JOURNAL OF PATHOLOGY, VOL.
168: 67-73 ( I 992)
IN SITU HYBRIDIZATION DEMONSTRATES THE STABILITY OF mRNA IN POST-MORTEM RAT TISSUES E. WALKER AND A. M. McNICOL
University Department of Pathology, Royal Infirmary, Castle Street, GIusgow G4 OSF, U.K. Received 21 January 1992 Accepted 7 April 1992
SUMMARY lnsitu hybridizationwas used to detect messenger RNA (mRNA)in a variety of rat tissues which were fixed in formalin either immediately after death or after a 24 h period of storage at 5°C. A synthetic polydeoxythymidine [poly d(T)] oligonucleotideprobe was used to demonstrate polyadenylated [poly (A)]mRNA in the small intestine, pancreas, liver, cerebellum,and pituitary. Of these tissues, only the liver showed a small reproducible reduction in hybridization signal following delayed fixation. Synthetic oligonucleotide probes complementary to albumin and pro-opiomelanocortin (POMC) mRNAs were hybridized to liver and pituitary, respectively. There was no significant reduction in hybridization signal in post-mortem tissues. The results suggest that some mRNAs may be remarkably stable under certain postmortem conditions and this should encourage the wider application of in situ hybridization techniques to post-mortem material. KEY
WORDS-Rat, post-mortem, in situ hybridization, oligonucleotide, mRNA, poly d(A), POMC, albumin, digoxigenin.
INTRODUCTION The detection of messenger RNA in routinely fixed and processed tissues by in situ hybridization is proving to be an extremely valuable research tool. The recent introduction of reliable and sensitive non-radioactive detection methods will mean more widespread application of the technique to routine histopathological problems.'.* However, as wider research and diagnostic applications of the technique are sought, it is important to identify variables that might influence the detectable levels of mRNA in human biopsy and autopsy specimens. Of particular importance in autopsy material is the effect of the delay between death and fixation of tissue. Some mRNA species have a short half-life3 and thus might logically be expected to decay rapidly after death due to the action of endogenous ribonucleases. The possible effects of post-mortem Addressee for correspondence: Dr E. Walker, University Department of Pathology, Royal Infirmary, Castle Street, Glasgow G4 OSF, U.K.
0022-341 7/92/09006747 $08.50 0 1992 by John Wiley & Sons, Ltd.
changes on mRNAs with a longer physiological half-life are less clear. Taylor et aL4 reported some degradation of RNA isolated from human and mouse brains which was related to both the post-mortem interval and the cause of death. Furthermore, individual mRNA species were reported to be degraded at different rates. On the other hand, several reports have demonstrated remarkable structural and functional stability of mRNA isolated from human liver' or from rodent and human brains at intervals up to 48 h after death."' There have been few formal studies on the effects of post-mortem change on the detection of mRNA by in situ hybridization. Terenghi et aL9 reported successful in situ hybridization for neuropeptide Y mRNA in human brain fixed up to 4 h after death, but Hoefler el a!." reported that a delay of I h between tissue collection and fixation resulted in total loss of hybridization signal for somatostatin mRNA in rat brain. Other studies'' have also suggested poor results in post-mortem tissues. To date, only one experimental studyI2 has specifically
68
E. WALKER AND A. M. McNICOL
set out to assess post-mortem changes in mRNA using in situ hybridization. These workers noted a significant decrease in vasopressin mRNA levels in rat brains fixed 24 h after death. Thus, there is an apparent discrepancy between some isolation studies which appear to show mRNA stability and the few in situ hybridization studies which appear to show relatively rapid postmortem mRNA degradation. To investigate this further, we have made direct comparisons between in situ hybridization signals in a variety of rat tissues fixed either immediately after death or after a postmortem delay of 24 h. This time interval was chosen because the majority of hospital autopsies are carried out within this period. Three synthetic oligonucleotide probes were used-poly d(T) for the assessment of total mRNA in a variety of tissues,” and albumin and pro-opiomelanocortin (POMC) probes for assessing specific mRNAs in the liver and pituitary, respectively. MATERIALS AND METHODS Animals Four male Sprague-Dawley rats (500-600 g) were killed by cervical dislocation and the intact bodies were kept in a refrigerator at 5-10°C for 24 h. After this time, organs were dissected and either fixed whole (pituitary) or in 5-10 mm thick blocks (cerebellum, liver, small intestine, and pancreas). At the same time, four further animals were killed and the tissues dissected and fixed immediately. All tissue samples were fixed for 24 h in 10 per cent neutral buffered formalin followed by standard processing and embedding in paraffin wax. Probes and digoxigenin labelling procedure Four synthetic oligonucleotide probes were used in this study:
(1) 20 mer poly d(T) probe (Pharmacia); (2) 20 mer poly d(A) probe (Pharmacia); (3) 24 mer probe complementary to the region of rat POMC mRNA coding for the ACTH 4 - 1 1 sequence;13and (4) 24mer probe complementary to rat albumin mRNA (kindly donated by D r K. Hillan, Department of Pathology, Western Infirmary, Glasgow). All probes were labelled at their 3’ ends with digoxigenin- 1 1-dUTP by a method previously
