previous phd theses: application of microarrayalbert szent-györgyi ph.d. library 1 liliána z....

Albert Szent-GyörGyi Ph.D. Library

Sze g e d, 2 0 1 1

LiLiána Z. Fehér msc

DNS-Chip Laboratorium

Biological Research Center

Hungarian Academy of Sciences

Application of microarray and quantitative real-time PCR methods in oncology

M269

Previous PhD Theses:

260. Zsolt Kopniczky MD:The ef fect of neurosurgical ablation of entorhinal cor tex

261. Attila Kovács MD: Polymorphism of HLA class II alleles and tumor necrosis factor alpha promoter alleles in Hungarian patients with systemic lupus er ythematosus and with primar y Sjögren's syndrome

262. Ádám Kuncz MD: The treatment of trigeminal neuralgia with microvascular decompression. The role of magnetic resonance angiography in the indication of surgical treatment in trigeminal

neuralgia

263. András Palotás MD: BiochemiCa2+l studies in Alzheimer ’s disease: the APParent diagnostic and therapeutic role of antipsychotics on calcium homeostasis and amyloid-precursor-protein

metabolism

264. István Pelsőczi-Kovács DDS: New Method for Modif ication of a Biomaterial Sur face to Improve its Osseointegration

265. Réka Szakács MD: Neuronal proto oncogene expression in the pharmacological assessment of the synaptic mechanisms of the hippocampus in rats

266. Kinga Szigeti MD: Charcot-Marie-Tooth disease and related peripheral neuropathies: the role of genetic testing

267. Zsolt Szigeti MD: Ef fects of preconditioning on myocardial regional contractility during low-f low ischaemia; the possible role of nitric oxide

268. Tünde Vezér MD: Experimental investigation of the behavioural toxicity of an environmental pollutant heavy metal, manganese

Albert Szent-Györgyi Ph.D. Library

DNS-Chip Laboratorium, Biological Research Center, Szeged

No. M 269 Title: Application of microarray and quantitative real-timePCR methods in oncology Ph.D. candidate: Liliána Fehér MSc Ph.D. supervisor: László Puskás MD, PhD Date of public defence: May 29, 2006 Ph.D. committee:

chairman: Imre Boros, MSc, PhD, DSc official reviewers: Judit Deák, MD, PhD

Ernő Duda, MSc, PhD, DSc. committee members: László Krenács, MD, PhD

László Tiszlavicz, MD, PhD Miklós Sántha MSc, PhD The history of the University of Szeged dates back to 1581 when István Báthory, the Prince of Transylvania founded a higher education institution in Kolozsvár (Cluj-Napoca) which became prestigious in a short time. Due to its professors well-known all around Europe it provided high standard education and also had the right to confer baccalaureate and master’s degrees. Moreover, it was the only institute for higher education in Hungary at the end of 16th century. Later Maria Theresia entrusted the Piarists to reorganize the institution as a result of which the Faculty of Medicine-Surgery was established in 1775. Later on, these served as the basis for the Hungarian Royal University of Kolozsvár, founded by Francis Joseph I in 1872. It was renamed after the king in 1881 and bore his name until 1940. The institution moved to Szeged in 1921.

Nowadays, the Faculty of Medicine, University of Szeged is one of the most outstanding medical schools in Hungary teaching health sciences in three languages. The Faculty has excellent scientific laboratories performing high standard researches supported by national and international grants. Students have a wide range of opportunities to join scientific research activities during the time of their studies. Experience gained during university years help many students to become successful researchers all around the world. The Faculty has four Ph.D. Doctoral Schools in which more than a hundred supervisors offer dissertation topic proposals. Notably, annually there are approximately 40 defended Ph.D. dissertations and graduations. It gained high reputation in research, education and practice of medical sciences. We are all proud of Albert Szent-Györgyi, former professor and dean of the Faculty who was awarded Nobel Prize in 1937 for his research in Szeged. He is an idol both for lecturers and students, presenting the idea that world-wide results can be achieved in Hungary and Szeged.

© Liliána Fehér, 2011

Responsible for edition: Dean of the Faculty of Medicine: Prof. László Vécsei MHAS Dean of the Faculty of Pharmacy: Prof. Ferenc Fülöp MHAS

ISSN 1417-0620

ME DI CI NA

Editors: Dr. Péter Hegyi and Dr. Gerda SzakonyiTechnical editor: Imre DócziCover page image: Factory Creative StudioSize: 7,5 (A/5) sheets

Faculty of Medicine

University of Szeged

Szeged, Hungary

Application of microarray and

quantitative real-time PCR

methods in oncology

Ph.D. Thesis

2006.

Liliána Z. Fehér

Biological Research Center

of the Hungaryan Academy of Sciences

Laboratory of Functional Genomics

CONTENT

I. PUBLICATIONS RELATED TO THE THESIS ............................................................................................ 1 II. AIMS .................................................................................................................................................................. 1 IV. ABSTRACT ..................................................................................................................................................... 3 V. INTRODUCTION ............................................................................................................................................. 6

5.1. Functional genomics in oncology ................................................................................................................ 6 5.2. Quantitative real-time PCR .......................................................................................................................... 8

5.2.1. Applications of QRT-PCR ................................................................................................................... 8 5.3. Whole genome amplification techniques ..................................................................................................... 9

5.3.1. Strand or multiple displacement amplification (SDA or MDA) and T7-based linear amplification .. 10 5.3.2. PCR-based genomic DNA amplification methods ............................................................................. 10 5.3.2.1. Degenerate oligonucleotide-primed PCR ........................................................................................ 11

5.4. DNA microarray techniques and comparative genome hybridization ....................................................... 13 5.4.1. DNA microarray technology .............................................................................................................. 13 5.4.2. Development of CGH method ............................................................................................................ 14 5.4.3. CGH combined with microarray technique ........................................................................................ 14

5.5. Molecular pathomechanism of different tumour types .............................................................................. 15 5.5.1. Merkel cell carcinoma ........................................................................................................................ 15 5.5.2. Classification of the thyroid carcinomas ............................................................................................ 16 5.5.2.1. Papillary and anaplastic thyroid carcinoma ..................................................................................... 16 5.5.2.2. Diagnostic molecular markers in papillary and anaplastic thyroid carcinomas ............................... 17 5.5.3. The effect of αIIbβ3 integrin on increased angiogenesis and tumour growth in melanoma ............... 18

VI. MATERIALS AND METHODS .................................................................................................................. 21 6.1. Tissue specimen and genomic DNA isolation ........................................................................................... 21

6.1.1. Samples for the improved DOP-PCR technique ................................................................................ 21 6.1.2. Samples for the Merkel cell carcinoma study ..................................................................................... 21 6.1.3. Papillary thyroid carcinoma samples .................................................................................................. 23

6.2. DNA and RNA quantity measurement ...................................................................................................... 24 6.3. Quantitative real-time PCR ........................................................................................................................ 24

6.3.1. Preparation of DOP-PCR amplified and overamplified genomic DNA ............................................. 24 6.3.2. Determination of the differences in relative copy numbers of the genome ........................................ 25

6.4. Preparation of microarray probes............................................................................................................... 26 6.4.1. Preparation of genomic DNA probes for array hybridization of MCC .............................................. 26 6.4.2. Preparation of genomic DNA probes for array hybridization of PTC samples .................................. 27 6.4.3. Generation of microarray probes for human melanoma cells ............................................................. 27

6.5. Microarray construction and array hybridization ....................................................................................... 27 6.5.1. Human cDNA microarray construction .............................................................................................. 27 6.5.2. CGH-array hybridization in the case of MCC .................................................................................... 27

6.5.3. Array hybridization in the case of thyroid tumours and human melanoma cells .................................... 28 6.6. Array data analysis .................................................................................................................................... 28

VII. RESULTS AND DISCUSSION .................................................................................................................. 29 7.1. Improved DOP-PCR-based representational whole genome amplification using QRT-PCR ................... 29 7.2. A second field metachronous MCC of the lip and the palatine tonsil ........................................................ 33 7.3. Determination of several copy number changes in different genomic DNA of thyroid carcinoma based on disease course ................................................................................................................................................... 38 7.4. QRT-PCR-based exponential amplification for microarray gene expression profiling ............................. 42 7.5. Analysis of the gene expression pattern of parallel expression of αIIbβ3 and αvβ3 integrins in human melanoma cells using microarray technology ................................................................................................... 43

VIII. CONCLUSIONS ......................................................................................................................................... 44 IX. ACKNOWLEDGMENTS ............................................................................................................................. 45 X. REFERENCES ................................................................................................................................................ 45

1

I. PUBLICATIONS RELATED TO THE THESIS

[1] Liliána Z. Fehér, Margit Balázs, János Z. Kelmen, István Németh, Zoltán Varga-Orvos,

László G. Puskás (2006) Improved DOP-PCR-based representational whole-genome

amplification using quantitative real-time PCR. Diagn. Mol. Pathol. 15:43-8.

[2] Judit Nagy, Liliána Z. Fehér, István Sonkodi, József Lesznyák, Béla Iványi, László G.

Puskás (2005) A second field metachronous Merkel cell carcinoma of the lip and the

palatine tonsil confirmed by microarray-based comparative genomic hybridization.

Virchows Arch. 446:278-86.

[3] Balázs Döme, Erzsébet Rásó, Judit Dobos, Livia Mészáros, Norbert Varga, László G.

Puskás, Liliána Z. Fehér, Tamár Lőrincz, Andrea Ladányi, Mohit Trikha, Kenneth V.

Honn, József Tímár (2005) Parallel expression of alphaIIbbeta3 and alphavbeta3 integrins

in human melanoma cells upregulates bFGF expression and promotes their angiogenic

phenotype. Int. J. Cancer. 116:27-35.

[4] Zsolt B. Nagy, János Z. Kelemen, Liliána Z. Fehér, Ágnes Zvara, Kata Juhász, László G.

Puskás (2005) Real-time polymerase chain reaction-based exponential sample

amplification for microarray gene expression profiling. Anal. Biochem. 337:76-83.

