| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Biology and Genetics |
Department of Molecular Physiology and Biological Physics, University of Virginia, Charlottesville, Virginia 22908 [M. A. H., J. W. G., J. J. G., D. T.]; Ludwig Institute for Cancer Research, Department of Medicine, Cancer Center, School of Medicine, University of California at San Diego, La Jolla, California [K. C. A., C. V.]; and Department of Pathology, Children’s Memorial Hospital, Northwestern University, Chicago, Illinois [E. J. P.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We have recently addressed this gap by introducing a model system based on the T24 (12) cell line, a widely used human bladder cancer cell line, and a spontaneously occurring lineage-related variant cell line, which we designate T24T (13) . T24 and T24T exhibit many striking phenotypic differences in growth, invasive and metastatic ability (14) . In view of these findings, we believe that these paired cell lines constitute a good model of bladder cancer progression, lending itself to detailed genetic characterization and molecular hypothesis testing. To further develop and characterize these two cell lines, we have undertaken a detailed comparison of their phenotypic and genetic characteristics.
Using a genome-wide approach to begin the hunt for specific defects that account for the phenotypic differences between T24 and T24T, we have previously compared the karyotypes of T24 and T24T and performed a G-banding analysis of marker chromosomes (13) . From this evidence, six chromosomal abnormalities were found to be common to T24 and T24T. In addition, T24 had five alterations unique to it, and between 2 and 10 unidentifiable marker chromosomes. T24T possessed four unique structural changes with six to nine unidentified markers. Therefore, conventional cytogenetic techniques did not provide the complete and accurate inventory of chromosomal alterations needed to mechanistically evaluate the relevance of each in the conversion to the T24T phenotype.
In this study, we have taken full advantage of the strengths of both SKY and CGH to fully elucidate the chromosomal complement of the T24 and T24T cell lines. The power of our synergistic approach is highlighted by the ability to assign an identity to all of the previously anonymous marker chromosomes and reassign the origin of some markers that were mistakenly assigned using conventional methods (13) . In addition, we have used oligonucleotide microarrays to measure gene expression in both cell lines to develop a novel PEP method using cytogenetic mapping information available from public databases to assess the functional consequences of chromosome imbalances and rearrangements obtained from the SKY and differential CGH. This approach has allowed important insights of how genomic structural changes affect gene expression on a regional scale and helped refine and focus the search for candidate genes associated with the invasive and metastatic phenotype.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Chromosomes were counterstained with DAPI/antifade solution supplied with the SKY kit. Spectral imaging was achieved using a SpectraCube system (Applied Spectral Imaging) mounted on a Nikon E600 microscope, viewed through a x60 oil-immersion plan-apo objective illuminated by a xenon lamp (Opti Quip, Highland Mills, NY). Chromosome classification was performed with Sky View software (Applied Spectral Imaging). DAPI (Vysis, Downers Grove, IL)-stained images were captured separately and subjected to electronic image inversion and contrast enhanced for alignment with spectral representations using Sky View software. Ten SKY images with their companion-inverted DAPI images were analyzed and compared with the previously published G-banded karyotypes (13) .
FISH.
Metaphase spreads were prepared by standard cytogenetic methods. Whole chromosome paint probes for chromosomes 4, 8, 12, and 18 were obtained from Vysis, and the Y satellite DYZ3 probe was obtained from Oncor (Gaithersburg, MD). Probes were applied to the slides and incubated for 10 min at 80°C on a slide warmer to achieve the codenaturation of both probe and target DNA. Hybridization was allowed to proceed for a minimum of 16 h in a dark, humid chamber at 37°C. The slides were washed according to the probe manufacturer’s directions using the formamide-wash procedure, and the metaphases were counterstained with DAPI. Slides were examined and photographed using a x100 objective on a Zeiss Axiophot microscope equipped with a 100-W mercury vapor lamp and appropriate UV filter combinations. Images were recorded using KODAKJ P1600 slide film and developed with push 2 (1600) processing.
CGH.
