Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Variation in expression of genes used for normalization of Northern blots after induction of cell proliferation

Variation in expression of genes used for normalization of Northern blots after induction of cell... Chemical Industry Institute of Toxicology, PO Box 12137, Research Triangle Park, North Carolina, USA (Received 1 1 January 1993; revision accepted 22 March 1993) Abstract. Quantitative knowledge of gene expression can provide valuable information for understanding the action of chemicals that alter cell proliferation and cancer. Accurate quantification of mRNA levels requires the normalization of the gene of interest to a gene with transcriptional levels that do not vary through the cell cycle or with a particular treatment. Changes in expression were examined in proliferating or non-proliferating rat liver for three constitutively expressed ‘housekeeping’ genes commonly used to normalize mRNA levels from Northern blots. In addition, a direct method of quantifying poly(A)+ mRNA by hybridization with a radiolabelled polythymidylate-poly(T)-probe was compared with traditional methods. Hepatocyte cytolethality and a subsequent wave of hepatocyte proliferation were induced in male Fischer-344 rats by treatment with a single gavage dose of carbon tetrachloride. Induced cell proliferation peaked at 48 h after treatment. Expression of the housekeeping genes (GAPDH) and albumin, as well as actin, glyceraldehyde-3-phosphate-dehydrogenase the proto-oncogene H-ras, was determined by Northern blot analysis at times from 0.5 h to 4 days after treatment. Time-dependent changes were observed in the expression of these genes relative to the levels observed in the untreated control animals. Actin expression peaked at 3.4-fold over control and GAPDH expression was increased by 1.9-fold over control. Albumin mRNA levels varied the least, 1.4-fold over control, indicating that this gene is more appropriate than actin or GAPDH for normalization of proto-oncogene expression under experimental conditions that induce cell proliferation in rat liver. The direct quantification of poly(A)+ mRNA using a poly(T) probe was not influenced by the induction of cell proliferation. This method may be useful when the expression of housekeeping genes is affected by treatment. Chemical carcinogens act by both genotoxic and nongenotoxic mechanisms. Carcinogenic activity of nongenotoxic carcinogens is often associated with the sustained induction of cytolethality and cell proliferation (Butterworth & Goldsworthy 1991). The altered expression of oncogenes and other genes associated with growth control has provided insights into the mechanisms involved in chemical carcinogenesis (Braun et al. 1988, Bailey et al. 1989, Mead & Fausto 1989). Detection of mRNA levels by Northern blot analysis is used to evaluate changes in transcription following chemical exposure. Although numerous studies have reported increases Correspondence: M S. M. Goldsworthy, Pathology Associates Inc., 491 5 D Prospectus Dr., Durham, NC 2771 3, s USA. S. M. Goldsworthy et al. in specific gene expression following chemical treatment (Fausto & Shank 1983, Goyette et al. 1984, Hsieh, Peraino & Weinstein 1988), the quantification of this endpoint remains difficult. Using optical density values to determine the amount of mRNA present in a sample is unreliable because of varying amounts of contaminating ribosomal RNA in preparations enriched for poly(A)+ mRNA. Also, variations in gel loading and transfer are not accounted for by this measurement. To overcome these problems, gene expression data are often normalized to a constitutively expressed housekeeping gene. Three genes traditionally used for normalization are actin (Cleveland et al. 1980), glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) (Fort el ul. 1985), and albumin (Sargent, Yang & Bonner 1981). In order to interpret an alteration of expression in the gene of interest correctly, it is important that the gene used for normalization does not vary with cell cycle or chemical treatment. Current studies in this laboratory involve the study of proto-oncogene expression in the target tissues of rodents following treatment with carcinogenic chemicals that cause cell proliferation (Butterworth et al. 1992). Because some proto-oncogenes are expressed only at very low levels, it was necessary to enrich for poly(A)+ mRNA. To determine the best method for quantification of gene expression data in poly(A)+ mRNA samples, we compared the expression of actin, GAPDH and albumin in the livers ofrats after treatment with carbon tetrachloride. Further, the expression of H-ras, a gene known to increase after carbon tetrachloride treatment (Goyette et ul. 1983) was normalized to each of the housekeeping genes. We also investigated an alternative method in which mRNA was quantitated by hybridization of a 35Slabelled polythymidylate-poly(T)probe to the poly(A)+ tails of mRNA samples on a dot blot (Hollander & Fornace 1990). MATERIALS A N D METHODS Animals and treatment This study was approved by the CIIT Institutional Animal Care and Use Committee and was conducted under NIH guidelines. Twelve-week-old male F344 rats were purchased from Charles River Breeding Laboratories Inc. (Raleigh, NC, USA). After a 14-day quarantine period, animals were determined