Genetics / Screens - Approaches
These projects combine expression strategies with novel assays and screens to identify genes that regulate cell migration. Drosophila genetics provides an invaluable approach to identifying sets of genes that regulate invasive behavior. Expression cloning approaches using cDNA libraries and high throughput image-based technologies will identify gene products that promote migratory behavior. These high throughput approaches, when used with cDNA libraries expressing gene products that are fused to GFP, can also be used to identify proteins that localize in or perturb the structure of migration organelles. They can also be used to screen chemical libraries for new migration-related reagents and screening of migration-related siRNA libraries.
Genetic Approaches
The fruit fly, drosophila melanogaster (D. melanogaster) and the nematode, Caenorhabditis elegans (C. elegans) are two highly tractable genetic organisms which the Consortium is using to identify genes involved in migratory behaviour. In the case of D. melanogaster comparisons are being made between cells in the drosophila overy that exhibit migratory behaviour and those that do not, while in C. elegans the migratory behaviour of distal tip cells is being disrupted using RNAi approaches in an effort to identify key migratory genes.
D. melanogaster
In Drosophila ovary, there are a small group of follicle cells which develop invasive behavior in response to specific transcriptional factors. The egg chamber of the Drosophila ovary contains 16 central germline cells surrounded by a monolayer epithelium of about a thousand follicle cells. At stage 9 of oogenesis, follicle cells at the extreme anterior of the egg chamber differentiate from neighboring epithelial cells and invade the nurse cell cluster, migrating a significant distance to the front of the oocyte. Their migration stops at the border between the oocyte and nurse cells, consequently these migratory cells are referred to as border cells (see figure 1 below). The Montell laboratory has been using this elegant model to identify genes that are required for cells to acquire invasive/migratory behavior (Montell, 1999 & 2003).
Methods have been developed to purify the migratory cells from ovaries dissected from wild-type flies or from mutants lacking a specific transcription factor (see figures 2 & 3 below).
RNA is then isolated from the purified cells, amplified if necessary and then reverse transcribed into cDNA. Labeled cRNA is then transcribed and used to interrogate whole genome micro-arrays. In this way, wild-type and mutant patterns of gene expression are compared, and the downstream targets of key transcription factors required for border cell invasive behavior are being identified.
The functional significance of these genes/proteins in border cell invasion is being investigated, individually and in combination. So far this approach has provided a pool of candidate genes for further study and identified new guidance receptors as well as genes not previously known to function in border cell migration.
Figure 1 - Multiple signals control border cell invasive behavior. Multiple extracellular signals, originating from a variety of different cell types, activate transcriptional programs required for proper border cell migration. The transcriptional targets however are largely unknown. |
Figure 2 - Cell types labeled and purified from dissected ovaries for microarray analysis A) Schematic representation of the GAL4-regulated expression of the chimeric protein composed of the extracellular domain of the mouse CD8 antigen (mCD8), a transmembrane domain and GFP. B) Border cells and centripetal cells, two migratory cell types, are marked for purification in slboGAL4;UASmCD8-GFP flies. C) Only border cells are marked in c522GAL4;UAS-mCD8-GFP flies. D) All follicle cells are labeled in tubulin-GAL4;UAS-mCD8-GFP flies. |
Figure 3. The migratory cells are purified by first dissecting the ovaries (step 1 above) then dissociating the cells (step 2 above) and finally purifying them (step 3) using magnetic beads coated with anti-CD8 antibodies. |
Microarray screen to identify genes upregulated in migrating cells Out of the 900 epithelial follicle cells, only the six to ten border cells and then the ~100 centripetal cells migrate. What are the differences in gene expression between migratory and non-migratory cells? To address this question we have developed methods for separating migratory cells from non-migratory cells from dissected ovaries. Then we isolated RNA from the separated cell populations, generated cDNA and then labeled cRNA from these cells and hybridized Affymetrix whole genome microarrays in order to compare patterns of gene expression. These experiments have been completed.
return to top
Microarray screen to identify downstream targets of transcription factors required for cell migration From our earlier genetic approaches we found that several transcription factors are required specifically for border cell migration however we do not know many of the critical downstream targets of these factors. In order to identify the targets, we purified migratory cells from wild-type ovaries or from ovaries lacking the function of a specific transcription factor required for migration. We then hybridized Affymetrix GeneChips with labeled cRNA made from these cells and compared wild-type and mutant gene expression patterns. To date we have completed several different microarray analyses, comparing patterns of gene expression in wild-type cells and several different mutants. We are in the process of annotating the data for inclusion on the CMC website.
