Conceived and designed the experiments: HI YS JSB. Performed the experiments: HI YS FF NI ZW YL. Analyzed the data: HI YS LS ZW LZ JY ALR KP GM WCH JSB. Contributed reagents/materials/analysis tools: HI YS LS YL CBE SN KP. Wrote the paper: HI LS JSB.
This work was partially funded by a grant from Sanofi-Aventis to Joan Brugge. We declare that there are no restrictions on sharing any of the data or reagents used in this study in accordance with the policies of PLoS ONE.
Proteins that are required for anchorage-independent survival of tumor cells represent attractive targets for therapeutic intervention since this property is believed to be critical for survival of tumor cells displaced from their natural niches. Anchorage-independent survival is induced by growth factor receptor hyperactivation in many cell types. We aimed to identify molecules that critically regulate IGF-1-induced anchorage-independent survival.
We conducted a high-throughput siRNA screen and identified PTK6 as a critical component of IGF-1 receptor (IGF-1R)-induced anchorage-independent survival of mammary epithelial cells. PTK6 downregulation induces apoptosis of breast and ovarian cancer cells deprived of matrix attachment, whereas its overexpression enhances survival. Reverse-phase protein arrays and subsequent analyses revealed that PTK6 forms a complex with IGF-1R and the adaptor protein IRS-1, and modulates anchorage-independent survival by regulating IGF-1R expression and phosphorylation. PTK6 is highly expressed not only in the previously reported Her2+ breast cancer subtype, but also in high grade ER+, Luminal B tumors and high expression is associated with adverse outcomes.
These findings highlight PTK6 as a critical regulator of anchorage-independent survival of breast and ovarian tumor cells via modulation of IGF-1 receptor signaling, thus supporting PTK6 as a potential therapeutic target for multiple tumor types. The combined genomic and proteomic approaches in this report provide an effective strategy for identifying oncogenes and their mechanism of action.
Adhesion to extracellular matrix (ECM) provides epithelial cells with critical cues about their environment that are required for their proliferation, survival and tissue organization. Loss of attachment to matrix compromises viability of normal epithelial cells through a variety of mechanisms that help preserve tissue homeostasis and prevent aberrant growth (reviewed in
Tumor cells adopt several different strategies to evade anoikis including: (1) activation of survival pathways such as those regulated by Erk/MAPK and Akt through oncogenic mutations or constitutive growth factor receptor activation; (2) modulation of expression or activity of anti-apoptotic and pro-apoptotic proteins including Bcl2 family members, and (3) altered expression and engagement of integrins by basement membrane proteins produced via autocrine mechanisms (reviewed in
As a complement to these gain-of-function screens, loss-of-function screens also provide insight into mechanisms that are necessary for anoikis suppression and identify potential targets for therapeutic intervention. Screens utilizing small molecule inhibitors have previously been reported
IGF-1R has been shown to be expressed in the majority of human breast cancers with evidence of sporadic amplification in a small proportion of these cases
In contrast to non-transformed MCF-10A mammary epithelial cells, MCF-10A cells overexpressing IGF-1 receptor (IGF-1R cells) exhibit enhanced survival in suspension cultures with IGF-1 stimulation (
A) IGF-1R hyperstimulation enhances anchorage-independent survival of MCF-10A cells. MCF-10A cells overexpressing vector control or IGF-1R (IGF-1R) were cultured on adherent plates or in suspension cultures for 48 hours. The cells were collected and cell death was assessed by measuring cytoplasmic histone and DNA content in triplicate wells (Roche Diagnostics). Higher values indicate more death. The fold change is the average of three independent experiments. Error bars indicate standard deviation. Western blot analysis shows relative IGF-1R levels in the cell lines. B) Candidate siRNAs that are critical for IGF-1R-induced anchorage-independent survival were identified by a loss-of-function screen. In the primary screen, a library of siRNA oligonucleotides targeting all human kinases was screened using the IGF1R overexpressing MCF10A line. Alamar blue reduction was assayed. Positives from the primary screen were counter-screened in MCF10A cells cultured in 2% Matrigel in suspension. The screen positives that showed a differential effect in IGF1R-overexpressing cells compared to the MCF10A cells in Matrigel were selected for validation using additional sequences.
