Proceedings of the National Academy of SciencesVol. 68, No. 2, pp. 425-429, February 1971

Restoration of Several Morphological Characteristics ofNormal Fibroblasts in Sarcoma Cells Treated withAdenosine-3':5'-Cyclic Monophosphate and Its Derivatives

GEORGE S. JOHNSON*, ROBERT -M. FRIEDIMANt, AND IRA PASTAN** Laboratory of Molecular Biology and t Laboratory of Pathology, National Cancer Institute, NationalInstitutes of Health, Bethesda, Maryland 20014

Communicated by C. B. Anfinsen, ANovember 23, 19,0

ABSTRACT Sarcoma cells growing in tissue culturehave morphological and growth characteristics differentthan normal fibroblasts. Several of the morphological andgrowth characteristics of normal fibroblasts are regainedwhen the cells are incubated with dibutyryl-cyclic AMPor butyryl-cyclic AMP (0.1-1 mM), or cyclic AMP (3 mM)plus theophylline (1 mM), but not with ATP, ADP, AMP,adenine, or adenosine (1-7 mM). The cell bodies becomeelongated; distinct narrow processes are formed. Withprolonged incubation, the cells show less tendency to pileup or become polygonal. Further, L-929 and Rous sarcoma-transformed hamster cells orient in parallel arrays char-acteristic of contact inhibition. The cells retain theiraltered morphology as long as the butyryl-cyclic AMP ispresent, but revert after its removal. Experiments withcycloheximide, puromycin, and actinomycin D indicatethat protein synthesis, but not RNA synthesis, is requiredfor the response. Microtubular proteins may be involved.No response is observed with normal fibroblasts or withvarious epithelial cells. The data suggest that cyclicAMP may be an important factor in the determination ofmorphology of normal fibroblasts and this function maybe lost or altered during transformation.

Cyclic AIP cANIP has a clearly defined role as a "secondmessenger" in the action of many hormones (1). Since allnormal cells contain cAMP, it seems likely that the cyclicnucleotide serves other roles as well. A question of particularinterest is whether cA.MP metabolism is altered in cancer.

cAMIP and its derivative, V6,02 '-dibutyryl cyclic ANIP [(But)2-cAMIP] inhibit the growth of several types of tumor cells, butnot the parent, untransformed cells (2). Also, polyoma-trans-formed cells appear to have less adenyl cyclase than theuntransformed parent cells (3). These findings suggest thatcAMP may have a role in the altered properties of tumor cells.

Fibroblasts that have been transformed into sarcomas eitherspontaneously or after treatment with chemical carcinogens or

oncogenic viruses differ morphologically from their parentcells. In this studv, we have examined the effects of cA'MP andits derivatives, (But)2-cAMIP and N6-monobutyryl cyclic AMIP(N-But-cA'MP), on the morphology of various normal andsarcoma cells. We find that after treatment with (But)2-cAMIP, A-Bu+-eAMIP, or cAMP with theophylline, butnot ATP, ADP, A'MP, adenosine, or adenine, sarcoma cells ap-parently regain several morphological characteristics of normalfibro-blasts. However, the nucleotide has no effect on normalfibro-blasts or on various epithelial cells.

425

MATERIALS AND METHODS

The cells were grown in Eagle's minimal essential media with10% fetal calf serum (Grand Island Biologicals, Grand Island,N. Y.), glutamine (2 mM), penicillin (50 units/ml), and myco-statin (50 units/ml). Cells were incubated at 370C in a hu-midified atmosphere containing 5% CO2. The cells were rou-tinely grown in 60-mm tissue culture dishes (Falcon Plastics,Los Angeles, Calif.).

Fresh aqueous solutions of (But)2-cAMP (Schwarz Bio-chemicals) had a strong odor characteristic of butyric acid,suggesting that the relatively unstable 02'-butyryl ester groupwas hydrolyzed, and some preparations were found to be amixture of (But)2-cAMP and N-But-cAMP. Therefore, theresidual (But)2-cAMP was converted to N-But-cAMP bytreatment with alkali, and the butyric acid was removed byextraction with ether (4). Other samples of (But)2-cAMPcontained little N-But2-cAMIP but large quantities of butyricacid. The latter was removed by extraction with ether. Theconcentration of N-But-cAIP was determined using theabsorbance at 272 nm and the molar extinction of 15,100 (4).

