c-Kit is required for growth and survival of the cells of origin of Brca1-mutation-associated breast cancer

Article metrics


BRCA1 mutation-associated breast cancer originates in oestrogen receptor-alpha-negative (ER) progenitors in the mammary luminal epithelium. These cells also express high levels of the Kit gene and a recent study demonstrated a correlation between Brca1 loss and Kit over-expression in the mammary epithelium. However, the functional significance of c-Kit expression in the mammary gland is unknown. To address this, c-Kit and c-Kit+ mammary epithelial subsets were isolated by flow cytometry, characterised for expression of lineage-specific cell markers and functionally analysed by in vitro colony forming and in vivo transplantation assays. The results confirm that the majority of luminal ER progenitors are c-Kit+, but also that most stem cells and the differentiated cell populations are c-Kit. A subset of c-Kit+ cells with high proliferative potential was found in the luminal ER+ population, however, suggesting the existence of a distinct luminal ER+ progenitor cell type. Analysis of mouse Brca1 mammary tumours demonstrated that they expressed Kit and its downstream effector Lyn at levels comparable to the most strongly c-Kit+ luminal ER progenitors. Consistent with c-Kit being a progenitor cell marker, in vitro three-dimensional differentiation of c-Kit+ cells resulted in a loss of c-Kit expression, whereas c-Kit over-expression prevented normal differentiation in vivo. Furthermore, c-Kit was a functional marker of proliferative potential, as c-Kit inhibition by short hairpin knockdown prevented normal epithelial growth and caused cells to undergo apoptosis. Therefore, c-Kit defines distinct progenitor populations in the mammary epithelium and is critical for mammary progenitor survival and proliferation. Importantly, c-Kit is only the second mammary epithelial stem/progenitor marker to be shown to have a functional role in the mammary epithelium and the first marker to be shown to be required for progenitor cell function. The c-Kit signalling network has potential as a target for therapy and/or prevention in BRCA1-associated breast cancer.


The identification and molecular characterisation of stem, progenitor and differentiated cell hierarchies in complex tissues is key to understanding not only normal cellular homoeostasis but also the origins of cancer (Molyneux et al., 2010). Identifying the key biological processes common to tumours and the cells from which they develop will help define the pathways/networks required for tumour initiation and maintenance. Such pathways are potential targets for therapeutic and preventative intervention.

In the adult breast, the epithelium is organised into two major cell types defined by their spatial location, the luminal and basal epithelial cells. However, mammary epithelial cells can also be defined functionally. Stem cells regenerate the epithelium with high efficiency when injected into a ‘cleared’ mammary fat pad (one free of endogenous epithelium). Progenitors are cells derived from stem cells that have high in vitro colony forming and proliferative potential, but limited ability to repopulate a cleared fat pad. Terminally differentiated cells do not transplant or form colonies in vitro (Shackleton et al., 2006; Stingl et al., 2006; Asselin-Labat et al., 2007; Sleeman et al., 2007). Using flow cytometry and functional assays, stem cells have been localised to the basal cell layer, whereas progenitors are found in large numbers in the luminal population (Shackleton et al., 2006; Stingl et al., 2006; Asselin-Labat et al., 2007; Sleeman et al., 2007; Taddei et al., 2008). We have recently demonstrated that oestrogen receptor-alpha-negative (ER) luminal progenitors are the origin of BRCA1-associated breast cancers and likely of the majority of ‘basal-like’ breast cancers not linked to a family history (Molyneux et al., 2010). Understanding the molecular regulation of these cells in particular is therefore important for identifying novel approaches to breast cancer prevention and therapy.

Gene expression analysis showed that Kit was highly expressed by luminal ER progenitors in both mouse (Kendrick et al., 2008) and human (Lim et al., 2009). Furthermore, Brca1 loss in mouse mammary epithelial cells resulted in defective differentiation and increased c-Kit expression (Smart et al., 2011). However, a role for c-Kit in directly regulating mammary epithelial differentiation could not be determined. Therefore, the function of c-Kit in the mammary epithelium remains unclear.

We have analysed c-Kit-expressing subpopulations within the mammary epithelial stem, progenitor and differentiated cells. We demonstrate that the luminal ER progenitors, the cells of origin of Brca1 mutation-associated breast cancer, are strongly c-Kit+. Expression of c-Kit is also seen in a subpopulation of luminal ER+ cells. Using functional assays, we show that c-Kit expression marks progenitors with high in vitro proliferative potential, but that the majority of mammary stem cells (MaSCs) are c-Kit. We demonstrate that in a mouse model of Brca1 breast cancer that phenocopies the human disease (Molyneux et al., 2010), the tumours express Kit and its downstream effector Lyn at levels comparable to those luminal ER progenitors, which are most strongly c-Kit+. Importantly, we demonstrate that c-Kit is required for mammary epithelial cell survival. This makes the c-Kit signalling network not only a key component in the regulation of normal mammary epithelial homoeostasis, but also a potential target for therapy and prevention in BRCA1 mutation carriers and potentially other at-risk groups.


The majority of c-Kit+ mammary epithelial cells are luminal ER progenitors

To identify c-Kit+ populations and enable functional characterisation, freshly harvested primary mammary cells were stained for CD24, Sca-1, CD49f and c-Kit expression and analysed by flow cytometry. As previously, CD24+/Low basal epithelial cells and CD24+/High luminal epithelial cells were identified (Sleeman et al., 2006). Within the basal population, a CD24+/Low Sca-1 CD49fLow population (previously identified as mainly myoepithelial cells) and a CD24+/Low Sca-1 CD49fHigh population (previously shown to be highly enriched for mammary stem cells) could be recognised (Stingl et al., 2006; Sleeman et al., 2007). Within the luminal population, CD24+/High Sca-1 luminal cells (mainly ER progenitors) and CD24+/High Sca-1+ luminal cells (mainly differentiated ER+ cells) could be distinguished (Figure 1a and Supplementary Figure 1) (Sleeman et al., 2007; Molyneux et al., 2010).

Figure 1

c-Kit is predominantly expressed by luminal Sca-1 cells in the mammary epithelium. (a) Flow cytometry dot plots of primary mouse mammary epithelial cells isolated from 10 week virgin FVB mice and stained with antibodies against CD24, Sca-1, CD49f and c-Kit. Only the epithelial cells are shown (for the full gating strategy used to define the epithelial-only compartment see Supplementary Figure 1). In the upper series of plots, the c-Kit status of the luminal subpopulations was determined by first plotting CD24 against Sca-1 staining (left hand plot), to define the CD24+/High Sca-1 (Gate I; green) and CD24+/High Sca-1+ (Gate II; blue) luminal cells. Dot plots of c-Kit against CD24 staining were then drawn for each of these gated populations (middle and right hand plots). The numbers indicate the proportion of c-Kit or c-Kit+ cells as a percentage of the total cells in that gate (±s.d.; n=five independent sorts). In the lower series of plots, CD49f was plotted against CD24 (data from the same samples) to similarly define the CD24+/Low Sca-1 CD49fLow myoepithelial cells (Gate III; red) and CD24+/Low Sca-1 CD49fHigh mammary stem cell enriched population (MaSCs; Gate IV; black). c-Kit and CD24 dot plots were then drawn as before. (b) Table of frequencies of each population as a percentage of the total epithelium (±s.d.) calculated from percentages shown in (a). (c) Levels of Kit and Kitl gene expression in the different mammary epithelial subpopulations as determined by qPCR relative to comparator (the luminal Sca-1+ c-Kit+ population). Data are presented as fold expression levels±95% confidence intervals (n=three independently harvested isolates of each cell population). *Kit gene expression was undetectable in these populations in all three independent isolates. **Kit gene expression was only detected (at very low levels) in one of three isolates of the luminal Sca-1+ c-Kit population. No error bars are therefore shown for this sample.

