- Email: [email protected]
Generation of locus-specific probes for interphase fluorescence in situ hybridisation—application in Barrett's esophagus
Experimental and Molecular Pathology 77 (2004) 26 – 33 www.elsevier.com/locate/yexmp
Generation of locus-specific probes for interphase fluorescence in situ hybridisation—application in Barrett’s esophagus S.H. Doak, a,* D. Saidely, a G.J.S. Jenkins, a E.M. Parry, a A.P. Griffiths, b J.N. Baxter, c and J.M. Parry a a
Human Molecular Pathology Group, School of Biological Sciences, University of Wales Swansea, Swansea, SA2 8PP, UK b Department of Pathology, Morriston Hospital, Swansea SA6 6NL, UK c Department of Surgery, Morriston Hospital, Swansea SA6 6NL, UK Received 1 April 2004 Available online 2 June 2004
Abstract Despite the wide range of probes commercially available for interphase fluorescence in situ hybridisation (FISH), the supply of locusspecific probes is limited to genes or chromosomal regions commonly altered in genetic diseases or during carcinogenesis. Generation of these probes is therefore desirable to accommodate individual research requirements. Hence, we detail the methodology required to design and produce custom locus-specific interphase FISH probes for any human genomic region of interest and their application was illustrated in cytogenetic investigations of Barrett’s tumourigenesis. Previously utilising FISH, we observed that Barrett’s tissues demonstrated chromosome 4 hyperploidy [Gut 52 (2003) 623], but as centromeric probes were used in this analysis, it was not known if the whole chromosome was amplified. We consequently generated single-copy sequence probes for the 4p16.3 and 4q35.1 subtelomeric loci. Multicolour FISH was subsequently performed on interphase preparations originating from patients with Barrett’s esophagus at varying histological grades, thus demonstrating the whole region of chromosome 4 was amplified within the tissues. Additionally, probes for the DNA methyltransferase genes were produced to determine if gene dosage alterations were responsible for increasing methylation activity during Barrett’s neoplastic progression. No significant alterations at the DNMT1 and DNMT3a loci were detected. An increased copy number of these genes is therefore not the basis for the hypermethylation commonly observed in this premalignant lesion. D 2004 Elsevier Inc. All rigths reserved. Keywords: Barrett’s esophagus; DNMT1; DNMT3a; Fluorescence in situ hybridisation; Locus-specific probes
Introduction Fluorescence in situ hybridisation (FISH) is a powerful cytogenetic tool for directly localising and visualising unique regions of the genome at the single cell level. It has been widely exploited in many applications ranging from environmental genotoxicity studies to DNA mapping and has proved to be a vital source of genetic information in cancer research, providing data on whole chromosome copy number changes down to deletions or amplifications of single genes (Eastmond et al., 1995; Popescu and Zimonjic, 1997; Werner et al., 1997). Historically, conventional cytogenetic analysis has relied on the ability to culture cells to * Corresponding author. Human Molecular Pathology Group, School of Biological Sciences, University of Wales Swansea, Singleton Park, Swansea, SA2 8PP, UK. Fax: +44-1792-295447. E-mail address: [email protected] (S.H. Doak). 0014-4800/$ - see front matter D 2004 Elsevier Inc. All rigths reserved. doi:10.1016/j.yexmp.2004.04.001
obtain metaphase preparations. However, FISH has dramatically improved this situation as the technique can be performed on interphase cells (Werner et al., 1997), permitting the acquisition of cytogenetic data from tissues that previously have not been amenable to such analysis, hence revolutionising solid tumour cytogenetics. Due to the increasing popularity of FISH, many types of ready-labelled probes are commercially available, designed for diagnostic purposes to determine the spatial location of specific nucleotide sequences. FISH probes can be used to paint the whole genome (i.e., SKY—spectral karyotyping; M-FISH—multicolour FISH) or sections such as individual chromosomes (whole chromosome paints) and separate chromosome arms (partial chromosome paints). Alternatively, a much smaller chromosomal region may be detected with centromeric, telomeric or locus-specific probes. FISH probes for whole chromosomes or the centromeric and telomeric regions are readily available for each human
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
chromosome. However, the range of commercial locusspecific probes is considerably more limited. FISH probes for genes or loci that are commonly altered through amplification, deletion or translocations in genetic diseases or during carcinogenesis can be purchased from various companies, but these only span a very small proportion of the genome and in our experience the quality is variable. Hence, probes for specific single-copy sequences need to be readily generated to accommodate individual research requirements. We therefore present the detailed methodology required to design and generate such locus-specific FISH probes, with specific regard to their use in whole interphase cell preparations such as those commonly used in diagnostic pathology (e.g., touch preparations, cell smears, fine needle aspirates and tissue sections). In addition, the application of these probes is illustrated in cancer research, examining the cytogenetic alterations that accompany neoplastic progression of the premalignant lesion, Barrett’s esophagus to an adenocarcinoma. Esophageal adenocarcinoma cells have been successfully cultured (De Both et al., 2001; Rockett et al., 1997), but to date productive Barrett’s esophagus cell cultures have been few and far between due to contamination problems and an inability to generate immortal cell lines (Khan et al., 1997; Palanca-Wessels et al., 1998). Efforts have been made in our laboratory to generate Barrett’s cell lines from biopsies originating from patients at various stages of progression. The cells grew very slowly for several weeks but they always died before they could be passaged. As a result, metaphase cell preparations could not be obtained from premalignant Barrett’s samples; thus, cytogenetic studies on these specimens can only be performed with interphase cells. A previous report from our laboratory utilising commercial centromeric probes on cytology preparations demonstrated that aneuploidy was a prevalent event at all Barrett’s histological grades, with chromosome 4 hyperploidy representing the earliest and most prominent alteration (Doak et al., 2003). However, we could not be sure the whole chromosome was actually amplified. Commercially available telomeric probes were not suitable for the interphase cell substrate as the resultant signals were too obscure (due to the small probe size optimised for metaphase spreads) for efficient and reliable scoring. Chromosome 4 subtelomeric locus-specific probes that generated a bright compact signal within interphase cells were thus initially produced in the present study to assess whether or not the whole chromosome was indeed amplified in Barrett’s esophagus. One of the most prominent epigenetic changes that arises during tumourigenesis is the hypermethylation of gene control regions, which acts to silence their expression. Aberrant methylation patterns have been implicated in the pathogenesis of Barrett’s esophagus, particularly involving epigenetic silencing of the p16 gene (Bian et al., 2002; Eads et al., 2001; Klump et al., 1998). To date, three DNA methyltransferase enzymes (DNMT1, 3a and 3b) have been
27
identified in mammalian cells (El-Deiry et al., 1991; Okano et al., 1998), and their overexpression has been correlated with the development of leukemia (Mizuno et al., 2001), colorectal (De Marzo et al., 1999; Jubb et al., 2001) and ovarian cancers (Ahluwalia et al., 2001). However, their association with altered methylation patterns in Barrett’s tissues is presently unknown. We have previously reported that an amplified chromosome 20 copy number accompanied Barrett’s neoplastic progression (Doak et al., 2003). This has also been noted in several other cytogenetic analyses with high frequency gains reported at 20q11.2.2– 20q13.1 (Riegman et al., 2001; Varis et al., 2001; Walch et al., 2000). This is of particular interest as the gene for DNMT3b is located at 20q11.2, thus suggesting these alterations could lead to up-regulation of this methylase enzyme. The present study was therefore additionally aimed at investigating the status of the other DNMT genes [i.e., the DNMT1 (19p13.2) and DNMT3a (2p23) genes] at the chromosomal level during Barrett’s neoplastic progression, using interphase FISH. A FISH probe was commercially available for the DNMT3a locus, but not for the DNMT1 gene; hence, this probe had to be synthesised and optimised.
Materials and methods Patient samples A range of interphase preparations from patients with varying Barrett’s histological grades and degrees of chromosome 4 aneuploidy were selected from our previously analysed sample series (Doak et al., 2003) to assess the extent of the chromosomal aberrations. Four Barrett’s metaplasia (BM), three low-grade dysplasia (LGD), three high-grade dysplasia (HGD), and two esophageal adenocarcinomas (EA) were analysed. Of these, three did not display chromosome 4 aneuploidy, one presented loss of chromosome 4 and the remaining eight demonstrated chromosome 4 hyperploidy to varying degrees. Ten control slides (interphase cells originating from an area of normal squamous cell epithelium) were also included in the analysis to determine the background levels of hybridisation signal loss or gain. Additionally, endoscopic cytology brushings were collected from consenting patients attending endoscopy clinics with a special interest in Barrett’s esophagus at Morriston District General Hospital, Swansea, UK, following ethical approval. Brushings from the Barrett’s segment and an area of normal squamous epithelium were obtained from 12 patients with Barrett’s metaplasia, 7 low-grade dysplasia (LGD) and 4 high-grade dysplasia (HGD) for the DNMT analysis. The male –female ratio was 3:1, while the ages ranged from 38 to 85 years with a median of 62 years. Interphase cell preparations were subsequently generated and pretreated as previously described (Doak et al., 2003).
