PEERY, C. V., JOHNSON, G. S., AND PASTAN, I. (1971) J. Biol. Chem. 246, i785-5790. TEMIN, H. M. (1967) in Growth Regulating. Substances for Ani- mal Cells ...
Vol.
THE JOURNAL cm B~o~oorcar. CHEMISTRY 247, No. 21, Issue of November 10, PP. 7082-7087, Printed in U.S.A.
Regulation Adenosine
1972
of Cell Growth by Cyclic 3’, S-Monophosphate
EFFECT OF CELL DENSITY AND AGENTS WHICH 3’) 5’-MONOPHOSPHATE LEVELS IN FIBROBLASTS*
ALTER
CELL
GROWTH
ON CYCLIC
(Received for publication, JACQUES
OTTEN,~
GEORGE
S.
JOHNSON,~
AND
IRA
SUMMARY
Several lines of evidence suggest that cyclic adenosine 3’,5’monophosphate is involved in the regulation of growth and other properties of fibroblasts. Cyclic AMP’ levels are 3- to 4-fold higher in confluent 3T3-4 cells than in logarithmically growing cells, under conditions where the medium was frequently changed to prevent depletion of serum factors (2). The cyclic AMP * Portions of this manuscript have been published form (1). $ Recipient of National Institutes of Health 1 F05TW
in abstract Fellowship
1714-02.
3 Leukemia Society of America, Special Fellow. 1 The abbreviations used are: cyclic AMP, cyclic 3’,5’-monophosphate; (but)zcAMP, N62’-0-dibutyryl 3’,5’-monophosphate; PGE,, prostaglandin El.
adenosine adenosine
May 11, 1972)
PASTAN
From the Laboratory of Molecular Biology, National Cancer Institute, National Bethesda, Maryland 20014
Growth stops at confluency in mouse 3T3 cells and in MA308 human diploid fibroblasts even with frequent changes of media. Accompanying this cessation of growth is an increase in the endogenous levels of cyclic adenosine 3’,5’ monophosphate (cyclic AMP). In WI38 human diploid fibroblasts cyclic AMP levels do not rise as the cells grow to high densities with frequent media renewals: however, if the medium is not replenished, cyclic AMP levels rise as growth slows, presumably due to depletion of some serum factor. Initiation of DNA synthesis in confluent 3T3 cells by the addition of fresh serum is preceded by a decrease in cyclic AMP levels. The cyclic AMP levels also decrease following addition of trypsin or insulin, agents which stimulate DNA synthesis and cell growth. Prostaglandin El (PGE1) which inhibits cell growth in L-929 cells, raises cyclic AMP levels. Insulin alone has no effect, but when added with PGEi it will partially prevent the elevation of cyclic AMP levels and corresponding inhibition of growth. The results give further support for the proposal that cyclic AMP regulates cell growth and mediates contact inhibition of growth.
ADENOSINE
Institutes of HeaJth,
levels in slowly growing cells are higher than in more rapidly growing cells (2, 3), and addition of (but)icAMP (4) or PGEr (5), which activates adenylate cyclase in cell extracts (6), decreases the growth rate. To understand better the relationship between cyclic AMP and growth we have measured cyclic AMP levels in cells treated with various substances known to alter the growth of fibroblasts. We find that addition of fresh serum or trypsin, which stimulate DNA synthesis and cell growth (7-Q), rapidly lowers cyclic AMP levels, whereas PGEi, which inhibits growth (5), elevates cyclic AMP levels. Two contact inhibited cells lines, in addition to the 31’3-4 cells, increase their cyclic AMP levels at confluency. These results give further evidence that cyclic AMP regulates logarithmic growth and mediates contact inhibition of growth in fibroblasts. MATERIALS
AND
METHODS
Cell lines and growth media are listed in Table I. Penicillin (50 pg per ml) and streptomycin (50 units per ml) were added to all media and glutamine (2 mM) was added to the Eagles minimal essential medium. The cells were grown in 50-cm2 plastic tissue culture dishes with 10 ml of growth media in a 5% COZhumidified atmosphere at 37”. Unless stated otherwise, the medium was changed 24 hours after planting and then every other day, even after confluency was reached. Cyclic AMP was extracted 24 hours after the previous change of medium unless otherwise indicated. Approximately 10’ cells were needed for each complete cyclic AMP assay, therefore the number of