|Home | About | Journals | Submit | Contact Us | Français|
Mouse models for hepatocellular carcinoma (HCC) provide an experimental ground to dissect the genetic and biological complexities of human liver cancer and enhance our ability to gain insights into the relevance of candidate cancer genes. We examined, using spectral karyotyping (SKY) and array-based CGH (aCGH), seven cell lines derived from HCC spontaneously developed in transgenic c-Myc animals (c-Myc), and four cell lines established from tumors induced in nude mice by inoculation with the original c-Myc cells (nuMyc). All of those cell lines exhibited gain of material from chromosomes 5, 6, 8, 10, 11, 15 and 19, and DNA copy-number loss from chromosomes 2, 4, 7, 9, 12, 14, and X. In addition, several recurrent chromosome reorganizations were found including del(3), t(3;8), del(4) t(4;11), t(6;5), del(7), del(8), del(9), t(10;14), del(11) and del(16). Chromosome breakpoints underlying rearrangements clustered in the regions previously identified as important for the early stages of c-Myc induced hepatocarcinogenesis. The results strongly suggest the importance of recurrent breakage and loss of chromosomes 4, 9 and 14, and gain of chromosomes 15 and 19 in mouse liver neoplasia. Genomic changes observed in c-Myc HCC cell lines are also recurrent in HCC developed in a different other transgenic mouse models, in mouse spontaneous HCC and derivative cell lines, as well as in preneoplastic liver lesions induced with chemical carcinogens. Overall, our results demonstrate selective, non-random genomic changes involving chromosomal regions homologous to those implicated in human HCC.
Cancer arises from a single precursor cell by cumulative acquisition of multiple genetic and epigenetic alterations. Genomic reorganizations are recognized as important in the pathogenesis of cancer, as a large number of recurrent and specific structural alterations resulting in oncogenic stimulation or in loss of tumor suppressive functions have been identified in leukemias, lymphomas, and solid tumors [1–3]. In the past years, there was much progress in the development of fluorescence in situ hybridizations (FISH)-based molecular cytogenetic methods, such as comparative genomic hybridization (CGH), spectral karyotyping (SKY) and high-resolution array-based CGH (aCGH), which allowed identification of several chromosomal regions of recurrent genomic imbalances in human hepatocellular carcinoma (HCC), involving genes implicated in regulatory pathways that control genomic integrity, cell proliferation and differentiation .
Increasing availability of animal models for various disorders point to their effectiveness as tool to identify and characterize critical pathogenic modifications relevant to human disease. In particular, mouse models of tumor suppressor genes and oncogenes are extremely useful to investigate various aspects of cancer genetics and preclinical therapeutic evaluation. Tumor-prone transgenic mice with uniform genetic background offer the distinct opportunity for identifying chromosome and gene alterations associated with early stages of neoplastic development and tumor progression.
In one such example, independent analyses of liver cancer development in WHV/c-Myc and c-Myc/TGFα transgenic mouse models [5–7] have clearly demonstrated the high oncogenic impact of the c-Myc gene activation. However, the examinations of the genomic changes in liver tumors from genetically engineered animals are scarce [8, 9]). In this study, we combined SKY and aCGH to dissect genomic alterations in HCC cell lines (c-Myc cells) established from tumors spontaneously developed in transgenic mice overexpressing c-Myc oncogene controlled by albumin enhancer/promoter. In addition, to further analyze these cells and see if they elicit tumorigenicy, retain their karyotype, or if emerging new populations are selected under in vivo environment, we inoculated nude mice with cells from all c-Myc lines and from thus caused tumors derived several new cell lines (nuMyc). We found that explanted nuMyc cells, like original c-Myc cells, displayed similar non-random chromosome alterations, but also exhibited an increased DNA copy-number changes. Comparison of our data and previously published findings [10, 11] revealed that a limited number of chromosomal changes are selected during liver tumor development
Seven HCC cell lines (604D.02.T1.CL1, 604D02T1CL4, 604D02T2CL1, 620A03T2CL3, 614B02T1CL4, 648A01T1CL3 and 654A03T1) were isolated and cloned in our laboratory from spontaneous HCC developed in c-Myc transgenic mice described elsewhere [5, 7]. In addition, four cell lines, named nuMyc-1, nuMyc-3, nuMyc-4 and nuMyc-6, were derived from tumors developed in athymic nude mice after s.c. transplantation of the original HCC cell lines, 654A03T1, 604D02T1CL4, 604D02T2CL1 and 620A03T2CL3, respectively.
