Identification of tumor suppressive activity by irradiation microcell-mediated chromosome transfer and involvement of alpha B-crystallin in nasopharyngeal carcinoma

Authors

  • Hong Lok Lung,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Cathy Carfield Lo,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Carmen Chak Lui Wong,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Arthur Kwok Leung Cheung,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Ka Fu Cheong,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Nathalie Wong,

    1. Department of Anatomical and Cellular Pathology, The Chinese University of Hong Kong, Shatin, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Fung Mei Kwong,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • King Chi Chan,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Evan Wai Lok Law,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Sai Wah Tsao,

    1. Department of Anatomy, University of Hong Kong, Pokfulam, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Daniel Chua,

    1. Department of Clinical Oncology, University of Hong Kong, Pokfulam, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Jonathan Shuntong Sham,

    1. Department of Clinical Oncology, University of Hong Kong, Pokfulam, Hong Kong (SAR), People's Republic of China
    Search for more papers by this author
  • Yue Cheng,

    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    2. Genetics Branch, National Cancer Institute, National Naval Medical Center, Bethesda, MD
    Search for more papers by this author
  • Eric J. Stanbridge,

    1. Department of Microbiology and Molecular Genetics, University of California, Irvine, CA
    Search for more papers by this author
  • Gavin P. Robertson,

    1. Department of Pharmacology, The Pennsylvania State College of Medicine, Hershey, PA
    Search for more papers by this author
  • Maria Li Lung

    Corresponding author
    1. Department of Biology and Center for Cancer Research, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (SAR), People's Republic of China
    • Department of Biology, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
    Search for more papers by this author
    • Fax: +852-2358-1559.


Abstract

In previous studies, we successfully refined nasopharyngeal carcinoma (NPC) critical regions (CRs) mapping to chromosome 11q13 and 11q22-23. The chromosome 11 fragment containing the 1.8 Mb NPC CR at 11q13 (CR1), the CR at 11q22.3 mapped near D11S2000 (CR2), part of the CR at 11q23.1-11q23.2 overlapping with D11S1300 and D11S1391 (CR3), and the CR at celladhesionmolecule 1 (CADM1) locus (CR4), was chosen as the chromosome 11 donor cell line for the present study. Gamma irradiation was applied to cleave this truncated chromosome into smaller fragments and a new panel of donor cells containing further deleted fragments was produced. Subclones XMCH3.2 and XMCH3.4 were chosen for subsequent transfer to HONE1 cells; each contains a single copy of deleted chromosome 11 fragment with or without CR2 and the THY1 locus, previously shown to be involved in NPC. Both resultant chromosome 11 fragments in XMCH3.2 and XMCH3.4 caused tumor suppression. The association of alpha B-crystallin (CRYAB), a gene identified as being differentially expressed by gene profiling of NPC and an immortalized nasopharyngeal epithelial cell line, and which is located near CR3, was found to be associated with tumor suppression in all the tumor-suppressive hybrids. In addition, the expression level of this gene was down-regulated in the 7 NPC cell lines and in 5 out of 14 normal/tumor tissue pairs in the present study. Both promoter hypermethylation and allelic loss may be involved in the inactivation of this gene, suggesting its possible role in NPC development. © 2007 Wiley-Liss, Inc.

Nasopharyngeal carcinoma (NPC) is a malignancy that is common in Southern China and Southern Asia populations among the ethnic Chinese.1–3 Epstein-Barr virus (EBV), genetic predisposition, dietary and environmental factors are all believed to be involved in NPC development. Losses or inactivation of tumor suppressor genes (TSGs) are an important step in tumor development. By using a monochromosome transfer approach, it has been shown previously that critical regions (CRs) responsible for tumor suppression are localized to chromosome 11. By detailed microsatellite genotyping and comparative BAC FISH analyses on nontumorigenic hybrids and their tumor segregants (TSs), these CRs were mapped to 11q13 and 11q22-23.4 A chromosomal interval of 1.8 Mb within 11q13 and 3 chromosomal regions of 0.36, 0.44 and 0.30 Mb within 11q22-23 were observed to be commonly lost in the TSs. Selective loss or inactivation of wild type alleles from the normal donor chromosome 11 may occur under natural selection pressure. The 1.8 Mb CR at 11q13 mapped near D11S913, where the multiple endocrine neoplasia type 1 (MEN1) gene is located.5 The three CRs at 11q22-23 of 0.36, 0.44 and 0.30 Mb mapped near D11S2000, D11S1300, D11S1391 and D11S4484, respectively. The 11q22-23 region contains a critical tumor suppressive region with TSG activity, as observed in many kinds of cancer, such as melanoma, breast, ovarian, lung, cervical, bladder, colorectal and prostate cancers. Within this region, a tumor suppressor celladhesionmolecule 1 (CADM1, formerly called TSLC1) near D11S4484 (11q23.2-23.3) is also important in NPC development.6 In our recent study, tumorigenicity assay results show that the activation of CADM1 suppresses tumor formation in nude mice, inhibits cell growth and induces apoptosis in the absence of serum. These findings suggest that CADM1 is a TSG in NPC, which is significantly associated with lymph node metastases.7

