Comprehensive molecular cytogenetic characterization of cervical cancer cell lines

We applied a combination of molecular cytogenetic methods, including comparative genomic hybridization (CGH), spectral karyotyping (SKY), and fluorescence in situ hybridization (FISH), to characterize the genetic aberrations in eight widely used cervical cancer (CC) cell lines. CGH identified the most frequent chromosomal losses including 2q, 3p, 4q, 6q, 8p, 9p, 10p, 13q, and 18q; gains including 3q, 5p, 5q, 8q, 9q, 11q, 14q, 16q, 17q, and 20q; and high‐level chromosomal amplification at 3q21, 7p11, 8q23–q24, 10q21, 11q13, 16q23–q24, 20q11.2, and 20q13. Several recurrent structural chromosomal rearrangements, including der(5)t(5;8)(p13;q23) and i(5)(p10); deletions affecting chromosome bands 5p11, 5q11, and 11q23; and breakpoint clusters at 2q31, 3p10, 3q25, 5p13, 5q11, 7q11.2, 7q22, 8p11.2, 8q11.2, 10p11.2, 11p11.2, 14q10, 15q10, 18q21, and 22q11.2 were identified by SKY. We detected integration of HPV16 sequences by FISH on the derivative chromosomes involving bands 18p10 and 18p11 in cell line C‐4I, 2p16, 5q21, 5q23, 6q, 8q24, 10, 11p11, 15q, and 18p11 in Ca Ski, and normal chromosome 17 at 17p13 in ME‐180. FISH analysis was also used further to determine the copy number changes of PIKA3CA and MYC. This comprehensive cytogenetic characterization of eight CC cell lines enhances their utility in experimental studies aimed at gene discovery and functional analysis. © 2003 Wiley‐Liss, Inc.


INTRODUCTION
Cervical cancer (CC) is the second most common gynecologic malignancy in the world in both incidence and mortality. Infection with human papilloma virus (HPV) is considered an initiating factor in CC carcinogenesis (Brisson et al., 1994). Although 95% of the patients with precancerous lesions harbor HPV, only a small fraction of the cases eventually progress to invasive cancer (zur Hausen, 2000). Therefore, HPV infection alone was considered insufficient for malignant conversion, suggesting a role of other genetic changes in the development of CC. Identification of such genetic alterations is critical for our understanding of the molecular basis of CC development. Thus, cell lines provide an excellent and continuous source of tumor material for such a genetic characterization.
Conventional cytogenetic analysis has identified several recurrent structural chromosomal changes involving chromosomes 1, 3, 5, 17, and the X chromosome in CC (Atkin, 1997). A precise cytogenetic characterization in CC was not feasible because of the difficulty in identifying complex chromosome changes in primary tumors. Recently, however, several molecular cytogenetic techniques, such as fluorescence in situ hybridization (FISH), comparative genomic hybridization (CGH), and spectral karyotyping (SKY)/multicolor-FISH (m-FISH), have been used to illustrate the precise genetic aberrations in tumor genomes (Kallioniemi et al., 1992;Schrö ck et al., 1996;Speicher et al., 1996;Singh et al., 2001;Rao et al., 2002). Several cell lines derived from patients with CC are available that provide a renewable source of material for genetic analysis and that have been used in various studies. Therefore, complete cytogenetic characterization of CC lines will provide useful genetic information for multiple studies.
Here, we characterized eight widely used CC cell lines by G-banding, CGH, SKY, and FISH, and we present their comprehensive molecular cytogenetic profiling. These cell lines have previously not been subjected to detailed cytogenetic characterization. We identified several chromosomal regions that may be important in tumor de-velopment. Losses of 2q and gains of 3q, 5p, and 8q were significant. Notably, the combined approach of CGH, SKY, and FISH analysis defined the origin of 2q loss and gain of 3q, 5p, and 8q. In addition, we identified several amplified chromosomal regions: 3q21, 7p11, 8q23-q24, 10q21, 11q13, 16q23-q24, 20q11.2, and 20q13.1. Through FISH, we have also shown the integration of HPV16 sequences in three of the eight cell lines. This comprehensive characterization of these cell lines augments their utility in investigations targeted at gene discovery and functional analysis in CC.

