Reverse Transformation of Harvey Murine Sarcoma Virus-transformed NIH/ST3 Cells by Site-selective Cyclic AMP Analogs*

Eighteen site-selective cAMP analogs modified at either the C-8 position or the C-6 position were tested for their growth regulatory effects on the Harvey mu- rine sarcoma virus-transformed NIH/3T3 clone 13- 3B-4 cells grown in a serum-free defined medium. All 18 analogs, when tested individually, exhibited an ap- preciable growth inhibitory effect at micromolar concentrations. The most potent growth inhibitory analogs contained a thio moiety at the C-8 position. In general, C-6 ana- logs required 5-10-fold greater concentrations than C-8 analogs to produce the same degree of growth inhibition. The growth inhibition induced by these an- alogs was accompanied by a change in cell morphology; cells treated with the analogs exhibited the morphology characteristic of untransformed fibroblasts, while untreated cells retained a transformed phenotype. The regulatory subunit of CAMP-dependent protein kinase, the cAMP receptor protein, has two different intrachain cAMP binding sites, and cAMP analogs modified at the C-8 position (C-8 analogs) are generally selective for Site 1, while analogs modified at the C-6 position (C-6 analogs) are generally selective for Site 2. Thus, C-8 and C-6 analogs were tested in combina- tion to enhance the growth regulatory effect. Both growth inhibition and morphological change were en-hanced synergistically by a

Eighteen site-selective cAMP analogs modified at either the C-8 position or the C-6 position were tested for their growth regulatory effects on the Harvey murine sarcoma virus-transformed NIH/3T3 clone 13-3B-4 cells grown in a serum-free defined medium. All 18 analogs, when tested individually, exhibited an appreciable growth inhibitory effect at micromolar concentrations.
The most potent growth inhibitory analogs contained a thio moiety at the C-8 position. In general, C-6 analogs required 5-10-fold greater concentrations than C-8 analogs to produce the same degree of growth inhibition. The growth inhibition induced by these analogs was accompanied by a change in cell morphology; cells treated with the analogs exhibited the morphology characteristic of untransformed fibroblasts, while untreated cells retained a transformed phenotype.
The regulatory subunit of CAMP-dependent protein kinase, the cAMP receptor protein, has two different intrachain cAMP binding sites, and cAMP analogs modified at the C-8 position (C-8 analogs) are generally selective for Site 1, while analogs modified at the C-6 position (C-6 analogs) are generally selective for Site 2. Thus, C-8 and C-6 analogs were tested in combination to enhance the growth regulatory effect. Both growth inhibition and morphological change were enhanced synergistically by a combination of the C-6 and C-8 analogs. Two C-6 analogs or two C-8 analogs added together did not cause synergism.
For both growth inhibition and phenotypic change, C-8 thio analogs acted far more synergistically than C-8 amino analogs when cells were treated in combination with C-6 analogs, suggesting a response of the R" rather than the R' cAMP receptor protein. DEAEcellulose chromatography revealed that the growth inhibition, in fact, correlates with an increase of the R" cAMP receptor protein and a decrease of the R' receptor protein. The growth inhibitory effect of the siteselective analogs was not due to the cytotoxic effect of adenosine metabolites as shown by the different behavior of 8-Cl-CAMP compared with 8-Cl-adenosine in 1) cell cycle effects and 2) release from growth inhibition.
It is concluded that the observed growth inhibition and phenotypic reversion of 13-3B-4 cells is most * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
11 To whom correspondence should be addressed National Cancer Institute, National Institutes of Health, Bldg. 10, Rm. 5B38, Bethesda, MD 20892. likely mediated through the cellular effector, the R" cAMP receptor protein.
