0013-7227/78/1024-1155$02.00/0 Endocrinology Copyright © 1978 by The Endocrine Society
Vol. 102, No. 4 Printed in U.S.A.
Effect of Glucagon and Insulin on the Growth of Neonatal Rat Hepatocytes in Primary Tissue Culture UBALDO ARMATO,* ENRICA DRAGHI, AND PAOLA G. ANDREIS Tissue Culture Laboratory, Department of Human Anatomy, University ofPadova, 1-35100, Padova, Italy ABSTRACT. Commercial (bovine-porcine) glucagon added in a single dose between 10~12 and 10~7 M to neonatal rat hepatocytes in primary cultures with subsequent incubation for 20-24 h, stimulated their entry into the DNA synthesis phase as revealed by [3H]thymidine-labeling and radioautography; about 14 h of incubation was required before an effect was observed. Commercial (bovine) insulin at doses between 10"" and 10~7 M apparently stimulated the entry of hepatocytes into S phase. However, insulin's effect, which needed 20 h for induction, was due to the release of a wave of synchronized hepatocytes from an earlier produced block near the Gi/S boundary of their growth-division cycle. Equimolar mixtures of glucagon with insulin from 10~l5-10~7 M increased the fraction of hepatocytes synthesizing DNA first at 4-8 h, and then at 20-24 h. Effective doses of glucagon, insulin, and glucagon plus insulin also increased the entry of hepatocytes into mitosis, as found after a 4-h incubation with colchicine
R
OUS AND LARIMORE (1) assumed that some hepatotrophic factors were present in the portal venous blood because liver failure resulted in the dog from Eck's fistula (i.e., portacaval deviation). This hepatotrophic theory has been either forgotten or questioned (2) until recent complex surgical studies revived it (3-12), showing that insulin, either by itself or associated with glucagon, did affect the size, chemical composition, and, once mitotic activation had occurred (8, 9), the proliferative activity of hepatocytes. Despite these observations, the regulation of liver growth and the roles therein of insulin and glucagon are still highly controversial (2, 8-12). As in vivo studies may be complicated by simultaneous complex neuro-humoral interactions, we thought it worthwhile to use primary cultures of differentiated neonatal rat hepatocytes (13-17) to try to identify the enReceived February 16, 1977. * To whom all correspondence and requests for reprints should be addressed at: Istituto di Anatomia Umana Normale, Via A. Gabelli 65,1-35100, Padova, Italy.
(0.1 mM). Withholding inactivated fetal bovine serum from the growth medium for 24 h did not change the mitotic activity either of the untreated or of the glucagon- and glucagon plus insulin-stimulated hepatocytes, but it increased the proliferogenic effect of bovine insulin. Highly purified crystalline (porcine) glucagon, insulin, and glucagon plus insulin also stimulated the growth of hepatocytes in the presence or absence of serum. Finally, equimolar (10~H M) mixtures of glucagon with (BuhcGMP and of insulin with (Bu>2cAMP increased the hepatocytic replication as efficiently as did glucagon plus insulin at the same dose. The present results show that glucagon and insulin are synergistic, intracycle regulators of the growth of neonatal rat hepatocytes. They also suggest that cyclic nucleotides may mediate at least partly the hepatotrophic effects of the pancreatic hormones. (Endocrinology 102: 1155, 1978)
docrine factors participating in the regulation of liver growth. With such an in vitro model, we first tested the feasibility of insulin and glucagon as intracycle regulators of liver cell replication. Our data show that glucagon and insulin are major synergistic positive stimulators of the growth of cycling neonatal rat hepatocytes, and they also suggest that cyclic nucleotides may participate in the hepatotrophic action of the pancreatic hormones. Materials and Methods The following materials were obtained: Hanks' Basal Salt Solution (BSS) and Eagle's Minimum Essential Medium (MEM) (Wellcome); adult and fetal bovine serum (Flow Laboratories); trypsin 1:250 powder (Difco); collagenase type I (from Clostridium histolyticum), hyaluronidase type II (from ovine testes), bovine-porcine glucagon, bovine insulin, cAMP, (Bu)acAMP, cGMP, (Bu)2cGMP, butyrate, thymidine, and colchicine (Sigma); cephaloridin (Eli Lilly); streptomycin (Farmitalia); nystatin (Squibb); [me£/iy/-3H]thymidine (specific ac-
1155
