Mammary gland specific hEGF receptor transgene expression induces neoplasia and inhibits differentiation

Abstract

The epidermal growth factor receptor (EGFR) is overexpressed in about 48% of human breast cancer tissues. To analyse the role of the EGFR in mammary tumor development we generated transgenic mice expressing the human EGFR under the control of either the MMTV-LTR (MHERc) or the β-lactoglobulin promoter (BLGHERc). The BLGHERc-transgene was expressed exclusively in the female mammary gland, whereas the MHERc transgene was expressed more promiscuously in other organs, such as ovary, salivary gland and testis. Female virgin and lactating transgenic mice of both strains have impaired mammary gland development. Virgin EGFR transgenic mice developed mammary epithelial hyperplasias, whereas in lactating animals progression to dysplasias and tubular adenocarcinomas was observed. In both strains the number of dysplasias increased after multiple pregnancies. The transgene expression pattern was heterogeneous, but generally restricted to regions of impaired mammary gland development. Highest EGFR transgene expression was observed in adenocarcinomas. By using a whole mount organ culture system to study the differentiation potential of the mammary epithelium, we observed a reduced number of fully developed alveoli and a decrease in whey acidic protein expression. Taken together, EGFR overexpression results in a dramatic effect of impaired mammary gland development in vitro as well as in vivo, reducing the differentiation potential of the mammary epithelium and inducing epithelial cell transformation.

Introduction

Growth, development and differentiation of mammary epithelial tissues are multi-stage processes that are driven by a combination of autocrine systemic ovarian and pituitary hormones as well as paracrine signaling factors (Lee et al., 1995; Spitzer et al., 1995). There is increasing evidence that genetic alterations in growth factor signaling pathways are closely related to developmental abnormalities and to chronic diseases including cancer (Aaronson et al., 1991; Yee et al., 1989). The EGFR is a member of the type I receptor tyrosine kinase family (RTK) (Hackel et al., 1999) and can be stimulated upon autocrine and/or paracrine interactions with corresponding ligands such as EGF, transforming growth factor alpha (TGFα) and amphiregulin (Normanno et al., 1995). Overexpression of the EGFR is associated with enhanced metastatic potential in experimental studies, and poor prognosis in human breast cancer (Harris and Nicholson 1987; Klijn et al., 1992). The EGFR is overexpressed in about 48% of primary human breast tumors (Klijn et al., 1992). Cellular transformation mediated by ligand activation of the EGFR has been experimentally demonstrated in rodent fibroblasts transfected with the EGFR and mammary epithelial cells following the constitutive overproduction of EGF or TGFα (Di Fiore et al., 1987a; Tortora et al., 1989).

Therefore, it has been concluded that EGFR can facilitate cellular transformation. An increase in TGFα synthesis and secretion generally occurs in several types of human carcinoma cell lines, in primary human tumors and in fibroblasts or in mammary epithelial cell lines that have been transformed with a number of different oncogenes such as point mutated c-Ha- or Ki-ras genes (Di Marco et al., 1989; Salomon et al., 1991). Furthermore, blocking EGFR expression by introducing an EGFR antisense construct or treating cells with an EGFR neutralizing antibody inhibits colony formation in semisolid medium as well as growth of tumors cells as xenografts in nude mice (Baselga and Mendelsohn, 1994; Chakrabarty et al., 1995; Fan et al., 1993; Liu et al., 1995; Rajagopal et al., 1995).

Transgenic mice represent a useful model to assess tissue specific action of oncogenes or growth factors in vivo and to study their role in breast cancer formation (Hinrichs et al., 1991). Transgenic mice expressing various oncogenes specifically in the mammary gland have been generated by employing either the mouse mammary tumor virus (MMTV)-LTR, or milk protein specific promoters such as the whey acidic protein, β-lactalbumin or β-lactoglobulin promoters (for review see Muller, 1991). In the majority of these cases, mammary tumors develop in a stochastic fashion with a latency period of 5–10 months (Andres et al., 1988; Bouchard et al., 1989; Goldenring et al., 1993; Krane and Leder, 1996). These data suggest that overexpression of individual oncogenes is necessary but not sufficient for tumor induction. Additional secondary genetic events are likely to contribute to the tumor phenotype. In this respect, mammary gland specific TGFα overexpression leads to the development of mammary adenocarcinomas that also exhibit an increase in EGFR expression, supporting the hypothesis of an autocrine mechanism for transformation of the mammary epithelium by the EGFR (Matsui et al., 1990). Therefore an increase in the expression of the EGF-receptor and/or its ligands in mammary tissue might be a primary cause for tumor formation. However, direct experimental validation of whether EGFR overexpression can contribute to mammary tumorigenesis is lacking.

