DEVELOPMENTAL GENETICS 11:318-325 (1990)
In Situ Hybridization Studies on Murine Catalase mRNA Expression During Embryonic Development D.L. REIMER AND S.M. SINGH Genetics Laboratories, Department of Zoology (D.L.R., S.M.S.) and Division of Medical Genetics (S.M.S.), University of Western Ontario, London, Ontario, Canada In situ hybridization using nuABSTRACT cleic acid probes was used to detect cell- and tissue-specific transcriptjs) of embryonic genes during development and differentiation. This highly sensitive technique has the potential to provide valuable information on the regulation of lowabundance housekeeping genes during development. We have determined the experimental conditions required to detect the catalase message in adult mouse liver. Catalase effects the breakdown of H , 0 2 to O2 and H,O and offers protection against the toxic effects of oxygen radicals. We used a cloned 550 bp BamHI-Pstl fragment from a mouse catalase cDNA (pMCT-1) to generate 35Slabeled sense and antisense riboprobes. The experimental conditions used were sensitive enough to quantitate the abundance of silver grains generated by the antisense riboprobe on the adult liver, a tissue known to be positive for this message. The hybridization protocol was applied to serial sections of 13- and 18-day-old mouse embryos. The results suggest that the catalase expression in the liver and brain begins with somite formation and increases with development and differentiation. O n the other hand, this message appears to be absent in mesenchyme, particularly in day 13 embryos. The message in positive tissues appears evenly distributed throughout the cell. The observed expression of the catalase message in the adult liver is approximately six times that in the embryonic liver. It is compatible with the enzyme activity results and emphasizes the sensitivity of the in situ hybridization method (over northern blot, etc.) used in this study.
Key words: Mice, housekeeping genes, liver, tissue specificity
hybridization analysis on cells or tissue sections with labeled DNA or RNA probe(s). The latter is particularly suitable for identifying specific cell types and individual cell components that harbor the gene-specific message. In situ hybridization methodology has been extensively used to evaluate the spatial and temporal patterns of gene expression during embryonic development in a variety of species including pre- and postimplantation mouse embryos [Anderson and Axel, 1985; Utset et al., 1987; Dean et al., 1989; Nisson et al., 19891. The technical improvements, including 35S-labeled cRNA probes, now make this technique sensitive enough to be able to detect two to ten target molecules per cell and permit reliable quantitation of any message [Davis et al., 1986; Simmons et al., 19891. The use of the sense transcript provides a suitable negative control imperative to such studies and the cRNA-mRNA hybrids are more stable than cDNA-mRNA hybrids and can therefore be subjected to higher stringency washes needed to eliminate nonspecific background binding. As a result, this protocol has the potential to be useful in studies on developmental regulation of low-abundance mRNAs in cells and tissues. Catalase (EC 1.11.1.6; hydrogen-peroxide: hydrogenperoxide oxidoreductase) is a housekeeping enzyme that provides protection against toxic effects of oxygen radicals generated through normal cellular metabolism. It is located primarily in the peroxisomes of eukaryotic cells and catalyzes the breakdown of H,O, to O2 and H,O [Aebi, 19841. In mice, it is coded for by a single structural gene [Dickerman et al., 1968; Holmes and Duley, 19751 and has been shown to be developmentally regulated [El-Hage and Singh, 1989, 19901. The catalase transcript is detectable as early a s 8 days postconceptus, with the beginning of somite formation. Although these studies, which rely on northern and slot blot hybridization, demonstrated that the expression of this gene is significantly higher in the develop-
INTRODUCTION Studies on gene expression are dependent on the availability of appropriate biological material and sensitive molecular technology. In general, the analysis of gene expression a t the level of mRNA (transcription/ pretranslation) relies on northern, slot blot, or in situ
Received for publication J u n e 4, 1990; accepted August 13, 1990. Address reprint requests to S. M. Singh, Genetics Laboratories, Department of Zoology, University of Western Ontario, London, Ontario, Canada N6A 5B7.
