Terminal coma affects messenger RNA detection in post mortem human temporal cortex

Molecular Brain Research, 9 (1991) 161-164 Elsevier

161

BRESM 80070

Terminal coma affects messenger RNA detection in post mortem human temporal cortex P.J. Harrison 1'2, A.W. Procter 3'4, A.J.L. Barton 1, S.L. Lowe 3, A. Najlerahim 5, P.H.F. Bertolucci 3, D.M. Bowen 3 and R.C.A. Pearson 5 Departments of 1Anatomy and Cell Biology, and ZPsychiatry, St. Mary's Hospital Medical School, London (U.K.), SMiriam Marks Department of Neurochemistry, Institute of Neurology, London (U.K.), 4Department of Psychiatry, UMDS-Guy's Hospital Campus, London (U.K.) and SDepartment of Biomedical Science, The University, Sheffield ( U. K. ) (Accepted 24 July 1990) Key words: In situ hybridization histochemistry; Alzheimer's disease; Muscarinic receptor; Glutamate decarboxylase; Poly(A); Agonal state

In situ hybridization histochemistry has been used to study the amount of M1 muscarinic receptor mRNA in temporal cortex from subjects with Alzheimer's disease and other neurodegenerative disorders, where the duration of terminal coma was known. Total polyadenylated mRNA and glutamate deearboxylase activity were also measured. Both muscarinic receptor mRNA and enzyme activity showed a significant decline with increasing duration of terminal coma, but were not related to diagnosis. Polyadenylated mRNA signal did not show an association with coma. These data indicate the need to consider the nature of the terminal illness in post mortem studies of mRNA as well as for neurochemical research.

Post mortem human brain is often used for neurochemical research, and increasingly so for in situ hybridization histochemistry (ISHH). An important consideration in both instances is to separate disease-specific changes from the non-specific effects of factors such as post mortem delay (PMD) and the nature of the terminal illness (called agonal state). The latter is known markedly to affect some enzyme activities measured post mortem 7' 16 as well as other neuroehemieal parameters 5. Terminal coma, hypoxia and pyrexia are all potential confounding variables. Recently it has been suggested that such considerations may also apply to detection of m R N A and contribute significantly to the variable yields of RNA obtained from otherwise similar brains s. With the extension of I S H H to include post mortem material, and especially given the attempts to make quantitative comparisons of m R N A s between Alzheimer's disease (AD) and control brains, we have investigated the effect of agonal state on detection of a specific m R N A , that coding for the M 1 subtype muscarinic receptor. Polyadenylate m R N A (poly(A) + m R N A ) and glutamate decarboxylase ( G A D ) activity were also estimated in the same tissue samples. For this study, 14 pathologically confirmed A D brains were used together with 6 other neuropsychiatric cases and one normal control. Details of most of this series have been reportedt3; Table I summarises the clinicopa-

thological data and describes the additional cases used in the present study. Brains were removed shortly after death; one hemisphere was placed in formalin for diagnosis, the other cut coronaUy into 1 cm slices. These were placed in cold modified Krebs-Ringer phosphate buffer, transferred to the laboratory in an insulated box, and frozen at -70 °C19, From these slices, blocks of middle temporal gyrus were taken. Ten-/~m-thick cryostat sections were cut for ISHH; for assay of G A D activity, 200 mg of tissue, containing all cortical layers and free of meninges and white matter, was dissected from the blocks. For ISHH, sections were pretreated and hybridized as described previously 8'9. Briefly, for M 1 receptor ISHH, an oligonucleotide complementary to bases 1690-1719 of the published sequence 1 was 3" tail labelled using a kit (NEN DuPont) with [35S]dATE 1 x 10 6 cpm of labelled probe was added to each section in 100 gl of buffer, comprising 4 x SSC, 1 x Denhardt's solution, 5 mM NaEPO4, 10% dextran sulphate, 50% formamide, 1 mM E D T A and 10 mM dithiothreitol, plus 100 /~g/ml of denatured salmon sperm D N A , poly(A) and yeast tRNA. Incubation was at 30 °C for 18 h followed by washing in 1 x SSC at 53 °C (3 x 20 min) and room temperature (2 x 60 min). Sections were apposed to film (Hyperfilm betamax, Amersham) for 21 days. Controls for I S H H consisted of: (1) concurrent and identical