to be virus free by standard murine antibody determination tests (Microbiological Associates, Bethesda, MD, USA). Animals were randomized by body weight and housed five per cage in polycarbonate shoe box cages with filter tops. Conditions were 22 f 1°C and 50 k 10'%1 humidity with a 12 h light/dark cycle. Control and treated animals received NIH-07 diet and water ad libitum. Five rats per timepoint were treated by gavage with a single dose of 2 g/kg body weight of carbon tetrachloride (Aldrich Chemical Co., Milwaukee, WI, USA). All animals except those killed at the 0.5, 1 and 2 h timepoints were injected with tritiated thymidine (7.4 x lo7 Bq/kg) 2 h before sacrifice. Animals were anaesthetized with isofluorane and killed by exsanguination at 0.5, 1,2,6, 12, 18,24,36 h, 2 and 4 days after treatment. Control animals were given a single dose of corn oil by gavage (2 mYkg body weight) and killed 24 h later. Livers were removed, weighed, quickly frozen in liquid nitrogen, and stored at -70°C. Histoautoradiography Unstained deparaffinized liver sections were dipped in Kodak autoradiographic emulsion type NTB2 (Eastman Kodak Co., Rochester, NY, USA) at 42°C and dried overnight at room temperature, then exposed in light-tight desiccator boxes at -20°C. Exposures ranged from 9 to 16 weeks. Test slides were developed periodically to determine optimal exposure time for each batch of slides. After exposure, slides were incubated for 4 min at 13°C in Kodak D-19 Gene expression and cell proliferation developer, rinsed, then fixed for 3 min. Slides were rinsed in gently running water at 13°C for 20min, stained with haematoxylin and eosin, air dried and cover slipped (Eldridge et al. 1990). Scoring of labelled nuclei Computer-generated random fields were used for scoring [3H]dT-labelled hepatocyte nuclei by light microscopy. At least 2000 nuclei from the left lobe were counted for each liver. The labelling index (LI) was calculated by dividing the number of labelled hepatocyte nuclei by the total number of hepatocyte nuclei counted, and the result was expressed as a percentage. Cells judged to be in S phase were easily identified as nuclei containing at least 10 silver grains. A section of duodenum was included on each slide as a positive control for systemic delivery of the DNA precursor label. Values given are mean (tSEM, animal-to-animal variation). mRNA isolation Total cellular RNA was isolated from rat liver (n=3) by the method of Chomczynski & Sacchi (1987) with slight modifications (Puissant & Houdebine 1990).Approximately 0.5 g of frozen liver was homogenized with an SDT Tissumizer (Tekmar, Cincinnati, OH, USA) in 5 ml of a 4 M guanidine thiocyanate, 25 mM sodium citrate, 0.5'%) sarcosyl, 0.1 M 2-mercaptoethanol solution. RNA was extracted with 5 ml water-saturated phenol, 1 ml chloroform : isoamyl alcohol (49 : l ) , and 0.5 ml 2 M sodium acetate (pH 4) for 25 min at 4°C. The aqueous and organic phases were separated by centrifugation at 10000 g and the RNA was precipitated from the upper aqueous phase with 1 volume of isopropanol at -20°C. This precipitate was recovered by centrifugation at 6000 g, resuspended in 4 M lithium chloride and repelleted at 1000 g, then resuspended in 2 ml lOmM TRIS (pH 7.3, 0.5% SDS, 1 mM EDTA buffer, and extracted with 1 volume of chloroform. The resulting aqueous phase was precipitated with 1 volume isopropanol at -20°C overnight. The total cellular RNA was pelleted at 12000 g and resuspended in 0.5 ml diethylpyrocarbonate (DEPC) treated water. To purify poly(A)+ mRNA, the RNA sample was made0.5 M in KCl and lOmM in TRIS (pH 7.5) and loaded on a 0.25 g oligo(dT) cellulose column (Pharmacia, Piscataway, NJ, USA) equilibratedwiththesamebuffer.Unbound RNA waselutedwith l 0 0 m KCI, 1 0 m ~ T R 1 S ( p H ~ 7.5) until the A,,, of the effluent was less than 0.05. Bound poly(A)+ mRNA was eluted with 2 ml of 40°C 10 mM TRIS (pH 7.5) in 500 pl aliquots. Samples were precipitated overnight with 0.1 volume 3 M sodium acetate (pH 5.3) and 2.5 volumes of ethanol, recovered by centrifugation at 12000 g for 30min, then resuspended in 0.5 ml DEPC-treated water (Sambrook, Fritsch & Maniatis 1989). Concentrations of mRNA were determined spectrophotometrically by measuring absorbance at 260 nm, and the mRNA was reprecipitated with ethanol and sodium acetate overnight. The poly(A)+ mRNA was recovered by centrifugation at 4°C for 20 min, dried under vacuum, and resuspended at a concentration of 1 pg/,d based on the absorbance at 260 nm. Northern blot analysis For Northern blot analysis, 5 pg of poly(A)+ mRNA was denatured at 65°C in 21 pI 2 . 