Chromosome preparations obtained from exponentially growing cultures were used for SKY hybridization according to a standard protocol with minor modifications as previously described in detail . Six to ten complete sets of karyotypes, i.e. spectral, classified and G-banded, were established and used for analysis of each line. Changes, both structural and numeric, observed in two or more karyotypes, were considered as recurrent for a given line. Numerical changes were determined based on ploidy level and representation in chromosome rearrangements. The composite karyotype of each cell line was established according to official cytogenetic nomenclature .
For aCGH, mouse genomic DNA was isolated with a QIAamp DNA Mini Kit according to manufacturer protocol (Qiagen, Valencia, CA). Test and reference (Promega, Madison, WI) DNAs were labeled with Cy3 or Cy5 fluorescent dyes (Pharmacia, Piscataway, NJ) according to BioPrime Array CGH Genomic Labeling protocol (Invitrogen, Carlsbad, CA) and cleaned using Microcon YM-30 filters (Millipore, Billerica, MA). Hybridization was carried out using Mouse Genome CGH Microarray 44A from Agilent Technologies (Santa Clara, CA) according to CGH Procedures for Genomic DNA Analysis (Agilent Technologies). Slides were hybridized for 20 hours, washed, scanned with an Agilent microarray scanner, and data were analyzed using Feature Extraction® and CGH Analytics® software packages (Agilent Technologies). Dye-reversal experiments, with reciprocal labeling of the test and reference DNA, were performed for each experiment to ensure the test’s reliability.
All seven cell lines isolated from HCC developed in c-Myc transgenic mice (c-Myc cells) were aneuploid with chromosome number ranging from hypodiploid (32) to hypertetraploid (87). Cell lines derived from tumors grown after inoculation of nude mice with original c-Myc cell lines (nuMyc cells) were also aneuploid but ranged from hyperdiploid (41) to hypersextaploid (137). Composite karyotypes (cp) of seven c-Myc and four nuMyc cell lines are presented in Tables 1 and and2,2, respectively.
The advantage of combining SKY and aCGH to examine the genomic alterations in these HCC cell lines is clearly illustrated in Figure 1. Generally, the comparison of the data obtained through SKY and aCGH analyses showed a remarkable concordance and enabled precise annotation of gains and/or losses, such as small DNA gains in proximal regions of chromosomes 1 and 11 in 604D.02.T2.CL1 (Fig. 1a, b), or increased DNA copy number in the distal region of chromosome 5 in 620A.03.T2.CL3 resulting from an increased number of derivatives of chromosome 6 bearing translocated chromosome 5 material (Fig. 1c, d).
Both the original (c-Myc) and secondary (NuMyc) cells exhibited aneuploidy and a variety of structural chromosomal rearrangements which, taken together, contributed to abnormalities in DNA copy numbers involving several chromosomes. Interestingly, very similar pattern of chromosomal gains and losses characterized both types of cell lines (Table 3); gain of material from chromosomes 5, 6, 8, 10, 11, 15 and 19, and DNA copy-number loss from chromosomes 2, 4, 7, 9, 12, 14, and X. Most frequent gains involved chromosomes 6, 10, 11, 15 and 19; in one cell line we identified 6–20 double minutes (DM) originating from chromosome 19. Losses occurred repeatedly on chromosomes 4, 9, 14 and X.
