Abstract
A better understanding of biological phenomena involving modulations of gene expression requires quantitative analysis of the expression of several genes in the same structure. For this purpose, we have developed a novel in situ hybridization method to quantify two different mRNA species in the same tissue section simultaneously. Two probes labeled with radioelements of significantly different energies (3H and 33P or 35S) were used to detect the mRNA species. Radioactive images corresponding to the detected mRNA species were acquired with a Micro Imager, a real-time, high-resolution digital autoradiography system. An algorithm was used to process the data such that the initial radioactive image acquired was filtered into two subimages, each representative of the hybridization result specific to one probe. This novel method allows local discrimination and quantification of the respective contributions of each label to each pixel and can therefore be used for quantitative analysis of two mRNAs with a resolution of 15–20 μm.
I
In 1994, we described a new ISH approach based on the direct detection of radioactive emission by using the high resolution of a radio imager to analyze mRNA expression in brain tissue sections (Charon et al. 1990; Lanièce et al. 1994, 1998). The main advantage of this approach over standard autoradiographic approaches is the possibility of quantifying mRNA in real time and with a high dynamic range (104), leading to cellular resolution in shorter delays. Recently, we have improved the use of this device by developing adequate signal acquisition and processing algorithms to discriminate different radioactive emission spectra obtained simultaneously. Here, we demonstrate simultaneous ISH of two radioactive probes on the same tissue section, each probe being labeled with radioelements of significantly different energies (3H and 33P or 35S). We also demonstrate that this allows quantitative analysis of two mRNAs in a single section.
Materials and Methods
Tissue Preparation
One male adult Sprague-Dawley rat (Iffa Credo; L'Arbresle, France), weighing between 350 and 400 g, was anesthetized with urethane carbamate (1.5 mg/kg), and placed in a stereotaxic frame for electric stimulation. The animal was sacrified and its brain was extracted and frozen in isopentane at −60C. Coronal sections (20-μm-thick) were cut on a cryostat at −22C. Sections were mounted on Superfrost Plus slides and stored at −80C. All experimental procedures were carried out in accordance with the European Communities Council Directive (24.xi.1986) and with the guidelines of the CNRS and the French Agricultural and Forestry Ministry (decree 87848, license number A91429).
Double Radioactive ISH
Two oligonucleotide probes were used for these experiments. One is complementary to part of the syntaxin 1B sequence (35-mer oligonucleotide sequence 5′-GAT GTG TGG GGA GGG TCC TGG GGA AGA GAA GGG TA-3′) and the other to part of the Homer sequence (39-mer oligonucleotide sequence 5′-GGT CAG TTC CAT CTT CTC CTG CGA CTT CTC CTT TGC CAG-3′). Oligonucleotides were synthesized in house on a Beckman Oligo 1000DNA synthesizer. The probes were 3′ end-labeled with [35S]-deoxyadenosine triphosphate (Amersham; Orsay, France) or [3H]-deoxycytosine triphosphate (Amersham) in a tailing reaction, using terminal deoxynucleotide transferase (Amersham) according to the manufacturer's instructions. The specific activity after labeling was between 1 × 108 and 3 × 109 cpm/μg for each probe.
Coronal brain sections (20-μm-thick) were postfixed in 4% paraformaldehyde in PBS, then washed three times for 10 min in PBS baths and dried in a 95% ethanol bath immediately before hybridization. The hybridization solution was composed of 50% Amersham in situ hybridization buffer, 40% formamide (Eurobio; Les Ulis, France), 0.1 M dithiothreitol (DTT) (Euromedex; Souffel Weyersheim, France), and 0.5 mg/ml poly(A) (Roche; Saint Quentin Fallavier, France). Both probes were diluted 1:100 in the hybridization solution and 75 μl of the mixture was applied to each brain slice. Sections were incubated overnight at 50C under Fuji parafilm coverslips, then washed twice for 15 min in 1 × standard saline citrate (SSC)/10 mM DTT at 53C, twice for 15 min in 0.5 × SSC/10 mM DTT at 53C, and once in 0.5 × SCC/10 mM DTT at room temperature and then dried in a 95% ethanol bath. Radioactive signals from the sections were acquired with a Micro-Imager (Biospace Mesures; Paris, France), which is a real time, high-resolution digital autoradiography system.