To obtain direct functional evidence of the other three NPC CR tumor suppressing activities, we selected a chromosome 11 fragment from donor cell 22,8 to use as a donor chromosome for γ-irradiation microcell-mediated chromosome transfer (XMMCT), as most of its p arm is already deleted. This chromosome 11 fragment contains the 1.8 Mb NPC CR at 11q13 mapped near D11S913 (CR1), the CR at 11q22.3 mapped near D11S2000 (CR2), part of the CR at 11q23.1-11q23.2 overlapping with D11S1300 and D11S1391 (CR3) and the CR at 11q23.2-23.3 nearby D11S4484 (CR4), respectively. The irradiated chromosome 11 fragments were transferred into the mouse A9 fibroblast cells to establish a new panel of donor cells containing truncated chromosome 11 fragments. The truncated chromosome fragments with unique genomic patterns and CR(s) deleted were then transferred into the NPC HONE1 cell line and the effects on the tumorigenicity were studied. From microarray analysis of the gene expression profile of NPC cell lines versus the immortalized normal nasopharyngeal epithelial cell line NP460, CRYAB located nearby CR3, was one of the candidates that showed down-regulation in all the NPC cell lines tested. The association of CRYAB with the tumorigenicity of the microcell hybrids (MCHs) containing the truncated chromosome fragments, its mechanism of inactivation, and the possible role of this gene in NPC were studied.

Abbreviations:

BAC, bacterial artificial chromosome; CRYAB, alpha B-crystallin; FISH, fluorescence in situ hybridization; MCH, microcell hybrid; NPC, nasopharyngeal carcinoma; TSG, tumor suppressor gene; XMMCT, irradiation microcell-mediated chromosome transfer.

Material and methods

Cell lines and culture conditions

The donor cell line 22, which contains a truncated human chromosome 11 tagged with a neomycin resistance gene8 and MCH cell lines were maintained in DMEM/10%FCS/G418 (Geneticin) at 500 μg/ml. The recipient NPC HONE1 cell line and other NPC cell lines were maintained as previously described.9 The immortalized nasopharyngeal epithelial cell line NP460 was cultured as described.10 The mouse donor hybrid cell line, MCH 556.15, containing an intact human chromosome 11, was maintained as previously described.4

PCR-based microsatellite analysis

Microsatellite typing of recipient, donor and MCH cell lines was performed using semi-automated fluorescent PCR-based analysis on an Applied BioSystems Prism 3100 Genetic Analyzer (Perkin-Elmer Applied Biosystems, Foster City, CA), as previously described.6 Primers for microsatellite analysis were synthesized (Invitrogen, Carlsbad, CA) according to previous reports4 and the Genome Database (http://www.gdb.org). Twenty-seven microsatellite markers, spanning the 121.7 Mb lengths between D11S1363 and D11S4094 were used. Molecular and cytogenetic map positions were obtained from NCBI (http://www.ncbi.gov) and are shown in Figure 1.

Figure 1.

Microsatellite analysis of cell lines, HONE1, 22, XMCH3.2 and XMCH3.4, and the corresponding HONE1/chromosome 11 hybrids. (○, presence of marker; •, loss of marker; U, uninformative) The four CRs identified in our previous study (Lung et al., Ref.7) and the approximate positions for BACs (color of probe: O, orange; G, green; P, purple) utilized in this study and for genes CRYAB, CADM1 and THY1 are shown. The donor cells, XMCH3.2 and XMCH3.4, differ in the retention of CR2 and THY1 and all their corresponding MCHs were tumor-suppressive. T, tumorigenic; S, suppressed tumorigenic potential.

For comparative genotyping analysis between MCHs and TSs at the CRYAB locus and the nearby region, overlapping and nearby microsatellite markers, D11S1300, RH17897, D11S3407, STS-S67070, WI17363, D11S4097 and D11S614, were mapped according to the Genome Database (http://www.gdb.org), NCBI (http://www.nih.ncbi.gov) and Ensembl Genome Browser (www.ensembl.org), as shown in Figure 7a. Each allele was scored by comparing the ratio of the signal intensities between the MCH and its corresponding tumor segregant. When the value was less than 0.8 or larger than 1.2, the allele was scored as allelic loss.6 For loss of heterozygosity (LOH) studies of NPC biopsies, each allele was scored by comparing the ratio of the signal intensities between the tumor tissue and its normal counterpart. The values for peak areas of the two alleles in the paired normal and tumor samples were used for calculation of allele loss based on the formula T1:N1/T2:N2, where T1 and N1 are the area values of the lower allele peaks and T2 and N2 are the area values of the higher allele peaks for the tumor (T) and normal (N) samples. LOH was defined as having a ratio of less than or equal to 0.5, to allow for contaminating normal cells in the tumor sample. All assays were performed at least twice and the mean value was taken.

