Skip to main content
NCBI home page
Genome Research logoLink to Genome Research
. 2002 Jan;12(1):112–121. doi: 10.1101/gr.174202

Expression Profiles of 290 ESTs Mapped to Chromosome 3 in Human Epithelial Ovarian Cancer Cell Lines Using DNA Expression Oligonucleotide Microarrays

Emily N Manderson 1, Anne-Marie Mes-Masson 2,3, Jaroslav Novak 5, Peter D Lee 1,5, Diane Provencher 2,4, Thomas J Hudson 1,5,6,7, Patricia N Tonin 1,6,7,8
PMCID: PMC155251  PMID: 11779836

Abstract

We have investigated previously the utility of oligonucleotide expression microarray technology in an analysis of four spontaneously transformed epithelial ovarian cancer (EOC) cell lines, TOV-21G, TOV-81D, OV-90, and TOV-112D. Here, we examine the expression of 290 expressed sequence tags (ESTs) that map to human chromosome 3 in a primary culture derived from normal ovarian surface epithelium (NOSE), NOV-31, and the four spontaneously transformed EOC cell lines. One of these cell lines, OV-90, harbors a deletion of an entire chromosome 3p arm. Whereas the most aggressive cell lines (OV-90, TOV-112D, and TOV-21G) exhibited the highest levels of expression, assessed by the mean of expression values of all ESTs, OV-90 showed the lowest mean of expression of ESTs that map to the 3p arm in comparison with TOV-112D and TOV-21G. This difference in expression profile of 3p ESTs in OV-90 is also reflected in the ratio of expression of ESTs on 3p versus the 3q arm and in that the expression values of ESTs that map to 3p were more often lower than higher in OV-90 in two-way comparisons with NOV-31, TOV-21G, and TOV-112D. The loss of a 3p arm does not affect the pattern of differential expression in analyses based on the range of numeric expression values of each EST or fold differences in expression for each EST in comparison with NOV-31. However, 25 differentially expressed ESTs were identified on the basis of threefold differences in expression values between NOV-31 and any EOC cell line; and six of these ESTs were differentially expressed uniquely in OV-90. The investigation of these ESTs could facilitate the identification of novel chromosome 3 genes implicated in ovarian tumorigenesis.

Epithelial ovarian cancer (EOC) is the second most common malignancy of the female genital tract and is the fifth most common cancer in women (National Cancer Institute of Canada 2000). The prognosis of this disease is usually poor, as reflected in the high proportion of incident cases that result in death, as well as the fact that the 5 yr survival rate is just under 30%. This is predominantly due to the fact that EOC is often diagnosed at a late stage because of the lack of early warning symptoms. As a result, the molecular events underlying ovarian tumorigenesis remain largely unknown.

Cytogenetic studies have revealed that chromosome 3 is frequently rearranged in EOC, and these rearrangements often involve deletions of the short arm of chromosome 3 (Mertens et al. 1997). Loss of chromosome 3p in EOC has also been detected through loss of heterozygosity (LOH) studies and has been reported to occur at a frequency ranging from 13%–52%, at various loci tested (Ehlen and Dubeau 1990; Yang-Feng et al. 1992; Dodson et al. 1993; Cheng et al. 1996; Lounis et al. 1998; Fullwood et al. 1999). Functional evidence in support of a 3p gene important in EOC comes from a study by Rimessi et al. (1994) showing that microcell-mediated chromosome transfer (MMCT) of chromosome 3 into an ovarian cancer cell line induced senescence and growth arrest as well as suppression of tumorigenicity. Frequent over-representation of 3q26 in ovarian cancer has been detected by comparative genomic hybridization (CGH) in several studies (Arnold et al. 1996; Sonoda et al. 1997; Sugita et al. 2000). CGH also showed that PIK3CA (3q26) is frequently increased in copy number, which may lead to increased transcription in ovarian cancer (Shayesteh et al. 1999). The combined results of cytogenetic, LOH, MMCT, and CGH analyses suggest that there is one or more genes located on chromosome 3 that are implicated in ovarian tumorigenesis.

We have utilized an in vitro model system to study the molecular genetic events important in EOC. This model system is based on the establishment of four spontaneously transformed EOC cell lines (TOV-21G, TOV-81D, OV-90, and TOV-112D) that display the phenotypes of the original tumors from which they were derived (Provencher et al. 2000). Recently, we analyzed this EOC model system using the Hs6000 DNA expression microarrays from Affymetrix (Tonin et al. 2001). Our results indicated that this new technology is a valid approach to study ovarian cancer in that the patterns of gene expression detected by the microarray are consistent with the phenotypes of the EOC cell lines.

The objective of this study is to examine the expression of ESTs on chromosome 3 in human EOC cell lines using DNA expression microarrays. Here we describe the expression patterns of 290 ESTs that map to chromosome 3 in a primary culture derived from normal ovarian surface epithelium (NOSE), NOV-31, and the four spontaneously transformed EOC cell lines. The OV-90 cell line has been shown by both karyotype and LOH analyses to harbor the complete loss of one chromosome 3p arm (Lounis et al. 1998; Provencher et al. 2000). Two methods of analysis were applied to analyze the expression data collected using the Hs6000 microarray. The first method is based on examining the overall expression of ESTs that map to chromosome 3, whereas the second method is based on examining the expression of individual ESTs in the cell lines in comparison with NOV-31.

RESULTS

A list of 290 ESTs that map uniquely to chromosome 3 was created by verifying the map information listed on the UniGene-Homo sapiens database in May 2000. There is an almost equal number of ESTs that map to the chromosome 3p arm and the chromosome 3q arm, 140 and 150 respectively. We analyzed the expression data of 290 ESTs that map to chromosome 3 in two different ways. First, we examined the overall expression of ESTs in the four EOC cell lines and NOV-31; secondly we identified ESTs that are differentially expressed between the EOC cell lines and NOV-31.