culture dishes used ranged from three for the determinat.ions in heavier cultures to up to 15 dishes for the lighter cultures. The pH of all solutions was measured at 25”. PGE, was a generous gift from J. Pike, Upjohn Co., Kalamazoo, Mich. Insulin (glucagon-free) was obtained from Pierre Freychet, National Institutes of Health. Beef heart cyclic AMP phosphodiesterase was purchased from Boehringer-Mannheim. Cyclic AMP Extraction and Purification-In a variety of cells the intracellular cyclic AMP levels vary rapidly in response to changing conditions in the extracellular medium. In hormoneresponsive tissues, measurable changes in cyclic AMP concentration may be observed within a period of less than a minute. In the brain, asphyxia increases cyclic AMP levels in some sec-
7082
7083 TABLE I Cell lines and growth media Cell
line
Growth
medium
Serum
Mouse 3T3-4
H. Green
Dulbecco’s
Mouse Mouse Mouse Mouse
G. Todaro Flow Laboratories0 G. Todaro S. Baron
Dulbecco’s modified Eagle’s medium Dulbecco’s modified Eagle’s medium Dulbecco’s modified Eagle’s medium Eagle’s minimal essential medium
Microbiological Assoc. Microbiological Assoc. NIH tissue culture section
Eagle’s minimal essential medium Eagle’s minimal essential medium Diploid growth medium
32’3-42 3T3SV40 ClX Balb 3T3 ClA31 L-929
Mouse embryo fibroblasts Human MA-308b Human WI-3%
0 3T3-SV40 were originally obtained from G. Todaro. b MA-308 is an established human diploid cell line originally c Human WI-38 were studied at their 20 to 22 passages.
modified Eagle’s medium
10% calf serum (Flow Laboratories) 10% calf serum 10% calf serum 10% calf serum 10% fetal calf serum (Flow Laboratories) 10% fetal calf serum 10% fetal calf serum 10% fetal calf serum
derived from a nasal polyp.
onds (10). For this reason, care was taken to extract cyclic AMP from dishes which had been kept in steady condition with respect to temperature and pH and without any rinsing of the cells between the removal of the medium and the addition of trichloroacetic acid. Culture medium was removed by aspiration and the cell layer immediately covered with 1 ml of icecold 5% (w/v) trichloroacetic acid. The dish was placed on ice, the cells removed with a rubber policeman, and the dish rinsed once with 0.5 ml 5% trichloroacetic acid. When an extract was prepared from two or more dishes, the trichloroacetic acid rinse was stepwise transferred from one dish to the next; 0.25 pmole of cyclic [3H]AMP (25 Ci per mM) was added to the extract to measure the recovery of cyclic AMP through the purification procedure. The suspension was centrifuged for 10 min at 10,000 x g and the supernatant was removed. After the addition of 0.2 ml of 1 N HCl, the supernatant was extracted five times with 3 volumes of water-saturated ether and lyophilized. The dry residue was dissolved in 0.3 ml of 0.086 M Tris-HCl pH 8.5, containing 4 mM MgClz and 150 pg per ml of bovine serum albumin. Purification of the cyclic AMP was necessary because the culture medium contains factors which behave like cyclic AMP in the binding assay. In addition one-half of each sample was treated with cyclic AMP phosphodiesterase before assaying to show that the material being measured was cyclic AMP. Thus, half of the sample was treated with 10 pg of phosphodiesterase for 4 hours at 30”. The enzyme had been dialyzed for 3 hours against Tris-HCl (0.086 M, pH 8.5) before use and stored in frozen aliquots. In this condition, 95 to 100% of the cyclic AMP is destroyed as determined by the BaS04 precipitation method of Krishna (11). After heating at 100” for 5 min, both phosphodiesterase-treated and untreated were treated with 1 ml 0.1 N HCl and passed over a column (0.7 x 10 cm) of Dowex 5OW-X8, 200 to 400 mesh in the H+ form. The column was rinsed with 6 ml of 0.1 N HCl. Cyclic AMP was eluted between 12 and 17 ml of HzO; recovery of cyclic taH]AMP varied from 75 to 100%. The eluate was lyophilixed and dissolved in 0.05 M sodium acetate, pH 4.0. Cyclic AMP Assay-Cyclic AMP was measured by the previously described modification of Gilman’s method (2, 10). In this assay unlabeled and cyclic [aH]AMP compete for binding to beef muscle protein kinase. It was found that the Dowex resin contains small amounts of substances which are eluted in the same faction as cyclic AMP and are indistinguishable from cyclic AMP in the binding assay. Therefore, in each assay, a