Fluorescence in situ hybridization

Confirmation of an extra chromosome 11 fragment in the hybrids was determined by fluorescence in situ hybridization (FISH) using WCP 11 SpectrumGreen™ and chromosome 11-specific probes (Vysis, Downers Grove, IL). Hybridization and fluorescence signal capture was performed as described previously.5, 6

Multi-color BAC FISH analysis

Three human BAC clones (RPCI-11 Human BAC Library and California Institute of Technology Human BAC Library) were obtained from Research Genetics (Huntsville, AL) and include RP11-482O1 (11q13.1), RP11-451F15 (11q22.1) and RP11-521L22 (11q23.3), as seen in Figure 1. Molecular and cytogenetic map positions were obtained from the Human Genome database from NCBI. Dual-color FISH analysis was performed according to the previous method described.11 In brief, BACs RP11-482O1 and 451F15 were directly labeled with Spectrum Orange-dUTP (Vysis, Downers Grove, IL) and Spectrum Green-dUTP (Vysis, Downers Grove, IL), respectively, while RP11-521L22 was indirectly labeled with biotin and visualized using avidin-conjugated Texas Red. The three probes were hybridized simultaneously on a single metaphase slide and they were verified for cytogenetic locations on normal human metaphase chromosomes prior to mapping study of the hybrid cells (data not shown). Hybridized signals were evaluated using a Leitz DMRB (Leica, Wetzlar, Germany) fluorescence microscope. At least 20 well-spread metaphases were examined for each cell line and duplicated experiments were performed for each cell line.

Microcell-mediated chromosome transfer and DNA slot blot assay

MMCT and DNA slot blot assay were performed as previously described.12

X-irradiated microcell-mediated chromosome transfer

Gamma irradiation was used to further produce smaller chromosome 11 fragments from donor cell 22. After microcells were obtained after centrifugation, they were exposed to 1,000 rads of γ-irradiation from a cesium 137 source (Gammacell-1000 Elite Irradiator; MDS Nordion, Kanata, Ontario, Canada) and the resultant microcells were then fused with mouse A9 cells as in normal MMCT.8

In vivo tumorigenicity assay

The tumorigenicity of each cell line was assayed by subcutaneous injection as described in Lung et al.7 In brief, 1 × 107 cells were injected into 4- to 8-week-old female athymic Balb/c Nu/Nu mice. A total of 6 sites were tested for each cell line and tumor growth in animals was checked weekly.

NPC tissue specimens

Matched normal nasopharyngeal and NPC biopsies from 15 NPC patients were collected at Queen Mary Hospital in Hong Kong between January 2006 and December 2006 and directly stored in RNAlater solution (Ambion, Austin, TX). Ages of the Hong Kong patients ranged from 28 to 76 years (mean: 51.3-years-old), with a male to female ratio of 2:1. The tumor specimens encompassed 6 primary NPC cases without metastasis and 9 lymph node metastatic NPC. They were classified as undifferentiated squamous cell carcinoma.

Reverse transcription-polymerase chain reaction

Total RNA was extracted from cells using the RNeasy Midi Kit (Qiagen, CA). One microgram of total RNA was reverse transcribed with 200 U of SuperScriptII reverse transcriptase (Gibco, NY). For amplification of alpha B-crystallin (CRYAB), the primer sequences were 5′-TTCTTCGGAGAGCACCTGTT-3′ and 5′-TTTTCCATGCACCTCAATCA-3′. For PCR amplification of specific cDNA derived from glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene, a pair of primers was used as previously described.7 PCR was performed in a GeneAmp PCR System 9700 programmable thermal controller.

Methylation-specific PCR analysis

The design of primer pairs for methylated and unmethylated DNA was based on MethPrimer (www.urogene.org/methprimer) primer design program.13 The sequences were 5′-GAGGAGGAGGGGTATATTACGTC-3′ and 5′-AACACAAACACGTCTAAACCGA-3′ for methylation primers and 5′-GAGGAGGAGGGGTATATTATGTTG-3′ and 5′-AAAAAACACAAACACATCTAAACCA-3′ for unmethylated primers. In brief, 1 μg genomic DNA was bisulfite-modified using the CpGenome™ DNA Modification Kit (Chemicon International, Temecula, CA) and modified DNA was subjected to PCR. CpGenome™ Universal Methylated DNA (Chemicon International, Temecula, CA) was used as a positive control for methylated DNA. For each PCR, negative controls without DNA templates were included.

5-Aza-2′-deoxycytidine treatment

The NPC cell lines were treated with 5 μM 5-aza-2′-deoxycytidine (Sigma, St Louis, MO). The medium was changed every 24hr. After incubation for 4 days, HONE1 cells were harvested for total RNA extraction.

