- Email: [email protected]
Simultaneous detection of glutamic acid decarboxylase and reelin mRNA in adult rat neurons using in situ hybridization and immunofluorescence
Brain Research Protocols 3 Ž1998. 155–160
Protocol
Simultaneous detection of glutamic acid decarboxylase and reelin mRNA in adult rat neurons using in situ hybridization and immunofluorescence Christine Pesold a
a, )
, Maria G. Pisu a , Francesco Impagnatiello a , Doncho P. Uzunov a , Hector J. Caruncho a,b
Psychiatric Institute, Department of Psychiatry, College of Medicine, UniÕersity of Illinois at Chicago, 1601 West Taylor Street, M r C 912, Chicago, IL 60612, USA b Department of Fundamental Biology, UniÕersity of Santiago de Compostela, Galicia, Spain Accepted 10 July 1998
Abstract The combination of in situ hybridization and immunocytochemical technique is an important tool to detail the biochemical phenotype of individual neurons. In this work, we have developed a double fluorescence method to show the presence of reelin mRNA in GABAergic cells. This was achieved by demonstrating the colocalization of glutamic acid decarboxylase67 , the synthesizing enzyme for GABA, with the mRNA for reelin, a novel factor involved in brain development and possibly the maintenance of the synaptic organization of layered structures in adult brain. The results demonstrated that reelin is expressed primarily in GABAergic cells in the adult rat cerebrum, but not in the cerebellum. q 1998 Elsevier Science B.V. All rights reserved.
Themes: Development and regeneration Topics: Neurotrophic factors: expression and regulation Keywords: Reelin; Neurotrophic factor; In situ hybridization; Immunocytochemistry
1. Introduction Reelin is a protein produced by pioneer neurons that plays an important role in the development of the layered structures of the brain w2–4,8,9x. In addition, reelin has recently also been shown to be expressed in cortical, hippocampal, and cerebellar neurons of adult brain w4,10,11x. Reelin is an intracellular protein that appears to be released from the somata and neurites of Cajal–Retzius neurons during development and transported to axon terminals in adult brain w10x. Since reelin immunolabelling is not restricted to the cell somata, and in fact, in some cases, the somata intensity of labelling is very faint Ži.e., cerebellar granule cells., the detection of reelin mRNA by in situ hybridization could, therefore, be a more suitable approach to characterize the neurons in which it is produced.
To determine if reelin is expressed in GABAergic neurons, we developed a method of double labelling with immunofluorescence and in situ hybridization visualized with fluorescence for easy detection of colocalization of reelin mRNA and glutamic acid decarboxylase67 ŽGAD67 . protein.
2. Type of research Ø Determination of the proportion of reelin mRNA-containing cells that are GABAergic. Ø Can be adapted to determine the colocalization of other markers in neurons.
3. Time required )
Corresponding author. [email protected]
Fax:
q 1-312-433-8369;
E-mail:
Ø Perfusion and fixation: 24 h Ø Sucrose embedding: 48–72 h
1385-299Xr98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 5 - 2 9 9 X Ž 9 8 . 0 0 0 3 6 - 1
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
156
Ø Cryostat sectioning: variable Ø In situ hybridization immunofluorescence protocol: 3 days
4. Materials Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø
Adult Fisher rats: Harlan Chloral hydrate: Sigma, St. Louis, MO Paraformaldehyde: Fisher Scientific Sucrose: Fisher Scientific Cryostat: Microm, HM 505E Diethylpyrocarbonate ŽDEPC.: Sigma, St. Louis, MO Tween 20: Sigma, St. Louis, MO Formamide: Sigma, St. Louis, MO Sodium Chloride ŽNaCl.: Fisher Scientific Sodium Citrate: Fisher Scientific Saline Sodium Citrate ŽSSC; 20 = .: Sigma, St. Louis, MO Dextran Sulfate: Sigma, St. Louis, MO Salmon Sperm: Gibco, Grand Island, NY Yeast tRNA: Boehringer Mannheim, Indianapolis, IN Polyvinylpyrrolidone: Sigma, St. Louis, MO Ficoll: Sigma, St. Louis, MO Dithiothreitol ŽDTT.: Sigma, St. Louis, MO Digoxygenin-oligonucleotide tailing kit: Boehringer Mannheim, Indianapolis, IN Bovine serum albumin: Sigma, St. Louis, MO Mouse anti-digoxygenin: Boehringer Mannheim, Indianapolis, IN Rabbit anti-GAD67 : Chemicon, Temecula, CA Digoxygenin-labelled sheep anti-mouse: Boehringer Mannheim, Indianapolis, IN Fluorescein-labelled sheep anti-digoxygenin: Boehringer Mannheim, Indianapolis, IN Rhodamine-labelled goat anti-rabbit: Chemicon, Temecula, CA Mouse anti-biotin: Boehringer Mannheim, Indianapolis, IN Biotinylated goat anti-mouse: Vector Laboratories, Burlingame, CA Fluorescein-conjugated streptavidin: Boehringer Mannheim, Indianapolis, IN Biotinylated oligoprobe: DNA International, Lake Oswego, OR Glycerol gelatin: Sigma, St. Louis, MO 1,4-Diazabicyclo-w2.2.2xoctane ŽDABCO.: Sigma, St. Louis, MO Leica TCS-NT Confocal laser scanning microscope
5. Detailed procedure Ži. Anaesthetize rats with 15% chloral hydrate and perfuse them with 100 ml of phosphate-buffered saline
