MATERIALS AND METHODS

DNA extraction

DNA was extracted from the two cell lines Colo320 and T47D using the Purgene kit (Gentra Systems, Minnesota, USA), according to the manufacturer’s protocol. In brief, the cell pellet was lysed with 600 μl of Purgene cell lysis solution with the addition of 20 μl proteinase K (20 mg/ml; Gibco BRL, Gaithersburg, Maryland, USA). The samples were digested at 56°C for 16 hours. To obtain RNA free DNA, samples were treated with 3 μl RNase A (4 mg/ml; Gentra Systems), according to the manufacturer’s instruction. Proteins were precipitated by the addition of 200 μl protein precipitation solution. Samples were then vortexed and centrifuged at 16 000 ×g for five minutes. The supernatant was transferred to a 1.5 ml eppendorf tube containing 600 μl isopropanol. The tube containing the mixture was inverted and the samples were centrifuged at 16 000 ×g for five minutes. The DNA pellets obtained were washed with 70% alcohol and vacuum dried. The DNA pellet was dissolved in 70 μl TE buffer (10mM Tris HCl, pH 7.0, 1mM EDTA, pH 8.0). The DNA concentration was measured with an ultraviolet spectroscope. Normal DNA for CGH experiments was extracted from the blood of a healthy donor using the above protocol, except that the red blood cells were lysed with Purgene RBC lysis solution (Gentra Systems). The white blood cells were lysed with Purgene cell lysis solution (Gentra Systems).

Labelling of DNA for CGH

Cell line DNA (2 μg aliquots) was labelled using biotin nick translation reagent (Roche, Hamburg, Germany), according to manufacturer’s instructions. The labelling reaction was carried out at 15°C for 75 minutes and the reaction was stopped by the addition of 2 μl of 0.5M EDTA (pH 8.0). Samples were then heated to 72°C for 10 minutes. Normal DNA was labelled as above with digoxigenin nick translation reagent (Roche, Mannheim, Germany).

Comparative genomic hybridisation

The test and reference labelled DNAs were mixed in equal concentrations (∼ 2.0 μg) and precipitated with 100 μg human Cot1 DNA (Invitrogen). The CGH probe was pelleted by centrifugation, washed with 70% alcohol, and vacuum dried. The probe was dissolved in 16 μl CGH hybridisation buffer (50% dextran sulfate in 2× saline sodium citrate (SSC) containing 50% formamide). The probe was denatured at 75°C for 10 minutes and reannealed at 37°C for one hour. CGH metaphase slides were obtained from Vysis Inc (Illinois, USA) and metaphases were marked under a phase contrast microscope and incubated in denaturation solution (70% deionised formamide in 2× SSC) at 73°C for three minutes. The slides were then dehydrated in a series of alcohol (70%, 80%, and 100%) for three minutes each and left to dry at 37°C for 10 minutes. After applying the probe, slides were incubated at 37°C in a humid environment for 72 hours. The slides were finally washed with 70% formamide in 2× SSC buffer and stained with fluorescent antibodies and diamidinophenylindole (DAPI), as described previously.² Images were captured by CCD camera (JVL, JAI Corporation, Japan) using DAPI, fluorescein isothiocyanate, and Texas red filters mounted on Axioplan 2 imaging system (Carl Zeiss, Hamburg, Germany). The images were then analysed by ISIS^™ software (Metasystems, Hamburg, Germany). Typically, 10 to 20 metaphases were captured using the appropriate filters. The images were digitally stored and the green to red and red to green ratios were calculated. Amplifications were scored if the median green to red ratio was above or equal to the 1.25 threshold, and deletions (only for CGH) were below or equal to the 0.75 threshold. In all experiments, separate normal versus normal DNA hybridisation controls were also included. A genetic aberration was scored if the probability representing the 99% confidence interval of the tumour versus the normal profiles was out with the 99% confidence interval of the normal versus normal control profiles.

RNA extraction

Cell lines at 70% confluency were harvested and then transferred to DEPC treated 1.5 ml eppendorf tubes. The cells were washed twice with phosphate buffered saline and centrifuged at 165 ×g for five minutes at 4°C. The cells were lysed with 500 μl of Trizol (Life Technologies, Hamburg, Germany). Chloroform (500 μl) was then added, and the mixture was vortexed and then centrifuged at 4°C, 6800 ×g for 30 minutes. The supernatant was transferred into a clean DEPC treated 1.5 ml eppendorf tube and an equal volume of isopropanol was added for precipitation. The mixture was kept on ice for 15 minutes and then centrifuged at 4°C, 20 800 ×g rpm for 30 minutes. The RNA pellet was washed with 70% alcohol, vacuum dried, and dissolved in 70 μl of DEPC treated water.

