Characterization of heterotypic interaction effects in vitro to deconvolute global gene expression profiles in cancer
© Buess et al.; licensee BioMed Central Ltd. 2007
Received: 26 March 2007
Accepted: 14 September 2007
Published: 14 September 2007
Perturbations in cell-cell interactions are a key feature of cancer. However, little is known about the systematic effects of cell-cell interaction on global gene expression in cancer.
We used an ex vivo model to simulate tumor-stroma interaction by systematically co-cultivating breast cancer cells with stromal fibroblasts and determined associated gene expression changes with cDNA microarrays. In the complex picture of epithelial-mesenchymal interaction effects, a prominent characteristic was an induction of interferon-response genes (IRGs) in a subset of cancer cells. In close proximity to these cancer cells, the fibroblasts secreted type I interferons, which, in turn, induced expression of the IRGs in the tumor cells. Paralleling this model, immunohistochemical analysis of human breast cancer tissues showed that STAT1, the key transcriptional activator of the IRGs, and itself an IRG, was expressed in a subset of the cancers, with a striking pattern of elevated expression in the cancer cells in close proximity to the stroma. In vivo, expression of the IRGs was remarkably coherent, providing a basis for segregation of 295 early-stage breast cancers into two groups. Tumors with high compared to low expression levels of IRGs were associated with significantly shorter overall survival; 59% versus 80% at 10 years (log-rank p = 0.001).
In an effort to deconvolute global gene expression profiles of breast cancer by systematic characterization of heterotypic interaction effects in vitro, we found that an interaction between some breast cancer cells and stromal fibroblasts can induce an interferon-response, and that this response may be associated with a greater propensity for tumor progression.
Communication between different cell types is fundamental for the development and homeostasis of multi-cellular organisms. Cells of different origin communicate in a network of interactions via proteins, peptides, small molecular signals, the extracellular matrix and direct cell-cell contact. These heterotypic interactions provide information that is necessary for the regulation of the gene expression programs in normal development , differentiation , topologic organization  and homeostasis  of complex tissue structures. Given the important physiological role of intercellular communication to maintain the delicate dynamic equilibrium of a normal tissue, it is not surprising that aberrant cell-cell interaction signals have been implicated in cancer development and progression [5–10]. Although the characteristics and roots of the heterotypic interaction effects are fundamental aspects of normal physiology and disease, they have not been systematically explored.
In cancer biology, there is increasing evidence for the importance of the interaction between the malignant epithelial cells and the surrounding stromal cells . Tumors are not merely aggregates of malignant cells but are in many respects organ-like structures, which include host stromal cells, such as fibroblasts, endothelial cells and so on, with which the malignant cells themselves intermingle and interact. Inductive interactions between these different cell lineages can play not only a morphogenetic role but also an important mechanistic role in the pathogenesis and progression of malignancy. Co-inoculation of stromal cells with pre-malignant or malignant epithelial cells can increase tumorigenicity and the capacity to metastasize for a variety of tumor types [11, 12], including breast cancer . On the molecular level, results from the knockout of single genes have demonstrated the importance of specific signaling pathways in the tumor-stroma interaction. For example, conditional inactivation of the transforming growth factor (TGF)-β receptor type II in stromal cells led to development of epithelial cancer of the prostate and forestomach in mice . In the mammary gland, site-specific knockout of TGF-β receptor type II in stromal fibroblasts led to defective mammary ductal development and increased carcinoma growth and metastasis . Experiments exploring the interaction of tumor with stromal cells in vitro have revealed changes in expression of several genes involved in cancer [16–18]. These effects reveal the significance of one specific signaling mechanism, but a more complete overview of the molecular systems that mediate these cell-cell interaction effects remains to be revealed.
Biopsy samples of human carcinoma frequently contain both malignant cells and stromal cells. Since gene expression profiles of human cancer are generally derived from these mixed cell populations of grossly dissected tissues, the effects of heterotypic interactions among the cells in the tumor tissue are expected to leave their traces in the global gene expression profiles. Datasets representing expression profiles of thousands of genes in collections of benign and malignant tissues from hundreds of patients have steadily grown in recent years and might be a rich latent source of insights into heterotypic interaction effects on global gene expression. The superposition of the cell specific profiles, however, results in complex gene expression patterns that are difficult to interpret. In breast cancer, Allinen et al.  attempted to resolve this complexity by fractionating the tissue using cell-surface markers to separate different cell types. This led to the identification of cell type specific gene expression profiles. As a result of this analysis they suggested that a myofibroblast expression of CXCL14 and CXCL12, which can bind to the respective receptor CXCR4 on the epithelial cells, is a specific tumor promoting mechanism leading to enhanced proliferation, invasion and metastasis. In a different approach to search for the relevance of stromal signals in cancer data, West et al.  identified stromal-cell specific gene expression signatures in breast cancer using gene expression data from fibroblastic tumors as in vivo models of homogenous populations of malignant mesenchymal cells. Based on stromal-cell specific signatures they were able to segregate breast cancer samples into two subgroups with distinct clinical outcome.
A further layer of complexity, in addition to the simple additive effects of the involved cells, might arise from the effects on gene expression profiles induced by heterotypic cell-cell interactions. The deconvolution of these intercellular signaling effects poses an even greater challenge, since they result in supra-additive non-linear behaviors, which are hard to disentangle and distinguish from the cell-intrinsic regulatory processes. These cell interaction effects might account for a significant proportion of the unrevealed information in the gene expression data from tissue specimens. Given the evidence that interactions between cells can play critical roles in tumor progression, such data might be even more meaningful than prominent expression patterns that are driven by the proportional representation of a given cell type in a tissue .
The primary aim of this work was to survey and characterize the effects of cell-cell interaction in an attempt to disentangle the complex network of intercellular signaling in a multi-cellular tissue and specifically in breast cancer. To extract the information about tumor-stroma interaction from global gene expression profiles of cancer tissue, we applied an approach based on in vitro modeling combined with subsequent testing of the in vitro findings in published cancer datasets. Observation of fundamental biological processes in vitro, such as the cell cycle  and the reaction of fibroblasts to serum [23, 24], or observation of the common response of different cell types to hypoxic conditions  has proven to be a worthwhile approach to better understand complex biological mechanisms underlying global gene expression profiles in human cancer. Using a simple ex vivo co-culture system allowed us to address a few basic questions about heterotypic cell-cell interactions. First, is global gene expression in a co-culture setting different from the expression in monoculture and, if so, in which respect is it different? Second, how do the responses to co-culture differ among different cell combinations? Third, are the in vitro observations transferable in vivo using published gene expression datasets from human tissue specimens? We analyzed heterotypic interaction effects in stromal fibroblasts and a diverse set of benign and malignant breast epithelial cells in a mixed co-culture setting by measuring changes in global gene expression using DNA microarrays. The global view of the gene expression responses facilitated the identification of specific changes and pathways underlying these effects. Gene expression signatures paralleling a response to heterotypic interaction in this ex vivo model were shared by clinically distinct subgroups of breast cancer.
