WAYNE B. ANDERSON, THOMAS R. RUSSELL, RICHARD A. CARCHMAN*, AND IRA PASTAN. Laboratory of Molecular Biology, National Cancer Institute, ...
Proc. Nat. Acad. Sci. USA Vol. 70, No. 12, Part II, pp. 3802-3805, December 1973
Interrelationship Between Adenylate Cyclase Activity, Adenosine 3': 5' Cyclic Monophosphate Phosphodiesterase Activity, Adenosine 3': 5' Cyclic Monophosphate Levels, and Growth of Cells in Culture (contact inhibition of growth/NRK cells/CEF)
WAYNE B. ANDERSON, THOMAS R. RUSSELL, RICHARD A. CARCHMAN*, AND IRA PASTAN Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20014
Communicated by E. R. Stadtman, August 28, 1973 ABSTRACT To investigate how cell population density influences the intracellular concentration of cyclic AMP we have measured adenylate cyclase and cyclic AMP phosphodiesterase activities and cyclic AMP levels at various stages of cell density in normal rat-kidney (NRK) cells, which exhibit contact-inhibition of growth, and in normal chick-embryo fibroblasts (CEF), which do not show contact inhibition of growth under our conditions. Until NRK cells reach confluency, both activities increase with increasing cell population and cyclic AMP levels are low. As NRK cells reach confluency, cyclic AMP phosphodiesterase activity decreases somewhat whereas adenylate cyclase activity continues to rise. This increase in synthetic ability is accompanied by the increase in cyclic AMP levels which occurs in these cells at confluency. In CEF grown in 5% serum where density-dependent inhibition of growth is not observed, both adenylate cyclase and cyclic AMP phosphodiesterase activities increase proportionately with increasing cell population density. No significant alteration occurs in the ratio between these two enzyme activities and no change is observed in cyclic AMP levels. The NaF-stimulated activity in NRK cells increases with increasing cell density until the cells reach confluency; thereafter the NaF-stimulated activity remains constant. In contrast, the NaF-stimulated &ctivity observed in CEF does not vary appreciably between light and heavy density. The observed changes in the enzymes of cyclic AMP metabolism accurately reflect the changes in cyclic AMP concentration as a function of cell population density. The data indicate that these two enzyme activities respond to increasing cell density to elicit a rise in intracellular cyclic AMP levels. The elevated cyclic AMP levels are thought to be involved in the regulation of cellular growth rate and the mediation of contact inhibition of growth.
is arrested at confluency can be released from this contactinhibited state by the addition of fresh serum (6), insulin (7), or brief treatment with trypsin or other proteases (8, 9). Such treatments have been shown to result in rapidly decreased levels of cyclic AMP (5, 10, 11). The involvement of cyclic AMP in regulating DNA synthesis has been implicated in studies with 3T3 cells (11, 12). The level of cyclic AMP in the cell is determined in large measure by its rate of synthesis, catalyzed by the enzyme adenylate cyclase, and by its rate of hydrolysis, catalyzed by the enzyme cyclic AMP phosphodiesterase. Usually only small amounts of cyclic AMP are released into the medium. Thus it is of importance to know if these two enzyme activities are altered as a function of increasing cell density. Previous reports have indicated that adenylate cyclase activity is increased as cell density is increased in 3T3 and Chang liver cells (13) and in human diploid fibroblasts (14). Other studies have shown that cyclic AMP phosphodiesterase activity increases at confluency in L cells (15) and in 3T34 cells (16). Since both enzyme activities are involved in maintaining intracellular cyclic AMP levels, we have measured both activities simultaneously. To study the relationship between increasing cell population density and altered cyclic AMP metabolism, we have used normal rat-kidney (NRK) fibroblasts and normal chickembryo fibroblasts (CEF). One of these, NRK cells, exhibits density-dependent inhibition of growth. The other, CEF, does not display density-dependent inhibition of growth when maintained in 5% serum but grows to a high density with no decrease in growth rate. We find cyclic AMP levels to be elevated in NRK cells at confluency, and this Gates with an increase in adenylate cyclase activity in Us of a small rise noted in phosphodiesterase activity. In chick cells in which the adenylate cyclase and phosphodiesterase activities both rise in prallel with increasing cell density, the intracellular cyclic AMP levels remain constant.