ŽPBS: 137 mM NaCl, 2.7 mM KCl, 5 mM Na 2 HPO4 , 1.7 mM KH 2 PO4 , pH 7.4., followed by 100 ml of cold fixative Ž4% paraformaldehyde in PBS, pH 7.4.. Žii. Carefully remove brains and post-fix them in 4% paraformaldehyde in PBS for 24 h at 48C. Žiii. Embed brains in a 30% sucrose solution in PBS at 48C for 2 to 3 days, or until the brains sink completely. Živ. Cut brains in a cryostat at 20 mm thickness and collect sections in cold PBS and rinse for 3 h. For long-term storage at y208C, collect sections in an anti-freezing solution containing 80 ml phosphate buffer Ž2.7 mM KCl, 5 mM Na 2 HPO4 , 1.7 mM KH 2 PO4 , pH 7.4., 60 ml glycerol and 60 ml ethylene glycol. Žv. To perform the double in situ hybridization-immunolabelling protocol, wash the sections in cold DEPC– PBS with 0.02% Tween 20 for 1 h, followed by a second wash in cold DEPC–PBS alone ŽDEPC–PBS has to be autoclaved prior to use, to destroy the DEPC molecules.. All relevant materials should be autoclaved or pre-sterilized Že.g., tubes, pipette tips, buffers. and gloves and masks should be worn until the mRNA-oligoprobe hybrid is complete Žstep vii., to prevent RNAse contamination. Žvi. Hybridize free-floating sections overnight at 428C with a hybridization buffer containing 50 pmolrml of an antisense oligoprobe for reelin Ž1264–1311 bp; base numbers are given according to the mouse reelin coding sequence, accession aU24703. biotinylated at positions 1267,1278,1289, 1299, 1307 ŽDNA International, Lake Oswego, OR.. Hybridization buffer: 50% deionized formamide; 2 = SSC Ž0.3 M NaCl, 0.03 M sodium citrate, pH 7.0.; 10% dextran sulfate ŽMW 500 K.; 500 mgrml sheared salmon sperm DNA; 250 mgrml yeast tRNA; 1 = Denhardt’s solution Ž0.02% polyvinylpyrrolidone, 0.02% Ficoll, and 0.02% acetylated BSA.; 100 mM DTT; add autoclaved 0.1% DEPCP H 2 O to make final volume. Žvii. After the hybridization step, wash the samples four times for 15 min in 1 = SSC Ž150 mM NaCl; 15 mM sodium citrate, pH 7. at room temperature ŽRT., followed by three 15-min washes in PBS at RT. Žviii. Incubate sections at RT for 30 min in a blocking solution containing 1% bovine serum albumin ŽBSA. in PBS. Žix. Incubate overnight at 48C in a solution of 1% BSA in PBS containing a primary antibody against biotin Žmouse anti-biotin monoclonal antibody from Boehringer Mannheim, 2 mgrml., and a primary antibody raised to recognize GAD67 Žrabbit polyclonal anti-GAD67 from Chemicon, 1:2000 w7x.. Žx. The next day, continue incubation in primary antibodies for 2 h at RT. Žxi. Wash sections twice for 15 min in 1% BSA in PBS at RT. Žxii. Incubate sections for 1 h at RT in a solution of 1% BSA in PBS containing a secondary antibody that will recognize the one labelling biotin Žbiotinylated goat antimouse antibody from Vector Labs, 1:250..
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
Žxiii. Wash sections twice for 15 min in 1% BSA in PBS at RT. Žxiv. For the final incubation step, incubate sections in the dark for 1 h at RT in a 1% BSA in PBS solution containing a fluorescein-labelled compound that will identify biotin Žstreptavidin–fluorescein from Boehringer Mannheim, 40 mgrml., and a rhodamine-tagged antibody that will recognize the antibody labelling GAD67 Žgoat anti-rabbit conjugated with rhodamine from Chemicon, 1:100.. In this way, the reelin mRNA will be revealed by the presence of fluorescein, and the GAD67 protein will be revealed by the presence of rhodamine. Žxv. Wash sections twice for 15 min in 1% BSA, twice in PBS for 15 min, and 15 min in H 2 O. Žxvi. Mount floating sections onto slides and let dry thoroughly in the dark Žminimum 2 h.. Žxvii. Warm the glycerol–gelatin mounting media ŽSigma. to 428C and add 25 mgrml of anti-fading agent ŽDABCO, Sigma., mixing gently until dissolved. While still warm, apply several drops onto slides and cover quickly with cover glass, applying pressure to squeeze out excess mounting media before it cools and hardens. Once cooled, sections are ready to be observed under a fluorescent microscope andror studied with a laser confocal microscope ŽLeica TCS-NT.. Žxviii. Colocalization of reelin mRNA with the GAD67 protein can be determined with a fluorescent microscope by switching back and forth between a filter that will allow the visualization of fluorescein, and one that will allow rhodamine visualization. One can begin by determining the number of reelin mRNA-positive cells that contain GAD67 , and conversely the number of GAD67 -positive cells that contain reelin. If a confocal microscope is used, simultaneous visualization of reelin mRNA and GAD67 protein can be achieved by scanning with an ArKr laser 488 nm line to excite fluorescein, and 568 nm line to excite rhodamine, using a dichroic beam splitter RSP 580 and a long-pass filter LP 590.
6. Results Using this approach, it is possible to analyze the distribution of reelin mRNA expressing neurons and their colocalization with GAD67 in adult rat brain w10x. Using the appropriate filters, reelin mRNA appears as small green points in the cell somata, while GAD67 expression can be seen as red, filling the entire cell somata and proximal dendrites, as well as some more distal axon terminals or dendritic endings Žsee Fig. 1.. In fact, one should be careful when examining the sections because in some cases Ži.e., cortical and hippocampal pyramidal cells., non-GABAergic cells are almost completely surrounded by punctate GAD67 immunoreactivity Žpossibly GABAergic terminals. and can easily be mistaken for GABAergic
157
neurons. No reelin labelling was observed when we used a sense oligoprobe corresponding to the same base pairs. This in situ hybridization and immunocytochemistry double fluorescence approach revealed that most of the reelin-positive cells in the hippocampus are GABAergic Ž93–97%., whereas in the neocortex, only about 70% of the reelin-positive cells are GABAergic Žsee Fig. 1.. Furthermore, not all GABAergic cells express reelin, as shown by the lack of reelin mRNA detected in many GAD67 -positive neurons.
