Retroviral vector-mediated overexpression of c-myc in embryonic bursal precursors induces multi-staged tumorigenesis beginning with preneoplastic-transformed follicles (TF) and progressing to clonal metastatic B-cell lymphomas. Using a 13K chicken cDNA microarray, specifically enriched for chicken immune system expressed sequence tagged (ESTs), we carried out array-based comparative genomic hybridization (array-CGH) and detected significant DNA copy number change at many loci on most or all chromosomes in both early TF and end-stage lymphomas. Formation of long palindromes, through breakage–fusion–bridge cycles, is thought to play an early role in gene amplification. Employing genome-wide analysis of palindrome formation (GAPF), we detected extensive palindrome formation in early TF and end-stage lymphomas. The population of loci showing amplification by array-CGH was enriched for palindromes detected by GAPF providing strong evidence for genetic instability early in Myc-induced tumorigenesis and further support for the role of palindromes in gene amplification. Comparing gene copy number change and RNA expression changes profiled on the same cDNA array, we detected very little consistent contribution of gene copy number change to RNA expression changes. Palindromic loci in TF and tumors, however, were expressed, many at high levels, suggesting an abundance of RNA species with long double-stranded segments generated during tumorigenesis.
Deregulated expression of c-myc is implicated in many human neoplasms and experimental animal tumor systems (Grandori et al., 2000). Avian leukosis virus-induced lymphomas in the bursa of Fabricius, reported nearly 100 years ago (Ellerman and Bang, 1908), are classical Myc oncogene-mediated tumors that progress from preneoplastic-transformed bursal follicles (TF) to clonal (Neiman et al., 1980) metastatic lymphomas (Neiman, 1994). Experimentally, retroviral vector-mediated constitutive expression of Myc in transplantable bursal stem cells has been employed to generate transformed follicles and end-stage lymphomas for analysis of several key early neoplastic phenotypes, including expansion of stem cell populations (Neiman et al., 1985; Thompson et al., 1987), altered programmed cell death (Neiman et al., 1991; Brown et al., 2004) and angiogenesis (Brandvold et al., 2000). Genetic instability is thought to be important early in the development of neoplasms. Gene amplification and gross chromosomal instability has been reported in cultured fibroblasts subjected to conditional overexpression of c-myc (Mai et al., 1996a, 1996b; Felsher and Bishop, 1999). A bursal lymphoma-derived cell line, DT40, shows modal karyotype differences and cytogenetic instability with time and media conditions (Chang and Delany, 2004).
The development of chicken cDNA microarrays (Neiman et al., 2001; Burnside et al., 2005), together with other tools made available by release of a draft chicken genome sequence (Hillier et al., 2004), allowed a genome-wide analysis of instability in this experimental system. Two different array-based techniques were employed to examine changes in DNA, and correlated with RNA expression change, in both preneoplastic TF and end-stage metastatic lymphomas. Comparative genome hybridization using cDNA microarrays (array-CGH) (Pollack et al., 1999) has been used successfully in detecting and mapping DNA copy number change including gene amplification even at small increases in copy number. The second technique, genome-wide analysis of palindrome formation (GAPF), detects long palindromes in DNA thought to be important early intermediates in gene amplification (Tanaka et al., 2005). Applying these approaches, we detected extensive gene copy number change and palindrome formation early in the development of Myc-induced bursal lymphomas. We mapped chromosomal positions of palindromes and amplifications in TF and lymphomas and analysed the relationship between copy number change, palindromes and RNA expression patterns.
Array-comparative genome hybridization analysis of DNA copy number change during Myc-induced bursal lymphomagenesis
Preneoplastic bursal TF and derivative lymphomas metastatic to the liver were induced in in-bred histocompatible SC-1 chickens with an Myc-transducing defective retroviral vector generally as described previously (Neiman et al., 2001). The principal change was the use of a modified LmycSN defective vector in which the v-myc gene was replaced by a c-myc gene obtained from an RCAS vector (Hughes et al., 1987), the promoter driving c-myc replaced by a promoter from the long terminal repeat (LTR) of the recovered avian myelocytomatosis virus HB1 (Enrietto et al., 1983) and the selectable neo gene replaced by a green fluorescence protein marker. We have found that this HB1-derived promoter is highly active in chicken bursal cells (data not shown). This modified vector, called LGH-myc, is diagramed in Figure 1.