28
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
All sample slides were coded; hence, subsequent experimentation and scoring was performed with no knowledge of histological grade or previous chromosome 4 cytogenetic status.
tetramethylrhodamine-dUTP (Roche Diagnostics Ltd.). All nick translation products were subsequently purified with the QIAquick PCR Purification Kit (Qiagen) according to the manufacturer’s instructions. However, the probes were finally eluted in 30 Al of distilled water.
Cosmid or BAC clone preparation Fluorescence in situ hybridisation The vectors required to generate the FISH probes were selected using the Human Genome Resources at www.ncbi. nih.gov/genome/guide/human/. The appropriate chromosome was selected on the ‘‘Browse Your Genome for Clones’’ option, and then the region of interest was zoomed into until all adjacent bacterial artificial chromosome (BAC) clones were shown. Distributor information for each BAC clone was given when selected from the maps. To view cosmid clones available within the region of interest, the ‘‘Component’’ selection on the master map was chosen. Adjacent clones spanning a region of approximately 200 kb from the areas of interest were subsequently selected. Adjacent human BAC clones from the 4p16.3 and 4q35.1 regions for the subtelomeric chromosome 4 probes were purchased from the BACPAC Resource Centre at Childrens Hospital Oakland Research Institute (California, USA). Human cosmid vectors were purchased from the UK Human Genome Mapping Project Resource Centre (Cambridge, UK) to generate the DNMT1 gene-specific probe (accession number AF180682). On arrival, each colony was cultured in sterile LB broth supplemented with 20 Ag/ml chloramphenicol or 25 Ag/ml kanamycin (Sigma, Dorset, UK) for the BAC or cosmid clones, respectively. A 30-ml cell suspension per clone was centrifuged and the BAC or cosmid vector was extracted from the cell pellet using the QIAprep Spin Miniprep Kit (Qiagen, West Sussex, UK) according to manufacturer’s instructions with slight modifications to accommodate the large vector sizes. The volume of the first three solutions used were doubled to ensure complete lysis of all cells within the pellet, and the vector DNA elution recommendations were altered to two applications of 30 Al of distilled water at 70jC left on the columns’ membrane for 1 min before centrifugation. This procedure improved the suspension of the vector and increased subsequent yields. Nick translation One microgram of the BAC or cosmid DNA was labelled by nick translation using the Nick Translation System from Invitrogen (Paisley, UK) according to the manufacturer’s recommendations. However, the incubation time was extended to 2 h 30 min following empirical time series experiments to generate labelled fragments between 200 and 800 bp in size. The BAC clones for the 4p16.3 probe and cosmids for the DNMT1 probe were labelled with fluorescein-dUTP (Roche Diagnostics Ltd., East Sussex) during nick translation, while the 4q35.1 BAC clones were labelled with
For the DNMT analysis, FISH was performed with the commercially available Locus-Specific Identifier (LSI) 2p23 probe (Vysis, Surrey, UK) for DNMT3a, as previously described (Doak et al., 2003). The second probe for this analysis (DNMT1 gene) and the two required for the assessment of chromosome 4 hyperploidy in Barrett’s esophagus were produced in-house. Two hundred nanograms of the BAC or cosmid nick translation products for each region (4p16.3, 4q35.1, DNMT1) were pooled and 20 Cot-1 DNA (Roche Diagnostics Ltd.) was added to each of the three preparations. The resultant probe mixes were precipitated and the subsequent DNA pellets were each resuspended in 5 Al of hybridisation buffer consisting of 50% formamide, 2 SSC, 10% dextran sulphate, 1 Ag salmon sperm and water to provide a 10-Al total volume. Before denaturation, target slides were dehydrated in an ethanol series: 70%, 80% and 95% for 2 min each. Once dried, the slides were denatured in 70% formamide/2 SSC at 75jC for 5 min, quenched in an ice-cold ethanol series (2 min in each of 70%, 80% and 95% ethanol) and subsequently air-dried. The probe was denatured at 75jC for 10 min and then incubated at 37jC for 20 min to allow the preannealing of repetitive sequences. The probes were hybridised to the interphase preparations overnight at 37jC in a humidified chamber. The slides were washed in 0.4 SSC/0.3% NP-40 at 73jC for 2 min, 2 SSC/0.1% NP-40 at ambient temperature for 30 s and left to air-dry in the dark. Nuclei were counterstained with 10 Al of 125 ng/ml DAPI II (Vysis) and slides were viewed under an Olympus BX 50 fluorescence microscope using a UplanF1 100/1.3 oil objective. Probes were visualised with single- and multiple-bandpass filter sets, while images were captured with a cooled CCD (chargedcoupled device) camera and analysed with the MacProbe version 4.1 software (Applied Imaging, Newcastle Upon Tyne, UK). Scoring was performed as previously described in Doak et al. (2003). The mean signal gains or losses plus three standard deviations in control samples were used to establish significant cytogenetic alterations. Consequently, the cutoff percentage of nuclei defining significant abnormal signal losses was 1.6%, 1.6%, 2.2% and 3.1% for the 2p23 (DNMT3a), DNMT1, 4p16.3 and 4q35.1 probes, respectively. The cutoff for abnormal signal gains with the 2p23 and DNMT1 probes were 1.6% and 1.0%, respectively. However, no signal gains were observed in the control samples with the 4p16.3 and 4q35.1 probes.
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
Results FISH optimisation All initial FISH reactions to optimise the hybridisation of the probes produced were performed on metaphase preparations from human male lymphocytes to ensure that they were hybridising to the correct region(s) of the genome. Initial experiments were designed to ascertain the size of the labelled probe required to generate a signal that was sufficiently strong to ensure visualisation in interphase
29
cells, where the chromosomes are not condensed (as in metaphases). Five BAC clones from the 4p16.3 and 4q35.1 regions were selected and purchased. The BAC DNA was extracted, each was individually labelled via nick translation, and then 200 ng of each were pooled to generate probes ranging from 50 to 250 kb, which were subsequently hybridised to metaphase preparations. Figs. 1a and b display the signal intensities obtained when varying the probe size, and it could be consequently concluded that although signals were visible with 100 kb probes, those >150 kb generated easily identifiable signals. Twenty
Fig. 1. FISH probe optimisation and application. (A) Signals at 4q35.3 generated from a 120-kb probe. (B) Signals at 4q35.3 generated from a 200-kb probe. (C) Comparison between the commercially available LSI 2p23 probe (red) and the DNMT1 probe produced in the present study from cosmid clones. (D) Cohybridisation of the 4p16.3 (green) and 4q35.1 (red) locus-specific probes. (E) Scattered 4p16.3 and 4q35.1 signals. (F) More compact 4p16.3 and 4q35.1 signals in interphase nuclei. (G and H) Whole chromosome 4 amplification within Barrett’s interphase nuclei.
30
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
metaphases were examined, all of which displayed probe hybridisation signals in the correct location with no cross hybridisation, thereby demonstrating the high specificity and sensitivity of the probes produced. This procedure was also repeated with the cosmid clones for DNMT1 and a strong signal resulted, which was comparable to commercially available locus-specific FISH probes (Fig. 1c). Before FISH, both the probe and target samples were denatured. This can either be done separately or both probe and target may be denatured together. Both techniques were assessed to determine which provided stronger signals. The probe and metaphase target were denatured individually using a standardised formamide-based technique (VarellaGarcia et al., 1998; von Bergh et al., 2000), while the probe and a second metaphase target were codenatured for 2 min on a 75jC hotplate. Upon examination, signals achieved by individually denaturing the probe and sample was considerably stronger than when using codenaturation. Hence, in all subsequent FISH experiments, the formamide-based technique was utilised. As the degree of the probes’ nonspecific hybridisation is difficult to predict, stringency washes also needed careful adjustment through trial and error to optimise the intensity of the FISH signals. Low to high stringency washes were performed on several slides that had been hybridised with the same probe. Subsequent analysis demonstrated signal intensities were strongest with a relatively low stringency wash that was subsequently used for all subsequent FISH experiments (as detailed in Materials and methods). For the chromosome 4 study, it was more desirable to use multicolour FISH (i.e., simultaneously hybridise both probes designed for the 4p16.3 and 4q35.1 regions) as this method conserved both time and precious samples. It therefore had to be ensured that the two probes produced could be successfully cohybridised without losing signal intensity and quality. To investigate, the probes for the 4p16.3 and 4q35.1 regions were combined, precipitated and resuspended in hybridisation buffer. The subsequent probe was then hybridised to a metaphase preparation and consequently analysed. Cohybridisation of the two probes was successful as can be seen in Fig. 1d, with no compromise to signal size or intensity; hence, the use of multicolour FISH was demonstrated to be feasible with the single-copy sequence probes. Once optimised on metaphase preparations, the probes and techniques developed were tested on interphase nuclei. The probes originating from BAC and cosmid clones were therefore applied to interphase preparations, but a very scattered signal was detected with the BAC probes (Fig. 1e) due to their size and gaps between their mapped positions—a characteristic that could not be seen on the metaphase preparations, where DNA is considerably more condensed. To avoid this scatter, only slightly overlapping clones spanning approximately 150 kb DNA were selected and resulted in stronger and more compact signals in interphase nuclei comparable to commercial probes (Fig. 1f).
