The Prostate 20:151-158 (1992)
Interactions Between Epidermal Growth Factor-Mediated Autocrine Regulation and Linoleic Acid-Stimulated Growth of a Human Prostate Cancer Cell Line Jeanne M. Connolly and David P. Rose Division of Nutrition and Endocrinology, Naylor Dana Institute for Disease Prevention, American Health Foundation, Valhalla, New York Human prostate cancer (PC) cell lines possess epidermal growth factor (EGF) receptors and secrete EGF-related polypeptides. We used an EGF receptor-blocking antibody (antiEGF.R) to demonstrate a functional autocrine loop, as well as the interaction between this and the effects of linoleic acid (LA), an omega-6 fatty acid, on PC cell growth. The anti-EGF.R competed effectively with ['=I]EGF for receptors on DU145 PC cells, and on a high-passage DU145 variant (DU145M);when added to the culture medium, it suppressed both DU145 and DU145M cell growth in a dose-dependent manner. LA, a precursor for eicosanoid synthesis, had little effect on DU145 cell growth rate but stimulated DU145M growth in a concentration-related manner over a range of 0.25-2.0 pg/ml. Anti-EGF.R M) caused suppression of LA-stimulated growth of DU145M cells in serum-free medium, which was prevented by the addition of 2 nM EGF. We conclude that an EGF.Rmediated autocrine loop is involved in PC cell growth regulation and that at least one site of action may be the synthesis of eicosanoids from their LA precursor.
Key words: DU145 cells, omega4 fatty acids, growth factors, eicospnoids
INTRODUCTION A comparison of age-adjusted prostate cancer (PC) mortality rates with the corresponding frequency of prostatic carcinoma in situ in different countries leads to the hypothesis that environmental factors, perhaps dietary in nature, may increase the risk of developing clinically manifest disease [l-41. On an international basis, a positive correlation exists between the percentage calories consumed as fat and PC mortality rate [5,6], while the consumption of the omega-6 polyunsaturated fatty acid (n-6 FA) linoleic acid (LA) has been epidemiologically linked to the rise in prostate cancer mortality rates in Western countries [7]. In support of these ecological data, we have shown that LA stimulates the growth of two PC cell lines in culture: the androgen-independent PC-3 cell [8], and the androgen-responsive LNCaP cell line (unpublished data). In other studies, we demonstrated that another human PC cell line, DU145,
Received for publication June 10, 1991; accepted November 13, 1991. Address reprint requests to Dr. D.P. Rose, Division of Nutrition and Endocrinology, Naylor Dana Institute for Disease Prevention, Valhalla, NY 10595. 0 1992 Wdey-Liss, Inc.
152
COMOUYand Rose
possesses a functional epidermal growth factor (EGF) autocrine loop [9]. A variant of the DU145 cell line has emerged spontaneously after extensive passage in our laboratory, the growth of which is also stimulated by autocrine EGF secretion. However, unlike the parental cell line, its growth in serum-free medium is also enhanced in the presence of LA. We now report that the LA-stimulated growth in these PC cells requires an intact EGF autocrine loop, most likely because, as in some other cell types, EGF regulates the mobilization of arachidonic acid, an n-6 FA synthesized from LA, from cell membrane phospholipids, and hence production of eicosanoids which act as second messengers in the mitogenic process [lo-121. MATERIALS AND METHODS Cells and Cell Culture The DU145 cell line was obtained from the American Type Culture Collection (Rockville, MD) and maintained in RPMI-1640 medium (GIBCO, Grand Island, NY) with 5 % fetal bovine serum (FBS)(Hyclone, Logan, UT), and 100,OOO U/L penicillin and 100 mg/L streptomycin (Sigma, St. Louis, MO) in a 95% air/5% CO, incubator at 37°C. Passage numbers below 80 were used for all experiments with the original line. The DU145 variant (designated DU145M) emerged after prolonged passage (passage > 120) and has maintained LA responsiveness for more than 25 subsequent passages. All cells routinely tested negative for mycoplasma. Binding Experiments With Anti-EGF Receptor (EGF.R) Antibody