7. Discussion This in situ hybridization and immunocytochemistry double fluorescence technique has proved to be successful for determining which proportion of reelin-positive neurons are GABAergic. This protocol has been tried in both adult and newborn rats and yields very good results w10x. While this protocol has not been tried in other animal species, it should produce similar results, providing that the antisense oligoprobe is designed against a homologous part of the reelin mRNA sequence. In general, the use of fluorescence for confocal microscopy is not ideal for aged mammalian or human brain tissue due to the confound introduced by the highly auto-fluorescent lipofuscins. 7.1. General There are several steps in the protocol that could be considered essential for the success of the experiment. Ø Our experience showed that the use of paraformaldehyde perfused tissue improved the immunolabelling results without affecting the in situ hybridization labelling. Ø To improve permeability of the oligoprobe and the antibodies, we also used Tween 20 in the initial washing step. Omission of the detergent will lower the intensity of labelling. Ø The use of free-floating sections will improve the permeability of the oligoprobe due to the increased exposure of the labelled oligoprobe to the section surfaces. Generally, when using a conventional in situ hybridization approach with a radioactive probe, this is not necessary since labelling detection can be enhanced by simply increasing the exposure time of the radiolabelled sections to the emulsion. In contrast, the intensity of the fluorescent label will depend on the number of fluorophores bound to the probe which is optimized when using free-floating sections. In our experience, using sections mounted on slides yielded poor fluorescent labelling. Ø Amplification steps are necessary to increase the fluorescent signal for reelin mRNA. Instead of simply using a primary antibody conjugated with fluorescein to recognize the biotin-labelled probe, we used two intermediate antibody steps for the purpose of amplifying the
158 C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160 Fig. 1. Colocalization of reelin mRNA and GAD67 in adult rat brain. Confocal microscope images of reelin mRNA detected by in situ hybridization visualized with fluorescein Žgreen., and GAD67 immunolabelling visualized with rhodamine Žred., in adult rat neocortex Žtop. and hippocampus Žbottom.. The overlay of reelin and GAD67 distribution shows that in the cortex, mRNA for reelin is found both in GABAergic and non-GABAergic cells, whereas in the hippocampus ŽCA2., all reelin-positive cells are GABAergic. Note also the different labelling pattern of GAD67 in the cytoplasm of GABAergic cells, and that of the putative GAD67 axon terminals or dendritic endings surrounding the cell bodies of pyramidal cells in the hippocampus.
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
number of fluorescent molecules that will recognize the reelin oligoprobe. When omitting the amplification steps, the fluorescence obtained is faint, at best. Ø In the in situ hybridization immunocytochemistry protocol described in this paper, in situ hybridization is performed first. The mRNA-oligoprobe hybrids which are formed in the first steps of the protocol are very stable and are not affected by RNAses. The protocol for the immunolabelling which is performed in the second step of the paper can, therefore, be carried out without any restrictions with respect to temperature Žsee below.. Ø The antisense oligoprobe used in this protocol corresponds to reelin base pairs 1264–1311. We have also obtained similar results using an antisense oligoprobe for reelin corresponding to base pairs 10225–10282. Whichever oligoprobe is chosen, one should include an experimental control condition in which sections are exposed to the corresponding sense oligoprobe. Ø GABA is synthesized in the adult brain by two forms of glutamic acid decarboxylase: GAD67 and GAD65 . However, this protocol proposes the use of an antibody against GAD67 ŽChemicon. to reveal GABAergic cells, since this antibody reveals the presence of GAD67 in cell bodies, as well as dendrites, whereas antibodies directed against GAD65 Ži.e., GAD-6 from Boehringer Mannheim. show that it is present mostly in nerve endings w7x. 7.2. AlternatiÕe and support protocols There are several protocols in the literature describing the use of a combination of in situ hybridization and immunohistochemistry w1,5,6,12x. Ø Some in situ hybridization immunohistochemistry protocols recommend performing the immunolabelling steps first, followed by the in situ hybridization to assure that the immunolabelling of antigens will not be compromised by some of the agents used in the in situ protocol w1,6x. However, the use of some variations with respect to conventional in situ hybridization protocols, will reduce this problem, allowing a nice double labelling w4,10x. When performing the immunolabelling first however, one must be careful to carry out all of the steps at 48C without shaking, to preserve the integrity of the mRNAs for the ensuing in situ hybridization steps. Ø Some of the protocols combining in situ hybridization and immunocytochemistry use a radioactive oligoprobe for the in situ, and a biotinylated secondary antibody to detect the immunocytochemical reaction with 3-3X-diaminobenzidine ŽDAB.. The main disadvantages with this approach include the time necessary to expose the radioactively-labelled sections to the emulsion, and the radioactive waste. Ø Alternatively, other protocols use a double colorimetric approach for the combination of immunocytochemistry and in situ hybridization, using digoxygenin-labelled oligo-
159
probes w1,5,6,12x. While this approach yields comparable results to our double fluorescent method, the use of fluorescence labelling allows one to analyze the samples at almost subcellular resolution with the help of a laser confocal microscope. Ø The use of fluorescein-conjugated antibodies can be substituted in this protocol by the use of antibodies conjugated to carbocyanin 2 ŽCy2., which has similar spectral properties but is considerably brighter and more stable. Similarly, the use of rhodamine-conjugated anti-rabbit can be substituted by a Texas Red-conjugated anti-rabbit, which has a higher emission spectra and can, therefore, be better distinguished from fluorescein when using a laser scanning confocal microscope. Ø Additional amplification steps can be included to increase the labelling of reelin, such as biotinyl–tyramide, as well as fluorescent antibodies against streptavidin. Ø Finally, this protocol describes an in situ hybridization using an oligoprobe labelled with biotin, rather than the more commonly used digoxygenin-labelled probe. One advantage of using an oligoprobe where biotin is chemically bound in specific positions within the sequence, is that it gives a much stronger signal, due in part to the very high affinity of streptavidin for biotin, and to the lesser steric hindrance caused by the biotin–streptavidin interaction. Another advantage is that it is sometimes difficult to assure successful end-labelling of the oligoprobe with digoxygenin. However, in some cases, the biotinylation of the probe may interfere with hybridization. Alternatively, one can end-label the reelin oligonucleotide probe with digoxygenin and change several of the subsequent labelling steps; however, the immunocytochemistry protocol remains unchanged. When using a digoxygenin-labelled oligoprobe, follow the same protocol described in this paper with the exception of the following appropriately numbered steps mentioned below. Žvi. Hybridize free-floating sections overnight at 428C in a hybridization buffer Žsee above. containing 50 pmolrml of a digoxygenin-labelled antisense oligoprobe corresponding to the bp 1264–1301 of the rat reelin sequence Ž5X-GCT CCC AAG ACA TCC CAT CCT CTG GGC TGG CTC TCA AAT TCT TCT G-3X .. To increase the sensitivity of this procedure, one can use a mixture of oligoprobes that are complimentary to the different regions of the target mRNA. Žix. Incubate overnight at 48C in a solution of 1% BSA in PBS containing a primary antibody against digoxygenin Žmouse anti-digoxygenin monoclonal antibody from Boehringer Mannheim, 0.4 mgrml., and a primary antibody raised to recognize GAD67 Žrabbit polyclonal antiGAD67 from Chemicon, 1:2000.. Žxii. Incubate sections for 1 h at RT in a solution of 1% BSA in PBS containing a secondary antibody that will recognize the one labelling digoxygenin Žanti-mouse antibody conjugated with digoxygenin from Boehringer Mannheim, 40 mgrml..