Correspondence: R.C.A. Pearson, Department of Biomedical Science, The University, Sheffield S10 2TN, U.K. 0169-328X/91/$03.50 © 1991 Elsevier Science Publishers B.V. (Biomedical Division)

162 hybridization to adjacent sections with the sense strand probe, (2) ribonuclease pretreatment of sections, and (3) a Northern blot of 30 pg total R N A from frontal cortex hybridized with [32p]dATP-labelled probe 15. For detection of poly(A) tails of m R N A 1°, a 30 base deoxythymidine probe ('oligo(dT)') was labelled with [35S]dTTP, together with the sense strand (30 bases of adenosine, labelled with [35S]dATP). 1.75 x 106 cpm of labelled probe were added to each section in 100 pl of 2 × SSC, 20% formamide and 10 mM dithiothreitol. Incubation and washing were at 25 °C and 50 °C respectively, for the same times as for M1 receptor ISHH. Exposure to film was for 7.5 min. I S H H was quantified densitometrically on autoradiograms blind to clinical details using an image analysis apparatus as described previously 9 (Image Manager; Sight Systems). For each section, the mean grey density value through the depth of the cortical grey matter was determined. Since the instrument assigns numbers from 0 (black; strongest signal) to 255 (white; no signal), this density value was divided into 1000, so that higher numerical values indicate increasing hybridization. The reading arising from sense strand hybridization was deducted from that of the antisense probe for each case to produce a final value representing specific hybridization. All values fell within the linear part of the apparatus' response curve relating radioactivity exposure to observed grey density value. For determination of G A D activity, tissue was homogenized in 5 volumes of I0 mM potassium phosphate buffer (pH 6.4) using a glass-teflon motorized homogenizer assembly. G A D activity was assayed using a radioactive CO2 trapping method with slight modification: 50 mM L-[U-laC]glutamic acid was used as substrate

and incubations were carried out for 1 h at 37 °C; total protein was measured colourimetrically 4. The muscarinic receptor I S H H controls confirmed hybridization specificity; the Northern b l o t revealed a single band of 2.7 kb 8'9, sense strand hybridization produced a low, uniform signal (Fig. 1), and ribonuclease pretreatment abolished significant hybridization (not shown). For oligo(dT) I S H H , the oligo(dA) sense strand produced no detectable signal (not shown). A significant inverse correlation was found between the densitometric amount of M1 receptor m R N A and duration of terminal coma (r = -0.79, P < 0.001; Table II); a similar correlation existed for the A D cases taken separately (r = -0.70, P < 0.01). Subdividing the length of coma suggested the major decline occurred with a duration greater than 24 h (Table II). There was no significant difference in amount of M~ receptor m R N A

TABLE ! Clinicopathological details o f cases ~

Age (years) Sex Coma (h) PMD (h)

Alzheimer' s disease (n=14)

Non-Alzheimer's disease b (n=7)

73 ± 10 2M, 12F 65 ± 93 2.4 ± 0.7

74 + 8 4M, 3F 68 + 90 1.8 + 0.6

Values are mean + S.D. a Additional AD cases from those of ref. 12: AP 13 (78 years), AP16 (86 years), AP21 (89 yeats) and AP22 (80 years), with duration (h) of terminal coma (PMD, h, in parentheses) of 12 (2), 20 (1.5), 24 (1.25) and 20 (2.5), respectively. Coma length was not determined for a subject with Pick's disease (AP19, 71 years, 2 h PMD) or for 3 AD cases, AP12, AP15 and AP18, with ages (PMD, h, in parentheses) of 67 (3.5), 65 (3) and 62 (3.5), respectively, b Diagnoses: Pick's disease (2), multi-infarct dementia, depression, dementia with cortical plaques but no tangles, multisystem atrophy, normal control.

