- Email: [email protected]
Retrieval of mRNA from paraffin-embedded human infant brain tissue for non-radioactive in situ hybridization using oligonucleotides
Journal of Neuroscience Methods 115 (2002) 129 /136 www.elsevier.com/locate/jneumeth
Retrieval of mRNA from paraffin-embedded human infant brain tissue for non-radioactive in situ hybridization using oligonucleotides Bronwyn L. Relf a, Rita Machaalani a, Karen A. Waters a,b,c,* a
Department of Medicine, Room 206, Blackburn Building, DO6, The University of Sydney, Sydney, NSW 2006, Australia b Paediatrics and Child Health, The University of Sydney, Sydney, NSW 2006, Australia c The Childrens Hospital, Westmead, Sydney, Australia Received 9 July 2001; received in revised form 20 December 2001; accepted 20 December 2001
Abstract In situ hybridization (ISH) is used to examine the spatiotemporal distribution of gene expression in a range of tissues. Neuroscience research in human brain tissue requires techniques that can be used in formalin fixed and paraffin-embedded tissue rather than frozen tissue which is recommended, but difficult to obtain. This study presents a method for non-radioactive (DIG) ISH for detecting NR1 gene expression, in human infant brain tissue. We compared three pre-treatment effects, protease digestion, autoclaving (in citrate and Tris/EDTA buffer) and microwaving (in citrate and Tris/EDTA buffer). Tissue had been fixed in formalin for 2 /12 weeks. Results were compared for the hybridization and background signal intensities, and tissue morphology. We found that optimum results were obtained using 12-min microwave pre-treatment in Tris/EDTA buffer. This method produced optimum signal to background ratio in infant and adult tissue, preserved tissue morphology, and was suitable for use across a broad range of fixation times. # 2002 Elsevier Science B.V. All rights reserved. Keywords: In situ hybridization; Oligonucleotide; Non-radioactive; mRNA; Infant; Brainstem; Microwave
1. Introduction In situ hybridization (ISH) is used widely to examine the spatiotemporal distribution of gene expression in tissue. Recent studies have used ISH to identify gene expression differences in disease states of the brain, when compared to normal. Frozen tissue preparations are preferred over paraffin embedded tissue, because of the superior hybridization signal obtained (Oliver et al., 1997b). However, post-mortem brain tissue from humans is difficult to obtain, and is most often fixed so that frozen tissue is unavailable despite its perceived benefits. Research applications would be greatly extended if the methods of analysis were adapted for use in fixed tissue. Formalin, the most common fixative used, helps retain nucleic acids but forms extensive cross-linkages * Corresponding author. Tel.: 61-2-9351-5165; fax: 61-2-95503851. E-mail addresses: [email protected], karenw2@chw. edu.au (K.A. Waters).
in the process, hindering probe penetration (Eastwood and Harrison, 1999). The longer the period of fixation the greater the amount of cross-linkages formed and the harsher the pre-treatments required to expose the retained RNA to the probe. Pre-treatments that have been used to remove cross-linkages include proteinase K digestion (Bayer et al., 1995; Oliver et al., 1997a,b), hydrated autoclaving (Oliver et al., 1997a,b; Eastwood and Harrison, 1999) and microwave treatment in buffer solution (Lan et al., 1996; McMahon and McQuaid, 1996; Oliver et al., 1997a,b). Enzyme digestion with proteases, such as proteinase K, have been popular, but to achieve optimum results the conditions for digestion must be altered to suit the type of tissue and the duration of prior fixation. In addition, with enzymatic digestion the unmasking of RNA often occurs at the expense of tissue and cellular morphology (Lan et al., 1996). This has led to the use of hydrated autoclaving or microwaving as alternative pretreatments (Lan et al., 1996; McMahon and McQuaid, 1996; Oliver et al., 1997a,b; Eastwood and Harrison, 1999). Finally, radioactive labels have traditionally been
0165-0270/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 0 2 7 0 ( 0 2 ) 0 0 0 0 3 - 1
130
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
used for ISH. However, non-radioactive methods are increasingly popular because they produce faster results, the labelled probes are more stable, results are more consistent, and there is less biohazard to laboratory staff (McQuaid et al., 1995). Our laboratory has a specific interest in studying the neuropathology of sudden, unexplained death in infancy. The aim of this study was to determine the optimal pre-treatment for retrieving cellular mRNA from paraffin embedded human infant brainstem and hippocampal tissue. We compared several pre-treatment regimes on the hybridization signal intensity, background staining and tissue morphology in non-radioactive (digoxigenin) ISH to identify the NR1 subunit of the N -methyl-D-aspartate (NMDA) receptor in infant brain tissue.
2. Materials and methods 2.1. Tissue acquisition Archival, paraffin-embedded human infant (age 1 month to 1 year) brainstem tissue was obtained with the permission of the NSW Coroner. At the time of post-mortem, none of the cases had any neuropathology identified. The tissue was routinely fixed in 10% formalin, prior to paraffin embedding for routine diagnostic studies at the Institute of Forensic Medicine. Adult hippocampal tissue was obtained from the NSW Tissue Resource Centre and was used to confirm the method of Bayer et al. (1995) and subsequently, as positive controls. The study was approved by the Human Ethics Committee of the University of Sydney. All tissue sections were cut at 7 mm thickness and mounted on slides coated with 3-aminopropyltriethoxysilane, then dried at 37 8C for 24 h prior to in situ hybridization. 2.2. Oligonucleotide probes Synthetic oligonucleotide probes (Gibco BRL, Melbourne, Australia), complementary to the sub-unit specific region between transmembrane region I and transmembrane region II, were designed from the human NR1 gene (Karp et al., 1993; Planells-Cases et al., 1993). The probes had the following sequences: 5? CTC CTC CTC CTC GCT GTT CAC CTT GAA CCG GCC GAA GGG GCT GAA 3? (antisense probe, complementary to sequences encoding amino acid residues 565/579 of the mature NR1 protein) and 5? TTC AGC CCC TTC GGC CGG TTC AAG GTG AAC AGC GAG GAG GAG GAG 3? (NR-1 sense, reverse complement of NR-1 antisense oligonucleotide). The probes were designed to correspond to the most abundant isoform NR1A. This probe length was chosen
to optimise penetration in formalin-fixed tissue, whilst maintaining specificity. Anti-sense and sense probes were labelled by 3?-tailing with digoxigenin-11-dUTP according to the DIG Oligonucleotide tailing kit from Boehringer Mannheim (Mannheim, Germany). The reaction was incubated at 37 8C for 1 h. Labelled oligonucleotide was purified by ethanol precipitation. Probe concentration was determined by quantification against known standards on Hybond N filters, as described in the DIG System User’s Guide for Filter Hybridization (Boehringer, Mannheim, Germany). 2.3. Pre-treatment of sections prior to in situ hybridization All reactions were performed at room temperature unless noted otherwise. Solutions were made up in diethyl pyrocarbonate (DEPC) treated water. Sections were deparaffinised in xylene (2 /10 min), taken through a graded series of ethanols (1 /5 min in 100, 95, 75, 50 and 30%) and washed in DEPC H2O (2 /5 min). From this point sections were subjected to one of three different pre-treatments: 1.