To address this question we generated transgenic mouse lines expressing the hEGFR under the control of two mammary gland directed promoters and studied mammary gland growth and functional differentiation in vivo as well as in vitro. We demonstrate that hEGFR expression can cause hyperplastic, dysplastic and neoplastic changes in mammary epithelia of transgenic mice. By employing whole mount mammary gland organ culture, we found that the expression of the hEGFR suppresses functional and morphological differentiation of the mouse mammary epithelium.

Results

Generation of transgenic mice

To directly investigate the role of EGFR overexpression in growth, development and tumorigenesis of the mammary gland, we created transgenic mice expressing the hEGFR under the mammary gland specific control of the MMTV/LTR or the β-lactoglobulin (BLG) promoter. Two hybrid genes, the MMTV-HERc-SV40 polyA+ (MHERc) and the β-lactoglobulin-HERc (BLGHERc) were constructed (Figure 1a,c). In addition to the 5′- and 3′-untranslated regions of the β-lactoglobulin gene, the expression cassette contains a truncated exon/intron sequences of this gene (Figure 1b). The human EGFR BLGHERc expression construct was co-injected with the complete sheep β-lactoglobulin gene containing an in-frame mutation to generate a premature translational stop (Figure 1b,c), preventing BLG synthesis. Co-injection of these two constructs results in co-integration of both cDNA's into the mouse genome and should stabilize the hEGFR mRNA (Yull et al., 1996).

Figure 1
figure1

Structure of hEGFR expression constructs. (a) A 1.01 kb MMTV/LTR promoter/enhancer fragment was fused to the hEGFR cDNA which was linked to the SV 40 polyA+. (b) pBLGdeltatrans, is a modified entire BLG gene with a frameshift mutation creating a translational stop (see Materials and methods). (c) PBJ41 with inserted hEGFR cDNA. The minigene has some modifications compared to the non-coding region of BLG gene (see Materials and methods)

Transgenic mice were identified by PCR analysis by employing specific hEGFR primers. hEGFR positive mice were further verified by Southern blot analysis using a hEGFR specific 1.1 kb DNA probe (data not shown). DNA from BLGHERc positive mice was additionally analysed by PCR detecting a DNA sequence of the first exon of the sheep BLG sequence not present in the BLGHERc expression construct. All BLGHERc founder mice were shown to be bi-transgenic for both the BLGHERc expression construct and the silent β-lactoglobulin gene. For both the MHERc and the BLGHERc expression construct we identified two founder mice which were used to generate four independent transgenic lines: line 6903 and line 7343 for the MHERc strain and line 28 and line 116 for the BLGHERc strain.

Females of both the MHERc and the BLGHERc strains were able to wean their offsprings. Males of one MHERc strain (line 7343) were infertile.

Expression analysis

hEGFR transgene expression was analysed by RT–PCR (Table 1). The specificity of the RT–PCR reaction for the transgene was confirmed by using total RNA from wild type liver expressing high amounts of endogenous mouse EGFR mRNA which yielded no amplificated product. hEGFR mRNA was expressed in mammary glands of both virgin and lactating animals from the MHERc strains. Compared to lactating animals approximately 10 times more mRNA had to be used for RT–PCR to detect transgene expression in virgin MHERc mice (data not shown). In addition, transgene expression was observed in testis, ovary and male salivary gland. No transgene expression was found in male mammary gland of MHERc mice, in uterus and in liver (Table 1). For the BLGHERc strains, a different expression pattern of the transgene was observed. hEGFR mRNA was expressed only in the lactating mammary gland which generated a strong RT–PCR signal. No other organ was found positive for the transgene expression (Table 1). Highest expression was found in tumor tissue as shown by Northern analysis (Figure 2a) and immunohistochemistry (Figure 5).

Table 1 Expression of hEGFR transgene in mouse tissue analysed by RT–PCR
Figure 2
figure2

Northern blot analysis of mammary gland tissue and whole mount stainings. (a) 15 μg total RNA from transgenic mammary glands with hyperplastic nodules (#2254, #1746), a mammary gland tumor (#2252) and a wild type mammary gland. The blot was hybridized to a hEGFR specific cDNA probe and the transgene-specific mRNAs are indicated by an arrow. (b) The same blot as in a was hybridized to a GAPDH specific cDNA probe as a loading control. (c) Carmine-red stained mammary gland preparation from a wild type female mouse. (d) Carmine-red stained mammary gland preparation from a transgenic female mouse of the BLGHERc strain

Figure 5
figure3

Immunohistochemical analysis of hEGFR expression in mammary glands obtained from transgenic mice. (a) Lactating control gland, omitting primary antibody. (b) Atrophic region within lactating gland demonstrating strong hEGFR expression (BLGHERc line 28, after one pregnancy). (c) MHERc, dysplasia (line 6903), (d) BLGHERc, dysplasia (line 116, after two pregnancies), (e) Atrophic region adjacent to lactating alveoli (MHERc line 6903, after three pregnancies) and (f) invasive adenocarcinoma of BLGHERc gland (line 116, compare to Figure 3g). All pictures are taken with 200-fold magnification

Whole mount mammary gland preparations

To gain more information about the influence of EGFR overexpression on phenotypic differences between wild type and transgenic animals, mammary gland whole-mount preparations were made (Figure 2c,d. The mammary tree of the transgenic animals was grossly abnormal (Figure 2d). They showed enlarged ducts and multiple hyperplastic alveolar nodules, whereas the wild type glands appeared normal (Figure 2c).