described in detail.I3Labelled probe was lyophilized and stored at - 20°C before use. In situ hybridization Tissue sections (4 pm) were cut onto glass slides coated with aminopropyltriethoxysilane (Sigma). The pre-hybridization treatments, in situ hybridization conditions, and post-hybridization washes were identical to those described in detail by Farquharson et al.,I3 the only difference being the omission of the 10 min pre-hybridization treatment in 0.24 per cent acetic anhydride in 0.1 M triethanolamine (pH 7.5) in the present study. Digoxigenin detect ion Following the post-hybridization washes and a 5 min wash in buffer 1 (0.1 M Tris buffer, pH 7.5, with 0.15 M sodium chloride), alkaline phosphatase-conjugated sheep anti-digoxigenin antibody (Boehringer Mannheim) diluted 1500 with 20 per cent normal rat serum in buffer 1 was applied to each section for 3 h. Unbound conjugate was removed by two 20 min washes in buffer 1 followed by one 5 min wash in 0.1 M Tris buffer, pH 9.5, containing 0.1 M sodium chloride and 0.05 M magnesium chloride. Binding was detected by overnight incubation in nitroblue tetrazolium/ bromochloroindolyl-phosphate containing levamisole (80 mg/l) to inhibit endogenous alkaline phosphatase. Slides were counterstained in haematoxylin and mounted using Glycergel (Dako). The following controls were performed:
Sections were processed as described above but in the absence of labelled probe. (2) Sections were incubated with RNAase A (Boehringer Mannheim) at 5 mg/mlin 2 x SSC (1 x SSC = 0.15 M sodium chloride, 0.015 M sodium citrate) for 2 h at 37°C prior to hybridization. (3) Sections were hybridized with a poly d(A) probe to act as a sense control for the poly d(T) probe. (1)
RESULTS On the basis of preliminary studies, tissues were selected which gave a reproducible pattern of staining with the poly d(T) probe. In addition, pituitary gland and liver were included for assessment of
mRNA STABILITY IN POST-MORTEM TISSUES
specific mRNAs. In all tissues and with all probes, pretreatment with RNAase abolished or significantly reduced the positive signal. Hybridization with the poly d(A) probe was negative or produced only background or minor nuclear staining in all the tissues tested. Table I summarizes the results from all in situ hybridization experiments. Table I-Effects of 24 h post-mortem delay in fixation on in situ hybridization signals in rat tissues Result (No. of expts)
Probe
Tissue
Poly d(T) Poly d(T) Poly d(T)
Small intestine Pancreas Brain
Poly d(T)
Liver
Poly d(T)
Pituitary
Albumin
Liver
POMC
Pituitary
NS ( 6 / 6 ) NS (3/3) NS (4/6) M R (2/6) NS ~ 7 ) SR (5/7) NS (6/7) M U (1 /7) NS (3/4) ~ ~ ( 1 / 4 )
NS (6/6)
NS = No significant reduction in hybridization signal in postmortem tissue. M R = Minor reduction in hybridization signal in post-mortem tissue. SR = Significant reduction in hybridization signal in post-mortem tissue. POMC =Pro-opiomelanocortin.
Small intestine-poly
d ( T ) probe
Small intestinal mucosa showed the strongest most reproducible staining of all tissues tested with the poly d(T) probe. In fresh tissue, the strongest cytoplasmic signal was present at the crypt bases and there was an overall reduction in epithelial staining towards the mucosal surface, although occasional single cells still stained strongly (Fig. 1a). Cells within the lamina propria, such as plasma cells, also showed strong cytoplasmic staining. Equivalent tissue fixed 24 h after death showed similar intense staining of crypts and cells within the lamina propria (Fig. 1b). Areas of tissue showing advanced autolytic change, such as the tips of villi, showed weaker hybridization signals. There was no significant difference in the poly d(T) hybridization signal when fresh and post-mortem tissues were compared.
69
Pancreas-poly d ( T ) probe
This tissue also showed equally strong cytoplasmic staining with the poly d(T) probe in both fresh and post-mortem tissue. Brain-poly
d( T ) probe
Preliminary experiments showed that the best hybridization signal was obtained with the large Purkinje cells of the cerebellar cortex and hence all the results discussed here refer to sections of cerebellum. Overall, both fresh and post-mortem tissues showed equally strong cytoplasmic staining of Purkinje cell cytoplasm with some additional nuclear staining. Liver-poly
d ( T ) and albumin probes
Hybridization of liver sections with the poly d(T) probe produced an unusual staining pattern (Fig. 2a). There was intense hepatocyte staining in periportal areas (zone 1) with patchy intense staining of single cells or small groups of cells in zones 2 and 3 . Post-mortem tissue (Fig. 2b) showed a similar pattern of staining but in five out of seven experiments there was a significant reduction in the intensity of staining. Hybridization of liver tissue with the albumin probe produced clear cytoplasmic staining of hepatocytes which was strongest in periportal areas (zone 1 ) with progressive reduction through zones 2 and 3 (Fig. 2c). Post-mortem tissue showed a similar pattern, although there was a tendency towards more widespread staining in zones 2 and 3 (Fig. 2d), but overall there was no reduction in hybridization signal in post-mortem tissue. Pituitary-poly
d( T ) and POMC probes
Pituitary tissue showed variable staining with the poly d(T) probe. The intermediate and posterior lobes showed no significant hybridization signal over a wide range of proteinase K concentrations. The most consistent pattern seen in the anterior lobe was of patchy staining with increased intensity in cells surrounding small blood vessels. Post-mortem tissue showed a similar pattern but no consistent loss of hybridization signal. The POMC probe gave strong cytoplasmic staining in both intermediate lobe cells and anterior lobe corticotrophs, and there were no differences between fresh and post-mortem tissues (Figs 3a and 3b).
