Molecular Brain Research, 10 (1991) 71-81 © 1991 Elsevier Science Publishers B.V. (Biomedical Division) 0169-328X/91/$03.50 ADONIS 0169328X9170283W

71

BRESM 70283

Alzheimer's disease: specific increases in a G protein subunit (Gsa) m R N A in hippocampal and cortical neurons P.J. Harrison 1'2, A.J.L. Barton 3, B. McDonald 4 and R . C . A . Pearson 3 Departments of lAnatomy and 2Psychiatry, St. Mary's Hospital Medical School, London (U.K.), 3Department of Biomedical Science, The University of Sheffield (U.K.) and ¢Department of Neuropathology, Radcliffe Infirmary, Oxford (U.K.) (Accepted 4 December 1990)

Key words: Dementia; GTP binding protein; Gene expression; In situ hybridization histochemistry; Second messenger; Signal transduction

The GTP binding protein, Gs, activates adenyl cyclase in direct response to stimulation of several neurotransmitter receptors. In situ hybridization histochemistry (ISHH) with a asS-labelled oligonucleotide has been used to detect the mRNA encoding the a subunit of G s (Gsa) in human hippocampus, temporal and visual cortices and cerebellum, and its level has been compared between Alzheimer's disease (AD) and control brains. A marked regional increase was found in the hippocampus of AD cases. Analysis of levels of Gsa mRNA in individual constituent pyramidal cells confirmed this increase (3 to 4-fold in densitometric units) in hippocampal fields CA1, CA3 and CA4, as well as in temporal cortex. Levels of Gsa mRNA were also determined relative to total poly(A)+ mRNA in the same cell populations in each case. Gene-specific elevation of Gsa mRNA was thereby confirmed in hippocampal fields, and also in temporal cortex. No changes were seen in visual cortex. The increase in Gsa mRNA may represent a response by AD neurons in affected areas to receptor alterations, or to an abnormality in receptor-G protein coupling. Alternatively, altered G protein gene expression might be a pathogenic event underlying changes in linked receptor populations. INTRODUCTION The G proteins are a family of cytoplasmic-facing membrane-associated proteins that play a central role in signal transduction. They share the ability to hydrolyze GTP, an event producing a conformational change in the protein which in turn causes stimulation or inhibition of linked effector molecules 2°'23"44"52'6°. G proteins are directly coupled to many receptors 11, such that occupancy of the receptor by its ligand activates the G protein and sets in motion a chain of events within the cell. Three main receptor-linked G proteins are known, Gs, G i and G o , encoded by separate genes and which differ in their associated receptors and second messengers 62. G s is so called because it stimulates adenyl cyclase, and thereby promotes many phosphorylation reactions, in turn producing diverse cellular responses. Neurotransmitter receptors which are coupled to G s include fl-adrenergic (both fll and f12), 5-HT2, dopamine (D 0 , histamine (H2), adenosine (A2), vasoactive intestinal polypeptide (VIP) and certain neuropeptides, and possibly one of the muscarinic receptors (M2) 3. Given that multiple receptors are affected in A D brain, including some of those which are linked to G s (see discussion), a systematic study in A D of the G proteins is indicated.

Such work may help clarify c o m m o n mechanisms underlying receptor abnormalities, as well as indicate the intactness of intracellular signalling in postsynaptic neurons. G proteins are heterotrimers, sharing two subunits (fl and ),), but differing in the composition of the a subunit. In the case of Gs, the a subunit exists in two main forms, of molecular weight 45 and 52 kDa, arising from differential splicing of a c o m m o n m R N A 9'51 whose c D N A has been sequenced 9'25'33. Other variants of G s are also recognised 31'37. The specificity of the alpha subunit for each G protein allows hybridization probes directed against a particular G a species to be used to determine the pattern and amount of expression of m R N A encoding that m e m b e r of the G protein family. The present study has used I S H H with an antisense oligonucleotide to focus initially on the m R N A encoding the alpha subunit of Gs, Gsa, in the control and A D brain. In this way, detailed information can be obtained as to the nature and extent of changes in this m R N A in neurons in A D . MATERIALS AND METHODS The details of the cases used in this study are given in Table I. All

Correspondence: R.C.A. Pearson, Department of Biomedical Science, The University, Sheffield S10 2TN, U.K.

