Hormonal Activation of Adenylate Cyclase in Mouse Melanoma Metastatic Variants R. M. NILES AND J. S. MAKARSKI Boston University School of Medicine, Division of Surgery and Department of Biochemistry, Rm. 908 Instructional Bldg., 80 East Concord Street. Boston, Massachusetts 021 18

ABSTRACT The ability of melanocyte stimulating hormone (MSH), adrenocorticotropic hormone (ACTH), and prostaglandin El (PGE,) to stimulate the accumulation of cyclic AMP was examined in intact mouse melanoma cells of varying metastatic potential. Fl cells (low metastatic potential) had significantly greater cyclic AMP levels in response to all three hormones than F5 (intermediate metastatic potential) and Fl0 (high metastatic potential) cells. The ranking of the response was as follows: MSH, F, > F5> Flo,ACTH, F, > F5> Flo,PGE, Fl > Flo> F5. In contrast to the above, the degree of hormonal stimulation of adenylate cyclase in broken cell preparations was virtually identical in all three melanoma cell lines. Control enzyme activity was depressed in both F5 and F,, relative to F,. The conflicting results between studies of intact vs. broken cell preparations could not be explained by increased cyclic AMP phosphodiesterase activity in F5 and Flocells. We conclude that as the melanoma cells increase in metastatic potential, there is a significant loss in the ability of their cyclic AMP system to respond appropriately to hormonal stimuli. Studies on the mechanism of tumor cell cently been described (Nicolson et al., '77). We metastasis have been hampered by the lack of have sought to examine changes in the cell a suitable experimental model. Recently a surface of three metastatic variants from the series of mouse melanoma variants were se- viewpoint of function. In particular, we have lected which differ in their ability to produce examined the ability of hormones which inlung metastases (Fidler, '73). Cells of high teract with cell surface receptors to activate a metastatic potential not only formed more plasma membrane bound enzyme, adenylate metastases, but were also more invasive a t the cyclase. primary tumor site (Fidler, '75). Thus, these MATERIALS AND METHODS tumor cell lines are useful for studying cell Cell lines and culturing properties which render one line highly metaB-16 F,, F5 and F,, mouse melanoma cells static. Several biochemical differences bewere obtained through the courtesy of Doctor tween cells of high and low metastatic potential have been reported (Boseman et al., '73). 1. J. Fidler, Frederick Cancer Research LaboAmong these were electrophoretic mobility, ratories, Frederick, Maryland. Stock cultures levels of degradative enzymes, surface glyco- of these cells were routinely maintained in protein glycosyl transferases, glycosidases, Minimal Essential Medium (MEM) containand proteases. In other studies (Nicolson and ing Earle's salts, non-essential amino acids, Winkelhake, '75) i t was observed that the vitamin solution, L-glutamine (2 mM), sodium more highly metastatic melanoma cells aggre- pyruvate (1 mM), 50 pg/ml streptomycin sul10% heat gated extremely well with dispersed lung cells fate, 50 unitslml penicillin G, from the same animal. From these results it inactivated fetal bovine serum (GIBCO), adwas postulated that the surface characteris- justed to a final pH of 7.1, and grown in a tics changed as the cells became more meta- 37"C, 95%air - 5%GO2 humidified incubator. static, thereby enabling the cells to recognize The culture media and cells where checked for and adhere to the target organ. Additional bacterial contamination by inoculation in Received Jan. 9, '78. Accepted Mar. 15, '78. properties of these melanoma lines have re-


J. CELL. PHYSIOL. (1978)96: 366-360.




thioglycolate broth and incubation a t 37°C. Possible mycoplasma contamination was checked by the uridinehracil ratio method (Schneider et al., ’74). After 10-15passages in vitro, the cells were tested to assure that they retained their differential metastatic phenotype. This was accomplished by injecting the tail vein of C57/ B1 mice with 1 x lo5 cells of the appropriate cell line, sacrificing the mice one month following inoculation, and counting the number of pulmonary tumor nodules. Intact cell stimulation