70
E. WALKER AND A. M. McNICOL
Fig. I-Sections of rat small intestine hybridized with the poly d(T) probe, Proteinase K (PK) 20pg/ml. Tissues were fixed (a) immediately after death or (b) after a 24 h post-mortem delay. There is no significant difference in staining of post-mortem tissue
DISCUSSION
to those shown in previous studies using human tissue,""' but Pringle et aI." also showed good A large proportion of eukaryotic mRNA is staining in kidney and lymphoid tissue. Our rather synthesized with a 3' tail of 20-250 adenosine unusual but nonetheless reproducible pattern of nucleotides which is added to the mRNA molecule staining produced by hybridization of poly d(T) to in the nu~1eus.l~ This, together with various binding liver has not previously been reported. There are proteins, may play a role in the processing and several possible explanations for the differences in transport of mRNA, the regulation of its trans- signal obtained with poly d(T). They may reflect real lational efficiency, and also its stability or half- differences in the levels of poly (A) mRNA, since life.'5-" Degradation of mRNA molecules is usually some tissues are known to synthesize a significant preceded by removal of the poly (A) tail sequence15 proportion of mRNA without a poly (A) tail." and thus poly (A) sequences should be present in Alternatively, they may be caused by species and association with intact mRNA molecules in all cells. tissue-specific differences in proteins which bind to Since poly d(T) oligonucleotide probes are easy and mRNA, including the poly (A) tail,"-" and thus cheap to synthesize, they have been proposed as a may restrict probe access.2o Our findings warrant caution in the generalized useful means of identifying general levels of mRNA in fixed tissues by in situ hybridization.",'* use of poly d(T) probes to assess whether tissues are In this study, reproducible staining patterns with suitable for in situ hybridization to detect specific the poly d(T) probe were found in the small intes- mRNAs. This is illustrated particularly in the intertine, pancreas, cerebellum, and liver. Anterior mediate lobe of the pituitary gland, where poly pituitary produced a rather variable pattern whilst d(T) hybridization produced no signal, while other tissues considered in preliminary experiments POMC mRNA gives a positive signal even without (kidney, adrenal, spleen) gave no consistent proteinase K t r e a t m ~ n t . ' ~ hybridization signal even in freshly fixed tissues. For our comparative studies, we chose tissues Our results for intestine and cerebellum are similar which produced reproducible staining specific for
mRNA STABILITY IN POST-MORTEM TISSUES
Fig. 2-Sections of rat liver, all PK 60pg/ml. (a) Poly d(T) probe, tissue fixed immediately after death. (b) Poly d(T) probe, tissue fixed after a 24 h post-mortem delay. The post-mortem tissue shows a pattern of staining similar to that of the fresh tissue but the intensity is reduced. (c) Albumin probe, tissue fixed immediately after death. (d) Albumin probe, tissue fixed after a 24 h post-mortem delay. There is no significant reduction in staining of post-mortem tissue
71
72
E. WALKER AND A. M. McNICOL
Fig. 3-Sections of rat pituitary hybridized with the POMC probe, PK lOpg/ ml. Tissues were fixed (a) immediately after death or (b) after a 24 h postmortem delay. There is no significant reduction in hybridization signal in intermediate lobe cells or anterior lobe corticotrophs in post-mortem tissue
poly (A) mRNA as shown by poly d(A) sense probe and RNAase studies. To rule out age and sexrelated differences, or circadian variations, tissues were taken from animals of identical age and sex and which were also killed at the same time of day. Consequently, we showed that there was no significant loss of poly d(T) hybridization signal when freshly fixed samples of small intestine, pancreas, cerebellum, and anterior pituitary were compared directly with similar tissues fixed 24 h after death. Some preliminary experiments (unpublished observations) were also performed on tissue fixed 48 h after death and these still showed significant staining with poly d(T) which was either similar to fresh tissue (small intestine, pituitary) or was similar to staining at 24 h post-mortem (cerebellum). Liver was the only tissue to show a consistent reduction in poly d(T) hybridization signal in tissue fixed 24 h after death, suggesting degradation of mRNA. Stable eukaryotic mRNAs are considered typically to have a half-life of around 24h,” and that of POMC mRNA, for example, has been calculated at between 14 and 24h.” Our studies consistently demonstrated no loss of in situ hybridization signal
for this mRNA over a 24 h post-mortem period. Assessment of another stable mRNA, for albumin, in liver yielded similar results. This is in contrast to the changes seen in poly d(T) labelling of liver sections. If post-mortem degradation of mRNA is taking place, then it is likely that the 3’ poly (A) sequence will be first to d i ~ a p p e a r ,thus ’ ~ accounting for the reduced poly d(T) hybridization signal, whilst the remaining coding sequence is still detectable with the albumin-specific probe. Our experimental animals were refrigerated soon after death and being of small size, all tissues would have cooled rapidly to 5°C. This situation may not be entirely comparable to the human situation, where, firstly, delays inevitably occur prior to refrigeration of bodies and, secondly, the greater size will mean slower cooling of internal organs. Finger et a1.’ and Johnson et al.’ tried to simulate human post-mortem conditions in liver and brain tissue, respectively. Both groups found no significant effect of the post-mortem interval or storage temperatures on the structure and function of purified mRNA. In an experimental study of vasopressin mRNA in post-mortem rat brains, Arai et