The cells were plated at 1.5 X lo4cells/ml/well in plastic 24-well culture plates (Corning Inc., Corning, NY) in 5% FBS-supplemented phenol red-free (PRF) RPMI1640. After 48 h of incubation at 37°C and at approximately 70% confluence, the cells were refed; 2 h before beginning a binding experiment, the medium was changed to PRF RPMI-1640 containing 0.1 % delipidized bovine serum albumin (DBSA) (Collaborative Research, Lexington, MA). At time zero, the medium was again changed to incorporate one of the competitors, human recombinant EGF (hrEGF; Biomedical Technologies, Inc., Stoughton,MA), or monoclonal antibody to human EGF.R (Anti-EGF.R; Upstate Biotechnology, Lake Placid, NY). ['251]EGF (specific activity 125pCi/p,g,Biomedical Technologies, Inc.) in 100 p1 of RPMI-1640 plus 0.1% DBSA was immediately added to each well at a final concentration of 5-8 x lo-" M, mixed, and incubation performed at 0°C (ice bath) for 2 h. All concentrations of each competing ligand were set up in triplicate. A range of concentrations of nonspecific mouse IgG 1 was included as a negative control, together with triplicate wells to provide for the total binding and total cell counts. Growth Experiments Cells were plated as for binding experiments. After 24 h of incubation to permit cell attachment, the medium was changed to PRF RPMI-1640 containing either 3% FBS, or 10 pg/d insulin (Sigma) plus 1.25 mg/ml DBSA, as designated for the individual experiments described below. Additions of Anti-EGF.R and/or LA (as complex with DBSA; Sigma) were also made at this time. The cultures were examined with an invert microscope daily to exclude the possibility of significant cell loss due to detachment, and incubated for 3 days, after which they were harvested by
Linoleic Acid, EGF, and Prostate Cancer Cells
153
100
P
.c
E
80
a
-
;
60
I-
L
-!
40
c
0
#
2o
0
ole-"
lo-'' 104 anti-EGF.R (MI
lo-'
Fig. 1. Competitionbetween [1251]EGFand EGF.R monoclonal antibody for binding to DU145 cells ( 0 ) and DU145M (0)cells. Binding is expressed as percentage of maximum [1251]EGFbinding. The reduction in percentage of [lZSI]EGFbinding was statistically significant with antibody concentrations of greater than 1 X lo-'' M.
trypsinization and counted with an electronic particle counter (model F, Coulter Electronics, Hialeah, FL). Statistical Comparisons These were carried out by Student's unpaired t-test; values of P regarded as significant.
< 0.05 were
RESULTS Competitive Binding Experiments
These showed that anti-EGF.R was an effective competitive blocker of specific ['251]EGFbinding to both the original DU145 and the DU145M prostate cancer cells (Fig. 1). Maximal inhibition of binding plateaued in each case at an antibody concentration of 5 X M, with no significant difference in ['241]EGF binding displacement between the cell types. Nonspecific IgGl had no effect on the binding of labeled EGF to either of the cell types, while lo-' M EGF effectively prevented binding (data not shown). Growth Experiments-Anti-EGF.R
Titration
In the first series of growth experiments, 1.5 X lo4 cells of each type were plated, allowed to attach in 5% FBS-RPMI-1640 for 24 h, and then cultured in RPMI-1640 supplemented with 3% FBS for 3 days with the addition to triplicate M, and wells of anti-EGF.R at concentrations ranging from 1 x lo-" to 5 x with an additional triplicate set of positive growth wells containing no antibody. In a 'parallel set of experiments, DU145 or DU145M cells were plated in the same manner, but cultured in serum-free (SF)insulin and DBSA-containing culture medium with or without the antibody.
