Molecular and Cellular Biochemistry 149/150: 287-292, 1995. �9 1995 Klawer Academic Publishers. Printed in the Netherlands.

Disturbances in signal transduction mechanisms in Alzheimer's disease

Christopher J. Fowler 1,2, Richard E Cowburn 1, Anita Garlind 1, Bengt Winblad 1 and Cora O'Neill 1,3 1Alzheimer's Disease Research Centre, Department of Geriatric Medicine, Karolinska Institute, Huddinge University Hospital, S-141 86 Huddinge, Sweden," 2Astra Pain Control AB, Preclinical R and D, Novum Unit, S-14157 Huddinge, Sweden and 3Department of Biochemistry, University College, Lee Maltings, Cork, Ireland

Abstract

Many of the treatments directed towards alleviation of symptoms in Alzheimer's disease assume that target receptor systems are functionally intact. However, there is now considerable evidence that this is not the case. In human post-mortem brain tis- sue samples, the function of the GTP-binding protein G s in regulating adenylyl cyclase is severely disabled, whereas that of G~ is intact. This difference in the function of the two G-protein types is also found in G-protein regulation of high- and low- affinity receptor recognition site populations. Measurement of G-protein densities using selective antibodies has indicated that the dysfunction in Gs-stimulation of cAMP production correlates with the ratio of the large to small molecular weight isoforms of the Get subunit. With respect to intracellular second messenger effects, there is a dramatic decrease in the density of brain receptor recognition sites for Ins(1,4,5)P 3 that is not accompanied by a corresponding change in the Ins(1,3,4,5)P 4 recognition site density. Protein kinase C function is also altered inAlzheimer's disease, a finding that may be of importance for the control of [3-amyloid production. These studies indicate that signal transduction processes are severely compromised in Alzheimer's disease. Some of these disturbances are also seen in cultured fibroblasts fromAlzheimer's disease patients, indicating that they are neither restricted to areas of histopathological change, nor non-specific changes found late in the course cff the disease. Cellular models to investigate the relation between amyloid production and deficits in signal transduction are also discussed. (Mol Cell Biochem 149/150: 287-292, 1995)

Key words: Alzheimer's disease, receptors, G-proteins. adenylyl cyclase, phosphoinositide breakdown, inositol (1,4,5)- trisphosphate, protein kinase C, 13-amyloid

Abbreviations: GTP3,S - guanosine-5'-O-(3-thiotriphosphate); Gpp[NH]p - 5'-guanylylimidodiphosphate; Ins - inositol; P - phosphate; CGP- 12177 - (+)-4-(3-t-butylamino-2-hydroxypropoxy)-[5,7-3H]benzimidazol-2-one

Introduction

Alzheimer's disease is a dementia disorder that afflicts approxi.- mately 5% of the population over 65 years of age. The ever- increasing proportion of elderly in industrialised countries means that the care of patients with this disease, for which there is at present no satisfactory therapy, places an increasing burden both on caregivers and upon economic resources.

Over the last few years, considerable advances in our

knowledge of the integrity of cellular signal transduction mechanisms inAlzheimer's disease have been made. Several studies have reported dysfunctions in GYP binding ( 'G-') proteins and in the actions ofintracellular second messengers. Moreover, there is evidence that such changes are not re- stricted merely to the late stages of the disease. These changes, together with a consideration of their implications for the disease process and its treatment, are the subject of the present article.

Address for offprints." C.J. Fowler~ Department of Pharmacology, Umegt University, S-901 87 Umefi, Sweden

288

Changes at the level of the receptor recognition site

The simplicity of the radioligand binding technique and the availability of selective ligands for a wide variety of receptor systems have allowed a vast literature to be built up concern- ing levels of receptor recognition sites inAlzheimer's disease (for review, see e.g. [1]). It should be pointed out, however, that changes in the levels of receptor recognition sitesper se are very hard to interpret given both that the assay is non- functional, and that differences in receptor reserve from cell to cell can mean that a small change in one system is of con- siderable functional importance, whereas a large change in another system is not. In consequence, we have restricted this article to the consideration of changes in the receptor - G- protein coupled systems and to changes in the cellular effects of the second messengers produced as a result of receptor stimulation.

