Abstract
We applied in situ hybridization and the TUNEL technique to free-floating (vibratomed) sections of embryonic and postnatal mouse CNS. Full-length cDNAs specific for oligodendrocyte- or astrocyte-specific genes were labeled with digoxigenin using the random primer method. With paraformaldehyde-fixed sections, the nonradioactive in situ hybridization method provides detection of individual, very small glial progenitor cells in embryonic development. Small, isolated cells expressing oligodendrocyte specific messages can be detected in the neuroepithelium at embryonic and postnatal stages. The technique can be completed within 3 days and is as sensitive as the radioactive method. Likewise, the TUNEL method using DAB as the chromogen on free-floating sections provides excellent resolution. These DAB-stained sections can be embedded in plastic and thin-sectioned to visualize the ultrastructure of apoptotic cells. Both in situ hybridization and TUNEL methods can be applied to the same section, the tissue embedded in plastic, and semithin sections cut. The high resolution obtained with this combined procedure makes it possible to determine whether brain cells expressing glia-specific messages are undergoing apoptosis.
Keywords
A
We have found that the use of the nonradioactive in situ hybridization technique with free-floating (vibratomed) sections is extrememly sensitive and provides excellent cellular resolution of glia-specific mRNAs. These free-floating sections can be embedded in plastic and semithin-sectioned to provide even better resolution. Because of our interest in glial cell death during development (Skoff 1995), we have also combined the in situ hybridization with the terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) technique. The free-floating sections offer considerable flexibility because other techniques can be easily combined with the nonradioactive in situ hybridization.
Materials and Methods
Tissue Preparation
Timed-pregnant C57BL/6J mice were obtained from Charles River Labs (Wilmington, MA). The first day of gestation was defined as the day the vaginal plug was found. Postnatal mice were obtained from our breeding colony maintained by the Division of Laboratory Animal Resources, a federally approved laboratory facility. Embryonic (El4-20) and 0-74 days postnatal (DPN) mice were anesthetized with chloral hydrate and then perfused under pressure with fresh 4% paraformaldehyde (Electron Microscopy Sciences; Fort Washington, PA) in 0.1 M PBS, pH 7.4. Brains and cervical spinal cords were removed immediately after perfusion and placed in fresh 4% paraformaldehyde overnight at 4C. Fifty μm transverse sections were cut with a vibratome (Technical Products International; St Louis, MO) and kept in 4% paraformaldehyde in 24-well plates until used for in situ hybridization or TUNEL studies. Gloves were worn during sectioning for in situ hybridization, but after this step no additional sterile procedures were used.
In Situ Hybridization
Proteolipid protein (PLP) cDNA clone 68 (Sorg et al. 1987) and myelin basic protein (MBP) cDNA clone M72 (Newman et al. 1987), both kindly provided by A.T. Campagnoni, were removed from their vectors using an EcoR1 digest to produce a 2.4- and 2.2-KB fragment for PLP and MBP, respectively. The GFAP cDNA clone (Lewis et al. 1984) was recloned into pGEM32 and cut from its vector with a SalI and HindIII digest, yielding a 1.1-KB fragment. The digest was run out on a 1% agarose gel to separate the probe DNA from the vector and was cleaned with QIAquick Gel Extraction Kit (QIAGEN; Chatsworth, CA) and resuspended in 120 μl of dH2O. Five μl of this DNA was random primer-labeled with digoxigenin (DIG)-UTP (Boehringer Mannheim; Indianapolis, IN). The labeling was done according to the manufacturer's protocol except that the incubation was carried out overnight and precipitation was done with Naacetate overnight at −70C. Two μl of this labeled DNA was boiled for 10 min, snap-cooled, and then added to 100 μl of prehybridization solution to make the hybridization solution. Labeling of small amounts of probe (1-2 μg) provided the most intense signal, whereas labeling of larger amounts always produced a reduced signal.
