Clinical Chemistry Link to Randox Laboratories Web Site
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Clinical Chemistry 51: 27-34, 2005. First published October 28, 2004; 10.1373/clinchem.2004.038620
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Data Supplement
Right arrow All Versions of this Article:
clinchem.2004.038620v1
51/1/27    most recent
Right arrow Submit an electronic Letter to
the Editor about this paper
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sgarlato, G. D.
Right arrow Articles by Sussman, H. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sgarlato, G. D.
Right arrow Articles by Sussman, H. H.
Related Collections
Right arrow Molecular Diagnostics and Genetics
(Clinical Chemistry. 2005;51:27-34.)
© 2005 American Association for Clinical Chemistry, Inc.

Molecular Diagnostics and Genetics

Panel of Genes Transcriptionally Up-regulated in Squamous Cell Carcinoma of the Cervix Identified by Representational Difference Analysis, Confirmed by Macroarray, and Validated by Real-Time Quantitative Reverse Transcription-PCR

Gregory D. Sgarlato1,1, Catharine L. Eastman1,1 and Howard H. Sussmana,1

1 Department of Pathology, Stanford University, 300 Pasteur Dr., Stanford, CA 94305.

aAuthor for correspondence. Fax 650-725-6902; e-mail hsussman{at}stanford.edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: The Pap smear is currently the most widely used method of screening for squamous cell carcinoma of the cervix (SCCC). Because it is based on cell morphology, it is subject to variability in interpretation. Sensitive molecular markers capable of differentiating cancerous samples from noncancerous ones would be beneficial in this regard.

Methods: We performed representational difference analysis (RDA) using paired, noncancerous (normal) and cancerous (disease) tissues taken from the same specimen obtained from a single patient with a confirmed diagnosis of SCCC. Linearly amplified cDNA from normal and diseased tissues of the original patient and seven others were hybridized to DNA macroarrays containing the candidate gene transcript fragments. Real-time quantitative reverse transcription-PCR was used to validate the macroarray results.

Results: RDA identified a candidate pool of 65 transcript fragments up-regulated in diseased tissue compared with normal tissue. Forty-one transcripts were found to be up-regulated in diseased compared with normal tissue in at least one half the patients by macroarray hybridization. Eleven of those genes were selected for real-time quantitative reverse transcription-PCR analysis, and all were confirmed as transcriptionally up-regulated in cancer compared with normal tissue in at least one half the patients.

Conclusions: RDA using tissues from a single patient identified gene fragments confirmed to be transcriptionally up-regulated in SCCC both in the original patient and in seven others. The confirmed genes have a variety of functions and also have the potential to serve as diagnostic or prognostic markers.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Squamous cell carcinoma of the cervix (SCCC)2 is the second most common malignancy in women worldwide (1). SCCC is by far the most common histologic type of cervical cancer. The Pap test, based on cytologic examination of vaginal exfoliated cells, has reduced the incidence and mortality of cervical cancer by 60–70% in areas where it has been used in routine screening programs. However, where no Pap screening programs are in place or where a population does not participate in screening programs, the incidence and mortality of the disease remain high (1). A limitation of the Pap test is that it is morphologically based, and the accuracy can be problematic because of preanalytical processing and interpretive errors. There is interobserver variation in the reading and classifying of the cytologic smears. Molecular-based testing for high-risk human papillomavirus (HPV) strains is mostly performed when Pap tests are inconclusive and is generally used in conjunction with liquid-based cytologic methods. These tests are still being investigated in large studies to further determine their usefulness (2).

Current guidelines for managing patients with atypical squamous cells call for assigning these cases to Pap subcategories that distinguish the cases that have a high risk for invasive carcinoma (ASC-H; HSIL) from the cases of undetermined significance (ASC-US) (3). A molecular test based on multiple diagnostic markers that are associated with the cancer phenotype potentially could identify SCCC with higher specificity than currently available tests. Furthermore, the identification of a subset of the genes expressed in SCCC could be helpful in subcategory assignment.

