Characterization of the growth of murine fibroblasts that express human insulin receptors

EXPERIMENTAL

CELL

RESEARCH

198,

31-39 (1990)

aracterization of the Growth of Murin That Express Human insulin Ii. Interaction of Insulin with Other

rowth Factors

PAUL A. RANDAZZO AND LEONARD JARETFT~ Department

of Pathology

and Laboratory Medicine, University of Pennsylvania School of Medicine, 36th and Hamilton Walk, Philadelphia, Fennsyluania 19104

murine 3T3 fibroblasts [l-8]. This latter ceil type is an attractive system for studying mechanisms of insulinstimulated growth because the cells have been well characterized [8]. Insulin and insulin-like growth factor-l (IGF-1) have both been found to act as progression factors, inducing DNA synthesis either when a second growth factor such as epidermal growth factor (EGF2) is present or when cells have been made competent by, for instance, treatment with platelet-derived growth factor (PDGF) or by oncogenic transformation [S-10]. However, since the contribution of insulin receptor, compared to IGF-1 receptor, to the growth response is small in 3T3 fibroblasts [g-11], examination of an insulin receptor specific pathway may be difficult. The expression of additional insulin receptors was thought to be a means of augmenting insulin receptor-mediated growth and, consequently, to aid in the examination of insulin receptor-specific mechanisms. Hence, insulin receptormediated growth has been examined in 3T3 fibroblasts which ectopically express human insulin receptor consequent to transfection with an expression vector containing human insulin receptor cDNA (NIH 3T3/HIR cells) [ll-141. Insulin, in the absence of a second mitogen, was found to induce growth to the same extent as did 10% serum. Because this behavior was atypical for the effect of a progression factor on an apparently quiescent cell, the studies described in this report were undertaken to further characterize the growth of NIH 3T31 HIR cells. The requirement for a single growth factor was found to be specific for insulin. Insulin’s interactions with other growth factors were also studied in an attempt to determine if insulin-stimulated growth was independent of other mitogens and to discriminate among possible mechanisms by which insulin may be

The effects of insulin-like growth factor-l (IGF-l), epidermal growth factor (EGF), platelet-derived growth factor (PDGF), and insulin on DNA synthesis were studied in murine fibroblasts transfected with an expression vector containing human insulin receptor c:DNA (NIH 3T3/HIR) and the parental NIH 3T3 cells. In NIH 3T3/HIR cells, individual growth factors in serum-free medium stimulated DNA synthesis with the following relative efficacies: insulin > 10% fetal calf serum > PDGF > IGF-1% EGF. In comparison, the relative efficacies of these factors in stimulating DNA synthesis by NIH 3T3 cells were 10% fetal calf serum > PDGF > EGF $ IGF-1 = insulin. In NIH 3T3/HIR cells, EGF was synergistic with l-10 rig/ml insulin but not with 100 rig/ml insulin or more. Synergy of PDGF or IGF-1 with insulin was not detected. In the parental NIH 3T3 cells, insulin and IGF-1 were found to be synergistic with EGF (1 rig/ml), PDGF (100 nglml), and PDGF plus EGF. In NIH 3T3/HIR cells, the lack of interaction of insulin with other growth factors was also observed when the percentage of cells synthesizing DNA was examined. Despite insulin’s inducing only 60% of NIH 3T3/HIR cells to incorporate thymidine, addition of PDGF, EGF, or PDGF plus EGF had no further effect. In contrast, combinations of growth factors resulted in 95% of the parental NIH 3T3 cells synthesizing DNA. The independence of insulin-stimulated DNA synthesis from other mitogens in the NIH 3T3/HIR cells is atypical for progression factor-stimulated DNA synthesis and is thought to be partly the result of insulin receptor expression in an inappropriate context or quantity. 0 1990 Academic Press, Inc. -

INTRODUCTION Insulin receptor has been found to mediate a proliferative response in a variety of cultured cells, including ’ To whom correspondence dressed.

and reprint

requests

should

‘Abbreviations used: IGF-I, insulin-like growth factor-l; PDGF, platelet-derived growth factor; EGF, epidermal growth factor; DMEM, Dulbecco’s modified Eagle’s medium; NIH 3T3/HIR, NIH 3T3 murine fibroblasts transfected with an expression vector containing human insulin receptor cDNA.

be ad-

31

‘M14-4827/90

$3.00

Copyright Q 1WO by Academic Press, Inc. All rights of reprodurtion in an:, form reserved.

32

RANDAZZO

stimulating growth. Submaximally stimulatory concentrations of insulin were found to interact with EGF in stimulating DNA synthesis. However, maximally stimulatory concentrations of insulin were found not to synergize with either PDGF or EGF, despite inducing only 60% of the cells to synthesize DNA. Insulin’s lack of a requirement for a second mitogen and the lack of synergy with other growth factors is atypical for a progression factor. The NIH 3T3/HIR may be an example of aberrant growth, resulting from the expression of a normal protein, in this case insulin receptor, in an inappropriate quantity or context. MATERIALS