2 ~ formaldehyde, 50% formamide, 1 x formaldehyde gel running buffer (5 Prime 3 Prime, West Chester, PA, USA). The mRNA was electrophoresed through a 1%) agarose, 2.2 M formaldehyde gel at 25 V for 18 h. mRNA was transferred to a Hybond-N membrane (Amersham, Arlington Heights, IL, USA) by capillary action using 20 x sodium-sodium citrate buffer (SSC) (Oncor Inc., Gaithersberg, MD. USA) and bound to the membrane by exposure to 1200 J/m2 UV in a Stratalinker (Stratagene, La Jolla, CA, USA). DNA probes were labelled with 32P dCTP with a specific activity of 1.1 1 x l o i 4Bq/mM S. M. Goldsworthy et al. (Amersham, Arlington Heights, IL, USA) to a specific activity of greater than 1 x 108c.p.m./plby nick translation (Rigby et al. 1977). H-ras (rat, 0.730 kb Sstl-Pstl fragment) and actin (chicken, 0.770 kb Hincll-Taql fragment) probes were obtained from Oncor (Gaithersberg, MD. USA), GAPDH (chicken, 1.4 kb Pstl fragment) was a gift from Robert Schwartz (Baylor College, TX, USA), and serum albumin (rat, 1.0 kb Pstl fragment) was a gift from Larry Kier (Monsanto Corp., St Louis, MO, USA). All prehybridizations and hybridizations were performed in a Robbins hybridization incubator (Robbins ScientificCorp., Sunnyvale, CA, USA). In this system, membranes were rolled up and placed, RNA-side in, in screw-topped, 5 cm diameter glass tubes. The tubes were placed in a horizontal rotor assembly that assures uniform distribution of the hybridization fluid. Filters were prehybridized in the incubator for 4 h at the hybridization temperature in 3 ml Hybrisol I (Oncor, Gaithersberg, MD, USA) with 0.02 mg/ml denatured herring sperm DNA. Denatured labelled probe (500 000 Cerenkov counts) was added directly to the prehybridization fluid, and the membrane was incubated for 2 4 6 4 h. Hybridization temperatures were 42°C for albumin and 45°C for H-ras, GAPDH and actin. Hybridization fluid was discarded and the membranes rinsed once in the glass tubes with 5 x SSC. Blots were then washed at the hybridization temperature in 50 ml SSC : 0.1% SDS for 30 min, then in 50 mlO.1 SSC : 0.1% SDS for lC15 min. All washes were performed in the hybridization incubator. Membranes were sealed in bags and exposed at -70°C using XAR2 film (Eastman Kodak Co., Rochester, NY, USA) and Du Pont Cronex intensifying screens (Du Pont, Wilmington, DE, USA). To avoid any variation introduced by stripping and re-probing of blots, identical blots run in parallel were probed for the housekeeping or H-rus genes and the optical density of the bands was quantitated. Quantification of the band intensity in units of absorbance times area (AU x mm) was performed on a LKB Ultroscan XL Laser Densitometer (Pharmacia LKB Biotechnology, Uppsala, Sweden) according to the manufacturer’s recommendations. Data were transferred to a IBM PS/2 and analysed using GelScan XL software (Pharmacia LKB Biotechnology, Uppsala, Sweden). The H-ras value was then divided by the corresponding value for actin, GAPDH, albumin or poly(T) to determine the expression relative to each gene. The mRNA level in the 24 h corn oil control animals was considered to be the control value. Other gene expression studies in this laboratory have employed corn oil controls from different time points and indicate little, if any, change in baseline expression values in control animals with time. Each data set was then normalized to the control value (set equal to 1) and expressed as fold increase over control. Poly(T) probe synthesis and hybridization Synthesis of and hybridization to the poly(T) probe were performed using the method of Hollander & Fornace (1990). Briefly, 1 pg of the mRNA sample was diluted in 99 p1 of 50 m M NaH,PO, (pH 6.8),denatured at 65°C for 5 min and placed on ice. The sample was then pipetted directly onto a Hybond-N membrane in a slot-blot manifold under vacuum. Membranes were crosslinked in a Stratalinker. Each blot included a standard curve ranging from 0.5 to 1.5 p g control mRNA to assess signal linearity. The poly(Tj probe was synthesized using AMV reverse transcriptase (BioRad, Richmond, CA, USA) in the presence of 35S-TTPwith a poly(Aj RNA template. Hybridizations were done at 42°C in a Robbins hybridization oven, as described above, using 1 x lo6 c.p.m./ml 35S-poly(T) probe and 1pg/ml unlabelled poly(T). For quantification, each slot was cut out and counted in a 14C/35S window in 5 ml of Ecolume scintillation fluor. The results werenormalized to the value for liver from control animals. Gene expression and cell prolijeration 1412 X Q 10 - c .- .E 8 rn 5 n 64- m 1 *k km € Figure 1. Induced regenerative cell proliferation following treatment with carbon tetrachloride. Male F-344rats (n=5) were treated with a single dose of 2 g/kg of carbon tetrachloride and killed at various times after treatment. Animals were injected with 3Hthymidine 2 h before sacrifice and hepatocytes in S-phase visualized and quantitated by autoradiographic analysis of histological sections of the left liver lobe. Labelling index (LI) was calculated as the per- RESULTS Male Fischer rats were treated with a single dose of carbon tetrachloride. No animals were lost due to this treatment and no increases in body weights or liver weights were observed. The time course for regenerative cell proliferation and the expression of three housekeeping genes were examined. The expression of the proto-oncogene H-ras is known to increase with induced cell proliferation, and the time course for changes in H-ras mRNA levels was also determined. The effect of normalizing H-ras mRNA values to each of the different housekeeping genes was then contrasted. The wave of regenerative cell proliferation following treatment with carbon tetrachloride is illustrated in Figure 1. An increase in labelled nuclei (LI) was detected at 2 4 4 8 h, with peak expression (1 1.7% rf: 1.41) a t 48 h after treatment. Figure 2a shows the raw densitometric data for H-ras, actin, GAPDH and albumin. For comparison, the control value was set equal to 1 and each data set expressed relative to control. 