Large DNA fragments (such as those cloned within BAC vectors) contain long stretches of repetitive sequences. Before probe hybridisation, these sequences must be blocked to prevent them from annealing to large regions of the target DNA, thereby reducing nonspecific hybridisation. This is achieved by incubation with an unlabelled competitor DNA, which anneals to the repetitive sequences in both the sample and probe, without reducing the probes’ hybridisation efficiency to its intended target. A frequently used block is human cot-1 DNA, a modified form of total genomic DNA that is enriched for the repetitive sequences. Although prehybridisation with cot-1 DNA reduces nonspecific background, the concentration utilised must be adjusted accordingly. When too high, it interferes with probe hybridisation, and when too low the background noise is strong. A series of FISH experiments with cot-1 DNA at 10, 20 or 50 of the total probe quantity were therefore performed. The lowest concentration that gave the least background hybridisation was 20 and hence was utilised in all subsequent FISH experiments. Chromosome 4 hyperploidy This analysis demonstrated that chromosome 4 is indeed hyperploid in Barrett’s esophagus as both the 4p16.3 and 4q35.1 probes had the same copy number in all samples. A majority of the hyperploid nuclei detected had three copies of 4p16.3 and 4q35.1, but in some samples (particularly at the more advanced stages) up to seven copies of each were identified, as illustrated in Figs. 1g and h. Occasionally, nuclei were detected with 3 4p16.3 and 2 4q35.1q signals (or vice versa), but the number of such nuclei within the samples was always very low (a maximum of five nuclei per sample). Hence, inadequate probe hybridisation within these nuclei may have been at fault rather than amplification of just one chromosome arm. Table 1 compares the percentage of aberrant nuclei previously detected with the commercially available CEP 4 (Vysis) centromeric probe (Doak et al., 2003) and with the locus-specific 4p16.3 and 4q35.1 probes generated in the present study. As can be seen, there was a clear positive correlation between the results generated with both probe types. In all but two of the samples, there was no significant difference detected between the proportion of abnormal nuclei identified with the CEP 4 and locus-specific 4p/q probes (using the v2 statistic, P < 0.05). Samples 4 and 22 were the two in which a significant difference was detected; however, both probe types demonstrated chromosome 4 hyperploidy within the sample. It was the proportion of sample cells displaying the abnormality that differed. Cytogenetic status of the DNMT1 and DNMT3a loci The DNMT3a locus was analysed first due to the commercial availability of the probe. However, significant
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33 Table 1 Percentage of cells displaying chromosome 4 aneuploidy using a commercially available probe for its centromeric region (CEP 4; Vysis) and the probes produced for 4p16.3 and 4q35.1 Slide
Pathology
Chromosome 4 loss a
24 26 41 43 120 122 126 134 136 140 4 11 49 86 29 40 55 1 10 91 15 22
Control Control Control Control Control Control Control Control Control Control BM BM BM BM LGD LGD LGD HGD HGD HGD EA EA
b
Chromosome 4 gain
CEP 4
4p/q
CEP 4a
4p/qb
NN NN NN NN NN NN NN NN NN NN 1.2 1.3 1.3 0.0 0.5 0.8 0.7 4.7 3.0 0.3 1.5 1.0
1.1 1.4 1.7 1.8 1.3 1.1 1.4 1.8 0.6 1.1 1.8 1.3 1.1 0.9 0.9 1.3 1.3 1.1 0.8 0.5 1.8 0.7
NN NN NN NN NN NN NN NN NN NN 22.8 0.0 0.0 25.3 25.5 28.5 3.2 1.7 25.3 50.6 32.4 54.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 12.6 0.0 0.0 21.5 23.5 19.8 5.0 2.5 20.3 43.0 28.5 26.2
NN: data were not necessary, as controls were included to examine background hybridisation for the 4p/q probes. a Data were obtained from our previous study (Doak et al., 2003). b Average percentage of 4p16.3 and 4q35.1 signal losses or gains.
copy number alterations were very rarely detected. No samples demonstrated loss of DNMT3a while gains (3 signals) were only observed in four samples (1 metaplasia; 1 LGD; 2 HGD), but these were never noted in more than 3% of cells. Due to the low number of alterations detected at the DNMT3a locus, a preliminary investigation including only the HGDs was analysed with the DNMT1 probe. Again, abnormalities were rare; one sample demonstrated significant signal loss in 4% of nuclei and one sample had 3 DNMT1 signals in 2.5% of nuclei. Due to the lack of alterations at this advanced stage of Barrett’s progression, examination of the earlier grades was not deemed necessary.
Discussion Commercially available single-copy sequence probes are very limited but are valuable for research and are an important tool in identifying new molecular biomarkers for disease. We have therefore detailed the protocols necessary for designing and generating custom locus-specific probes for interphase FISH. In addition, we have demonstrated their use on cytology preparations, such as those commonly utilised in routine diagnostic pathology. This technique for probe generation considerably reduces the costs of FISH when compared to the expense of purchasing commercially
31
available probes. Hence, our results indicate this approach of in-house probe preparation is practical and cost-effective. Long stretches of DNA can be obtained from cloned sequences to generate FISH probes for targeting specific chromosomal regions and a number of vectors with varying insert sizes exist: cosmids—35 to 45 kb; BAC—70 to 100 kb; PAC—100 to 300 kb; YAC—100 kb to 2 Mb. Large quantities of high purity, double stranded DNA can be efficiently isolated from host cells (in which the vector is amplified), and clones containing DNA fragments spanning the whole genome are readily available from sequencing institutions. Cloned DNA was therefore the favoured substrate for the generation of locus-specific probes, but the smaller vectors were chosen (cosmids and BACs), as they are easier to label and contain less redundant DNA. In the first instance, nick translation was used to differentially label human DNA cloned within BAC vectors, originating from the 4p16.3 and 4q35.1 loci. Following optimisation of hybridisation conditions and stringency washes, multicolour FISH was subsequently performed on interphase preparations with known chromosome 4 aberrations, as determined by our previous analysis (Doak et al., 2003). A strong correlation between the data generated by the commercially available Vysis probe (Doak et al., 2003) and those recorded for the 4p16.3 and 4q35.1 locus-specific probes existed. In all cases, the chromosome 4 status (i.e., hyperploid/monosomic/diploid) determined by the two types of probes was concordant. In 10/12 (83%) of the cases examined, no significant differences between the percentages of aberrant cells were detected. For the two cases in which there was a significant difference, the proportion of sample cells displaying the abnormality differed slightly. This variation may have been due to the use of different interphase preparations; therefore, a different population of cells was analysed by each probe type. However, despite this, chromosome 4 hyperploidy was still detected in these samples with both the CEP 4 and 4p/q probes. Hence, the 4p16.3 and 4q35.1 locus-specific probes generated established that the whole chromosome 4 structure was amplified in Barrett’s esophagus. This investigation was additionally aimed at determining the significance of gene dosage imbalances at the DNMT loci during Barrett’s progression, as abnormal methylation patterns have been implicated in Barrett’s tumourigenesis (Eads et al., 2000, 2001). The DNMT enzymes are responsible for establishing and maintaining DNA methylation patterns in eukaryotes (Okano et al., 1998, 1999). Hence, it has been speculated that increases in their activity may be accountable for epigenetic silencing during human carcinogenesis (Antoun et al., 2000; Saito et al., 2003; Vertino et al., 1996). Previous Barrett’s cytogenetic investigations have detected amplifications at the DNMT3b locus (20q11.2; Riegman et al., 2001; Varis et al., 2001; Walch et al., 2000), but reports regarding the DNMT3a and DNMT1 loci are conflicting (Riegman et al., 2001; Van Dekken et al.,
32
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
1999; Walch et al., 2000), they were therefore assessed in the present study. A FISH probe was commercially available for DNMT3a, but not for the DNMT1 gene; hence, a probe for this region had to be produced. Following the success of the chromosome 4 locus-specific probes, a fluorescently labelled probe for the DNMT1 locus was produced. However, cosmid clones were utilised to generate the probe, as the appropriate BAC clones were not available for the region of interest. Once developed and optimised, this probe, together with the commercially available probe for the DNMT3a locus, was applied to interphase Barrett’s samples with various histological grades to assess the status of the DNMT genes at the chromosomal level. This analysis subsequently determined that no significant copy number changes were present at the DNMT1 and DNMT3a loci. In a small number of cases, gains of the loci were noted, but they were in less than 4% of cells and predominantly arose in HGD. Hence, it is most probable that these observations are a consequence of the general increase in aneuploidy and genetic instability that accompanies Barrett’s neoplastic progression (Croft et al., 2002; Doak et al., 2003; Riegman et al., 2001). In summary, this report details the methodology required to generate single-copy locus-specific interphase FISH probes for any genomic region of interest, hence broadening cytogenetic applications in the research of solid tumours or diseased tissues and making the technique more amenable to diagnostic cytopathology in the future. Two probes generated for the subtelomeric regions of chromosome 4 were used to verify the amplification identified in a previous investigation actually involved the whole chromosome. In addition, a FISH probe was produced and subsequently implemented to facilitate the assessment of DNMT gene dosages during Barrett’s associated neoplastic progression. Copy number alterations of the loci at which the DNMT1 and DNMT3a genes are located were very rarely observed, indicating amplifications or deletions of these genes are unlikely to be responsible for the aberrant DNMT activity observed in Barrett’s esophagus. Alterations in gene expression or the enzymes’ biochemical properties are alternative mechanisms that may drive hypermethylation during Barrett’s tumourigenesis but have yet to be assessed.