2.
3.
Enzymatic digestion: sections were digested in 50, 100 or 200 mg/ml proteinase K (Boehringer Mannheim, Germany) made up in phosphate buffered saline (PBS) for 15 min at 37 8C. Digestion was halted by washing in PBS containing 2 mg/ml glycine (2 /5 min). Autoclaving: sections were placed in either 10 mM citrate buffer, pH 6.0 or in Tris/EDTA buffer (1 mM EDTA, 1 mM sodium citrate, 2 mM Tris; pH 9.0), autoclaved at 120 8C/2 bar for 40 min, cooled to 70 8C and washed in DEPC H2O (1 /15 min). Microwaving: sections were either placed in 10 mM citrate buffer or in Tris/EDTA buffer and microwaved at full power on a rotating turntable, uninterrupted for 12 min (Samsung 850 W, Korea). Sections were allowed to cool in the buffer for 15 min and DEPC H2O was added gradually over 5 min to reduce the buffer to room temperature. Sections were then washed in DEPC H2O (2 /5 min).
2.4. In situ hybridization All sections were subsequently washed in PBS (1 /5 min) and acetylated in 0.25% acetic anhydride in 0.1 M triethanolamine, pH 8.0 for 10 min. After washing in PBS (5 min), PAP pen (Zymed laboratories, San Francisco) was applied around the sections and the sections rinsed in PBS (5 min). Sections were prehybridized in 4/standard saline citrate (SSC)/50% formamide for 2 h at 37 8C, and rinsed in 2/SSC (5 min)
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
prior to overnight hybridization at 37 8C in a humidified chamber. Specifically, 50 ng of antisense probe in 100 ml of hybridization buffer was added just prior to hybridization, during which the sections were coverslipped with Parafilm. The composition of the hybridization solution was 18% formamide, 2 /SSC, 1 / Denhardt’s solution, 10% dextran sulfate, 50 mM dithiothreitol (DTT), 250 mg/ml yeast t-RNA, 100 mg/ ml polyadenylic acid and 500 mg/ml denatured and sheared salmon sperm DNA. Nonspecific hybridization was examined including slides that had been incubated with either the sense probe or hybridization solution only.
2.5. Post-hybridization Following hybridization, sections were washed in 2 / SSC at room temperature (15 min) until the parafilm washed off the section. Washes were all 15 min duration, and included 2 /SSC at 37 8C, 1/SSC at 37 8C, then at room temperature, 1/SSC, 0.5 /SSC and 0.25 / SSC.
2.6. Immunological detection Sections were then treated with either concentrated (100%) or diluted (10%) immunological detection buffers. After three washes with either 100% digoxigenin buffer (DB; 100 mM maleic acid, 150 mM NaCl, pH 7.5), or 10% DB, for 5 min each, blocking solution was added for 1 h at 37 8C. The composition of the blocking solution was 10% w/v blocking reagent [BR] (Roche Diagnostics), and 2.5% normal sheep serum (NSS) in the relevant concentration of DB. Anti-DIG alkaline phosphatase antibody (1:250 dilution in DB, containing 10% BR and 1% NSS) was added after blocking and the sections were incubated for 2 h on a rocking platform at room temperature. Non-specific staining was examined by the inclusion of sections without antibody treatment. All sections were washed in DB (2 /10 min) and then in either 100% digoxigenin detection buffer (DDB; 100 mM Tris, 100 mM NaCl, 50 mM MgCl2, pH 9.5) or 10% DDB for 10 min. The alkaline phosphatase color reaction buffer mixture was added according to the specifications of the manufacturer (Boehringer Mannheim). This buffer contained 35 ml nitroblue tetrazolium [NBT], and 45 ml 5-bromo-4-chloro-3-indolylphosphate [BCIP] in 10 ml DDB. Sections were incubated overnight (16 /18 h) in a humidified dark box. The color reaction was terminated by washing in Tris buffer (10 mM Tris, 0.1 mM EDTA, pH 8.0) (5 min) and distilled H2O (5 min). Sections were then mounted in Aquamount solution (BDH Laboratory Supplies Poole, UK).
131
2.7. Analysis of in situ hybridization results Pre-treatment effects on the signal intensity, background levels and tissue morphology were evaluated visually (under a light microscope) and compared using a visual scaling system. The following semiquantitative score was used for signal intensity: 0, negative; /, weak; //, moderate; ///, strong; and ////, very strong. Background staining was scored: /, low; //, medium; and ///, high. For tissue morphology the scores were: 0, no change; /, some tissue disruption; //, marked tissue disruption. Quantification by optical density measurement was only made for the sections included in Fig. 1, to provide representative measurements from each pre-treatment, and study group. Briefly, microscopy images were captured by a Sensicam (PCO Computer optics, Kelheim, Germany) attached to a microscope (Nikon Eclipse E800, Tokyo, Japan). The images were digitised, and optical density (OD) values were calculated using commercially available software (Zeiss KS 400, Munich, Germany). Optical density was measured for all neurones and five randomly selected background areas, for each section shown in Fig. 1. The study groups comprised tissue from the CA3 region of the adult hippocampus, and a hypoglossal nucleus at the level of the closed medulla from infant brainstem. The signal to background ratio was calculated by dividing the mean cellular optical density value by the mean background optical density value for each section. Statistical comparisons were performed using SPSS (V10 for Windows, Chicago). Optical density values were compared amongst groups (infant brainstem or adult hippocampus), signal and background intensity, and pre-treatment (proteinase K, autoclaving, or microwaving), as factors in a general linear model. Results are presented as mean9/S.E., and a P -value of B/0.05 was considered significant.
3. Results A total of 29 cases (25 infants and 4 adults) were studied, using one to two sections per case. The mean age was 3.79/2.6 months (range 1 /10) for infants, and 65.89/14.5 years (range 46 /78) for adults. Fixation time was 6.69/3.5 weeks (range 2/12) for infants, and 5.39/ 2.2 weeks (range 3 /8) for adults. To assess each pretreatment regime, qualitative analyses was performed for hybridization signal intensity, background staining and tissue morphology. See Table 1. The optical density measurements confirmed our visual scaling results and permitted quantification of the results. Note that within the range of fixation times we tested, no change was required in the protocol.
132
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
Fig. 1. Effect of pre-treatment on signal strength and background hybridization in adult hippocampal tissue (CA3 region, left column, panels A /C) and infant brainstem (hypoglossal nucleus at closed medulla, right column, panels D /F). Sections digested with 200 mg/ml proteinase K (A) and (D); sections autoclaved in Tris/EDTA buffer (B) and (E); Sections microwaved in Tris/EDTA buffer (C) and (F).