AND

METHODS

Materials. NIH 3T3 fibroblasts expressing approximately 10’ human insulin receptors/cell [14] consequent to transfection with an expression vector containing human insulin receptor cDNA (NIH 3TJ/HIR) and the parent NIH 3T3 fibroblasts which express approximately 1400 insulin receptors/cell [ 141 have been described [ 111 and were kindly provided by Dr. Jonathan Whittaker, State University of New York, Stonybrook. Dulbecco’s modified Eagle’s medium with 1.0 g glucose/liter and 110 mM pyruvate (DMEM), glutamine, gentamytin, bovine fetal calf serum, ethylenediaminetetraacetic acid (EDTA) and trypsin were from GIBCO. Culture Hasks and four-well Lab-Tek glass slides were from NUNC. Twenty-four-well polystyrene culture plates were from Costar and Falcon. Porcine insulin was from Sigma and was filter sterilized. Insulin-like growth factor-l (IGF-l), epiderma1 growth factor (EGF), and porcine platelet-derived growth factor (PDGF) were from ICN. Receptor-quality porcine PDGF was from Mallinckrodt. ‘2”I-labeled insulin, prepared as described in Hammons and Jarett [ 151, was kindly provided by Robert M. Smith. ‘%I-labeled IGF-1 and iz’I-labeled EGF were from Amersham. Insulin-free bovine serum albumin was from United States Biochemical Corp. All other reagents were from Sigma. Methods. Stock cultures were grown in T-17*5 Basks using DMEM with 10% fetal calf serum, 4 mM glutamine, and 50 mg gentamycin/liter in a humidified 10% CO,/90% air atmosphere. For experiments, cells were seeded in 24-well dishes (17-mm-diameter wells) or four-well Lab-Tek slides at a density of approximately 5 X 103 cells/ well using 1 ml of DMEM with 10% fetal calf serum, 4 mM glutamine, and 50 mg gentamycin/liter per well. On the third day, the medium was replaced with fresh medium. On Day 6, the cells were at conHuence for more than 1 day and experimental conditions were established. The monolayers were rinsed two times with phosphate-buffered saline (IO mh4 sodium phosphate, pH 7.4, in 0.15 M sodium chloride) and the medium was replaced with serum-free medium (DMEM with 0.1% bovine serum albumin (BSA), 4 mM glutamine, and 50 mg gentamycinlliter) with the indicated additions or, as a control, with DMEM containing 10% fetal calf serum, 4 mM glutamine, and .50 mg gentamycin/liter. EGF and IGF-1 binding to confluent monolayers was determined as described [ 111. The data were analyzed by nonlinear least squares fitting of a one-site binding model with nonspecific binding using the computer program Enzfitter 1161. DNA synthesis was determined by [methyl-3H]thymidine incorporation into trichloroacetic acid (TCA)-precipitable material. Seventeen and one-half hours after establishing the indicated experimental conditions, 2 &!i of [aH]thymidine (1 Ci/mmol) was added to each well. The cells were incubated for 1 h at 37°C and the reaction was stopped by aspirating the medium and adding 10% trichloroacetic acid. The monolayer was washed three times with TCA and then dissolved in 500 ~1 0.1 M NaOH. A 200-~1 sample of the solution was taken, neutralized with HCI, and mixed with 4 ml of Aquasol scintil-

AND

JARE’IT

lant. Radioisotope content was determined by scintillation spectroscopy using a Beckman LS 2800. The percentage of cells, grown on Lab-Tek slides, incorporating [3H]thymidine during 24 h was determined by autoradiography as described [ 171. Insulin and IGF-1 degradation were estimated as the percentage of ‘%I-labeled ligand precipitated by 5% TCA. A tracer quantity of iodinated insulin (2000 cpm/well, SA = 150 &i/rig) or IGF-1 (2000 cpm/well, SA = 333 rC!ilng) was added with the indicated quantities of the relevant unlabeled ligand. After the indicated times, the medium was withdrawn and added to an equal volume of cold 10% TCA, and the suspension was centrifuged at 2OUOgfor 10 min. The gamma radiation in the supernatant plus pellet and in the pellet alone was determined using a Beckman gamma 4000 counter. Data are reported as the mean + SEM for the indicated number of experiments performed in triplicate unless otherwise indicated. Statistical analyses are performed as described in Refs. [18] and [19] using the software provided with Ref. [19] or using Systat 1201.

RESULTS

To determine the role of insulin receptor on cell growth, NIH 3T3/HIR cells, which express approximately lo6 insulin receptors/cell, and the parental NIH 3T3 cells, which express approximately 1400 receptors/ cell, were examined. Insulin, in the absence of a second mitogen, was found not to affect the growth of NIH 3T3 cells but was as efficacious as 10% serum in inducing apparently quiescent NIH 3T3/HIR cells to grow [14]. However, in density arrested cells, normally two or more mitogens are required to achieve a response of this magnitude. The effect of other growth factors on DNA synthesis in confluent monolayers of NIH 3T3/HIR and parental NIH 3T3 cells was therefore studied to determine if independence from a second mitogen was specific for insulin. IGF-1, the first growth factor examined, was found not to affect DNA synthesis in the parental NIH 3T3 cells and was found to be less potent and to induce DNA synthesis to a lesser extent than did insulin in the NIH 3T3/HIR cells. As shown in Fig. 1,” the rate of DNA synthesis in the NIH 3T3 cells was not affected by 1 to 500 rig/ml IGF-1. In the NIH 3T3/HIRcells, IGF-1 stimulated DNA synthesis to 43 f 8.5% the rate induced by 10% serum, and a maximal effect was not observed at 500 rig/ml of IGF-1. In contrast, insulin stimulated DNA synthesis in the NIH 3T3/HIR cells to approximately 120% the rate induced by serum [ 141 and a maximum response was achieved at 100 rig/ml of insulin

[12-141. ’ In these studies thymidine incorporated during 1 h was used as an estimate of DNA synthesis as described under Materials and Methods. The data are expressed as a percentage of serum-induced DNA synthesis which is calculated from the thymidine incorporation data as (cpm of thymidine incorporated in the presence of the indicated addition - cpm incorporated into monolayer maintained in serum free medium)/(cpm incorporated in presence of DMEM containing 10% fetal calf serum - cpm incorporated in serum free medium).

