Expression of receptors for epidermal growth factor and insulin-like growth factor I by ZR-75-1 human breast cancer cell variants is inversely related: the effect of steroid hormones on insulin-like growth factor I receptor expression.

We have investigated the expression of insulin-like growth factor I receptors (IGFR) by the ZR-75-1 human breast cancer cell line and tamoxifen-resistant (ZR-75-9a1) and oestrogen-independent (ZR-PR-LT) variants. ZR-75-1 cells expressed 6633+/-953 receptors per cell,(K(d) 0.24+/-0.06 nM). IGFR expression was reduced in ZR-75-9a1 cells (1180+/-614 receptors per cell, K(d) 0.13+/-0.05) and increased in the ZR-PR-LT cell line (18 430+/-3210 receptors per cell, K(d) 0.24+/-17). A comparison of these data with previously published findings for epidermal growth factor receptor (EGFR) expression by these cell lines revealed that IGFR and EGFR expression are inversely related in the variant lines whereas ZR-75-1 cells express similar numbers of both receptors. Since the changes in IGFR expression observed are associated with changes in steroid hormone receptor status, we also investigated the effects of oestradiol, the synthetic progestin ORG 2058 and dexamethasone on IGFR expression. Oestradiol increased IGFR expression only in the ZR-75-1 cell line. Low concentrations of ORG 2058 increased IGFR levels in the two cell lines positive for progesterone receptor (ZR-75-1 and ZR-PR-LT). High concentrations of ORG 2058 increased IGFR expression in all cell lines, as did dexamethasone. These data suggest that EGFR and IGFR expression may be linked in breast cancer, and that EGFR/IGFR ratios in breast cancer may be a more sensitive prognostic indicator than EGFR expression alone. Regardless of basal IGFR expression by the cell studied, ORG 2058 increased IGFR expression, possibly via both the progesterone and glucocorticoid receptors.

Epidermal growth factor (EGF) and insulin-like growth factor I (IGF-I) are potent mitogens in human breast cancer and both act via membrane-associated receptors with intrinsic tyrosine kinase activity. The influence of IGF-I in regulating breast cancer cell proliferation appears to be under steroid hormone control at a number of levels. Oestrogen has been reported to up-regulate IGF-I receptor (IGFR) expression, possibly sensitising tumour cells to the mitogenic effect of IGF-I (Stewart et al., 1990) and a positive relationship between oestrogen receptor (ER) and IGFR expression has been reported (Pekonen et al., 1988). Conversely, progestins appear to down-regulate IGFR numbers (Papa et al., 1991;Owens et al., 1993). Tamoxifen has been reported to reduce circulating levels of IGF-I (Pollack et al., 1992), and the pattern of expression of IGF binding proteins (IGFBPs), which can both attenuate and potentiate the actions of IGF-I, is influenced by both oestrogens and antioestrogens (Lonning, 1992;Owens et al., 1993;Lahti et al., 1994;Manni et al., 1994). It is generally accepted that tumour expression of receptors for epidermal growth factor (EGFR) is a powerful prognostic indicator in human breast cancer, with high EGFR numbers being associated with low ER content and poor clinical prognosis (Harris, 1989). Recent evidence has suggested that the EGF and IGF receptor systems may influence one another at several levels. Administration of EGF or oestradiol to ovariectomised mice increases uterine IGF-I mRNA production (Hana and Murphy, 1994) suggesting that activation of the IGF receptor system may be a common down-stream event for both oestradioland EGF-induced cell proliferation. It has been reported that EGF can regulate IGFBP expression (Andreatta van Leyen et al., 1994;Hembree et al., 1994). The relationship between EGFR and IGFR expression is less clear. Since a positive ER status is associated with a positivity for IGFR (Pekonen et al., 1988) and low EGFR levels (Harris, 1989), an inverse relationship between EGFR and IGFR expression might be anticipated. However such a relationship has not been clearly demonstrated to date (Pekonen et al., 1988).
We have previously shown that acquired tamoxifen resistance accompanied by loss of detectable ERs and progesterone receptors (PGRs) in the ZR-75-1 human breast cancer cell line (Van den Berg et al., 1989) is associated with an increase in EGFR expression (Long et al., 1992), in agreement with clinical findings (Harris, 1989). Conversely, an oestrogen-independent variant of the same cell line, which constitutively expresses high numbers of PGRs (Van den Berg et al., 1990), has a much reduced EGFR content (Long et al., 1992). In this study we have characterised IGFR expression by these cell lines and report that IGFR expression is inversely related to EGFR expression. In light of an accompanying positive association between IGFR and PGR expression, we have also investigated the effects of progestins on IGFR expression in these cell lines.