Imaging Equipment for Radiolabeled Tissue Sections
To analyze the double radiolabeling in the sections, a thin foil of scintillating paper is brought into contact with the sections. β-Particles emitted by the sections are identified through acquisition of the light spot emissions in the scintillating foil by a CCD camera coupled to an image intensifer. The result of the acquisition is displayed live on a computer. During the acquisition, radioactive images can be saved at any time to be analyzed. Acquisition is stopped once the number of disintegrations acquired is statistically sufficient. The filter processing allows discrimination and quantification in each pixel of the respective contributions of the two radioelements of significantly different energies.
Results
To show the feasibility of simultaneous ISH of two radioactive probes on the same section, electric stimulations were used for neuronal activation in one side of a rat brain. The expression of the two genes studied, Homer and syntaxin 1B, which are differentially regulated, was followed. Because this work aimed only to validate the technique, the physiological implications of the findings are not addressed.
The principle of the double-labeling ISH technique is illustrated in Figure 1. [35S]-dATP and [3H]-dCTP were chosen to label two different probes that were simultaneously hybridized to a single tissue section. The Micro Imager was used to acquire the signal from the hybridized section in a single step. The initial image was consequently filtered to segregate the image corresponding to [3H]-β disintegrations (Figure 2C) from that corresponding to [35S]-β disintegrations (Figure 2E). The quantitative data for both 3H and 35S labeling were incorporated into a single image (Figure 2A). In this figure, green corresponds to the cells that contain only the mRNA detected by the 3H-labeled probe, red to those that contain only the mRNA detected by the 35S-labeled probe, and shades of yellow to those that contain both.
Principle of the double-labeling technique in ISH. Two differently labeled probes are simultaneously hybridized on the same tissue section. After washing, the section is read by the Micro Imager. The initial image acquired is then filtered to segregate the image corresponding to [3H]-β disintegrations from that corresponding to [32P/35S/33P]-β disintegrations.
To control the filtering segregating [35S]-β disintegrations from the [3H]-β-disintegrations, three control dots were spotted by hand on the slide. The dots contained, respectively, the 3H-labeled probe (200 cpm), a mix of the 3H (200 cpm)- and the 35S (200 cpm)-labeled probes, and the 35S-labeled probe (200 cpm). All three spots are observed in the image with both labels (Figure 2A) and only two dots after filtering, as expected (Figure 2C and 2E). Quantification of the radioactivity emitted by each dot before and after filtering gave values in accordance with the amount of radioactivity spotted.
The expression of the two mRNAs along a line drawn on the section is quantitatively analyzed in Figure 2 for illustration. The respective contribution of each label to each pixel along this line is shown on graphs (Figure 2B, 2D, and 2F). From the graphs, cells that differentially expressed the two mRNA species are clearly identified and others expressed them at a similar level. This novel method allows quantitative comparison of the expression of these mRNAs in different cells. For example, the cells indicated by Arrow 4 expressed about five times as much mRNA hybridizing with the 3H-labeled probe as the cells indicated by Arrow 3. They also contain large amounts of mRNA detected by the 35S-labeled probe, whereas the amount in the cells indicated by Arrow 3 is barely detectable (Figure 2).
Discussion
A number of ISH protocols have been developed. They use either enzymatically synthesized RNA and DNA probes or chemically synthesized DNA probes (“oligodeoxynucleotide” probes). Standard protocols use either nonradioactive or radioactively labeled probes. The method of signal detection used depends on the required level of resolution and sensitivity and also on the physiological context (for reviews see Valentino et al. 1987; Chesselet 1990; Wilkinson 1994).
Nonradioactive probes are mainly used for anatomic analyses of gene expression because they provide the greatest spatial resolution and they allow detection of several mRNAs in the same tissue section (peroxidase/alkaline phosphatase) (Dumas et al. 1992), in the same cell (fluorescence) (Nederlof et al. 1990), and even in a confocal microscopic field for subcellular discrimination (Paratore et al. 1999). Moreover, the results are obtained rapidly (1 or 2 days). However, nonradioactive probes do not provide quantitative results concerning the level of gene expression and are useful only for identification of the cells that contain a particular mRNA or DNA (for a review see Wilkinson 1994).