m R N A STABILITY IN P(XT-MORTEM TISSUES
73
6 . Morrison M R , Giffin WST. The isolation and in vitro translation of al.” showed that the in situ hybridization signal was undegraded messenger RNAs from human postmortem brain. Anal still strong even after 8 h storage at room temperaBiochem 1981; 113: 318-324. ture. Some ~ ~ r k e rconsider ~ ~ ~that ~ other ~ - ~ 7.~ Johnson , ~ SA,~ Morgan DG. Finch CE. Extensive postmortem stability of RNA from rat and human brain. J Neurosci Res 1986; 1 6 267-280. factors such as cause of death, agonal state, and 8. Kobayashi H, Sakimura K, Kuwano R, er a/. Stability of messenger dissection or handling of tissues prior to fixation RNA in postmortem human brains and construction of human brain cDNA libraries. J Mol Nrurosci 1990; 2: 29-34. are more important factors than the post-mortem G , Polak JM. Hamid Q, ~t a/. Localization of ueuropeptide storage interval. These were not tested in our study. 9. Terenghi Y mRNA in neurons of human cerebral cortex by means of ifi siru We would propose, therefore, that in situ hybridization with a complementary RNA probe. Proc Nail Acad Sci USA 1987;84:7315-7318. hybridization may usefully be applied to some 10. Hoefler H. Childers H, Montminy M R , et a/. I n siiu hybridization tissues fixed up to at least 24 h after death for the methods for the detection of somatostatin mRNA in tissue seclions using antisense RNA probes. Histochrm J 1986; 1 8 597-604. qualitative detection of stable mRNAs of moderate Pringle JH, Primrose L, Kind CN, Talbot IC. Lauder I. fn situ to high abundance. However, it is worth emphasiz- I I . hybridization demonstration of poly-adenylated RNA sequences in ing that the results should be validated for each formalin-fixed paraffin sections using a biotinylated oligonucleotide poly d(T) probe. J Parhol1989; I58 279-286. specific mRNA, tissue type, and set of conditions 12. Arai H , Noguchi I , Sagi N, Moroji T, lizuka R. A study of nontested . isotopic in silu hybridization histochemistry on postmortem changes
ACKNOWLEDGEMENTS
We thank Kenneth Hillan for the albumin probe, Rona Harvie and Maura Farquharson for helpful advice, and Joan Cramb for typing the manuscript.
REFERENCES I. Akhtar N. Ruprai A, Pringle JH, el ul. in situ hybridization detection
2.
3. 4.
5.
of‘light chain mRNA in routine bone marrow trephines from palients with suspected myeloma. Br J Harmatol1989; 7 3 296301. Walker E, Robertson AG, Boorman JG, McNicol AM. Primary cutaneous plasmacytoma. The use of in situ hybridization t o detect monoclonal immunoglobulin light chain mRNA. Hi,s/upa/holugy 1992; 20: 135-138. Brawerman G. Determinants of messenger RNA stability. Cell 1987; 48:5 6 . Taylor GR. Carter GI, Crow TJ, ct al. Recovery and measurement of specific R N A species from postmortem brain tissue: a general reduction in Alzheimer’s disease detected by molecular hybridization. E r p M o l Pathol 1986;44: I 1 1-1 I6. Finger JM, Mercer JFB, Cotton RGH, Danks DM. Stability of protein and mRNA in human postmortem liver-analysis by twodimensional gel electrophoresis. Clin Chim Acta 1987; 170: 209 -21 8.
in vasopressin mRNA in rat brain. Neurosci Lrtt 1989; 103: 127-132. 13. Farquharson M, Harvie R, McNicol AM. Detection of messenger RNA using a digoxigenin end labellcd oligodeoxynucleotide probe. J Clin Parhol 1990; 4 3 424428. 14. Wickens M. How the messenger got its tail: addition ofpoly (A) in the nucleus. Trends Bioehem Sci 1990; I 5 277--28 I . 15. Bernstein P, Ross J. Poly (A), poly (A) binding protein and the regulation of mRNA stability. Trends Biochem Sci 1989; 1 4 375-377. 16. Munroe D, Jacobson A. Tales of poly (A): a review. Gene 1990; 91: 151-1 58. 17. Saini KS, Summerhayes IC, Thomas P. Molecular events regulating messenger R N A stability in eukaryotea. Mol Cell Biochem 1990; 96: 15-23. 18. Talhot IC, Primrose L, Pringle JH. Functional activity of intestinal epithelium demonstrated by mRNA in .si/u hybridization. J Pirrhol 1989: 158: 287-291. 19. ChdUdhdri N, Hahn WE. Genetic expresaion in the developing brain. Science 1983: 220 924-928. 20. Harrison PJ, Pearson RCA. I n situ hybridization histochemistry and the study ofgene expression in the human brain. Prug. Neuruhiol1990; 34271-312. 21. Puckett L, Chambers S, Darnell JE. Short-lived messenger RNA in HeLa cells and its impact o n the kinetics of accumulation of cytoplasmic polyadenylate. Pruc No// Acud Sci USA 1975; 72: 389-393. 22. Birnberg NC, Lissitzky J-C, Hinman M, Herbert E. Glucocorticoids regulate proopiomelanocortin gene expression in vivo at the levels of transcription and secretion. Proc. Nut/ Acad Sci USA 1983: 80: 6982-6986. 23. Terenghi G, Fallon RA. Techniques and applicationa of m sirir hybridization. Curr ‘Fop Pufhol 1990; 82: 289-337. 24. Harrison PJ. Procter AW, Barton AJL, rt a/. Terminal coma affects messenger R N A detection in postmortem human temporal cortex. MolBrain Rrs 1991;9 161-164.