DNA was extracted using routine methods with phenol/chloroform after SDS/proteinase K digestion (15)
. DNA (1 µg) extracted from T24T and T24 were direct labeled with fluorescein-dUTP using DNase and polymerase I with dATP, dGTP, dCTP (20 mM), Tris (50 mM), MgCl (5 mM), mercaptoethanol (10 mM), and BSA (10 µg/ml) at 15°C. The amount of each enzyme and the reaction time were adjusted to achieve fragment lengths of 600-1500 bp on a 1% agarose gel. Normal male reference DNA derived from lymphocytes, as well as DNA extracted from T24, were direct labeled with Texas red-5-dUTP using the same procedure. Three hundred to 500 ng reference and either the T24T or T24 DNA were coprecipitated with 20 µg Cot-1 DNA (for the comparisons to PBL). Similarly, 300–500 ng each of fluorescein-labeled T24T and Texas red-labeled T24 were coprecititated with Cot-1 DNA (for the T24T versus T24 comparison). The pellets were resuspended in hybridization medium (50% formamide, 2x SSC, and 10% dextran sulfate), and the probes were denatured at 75°C for 10 min and subsequently partially reannealed at 37°C for 30 min. The probes were hybridized to slides containing methotrexate synchronized reference male metaphases that had been previously aged 5 days and denatured at 72°C for 2 min in 70% formamide, 2x SSC (pH 7). The slides were hybridized for 3 days at 37°C in a moist chamber. After washing for 5 min each in 2x SSC at 72°C, 37°C, and water at room temperature, the slides were counterstained with 0.1 µg/ml DAPI in antifade.
Gray level images were acquired for each fluorescent dye with a charge-coupled device camera on a Zeiss Axioscope epifluoresence microscope using the Applied Imaging Corporation’s dedicated Cytovision software and hardware. Chromosomes were identified using reverse DAPI banding. The background fluorescence was subtracted and the green:red ratio of each entire metaphase was normalized to 1.0. Data from at least 12 representatives of each chromosome were combined to generate an average ratio profile and SE for each chromosome.
The interpretation of gains and losses by CGH are difficult in samples with abundant abnormalities, particularly when the samples are not diploid. The thresholds used for the interpretation of gains and losses by CGH must take into account the DNA content of the sample. Because the samples analyzed were hypertriploid when comparing the sample with PBLs, we chose the upper and lower thresholds of 
1.20 and 
0.80, respectively. These reflect rather stringent thresholds when interpreting gains and losses for hypertriploid samples. When comparing T24 to T24T, thresholds of >1.25 and <0.75 were used (16)
. In addition, sharp deviations of the profile were used in the interpretation, which in some cases broadens the regions listed beyond the points where the profile strictly exceeds the thresholds.
Microarray Analysis.
Affymetrix oligonucleotide microarray gene expression analysis was performed by the Molecular Biology Core Facility of the University of Virginia. Analysis was performed essentially as described in the manufacturer’s instructions [Affymetrix (2001) Microarray Analysis Suite User’s Manual]. Briefly, cRNA was prepared from 8 µg of total RNA, hybridized to HuGeneFL Affymetrix oligonucleotide arrays, which contains 
6800 human genes or expressed sequence tags. Total RNA for microarray analysis was isolated from subconfluent cultures of T24T and T24 cells using RNAEasy minicolumns (Qiagen) according to the manufacturer’s directions. After washing in a fluidic station, the arrays were scanned with a 2.5-µm resolution HP Microarray Scanner (Hewlett Packard). Scanned images were first examined for visible defects and then checked for the fitness of the gritting. When passed, the image file was analyzed to generate the Cel data file. From this point on, a coordination of two paths of analysis was carried on using the Affymetrix Microarray Analysis Suite 5.0 (MAS 5.0; Affymetrix, Santa Clara, CA) and the Dchip software (17
, 18)
. The detection of a particular gene, present, absent, or marginal, was made using the MAS 5.0, those detection calls as well as the Cel data file were later imported into and used by the Dchip program. Scatter plots were also generated using this software to inspect the degree of changes of the samples under comparison. Quantitation of the genes was obtained using Dchip, which applied the model-based approach to derive the probe sensitivity index and expression index. The two indices combined to quantify a particular gene. When certain probes or transcripts deviated from the model to a set extent, they were excluded from the quantitation process. Normalization of the arrays was performed using the invariant set approach.
Expression Profiling by Genomic Position.