INSULIN

AS A GROWTH

The relative magnitude and sensitivities of the IGF-1 and insulin responses cannot be attributed to differences in ligand degradation rates. lz51-labeled ligand was added with 1, 10, or 100 rig/ml of unlabeled ligand to monolayers and the percentage of ligand that remained intact was estimated by the percentage of ligand that remained acid precipitable. NIH 3T3/HIR cells degrade 10 rig/ml insulin threefold faster than 10 rig/ml IGF-1. At 8 h, 60 + 4% (n = 9) of the insulin is degraded compared to 20 -t 3% (n = 3) of the IGF-1 and by 24 h 84 :r 2% (n = 9) of the insulin is degraded whereas 38 + 5% (n = 3) of the IGF-1 was degraded. Degradation rates rneasured at 1 and 100 rig/ml of ligand were indistinguishable from those measured for 10 rig/ml ligand (data not shown). The lack of IGF-1 and insulin responses in the parental cells also cannot be attributed to rapid degradation of the ligands. The parental cells degrade IGF-1 and insulin at one-half the rate at which NIH 3T3/HIR cells degrade IGF-1 and insulin. At 8 h NIH 3T3 cells had degraded 36 f 7% (n = 9) and 8 * 3% (n = 3) of 10 rig/ml insulin and IGF-1, respectively. Degradation rates at 1 and 100 rig/ml ligand were similar (data not shown). The difference between the NIH 3T3/HIR and parental cells’ responses to IGF-1 and insulin is not likely attributable to differences in IGF-I receptor expression. The parental cells were found to contain 3.8 X lo5 f 4 x lo4 IGF-1 binding sites/cell with an affinity of 7.3 10.7 nM compared to 1.5 X lo” f 1.8 X lo4 binding sites/cell with an affinity of 2.7 & 0.4 nM in the NIH 3T3/HIR cells. [‘2”I]-labeled IGF-1 (50 pg/ml) was displaced from both cell types by insulin (data not shown). In the NIH 3T3/HIR cells, the displacement was biphasic. Thirty-five percent of the IGF-1 was displaced by 50

6o 1

-10

.', t

10

Concentration

.I :00

of IGF-1

I,

I

1000

(rig/ml)

FIG. 1. The effect of IGF-1 on the incorporation of [aH]thymidine into NIH 3T3 and NIH 3T3/HIR cells. The indicated concentrat ions of IGF-1, in serum-free medium, were added to confluent monolayers, and thymidine incorporation was determined 17.5 h later as described under Materials and Methods. Data are from eight and three experiments for the NIH 3T3/HIR and NIH 3T3 cells, respectively, and are analyzed by anova foliowed by the Student-Newmana significant difference compared to Keull test 1191. * indicates serum-free medium, I’ -=c0.06.

FACTOR,

33

II

0.01

0.1

EGF Concentration

,

IO

(ng/mil

FIG. 2. EGF-stimulated [3H]thymidine incorporation into NIH 3T3 and NIH 3T3/HIR cells. The indicated concentrations of EGF were added to confluent monolayers in serum-free medium and thymidine incorporation was determined 17.5 h Inter as described under Materials and Methods. Data are from five and three experiments for the NIH 3T3/HIR and NIH 3’1‘3 cells, respectively, and are analyzed by anova followed by the Student-Newman-Keuil test [19]. * indicates a significant difference compared to serum-free medium, P < 0.05.

to 100 nglmi insulin and was presumably bound to the insulin receptor.* Sixty-five percent was displaced by greater than 1 pg/ml of insulin with half-maximal displacement at 5-10 pg/ml of insulin. This 65% was presumably bound to IGF-1 receptor. In the parental NIH 3T3 ceils, displacement was monophasic with half-maximal displacement occurring at an insulin concentration of 5-10 pg/ml. No IGF-1 was displaced with 100 rig/ml insulin. The NIH 3T3/HIR cells were less responsive to EGF than were the parental cells. As shown in Fig. 2, EGF induced DNA synthesis in the parental cells with a halfmaximal effect occurring at 1 rig/ml (1.7 X 10-‘” Mf EGF and a maximal effect that was 33% the effect of serum. EGF did not significantly increase DNA synthesis by the NIH 3T3/HIR cells. The difference in EGF responses between the cells may be attributable to differences in receptor number, as the NIH 3T3 parental cells contained 2400 & 500 receptors/cell with an affinity of 5.35 X lo-” + 3.03 X 10-l’ M and the NIH 3T3/ HIR cells contained 790 -+ 350 receptors!cell with an affinity of 2.83 X 10-l’ -t 1.55 X 10. I1 LM. The effect of PDGF on DNA synthesis by NIH 3T3/ HIR and the parental NIH 3T3 cells was similar. As shown in Fig. 3, in both cell lines PDGF induced DNA synthesis to 60% the extent of that induced by serum and the half-maximal effect occurred at a concentration of 0.15 U/ml. Hence, of the single mitogens tested in the NIH 3T3/HIR cells, insulin stimulated DNA synthesis ’ This relatively large fraction of IGF-1 bound to insulin receptor despite the low affinity of insulin receptor for IGF-1 is expected because of the high concentration of insulin receptor compared to IGF-1 receptor.