Sections were washed in 0.1 M PBS, pH 7.4, three times for 10 min at room temperature (RT), treated with 0.3% Triton X-100 (ICN Biomedicals; Aurora, OH) in 0.1 M PBS for 15 min at RT, and washed in 0.1 M PBS. The sections were treated with proteinase K (GIBCO; Grand Island, NY) 0.3 μg/ml in 0.1 M Tris, pH 8.0, 0.05 M EDTA for 20 min at 37C. The sections were washed in 0.1 M glycine in 0.1 M PBS, pH 7.4, for 5 min at RT, washed in 0.1 M PBS for 1 min, and then incubated in 0.25% acetic anhydride (Sigma; St Louis, MO) in 0.1 M triethanolamine, pH 8.0, and 0.9% NaCl for 10 min at RT. After a wash in sterile dH2O for 10 min, the sections were incubated with prehybridization solution (50% formamide, 2 × SSC, 0.05 g/ml dextran sulfate, 1 × Denhardt's, and 0.1 mg/ml salmon testis DNA) for 1 hr at 52C. The prehybridization solution was then replaced with hybridization solution.
Sections were incubated overnight for 12-16 hr at 52C. Sections were washed twice for 30 min at 37C in 50% for-mamide, 2 × SSC, then washed two more times in 50% for-mamide, 1 × SSC for 30 min at 37C. The final washes consisted of 1 × SSC for 30 min at RT and 0.5 × SSC for 30 min at RT. The sections were then treated with Sl nuclease (GIBCO) 700 U/100 μl solution for 15 min at 37C. Next, they were washed in 0.1 M Tris twice for 30 min, pH 7.5, incubated in anti-DIG (Boehringer Mannheim) at 1:1000 in 1 × Denhardt's in Tris, pH 7.5, for 1 hr and 15 min, then washed four times for 15 min in Tris, pH 7.5, at RT. The sections were next washed in Tris, pH 9.4, NaCl:MgCl2 (20:1) for 2 min, and color-developed in Tris, pH 9.4, NaCl:MgCl2 (20:1) in the presence of nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP) (Boehringer Mannheim) for 30 min to 12 hr. Sections were rinsed several times and then mounted on glass slides with Aqua Poly/Mount (Polysciences; Warrington, PA).
Controls for in situ hybridization were the addition of hybridization solution without the DNA probe during the hybridization step. No staining was found in the control tissue. Our previous in situ hybridiztion studies (Maki et al. 1997; Granneman et al. 1998) with pBR322 DIG-labeled cDNA produced no signal using a similar procedure, and RNAse treatment also abolished the signal.
TUNEL Procedure
Tissue used for the TUNEL procedure was treated in the same manner as described for in situ hybridization, up to the first dH2O step. After this step the tissue was washed in 0.1 M PBS, pH 7.4, for 5 min, treated with 0.2% H2O2 for 20 min, rinsed in 0.1 M PBS for 5 min, and then incubated in TUNEL reaction mixture using the In Situ Cell Death Detection Kit, POD (Boehringer Mannheim) for 1 hr at 37C. The tissue was then rinsed in 0.1 M PBS three times for 5 min and incubated in Converter-peroxidase (POD) from the In Situ Cell Death Detection Kit (Boehringer Mannheim) for 30 min at 37C, rinsed in 0.1 M PBS three times for 5 min, and color-developed with SIGMA FAST, a diaminobenzidine (DAB) POD substrate (Sigma). If TUNEL-treated sections were subsequently used for in situ hybridization, they were rinsed in 0.1 M PBS twice for 5 min, washed in dH2O for 10 min, and then incubated in prehybridization solution and the remaining steps performed as stated above.