Embryonic day 15 (E15) bursal cells were infected ex vivo with LGH-myc and used to reconstitute cyclophosphamide-ablated bursas of recipient birds (Eskola and Toivanen, 1974). By 4 weeks after hatching, the only transplantable stem cells in reconstituted bursa were Myc-transformed TF cells (Neiman et al., 1985). Reconstitution of ablated bursas in secondary recipients with cells from these bursas produced TF by 4 weeks and metastatic lymphomas by 4–8 weeks, which were used to prepare DNA and RNA for microarray experiments.
Reference DNA for array-CGH was prepared from a normal bursa reconstituted with E15 embryonic cells transduced with an empty defective vector (LGH-X). Experimental samples from metastatic lymphomas dissected from the livers of three birds were separately analysed and included measurements for both dye-labeling orientations. Transformed follicle samples from six different transplant experiments were used to test for early-stage copy number change. Variation in the array-CGH assay was estimated by independently comparing two different preparations of DNA from normal reconstituted bursas against the common reference. The background subtracted, log2-transformed data sets were averaged (mimicking the processing of the reversed dye pairs in the tumor samples) and a standard deviation (s.d.=0.13) was calculated. This protocol was slightly modified to accommodate the single-chip TF results, where an s.d.=0.2 was employed to estimate the assay variation for a single chip. In both cases, gains or losses at a given locus were scored significant if the normalized fluorescence log2 ratio exceeded 4 × s.d.
With respect to sex chromosomes, female chickens are hemizygous for Z (ZW) whereas males are homozygous for Z. We began by comparing fluorescence ratios from cDNAs mapped to the Z-chromosome and those from autosomes on the same array in array-CGH experiments from the three different tumor samples from female birds using the same reference male normal DNA. The results of log2 fluorescence ratios, shown in Figure 2, indicate the expected median value of about −0.5 for Z chromosome genes compared to autosomal genes in these female samples. These results indicated that, at least with respect to average behavior of large numbers of loci, our array-CGH experiments were sensitive to single-copy changes. However, to avoid complications in analysis when reference and experimental sample were of different sex, we limited the subsequent analysis to autosomal loci.
Table 1 summarizes the copy number changes we detected by array-CGH in preneoplastic TF and end-stage metastatic tumors. Gains or losses at a given locus were scored significant if the normalized fluorescence log2 ratio exceeded 4 s.d. – the threshold established from the variation estimate described above. As enumerated in the table, substantial numbers of loci with copy number change were detected in both TF and tumor DNA. Most of the amplifications were low level, ranging from 1 to 5 extra copies. Importantly, copy number change of individual genes appeared to be largely a random process. There was considerable variation in numbers of affected loci in different samples, and little or no correlation between amplification and deletion. We detected only a handful of expressed sequence tagged (EST) loci that were changed, that is, either amplified (16) or deleted (24), in common among the three tumors, and no loci were found in common between TF samples, and between TF and tumor DNA. Based on chromosome position assignments of ESTs from the current chicken genome sequence database (Hillier et al., 2004), we mapped the positions of apparently amplified loci, which appear on all autosomes for which sequence has been determined. Figure 6a shows an example for chromosome 1. Identification, annotation and fluorescence ratios for the loci amplified or deleted in common among the three tumors are given in Supplementary Table 1.
Genome-wide analysis of palindrome formation during tumorigenesis
Long inverted repeats or palindromes have been detected by virtue of their ability to form ‘snap-back’ DNA and found in well-characterized tumor-specific amplicons, for example, at the c-myc locus (Ford and Fried, 1986). Palindrome formation is a conserved, rate-limiting step in eukaryotic gene amplification (Yasuda and Yao, 1991; Butler et al., 1996; Tanaka et al., 2002). Recently, a new technique for genome-wide detection of DNA palindromes called GAPF has demonstrated widespread occurrence of palindromes in human cancer cells and provided further evidence that such structures may lead to gene amplification (Tanaka et al., 2005). We therefore employed a modified version of GAPF as a second approach to analysing genomic instability in this experimental system.
