Loss of Nuclear Cyclic AMP Binding in Cyclic AMP- unresponsive Walker 256 Mammary Carcinoma*

A marked increase of cyclic AMP-binding and protein kinase activities occurs in the nuclei of N6,02’-dibutyryl adenosine 3’:5’-monophosphate (Bt,cAMP)-responsive Walker 256 mammary carcinoma (W256) following incubation of the tumor slices with CAMP in uitro. The macromolecular fraction containing [“HIcAMP in the nuclei can be extracted with 1.0 M KC1 and identified by acrylamide gel electrophoresis. Cytoplasmic origin of these increased nuclear CAMP-binding and protein kinase activities is suggested by the following observations: (a) cytoplasmic and nuclear CAMP-binding and protein kinase activities are inversely related during the CAMP stimulation of tumor slices; (b) the sequential transfer of CAMP-binding proteins and protein kinase to the nucleus is a temperature-dependent process; and (c) an initial interaction of CAMP with cytoplasm is an absolute prerequisite for the nuclei binding in vitro. The nuclear translocation of CAMP-binding proteins and protein kinase is greatly diminished in the other type (Bt,cAMP-unresponsive) of W256, which grows during the administration of B&CAMP in duo. The experiments using a cell-free system show that cytoplasmic CAMP-binding protein . CAMP complex from responsive W256 binds to isolated nuclei from both responsive and unresponsive tumors, whereas the complex from the unresponsive tumor binds neither nuclei. These results suggest that the lack of nuclear accumulation of CAMP-binding proteins and protein kinase observed in unresponsive W256 could have been due to a defect in cytoplasmic CAMP-binding proteins which fail to interact with nuclear components. Cyclic AMP-binding proteins of unresponsive W256 also fail to respond to endogenously generated CAMP: e.g. when tumor slices are incubated with prostaglandin E, (PGE,) in vitro CAMP-binding proteins in unresponsive W256 do not respond to the PGE, stimulus as do the binding proteins in responsive W256, despite a significant elevation of the CAMP level in the tumor slices. These results suggest that a molecular lesion in CAMP-binding proteins can be a cause of B&CAMP unre-

Cytoplasmic origin of these increased nuclear CAMP-binding and protein kinase activities is suggested by the following observations: (a) cytoplasmic and nuclear CAMP-binding and protein kinase activities are inversely related during the CAMP stimulation of tumor slices; (b) the sequential transfer of CAMP-binding proteins and protein kinase to the nucleus is a temperature-dependent process; and (c) an initial interaction of CAMP with cytoplasm is an absolute prerequisite for the nuclei binding in vitro.
The nuclear translocation of CAMP-binding proteins and protein kinase is greatly diminished in the other type (Bt,cAMP-unresponsive) of W256, which grows during the administration of B&CAMP in duo. The experiments using a cell-free system show that cytoplasmic CAMP-binding protein . CAMP complex from responsive W256 binds to isolated nuclei from both responsive and unresponsive tumors, whereas the complex from the unresponsive tumor binds neither nuclei. These results suggest that the lack of nuclear accumulation of CAMP-binding proteins and protein kinase observed in unresponsive W256 could have been due to a defect in cytoplasmic CAMP-binding proteins which fail to interact with nuclear components. Cyclic AMP-binding proteins of unresponsive W256 also fail to respond to endogenously generated CAMP: e.g. when tumor slices are incubated with prostaglandin E, (PGE,) in vitro CAMP-binding proteins in unresponsive W256 do not respond to the PGE, stimulus as do the binding proteins in responsive W256, despite a significant elevation of the CAMP level in the tumor slices. These results suggest that a molecular lesion in CAMP-binding proteins can be a cause of B&CAMP unre-* This is Paper 6 in a series in the role of cyclic AMP in neoplastic cell growth and regression from the Laboratory of Pathophysiology, National Cancer Institute, National Institutes of Health, Bethesda, Md. 20014. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "'uduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom all correspondence should be addressed.
sponsiveness of one cell population of Walker 256 mammary carcinoma.