72 AD cases were diagnosed clinically and satisfied standard histopathological criteria 36. Senile plaque counts, made in temporal cortex (Brodmann area 21), were available for 16 of the AD cases. The non-AD dementias comprised Pick's disease, Huntington's chorea, Gerstmann-Str/iussler syndrome and a non-AD familial dementia lacking definitive neuropathology. All cases were rated for agonai state, since this has been shown to affect preservation of some mRNAs 28. Agonal state was determined by case note analysis. A score of 1 was given where death was rapid, 4 if death occurred after prolonged coma, and 2 or 3 for intermediate cases which had suffered brief coma, fever or dehydration. At post mortem, blocks were taken from hippocampus, middle temporal gyrus (area 21), visual cortex (area 17) and cerebellar cortex, although not all areas were available from each case. Blocks were embedded and stored at -70 °C for 3-18 months. Ten-/zm-thick fresh frozen sections were collected on subbed slides and processed for in situ hybridization histochemistry (ISHH) as previously described 26'43. For hybridization, 1.6 x 10 6 cpm of labelled probe were added to each section in 100/d of buffer43. Sections from each region were hybridized in triplicate. Incubation was carried out for 18 h at 35 °C, followed by washing steps in 1 × SSC (3 x 20 min at 57 °C, 2 x 60 min at room temperature). The probe consisted of a 45 mer directed against bases 476-520 of the rat Gsa sequence 33, corresponding to bases 309-353 of the human cDNA 25, and is predicted to detect both major Gsa mRNA transcripts. The same probe has been used previously for ISHH and is directed against a r e # o n of the gene divergent from the other G a mRNAs 39. The published human Gsa cDNA shows two mismatches from this sequence25; the hybridization and wash temperature given above allow this mismatch, and are 15 °C below the corrected melt temperature ~6. Controls comprised the following: (1) The specificity of the probe was determined by Northern blotting as described43 using [32p]. dATP-labelled probe hybridized to 25/~g total RNA extracted 12from the frontal cortex of an AD patient. (2) Concurrent hybridization to adjacent sections under identical conditions with the complementary sense strand oligonucleotide. (3) Pretreatment of sections with ribonuclease43. (4) Addition of 100-fold excess unlabelled probe. Hybridized sections were placed against tritium-sensitive film (Hyperfilm betamax, Amersham) at room temperature for 21 days to generate autoradiograms. After resubbing in 0.1% gelatin/0.01% chrom alum, sections were subsequently dipped in emulsion (Ilford K5, mixed 1:1 with 2% glycerol at 43 °C) and exposed for 63 days at 4 °C. Finally, sections were lightly counterstained with toluidine blue and coverslipped. Quantitative assessment of Gsa mRNA was made at both the regional and cellular level using an image analysis system (Image Manager PC, Sight Systems) essentially as previously described 2' 27,29. Briefly, on autoradiograms, mean grey density (MGD) was calculated, blind to case details, through the depth of the grey

matter in temporal and visual cortex, over the granule cell layer of the cerebellum, over the stratum granulosum of the dentate gyrus, and over the stratum pyramidale of hippocampal fields CA4, CA3 and CA1. The mean value from each set of triplicates was used for subsequent calculations. The original MGD values, from 0, (black, maximum signal), to 255 (white, no signal), were divided into 255 and converted to a logarithm. This produces a linear scale (data not shown), whereby a doubling of numerical value corresponds to a doubling of radioactivity over the range of grey densities occurring in these experiments. The value for the sense strand hybridization, representing background signal, was also converted to a logarithm and subtracted from the antisense-derived value. The final measure, called 'corrected grey density', is used to produce the units in Figures 5 and 6, and is thus given by the formula: Corrected grey density = loglo(255/MGDantisense) - Ioglo(255/MGD~n~) In order to assess cellular Gsa mRNA content, dipped sections were viewed under dark field microscopy and the image projected onto a video monitor. The MGD was determined over 20-30 pyramidal neurons in each area, except in cerebellum where Purkinje cells were measured. Cell types were identified on the basis of their size, shape and position under bright-field illumination. The sense strand-derived signal was again subtracted from each antisense value. Cellular MGD was calibrated against grain counts per cell, and showed a linear relationshipZ9; thus, MGD at a cellular level was not converted to a logarithmic scale. Finally, the mean neuronal areas measured for the determination of MGD was calculated in each case, to identify possible differences in cell size between cases and controls which might confound interpretation of MGD in each group. In addition, levels of Gsa mRNA were related to total polyadenylated mRNA in each region or cell population for every case where poly(A) ÷ mRNA data were available. Poly(A) ÷ mRNA quantitation was determined using oligo(dT) ISHH as described 27' zg. Data are thus presented both for Gsa mRNA alone, and Gsa mRNA as a ratio of poly(A) ÷ mRNA, in order to identify whether changes in quantities of the specific mRNA are parallel to, or differentially altered from, that of poly(A) ÷ mRNA as a whole. Data was analysed using two-tailed t-tests and Spearman's correlation.

RESULTS Gsa mRNA

w a s d e t e c t e d in e a c h a r e a i n v e s t i g a t e d .

Specificity o f h y b r i d i z a t i o n w a s c o n f i r m e d b y t h e c o n trols: (1) N o r t h e r n b l o t t i n g d e t e c t e d a s t r o n g b a n d o f 1.9 k b a n d a w e a k e r b a n d o f 1.6 k b (Fig. 1), c o r r e s p o n d i n g to t h e m a j o r G s a

TABLE I

produced Clinicopathological details o f cases

a uniform,

m i n i m a l l e v e l o f s i g n a l in e a c h

r e g i o n (Figs. 2 D a n d 3 B ) . (3) R i b o n u c l e a s e p r e t r e a t m e n t

Values are mean + S.E.M. aSee text for diagnoses, bSee text for details. Agonal state not known for 1 AD case or for 2 non-AD dementias.

o r e x c e s s u n l a b e l l e d p r o b e a b o l i s h e d s i g n i f i c a n t signal (not shown). Molecular

