Altered beta-adrenergic receptor-stimulated cAMP formation in cultured skin fibroblasts from Alzheimer donors.

An alteration in signal transduction systems in Alzheimer's disease would likely be of pathophysiological significance, because these steps are critical to normal brain function. Since dynamic processes are difficult to study in autopsied brain, the current studies utilized cultured skin fibroblasts. The beta-adrenergic-stimulated increase in cAMP was reduced approximately 80% in fibroblasts from Alzheimer's disease compared with age-matched controls. The deficit in Alzheimer fibroblasts in response to various adrenergic agonists paralleled their beta-adrenergic potency, and enhancement of cAMP accumulation by a non-adrenergic agonist, such as prostaglandin E1, was similar in Alzheimer and control fibroblasts. Diminished adenylate cyclase activity did not underlie these abnormalities, since direct stimulation of adenylate cyclase by forskolin elevated cAMP production equally in Alzheimer and control fibroblasts. Cholera toxin equally stimulated cAMP formation in Alzheimer and control fibroblasts. Moreover, cholera toxin partially reduced isoproterenol-induced cAMP deficit in Alzheimer fibroblasts. Pertussis toxin, on the other hand, did not alter the Alzheimer deficits. The results suggest either that the coupling of the GTP-binding protein(s) to the beta-adrenergic receptor is abnormal or that the sensitivity of receptor is altered with Alzheimer's disease. Further, any hypothesis about Alzheimer's disease must explain why a reduced beta-adrenergic-stimulated cAMP formation persists in tissue culture.

Several lines of evidence suggest that signal transduction systems are altered in Alzheimer's disease and that these changes can be related to the pathophysiology of the disease (1)(2)(3). These abnormalities in signal transduction may underlie changes in neurotransmitter homeostasis ( 3 4 , oxidative metabolism (1,6,7), kinases and phosphatases (8)(9)(10)(11)(12), changes in processing of amyloid (13,14), and tau proteins (15)(16)(17). Examination of these processes in autopsied brain is difficult because of the postmortem lability of these processes. The hypothesis that Alzheimer's disease is a systemic disorder (1) in combination with the proposal that altered genes or gene regulation (18) may be important suggest that cells from the peripheral tissue may be used to examine primary molecular deficits.
Cultured skin fibroblasts provide a valuable tool for the examination of dynamic cellular and molecular aspects of diseases. Numerous abnormalities persist in cultured fibro-*This research was supported by National Institute on Aging Grants AGO3853 and AG09014. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ''advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. blasts from Alzheimer subjects and from Down Syndrome patients, who predictably develop Alzheimer like neuropathology (19). The reported changes in Alzheimer fibroblasts include alterations in DNA fragility (20), diminished oxidative metabolism including changes in enzyme levels (3,6,7,21), and abnormal signal transduction. The latter include alterations in cytosolic free calcium (22)(23)(24)(25)(26), calcium uptake (27), cAMP (28), the phosphatidylinositol cascade (25,29), and protein kinase C (30). The persistent abnormalities in tissue culture likely reflect genetic differences and may be useful in the design of therapeutic approaches, as well as for exploring detailed molecular mechanisms.
Thus, in the current experiments, one of the signal transduction systems, the receptor-stimulated cAMP formation, was examined in detail. Studies on the receptor-mediated induction of cAMP in Alzheimer's disease have been limited and controversial, and the molecular mechanism of reported changes has not been examined. An increased cAMP response in Alzheimer fibroblasts to a single concentration of isoproterenol at a single time point has been reported (28). On the other hand, others report no difference in the adrenergic response in Alzheimer fibroblasts (31), and PGE1' induction of cAMP was less in lymphocytes from Alzheimer patients than age-matched controls (32). Therefore, the current studies were designed to dissect the signal transduction pathways that may account for the alteration of @-adrenergic receptorstimulated formation of cAMP in Alzheimer fibroblasts and the regulation of these processes by G-proteins. An understanding of the molecular mechanism will likely reveal why the previous studies have been so equivocal. Preliminary results of this study have been presented as an abstract (33).
