JOURNAL OF CELLULAR PHYSIOLOGY 144:lS-25 (1990)
Altered Cell Cycle Responses to Insulin-Like Growth Factor I, but Not Platelet-Derived Growth Factor and Epidermal Growth Factor, in Senescing Human Fibroblasts Y U H C H Y A U CHEN AND PETER S. RABINOVITCH* Department of Pathology, SM-30, University of Washington, Seattle, Washington 98 195
Human diploid fibroblasts (HDF) were used to study aging-related changes in the pro1iferative response to platelet-derived growth factor (PDGF), epidermal growth factor ( E L F ) , and insulin-like growth factor I (IGF-I, somatomedin-C) in serumfree, chemically defined culture medium. Cell cycle kinetic parameters were determined by using 5-bromodeoxyuridine incorporation and flow cytometric analysis with the D N A stain Hoechst 33258. This allowed analysis of the growth factor response to be focussed exclusively upon of the cycling faction of cells within the culture, even in senescent cell cultures which contained predominantly nondividing cells. PDGF and EGF exert their primary effect upon regulation of the proportion of cycling cells in the culture. The doses of PDGF and EGF that produced a half-maximal cycling fraction, analogous to K,,,, showed no large or consistent difference between young- and old-passage cells. In contrast, IGF-I primarily affects the rate of transition of cells from G, into S phase, and the dose of IGF-I which produced a half-maximal rate of G, exit increased up to 130-fold in older-passage cells. Unexpectedly, supraphysiologic concentrations of IGF-I were found to increase the G , exit rate of the dividing subpopulation of cells in older-passage cultures to rates higher than those seen in young cultures. In summary, among cells capable of cycling in aging cultures, there were few changes in the regulation of the growth fraction by PDGF and EGF, but there was a greatly increased dependence on IGF-I for regulation of the rate of entry into S phase. The slower growth of the dividing population of cells in aging cultures may be related to a reauirement for IGF-I at levels which are greatly above those usually supplied.
Human diploid fibroblasts (HDF) have a limited life span in culture (Hayflick and Moorhead, 1961; Hayflick, 1965). This replicative life span is inversely related to the age of the donor (Martin et al., 1970; Schneider and Mitsui, 1976) and directly related to the maximum life span of the species (Rohme, 1981). HDF from subjects with certain progeric syndromes (premature aging) also have shortened life span and diminished growth potential in vitro (Goldstein, 1969; Martin et al., 1970; Salk et al., 1981). Cellular senescence has therefore been used as a model system to study age-related changes in vitro. With serial passages in culture, HDF become progressively less responsive t o mitogenic stimuli, showing a gradual increase in the fraction of cells arrested in a nondividing state (Cristofalo and Sharf, 1973; Matsumura et al., 1979) and increased length and heterogeneity of the GI phase of the cell cycle of the dividing subpopulation in cellular senescence (Grove and Cristofalo, 1976). An assay of cell kinetics using 5-bromodeoxyuridine (BrdU), the DNA-specific dye Hoechst 33258, and flow cytometry is especially suited to an investigation of the cell cycle kinetics of the dividing cells within a cell culture that also contains a nondividing population (Rabinovitch et 1990 WILEY-LIB, INC.
al., 1988). Analysis of aging HDF cell kinetics using BrdU-Hoechst flow cytometry also demonstrated an increased fraction of noncycling cells, and a slower rate of G, to S-phase transition among the cycling cells in senescent serum-treated cultures (Rabinovitch, 1983). We have recently described roles of platelet-derived growth factor (PDGF), epidermal growth factor (EGF), and insulin-like growth factor I (IGF-I) in regulating distinct cell-cycle parameters of HDF in serum-free, chemically defined medium (Chen and Rabinovitch, 1989). The BrdU-Hoechst flow cytometric analysis shows that the predominant effect of PDGF and EGF is upon the fraction of cycling cells (growth fraction). In contrast, IGF-I regulates the rate of exit from GI into S phase without an apparent effect upon the cycling fraction. In the present studies, we hoped that such an analysis of the response of senescing HDF to defined growth factors might yield new insight into the mechanisms responsible for the loss of proliferative potential in senescence. Received June 19, 1989; accepted February 26, 1990.