Transformation of cells with RNA tumor viruses results in marked changes in cell morphology, growth-related properties, and numerous cellular components associated with transformation (1). It has been shown that one of the cell components that changes rapidly with transformation is the intracellular cAMP level and that reversal of transformation can be obtained by treatments with cAMP analogs (2). It was suggested that cAMP analogs work by increasing the cellular cAMP concentration. This is brought about by inhibition of phosphodiesterase through competition between the analog and endogenous CAMP. Thus, raising the cellular cAMP level is the analog effect generally believed to be involved in the reverse transformation process. There are, however, a number of studies (3)(4)(5) that indicate that a decrease or increase of cellular cAMP does not correlate with transformation or the reverse transformation process, suggesting that cellular effector(s) other than endogenous cAMP may be involved in the cAMP regulation of cell growth. cAMP in mammalian cells functions by binding to its receptor protein, the regulatory subunit of CAMP-dependent protein kinase (6, 7). Two distinct isozymes, type I and type I1 protein kinases, having different regulatory subunits (R', R") but an identical catalytic subunit have been identified (8,9). Differential expression of these isozymes has been shown to be linked to regulation of cell growth and differentiation Because a mixture of type I and type I1 kinase isozymes is present in most mammalian cells (8,9), selective modulation of these isozymes in intact cells may be a crucial function of CAMP. All past studies of the cAMP regulation of cell growth employed either a few earlier known cAMP analogs that require unphysiologically high concentrations (millimolar) or agents that raise cellular cAMP to abnormally and continuously high levels (14-16). Under these experimental conditions, separate modulation of type I and type I1 kinase isozyme is not possible, since cAMP at high levels activates both isozymes maximally and equally without discrimination (6,9,17). Each regulatory subunit (R', R") of protein kinase isozyme contains two types of binding sites for CAMP, Site 1 and Site 2 (18,19). These cAMP binding sites can be differentiated based on their cAMP dissociation rates (18,19) and CAMP analog specificity (19). Site 1 is characterized by a slower dissociation rate and has a relative selectivity for CAMP analogs modified at the C-8 position on the adenine ring (C-8 analogs), while Site 2 has a faster dissociation rate and is more selective for analogs modified at the C-6 position (10-13).

409
(C-6 analogs). The Site 1-and Site 2-selective analog binding experiments established that binding of cyclic nucleotides at either site stimulates binding at the other site for both type I and type I1 protein kinase isozymes (20,21). Furthermore, the site-selective analogs in appropriate combinations demonstrate synergism of binding and show specificity toward either type I or type I1 kinase (22,23). This unique binding specificity of site-selective cAMP analogs is not mimicked by cAMP itself or by previously studied analogs.
Despite extensive studies in vitro, the studies in vivo of the effect of site-selective analogs in intact cells or tissues have been scarce (24)(25)(26)(27). In the present studies, we, therefore, investigated the effect of site-selective cAMP analogs on the growth and morphology of the Harvey murine sarcoma virus DNA transfectant, the NIH/3T3 clone 13-3B-4 cells. The experiments were carried out with the cells grown in a serumfree defined medium to avoid any possible interference of the analog effect that can be caused by the uncharacterized factors present in serum. By using site-selective analogs on the intact transformed cells, we can study the cellular cyclic nucleotide effector mechanisms involved in growth control because: 1) a new test is available to establish whether or not the cAMP receptor protein is the mediator of the response; 2) the analog binding to Site 1 and 2 is cooperative, and therefore, the analog sensitivity toward synergism for the binding to the cAMP receptor protein in intact cells can be measured 3) if the analog combinations demonstrate synergism, lower total analog concentrations can be used to achieve the same cellular response obtained by using a single analog; and 4) we can selectively modulate one of two protein kinase isozymes present in intact cells to correlate the analog effect with a particular protein kinase isozyme.