IN SITU HYBRIDIZATION STUDIES IN EMBRYOS 5'
A
P S
B
319
3' % Bg Xb , \
I
*
I I I I I
',
2.0 k b
*
\
\
, \ \
B
--II-' SP6
Blg
P
. -1
0.5 k b
T7
Fig. 1. Restriction map of the 2.0 kb mouse catalase cDNA (pMCT-1)and the strategy used to subclone the 0.5 kb BamHI-PstI fragment into the pGem-2 riboprobe vector. B, BamHI; P, PstI; Bg, BglII; S,SphI; X, XhoI; N, NsiI; A, ApaI; Xb, XbaI. The T7 and SP6 promotors were used to generate the antisense and sense riboprobes, respectively
ing liver than in the carcass, they do not identify specific tissues or cell types that are involved in the synthesis of the catalase message. Here we have investigated the spatial and temporal pattern of catalase gene expression in mouse embryos using in situ hybridization. We have applied 35S-labeled single-stranded riboprobes on serial paraffin sections of mouse embryos under experimental conditions determined during this investigation. The results were quantified, and the temporal and spatial pattern of expression of this gene during development was evaluated.
MATERIALS AND METHODS Tissue Collection, Fixation, and Sectioning Female mice, 11 weeks of age [Swiss Webster (SW)] were superovulated by intraperitoneal (i.p.) injection of 5 U pregnant mare's serum gonadotropin (PMSG; Sigma), followed by 5 U human chorionic gonadotropin (HCG; Sigma) 48 h r later. Each female was placed with a mature SW male overnight. The presence of a copulatory plug the following morning indicated successful mating (day 1of gestation). Pregnant females were sacrificed in order to collect fetuses representing days 13 and 18 postconception. Fetuses were dissected out in Sorenson's buffer (pH 7.0) and immersion fixed in buffered 4% paraformaldehyde overnight at 4°C. Adult liver was removed and cut into 2 cm cubes and immersion fixed a s well. The tissues were subsequently processed the following day on a 16 hr cycle on a Histomatic Tissue Processor (Fisher model 1661, and embedded in paraffin. Adult liver and sagittal sections of fetuses were cut a t 6 p.m on a microtome and collected on gelatinized and poly-L-lysine (Sigma)-coated microscope slides. In Situ Hybridization In situ hybridization was performed using modifications from Simmons et al. f19891. Deparaffinized slides
were pretreated by sequential immersion in 0.2 N HCl for 20 min, 2 x SSC (1x SSC = 0.15 M NaC1,0.015 M sodium citrate, pH 7.4) for 30 min, proteinase K (Boehringer Mannheim Canada), 25 pgiml for 15 min a t 37"C, acetic anhydride in 0.1 M triethanolamine HCl buffer (pH 8.0) for 10 min, rehydrated in graded alcohols, and air dried. A cDNA clone containing the 2.0 kb 3' end of the mouse catalase gene (pMCT-1) was generously donated by Dr. Robert Korneluk (Ottawa, Ontario, Canada). This cDNA clone was excised with BamHI and PstI (to remove poly-A' tail), and the resulting 550 bp insert was subcloned into the pGem-2 vector (Promega). The restriction map of the clone and the cloning strategy used in this study is shown in Figure 1. Radiolabeled RNA probes were synthesized in the presence of 35SUTP (Amersham) using a n RNA transcription kit (Promega). In this construct, T7 promotor-driven transcription yields antisense RNA that is complimentary to catalase mRNA. On the other hand, the SP6 promotor-driven transcription will yield sense mRNA to be used a s a negative control. The labeled probes were reduced by alkaline hydrolysis [Cox et al., 19841 and had specific activities of 1.9-2.9 x lo9 cpmibg. The hybridization mixture contained 0.2 pgiml '35Slabeled RNA transcripts, 50 mM dithiothreitol (DTT), 0.3 M NaCl, 50% (viv) deionized formamide, 10% (wiv) dextran sulfate, 0.2 mgiml sheared salmon sperm DNA, 0.125 mgiml tRNA, 0.02% (wiv) Ficoll, 0.02% (wiv) polyvinylpyrrolidone, 10 mM Tris (pH 7.4), 1 mM EDTA, and 0.1% Triton X-100. The mixture was applied to pretreated slides, coverslipped, and incubated a t 45°C for 4 hr. Following hybridization, the slides were washed twice in 4 x SSC with 5 mM DTT for 10 min, once in 50 pgiml RNAse A (Boehringer Mannheim Canada) in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 30 min a t 37"C, twice in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 15 min a t 37"C, twice in 0 . 1 ~ SSC
320
REIMER AND SINGH
with 5 mM DTT and 0.1% SDS for 10 min a t 50°C, 2 x SSC for 10 min, dehydrated in graded alcohols (each containing 0.3 M ammonium acetate), and air dried. Slides were dipped in NTB-2 nuclear track emulsion (Eastman Kodak Company, Rochester, NY) diluted (1: 1) with 0.6 M ammonium acetate and exposed for 2 days a t 4°C. The slides were subsequently developed in D19 (Eastman Kodak Company) for 5 min a t 15"C, fixed with 30% sodium thiosulfate for 5 min, and counterstained with Ehrlich's hematoxylin. Slides were prepared from three embryos at two developmental stages each. They were evaluated for a general pattern of tissue morphology and silver grains on different cells and tissue types. Although there was a tissue-specific distribution for silver grains a t both development stages, conclusions about the expression of the catalase gene could only be made following quantitative evaluation of such observations. Data was therefore collected on a representative slide for the adult liver and day 13 and day 18 embryos by quantitating the cell density and number of silver grains.