Disruption in G-protein function

There is now considerable evidence that there is a Gfprotein dysfunction inAlzheimer's disease. In our original study [2], we demonstrated that GTPu cAMP formation was greatly reduced in all regions tested in post-mortem

samples fromAlzheimer's disease cases compared with con- trols. These regions included those associated with the end- stage histopathological changes of neuritic plaque and neurofibrillary tangle accumulation (temporal and frontal cortices, angular gyrus), but were also seen in brain regions considered to be more mildly affected (occipital cortex, cer- ebellum). A subsequent study [3] extended these findings to the hippocampus. The response to stimulation by forskolin was unaffected, suggestive of an impairment in G s function rather than an alteration in intrinsic enzyme activity.

G c~ mRNA levels have been shown to be increased in Alzheimer's disease [4]. In our hands, the levels of the large ( G . 0 or small (G_~) molecular weight Gs~ subunit types were not significantly changed (Fig. 1). Rather, the ratio of G L: G s subunit types was significantly increased [3]. Furthermore, this correlated with the relative stimulation produced by GTPTS, suggesting that in Alzheimer's disease the reduction in G -mediated cAMP production may be re- lated to a change in the relative amounts of the G subunits. Finally, a disruption in the ability of Gpp[NH]p to affect the agonist inhibition curve of the binding of [3H]CGP-12177 to [3~-adrenoceptors [5] is consistent with a disruption in G function.

Deficits in cAMP production have also been reported by others. Recently, Schnecko et al. [6] found large reductions

A~ B~

250- .6

20(}-

~150-

"~ 100-

50-

0- 0 1 tOO

[Forskolin] (/aM)

2oo - T 175- T T~ --'4-x ~ ~ 60-

15o ~" ~ 45-

xa 125-

1oo- i 3o- ~ 15-

75- i i

0.01 0.I 1 10 IGTI'~SI (laM)

C ~

75-, Gsa-L Gsa-S

i i i h

C A C A

D. 100-

85-

~ 70-

"d s. ~ 55-

J~

40-

0.01 O l J 10 IGpp[NI tip] (ltMI

1011

E ~

75-

.~" 60"

~,45-

�9 r-, 30-

~ e 15-

Gia Gia-1

T

I i

C A C A

Fig. 1. Temporal cortical G-protein function in Alzheimer's disease. Panel A, stimulation of cAMP production by forskolin (drawn from the data given in Table 4 of Ref. [2]); Panel B, stimulation of cAMP production by GTPyS (calculated from original data presented as prnol cAMP/mg protein/rain in Fig. 1A of Ref. [2]) under conditions favouring G s activity. Panel C, densitometric analyses of G_L and G_s levels (drawn from the data given in Table 2 of Ref. [3]). Panel D, inhibition of cAMP production by Gpp[NH]p (calculated from original data presented as pmol cAMP/rag protein/rain in Fig. 1A of Ref. [9]) under conditions favouring G~ activity. Panel E, densitometric analyses orGy, and G~a_~ levels (drawn from the data given in Table 2 of Ref. [3]). Synaptic membrane preparations from control (r-I, A) and Alzheimer's disease (IESI, T) autopsy cases were used. Data are means • S.E.M, n = 6-8.

in basal, GTP and Gpp[NH]p-stimulated cAMP production in the hippocampus and cerebellum in Alzheimer's disease. These authors, however, found no significant changes in the responses to forskolin. Given that the assays contained a nucleoside triphosphate regenerating system, one possibility is that there is sufficient GDP (and/or GTP) present in the crude membrane preparations used to produce GTP in situ, so that the basal responses are to a certain degree GTP-stimu- lated. This would be a consequence of the stated intention of the authors to process the membranes as little as possible 'in order to stay as close as possible to the "physiological situa- tion"' [6]. Certainly, addition of GDP to membranes will produce a cAMP response, presumably via its conversion to GTP [7]. Thus, the results of Schnecko et al. [6] are also consistent with a deficit in G s function.

Such a explanation might also hold for the frontal cortical data reported in the study of Ross et al. [8], who found a decreased basal and fluoroaluminate-stimulated cAMP pro- duction in this region. Although the lack of data for stimula- tion with forskolin make for difficulties of interpretation, the finding of Ross et aI. [8] that both basal and fluoroaluminate- stimulated cAMP production were not decreased in the hip- pocampus is discrepant with our data. The authors found no significant changes in either Gs-L or G s levels in either region, a result also seen in frontal cerebrocortical membranes by Wang and Friedman [9]. It is possible that the degree of intactness of G function, particularly in the hippocampus, may reflect sample selection. Alternatively, as pointed out by Ross et al. [8], differences in the composition of the adenylyl cyclase assay medium might be of importance in this respect.