Many lines of normal cells growing in tissue culture stop growing or grow at a greatly reduced rate once the cells reach confluency. This process is known as "contact inhibition" (1, 2) or "density-dependent inhibition" (3) of growth. Although the detailed mechanism by which the cell controls its rate of growth and DNA synthesis is unknown, an everincreasing body of evidence indicates that the intracellular concentration of cyclic AMP is involved in this regulation. If cyclic AMP is involved as a mediator of density-dependent inhibition of growth, the intracellular levels of this cyclic nucleotide should be elevated when normal cells reach confluency. Indeed, this has been found to be the case in the 3T34 and 3T3-42 lines of mouse fibroblasts and in MA 308 human diploid fibroblasts (4, 5). Normal fibroblasts whose growth
MATERIALS AND METHODS
CEF were prepared as described (17) and grown in Eagle's minimal essential medium supplemented with D-glucose (1 g per liter), sodium pyruvate (5 mM), 5% fetal bovine serum, 10% tryptose phosphate broth, penicillin (50 units per ml), streptomycin (50 ug per ml), and tylosine (50 ug per ml). Secondary chick fibroblasts were used in these studies and were grown at 390C in a humidified 5% CO2 atmosphere. The growth medium was changed every 24 hrs. CEF do not show density-dependent inhibition under these conditions. NRK cells were initially obtained from Dr. E. Scolnick
Abbreviations: CEF, chick-embryo fibroblasts; NRK fibroblasts, normal rat-kidney fibroblasts. * Fellow, National Cancer Institute, Public Health Service 1 F02 CA 54587-01.
3802
Proc. Nat. Acad. Sci. USA 70
(1973)
Cyclic AMP in Cultured Cells
and were grown in Dulbecco-Vogt modified Eagle's medium with 10% calf serum. NRK cells were grown at 370C in a 5% C02 humidified atmosphere with the growth medium changed every 48 hr. For the enzyme assays, the cells were washed and harvested, and crude homogenates were prepared as described (18). Adenylate cyclase activity was measured as described (18), with 5 mM MgCl2 and either 0.2 mM or 2.0 mM ATP as indicated. Cyclic nucleotide phosphodiesterase was assayed at low (0.12 or 0.5 AM) or high (18 EM) cyclic AMP by the method of Thompson and Appleman (19). For the determination of cyclic AMP, the medium was aspirated from the plates and 1 ml of cold 5% CC13COOH containing 1 X 104 cpm of 10.0 O
.5.0 A
I2.0,
35
/
30~
_
~~~~~~25~
0
1
~0.5
"
20
Smr~ 10-
0.2
5
0.1
0
1250 -
250 -
5000-
XF 124 200 -20
-
B
_ 750
.750 + -
500
250 -
cyclic ['H]AMP (21 Ci/mmole) was added. Cyclic AMP was measured by the method of Steiner et al. (20) as modified by D'Armiento et al. (21). Protein measurements were made by the method of Lowry et al. (22), with bovine-serum albumin as the standard. RESULTS
Influence of Cell Population Density on Cyclic AMP Levels, Adenylate Cyclase, and Cyclic AMP Phosphodiesterase in NRK Cells. NRK cells are a contact-inhibited line that grow logarithmically with a doubling time of about 15 hr until the cells approach confluency. This occurs after about 5 days when the cells have been seeded at 2,000 cells/cm2 (Fig. 1A). By day 7, the growth rate is declining and it continues to decline with increasing population density until growth ceases completely by day 17. We have elected to present our growth data as mg of protein per dish since the enzyme activities and cyclic AMP content are expressed per mg of protein. Similar curves are obtained whether cell growth is expressed as number of cells per dish, mg of nucleic acid per dish, or mg of protein per dish. It is evident that the cyclic AMP concentration remained low and relatively unchanged during the period of logarithmic growth (Fig. 1A). On day 7, when the growth rate is declining,
24
10 r.
2
-
en
I6
o50
U_~~200 .
3803
D
12 Cl)
4
0
-o
0
0
-4
M Cr
0
0
0rn
2500 .500
C
60
2
'~50
L~2000- uj400 ~~I500 ~~~300
d3
1I000. 22200 -0 20 10
500 -100 0
0
2 34 DAYS AFTER PLANTING
P.
