The Brn-3b POU family transcription factor regulates the ... - Nature

Oncogene (2001) 20, 4961 ± 4971 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

The Brn-3b POU family transcription factor regulates the cellular growth, proliferation, and anchorage dependence of MCF7 human breast cancer cells Jonathon Hancock Dennis1, Vishwanie Budhram-Mahadeo1 and David Seymour Latchman*,1 1

Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK

The Brn-3b POU domain containing transcription factor is expressed in the developing sensory nervous system as well as in epithelial cells of the breast, cervix, and testes. Brn-3b functionally interacts with the estrogen receptor (ER) and in association with the ER, regulates transcription from estrogen responsive genes. In addition, Brn-3b expression is elevated in breast tumours compared to levels in normal mammary cells. To explore the role of Brn-3b in breast cancer, we established stable cell lines derived from the MCF7 human breast cancer cell line which had been transfected with Brn-3b sense or anti-sense constructs. The Brn-3b over-expressing cell lines exhibited increased growth rate, reached con¯uence at a higher saturation density, had higher proliferative activity, and an enhanced ability to form colonies in soft agar when compared to the control empty vector transfected cells. Likewise, the Brn-3b anti-sense cell lines showed reduced cellular growth and proliferation, reached con¯uence at a lower density, and exhibited a decreased ability to form colonies in soft agar when compared to the vector controls. Five to ten per cent of the Brn-3b over-expressing cells exhibited a severely altered morphology characterized by reduced adherence to tissue culture plastic, increased cell size, and a vacuolar cell shape. These results thus further indicate a role for the Brn-3b transcription factor in regulating mammary cell growth and suggest that its elevation in breast cancer is of functional signi®cance. Oncogene (2001) 20, 4961 ± 4971. Keywords: Brn-3; MCF7; stable cell line; breast cancer; transcription factor Introduction The closely related but distinct Brn-3a and Brn-3b POU domain containing proteins were originally identi®ed in the cells of the developing sensory nervous system where they regulate neuronal proliferation and dierentiation (Lillycrop et al., 1992; He et al., 1989; Gerrero et al., 1993; Turner et al., 1994). Brn-3

*Correspondence: DS Latchman Received 15 December 2000; revised 27 March 2001; accepted 27 April 2001

proteins have since, however, been detected in tissues of the reproductive tract including the testes, cervix, and breast (Ndisang et al., 1999; Budhram-Mahadeo et al., 1998). Brn-3a and Brn-3b are encoded by two distinct genes located on separate chromosomes. Both of these genes are subject to alternative splicing. When the Brn3a and Brn-3b genes are transcribed from a downstream promoter they produce an intronless short form of the transcription factor. Likewise, when either of the genes is transcribed from an upstream promoter, a long RNA including a single intron is produced. Once this intron is removed, this longer RNA results in the translation of long forms of Brn-3a or Brn-3b which contain additional N-terminal sequences (Theil et al., 1993). In the case of Brn-3a this N-terminal sequence is responsible for speci®c transcriptional activity. Activation of the a-internexin promoter for example, requires the N-terminal domain of Brn-3a (BudhramMahadeo et al., 1995) while this domain is not necessary for the activation of test promoters containing Brn-3 sites (Morris et al., 1994). Thus the alternative splice forms of the Brn-3 proteins encode functionally distinct transcription factors (Liu et al., 1996). We have shown that Brn-3b functionally interacts with and potentiates the transcriptional eects of the ER (Budhram-Mahadeo et al., 1998), which is also expressed in tissues of the reproductive tract. The POU domains of both Brn-3a and Brn-3b interact with the DNA-binding domain of the ER in anity chromatography pull-down experiments. This interaction was con®rmed as physiologically relevant in coimmunoprecipitation experiments from rat brain and ovary. In addition, electromobility shift assays have shown that both Brn-3a and Brn-3b increase the anity of the ER's binding to the estrogen response element. Finally, Brn-3a and Brn-3b modulate the activity of promoters containing an estrogen responsive element: the vitellogenin gene promoter and an arti®cial ERE-containing promoter. In the presence of 17-b estradiol, Brn-3a acts as a weak repressor and Brn-3b acts as a strong activator on these promoters (Budhram-Mahadeo et al., 1998). The ER has long been known to regulate the growth and dierentiation of female reproductive organs (Endoh et al., 1999; Kato et al., 1995), play a role in