Consistent with the transcriptomic data (Kendrick et al., 2008), the majority (up to 90%) of luminal Sca-1 progenitors were c-Kit+ (equal to 50% of all mammary epithelial cells). A small number of these (designated luminal Sca-1 c-Kit+/High) had very high levels of c-Kit expression. In contrast, only approximately 20% of the luminal Sca-1+ cells were c-Kit+ (3% of the total epithelium). The majority of basal cells were c-Kit. However, occasional c-Kit+ cells could be detected in both CD24+/Low Sca-1 CD49fLow and CD24+/Low Sca-1 CD49fHigh populations. It was unclear at this stage whether they represented minor contamination from the luminal c-Kit+ populations, a biologically relevant group of cells or a sorting artefact. These rare cells were termed basal CD49fLow c-Kit+ and basal CD49fHigh c-Kit+ to distinguish them from the majority basal CD49fLow c-Kit myoepithelial and basal CD49fHigh c-Kit mammary stem cells, respectively (Figures 1a and b).

To confirm the flow cytometry data, analysis of Kit gene expression by quantitative real-time RT–PCR (qPCR) was carried out on each mammary epithelial subpopulation. Kit was undetectable in the basal myoepithelial cells and MaSCs, but its expression could be detected in the rare basal CD49fLow c-Kit+ and basal CD49fHigh c-Kit+ cells. As expected, Kit was detected at only very low levels (and in only one of three independent isolates) in the luminal Sca-1+ c-Kit population, but was robustly expressed in the luminal Sca-1+ c-Kit+ cells. Surprisingly, however, while Kit was strongly expressed in the luminal Sca-1 c-Kit+/Low and luminal Sca-1 c-Kit+/High populations, it was also detected in the luminal Sca-1 c-Kit cells at levels comparable to the luminal Sca-1+ c-Kit+ population, suggesting post-transcriptional regulation.

In addition, levels of the c-Kit receptor ligand, Kit ligand/Stem cell factor (Kitl), were measured (Kent et al., 2008). Kitl expression could be detected in all cells; however, it was most strongly expressed by the basal populations. Given the mainly luminal pattern of c-Kit expression, this suggests that c-Kit signalling in the mammary epithelium is primarily driven by paracrine signals from the basal to the luminal cells (Figure 1c).

c-Kit+ expression defines cells with high clonogenic potential and which express markers of multiple cell lineages

Previously, luminal CD24+/High Sca-1 progenitors were found to have the highest in vitro colony forming potential of any mammary epithelial subpopulation (Sleeman et al., 2007). To determine whether colony forming potential correlated with c-Kit+ expression, the relative in vitro colony forming abilities of c-Kit+ and c-Kit cells from the mammary epithelial subpopulations were compared in a single cell per well 96-well plate assay. The results showed that in the two luminal populations, the c-Kit+ cells had significantly higher (P<0.01) colony forming ability relative to the c-Kit cells. Basal CD49fLow c-Kit+ cells also had a slight, but significant (P<0.05) higher colony forming ability relative to the other basal populations (Figure 2a). Given the scarcity of these cells, however, it was clear that the majority of in vitro colony-forming activity in the mammary epithelium was associated with luminal c-Kit+ cells.

Figure 2

c-Kit expression defines cells with progenitor features. (a) Colony forming efficiency in single-cell cloning 96-well plate assays of mammary epithelial subpopulations separated by c-Kit status. Data are presented as mean percentage colony forming ability ±s.d. (n=three independent experiments each of which analysed three 96-well plates of each cell population). *P<0.05, **P<0.01 (t-tests on Log10-transformed data). (b) Heat map of qPCR gene expression data for mammary cell subpopulations. Relative gene expression levels (n=3 independent cell preparations) over the comparator population (luminal Sca-1 c-Kit+/Low cells for all genes except Trp63 and ΔNp63, for which the MaSCs were the comparator) were used to determine relationships between the mammary epithelial subpopulations by unsupervised hierarchical clustering. All relative expression data together with 95% confidence intervals are given in Supplementary Table 1. Grey boxes indicate samples in which the gene was undetectable. (c) Fold expression±95% confidence intervals in mammary epithelial subpopulations (n=3 independent cell preparations) over the comparator population (as above) of genes associated with the basal (Krt14, Itga6 and Trp63/ΔNp63) and luminal (Krt18, Esr1) lineages (Kendrick et al., 2008). *Gene expression was undetectable in these populations in all three independent isolates. **Gene expression was only detected (at very low levels) in two of three isolates of the luminal Sca-1+ c-Kit population. No error bars are therefore shown for this sample.

The high in vitro clonogenic potential of c-Kit+ cells is consistent with a progenitor identity. Plasticity of luminal progenitor differentiation is thought to underlie the ‘basal-like’ gene expression profile observed in tumours derived from these cells (Molyneux et al., 2010). Therefore, we examined whether mammary epithelial populations showed evidence of differentiative plasticity by analysing expression of the basal marker Keratin 14 (K14), the luminal marker Keratin 18 (K18) and ER in freshly isolated subpopulations sorted as single cells on to slides. The results (Table 1) demonstrated that the c-Kit and c-Kit+ luminal Sca-1 cells, and the c-Kit+ luminal Sca-1+ cells, were weakly K14-positive (K14+/−) as well as K18-positive (K18+), suggesting potential for differentiative plasticity. The luminal Sca-1+ c-Kit cells, however, were K18+ only, consistent with terminal differentiation. Luminal markers could not be detected in the basal CD49fLow c-Kit+ cells but the basal CD49fHigh c-Kit+ cells were K14+ and also K18+/−. Both the myoepithelial cells and MaSCs were K14+ but K18. Nuclear ER staining was found in both the luminal Sca-1+ c-Kit+ and luminal Sca-1+ c-Kit cells, supporting a progenitor—differentiated cell relationship between them. However, nuclear ER staining was also found in the luminal Sca-1 c-Kit cells.

Table 1 Staining of mammary epithelial subpopulations for Keratin 14 (K14), Keratin 18 (K18) and ER (data from three independent cell preparations)

To investigate further the relationships between the populations, a qPCR-based gene expression analysis was carried out on a panel of 33 genes. The panel included genes previously shown to be specifically expressed in basal and luminal cells, as well as genes thought to be expressed uniformly throughout the mammary epithelium (Kendrick et al., 2008). Unsupervised hierarchical clustering of the data was carried out (Figure 2b; the full data set including relative expression levels and 95% confidence intervals for all genes in all populations is given in Supplementary Table 1; the expected luminal or basal expression pattern of each gene based on our previous transcriptomic and qPCR analysis is also shown) and patterns of expression of individual genes were analysed (Figure 2c). The clustering analysis divided the populations into three main branches. In one branch were the MaSCs together with the basal CD49fLow c-Kit+ and basal CD49fHigh c-Kit+ cells. This suggested that the rare basal c-Kit+ cells, despite their phenotypic differences, were in fact closely related to the MaSCs. The luminal populations were clustered on a second branch, with the two luminal Sca-1+ populations on one sub-branch and the three luminal Sca-1 populations on another. Finally, the myoepithelial cells were on a branch on their own.

Looking at individual genes in more detail, analysis of the expression patterns of Krt14, Krt18 and Esr1 largely supported the K14, K18 and ER staining on slides, although in populations which showed no staining for these markers, low levels of gene expression could always be detected, consistent with previous observations (Sleeman et al., 2007; Kendrick et al., 2008). As expected, the basal populations expressed the highest levels of Krt14, while the luminal populations expressed very low levels. Consistent with the staining of cells on slides, Krt14 expression was lowest in the luminal Sca-1+ c-Kit cells, in which K14 staining could not be detected, with slightly higher levels in the luminal Sca-1+ c-Kit+ and luminal Sca-1 c-Kit, c-Kit+/Low and c-Kit+/High populations, which were weakly K14-positive. Krt18 expression was highest in the luminal populations and lowest in the basal populations. Esr1 gene expression was highest in the luminal Sca-1+ c-Kit+, luminal Sca-1+ c-Kit and the luminal Sca-1 c-Kit cells, consistent with the ER staining. Remarkably, in the basal CD49fLow c-Kit+ cells Esr1 was expressed at high levels, despite being undetected by staining.

Consistent with the clustering data, but in contrast with their flow cytometry profile, the basal CD49fLow c-Kit+ cells expressed high levels of Itga6 (the gene coding for CD49f/α6 integrin) comparable to those of the MaSCs and basal CD49fHigh c-Kit+ cells. Furthermore, analysis of Trp63 levels with two different probes, one of which detects both the full length and ΔN splice isoforms, and the other only the ΔN isoform, demonstrated that MaSCs express mainly full length Trp63, while the myoepithelial cells likely express mainly ΔNp63 (Figure 2c). The significance of this is unclear at this time. Overall, the gene expression patterns were consistent with the basal or luminal identities of the cell populations (the expression of Esr1 in the CD49fLow c-Kit+ cells being one of the notable exceptions; Supplementary Table 1) (Kendrick et al., 2008).