Human skin fibroblasts from age-matched controls (64 k 3 years) and Alzheimer subjects (59 k 2 years) were studied. The cells were obtained from the Human Genetic Mutant Cell Repository at the Coriell Institute for Medical Research in Camden, NJ. The cell lines that were used in these studies were from apparently normal agematched donors (AG6846, AG9878, AG4560, AG8044, AG9553, AG4260, AG6241, AG6842) and Alzheimer (AG4401, AG4402, AG6265, AG4400, AG7377, AG6848, AG6264, AG8170) donors. No cell line from subjects listed as "at r i s k for familial Alzheimer's disease was studied. All cell lines were studied a t comparable cumulative population doubling time (20.3 t 1.5) and at early passage The abbreviations used are: PGE,, prostaglandin El; CTX, cholera toxin; PTX, pertussis toxin; G-protein, GTP-binding protein; ANOVA, analysis of variance; SNK, Student-Newman-Keul's test; Gpp(NH)p, guanyl-5"yl imidodiphosphate.
numbers. Thus, cells were matched both for the chronological age of the donor and age in culture. The fibroblasts were grown to confluence in Dulbecco's modified Eagle's media supplemented with 10% fetal bovine serum under standardized conditions (34). Cells were subcultured in 35-mm Petri dishes at a seeding density of 1 X lo4 cells/cm2 and incubated at 37 "C in a humidified incubator with 5% COZ and 95% air for 7 days. Media were replaced 4 days after subculturing. On the day of experiment, the Dulbecco's modified Eagle's media were removed, and the cells were washed twice with calcium-free phosphate-buffered saline (NaC1, 170 mM; KCI, 3.4 mM; Na2HP04, 10 mM; KHPPO,, 1.8 mM) at 37 "C and were then incubated with agonists in the same media in the presence of 10 mM theophylline for the indicated times. The reactions were stopped by aspirating the agonist solution and adding 1 ml of ice-cold 0.2 N percholoric acid (HCIO,) containing 4.5 mM EDTA. Cells were removed by scraping and transferred into microcentrifuge tubes. The Petri dish was rinsed with an additional 0.5 ml of the HClO, containing EDTA, and the rinse was added to the microcentrifuge tubes for a total volume of 1.5 ml. Tubes were centrifuged at 15,000 X g in an Eppendorf microcentrifuge for 15 min at 4 "C. The supernatants were neutralized with 1 M KOH, 0.4 M imidazole, 0.4 M KC1 and centrifuged at 15,000 X g for 5 min at 4 "C. The concentration of cAMP in the neutralized supernatant was measured by radioreceptor binding assay (Amersham kit). The acid-precipitable protein was solubilized with 1 N NaOH and measured by the Bradford (35) method.

Isoproterenol-induced CAMP Formation in Control and Alz-
heimer Fibroblasts-Fibroblasts from six age-matched and six Alzheimer donors were stimulated with 10 ~L M isoproterenol for 10 min (Fig. 1). All Alzheimer fibroblast cell lines were less responsive to isoproterenol stimulation than any of the age-matched controls. The temporal response of the isoproterenol-induced cAMP formation was examined in preliminary experiments to determine if the control-Alzheimer differences varied with time. The pattern was similar in one age- matched control (AG4260) and one Alzheimer (7377) cell line except that the magnitude of the response was much less in Alzheimer fibroblasts (data not shown). Isoproterenol-induced cAMP accumulation increased from 1 to 5 min, peaked between 5 and 10 min, and declined by 15 min. Thus, 10-min incubations were utilized for all subsequent experiments.
Dose-response curves tested whether the Alzheimer and control fibroblasts responded similarly at other isoproterenol concentrations (Fig. 2). Increasing the isoproterenol concentration from 10 nM to 100 pM elevated the formation of CAMP in age-matched controls ( n = 4) in a dose-dependent manner, and the responses were plateaued at 10 p~. On the other hand, the formation of cAMP in response to isoproterenol was greatly reduced at all concentrations examined in the Alzheimer lines ( n = 5). For example, increasing isoproterenol from 10 nM to 100 p~ elevated cAMP at 10 min in the controls from 114 f 64 to 1056 f 402 pmol/mg of protein, whereas in the Alzheimer fibroblasts, the rise was only from 88 f 40 to 215 2 33 pmol/mg of protein.
Comparison of the IsoproterenoE-induced Increase in CAMP with That of Other Adrenergic Agonists and Non-adrenergic Agonists-A comparison of the cAMP response to isoproterenol with that of other adrenergic agonists further supported the suggestion of a deficit in the coupling of @-adrenergic receptors to cAMP formation in Alzheimer fibroblasts (Fig.  3). The magnitude of the cAMP accumulation and the reduction due to Alzheimer's disease parallel the compounds' @adrenergic potency (i.e. isoproterenol > epinephrine > norepinephrine (36)). The corresponding cAMP (pmol/mg of protein in 10 min) responses in controls ( n = 3) were 3357 f 428,1814 f 286, and 355 f 73. The reduction in the Alzheimer fibroblasts ( n = 3) compared with controls was 80% with the pure @-agonist isoproterenol, 73% with epinephrine, and 33% with norepinephrine (Fig. 3).