Brn-3b transcription factor in human breast cancer cells JH Dennis et al

4962

the progression of human breast cancers (Dickson and Lippman, 1987; Dickson et al., 1987; Lippman et al., 1987; Fuqua, 1992; Fuqua et al., 1992; Elledge and Osborne, 1997), and consistently increase the estrogendependent proliferation of breast cancer cell lines which express it (Reddel et al., 1983; Taylor et al., 1983; Sutherland et al., 1983a,b; Reddel and Sutherland, 1983; Murphy and Sutherland, 1983). Brn-3 proteins are associated with the balance between proliferation and dierentiation of the ND7 and other neuronal cell lines (Latchman, 1999). Speci®cally, while Brn-3a activates promoters associated with neuronal dierentiation, Brn-3b not only represses these same promoters, but also is able to negate the stimulatory eects of Brn-3a on these promoters (Smith et al., 1997). Likewise, when ND7 cells and other neuronal cell lines are induced to dierentiate, the levels of Brn3b fall and the levels of Brn-3a rise (Smith and Latchman, 1996). Finally, arti®cial over-expression of Brn-3a induces neurite outgrowth in the ND7 cell line, and arti®cial over-expression of Brn-3b inhibits neuronal process outgrowth and neuronal speci®c gene activation, thus preventing the dierentiation of these cells (Smith et al., 1997). These results suggest that a role for Brn-3b may be to maintain these cells in a proliferative state. With these two scenarios in mind, it is interesting to explore the role of Brn-3b in estrogen responsive cancers. Brn-3a and Brn-3b have been detected in the MCF7 human breast adenocarcinoma cell line (BudhramMahadeo et al., 1998). Moreover, both Brn-3a and Brn-3b are expressed in primary breast tissue including: normal mammary cells, benign tumours, and malignant tumours. Brn-3b levels were previously studied in primary breast sample tumours by RT ± PCR and Western blot immunodetection. These experiments showed that the levels of Brn-3b, but not the levels of Brn-3a, were enhanced in tumours compared to the levels in normal mammary cells (Budhram-Mahadeo et al., 1999). To evaluate the possible functional signi®cance of this over-expression, we investigated the eect of arti®cially manipulating Brn-3b levels on the growth of the well-characterized, estrogen-responsive, MCF7 human breast adenocarcinoma cell line. Here we show that Brn-3b is capable of increasing the cellular growth, proliferation, and endocrine response of this cell line.

Results Establishment and confirmation of stable transformants To establish stable cell transformants with enhanced and reduced levels of Brn-3b, the human breast adenocarcinoma cell line MCF7 was selected. This cell line is well known to be estrogen responsive, and previous work in the lab has shown that MCF7 cells express Brn-3b (Budhram-Mahadeo et al., 1998). Proliferating cells were transfected with the short form Oncogene

of Brn-3b in pLTR, empty pLTR, or pJ4 expressing the ®rst 300 bases of Brn-3b in an anti-sense orientation using the calcium phosphate method. All transfections included pCi Neo at a ®vefold reduced concentration, compared to the expression constructs, for selection of individual clones with antibiotic resistance. There was no signi®cant dierence in the eciency of recovery, or total colony formation when the cells were transfected with vector alone, vector expressing Brn-3b, or vector expressing the Brn-3b antisense factor. Twelve independent G-418-resistant clones were selected from each transfection and grown up for further analysis. Eleven of the twelve selected Brn-3b anti-sense clones grew up. RT ± PCR with a set of primers corresponding to the polylinker of the pJ4 vector and a region within the anti-sense construct (and therefore speci®c for the exogenous transcript and not endogenous Brn-3b) were used to con®rm the expression of Brn-3b anti-sense transcripts (Figure 1a). Two of these clones (hereafter called 3bAS1 and 3bAS2) expressed the construct at high levels. These two independent clones were grown up for further analysis. Ten of the twelve Brn-3b over-expressing clones were viable and each was tested by RT ± PCR with a set of primers corresponding to the Brn-3b transcript (Figure 1b). Because the primer pairs used in this analysis could amplify both the endogenous and ectopically expressed Brn-3b, cycle numbers were kept low (18 cycles) to identify clones which overexpressed Brn-3b transcripts. Two of these clones (hereafter called 3bO1 and 3bO2) expressed the transcript at high levels and were grown up for further analysis. Having shown that the cell lines expressed the appropriate RNA transcripts, it was next determined whether over-expression of Brn-3b mRNA led to a rise in Brn-3b protein levels, and likewise, if the levels of expression of the anti-sense construct led to a decrease in endogenous Brn-3b. Eighty per cent con¯uent samples of each clonally selected cell line were harvested in cell lysis buer. Sixty micrograms of total cell protein was run on a SDS/15% polyacrylamide gel, blotted, and incubated with a Brn-3b speci®c antibody (Figure 1c and Table 1). The over-expressing clones showed not only an increase in the short form of Brn3b (33 kDa) but surprisingly also a smaller increase in the endogenous long form of Brn-3b (46 kDa). The anti-sense clones showed a reduction in the endogenous levels of both long and short forms of Brn-3b as expected since the anti-sense construct is derived from a region common to the mRNAs encoding these two forms. Modulation of Brn-3b expression results in altered cell growth in monolayers Having established that the clones did exhibit the appropriate changes in Brn-3b expression, we wished to test the eect of this alteration on cell growth, saturation density growth limitation, and proliferation.


