Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 May 25.
Published in final edited form as: In Vitro Toxicol. 1997;10(3):295–308.

Detection of Cytochrome P450 mRNA in Tissue Sections and Cell Lines Using Enzyme-Labeled Fluorescence In Situ Hybridization

CATHERINE VILLAROMAN 1, FEDERICO M FARIN 1, JASPREET S SIDHU 1, DOLPHINE ODA 2, CURTIS J OMIECINSKI 1
PMCID: PMC3360469  NIHMSID: NIHMS374868  PMID: 22639489

Abstract

Cytochrome P450s (P450s) constitute a superfamily of enzymes that metabolize a broad array of xenobiotics. The ability to measure basal and induced levels of P450 mRNA in specific cells and tissues should provide valuable insight regarding the functional role and heterogeneous expression of these enzymes in chemically related diseases. Methodologies for detecting cell-specific mRNA expression patterns typically rely on radiolabeled probes and photographic emulsions, often coupled with long exposure times. These studies were conducted to evaluate an enzyme-labeled fluorescence (ELF) in situ hybridization technique to detect specific P450 mRNA. Deparaffinized, formalin-fixed tissue sections and cells from culture were incubated for 12 hours with 5′-biotinylated 20-base DNA oligomer probes (20-mer). Specific hybridization was detected using a streptavidin alkaline-phosphatase conjugate followed by incubation with the ELF substrate, yielding a bright, yellow-green fluorescent signal. In this study, utility of the technique was demonstrated using cultured rat hepatorna cells, and tissue sections from rat liver and human oral epithelium. Ribonuclease A pretreatment of the sample, omission of the probe, competition with a nonbiotinylated oligomer, and the use of only partially homologous probes served as negative controls to demonstrate the specificity of the hybridization signal. Our results clearly demonstrated the ability of ELF in situ hybridization to discriminately detect cell-specific P450 mRNA in tissue sections and cultured cells. This technique eliminates the use of radioactivity and enables in situ detection of mRNAs with relative ease, efficiency, specificity, and high sensitivity.

INTRODUCTION

Cytochrome P450s (P450s, CYPs) are a superfamily of heme-containing enzymes responsible for catalyzing the oxidative metabolism of a vast array of endogenous and xenobiotic substrates. These enzyme systems are localized principally in the cellular smooth endoplasmic reticulum and are widely distributed among cell types. Although the liver is in the major organ determining total body disposition of xenobiotic metabolites, P450s in extrahepatic tissues likely play significant roles in mediating organ-specific toxicity and carcinogenicity (Gram et al., 1986; Farin et al., 1995).

In situ nucleic acid hybridization (ISH) is a sensitive tool for exploring the cellular and tissue-specific expression of genes. Conventional ISH typically employs radiolabeled nucleic acid probes together with radiosensitive emulsions. The resulting signals often require days or even weeks to visualize. In cases of low gene expression level, distinguishing between background and low levels of reduced silver grains in the in situ autoradiographs can be difficult (Hassett et al., 1989). Furthermore, radioactive probes are usually unstable and require special handling precautions, thus making it desirable to have an alternative but equally sensitive method for detecting or localizing specific nucleic acid sequences (Holm et al., 1992).

Fortunately, nonradioactive ISH methods have emerged over the Past decade. These techniques are rapid, usually producing results in 1–4 working days, and provide improved spatial resolution of the hybridization product (Holm et al., 1992; speel et al., 1992). The sensitivity of nonradioactive techniques appears comparable to that of radioactive procedures (Syrjanen et al., 1988; Holm et al., 1992; Speel et al., 1992). Moreover, non-isotopic in situ hybridization allows the simultaneous analysis of multiple probes, which if combined with fluorophore-labeled antibodies could also permit a concomitant visualization of both protein and mRNA (Nederlof et al., 1989; Speel et al., 1992; Larison et al., 1995).

In this study we developed the sensitive enzyme-labeled fluorescence in situ hybridization assay to examine mRNA expression profiles for rat and human CYP1A1, using cultured cell lines as in vitro models for P450 induction as well as paraffin-embedded tissue sections of both human and rat origin. Furthermore, by combining immunocytochemical methods and ISH we provide a compelling visual assessment of the expression of the P4501A1 gene in individual cells both at the protein and mRNA levels.

MATERIALS AND METHODS

Cell Culture Materials

Dulbecco's modified Eagle's:Ham's F12 (DMEM:F12), Dulbecco's phosphate buffered saline (PBS), as well as other cell culture media constituents, were procured from Gibco BRL (Grand Island, NY). Nu-Serum™ was obtained from Collaborative Research, Inc. (Bedford, MA). Falcon 60-mm tissue culture dishes were acquired from Becton Dickinson and Company (Franklin Lakes, NJ). Beta-naphthoflavone (β-NF) was purchased from Aldrich Chemical Co. (Milwaukee, WI), and levamisole was acquired from Sigma Chemical Co. (St. Louis, MO). All other chemicals and biochemicals were of the highest grade available from commercial sources.

Formalin-fixed, Paraffin-Embedded Tissue Sections

Formalin-fixed tissue sections, 5-μm thick, were cut from paraffin blocks and mounted on poly-L-lysine coated slides (Stat Path, Riderwood, MD). Tissue sections of oral epithelium were obtained from the Oral and Maxillofacial Pathology Biopsy Service archives at the University of Washington. Sprague-Dawley rat liver tissue was isolated 1 day after an 80 mg/kg intraperitoneal β-NF injection, whereas the controls were injected with saline for 1 day.

