|Home | About | Journals | Submit | Contact Us | Français|
The Karpas-620 human myeloma cell line (HMCL) expresses high levels of Cyclin D1 (CCND1), but has a der(8)t(8;11) and a der(14)t(8;14), and not a conventional t(11;14). Fluorescent in situ hybridization (FISH) and array comparative genomic hybridization (aCGH) studies suggest that der(14)t(11;14) from a primary translocation underwent a secondary translocation with chromosome 8 to generate der(8)t(8;;11) and der(14)t(8;;14). Both secondary derivatives share extensive identical sequences from chromosomes 8, 11, and 14, including MYC and the 3′ IgH enhancers. Der(14), with MYC located ~700 kb telomeric to the 3′ IGH enhancer, expresses MYC. By contrast, der(8), with both CCND1 and MYC repositioned near a 3′ IGH enhancer, expresses CCND1, which is telomeric of the enhancer, but not MYC, which is centromeric to the enhancer. The secondary translocation that dysregulated MYC resulted in extensive regions from both donor chromosomes being transmitted to both derivative chromosomes, suggesting a defect in DNA recombination or repair in the myeloma tumor cell.
Multiple myeloma (MM) includes both primary and secondary IGH translocations [1,2]. Primary IGH translocations, which are thought to occur when B lymphocytes pass through germinal centers, appear to be mediated mostly by errors in IGH switch recombination, with breakpoints occurring within or near IGH switch regions; but less often they appear to be mediated by errors in somatic hypermutation, with breakpoints occurring within or near JH regions [2,3]. These mostly simple reciprocal translocations involve five recurrent partner chromosomal loci and oncogenes: 11q13 (Cyclin D1 (CCND1)), 15%; 6p21 (CCND2), 3%; 4p16 (MMSET and FGFR3), 15%; 16q23 (MAF), 5%; and 20q12 (MAFB), 2%. Secondary IGH translocations, which can occur during any stage of tumor progression, rarely occur near or within JH or IGH switch regions, and typically are complex rearrangements or insertions. Genomic rearrangements involving a MYC gene (C-»N-»L-), which occur as a very late event during tumor progression, provide the best-studied examples of secondary IGH translocations in MM [4,5].
The Karpas-620 human myeloma cell line (HMCL)  expresses high levels of CCND1 RNA (M. Kuehl, unpublished). The triploid karyotype includes unbalanced translocations that involve chromosomes 8, 11, and 14, with two copies each of der(8)t(8;11)(q24;q13), der(14)t(8;14)(q24;q32), and der(14)t(1;14)(q11;q32), plus other marker chromosomes, all of which were present in the original tumor cells [6,7]. By fluorescent in situ hybridization (FISH) analyses, we have shown previously that MYC co-localizes with IGH sequences on der(14), whereas both CCND1 and MYC co-localize with IGH sequences on der(8) [5,8].
Hypothesizing that there might have been a primary t(11;14) translocation, followed by a complex secondary translocation that involved MYC, we identified and cloned an 11;14 translocation breakpoint fragment. We then performed FISH and array comparative genomic hybridization (aCGH) mapping studies to elucidate the structures of the rearranged chromosomes that contain MYC or CCND1 sequences. In addition,we determined which chromosomes express MYC and CCND1 RNA.
Karpas-620 cells (108) were fused with SP2/0 mouse plasmacytoma cells (5 × 107), and somatic cell hybrids were selected in RPMI 1640 medium supplemented with 10% fetal calf serum, 0.1 mM hypoxanthine, 0.4 µM aminopterin, 16 µM thymidine and 1 µM ouabain . There were 60 positive wells, all of which contained mouse myc, including all possible expression patterns for human MYC and CCND1 RNA. Subclones of the hybrids were made by limiting dilution in 24 well macrotiter plates. Somatic cell hybrid clones and subclones that contain the human MYC gene were identified by a competitive PCR reaction. A human exon 2 primer (GTCAAGCTTAGACTGCCTCCCGCTTTGTGT) and mouse intron 1 primer (TTGGAAGTACAGCACGCTGAA), together with a human/mouse primer (GTAGTCGAGGTCATAGTTCCTGTT), generated 105 and 184 bp products fromthe human and mouse genomic DNA, respectively. A competitive RT.PCR assay also was used to determine the expression of human MYC RNA. An exon 2 human/mouse primer (CCAGGACTGTATGTGGAGCG), together with exon 3 human (GAGGTTTGCTGTGGCCTCCAG) and mouse (TGTGTGTCCGCCTCTTGTCG) primers, generated 482 and 685 bp products, respectively. An RT.PCR assay specific for human CCND1 used two oligonucleotides (CTGGCCATGAACTACCTGGA and GTCACACTTGATCACTCTGG) that generated a 483 bp product. The presence of Karpas-620 chromosomes – normal 8 and 11, der(8)t(8;11), and der(14)t(8;14) – in the somatic cell hybrid clones and subclones were determined by FISH analyses.