Number Age(y) Range Sex Post mortem interval (h) Range Agonalstate score b

t r a n s c r i p t s . (2) S e n s e s t r a n d I S H H

Alzheimer's Normal disease controls

Non-AD dementias a

22 73+2.8 44-92 9M, 13F 38.8 + 4.6 6-86 2.4 + 0.2

4 64+2 57-74 1M, 3F 26 + 3 19-36 2.0

13 69.2+3.2 51-96 9M, 4F 41.3 + 8.1 7-70 2.6 + 0.1

other

Additionally,

Biology

known

a search of the European

Library

human

mRNA

database (G

showed

protein

that no

subunit

or

o t h e r w i s e ) c o n t a i n s a 45 b a s e s e q u e n c e 6 o r less mismatches from the Gsa oligonucleotide. Together, these c o n t r o l s e n s u r e t h a t t h e signal r e s u l t s f r o m h y b r i d i z a t i o n to G s a m R N A a n d n o o t h e r . Signal was c o n s i s t e n t w i t h i n e a c h set o f t r i p l i c a t e h y b r i d i z a t i o n s , w i t h a n i n t e r - s e c t i o n v a r i a t i o n o f less t h a n 2 0 % , a n d u s u a l y less t h a n 1 0 % . N o c o r r e l a t i o n s w e r e

73 seen between signal strength and age, storage time, agonal state or post mortem interval. Neither did temporal cortex Gsa mRNA within the AD group show an association with senile plaque counts made in this area.

Distribution of Gsa mRNA in human brain The regional localization of Gsa mRNA in normal brain is in broad agreement with that described by Mengod et al. 41 who used a 32p-labelled cDNA probe. In cerebellum, signal is concentrated over .the granule cell layer, with minimal signal over the molecular layer (Fig. 2F). In the hippocampal formation, signal is present over the stratum granulosum of the dentate gyrus, and within the stratum pyramidale of CA fields and subiculum (Fig. 2A,B). Temporal cortex shows positive hybridization throughout the grey matter, with enhanced signal over laminae III/V (Fig. 2C). This distribution is similar in visual cortex, but with higher overall levels and a more pronounced accentuation in lamina V/VI (Fig. 2E); in this regard, our findings differ from Mengod and colleagues 41, who found a more homogeneous laminar distribution in area 17. No signal above background was observed in the white matter. As predicted from cerebellar autoradiograms, signal at the cellular level was seen over granule cells, but was also present over Purkinje cells (which do not show significant Gsa mRNA in the rat39), together with labelling of many scattered cells in the molecular layer (Fig. 4B). In the hippocampus, granule cells of the dentate gyrus showed positive hybridization, as did many pyramidal cells and some smaller neurons in the CA fields (Fig. 3A,C). In the cortical areas, signal was greatest over pyramidal cells, although many smaller cells were also labelled. Notably, in both hippocampus and neocortex, not all pyramidal cells were labelled to the same degree. Sometimes a heavily-labelled pyramidal neuron was present next to a moderately or minimally labelled one, especially in visual cortex (Fig. 4A), indicating possible heterogeneity of pyramidal cells in terms of their Gsa expression. Dendritic localization of Gsa mRNA, as reported for some other mRNAs 4'21, is suggested by the presence of grains over proximal dendrites (Fig. 3). Signal over the neuropil was consistently greater than background, also supporting the possibility of dendritic (and glial) Gsa mRNA occurring in un-counterstained cell processes (Figs. 3 and 4A). No hybridization signal was seen over vasculature or pia. A marked individual variability in hybridization signal was seen between different regions of a single brain as well as between one brain and another. For example, in some cases, no signal was seen over the dentate gyrus despite strong signal in CA fields; in others, one of the

Fig. 1. Hybridization of [32p]dATP-labelled Gsa probe to a Northern blot of RNA from AD frontal cortex showing doublet band of 1.6 and 1.9 kb.

CA fields would show much stronger hybridization than the rest. Other brains showed strong signal in the neocortex but weak signal in cerebellum, or vice versa. These differences exceed the brain-to-brain variation seen with all mRNA studies of human post mortem tissue, including results with other probes used for ISHH on the present collection (refs. 2, 27, 29, and unpublished observations). This variation could not be explained by any identifiable factor such as age, sex, agonal state, post mortem delay or total poly(A) + mRNA content. It may therefore reflect genuine differences in expression of Gsa mRNA between individuals, although correlation with another, unidentified, confounding variable (e.g. corticosteroid levels53) cannot be ruled out. AD cases showed similar qualitative variation, and no differences in the spatial distribution of Gsa mRNA between AD cases and controls were found.

Quantitative analysis of autoradiograms Despite such variations, on a regional basis, significant increases in Gsa mRNA in AD were found in the hippocampus, but not in temporal and visual cortices or cerebellum, although the latter two areas showed a similar trend (Fig. 5). In densitometric terms, the increases were: dentate gyrus, x l . 9 (P < 0.02); CA4, x3.1 (P < 0.01); CA3, x3.1 (P < 0.01) and CA1, x4.4 (P < 0.02). Expressed as a ratio of poly(A) ÷ mRNA in each region, Gsa mRNA in the CA fields shows a similar

74

Fig. 2. Regional distribution of Gsa mRNA in human brain. A: hippocampus (control); B: hippocampus (AD); DG, stratum granulosum of the dentate gyrus; CA3, stratum pyramidale of the CA3 field. C: temporal cortex (AD). D: temporal cortex, sense strand hybridization. E: visual cortex (control); F: cerebellum. Pictures are printed directly from autoradiograms; increasing whiteness indicates increasing hybridization.

75 increase, although of greater magnitude in CA1 ( x 6 , P < 0.01). Relative to poly(A) + m R N A , the trend towards increased Gs a m R N A in the cerebellum in A D becomes significant ( x 2 . 0 , P < 0.01), since cerebellar poly(A) +

m R N A is also reduced in A D cases 29. The few n o n - A D dementias showed similar trends to that seen in the A D cases, both when expressed as Gs a m R N A alone or as G s a m R N A / p o l y ( A ) + m R N A (Figs.