Cell Culture-13-3B-4 cells, an NIH/3T3 clone that had been transfected with Harvey murine sarcoma virus DNA (kindly provided by Dr. D. R. Lowy, National Cancer Institute), were carried in Dulbecco's modified Eagle's medium containing penicillin (100 units/ ml) and streptomycin (100 pg/ml) and supplemented with 10% fetal bovine serum. Cells were then grown in 60-mm Petri dishes that had been coated with poly-D-lysine in the absence or presence of additives (CAMP analogs) in serum-free chemically defined medium. The serum free medium was composed of Dulbecco's modified Eagle's medium and Ham's F-12 (nutrient mixture F-12 Ham) in a ratio of 75:25 and supplemented with bovine insulin (5 pglml), transferrin (5 pg/ml), histidine HCl (42 pg/ml), glutamine (292 pglml), penicillin (100 units/ml), streptomycin (100 pg/ml), and HEPES (20 mM, pH 7.3); the medium was changed every 48 h, and the additives were provided every 48 h. The cells were grown in humidified incubators in an atmosphere of 10% CO,.
For cell growth experiments, 2 X lo5 cells/60-mm dish were seeded in serum-containing medium and, 24 h later (day zero), the medium was changed to serum-free medium, and the additives were added then and every 48 h thereafter. At the desired times, cell counts in duplicate were performed on a Coulter counter after harvesting cells with gentle trypsinization.
For morphological studies, 5 X lo5 cells were plated onto the shifted to serum-free medium as in the cell count experiment. Photographs were taken using a Leitz inverted microscope.
Cell Cycle Analysis-DNA histograms were generated on the FACS I1 (Becton Dickinson, Sunnyvale, CA) by using the DNA intercalating dye propidium iodide as described by Braylan et al. (28). The percentage of cells in each cycle phase was calculated by using a PDP-11/34 computer (Digital Equipment Corp., Maynard, MA) and software (Division of Computer Research and Technology, NIH) as previously described (29).
DEAE-cellulose Chromatography of Protein Kinase-The CAMPdependent protein kinase holoenzymes and the regulatory subunits of 13-3B-4 cells were separated using DEAE-cellulose according to the method of Robinson-Steiner and Corbin (22). The cell pellets (2-4 X lo' cells), after two washes with phosphate-buffered saline, were hand homogenized with a Dounce homogenizer (60 strokes). The homogenates were centrifuged for 20 min at 10,000 X g. The resulting supernatants (2-2.5 ml) were loaded on a 0.9 X 5.0-cm column preequilibrated with Buffer B. After washing, the column was eluted using a 60-ml total volume gradient from 0 to 0.4 M NaCl in Buffer B with a 1.0-1.2 ml fraction volume.
Protein Kinase Assay-The activity of the CAMP-dependent protein kinase was determined by the method previously described (22).
The reaction was stopped by the addition of 1.0 ml of ice-cold 20% trichloroacetic acid. After standing in an ice bath for 30 min, the samples were passed through Millipore filters (type HA, HAWP 02500). The filters were washed five times with 1 ml of 5% trichloroacetic acid and then counted with liquid scintillation in 7 ml of Filtron-X (National Diagnostics, Somerville, NJ).
CAMP Binding Assay-CAMP binding activity was measured by the method previously described (8) at cAMP exchange conditions (30, 31). The reaction mixture contained a 50-pl sample and a 50-pl solution of 100 mM potassium phosphate (pH 6.8), 2 mM EDTA, 10 mM theophylline, 2 M NaCl, 2 p M [3H]cAMP (-20 cpm/fmol), and f 2 mM CAMP. After incubation at 23 "C for 1 h, the reaction was stopped by filtering through Millipore filters (type HA), and the filters were washed five times with 1.0 ml of potassium phosphate (10 mM, pH 6.8) and then counted as for the protein kinase assay.