PTK6 was identified as a “high-confidence” candidate gene because multiple siRNA oligonucleotides caused a significant, preferential inhibition of IGF-1R-driven anchorage-independent viability when compared to parental MCF-10A cells. PTK6 was also considered an attractive candidate for further study because it is highly expressed in multiple tumor types, including breast and ovarian tumors
A) An shRNA vector targeting PTK6 reverses IGF-1R-induced anchorage-independent survival. (Left) Western analysis confirmed downregulation of PTK6 expression and demonstrated increased PARP cleavage with PTK6 downregulation in IGF-1R cells in suspension cultures. (Middle) IGF-1R cells expressing control or PTK6 shRNA vector were cultured as attached cells or in suspension cultures in the presence of IGF-1. Cell death was assessed after 48 hours as in
To more specifically determine whether PTK6 downregulation induces cell death of IGF-1R cells in suspension cultures, we utilized Cell Death ELISA assays and monitored cleavage of PARP, a caspase target. Downregulating PTK6 by either transfecting siRNA oligonucleotides or expressing an shRNA vector enhanced death of IGF-1R cells in suspension with enhanced detection of cleaved PARP product (
Acquisition of anchorage-independent survival is believed to play a role in the abnormal accumulation of cells in the lumen of terminal breast acinar structures, a property that is commonly observed in human breast lesions like ductal carcinoma in situ (DCIS). Consistent with this, we have found that filling of the matrix-free lumen of mammary acinar structures in 3D culture models requires induction of both hyperproliferation and anti-apoptotic activity
We and others have previously reported that IGF-1R hyperstimulation of breast epithelial cells results in the formation of abnormally large structures with filled lumen due to hyperproliferation and suppression of death of the centrally localized cells in the presumptive luminal space
We then examined the effect of PTK6 downregulation on IGF-1R-induced luminal filling in 3D cultures. Compared to acini expressing vector control, PTK6 downregulation resulted in a significantly higher percentage of structures with clear lumen (
To complement the loss-of-function studies, we examined the effect of overexpressing PTK6 in both MCF-10A cells and immortalized human mammary epithelial cells (HMLE). Overexpression of PTK6 in parental MCF-10A cells using either a FLAG-tagged wild type PTK6 (F-PTK6) or myristylated, membrane targeted PTK6 (MF-PTK6), was sufficient for a modest, but significant, suppression of cell death upon detachment from matrix (
A) Overexpression of PTK6 enhances anchorage-independent survival. MCF-10A cells overexpressing vector control or FLAG-tagged PTK6, with or without a myristylation sequence (MF-PTK6 or F-PTK6, respectively), were cultured in suspension for 48 hours. Cell death was assessed as in
Overexpression of PTK6 alone (MF-PTK6 or F-PTK6) in HMLE was not sufficient to significantly promote colony formation in soft agar, another measure of anchorage-independent survival (
We sought to determine whether PTK6 plays a critical role in modulating anchorage-independent survival of breast and ovarian cancer cells in which PTK6 is highly expressed, particularly those which co-express IGF-1R. Examination of a panel of cell lines revealed that PTK6 was abundantly expressed in several Her2+ breast cancer lines, as previously reported (
A) PTK6 downregulation inhibits survival of Her2+ cells. The effect of PTK6 downregulation on anchorage-independent survival of MCF-10A cells overexpressing ErbB2 was assessed. Cell death was assessed as in
Downregulation of PTK6 also prevented IGF-1 stimulated anoikis resistance of DOV-13 ovarian cancer cells, which co-express PTK6 and IGF-1R (
To begin to elucidate the mechanisms by which PTK6 modulates anchorage-independent survival, we utilized reverse-phase protein arrays (RPPA) to identify signaling molecules that are altered by downregulation or overexpression of PTK6 (
A) Reverse-phase protein arrays (RPPA) reveal changes in IGF-1-stimulated signaling with PTK6 downregulation. RPPA analysis was performed on lysates of IGF-1R cells overexpressing Bcl2 (IGF-1R/Bcl2) that were transfected with control or PTK6 siRNA oligonucleotides in suspension cultures containing IGF-1 (100ng/ml). The heatmap represents values normalized to luciferase control from transfections performed in triplicate. B) shRNA-mediated downregulation of PTK6 in MCF-7 cells decreases ligand-stimulated IGF-1R phosphorylation. MCF-7 cells expressing control or PTK6 shRNA were stimulated in suspension cultures for the indicated times with 100ng/ml IGF-1. Lysates were probed with the indicated antisera. C) DOV-13 cells cultured on adherent plates were stimulated with 100ng/ml IGF-1 for 20 minutes. Lysates were probed with the indicated antisera. D) PTK6 downregulation inhibits IGF-1- stimulated activation of Akt and Erk/MAPK. Lysates of MCF-7 cells stimulated with the indicated concentrations of IGF-1 for 60 minutes (left panel) or 15 minutes (right panel) in suspension cultures were probed with the indicated antisera.