Using a different approach, Wang and Friedman [9] dem- onstrated a dysfunction in G s in post-mortem frontal cere- brocortical Alzheimer's disease samples. These authors measured the levels of basal and isoprenaline-stimulated immunoprecipitation of [35S]GTPyS labelled Gs~ proteins, and found that whilst there was a modest (-10%) reduction in basal labelling in theAlzheimer samples with respect to con- trol samples, the stimulation of binding produced by 10 laM isoprenaline was reduced by 85% in the Alzheimer samples.

Unlike the situation with G, the function of the G~ protein, as assessed either by measuring inhibitory effects of Gpp- [NH]p and cyclohexyladenosine upon cAMP production [6, 10], or by measuring the inhibitory effects of Gpp[NH]p upon agonist radioligand binding to the G~ coupled ct2-adrenergic, 5-HTjA-serotoninergic and ~c-opioid receptors [11-13], is preserved in Alzheimer's disease. On the other hand, levels of the G~_2 and particularly the G~_~ subunit are dramatically decreased in the Alzheimer's disease temporal cortex [3; see Fig. 1] but not in the frontal cortex or hippocampus [3, 8, 9].

The discrepancy between levels and function seen in our studies can be explained in two ways. It is possible that there is a considerable excess of this G~ subunits in temporal

289

cortical membranes sufficient even in the Alzheimer's dis- ease cases for functional adenylyl cyclase inhibition and receptor-Gj interactions. Alternatively, deficits in the G~ subunits and particularly the Gia I subunit in theAlzheimer's disease temporal cortex may be linked to effectors other than adenylyl cyclase. Such effectors may include the regulation of ion channel and phospholipase C function [ 14,15] (for fur- ther discussion, see Ref. 3). This might explain the finding that the stimulation by the muscarinic agonist carbachol of immunoprecipitation of [3sS]GTPTS labelled G~ proteins is greatly reduced in Alzheimer's disease frontal cortical samples [9].

Little is known about signalling mediated via the phos- phoinositide pathway in the brain in Alzheimer's disease, mainly due to the difficulties of measuring this parameter in post-mortem tissue. However, a recent article reported that there was a reduction in the ability of the muscarinic agonist carbachol to stimulate guanine nucleotide-dependent [3H]- phosphatidylinositol-4,5-bisphosphate hydrolysis in autopsy samples from Alzheimer patients compared with controls [ 16]. However, the stimulations were small ( 12.9 -+ 3.0% for controls and 5.1 + i .5% for theAlzheimer cases) and the basal phospholipase C activity rather variable, particularly in the controls, so it is perhaps unwise to draw far-reaching conclu- sions from these data.

Impaired functionality o f second messenger systems

Even if receptor-mediated production of second messenger is preserved in Alzheimer's disease, a de facto functional im- pairment will be present if the second messengers themselves cannot function properly. The most dramatic of such changes was reported by Young et al., [17] who demonstrated large 50-90% deficits in the densities ofIns(1,4,5)P 3 binding sites in the Alzheimer brain. Even though their assay method was not optimal [18], their finding has been confirmed (Garlind et al., manuscript in preparation; see Fig. 2).

Ins(1,4,5)P 3 and Ins(1,3,4,5)P4, the latter of which is pro- duced from Ins(1,4,5)P 3 by a specific 3-kinase, act syner- gistically in mobilising calcium [19]. Such synergy raises the theoretical possibility that the deficit in Ins(1,4,5)P 3 receptors can be compensated by an increased function of Ins(1,3,4,5)P 4

receptors. However, no changes either in the density of [3H]Ins(1,3,4,5)P4 binding sites or in their affinity for this ligand were seen in Alzheimer's disease (Garlind et al., in press; see Fig. 2).

The other second messenger produced upon activation of the phosphoinositide pathway is diacylglycerol, which acti- vates protein kinase C isoforms (for review, see [20]). Dis- turbances in protein kinase C function inAlzheimer's disease have been reported in a number of studies (for reviews, see [21,221).