Examination of the Overall Expression of ESTs That Map to Chromosome 3

We calculated the mean of the expression values of ESTs (mean expression) that map to the entire chromosome 3 homolog and to each arm for each sample (Fig. 1A). The results show an increase in overall expression of ESTs that map to chromosome 3 in the three aggressive cell lines, OV-90, TOV-112D, and TOV-21G in comparison with the primary culture of NOSE, NOV-31. The least aggressive cell line, TOV-81D, shows levels of expression similar to that of NOV-31. The mean of the expression values of ESTs on the 3p arm for cell line OV-90 is greater than that for NOV-31, however, it is lower than the mean for the two other aggressive cell lines. There also appears to be an increase in the expression of ESTs that map to the 3q arm in OV-90 in comparison with NOV-31 and the other cell lines.

Figure 1.

Figure 1

Open in a new tab

Analysis of the mean of the expression values of chromosome 3 mapped ESTs. The mean was calculated by adding the expression values of all ESTs and dividing by the total number of ESTs for each sample. (A) The mean of the expression values of ESTs that mapped to the entire chromosome 3 and each arm separately was calculated for NOV-31 and the four EOC cell lines. (B) A comparison was made between NOV-31 and the EOC cell lines by dividing the mean of the expression values for each cell line by the mean of the expression values for NOV-31. (C) The results of comparisons of the mean of the expression values of ESTs that map to the 3p arm to the mean of the expression values of ESTs that map to the 3q arm. The ratio was calculated by dividing the mean of the expression values of ESTs that mapped to the 3p arm by that of the 3q arm for each sample.

In an alternate analysis, we compared the mean of the expression values of ESTs for each cell line with that of NOV-31 (Fig. 1B). The results show that TOV-81D, TOV-112D, and TOV-21G have the same level of decrease or increase in expression in comparison with NOV-31, whether one examines the mean of expression of ESTs on the entire homolog or each arm separately. However, for OV-90, the ratio for ESTs that map to the 3p arm is lower than the ratio for the 3q arm. In Figure 1C, we compared the mean expression of ESTs on the 3p arm with that of the mean expression of ESTs on the 3q arm in each sample. This analysis was performed to establish whether there is a difference in the overall expression of ESTs between the two arms in the four EOC samples and NOV-31. This analysis has the advantage of comparing the expression of ESTs within a given sample (one microarray and one scan), instead of comparing across samples (multiple microarrays and multiple scans). The results show that the ratio approaches 1.0 for all samples with the exception of OV-90, for which the ratio is 0.74. Figure 1C also indicates that the ratio of the mean expression of 3p ESTs to the mean expression of 3q ESTs is the same in NOV-31 and the EOC cell lines, with the exception of OV-90.

Two-way comparisons were performed between OV-90 and each of the other samples to examine the effect of the loss of one 3p arm on the expression of individual ESTs in cell line OV-90. We counted the number of ESTs on each arm for which the expression value in OV-90 was either higher or lower than that of the other sample (Fig. 2). The results for the ESTs that map to the 3p arm (Fig. 2A) show that there are more ESTs for which the expression value in OV-90 is lower than the value in NOV-31, TOV-112D, or TOV-21G. In the majority of cases in which the expression value is lower in OV-90, the magnitude of the difference is <50. The expression values of ESTs on the 3p arm are more often higher in OV-90 than in TOV-81D, and this is consistent with the finding that TOV-81D has the lowest mean of expression values of ESTs that map to chromosome 3 (Fig. 1A). The results for ESTs that map to chromosome 3q (Fig. 2B) show that there are more instances in which the expression value of an EST is higher in OV-90 in the two-way comparisons with all other samples; and this result is consistent with the observation that OV-90 has the highest mean of expression values of ESTs that map to chromosome 3q (Fig. 1A).

Figure 2.

Figure 2

Open in a new tab

Two-way comparisons of the numeric expression value of individual ESTs between OV-90 and the other samples. The number of ESTs for which the expression value in OV-90 was lower than, higher than, or equal to the value in the sample with which it was being compared was counted. Expression values that were negative or zero were standardized to a minimum value of one for the purpose of making comparisons. The analysis was performed for ESTs that mapped to the 3p arm (A) and for ESTs that mapped to the 3q arm (B).

Comparison of the Expression of Individual ESTs Between the EOC Cell Lines and NOV-31

The methods of analysis used to examine the expression profile of individual ESTs that map to chromosome 3 are based on differences in the range of expression values for each EST or on ESTs displaying at least a threefold difference in expression in any EOC cell line in comparison with NOV-31. A representative sample of comparisons using either method is shown in Figure 3, and the analysis of the entire data set is available at http://genome.mcgill.ca/ovarian.

Figure 3.

Figure 3

Open in a new tab

A schematic representation of selected ESTs that are differentially expressed on the basis of their range of numeric expression value (Panel A) or at least a threefold difference in expression between NOV-31 and any EOC cell line (Panel B). The ESTs are displayed from the 3p telomere to the 3q telomere in order of genetic and cytogenetic map position on the basis of information retrieved from the UniGene-Homo sapiens database, NCBI. A box around two ESTs indicates that they represent the same gene. The figure that displays the expression profiles for all 290 ESTs is available at http://genome.mcgill.ca/ovarian.

Figure 3, panel A displays the ESTs that are differentially expressed based on their numerically assigned expression values categorized on the range of expression divided into five intervals: 1–49, 50–199, 200–499, 500–1000, and >1000 (Fig. 3). A total of 114 ESTs are differentially expressed based on the range of numeric expression value. The number of ESTs that are differentially expressed between NOV-31 and each cell line is listed in Table 1. In the two-way comparisons, a similar number of ESTs were differentially expressed between NOV-31 and each aggressive cell line, whereas the fewest number of differences were observed between NOV-31 and TOV-81D (Table 1).

Table 1.

