM, Willson JK, Markowitz S, Kinzler KW, Vogelstein B, Velculescu VE: Mutational analysis of the tyrosine phosphatome in colorectal cancers. Science. 2004, 304 (5674): 1164-1166. 10.1126/science.1096096.CrossRefPubMedGoogle Scholar
- 10.Don AS, Kisker O, Dilda P, Donoghue N, Zhao X, Decollogne S, Creighton B, Flynn E, Folkman J, Hogg PJ: A peptide trivalent arsenical inhibits tumor angiogenesis by perturbing mitochondrial function in angiogenic endothelial cells. Cancer Cell. 2003, 3 (5): 497-509. 10.1016/S1535-6108(03)00109-0.CrossRefPubMedGoogle Scholar
- 18.Dilda PJ, Decollogne S, Rossiter-Thornton M, Hogg PJ: Para to ortho repositioning of the arsenical moiety of the angiogenesis inhibitor 4-(N-(S-glutathionylacetyl)amino)phenylarsenoxide results in a markedly increased cellular accumulation and antiproliferative activity. Cancer Res. 2005, 65 (24): 11729-11734. 10.1158/0008-5472.CAN-05-2797.CrossRefPubMedGoogle Scholar
- 21.Ghogomu SM, Vanvenrooy S, Ritthaler M, Wedlich D, Gradl D: Hic-5, a novel repressor of Lef/Tcf driven transcription. J Biol Chem. 2005Google Scholar
- 22.Shibanuma M, Kim-Kaneyama JR, Ishino K, Sakamoto N, Hishiki T, Yamaguchi K, Mori K, Mashimo J, Nose K: Hic-5 communicates between focal adhesions and the nucleus through oxidant-sensitive nuclear export signal. Mol Biol Cell. 2003, 14 (3): 1158-1171. 10.1091/mbc.02-06-0099.CrossRefPubMedPubMedCentralGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2407/6/155/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.