Identification of heterotypic interaction effects
To further validate the results obtained by cDNA microarray analysis OAS2 transcript levels were measured by quantitative real time PCR (Figure A in Additional data file 2). Moreover, for STAT1 the increase in transcript levels in co-culture (2.8-fold) was paralleled by an increase in STAT1 protein as detected by fluorescence assisted cell sorting (FACS) analysis (Figure B in Additional data file 2).
Since breast cancer is a clinically and molecularly heterogeneous disease, we selected a broad spectrum of different breast cancer cell lines to sample this heterogeneity and explored the effects of heterotypic culture looking for subtype-specific and shared response patterns. We focused on epithelial-mesenchymal interactions co-cultivating fibroblasts of different origins (HTB125 (breast stromal fibroblast), HDF (fibroblast from breast skin) and CCL-171 (embryonic lung fibroblast)), in combination with normal breast epithelial cells (human mammary epithelial cells (HMECs)) and seven widely used breast cancer cell lines.
There was obvious heterogeneity in the responses of different pairs of cells to co-cultivation. The patterns of gene expression changes due to co-cultivation were mainly determined by the type of the epithelial cell involved whereas the origin of the fibroblasts had a minor influence. Importantly, CCL-171, a lung fibroblast, and HTB125, a fibroblast derived from the breast of a cancer patient, induced distinct but very similar interferon responses in co-cultivation with different epithelial cells (Figure 2c). To highlight consistent features of the responses of distinct normal or malignant epithelial cells, representing the distinct types of breast cancer, to co-cultivation with fibroblasts, we collapsed our data into eight groups, one group for each epithelial cell co-cultured with three different types of fibroblasts. There were 3,000 genes that showed a significant reproducible change (FDR < 1%) in transcript levels in response to co-culture in at least one of the groups. Clustering the averaged values of co-culture-induced changes for each group revealed specific and shared effects (Figure 2d). For several cell combinations, co-cultures led to an induction of smooth muscle actin (ACTA2), myosin regulatory light chain interacting protein (MYLIP), myosin, light polypeptide kinase (MYL), myosin regulatory light chain 2, smooth muscle isoform (MYL9), calponin 2 (CNN2) and fibronectin (FN1). Induction of these genes has previously been described to be associated with the acquisition of a myofibroblast phenotype . The ability of the tumor cells to induce this 'myofibroblast' expression program varied among the breast cancer cell lines; the strongest effect was seen with MCF7 cells. In a previous study, conditioned medium of MCF7 cells was shown to induce a myofibroblast phenotype . Targets of the TGF-β pathway, such as the gene encoding latent transforming growth factor beta binding protein LTBP2 and transforming growth factor induced gene TGFBI, were induced in parallel with the 'myofibroblast response'. In fact, TGF-β has previously been shown to induce a 'myofibroblast' phenotype , suggesting that the response observed in these co-cultures might be mediated by the TGF-β pathway.
The most consistent coordinated response, however, was an induction of interferon-associated genes by cultivation of fibroblasts with four of the seven breast cancer cell lines. This response was seen in the co-cultures involving the estrogen-receptor negative breast cancer cell lines MDA-MB231, MDA-MB436, Hs578T and BT549, but neither in HMECs nor in the estrogen-receptor positive breast cancer cells MCF7, T47D and SKBR-3. Although the gene expression profiles of these epithelial cells grown as monocultures reflected their molecular differences, including some consistent differences between the estrogen-receptor negative and estrogen-receptor positive breast cancer cell lines, there were no consistent differences between these groups in baseline expression of the interferon-induced genes in the monocultures. The cell-type specificity is a strong hint that the interferon-response activation is a specific effect of heterotypic interaction. Since we compared the gene-expression responses in the co-cultures with the responses in the corresponding monocultures kept under the same conditions, we can exclude responses to serum stimulation or withdrawal as sources of the interferon response observed in these experiments. The response does not represent an effect of crowding, which is a known inducer of an interferon response , since the cell density in our experiments was maintained below the threshold at which the interferon response genes were turned on (data not shown). Furthermore, we were unable to identify any infective agent in any of the cultures despite extensive testing for mycoplasma, reverse transcriptase activity and viral transcripts, using microarrays that provide a broad survey of human viruses  (data not shown). The consistent cell-type specific, coordinated response suggests that it depends on a specific physiological feature shared among the estrogen-receptor negative human breast cancers, which is retained in long-term culture, enabling them to activate this specific response upon contact with stromal cells.
Localizing expression of interferon-response genes to breast cancer cell lines
Induction of interferon in co-culture
Genomic analysis of epithelial-mesenchymal interaction effects in human cancers
As a first assessment of its potential biological relevance, we compared distant metastasis-free survival and overall-specific survival between the two groups distinguished by the interferon-response genes (Figure 6b). We found that tumors with high expression levels of interferon-response genes had a significantly shorter metastasis-free survival (p = 0.0014; 58% at 10 years) and overall survival (p = 0.001; 59% at 10 years) than tumors with low expression levels (metastasis free survival, 74% at 10 years; overall survival, 80% at 10 years).
The same trend toward unfavorable outcome in patients with cancers showing high levels of interferon-response gene transcripts (p = 0.067) could be seen in an analysis of published data from advanced-stage breast cancers [36, 37]. As a metric that can be compared to known prognostic parameters and applied to other prospectively collected samples, we defined an 'interferon-response score' by averaging the gene expression levels for the 42 genes of the interferon-response gene list. The interferon response did not significantly correlate with clinical parameters such as age of the patient, tumor size, nodal stage or angio-invasion. It was, however, very significantly correlated with tumor grade and estrogen receptor status (p < 10-6; Additional data file 4), paralleling our in vitro findings that cell lines representing estrogen-receptor negative tumors preferentially induce the interferon-response genes in co-culture.
We also investigated the relationship between the interferon-response gene signature and three previously identified gene-expression signatures, which were useful prognosticators in this dataset. The first signature is a set of 70 genes , which was identified in a supervised analysis of a subset of the NKI early stage breast cancer dataset , to predict freedom from metastasis at 5 years. The second signature was identified in vitro by exposing fibroblasts to serum to mimic a wound response, and has been shown to predict risk of progression . The third signature, the response to hypoxia in vitro , is also associated with a poor prognosis. The interferon signature was only very weakly correlated with either the wound signature or the hypoxia signature, and moderately correlated with the 70-gene prognostic profile, whereas the wound signature and the 70-gene score were more strongly correlated to one another (Figure 7). Thus, the interferon response appears to be a distinct feature of breast cancer biology, identifying a subgroup of cancers with a higher propensity for progression.