160
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
Žxiv. For the final incubation step, incubate sections in the dark for 1 h at RT in a 1% BSA in PBS solution containing a fluorescein-labelled antibody against digoxygenin Žsheep anti-digoxygenin conjugated with fluorescein from Boehringer Mannheim, 0.4 mgrml., and a rhodamine-tagged antibody that will recognize the antibody labelling GAD67 Žgoat anti-rabbit conjugated with rhodamine from Chemicon, 1:100.. In this way, the reelin mRNA will be revealed by the presence of fluorescein, and the GAD67 protein will be revealed by the presence of rhodamine. 7.3. Troubleshooting This protocol should yield good results if followed closely. However, deviations from the protocol or slight oversights may lead to problems. The following is a list of potential factors that may contribute to such problems. Ø Little or no signal can occur for several reasons: RNA can be destroyed if the sections are not kept at 48C prior to hybridization; amplification steps may be insufficient; secondary antibody raised against species different from that in which primary antibody was raised. Ø Too much background: Insufficient washing steps; antibody concentration too high; RNAse contamination due to lack of sterility; lack of specificity resulting from too numerous amplification steps.
8. Quick procedure Ži–ii. Perfuse rats with 4% paraformaldehyde, remove brains and post-fix for 24 h. Žiii–iv. Cut 20 mm cryostat sections through 30% sucrose embedded brains, and rinse in PBS. Žv–vi. Permeabilize sections in 0.02% Tween 20 in DEPC–PBS Ž1 h., rinse in DEPC–PBS, and hybridize overnight in a cocktail containing the biotinylated oligoprobe. Žvii–x. Following several high stringency rinses in SSC and PBS, block sections with 1% BSA Ž30 min., and incubate overnight at 48C followed by 2 h at RT in mouse anti-biotin Ž2 mgrml. and rabbit anti-GAD67 Ž1:2000.. Žxi–xii. Following two 15-min washes in 1% BSA in PBS ŽRT., incubate sections for 1 h at RT with biotinylated anti-mouse IgG Ž1:250 in 1% BSA.. Žxiii–xiv. Following two 15-min washes in 1% BSA in PBS, incubate sections in the dark for 1 h with streptavidin–fluorescein Ž40 mgrml. and rhodamine-labelled goat anti-rabbit Ž1:100. in 1% BSA in PBS. Žxv–xvi. Wash sections thoroughly with 1% BSA, PBS, and water before mounting onto slides.
Žxvii–xviii. Once dry, cover with anti-fading mounting media, coverslip, and study with a Leica laser confocal microscope.
9. Essential literature reference Refs. w1,5,6,10,12x.
References w1x M. Brahic, M. Ozden, Simultaneous detection of cellular RNA and proteins, in: D.G. Wilkinson ŽEd.., In Situ Hybridization. A Practical Approach, Oxford Univ Press, New York, 1992, pp. 85–104. w2x G. D’Arcangelo, G.G. Miao, S.C. Chen, H.D. Soares, J.I. Morgan, T. Curran, A protein related to extracellular matrix proteins deleted in the mouse mutant reeler, Nature 374 Ž1995. 719–723. w3x G. D’Arcangelo, K. Nakajima, T. Miyata, M. Ogawa, K. Mikoshiba, T. Curran, Reelin is a glycoprotein recognized by the CR-50 monoclonal antibody, J. Neurosci. 17 Ž1997. 23–31. w4x J.A. Del Rio, B. Heimrich, V. Borrell, E. Forster, A. Drakew, S. Alcantara, K. Nakajima, T. Miyata, M. Ogawa, K. Mikoshiba, P. Derer, M. Frotscher, E. Soriano, A role for Cajal–Retzius cells and reelin in the development of hippocampal connections, Nature 385 Ž1997. 70–74. w5x P.C. Emson, B. Heppelmann, S.J. Augood, Development of techniques to combine isotopic and nonisotopic in situ hybridization and immunocytochemistry for phenotypic characterization of individual neurons, in: J.H. Eberwine, K.L. Valentino, J.D. Barchas ŽEds.., In Situ Hybridization in Neurobiology. Advances in Methodology, Oxford Univ Press, New York, 1994, 43–62. w6x M. Jaber, M. Cador, B. Dumartin, E. Normand, L. Stinus, B. Bloch, Acute and chronic amphetamine treatments differently regulate neuropeptide messenger RNA levels and FOS immunoreactivity in rat striatal neurons, Neuroscience 65 Ž1995. 1041–1050. w7x D.L. Kaufman, C.R. Houser, A.J. Tobin, Two forms of the g-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions, J. Neurochem. 56 Ž1991. 720–723. w8x T. Miyata, K. Nakajima, J. Aruga, S. Takahashi, K. Ikenaka, K. Mikoshiba, M. Ogawa, Distribution of a reeler gene-related antigen in the developing cerebellum: an immunohistochemical study with an allogenic antibody CR-50 on normal and reeler mice, J. Comp. Neurol. 372 Ž1996. 215–228. w9x M. Ogawa, T. Miyata, K. Nakajima, K. Yagyu, M. Seike, K. Ikenaka, H. Yamamoto, K. Mikoshiba, The reeler gene-associated antigen on Cajal–Retzius neurons is a crucial molecule for laminar organization of cortical neurons, Neuron 14 Ž1995. 899–912. w10x C. Pesold, F. Impagnatiello, M.G. Pisu, D.P. Uzunov, E. Costa, A. Guidotti, H.J. Caruncho, Reelin is preferentially expressed in neurons synthesizing g-aminobutyric acid ŽGABA. in cortex and hippocampus of adult rats, Proc. Natl. Acad. Sci. USA 95 Ž1998. 3221–3226. w11x S.N. Schiffmann, B. Bernier, A.M. Goffinet, Reelin mRNA expression during mouse brain development, Eur. J. Neurosci. 9 Ž1997. 1055–1071. w12x P. Wahle, Combining non-radioactive in situ hybridization with immunohistological and anatomical techniques, in: W. Wisden, B.J. Morris ŽEds.., In Situ Hybridization Protocols For The Brain, Academic Press, London, 1994, pp. 98–120.