In contrast, radioactive labeling allows precise measurement of the level of gene expression (for reviews see Valentino et al. 1987; Chesselet 1990). Four isotopes can be used for labeling probes: 3H, 35S, 33P, and 32P. Various methods are used to quantify mRNA: classical autoradiographic methods (film and emulsion) (Dumas et al. 1990, 1996); indirect detection through storage in phosphor screens (Ito et al. 1995); and direct detection through a solid scintillator sheet coupled to a CCD camera (Micro Imager) (Lanièce et al. 1994, 1998). Each of these methods has both advantages and drawbacks.
For analysis of the regional distribution of mRNA, storage phosphor screens [resolution of 80 μm (3H) and 180 μm (35S/14C)] and autoradiographic films (20–30-μm) allow quantification of signals with exposure times of several days to weeks for films and eightfold less for storage phosphor screens. To detect mRNA in individual cells, the hybridized sections are usually dipped in nuclear emulsion: the amount of the mRNA can be quantified at a cellular level by counting grains. The exposure time required for this technique is often long, from several weeks to several months depending on the amount of mRNA in the tissue (for a review see Valentino et al. 1987). These three radioactive techniques cannot be used for simultaneous analysis of two mRNA species in a single section.
Here, we demonstrate that the Micro Imager, in contrast, allows quantitative co-detection. Moreover, this is performed in real time, with a high dynamic range (104), satisfactory resolution (15 μm), and exposure times 10 times shorter than autoradiographic films and 50 times shorter than emulsion (Lanièce et al. 1994). The high dynamic range of the Micro Imager allows the comparative analysis of weak and strong signals on the same tissue section, such expression profiles being commonly observed in the CNS. The accuracy is better than 5% without underexposure owing to the direct particle counting principle of the instrument in real time such that acquisition can be halted at the appropriate time. Very small variations of expression for several genes can therefore be measured with high accuracy on a same section.
Our ISH experiments, performed with two different labeled probes (3H/35S), demonstrate the feasibility of double-labeling procedures to study simultaneously the expression of different mRNA species in a single tissue section. To our knowledge, this is the first article reporting ISH detection of more than one transcript and the quantification of their respective expression, allowing the comparison of expression of several genes at the cellular level. The findings with this approach were compared with those obtained by independent single-labeling ISH experiments on adjacent sections. As expected, the expression patterns observed were qualitatively and quantitatively similar.
Visualization (arbitrary colors) of the results of a double radioactive ISH. (
Two probes hybridized on the same section can be distinguished from each other only if the radioisotopes used to label them have different emission energy spectra. We labeled one probe with 3H and the other with either 33P or 35S. 35S and 33P have similar spectra but different half-lives. However, the 3H energy spectrum is clearly different from those of 33P and 35S (Valentino et al. 1987). The disintegration half-life of 3H is more than 1 log (10 times) longer than those of 33P and 35S. Therefore, the frequency of disintegration events is much lower with 3H for a given amount of isotope and is, in part, the reason for the long exposure times commonly used with 3H labeling (such as in autoradiographic techniques). For the double-labeling technique, it is crucial that both labeling signals are simultaneously acquired. However, when separately adapting the probe labeling procedures for each of these radioisotopes, we were able to establish a protocol in which acquisition times were equivalent for both 3H and the other isotopes. This was made possible by the specificity of the acquisition principle used by the Micro Imager. Because it is based on a particle counting method, it is equally sensitive to β disintegrations of all energies, including 3H. This feature allows a single acquisition of the images corresponding to the 3H and 33P (or 35S) isotopes. Discriminating a third isotope, such as 32P, from both 3H and 33P/35S is also feasible with adequate adaptation of signal acquisition software.
In situ hybridization has already made an enormous contribution to our understanding of how cellular events interrelate and how mRNA is organized, spliced, and transported. Radioactive detection may now further improve the power of this approach and is suitable for gene expression screening on tissue sections. It may also allow novel types of experiments, e.g., co-detection of an mRNA species (with a radiolabeled nucleotide probe) and a protein (with a 125I-radiolabeled antibody). Furthermore, the co-detection of two radioactively labeled species could be used in conjunction with the detection of other molecules using nonradioactive labeled probes or reagents. This would allow the quantitative and qualitative analysis of five markers on a single tissue section, two of them being labeled with radioactive molecules.
Footnotes
Acknowledgements
Supported by the Centre National de la Recherche Scientifique and the Conseil Régional de l'Ile de France. H. Salin was supported by a PhD studentship from the Ministère de l'Education Nationale de la Recherche et de la Technologie.