Full length cDNA synthesis and amplification

Total RNA (1.0 μg) was reverse transcribed into cDNA using the primers 5‘- AAGCAGTGGTATCAACGCAGAGTACGCGGG -3‘ and 5‘-AAGCAGTGGTATC AACGCAG AGTACT₍₃₀₎N_-1N-3‘ (Clontech, Palo Alto, California, USA). The samples were first denatured with the above two primers at 70°C for two minutes and mixed with 5× first strand buffer (250mM Tris/HCl, 365mM KCl, 30mM MgCl₂), dithiothreitol (20mM), 50× dNTP, and PowerScript reverse transcriptase (Clontech). The reaction mixture was incubated at 42°C for one hour. The first strand reaction was diluted with 40 μl of TE buffer (10mM Tris, 1mM EDTA). The reaction was stopped by heating at 72°C for seven minutes. cDNA amplification and labelling were performed in the next step by Advantage 2 PCR (Clontech). A 2 μl aliquot of the cDNA sample was added to the master mix containing 10× Advantage PCR buffer, 50× dNTP (10mM), 0.1mM biotin 16-dUTP (Roche Boehringer), 50× Advantage 2 polymerase mix, and a 10 μM of 5‘ PCR primer (5‘-AAGCAGTGGTATCAACGCAGAGT-3‘; Clontech) in a 100 μl total volume. PCR was performed using initial denaturation at 95°C for one minute, followed by 20 cycles of 95°C for five seconds, 65°C for five seconds, and 68°C for six minutes. A 10 μl aliquot of the PCR product was run on a 1.2% agarose gel, and produced a smear of cDNA ranging from 0.4 to 6 kbp.

Expressive genomic hybridisation

Biotin labelled amplified cDNA from the cell line and the digoxigenin labelled normal reference male DNA were mixed in equal parts and precipitated with 100 μg of human Cot-1 DNA (Invitrogen). The probe was dissolved in CGH hybridisation buffer (50% formamide, 50% dextran sulfate in 2× SSC), denatured, reannealed, and hybridised to denatured normal metaphase CGH slides using the procedure described above. To validate this method, we included two control experiments. In the first control, the labels were reversed: the cell line cDNA was labelled with digoxigenin and the normal DNA was labelled with biotin, so that the same EGH profiles would be expected, but in reverse colours. In the second control, we hybridised the cell line cDNA labelled with biotin with the same cell line cDNA labelled with digoxigenin to ensure maximum specificity of the signals. The slides were processed as described for CGH. The images were digitally stored and the green to red and red to green ratios were calculated. Overexpression was scored if the median green to red ratio was above or equal to the 1.25 threshold and underexpression when the green to red ratio was below or equal to the 0.75 threshold.

RESULTS AND DISCUSSION

CGH detected several genetic aberrations in the Colo320 cell line, including amplification of 1p21–q24, 2p25.3–q22, 3q13.1–q29, 4p16–q26, 6p23–p25, 6q15–q27, 8q23–q24.3, 9q21.3–q34.3, 11p12–15.5, 12p13.3–q15, 12q23–24.2, 13p13–q34, 14p11.2–13, 17q21–25, 18q11.1, and 22p12–q12.3, in addition to an increase in copy number of chromosomes 16 and 20. Deletions detected by CGH in this cell line were at 5p14–15.3, 5p11, 5q33.3–ter, 7p11.1–p13, 8q12, 10p11.2–12.3, 10p13–14, 11q12–13.1, 11q23.1–23.3, 11q24, 17p11.2–12, and 21p13–ter (fig 1). The profile generated is in accordance with previously published data.^14–16 Similarly, the CGH profile of the T47D cell line showed amplification of 1q11–q43, 3p21.3–25, 3q24–29, 5p15.3–q11.3, 5q33.1–35.3, 6p22.1–p25, 7p21.1–q21.1, 8p12–qter, 10q11.2–24.2, 11p15–qter, 12q11.1–qter, 14p13–q23, 15q21.3–qter, 17q11–q21.3, and 21q22.2–qter, in addition to increased copy number of chromosome 20. The areas containing deletions in this cell line were 2p21–pter, 3p11.1–14.1, 3q13.1–13.3, 9p21–p23, 12p12.3–pter, 16p11.2–13.1, 18q12.3–qter, and 19q13.1–13.2 (fig 2). Although CGH analysis on this cell line has not been reported, we chose to use it because its DNA and expression profiles have been previously established using microarrays.^15,17 We then generated cDNA probes from the same cell lines (as described earlier) and applied them to metaphase chromosomes. SMART cDNA synthesis technology has not been used previously to generate global hybridisation probes. DOP-PCR has been used extensively for the amplification of nucleic acids in CGH,^18,19 but because contaminating genomic DNA can be amplified by DOP-PCR, cDNA amplified by DOP-PCR should be treated with DNase before amplification. In contrast, the SMART method requires no prior DNase treatment because one of the primers has a poly T chain that anneals to mRNA only. In addition, SMART cDNA amplification maintains the relative degrees of expression of the different transcripts.¹³