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
133
Table 1 Visual scale comparison of pre-treatment effects on signal intensity, background level and tissue morphology in adult and infant brain tissue Pre-treatment
Adult hippocampus
Infant brainstem
Signal
BG
Morphology
Signal
BG
Morphology
Proteinase K 50 mg/ml 100 mg/ml 200 mg/ml
0 0
0 0
0 0
0 0
Autoclave Tris/EDTA Citrate
Microwave Tris/EDTA Citrate
0
0 0
First column: 0, no signal; , weak; , moderate; , strong signal intensity; , very strong signal intensity and non specific staining. Second column: , low; , medium; , high background. Third column: 0, no effect on tissue morphology; , some tissue disruption; , marked tissue disruption. BG, background.
3.1. Effect of pre-treatment on signal intensity No positive hybridization signal was observed in sections hybridized with the sense probe, where probe was omitted or in any sections digested with 50 or 100 mg/ml proteinase K. Positive hybridization signals were observed in all other sections. Signal intensity varied amongst the pre-treatment groups and between adult and infant tissue. Generally, staining was localised in the nucleolus and cytoplasm but some nuclear and dendritic process staining was also observed. Similar patterns of hybridization have been reported by others using DIGlabelled oligonucleotide probes (Dirks et al., 1993; Bayer et al., 1995). In adult hippocampal tissue, the lowest signal intensity was observed after digestion with 200 mg/ml proteinase K (Fig. 1A and Table 1). Autoclaving with Tris/EDTA buffer gave a high, but non-specific, signal intensity, staining all cells including neurones and glia (Fig. 1B). Autoclaving with citrate buffer gave a medium signal intensity which was also non-specific (Table 1). Microwaving with either Tris/EDTA (Fig. 1C) or citrate buffer produced high signal intensities, which was localised to neuronal cells. In infant tissue, the highest signal intensities were observed when the sections were digested with 200 mg/ml proteinase K (Fig. 1D) or autoclaved in Tris/EDTA (Fig. 1E) or citrate buffer but in each case the staining was also non-specific. Microwaving produced a high signal intensity in Tris/EDTA buffer (Fig. 1F) and medium staining in citrate buffer, with the staining remaining specific to neurones in each case. Quantification of the pre-treatment effects on signal intensity, using optical density measurements, is shown in Fig. 2A.
Fig. 2. I, Adult hippocampal tissue; j, infant brainstem tissue. OD, optical density. Pre-treatments include protK 200 mg/ml proteinase K, AC tris autoclaved in Tris/EDTA buffer, and micro tris microwaved in Tris/EDTA buffer. (A) Optical density (OD) values of the neuronal hybridization signal. (B) Optical density (OD) values of the background tissue. (C) Signal to background ratio amongst the three pre-treatments.
134
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
Statistical analysis of the optical density measures showed significant differences amongst each of the pretreatments. Overall mean value for proteinase K/ 0.579/0.01, autoclaved in Tris/EDTA /1.289/0.05, and microwaved in Tris/EDTA /0.939/0.02 (P B/ 0.001 amongst all pre-treatments). These values reflect optical density in 52, 54 and 47 cells, respectively. 3.2. Effect of pre-treatment on background levels For both the adult and infant tissue, low background levels were observed after digestion in proteinase K at either 50 or 100 mg/ml. However, the absence of a hybridization signal in these sections indicated that the only color substrate binding was to antibody attached non-specifically to the tissue. The level of background staining seen in these sections was used as the baseline for comparing background levels after the other pretreatment regimes. In the adult hippocampus, medium background levels were observed in sections digested with 200 mg/ml proteinase K (Fig. 1A), and those autoclaved or microwaved in citrate buffer. High background levels were observed in sections autoclaved with Tris/EDTA buffer (Fig. 1B). A low background level, equivalent to baseline, was observed in the sections when microwaved in Tris/EDTA buffer (Fig. 1C). In the infant tissue, high background was seen after autoclaving in both citrate and Tris/EDTA (Fig. 1E, Table 1), whereas microwaving (Fig. 1F) and digestion with 200 mg/ml proteinase K (Fig. 1D) gave background levels equivalent to baseline (Table 1). Quantification of the pre-treatment effects on background levels, using optical density measurements, is shown in Fig. 2B. 3.3. Signal to background ratio In both the adult and infant brain tissue, the best (highest) signal to background ratio was obtained by microwaving in Tris/EDTA buffer. The worst (lowest) signal to background ratio for both adult and infant tissue was obtained in the Tris/EDTA autoclaved sections. Quantification of the pre-treatment effects on signal to background ratio, using optical density measurements, is shown in Fig. 2C. Statistical analyses were not performed for the signal to background ratio. Differences between signal and background were significant for all pre-treatment groups. Overall signal intensity (for all pre-treatments) was 0.869/0.04 and 1.029/0.05 for adult hippocampus and infant brainstem, respectively (P B/0.001). Overall background intensity was 0.379/0.08 and 0.319/0.06 for adult hippocampus and infant brainstem, respectively. Differences were significant for signal versus background P B/0.001, and interaction between signal and group (adult hippocampus vs. infant brainstem, P /0.005).
3.4. Effect of pre-treatment on tissue morphology Tissue morphology was assessed qualitatively. The parameters that were evaluated were the presence/ absence of identifiable cellular structures such as the nucleus, nucleolus, and indications that the cellular and nuclear membranes were intact. Tissue loss from the slide was also evaluated. Pre-treatment effects on tissue morphology were similar between adult and infant sections. Amongst the different pre-treatments autoclaving caused the greatest tissue disruption, with significant loss of tissue from slides. Digestion with 200 mg/ml proteinase K caused some tissue degradation, although less than that seen in autoclaved sections. In general, microwaving had no effect on tissue morphology, although some loss of tissue was observed from slides of adult brain tissue that were microwaved in citrate buffer. 3.5. Effect of dilution of the immunological detection buffers Dilution of the digoxygenin buffer (DB) and digoxygenin detection buffer (DDB) to 10%, reduced background staining in the tissue sections whilst maintaining signal strength and specificity. In contrast, washing with 100% DB and DDB caused higher levels of background staining and more non-specific staining of glial and neuronal cells. Tissue morphology also appeared to be better preserved in sections washed in the diluted buffers (Fig. 3).
4. Discussion In this study, three pre-treatment regimes were compared in formalin fixed and paraffin embedded human brain tissue, from adults and infants. The aim was to optimise the methods for detecting hybridization signal in fixed tissue, compared to previously published methods recommending the use of hydrated autoclaving (Oliver et al., 1997a,b; Eastwood and Harrison, 1999). Hydrated autoclaving was not suitable for use in infant brain tissue, because of significant degradation of tissue morphology, high background staining and high levels of non-specific staining. The new methods presented in this study resolved those problems in infant brain tissue, but also reduced the signal to noise ratio, and improved tissue morphology in the adult tissue sections, compared to the autoclaving method. The changes to pre-treatment that were required to examine infant brain tissue may relate to maturational differences between adult and infant tissues, particularly the level of myelination. The results of our study suggest that pre-treatments suitable for adult tissue were too harsh for infant brain tissue. The optimal pre-treatment gave the highest signal
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
135
Fig. 3. Effect of dilution of immunological buffers on hybridization signal strength, and background hybridization in infant brainstem tissue (hypoglossal nucleus). (A) Sections washed with concentrated (100%) DB and DDB. (B) Sections washed with diluted (10%) DB and DDB. (C) A section treated with sense probe, using diluted (10%) DB and DDB. All sections illustrated here were microwaved in Tris/EDTA buffer.