*

34

RANDAZZO

. NIH 3TWHlR

O-I 0.01

0.1

PDGF Concentration

1.0

10

(U/ml)

FIG. 3. PDGF-stimulated [3H]thymidine incorporation into NIH 3T3 and NIH 3T3/HIR cells. The indicated concentrations of PDGF in serum-free medium were added to confluent monolayers and thymidine incorporation was determined 17.5 h later as described under Materials and Methods. Data are from three experiments and are analyzed by anova followed by the Student-Newman-Keull test [191.* indicates a significant difference compared to serum-free medium, P < 0.05. The data for one experiment using receptor-quality PDGF (from Mallinckrodt) were similar and did not reveal differences between the cells.

to the greatest extent and was the only single mitogen stimulating DNA synthesis to the same extent as serum. In the’parental NIH 3T3 cells, PDGF was the most efficacious single mitogen which is typical for 3T3 fibroblasts [8]. Individual mitogens usually are synergistic in stimulating DNA synthesis in 3T3 fibroblasts [8,9]. The relative magnitude of insulin-stimulated DNA synthesis in the NIH 3T3/HIR, however, suggested that the insulin response in these cells might be independent of other growth factors. We therefore examined the interaction of insulin with other mitogens. EGF was found to be synergistic in stimulating DNA synthesis with insulin in the parental cells and with 1 to 10 rig/ml insulin in the NIH 3T3/HIR cells. As shown in Fig. 4, in the parental NIH 3T3 cells, 1 rig/ml EGF in serum-free medium induced DNA synthesis to 43% of that elicited by serum. The addition of insulin increased DNA synthesis to 148 f 11% of that elicited by serum. Saturation of the response was not observed at the highest concentration of insulin used (1000 rig/ml). In the NIH 3T3/HIR cells, in the presence of 1 rig/ml of EGF, insulin induced DNA synthesis with a half-maximal response occurring at an insulin concentration of 3 rig/ml (35 rig/ml in the absence of EGF) and a maximum response that was 115% of the response elicited by serum, which was not different from the maximal response occurring in the absence of EGF [ 141. The effect of 0.1 and 10 rig/ml of EGF on insulin-induced DNA synthesis by the NIH 3T3/HIR cells was indistinguishable from the effect of 1 rig/ml of EGF (data not shown).

AND

JARETT

The effect of IGF-1 in the presence of EGF (1 rig/ml) on DNA synthesis was consistent with insulin functioning through the insulin receptor in the NIH 3T3/HIR. In the absence of a second mitogen, insulin-stimulated DNA synthesis was found to be mediated by the insulin receptor [13]. In the presence of EGF, however, the effect of insulin might be mediated through the IGF-1 receptor. To address this possibility, the relative sensitivities of the cells to IGF-1 and insulin in the presence of EGF were examined. In the parental NIH 3T3 cells, IGF-1 was synergistic with EGF. EGF alone, as shown in Fig. 5, stimulated DNA synthesis to 35 f 4% of the serum-stimulated rate. The addition of IGF-1 increased DNA synthesis to 123 + 14% of the serum-stimulated rate with a half maximal effect occurring at an IGF-1 concentration of 2 rig/ml. In contrast, in the presence of

150

A. NIH 3T3

0

0.1

1

10

Insulin Concentration

150

100

1000

hg/ml)

1

6. NIH 3T3/HIR

0.1

1

10

Insulin Concentration

100

1000

(rig/ml)

FIG. 4. The effect of EGF on insulin-induced [3H]thymidine incorporation into NIH 3T3 and NIH 3T3/HIR cells. One nanogram per milliliter of EGF and the indicated concentrations of insulin were added to confluent monolayers of cells in serum-free medium and [3H]thymidine incorporation was measured 17.5 h later as described under Materials and Methods. The data are from three and five experiments in the presence of EGF for the NIH 3T3 and NIH 3T3/ HIR cells, respectively. The data in the absence of EGF are from three and eight experiments for the NIH 3T3 and NIH 3T3/HIR cells, respectively, and are redrawn from Ref. [14]. A and B show the data from the NIH 3T3 and NIH 3T3/HIR cells, respectively. The data from each panel were analyzed by two-way anova. Positive interaction between EGF and insulin occurred in both cell types (P < 0.05).

INSULIN

AS A GROWTH

EGF, 500 rig/ml IGF-1 stimulated the NIH 3T3/HIR cells to synthesize DNA at 83% the rate stimulated by serum and saturation of the response was not observed. These data suggest that in the parental cells, insulin and IGF-1 mediate their effects through the IGF-1 receptor and in the NIH 3T3/HIR cells insulin-stimulated DNA synthesis is mediated through the insulin receptor. PDGF and insulin were weakly synergistic in stimulating DNA synthesis in the NIH 3T3 cells. A positive interaction between these ligands was not identified in the NIH 3T3/HIR cells. As shown in Fig. 6, in the parental cells PDGF stimulated DNA to 35 ? 4% the rate induced by serum, and the addition of insulin increased DNA synthesis to 46 +- 4% of the serum-stimulated rate. In the NIH 3T3/HIR cells, PDGF stimulated DNA syn-

150 1 E 2 ”2: 7120 xz & g

E’ACTOK,

lso

35

II

1 B. NIH 373/HIR

A. NM 3T3

-

n

.