In addition to DNA copy-number changes, we identified a number of structural chromosomal rearrangements (Fig. 2) such as del(3), t(3;8), del(4) t(4;11), t(6;5), del(7), del(8), del(9), t(10;14), del(11) and del(16). Reciprocal translocation t(5;6)(G2;G1) was invariably present in all cell lines although number of copies of individual derivative chromosomes 5 and 6 varied. Distal part of chromosome 5 in derivative chromosome 6 is the site of c-Myc transgene integration . In addition, there were two major deletions such as frequent unbalanced recurrent translocations t(3;8)(F1;D1), t(4;11)(D2;B5) with variant t(4;11)(D2;B1B5), and t(10;14)(B1;E3). Deletions of chromosomes 4, 8, 9 and 11 frequently occurred at different breakpoints;breakpoints on chromosome 4 clustered in regions A, C and D (Fig. 3) whereas chromosomes 8, 9 and 11 broke frequently in regions B-C, A-B and A-B, respectively. The breakpoints in translocations involving chromosomes 4 and 11 occurred at some of the sites observed in those chromosomes’ respective deletions, namely within the regions 4D2 and 11B4, respectively, whereas breakpoint at chromosome 14 was identified as 14E3.
To identify chromosome changes associated with tumor progression, nude mice were s.c. injected with the original c-Myc cells from all lines, and thus initiated tumors were used to isolate and establish secondary nuMyc cell lines. SKY and aCGH analysis of those cells revealed that a proportion of numerical and structural alterations observed in the original (c-Myc) cells remained preserved in the secondary (nuMyc) cells as well. Yet, secondary cells (Fig. 4b, d, f) showed an increased incidence of DNA copy-number changes compared to the corresponding original c-Myc cell lines (Fig. 4a, c, e) and in some cases, almost a completely new pattern of gains and/or losses (Fig. 4a, b).
Not surprisingly, they also lost some of the preexisting, and exhibited a few novel structural rearrangements such as translocations t(1;8), t(2;9), t(4;5) and t(14;19), which were not detected in the original c-Myc cells. Two examples of this pattern of changes are presented in Table 4.
The molecular cytogenetic analysis of cell lines derived from HCC developed spontaneously in c-Myc transgenic mice, and from nude-mice tumors caused by HCC cells injection, identified chromosome changes nonrandomly involving specific chromosomes and chromosome sites which corresponded to the breakpoints described in early hepatic lesions . Partial or complete gain of chromosomes 15 and 19 and loss of chromosomes 4, 9, and 14 were the most common alterations found in the c-Myc HCC cell lines. The same alterations were also demonstrated in HCC from WHV/c-Myc transgenic mice and HCC induced by chemical carcinogens by CGH and loss of heterozygosity (LOH) analyses [11, 14, 15]). More importantly, nonrandom and specific alterations detected in our and other studies involve chromosome regions containing cancer-related genes in human HCC. Together these findings suggest that a limited number of genomic alterations selected in the majority of tumors are likely to be critical in hepatocarcinogenesis.
The most common gain involves chromosome 15 that has large parts of homology with human chromosome 8q. Gain of 8q is highly recurrent in human HCC as well as in several other cancers. Invariably, the CGH pattern of chromosome 8 consists of the long arm gain and frequent amplification of the distal region where c-Myc gene is located . Since both original (c-Myc) and secondary (nuMyc) cell lines express the c-Myc transgene, the increased copy number of chromosome 15 might enhance the effect of this powerful oncogene in the development of tumors. Gain at 8q24 and c-Myc gene amplification in human or murine cancer cells, including HCC is often due to the isochromosome formation, Robertsonian translocations or complex rearrangements of 8q [17–19]). Recently, a comprehensive study using a “mosaic” mouse model for HCC demonstrated that knockdown of tumor suppressor gene DLC1 cooperates with c-Myc to induce HCC .