II. AIMS

[1] Development of an improved degenerate oligonucleotide-primed (DOP) PCR-based representational whole-genome amplification method using quantitative real-time PCR (QRT-PCR).

[2] Identification of an uncommon metastasis of Merkel cell carcinoma (MCC) cell carcinoma with the improved whole genome amplification and comparative genome hybridization (CGH) technique with DNA microarrays.

[3] Determining changes in chromosome copy numbers of papillary thyroid carcinomas (PTC) of good disease outcome and an aggressive type of PTC derived from the same patient using CGH-microarray and QRT-PCR techniques.

[4] Identification of gene markers based on copy number changes for classification PTCs of different disesase courses.

[5] Determining the angiogenic phenotype of αIIbβ3 integrin-transduced human melanoma cells expressing integrin αvβ3 using DNA-microarray technology to reveal the underlying pathomechanism.

2

III. ABBREVIATIONS AKAP13: A-kinase anchor protein 13 Arf6: ADP-ribosylation factor 6 APTC: An aggressive type of papillary thyroid carcinoma ATC: Anaplastic thyroid cancer BAC: Bacterial artificial chromosome Bcl-2: B-cell CLL/lymphoma 2 bFGF: Basic fibroblast growth factor BRAF: v-raf murine sarcoma viral oncogene homolog B1 CGH: Comparative genome hybridization CHO: Chinese hamster ovary COX-2: Prostaglandin-endoperoxide synthase 2 Dbl: MCF.2 cell line derived transforming sequence DGGE: Denaturing gradient gel electrophoresis DOP-PCR: Degenerate oligonucleotide-primed PCR EIF4EBP3: Eukaryotic initiation factor 4E-binding protein 3 FGF7: Fibroblast growth factor 7 FISH: Fluorescent in situ hybridization FTC: Follicular thyroid carcinoma GAP: GTPase-activating protein GEF: Guanine nucleotide exchange factor GPTC: Papillary thyroid carcinoma of good disease outcome HRAS: v-Ha-ras Harvey rat sarcoma viral oncogene homolog IRS-PCR: Infrequent-restriction-site polymerase chain reaction IPTC: Intermediate papillary thyroid carcinoma Ki-67: Antigen identified by monoclonal antibody Ki-67 KRAS: v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog LA-PCR: Linker adapter polymerase chain reaction Lbc: Lymphoid blast crisis oncogene LMP: Ligation mediated polymerase chain reaction LOH: Loss of heterozygosity MAPK: Mitogen-activated protein kinase MAPKK: Mitogen-activated protein kinase kinase MCC: Merkel cell carcinoma MDA: Multiple displacement amplification NRAS: Neuroblastoma RAS viral (v-ras) oncogene homolog PAC: P1-derived artificial chromosome PEP: Primer extension preamplification PIK3CA: Phosphatydilinositol 3-kinase, catalytic, alpha polypeptide PI3K: Phosphatidylinositol 3-kinase PTC: Papillary thyroid cancer PTEN: Phosphatase and tensin homolog QRT-PCR: Quantitative real-time PCR RET: REarranged during Transfection proto-oncogene RT: Reverse transcription RT-PCR: Reverse transcription-polymerase chain reaction SDA: Strand displacement amplification SKY: Spectral karyotyping SNP: Single-nucleotide polymorphism SSCP: Single-stranded confirmational polymorphism TB10: Thymosin beta 10 TLAD: T7-based linear amplification TNM: Tumour, node, metastasis TPO: Thyroid peroxidase Tre-2: Ubiquitin specific peptidase 6 TSH: Thyroid stimulating hormone VEGF-C: Vascular endothelial growth factor C vWF: von Willebrand factor

3

IV. ABSTRACT

The technological advances in molecular biology have proven invaluable to the

understanding of the pathogenesis of human cancer. The application of molecular technology

to the study of cancer has not only led to advances in tumour diagnosis, but has also provided

markers for the assessment of prognosis and disease progression. The amount of genomic

DNA from archived clinical samples available for genetic studies can often be limited.

Diverse whole genome amplification methods are applied to provide a sufficient amount of

DNA for CGH, single nucleotide polymorphism (SNP) and microsatellite analyses. CGH

allows identification of changes in relative copy number of DNA sequences (gains and losses)

and provides information on the primary genetic changes. Currently the microarray-based

CGH serves as an excellent method of display. In this thesis an improved DOP-PCR

technique, and its application in the case of two clinical studies are described. The effect of

αIIbβ3 integrin transfection onto gene expression pattern of human melanoma cells was also

analysed.

In genomic studies the reliability of the amplification techniques is particularly

important. In PCR-based approaches, the plateau-effect can seriously alter the original relative

copy number of certain chromosomal regions. To eliminate this distorting effect, we

improved the standard DOP-PCR technique by following the amplification status with QRT-

PCR. With real-time detection of the products, we managed to eliminate DNA over-

amplification. Probes were generated from 10 different tumour samples by using non-

amplified and amplified genomic DNA with DOP-PCR and DOP-PCR combined with QRT-

PCR. To demonstrate the reliability of the QRT-PCR based amplification protocol, altogether

152 relative copy number changes of 44 regions were determined. There was a 85.6%

concordance in copy number alterations between the QRT-PCR protocol and the non-

amplified samples, while this value was only 63.8% for the traditional DOP-PCR. Our results

demonstrate that this protocol preserves the original copy number of different chromosomal

regions in amplified genomic DNA – more accurately than standard DOP-PCR techniques.

This improved genomic amplification technique was also applied in two clinical cases.

In the first case, MCC was diagnosed in a 79-year-old Caucasian woman. This

malignant, neuroendocrin tumour was localized to the upper lip. After successful cryosurgery

and a 7-year tumour-free period, a new tumour developed in her palatine tonsil. The regional

4

lymph nodes were devoid of metastasis. After a long tumour-free period, an anaplastic

carcinoma with neuroendocrine features developed in her palatine tonsil, raising the

possibility of a late haematogenous metastasis, a second field tumour, or a second primary

tumour. The genomic DNA obtained from the paraffin-embedded two tumours and unaffected

lymphatic tissue was amplified with our improved DOP-PCR protocol for CGH and DNA

microarray techniques to assess the genetic relationship of the tumours. The partly similar and

partly different molecular patterns indicated a genetic relationship between the tumours, and

excluded the possibility that the tonsillar tumour was a metastasis. Our results suggest that a

genetically altered field was the reason for the development of the tonsillar cancer; thus, it can

be pathogenetically regarded as a second field tumour.

The improved DOP-PCR technique was also employed for the detection of molecular

patterns of PTC of different disease courses. PTC is the most common type of thyroid

cancers, making up about 80% of all types. Unlike some other tumours, the generally

excellent outcome for PTC is usually not affected by spread of the cancer to the lymph nodes.

All the same in some cases it transforms into a dedifferentiated invasive tumour with an

unfavourable outcome. Our aim was to identify gene copy number alteration as a marker or

set of markers in order to predict individual patient outcome more effectively than existing

staging methods; for this reason, we performed CGH-array analysis of primary- and terminal-

stage APTCs, comparing them with PTC of good disease outcome (GPTC) obtained from

paraffin-embedded tissue. On the basis of CGH-array and confirming QRT-PCR results, six

regions were selected for further investigation of 34 PTCs of different disease courses. Two

groups were constructed according to the disease groups: ’A’ group: 10 cases of GPTCs; ’B’

group: 14 cases of intermediate (IPTC) and 10 cases of APTCs (I/APTC). In the present

study, Tre-2 oncogene was found to be over-represented in 80% of both the GPTC and the

I/APTC groups. Thymosin beta 10 (TB10) was amplified in 62.5% of APTC, whereas in the

case of the GPTC group, the ratio of DNA gains was only 20%. A similarly patterned

difference, but with lower percentage rates, could be detected in our examined Image EST

region, which is mapped to the AKAP13 gene. Two regions were under-represented in a

different proportion: eukaryotic initiation factor 4E-binding protein 3 (EIF4EBP3) and

KIAA0549 genes. The sixth examined gene, namely fibroblast growth factor 7 (FGF7),

demonstrated both DNA gains and losses with a smaller rate.

5

Copy number changes of the above mentioned six loci alone are still not sufficient for

classifying the tumours; however, on the basis of our results, the synchronous presence of the

amplified TB10 and Image EST (Acc. No. R78712) regions could serve a good diagnostic –

or possibly prognostic – marker for aggressive thyroid carcinoma.

In our other study we developed a cDNA amplification method within which –

similarly to the whole genomic DNA amplification – the course of amplification was

followed using QRT-PCR. With this approach low quantity RNA sample could also be

reliably analysed, which is especially important in assessing gene expression pattern analysis

for instance the molecular mechanisms in different tumours.

In cancerous tissues new vessels are required to provide cancer cells with nutrients and

are targets for invading cancer cells themselves. The transfection of the platelet integrin

αIIbβ3 into human melanoma cells expressing integrin αvβ3 promoted their in vivo (but not in

vitro) growth and cell survival. To reveal the underlying pathomechanism, we have analysed

the angiogenic phenotype of αIIbβ3 integrin-transduced human melanoma cells expressing

integrin αvβ3. The cDNA microarray analysis of the 19H cells revealed 12 down-regulated

and 36 up-regulated genes. Three out of nineteen known genes, up-regulated significantly in

αIIbβ3-transfected 19H melanoma cells, are endothelial cell-specific genes: CD34, endothelin

receptor B, and prostaglandin I-2 synthase. We propose that the illegitimate expression of

αIIbβ3 integrin in human melanoma cells already expressing αvβ3 integrin may alter their in

vivo growth properties due to the modulation of their angiogenic phenotype.

6

V. INTRODUCTION

5.1. Functional genomics in oncology

Dramatic technological advances have resulted in overwhelming information on

biological systems. Advances in manipulating and sequencing DNA triggered the last such

leap. More than twenty years ago this technology began finding its way into biology

laboratories and multiple landmark discoveries have followed, including the sequencing of the

genomes of several prokaryotic and eukaryotic species. This, in turn, spawned a new interest

in bioinformatics, structural biology, and high-throughput methods that would allow scientists

to look at the response of the entire genome to physiological, pharmacological, and

pathological changes.