4963

Figure 1 Overexpression of Brn-3b constructs in MCF7 cells. (a, b) RT ± PCR of mRNA from MCF7 clones selected by G418 resistance. (a) Lane 1: molecular weight markers, lanes 2 and 3: control clones transfected with empty pLTR, lanes 4 ± 10: clones transfected with pJ4 Brn-3b antisense, lane 11: transfected plasmid positive control, lane 12: no reverse transcriptase negative control. (b) Lanes 1 and 2: control clones transfected with empty pLTR, lanes 3 ± 8: clones transfected with pLTR Brn-3b short, lanes 9 and 10: clones transfected with pJ4 Brn-3b antisense, lane 11: no reverse transcriptase negative control, lane 12 transfected plasmid positive control. (c) Top panel: Immunoblot of selected Brn-3b clones in MCF7 cells. Eighty per cent con¯uent samples of each clonally selected cell line were harvested in cell lysis buer and 60 mg of total cellular protein was run on a SDS/15% polyacrylamide gel, blotted to nylon membrane and incubated with a Brn-3b speci®c antibody. Lanes 1 and 2: clones transfected with pLTR Brn-3b short, lanes 3 and 4: control clones transfected with empty pLTR, lanes 5 and 6: clones transfected with pJ4 Brn3b antisense. Both the long and short splice forms of Brn-3b are shown on the blot. The bottom panel shows an identically prepared immunoblot probed with anti-actin antibody

Table 1 Expression of Brn-3b in MCF7 clonal cell lines as determined by immunoblotting

Clone

Densitometry (arbitrary units) Brn-3b Brn-3b long short

pLTR Brn-3b O 1 pLTR Brn-3b O 2 pLTR V 1 pLTR V 2 pJ4 Brn-3b AS 1 pJ4 Brn-3b AS 2

31440 39539 8062 15271 1763 4919

41149 45958 3670 4072 0 0

Fold expression Brn-3b Brn-3b long short 2.6 3.3 1.0 1.0 0.2 0.4

10.6 11.9 1.0 1.0 0.0 0.0

Growth curve study The Brn-3b over-expressing, vector control, and Brn-3b anti-sense clones were studied to determine eects on cell growth (Figure 2

and Table 2). Alterations in the levels of Brn-3b in MCF7 cell lines has a signi®cant eect on both population doubling time as well as cellular density at plateau phase. The clone 3bO1 over-expressing Brn3b reached a plateau at an average density of 7.63+1.666105 cells per cm2 after 14 days. The clone 3bO2 over-expressing Brn-3b reached a plateau at an average density of 7.43+1.666105 cells per cm2 after 14 days. The pLTR control clone V1 reached a plateau at an average density of 4.18+0.966105 cells per cm2 after 12 days. The pLTR control clone V2 reached a plateau at an average density of 4.25+0.546105 cells per cm2 after 12 days. The clone 3bAS1 over-expressing the anti-sense construct of Brn-3b reached a plateau at an average density of 2.93+0.906105 cells per cm2 after 10 days. The clone 3bAS2 over-expressing the anti-sense construct of Brn-3b reached a plateau at an Oncogene


4964

Figure 2 Growth rate of MCF7 stable clones. Two clones representing each of the overexpressing (3bO1 and 3bO2), control (V1 and V2), and antisense (3bAS1 and 3bAS2) cell lines were used in each of these experiments. The number of cells at each timepoint represents the mean of three independent experiments+the standard deviation of the mean. *Represent statistically signi®cant dierences (P50.05) when overexpressing or antisense cell lines were compared to empty vector controls

Table 2 The growth parameters of the Brn-3b altered MCF7 clonal cell lines

Clone pLTR Brn-3b O 1 pLTR Brn-3b O 2 pLTR V 1 pLTR V 2 pJ4 Brn-3b AS 1 pJ4 Brn-3b AS 2

Days to plateau

Plateau phase cell number (6105/cm2)

Saturation density cell number (6105/cm2)

14 14 12 12 10 10

7.63+1.66 7.43+1.66 4.18+0.96 4.25+0.54 2.93+0.90 2.13+0.60

8.63+0.89 8.81+1.08 3.95+0.82 4.31+0.52 3.83+0.81 2.64+0.84

Anchorage independent growth Colonies Colonies (number/plate) (size) 229+44

M

188+39

S±M

51+9

S

The cell density at plateau was determined using the cell number at the day 14 time point (the ®rst time point at which all cell lines were in plateau phase) of the growth curve. The saturation density was determined from the mean of all four time points taken. Colonies in soft agar were counted directly using an inverted microscope