EGF. 1 rig/ml EGF absent

go-

P; 0

‘c :: 60.0 6 En -30-

1

10

too

1300

(q/ml)

v

c o

0.1

Insulin Concentration

&p-----*--A

0

0.1

1

10

IGF-1 Concentration

100 1

100

B. NIH 3T3/HIR

n

1000

hg/ml)

1

EGF. 1 rig/ml

A EGF absent

IGF-1 Concentration

(rig/ml)

FIG. 5. The effect of EGF on IGF- 1-induced [‘HJthymidine incorporation into NIH 31’3 and NIH 3T3/HIR cells. One nanogram per milliliter of EGF and the indicated concentrations of IGF-1 were added to confluent monolayers of cells, and [3H]thymidine incorporaSon was measured 17.5 h later as described under Materials and Methods. The data are from three experiments in the presence of EGF. The data in the absence of EGF are from three and eight experiments for the NIH 3T3 and NIH 3T3/HIR cells, respectively, and are redrawn from Fig. 3. A and B show the data from the NIH 3T3 and NIH 3T3/HIR cells, respectively. The data from each panel were analyzed by two-way anova. Positive interaction between EGF and IGF-1 occurred in both cell types (P < 0.05).

FIG. 6. The effect of PDGF on insulin induced i:‘H]thymidine incorporation into NIH 3T3 and NIH :3T3/HIR cells. One hundred nanograms per milliliter of PDGF and the indicated concentrations of insulin were added to confluent monolayers of cells in serum-free medium, and [‘Hlthymidine incorporation. was measured 17.5 h later aa described under Materials and Methods. The data are from oae experiment performed in triplicate in the presence of PDGF and are reported as the means z SD. The data in the absence of PDGF are from three and eight experiments for the NiH 3’1’3 and NIH 3T3/ HIR cells, respectively, and are redrawn from Ref. ;141. A and B show the data from the NIH 31’3 and NIH 3T3/HIIZ cells, respectively. The data from each pane1 were analyzed hy two-way anova. Positive interaction between PDGF and insulin occurred in the parental NIH 3T3 cells (P -C 0.05) but was not detect.ed in the NIH :3T:!!HIR cells.

thesis to 78 + 7% of the serum-stimulated rate and addition of‘insulin increased this response to 131-+ 9% of the serum-stimulated rate, which was not significantly greater than the maximal response induced by insulin in the absence of FDGF. In the NIH 3T3/HIR cells, IGF-; was additive with suboptimal concentrations of insulin but did not augment the maximal response elicited by serum, as shown in Fig. 7. Combinations of insulin and IGF-1 did not detectably induce DNA synthesis in the parental cells (see Table 1). These data would be expected if the iigands were using the same receptor or postreceptor pathways. The effect. of combinations of two or more mitogens revealed additional differences between the NIH 3T3

36

RANDAZZO

AND

JARETT

TABLE Interaction Stimulating

2

of PDGF, IGF-1, EGF, and Insulin in DNA Synthesis in NIH 3T3/HIR Cells [3H]Thymidine incorporation (% of serum response) Insulin (500 rig/ml)

(n = 7)

Insulin (100 rig/ml) (n = 4)

64 -+ 9 16 + 3 3k3 108 + 9 78 AI 9 47 2 8 112 t6

109 t 5 146 -t 6* 115+ 7 114+ 4 147 * 10* 133 + 7 112k 5 11s* 4

92 f 100 -t 91 + 94 + 1012 99 k 94 + 89 +

No addition

Insulin

Concentration

hg/ml)

FIG. 7. The effect of IGF-1 on insulin-induced [3H]thymidine incorporation into NIH 3T3/HIR cells. One hundred nanograms per milliliter of IGF-1 and the indicated concentrations of insulin were added to confluent monolayers of cells in serum-free medium and [3H]thymidine incorporation was measured 17.5 h later as described under Materials and Methods. The data are from three experiments performed in the presence of IGF-1. The data in the absence of IGF-1 are from eight experiments and are redrawn from Ref. [14]. The data were analyzed by two-way anova. No positive interaction was identified.

and NIH 3T3/HIR cells. PDGF, IGF-1, EGF, and insulin were added at maximally effective concentrations in the combinations indicated in Tables 1 and 2, and DNA synthetic rate was determined. In the parental cells (Table l), IGF-1 was synergistic with EGF, PDGF, and EGF plus PDGF in stimulating DNA synthesis. A maximum response occurred in the presence of all three mitogens with a DNA synthesis rate that was 254 + 16%

TABLE

1

Interaction of PDGF, IGF-1, EGF, and Insulin in Stimulating DNA Synthesis in NIH 3T3 Cells [3H]Thymidine incorporation (% of serum response) No addition No addition PDGF IGF-1 EGF PDGF + IGF PDGF + EGF IGF + EGF PDGF + IGF + EGF

47 * 1% 32 +86k 200 f 139 f 254 -+

15 1 13 9 11* 8* 16*

Insulin

(100 rig/ml)

-2+64 k 2k 102 f 95+244 2 170 rt 254 -t

1 16 2 22* 6 14 12 29

Note. PDGF (100 rig/ml), IGF-1 (10 rig/ml), EGF (1 rig/ml), insulin, or the indicated combinations were added in serum-free medium to confluent monolayers of NIH 3T3 cells, and [3H]thymidine incorporation was determined 18 h later. The data are from three experiments performed in duplicate and are presented as the ,percentage of the response induced by 10% fetal calf serum (see footnote 2). The data were analyzed by four-way anova. * indicates significant positive interaction with P < 0.05 using the Bonferoni correction [18, 201.