Acknowledgment SH Doak was funded by a studentship from the Tenovus cancer charity.
References Ahluwalia, A., Hurteau, J.A., Bigsby, R.M., Nephew, K.P., 2001. DNA methylation in ovarian cancer: II. Expression of DNA methyltransferases in ovarian cancer cell lines and normal ovarian epithelial cells. Gynecol. Oncol. 82, 299 – 304. Antoun, G., Baylin, S.B., Ali-Osman, F., 2000. DNA methyltransferase
levels in altered CpG methylation in the total genome and in the GSTP1 gene in human glioma cells transfected with sense and antisense DNA methyltransferase cDNA. J. Cell. Biochem. 77, 372 – 381. Bian, Y.S., Osterheld, M.C., Fontolliet, C., Bosman, F.T., Benhattar, J., 2002. p16 inactivation by methylation of the CDKN2A promoter occurs early during neoplastic progression in Barrett’s esophagus. Gastroenterology 122, 1113 – 1121. Croft, J., Parry, E.M., Jenkins, G.J.S. et al., 2002. Analysis of the premalignant stages of Barrett’s esophagus through to adenocarcinoma by comparative genomic hybridisation. Eur. J. Gastroenterol. Hepatol. 14, 1179 – 1186. Doak, S.H., Jenkins, G.J.S., Parry, E.M., D’Souza, F.R., Griffiths, A.P., Toffazal, N., Shah, V., Baxter, J.N., Parry, J.M., 2003. Chromosome 4 hyperploidy represents an early genetic aberration in pre-malignant Barrett’s oesophagus. Gut 52, 623 – 629. De Both, N.J., Wijnhoven, B.P.L., Sleddens, H.F.B.M., Tilanus, H.W., Dinjens, W.N.M., 2001. Establishment of cell lines from adenocarcinomas of the esophagus and gastric cardia growing in vivo and in vitro. Virchows Arch.—An International Journal of Pathology 438, 451 – 456. De Marzo, A.M., Marchi, V.L., Yang, E.S., Veeraswamy, R., Lin, X.H., Nelson, W.G., 1999. Abnormal regulation of DNA methyltransferase expression during colorectal carcinogenesis. Cancer Res. 59, 3855 – 3860. Eads, C.A., Lord, R.V., Kurumboor, S.K., Wickramasinghe, K., Skinner, M.L., Long, T.I., Peters, J.H., DeMeester, T.R., Danenberg, K.D., Danenberg, P.V., Laird, P.W., Skinner, K.A., 2000. Fields of aberrant CpG island hypermethylation in Barrett’s esophagus and associated adenocarcinoma. Cancer Res. 60, 5021 – 5026. Eads, C.A., Lord, R.V., Wickramasinghe, K., Long, T.I., Kurumboor, S.K., Bernstein, L., Peters, J.H., DeMeester, S.R., DeMeester, T.R., Skinner, K.A., Laird, P.W., 2001. Epigenetic patterns in the progression of esophageal adenocarcinoma. Cancer Res. 61, 3410 – 3418. Eastmond, D.A., Schuler, M., Rupa, D.S., 1995. Advantages and limitations of using fluorescence in situ hybridisation for the detection of aneuploidy in interphase human cells. Mutat. Res. 348, 153 – 162. El-Deiry, W.S., Nelkin, B.D., Celano, P., Chiu Yen, R.W., Falco, J.P., Hamilton, S.R., Baylin, S.B., 1991. High expression of the DNA methyltransferase gene characterises human neoplastic cells and progression stages of colon cancer. Proc. Natl. Acad. Sci. 88, 3470 – 3474. Jubb, A.M., Bell, S.M., Quirke, P., 2001. Methylation and colorectal cancer. J. Pathol. 195, 111 – 134. Khan, S.M., Pillay, S.P., Papadimos, D., Yong, J.W.K., Roberts, H.J.V., Crawford, D.H., 1997. Barrett’s esophagus: a technique for the culture of Barrett’s oesophageal cells. J. Gastroenterol. Hepatol. 12, 606 – 611. Klump, B., Hsieh, C.-J., Holzmann, K., Gregor, M., Porschen, R., 1998. Hypermethylation of the CDKN2/p16 promoter during neoplastic progression in Barrett’s esophagus. Gastroenterology 115, 1381 – 1386. Mizuno, S., Chijiwa, T., Okamura, T., Akashi, K., Fukumaki, Y., Niho, Y., Sasaki, H., 2001. Expression of DNA methyltransferases DNMT1, 3A and 3B in normal hematopoiesis and in acute and chronic myelogenous leukaemia. Blood 97, 1172 – 1179. Okano, M., Xie, S., Li, E., 1998. Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat. Genet. 19, 219 – 220. Okano, M., Bell, D.W., Haber, D.A., Li, E., 1999. DNA methyltransferases DNMT3a and DNMT3b are essential for de novo methylation and mammalian development. Cell 99, 247 – 257. Palanca-Wessels, M.C., Barrett, M.T., Galipeau, P.C., Rohrer, K.L., Reid, B.J., Rabinovitch, P.S., 1998. Genetic analysis of long-term Barrett’s esophagus epithelial cultures exhibiting cytogenetic and ploidy abnormalities. Gastroenterology 114, 295 – 304. Popescu, N.C., Zimonjic, D.B., 1997. Molecular cytogenetic characterisation of cancer cell alterations. Cancer Genet. Cytogenet. 93, 10 – 21. Riegman, P.H.J., Vissers, K.J., Alers, J.C., Geelen, E., Hop, W.C.J., Tilanus, H.W., Van Dekken, H., 2001. Genomic alterations in malignant transformation of Barrett’s esophagus. Cancer Res. 61, 3164 – 3170.
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33 Rockett, J.C., Larkin, K., Darnton, S.J., Morris, A.G., Matthews, H.R., 1997. Five newly established esophageal carcinoma cell lines: phenotypic and immunological characterization. Br. J. Cancer 75, 258 – 263. Saito, Y., Kanai, Y., Nakagawa, T., Sakamoto, M., Saito, H., Ishi, H., Hirohashi, S., 2003. Increased protein expression of DNA methyltransferase (DNMT) 1 is significantly correlated with the malignant potential and poor prognosis of human hepatocellular carcinomas. Int. J. Cancer 105, 527 – 532. Van Dekken, H., Vissers, C.J., Tilanus, H.W., Tanke, H.J., Rosenberg, C., 1999. Clonal analysis of a case of multifocal esophageal (Barrett’s) adenocarcinoma by comparative genome hybridisation. J. Pathol. 188, 263 – 266. Varella-Garcia, M., Gemmill, R.M., Rabenhorst, S.H., Lotto, A., Drabkin, H.A., Archer, P.A., Franklin, W.A., 1998. Chromosomal duplication accompanies allelic loss in non-small cell lung carcinoma. Cancer Res. 58, 4701 – 4707. Varis, A., Puolakkainen, P., Savolainen, H., Kokkola, A., Salo, J., Nieminen, O., Nordling, S., Knuutila, S., 2001. DNA copy number profiling in esophageal Barrett’s adenocarcinoma: comparison with gastric
33
carcinoma and esophageal squamous cell carcinoma. Cancer Genet. Cytogenet. 127, 53 – 58. Vertino, P.M., Yen, R.W.C., Gao, J., Baylin, S.B., 1996. De novo methylation of CpG island sequences in human fibroblasts over-expressing DNA (cytosine-5) methyltransferase. Mol. Cell. Biol. 16, 4555 – 4565. von Bergh, A., Emanuel, B., van Zelderen-Bhola, S., Smetsers, T., van Soest, R., Stul, M., Vranckx, H., Schuuring, E., Hagemeijer, A., Kluin, 2000. A DNA probe combination for improved detection of MLL/1 1q23 breakpoints by double-color interphase-FISH in acute leukemias. Genes Chromosomes Cancer 28, 14 – 22. Walch, A.K., Zitzelsberger, H.F., Bruch, J., Keller, G., Angermeier, D., Aubele, M.M., Mueller, J., Stein, H., Braselmann, H., Siewert, J.R., Hofler, H., Werner, M., 2000. Chromosomal imbalances in Barrett’s adenocarcinoma and the metaplasia – dysplasia – carcinoma sequence. Am. J. Pathol. 156, 555 – 566. Werner, M., Ludwig, W., Aubele, M., Nolte, M., Zitzelsberger, H., Komminoth, P., 1997. Interphase cytogenetics in pathology: principles, methods and applications of fluorescence in situ hybridisation (FISH). Histochem. Cell Biol. 108, 381 – 390.