These data demonstrate that the large majority of c-Kit+ cells in the mammary epithelium are found in the luminal cell layers, and primarily in the luminal Sca-1 ER populations. Furthermore, c-Kit+ mammary epithelial cells have high proliferative capacity and show plasticity of developmental potential consistent with a progenitor phenotype.

c-Kit+ mammary epithelial cells behave as progenitors in three-dimensional culture

In three-dimensional culture on Matrigel, mammary epithelial progenitors form small and simple acini with a single layer of epithelial cells. In contrast, stem cell-derived colonies on Matrigel are large and tend to be solid, complex structures (Stingl et al., 2006). To assess the response of c-Kit+ and c-Kit mammary epithelial cells to three-dimensional culture, freshly isolated populations were plated at low density, in equal numbers, on growth factor-reduced Matrigel. Within 24–48 h of being placed in culture, c-Kit+ cells began to form colonies. In contrast, colony formation by c-Kit cells was observed only after 4–5 days in culture. After 20 days in culture, c-Kit cells had generated approximately five-fold fewer colonies than c-Kit+ cells (Figure 3a). However, those colonies that did grow from the c-Kit cells were significantly larger than colonies derived from c-Kit+ cells, and tended to be solid multilayered structures. In contrast, the smaller c-Kit+ cell-derived structures tended to have a typical differentiated acinus-like morphology, consisting of a single layer of cells around a lumen (Figures 3b–d). qPCR analysis of the c-Kit+ cell-derived colonies demonstrated that they had significantly lower levels of Kit expression compared with their cells of origin. In contrast, c-Kit cell-derived colonies had significantly upregulated levels of both Kit and Kitl gene expression compared with the freshly isolated parental cells (Figure 3e). Therefore, not only do c-Kit+ cells behave as progenitors in three-dimensional culture, but a subset of c-Kit cells form colonies resembling stem cell-derived colonies. Furthermore, the data suggest that differentiation of c-Kit+ cells on Matrigel is associated with downregulation of c-Kit expression.

Figure 3

c-Kit+ cells have a distinct colony morphology in three-dimensional culture. (a) Mean (±s.d.) number of colonies per field of view in Matrigel cultures of c-Kit and c-Kit+ mammary epithelial cells. Data from three independent cell preparations after 20 days in culture (each of which had three separate fields of view in each of three separate wells counted). Identical numbers of cells were plated per well in all cultures. (b) Mean (±s.d.) area of colonies in Matrigel cultures of c-Kit and c-Kit+ mammary epithelial cells. Data from three independent cell preparations after 20 days in culture (each of which had three separate fields of view in each of three separate wells analysed). (c, d) Haematoxylin and eosin sections of c-Kit origin (c) and c-Kit+ origin (d) colonies. Bar=20 μm. (e) qPCR analysis of Kit and Kitl gene expression in c-Kit and c-Kit+ origin colonies compared with freshly isolated parental c-Kit and c-Kit+ cells. Data are mean fold expression levels over the parental populations±95% confidence intervals. Data from three independent cell preparations after 20 days in culture.

In vivo transplantation potential is highest in c-Kit basal cells

Our data suggest a model in which mammary progenitors are c-Kit+, while the c-Kit fraction of the mammary epithelium includes both terminally differentiated cells and stem cells. To provide in vivo functional evidence for this model, we determined the transplantation potential of mammary epithelial subpopulations prospectively isolated on the basis of CD24 Sca-1 CD49f and c-Kit staining (Figures 4a–f).

Figure 4

MaSCs are c-Kit. (a) Results of limited dilution transplant series for mammary epithelial cells separated on the basis of CD24, Sca-1 and c-Kit staining. The number of successful outgrowths as a fraction of the number of cleared fat pads transplanted is given. The extent to which each outgrowth filled the fat pad is indicated by the ‘pie chart’ symbols. The χ2 statistic is a test for internal consistency in the data. Non-significant values indicate that all data sets are consistent with the Poisson distribution assumed to calculate the stem cell frequency. Data are from eight independent transplant sessions. (b) Carmine stained wholemount example of a transplanted mammary epithelial outgrowth. Bar=3.5 mm. (cf) Serial sections through epithelium from a successful transplant showing haematoxylin and eosin staining (c), K14 staining (d), K18 staining (e) and ER staining (f). All outgrowths, regardless of origin, contained basal, luminal ER and luminal ER+ cells. Bars=25 μm. (g) Results of single cell transplants for MaSCs and basal CD49fHigh c-Kit+ cells. Data are from five independent transplant sessions.

Transplants of basal epithelial cells (CD24+/Low Sca-1) separated by c-Kit expression showed that the total basal c-Kit fraction was highly enriched for transplantable cells, whereas the rare basal c-Kit+ cells (which are mainly CD49fLow c-Kit+; Figure 1b) had poor transplantation activity. No transplants were observed from either the luminal Sca-1 c-Kit or the luminal Sca-1+ c-Kit populations. In contrast, when c-Kit+ cells from the luminal populations were transplanted, some outgrowths were observed (Figure 4a), especially when larger cell numbers were used. Importantly, histological analysis and immunohistochemical staining of sections through all outgrowths demonstrated that outgrowth formation was always associated with the ability to generate myoepithelial, luminal ER and luminal ER+ cells (Figures 4b–f).

Owing to their scarcity, basal CD49fHigh c-Kit+ cells could not be assessed by standard limiting dilution transplants. Therefore, to define the transplantation potential of these cells, MaSCs and basal CD49fHigh c-Kit+ cells were transplanted at one cell per fat pad into fat pads of syngeneic FVB mice. Of the 34 fat pads that received a single MaSC, 9 produced outgrowths, the highest enrichment of a mammary epithelial subpopulation for MaSCs reported to date. The basal c-Kit+ transplants produced fewer outgrowths, 4 out of 34, but this difference was non-significant (P>0.05 by χ2-test) (Figure 4g). This contrasts with the poor transplantation ability of c-Kit+ cells isolated from the total basal epithelial population (CD24+/Low Sca-1) without CD49f fractionation (Figure 4a).

These results demonstrate that basal MaSCs are c-Kit. However, rare basal CD49fHigh c-Kit+ cells that are related to the MaSCs can be isolated (Figure 2b), which have good transplantation potential and which also express the luminal marker K18 at low levels. The basal CD49fLow c-Kit+ cells, which are also apparently related to the MaSCs, have poor transplant potential and do not express K18. By contrast, the limited transplantation capacity which has been previously observed in luminal mammary epithelial cells (Sleeman et al., 2007) is a property of the luminal c-Kit+ fractions, supporting a progenitor function for these cells.

Brca1 mammary tumours have a progenitor-like phenotype

We have previously demonstrated that Brca1 mutation-associated breast cancers, and the majority of sporadic ‘basal-like’ breast cancers, develop from luminal Sca-1 ER progenitors. We also demonstrated that high expression of the Src-family kinase Lyn is associated with ‘basal-like’ tumours (Molyneux et al., 2010). As Lyn is a downstream effector of c-Kit, we examined its expression pattern in the normal mammary epithelial subpopulations we have defined by CD24 Sca-1 CD49f and c-Kit staining. Lyn expression could be detected in all populations, although it was detected most robustly and reproducibly in the luminal populations (Figure 5a). The relative expression levels of Lyn in the different luminal populations correlated with Kit expression levels (Figure 1c), while there was no such correlation in the basal populations.

