The mRNA-binding protein HuR promotes hypoxia-induced chemoresistance through posttranscriptional regulation of the proto-oncogene PIM1 in pancreatic cancer cells.

Previously, it has been shown that pancreatic ductal adenocarcinoma (PDA) tumors exhibit high levels of hypoxia, characterized by low oxygen pressure (pO2) and decreased O2 intracellular perfusion. Chronic hypoxia is strongly associated with resistance to cytotoxic chemotherapy and chemoradiation in an understudied phenomenon known as hypoxia-induced chemoresistance. The hypoxia-inducible, pro-oncogenic, serine-threonine kinase PIM1 (Proviral Integration site for Moloney murine leukemia virus 1) has emerged as a key regulator of hypoxia-induced chemoresistance in PDA and other cancers. Although its role in therapeutic resistance has been described previously, the molecular mechanism behind PIM1 overexpression in PDA is unknown. Here, we demonstrate that cis-acting AU-rich elements (ARE) present within a 38-base pair region of the PIM1 mRNA 3'-untranslated region mediate a regulatory interaction with the mRNA stability factor HuR (Hu antigen R) in the context of tumor hypoxia. Predominantly expressed in the nucleus in PDA cells, HuR translocates to the cytoplasm in response to hypoxic stress and stabilizes the PIM1 mRNA transcript, resulting in PIM1 protein overexpression. A reverse-phase protein array revealed that HuR-mediated regulation of PIM1 protects cells from hypoxic stress through phosphorylation and inactivation of the apoptotic effector BAD and activation of MEK1/2. Importantly, pharmacological inhibition of HuR by MS-444 inhibits HuR homodimerization and its cytoplasmic translocation, abrogates hypoxia-induced PIM1 overexpression and markedly enhances PDA cell sensitivity to oxaliplatin and 5-fluorouracil under physiologic low oxygen conditions. Taken together, these results support the notion that HuR has prosurvival properties in PDA cells by enabling them with growth advantages in stressful tumor microenvironment niches. Accordingly, these studies provide evidence that therapeutic disruption of HuR's regulation of PIM1 may be a key strategy in breaking an elusive chemotherapeutic resistance mechanism acquired by PDA cells that reside in hypoxic PDA microenvironments.

Acquisition of estrogen independence induces TOB1-related mechanisms supporting breast cancer cell proliferation.

Resistance to therapies targeting the estrogen pathway remains a challenge in the treatment of estrogen receptor-positive breast cancer. To address this challenge, a systems biology approach was used. A library of small interfering RNAs targeting an estrogen receptor (ER)- and aromatase-centered network identified 46 genes that are dispensable in estrogen-dependent MCF7 cells, but are selectively required for the survival of estrogen-independent MCF7-derived cells and multiple additional estrogen-independent breast cancer cell lines. Integration of this information identified a tumor suppressor gene TOB1 as a critical determinant of estrogen-independent ER-positive breast cell survival. Depletion of TOB1 selectively promoted G1 phase arrest and sensitivity to AKT and mammalian target of rapmycin (mTOR) inhibitors in estrogen-independent cells but not in estrogen-dependent cells. Phosphoproteomic profiles from reverse-phase protein array analysis supported by mRNA profiling identified a significant signaling network reprogramming by TOB1 that differed in estrogen-sensitive and estrogen-resistant cell lines. These data support a novel function for TOB1 in mediating survival of estrogen-independent breast cancers. These studies also provide evidence for combining TOB1 inhibition and AKT/mTOR inhibition as a therapeutic strategy, with potential translational significance for the management of patients with ER-positive breast cancers.

Application of molecular technologies for phosphoproteomic analysis of clinical samples.

The integration of small kinase inhibitors and monoclonal antibodies into oncological practice has opened a new paradigm for treating cancer patients. As proteins are the direct targets of the new generations of targeted therapeutics, many of which are kinase/enzymatic inhibitors, there is an increasing interest in developing technologies capable of monitoring post-translational changes of the human proteome for the identification of new predictive, prognostic and therapeutic biomarkers. It is well known that the vast majority of the activation/deactivation of these drug targets is driven by phosphorylation. This review provides a description of the main proteomic platforms (planar and bead array, reverse phase protein microarray, phosphoflow, AQUA and mass spectrometry) that have successfully been used for measuring changes in phosphorylation level of drug targets and downstream substrates using clinical specimens. Major emphasis was given to the strengths and weaknesses of the different platforms and to the major barriers that are associated with the analysis of the phosphoproteome. Finally, a number of examples of application of the above-mentioned technologies in the clinical setting are reported.