Data Collection and Statistical Analysis Each slide was evaluated using bright-field microscopy for characterization of cell types, cellular structures, and number of cells in a defined area. Dark-field microscopy was used to count silver grains per defined area on the same slides. Different numbers of silver grains were observed on different areas of each slide following hybridization with the labeled sense and the antisense riboprobes. Grains were counted in defined areas with and without (blank) tissue on the same slide. It was apparent that the number of silver grains was significantly higher on the tissue-containing areas particularly with hybridization using the antisense probe. Silver grain counts on tissues were corrected for grains on a representative area of the same slide without tissue (blank). Five random areas were counted for the number of silver grains on day 13 and day 18 embryonic tissues, whereas ten such numbers were generated for the adult liver. For each count of silver grains for a n area, the number of cells in the same area was recorded using bright-field microscopy. It is therefore possible to express these observations as grains per cell for all experiments. The difference between grain counts generated by antisense and sense 35S-labeled riboprobes was used to assess the expression of the catalase message in a tissue at a given stage of development. Significant differences in all pairwise comparisons were evaluated using Student's t test.
RESULTS In situ hybridization has been used to analyze genespecific expression during embryonic development, for high [Cox et aZ., 19841 and low [Davis et al., 1986; Simmons et al., 19891 abundance mRNAs. We used this technology to study the tissue-specific expression of a
low abundance mRNA of a housekeeping gene (catalase) during embryonic development in mice. To address the technical success of this protocol, paraffin sections of adult mouse liver (which is known to express catalase) were hybridized with 35S-labeled sense (negative control) and antisense mouse catalase riboprobes. A hematoxylin and eosin-stained section of the adult mouse liver (Fig. 2A) was used to evaluate the morphology and quantitate cell numbers per area. It shows epithelial cells, numerous sinusoids, and fully differentiated and relatively large hepatocytes. Molecular hybridization of this tissue section with the sense probe revealed few silver grains under dark-field (Fig. 2B). Here, the grain counts on the tissue differ minimally from the number of silver grains off the tissue. When a serial section of this tissue was hybridized with the antisense riboprobe (Fig. 2C), a n abundance of silver grains is observed. Furthermore, a bright-field evaluation of this section following counterstaining with hematoxylin (to enable identification of the nuclei of individual cells) shows cell density and a n apparent distribution of silver grains on the nucleus a s well as on the cytoplasm (Fig. 2D). These results suggest that hybridizable catalase mRNA is present in the adult liver and the transcript for this housekeeping gene in this tissue is detectable throughout the cell. We have applied the same experimental conditions to study the spatial and temporal expression of the catalase gene during development in mice. Paraffin sections from 13 day and 18 day embryos were subjected to this in situ hybridization and evaluation protocol. In day 13 and day 18 embryos, the silver grains were particularly localized to the liver, followed by brain cells or their progenitors. The grain counts on these two tissues were quantified and evaluated a t these developmental stages and are presented below. Figure 3 shows some representative tissue differences seen with the application of antisense riboprobe in developing embryos. For example, higher silver grain counts on a region of nucleated red blood cells in the tail bud of day 13 embryo were observed compared with the surrounding mesenchyme cells [Fig. 3A (dark-field) and B (brightfield)]. One may therefore conclude that the undifferentiated tissue (e.g., mesenchyme) has lower catalase mRNA level than nucleated red blood cells a t this stage of development. A similar tissue difference in grain counts on day 18 embryos is seen in Figure 3C (darkfield) and D (bright-field). The number of silver grains is elevated on the liver compared with the surrounding tissue types, including gut. An appropriate comparison of the same tissue a t different developmental stages, however, is needed to evaluate accurately the developmental expression of the catalase gene in mice. Molecular hybridizations of sense and antisense riboprobes to day 13 and day 18 embryo sections were analyzed in detail for developmental expression of catalase in the liver, brain (Fig. 41, and mesenchyme. The liver and brain are easily identified and reasonably
IN SITU HYBRIDIZATION STUDIES IN EMBRYOS
321
Fig. 2. In situ hybridization analysis of catalase expression in adult mouse liver. A Bright-field, stained with hematoxylin and eosin shows numerous sinusoids (thick arrows) and large hepatocytes (thin arrow). Serial sections (dark-field)hybridized with the sense probe (B) and the antisense probe (C). D Bright-field of C to show the distribution of the silver grains on the cells. Bar = 100 pm.
