Interactions between oestradiol and epidermal growth factor in
`endometrial stromal proliferation and differentiation
`S. J. Mellor and E. J. Thomas
`Department of Obstetrics and Gynaecology, University of Southampton, Princess Anne Hospital,
`Coxford Road, Southampton SOl6 5YA, UK
`
`The relationship between oestradiol and epidermal growth factor (EGF) in the control of
`endometrial proliferation and differentiation in cultures of human endometrial stromal
`cells was investigated. Oestradiol at a concentration of 10 nmol
`1 \m=-\1increased the
`incorporation of both [3H]thymidine and [3H]leucine but the differences were significantly
`different from control only for [3H]leucine incorporation. Concentrations of 0.16, 1.6 and
`16 nmol EGF 1 \m=-\1 significantly increased both [3H]thymidine (P < 0.01) and [3H]leucine
`inhibited any
`incorporation (P<0.01). The pure steroidal antioestrogen,
`ICI 182,780,
`increase in [3H]thymidine and [3H]leucine incorporation stimulated by oestradiol
`in
`endometrial stroma. The monoclonal antibody, ICR 16, directed against the EGF receptor
`indicating that, in this model system,
`did not inhibit the oestradiol action in stromal cells,
`oestradiol does not act by inducing synthesis or release of EGF. However, ICI 182,780
`potently inhibited the incorporation of [3H]thymidine stimulated by EGF in endometrial
`stromal cells, suggesting interdependence between oestradiol and EGF in the control of
`endometrial stromal proliferation. Oestrogen-free conditioned medium from endometrial
`[3H]thymidine or
`[3H]leucine incorporation,
`stromal cultures did not stimulate either
`suggesting that oestradiol did not stimulate the secretion of a trophic factor
`from
`endometrial stromal cells.
`
`Introduction
`The precise mechanisms involved in endometrial proliferation
`have been studied intensely in recent years, owing to the
`emerging consensus that oestrogen may not act only through
`its receptor. Many workers have failed to demonstrate a
`proliferative response to oestrogen in endometrial cell cultures
`in vitro, of either epithelial cells or stromal cells (Casimiri el al,
`1980; Fleming and Gurpide, 1982; Alkhalaf et al, 1991; Haining
`et al, 1991; Uchima et al, 1991), although oestrogen will cause
`proliferation of either epithelial cells or stromal cells under
`certain culture conditions (Gerschenson et al, 1981; Irwin et al,
`1991; Olive et al, 1991). One potential candidate as the agent
`which controls endometrial proliferation is epidermal growth
`factor (EGF). Administration of EGF to ovariectomized mice
`induces uterine and vaginal growth equivalent
`to that
`in
`control mice given oestradiol, and this action of oestradiol
`is
`inhibited by an antibody directed against EGF (Nelson et al,
`In this model, EGF also induces the expression of
`1991).
`lactoferrin, a major oestrogen-inducible secretory protein,
`in
`mice. EGF (Haining et al, 1991) and its receptor (EGF-R; Smith
`et al, 1991) have been identified in human endometrium.
`the concentrations of EGF
`Studies have indicated that
`(Gonzalez et al, 1984) and EGF-R (Bonaccorsi et al, 1989;
`Stancel et al, 1990; Taketani and Mizuno, 1991; Troche et al,
`*Correspondence.
`Received 20 December 1994.
`
`1991) are regulated by oestrogen in the endometrium. Trans¬
`forming growth factor (TGF- ) has also been shown to
`mediate oestrogen action in the mouse uterus (Nelson el al,
`1992).
`this study was to investigate further
`the
`The aim of
`relative roles of oestradiol and EGF in endometrial prolifera¬
`tion. Any possible interdependence between oestrogen and
`EGF in endometrial growth was examined using specific
`their action in a model system of cultured
`inhibitors of
`human endometrial stromal cells. Endometrial stromal cells
`were chosen because they are known to express oestrogen
`and EGF receptors, and because they yield an extremely pure
`population of cells, which facilitates experimental design and
`interpretation. The actions of oestradiol were inhibited by
`using the pure specific antioestrogen ICI 182,780 (Wakeling
`ICR16, a monoclonal antibody directed against
`et al, 1991).
`(Modjtahedi et al, 1993), was used to
`the EGF receptor
`inhibit EGF action. This strategy enabled examination of
`two further hypotheses. First,
`if oestrogen acts on the
`endometrium to release a trophic factor such as EGF, which
`then stimulates replication of endometrial stromal cells,
`then
`treatment of stromal cultures with ICI 182,780 should inhibit
`production of this factor and thus any proliferative effect of
`such a factor should be abolished. Second, if oestrogen causes
`endometrial growth indirectly via the action of EGF,
`then
`cell cultures with the anti-EGF-R
`treatment of stromal
`antibody should inhibit any actions of oestrogen.