Pilot phase I/II personalized therapy trial for metastatic colorectal cancer: evaluating the feasibility of protein pathway activation mapping for stratifying patients to therapy with imatinib and panitumumab.

This nonrandomized phase I/II trial assessed the efficacy/tolerability of imatinib plus panitumumab in patients affected by metastatic colorectal cancer (mCRC) after stratification to treatment by selection of activated imatinib drug targets using reverse-phase protein array (RPPA). mCRC patients presenting with a biopsiable liver metastasis were enrolled. Allocation to the experimental and control arms was established using functional pathway activation mapping of c-Kit, PDGFR, and c-Abl phosphorylation by RPPA. The experimental arm received run-in escalation therapy with imatinib followed by panitumumab. The control arm received panitumumab alone. Seven patients were enrolled in the study. For three of the seven patients, sequential pre- and post-treatment biopsies were used to evaluate the effect of the therapeutic compounds on the drug targets and substrates. A decrease in the activation level of the drug targets and downstream substrates was observed in two of three patients. Combination therapy increased the activation of the AKT-mTOR pathway and several receptor tyrosine kinases. This study proposes a novel methodology for stratifying patients to personalized treatment based on the activation level of the drug targets. This workflow provides the ability to monitor changes in the signaling pathways after the administration of targeted therapies and to identify compensatory mechanisms.

Combined PDK1 and CHK1 inhibition is required to kill glioblastoma stem-like cells in vitro and in vivo.

Glioblastoma (GBM) is the most common and deadly adult brain tumor. Despite aggressive surgery, radiation, and chemotherapy, the life expectancy of patients diagnosed with GBM is ∼14 months. The extremely aggressive nature of GBM results from glioblastoma stem-like cells (GSCs) that sustain GBM growth, survive intensive chemotherapy, and give rise to tumor recurrence. There is accumulating evidence revealing that GSC resilience is because of concomitant activation of multiple survival pathways. In order to decode the signal transduction networks responsible for the malignant properties of GSCs, we analyzed a collection of GSC lines using a dual, but complementary, experimental approach, that is, reverse-phase protein microarrays (RPPMs) and kinase inhibitor library screening. We treated GSCs in vitro with clinically relevant concentrations of temozolomide (TMZ) and performed RPPM to detect changes in phosphorylation patterns that could be associated with resistance. In addition, we screened GSCs in vitro with a library of protein and lipid kinase inhibitors to identify specific targets involved in GSC survival and proliferation. We show that GSCs are relatively insensitive to TMZ treatment in terms of pathway activation and, although displaying heterogeneous individual phospho-proteomic profiles, most GSCs are resistant to specific inhibition of the major signaling pathways involved in cell survival and proliferation. However, simultaneous multipathway inhibition by the staurosporin derivative UCN-01 results in remarkable inhibition of GSC growth in vitro. The activity of UCN-01 on GSCs was confirmed in two in vivo models of GBM growth. Finally, we used RPPM to study the molecular and functional effects of UCN-01 and demonstrated that the sensitivity to UCN-01 correlates with activation of survival signals mediated by PDK1 and the DNA damage response initiated by CHK1. Taken together, our results suggest that a combined inhibition of PDK1 and CHK1 represents a potentially effective therapeutic approach to reduce the growth of human GBM.

Feedbacks and adaptive capabilities of the PI3K/Akt/mTOR axis in acute myeloid leukemia revealed by pathway selective inhibition and phosphoproteome analysis.

Acute myeloid leukemia (AML) primary cells express high levels of phosphorylated Akt, a master regulator of cellular functions regarded as a promising drug target. By means of reverse phase protein arrays, we examined the response of 80 samples of primary cells from AML patients to selective inhibitors of the phosphatidylinositol 3 kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) axis. We confirm that >60% of the samples analyzed are characterized by high pathway phosphorylation. Unexpectedly, however, we show here that targeting Akt and mTOR with the specific inhibitors Akti 1/2 and Torin1, alone or in combination, result in paradoxical Akt phosphorylation and activation of downstream signaling in 70% of the samples. Indeed, we demonstrate that cropping Akt or mTOR activity can stabilize the Akt/mTOR downstream effectors Forkhead box O and insulin receptor substrate-1, which in turn potentiate signaling through upregulation of the expression/phosphorylation of selected growth factor receptor tyrosine kinases (RTKs). Activation of RTKs in turn reactivates PI3K and downstream signaling, thus overruling the action of the drugs. We finally demonstrate that dual inhibition of Akt and RTKs displays strong synergistic cytotoxic effects in AML cells and downmodulates Akt signaling to a much greater extent than either drug alone, and should therefore be explored in AML clinical setting.