Number of ESTs That Map to Chromosome 3 and are Differentially Expressed Between NOV-31 and Each EOC Cell Linea

No. of differences based on range of numeric expression value No. of three-fold or greater differences in expression
NOV-31 vs. TOV-81D 37  2
NOV-31 vs. OV-90 62 10
NOV-31 vs. TOV-112D 56 10
NOV-31 vs. TOV-21G 64 12
Open in a new tab
a

The number of differentially expressed ESTs was calculated on the basis of two analyses performed in Fig. 3. The first column lists the number of ESTs that were differentially expressed based on their numerically assigned expression values that were categorized on the basis of range of expression divided into the following intervals: 1–49, 50–199, 200–499, 500–1000 and >1000. The second column lists the number of ESTs that displayed at least a three-fold difference in expression between NOV-31 and any EOC cell line. The fold difference calculations were performed after the data was standardized to a minimum value of 50. 

The second method involves calculating the number of ESTs that are differentially expressed on the basis of at least a threefold difference in expression between any EOC cell line and NOV-31 (Fig. 3, panel B). As the expression values that fall below 50 have a high degree of variability (T.J. Hudson, pers. comm.), we calculated the fold difference after standardizing all data to a minimum value of 50. A total of 25 ESTs (Table 2) are differentially expressed with at least a threefold difference, which represents 9% of the total ESTs that were examined (two ESTs represent the transferrin receptor gene). The number of ESTs that are differentially expressed between NOV-31 and each cell line is listed in Table 1. There are a similar number of ESTs that are differentially expressed in the two-way comparisons between NOV-31 and each aggressive cell line, and the fewest number of ESTs with a threefold difference in expression were detected in the comparison between TOV-81D and NOV-31. Only six ESTs are differentially expressed uniquely in the comparison between NOV-31 and OV-90 (Fig. 3, panel B).

Table 2.

A List of ESTs That Were Found to be Differentially Expressed Based on Their Range of Numeric Expression Value and At Least a Threefold Difference in Expression Between NOV-31 and Any EOC Cell Line

Accession no. Gene symbol Description Functionb Expression values and reliability scoresa
NOV-31 TOV-81D TOV-112D OV-90 TOV-21G
D42043 KIAA0084 Hypothetical protein kiaa0084 Unknown. 181 P 178 P 11 A 28 P 55 P
X77777 VIPR1 vasoactive intestinal peptide receptor 1 This is a receptor for vasoactive intestinal polypeptide. The activity of this receptor is mediated by G-proteins which activate adenylyl cyclase. 192 P −27 A −41 A −93 A −18 A
M83751 ARP arginine rich protein Unknown. 112 P 213 P 166 P 876 P 108 A
R15944 GC20 translation factor sui1 homolog Translation factor. 90 P 40 P 288 P 114 P 112 P
R42501 IMPDH2 Inosine monophosphate dehydrogenase-2 IMP is the rate limiting enzyme in the de novo synthesis of guanine nucleotides and is involved in the regulation of cell growth. It may also have a role in the development of malignancy and the growth progression of some tumors. 7 A 50 P 217 P 14 A 43 A
R55310 UQCRC1 ubiquinol-cytochrome c reductase core protein I This is a component of the ubiquinol-cytochrome c reductase complex (complex iii or cytochrome b-c1 complex), which is part of the mitochondrial respiratory chain. This protein may mediate formation of the complex between cytochromes c and c1. 23 A 43 M 160 P 50 M −257 A
R46069 EST Unknown. 44 P 15 A 162 P 80 P −3 A
H05285 QARS glutamine-tRNA synthetase Part of a multisubunit complex that groups tRNA ligases for arg, asp, glu, gln, ile, leu, lys, met and pro. 32 P 15 A 168 P 32 P 79 P
T70046 FLNB filamin B, beta (actin-binding protein-278) Dimeric actin cross-linking phosphoprotein of the peripheral cytoplasm, interacts with the cytoplasmic tail of glycoprotein GP1BA. 320 P 190 P 61 P 944 P 170 P
R88575 TKT transketolase (WernickeKorsakoff syndrome) Key enzyme in the pentose-phosphate pathway. 36 P 59 P 137 P 113 P 417 P
X12892 PROS1 protein S (alpha) Protein S, alpha coagulation factor (cofactor of proC), vitamin K-dependent, ligand for TYRO3, AXL family of receptor tyrosine kinase. 8 A 32 P 77 P 73 P 241 P
H56192 NR1D2 nuclear receptor subfamily 1, group D, member 2 Binds to the sequences 5′-aatgtaggtca-3′ and 5′-ataactaggtca-3′. Acts as a potent competitive repressor of ror α function (by similarity). 10 A 10 A 15 P 7 A 609 P
M37984 TNNC1 troponin C1, slow Troponin is the central regulatory protein of striated muscle contraction. 10 A 17 A 50 P 98 P 205 P
X04828 GNAI2 guanine nucleotide binding protein (G protein), alpha inhibiting activity polypeptide 2 Guanine nucleotide-binding proteins (G-proteins) are involved as modulators or transducers in various transmembrane signaling systems. 198 P 122 P 101 P 84 A 16 A
H11650 ARF4 ADP-ribosylation factor 4 GTP-binding protein that functions as an allosteric activator of the cholera toxin catalytic subunit, an ADP-ribosyltransferase. Involved in protein trafficking. 391 P 548 P 225 P 228 P 19 A
T55741 MYLK myosin, light polypeptide kinase Phosphorylates specific serine in N terminus of a myosin light chain. 290 P 35 P 13 A −4 A −3 A
Y00281 RPN1 ribophorin 1 Essential subunit of N-oligosaccharyl transferase. 147 P 167 P 203 P 556 P 178 P
H22939 GYG glycogenin Protein primer required for the biogenesis of muscle glycogen. Elongation of the primer is effected by glycogen synthase. 191 M 113 P 210 M 57 A 120 P
H41528 DLG1 discs, large (Drosophila) homolog 1 Interacts with the cytoplasmic tail of nmda receptor subunits. Associates with protein 4.1 154 P 61 P 49 P 29 A 56 P
R23889X01060 TFRC transferrin receptor Cellular uptake of iron occurs via receptor-mediated endocytosis of ligand-occupied transferrin receptor. 254 P  80 P  392 P 237 P 114 P 45 P  780 P  288 P  283 P  151 P 
R22228 HRY hairy (Drosophila)homolog Drosophila hairy pair rule gene homolog, also homolog to rat RHL, an immediated early response gene to growth factor. −5 A 1 A 22 P 332 P 27 P
U25441 DRD3 dopamine receptor D3 This is one of five types (d1 to d5) of receptors for dopamine. The activity of this receptor is mediated by G-proteins which inhibit adenylyl cyclase. −13 A 10 A −31 A −4 A 175 A
H20503 NDUFB7 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 7 (18 kD, B18) Transfer of electrons from NADH to the respiratory chain. The immediate electron acceptor for the enzyme is believed to be ubiquinone. 28 P 12 A 48 P 62 P 294 P
H88250 SFRS10 splicing factor, arginine/serine-rich (transformer 2 Drosophila homolog) 10 Splicing factor. 41 P 45 P 148 P 116 P 156 P
Open in a new tab
a