As an internal control for the formation of snap-back DNA in our version of GAPF, we added to DNA samples used for analysis a one-copy equivalent of a plasmid construct carrying an inverted 1.8 kb repeat of a human oxidative stress response gene (OSR-1) cDNA sequence, (pTOI; Figure 3a). Eight human OSR-1 cDNA spots were added to different areas of the chicken cDNA array. As depicted in Figure 4, DNA from the three tumors and three of the TF samples was heat denatured, cooled to allow snap-back formation and single-stranded DNA degraded with Mung Bean nuclease. Remaining double-stranded snap-back product was digested with DPN-11 and ligated to DPN-11 primers for expansion by polymerase chain reaction (PCR) with Cy-5 dye-labeled primers. Reference normal DNA, to which a control OSR-1 direct repeat construct was added (pTOD, Figure 3a), was processed in the same way using Cy-3 dye-labeled primers, mixed with the tumor or TF snap-back PCR product and hybridized to the chicken cDNA microarray.
Figure 3b gives representative GAPF results for eight OSR-1 spots on the array. The table shows a 12–25-fold increase of OSR-1 signal from the pTOI construct over the low-level signal from the pTOD control direct repeat, indicating successful snap-back formation and PCR amplification from a one-copy palindrome in this experiment. We required at least a 12-fold increase of pTOI signals over pTOD for GAPF experiments to be included in this analysis. Figure 5 depicts two representative scan images of six of the 31 array blocks in GAPF experiments with a tumor and a TF sample. The red Cy-5 signals indicate palindrome formation in each of these samples and not in the normal reference DNA. Genome-wide analysis of palindrome formation experiments with reference DNA and other normal DNA samples in place of tumor or TF DNAs gave virtually no signals (not shown). The GAPF technique used, that is, the microarray and use of one restriction enzyme, is not saturating. Therefore, the presence of some long DNA palindromes in normal chicken DNA is not excluded.
These experiments indicated widespread palindrome formation in all of the tumor and TF DNA samples tested. As was carried out for the amplified loci, we mapped the palindromes that we detected in the three tumor samples on chicken autosomes, depicted for chromosome 1 in Figure 6a. Figure 6b shows a table enumerating numbers of GAPF-detected palindromes and array-CGH-detected amplified loci in each of the three tumor samples and the number of loci scoring in both assays. Only a limited number of palindromic loci showed amplification by array-CGH, as described previously (Tanaka et al., 2005). The failure to detect palindromes at amplified loci with our GAPF technique would be expected for a number of reasons including false negatives owing to the requirement for DPN-11 sites. Nevertheless, there was a statistical association (using a hyper-geometric distribution probability, not shown) between array-CGH detected amplified loci and palindromes consistent with previous reports (Tanaka et al., 2005). Both tumor-associated amplicons and palindromes have been described as extending over 100 kb or more (Looney and Hamlin, 1987; Ma et al., 1993; Feo et al., 1996; Mangano et al., 1998). We therefore extended the analysis to detect regional clustering of results for the two kinds of analyses. Each chromosome was segmented into consecutive 100 kb units (blocks). Genome-wide analysis of palindrome formation and array-CGH data for each tumor were co-evaluated and scored positive if one or more loci within a block were called positive for both palindromes by GAPF and locus amplification by array-CGH. That is, overlap resolution was expanded from the gene to the block level. To address the significance of a co-occurrence within a given block, data from each tumor was randomly redistributed onto our chromosome map and repeated for 100 000 iterations. A positive association between palindromes and amplification was identified for cases where the tumor data scored positive for a given block and the block probability, P⩽0.05. As also depicted in Figure 3b, significant clusters of palindromes and amplified loci were mapped on each chicken chromosome for each of the three tumors. Finally, the figure shows a Venn diagram enumerating the numbers of these clusters and sharing between each of the tumors. Our results support the notion that palindromes can be substrates for gene amplification, but detected limited, if any, selection for these events during tumor progression in this model
RNA expression change and relationship to DNA copy number change