Oligomer Probes

High performance liquid chromatography-purified, 5′-biotin conjugated oligomer probes were purchased from Genosys Biotechnologies, Inc. (The Woodlands, TX). Selection of the antisense oligomers was based on published cDNA sequences in the GenBank/EMBL database. The rat CYP1A1 20-mer, 5′-GGACCAGAAGACCGCATCTG-3′, is complementary to rat CYP1A1 mRNA, position 5603–5622 of the complete gene sequence (Sogawa et al., 1984). Another 20-mer, 5′-CAGGCAGGATCCCTTAGGCT-3′, is homologous to the human CYP1A1 mRNA sequence at position 7120–7139 (Jaiswal et al., 1985). A nonbiotinylated rat CYP1A1 oligonucleotide was also purchased from Genosys with an identical 20-base sequence as the rat CYP1A1 biotinylated oligomer described above. Probes were diluted with sterile water and aliquoted as stock concentrations of approximately 50 μg/ml and stored at −20°C.

Culture of Rat Hepatoma Cells

Rat hepatoma H4IIE cells were obtained from the American Type Culture Collection (Rockville, MD). Rat hepatoma Fao cells were a kind gift from Dr. John Koontz (University of Tennessee, Knoxville, TN). Cell culture conditions were similar to those described previously (Sidhu et al., 1993). These cells were grown as continuous monolayer cultures in DMEM:F12 (1 : 11), supplemented with 1% penicillin-streptomycin, 2-mM glutamine, 2.2-mg/ml tissue culture grade sodium bicarbonate, and 9% Nu-Serum™ in 75-cm2 tissue culture flasks (Becton Dickinson). After reaching confluency, the cells were seeded on 60-mm Falcon tissue culture dishes using nonenzymatic cell dissociation buffer (Sigma). These cells were propagated in 3 ml of medium; after attaining confluency, cell cultures were maintained in serum-free William's E medium overnight (Sidhu et al., 1994). Treated cells were then incubated for 24 hours with 22-μM β-NF, dissolved in 0.05% (v/v) cell culture grade dimethyl sulfoxide (Sigma). Cells were then ready for ELF in situ hybridization.

In situ Hybridization/ELF Detection

Enzyme-labeled fluorescence ISH experiments were performed on control and β-NF-exposed rat hepatoma cell lines Fao and H4IIE. After inducer treatment, control and β-NF-induced cell cultures were washed twice for 5 min with PBS preceding fixation with 4% formaldehyde, 5% acetic acid in 0.9% (w/v) NaCl for 30 min. Experiments were performed at room temperature unless otherwise stated. Cells were dehydrated through a graded series of ethanol treatments (70, 85, and 100%), immersed in methanol for 10 min to remove residual lipids, then rehydrated sequentially through the same series of ethanol. After washing with PBS for 5 min, cells were postfixed for 10 min with 1% formaldehyde buffered in PBS, followed by another 5 min PBS wash. Hybridization buffer (2 ml) containing 1× Denhardt's (0.02% polyvinylpyrrolidone, 0.02% Ficoll type 400 and 0.02% BSA), 5× SSC (1× SSC = 0.15 M sodium chloride, 0.015 M sodium citrate, pH 7.4), 0.025 M sodium phosphate, pH 6.5, 20 μg/ml poly A, 1 U/μl RNasin (Promega), and diethylpyrocarbonate-treated distilled water (DEPC H2O) was added to the sample dishes, and placed in a humidified chamber at 55°C for 2 hours as a prehybridization step. Subsequently, 1 ml of this same hybridization solution containing 4.5 μg/ml of a biotinylated probe was added to the culture dishes, leading to a final concentration of 1.5 μg/ml. The cells were incubated for 12 hours at 55°C. After hybridization, unbound probe was removed by washing three times, 10 min each, with 4× SSC at room temperature. All ELF buffers and substrates were procured from Molecular Probes, Inc. (Eugene, OR). Samples were rinsed for 5 min in ELF wash buffer, followed by a 30-min incubation in the ELF blocking reagent with a plastic coverslip. Cells were then treated with 2 mM of levamisole diluted in blocking reagent and a coverslip was applied for 45 min. This step inhibits endogenous alkaline phosphatase activity. After washing for 5 min with PBS, the streptavidin-alkaline phosphatase conjugate (Boerhinger-Mannheim, Mannheim, Germany), diluted 1:100 in ELF blocking reagent, was applied to the samples along with a coverslip for 30 min. Cells were then washed twice with ELF wash buffer, for 5 min each, proceeded by development with filtered 1:10 dilution of the ELF substrate working solution. Hoechst 33342, a blue fluorescent nucleic acid stain, was diluted with DEPC water to a concentration of 1.8 μg/ml, and then placed on the samples for 10 min. After rinsing samples twice for 5 min each with ELF wash buffer, two drops of mounting medium were added along with a glass coverslip. Signals were visualized as a bright yellow-green precipitate under an ultraviolet (UV) microscope using an amino methyl coumarin acetic acid (AMCA) filter set (excitation 365 ± 8 nm; emission 450 ± 33 nm) on a Nikon cube (Meridian Instrument Company, Inc., Auburn, WA). Photographs were taken using Ektachrome Elite 400D color slide film (Eastman Kodak Company, Rochester, NY). The inducibility of CYP1A1 following treatment with β-NF was assessed qualitatively from the changes in the levels of signal as compared to control cultures.