Figure 5

c-Kit signalling is required for mammary epithelial progenitor growth. (a) Levels of Lyn gene expression in mammary epithelial subpopulations relative to comparator (the luminal Sca-1+ c-Kit+ population). Data are presented as fold expression levels±95% confidence intervals (n=three independently harvested isolates of each cell population). *Lyn gene expression was only detected in one of three isolates of these populations. No error bars are therefore shown. **Lyn gene expression was only detected in two of three isolates of these populations. No error bars are therefore shown. (b) Kit and Lyn gene expression in normal luminal Sca-1 c-Kit, c-Kit+/Low and c-Kit+/High populations, as well as in five luminal Sca-1-origin Blg-Cre Brca1f/fp53+/− mammary tumours (Molyneux et al., 2010). (c) Kit gene expression levels in primary mammary epithelial cells in in vitro monolayer or three-dimensional culture 2 days after transduction with control or Kit-knockdown lentiviruses. Data expressed as fold expression (mean±95% confidence intervals; data from three independent experiments) over comparator population (the control transduced cells). *=Kit expression was below the limits of detection in these samples. (d) Mammary epithelial cell growth six days after transduction with control or Kit-knockdown lentiviruses. Data expressed as percentage (mean±s.d.; data from three independent experiments) of cells surviving relative to control transduced monolayer or three-dimensional cultures. (e) Apoptosis detected by indirect TUNEL labelling (indicated by red stain) in control untreated, staurosporine treated (1 μm/ml for four hours) or transduced with control, shKit1 or shKit2 viruses (and then cultured for four days) monolayer primary mammary epithelial cells. Nuclei are stained blue with DAPI. The virus transduced cells also express GFP at low levels. Bar=40 μm. (f) Nuclear morphology of control-, shKit1- and shKit2-transduced mammary acini cultured on Matrigel. Note the disrupted acinar morphology in the knockdown cultures and pyknotic nuclei. Bar=20 μm.

Expression of both Kit and Lyn were next determined by qPCR in Blg-Cre Brca1flox/floxp53+/− mouse mammary tumours, which are derived from the luminal ER progenitors and phenocopy human BRCA1 breast cancers (Molyneux et al., 2010). Both Kit and Lyn were expressed at levels comparable to those in the luminal Sca-1 c-Kit+/High progenitors (Figure 5b) and staining of the tumours confirmed the presence of Lyn-positive neoplastic cells (data not shown). These findings were consistent with the progenitor origins of Blg-Cre Brca1flox/floxp53+/− mouse mammary tumours and support the hypothesis that these tumours have a strong mammary progenitor-like phenotype.

c-Kit activity is functionally required for progenitor growth and survival

To determine whether c-Kit activity is functionally required for progenitor activity, primary mammary cells grown in both monolayer culture and as three-dimensional acini on Matrigel were transduced with lentiviruses expressing either of two different short hairpin sequences (shKit1 and shKit2) or a control virus. Cells were harvested after 2 days to determine the extent of c-Kit knockdown and cell numbers were determined in parallel cultures after 6 days. In monolayer, cultures transduced with shKit1 expressed approximately thirty-fold less Kit mRNA than control cells. Cultures transduced with shKit2 expressed approximately six-fold less Kit mRNA than control. In three-dimensional culture, however, Kit expression was undetectable in both shKit1 and shKit2 cultures (Figure 5c). After 6 days in monolayer culture, the overall number of cells in the shKit2-transduced cultures was reduced by 10% compared with control. However, in the shKit1-transduced cultures the number of surviving cells was reduced by 80% compared with control, consistent with the extent of knockdown with this virus. On Matrigel, cell numbers were reduced by >95% compared with control in both shKit1- and shKit2-transduced cultures after 6 days (Figure 5d).

To determine whether c-Kit knockdown blocked cell proliferation or compromised cellular viability, the extent of apoptosis in treated cultures was measured. In monolayer culture, shKit1 transduction caused extensive apoptosis at levels comparable to staurosporine-treated positive controls (Figure 5e). Transduction with shKit2 also resulted in apoptosis, although at lower levels than shKit1 (Figure 5e). In three-dimensional culture, c-Kit knockdown also caused extensive apoptosis, as demonstrated by disrupted acinar morphology and presence of numerous pyknotic nuclei (Figure 5f, arrowheads).

To determine the effects of deregulated c-Kit over-expression in vivo, primary mammary epithelial cells were transduced with either a lentivirus carrying a c-Kit (GNNK+ isoform) cDNA, as well as green fluorescent protein (GFP) in a bicistronic expression cassette, or a control virus coding for GFP only (Kendrick et al., 2008) and transplanted into cleared fat pads. Flow cytometric analysis confirmed that cells transduced with the c-Kit virus had higher levels of c-Kit expression than control cells (Supplementary Figure 3a). After 8 weeks, fat pads were isolated and examined for the presence of GFP+ outgrowths. Of the 24 fat pads transplanted with cells overexpressing c-Kit in three separate experiments, only two gave rise to very small GFP+ transplants (8%). In contrast, transplantation of control cells gave rise to 16 extensive GFP+ outgrowths out of 27 fat pads (59%) transplanted (P<0.01; χ2-test) (Supplementary Figure 3b). Therefore, although c-Kit expression is a marker of mammary epithelial progenitors and is required for their survival, forced c-Kit over-expression prevents cells from contributing to outgrowths in cleared fat pads. The proper regulation of c-Kit is therefore required for normal mammary development.


The origins of the different subtypes of breast cancer can be explained, in part, by transformation of different stem and/or progenitor cell types (Dontu et al., 2003; Molyneux et al., 2007, 2010). We have shown that ER progenitors in the luminal cell layer of the mammary epithelium are the cells of origin of BRCA1-associated breast cancers (Molyneux et al., 2010). Here, we demonstrate that c-Kit is a marker of these progenitors, consistent with data from the human (Lim et al., 2009). Unlike in the human, however, a c-Kit+ luminal ER+ progenitor population was also identified. Such cells have been suggested as the origin of ER-positive breast cancer (Dontu et al., 2003; Molyneux et al., 2007; Arendt et al., 2011), but until now it has not proven possible to prospectively isolate them. Importantly, c-Kit was required for progenitor function and is therefore only the second mammary stem/progenitor marker, after β1 integrin (Taddei et al., 2008), to be shown to have a functional role in the gland.

Luminal Sca-1 ER c-Kit+ cells formed >50% of the epithelial content of the mammary gland, consistent with a report showing reciprocal staining of c-KIT and ER in the human breast (Westbury et al., 2009). c-Kit+ luminal cells expressed markers of multiple lineages, as has been reported for a subset of luminal cells in the human (Gusterson et al., 2005; Lim et al., 2009), and had high in vitro proliferative potential. When transplanted, the limited ability to form small outgrowths in cleared fat pads that we previously observed in the luminal epithelial populations (Sleeman et al., 2007) was found to be a property of c-Kit+ cells. Together these results suggest that c-Kit expression principally marks luminal progenitors with high proliferative potential and the capacity for multilineage differentiation. This is a huge reservoir of potential proliferative capacity in a tissue in which there is normally little proliferation (Clarke et al., 1997; Fernandez-Gonzalez et al., 2009). This potential is required, however, during the massive growth stimulus, which occurs during pregnancy, when epithelial cell numbers can increase up to 20-fold (Kordon and Smith, 1998). Importantly, the multilineage differentiation potential of these luminal progenitors is consistent with their ability to give rise to tumours with basal features (Molyneux et al., 2010).

The large majority of basal cells, both CD49fLow myoepithelial cells and CD49fHigh MaSCs, were c-Kit. However, rare c-Kit+ cells were observed in both the CD49fLow and CD49fHigh populations. These rare cells were too scarce to be confidently classed as distinct populations, however, analysis of staining for K14, K18 and ER, as well as gene expression analysis suggested that while they did have similarities to other mammary epithelial populations, they also had their own unique features. Gene expression analysis confirmed that both the CD49fLow and CD49fHigh c-Kit+ cells expressed Kit, whereas Kit was undetectable in the majority of basal cells. Their strong K14 staining argued against these rare cells being contaminants from the luminal c-Kit+ populations. The basal CD49fLow c-Kit+ cells had similarities to myoepithelial cells, including being CD49fLow, K14+ and K18. However, although they were CD49fLow, they expressed Itga6 at levels comparable to CD49fHigh cells. They also expressed luminal associated genes, such as Esr1, although they did not stain for ER. The basal CD49fHigh c-Kit+ cells had similarities to MaSCs, being CD49fHigh and K14+, as well as to luminal cells, being weakly K18+. Crucially, however, gene expression analysis and clustering demonstrated that both the basal CD49fLow c-Kit+ and basal CD49fHigh c-Kit+ cells were in fact most closely related to the MaSCs. This suggests that the occasional c-Kit+ basal cells may be variant MaSCs undergoing ‘lineage priming’, in which stem/progenitor cells express genes associated with their differentiated daughter populations (Huang et al., 2007; Mansson et al., 2007). Such cells are unlikely to comprise a stable population, but will revert to a canonical stem cell state if they do not receive signals that promote their differentiation.