Analysis of gene expression by chromosome position was performed in the following manner. Cytogenetic positions for each probe set were first determined by obtaining the LocusLink ID from the annotation information provided by the Dchip software. The LocusLink file "LL Out"5
from the National Center for Biotechnology Information provided the cytogenetic range to which each probe set or locus has been mapped. A relational database created in Microsoft Access was used to join the mapping information to the probe set. In the following analysis, probe sets that have not been localized to a subdivision of a chromosome in the latest iteration of LocusLink were not included. Therefore, 496 probe sets were eliminated from additional analysis of the 6084 probe sets arising from the Dchip data.
The physical size for chromosomes and cytogenetic bands and subbands were obtained from the file ISCN800-abc.6 The values ISCN-top and ISCN-bottom in this file represent numerical ranges for the chromosome (and the cytogenetic bands and subbands that comprise it) that are proportional to their respective lengths. The ISCN-bottom number for each chromosome (its total length) was divided by 100 to define a series of positional indices for every chromosome. Thus, the number of indices representing a chromosome is proportional to its size, and each index is the same constant length, regardless of its position in the genome. For example, chromosome 1 was divided into 150 positional indices, and chromosome 22 was divided into 34 indices. The expression data for each probe set was then joined to the range of positional indices corresponding to the cytogenetic localization for that locus in the database.
The gene expression arising from any loci that might lie within each positional index was then computed. The expression level for every probe set in each cell line was divided by the number of indices that comprise the cytogenetic range to which it has been mapped. This value represents a weighted expression level based upon the precision of mapping data available for each loci. Then for each individual index, the weighted expression levels for all loci, the cytogenetic position that overlapped the index, were summed. These values were computed for the expression data from both the T24 and T24T cell lines. For each positional index, the log value of T24T expression/T24 expression was computed. To graphically represent this data, the information from the database was used to construct a pivot table in a Microsoft Excel spreadsheet. Graphs were created showing the log ratio of T24T expression/T24 expression versus positional index. For display purposes, cytogenetic bands are labeled on the graphs as opposed to the underlying positional indices. It should be noted, however, that there is one data point on the graphs for each index and that each cytogenetic band is composed of several indices. A table showing cytogenetic bands with the largest differences in gene expression between T24T and T24 was also created. For this table, the log ratio of the weighted expression level of each probe set in the positional indices that underlie that cytogenetic band were summed for each cell line, and the log ratio of T24T versus T24 was computed. Therefore a positive log ratio represents overexpression in T24T relative to T24 and a negative ratio represents overexpression by T24.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. These particular spreads were chosen because they exhibited the widest variety of the observed translocations for each cell type. To define chromosomal breakpoints, the inverted DAPI-banded images were aligned with the SKY-classified chromosomes. When breakpoints could not be easily identified using SKY, the identity of the fragment was established, when possible, using the CGH results (Fig. 2)
. For certain difficult, complex translocations, and as a quality check of the SKY results, individual chromosome paint probes were used in FISH for some chromosomes. SKY analysis of both T24 and T24T cells generated a complete karyotype for each cell line leaving no markers of unknown origin. The SKY results are based on a complete analysis of 10 metaphase spreads. A summary of the SKY results compared with the previously published karyotypes (13)
for T24 is shown in Table 1
, and the results for T24T appear in Table 2
. To be included in Tables 1
and 2
, the chromosome aberration had to be observed in at least 3 of 10 metaphases. In some cases, we were unable to determine the exact chromosomal breakpoints by SKY and inverted DAPI banding alone (indicated by (S) in the tables). However, in cases where a consistent aberration resulted in a significant genetic copy number change, the CGH ratio profile was used to establish the identity of the breakpoints [indicated by (C) in the tables].
|
|
|
|
18 abnormal chromosomes/cell. The composite karyotype based on SKY is 82, X, +1, +1, +3, +4, +5, +5, +6, +7, +8, +10, +11, +11, +12, +15, +16, +16, +17, +17, +19, +20, +20, +20, +22, t(2:11), t(2:20), t(5;11), iso 5p, t(7;8)a, t(12;18), t(X;13), i(13q), i(14q), t(15;18), t(X;21), 22q+(?), Xp+, del(5p), del(9p)a, del(9p)b, del(10p), del(15). We previously undertook (13)
a G-band analysis of T24 and identified five chromosomal alterations unique to T24: del (2)(p16), del (10)(p12), +11, i(14)(q10), add(X)(p22.3). SKY allowed us to confirm and/or redefine the classification of all of these markers. We redefined the del(2)(p16) as t(2;11) consisting of 2q and 11p. We confirmed the presence of del(10)(p11), +11, i(14)(q(10), and add(X)(p22.3). In addition we identified several additional T24-specific aberrant chromosomes listed in the composite karyotype above, which account for the remaining unidentified marker chromosomes in the previously published karyotypes.