Numeric expression values before standardization and the reliability score for each expression value were generated by the Affymetrix GeneChip software based on hybridization signals for probe sets representing each EST. 

b

The function was retrieved from the Weizman Gene Card database (www.bioinfo.weizman.ca). 

DISCUSSION

In this study, we analyzed both the overall expression and the individual expression profiles of 290 ESTs that map to chromosome 3 in four EOC cell lines and NOV-31. When we examined the overall expression of ESTs that map to the entire chromosome 3 and each arm separately, we observed a trend to over-expression of ESTs in the three aggressive cell lines. However, OV-90 shows the lowest mean of expression of ESTs that map to the 3p arm in comparison with TOV-21G and TOV-112D. This difference in expression profile of 3p ESTs in OV-90 is also revealed by the ratio of expression of ESTs on the 3p arm versus the 3q arm and the finding that the expression values of ESTs that map to 3p are more often lower than higher in OV-90 in two-way comparisons with NOV-31, TOV-21G, and TOV-112D. The loss of a 3p arm in OV-90 does not affect the pattern of expression in analyses on the basis of the range of numeric expression values of each EST or fold differences in expression for each EST in comparison with NOV-31. Similar results were observed in an analysis of the impact of trisomy 17 on expression profiles of EOC cell lines using oligonucleotide microarrays (P.N. Tonin, unpubl.).

The fewest number of differences were observed between the expression profiles of NOV-31 and the least aggressive cell line, TOV-81D, in all methods of analysis. This is consistent with our previous findings that TOV-81D displayed growth characteristics similar to that of NOSE in that it is unable to form colonies in soft agarose or to form tumors in nude mouse xenografts (Provencher et al. 2000). This cell line was derived from an ovarian tumor sample from a patient with indolent disease with >9-yr survival (Provencher et al. 2000). The similar expression profiles of TOV-81D and NOV-31 of chromosome 3 ESTs is also consistent with the analysis of 6416 ESTs represented on the Hs6000 expression microarrays regardless of their chromosome location (Tonin et al. 2001).

In the examination of the expression profiles of the 290 ESTs, we identified 114 of these ESTs as differentially expressed on the basis of their range of numerical expression values, and 25 ESTs as differentially expressed with at least a threefold difference between NOV-31 and any EOC cell line. The majority of the ESTs (n = 92) that are not differentially expressed have expression values below 50 across all samples and have a low-reliability score (ambiguous) on the basis of the Affymetrix GeneChip software analysis. Noteworthy is that these ESTs were expressed with a high-reliability score in other human tissues and cell lines in independent experiments by use of the Hs6000 microarray (P.D. Lee, pers. comm.). One interpretation is that these ESTs are expressed at a low level in both NOSE and the four EOC cell lines. Seven ESTs were expressed at levels >50 with a high-reliability score, and are also not differentially expressed on the basis of our analyses. As all of these ESTs are not differentially expressed in EOC cell lines and NOV-31, they may not play a significant role in ovarian tumorigenesis.

Several ESTs are differentially expressed on the basis of both methods of analysis displayed in Figure 3 and, therefore, may be considered as interesting candidates for further analysis (Table 2). For example, the EST representing the arginine-rich protein (ARP) is overexpressed in OV-90 (Fig. 3). This gene has been implicated as a possible oncogene, in that mutations were identified in the trinucleotide repeat in 11 of 37 pancreatic tumors (Shridhar et al. 1997). However, others have proposed that these sequence changes represent normal variations rather than tumor-specific changes (Evron et al. 1997; Tanaka et al. 2000). The region that was found to be variable was sequenced in OV-90 and no sequence variations were detectable (E.N. Manderson, unpubl.). Another example is the EST representing the gene for guanine nucleotide-binding protein α-inhibiting polypeptide 2 (GNAI2 or Gi2) that was found to be underexpressed threefold in TOV-21G in comparison with NOV-31 (Fig. 3). Lyons et al. (1990) reported mutations of GNAI2 in 3 of 10 endocrine tumors of the ovary. However, the relationship of these mutations to the expression levels of mRNA was not examined in their study. The mutations were found in Arg 179, a conserved amino acid among G-proteins. This arginine is the cognate of Arg 201, which is the target for cholera toxin-catalyzed adenosine diphosphate (ADP)-ribosylation of the α subunit of the G-protein Gs (Cassel and Pfeuffer 1978; Abood et al. 1982). Interestingly, in TOV-21G, the EST that represents the gene for ADP-ribosylation factor 4 (ARF-4) is also down-regulated (Fig. 3). ARF-4 is a guanine nucleotide-binding protein that stimulates the ADP-ribosyltransferase activity of the cholera toxin in vitro (Lebeda and Haun 1999). Two other examples are the ESTs representing vasoactive intestinal peptide receptor 1 and myosin light polypeptide kinase that are consistently underexpressed with a threefold difference in the four EOC cell lines in comparison with NOV-31, and therefore may play a role in EOC (Fig. 3).