No addition PDGF IGF-1 EGF PDGF + IGF PDGF + EGF IGF + EGF PDGF + IGF + EGF

(n = 3) 4 6 4 3 1 1 3 2

Note. PDGF (100 rig/ml), IGF-1 (10 rig/ml), EGF (1 rig/ml), insulin, or the indicated combinations were added in serum-free medium to confluent monolayers of NIH 3T3/HIR cells, and [3H]thymidine incorporation was determined 17.5 h later. The number of experiments, performed in duplicate, is indicated in parentheses. The data are presented as the percentage of the response induced by 10% serum (see footnote 2) and were analyzed by four-way anova. No significant positive interactions were identified. * indicates significantly greater than insulin alone as determined using a Bonferoni t-test [18, 201 with P < 0.05.

the serum-stimulated rate. Insulin showed the same trends of synergy with EGF and PDGF as did IGF-1. Insulin and IGF-1 were not synergistic. In contrast with the parental NIH 3T3 cells, in the NIH 3T3/HIR cells, as shown in Table 2, at a concentration of 500 rig/ml, insulin was not synergistic with any combinations of PDGF, EGF, or IGF-1 and at a concentration of 100 rig/ml of insulin was weakly additive with PDGF, but the addition of other mitogens did not augment this response. These data were corroborated by experiments examining the percentage of cells incorporating 13H]thymidine over 24 h. As shown in Table 3, IGF-1 and insulin alone did not affect the percentage of the parental NIH 3T3 cells incorporating thymidine but were synergistic with EGF and PDGF. A maximum of 95% of the NIH 3T3 parental cells incorporating label was observed when three mitogens were present. In contrast, insulin induced 56% of the NIH 3T3/HIR cells to incorporate thymidine and addition of other mitogens had no further effect. DISCUSSION These studies examined the growth of NIH 3T3/HIR cells in the presence of the mitogens PDGF, EGF, IGF1, and insulin. Previously, insulin in the absence of other mitogens had been shown to stimulate the growth of these cells to the same extent as did serum [ 141. Since

INSULIN

TABLE

AS A GROWTH

3

Effect of PDGF, I(;F-1, EGF, and Insulin on the Labeling Index of NIH 3T3 and NIH 3T3/HIR Cells [“H]Thymidine

incorporation

NIH XT:: No addition No addition F’DGF IGF-1 EGF PDCF t IGF PIXF + EGF IGF 7 EGF E’DGF + IGF ! JXP

7 44 il 16 70 89 Sl 92

NlH

Insulin .s.5 68 7.1 41 7.5 95 43 95

(5% of cells) BTG/HIR

No addition 2.6 26 10 3.1 35 25 12 41

Insulin S6 59 4.5 46 45 53 -2 2

N&e. I’DGF (100 rig/ml). IGF-1 (10 rig/ml), EGF (1 rig/ml), insulin (100 rig/ml), or the indicated combinations were added in serum-free medium to confluent monolayers of cells and the percentage of cells incorporating [“Hlthymidine during 24 h was determined by autoradiography. Serum induced 45% of the NJH OTB/HIR cells and 65x7 of the NIH ST3 cells to incorporate [“H]thymidinc. These data are from cne experiment.

a response of this magnitude was unexpected, these studies attempted to determine if the cells were atypically responsive to growth factors in general or if the response was specific for insulin. The NIH 3T3/HIR cells were found to synthesize DNA to a greater extent in response to insulin than in response to serum, PDGF, IGF-1, and EGF. These cells are not believed to be transformed independently of insulin receptor activity since t.hey appear capable of achieving density arrest 1141. In contrast, DNA synthesis in the parental cells was induced to the greatest extent by serum. The efficacy of single mitogens was of the relative order PDGF IS EGF > insulin = IGF-1 which is typical for 3T3 fibroblasts [8]. The interaction of PDGF, EGF, and IGF-1 with insulin was studied to obtain further evidence for the independence of insulin-stimulated DNA synthesis from other growth factors in NIH 3T3/HIR cells. In contrast to the NIH 3T3 parental cells in which PDGF, EGF, j;GF- 1, and insulin synergized in stimulating DNA synthesis, additional growth factors did not augment insui.in-stimulated DNA synthesis in the NIH 3T3/HIR cells. However, insulin’s lack of synergy with other mitogens is difficult to interpret. These data could be explained by at least two mechanisms which are not mutually exclusive. First, the cells may achieve their maximum growth rate in response to insulin. Second, insulin may inhibit other effector pathways. The maximum laheling index of 60%, the decreased binding of EGF relative to the parental cells, and the decreased responsiveness to EGF and combinations of EGF, PDGF, and IGF-1 relative to the parental cells suggest that growth

