Skip to main content
NCBI home page
Molecular Oncology logoLink to Molecular Oncology
. 2008 Oct 2;2(4):327–339. doi: 10.1016/j.molonc.2008.09.004

CAMK1D amplification implicated in epithelial–mesenchymal transition in basal‐like breast cancer

Anna Bergamaschi 1,2, Young H Kim 2, Kevin A Kwei 2, Yoon La Choi 2,3, Melanie Bocanegra 2, Anita Langerød 1, Wonshik Han 4, Dong-Young Noh 4, David G Huntsman 5, Stefanie S Jeffrey 6, Anne-Lise Børresen-Dale 1,7, Jonathan R Pollack 2,✉
PMCID: PMC2653212  NIHMSID: NIHMS83121  PMID: 19383354

Abstract

Breast cancer exhibits clinical and molecular heterogeneity, where expression profiling studies have identified five major molecular subtypes. The basal‐like subtype, expressing basal epithelial markers and negative for estrogen receptor (ER), progesterone receptor (PR) and HER2, is associated with higher overall levels of DNA copy number alteration (CNA), specific CNAs (like gain on chromosome 10p), and poor prognosis. Discovering the molecular genetic basis of tumor subtypes may provide new opportunities for therapy. To identify the driver oncogene on 10p associated with basal‐like tumors, we analyzed genomic profiles of 172 breast carcinomas. The smallest shared region of gain spanned just seven genes at 10p13, including calcium/calmodulin‐dependent protein kinase ID (CAMK1D), functioning in intracellular signaling but not previously linked to cancer. By microarray, CAMK1D was overexpressed when amplified, and by immunohistochemistry exhibited elevated expression in invasive carcinomas compared to carcinoma in situ. Engineered overexpression of CAMK1D in non‐tumorigenic breast epithelial cells led to increased cell proliferation, and molecular and phenotypic alterations indicative of epithelial–mesenchymal transition (EMT), including loss of cell–cell adhesions and increased cell migration and invasion. Our findings identify CAMK1D as a novel amplified oncogene linked to EMT in breast cancer, and as a potential therapeutic target with particular relevance to clinically unfavorable basal‐like tumors.

Keywords: Basal-like breast cancer, Genomic profiling, DNA amplification, Oncogene, Epithelial‐mesenchymal transition, EMT

1. Introduction

Breast cancer is the most frequently diagnosed cancer and second leading cause of cancer death among women in the United States (Jemal et al., 2007). Breast cancer is known to comprise distinct molecular entities, where scoring expression of specific molecular markers like ER, PR and HER2 has implications for prognostication and selection of therapy (Subramaniam et al., 2005).

More recently, DNA microarray‐based expression profiling has suggested a refined classification, distinguishing five major subtypes based on different patterns of gene expression (Perou et al., 2000; Sorlie et al., 2001). Luminal‐A and ‐B subtypes, so‐called because they share expression markers with the luminal epithelial layer of normal breast ducts, are ER positive, with luminal‐B (also known as ‘highly proliferating luminal’ (Langerod et al., 2007)) exhibiting higher proliferation rates and less favorable prognosis than luminal‐A subtype. The basal‐like subtype, sharing expression markers with the basal (myoepithelial) layer of normal breast ducts, is ER negative and associated with particularly poor outcome in most but not all studies (Chin et al., 2006; Sorlie et al., 2001). The ERBB2 subtype is associated with overexpression of genes co‐amplified with ERBB2 (encoding HER2) on chromosome region 17q12–q21, and the normal‐like subtype shares expression patterns with normal breast tissue specimens.

The nature of the gene expression patterns suggests a possible different cell type of origin for the various subtypes, for example basal vs. luminal epithelial progenitor cells (Boecker et al., 2003). Recent studies have also shown the different gene expression subtypes to be associated with different overall levels of DNA copy number alteration (CNA), implying distinct underlying genomic instabilities, as well as different specific CNAs, implying progression via distinct sets of known or novel cancer genes (Bergamaschi et al., 2006; Chin et al., 2006; Hicks et al., 2006; Lai et al., 2007). The ERBB2 subtype is defined in part by amplification on 17q (ERBB2). Luminal‐A is associated with gain on 1q and 16p. Luminal‐B has more overall high‐level DNA amplification and specific amplifications on 8q (MYC) and 20q (ZNF217). Basal‐like tumors have more overall low‐level chromosome segment gain/loss, as well as specific gain on 10p, and loss on 4q and 5q among other CNAs.

Discovering the driver genes at these subtype‐specific loci is key to understanding the pathobiology of the different tumor subtypes, and may suggest new strategies for therapy. Here, we focus on the 10p amplicon associated with basal‐like tumors. We use genomic profiling and functional studies to identify CAMK1D as a candidate amplified oncogene contributing to cell proliferation, and to the loss of cell–cell adhesion and increased motility/invasion that characterize the epithelial–mesenchymal transition (EMT) phenotype linked to breast cancer progression.

2. Results

2.1. Genomic profiling defines 10p13 gain spanning CAMK1D

Prior array CGH studies identified 10p gain to be associated with basal‐like breast tumors (Bergamaschi et al., 2006; Chin et al., 2006). To delimit the smallest common region of gain, we analyzed cDNA array CGH genomic profiles from 172 primary breast carcinomas (clinicopathological characteristics are summarized in Table 1), including 87 specimens newly profiled for this study. In total, gain on 10p occurred in 19 of 172 (11%) breast cancers, and in 11 of 28 (39%) basal‐like tumors compared to eight of 136 (6%) assignable non‐basal‐like tumors (P<0.001, Fisher's exact test).

Table 1.

Clinicopathological features of tumor sets

DOXO‐FUMI (n=85) ULL (n=50) Korean (n=37)
Age (median) 64 67.6 41
Grade
Low grade (I and II) (%) 45 (54) 37 (74) 20 (56)
High grade (III) (%) 38 (46) 13 (26) 16 (44)
Unknown 2 0 1
TNM classification: T
T1 (%) 0 7 (14) 10 (27)
T2 (%) 2 (2) 32 (66) 22 (59)
T3 (%) 53 (63) 5 (10) 5 (14)
T4 (%) 29 (35) 5 (10) 0 (0)
Unknown 1 1 0
Estrogen receptor
Positive (%) 68 (83) 26 (54) 17 (46)
Negative (%) 14 (17) 22 (46) 20 (54)
Unknown 3 2 0
TP53
Wildtype (%) 44 (52) 37 (74) 22 (59)
Mutant (%) 41 (48) 13 (26) 15 (41)
Unknown 0 0 0
TNM classification: N
Positive (%) 62 (76) 19 (42) 27 (73)
Negative (%) 20 (24) 26 (58) 10 (27)
Unknown 3 5 0
Histology
Ductal (%) 74 (87) 29 (59) 37 (100)
Lobular (%) 7 (8) 15 (31) 0 (0)
Others (%) 4 (5) 5 (10) 0 (0)
Unknown 0 1 0
Subtypes
Luminal A (%) 35 (41) 16 (36) 17 (50)
Luminal B (%) 15 (18) 6 (13) 2 (6)
ErbB2 (%) 21 (25) 7 (16) 8 (24)
Basal‐like (%) 12 (14) 10 (22) 6 (18)
Normal‐like (%) 2 (2) 6 (13) 1 (3)
Indeterminate 0 5 3
Open in a new tab