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
1156
ARMATO, DRAGHI, AND ANDREIS
tivity, 29 Ci/mmol) (Radiochemical Centre); K5 Nuclear Emulsion and Hypam® (Ilford-Ciba); D19a developer (Kodak). Samples of very highly purified crystalline porcine glucagon and insulin were the kind gift of Dr. Robert J. Hosley of Lilly Research Laboratories. Preparation of liver cultures and labeling with pHJthymidine Primary cultures of enzymatically dispersed hepatocytes were prepared by using pooled livers taken aseptically from 4-day-old Wistar albino rats (Morini), male and female littermates, as previously detailed (13-17). The basic (control) medium consisted of Eagle's MEM containing 10% (vol/vol) fetal bovine serum (inactivated at 56 C for 30 min), aqueous cephaloridin and streptomycin (both at 50 jug/ml), nystatin (25 IU/ml), and 1-2 ml NaHCO3 (8% wt/vol) to bring the final pH to 7.2-7.3. The cultures were incubated at 35 C in a gas phase of 5% (vol/vol) CO2 in air and checked every day for viability and cell growth according to Armato and coworkers (13-17). To estimate the size of the hepatocyte population that enters the S phase, groups of liver cultures on their 4th day in vitro were treated for 24 h with the following compounds directly dissolved in the growth medium in doses ranging from 10~15-10~17 M: (a) commercial glucagon or insulin; (b) equimolar combinations of glucagon with (i) insulin, (ii) cAMP, or (iii) cGMP; (c) equimolar mixtures of insulin with (i) cAMP or (ii) cGMP. After treatment, the cultures were labeled for 1 h at 35 C with 1.0 /iCi/ml [metfiyZ-3H]thymidine, fixed and processed for radioautography (14-16). In other experiments, in order to determine the timing of the onset of entry of cells into DNA synthesis in the presence of the most effective doses of commercial glucagon (10~8 M), insulin (10~10 M), or glucagon plus equimolar insulin (10~14 M) observations were made every 4 h for 24 h. Radioautography and evaluation of entry of hepatocytes into DNA synthesis Radioautographs were prepared as already described (14-16). To reduce the incubation time needed for radioautography to 10 days, [3H]thymidine of higher specific activity and in twice the amount formerly employed (14-16) was used in the present experiments. Phase S hepatocytes were recognized in developed and hematoxylin-eosinstained specimens by optical counts at 1200 X magnification. To test the specificity of the effects of glucagon and insulin for hepatocytes, the percent-
Endo • 1978 Vol 102 • No 4
age of phase S fibroblast-like cells was also estimated in liver cultures exposed to the most effective doses of these hormones used alone and in equimolar combinations. Estimation of the numbers of hepatocytes entering mitosis To evaluate the entry of cultured hepatocytes into mitosis, groups of 4-day-old cultures were treated for 20 h with the same test agents as were used for the [3H]thymidine-labeled experiments. After treatment, the specimens were incubated for 4 h with normal medium containing colchicine (0.12 mM) and ethanol (1.0 X 10"4% vol/vol). In fixed and Feulgen-stained cultures, the percentage of the hepatocyte population stopped in the so called "Cmetaphase" was determined as previously described (15, 16). Similar ancillary experiments tested the effects, if any, of heat-inactivated fetal bovine serum, usually added to the growth medium, on the hormonal actions observed. Hence, selected amounts of both commercial and highly purified crystalline glucagon (10~8 M), insulin (10~10 M), and equimolar glucagon plus insulin (10~14 M) were dissolved into Eagle's MEM with no serum added. The resulting media were added to liver cultures for 20 h before incubating the specimens with colchicine for 4 h. Finally, to investigate if cyclic nucleotides can replace, at least partly, pancreatic hormones, liver cultures were incubated for 20 h with combinations of either commercial or highly purified crystalline (a) glucagon (10~14 M) with (i) (Bu)2cGMP (10"14 M), (ii) butyrate (10~14 M), or (iii) cGMP (10"11 M); (b) insulin (10~14 M) with (i) (Bu)2cAMP (lO"14 M), (ii) butyrate (10"14 M), or (iii) cAMP (10" u M) before the 4-h incubation with colchicine. In every experiment, groups of untreated cultures derived from the same animal and fed with normal growth medium with or without inactivated fetal bovine serum were run in parallel as controls.