Cell Lines
The cell lines (Ca Ski, C-33A, C-4I, HT-3, ME-180, MS751, SiHa, and SW756) were obtained from the American Type Culture Collection (ATCC, Manassas, VA), and were grown according to the supplier's recommendations. Chromosome preparations and G-banded karyotypes were made by standard protocols. Clonal chromosomal abnormalities identified by G-banding were described according to An International System for Human Cytogenetic Nomenclature (ISCN, 1995). Clonality was described by detection of two cells with the same structural abnormality or chromosomal gain and loss of the same chromosome in three cells. The metaphase spreads were aged at room temperature for 1 week for SKY. The consensus karyotype was generated for each cell line on the basis of G-banding and SKY analysis.

CGH
High molecular weight DNA was isolated from cell lines and normal placenta by standard methods and subjected to CGH according to a previously published method, with some modifications (Kallioniemi et al., 1992). The metaphase preparations were captured and processed by use of the Quantitative Image Processing System (Applied Imaging, Santa Clara, CA).

SKY
The cocktail of human chromosome paints was obtained from Applied Spectral Imaging (ASI, Carlsbad, CA). Hybridization and detection were carried out according to the manufacturer's protocol, with slight modifications. Chromosomes were counterstained with 4Ј,6-diamidino-2-phenylindole (DAPI). For each case, at least five metaphase cells were analyzed by SKY. Images were acquired with a SD200 Spectra cube (ASI) mounted on a Zeiss Axioplan II microscope (Zeiss, Jena, Germany) by use of a custom-designed optical filter (SKY-1; Chroma Technology, Brattleboro, VT) and analyzed by use of SKY View 1.2 software (ASI).
The breakpoints on the SKY-painted chromosomes were determined by comparison of corresponding inverted-DAPI banding of the same chromosome and by comparison with the G-banded karyotype for each cell line. By this method, we were able to define the breakpoints on add and der chromosomes, but were unable to assign the precise breakpoints of chromosomal segments from partner chromosomes that generated the add or der chromosomes. A breakpoint was considered to be recurrent if it was identified in two or more cases (Rao et al., 1998). A breakpoint cluster was defined as the occurrence of four or more breakpoints in the same chromosomal band.

FISH
FISH was performed to identify the HPV16 integration sites and copy number changes of MYC (8q24) and PIKA3CA (3q26.3) in eight CC cell lines. The plasmid containing HPV16 sequences and the BAC clone (386M7) containing PIKA3CA was labeled by nick translation with Spectrum Red/ Spectrum Green (Vysis, Downers Grove, IL). The digoxigenin-labeled MYC probe was obtained from Ventana (Tucson, AZ). Hybridization and detection were performed according to the manufacturer's protocols.