A large number of investigators, using expression microarray profiling and other techniques, have reported extensive transcriptional change in response to Myc (Zeller et al., 2003), characterized by increased expression of many genes known to influence both cell growth and proliferation. We have reported similar expression profiles in TF and end-stage lymphomas in this experimental system (Neiman et al., 2001, 2003). The development of our expanded chicken cDNA microarray, coupled with technology for RNA amplification from limited tissue samples (Van Gelder et al., 1990), provided an opportunity to extend our previous analysis of RNA expression change and determine the relationship of such changes to alteration of gene copy number in the same experimental samples. Reference amplified RNA (aRNA) was prepared from E15 bursas, representing the target embryonic bursal cell population for Myc-induced neoplasia. Inter-array variability for our expression profiling assay was determined by co-hybridizing the E15 reference preparation against a second E15 preparation and assessing the distribution of the filtered, normalized log2 ratio values. Accordingly, an initial ±2-fold (s.d.>1.3) threshold was applied to all valid log2 ratio values. To further enhance the stringency of the assay, we subsequently required that this threshold be exceeded in all five biologically independent TF vs E15 reference expression profile data sets. Changes in expression level potentially owing to partial maturation of Myc-transformed E15 target cells were assessed by microarray comparison of E15 expression with E16 and E18 normal bursa, and genes where the relative expression levels exceeded the ±2-fold threshold, we removed from the final TF signature. Shown in the Venn diagram in Figure 7, there remained a robust expression signature common to the TF preparations predominantly with respect to upregulated genes. The identities and functions of known genes with altered expression in TF are similar to previously reported profiles (Neiman et al., 2001) and recorded in Supplementary Table 2.
We then selected aRNA from one of the TF preparations to use as a reference for determining the profile of expression change from preneoplastic TF to lymphomas metastatic to the liver. RNA from dissected tumors from the same three experimental birds used for array-CGH and GAPF were compared with the reference TF aRNA in duplicate with Cy-dye swaps. In this case, the intra-array technical variation was estimated from a co-hybridization of the TF reference RNA to itself and we further assessed the distribution of the filtered, normalized log2 ratio values. As before, an initial ±2-fold (s.d.>5.6) threshold was applied to the appropriate log2 ratio values. Furthermore, we required the additional criteria that the relative fold-change threshold be exceeded in at least five of the six metastatic tumors vs TF reference expression profile data sets (i.e., three independent tumor samples assayed separately using both labeling orientations). Moreover, to minimize artifacts from contaminating normal liver cells, relative gene expression that exceeded ±2-fold in liver in comparison with the TF reference were filtered by microarray expression analysis of normal chicken liver. As shown in Figure 7, a smaller number of genes showed consistent increased expression in metastatic tumors compared with TF cells, whereas a strikingly increased number of genes appeared to show consistent decreased expression during tumor progression. These were different than the upregulated set seen in TF formation, and contained examples of genes known to be involved in apoptotic cell death and in tumor suppression. Supplementary Table 2 records gene annotation for this signature as well.
Of relevance to this report, we compared these results with the array-CGH and GAPF results. One striking distinction is the detection of consistent patterns of gene expression change during Myc-induced tumorigenesis compared with the similarly prevalent but more random palindrome formation and gene copy number change. Given this observation, it was not surprising that we found little or no consistent relationship between DNA copy number and expression. Figure 8 is a data scatter plot relating copy number change and RNA expression level in five TF samples. The results for one TF sample (solid symbols) show some genes in which both copy number and expression level varied together, for example, amplified and overexpressed genes appeared in the upper right box. However, when date from all five samples were averaged (open symbols), few if any genes scored in this way because of the lack of consistent copy number change for genes among the five experimental samples. In contrast, genes scoring for over- and underexpression in this analysis were not eliminated by averaging because of the consistent expression changes detected in the TF samples (Figure 7). The figure supports our conclusion that consistent changes in expression of very few genes were directly owing to gene amplification or deletion.
Expression of palindromic loci
The formation of long inverted repeats could have effects on transcription of affected loci, or if transcribed, on the function of product RNA. Figure 9 is a box plot of data from three metastatic tumors that compares transcription levels from palindromic loci with transcription from the rest of the genes on the same array. The results indicate that, as a group, transcription units with palindromes expressed RNA and did so at levels in the same range as those from other loci represented on the chip. Although palindromic loci are expressed, our analysis did not compare RNA expression levels before and after palindrome formation. There was, however, no overlap between the list of genes with diminished RNA expression during tumor progression and the loci with palindromes that we detected. Thus, direct inhibition of transcription by DNA palindromes was not responsible for the transcription repression signature.