Statistical analysis Statistical significance was determined by using Student's t test for unpaired samples. Only differences between experimental groups with P < 0.05 were considered as significant. Where appropriate, curve fitting was performed by weighted regression analysis of the data (18). To evaluate and compare the total 24-h effect of the most active doses of commercial glucagon and insulin, used singly and in equimolar combination, the areas beneath the corresponding curves were approximated according to the equation:
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
GLUCAGON, INSULIN, AND HEPATOCYTE GROWTH
1157
f(Xn)l =
where Ii is the area and h is the interval between equidistant Xo, xi • • xn points. Because no difference was found between the control groups at the various points, the controls were pooled and the area beneath their curve calculated by assuming that f(xo) was equal to f(xi) • • f(xn).
Results Primary cultures of neonatal rat liver cells Primary cultures derived from dissociated neonatal rat liver cells contained hepatocytes (28% of the total population) and also several types of mesenchymal cells (13-17). We considered only the hepatocytes. They were grouped mostly in monolayered homogeneous islets of 10-500 elements (Fig. la). Hepatocytes were shown to contain, at least up to the 25th day in vitro, several types of exportable serum proteins, including proalbumin and fibrinogen (Fig. lb) (17,19). If exposed to phenobarbitone, they increased significantly their capacity to survive cytotoxic concentrations of vinblastine sulfate (20). Unlike the coexisting fibroblast-like cells and the so called "clear cells", hepatocytes took up intensely and specifically [3H]bilirubin from the growth medium (17). Finally, their incorporation of [3H] leucine was affected by cAMP in a way distinct from that of the neighboring mesenchymal cells (13). Entry of hepatocytes into DNA synthesis and mitosis On the fifth day in vitro, the untreated (control) hepatocytes entered moderately the S and M stages of their mitotic cell cycle (Figs. 2-9). After a 20-24-h exposure to commercial (bovine-porcine) glucagon at physiological and higher doses (21), the cycling hepatocytes entered at an increased rate into both DNA synthesis and mitosis. Commercial (bovine) insulin at physiological and higher concentrations (21) appeared to stimulate the entry of hepatocytes into both S and M phases, but was less active than glucagon. Equimolar combinations of these two hormones at all the
FlG. 1. Neonatal rat liver cells in culture, a. An islet of primary rat hepatocytes (H) on the fifth day in vitro. Cells attached to the polyethylene disc were fixed in glutaraldehyde (2.5% wt/vol) and stained with uranyl acetate. The cytoplasm of parenchymal cells, although resembling a honeycomb because of liposome extraction during histological processing, is much darker, due to its richness in organelles, than that of the neighboring fibroblast-like cells (F). Wandering macrophages (M), which move often superficial to the parenchymal cells (X500). b. A 6-day-old primary rat liver culture fixed for 20 min at -30 C in absolute acetone, incubated for 1 h at 37 C with goat antibodies to adult rat fibrinogen, and then stained for 1 h with fluorescein-conjugated rabbit anti-goat globulins. Hepatocytes, whose cytoplasm fluoresces brightly, are arranged in an irregularly shaped islet, surrounded by negative (hence, not detectable) mesenchymal cells. Rather similar immunofluorescent pictures are obtained when antisera to adult rat albumin or total serum proteins were used (X300).