A number of ESTs that are differentially expressed on the basis of either their range of numeric expression values or fold difference in expression between NOV-31 and any EOC cell line have been implicated previously in cancer on the basis of conventional methods of gene expression analysis. For example, the EST representing catenin β 1 (CTNNB1) was found to be overexpressed in TOV-21G, on the basis of its numeric expression values (Fig. 3, Panel A). The expression of CTNNB1 was found to be variable among 34 ovarian adenocarcinomas by immunohistochemistry, which is consistent with our results (Davies et al. 1998). Sagae et al. (1999) detected mutations in one mucinous and four endometrioid ovarian adenocarcinomas that resulted in an accumulation of the protein within the cells. These findings suggest that increased levels of catenin β 1 protein could be involved in the development of EOC by escaping regulation by the APC protein, and thus may play a role in the tumorigenesis of cell line TOV-21G. Therefore, we recommend analyzing the expression data using both methods to identify ESTs that may be important in ovarian tumorigenesis.

A review of the literature revealed that the expression of some of the genes represented on the microarray has been studied previously in EOC by use of conventional techniques such as Northern blot analysis and immunohistochemistry (Table 3). For some genes, the results obtained by the microarray are consistent with the level of expression reported in the literature. For example, topoisomerase 2 β (TOP2B) expression was found to vary in 15 tumors by Northern blot analysis (Withoff et al. 1999) and in 37 tumors by RNAse protection assay (Cornarotti et al. 1996). We also observed variable expression of TOP2B in the EOC cell lines by microarray analysis (Fig. 3). In some cases, the expression profile of some genes reported in the literature was not concordant with levels of detection by microarray analysis. For example, the EST representing transforming growth factor β receptor type 2 (TGFBR2) revealed an absence of expression in the normal sample and the four EOC cell lines (Fig. 3). However, we have reported that this gene is expressed in the four EOC cell lines by Northern blot and RT–PCR analyses (Manderson et al. 2000). Also, others have detected its expression in ovarian tumors and cell lines by RT–PCR and in situ hybridization (Henriksen et al. 1995; Bartlett et al. 1997; Hu et al. 2000). The probe set for the Hs6000 microarray was designed on the basis of the complete cDNA of TGFBR2 (Affymetrix). The discordance in the microarray results and the Northern blot analysis serves to illustrate the limitations of microarray technology.

Table 3.

Correlation Between Microarray Expression Analysis and Expression Studies in EOC Using Conventional Techniques Such as Northern Blot, RT-PCR or Immunohistochemistry

Accession no. Gene symbol Name Data reported in literature Expression values and reliability scoresa
NOV-31 TOV-81D TOV-112D OV-90 TOV-21G
Z15115 TOP2B topoisomerase 2 β Northern blot analysis revealed variable expression in 15 solid ovarian tumors (Withoff et al. 1999). Intertumoral variability of TOP2B mRNA levels was detected by RNAse protection assay and no significant increase in expression was detected in tumors compared to normal ovaries (Cornarotti et al. 1996). 49 P 47 P 117 P 53 P 116 P
M85079 TGFBR2 transforming growth factor beta receptor type 2 Expression of TGFBR2 was detected in 14 cell lines by RT–PCR (Hu et al. 2000). Northern blot analysis revealed differential expression of TGFBR2 in four EOC cell lines and one solid tumor (Manderson et al, 2000). Expression was detected in 92% of tumors by RT–PCR (Bartlett et al. 1997). −53 A −70 A −67 A −170 A −62 A
R41937 MLH1 MutL (E. coli) homolog 1 Immunohistochemistry revealed positive staining in a large fraction of malignant cells in 54 solid tumors (Samimi et al. 2000). Colella et al (1998) were unable to detect MLH1 protein by Western blot in four of 20 tumors and 5 of 9 cell lines. 10 A 20 M 34 P 22 P 20 P
X87838 CTNNB1 catenin β 1 Four of twenty-seven tumors showed positive staining by immunohistochemistry, all four harbored mutations in the gene (Sagae et al. 1999). Expression of CTNNB1 was found to be variable among 34 adenocarcinomas of the ovary by immunohistochemistry (Davies et al. 1998). 192 P 98 P 153 P 155 P 335 P
R36644 ACVR2B activin A receptor, type IIB Expression was observed by RT–PCR in 100 percent of tumor samples (Zheng et al. 1998). Western blot detected ACVR2B protein in OVCAR-3 and SKOV3 (Ito et al. 2000). 44 M 29 A 100 P 92 A 53 M
M81933 CDC25A cell division cycle 25A CDC25A mRNA was detected by RT–PCR in nine of 14 cell lines (Hu et al. 2000). 43 P 16 A 90 P 60 A 36 A
H59260 4 A −6 A 12 A 5 A −4 A
R43361 RYK RYK receptor-like tyrosine kinase In situ hybridization revealed minimal to absent expression of RYK in the 29 P 39 P 33 A 92 P 64 P
X69970  surface epithelium of normal ovaries. Expression was elevated primarily in the epithelium of malignant tumors (Wang et al. 1996). 25 P 39 M 31 P 69 P 27 M
Z29090 PIK3CA phosphatidylinositol 3-kinase catalytic subunit, alpha isoform Increased transcription detected by quantitative PCR in one ovarian cancer cell line (Shayesteh et al. 1999). 10 A −3 A 11 A 3 A 14 A
R23889 TFRC transferrin receptor Immunohistochemistry revealed heterogenous expression of TFRC 254 P 392 P 114 P 780 P 283 P
X01060  (Berchuck et al. 1990). 80 P 237 P 45 P 288 P 151 P
H60074 10 A 15 A 24 A 6 A −29 A
M11507 7 A −8 A 7 A −1 A −4 A
H26419 APOD apolipoprotein D 18 of 68 ovarian tumors stained positively for APOD expression (Vazquez et al. 2000). 28 A −6 A 19 A 29 A 33 A
Open in a new tab
*

Numeric expression values before standardization and the reliability score for each expression value were generated by the Affymetrix GeneChip software based on hybridization signals for probe sets representing each EST. 