Fig. 3. Representative tissue differences seen with antisense riboprobe on mouse embryos. Dark-field (A) and bright-field (B) photographs of the tail region of a day 13 embryo showing specific hybridization to nucleated red blood cells (arrow);m, mesenchyme; dark-field (C) and bright-field (D) representation of a day 18 embryo showing elevated numbers of silver grains on liver (1) compared with gut (g).Bar = 200 km.
differentiated at the two developmental stages studied and show relatively high silver grain counts using the antisense probe compared with the sense control. In day 13 embryonic liver (Fig. 4A, bright-field), there are
numerous sinusoids, and the tissue is hemopoetic, highly dividing, and rapidly differentiating. The serial section hybridization of this tissue by the sense and antisense riboprobe showed that relative grain counts
322
REIMER AND SINGH
LIVER Fig. 4. Developmental expression of catalase in liver and brain of mouse embryos. Bright-field analysis of day 13 liver (A) shows sinusoids (arrow)and differentiating cells; silver grains following hybridization with the antisense riboprobe (B) dark-field; day 18 liver (C) bright-field, relatively differentiated tissue with sinusoids (arrow);D
Dark-field hybridization with the antisense riboprobe. Brain tissue of day 13 embryos (E, bright-field) and with silver grains following antisense riboprobe hybridization (F, dark-field); day 18 brain bright field ( G )and dark field (H). Bar = 100 pm.
were significantly higher with the antisense probe (Fig. 4B) compared with the sense probe (not shown). In comparison, the 18 day embryonic liver is more differentiated, with higher number of cells in a given area but still hemopoetic (Fig. 4C, bright-field). A number of sinusoids are also visible a t this stage in this tissue. The in situ hybridization analysis shows that, a s in day 13 embryos, the grain density is higher with the antisense probe (Fig. 4D) than with the sense probe. It may be pointed out that the grain count on the adult liver cells (Fig. 2C) is remarkably elevated compared with the embryonic stages. Furthermore, the density of silver grains on the 18 day embryonic liver (Fig. 4D) is higher than that of the day 13 embryos (Fig. 4B). The brain of the 13 day embryos possesses actively
dividing progenitors in their initial stages of differentiation with clear gaps and migrating cells (Fig. 4E, bright-field). Subsequent observations of in situ hybridizations on such embryos shows that the gene-specific riboprobe is able to detect expression of the catalase gene. Here, the silver grains generated by the antisense probe (Fig. 4F) are evident a t a lower level than those in the liver. The brain of the day 18 embryos is highly differentiated [Fig. 4G (bright-field), a s a n example], and individual cell types can be recognized. Here, the cells are densely packed, having reached their level of differentiation and organization with neural development nearing completion. The in situ hybridization on serial sections yielded a higher number of cell-specific silver grains on the brain cells a t this
IN SITU HYBRIDIZATION STUDIES IN EMBRYOS
323
TABLE 1. Silver Grain Counts on Different Tissues Following In Situ Hybridization With Antisense and Sense 35S-Labeled Riboprobes* Dev. stage 13 Day
Tissue Liver
18 Day
Brain Mesenchyme Liver Brain Mesenchyme
Adult
Liver
Grains per 3,600 pm2 Sense Antisense 50.40 i 3.97 22.80 i 0.86 30.20 2 1.20 93.60 2 4.07 26.60 2 1.25 35.80 2 2.65 77.80 2 1.76
96.60 i 4.53 35.60 i 2.71 29.80 2 2.74 203.80 2 5.50 84.80 i 4.21 65.20 i 3.34 166.00 i 4.39
#Cells/3600 pm2 65.2 i 1.62 72.0 2 1.00 64.0 2 2.68 99.0 2 2.47 45.8 i 1.20 64.6 i 1.03 13.4 i 0.45
Grains per cell Sense Antisense 0.78 2 0.08 0.32 2 0.01 0.47 i 0.03 0.94 i 0.06 0.58 2 0.02 0.56 i 0.05 5.88 2 0.28
P
1.49 i 0.09 0.49 i_ 0.04 0.47 2 0.04 2.06 2 0.07 1.86 i 0.14 1.01 t 0.04 12.52 i 0.53
In Situ Hybridization Studies on Murine Catalase mRNA Expression During Embryonic Development D.L. REIMER AND S.M. SINGH Genetics Laboratories, Department of Zoology (D.L.R., S.M.S.) and Division of Medical Genetics (S.M.S.), University of Western Ontario, London, Ontario, Canada In situ hybridization using nuABSTRACT cleic acid probes was used to detect cell- and tissue-specific transcriptjs) of embryonic genes during development and differentiation. This highly sensitive technique has the potential to provide valuable information on the regulation of lowabundance housekeeping genes during development. We have determined the experimental conditions required to detect the catalase message in adult mouse liver. Catalase effects the breakdown of H , 0 2 to O2 and H,O and offers protection against the