4 r 3c Time (ti) Time (h) Figure 2. a Raw densitometric data expressed as fold increase over control for H-ras (H),actin ( O ) ,GAPDH ( A )and albumin (0) from Northern blots of rat liver mRNA after carbon tetrachloride treatment ( n = 3 ) . Control values were set equal to 1 and data expressed as fold increase over control. Note the concurrent increases in H-rus, actin and GAPDH at 36 h. b Effect of normalization of H-ras expression data by GAPDH (A), actin (0) and albumin (0). The concurrent increases in actin and GAPDH (as shown in a) cause the increase in H-ras to be minimized when these genes were used for normalization. S. M. Goldsworthy et al. Increased expression of H-ras was seen beginning at 18 h, peaking at 1.6-fold over the control at 36 h, then returning to control levels by 4 days. Interestingly, both actin and GAPDH also exhibited increased expression following a timecourse similar to H-rus. peaking at 3.4-fold and 1 .Pfold over the zero time point control, respectively. Expression of albumin showed some variability at the early timepoints, but returned to control levels by 12hand varied far less than the other genes. Similar profiles of mRNA expression were observed for H-ras, GAPDH and albumin after treatment of male Fischer rats with WY-14,643or partial hepatectomy, two treatments that also induce hepatocyte cell proliferation (data not shown). Figure 2b illustrates the effect on data analysis when H-rasexpression is normalized to each of the housekeeping genes. When actin or GAPDH was used to normalize H-rcrs expression, the increases in H-rus were cancelled out by the concurrent increase in expression of these genes, leading to the conclusion that there was no apparent increased H-rus expression. In contrast, normalization to albumin revealed a peak of expression of H-ras from 18 to 48 h. Based on the results shown in Figure 2a, albumin was the more appropriate housekeeping gene to use. Figure 3 illustrated the use of a radiolabelled poly(T) probe to normalize proto-oncogene expression data. A standard curve for total amount of poly(A)+ mRNA v. counts on the dot blot is shown in Figure 3a. This illustrates that there is a linear relationship between poly(A)+ mRNA and counts in the range of 500 to 1500 ng of RNA. Figure 3b shows the results of quantification of mRNA samples from the carbon tetrachloride experiment using hybridization to a 35S-labelledpoly(T) probe. The amount of mRNA is consistent among the samples. When the poly(T) probe data in Figure 3b is used to normalize H-rus gene expression data (Figure 3c), the curve is not influenced by changes in other genes and therefore shows the true curve for increased H-ras expression. The curve is similar to the H-ras data normalized to albumin (Figure 2b). 12 W '.Or E: 1086 ' u 4 /. I I 12 24 36 48 60 72 84 Time ( h ) Figure 3. a Representative standard curve for a polythymidylate probe dot blot hybridization showing linearity of 35S-poly(T)signal. b Quantitation of inRNA from carbon tetrachloride treated rat liver as determined by poly(T) hybridization (n=3). Control mRNA is at a concentration of 1 &pl based on the absorbance at 260. Sample concentrations are expressed relative to control. c H-ras expression normalized to mRNA concentrations as shown in b. Time (h) DISCUSSION Quantification of gene expression is required for increased understanding of cell cycle regulation and for defining the changes in gene expression during stimulation of cell proliferation. However, the very nature of these changing patterns of expression during cell growth makes such quantification difficult. Rigorous quantification of gene expression requires normalization to total Gene expression and cell proliferation mRNA or to the amount of mRNA of a gene that is constitutively expressed and unchanging during liver regeneration. The data presented here confirm an increased expression of H-ras associated with regenerative cell proliferation seen following liver injury with carbon tetrachloride (Goyette et al. 1984). These results also warn that use of an inappropriate gene for normalization of proto-oncogene expression may lead to a misinterpretation of gene expression data. For the conditions presented here, in which carbon tetrachloride is used to induce hepatcellular proliferation, albumin appears to be an appropriate housekeeping gene for normalization of gene expression. Changes in the expression of albumin (Panduro et al. 1986, Gluck et al. 1992)and actin (Kier, personal communication, Yin, Davison & Tsang 1992) have been reported with other treatment regimens and rodent strains. These variations emphasize the importance of examining expression of the gene used for normalization under each set of experimental conditions. Poly(T) probe dot blot analysis also proved reliable for quantification of total poly(A)+ mRNA. This method, however, does not correct for variations in gel loading or mRNA transfer, but these types of errors can be minimized by the use of multiple samples. Although the polythymidylate probe method is more labour intensive, it is the method of choice when changes in expression of housekeeping genes prohibit their use for normalization. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cell Proliferation Wiley

Variation in expression of genes used for normalization of Northern blots after induction of cell proliferation

Download PDF
8 pages

Loading next page...