The identity of the luminal Sca-1 c-Kit cells is more difficult to interpret. These cells expressed levels of the Kit gene comparable to other Kit+ populations (Figure 1c), but were not significantly contaminated by luminal Sca-1 c-Kit+ ER cells as they were ER+ (Table 1). Furthermore, they clustered with the luminal Sca-1 c-Kit+ ER cells (Figure 2b), but also expressed genes in common with the luminal Sca-1+ ER+ cells, such as Prom1 and Prlr (Supplementary Table 1). We suggest that they may represent an alternative route for generating ER+ cells from luminal Sca-1 progenitors, rather than from the luminal Sca-1+ ER+ progenitor population. This is supported by lineage tracing experiments, which suggest that a small minority of luminal Sca-1+ cells are indeed derived from the luminal Sca-1 population and not from basal stem cells (Molyneux et al., 2010). A potential mammary epithelial cell hierarchy based on these data is shown in Supplementary Figure 3a.

Not only was c-Kit a marker of progenitors, but also its inhibition in vitro by shRNA knockdown resulted in their apoptosis. The effects of c-Kit loss of function in vivo are more difficult to determine. Homozygotes for naturally occurring mutations in Kit and Kitl (the dominant white-spotting and steel mutants of mice, respectively) are often lethal and those that do survive are sterile and anaemic (Baxter et al., 2004). In W-sash mice, which carry a hypomorphic Kit allele, mammary ductal outgrowth but not differentiation were impaired (Lilla and Werb, 2010). These mice had decreased mammary epithelial proliferation, however, it was not possible to distinguish cell autonomous from non-cell autonomous (such as mast cell depletion) effects of c-Kit loss. Our findings confirm a cell-autonomous role for c-Kit in the mammary epithelium although it may be that in vivo, interactions with the microenvironment alter the response to c-Kit inhibition in proliferating mammary cells from apoptosis to proliferative arrest.

Forced over-expression of c-Kit was also deleterious, as it blocked the ability of primary cells to form outgrowths in cleared mammary fat pads. It may be that forced c-Kit expression prevented differentiation and this stopped normal outgrowths from developing (for example, by preventing the appearance of functionally mature luminal Sca-1+ ER+ cells required for the paracrine interactions necessary for ductal extension) (Mallepell et al., 2006). Consistent with this, downregulation of c-Kit expression was observed when c-Kit+ cells were cultured under conditions that generate an in vivo-like morphology. These data therefore suggest that c-Kit activity must be downregulated to allow normal differentiation.

Signalling pathways downstream of c-Kit are complex (Roskoski, 2005a, 2005b). This complexity is increased by the existence of two different splice variants of c-Kit, GNNK and GNNK+. Both are thought to be present in all cells in which c-Kit is expressed (Hayashi et al., 1991; Reith et al., 1991). The two isoforms have different signalling activities, although these appear to be cell-type specific and possibly context dependent. In NIH3T3 cells (which have low levels of GAB2), both isoforms activate PI3K/Akt signalling but the GNNK variant also strongly activates MAPK signalling in a Src-dependent manner (Caruana et al., 1999; Voytyuk et al., 2003; Young et al., 2007). However, in haematopoietic cells, PI3K activity is at least partly dependent on Src-family kinase mediated phosphorylation of GAB2 and this can lead to isoform-specific c-Kit activation of PI3K signalling (Yu et al., 2006; Sun et al., 2008). In NIH3T3 cells both isoforms promote resistance to anoikis, but GNNK also strongly promotes loss of contact inhibition and increases tumourigenicity (Caruana et al., 1999). It is unclear at this time, which isoform predominates in mammary epithelial cells or whether the balance of isoforms is different between the subpopulations. We used the GNNK+ variant, which is less strongly transforming, for our over-expression studies.

Kit ligand/SCF is also found in two isoforms which differ by the presence or absence of exon 6. It localises to the cell membrane, where the isoform that includes exon 6 can undergo proteolytic cleavage of the extracellular domain to release soluble SCF. The isoform that lacks exon 6 cannot be cleaved and remains membrane bound (Miyazawa et al., 1995; Gommerman et al., 2000; Driessen et al., 2003). The soluble form of SCF rapidly activates c-Kit but the signalling activity quickly decays owing to receptor internalisation. The membrane bound form gives a more persistent signalling activity, possibly owing to increased stability of the receptor-ligand complex and reduced internalisation (Miyazawa et al., 1995; Gommerman et al., 1997). In vivo, only the membrane bound form supports survival of long-term haematopoietic progenitors (Gommerman et al., 2000; Driessen et al., 2003). The Steel-Dickie mutation in mice generates an SCF protein lacking both transmembrane and cytoplasmic domains and results in production of only soluble SCF (Brannan et al., 1991), supporting the importance of the membrane bound form in a number of biological processes. It is not yet clear which form predominates in the mammary epithelium.

Anti-c-KIT therapy has had limited success in human breast cancer treatment (Barton et al., 2006; Chew et al., 2006; Haley et al., 2007; Waterhouse et al., 2008), suggesting that in transformed mammary cells the c-KIT signalling network may be activated downstream of the receptor. Consistent with this is the strong association of basal-like breast cancers with high expression of the Src-family kinase member Lyn (Molyneux et al., 2010), a transducer of c-Kit signalling (Linnekin et al., 1997). Previous transcriptomic analysis suggested that Lyn is strongly expressed in the normal luminal ER progenitors (Kendrick et al., 2008). Here, analysis of Lyn expression in the mammary subpopulations by qPCR demonstrated that in the luminal cells, but not the basal cells, the relative levels of Kit and Lyn expression had a very similar pattern. Although correlative, these data support a role for Lyn as a transducer of c-Kit signalling in the mammary luminal epithelium. Further work will be needed to delineate the complexities of c-Kit signalling in the different luminal progenitor populations.

Kit and Lyn were expressed in mouse Brca1 tumours derived from the luminal Sca-1 ER progenitor population (Molyneux et al., 2010) at levels comparable to those seen in the c-Kit+/High fraction of that population. This suggests that Brca1 tumours are blocked in a progenitor-like state. The expression of basal markers at low levels in the normal c-Kit+ luminal progenitors (Table 1) and the presence in the tumours of neoplastic cells with both luminal and basal features, would be consistent with such a block. Lyn is a potential regulator of Brca1, acting via Akt (Linnekin et al., 1997; Altiok et al., 1999; Shivakrupa and Linnekin, 2005). Brca1 is itself a potential regulator of Lyn, acting via CDC2 (Pathan et al., 1996; Wang et al., 1997). Furthermore, the downstream transducer of c-Kit signalling, Stat5a, is negatively regulated by Brca1 (Brizzi et al., 1999; Vidarsson et al., 2002), while Stat1 and Jak2 are positively regulated by both Brca1 and c-Kit (Deberry et al., 1997; Chin et al., 1998; Ouchi et al., 2000; Gao et al., 2001). Taken together, these studies suggest a model in which Brca1, c-Kit and Lyn function in a feedback loop that modulates Jak/Stat survival signalling (Supplementary Figure 3b). Although this remains to be directly tested in the mammary epithelium, in a recent study Brca1 loss in mouse mammary epithelial cells was found to result in defective differentiation and increased Kit expression (Smart et al., 2011), consistent with our data on Kit and Lyn expression in mouse Brca1 tumours.

Our data show that the c-Kit signalling pathway is required for the normal function of mammary epithelial progenitors. Regulation of this pathway by Brca1 activity may explain why these cells are sensitive to transformation by Brca1 loss, while over-expression of Lyn in basal-like breast cancers may explain why anti-c-Kit therapy has had poor results in breast cancer. Understanding how c-Kit and Brca1 regulate each other has important implications for therapeutic and preventative approaches in basal-like breast cancer.