doses tested enhanced intensely the DNA synthetic and mitotic activities of hepatocytes and were particularly more effective than
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
ARMATO, DRAGHI, AND ANDREIS
1158
9
-
8
-
Kndo • 1978 Vol 102 • No 4
insulm plus , glucagon
13
12
1 1
o
7—
if
10
6 -
o o X M a. i
"2 "*
fs
•8 I
•g o
5 -
-I •|i!
II 5""
_
a
(a)
rb
I 15
10
-14
10
I 13
10
-12
10
11
10
10-9
10
-8
-7
14
13
12
10 10 10 10 10 10 10 hormone(s) concentration (moles/I of medium)
10
10
9
10
8
10
-7
10
FIG. 2. Effect of glucagon, insulin, and equimolar mixtures of the two hormones on the entry into DNA synthesis (a) and mitosis (b) of cultured cycling hepatocytes. D, Controls. The points are means ± SE of the values from 8-11 distinct cultures.
either hormone alone at concentrations of 10~12 M or below (Fig. 2a, b). The results in Fig. 3a, b show that commercial glucagon (10~8 M) increased the flow of hepatocytes into DNA synthesis after about 14 h and into mitosis after 10 h of continued treatment. Insulin (10~10 M) significantly decreased the entry of hepatocytes into S phase between the 8th and the 16th hour of treatment and then increased it from the 20th hour onwards. The mitotic activity of hepatic cells was also stimulated by insulin after 14 h. Treatment with an equimolar (10~14 M) combination of the two hormones by 4 h had effected an early increase in the entry into DNA synthesis and a second increase after 20 h. This hormonal combination also elicited two peaks of dividing hepatocytes after 4 and 24 h. The various control groups of the exper-
iments shown in Fig. 3a, b were pooled and put on the ordinate because no notable difference was found among the groups at the different time points. This finding is in keeping with in vivo studies showing that nyctohemeral changes in DNA synthetic and mitotic activities of rat hepatocytes can be observed only after the first 3 weeks of postnatal life (22). Mathematical analysis of the data in Fig. 3, a and b, shows (Table 1) that glucagon did truly increase the fractions of hepatocytes entering S and M within 24 h, whereas insulin did not change the total 24-h amount of DNAsynthesizing cells while augmenting the 24-h fraction entering mitosis. Table 1 also demonstrates that the combination of glucagon with insulin at 10"14 M was 4-6 orders of magnitude more effective than either hormone alone in increasing the number of DNA-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
GLUCAGON, INSULIN, AND HEPATOCYTE GROWTH
1159
4.5 -
FIG. 3. The timing of the onset of stimulation by glucagon (10~H M; • ) , insulin (10"'° M; ©), and by an equimolar mixture of the two hormones (lO"14 M; • ) of the transit of hepatocytes into the S (a) and M (b) stages of their mitotic cycle. *, Controls. Mitotic activity is the sum of the fractions of cells which were in prophase, metaphase, anaphase, and telophase at the time of the exposure of the cultures to [•'HJthymidine. The points are means ± SE of the values from four distinct experiments.
8
12
16 20 24 0 4 8 hours of exposure to hormone(s)
12
16 20
24
TABLE 1. The total 24-h effects of glucagon and insulin on the entry of rat hepatocytes into DNA synthesis and mitosis phases of the cell cycle Treatment None
Glucagon (10~8 M ) A Insulin (10"10 M ) A Ai
Glucagon plus insulin (lO"14 M)
Phase S hepatocytes 6360 ± 240 8956 ± 555 + 40.8 6403 ± 315 + 0.7 -28.5 9615 ± 605
P
Vol. 102, No. 4 Printed in U.S.A.