The aim of this study was to identify genes that are transcriptionally up-regulated in SCCC. The identification of these genes may provide insight into the understanding of the biology of SCCC, and the genes identified have potential use in diagnosis. In contrast to recently published microarray-based studies on SCCC (4)(5), the research presented here uses representational difference analysis (RDA) to isolate a relatively small candidate pool of transcripts up-regulated in diseased vs nondiseased tissue in a single patient. RDA has been used to identify potentially up-regulated transcripts in other cancers (6). The selected pool of candidate transcripts is then screened by comparative hybridization on DNA macroarrays with amplified cDNA from the original patient from which they were derived and seven other patients. Real-time quantitative reverse transcription-PCR (RT-PCR) is firmly established as a highly sensitive gene-specific method for determining concentrations of selected gene transcripts (7) and was used here to confirm the transcriptional up-regulation of several of the genes identified by the RDA procedure across multiple patients.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
patient specimens
Tissue specimens were obtained from ILSBio or Genomics Collaborative. All patient samples were obtained in Vietnam and collected with patient consent in compliance with the company Institutional Review Boards and with the Code of Federal Regulations 45CFR46.101B. All specimens were anonymized by ILSBio and Genomics Collaborative. Paired SCCC (disease) and noncancer (normal) tissues were taken from single-patient surgical specimens that had been frozen in liquid nitrogen within 30 min of extirpation. Microscope slides were reviewed by a pathologist for diagnosis and staging, and a pathology report was received with each tissue specimen.

rna isolation
Frozen tissue samples (150 mg) were ground to a fine powder under liquid nitrogen. Total RNA was isolated from the powder by use of an RNEasy Midi Kit (Qiagen). Samples were treated with DNase during the column purification procedure. Total RNA samples were analyzed with an Agilent 2100 Bioanalyzer system for 18S and 28S band integrity, quantified by absorbance at 280 nm (A280), and checked for purity by the A260/A280 ratio.

POLYA rna isolation
Messenger RNA was isolated from total RNA of three patients (patients 1, 3, and 4) with magnetic bead isolation reagents (Boehringer-Mannheim) essentially according to manufacturer’s instructions. The bound mRNA was washed extensively with high-salt buffer and eluted with water. The purity and quantity of mRNA were estimated by A260/A280 readings.

CDNA synthesis
cDNA was synthesized according to one of two methods. The first method essentially followed that outlined by Gubler (8), using 2 µmol/L dT18-NOT-B primer [5'biotin-CACACACACACACAGGGCCGCT( (18))-3'] with polyA mRNA from normal and disease tissues from patients 1, 3, and 4. In the second method, 5 µg of total RNA from normal and disease tissues of patients 2, 3, 4, 5, 6, 7, and 8 was used as template in the Roche cDNA Synthesis System according to the manufacturer’s instructions and with use of the dT18-NOT-B primer.

rda subtraction
RDA protocols were carried out as described by Hubank and Schatz (9) with cDNA from normal and disease tissues of patient 1. Normal and disease amplicons were subsequently used to generate melt depletion normal and melt depletion disease amplicons. Subtraction-hybridization reactions were performed with reduced amounts of amplicon (10). Hybridization reactions used normal, disease, melt depletion normal, or melt depletion disease amplicon as driver. Two rounds of subtraction were performed with tester/driver ratios of 1:80 and 1:400. RDA products from the second round of hybridization from reactions using normal or melt depletion normal driver conditions were shotgun-cloned into the BamHI site of pBluescript II KS+. Three groups of 96 clones were selected for analysis. Plasmid DNA was purified with Qiagen plasmid miniprep columns according to the manufacturer’s instructions and sequenced with use of Big Dye PCR (Applied Biosystems) reactions with either T3 22mer (5'-GAAATTAACCCTCACTAAAGGG-3') or T7 22mer (5'-GTAATACGACTCACTATAGGGC-3'). Plasmids were confirmed by PCR amplification of the T3-T7 region of pBluescript II KS+ with T3 22mer and T7 22mer primers using standard protocols.

amplicon probe synthesis
We digested 10–50% of each cDNA reaction (normal and disease from all patients) with DpnII and ligated the digest to an excess of R-Bgl-12/24 linker (9). The resulting linked cDNA was amplified essentially as described in the RDA amplicon generation protocol (9) with the following modifications. All amplifications contained 5 U of Taq polymerase/100-µL reaction and 100 pM of R-Bgl-24 primer. The number of amplification cycles was determined based on the amount of cDNA used as template (18 cycles for 1 µL of 6 mg/L target). The cDNA concentrations were estimated based on the amount of total RNA used for cDNA synthesis, with use of values of 2% polyA mRNA and 100% cDNA synthesis efficiency. Eight reactions each were performed with normal and disease cDNA from each patient as template. Normal and disease amplicons were pooled separately, extracted with phenol–chloroform, precipitated with ethanol, resuspended, quantified by A260, and checked for purity by the A260/A280 ratio.