blank eluate from the same batch of resin was used as a control to determine the interference due to the resin. This interference varied with each batch of Dowex, but was usually equivalent to 10 pmoles of cyclic AMP per total eluate. In each experiment the phosphodiesterase-treated and nontreated parts of the sample were assayed in parallel and the difference was taken as the content of cyclic AMP. The data is expressed as micromoles per liter of cell volume. The number of cells was determined from two or more separate dishes grown in parallel to those used for the cyclic AMP assay. The mean cell volume was obtained from the packed cell volume of a known number of cells. The packed cell volume was measured in a microhematocrit centrifuge (Drummond Scientific Co., Brooman, Penna.), and the cell number was measured in a Coulter counter. DNA Synthesis-DNA synthesis was evaluated in 3T3-4 cells grown in 20.cm2 plastic dishes to confluency, kept for 4 days in 4 ml of media without media change, and then treated by the addition of 0.8 ml of calf serum. At various times after the addition of serum, 0.01 PCi of [aH]thymidine was added. One hour later the medium was removed, the cells were rinsed three times with a cold solution of 0.15 M NaCl, 0.01 M sodium phosphate, pH 7.4, and treated with 10% ice-cold trichloroacetic acid, the cells were removed with a clean rubber stopper and centrifuged for 10 min at 3,000 X g. The pellet was washed with 10% ice-cold trichloroacetic acid and recentrifuged. The precipitate was dissolved in 0.5 ml of Nuclear Chicago Solubilizer and counted in 10 ml of Liquifluor-toluene mixture in a liquid scintillation counter. RESULTS
Contact Inhibition of Growth-Mouse 31’3-4 cells grown logarithmically, usually to a density of 50,000 per cm2 where growth stops (Fig. l), even with frequent medium changes. During logarithmic growth the cyclic AMP concentration remains low, increases when the cells approach confluency, and rises to even higher levels several days after the cells have stopped dividing. This increase in cyclic AMP concentration at confluency has also been found in 31’3-42 cells, a subclone of 31’3-4 cells, and in MA308 cells (Table II). In cell lines not displaying contact inhibition of growth, cyclic AMP is highest during the early logarithmic phase of the growth and falls when a high density of cells is reached (2). WI 38 human diploid fibroblasts (12) grow to a higher cell density than 3T3 cells (Table III). When the media is changed
7084 I
1
50
I
I
I
I
7
300
6.0
I
5.0
I (D ‘g x
4.0
;4 0
3.0
200
z
150
2 # a
100
s:
SO
E F ::
60
9
4
I
3 I
2 3
4
5
20
II Cyclic AMP concentration in contact-inhibited Jibroblasts 31‘3-4 and 31’3-42 cells were planted as described in the legend for Fig. 1. MA 308 cells were planted at a density of 2 X 104 cells per cm2. The cyclic AMP concentration during logarithmic growth was determined after 3 days growth and at a cell density of 3.5 X lo4 cells per cm2. The confluent point was determined after 6 days growth at a density of 6 X lo4 cells per cm2. Media in all experiments were changed every 48 hours and 24 hours before assaying. TABLE
Logarithmigpowth:
cyclic
x 10-s M
31’3-4 31’3-42 MA 308 a Mean f
2.7 f 2.8 f 0.8 f
0.3 0.4 0.2
I
x 108 )W cma
21 132 227
0.7 0.7 0.3
Cyclic AMP x 10-e M
5 12 16
I IO DAYS
, 15 AFTER
I 20 PLANTING
I 25
I 30
FIG. 2. Cyclic AMP concentration in starved WI38 cells. Cells were planted at a density of 2 X lo4 cells per cme and the medium was changed only once, 24 hours after planting.
m
x 10-6 M
11.0 f 6.4 f 2.6 f
TABLE III Cyclic AMP concentration in WI 58 cells fed regularly The cells were planted at a density of 2 X lo4 cells per cm2. The medium was changed every 48 hours and 24 hours before cyclic AMP extraction. Days after planting
I 5
I
Contact-inhibition of growth: cyclic AMP
S.E.