The smallest region of common gain (Figure 1A) spanned ∼12Mb within cytoband 10p13 and included just seven RefSeq (Pruitt et al., 2003) genes: CAMK1D (calcium/calmodulin‐dependent protein kinase ID), CCDC3 (coiled‐coil domain containing 3), OPTN (optineurin), MCM10 (minichromosome maintenance complex component 10), PHYH (phytanoyl‐CoA 2‐hydroxylase), SEPHS1 (selenophosphate synthetase 1), and C10orf30 (chromosome 10 open reading frame 30). CAMK1D (also called CaMKIδ and CKLiK) is a member of the Ca2+/calmodulin‐dependent protein kinase (CaMK) family (Soderling, 1999), which transduces intracellular calcium signals to affect diverse cellular processes. Though a role of CaMKs in cancer remains largely unexplored, kinases are proven cancer drug targets (Sawyers, 2002), and therefore we sought to determine a possible functional connection between CAMK1D amplification/overexpression and breast cancer.

Figure 1.

Figure 1

Open in a new tab

CAMK1D at 10p13 is amplified and overexpressed in breast cancer. (A) Genomic profiles by CGH on cDNA microarrays of primary breast tumors, spanning cytoband 10p13. Each column represents a different tumor (cohort and expression subtype indicated), and each row represents a different gene ordered by chromosome position. Red indicates positive tumor/normal aCGH ratios (scale shown), and samples called gained at 10p13 are marked below by closed circle. Key samples defining the smallest common region of gain are marked by an asterisk. Genes and ESTs (IMAGE clone ID shown) residing within the amplicon core are labeled; CCDC3 and MCM10 (asterisked) were not present on the array but reside where shown. Samples assayed by Q‐PCR having gain (closed arrow) or no gain (open arrow) are indicated; note, the gain in sample “4” was missed by the CLAC CGH calling algorithm. (B) Q‐PCR validation of CAMK1D gain. Note, hybridization measurements by CGH tend to underestimate true CNA ratios (Pollack et al., 1999). (C)CAMK1D transcript levels, analyzed by microarray for the 120 (of 172) well‐measured specimens, are elevated in breast cancer specimens with DNA gain compared to no gain at 10p13. Box plots show 25th, 50th and 75th percentiles; P‐value (Mann–Whitney U‐test) indicated. (D) Immunohistochemical staining of CAMK1D expression in breast cancer. Shown are representative IHC stains for CAMK1D protein expression in (left): normal breast ducts; note prominent staining of basal layer (closed arrows); staining in terminal ductal lobular units was not confined to the basal layer (not shown); (center): ductal carcinoma in situ (DCIS) (outlined by open arrows), inadvertently sampled along with invasive cancer on the TMAs; and (right): a strongly positive staining invasive ductal carcinoma.

We first validated CAMK1D gain in a subset of samples by quantitative‐PCR (Q‐PCR) (Figure 1B). Next, consistent with a possible oncogenic role, we found by microarray CAMK1D to be overexpressed in specimens with gain (Figure 1C). To evaluate CAMK1D protein expression in breast samples, we performed immunohistochemistry (IHC) using tissue microarrays (TMAs) that included cases of normal breast, ductal carcinoma in situ (DCIS), and invasive breast carcinoma (Figure 1D). In normal breast ducts, CAMK1D expression was prominent in the basal epithelial layer. Immunostaining was also more pronounced in invasive cancer compared to DCIS, and overall 56% of invasive breast cancers exhibited moderate/strong staining. Despite that 10p gain was more common in basal‐like tumors, CAMK1D immunostaining was not more frequent in basal‐like tumors compared to all other subtypes combined (subtypes defined by IHC markers, see Section 4) (Table 2; P=0.53, Fisher's exact test). There was also no significant association between CAMK1D expression and prognosis among the 259 scorable cases with clinical follow‐up (Figure 2; P=0.67, log rank test).

Table 2.

CAMK1D immunostaining in basal‐like subtype

CAMK1D staining intensity Basal‐like casesa,b (%) Other subtypes (%) Total
Negative 18 (9.7) 6 (22.2) 24
Weak 63 (33.9) 8 (29.6) 71
Moderate 57 (30.6) 8 (29.6) 65
Strong 48 (25.8) 5 (18.5) 53
Total 186 (100) 27 (100) 213
Open in a new tab
a

Basal=ER−, PR−, HER2−, and CK5+ or EGFR+.

b

P =0.53 (Fisher's exact test; mod/strong staining in basal‐like vs. non‐basal‐like).

Figure 2.

Figure 2

Open in a new tab

CAMK1D expression and patient outcome. CAMK1D expression, scored by IHC on TMAs, is not associated with poor prognosis. Kaplan–Meier analysis of overall survival; log rank test P‐value shown. Performed on 259 scorable cases from Vancouver General Hospital; 115 cases scored negative/weak, and 144 cases moderate/strong.

2.2. CAMK1D promotes cell proliferation and epithelial–mesenchymal transition

To investigate a functional role of CAMK1D amplification/overexpression, we engineered the stable overexpression of CAMK1D (both native and V5 epitope tagged) in MCF10A cells, an immortalized but non‐tumorigenic breast epithelial cell line with undetectable CAMK1D expression (Figure 3A), and no CNA at 10p by array CGH (not shown). Stable expression of CAMK1D and CAMK1D‐V5 was verified by Western blot (Figure 3A). Since the tagged protein exhibited higher expression levels in stably transfected cell pools, we chose this construct for more detailed characterization in the studies described below (though we also verified comparable effects of the native untagged form in selected assays).

Figure 3.

Figure 3

Open in a new tab

CAMK1D overexpression in MCF10A cells promotes cell proliferation. (A) Western blot analysis confirming CAMK1D expression (43kD) in MCF10A cells stably transfected with CAMK1D (native or V5 epitope tagged) construct compared to empty vector control. GAPDH serves as a loading control. (B) CAMK1D‐V5 expression enhances MCF10A cell proliferation, measured by WST‐1 assay, compared to empty‐vector control. *, P<0.05; **, P<0.01 (Student's t‐test). (C) CAMK1D expression promotes MCF10A cell‐cycle progression, evidenced by increased S‐phase fraction (with decreased G1 fraction) following BrdU labeling, quantified by flow cytometry (representative plots shown). *, P<0.05 (Student's t‐test).

To test a role of CAMK1D in cell proliferation, we measured cell growth (using metabolic activity as a surrogate), and cell‐cycle progression by BrdU incorporation. MCF10A cells expressing CAMK1D exhibited significantly increased growth compared to vector‐alone control cells (Figure 3B), as well as increased cell‐cycle progression (evidenced by increased S‐phase fraction) (Figure 3C).