Competition experiments involved adding a 60-fold excess of nonbiotinylated CYP1A1 probe into the same hybridization solution containing a much lower concentration of the biotin-la-beled CYP1A1 probe. Another negative control experiment involved treating the cells with RNase A (Boehringer Mannheim). Following the fixation step with 4% formaldehyde, 5% acetic acid in 0.9% (w/v) NaCl, cells were washed with 2× SSC twice for 4 min each. The cells were subsequently exposed to 150 μg/ml of RNase A diluted in 2× SSC and then incubated at 45°C for 60 min. After two washings with 2× SSC, the cells were subjected to the same procedure as described above, resuming with the ethanol dehydration step.

We optimized the use of the ELF in situ hybridization method empirically. High probe concentrations were used to achieve optimal signal amplification and rapid results. To produce similar results with lower dilutions of probe, a longer development time was required (data not shown). Several investigators have previously established the importance of maximizing sensitivity when using chemically biotinylated probes over other labels, such as digoxygenin, for detecting mRNA by ISH (Herrington et al., 1989; Holm et al., 1992; Larsson and Hougaard, 1993; Arai et al., 1988). A variety of other parameters involved with the ISH procedure were tested to ensure retention of RNA and integrity of cellular morphology. For both paraffin-embedded sections and cultured rat hepatoma cells, background signals were much lower after hybridization at 55°C versus the lower temperatures tested (data not shown). Acetic anhydride and Denhardt's solution were used to help eliminate nonspecific binding of the biotinylated probes. Filtering of the ELF reagents improved the quality of the results by further reducing background signals and ensuring purity of the solutions.

P450 mRNA expression was also examined in formalin-fixed, paraffin-embedded tissue sections using similar procedures as with cultured cells. First, tissue sections were deparaffinized with xylene twice for 15 min each. Slides were then rehydrated by passing them through decreasing concentrations of ethanol (2 × 100, 97, 95, 85, 70, and 50%), 4 min each. After a 5-min bath in PBS, samples were incubated with 0.2 N HCl for 20 min, followed by a 15-min wash with 0.3% triton X-100 in PBS. Sections were then postfixed in freshly prepared 4% formaldehyde in PBS for at least 5 min. To acetylate the amino groups in the tissue in order to reduce the nonspecific binding of the oligomer, slides were incubated in 0.25% acetic anhydride (v/v) containing 0.1 M triethanolamine-HCl buffer, pH 8.0, for 10 min at room temperature. Probe clips (Research Products International Corp., Mount Prospect, IL) with siliconized gaskets were used to hold 500 μl of the prehybridization solution, which was then overlaid and hermetically sealed onto tissue sections. Samples were incubated for 2 hours in a humidified chamber at 55°C. Subsequently, 500 μl of the same solution but containing a probe concentration of approximately 3.0 μg/ml was then placed over tissues and hybridized for 12 hours at 55°C. After hybridization, the probe clips were carefully removed, followed by three 10-min washes in 4× SSC at room temperature. Posthybridization and ELF detection steps were similar to those aforementioned. Negative controls included omission of probe from the hybridization buffer, or substitution of the human CYP1A1 probe with the 35% mismatching rat CYP1A1 probe when working with human oral epithelium.

To ensure reproducibility of the in situ results for all tissue sections and cell lines examined, the ELF technique was performed at least five times on each sample.

ELF in situ Hybridization and Immunocytochemisty

Characterization of the CYP1A1 antibody and the peroxidase immunocytochemistry method was described previously (Farin and Omiecinski, 1993). The protocol for simultaneous detection of CYP1A1 protein and mRNA followed the same steps as those described above for paraffin-embedded sections, with the exception that immunohistochemistry was performed prior to ELF detection. After the stringent 4× SSC wash steps, samples were rinsed twice for 5 min in PBS then treated with 4% normal goat serum (NGS) (Sigma) diluted in PBS for 30 min to decrease nonspecific binding of antibody. The sections were later washed in PBS and exposed for 60 min to 1:200 rabbit polyclonal primary CYP1A1 antibody diluted in 4% NGS. Test sections were incubated with the primary antibody at dilutions ranging from 1:50 to 1:400 to determine the concentration for optimal staining (data not shown). Unbound primary antibody was removed by a thorough rinse in PBS, with subsequent application of 0.05 μg/ml rhodamine-labeled anti-rabbit IgG secondary antibody (Boehringer Mannheim, Indianapolis, IN) diluted with 4% NGS in PBS for 60 min. The unbound secondary antibody was further washed with PBS. ELF detection steps ensued thereafter. Sections incubated without primary antibody served as a negative control for the specificity of the immunostaining reaction. This triple-labeled staining procedure, including the blue fluorescent nucleic acid stain, Hoechst 33342, the rhodamine-labeled secondary antibody, and the fluorescent signal produced by the enzymatic cleavage of the ELF substrate, can be observed by interchanging filters on the microscope and double-exposing the film. The typical exposure times were 2–3 sec for the ELF and Hoechst signals and 30–40 sec for the rhodamine signal. The filter combination used was FITC/Texas Red dual band filter and the AMCA filter set (Meridian Instrument Company, Inc., Auburn, WA).

RESULTS

Cultured Rat Hepatoma Cells

Figure 1 illustrates ELF in situ hybridization detection of CYP1A1 mRNA using H4IIE (panels A–C) and Fao (panels D–F) rat hepatoma cultures with and without exposure to the prototypic CYP1A1 inducer, β-NF. After hybridization with the rat CYP1A1 biotinylated probe, a low-level ELF signal was detected in control cultures of H4IIE (Figure 1A) and Fao (Figure 1D). Figures 1B and 1E show an increased level of ELF signal after the cell cultures were exposed to 22-μM β-NF for 24 hours compared to the control cultures (Figures 1A and 1D). The ELF signal was generally dispersed throughout the cytoplasm, although certain cells exhibited a more focal pattern. ELF signal emanating from the nucleus was not observed. Omission of the rat biotin-conjugated CYP1A1 probe from the hybridization protocol resulted in the absence of ELF precipitate in both β-NF-treated H4IIE (Figure 1C) and Fao (Figure IF) cultures.