290

A . [3HlIns(1,4,5)P3 b i n d i n g

150-

e~ 120-

90 -

.~ 60- 7,

~= 30-

r

Temporal Frontai Cortex Cortex

T

C A C A

Cerebellum

T

II C A

B. [3H]Ins(1,3,4,5)P4 binding

1200-

g lO00-

~a 800-

~.~ 600-

"~ 4130- ,12

200 -

&

Temporal Frontal Cerebellum Cortex Cortex

T

C A C A C A

Fig. 2. [3H]Ins(l,4,5)P3 (Panel A) and [3H]Ins(l,3,4,5)P4 (Panel B) to synaptic membrane preparations from control (r-l) andAlzheimer's disease (~7-~) autopsy cases. In the case of [3H]Ins(l,4,5)P 3, the mean ligand concentrations were 1.73, 1.86 and 0.33 nM for temporal cortex, frontal cortex and cerebellum, respectively. The lower ligand concentration for the cerebellum is shown here to allow the data to be placed on the same graph. Similar large decreases were seen for the temporal cortex at 0.06 and 0.34 nM ligand concentrations; for the frontal cortex at 0.07 and 0.37 nM ligand concentrations; and for the cerebellum at 0.07 and 1.67 nM ligand concentrations. In the case of [3H]Ins(1,3,4,5)P4, the mean ligand concentrations were 0.62, 0.59 and 0.66 nM for temporal cortex, frontal cortex and cerebellum, respectively. Data are means 4- S.E.M., n = 6. Garlind et al., in press.

Consequences o f disturbed signal transduction mecha-

nisms in Alzhe imer ' s disease

From the above discussion, it is clear that there are a number of disturbances in the intracellular signalling pathways in post-mortem brain samples from individuals with Alzheim- er 's disease. I f the assumption is made that such changes are not confined to the final stages of the disease (see below), then two important consequences should be considered: a) revision of treatment strategies

Most of the treatment strategies for Alzheimer's disease have been based upon the idea ofneurotransmitter replace-

ment, whereby either a neurotransmitter precursor, direct acting agonist, or a neurotransmitter metabolism blocker have been given to the patient. Such an approach assumes that the receptors targeted by the drug are functionally intact. The experiments reviewed here indicate that this may not be the case for a number of neurotransmitter receptor types, thus providing an explanation for the gen- erally disappointing outcome of clinical trials based on the tenet of neurotransmitter replacement [23].

b) implication for 13-amyloid accumulation It is now clear that 13-amyloid accumulation is a central feature in the pathogenesis ofAlzheimer's disease. Genetic defects in the parent amyloid precursor protein have been found to cause some cases of familialAlzheimer's disease [24]. Cells transfected with amyloid precursor protein (APP) containing one of these familial mutations (i.e. the App67~ show alteredAPP metabolism favour- ing amyloid secretion [25, 26]. Recent work has demonstrated that amyloid metabolism

can be regulated by protein kinase C. Thus, Hung et aI. [27] reported that activation of protein kinase C decreased the formation in vitro of the 39-43 amino acid amyloid [3-protein in cells transfected with the [3APP 695 isoform. Such a result was seen both via direct activation with phorbol esters and indirectly via stimulation of muscarinic ml receptors cou- pled to the phosphoinositide signal transduction pathway [27]. I f protein kinase C activation controls [3-amyloid pro- duction, then a disturbance in signal transduction mecha- nisms may lead to a pathologically high level of [3-amyloid production. The deleterious effects of [3-amyloid in turn on cellular signalling (see [28]) would further exacerbate the situation.

Such a conclusion, however, relies on the assumption that the signal transduction deficits precede the amyloid precipi- tation found in the Alzheimer's disease brain. Masliah et al.

[29] have shown that diffuse plaques immunostained with anti-protein kinase C [3II antibodies, suggesting an early in- volvement of this isoform in Alzheimer's disease pathology. Evidence that adenylyl cyclase-mediated signal transduction may also be affected early in the course of the disease has come from a recent paper by Huang and Gibson [30] who demonstrated a dramatic deficit in cAMP production in re- sponse to isoprenaline, but not forskolin, stimulation in cul- tured fibroblasts fromAlzheimer's disease patients compared with controls. This group and others have previously reported thatAlzheimer's disease fibroblasts show a number of abnor- malities in calcium homeostasis and protein kinase function (see [31, 32]). Such findings implicate changes in signal transduction prior to the final stages of the disease. In ad- dition, they provide further support for the notion that a number of biochemical dysfunctions inAlzheimer's disease are not confined to the brain alone, but are also seen in the periphery.