A candidate anti-HIV reservoir compound, auranofin, exerts a selective 'anti-memory' effect by exploiting the baseline oxidative status of lymphocytes.

Central memory (T(CM)) and transitional memory (T(TM)) CD4(+) T cells are known to be the major cellular reservoirs for HIV, as these cells can harbor a transcriptionally silent form of viral DNA that is not targeted by either the immune system or current antiretroviral drug regimens. In the present study, we explored the molecular bases of the anti-HIV reservoir effects of auranofin (AF), a pro-oxidant gold-based drug and a candidate compound for a cure of AIDS. We here show that T(CM) and T(TM) lymphocytes have lower baseline antioxidant defenses as compared with their naive counterpart. These differences are mirrored by the effects exerted by AF on T-lymphocytes: AF was able to exert a pro-differentiating and pro-apoptotic effect, which was more pronounced in the memory subsets. AF induced an early activation of the p38 mitogen-activated protein kinase (p38 MAPK) followed by mitochondrial depolarization and a final burst in intracellular peroxides. The pro-differentiating effect was characterized by a downregulation of the CD27 marker expression. Interestingly, AF-induced apoptosis was inhibited by pyruvate, a well-known peroxide scavenger, but pyruvate did not inhibit the pro-differentiating effect of AF, indicating that the pro-apoptotic and pro-differentiating effects involve different pathways. In conclusion, our results demonstrate that AF selectively targets the T(CM)/T(TM) lymphocyte subsets, which encompass the HIV reservoir, by affecting redox-sensitive cell death pathways.

Individualized therapy for metastatic colorectal cancer.

Systemic therapeutic efficacy is central to determining the outcome of patients with metastatic colorectal cancer (CRC). In these patients, there is a critical need for predictive biomarkers to optimize efficacy whilst minimizing toxicity. The integration of a new generation of molecularly targeted drugs into the treatment of CRC, coupled with the development of sophisticated technologies for individual tumours as well as patient molecular profiling, underlines the potential for personalized medicine. In this review, we focus on the latest progress made within the genomic and proteomic fields, concerning predictive biomarkers for individualized therapy in metastatic CRC.

Aurora kinases in childhood acute leukemia: the promise of aurora B as therapeutic target.

We investigated the effects of targeting the mitotic regulators aurora kinase A and B in pediatric acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). Aurora protein expression levels in pediatric ALL and AML patient samples were determined by western blot and reverse phase protein array. Both kinases were overexpressed in ALL and AML patients (P<0.0002), especially in E2A-PBX1-translocated ALL cases (P<0.002), compared with normal bone-marrow mononuclear cells. Aurora kinase expression was silenced in leukemic cell lines using short hairpin RNAs and locked nucleic acid-based mRNA antagonists. Aurora B knockdown resulted in proliferation arrest and apoptosis, whereas aurora A knockdown caused no or only minor growth delay. Most tested cell lines were highly sensitive to the AURKB-selective inhibitor barasertib-hydroxyquinazoline-pyrazol-anilide (AZD1152-HQPA) in the nanomolar range, as tested with an MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay. But most importantly, primary ALL cells with a high aurora B protein expression, especially E2A-PBX1-positive cases, were sensitive as well. In adult AML early clinical trials, clear responses are observed with barasertib. Here we show that inhibition of aurora B, more than aurora A, has an antiproliferative and pro-apoptotic effect on acute leukemia cells, indicating that particularly targeting aurora B may offer a new strategy to treat pediatric ALL and AML.

AMPK inhibition enhances apoptosis in MLL-rearranged pediatric B-acute lymphoblastic leukemia cells.