n
N 0
-o-o
0
FIR. 1. The cyclic AMP concentration, adenylate cyclase activity, and cyclic AMP phosphodiesterase activity in NRK cells as a function of increasing cell population density. NRK cells were planted in 100-mm petri dishes at 1 X 106 cells per dish and were grown at 370C. Media were changed every 48 hr. At the indicated times, cells were harvested and prepared for assay. (A) The increase in cell population density is indicated by the change in total mg of protein per dish (0--- 0). Cyclic AMP content (U--) is given as pmoles of cAMP per mg of protein. (B) Adenylate cyclase activity was determined with low substrate concentration (0.2 mM ATP) in either the presence (0-0) or absence (0-) of 10 mM NaF. Cyclic AMP phosphodiesterase activity (A-- -A) was assayed with 0.12 jsM cAMP as the low substrate concentration. (C) Adenylate cyclase activity was measured with high ATP concentration (2.0 mM) in either the presence (0-a) or absence (0-0) of 10 mM NaF. Cyclic AMP phosphodiesterase activity (A- - -A) was determined with 18 ,uM cAMP as substrate. Adenylate cyclase activity is expressed as pmoles of cAMP formed per 10 min per mg of protein. Cyclic AMP phosphodiesterase activity is expressed as pmoles of cAMP hydrolyzed per min per mg of protein.
0 0 cn
C/) --4 -4
5 10 15 20 DAYS AFTER PLANTING
FIG. 2. Cyclic AMP concentration, adenylate cyclase activity, and cyclic AMP phosphodiesterase activity in CEF as a function of increasing cell population density. CEF were planted in 100mm petri dishes and grown at 390C. The media were changed every 24 hr. At the indicated times, cells were harvested and prepared for assay and cyclic AMP levels and enzyme activities were measured. The legends for parts A, B, and C are identical to those given in the legend to Fig. 1 except that the cyclic AMP phosphodiesterase activity in B was assayed at 0.5 ,&M cAMP.
3804
Biochemistry: Anderson et al.
Proc. Nat. Acad. Sci. USA 70 0
_ zO 0 X -< I r,
-
20
0
Scn
(1973)
rIn
mn
)k
10
A/
//
A.a_
(A Co
mn m
-_
4.0-0. w) w
-J
Ia
A-
0-
5
_I
a
.
1-
0
m
_%m! 2
3
4
5
0
2
3
4
5
MG PROTEIN/DISH
FIG. 3. Correlation of cyclic AMP levels (A---A) with the ratio of adenylate cyclase to cyclic AMP phosphodiesterase activities (@-@) measured in NRK (A) and CEF (B) cells. Data were taken from Figs. 1 and 2 and plotted as a ratio of the two enzyme activities. Adenylate cyclase activity was measured with 2.0 mM ATP and cyclic AMP phosphodiesterase activity was determined with 18 AM cAMP. Ratio = pmoles of cyclic AMP formed per 10 min per mg of protein/pmoles of cyclic AMP hydrolyzed per min per mg of protein.
the cyclic AMP level is rising and it continues to increase over the next several days. This increase coincides quite well with the decreased growth rate and the eventual stationary cell population. The cyclic AMP content was determined in a growth study separate from the one presented in Fig. 1A. The cells used to measure cyclic AMP levels grew somewhat more rapidly and reached a stationary population density 12 days after planting. As a result, when the data are plotted against days after planting, the cyclic AMP concentrations determined at high density are shifted slightly to the left relative to the growth curve in Fig. 1A. However, excellent correlation is observed if the data are presented as a function of mg of protein per dish, as is done in Fig. 3A. In our study of adenylate cyclase and phosphodiesterase we measured their activities at both low and high substrate concentrations, to determine if any change in activity observed might be due to a change in affinity for substrate or, in addition, in the case of the phosphodiesterase, if one of the two forms of the enzyme might be selectively altered as a function of increasing cell density (23-26). The adenylate cyclase from NRK cells measured at 1.2 and 4 mg of protein per dish has an apparent KmATP -0.25 mM (M. Gallo, unpublished data). Thus, we assayed the enzyme at a substrate concentration similar to this Km value (0.2 mM ATP) and ten times this value (2.0 mM ATP). Cyclic nucleotide phosphodiesterase was assayed at 0.12 and 18 MAM cyclic AMP. Basal adenylate cyclase activity showed a progressive increase with increasing cell population through day 17 which resulted in a 3- to 4-fold increase in enzyme activity (Fig. 1B and C). Cyclic AMP phosphodiesterase activity also increased With increasing cell population, but only through day 6 (Fig. 1B and C). On day 7 the activity declined somewhat and then leveled off and remained constant through day 17. Similar activity profiles were observed whether the enzymes were measured at low or high substrate concentrations. Of particular significance is the finding that as the two enzyme activities began to diverge at day 7, cyclic AMP levels began to rise and cell growth rate to decline.