In culturing CAMK1D‐expressing MCF10A cells, we observed an altered morphology (Figure 4A). At sub‐confluent densities, MCF10A and MCF10A‐empty vector control cells grew with a characteristic “cobblestone” pattern reflecting cell–cell adhesions characteristic of epithelia. In contrast, MCF10A‐CAMK1D cells at the same density tended to grow with fewer cell–cell contacts, and with higher numbers of spindle‐shapes more characteristic of mesenchymal/fibroblast cells.

Figure 4.

Figure 4

Open in a new tab

CAMK1D overexpression in MCF10A promotes features of EMT. (A) MCF10A cells stably expressing CAMK1D‐V5 (right) (or native CAMK1D, not shown) exhibit increased morphologic features of EMT (i.e. spindle‐like morphology, reduced cell–cell contacts), compared to untransfected (left) or empty‐vector control cells (center) plated at the same density. Phase contrast images captured at 4× magnification. (B) CAMK1D overexpression in MCF10A cells is associated with decreased E‐cadherin (epithelial marker) and increased vimentin (mesenchymal marker) expression, measured by Western blot, compared to untransfected or empty‐vector control cells. GAPDH serves as a loading control. (C) CAMK1D expression enhances cell migration, measured by wound healing assay. The slide surface was scratched (time 0) then photographed at 12h. Margins of the original scratch are shown superimposed, using a felt pen mark for orientation. Enhanced migration is evident by the greater cell numbers migrated to fill the gap in CAMK1D‐expressing MCF10A cells (right) compared to empty‐vector control (left). (D) CAMK1D‐V5 expression (or native CAMK1D, not shown) promotes MCF10A cell invasion, quantified by Boyden chamber/Matrigel assay, compared to empty‐vector control. *, P<0.05; **, P<0.01 (Student's t‐test). Insets show representative fields of invasive cells (crystal violet‐stained). Note, lamellipodia‐like structures (arrows) are more prominent in CAMK1D‐expressing invasive cells.

A loss of cell–cell adhesion is one of the features of cells undergoing epithelial to mesenchymal transition (EMT), a regulated process during normal development but thought to be pathologically appropriated by cancer cells to effectuate invasion and metastasis (Thiery, 2002). Therefore, we characterized the relation between CAMK1D expression and other features of EMT. By Western blot, CAMK1D expression was associated with molecular hallmarks of EMT, namely decreased expression of the cell–cell adhesion glycoprotein E‐cadherin (CDH1), and increased expression of the mesenchymal intermediate filament vimentin (VIM) (Figure 4B). Increased vimentin expression in MCF10A‐CAMK1D cells was confirmed by immunofluorescence (Figure 5). MCF10A‐CAMK1D cells also exhibited increased cell migration (visualized by scratch assay) (Figure 4C), and increased invasion (quantified by Boyden chamber Matrigel assay) (P<0.01, Student's t‐test) (Figure 4D), both key phenotypic features of EMT.

Figure 5.

Figure 5

Open in a new tab

CAMK1D expression drives increased vimentin levels. Vimentin expression by immunofluorescence in (A) CAMK1D‐V5 overexpressing MCF10A cells and (B) empty‐vector control cells, imaged at 40× magnification with identical exposure times. Immunofluorescence images are pseudocolored as follows: CAMK1D (red), vimentin (green), TO‐PRO DNAstain (blue). Note increased vimentin expression in cells expressing (closed arrows) compared to not expressing (open arrows) CAMK1D‐V5.

2.3. Expression analysis implicates CREB pathway

To gain additional insight into CAMK1D‐modulated cell phenotypes, we compared gene expression patterns between CAMK1D‐expressing and non‐expressing MCF10A cells using whole‐genome oligonucleotide microarrays. Gene set enrichment analysis (GSEA) (Subramanian et al., 2005) was applied to discover differentially modulated biological processes. GSEA identified several functional gene sets significantly enriched in MCF10A cells expressing CAMK1D (Table 3), including gene sets relating to cell proliferation and EMT. The analysis also revealed regulatory motif gene sets, many relating to E2F transcription factors known to function in cell‐cycle progression (Table 3).

Table 3.

GSEA‐derived significant gene sets

NAME NESa FDRb
FUNCTIONAL PROTEASOME_PATHWAY 2.11 0.00
GLUT_DOWN 2.08 0.00
HTERT_UP 2.07 0.00
LEU_DOWN 2.01 0.01
AMI_PATHWAY 1.99 0.01
RAP_DOWN 1.98 0.01
CSK_PATHWAY 1.98 0.01
PROTEASOME_DEGRADATION 1.98 0.01
SIG_REGULATION_OF_THE_ACTIN_CYTOSKELETON_BY_RHO_GTPASESc 1.88 0.02
GLUCOSE_DOWN 1.84 0.03
CELL_CYCLEd 1.81 0.03
NDKDYNAMIN_PATHWAY 1.77 0.05
ANDROGEN_UP_GENES 1.76 0.05
ADULT_LIVER_VS_FETAL_LIVER_GNF2 1.71 0.08
KRAS_TOP100_KNOCKDOWN 1.70 0.08
NO1_PATHWAY 1.67 0.10
TGF_BETA_SIGNALING_PATHWAYc 1.67 0.09
CR_CELL_CYCLEd 1.66 0.10
G13_SIGNALING_PATHWAY 1.64 0.10
SHH_LISA 1.63 0.10
mRNA_SPLICING 1.63 0.10
AR_MOUSE 1.62 0.10
mRNA_PROCESSING 1.62 0.10
UPREG_BY_HOXA9 1.61 0.10
SA_G1_AND_S_PHASESd 1.57 0.13
AR_MOUSE_PLUS_TESTO_FROM_NETAFFX 1.56 0.14
EMT_UPc 1.56 0.13
GLUCO 1.56 0.13
FRASOR_ER_UP 1.56 0.13
CR_CAM 1.55 0.13
AR_ORTHOS_MAPPED_TO_U133_VIA_NETAFFX 1.53 0.14
REGULATORY MOTIF KRCTCNNNNMANAGC_UNKNOWN 2.66 0.00
TTTNNANAGCYR_UNKNOWN 2.46 0.00
V$ZIC3_01 1.91 0.13
V$E2F_Q6_01 1.88 0.13
V$E2F1_Q3 1.88 0.10
V$E2F1DP2_01 1.86 0.11
V$E2F_02 1.85 0.10
V$E2F1DP1_01 1.85 0.09
V$E2F_Q4_01 1.83 0.10
V$E2F4DP1_01 1.83 0.09
GATTGGY_V$NFY_Q6_01 1.83 0.08
V$E2F_03 1.82 0.08
V$E2F4DP2_01 1.82 0.08
SGCGSSAAA_V$E2F1DP2_01 1.81 0.09
V$E2F1_Q4_01 1.79 0.10
V$E2F_Q3_01 1.78 0.10
GTGACTT,MIR‐224 1.78 0.10
V$CEBPDELTA_Q6 1.77 0.10
V$E2F_Q3 1.76 0.10
AGGAAGC,MIR‐516‐3P 1.74 0.12
V$OCT_Q6 1.71 0.14
V$CREB_Q2_01 1.71 0.14
V$E2F1_Q6 1.69 0.15
Open in a new tab
a

Normalized enrichment score.

b

False discovery rate.

c

EMT‐relevant.

d

Cell proliferation‐relevant.