*To whom reprint requests/correspondence should be addressed.
MlTOGEN RESPONSE IN SE:NESCING FlBROBLASTS
MATERIALS AND METHODS Cell strains and culture The human diploid fibroblast cell strain designated 79-81 was explanted from a skin biopsy of a 27-year-old male and had a maximum in vitro life span of 65 cumulative populations doublings. Strain 78-89 was derived from a skin sample of a 36-week gestational age female abortus and had a maximum life span of 61 cumulative population doublings. Both strains were karyotypically normal. Adult and fetal strains were chosen in order to examine two cell strains of different origin. Cells were routinely cultured in Earle’s minimum essential medium supplemented with 26 mM sodium bicarbonate and 10% heat-inactivated fetal bovine serum (FBS) at 37”C, 5% C02. In one experiment, cells were gamma irradiated (5,000 rads) to inhibit subsequent replication. Tests for mycoplasma were negative by staining with 4,6-diamidino-2-phenylindole (Russell et al., 1975).
19
quently; thus, 3-d serum deprivation was used for these cells. After serum deprivation, the experimental protocol was initiated, washing each dish twice with DMEM/ F-12 without serum, and then refeeding each dish with serum-free experimental medium containing BrdU, deoxycytidine, and purified growth factors in the indicated combinations and concentrations. The experimental media were replaced every second d thereafter to avoid depletion of growth factors. These cultures were exposed only to 580-590 nm light (model DUB safelight, Thomas Instrument, Charlottesville, VA) so as to avoid BrdU-mediated phototoxicity. At the time points described, triplicate dishes were harvested for flow cytometry.
Flow cytometry and cell cycle kinetic analysis Analysis by BrdU-Hoechst flow cytometry was performed as previously described (Rabinovitch et al., 1988). In brief, cells were stained with Hoechst 33258 (2.9 pg/ml) and ethidium bromide (1 pgiml) and analyzed by using a n ICP-22 flow cytometer (Ortho DiagGrowth factors, chemicals, and medium nostic Systems, Westwood, MA) interfaced to a PDP Mouse EGF (tissue culture grade) was purchased 11/03 computer (Digital Equipment, Maynard, MA). from Collaborative Research (Waltham, MA). Purified The extent of BrdU substitution into DNA was meaPDGF was a kind gift from Dr. Russell Ross (Univer- sured by the degree of quenching of “blue” Hoechst sity of Washington). IGF-I (recombinant) was pur- 33258 fluorescence, while the “red” ethidium bromide chased from Amgen Biologicals (Thousand Oaks, CA). fluorescence yielded cell cycle distributions unaffected Porcine insulin, human transferrin, dexamethasone, by BrdU incorporation. This allowed quantitation of bovine serum albumin (BSA), BrdU, and deoxycytidine the cell numbers in G,, S, and G2/M compartments of were purchased from Sigma Chemical Co. (St. Louis, three successive cell cycles. From this information the MO). fraction of initially plated cells that had never left Go/ Dulbecco’s modified Eagle’s medium (DMEM) and G, (non-BrdU-labeled cells) was calculated (RabinoHam’s F-12 Nutrient Mixture, Hepes buffer, sodium vitch et al., 1988). bicarbonate, and FBS were purchased from Grand The cell kinetic data was plotted a s log percent G,/G, Island Biological Company. Trace elements (MnSO,. cells vs. time. A model was used which is able to ex5H,O, (NH,),Mo,O,,, NiCl,.GH,O, H,SeO, Na,SiO,. plain the heterogeneity in the lengths of the GJG, 9H,O, SnC1,.2H,O, and NH,V03) were purchased from phase; the model is based on findings of a n exponential Johnson Matthey Chemicals Ltd., (London, England). rate of exit of cells from G,/G, (Smith and Martin, Hoechst 33258 and ethidium bromide were purchased 1973), as modified to include a noncycling fraction of from Calbiochem (San Diego, CAI. cells (Rabinovitch, 1983): 76 Cells in Initial GJG, = ~lOO-n)lO-r’t-Tb’ + n (for t >‘ Tb) Defined medium and serum-free culture (’% G,,/G, Cells = 100 f o r t
Altered Cell Cycle Responses to Insulin-Like Growth Factor I, but Not Platelet-Derived Growth Factor and Epidermal Growth Factor, in Senescing Human Fibroblasts Y U H C H Y A U CHEN AND PETER S. RABINOVITCH* Department of Pathology, SM-30, University of Washington, Seattle, Washington 98 195