The development of our understanding of the cell and molecular biology of the

neoplastic process has closely paralleled advances in our understanding of tumour cell and

molecular biology of normal cells. The vast majority of the phenotypic and genotypic

alterations seen in neoplasia—i.e., cells in the stage of progression—are a direct result of the

primary characteristic of this stage, evolving karyotypic instability. The discoveries of

chemical, physical, and biological carcinogenic agents and their actions have been among the

most exciting and significant in our understanding of the causation of cancer and of many

aspects of its prevention, but nothing has intrigued the biological scientist more than the

molecular differences between cancer cells and normal cells.

Many researchers have described a number of processes and functions in the cell and

in molecular biology of preneoplasia and neoplasia, namely the following: glycolysis,

regulation of gene expression, signal transduction, transcription factors, cell cycle, DNA

methylation, apoptosis, telomerase mechanism, unique or mutant protein expression, unique

lipids and carbohydrates. However, it is important to note that there is no single component of

any of these processes that is ubiquitously abnormal in all neoplasms. However, the majority

of neoplasms exhibit at least one defective process in each of those listed. A comparison of

the cellular and molecular biology of preneoplasia with that of neoplasia does, however,

emphasize the importance of an understanding of the molecular transition between these two

stages as critical to our understanding of the ultimate formation of the cancer cell.

7

For the identification of chromosomal abnormalities, high-throughput methods are

becoming more and more routine and they are comprising of fluorescent in situ hybridization

(FISH) (1, 2), spectral karyotyping (SKY) (3), CGH (4), and microsatellite analysis (5). FISH

has a prominent role in the molecular analysis of cancer and can be used for the detection of

numerical and structural chromosomal abnormalities. The recently described SKY, in which

all human metaphase chromosomes are visualized in specific colors, allows for the definition

of all chromosomal rearrangements and marker chromosomes in a tumour cell. Protocols for

the detection of chromosomal rearrangements by PCR and RT-PCR are described, as well as

the technique of DNA fingerprinting, a powerful tool for studying somatic genetic alterations

in tumourigenesis. A number of approaches to identify mutations are detailed, and include

single-stranded confirmational polymorphism (SSCP) (6), denaturing gradient gel

electrophoresis (DGGE) (7), the nonisotopic RNase cleavage assay (8), the protein truncation

assay (9), and DNA sequencing (10). A change in DNA methylation status is commonly

observed in cancer, and specific methodology for methylation analysis (11) is also provided

by this volume. A reduction in telomere length, together with expression of the telomere

maintenance enzyme, telomerase, has been described in case of a wide range of human

cancers (12). The research field of the spotted DNA microarray is dealing with the structure

and activity of genomes and global relationships between genotype and phenotype. The birth

of the field can be traced back to three seminal papers (13-15). Similar interminable

approaches have been used to monitor relative protein levels (16), protein functions (17),

cellular activities (18), and molecular interactions (19). The analysis of gene expression

represents a key area of research in the study of human cancer. Global RNA expression

analysis using microarray technology (20) allows the identification of genes that are

differentially expressed in tumour versus normal tissues (21). This is a powerful approach for

identifying genes that are central to disease development or progression and can also identify

new prognostic markers.

In this thesis we describe a new QRT-PCR-based DOP-PCR method which is capable

of preserving the original copy number of different chromosomal regions in amplified

genomic DNA. Furthermore, genomic imbalances were detected by CGH with microarray to

assess the genetic relationship between MCC and later appearing tonsillar tumour, in addition

to molecular genetic markers in the case of PTCs of different disease courses. Finally, the

8

gene expression pattern of αIIbβ3 integrin transfected human melanoma cells was

demonstrated to define molecular pathomechanism of increased melanoma growth and

angiogenesis by DNA microarray technique.

5.2. Quantitative real-time PCR

5.2.1. Applications of QRT-PCR The ability to monitor the real-time progress of the PCR completely revolutionizes the

way one approaches PCR-based quantification of DNA and RNA. The development of

fluorescent detection systems, capable of monitoring PCR product accumulation has greatly

improved the reliability of QRT-PCR. Quantitative PCR is aimed to determine the absolute or

relative amounts of RNA or DNA sequences in a given sample. There is a well-established

inverse correlation between the template concentration and the duration of the lag phase of the

PCR. An entire generation of automated systems has been developed to exploit this. Using

any of several different chemistries, the accumulation of PCR product is monitored until a

predetermined threshold is reached. This lag time is then compared with a standard curve and

a concentration calculated (22). Reactions are characterized by the point in time during

cycling when amplification of a PCR product is first detected rather than the amount of PCR

product accumulated after a fixed number of cycles. The higher the starting copy number of

the nucleic acid target, the sooner a significant increase in fluorescence is observed.

Advantages of this method not only include the elimination of differences in reaction mix

volumes and conditions that may occur in separate samples, but importantly it also allows

reference of RNA or DNA quantitation to an internal control. Consequently, this system of

quantitation has the advantage in that it allows the quantitation of multiple sequences from a

single sample and a single experiment of amplification without the need of standardization for

each sequence.

The QRT-PCR technology is now widely used, especially in such areas as therapeutic

monitoring at the assessment of residual disease after treatment of leukemia and lymphoma

(23-25), quality control, disease diagnosis (26), the detection of viral nucleic acids, regulation

of gene expression (27), and the detection of gene amplification or deletion and of aneuploidy.

9

5.2.2. Determination of copy number changes in whole genome and validation of CGH-

array profiles

PCR and QRT-PCR are now well-established methods for the study of nucleic acid

sequences in histological material. QRT-PCR is the reliable detection and measurement of

products generated during each cycle of the PCR which are directly in proportion to the

amount of template prior to the start of the PCR process.

The applications of QRT-PCR are diverse and numerous, including such areas as

DNA/cDNA copy number measurement in genomic or viral DNA/RNA, functional genomics

including mRNA expression analysis, allelic discrimination assays and confirmation of

microarray data. Upstream modifications to the technique, including the use of laser capture

microdissection, add an extra dimension in facilitating the production of homogenous cell

populations from complex tissue sections for more accurate quantitative analysis.

The challenge is to put the technology to full use by identification of the genetic

changes that are important in human disease. Knowledge of these changes will permit the

establishment of rapid, cost-effective tests for diagnosis and monitoring of disease as well as

providing a basic understanding of their underlying causes. These developments will be

enhanced by unravelling of the complete human genome and by improvements in techniques

for gene amplification using histological samples.

Data obtained from the microarray experiments need validation by independent

techniques like QRT-PCR or metaphase CGH. To determine the relative copy number ratios

of the genome and in order to strengthen the CGH-array data, QRT-PCR was applied in this

study.

5.3. Whole genome amplification techniques

High-throughput genomic analysis requires large amounts of template; however, the

typical yield of DNA from individual samples can be often limited. The primary drawback of

the use of laser capture microscopy in DNA analysis is that microdissections yield insufficient

nucleic acids due to low genomic DNA recovered from small sample sizes. With samples

such as these, the ability to conduct sample amplification becomes imperative, to ensure that

enough material is available for genetic analysis. There are three basic techniques to amplify

10

the entire genomic DNA: PCR-based approaches (28-38), strand or multiple displacement

amplification (SDA or MDA) (39) and T7-based linear amplification (TLAD) (40).

5.3.1. Strand or multiple displacement amplification (SDA or MDA) and T7-based linear

amplification

The SDA and MDA isothermal methods employ the unique biochemical properties of

Phi29 DNA polymerase to amplify linear DNA. These techniques are based on the rolling

circle amplification by which circular DNA molecules such as plasmids or viruses frequently

replicate. These rolling circle amplification methods were initially adapted for the

amplification of large circular DNA templates (41, 42) and more recently for the

amplification of genomic DNA (39). The drawback of SDA is that the efficiency of

amplification depends on the length of the template. T7-based linear amplification is

applicable for the amplification of genomic DNA of varying fragment size and quality, and it

is based on a protocol devised by Phillips and Eberwine for the amplification of mRNA for

cDNA microarrays (40, 43).

5.3.2. PCR-based genomic DNA amplification methods

The inter-spread repetitive sequence PCR (38), was the first PCR-based genome

amplification method, which used primers designed to anneal to Alu repeats within the

genome. This strategy was further improved to amplify specific DNA flanked by Alu repeats

by using the so-called Restricted-PCR technology (44). An alternative method (32), termed

the linker adapter technique PCR (LA-PCR), involved restriction digestion of the DNA and

ligation of adaptors to serve as priming sites for subsequent PCR. In principle the ligation

mediated PCR (LMP) is very similar to the LA-PCR method (32, 33). Nonetheless, this

approach has not been widely applied, likely due to technical difficultes involved. PCR-based

genome amplification methods, primer extension preamplification (PEP) (34, 35) and DOP-

PCR (36, 37) techniques have been developed for the amplification of genomic DNA starting

from as little as a single cell. The original PEP protocol involves a 50 cycle PCR program

using a 15 bp long oligonucleotide primer; however, via this approach only ~78% of the

genome from a single haploid cell can be copied (34).

DOP-PCR and PEP protocols amplify DNA in an exponential fashion, making them

highly susceptible to biases, and they have a tendency to specifically over-represent or down-

11

represent certain sequences. The major disadvantage of PCR-based protocols is that this could

distort the ratio of the starting genomic imbalances (45). Recently, Wang et al. have described

a modified DOP-PCR protocol for balanced amplification of genomic DNA (46), however the

amplification status has not been followed during PCR. In their protocol genomic DNA was

digested with a restriction enzyme and tagged linkers were ligated to the products before

amplification.

5.3.2.1. Degenerate oligonucleotide-primed PCR

The DOP-PCR technique is increasingly being applied for simultaneous amplification

of multiple loci in target DNA using oligonucleotide primers of partially degenerate

sequences (36); however, the amplified DNA does not cover the genome completely.

Contrary to other PCR-based amplification methods (Alu-PCR, IRS-PCR), DOP-PCR is a

species-independent technique. The protocol is adapted for the use of microdissected,

formalin-fixed and paraffin-embedded tissue in comprehensive genetic analyses (47).