STAT1 protein expression in a second independent breast cancer dataset
The main objective of this study was to examine and characterize the effects of heterotypic cellular interaction, to gain insight into the underlying biology of these effects in normal mammary tissue and breast cancer. To isolate specific, direct interactions from more complex interactions involving multiple cell types in a whole tissue or organism we used a simple ex vivo co-culture model. Since some important heterotypic interactions can require direct cell-cell contact, we focused on a co-culture model where the two cell types were mixed. A challenge in the analysis of a mixed co-culture model is the separation of the interaction effects induced by signal exchange between the two cell types from the simple additive combination of their intrinsic gene expression patterns in the overall gene expression profile of the co-culture. Our strategy of normalizing for the simple additive effects based on a linear regression model proved to be advantageous, since it does not depend on prior knowledge of the exact proportional contribution of the different cell types to the superposed gene expression pattern. A similar approach has been described to define the proportional contribution of different cell cycle states in a mix of cells, although without taking into account interaction effects . This strategy was effective in isolating the cell-cell interaction effects on gene expression.
We examined the effects on global gene expression of the molecular crosstalk between stromal fibroblasts and each of a diverse set of breast cancer cell lines or normal breast epithelial cells as they interact in vitro. Not unexpectedly, the picture of heterotypic interaction effects that emerged from combinatorial co-cultivation of multiple different cell types was complex, reflecting the different abilities of normal and malignant cells to send and to respond to extrinsic signals. The overall pattern of gene expression changes were dominated by the type of epithelial cells. Against our expectations, which were based on the knowledge that fibroblasts from different parts of the body show distinct gene expression patterns  leading to different physiological properties that persist through many passages of in vitro cultivation, the source of the fibroblast had only a minor influence on gene expression responses to heterotypic interaction in our co-culture system.
We cannot exclude the possibility that fibroblasts isolated from within a tumor might show additional specific interaction effects. Nevertheless, it would be surprising if carcinoma associated fibroblasts failed to show the strong effects that we consistently observed in co-cultures with fibroblasts of diverse origin. We recognize that these experiments might be insufficient to detect subtle differences between co-cultures involving different types of fibroblasts. To rigorously evaluate these differences a more extensive survey of co-culture conditions would be needed.
In our co-culture system a subset of tumor-stroma combinations showed induction of a set of genes characteristic of a 'myofibroblast' phenotype. In the same co-cultures, target genes of the TGF-β pathway were induced in parallel. This coordinated induction is in line with reports describing TGF-β as the major trigger of a 'myofibroblast' phenotype . In vivo, activation of a contractile 'myofibroblast' phenotype in the tumor stroma occurs in a subgroup of patients, leading to shrinkage of the tumor environment causing skin dimpling and nipple retraction, both cardinal signs indicative of breast cancer. This example demonstrates how the analysis of heterotypic interaction effects allows inferring signaling pathways involved in specific physiological and morphological changes of importance in breast cancer.
The most prominent recurring theme arising from the heterotypic interactions we examined was the induction of an interferon-response program in cell lines derived from estrogen-receptor negative breast cancers upon co-culture with fibroblasts. Interferon-response genes showed a strikingly coordinated variation in expression in an analysis of diverse tumors and multiple datasets. Differential regulation of the interferon response genes has been observed in many human malignancies, including leukemias , ovarian cancer , gastric cancer , lung cancer  and breast cancer [30, 43]. In breast cancer, in an attempt to validate the previously described intrinsic gene signatures [36, 37], Hu et al.  assigned a small group of tumors with very high gene expression levels known to be induced by interferon as the 'interferon subtype' with a poor clinical outcome. Despite its common occurrence, the origin and the consequences of this phenomenon are unknown. Some reports have proposed that this program might reflect a viral infection or invasion of inflammatory cells in response to the tumor . Our data suggest that the interferon response is not necessarily dependent on immune cells since our in vitro co-culture system comprises only fibroblasts and epithelial cells and no immune cells. Despite considerable effort to identify infective agents, we could not find any evidence for an infection in our cell culture causing the interferon response. Without excluding these possibilities, we propose that in a subset of breast cancer, the interferon response arises as an effect of the interaction of the malignant epithelial cells with the stroma.
At a first glance, the proposed link between interferon signaling and tumor-stroma interaction is surprising. However, interferons are pleiotropic cytokines, and while best known for their function as a viral defense mechanism they are also involved in other biological processes , such as the induction of cell cycle arrest, apoptosis, cell differentiation, immune stimulation and regulation of bone metabolism . The induction of interferons at the interface between tumor cells and the surrounding stroma may have profound biological significance. In response to viral infection, induction of the interferon response genes, such as EIF2AK2, can lead to a global arrest of translation and subsequent apoptosis . Interferon treatment has an anti-proliferative effect in some cultured cancer cells, and some human cancers shrink in response to interferon , leading to the speculation that an interferon response might be linked to a better prognosis . In fact, our results show the opposite effect; patients with breast cancers displaying high interferon-response gene expression were 1.7 times more likely (95% confidence interval 1.1-2.6; p = 0.018) to develop metastasis and 1.8 times more likely to die of the disease (95% confidence interval 1.2-2.7; p = 0.006) than patients with tumors showing low expression levels of the interferon-response genes. Similar results have been reported by others. For example, an increase in EIF2AK2 expression and activity during tumor progression had been described in melanoma and colorectal cancer . In breast cancer cells EIF2AK2 was elevated compared to normal breast epithelial cells . Also, IFI 27, known to be inducible by IFNα, is frequently over-expressed in breast cancer . IFITM1 over-expression in gastric cancer cells was reported to enhance migration and invasion in vitro . These findings along with the observation that interferon response gene expression in cancer is highly coordinated, suggests the possibility that the interferon response program can promote cancer progression.
The role of STAT1, the main regulator of the interferon response genes, is controversial too. Our finding of a worse outcome for patients with tumors with high levels of STAT1 protein expression support our results of STAT1 mRNA expression levels and is in accordance with the expression levels of the full set of interferon response genes. The mechanism for the negative association between interferon response gene induction and patient outcome is not yet understood. Several mechanisms are possible. Up-regulation of STAT1 was found to be associated with resistance to radiotherapy . IFITM1 was reported in another model to be involved in IFNα induced radioresistance . For patients treated with radiotherapy, high expression of interferon response genes leading to radioresistance of the tumors could contribute to an unfavorable outcome compared to the more radiosensitive tumors. Since all patients receiving breast conserving therapy from the NKI dataset underwent adjuvant radiotherapy, this hypothesis cannot be further substantiated from our data because of the lack of an appropriate control group. Another possible mechanism, independent from an effect on therapeutic efficacy, could be mediated by an effect on invasiveness of the tumor. Up-regulation of STAT1 has been reported in breast cancer micrometastasis in the bone marrow , suggesting a more metastasis-prone phenotype.
The finding that an interferon response can be induced in response to tumor-stroma interaction raises questions for further inquiry. First, our results do not allow us to distinguish whether the interferon response has any role in contributing to tumor progression or is merely an incidental feature of certain cancers that tend to be more aggressive. If it should turn out that the interferon response has a contributing role in progression and metastasis in some tumors, therapeutic application of interferon might be detrimental in such cancers. Indeed, blocking the interferon response induction as a therapeutic target using antibodies or small molecule inhibitors might be beneficial in this situation. Second, apart from infectious agents, molecular signals that can induce interferon secretion are not well defined and the signals that induce interferon secretion in the stromal fibroblasts in our system are still to be discovered. One molecular mechanism for induction of interferon could be stimulation of a member of the toll like receptor family by a tumor cell associated ligand. Endogenous ligands for toll like receptors have been proposed, but further studies are needed to prove their existence . In our experimental system close cell-cell contact was needed to induce the interferon-response, suggesting that a short-range signaling mechanism, perhaps involving a cell-surface ligand, might be involved.