168: 67-73 ( I 992)
IN SITU HYBRIDIZATION DEMONSTRATES THE STABILITY OF mRNA IN POST-MORTEM RAT TISSUES E. WALKER AND A. M. McNICOL
University Department of Pathology, Royal Infirmary, Castle Street, GIusgow G4 OSF, U.K. Received 21 January 1992 Accepted 7 April 1992
SUMMARY lnsitu hybridizationwas used to detect messenger RNA (mRNA)in a variety of rat tissues which were fixed in formalin either immediately after death or after a 24 h period of storage at 5°C. A synthetic polydeoxythymidine [poly d(T)] oligonucleotideprobe was used to demonstrate polyadenylated [poly (A)]mRNA in the small intestine, pancreas, liver, cerebellum,and pituitary. Of these tissues, only the liver showed a small reproducible reduction in hybridization signal following delayed fixation. Synthetic oligonucleotide probes complementary to albumin and pro-opiomelanocortin (POMC) mRNAs were hybridized to liver and pituitary, respectively. There was no significant reduction in hybridization signal in post-mortem tissues. The results suggest that some mRNAs may be remarkably stable under certain postmortem conditions and this should encourage the wider application of in situ hybridization techniques to post-mortem material. KEY
WORDS-Rat, post-mortem, in situ hybridization, oligonucleotide, mRNA, poly d(A), POMC, albumin, digoxigenin.
INTRODUCTION The detection of messenger RNA in routinely fixed and processed tissues by in situ hybridization is proving to be an extremely valuable research tool. The recent introduction of reliable and sensitive non-radioactive detection methods will mean more widespread application of the technique to routine histopathological problems.'.* However, as wider research and diagnostic applications of the technique are sought, it is important to identify variables that might influence the detectable levels of mRNA in human biopsy and autopsy specimens. Of particular importance in autopsy material is the effect of the delay between death and fixation of tissue. Some mRNA species have a short half-life3 and thus might logically be expected to decay rapidly after death due to the action of endogenous ribonucleases. The possible effects of post-mortem Addressee for correspondence: Dr E. Walker, University Department of Pathology, Royal Infirmary, Castle Street, Glasgow G4 OSF, U.K.
0022-341 7/92/09006747 $08.50 0 1992 by John Wiley & Sons, Ltd.
changes on mRNAs with a longer physiological half-life are less clear. Taylor et aL4 reported some degradation of RNA isolated from human and mouse brains which was related to both the post-mortem interval and the cause of death. Furthermore, individual mRNA species were reported to be degraded at different rates. On the other hand, several reports have demonstrated remarkable structural and functional stability of mRNA isolated from human liver' or from rodent and human brains at intervals up to 48 h after death."' There have been few formal studies on the effects of post-mortem change on the detection of mRNA by in situ hybridization. Terenghi et aL9 reported successful in situ hybridization for neuropeptide Y mRNA in human brain fixed up to 4 h after death, but Hoefler el a!." reported that a delay of I h between tissue collection and fixation resulted in total loss of hybridization signal for somatostatin mRNA in rat brain. Other studies'' have also suggested poor results in post-mortem tissues. To date, only one experimental studyI2 has specifically
68
E. WALKER AND A. M. McNICOL
set out to assess post-mortem changes in mRNA using in situ hybridization. These workers noted a significant decrease in vasopressin mRNA levels in rat brains fixed 24 h after death. Thus, there is an apparent discrepancy between some isolation studies which appear to show mRNA stability and the few in situ hybridization studies which appear to show relatively rapid postmortem mRNA degradation. To investigate this further, we have made direct comparisons between in situ hybridization signals in a variety of rat tissues fixed either immediately after death or after a postmortem delay of 24 h. This time interval was chosen because the majority of hospital autopsies are carried out within this period. Three synthetic oligonucleotide probes were used-poly d(T) for the assessment of total mRNA in a variety of tissues,” and albumin and pro-opiomelanocortin (POMC) probes for assessing specific mRNAs in the liver and pituitary, respectively. MATERIALS AND METHODS Animals Four male Sprague-Dawley rats (500-600 g) were killed by cervical dislocation and the intact bodies were kept in a refrigerator at 5-10°C for 24 h. After this time, organs were dissected and either fixed whole (pituitary) or in 5-10 mm thick blocks (cerebellum, liver, small intestine, and pancreas). At the same time, four further animals were killed and the tissues dissected and fixed immediately. All tissue samples were fixed for 24 h in 10 per cent neutral buffered formalin followed by standard processing and embedding in paraffin wax. Probes and digoxigenin labelling procedure Four synthetic oligonucleotide probes were used in this study:
(1) 20 mer poly d(T) probe (Pharmacia); (2) 20 mer poly d(A) probe (Pharmacia); (3) 24 mer probe complementary to the region of rat POMC mRNA coding for the ACTH 4 - 1 1 sequence;13and (4) 24mer probe complementary to rat albumin mRNA (kindly donated by D r K. Hillan, Department of Pathology, Western Infirmary, Glasgow). All probes were labelled at their 3’ ends with digoxigenin- 1 1-dUTP by a method previously