 
/lp/wiley/variation-in-expression-of-genes-used-for-normalization-of-northern-45gvenP4n0

References (21)

Publisher
Wiley
Copyright
1993 Blackwell Science Limited
ISSN
0960-7722
eISSN
1365-2184
DOI
10.1111/j.1365-2184.1993.tb00029.x
Publisher site
See Article on Publisher Site

Abstract

Chemical Industry Institute of Toxicology, PO Box 12137, Research Triangle Park, North Carolina, USA (Received 1 1 January 1993; revision accepted 22 March 1993) Abstract. Quantitative knowledge of gene expression can provide valuable information for understanding the action of chemicals that alter cell proliferation and cancer. Accurate quantification of mRNA levels requires the normalization of the gene of interest to a gene with transcriptional levels that do not vary through the cell cycle or with a particular treatment. Changes in expression were examined in proliferating or non-proliferating rat liver for three constitutively expressed ‘housekeeping’ genes commonly used to normalize mRNA levels from Northern blots. In addition, a direct method of quantifying poly(A)+ mRNA by hybridization with a radiolabelled polythymidylate-poly(T)-probe was compared with traditional methods. Hepatocyte cytolethality and a subsequent wave of hepatocyte proliferation were induced in male Fischer-344 rats by treatment with a single gavage dose of carbon tetrachloride. Induced cell proliferation peaked at 48 h after treatment. Expression of the housekeeping genes (GAPDH) and albumin, as well as actin, glyceraldehyde-3-phosphate-dehydrogenase the proto-oncogene H-ras, was determined by Northern blot analysis at times from 0.5 h to 4 days after treatment. Time-dependent changes were observed in the expression of these genes relative to the levels observed in the untreated control animals. Actin expression peaked at 3.4-fold over control and GAPDH expression was increased by 1.9-fold over control. Albumin mRNA levels varied the least, 1.4-fold over control, indicating that this gene is more appropriate than actin or GAPDH for normalization of proto-oncogene expression under experimental conditions that induce cell proliferation in rat liver. The direct quantification of poly(A)+ mRNA using a poly(T) probe was not influenced by the induction of cell proliferation. This method may be useful when the expression of housekeeping genes is affected by treatment. Chemical carcinogens act by both genotoxic and nongenotoxic mechanisms. Carcinogenic activity of nongenotoxic carcinogens is often associated with the sustained induction of cytolethality and cell proliferation (Butterworth & Goldsworthy 1991). The altered expression of oncogenes and other genes associated with growth control has provided insights into the mechanisms involved in chemical carcinogenesis (Braun et al. 1988, Bailey et al. 1989, Mead & Fausto 1989). Detection of mRNA levels by Northern blot analysis is used to evaluate changes in transcription following chemical exposure. Although numerous studies have reported increases Correspondence: M S. M. Goldsworthy, Pathology Associates Inc., 491 5 D Prospectus Dr., Durham, NC 2771 3, s USA. S. M. Goldsworthy et al. in specific gene expression following chemical treatment (Fausto & Shank 1983, Goyette et al. 1984, Hsieh, Peraino & Weinstein 1988), the quantification of this endpoint remains difficult. Using optical density values to determine the amount of mRNA present in a sample is unreliable because of varying amounts of contaminating ribosomal RNA in preparations enriched for poly(A)+ mRNA. Also, variations in gel loading and transfer are not accounted for by this measurement. To overcome these problems, gene expression data are often normalized to a constitutively expressed housekeeping gene. Three genes traditionally used for normalization are actin (Cleveland et al. 1980), glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) (Fort el ul. 1985), and albumin (Sargent, Yang & Bonner 1981). In order to interpret an alteration of expression in the gene of interest correctly, it is important that the gene used for normalization does not vary with cell cycle or chemical treatment. Current studies in this laboratory involve the study of proto-oncogene expression in the target tissues of rodents following treatment with carcinogenic chemicals that cause cell proliferation (Butterworth et al. 1992). Because some proto-oncogenes are expressed only at very low levels, it was necessary to enrich for poly(A)+ mRNA. To determine the best method for quantification of gene expression data in poly(A)+ mRNA samples, we compared the expression of actin, GAPDH and albumin in the livers ofrats after treatment with carbon tetrachloride. Further, the expression of H-ras, a gene known to increase after carbon tetrachloride treatment (Goyette et ul. 1983) was normalized to each of the housekeeping genes. We also investigated an alternative method in which mRNA was quantitated by hybridization of a 35Slabelled polythymidylate-poly(T)probe to the poly(A)+ tails of mRNA samples on a dot blot (Hollander & Fornace 1990). MATERIALS A N D METHODS Animals and treatment This study was approved by the CIIT Institutional Animal Care and Use Committee and was conducted under NIH guidelines. Twelve-week-old male F344 rats were purchased from Charles River Breeding Laboratories Inc. (Raleigh, NC, USA). After