Effect of CAMP Analog
Concentrations on Growth Inhibition-A variety of cAMP analogs, modified at the C-6 position or the C-8 position of the adenine moiety, at various concentrations, were tested for their growth inhibitory effect on 13-3B-4 cells. Fig. 1 shows the dose-response curves of 16 representative analogs. Generally, the analogs fit into two major categories based on their dose-response characteristics. The effects of analogs containing a thio derivative (8-thiop-chlorophenyl-CAMP, 8-thiomethyl-cAMP, 8-thioisopropyl CAMP), 8-Cl-cAMP, and 8-Br-CAMP varied directly linear with the concentrations (Fig. 1, A and B). Analogs that were modified at the C-6 position (Fig. 1, C and D) and 8-amino cAMP (Fig. U) displayed hyperbolic dose-response curves.
In general, the C-8 analogs exhibited 5-lO-fold greater inhibitory effect compared with the C-6 analogs at given concenof 21 cyclic nucleotide analogs. The analogs are listed in order from the most to the least potent for growth inhibition for a given modification on the adenine ring. Analogs modified with a thio moiety at the C-8 position, such as 8-thio-pchlorophenyl-, 8-thiomethyl-, and 8-thioisopropyl-cAMP, were the most potent inhibitors, exhibiting over 60% inhibi-  Table  I shows values for the growth inhibitory effect  (20,21). It was further demonstrated that two such site-selective analogs could also be used in combination to synergistically activate protein kinase (22,23). Fig. 2 shows an experiment in which the Site 2-selective analog N6-butyryl-CAMP and the Site 1-selective ' 1 0 0 analog 8-thiomethyl-CAMP were added alone, and in combination, to 13-3B-4 cells in culture. In Fig. 2  combination divided by the sum of the net individual analog effects on growth inhibition. When N'-butyryl-cAMP (1-50 p~) and 8-thiomethyl-CAMP (2.5-10 p~) were added alone to the cell culture, linear dose-response data were seen (Fig.  2, open bars). When 2.5 p~ 8-thiomethyl-CAMP was kept constant and the concentration of N6-butyryl-CAMP was varied, the synergism quotient was directly proportional to the concentration of Nfi-butyryl-CAMP (Fig. 2 A , inset). At maximal growth inhibition, the synergism quotient approached 1.0 (data not shown). However, when 25 PM N6butyryl-CAMP was added with increasing concentrations of 8-thiomethyl-cAMP, a relatively constant degree of synergism was seen (Fig. 2B, inset). The apparent maximum synergism quotient of about 1.5 was not increased when a lower (0.1 p~) or higher (50 p~) (data not shown) concentration of 8-thiomethyl-CAMP was added in combination with N6-butyryl-CAMP. Thus, the magnitude of synergism seen between Nfibutyryl-CAMP and 8-thiomethyl-CAMP was not dependent on the concentration of 8-thiomethyl-CAMP but dependent on the concentration of Nfi-butyryl-CAMP. Experiments similar to those in Fig. 2 carried out with N6-butyryl-CAMP and the other Site 1-selective analog, 8-thio-p-chlorophenyl-CAMP, produced similar results. An increasing magnitude of synergism was obtained when the concentration of N6-butyryl-CAMP was increased from 1 to 50 p~ in the presence of 2.5 p~ 8-thio-p-chlorophenyl-CAMP. Nfi-Carbamoylphenyl-cAMP (25 p~) in combination with 8-thiomethyl-CAMP (2.5 p~) also produced the synergism of growth inhibition (synergism quotient = 1.9). The Nfi analogs exhibited synergism with another C-8 thio derivative, 8-thioisopropyl-CAMP (synergism quotient = 1.5), and with 8-Cl-CAMP and 8-Br-CAMP (synergism quotient = 1.5-1.7). Only a limited degree of synergism was expressed, however, when N6 analogs were combined with 8-amino derivatives (average synergism = 1.2). Thus, the C-6 analogs acted far more synergistically when in combination with 8thio-CAMP analogs than with 8-amino derivatives.
