Separation of effects of adenosine on energy metabolism from those on cyclic AMP in rat thymic lymphocytes.

In rat thymic lymphocytes incubated for 2 h without exogenous energy-providing substrate, adenosine may be substituted for glucose as a means of maximally restoring energy metabolism and those cellular functions whose rates are sensitive to small changes in the energy balance, such as protein synthesis and uridine utilization for RNA synthesis (Nordeen, S. K., and Young, D. A. (197615. Biol. Chem. 251, ‘7295-7303). Since effects of adenosine in thymocytes and other cells have frequently been attributed to changes in cyclic AMP, this report investigates its possible involvement in these glucose-like restorative actions of adenosine. Although the same range of doses of adenosine effective at raising cyclic AMP also elicit roughly parallel stimulations of protein synthesis and uridine utilization, further results dissociate the restorative actions from those on cyclic AMP. (a) Other purine nucleosides mimic the glucose-like actions of adenosine without increasing cyclic AMP; (b) conversely, prostaglandin E, mimics the cyclic AMP response without restoring energy metabolism or energy-dependent functions; and (c) potentiation of the cyclic AMP response, either by inhibiting phosphodiesterase or adenosine deaminase, does not enhance the restorative response to a range of doses of adenosine. Finally, cyclic AMP-mediated glycogenolysis cannot account for the glucose-like effects since addition of adenosine increases, not decreases, levels of glycogen. Several of the results suggest that metabolism of adenosine itself as an energy-providing substrate might account for its glucose-like effects. Further experiments which reveal that carbon from the ribose moiety is metabolized to CO, in quantities comparable to carbon from glucose support this proposal. Although cyclic AMP does not appear to be involved in the restorative actions of adenosine on cellu-

In rat thymic lymphocytes incubated for 2 h without exogenous energy-providing substrate, adenosine may be substituted for glucose as a means of maximally restoring energy metabolism and those cellular functions whose rates are sensitive to small changes in the energy balance, such as protein synthesis and uridine utilization for RNA synthesis (Nordeen, S. K., and Young, D. A. (197615. Biol. Chem. 251, '7295-7303). Since effects of adenosine in thymocytes and other cells have frequently been attributed to changes in cyclic AMP, this report investigates its possible involvement in these glucose-like restorative actions of adenosine. Although the same range of doses of adenosine effective at raising cyclic AMP also elicit roughly parallel stimulations of protein synthesis and uridine utilization, further results dissociate the restorative actions from those on cyclic AMP. (a) Other purine nucleosides mimic the glucose-like actions of adenosine without increasing cyclic AMP; (b) conversely, prostaglandin E, mimics the cyclic AMP response without restoring energy metabolism or energy-dependent functions; and (c) potentiation of the cyclic AMP response, either by inhibiting phosphodiesterase or adenosine deaminase, does not enhance the restorative response to a range of doses of adenosine. Finally, cyclic AMP-mediated glycogenolysis cannot account for the glucose-like effects since addition of adenosine increases, not decreases, levels of glycogen.
Several of the results suggest that metabolism of adenosine itself as an energy-providing substrate might account for its glucose-like effects. Further experiments which reveal that carbon from the ribose moiety is metabolized to CO, in quantities comparable to carbon from glucose support this proposal. Although cyclic AMP does not appear to be involved in the restorative actions of adenosine on cellu-

Relationships between Small Shifts in Levels or Ratios of Adenine Nucleotides and Changes
in Rates ofProtein Synthesis and Uridine Utilization -As detailed in previous publications from this laboratory (13)(14)(15)(16), isolated thymus cells provided with glucose maintain rather constant levels of ATP, ADP, AMP, and rates of protein synthesis and uridine utilization. In the absence of glucose (or other energy-providing substrate), ADP and AMP begin to rise, and ATP declines slowly. The energy charge falls by about 7% by 2 h. There is also an associated but much larger decline (roughly by 75 to 80%) in some energy-dependent processes, such as protein synthesis and uridine utilization.