Mammary histology of virgin mice

Histological examination revealed that mammary glands of control virgin mice aged 6–12 weeks consist of adipose tissue surrounding small ducts lined with unilayered cuboid epithelial cells. These ducts are surrounded by myoepithelial cells and connective tissue (Figure 3a1). In contrast, both MHERc and BLGHERc transgenic virgin mice of the same age showed scattered clusters of small ducts lined by multilayered cuboidal epithelium and myoepithelium (Figure 3b,c). Although the diameter was similar to the ducts of control mice (Figure 3a1), the lumens of ducts evaluated in glands obtained from MHERc mice were not always apparent as they were occluded with epithelial cells (Figure 3b). In the virgin hEGFR transgenic mice, the number of ductal clusters per microscopic field (36≈plus;4) exceeded that of control nontransgenic mice (15≈plus;5, compare to Figure 3a1). Enhanced proliferative activity (more than twofold) of the epithelium as evidenced by PCNA staining was observed for the transgenic MHERc virgin mice (Figure 4). These data demonstrate an increased proliferation rate of the mammary epithelium found in virgin MHERc and BLGHERc transgenic mice. Furthermore, the ductal epithelium in glands of virgin MHERc mice was multilayered, probably corresponding to a higher mitotic index as assessed by PCNA labeling (Figure 4).

Figure 3
figure4

Histological examination of mammary glands from control and transgenic mice. (a1) Virgin control mouse. Note the appearance of only a few ducts surrounded by monolayered epithelium. (b) Virgin MHERc mouse (line 6903). Note the high number of epithelial islands with lumen mostly disappeared surrounded by multilayered epithelium. The number of ducts was counted in five microscopic fields. These histological patterns was shown to be very similar to that of the regressed mammary gland of both MHERc (a2) and BLGHERc (not shown) mice after three pregnancies. (c) Virgin BLGHERc mouse (line 116). Note the multilayered epithelium surrounding the lumen of ducts. (d) Lactating control gland. (e) Lactating gland of MHERc mouse (line 6903, after three pregnancies). Note the hyperplastic epithelium with adjacent trophic resting structures. (f) Lactating gland of BLGHERc mouse with hyperplastic alveoli (line 116). Note the multilayered epithelium and reduced lumen of the alveoli. (g) Mammary adenocarcinoma of a multiparous BLGHERc mouse (line 116, after three pregnancies). (h) Large dysplastic area of a mammary gland obtained from a MHERc mouse (line 6903, after two pregnancies). (i) Invasive mammary gland adenocarcinoma obtained from a multiparous MHERc mouse (line 7343, after four pregnancies). All pictures are taken with 200-fold magnification, with the exception of a2 where the magnification is 100-fold

Figure 4
figure5

Labeling index of mammary epithelial cells of hEGFR virgin transgenic and control mice. Labeling index was determined by analysing PCNA expression using a anti-PCNA specific antibody. Each bar represents the average number of mitotic cells counted in five independent microscopic fields

Mammary histology of lactating mice

The majority of lactating mammary glands from both MHERc and BLGHERc transgenic mice after three pregnancies exhibited glandular epithelial tissue corresponding to that of the control mice (Figure 3d). However, an overall increase in epithelial cell proliferation within lobules was observed in multilayered lobular hyperplastic and dysplastic epithelial ducts (Figure 3e,f, compare to the control represented in Figure 3d). Mammary glands were obtained from hEGFR transgenic mice on days 5, 10 and 16 of lactation after up to four pregnancies. In some lobules the transgenic mammary epithelium was multilayered and the typical cuboidal morphology of these cells was lost with cells being packed in clusters (Figure 3e). Regressed MHERc mammary gland showed a similar histology as the virgin gland (Figure 3a2).

The MHERc and BLGHERc strains exhibited a similar histology in the lactating mammary gland. In both transgenic strains, the number of poorly differentiated areas and areas consisting of multilayered lobular epithelium increased with the number of pregnancies. In addition, the number of cells located within the poorly differentiated clusters was increased with the number of pregnancies, which correlates with the high proliferation index in these areas. In one out of six multiparous BLGMHERc and in two out of six multiparous MHERc transgenic mice exhibiting dysplasia, small tubular adenocarcinomas were found (Figure 3g,h,i). The tumors exhibited nuclear anisomorphy and showed only irregular separation from adjacent non-carcinoma tissue.