Whether the reduction of receptor-mediated cAMP formation in Alzheimer fibroblasts was specifically linked to the @adrenergic receptor was examined (Fig. 4). Control and Alzheimer fibroblasts were stimulated with isoproterenol or a variety of non-adrenergic agonists: histamine, carbachol, bra-

FIG. 2. Accumulation of cAMP with various isoproterenol concentrations in age-matched and Alzheimer fibroblasts.
On the day of the experiment, the media were replaced with phosphatebuffered saline containing 10 mM theophylline and stimulated with various isoproterenol concentrations (10 nM-100 p M ) for 10 min.
Significant differences ( p < 0.05) between groups were determined by ANOVA (F(9,63) = 4.2) followed by SNK. The asterisk denotes that Alzheimer fibroblasts differ significantly from age-matched control cells at the same isoproterenol concentration. On the day of the experiment, the media were replaced with phosphate-buffered saline with 10 mM theophylline and stimulated with the same concentration (10 pM) of isoproterenol (ISO), epinephrine ( E P ) , or norepinephrine (NEP) for 10 min. Data (mean f S.E.) were obtained from two separate experiments in triplicate or quadruplicate on age-matched (AG4260, AG8044, AG9878) and Alzheimer (AG6848, AG7377, AG8170) cell lines. Significant differences between groups were determined by ANOVA (F(5,37) = 27) followed by SNK. The asterisk denotes that Alzheimer fibroblasts differ significantly ( p < 0.05) from age-matched control cells. Data (mean f S.E.) for isoproterenol and PGE were from two separate experiments in triplicate on age-matched (AG4260, AG8044, AG9878, AG6846) and Alzheimer (AG7377, AG4401, AG6848, AG4402, AG8170) cell lines. Data for the other agonists were from age-matched (AG8044, AG6846, AG4260) and Alzheimer (AG7377, AG4402, AG8170) cell lines. Significant differences between groups were determined by ANOVA (F(11,60) = 58) followed by SNK. The asterisk denotes that Alzheimer fibroblasts differ significantly ( p < 0.05) from age-matched control cells.
dykinin, and PGE1. Isoproterenol and PGEl but not other agonists examined enhanced cAMP formation in fibroblasts (Fig. 4). Although the production of cAMP by isoproterenol declined by 75% in Alzheimer fibroblasts, the cAMP response to PGEl was identical in control (n = 4) and Alzheimer fibroblasts (n = 5; 8758 f 998 versus 8116 f 597 pmol/mg of protein for 10 min). Dose-response curves with PGEl were examined to test the possibility that a difference between control and Alzheimer fibroblasts might occur at other concentrations (Fig. 5 )  on age-matched (AG4260) and Alzheimer (AG7377) fibroblasts. Significant differences between groups were assessed by ANOVA (F(9,53) = 58) followed by SNK.
Significant effects of PGE, ( p < 0.05) were obtained in age-matched and Alzheimer fibroblasts, but no difference between age-matched control and Alzheimer fibroblasts occurred at any concentration examined. and Alzheimer (AG4401, AG6848, AG4402, AG8170) cell lines. Significant differences between groups were determined by ANOVA (F(5,36) = 30) followed by SNK. Significant effects ( p < 0.05) of forskolin were obtained in fibroblasts from both groups, but no difference was obtained between agematched control and Alzheimer fibroblasts at any concentration examined.
identical in Alzheimer and age-matched fibroblasts. Since PGEl does not utilize p-adrenergic receptors (37), the results further support the suggestion of a selective alteration of padrenoceptor-coupling to cAMP formation with Alzheimer's d Isease.
Forskolin Stimulated cAMP Formation in Fibroblasts-Forskolin, an activator of adenylate cyclase, was utilized to test the possibility that reduction of P-adrenergic coupled to cAMP formation with Alzheimer's disease is mediated by diminished activity of adenylate cyclase (Fig. 6). No response occurred at 0.1 PM forskolin, while higher forskolin concen-trations caused a parallel elevation in cAMP concentrations. Since fibroblasts from control ( n = 4) and Alzheimer donors (n = 4) responded similarly to all concentrations of forskolin, reduced adenylate cyclase does not appear to underlie the deficit in P-adrenergic-stimulated formation of CAMP.