Cell density
$ a P N
2.0
6 7 8 9 IO II 12 13 14 15 I6 17 I6 DAYS AFTER PLANTING
FIG. 1. The cyclic AMP content of 3T3-4 cells. Cells were planted at a density of 2 X lo3 cells per cm2. The medium was changed every 48 hours and 24 hours before the cyclic AMP analysis. Each point on the growth curve is an average of the cell count from two dishes.
Cell line
40
0.85 f 1.03 f 0.90 f
0.06 0.04 0.06
every other day, the cyclic AMP concentration remains constant. However, when the medium is not replenished the cyclic AMP levels begin to rise as the growth decreases and after 30 days reaches a value about six times higher than that during logarithmic growth (Fig. 2). Effect of Serum on Contact-inhibited ST3 Cells -In confluent 3T3 cells, the amount of DNA synthesis and the percentage of cells in mitosis are negligible. The addition of fresh serum induces, after a lag period of about 14 hours, an abrupt increase in DNA synthesis (Fig. 3). In three experiments, of which that
IO
30M MINUTES
3
5
10 15 20 HOURS TIME AFTER SERUM ADDITION
0
25
3. Cyclic AMP concentration and DNA 3T3-4 cells treated with serum. Cells density of 2 X lo3 cells per cm%and the medium 48 hours until confluency was reached. They for 4 days before the addition of 2 ml of fresh FIG. fluent
30
synthesis in conwere planted at a was changed every were then not fed serum.
illustrated in Fig. 3 is an example, the addition of serum is followed by a rapid decrease in cyclic AMP concentration. Very low values are reached at about 30 min. The cyclic AMP levels then increase towards their initial level. (These experiments were done with a different isolate of 31‘3-4 cells than that used in Fig. 1. The reasons for the difference in initial cyclic AMP levels are unknown.) A similar effect of serum also has been observed with 3T3-42 cells. After numerous passages one isolate of 31‘3-4 cells changed its responsiveness toward serum stimulation. At this time the cells grew in the presence of 10% serum to a saturation density of 50,000 cells per cm2 but, after confluency was reached and the medium allowed to be depleted, the addition of fresh serum did not induce further DNA synthesis (Table IV). This refractiveness to serum stimulation was checked in three successive experiments. In this same cell line, the addition of serum did not change the cyclic AMP concentration (Table IV). E$ect of Trypsin and Insulin-The addition of trypsin (70 pg per ml) (Table V) to confluent 3T3 cells also results in a decrease in cyclic AMP levels and the kinetics of this decrease is comparable to that induced by serum (Fig. 3). Similar results are
7085
Efect
of serum
TABLE IV to con$uent ST3 cells refractory
addition
12
to serum
r
stimulation
Cells were planted and grown as described in the legend for Fig. 3. DNA synthesis was determined as described under “Materials and Methods.” Time after serum addition
Cyclic AMP
..
Time after serum addition
[aH]Thymidine incorporation
x 10-s M 0 1 10 30 1 3 6 12
min min min hour hours hours hours
-
0.84 0.72 0.75 0.77 0.81 0.95 1.84 1.31
f f f f f f f A
CM 360
12 hours 15 hours 18 hours 21 hours 23 hours 27 hours 29 hours 31 hours
0.03 0.05 0.02 0.05 0.02 0.04 0.06 0.04
290 380 570 825 620
810
TABLE V Effect
of trypsin
on cyclic
AMP
concentration
j
480
in conjluent
ST3 cells
Cells were planted and grown as described in the legend for Fig. 3. Then trypsin (70 ~8 per ml) was added and cyclic AMP levels were measured.
30 MINUTE:
FIG. 4. Effect of PGEi and insulin on the cyclic AMP levels in L-929 cells. Cells were planted at a density of 5 x 163 cells per cm2. Medium was changed every 48 hours and 24 hours before the assays. After 5 days growth PGEr (10 fig per ml) or insulin (125 milliunits per ml), or both, was added and cyclic AMP was assayed at the indicated times. Ethanol (0.2% (v/v) final concentration) was added to the control medium.