average density of 2.13+0.606105 cells per cm2 after 10 days. Using the w2 statistical test, the individual cell counts for the Brn-3b over-expressing and Brn-3b antisense cell lines were compared to the pLTR control cell line cell counts at each time point. All time points after the initial lag phase of growth were statistically signi®cant from the expected growth rate gauged by the pLTR controls. The Brn-3b over-expressing cell lines grew signi®cantly faster, and the Brn-3b anti-sense signi®cantly slower than the pLTR control cell lines (all P-values 50.00001). Saturation density limitation of growth study In addition to the signi®cant dierences in each of the representative cell lines' growth, the dierences in cell density at the plateau phase of the cellular growth Oncogene

curves are also similarly dierent. A saturation density limitation growth study was therefore carried out to exclude the possibility that the dierent saturation densities observed in the initial studies were due to slow growth or exhaustion of medium. Cells were plated out at near con¯uence onto coverslips, grown for 48 h, and then removed to large dishes containing excess medium which was frequently changed. After 10 days the cells were harvested and counted every 2 days for 8 more days. Three coverslips of each clone were used for each time point (data not shown). The saturation densities of each clone were similar to the plateau densities for each clone shown in the growth curves. Using the w2 statistical test, the individual cell counts for the Brn-3b over-expressing and Brn-3b antisense cell lines were compared to the pLTR control cell


line cell counts at each time point. The Brn-3b overexpressing cell lines grew to signi®cantly higher densities (P-values50.00001) at all of the time points, and the Brn-3b anti-sense to signi®cantly lower than the pLTR control cell lines (P-values50.05) in three out of four of the time points (data summarized in Table 2). In summary, the Brn-3b over-expressing clones had saturation densities over twice that of the control and anti-sense cell lines, while the control cell lines generally grew to higher densities than the anti-sense cell lines. Tritiated-thymidine incorporation study To investigate if the growth dierences seen in the cell lines were the result of dierences in the rate of proliferation in the cell lines, incorporation of tritiated thymidine in the replicating cells was used as a measure of proliferative activity. Cells were incubated with tritiated thymidine for 1 h, then harvested and counted on a scintillation counter (Figure 3). It is clear from these studies that these cell lines proliferate at a rate commensurate with the level of Brn-3b protein, with the Brn-3b over-expressing cell lines consistently proliferating signi®cantly more than, and the Brn-3b anti-sense cell lines consistently proliferating signi®cantly less than, the pLTR control cell lines. Hence the dierences in growth of the dierent cell lines as measured by cell number appear to be primarily dependent on dierences in the rate of cell division rather than, for example, the amount of apoptosis.

4965

Modulation of Brn-3b expression results in altered anchorage-independent growth The ability of cells to exhibit anchorage-independent growth is a hallmark of the transformed cell phenotype and is necessary for tumour growth in vivo. We further tested the properties of these clones by measuring the ability of the clones to form colonies in a soft agar assay. Dierences were observed in this experiment in both the number and size of colonies formed (Figure 4 and Figure 5D-F and Table 2). In full growth medium containing 0.35% agarose the Brn-3b over-expressing and pLTR empty control vector cell lines had 229+44 and 188+39 colonies respectively. In general, the Brn3b over-expressing colonies were consistently larger than the pLTR control colonies. The Brn-3b anti-sense cells produced 51+9 colonies that were consistently smaller in size than the over-expressing or control colonies. Modulation of Brn-3b expression results in altered cellular morphology Morphological examination of the clonal cell lines revealed changes in the shape of the cells. Five to 10 per cent of the Brn-3b over-expressing cells had a severely altered cell shape. These cells appeared to adhere to other MCF7 cells but were never seen adhering to the tissue culture plastic. The cells were 1 ± 5 times the size of normal cells. These large cells had a vacuolar-type appearance with the cellular machinery appearing as a crescent in the side of the cell. Finally, these cells were

Figure 3 Tritiated thymidine incorporation proliferation studies. Two clones representing each of the overexpressing (3bO1 and 3bO2), control (V1 and V2), and antisense (3bAS1 and 3bAS2) cell lines were used in each of these experiments. The number of counts for each cell line represents the mean of at least three independent experiments+the standard deviation of the mean. *Represent statistically signi®cant dierences (P50.05) when overexpressing or antisense cell lines were compared to empty vector controls Oncogene