As is shown in Figure 1/a, primers used for DOP-PCR contain a variable binding

region with a sequence of six randomly combined nucleotides which is flanked by a stringent

primer binding region at the 5’-end (ten nucleotides) and a main primer binding region at the

3’-end (six nucleotides) which defines the specificity of the reaction.

Figure 1/a Sequence of the DOP-PCR primer

The amplification procedure is divided into two steps (Figure 1/b). During the first

step, eight PCR cycles are performed at low stringency conditions (annealing temperature, Ta:

30°C). These conditions allow a frequent annealing of the DOP primer, which is primarily

mediated by the short specific sequence at its 3’-end and by the adjacent central cassette of

degenerated nucleotides. Eventually a pool of many different DNA sequences is generated

which should represent a statistically distributed subpopulation of the genomic DNA. During

the second step a more stringent annealing temperature is applied and the annealing

CCGACTCGAG5’- NNNNNN ATGTGG -3’

Stringent primer binding region

Variable binding region

Main binding region




Main binding region





Main binding region




Main binding region


12

temperature is raised to 62°C, allowing amplification only after specific base pairing of the

full length primer sequence. After complete DOP-PCR reaction, a high amount of similar

DNA fragments will be generated with DOP-primer sequences at both ends.

Figure 1/b Reaction mechanism of DOP-PCR In our improved protocol of DOP-PCR, direct amplification of genomic DNA can be

performed without any pretreatment of the samples. Our protocol includes QRT-PCR to

follow the cycling status which results in a more reliable and reproducible sample

amplification.

DOP-PCR is a simple method for whole genome amplification in comprehensive

genetic studies (21) such as metaphase CGH (4), array CGH (48), SNP genotyping (49),

microsatellite genotyping (5), SSCP (6), loss of heterozygosity (LOH) (50), and sequence

analysis (10), that is increasingly being applied for simultaneous amplification of multiple

loci in target DNA. DOP-PCR utilizes oligonucleotide primers with partially degenerate

sequences close to their 3’ end (36). DOP-PCR is a species-independent approach, which

produce sufficient amounts of DNA for analysing clinical tumour samples, for prenatal

diagnosis and for many other investigations. A comparison of several DOP-PCR methods is

also described by Larsen et al. (47). Not only the amount but also the quality of genomic

DNA obtained from various samples is impediment for further studies. For example, paraffin-

embedded tissues often provide partially degraded low-quality DNA. Unfortunately, PCR-

First reaction step under low stringency conditions

Second reaction step under high stringency conditions





13

based protocols often result in misrepresentation of the genome, with some regions over-

represented or even lost. The reliability of PCR-based amplification mostly depends on the

exponential phase during cycling. At the late cycles of amplification, the original copy

number of chromosomal regions can be changed and the method provides false results. The

plateau effect (51, 52) is the result of a marked shift of the overall mass balance in favor of

the reaction product. Different factors seem to be responsible for the attainment of the plateau

rather than a single factor or parameter, as re-addition of presumably exhausted reagents at

late cycles did not cause the reaction to proceed with increased efficiency (53). The plateau

effect is caused by the accumulation of product molecules, usually it is the consequence of the

significant degree of annealing between complementary strands of the products, rather than

between the primers and template. Furthermore, the finite amount of enzyme molecules

present is unable to extend all the primer-template complex in the given extension time (54).

5.4. DNA microarray techniques and comparative genome hybridization

5.4.1. DNA microarray technology

Microarray technologies as a whole provide new tools that transform the way

scientific experiments are carried out. In place of conducting experiments based on results

from one or a few genes, while microarrays allow for the simultaneous interrogation of

hundreds or thousands or tens of thousands of genes.

Microarrays are microscope slides that contain an ordered series of samples (DNA,

RNA, protein, tissue). The type of microarray depends upon the material placed onto the

slide: DNA microarray, RNA microarray; protein microarray; and tissue microarray. Since

the samples are arranged in an ordered fashion, data obtained from the microarray can be

traced back to any of the samples. This means that genes on the microarray are addressable.

The number of ordered samples on a microarray can reach up to a hundred thousand. A

typical microarray contains several thousands of addressable genes. DNA printed or spotted

onto the slides can be chemically synthesized long oligonucleotides or enzymatically

generated PCR products. The slides contain chemically reactive groups (typically aldehydes

or primary amines) that help stabilizing the DNA onto the slide, either by covalent bonds or

electrostatic interactions. An alternative DNA chip technology allows the DNA to be

synthesized directly onto the slide itself by a photolithographic process. This process has been

commercialized and is widely available.

14

DNA microarrays are used to determine the expression levels of genes in a sample

(expression profiling) and have been highly successful in a variety of genomic analyses (55),

ranging from the detection of SNPs (56) to functional genomics (57-59). Furthermore, DNA

microarrays can be used to detect and map copy number changes in regions of the genome.

5.4.2. Development of CGH method

CGH is an in situ hybridization technique used to characterise chromosomal

abnormalities where there is a loss (deletion) or gain (amplification) of genetic material (4). It

is based on the competitive hybridization of labelled tumour DNA and normal DNA to

normal metaphase chromosomes. CGH is not useful for the detection of balanced

translocations or inversions where there is no overall gain or loss of material. The main

advantage of CGH compared with conventional cytogenetics is that it is not necessary to

obtain metaphase chromosome spreads from the tumour under investigation. Thus, CGH is

ideally suited to the analysis of samples where it is difficult to obtain good quality metaphase

chromosomes. Unlike conventional cytogenetics, CGH is applicable to the study of archival

tumour specimens (frozen and paraffin). Recently, using a DOP-PCR method it has been

possible to amplify sufficient DNA for CGH even from minute amounts of starting material

and in turn this means that CGH can now be regularly performed on microdissected samples.

It is now possible to microdissect small pre-malignant lesions and to begin to unravel the

genetic changes that might characterise the early changes leading to malignancy.

5.4.3. CGH combined with microarray technique

CGH has already showed a significant impact on the field of cancer cytogenetics as a

powerful tool for detection of chromosome copy number aberrations even in epithelial solid

tumours, in the case of which it is hard to elucidate tumour specific genome alterations by

conventional cytogenetic methods. However, the resolution of CGH to metaphase

chromosomes can provide only limited resolution at 5-10 Mb level for detection of copy

number losses and gains, and at 2Mb for amplifications. To circumvent this limitation, an

innovative strategy, called matrix-CGH (60) or array-based CGH (61) has been devised. CGH

using cDNA microarray was also reported to be useful for detection and mapping of gene

amplification and homozygous deletions (62). In array CGH, chromosomal targets are

replaced by arrays consisting of well-defined genomic clones such as BAC, PAC or cosmid

15

clones, which are spotted onto a microscopic slide-glass using robotic devices. Since the

clones spotted on slide-glass contain sequence information directly connecting with the

genome database, we can easily obtain particular biological aspects of genes mapped within

regions involved in copy number aberrations detected by array-CGH, facilitating

identification of genes responsible for cancer.

5.5. Molecular pathomechanism of different tumour types

5.5.1. Merkel cell carcinoma

The Merkel cells are found in the skin and in those parts of the mucosa derived from

the ectoderm (63). These cells are the origin of a rare, malignant neuroendocrine tumour that

occurs predominantly in the sun-exposed areas of the skin called MCC (64-70). Local

reccurence, and regional or distant metastasis generally develop within a short period of time.

Oropharyngeal metastasis is very rare, and metastasis to the palatine tonsil has been described

in only one case (71). A perioral or intraoral localization of the MCC is very infrequent: (64,

72) to date, 10 cases of MCC of the lip (73) and 14 cases of intraoral MCCs have been

described (64, 74, 75). During the past decade, our co-workers at University of Szeged treated

and followed up an elderly woman with MCC of the upper lip. After a long tumour-free

period, an anaplastic carcinoma with neuroendocrine features developed in her palatine tonsil,

raising the possibility of a late haematogenous metastasis, a second field tumour, or a second

primary tumour.

The term ”secondary field tumour” was introduced in 1953 by Slaughter et al. who

examined slides from patients with head and neck cancer (76). It was observed that all of the

epithelium beyond the margins of the tumours displayed histologic alterations, and 11% of the

patients were found to have more than one independent area of malignancy. The authors

concluded that the mucosa of the head and neck had undergone a change, perhaps due to

carcinogen exposure, and was therefore more susceptible to the development of foci of

malignant transformation. Organs in which field cancerization has been described since then

are the oral cavity, oropharynx, larynx, lung, oesophagus, colon, skin, vulva, cervix, breast,

renal pelvis, ureter and bladder (77-80).

The determination of molecular patterns of first and second tumours has become a

valuable tool to explore the relationship between them: similar aberrations indicate a

metastasis, partly different aberrations indicate a second field tumour, and different

16

aberrations denote a second primary tumour (78-80). We identified the molecular patterns

with CGH with microarray technology. Array-based CGH consists of a series of mapped

artificial bacterial chromosomes or sequenced human cDNA clones on glass slides, to which

DNA from test and control samples is hybridized (81-83).

Previous CGH studies of MCCs have identified divergent regions that are affected, with

a small number of similarities between the samples analysed. Recurrent chromosomal

imbalances detected by CGH analysis were loss of 3p, 4p15-pter, 10q, 13q and 17p and gains

of 1q, 3q, 5p, 6, 8q, 18 and 20 (84-87). In this study, we detected some of the previously

found regions and also numerous novel regions with chromosomal imbalances.

5.5.2. Classification of the thyroid carcinomas

There are four types of thyroid cancer some of which are much more common than

others: Papillary and/or mixed papillary/follicular thyroid carcinoma (PTC) (78%); Follicular

and/or Hürthle cell (FTC) (17%); Medullary (4%); Anaplastic (ATC) (1%). Thyroid cancer

can occur in any age group, although it is most common after age 30 and its aggressiveness

increases significantly in older patients. The majority of patients present with a nodule on

their thyroid and far less than 1% of all thyroid nodules are malignant.