The serine/threonine kinase AMP-activated protein kinase (AMPK) and its downstream effectors, including endothelial nitric oxide synthase and BCL-2, are hyperactivated in B-cell precursor-acute lymphoblastic leukemia (BCP-ALL) cells with MLL gene rearrangements. We investigated the role of activated AMPK in supporting leukemic cell survival and evaluated AMPK as a potential drug target. Exposure of leukemic cells to the commercial AMPK inhibitor compound C resulted in massive apoptosis only in cells with MLL gene rearrangements. These results were confirmed by targeting AMPK with specific short hairpin RNAs. Compound C-induced apoptosis was associated with mitochondrial membrane depolarization, reactive oxygen species production, cytochrome c release and caspases cleavage, indicating intrinsic apoptosis pathway activation. Treatment with low concentrations of compound C resulted in a strong antileukemic activity, together with cytochrome c release and cleavage of caspases and poly(ADP-ribose) polymerase, also in MLL-rearranged primary BCP-ALL samples. Moreover, AMPK inhibition in MLL-rearranged cell lines synergistically enhanced the antiproliferative effects of vincristine, daunorubicin, cytarabine, dexamethasone and L-asparaginase in most of the evaluated conditions. Taken together, these results indicate that the activation of the AMPK pathway directly contributes to the survival of MLL-rearranged BCP-ALL cells and AMPK inhibitors could represent a new therapeutic strategy for this high-risk leukemia.

BCAR4 induces antioestrogen resistance but sensitises breast cancer to lapatinib.

BACKGROUND: High BCAR4 and ERBB2 mRNA levels in primary breast cancer associate with tamoxifen resistance and poor patient outcome. We determined whether BCAR4 expression sensitises breast cancer cells to lapatinib, and identifies a subgroup of patients who possibly may benefit from ERBB2-targeted therapies despite having tumours with low ERBB2 expression. METHODS: Proliferation assays were applied to determine the effect of BCAR4 expression on lapatinib treatment. Changes in cell signalling were quantified with reverse-phase protein microarrays. Quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) of ERBB2 and BCAR4 was performed in 1418 primary breast cancers. Combined BCAR4 and ERBB2 mRNA levels were evaluated for association with progression-free survival (PFS) in 293 oestrogen receptor-α (ER)-positive patients receiving tamoxifen as first-line monotherapy for recurrent disease. RESULTS: BCAR4 expression strongly sensitised ZR-75-1 and MCF7 breast cancer cells to the combination of lapatinib and antioestrogens. Lapatinib interfered with phosphorylation of ERBB2 and its downstream mediators AKT, FAK, SHC, STAT5, and STAT6. Reverse transcriptase-PCR analysis showed that 27.6% of the breast cancers were positive for BCAR4 and 22% expressed also low levels of ERBB2. The clinical significance of combining BCAR4 and ERBB2 mRNA status was underscored by the finding that the group of patients having BCAR4-positive/ERBB2-low-expressing cancers had a shorter PFS on tamoxifen treatment than the BCAR4-negative group. CONCLUSION: This study shows that BCAR4 expression identifies a subgroup of ER-positive breast cancer patients without overexpression of ERBB2 who have a poor outcome and might benefit from combined ERBB2-targeted and antioestrogen therapy.

Chemoresistant colorectal cancer cells and cancer stem cells mediate growth and survival of bystander cells.

BACKGROUND: Recent studies suggest that cancer stem cells (CSCs) mediate chemoresistance, but interestingly, only a small percentage of cells in a resistant tumour are CSCs; this suggests that non-CSCs survive by other means. We hypothesised that chemoresistant colorectal cancer (CRC) cells generate soluble factors that enhance survival of chemonaive tumour cells. METHODS: Chemoresistant CRC cells were generated by serial passage in oxaliplatin (Ox cells). Conditioned media (CM) was collected from parental and oxaliplatin-resistant (OxR) cells. CRC cells were treated with CM and growth and survival were assessed. Tumour growth rates were determined in nude mice after cells were treated with CM. Mass spectrometry (MS) identified proteins in CM. Reverse phase protein microarray assays determined signalling effects of CM in parental cells. RESULTS: Oxaliplatin-resistant CM increased survival of chemo-naive cells. CSC CM also increased growth of parental cells. Parental and OxR mixed tumours grew larger than tumours composed of parental or OxR cells alone. Mass spectrometry detected unique survival-promoting factors in OxR CM compared with parental CM. Cells treated with OxR CM demonstrated early phosphorylation of EGFR and MEK1, with later upregulation of total Akt .We identified progranulin as a potential mediator of chemoresistance. CONCLUSION: Chemoresistant tumour cells and CSCs may promote resistance through soluble factors that mediate survival in otherwise chemosensitive tumour cells.