Materials and methods

Preparation of mammary epithelial cells

All animal work was carried out under UK Home Office project and personal licences following local ethical approval and in accordance with local and national guidelines. Single cells were prepared from fourth mammary fat pads of 10–week-old virgin female FVB mice as described (Smalley, 2010). Cell suspensions at 106 cells/ml were stained with combinations of anti-CD24-FITC (1.0 μg/ml; BD Biosciences, Oxford, UK), anti-Sca-1-APC (1.0 μg/ml; eBioscience, Hatfield, UK), anti-CD45-PE-Cy7 (1.0 μg/ml; BD Biosciences), anti-CD49f-PE-Cy5 (5.0 μl/ml; BD Biosciences) and anti-c-Kit-PE (1.0 μg/ml; BD Biosciences). Cells were sorted on a FACSAria (Becton Dickenson, Oxford, UK) and dead cells, CD45+ leucocytes and non-single cells were excluded as shown in Supplementary Figure 1 (Sleeman et al., 2006). Cells incubated in non-specific IgG were used to set the limits that defined negative and positive staining for each antibody. Within the CD24+ population, two separate cell populations were annotated as CD24+/High and CD24+/Low as previously described (Sleeman et al., 2006). Within the basal CD49f+ population CD49fHigh mammary stem cell enriched cells were defined by previous understanding and with reference to the outermost contours of a 5% interval linear density contour plot as described (Supplementary Figure 1) (Britt et al., 2009). For consistency, c-Kit+/High cells within the luminal Sca-1 c-Kit+ population were defined with reference to the outermost contours of a 5% interval linear density contour plot.

Gene expression analysis by quantitative real-time rtPCR

Freshly sorted normal cells were resuspended in RLT buffer (Qiagen, Crawley, West Sussex, UK) and stored at −80 °C until required for RNA extraction. RNA was isolated from cultured cells with Trizol (Invitrogen, Paisley, UK). qPCR reactions were performed largely as previously described (Kendrick et al., 2008) using TAQMAN Assays-on-Demand probes (Supplementary Table 3). Actb (β-actin) was used as an endogenous control and results calculated using the Δ-ΔCt method. Data were expressed as the mean fold gene expression difference in three independently isolated cell preparations over a comparator sample with 95% confidence intervals.

The Blg-Cre Brca1flox/floxp53+/− mouse model and the tumours it generates have been fully described previously (Molyneux et al., 2010). For analysis of Kit and Lyn gene expression by qPCR, archived RNA samples were used. Previous analysis of these samples with a probe targeting Brca1 exon 22 (contained within the floxed region) had demonstrated a very high contribution of neoplastic cells and minimal normal tissue contamination (Molyneux et al., 2010).

For unsupervised hierarchical clustering of qPCR gene expression, data on epithelial subpopulations and median-centred data for mean fold expression levels (n=three independent samples) over the comparator population were uploaded into CLADIST (http://www.brcabase.icr.ac.uk/software/cladist.jsp). Average linkage clustering with a Euclidean distance similarity metric was used.

Lentivirus production

Virus over-expressing the GNNK+ isoform of c-Kit was generated using the Tronolab (http://www.tronolab.epfl.ch/) pWPI system (Kendrick et al., 2008). Oligonucleotide pairs for shKit1 (CACCGCCTTCCCTTGTTTGGCAAAGCGAACTTTGCCAAACAAGGGAAGGC and AAAAGCCTTCCCTTGTTTGGCAAAGTTCGCTTTGCCAAACAAGGGAAGGC) and shKit2 (CACCGCAGAAGCGGCACTTTATAAGCGAACTTATAAAGTGCCGCTTCTGC and AAAAGCAGAAGCGGCACTTTATAAGTTCGCTTATAAAGTGCCGCTTCTGC) were ligated into pENTR/U6 Gateway system entry vector (Invitrogen) according to manufacturer's instructions. Hairpin sequences were verified and then transferred, together with the U6 promoter, into a Gateway-modified pSEW lentiviral vector backbone (Vafaizadeh et al., 2010) by LR reaction (Invitrogen).

Viral supernatants were generated as described (Kendrick et al., 2008). shKit1 and shKit2 viruses were concentrated with Lenti-X concentrator before use (Clontech Europe, Saint-Germain-en-Laye, France). Supernatants were stored at −80 °C until use. Primary mammary cells were transduced with virus using the suspension method as described (Kendrick et al., 2008). Relative titres of viral supernatants were determined by transducing NIH 3T3 cells and analysing the number of GFP+ cells after 6 days. The relative titre of the pWPI-c-Kit over-expression virus was found to be 60% of that of the control. The relative titres of the shKit1 and shKit2 viruses were found to be 181 and 127% that of the control (which was not concentrated before use), respectively.

Immunofluorescence staining

Immunophenotyping of cell populations sorted on to slides and of mouse mammary epithelial cell clones grown on glass coverslips was carried out as described (Sleeman et al., 2007) with the exception that cells sorted on to slides were incubated with primary antibody either for 40 min at room temperature or overnight at 4 °C. Antibodies used were LLOO2 for K14 (0.26 μg/ml; Abcam, Cambridge, UK), Ks18.04 for K8/18 (1:2 dilution; Progen Biotechnik, Heidelberg, Germany) and ID5 for ERα (9.9 μg/ml; DAKO, Ely, UK). Negative controls with isotype-matched clones gave no staining.

Staining intensities of K14 and K18 were assessed by comparison with unstained controls using identical settings on a Leica TCS SP2 confocal microscope (Leica Microsystems, Milton Keynes, UK). Images captured with Leica confocal software were used for scoring staining. The following criteria were used to score cells as negative, weakly positive or positively stained. In ‘negative’ populations, no staining could be distinguished above background using either protocol. In populations consisting of cells with ‘weak’ staining, such staining could only be distinguished above the background in cells stained overnight at 4 °C. In samples stained at room temperature for 40 min, staining could not be distinguished above background. In populations scored as ‘positive’, the staining was still distinct even with the shorter incubation time.

For Lyn staining in tumour samples, fresh sections were cut from archived formalin-fixed and paraffin-embedded tissue. Dewaxed and re-hydrated sections underwent antigen retrieval in citrate buffer (0.01 M, pH 6.0) for 18 min in a microwave (900W) before blocking in 0.05% hydrogen peroxide solution then protein block (Dako UK Ltd, Cambridgeshire, UK) for 30 min. The Lyn primary antibody (ab53141, rabbit polyclonal, Abcam; diluted 1:200) was incubated overnight at 4 °C. Detection was carried out using the ABC method (ABC, Vector Laboratories, Peterborough, UK). Negative controls were performed using the same protocols with substitution of the primary antibody with IgG-matched controls.

In vitro analysis of mammary progenitors

Analysis of in vitro colony forming potential of single cells was carried out in a 96-well plate format as described (Sleeman et al., 2007). Single cellularity of sorted populations was ensured using time-of-flight gating on both forward and side scatter. Statistical comparisons were carried out using t-tests of Log10 transformed data.

For three-dimensional culture of mammary epithelial cells, 24-well plates were prepared with 200 μl growth factor reduced Matrigel (BD Biosciences) per well. Wells were seeded at 104 mammary epithelial cells per well in 1 ml/well of 1:1 DMEM:Ham's F12 medium (Invitrogen) with 10% FCS (PAA, Yeovil, UK), 5 ug/ml insulin (Sigma, Poole, UK), 10 ng/ml cholera toxin (Sigma) and 10 ng/ml epidermal growth factor (Sigma). Cultures were maintained at 37 °C in a 5%O2 atmosphere in a Heraeus BB6060 low oxygen incubator (Thermo Fisher Scientific, Leicestershire, UK) for up to 20 days. Numbers and sizes of colonies were determined from three fields of view of photomicrographs of cultures taken at identical magnifications. Colony areas were determined by image analysis using ImageJ image analysis software (http://www.rsbweb.nih.gov/ij/) or Adobe Photoshop CS4 Extended (Adobe Systems, Uxbridge, UK). Owing to the optical properties of Matrigel cultures and the necessity to change focal planes, when examining the centre as opposed to the edges of wells, it was not possible to take a single photograph of an entire well for automated analysis.

For haematoxylin and eosin stained sections of three-dimensional colonies, gels were fixed overnight in 4% paraformaldehyde in PBS at room temperature. The fixed Matrigel containing the colonies was then removed from the wells and subsequently processed by standard histological techniques.