However, another functionally significant consequence of transcription of palindromes is possible. Hairpin structures would be likely to form in such transcripts, suggesting that RNAs with long double-stranded segments were generated during tumorigenesis. The data indicate that numerous loci generated such novel RNA structures. As neither the array nor the method of detection are known to be saturating for all palindromes formed or expressed, our data probably underestimate the abundance of such novel segments of double-stranded RNA.
Genetic instability, including regional genomic amplification, is a common tumor phenotype thought to contribute to neoplastic progression (Tlsty, 1997; Lengauer et al., 1998). Our array-CGH analysis of this Myc oncogene-mediated experimental tumor system is consistent with preceding cytogenetic studies (Mai et al., 1996a, 1996b; Felsher and Bishop, 1999), demonstrating that such changes occur in preneoplastic stages in vivo, and adds to the body of work indicating widespread genetic change early in the pathogenesis of many forms of cancer. We have not carried out cytogenetic analyses of instability on these in vivo preneoplastic and neoplastic tissues. However, such studies have recently been carried out on a cultured myc-induced bursal lymphoma, which demonstrate chromosomal instability (Chang and Delany, 2004). Furthermore, array-CGH and GAPF analyses of the same cell line yielded results consistent with the published cytogenetic changes and indistinguishable from those described here for in vivo lymphomas (Neiman et al., 2006).
By extending the analysis to palindrome formation, we confirmed the observation of this widespread structural change and its linkage to gene amplification, recently reported in human tumors (Tanaka et al., 2005). The use of a qualitative assay for structural change in DNA relevant to tumor development should be a powerful tool for further analysis in models and early cancer detection in humans. The structure of palindromic loci arising during tumorigenesis would be of interest. For example, pre-existing short inverted repeats have been shown, in model experiments in animal cells, to nucleate long palindrome formation at the sites of DNA strand breaks (Tanaka et al., 2002). Palindrome containing amplicons at the dhfr locus in Chinese hamster ovary cells (Looney and Hamlin, 1987; Ma et al., 1993) and the c-myc locus in HL-60 human leukemia cells (Feo et al., 1996; Mangano et al., 1998) have been cloned and characterized. It remains to be determined if palindromic loci arising early in the formation of these myc-induced neoplasms are as complex as was found in these pioneering studies.
The question arises, however, of the relationship of palindrome formation and copy number change to the mechanism by which Myc induces cancer in this model system. At the level of individual genes, there appears to be little evidence that changes in RNA expression are directly caused by copy number change. Thus, Myc-induced bursal lymphomas may be caused primarily by the direct and indirect effects of Myc as a pleiotropic transcription factor. Such a conclusion begs the question of whether or not Myc-induced genetic instability plays a greater role in the many clinical and other experimental tumor systems in which deregulated myc expression occurs.
Whether or not there is a direct effect of copy number instability on Myc-induced neoplasia in any given tumor system, the detection of RNA expression from widespread DNA palindromes formed early during tumorigenesis raises a further novel possibility. Although not demonstrated directly in this analysis, a strong possibility exists that many RNA species bearing novel duplex segments are generated as a result of this structural change. Such RNA duplexes could have functional consequences. They could, for example, activate interferon pathway responses and/or directly inhibit translation of tumor suppressors. In addition, long well-matched segments of double-stranded RNA could well be processed into functional siRNAs. Normally occurring micro-RNAs have been implicated in the regulation of important processes relevant to neoplasia such as cell division (Kittler et al., 2004), as well as serving as targets of c-Myc regulation (O'Donnell et al., 2005) and providing expression profiles for the classification of human tumors (Lu et al., 2005). In this case, we are suggesting the generation of novel siRNAs during tumor progression, some of which might target important genes for repression. The large number of genes with diminished expression, which we detected during tumor progression in this model (Figure 7), provides a candidate list for experimental detection of this possible novel mechanism in tumorigenesis.