GSEA of regulatory motif sets also identified enrichment of CREB (cAMP responsive element binding protein) binding sites. CREB is a ubiquitous bZIP (basic and leucine zipper) transcription factor that functions as a homodimer, or heterodimer with the related protein ATF1, to activate the transcription of genes modulating diverse processes including cell proliferation and differentiation (Shaywitz et al., 1999). CREB itself is activated by a diverse set of stimuli via upstream kinases, including CaMKs (Hook et al., 2001). Consistent with the GSEA result, Western blot analysis identified increased phospho‐ser133 CREB (the activated form) in MCF10A cells expressing CAMK1D (Figure 6).

Figure 6.

Figure 6

Open in a new tab

CAMK1D ecpression is associated with CREB pathway activation. Phospho‐Ser133 CREB (active form) and phospho‐ATF1 (detected by antibody cross‐reactivity) levels, by Western blot, are increased in MCF10A cells overexpressing CAMK1D‐V5 compared to empty‐vector control. Right‐most lane contains phosphorylated‐CREB positive control cell extracts (SK‐N‐MC cells prepared with IBMX and forskolin treatment; Cell Signaling). GAPDH serves as a loading control.

3. Discussion

Prior studies have defined molecular subtypes of breast cancer with different patterns of gene expression and CNAs, and with different clinical behaviors. One of these subtypes, the basal‐like, represents a “triple negative” (i.e. ER−, PR−, HER2−) class (Kreike et al., 2007) with poorly understood pathogenesis and unfavorable outcome. Recurrent CNAs selective for basal‐like tumors provide a starting point to discover the underlying molecular genetic alterations. Here, we undertook a genomic profiling study to pinpoint the driver oncogene within the 10p gain recurring in basal‐like tumors.

Genomic profiling confirmed an association of 10p gain with basal‐like breast tumors, and delimited a smallest region of shared gain to a ∼12Mb region within 10p13 spanning seven genes. One of these, CAMK1D, was selected for further study because of its function in intracellular signaling, and because it belongs to a “druggable” class of proteins. CAMK1D belongs to the CaMK family of Ca2+/calmodulin‐regulated serine/threonine protein kinases that includes the multifunctional kinases CaMKK (itself a co‐activator of other CaMKs), CaMKI, CaMKII and CaMKIV (Hook et al., 2001; Soderling, 1999). CaMKs have been studied mainly in neurons and lymphocytes, and function in diverse cellular processes (though comparatively few downstream targets and effectors have been characterized). CAMK1D, sharing 77% amino acid homology with CAMKI, was first cloned from granulocytes, where it functions in chemo‐attractant response (Verploegen et al., 2000, 2005), and was later cloned from HeLa cells (Ishikawa et al., 2003). The role of CaMKs in cancer remains largely unexplored, though recently CaMKI was shown to modulate cell proliferation in MCF7 breast cancer cells (Rodriguez‐Mora et al., 2005).

To explore a functional role of CAMK1D overexpression in cancer, we engineered its overexpression in MCF10A cells, an immortalized but non‐tumorigenic breast epithelial cell line. Of note, in addition to being a widely used cell line for overexpression studies, MCF10A cells also express basal epithelial markers (Ross et al., 2001), and therefore may provide a particularly relevant context for these studies. Overexpression of CAMK1D promoted cell proliferation, and induced features of EMT, namely loss of cell–cell contacts, a switch from E‐cadherin to vimentin expression, increased cell migration and invasion, and (by GSEA) characteristic gene expression changes. EMT occurs during normal development, where epithelia at specific sites give rise to loosely‐knit or migratory mesenchymal cells (Thiery, 2002). It is widely believed the same pathways are usurped by cancer cells to invade tissue and metastasize.

How might CAMK1D enhance cell proliferation? The related protein CaMKI has been shown to activate cyclin D1/cdk4 complexes in fibroblasts (Kahl et al., 2004), and to regulate cyclin D1 levels in MCF7 breast cancer cells (Rodriguez‐Mora et al., 2005). It remains to be determined whether CAMK1D might promote cell proliferation through similar actions on cyclin D1/cdk4 (upstream of E2F). Of note, we observed CREB activation in MCF10A cells overexpressing CAMK1D, and the CCND1 (encoding cyclin D1) promoter contains a CREB binding sequence (Herber et al., 1994). It is therefore tempting to speculate that CREB is a direct target of CAMK1D, and that the CREB pathway links CAMK1D overexpression with cell proliferation.

Still less is known regarding a connection between CaMKs and EMT. Molecular processes causally connected with EMT include signaling pathways (e.g. TGFβ, Wnt), transcription factors (e.g. Snail, Twist), cell–cell adhesion proteins (e.g. E‐cadherin), cytoskeletal modulators (e.g. Rho family) and extracellular proteases (e.g. matrix metalloproteinases) (Thompson et al., 2005). Of interest, some of these pathways (e.g. TGFβ signaling, regulation of actin cytoskeleton by Rho GTPases) are found among the functional gene sets associated with CAMK1D expression. Clearly, additional studies are needed to determine where CAMK1D might mechanistically intersect these processes to promote EMT, and whether CREB might function as a mediator.

Our immunostaining studies of breast tissues indicate CAMK1D expression is restricted primarily to the basal layer in normal breast ducts. This finding is intriguing given the association between 10p gain and basal‐like tumors. Strong staining was observed in 56% of breast tumors, but not noted in DCIS. The transition from DCIS to invasive cancer is associated with loss of cell–cell contacts, and migration/invasion, features of EMT. Therefore, finding high‐level CAMK1D expression in invasive cancer but not DCIS is consistent with our cell culture studies assigning a function of CAMK1D in EMT.

Given the connection between 10p gain and basal‐like tumors, however, an unexpected finding was that CAMK1D protein (by IHC) was not selectively expressed in the basal‐like subtype. One possibility is that CAMK1D plays a pathogenic role across tumor subtypes, but the mechanism for its overexpression in basal‐like tumors is low‐level gain. Consistent with this possibility, basal‐like tumors are known to be associated with increased levels of segmental gain/loss, which appears to be a preferred underlying mechanism driving altered gene expression (Bergamaschi et al., 2006; Chin et al., 2006). Another possibility is that while CAMK1D amplification/overexpression may contribute to invasive oncogenic phenotypes, a different gene residing at 10p13 is associated with oncogenesis specifically in the basal‐like subtype.

While additional studies are needed to clarify the link to basal‐like cancers, our studies clearly identify CAMK1D amplification and overexpression in a subset of breast cancers, and define a connection to cell proliferation and EMT. As EMT is believed to be integral to invasion and metastasis, CAMK1D represents an attractive target for therapy. Even more so, since as a kinase it belongs to an enzyme class with a proven track record with molecularly directed therapies, as shown by recent successes targeting ERBB2, BCR‐ABL and EGFR (Sawyers, 2002). In summary, our findings identify CAMK1D amplification/overexpression in breast cancer, and define CAMK1D as a new candidate target for molecular‐targeted therapy.