toxic effects of oxygen radicals. We used a cloned 550 bp BamHI-Pstl fragment from a mouse catalase cDNA (pMCT-1) to generate 35Slabeled sense and antisense riboprobes. The experimental conditions used were sensitive enough to quantitate the abundance of silver grains generated by the antisense riboprobe on the adult liver, a tissue known to be positive for this message. The hybridization protocol was applied to serial sections of 13- and 18-day-old mouse embryos. The results suggest that the catalase expression in the liver and brain begins with somite formation and increases with development and differentiation. O n the other hand, this message appears to be absent in mesenchyme, particularly in day 13 embryos. The message in positive tissues appears evenly distributed throughout the cell. The observed expression of the catalase message in the adult liver is approximately six times that in the embryonic liver. It is compatible with the enzyme activity results and emphasizes the sensitivity of the in situ hybridization method (over northern blot, etc.) used in this study.
Key words: Mice, housekeeping genes, liver, tissue specificity
hybridization analysis on cells or tissue sections with labeled DNA or RNA probe(s). The latter is particularly suitable for identifying specific cell types and individual cell components that harbor the gene-specific message. In situ hybridization methodology has been extensively used to evaluate the spatial and temporal patterns of gene expression during embryonic development in a variety of species including pre- and postimplantation mouse embryos [Anderson and Axel, 1985; Utset et al., 1987; Dean et al., 1989; Nisson et al., 19891. The technical improvements, including 35S-labeled cRNA probes, now make this technique sensitive enough to be able to detect two to ten target molecules per cell and permit reliable quantitation of any message [Davis et al., 1986; Simmons et al., 19891. The use of the sense transcript provides a suitable negative control imperative to such studies and the cRNA-mRNA hybrids are more stable than cDNA-mRNA hybrids and can therefore be subjected to higher stringency washes needed to eliminate nonspecific background binding. As a result, this protocol has the potential to be useful in studies on developmental regulation of low-abundance mRNAs in cells and tissues. Catalase (EC 1.11.1.6; hydrogen-peroxide: hydrogenperoxide oxidoreductase) is a housekeeping enzyme that provides protection against toxic effects of oxygen radicals generated through normal cellular metabolism. It is located primarily in the peroxisomes of eukaryotic cells and catalyzes the breakdown of H,O, to O2 and H,O [Aebi, 19841. In mice, it is coded for by a single structural gene [Dickerman et al., 1968; Holmes and Duley, 19751 and has been shown to be developmentally regulated [El-Hage and Singh, 1989, 19901. The catalase transcript is detectable as early a s 8 days postconceptus, with the beginning of somite formation. Although these studies, which rely on northern and slot blot hybridization, demonstrated that the expression of this gene is significantly higher in the develop-
INTRODUCTION Studies on gene expression are dependent on the availability of appropriate biological material and sensitive molecular technology. In general, the analysis of gene expression a t the level of mRNA (transcription/ pretranslation) relies on northern, slot blot, or in situ
Received for publication J u n e 4, 1990; accepted August 13, 1990. Address reprint requests to S. M. Singh, Genetics Laboratories, Department of Zoology, University of Western Ontario, London, Ontario, Canada N6A 5B7.
IN SITU HYBRIDIZATION STUDIES IN EMBRYOS 5'
A
P S
B
319
3' % Bg Xb , \
I
*
I I I I I
',
2.0 k b
*
\
\
, \ \
B
--II-' SP6
Blg
P
. -1
0.5 k b
T7
Fig. 1. Restriction map of the 2.0 kb mouse catalase cDNA (pMCT-1)and the strategy used to subclone the 0.5 kb BamHI-PstI fragment into the pGem-2 riboprobe vector. B, BamHI; P, PstI; Bg, BglII; S,SphI; X, XhoI; N, NsiI; A, ApaI; Xb, XbaI. The T7 and SP6 promotors were used to generate the antisense and sense riboprobes, respectively
ing liver than in the carcass, they do not identify specific tissues or cell types that are involved in the synthesis of the catalase message. Here we have investigated the spatial and temporal pattern of catalase gene expression in mouse embryos using in situ hybridization. We have applied 35S-labeled single-stranded riboprobes on serial paraffin sections of mouse embryos under experimental conditions determined during this investigation. The results were quantified, and the temporal and spatial pattern of expression of this gene during development was evaluated.