Fig. 3. Cellular localization of Gsa mRNA in control hippocampus (A-C) and AD temporal cortex (D). A: CA3; B: CA3, sense strand hybridization; C: CA1,; D: temporal cortex (lamina V). Note the increased grains present over the neuropil in A, C and D compared to B, and the grains over proximal dendrites (arrowed in C and D), suggesting dendritic localization of Gsa mRNA. Bar = 25/zm.

76

4

B 0

A

77

0.6000.500

2.400]

t

z ~ 2.000]

0.400J

s

0.200]

ii!

~~,~~d~1.200] @

iiii11

CA1

Fig. 5. Quantitation of Gsa autoradiograms. Open bars: normal controls; hatched bars: AD; solid bars: non-AD dementias. Values are mean + S.E.M. CA1, CA3, CA4, stratum pyramidale of hippocampal CA fields; DG, stratum granulosum of the dentate gyms; TC, middle temporal gyrus (area 21); VC, visual cortex (area 17); Cer, cerebellum. In hippoeampal fields, n = 12 (controls), n = 13 (AD), n = 3 (non-AD dementias). In neocnrtical areas, n = 13 (controls), n = 17 (AD), n = 4 (non-AD dementias). In cerebellum, n = 13 (controls), n = 21 (AD), n = 3 (non-AD dementias). **P < 0.02, ***P < 0.01, AD vs. controls.

5 and 6). No statistical analysis of these cases was attempted due to their small number.

0.8001 CA1

li

CA3 CA4 DG TC VC

Cer

Fig. 6. Quantitation of autoradiograms, expressing G s a m R N A as a ratio of poly(A) + m R N A . Abbreviations and symbols as for Fig. 5. In hippocampus, n = 12 (controls), n = 10 (AD), n = 3 (non-AD dementias). In temporal cortex, n = 11 (controls), n = 17 (AD), n = 3 (non-AD dementias). In cerebellum, n = 13 (controls), n = 19 (AD), n = 3 (non-AD dementias). *P < 0.05, **P < 0.02, ***P < 0.01, A D vs. controls.

the high packing density of granule cells. In A D temporal cortex, pyramidal cells also contained significantly more G s a m R N A (x2.7, P < 0.05). In cerebellum, Purkinje cells in A D cases had unchanged G s a m R N A . Similar trends were also present for the n o n - A D dementias (Fig.

Quantitative analysis of Gs a m R N A in neurons

7).

The particular value of I S H H in the study of gene expression compared to other techniques is its ability to address gene expression, at a cellular resolution, and thus provide a much finer grain of detail than is possible with regional analyses such as Northern blotting 26. Moreover, quantitation of autoradiograms such as that described above, although valuable, is affected by cell density and changes in cell populations (for example, the ratio of pyramidal cells to glia), both of which parameters may change in A D . Thus, analysis of G s a m R N A per pyramidal neuron in each area assessed above was performed. The area of measured neurons did not vary in any region between A D and controls (data not shown); thus, differences in cellular grey density between these groups reflect differences in absolute mean grain count per cell. In the hippocampus, the location and extent of increased G s a m R N A was similar per pyramidal cell to that seen regionally (Fig. 7). Thus, in CA4, A D cases had a mean increase compared to controls of x2.3 (P < 0.01), in CA3 the increase was x 1.9 (P < 0.05), and in CA1, x2.7 (P < 0.01). The dentate gyrus could not be measured at the cellular level because of its small size and

When related to poly(A) ÷ m R N A content of pyramidal cells in each region (Fig. 8), G s a m R N A showed significant increases in pyramidal neurons of the C A fields (all approximately 4-fold, P < 0.01) and in temporal cortex ( x 3.1, P < 0.02). No significant changes in G s a m R N A were seen in visual cortex. No cellular poly(A) ÷ m R N A data were available for Purkinje cells, or for the n o n - A D dementias. DISCUSSION

The data demonstrate significant increases in m R N A encoding the alpha subunit of the stimulatory G protein, G s, in A D brain. These increases are present in pyramidal cells, with the greatest elevation in hippocampal C A fields, a moderate increase in temporal cortex, and no change in visual cortex. The regional selectivity of the findings parallels the extent of disease involvement of these areas, and suggests that the increases are related in some. way to the structural pathology of A D . However, the nature of this relationship is unclear and likely to be complex, given that G s a m R N A levels did not correlate with severity of pathology as determined by senile plaque

Fig. 4. Cellular distribution of G s a m R N A in (A) visual cortex (lamina V/VI); (B) cerebellum. In A, some pyramidal cells show heavy labelling (h), whilst others show moderate or light labelling (m). In B, both Purkinje cells (pc) and cells in the molecular layer (arrowed) show positive hybridization. Signal over the granule cell layer (gcl) is barely visible due to counterstaining. Bar = 25/xm.