4966

Figure 4 Anchorage-independent growth studies. Clones representing each of the overexpressing (3bO1), control (V1), and antisense (3bAS1) cell lines were used in each of these experiments. Cells were grown in full growth medium containing 0.35% agarose and colonies containing greater than 32 cells were counted using an inverted microscope after 21 days. The number of colonies scored for each cell line represents the mean of three independent experiments+the standard deviation of the mean. *Represent statistically signi®cant dierences (P50.05) when overexpressing or antisense cell lines were compared to empty vector controls

less adherent to tissue culture plastic as well as to other cells (Figure 5a). This morphology existed in the colonies grown in soft agar as well (Figure 5d). Neither the MCF7 cells transfected with the pLTR empty control plasmid (Figure 5b,e) nor those transfected with the Brn-3b anti-sense construct (Figure 5c,f) had an appearance varied from wild-type MCF7 cells. Modulation of Brn-3b expression results in altered response to estrogen Eects of estrogen and estrogen analogues on MCF7 cells Because MCF7 cells are well documented to be estrogen responsive, and we have previously shown that Brn-3b functionally interacts with the ER (Budhram-Mahadeo et al., 1998), it was necessary to see if the altered levels of Brn-3b resulted in dierences in estrogen responsiveness. In general, estrogen stimulation of MCF7 cell growth is associated with an increase in growth fraction, a shortening of cell cycle time, and a decrease in cell death (Sutherland et al., 1983a). Tamoxifen treatment of MCF7 cells causes an accumulation of cells in G0/G1 phase in the cell cycle with a commensurate decrease of cells in S phase (Sutherland et al., 1983b). In order to elucidate the pathways through which the altered levels of Brn-3b may produce their eect, the eect of treatment with Oncogene

17 b-estradiol (E2) was tested on each of the MCF7 clonal cell lines. Changes in growth in monolayers was monitored. In the previous set of experiments, cells were grown in full growth medium containing phenolsulphonthalein (phenol-red) pH indicator and fetal calf serum. In order to determine the eects of 17-b estradiol, this set of experiments was carried out in both full growth medium, as well as medium without the estrogenic phenol red pH indicator prepared with serum pretreated with dextran-coated charcoal to remove endogenous estrogens (stripped serum medium). Eects on growth Cells were plated and grown in the presence of estradiol, vehicle control, or tamoxifen in full growth medium (Figure 6a) as well as in stripped serum medium (Figure 6b). The addition of E2 to full growth medium increased the growth of all the MCF7 clonal cell lines. Growth dierences in which the Brn-3b over-expressing cell lines grew more than the empty vector controls, which grew more than the Brn-3b anti-sense cell lines, could be detected under all conditions in both types of media. Hence, the dierent growth rates of the cells with dierent Brn-3b levels can be observed under all conditions with or without added E2. Interestingly, however, the strongest response to E2 was observed in the Brn-3b over-expressing cells in stripped serum, indicating that these cells show enhanced estrogen


4967

Figure 5 Photomicrographs (all taken at610 magni®cation) of MCF7 clonal cell lines (a ± c) grown in monolayer, and (d ± f) grown in soft agar. (a ± c) Cells were plated in full growth medium and allowed to grow for 3 days at which point the medium was replaced with fresh medium. These photomicrographs were taken on the ®fth day post-plating. (d ± f) Cells were plated in full growth medium containing 0.35% agarose and allowed to grow for 21 days. These photomicrographs were taken on the 22nd day post-plating. In (a) and (d) the é represents cells with altered morphology

responsiveness (Figure 6b). This extends our previous ®nding that Brn-3b can enhance the response of an estrogen-dependent reporter construct in co-transfection assays (Budhram-Mahadeo et al., 1999). To further investigate the role of the estrogen receptor in the eects we observed, we also monitored the growth of the dierent cells in the presence of the estrogen antagonist, Tamoxifen. As indicated in Figure 6, all the cell lines showed decreased growth in the presence of Tamoxifen in the experiments carried out in full growth medium and decreases were also observed for the control and anti-sense cells in the estradiol-free medium. Importantly however, the growth rate dierences between the dierent cell lines were still maintained even in the presence of the estrogen antagonist both in full growth medium and the estradiol-free medium. Hence the cells overexpressing Brn-3b continue to grow faster than control

cells even under the conditions, with reduced growth still being observed in the anti-sense cells. Similar dierences between the dierent cell lines with the dierent treatments were also observed in assays of cellular proliferation using tritiated thymidine (data not shown). Discussion Overexpression of Brn-3b in ND7 neuronal cells inhibits neuronal process outgrowth and synaptic vesicle gene expression and allows continued proliferation under conditions which would normally induce dierentiation (Smith et al., 1997). Moreover, in epithelial cells, Brn-3b can interact with the ER to potentiate the transcriptional eects of the ER, and the ER enhances estradiol-stimulated MCF7 cell proliferaOncogene


4968

Figure 6 Endocrine characteristics of MCF7 stable clones as measured by cell counts. Two clones representing each of the overexpressing (3bO1 and 3bO2), control (V1 and V2), and antisense (3bAS1 and 3bAS2) cell lines were used in each of these experiments. Cells were grown in full growth medium (FGM) or phenol-redless medium containing dextran charcoal-stripped serum (SS) for 48 h and subsequently treated with 17-b estradiol (200 ng/ml), ethanol vehicle, or Tamoxifen (200 ng/ml). Cells were trypsinized and counted on a haemocytometer at the end of 2 days stimulation. The number of cells for each cell line represents the mean of three independent experiments+the standard deviation of the mean. (a) Cells grown in full growth medium (FGM), (b) cells grown in phenol-redless medium containing dextran charcoal-stripped serum (SS)

tion (Budhram-Mahadeo et al., 1998). In vivo, the levels of Brn-3b are elevated in breast tumour samples when compared to normal breast tissue samples. Thus, we have manipulated the levels of Brn-3b in the MCF7 human breast adenocarcinoma cell line to investigate the functional role of Brn-3b in breast cancer. Oncogene