For measurement of the intact cell response t o hormones, cells were seeded a t 1.0 x lo5 in 60 mm tissue culture dishes (Falcon) containing 3 ml of MEM +lo% fetal bovine serum. After 24 hours the cultures were refed with 2.7 ml of the same media, and the following day (48 hours after seeding) the experiments were initiated by adding 300 p1 of the hormone dissolved in MEM or ethanol. Cell densities a t the time of stimulation ranged from 2.0-5.0 x lo5 cells/dish (subconfluent). After incubation a t 37°C for various time periods, the reaction was terminated by aspirating the media, washing the cells twice with saline, and adding 1ml of 3.5%perchloric acid (PCA). The PCA extract was taken from the plates, centrifuged to remove cellular debris, and the supernatant neutralized with 3 N KOH. The resultant salt precipitate was removed by centrifugation and the nucleotides in the supernatant acetylated by use of triethylamine and acetic anhydride (Harper and Brooker, ’75). Cyclic AMP was then quantitated by use of the radioimmunoassay technique (Steiner et al., ’69). Authenticity of cyclic AMP was tested by digestion of one half of the sample with purified beef heart cyclic AMP phosphodiesterase (Boerhinger Mannheim). In all cases assay of the digested sample indicated a t least 98%hydrolysis. At the time of hormonal stimulation replicate plates were processed for cell counts by removing the cells from the dish with 0.25% viocase (pancreatin, GIBCO) and enumerated through the use of a model B Coulter counter. Adenylate cyclase assay For adenylate cyclase assay, cells were seeded in 90-mm Falcon tissue culture dishes. All cultures had their media replenished the day before the experiments. Cells were processed for assay by removing the media, wash-

ing the monolayer two times with 0.025 M Tris 0.25 M sucrose pH 7.4, and then scraping the cells from the plate with a rubber policeman. The plates were rinsed with 1 ml of washing buffer and this plus the cell suspension were transferred to a plastic conical centrifuge tube for sonication (Bronson Sonifier, Heat Systems, New York). It was determined experimentally that ten seconds of sonication a t setting 5 ruptured the majority of cells, while preserving the optimal adenylate cyclase responsiveness. Enzyme activity was then measured by the method of Salomon, Londos and Rodbell (’74), as modified by Niles et al. (‘77).


Cyclic nucleotide phosphodiesterase activity

Cells were prepared for phosphodiesterase (PDE) assay as described above for adenylate cyclase assay except that prior to incubation, the homogenate was adjusted to 2 mM MgC12. Enzyme activity was measured using 100 pM or 1 pM cyclic AMP along with negligible amounts 3H cyclic AMP (1pCi/assay tube, 28 Ci/mmole, New England Nuclear), 40 mM Tris-HC1pH 7.4, and 2 mM MgSO, in a final reaction volume of 300 p1 (Chlapowski and Butcher, ’73). The reaction was initiated by addition of 100 p1 of enzyme and the mixtures incubated for 20 minutes at 3OOC. After this time 100 pl of snake venom (1 mg/ml) was added and the reaction mixture incubated ten minutes further. The reaction was terminated by addition of 500 p1 of “stopping solution” containing 0.02 pCi 14Cadenosine, 1mM adenosine, and 0.06 M EDTA. Reaction mixtures were then loaded onto columns (0.6 x 4 cm) containing 1ml of Dowex AG 1 x 2 (Bio-Rad), previously equilibrated with 10 mM Tris-HC1 pH 7.4 buffer. The runoff plus a 3.1-ml wash of 10 mM nis-HC1 pH 7.4 buffer was collected and 500 pl of this fraction counted in 5 ml of Aquasol for recovery of adenosine 14Cand production of adenosine 3H from cyclic AMP. RESULTS

The results from injection of the three melanoma cell lines into their syngeneic hosts showed that in general they retained their differential metastatic phenotype after extensive passaging in vitro. Pulmonary tumor nodule counts were as follows: F, - 10 % 3.5, F5 - 56 % 10.8, Flo- 75% 25.4. It can be readily ascertained from this data that although it was easy to differentiate between F, and F, or Fl0,it sometimes was difficult to differentiate



metastatic potential Flo cells. The F5 cells have a 2.3-fold higher cyclic AMP values than All three metastatic variants responded to the Flocells. A similar experiment with ACTH melanocyte stimulating hormone (MSH), (fig. 1: Panel B), a peptide hormone which is adrenocorticotropic hormone (ACTH) and structurally similar to MSH, revealed the prostaglandin El (PGE,) by markedly increas- same pattern of response as that obtained ing their cyclic AMP levels (table 1). Cate- with MSH. Again the peak response in F, and cholamines (isoproterenol, epinephrine, and F5 cells is obtained after 30 minutes exposure norepinephrine) did not exert any pronounced to the hormone; however, Flocells have a peak stimulatory effect. Although there were no response t h a t occurs between 10 to 30 qualitative differences in the response of the minutes. A t the 30-minute time point, the cyvariants to hormones, there were significant clic AMP levels of F, are 2-fold higher than F5 quantitative differences. To examine these and 8-fold higher than Flo.Incubation of these differences further, we tested the response of cells with PGE, showed a different ranking in the cell lines to MSH, ACTH, and PGE, over a the response (fig. 1: Panel C). The maximum varying time period. Panel A of figure 1 accumulation of cyclic AMP in all three varidepicts the response of all three variant mela- ants occurred between 10 to 30 minutes, and noma lines to MSH over a period of 60 declined thereafter. Although the F, cells minutes. All three lines have a peak response again showed the highest accumulation of cya t approximately 30 minutes with subsequent clic AMP in response to the hormone, the Flo decline during the next 30 minutes. However, cells in this particular hormonal incubation a t the 30-minute time point there is a con- accumulated more cyclic AMP than the F5 siderable quantitative difference. The low cells. Thus, the intensity of the response to metastatic potential F, cells have a 1.5-fold PGE, in F5 and Flocells is opposite that obgreater amount of cyclic AMP than the inter- tained with MSH and ACTH. The hormonal sensitivity of adenylate cymediate metastatic potential F5 cells, and a 3.6-fold higher cyclic AMP level than the high clase in broken cell preparations mimicked