4. Experimental procedures

4.1. Breast cancer specimens

One hundred seventy two breast cancer specimens were studied from three cohorts: (i) 85 Norwegian patients with locally advanced (T3/T4 and/or N2) breast cancer, receiving doxorubicin (Doxo) or 5‐fluorouracil/mitomycin C (FUMI) neoadjuvant therapy (Geisler et al., 2001, 2003); (ii) 50 Norwegian patients from a population‐based series (ULL) (Kapp et al., 2006; Langerod et al., 2007; Zhao et al., 2004); and (iii) 37 Korean patients (Kapp et al., 2006). All specimens were collected with patient informed consent and IRB approval from participating institutions. Clinicopathological characteristics of specimens are summarized in Table 1. cDNA microarray‐based gene expression profiling data for these tumor specimens has been reported elsewhere (Kapp et al., 2006, [Link], 2001, 2003, 2004). Gene expression subtypes were confidently assignable to 168 of the 172 samples using the nearest centroid method (Sorlie et al., 2003).

4.2. Array‐based comparative genomic hybridization (array CGH)

cDNA microarrays were obtained from the Stanford Functional Genomics Facility and included 39,632 human cDNAs, representing 22,488 mapped human genes (18,040 UniGene clusters (Schuler, 1997)), together with 4112 additional mapped ESTs not assigned UniGene IDs. We performed array CGH according to our published protocols (Pollack et al., 1999; Pollack et al., 2002). Briefly, 4μg of genomic DNA from each tumor specimen was random‐primer labeled with Cy5 and co‐hybridized to the microarray along with 4μg of Cy3‐labeled normal female leukocyte reference DNA from a single donor. Following overnight hybridization and washing, arrays were imaged using a GenePix 4000B scanner (Molecular Devices, Union City, CA). Fluorescence ratios were extracted using SpotReader software (Niles Scientific, Portola Valley, CA), and the data uploaded into the Stanford Microarray Database (Demeter et al., 2007) for storage, retrieval and analysis. Array CGH profiles for 85 of the 172 samples were previously reported (DOXO/FUMI cohort) (Bergamaschi et al., 2006). The complete microarray dataset is available at the Stanford Microarray Database (http://smd.stanford.edu) and at the Gene Expression Omnibus (accession GSE12827).

4.3. Array CGH analysis

Background‐subtracted fluorescence ratios were normalized for each array by setting the average log fluorescence ratio for all array elements equal to 0. Genes were considered reliably measured if the fluorescence intensity for the Cy3 reference channel was at least 1.4‐fold above background. Map positions for arrayed cDNA clones were assigned using the NCBI genome assembly, accessed through the UCSC genome browser database (NCBI Build 36). For genes represented by multiple arrayed cDNAs, the average fluorescence ratio was used. DNA gains and losses were identified using the CLuster Along Chromosomes method (CLAC; http://www‐stat.stanford.edu/∼wp57/CGH‐Miner) (Wang et al., 2005).

4.4. Quantitative‐PCR

Quantitative (Q)‐PCR was performed using Applied Biosystems' (Foster City, CA) TaqMan probes and reagents on an ABI 7500 sequence detection system as per the manufacturer's instructions. PCR was initiated at 95°C for 15min (to activate the modified Taq polymerase), followed by a 40 cycle amplification (95°C 15s, 58°C 30s, 72°C 30s). Melting curve analysis was performed to ensure specific PCR product while excluding primer dimers. We used the comparative CT method (Livak et al., 2001) to calculate relative CAMK1D DNA levels normalized to GAPDH (a gene located outside the 10p13 amplicon and not exhibiting CNA), which we then expressed as a ratio to the C T value of CAMK1D (also normalized to GAPDH) obtained from normal DNA. PCR primer and probe sequences were as follows: CAMK1D forward 5′‐CATAGGACTGGAAGACCGAAGTTTT, reverse CTCGAGTCAGTACAGTTTGTGAGAA, TaqMan probe FAM‐CCACTGCAATTCTG; GAPDH forward 5′‐AAATGTCACCGGGAGGATTGG, reverse 5′‐GGAGATCTGGTTTCCGGAAGAC, TaqMan probe FAM‐CCTGCCCTTCTCCC.

4.5. Immunohistochemistry (IHC)

IHC was carried out on a tissue microarray (TMA) containing 259 scorable cases (each represented by duplicate 0.6mm cores) of invasive breast cancer, representing sequential archival cases from Vancouver General Hospital during the period 1974–1995 (Makretsov et al., 2003). A 4μm section was cut from the tissue microarray block, de‐paraffinized in Citrisolv (Fisher Scientific, Pittsburgh, PA), and hydrated in a graded series of alcohol solutions. Heat‐induced antigen retrieval was performed by microwave pretreatment in citrate buffer (1mM, pH 6.0) for 15min before staining. Endogenous peroxidase was blocked by preincubation with 1% hydrogen peroxide in phosphate‐buffered saline. Anti‐CAMK1D rabbit polyclonal antibody (AbCam Inc., Cambridge, MA) was used at 1:50 dilution for 1h. Chromogenic detection was carried out using a peroxidase‐conjugated secondary antibody and DAB reagents provided with the Envision detection kit (Dako, Carpinteria, CA). Staining intensity was scored by a pathologist (Y.C.) as absent, weak, moderate or strong (0–3 scale). Basal‐like tumors were defined as cases negative for ER, PR and HER2, but positive for CK5 or HER1 by IHC (Carey et al., 2006).

4.6. Stably‐transfected cell lines

The immortalized non‐tumorigenic breast epithelial cell line MCF10A was obtained from the American Type Culture Collection (Manassas, VA), and cultured in MEGM media (Cambrex, East Rutherford, NJ). A full length CAMK1D cDNA clone (TC100851) was obtained from Origene Technologies (Rockville, MD) and subcloned into pcDNA6/V5‐His (Invitrogen, Carlsbad, CA) to create two sequence‐verified expression constructs, one with an in‐frame C‐terminal V5 epitope tag (CAMK1D‐V5) and one without (CAMK1D). To create stably‐transfected cell pools, MCF10A cells (2.0×105 cells per six‐well plate well) were transfected with 4μg of either empty vector, CAMK1D‐V5 or CAMK1D expression construct using LipofectAMINE 2000 reagent (Invitrogen). One day post‐transfection, cells were seeded at 2.0×104 cells per well in six‐well plates, and the following day 10μg/ml blasticidin (Invitrogen) was added to the culture medium, and cells selected for 10–12days.