154
COMOUYand Rose TABLE I. Growth of DU145 and DU145M Cells in Medium Containing Fetal B o v k Serum (3% FBS), or Serum-Free (SF) Medium With or Without Linoleic Acid (0.75 pg/d LA) Celldwell Treatment 3% FBS
SF medium SF + 0.75 udml LA
DU145 cells 5.3 6.7
* 0.2 * 0.3
6.5 2 0.9
X
lo4 (mean
2
SD)
DU145M cells 12.7 5.7 7.5
k
0.6*
k
0.6**
* 0.6
*DU145M cell number significantly greater than DU145 cell number after 3 days culture, P
Interactions Between Epidermal Growth Factor-Mediated Autocrine Regulation and Linoleic Acid-Stimulated Growth of a Human Prostate Cancer Cell Line Jeanne M. Connolly and David P. Rose Division of Nutrition and Endocrinology, Naylor Dana Institute for Disease Prevention, American Health Foundation, Valhalla, New York Human prostate cancer (PC) cell lines possess epidermal growth factor (EGF) receptors and secrete EGF-related polypeptides. We used an EGF receptor-blocking antibody (antiEGF.R) to demonstrate a functional autocrine loop, as well as the interaction between this and the effects of linoleic acid (LA), an omega-6 fatty acid, on PC cell growth. The anti-EGF.R competed effectively with ['=I]EGF for receptors on DU145 PC cells, and on a high-passage DU145 variant (DU145M);when added to the culture medium, it suppressed both DU145 and DU145M cell growth in a dose-dependent manner. LA, a precursor for eicosanoid synthesis, had little effect on DU145 cell growth rate but stimulated DU145M growth in a concentration-related manner over a range of 0.25-2.0 pg/ml. Anti-EGF.R M) caused suppression of LA-stimulated growth of DU145M cells in serum-free medium, which was prevented by the addition of 2 nM EGF. We conclude that an EGF.Rmediated autocrine loop is involved in PC cell growth regulation and that at least one site of action may be the synthesis of eicosanoids from their LA precursor.
Key words: DU145 cells, omega4 fatty acids, growth factors, eicospnoids
INTRODUCTION A comparison of age-adjusted prostate cancer (PC) mortality rates with the corresponding frequency of prostatic carcinoma in situ in different countries leads to the hypothesis that environmental factors, perhaps dietary in nature, may increase the risk of developing clinically manifest disease [l-41. On an international basis, a positive correlation exists between the percentage calories consumed as fat and PC mortality rate [5,6], while the consumption of the omega-6 polyunsaturated fatty acid (n-6 FA) linoleic acid (LA) has been epidemiologically linked to the rise in prostate cancer mortality rates in Western countries [7]. In support of these ecological data, we have shown that LA stimulates the growth of two PC cell lines in culture: the androgen-independent PC-3 cell [8], and the androgen-responsive LNCaP cell line (unpublished data). In other studies, we demonstrated that another human PC cell line, DU145,
Received for publication June 10, 1991; accepted November 13, 1991. Address reprint requests to Dr. D.P. Rose, Division of Nutrition and Endocrinology, Naylor Dana Institute for Disease Prevention, Valhalla, NY 10595. 0 1992 Wdey-Liss, Inc.
152
COMOUYand Rose
possesses a functional epidermal growth factor (EGF) autocrine loop [9]. A variant of the DU145 cell line has emerged spontaneously after extensive passage in our laboratory, the growth of which is also stimulated by autocrine EGF secretion. However, unlike the parental cell line, its growth in serum-free medium is also enhanced in the presence of LA. We now report that the LA-stimulated growth in these PC cells requires an intact EGF autocrine loop, most likely because, as in some other cell types, EGF regulates the mobilization of arachidonic acid, an n-6 FA synthesized from LA, from cell membrane phospholipids, and hence production of eicosanoids which act as second messengers in the mitogenic process [lo-121. MATERIALS AND METHODS Cells and Cell Culture The DU145 cell line was obtained from the American Type Culture Collection (Rockville, MD) and maintained in RPMI-1640 medium (GIBCO, Grand Island, NY) with 5 % fetal bovine serum (FBS)(Hyclone, Logan, UT), and 100,OOO U/L penicillin and 100 mg/L streptomycin (Sigma, St. Louis, MO) in a 95% air/5% CO, incubator at 37°C. Passage numbers below 80 were used for all experiments with the original line. The DU145 variant (designated DU145M) emerged after prolonged passage (passage > 120) and has maintained LA responsiveness for more than 25 subsequent passages. All cells routinely tested negative for mycoplasma. Binding Experiments With Anti-EGF Receptor (EGF.R) Antibody