As can be seen in Fig. 1, addition of glucose (or alternatively adenosine) at 2 h rapidly restores cellular energy charge to the level observed in cells that had been provided glucose (or adenosine) all along (compare open uersus closed symbols). With adenosine, the total adenine nucleotide pool expands for about 10 min before reaching a new steady state level as can best be seen in Fig. 1 by comparing levels of ATP in the presence of glucose uersus those with adenosine. Yet despite the higher levels of ATP and the expansion of the total adenine nucleotide pool with adenosine compared to glucose, the energy charge is restored to nearly identical values with equal rapidity (within 5 min).
Many experiments, such as that in Fig. 2, have also shown that in conjunction with these rather subtle increases in energy charge there is a rapid re-establishment of maximal rates of protein synthesis' and uridine uptake (a 3-to 6-fold increase). Thus, measurements of protein synthesis and uridine utilization provide the most sensitive indices of small changes in cellular energy metabolism. Data in Fig. 2 also show that adenosine approaches the efficacy of glucose at re-establishing maximal rates of protein synthesis and uridine utilization, but addition of adenosine together with glucose does not yield an additive stimulation of either. In fact, adenosine exerts a transient suppression of glucose-stimulated uridine utilization. This is the result of a short lived, dose-dependent, inhibitory effect of adenosine on uridine incorporation into RNA as previously noted (13). After this initial inhibition (5 to 10 min), rates of uridine utilization do tend to reflect the changes in energy balance.
Effects of Adenosine and Analogues on Cyclic AMP and on Rates of Energy-dependent Functions -Experiments such as that in Fig. 3 show that levels of adenosine that restore rapid protein synthesis and uridine utilization also increase levels of cyclic AMP. However, a closer comparison of dose responses reveals only a rough correlation between levels of cyclic AMP and observed rates of protein synthesis. For example, levels of adenosine between 0.25 and 2.5 mM are about equally effective at restoring protein synthesis (left) and initially (at 5 min) increasing cyclic AMP accumulation (right). At later times, however, the higher doses, e.g. 2.5 mM, further increase cyclic AMP without concurrent increases in protein synthesis. As can also be seen from Fig. 3  without effect on basal levels of cyclic AMP, starved cells exhibit cyclic AMP levels identical with those seen in cells provided glucose from the start, or at later times." To further assess a possible role of cyclic AMP, the ability of adenosine to restore biosynthetic processes was compared with that of added cyclic AMP, or various adenosine analogues (Fig. 4). Cyclic AMP itself (at concentrations up to 1 mM) fails to increase protein synthesis or uridine utilization; however, this failure may possibly be attributed to the inability of added cyclic AMP to penetrate thymus cells (23). Although dibutyryl cyclic AMP (at 1 mM) apparently stimulates these energydependent functions somewhat, this may not be due to its cyclic AMP-like effects, but instead to the energy-yielding metabolism of butyrate derived from deacylation of the dibutyryl derivative. 4 In Fig. 4, inosine, guanosine, AMP, and ATP (also at concentrations of 1 mM) all are effective at restoring rates of energy-dependent processes. AMP and ATP, like adenosine, also rapidly raise levels of cyclic AMP. Yet with ATP, the restoration of protein synthesis and uridine uptake is substantially delayed, suggesting that the increase in cyclic AMP is itself either insufficient or incidental to the restoration; instead further metabolism of the ATP (perhaps to adenosine) may be required." The concept that the rise in cyclic AMP is incidental to the restorative actions of adenosine is supported by further data which indicate that other purine nucleosides, inosine and guanosine, are equally as effective at supporting protein synthesis and uridine utilization without 3  altering intracellular levels of cyclic AMP. The purine nucleoside configuration appears to be integral to the ability to support thymocyte metabolism because neither the pyrimidine nucleoside, uridine, nor the purine base, adenine, stimulates protein synthesis or uridine incorporation.