Statistical analysis

Statistical comparison of tumor volume in HONE1 to that in the MCHs 6 weeks after injection was performed with the independent t test. The χ2 and Fisher's Exact test were used for analysis of significant differences in CRYAB expression levels detected by RT-PCR between primary and metastatic NPC. Differences were considered statistically significant for p-values <0.05.

Results

Analysis of chromosome 11 fragment in cell line 22 and its microcell hybrids

The cell line 22 containing a truncated chromosome 11 (Ref.8) was selected as chromosome 11 donor cells for the present study. Microsatellite analysis shows that most of the p arm of the truncated chromosome 11 in cell line 22 was deleted, as indicated by the absence of the region between markers D11S1363 and D11S904. In addition, microdeletions were observed at 11q23.2 and 11q24.1, where markers D11S1391 and D11S4094 map, respectively (Fig. 1). Hence, this donor chromosome 11 contains the 1.8 Mb NPC critical region (CR) at 11q13 mapped near D11S9135 (CR1), the CR at 11q22.3 mapped near D11S2000 (CR2), part of the CR at 11q23.1-11q23.2 overlapping with D11S1300 and D11S1391 (CR3), and the CR at CADM1 as indicated by D11S4484 (CR4),6 respectively. In addition, the THY1 locus is present in this chromosome 11 fragment, as indicated by the markers D11S4129 and D11S4171. FISH analysis confirms donor cell 22 contains a single discrete copy of the truncated chromosome 11 in the A9 background (Fig. 2a).

Figure 2.

FISH analysis of parental HONE1 cells, donor cells and HONE1/chromosome 11 hybrids. (a–c) Whole chromosome painting of donor cell lines 22, XMCH3.2 and XMCH3.4. (d, e) Multi-color FISH analysis for XMCH3.2 and XMCH3.4 with BACs RP11-482O1 (orange), RP11-451F15 (green) and RP11-521L22 (purple). The exogenous chromosome 11 is indicated by an arrow (▾) and a magnified image of this chromosome is shown. (f–i) Whole chromosome painting of HONE1 cells and HONE1/chromosome 11 hybrids shows an extra copy of chromosome 11 is present in the representative MCHs, as indicated by an arrow (▾), when compared with HONE1. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Production of discrete radiation-reduced monochromosome hybrids

After γ-irradiation was applied to donor cell 22 and MMCT was performed to transfer the truncated fragments to A9 cells, 83 clones were obtained. Out of the 83 XMCH clones, DNAs from 50 clones were extracted and screened by DNA microsatellite analysis. Only clones with unique genomic patterns were chosen for further screening. By FISH analysis, clone XMCH3 was shown to be multi-clonal (data not shown); subcloning was performed to separate the individual subclones. By genotyping and FISH analyses, the two subclones of XMCH3, XMCH3.2 and XMCH3.4, contain a single copy chromosome 11 fragment (Figs. 2b and 2c) with further deletions at 11q21-23 and at the telomere. The 15.8 Mb region between D11S2002 and D11S901 at 11q13.1 to 11q14 was deleted in both subclones (Fig. 1). In addition to this, the chromosome 11 fragment in XMCH3.4 was further deleted at the 8.6 Mb regions between D11S2000 and 11S1294 and at the THY1 locus, a 4.8 Mb region mapped by D11S4129 and D11S4171 (Fig. 1). Hence, the subclones XMCH3.2 and XMCH3.4 differ in the retention of CR2 and THY1. This is confirmed by multi-color BAC FISH. The green signal detected in XMCH3.2 with BAC probe 451F15, which maps near to D11S2000, was not detected in XMCH3.4 (Fig. 2e). The BAC FISH results also showed that there were two copies of orange signals detected with the probe RP11-48201 mapped to 11q13.1 in both subclones (Figs. 2d and 2e).

By comparative study of these two chromosome 11 fragments, functional evidence for the importance of CR2 and THY1 were obtained. Thus, the two subclones were used as donor cells in MMCT for the transfer of the fragmented chromosome 11 into HONE1 cells.

Transfer of chromosome 11 fragments into HONE1 cells

To determine if the chromosome 11 fragments in donor cell 22, XMCH3.2 and XMCH3.4 are tumor suppressive, MMCT was performed to transfer the exogenous chromosome 11 to the NPC HONE1 cells. DNA from all hybrids was subsequently subjected to slot blot assays. Mouse A9 DNA was used as a probe to exclude any possible contamination of residual mouse DNA during the chromosome transfers (data not shown). Only mouse DNA-free clones were utilized for further screening. By microsatellite genotyping and FISH analyses, 3 clones from each donor cell were verified to contain an extra human chromosome 11 with retention of most of the transferred chromosome (Figs. 1 and 2).