return to top
High Throughput Screening in Drosophila S2 Cells Drosophila constitutes an ideal model system for RNAi screening studies for gene discovery, given its fully sequenced genome and relatively low level of gene redundancy. Using such an approach in the Vale laboratory, the two experimental strengths of Drosophila as a model organism for the study of cell shape and motility (the ability to carry out whole genome RNAi screens in cultured S2 cells and forward genetic screening in vivo) are being exploited. Additionally, Drosophila S2 cells take up dsRNA molecules very robustly and without triggering an interferon response. Moreover, genes identified based on RNAi screening studies can easily be depleted in specific cell types in transgenic flies to analyze their effects on cell migration in vivo. Additionally, the Vale lab proposes to employ newly developed protein-trapping methods to identify proteins that are expressed asymmetrically within migrating cells in vivo
.
Preliminary data: The Vale lab has been using Drosophila S2 cells in tissue culture as a model system for studying cellular morphology. They have found that S2 cells spread and form actin rich lamellae and a dense microtubule network when plated on Concanavalin A. They have used this system to perform an RNAi screen of ~100 known actin-associated genes (Rogers et al., 2003). This study confirmed the activities of many proteins, but also revealed unexpected activities for several proteins (e.g. cyclase-associated protein and the Scar inhibitory complex). They have extended this to screen over 7200 Drosophila genes (the subset of genes that have homologues in mammals or C. elegans) to identify genes that alter overall cell morphology or specific features of the actin and microtubule cytoskeletons. In contrast to screens with S2 cells from other laboratories, the Vale lab examined cell spreading and has performed high resolution imaging with a 63X, n.a. 1.4 objective lens. The screen identified many of the genes known to regulate the actin cytoskeleton (e.g. all subunits of Arp2/3, Scar, upstream regulators such as Dock180, rac, and downstream effectors of actin such as profilin, cofilin, etc). However, in addition to these known genes, many novel or unexpected genes were found to be important for actin-based cell morphology and the Vale lab is working to characterize these proteins further. Figure 1 shows typical results seen in the RNAi morphology screen
Figure 1: Representative images of S2 cells after dsRNA treatment. The cells in the first panel show wildtype morphology after RNAi. The second and third panels show severe actin (red) and microtubule (green) morphology defects after the indicated genes were depleted. CG15415 is a completely novel gene, and CG5179 is CDK9, a kinase that is thought to regulate transcription.
C. elegans
Gonad morphogenesis in C. elegans depends on the migration of two specialized leader cells called the distal tip cells (DTCs). As larval development proceeds, the DTCs migrate out from the gonad primordium along the ventral extracellular matrix (ECM), one DTC migrates toward the anterior and one toward the posterior of the animal. In response to specific signals, each DTC then turns to the dorsal side and migrates back toward the middle, resulting in a mirror image U-shaped gonad. The ECM provides a substrate for migration of the DTCs and presents a variety of directional cues that in conjunction with other factors control the timing and positions of turning and stopping of these migrating cells. The shape of each gonad arm reflects the migratory path taken by the DTC during larval development. Therefore, if the DTC fails to migrate or follows an aberrant path, malformation of the gonad arm will result. Such defects are easily scored, and provide the basis for our screen for genes involved in cell migration.
RNAi screens have been established in C. elegans, to identify genes required for cell migration including genes that function downstream of integrins. In RNAi, double-stranded RNA (dsRNA) introduced into larvae or adults activates an enzyme pathway that eliminates RNAs homologous to the dsRNA (Fire et al., 1998). RNA phenocopies loss-of-function or reduction-of-function phenotypes for the gene in question. Double-stranded RNAs are introduced by feeding nematodes bacteria that are expressing the dsRNA under control of an IPTG-inducible promotor. The feeding method in particular lends itself to large scale screens and allows analysis of phenotypes at particular stages or over multiple generations.
Using a library of approximately 80% of C. elegans open reading frames, each contained in a vector optimized for bacterial mediated RNAi (Kamath et al., 2003), key components involved in distal tip cell (DTC) migration are being identified. RNAi is combined with simple light microscopy analusis in which adult hermaphrodites are examined for clear areas within the body cavity (Figure 1, A B). The clear appearance results from displacement of the intestine due to inappropriate turns of the gonad arms, gonad distension, or other morphological defects. For further analysis, animals showing clear areas are transferred to slides and examined by Nomarski microscopy. The path taken by the migrating DTC can be inferred from the resulting shape of the gonad tube (Figure 1 C,D). Screening is carried out on 24-well plates allowing analysis of about 150 clones at a time, a small percentage also being examined at higher magnification. Initial studies using this system (Cram et al., 2003) prove proof of principle, that these screens can identify a set of genes required for DTC migartion and further can define a relationship between integrins and cell migration in vivo.