`
`MYLAN PHARMS. INC. EXHIBIT 1015 PAGE 1
`
`

`
`Materials and Methods
`
`I
`
`Materials
`Human recombinant EGF was obtained from Boehringer
`Mannheim (Lewes). Collagenase type XI,
`insulin, oestradiol,
`Hepes, L-glutamine,
`thymidine and leucine were all obtained
`from Sigma (Poole). Percoli was obtained from Pharmacia
`(Milton Keynes). Hank's balanced salt solution without
`red and 10 000 iu penicillin
`calcium, magnesium or phenol
`ml ~ :-10 000 pg streptomycin ml ~
`were obtained from
`Northumbria Biologicals Ltd (Cramlington). Dulbecco's modi¬
`fied Eagle's medium, Ham's nutrient mixture F12, amphotericin
` (fungizone), gentamycin and nonessential amino acids were
`obtained from Gibco (Paisley). A custom-made mixture of 1:1
`thymidine,
`leucine or phenol
`red was
`DMEM:F12 without
`also supplied by Gibco. Charcoal-treated fetal calf serum was
`obtained from Imperial Laboratories (Andover). Trichloroacetic
`acid (TCA) was provided by BDH (Poole) and Optiscint HiSafe
`from Wallac (Milton Keynes). All other chemicals were
`obtained from either BDH or Sigma.
`[3H]thymidine, L-(4,5)[3H]leucine and the Biotrak
`Methyl
`EGF ELISA kit were all purchased from Amersham Inter¬
`national (Amersham, Bucks).
`(7a[9-(4,4,5,5,5-pentafluoropentylsulphinyl)
`ICI
`182,780
`nonyl]estra-l,3,5(10)-triene-3,17ß-diol) has potent and pure
`antioestrogenic activity both in vitro (Wakeling et al, 1991) and
`in vivo (Thomas et al, 1994). The rat monoclonal antibody,
`inhibits EGF activity by blocking the receptor
`ICR16,
`(Modjtahedi el al, 1993). The control antibody used was
`ALN/ll/53, as described by Modjhedi et al (1993).
`
`Tissue samples
`Endometrial biopsies were obtained from patients under¬
`going either hysterectomy or surgical procedures for benign
`gynaecological disorders using a Pipelle suction curette
`(Eurosurgical; Cranleigh). None of the patients was receiving
`any hormonal
`treatment before surgery, and the operations
`were performed for menorrhagia or fibroids. The phase of the
`menstrual cycle at which the specimens were obtained was
`determined by the date of the last menstrual period. Of the
`samples used in the experiments, 57% were from day 6 to day
`14 of the cycle and the remaining 43% from day 15 to day 28.
`Tissue samples were collected in Hank's balanced salt solution
`without calcium, magnesium or phenol red, supplemented with
`20 mmol Hepes 1 ~ \ 100 iu penicillin ml ~ 1, 100 pg strepto¬
`mycin ml_ 1 and 5 pg amphotericin ml ~
`(medium A).
`Further processing of the tissue samples occurred in the cell
`culture laboratory.
`AH patients gave their consent to the collection of tissue
`specimens. Ethical permission for this study was granted by
`Southampton and South West Hampshire Health Authority/
`University of Southampton Joint Ethics Committee.
`
`1
`
`Isolation and culture of endometrial stromal cells
`Endometrial stromal cells uncontaminated with glandular
`epithelial cells were isolated using the method described by
`
`~
`
`Mellor and Thomas (1994). Endometrial tissue was minced and
`digested with 1 mg collagenase type XI ml
`until individual
`endometrial glands were released from the preparation. Un¬
`digested pieces of tissue and endometrial glands were removed
`from the digest by a two-step filtration procedure through
`45 pm and 10 pm nylon meshes (Lockertex, Warrington). The
`filtrate was centrifuged (at 500 # for 15 min) over 60% Percoli
`to remove contaminating red blood cells. After washing the
`cells contained in the band at the interface of the Percoli and
`medium A, cells were resuspended in 1:1 DMEMT12 contain¬
`ing 20 mmol Hepes 1 ~ l, 100 pg streptomycin ml ~ , 100 iu
`penicillin ml ~
`, 5 pg amphotericin ml
`~ , 1% (v/v) non-
`essential amino acids, 50 pg gentamycin ml-1, 10 mmol
`L-glutamine 1~1, 10 pg insulin ml ~
`1 and 5% (v/v) charcoal-
`treated fetal calf serum (medium C). After counting by
`haemocytometry, cells were plated out at a concentration of
`1 IO5 ml ~
`in 24-well plates (Nunc, Paisley). The stromal
`cells were allowed to adhere overnight in a humid atmosphere
`of 95% C02:5% 02 at 37°C. On the following day,
`this
`medium was removed and replaced with the custom-made
`phenol red-free 1:1 DMEM:F12 containing all
`the additions
`listed above. Cells were incubated in this medium for a further
`3 days until experiments began. The medium was replenished
`every 3 days.