In conclusion, we have shown that patterns of gene expression assayed by oligonucleotide microarrays can be related to chromosomal anomalies. However, simple chromosomal anomalies, such as allelic loss, may not affect the patterns of expression in analyses based on range of numeric expression value or fold difference in expression values. This result is encouraging because it indicates that if one is using microarray technology to identify candidate genes on the basis of differential expression in a sample with LOH, chromosome loss may not impact significantly on the comparative analyses between normal versus tumor tissue. This is exemplified by our observation that we identified a subset of ESTs that are differentially expressed between NOV-31 and any of the four EOC cell lines, and only six of these are uniquely expressed with a threefold difference in expression between NOV-31 and OV-90. In addition, we were able to identify two ESTs that are underexpressed in all four EOC cell lines in comparison with NOV-31. It will be important to validate the differential expression patterns of candidate chromosome 3 ESTs by examining their expression profiles in a larger panel of EOC tumor samples and cell lines to determine whether they play a significant role in ovarian tumorigenesis.

METHODS

Source of RNA

RNA was extracted from a primary culture of NOSE (NOV-31) and four spontaneously transformed EOC cell lines (TOV-21G, TOV-81D, OV-90, and TOV-112D) with TRIsol reagent (GIBCO-BRL, Life Technologies, Inc.) as described in Tonin et al. (2001). Briefly, the EOC cell lines were maintained in OSE medium consisting of 50:50 medium 199:105 (Sigma), supplemented with 10% FBS and 2.5 μg/mL of fungizone and 50 μg/mL of gentamicin (Kruk et al. 1990). The four EOC cell lines have been characterized extensively and described elsewhere (Provencher et al. 2000).

Microarray Data Collection and Analysis

The expression data was derived by use of the Hs6000 DNA microarray (Affymetrix) as described in Tonin et al. (2001). In summary, the hybridization target was prepared from 20 μg of total RNA that was reverse transcribed to double-stranded cDNA using oligo-dT primer containing a T7 RNA polymerase-binding site. This cDNA was then in vitro transcribed to cRNA using biotinylated dUTP and dCTP. The resulting cRNA target represents a labeled 50–1000-fold linear amplification of the cDNA sample. To reduce secondary structure, the target was fragmented in 40 mM Tris acetate, 100 mM potassium acetate, and 30 mM MgCl, (pH 8.1) at 95°C. Hybridizations were performed with 15 μg of target on Affymetrix Hs6000 oligonucleotide microarrays containing probe sets for 6416 human genes (5223 known genes and 1193 ESTs). Following washing and staining, microarrays were scanned using a Hewlett Packard GeneArray scanner. Gene expression levels were calculated for each EST from the scanned image by Affymetrix GeneChip software, giving a single average difference ratio (expression value) across 20 probe pairs as well as a reliability score [Ambiguous (A), Present (P) or Marginal (M) based on the variability of hybridization with each probe set]. Expression values that were negative or zero were standardized to a minimum value of one. A query was designed to retrieve the corresponding UniGene name (cluster) for each GenBank accession number that corresponded to each EST represented on the microarray from UniGene-Homo sapiens, National Center for Biotechnology Information (NCBI) (www.ncbi.nlm.nih.gov/UniGene) in November 1999. Additional information was retrieved for each cluster, including the gene symbol, description of the gene, cytogenetic position, and location on Gene Map 98. A list of ESTs that map to chromosome 3 was generated from the mapping information collected for all ESTs on the microarray, and all information was verified and updated using the UniGene-Homo sapiens (NCBI) website in May 2000. The genetic map location of each EST was determined by the positions in cM of markers on chromosome 3 obtained from a comprehensive genetic map (Dib et al. 1996), which is available on the Genethon server (www.genethon.fr).

WEB SITE REFERENCES

www.bioinfo.weizman.ca

www.ncbi.nlm.nih.gov/UniGene

www.genethon.fr

Acknowledgments

We thank Todd Golub for microarray analysis, Louise Champoux for technical assistance, and Nadège Presneau for her insight and comments. This research was supported by a grant from the Medical Research Council of Canada (MRC) (to P.N.T., A.-M.M.-M., and D.P.), a grant from the MRC (to A.-M.M.-M., P.N.T., T.J.H., and D.P.), and a grant from the Canadian Genetic Diseases Network, Federal Networks of Centres of Excellence Program (to T.J.H.). E.N.M. is the recipient of a studentship from the Natural Sciences and Engineering Research Council of Canada, the Canadian Institutes of Health Research and the Fonds de Recherche en Santé du Québec (FRSQ); A.-M.M.-M. is a recipient of a Chercheur National Fellowship from the FRSQ; D.P. is a recipient of a Chercheur-Cliniciens Senior from the FRSQ; T.J.H. is a recipient of a Clinician Scientist Award from the MRC, and P.N.T. is a scholar of the MRC and the Cancer Research Society, Inc.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.

Footnotes

E-MAIL tonin@med.mcgill.ca; FAX (514) 934-8273.

Article and publication are at http://www.genome.org/cgi/doi/10.1101/gr.174202.