Interaction of G-protein Toxins with Deficits in @-Adrenoceptor Induced CAMP Formation in Alzheimer Fibroblasts-The selective reduction of @-adrenergic receptor-coupled cAMP formation with normal adenylate cyclase activity suggested an abnormality between the receptors and G-proteins.
To determine whether CTX-sensitive G-proteins are involved, the effects of CTX were compared in control (n = 4) and Alzheimer ( n = 5) cell lines in the presence or absence of isoproterenol. As in previous experiments, formation of cAMP after a 10-min isoproterenol stimulation declined 88% with Alzheimer's disease (Fig. 7, ISO). Treatment with CTX for 40 min stimulated cAMP accumulation equally in Alzheimer and age-matched fibroblasts (Fig. 7, CTX). The magnitude of the responses by CTX was similar to that of isoproterenol in the age-matched group. Stimulation with isoproterenol for 10 min after pretreatment with CTX for 30 min increased cAMP significantly both in control and in Alzheimer fibroblasts. The responses in Alzheimer lines were only 40% less than in the control group as compared with 88% reduction in the absence of CTX (Fig. 7, CTX + ISO). Thus, the results suggest that abnormality between receptor and G-proteins in Alzheimer fibroblasts is partially abolished by manipulation of G-proteins with CTX.
T o determine whether PTX-sensitive G-proteins were involved in the deficit, the effects of PTX on basal and isoproterenol-stimulated formation of cAMP were evaluated. P T X did not affect basal cAMP formation ( Fig. 7; PTX). Pretreatment of the fibroblasts with P T X for 30 min reduced the isoproterenol-stimulated cAMP formation in controls ( Fig. 7 FIG. 7. Interaction of G-protein toxins with isoproterenolinduced CAMP formation in control and Alzheimer fibroblasts. On the day of the experiment, the media were replaced with phosphate-buffered saline with 10 mM theophylline and stimulated with CTX (1 pg/rnl) or PTX (150 ng/ml). After 30 min, phosphatebuffered saline with or without isoproterenol was added, and the reaction was stopped 10 min later. Data (mean f S.E.) were obtained from two separate experiments in triplicate or quadruplicate on agematched (AG4260, AG8044, AG6842, AG9878) and Alzheimer cell lines (AG7377, AG6848, AG4401, AG4402, AG8170). Significant differences between groups were determined by ANOVA (F(9,86) = 8.3) followed by SNK. The asterisk denotes a significant difference ( p < 0.05) between age-matched control and Alzheimer fibroblasts. t, a significant difference between the age-matched control with isoproterenol alone and age-matched control with other treatments; $, a significant difference between the Alzheimer fibroblasts with isoproterenol and the Alzheirner group with other treatments. 7, PTX + ISO), which suggests the involvement of a PTXsensitive G-protein in the activation of the cAMP response. Pretreatment of Alzheimer fibroblasts with P T X also diminished the response to isoproterenol in controls 12% (Fig. 7, IS0 versus PTX + E O ) , although the difference was not statistically significant. The percent reduction in the isoproterenol-stimulated formation of cAMP in Alzheimer fibroblasts compared with the control was approximately 88% (Fig.  7, ISO), and this difference was not altered by pretreatment with P T X (Fig. 7, IS0 + PTX). Thus, PTX-sensitive Gproteins occur in fibroblasts and interact with cAMP induction by @-adrenergic receptors, but manipulation of these Gproteins by P T X did not alter the difference between control and Alzheimer fibroblasts.

DISCUSSION
These results support the hypothesis that the deficit in the 0-adrenergic receptor-stimulated formation of cAMP with Alzheimer's disease is because of faulty coupling between the receptor, G-proteins, and adenylate cyclase. The studies demonstrate a selective reduction of cAMP formation in response to P-adrenergic receptor stimulation in Alzheimer fibroblasts. The decline in cAMP response paralleled the order of Pagonist potency. Furthermore, PGEl stimulated cAMP formation equivalently in control and Alzheimer fibroblasts. The possibility that the alteration was mediated by a reduction in the catalytic ability of adenylate cyclase was eliminated, because forskolin, which directly activates adenylate cyclase (38), equally stimulated cAMP formation in Alzheimer and control fibroblasts. Furthermore, CTX equally stimulated cAMP formation in age-matched control and Alzheimer fibroblasts and partially abolished the deficit of P-adrenoceptor-coupled cAMP formation in Alzheimer fibroblasts. The results suggest that the abnormality is due to faulty coupling of CTX-sensitive G-proteins to the receptor rather than to the adenylate cyclase. Similarly, the results with autopsied brain showed that forskolin induced cAMP similarly in agematched and Alzheimer patients, but GTP?,-stimulated cAMP was impaired in Alzheimer brain (39). Thus, studies from brain and fibroblasts suggest a functional uncoupling of G-proteins to P-adrenergic receptor-mediated induction of CAMP.