Human diploid fibroblasts (HDF) were used to study aging-related changes in the pro1iferative response to platelet-derived growth factor (PDGF), epidermal growth factor ( E L F ) , and insulin-like growth factor I (IGF-I, somatomedin-C) in serumfree, chemically defined culture medium. Cell cycle kinetic parameters were determined by using 5-bromodeoxyuridine incorporation and flow cytometric analysis with the D N A stain Hoechst 33258. This allowed analysis of the growth factor response to be focussed exclusively upon of the cycling faction of cells within the culture, even in senescent cell cultures which contained predominantly nondividing cells. PDGF and EGF exert their primary effect upon regulation of the proportion of cycling cells in the culture. The doses of PDGF and EGF that produced a half-maximal cycling fraction, analogous to K,,,, showed no large or consistent difference between young- and old-passage cells. In contrast, IGF-I primarily affects the rate of transition of cells from G, into S phase, and the dose of IGF-I which produced a half-maximal rate of G, exit increased up to 130-fold in older-passage cells. Unexpectedly, supraphysiologic concentrations of IGF-I were found to increase the G , exit rate of the dividing subpopulation of cells in older-passage cultures to rates higher than those seen in young cultures. In summary, among cells capable of cycling in aging cultures, there were few changes in the regulation of the growth fraction by PDGF and EGF, but there was a greatly increased dependence on IGF-I for regulation of the rate of entry into S phase. The slower growth of the dividing population of cells in aging cultures may be related to a reauirement for IGF-I at levels which are greatly above those usually supplied.
Human diploid fibroblasts (HDF) have a limited life span in culture (Hayflick and Moorhead, 1961; Hayflick, 1965). This replicative life span is inversely related to the age of the donor (Martin et al., 1970; Schneider and Mitsui, 1976) and directly related to the maximum life span of the species (Rohme, 1981). HDF from subjects with certain progeric syndromes (premature aging) also have shortened life span and diminished growth potential in vitro (Goldstein, 1969; Martin et al., 1970; Salk et al., 1981). Cellular senescence has therefore been used as a model system to study age-related changes in vitro. With serial passages in culture, HDF become progressively less responsive t o mitogenic stimuli, showing a gradual increase in the fraction of cells arrested in a nondividing state (Cristofalo and Sharf, 1973; Matsumura et al., 1979) and increased length and heterogeneity of the GI phase of the cell cycle of the dividing subpopulation in cellular senescence (Grove and Cristofalo, 1976). An assay of cell kinetics using 5-bromodeoxyuridine (BrdU), the DNA-specific dye Hoechst 33258, and flow cytometry is especially suited to an investigation of the cell cycle kinetics of the dividing cells within a cell culture that also contains a nondividing population (Rabinovitch et 1990 WILEY-LIB, INC.
al., 1988). Analysis of aging HDF cell kinetics using BrdU-Hoechst flow cytometry also demonstrated an increased fraction of noncycling cells, and a slower rate of G, to S-phase transition among the cycling cells in senescent serum-treated cultures (Rabinovitch, 1983). We have recently described roles of platelet-derived growth factor (PDGF), epidermal growth factor (EGF), and insulin-like growth factor I (IGF-I) in regulating distinct cell-cycle parameters of HDF in serum-free, chemically defined medium (Chen and Rabinovitch, 1989). The BrdU-Hoechst flow cytometric analysis shows that the predominant effect of PDGF and EGF is upon the fraction of cycling cells (growth fraction). In contrast, IGF-I regulates the rate of exit from GI into S phase without an apparent effect upon the cycling fraction. In the present studies, we hoped that such an analysis of the response of senescing HDF to defined growth factors might yield new insight into the mechanisms responsible for the loss of proliferative potential in senescence. Received June 19, 1989; accepted February 26, 1990.