T24T cells are hypertriploid with a modal number of 82–83 chromosomes and 18 abnormal chromosomes/cell. On the basis of our SKY analysis, T24T has a composite karyotype of 82–83, X, +1, +1, +2, +3, +4, +5, +5, +6, +6, +7, -8, -10, +11, +12, +13, +14, -15, +17, +19, +20, +20, +21, t(2;3), t(2;10), t(3;5), t(7;8)a, t(7:8)b, t(8;17), t(8;10;14), i(10q), t(11;22), 2xt(15;18), t(16;20), del(2), del(9p), del(11), 2xdel(15), del(16). The G-band analysis of T24T (13)
identified four chromosomal alterations unique to T24T: i(10)(q10), add(10)(p12), -15, del(X)(q23). SKY allowed us to confirm and/or redefine the classification of all of these markers. We redefined the add(10)(p12) as t(8;10;14)(q24; p11; q24) and del(X)(q23) as t(X;2). We confirmed the presence of i(10)(q10) and -15. Our previous karyotyping and G banding identified several chromosome alterations shared by both T24 and T24T: del(5)(p14), del(8)(p21), del(9)(p13), t(15;18)(q21;p11.2), add(21)(p11), +20. In this study, we were able to detect the presence of del(5)(p14) and the add(21)(p11) only in T24. However, based on our SKY analysis, we were able to confirm and/or redefine the classification of the rest of the reported chromosomal alterations. We redefined the del(8)(p21) as a t(7;8) and confirmed the presence of del(9)(p13), t(15;18)(q21;p11.2), and +20 in both cell lines.
CGH detected many genetic gains and losses in both cell lines compared with normal lymphocyte (tonsil) DNA (Fig. 2, A and B)
, the details of which are listed in Tables 1
and 2
. For analysis of CGH results, the normal chromosomal complement of T24 and T24T was assumed to be triploid. Gains common to both cell lines included regions on chromosome 1, 5p, 5q, 7q, 8q, 9q, 10q, 11p, 11q, 12q, 20, and 21q. Common losses were found at 9p and 10p. Differential labeling in chromosomal regions of distal 1p, 19, and 22 has been reported to result in artifactual deviations of the CGH profile. Unequal incorporation of fluorescein and Texas red-conjugated nucleotides into the sample DNAs from these regions was shown to cause these discrepancies (19)
. Therefore, gains in these regions must be confirmed using other techniques. In addition, deviations of the profile at centromeric regions are frequently observed in CGH and are probably because of suppression of hybridization to repetitive regions. These idiosyncrasies have been noted in the tables, and for these reasons, some apparent gains and losses have been omitted from the tables.
Given the phenotypic differences between T24 and T24T cells that we have previously reported, we were very interested in chromosomal copy number imbalances between these two cell lines. The comparative CGH shown in Fig. 2C
and summarized in Table 3
was performed to directly compare the chromosomal complement of the two lines. This comparative CGH was in close agreement with the differences that could be inferred from the separate CGH analyses performed in comparison to normal lymphocytes described above. Differences in T24T in comparison to T24 included gains on, 2p, 2q, 3p, 3q, 6, 7q, 10q, 12q, 17, 18q, 20, and 21q and losses on 4q, 8p, 8q, 10p, 11q, 13q, 14q, 22 (loss of 22 is suspected because of the differential labeling noted above), and X.
|
. Cytogenetic bands that had the greatest overall expression differences are listed in Table 4
, and this is compared with differential CGH results shown in Table 3
.
|
|
with those regions with chromosomal imbalances in Table 3
or with the presence of rearrangements detected by SKY in Tables 1
and 2
. Although some of the largest expression differences by cytogenetic band seem to correlate with genomic structural changes, many do not. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study demonstrates how SKY, in conjunction with CGH, can resolve and identify the complex karyotypes encountered when examining cancer cells from solid tumors. Not only was this approach able to assign an identity to every unidentified marker chromosome found in our previous study using conventional G banding, it also revealed the deficiencies in this technique when used in isolation. In combination, SKY and CGH are complementary techniques that can both confirm each other and provide overlapping nonredundant information.