FACTOR,

II

37

of the NIH 3T3/HIR is attenuated, although whether this is the result of clonal variation, decreased EGF receptor expression, or increased insulin receptor expression cannot be established with the available data. Thus, whether the lack of synergy is due to independence of insulin from other growth factors or due to inhibitory interaction is not clear. Regardless of the mechanism, the magnitude of the effect of insulin in the absence of a second mitogen and the lack of synergy with other growth factors is atypical for a progression factor. The interactions of mitogens in st.imuiating DNA synthesis were also studied to begin addressing the mechanism by which insulin caused growth. Patterns of synergy can sometimes allow one to distinguish between mechanisms. For example, in the parental NIH 3T3 cells, insulin and IGF-1 were synergistic with EGF but insulin was not synergistic with IGF-1, suggesting that insulin and IGF-1 used similar signal transduction mechanisms which were distinct from the mechanisms used by EGF. In the NIH 3T3/HIR cells, EGF, which had no effect in the absence of other growth factors, increased the effect of l-10 rig/ml insulin on DNA synthesis. This synergy suggests that insulin, at concentrations of l-10 rig/ml, used different effecters than did EGF. Additivity with IGF-1 at submaximally stimulatory concentrations of insulin without augmentation of the maximum insulin response suggests that IGF-1 and insulin were using similar mechanisms. The lack of synergy of maximally stimulatory concentrations of insulin with other growth factors cannot be interpreted in such mechanistic terms and, specifically, the hypothesis that insulin receptors use pathways normally part of the signal transduction apparatus for other grcwth factors, thought to be a possible explanation fcr the insulin-induced growth of the NIH 3T3/HIR cells I141, cannot be ruled out. Insulin is thought to be exerting its effect,s by binding insulin receptor in the NIH 3T3/HIR cells. This conclusion is based on studies using monoclonal antibodies to the IGF-1 receptor and insulin receptor [13]. The data presented here provide additional evidence supporting this conclusion. IGF-1, in the absence or presence of a second mitogen, was found to stimulate DNA synthesis to a lesser extent and with less potency than did insulin in the NIH 3T3/HIR cells. If insulin were functioning through IGF-1 receptor, IGF-? would have been expected to stimulate DNA synthesis to as great an extent as insulin and with greater potency than insulin. These differences cannot be explained by ligand degradation since insulin is degraded two- to threefold faster than is IGF-1 by the NIH 3T3/HIR cells. The relative sensitivity of the KIH 3T3/HIR celis to IGF-1 and insulin suggests that both ligands induced DNA synthesis by binding the insulin receptor in the NIH 3T3/HIR ce!ls, although the possibility that IGF-1 may induce DNA

38

RANDAZZO

synthesis by binding a hybrid IGF-1 receptor-insulin receptor [21] has not been ruled out. Insulin’s effects, however, are not believed to be mediated by such a hybrid receptor for two reasons. First, if insulin and IGF-1 were binding the same receptor, which had similar affinities for both ligands, then the two ligands should have similar efficacies and potencies in stimulating DNA synthesis. Second, using autophosphorylation as a functional assay, IGF-1 but not insulin has been found to activate a hybrid receptor [21]. Binding studies were also consistent with insulin-receptor mediating insulinstimulated DNA synthesis in the NIH 3T3/HIR cells. The parental cells contained approximately twofold more IGF-1 receptors than did the NIH 3T3/HIR cells; hence the differential responses of the two cells are unlikely secondary to a difference in IGF-1 receptor expression. Also, insulin binding paralleled the insulin concentration dependence of DNA synthesis by the NIH 3T3/HIR cells, whereas the concentration of IGF1 required for a half-maximal effect was greater that the apparent Kd of IGF-1 binding. These data suggest that insulin stimulated DNA synthesis by binding the insulin receptor. In contrast, in the parental cells neither insulin nor IGF-1 stimulated DNA synthesis in the absence of a second mitogen and, in the presence of EGF, IGF-1 was one to two orders of magnitude more potent than insulin in stimulating DNA synthesis. These data suggest that both ligands were functioning through the IGF-1 receptor in the parental cells. Analysis of the insulin concentration dependence for [3H]thymidine incorporation in the NIH 3T3/HIR cells reveals that a half-maximal DNA synthesis response occurred at an insulin concentration of 35 rig/ml or 5.8 X lo-’ M. This concentration is greater than the Kd of the insulin-insulin receptor complex, for which values between 5 X lo-l1 M [22] and 1 X lo-’ M [15] have been reported; this is apparently inconsistent with an interpretation of the insulin response being mediated through the insulin receptor. However, two phenomena may explain the concentration dependence curve. First, insulin degradation can at least partly account for the discrepancy. At insulin concentration of 100 rig/ml or less, the NIH 3T3/HIR cells degrade more than 50% of the insulin by 8 h and, considering that the methods employed underestimate insulin degradation, likely deplete insulin within 24 h. Thus, the initial insulin concentrations overestimate the insulin concentrations to which the cells are exposed during the course of the experiments. The relative low sensitivity to insulin might also be related to the high number of receptors in the NIH 3T3/HIR cells. If, as previously suggested [ 141, the effect of insulin on growth is dependent on a high concentration of activated receptor molecules, then a high fraction of receptors must be occupied to elicit an effect. In the 17-mm well, one can calculate that there are approximately 2 X 10-l’ mol of receptor. If there were