RNAi screen for distal tip cell migration defects The is a genome-wide screen based on RNA interference (RNAi) being conducted to identify genes that are involved in distal tip cell (DTC) migration. In RNAi, double-stranded RNA (dsRNA) introduced into larvae or adults activates an enzymatic pathway that eliminates RNAs homologous to the dsRNA (Fire et al. 1998). The DTCs are specialized leader cells that direct the formation of the gonad. The path taken by the migrating DTC can be inferred from the resulting shape of the gonad tube, and defects in DTC migration lead to characteristic clear areas within the body cavity that are easily scored by light microscopy. We have obtained a library of 16,750 C. elegans open reading frames, each contained in a vector optimized for bacterial mediated RNAi (Kamath et al. 2003) (http://www.wormbase.org). Screening is carried out on multi-well plates allowing us to analyze about 150 clones at a time. A small percentage of these are examined at higher magnification.
Significant progress screening the RNAi library is being made. Some of the isolated genes include talin (unc-35)(Cram et al. 2003) , various tubulins (e.g., tba-1, tba-2), Rac (ced-10), transcription factors such as daughterless (hlh-2) and hlh-12, and the plakin shortstop (vab-10). Based on progress so far, approximately 100 genes required for correct migration of DTCs are expected. Genes identified in this screen will be grouped into categories by molecular function and expression pattern. Several genes will be selected for more detailed analysis using RNAi and available mutant alleles. Selection of candidates for further study will be coordinated with other projects supported by the Migration Consortium.
Determine cell autonomy of 99 DTC migration genes Distal tip cell (DTC) migration depends on genes expressed in the DTCs themselves but also is influenced by the body wall muscle cells that produce a basement membrane migration path. To determine the tissue requirements for the genes in our cell migration gene panel (Cram et al., 2006), we are planning to use DTC-specific RNAi to knockdown each gene in the DTCs only. In C. elegans, rde-1 is required for RNA interference in C. elegans but rde-1(ne219) mutants do not have other defects. Tissue-specific expression of rde-1 in the mutant background can be used to rescue RNAi in specific tissues. We have generated a C. elegans transgenic strain (rde-1(ne219); lag-2p::rde-1::gfp) for this purpose. Under control of the lag-2 promoter, rde-1 is expressed primarily in the DTCs. Preliminary RNAi experiments using this strain show the efficacy of this strategy. After further characterization of the strain, we will begin to screen the 99 migration genes for DTC-specific effects.
Gene and Chemical Screens
This approach encompases several different project. Using a range of different cell types in combination with a number of different gene (cDNA or siRNA) or chemical libraries, genes and compounds are being identified, that (i) alter migratory behavior , (ii) perturb focal adhesions and actin cytoskeleton or localize to migration related structures.
Alteration of Migratory Behavior
A transwell migration assay is used to determine altered migratory behavior in one series of studies, while an automated high throughput microscopy approach is used in conjunction with a phagokinetic assay to determines migratory behaviour in another series of experiments.
Transwell migration assay. This approach involves functional screening of both biased and unbiased cDNA libraries expressed in the MCF10A mammary epithelial cell. Under normal conditions these cells are poorly migratory, but sensitization of the cells by overexpressing c-ErbB2 or stimulation with epidermal growth factor (EGF), provides conditions for evaluating enhanced migration or inhibition of EGF-stimulated migration. Each individual clone is expressed in the MCF10A cells via retroviral infection and evaluated for its effect on the migration of these cells through a microcellulose filter (see figure below). For more information see Seton-Rogers et al., 2004 and Witt et al., 2006.
| Migration screen with BC1000 cDNA library | |
 |
|
| Epithelial cells infected with retroviruses encoding cDNAs of interest | Transwell migration assay |
Assay Details: Assays were performed in MCF-10As expressing the inducibly activated ErbB2 receptor (10A.B2) that contains the extracellular and transmembrane domain of p75low affinity NGF receptor, the cytoplasmic domain of ErbB2, an HA tag and FKBP (Muthuswamy et al., 2001). Activation of the chimeric ErbB2 receptor using the dimerizing ligand AP1510 in 10A.B2 cells in the absence of EGF induces proliferation and multi-acini formation, but not cell migration (Seton-Rogers et al., 2004).