78 110100 9O

80z

70

~

60-

~

50-

~

40-

L4 C~

:~ 3020 10CA1

CA3

CA4

TC

VC

Cer

Fig. 7. Q u a n t i t a t i o n of G s a m R N A h y b r i d i z a t i o n per cell. A b b r e v i a t i o n s and s y m b o l s as for Fig. 5. In e a c h a r e a , grain d e n s i t y was m e a s u r e d o v e r p y r a m i d a l cells, e x c e p t in c e r e b e l l u m w h e r e P u r k i n j e cells w e r e m e a s u r e d . In C A 1 , n = 11 (controls), n = 9 ( A D ) , n = 3 ( n o n - A D d e m e n t i a s ) ; in C A 3 , n = 9 (controls), n = 8 ( A D ) , n = 3 ( n o n - A D d e m e n t i a s ) ; in C A 4 , n = 10 (controls), n = 12 ( A D ) , n = 3 ( n o n - A D d e m e n t i a s ) . N u m b e r s v a r y t h r o u g h the C A fields due to p l a n e of section. In n e o c o r t i c a l areas, n = 10 (controls), n = 13 ( A D ) , n = 4 ( n o n - A D d e m e n t i a s ) . In c e r e b e l l u m , n = 12 (controls), n = 21 ( A D ) . * P < 0.05, ***P < 0.01, A D vs. controls.

counts, and given that an increase also occurred in the cerebellum, whose involvement in AD is less prominent.

Mechanisms underlying the increase in Gsa mRNA It is likely that enhanced Gsa mRNA reflects a response to changes in receptor densities or affinities, with the convergence of multiple receptors on Gs amplifying such a response and accounting for the magnitude of the increases. Data concerning Gs-linked receptors in AD are, however, limited, and do not indicate a uniform alteration. Reductions in 5-HT recep-

1300 ] 1.6001

o o.soo 1

~

o.4oo1 CA3

CA4

m TC

VC

Fig. 8. Q u a n t i t a t i o n of c e l l u l a r G s a m R N A relative to p o l y ( A ) + m R N A c o n t e n t of the s a m e cell p o p u l a t i o n . A b b r e v i a t i o n s and s y m b o l s as for Fig. 5. In C A I , n = 10 (controls), n = 6 ( A D ) ; in C A 3 , n = 9 (controls), n = 6 ( A D ) ; in C A 4 , n = 10 (controls), n = 8 ( A D ) . In t e m p o r a l cortex, n = 8 (controls), n = 11 ( A D ) . In visual c o r t e x , n = 8 (controls), n = 14 ( A D ) . A D and control g r o u p s r e m a i n m a t c h e d for age a n d post m o r t e m delay in e a c h area. **P < 0.02, ***P < 0.01.

tors are well established 6'7J4'49, whereas fl-receptors show no overall change 6'35 although f12 subtypes may be increased 35'55. The other receptors have not been investigated in any detail. Even if Gs-linked receptor densities show no consistent changes, alterations may occur in the coupling of a receptor to G s. In support of this, Smith et al. 56 reported preliminary evidence of impaired G protein interactions with muscarinic receptors in AD hippocampus, and de Keyser et al. 17 have shown abnormal G protein-D1 receptor function in AD frontal cortex. It is possible that such changes are related to the increased membrane fluidity which occurs in AD cases T M . However, all suggestions of putative compensatory or responsive mechanisms must remain tentative since there is little information concerning the effect of receptor down- or up-regulation on G proteins. More investigation of these processes and their possible derangement in disease is indicated 54. Given the similar increases in Gsa mRNA seen in the few non-AD dementias, the possibility that the changes represent a reaction to chronic neuronal insults or deafferentation must be considered. Such a response could be beneficial or detrimental. Similar suggestions have been put forward to account for elevations of tubulin 22, tau 2 and muscarinic receptor m R N A s 27"28 in AD, as well as for increased protein kinase C isoenzymes4°. Interestingly, cardiac Gsa and Gia mRNAs (but not Goa) are elevated in heart failure Is, whilst experimental lesions in the rat brain are also accompanied by increases in Gsa mRNA in surrounding neurons (A. Najlerahim, P,J.H., A.J.L.B. and R.C.A.P., unpublished observations). These findings support the interpretation that elevations in G protein expression may accompany cellular injury. Alternatively, it is possible that abnormalities in G proteins, including their expression, might be the proximate cause of, not response to, multiple changes in their linked receptors. A single change in G s regulation would be expected to have multiple and diverse consequences for the cell. Several human diseases are now recognised in which mutations or aberrant expression of a G protein are a central aetiological factor 38,57. Finally, it is becoming apparent that G proteins are involved in cellular processes apart from classical signal transduction at the cell membrane 5'45, and their structural and functional similarities with ras oncogene products and growth proteins are of interest 1"3°"59 Whilst Gs has yet to be directly implicated in any of these additional processes, they demonstrate the possibility that it may have such roles, and therefore that the observed increases in AD could be related to alterations in any of these additional functions.

79 Consequences o f increased Gsa m R N A

Whatever the reason for the change in Gsa m R N A , it is probable that affected cells will have significant consequences arising from it. An increase in m R N A of this magnitude, assuming it is not matched by reductions in translational activity, would be predicted to be accompanied by elevations in G s protein synthesis, which in turn would enhance the diverse functions subserved by this protein. For example, abnormal and excessive phosphorylation of proteins is a feature of AD 24'63, and this would be promoted by elevated adenyl cyclase. Direct measurements of adenyl cyclase activities in A D indeed suggest that increases occur, in keeping with the present data 15 (but see ref. 46). Since G proteins themselves are a target for phosphorylation, which enhances their activity 47, increased G s synthesis may promote G s phosphorylation and exacerbate the dysfunction. Moreover, Gs is stimulated by aluminofluoride 6°, providing another potential mechanism for phosphorylation, which is of interest given that aluminium has been implicated in A D 48'5° and is known to promote phosphorylation of cytoskeletal proteins 32. Since G s is now thought to activate calcium channels in addition to its effect on adenyl cyclase ~°'58, an alternative consequence of increased G s might be the promotion of calcium influx into neurons, contributing to neurotoxicity ~3. The importance of non-cAMP mediated effects of G s are indirectly supported by the poor concordance between the localization of Gsa m R N A and that of adenyl cyclase (present study, and refs. 8, 39, 41). Conclusions