MATERIALS AND METHODS Tissue Collection, Fixation, and Sectioning Female mice, 11 weeks of age [Swiss Webster (SW)] were superovulated by intraperitoneal (i.p.) injection of 5 U pregnant mare's serum gonadotropin (PMSG; Sigma), followed by 5 U human chorionic gonadotropin (HCG; Sigma) 48 h r later. Each female was placed with a mature SW male overnight. The presence of a copulatory plug the following morning indicated successful mating (day 1of gestation). Pregnant females were sacrificed in order to collect fetuses representing days 13 and 18 postconception. Fetuses were dissected out in Sorenson's buffer (pH 7.0) and immersion fixed in buffered 4% paraformaldehyde overnight at 4°C. Adult liver was removed and cut into 2 cm cubes and immersion fixed a s well. The tissues were subsequently processed the following day on a 16 hr cycle on a Histomatic Tissue Processor (Fisher model 1661, and embedded in paraffin. Adult liver and sagittal sections of fetuses were cut a t 6 p.m on a microtome and collected on gelatinized and poly-L-lysine (Sigma)-coated microscope slides. In Situ Hybridization In situ hybridization was performed using modifications from Simmons et al. f19891. Deparaffinized slides
were pretreated by sequential immersion in 0.2 N HCl for 20 min, 2 x SSC (1x SSC = 0.15 M NaC1,0.015 M sodium citrate, pH 7.4) for 30 min, proteinase K (Boehringer Mannheim Canada), 25 pgiml for 15 min a t 37"C, acetic anhydride in 0.1 M triethanolamine HCl buffer (pH 8.0) for 10 min, rehydrated in graded alcohols, and air dried. A cDNA clone containing the 2.0 kb 3' end of the mouse catalase gene (pMCT-1) was generously donated by Dr. Robert Korneluk (Ottawa, Ontario, Canada). This cDNA clone was excised with BamHI and PstI (to remove poly-A' tail), and the resulting 550 bp insert was subcloned into the pGem-2 vector (Promega). The restriction map of the clone and the cloning strategy used in this study is shown in Figure 1. Radiolabeled RNA probes were synthesized in the presence of 35SUTP (Amersham) using a n RNA transcription kit (Promega). In this construct, T7 promotor-driven transcription yields antisense RNA that is complimentary to catalase mRNA. On the other hand, the SP6 promotor-driven transcription will yield sense mRNA to be used a s a negative control. The labeled probes were reduced by alkaline hydrolysis [Cox et al., 19841 and had specific activities of 1.9-2.9 x lo9 cpmibg. The hybridization mixture contained 0.2 pgiml '35Slabeled RNA transcripts, 50 mM dithiothreitol (DTT), 0.3 M NaCl, 50% (viv) deionized formamide, 10% (wiv) dextran sulfate, 0.2 mgiml sheared salmon sperm DNA, 0.125 mgiml tRNA, 0.02% (wiv) Ficoll, 0.02% (wiv) polyvinylpyrrolidone, 10 mM Tris (pH 7.4), 1 mM EDTA, and 0.1% Triton X-100. The mixture was applied to pretreated slides, coverslipped, and incubated a t 45°C for 4 hr. Following hybridization, the slides were washed twice in 4 x SSC with 5 mM DTT for 10 min, once in 50 pgiml RNAse A (Boehringer Mannheim Canada) in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 30 min a t 37"C, twice in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 15 min a t 37"C, twice in 0 . 1 ~ SSC
320
REIMER AND SINGH
with 5 mM DTT and 0.1% SDS for 10 min a t 50°C, 2 x SSC for 10 min, dehydrated in graded alcohols (each containing 0.3 M ammonium acetate), and air dried. Slides were dipped in NTB-2 nuclear track emulsion (Eastman Kodak Company, Rochester, NY) diluted (1: 1) with 0.6 M ammonium acetate and exposed for 2 days a t 4°C. The slides were subsequently developed in D19 (Eastman Kodak Company) for 5 min a t 15"C, fixed with 30% sodium thiosulfate for 5 min, and counterstained with Ehrlich's hematoxylin. Slides were prepared from three embryos at two developmental stages each. They were evaluated for a general pattern of tissue morphology and silver grains on different cells and tissue types. Although there was a tissue-specific distribution for silver grains a t both development stages, conclusions about the expression of the catalase gene could only be made following quantitative evaluation of such observations. Data was therefore collected on a representative slide for the adult liver and day 13 and day 18 embryos by quantitating the cell density and number of silver grains.