Cell lines
The ZR-75-1 human breast cancer cell line was obtained from Flow Laboratories, Irvine, UK. Cells were routinely maintained in RPMI-1640 medium supplemented with 5% fetal calf serum, 100 IU ml-' penicillin and 100 tg ml-1 streptomycin. ZR-75-9al cells are a tamoxifen-resistant variant of ZR-75-1 (Van den Berg et al., 1989) routinely maintained in RPMI-1640 medium in the presence of 8 4UM tamoxifen. All experiments described were carried out on cells that had been grown in tamoxifen-free medium for at least 3 days.
ZR-PR-LT cells are oestrogen independent (Van den Berg et al., 1990) and routinely maintained in medium lacking known oestrogenic activity, (RPMI-1 640 medium lacking 478 phenol red and supplemented with heat-treated and dextrancoated charcoal-stripped fetal calf serum). Steroid and EGF receptor status of the cell lines is routinely determined and the phenotypes originally reported ( Van den Berg et al., 1989Long et al., 1992) have proved to be stable.
Radioiodination of IGF-I Receptor grade IGF-I (10,g, Penisula Laboratories, St. Helens, UK) was iodinated using the iodogen method (Fraker and Speck, 1978) as previously described for EGF (Long et al., 1992). The radioiodination mixture was fractionated by reverse-phase high-performance liquid chromatography (HPLC) using a Waters Associates gradient system (Milford, MA, USA) fitted with an analytical Vydac C4 column. The eluting gradient was trifluoroacetic acid (TFA)/water (0.05%/99.95% v/v) to TFA/water/acetonitrile (0.05%/29.95%/70.0% v/v) and the flow rate was 1 ml min-'. Fractions (0.5 ml) were collected and 25 pl aliquots were taken for determination of radioactivity. Three major iodinated peak fractions were identified and the fractions covering these peaks were pooled, aliquoted and stored at -20°C. These peaks were assumed to represent monoiodination of IGF-I at each of the tyrosine residues present in the peptide. Of the three iodinated peaks separated by HPLC, two demonstrated similar specific binding to ZR-75-1 cells and variants. The third peak showed no specific binding and may contain IGF-I iodinated at Tyr', which has previously been shown to have low affinity for the IGF-I receptor (Schaffer et al., 1993).
['25I1]IGF-I binding to ZR-75-1 cells and variants IGFR expression by cells was determined using a whole cell binding assay as previously described (Long et al., 1992). Cells (2 x 105) were plated into 24-place multiwell dishes and allowed to attach for 24 h. ['25I]IGF-I binding was determined by replacing the medium with RPMI-1640 medium (0.5 ml) supplemented with 1% bovine serum albumin containing ['25I]IGF-I (0.01 -0.33 nM) in the absence or presence of a 100-fold excess of non-labelled IGF-I to determine non-specific binding. Following a 1 h incubation at 4°C medium was removed and wells rinsed twice with ice-cold phosphate-buffered saline. Aliquots of 500,l of IM sodium hydroxide were added to each well and plates incubated for 1 h at 37°C to extract radioactivity. Radioactivity was determined by scintillation counting, (LKB Wallac 1410 LSC). Maximum binding capacity (Bmax) and affinity (Kd) were calculated after linearisation of specific binding data (Keightly and Cressie, 1980).

Steroid hormone modulation of ['25I
]IGF-I binding Cells (5 x 104) were plated into 24-place multiwell dishes and allowed to attach for 24 h. Medium was then replaced with medium containing a range of concentrations of oestradiol, the synthetic progestin ORG 2058 (Amersham International) or dexamethasone (Sigma-Aldrich, Poole, Dorset, UK). Cells were incubated for 6 days and [125I]IGF-I binding assessed as described above using a single concentration of ['25I]IGF-I (0.2 nM). Data is presented as ['251]IGF-I binding as a percentage of control after correcting for changes in cell numbers in the treated groups. Oestradiol treatment (10-8_ 10-7 M) increased ZR-75-1 cell numbers by <10%, decreased ZR-PR-LT cell numbers by 10-40% (Van den Berg et al., 1990) and had no significant effect on ZR-75-9al cells (Van den Berg et al., 1989). Dexamethasone and ORG 2058 reduced cell numbers in all cell lines by 5-40% (Van den Berg et al., 1993).