described in detail.I3Labelled probe was lyophilized and stored at - 20°C before use. In situ hybridization Tissue sections (4 pm) were cut onto glass slides coated with aminopropyltriethoxysilane (Sigma). The pre-hybridization treatments, in situ hybridization conditions, and post-hybridization washes were identical to those described in detail by Farquharson et al.,I3 the only difference being the omission of the 10 min pre-hybridization treatment in 0.24 per cent acetic anhydride in 0.1 M triethanolamine (pH 7.5) in the present study. Digoxigenin detect ion Following the post-hybridization washes and a 5 min wash in buffer 1 (0.1 M Tris buffer, pH 7.5, with 0.15 M sodium chloride), alkaline phosphatase-conjugated sheep anti-digoxigenin antibody (Boehringer Mannheim) diluted 1500 with 20 per cent normal rat serum in buffer 1 was applied to each section for 3 h. Unbound conjugate was removed by two 20 min washes in buffer 1 followed by one 5 min wash in 0.1 M Tris buffer, pH 9.5, containing 0.1 M sodium chloride and 0.05 M magnesium chloride. Binding was detected by overnight incubation in nitroblue tetrazolium/ bromochloroindolyl-phosphate containing levamisole (80 mg/l) to inhibit endogenous alkaline phosphatase. Slides were counterstained in haematoxylin and mounted using Glycergel (Dako). The following controls were performed:
Sections were processed as described above but in the absence of labelled probe. (2) Sections were incubated with RNAase A (Boehringer Mannheim) at 5 mg/mlin 2 x SSC (1 x SSC = 0.15 M sodium chloride, 0.015 M sodium citrate) for 2 h at 37°C prior to hybridization. (3) Sections were hybridized with a poly d(A) probe to act as a sense control for the poly d(T) probe. (1)
RESULTS On the basis of preliminary studies, tissues were selected which gave a reproducible pattern of staining with the poly d(T) probe. In addition, pituitary gland and liver were included for assessment of
mRNA STABILITY IN POST-MORTEM TISSUES
specific mRNAs. In all tissues and with all probes, pretreatment with RNAase abolished or significantly reduced the positive signal. Hybridization with the poly d(A) probe was negative or produced only background or minor nuclear staining in all the tissues tested. Table I summarizes the results from all in situ hybridization experiments. Table I-Effects of 24 h post-mortem delay in fixation on in situ hybridization signals in rat tissues Result (No. of expts)
Probe
Tissue
Poly d(T) Poly d(T) Poly d(T)
Small intestine Pancreas Brain
Poly d(T)
Liver
Poly d(T)
Pituitary
Albumin
Liver
POMC
Pituitary
NS ( 6 / 6 ) NS (3/3) NS (4/6) M R (2/6) NS ~ 7 ) SR (5/7) NS (6/7) M U (1 /7) NS (3/4) ~ ~ ( 1 / 4 )
NS (6/6)
NS = No significant reduction in hybridization signal in postmortem tissue. M R = Minor reduction in hybridization signal in post-mortem tissue. SR = Significant reduction in hybridization signal in post-mortem tissue. POMC =Pro-opiomelanocortin.
Small intestine-poly
d ( T ) probe
Small intestinal mucosa showed the strongest most reproducible staining of all tissues tested with the poly d(T) probe. In fresh tissue, the strongest cytoplasmic signal was present at the crypt bases and there was an overall reduction in epithelial staining towards the mucosal surface, although occasional single cells still stained strongly (Fig. 1a). Cells within the lamina propria, such as plasma cells, also showed strong cytoplasmic staining. Equivalent tissue fixed 24 h after death showed similar intense staining of crypts and cells within the lamina propria (Fig. 1b). Areas of tissue showing advanced autolytic change, such as the tips of villi, showed weaker hybridization signals. There was no significant difference in the poly d(T) hybridization signal when fresh and post-mortem tissues were compared.
69
Pancreas-poly d ( T ) probe
This tissue also showed equally strong cytoplasmic staining with the poly d(T) probe in both fresh and post-mortem tissue. Brain-poly
d( T ) probe
Preliminary experiments showed that the best hybridization signal was obtained with the large Purkinje cells of the cerebellar cortex and hence all the results discussed here refer to sections of cerebellum. Overall, both fresh and post-mortem tissues showed equally strong cytoplasmic staining of Purkinje cell cytoplasm with some additional nuclear staining. Liver-poly
d ( T ) and albumin probes
Hybridization of liver sections with the poly d(T) probe produced an unusual staining pattern (Fig. 2a). There was intense hepatocyte staining in periportal areas (zone 1) with patchy intense staining of single cells or small groups of cells in zones 2 and 3 . Post-mortem tissue (Fig. 2b) showed a similar pattern of staining but in five out of seven experiments there was a significant reduction in the intensity of staining. Hybridization of liver tissue with the albumin probe produced clear cytoplasmic staining of hepatocytes which was strongest in periportal areas (zone 1 ) with progressive reduction through zones 2 and 3 (Fig. 2c). Post-mortem tissue showed a similar pattern, although there was a tendency towards more widespread staining in zones 2 and 3 (Fig. 2d), but overall there was no reduction in hybridization signal in post-mortem tissue. Pituitary-poly
d( T ) and POMC probes
Pituitary tissue showed variable staining with the poly d(T) probe. The intermediate and posterior lobes showed no significant hybridization signal over a wide range of proteinase K concentrations. The most consistent pattern seen in the anterior lobe was of patchy staining with increased intensity in cells surrounding small blood vessels. Post-mortem tissue showed a similar pattern but no consistent loss of hybridization signal. The POMC probe gave strong cytoplasmic staining in both intermediate lobe cells and anterior lobe corticotrophs, and there were no differences between fresh and post-mortem tissues (Figs 3a and 3b).