Tumorigenicity assays and tumor segregant analyses

Tumorigenicity was suppressed by an exogenous copy of chromosome 11 for every hybrid clone derived from the 3 donor cells (Fig. 3). However, no significant differences in growth kinetics were found in the tumors from cell lines containing the three chromosome 11 fragments (Table I). The HONE1 cells are highly tumorigenic; palpable tumors form within 7–21 days of injection of these cells in 100% of nude mice. All chromosome 11 MCHs exhibited a delayed lag period in forming tumors, in which tumors were not observed after 12 weeks postinjection and their growth kinetics are significantly different to those of HONE1 (p-values <0.005).

Figure 3.

Tumor growth kinetics of HONE1 and MCH cell lines derived from donors 22, XMCH3.2 and XMCH3.4 in nude mice. The curves represent an average tumor volume of all 6 sites inoculated for each cell population.

Table I. Tumorigenicity Assays of Parental Hone1 and Chromosome 11 XMCH Clones in Nude Mice
Cell lineIdentificationTumor formation, No of tumors/No of sitesTime to appearance of tumors (days)p-Value1
  • 1

    p-Value obtained by comparison with HONE1.

  • 2

    p-Value obtained by comparison with the average tumor growth kinetics of MCH22C1, MCH22C2 and MCH22C4.

HONE1Parental NPC cells6/67–21
MCH22C1HONE1× 223/649–56<0.005
MCH22C2HONE1× 224/656–70<0.005
MCH22C4HONE1× 222/677–84<0.005
  Average60–70
MCH3.2C1HONE1× XMCH3.25/642–63<0.005
MCH3.2C2HONE1× XMCH3.20/6<0.005
MCH3.2C4HONE1× XMCH3.22/677–84<0.005
  Average59.5–73.50.252
MCH3.4C1HONE1× XMCH3.43/656–77<0.005
MCH3.4C2HONE1× XMCH3.44/663–77<0.005
MCH3.4C3HONE1× XMCH3.40/6<0.005
  Average59.5–770.152

RT-PCR analysis of CRYAB and THY1 gene expression in HONE1 cells, chromosome 11 MCHs and their TSs

From microarray analysis of the gene expression profile of NPC cell lines versus the immortalized nasopharyngeal epithelial cell line NP460, CRYAB was one of the candidates in chromosome 11 showing down-regulation in all the NPC cell lines tested (data not shown). In addition to its abnormal expression in NPC cell lines, CRYAB is mapped near D11S4097 at 11q23 and is present in all the 3 chromosome fragments in the current study (Fig. 1); hence, we attempted to see if the tumor suppressive effect of those fragments was associated with this gene. The gene expression of CRYAB in the MCHs and HONE1 cells was analyzed by RT-PCR; results show that CRYAB expression was up-regulated in all MCHs after the transfer (Fig. 4a). On the other hand, THY1 gene expression, identified in our previous study, was up-regulated in MCHs derived from donor cell line 22 and XMCH3.2, but up-regulation was not detected in any MCH3.4 clones (Fig. 4a), as THY1 is physically deleted in XMCH3.4 (Fig. 1). The gene expression of NP460 served as a positive control.

Figure 4.

(a) RT-PCR analysis of gene expression levels of CRYAB and THY1 in HONE1 and MCHs derived from donors 22, XMCH3.2 and XMCH3.4. The immortalized normal nasopharyngeal epithelial cell line NP460 was used as a positive control. CRYAB gene expression was up-regulated after the transfer of chromosome fragments from cell lines 22, XMCH3.2 and XMCH3.4, respectively, and THY1 was not expressed in MCHs derived from XMCH3.4. (b) Comparison of CRYAB and THY1 gene expression in MCHs and their TSs. The frequency of decrease of CRYAB gene expression is 13 out of 15 TSs (87%), whereas for THY1, the frequency is 4 out of 12 (33%). The DNA ladder (L) indicates the sizes of the PCR bands, which are shown on the right (bp). GAPDH served as an internal control.

To check the expression of newly transferred genes following tumor formation, cells from tumors were reconstituted in culture medium and TSs were obtained. The loss of gene expression in TSs is correlated with the growth of tumors in animals, consistent with the hypothesis that small CRs, which could not be detected by cytogenetics, were selectively lost in tumor reconstitutes.4, 13 Gene expression of CRYAB and THY1 in MCH22C1, MCH22C2, MCH22C4, MCH3.2C1 and MCH3.4C1 and their corresponding TSs was analyzed. Figure 4 shows that CRYAB gene expression was decreased in all TSs except MCH3.4-TS1 and MCH3.4-TS2, when compared with the respective MCHs. Complete loss was observed in MCH3.4-TS3 (Fig. 4b). The frequency of decrease of CRYAB gene expression is 13 out of 15 TSs (87%), whereas for THY1, the frequency is 4 out of 12 (33%). The decrease of THY1 expression was observed in MCH22C1-TS2, MCH3.2C1-TS1, MCH3.2C1-TS3 and MCH3.2C1-TS5 (Fig. 4b).