To test the size range of the mRNA amplification procedure, the PCR products were run on a 1.2% agarose gel and showed smeared fragments ranging from 400 to 6000 bp (fig 3). To ensure that DNA is not amplified by this method, an RNA free DNA sample was also amplified and no product was seen, as was expected (fig 3, lane 3). Because the hybridisation of regular cDNA on to metaphase slides produced no detectable signals, we first tested whether the amplified cDNA can generate detectable signals on corresponding chromosomes. To do this, the amplified test samples were hybridised to metaphase slides without reference nucleic acid. Figure 4A shows the highly efficient hybridisation signals seen using this method. To quantitate the expression relative to a reference point, normal genomic reference DNA was co-hybridised with the amplified cDNA (fig 4B). CGH profiles generated from the Colo320 and T47D cell lines were then compared with the results obtained by EGH from the same cell lines (fig 4C). The results of copy number changes/amplifications versus expression profiles for the Colo320 and T47D cell lines are summarised in figs 5 and 6, respectively. The profiles generated can be subclassified into the following categories of chromosome regions: (1) amplified and totally or partially expressed; (2) not amplified but highly expressed; and (3) amplified but not expressed. For the Colo320 cell line, category 1 areas comprised: 1p21–1p13.3, 1q12–1q24, 2p13–p22, 2q12, 3q21–29, 4p14–q26, 6q15–q18, 6q22, 6q23.3–q25, 8q23–24.3, 12p13–p11.2, 13p11.2–q33, 16q11.2–q12.2, and 17q21–q25; category 2 areas comprised: 2q24.1–q33, 5p13–p12, 5q13–q23.3, 7q22–31.3, 12q21.3–q23, 14q11.2–q24, and 17q11.2–q21; and category 3 areas comprised: 1p13.3–q12, 2p25–p22, 2p13–q12, 2q12.3–q22, 3q13.1–q21, 4p16–p14, 6p25–23, 6q17–q22, 6q22.3–23.3, 6q25–q27, 8q24.1–q24.3, 9q21.3–q34.3, 11p15.5–p12, 12p13.3–p13, 12p11.2–q14, 12q23–q24.2, 13p13–p11.2, 13q33–q34, 14p13–q11, 16p13.3–q11.2, 16q13–q24, 20p13–q11.2, and 22p12–q12 (fig 5). Similarly, for the T47D cell line category 1 chromosomal areas comprised: 1q22–q42, 3p24–p22, 3q25–q29, 5p15.3–q11.2, 6p22, 6q24–q25, 7p15–p12, 8p12–p11.2, 8q13–q24.1, 10q21.2, 11p15.1–p11.2, 11q14–q23, 11q24–q25, 12q13–q24.1, 14p12–q23, 15q21–q24, and 15q26; category 2 areas comprised: 5q11.2–q13, 7q21–q22, 12p12–p11.2, 13q14–q21.1, and 21q11–q21.3; and category 3 areas comprised: 1q12–q22, 1q42.1–q43, 3p25–24.1, 3p22–p21.3, 3q24–q25.1, 5q33–q35, 6p25–p22.1, 7p21.1–p15, 7p12–q21.1, 8q11.2–q13, 8q24.1–24.3, 10q11.2–q21, 10q21.3–24.1, 11p15.2–p15.1, 11p11.2–q14, 11q23.3, 11q24–q25, 12q12–q13, 12q24.1–q24.3, 15q24–q25.5, 15q26–qter, 17q11.2–q21.3, 20q11.2–qter, and 21q22–qter (fig 6).

For both cell lines, the expression profiles were confirmed by reverse labelling of the cDNA and the reference DNA (fig 4 B,C). In addition, neither of the cell lines expressed sequences that localise to the Y chromosome because they were developed from female patients (fig 4B, see Y chromosome profile). As expected, none of the deleted areas detected by CGH in both cell line was overexpressed.

“Expressive genomic hybridisation can be used to identify and localise important genes that could have been missed as potential oncogenes by comparative genomic hybridisation”

Our results demonstrate that genetic amplifications closely correlate with gene overexpression, which is in accordance with previously published data. However, the results also show that, in many chromosomes, amplifications do not necessarily result in overexpression. It could be that these areas contain a few genes that are overexpressed but their detection is beyond the limits of this method. Nevertheless, the fact that overexpression was clearly demonstrated in many amplified chromosomal regions but not in others suggests a cellular mechanism of gene expression control. This result is supported by previous work, which found a limited number of genes with low levels of overexpression or even larger number of genes underexpressed within amplicons detected by CGH.^10,16,20 These controversies will remain until the development of single arrays that carry the complete human genome. Until then, microarray based strategies suffer from an important limitation; namely, that they can only provide information on genes that have been previously selected for spotting on arrays. Although other published techniques have been described to circumvent such limitations,²¹ the technique described here offers a method for focusing on and selecting genes based on their chromosomal location to be investigated further by microarrays. In addition, because EGH localises overexpressed transcripts to particular chromosomes without them being identified as amplicons, this method can be used to identify and localise important genes that could have been missed as potential oncogenes by CGH.

EGH allows the development of a global expression signature for a cell line or tissue at the cytogenetic level. Because ultimately expression is what determines the characteristics of individual tumours, we envisage that this technique could be used for the localisation of expression sequences and their relevance to tumour progression. This is important in view of our finding that some amplified sequences are not necessarily overexpressed and that some overexpressed sequences are generated by mechanisms other than genomic amplification.