Data Collection and Statistical Analysis Each slide was evaluated using bright-field microscopy for characterization of cell types, cellular structures, and number of cells in a defined area. Dark-field microscopy was used to count silver grains per defined area on the same slides. Different numbers of silver grains were observed on different areas of each slide following hybridization with the labeled sense and the antisense riboprobes. Grains were counted in defined areas with and without (blank) tissue on the same slide. It was apparent that the number of silver grains was significantly higher on the tissue-containing areas particularly with hybridization using the antisense probe. Silver grain counts on tissues were corrected for grains on a representative area of the same slide without tissue (blank). Five random areas were counted for the number of silver grains on day 13 and day 18 embryonic tissues, whereas ten such numbers were generated for the adult liver. For each count of silver grains for a n area, the number of cells in the same area was recorded using bright-field microscopy. It is therefore possible to express these observations as grains per cell for all experiments. The difference between grain counts generated by antisense and sense 35S-labeled riboprobes was used to assess the expression of the catalase message in a tissue at a given stage of development. Significant differences in all pairwise comparisons were evaluated using Student's t test.
RESULTS In situ hybridization has been used to analyze genespecific expression during embryonic development, for high [Cox et aZ., 19841 and low [Davis et al., 1986; Simmons et al., 19891 abundance mRNAs. We used this technology to study the tissue-specific expression of a
low abundance mRNA of a housekeeping gene (catalase) during embryonic development in mice. To address the technical success of this protocol, paraffin sections of adult mouse liver (which is known to express catalase) were hybridized with 35S-labeled sense (negative control) and antisense mouse catalase riboprobes. A hematoxylin and eosin-stained section of the adult mouse liver (Fig. 2A) was used to evaluate the morphology and quantitate cell numbers per area. It shows epithelial cells, numerous sinusoids, and fully differentiated and relatively large hepatocytes. Molecular hybridization of this tissue section with the sense probe revealed few silver grains under dark-field (Fig. 2B). Here, the grain counts on the tissue differ minimally from the number of silver grains off the tissue. When a serial section of this tissue was hybridized with the antisense riboprobe (Fig. 2C), a n abundance of silver grains is observed. Furthermore, a bright-field evaluation of this section following counterstaining with hematoxylin (to enable identification of the nuclei of individual cells) shows cell density and a n apparent distribution of silver grains on the nucleus a s well as on the cytoplasm (Fig. 2D). These results suggest that hybridizable catalase mRNA is present in the adult liver and the transcript for this housekeeping gene in this tissue is detectable throughout the cell. We have applied the same experimental conditions to study the spatial and temporal expression of the catalase gene during development in mice. Paraffin sections from 13 day and 18 day embryos were subjected to this in situ hybridization and evaluation protocol. In day 13 and day 18 embryos, the silver grains were particularly localized to the liver, followed by brain cells or their progenitors. The grain counts on these two tissues were quantified and evaluated a t these developmental stages and are presented below. Figure 3 shows some representative tissue differences seen with the application of antisense riboprobe in developing embryos. For example, higher silver grain counts on a region of nucleated red blood cells in the tail bud of day 13 embryo were observed compared with the surrounding mesenchyme cells [Fig. 3A (dark-field) and B (brightfield)]. One may therefore conclude that the undifferentiated tissue (e.g., mesenchyme) has lower catalase mRNA level than nucleated red blood cells a t this stage of development. A similar tissue difference in grain counts on day 18 embryos is seen in Figure 3C (darkfield) and D (bright-field). The number of silver grains is elevated on the liver compared with the surrounding tissue types, including gut. An appropriate comparison of the same tissue a t different developmental stages, however, is needed to evaluate accurately the developmental expression of the catalase gene in mice. Molecular hybridizations of sense and antisense riboprobes to day 13 and day 18 embryo sections were analyzed in detail for developmental expression of catalase in the liver, brain (Fig. 41, and mesenchyme. The liver and brain are easily identified and reasonably
IN SITU HYBRIDIZATION STUDIES IN EMBRYOS
321
Fig. 2. In situ hybridization analysis of catalase expression in adult mouse liver. A Bright-field, stained with hematoxylin and eosin shows numerous sinusoids (thick arrows) and large hepatocytes (thin arrow). Serial sections (dark-field)hybridized with the sense probe (B) and the antisense probe (C). D Bright-field of C to show the distribution of the silver grains on the cells. Bar = 100 pm.