4.7. Western blot analysis

Total cellular protein was extracted using 1× RIPA Lysis buffer (Upstate/Chemicon, San Francisco, CA) supplemented with 1× complete protease inhibitor (Roche, Indianapolis, IN), 0.1mM sodium orthovanadate, 1mM sodium fluoride, and 1mM PMSF, and protein was quantified using the DC Protein Assay (Bio‐Rad, Hercules, CA). For Western blot, 25μg protein lysate was electrophoresed on a 4–15% Tris/glycine polyacrylamide gradient gel (Bio‐Rad) and transferred to PVDF membrane (Bio‐Rad). After blocking in TBS‐T buffer (20mM Tris–HCl (pH 7.4), 0.15M NaCl, and 0.1% Tween 20) with 5% dry milk for 30min, blots were incubated sequentially with primary antibody for 90min and secondary antibody for 45min, each at room temperature in TBS‐T buffer. Primary antibodies were used as follows: anti‐CAMK1D (1:500, AbCam), anti‐V5 antibody (1:5000, Invitrogen), anti‐Vimentin (1:1000, Santa Cruz Biotechnologies, Santa Cruz, CA), anti‐E‐Cadherin (Zymed/Invitrogen). Secondary antibodies were used as follows: HRP‐conjugated anti‐mouse IgG (1:20,000, Pierce, Rockford, IL), and HRP‐conjugated anti‐rabbit IgG (1:20,000, Pierce). Detection was carried out using the ECL kit (GE Healthcare, Piscataway, NJ). Levels of Ser‐133 phosphorylated CREB were detected using phopho‐CREB (Ser133) rabbit monoclonal antibody according to the manufacturer's protocol (Cell Signaling Technology, Danvers, MA).

4.8. Cell proliferation assay

Cell proliferation was quantified by colorimetry based on the metabolic cleavage of the tetrazolium salt WST‐1 in viable cells, according to the manufacturer's protocol (Roche Applied Science, Indianapolis, IN). Briefly, WST‐1 reagent was added at 1/10th the culture volume and incubated at 37°C for 30min. Absorbance was then measured at 450nm with reference to 650nm using a Spectra Max 190 plate reader (Molecular Devices, Sunnyvale, CA). Assays were performed in triplicate and average (±1 SD) OD reported.

4.9. Cell‐cycle analysis

Cell‐cycle distribution analysis was performed by flow cytometry using the BrdU‐FITC Flow kit (BD Biosciences, San Jose, CA) per the manufacturer's instructions. Cells were incubated with 10μM BrdU at 37°C for 4h, then fixed and permeabilized with Cytofix/Cytoperm buffer (BD Biosciences). Cellular DNA was treated with DNase at 37°C for 1h to expose incorporated BrdU, then cells were stained with anti‐BrdU FITC antibody (to quantify incorporated BrdU) and 7‐aminoactinomycin D (7‐AAD; to quantify total DNA content). A total of 10,000 events were scored by FACSCalibur (BD Biosciences) and analyzed using CellQuest software (BD Biosciences). Assays were performed in triplicate and average (±1 SD) cell‐cycle fractions reported.

4.10. Cell migration and invasion assay

Cell migration was quantified using a wound‐healing (scratch) assay. A total of 1.0×105 cells were seeded onto Lab‐Tek II chamber slides (Nalge Nunc, Rochester, NY) and 24h later the cell monolayer was scratched using a 200μl pipette tip. Cell migration into the wound was documented by phase contrast microscopy at times 0, 6, and 12h. Matrigel invasion assays were carried out using modified Boyden chambers with polycarbonate Nucleopore membrane (Corning, Corning, NY). Precoated filters (8μm pore size, Matrigel 100μg/cm2) were rehydrated with 500μl MEGM medium, and then 2×104 cells in 500μl MEGM medium were seeded into the upper chamber. Following incubation for 24, 48 and 72h at 37°C, cells were fixed in 10% formalin buffer and stained with crystal violet. Non‐invaded cells on the upper surface of the filter were removed with a cotton swab, and invasiveness quantified by counting stained cells. Invasion assays were performed in duplicate and the average (±1 SD) cell count reported.

4.11. Immunofluorescence

Cells (2.0×104) were plated onto Lab‐Tek II chamber slides, fixed in 1.5% paraformaldehyde for 15min and blocked (5% BSA, 0.3% Triton in PBS) for 1h. Primary antibodies (anti‐CAMK1D and anti‐Vimentin) were used at 1:50 dilution and incubated for 1h. Secondary antibodies (goat anti‐mouse Alexa468 and goat anti‐rabbit‐Alexa548; Invitrogen) were diluted 1:200 in blocking buffer and incubated for 1h at room temperature. Nuclei were counterstained with TO‐PRO (Invitrogen). Images were acquired using an epifluorescent microscope (Axioplan2; Carl Zeiss MicroImaging, Inc.), digital camera (ORCA‐ER C4742‐95; Hamamatsu Photonics), and image acquisition software (OpenLab 4.0.2; Improvision, Waltham, MA).

4.12. CAMK1D‐associated gene expression profiles

HEEBO (human exonic evidence‐based oligonucleotide) microarrays were obtained from the Stanford Functional Genomics Facility and contained 70‐mer probes representing 24,207 different human genes spotted on Nexterion E epoxysilane slides (Schott, Elmsford, NY). Before hybridization, slides were processed according to the manufacturer's instructions (Schott). Fifty μg Trizol‐prepared total RNA from test and reference (pool of 11 cancer cell lines) samples were each EtOH‐precipitated overnight and dissolved in 14.4μl RNase‐free H20. Two μl oligo‐dT primer (2.5μg/μl stock) (Qiagen, Valencia, CA) and 2μl of pdN6 random hexamer (2.3μg/μl stock) (GE Healthcare) were added, and samples denatured for 10min at 70°C, followed by a 5min incubation on ice. For cDNA synthesis, 11.6μl 1× reverse transcriptase (RT) master mix [6μl 5× SuperScript III first‐strand buffer (Invitrogen); 3μl 0.1M dithiothreitol (DTT); 0.6μl 50× amino‐allyl dUTP (aa‐dUTP) mix (Ambion, Austin, TX); 2μl SuperScript III reverse transcriptase (200U/μl) (Invitrogen)] was added to each RNA sample, then incubated for 2h at 48°C. Following incubation, remaining RNA was hydrolyzed by adding 13μl 1N NaOH and incubating at 67°C for 10min, followed by neutralization with 50μl 1M HEPES buffer (pH 7.3). Samples were purified using the Qiagen MinElute Reaction Cleanup Kit (Qiagen), and cDNA eluted in 25μl 10mM sodium phosphate pH 8.5. Test and reference samples were labeled respectively by addition of 25μl N‐hydroxysuccinimide ester (NHS)‐conjugated Cy5 and Cy3 dyes (40nmol/50μl DMSO stock) (GE Healthcare), followed by incubation at room temperature for 90min in a light‐protected environment. Following labeling, test and reference samples were pooled together and unincorporated dyes removed using the Qiagen QIAquick PCR Purification Kit. Samples were eluted in 25μl elution buffer, then ddH20 added to a final volume of 223μl. Twenty μl human Cot‐1 DNA (1mg/ml stock) (Invitrogen), 5μl yeast tRNA (5mg/ml stock) (Invitrogen), 2μl polyA (10mg/ml stock) (Sigma Aldrich, Saint Louis, MO), and 250μl 2× Hybridization Buffer (Agilent Technologies, Santa Clara, CA) were added to bring the final hybridization volume of 500μl. Samples were then denatured for 3min in boiling H20, incubated for 30min at 37°C, then hybridized to HEEBO microarrays using an Agilent hybridization gasket/chamber set in a rotating oven (Agilent Technologies), at 20rpm for 40h at 65°C. Following hybridization, arrays were imaged and data acquired as for array CGH above. Gene set enrichment analysis (GSEA) was performed comparing MCF10A cells expressing (CAMK1D, CAMK1D‐V5) vs. not expressing (empty‐vector and non‐transfected) CAMK1D, using 522 functional gene sets (C2v1) and 837 regulatory motif gene sets (C2v2) as described (Subramanian et al., 2005, 2007).