FIG. 1.

FIG. 1

Open in a new tab

Enzyme-labeled fluorescence (ELF) in situ hybridization using a rat biotin-conjugated CYP1A1 oligomer probe with cultured rat hepatoma cells. (A9 untreated H4IIE cells probed with biotinylated CYP1A1 oligomer. (B) 22-μM beta-naphoflavone (β-NF)-treated H4IIE cells hybridized with biotin-labeled CYP1A1 oligomer (C) 22-μM β-NF exposed H4IIE cells lacking the biotinylated CYP1A1 probe from the hybridization solution. (D) Unexposed Fao cells probed with biotinylated CYP1A1 oligomer. (E) 22-μM β-NF-treated Fao cells hybridized with biotin-labeled CYP1A1 oligomer (F) 22-μM β-NF-exposed Fao cells without the biotinylated CYP1A1 probe. Magnification 400×.

The data in Figure 2 present various negative controls to demonstrate the specificity of the ELF in situ technique. The H4IIE cultures (Figures 2A–C) were exposed to β-NF and subjected to ELF in situ hybridization using the biotinylated rat CYP1A1 oligomer. Figure 2A shows a diffuse fluorescent precipitate abundantly localized over the cytoplasm of β-NF-treated H4IIE cultures. In contrast, the fluorescent signal produced by hybridization with the biotin-conjugated rat CYP1A1 probe was completely abolished when β-NF treated H4IIE cells were pretreated with 150 μg/ml of RNase A (Figure 2B). Moreover, the fluorescent ELF signal was not observed when using the biotin-labeled rat CYP1A1 probe along with a 60-fold excess of an identical, non-biotinylated rat CYP1A1 probe (Figure 2C). These results provided additional confirmation that the ELF in situ hybridization signal was not spurious (Bursztajn et al., 1990; Larsson and Houggard, 1993; Wilcox, 1993).

FIG. 2.

FIG. 2

Open in a new tab

Enzyme-labeled fluorescence (ELF) in situ hybridization control experiments utilizing cultured rat hepatoma cells and formalin-fixed, paraffin-embedded human tissue sections. (A) 22-μM beta-napthoflavone (β-NF)-induced H4IIE rat cells hybridized with rat biotin-conjugated CYP1A1 oligomer. (B) 22-μM β-NF-treated H4IIe rat cells probed with the rat biotin-labeled CYP1A1 20-mer and pretreated with 150 μ/ml of RNase A. (C) 22-μM β-NF-exposed H4IIE rat cells hybridized with rat biotinylated CYP1A1 2Gmer plus 60-fold excess nonbiotinylated rat CYP1A1 probe. (D) Poorly differentiated squamous cell carcinoma of human oral epithelium probed with human CYP1A1 biotin-labeled 20-mer. (E) Poorly differentiated squamous cell carcinoma of human oral epithelium hybridized with the partially homologous rat CYP1A1 biotin-conjugated oligomer probe. (F) Poorly differentiated squamous cell carcinoma of human oral epithelium without the addition of a biotinylated oligomer to the hybridization buffer. Magnification 400×.

Hybridizations were conducted incorporating the partially homologous biotinylated rat CYP1A1 probe (7/20 mismatches with the human CYP1A1 mRNA target sequence) using a tissue section of poorly differentiated squamous cell carcinoma of human oral epithelium. These experiments resulted in no observable ELF precipitate (Figure 2E). In parallel, the same human oral epithelium lesion was incubated with a human biotin-labeled CYP1A1 probe. The latter probe resulted in ELF signal primarily localized over the well-differentiated cells (Figure 2D). Elimination of the probe from the hybridization protocol failed to yield any detectable yellow-green ELF precipitate (Figure 2F).

Paraffin-Embedded Tissue Sections

The specific cell types that express CYP1A1 mRNA and protein were identified in normal tissue samples of human oral epithelium as well as in rat liver sections. Figures 3A–C demonstrate the immunocytochemical staining pattern of normal oral epithelium using a peroxidase-conjugated anti-rabbit IgG secondary antibody directed against the rabbit polyclonal primary CYP1A1 IgG antibody. Strong immunoreactivity for CYP1A1 was evident across the upper, more differentiated spinous and granular epithelial cell layers, whereas only weak reactivity was seen in the undifferentiated basal and parabasal regions (Figure 3A, higher magnification in 3B). When the primary antibody was eliminated from the immunocytochemical protocol, staining for the CYP1A1 protein was completely absent (Figure 3C).

FIG. 3.

FIG. 3

Open in a new tab

Immunocytochernistry and enzyme-labeled fluorescence (ELF) in situ hybridization on paraffin-embedded, normal human oral epithelium tissue sections. A–C, Cellular localization of cytochrome P4501A1 protein. (A) (100×) Section exposed to the rabbit polyclonal primary CYP1A1 antibody and a peroxidase-conjugated anti-rabbit IgG secondary antibody. (B) (400×) Higher magnification of the same section displayed in A. (C) (100×) Section lacking the rabbit polyclonal primary CYP1A1 antibody. (D–F) Cellular localization of cytochrome P4501A1 mRNA (D) (100×) Section probed with the human biotin-conjugated CYP1A1 oligomer. (E)(400×) Higher magmfication of the same section pictured in D (F) (100×) Tissue section minus the human biotinylated CYP1A1 probe from the hybridization protocol.