a 14-day quarantine period, animals were determined to be virus free by standard murine antibody determination tests (Microbiological Associates, Bethesda, MD, USA). Animals were randomized by body weight and housed five per cage in polycarbonate shoe box cages with filter tops. Conditions were 22 f 1°C and 50 k 10'%1 humidity with a 12 h light/dark cycle. Control and treated animals received NIH-07 diet and water ad libitum. Five rats per timepoint were treated by gavage with a single dose of 2 g/kg body weight of carbon tetrachloride (Aldrich Chemical Co., Milwaukee, WI, USA). All animals except those killed at the 0.5, 1 and 2 h timepoints were injected with tritiated thymidine (7.4 x lo7 Bq/kg) 2 h before sacrifice. Animals were anaesthetized with isofluorane and killed by exsanguination at 0.5, 1,2,6, 12, 18,24,36 h, 2 and 4 days after treatment. Control animals were given a single dose of corn oil by gavage (2 mYkg body weight) and killed 24 h later. Livers were removed, weighed, quickly frozen in liquid nitrogen, and stored at -70°C. Histoautoradiography Unstained deparaffinized liver sections were dipped in Kodak autoradiographic emulsion type NTB2 (Eastman Kodak Co., Rochester, NY, USA) at 42°C and dried overnight at room temperature, then exposed in light-tight desiccator boxes at -20°C. Exposures ranged from 9 to 16 weeks. Test slides were developed periodically to determine optimal exposure time for each batch of slides. After exposure, slides were incubated for 4 min at 13°C in Kodak D-19 Gene expression and cell proliferation developer, rinsed, then fixed for 3 min. Slides were rinsed in gently running water at 13°C for 20min, stained with haematoxylin and eosin, air dried and cover slipped (Eldridge et al. 1990). Scoring of labelled nuclei Computer-generated random fields were used for scoring [3H]dT-labelled hepatocyte nuclei by light microscopy. At least 2000 nuclei from the left lobe were counted for each liver. The labelling index (LI) was calculated by dividing the number of labelled hepatocyte nuclei by the total number of hepatocyte nuclei counted, and the result was expressed as a percentage. Cells judged to be in S phase were easily identified as nuclei containing at least 10 silver grains. A section of duodenum was included on each slide as a positive control for systemic delivery of the DNA precursor label. Values given are mean (tSEM, animal-to-animal variation). mRNA isolation Total cellular RNA was isolated from rat liver (n=3) by the method of Chomczynski & Sacchi (1987) with slight modifications (Puissant & Houdebine 1990).Approximately 0.5 g of frozen liver was homogenized with an SDT Tissumizer (Tekmar, Cincinnati, OH, USA) in 5 ml of a 4 M guanidine thiocyanate, 25 mM sodium citrate, 0.5'%) sarcosyl, 0.1 M 2-mercaptoethanol solution. RNA was extracted with 5 ml water-saturated phenol, 1 ml chloroform : isoamyl alcohol (49 : l ) , and 0.5 ml 2 M sodium acetate (pH 4) for 25 min at 4°C. The aqueous and organic phases were separated by centrifugation at 10000 g and the RNA was precipitated from the upper aqueous phase with 1 volume of isopropanol at -20°C. This precipitate was recovered by centrifugation at 6000 g, resuspended in 4 M lithium chloride and repelleted at 1000 g, then resuspended in 2 ml lOmM TRIS (pH 7.3, 0.5% SDS, 1 mM EDTA buffer, and extracted with 1 volume of chloroform. The resulting aqueous phase was precipitated with 1 volume isopropanol at -20°C overnight. The total cellular RNA was pelleted at 12000 g and resuspended in 0.5 ml diethylpyrocarbonate (DEPC) treated water. To purify poly(A)+ mRNA, the RNA sample was made0.5 M in KCl and lOmM in TRIS (pH 7.5) and loaded on a 0.25 g oligo(dT) cellulose column (Pharmacia, Piscataway, NJ, USA) equilibratedwiththesamebuffer.Unbound RNA waselutedwith l 0 0 m KCI, 1 0 m ~ T R 1 S ( p H ~ 7.5) until the A,,, of the effluent was less than 0.05. Bound poly(A)+ mRNA was eluted with 2 ml of 40°C 10 mM TRIS (pH 7.5) in 500 pl aliquots. Samples were precipitated overnight with 0.1 volume 3 M sodium acetate (pH 5.3) and 2.5 volumes of ethanol, recovered by centrifugation at 12000 g for 30min, then resuspended in 0.5 ml DEPC-treated water (Sambrook, Fritsch & Maniatis 1989). Concentrations of mRNA were determined spectrophotometrically by measuring absorbance at 260 nm, and the mRNA was reprecipitated with ethanol and sodium acetate overnight. The poly(A)+ mRNA was recovered by centrifugation at 4°C for 20 min, dried under vacuum, and resuspended at a concentration of 1 pg/,d based on the absorbance at 260 nm. Northern blot analysis For Northern blot analysis, 5 pg of poly(A)+ mRNA was denatured at 65°C in 21 pI 2 . 