Although numerous biochemical abnormalities including receptor-mediated cAMP formation (28,31), cytosolic-free calcium (23-26, 40, 41), and receptor-mediated phosphoinositide cascade (25,29) have been documented in Alzheimer fibroblasts, many discrepant results exist. For instance, cytosolic-free calcium concentrations in response to agonists in Alzheimer fibroblasts have been reported to decrease (23, 24, 411, increase (401, or not change (25, 26). The phosphoinositide cascade may be enhanced (29) or unchanged (25) under different conditions. The current results on the P-adrenergicstimulated formation of cAMP disagree with previous studies. @-Adrenergic-coupled formation of cAMP has been reported to increase (28) or not change in Alzheimer fibroblasts (31). Different cell culture methods, degree of cell confluence, serum concentrations, growth conditions, and tissue preparations may account for the difference. To deal with the possible technical differences between laboratories, the tissue culture methods have been standardized to a high degree (34). In addition, the molecular basis of these changes has been characterized with temporal and dose-response curves, multiple agonists studies, and with G-protein toxins. The controversy between laboratories will likely diminish when the primary molecular mechanism is identified.
Diminished P-adrenoceptor number, receptor biosynthesis,  (29). Whether the reduction in the P-adrenergic receptor-coupled adenylate cyclase in Alzheimer fibroblasts is mediated by abnormal protein kinase C requires further investigation. On the other hand, glucocorticoids regulate the efficiency of the coupling between the @-adrenergic receptor and G-proteins by increasing receptor mRNA (43), and abnormal glucocorticoid function has been implicated in Alzheimer's disease (49). Reduction of @-adrenergic receptor number (i.e. diminished mRNA formation) or structural alteration of the receptor in Alzheimer fibroblasts may explain the poor functional coupling between the receptor and Gproteins. The number and affinity of @-adrenergic receptors have not yet been determined in Alzheimer fibroblasts. Although total @-adrenergic receptor number is not changed in Alzheimer's disease brain (50), the number of p2 subtype receptors increases (50-52).
The results with CTX delineate the possible molecular mechanism of the alterations in the coupling of G-proteins to the @-adrenergic receptor. CTX increased cAMP without receptor activation and modified the response to receptor stimulation similarly in Alzheimer and control fibroblasts, which shows that CTX-sensitive G-proteins are functional in the Alzheimer fibroblasts. CTX blocks GTPase and thus activates the stimulatory G-proteins and their coupling to adenylate cyclase (53, 54). The equal induction of cAMP formation in control and Alzheimer cells by CTX suggests that the G-proteins are normal in the Alzheimer cell lines and that the coupling of the G-proteins to adenylate cyclase is normal. Thus, the defect is the coupling of the receptor to the G-proteins. An alternative explanation is that Alzheimer cells have excess GTPase and that CTX reverses the Alzheimer deficit, since the GTPase is no longer regulatory.
The results with P T X further define the role of G-proteins in the Alzheimer deficit. The impairment of the isoproterenolstimulated increase in cAMP by P T X demonstrates that PTX-sensitive G-proteins exist in human skin fibroblasts. P T X also impaired receptor-mediated cAMP formation in other cells (e.g. human promonocyte and 3T3 fibroblasts (55)). The observation that the percent reduction in Alzheimer cells was similar in the presence and absence of P T X suggests that alterations in PTX-sensitive G-proteins do not underlie the Alzheimer deficit.
Abnormalities in G-proteins might cause multiple changes in G-protein-linked receptors. Indeed, G a mRNA increases in Alzheimer hippocampus and cortex (56). Since G, regulates adenylate cyclase activity and receptor affinity (57), changes in G, would alter the coupling between receptors and Gproteins. Impaired interactions of G-proteins with muscarinic receptors in Alzheimer hippocampus (58, 59) and with dopamine receptor in Alzheimer frontal cortex (60) have been reported. Furthermore, Gpp(NH)p increases high affinity agonist binding coefficient in control thalamus but has no effect on that from Alzheimer's disease, suggesting that muscarinic receptors coupling to G-proteins seemed to be altered in the brain of Alzheimer's disease (61). Thus, a body of evidence is accumulating that the coupling of G-proteins to receptors may be altered in Alzheimer's disease. Any hypothesis about Alzheimer's disease must explain why a reduced P-adrenergicstimulated cAMP formation persists in tissue culture.