between the metastatic potential of F5 and



Cyclic AMP levels of mouse melanoma variants in response to hormones Picomoles cyclic AMPI1O6 cells Hormone

Control MSH 0.2 pg/ml ACTH 0.4U/ml PGE, 5 x 10-6M Isoproterenol 1 X lo-' M Epinephrine 1 X lo-' M



3.0020.60 5,4342382 1,361239.3 3,7672930 11.220.7 5.320.5

2.3620.22 3,5522108 902k 15.6 18625.9 6.0323.1 5.220.7



_ .. 2.0220.12 1,51627.0 35926.3 1,4212306 6,4C0.2 6.020.8


Cells were prepared and assayed for cyclic AMP levels in response to the hormones as described in MATERIALS AND METHODS. All values are represented as the mean S.E.M. where N = 4.



Adenylate cyclase activity in broken-cell preparations of mouse melanoma variants Picomoles cyclic AMP/minI@g DNA

Control NaF








55.9 287.7 208.9 171.8 125.4


53.7 186.1 223.4 177.8 108.0


5.2 3.7 3.1 2.2

3.5 4.2 3.3 2.0

22.9 86.3 82.1 72.8 58.9

3.8 3.6 3.2 2.6


Cells were prepared and assayed for adenylate cyclaae activity as described in MATERIALS AND METHODS. DNA wan quan. titated by the diphenylamine reaction (Burton, '561. The data represented is the average of duplicate determinations. The entire experiment has been repealed two additional times wth similar reaults. NaF - 2 x 10.) M, MSH - 0.2 pg/ml. ACTH - 0.04 Ulml, PGE, - 5 x l o 4 M, R' - ratio of stimulated to control activity.








do I




Time (Min.)



Time (Min.1





Time (Min.)

Fig. 1 The effect of hormones on cyclic AMP levels in mouse melanoma variants as a function of time of incubation. Cells were prepared for assays and cyclic AMP levels quantitated as described in the text. Panel F,, 0 --- 0 Fs, A-A A, a-MSH 0.2 fig/ml. Panel B, ACTH 0.04 U/ml.Panel C, PGE, 5 X 10-‘M. 0-0 Flo. The data are represented as the mean of triplicate plates of cells, each assayed in duplicate. The bars above and below each data point represent S.E.M. TABLE 3

Cyclic nucleotide phosphodiesterase aetivity in mouse melanoma variants Nanomolee adenosinelminlpg DNA Concentration of cyclic AMP

1 x 10-4 M 1X 10-6M






286 4.4

275 3.4

Cells were prepared and assayed for cyclic nucleotide phosphodiesterase activity as described in MATERIALS AND METHODS. The data is expressed as the average of duplicate determinations. The entire experiment was repeated two additional times with similar results.

the results obtained with intact cells (i.e., MSH > ACTH > PGE,). However, the degree of stimulation (ratio of stimulated/control) of the enzyme by each individual hormone was

approximately the same in all three cell lines. The basal specific activity of Flowas less than one half that of F5 and F1.Additionally, the ratio of NaF-stimulated activity to control activity was 40-50%greater in F, than in F5 or FlO.