Synergism of growth inhibition was only seen when a Site 1-selective analog was added with a Site 2-selective analog but not when two Site 1-selective or two Site 2-selective analogs were combined (Table 11). When two C-6 analogs or two C-8 analogs were combined such that alone they exhibit   growth inhibition of 10-30% each, their synergism quotients cells treated with Bt2cAMP (500 p~) (panel C) exhibited a were less than or equal to 1 (Table 11). morphology characteristic of untransformed fibroblasts (panel Effect of CAMP Analogs on Cell Morphology-The growth A), while the untreated cells retained a transformed phenoinhibitory effect of cAMP analogs in 13-3B-4 cells correlated type (panel 23). Treatment of cells with N6-butyryl-CAMP with a change in the cell morphology. As shown in Fig. 3, (100 p~)  The synergistic growth inhibitory effect of the C-6 and C-8 analog combination was also reflected in the change in the cell morphology. As shown in Fig. 4, treatment of cells with either 100 pM BtcAMP (Fig. 4B) or 10 p~ 8-Br-CAMP (Fig.  40) for 4 days did not induce a change in the cell morphology; cells were round and refractile and eventually floated away from the substratum as did the untreated transformed cells (Fig. 4A). When the cells were treated in a combination of Bt,cAMP (100 PM) and 8-Br-CAMP (10 p~) , however, the phenotypic transformation was inhibited (Fig. 4E); cells were flat and exhibited contact-inhibited monolayers just as cells treated with a high concentration of Bt2cAMP (500 p~) alone (Fig. 4C). The same synergism was demonstrated between the other C-6 analog, N6-butyryl-CAMP (50 p~) , and the C-8 analog, 8-thiomethyl-CAMP (5 p~) (Fig. 4F), as well as with other combinations of C-6 and C-8 analogs that demonstrated the synergism of growth inhibition.

Site-selective cAMP Analogs in
Effect of 8-Cl-CAMP and 8-C1-adenosine on Growth and Cell Cycle Progression-We examined whether the growth inhibitory effect of the site-selective cAMP analog reflects a cytotoxic effect due to an adenosine metabolite.   The paired numbers were derived from two separate experiments.
mediated through two different mechanisms, the former by a decrease in the rate of replication without affecting cell viability and the latter by cell killing. We examined whether the reduced cell proliferation observed in 13-3B-4 cells after treatment with the analogs is due to a specific block in one phase of the cell cycle. As shown in Table 111, the fractions of cells in GI, S, and G2/M phases were not appreciably different between the control cells (untreated) and the cells treated with N'-butyryl-cAMP (50 p M ) N6,02'-Bt2-cAMP (1 mM), or 8-C1-CAMP. Thus, the inhibition of cell growth induced by the cAMP analogs was not associated with a specific block in one phase of the cell cycle. However, 8-C1-adenosine treatment induced an appreciable increase of the cell population in GI phase with a marked reduction in S phase (Table 111).
These data, combined with those from the release experiments (Fig. 5), confirm that the growth inhibition produced by 8-C1-cAMP treatment was not due to its adenosine metabolite. In fact, by high performance liquid chromatography analyses, 8-C1-adenosine was not detected in either cell extracts or medium from the cells treated with 8-C1-CAMP for 48-72 h (data not shown).
Effect of cAMP Analogs on the Levels of R' and R" cAMP Receptor Proteins-The synergistic effect demonstrated on the growth inhibition and phenotypic reversion of 13-3B-4 cells by the C-6 and C-&thio derivatives of cAMP analogs in combination indicated a response of type I1 protein kinase rather than type I kinase present in the cells.