Molecular interactions between epithelial and mesenchymal cells represent only a small part of the molecular conversation among all the interacting cells in the breast cancer microenvironment. The approach used in this work, employing an ex vivo model to develop gene expression signatures as an experimentally tractable window on the more complex interactions in vivo can be deliberately extended to other cells types, such as endothelial, inflammatory and immune cells. This technique may allow us to explore complex interactions among the multiple molecules operating in these cells to orchestrate the process of cancer progression and metastasis. Our experience suggests that in vitro modeling of specific processes and features of the tumor microenvironment can provide a valuable interpretive framework for analyzing the gene expression patterns in more complex heterogeneous in vivo samples and identify effects of heterotypic cellular interactions.
Materials and methods
HMECs (Cambrex Bio Science Walkersville, Inc.,Walkersville, MD, USA) were expanded in mammary epithelial basal medium supplemented with bovine pituitary extract, human epithelial growth factor, insulin and antibiotics (Clonetics, Cambrex Bio Science Walkersville, Inc.). MCF-7, T47D, MDA-MB231, MDA-MB436, SKBR-3, Hs578T, BT549, CCL-171, HTB-125 (ATTC) and HDF (Cambrex Bio Science Walkersville, Inc.) were propagated in DMEM supplemented with 10% FBS (HyClone, Logan, UT, USA), glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin (GIBCO, Grand Island, NY, USA). For co-culture experiments the cells were cultivated for 48 h at 50,000 cells/cm2 in endothelial basal media (Cambrex Bio Science Rockland, Inc., Rockland, ME, USA) with 0.2% FBS, which was a good universal medium for all cells involved. Separated co-cultures were kept in Transwell ® chambers with a 0.4 μm pore size (Costar, Corning Inc., Corning, NY, USA). The cells for analysis were always harvested from the bottom well and reciprocal interactions were tested. Cells negatively tested for mycoplasma infection using MycoAlert™ (Cambrex Bio Science Rockland, Inc.) and VenorGem ® (Sigma, Saint Louis, MO) mycoplasma detection kits used according to the manufacturers' instructions.
Cells were fixed and stained using the Cytofix/Cytoperm™ Kit (BD Biosciences, San Jose, CA, USA) according to the manufacturer's instructions using 20 μg/ml STAT1a mAB (Abcam, Cambridge, MA, USA) and a fluorescein-5-isothyocyanate labeled goat anti-mouse IgG (Sigma-Aldrich, St Louis, MO, USA) for detection. Goat serum 1:200 was used for blocking. Analytical flow cytometry was done on a modified dual laser LSRScan (BD Immunocytometry Systems, San Diego, CA, USA) in the Shared FACS Facility, Center for Molecular and Genetic Medicine at Stanford, using FlowJo software (TreeStar, Ashland, OR, USA) for data analysis.
For FACS sorting, cells were stained with the lipophilic carbocyanine dye DiO (Vybrant ® DiO cell-labeling solution, Molecular Probes™ Invitrogen, Eugene, OR, USA) in serum-free DMEM medium for 20 minutes and washed three times in calcium- and magnesium-free PBS according to the manufacturer's instructions before co-culturing. After 48 h, the cells were detached by incubation in Trypsin/EDTA (GIBCO, Grand Island, NY, USA) for 3 minutes and washed in ice-cold PBS and then immediately put on ice. Cell sorting was done on a MoFlow cell sorter (Becton Dickinson, Mountain View, CA, USA) in the Shared FACS Facility, Center for Molecular and Genetic Medicine at Stanford. The sorted cells were harvested in TRIZOL ® LS Reagent (Invitrogen, Carlsbad, CA, USA). FlowJo software (TreeSTAR) was used for data analysis.
RNA isolation and amplification
After discarding the culture medium and washing the cell layer once with PBS, total RNA was isolated by lysing the cells in the culture dish with RLT buffer (Qiagen, Valencia, CA, USA) and extraction with the RNeasy ® Mini Kit (Qiagen). Total RNA (500 ng) was amplified using the Message Amp ™ II aRNA Kit (Ambion, Austin, TX, USA). The amplification product was checked for integrity by electrophoresis in a 1% agarose gel in MOPS buffer.
cDNA microarrays and hybridization
We used human cDNA microarrays containing 40,700 elements that represent 24,472 unique genes based on Unique Clusters. Arrays were produced at the Stanford Functional Genomic Facility. Complete details regarding the clones on the arrays may be found at Stanford: functional genomics facility . cDNA produced from 6 μg amplified RNA were hybridized to the array in a two-color comparative format, with the experimental samples labeled with one fluorophore (Cy5) and a reference pool of mRNA (Universal human reference, Stratagene, La Jolla, CA, USA) labeled with a second fluorophore (Cy3). Fluorescent dyes were purchased from Amersham Pharmacia Biotech (Piscataway, NJ, USA). Hybridizations were carried out using the standard protocol described previously .
Data analysis and clustering
Array images were scanned using an Axon Scanner 4000B (Axon Instruments, Union City, CA, USA), and image analysis was performed using Genepix Pro version 5.0 220.127.116.11 (Axon Instruments). The raw data files were stored in the Stanford Microarray Database ; the data used for the paper are available at the accompanying website . Data were expressed as the log2 ratio of fluorescence intensities of the sample and the reference, for each element on the array.
The (Cy5/Cy3) ratio is defined in the Stanford Microarray Database as the normalized ratio of the background-corrected intensities. Spots with aberrant measurements due to obvious array artifacts or poor technical quality were manually flagged and removed from further analysis. A filter was applied to omit measurements where the fluorescent signal from the DNA spot was less than 50% above the measured background fluorescence surrounding the printed DNA spot in either the Cy3 or Cy5 channel. Genes that did not meet these criteria for at least 80% of the measurements across the experimental samples were excluded from further analysis. Valid data were filtered to exclude elements that did not have at least a three-fold deviation from the mean in at least three samples. Data were evaluated by unsupervised hierarchical clustering  and SAM  and displayed using Treeview .
Determination of the heterotypic interaction effect on gene expression
To facilitate the identification of heterotypic interaction effects on global gene expression in a mixed co-culture experiment, we normalized the gene expression data based on the proportional contribution of each cell type to transcript abundance. Given that the average gene does not change due to heterotypic interaction and that there are simple additive effects to account for, we used a linear regression fit for normalization. To determine the contribution of each cell type to the combined gene expression pattern in the linear regression model, the expression levels of the monocultures are the predictors and the expression levels of the co-culture, the response.