Here we have clearly shown successful manipulation of Brn-3b expression at both the RNA and protein levels. Interestingly, in the Brn-3b short over-expressing cells we noticed an increase in the protein levels of the long form of Brn-3b. This gives an indication that Brn3b may be capable of autoregulating its own


expression. This result is interesting in light of the fact that autoregulatory circuits control the expression of several homeobox genes (Awgulewitsch and Jacobs, 1992; Bergson and McGinnis, 1990; Malicki et al., 1992; Popperl and Featherstone, 1992; Packer et al., 1998). A bipartite DNA-binding domain containing a POU-homeodomain and a POU-speci®c domain characterize the POU domain family of transcription factors. Thus, this Brn-3b autoregulation may be the result of an evolutionarily conserved homeodomain autoregulatory circuit and should be investigated. As expected, stable expression of the Brn-3b anti-sense construct resulted in a decrease in the levels of both the long and short forms of Brn-3b as the anti-sense gene product covered an area common to both long and short isoforms of the Brn-3b gene. The functional eects of the manipulated levels of Brn-3b are clear. The increased and decreased levels of Brn-3b have signi®cant eects on the growth characteristics of the cells grown in monolayers as well as cells grown as anchorage-independent colonies. The cells over-expressing Brn-3b short consistently grew faster and to higher densities when compared to the mock transfected controls and the clones with reduced levels of Brn-3b. This result was consistent for cells grown in the absence or presence of 17-b estradiol. In addition, the Brn-3b over-expressing cells exhibited an anchorage-independent growth advantage, while the cells with reduced levels of Brn-3b were anchorageindependent growth inhibited. Brn-3b is able to maintain ND7 cells in a proliferative state, and prevent dierentiation. In addition, this transcription factor is able to act as an active repressor of neuronal speci®c gene activation (Smith et al., 1997). Thus, it would be interesting to investigate how Brn-3b transcriptionally mediates the growth-enhanced phenotype seen in the Brn-3b overexpressing cells. As Brn-3b is able to functionally interact with the estrogen receptor and potentiate the transcription of genes containing an estrogen responsive element, it would be logical to investigate the gene expression of estrogen responsive genes which are altered in breast cancer (i.e. ER or Hsp27) in each of the Brn-3b manipulated cell lines. We have previously demonstrated that Brn-3b can interact with the estrogen receptor and enhance its eects on estrogen-dependent genes (BudhramMahadeo et al., 1998). Thus, some of the eects we have observed may be due to the interaction of Brn-3b with the estrogen receptor. Indeed, we observed a much stronger response to estrogen in the Brn-3b over-expressing cells grown in stripped serum. Despite this, however, alterations in estrogen responsiveness, mediated via altered Brn-3b levels, cannot be the sole explanation for the eects we observed. Thus, growth rate dierences between the Brn-3b over-expressing cells, control cells, and Brn3b anti-sense cells were observed under all conditions regardless of the presence of estrogen, either added additionally or present in serum, or with the addition of the estrogen antagonist Tamoxifen.

Hence it appears that the eects we observed are mediated, at least in part, via a direct eect of Brn3b on cellular growth, independent of the estrogen receptor. In summary therefore, Brn-3b plays a key role in the growth and proliferation of human breast cancer cells in vitro and enhances the ability to grow in an anchorage independent manner which is necessary for tumour growth in vivo. This work provides a ®rm link to the previous studies that have shown Brn-3b is overexpressed in primary breast cancer tissue samples. Further investigation is, however, necessary to determine the pathways and transcriptional mechanisms aected by the altered levels of Brn-3b in vivo. The further analysis of Brn-3b should greatly enhance our understanding of the mechanisms of human breast cancer growth, proliferation, and endocrine response and may potentially identify Brn-3b as a therapeutic target for breast cancer.