to background ratio whilst still maintaining tissue morphology in both adult and infant brain tissue. The advantages of using digoxigenin labelled probes permitted a 3-day development time compared to several weeks for radioactive probes, and the signal intensity could be quantified using optical density measurements. Microwaving is a suitable pre-treatment for in situ hybridization (McMahon and McQuaid, 1996), to prepare sections where enzyme digestion or hydrated autoclaving would otherwise be used. The heat produced by microwaving, or autoclaving, disrupts crosslinkages formed during formalin fixation (McMahon and McQuaid, 1996; Oliver et al., 1997a,b; Eastwood and Harrison, 1999). During microwaving, heat is produced by the oscillation of dipolar molecules (such as water and the side chains of protein molecules), that is continuous and proportional to the energy applied (Leong et al., 1988). The argument that uneven heating, and evaporation of the liquid during microwaving, produces uneven disruption of the cross-linkages was not confirmed in our results, where morphology and signal distribution were consistent throughout the sections.
The three major advantages with microwaving were lower signal to background ratio, preservation of tissue morphology, and shorter treatment time. To compensate for the same factors using enzyme digestion would require adjustments of time, temperature and of enzyme concentrations. Note also, that within the range of fixation times we tested, no change was required in the protocol, a finding that we attribute to the fact that we used microwaving in the pre-treatment regime. Thus microwaving enabled probe access to the target mRNA by breaking down the extensive cross-linkages formed during formalin fixation whilst preserving tissue morphology. Effects of changes in the buffer solutions may have several origins. The Tris/EDTA buffer, with a pH of 9.0, proved optimal during microwaving pre-treatment. This buffer gave higher signal intensity in all tissue types with less background compared to the citrate buffer. Studies have shown that the pH of the retrieval solution is an important factor where a high pH (pH 8 /10) largely improved staining results in immunohistochemistry (Shi et al., 1995). Citrate buffer, pH 6.0, not only has a low pH but utilises NaOH which has been noted to cause
136
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
tissue loss from slides during microwaving (Evers et al., 1998). The different salt concentration of the buffers may also have contributed to their effects on signal strength. Tris/EDTA has a total salt concentration of 4 mM compared to citrate that has a salt concentration of 10 mM. The effects of diluting the immunological detection buffers, DB and DDB, are not well understood. However, research suggests that the reduction in salt concentrations may modify the hydrophobicity of the polypeptide chains which acts by reducing probe binding to free proteins (McMahon and McQuaid, 1996). Some non-specific staining of pre-mRNA was found in the nucleus as a consequence of our use of an oligonucleotide probe. Dirks et al. (1993), suggest that nuclear and nucleolar localization is almost certainly the result of using an oligonucleotide probe, since this also detects DNA and pre-mRNA sequences located in the cell nucleus. In contrast, RNA probes detect only mature mRNA sequences. The oligonucleotide probe we used did not span the intron, although this method may be an alternative means of detecting only mature RNA with cytoplasmic localisation. The oligonucleotide sequence we used was taken from Bayer et al. (1995) and we found and reported the same localisation of the probe as they did. To our knowledge, this is the first documentation of non-radioactive, in situ hybridization of digoxigenin labelled oligonucleotide probes for mRNA of NR1 receptor in the infant human brain. Pre-treating human brain tissue for ISH by microwaving for 12 min in Tris/ EDTA buffer permitted the study of human infant tissue that had been formalin fixed and paraffin embedded. Using the same methods on adult brain tissue improved hybridization signals, maintained minimum background signal, and maintained tissue morphology compared to previously published methods.
Acknowledgements The authors would like to acknowledge Dennis Dwarte for technical assistance in evaluating and quantifying hybridization results. The human adult hippocampal tissue was obtained from the NSW Tissue Resource Centre which is funded by NISAD and NHMRC network for brain research into mental
disorders with Central Sydney Area Health Service (CSAHS) approval (X98-0216). The project was funded by the National SIDS Council of Australia, the Ramaciotti Foundation, and Rita Machaalani is supported by a scholarship from CHATA, NSW.
References Bayer TA, Wiestler OD, Wolf HK. Hippocampal loss of N -methyl-Daspartate receptor subunit 1 mRNA in chronic temporal lobe epilepsy. Acta Neuropathol (Berl) 1995;89:446 /50. Dirks RW, van de Rijke FM, Fujishita S, van der Ploeg M, Raap AK. Methodologies for specific intron and exon RNA localization in cultured cells by haptenized and fluorochromized probes. J Cell Sci 1993;104:1187 /97. Eastwood SL, Harrison PJ. Detection and quantification of hippocampal synaptophysin messenger RNA in schizophrenia using autoclaved, formalin-fixed, paraffin wax-embedded sections. Neuroscience 1999;93:99 /106. Evers P, Uylings HB, Suurmeijer AJ. Antigen retrieval in formaldehyde-fixed human brain tissue. Methods 1998;15:133 /40. Karp SJ, Masu M, Eki T, Ozawa K, Nakanishi S. Molecular cloning and chromosomal localization of the key subunit of the human N methyl-D-aspartate receptor. J Biol Chem 1993;268:3728 /33. Lan HY, Mu W, Ng YY, Nikolic-Paterson DJ, Atkins RC. A simple, reliable, and sensitive method for nonradioactive in situ hybridization: use of microwave heating to improve hybridization efficiency and preserve tissue morphology. J Histochem Cytochem 1996;44:281 /7. Leong AS, Milios J, Duncis CG. Antigen preservation in microwaveirradiated tissues: a comparison with formaldehyde fixation. J Pathol 1988;156:275 /82. McMahon J, McQuaid S. The use of microwave irradiation as a pretreatment to in situ hybridization for the detection of measles virus and chicken anaemia virus in formalin-fixed paraffin-embedded tissue. Histochem J 1996;28:157 /64. McQuaid S, McMahon J, Allan GM. A comparison of digoxigenin and biotin labelled DNA and RNA probes for in situ hybridization. Biotech Histochem 1995;70:147 /54. Oliver KR, Heavens RP, Sirinathsinghji DJ. Quantitative comparison of pre-treatment regimens used to sensitize in situ hybridization using oligonucleotide probes on paraffin-embedded brain tissue. J Histochem Cytochem 1997;45:1707 /13. Oliver KR, Wainwright A, Heavens RP, Sirinathsinghji DJ. Retrieval of cellular mRNA in paraffin-embedded human brain using hydrated autoclaving. J Neurosci Methods 1997;77:169 /74. Planells-Cases R, Sun W, Ferrer-Montiel AV, Montal M. Molecular cloning, functional expression, and pharmacological characterization of an N -methyl-D-aspartate receptor subunit from human brain. Proc Natl Acad Sci USA 1993;90:5057 /61. Shi SR, Imam SA, Young L, Cote RJ, Taylor CR. Antigen retrieval immunohistochemistry under the influence of pH using monoclonal antibodies. J Histochem Cytochem 1995;43:193 /201.