Purified amplicons were biotin-labeled to high specificity with use of BioPrime labeling reagents (Invitrogen). The manufacturer’s instructions were followed with the following exceptions: 1 µg of template was used, and the label reactions were incubated for 90–120 min. The biotinylated product was purified away from free deoxynucleotide triphosphates and primers with Chromaspin TE-100 size exclusion columns (BDBioscience/Clontech) pre-equilibrated with 2x standard saline citrate (SSC) containing 1 g/L sodium dodecyl sulfate (SDS). The mean yield for probe synthesis was 10–12 µg per reaction as determined by biotin quantification using probe biotinylation reagents (KPL).

dna macroarray synthesis and hybridization
Paired DNA macroarrays with identical spot patterns were prepared with a Bio-Blot apparatus (Bio-Rad) according to the manufacturer’s instructions. We denatured 1600 ng of each DNA sample, including the MCS region of pBluescript II KS+, in a solution of 0.4 mol/L NaOH in 2x SSC in a total volume of 440 µL and applied 110 µL to paired spots on duplicate positively charged nylon membranes (Sigma-Aldrich). For all of the 65 gene fragments to be analyzed by macroarray in duplicate, it was necessary to use two separate pairs of membranes. Macroarrays were cross-linked in a UV Stratalinker (Stratagene) at a setting of 1200 x 100 µJ.

We normalized macroarray hybridization experiments by adding equal masses of normal and disease probe, as determined by biotin quantification, to the hybridization reactions. Macroarrays were prehybridized in a roller bottle oven at 50 °C in 8 mL of 330 mL/L formamide in 2x SSC plus 200 µg/L sheared salmon sperm DNA and 1.25 mg/L DpnII-digested PCR product of the pBluescript II KS+ MCS for at least 1 h. We denatured 1–4 µg of normal or disease probe and added it to the reaction. Hybridizations were performed at 50 °C for >40 h. The macroarrays were subjected to stringent wash conditions (three 30-min washes with 2x SSC containing 1 g/L SDS at 50 °C; one 30-min wash with 0.2x SSC containing 1 g/L SDS at 45 °C; and one 1-h wash in 2x SSC at room temperature) and developed with DNA Detector HRPO reagents (KPL) essentially according to the manufacturer’s instructions. Wash times were increased to 10–15 min, and the KPL chemiluminescent substrate was replaced with SuperSignal West Dura Extended substrate (Pierce). Luminescence was captured with Bio-Max film (Kodak). Exposure times ranged from 1 s to 20 min.

semiquantitative dna macroarray analysis
Films were scanned as 300-dpi TIFF files on a Perfection 1250 flatbed printer (Epson America). Images were analyzed for integrated absorbance (intensity) in dot-blot GelPro 3.1 (Media Cybernetics) with the dot-blot analysis tools. Dot diameter was set at 90, and background close to the dot was subtracted. Mean intensities for pairs of dots were recorded for normal and disease for each exposure. We calculated semiquantitative fold expression values for each gene by dividing mean disease intensity by mean normal intensity at each exposure. Final values were chosen as those farthest from 1, obtained preferably from exposures at which both normal and disease intensities were above background.

real-time quantitative rt-pcr
We selected 11 genes for real-time quantitative RT-PCR analysis using the TaqMan system (ABI). Gene fragments were selected as candidates for analysis if DNA macroarray analysis indicated transcriptional up-regulation in at least four of the eight patients. External primers and dual-labeled 6-carboxyfluorescein/6-carboxytetramethylrhodamine internal probes were designed using Primer Express software (ABI) based on the sequences of the gene fragments isolated by the RDA procedure. The sequences of the primers and probes used are listed in Table 1 of the Data Supplement that accompanies the online version of this article athttp://www.clinchem.org/content/vol51/issue1/.


View this table:
[in this window]
[in a new window]
  		Table 1. Frequency of transcriptional up-regulation for 11 genes analyzed by DNA macroarray.

Template consisted of a 1:10 dilution of double-stranded cDNA in tRNA buffer [10 mM Tris (pH 8.0), 5 mg/L purified yeast tRNA], except for patient 1, for whom amplicon was used. All patient cDNA samples (normal and disease) were normalized based on equal target input of total RNA into cDNA reactions (5 µg). Patient 1 normal and disease amplicons were normalized by concentration calculated based on A260 absorbance, before dilution to 0.2 mg/L with tRNA buffer. Individual amplifications were performed in duplicate 30-µL reactions containing 90 nM external primers, 25 nM reporter probe, and 1.5 µL of template. Gene-specific quantitative calibrators consisted of purified PCR products of the individual gene fragments isolated by the RDA protocols. PCR products were purified and diluted in tRNA buffer to establish a dilution series of 2 x 107, 2 x 106, and 2 x 105 copies/µL for each gene fragment assay. Calibrators were tested for uniform differences between threshold values of 1:10 dilutions before use. Each assay tested two genes in eight patients. Each gene fragment was analyzed at least twice.