Homologous control of protein signaling networks.

In a previous paper we introduced a method called augmented sparse reconstruction (ASR) that identifies links among nodes of ordinary differential equation networks, given a small set of observed trajectories with various initial conditions. The main purpose of that technique was to reconstruct intracellular protein signaling networks. In this paper we show that a recursive augmented sparse reconstruction generates artificial networks that are homologous to a large, reference network, in the sense that kinase inhibition of several reactions in the network alters the trajectories of a sizable number of proteins in comparable ways for reference and reconstructed networks. We show this result using a large in-silico model of the epidermal growth factor receptor (EGF-R) driven signaling cascade to generate the data used in the reconstruction algorithm. The most significant consequence of this observed homology is that a nearly optimal combinatorial dosage of kinase inhibitors can be inferred, for many nodes, from the reconstructed network, a result potentially useful for a variety of applications in personalized medicine.

Platform for establishing interlaboratory reproducibility of selected reaction monitoring-based mass spectrometry peptide assays.

Mass spectrometry (MS) is an attractive alternative to quantification of proteins by immunoassays, particularly for protein biomarkers of clinical relevance. Reliable quantification requires that the MS-based assays are robust, selective, and reproducible. Thus, the development of standardized protocols is essential to introduce MS into clinical research laboratories. The aim of this study was to establish a complete workflow for assessing the transferability and reproducibility of selected reaction monitoring (SRM) assays between clinical research laboratories. Four independent laboratories in North America, using identical triple-quadrupole mass spectrometers (Quantum Ultra, Thermo), were provided with standard protocols and instrumentation settings to analyze unknown samples and internal standards in a digested plasma matrix to quantify 51 peptides from 39 human proteins using a multiplexed SRM assay. The interlaboratory coefficient of variation (CV) was less than 10% for 25 of 39 peptides quantified (12 peptides were not quantified based upon hydrophobicity) and exhibited CVs less than 20% for the remaining peptides. In this report, we demonstrate that previously developed research platforms for SRM assays can be improved and optimized for deployment in clinical research environments.

NOTCH1 and/or FBXW7 mutations predict for initial good prednisone response but not for improved outcome in pediatric T-cell acute lymphoblastic leukemia patients treated on DCOG or COALL protocols.

Aberrant activation of the NOTCH1 pathway by inactivating and activating mutations in NOTCH1 or FBXW7 is a frequent phenomenon in T-cell acute lymphoblastic leukemia (T-ALL). We retrospectively investigated the relevance of NOTCH1/FBXW7 mutations for pediatric T-ALL patients enrolled on Dutch Childhood Oncology Group (DCOG) ALL7/8 or ALL9 or the German Co-Operative Study Group for Childhood Acute Lymphoblastic Leukemia study (COALL-97) protocols. NOTCH1-activating mutations were identified in 63% of patients. NOTCH1 mutations affected the heterodimerization, the juxtamembrane and/or the PEST domains, but not the RBP-J-κ-associated module, the ankyrin repeats or the transactivation domain. Reverse-phase protein microarray data confirmed that NOTCH1 and FBXW7 mutations resulted in increased intracellular NOTCH1 levels in primary T-ALL biopsies. Based on microarray expression analysis, NOTCH1/FBXW7 mutations were associated with activation of NOTCH1 direct target genes including HES1, DTX1, NOTCH3, PTCRA but not cMYC. NOTCH1/FBXW7 mutations were associated with TLX3 rearrangements, but were less frequently identified in TAL1- or LMO2-rearranged cases. NOTCH1-activating mutations were less frequently associated with mature T-cell developmental stage. Mutations were associated with a good initial in vivo prednisone response, but were not associated with a superior outcome in the DCOG and COALL cohorts. Comparing our data with other studies, we conclude that the prognostic significance for NOTCH1/FBXW7 mutations is not consistent and may depend on the treatment protocol given.

Nanoparticle technology: addressing the fundamental roadblocks to protein biomarker discovery.