Injection ofN6,02'-dibutyryl adenosine 3':5'-monophosphate (Bt,cAMP) into rats bearing Walker 256 mammary carcinoma (W256) produces regression of one type of W256' (Bt,cAMPresponsive) and CAMP-binding proteins appeared to play a major role in this regression (1). It has also been shown (2) that during Bt,cAMP treatment in uivo, CAMP-binding proteins and protein kinase located in the cytoplasm accumulated in the nuclei of the regressing tumor but not in the nuclei of the nonregressing tumor (Bt,cAMP-unresponsive). The present studies explore further the mechanism of Bt,cAMP unresponsiveness of a W256 cell population in both an in vitro system of tumor slices and a cell-free system. Results suggest that the lack of nuclear binding in B&AMPunresponsive W256 is due to a defect in cytoplasmic CAMPbinding proteins. Bt,cAMP administration in uivo (13) showed a nuclear accumulation of CAMP-binding proteins during tumor regression (2). To examine the interaction of CAMP with cytoplasmic and nuclear binding molecules under more defined conditions in vitro, tumor slices from B&AMP-responsive and -unresponsive W256 were incubated with ["HIcAMP and the intracellular distribution of I:'H]cAMP was followed. The temporal sequence of radioactivity uptake by the cytoplasm and the crude nuclear fraction during incubation is shown in Fig. 1. When responsive tumor slices were incubated at 30", the tritium was initially present almost exclusively in the cytoplasm and relatively little was detected in the nuclear fraction. Upon continued incubation, there was a progressive decrease in cytoplasmic "H and an increase in nuclear "H, suggesting that the CAMP was sequentially transferred from the cytoplasm to the nucleus. Such changes were not observed when B&CAMPunresponsive tumor slices were incubated under the same conditions, although the maximum cytoplasmic 3H was about 55% of that found in the responsive tumor slices. At 0" incubation, both responsive and unresponsive tumor slices incorporated the radioactivity continuously into the cytoplasm but not appreciably into the nucleus. This suggests that the increase of I:'H]cAMP into nucleus may be a temperature-dependent process. protein complex, responsive tumor slices were incubated in uitro, in the same manner as described in the legend to Fig. 1, and the labeled cytoplasmic and nuclear binding components were identified by electrophoresis on polyacrylamide gel (2). As shown in Fig. 2 an increased amount of I"HlcAMP was bound to the nuclear components, while the amount of ["HIcAMP bound to the cytoplasm decreased. The difference in apparent mobility between cytoplasmic and nuclear CAMP-binding components shown in Fig. 2  incubated at 30" in the presence or absence of unlabeled CAMP and the enzyme activity was measured in the isolated cytosols and nuclear extracts (Fig. 3). In both responsive and unresponsive tumor slices, protein kinase activity was present mainly in the cytoplasm with relatively little activity in the nuclei at zero time of incubation. Upon incubation of responsive tumor slices in the presence of CAMP, there was a progressive decrease in cytoplasmic protein kinase activity and an increase of nuclear protein kinase activity, suggesting that protein kinase was subsequently transferred to the nucleus.3 This sequential transfer of cytoplasmic protein kinase into the nuclear fraction was shown to be a CAMP-dependent process since the incubation of tumor slices in the absence of CAMP resulted in only a small increase of nuclear kinase activity. Under the same conditions nuclear protein kinase activity in the unresponsive tumor slices did not increase and cytoplasmic kinase activity remained the same. The examination of substrate specificity and effect of protein kinase inhibitor protein (15) on the increased nuclear protein kinase of responsive tumor slices (data not shown) suggested that the increased enzyme is probably the catalytic subunit derived from the cytoplasmic CAMP-dependent protein kinase (16)(17)(18) Fig. 2) was found in the responsive tumor nuclear extract but not in the nuclear extract of the unresponsive tumor. The radioactivity peak in the nuclear extract of the responsive tumor was higher when the nuclei were incubated at 0" rather than at 23", and the incubation at 30" resulted in a marked decrease of the radioactive component (data not shown). By contrast, the [3H]cAMP-binding protein complex was readily extracted from the nuclei after incubation of whole tumor slices with [3H]cAMP at 30" (Fig. 2). This discrepancy may be due to the susceptibility of the CAMP-binding protein complex to proteolytic attack under the conditions of the cell-6352 Nuclear Cyclic AMP Binding and Cyclic AMP Responsiveness in Vivo free system in