Acknowledgments

We wish to thank the SFGF for microarray manufacture, the SMD for database support, Alessandra Sacco for assistance with IF studies, and Dmitry Turbin and Sam Leung for assistance with TMA data analysis. We also thank the members of the Pollack lab for helpful discussion. This work was supported in part by grants from the NIH, CA112016 (J.R.P.); the California Breast Cancer Research Program, 8KB‐0135 (J.R.P.) and 11IB‐0175 (S.S.J); the Norwegian Research Council, NFR, 155218/300 (A.L.B.D.); and the Korea Health 21 R&D Project, Ministry of Health & Welfare R.O.K, 01‐PJ3‐PG6‐01GN07‐0004 (W.H. and D‐Y.N.). A.B. is a fellow of The Norwegian Cancer Society. D.G.H is a Michael Smith Foundation for Health Research Scholar. Funding for the TMA analysis was provided in part by an unrestricted educational grant from sanofi‐aventis Canada. The authors declare no conflicts of interest.

Bergamaschi Anna, Kim Young H., Kwei Kevin A., La Choi Yoon, Bocanegra Melanie, Langerød Anita, Han Wonshik, Noh Dong-Young, Huntsman David G., Jeffrey Stefanie S., Børresen-Dale Anne-Lise, Pollack Jonathan R., (2008), CAMK1D amplification implicated in epithelial–mesenchymal transition in basal-like breast cancer, Molecular Oncology, 2, doi: 10.1016/j.molonc.2008.09.004.