Calculated copy numbers based on each gene-specific quantitative calibrator were exported into Excel (Microsoft) for statistical analysis. Copy numbers for each patient sample (normal and disease) were averaged and analyzed. Data always consisted of at least three values per sample, and each dataset was confirmed to have a CV <40%. Ratios of transcriptional up-regulation (disease/normal) were calculated from the mean copy numbers for each patient for each gene.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
rda
RDA was performed with disease and normal tissues from a single patient (patient 1) diagnosed with nonkeratinizing SCCC. We picked 288 clones, which were found to contain fragments matching portions of 65 different genes. Sixty-two of these were human genes, 4 of which were novel transcripts. Eleven gene fragments are listed in Table 1Up ; all isolated fragments are listed in Table 2 of the online Data Supplement.


View this table:
[in this window]
[in a new window]
  		Table 2. Real-time quantitative PCR analysis of 11 selected genes.1

real-time quantitative rt-pcr validation of relative expression ratios in amplicon
We performed an experiment to determine whether the amplicons resulting from ~20 000-fold amplification of cDNA (e.g., 50 µg of amplicon synthesized from 2.5 ng of cDNA) maintained the same relative ratios of disease/normal expression as in the original cDNA. Double-stranded cDNA was synthesized from normal and disease samples from patient 6 with use of 6 µg of total RNA and a novel polyT18-based primer in the Roche cDNA Synthesis System, and normal and disease amplicons were synthesized as described in the Materials and Methods. Three gene fragments (CCNB1, ZWINT, and SPINT2) were each tested in parallel with actin by real-time quantitative RT-PCR using normal and disease cDNA and normal and disease amplicons as template. The mean CV for calculated copy number was 7.1% (range, 3.6–14%). Disease/normal ratios for the test genes were actin-corrected and compared with respect to template. For CCNB1, the cDNA ratio was 3.38, whereas the amplicon ratio was 4.63 (a 37.0% difference compared with the cDNA ratio). For ZWINT, the cDNA ratio was 3.12, whereas the amplicon ratio was 2.42 (22.4% difference). For SPINT2, the cDNA ratio was 2.96, whereas the amplicon ratio was 2.65 (10.5% difference).

dna macroarray analysis
Biotinylated normal and disease amplicon probes from patient 1 and seven additional cervical cancer patients were hybridized to arrays of PCR products representing the RDA fragments. An example of one visualized and analyzed macroarray is shown in Fig. 1 . Several pairs of dots are substantially darker in Fig. 1B (hybridized to disease amplicon) than the corresponding pairs in Fig. 1A (hybridized to normal amplicon). Additionally, several pairs of dots are easily visible in Fig. 1B that are not visible in Fig. 1A . The boxed pairs of dots indicate three gene fragments that are transcriptionally up-regulated in the patient shown here (patient 3; not all such genes are indicated). Mean intensity values for the pairs of dots in panels A and B and the calculated fold change in expression (disease/normal) are displayed in Fig. 1 of the online Data Supplement. It should be noted that Fig. 1 displays a single exposure and provides an example of how the analysis is performed, whereas the data presented in Table 1Up reflect data from multiple exposure times, generally ranging from 1 s to 20 min. The range of exposure times was required to capture the wide range of transcript concentrations seen among the genes isolated by the RDA procedure.



View larger version (23K):
[in this window]
[in a new window]
  		Figure 1. Example of DNA macroarray analysis.

(A and B), duplicate DNA macroarrays hybridized to biotin-labeled amplified normal and disease cDNA, respectively, from patient 3 and exposed to film for 20 min. Each pair of dots contains a single gene fragment. Boxes indicate three genes that are present at higher concentrations in disease than normal, indicating transcriptional up-regulation (not all such genes are indicated). Boxed genes, reading from top left, down, and then across, are KRT14, NDRG1, and NQO1. The intensities of the dots containing a fragment from NDRG1 (bottommost box) in A are 364 and 484 integrated absorbance (IOD) units, giving a mean value of 425 IOD units. The intensities of the same dots in B are 2125 and 2224 IOD units, giving a mean of 2175 IOD units. The disease/normal ratio for this gene at this exposure is calculated as 2175/425, or 5.1.