Retrieval of mRNA from paraffin-embedded human infant brain tissue for non-radioactive in situ hybridization using oligonucleotides Bronwyn L. Relf a, Rita Machaalani a, Karen A. Waters a,b,c,* a
Department of Medicine, Room 206, Blackburn Building, DO6, The University of Sydney, Sydney, NSW 2006, Australia b Paediatrics and Child Health, The University of Sydney, Sydney, NSW 2006, Australia c The Childrens Hospital, Westmead, Sydney, Australia Received 9 July 2001; received in revised form 20 December 2001; accepted 20 December 2001
Abstract In situ hybridization (ISH) is used to examine the spatiotemporal distribution of gene expression in a range of tissues. Neuroscience research in human brain tissue requires techniques that can be used in formalin fixed and paraffin-embedded tissue rather than frozen tissue which is recommended, but difficult to obtain. This study presents a method for non-radioactive (DIG) ISH for detecting NR1 gene expression, in human infant brain tissue. We compared three pre-treatment effects, protease digestion, autoclaving (in citrate and Tris/EDTA buffer) and microwaving (in citrate and Tris/EDTA buffer). Tissue had been fixed in formalin for 2 /12 weeks. Results were compared for the hybridization and background signal intensities, and tissue morphology. We found that optimum results were obtained using 12-min microwave pre-treatment in Tris/EDTA buffer. This method produced optimum signal to background ratio in infant and adult tissue, preserved tissue morphology, and was suitable for use across a broad range of fixation times. # 2002 Elsevier Science B.V. All rights reserved. Keywords: In situ hybridization; Oligonucleotide; Non-radioactive; mRNA; Infant; Brainstem; Microwave
1. Introduction In situ hybridization (ISH) is used widely to examine the spatiotemporal distribution of gene expression in tissue. Recent studies have used ISH to identify gene expression differences in disease states of the brain, when compared to normal. Frozen tissue preparations are preferred over paraffin embedded tissue, because of the superior hybridization signal obtained (Oliver et al., 1997b). However, post-mortem brain tissue from humans is difficult to obtain, and is most often fixed so that frozen tissue is unavailable despite its perceived benefits. Research applications would be greatly extended if the methods of analysis were adapted for use in fixed tissue. Formalin, the most common fixative used, helps retain nucleic acids but forms extensive cross-linkages * Corresponding author. Tel.: 61-2-9351-5165; fax: 61-2-95503851. E-mail addresses: [email protected], karenw2@chw. edu.au (K.A. Waters).
in the process, hindering probe penetration (Eastwood and Harrison, 1999). The longer the period of fixation the greater the amount of cross-linkages formed and the harsher the pre-treatments required to expose the retained RNA to the probe. Pre-treatments that have been used to remove cross-linkages include proteinase K digestion (Bayer et al., 1995; Oliver et al., 1997a,b), hydrated autoclaving (Oliver et al., 1997a,b; Eastwood and Harrison, 1999) and microwave treatment in buffer solution (Lan et al., 1996; McMahon and McQuaid, 1996; Oliver et al., 1997a,b). Enzyme digestion with proteases, such as proteinase K, have been popular, but to achieve optimum results the conditions for digestion must be altered to suit the type of tissue and the duration of prior fixation. In addition, with enzymatic digestion the unmasking of RNA often occurs at the expense of tissue and cellular morphology (Lan et al., 1996). This has led to the use of hydrated autoclaving or microwaving as alternative pretreatments (Lan et al., 1996; McMahon and McQuaid, 1996; Oliver et al., 1997a,b; Eastwood and Harrison, 1999). Finally, radioactive labels have traditionally been
0165-0270/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 0 2 7 0 ( 0 2 ) 0 0 0 0 3 - 1
130
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
used for ISH. However, non-radioactive methods are increasingly popular because they produce faster results, the labelled probes are more stable, results are more consistent, and there is less biohazard to laboratory staff (McQuaid et al., 1995). Our laboratory has a specific interest in studying the neuropathology of sudden, unexplained death in infancy. The aim of this study was to determine the optimal pre-treatment for retrieving cellular mRNA from paraffin embedded human infant brainstem and hippocampal tissue. We compared several pre-treatment regimes on the hybridization signal intensity, background staining and tissue morphology in non-radioactive (digoxigenin) ISH to identify the NR1 subunit of the N -methyl-D-aspartate (NMDA) receptor in infant brain tissue.
2. Materials and methods 2.1. Tissue acquisition Archival, paraffin-embedded human infant (age 1 month to 1 year) brainstem tissue was obtained with the permission of the NSW Coroner. At the time of post-mortem, none of the cases had any neuropathology identified. The tissue was routinely fixed in 10% formalin, prior to paraffin embedding for routine diagnostic studies at the Institute of Forensic Medicine. Adult hippocampal tissue was obtained from the NSW Tissue Resource Centre and was used to confirm the method of Bayer et al. (1995) and subsequently, as positive controls. The study was approved by the Human Ethics Committee of the University of Sydney. All tissue sections were cut at 7 mm thickness and mounted on slides coated with 3-aminopropyltriethoxysilane, then dried at 37 8C for 24 h prior to in situ hybridization. 2.2. Oligonucleotide probes Synthetic oligonucleotide probes (Gibco BRL, Melbourne, Australia), complementary to the sub-unit specific region between transmembrane region I and transmembrane region II, were designed from the human NR1 gene (Karp et al., 1993; Planells-Cases et al., 1993). The probes had the following sequences: 5? CTC CTC CTC CTC GCT GTT CAC CTT GAA CCG GCC GAA GGG GCT GAA 3? (antisense probe, complementary to sequences encoding amino acid residues 565/579 of the mature NR1 protein) and 5? TTC AGC CCC TTC GGC CGG TTC AAG GTG AAC AGC GAG GAG GAG GAG 3? (NR-1 sense, reverse complement of NR-1 antisense oligonucleotide). The probes were designed to correspond to the most abundant isoform NR1A. This probe length was chosen
to optimise penetration in formalin-fixed tissue, whilst maintaining specificity. Anti-sense and sense probes were labelled by 3?-tailing with digoxigenin-11-dUTP according to the DIG Oligonucleotide tailing kit from Boehringer Mannheim (Mannheim, Germany). The reaction was incubated at 37 8C for 1 h. Labelled oligonucleotide was purified by ethanol precipitation. Probe concentration was determined by quantification against known standards on Hybond N filters, as described in the DIG System User’s Guide for Filter Hybridization (Boehringer, Mannheim, Germany). 2.3. Pre-treatment of sections prior to in situ hybridization All reactions were performed at room temperature unless noted otherwise. Solutions were made up in diethyl pyrocarbonate (DEPC) treated water. Sections were deparaffinised in xylene (2 /10 min), taken through a graded series of ethanols (1 /5 min in 100, 95, 75, 50 and 30%) and washed in DEPC H2O (2 /5 min). From this point sections were subjected to one of three different pre-treatments: 1.
2.
3.