-
Cyclic AMP Time 3T3-4
I x 10-s M
min 0 10 30 60
1.73 0.5
f f
0.04 0.02
1.75
f
0.06
5.2 0.55 3.2 4.1
3T3-42
f f f f
0.1 0.02 0.2 0.1
obtained with 3T3-4 and 3T3-42 cells. Under identical conditions insulin (125 milliunits per ml) decreases the concentration of cyclic AMP in 3T3-4 cells from 1.73 f 0.04 to 1.00 f 0.03 PM in 10 min and in 3T3-42 cells from 5.2 f 0.1 to 3.5 f 0.1 PM in 30 min. E$ect of PGEl and Insulin on L-929 Cells-PGEi inhibits the growth of L-929 cells (5). Control cells have a doubling time of 20 hours. In the presence of PGEi (10 pg per ml) the doubling time is increased to 28 hours. Insulin (125 milliunits per ml) has no effect by itself but does partially block the effect of PGEi; L-929 cells in the presence of both compounds have a doubling time of 23 hours. Correlating with the growth data, PGEi greatly increases the levels of cyclic AMP; insulin by itself has na effect but does block partially the elevation of cyclic AMP levels by PGEi (Fig. 4). The concentration of cyclic AMP increases with increasing concentrations of PGEi (Table VI). Increased levels at 30 min are observed with concentrations of PGEi as low as 1 pg per ml, however the elevated levels are not sustained for more than 6 hours. With 50 pg per ml of PGEi, increased levels are observed for at least 24 hours (Table VII). Recently Manganiello and Vaughn reported that PGEi raised cyclic AMP levels in nongrowing fibroblasts (13). Effect of PGEl on Unresponsive Cell Lines-Most lines of fibroblasts contain an adenylate cyclase activated by PGEi (6). However, 3T3SV40 Clone X (6) and Balb 3T3 CIA31 cells recently obtained from G. Todaro have an adenylate cyclase which is not activated by PGE1.2 (Another Balb 3T3 line pre2 C. Perry, unpublished
data.
TABLE VI PGEI on cyclic AMP concentration in L-929 cells Cells were planted and grown as described in the legend for Fig. 4. Cyclic AMP was assayed 30 min after the addition of PGEi. Effect
of
PGEI concentration
Cyclic AMP concentration
m-/ml 0 1 5 20 50
x
10-s M
2.50 5.31 7.10 16.50 18.30
f f f f f
0.12 0.24 0.30 0.41 0.50
times
after
TABLE VII Cyclic
AMP
levels
in
Cells were planted Fig. 4. Time
L-929
cells at various with PGE,
and grown
as described
treatment
in the legend for .--_
Control (0.5% ethanol)
+PGEl
(SOl.lg per ml)
x 10-o M
5 min 30 min 60 min 5 hours 24 hours
2.3 2.8
f f
0.01 0.07
20.2 22.0
3.1 f 1.9 f
0.1 0.09
13.4 & 0.3
2.3
0.08
i
5.7 4.9
f f f f
0.3 0.6 0.1 0.3
viously studied did have an adenylate cyclase activated by PGEi (B).) These lines gave us an opportunity to see whether prostaglandins only affected cell behavior by increasing the activity of adenylate cyclase and thereby cyclic AMP levels. When cyclic AMP levels were measured in these two lines, no effect on cyclic AMP levels was observed. In the SV40 cells the cyclic AMP level was 0.54 f 0.04 pmole without PGEi, 0.54 + 0.04 5 min after the addition of 50 pg per ml of PGEi, and 0.39 f 0.05 25 min later. With the Balb 3T3 cells the level
7086 was 2.13 f 0.09 pmole without PGEI, 2.30 =t 0.20 5 min after PGEl, and 1.95 =t 0.17 25 min later. Neither line showed a morphologic response to PGEI. Further, the doubling time of the 3T3 SV40 cells was 12 hours in the presence or absence of 50 pg per ml of PGEl (14). DISCUSSION Regulation of cell growth is poorly understood. Many cells divide in logarithmic phase until confluency, at which time growth ceases or continues at a diminished rate. The growth properties are influenced by components of the growth medium such as ionic composition (15, 16) pH (17), serum concentration (8), or by addition of trypsin (9), insulin (7), or prostaglandins (5). Cell growth of most kinds of cells is stopped or severely retarded by the use of depleted medium; the addition of fresh serum to the cultures results in a resumption of cell growth. Recent work has suggested that cyclic AMP may regulate cell growth. In this study we have analyzed the cyclic AMP content and the kinetics of cyclic AMP change under conditions which alter growth. In 3T3 cells and MA308 cells, cyclic AMP levels rise at confluency even with frequent changes of medium. However, in WI38 cells, cyclic AMP levels rise at higher cell density only with prolonged growth without media change, This appears to be result of serum starvation rather than cell density because cells grown to a similar density with frequent media