5.5.2.1. Papillary and anaplastic thyroid carcinoma

The most frequent thyroid cancer is PTC, and this type is one of the most curable

cancers with ten year’ survival rates (at about 80-90%). Cervical metastasis (spread to lymph

nodes in the neck) are present in 50% of small tumours and in over 75% of the larger thyroid

cancers. The presence of lymph node metastasis in these cervical areas causes a higher

recurrence rate but not a higher mortality rate. Distant metastasis (spread) is uncommon in

PTC, but during follow-up 6 to 11% of PTC patients develop distant metastases. Although

patients may live for long time with distant metastases, this does significantly worsen

prognosis. About one-third of patients with distant metastases survive for 10 years. That is

two-third of them die in ten years. An other form of aggressive disease outcome is local

recurrance and invasion. Age is important for both predicting development of distant

metastases and for influencing long-term survival.

The least common type of thyroid cancer is ATC which has a very poor prognosis. ATC

tends to be found after it has spread and is not cured in most cases. Anaplastic tumour has a

17

very low cure rate and most patients with ATC do not live one year from the day they are

diagnosed. ATC often arises within a more differentiated thyroid cancer or even within a

goiter. It seems a generally accepted hypothesis, that they originated from differentiated

thyroid cancers. Cervical metastasis (spread of the cancer to lymph nodes in the neck) are

present in the vast majority (over 90%) of cases at the time of diagnosis. The presence of

lymph node metastasis in these cervical areas causes a higher recurrence rate and is predictive

of a high mortality rate. Anaplastic cancers invade adjacent structures and metastasize

extensively to cervical lymph nodes and distant organs such as lung and bone.

5.5.2.2. Diagnostic molecular markers in papillary and anaplastic thyroid carcinomas

Tumour markers can be used to screen a healthy population or a high-risk population

for the presence of cancer, to make a diagnosis of cancer, to determine the prognosis of the

patient, or to monitor the course in a patient during follow-up. Several studies have

investigated tumour markers in PTC, and some studies have reported them as promising for

predicting prognosis, but no warranted markers for PTC have been established. At present

thyroglobulin is the only tumour marker of PTC being routinely used to determine the

effectiveness of thyroid cancer treatment and to screen for recurrence.

To identify prognostic genetic markers (polymorphic microsatellites, SNPs, changes in

gene expression or in gene copy number) which are associated with poor prognosis has been

in the focus of several studies. Recently Siironen and co-workers analyzed the expression of

COX-2, MMP-2, VEGF-C, Bcl-2, Ki-67 and p21 by immunohistochemistry in 72 samples (36

with aggressive and 36 with benign form) (88), and Gerdes et al. have demonstrated that Ki-

67 expression associated with tumour grade and number of mitoses (89). None of these

proteins showed superior classification over TNM. Therefore more studies and more robust

techniques are needed to predict the progression of the agressive form.

The frequency of LOH on each chromosome arm was determined in 24 patients

samples with poor prognosis and in 45 survived cases by Kitamura et al. (90). They found

significantly higher frequencies of LOH on 1q, 4p, 7q, 9p, 9q and 16q loci among the poor

outcome patient population, however finer resolution of possible genetic markers were not

identified. Loss of heterozigocity at the TPO gene locus has been implicated as a cause of the

organification defect typical of benign and malignant thyroid tumours (91, 92). Amplification

18

of the phosphatidylinositol 3-kinase (PIK3CA) gene is relatively common and may be a novel

mechanism in activating the PI3K/Akt pathway in some thyroid tumours (93).

All three RAS genes (NRAS, KRAS, HRAS) can be mutated in PTC (94), and Ras

mutations and RET/PTC rearrangements are associated with aggressive tumour phenotypes

(95-99). The RET gene is found to be rearranged in approxymately 40% of PTCs (100-103).

Some authors have desribed BRAF mutation in thyroid tumours (104-110). The majority of

ATCs with papillary components are derived from BRAF-mutated PTC, and this indirectly

supports the notion that PTC with RET/PTC rearrangements do not progress to ATC.

RET/PTC rearrangements activate multiple downstream signaling pathways, including the

RAS-RAF-MAPKK-MAPK pathway, do not progress to ATC; whereas PTC with BRAF

mutation which activates fewer signaling pathways than RET/PTC, can progress to ATC

(111). Expression of RET/PTC1 and RET/PTC3 oncogenes was identified in both benign and

malignant studies (112), while RET protein expression has been evaluated in PTCs (113).

TRK gene chromosomal rearrangements are found in 10% of PTCs (114, 115). The tumour

suppressor gene, p53, is rarely mutated in well differentiated PTC and FTC, but is mutated at

a moderate rate in poorly differentiated PTC and at a very high rate in ATC (116-119) but not

in its PTC precursor. This indicates that development of ATC is due to further p53 mutation

in the setting of an oncogenic mutation such as BRAF (111).

Since no ideal examination method or protocol has been developed by which the

circumstances of development and outcome of PTC were able to be defined exactly – which

is necessary for the use of aimed therapeutic medicine –, we performed CGH-array analysis of

a primary- and terminal-stage APTC, comparing it to GPTC obtained from paraffin-

embedded tissue to identify gene copy number alteration as markers which might predict the

outcome of an individual patient better than done by TNM classification alone.

5.5.3. The effect of αIIbβ3 integrin on increased angiogenesis and tumour growth in melanoma

Progression of human melanoma is a highly efficient process compared to other

malignant tumours since a few millimeters thick tumour (a very small primary in the case of

other malignancies) can colonize regional lymph nodes or visceral organs. It is therefore

extremely important to understand the molecular mechanisms behind this highly aggressive

behavior. Previous studies indicated that the autocrine growth regulation of melanoma

19

acquired at the switch from a less invasive radial to a more invasive vertical growth phase

involves bFGF expression (120, 121).

Some studies have demonstrated that increased expression of αvβ3 integrin in

activated endothelial cells promotes angiogenesis and its blockade promotes endothelial cell

apoptosis (122-124).

Integrins are ubiquitous transmembrane α/β heterodimers that mediate diverse

processes requiring cell-matrix and cell-cell interactions such as tissue migration during

embryogenesis, cellular adhesion, cancer metastases, and lymphocyte helper and killer cell

functions (125). Platelets express 3 members of the ß1 subfamily (αIIβ1, αvβ1, and αv1β1) that

support platelet adhesion to the ECM proteins collagen, fibronectin, and laminin, respectively

(126-129), and both members of the β3 subfamily (αvβ3 and αIIbβ3). Although αvβ3

mediates platelet adhesion to osteopontin and vitronectin in vitro (130, 131), it is uncertain

whether it plays a role in platelet function in vivo. By contrast, αIIbβ3, a receptor for

fibrinogen, vWF, fibronectin, and vitronectin, is absolutely required for platelet aggregation.

Because αIIbβ3 plays an indispensable role in hemostasis and thrombosis, it is among the

most intensively studied integrins. Expression of αIIbβ3 is restricted to cells of the

megakaryocyte lineage. In megakaryocytes, αIIbβ3 is assembled from αIIb and β3 precursors

in the endoplasmic reticulum (132) and undergoes posttranslational processing in the Golgi

complex, where αIIb is cleaved into heavy and light chains (133). αIIbβ3 can support the

adhesion of unstimulated platelets to many of its ligands when they are immobilized in vitro,

platelet stimulation is required to enable αIIbβ3 to mediate platelet aggregation by binding

soluble fibrinogen and vWF (134).

Different integrins may crosstalk with each other, thereby regulating each other's

function. For instance αIIbβ3-mediated signaling in CHO cells influences β1 integrin-

mediated cell adhesion and spreading (135). This phenomenon of transdominant signaling

appears to involve MAPK activity. Activation of myosin light chain and lipid kinases by the

MAP kinase pathway and the Rho family of GTPase is modulated by integrin activity, which

provides a link between the extracellular matrix and effect on cell shape and motility that

ultimately influences tumour cell invasion. Although the relationship between integrin

receptors and metastasis is variable, an increase in β3 integrin expression directly correlates

with progression of melanoma (136, 137).

20

Studies have demonstrated that the megakaryocyte specific αIIbβ3 integrin is

expressed in cultured cell lines derived from solid tumours and this integrin is expressed both

in vitro and in vivo (138-144). Compared to expression of αvβ3, expression of αIIbβ3 is much

lower and decreases with cell culture in vitro (145).

The αIIbβ3 integrin (GpIIbIIIa) is the predominant adhesion receptor of platelets and

the expression of the integrin αIIb chain is megakariocyte-specific. It is involved primarily in

platelet activation, since it is expressed in a low affinity state and binds fibrinogen only upon

activation. The illegitimate expression of αIIbβ3 integrin in human melanoma has been

described (146, 147), involved in the same phases of tumour progression as αvβ3.

The transduction of αIIbβ3 into αvβ3-expressing human melanoma cells did not affect

the in vitro growth, but promoted the in vivo growth of tumour cells due to decreased

apoptosis (147). We have postulated that one possible factor behind the increased in vivo

growth is vascularization. Therefore, we have analysed the gene expression changes of

αIIbβ3-transfected human melanoma clones with special attention to their angiogenic

phenotype.

21

VI. MATERIALS AND METHODS

6.1. Tissue specimen and genomic DNA isolation

6.1.1. Samples for the improved DOP-PCR technique

Melanoma tissues were obtained from the Department of Dermatology, University of

Debrecen, Hungary. Primary melanoma (location: face) and metastatic tumour (location:

lymph node) were removed from the same patient at the same time. The relative amount of

tumour cells was a minimum of 70% in all lesions as confirmed after hematoxylin and eosin

staining by light microscopy. High–molecular-weight DNA was extracted according to a

standard procedure starting from 50 mg of tissue (46, 148). Chromosomal CGH and image

analyses were performed as described earlier in details elsewhere (148).

Epidermoid, bronchioloalveolar lung carcinoma, 2 renal cell carcinoma, and 2

colorectal carcinoma specimens were obtained from the Institute of Pathology, Szeged

University. Conn and Cushing adenoma samples were obtained from Semmelweis University,

Budapest. DNA was purified by using the DNA purification kit from Macherey-Nagel

(Düren, Germany) according to the manufacturer’s instructions.