For the most part, the SKY and CGH results from this study were in agreement, confirming the validity of the results. In those instances where CGH did not provide unequivocal support of the SKY data, FISH with single chromosome painting probes was performed to make an unambiguous chromosomal assignment (i.e., the multiple translocations involving chromosome 8.) The FISH data were also useful for resolving chromosomes that have significant spectral overlaps with the SKY probes. Some of the differences between SKY and CGH can be attributed to the size of cell populations being examined and the reporting of the results. For example, we have chosen to include in the tables, translocations that occur in at least 3 of 10 spreads examined by SKY for each cell line. It is possible that translocations found in only 30% of the cells might not be detectable as significant net copy changes by CGH. Also, gains or losses of subchromosomal regions that are not involved in translocation to another chromosome might go undetected by SKY but would be detected by CGH.
Several of the chromosomal abnormalities found in human bladder tumor specimens are also found in both T24 and T24T, including losses on 9p (6 , 10 , 25) and gains on 5p (11 , 25, 26, 27) , 8q (7 , 25, 26, 27) , and both arms of 20 (11 , 25, 26, 27, 28) . However, the most intriguing aspects of this study are defects found only in T24T that have been correlated with tumor progression by others. SKY found two translocations of chromosome 3 in T24T that were not present in T24, and CGH confirmed a gain of 3p. Richter et al. (7) have noted an association of 3p22-24 amplifications and subsequent tumor progression. Losses on 4q have also been correlated with advanced stage and grade (2) , and we found loss of chromosome 4 with greater loss at 4q26-q32 in T24T relative to T24. Koo et al. (25) have also noted losses of 4q13-23 but did not associate these with progression. Amplification of 6p has previously been noted in bladder cancer (11 , 25 , 29) , and both SKY and CGH found an increase in chromosome 6 in T24T relative to T24. Our results also indicate a greater loss of 8p in T24T, which coincides with observations that losses on 8p are correlated with higher tumor stage and grade (8 , 9) . Whether these and other genomic structural differences are merely correlative or are causative for the invasive phenotype can be determined by using gene transfer techniques in the T24/T24T model. Not surprisingly, our T24/T24T progression model does not have all genomic changes associated with human bladder cancer progression such as gains on 1q and losses from 2q, 10q, 11p, 17p, and 18q (6 , 7 , 10 , 27) . Also, an isochromosome for 5p, which has been associated with progression, was found in T24 but not T24T (10) .
A common assumption is that chromosomal rearrangements and imbalances seen in cancer cells result in differences in gene expression, resulting in an altered cellular phenotype. However, it is by no means clear what proportion of the genes within an affected region of a genomic structural change are altered at the level of expression or regulation. In this study, we have applied the tools now available for investigating the genome of cancer cells and find that the results challenge a simplistic view of these relationships. Several of the differences noted above between T24 and T24T cells, which have been implicated in bladder cancer progression, are manifested on the positional expression profiles (i.e., the decrease in copy number and expression for 8p). In other cases, there is a lack of correlation between chromosomal imbalances and the expression profile such as seen on 10q. Some discrepancies may either indicate an idiosyncrasy in the techniques or that biological mechanisms such as gene silencing are at work. For example, we observed a gain of 3p24 in T24T relative to T24 but no expression difference (log of T24T/T24 expression = -0.008). However, 3p26 showed the highest differential increase in gene expression for T24T relative to T24 of any cytogenetic band (Table 4)
. It is possible that the peak in expression seen at 3p26 will shift as cytogenetic localization of the genes that we have placed in this interval are updated. Alternatively, gains in one chromosomal region could possibly affect expression in adjacent regions.
The above example illustrates the weakness of current attempts at PEP. This is because annotation of the human genome sequence for cytogenetic positions is relatively incomplete and in many cases imprecise. The power of our method resides in the fact that as the annotation process catches up to the sequence data in the public databases such as LocusLink, it will be easy to update and refine our expression profiles. Indeed, when all of the sequencing gaps have been filled and the genome is fully annotated for cytogenetic markers, all of the loci will be weighed evenly. One foreseeable consequence of this refinement is that some differences in the expression profiles may become more pronounced, whereas others may diminish. Some shifting of the profiles along the chromosomes may also occur, but we purposely devised the weighting scheme to keep this to a minimum.