The Boehringer Mannheim TUNEL kit was compared to the Apop Tag TUNEL kit (ONCOR; Gaithersburg, MD). Tissue preparation was the same for both kits. For the Apop Tag kit, the tissue was placed in the Equilibration Buffer (ONCOR) for 5 min, incubated in working-strength terminal deoxynucleotidyl transferase enzyme (TdT) for 1 hr at 37C, incubated in working-strength stop/wash buffer for 30 min at 37C, rinsed in 0.1 M PBS three times for 5 min, and then developed with SIGMA FAST, DAB-POD substrate (Sigma). The tissue was then treated for in situ hybridization as described above. Both TUNEL kits yielded similiar results, a finding similar to a recent report comparing different TUNEL kits (Labat-Moleur et al. 1998). A comparison of the sensitivity of the TUNEL method to detect all pyknotic glial cells in normal and myelin-deficient mutants is discussed in Skoff (1995). Because the Apop Tag kit incorporates DIG-nucleotide residues into the DNA and is detected with anti-DIG conjugated to POD, the use of DIG/anti-DIG conjugated to alkaline phosphatase in the in situ procedure appeared to increase the reaction product of the TUNEL procedure. The In Situ Cell Death Detection Kit, POD incorporates flourescein-labeled residues into the DNA and is detected by an anti-flourescein antibody conjugated to POD. Therefore, the sequential treatment of the tissue to TUNEL and to in situ hybridization did not increase the labeling of the TUNEL-labeled cells.
Controls for the TUNEL procedure were treated in the same manner as the test samples except that the TdT enzyme was omitted from the reaction mixtures in both kits and was replaced with dH2O. No labeling was found in the controls.
TUNEL-processed or combined TUNEL-in situ sections were further processed for plastic embedding in Araldite 502 (Electron Microscopy Sciences) plastic. Sections were rinsed several times in 0.1 M Tris, pH 7.5, dehydrated in a graded series of ETOH followed by three changes in 100% propylene oxide over 30 mins, and stored overnight in 50% propylene oxide/50% Araldite 502 in a dessicator at RT. The next day the Araldite was replaced with fresh 100% plastic, kept dessicated at RT for 3-4 hr, and then embedded in Araldite. Two-μm sections were cut on an AO/Reichert Ultracut E (Reichert-Jung; Vienna, Austria) microtome. Silver-gold ultrathin sections were cut from blocks of TUNEL-processed tissue and photographed on a JEOL 1010 transmission electron microscope. All light micrographs were taken on a Leitz microscope using Kodak 160T tungsten film or TMAX100.
Results
Perfusion with 4% paraformaldehyde and routine fixation methods provided excellent preservation of tissue for both the in situ hybridization and TUNEL studies. Low-magnification pictures of brain and spinal cord showed intense labeling for myelin protein genes in the mouse CNS using 50-μm free-floating sections (Figures 1A and 1B). The distribution of PLP-and MBP-labeled cells in the brain and spinal cord essentially overlapped from the earliest time point studied (El4) to the latest time point (P74). Embryonic sections processed for myelin gene expression were generally developed twice as long as postnatal sections to expose the labeled cells. This suggests that the amount of message was several-fold less in younger than in older animals. Expression of GFAP mRNA was also examined to ascertain whether the free-floating section method was amenable to study the time of expression of other CNS genes. Intense signal for GFAP message was found in P0 mice, but the distribution of GFAP-labeled cells (Figure 1C) was different from that of PLP- and MBP-labeled cells. At this age, GFAP mRNA was expressed in many cells near the midline. These GFAP+ cells were frequently present in the neuroepithelium bordering white matter tracts and near the gray matter adjacent to the pia-glial limitans. These astrocytes had more oval or rectangular perikarya than the oligodendrocytes, which were usually spherical.