The results of the present study are clearcut, but the interpretations of them are many and speculative. It therefore seems appropriate to end with a number of suggestions for future work to resolve these uncertainties. A detailed study of all G proteins, including other G a species and the fl and y subunits, is indicated to determine the specificity of the alterations in AD. Should REFERENCES 1 Barbacid, M., ras genes, Annu. Rev. Biochem., 56 (1987) 779-827. 2 Barton, A.J.L., Harrison, P.J., Najlerahim, A. et al., Increased tau messenger RNA in Alzheimer's disease hippocampus, Am. J. Pathol., 137 (1990) 497-502. 3 Birnbaumer, L., Abramowitz, J. and Brown, A.M., Receptoreffector coupling by G proteins, Biochim. Biophys. Acta, 1031 (1990) 163-224. 4 Bloch, B., Guitteny, A.F., Normand, E. and Chouham, S., Presence of neuropeptide messenger RNAs in neuronal processes, Neurosci. Lett., 109 (1990) 259-264. 5 Bourne, H.R., Do GTPases direct membrane traffic in secretion?, Cell, 53 (1988) 669-671. 6 Bowen, D.M., Allen, S.J., Benton, J.S. et al., Biochemical assessment of serotonergic and cholinergic dysfunction and

all G protein subunits show increases in their mRNAs, it is less likely that any one is central to the pathogenesis of the disease; conversely, should the others show no change, or are found to be decreased, the significance of the present data will be enhanced. I S H H using probes specific to the alternative splice variants of Gs a m R N A will identify detailed changes in their differential expression, which may be important given the non-identical properties of the encoded isoforms 2°'34. Correlating I S H H with immunocytochemical detection of G protein subunits will clarify the relationship between m R N A and protein, whilst studies with double-labelling I S H H will show colocalization of G protein m R N A species in individual cells. Identification of features which distinguish strongly Gsa-positive pyramidal neurons from others will also be valuable, and may relate to other measures of vulnerability to degeneration 42,61. Neurochemical studies in A D have focussed upon transmitters, their synthetic enzymes, and their receptors. The demonstration that a major G protein is also significantly altered by the disease emphasizes the need to investigate other steps in neuronal signalling, since the crucial abnormality may be at any point in the process 19. The data also emphasize that the functional integrity of such a system is not guaranteed merely by unchanged transmitter and receptor parameters. Indeed, changes in second messengers such as G proteins, which serve as convergent, divergent and integrative signalling molecules 52, may be of special importance. This, in turn, has implications for drug therapies which may be considered, since replacement strategies are unlikely to be successful in the presence of major alterations at any point along the signal transduction pathway.

Acknowledgements. This work was supported by the Wellcome Trust (U.K.). P.J.H. is a Medical Research Council (U.K.) Training Fellow. A.J.L.B. is Cottrell Fellow of Research into Aging. We thank Dr. Margaret Esiri for additional neuropathology and Josie Heffernan for expert technical assistance.

cerebral atrophy in Alzheimer's disease, J. Neurochem., 41 (1983) 266-272. 7 Bowen, D.M., Najlerahim, A., Procter, A.W., Francis, P.T. and Murphy, E., Circumscribed changes of the cerebral cortex in neuropsychiatric disorders of later life, Proc. Natl. Acad. Sci. U.S.A., 86 (1989) 9504-9508. 8 Brann, M.R., Collins, R.M. and Spiegel, A., Localization of mRNAs encoding the a-subunits of signal-transducing G-proteins within rat brain and among peripheral tissues, FEBS Lett., 222 (1987) 191-198. 9 Bray, P., Carter, A., Simons, C. et al., Human eDNA clones for four species of G~,s signal transduction proteins, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 8893-8897. 10 Brown, A.M. and Birnbaumer, L., Ionic channels and their regulation by G protein subunits, Annu. Rev. Physiol., 52 (1990) 197-213. 11 Caron. M.G., The guanine nucleotide regulatory protein-