4969

Materials and methods Establishment and confirmation of pLTR Brn-3b short and pJ4 Brn-3b anti-sense cell lines The pLTR vector and the pLTR construct encoding recombinant Brn-3b short were the kind gift of Dr T Moroy. The Brn-3b anti-sense construct was prepared by cloning the ®rst 300 bases of Brn-3b short in reverse orientation into the SalI/BamHI sites of the pJ4 omega mammalian expression vector. Stable cell lines expressing these constructs as well as control cell lines containing empty pLTR (all under the control of the Moloney Murine Leukaemia Virus promoter) were prepared, selected, and isolated as follows. MCF7 cells (56105) in each well of a six-well dish were transfected using pLTR Brn-3b short, pLTR empty, or pJ4 Brn-3b anti-sense (5 mg each DNA) using the calcium phosphate method. In each of the transfections empty pCi Neo (1 mg DNA) was cotransfected for antibiotic selection of successfully transfected cells. Cells were allowed to grow for 72 h post-transfection. Medium containing 800 mg/ml of G418 sulphate was used to select antibiotic resistant clones. After 14 days, at least 12 single colonies representing each of the three constructs were picked using a pipette and allowed to grow in individual wells of a 24-well dish. Surviving cells were grown to con¯uence, and tested for Brn-3b expression. Initially clonal cell lines were examined using RT ± PCR. Primers corresponding to the polylinker of pJ4 5'-AGATCTGGTACCATCGAT-3' and the Brn-3b anti-sense sequence 5'GCTCTCATCAAAGCTTC-3' were used to con®rm the stable integration of the anti-sense construct. Primers corresponding to the POU domain of mouse Brn-3b 5'GGTCTGCATCCACGTCGCTC-3', and 5'-GGAGGGCGAGCTGCTTGAGC-3' were used to con®rm the stable integration of the Brn-3b construct. To examine the levels of Brn-3b protein expression, selected colonies were grown in six-well plates to 80% con¯uence and harvested directly in SDS cell lysis buer (3% SDS, 50 mM Tris pH 6.8, 10% glycerol, and 1% b-mercaptoethanol). Protein concentration of cellular extracts were determined using Bradford chemistry, and 60 mg of total cellular protein was run on an SDS/15% polyacrylamide gel, blotted to nylon membrane, and incubated with a Brn-3b speci®c antibody. Oncogene


4970

Cellular growth conditions Cells were grown in either full growth medium (DMEM, 10% foetal calf serum, 1% non-essential amino acids, 800 mg/ml G418 sulphate), or in stripped serum medium (DMEM without phenol red, 10% dextran charcoal-stripped foetal calf serum, 1% non-essential amino acids, 800 mg/ml G418 sulphate). The eects of 17-b estradiol and Tamoxifen were tested at concentrations of 200 ng/ml of medium. Control experiments contained an equal volume of ethanol vehicle. Cellular growth curves and saturation density limitation studies Two MCF7 clonal cell lines representing each of the stably integrated constructs were grown to 80% con¯uence, trypsinized, counted and plated into 24-well plates using 1 ml of cells at 16104 cells per ml. The medium was changed every 3 days for 20 days. Triplicate wells were counted for each time point. For the saturation density limitation of growth study, two MCF7 clonal cell lines representing each of the stably integrated constructs were grown to 80% con¯uence, trypsinized, counted and plated into 24-well plates containing 13 mm diameter coverslips in each well using 1 ml of cells at 16105 cells per ml. After 48 h, the coverslips were transferred to 6-well dishes containing 5 ml of medium. The medium was changed every 3 days for 16 days. Ten days post-transfer to the 6-well plates, triplicate wells were counted for each time point every 48 h. Tritiated thymidine incorporation proliferation studies Two MCF7 clonal cell lines representing each of the stably integrated constructs were grown to 80% con¯uence, trypsinized, counted and plated into 96-well plates using

200 ml of cells at 16103 cells/ml. Cells were grown for the indicated amount of time. At the appropriate time point, cells were pulsed with 10 ml of 50 mCi/ml [3H] thymidine for 1 h. Cells were then trypsinized and then harvested into glass ®lters, and washed using a Tomtec cell harvester. The glass ®lters were then treated with scintillant and counted on a Nuclear Enterprises NE10600 scintillation counter. Anchorage-independent growth Anchorage-independent growth studies were carried out in soft agar in 6-well plates. A base layer of 2 ml of growth medium containing 0.7% agarose, was covered with 1 ml of growth medium containing 0.35% agarose and 16105 cells per ml. Finally, 2 mls growth medium containing 0.7% agarose was layered on top of the cells. Colonies were allowed to grow for 21 days. After 21 days, colonies comprised of at least 32 cells were counted on an inverted microscope. Statistics Data are presented as the mean+the standard error of the mean. Experimental cell line data were analysed using the w2 statistical test to determine the statistical signi®cance of dierences from vector control (expected) cell data.

Acknowledgements We thank Tarik Moroy for the gift of pLTR and Brn-3b recombinant clones, Martin Smith for the gift of pJ4 Brn3b anti-sense constructs, and Benjamin Chain and Walter Low for assistance with the proliferation studies. This work was supported by the USA Department of Defence, Breast Cancer Research Program.