RESULTSCharacteristics of normal, transformed, and (But)2-cAMP-and N-But-cAMP-treated cells

The morphological characteristics of untransformed mousefibroblasts are shown in Fig. 1. The cells typically have elon-gated cell bodies and distinct narrow processes. As they multi-ply and crowd together, they line up with the long axis of eachcell body in parallel alignment. However, L-929 cells [a mousefibroblast line chemically transformed by 3-methyleholan-threne "in vitro" (5)] show altered morphological and growthcharacteristics. As they multiply and crowd together, theylose their elongated cell bodies and narrow processes and be-come polygonal in shape. Extensive piling up of the cells isobserved with prolonged incubation (Fig. 2). However, duringtreatment with N-But-cAMP (1 miM) the cells regain severalof the morphological and growth characteristics of normalfibroblasts (Fig. 3). The cells align in parallel and retain theirspindly appearance. Little, if any, piling up of the cells is ob-served.We have also studied the effects of (But)2-cA.MP and N-

But-cA.MP on the morphological and orowth characteristics ofother transformed cells. XC cells [ a cell line obtained from aRous sarcoma virus-induced rat sarcoma (6) 1, when grown intissue culture at low cell (population) denl4hiies, have a slightly

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

021

426 Biochemistry: Johnson et al.

(Upper left) FIG. 1. Mouse-embryo fibroblasts were prepared from secondary cultures which were originally obtained from mid-termmouse embryos (New Zealand White strain). Cells were planted at 4 X 104 cells/cm2, and medium was changed after 48 hr. The photo-graph was taken 97 hr after planting (phase contrast, X260).

(Upper right) FIG. 2. L cells Were planted at 6 X 104 cells/cm2 and the medium was changed every 24 hr. The photograph was taken120 hr after planting (phase contrast, X 260).(Lower left) FIG. 3. L cells were planted at 6 X 104 cells/cm2 and were grown in the presence of 1.2 mM N-But-cAMP. The medium

was changed every 24 hr. The photograph was taken 120 hr after planting (phase contrast X 260).(Lower right) FIG. 4. XC cells were planted at 1.7 X 104 cells/cm2. The photograph was taken 30 hr after planting (phase contrast,

X260).

elongated cell body and occasional narrow processes (Fig. 4).However, when the XC cells multiply and crowd together(Fig. 5), they lose their spindly appearance and become po-lygonal, as do the L cells. The processes are no longer apparent,and extensive piling up of the cells is observed during pro-longed incubation.Within 3-5 hr after the addition of 1 mM N-But-cAMP or

(But)2-cAMP, the XC cells begin to show an altered morphol-ogy. This effect is most apparent when the cells are lightlyseeded, but is also seen when the cell layer is heavy. After a30-hr incubation, the cells are dramatically altered (Fig. 6).The cell bodies are elongated, and numerous long, narrowprocesses, some exceeding 100 micrometers in length, aredistinctly apparent. Figs. 5 and 7 show the same control andtreated cultures after a 78-hr incubation. The treated cellsretain their elongated cell bodies and long, narrow processes,

whereas the untreated cells have become polygonal and losttheir processes.During continued growth, the treated cells show a de-

creased tendency to pile up, become polygonal, or lose theirprocesses, but no parallel orientation of the cells could be de-tected. They are also more resistant to removal from the sub-stratum by trypsin. When the treated cell layers are split andthe cells are transferred to fresh tissue culture dishes, theycontinue to grow and maintain their altered morphology for upto 10 days, as long as N-But-cAMP or (But)2-cAMP is present.Longer incubations have not been tested.

Cell viability was evaluated by trypan blue staining.Treatment with N-But-cAMP or (But)2-cAMP does not affectthe viability of the cells. Both control and treated cellsanalyzed daily for 4 days are approximately 95% viable at eachanalysis.

Proc. Nat. Acad. Sci. USA

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

021

Cyclic AMNIP and Cell Surfaces 427

(Upper left) FIG. 5. Same XC cell culture as shown in Fig. 4. Medium was changed every 24 hr. The photograph was taken 75 hrafter planting (phase contrast, X 260).

(Upper right) FIG. 6. XC cells grown as described in legend for Fig. 4 with 1.2 mMNT N-But-cANIP added at planting. The photograph wastaken 30 hr after planting (phase contrast, X 260).