Influence of Cell Population Density on Cyclic AMP Levels, Adenylate Cyclase, and Cyclic AMP Phosphodiesterase in CEF. The growth of CEF cultured in medium with 5%o fetal calf
is markedly different from that of NRK cells. CEF continued in logarithmic growth to high density and thus showed no evidence for density-dependent inhibition of growth (Fig. 2A). In these cells we found that the cyclic AMP content did not change appreciably from sparse to heavy cell density (Fig. 2A). For good growth, the CEF were planted at about 100,000 cells/cm2 and, therefore, their growth could only be studied over a few days. The adenylate cyclase of CEF has an apparent Km ATP 0.23 mM (18), which is similar to that of NRK cells, and was assayed at 0.2 mM and 2.0 mM ATP. Russell and Pastan (27) have shown that the cyclic nucleotide phosphodiesterase of CEF is found in both particulate and soluble fractions. Different catalytic forms of this enzyme exist which exhibit apparent Km values of 0.5 AM and 5 MM for cyclic AMP. In these studies we have used a cyclic AMP concentration of 0.5 MM to measure the contribution of the low Km form of the enzyme and 18 MM cyclic AMP to determine total phosphodiesterase activity. We observed about a 2-fold increase in basal adenylate cyclase activity with increasing cell density between days 1 and 3 (Fig. 2B and C). Cyclic AMP phosphodiesterase activity also increased in activity between days 1 and 3. The increase in activity of these two enzymes followed a similar time course, and apparently the increases counteracted one another so that no change in cyclic AMP concentration occurred. The decrease in enzyme activities and cyclic AMP levels at day 4 is probably a reflection of an unhealthy cell population at this high density, for the cells were becoming vacuolated and beginning to die at this time. serum
Interrelationship Between the Ratio of Adenylate Cyclase to Cyclic AMP Phosphodiesterase and Cyclic AMP Levels. To illustrate more clearly the close correlation between adenylate cyclase activity, cyclic AMP phosphodiesterase activity, and cyclic AMP content as a function of cell population density, we have plotted the ratio of the two enzyme activities along with cyclic AMP levels in Fig. 3A and B. In NRK cells the enzyme ratio showed little change until the cells reach confluency at about 2 mg of protein per dish (Fig. OA). At confluency, the adenylate cyclase activity in NRK cells continued to increase, whereas the phosphodiesterase activity
Proc. Nat. Acad. Sci. USA 70
(1973)
did not, and this was reflected in the increased ratio of enzyme activities as cell growth slowed. The cyclic AMP levels mirrorthis increase in adenylate cyclase activity relative to phosphodiesterase activity. The CEF, which do not show density-dependent inhibition of growth under these conditions, showed very little change in enzyme ratio at confluency (Fig. 3B), and similarity the cyclic AMP levels were relatively unchanged with increasing cell protein per dish. NaF-Activated Adenylate Cyclae. We also measured adenylate cyclase in the presence of the activator, NaF. In NRK cells the NaF-stimulated activity increased over the period from day 3 to day 7, but then remained constant as the cells approached the stationary phase of the growth curve. On the other hand, the specific activity of the NaF-activated enzyme from CEF did not change throughout the course of the study, which is similar to results previously reported for 3T3 and Chang liver cells (13). DISCUSSION The mechanism by which normal cells regulate their rate of growth is very complex. One approach to understanding growth regulation is to identify substances that regulate cell growth. One of these appears to be cyclic AMP. Previous studies in our laboratory have indicated that cyclic AMP levels rise as contact-inhibited cells approach confluency (4, 5). Our present results with NRK cells substantiate this. In CEF, which do not exhibit density-dependent inhibition of growth when grown in 5% serum, cyclic AMP levels fail to rise even when the CEF grow to high cell density. Basal adenylate cyclase activity is low in rapidly growing cells and is enhanced as the population density increases in both NRK cells and CEF. In both cell types it reaches the same ultimate specific activity (Fig. 1 and 2). One apparent difference in the two cell types is the behavior of phosphodiesterase activity, which rises steadily in the CEF but fails to continue to rise in NRK cells. We believe this dissociation in activities accounts for the rise in cyclic AMP in NRK cells. Another difference between NRK