6.1.2. Samples for the Merkel cell carcinoma study

Between 1970 and 2003, 4418 patients with malignant orofacial tumours were treated

in Department of Dentistry and Oral Surgery at University of Szeged. One of these patients

suffered from MCC. A 79-year-old Caucasian woman presented with a tumour in her upper

lip. She was an outdoor worker and never smoked. She was disturbed only about the

aesthetics. The physical examination revealed a 2–3-cm firm, compact mass in the skin of the

upper lip on the left side. It was sharply separated from its surroundings and had a

teleangiectatic surface (Figure 2).

Figure 2. MCC developed in the skin of the upper lip

22

The tumour was classified as stage T2. No regional lymph-node metastasis was detected. Surgical resection of the tumour and removal of the regional lymph nodes were suggested, but she refused it. As an alternate therapy, a biopsy was taken, and cryotherapy was applied according to the standard protocol. The histological and immunohistochemical characterization was done in the Department of Pathology at Medical University of Szeged. Light microscopy revealed that the tumour was confined to the dermis and was sharply separated from the epidermis. The tumour cells were arranged in nests and cords. Cytologically, the tumour cells were monomorphic, the nuclei were round and the chromatin pattern was finely granular and “dusty”. The nucleoli were small; often two or even three were detected. The cytoplasm was scanty. The mitotic rate was high (six to ten figures/ high-power field). The Grimelius staining was negative. The results of immunostainings are shown in Table 1. The histological and immunohistochemical features pointed to MCC of the lip. Following the histological diagnosis, distant metastases in the lungs and bones were searched for, with negative results.

Immunohistochemical

markers

Lip

tumour

Tonsillar

tumour

AE1/AE3 +++ +++

Cytokeratin 20 20

Paranuclear dots +++ ++

Chromogranin ++ +

Synaptophysin - -

TTF - -

HMB-45 - -

CEA - -

Lymphoid markers - -

Table 1. Immunohistochemical findings In the 7-year follow-up, the patient was in a tumour-free condition. After 7 years, she

was admitted to the County Hospital in Kecskemét with a left palatine tonsillar tumour. This tumour was resected in part, and the regional lymph nodes were removed. Histologically, the tumour displayed features of anaplastic carcinoma.

The morphological appearance of the tumour resembled that of the MCC of the lip. A further excision was performed 3 months later. The morphological and immunohistochemical

23

features of the tonsillar tumour were essentially the same as observed 3 months earlier. The patient died at home on the 20th postoperative day. Autopsy was not performed.

Paraffin-embedded tumour tissues and unaffected lymph nodes (200-500 mg) were deparaffinated in hexane, and washed with ethanol. DNA was purified by using the DNA purification kit from Macherey-Nagel (Düren, Germany) according to the manufacturer’s instructions.

6.1.3. Papillary thyroid carcinoma samples PTC was detected in a 63-year-old male patient; carcinoma was confirmed by

cytology and other investigations. Total thyreoidectomia and lege artis ablatio with 131I were performed; unfortunately, after eight months, local recidiva developed. A necessary reoperation was carried out by the University of Debrecen; despite the operation, however, the tumour spread into the pharynx, main vessels, and pleura, and the patient died three months later. On the basis of histology results, anaplastic dedifferentiation was detected. The primary- and terminal-stage APTC samples obtained from the above patient were analysed compared GPTC and normal thyroid tissue in further QRT-PCR studies. The disease course of the female patient with GPTC providing control samples in some experiments was followed strictly for five years after operation and radioiodine ablation.

Other PTC samples included in further studies and normal thyroid tissues were obtained from the Department of Pathology University of Debrecen, Jósa András County Hospital in Nyíregyháza, Erzsébet Hospital in Hódmezővásárhely, County Hospital in Kecskemét and Pándy K County Hospital in Gyula. Two disease course groups of our cases were created: the GPTC (group A): patients with no recurrance and no metastases, all of them are disease free up to now. There were 10 patients of GPTC, all of them were female. Mean age was 42.1 years (range: 27-52 years) at the time of the operation. Median follow-up was 4.8 years (range: 3-7). The second group was the I/APTC (group B) (14 intermediate – i.e. metastasis developed only in cervical lymph nodes, by which the survival rate is much higher than by APTC, and 10 malignant APTCs. This group B consisted of 24 patients (17 female, 7 male). Patients with cervical (12) and regional (mediastinal) lymph node (2), pulmonary (6), bone (3) metastases, trachea infiltration (1), and local invasion (4) in different combination. The mean age of patients of the I/APTC group was 48.3 years (range: 13-76 years) at the time of the operation. Median follow-up was 6.8 years (range: 9 months-17 years). Eight of them died, and the mean survival time following operation was 5.6 years (range: 9 months-22 years).

24

Paraffin-embedded tissues (500 mg) were deparaffinated in hexane and washed with ethanol. DNA was purified by the DNA purification kit from Macherey-Nagel (Düren, Germany), according to the manufacturer’s instructions.

6.2. DNA and RNA quantity measurement

The quantity of genomic DNA and total RNA was assessed spectrophotometrically by NanoDrop (Rockland, DE, USA). Genomic DNA was used for microarray analysis and for QRT-PCR.

The WM983B cell line that does not express αIIbβ3 on the cell surface was a kindly provided to us by M. Herlyn (The Wistar Institute, Philadelphia, PA). Mock transfected WM983B cells (3.1P) and αIIb and β3 transfected WM983B cells (19L and 19H) have been described in a study by Trikha M. et al. (147). 19H transfected cell line was subcloned by limited dilution to obtain 7D7 and 8F3 subclones. The parental cell line was grown in RPMI-1640 medium supplemented with 5% FBS and antibiotics (Sigma Chemical Co., St. Louis, MO) in a 5% CO2 humidified atmosphere at 37°C. Transfected cells were maintained in medium containing G418 (Gibco BRL, Gaithersburg, MD). RNA isolation from 5x106 cells was carried out with the RNA isolation kit of Macherey-Nagel (Düren, Germany) according the manufacturer’s instruction. The RNA concentration was assessed spectrophotometrically by NanoDrop (Rockland, DE, USA) and the quality was checked by electrophoresis. RNA was stored at -80°C in the presence 30 U of Prime RNAse inhibitor (Fermentas, Vilnius, Lithuania). Total RNA was used for RT, microarray analysis and for QRT-PCR.

6.3. Quantitative real-time PCR

6.3.1. Preparation of DOP-PCR amplified and overamplified genomic DNA

Genomic DNA (20 ng from native tissue and 100 ng from paraffin embedded tissue) was amplified with a modified version of the DOP-PCR protocol by using a RotorGene 3000 QRT-PCR instrument, (Corbett Research, Mortlake, Australia), to follow the amplification. The DNA concentration was assessed spectrophotometrically by NanoDrop (Rockland, DE, USA). Reactions were performed in a total volume of 100 µl. The cycling parameters were as follows: heat start at 95 oC (15 min); 8 cycles denaturation at 94 oC (50 s); annealing for 2 min from 45 oC to 72 oC with a 0.2 oC/s ramp; and, extension at 72 oC (90 s). After 8 cycles, the reaction mix was divided into two 50 µl aliquots and SybrGreen was added to both (1x final concentration, Molecular Probes, Eugene, USA). The following cycling protocol was

25

performed in a real-time PCR instrument: denaturation at 95 oC (40 s); annealing at 58 oC (60s); and, extension at 72 oC for 80 s. To avoid over-amplification of the products, cycling was terminated just before the reaction reached the plateau (i.e., at 13-15 cycles); in contrast, the over-amplified genomic DNA was obtained from PCR with 21 cycles.

The PCR products were purified with PCR-purification columns (Bioneer, Daejeon, Korea). In the reactions, UN primer (5’-CCGACTCGAGNNNNNNATGTGG-3') was used in 1 µM concentration (36). The reactions were performed with ExTaq DNA polymerase in 1X buffer (Takara, Tokyo, Japan) in the case of MCC, while the reactions were performed with 1x ABsolute QPCR Mixes (ABgene, Surrey, UK) in case of confirmation of improved DOP-PCR and PTC.

6.3.2. Determination of the differences in relative copy numbers of the genome

QRT-PCR was carried out in a final volume of 20 µl with 20 or 100 ng (DNA obtained from native or paraffin-embedded tissues accordingly) non-amplified or amplified genomic DNA in a RotorGene 3000 instrument (Corbett Research, Mortlake, Australia). The reactions were performed with 5 pmol/each forward and reverse gene-specific primers in 1x ABsolute QPCR Mix (ABgene, Surrey, UK) with SybrGreen (1 µM final concentration, Molecular Probes, Eugene, USA) with the following protocol: heat start for 15 min at 95 °C, 45 cycles of 25 s denaturation at 95 °C, 25 s annealing at 59 oC and 25 s extension at 72 oC. Fluorescent signals were collected after each extension step at 72 oC. Curves were analysed with the RotorGene software, using dynamic tube and slope correction methods, while ignoring data from cycles close to baseline. Primers were designed by using the PrimerExpress software (Applied Biosystems, Foster City, CA, USA). Relative ratios were normalized to the Ct values of the metallopeptidase 1 gene as an internal control - in the case of primary and metastatic melanoma samples confirming improved DOP-PCR technique -, and calculated with the Pfaffl method (149). The sequences of PCR primers used in this study were deposited to http://160.114.60.33/Data1/. All the PCRs were performed three times in separate runs.

26

6.3.3. Confirmation of CGH–microarray results

The confirmatory QRT-PCR was performed on a RotorGene 3000 instrument with

previously described protocol (detailed in the 6.3.2. subsection of this thesis). Relative ratios

were normalized to the Ct values obtained with the dihydrofolate reductase gene probe in the

case of MCC and alpha1-fetoprotein transcript factor and H3 histone (family 3A) in the case

of PTCs and calculated with the Pfaffl method (149). PCR primers used are listed in Table 2.