Immunohistochemistry

Elevated hEGFR expression was detected in transgenic mammary glands, as evidenced by enhanced immunohistochemical staining of the epithelial cells within the ducts (Figure 5b–d). hEGFR expression in these regions was heterogeneous localized in discrete cellular clusters. In addition histological samples contained areas, which consisted of small ducts with unilayered epithelium surrounded by myoepithelium (Figure 5e). These regions appeared to resemble non-differentiated resting lobules with lower proliferation index of approximately 30–50% that had been counted by immunohistochemical proliferation cell nuclear antigen (PCNA) detection (data not shown). In these areas hEGFR expression was higher than in normal differentiated areas of the mammary gland where no immunostaining was detected (Figure 5e). The adenocarcinomas were characterized by high hEGFR expression as demonstrated by immunohistochemistry (Figure 5f) and northern analysis (Figure 2a).

Whole mount organ cultures of mammary glands

To ascertain whether hEGFR overexpression might influence mammary epithelial differentiation, whole mount organ cultures of the second thoracic mammary gland were obtained from progesterone and estradiol primed 4–6 week-old virgin female transgenic or control mice. To induce mammary gland differentiation in vitro mammary glands were cultured in a lactogenic hormone mixture, containing aldosterone, prolactin, insulin and hydrocortisone (APIH). Glands obtained from control mice that were cultured for 7 days in APIH medium developed ducts which terminated in lobuli consisting of fully developed alveoli (Figure 6a). However, mammary glands cultured in medium, containing aldosterone and insulin (AI-medium), developed ducts with little branching (data not shown). Functional differentiation was analysed by assessing whey acidic protein (WAP)-mRNA expression by Northern blot analysis (Figure 7). Control glands cultured in AI-medium did not express WAP (data not shown) whereas APIH-medium induced WAP expression (Figure 7, lane 1). The addition of EGF to the lactogenic APIH-medium decreased branching of the ducts and development of alveoli whereas the number of ducts was comparable to that in glands cultured in lactogenic medium without EGF (Figure 6b). In these EGF treated glands, WAP expression was down regulated (Figure 7, lane 2).

Figure 6
figure6

Differentiation study of the mammary gland using whole mount organ culture obtained from transgenic and control mice. Second thoracic mammary gland of transgenic and control mice were cultured for 7 days under different medium conditions. (a) Control gland cultured in lactogenic APIH-medium. (b) Control gland cultured in lactogenic APIH-medium in the presence of 10 ng/ml mEGF, note that morphologic differentiation is strongly inhibited. Both MHERc (c, line 6903) and BLGHERc (d, line 116) transgenic glands are cultured in lactogenic APIH-medium, note the complete reduction of alveoli and only poor branching of the main ducts

Figure 7
figure7

WAP expression analysed by Northern blot analysis. 15 μg total RNA were purified from second thoracic mammary gland of control and transgenic mice after 7 days in culture. WAP- and β-actin specific probes were employed. Tissue samples used in lanes 1–4 are identical to a–d in Figure 5. Whole mount culture has been analysed. Lane 1: control gland cultured in lactogenic APIH-medium, lane 2: control gland cultured in lactogenic APIH-medium supplemented with 10 ng/ml mEGF, lane 3: MHERc transgenic gland (line 6903), and lane 4: BLGHERc transgenic (line 116), both cultured in lactogenic APIH-medium. Hybridization with an actin probe was used as control to normalize the WAP expression

Mammary glands were obtained from both MHERc and BLGHERc transgenic virgin mice and cultured in lactogenic APIH-medium. These glands had reduced numbers of secondary ducts and alveoli (Figure 6c,d). Little morphological differentiation of these glands was observed. Only main ducts showing a few branches containing short secondary ducts with a few immature lobulo-aveolar structures developed (Figure 6c,d). The morphology of these undifferentiated glands was identical to glands obtained from nontransgenic mice that had been cultured in APIH-lactogenic medium containing EGF (compare to Figure 6b). Furthermore, functional differentiation as assessed by WAP mRNA expression was inhibited. WAP expression is impeded in both MHERc and BLGHERc transgenic glands after 7 days of culture (Figure 7, lanes 3, 4). Adding EGF to the lactogenic APIH-medium did not lead to further inhibition of functional and morphological differentiation (data not shown) suggesting that hEGFR overexpression is sufficient to block functional and morphological differentiation of the mammary gland.

Discussion

Several lines of evidence suggest an oncogenic role for the EGF-receptor in mammary tumors. EGFR overexpression could be a primary event in inducing transformation or a secondary modification facilitating growth and transformation, which had been demonstrated in in vitro systems (Di Fiore et al., 1987b). Together with other genetic alterations such as c-erbB-2 overexpression or estrogen receptor status, EGFR overexpression, due in large part to gene amplification, correlates with a poor prognosis for overall patient survival (Al-Kasspooles et al., 1993; Borg et al., 1991; Fox et al., 1994; Gramlich et al., 1994). Furthermore, EGFR overexpression can cooperate with other tyrosine kinase receptors to induce cell transformation (Mullaney and Skinner, 1992). These data suggest that the EGFR is one important factor for inducing tumorigenesis. However, the causal role played by the EGFR in the development of mammary tumors has yet to be defined.