The relative proportions of free R' and R" and holoprotein kinases, type I and type 11, were determined using DEAEcellulose chromatography. Chromatography of the cytosols from treated and untreated 13-3B-4 cells is shown in Fig. 6. Catalytic subunit was eluted before the start of the NaCl gradient (results not shown). The untreated cells (Fig. 6a) showed two major peaks (peaks 1 and 2) of CAMP-dependent protein kinase activity that were coincident with peaks of cAMP binding activity. Peak 1 eluted at 0.07 M NaCl, and peak 2 eluted at 0.22 M NaC1, and the kinase and binding activities of peak 1 were -3-fold that of peak 2. In addition, there were two minor cAMP binding peaks (peaks 3 and 4) with no CAMP-dependent protein kinase activity, eluted at 0.13 and 0.30 M NaCl, respectively. Radioautography after photoaffinity labeling the fractions of the eluents with 8-a~ido-[~*P]cAMP (32) and performing NaDodSO,-polyacrylamide gel electrophoresis (33) showed that peaks 1 and 3 contained R', whereas peaks 2 and 4 contained R" (Fig. 6, c and d). These results suggest that peaks 1 and 2 are similar to types I and I1 holoprotein kinases, and peaks 3 and 4 are a l 100 n the chromatographic pattern was considerably altered (Fig. 66). Both CAMP-dependent protein kinase activity and cAMP binding activity of peak 1 decreased to 50% of those in the untreated malogs in Growth Control 415 cells, while the CAMP-stimulated kinase and cAMP binding activities of peak 2 increased to 2-and %fold, respectively, over that of untreated cells. In addition, peak 4 cAMP binding activity increased to 3-fold over that of untreated control cells, while peak 3 remained without appreciable change.

Site-selective cAMP An
Thus, decrease of type I holoenzyme (peak 1) was accompanied by an increase of both type I1 holoenzyme (peak 2) and R" subunit (peak 4). A similar change in the elution profile was observed when the cells were treated with other combinations of N6 + C-8 analog that showed synergism in growth inhibition, whereas each of these analogs alone exhibiting little or no growth inhibition caused no apparent change in the elution profile (data not shown). Thus, the same synergism of N6 + C-8 analog combination was observed in protein kinase activity as that observed for growth inhibition and phenotypic change. When intact untreated cells were washed just before collection with the analog combinationcontaining medium, the elution profile was the same as that of cytosol from untreated cells. Thus, the change in peaks 1, 2, and 4 observed in the analog-treated cells was not a consequence of residual analog from the medium interacting with cytosol during cell homogenization. The increase in peaks 2 and 4 observed after the analog treatment suggests that the analog caused both dissociation and increase of type I1 protein kinase. Furthermore, the presence in the treated cells of a considerable amount of type I1 holoenzyme (peak 2) suggests that peak 2, at least in part, may contain a partially dissociated form of holoenzyme, such as R& (36), which may not be resolved from RZC2 (8) by DEAE-cellulose chromatography.

DISCUSSION
The results presented here are the first unequivocal demonstration that site-selective cAMP analogs are capable of exerting a major regulatory effect on the growth of transformed fibroblastic cells.
We demonstrated here that the site-selective cAMP analogs regulated cell growth a t micromolar concentrations. Previous studies of CAMP effect on cell growth, using the earlier studied analogs, reported the effective analog concentrations in an unphysiologic millimolar range (14-16). Thus, the new siteselective analogs are active in their in vivo effect at much lower concentrations.
The efficacy of the analogs tested in the present studies can be related to three basic analog properties: high lipophilicity, low K, for protein kinase activation, and stability to low K , cAMP phosphodiesterase hydrolysis. Analogs with hydrophobic substituents such as an alkyl or aryl group have been shown to be highly lipophilic, whereas analogs with more hydrophilic substituents such as 8-amino derivatives show the lowest lipophilicity (25). Our data showed no clear correlation between the hydrophilic or hydrophobic substituents of analogs and the efficacy for growth inhibition (Table I). Likewise, the growth inhibitory effect of the analogs did not appear to be directly related to the resistance of the analogs toward the phosphodiesterase. The low K , phosphodiesterase I, values for N6 derivatives are generally much greater than those of C-8 derivatives (25, 37). However, C-8 analogs, which are much less resistant than N6 analogs toward phosphodiesterase, showed the greater growth inhibitory effect. The synergistic effect demonstrated by C-6 and C-8 analog combinations further argues against the role of phosphodiesterase in the analog effect. These analogs in combination exerted growth inhibition a t concentrations at least one-tenth below the reported (25) I, for the low K, phosphodiesterase; a t these low concentrations, the analogs would not be metabo-

Site-selective CAMP Analogs in Growth Control
lized to produce toxic products. The absence of a role for adenosine metabolites in the analog effect was also experimentally demonstrated by direct comparison of the effect of 8-C1-adenosine to that of 8-C1-CAMP. The different behavior between 8-C1-adenosine and 8-Cl-CAMP was demonstrated in 1) cell cycle effects and 2) release from growth inhibition. It seems, therefore, highly unlikely that the phosphodiesterase significantly contributes to the analog-mediated growth inhibition.