Fig. 1. Representative autoradiogram of MI receptor mRNA in temporal cortex hybridized with (A) antisense, and (B) sense strand. Pictures are printed directly from film under identical conditions; increasing whiteness indicates increasing signal.

163 TABLE II Effect o f coma on mascarinic receptor m R N A , poly(A) + m R N A and G A D activity

Values are mean + S.D. a Spearman rank correlation coefficient. b Arbitrary grey density units, c nmol/mg protein/h. Correlation Mean signalb ( m R N A ) or activity c with coma GAD length a 24 h coma coma

M~ receptor mRNA Poly(A) ÷ mRNA GAD activity

r = -0.79 P < 0.001 r =-0.11 N.S. r = -0.48 P < 0.05

3.9 _+ 1.9 (n = 4)

3.4 +_1.5 (n = 5)

1.7 + 0.6* (n = 7)

5.1+0.3 5.2+0.6 5.2+0.5 (n = 4) (n = 5) (n = 7) 36.5 + 17.2 13.2 + 6.4* 13.2 + 12.2 (n = 5) (n = 4) (n = 8)

*P < 0.05, compared to less than 1 h coma (Mann-Whitney).

b e t w e e n A D cases and o t h e r neurological diseases (Table III); neither did the A D cases have a different mean c o m a duration from the other cases (Table I). I S H H signal did not relate to age or sex of the patient, nor to P M D (not shown). A similar association was found b e t w e e n coma duration and G A D activity (r = - 0 . 4 8 , P < 0.05). Unlike the case with muscarinic receptor m R N A , this decline occurred within 24 h of coma and then p l a t e a u e d (Table III). G A D activity also correlated with M 1 r e c e p t o r I S H H signal (r = 0.52, P < 0.02). T h e r e was no correlation between p o l y ( A ) ÷ m R N A signal and c o m a o r diagnosis (Tables II and III). A g o n a l state has often been implicated as a cause of variability in neurochemical p a r a m e t e r s between otherwise similar cases 3-6"14'18'2°. H o w e v e r , there has been little firm d a t a to support this contention in the case of R N A studies. This r e p o r t provides direct evidence that agonal state, at least in terms of terminal coma, is indeed a d e t e r m i n a n t of m R N A preservation as well as of enzyme activity. The need accurately to document, and to a t t e m p t to control for, agonal state, is a p p a r e n t from

TABLE III Muscarinic receptor m R N A , p o l y ( A ) + m R N A and G A D activity in A D and n o n - A D cases A D cases (n = 14)

Muscarinic receptor mRNA b Poly(A)÷ mRNA b GAD activityc

2.4 + 1.3 5.1 + 0.6 12.8 + 8.2

Non-A D cases a (n=6)

N.S. N.S. N.S.

3.2 + 1.8 5.1 + 0.3 17.4 + 13.7

Values are mean + S.D. ~ Excluding normal control, b Arbitrary grey density units, c nmoi/mg proteirdh.