Fig. 3. Representative tissue differences seen with antisense riboprobe on mouse embryos. Dark-field (A) and bright-field (B) photographs of the tail region of a day 13 embryo showing specific hybridization to nucleated red blood cells (arrow);m, mesenchyme; dark-field (C) and bright-field (D) representation of a day 18 embryo showing elevated numbers of silver grains on liver (1) compared with gut (g).Bar = 200 km.
differentiated at the two developmental stages studied and show relatively high silver grain counts using the antisense probe compared with the sense control. In day 13 embryonic liver (Fig. 4A, bright-field), there are
numerous sinusoids, and the tissue is hemopoetic, highly dividing, and rapidly differentiating. The serial section hybridization of this tissue by the sense and antisense riboprobe showed that relative grain counts
322
REIMER AND SINGH
LIVER Fig. 4. Developmental expression of catalase in liver and brain of mouse embryos. Bright-field analysis of day 13 liver (A) shows sinusoids (arrow)and differentiating cells; silver grains following hybridization with the antisense riboprobe (B) dark-field; day 18 liver (C) bright-field, relatively differentiated tissue with sinusoids (arrow);D
Dark-field hybridization with the antisense riboprobe. Brain tissue of day 13 embryos (E, bright-field) and with silver grains following antisense riboprobe hybridization (F, dark-field); day 18 brain bright field ( G )and dark field (H). Bar = 100 pm.
were significantly higher with the antisense probe (Fig. 4B) compared with the sense probe (not shown). In comparison, the 18 day embryonic liver is more differentiated, with higher number of cells in a given area but still hemopoetic (Fig. 4C, bright-field). A number of sinusoids are also visible a t this stage in this tissue. The in situ hybridization analysis shows that, a s in day 13 embryos, the grain density is higher with the antisense probe (Fig. 4D) than with the sense probe. It may be pointed out that the grain count on the adult liver cells (Fig. 2C) is remarkably elevated compared with the embryonic stages. Furthermore, the density of silver grains on the 18 day embryonic liver (Fig. 4D) is higher than that of the day 13 embryos (Fig. 4B). The brain of the 13 day embryos possesses actively
dividing progenitors in their initial stages of differentiation with clear gaps and migrating cells (Fig. 4E, bright-field). Subsequent observations of in situ hybridizations on such embryos shows that the gene-specific riboprobe is able to detect expression of the catalase gene. Here, the silver grains generated by the antisense probe (Fig. 4F) are evident a t a lower level than those in the liver. The brain of the day 18 embryos is highly differentiated [Fig. 4G (bright-field), a s a n example], and individual cell types can be recognized. Here, the cells are densely packed, having reached their level of differentiation and organization with neural development nearing completion. The in situ hybridization on serial sections yielded a higher number of cell-specific silver grains on the brain cells a t this
IN SITU HYBRIDIZATION STUDIES IN EMBRYOS
323
TABLE 1. Silver Grain Counts on Different Tissues Following In Situ Hybridization With Antisense and Sense 35S-Labeled Riboprobes* Dev. stage 13 Day
Tissue Liver
18 Day
Brain Mesenchyme Liver Brain Mesenchyme
Adult
Liver
Grains per 3,600 pm2 Sense Antisense 50.40 i 3.97 22.80 i 0.86 30.20 2 1.20 93.60 2 4.07 26.60 2 1.25 35.80 2 2.65 77.80 2 1.76
96.60 i 4.53 35.60 i 2.71 29.80 2 2.74 203.80 2 5.50 84.80 i 4.21 65.20 i 3.34 166.00 i 4.39
#Cells/3600 pm2 65.2 i 1.62 72.0 2 1.00 64.0 2 2.68 99.0 2 2.47 45.8 i 1.20 64.6 i 1.03 13.4 i 0.45
Grains per cell Sense Antisense 0.78 2 0.08 0.32 2 0.01 0.47 i 0.03 0.94 i 0.06 0.58 2 0.02 0.56 i 0.05 5.88 2 0.28
P
1.49 i 0.09 0.49 i_ 0.04 0.47 2 0.04 2.06 2 0.07 1.86 i 0.14 1.01 t 0.04 12.52 i 0.53