A more experimental way of determining the relationship between signal transduction deficits and amyloid production is to use cultured cells. Cells transfected with 13APP isoforms have been reported [27] and are ideal in this respect. An al- ternative is the use of a cell line that will produce !3-amyloid under the right conditions. One such example is the human SH-SY5Y neuroblastoma cell line: following treatment of the cells with retinoic acid, which causes them to differentiate to a neuronal phenotype, there is a large increase in the mRNA coding for 13APP 695, peaking at 4 days after treatment [33]. Much shorter incubation times with retinoic acid, however, reduce the phosphoinositide breakdown response to mus- carinic receptor stimulation in these cells [34]. Whether such a result is purely coincidental or else is indicative that in some cells the signal transduction deficits precede the pathologi- cal changes in amyloid production awaits elucidation.

Acknowledgements

The authors would like to thank the Swedish Medical Re- search Council, Axelssons Johnssons, Gamin Tj~inarinnor, Sigurd and Elsa Goljes Minne, Torsten and Ragnar SOder- berg's and Hans and Loo Osterman's foundations for grants that made the studies reviewed here possible. In addition, we are indebted to Dr. Rivka Ravid and her colleagues at the Netherlands Brain Bank who provided much of the tissue used in these studies.

References

t. Morgan DG, May PC, Finch CE: Neurotransmitter receptors in normaI human aging and Alzheimer's disease. In: A.K. Sen and T. Lee (eds). Receptors and Ligands in Neurological Disorders. Cambridge University Press, Cambridge~ UK, 1988, pp 120-147

2. Cowburn RF, O'Neill C, Ravid R, Alafuzoff I, Winblad B, Fowler C J: Adenylyl cyclase activity in postmortem human brain: evidence of altered G protein mediation in Alzheimer's disease. J Neurochem 58: 1409-1419, 1992

3. O'Neill C, Wiehager B, Fowler C J, Ravid R, Winblad B, Cowburn RF: Regionally selective alterations in G protein subunit levels in the Alzheimer's disease brain. Brain Res 636: 193-201, 1994

4. Harrison PJ, Barton AJL, McDonald B, Pearson RCA: Alzheimer's disease: specific increases in a G protein subunit ( G ) mRNA in hippocampal and cortical neurons. Mol Brain Res 10: 71-81, 1991

5. Cowburn RF, Vestling M, Fowler C J, Ravid R, Winblad B, O'Neill C: Disrupted [3<adrenoceptor-G protein coupling in the temporal cortex of patients with Alzheimer's disease. Neurosci Letts 155: 163-166, 1993

6. Schnecko A, Witte K, Bohl J, Ohm T, Lemmer B: Adenylyl cyclase activity in Atzheimer's disease brain: stimulatory and inhibitory signal transduction pathways are differently affected. Brain Res 644: 291-296, 1994

7. Cowburn RF, Garlind A, O'Neill C, Alafuzoff l, Winblad B, Fowler C J: Characterization and regional distribution of adenylyl cyclase

291

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

activity from human brain. Neurochem Int 18: 389-398, 1991 8. Ross BM, McLaughlin M, Roberts M, Milligan G, McCullogh J,

K.nowler JT: Alterations in the activity of adenylate cyclase and high affinity GTPase in Alzheimer's disease. Brain Res 622: 35-42, 1993

9. Wang H-Y, Friedman E: Receptor-mediated activation of G proteins is reduced in postmortem brains from Alzheimer's disease patients. Neurosci Letts 173: 37-39, 1994 Cowburn RF, O'Neil l C, Ravid R, Winblad B, Fowler C J: Preservation of Gi-protein inhibited adenylyl cyctase activity in the brains of patients withAlzheimer's disease. Neurosci Letts 141: 16--20, 1992