A) PTK6 overexpression increases phosphorylation of IGF-1R. (Left) MCF-10A cells overexpressing a constitutively active (CA) or kinase-inactive (KD) PTK6 were grown in suspension cultures for 48 hours in the presence of IGF-1 (100ng/ml). Lysates were prepared and probed using RPPA. The values represent the average normalized expression values derived from triplicate samples. Error bars indicate standard deviation. (Right) Lysates prepared from suspension cultures of MCF-10A cells overexpressing vector control, MF-PTK6 or kinase-inactive MF-PTK6 (KD-MFPTK6) were probed with the indicated antisera. B) Lysates prepared from suspension cultures of MCF-10A cells overexpressing vector control, MF-PTK6 or wild type Akt in the presence of IGF-1 (100ng/ml) were probed with the indicated antisera. C) PTK6 co-precipitates with IGF-1R and IRS-1 adapter protein. Lysates of MCF-10A cells overexpressing IGF-1R and wild type PTK6 were immunoprecipitated with the indicated antisera and immunoblotted with anti-PTK6 antibody. D) Anchorage-independent survival stimulated by PTK6 overexpression is suppressed by IGF-1R kinase inhibition. MCF-10A cells overexpressing vector control or MF-PTK6 were grown in suspension cultures containing IGF-1 (100ng/ml) for 48 hours in the absence or presence of BMS-536924 at the indicated concentrations. Cell death was assessed as described in
The effect of overexpressing PTK6 on IGF-1R phosphorylation was also examined (
Previous studies have shown that PTK6 is able to form complexes with EGF receptor family members and PTK6 couples these receptors to downstream signaling pathways
To determine whether IGF-1R kinase activity is critical for the suppression of anoikis induced by PTK6 overexpression, we assessed the effect of BMS-536924, an IGF-1R/insulin receptor inhibitor, on the anchorage-independent survival of MCF-10A cells overexpressing MF-PTK6. Cells were cultured in suspension in the presence of sub-saturating doses of the inhibitor, which reversed the enhanced survival induced by PTK6 (
PTK6 may promote human cancer growth by modulating signaling via growth factors receptors, such as IGF-1R and ErbB2. PTK6 has been reported to be expressed in multiple tumor types, including breast and ovarian cancer
A) (Top) SNP array analyses of 93 breast tumors demonstrates PTK6 copy number gain in a subet of tumors.The x-axis represents individual breast tumor specimens; the y-axis represents copy number. The red line indicates copy number = 3 (Bottom) Correlation analysis with other amplified genes is shown. The table lists the number of tumors with PTK6 copy number gain (copy number≥3.0), the Pearson product-moment correlation for each pair of variables and the associated P-values. B) FISH analysis demonstrates PTK6 copy number gain. Two human breast tumors were selected for FISH analyses based on high PTK6 transcript expression levels in microarray analyses. FISH analysis was performed using a probe specific for PTK6 (green) or a control chromosome 20 centromeric probe (pink). C) PTK6 mRNA is differentially expressed in human breast cancer subtypes. Publicly available datasets were used to compare levels of PTK6 mRNA in breast cancer subtypes (basal, HER2/ERBB2+, Luminal A, Luminal B) and normal samples. P-values were calculated using ANOVA in JMP 7.0 software. D) PTK6 mRNA level is a predictor of prognosis in human breast cancer. Kaplan-Meier curves depict the probability of breast cancer recurrence based on relative levels of PTK6 transcript expression. Probabilities for the entire cohort (top) as well as the ER+ subgroup (bottom) were determined. The Cox regression test evaluates the association of PTK6 level with patient outcome treating the level of PTK6 as a continuous variable. The log-rank test evaluates whether there are significant differences between any of the three groups. Data sets from Van't Veer et al. (Van't Veer et al., 2002) and Wang et al.(Wang et al., 2005) were analyzed.