these data. This is especially true for quantitative I S H H studies with brains affected by disorders in which patients frequently die after terminal c o m a , pyrexia and hypoxia, and which are c o m p a r e d to n o r m a l controls whose agonal state on average may be better. In contrast, P M D appears to be less significant, since delays of 72 h or more are still satisfactory for I S H H 8. I n d e e d , we have found no decline in muscarinic r e c e p t o r m R N A between the present brains, with a m e a n P M D of 2.4 h, and a n o t h e r series processed after 30 h 9. These findings also reaffirm the value of a positive control group with which to c o m p a r e A D cases. A significant gene-specific increase in muscarinic r e c e p t o r m R N A has been r e p o r t e d in A D t e m p o r a l cortex c o m p a r e d to n o r m a l controls 9. H o w e v e r , in the present study, the amount of this m R N A was at least as high in non-AD neurological cases as in A D (Table III). This suggests that the amounts found here in both disease groups are elevated c o m p a r e d to n o r m a l controls, and m a y be the result of chronic brain disease p e r se, or the involvement of the basal cholinergic forebrain in n o n - A D pathologies 17. In either event, the rise in muscarinic r e c e p t o r m R N A in A D is not a disease-specific change. Conversely, the fact that this m R N A is increased in neurological disorders c o m p a r e d to normal controls, despite its decline with coma, m a k e s the increase a more robust finding, given that c o m a is likely to be more m a r k e d amongst neurological cases. T h e absence of a normal control group in the p r e s e n t study, however, leaves open the possibility that coma differentially affects m R N A in diseased brains in comparison with neurologically intact ones. It would be necessary, although difficult, to collect prospectively assessed, rapid autopsy brains from n o r m a l controls dying after varying coma times in o r d e r to answer this question directly. P o l y ( A ) ÷ m R N A did not decline with coma in parallel with that of M 1 r e c e p t o r m R N A . This is not due to an inability of oligo(dT) I S H H to detect changes in p o l y ( A ) ÷ m R N A , since the p r e s e n t m e t h o d o l o g y has identified reductions b e t w e e n A D brains and controls elsewhere 2'1°. It is possible that c o m a is associated with increased or de novo transcription of genes which c o m p e n s a t e for the decline in o t h e r m R N A s ; candidates for such an increase include heat shock genes 12 and proto-oncogenes 11, which are both elevated by other factors associated with a p o o r agonal state. Alternatively, given the unchanged a m o u n t of p o l y ( A ) ÷ m R N A found here despite p r o l o n g e d coma, it may prove that the production or stability of m R N A as a whole is n o t m a r k e d l y influenced by coma. If this is the case, the present reduction in M 1 r e c e p t o r m R N A (which will have an undetectable effect on a m o u n t of p o l y ( A ) ÷ m R N A ) is likely to be an example of a m o r e specific effect of coma

164 selectively decreasing certain low abundance m R N A s . These opposing interpretations emphasize the need to investigate p r e m o r t e m effects on any individual m R N A to be studied in post m o r t e m brain; it may well prove that, like enzymes, different m R N A s vary in their responses to terminal coma and sensitivities to o t h e r agonal state factors. I n d e e d , recent data (unpublished observations) suggest this to be so. The present data indicate that I S H H studies on post m o r t e m brain should include documentation and att e m p t e d matching of agonal state; in this way, effects of the underlying disease can be distinguished from nonspecific events. A p a r t from coma, other agonal state factors such as fever, acidosis and medication should be noted, since all m a y affect gene expression and protein synthesis s. The i m p o r t a n c e of clinicians' cooperation in this evaluation becomes a p p a r e n t 19. A s with other tech-

We thank a number of colleagues, in particular Dr. R. Doshi and Prof. E. Murphy, for help in collection and classification of cases. The research was supported by the Wellcome Trust, Halley-Stewart Trust and the Brain Research Trust ('Miriam Marks Department'). P.J.H. is an M.R.C. Training Fellow. A.J.L.B. is Cottrell Fellow of Research into Ageing. P.H.EB. is supported by Consello Nacional Para O Desenvolvimento Cientifico e Technologica (Brazil).