Materials and methods
Tumor induction, tissue processing and preparation of DNA and RNA
Normal reference and control nucleic acids were extracted from embryonic bursas from E15, E16, E18 and E21 (hatch day) inbred SC-1 white leghorn eggs (and from bursa and liver from 2-week-old post-hatch chickens). Induction of preneoplastic TF and metastatic tumors was carried out as described (Neiman et al., 2001). Briefly, E15 embryonic bursal cells were co-cultivated overnight with PG-13 packaging cells (Miller et al., 1991) producing the defective LGH-myc-transducing virus (Figure 1), and 106 infected cells injected intravenously into E18 embryos with cyclophosphamide-ablated bursas (Eskola and Toivanen, 1974). After 4 weeks, reconstituted bursas with TF were harvested and used to reconstitute ablated secondary recipients, producing bursas with essentially pure populations of Myc-transformed follicles (Neiman et al., 1985; Thompson et al., 1987) 4–8 weeks after transplantation. These bursas were extracted as were metastatic lymphomas dissected from livers in the same birds. Histological confirmation was carried out on all experimental samples.
Tissue samples were divided for RNA and DNA extraction. Total RNA was prepared with RNA-Bee reagent (Tel-Test Inc., Friendswood, TX, USA) as per the manufacturer's instructions. Amplified RNA was prepared with a MessageAmp kit (Ambion, Austin, TX, USA) as per the manufacturer's instructions. DNA was prepared as described (Neiman et al., 1975).
13K chicken cDNA microarray
The cDNA glass-slide microarray used for these studies has been described in detail (Burnside et al., 2005). Briefly, the array was based on an earlier 3500 cDNA microarray composed of ESTs from the bursal lymphoma-derived cell line DT-40, and a number of chicken immune tissue-derived cDNA libraries (Neiman et al., 2001, 2003). This base clone set was expanded with about 9000 non-overlapping ESTs from diverse embryonic and adult chicken tissues from the Biotechnology and Biological Sciences Research Council (BBSRC) chicken cDNA project (Boardman et al., 2002) to include 13 007 usable features and 11 447 non-overlapping ESTs. With respect to genes expressed in bursa, 7433 features gave detectable signals (signal/noise >3-fold) with Cy dye-labeled bursal aRNA.
Array-comparative genome hybridization
For array-CGH, DNA was digested overnight with DPN-11 to produce fragments ranging from about 500 to 1500 bp, cleaned with a QIAquick PCR purification kit (Qiagen, Valencia, CA, USA), ethanol precipitated and adjusted to a concentration of 0.4 μg/μl in elution buffer (10 mM Tris, pH 8.0). Five micrograms of DPN-11 fragments from normal reference DNA were labeled with Cy-3 and experimental sample DNA with Cy-5 mixed and hybridized to the cDNA array generally as described (Pollack et al., 1999). The array-CGH hybridization, wash and scan parameters employed in this study are identical to those used to generate expression profiles (Neiman et al., 2001 and 2003). For each array-CGH measurement, spot intensity signals were filtered and removed if the values did not exceed 3 s.d. above the background signal in at least one signal channel or if a spot was flagged as questionable by the GenePix Pro software. Spot-level ratios (Cy-5/Cy-3) were log2 transformed and normalized using a using a block-level Lowess methodology (Yang et al., 2002).
Genome-wide analysis of palindrome formation
For construction of control direct and inverted repeat plasmids pTOD and pTO1 (Figure 3a), the OSR-1 cDNA (Genbank Accession No. R98985) cloned in pT7T3D with a Pac1 and EcoR1-modified polylinker (GE Healthcare Life Sciences, Piscataway, NJ, USA) was digested with Pac1, blunt-ended with Klenow DNA polymerase and ligated in either orientation to a blunt-ended, 1.6 kb. AlfII – EcoR1 fragment of OSR-1 from the same plasmid.