Protocol
Simultaneous detection of glutamic acid decarboxylase and reelin mRNA in adult rat neurons using in situ hybridization and immunofluorescence Christine Pesold a
a, )
, Maria G. Pisu a , Francesco Impagnatiello a , Doncho P. Uzunov a , Hector J. Caruncho a,b
Psychiatric Institute, Department of Psychiatry, College of Medicine, UniÕersity of Illinois at Chicago, 1601 West Taylor Street, M r C 912, Chicago, IL 60612, USA b Department of Fundamental Biology, UniÕersity of Santiago de Compostela, Galicia, Spain Accepted 10 July 1998
Abstract The combination of in situ hybridization and immunocytochemical technique is an important tool to detail the biochemical phenotype of individual neurons. In this work, we have developed a double fluorescence method to show the presence of reelin mRNA in GABAergic cells. This was achieved by demonstrating the colocalization of glutamic acid decarboxylase67 , the synthesizing enzyme for GABA, with the mRNA for reelin, a novel factor involved in brain development and possibly the maintenance of the synaptic organization of layered structures in adult brain. The results demonstrated that reelin is expressed primarily in GABAergic cells in the adult rat cerebrum, but not in the cerebellum. q 1998 Elsevier Science B.V. All rights reserved.
Themes: Development and regeneration Topics: Neurotrophic factors: expression and regulation Keywords: Reelin; Neurotrophic factor; In situ hybridization; Immunocytochemistry
1. Introduction Reelin is a protein produced by pioneer neurons that plays an important role in the development of the layered structures of the brain w2–4,8,9x. In addition, reelin has recently also been shown to be expressed in cortical, hippocampal, and cerebellar neurons of adult brain w4,10,11x. Reelin is an intracellular protein that appears to be released from the somata and neurites of Cajal–Retzius neurons during development and transported to axon terminals in adult brain w10x. Since reelin immunolabelling is not restricted to the cell somata, and in fact, in some cases, the somata intensity of labelling is very faint Ži.e., cerebellar granule cells., the detection of reelin mRNA by in situ hybridization could, therefore, be a more suitable approach to characterize the neurons in which it is produced.
To determine if reelin is expressed in GABAergic neurons, we developed a method of double labelling with immunofluorescence and in situ hybridization visualized with fluorescence for easy detection of colocalization of reelin mRNA and glutamic acid decarboxylase67 ŽGAD67 . protein.
2. Type of research Ø Determination of the proportion of reelin mRNA-containing cells that are GABAergic. Ø Can be adapted to determine the colocalization of other markers in neurons.
3. Time required )
Corresponding author. [email protected]
Fax:
q 1-312-433-8369;
E-mail:
Ø Perfusion and fixation: 24 h Ø Sucrose embedding: 48–72 h
1385-299Xr98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 5 - 2 9 9 X Ž 9 8 . 0 0 0 3 6 - 1
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
156
Ø Cryostat sectioning: variable Ø In situ hybridization immunofluorescence protocol: 3 days
4. Materials Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø Ø
Adult Fisher rats: Harlan Chloral hydrate: Sigma, St. Louis, MO Paraformaldehyde: Fisher Scientific Sucrose: Fisher Scientific Cryostat: Microm, HM 505E Diethylpyrocarbonate ŽDEPC.: Sigma, St. Louis, MO Tween 20: Sigma, St. Louis, MO Formamide: Sigma, St. Louis, MO Sodium Chloride ŽNaCl.: Fisher Scientific Sodium Citrate: Fisher Scientific Saline Sodium Citrate ŽSSC; 20 = .: Sigma, St. Louis, MO Dextran Sulfate: Sigma, St. Louis, MO Salmon Sperm: Gibco, Grand Island, NY Yeast tRNA: Boehringer Mannheim, Indianapolis, IN Polyvinylpyrrolidone: Sigma, St. Louis, MO Ficoll: Sigma, St. Louis, MO Dithiothreitol ŽDTT.: Sigma, St. Louis, MO Digoxygenin-oligonucleotide tailing kit: Boehringer Mannheim, Indianapolis, IN Bovine serum albumin: Sigma, St. Louis, MO Mouse anti-digoxygenin: Boehringer Mannheim, Indianapolis, IN Rabbit anti-GAD67 : Chemicon, Temecula, CA Digoxygenin-labelled sheep anti-mouse: Boehringer Mannheim, Indianapolis, IN Fluorescein-labelled sheep anti-digoxygenin: Boehringer Mannheim, Indianapolis, IN Rhodamine-labelled goat anti-rabbit: Chemicon, Temecula, CA Mouse anti-biotin: Boehringer Mannheim, Indianapolis, IN Biotinylated goat anti-mouse: Vector Laboratories, Burlingame, CA Fluorescein-conjugated streptavidin: Boehringer Mannheim, Indianapolis, IN Biotinylated oligoprobe: DNA International, Lake Oswego, OR Glycerol gelatin: Sigma, St. Louis, MO 1,4-Diazabicyclo-w2.2.2xoctane ŽDABCO.: Sigma, St. Louis, MO Leica TCS-NT Confocal laser scanning microscope
5. Detailed procedure Ži. Anaesthetize rats with 15% chloral hydrate and perfuse them with 100 ml of phosphate-buffered saline