O'Neill C, Fowler C J, Wiehager B, Cowburn RF, Alafuzoff 1, Winblad B: Coupling of human cerebral cortical %-adrenoeeptors to GTP- binding proteins in Alzheimer's disease. Brain Res. 563: 39-43, 1991 O'Neill C, Cowburn RF, Wiehager B, Alafuzoffl, Winblad B, Fowler C J: Preservation of 5-hydroxytryptamin%~ receptor-G protein interactions in the cerebral cortex of patients withAlzheimer's disease. Neurosci Letts 133: 15-19, 1991 GarlindA, Cowburn RE Wiehager B, Ravid R, Winblad B, Fowler C J: Preservation of Kappa 1 opioid receptor recognition site density and regulation by G-proteins in the temporal cortex of patients with Alzheimer's disease. Neurosci Letts 185:131-134, 1995 Schulz G, Rosenthal W, Herscheler J, Trautwein W: Role of G proteins in calcium channel modulation. Ann Rev Physiol 52: 275-292, 1990 Watkins DC, Moxham CM, Morris A J, Malbon CC: Suppression of G~ 2 enhances phospholipase C signalling. Biochem J 299: 593-596, 1994 Ferrari-DiLeo G, Flynn DD: Diminished muscarinic receptor~ stimulated [3H]-PIP2 hydrolysis in Alzheimer's disease. Life Sci 53: PL439-PL444, 1993 Young LT, Kish S J, Li PP, Warsh J: Decreased brain 3H-inositol 1,4,5- trisphosphate binding in Alzheimer's disease. Neurosci Letts 94: 198-202, 1988 Garlind A, Cowburn RF, Fowler C J: Characterization of [3H]inositol 1,4,5-trisphosphate binding sites in human temporal cortical and cerebellar membranes. Neurochem Int 24: 73-80, I994 Downes CP, MacPhee CH: myo-lnositol metabolites as cellular signals. Eur J Biochem 193:1-18, 1990 Nishizuka Y: The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature 334:66 l~565 Saitoh T, Cole G, Huynh TV: Aberrant protein kinase C cascades in Alzheimer's disease. In: J.M. Lauder (ed.). Molecular Aspects of Development and Aging of the Nervous System. Plenum Press, New York, 1990, pp 301-310 Rogue P: Prot6ine kinase C et maladie d'Alzheimer. Alzheimer Actualites No. 74: 6-8, 1993 Whalley L J: Drug treatments of dementia. Br J Psychiat 155:595-611, 1989 Ashall F, Goate AM: Role of the [3-amyloid precursor protein in Alzheimer's disease. TIBS 19: 42-46, 1994 Citron M, Oltersdorf T, Haass C, McConlogue L, Hung AY, Seubert P, Vigo-Pelfrey C, Lieberburg I, Selkoe D J: Mutation of the [3-amyloid precursor protein in familial Alzheimer's disease increases [3-protein production. Nature 360: 672~574, 1992 Cai X-D, GoldeTE, Younkin SM: Release of excess amyloid [3 protein from a mutant amyloid [3 protein precursor. Science 259: 514-516, 1993 Hung AY, Haass C, Nitsch RM, Qiu WQ, Citron M, Wurtman R J, Growdon JH, Selkoe DJ: Activation of protein kinase C inhibits cellular production of the amyloid [3-protein. J Biol Chem 268: 22959--22962, 1993 Takenouchi T, Munekata E: Inhibitory effects of [3-amyloid peptides on nicotine-induced Ca 2+ influx in PCI2h cells in culture. Neurosci

292

Letts 173: 147-150, 1994 29. Masliah E, Cole GM, Hansen LA, Mallory M, Albright T, Terry RD,

Saitoh T: Protein kinase C alteration is an early biochemical marker in AIzheimer's disease. J Neurosci 11: 2759-2767, 1991

30. Huang H-M, Gibson GE: Altered ~-adrenergic receptor-stimulated cAMP formation in cultured skin fibroblasts from Alzheimer donors. J Biol Chem 268: 14616-14621, 1993

31. Blass JP, Gibson GE, Black RS, Ko L-w, Sheu K-FR: Cultured skin fibroblasts, 'Alzheimer antigens,' and cellular calcium homeostasis in Alzheimer's disease. In: C.J. Fowler, L.A. Carlson, C.-G. Gottfries and B. Winblad (eds). Biological Markers in Dementia of Alzheimer Type. Smith Gordon and Co Ltd, London, 1990, pp 203-214

32. BruelA, Cherqui G, Cotumelli S, Margelin D, Roudier M, Sinet P-M, Prieur M, Perignon J-L, Delabar J: Reduced protein kinase C in sporadic Alzheimer's disease fibroblasts. Neurosci Letts 133: 89-92, 1991

33. Konig G, Masters CL, Beyreuther K: Retinoic acid induced differentiated neuroblastoma cells show increased expression of the [3A4 amyloid gene of Alzheimer's disease and an altered splicing pattern. FEBS Letts 269: 305-310, 1990

34. Fowler CJ, O'Neill C, Almqvist P, Nilsson S, Wiehager B, Winblad B: Muscarinic receptors coupled to inositol phospholipid breakdown in human SH-SY5Y neuroblastoma cells: effect of retinoic acid- induced differentiation. Neurochem Int 15: 73-79, 1989