In addition, neither uridine nor adenine increases cyclic AMP. Other data (not shown) indicate that neither prostaglandin E, (1 PM), which elicits a cyclic AMP response similar to that of 1 mM adenosine (121, nor ribose, nor a combination of adenine and ribose can mimic the restorative actions of adenosine. Experiments that Rule Out Involvement of Glycogenolysis in Glucose-like Effects ofAdenosine -As discussed earlier (see the introduction), glycogenolysis seems a likely mechanism by which adenosine, through its stimulation of cyclic AMP, could rapidly increase levels of glucose-6-P and support energy metabolism and energy-dependent functions. However, experiments like those in Fig. 5 demonstrate that net breakdown of glycogen cannot explain these actions, since addition of adenosine (squares) at the start of the incubation, like glucose Wangles), largely prevents the decline in levels of glycogen seen in substrate-deprived cells (circles). Furthermore, addition of adenosine at 2 h leads to a net gain of glycogen.
Potent&ion of Cyclic AMP Response does not Enhance Restorative Effects of Adenosine -Experiments in which the actions of adenosine on cyclic AMP are independently enhanced further dissociate these actions from the restorative effects. As shown by the shaded bars in Fig. 6, Ro20-1724, an inhibitor of phosphodiesterase, potentiates the cyclic AMP response to doses of adenosine which range from otherwise ineffective to optimal; however, the rate of protein synthesis is not similarly enhanced (compare closed versus open (control) symbols).
A second means of augmenting the cyclic AMP response is to inhibit the degradation of adenosine with EHNA, an inhibitor of adenosine deaminase. Without EHNA, adenosine at a dose of 1 mM is undetectable in the medium 10 min aRer its addition (cf. Fig. 7). One might infer from this that adenosine is no longer required after 10 min, having already generated prevented with EHNA then no metabolites could accumulate, some signal which leads to the restoration and maintenance of and thus no restorative effects would be seen. Conversely, the cellular metabolism. Alternatively, metabolites derived from first mechanism predicts that preservation of adenosine adenosine may be responsible for the glucose-like effects. The should augment (or at least not change) adenosine's restoralatter mechanism predicts that if deamination of adenosine is tive actions on thymocyte metabolism. EHNA at 1 pM par- tially blocks deamination of adenosine; however, 100 pM is required for nearly total blockade (cfi Fig. 7). At this dose, the inhibitor markedly decreases the amount of adenosine necessary to elicit a cyclic AMP response (Fig. 8, closed symbols). For example, 10 pM adenosine (squares) now elicits a response, and 100 PM (circles) is maximally effective since 1 mM (diamonds) produces no further increase. Without EHNA, such a response may also be observed if adenosine is s&iciently increased (to at least 2.5 mM; cf. Fig. 3), which implies that EHNA is acting as suggested as opposed to other potential mechanisms such as inhibition of phosphodiesterase. In contrast, when the cyclic AMP response to adenosine has been thus enhanced with EHNA, the restorative effects are abolished. For example, EHNA prevents the restoration of energy charge (here from 0.865 to 0.918, Table I) and of protein synthesis and uridine utilization (Fig. 9, closed circles).