Gene expression of CRYAB in NPC cell lines and biopsy specimens

In comparison with the gene expression of NP460, CRYAB is down-regulated in all seven NPC cell lines in the current study; the gene expression in HONE1, HNE1 and C666 was barely detectable after amplification (Fig. 5a). These results are in agreement with our unpublished microarray data. To investigate the clinical significance of CRYAB in NPC, 15 NPC normal/tumor pairs were analyzed for CRYAB gene expression using RT-PCR. A total of 7/15 pairs (47%) (samples 24, 27, 29, 31, 33, 42 and 48) showed no or decreased CRYAB expression, when compared with the normal counterpart (Fig. 5b). The frequency of down-regulated expression of CRYAB in lymph node metastatic NPC was 67% (6/9), which was higher than that in primary NPC without metastasis (17%) (1/6), but the difference is not statistically significant (p = 0.08) because of the small sample size.

Figure 5.

(a) Gene expression analysis of transcript levels of CRYAB in NPC cell lines. NP460 serves as a positive normal control. (b) CRYAB gene expression analysis of NPC specimens. Fifteen pairs of normal and tumor tissues are included. The DNA ladder (L) indicates the sizes of the PCR bands, which are shown on the right (bp). GAPDH served as an internal control.

CRYAB promoter methylation analysis in NPC cell lines, biopsy specimens and chromosome 11 MCHs and their TSs

To investigate the inactivation mechanism of CRYAB in NPC, we studied the methylation status of CRYAB promoter in the NPC cell lines and biopsies, and a tumor revertant by MSP analysis. Figure 6a shows amplification of methylated sequences in the HONE1, HK1, HNE1, CNE2, SUNE1, CNE2 and C666 cells, but not in CNE1 cells, while an unmethylated signal was observed in the normal immortalized nasopharyngeal epithelial cell NP460 and the intact chromosome 11 donor cell line MCH556.15. Promoter methylation of CRYAB was observed in 6 out of 7 NPC cell lines in the present study. Subsequent reexpression of the gene in HONE1, HK1 and SUNE1 cells, but not in CNE1 cells, was observed after treatment with the demethylation agent 5-aza-2′-deoxycytidine (Fig. 6b). For the NPC biopsies, T24, T27, T31, T42 and T48, showing the most down-regulated CRYAB gene expression (Fig. 5b), were chosen for MSP analysis. Figure 6a shows that one methylated allele of the CRYAB promoter was detected in all the analyzed tumor tissues, whereas both alleles in all normal tissues, except N42, were unmethylated. A faint band of the methylated allele was observed in N42 and is most likely explained by minor contamination of tumor cells in the normal tissue specimen.

Figure 6.

(a) MSP analysis of CRYAB promoter methylation in seven NPC cell lines (HONE1, HK1, HNE1, CNE1, CNE2, SUNE1 and C666), the immortalized normal nasopharyngeal epithelial cell line NP460, MCH556.15, the universal methylated DNA, the NPC tumor and normal paired tissues (samples 24, 27, 31, 42 and 48), and the chromosome 11 hybrid MCH22C2 and its corresponding TS MCH22C2-TS4. Sizes of PCR products are shown on the right. (b) Reexpression of CRYAB in 4 NPC cell lines (HONE1, HK1, CNE1 and SUNE1) and 5 TSs after treatment with 5 μM 5-aza-2′-deoxycytidine was monitored by RT-PCR analysis.

Five tumor revertants (MCH22C1-TS3, MCH22C2-TS4, MCH22C4-TS1, MCH3.2C1-TS1 and MCH3.4C1TS3), showing the highest down-regulation of CRYAB gene expression (Fig. 4b), were chosen for the CRYAB promoter methylation analysis. The level of gene expression of CRYAB in MCH22C1-TS3, MCH22C2-TS4 and MCH3.4C1TS3 was significantly restored after treatment with the demethylation agent 5-aza-2′-deoxycytidine (Fig. 6b). When comparing the methylation status among MCH22C2-TS4, its corresponding hybrid cell line MCH22C2, and the recipient HONE1 cells, the signal intensity of the methylation allele was barely detectable in MCH22C2 and increased to a similar level to HONE1 in its revertant MCH22C2-TS4 (Fig. 6a). Hence, promoter hypermethylation of CRYAB is one of the inactivation mechanisms in both NPC cell lines and clinical biopsies.

CRYAB allelic loss analysis by microsatellite analysis in chromosome 11 MCHs and their TSs and NPC biopsy specimens