changes retain the low levels of cyclic AMP. The differences between these cells lines leading to these differences in the change of cyclic AMP levels are not clear at the present time. Cyclic AMP levels fall appreciably within minutes after addition of serum to confluent 3T3-4 cells. About 15 hours later the cells undergo increased DNA synthesis. A regulatory function of cyclic AMP in the induction of this DNA synthesis is suggested by the following evidence. First, only a 4-hour contact with fresh medium to serum-starved cells was necessary to initiate the DNA synthesis (18). This suggests that an early event following addition of fresh serum “triggered” the eventual increased DNA synthesis. Second, the presence of (but)zcAMP in the fresh medium will prevent the DNA synthesis upon addition of fresh medium to serum-starved cells (19), or to confluent 3T3 cells which were trypsinized and planted a lighter cell density to initiate DNA synthesis and cell growth.a Third, the addition of trypsin rapidly lowers cyclic AMP levels. Finally, in the variant line of 3T3 cells which does not synthesize DNA upon addition of fresh serum, there also is no rapid fall in cyclic AMP levels (Table IV). The fact that subst’ances which stimulate cell growth decrease the concentration of cyclic AMP, and that PGEi which elevates cyclic AMP levels inhibits growth, strongly suggests that endogenous levels of cyclic AMP are involved in the regulation of growth of fibroblasts. Numerous investigators have reported evidence that cyclic AMP may mediate contact inhibition of growth. Burk first proposed this concept when he observed that theophylline, an inhibitor of phosphodiesterase, would inhibit growth and lower the saturation density of transformed cells (20). We later observed that transformed fibroblasts acquired a more spindly appearance and often grew in parallel arrays, both characteristics of confluent fibroblasts, during the treatment with (but)zcAMP (21). Similar observations were made in Chinese hamster ovary cells (22) but growth data were not present in either report. Later Sheppard reported (23) that (but)zcAMP would restore the saturation density of spontaneously transformed and polyoma 3 M. Willingham,
unpublished
data.
virus-transformed 3T3 cells to that of the parent cells. We have observed that (but)scAMP lowered the saturation density of untransformed 3T3 cells but it did not decrease the saturation density of transformed cells (4). If cyclic AMP is the intracellular mediator of “contact inhibition of growth,” the endogenous levels of cyclic AMP should be elevated at confluency. During analysis of cyclic AMP levels in numerous cell lines, we substantiated this prediction in 3T3-4 cells, a contact-inhibited cell line, but not in noncontactinhibited lines, under conditions of frequent media renewal (2). The results of this study shown that 31’3-42 and MA-308 cells also elevate the intracellular levels of cyclic AMP at confluency. Heidrich and Ryan reported a similar elevation in cyclic AMP levels at confluency with L cells (24). This finding was difficult to interpret however because L cells do not show contact inhibition of growth. It is possible that the effects they observed were due to serum depletion. Recently Sheppard reported that he was unable to show an elevation in cyclic AMP levels at confluency with 3T3 cells (3). In those experiments he rinsed his cells before extra,ction. Since cyclic AMP levels in other cell types are known to change rapidly when the cells are manipulated, differences in the procedure could explain the results (see “Materials and Methods”). Also subtle differences in the cell lines could be possible. Sheppard also found that serum and other factors affected cyclic AMP levels. The addition of fresh serum has numerous effects on fibroblasts growing in culture. They become more motile, less adhesive, less spindly in morphology and are stimulated to synthesize DNA (8, 25). Considering that addition of agents which elevate levels of cyclic AMP also decrease cell motility (26), increase cell adhesiveness to the substratum (27), elongate the cell (21), and prevent DNA synthesis, the lowering of the intracellular levels of cyclic AMP by serum should be considered as the probable mediator of these effects. Acknowledgments-We acknowledge the technical Patricia Middleton and Elizabeth Lovelace.