70
E. WALKER AND A. M. McNICOL
Fig. I-Sections of rat small intestine hybridized with the poly d(T) probe, Proteinase K (PK) 20pg/ml. Tissues were fixed (a) immediately after death or (b) after a 24 h post-mortem delay. There is no significant difference in staining of post-mortem tissue
DISCUSSION
to those shown in previous studies using human tissue,""' but Pringle et aI." also showed good A large proportion of eukaryotic mRNA is staining in kidney and lymphoid tissue. Our rather synthesized with a 3' tail of 20-250 adenosine unusual but nonetheless reproducible pattern of nucleotides which is added to the mRNA molecule staining produced by hybridization of poly d(T) to in the nu~1eus.l~ This, together with various binding liver has not previously been reported. There are proteins, may play a role in the processing and several possible explanations for the differences in transport of mRNA, the regulation of its trans- signal obtained with poly d(T). They may reflect real lational efficiency, and also its stability or half- differences in the levels of poly (A) mRNA, since life.'5-" Degradation of mRNA molecules is usually some tissues are known to synthesize a significant preceded by removal of the poly (A) tail sequence15 proportion of mRNA without a poly (A) tail." and thus poly (A) sequences should be present in Alternatively, they may be caused by species and association with intact mRNA molecules in all cells. tissue-specific differences in proteins which bind to Since poly d(T) oligonucleotide probes are easy and mRNA, including the poly (A) tail,"-" and thus cheap to synthesize, they have been proposed as a may restrict probe access.2o Our findings warrant caution in the generalized useful means of identifying general levels of mRNA in fixed tissues by in situ hybridization.",'* use of poly d(T) probes to assess whether tissues are In this study, reproducible staining patterns with suitable for in situ hybridization to detect specific the poly d(T) probe were found in the small intes- mRNAs. This is illustrated particularly in the intertine, pancreas, cerebellum, and liver. Anterior mediate lobe of the pituitary gland, where poly pituitary produced a rather variable pattern whilst d(T) hybridization produced no signal, while other tissues considered in preliminary experiments POMC mRNA gives a positive signal even without (kidney, adrenal, spleen) gave no consistent proteinase K t r e a t m ~ n t . ' ~ hybridization signal even in freshly fixed tissues. For our comparative studies, we chose tissues Our results for intestine and cerebellum are similar which produced reproducible staining specific for
mRNA STABILITY IN POST-MORTEM TISSUES
Fig. 2-Sections of rat liver, all PK 60pg/ml. (a) Poly d(T) probe, tissue fixed immediately after death. (b) Poly d(T) probe, tissue fixed after a 24 h post-mortem delay. The post-mortem tissue shows a pattern of staining similar to that of the fresh tissue but the intensity is reduced. (c) Albumin probe, tissue fixed immediately after death. (d) Albumin probe, tissue fixed after a 24 h post-mortem delay. There is no significant reduction in staining of post-mortem tissue
71
72
E. WALKER AND A. M. McNICOL
Fig. 3-Sections of rat pituitary hybridized with the POMC probe, PK lOpg/ ml. Tissues were fixed (a) immediately after death or (b) after a 24 h postmortem delay. There is no significant reduction in hybridization signal in intermediate lobe cells or anterior lobe corticotrophs in post-mortem tissue
poly (A) mRNA as shown by poly d(A) sense probe and RNAase studies. To rule out age and sexrelated differences, or circadian variations, tissues were taken from animals of identical age and sex and which were also killed at the same time of day. Consequently, we showed that there was no significant loss of poly d(T) hybridization signal when freshly fixed samples of small intestine, pancreas, cerebellum, and anterior pituitary were compared directly with similar tissues fixed 24 h after death. Some preliminary experiments (unpublished observations) were also performed on tissue fixed 48 h after death and these still showed significant staining with poly d(T) which was either similar to fresh tissue (small intestine, pituitary) or was similar to staining at 24 h post-mortem (cerebellum). Liver was the only tissue to show a consistent reduction in poly d(T) hybridization signal in tissue fixed 24 h after death, suggesting degradation of mRNA. Stable eukaryotic mRNAs are considered typically to have a half-life of around 24h,” and that of POMC mRNA, for example, has been calculated at between 14 and 24h.” Our studies consistently demonstrated no loss of in situ hybridization signal
for this mRNA over a 24 h post-mortem period. Assessment of another stable mRNA, for albumin, in liver yielded similar results. This is in contrast to the changes seen in poly d(T) labelling of liver sections. If post-mortem degradation of mRNA is taking place, then it is likely that the 3’ poly (A) sequence will be first to d i ~ a p p e a r ,thus ’ ~ accounting for the reduced poly d(T) hybridization signal, whilst the remaining coding sequence is still detectable with the albumin-specific probe. Our experimental animals were refrigerated soon after death and being of small size, all tissues would have cooled rapidly to 5°C. This situation may not be entirely comparable to the human situation, where, firstly, delays inevitably occur prior to refrigeration of bodies and, secondly, the greater size will mean slower cooling of internal organs. Finger et a1.’ and Johnson et al.’ tried to simulate human post-mortem conditions in liver and brain tissue, respectively. Both groups found no significant effect of the post-mortem interval or storage temperatures on the structure and function of purified mRNA. In an experimental study of vasopressin mRNA in post-mortem rat brains, Arai et