Effect of Glucagon and Insulin on the Growth of Neonatal Rat Hepatocytes in Primary Tissue Culture UBALDO ARMATO,* ENRICA DRAGHI, AND PAOLA G. ANDREIS Tissue Culture Laboratory, Department of Human Anatomy, University ofPadova, 1-35100, Padova, Italy ABSTRACT. Commercial (bovine-porcine) glucagon added in a single dose between 10~12 and 10~7 M to neonatal rat hepatocytes in primary cultures with subsequent incubation for 20-24 h, stimulated their entry into the DNA synthesis phase as revealed by [3H]thymidine-labeling and radioautography; about 14 h of incubation was required before an effect was observed. Commercial (bovine) insulin at doses between 10"" and 10~7 M apparently stimulated the entry of hepatocytes into S phase. However, insulin's effect, which needed 20 h for induction, was due to the release of a wave of synchronized hepatocytes from an earlier produced block near the Gi/S boundary of their growth-division cycle. Equimolar mixtures of glucagon with insulin from 10~l5-10~7 M increased the fraction of hepatocytes synthesizing DNA first at 4-8 h, and then at 20-24 h. Effective doses of glucagon, insulin, and glucagon plus insulin also increased the entry of hepatocytes into mitosis, as found after a 4-h incubation with colchicine
R
OUS AND LARIMORE (1) assumed that some hepatotrophic factors were present in the portal venous blood because liver failure resulted in the dog from Eck's fistula (i.e., portacaval deviation). This hepatotrophic theory has been either forgotten or questioned (2) until recent complex surgical studies revived it (3-12), showing that insulin, either by itself or associated with glucagon, did affect the size, chemical composition, and, once mitotic activation had occurred (8, 9), the proliferative activity of hepatocytes. Despite these observations, the regulation of liver growth and the roles therein of insulin and glucagon are still highly controversial (2, 8-12). As in vivo studies may be complicated by simultaneous complex neuro-humoral interactions, we thought it worthwhile to use primary cultures of differentiated neonatal rat hepatocytes (13-17) to try to identify the enReceived February 16, 1977. * To whom all correspondence and requests for reprints should be addressed at: Istituto di Anatomia Umana Normale, Via A. Gabelli 65,1-35100, Padova, Italy.
(0.1 mM). Withholding inactivated fetal bovine serum from the growth medium for 24 h did not change the mitotic activity either of the untreated or of the glucagon- and glucagon plus insulin-stimulated hepatocytes, but it increased the proliferogenic effect of bovine insulin. Highly purified crystalline (porcine) glucagon, insulin, and glucagon plus insulin also stimulated the growth of hepatocytes in the presence or absence of serum. Finally, equimolar (10~H M) mixtures of glucagon with (BuhcGMP and of insulin with (Bu>2cAMP increased the hepatocytic replication as efficiently as did glucagon plus insulin at the same dose. The present results show that glucagon and insulin are synergistic, intracycle regulators of the growth of neonatal rat hepatocytes. They also suggest that cyclic nucleotides may mediate at least partly the hepatotrophic effects of the pancreatic hormones. (Endocrinology 102: 1155, 1978)
docrine factors participating in the regulation of liver growth. With such an in vitro model, we first tested the feasibility of insulin and glucagon as intracycle regulators of liver cell replication. Our data show that glucagon and insulin are major synergistic positive stimulators of the growth of cycling neonatal rat hepatocytes, and they also suggest that cyclic nucleotides may participate in the hepatotrophic action of the pancreatic hormones. Materials and Methods The following materials were obtained: Hanks' Basal Salt Solution (BSS) and Eagle's Minimum Essential Medium (MEM) (Wellcome); adult and fetal bovine serum (Flow Laboratories); trypsin 1:250 powder (Difco); collagenase type I (from Clostridium histolyticum), hyaluronidase type II (from ovine testes), bovine-porcine glucagon, bovine insulin, cAMP, (Bu)acAMP, cGMP, (Bu)2cGMP, butyrate, thymidine, and colchicine (Sigma); cephaloridin (Eli Lilly); streptomycin (Farmitalia); nystatin (Squibb); [me£/iy/-3H]thymidine (specific ac-
1155
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
1156