REFERENCES

  1. Abood ME, Hurley JB, Pappone M-C, Bourne HR, Stryer L. Functional homology between signal-coupling proteins. Cholera toxin inactivates the GTPase activity of transducin. J Biol Chem. 1982;257:10540–10543. [PubMed] [Google Scholar]
  2. Arnold N, Hagele L, Walz L, Schempp W, Pfisterer J, Bauknecht T, Kiechle M. Overrepresentation of 3q and 8q material and loss of 18q material are recurrent findings in advanced human ovarian cancer. Genes Chromosomes Cancer. 1996;16:46–54. doi: 10.1002/(SICI)1098-2264(199605)16:1<46::AID-GCC7>3.0.CO;2-3. [DOI] [PubMed] [Google Scholar]
  3. Bartlett JM, Langdon SP, Scott WN, Love SB, Miller EP, Katsaros D, Smyth JF, Miller WR. Transforming growth factor-β isoform expression in human ovarian tumours. Eur J Cancer. 1997;33:2397–2403. doi: 10.1016/s0959-8049(97)00304-3. [DOI] [PubMed] [Google Scholar]
  4. Berchuck A, Olt GJ, Soisson AP, Kamel A, Soper JT, Boyer CM, Clarke-Pearson DL, Leslie DS, Bast RC., Jr Heterogeneity of antigen expression in advanced epithelial ovarian cancer. Am J Obstet Gynecol. 1990;162:883–888. doi: 10.1016/0002-9378(90)91288-n. [DOI] [PubMed] [Google Scholar]
  5. Cassel D, Pfeuffer T. Mechanism of cholera toxin action: Covalent modification of the guanyl nucleotide-binding protein of the adenylate cyclase system. Proc Natl Acad Sci. 1978;75:2669–2673. doi: 10.1073/pnas.75.6.2669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cheng PC, Gosewehr JA, Kim TM, Velicescu M, Wan M, Zheng J, Felix JC, Cofer KF, Luo P, Biela BH, et al. Potential role of the inactivated X chromosome in ovarian epithelial tumor development. J Natl Cancer Inst. 1996;88:510–518. doi: 10.1093/jnci/88.8.510. [DOI] [PubMed] [Google Scholar]
  7. Colella G, Vikhanskaya F, Codegoni AM, Bonazzi C, D'Incalci M, Broggini M. hMLH1 and hMSH2 expression and BAX frameshift mutations in ovarian cancer cell lines and tumors. Carcinogenesis. 1998;19:691–694. doi: 10.1093/carcin/19.4.691. [DOI] [PubMed] [Google Scholar]
  8. Cornarotti M, Capranico G, Bohm S, Oriana S, Spatti GB, Mariani L, Ballabio G, Zunino F. Gene expression of DNA topoisomerases I, II α and II β and response to cisplatin-based chemotherapy in advanced ovarian carcinoma. Int J Cancer. 1996;67:479–484. doi: 10.1002/(SICI)1097-0215(19960807)67:4<479::AID-IJC3>3.0.CO;2-P. [DOI] [PubMed] [Google Scholar]
  9. Davies BR, Worsley SD, Ponder BA. Expression of E-cadherin, α-catenin and β-catenin in normal ovarian surface epithelium and epithelial ovarian cancers. Histopathology. 1998;32:69–80. doi: 10.1046/j.1365-2559.1998.00341.x. [DOI] [PubMed] [Google Scholar]
  10. Dib C, Faure S, Fizames C, Samson D, Drouot N, Vignal A, Millasseau P, Marc S, Hazan J, Seboun E, et al. A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature. 1996;380:152–154. doi: 10.1038/380152a0. [DOI] [PubMed] [Google Scholar]
  11. Dodson MK, Hartmann LC, Cliby WA, DeLacey KA, Keeney GL, Ritland SR, Su JQ, Podratz KC, Jenkins RB. Comparison of loss of heterozygosity patterns in invasive low-grade and high-grade epithelial ovarian carcinomas. Cancer Res. 1993;53:4456–4460. [PubMed] [Google Scholar]
  12. Ehlen T, Dubeau L. Loss of heterozygosity on chromosomal segments 3p, 6q and 11p in human ovarian carcinomas. Oncogene. 1990;5:219–223. [PubMed] [Google Scholar]
  13. Evron E, Cairns P, Halachmi N, Ahrendt SA, Reed AL, Sidransky D. Normal polymorphism in the incomplete trinucleotide repeat of the arginine-rich protein gene. Cancer Res. 1997;57:2888–2889. [PubMed] [Google Scholar]
  14. Fullwood P, Marchini S, Rader JS, Martinez A, Macartney D, Broggini M, Morelli C, Barbanti-Brodano G, Maher ER, Latif F. Detailed genetic and physical mapping of tumor suppressor loci on chromosome 3p in ovarian cancer. Cancer Res. 1999;59:4662–4667. [PubMed] [Google Scholar]
  15. Henriksen R, Gobl A, Wilander E, Oberg K, Miyazono K, Funa K. Expression and prognostic significance of TGF-β isotypes, latent TGF-β 1 binding protein, TGF β type I and type II receptors, and endoglin in normal ovary and ovarian neoplasms. Lab Invest. 1995;73:213–220. [PubMed] [Google Scholar]
  16. Hu W, Wu W, Nash MA, Freedman RS, Kavanagh JJ, Verschraegen CF. Anomalies of the TGF-β postreceptor signaling pathway in ovarian cancer cell lines. Anticancer Res. 2000;20:729–733. [PubMed] [Google Scholar]
  17. Ito I, Minegishi T, Fukuda J, Shinozaki H, Auersperg N, Leung PC. Presence of activin signal transduction in normal ovarian cells and epithelial ovarian carcinoma. Br J Cancer. 2000;82:1415–1420. doi: 10.1054/bjoc.1999.1127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kruk PA, Maines-bandiera SL, Auersperg N. A simplified method to culture human ovarian surface epithelium. Lab Invest. 1990;63:132–136. [PubMed] [Google Scholar]
  19. Lebeda RA, Haun RS. Cloning and characterization of the human ADP-ribosylation factor 4 gene. Gene. 1999;237:209–214. doi: 10.1016/s0378-1119(99)00290-5. [DOI] [PubMed] [Google Scholar]
  20. Lounis H, Mes-Masson A-M, Dion F, Bradley WE, Seymour RJ, Provencher D, Tonin PN. Mapping of chromosome 3p deletions in human epithelial ovarian tumors. Oncogene. 1998;17:2359–2365. doi: 10.1038/sj.onc.1202152. [DOI] [PubMed] [Google Scholar]