m R N A STABILITY IN P(XT-MORTEM TISSUES
73
6 . Morrison M R , Giffin WST. The isolation and in vitro translation of al.” showed that the in situ hybridization signal was undegraded messenger RNAs from human postmortem brain. Anal still strong even after 8 h storage at room temperaBiochem 1981; 113: 318-324. ture. Some ~ ~ r k e rconsider ~ ~ ~that ~ other ~ - ~ 7.~ Johnson , ~ SA,~ Morgan DG. Finch CE. Extensive postmortem stability of RNA from rat and human brain. J Neurosci Res 1986; 1 6 267-280. factors such as cause of death, agonal state, and 8. Kobayashi H, Sakimura K, Kuwano R, er a/. Stability of messenger dissection or handling of tissues prior to fixation RNA in postmortem human brains and construction of human brain cDNA libraries. J Mol Nrurosci 1990; 2: 29-34. are more important factors than the post-mortem G , Polak JM. Hamid Q, ~t a/. Localization of ueuropeptide storage interval. These were not tested in our study. 9. Terenghi Y mRNA in neurons of human cerebral cortex by means of ifi siru We would propose, therefore, that in situ hybridization with a complementary RNA probe. Proc Nail Acad Sci USA 1987;84:7315-7318. hybridization may usefully be applied to some 10. Hoefler H. Childers H, Montminy M R , et a/. I n siiu hybridization tissues fixed up to at least 24 h after death for the methods for the detection of somatostatin mRNA in tissue seclions using antisense RNA probes. Histochrm J 1986; 1 8 597-604. qualitative detection of stable mRNAs of moderate Pringle JH, Primrose L, Kind CN, Talbot IC. Lauder I. fn situ to high abundance. However, it is worth emphasiz- I I . hybridization demonstration of poly-adenylated RNA sequences in ing that the results should be validated for each formalin-fixed paraffin sections using a biotinylated oligonucleotide poly d(T) probe. J Parhol1989; I58 279-286. specific mRNA, tissue type, and set of conditions 12. Arai H , Noguchi I , Sagi N, Moroji T, lizuka R. A study of nontested . isotopic in silu hybridization histochemistry on postmortem changes
ACKNOWLEDGEMENTS
We thank Kenneth Hillan for the albumin probe, Rona Harvie and Maura Farquharson for helpful advice, and Joan Cramb for typing the manuscript.
REFERENCES I. Akhtar N. Ruprai A, Pringle JH, el ul. in situ hybridization detection
2.
3. 4.
5.
of‘light chain mRNA in routine bone marrow trephines from palients with suspected myeloma. Br J Harmatol1989; 7 3 296301. Walker E, Robertson AG, Boorman JG, McNicol AM. Primary cutaneous plasmacytoma. The use of in situ hybridization t o detect monoclonal immunoglobulin light chain mRNA. Hi,s/upa/holugy 1992; 20: 135-138. Brawerman G. Determinants of messenger RNA stability. Cell 1987; 48:5 6 . Taylor GR. Carter GI, Crow TJ, ct al. Recovery and measurement of specific R N A species from postmortem brain tissue: a general reduction in Alzheimer’s disease detected by molecular hybridization. E r p M o l Pathol 1986;44: I 1 1-1 I6. Finger JM, Mercer JFB, Cotton RGH, Danks DM. Stability of protein and mRNA in human postmortem liver-analysis by twodimensional gel electrophoresis. Clin Chim Acta 1987; 170: 209 -21 8.
in vasopressin mRNA in rat brain. Neurosci Lrtt 1989; 103: 127-132. 13. Farquharson M, Harvie R, McNicol AM. Detection of messenger RNA using a digoxigenin end labellcd oligodeoxynucleotide probe. J Clin Parhol 1990; 4 3 424428. 14. Wickens M. How the messenger got its tail: addition ofpoly (A) in the nucleus. Trends Bioehem Sci 1990; I 5 277--28 I . 15. Bernstein P, Ross J. Poly (A), poly (A) binding protein and the regulation of mRNA stability. Trends Biochem Sci 1989; 1 4 375-377. 16. Munroe D, Jacobson A. Tales of poly (A): a review. Gene 1990; 91: 151-1 58. 17. Saini KS, Summerhayes IC, Thomas P. Molecular events regulating messenger R N A stability in eukaryotea. Mol Cell Biochem 1990; 96: 15-23. 18. Talhot IC, Primrose L, Pringle JH. Functional activity of intestinal epithelium demonstrated by mRNA in .si/u hybridization. J Pirrhol 1989: 158: 287-291. 19. ChdUdhdri N, Hahn WE. Genetic expresaion in the developing brain. Science 1983: 220 924-928. 20. Harrison PJ, Pearson RCA. I n situ hybridization histochemistry and the study ofgene expression in the human brain. Prug. Neuruhiol1990; 34271-312. 21. Puckett L, Chambers S, Darnell JE. Short-lived messenger RNA in HeLa cells and its impact o n the kinetics of accumulation of cytoplasmic polyadenylate. Pruc No// Acud Sci USA 1975; 72: 389-393. 22. Birnberg NC, Lissitzky J-C, Hinman M, Herbert E. Glucocorticoids regulate proopiomelanocortin gene expression in vivo at the levels of transcription and secretion. Proc. Nut/ Acad Sci USA 1983: 80: 6982-6986. 23. Terenghi G, Fallon RA. Techniques and applicationa of m sirir hybridization. Curr ‘Fop Pufhol 1990; 82: 289-337. 24. Harrison PJ. Procter AW, Barton AJL, rt a/. Terminal coma affects messenger R N A detection in postmortem human temporal cortex. MolBrain Rrs 1991;9 161-164.