Selected results of the DNA macroarray analysis of the gene fragments in the eight patients examined are shown in Table 1Up . The 11 genes shown are all transcriptionally up-regulated in at least one half of all the patients. These genes plus an additional 30 that are also transcriptionally up-regulated in at least one half the patients are listed as group I in Table 2Up of the online Data Supplement. The remaining 24 genes are transcriptionally up-regulated in less than one half the patients as determined by DNA macroarray analysis (group II in Table 2Up of the online Data Supplement). It should be noted that many of the gene fragments listed in group II were not detected in all patients by DNA macroarray analysis. Such a lack of detection does not necessarily indicate that the gene fragments are not transcriptionally up-regulated in those patients, merely that the transcript concentrations were too low to be detected by this method.

real-time quantitative rt-pcr analysis of selected genes
To validate the expression array, we analyzed 11 of the genes that were indicated as transcriptionally up-regulated in at least one half the patients by DNA macroarray analysis by real-time quantitative RT-PCR. The genes chosen are those listed in Table 1Up . Table 2Up summarizes the results of the real-time quantitative RT-PCR analysis. All of the genes shown in Table 2Up were transcriptionally up-regulated by 1.8-fold or greater in at least four of the eight patients in disease vs normal tissue. Ten of those 11 genes were transcriptionally up-regulated by 1.8-fold or greater in at least six of the eight patients. For two patients, patients 7 and 8, real-time quantitative RT-PCR analysis showed transcriptional up-regulation in one half or fewer of the genes examined, whereas all other patients showed transcriptional up-regulation in at least two thirds of the genes examined. The CV for the replicate copy numbers for these genes ranged from 2.3% to 38%, with a mean value of 17%.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
This study was directed toward investigating the phenotype of SCCC by examining differences in expression between normal (noncancerous) and disease (cancerous) cervical tissue. Pursuant to this goal, we identified a panel of genes that are transcriptionally up-regulated in SCCC. A candidate group of 65 genes was identified by RDA using normal and disease tissues from a single patient. Amplicon probes were generated from normal and disease tissues in seven additional patients with SCCC and were used to confirm the transcriptional up-regulation of this diverse gene set. The small amount of cDNA needed to generate the amplicon probe (<10 ng) for each patient sample allowed the remaining cDNA to be used in confirmatory real-time quantitative RT-PCR experiments. Forty-one of the 65 genes identified by RDA were transcriptionally up-regulated in at least four of the eight patients as determined by comparative DNA macroarray hybridization analysis. Of the 11 genes examined by real-time quantitative RT-PCR, 10 were confirmed to be transcriptionally up-regulated in 75% of the patients and 1 gene, OAZ1, was confirmed to be transcriptionally up-regulated in 50% of the patients. The genes identified in this report may be useful in diagnostic applications.

RDA subtraction using normal and disease tissues from a single patient reduced the transcriptome complexity and allowed the isolation of key candidates with the screening of relatively few clones (288). Other studies using RDA to isolate genes of interest have used pooled samples from several patients (11) or used tissue culture samples (12). DNA macroarray analysis of the gene fragments isolated in the RDA protocols showed that more than two thirds of these gene fragments appear to be transcriptionally up-regulated in at least 50% of patients. This result demonstrates the power of RDA to isolate a small number of genes of interest. This power is further demonstrated by the identification of four transcripts that were previously unknown and an additional four that were not represented on human arrays manufactured by Affymetrix at the time of the study. The transcript concentrations of most of the genes in this group were too low to be detected by DNA macroarray. Relative transcript concentrations of these genes could be determined by the sensitive methodology of real-time quantitative RT-PCR analysis.

Normal and disease amplicons from patient 6, which were used to generate biotinylated probe for hybridization experiments, were directly compared with the original normal and disease cDNA by real-time quantitative RT-PCR. The results showed that these amplicons had fold expression ratios (disease/normal) similar to those for the cDNAs. The mean percentage difference between the ratios was 23.3%, which is very small considering the high degree of amplification (~20 000-fold) and the large increase in testable material. The validated amplicons used here have potential for use in array hybridization and other expression analyses and diagnostic platforms, particularly in cases where the original source material is limiting.

Comparative hybridization of DNA macroarrays is identical in concept to comparative microarray hybridization and carries similar potentials and dangers as outlined by Russo et al. (13). Macroarrays have a limited number of spots available on each blot and thus limit the number of replicates possible for each gene. The macroarrays in this study consisted of relatively long DNA sequences (120 bp or more) and so present opportunities for cross-hybridization. cDNA-based microarrays share this quality, but oligonucleotide-based microarrays do not. Macroarrays have some advantages over commercial microarrays. Macroarrays are inexpensive, straightforward to synthesize and use in a small laboratory, and can be stripped and reused several times. Macroarrays also allow the selective screening of a small number of genes, such as those isolated by RDA.