Paraffin-embedded human oral epithelial tissue sections, similar to those used for immunocytochemical staining, were incorporated for ELF in situ hybridization using the human biotinylated CYP1A1 oligomer. A strong ELF signal was distributed over the upper, more differentiated spinous and granular layers of epithelium as compared to the absence or weak fluorescent precipitate found in the lower, undifferentiated basal cell layer (Figure 3D, higher magnification in 3E). This pattern was quite similar to that observed for the CYP1A1 protein described above and shown in figures 3A and 3B. When the biotin-labeled 20-mer was removed from the hybridization solution, the ELF signal was not detectable (Figure 3F).

Paraffin-embedded rat liver sections obtained from a β-NF-treated animal also were examined for CYP1A1 mRNA expression using the rat biotin-labeled CYP1A1 20-mer. ELF in situ hybridization with these β-NF-exposed sections presented a widespread but heterogeneous distribution of the resultant ELF signal. The cells surrounding hepatic venules exhibited a higher density of ELF precipitate than the periportal hepatocytes (data not shown). In comparison, paraffin-embedded rat liver sections isolated from untreated animals displayed less reactive cells with a substantially reduced ELF signal (data not shown). Deletion of the rat biotin-conjugated CYP1A1 probe from the hybridization protocol yielded no signal on similar sections (data not shown). Ultimately, CYP1A1 protein and mRNA were simultaneously detected in these sections using both ELF in situ hybridization and immunocytochemistry, as described in the next section.

ELF in situ Hybridization and Immunohistochemisty

Examining CYP1A1 mRNA and protein expression in paraffin-embedded rat liver sections demonstrated the full potential of combining two methods, ELF in situ hybridization and immunocytochemistry. The immunoreactivity is delineated by red fluorescence, whereas hybridization is depicted by the yellow-green precipitate. ELF in situ hybridization incorporating the rat biotinylated CYP1A1 20-mer on liver sections isolated from a β-NF treated rat exhibited focal ELF precipitate over hepatocytes. Immediately following the ELF in situ procedure, immunocytochemistry using a specific CYP1A1 antibody and a rhodamine-conjugated secondary antibody on the same rat liver section yielded a diffuse, fluorescent rhodamine signal within the cytoplasm of cells. Together, these two methods produced a high resolution, triple-label image of hepatic nuclei, CYP1A1 protein and mRNA (Figure 4).

FIG. 4.

FIG. 4

Open in a new tab

Combined enzyme-labeled fluorescence (ELF) in situ hybridization and irnmunocytochemistry on a formalin-fixed, paraffin-embedded rat liver tissue section. Nuclei are visualized with the blue fluorescent nucleic acid stain, Hoechst 33342. Immunocytochemical localization of the CYPlAl protein is delineated by the red fluorescent stain, while the distribution of CYPlAl mRNA is depicted by the yellow-green, fluorescent signal. Immunocytochemistry was performed using a rabbit polyclonal primary CYPlAl antibody and a rhodamine-conjugated anti-rabbit IgG secondary antibody. Subsequently, on the same rat liver section, ELF in situ hybridization was performed using a rat-specific biotin-labeled CYPlAl 20 base-oligomer probe. Magnification, 400×.

DISCUSSION

In the current study, we optimized a fluorescent technique enabling in situ localization of specific cytochrome P450 mRNAs using biotin-labeled oligonucleotide probes coupled with the ELF alkaline phosphatase substrate, 2-(5′-chloro-2′-phosphoryloxyphenyl)-6-chloro-4-[3H]-quinazoli-none, detection system. The alkaline phosphatase can steadily precipitate the ELF substrate over time, so that the fluorescent signal is continually amplified, resulting in a highly sensitive detection system (Larison et al., 1995). With the ELF protocol, nonisotopic detection and amplification in cells or tissue sections yield a brilliant, yellow-green fluorescent signal, with the developing time taking only minutes. As a result, the entire ELF process can be performed in less than 2 days and without any associated radioactive hazard. In addition to a brighter signal and excellent spatial resolution, background staining is negligible. The highly fluorescent and photostable signal produced by the cleavage of the ELF substrate enhances the simplicity of using the ELF in situ hybridization protocol to detect specific P450s.

We provide evidence that the ELF in situ assay, using 20-base oligomers and optimized hybridization conditions for the different tissue and cell lines, can achieve the sensitivity and specificity required for detecting P4501A1 mRNA in both hepatic and extrahepatic cell types.

In terms of sensitivity, the ELF in situ hybridization technique has the ability to determine induction levels of mRNA in cultured cells where the level of enzymes are either too low or too focally distributed to be easily characterized by other techniques (Foster et al., 1986). Our data, when using a rat-specific CYP1A1 biotinylated 20-mer, demonstrate the ability of cultured Fao and H4IIE cells to respond to xenobiotic challenge, specifically by exhibiting marked increases in CYP1A1 mRNA expression upon β-NF exposure (Figures 1B and 1E). It is noteworthy that the low levels of constitutive CYP1A1 expression also were detected in uninduced cultures of both H4IIE and Fao cells (Figures 1A and 1D) with the technique, demonstrating the exquisite sensitivity of the ELF in situ hybridization procedure. Comparative analyses with more standard procedures have detected increased levels of mRNA only after exposure to prototypic inducers. For example, Xu and Bresnick (1990) conducted Northern blot analyses of CYP1A1 induction in cultured rat hepatoma H4IIE cells treated with either benzo[a]pyrene, 3-methylcholanthrene, or 2,3,7,8-tetrachlorodibenzofuran. In the absence of inducer treatment, they could not detect a positive signal. Similarly, using Northern blot techniques, Raha et al. (1991) detected induced, but not constitutive expression of CYP1A1 mRNA in H4IIE cells. The comparative results from these studies further support the sensitivity of our ELF assay for detection of induced levels of P450s and of even the extremely low levels of mRNA expression found in uninduced cells.