Statistical analysis
All experiments were carried out in triplicate and data analysed by a one-way analysis of variance using the Student Newman -Keuls test. Figure 1 shows the concentration-dependent binding of ['251]IGF-I to ZR-75-1 cells. Non-specific binding was typically less than 10% and linearisation of binding data suggested a single class of specific binding site. Maximum binding capacity and ligand affinity for the receptor for the three cell lines studied is shown in Table I. The oestrogen independent ZR-PR-LT line expressed approximately three times the number of IGFR compared with the parent ZR-75-1 line. Conversely, IGFR numbers were greatly reduced in the tamoxifen-resistant ZR-75-9al line, there being a more than 15-fold difference between IGFR numbers in this cell line compared with ZR-PR-LT cells. These changes in IGFR expression were not accompanied by any significant change in ligand affinity for the receptor (Table I). Figure 2 compares IGFR expression in the three cell lines with previously published data for EGFR expression (Long et al., 1992). It can be seen that there is a clear inverse relationship between IGFR and EGFR expression in the variant cell lines ZR-PR-LT and ZR-75-9al while ZR-75-1 cells express similar numbers of both receptors.    We have previously shown that acquisition of tamoxifen resistance as exemplified by the ZR-75-9al human breast cancer cell line is associated with an increase in the expression of EGFR compared with the parent ZR-75-1 line (Long et al., 1992) and loss of ER and PGR (Van den Berg et al., 1989). This finding was in accord with clinical observations that ER-negative/EGFR-positive human breast cancers have a poor prognosis and are resistant to tamoxifen treatment (Nicholson, 1988;Harris, 1989). In contrast, the oestrogenindependent ZR-PR-LT line has much reduced EGFR numbers accompanying elevated PGR expression (Van den Berg, 1990). In this study we have shown that these changes in EGFR expression associated with tamoxifen resistance and oestrogen independence respectively are paralleled by opposite changes in IGFR expression (Table I). While the parent cell line expresses similar numbers of EGFR and IGFR, in the variant lines, EGFR and IGFR expression is inversely related (Figure 2). These data suggest that EGFR and IGFR expression in ZR-75-1 cells are linked, with changes in the level of expression of one receptor being reflected by an opposite change in the expression of the other. There is evidence to support the concept of receptor crosstalk for EGFR and IGFR. Administration of EGF to ovariectomised mice increases uterine IGF-I mRNA production (Hana and Murphy, 1994) and it has been reported that EGF can regulate IGFBP-3 expression, thus sensitising cells to the effects of IGF-I. (Andreatta van Leyen et al., 1994;Hembree et al., 1994). It is possible that EGF sensitising of cells to IGF may occur more effectively in the face of high EGFR/low IGFR levels, while sensitisation would be less effective when EGFR numbers are low, and perhaps unnecessary when corresponding IGFR numbers are high.

Results
Since there is good evidence that ER and EGFR expression is inversely related in breast cancer, (Nicholson, 1988;Harris, 1989) and there is a positive relationship between ER and IGFR expression (Pekonen et al., 1988;Railo et al., 1994), it might be expected that an inverse relationship between IGFR and EGFR expression would exist. Clinical studies have failed to establish such a relationship (Pekonen et al., 1988;Foekens et al., 1989). The reasons for this are unknown, but may be due to inappropriate cut-off points for receptor positivity, occupation of receptors by endogeneous ligands and other factors. To our knowledge, our data are the first to show an inverse relationship between IGFR and EGFR expression in human breast cancer cell lines in vitro. Whereas EGFR expression is elevated only 3-fold in the ZR-75-9al tamoxifen-resistant line compared with the parent line (Figure 2), the EGFR/IGFR ratios in the two cell lines are 0.75 and 12.5 respectively. Should such a relationship exist in vivo it is possible that an EGFR/IGFR ratio may provide a more sensitive prognostic indicator for antioestrogen resistance than EGFR expression alone.
We have shown that oestradiol increases IGF binding by the oestrogen-sensitive ZR-75-1 cell line, (Figure 3), in agreement with earlier studies (Stewart et al., 1990). As expected, oestradiol was without effect in the ZR-75-9al line, which lacks oestrogen receptors (Van den Berg et al., 1989). Although the mechanism is unknown, the observation that low concentrations of oestradiol reduces ['25I]IGF-I binding by ZR-PR-LT cells would be consistent with our earlier findings that oestradiol inhibits ZR-PR-LT cell proliferation (Van den Berg et al., 1990). Elevated expression of IGFR by the oestrogen-independent ZR-PR-LT line is consistent with the observation that another oestradiol-induced protein (PGR) is also overexpressed in this cell line in the absence of oestrogenic stimulation (Van den Berg et al., 1990).
[1251]IGF-I binding is also increased in all cell lines by the synthetic progestin ORG 2058, (Figure 4) [I251]IGF-I binding seems to be primarily glucocorticoid receptor mediated. Confirmation of the relative contributions of PGR and the glucocorticoid receptor in mediating the effects of ORG 2058 described will require the use of specific progestin and glucocorticoid antagonists.  (Papa et al., 1991). However, a different cell line (T47D) and different progestins (progesterone and R5020) were used in the latter study and the down regulation of IGFR reported was attributed to a progestin-induced increase in IGF-II secretion. The failure of the present study to demonstrate IGFR down regulation may in part be explained by the report that ZR-75-1 cells do not secrete IGF-II (Osborne et al., 1989). Taken together with these earlier studies, our findings emphasise the complexity of potential interactions between steroid and peptide growth factor receptors. To our knowledge, this report is the first to indicate that IGFR expression may also be increased by glucocorticoids, and that high concentrations of a progestin may increase ['251]IGF-I binding via the glucocorticoid receptor. The physiological significance of these findings is unclear, as progestins and glucocorticoids are generally growth inhibitory towards breast cancer cells in vitro and down regulation of a receptor for a potent mitogen such as IGF-I would be more consistent with these anti-proliferative effects. In this context it is of interest that progestins and glucocorticoids have also been shown to increase EGFR expression in a number of human breast cancer cell lines (Ewing et al., 1989).
In conclusion, we have demonstrated an inverse relationship between EGFR and IGFR receptor expression by human breast cancer cells in vitro. Regardless of basal IGFR expression by the cell lines studied, [1251]IGF-I binding is increased following exposure to a progestin and this effect may be mediated via both PGRs and glucocorticoid receptors.