1 Allard, W.J., Sigal, I.S. and Dixon, R.A.E, Sequence of the gene encoding the huron M1 muscarinic acetylcholine receptor, Nucl. Acids Res., 15 (1987) 10604. 2 Barton, A.J.L., Harrison, P.J., Najlerahim, A., Hefferman, J., McDonald, B., Robinson, J.R., Davies, D.C., Harrison, W.J., Mitra, P., Hardy, J.A. and Pearson, R.C.A., Increased tau messenger RNA in Alzheimer's disease hippocampus, Am. J. PathoL, in press. 3 Bird, E.D., Gale, J.S. and Spokes, E.G., Huntington's chorea: postmortem activity of enzymes involved in cerebral glucose metabolism, J. Neurochem., 29 (1977) 539-545. 4 Bowen, D.M., Smith, C.B., White, P. and Davison, A.N., Neurotransmitter-related enzymes and indices of hypoxia in senile dementia and other abiotrophies, Brain, 99 (1976) 459-496. 5 Bowen, D.M., Smith, C.B., White, P., Goodhardt, M.J., Spillane, J.A., Flack, R.H.A. and Davison, A.N., Chemical pathology of the organic dementias. I. Validity of biochemical measurements on human postmortem brain specimens, Brain, 100 (1977) 397-406. 6 Butterwortb, J., Yates, C.M. and Simpson, J., Phosphateactivated glutaminase in relation to Huntington's disease and agonal state, J. Neurochem., 41 (1983) 440-447. 7 Dodd, P.R., Hambley, J.W., Cowburn, R.E and Hardy, J.A., A comparison of methodologies for the study of functional transmitter neurochemistry in human brain, J. Neurochem., 50 (1988) 1333-1345. 8 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. 9 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., in press. 10 Harrison, P.J., Barton, A.J.L., Najlerahim, A., McDonald, B.

and Pearson, R.C.A., Regional and neuronal changes in polyadenylated messenger RNA levels in Alzheimer's disease brain, submitted. 11 Jorgensen, M.B., Deckert, J., Wright, D.C. and Gehlert, D.R., Delayed c-los proto-oncogene expression in the rat hippocampus induced by transient global cerebral ischaemia: an in situ hybridization study, Brain Res., 484 (1989) 393-398. 12 Lindquist, S., The heat-shock response, Annu. Rev. Biochem., 55 (1986) 1151-1191. 13 Lowe, S.L., Francis, P.T., Procter, A.W., Palmer, A.M., Davison, A.N. and Bowen, D.M., Gamma-aminobutyric acid concentration in brain tissue at two stages of Alzheimer's disease, Brain, 111 (1988) 785-789. 14 McGeer, P.L., McGeer, E.G. and Wada, J.A., Glutamic acid decarboxylase in Parkinson's disease and epilepsy, Neurology, 21 (1971) 1000-1007. 15 Najlerahim, A., Barton, A.J.L., Harrison, P.J., Heffernan, J. and Pearson, R.C.A., Messenger RNA encoding the D2 dopaminergic receptor detected in rat brain by in situ hybridization histochemistry, FEBS Lea., 255 (1989) 335-339. 16 Palmer, A.M., Lowe, S.L., Francis, P.T. and Bowen, D.M., Are postmorten biochemical studies of human brain worthwhile?, Biochem. Soc. Trans., 16 (1987) 472-475. 17 Pearson, R.C.A. and Powell, T.P.S., The neuroanatomy of Alzheimer's disease, Rev. Neurosci., 2 (1989) 101-123. 18 Perry, E.K., Perry, R.H. and Tomlinson, B.E., The influence of agonal status on some neurochemieal activities of post mortem human brain tissue, Neurosci. Lett., 29 (1982) 303-307. 19 Procter, A., Doshi, B., Bowen, D. and Murphy, E., Rapid autopsy brains for biochemical research: experiences in establishing a programme, Int. J. Geriatr. Psychiat., in press. 20 Spokes, E., An analysis of factors influencing measurements of dopamine, noradrenaline, glutamate decarboxylase and choline acetyltransferase in human post mortem brain tissue, Brain, 102 (1979) 333-346.

niques, I S H H studies should also include a p p r o p r i a t e positive controls w h e n e v e r possible to separate changes specific to a particular disorder from those reflecting chronic brain disease. A d d i t i o n a l l y , the concomitant use of a measure of overall gene expression, such as oligo(dT) I S H H , helps identify m R N A s which are differentially altered in a disease, as o p p o s e d to those which merely change in t a n d e m with total gene expression. Only by attending to all these variables will I S H H studies contribute maximally to the understanding of human brain disorders.