These cells were infected with VSV G pseudotyped retroviruses expressing a subset of genes in the (BC1000 collection). 10A.B2 cells expressing genes from the BC1000 collection were starved overnight in assay media (MCF-10A media containing no EGF and only 1% serum) (Debnath et al., 2003). 1x105 cells were added to the top chambers of 24-well transwell plates (BD, 8 µm pore size), and assay media with 500nM AP1510 (ErbB2 dimerizer) was added to the bottom chambers. After overnight incubation, top (non-migrated) cells were removed, and bottom (migrated) cells were fixed and stained with crystal violet. Genes were scored as a hit if there was an observed increase the number of migrating cells compared to vector alone. Fold-change compared to vector alone was assessed visually. Experiments were repeated a minimum of three times for genes that scored as a hit, and considered a validated hit if a migratory phenotype was observed in at least two of the three experiments (or >50% if more than 3 repeats were done). Validated hits were also tested under similar conditions without the addition of 500nM AP1510.
MCF7 cDNA library screen; A cDNA library from the breast cancer cell line MCF7 was constructed in a new retroviral expression system (pEYK) constructed by collaborators at MIT (George Daley and Eugene Koh). The system was designed to facilitate easy and efficient recovery of a clone of interest from the viral construct which is integrated into the host DNA. This MCF7 retroviral expression library was used to infect the MCF10A mammary epithelial cell and select for infected cells that had an enhanced capacity to migrate. In addition, to sensitize the system, MCF10A cells were designed to co-express either the ErbB2 (Her2) or the Colony Stimulating Factor (CSF)-1 receptor tyrosine kinase. After several rounds of enrichment, clones are recovered and sequenced. For more information see Gunawardane et al., 2005.
| Migration screens with MCF7 cDNA library |
 |
Phagokinetic tracks (PKT) assay. This approach involves high throughput analysis of cell migration in 2-dimensions. This assay is performed in both the Geiger and Brugge laboratories and the protocols are very slightly different.
In the Geiger laboratory the screen is conducted in 96 well plates (glass bottom), coated with 10 µg/ml fibronectin and a monolayer of polystyrene microspheres (0.34 microns diameter). Conditions have been defined so that a highly uniform bead monolayer, can be remodeled by migrating cells and retain clearly visible migration tracks. Software has been developed for analysis of the PKT, including segmentation and automatic recognition of tracks, merger of “marginal” tracks in neighboring images and calculation of PKT parameters, including track net area (normalized per cell), persistence, branching, “dispersion” and width.
Phagokinetic tracks |
||
 |
||
|
Low Migration
|
High Migration | |
return to top
High throughput wound healing assay. To screen large numbers of siRNAs for their ability to modulate cell migration, we modified the classic scratch-wound assay by developing a robotic-driven pin to deliver a precise 1.5 x 4 mm scratch in confluent cell monolayers in 96-well plates. MCF-10A cells were transfected with siRNAs 16 hours post plating and the confluent monolayer was wounded 56 hours later. The extent of wound healing was evaluated after 12 hours when mock-transfected cells, which migrate as an epithelial cell sheet, had closed the wound 50-60%; this enabled detection of accelerated and impaired migration. To quantitate cell migration, the cells were labeled at the end of the migration period with rhodamine-conjugated phalloidin and imaged using an Applied Precision CellWorx automated fluorescent microscope. The extent of motility, termed the ‘area score’, was quantified using an algorithm we developed that identified and measured the unfilled area within a uniform region encompassing the wound. In addition, the wells were visually evaluated to provide a secondary assessment of motility. The impact of the siRNAs on proliferation and general metabolism was measured prior to wounding using Alamar Blue.
Morphological perturbation of focal adhesions & actin cytoskeleton
This screening approach utilizes a high throughput automated microscopy system to identify chemical compounds or siRNAs that specifically perturb focal adhesions or cytoskeletal structures. Reporter cell lines, uniformly expressing YFP-tagged components of the adhesion sites or the actin cytoskeleton (mainly paxillin-YFP) are plated in 384-well, thin bottom plates and incubated with the compound libraries (consisting of extracts from natrual sources or commercial synthetic compound libraries) or transfected with siRNA libraries. Treated cells are then fixed and examined with the screening automated microscope. The resulting fluorescence patterns are analyzed by pattern- recognition and analysis algorithms.