2 ~ formaldehyde, 50% formamide, 1 x formaldehyde gel running buffer (5 Prime 3 Prime, West Chester, PA, USA). The mRNA was electrophoresed through a 1%) agarose, 2.2 M formaldehyde gel at 25 V for 18 h. mRNA was transferred to a Hybond-N membrane (Amersham, Arlington Heights, IL, USA) by capillary action using 20 x sodium-sodium citrate buffer (SSC) (Oncor Inc., Gaithersberg, MD. USA) and bound to the membrane by exposure to 1200 J/m2 UV in a Stratalinker (Stratagene, La Jolla, CA, USA). DNA probes were labelled with 32P dCTP with a specific activity of 1.1 1 x l o i 4Bq/mM S. M. Goldsworthy et al. (Amersham, Arlington Heights, IL, USA) to a specific activity of greater than 1 x 108c.p.m./plby nick translation (Rigby et al. 1977). H-ras (rat, 0.730 kb Sstl-Pstl fragment) and actin (chicken, 0.770 kb Hincll-Taql fragment) probes were obtained from Oncor (Gaithersberg, MD. USA), GAPDH (chicken, 1.4 kb Pstl fragment) was a gift from Robert Schwartz (Baylor College, TX, USA), and serum albumin (rat, 1.0 kb Pstl fragment) was a gift from Larry Kier (Monsanto Corp., St Louis, MO, USA). All prehybridizations and hybridizations were performed in a Robbins hybridization incubator (Robbins ScientificCorp., Sunnyvale, CA, USA). In this system, membranes were rolled up and placed, RNA-side in, in screw-topped, 5 cm diameter glass tubes. The tubes were placed in a horizontal rotor assembly that assures uniform distribution of the hybridization fluid. Filters were prehybridized in the incubator for 4 h at the hybridization temperature in 3 ml Hybrisol I (Oncor, Gaithersberg, MD, USA) with 0.02 mg/ml denatured herring sperm DNA. Denatured labelled probe (500 000 Cerenkov counts) was added directly to the prehybridization fluid, and the membrane was incubated for 2 4 6 4 h. Hybridization temperatures were 42°C for albumin and 45°C for H-ras, GAPDH and actin. Hybridization fluid was discarded and the membranes rinsed once in the glass tubes with 5 x SSC. Blots were then washed at the hybridization temperature in 50 ml SSC : 0.1% SDS for 30 min, then in 50 mlO.1 SSC : 0.1% SDS for lC15 min. All washes were performed in the hybridization incubator. Membranes were sealed in bags and exposed at -70°C using XAR2 film (Eastman Kodak Co., Rochester, NY, USA) and Du Pont Cronex intensifying screens (Du Pont, Wilmington, DE, USA). To avoid any variation introduced by stripping and re-probing of blots, identical blots run in parallel were probed for the housekeeping or H-rus genes and the optical density of the bands was quantitated. Quantification of the band intensity in units of absorbance times area (AU x mm) was performed on a LKB Ultroscan XL Laser Densitometer (Pharmacia LKB Biotechnology, Uppsala, Sweden) according to the manufacturer’s recommendations. Data were transferred to a IBM PS/2 and analysed using GelScan XL software (Pharmacia LKB Biotechnology, Uppsala, Sweden). The H-ras value was then divided by the corresponding value for actin, GAPDH, albumin or poly(T) to determine the expression relative to each gene. The mRNA level in the 24 h corn oil control animals was considered to be the control value. Other gene expression studies in this laboratory have employed corn oil controls from different time points and indicate little, if any, change in baseline expression values in control animals with time. Each data set was then normalized to the control value (set equal to 1) and expressed as fold increase over control. Poly(T) probe synthesis and hybridization Synthesis of and hybridization to the poly(T) probe were performed using the method of Hollander & Fornace (1990). Briefly, 1 pg of the mRNA sample was diluted in 99 p1 of 50 m M NaH,PO, (pH 6.8),denatured at 65°C for 5 min and placed on ice. The sample was then pipetted directly onto a Hybond-N membrane in a slot-blot manifold under vacuum. Membranes were crosslinked in a Stratalinker. Each blot included a standard curve ranging from 0.5 to 1.5 p g control mRNA to assess signal linearity. The poly(Tj probe was synthesized using AMV reverse transcriptase (BioRad, Richmond, CA, USA) in the presence of 35S-TTPwith a poly(Aj RNA template. Hybridizations were done at 42°C in a Robbins hybridization oven, as described above, using 1 x lo6 c.p.m./ml 35S-poly(T) probe and 1pg/ml unlabelled poly(T). For quantification, each slot was cut out and counted in a 14C/35S window in 5 ml of Ecolume scintillation fluor. The results werenormalized to the value for liver from control animals. Gene expression and cell prolijeration 1412 X Q 10 - c .- .E 8 rn 5 n 64- m 1 *k km € Figure 1. Induced regenerative cell proliferation following treatment with carbon tetrachloride. Male F-344rats (n=5) were treated with a single dose of 2 g/kg of carbon tetrachloride and killed at various times after treatment. Animals were injected with 3Hthymidine 2 h before sacrifice and hepatocytes in S-phase visualized and quantitated by autoradiographic analysis of histological sections of the left liver lobe. Labelling index (LI) was calculated as the per- RESULTS Male Fischer rats were treated with a single dose of carbon tetrachloride. No animals were lost due to this treatment and no increases in body weights or liver weights were observed. The time course for regenerative cell proliferation and the expression of three housekeeping genes were examined. The expression of the proto-oncogene H-ras is known to increase with induced cell proliferation, and the time course for changes in H-ras mRNA levels was also determined. The effect of normalizing H-ras mRNA values to each of the different housekeeping genes was then contrasted. The wave of regenerative cell proliferation following treatment with carbon tetrachloride is illustrated in Figure 1. An increase in labelled nuclei (LI) was detected at 2 4 4 8 h, with peak expression (1 1.7% rf: 1.41) a t 48 h after treatment. Figure 2a shows the raw densitometric data for H-ras, actin, GAPDH and albumin. For comparison, the control value was set equal to 1 and each data set expressed relative to control. 