ŽPBS: 137 mM NaCl, 2.7 mM KCl, 5 mM Na 2 HPO4 , 1.7 mM KH 2 PO4 , pH 7.4., followed by 100 ml of cold fixative Ž4% paraformaldehyde in PBS, pH 7.4.. Žii. Carefully remove brains and post-fix them in 4% paraformaldehyde in PBS for 24 h at 48C. Žiii. Embed brains in a 30% sucrose solution in PBS at 48C for 2 to 3 days, or until the brains sink completely. Živ. Cut brains in a cryostat at 20 mm thickness and collect sections in cold PBS and rinse for 3 h. For long-term storage at y208C, collect sections in an anti-freezing solution containing 80 ml phosphate buffer Ž2.7 mM KCl, 5 mM Na 2 HPO4 , 1.7 mM KH 2 PO4 , pH 7.4., 60 ml glycerol and 60 ml ethylene glycol. Žv. To perform the double in situ hybridization-immunolabelling protocol, wash the sections in cold DEPC– PBS with 0.02% Tween 20 for 1 h, followed by a second wash in cold DEPC–PBS alone ŽDEPC–PBS has to be autoclaved prior to use, to destroy the DEPC molecules.. All relevant materials should be autoclaved or pre-sterilized Že.g., tubes, pipette tips, buffers. and gloves and masks should be worn until the mRNA-oligoprobe hybrid is complete Žstep vii., to prevent RNAse contamination. Žvi. Hybridize free-floating sections overnight at 428C with a hybridization buffer containing 50 pmolrml of an antisense oligoprobe for reelin Ž1264–1311 bp; base numbers are given according to the mouse reelin coding sequence, accession aU24703. biotinylated at positions 1267,1278,1289, 1299, 1307 ŽDNA International, Lake Oswego, OR.. Hybridization buffer: 50% deionized formamide; 2 = SSC Ž0.3 M NaCl, 0.03 M sodium citrate, pH 7.0.; 10% dextran sulfate ŽMW 500 K.; 500 mgrml sheared salmon sperm DNA; 250 mgrml yeast tRNA; 1 = Denhardt’s solution Ž0.02% polyvinylpyrrolidone, 0.02% Ficoll, and 0.02% acetylated BSA.; 100 mM DTT; add autoclaved 0.1% DEPCP H 2 O to make final volume. Žvii. After the hybridization step, wash the samples four times for 15 min in 1 = SSC Ž150 mM NaCl; 15 mM sodium citrate, pH 7. at room temperature ŽRT., followed by three 15-min washes in PBS at RT. Žviii. Incubate sections at RT for 30 min in a blocking solution containing 1% bovine serum albumin ŽBSA. in PBS. Žix. Incubate overnight at 48C in a solution of 1% BSA in PBS containing a primary antibody against biotin Žmouse anti-biotin monoclonal antibody from Boehringer Mannheim, 2 mgrml., and a primary antibody raised to recognize GAD67 Žrabbit polyclonal anti-GAD67 from Chemicon, 1:2000 w7x.. Žx. The next day, continue incubation in primary antibodies for 2 h at RT. Žxi. Wash sections twice for 15 min in 1% BSA in PBS at RT. Žxii. Incubate sections for 1 h at RT in a solution of 1% BSA in PBS containing a secondary antibody that will recognize the one labelling biotin Žbiotinylated goat antimouse antibody from Vector Labs, 1:250..
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
Žxiii. Wash sections twice for 15 min in 1% BSA in PBS at RT. Žxiv. For the final incubation step, incubate sections in the dark for 1 h at RT in a 1% BSA in PBS solution containing a fluorescein-labelled compound that will identify biotin Žstreptavidin–fluorescein from Boehringer Mannheim, 40 mgrml., and a rhodamine-tagged antibody that will recognize the antibody labelling GAD67 Žgoat anti-rabbit conjugated with rhodamine from Chemicon, 1:100.. In this way, the reelin mRNA will be revealed by the presence of fluorescein, and the GAD67 protein will be revealed by the presence of rhodamine. Žxv. Wash sections twice for 15 min in 1% BSA, twice in PBS for 15 min, and 15 min in H 2 O. Žxvi. Mount floating sections onto slides and let dry thoroughly in the dark Žminimum 2 h.. Žxvii. Warm the glycerol–gelatin mounting media ŽSigma. to 428C and add 25 mgrml of anti-fading agent ŽDABCO, Sigma., mixing gently until dissolved. While still warm, apply several drops onto slides and cover quickly with cover glass, applying pressure to squeeze out excess mounting media before it cools and hardens. Once cooled, sections are ready to be observed under a fluorescent microscope andror studied with a laser confocal microscope ŽLeica TCS-NT.. Žxviii. Colocalization of reelin mRNA with the GAD67 protein can be determined with a fluorescent microscope by switching back and forth between a filter that will allow the visualization of fluorescein, and one that will allow rhodamine visualization. One can begin by determining the number of reelin mRNA-positive cells that contain GAD67 , and conversely the number of GAD67 -positive cells that contain reelin. If a confocal microscope is used, simultaneous visualization of reelin mRNA and GAD67 protein can be achieved by scanning with an ArKr laser 488 nm line to excite fluorescein, and 568 nm line to excite rhodamine, using a dichroic beam splitter RSP 580 and a long-pass filter LP 590.
6. Results Using this approach, it is possible to analyze the distribution of reelin mRNA expressing neurons and their colocalization with GAD67 in adult rat brain w10x. Using the appropriate filters, reelin mRNA appears as small green points in the cell somata, while GAD67 expression can be seen as red, filling the entire cell somata and proximal dendrites, as well as some more distal axon terminals or dendritic endings Žsee Fig. 1.. In fact, one should be careful when examining the sections because in some cases Ži.e., cortical and hippocampal pyramidal cells., non-GABAergic cells are almost completely surrounded by punctate GAD67 immunoreactivity Žpossibly GABAergic terminals. and can easily be mistaken for GABAergic
157
neurons. No reelin labelling was observed when we used a sense oligoprobe corresponding to the same base pairs. This in situ hybridization and immunocytochemistry double fluorescence approach revealed that most of the reelin-positive cells in the hippocampus are GABAergic Ž93–97%., whereas in the neocortex, only about 70% of the reelin-positive cells are GABAergic Žsee Fig. 1.. Furthermore, not all GABAergic cells express reelin, as shown by the lack of reelin mRNA detected in many GAD67 -positive neurons.