Besides promoter hypermethylation, allelic loss is another common mechanism for TSG inactivation. Thus, microsatellite analysis was performed with overlapping and nearby informative markers, D11S1300, RH17897, D11S3407, WI17363, D11S4097 and D11S614, as shown in Figure 7a. The same five tumor revertants (MCH22C1-TS3, MCH22C2-TS4, MCH22C4-TS1, MCH3.2C1-TS1 and MCH3.4C1TS3) showing down-regulation of CRYAB expression in Figure 4b were chosen for this study. All five TSs show allelic loss at the CRYAB locus, as indicated by the markers RH17897 and D11S3407, and all the 5 TSs except MCH22C1-TS3 show allelic loss for all the informative CRYAB mapping markers (RH17897, D11S3407 and WI17363) (Fig. 7a). Representative results with the microsatellite marker D11S4097, showing 80% allelic loss in these TSs, are shown for a paired hybrid and tumor segregant in Figure 7b. Two distinct alleles are present in the recipient HONE1 with the D11S4097 microsatellite marker, while only one peak is observed in the donor cell line XMCH3.2. An increase in the peak height corresponding to the donor allele is evident in the MCH3.2C1 hybrid; its tumor segregant, MCH3.2C1-TS1, shows a decrease in the same peak compared to HONE1, as indicated. This demonstrates loss of both exogenous and endogenous loci. The allelic loss at the CRYAB locus was also observed in clinical NPC specimens for the matched tumor tissues of all chosen samples (T24, T27, T31, T42 and T48) with loss of CRYAB gene expression (Fig. 7a). The highest frequency of loss was observed with the markers D11S3407 and STS-S67070, within the CRYAB locus, and this is similar to the genotyping results of the TSs analysis (Fig. 7a). Representation of the allelic loss at the CRYAB mapping markers D11S3407, STS-S67070 and the nearby marker D11S4097, in the matched NPC tumor tissues is shown in Figure 7c. The genotyping results are not available for markers such as D11S1300 and D11S614 of NPC specimen pairs and some markers of sample 24, because of inadequate specimen DNA (Fig. 7a) for this extended analysis.

Figure 7.

(a) Genotyping results of D11S1300, RH17897, D11S3407, STS-S67070, WI17363, D11S4097 and D11S614 mapping to the CRYAB locus in TSs derived from HONE1 chromosome 11 hybrids and LOH analysis in NPC tumor and normal paired tissues (T24, T27, T31, T42 and T48). (○, presence of marker; •, loss of marker; U, uninformative; –, the data is not available) Microsatellite analysis of (b) MCH3.2C1 and MCH3.2C1-TS1 and (c) NPC specimens N27 and T27 at locus D11S4097 and N31 and T31 at loci D11S3407 and STS-S76070. The genotyping results show genetic losses in the TS and tumor tissues, as compared to the corresponding hybrid or normal tissue. Arrows (▾) indicate allelic loss.

Discussion

The current study has generated two useful truncated donor fragments of human chromosome 11 from a deleted chromosome 11 by γ-irradiation. By a refined multi-color BAC FISH analysis, the region of deletion was confirmed. In addition, duplication of the area around 11q13.1 was observed in both fragments and indeed this region is frequently amplified in various cancers.14–16 The truncated donor chromosome 11 fragments were subsequently transferred to HONE1 cells. As the even more deleted fragment missing CR2 in XMCH3.4 is still tumor-suppressing, the nude mouse assay suggests that our previously defined CRs for NPC, including CR1, CR3 and/or CR4, are responsible for the tumor suppressive effect. Moreover, THY1, a candidate TSG in NPC identified in our former study based on HONE1/intact chromosome 11 hybrids gene expression profiling analysis,13 was absent in the chromosome 11 fragment in XMCH3.4. THY1 was not expressed in the respective MCHs. The tumorigenicity assay shows that this truncated fragment was also tumor-suppressing. This fragment is believed to contain TSG(s) other than THY1. Our previous genotyping results showed a high frequency of allelic loss in the TSs derived from HONE1/intact chromosome 11 hybrids at the CADM1 locus (CR4) in the region of 11q23.2-3.6 In our recent study, we have provided strong evidence that CADM1 is a TSG in NPC.7 In the present RT-PCR analysis of the MCHs derived from the 3 donor cells, CADM1 was up-regulated in a moderate manner in 6 out of the 9 MCHs (MCH22C1, MCH22C2, MCH3.2C4, MCH3.4C1, MCH3.4C2 and MCH3.4C3; data not shown). This may partly explain why some MCHs are tumor-suppressing.

The gene CRYAB is located at 11q23.1 and is close to CR3. It was up-regulated dramatically in all suppressive MCHs, after the chromosome transfer and functional complementation for its abnormal low expression level in HONE1. The frequency of down-regulation of expression in the tumor revertants was 87%. This is significantly higher than that of THY1 (33%). The observation of gene expression down-regulation in some TSs provides evidence of significant heterogeneity and complexity in the genes involved within this chromosomal region and is in agreement with our early report.6 Loss of CRYAB expression was also observed in the NPC cell lines and tumor tissues of NPC patients in the current study. The loss of CRYAB gene expression seems to be correlated with the lymph node metastasis in NPC, but the sample size is small and needs further validation.