ARMATO, DRAGHI, AND ANDREIS
tivity, 29 Ci/mmol) (Radiochemical Centre); K5 Nuclear Emulsion and Hypam® (Ilford-Ciba); D19a developer (Kodak). Samples of very highly purified crystalline porcine glucagon and insulin were the kind gift of Dr. Robert J. Hosley of Lilly Research Laboratories. Preparation of liver cultures and labeling with pHJthymidine Primary cultures of enzymatically dispersed hepatocytes were prepared by using pooled livers taken aseptically from 4-day-old Wistar albino rats (Morini), male and female littermates, as previously detailed (13-17). The basic (control) medium consisted of Eagle's MEM containing 10% (vol/vol) fetal bovine serum (inactivated at 56 C for 30 min), aqueous cephaloridin and streptomycin (both at 50 jug/ml), nystatin (25 IU/ml), and 1-2 ml NaHCO3 (8% wt/vol) to bring the final pH to 7.2-7.3. The cultures were incubated at 35 C in a gas phase of 5% (vol/vol) CO2 in air and checked every day for viability and cell growth according to Armato and coworkers (13-17). To estimate the size of the hepatocyte population that enters the S phase, groups of liver cultures on their 4th day in vitro were treated for 24 h with the following compounds directly dissolved in the growth medium in doses ranging from 10~15-10~17 M: (a) commercial glucagon or insulin; (b) equimolar combinations of glucagon with (i) insulin, (ii) cAMP, or (iii) cGMP; (c) equimolar mixtures of insulin with (i) cAMP or (ii) cGMP. After treatment, the cultures were labeled for 1 h at 35 C with 1.0 /iCi/ml [metfiyZ-3H]thymidine, fixed and processed for radioautography (14-16). In other experiments, in order to determine the timing of the onset of entry of cells into DNA synthesis in the presence of the most effective doses of commercial glucagon (10~8 M), insulin (10~10 M), or glucagon plus equimolar insulin (10~14 M) observations were made every 4 h for 24 h. Radioautography and evaluation of entry of hepatocytes into DNA synthesis Radioautographs were prepared as already described (14-16). To reduce the incubation time needed for radioautography to 10 days, [3H]thymidine of higher specific activity and in twice the amount formerly employed (14-16) was used in the present experiments. Phase S hepatocytes were recognized in developed and hematoxylin-eosinstained specimens by optical counts at 1200 X magnification. To test the specificity of the effects of glucagon and insulin for hepatocytes, the percent-
Endo • 1978 Vol 102 • No 4
age of phase S fibroblast-like cells was also estimated in liver cultures exposed to the most effective doses of these hormones used alone and in equimolar combinations. Estimation of the numbers of hepatocytes entering mitosis To evaluate the entry of cultured hepatocytes into mitosis, groups of 4-day-old cultures were treated for 20 h with the same test agents as were used for the [3H]thymidine-labeled experiments. After treatment, the specimens were incubated for 4 h with normal medium containing colchicine (0.12 mM) and ethanol (1.0 X 10"4% vol/vol). In fixed and Feulgen-stained cultures, the percentage of the hepatocyte population stopped in the so called "Cmetaphase" was determined as previously described (15, 16). Similar ancillary experiments tested the effects, if any, of heat-inactivated fetal bovine serum, usually added to the growth medium, on the hormonal actions observed. Hence, selected amounts of both commercial and highly purified crystalline glucagon (10~8 M), insulin (10~10 M), and equimolar glucagon plus insulin (10~14 M) were dissolved into Eagle's MEM with no serum added. The resulting media were added to liver cultures for 20 h before incubating the specimens with colchicine for 4 h. Finally, to investigate if cyclic nucleotides can replace, at least partly, pancreatic hormones, liver cultures were incubated for 20 h with combinations of either commercial or highly purified crystalline (a) glucagon (10~14 M) with (i) (Bu)2cGMP (10"14 M), (ii) butyrate (10~14 M), or (iii) cGMP (10"11 M); (b) insulin (10~14 M) with (i) (Bu)2cAMP (lO"14 M), (ii) butyrate (10"14 M), or (iii) cAMP (10" u M) before the 4-h incubation with colchicine. In every experiment, groups of untreated cultures derived from the same animal and fed with normal growth medium with or without inactivated fetal bovine serum were run in parallel as controls.