Localization to migratory structures
In one approach the cDNA libraries to be screened are cloned into an expression vector encoding GFP, producing a GFP-fusion library which then can be screened for gene products that localize to focal adhesions or cytoskeletal structures. A high throughput microscope-based screening system is used to screen individual cDNAs expressed in appropriate cells, which have highly organized adhesions (e.g.,REF52) , in individual wells.
This strategy has already been used to identify new gene products in focal adhesions and cytoskeletal structures. Some of these encode known genes while others appear to be novel and merit further investigation.
Alterations in morphology are assayed by analyzing individual cells using computer-based pattern recognition and processing algorithms.
return to top
Libraries & Cells
The different cDNA, siRNA and chemical libraries as well as the various cell types being used in the gene and chemical screens are detailed here.
cDNA Libraries
Breast Cancer 1000 cDNA Library - In collaboration with researchers at the Harvard Institute of Proteomics, a collection of 1000 genes associated with breast cancer have been assembled and many have been cloned into a retroviral mammalian expression vector. The genes in the collection were identified through a combination of literature review, expression array data, and text-searching of known databases. (http://www.hip.harvard.edu/research/breast_cancer/index.htm)
MCF7 cDNA Library - A cDNA library from the breast cancer cell line MCF7 was constructed in a new retroviral expression system (pEYK) constructed by collaborators at MIT (George Daley and Eugene Koh).
MDA-MB-231 cDNA Library - A library containing cDNA obtained from highly motile MDA-MB-231 cells in pEYK3.1 retroviral vector. Prepared by the Brugge laboratory.
YFP-tagged arabidopsis cDNAs - A cDNA library was prepared from two weeks old Arabidopsis Columbia seedlings in pLPCX expression vector upstream to sequence encoding for YFP . Infection of 5000 clones into rat fibroblasts yielded several different localizations of the YFP-chimeras such as into nuclei, nucleoli, mitochondria, actin stress fibers and focal contacts.
GFP-tagged HT1080 cDNAs - The library with oligo-dT-primed cDNAs was prepared from human fibrosarcoma HT1080 cells and was inserted into a GFP expression vector. The resulting gene products were chimeras with the cDNA fused to the C terminus of the GFP. Each cDNA clone was transfected transiently into REF52 cells growing in glass-bottomed, 96-well plates and was screened by fluorescence microscopy for gene products that localize to specific subcellular structures.
return to top
siRNA Libraries
Three sets of Dharmacon SMART pool RNAs are being used (i) kinases, (ii) phosphatases and (iii) a custom set targeting genes predicted to be involved in regulating cell adhesion and migration.
More details on these will be posted as they become available. The sequence for all positive siRNAs will also be provided once available.
return to top
Chemical Libraries
A natural products library and two chemical compound libraries are being used the details of which are provided below.
Tel Aviv Natural Products Library - www.tau.ac.il/~nchts/
This is a library of extracts prepared from different natural sources (marine invertebrates, plants, fungi etc) was obtained from the HTS center of Tel Aviv University (Pros. M. Ilan and J. Kashman) and analyzed for effects on focal adhesions. About 2000 extracts were screened and positive extracts were validated and examined in greater detail. Noteworthy effects were obtained with about 20 extracts, and 4 of those are currently further purified and studied in depth. Images of all active extracts are available on the CMC website, and all interested researchers can contact Micha Ilan (Milan@post.tau.ac.il).
Chem Bridge & Chemical Diversity Libraries
Two compound libraries were purchased and partially tested for focal adhesion perturbing molecules. The libraries consist of well-defined molecules (library sizes are 10,000 and 55,000 compounds). So far we have screened close to 10,000 compounds and discovered ~100 compounds with potentially interesting effects. All images are displayed on the CMC website. A selection of those was validated and is currently investigated in detail. Researchers interested in pursuing the characterization of specific molecules should contact Benny Geiger (benny. geiger@weizmann.ac.il) .
return to top
Cell Types
HeLa-JW paxillin YFP-reporter
Cervical cancer cells stably infected with plasmid encoding paxillin fused to yellow fluorescent protein (YFP) .
MCF-10A
MCF-10A is a spontaneously immortalized, but nontransformed human mammary epithelial cell lline derived from the breast tissue of a 36-year old patient with fibrocystic changes (Soule et al., 1990). Additional information on conditions for growth and handling of these cells can be found in this reference (Debnath et al., 2003).