One copy per genome equivalent of pTOI was added to 100 ng of experimental sample (TF or tumor) genomic DNA in 10 μl of 0.1 M NaCl final volume, and the same amount of pTOD was added to 100 ng of reference normal DNA. Each DNA was then denatured at 100°C for 10 min, rapidly cooled on ice, for 1.5 min and digested with 8 U of Mung Bean nuclease (New England Biolabs, Beverly, MA, USA) in 16 μl × 1 nuclease buffer at 30°C for 30 min. Note that digestion conditions require optimization with different batches of Mung Bean nuclease. Reaction volume was expanded to 100 μl with 0.01 M Tris, pH 7.5, extracted with 100 μl phenol and chloroform 1:1 (vol:vol), and precipitated with ethanol in using a glycogen carrier (Roche, Applied Biosciences, Indianapolis, IN, USA). The pellet was digested for 5 min at 37°C with 200 U of DPN-II in 20 μl × 1 buffer plus bovine serum albumin, and the enzyme was inactivated at 65°C for 15 min.
DPN-11 linker ligation was carried out using DPN-11 linker oligonucleotides: 23 mer, 5′-IndexTermTTCACTACACACCTCATCCTTCT-3 and 12 mer, 5′-IndexTermGATCAGAAGGA-3 heated to 100°C and annealed by cooling to room temperature. Ligation reaction mix with 300 pM of annealed primers and 5 U of T4 DNA ligase (Invitrogen, Carlsbad, CA, USA), in 10 μl of ligase buffer, was added to the inactivated DPN-11 digests and incubated for 5 min at room temperature. For linker-mediated PCR, 300 pM of the 23-mer oligonucleotide labeled with either Cy-3 for reference DNA sample or Cy-5 for experimental samples was added to a PCR reaction mix in a final volume of 70 μl with 5 U of Taq polymerase (Invitrogen), preincubated at 72°C for 3 min and denatured at 94°C for 3 min. For amplification, 28 PCR cycles of 94°C for 0.5 min, 64°C for 0.5 min and 72°C for 2 min were carried out in a thermal cycler (Perkin-Elmer Model 9600) followed by 72°C for 5 min. Cy dye-labeled PCR products from reference and experimental samples were mixed, processed and hybridized to the 13K cDNA microarray as described for array-CGH above.
Five micrograms of aRNA from reference and experimental samples were labeled using random primers by indirect Cy dye labeling, mixed and hybridized to the cDNA microarray as described previously (Neiman et al., 2003). Array imaging was performed using the GenePix 4000B Microarray Scanner. Image quantification was performed using GenePix Pro v3.0 image analysis software. For each array, spot intensity signals were filtered and removed if the values did not exceed 3 s.d. above the background signal in at least one signal channel or if a spot was flagged as questionable by the GenePix Pro software. Spot-level ratios (Cy-5/Cy-3) were log2 transformed, and global loess normalization was applied using S-Plus (MathSoft, Cambridge, MA, USA).
Boardman E, Sanz-Ezquerro J, Overton IM, Burt DW, Bosch E, Fong WT et al. (2002). Curr Biol 12: 1965–1969.
Brandvold KA, Neiman P, Ruddell A . (2000). Oncogene 19: 2780–2785.
Brown CY, Bowers SJ, Loring G, Heberden C, Lee R-m, Neiman PE . (2004). Dev Comp Immunol 28: 619–634.
Burnside J, Neiman P, Tang J, Basom R, Talbot R, Aronszajn M et al. (2005). BMC Genom 6: 13.
Butler DK, Yasuda LE, Yao MC . (1996). Cell 87: 1115–1122.
Chang H, Delany ME . (2004). Chromosome Res 12: 299–307.
Ellerman V, Bang O . (1908). ZBL Bakt Parisit Infekt Hygiene 46: 595–609.
Enrietto PJ, Payne LN, Hayman MJ . (1983). Cell 35: 369–379.
Eskola J, Toivanen P . (1974). Cell Immunol 13: 459–471.
Felsher DW, Bishop JM . (1999). Proc Natl Acad Sci USA 96: 3940–3944.
Feo S, Di Liegro C, Mangano R, Read M, Fried M . (1996). Oncogene 13: 1521–1529.
Ford M, Fried M . (1986). Cell 45: 425–430.
Grandori C, Cowley SM, James LP, Eisenman RN . (2000). Annu Rev Cell Dev Biol 16: 653–699.
Hillier LW, Miller W, Birney E, Warren W, Hardison RC, Ponting CP et al. (2004). Nature 432: 695–716.
Hughes S, Greenhouse JJ, Petropoulos CJ, Sutrave P . (1987). J Virol 61: 3004–3012.