Results from control experiments demonstrated the specificity of the ELF in situ technique. Exclusion of the biotinylated probe, in a mock hybridization procedure, produced no detectable ELF signal. These results emphasized that other components of the ELF in situ hybridization protocol did not yield a fluorescent precipitate and verified the absence of interfering endogenous enzyme activities (Figures 1C, 1F, 2F, and 3F) (Larsson et al., 1988; Larsson and Houggard, 1990; 1993). Likewise, deletion of the primary antibody from the immunocytochemical procedure led to an absence of fluorescent rhodamine signal (Figure 3C), thus verifying the specificity of the immunoreaction. Cultures that were predigested with 150 μg/ml RNase A (Uhl et al., 1985; Larsson et al., 1988; Larsson and Houggard, 1990; 1993; Wilcox, 1993) yielded no ELF fluorescent signal (Figure 2B), proving that endogenous RNA is necessary for the hybridization reaction.

Specific detection of P450 transcripts is further supported by the substitution of the human probe with the mismatching rat probe (Larsson and Houggard, 1993). Our results demonstrated an absence of ELF signal when using the rat-specific CYP1A1 biotin-labeled probe, which is only partially homologous (65%) with the human CYP1A1 mRNA sequence, on human oral epithelium (Figure 2E). However, when the same rat probe was hybridized to target mRNA sequences in β-NF exposed rat hepatoma cells, an intense, yet diffuse, fluorescent signal was produced (Figures 1B and 2A). This particular control experiment using a mismatching oligomer excluded the presence of endogenous biotin and APase activity, and ensured that the probe was species-specific. Competition experiments were performed using the biotin-conjugated rat CYP1A1 probe along with a 60-fold excess of nonbiotinylated rat CYP1A1 probe. Adding increasing concentrations of unlabeled probe provoked a dose-dependent decrease and disappearance of the signal (data not shown). Fluorescent signal normally evident from the hybridization of the biotinylated probe was markedly reduced by inclusion of 60-fold molar excess of unlabeled oligomer (Figure 2C). The blocking of fluorescent signal by the excess concentration of unlabeled probe also demonstrated the lack of endogenous biotin in these cells (Uhl et al., 1985; Larsson et al., 1988; Larsson and Houggard, 1990).

A major advantage afforded by conducting ISH studies is the ability to precisely localize mRNA expression to individual cells in heterogeneous samples. Furthermore, since cellular integrity is maintained, additional biochemical and morphological characterization of cells are possible. For example, an association between P450 enzyme expression with tissue morphology and cell-specific pathological changes within a tissue might be recognized (Murray and Burke, 1995). These advantages of the in situ detection method are important for characterizing the regulation of gene expression to an extent not previously feasible with a heterogeneous population of cells (Singer et al., 1986; Sidhu et al., 1993; 1994), as is the case with traditional Northern blotting, PCR, or slot-blotting procedures.

Along these lines, CYP1A1 was readily assessed in paraffin-embedded rat liver and normal human oral epithelium tissue sections. The 5-μm sections we used in our experiments provided good resolution for morphological identification of the specific cells expressing mRNA. Our results with rat liver tissue sections isolated from a β-NF-exposed animal demonstrated that the ELF precipitate was diffused over the hepatocytes, yet it was primarily localized in cells surrounding the central vein (data not shown). This heterogeneous distribution of CYP1A1 is consistent with results from other investigators (Baron et al., 1981, 1982; van Sliedregt and van Bezooijen, 1990). Similar paraffin-embedded rat liver and human oral epithelium tissue sections were examined in parallel to evaluate the ability of the ELF in situ technique to localize CYP2E1 mRNA (data not shown).

Hybridization of the human biotinylated CYP1A1 probe to normal human oral epithelium tissue sections demonstrated a cell-specific localization of the ELF signal in the upper, more differentiated epithelial layer. This signal was not apparent in the basal and parabasal regions of the undifferentiated epithelium (Figures 3D and 3E). Previously, the lack of appropriate test systems have limited our understanding of oral cancer etiology (i.e., with respect to the specific cell types involved in the bioactivation events that appear to be associated with the disease process, in particular xenobiotic compounds derived from smoking tobacco and alcohol consumption). Use of the ELF in situ hybridization technique should further elucidate the distribution of P450 mRNA in single cells within an organ, as well as the contribution of these enzymes with respect to cell-specific toxicity and cancer.

We performed immunocytochemistry and the ELF in situ hybridization technique in combination in order to simultaneously localize CYP1A1 protein and mRNA. The nonradioactive detection of RNA using the ELF substrate along with a rhodamine-labeled secondary antibody to detect the CYP1A1 protein demonstrated the versatility of using the ELF in situ technique for multicolor applications. By simply alternating the fluorescent filters that excite and emit at specific wavelengths, we were able to obtain double or triple exposures of the same microscopic field. By combining both techniques, a high-resolution, triple-label image of nuclei, protein, and mRNA, with minimal background was achieved. A heterogeneity of CYP1A1 expression was apparent among some hepatocytes, indicated by a higher intensity of ELF and rhodamine signals over cells near the central vein as well as by some cells exhibiting mainly one signal greater than the other (data not shown). In Figure 4, occasional rare cells were either hybridization positive or only immunofluorescent. This observation might be due to differential slicing of the tissue section, with the possible consequence of some cells being punctured. However, our results overall, with rat liver tissue sections, showed that most reactive cells show co-localization of both CYP1A1 mRNA and CYP1A1 protein.