Expression of PLP, MBP, and GFAP mRNAs in brain using the nonradioactive in situ hybridization technique. (A) PLP in situ hybridization of a transversely sectioned P7 mouse brain at the level of the interventricular foramen. Intensely stained cells are predominantly present in white matter tracts, including the fornix (F) and corpus callosum (CC). The midline of the brain is to the right, with the hippocampus (H) medial to the lateral ventricle and the thalamus (T) to the left of the ventricle. (B) MBP in situ hybridization of an E18 cervical spinal cord. MBP+ cells are located along ventral and lateral edges of the spinal cord. The small dorsal funiculi also contain many MBP+ cells (arrows). At this age, occasional labeled cells are present in the ventral gray matter and around the central canal but not within the neuroepithelium. (C) Low magnification of a newborn brain probed for GFAP message. Many cells near the midline are intensely labeled, but few cells within the cortical gray matter are labeled at this time point. Cells at the base of the interhemispheric fissure (asterisk) are intensely labeled. In the corpus callosum (CC), GFAP+ cells predominate near the midline. Strong labeling is present in cells at the dorsal border of the lateral ventricles (V). These GFAP+ message-containing cells are frequently located within the neuroepithelium. Bars = 100 μm. (D) In situ hybridization of a 43-day-old brain, showing the dorsolateral zone of the lateral ventricle. The neuroepithelial layer is thickened at the dorsolateral border. An MBP+ cell (arrow) with two short processes is located within the neuroepithelium. (E) MBP expression at the level of the 3rd ventricle of an E18 mouse. MBP+ cells are located immediately adjacent to the neuroepithelium. The cytoplasm of these cells is intensely stained, but the nuclei exhibit minimal staining. Some of these small cells have short processes. (F) PLP in situ hybridization of an E17 brain sectioned just rostral to the optic chiasm. An isolated PLP+ cell (arrow) is located just outside the thickened neuroepithelium. The location of this cell is in the lateral ventricle (V) near the midline. (G) High magnification of an MBP+ mRNA cell from the spinal cord of a 74-day-old mouse. Focusing through the cell with the microscope shows that the nucleus (asterisk) has no staining but that the surrounding cytoplasm is intensely stained. Several processes arise from the perikaryon, two of which can be seen in this micrograph. MBP message extends along these long processes and is often aggregated into clumps (arrowheads) along the processes. (H) Combined MBP in situ hybridization and TUNEL staining of a 7-day-old mouse. This 2-μm plastic section is from the corpus callosum. MBP message (arrow) is blue due to NBT/BCIP staining, and TUNEL staining (arrowheads) is brown due to peroxidase reaction product. The TUNEL+ cell at the upper right is surrounded by a blue halo, suggesting that this dying cell is an oligodendrocyte progenitor. Message is located in cytoplasm and nucleus is unstained. Many TUNEL+ cells are located in the brain at this age and are in various stages of degeneration. Most nuclei (asterisks) are stained a faint brown, with no blue reaction product in their cytoplasm and no intense brown reaction in their nuclei. Bars = 10 μm.
Examination of 50-μm sections at higher magnification showed that the cytoplasm was intensely stained for both astrocyte- and oligodendrocyte-specific mRNAs, but the nucleus was unstained (Figure 1G). The distribution of MBP and PLP messages in oligodendrocytes was different, with MBP extending into the processes, whereas PLP message was confined primarily to the perikaryon. High-magnification photographs of MBP in situ hybridization slides showed that this message extended into the distal processes where they expanded to form the myelin sheath (Figure 1G). Along the oligodendrocyte processes, MBP message tended to be aggregated into distinct clusters. In contrast, PLP very rarely extended into the finer processes.
The low background and the well-preserved tissue permitted identification of message-expressing cells in precise relationship to other neural structures. For example, individual MBP+ and PLP+ cells were occasionally found scattered in the neuroepithelium of young adult animals, especially at the dorsolateral edge of the lateral ventricles (Figure 1D). Single, isolated PLP+ or MBP+ cells were found throughout the cerebrum in embryonic development (not shown). At E14-E18, some of these oligodendrocyte progenitor cells were located within the neuroepithelium (Figure 1F) and some were found in presumptive white matter tracts, but most were located just outside the neuroepithelium. Although these isolated cells had very little cytoplasm, many of them exhibited a ring of dense staining around the nucleus.