Specifically, a set of equations (1-n) is established (one per gene), as illustrated in the additional data file 5 in which the expression level of gene n (en, co-culture) in the mixed co-culture equals the fraction a of mRNA from cell type 1 times the relative expression level of gene n in type 1 mono-cultured cells plus the fraction (1-a) of mRNA from cell type 2 times the relative expression level of gene n in type 2 mono-cultured cells multiplied by the interaction coefficient In. We assume that the average gene is not influenced by heterotypic interaction in the mixed co-culture represented as I = 1. Since the dataset over e1-n is skewed, we empirically identified a linear regression fit based on Gamma errors and identity link as a good model to calculate a. Then the equations 1-n can then be solved for I1-n, which results in a profile of interaction effects for the genes1-n. These interaction effects can be analyzed in much the same way as conventional gene expression measurements.
Real time quantitative PCR
Total RNA (500 ng) was mixed with dT16 primer in a volume of 11 μl, incubated at 65°C for 10 minutes and immediately put on ice. Following addition of 100 units Superscript II reverse transcriptase (GIBCO, Carlsbad, CA, USA), reverse transcription was performed for 2 h at 42°C in 1× RT reaction buffer (GIBCO), 10 μM dithiothreitol, 500 μM dNTP (Amersham Biosciences, Pittsburgh, PA, USA) with 2.5 μM dT16 primer in a volume of 20 μl. PCR reactions were performed in a final volume of 20 μl with cDNA prepared from 20 ng RNA and a final concentration of 1× SYBR ®Green PCR Master Mix (ABI, Foster City, CA, USA) and 200 nM of each primer (sequences: GAPDH, forward GAAGGTGAAGGTCGGAGTC, reverse GAAGATGGTGATGGGATTTC; OAS2, forward GGAATACCTGAAGCCCTACGAA, reverse CCTGCAGACGTCACAGATGGT; IFNα, forward CCTCGCCCTTTGCTTTACTG, reverse GCCCAGAGAGCAGCTTGACT; IFNβ, forward ACCTCCGAAACTGAAGATCTCCTA, reverse TGCTGGTTGAAGAATGCTTGA). The reaction was run in an ABI 7700 Sequence Detection System with the following cycling conditions: 50°C for 2 minutes, 94°C for 10 minutes, then 40 cycles of 94°C for 15 s and 60°C for 60 s. For each gene a standard curve was prepared and triplicate measurements were performed for each sample.
Large biopsy or tissue microarray sections were cut from paraffin blocks, deparaffinized in xylene, and hydrated in a graded series of alcohol. The slides were pretreated with citrate buffer and a microwave step. Immunostaining was performed using the DAKO Envision+ System, Peroxidase DAB, (DAKO, Cambridgeshire, United Kingdom) for STAT1a monoclonal antibody (1:100 dilution; Abcam).
We stained 1,024 tissue cores from 521 donor blocks. Immunohistochemistry images were acquired with the BLISS Microscope System (Bacus Laboratories, Lombard, IL, USA). Staining results were assessed using a three-point scoring system, where 0 = invasive tumor cells present in the tissue core and no staining seen, 1 = invasive tumor cells present with weak staining intensity and/or < 20% of the cells stained, and 2 = invasive tumor cells present with strong staining in > 20% of the cells. Tissue cores that failed to adhere to the glass slide, did not contain invasive carcinoma or were otherwise uninterpretable were excluded. Scoring of the arrays was analyzed using the Deconvoluter software as previously described , with each sample receiving the higher of the scores for two replicate cores.
Human breast cancer dataset
The dataset for breast cancer contained 295 tumors analyzed on a 25,000 spot oligonucleotide array as described . In brief, patients were diagnosed and treated at the NKI for early stage breast cancer (stage I and II) between 1984 and 1995. The clinical data were updated in January 2005. The median follow-up for patients still alive is now 12.3 years.
The interferon response gene list consists of 42 genes represented by 49 image clones on the cDNA Stanford array. Clones having the same Unigene locus were removed. The gene sequences were mapped to spots on the NKI array using Unigene build number 184 (released on 9 June 2005) to give 29 unique spots. In order to overcome possible overweighting of clones from Unigene clusters that were matched to more than one probe on the NKI array, expression values derived from probes that were not matched to the same Unigene cluster were averaged. Expression measurements for each gene were mean centered. The resulting dataset was subjected to hierarchical clustering with average linkage clustering  and displayed with Treeview .
Distant metastasis was analyzed as first event only (distant metastasis-free probability). If a patient developed a local recurrence, axillary recurrence, contralateral breast cancer or a second primary cancer (except for non-melanoma skin cancer), she was censored at that time and subsequent distant metastases were not analyzed. This is based on the theoretical possibility that the locally recurrent or second primary cancers could be a source for distant metastases. An ipsilateral supra-clavicular recurrence was soon followed by a distant metastasis in all but one patient. An ipsilateral supra-clavicular recurrence was thus considered the first clinical evidence for metastatic disease for this analysis and patients were not censored at the time of ipsilateral supra-clavicular recurrence. Overall survival was analyzed based on death from any cause and patients were censored at last follow up. Kaplan-Meier survival curves were compared by the Cox/Mantel log/rank test using Winstat for Microsoft Excel (RFitch Software, Staufen, Germany). Multivariate analysis by the Cox proportional hazard method was performed using the software package SPSS R 11.5 (SPSS, Inc., Chicago, IL, USA).
A dataset of gene expression patterns from advanced breast cancers was described by Sorlie et al. [36, 37]. Expression data from 19 image clones representing the interferon response gene list were included in this dataset. Genes and samples were organized by hierarchical clustering. Relapse-free and overall survival were calculated as described above.
The independent breast cancer tissue microarray validation series is as described ; immunohistochemical scores for STAT1 were related to breast cancer-specific survival by Kaplan-Meier analysis with log-rank test.
Additional data files
The following additional data are available with the online version of this paper. Additional data file 1 is a table listing the interferon response genes. Additional data file 2 shows the expression of OAS2 measured by RT-PCR in the co-culture CCL171/MDA-MB-231 and the expression of STAT1 measured by immunofluorescent staining and FACS analysis. Additional data file 3 shows the analysis of the interferon response signature in advanced human breast cancers. Additional data file 4 shows box and scatter plots illustrating the correlation of the interferon score to clinical parameters with known prognostic significance. Additional data file 5 is an illustration of the linear regression model used to normalize for additive effects in the mixed co-culture gene expression data.
Dulbecco's modified Eagle's medium
eukaryotic translation initiation factor 2-alpha kinase
fluorescent assisted cell sorting
fetal bovine serum
false discovery rate
human mammary epithelial cell
interferon-induced protein with tetratricopeptide repeats
Netherlands Cancer Institute
significance analysis of microarray data
signal transducer and activator of transcription
transforming growth factor.
We would like to thank Michael Fero, John Collar, Elena Seraia and the staff of the Stanford Functional Genomics Facility for supplying us with the human cDNA microarrays used for this study. We thank our colleagues of the Brown lab for discussion, Janos Demeter, Jeremy Gollub, Gavin Sherlock, Catherine Bell and other staff of the Stanford Microarray Database. This work was supported by NIH grant CA111487 and by the HHMI and the Dutch Cancer Society NKB 2002-2575 (DSAN). POB is an investigator of the Howard Hughes Medical Institute. MB is a fellow of the Krebsliga beider Basel, the M. and W. Lichtenstein-Stiftung Basel, and the Huggenberger-Krebsstiftung, Zürich, Switzerland.