We therefore set out to generate transgenic mice expressing the human EGFR under mammary gland specific promoters to test the hypothesis that the overexpression of the EGFR represents a primary event in inducing mammary tumors. We have generated independent transgenic mouse lines expressing the hEGFR cDNA under the control of the MMTV-LTR (MHERc) and the β-lactoglobulin promoter (BLGHERc), respectively. Hybrid genes comprising cDNAs are sometimes poorly expressed or are completely silent in transgenic animals. To circumvent these problems we used a transgene rescue approach (Yull et al., 1996). The hEGFR cDNA was cloned into the pBJ41 ß-lactoglobulin minigene-like expression construct and co-injected with the full length modified, translational inactive genomic sequence of the ß-lactoglobulin gene into fertilized oocytes. This transgene rescue is associated with transcription of the rescuing gene but does not require its translation and is thought to result in the co-activation of the EGF-receptor expression vector. In this regard, all transgenic lines generated with the BLGHERc constructs were double transgenic for both the pBJ41BLGHERc and the pBLGΔtr.

The phenotype observed in the lactating mammary glands of BLGHERc and MHERc mice was similar indicating that the transgene effect is promoter independent. Surprisingly, we observed that there was impaired mammary epithelial development in virgin BLGHERc mice. This phenotype was characterized by multilayered ductal epithelium. It has been demonstrated that the WAP promoter can be activated in a subpopulation of mammary epithelial cells in the virgin gland even though its main expression occurs during lactation (Hennighausen et al., 1994). If there exists a similar expression pattern for the BLG promoter, the human EGFR transgene could be expressed already during early development of the virgin mammary epithelium resulting in ductal hyperplasia.

Although the hEGFR transgene was expressed in salivary gland, ovary and testis of MHERc mice no apparent phenotypic changes were observed in these organs (data not shown). This is in contrast to other MMTV-v-Ha-ras, MMTV-c-neu and MMTV-heregulin transgenic mice that develop hyperplasias and/or tumors in these organs expressing the oncogenic transgene (Krane and Leder, 1996; Sinn et al., 1987). Interestingly, in comparison to these animals the latency period in hEGFR overexpressing transgenic mice to develop dysplasias is longer and strongly correlated with the number of pregnancies. A potential explanation for these findings is that, in the hEGFR transgenic mice, the level of hEGFR associated tyrosine kinase activity is below a certain threshold level of expression required for rapid transformation of mammary epithelial cells. In this respect, no tumors have been detected in virgin MHERc mice.

During involution, hEGFR expression driven by the BLG promoter is inactive while the expression driven by the MMTV-LTR promoter was relatively low as assessed, which we ascertained by immunohistochemistry and RT–PCR. In this respect, we propose that many cells expressing high level of the hEGFR transgene are removed during involution. However, the histology of the regressed hEGFR transgenic mammary gland is different compared to regressed control mammary gland (Figure 3a2). It is very similar to the histology of virgin transgenic mammary gland (Figure 3b,c). Transgenic mammary glands of lactating MHERc and BLGHERc mice exhibit atrophic areas with increased labeling index and pronounced hEGFR expression. The number of such areas containing ducts which lack free unobstructed lumens increases with the number of pregnancies. We conclude that these structures are not fully regressed during involution. It is possible that these hEGFR expressing atrophic areas accumulate during each successive pregnancy and may represent a potential stem cell population for malignant transformation.

The result from our transgenic study suggests that overexpression of the hEGFR in the mouse mammary gland can predispose animals to mammary epithelial cell transformation with a moderate degree of tumor development. Independent of the promoter used tumors develop stochastically after multiple pregnancies. EGF, one of the EGFR ligands, is transiently expressed during mammary gland development. This might restrict the proliferative stimulus of activated EGFR transgene exclusively to pregnancy and lactation. Taking into account that most of mammary epithelial cells undergo apoptosis after lactation it is not surprising that in hEGFR transgenic mice breast tumor development is correlated with the number of pregnancies. Cross breeding of hEGFR transgenic mice with other transgenic mice, such as TGF-α may confirm this hypothesis. We would expect a reduction of latency period and enhanced tumor formation in these bi-transgenic mice.