vitro. Incubation of nuclei with [3H]cAMP and buffer alone produced no extractable binding protein from the nuclei at any temperature. The data in Fig. 4 indicate that cell-free nuclear binding of the CAMP-binding protein complex is greatly diminished in BQAMP-unresponsive W256. Binding of CAMP-binding Protein Complex to Nuclei from B&CAMP-responsive and -unresponsive W256-We next examined whether the failure of CAMP-binding proteins to bind to the nuclei of unresponsive tumor was due to defective cytoplasmic binding proteins or to a defect in the nuclear acceptor sites of these proteins. Cyclic AMP-binding protein 13H1cAMP complex derived from each responsive and unresponsive tumor cytosol was incubated with homologous or heterologous nuclei and the specifically bound L3H]cAMP in the washed nuclei was determined ( proteins, tumor slices were incubated with PGE, in vitro and CAMP-binding activity in the cytosol was measured during the incubation period. Fig. 5 shows the binding activity and CAMP content of the tumor slices. The incubation of tumor slices with PGE, at 30" resulted in a significant decrease of CAMP binding in responsive tumor cytosol, but the binding activity in unresponsive tumor slices did not change (Fig. 5B). The decrease in CAMP binding in responsive tumor slices may be due to an increase in endogenous binding of unlabeled CAMP which decreases the exogenous ["HIcAMP binding. This possibility was examined by performing the binding assay at 23" in order to enhance the CAMP exchange (3) by which the total binding sites (free sites and sites endogenously bound) could be mea--PGE, AT  6. Time course of total CAMP binding in Bt,cAMP-responsive and -unresponsive W256 exposed to PGE, in uitro. Tumor slices were incubated-either at 30" for i0 min or 4" for 19 h. The incubation medium and homogenization were the same as those described in the legend to Fia. 5. Total CAMP binding (PHIcAMP bindina in vitro to bo'th free binding sites and those si'tes-endogenously bound by the exchange (3, 5-7)) to tumor cytosol (105,000 x g, 1 h) was performed at 23" for 4 and 16 h, respectively, as described under "Experimental Procedures." Values are the mean of triplicate determinations.
sured. The time course of CAMP binding at 23" is shown in Fig.  6. At 4 h after the binding reaction, the CAMP-binding activities of the responsive tumor slices, incubated with PGE, at either 30" for 10 min or 4" for 19 h, were significantly lower than those of the controls (without PGE,). Thus, endogenous binding as well as the CAMP concentration appeared to be increased in these tumor slices (Fig. 5E). The binding activity of the tumor slices incubated with PGE, at 4" was greater than that of the control when the assay was carried out for 16 h, demonstrating an increase of total CAMP binding in the responsive tumor slices due to PGE, stimulation.
However, PGE, did not cause an increase in the total CAMP-binding activity of the unresponsive tumor slices despite an elevation of the CAMP concentration (Fig. 5E). The fact that the binding of unresponsive tumor cytosol was appreciably lower after 16 h assay than after 4 h suggested an instability of the binding proteins, as reported in the preceding paper (3).

DISCUSSION
The objective of this study was to assess the mechanism of BtcAMP unresponsiveness of one type of W256 mammary carcinoma by comparing it to a different type of W256 which is responsive to Bt+AMP treatment. The results obtained with Bt+AMP-responsive W256 support the hypothesis that the nuclear translocation of CAMP-binding proteins and protein kinase may be associated with the role of CAMP in growth control. Data obtained from experiments with tumor slices incubated in vitro with CAMP have shown: (a) an absolute prerequisite for cytoplasmic CAMP binding prior to the appearance of nuclei binding, (b) an inverse relationship between cytoplasmic and nuclear activities of CAMP binding and protein kinase, (c) a temperature-dependent transfer of CAMP-binding proteins and protein kinase into the nucleus, and (d) a similarity between protein kinase of stimulated nuclei and that of cytoplasm with respect to substrate specificity and inhibitory effect of protein kinase inhibitor protein on the enzyme. Resolution of CAMP-binding components in polyacrylamide gel electrophoresis, however, indicated a difference between cytoplasmic binding proteins and the binding proteins of stimulated nuclei. These data are similar to those obtained with cytoplasmic and nuclear binding proteins of responsive W256 after Bt,cAMP treatment in vivo (2). One possible explanation for this apparent physicochemical difference is that the binding proteins in the cytoplasm were modified at some step during their nuclear translocation.