(The latter is probably reduced to below basal rates because of the inhibitory effects of adenosine on uridine utilization mentioned earlier.) In Fig. 9 inosine-restored (squares 1, and basal (triangles 1 rates of synthetic fimctions are either uninhibited or only slightly inhibited. Measurements of adenine nucleotide levels in Table I also indicate that EHNA also has some inhibitory effect in the presence of glucose; however, this effect is much smaller than when adenosine is the restoring agent. Also, since the results (in Fig. 9 and Table I  were done to test this alternative to a quasi-hormonal, cyclic AMP-mediated mechanism. As can be seen in Table II   CO,-yielding reaction, only the ribose moiety of adenosine contributes to '%O, production. This contention is supported by results in Experiment E which show that virtually no "COI is detected when adenosine labeled only in the base moiety (carbon 8) is used. We interpret these results as demonstrating that the ribose provided by adenosine is metabolized by energy-yielding pathways about as effectively as glucose. The large inhibition of CO, production from adenosine by EHNA (Experiments D and Fl confirms the implication of the results in Fig. 9 and Table I, that deamination to inosine is necessary for the utilization of the ribose moiety as a substrate. Also, as this conclusion predicts, uniformly labeled inosine supports 14C0, production as effectively as adenosine (Experiments E and F). In repeated experiments, EHNA at 100 PM also inhibits '"CO, production from glucose and from inosine somewhat, but always to a much smaller extent than is seen with adenosine.

DISCUSSION
This report investigates mechanisms which may account for the previously reported glucose-like supportive actions of adenosine on thymus cell metabolism (13). Cyclic AMP, which may play an essential role in some adenosine actions including certain actions in lymphoid tissues (IO, 11, 26, 271, does not appear to be involved in the actions of adenosine on energy metabolism in thymic lymphocytes. Here adenosine is able to substitute for glucose by donating its ribose moiety for metabolism as an energy-providing substrate. Adenosine supports cellular metabolism, whereas ribose itself is ineffective, possibly because adenosine provides ribose in a utilizable form or serves as a means to carry ribose into the cell by a specific transport mechanism (or both).
First, adenosine is rapidly deaminated by thymus cells (Fig. 7). Second, enzymes catalyzing Reactions 2 and 3 are present in thymus (28). Third, both inosine and guanosine, substrates for purine nucleoside phosphorylase (291, mimic the supportive actions of adenosine (Fig. 4). Fourth, when an inhibitor of adenosine deaminase is present, then adenosine, a poor substrate itself for the nucleoside phosphorylase (291, is unable to restore either energy charge or macromolecular labeling (Fig. 9, Table I). Finally, metabolism of ribose-5-P by enzymes associated with the hexose monophosphate shunt, the glycolytic, and the glycogen synthetic pathways could account for the ability of adenosine to increase levels of glucose-6-P, lactate, and glycogen (Ref. 13 and Fig. 5).
Among the best studied actions of adenosine are its "insulinlike" effects on the metabolism of glucose and lipids in fat cells (5)(6)(7). Although adenosine, like insulin, inhibits catecholamine-induced lipolysis and, under some circumstances, increases glucose oxidation, it appears to work via a different mechanism than insulin (6,7). Conclusive evidence indicates that in fat cells adenosine is acting through modulation of cyclic AMP metabolism (here decreasing levels) and not as in thymus cells through its own further metabolism.
(a) The effects of adenosine are mimicked by an analogue, N-phenyli-sopropyl adenosine (which is not deaminated) and are not reproduced by inosine, the deamination product of adenosine (5); (b) inclusion of adenosine deaminase in the medium abolishes the effects of adenosine (6,7); and (c) such structurally diverse agents as prostaglandin E, and nicotinic acid which affect fat cell cyclic AMP-like adenosine, reproduce the effects of adenosine on the metabolism of glucose and lipids (5).
However, some actions of adenosine in other tissues may be attributable to the energy-yielding metabolism of the ribose moiety. Degradation of cyclic AMP to adenosine or the subsequent metabolism of adenosine (or both) may be involved in the ability of millimolar levels of either agent to increase levels of glycogen in HeLa cells, an effect opposite to that elicited by dibutyryl cyclic AMP (30), and to increase several biosynthetic parameters (including levels of glycogen) in substrate-deprived uterine horn (31). In the latter studies, other purine nucleotides or nucleosides (or both) also produced adenosine-like effects. The metabolism of adenosine as an energy source probably also accounts for the increase in lactate output in ascites tumor cells given adenosine (32) and contributes to the preservative effects of adenosine (and inosine) on the storage of red blood cells (33) (see also Ref. 34 for recent discussion).