References Awgulewitsch A and Jacobs D. (1992). Nature, 358, 341 ± 344. Bergson C and McGinnis W. (1990). EMBO J., 9, 4287 ± 4297. Budhram-Mahadeo V, Morris PJ, Lakin ND, Theil T, Ching GY, Lillycrop KA, Moroy T, Liem RK and Latchman DS. (1995). J. Biol. Chem., 270, 2853 ± 2858. Budhram-Mahadeo V, Ndisang D, Ward T, Weber BL and Latchman DS. (1999). Oncogene, 18, 6684 ± 6691. Budhram-Mahadeo V, Parker M and Latchman DS. (1998). Mol. Cell. Biol., 18, 1029 ± 1041. Dickson RB and Lippman ME. (1987). Endocr. Rev., 8, 29 ± 43. Dickson RB, Kasid A, Hu KK, Bates SE, Knabbe C, Bronzert D, Gelmann EP and Lippman ME. (1987). Proc. Natl. Acad. Sci. USA, 84, 837 ± 841. Elledge RM and Osborne CK. (1997). BMJ, 314, 1843 ± 1844. Endoh H, Maruyama K, Masuhiro Y, Kobayashi Y, Goto M, Tai H, Yanagisawa J, Metzger D, Hashimoto S and Kato S. (1999). Mol. Cell. Biol., 19, 5363 ± 5372. Fuqua SA. (1992). J. Natl. Cancer Inst., 84, 554 ± 555. Fuqua SA, Fitzgerald SD, Allred DC, Elledge RM, Nawaz Z, McDonnell DP, O'Malley BW, Greene GL and McGuire WL. (1992). Cancer Res., 52, 483 ± 486. Gerrero MR, McEvilly RJ, Turner E, Lin CR, O'Connell S, Jenne KJ, Hobbs MV and Rosenfeld MG. (1993). Proc. Natl. Acad. Sci. USA, 90, 10841 ± 10845. Oncogene

He X, Treacy MN, Simmons DM, Ingraham HA, Swanson LW and Rosenfeld MG. (1989). Nature, 340, 35 ± 41. Kato S, Endoh H, Masuhiro Y, Kitamoto T, Uchiyama S, Sasaki H, Masushige S, Gotoh Y, Nishida E, Kawashima H et al., (1995). Science, 270, 1491 ± 1494. Latchman DS. (1999). J. Cell. Physiol., 179, 126 ± 133. Lillycrop KA, Budrahan VS, Lakin ND, Terrenghi G, Wood JN, Polak JM and Latchman DS. (1992). Nucleic Acids Res., 20, 5093 ± 5096. Lippman ME, Bates S, Hu KK, Davidson N and Dickson RB. (1987). J. Lab. Clin. Med., 109, 230 ± 235. Liu YZ, Dawson SJ and Latchman DS. (1996). J. Mol. Neurosci., 7, 77 ± 85. Lou L, Bergson C and McGinnis W. (1995). Nucleic. Acids. Res., 23, 3481 ± 3487. Malicki J, Cianetti LC, Peschle C and McGinnis W. (1992). Nature, 358, 345 ± 347. Morris PJ, Theil T, Ring CJ, Lillycrop KA, Moroy T and Latchman DS. (1994). Mol. Cell. Biol., 14, 6907 ± 6914. Murphy LC and Sutherland RL. (1983). J. Clin. Endocrinol. Metab., 57, 373 ± 379. Ndisang D, Budhram-Mahadeo V and Latchman DS. (1999). J. Biol. Chem., 274, 28521 ± 28527. Packer AI, Crotty DA, Elwell VA and Wolgemuth DJ. (1998). Development, 125, 1991 ± 1998. Popperl H and Featherstone MS. (1992). EMBO J., 11, 3673 ± 3680.


Reddel RR and Sutherland RL. (1983). Eur. J. Cancer Clin. Oncol., 19, 1179 ± 1181. Reddel RR, Murphy LC and Sutherland RL. (1983). Cancer Res., 43, 4618 ± 4624. Smith MD and Latchman DS. (1996). Int. J. Cancer, 67, 653 ± 660. Smith MD, Dawson SJ and Latchman DS. (1997). J. Biol. Chem., 272, 1382 ± 1388. Sutherland RL, Hall RE and Taylor IW. (1983a). Cancer Res., 43, 3998 ± 4006.

Sutherland RL, Green MD, Hall RE, Reddel RR and Taylor IW. (1983b). Eur. J. Cancer Clin. Oncol., 19, 615 ± 621. Taylor IW, Hodson PJ, Green MD and Sutherland RL. (1983). Cancer Res., 43, 4007 ± 4010. Theil T, McLean Hunter S, Zornig M and Moroy T. (1993). Nucleic Acids Res., 21, 5921 ± 5929. Turner EE, Jenne KJ and Rosenfeld MG. (1994). Neuron, 12, 205 ± 218.

4971

Oncogene