The RNA FISH experiments to determine the sites of CCND1 RNA expression were done using a previously described protocol . A plasmid containing a 16 kb genomic DNA segment that included the CCND1 gene was used as a probe .
Three color FISH analyses were done on metaphase chromosomes from Karpas-620, using previously described procedures [5,8]. The EG3-5 probe includes sequences that are 20–35 kb centromeric to the Eα2 3′ IGH enhancer . The CH probe, which includes sequences that hybridize with both the Eα1 and Eα2 3′ IGH enhancers, and most other probes were described and referenced previously . Details regarding BAC probes used for mapping the CCND1 and MYC loci are available upon request.
Agilent Human Genome Microarray Kit 244A (Agilent Technologies, Santa Clara, CA, USA) and Agilent Custom 4x44k arrays were used in array CGH analysis. Agilent 244A array contains 236000 60-mer oligonucleotide probes for coding and non-coding genomic sequences with a median distance between probes 8.9 kb. Custom44k arrays were designed by selecting 39895 non-overlapping oligonucleotide probes from Agilent High Density CGH and ChIP Database Probes spanning 10 genomic regions of interest with an average distance 195 bp between probes (Design 016569; details are available on request).DNA labeling with Cy5- and Cy3-dUTP and array hybridization were done according to the Agilent “Oligonucleotide Array-Based CGH for Genomic DNA Analysis Protocol” (Version 5.0, Publication Number: G4410-90011 V.5.1, November 2007), starting with 3 µg of AluI- and RsaI-digested Karpas-620 and human male DNA. Arrays were scanned and data were extracted from array images by using Agilent Feature Extraction Software (version 9.0) with default settings. Datawere imported into R-2.6.0 language and statistical computing environment , and analyzed using snapCGH package  of BioConductor-1.9 . Background intensity for each spot was removed from foreground hybridization intensity using “minimum” method; data were transformed to log2 ratio of Cy5- and Cy3-signals, and normalized within array by subtracting median log2 ratio. Presumptive rearrangement points were determined with two segmentation algorithms that showed good concordance: Circular Binary Segmentation algorithm (as implemented in DNAcopy package [16,17]) and Heterogeneous Hidden Markov Model (BioHMM  available in the aCGH package ).
A Southern blot assay revealed that Karpas-620 DNA lacks sequences immediately 3′ of Sμ, and has only germline Sα sequences (not shown) . However, it has two different fragments that contain legitimate 5′Sμ-3′Sγ switch recombination junctions, 6.5 and 8.5 kb SphI fragments (Fig. 1B). A second Southern blot identified 13 and 16 kb BamHI fragments with a 5′Sμ probe, but only the 16 kb fragment hybridized with 5′σμ or JH probes (Fig. 1C). This suggested that the 13 kb fragment included an illegitimate recombination event that occurred within the 1.6 kb separating the distal ends of the 5′Sμ and 5′σμ probes (Fig. 1A). The 13 and 16 kb BamHI fragments were cloned into lambda DASH, selected, and sequenced. A legitimate SμSγ1 junction was identified in each case. The 16 kb fragment contained only IGH sequences, but the 13 kb fragment contained an 11;14 breakpoint (Fig. 1D, E) that was located 880 bp telomeric (5′) to Sμ, 343.2 kb centromeric to CCND1, and ~530 bp telomeric to MYEOV (11q13:68818198-68821329, NCBI build 35).
We identified four copies of CCND1 in Karpas-620, one each on two copies of der(8)t(8;11) and two copies of der(1)t(1;11)(q32;q13) (Table 1) . RNA FISH analyses showed that CCND1 RNA is expressed only from the two CCND1 loci that are associated with IGH sequences, i.e., der(8)t(8;11) (Fig. 2). Unfortunately, we were unable to identify MYC expression by RNA FISH assays. Instead, we made somatic cell hybrids between Karpas-620 and a mouse plasmacytoma cell line, and then analyzed hybrid subclones in an attempt to correlate specific chromosome content with human MYC and CCND1 RNA expression. As expected from the RNA FISH results, CCND1 RNA was expressed only by hybrid subclones that contain der(8)t(8;11) (data not shown). There are three kinds of chromosomes in Karpas-620 that contain the MYC gene [normal 8; der(8)t(8;11); der(14)t(8;14)], with MYC associated with IGH sequences in the latter two (Table 1) [7,8]. Hybrids containing only normal human chromosome 8 do not express human MYC RNA, a result that is consistent with the hypothesis that MYC is not expressed from a normal germline allele in most – if not all – HMCL and mouse plasmacytoma [4,22]. Somewhat surprisingly, however, hybrid subclones that contain only der(8) also do not express human MYC, whereas those that contain der(14) do express human MYC RNA (Table 2).