Clinically relevant biomarkers exist in blood and body fluids in extremely low concentrations, are masked by high abundance high molecular weight proteins, and often undergo degradation during collection and transport due to endogenous and exogenous proteinases. Nanoparticles composed of a N-isopropylacrylamide hydrogel core shell functionalized with internal affinity baits are a new technology that can address all of these critical analytical challenges for disease biomarker discovery and measurement. Core-shell, bait containing, nanoparticles can perform four functions in one step, in solution, in complex biologic fluids (e.g. blood or urine): a) molecular size sieving, b) complete exclusion of high abundance unwanted proteins, c) target analyte affinity sequestration, and d) complete protection of captured analytes from degradation. Targeted classes of protein analytes sequestered by the particles can be concentrated in small volumes to effectively amplify (up to 100 fold or greater depending on the starting sample volume) the sensitivity of mass spectrometry, western blotting, and immunoassays. The materials utilized for the manufacture of the particles are economical, stable overtime, and remain fully soluble in body fluids to achieve virtually 100 percent capture of all solution phase target proteins within a few minutes.

Augmented sparse reconstruction of protein signaling networks.

The problem of reconstructing and identifying intracellular protein signaling and biochemical networks is of critical importance in biology. We propose a mathematical approach called augmented sparse reconstruction for the identification of links among nodes of ordinary differential equation (ODE) networks, given a small set of observed trajectories with various initial conditions. As a test case, the method is applied to the epidermal growth factor receptor (EGFR) driven signaling cascade, a well-studied and clinically important signaling network. Our method builds a system of representation from a collection of trajectory integrals, selectively attenuating blocks of terms in the representation. The system of representation is then augmented with random vectors, and l(1) minimization is used to find sparse representations for the dynamical interactions of each node. After showing the performance of our method on a model of the EGFR protein network, we sketch briefly the potential future therapeutic applications of this approach.

Critical dependence of blood-borne biomarker concentrations on the half-lives of their carrier proteins.

Until recently, the low-abundance (LA) range of the serum proteome was an unexplored reservoir of diagnostic information. Today it is increasingly appreciated that a diagnostic goldmine of LA biomarkers resides in the blood stream in complexed association with more abundant higher molecular weight carrier proteins such as albumin and immunoglobulins. As we now look to the possibility of harvesting these LA biomarkers more efficiently through engineered nano-scale particles, mathematical approaches are needed in order to reveal the mechanisms by which blood carrier proteins act as molecular 'mops' for LA diagnostic cargo, and the functional relationships between bound LA biomarker concentrations and other variables of interest such as biomarker intravasation and clearance rates and protein half-lives in the bloodstream. Here we show, by simple mathematical modeling, how the relative abundance of large carrier proteins and their longer half-lives in the bloodstream work together to amplify the total blood concentration of these tiny biomarkers. The analysis further suggests that alterations in the production of biomarkers lead to gradual rather than immediate changes in biomarker levels in the blood circulation. The model analysis also points to the characteristics of artificial nano-particles that would render them more efficient harvesters of tumor biomarkers in the circulation, opening up possibilities for the early detection of curable disease, rather than simply better detection of advanced disease.

Signal pathway profiling of epithelial and stromal compartments of colonic carcinoma reveals epithelial-mesenchymal transition.

Molecular crosstalk, including reciprocal stimulation, is theorized to take place between epithelial cancer cells and surrounding non-neoplastic stromal cells. This is the rationale for stromal therapy, which could eliminate support of a cancer by its genetically stable stroma. Epithelial-stromal crosstalk is so far poorly documented in vivo, and cell cultures and animal experiments may not provide accurate models. The current study details stromal-epithelial signalling pathways in 35 human colon cancers, and compares them with matched normal tissues using quantitative proteomic microarrays. Lysates prepared from separately microdissected epithelium and stroma were analysed using antibodies against 61 cell signalling proteins, most of which recognize activated phospho-isoforms. Analyses using unsupervised and supervised statistical methods suggest that cell signalling pathway profiles in stroma and epithelium appear more similar to each other in tumours than in normal colon. This supports the concept that coordinated crosstalk occurs between epithelium and stroma in cancer and suggests epithelial-mesenchymal transition. Furthermore, the data herein suggest that it is driven by cell proliferation pathways and that, specifically, several key molecules within the mitogen-activated protein kinase pathway may play an important role. Given recent findings of epithelial-mesenchymal transition in therapy-resistant tumour epithelium, these findings could have therapeutic implications for colon cancer.