`
`1
`
`[3H]thymidine and [3H]leucine incorporation studies
`All experiments were performed in phenol red-free medium
`C. Oestradiol and ICI 182,780 were dissolved in ethanol, and
`appropriate dilutions prepared in culture medium. The final
`in all wells was adjusted to 0.01%
`concentration of ethanol
`(v/v). Experiments were commenced on day 4 of culture.
`Stromal cells were incubated with the agent(s) of
`interest
`(oestradiol, ICI 182,780, EGF or anti-EGF-R antibody) for 48 h.
`Experimental media are replaced after 24 h. Methyl [3H]thymi-
`dine or L-4,5-[3H]leucine was added for the final 18 h of the
`experiment at a concentration of 5 pCi ml~ J (185 kBq ml" J).
`At the conclusion of the experiment, the medium was removed;
`the cells were washed in medium A and removed from the
`plates with 0.05% (v/v) trypsin-0.02% (v/v) EDTA and by
`scraping. The radioactive macromolecules were precipitated by
`the addition of TCA to a final concentration of 5% (w/v)
`([3H]thymidine experiments) or 10% (w/v) ([3H]leucine experi¬
`ments). After at least 1 h at 4°C, the precipitates were collected
`on Whatman GF/C filters (BDH) and air-dried. The liquid
`scintillant Optiscint HiSafe was added and the radioactivity on
`the filters determined by scintillation counting (LKB). For the
`ICR16 antibody experiments,
`the stromal cells were preincu¬
`bated with the antibody for 2 h before to the addition of EGF.
`
`Preparation of conditioned media
`Cells were plated out in 75 cm2 flasks (Nunc) at an equiva¬
`lent cell density to that used for the experiments in 24-well
`plates. The same experimental protocol as in the labelling
`studies described above was used for
`the preparation of
`conditioned media. Cells were incubated in phenol
`red-free
`medium C under the following experimental conditions: con¬
`trol (ethanolic vehicle alone), 10 nmol oestradiol 1 ~ *, 10 nmol
`
`MYLAN PHARMS. INC. EXHIBIT 1015 PAGE 2
`
`

`
`Table 1. Effect of various concentrations of oestradiol and epidermal growth factor (EGF) on [3H]thymidine
`and [3H]leucine incorporation in endometrial stromal cell cultures
`
`Parameter
`
`[3H]thymidine incorporation
`
`[3H]leucine incorporation
`
`Oestradiol
`I ~1)
`(nmol
`
`Percentage of
`control (sem)
`
`EGF
`(nmol 1
`
`~ ')
`
`Percentage of
`control (sem)
`
`0.01
`0.1
`1
`10
`
`0.01
`0.1
`1
`10
`
`75.6 (10.9)
`(4.1)
`94
`130.1 (35)
`(19.7)
`129
`
`(8.8)
`95
`(5.8)
`107.5
`(3.3)
`101.4
`111.7 (6.2)*
`
`0.16
`1.6
`16
`
`0.16
`1.6
`16
`
`248
`238
`274
`
`(29)*
`(29)**
`(48)***
`
`(4.6)
`134.5
`(5)**
`142.2
`133.9 (4)**
`
`Data are expressed as percentage of the control value (control = 100%) and are a mean of four experiments for oestradiol and three
`for EGF.
`*P<0.05, **P<0.01, ***P< 0.001.
`
`\ or 100 nmol ICI
`1 plus 100 nmol ICI 182,780 1
`oestradiol 1
`182,780 1_1 alone. The media for each treatment were
`collected after 24 and 48 h and then pooled. Residual oestradiol
`and ICI 182,780, which might have interfered with the inter¬
`pretation of experimental data, were removed from the
`conditioned media by ultrafiltration using Centriprep concen¬
`trators (Amicon, Stonehouse, Gloucestershire) with a mem¬
`brane cut-off of 3 kDa. After ultrafiltration, the concentrate was
`diluted to the original volume using phenol red-free,
`insulin-
`free medium C that contained only 1% (w/v) charcoal-treated
`fetal calf serum, and filtered through 0.22 pm filters. Oestradiol
`assay (Serono Diagnostics, Fleet) performed on the conditioned
`revealed
`media that initially contained 10 nmol oestradiol 1 ~
`that, after ultrafiltration,
`the concentration of oestradiol was
`400 pmol 1 ~ I, a concentration that was ineffective on [3H]thy-
`[3H]leucine incorporation in the assay system
`midine or
`described here.