Nuclear accumulation of CAMP-binding proteins and protein kinase is greatly diminished in Bt,cAMP-unresponsive W256. Since unresponsive W256 continues to grow during Bt,cAMP treatment, a correlation can be made between the nuclear translocation of CAMP-binding proteins and protein kinase and CAMP-induced tumor regression. Dibutyryl CAMPunresponsive W256 has been shown (l-3) to contain altered CAMP-binding proteins as compared to those in responsive W256, although the unresponsive tumor exhibits a binding activity equivalent to -70% of that in the responsive tumor (Table I). It is conceivable that the altered CAMP-binding proteins in unresponsive W256 are unable to mediate the binding of CAMP to the nucleus, thus causing Bt,cAMP unresponsiveness. Results of experiments using a cell-free system support this possibility. It was found that cytoplasmic CAMPbinding protein. CAMP complex from responsive tumor binds to isolated nuclei from both responsive and unresponsive tumors, whereas the complex from the unresponsive tumor binds to neither nuclei. Thus it appears that the defect of the unresponsive tumor cannot be in the nucleus, but in the cytoplasmic binding protein system. As indicated by the experiment using mixed cytosols from both responsive and unresponsive tumors, the failure of nuclear binding in the unresponsive tumor is not due to a diffusible inhibitor present in the cytoplasm. Thus, the lesion in the unresponsive tumor lies in the binding molecules themselves. If the nuclear association of CAMP-binding protein. CAMP complex must be preceded by activation of the complex and penetration into nucleus, the altered binding molecule must theoretically be unable to carry out any of these processes. However, it is difficult to examine whether activation or binding itself is actually impaired in unresponsive W256, since no assay is yet available for determining an activation reaction independent of nuclear binding.
The above data are consistent with the following model of CAMP action presented schematically in Fig. 7. Exogenously supplied or endogenously generated CAMP (N) binds to the cytoplasmic binding proteins (holoenzyme) of protein kinase, R,C, or Rf2C2, consisting of two asymmetrical regulatory subunits with two globular catalytic subunits (28). This binding may induce the separation of catalytic subunits from the regulatory subunits, producing active protein kinase, C (32)(33)(34)(35)(36)(37)(38)(39) (CR-N), the "activated" state of the complex which is translocated into the nuclei. The complex binds through its R subunit to nuclear acceptor sites (A) consisting of DNA and chromatin-associated proteins. Because the binding site of C subunit in the chromatin is apparently blocked when the C subunit is combined in the complex, its release from the complex after the localization on chromatin may possibly occur by the mechanism of autophosphorylation (40)(41)(42).
The C subunit would then be free to interact with the adjacent genome. R and C subunits in the chromatin would function either independently or cooperatively to cause the eventual tumor cell regression. The hypothetical protein. CAMP complex (CR-N) is shown only with R from the The protein kinase was not isolated from the tumors, the structural symbols used are: AC = adenylate cyclase; N = cyclic AMP (nucleo-identification of the enzyme was deduced from the enzyme of normal tide); C = catalytic subunit of protein kinase; R and R' = regulatory tissues (28)(29)(30)(31). The alternative possibility of the nuclear entry by subunits of protein kinase (CAMP-binding proteins) of Bt2cAMP-the holoenzyme or R and C subunits is not shown in the scheme, responsive and -unresponsive tumors, respectively. A = nuclear since the data of these studies suggest that this is a less likely acceptor sites; -, indicates the sequence of events only. Since pure possibility in these tumors.
BQAMP-responsive tumor cell; the complex cannot be formed with R' molecule from the unresponsive tumor cell. This initial block may result in the failure of tumor regression in the unresponsive tumor.
Cyclic AMP-binding proteins in BtcAMP-unresponsive W256 are unable to respond either to exogenously supplied CAMP (1,2) or to the endogenously generated CAMP. Results of experiments on PGE, stimulation of tumor slices in vitro have demonstrated that CAMP-binding activity in unresponsive W256 tumor cytosol does not change during PGE, stimulation despite a significant elevation of CAMP content in the tumor slices.
The present studies indicate that a molecular lesion in cytoplasmic CAMP-binding proteins may be responsible for BQAMP unresponsiveness of a cell population of W256 mammary carcinoma. If a complex chain of events does underlie the action of CAMP in growth control, one should expect CAMP unresponsiveness to arise by some other means as well. However, the novel mechanism of CAMP action at the nuclear level, suggested by the nuclear binding of CAMP-binding protein.cAMP complex, does suggest its potential importance. Such a mechanism could delineate the possible inter-relationship between the actions of CAMP and steroid hormones in the growth control of hormone-dependent tumors.: It should be pointed out, however, that despite an apparent correlation between nuclear accumulation of CAMP-binding proteins and B&CAMP responsiveness in vivo, we still do not have conclusive evidence that the nuclear binding of CAMP plays a key role in tumor regression. The function of CAMP-binding proteins in the nuclei in conjunction with or without protein kinase is currently under investigation.