cells and CEF is the behavior of the adenylate cyclase activity measured in the presence of F-. The NaF-stimulated cyclase activity in NRK cells increases with increasing cell density until the cells reach confluency; thereafter this activity remains constant. In contrast, the NaF-stimulated adenylate cyclase in CEF does not vary appreciably between cells of light and heavy density. How are the levels of adenylate cyclase and phosphodiesterase regulated in cells? Cyclic AMP itself is known to be an inducer of cyclic AMP phosphodiesterase (16, 28). It is conceivable that at confluency in NRK cells the phosphodiesterase is either fully induced or other control functions take over to limit its activity while cyclase activity continues to increase. Increased synthesis of the catalytic portion of adenylate cyclase may account for the increase in both basal and NaF-stimulated activity during the rapid growth phase of NRK cells. However, it seems unlikely that synthesis of the catalytic portion of adenylate cyclase is involved in controlling adenylate cyclase activity in either confluent NRK cells or in CEF at any stage in their growth since basal activity increases while NaF-stimulated activity, which is thought to represent total potential activity, is relatively unchanged. Perhaps the synthesis or degradation of specific
Cyclic AMP in Cultured Cells
3805
activators or inhibitors of these two enzymes, or the enzymes' ability to respond to such factors, is altered with increasing cell-to-cell contact. Whatever the mechanisms that control the synthesis and activity of these enzymes, the results, as presented in Fig. 3, clearly illustrate that the observed changes in these enzyme activities accurately reflect the alterations that occur in intracellular cyclic AMP concentration. Of particular significance is the correlation between the net increase in the ability to synthesize cyclic AMP (divergence of the two enzyme activities) with the increased cyclic AMP levels and decreased growth rate of NRK cells at confluency. In contrast, transformation of CEF by the Schmidt-Ruppin strain of Rous sarcoma virus and of NRK cells by the Kirsten strain of murine sarcoma virus is accompanied by a failure of adenylate cyclase to rise in growing cells (unpublished data). Such studies give us a clue as to the means by which normal and abnormal cells regulate their growth. We thank Dr. G. S. Johnson for helpful discussions, and acknowledge the excellent technical assistance of Elizabeth Lovelace. 1. Abercrombie, M. (1962) Cold Spring Harbor Symp. Quant. Biol. 27, 427-431. 2. Golde, A. (1962) Virology 16, 9-20. 3. Stoker, M. G. P. & Rubin, H. (1967) Nature 215, 171-172. 4. Otten, J., Johnson, G. S. & Pastan, I. (1971) Biochem. Biophys. Res. Commun. 44, 1192-1198. 5. Otten, J., Johnson, G. S. & Pastan, I. (1972) J. Biol. Chem. 247, 7082-7087. 6. Todaro, G. J., Lazar, G. K. & Green, H. (1965) J. Cell. Comp. Physiol. 66, 325-333. 7. Temin H. (1967) J. Cell. Physiol. 69, 377-384. 8. Burger, M. M. (1970) Nature 227, 170-171. 9. Sefton, B. M. & Rubin, H. (1970) Nature 227, 843-845. 10. Sheppard, J. R. (1972) Nature New Biol. 236, 14-16. 11. Burger, M. M., Bombik, B. M., Breckenridge, B. M. & Sheppard, J. R. (1972) Nature New Biol. 239, 161-163. 12. Willingham, M. C., Johnson, G. S. & Pastan, I. (1972) Biochem. Biophys. Res. Commun. 48, 743-748. 13. Makman, M. H. (1971) Proc. Nat. Acad. Sci. USA 68,
2127-2130. 14. Zacchello, F., Benson, P. F., Giannelli, F. & McGuire, M. (1972) Biochem. J. 126, 27P. 15. Heidrick, M. L., & Ryan, W. L. (1971) Cancer Res. 31, 1313-1315. 16. D'Armiento, M., Johnson, G. S. & Pastan, I. (1972) Proc. Nat. Acad. Sci. USA 69, 459-462. 17. Nakata, Y. & Bader, J. P. (1968) Virology 36, 401-410. 18. Anderson, W. B., Johnson, G. S. & Pastan, I. (1973) Proc. Nat. Acad. Sci. USA 70, 1055-1059. 19. Thompson, W. J. & Appleman, M. M. (1971) Biochemistry 10, 311-316. 20. Steiner, A. L., Parker, L. W. & Kipnis, D. (1972) J. Biol. Chem. 247, 1106-1113. 21. D'Armiento, M., Johnson, G. S. & Pastan, I. (1973) Nature New Biol. 242, 78-80. 22. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, 265-275. 23. Rosen, 0. (1970) Arch. Biochem. Biophys. 137, 435-441. 24. Beavo, J., Hardman, J. G. & Sutherland, E. W. (1970) J. Biol. Chem. 245, 5649-5655. 25. Thompson, W. J. & Appleman, M. M. (1971) J. Biol. Chem. 246, 3145-3150. 26. Russell, T. R., Terasaki, W. & Appleman, M. M. (1973) J. Biol. Chem. 248, 1334-1340. 27. Russell, T. R. & Pastan, I. (1973) J. Biol. Chem. 248, 58355840. 28. Manganiello, V. & Vaughan, M. (1972) Proc. Nat. Acad.
Sci. CTSA 699' 269-273.