Statistical analysis Statistical significance was determined by using Student's t test for unpaired samples. Only differences between experimental groups with P < 0.05 were considered as significant. Where appropriate, curve fitting was performed by weighted regression analysis of the data (18). To evaluate and compare the total 24-h effect of the most active doses of commercial glucagon and insulin, used singly and in equimolar combination, the areas beneath the corresponding curves were approximated according to the equation:
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
GLUCAGON, INSULIN, AND HEPATOCYTE GROWTH
1157
f(Xn)l =
where Ii is the area and h is the interval between equidistant Xo, xi • • xn points. Because no difference was found between the control groups at the various points, the controls were pooled and the area beneath their curve calculated by assuming that f(xo) was equal to f(xi) • • f(xn).
Results Primary cultures of neonatal rat liver cells Primary cultures derived from dissociated neonatal rat liver cells contained hepatocytes (28% of the total population) and also several types of mesenchymal cells (13-17). We considered only the hepatocytes. They were grouped mostly in monolayered homogeneous islets of 10-500 elements (Fig. la). Hepatocytes were shown to contain, at least up to the 25th day in vitro, several types of exportable serum proteins, including proalbumin and fibrinogen (Fig. lb) (17,19). If exposed to phenobarbitone, they increased significantly their capacity to survive cytotoxic concentrations of vinblastine sulfate (20). Unlike the coexisting fibroblast-like cells and the so called "clear cells", hepatocytes took up intensely and specifically [3H]bilirubin from the growth medium (17). Finally, their incorporation of [3H] leucine was affected by cAMP in a way distinct from that of the neighboring mesenchymal cells (13). Entry of hepatocytes into DNA synthesis and mitosis On the fifth day in vitro, the untreated (control) hepatocytes entered moderately the S and M stages of their mitotic cell cycle (Figs. 2-9). After a 20-24-h exposure to commercial (bovine-porcine) glucagon at physiological and higher doses (21), the cycling hepatocytes entered at an increased rate into both DNA synthesis and mitosis. Commercial (bovine) insulin at physiological and higher concentrations (21) appeared to stimulate the entry of hepatocytes into both S and M phases, but was less active than glucagon. Equimolar combinations of these two hormones at all the
FlG. 1. Neonatal rat liver cells in culture, a. An islet of primary rat hepatocytes (H) on the fifth day in vitro. Cells attached to the polyethylene disc were fixed in glutaraldehyde (2.5% wt/vol) and stained with uranyl acetate. The cytoplasm of parenchymal cells, although resembling a honeycomb because of liposome extraction during histological processing, is much darker, due to its richness in organelles, than that of the neighboring fibroblast-like cells (F). Wandering macrophages (M), which move often superficial to the parenchymal cells (X500). b. A 6-day-old primary rat liver culture fixed for 20 min at -30 C in absolute acetone, incubated for 1 h at 37 C with goat antibodies to adult rat fibrinogen, and then stained for 1 h with fluorescein-conjugated rabbit anti-goat globulins. Hepatocytes, whose cytoplasm fluoresces brightly, are arranged in an irregularly shaped islet, surrounded by negative (hence, not detectable) mesenchymal cells. Rather similar immunofluorescent pictures are obtained when antisera to adult rat albumin or total serum proteins were used (X300).