4 r 3c Time (ti) Time (h) Figure 2. a Raw densitometric data expressed as fold increase over control for H-ras (H),actin ( O ) ,GAPDH ( A )and albumin (0) from Northern blots of rat liver mRNA after carbon tetrachloride treatment ( n = 3 ) . Control values were set equal to 1 and data expressed as fold increase over control. Note the concurrent increases in H-rus, actin and GAPDH at 36 h. b Effect of normalization of H-ras expression data by GAPDH (A), actin (0) and albumin (0). The concurrent increases in actin and GAPDH (as shown in a) cause the increase in H-ras to be minimized when these genes were used for normalization. S. M. Goldsworthy et al. Increased expression of H-ras was seen beginning at 18 h, peaking at 1.6-fold over the control at 36 h, then returning to control levels by 4 days. Interestingly, both actin and GAPDH also exhibited increased expression following a timecourse similar to H-rus. peaking at 3.4-fold and 1 .Pfold over the zero time point control, respectively. Expression of albumin showed some variability at the early timepoints, but returned to control levels by 12hand varied far less than the other genes. Similar profiles of mRNA expression were observed for H-ras, GAPDH and albumin after treatment of male Fischer rats with WY-14,643or partial hepatectomy, two treatments that also induce hepatocyte cell proliferation (data not shown). Figure 2b illustrates the effect on data analysis when H-rasexpression is normalized to each of the housekeeping genes. When actin or GAPDH was used to normalize H-rcrs expression, the increases in H-rus were cancelled out by the concurrent increase in expression of these genes, leading to the conclusion that there was no apparent increased H-rus expression. In contrast, normalization to albumin revealed a peak of expression of H-ras from 18 to 48 h. Based on the results shown in Figure 2a, albumin was the more appropriate housekeeping gene to use. Figure 3 illustrated the use of a radiolabelled poly(T) probe to normalize proto-oncogene expression data. A standard curve for total amount of poly(A)+ mRNA v. counts on the dot blot is shown in Figure 3a. This illustrates that there is a linear relationship between poly(A)+ mRNA and counts in the range of 500 to 1500 ng of RNA. Figure 3b shows the results of quantification of mRNA samples from the carbon tetrachloride experiment using hybridization to a 35S-labelledpoly(T) probe. The amount of mRNA is consistent among the samples. When the poly(T) probe data in Figure 3b is used to normalize H-rus gene expression data (Figure 3c), the curve is not influenced by changes in other genes and therefore shows the true curve for increased H-ras expression. The curve is similar to the H-ras data normalized to albumin (Figure 2b). 12 W '.Or E: 1086 ' u 4 /. I I 12 24 36 48 60 72 84 Time ( h ) Figure 3. a Representative standard curve for a polythymidylate probe dot blot hybridization showing linearity of 35S-poly(T)signal. b Quantitation of inRNA from carbon tetrachloride treated rat liver as determined by poly(T) hybridization (n=3). Control mRNA is at a concentration of 1 &pl based on the absorbance at 260. Sample concentrations are expressed relative to control. c H-ras expression normalized to mRNA concentrations as shown in b. Time (h) DISCUSSION Quantification of gene expression is required for increased understanding of cell cycle regulation and for defining the changes in gene expression during stimulation of cell proliferation. However, the very nature of these changing patterns of expression during cell growth makes such quantification difficult. Rigorous quantification of gene expression requires normalization to total Gene expression and cell proliferation mRNA or to the amount of mRNA of a gene that is constitutively expressed and unchanging during liver regeneration. The data presented here confirm an increased expression of H-ras associated with regenerative cell proliferation seen following liver injury with carbon tetrachloride (Goyette et al. 1984). These results also warn that use of an inappropriate gene for normalization of proto-oncogene expression may lead to a misinterpretation of gene expression data. For the conditions presented here, in which carbon tetrachloride is used to induce hepatcellular proliferation, albumin appears to be an appropriate housekeeping gene for normalization of gene expression. Changes in the expression of albumin (Panduro et al. 1986, Gluck et al. 1992)and actin (Kier, personal communication, Yin, Davison & Tsang 1992) have been reported with other treatment regimens and rodent strains. These variations emphasize the importance of examining expression of the gene used for normalization under each set of experimental conditions. Poly(T) probe dot blot analysis also proved reliable for quantification of total poly(A)+ mRNA. This method, however, does not correct for variations in gel loading or mRNA transfer, but these types of errors can be minimized by the use of multiple samples. Although the polythymidylate probe method is more labour intensive, it is the method of choice when changes in expression of housekeeping genes prohibit their use for normalization.

Journal

Cell Proliferation – Wiley

Published: Nov 1, 1993

There are no references for this article.