Eleven of the genes that were transcriptionally up-regulated by DNA macroarray analysis were analyzed by real-time quantitative RT-PCR and confirmed to be up-regulated in the cancerous specimen. These confirmatory results show that DNA macroarrays can be used in conjunction with RDA as a screening tool for identifying genes that are transcriptionally up-regulated.

As Fig. 2 shows, DNA macroarray analysis detected transcriptional up-regulation in 62 of 88 patient-gene data points (70.5%). Real-time quantitative RT-PCR detected transcriptional up-regulation in 67 of 88 data points (76.1%). Combined, the two methods detect transcriptional up-regulation in 79 of 88 data points (89.8%). The two methods agreed with each other in detecting transcriptional up-regulation in 59 of 88 data points (67.0%). Occasional disagreement between real-time quantitative RT-PCR and comparative hybridization results has been seen in other studies (14)(15). The increased sensitivity of real-time quantitative RT-PCR over DNA macroarray analysis accounts for some instances in which real-time quantitative RT-PCR indicated transcriptional up-regulation where the macroarray analysis did not. The increased specificity of real-time quantitative RT-PCR over DNA macroarray analysis likely accounts for some instances in which the DNA macroarray indicated transcriptional up-regulation where real-time quantitative RT-PCR did not. The primers and probes used in real-time quantitative RT-PCR are gene specific and thus analyze only one gene of a group of similar genes that may hybridize to a spot on the macroarray. Such cross-hybridization events on the macroarray may indicate transcriptional up-regulation of multiple genes in a family of genes with similar sequences. These similar genes could be examined individually using gene-specific primers and probes in real-time quantitative RT-PCR experiments.



View larger version (55K):
[in this window]
[in a new window]
  		Figure 2. Comparison of results from DNA macroarray and real-time quantitative RT-PCR experiments.

Filled circles, transcriptional up-regulation detected by both DNA macroarray (disease/normal ≥2.0) and real-time quantitative RT-PCR (disease/normal ≥1.8). Circles with filled top half, transcriptional up-regulation detected by DNA macroarray but not real-time quantitative RT-PCR. Circles with filled bottom half, transcriptional up-regulation detected by real-time quantitative RT-PCR but not DNA macroarray. Open circles, transcriptional up-regulation not detected by DNA macroarray or real-time quantitative RT-PCR. Failure to detect transcriptional up-regulation includes failure to detect transcript in normal and disease samples, equal amounts of transcript in normal and disease, and transcriptional down-regulation in disease vs normal. Genes are listed in order from those up-regulated in the greatest number of patients to least as determined by both DNA macroarray and real-time quantitative RT-PCR (filled circles). *, original RDA source.

Several genes in the set of 11 confirmed genes are known to be up-regulated or involved in other cancers: CCNB1 (cyclin B1) (16)(17), AURKB (aurora B kinase) (18), SPINT2 (serine protease inhibitor 2)(19), OAZ1 (ornithine decarboxylase antizyme 1) (20), and HPV16 E7 (the E7 protein of human papillomavirus 16) (21). (HPV16 E7 was detected in the disease specimens from six of the eight patients, including the original patient specimen used for RDA. No tests for other HPV genes were performed.) Two genes function in cell division: ZWINT (Zw10 interacting factor) may be involved in checkpoint function (22), and CDCA8 (cell division cycle associated protein 8) is coexpressed with other cell cycle genes, such as CDC2, CDC3, and cyclin (23). The other confirmed transcriptionally up-regulated genes may be associated with cervical disease: G1P2 (interferon-stimulated protein; 15-kDa) may be overproduced as a result of infection (24), and KRT14 and KRT16 (keratin 14 and 16) are produced at high concentrations in the keratinizing squamous epithelium of the cervix. Increased proliferation of tissue that naturally produces keratins is likely to produce increased concentrations of keratin; such an increase may be reflected at the transcript level.

A recent microarray study examining the transcriptional profiles of several stages of SCCC (4) independently identified two transcriptionally up-regulated genes that appear in this study: ARK2/AURKB, which is confirmed here to be transcriptionally up-regulated by real-time quantitative RT-PCR, and MYBL2 (v-myb-like 2), which appears to be transcriptionally up-regulated in four of eight patients by DNA macroarray analysis. Minichromosome maintenance protein 2 (MCM2), which is in the same functional family as two other genes (MCM4 and MCM6) identified in the study of Chen et al. (4), is also indicated as transcriptionally up-regulated by DNA macroarray analysis. No other genes isolated in this study appear either in the study of Chen et al. (4) or in a microarray study performed in Wong et al. (5). The genes identified in Table 2Up of the Data Supplement therefore add key elements to the picture of transcriptionally up-regulated genes in SCCC.