7. Discussion This in situ hybridization and immunocytochemistry double fluorescence technique has proved to be successful for determining which proportion of reelin-positive neurons are GABAergic. This protocol has been tried in both adult and newborn rats and yields very good results w10x. While this protocol has not been tried in other animal species, it should produce similar results, providing that the antisense oligoprobe is designed against a homologous part of the reelin mRNA sequence. In general, the use of fluorescence for confocal microscopy is not ideal for aged mammalian or human brain tissue due to the confound introduced by the highly auto-fluorescent lipofuscins. 7.1. General There are several steps in the protocol that could be considered essential for the success of the experiment. Ø Our experience showed that the use of paraformaldehyde perfused tissue improved the immunolabelling results without affecting the in situ hybridization labelling. Ø To improve permeability of the oligoprobe and the antibodies, we also used Tween 20 in the initial washing step. Omission of the detergent will lower the intensity of labelling. Ø The use of free-floating sections will improve the permeability of the oligoprobe due to the increased exposure of the labelled oligoprobe to the section surfaces. Generally, when using a conventional in situ hybridization approach with a radioactive probe, this is not necessary since labelling detection can be enhanced by simply increasing the exposure time of the radiolabelled sections to the emulsion. In contrast, the intensity of the fluorescent label will depend on the number of fluorophores bound to the probe which is optimized when using free-floating sections. In our experience, using sections mounted on slides yielded poor fluorescent labelling. Ø Amplification steps are necessary to increase the fluorescent signal for reelin mRNA. Instead of simply using a primary antibody conjugated with fluorescein to recognize the biotin-labelled probe, we used two intermediate antibody steps for the purpose of amplifying the
158 C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160 Fig. 1. Colocalization of reelin mRNA and GAD67 in adult rat brain. Confocal microscope images of reelin mRNA detected by in situ hybridization visualized with fluorescein Žgreen., and GAD67 immunolabelling visualized with rhodamine Žred., in adult rat neocortex Žtop. and hippocampus Žbottom.. The overlay of reelin and GAD67 distribution shows that in the cortex, mRNA for reelin is found both in GABAergic and non-GABAergic cells, whereas in the hippocampus ŽCA2., all reelin-positive cells are GABAergic. Note also the different labelling pattern of GAD67 in the cytoplasm of GABAergic cells, and that of the putative GAD67 axon terminals or dendritic endings surrounding the cell bodies of pyramidal cells in the hippocampus.
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
number of fluorescent molecules that will recognize the reelin oligoprobe. When omitting the amplification steps, the fluorescence obtained is faint, at best. Ø In the in situ hybridization immunocytochemistry protocol described in this paper, in situ hybridization is performed first. The mRNA-oligoprobe hybrids which are formed in the first steps of the protocol are very stable and are not affected by RNAses. The protocol for the immunolabelling which is performed in the second step of the paper can, therefore, be carried out without any restrictions with respect to temperature Žsee below.. Ø The antisense oligoprobe used in this protocol corresponds to reelin base pairs 1264–1311. We have also obtained similar results using an antisense oligoprobe for reelin corresponding to base pairs 10225–10282. Whichever oligoprobe is chosen, one should include an experimental control condition in which sections are exposed to the corresponding sense oligoprobe. Ø GABA is synthesized in the adult brain by two forms of glutamic acid decarboxylase: GAD67 and GAD65 . However, this protocol proposes the use of an antibody against GAD67 ŽChemicon. to reveal GABAergic cells, since this antibody reveals the presence of GAD67 in cell bodies, as well as dendrites, whereas antibodies directed against GAD65 Ži.e., GAD-6 from Boehringer Mannheim. show that it is present mostly in nerve endings w7x. 7.2. AlternatiÕe and support protocols There are several protocols in the literature describing the use of a combination of in situ hybridization and immunohistochemistry w1,5,6,12x. Ø Some in situ hybridization immunohistochemistry protocols recommend performing the immunolabelling steps first, followed by the in situ hybridization to assure that the immunolabelling of antigens will not be compromised by some of the agents used in the in situ protocol w1,6x. However, the use of some variations with respect to conventional in situ hybridization protocols, will reduce this problem, allowing a nice double labelling w4,10x. When performing the immunolabelling first however, one must be careful to carry out all of the steps at 48C without shaking, to preserve the integrity of the mRNAs for the ensuing in situ hybridization steps. Ø Some of the protocols combining in situ hybridization and immunocytochemistry use a radioactive oligoprobe for the in situ, and a biotinylated secondary antibody to detect the immunocytochemical reaction with 3-3X-diaminobenzidine ŽDAB.. The main disadvantages with this approach include the time necessary to expose the radioactively-labelled sections to the emulsion, and the radioactive waste. Ø Alternatively, other protocols use a double colorimetric approach for the combination of immunocytochemistry and in situ hybridization, using digoxygenin-labelled oligo-
159
probes w1,5,6,12x. While this approach yields comparable results to our double fluorescent method, the use of fluorescence labelling allows one to analyze the samples at almost subcellular resolution with the help of a laser confocal microscope. Ø The use of fluorescein-conjugated antibodies can be substituted in this protocol by the use of antibodies conjugated to carbocyanin 2 ŽCy2., which has similar spectral properties but is considerably brighter and more stable. Similarly, the use of rhodamine-conjugated anti-rabbit can be substituted by a Texas Red-conjugated anti-rabbit, which has a higher emission spectra and can, therefore, be better distinguished from fluorescein when using a laser scanning confocal microscope. Ø Additional amplification steps can be included to increase the labelling of reelin, such as biotinyl–tyramide, as well as fluorescent antibodies against streptavidin. Ø Finally, this protocol describes an in situ hybridization using an oligoprobe labelled with biotin, rather than the more commonly used digoxygenin-labelled probe. One advantage of using an oligoprobe where biotin is chemically bound in specific positions within the sequence, is that it gives a much stronger signal, due in part to the very high affinity of streptavidin for biotin, and to the lesser steric hindrance caused by the biotin–streptavidin interaction. Another advantage is that it is sometimes difficult to assure successful end-labelling of the oligoprobe with digoxygenin. However, in some cases, the biotinylation of the probe may interfere with hybridization. Alternatively, one can end-label the reelin oligonucleotide probe with digoxygenin and change several of the subsequent labelling steps; however, the immunocytochemistry protocol remains unchanged. When using a digoxygenin-labelled oligoprobe, follow the same protocol described in this paper with the exception of the following appropriately numbered steps mentioned below. Žvi. Hybridize free-floating sections overnight at 428C in a hybridization buffer Žsee above. containing 50 pmolrml of a digoxygenin-labelled antisense oligoprobe corresponding to the bp 1264–1301 of the rat reelin sequence Ž5X-GCT CCC AAG ACA TCC CAT CCT CTG GGC TGG CTC TCA AAT TCT TCT G-3X .. To increase the sensitivity of this procedure, one can use a mixture of oligoprobes that are complimentary to the different regions of the target mRNA. Žix. Incubate overnight at 48C in a solution of 1% BSA in PBS containing a primary antibody against digoxygenin Žmouse anti-digoxygenin monoclonal antibody from Boehringer Mannheim, 0.4 mgrml., and a primary antibody raised to recognize GAD67 Žrabbit polyclonal antiGAD67 from Chemicon, 1:2000.. Žxii. Incubate sections for 1 h at RT in a solution of 1% BSA in PBS containing a secondary antibody that will recognize the one labelling digoxygenin Žanti-mouse antibody conjugated with digoxygenin from Boehringer Mannheim, 40 mgrml..