  21. Lyons J, Landis CA, Harsh G, Vallar L, Grunewald K, Feichtinger H, Duh Q-Y, Clark OH, Kawasaki E, Bourne HR, et al. Two G protein oncogenes in human endocrine tumors. Science. 1990;249:655–659. doi: 10.1126/science.2116665. [DOI] [PubMed] [Google Scholar]
  22. Manderson EN, Mes-Masson A-M, Provencher D, Tonin PN. Mutations in a 10- bp polyadenine repeat of transforming growth factor β-receptor type II gene is an infrequent event in human epithelial ovarian cancer. Clin Genet. 2000;57:151–153. doi: 10.1034/j.1399-0004.2000.570211.x. [DOI] [PubMed] [Google Scholar]
  23. Mertens F, Johansson B, Hoglund M, Mitelman F. Chromosomal imbalance maps of malignant solid tumors: A cytogenetic survey of 3185 neoplasms. Cancer Res. 1997;57:2765–2780. [PubMed] [Google Scholar]
  24. National Cancer Institute of Canada: Canadian Cancer Statistics 2000. Toronto, Canada, 2000.
  25. Provencher DM, Lounis H, Champoux L, Tetrault M, Manderson EN, Wang JC, Eydoux P, Savoie R, Tonin PN, Mes-Masson A-M. Characterization of four novel epithelial ovarian cancer cell lines. In Vitro Cell Dev Biol Anim. 2000;36:357–361. doi: 10.1290/1071-2690(2000)036<0357:COFNEO>2.0.CO;2. [DOI] [PubMed] [Google Scholar]
  26. Rimessi P, Gualandi F, Morelli C, Trabanelli C, Wu Q, Possati L, Montesi M, Barrett JC, Barbanti-Brodano G. Transfer of human chromosome 3 to an ovarian carcinoma cell line identifies three regions on 3p involved in ovarian cancer. Oncogene. 1994;9:3467–3474. [PubMed] [Google Scholar]
  27. Sagae S, Kobayashi K, Nishioka Y, Sugimura M, Ishioka S, Nagata M, Terasawa K, Tokino T, Kudo R. Mutational analysis of β-catenin gene in Japanese ovarian carcinomas: Frequent mutations in endometrioid carcinomas. Jpn J Cancer Res. 1999;90:510–515. doi: 10.1111/j.1349-7006.1999.tb00777.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Samimi G, Fink D, Varki NM, Husain A, Hoskins WJ, Alberts DS, Howell SB. Analysis of MLH1 and MSH2 expression in ovarian cancer before and after platinum drug-based chemotherapy. Clin Cancer Res. 2000;6:1415–1421. [PubMed] [Google Scholar]
  29. Shayesteh L, Lu Y, Kuo W-L, Baldocchi R, Godfrey T, Collins C, Pinkel D, Powell B, Mills GB, Gray JW. PIK3CA is implicated as an oncogene in ovarian cancer. Nat Genet. 1999;21:99–102. doi: 10.1038/5042. [DOI] [PubMed] [Google Scholar]
  30. Shridhar V, Rivard S, Wang X, Shridhar R, Paisley C, Mullins C, Beirnat L, Dugan M, Sarkar F, Miller OJ, et al. Mutations in the arginine-rich protein gene (ARP) in pancreatic cancer. Oncogene. 1997;14:2213–2216. doi: 10.1038/sj.onc.1201054. [DOI] [PubMed] [Google Scholar]
  31. Sonoda G, Palazzo J, du Manoir S, Godwin AK, Feder M, Yakushiji M, Testa JR. Comparative genomic hybridization detects frequent over representation of chromosomal material from 3q26, 8q24, and 20q13 in human ovarian carcinomas. Genes Chromosomes Cancer. 1997;20:320–328. [PubMed] [Google Scholar]
  32. Sugita M, Tanaka N, Davidson S, Sekiya S, Varella-Garcia M, West J, Drabkin HA, Gemmill RM. Molecular definition of a small amplification domain within 3q26 in tumors of cervix, ovary, and lung. Cancer Genet Cytogenet. 2000;117:9–18. doi: 10.1016/s0165-4608(99)00135-1. [DOI] [PubMed] [Google Scholar]
  33. Tanaka H, Shimada Y, Harada H, Shinoda M, Hatooka S, Imamura M, Ishizaki K. Polymorphic variation of the ARP gene on 3p21 in Japanese esophageal cancer patients. Oncol Rep. 2000;7:591–593. doi: 10.3892/or.7.3.591. [DOI] [PubMed] [Google Scholar]
  34. Tonin PN, Hudson TJ, Rodier F, Bossolasco M, Lee PD, Novak J, Manderson EN, Provencher D, Mes-Masson A-M. Microarray analysis of gene expression mirrors the biology of an ovarian cancer model. Oncogene. 2001;20:6617–6626. doi: 10.1038/sj.onc.1204804. [DOI] [PubMed] [Google Scholar]
  35. Vazquez J, Gonzalez L, Merino A, Vizoso F. Expression and clinical significance of apolipoprotein D in epithelial ovarian carcinomas. Gynecol Oncol. 2000;76:340–347. doi: 10.1006/gyno.1999.5678. [DOI] [PubMed] [Google Scholar]
  36. Wang XC, Katso R, Butler R, Hanby AM, Poulsom R, Jones T, Sheer D, Ganesan TS. H-RYK, an unusual receptor kinase: Isolation and analysis of expression in ovarian cancer. Mol Med. 1996;2:189–203. [PMC free article] [PubMed] [Google Scholar]
  37. Withoff S, van der Zee AGJ, de Jong S, Hollema H, Smit EF, Mulder NH, de Vries EGE. DNA topoisomerase II α and -β expression in human ovarian cancer. Br J Cancer. 1999;79:748–753. doi: 10.1038/sj.bjc.6690120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yang-Feng TL, Li S, Han H, Schwartz PE. Frequent loss of heterozygosity on chromosomes Xp and 13q in human ovarian cancer. Int J Cancer. 1992;52:575–580. doi: 10.1002/ijc.2910520414. [DOI] [PubMed] [Google Scholar]
  39. Zheng W, Luo MP, Welt C, Lambert-Messerlian G, Sung CJ, Zhang Z, Ying SY, Schneyer AL, Lauchlan SC, Felix JC. Imbalanced expression of inhibin and activin subunits in primary epithelial ovarian cancer. Gynecol Oncol. 1998;69:23–31. doi: 10.1006/gyno.1998.4958. [DOI] [PubMed] [Google Scholar]

Articles from Genome Research are provided here courtesy of Cold Spring Harbor Laboratory Press

RESOURCES