All the PCRs were performed 4 times in separate runs. Table 2. Sequences of the oligonucleotides (5’-3’) used in the case of MCC, tonsillar tumour and thyroid

carcinomas

Gene (Accession No.) Chrom.

location Forward primer Reverse primer Product size (bp)

Steroid 5-alpha-reductase 2 (M74047 ) 2p23 CAGAAGCCCCAAGCAACTTT CCTTCTTGAACAGGTCCTGAGAA 69

Hepatocyte nuclear factor-3 alpha (U39840) 14q12-q13 CTCAAGAGTTGCTTGACCGAAA GGGCCATCTGTGGGTAGAGA 67

Zinc finger protein (AF020591) 19q13.43 GAAATTTCCCTGCCAGACCTT CATGAAGCCCCACTCTGAGAGT 73Androgen receptor (dihydro-testosterone receptor) (M23263)

Xq12 CGGAAATGATGGCAGAGATCA TGGGCTTGACTTTCCCAGAA 66

C8FW phosphoprotein (AJ000480) 8q24.13 GGACGATACCCCTTCCATGA GTCCACGCCGAATTTTGG 61Cardiac myosin binding protein-C (U91629) 11p11.2 GCCTAAATCCGAGCATCTGTTT TGCACTCTCAGGGAATTTGAGA 70EST (N22765) 15q14 CCTGATTCCAGAAAAGCAAGTGT CACTGAGATTACCGGGCATGA 76Cytochrome P450, subfamily XIA (M14565) 15q22 CTGGTGCAAGTGGCCATCTA AATTTTCCGGGTCGAAGAAGA 64EST sim. to phosphatidylcholin transfer protein (AA933627) 17q23 CACAAGGCTATGCACAAAGCA GGAAACTGAGGCGTCAAGATG 68

Dihydrofolate reductase (BC070280) 5q14 CTGTCATGGTTGGTTCGCTAAA TGCCGATGCCCATGTTC 60

Tre-2 (X63547) 17p13 TCCAGCGGCCCATTTG AGGGCGTGGAAAAACGAGAT 57Thymosin beta 10 (BC016731.1) 5q31.3 AATCGCCAGCTTCGATAAGG GGTCGGCAGGGTGTTCTTC 67Image EST (R78712) 15q25 GGTCATGTGTCGTGAAATATTATTGTT CCCGGCTGAGATTTTACATTTT 76Euk. Initiation factor 4EBP3 (AF038869) 5q31.3 ACCTGGCATGTGGAGTTACAGA TGGATGCCCCAGGAAGAG 70KIAA0549 (AB011121) 2q33 TTGCCTGGACAGTTACAGTTTCC AATCGGACAATTTAACGTTGTACTACTC 87Fibroblast growth factor 7 (M60828) 15q15-q21.1 GAACAAAATTTCTAATGCTGCTCAAG CATCAATCACTGTTGCTATCTTATATACAAG 93Alpha 1-fetoprotein transcription factor (U93553) 1q32.1 GGTGTCCAGGAACAAGTCAATG CTCTGTCTGCTGCGGGTAGTT 71

H3 histone, family 3A (BC081561) 1q41 TGCAGGAGGCAAGTGAGGC CTGGATGTCTTTTGGCATAATTGTT 101

Merkel cell carcinoma and tonsillar tumour

Thyroid carcinomas

6.4. Preparation of microarray probes

6.4.1. Preparation of genomic DNA probes for array hybridization of MCC

100 ng purified DNA was labelled with another round of PCR in the presence of Cy3-

dUCTP (0.05 mM) in a total volume of 50 µl with the same parameters for 20 cycles as in the

second PCR. In all these reactions, UN primer (5’-CCGACTCGAGNNNNNNATGTGG-3')

was used in 1 µM concentration. The reactions were performed with ExTaq DNA polymerase

27

in 1X buffer (Takara, Tokyo, Japan). The labelled PCR products were purified with the PCR

purification kit (Bioneer, Daejeon, Korea). The eluted DNA was dried in a Speed-Vacuum.

6.4.2. Preparation of genomic DNA probes for array hybridization of PTC samples Three microgramms of amplified genomic DNA was fragmented with AluI restriction

endonuclease, then it was tailed with dTTP resulting an oligo dT sequence at the 3’-ends of

the products. Afterwards, a special capture oligo was ligated to the tailed DNA according to

Genisphere DNA labeling system (Genisphere, Hatfield, PA).

6.4.3. Generation of microarray probes for human melanoma cells

Four microgramms of total RNA was reverse transcribed using poly-dT primed

Genisphere Expression Array 350 Detection system (Genisphrere, Hatfield, PA) in 20 µl total

volume using 20 Unit RNAsin (Fermentas), 1x first strand buffer and 200 Units of RNAse H

(-) point mutant M-MLV reverse transcriptase (Fermentas). All the other probe preparation

steps were done according the manufacturer’s instruction (Genisphere).

6.5. Microarray construction and array hybridization

6.5.1. Human cDNA microarray construction

Construction and use of microarrays were performed as previously described (43,

150). Briefly, 3200 cDNA inserts from human cDNA libraries (melanoma, lymphocytes,

heart and mixed tissue libraries) were amplified and purified with MultiScreen-PCR plate

(Millipore), resuspended in 50% dimethylsulfoxide/water and arrayed on FMB cDNA slides

(Full Moon BioSystems, Sunnyvale, CA) by using a MicroGrid Total Array System

(BioRobotics, Cambridge, UK) spotter with 16 pins in a 4 x 4 format. DNA elements were

deposited in duplicate. The diameter of each spot was approximately 200 µm. After printing,

DNA was UV crosslinked to the slides (Stratagene, Stratalinker, 700 mJ).

6.5.2. CGH-array hybridization in the case of MCC

The labelled DNA was reconstituted in ChipHybe hybridization buffer (Ventana Discovery, Tuchon, USA) containing 20 µg human Cot DNA and 5 µg salmon sperm DNA (Invitrogen). Hybridization was performed on human cDNA microarrays having 3200 gene-specific samples in duplicate in 200 µl by using the Ventana hybridization station (Ventana)

28

at 42 oC for 8 h. (151, 152). After hybridization, the slides were washed twice in 0.2X SSC at RT for 10 min, and then dried.

6.5.3. Array hybridization in the case of thyroid tumours and human melanoma cells

cDNA was hybridized onto human cDNA microarrays in a Ventana hybridization station (Ventana Discovery, Tuchon, State) by using the ‘‘antibody’’ protocol. First hybridization was performed at 42°C for 6 hr in ‘‘Chiphybe’’ hybridization buffer (Ventana) and then 2.5 µl of Cy5 capture reagents (thyroid tumours), while in other experiments 2.5 µl of each Cy5 and Cy3 capture reagents (MCC investigation) were added to the slides in 200 µl Chiphybe hybridization buffer (Ventana) and incubated at 42°C for 2 hr. After hybridization, the slides were washed in 0.2 x SSC twice at RT for 10 min and then dried and scanned. In the case of thyroid tumours the probe and control samples were hybridized onto separate slides, while they were hybridized onto the same cDNA microarrays in the case of MCC analysis.

6.6. Array data analysis

The presented data were calculated from the results of 4 data points obtained from two separate labelling and hybridization protocols. The slides were scanned with confocal laser scanner (ScanArray Lite, GSI Lumonics, Billerica, MA, USA). Image files were analysed with the GenePix Pro 3.0.5. program (Axon, Union City, CA, USA). The background corrected intensity data was filtered for flagged spots and weak signal. After automatic flagging, manual flagging was performed to exclude spots having irregularities, such as scratches and dust particles. Technical replicates on the same array were averaged. Data were excluded in cases where technical replicates were significantly different. Data obtained by the median of feature pixels were accepted only if 30 % of pixels were above 2 x standard deviation of the background intensity. Normalization was performed using the print-tip LOWESS method (153). Next we used the one-sample t-test in order to determine the genes to be regarded as changed in copy number. Logarithm was taken from each ratio to fulfill the t-test’s requirement for a normal distribution. Genes for which the mean of log-ratios across the biological replicates was equal to zero at a significance level α=0.05 are considered to have an unchanged copy number. On the other hand, genes having a p-value smaller than α and the average-fold change (over- or under-representation) of the four data points were at least 2.0-fold were considered as changes in DNA copy number or gene expression.

29

VII. RESULTS AND DISCUSSION

7.1. Improved DOP-PCR-based representational whole genome amplification using

QRT-PCR

The amount of genomic DNA available for genetic studies can often be limiting. DOP-

PCR is an appropriate method for overcoming these limitations by efficiently performing

whole genome amplification. We have presented a more reliable PCR-based genome

amplification procedure, developed as an alternative sample amplification method which

follows the amplification status by QRT-PCR. With real-time detection of the products, DNA

over-amplification could be diminished. After initial amplification of the genome, SybrGreen

dye is added to the reaction mix and a second amplification step is applied in a QRT-PCR

instrument. The cycling is performed until the reaction reaches the plateau (Figure 3).

Figure 3. Whole genomic DNA amplification using

QRT-PCR. Diagram of the primary, metastatic

melanoma tissue and normal peripheric lymphocyte. The

amplification halted at the 13th cycle, the amplified

genomic DNA (2) was generated in the exponential

phase of the reaction; the over-amplified genomic DNA

(1) was isolated from reactions halted at the 21st cycle;

Number 3 denotes for the non-template control.

To compare the original gene copy number of different templates generated by using the

traditional DOP-PCR and with the protocol using QRT-PCR, 44 genes were selected for

confirmation by using QRT-PCR analysis in case of two melanoma samples. Genes were

selected on the basis of metaphase CGH data performed earlier (154). The QRT-PCR was

performed on DNA samples obtained from primary and metastatic melanoma tissues resected

from the same patient along with normal DNA prepared from peripheral blood of a healthy

individual as a control. Altogether 88 relative copy numbers were determined on three

30

Table 3. Comparison of copy number changes of several gene loci obtained from primary and metastatic tumour samples relative to normal tissue by using non-amplified and different amplified genomic DNA as templates. Light gray: altered relative copy number at certain chromosomal regions in the amplified sample compared to control, non-amplified sample.