*To whom reprint requests/correspondence should be addressed.
MlTOGEN RESPONSE IN SE:NESCING FlBROBLASTS
MATERIALS AND METHODS Cell strains and culture The human diploid fibroblast cell strain designated 79-81 was explanted from a skin biopsy of a 27-year-old male and had a maximum in vitro life span of 65 cumulative populations doublings. Strain 78-89 was derived from a skin sample of a 36-week gestational age female abortus and had a maximum life span of 61 cumulative population doublings. Both strains were karyotypically normal. Adult and fetal strains were chosen in order to examine two cell strains of different origin. Cells were routinely cultured in Earle’s minimum essential medium supplemented with 26 mM sodium bicarbonate and 10% heat-inactivated fetal bovine serum (FBS) at 37”C, 5% C02. In one experiment, cells were gamma irradiated (5,000 rads) to inhibit subsequent replication. Tests for mycoplasma were negative by staining with 4,6-diamidino-2-phenylindole (Russell et al., 1975).
19
quently; thus, 3-d serum deprivation was used for these cells. After serum deprivation, the experimental protocol was initiated, washing each dish twice with DMEM/ F-12 without serum, and then refeeding each dish with serum-free experimental medium containing BrdU, deoxycytidine, and purified growth factors in the indicated combinations and concentrations. The experimental media were replaced every second d thereafter to avoid depletion of growth factors. These cultures were exposed only to 580-590 nm light (model DUB safelight, Thomas Instrument, Charlottesville, VA) so as to avoid BrdU-mediated phototoxicity. At the time points described, triplicate dishes were harvested for flow cytometry.
Flow cytometry and cell cycle kinetic analysis Analysis by BrdU-Hoechst flow cytometry was performed as previously described (Rabinovitch et al., 1988). In brief, cells were stained with Hoechst 33258 (2.9 pg/ml) and ethidium bromide (1 pgiml) and analyzed by using a n ICP-22 flow cytometer (Ortho DiagGrowth factors, chemicals, and medium nostic Systems, Westwood, MA) interfaced to a PDP Mouse EGF (tissue culture grade) was purchased 11/03 computer (Digital Equipment, Maynard, MA). from Collaborative Research (Waltham, MA). Purified The extent of BrdU substitution into DNA was meaPDGF was a kind gift from Dr. Russell Ross (Univer- sured by the degree of quenching of “blue” Hoechst sity of Washington). IGF-I (recombinant) was pur- 33258 fluorescence, while the “red” ethidium bromide chased from Amgen Biologicals (Thousand Oaks, CA). fluorescence yielded cell cycle distributions unaffected Porcine insulin, human transferrin, dexamethasone, by BrdU incorporation. This allowed quantitation of bovine serum albumin (BSA), BrdU, and deoxycytidine the cell numbers in G,, S, and G2/M compartments of were purchased from Sigma Chemical Co. (St. Louis, three successive cell cycles. From this information the MO). fraction of initially plated cells that had never left Go/ Dulbecco’s modified Eagle’s medium (DMEM) and G, (non-BrdU-labeled cells) was calculated (RabinoHam’s F-12 Nutrient Mixture, Hepes buffer, sodium vitch et al., 1988). bicarbonate, and FBS were purchased from Grand The cell kinetic data was plotted a s log percent G,/G, Island Biological Company. Trace elements (MnSO,. cells vs. time. A model was used which is able to ex5H,O, (NH,),Mo,O,,, NiCl,.GH,O, H,SeO, Na,SiO,. plain the heterogeneity in the lengths of the GJG, 9H,O, SnC1,.2H,O, and NH,V03) were purchased from phase; the model is based on findings of a n exponential Johnson Matthey Chemicals Ltd., (London, England). rate of exit of cells from G,/G, (Smith and Martin, Hoechst 33258 and ethidium bromide were purchased 1973), as modified to include a noncycling fraction of from Calbiochem (San Diego, CAI. cells (Rabinovitch, 1983): 76 Cells in Initial GJG, = ~lOO-n)lO-r’t-Tb’ + n (for t >‘ Tb) Defined medium and serum-free culture (’% G,,/G, Cells = 100 f o r t