Kittler R, Putz G, Pelletier L, Poser I, Heninger AK, Drechsel D et al. (2004). Nature 432: 1036–1040.
Lengauer C, Kinzler KW, Vogelstein B . (1998). Nature 396: 643–649.
Looney JE, Hamlin JL . (1987). Mol Cell Biol 7: 569–577.
Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D et al. (2005). Nature 435: 834–838.
Ma C, Martin S, Trask B, Hamlin JL . (1993). Genes Dev 7: 605–620.
Mai S, Fluri M, Siwarski D, Huppi K . (1996a). Chromosome Res 4: 365–371.
Mai S, Hanley-Hyde J, Fluri M . (1996b). Oncogene 12: 277–288.
Mangano R, Piddini E, Carramusa L, Duhig T, Feo S, Fried M . (1998). Oncogene 17: 2771–2777.
Miller AD, Garcia JV, vonSuhr N, Lynch CM, Wilson C, Eiden M . (1991). J Virol 65: 2220–2224.
Neiman P, Burnside J, Elsaesser K, Hwang H, Clurman BE, Kimmel R et al. (2006). In: Buerstedde J-M (ed). Reviews and Protocols in DT40 Research: Subcellular Biochemistry. Springer: Dordrecht, (in press).
Neiman P, Payne LN, Weiss RA . (1980). J Virol 34: 178–186.
Neiman P, Wolf C, Enrietto PJ, Cooper GM . (1985). Proc Natl Acad Sci USA 82: 222–236.
Neiman PE . (1994). Adv Immunol 56: 467–484.
Neiman PE, Grbic JJ, Polony TS, Kimmel R, Bowers SJ, Delrow J et al. (2003). Oncogene 22: 1073–1086.
Neiman PE, Purchase HG, Okazaki W . (1975). Cell 4: 311–319.
Neiman PE, Ruddell A, Jasoni C, Loring G, Thomas SJ, Brandvold KA et al. (2001). Proc Natl Acad Sci USA 98: 6378–6383.
Neiman PE, Thomas SJ, Loring G . (1991). Proc Natl Acad Sci USA 88: 5857–5861.
O'Donnell KA, Wentzel EA, Zeller KI, Dang CV, Mendell JT . (2005). Nature 435: 839–843.
Pollack JR, Perou CM, Alizadeh AA, Eisen MB, Pergamenschikov A, Williams CF et al. (1999). Nat Genet 23: 41–46.
Tanaka H, Bergstrom DA, Yao MC, Tapscott SJ . (2005). Nat Genet 37: 320–327.
Tanaka H, Tapscott SJ, Trask BJ, Yao MC . (2002). Proc Natl Acad Sci USA 99: 8772–8777.
Thompson CB, Humphries EH, Carlson LM, Chen C-LH, Neiman PE . (1987). Cell 51: 371–381.
Tlsty TD . (1997). Curr Top Microbiol Immunol 221: 37–46.
Van Gelder RN, von Zastrow ME, Yool A, Dement WC, Barchas JD, Eberwine JH . (1990). Proc Natl Acad Sci USA 87: 1663–1667.
Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J et al. (2002). Nucleic Acids Res 30: e15.
Yasuda LF, Yao MC . (1991). Cell 67: 505–516.
Zeller KI, Jegga AG, Aronow BJ, O'Donnell KA, Dang CV . (2003). Genome Biol 4: R69.
We thank Alana Ruddell, Brian Freie, Mark Groudine and Steve Tapscott for helpful comments. This work was supported by NIH Grants R01 CA-20068 and R01 CA-109365 to PEN.
About this article
Cite this article
Neiman, P., Kimmel, R., Icreverzi, A. et al. Genomic instability during Myc-induced lymphomagenesis in the bursa of Fabricius. Oncogene 25, 6325–6335 (2006). https://doi.org/10.1038/sj.onc.1209646
- Myc oncogene
- gene amplification
- DNA palindromes
PARP inhibitors enhance replication stress and cause mitotic catastrophe in MYCN-dependent neuroblastoma
The MRN complex is transcriptionally regulated by MYCN during neural cell proliferation to control replication stress
Cell Death & Differentiation (2016)
Transformation, genomic instability and senescence mediated by platelet/megakaryocyte glycoprotein Ibα