Generation of locus-specific probes for interphase fluorescence in situ hybridisation—application in Barrett’s esophagus S.H. Doak, a,* D. Saidely, a G.J.S. Jenkins, a E.M. Parry, a A.P. Griffiths, b J.N. Baxter, c and J.M. Parry a a
Human Molecular Pathology Group, School of Biological Sciences, University of Wales Swansea, Swansea, SA2 8PP, UK b Department of Pathology, Morriston Hospital, Swansea SA6 6NL, UK c Department of Surgery, Morriston Hospital, Swansea SA6 6NL, UK Received 1 April 2004 Available online 2 June 2004
Abstract Despite the wide range of probes commercially available for interphase fluorescence in situ hybridisation (FISH), the supply of locusspecific probes is limited to genes or chromosomal regions commonly altered in genetic diseases or during carcinogenesis. Generation of these probes is therefore desirable to accommodate individual research requirements. Hence, we detail the methodology required to design and produce custom locus-specific interphase FISH probes for any human genomic region of interest and their application was illustrated in cytogenetic investigations of Barrett’s tumourigenesis. Previously utilising FISH, we observed that Barrett’s tissues demonstrated chromosome 4 hyperploidy [Gut 52 (2003) 623], but as centromeric probes were used in this analysis, it was not known if the whole chromosome was amplified. We consequently generated single-copy sequence probes for the 4p16.3 and 4q35.1 subtelomeric loci. Multicolour FISH was subsequently performed on interphase preparations originating from patients with Barrett’s esophagus at varying histological grades, thus demonstrating the whole region of chromosome 4 was amplified within the tissues. Additionally, probes for the DNA methyltransferase genes were produced to determine if gene dosage alterations were responsible for increasing methylation activity during Barrett’s neoplastic progression. No significant alterations at the DNMT1 and DNMT3a loci were detected. An increased copy number of these genes is therefore not the basis for the hypermethylation commonly observed in this premalignant lesion. D 2004 Elsevier Inc. All rigths reserved. Keywords: Barrett’s esophagus; DNMT1; DNMT3a; Fluorescence in situ hybridisation; Locus-specific probes
Introduction Fluorescence in situ hybridisation (FISH) is a powerful cytogenetic tool for directly localising and visualising unique regions of the genome at the single cell level. It has been widely exploited in many applications ranging from environmental genotoxicity studies to DNA mapping and has proved to be a vital source of genetic information in cancer research, providing data on whole chromosome copy number changes down to deletions or amplifications of single genes (Eastmond et al., 1995; Popescu and Zimonjic, 1997; Werner et al., 1997). Historically, conventional cytogenetic analysis has relied on the ability to culture cells to * Corresponding author. Human Molecular Pathology Group, School of Biological Sciences, University of Wales Swansea, Singleton Park, Swansea, SA2 8PP, UK. Fax: +44-1792-295447. E-mail address: [email protected] (S.H. Doak). 0014-4800/$ - see front matter D 2004 Elsevier Inc. All rigths reserved. doi:10.1016/j.yexmp.2004.04.001
obtain metaphase preparations. However, FISH has dramatically improved this situation as the technique can be performed on interphase cells (Werner et al., 1997), permitting the acquisition of cytogenetic data from tissues that previously have not been amenable to such analysis, hence revolutionising solid tumour cytogenetics. Due to the increasing popularity of FISH, many types of ready-labelled probes are commercially available, designed for diagnostic purposes to determine the spatial location of specific nucleotide sequences. FISH probes can be used to paint the whole genome (i.e., SKY—spectral karyotyping; M-FISH—multicolour FISH) or sections such as individual chromosomes (whole chromosome paints) and separate chromosome arms (partial chromosome paints). Alternatively, a much smaller chromosomal region may be detected with centromeric, telomeric or locus-specific probes. FISH probes for whole chromosomes or the centromeric and telomeric regions are readily available for each human
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
chromosome. However, the range of commercial locusspecific probes is considerably more limited. FISH probes for genes or loci that are commonly altered through amplification, deletion or translocations in genetic diseases or during carcinogenesis can be purchased from various companies, but these only span a very small proportion of the genome and in our experience the quality is variable. Hence, probes for specific single-copy sequences need to be readily generated to accommodate individual research requirements. We therefore present the detailed methodology required to design and generate such locus-specific FISH probes, with specific regard to their use in whole interphase cell preparations such as those commonly used in diagnostic pathology (e.g., touch preparations, cell smears, fine needle aspirates and tissue sections). In addition, the application of these probes is illustrated in cancer research, examining the cytogenetic alterations that accompany neoplastic progression of the premalignant lesion, Barrett’s esophagus to an adenocarcinoma. Esophageal adenocarcinoma cells have been successfully cultured (De Both et al., 2001; Rockett et al., 1997), but to date productive Barrett’s esophagus cell cultures have been few and far between due to contamination problems and an inability to generate immortal cell lines (Khan et al., 1997; Palanca-Wessels et al., 1998). Efforts have been made in our laboratory to generate Barrett’s cell lines from biopsies originating from patients at various stages of progression. The cells grew very slowly for several weeks but they always died before they could be passaged. As a result, metaphase cell preparations could not be obtained from premalignant Barrett’s samples; thus, cytogenetic studies on these specimens can only be performed with interphase cells. A previous report from our laboratory utilising commercial centromeric probes on cytology preparations demonstrated that aneuploidy was a prevalent event at all Barrett’s histological grades, with chromosome 4 hyperploidy representing the earliest and most prominent alteration (Doak et al., 2003). However, we could not be sure the whole chromosome was actually amplified. Commercially available telomeric probes were not suitable for the interphase cell substrate as the resultant signals were too obscure (due to the small probe size optimised for metaphase spreads) for efficient and reliable scoring. Chromosome 4 subtelomeric locus-specific probes that generated a bright compact signal within interphase cells were thus initially produced in the present study to assess whether or not the whole chromosome was indeed amplified in Barrett’s esophagus. One of the most prominent epigenetic changes that arises during tumourigenesis is the hypermethylation of gene control regions, which acts to silence their expression. Aberrant methylation patterns have been implicated in the pathogenesis of Barrett’s esophagus, particularly involving epigenetic silencing of the p16 gene (Bian et al., 2002; Eads et al., 2001; Klump et al., 1998). To date, three DNA methyltransferase enzymes (DNMT1, 3a and 3b) have been
27
identified in mammalian cells (El-Deiry et al., 1991; Okano et al., 1998), and their overexpression has been correlated with the development of leukemia (Mizuno et al., 2001), colorectal (De Marzo et al., 1999; Jubb et al., 2001) and ovarian cancers (Ahluwalia et al., 2001). However, their association with altered methylation patterns in Barrett’s tissues is presently unknown. We have previously reported that an amplified chromosome 20 copy number accompanied Barrett’s neoplastic progression (Doak et al., 2003). This has also been noted in several other cytogenetic analyses with high frequency gains reported at 20q11.2.2– 20q13.1 (Riegman et al., 2001; Varis et al., 2001; Walch et al., 2000). This is of particular interest as the gene for DNMT3b is located at 20q11.2, thus suggesting these alterations could lead to up-regulation of this methylase enzyme. The present study was therefore additionally aimed at investigating the status of the other DNMT genes [i.e., the DNMT1 (19p13.2) and DNMT3a (2p23) genes] at the chromosomal level during Barrett’s neoplastic progression, using interphase FISH. A FISH probe was commercially available for the DNMT3a locus, but not for the DNMT1 gene; hence, this probe had to be synthesised and optimised.
Materials and methods Patient samples A range of interphase preparations from patients with varying Barrett’s histological grades and degrees of chromosome 4 aneuploidy were selected from our previously analysed sample series (Doak et al., 2003) to assess the extent of the chromosomal aberrations. Four Barrett’s metaplasia (BM), three low-grade dysplasia (LGD), three high-grade dysplasia (HGD), and two esophageal adenocarcinomas (EA) were analysed. Of these, three did not display chromosome 4 aneuploidy, one presented loss of chromosome 4 and the remaining eight demonstrated chromosome 4 hyperploidy to varying degrees. Ten control slides (interphase cells originating from an area of normal squamous cell epithelium) were also included in the analysis to determine the background levels of hybridisation signal loss or gain. Additionally, endoscopic cytology brushings were collected from consenting patients attending endoscopy clinics with a special interest in Barrett’s esophagus at Morriston District General Hospital, Swansea, UK, following ethical approval. Brushings from the Barrett’s segment and an area of normal squamous epithelium were obtained from 12 patients with Barrett’s metaplasia, 7 low-grade dysplasia (LGD) and 4 high-grade dysplasia (HGD) for the DNMT analysis. The male –female ratio was 3:1, while the ages ranged from 38 to 85 years with a median of 62 years. Interphase cell preparations were subsequently generated and pretreated as previously described (Doak et al., 2003).