Recently adenosine, a stimulator of adenyl cyclase in cultured astrocytoma cells (31, was used in conjunction with a phosphodiesterase inhibitor to study the effects of cyclic AMP on glycogen metabolism (35). Somewhat surprisingly, instead of a glycogenolytic response, adenosine partially reversed the glycogen-lowering effect of the phosphodiesterase inhibitor. While an accurate assessment is difficult, it is conceivable that the metabolism of adenosine as an energy source may have favored glycogen synthesis, especially since glycogen metabolism was shown to be regulated by the availability of energyproviding substrate.
In this same study, Passonneau and Crites (351 further showed glycogen metabolism to be additionally and independently regulated by cyclic AMP. Little is known about the regulation of glycogen metabolism in thymic lymphocytes, possibly because low levels are difficult to measure. However, in our experiments, adenosine can cause net deposition of glycogen in spite of substantial increases in cyclic AMP, suggesting that here cyclic AMP is not the controlling factor. Furthermore, experiments such as those in Fig. 5 suggest that, under conditions of stringent energy supply, there may be yet another factor involved in the control of glycogen metabolism; the rate of glycogen breakdown appears to be dependent upon the amount of glycogen itself. First order kinetics of glycogen catabolism are indicated by the linear fit obtained when levels of intracellular glycogen in substrate-deprived cells are plotted as a function of incubation time on a semilog scale (inset, Fig. 5). We have found that as breakdown slows sufficiently in substrate-deprived cells levels of ADP and AMP rise as ATP drops, and concurrently those biosynthetic and transport functions that are sensitive to small changes in adenine nucleotide ratios, e.g. protein synthesis and uridine utilization, first slow, and then are virtually shut down (cf. 13,14). The low levels of glycogen in thymus cells probably explain why rates of these processes begin to fall within 30 to 45 min unless the cells are maintained with exogenous energyproviding substrate.
In lymphoid tissues, the relationship between the metabolism of adenosine and possible physiological effects is of particular interest in light of the association of adenosine deaminase deficiency in man with a form of severe combined immunodefi- ciency (36)(37)(38). AMicted individuals, who usually experience chronic bacterial and viral infections early in infancy, have involuted thymuses and lymphocytes that respond poorly to mitogenic stimulation (39). Our observation that the sensitivity of the cyclic AMP response to adenosine is markedly increased by EHNA supports a suggestion of Wolberg et al. (10) that adenosine deaminase deficiency may leave lymphocytes unprotected from cyclic AMP-elevating actions, which in turn have been implicated experimentally in adenosine's immunosuppressive effects (10,11,26,271. In addition our findings together with those of another study (40), which imply that human lymphocytes undergoing blastogenesis are particularly sensitive to EHNA, further suggest that the increase in adenosine deaminase activity seen in vivo at& antigenic stimulation (41) may be a physiological mechanism for protecting proliferating lymphocytes from inhibition by adenosine. This proposal is also supported by observations of large increases in adenosine deaminase activity associated with stimulation of the immune system in disease states such as infectious mononucleosis (42).
Green and Chan reported that inhibition of de nova pyrimidine synthesis is responsible for the cytotoxicity of adenosine in cultured lymphoid cells. They suggested that adenosine deaminase deficiency may increase cellular susceptibility to pyrimidine starvation (43). Adenosine also inhibits pyrimidine uptake (44). We previously have observed an inhibition of the incorporation of uridine into RNA which normally is rapidly relieved as the adenosine is metabolized (13). However, as shown here, when EHNA is used to prevent deamination, the suppression of uridine incorporation is prolonged. This finding suggests that effects on pyrimidine metabolism as well as those on cyclic AMP might also be involved in the pathogenesis of immunodeficiency.