Gain of chromosome 19 is an early recurrent alteration during chemically induced hepatocarcinogenesis in the mouse as well as in the primary HCC and preneoplastic liver lesions [14–15]. Chromosome 19 carries susceptibility genes for liver tumors and possibly oncogenes involved in hepatocarcinogenesis . One of the examined c-Myc HCC cell lines contained 3–4 copies of normal chromosomes 19 and 5 to 25 double minute chromosomes (DM) derived from chromosome 19. DM and homogeneously stained regions are commonly associated with the acquisition of drug resistance and tumor progression and manifest DNA amplification [22, 23]. We microdissected and generated PCR probe from DM and localized the sites of DNA amplification at regions 19B1.1-B1.3 and 19D1-D2 . These regions harbor MXI1 and ms-Myc gene that interfere with c-Myc gene function, and RelA, a gene recognized to be important in regulation of malignant cell growth, survival and transformed phenotype [24–26]. We found that ms-Myc gene is located on chromosome X and not on chromosome 19 as was previously thought 
Deletion and LOH of chromosomes 4, 9 and 14 are highly recurrent events in HCC developed in both c-Myc and in WHV/c-Myc transgenic mice . Deletions of chromosome 4 occur at different regions. The distal portion of chromosome 4 was the most frequently deleted region. A similar pattern was observed in mouse breast tumors . Recurrent LOH of chromosome 4 affecting a region homologous to human chromosome 9 was detected in diethylnitrosoamine–induced-C3H/MSM mouse HCC . Mouse chromosome 4 and human chromosome 1 suppresses the malignant phenotype in cell fusions of normal and tumor cells [30, 31]). The mouse orthologues of at least three genes that are lost or mutated in several human cancers map to chromosome 4: the type I TGF-β receptor gene (TGFR1), the tightly linked genes for the p15INK4b (CDKN2B) and p16INK4a/p19ARF (CDKN2A) cyclin-dependent kinase inhibitors, and the p18INK4c cyclin-dependent kinase inhibitor gene (CDKN2C). The cyclin-dependent kinase inhibitors CDKN2/p16, p15 and p19 are deleted in murine HCC [32, also reviewed in 33].
Loss of chromosome 9B-ter was common in both original c-Myc cell lines and secondary nuMyc cells. Deletion of the same region has been observed in chemically induced liver tumors isolated from B6C3F1 mice . The distal portion of mouse chromosome 9 is homologous to human 3p, a region with the highest fragility in human genome encompassing the FHIT gene  The FHIT gene is expressed in normal human hepatic cells and is not expressed or is abnormally expressed in human and rodent primary HCC or HCC cell lines, suggesting the importance of the FHIT alterations in HCC development [36–38]. The transforming growth factor beta II receptor (TGF-β receptor II), mapped to mouse chromosome 9 and human 3p, is another candidate tumor susceptibility gene [39, 40]). The expression of TGF-β II receptor was found to be downregulated in human HCC and in 80% of the c-Myc/TGF-β tumors [41, 42]. Among bona fide and candidate tumor suppressor genes, on mouse chromosome 9 and human chromosome 3p, is also TMEM7 (transmembrane protein 7, RTP3) that shares substantial sequence homology with human and mouse 28-kDa interferon-alpha (IFN-α). Ectopic expression of TMEM7 in HCC lines inhibits HCC cell proliferation, colony formation, and cell migration in vitro and reduces tumor formation in nude mice. [43, 44].
Deletion of chromosome 14D-ter was the second most common deletion in c-Myc tumor cell lines. In contrast, deletion 14 was not observed in the early lesions isolated from the c-Myc/TGF-α HCC , suggesting that, similar to other mouse malignancies , deletion 14 occurs at later stages of tumor development. Region of deletions 14D-ter is homologous to human chromosome regions 8p21-23 and 13q14. Both regions are recurrently underrepresented in HCC  and the region 8p21-23, in particular, harbors several candidate and bona fide tumor suppressor genes [46, 47].
Significantly, near the reciprocal translocation t(5;6) with the breakpoint near the integration site of c-Myc transgene, which was present in all analyzed c-Myc and nuMyc cell lines, we identified a new gene . Analysis of the genomic DNA flanking the c-Myc transgene revealed that the integration of the transgene had induced an approximately a 40-kb deletion that included the first exon of a new gene encoding GTF1IRD1 (general transcription factor II-I repeat domain containing protein 1). Human GTF2IRD1 is one of the 16 genes present in approximately 1.6 Mb interval commonly deleted in Williams-Beuren syndrome (WBS). Therefore, the c-Myc transgenic mouse carrying the t(5;6) translocation, represents the first “knockout” of one of the genes present in the WBS critical region. GTF2IRD1 encodes a widely expressed, multi-functional helix-loop-helix transcription factor, which binds to pRb, though its role in carcinogenesis remains elusive .
In conclusion, our results underline the value of mouse models of HCC for identification of non-random and recurrent genomic alterations that are homologous to human and harbor a limited set of genes relevant to hepatocarcinogenesis
This research was supported by the Intramural Research Program of the National Cancer Institute, NIH.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.