In this pilot study, many gene fragments were isolated that are indicated as transcriptionally up-regulated in both the single patient from which they were isolated and 75% or more of all patients examined by DNA macroarray analysis. The genes listed in Table 2Up of the online Data Supplement are worthy of further study. Although some of these genes, such as MCM2 (25), NDRG1(26), CBR1 (27), and EIF4A (28), have been identified as transcriptional markers of cancer, others, such as CALML5, have not been identified as having roles in SCCC or in other cancers.

RDA performed with normal and disease tissues from a single patient identified a panel of transcriptionally up-regulated genes that were confirmed by use of amplified cDNA from seven other patients. The genes of interest in the panel are those that have a high correlation of expression in multiple patients. The presence of genes that do not have a high correlation of expression indicates the variable expression that may be a function of differences in neoplastic transformation and/or the growth characteristics of SCCC. One could potentially increase the size of this gene panel by performing RDA on additional SCCC patients and confirming the expression of newly identified fragments of genes of interest in an expanded number of patients. Panels of genes shown to be transcriptionally up-regulated in SCCC, such as those presented in this study, could improve the understanding of this disease and provide the basis for a diagnostic test.


   Acknowledgments
 
Sequencing reactions were run at the Stanford PAN facility or by our collaborators at Beckman Coulter (Fullerton, CA). This work was supported by a grant from Beckman Coulter, Inc. (2HMZ901).