The analog effect appears to be more related to low K,, for protein kinase activation. C-8 analogs, having lower or similar K, values (23,25) as CAMP, were 5-10 times more potent in growth inhibition than N6 analogs which possess 10 times or greater K, values (23,25) than CAMP. The growth inhibitory effect of the analogs was related to selective activation of type I1 protein kinase, that is the inhibition brought about an increase of R" and type I1 protein kinase with a decrease of R' and type I protein kinase. Thus, the analog efficacy correlated with its ability to selectively activate type I1 protein kinase over type I kinase. In fact, 8-thio analogs that show preferential activation of type I1 over type I kinase (23) exhibited greater potency of growth inhibition than 8-amino analogs that preferentially activate type I protein kinase (23). The synergistic effect of the C-6 and C-8 analog combination demonstrated on growth inhibition further supports that an analog's efficacy is dependent on its ability to selectively activate type I1 protein kinase. It has been shown that bovine heart type I1 and rat heart type I protein kinase exhibit a different C-8 analog specificity for stimulation of binding and synergism of activation when combined with a C-6 analog (22). Specifically, when used in combination with C-6 analogs, those analogs with a sulfur atom attached to C-8 act more synergistically for type I1 protein kinase, and the analogs with a nitrogen atom attached at C-8 exhibit greater synergism for type I protein kinase (22). Our results showing that the synergism of growth inhibition by N6 analogs when combined with 8-thio analogs far exceeds that by N6 analogs in combination with 8-amino derivatives, therefore, suggests a response of type I1 rather than type I protein kinase present in 13-3B-4 cells.
The synergistic effect of the N6 and C-8 analog combination demonstrated in 13-3B-4 cells is similar to that previously shown by Beebe et al. (25) in isolated rat adipocytes that primarily contain type I1 protein kinase isozyme. Beebe et al. (25) demonstrated that, for both lipolysis and protein kinase activation, C-8 thio analogs act more synergistically than C-8 amino analogs when incubated with adipocytes in combination with C-6 analogs, a characteristic of type I1 protein kinase. The efficacy of individual analogs in adipocytes versus 13-3B-4 cells, however, showed some inconsistencies. In 13-3B-4 cells, N6-butyryl-cAMP was a relatively potent growth inhibitor and N6-carbamoylpropyl-CAMP was a weak inhibitor, whereas, in adipocytes, N6-carbamoylpropyl-CAMP was the most potent activator of all C-6 and C-8 analogs tested, and N'-butyryl-cAMP was quite ineffective. It appears that, in both systems, the analogs exert their effects through binding to type I1 protein kinase. However, the efficacy of the analogs shown in lipolysis may not be directly comparable to the analog efficacy in the complex phenomenon of growth inhibition.
From these studies, we suggest that a mere decrease or increase in cellular cAMP does not determine cell transformation or reverse transformation, respectively, but that the cellular cAMP effector(s), CAMP-dependent protein kinase, plays an important role in these processes. Site-selective cAMP analogs thus provide us with an important biological tool for understanding the mechanism of cell transformation.