We did FISH mapping studies to determine the anatomy of rearranged chromosomes that contained sequences from chromosomes 8, 11, and 14. We hoped that this might shed some light on the events that created these rearranged chromosomes, but might also help us understand why MYC RNA is not expressed from der(8)t(8;11). Representative results (Table 1) led to the following conclusions. First, both der(8)t(8;11) and der(14)t(8;14) contain sequences from the MYC, CCND1, and IGH loci. Therefore, despite our inability to detect sequences from all three chromosomes using whole chromosome painting probes, these translocations are more correctly designated as der(8)t(8;;11) and der(14)t(8;;14). Second, there are sizable regions of chromosomes 8, 11, and 14 that are represented on both der(8) and der(14). This indicates that the process that generated these chromosomes was not simple reciprocal rearrangements but included duplication of extensive sequence from each chromosome. Finally, results with probes from the CCND1 locus suggest that the 11q13 breakpoint in der(11)t(11;13)(q13;q14) shares a simple, reciprocal relationship with the 11q13 breakpoint on der(8) and der(14).
We used Agilent 244K and custom arrays to determine the DNA content of the regions on chromosomes 8, 11, and 14 that include the breakpoints and duplicated sequences that are present on der(8)t(8;;11) and der(14)t(8;;14). Fig. 3A shows that there is a nearly 50% increase in DNA copy number of 828 kb of chromosome 8 sequences, including MYC plus sequences 357 kb centromeric and 471 kb telomeric to MYC. Fig. 3B shows that there is an approximately 50% increase in DNA copy number of 267 kb of chromosome 11 sequences that start at the cloned 11;14 breakpoint (estimated as 341 kb centromeric to CCND1 by array CGH, in good agreement with the cloned breakpoint that is 343.2 kb centromeric to CCND1), and extend to within 74 kb of CCND1. Fig. 3C shows a nearly 33% decrease in copy number of sequences that are ~300 kb centromeric to the Eα2 3′ IGH enhancer, suggesting that the duplicated region extends about 475 kb centromeric from the cloned 11;14 breakpoint that is located upstream from SμSγ1.
On der(14)t(8;;14), MYC appears to be located about 670 kb telomeric to the Eα1 3′ IGH enhancer, which is well within the distance over which this enhancer can act in plasma cell tumors . On der(8)t(8;;11), the CCND1 appears to be located nearly 400 kb telomeric to the Eα1 enhancer, whereas the MYC is located about 770 kb centromeric to the Eα2 enhancer. The lack of expression of MYC on der(8) might be explained by the possible occurrence of insulator sequences immediately centromeric to the Eα2 3′ IGH enhancer (E. Max, personal communication). However, a more likely explanation is that there are five genes (HOLE, CRIP1, CRIP2, MTA1, and PACS2) located between the Eα2 3′ IGH enhancer and the centromeric chromosome 14 breakpoint that is located about 470 kb from MYC (Fig. 4B). In any case, this is an example in which juxtaposition of an oncogene close to IG enhancer sequences is not associated with dysregulation of the oncogene.
As depicted in Fig. 4A, the structures of der(8)t(8;;11) (q24;q32;q13), der(14)t(8;;14)(q24;q13;q32), and t(11;13) (q13;q14), as determined by a combination of FISH mapping and array CGH, suggests that a primary t(11;14) translocation was followed by secondary translocations on both derivatives. The secondary translocation that converted der(11)t(11;14) to der(11)t(11;13) appears to be a simple albeit unbalanced translocation, lacking der(13). In contrast, the other secondary translocation converted der(14)t(11;14) to der(8)t(8;;11) and der(14)t(8;;14) by a process that included duplication of extensive sequences on chromosomes 8 (828 kb), 11 (267 kb), and 14 (475 kb) (Fig. 4B). One possible explanation is duplication of the donor chromosomes (e.g., G2), with double strand breaks at different sites in each of the four chromosomes, followed by heterologous joining of four ends and loss of the other four ends. Another possibility is that there were widely separated single strand breaks on both strands of der(14)t(11;14) and of chromosome 8 (arrowheads in Fig. 4B), followed by a repair process that duplicated extensive sequences at the two breakpoint ends of each chromosome, and then joining of heterologous ends. A third possibility is the involvement of a break-induced replication (BIR) mechanism, although it is difficult to envisage precisely how this might have happened [23,24]. It is thought that B cell specific DNA modification mechanisms (VDJ recombination, somatic hypermutation, and IgH switch recombination) are inactive in normal and tumor plasma cells [2,3]. Therefore, secondary MYC translocations, which occur as late progression events in myeloma tumors, are thought to be mediated by the poorly characterized mechanisms responsible for unbalanced translocations in most kinds of tumors. Even though the mechanism(s) responsible for this unusual secondary translocation are not clear, the occurrence of extensive regions of duplication transmitted from both donor chromosomes to both derivative chromosomes suggests the occurrence of a defect inDNA recombination or repair  in the myeloma tumor cell that gave rise to the Karpas-620 cell line.
This research was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research; and NIH grantsCA83208 and EB2060 (R.H.S.). The authors would like to acknowledge Leslie Brents for technical contributions, Ana Gabrea for help in preparation of the manuscript, and also Keith Caldecott, James Haber, and Richard Kolodner for helpful discussions.
Conflict of interest statement
The authors declare that there are no conflicts of interest.