Comparison of CGH and SKY Analysis
There was a good concordance between the number of genomic imbalances identified by CGH and unbalanced structural aberrations identified by SKY. An example of the comparison is depicted in Figure 2, which displays a loss of 2q (Ca Ski and C-4I) and 11q (Ca Ski) and loss and gain of 3p and 3q (ME-180) and 5p and 5q (Ca Ski).
An isochromosome for the short arm of chromosome 5 [i(5)(p10)] was noted by SKY in C-4I, SW756, and SiHa. The corresponding CGH profiles showed only a gain of chromosome arm 5p and no loss of 5q. This may be attributable to the translocation of 5q chromosomal material to other chromosomes in the genome. The loss of Xp and gain of Xq in cell line MS751 and loss of 16p and gain of 16q in Ca Ski can be correlated with the presence of i(X)(q10) and i(16)(q10), respectively. SKY identified i(12)(p10) in SW756, but the corresponding CGH profiles were normal; this discrepancy may be ascribed to the low abundance of this clone in the tumor genome.
In the majority of cases, the chromosomal amplification results from the unbalanced translocations material from one region to the other chromosomes in the genome except for SIHA, where it displayed the translocation, der(16)t(7;16)(q?;p11), and duplication of the 16q21q24 region. However, the amplification at 10q21 in ME-180 and 11q13 in MS751 was visible only by CGH.
Because we identified recurrent gains at 3q and 8q by CGH, we performed FISH analysis to estimate copy numbers in interphase nuclei and metaphase spreads through use of PIK3CA (3q26.3) and MYC (8q24) as probes ( Table 2). The distribution of copy number for MYC ranged from two to six copies, with more than five copies being noted in Ca Ski, ME-180, MS751, and SiHa. In the ME-180 cell line, we identified der(5)t(5;8)(p13;q23) by SKY, but FISH analysis identified this as a cryptic duplication of MYC on the der(5) chromosome (Fig. 3). This was consistent with the increased copy number identified by CGH at 8q23-q24.1. By use of PIK3CA (3q26.3) as a probe, FISH analysis revealed a copy number increase in C-4I (three to five copies), Ca Ski (two to four copies), and ME-180 (two to three copies). These cell lines exhibited gain of chromosome arm 3q by CGH.
An increased copy number for 3q was noted in four of the eight cell lines, and the common region mapped to 3q26 -q27. The 3q gain was identified in a diverse group of tumors including head and neck, cervix, lung, and esophagus, and it has been shown to associate with a progression from highgrade CIN to invasive CC (Heselmeyer et al., 1996). Although previous studies have identified several candidate genes on 3q26 -q27, controversy still exists regarding the gene(s) that are the target of amplification in CC. The PIK3CA locus, which encodes a catalytic subunit of phosphatidylinositol 3-kinase, has been implicated as a candidate oncogene at 3q26 in carcinomas of the cervix, ovary, and head and neck (Shayesteh et al., 1999;Ma et al., 2000;Singh et al., 2001). We recently showed that the negative effect of TP53 induction on cell survival involves the transcriptional inhibition of PIK3CA that is independent of PTEN activity in squamous cell carcinomas (Singh et al., 2002). Another gene, eIF-5A2, a member of the eukaryotic initiation factor, which plays a role in translation initiation, has been implicated as a candidate oncogene in ovarian cancer (Guan et al., 2001). Gains of 5p were noted in five of eight cell lines studied. This is suggestive of the presence of i(5)(p10) in the tumor genome and is consistent with previously published cytogenetic studies (Atkin, 1997). The telomerase reverse transcriptase (TERT) that maps to 5p15 has been shown to control the telomerase activity by an expression of its catalytic component. The TERT gene has been shown to be a putative oncogene and was amplified in several solid tumors, including cervical cancer (Greider, 1999;Zhang et al., 2002).
CC is shown to associate with HPV infection, particularly the high-risk types HPV16 and HPV18 (Canavan and Doshi, 2000). It has been established that the oncogenic HPV viral proteins E6 and E7 interact with and inactivate the tumor-suppressor proteins p53 and pRb1, thereby initiating a malignant transformation (zur Hausen, 2000). The HPV integration was found in a limited number of CC cell lines or primary tumors at normal or abnormal chromosomes. Recent molecular cytogenetic characterization of the HeLa cell line revealed HPV18 integration on normal chromosome 8 (8q24) and two derivative chromosomes involving the 8q24 band (Macville et al., 1999). In the present study, we noted HPV16 integration on normal chromosome 17 at 17p13 in the ME-180 cell line and two derivative chromosomes involving chromosome bands 18p10 and 18p11 in C-4I. In cell line Ca Ski, HPV integration was observed on one derivative chromosome involving the 8q24 band and another eight derivative chromosomes involving 2p16, 5q21, 5q23, 6q, 10, 11p11, 15q, and 18p11. This strongly suggests that the genomic reorganization is caused by HPV viral integration in cervical cancer. This is consistent with previously published correlation studies between the breakpoints in chromosomal rearrangements and HPV integration sites in cervical cancer (De Braekeleer et al., 1992).
In conclusion, we report here the precise and detailed characterization of eight frequently used