- Cunha GR, Hom YK: Role of mesenchymal-epithelial interactions in mammary gland development. J Mammary Gland Biol Neoplasia. 1996, 1: 21-35. 10.1007/BF02096300.PubMedView ArticleGoogle Scholar
- Levine JF, Stockdale FE: Cell-cell interactions promote mammary epithelial cell differentiation. J Cell Biol. 1985, 100: 1415-1422. 10.1083/jcb.100.5.1415.PubMedView ArticleGoogle Scholar
- Weaver M, Batts L, Hogan BL: Tissue interactions pattern the mesenchyme of the embryonic mouse lung. Dev Biol. 2003, 258: 169-184. 10.1016/S0012-1606(03)00117-9.PubMedView ArticleGoogle Scholar
- Howlett AR, Bissell MJ: The influence of tissue microenvironment (stroma and extracellular matrix) on the development and function of mammary epithelium. Epithelial Cell Biol. 1993, 2: 79-89.PubMedGoogle Scholar
- Howard B, Ashworth A: Signalling pathways implicated in early mammary gland morphogenesis and breast cancer. PLoS Genet. 2006, 2: e112-10.1371/journal.pgen.0020112.PubMedPubMed CentralView ArticleGoogle Scholar
- Bissell MJ, Radisky DC, Rizki A, Weaver VM, Petersen OW: The organizing principle: microenvironmental influences in the normal and malignant breast. Differentiation. 2002, 70: 537-546. 10.1046/j.1432-0436.2002.700907.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Bissell MJ, Radisky D: Putting tumours in context. Nat Rev Cancer. 2001, 1: 46-54. 10.1038/35094059.PubMedPubMed CentralView ArticleGoogle Scholar
- Hanahan D, Weinberg RA: The hallmarks of cancer. Cell. 2000, 100: 57-70. 10.1016/S0092-8674(00)81683-9.PubMedView ArticleGoogle Scholar
- Arias AM: Epithelial mesenchymal interactions in cancer and development. Cell. 2001, 105: 425-431. 10.1016/S0092-8674(01)00365-8.PubMedView ArticleGoogle Scholar
- Sneddon JB, Zhen HH, Montgomery K, van de Rijn M, Tward AD, West R, Gladstone H, Chang HY, Morganroth GS, Oro AE, Brown PO: Bone morphogenetic protein antagonist gremlin 1 is widely expressed by cancer-associated stromal cells and can promote tumor cell proliferation. Proc Natl Acad Sci USA. 2006, 103: 14842-14847. 10.1073/pnas.0606857103.PubMedPubMed CentralView ArticleGoogle Scholar
- Olumi AF, Grossfeld GD, Hayward SW, Carroll PR, Tlsty TD, Cunha GR: Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer Res. 1999, 59: 5002-5011.PubMedGoogle Scholar
- Camps JL, Chang SM, Hsu TC, Freeman MR, Hong SJ, Zhau HE, von Eschenbach AC, Chung LW: Fibroblast-mediated acceleration of human epithelial tumor growth in vivo. Proc Natl Acad Sci USA. 1990, 87: 75-79. 10.1073/pnas.87.1.75.PubMedPubMed CentralView ArticleGoogle Scholar
- Kuperwasser C, Chavarria T, Wu M, Magrane G, Gray JW, Carey L, Richardson A, Weinberg RA: Reconstruction of functionally normal and malignant human breast tissues in mice[see comment]. Proc Natl Acad Sci USA. 2004, 101: 4966-4971. 10.1073/pnas.0401064101.PubMedPubMed CentralView ArticleGoogle Scholar
- Bhowmick NA, Chytil A, Plieth D, Gorska AE, Dumont N, Shappell S, Washington MK, Neilson EG, Moses HL: TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science. 2004, 303: 848-851. 10.1126/science.1090922.PubMedView ArticleGoogle Scholar
- Cheng N, Bhowmick NA, Chytil A, Gorksa AE, Brown KA, Muraoka R, Arteaga CL, Neilson EG, Hayward SW, Moses HL: Loss of TGF-beta type II receptor in fibroblasts promotes mammary carcinoma growth and invasion through upregulation of TGF-alpha-, MSP- and HGF-mediated signaling networks. Oncogene. 2005, 24: 5053-5068. 10.1038/sj.onc.1208685.PubMedPubMed CentralView ArticleGoogle Scholar
- Petersen OW, Nielsen HL, Gudjonsson T, Villadsen R, Rank F, Niebuhr E, Bissell MJ, Ronnov-Jessen L: Epithelial to mesenchymal transition in human breast cancer can provide a nonmalignant stroma. Am J Pathol. 2003, 162: 391-402.PubMedPubMed CentralView ArticleGoogle Scholar
- Nielsen HL, Ronnov-Jessen L, Villadsen R, Petersen OW: Identification of EPSTI1, a novel gene induced by epithelial-stromal interaction in human breast cancer. Genomics. 2002, 79: 703-710. 10.1006/geno.2002.6755.PubMedView ArticleGoogle Scholar
- Khodarev NN, Yu J, Labay E, Darga T, Brown CK, Mauceri HJ, Yassari R, Gupta N, Weichselbaum RR: Tumour-endothelium interactions in co-culture: coordinated changes of gene expression profiles and phenotypic properties of endothelial cells. J Cell Sci. 2003, 116: 1013-1022. 10.1242/jcs.00281.PubMedView ArticleGoogle Scholar
- Allinen M, Beroukhim R, Cai L, Brennan C, Lahti-Domenici J, Huang H, Porter D, Hu M, Chin L, Richardson A, et al: Molecular characterization of the tumor microenvironment in breast cancer [see comment]. Cancer Cell. 2004, 6: 17-32. 10.1016/j.ccr.2004.06.010.PubMedView ArticleGoogle Scholar
- West RB, Nuyten DS, Subramanian S, Nielsen TO, Corless CL, Rubin BP, Montgomery K, Zhu S, Patel R, Hernandez-Boussard T, et al: Determination of stromal signatures in breast carcinoma. PLoS Biol. 2005, 3: e187-10.1371/journal.pbio.0030187.PubMedPubMed CentralView ArticleGoogle Scholar
- Whitney AR, Diehn M, Popper SJ, Alizadeh AA, Boldrick JC, Relman DA, Brown PO: Individuality and variation in gene expression patterns in human blood. Proc Natl Acad Sci USA. 2003, 100: 1896-1901. 10.1073/pnas.252784499.PubMedPubMed CentralView ArticleGoogle Scholar
- Lamb J, Ramaswamy S, Ford HL, Contreras B, Martinez RV, Kittrell FS, Zahnow CA, Patterson N, Golub TR, Ewen ME: A mechanism of cyclin D1 action encoded in the patterns of gene expression in human cancer. Cell. 2003, 114: 323-334. 10.1016/S0092-8674(03)00570-1.PubMedView ArticleGoogle Scholar
- Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JC, Trent JM, Staudt LM, Hudson J, Boguski MS, et al: The transcriptional program in the response of human fibroblasts to serum[see comment]. Science. 1999, 283: 83-87. 10.1126/science.283.5398.83.PubMedView ArticleGoogle Scholar
- Chang HY, Sneddon JB, Alizadeh AA, Sood R, West RB, Montgomery K, Chi JT, van de Rijn M, Botstein D, Brown PO: Gene expression signature of fibroblast serum response predicts human cancer progression: similarities between tumors and wounds. PLoS Biol. 2004, 2: E7-10.1371/journal.pbio.0020007.PubMedPubMed CentralView ArticleGoogle Scholar
- Chi JT, Wang Z, Nuyten DS, Rodriguez EH, Schaner ME, Salim A, Wang Y, Kristensen GB, Helland A, Borresen-Dale AL, et al: Gene expression programs in response to hypoxia: cell type specificity and prognostic significance in human cancers. PLoS Med. 2006, 3: e47-10.1371/journal.pmed.0030047.PubMedPubMed CentralView ArticleGoogle Scholar
- Tusher VG, Tibshirani R, Chu G: Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA. 2001, 98: 5116-5121. 10.1073/pnas.091062498.PubMedPubMed CentralView ArticleGoogle Scholar
- Lazard D, Sastre X, Frid MG, Glukhova MA, Thiery JP, Koteliansky VE: Expression of smooth muscle-specific proteins in myoepithelium and stromal myofibroblasts of normal and malignant human breast tissue. Proc Natl Acad Sci USA. 1993, 90: 999-1003. 10.1073/pnas.90.3.999.PubMedPubMed CentralView ArticleGoogle Scholar
- Valenti MT, Azzarello G, Balducci E, Sartore S, Sandri M, Manconi R, Sicari U, Bari M, Vinante O: Conditioned medium from MCF-7 cell line induces myofibroblast differentiation, decreased cell proliferation, and increased apoptosis in cultured normal fibroblasts but not in fibroblasts from malignant breast tissue. Histochem J. 2001, 33: 499-509. 10.1023/A:1014927305775.PubMedView ArticleGoogle Scholar
- Desmouliere A, Guyot C, Gabbiani G: The stroma reaction myofibroblast: a key player in the control of tumor cell behavior. Int J Dev Biol. 2004, 48: 509-517. 10.1387/ijdb.041802ad.PubMedView ArticleGoogle Scholar
- Perou CM, Jeffrey SS, van de Rijn M, Rees CA, Eisen MB, Ross DT, Pergamenschikov A, Williams CF, Zhu SX, Lee JC, et al: Distinctive gene expression patterns in human mammary epithelial cells and breast cancers. Proc Natl Acad Sci USA. 1999, 96: 9212-9217. 10.1073/pnas.96.16.9212.PubMedPubMed CentralView ArticleGoogle Scholar
- Wang D, Coscoy L, Zylberberg M, Avila PC, Boushey HA, Ganem D, DeRisi JL: Microarray-based detection and genotyping of viral pathogens. Proc Natl Acad Sci USA. 2002, 99: 15687-15692. 10.1073/pnas.242579699.PubMedPubMed CentralView ArticleGoogle Scholar
- Suzuki T, Urano T, Tsukui T, Horie-Inoue K, Moriya T, Ishida T, Muramatsu M, Ouchi Y, Sasano H, Inoue S: Estrogen-responsive finger protein as a new potential biomarker for breast cancer. Clin Cancer Res. 2005, 11: 6148-6154. 10.1158/1078-0432.CCR-05-0040.PubMedView ArticleGoogle Scholar
- Parrinello S, Coppe JP, Krtolica A, Campisi J: Stromal-epithelial interactions in aging and cancer: senescent fibroblasts alter epithelial cell differentiation. J Cell Sci. 2005, 118: 485-496. 10.1242/jcs.01635.PubMedView ArticleGoogle Scholar
- Orimo A, Gupta PB, Sgroi DC, Arenzana-Seisdedos F, Delaunay T, Naeem R, Carey VJ, Richardson AL, Weinberg RA: Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell. 2005, 121: 335-348. 10.1016/j.cell.2005.02.034.PubMedView ArticleGoogle Scholar
- van de Vijver MJ, He YD, van't Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ, et al: A gene-expression signature as a predictor of survival in breast cancer [see comment]. New Engl J Med. 2002, 347: 1999-2009. 10.1056/NEJMoa021967.PubMedView ArticleGoogle Scholar
- Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, et al: Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001, 98: 10869-10874. 10.1073/pnas.191367098.PubMedPubMed CentralView ArticleGoogle Scholar
- Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, Deng S, Johnsen H, Pesich R, Geisler S, et al: Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci USA. 2003, 100: 8418-8423. 10.1073/pnas.0932692100.PubMedPubMed CentralView ArticleGoogle Scholar
- van 't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, et al: Gene expression profiling predicts clinical outcome of breast cancer. Nature. 2002, 415: 530-536. 10.1038/415530a.PubMedView ArticleGoogle Scholar
- Chang HY, Nuyten DS, Sneddon JB, Hastie T, Tibshirani R, Sorlie T, Dai H, He YD, van't Veer LJ, Bartelink H, et al: Robustness, scalability, and integration of a wound-response gene expression signature in predicting breast cancer survival. Proc Natl Acad Sci USA. 2005, 102: 3738-3743. 10.1073/pnas.0409462102.PubMedPubMed CentralView ArticleGoogle Scholar
- Makretsov NA, Huntsman DG, Nielsen TO, Yorida E, Peacock M, Cheang MC, Dunn SE, Hayes M, van de Rijn M, Bajdik C, Gilks CB: Hierarchical clustering analysis of tissue microarray immunostaining data identifies prognostically significant groups of breast carcinoma. Clin Cancer Res. 2004, 10: 6143-6151. 10.1158/1078-0432.CCR-04-0429.PubMedView ArticleGoogle Scholar
- Lu P, Nakorchevskiy A, Marcotte EM: Expression deconvolution: a reinterpretation of DNA microarray data reveals dynamic changes in cell populations. Proc Natl Acad Sci USA. 2003, 100: 10370-10375. 10.1073/pnas.1832361100.PubMedPubMed CentralView ArticleGoogle Scholar
- Chang HY, Chi JT, Dudoit S, Bondre C, van de Rijn M, Botstein D, Brown PO: Diversity, topographic differentiation, and positional memory in human fibroblasts. Proc Natl Acad Sci USA. 2002, 99: 12877-12882. 10.1073/pnas.162488599.PubMedPubMed CentralView ArticleGoogle Scholar
- Einav U, Tabach Y, Getz G, Yitzhaky A, Ozbek U, Amariglio N, Izraeli S, Rechavi G, Domany E: Gene expression analysis reveals a strong signature of an interferon-induced pathway in childhood lymphoblastic leukemia as well as in breast and ovarian cancer. Oncogene. 2005, 24: 6367-6375.PubMedGoogle Scholar
- Schaner ME, Ross DT, Ciaravino G, Sorlie T, Troyanskaya O, Diehn M, Wang YC, Duran GE, Sikic TL, Caldeira S, et al: Gene expression patterns in ovarian carcinomas. Mol Biol Cell. 2003, 14: 4376-4386. 10.1091/mbc.E03-05-0279.PubMedPubMed CentralView ArticleGoogle Scholar
- Chen X, Leung SY, Yuen ST, Chu KM, Ji J, Li R, Chan AS, Law S, Troyanskaya OG, Wong J, et al: Variation in gene expression patterns in human gastric cancers. Mol Biol Cell. 2003, 14: 3208-3215. 10.1091/mbc.E02-12-0833.PubMedPubMed CentralView ArticleGoogle Scholar
- Garber ME, Troyanskaya OG, Schluens K, Petersen S, Thaesler Z, Pacyna-Gengelbach M, van de Rijn M, Rosen GD, Perou CM, Whyte RI, et al: Diversity of gene expression in adenocarcinoma of the lung [erratum appears in Proc Natl Acad Sci USA 99:1098]. Proc Natl Acad Sci USA. 2001, 98: 13784-13789. 10.1073/pnas.241500798.PubMedPubMed CentralView ArticleGoogle Scholar
- Hu Z, Fan C, Oh DS, Marron JS, He X, Qaqish BF, Livasy C, Carey LA, Reynolds E, Dressler L, et al: The molecular portraits of breast tumors are conserved across microarray platforms. BMC Genomics. 2006, 7: 96-10.1186/1471-2164-7-96.PubMedPubMed CentralView ArticleGoogle Scholar
- Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber RD: How cells respond to interferons. Annu Rev Biochem. 1998, 67: 227-264. 10.1146/annurev.biochem.67.1.227.PubMedView ArticleGoogle Scholar
- Takayanagi H, Kim S, Matsuo K, Suzuki H, Suzuki T, Sato K, Yokochi T, Oda H, Nakamura K, Ida N, et al: RANKL maintains bone homeostasis through c-Fos-dependent induction of interferon-beta. Nature. 2002, 416: 744-749. 10.1038/416744a.PubMedView ArticleGoogle Scholar
- Donze O, Jagus R, Koromilas AE, Hershey JW, Sonenberg N: Abrogation of translation initiation factor eIF-2 phosphorylation causes malignant transformation of NIH 3T3 cells. EMBO J. 1995, 14: 3828-3834.PubMedPubMed CentralGoogle Scholar
- Jonasch E, Haluska FG: Interferon in oncological practice: review of interferon biology, clinical applications, and toxicities. Oncologist. 2001, 6: 34-55. 10.1634/theoncologist.6-1-34.PubMedView ArticleGoogle Scholar
- Kim SH, Gunnery S, Choe JK, Mathews MB: Neoplastic progression in melanoma and colon cancer is associated with increased expression and activity of the interferon-inducible protein kinase, PKR. Oncogene. 2002, 21: 8741-8748. 10.1038/sj.onc.1205987.PubMedView ArticleGoogle Scholar
- Kim SH, Forman AP, Mathews MB, Gunnery S: Human breast cancer cells contain elevated levels and activity of the protein kinase, PKR. Oncogene. 2000, 19: 3086-3094. 10.1038/sj.onc.1203632.PubMedView ArticleGoogle Scholar
- Rasmussen UB, Wolf C, Mattei MG, Chenard MP, Bellocq JP, Chambon P, Rio MC, Basset P: Identification of a new interferon-alpha-inducible gene (p27) on human chromosome 14q32 and its expression in breast carcinoma. Cancer Res. 1993, 53: 4096-4101.PubMedGoogle Scholar
- Yang Y, Lee JH, Kim KY, Song HK, Kim JK, Yoon SR, Cho D, Song KS, Lee YH, Choi I: The interferon-inducible 9-27 gene modulates the susceptibility to natural killer cells and the invasiveness of gastric cancer cells. Cancer Lett. 2005, 221: 191-200. 10.1016/j.canlet.2004.08.022.PubMedView ArticleGoogle Scholar
- Khodarev NN, Beckett M, Labay E, Darga T, Roizman B, Weichselbaum RR: STAT1 is overexpressed in tumors selected for radioresistance and confers protection from radiation in transduced sensitive cells. Proc Natl Acad Sci USA. 2004, 101: 1714-1719. 10.1073/pnas.0308102100.PubMedPubMed CentralView ArticleGoogle Scholar
- Kita K, Sugaya S, Zhai L, Wu YP, Wano C, Chigira S, Nomura J, Takahashi S, Ichinose M, Suzuki N: Involvement of LEU13 in interferon-induced refractoriness of human RSa cells to cell killing by X rays. Radiat Res. 2003, 160: 302-308. 10.1667/RR3039.PubMedView ArticleGoogle Scholar
- Woelfle U, Cloos J, Sauter G, Riethdorf L, Janicke F, van Diest P, Brakenhoff R, Pantel K: Molecular signature associated with bone marrow micrometastasis in human breast cancer. Cancer Res. 2003, 63: 5679-5684.PubMedGoogle Scholar
- Ohashi K, Burkart V, Flohe S, Kolb H: Cutting edge: heat shock protein 60 is a putative endogenous ligand of the toll-like receptor-4 complex. J Immunol. 2000, 164: 558-561.PubMedView ArticleGoogle Scholar
- Stanford: Functional Genomics Facility. [http://www.microarray.org/sfgf/jsp/servicesFrame.jsp#productionArrays]
- Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, et al: Molecular portraits of human breast tumours. Nature. 2000, 406: 747-752. 10.1038/35021093.PubMedView ArticleGoogle Scholar
- Ball CA, Awad IA, Demeter J, Gollub J, Hebert JM, Hernandez-Boussard T, Jin H, Matese JC, Nitzberg M, Wymore F, et al: The Stanford Microarray Database accommodates additional microarray platforms and data formats. Nucleic Acids Res. 2005, 33: D580-582. 10.1093/nar/gki006.PubMedPubMed CentralView ArticleGoogle Scholar
- SMD : View Publication : Characterization of heterotypic interaction effects in vitro to deconvolute global gene expression profiles in cancer. [http://microarray-pubs.stanford.edu/tumor-stroma-interaction]
- Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA. 1998, 95: 14863-14868. 10.1073/pnas.95.25.14863.PubMedPubMed CentralView ArticleGoogle Scholar
- Eisen Lab: Software. [http://rana.lbl.gov/EisenSoftware.htm]
- Liu CL, Prapong W, Natkunam Y, Alizadeh A, Montgomery K, Gilks CB, van de Rijn M: Software tools for high-throughput analysis and archiving of immunohistochemistry staining data obtained with tissue microarrays. Am J Pathol. 2002, 161: 1557-1565.PubMedPubMed CentralView ArticleGoogle Scholar
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.