Two recent studies have addressed the question of whether gene copy number changes affect gene expression on a genomic scale. Lu et al. (30) developed a technique which they refer to as CESH. Analogous to CGH, CESH uses differentially labeled cDNAs from two test samples cohybridized to normal metaphase chromosome spreads to determine differential gene expression along a chromosome. An advantage of this technique is that the raw data directly reflects the gene expression for a given chromosomal region. However this technique has several limitations. One is that CESH was only able to detect changes in expression when several genes within a circumscribed region were differentially expressed by a considerable amount. Another limitation is that there is no direct method to identify the genes that are differentially expressed in a particular region. Herein is the fundamental advantage to our method of PEP. It is relatively simple to view the data on any scale desired from average differential expression across a band to differential expression of individual genes within a band. The only limitations on resolution are the number of genes on the array and the precision of the mapping data available from public databases. As mentioned above, as the resolution of both of these elements increases, our model for PEP will increase in precision as well.
Pollack et al. (31) have also devised an alternative method for PEP. These authors have developed a cDNA microarray-based CGH method. Specially constructed cDNA microarrays corresponding to clones mapped on a radiation hybrid panel were used to measure both DNA copy number and expression from the same set of genes. This approach has several distinct advantages, including the precision of the mapping data, a direct correlation between copy number of a genetic locus and its expression, and the potential for greater resolution than CGH for detecting chromosomal imbalances. However, this technique does require the production of specialized microarrays from radiation hybrid panels and protocols to increase sensitivity of the hybridization to the arrays. A significant advantage of our method, is that preexisting karyotypic data or CGH analysis can be compared with gene expression data from commercially available oligonucleotide arrays using relatively common and standardized techniques making this technique available to everyone.
The results from our combined multimodal approach suggest that structural and numerical differences in the chromosomal content of the T24 and T24T cell lines result in functional gene expression differences in some but not all cases. We are especially interested in the differences such as those on 8p, which has been implicated in bladder cancer progression in other studies. In addition, our results implicate chromosomal regions that have not been noted as frequently in other studies. For example, the numerical and expression level differences seen between 12p and 12q, as well as the X chromosome, suggest that a gene or genes on these chromosomes may play a causative role in differentiating the T24 and T24T phenotypes. Indeed, we have already shown that one candidate gene, RhoGDI2, which is located at chromosome 12p12.3, is expressed in T24 but not T24T and that its expression inversely correlates with the aggressiveness of other bladder cancer cell lines (14) . We are currently investigating this gene and others on chromosome 12 to determine whether they regulate the invasive phenotype. The presence of chromosomal regions that undergo numerical gains but do not show expression differences such as 10q is also strongly suggestive that a mechanism of gene silencing is at work in our model system. We are currently investigating the nature of this silencing and whether similar mechanisms play a role in human bladder cancer invasiveness.
Gene silencing by DNA methylation or imprinting may also explain another phenomenon that we have noticed in the positional expression profiles. The largest differences in gene expression by chromosomal region are predominantly attributable to overexpression in T24 relative to T24T, hence, the greater absolute values of the bands for T24 in Table 4
. This might argue for a role in silencing of highly expressed genes in T24T and not in T24. It might also explain why many gains in copy number in T24T cells are not reflected at the gene expression level. It is interesting to note that Liang et al. (32)
have recently reported that perturbation of DNA methylation has profound effects on gene expression and phenotype in at least one variant of T24 cells. In fact, abnormalities in gene expression levels because of epigenetic mechanisms like imprinting and methylation is a common characteristic of cancer in general (33
, 34)
. In addition, the consequences of genetic imprinting and the inequality of gene alleles is becoming increasingly evident as the role of uniparental disomy in some diseases other than cancer are discovered (35
, 36)
. It will be interesting to investigate whether DNA methylation patterns or uniparental origin of discrepant chromosomal regions correlate with gene expression patterns when viewed by chromosomal position and whether these phenomenon can explain the positional expression profiles that we have constructed here.