Mammary epithelial cell transplantation

Transplantation of freshly isolated primary mouse mammary epithelial cells at limiting dilutions into cleared fats pads was carried out as described (Sleeman et al., 2007; Britt et al., 2009). For single cell transplants, freshly isolated cells were directly sorted into the wells of low-volume (10 μl) Terasaki plates using the ‘single cell’ purity protocol. The presence of a cell was visually confirmed and then the full volume of each well injected. Fat pads were wholemounted, carmine stained and analysed 8 weeks after transplant as described (Sleeman et al., 2007; Britt et al., 2009). The χ2-test comparing the expected versus the actual number of outgrowths in transplanted fat pads was calculated using Excel.

Survival assays

Cell viability in monolayer cultures was quantified using the CellTitre Glo cell viability reagent (Promega, Southampton, Hampshire, UK), according to the manufacturer's instructions. For three-dimensional culture, the reagent was added directly to 24-well plates containing cells in Matrigel following media removal. Plates were left to equilibrate at room temperature for 20 min before lysates being transferred to 96-well plates. Assay plates were read in a luminescence plate reader (PerkinElmer, Cambridge, Cambridgeshire, UK).

Quantification of apoptotic cells

Quantification of apoptosis in monolayer cultures was carried out using the ApopTag Red indirect TUNEL labelling kit (Millipore, Watford, UK). Following treatment, both floating and adherent cell fractions were combined as cytospins and then analysed for an overall assessment of apoptotic fraction. Staurosporine-treated cells (1 μl/ml for four hours) were used as a positive control. Technical difficulties in recovering cells from three-dimensional Matrigel cultures meant the presence of apoptosis in these samples had to be determined by direct visualisation of nuclear morphology. Eight-well chamber slides (LabTek, Nunc, Roskilde, Denmark) were prepared with 40 μl of Matrigel (BD Biosciences) per well and seeded with 8000 cells per well. Cells were fixed in situ in 4% paraformaldehyde after 2 days, stained with DAPI and analysed on a Leica TCS SP2 confocal microscope using the immersion lens. Apoptosis was determined by the assessment of nuclear morphology.


  1. Altiok S, Batt D, Altiok N, Papautsky A, Downward J, Roberts TM et al. (1999). Heregulin induces phosphorylation of BRCA1 through phosphatidylinositol 3-Kinase/AKT in breast cancer cells. J Biol Chem 274: 32274–32278.

  2. Arendt LM, Rugowski DE, Grafwallner-Huseth TA, Garcia-Barchino MJ, Rui H, Schuler LA . (2011). Prolactin-induced mouse mammary carcinomas model estrogen resistant luminal breast cancer. Breast Cancer Res 13: R11.

  3. Asselin-Labat ML, Sutherland KD, Barker H, Thomas R, Shackleton M, Forrest NC et al. (2007). Gata-3 is an essential regulator of mammary-gland morphogenesis and luminal-cell differentiation. Nat Cell Biol 9: 201–209.

  4. Barton J, Liggett W, Mainwaring M, Hainsworth JD, Simons L, Spigel DR et al. (2006). Phase II pilot trial of imatinib mesylate with weekly docetaxel in metastatic breast cancer. J Clin Oncol 24: 10716.

  5. Baxter LL, Hou L, Loftus SK, Pavan WJ . (2004). Spotlight on spotted mice: a review of white spotting mouse mutants and associated human pigmentation disorders. Pigment Cell Res 17: 215–224.

  6. Brannan CI, Lyman SD, Williams DE, Eisenman J, Anderson DM, Cosman D et al. (1991). Steel-Dickie mutation encodes a c-kit ligand lacking transmembrane and cytoplasmic domains. Proc Natl Acad Sci USA 88: 4671–4674.

  7. Britt KL, Kendrick H, Regan JL, Molyneux G, Magnay FA, Ashworth A et al. (2009). Pregnancy in the mature adult mouse does not alter the proportion of mammary epithelial stem/progenitor cells. Breast Cancer Res 11: R20.

  8. Brizzi MF, Dentelli P, Rosso A, Yarden Y, Pegoraro L . (1999). STAT protein recruitment and activation in c-Kit deletion mutants. J Biol Chem 274: 16965–16972.

  9. Caruana G, Cambareri AC, Ashman LK . (1999). Isoforms of c-KIT differ in activation of signalling pathways and transformation of NIH3T3 fibroblasts. Oncogene 18: 5573–5581.

  10. Chew HK, Barlow W, Albain K, Lew D, Budd T, Allen G et al. (2006). SWOG 0338: A phase II trial of imatinib mesylate in combination with capecitabine in metastatic breast cancer. J Clin Oncol 24: 10529.

  11. Chin H, Arai A, Wakao H, Kamiyama R, Miyasaka N, Miura O . (1998). Lyn physically associates with the erythropoietin receptor and may play a role in activation of the Stat5 pathway. Blood 91: 3734–3745.

  12. Clarke RB, Howell A, Potten CS, Anderson E . (1997). Dissociation between steroid receptor expression and cell proliferation in the human breast. Cancer Res 57: 4987–4991.

  13. Deberry C, Mou S, Linnekin D . (1997). Stat1 associates with c-kit and is activated in response to stem cell factor. Biochem J 327 (Part 1): 73–80.

  14. Dontu G, Al-Hajj M, Abdallah WM, Clarke MF, Wicha MS . (2003). Stem cells in normal breast development and breast cancer. Cell Prolif 36 (Suppl 1): 59–72.

  15. Driessen RL, Johnston HM, Nilsson SK . (2003). Membrane-bound stem cell factor is a key regulator in the initial lodgment of stem cells within the endosteal marrow region. Exp Hematol 31: 1284–1291.

  16. Fernandez-Gonzalez R, Illa-Bochaca I, Welm B, Fleish MC, Werb Z, Ortiz-deSolorzano C et al. (2009). Mapping mammary gland architecture using multi-scale in situ analysis. Integr Biol 1: 80–89.

  17. Gao B, Shen X, Kunos G, Meng Q, Goldberg ID, Rosen EM et al. (2001). Constitutive activation of JAK-STAT3 signaling by BRCA1 in human prostate cancer cells. FEBS Lett 488: 179–184.

  18. Gommerman JL, Rottapel R, Berger SA . (1997). Phosphatidylinositol 3-kinase and Ca2+ influx dependence for ligand-stimulated internalization of the c-Kit receptor. J Biol Chem 272: 30519–30525.

  19. Gommerman JL, Sittaro D, Klebasz NZ, Williams DA, Berger SA . (2000). Differential stimulation of c-Kit mutants by membrane-bound and soluble Steel Factor correlates with leukemic potential. Blood 96: 3734–3742.

  20. Gusterson BA, Ross DT, Heath VJ, Stein T . (2005). Basal cytokeratins and their relationship to the cellular origin and functional classification of breast cancer. Breast Cancer Res 7: 143–148.

  21. Haley BB, Ashfaq R, DeHaas M, Ramaswami A, Sikder K, Tripathy D . (2007). A phase I/II study of imatinib and docetaxel as neoadjuvant therapy in locally advanced breast cancer. J Clin Oncol 25: 11039.

  22. Hayashi S, Kunisada T, Ogawa M, Yamaguchi K, Nishikawa S . (1991). Exon skipping by mutation of an authentic splice site of c-kit gene in W/W mouse. Nucleic Acids Res 19: 1267–1271.

  23. Huang S, Guo YP, May G, Enver T . (2007). Bifurcation dynamics in lineage-commitment in bipotent progenitor cells. Dev Biol 305: 695–713.

  24. Kendrick H, Regan JL, Magnay FA, Grigoriadis A, Mitsopoulos C, Zvelebil M et al. (2008). Transcriptome analysis of mammary epithelial subpopulations identifies novel determinants of lineage commitment and cell fate. BMC Genomics 9: 591.

  25. Kent D, Copley M, Benz C, Dykstra B, Bowie M, Eaves C . (2008). Regulation of hematopoietic stem cells by the steel factor/KIT signaling pathway. Clin Cancer Res 14: 1926–1930.

  26. Kordon EC, Smith GH . (1998). An entire functional mammary gland may comprise the progeny from a single cell. Development 125: 1921–1930.