AND

JARETT

100% binding of ligand to receptor, to occupy one-half of these receptors, one would need a minimum of 10-l’ mol of insulin or, for a volume of 1 ml, lo-’ M solution. This is a situation in which receptor concentration is equal to or greater than the Kd of the receptor-ligand complex and, consequently, the concentration of ligand required for occupancy of half of the receptors will be greater than the Kd [22]. As predicted by these considerations, studies of insulin binding to confluent monolayers of these cells yielded a Kd of 4 X lo-’ M [14]. In summary, in NIH 3T3/HIR cells, the rate of DNA synthesis induced by insulin was approximately equal to that induced by serum and was greater than that inducedby PDGF, IGF-1, and EGF. At maximally stimulatory concentrations of insulin, insulin did not synergize with either EGF, PDGF, or PDGF plus EGF. The maximum number of NIH 3T3/HIR cells synthesizing DNA was 60%. These results are in contrast to those obtained in the parental NIH 3T3 cells, in which insulin alone does not affect DNA synthesis; insulin synergizes with EGF, PDGF, and EGF plus PDGF and the combination of PDGF plus EGF plus insulin induces more than 90% of the cells to synthesize DNA. Hence, the behavior of insulin is atypical for a progression factor in the NIH 3T3/HIR cells. The effects of insulin may be produced by an unphysiologically high receptor concentrations or by the expression of a receptor in an inappropriate context, as has been suggested for other hormone receptors similarly expressed in NIH 3T3 cells [24-271. Current studies are underway to pursue further the growth characteristics of these cells and to address the mechanisms by which insulin stimulates growth. The authors thank Dr. Andrew Abler for extensive discussions and a critical evaluation of the data and manuscript and Dr. C. D. Scher for critical evaluation of the original manuscript. This research was supported by NIH Grants DK 28144 and DK 08484-01.

REFERENCES 1. Koontz, J. W., and Iwahashi, M. (1981) Science 211,947-949. 2. Hickman, J., and McElduff, A. (1989) Endocrinology 124, 701706. 3. Simonian, M. H., White, M. L., and Gill, G. N. (1982) Enokxridogy 111,919-922. 4. van Dijk, J. P., Tanswell, A. K., and Challis, J. R. G. (1986) J. Endocrin. 119,509-516. 5. Flier, J. S., Usher, P., and Moses, A. C. (1986) Proc. Natl. Acad. Sci. USA 83,664-668.

6. Conover, C. A., Hintz, R. L., and Rosenfeld,

R. G. (1985) J. Cell Physiol. 122, 133-141. 7. Corps, A. N., andBrown, K. D. (1988)Bicchem. J. 252,119-125. a. Rozengurt, E. (1986) Science 234,161-166. 9. Harrington, M. A., and Pledger, W. J. (1987) in Methods in Enzymology (Barnes, D., and Sirbasku, D. A., Eds.), Vol. 147, pp. 400-406, Academic Press, San Diego. 10. Stiles, C. D., Capone, G. T., Scher, C. D., Antoniades, H. N., van

INSULIN

11.

AS A GROWTH

Glanz, S. A. (1987) Primer McGraw-Hill, New York.

Whittaker, J., Okamato, A. K., Thys, R., Bell, G. I., Steiner, D. I., and Hofmann, C. A. (1987) Proc. NatE. Acad. Sci. USA 84,5237-

20.

Wilkinson, L. (1988) Systat: The System for Statistics: Inc., Evanston, IL.

5241.

21.

13.

Hofmann, C., Goldfine, I. D., and Whittaker, J. (1989) J. Biol. @hem. 264,8606-8611. Randazzo, I’. A., Morey, V. A., Polishook, A. K., and Jarett, L. Exp. Cell Res. 190, 25-30. Hammons, 6. T., and Jarett, L. (1980) Diabetes 29,475-486. Leatherbarrow, R. J. (1987) Enzfitter: A Non-linear Regression Data Analysis Program for the IBM-PC, Elsevier Science Publishers BV, Amsterdam. Baserga, R. (1989) in Cell Growth and Division, Baserga, R., Ed., pp. 6-11, IRL Press, New York.

17. 18.

39

19.

Hofmann, C., White, M. F., and Whittaker, ogy 124,701-706.

16. 16.

II

Wyk, J. J., and Pledger, W. J. (1979) Proc. Natl. Acad. Sci. USA 76, 1279-1283.

12.

14.

FACTOR,

Received December 27, 1989 Revised version received April

17, 1990

2nd ed., pp. 179-

C. P., Duronio,

2nd ed., pp. 30-96, Systat

V., and Jacobs, S. (4989) 9. Biol. Chem.

264,238-244.

J. (1989) Endocrinol-

Sokal, R. R., and Rohlf, F. J. (1981) Biometry, 207,321-371, W. H. Freeman, New York.

Moxham,

of Biostatistics,

22.

Cuatrecasas, P., and Hollenberg, Chem. 30,251-451.

23.

Chang, K. J., Jacobs, S., and Cuatrecasas, Biophy. Acta 406,294-303.

24.

Honegger, A., Dull, T. J., Szapary, D., Komoriya, A., Kris, R., Ullrich, A., and Schlessinger, J. (1988) EM30 J. ‘7, 3053-3060.

25.

Julius,

D., Livelle,

T. J., Jessell,

M. D. (1976) Adv. Protein P. (1975) Biochim.

T. M., and Axel,

R. (1989)

Scierzce244,1057-1062. 26.

Roussel, M. F., and Sherr, C. J. (1989) Proc. Natl. Acael. Sei. USA 86,7924-7927.

27.

Velu, T. J., Beguinot, L., Vass, W. C., Willingham, M. C., Merlino, 6. T., Pastan, I., and Lowy, D. .(1987)Science238,14081410.