80 coupled receptors for nucleosides, nucleotides, amino acids and amine neurotransmitters, Curr. Opinion Cell Biol., 1 (1989) 159-166. 12 Chirgwin, J.M., Przybyla, A.E., MacDonald, R. and Rutter, W.J., Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease, Biochemistry, 18 (1979) 125140. 13 Choi, D.W., Glutamate neurotoxicity and diseases of the nervous system, Neuron, 1 (1988) 623-634. 14 Cross, A.J., Serotonin in neurodegenerative disorders. In N.N. Osborne and M. Hamon (Eds.), Neuronal Serotonin, Wiley, New York, 1988, pp. 231-253. 15 Danielsson, E., Eckern~is, S.-A., Westlind-Danielsson, A. et al., VIP-sensitive adenylate cyclase, guanylate cyclase, muscarinic receptors, choline acetyltransferase and anticholinesterase, in brain tissue afflicted by Alzheimer's disease/senile dementia of the Alzheimer type, Neurobiol. Aging, 9 (1988) 153-162. 16 Davis, L.G., Dibner, M.D. and Batty, J.F., Basic Methods, in Molecular Biology, Elsevier, New York, 1986. 17 De Keyser, J., Ebinger, G. and Vauqelin, G., Dl-dopamine receptor abnormality in frontal cortex points to a functional alteration of cortical cell membranes in Alzheimer's disease, Arch. NeuroL, 47 (1990) 761-763. 18 Feldman, A.M., Cares, A.E., Bristow, M.R. and van Dop, C., Altered expression of a-subunits of G proteins in failing human hearts, J. Mol. Cell. Cardiol., 21 (1989) 359-365. 19 Fowler, C.J., O'Neill, C.,, Garlind, A. and Cowburn, R.E, Alzheimer's disease: is there a problem beyond recognition?, Trends Neurosci., 11 (1990) 183-184. 2(1 Freissmuth, M., Casey, P.J. and Gilman, A.G., G proteins control diverse pathways of transmembrane signalling, FASEB J., 3 (1989) 2125-2131. 21 Garner, C.C., Tucker, R.P. and Matus, A., Selective localization of messenger RNA for cytoskeletal protein MAP2 in dendrites, Nature, 336 (1988) 674-677. 22 Geddes, J.W, Wong, J., Choi, B.H., Kim, R.C., Cotman, C.W. and Miller, F.D., Increased expression of the embryonic form of a developmentally regulated mRNA in Alzheimer's disease, Neurosci. Lett., 109 (1990) 54-61. 23 Gilman, A.G., G proteins! transducers of receptor-generated signals, Annu. Rev. Biochem., 56 (1987) 615-649. 24 Grundke-Iqbal, I., lqbal, K., Tung, Y.-C., Quinlan, M., Wiln, H.M. and Binder, L.I., Abnormal phosphorylation Of the microtubule-associated protein tau in Alzheimer cytoskeletal pathology, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 4913-4917. 25 Harris, B.A., Complete cDNA sequence of a human stimulatory GTP-binding protein alpha subunit, Nucleic Acid Res.. 16 (1988) 3585. 26 Harrison, P.J. and Pearson, R.C.A., In situ hybridization histochemistry and the study of gene expression in the human brain, Prog. Neurobiol., 34 (1990) 271-312. 27 Harrison, P.J., Barton, A.J.L., Najlerahim, A., McDonald, B. and Pearson, R.C.A., Increased muscarinic receptor messenger RNA in Alzheimer's disease temporal cortex demonstrated by in situ hybridization histochemistry, Mol. Brain Res., 9 (1991) 15-21. 28 Harrison, P.J., Procter, A., Barton, A.J.L., Lowe, S i . , Najlerahim, A., Bertolucci, P.H.E, Bowen, D.M. and Pearson, R.C.A., Terminal coma affects messenger RNA detection in post mortem temporal cortex, Mol. Brain Res., 9 (1991) 161-164. 29 Harrison, P.J., Barton, A.J.L., Najlerahim, A., McDonald, B. and Pearson, R.C.A., Regional and neuronal reductions in polyadenylated messenger RNA in Alzheimer's disease, Psychol. Med., in press. 30 Hurley, J.B., Simon, M.I., Teplow, D.B., Robishaw, J.D. and Gilman, A.G., Homologies between signal transducing G protein and ras oncogene products, Science, 226 (1984) 860-862. 31 Ishikawa, Y., Bianchi, C., Nadal-Ginard, B. and Homey, C.J., Alternative promoter and 5" exon generates a novel Gs~