`A small aliquot of each conditioned medium was taken for
`EGF assay. The remainder was applied to fresh stromal cell
`cultures for [3H]thymidine incorporation studies using the basic
`experimental protocol.
`
`EGF assay
`Aliquots of the conditioned media were assayed for EGF
`with the EGF Biotrak ELISA.
`
`Western blotting
`Endometrial stromal cells in 75 cm2 flasks were treated with
`10 nmol oestradiol l-1 either alone or in combination with
`100 nmol ICI 182,780 1~ \ as in the protocol described above,
`for 48 h. At the end of the experiments, cells were removed by
`trypsinization and scraping and collected by centrifugation at
`110 £ for 5 min. The approximate volume of the pellet was
`determined and an 8 times excess of Tris-EDTA-molybdate
`(10 mmol Tris 1, 1.5 mmol EDTA \~\ 5.0 mmol
`buffer
`sodium molybdate 1 ~ :) containing 1 mmol monothioglycerol
`was added. After thorough vortex mixing, the tubes were
`1 ~
`
`*
`
`centrifuged at 4°C for 20 min at 2500 g. Cytosols were
`analysed by western blotting to detect the oestrogen receptor
`by S. Dauvois and M. G. Parker (Imperial Cancer Research
`Fund, London).
`
`Statistical analysis
`the data were analysed statistically using Student's
`All
`paired t test.
`
`Results
`
`Culture of endometrial stromal cells
`Characterization of endometrial stromal cells isolated using a
`two-step filtration method is discussed by Matthews et al
`(1992). Filtration of the tissue digest through a 10 pm filter
`results in preparations of stromal cells virtually uncontaminated
`with glandular epithelial cells. Cells were plated out in phenol
`red-containing medium C. After the cells had adhered to the
`plastic of the culture vessels (overnight), the cells were washed
`in medium A and incubated in phenol red-free medium C. The
`brief period in which the cells were exposed to phenol red did
`not affect cellular responses to oestradiol, since cells spent a
`further 3 days in phenol red-free medium C. The stage of the
`menstrual cycle at which the cells were isolated did not appear
`the experimental data, although the small numbers
`to affect
`used in each experiment may have concealed differences.
`
`Effect of oestradiol and EGF on [3H]thymidine and fHjleucine
`incorporation in endometrial stromal cells in culture
`1 increased both
`Table 1 shows that 10 nmol oestradiol
`1 ~
`[3H]thymidine and [3H]leucine incorporation, but that only the
`increase in [3H]leucine incorporation was statistically significant
`(P=0.15 and =0.04, respectively). EGF significantly stimu¬
`lated DNA and protein synthesis at both 1.6 (P<0.01) and
`
`MYLAN PHARMS. INC. EXHIBIT 1015 PAGE 3
`
`

`
`Fig. 2. Western blot of cytosolic extracts of human endometrial
`stromal cells and immunodetection of the oestrogen receptor. Track 1,
`control cells;
`1 ~1;
`track 2, 10 nmol oestradiol
`track 3, 10 nmol
`" 1; track 4, 100 nmol ICI
`oestradiol and 100 nmol ICI 182,780 1
`182,780 1~ l; track 5, positive control corresponding to approximately
`~ \
`200 fmol oestrogen receptor 1
`
`the oestrogen receptor was completely
`This induction of
`blocked by co-incubation of 10 nmol oestradiol
`1
`* and
`ICI 182,780 1. Administration of 100 nmol
`100 nmol
`ICI
`182,780 1_I alone did not induce expression of the oestrogen
`receptor.
`
`~
`
`Co-incubation of EGF with the anti-EGF-R antibody ICR16
`Table 2 shows that 5 pg ICR16 ml"1 inhibited by 50% the
`incorporation of [3H]thymidine stimulated by 1.6 nmol EGF
`completely abolished
`1 ~1
`and that
`the response was
`. Because of the large quantities of
`with 10 pg ICR16 ml ~
`antibodies required, a limited number of control wells were
`included to conserve supplies. A concentration of 5 pg ml ~
`1 of
`a control antibody in combination with 1.6 nmol EGF l-1 did
`not diminish the response to 1.6 nmol EGF 1 ~ 1. However,
`10 pg ml~ of the control antibody decreased the amount of
`radioactivity to 50% of the control value.
`
`110 -
`
`(a)
`
`100
`
`90
`
`-
`
`80
`
`70
`
`60
`
`10"
`
`10"
`
`10"
`
`10
`
`10"
`
`100
`
`(b)
`
`80
`
`60
`
`40
`
`20
`
`coo
`
`.
`
`g1
`
`CD9
`
`o
`
`Q.