160
C. Pesold et al.r Brain Research Protocols 3 (1998) 155–160
Žxiv. For the final incubation step, incubate sections in the dark for 1 h at RT in a 1% BSA in PBS solution containing a fluorescein-labelled antibody against digoxygenin Žsheep anti-digoxygenin conjugated with fluorescein from Boehringer Mannheim, 0.4 mgrml., and a rhodamine-tagged antibody that will recognize the antibody labelling GAD67 Žgoat anti-rabbit conjugated with rhodamine from Chemicon, 1:100.. In this way, the reelin mRNA will be revealed by the presence of fluorescein, and the GAD67 protein will be revealed by the presence of rhodamine. 7.3. Troubleshooting This protocol should yield good results if followed closely. However, deviations from the protocol or slight oversights may lead to problems. The following is a list of potential factors that may contribute to such problems. Ø Little or no signal can occur for several reasons: RNA can be destroyed if the sections are not kept at 48C prior to hybridization; amplification steps may be insufficient; secondary antibody raised against species different from that in which primary antibody was raised. Ø Too much background: Insufficient washing steps; antibody concentration too high; RNAse contamination due to lack of sterility; lack of specificity resulting from too numerous amplification steps.
8. Quick procedure Ži–ii. Perfuse rats with 4% paraformaldehyde, remove brains and post-fix for 24 h. Žiii–iv. Cut 20 mm cryostat sections through 30% sucrose embedded brains, and rinse in PBS. Žv–vi. Permeabilize sections in 0.02% Tween 20 in DEPC–PBS Ž1 h., rinse in DEPC–PBS, and hybridize overnight in a cocktail containing the biotinylated oligoprobe. Žvii–x. Following several high stringency rinses in SSC and PBS, block sections with 1% BSA Ž30 min., and incubate overnight at 48C followed by 2 h at RT in mouse anti-biotin Ž2 mgrml. and rabbit anti-GAD67 Ž1:2000.. Žxi–xii. Following two 15-min washes in 1% BSA in PBS ŽRT., incubate sections for 1 h at RT with biotinylated anti-mouse IgG Ž1:250 in 1% BSA.. Žxiii–xiv. Following two 15-min washes in 1% BSA in PBS, incubate sections in the dark for 1 h with streptavidin–fluorescein Ž40 mgrml. and rhodamine-labelled goat anti-rabbit Ž1:100. in 1% BSA in PBS. Žxv–xvi. Wash sections thoroughly with 1% BSA, PBS, and water before mounting onto slides.
Žxvii–xviii. Once dry, cover with anti-fading mounting media, coverslip, and study with a Leica laser confocal microscope.
9. Essential literature reference Refs. w1,5,6,10,12x.
References w1x M. Brahic, M. Ozden, Simultaneous detection of cellular RNA and proteins, in: D.G. Wilkinson ŽEd.., In Situ Hybridization. A Practical Approach, Oxford Univ Press, New York, 1992, pp. 85–104. w2x G. D’Arcangelo, G.G. Miao, S.C. Chen, H.D. Soares, J.I. Morgan, T. Curran, A protein related to extracellular matrix proteins deleted in the mouse mutant reeler, Nature 374 Ž1995. 719–723. w3x G. D’Arcangelo, K. Nakajima, T. Miyata, M. Ogawa, K. Mikoshiba, T. Curran, Reelin is a glycoprotein recognized by the CR-50 monoclonal antibody, J. Neurosci. 17 Ž1997. 23–31. w4x J.A. Del Rio, B. Heimrich, V. Borrell, E. Forster, A. Drakew, S. Alcantara, K. Nakajima, T. Miyata, M. Ogawa, K. Mikoshiba, P. Derer, M. Frotscher, E. Soriano, A role for Cajal–Retzius cells and reelin in the development of hippocampal connections, Nature 385 Ž1997. 70–74. w5x P.C. Emson, B. Heppelmann, S.J. Augood, Development of techniques to combine isotopic and nonisotopic in situ hybridization and immunocytochemistry for phenotypic characterization of individual neurons, in: J.H. Eberwine, K.L. Valentino, J.D. Barchas ŽEds.., In Situ Hybridization in Neurobiology. Advances in Methodology, Oxford Univ Press, New York, 1994, 43–62. w6x M. Jaber, M. Cador, B. Dumartin, E. Normand, L. Stinus, B. Bloch, Acute and chronic amphetamine treatments differently regulate neuropeptide messenger RNA levels and FOS immunoreactivity in rat striatal neurons, Neuroscience 65 Ž1995. 1041–1050. w7x D.L. Kaufman, C.R. Houser, A.J. Tobin, Two forms of the g-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions, J. Neurochem. 56 Ž1991. 720–723. w8x T. Miyata, K. Nakajima, J. Aruga, S. Takahashi, K. Ikenaka, K. Mikoshiba, M. Ogawa, Distribution of a reeler gene-related antigen in the developing cerebellum: an immunohistochemical study with an allogenic antibody CR-50 on normal and reeler mice, J. Comp. Neurol. 372 Ž1996. 215–228. w9x M. Ogawa, T. Miyata, K. Nakajima, K. Yagyu, M. Seike, K. Ikenaka, H. Yamamoto, K. Mikoshiba, The reeler gene-associated antigen on Cajal–Retzius neurons is a crucial molecule for laminar organization of cortical neurons, Neuron 14 Ž1995. 899–912. w10x C. Pesold, F. Impagnatiello, M.G. Pisu, D.P. Uzunov, E. Costa, A. Guidotti, H.J. Caruncho, Reelin is preferentially expressed in neurons synthesizing g-aminobutyric acid ŽGABA. in cortex and hippocampus of adult rats, Proc. Natl. Acad. Sci. USA 95 Ž1998. 3221–3226. w11x S.N. Schiffmann, B. Bernier, A.M. Goffinet, Reelin mRNA expression during mouse brain development, Eur. J. Neurosci. 9 Ž1997. 1055–1071. w12x P. Wahle, Combining non-radioactive in situ hybridization with immunohistological and anatomical techniques, in: W. Wisden, B.J. Morris ŽEds.., In Situ Hybridization Protocols For The Brain, Academic Press, London, 1994, pp. 98–120.