We examined the relative expression of PTK6 mRNA in different breast tumor subtypes (
Elevated levels of PTK6 mRNA expression are associated with prognostic significance. In two large breast cancer patient cohorts with long-term follow up, tumor specimens were stratified based on PTK6 expression levels. Patients with tumors with the highest expression levels of PTK6 mRNA transcript in each cohort were less likely to be recurrence-free over long-term follow up (
In contrast to normal epithelial cells, most cancer cells have acquired the ability to grow in an anchorage-independent manner. In this report, we identified PTK6 as a gene that is critically involved in IGF-1R-stimulated anchorage-independent survival using a highly validated siRNA screen. PTK6 was also found to be required for anchorage independence of breast and ovarian cancer cells. These findings were supported by gain-of-function studies showing that PTK6 is an effective “collaborative oncogene”, efficiently enhancing transformation in combination with other activated oncogenes. Insights into the mechanisms whereby PTK6 induces anoikis resistance were provided by high throughput protein arrays and subsequent analyses, which demonstrated that loss of PTK6 had broad effects on the major signaling pathways activated by IGF-1R as well as IGF-1R itself. Subsequent studies confirmed that PTK6 regulates the expression and the activity of IGF-1R. Effects of PTK6 on anchorage-independence could contribute to the poor outcomes associated with high PTK6 transcript levels that we found in multiple cohorts of breast cancer patients.
Our studies indicate that PTK6 regulates anchorage-independent survival through suppression of caspase-mediated apoptosis, as downregulation of PTK6 increased PARP cleavage and resulted in cell death that was inhibited by ZVAD-fmk. Interestingly, Harvey et al. recently reported that downregulation of PTK6 in the matrix-deprived T47-D breast cancer cells induces autophagy, but not apoptosis
RPPA analyses in this study demonstrated that AMPK and its target protein ACC were activated by PTK6 downregulation, indicating that loss of PTK6 also causes metabolic stress. In addition, although cell death induced by PTK6 downregulation was inhibited by ZVAD-fmk, the metabolic impairment was not (data not shown). While autophagy can be an adaptive survival mechanism for cells under stress it can also progress to cell death (reviewed in
Overexpression of PTK6 enhanced anchorage-independent survival. Although PTK6 is distantly related to the Src family of tyrosine kinases with an SH3, SH2 and catalytic domain, it lacks a native myristylation signal shared by most members of this class
While the ability of PTK6 to enhance signaling downstream of EGF receptor family kinases is well established
Whether the effects of PTK6 on IGF-1R function involve direct interactions with the receptor remain to be addressed. The ability of PTK6 to regulate the IGF-1 receptor does not appear to be dependent on a secreted factor as conditioned media from cells in which PTK6 expression was modulated failed to induce any change in IGF-1R (data not shown). The ability of PTK6 to co-precipitate with both the IGF-1 receptor and IRS-1 suggests that PTK6 is able to associate with the receptor complex. Interestingly, v-Src has been reported to directly phosphorylate the beta chain of the IGF-1 receptor at the corresponding autophosphorylation sites in rat fibroblasts with subsequent activation of IGF-1R kinase, and IGF-1R is required for v-Src transforming activity
Although both gain- and loss-of-function approaches used in this report highlighted a positive role for PTK6 in anchorage-independent survival, other studies utilizing mouse intestinal cells and rat fibroblasts suggest that PTK6 may be required for cell death triggered by specific stimuli such as DNA damage
PTK6 is expressed in multiple human tumor types, including breast and ovarian tumors
In our study we analyzed the prognostic significance of elevated PTK6 transcript expression in multiple large cohorts of patients with breast cancer. Our finding that elevated expression of PTK6 transcript correlates with adverse outcomes for patients with breast cancer in all cohorts analyzed is consistent with a recently reported correlation between PTK6 mRNA levels and tumor grade based on analyses of 44 tumors
PTK6 is an attractive potential therapeutic target for breast and ovarian cancer for several reasons. The rather restricted expression of PTK6 in normal tissues and the prevalence of PTK6 copy number gain and high expression in some breast and ovarian tumors improve the chance of selectively targeting malignant cells. PTK6 overexpression could contribute to resistance to currently available therapies, such as those targeting Her2 amplified tumors
MCF-10A, ZR751, MCF-7, BT474, DOV-13, Hey C2 and human immortalized mammary epithelial cells (HMEC) were obtained from ATCC (Manassas, Virginia). MCF-10A cells overexpressing IGF-1R receptor (IGF-1R) have been previously described
The constructs for PTK6 (kinase-active and kinase-inactive) were generated by cloning into pMSCV-puro. The construct encoding short hairpin RNA sequences targeting PTK6 was obtained from Open Biosystems (Catalog # TRCN0000021552). pBP-MF-PTK6 and pBP-F-PTK6 were obtained from the Harvard Institute of Proteomics (gift of Haley Hieronymus and Jesse Boehm). Viral packaging 293T or GPG-293T cells were transfected according to standard protocols. Viral supernatant was collected 36 and 60 hrs post-transfection. Immortalized HMECs or MCF-10A cells were infected in the presence of 2µg/ml polybrene with the viral supernatant or were spin-infected for 30 minutes at 2250 rpm. After three hours, media was changed on the target cells, which were then allowed to recover overnight, followed by another round of infection. Forty-eight hours post-infection, the target cells were exposed to puromycin (0.5–1 µg/ml for 48 hrs) to select for infected cells.