Enzymatic digestion: sections were digested in 50, 100 or 200 mg/ml proteinase K (Boehringer Mannheim, Germany) made up in phosphate buffered saline (PBS) for 15 min at 37 8C. Digestion was halted by washing in PBS containing 2 mg/ml glycine (2 /5 min). Autoclaving: sections were placed in either 10 mM citrate buffer, pH 6.0 or in Tris/EDTA buffer (1 mM EDTA, 1 mM sodium citrate, 2 mM Tris; pH 9.0), autoclaved at 120 8C/2 bar for 40 min, cooled to 70 8C and washed in DEPC H2O (1 /15 min). Microwaving: sections were either placed in 10 mM citrate buffer or in Tris/EDTA buffer and microwaved at full power on a rotating turntable, uninterrupted for 12 min (Samsung 850 W, Korea). Sections were allowed to cool in the buffer for 15 min and DEPC H2O was added gradually over 5 min to reduce the buffer to room temperature. Sections were then washed in DEPC H2O (2 /5 min).
2.4. In situ hybridization All sections were subsequently washed in PBS (1 /5 min) and acetylated in 0.25% acetic anhydride in 0.1 M triethanolamine, pH 8.0 for 10 min. After washing in PBS (5 min), PAP pen (Zymed laboratories, San Francisco) was applied around the sections and the sections rinsed in PBS (5 min). Sections were prehybridized in 4/standard saline citrate (SSC)/50% formamide for 2 h at 37 8C, and rinsed in 2/SSC (5 min)
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
prior to overnight hybridization at 37 8C in a humidified chamber. Specifically, 50 ng of antisense probe in 100 ml of hybridization buffer was added just prior to hybridization, during which the sections were coverslipped with Parafilm. The composition of the hybridization solution was 18% formamide, 2 /SSC, 1 / Denhardt’s solution, 10% dextran sulfate, 50 mM dithiothreitol (DTT), 250 mg/ml yeast t-RNA, 100 mg/ ml polyadenylic acid and 500 mg/ml denatured and sheared salmon sperm DNA. Nonspecific hybridization was examined including slides that had been incubated with either the sense probe or hybridization solution only.
2.5. Post-hybridization Following hybridization, sections were washed in 2 / SSC at room temperature (15 min) until the parafilm washed off the section. Washes were all 15 min duration, and included 2 /SSC at 37 8C, 1/SSC at 37 8C, then at room temperature, 1/SSC, 0.5 /SSC and 0.25 / SSC.
2.6. Immunological detection Sections were then treated with either concentrated (100%) or diluted (10%) immunological detection buffers. After three washes with either 100% digoxigenin buffer (DB; 100 mM maleic acid, 150 mM NaCl, pH 7.5), or 10% DB, for 5 min each, blocking solution was added for 1 h at 37 8C. The composition of the blocking solution was 10% w/v blocking reagent [BR] (Roche Diagnostics), and 2.5% normal sheep serum (NSS) in the relevant concentration of DB. Anti-DIG alkaline phosphatase antibody (1:250 dilution in DB, containing 10% BR and 1% NSS) was added after blocking and the sections were incubated for 2 h on a rocking platform at room temperature. Non-specific staining was examined by the inclusion of sections without antibody treatment. All sections were washed in DB (2 /10 min) and then in either 100% digoxigenin detection buffer (DDB; 100 mM Tris, 100 mM NaCl, 50 mM MgCl2, pH 9.5) or 10% DDB for 10 min. The alkaline phosphatase color reaction buffer mixture was added according to the specifications of the manufacturer (Boehringer Mannheim). This buffer contained 35 ml nitroblue tetrazolium [NBT], and 45 ml 5-bromo-4-chloro-3-indolylphosphate [BCIP] in 10 ml DDB. Sections were incubated overnight (16 /18 h) in a humidified dark box. The color reaction was terminated by washing in Tris buffer (10 mM Tris, 0.1 mM EDTA, pH 8.0) (5 min) and distilled H2O (5 min). Sections were then mounted in Aquamount solution (BDH Laboratory Supplies Poole, UK).
131
2.7. Analysis of in situ hybridization results Pre-treatment effects on the signal intensity, background levels and tissue morphology were evaluated visually (under a light microscope) and compared using a visual scaling system. The following semiquantitative score was used for signal intensity: 0, negative; /, weak; //, moderate; ///, strong; and ////, very strong. Background staining was scored: /, low; //, medium; and ///, high. For tissue morphology the scores were: 0, no change; /, some tissue disruption; //, marked tissue disruption. Quantification by optical density measurement was only made for the sections included in Fig. 1, to provide representative measurements from each pre-treatment, and study group. Briefly, microscopy images were captured by a Sensicam (PCO Computer optics, Kelheim, Germany) attached to a microscope (Nikon Eclipse E800, Tokyo, Japan). The images were digitised, and optical density (OD) values were calculated using commercially available software (Zeiss KS 400, Munich, Germany). Optical density was measured for all neurones and five randomly selected background areas, for each section shown in Fig. 1. The study groups comprised tissue from the CA3 region of the adult hippocampus, and a hypoglossal nucleus at the level of the closed medulla from infant brainstem. The signal to background ratio was calculated by dividing the mean cellular optical density value by the mean background optical density value for each section. Statistical comparisons were performed using SPSS (V10 for Windows, Chicago). Optical density values were compared amongst groups (infant brainstem or adult hippocampus), signal and background intensity, and pre-treatment (proteinase K, autoclaving, or microwaving), as factors in a general linear model. Results are presented as mean9/S.E., and a P -value of B/0.05 was considered significant.
3. Results A total of 29 cases (25 infants and 4 adults) were studied, using one to two sections per case. The mean age was 3.79/2.6 months (range 1 /10) for infants, and 65.89/14.5 years (range 46 /78) for adults. Fixation time was 6.69/3.5 weeks (range 2/12) for infants, and 5.39/ 2.2 weeks (range 3 /8) for adults. To assess each pretreatment regime, qualitative analyses was performed for hybridization signal intensity, background staining and tissue morphology. See Table 1. The optical density measurements confirmed our visual scaling results and permitted quantification of the results. Note that within the range of fixation times we tested, no change was required in the protocol.
132
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
Fig. 1. Effect of pre-treatment on signal strength and background hybridization in adult hippocampal tissue (CA3 region, left column, panels A /C) and infant brainstem (hypoglossal nucleus at closed medulla, right column, panels D /F). Sections digested with 200 mg/ml proteinase K (A) and (D); sections autoclaved in Tris/EDTA buffer (B) and (E); Sections microwaved in Tris/EDTA buffer (C) and (F).
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
133
Table 1 Visual scale comparison of pre-treatment effects on signal intensity, background level and tissue morphology in adult and infant brain tissue Pre-treatment
Adult hippocampus
Infant brainstem
Signal
BG
Morphology
Signal
BG
Morphology
Proteinase K 50 mg/ml 100 mg/ml 200 mg/ml
0 0
0 0
0 0
0 0
Autoclave Tris/EDTA Citrate
Microwave Tris/EDTA Citrate
0
0 0
First column: 0, no signal; , weak; , moderate; , strong signal intensity; , very strong signal intensity and non specific staining. Second column: , low; , medium; , high background. Third column: 0, no effect on tissue morphology; , some tissue disruption; , marked tissue disruption. BG, background.