`
`10
`
`10"
`1CT
`10
`10
`Log10 concentration of ICI 182,780 (mol I
`Fig. 1. Effect of various concentrations of ICI 182,780 and 10 nmol
`oestradiol 1 on (a) [3H]leucine and (b) [3H]thymidine incorporation
`in endometrial stromal cell cultures. Values are mean ( ± sem) and
`three experiments ([3H]thymidine) or six experiments
`represent
`(l3H]leucine). *P< 0.05, **P< 0.01 compared with 10 nmol oestradiol
`1 alone.
`1
`
`~
`
`1
`1 (P<0.01), but 0.16 nmol EGF 1
`16 nmol
`only [3H]thymidine incorporation (P < 0.05).
`
`1 augmented
`
`Inhibition of oestradiol-stimulated [3Hjthymidine and [3H]leucine
`incorporation by ICI 182,780
`When stromal cells were incubated with various concen¬
`trations of ICI 182,780 in the absence of oestradiol, [3H]thymi-
`dine and [3H]leucine incorporation did not differ from control
`values. Co-incubation with various doses of
`ICI 182,780
`inhibited any stimulation of 10 nmol oestradiol
`1
`J
`on both
`[3H]leucine and [3H]thymidine incorporation (Fig. 1). Further¬
`there was a suppression of
`this incorporation below
`more,
`control values, which was statistically significant at concen¬
`trations of ICI 182,780 of 100 nmol x and 10 nmol : for
`the [3H]thymidine experiments and 1000 nmol ICI 182,780 1" 1
`for the [3H]leucine experiments (P< 0.05). Figure 2 is a western
`blot analysis for oestrogen receptors showing that oestrogen
`receptors are absent
`from stromal cells treated with vehicle
`alone, but are induced by treatment of cells with 10 nmol
`oestradiol 1_I for 48 h, estimated at 25 fmol protein mg-1.
`
`~
`
`MYLAN PHARMS. INC. EXHIBIT 1015 PAGE 4
`
`

`
`Table 2. Effect of the antibody ICR16 on EGF-induced [3H]thy-
`midine incorporation in endometrial stromal cultures
`
`Addition
`
`1.6 nmol EGF *
`1.6 nmol EGF I" 75 µg ICRlóml"1
`7l0 µg ICR16 ml
`1.6 nmol EGF 1
`75 µ& control antibody ml
`1.6 nmol EGF 1
`10 µ control antibody ml" 2
`
`*
`
`"
`
`"
`
`"
`
`[3H]thymidine
`incorporation:
`percentage of
`control (sem)
`
`199
`141
`98
`163
`41.5
`
`(24.0)**
`(11.7)*
`(3.8)
`(4.6)***
`(0.45)****
`
`'
`
`"
`
`Data are expressed as percentage of the control value (control = 100%) and
`*P<0.05,
`**P<0.02,
`are means of
`***P<0.01,
`three experiments.
`****P< 0.001.
`
`was
`
`~
`
`'
`
`Co-incubation of oestradiol and ICR16 anti-EGF-R antibody
`Endometrial stromal cells were preincubated with 10 pg
`I
`for 2 h before 10 nmol oestradiol
`ICR16 ml ~
`I
`It was not possible to demonstrate
`added to the cultures.
`any blockage of oestradiol-induced increase in radiolabelled
`incorporation by the presence of
`the anti-EGF-R antibody
`at concentrations that effectively abolished the effects of
`1.6 nmol EGF 1
`. This may well have occurred because of the
`difficulty of demonstrating any stimulatory effects of 10 nmol
`on [ HJthymidine incorporation. However,
`oestradiol
`1 ~
`in
`[3H]leucine incorporation experiments,
`a concentration of
`10 pg ICR16 ml ~ did not reduce the stimulation of protein
`synthesis caused by 10 nmol oestradiol 1 ~ 1 (Table 3).
`
`1
`
`"
`
`
`
`Conditioned medium experiments
`Conditioned media were prepared from stromal cells treated
`with either vehicle alone, 10 nmol oestradiol 1~ , 10 nmol
`1 plus 100 nmol ICI 182,780 1
`or 100 nmol ICI
`oestradiol 1 ~
`182,780 1" 1 alone. These conditioned media were then applied
`to fresh stromal cultures and the effect of
`the conditioned
`media on [3H]thymidine incorporation was determined accord¬
`ing to the standard experimental protocol. There was no differ¬
`ence in the values for the amount of [3H]thymidine between
`stromal cells incubated in conditioned medium obtained from
`cells treated with ethanolic vehicle alone and those from cells
`treated with 10 nmol oestradiol 1"
`. No EGF was detected in
`any of the conditioned media (Table 4). Since the limit of
`~ , this confirms that the
`sensitivity of the assay was 0.2 pg ml
`[3H]thymidine counts above control
`lack of stimulation of
`values by conditioned media from oestradiol-treated stromal
`cells is due to the absence of any oestradiol-induced synthesis
`or release of EGF. Conditioned media from cells treated with a
`1 ~: and 100 nmol
`combination of 10 nmol oestradiol
`ICI
`182,780 1_I seemed to depress the amount of radioactivity
`present to below that of cells treated with vehicle alone.