The discrepancy between the ability of hormones to increase cyclic AMP levels in intact cells vs. broken-cell preparations may have resulted from a much higher cyclic nucleotide phosphodiesterase (PDE) activity in Fg and Flo as compared to F,. In broken-cell adenylate cyclase assays, PDE activity was negated by the “trapping” of radioactive cyclic AMP by an excess of cold cyclic AMP, while with intact cells PDE activity was not controlled. However, an analysis of the PDE enzyme in the homogenates of the melanoma variants


(table 3) showed nearly equivalent activity a t both high and low substrate concentrations. Therefore, i t appears unlikely that differences in PDE activity can account for the conflicti n g results of intact and broken cell studies.

potential of these cells. The causal relationship between altered cell surface function (hormonal stimulation) and cell behavior (propensity for metastasis) remains to be established. ACKNOWLEDGMENTS


The reasons for t h e discrepancy between intact and broken-cell experiments are not immediately apparent. One possible explanation may lie in the nature of the receptors, i.e., all three variants may contain the same number of hormone receptors although these may be "masked" (altered membrane configuration) in the case of F5 and Flo,and "unmasked" by the homogenization procedure. Alternatively, if one assumes that F, cells contain a greater number of hormone receptors than F5 or Flo, then the homogenization process might destroy a proportionally larger number of receptors in F1, thus causing differences in hormonal responsiveness among the cell lines to be minimal. Direct measurement of the number of MSH receptors on the surface of intact Fi, F5 and Flo cells is currently being determined. One factor which may contribute to t h e altered response of Flois the lower specific activity of the catalytic moiety of adenylate cyclase, as shown by decreased basal and NaF activity. Also, it should be noted that adenylate cyclase measurements were performed on all cell lines at the time of confluence. Other investigators (Boseman e t al., '73) have observed that biochemical differences between F, and Flo cells were lost when comparisons were made at high cell densities. Despite the observed difference between intact and broken cell-studies, we feel that results from intact cell preparations provide the best approximation to physiological mechanisms and hormone action' The intact response to hormones has been tested through many passages and is a stable characteristic of these mel&ma variants. Thus, we have esmarker to add another to those described previously (Baseman et al., '73), which correlates with the metastatic


We wish to express our appreciation to Doctor I. J. Fidler, Frederick Cancer Research Laboratories, Frederick, Maryland for supplying the melanoma variants, to Mary Logue for superb technical assistance, and to Martha Critz for typing and editing the manuscript. This work was supported by Grant CA 18913 from the NCI. LITERATURE CITED Boseman, H. G., G. R. Bieber, A. E. Brown, K. R. Case, D. M. Gersten, T. W. Kimmerer and A. Lione 1973 Biochemical parameters correlated with tumor cell implantation. Nature, 246; 487-489. Burton, K. 1956 A study of the conditions and mechanism of t h e diphenylamine reaction for t h e colorimetric estimation of deoxyribonucleic acid. Biochem., J., 62: 315-323. Chlapowski, F.J.,and R. W. Butcher 1973 Subcellular distribution of adenyl cyclase and phosphodiesterase in Acanthamoeba palestinensis. Biochim. Biophys. Acta, 309: 138-148. Fidler, I. J. 1973 Selection of successive tumor lines for metastasis. Nature, New Biol., 242: 148-149. 1975 Biological behavior of malignant melanoma cells correlated to their survival in uivo. Cancer Res., 35: 218-224. Harper, J. R., and G. Brooker 1975 Femtomole sensitive radioimmunoassay for cyclic AMP and cyclic GMP after 2'0 acetylation by acetic anhydride in aqueous solution. J. Cyclic Nucleotide Res., 1: 207-218. Nicolson, G. L., C. R. Birdwell, K. W. Brunson, J. S . Robbins, G. Beattie and I. J. Fidler 1977 Cell interactions in the metastatic process: Some cell surface properties associated with successful blood-borne tumor spread. In: Cell and Tissue Interactions. M. M.Burger and J. Lash, eds. Raven Press, New York, pp. 225-241. Nicolson, G. L., and J. L. Winkelhake 1975 Organ specificity of blood-bourne tumour metastasis determined by cell adhesion? Nature, 255: 230-232. Niles, R. M., J. S. Makarski, N. Ballinger, H. Kim and A. M. Rutenburg 1977 Adenylate cyclase activity in cultured epithelial cells. In Vitro, 13; 467-471. Salomon, Y., C. London and M. Rodbell 1974 A highly sensitive adenylate cyclase assay. Anal. Biochem.,58: 541-548. Steiner, A. L., D. M. Kipnis, R. Utiger and C. Parker 1969 Radioimmunoassay for the measurement of adenosine 3 , 5' cyclic phosphate. Proc. Nat. Acad. Sci., 64: 367-373.


Hormonal activation of adenylate cyclase in mouse melanoma metastatic variants.

Hormonal Activation of Adenylate Cyclase in Mouse Melanoma Metastatic Variants R. M. NILES AND J. S. MAKARSKI Boston University School of Medicine, Di...
393KB Sizes 0 Downloads 0 Views