3.1. Effect of pre-treatment on signal intensity No positive hybridization signal was observed in sections hybridized with the sense probe, where probe was omitted or in any sections digested with 50 or 100 mg/ml proteinase K. Positive hybridization signals were observed in all other sections. Signal intensity varied amongst the pre-treatment groups and between adult and infant tissue. Generally, staining was localised in the nucleolus and cytoplasm but some nuclear and dendritic process staining was also observed. Similar patterns of hybridization have been reported by others using DIGlabelled oligonucleotide probes (Dirks et al., 1993; Bayer et al., 1995). In adult hippocampal tissue, the lowest signal intensity was observed after digestion with 200 mg/ml proteinase K (Fig. 1A and Table 1). Autoclaving with Tris/EDTA buffer gave a high, but non-specific, signal intensity, staining all cells including neurones and glia (Fig. 1B). Autoclaving with citrate buffer gave a medium signal intensity which was also non-specific (Table 1). Microwaving with either Tris/EDTA (Fig. 1C) or citrate buffer produced high signal intensities, which was localised to neuronal cells. In infant tissue, the highest signal intensities were observed when the sections were digested with 200 mg/ml proteinase K (Fig. 1D) or autoclaved in Tris/EDTA (Fig. 1E) or citrate buffer but in each case the staining was also non-specific. Microwaving produced a high signal intensity in Tris/EDTA buffer (Fig. 1F) and medium staining in citrate buffer, with the staining remaining specific to neurones in each case. Quantification of the pre-treatment effects on signal intensity, using optical density measurements, is shown in Fig. 2A.
Fig. 2. I, Adult hippocampal tissue; j, infant brainstem tissue. OD, optical density. Pre-treatments include protK 200 mg/ml proteinase K, AC tris autoclaved in Tris/EDTA buffer, and micro tris microwaved in Tris/EDTA buffer. (A) Optical density (OD) values of the neuronal hybridization signal. (B) Optical density (OD) values of the background tissue. (C) Signal to background ratio amongst the three pre-treatments.
134
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
Statistical analysis of the optical density measures showed significant differences amongst each of the pretreatments. Overall mean value for proteinase K/ 0.579/0.01, autoclaved in Tris/EDTA /1.289/0.05, and microwaved in Tris/EDTA /0.939/0.02 (P B/ 0.001 amongst all pre-treatments). These values reflect optical density in 52, 54 and 47 cells, respectively. 3.2. Effect of pre-treatment on background levels For both the adult and infant tissue, low background levels were observed after digestion in proteinase K at either 50 or 100 mg/ml. However, the absence of a hybridization signal in these sections indicated that the only color substrate binding was to antibody attached non-specifically to the tissue. The level of background staining seen in these sections was used as the baseline for comparing background levels after the other pretreatment regimes. In the adult hippocampus, medium background levels were observed in sections digested with 200 mg/ml proteinase K (Fig. 1A), and those autoclaved or microwaved in citrate buffer. High background levels were observed in sections autoclaved with Tris/EDTA buffer (Fig. 1B). A low background level, equivalent to baseline, was observed in the sections when microwaved in Tris/EDTA buffer (Fig. 1C). In the infant tissue, high background was seen after autoclaving in both citrate and Tris/EDTA (Fig. 1E, Table 1), whereas microwaving (Fig. 1F) and digestion with 200 mg/ml proteinase K (Fig. 1D) gave background levels equivalent to baseline (Table 1). Quantification of the pre-treatment effects on background levels, using optical density measurements, is shown in Fig. 2B. 3.3. Signal to background ratio In both the adult and infant brain tissue, the best (highest) signal to background ratio was obtained by microwaving in Tris/EDTA buffer. The worst (lowest) signal to background ratio for both adult and infant tissue was obtained in the Tris/EDTA autoclaved sections. Quantification of the pre-treatment effects on signal to background ratio, using optical density measurements, is shown in Fig. 2C. Statistical analyses were not performed for the signal to background ratio. Differences between signal and background were significant for all pre-treatment groups. Overall signal intensity (for all pre-treatments) was 0.869/0.04 and 1.029/0.05 for adult hippocampus and infant brainstem, respectively (P B/0.001). Overall background intensity was 0.379/0.08 and 0.319/0.06 for adult hippocampus and infant brainstem, respectively. Differences were significant for signal versus background P B/0.001, and interaction between signal and group (adult hippocampus vs. infant brainstem, P /0.005).
3.4. Effect of pre-treatment on tissue morphology Tissue morphology was assessed qualitatively. The parameters that were evaluated were the presence/ absence of identifiable cellular structures such as the nucleus, nucleolus, and indications that the cellular and nuclear membranes were intact. Tissue loss from the slide was also evaluated. Pre-treatment effects on tissue morphology were similar between adult and infant sections. Amongst the different pre-treatments autoclaving caused the greatest tissue disruption, with significant loss of tissue from slides. Digestion with 200 mg/ml proteinase K caused some tissue degradation, although less than that seen in autoclaved sections. In general, microwaving had no effect on tissue morphology, although some loss of tissue was observed from slides of adult brain tissue that were microwaved in citrate buffer. 3.5. Effect of dilution of the immunological detection buffers Dilution of the digoxygenin buffer (DB) and digoxygenin detection buffer (DDB) to 10%, reduced background staining in the tissue sections whilst maintaining signal strength and specificity. In contrast, washing with 100% DB and DDB caused higher levels of background staining and more non-specific staining of glial and neuronal cells. Tissue morphology also appeared to be better preserved in sections washed in the diluted buffers (Fig. 3).
4. Discussion In this study, three pre-treatment regimes were compared in formalin fixed and paraffin embedded human brain tissue, from adults and infants. The aim was to optimise the methods for detecting hybridization signal in fixed tissue, compared to previously published methods recommending the use of hydrated autoclaving (Oliver et al., 1997a,b; Eastwood and Harrison, 1999). Hydrated autoclaving was not suitable for use in infant brain tissue, because of significant degradation of tissue morphology, high background staining and high levels of non-specific staining. The new methods presented in this study resolved those problems in infant brain tissue, but also reduced the signal to noise ratio, and improved tissue morphology in the adult tissue sections, compared to the autoclaving method. The changes to pre-treatment that were required to examine infant brain tissue may relate to maturational differences between adult and infant tissues, particularly the level of myelination. The results of our study suggest that pre-treatments suitable for adult tissue were too harsh for infant brain tissue. The optimal pre-treatment gave the highest signal
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
135
Fig. 3. Effect of dilution of immunological buffers on hybridization signal strength, and background hybridization in infant brainstem tissue (hypoglossal nucleus). (A) Sections washed with concentrated (100%) DB and DDB. (B) Sections washed with diluted (10%) DB and DDB. (C) A section treated with sense probe, using diluted (10%) DB and DDB. All sections illustrated here were microwaved in Tris/EDTA buffer.
to background ratio whilst still maintaining tissue morphology in both adult and infant brain tissue. The advantages of using digoxigenin labelled probes permitted a 3-day development time compared to several weeks for radioactive probes, and the signal intensity could be quantified using optical density measurements. Microwaving is a suitable pre-treatment for in situ hybridization (McMahon and McQuaid, 1996), to prepare sections where enzyme digestion or hydrated autoclaving would otherwise be used. The heat produced by microwaving, or autoclaving, disrupts crosslinkages formed during formalin fixation (McMahon and McQuaid, 1996; Oliver et al., 1997a,b; Eastwood and Harrison, 1999). During microwaving, heat is produced by the oscillation of dipolar molecules (such as water and the side chains of protein molecules), that is continuous and proportional to the energy applied (Leong et al., 1988). The argument that uneven heating, and evaporation of the liquid during microwaving, produces uneven disruption of the cross-linkages was not confirmed in our results, where morphology and signal distribution were consistent throughout the sections.