(Lower left) FIG. 7. Same XC cell culture as shown in Fig. 6. Medium was changed every 24 hr. The photograph was taken 75 hrafter plan ting (phase contrast, X 260).

(Lower right) FIG. 8. XC ce'ls planted at 4.2 X 104 cells/cm2 and grown in the presence of 1 mAM sodium butyrate, pH 7.2. Mediumwas changed every 24 hr. The photograph was taken 72 hr after planting (phase contrast X260).

Specificity of the cell type

A number of different cell lines have been tested. A similarelongation of the cell body, formation of long, narrow pro-cesses, and parallel alignment of cells are observed with a Roussarcoma (Schmidt-Ruppin) hamster tumor (Flow Laborato-ries, Rockville, Md.) during incubation with N-But-cAMP or(But)2-cAMP. They also produce similar effects to those de-scribed for theXC cells in a polyoma-transformed mouse fibro-blast (strain AL/N) and a human osteosarcoma. The processformation in the osteosarcoma is particularly striking withsome processes exceeding 200 micrometers in length. However,no alterations are detected with untransformed primarymouse-embryo fibroblasts (NZW or NIH Swiss strains) orBHK-21 cells.No alterations in the morphology of various cell lines of

epithelial origin are observed even upon treatment for up to 4

days. Those tested are the African green monkey kidney(VERO), HeLa, human choriocarcinoma (7), and mousemammary tumor cell lines. This finding suggests that theeffect may be specific for transformed cells of fibroblasticorigin.Specificity of the effectorCommercial samples of (But)2-cAMP produce the same mor-phological changes as N-But-cAMP. However, such prepa-rations are often a mixture of (But)2-cAMP, N-But-cAMP,and butyrate. Therefore, we decided to perform our studieswith both (But)rcAMP and N-But-cAMP, which is a morestable compound. In a typical experiment with XC cells,some morphologic alterations occur at concentrations as lowas 0.1 mM N-But-cAMP and increasing effects are notedup to 1 mM. cAMP has a small effect but usually only atconcentrations of 7 mM.

Vol. 68, 1971

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

021

428 Biochemistry: Johnson et al.

Theophylline, which inhibits phosphodiesterase and therebyelevates intracellular cAMP concentrations, was also tested.At 1 mM, theophylline alone produces a marginal response.Theophylline strikingly potentiates the action of N-But-cAMP (0.1-1 mM) or cAMP (3 mM), and alterations areobserved with N-But-cAMP at concentrations as low as 0.01mM.A number of other purines were tested over a range of 1-7

mM and over a period of 1-4 days. ADP (1-7 mM) is ineffec-tive. ATP (1-3 mM) has no effect, but 7 mM ATP causesextensive cell death after 48 hr. AMP, adenosine, and adenineat 1 mM have no effect, but at 3-7 mM promote extensiveformation of granules and vacuoles in the cells as well as celldeath. After 4 days, a few surviving cells have irregular celloutlines and occasional long processes. The appearance isquite different from the N-But-cAMP-treated cells. Theophyl-line (1 mM) in combination with 3 mM AMP or ADP is alsoineffective.

Epinephrine and isoproterenol (10-6-10-4 M) are ineffec-tive when tested alone or with 1 mM theophylline. Sodiumbutyrate affects cell morphology, but its effects are easilydistinguished from those of N-But-cAMP (Fig. 8). Within 24hr of the addition of 1 mM sodium butyrate, but not 0.1 mM,the cells become considerably flattened and the cell outlinesbecome indistinct. The butyrate does not appear to alter therate of growth of the cells even if it is added at the time ofplanting.

Effects on macromolecular synthesisSince the morphologic response to N-But-cAMP was evidentat 5 hr, we measured the synthesis of DNA, RNA, and pro-

TABLE 1. Effects of N-But-cAMP on macromolecular synthesis

Control,uCi/,ug Protein (%)

[3H I thymidineControl 7.8 + 1.3N,'-But-cAMP treated 8.6 + 0.6 110Control 9.9 ± 0.4N'-But-cAMP treated 8.7 ± 0.7 88

[3H] leucineControl 127 + 8N-But-cAMP treated 91 + 7 71Control 34 + 1.3N-But-cAMP treated 30 + 1.4 88