Another major advantage of the perfused, free-floating sections is the ability to resolve individual cells and to quantify them. For example, a cluster of MBP+ cells could be identified at the base of the third ventricle at E17 (Figure 1E) and, by focusing through the section, it was possible to quantify the number of stained cells. The boundaries of the neuroepithelium were visible in this section, and in this photograph the MBP+ cells were not within the neuroepithelium proper but were found immediately adjacent to it. These cells had already begun to develop short processes that contain small amounts of message.
TUNEL staining using free-floating sections also provides excellent cellular resolution of apoptotic cells. An advantage of using free-floating sections is that the TUNEL-stained sections can be combined with other techniques, including in situ hybridization, or they can be embedded in plastic for ultrastructural analyses. TUNEL staining of mouse cerebrum during the first postnatal week showed massive numbers of apoptotic cells in certain regions of the cerebrum, whereas other areas of the cerebrum had virtually no TUNEL+ cells (Figure 2A). Higher-magnification photomicrographs (Figure 2B) of apoptotic regions showed that most TUNEL+ cells had a signet ring morphology characteristic of apoptotic cells found in other tissues (e.g., Labat-Moleur et al. 1998). After the first postnatal week, the number of TUNEL+ cells declined dramatically, and virtually all apoptotic cells were restricted to the subventricular region in the older rodents (Skoff 1995). Transmission electron microscopy of paraformaldeyhyde-fixed tissue from these cerebral areas confirmed the presence of scores of apoptotic cells (Figure 3A). Many of these apoptotic cells were engulfed within cells having the morphology of microglial cells. Electron microscopic examination of the TUNEL-processed tissue showed that the DAB reaction product can be visualized in ultrathin sections. As predicted from the light microscopic tissue, fragmented DNA was found in a variety of forms. The large, dense DAB+ aggregates were presumably the cells' nuclei (Figure 3B). Small bits and pieces of DAB+ material were often found within the cytoplasm of phagocytic cells and indicate nuclear DNA breakdown.
(A) Light micrograph of 1-day postnatal cerebrum rostral to the hippocampus, illustrating the results of the TUNEL method with free-floating sections. TUNEL+ cells in this low-magnification picture appear as dark spots in certain structures of the cerebrum, whereas other regions have virtually none. Apoptotic cells are abundant in the cingulum near the interhemispheric sulcus (∗) and in the basal ganglia (BG). The cerebral cortex (CC) has few apoptotic cells. The edge of the ventricle (V) is at the bottom of the picture; dorsal is at the top and ventral at the bottom of the picture. Bar = 100μm. (B) Higher magnification of corpus callosum, which has an abundance of TUNEL+ cells. Apoptotic cells exhibit a ring morphology, with densely stained outer cores that surround more lightly stained inner cores (arrows). Many apoptotic cells are out of the plane of focus in the 50-μm section and appear as dense spheres. Bar = 10 μm.
(A) Uranyl acetate/lead citrate conventional electron micrograph of 1-day postnatal cerebrum around lateral ventricle where apoptotic cells are abundant (see Figure 2). Two cells (asterisks) at different stages of apoptosis are engulfed by a phagocytic cell. The upper of the two degenerating cells exhibits the classical features of an apoptotic cell, with marginated chromatin and intact cytoplasm. The cytoplasm of the phagocytic cell has stringy endoplasmic reticulum (ER) and inclusions (arrows) typical of microglial cells. The nucleus (N) of this microglial cell has been grazed so that the nuclear membrane is obliquely sectioned. (B) Electron micrograph of TUNEL-processed tissue from a 1-day postnatal brain. Electron-dense DAB reaction product (asterisks) indicates location of fragmented DNA. The remants of the large TUNEL+ cell has a dense central mass surrounded by smaller aggregates of fragmented nuclear debris. The appearance of this degenerating cell is similar to that of the cell illustrated in A. Bars = 1 μm.