The cells were plated at 1.5 X lo4cells/ml/well in plastic 24-well culture plates (Corning Inc., Corning, NY) in 5% FBS-supplemented phenol red-free (PRF) RPMI1640. After 48 h of incubation at 37°C and at approximately 70% confluence, the cells were refed; 2 h before beginning a binding experiment, the medium was changed to PRF RPMI-1640 containing 0.1 % delipidized bovine serum albumin (DBSA) (Collaborative Research, Lexington, MA). At time zero, the medium was again changed to incorporate one of the competitors, human recombinant EGF (hrEGF; Biomedical Technologies, Inc., Stoughton,MA), or monoclonal antibody to human EGF.R (Anti-EGF.R; Upstate Biotechnology, Lake Placid, NY). ['251]EGF (specific activity 125pCi/p,g,Biomedical Technologies, Inc.) in 100 p1 of RPMI-1640 plus 0.1% DBSA was immediately added to each well at a final concentration of 5-8 x lo-" M, mixed, and incubation performed at 0°C (ice bath) for 2 h. All concentrations of each competing ligand were set up in triplicate. A range of concentrations of nonspecific mouse IgG 1 was included as a negative control, together with triplicate wells to provide for the total binding and total cell counts. Growth Experiments Cells were plated as for binding experiments. After 24 h of incubation to permit cell attachment, the medium was changed to PRF RPMI-1640 containing either 3% FBS, or 10 pg/d insulin (Sigma) plus 1.25 mg/ml DBSA, as designated for the individual experiments described below. Additions of Anti-EGF.R and/or LA (as complex with DBSA; Sigma) were also made at this time. The cultures were examined with an invert microscope daily to exclude the possibility of significant cell loss due to detachment, and incubated for 3 days, after which they were harvested by
Linoleic Acid, EGF, and Prostate Cancer Cells
153
100
P
.c
E
80
a
-
;
60
I-
L
-!
40
c
0
#
2o
0
ole-"
lo-'' 104 anti-EGF.R (MI
lo-'
Fig. 1. Competitionbetween [1251]EGFand EGF.R monoclonal antibody for binding to DU145 cells ( 0 ) and DU145M (0)cells. Binding is expressed as percentage of maximum [1251]EGFbinding. The reduction in percentage of [lZSI]EGFbinding was statistically significant with antibody concentrations of greater than 1 X lo-'' M.
trypsinization and counted with an electronic particle counter (model F, Coulter Electronics, Hialeah, FL). Statistical Comparisons These were carried out by Student's unpaired t-test; values of P regarded as significant.
< 0.05 were
RESULTS Competitive Binding Experiments
These showed that anti-EGF.R was an effective competitive blocker of specific ['251]EGFbinding to both the original DU145 and the DU145M prostate cancer cells (Fig. 1). Maximal inhibition of binding plateaued in each case at an antibody concentration of 5 X M, with no significant difference in ['241]EGF binding displacement between the cell types. Nonspecific IgGl had no effect on the binding of labeled EGF to either of the cell types, while lo-' M EGF effectively prevented binding (data not shown). Growth Experiments-Anti-EGF.R
Titration
In the first series of growth experiments, 1.5 X lo4 cells of each type were plated, allowed to attach in 5% FBS-RPMI-1640 for 24 h, and then cultured in RPMI-1640 supplemented with 3% FBS for 3 days with the addition to triplicate M, and wells of anti-EGF.R at concentrations ranging from 1 x lo-" to 5 x with an additional triplicate set of positive growth wells containing no antibody. In a 'parallel set of experiments, DU145 or DU145M cells were plated in the same manner, but cultured in serum-free (SF)insulin and DBSA-containing culture medium with or without the antibody.
154
COMOUYand Rose TABLE I. Growth of DU145 and DU145M Cells in Medium Containing Fetal B o v k Serum (3% FBS), or Serum-Free (SF) Medium With or Without Linoleic Acid (0.75 pg/d LA) Celldwell Treatment 3% FBS
SF medium SF + 0.75 udml LA
DU145 cells 5.3 6.7
* 0.2 * 0.3
6.5 2 0.9
X
lo4 (mean
2
SD)
DU145M cells 12.7 5.7 7.5
k
0.6*
k
0.6**
* 0.6
*DU145M cell number significantly greater than DU145 cell number after 3 days culture, P