This study demonstrates that overexpression of the hEGFR is an important factor in promoting mammary epithelial transformation. EGF blocks the differentiation inducing effect of mammary derived growth inhibitor while stimulating monolayer growth and colony formation in soft-agar (Spitzer et al., 1995). Furthermore, hEGFR overexpression can also block functional differentiation in transgenic whole organ cultures. EGFR transgene overexpression inhibited ductal branching, development of lobulo-alveolar structures and WAP expression. In fact the morphology of the transgenic mammary gland is indistinguishable to wild type glands that have been cultured in lactogenic APIH-medium containing EGF. Therefore chronic stimulation of the EGFR pathway inhibits not only morphological differentiation of the mammary gland, but also functional differentiation while promoting epithelial cell proliferation, thereby facilitating cellular transformation and subsequent tumorigenesis. Additional genetic alterations are likely to contribute to the tumor phenotype in vivo in these hEGFR transgenic mice.

Materials and methods

DNA construction

Two constructs containing different mammary gland specific promoters driving the human EGFR cDNA were used for generating transgenic mice (Figure 1).

First, a 1.01 kb MMTV-LTR promoter/enhancer fragment was obtained by HincII/HindIII digestion from the pTG1 plasmid (kindly provided by Dr M Treuner, MDC-Berlin, Germany) and then cloned into the HindIII site of pBluescript plasmid (Stratagene, UK). The hEGFR cDNA linked to a SV 40 polyA+ site was obtained by XbaI/SacII digestion from the CVN/HERc plasmid (kindly provided by Dr A Ullrich, Max-Planck-Institute, Munich, Germany). The resulting 4 kb hEGFR cDNA (HERc), was then cloned into MMTV/LTR-pBluescript vector at the XbaI and SacII, resulting in the plasmid named pMHERc.

Second, two different constructs based upon the ß-lactoglobulin promoter (BLG) were used for a co-injection strategy (Yull et al., 1996). The plasmid pBJ41 (kindly provided by Dr J Clark, Rosslin Institute, Edinburgh) is an intronless minigene derived from the 10.5 kb sheep BLG gene with an EcoRV site introduced into the 5′ untranslated region (Whitelaw et al., 1991). In brief, the hEGFR cDNA was cut as described above from the CVN/HERc plasmid vector and was filled in by Klenow (Life Technologies, UK). Following alkaline phosphatase treatment, the 4.0 kb fragment was inserted into the EcoRV site of the pBJ41 plasmid, resulting in the plasmid pBJ41BLGHERc (Archibald et al., 1990; Barash et al., 1994; McClenaghan et al., 1995; Yull et al., 1996). All expression constructs were prepared from plasmid DNA as described previously (Theuring et al., 1995).

The construct BLGdeltatrans (BLGΔtr) which was co-injected with the pBJ41BLGHERc, was created from the BLG gene (10.5-kb SalI-XbaI-fragment pSSltgXS; (Simons et al., 1987)) by PCR-mediated mutagenesis. One mutation was the insertion of a TC after ATGA, where ATG is the initiation codon. This creates a Bcl I restriction site and a frameshift, which renders a TGA 66 bp downstream of the initiation into a stop codon by putting it into reading frame. In addition, adenine, preceding this TGA, was changed to thymidine in order to destroy a potential in-frame re-initiation site which might have led to the production of a non-secreted BLG. The PCR-amplified region was sequenced. Before establishing transgenic mice, the BLGdeltatrans (pBLGΔtr) construct was stably transfected into HC11 mouse mammary epithelial cells. Western blotting of transfected HC11 cell extracts failed to detect any sheep BLG expression (data not shown).

Generation and identification of transgenic mice

For DNA microinjection, pMHERc was released by KpnI/SacII digestion. The constructs pBJ 41BLGHERc and pBLGΔtr were obtained from the plasmids by SalI/XbaI digestion. Outbred NMRI mice were used for microinjection. Transgenic mice were generated and identified by previously described methods (Theuring et al., 1995).

MHERc transgenic mice

Transgene integration was analysed by PCR employing specific human EGFR primers: 5′-GAT CGG CCT CTT CAT GCG 3′, position 2174–2191 and 5′-TTC TTT CAT CCC CCT GAA TG-3′, position 3126–3146. This resulted in a characteristic band of about 1.0 kb.

BLGHERc transgenic mice

In parallel to specific human EGFR primers, specific sheep ß-lactoglobulin primers were used for PCR analysis: 5′-GCT TCT GGG GTC TAC CAG GAA-3′, position 579–599, and 5′-TCG TGC TTC TGA GCT CTG CA-3′, position 806–825, resulting in a 240 bp DNA band. Mice that were positive for both hEGFR and sheep ß-lactoglobulin sequences was considered to bi-transgenic.

RT–PCR and Northern blot analysis

Expression analysis in transgenic mice was performed by reverse transcription RT–PCR and Northern blot analysis. RNA was prepared by a single step method (Chomczynski and Sacchi, 1987). For RT–PCR analysis 25 μg of total RNA were converted to cDNA by reverse transcriptase (Life Technologies, UK). The PCR was performed as described before using ampliTaq DNA polymerase (Life Technologies, UK). For Northern analysis RNA was separated on a 1.25% agarose formaldehyde gel and transferred to a hybond N nylon membrane (Amersham). To detect specific transcripts, 32P-labeled cDNA probes were used for hybridization on the membranes. A 1.0 kb WAP cDNA fragment (kindly provided by Dr M Treuner, Max Delbrück Center, Berlin, Germany) was used for analysing WAP expression. For transgene expression a 1.8 kb hEGFR specific cDNA fragment was used.