28
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
All sample slides were coded; hence, subsequent experimentation and scoring was performed with no knowledge of histological grade or previous chromosome 4 cytogenetic status.
tetramethylrhodamine-dUTP (Roche Diagnostics Ltd.). All nick translation products were subsequently purified with the QIAquick PCR Purification Kit (Qiagen) according to the manufacturer’s instructions. However, the probes were finally eluted in 30 Al of distilled water.
Cosmid or BAC clone preparation Fluorescence in situ hybridisation The vectors required to generate the FISH probes were selected using the Human Genome Resources at www.ncbi. nih.gov/genome/guide/human/. The appropriate chromosome was selected on the ‘‘Browse Your Genome for Clones’’ option, and then the region of interest was zoomed into until all adjacent bacterial artificial chromosome (BAC) clones were shown. Distributor information for each BAC clone was given when selected from the maps. To view cosmid clones available within the region of interest, the ‘‘Component’’ selection on the master map was chosen. Adjacent clones spanning a region of approximately 200 kb from the areas of interest were subsequently selected. Adjacent human BAC clones from the 4p16.3 and 4q35.1 regions for the subtelomeric chromosome 4 probes were purchased from the BACPAC Resource Centre at Childrens Hospital Oakland Research Institute (California, USA). Human cosmid vectors were purchased from the UK Human Genome Mapping Project Resource Centre (Cambridge, UK) to generate the DNMT1 gene-specific probe (accession number AF180682). On arrival, each colony was cultured in sterile LB broth supplemented with 20 Ag/ml chloramphenicol or 25 Ag/ml kanamycin (Sigma, Dorset, UK) for the BAC or cosmid clones, respectively. A 30-ml cell suspension per clone was centrifuged and the BAC or cosmid vector was extracted from the cell pellet using the QIAprep Spin Miniprep Kit (Qiagen, West Sussex, UK) according to manufacturer’s instructions with slight modifications to accommodate the large vector sizes. The volume of the first three solutions used were doubled to ensure complete lysis of all cells within the pellet, and the vector DNA elution recommendations were altered to two applications of 30 Al of distilled water at 70jC left on the columns’ membrane for 1 min before centrifugation. This procedure improved the suspension of the vector and increased subsequent yields. Nick translation One microgram of the BAC or cosmid DNA was labelled by nick translation using the Nick Translation System from Invitrogen (Paisley, UK) according to the manufacturer’s recommendations. However, the incubation time was extended to 2 h 30 min following empirical time series experiments to generate labelled fragments between 200 and 800 bp in size. The BAC clones for the 4p16.3 probe and cosmids for the DNMT1 probe were labelled with fluorescein-dUTP (Roche Diagnostics Ltd., East Sussex) during nick translation, while the 4q35.1 BAC clones were labelled with
For the DNMT analysis, FISH was performed with the commercially available Locus-Specific Identifier (LSI) 2p23 probe (Vysis, Surrey, UK) for DNMT3a, as previously described (Doak et al., 2003). The second probe for this analysis (DNMT1 gene) and the two required for the assessment of chromosome 4 hyperploidy in Barrett’s esophagus were produced in-house. Two hundred nanograms of the BAC or cosmid nick translation products for each region (4p16.3, 4q35.1, DNMT1) were pooled and 20 Cot-1 DNA (Roche Diagnostics Ltd.) was added to each of the three preparations. The resultant probe mixes were precipitated and the subsequent DNA pellets were each resuspended in 5 Al of hybridisation buffer consisting of 50% formamide, 2 SSC, 10% dextran sulphate, 1 Ag salmon sperm and water to provide a 10-Al total volume. Before denaturation, target slides were dehydrated in an ethanol series: 70%, 80% and 95% for 2 min each. Once dried, the slides were denatured in 70% formamide/2 SSC at 75jC for 5 min, quenched in an ice-cold ethanol series (2 min in each of 70%, 80% and 95% ethanol) and subsequently air-dried. The probe was denatured at 75jC for 10 min and then incubated at 37jC for 20 min to allow the preannealing of repetitive sequences. The probes were hybridised to the interphase preparations overnight at 37jC in a humidified chamber. The slides were washed in 0.4 SSC/0.3% NP-40 at 73jC for 2 min, 2 SSC/0.1% NP-40 at ambient temperature for 30 s and left to air-dry in the dark. Nuclei were counterstained with 10 Al of 125 ng/ml DAPI II (Vysis) and slides were viewed under an Olympus BX 50 fluorescence microscope using a UplanF1 100/1.3 oil objective. Probes were visualised with single- and multiple-bandpass filter sets, while images were captured with a cooled CCD (chargedcoupled device) camera and analysed with the MacProbe version 4.1 software (Applied Imaging, Newcastle Upon Tyne, UK). Scoring was performed as previously described in Doak et al. (2003). The mean signal gains or losses plus three standard deviations in control samples were used to establish significant cytogenetic alterations. Consequently, the cutoff percentage of nuclei defining significant abnormal signal losses was 1.6%, 1.6%, 2.2% and 3.1% for the 2p23 (DNMT3a), DNMT1, 4p16.3 and 4q35.1 probes, respectively. The cutoff for abnormal signal gains with the 2p23 and DNMT1 probes were 1.6% and 1.0%, respectively. However, no signal gains were observed in the control samples with the 4p16.3 and 4q35.1 probes.
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
Results FISH optimisation All initial FISH reactions to optimise the hybridisation of the probes produced were performed on metaphase preparations from human male lymphocytes to ensure that they were hybridising to the correct region(s) of the genome. Initial experiments were designed to ascertain the size of the labelled probe required to generate a signal that was sufficiently strong to ensure visualisation in interphase
29
cells, where the chromosomes are not condensed (as in metaphases). Five BAC clones from the 4p16.3 and 4q35.1 regions were selected and purchased. The BAC DNA was extracted, each was individually labelled via nick translation, and then 200 ng of each were pooled to generate probes ranging from 50 to 250 kb, which were subsequently hybridised to metaphase preparations. Figs. 1a and b display the signal intensities obtained when varying the probe size, and it could be consequently concluded that although signals were visible with 100 kb probes, those >150 kb generated easily identifiable signals. Twenty
Fig. 1. FISH probe optimisation and application. (A) Signals at 4q35.3 generated from a 120-kb probe. (B) Signals at 4q35.3 generated from a 200-kb probe. (C) Comparison between the commercially available LSI 2p23 probe (red) and the DNMT1 probe produced in the present study from cosmid clones. (D) Cohybridisation of the 4p16.3 (green) and 4q35.1 (red) locus-specific probes. (E) Scattered 4p16.3 and 4q35.1 signals. (F) More compact 4p16.3 and 4q35.1 signals in interphase nuclei. (G and H) Whole chromosome 4 amplification within Barrett’s interphase nuclei.
30
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
metaphases were examined, all of which displayed probe hybridisation signals in the correct location with no cross hybridisation, thereby demonstrating the high specificity and sensitivity of the probes produced. This procedure was also repeated with the cosmid clones for DNMT1 and a strong signal resulted, which was comparable to commercially available locus-specific FISH probes (Fig. 1c). Before FISH, both the probe and target samples were denatured. This can either be done separately or both probe and target may be denatured together. Both techniques were assessed to determine which provided stronger signals. The probe and metaphase target were denatured individually using a standardised formamide-based technique (VarellaGarcia et al., 1998; von Bergh et al., 2000), while the probe and a second metaphase target were codenatured for 2 min on a 75jC hotplate. Upon examination, signals achieved by individually denaturing the probe and sample was considerably stronger than when using codenaturation. Hence, in all subsequent FISH experiments, the formamide-based technique was utilised. As the degree of the probes’ nonspecific hybridisation is difficult to predict, stringency washes also needed careful adjustment through trial and error to optimise the intensity of the FISH signals. Low to high stringency washes were performed on several slides that had been hybridised with the same probe. Subsequent analysis demonstrated signal intensities were strongest with a relatively low stringency wash that was subsequently used for all subsequent FISH experiments (as detailed in Materials and methods). For the chromosome 4 study, it was more desirable to use multicolour FISH (i.e., simultaneously hybridise both probes designed for the 4p16.3 and 4q35.1 regions) as this method conserved both time and precious samples. It therefore had to be ensured that the two probes produced could be successfully cohybridised without losing signal intensity and quality. To investigate, the probes for the 4p16.3 and 4q35.1 regions were combined, precipitated and resuspended in hybridisation buffer. The subsequent probe was then hybridised to a metaphase preparation and consequently analysed. Cohybridisation of the two probes was successful as can be seen in Fig. 1d, with no compromise to signal size or intensity; hence, the use of multicolour FISH was demonstrated to be feasible with the single-copy sequence probes. Once optimised on metaphase preparations, the probes and techniques developed were tested on interphase nuclei. The probes originating from BAC and cosmid clones were therefore applied to interphase preparations, but a very scattered signal was detected with the BAC probes (Fig. 1e) due to their size and gaps between their mapped positions—a characteristic that could not be seen on the metaphase preparations, where DNA is considerably more condensed. To avoid this scatter, only slightly overlapping clones spanning approximately 150 kb DNA were selected and resulted in stronger and more compact signals in interphase nuclei comparable to commercial probes (Fig. 1f).