  27. Lilla JN, Werb Z . (2010). Mast cells contribute to the stromal microenvironment in mammary gland branching morphogenesis. Dev Biol 337: 124–133.

  28. Lim E, Vaillant F, Wu D, Forrest NC, Pal B, Hart AH et al. (2009). Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 15: 907–913.

  29. Linnekin D, DeBerry CS, Mou S . (1997). Lyn associates with the juxtamembrane region of c-Kit and is activated by stem cell factor in hematopoietic cell lines and normal progenitor cells. J Biol Chem 272: 27450–27455.

  30. Mallepell S, Krust A, Chambon P, Brisken C . (2006). Paracrine signaling through the epithelial estrogen receptor alpha is required for proliferation and morphogenesis in the mammary gland. Proc Natl Acad Sci USA 103: 2196–2201.

  31. Mansson R, Hultquist A, Luc S, Yang L, Anderson K, Kharazi S et al. (2007). Molecular evidence for hierarchical transcriptional lineage priming in fetal and adult stem cells and multipotent progenitors. Immunity 26: 407–419.

  32. Miyazawa K, Williams DA, Gotoh A, Nishimaki J, Broxmeyer HE, Toyama K . (1995). Membrane-bound Steel factor induces more persistent tyrosine kinase activation and longer life span of c-kit gene-encoded protein than its soluble form. Blood 85: 641–649.

  33. Molyneux G, Geyer FC, Magnay FA, McCarthy A, Kendrick H, Natrajan R et al. (2010). BRCA1 Basal-like Breast Cancers Originate from Luminal Epithelial Progenitors and Not from Basal Stem Cells. Cell Stem Cell 7: 403–417.

  34. Molyneux G, Regan J, Smalley MJ . (2007). Mammary stem cells and breast cancer. Cell Mol Life Sci 64: 3248–3260.

  35. Ouchi T, Lee SW, Ouchi M, Aaronson SA, Horvath CM . (2000). Collaboration of signal transducer and activator of transcription 1 (STAT1) and BRCA1 in differential regulation of IFN-gamma target genes. Proc Natl Acad Sci USA 97: 5208–5213.

  36. Pathan NI, Geahlen RL, Harrison ML . (1996). The protein-tyrosine kinase Lck associates with and is phosphorylated by Cdc2. J Biol Chem 271: 27517–27523.

  37. Reith AD, Ellis C, Lyman SD, Anderson DM, Williams DE, Bernstein A et al. (1991). Signal transduction by normal isoforms and W mutant variants of the Kit receptor tyrosine kinase. EMBO J 10: 2451–2459.

  38. Roskoski Jr R . (2005a). Structure and regulation of Kit protein-tyrosine kinase—the stem cell factor receptor. Biochem Biophys Res Commun 338: 1307–1315.

  39. Roskoski Jr R . (2005b). Signaling by Kit protein-tyrosine kinase—the stem cell factor receptor. Biochem Biophys Res Commun 337: 1–13.

  40. Shackleton M, Vaillant F, Simpson KJ, Stingl J, Smyth GK, Asselin-Labat ML et al. (2006). Generation of a functional mammary gland from a single stem cell. Nature 439: 84–88.

  41. Shivakrupa R, Linnekin D . (2005). Lyn contributes to regulation of multiple Kit-dependent signaling pathways in murine bone marrow mast cells. Cell Signal 17: 103–109.

  42. Sleeman KE, Kendrick H, Robertson D, Isacke CM, Ashworth A, Smalley MJ . (2007). Dissociation of estrogen receptor expression and in vivo stem cell activity in the mammary gland. J Cell Biol 176: 19–26.

  43. Sleeman KE, Kendrick H, Ashworth A, Isacke CM, Smalley MJ . (2006). CD24 staining of mouse mammary gland cells defines luminal epithelial, myoepithelial/basal and non-epithelial cells. Breast Cancer Res 8: R7.

  44. Smalley MJ . (2010). Isolation, culture and analysis of mouse mammary epithelial cells. In: Ward A, Tosh D (eds). Methods in Molecular Biology. Springer: Berlin, pp 139–170.

  45. Smart CE, Wronski A, French JD, Edwards SL, Asselin-Labat ML, Waddell N et al. (2011). Analysis of Brca1-deficient mouse mammary glands reveals reciprocal regulation of Brca1 and c-kit. Oncogene 30: 1597–1607.

  46. Stingl J, Eirew P, Ricketson I, Shackleton M, Vaillant F, Choi D et al. (2006). Purification and unique properties of mammary epithelial stem cells. Nature 439: 993–997.

  47. Sun J, Pedersen M, Ronnstrand L . (2008). Gab2 is involved in differential phosphoinositide 3-kinase signaling by two splice forms of c-Kit. J Biol Chem 283: 27444–27451.

  48. Taddei I, Deugnier MA, Faraldo MM, Petit V, Bouvard D, Medina D et al. (2008). Beta1 integrin deletion from the basal compartment of the mammary epithelium affects stem cells. Nat Cell Biol 10: 716–722.

  49. Vafaizadeh V, Klemmt P, Brendel C, Weber K, Doebele C, Britt K et al. (2010). Mammary epithelial reconstitution with gene-modified stem cells assigns roles to Stat5 in luminal alveolar cell fate decisions, differentiation, involution, and mammary tumor formation. Stem Cells 28: 928–938.

  50. Vidarsson H, Mikaelsdottir EK, Rafnar T, Bertwistle D, Ashworth A, Eyfjord JE et al. (2002). BRCA1 and BRCA2 bind Stat5a and suppress its transcriptional activity. FEBS Lett 532: 247–252.

  51. Voytyuk O, Lennartsson J, Mogi A, Caruana G, Courtneidge S, Ashman LK et al. (2003). Src family kinases are involved in the differential signaling from two splice forms of c-Kit. J Biol Chem 278: 9159–9166.

  52. Wang H, Shao N, Ding QM, Cui J, Reddy ES, Rao VN . (1997). BRCA1 proteins are transported to the nucleus in the absence of serum and splice variants BRCA1a, BRCA1b are tyrosine phosphoproteins that associate with E2F, cyclins and cyclin dependent kinases. Oncogene 15: 143–157.

  53. Waterhouse DM, Mainwaring M, Barton J, Webb C, Markus TM, Spigel DR et al. (2008). Phase II pilot results of imatinib mesylate with weekly docetaxel in metastatic breast cancer. J Clin Oncol 26: May 20 Suppl; abstr 1090.

  54. Westbury CB, Reis-Filho JS, Dexter T, Mahler-Araujo B, Fenwick K, Iravani M et al. (2009). Genome-wide transcriptomic profiling of microdissected human breast tissue reveals differential expression of KIT (c-Kit, CD117) and oestrogen receptor-alpha (ERalpha) in response to therapeutic radiation. J Pathol 219: 131–140.

  55. Young SM, Cambareri AC, Odell A, Geary SM, Ashman LK . (2007). Early myeloid cells expressing c-KIT isoforms differ in signal transduction, survival and chemotactic responses to Stem Cell Factor. Cell Signal 19: 2572–2581.

  56. Yu M, Luo J, Yang W, Wang Y, Mizuki M, Kanakura Y et al. (2006). The scaffolding adapter Gab2, via Shp-2, regulates kit-evoked mast cell proliferation by activating the Rac/JNK pathway. J Biol Chem 281: 28615–28626.

Download references


We thank Eric So for the c-Kit cDNA, Fredrik Wallberg for help with flow cytometry analysis and Jorge Reis-Filho and Arno Gauthier for assistance with assessment of Lyn staining on mouse tumours. This work was funded by Breakthrough Breast Cancer. We acknowledge NHS funding to the NIHR Biomedical Research Centre. This work was funded by Breakthrough Breast Cancer. We acknowledge NHS funding to the NIHR Biomedical Research Centre.

Author information

Correspondence to M J Smalley.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Regan, J., Kendrick, H., Magnay, F. et al. c-Kit is required for growth and survival of the cells of origin of Brca1-mutation-associated breast cancer. Oncogene 31, 869–883 (2012) doi:10.1038/onc.2011.289

Download citation


  • Brca1
  • c-Kit
  • mammary
  • progenitor

Further reading