References

  1. Bergamaschi, A. , Kim, Y.H. , Wang, P. , Sorlie, T. , Hernandez-Boussard, T. , Lonning, P.E. , Tibshirani, R. , Borresen-Dale, A.L. , Pollack, J.R. , 2006. Distinct patterns of DNA copy number alteration are associated with different clinicopathological features and gene-expression subtypes of breast cancer. Genes Chromosomes Cancer 45, 1033–1040. [DOI] [PubMed] [Google Scholar]
  2. Boecker, W. , Buerger, H. , 2003. Evidence of progenitor cells of glandular and myoepithelial cell lineages in the human adult female breast epithelium: a new progenitor (adult stem) cell concept. Cell Prolif 36, (Suppl. 1) 73–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Carey, L.A. , Perou, C.M. , Livasy, C.A. , Dressler, L.G. , Cowan, D. , Conway, K. , Karaca, G. , Troester, M.A. , Tse, C.K. , Edmiston, S. , 2006. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA 295, 2492–2502. [DOI] [PubMed] [Google Scholar]
  4. Chin, K. , DeVries, S. , Fridlyand, J. , Spellman, P.T. , Roydasgupta, R. , Kuo, W.L. , Lapuk, A. , Neve, R.M. , Qian, Z. , Ryder, T. , 2006. Genomic and transcriptional aberrations linked to breast cancer pathophysiologies. Cancer Cell 10, 529–541. [DOI] [PubMed] [Google Scholar]
  5. Demeter, J. , Beauheim, C. , Gollub, J. , Hernandez-Boussard, T. , Jin, H. , Maier, D. , Matese, J.C. , Nitzberg, M. , Wymore, F. , Zachariah, Z.K. , 2007. The Stanford Microarray Database: implementation of new analysis tools and open source release of software. Nucleic Acids Res. 35, D766–D770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Geisler, S. , Lonning, P.E. , Aas, T. , Johnsen, H. , Fluge, O. , Haugen, D.F. , Lillehaug, J.R. , Akslen, L.A. , Borresen-Dale, A.L. , 2001. Influence of TP53 gene alterations and c-erbB-2 expression on the response to treatment with doxorubicin in locally advanced breast cancer. Cancer Res. 61, 2505–2512. [PubMed] [Google Scholar]
  7. Geisler, S. , Borresen-Dale, A.L. , Johnsen, H. , Aas, T. , Geisler, J. , Akslen, L.A. , Anker, G. , Lonning, P.E. , 2003. TP53 gene mutations predict the response to neoadjuvant treatment with 5-fluorouracil and mitomycin in locally advanced breast cancer. Clin. Cancer Res. 9, 5582–5588. [PubMed] [Google Scholar]
  8. Herber, B. , Truss, M. , Beato, M. , Muller, R. , 1994. Inducible regulatory elements in the human cyclin D1 promoter. Oncogene 9, 1295–1304. [PubMed] [Google Scholar]
  9. Hicks, J. , Krasnitz, A. , Lakshmi, B. , Navin, N.E. , Riggs, M. , Leibu, E. , Esposito, D. , Alexander, J. , Troge, J. , Grubor, V. , 2006. Novel patterns of genome rearrangement and their association with survival in breast cancer. Genome Res. 16, 1465–1479. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hook, S.S. , Means, A.R. , 2001. Ca(2+)/CaM-dependent kinases: from activation to function. Annu. Rev. Pharmacol. Toxicol 41, 471–505. [DOI] [PubMed] [Google Scholar]
  11. Ishikawa, Y. , Tokumitsu, H. , Inuzuka, H. , Murata-Hori, M. , Hosoya, H. , Kobayashi, R. , 2003. Identification and characterization of novel components of a Ca2+/calmodulin-dependent protein kinase cascade in HeLa cells. FEBS Lett. 550, 57–63. [DOI] [PubMed] [Google Scholar]
  12. Jemal, A. , Siegel, R. , Ward, E. , Murray, T. , Xu, J. , Thun, M.J. , 2007. Cancer statistics, 2007. CA Cancer J. Clin 57, 43–66. [DOI] [PubMed] [Google Scholar]
  13. Kahl, C.R. , Means, A.R. , 2004. Regulation of cyclin D1/Cdk4 complexes by calcium/calmodulin-dependent protein kinase I. J. Biol. Chem. 279, 15411–15419. [DOI] [PubMed] [Google Scholar]
  14. Kapp, A.V. , Jeffrey, S.S. , Langerod, A. , Borresen-Dale, A.L. , Han, W. , Noh, D.Y. , Bukholm, I.R. , Nicolau, M. , Brown, P.O. , Tibshirani, R. , 2006. Discovery and validation of breast cancer subtypes. BMC Genomics 7, 231 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kreike, B., van Kouwenhove, M., Horlings, H., Weigelt, B., Bartelink, H., Van de Vijver, M.J., 2007. Gene expression profiling and histopathological characterization of triple negative/basal-like breast carcinomas. Breast Cancer Res. 9, R65. [DOI] [PMC free article] [PubMed]
  16. Lai, C. , Horlings, H.H. , van de Vijver, M.M. , van Beers, E.H. , Nederlof, P.M. , Wessels, L.F. , Reinders, M.J. , 2007. SIRAC: supervised identification of regions of aberration in aCGH datasets. BMC Bioinform 8, 422 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Langerod, A., Zhao, H., Borgan, O., Nesland, J.M., Bukholm, I.R., Ikdahl, T., Karesen, R., Borresen-Dale, A.L., Jeffrey, S.S., 2007. TP53 mutation status and gene expression profiles are powerful prognostic markers of breast cancer. Breast Cancer Res. 9, R30. [DOI] [PMC free article] [PubMed]
  18. Livak, K.J. , Schmittgen, T.D. , 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) Method. Methods 25, 402–408. [DOI] [PubMed] [Google Scholar]
  19. Makretsov, N. , Gilks, C.B. , Coldman, A.J. , Hayes, M. , Huntsman, D. , 2003. Tissue microarray analysis of neuroendocrine differentiation and its prognostic significance in breast cancer. Hum. Pathol 34, 1001–1008. [DOI] [PubMed] [Google Scholar]
  20. Perou, C.M. , Sorlie, T. , Eisen, M.B. , van de Rijn, M. , Jeffrey, S.S. , Rees, C.A. , Pollack, J.R. , Ross, D.T. , Johnsen, H. , Akslen, L.A. , 2000. Molecular portraits of human breast tumours. Nature 406, 747–752. [DOI] [PubMed] [Google Scholar]
  21. Pollack, J.R. , Perou, C.M. , Alizadeh, A.A. , Eisen, M.B. , Pergamenschikov, A. , Williams, C.F. , Jeffrey, S.S. , Botstein, D. , Brown, P.O. , 1999. Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nat. Genet. 23, 41–46. [DOI] [PubMed] [Google Scholar]
  22. Pollack, J.R. , Sorlie, T. , Perou, C.M. , Rees, C.A. , Jeffrey, S.S. , Lonning, P.E. , Tibshirani, R. , Botstein, D. , Borresen-Dale, A.L. , Brown, P.O. , 2002. Microarray analysis reveals a major direct role of DNA copy number alteration in the transcriptional program of human breast tumors. Proc. Natl. Acad. Sci. USA 99, 12963–12968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Pruitt, K.D. , Tatusova, T. , Maglott, D.R. , 2003. NCBI reference sequence project: update and current status. Nucleic Acids Res. 31, 34–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Rodriguez-Mora, O.G. , LaHair, M.M. , McCubrey, J.A. , Franklin, R.A. , 2005. Calcium/calmodulin-dependent kinase I and calcium/calmodulin-dependent kinase kinase participate in the control of cell cycle progression in MCF-7 human breast cancer cells. Cancer Res. 65, 5408–5416. [DOI] [PubMed] [Google Scholar]
  25. Ross, D.T. , Perou, C.M. , 2001. A comparison of gene expression signatures from breast tumors and breast tissue derived cell lines. Dis Markers 17, 99–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Sawyers, C.L. , 2002. Rational therapeutic intervention in cancer: kinases as drug targets. Curr. Opin. Genet. Dev. 12, 111–115. [DOI] [PubMed] [Google Scholar]
  27. Schuler, G.D. , 1997. Pieces of the puzzle: expressed sequence tags and the catalog of human genes. J. Mol. Med 75, 694–698. [DOI] [PubMed] [Google Scholar]
  28. Shaywitz, A.J. , Greenberg, M.E. , 1999. CREB: a stimulus-induced transcription factor activated by a diverse array of extracellular signals. Annu. Rev. Biochem 68, 821–861. [DOI] [PubMed] [Google Scholar]
  29. Soderling, T.R. , 1999. The Ca-calmodulin-dependent protein kinase cascade. Trends Biochem. Sci. 24, 232–236. [DOI] [PubMed] [Google Scholar]
  30. Sorlie, T. , Perou, C.M. , Tibshirani, R. , Aas, T. , Geisler, S. , Johnsen, H. , Hastie, T. , Eisen, M.B. , van de Rijn, M. , Jeffrey, S.S. , 2001. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc. Natl. Acad. Sci. USA 98, 10869–10874. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Sorlie, T. , Tibshirani, R. , Parker, J. , Hastie, T. , Marron, J.S. , Nobel, A. , Deng, S. , Johnsen, H. , Pesich, R. , Geisler, S. , 2003. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc. Natl. Acad. Sci. USA 100, 8418–8423. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Subramaniam, D.S. , Isaacs, C. , 2005. Utilizing prognostic and predictive factors in breast cancer. Curr. Treat. Options Oncol 6, 147–159. [DOI] [PubMed] [Google Scholar]
  33. Subramanian, A. , Tamayo, P. , Mootha, V.K. , Mukherjee, S. , Ebert, B.L. , Gillette, M.A. , Paulovich, A. , Pomeroy, S.L. , Golub, T.R. , Lander, E.S. , 2005. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102, 15545–15550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Subramanian, A. , Kuehn, H. , Gould, J. , Tamayo, P. , Mesirov, J.P. , 2007. GSEA-P: a desktop application for Gene Set Enrichment Analysis. Bioinformatics 23, (23) 3251–3253. [DOI] [PubMed] [Google Scholar]
  35. Thiery, J.P. , 2002. Epithelial–mesenchymal transitions in tumour progression. Nat. Rev. Cancer 2, 442–454. [DOI] [PubMed] [Google Scholar]
  36. Thompson, E.W. , Newgreen, D.F. , Tarin, D. , 2005. Carcinoma invasion and metastasis: a role for epithelial–mesenchymal transition?. Cancer Res. 65, 5991–5995. discussion 5995 [DOI] [PubMed] [Google Scholar]
  37. Verploegen, S. , Lammers, J.W. , Koenderman, L. , Coffer, P.J. , 2000. Identification and characterization of CKLiK, a novel granulocyte Ca(++)/calmodulin-dependent kinase. Blood 96, 3215–3223. [PubMed] [Google Scholar]
  38. Verploegen, S. , Ulfman, L. , van Deutekom, H.W. , van Aalst, C. , Honing, H. , Lammers, J.W. , Koenderman, L. , Coffer, P.J. , 2005. Characterization of the role of CaMKI-like kinase (CKLiK) in human granulocyte function. Blood 106, 1076–1083. [DOI] [PubMed] [Google Scholar]
  39. Wang, P. , Kim, Y. , Pollack, J. , Narasimhan, B. , Tibshirani, R. , 2005. A method for calling gains and losses in array CGH data. Biostatistics 6, 45–58. [DOI] [PubMed] [Google Scholar]
  40. Zhao, H. , Langerod, A. , Ji, Y. , Nowels, K.W. , Nesland, J.M. , Tibshirani, R. , Bukholm, I.K. , Karesen, R. , Botstein, D. , Borresen-Dale, A.L. , 2004. Different gene expression patterns in invasive lobular and ductal carcinomas of the breast. Mol. Biol. Cell 15, 2523–2536. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Molecular Oncology are provided here courtesy of Wiley

RESOURCES