Histological evaluation

Tissue for histology and immunohistology was fixed in 10% buffered formalin, embedded in paraffin, sectioned at 4 μm and stained with hematoxilin and eosin. The number of ducts was counted under the microscope (100-fold enlargement). Five random fields (microscopic views) were selected for one data point.

Immunohistochemistry

For hEGFR protein detection, paraffin embedded tissue was sectioned at 4 μm and deparaffinized. Immunostaining was performed using a biotin-streptavidin amplified detection system (Amersham, Germany). The hEGFR was detected specifically using a sheep anti-hEGFR peptide specific polyclonal antibody (Amersham, Germany). For antibody binding detection, a peroxidase (POD) conjugated secondary antibody combined with the metal enhanced DAB system was used (Pierce, UK). PCNA staining was performed using the AB-1 monoclonal antibody (Oncogene Science, CA USA). As described above, the POD system was used for detection (five microscopic fields were counted for labeled nuclei). All sections were counterstained with Meyer's hematoxilin.

Whole mount staining

Whole-mounted mammary glands were fixed in Ethanol/CHCl3/acetic acid (6 : 3 : 1), rehydrated and stained in carmine red solution [0.2% carmine red dye (w/v), 10 mM AlK(SO4)2]. After staining the slides are dehydrated and the glands were cleared with xylene and stored in methyl salicylate.

Whole mount organ culture

To culture whole second thoracic mammary glands of estradiol and progesterone-primed virgin mice in serum-free medium, we adapted the original protocol with minor modifications (Banerjee et al., 1973). Four 6 week-old virgin female NMRI mice were injected subcutaneously for 9 consecutive days with progesterone (1 mg) and estradiol (1 μg) (Sigma Chemicals Co., St. Louis, MO, USA) dissolved in gum arabicum/0.9% NaCl. The second pair of thoracic glands, including the primary duct, were excised and cultivated for 7 days on perforated cyclopore membranes (Falcon Plastics, Cocksville, MD, USA) in 2 ml medium 199 with Hank's salts (GIBCO, BRL, Gaithersburgh, MD, USA), gentamycin (40 μg/ml), aldosterone, bovine prolactin, insulin (each at 5 μg/ml: Sigma), and hydrocortisone (1 μg/ml: Merck, Darmstadt, Germany), referred to as APIH lactogenic medium. For some experiments hEGF (10 ng/ml: Sigma Chemicals Co., St. Louis, MO, USA) was added to the APIH medium. Culture medium was changed every 2 days. Three pairs of contralateral glands were examined routinely, one gland was used as control. Formalin-fixed whole glands were stained with hematoxilin as described (Rivera, 1967).

References

  1. Aaronson SA . 1993 Science 254: 1146–1153

  2. Al-Kasspooles M, Moore JH, Orringer MB and Beer DG . 1993 Int J Cancer 54: 213–219

  3. Andres AC, Van der Walk MA, Schonenberger CA, Fluciger F, LeMeur M, Gerlinger P and Groner B . 1988 Genes Dev 6: 1486–1495

  4. Archibald AL, McClenaghan M, Hornsey V, Simons JP and Clark AJ . 1990 Proc Natl Acad Sci USA 87: 5178–5182

  5. Banerjee MR, Wood BG and Kinder DL . 1973 In Vitro Cell Dev Biol 9: 129–133

  6. Barash IA, Faerman T, Ratovitsky R, Puzis M, Nathan DR, Hurwitz and Shani M . 1994 Transgenic Res 3: 141–151

  7. Baselga J and Mendelsohn J . 1994 Breast Cancer Res Treat 29: 127–138

  8. Borg A, Baldetorp B, Ferno M, Killander D, Olsson H and Sigurdsson H . 1991 Oncogene 6: 137–143

  9. Bouchard L, Lamarre L, Tremblay PJ and Jolicouer P . 1989 Cell 57: 931–936

  10. Chakrabarty S, Rajagopal S and Huang S . 1995 Clin Exp Metastasis 13: 191–195

  11. Chomczynski P and Sacchi N . 1987 Anal Biochem 162: 156–159

  12. Di Fiore PP, Pierce JH, Fleming TP, Hazan R, Ullrich A, King CR, Schlessinger J and Aaronson SA . 1987a Cell 51: 1063–1070