Likewise, when the human oral epithelial sections were assessed immunocytochemically, the immunochemical staining was predominately localized in the upper epithelial layer of differentiated cells, with minimal staining observed in the lower, and less differentiated epithelial layer (Figures 3A and 3B). These results were consistent with those obtained using the ELF in situ hybridization technique.

Farin et al. (1995) demonstrated gene expression of CYP1A1 and CYP2E1, among other P450s, in oral epithelial-derived cell lines. They also detected the presence of these enzymes in other squamous epithelial cell cultures, such as foreskin- and cervical-derived cell lines. Their results further emphasize that, although the liver is the major site of xenobiotic biotransformation, extrahepatic tissues can also catalyze the formation of reactive intermediates, that may ultimately lead to tissue-specific toxicity or cancer.

The continued development of nonradioactive in situ hybridization methodologies, when used in conjunction with discriminatory oligomer probes, will allow a detailed analysis of P450 expression profiles in different tissues and cell lines under a variety of chemical exposures. The use of such procedures, specifically ELF in situ hybridization, should further elucidate the contribution of cytochrome P450 enzymes from specific cells within an organ and may prove valuable in improving predictions of target organ toxicities and cancers.

ACKNOWLEDGMENTS

This work was supported by NIEH Grants ES-04696 and ES-07033. CJO is a recipient of the Burroughs Wellcome Fund Scholar Award in Toxicology.