assistance of
REFERENCES 1. OTTLN, J., JOHNSON, G. S , AND PASTAN, I. (1972) Fed. Proc. 31,440 Abs 2. OTTEN, J., JOHNSON, G. S., AND PASTAN, I. (1971) &o&em. Biophys. Res. Commun. 44, 1192-1198 3. SHEPPARD, J. R. (1972) Arature iVew Biol. 236, 14 4. JOHNSON, G. S., AND PASTAN, I. (1972) J. Nat. Cancer Inst. 48, 1377 5. JOHNSON, G. S., AND PASTAN, I. (1971) J. Nat. Cancer Inst. 47, 1357-1364 6. PEERY, C. V., JOHNSON, G. S., AND PASTAN, I. (1971) J. Biol. Chem. 246, i785-5790 7. TEMIN, H. M. (1967) in Growth Regulating Substances for Animal Cells in Culture, (DEFENDI, V., AND STOKER, M., eds)
p. 103, Wistar Institute Press, Philadelphia G..MATSUYA.Y..BLOOM.S..ROBBINS. A.. ANDGREEN. H. (1967j in Growth Regulating Substances jar inimk Cells’ in Culture (DEEXNDI, V., AND STOKER, M., eds) p.87, Wistar Institute Press, Philadelphia 9. BURGER, M. M. (1970) Nature 227, 170-171 10. GILMAN, A. G. (1970) Proc. Nat. Acad. Sci. U. S. A. 67, 305 11. KRISHNA, G., WEISS, B., AND BRODIE, B. B. (1968) J. Phar8. TODARO.
macol. 12. 13.
Exp.
Ther.
163, 379-385
HAYFLICK, L. (1965) Exp. Cell Res. 37, 614636 MANGANIELLO, V., AND VAUGHN, M. (1972) Proc. Sci.
Nat.
Acad.
U. S. A. 69, 269
14. JOHNSON, G. S., PASTAN, I., PEERY, C. V., OTTEN, J., AND WILLINGHAM, M. (1972) in Prostaglandins in Cellular Biology and the ZnJlammatory Process (RAM~~ELL, P., AND PHARRISS, B., eds) p. 195, Plenum Press, New York
7087 15. ORR, C. W., YOSNIHAWA-FUKADA, M., AND EBERT, J. D. (1972) Proc. Nat. Acad. Sci. U. S. A. 69, 243 16. RUBIN, H. (1972) Proc. Nat. Acad. Sci. U. S. A. 69, 712 17. CECCARINI, C., AND ENGLE, H. (1971) Proc. Nat. Acad. Sci. U. S. A. 68, 229 18. TODARO, G. J., LAZAR, G. K., AND GREEN, H. (1965) J. Cell. Comp. Physiol. 66, 325 19. FRANK, W. (1972) Exp. Cell Res. 71, 238 20. BURK, R. R. (1968) Nature 219, 1272-1275 21. JOHNSON, G. S., FRIEDMAN, R. M., AND PASTAN, I. (1971) Proc. Nat. Acad. Sci. U. S. A. 68, 425
22. HSIE, A. W., AND PUCK, T. T. (1971) Proc. Nat. Acad. Sci. U. S. A. 68,358 23. SHEPPARD, J. R. (1971) Proc. Nat. Acad. Sci. U. S. A. 68, 1316 24. HEIDRICK, M. L., AND RYAN, W. L. (1971) Cancer Res. 31, 1313-1315 25. WEISS, L. (1964) Exp. Cell Res. 33, 277-288 26. JOHNSON, G. S., MORGAN, W. D., AND PASTAN, I. (1972) Nature 236, 54 27. JOHNSON, G. S., AND PASTAN, I. (1972) Nature New Biol. 236, 247