MCF-7
Human breast adenocarcinoma; Non metastatic, low motile cells, which form typical epithelial colonies.
MDA-MB-231
Human breast adenocarcinoma cell line(Cailleau et al., 1974). Highly migratory in vitro and tumorigenic in mice - forms poorly differentiated invasive, adenocarcinomas that metastasize to several organs.
REF-52 (with or without YFP reporter)
Rat embryo fibroblasts (REF52).
Rat embryonic fibroblasts (REF-52) with YFP reporter - stably infected with plasmid encoding paxillin fused to yellow fluorescent protein (YFP) . Cells with high YFP fluorescence were sorted by FACS and a relatively homogenous cell line was obtained by single cell cloning.
 |
return to top
Text References
Cailleau R, Young R, Olive M, Reeves WJ Jr. 1974. Breast tumor cell lines from pleural effusions. J Natl Cancer Inst. 1974 Sep;53(3):661-74. PubMed
Cram EJ, Clark SG, Schwarzbauer JE. 2003. Talin loss-of-function uncovers roles in cell contractility and migration in C. elegans. J Cell Sci. 116(Pt 19):3871-8. PubMed
Cram EJ, Shang H, Schwarzbauer JE. A systematic RNA interference screen reveals a cell migration gene network in C. elegans. 2006; 119(Pt 23):4811-8. PubMed | CMC Update article.
Debnath J, Muthuswamy SK, Brugge JS. 2003. Morphogenesis and oncogenesis Of MCF-10A mammary epithelial acini grown In three-dimensional basement membrane cultures. Methods, 30 (3)256-268. PubMed
Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, Mello CC. 1998. Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature. 391(6669):806-11. PubMed.
Gunawardane RN, Sgroi DC, Wrobel CN, Koh E, Daley GQ, Brugge JS. Novel role for PDEF in epithelial cell migration and invasion. Cancer Res. 2005 Dec 15;65(24):11572-80. PubMed
Hynes RO. 2002. Integrins. Bidirectional, allosteric signaling machines. Cell 110: 673-679. PubMed
Kamath RS, Fraser AG, Dong Y, Poulin G, Durbin R, Gotta M, Kanapin A, Le Bot N, Moreno S, Sohrmann M, Welchman DP, Zipperlen P, Ahringer J. 2003. Systematic functional analysis of the Caenorhabditis elegans genome using RNAi. Nature. 2003 421(6920):231-7. PubMed.
Koh EY, Chen T, Daley GQ. 2002. Novel retroviral vectors to facilitate expression screens in mammalian cells. Nucleic Acids Res. 30(24):e142. PubMed
Lee M, Cram EJ, Shen B, Schwarzbauer JE. 2001. Roles for beta(pat-3) integrins in development and function of Caenorhabditis elegans muscles and gonads. J. Biol. Chem. 276(39): 36404-10. PubMed
Montell DJ. 1999. The genetics of cell migration in Drosophila melanogaster and Caenorhabditis elegans development. Development, 126:3035-3046. PubMed
Montell DJ. 2003. Border-cell migration: the race is on. Nat Rev Mol Cell Biol. 4(1):13-24. PubMed
Muthuswamy SK, Li D, Lelievre S, Bissell MJ, Brugge JS. 2001. ErbB2, but not ErbB1, reinitiates proliferation and induces luminal repopulation in epithelial acini. Nat Cell Biol. 3(9):785-92. PubMed
Seton-Rogers SE, Lu Y, Hines LM, Koundinya M, LaBaer J, Muthuswamy SK, Brugge JS. 2004. Cooperation of the ErbB2 receptor and transforming growth factor beta in induction of migration and invasion in mammary epithelial cells. Proc Natl Acad Sci U S A. 2004 Feb 3;101(5):1257-62. PubMed
Soule HD, Maloney TM, Wolman SR, Peterson WD Jr, Brenz R, McGrath CM, Russo J, Pauley RJ, Jones RF, Brooks SC. 1990. Isolation and characterization of a spontaneously immortalized human breast epithelial cell line, MCF-10. Cancer Res. 50(18):6075-86. PubMed
Witt A, Hines L. Collins N, Hu Y, Gunawardane R, Moriera D, Raphael J, Jepson D, Khoundinya M, Rolf A, Taron B, Isakoff SJ, Brugge JS, LaBaer J. A Functional Proteomics Approach to Investigate the Biological Activities of cDNAs Implicated in Breast Cancer, J. Proteomics Research, in press.