mRNA, J. Biol. Chem., 265 (1990) 8458-8462. 32 Johnson, G.V. and Hope, R.S., Phosphorylation of rat brain cytoskeletal proteins is increased after orally administered aluminum, Brain Res., 456 (1988) 95-103. 33 Jones, D.T. and Reed, R.R., Molecular cloning of five GTPbinding protein cDNA species from rat olfactory neuroepithelium, J. Biol. Chem., 262 (1987) 14241-14249, 34 Jones, D.T., Masters, S.B., Bourne, H.R. and Reed, R.R., Biochemical characterization of three stimulatory GTP-binding proteins. The large and small forms of G s and the olfactoryspecific G-protein, Go~f, J. Biol. Chem., 265 (1990) 2671-2676. 35 Kalaria, R.N., Andorn, A.C., Tabaton, M., Whitehouse, P.J., Harik, S.L. and Unnerstall, J.R., Adrenergic receptors in aging and Alzheimer's disease: increased /3z-receptors in prefrontal cortex and hippocampus, J. Neurochem., 53 (1989) 1772-1781. 36 Khachaturian, Z.S., Diagnosis of Alzheimer's disease, Arch. Neurol., 42 (1985) 1097-1105. 37 Kosaza, T., Itoh, H., Tsukamoto, T. and Kaziro, Y., Isolation and characterization of the human Gsa gene, Proc. Natl. Acad. Sci. U.S.A., 85 (1988) 2081-2085. 38 Landis, C.A., Masters, S.B., Spada, A., Pace, A.M., Bourne, H.M. and Vallar, L., GTPase inhibiting mutations activate the chain of G s and stimulate adenyl cyclase in human pituitary tumours, Nature, 340 (1989) 692-696. 39 Largent, B.L., Jones, D.T., Reed, R.R., Pearson, R.C.A. and Snyder, S.H., G protein mRNA mapped in rat brain by in situ hybridization, Proc. Natl. Acad. Sci. U.S.A., 85 (1988) 28642868. 40 Masliah, E., Cole, G., Shimohama, S. et al., Differential involvement of protein kinase C isozymes in Alzheimer's disease, J. Neurosci., 10 (1990) 2113-2124. 41 Mengod, G., Palacios, J.M., Probst, A. and Harris, B., Regional distribution of the expression of a human stimulatory GTP-binding protein a-subunit in the human brain studied by in situ hybridization, Mol. Brain Res., 4 (1988) 23-29. 42 Morrison, J.H., Lewis, D.A., Campbell, M.J., Huntley, G.W., Benson, D.L. and Bouras, C., A monoclonal antibody to non-phosphorylated neurofilament protein marks the vulnerable cortical neurons in Alzheimer's disease, Brain Res., 416 (1987) 331-336. 43 Najlerahim, A., Harrison, P.J., Barton, A.J.L., Heffernan, J. and Pearson, R.C.A., Distribution of messenger RNAs encoding the enzymes glutaminase, aspartate aminotransferase and glutamic acid decarboxylase in rat brain, Mol. Brain Res., 7 (1990) 317-333. 44 Neer, E.J. and Clapham, D.E., Roles of G protein subunits in transmembrane signalling, Nature, 333 (1988) 129-134. 45 Ngsee, J.K., Miller, K., Wendland, B. and Scheller, R.H,, Multiple GTP-binding proteins from cholinergic synaptic vesicles, J. Neurosci.. 10 (1990) 317-322. 46 Ohm, rEG., B6hl, J. and Lemmer, B., Reduced cAMP-signal transduction in postmortem hippocampus of demented old people. In K. Iqbal, H.M. Wisniewski and B. Winblad (Eds.), Alzheimer's Disease and Related Disorders, Liss, New York, 1989, pp. 501-509. 47 Otero, A. de S., Transphosphorylation and G protein activation, Biochem. Pharmacol., 39 (1990) 1399-1404. 48 Perl, D.P. and Brody, A.R., Alzheimer's disease: X-ray spectometric evidence of aluminium accumulation in neurofibrillary tangle-bearing neurons, Science, 208 (1980) 297-299. 49 Reynolds, G.P., Arnold, L., Rossor, M.N., Iversen, L.L,, Mountjoy, C.Q. and Roth, M., Reduced binding of [3H]ketanserin to cortical 5-HT 2 receptors in senile dementia of the Alzheimer type, Neurosci. Lett.. 44 (1984) 47-51. 5(1 Roberts, E., Alzheimer's disease may begin in the nose and be caused by aluminosilicates, Neurobiol. Aging, 7 (1986) 561567. 5t Robishaw, J.D., Smigel, M.D. and Gilman, A.G., Molecular basis for two forms of the G protein that stimulates adenylate cyclase, J. Biol. ('hem., 26l (1986) 9587-9590.

81 52 Ross, E.M., Signal sorting and amplification through G proteincoupled receptors, Neuron, 3 (1989) 141-152. 53 Saito, N., Guitart, X., Hayward, M., Tallman, J.F., Duman, R.S. and Nestler, E.J., Corticosterone differentially regulates the expression of Gs~ and Gia messenger RNA and protein in rat cerebral cortex, Proc. Natl. Acad. Sci. U.S.A., 86 (1989) 3906-3910. 54 Severson, J.A., Neurotransmitter receptor interactions with G proteins: a critical link in receptor response mechanisms, Neurobiol. Aging, 9 (1988) 67-68. 55 Shimohama, S., Taniguchi, T., Fujiwara, M. and Kameyama, M., Changes in fl-adrenergic receptor subtypes in Alzheimertype dementia, J. Neurochem., 48 (1987) 1215-1221. 56 Smith, C.J., Perry, E.K., Perry, R.H., Fairbairn, A.E and Birdsall, N.J.M., Guanine nucleotide modulation of muscarinic cholinergic receptor binding in post mortem human brain - - a preliminary study in Alzheimer's disease, Neurosci. Lett., 82 (1987) 227-232. 57 Spiegel, A.M., Albright's hereditary osteodystrophy and defective G proteins, New Engl. J. Med., 322 (1990) 1461-1462. 58 Sternweis, P.C. and Pang, I.-H., The G-protein-channel con-

nection, Trends Neurosci., 13 (1990) 122-126. 59 Strittmatter, S.M., Valenzuela, D., Kennedy, T.E., Neer, E.J. and Fishman, M.C., GOis a major growth cone protein subject to regulation by GAP-43, Nature, 344 (1990) 836-841. 60 Stryer, L. and Bourne, H.R., G proteins: a family of signal transducers, Annu. Rev. Cell Biol., 2 (1986) 391-419. 61 Wolozin, B.L., Pruchnicki, A., Dickson, D.W. and Davies, P., A neuronal antigen in the brains of Alzheimer patients, Science, 232 (1986) 648-650. 62 Worley, P.F., Baraban, J.M. and Snyder, S.H., Beyond receptors: multiple second-messenger systems in brain, Ann. Neurol., 21 (1987) 217-229. 63 Zhang, H., Sternberger, N.H., Rubinstein, L.J., Herman, M.M., Binder, L.I. and Sternberger, L.A., Abnormal processing of multiple proteins in Alzheimer disease, Proc. Natl. Acad. Sci. U.S.A., 86 (1989) 8045-8049. 64 Zubenko, G.S., Cohen, B.M., Reynolds, C.F., Boiler, E, Malinakova, I. and Keefe, N., Platelet membrane fluidity in Alzheimer's disease and major depression, Am. J. Psychiatry, 144 (1987) 860-868.