The free-floating sections can be processed first for TUNEL staining and then for in situ hybridization. After completion of the TUNEL procedure, the sections were visualized under an inverted microscope to confirm that the TUNEL staining was strongly positive. These sections were embedded in plastic and re-sectioned at 1-2 μm to visualize the location of reaction product. mRNA was visualized as a blue reaction product within the cytoplasm of cells, whereas the TUNEL staining appeared as a brown reaction product (Figure 1H). TUNEL staining was rarely localized in the large MBP+ and PLP+ cells, suggesting that the vast majority of apoptotic cells are not mature oligodendrocytes expressing large amounts of myelin messages. However, small TUNEL+ cells sometimes exhibited a blue reaction product (either MBP or PLP mRNA) around their nuclei (Figure 1H), strongly suggesting that apoptosis in the oligodendrocyte lineage was primarily restricted to immature oligodendrocytes.
Discussion
The nonradioactive in situ hybridization protocol used for tissue sections is essentially similar to the nonradioactive in situ hybridizaton procedure used on our glial cell cultures (Maki et al. 1997; Granneman et al. 1998). We found that labeling small amounts of probe gave the most intense staining of cells, and our attempts to label larger amounts decreased the strength of the signal (see Materials and Methods). By labeling small aliquots of probe (1-2 μg), we were able to detect myelin-specific messages within the neuroepithelium as early as was described in other studies using radioactive probes (see below). Although S1 nuclease treatment of our glial cell cultures decreased background and improved the appearance of the in situ hybridizations, this step was absolutely essential for tissue sections. Otherwise, the background was very high and the material essentially uninterpretable. Triton X-100 (0.3% at RT for 15 min) and proteinase K pretreatment (0.3 mg/ml at 37C for 20 min) were also required. Increasing their concentrations and/or the times by roughly twofold had no noticeable effects on the strength of the signal and began to tear up the sections. The concentration of proteinase K was considerably less than that usually used with paraffin sections (e.g., Guiot and Rahier 1995; Oliver et al. 1997). The difference is probably due to using fresh vs paraffin-embedded sections. However, the concentrations of Triton X-100 and proteinase K were adequate for complete penetration of the DIG-labeled probes into the 50-μm sections: the signal within cells in the middle of the section was as intense as in the labeled cells at both surfaces. This finding was in contrast to our and other immunocytochemical studies in which only the top and bottom 10-15 μm of a 50-μm or thicker free-floating brain section were stained, presumably due to penetration of immunoglobulin complexes.
Nonradioactive in situ hybridization applied to free-floating sections offers several advantages compared to the radioactive method. With radioactive in situ hybridization, message-bearing cells that are small and scattered throughout a tissue are likely to be mistaken for background or difficult to identify as positively labeled. Accordingly, the distribution of message-bearing cells within a particular tissue may be incomplete. This is particularly the case with neuroglial progenitor cells which are very small and often are not clustered throughout the embryonic brain. Isolated message-bearing cells in the neuroepithelium of adult animals are also likely to be overlooked.