A library consisting of two individual siRNA sequences targeting human kinases were obtained from Qiagen (v.1.0). MCF-10A cells overexpressing IGF-1R were trypsinized and plated in 96 well plates pre-coated with polyhema (Sigma). Cells were transfected using Oligofectamine in quadruplicate wells with individual siRNA duplexes from the Qiagen kinase siRNA library. Cells were cultured for an additional 72 hours in suspension and assayed for Alamar reducing potential on a multiplate reader according to the manufacturer's protocol.
In the counter screen, parental MCF-10A cells cultured on polyhema plates were transfected with siRNA duplexes. After four hours the media was supplemented with Matrigel™ to a final concentration of 2%. After an additional 72 hour incubation, Alamar blue reduction was assayed as above. The ratio between the Alamar blue score in IGF1R-overexpressing cells to the score for MCF10A cells grown in Matrigel was calculated. This ratio was used to select candidates for testing additional siRNA reagents.
Cell death in suspension cultures was assessed using the Cell Death ELISA kit (Roche Diagnostics).according to the manufacturer's instructions.
The bottom layer of soft agar contained 0.6% agar with DMEM and 5% serum. Fifteen thousand cells were seeded per well on the top layer that contained 0.3% agar in MEGM. Cells were seeded in triplicate wells of six-well plates. Wells were re-fed every two weeks and colonies were counted at four weeks post-plating. At each time point, an image of each well was taken at 6× magnification and processed using the software Image J. Colonies larger than 50 sq. pixels were counted.
Breast tumors for SNP array analysis were collected from de-identified cases using protocol #98-229 approved by the Institutional Review Board of the Dana-Farber Cancer Institute. SNP array analyses were performed by the Dana-Farber Microarray Core and by the Broad Institute using Affymetrix 250K StyI arrays, according to methods described by Nikolsky et al
A fosmid probe specific for PTK6 and a Chromosome 20 centromere probe were obtained from CHORI/BacPac library. Hybridization and analyses were performed by the Dana-Farber/Harvard Cancer Center Cytogenetics Core Facility (P30 CA006516).
Boxplots showing the level of PTK6 mRNA in breast tumor subtypes were derived from three data sets: (1) GSE1992 downloaded from the Gene Expression Omnibus (GEO,
For GSE5460, raw expression values obtained from Affymetrix GENECHIP software were additionally analyzed using DNA-Chip analyzer (dChip) custom software (
For the Van't Veer data set, normalized PTK6 expression values and time to recurrence data were downloaded from the Rosetta website (
For the Wang set, PTK6 expression values and time to recurrence data were downloaded from the Gene Expression Omnibus (GEO, GSE2034). Samples were divided into three equal tertiles: 95 tumors with highest PTK6 level (>242.3); 96 tumors with intermediate PTK6 level (between 126.3 and 243.7) and 95 tumors with lowest PTK6 level (<126.9). For the ER+ plots, only the 209 samples classified as ER+ by Wang et al. were used.
For all survival curves, the y-axis (probability) is defined as the frequency of relapse-free survival. Kaplan-Meier analyses were carried out using survival package within R language. P-values were derived using the log-rank test and by fitting a Cox proportional hazards regression model. Multivariate Cox proportional hazards analysis was performed for PTK6 and clinical prognostic factors including estrogen receptor (ER),HER2 oncogene, axillary lymph node involvement.