`
`Co-incubation of EGF and ICI 182,780
`Stromal cells were incubated with 1.6 nmol EGF 1_I and
`ICI 182,780 using the standard
`various concentrations of
`
`experimental protocol.
`ICI 182,780 inhibited EGF action in
`endometrial stromal cells (Fig. 3). All the concentrations of ICI
`182,780 tested were effective at reducing the stimulation of
`[3H]thymidine incorporation into DNA by 1.6 nmol EGF 1_1
`(P < 0.05). There was no significant reduction in EGF-induced
`stimulation of [3H]leucine incorporation by ICI 182,780.
`
`Discussion
`This study describes a series of experiments carried out
`to
`clarify the interactions and interdependence between oestradiol
`and EGF in the endometrium by delineating their individual
`actions and interactive roles, using human endometrial stromal
`cells as a model. By using specific inhibitors of the action of
`the hypothesis that oestrogen causes
`oestradiol and EGF,
`the human endometrium by
`growth and differentiation of
`inducing the synthesis or release of the polypeptide growth
`factor, EGF, was tested.
`In pure cultures of human endometrial stroma cells, 10 nmol
`J had a small stimulatory effect on DNA and
`oestradiol
`1 ~
`in these cells, EGF was potently
`protein synthesis. However,
`mitogenic. Western blotting demonstrated that oestradiol
`administration induced oestrogen receptor expression in the
`stromal cells, so it appears that
`the absence of a mitogenic
`response to oestradiol was not due to a deficiency in oestrogen
`receptors. Uchima et al (1991) were unable to demonstrate a
`proliferative response to oestradiol
`in uterine epithelial cells
`despite the presence of functional oestrogen receptors. Other
`authors have reported the apparent dichotomy that oestrogen
`can induce protein synthesis or progesterone receptor expres¬
`increase in cellular proliferation
`sion, without a concomitant
`(Casimiri el al, 1980; Aronica and Katzenellenbogen, 1991;
`Uchima et al, 1991).
`It was not possible to demonstrate that
`the activity of
`in endometrial stromal cells was inhibited by an
`oestradiol
`antibody directed against the EGF-R. Under the same experi¬
`mental conditions in which the mitogenic activity of EGF was
`completely abolished, no inhibition of oestradiol action by the
`anti-EGF-R antibody was observed. Similarly,
`there was no
`induced synthesis or
`evidence that
`the action of oestradiol
`release of EGF, since EGF was not detected in conditioned
`media from stromal cells that had been treated with 10 nmol
`If oestradiol acted through a cascade mech¬
`oestradiol 1~ .
`anism, then it might be expected that a small amount of EGF
`would have a large effect on endometrial growth. However,
`any EGF in the oestradiol-conditioned medium would have
`been present at concentrations of less than 0.2 pg ml
`, the
`the kit used. Since the conditioned
`limit of sensitivity of
`effect on [3H]thymidine incorporation
`medium had no
`when applied to fresh stromal cell cultures, either EGF can
`cause proliferation only at concentrations of greater
`than
`or no suitable stimulatory growth factors were
`0.2 pg ml ~
`these experimental
`induced by oestradiol
`treatment under
`conditions.
`The experiments performed with ICI 182,780 alone confirm
`the classification of ICI 182,780 as a pure steroidal antioestro-
`gen (Wakeling et al, 1989, 1991; Dukes et al, 1992, 1993;
`Wade et al, 1993).
`ICI 182,780 alone in stromal celi culture
`did not demonstrate any oestrogenic effect and inhibited
`
`J
`
`~
`
`MYLAN PHARMS. INC. EXHIBIT 1015 PAGE 5
`
`

`
`Table 3. Effect of co-incubating endometrial stromal cells with 10 nmol oestradiol
`ICR16
`
`I
`
`l and the antibody
`
`Addition
`
`10 nmol oestradiol 1
`~ 7l0 µg ICR16 ml
`10 nmol oestradiol 1
`Vehicle/10 µg ICR16 ml
`7l0 µg control antibody ml
`10 nmol oestradiol 1
`Vehicle/10 µg control antibody ml
`
`1
`
`~
`
`'
`
`"
`
`3Hjthymidine: % of
`control (sem)
`
`3H]leucine: % of
`control (sem)
`
`121.3 (17.3)
`(9.2)
`115
`76.9 (13.5)
`(14.4)
`102
`80.3 (15.3)
`
`(6.2)*
`132
`125.8 (4.0)*
`(4.5)
`107
`(6.4)*
`130
`(7.5)
`113
`
`"
`
`'
`
`"
`
`
`
`~
`
`Values are expressed as percentage of the control value (control = 100%) and are means of three experiments, each performed in
`quadruplicate. *P < 0.02.