REFERENCES

  1. ARAI H, EMSON PC, AGRAWAL S, CHRISTODONLON C, GAIT MJ. In situ hybridization histochemistry: localization of vasopressin mRNA in rat brain using a biotinylated oligonucleotide probe. Mol. Brain Res. 1988;4:63–69. doi: 10.1016/0169-328x(88)90019-8. [DOI] [PubMed] [Google Scholar]
  2. BARON J, REDICK JA, GUENGERICH FP. An immunohistochemical study on the localizations and distributions of phenobarbital- and 3-methylcholanthrene-inducible cytochromes P450 within the livers of untreated rats. J. Biol. Chem. 1981;256:5931–5937. [PubMed] [Google Scholar]
  3. BARON J, REDICK JA, GUENGERICH FP. Effects of 3-methylcholanthrene, beta-naphthoflavone, and phenobarbital on the 3-methylcholanthrene-inducible isozyme of cytochrome P-450 within centrilobular, midzonal, and periportal hepatocytes. J. Biol. Chem. 1982;257:953–957. [PubMed] [Google Scholar]
  4. BURSZTAJN S, BERMAN SA, GILBERT W. Simultaneous visualization of neuronal protein and receptor mRNA. Biotechniques. 1990;9:440–449. [PubMed] [Google Scholar]
  5. FARIN FM, OMIECINSKI CJ. Regiospecific expression of cytochrome P450s microsomal epoxide hydrolase in human brain tissue. J. Toxicol. Environ. Health. 1993;40:323–341. doi: 10.1080/15287399309531797. [DOI] [PubMed] [Google Scholar]
  6. FARIN FM, BIGLER LG, ODA D, MCDOUGALL JK, OMIECINSKI CJ. Expression of cytochrome P450 and microsomal epoxide hydrolase in cervical and oral epithelial cells immortalized by human papillomavirus type 16 E6/E7 genes. Carcinogenesis. 1995;16:1391–1401. doi: 10.1093/carcin/16.6.1391. [DOI] [PubMed] [Google Scholar]
  7. FOSTER JR, ELCOMBE CR, BOOBIS AR, et al. Immunocytochemical localization of cytochrome P450 in hepatic and extra-hepatic tissues of the rat with a monoclonal antibody against cytochrome P450c. Biochem. Pharmacol. 1986;35:4543–4554. doi: 10.1016/0006-2952(86)90777-x. [DOI] [PubMed] [Google Scholar]
  8. GRAM TE, OKINE LK, GRAM RA. The metabolism of xenobiotics by certain extra-hepatic organs and its relation to toxicity. Annu. Rev. Pharmacol. Toxicol. 1986;26:259–291. doi: 10.1146/annurev.pa.26.040186.001355. [DOI] [PubMed] [Google Scholar]
  9. HASSETT C, LUCHTEL DL, OMIECINSKI CJ. Hepatic expression of rat P450 mRNA assessed by in situ hybridization to oligomer probes. DNA. 1989;8:29–37. doi: 10.1089/dna.1.1989.8.29. [DOI] [PubMed] [Google Scholar]
  10. HERRINGTON CS, BURNS J, GRAHAM AK, EVANS M, MCGEE JO. Interphase cytogenetics using biotin and digoxigenin labeled probes I: relative sensitivity of both reporter molecules for detection of HPV 16 in CaSki cells. Clin. Pathol. 1989;42:592–600. doi: 10.1136/jcp.42.6.592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. HOLM R, KARLSEN F, NESLAND JM. In situ hybridization with nonisotopic probes using different detection systems. Modern Pathology. 1992;5:315–319. [PubMed] [Google Scholar]
  12. JAISWAL AK, GONZALEZ FJ, NEBERT DW. Human P1-450 gene sequence and correlation of mRNA with genetic differences in benzo[a]pyrene metabolism. Nucleic Acids Res. 1985;13:4503–4520. doi: 10.1093/nar/13.12.4503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. LARISON KD, BREMILLER R, WELLS KS, CLEMENTS I, HAUGLAND RP. Use of a new fluorogenic phosphatase substrate in immunohistochemical applications. J. Histo. and Cytochem. 1995;43:77–83. doi: 10.1177/43.1.7822768. [DOI] [PubMed] [Google Scholar]
  14. LARSSON LI, CHRISTENSEN T, DALBOGE H. Detection of proopiomelanocortin rnRNA by in situ hybridization, using a biotinylated oligodeoxynucleotide probe and avidin-alkaline phosphatase histochemistry. Histochemistry. 1988;89:109–116. doi: 10.1007/BF00489913. [DOI] [PubMed] [Google Scholar]
  15. LARSSON LI, HOUGAARD DM. Optimization of non-radioactive in situ hybridization: image analysis of varying pretreatment, hybridization and probe labelling conditions. Histochemistry. 1990;93:347–354. doi: 10.1007/BF00315849. [DOI] [PubMed] [Google Scholar]
  16. LARSSON LI, HOUGAARD DM. Sensitive detection of rat gastrin mRNA by in situ hybridization with chemically biotinylated oligodeoxynucleotides: validation, quantitation, and double-staining studies. J. Histol. and Cytochem. 1993;41:157–163. doi: 10.1177/41.2.8419457. [DOI] [PubMed] [Google Scholar]
  17. MURRAY GI, BURKE MD. Cornmentary: Immunohistochemistry of drug-metabolizing enzymes. Biochem. Pharmacol. 1995;50:895–903. doi: 10.1016/0006-2952(95)00136-n. [DOI] [PubMed] [Google Scholar]
  18. NEDERLOF PM, ROBINSON D, ABUKNESHA R, et al. Three-color fluorescence in situ hybridization for the simultaneous detection of multiple nucleic acid sequences. Cytometry. 1989;10:20–27. doi: 10.1002/cyto.990100105. [DOI] [PubMed] [Google Scholar]
  19. RAHA A, REDDY V, XU LC, HOUSER WH, BRESNICK E. Presence of the 4 S polycyclic hydrocarbon-binding protein in H4IIE cells. Toxicology. 1991;66:175–186. doi: 10.1016/0300-483x(91)90217-o. [DOI] [PubMed] [Google Scholar]
  20. SIDHU JS, KAVANAGH TJ, REILLY MT, OMIECINSKI CJ. Direct determination of functional activity of cytochrome P4501A1 and NADPH DT-diaphorase in hepatoma cell lines using noninvasive scanning laser cytometry. J. Toxicol. Environ. Health. 1993;40:177–194. doi: 10.1080/15287399309531786. [DOI] [PubMed] [Google Scholar]
  21. SIDHU JS, FARIN FM, KAVANAGH TJ, OMIECINSKI CJ. Effect of tissue-culture substratum and extracellular matrix (ECM) overlay on liver-selective and xenobiotic inducible gene expression in primary rat hepatocytes. In Vitro Toxicol. 1994;7:225–242. [PMC free article] [PubMed] [Google Scholar]
  22. SINGER RH, LAWRENCE JB, VILLNAVE C. Optimization of in situ hybridization using isotopic and non-isotopic detection methods. Biotechniques. 1986;4:230–250. [Google Scholar]
  23. SOGAWA K, GOTOH O, KAWAJRI K, FUJI-KURIYAMA Y. Distinct organization of methylcholanthrene- and phenobarbital-inducible cytochrome P450 genes in the rat. Proc. Natl. Acad. Sci. USA. 1984;81:5066–5070. doi: 10.1073/pnas.81.16.5066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. SPEEL EJM, SCHUTTE B, WIEGANT J, RAMAEKERS FCS, HOPMAN AHN. A novel fluorescence detection method for in situ hybridization, based on the alkaline phosphatasefast red reaction. J. Histol. Cytochem. 1992;40:1299–1308. doi: 10.1177/40.9.1506667. [DOI] [PubMed] [Google Scholar]
  25. SYRJANEN S, PARTANEN P, MANTYJARVI R, SYRJANEN K. Sensitivity of in situ hybridization techniques using biotin- and 35S-labeled human papillomavirus (HPV) DNA probes. J. Virol. Methods. 1988;19:225–238. doi: 10.1016/0166-0934(88)90017-1. [DOI] [PubMed] [Google Scholar]
  26. UHL GR, ZINGG HH, HABENER JF. Vasopressin mRNA in situ hybridization: localization and regulation studied with oligonucleotide cDNA probes in normal and Brattleboro rat hypothalamus. Proc. Natl. Acad. Sci. USA. 1985;82:5555–5559. doi: 10.1073/pnas.82.16.5555. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. VAN SLIEDREGT A, VAN BEZOOIJEN CFA. Effect of different doses of 3-methylcholanthrene on the localization of the 3-methylcholanthrene-inducible isoenzymes of cytochrome P450 within the centrilobular and periportal zones of the rat liver. Biochem. Pharmacol. 1990;39:1703–1708. doi: 10.1016/0006-2952(90)90114-z. [DOI] [PubMed] [Google Scholar]
  28. WILCOX JN. Fundamental principles of in situ hybridization. J. Histochem. Cytochem. 1993;41:1725–1733. doi: 10.1177/41.12.8245419. [DOI] [PubMed] [Google Scholar]
  29. XU LC, BRESNICK E. Induction of cy tochrome P450IA1 in rat hepatoma cell by polycyclic hydrocarbons and a dioxin. Biochem. Phar macol. 1990;40:1399–1403. doi: 10.1016/0006-2952(90)90409-e. [DOI] [PubMed] [Google Scholar]

RESOURCES