Large DNA fragments (such as those cloned within BAC vectors) contain long stretches of repetitive sequences. Before probe hybridisation, these sequences must be blocked to prevent them from annealing to large regions of the target DNA, thereby reducing nonspecific hybridisation. This is achieved by incubation with an unlabelled competitor DNA, which anneals to the repetitive sequences in both the sample and probe, without reducing the probes’ hybridisation efficiency to its intended target. A frequently used block is human cot-1 DNA, a modified form of total genomic DNA that is enriched for the repetitive sequences. Although prehybridisation with cot-1 DNA reduces nonspecific background, the concentration utilised must be adjusted accordingly. When too high, it interferes with probe hybridisation, and when too low the background noise is strong. A series of FISH experiments with cot-1 DNA at 10, 20 or 50 of the total probe quantity were therefore performed. The lowest concentration that gave the least background hybridisation was 20 and hence was utilised in all subsequent FISH experiments. Chromosome 4 hyperploidy This analysis demonstrated that chromosome 4 is indeed hyperploid in Barrett’s esophagus as both the 4p16.3 and 4q35.1 probes had the same copy number in all samples. A majority of the hyperploid nuclei detected had three copies of 4p16.3 and 4q35.1, but in some samples (particularly at the more advanced stages) up to seven copies of each were identified, as illustrated in Figs. 1g and h. Occasionally, nuclei were detected with 3 4p16.3 and 2 4q35.1q signals (or vice versa), but the number of such nuclei within the samples was always very low (a maximum of five nuclei per sample). Hence, inadequate probe hybridisation within these nuclei may have been at fault rather than amplification of just one chromosome arm. Table 1 compares the percentage of aberrant nuclei previously detected with the commercially available CEP 4 (Vysis) centromeric probe (Doak et al., 2003) and with the locus-specific 4p16.3 and 4q35.1 probes generated in the present study. As can be seen, there was a clear positive correlation between the results generated with both probe types. In all but two of the samples, there was no significant difference detected between the proportion of abnormal nuclei identified with the CEP 4 and locus-specific 4p/q probes (using the v2 statistic, P < 0.05). Samples 4 and 22 were the two in which a significant difference was detected; however, both probe types demonstrated chromosome 4 hyperploidy within the sample. It was the proportion of sample cells displaying the abnormality that differed. Cytogenetic status of the DNMT1 and DNMT3a loci The DNMT3a locus was analysed first due to the commercial availability of the probe. However, significant
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33 Table 1 Percentage of cells displaying chromosome 4 aneuploidy using a commercially available probe for its centromeric region (CEP 4; Vysis) and the probes produced for 4p16.3 and 4q35.1 Slide
Pathology
Chromosome 4 loss a
24 26 41 43 120 122 126 134 136 140 4 11 49 86 29 40 55 1 10 91 15 22
Control Control Control Control Control Control Control Control Control Control BM BM BM BM LGD LGD LGD HGD HGD HGD EA EA
b
Chromosome 4 gain
CEP 4
4p/q
CEP 4a
4p/qb
NN NN NN NN NN NN NN NN NN NN 1.2 1.3 1.3 0.0 0.5 0.8 0.7 4.7 3.0 0.3 1.5 1.0
1.1 1.4 1.7 1.8 1.3 1.1 1.4 1.8 0.6 1.1 1.8 1.3 1.1 0.9 0.9 1.3 1.3 1.1 0.8 0.5 1.8 0.7
NN NN NN NN NN NN NN NN NN NN 22.8 0.0 0.0 25.3 25.5 28.5 3.2 1.7 25.3 50.6 32.4 54.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 12.6 0.0 0.0 21.5 23.5 19.8 5.0 2.5 20.3 43.0 28.5 26.2
NN: data were not necessary, as controls were included to examine background hybridisation for the 4p/q probes. a Data were obtained from our previous study (Doak et al., 2003). b Average percentage of 4p16.3 and 4q35.1 signal losses or gains.
copy number alterations were very rarely detected. No samples demonstrated loss of DNMT3a while gains (3 signals) were only observed in four samples (1 metaplasia; 1 LGD; 2 HGD), but these were never noted in more than 3% of cells. Due to the low number of alterations detected at the DNMT3a locus, a preliminary investigation including only the HGDs was analysed with the DNMT1 probe. Again, abnormalities were rare; one sample demonstrated significant signal loss in 4% of nuclei and one sample had 3 DNMT1 signals in 2.5% of nuclei. Due to the lack of alterations at this advanced stage of Barrett’s progression, examination of the earlier grades was not deemed necessary.
Discussion Commercially available single-copy sequence probes are very limited but are valuable for research and are an important tool in identifying new molecular biomarkers for disease. We have therefore detailed the protocols necessary for designing and generating custom locus-specific probes for interphase FISH. In addition, we have demonstrated their use on cytology preparations, such as those commonly utilised in routine diagnostic pathology. This technique for probe generation considerably reduces the costs of FISH when compared to the expense of purchasing commercially
31
available probes. Hence, our results indicate this approach of in-house probe preparation is practical and cost-effective. Long stretches of DNA can be obtained from cloned sequences to generate FISH probes for targeting specific chromosomal regions and a number of vectors with varying insert sizes exist: cosmids—35 to 45 kb; BAC—70 to 100 kb; PAC—100 to 300 kb; YAC—100 kb to 2 Mb. Large quantities of high purity, double stranded DNA can be efficiently isolated from host cells (in which the vector is amplified), and clones containing DNA fragments spanning the whole genome are readily available from sequencing institutions. Cloned DNA was therefore the favoured substrate for the generation of locus-specific probes, but the smaller vectors were chosen (cosmids and BACs), as they are easier to label and contain less redundant DNA. In the first instance, nick translation was used to differentially label human DNA cloned within BAC vectors, originating from the 4p16.3 and 4q35.1 loci. Following optimisation of hybridisation conditions and stringency washes, multicolour FISH was subsequently performed on interphase preparations with known chromosome 4 aberrations, as determined by our previous analysis (Doak et al., 2003). A strong correlation between the data generated by the commercially available Vysis probe (Doak et al., 2003) and those recorded for the 4p16.3 and 4q35.1 locus-specific probes existed. In all cases, the chromosome 4 status (i.e., hyperploid/monosomic/diploid) determined by the two types of probes was concordant. In 10/12 (83%) of the cases examined, no significant differences between the percentages of aberrant cells were detected. For the two cases in which there was a significant difference, the proportion of sample cells displaying the abnormality differed slightly. This variation may have been due to the use of different interphase preparations; therefore, a different population of cells was analysed by each probe type. However, despite this, chromosome 4 hyperploidy was still detected in these samples with both the CEP 4 and 4p/q probes. Hence, the 4p16.3 and 4q35.1 locus-specific probes generated established that the whole chromosome 4 structure was amplified in Barrett’s esophagus. This investigation was additionally aimed at determining the significance of gene dosage imbalances at the DNMT loci during Barrett’s progression, as abnormal methylation patterns have been implicated in Barrett’s tumourigenesis (Eads et al., 2000, 2001). The DNMT enzymes are responsible for establishing and maintaining DNA methylation patterns in eukaryotes (Okano et al., 1998, 1999). Hence, it has been speculated that increases in their activity may be accountable for epigenetic silencing during human carcinogenesis (Antoun et al., 2000; Saito et al., 2003; Vertino et al., 1996). Previous Barrett’s cytogenetic investigations have detected amplifications at the DNMT3b locus (20q11.2; Riegman et al., 2001; Varis et al., 2001; Walch et al., 2000), but reports regarding the DNMT3a and DNMT1 loci are conflicting (Riegman et al., 2001; Van Dekken et al.,
32
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33
1999; Walch et al., 2000), they were therefore assessed in the present study. A FISH probe was commercially available for DNMT3a, but not for the DNMT1 gene; hence, a probe for this region had to be produced. Following the success of the chromosome 4 locus-specific probes, a fluorescently labelled probe for the DNMT1 locus was produced. However, cosmid clones were utilised to generate the probe, as the appropriate BAC clones were not available for the region of interest. Once developed and optimised, this probe, together with the commercially available probe for the DNMT3a locus, was applied to interphase Barrett’s samples with various histological grades to assess the status of the DNMT genes at the chromosomal level. This analysis subsequently determined that no significant copy number changes were present at the DNMT1 and DNMT3a loci. In a small number of cases, gains of the loci were noted, but they were in less than 4% of cells and predominantly arose in HGD. Hence, it is most probable that these observations are a consequence of the general increase in aneuploidy and genetic instability that accompanies Barrett’s neoplastic progression (Croft et al., 2002; Doak et al., 2003; Riegman et al., 2001). In summary, this report details the methodology required to generate single-copy locus-specific interphase FISH probes for any genomic region of interest, hence broadening cytogenetic applications in the research of solid tumours or diseased tissues and making the technique more amenable to diagnostic cytopathology in the future. Two probes generated for the subtelomeric regions of chromosome 4 were used to verify the amplification identified in a previous investigation actually involved the whole chromosome. In addition, a FISH probe was produced and subsequently implemented to facilitate the assessment of DNMT gene dosages during Barrett’s associated neoplastic progression. Copy number alterations of the loci at which the DNMT1 and DNMT3a genes are located were very rarely observed, indicating amplifications or deletions of these genes are unlikely to be responsible for the aberrant DNMT activity observed in Barrett’s esophagus. Alterations in gene expression or the enzymes’ biochemical properties are alternative mechanisms that may drive hypermethylation during Barrett’s tumourigenesis but have yet to be assessed.