   Footnotes
 
1 These authors contributed equally to this work.

2 Nonstandard abbreviations: SCCC, squamous cell carcinoma of the cervix; HPV, human papillomavirus; RDA, representational difference analysis; RT-PCR, reverse transcription-PCR; SSC, standard saline citrate; SDS, sodium dodecyl sulfate; and MCM, minichromosome maintenance protein.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Parkin DM. Global cancer statistics in the year 2000 [Review]. Lancet Oncol 2001;2:533-543.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  2. Bulkmans NW, Rozendaal L, Snijders PJ, Voorhorst FJ, Boeke AJ, Zandwijken GR, et al. POBASCAM, a population-based randomized controlled trial for implementation of high-risk HPV testing in cervical screening: design, methods and baseline data of 44,102 women. Int J Cancer 2004;110:94-101.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  3. Wright TC, Jr, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ. 2001 consensus guidelines for the management of women with cervical cytological abnormalities. JAMA 2002;287:2120-2129.[Abstract/Free Full Text]
  4. Chen Y, Miller C, Mosher R, Zhao X, Deeds J, Morrissey M, et al. Identification of cervical cancer markers by cDNA and tissue microarrays. Cancer Res 2003;63:1927-1935.[Abstract/Free Full Text]
  5. Wong YF, Selvanayagam ZE, Wei N, Porter J, Vittal R, Hu R, et al. Expression genomics of cervical cancer: molecular classification and prediction of radiotherapy response by DNA microarray. Clin Cancer Res 2003;9:5486-5492.[Abstract/Free Full Text]
  6. Frohme M, Scharm B, Delius H, Knecht R, Hoheisel JD. Use of representational difference analysis and cDNA arrays for transcriptional profiling of tumor tissue. Ann N Y Acad Sci 2000;910:85-104discussion 104–5.[ISI][Medline] [Order article via Infotrieve]
  7. Mocellin S, Rossi CR, Pilati P, Nitti D, Marincola FM. Quantitative real-time PCR: a powerful ally in cancer research [Review]. Trends Mol Med 2003;9:189-195.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  8. Gubler U. Second-strand cDNA synthesis: classical method. Methods Enzymol 1987;152:325-329.[ISI][Medline] [Order article via Infotrieve]
  9. Hubank M, Schatz DG. Identifying differences in mRNA expression by representational difference analysis of cDNA. Nucleic Acids Res 1994;22:5640-5648.[Abstract/Free Full Text]
  10. Welford SM, Gregg J, Chen E, Garrison D, Sorensen PH, Denny CT, et al. Detection of differentially expressed genes in primary tumor tissues using representational differences analysis coupled to microarray hybridization. Nucleic Acids Res 1998;26:3059-3065.[Abstract/Free Full Text]
  11. Gress TM, Wallrapp C, Frohme M, Muller-Pillasch F, Lacher U, Friess H, et al. Identification of genes with specific expression in pancreatic cancer by cDNA representational difference analysis. Genes Chromosomes Cancer 1997;19:97-103.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  12. Niwa H, Harrison LC, DeAizpurua HJ, Cram DS. Identification of pancreatic ß cell-related genes by representational difference analysis. Endocrinology 1997;138:1419-1426.[Abstract/Free Full Text]
  13. Russo G, Zegar C, Giordano A. Advantages and limitations of microarray technology in human cancer [Review]. Oncogene 2003;22:6497-6507.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  14. Rajeevan MS, Vernon SD, Taysavang N, Unger ER. Validation of array-based gene expression profiles by real-time (kinetic) RT-PCR. J Mol Diagn 2001;3:26-31.[Abstract/Free Full Text]
  15. Alevizos I, Mahadevappa M, Zhang X, Ohyama H, Kohno Y, Posner M, et al. Oral cancer in vivo gene expression profiling assisted by laser capture microdissection and microarray analysis. Oncogene 2001;20:6196-6204.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  16. Li JQ, Kubo A, Wu F, Usuki H, Fujita J, Bandoh S, et al. Cyclin B1, unlike cyclin G1, increases significantly during colorectal carcinogenesis and during later metastasis to lymph nodes. Int J Oncol 2003;22:1101-1110.[ISI][Medline] [Order article via Infotrieve]
  17. Barrett KL, Demiranda D, Katula KS. Cyclin b1 promoter activity and functional cdk1 complex formation in G1 phase of human breast cancer cells. Cell Biol Int 2002;26:19-28.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  18. Warner SL, Bearss DJ, Han H, Von Hoff DD. Targeting Aurora-2 kinase in cancer [Review]. Mol Cancer Ther 2003;2:589-595.[Abstract/Free Full Text]
  19. Tanaka Y, Kobayashi H, Suzuki M, Kanayama N, Yamakawa T, Morishita H, et al. Reduced bikunin gene expression as a factor of poor prognosis in ovarian carcinoma. Cancer 2003;98:424-430.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  20. Feith DJ, Shantz LM, Pegg AE. Targeted antizyme expression in the skin of transgenic mice reduces tumor promoter induction of ornithine decarboxylase and decreases sensitivity to chemical carcinogenesis. Cancer Res 2001;61:6073-6081.[Abstract/Free Full Text]
  21. Munger K, Howley PM. Human papillomavirus immortalization and transformation functions [Review]. Virus Res 2002;89:213-228.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  22. Starr DA, Saffery R, Li Z, Simpson AE, Choo KH, Yen TJ, et al. HZwint-1, a novel human kinetochore component that interacts with HZW10. J Cell Sci 2000;113(Pt 11):1939-1950.[Abstract]
  23. Walker MG. Drug target discovery by gene expression analysis: cell cycle genes. Curr Cancer Drug Targets 2001;1:73-83.[CrossRef][Medline] [Order article via Infotrieve]
  24. Meraro D, Gleit-Kielmanowicz M, Hauser H, Levi BZ. IFN-stimulated gene 15 is synergistically activated through interactions between the myelocyte/lymphocyte-specific transcription factors, PU.1, IFN regulatory factor-8/IFN consensus sequence binding protein, and IFN regulatory factor-4: characterization of a new subtype of IFN-stimulated response element. J Immunol 2002;168:6224-6231.[Abstract/Free Full Text]
  25. Gonzalez MA, Pinder SE, Callagy G, Vowler SL, Morris LS, Bird K, et al. Minichromosome maintenance protein 2 is a strong independent prognostic marker in breast cancer. J Clin Oncol 2003;21:4306-4313.[Abstract/Free Full Text]
  26. Mackay A, Jones C, Dexter T, Silva RL, Bulmer K, Jones A, et al. cDNA microarray analysis of genes associated with ERBB2 (HER2/neu) overexpression in human mammary luminal epithelial cells. Oncogene 2003;22:2680-2688.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  27. Umemoto M, Yokoyama Y, Sato S, Tsuchida S, Al-Mulla F, Saito Y. Carbonyl reductase as a significant predictor of survival and lymph node metastasis in epithelial ovarian cancer. Br J Cancer 2001;85:1032-1036.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  28. Ji P, Diederichs S, Wang W, Boing S, Metzger R, Schneider PM, et al. MALAT-1, a novel noncoding RNA, and thymosin ß4 predict metastasis and survival in early-stage non-small cell lung cancer. Oncogene 2003;22:8031-8041.[CrossRef][ISI][Medline] [Order article via Infotrieve]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Data Supplement
Right arrow All Versions of this Article:
clinchem.2004.038620v1
51/1/27    most recent
Right arrow Submit an electronic Letter to
the Editor about this paper
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sgarlato, G. D.
Right arrow Articles by Sussman, H. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sgarlato, G. D.
Right arrow Articles by Sussman, H. H.
Related Collections
Right arrow Molecular Diagnostics and Genetics

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS