An integrated framework to model cellular phenotype as a component of biochemical networks.

Identification of regulatory molecules in signaling pathways is critical for understanding cellular behavior. Given the complexity of the transcriptional gene network, the relationship between molecular expression and phenotype is difficult to determine using reductionist experimental methods. Computational models provide the means to characterize regulatory mechanisms and predict phenotype in the context of gene networks. Integrating gene expression data with phenotypic data in transcriptional network models enables systematic identification of critical molecules in a biological network. We developed an approach based on fuzzy logic to model cell budding in Saccharomyces cerevisiae using time series expression microarray data of the cell cycle. Cell budding is a phenotype of viable cells undergoing division. Predicted interactions between gene expression and phenotype reflected known biological relationships. Dynamic simulation analysis reproduced the behavior of the yeast cell cycle and accurately identified genes and interactions which are essential for cell viability.

Cyclin D1/cyclin-dependent kinase 4 interacts with filamin A and affects the migration and invasion potential of breast cancer cells.

Cyclin D1 belongs to a family of proteins that regulate progression through the G1-S phase of the cell cycle by binding to cyclin-dependent kinase (cdk)-4 to phosphorylate the retinoblastoma protein and release E2F transcription factors for progression through cell cycle. Several cancers, including breast, colon, and prostate, overexpress the cyclin D1 gene. However, the correlation of cyclin D1 overexpression with E2F target gene regulation or of cdk-dependent cyclin D1 activity with tumor development has not been identified. This suggests that the role of cyclin D1 in oncogenesis may be independent of its function as a cell cycle regulator. One such function is the role of cyclin D1 in cell adhesion and motility. Filamin A (FLNa), a member of the actin-binding filamin protein family, regulates signaling events involved in cell motility and invasion. FLNa has also been associated with a variety of cancers including lung cancer, prostate cancer, melanoma, human bladder cancer, and neuroblastoma. We hypothesized that elevated cyclin D1 facilitates motility in the invasive MDA-MB-231 breast cancer cell line. We show that MDA-MB-231 motility is affected by disturbing cyclin D1 levels or cyclin D1-cdk4/6 kinase activity. Using mass spectrometry, we find that cyclin D1 and FLNa coimmunoprecipitate and that lower levels of cyclin D1 are associated with decreased phosphorylation of FLNa at Ser2152 and Ser1459. We also identify many proteins related to cytoskeletal function, biomolecular synthesis, organelle biogenesis, and calcium regulation whose levels of expression change concomitant with decreased cell motility induced by decreased cyclin D1 and cyclin D1-cdk4/6 activities.

The cell fate determination factor DACH1 is expressed in estrogen receptor-alpha-positive breast cancer and represses estrogen receptor-alpha signaling.

The Dachshund (dac) gene, initially cloned as a dominant inhibitor of the Drosophila hyperactive EGFR mutant ellipse, encodes a key component of the cell fate determination pathway involved in Drosophila eye development. Analysis of more than 2,200 breast cancer samples showed improved survival by some 40 months in patients whose tumors expressed DACH1. Herein, DACH1 and estrogen receptor-alpha (ERalpha) expressions were inversely correlated in human breast cancer. DACH1 bound and inhibited ERalpha function. Nuclear DACH1 expression inhibited estradiol (E(2))-induced DNA synthesis and cellular proliferation. DACH1 bound ERalpha in immunoprecipitation-Western blotting, associated with ERalpha in chromatin immunoprecipitation, and inhibited ERalpha transcriptional activity, requiring a conserved DS domain. Proteomic analysis identified proline, glutamic acid, and leucine rich protein 1 (PELP1) as a DACH1-binding protein. The DACH1 COOH terminus was required for binding to PELP1. DACH1 inhibited induction of ERalpha signaling. E(2) recruited ERalpha and disengaged corepressors from DACH1 at an endogenous ER response element, allowing PELP1 to serve as an ERalpha coactivator. DACH1 expression, which is lost in poor prognosis human breast cancer, functions as an endogenous inhibitor of ERalpha function.

Human breast cancer-associated fibroblasts (CAFs) show caveolin-1 downregulation and RB tumor suppressor functional inactivation: Implications for the response to hormonal therapy.

It is becoming increasingly apparent that the tumor microenvironment plays a critical role in human breast cancer onset and progression. Therefore, we isolated cancer-associated fibroblasts (CAFs) from human breast cancer lesions and studied their properties, as compared with normal mammary fibroblasts (NFs) isolated from the same patient. Here, we demonstrate that 8 out of 11 CAFs show dramatic downregulation of caveolin-1 (Cav-1) protein expression; Cav-1 is a well-established marker that is normally decreased during the oncogenic transformation of fibroblasts. Next, we performed gene expression profiling studies (DNA microarray) and established a CAF gene expression signature. Interestingly, the expression signature associated with CAFs encompasses a large number of genes that are regulated via the RB-pathway. The CAF gene signature is also predictive of poor clinical outcome in breast cancer patients that were treated with tamoxifen mono-therapy, indicating that CAFs may be useful for predicting the response to hormonal therapy. Finally, we show that replacement of Cav-1 expression in CAFs (using a cell-permeable peptide approach) is sufficient to revert their hyper-proliferative phenotype and prevent RB hyper-phosphorylation. Taken together, these studies highlight the critical role of Cav-1 downregulation in maintaining the abnormal phenotype of human breast cancer-associated fibroblasts.

Phosphatidylcholine removal from brain lipid extracts expands lipid detection and enhances phosphoinositide quantification by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry.

The utilization of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry for the analytical detection and quantification of phosphoinositides and other lipids in lipid extracts from biological samples was explored. Since phosphatidylcholine species in crude extracts have been shown to cause ion suppression of the MS signals for other lipids, a minicolumn of a silica gel cation exchanger was used to adsorb the cationic lipids including the phosphatidylcholine species from the chloroform phase of fetal and adult murine brain extracts. In positive ion mode, lipid peaks that had been completely suppressed in the crude extract became readily detectable and quantifiable in the flow-through fraction from the column. In negative ion mode, improved sensitivity made it possible to readily detect and measure phosphatidylinositol-4,5-bisphosphate (PIP(2)) which had been only marginally detectable before the fractionation. By incorporating an internal standard into the samples, the relative MALDI-TOF MS signals obtained for increasing concentrations of mammalian phosphatidylinositol (PtdIns) increased linearly with correlation coefficients >0.95. Using strong cation exchange minicolumn treated extracts, the levels of PtdIns and PIP(2) in adult and fetal murine brains were measured and compared. The removal of cationic lipids from the chloroform-methanol murine brain extracts resulted in improved overall detection of neutral and anionic lipids and quantification of phosphoinositides by MALDI-TOF MS.

A signature-based method to distinguish time-of-flight secondary-ion mass spectra from biological samples.

Time-Of-Flight Mass Spectrometry (TOF-SIMS) was used to determine elemental and biomolecular ions from isolated protein samples. We identified a set of 23 mass-to-charge ratio (m/z) peaks that represent signatures for distinguishing biological samples. The 23 peaks were identified by Singular Value Decomposition (SVD) and Canonical Analysis (CA) to find the underlying structure in the complex mass-spectra data sets. From this modified data, SVD was used to identify sets of m/z peaks, and we used these patterns from the TOF-SIMS data to predict the biological source from which individual mass spectra were generated. The signatures were validated using an additional data set different from the initial training set used to identify the signatures. We present a simple method to identify multiple variables required for sample classification based on mass spectra that avoids overfit. This is important in a variety of studies using mass spectrometry, including the ability to identify proteins in complex mixtures and for the identification of new biomarkers.

A three-dimensional model of intercellular calcium signaling in epithelial cells.

We have developed a fully three-dimensional (3D) model of calcium signaling in epithelial cells based on a set of reaction diffusion equations that are solved on a large-scale finite-element code in three dimensions. We have explicitly included the cellular compartments including the cell nucleus, cytoplasm, and gap junctions. The model allows for buffering of free Ca2+, calcium-induced calcium release, and the explicit inclusion of mobile buffers. To make quantitative comparisons to experimental results, we used fluorescence microscopy images of cells to generate an accurate mesh describing cell morphology. We found that Ca2+ wave propagation through the tissue is a function of both initial conditions used to start the wave and various geometrical parameters that affect propagation such as gap junction density and distribution, and the presence of nuclei. The exogenous dyes used in experimental imaging also affect wave propagation.

Linear fuzzy gene network models obtained from microarray data by exhaustive search.