`
`(a)
`
`10
`(b)
`
`-
`
`-
`
`-
`
`120
`
`110
`
`100
`
`90
`
`80
`
`70
`
`100
`
`80
`
`60
`
`40
`
`20
`
`10"
`
`10
`
`10"
`
`o
`
`. C
`
`3
`LU
`
` >
`
`
`>«
`
`2
`
` .
`
`VA-i-1-1-1
`10~9
`10"7
` ^8
`10"6
`Log10 concentration of ICI 182,780 (mol 1)
`Fig. 3. Effect of co-incubation of 1.6 nmol EGF 1
`: with various
`concentrations of ICI 182,780 on (a) [3H]leucine and (b) [3H]thymidine
`incorporation in endometrial stromal cultures. The data are expressed
`as means ( ± sem) of percentages of the EGF response alone (100%).
`*P<0.05.
`
`"
`
`ICI 164,384 also blocks EGF-induced uterine growth
`1993).
`and diminishes EGF-induced phosphatidylinositol
`lipid turn¬
`over in ovariectomized mice (Ignar-Trowbridge et al, 1992).
`The absence of inhibition of EGF-induced protein synthesis by
`ICI 182,780 in human endometrial stroma cells is interesting in
`view of the failure of ICI 164,384 to inhibit phosphatidyl¬
`lipid turnover completely (Ignar-Trowbridge et al,
`inositol
`1992).
`The antiproliferative effect of antioestrogens on growth
`factors appears to be mediated via the oestrogen receptor. In
`
`Table 4. Effect of conditioned media (CM) from various sources
`on [3H]thymidine incorporation in endometrial stromal cells
`
`Source of conditioned medium
`
`[3H]thymidine incorporation:
`% of control (sem)
`
`~
`
`'
`
`CM from cells treated with
`10 nmol oestradiol I
`CM from cells treated with
`10 nmol oestradiol 1
`~ 7
`100 nmol ICI 182,789 l"1
`CM from cells treated with
`100 nmol ICI 182,780 '
`
`102.7 (14.9)
`
`79.9 (3.0)
`
`97.5
`
`(9.5)
`
`Values are means ( ± sem) of
`three experiments. Data are expressed as
`percentages of the control value (control = 100%). No value was significantly
`different from the control.
`
`[3H]leucine incorpor¬
`oestradiol-stimulated [3H]thymidine or
`ation, nor did it induce expression of the oestrogen receptor.
`Potent anti-oestrogenic activity of
`ICI 182,780 in vivo on
`endometrial proliferation in premenopausal women was
`reported by Thomas el al (1994).
`When 1.6 nmol EGF l"1 was incubated with various con¬
`centrations of ICI 182,780,
`it was found that
`ICI 182,780
`profoundly inhibited EGF-induced [3H]thymidine but not
`[3H]leucine incorporation. This compares with the EGF-
`anti-EGF-R antibody experiments,
`in which the anti-EGF-R
`antibody did not markedly inhibit protein synthesis. Such inter¬
`actions between antioestrogens and growth factors have been
`noted previously. Proliferation of breast cancer MCF-7 cells
`induced by insulin or EGF is inhibited by 4-hydroxytamoxifen
`(Vignon et al, 1987).
`ICI 164,384 inhibits the proliferative
`action of both IGF-I and transforming growth factor in
`MCF-7 cells, and is more effective at blocking MCF-7 growth
`than 4-hydroxytamoxifen (Wakeling et al, 1989). The EGF-
`and IGF-I-induction of pS2 and cathepsin D mRNAs in
`MCF-7 cells is inhibited by ICI 164,384 (Chalbos et al,
`1993). Progesterone receptors are induced by EGF, but not
`by other growth factors, and this effect
`is inhibited by
`4-hydroxytamoxifen (Sumida
`and Pasqualini,
`1989).
`ICI
`164,384 inhibits EGF- and oestradiol-stimulated transcriptional
`activation in oestrogen receptor-transfected Ishikawa human
`endometrial adenocarcinoma cells (Ignar-Trowbridge et al,
`
`MYLAN PHARMS. INC. EXHIBIT 1015 PAGE 6
`
`

`
`MCF-7 cells, there is a strong correlation between the growth
`factor-induced antiproliferative activity and the affinity of
`several non-steroidal antioestrogens for the oestrogen receptor
`(Vignon et al, 1987).