Acknowledgment SH Doak was funded by a studentship from the Tenovus cancer charity.
References Ahluwalia, A., Hurteau, J.A., Bigsby, R.M., Nephew, K.P., 2001. DNA methylation in ovarian cancer: II. Expression of DNA methyltransferases in ovarian cancer cell lines and normal ovarian epithelial cells. Gynecol. Oncol. 82, 299 – 304. Antoun, G., Baylin, S.B., Ali-Osman, F., 2000. DNA methyltransferase
levels in altered CpG methylation in the total genome and in the GSTP1 gene in human glioma cells transfected with sense and antisense DNA methyltransferase cDNA. J. Cell. Biochem. 77, 372 – 381. Bian, Y.S., Osterheld, M.C., Fontolliet, C., Bosman, F.T., Benhattar, J., 2002. p16 inactivation by methylation of the CDKN2A promoter occurs early during neoplastic progression in Barrett’s esophagus. Gastroenterology 122, 1113 – 1121. Croft, J., Parry, E.M., Jenkins, G.J.S. et al., 2002. Analysis of the premalignant stages of Barrett’s esophagus through to adenocarcinoma by comparative genomic hybridisation. Eur. J. Gastroenterol. Hepatol. 14, 1179 – 1186. Doak, S.H., Jenkins, G.J.S., Parry, E.M., D’Souza, F.R., Griffiths, A.P., Toffazal, N., Shah, V., Baxter, J.N., Parry, J.M., 2003. Chromosome 4 hyperploidy represents an early genetic aberration in pre-malignant Barrett’s oesophagus. Gut 52, 623 – 629. De Both, N.J., Wijnhoven, B.P.L., Sleddens, H.F.B.M., Tilanus, H.W., Dinjens, W.N.M., 2001. Establishment of cell lines from adenocarcinomas of the esophagus and gastric cardia growing in vivo and in vitro. Virchows Arch.—An International Journal of Pathology 438, 451 – 456. De Marzo, A.M., Marchi, V.L., Yang, E.S., Veeraswamy, R., Lin, X.H., Nelson, W.G., 1999. Abnormal regulation of DNA methyltransferase expression during colorectal carcinogenesis. Cancer Res. 59, 3855 – 3860. Eads, C.A., Lord, R.V., Kurumboor, S.K., Wickramasinghe, K., Skinner, M.L., Long, T.I., Peters, J.H., DeMeester, T.R., Danenberg, K.D., Danenberg, P.V., Laird, P.W., Skinner, K.A., 2000. Fields of aberrant CpG island hypermethylation in Barrett’s esophagus and associated adenocarcinoma. Cancer Res. 60, 5021 – 5026. Eads, C.A., Lord, R.V., Wickramasinghe, K., Long, T.I., Kurumboor, S.K., Bernstein, L., Peters, J.H., DeMeester, S.R., DeMeester, T.R., Skinner, K.A., Laird, P.W., 2001. Epigenetic patterns in the progression of esophageal adenocarcinoma. Cancer Res. 61, 3410 – 3418. Eastmond, D.A., Schuler, M., Rupa, D.S., 1995. Advantages and limitations of using fluorescence in situ hybridisation for the detection of aneuploidy in interphase human cells. Mutat. Res. 348, 153 – 162. El-Deiry, W.S., Nelkin, B.D., Celano, P., Chiu Yen, R.W., Falco, J.P., Hamilton, S.R., Baylin, S.B., 1991. High expression of the DNA methyltransferase gene characterises human neoplastic cells and progression stages of colon cancer. Proc. Natl. Acad. Sci. 88, 3470 – 3474. Jubb, A.M., Bell, S.M., Quirke, P., 2001. Methylation and colorectal cancer. J. Pathol. 195, 111 – 134. Khan, S.M., Pillay, S.P., Papadimos, D., Yong, J.W.K., Roberts, H.J.V., Crawford, D.H., 1997. Barrett’s esophagus: a technique for the culture of Barrett’s oesophageal cells. J. Gastroenterol. Hepatol. 12, 606 – 611. Klump, B., Hsieh, C.-J., Holzmann, K., Gregor, M., Porschen, R., 1998. Hypermethylation of the CDKN2/p16 promoter during neoplastic progression in Barrett’s esophagus. Gastroenterology 115, 1381 – 1386. Mizuno, S., Chijiwa, T., Okamura, T., Akashi, K., Fukumaki, Y., Niho, Y., Sasaki, H., 2001. Expression of DNA methyltransferases DNMT1, 3A and 3B in normal hematopoiesis and in acute and chronic myelogenous leukaemia. Blood 97, 1172 – 1179. Okano, M., Xie, S., Li, E., 1998. Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat. Genet. 19, 219 – 220. Okano, M., Bell, D.W., Haber, D.A., Li, E., 1999. DNA methyltransferases DNMT3a and DNMT3b are essential for de novo methylation and mammalian development. Cell 99, 247 – 257. Palanca-Wessels, M.C., Barrett, M.T., Galipeau, P.C., Rohrer, K.L., Reid, B.J., Rabinovitch, P.S., 1998. Genetic analysis of long-term Barrett’s esophagus epithelial cultures exhibiting cytogenetic and ploidy abnormalities. Gastroenterology 114, 295 – 304. Popescu, N.C., Zimonjic, D.B., 1997. Molecular cytogenetic characterisation of cancer cell alterations. Cancer Genet. Cytogenet. 93, 10 – 21. Riegman, P.H.J., Vissers, K.J., Alers, J.C., Geelen, E., Hop, W.C.J., Tilanus, H.W., Van Dekken, H., 2001. Genomic alterations in malignant transformation of Barrett’s esophagus. Cancer Res. 61, 3164 – 3170.
S.H. Doak et al. / Experimental and Molecular Pathology 77 (2004) 26–33 Rockett, J.C., Larkin, K., Darnton, S.J., Morris, A.G., Matthews, H.R., 1997. Five newly established esophageal carcinoma cell lines: phenotypic and immunological characterization. Br. J. Cancer 75, 258 – 263. Saito, Y., Kanai, Y., Nakagawa, T., Sakamoto, M., Saito, H., Ishi, H., Hirohashi, S., 2003. Increased protein expression of DNA methyltransferase (DNMT) 1 is significantly correlated with the malignant potential and poor prognosis of human hepatocellular carcinomas. Int. J. Cancer 105, 527 – 532. Van Dekken, H., Vissers, C.J., Tilanus, H.W., Tanke, H.J., Rosenberg, C., 1999. Clonal analysis of a case of multifocal esophageal (Barrett’s) adenocarcinoma by comparative genome hybridisation. J. Pathol. 188, 263 – 266. Varella-Garcia, M., Gemmill, R.M., Rabenhorst, S.H., Lotto, A., Drabkin, H.A., Archer, P.A., Franklin, W.A., 1998. Chromosomal duplication accompanies allelic loss in non-small cell lung carcinoma. Cancer Res. 58, 4701 – 4707. Varis, A., Puolakkainen, P., Savolainen, H., Kokkola, A., Salo, J., Nieminen, O., Nordling, S., Knuutila, S., 2001. DNA copy number profiling in esophageal Barrett’s adenocarcinoma: comparison with gastric
33
carcinoma and esophageal squamous cell carcinoma. Cancer Genet. Cytogenet. 127, 53 – 58. Vertino, P.M., Yen, R.W.C., Gao, J., Baylin, S.B., 1996. De novo methylation of CpG island sequences in human fibroblasts over-expressing DNA (cytosine-5) methyltransferase. Mol. Cell. Biol. 16, 4555 – 4565. von Bergh, A., Emanuel, B., van Zelderen-Bhola, S., Smetsers, T., van Soest, R., Stul, M., Vranckx, H., Schuuring, E., Hagemeijer, A., Kluin, 2000. A DNA probe combination for improved detection of MLL/1 1q23 breakpoints by double-color interphase-FISH in acute leukemias. Genes Chromosomes Cancer 28, 14 – 22. Walch, A.K., Zitzelsberger, H.F., Bruch, J., Keller, G., Angermeier, D., Aubele, M.M., Mueller, J., Stein, H., Braselmann, H., Siewert, J.R., Hofler, H., Werner, M., 2000. Chromosomal imbalances in Barrett’s adenocarcinoma and the metaplasia – dysplasia – carcinoma sequence. Am. J. Pathol. 156, 555 – 566. Werner, M., Ludwig, W., Aubele, M., Nolte, M., Zitzelsberger, H., Komminoth, P., 1997. Interphase cytogenetics in pathology: principles, methods and applications of fluorescence in situ hybridisation (FISH). Histochem. Cell Biol. 108, 381 – 390.