In conclusion, we have used a novel method of PEP to define regions of the genome that are functionally altered in cancer progression. In this manner, we hope to use this information to discover individual genes that control the switch from superficial to invasive bladder cancer. Additionally, this type of PEP might also be used to investigate epigenetic alterations and mechanisms on a genomic scale. Additional development of these techniques and testing of candidate genes will be carried out in the T24/T24T model system, which accurately reflects these important cancer phenotypes.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported, in part, by NIH CA75115-01, a scholarship from the Kimmel Foundation (to D. T.), and an NIH training grant (to J. J. G.). 
2 These authors contributed equally to this manuscript.
3 To whom requests for reprints should be addressed, at Box 800422, University of Virginia Health Sciences Center, Charlottesville, VA 22908. Phone: (434) 924-0042; Fax: (434) 982-3652; E-mail: dt9d{at}virginia.edu
4 The abbreviations used are: SKY, spectral karyotyping; CGH, comparative genomic hybridization; PEP, positional expression profiling; DAPI, 4',6-diamidino-2-phenylindole; FISH, fluorescence in situ hybridization; PBL, peripheral blood lymphocyte; CESH, comparative expressed sequence hybridization;.
5 Internet address: ftp://ncbi.nlm.nih.gov/refseq/LocusLink/LL.out_hs.gz.
6 Internet address: ftp://ncbi.nlm.nih.gov/genomes/H_sapiens/maps/mapview/ISCN800_abc.
Received 5/28/02. Accepted 9/25/02.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. C. Smith, B. Nicholson, M. Nitz, H. F. Frierson Jr, M. Smolkin, G. Hampton, W. El-Rifai, and D. Theodorescu Profiling Bladder Cancer Organ Site-Specific Metastasis Identifies LAMC2 as a Novel Biomarker of Hematogenous Dissemination Am. J. Pathol., February 1, 2009; 174(2): 371 - 379. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. C. Black, G. A. Brown, T. Inamoto, M. Shrader, A. Arora, A. O. Siefker-Radtke, L. Adam, D. Theodorescu, X. Wu, M. F. Munsell, et al. Sensitivity to Epidermal Growth Factor Receptor Inhibitor Requires E-Cadherin Expression in Urothelial Carcinoma Cells Clin. Cancer Res., March 1, 2008; 14(5): 1478 - 1486. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z Wu, K. McRoberts, and D Theodorescu The role of PTEN in prostate cancer cell tropism to the bone micro-environment Carcinogenesis, July 1, 2007; 28(7): 1393 - 1400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Smith, G. Oxford, A. S. Baras, C. Owens, D. Havaleshko, D. L. Brautigan, M. K. Safo, and D. Theodorescu Expression of Ral GTPases, Their Effectors, and Activators in Human Bladder Cancer Clin. Cancer Res., July 1, 2007; 13(13): 3803 - 3813. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Gao, K. Furge, J. Koeman, K. Dykema, Y. Su, M. L. Cutler, A. Werts, P. Haak, and G. F. Vande Woude Chromosome instability, chromosome transcriptome, and clonal evolution of tumor cell populations PNAS, May 22, 2007; 104(21): 8995 - 9000. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Havaleshko, H. Cho, M. Conaway, C. R. Owens, G. Hampton, J. K. Lee, and D. Theodorescu Prediction of drug combination chemosensitivity in human bladder cancer Mol. Cancer Ther., February 1, 2007; 6(2): 578 - 586. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T Yanagawa, T Funasaka, S Tsutsumi, H Watanabe, and A Raz Novel roles of the autocrine motility factor/phosphoglucose isomerase in tumor malignancy Endocr. Relat. Cancer, December 1, 2004; 11(4): 749 - 759. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. A. Furge, K. A. Lucas, M. Takahashi, J. Sugimura, E. J. Kort, H.-o. Kanayama, S. Kagawa, P. Hoekstra, J. Curry, X. J. Yang, et al. Robust Classification of Renal Cell Carcinoma Based on Gene Expression Data and Predicted Cytogenetic Profiles Cancer Res., June 15, 2004; 64(12): 4117 - 4121. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. C. Doherty, S. R. McKeown, V. McKelvey-Martin, C. S. Downes, A. Atala, J. J. Yoo, D. A. Simpson, and W. K. Kaufmann Cell Cycle Checkpoint Function in Bladder Cancer J Natl Cancer Inst, December 17, 2003; 95(24): 1859 - 1868. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