  13. Di Fiore PP, Pierce JH, Kraus MH, Segatto O, King CR and Aaronson SA . 1987b Science 237: 178–182

  14. Di Marco E, Pierce JH, Fleming TP, Kraus MH, Molloy CJ, Aaronson SA and Di Fiore PP . 1989 Oncogene 4: 831–838

  15. Fan Z, Masui H, Altas I and Mendelsohn J . 1993 Cancer Res 53: 4322–4328

  16. Fox SB, Smith K, Hollyer J, Greenall M, Hastrich D and Harris AL . 1994 Breast Cancer Res Treat 29: 41–49

  17. Goldenring JR, Tsunoda Y, Stoch SA, Coffey R and Modlin IM . 1993 Regul Pept 43: 37–44

  18. Gramlich TL, Cohen C, Fritsch C, Derose DB and Gansler T . 1994 Am J Clin Pathol 101: 493–499

  19. Hackel PO, Zwick E, Prenzel N and Ullrich A . 1999 Curr Opin Cell Biol 11: 184–189

  20. Harris AL and Nicholson S . 1987 Epidermal growth factor in breast cancer In Lippman ME and Dickson RB eds Regulatory Mechanisms in Breast Cancer Kluwer Boston pp 93–118

  21. Hennighausen L, Mcknight R, Burdon T, Baik M, Wall RJ and Smith GH . 1994 Cell Growth Differ 5: 607–613

  22. Hinrichs SH, Fontes JD, Bills ND and Schneider PD . 1991 Princess Takamatsu Symp 22: 259–274

  23. Klijn JG, Berns PM, Schmitz PI and Foekens JA . 1992 Endocr Rev 13: 3–17

  24. Krane IM and Leder P . 1996 Oncogene 12: 1781–1788

  25. Lee PP, Darcy KM, Shudo K and Ip MM . 1995 Endocrinol 136: 1718–1730

  26. Liu T, Chen J and Zeng C . 1995 Chin Med J Engl 108: 653–659

  27. Matsui Y, Halter S, Holt JT, Hogan BLM and Coffey RJ . 1990 Cell 61: 1137–1146

  28. McClenaghan M, Springbett A, Wallace RM, Wilde CJ and Clark AJ . 1995 Biochem J 310: 637–641

  29. Mullaney BP and Skinner MK . 1992 Mol Endocrinol 6: 2103–2113

  30. Muller WJ . 1991 Cancer Met Rev 10: 217–227

  31. Normanno N, Kim N, Wen D, Smith K, Harris AL, Plowman G, Colletta G, Ciardiello F and Salomon DS . 1995 Breast Cancer Res Treat 35: 293–297

  32. Rajagopal S, Huang S, Moskal TL, Lee BN, ElNaggar AK and Chakrabarty S . 1995 Int J Cancer 62: 661–667

  33. Rivera EM . 1967 Methods Mammal Embryol 31: 442–471

  34. Salomon DS, Ciardiello F, Valverius EM and Kim N . 1991 Cancer Treat Res 53: 107–157

  35. Simons JP, McClenaghan M and Clark AJ . 1987 β-lactoglobulin in transgenic mice Nature 328: 530–532

  36. Sinn E, Muller W, Pattengale P, Tepler I, Wallace R and Leder P . 1987 Cell 49: 465–475

  37. Spitzer E, Zschiesche W, Binas B, Grosse R and Erdmann B . 1995 J Cell Biochem 57: 495–508

  38. Theuring F, Aguzzi A, Kropp C, Wohn KD, Hoffmann S and Schleuning WD . 1995 Fibrinol Proteol 9: 277–287

  39. Tortora G, Ciardiello F, Ally S, Clair T, Salomon DS and Cho-Chung YS . 1989 FEBS Lett 242: 363–367

  40. Whitelaw CBA, Archibald AA, Harris S and McClenaghan M . 1991 Transgenic Res 1: 3–13

  41. Yee LD, Kacinski BM and Carter D . 1989 Semin Diagn Pathol 6: 110–125

  42. Yull F, Binas B, Harold G, Wallace R and Clark AJ . 1996 Transgenic Res 6: 7–11

Download references

Acknowledgements

The authors would like to give their gratitude to Mrs Renate Franke for excellent technical assistance in histology and immunohistochemistry. We thank Drs AJ Clark. (Rosslin Institute, Edinburgh, Scotland), M Treuner (MDC Berlin, Germany) and A Ullrich (Max-Planck-Insitute, Munich, Germany), for providing plasmids, and Dr R Junker (Novartis Inc., Basel, Switzerland) for additional pathological evaluation. Finally we thank Drs B Davies, R Schäfer and DS Salomon for critical reading the manuscript and several discussions promoting to finish the manuscript.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Franz Theuring.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Brandt, R., Eisenbrandt, R., Leenders, F. et al. Mammary gland specific hEGF receptor transgene expression induces neoplasia and inhibits differentiation. Oncogene 19, 2129–2137 (2000). https://doi.org/10.1038/sj.onc.1203520

Download citation

Keywords

  • adenocarcinoma
  • epidermal growth factor receptor
  • mammary gland
  • transgenic mice
  • differentiation

Further reading