[3HI uridineControl 0.90g 0.09N-But-cAMP treated 0.80 + 0.03 89

XC cells were planted at 4.2-5.3 X 104 cells/cm2 and the me-dium was changed at 24 hr. After 48 hr, the medium was changedagain and the cells were incubated for 5 hr at 37°C in 1.2 mMN-But-cAMP. Treated and untreated cells were labeled for 10 minwith [3H] uridine (2 ,uCi/ml) or 45 min with [3HI thymidine(2 ,uCi/ml). For the [3H]leucine incorporation, the cells wererinsed 4 times with leucine-free medium without serum after theincubation with N-But-cAMP, and then labeled for 10 min with[3H]leucine (4,uCi/ml) in the same media with N-But-cANIP. Analiquot of the cells was precipitated with 10% trichloroacetic acid,washed with cold ethanol, dissolved in NCS (Nuclear ChicagoSolubilizer), and counted in a Beckman liquid scintillationcounter. Protein was determined by the method of Lowry et al.(14). Each number represents an average of 3 or 4 determina-

tein at that time. Table 1 shows the results of these studies.Treatment with 1.2 mM N-But-cAMP does not significantlyalter the rate of DNA, RNA, or protein synthesis. However,N-But-cAMP at 1 mM was found to inhibit the doublingtime of the cells by about 40%. Presumably, if we waitedlonger to measure DNA, RNA, or protein synthesis, thesewould have been decreased.

Requirements for RNA and protein synthesis

3 hr after planting, XC cells still appear rounded and are notfully in contact with the substratum; by 8 hr, the cells becomeflattened and appear similar to the cells shown in Fig. 2.When cycloheximide (10 ,ug/ml) is added at 3 hr, the cells failto flatten out over the ensuing 5 hr and N-But-cAMP failsto induce cell elongation or process formation. In contrast,if actinomycin D (1-3 jig/ml) is added at 3 hr, the cells adhereand flatten out normally and N-But-cAMP induces cell elon-gation and process formation as well as in the absence ofactinomycin D. Cycloheximide and puromycin (20 ,g/ml)also abolished the action of N-But-cAMP or (But)2-cAMPon a more established, crowded cell layer such as that shownin Fig. 5. These studies suggest that protein synthesis, butprobably not new RNA synthesis, is required for the morpho-logic response.

New cell division does not appear to be required for thecyclic AMP analogues to alter cell morphology since most XCcells show a response by 5 hr, whereas the doubling time ofXC cells is about 18 hr.

Reversal of response

Cells treated with N-But-cAMP revert to the morphologyof untreated cells upon removal of the effector. No reversal isnoted within 5 hr, but by 24 hr the cells lose all their long,narrow processes, and most become polygonal in shape. Suchcells grow normally upon subculture. Like the induction pro-

cess, the reversion to normal morphology appears to requireprotein synthesis. If cycloheximide (13 ,ug/ml) is present dur-ing the 24 hr after the removal of N-But-cAMP, most cells re-

tain their elongated appearance. However, a number ofnarrow processes are lost.

Evidence for involvement of microtubular proteins

Microtubular proteins have several functions in mammaliancells. Among these are the formation of spindle fibers in mito-sis and an ectoplasmic "cytoskeleton" which may govern theexternal cellular configuration (8). Microtubular proteins mayalso be involved in cell spreading and motility (9), and in theformation and maintenance of cell processes (10). Certainnaturally occurring alkaloids preferentially bind to microtubu-lar proteins and disrupt their extended structures. One ofthese, colchicine, was tested. At a concentration of 5 or 50ug/ml, colchicine prevents the development of processes inXC cells treated with N-But-cAMP. If colchicine is addedafter processes are induced by N-But-cAMP, these processesdisappear within 1 hr. Further, the processes present in un-

treated cells are abolished by colchicine treatment. Similareffects of colchicine on untreated fibroblasts have been re-

ported (10).

DISCUSSION

Normal fibroblasts show a definite migration in a monolayer.Inhibition of this movement, termed contact inhibition (of

tions.