`In this model system, oestradiol can
`counter the anti-insulin effect of 4-hydroxytamoxifen. Further¬
`line that is oestrogen receptor-
`in a breast cancer cell
`more,
`negative, EGF- and insulin-induced growth is not inhibited by
`4-hydroxytamoxifen (Vignon et al, 1987). Inhibition of IGF-I
`induced pS2 and cathepsin D mRNA expression is oestrogen
`receptor specific (Chalbos et al, 1993). In the ovariectomized
`mouse model, EGF can mimic the action of oestrogen by
`causing enhanced nuclear localization of the oestrogen recep¬
`tor, and the nuclear oestrogen receptor from EGF-treated mice
`forms similar complexes with the oestrogen response element
`from diethylstilboestrol-treated mice
`those
`(Ignar-
`to
`Trowbridge et al, 1992). Induction of transcriptional activation
`of the oestrogen response element by EGF and inhibition of
`this phenomenon by ICI 164,384 are dependent on both the
`oestrogen receptor and the EGF-R (Ignar-Trowbridge et al,
`1993).
`If these observations are extended to the present study, then
`the inhibition by ICI 182,780 of EGF-induced proliferation of
`endometrial stromal cells may also be mediated via the
`oestrogen receptor. How this is achieved is not clear. The
`oestrogen receptor can bind to DNA in the absence of ligand
`(Reese and Katzenellenbogen, 1992) and also when ICI 164,384
`is bound to the oestrogen receptor (Gibson et al, 1991; Reese
`and Katzenellenbogen, 1992). Binding of the oestrogen recep¬
`tor to the DNA is insufficient
`to cause gene transcription
`(Reese and Katzenellenbogen, 1992), so it may be that an
`endpoint of EGF-R signal transduction is an alteration in the
`oestrogen receptor such that
`it can bind to the oestrogen
`response element and lead to gene transcription. EGF is known
`to increase the concentration of nuclear oestrogen receptor and
`to promote the formation of a nuclear oestrogen receptor
`species thought previously to be induced by oestradiol (Ignar-
`Trowbridge et al, 1992), and it may be that these two events
`are either a consequence of an alteration of
`the oestrogen
`receptor or occur before binding of the oestrogen receptor to
`the oestrogen response element.
`Targets of gene transcription as a result of oestradiol and
`EGF action may include the cellular oncogenes such as c-fos,
`c-jun and c-myc. EGF induces c-jun, c-fos and jun-B mRNAS in
`MCF-7 breast cancer cells (Davidson et al, 1993). In rat uteris,
`c-myc, c-fos, c-jun, jun and jun D expression are all
`increased
`by oestradiol
`treatment (Murphy et al, 1987; Stancel et al,
`1991; Chiappetta et al, 1992).
`In uterine epithelial cells,
`oestradiol, EGF and insulin stimulates c-fos expression in
`combination, but not separately (Jouvenot et al, 1990). Poten¬
`tial oestrogen response elements have been identified in the
`c-/os oncogene (Hyder el al, 1991) and c-myc (Dubik and Shiu,
`1992). Since Fos-Jun and Jun-Jun dimers act as transcriptional
`regulators,
`those proto-oncogenes with upstream oestrogen
`responsive elements may be inhibited by ICI
`182,780,
`where the ICI 182,780 is inhibiting oestradiol or EGF action.
`Specificity of oestradiol or EGF or other growth factor action
`may result from differential stimulation of cellular oncogenes
`by these factors, such that the response to individual stimuli
`may depend on whether
`the target genes activated are
`stimulated to transcribe by Fos-Jun or Jun-Jun. Additionally,
`
`Fos and Jun may act
`in concert with the oestrogen receptor
`to regulate transcription of hormone-sensitive genes (Gaub
`et al, 1990).
`transduction pathways of EGF action
`The various signal
`may explain why no inhibition of protein synthesis by the
`anti-EGF-R antibody,
`ICR16, was observed.
`It may be that
`the induction of cellular proliferation and cellular protein
`synthesis are caused by two different signal
`transduction
`mechanisms and that, although the ICR16 antibody blocked
`the transduction pathway leading to cellular proliferation, the
`antibody may not be able to block transduction by the
`the pathway resulting in increased protein syn¬
`EGF-R of
`thesis. This may also explain why ICI 182,780 failed to
`inhibit EGF-induced protein synthesis. The pathway resulting
`in increased protein synthesis by EGF may not
`involve the
`oestrogen receptor and the oestrogen response element, but
`occur via another route. This would explain the observation
`ICI 164,384 fails to completely block phosphatidyl-
`that
`induced by EGF (Ignar-Trowbridge
`lipid turnover
`inositol
`et al, 1992).
`Although release of EGF could not be demonstrated,
`it is
`possible that oestradiol may increase expression of the EGF-R
`(Bonaccorsi et al, 1989; Stancel et al, 1990; Taketani and
`Mizuno, 1991; Troche et al, 19