Promoter hypermethylation contributes to the loss of CRYAB expression in NPC cell lines in the present study. At least one allele was methylated in most NPC cell lines and demethylation restored the CRYAB gene expression in the NPC cell lines with down-regulated CRYAB expression. Five NPC biopsies with down-regulated CRYAB gene expression were also hypermethylated in the CRYAB promoter region. The gene expression of CRYAB was restored in three out of five TSs after the demethylation treatment and the induction is highest in MCH22C2-TS4. The higher ratio of unmethylated allele in its corresponding hybrid clone MCH22C2 can be explained by the existence of an extra normal chromosome 11 in this cell line. Its methylation status was altered with reappearance of the methylated allele in the tumor revertant. Hence, promoter hypermethylation is one mechanism of inactivation of CRYAB expression in NPC cell lines, TSs and tumor tissues. Similarly, promoter hypermethylation also contributes to the silencing of CRYAB in anaplastic thyroid carcinomas, as previously reported.17 Allelic loss at the CRYAB locus was also observed in all TSs tested and 5 informative NPC tissues. Three TSs (MCH22C1-TS3, MCH22C4-TS1 and MCH3.4C1-TS3) and 5 tumor tissues (T24, T27, T31, T42 and T48) show CRYAB promoter hypermethylation with allelic loss. It is likely that both allelic loss and promoter hypermethylation contribute to the inactivation of CRYAB gene expression in NPC. The allelic loss at markers D11S3407 and STS-S67070 in the tumor tissue of specimen sample 31 did not show a complete deletion of one allele and the unmethylated allele was still observed in the methylation analysis. This may be due to either contaminating normal cells or the existence of heterogeneity within the tumors, as reported by Gonzalgo et al.18

Molecular and cytogenetic analyses reveal extensive 11q22-24 losses in NPC1, 19 and implicate putative TSGs as having a critical role in NPC development. The possibility that a cluster of TSGs located at 11q22-23 may be involved in NPC tumorigenesis, as indicated by identification of TSGs in the present and our previous studies, may be akin to the situation observed where multiple candidate TSGs mapping to the 3p21.3 region are involved in several human cancers.20 During cancer progression, loss of multiple TSG functions has been reported in numerous instances and correction of any one of the TSG defects is sufficient for the inhibition of tumorigenicity.21 No significant differences in tumor growth kinetics of MCHs with and without THY1 expression might result from the exogenous expression of CRYAB in those MCHs. The expression of THY1 and CRYAB is simultaneously down-regulated in tumor revertants MCH3.2C1-TS1, -TS3 and -TS5 and this may represent multiple loss of TSG functions in NPC. On the other hand, down-regulated CRYAB gene expression in MCH22C1-TS3, MCH22C2-TS2, MCH22C2-TS4 and MCH22C4-TS1 was observed with unchanged THY1 expression; these results indicate the importance of CRYAB in tumor formation in the tumor revertants. The loss of expression of CRYAB may represent an early response in tumor formation in nude mice and, thus, the expression of this gene is strongly related to the tumorigenesis in nude mice. Taken together, CRYAB fulfills the criteria of being another candidate TSG in NPC. Currently, functional studies are underway to elucidate its role in NPC.

CRYAB is a major lens structural protein of the vertebrate eye for light refraction. This protein also functions as a molecular chaperone to protect proteins from aggregation.22CRYAB is a member of the small heat shock protein family and is constitutively expressed in the lens of the eyes, myocardial cells and kidney epithelium. It also has an extensive tissue distribution outside the lens, but its functions in nonlens tissues are unclear. The knockout of this protein in mice results in hyperproliferative lens epithelial cells. The loss of CRYAB is associated with an increase in genomic instability. These observations suggest that CRYAB may play an important nuclear role in maintaining genomic integrity.23 In a study by Stronach et al.,24 using the combined approach of MMCT and expression difference analysis, CRYAB was also identified as one of the candidate genes for chromosome 11-mediated ovarian tumor suppression. In addition, they showed that low expression of CRYAB was significantly associated with adverse patient survival. The down-regulation expression of CRYAB has also been reported in testicular and breast tumors, when compared with their normal counterparts.25, 26 In contrast, over-expression of CRYAB is frequently observed in various cancers such as nonneuronal brain tumors, renal cell carcinoma and breast cancer.27–29 The tissue-specific regulation and expression of this gene may vary in different tissues. Functional evidence is required to determine whether it is an oncogene or a TSG in the specific cancer type.

In our previous study of analyzing the loss of chromosome regions in NPC TSs derived from HONE1/intact chromosome 11 MCHs, we successfully refined a TS region to 1.8 Mb at chromosome band 11q13 (CR1).5 In the current report, we also show that CR1 could exhibit tumor-suppressing activity; however, at the moment no good candidate TSG at 11q13 has been identified in NPC. The association described here between low CRYAB transcript expression suggests that loss of CRYAB expression may indeed be involved in promoting NPC tumorigenesis or tumor progression and functional assays are required to determine the exact role of this gene in NPC.

Acknowledgements

The authors thank the Research Grants Council of the Hong Kong Special Administrative Region, People's Republic of China, Grant number CA03/04.SC01 to MLL for the financial support of this work.

Ancillary