Proc. Nat. Acad. Sci. USA

movement), is found upon contact of such cells with each other

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

021

Cyclic AMP and Cell Surfaces 429

(11). As a consequence of this inhibition, the cells tend toorient in parallel arrays. Further, for poorly understood rea-sons, growth and cell division cease at confluency, i.e., whenall the available surface is covered with cells. This is commonlyreferred to as contact inhibition of growth (12). In general,tumor cells and transformed cells display less contact inhibi-tion of movement and growth, and cell division does not ceaseat confluency. Frequently, such cells lose their spindly appear-ance and distinct narrow processes, become polygonal, andextensive piling up of the cells is observed.Transformed fibroblasts treated with (But)2-cAMP and N-

But-cAMP acquire some morphologic characteristics of un-transformed cells. They develop a more spindly appearancewith an elongated cell body and distinct long narrow pro-cesses, arrange in parallel arrays, and show a decreasedtendency to become polygonal and pile up. This responseclearly requires protein synthesis, involves microtubular pro-teins, and is freely reversible upon removal of the effector.The mechanisms for contact inhibition of cell movement or

cell growth are unknown, but an "emitter-acceptor" mech-anism has been suggested (13). This model requires that whencells grow close together or come in contact with one another,a signal is transferred from one cell to the next. Possibly,transformed cells are deficient in either the emitter or the ac-ceptor function. We suggest that the signal for inhibition ofcell movement and growth may be mediated through cAMP innormal cells. Cell contact may activate adenyl cyclase bymeans of a diffusable substance on the plasma membrane orthrough a direct contact of membrane components. This wouldlead to elevated concentrations of cAMP, which in turn affectcell movement or growth. Transformed cells could be alteredin this function by a decrease in adenyl cyclase activity, a lackof the appropriate diffusable substance, or by altered surfaceproperties that make the proper contacts impossible or inef-ficient. Cell morphology at low cell densities could be influ-enced by substances in the growth medium which would altercyclic AMP levels. Measurements of cyclic AMP concentra-tions in normal and transformed cells are needed to evaluatethis hypothesis.

As previously discussed, N-But-cAMP or (But)2-cAMPreverses some of the morphological and growth abnormalitiesof transformed cells; particularly, they restore them to aspindle-like shape, and with L cells, cause parallel alignment.These are properties of normal cells. We have no informationas to how (But)2-cAMP, N-But-cAMP, or other cyclic AMPanalogues would affect the behavior of sarcoma cells in vivo,but are currently conducting such studies.We are grateful to Dr. Alan Rabson for his helpful suggestions

and for supplying the human osteosarcoma, Dr. Peter Kohlerfor supplying the human choriocarcinoma and mouse mammarytumor, Drs. J. Hartley and W. Rowe for supplying the XC cells,Dr. K. K. Takemoto for supplying the polyoma-transformedmouse fibroblasts, and Dr. S. Baron for supplying the L cells. Dr.Johnson is supported by NIH postdoctoral fellowship no. 1-F02-GA35507-01.

1. Robison, G. A., R. W. Butcher, and E. W. Sutherland,Annu. Rev. Biochem., 37, 149 (1968).

2. Ryan, W. L., and M. L. Heidrick, Science, 162, 1484(1968); Gericke, D., and C. Prakash, Hoppe-Seyler's Z. Physiol.Chem., 350, 1469 (1969); Heidrick, M. L., and W. L. Ryan,Cancer Res., 30, 376 (1970).

3. Burk, R. R., Nature, 219, 1272 (1968).4. Falbriard, J. G., T. Pasternak, and E. W. Sutherland,

Biochim. Biophys. Acta, 148, 99 (1967).5. Earle, W. R., E. L. Schilling, T. H. Stark, N. P. Straus,

M. F. Brown, and E. Shelton, J. Nat. Cancer Inst., 4, 165 (1943).6. Svoboda, J., Nature, 186, 980 (1960); Svoboda, J., P.

Chyle, D. Simkovic, and J. Hilgert, Folia Biol. (Prague), 9,77 (1963).

7. Kohler, P. O., W. E. Bridson, T. Vanha-Pertula, andJ. M. Hammond, in Program of the 61st Annual Meeting of theEndocrine Society (New York, 1969), p. 52.

8. Buckley, I. K., and K. R. Porter, Protoplasma, 64,349 (1967).

9. Carter, S. B., Nature, 213, 261 (1967).10. Goldman, R. D., and E. A. C. Follett, Exp. Cell Res., 57,

263 (1969).11. Abercrombie, M., and J. E. M. Heaysman, Exp. Cell.

Res., 6, 293 (1954).12. Stoker, M. G. P., and H. Rubin, Nature, 215, 171 (1967).13. Stoker, M., Virology, 24, 165 (1964); Abercrombie, M.,

In Vitro, 6, 128 (1970).14. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.

Randall, J. Biol. Chem., 193, 265 (1951).

Vol. 68, 1971

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

021