In the embryonic brain, clusters of cells expressing putative oligodendrocyte-specific messages have been described along the ventral neuraxis in embryonic brain using radioactive in situ hybridization (Timsit et al. 1992,1995; Pringle and Richardson 1993; Yu et al. 1994; Peyron et al. 1997) or in transgenic animals expressing different myelin protein-lac Z fusion trans-genes (Foran and Peterson 1992; Wight et al. 1993; Spassky et al. 1998). Although these methodologies have revealed the general location of platelet-derived growth factor-α receptor (PDGFαR)-, 2′,3′-cyclic-nucleotide 3′ phosphodiesterase (CNPase)-, MBP-and PLP/DM20-expressing cells, it is virtually impossible to determine numbers of message-bearing cells in a particular location and, most importantly, whether such cells are present in the neuroepithelium where oligodendrocyte progenitors must arise (e.g., Noll and Miller 1993; Pringle and Richardson 1993; Yu et al. 1994; Timsit et al. 1995; Hall et al. 1996). To definitively determine the origin of oligodendrocytes, it is essential to know whether cells expressing myelin messages are located inside or outside the neuroepithelium. Cells expressing myelin messages located outside the neuroepithelium may, in fact, have been born in another region of the cerebrum and simply migrated to another cerebral region using the boundaries of the neuroepithelium as a pathway. In agreement with previous studies, this report using nonradioactive in situ hybridization confirms clusters of oligodendrocyte progenitors along the ventral midline axis in embryonic development. However, we also find isolated message-bearing cells located adjacent to the lateral ventricles, 3rd and 4th ventricles, and even throughout the cerebral parenchyma beginning at El6. We have begun to study El4 embryos and find intensely stained MBP+ and PLP+ cells within the neuroepithelium of the 4th ventricle and in the spinal cord, strongly suggesting that oligodendrocyte progenitors can be identified much earlier with this technique. The location of these cells within the neuroepithelium, their numbers, and their presumed migration pattern at different embryonic ages will be documented in a detailed report. Our finding of putative oligodendrocyte progenitors in the embryonic brain outside of the ventral neuraxis is in agreement with a transplantation experiment showing that oligodendrocyte progenitors must be spread widely throughout the embryonic brain (Hardy and Friedrich 1996) and that mainly those in the ventral portion of the brain differentiate into oligodendrocytes.
Comparison of PLP/DM20 gene expression of embryonic spinal cord at E14-E18 using the radioactive and nonradioactive methods shows a similar pattern of labeled cells (Yu et al. 1994; Richardson et al. 1996), but the nonradioactive method detects more labeled cells in and around the neuroepithelium. Comparison of MBP gene expression of mid-embryonic brain with the two techniques is not possible, as in situ hybridization with radioactively labeled MBP probes has not been performed to the best of our knowledge. With the nonradioactive method, we are able to detect MBP+ and PLP/DM20+ stained cells within the E14 spinal cord neuroepithelium, but detection of these myelin message-bearing cells located within the neuroepithelium has not yet been demonstrated in rodents with the radioactive method. Certainly, the nonradioactive method provides much better resolution than the radioactive method and is as sensitive as the radioactive method on the basis of published reports.
In this study, 1-2-KB CDNAS were labeled with digoxigenin using the random primer method, whereas riboprobes and partial cDNA clones were radiolabeled and used in most other studies examining glial gene expression (e.g., Kristensson et al. 1986). The combination of several different factors, including the labeling method and the use of perfused material, all probably contribute to making the nonradioactive in situ hybridization as sensitive as the radioactive method for detecting glia-specific messages. The resolution obtained with the nonradioactive method clearly reveals the differential distribution of MBP and PLP messages within oligodendrocytes in the animal. It is well known from biochemical studies that MBP message co-localizes with myelin preparations, and that in tissue culture MBP message extends into the fine branching processes of oligodendrocytes. This study confirms that MBP message extends into the distalmost processes of oligodendrocytes in animals. The mechanisms regulating transport of MBP message into oligodendrocyte processes is reviewed by Brophy et al. (1993). The ability to easily discriminate differences in MBP and PLP mRNA cellular localization with this technique can be exploited to study mutant animals and experimentally altered animals in which transcription and transport of myelin messages may be altered. The rapidity of the entire procedure and its capacity to be combined with other techniques makes this a versatile technique that should be applicable to other tissues in addition to the nervous system.
Footnotes
Acknowledgements
Supported by the National Multiple Sclerosis Society.
We thank Ms C. Perry for performing the light microscopic TUNEL staining.