BACKGROUND: Recent technological advances in high-throughput data collection allow for experimental study of increasingly complex systems on the scale of the whole cellular genome and proteome. Gene network models are needed to interpret the resulting large and complex data sets. Rationally designed perturbations (e.g., gene knock-outs) can be used to iteratively refine hypothetical models, suggesting an approach for high-throughput biological system analysis. We introduce an approach to gene network modeling based on a scalable linear variant of fuzzy logic: a framework with greater resolution than Boolean logic models, but which, while still semi-quantitative, does not require the precise parameter measurement needed for chemical kinetics-based modeling. RESULTS: We demonstrated our approach with exhaustive search for fuzzy gene interaction models that best fit transcription measurements by microarray of twelve selected genes regulating the yeast cell cycle. Applying an efficient, universally applicable data normalization and fuzzification scheme, the search converged to a small number of models that individually predict experimental data within an error tolerance. Because only gene transcription levels are used to develop the models, they include both direct and indirect regulation of genes. CONCLUSION: Biological relationships in the best-fitting fuzzy gene network models successfully recover direct and indirect interactions predicted from previous knowledge to result in transcriptional correlation. Fuzzy models fit on one yeast cell cycle data set robustly predict another experimental data set for the same system. Linear fuzzy gene networks and exhaustive rule search are the first steps towards a framework for an integrated modeling and experiment approach to high-throughput "reverse engineering" of complex biological systems.

Retinal pigment epithelial function: a role for CFTR?

In the vertebrate eye, the photoreceptor outer segments and the apical membrane of the retinal pigment epithelium (RPE) are separated by a small extracellular (subretinal) space whose volume and chemical composition varies in the light and dark. Light onset triggers relatively fast (ms) retinal responses and much slower voltage and resistance changes (s to min) at the apical and basolateral membranes of the RPE. Two of these slow RPE responses, the fast oscillation (FO) and the light peak, are measured clinically as part of the electrooculogram (EOG). Both EOG responses are mediated in part by apical and basolateral membranes proteins that form a pathway for the movement of salt and osmotically obliged fluid across the RPE, from retina to choroid. This transport pathway serves to control the volume and chemical composition of the subretinal and choroidal extracellular spaces. In human fetal RPE, we have identified one of these proteins, the cystic fibrosis transmembrane conductance regulator (CFTR) by RT-PCR, immunolocalization, and electrophysiological techniques. Evidence is presented to suggest that the FO component of the EOG is mediated directly or indirectly by CFTR.

Age-dependent changes in breast cancer hormone receptors and oxidant stress markers.

Breast cancer incidence increases with age but this relationship has not been fully explored with regard to expression of estrogen receptor (ER) and ER-inducible genes (PR, pS2, Bcl2, cathepsin D), or the age-dependence of oxidant stress markers that also affect ER-inducible gene expression. In this three-part study, we first correlated age at diagnosis with expression of breast cancer markers ER, PR, pS2, Bcl2, and cathepsin D, quantitated by enzyme immunoassays from a European collective of approximately 3000 cryobanked primary breast cancers and approximately 300 adjacent non-malignant breast tissues. Results were then compared with ER and PR data reported to the SEER registry for 83,541 US cancers diagnosed during 1992-1997. Lastly, a homogeneous subset of 70 ER-positive tumors preselected from the European collective was blindly analyzed for age-specific changes in the DNA-binding content of redox-sensitive transcriprtion factors, AP1 and Sp1, and the oxidant stress-activated protein kinase, phosphorylated(P)-Erk5. Increases in breast tumor ER from patients aged <30 to >80 years mirrored 10-fold lower increases in non-malignant breast tissue ER content up to age 60, rising faster thereafter and reaching a near 25-fold differential between malignant and non-malignant breast tissue by age 80. ER-inducible markers PR, pS2, Bcl2, and cathepsin D were overexpressed in tumors relative to non-malignant breast tissue but, unlike ER, did not increase with patient age. While SEER data demonstrated that the increase in US breast cancer incidence rates after age 50 is confined to ER-positive tumors in patients of all ethnic subsets, these patients also showed a striking increase in the proportion of higher-risk ER-positive/PR-negative breast cancers arising after age 50. Mechanistically essential for ER-inducible PR expression, Sp1 DNA-binding function (but not Sp1 content) was lost with age in ER-positive tumors; and this functional defect correlated with increased tumor content of the oxidant stress marker, P-Erk5. Altogether these findings support two hypotheses: (i) dysregulated ER expression underlies the age-specific increase in breast cancer incidence after age 50; and (ii) oxidative stress and loss of Sp1 DNA-binding may contribute to an increasing incidence in higher-risk ER-positive/PR-negative breast cancers with aging.