The three major advantages with microwaving were lower signal to background ratio, preservation of tissue morphology, and shorter treatment time. To compensate for the same factors using enzyme digestion would require adjustments of time, temperature and of enzyme concentrations. Note also, that within the range of fixation times we tested, no change was required in the protocol, a finding that we attribute to the fact that we used microwaving in the pre-treatment regime. Thus microwaving enabled probe access to the target mRNA by breaking down the extensive cross-linkages formed during formalin fixation whilst preserving tissue morphology. Effects of changes in the buffer solutions may have several origins. The Tris/EDTA buffer, with a pH of 9.0, proved optimal during microwaving pre-treatment. This buffer gave higher signal intensity in all tissue types with less background compared to the citrate buffer. Studies have shown that the pH of the retrieval solution is an important factor where a high pH (pH 8 /10) largely improved staining results in immunohistochemistry (Shi et al., 1995). Citrate buffer, pH 6.0, not only has a low pH but utilises NaOH which has been noted to cause
136
B.L. Relf et al. / Journal of Neuroscience Methods 115 (2002) 129 /136
tissue loss from slides during microwaving (Evers et al., 1998). The different salt concentration of the buffers may also have contributed to their effects on signal strength. Tris/EDTA has a total salt concentration of 4 mM compared to citrate that has a salt concentration of 10 mM. The effects of diluting the immunological detection buffers, DB and DDB, are not well understood. However, research suggests that the reduction in salt concentrations may modify the hydrophobicity of the polypeptide chains which acts by reducing probe binding to free proteins (McMahon and McQuaid, 1996). Some non-specific staining of pre-mRNA was found in the nucleus as a consequence of our use of an oligonucleotide probe. Dirks et al. (1993), suggest that nuclear and nucleolar localization is almost certainly the result of using an oligonucleotide probe, since this also detects DNA and pre-mRNA sequences located in the cell nucleus. In contrast, RNA probes detect only mature mRNA sequences. The oligonucleotide probe we used did not span the intron, although this method may be an alternative means of detecting only mature RNA with cytoplasmic localisation. The oligonucleotide sequence we used was taken from Bayer et al. (1995) and we found and reported the same localisation of the probe as they did. To our knowledge, this is the first documentation of non-radioactive, in situ hybridization of digoxigenin labelled oligonucleotide probes for mRNA of NR1 receptor in the infant human brain. Pre-treating human brain tissue for ISH by microwaving for 12 min in Tris/ EDTA buffer permitted the study of human infant tissue that had been formalin fixed and paraffin embedded. Using the same methods on adult brain tissue improved hybridization signals, maintained minimum background signal, and maintained tissue morphology compared to previously published methods.
Acknowledgements The authors would like to acknowledge Dennis Dwarte for technical assistance in evaluating and quantifying hybridization results. The human adult hippocampal tissue was obtained from the NSW Tissue Resource Centre which is funded by NISAD and NHMRC network for brain research into mental
disorders with Central Sydney Area Health Service (CSAHS) approval (X98-0216). The project was funded by the National SIDS Council of Australia, the Ramaciotti Foundation, and Rita Machaalani is supported by a scholarship from CHATA, NSW.
References Bayer TA, Wiestler OD, Wolf HK. Hippocampal loss of N -methyl-Daspartate receptor subunit 1 mRNA in chronic temporal lobe epilepsy. Acta Neuropathol (Berl) 1995;89:446 /50. Dirks RW, van de Rijke FM, Fujishita S, van der Ploeg M, Raap AK. Methodologies for specific intron and exon RNA localization in cultured cells by haptenized and fluorochromized probes. J Cell Sci 1993;104:1187 /97. Eastwood SL, Harrison PJ. Detection and quantification of hippocampal synaptophysin messenger RNA in schizophrenia using autoclaved, formalin-fixed, paraffin wax-embedded sections. Neuroscience 1999;93:99 /106. Evers P, Uylings HB, Suurmeijer AJ. Antigen retrieval in formaldehyde-fixed human brain tissue. Methods 1998;15:133 /40. Karp SJ, Masu M, Eki T, Ozawa K, Nakanishi S. Molecular cloning and chromosomal localization of the key subunit of the human N methyl-D-aspartate receptor. J Biol Chem 1993;268:3728 /33. Lan HY, Mu W, Ng YY, Nikolic-Paterson DJ, Atkins RC. A simple, reliable, and sensitive method for nonradioactive in situ hybridization: use of microwave heating to improve hybridization efficiency and preserve tissue morphology. J Histochem Cytochem 1996;44:281 /7. Leong AS, Milios J, Duncis CG. Antigen preservation in microwaveirradiated tissues: a comparison with formaldehyde fixation. J Pathol 1988;156:275 /82. McMahon J, McQuaid S. The use of microwave irradiation as a pretreatment to in situ hybridization for the detection of measles virus and chicken anaemia virus in formalin-fixed paraffin-embedded tissue. Histochem J 1996;28:157 /64. McQuaid S, McMahon J, Allan GM. A comparison of digoxigenin and biotin labelled DNA and RNA probes for in situ hybridization. Biotech Histochem 1995;70:147 /54. Oliver KR, Heavens RP, Sirinathsinghji DJ. Quantitative comparison of pre-treatment regimens used to sensitize in situ hybridization using oligonucleotide probes on paraffin-embedded brain tissue. J Histochem Cytochem 1997;45:1707 /13. Oliver KR, Wainwright A, Heavens RP, Sirinathsinghji DJ. Retrieval of cellular mRNA in paraffin-embedded human brain using hydrated autoclaving. J Neurosci Methods 1997;77:169 /74. Planells-Cases R, Sun W, Ferrer-Montiel AV, Montal M. Molecular cloning, functional expression, and pharmacological characterization of an N -methyl-D-aspartate receptor subunit from human brain. Proc Natl Acad Sci USA 1993;90:5057 /61. Shi SR, Imam SA, Young L, Cote RJ, Taylor CR. Antigen retrieval immunohistochemistry under the influence of pH using monoclonal antibodies. J Histochem Cytochem 1995;43:193 /201.