Cells were lysed in RIPA lysis buffer (1%Triton X-100, 1%NaDOC, 0.1%SDS, 20mM Tris pH 7.5, 150 mM NaCl, 1mM EDTA) or RPPA lysis buffer (1% Triton X-100, 50mM HEPES, 150mM NaCl, 1.5mM MgCl2, 1mM EDTA, 100mM NaF, 10mM NaPP, 10% glycerol, 1mM PMSF, 1mM Na3V04) supplemented with protease and phosphatase inhibitors (1µg/ml leupeptin, 1 µg/ml aprotinin, 1 µg/ml pepstatin, 10 µg/ml PMSF, 1 mM NaVO4, 1mM NaF) for 30 minutes at 4°C. Lysates were clarified by centrifugation and supernatants were collected. Proteins were resolved by 8–10% SDS-PAGE gel electrophoresis and immunoblotted using standard techniques.
Serially diluted cell lysates were printed on nitrocellulose-coated slides and probed with specific antibodies as previously reported
Cells grown in monolayer cultures, as well as 3D acinar structures were visualized at 20°C using a Nikon TE300 microscope equipped with a CCD camera, using a 4×/0.13 objective. Images were acquired using ImageJ software, converted to TIFF images and arranged using Adobe Photoshop 7.0.
Acinar structures were fixed in 2% formalin (Sigma) at room temperature for 20 minutes and permeabilized in 0.5% Triton X-100 in PBS for 10 min at 4°C. Immunostaining of acinar structures was carried out as previously described
siRNAs targeting PTK6 identified by the screen inhibit anchorage-independent survival of IGF-1R cells. A) Parental and IGF-1R overexpressing MCF-10A cells transfected with control or PTK6 siRNA (OTP, Dharmacon) were cultured in suspension for 48 hours. Lysates were probed with the indicated anti-sera. B) Western analysis demonstrates down regulation of PTK6 expression in IGF-1R cells by siRNAs targeting PTK6 identified by the screen to induce changes in viability. IGF-1R cells transfected with multiple PTK6 siRNAs, including a pool of siRNAs from a distinct vendor (OTP, Dharmacon), were cultured in suspension for 48 hours. Lysates were prepared and probed with the indicated anti-sera. C) siRNAs targeting PTK6 inhibit anchorage-independent survival of IGF-1R cells. IGF-1R cells transfected with multiple PTK6 siRNAs, including a pool of siRNAs from a distinct vendor (OTP, Dharmacon), were cultured in suspension for 48 hours. Cell death was assessed as described in
(0.24 MB TIF)
Overexpression of PTK6 alone in HMEC does not enhance growth in soft agar. Immortalized human mammary epithelial cells (HMLE) overexpressing vector control, F-PTK6 or MF-PTK6, were seeded in soft-agar, and colony formation was assessed after four weeks. A representative experiment of three replicate experiments using triplicate wells for each cell line is shown.
(0.08 MB TIF)
Expression of PTK6 in breast and ovarian cancer cell lines. A) Western blot analysis was performed to assess levels of PTK6 protein expression in a panel of breast cancer cell lines representative of different subtypes, as classified in Neve et al.
(0.53 MB TIF)
Effect of IGF-1R kinase inhibitor on IGF-1R signaling. IGF-1R overexpressing MCF-10A cells were treated in suspension cultures containing IGF-1 (100ng/ml) for 48 hours with the indicated concentrations of IGF-1 receptor kinase inhibitor (BMS-536924; Selleck). Lysates were probed with the indicated antisera to assess IGF-1R autophosphorylation and phosphorylation of Akt.
(0.28 MB TIF)
Cluster analysis of chromosome 20 copy number alterations in breast cancers. The heatmap is based on SNP array copy number data from 93 breast tumor samples with 511 chromosome 20 genes (ordered from p arm to q arm). Only data above log2 value 0.5 or below −0.5 were used (representing copy numbers above 2.8 or below 1.4). Left, cluster of the entire chromosome 20; right, magnification of the yellow box. Dark blue, red, and light blue colors represent losses, gains, and no significant change in copy number.
(1.07 MB TIF)
Higher levels of PTK6 mRNA are associated with clinical measures of poor prognosis in multiple breast cancer data sets. Box plots and data are from the Oncomine database (
(0.65 MB TIF)
We thank Zhenyang Jiang and Shawn Hakim for technical assistance and Loling Song for helpful discussions. We are grateful to Haley Hieronymus, Jesse Boehm and the Harvard Institute of Proteomics for the pBP kinase vectors.