doses tested enhanced intensely the DNA synthetic and mitotic activities of hepatocytes and were particularly more effective than
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
ARMATO, DRAGHI, AND ANDREIS
1158
9
-
8
-
Kndo • 1978 Vol 102 • No 4
insulm plus , glucagon
13
12
1 1
o
7—
if
10
6 -
o o X M a. i
"2 "*
fs
•8 I
•g o
5 -
-I •|i!
II 5""
_
a
(a)
rb
I 15
10
-14
10
I 13
10
-12
10
11
10
10-9
10
-8
-7
14
13
12
10 10 10 10 10 10 10 hormone(s) concentration (moles/I of medium)
10
10
9
10
8
10
-7
10
FIG. 2. Effect of glucagon, insulin, and equimolar mixtures of the two hormones on the entry into DNA synthesis (a) and mitosis (b) of cultured cycling hepatocytes. D, Controls. The points are means ± SE of the values from 8-11 distinct cultures.
either hormone alone at concentrations of 10~12 M or below (Fig. 2a, b). The results in Fig. 3a, b show that commercial glucagon (10~8 M) increased the flow of hepatocytes into DNA synthesis after about 14 h and into mitosis after 10 h of continued treatment. Insulin (10~10 M) significantly decreased the entry of hepatocytes into S phase between the 8th and the 16th hour of treatment and then increased it from the 20th hour onwards. The mitotic activity of hepatic cells was also stimulated by insulin after 14 h. Treatment with an equimolar (10~14 M) combination of the two hormones by 4 h had effected an early increase in the entry into DNA synthesis and a second increase after 20 h. This hormonal combination also elicited two peaks of dividing hepatocytes after 4 and 24 h. The various control groups of the exper-
iments shown in Fig. 3a, b were pooled and put on the ordinate because no notable difference was found among the groups at the different time points. This finding is in keeping with in vivo studies showing that nyctohemeral changes in DNA synthetic and mitotic activities of rat hepatocytes can be observed only after the first 3 weeks of postnatal life (22). Mathematical analysis of the data in Fig. 3, a and b, shows (Table 1) that glucagon did truly increase the fractions of hepatocytes entering S and M within 24 h, whereas insulin did not change the total 24-h amount of DNAsynthesizing cells while augmenting the 24-h fraction entering mitosis. Table 1 also demonstrates that the combination of glucagon with insulin at 10"14 M was 4-6 orders of magnitude more effective than either hormone alone in increasing the number of DNA-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 10 June 2015. at 17:00 For personal use only. No other uses without permission. . All rights reserved.
GLUCAGON, INSULIN, AND HEPATOCYTE GROWTH
1159
4.5 -
FIG. 3. The timing of the onset of stimulation by glucagon (10~H M; • ) , insulin (10"'° M; ©), and by an equimolar mixture of the two hormones (lO"14 M; • ) of the transit of hepatocytes into the S (a) and M (b) stages of their mitotic cycle. *, Controls. Mitotic activity is the sum of the fractions of cells which were in prophase, metaphase, anaphase, and telophase at the time of the exposure of the cultures to [•'HJthymidine. The points are means ± SE of the values from four distinct experiments.
8
12
16 20 24 0 4 8 hours of exposure to hormone(s)
12
16 20
24
TABLE 1. The total 24-h effects of glucagon and insulin on the entry of rat hepatocytes into DNA synthesis and mitosis phases of the cell cycle Treatment None
Glucagon (10~8 M ) A Insulin (10"10 M ) A Ai
Glucagon plus insulin (lO"14 M)
Phase S hepatocytes 6360 ± 240 8956 ± 555 + 40.8 6403 ± 315 + 0.7 -28.5 9615 ± 605
P