Chrom. Location Gene product (Acession No.)Loss of chromosomal regions

primary 0,43 (0,21) 0,27 (0,28) 2,20 (0,39)metastatic 0,44 (0,12) 0,27 (0,03) 0,31 (0,13)primary 0,50 (0,07) 0,39 (0,03) 0,31 (0,05)metastatic 0,48 (0,01) 0,33 (0,03) 0,48 (0,07)primary 0,50 (0,08) 0,50 (0,06) 1,77 (0,10)metastatic 0,51 (0,02) 0,44 (0,12) 1,44 (0,00)

2p23 Anaplastic lymphoma kinase (Ki-1) (NM_004304) metastatic 0,57 (0,03) 0,56 (0,28) 0,28 (0,13) deletion10q26.13-q26.3 dedicator of cytokinesis 1 (NM_001380) primary 0,58 (0,34) 0,52 (0,06) 0,74 (0,05) deletion

2p23 Anaplastic lymphoma kinase (Ki-1) (NM_004304) primary 1,11 (0,07) 0,68 (0,03) 2,77 (0,95) no changeprimary 1,16 (0,09) 0,50 (0,17) 0,74 (0,16) no changemetastatic 0,60 (0,01) 0,40 (0,09) 0,54 (0,06) deletionprimary 0,87 (0,04) 0,52 (0,03) 1,47 (0,05) no changemetastatic 0,64 (0,06) 0,59 (0,02) 1,25 (0,07) deletionprimary 0,83 (0,02) 0,53 (0,22) 1,63 (0,93)metastatic 0,85 (0,16) 0,58 (0,25) 0,69 (0,29)primary 0,89 (0,06) 0,92 (0,03) 0,90 (0,12)metastatic 0,82 (0,31) 1,13 (0,08) 0,72 (0,06)primary 0,76 (0,04) 0,86 (0,38) 0,90 (0,12)metastatic 0,90 (0,07) 0,82 (0,32) 0,75 (0,10)primary 1,07 (0,11) 0,69 (0,14) 0,48 (0,12)metastatic 1,33 (0,18) 0,90 (0,36) 0,49 (0,14)primary 1,01 (0,10) 0,49 (0,16) 0,66 (0,25)metastatic 1,17 (0,32) 0,59 (0,65) 0,65 (0,15)primary 0,86 (0,01) 1,10 (0,02) 1,27 (0,23)metastatic 0,77 (0,07) 1,03 (0,2) 0,57 (0,05)primary 1,03 (0,11) 0,99 (0,16) 0,99 (0,31)metastatic 1,14 (0,17) 1,11 (0,06) 1,24 (1,03)primary 1,20 (0,19) 0,97 (0,16) 0,48 (0,06)metastatic 1,18 (0,20) 0,91 (0,05) 0,35 (0,30)primary 1,05 (0,23) 0,82 (0,07) 0,87 (0,06)metastatic 1,31 (0,04) 1,10 (0,03) 0,41 (0,06)primary 1,18 (0,12) 0,51 (0,02) 1,01 (0,55)metastatic 1,03 (0,10) 0,56 (0,01) 0,44 (0,06)primary 1,19 (0,06) 0,80 (0,03) 1,22 (0,15)metastatic 1,31 (0,05) 1,11 (0,06) 0,93 (0,12)primary 1,24 (0,09) 0,82 (0,01) 1,34 (0,03)metastatic 1,43 (0,04) 1,12 (0,16) 0,99 (0,31)primary 1,15 (0,12) 0,71 (0,03) 1,22 (0,10)metastatic 1,14 (0,04) 0,89 (0,06) 1,17 (0,10)primary 0,90 (0,14) 0,35 (0,27) 0,62 (0,25)metastatic 0,98 (0,19) 0,60 (0,48) 0,80 (0,04)primary 1,23 (0,15) 1,10 (0,11) 1,17 (0,24)metastatic 1,30 (0,07) 1,05 (0,11) 1,14 (0,10)primary 1,03 (0,13) 0,84 (0,20) 0,67 (0,13)metastatic 1,14 (0,13) 0,97 (0,14) 0,63 (0,03)primary 1,35 (0,13) 1,31 (0,14) 2,27 (2,07)metastatic 1,22 (0,10) 1,42 (0,05) 2,16 (1,91)primary 0,89 (0,12) 0,68 (0,11) 0,91 (0,02)metastatic 0,82 (0,10) 1,02 (0,05) 0,48 (0,06)primary 1,28 (0,04) 2,00 (0,17) 0,82 (0,11)metastatic 1,34 (0,12) 1,73 (0,03) 0,78 (0,15)primary 1,24 (0,06) 1,12 (0,10) 0,99 (0,41)metastatic 1,11 (0,04) 0,93 (0,18) 0,45 (0,06)primary 1,05 (0,18) 0,67 (0,01) 0,50 (0,06)metastatic 1,03 (0,11) 0,83 (0,07) 0,51 (0,02)primary 1,10 (0,37) 1,37 (0,09) 1,41 (0,11)metastatic 1,11 (0,22) 1,31 (0,05) 0,75 (0,24)primary 1,46 (0,12) 0,68 (0,10) 0,50 (0,01)metastatic 1,41 (0,04) 0,87 (0,08) 0,55 (0,08)primary 1,22 (0,09) 0,76 (0,03) 0,68 (0,25)metastatic 1,22 (0,17) 0,91 (0,03) 0,65 (0,05)primary 0,96 (0,00) 0,60 (0,08) 0,82 (0,11)metastatic 0,91 (0,25) 1,00 (0,00) 1,01 (0,05)primary 1,25 (0,11) 0,49 ( 0,02) 1,11 (0,34)metastatic 1,40 (0,12) 0,50 (0,07) 1,31 (0,03)primary 1,20 (0,06) 1,31 (0,21) 0,80 (0,11) no changemetastatic 1,13 (0,07) 0,97 (0,24) 2,89 (0,12) amplification

7p15 CG1 (U97198) primary 1,14 (0,13) 0,79 (0,05) 1,71 (0,30) no change4p14 Amyloid beta precursor prot. binding B2 (BC027946) primary 1,03 (0,11) 1,49 (0,07) 0,55 (0,11) amplification3q26.31 N-acetylated a-linked acidic dipeptidase 2 (AF040990) primary 1,34 (0,09) 1,05 (0,17) 1,65 (0,40) no change

7p15 CG1 (U97198) metastatic 1,52 (0,08) 0,92 (0,00) 2,63 (0,78) no change4p14 Amyloid beta precursor protein-binding B2 (BC027946) metastatic 1,50 (0,13) 1,51 (0,08) 0,81 (0,03) amplification10q26.13-q26.3 dedicator of cytokinesis 1 (NM_001380) metastatic 1,52 (0,07) 0,43 (0,00) 0,86 (0,24) deletion3q26.31 N-acetylated a-linked acidic dipeptidase 2 (AF040990) metastatic 1,56 (0,06) 1,67 (0,18) 1,09 (0,15) amplification

primary 1,80 (0,21) 0,80 (0,04) 1,14 (0,19)metastatic 2,08 (0,20) 0,86 (0,09) 1,48 (0,02)primary 1,71 (0,05) 1,60 (0,32) 1,23 (0,16)metastatic 1,58 (0,05) 1,69 (0,31) 2,33 (0,28)primary 1,56 (0,15) 1,91 (0,03) 1,74 (0,03)metastatic 1,54 (0,12) 2,48 (0,65) 0,71 (0,38)primary 1,61 (0,21) 1,64 (0,17) 1,09 (0,10)metastatic 1,62 (0,21) 1,72 (0,62) 0,85 (0,07)primary 3,12 (1,10) 1,61 (0,05) 1,69 (0,22)metastatic 2,02 (0,40) 1,49 (0,12) 0,95 (0,20)primary 1,69 (0,06) 1,74 (0,04) 4,23 (1,98)metastatic 2,08 (0,22) 2,14 (0,06) 0,62 (0,02)primary 1,75 (0,03) 1,65 (0,02) 4,77 (2,47)metastatic 2,64 (1,08) 2,46 (0,67) 1,73 (0,56)

Metaphase CGHNon-amplified SD

DOP-PCR+ QRT-PCR SD

DOP-PCR SD

Roundabout, axon guidance receptor (BC057243) 3p12

Type of melanoma

2q31.1 Recepin (U03644)

15q25 Image EST R78712 (R78712)

2q32-33 amyotrophic lateral sclerosis 2 (AB011121)

14q31

no change

20q13.12 Image EST 364782 (AA034412) no change

5q21.3 F-box, leucine-rich repeat protein 17 (BC063316) amplification

6p12.2 Polycystic kidney, hepatic disease 1 (AY074797) amplification

17q11.2-q12 Amiloride-sensitive cation channel 1(BC075043) amplification

17q23

no change

no change

no change

no change

no change

no change

no change

no change

no change

no change

no change

no change

no change

no change

deletion

deletion

deletion

no change

deletion

no change

no change

no change

no change

no change

Galactocerebrosidase (D86181)

19q13.3-4 Electron-transfer-flavoprotein b (X71129)

1p33-34.1 Mitotic centromere-associated kinesin (U63743)

15q23-25 Neuromedin B precursor (M21551)

15q15-q21.1 Fibroblast growth factor 7 (M60828)

7q31 Met proto-oncogene (J02958)

RNA pol. transcriptional regulation mediator (U78082)

16p12-13.2 Epithelial amiloride-sensitive Na channel g (X87160)

Ribonuclease k6 precursor gene (U64998)

14q22.3-q23.1 Retinoblastoma-binding protein 1 (S66427)

14q23-24.1

5q11.2 Integrin, alpha 2 (M28249)

12q14.3

17p13

17p13.1

11p15.5

10p12.1

2p25.3

Chaperonin cont. t-complex polypept. 1b (AF026293)

Human clone 23665 (U90913)

15S-lipoxygenase (U78294)

p27-BWR1B (AF037066)

G protein-coupled receptor 158 (AY528411)

Myelin transcription factor 1-like (BC071612)

Genes with no change

9p23-p24.3 Prot. tyrosine phosphatase, receptor D (NM_130392)

Breast carcinoma amplified sequence 3 (BC001250) amplification

1p36.21 Cardiodilatin-atrial natriuretic factor (AL021155) amplification

6p25.2 Cell death protein (RIP) (U50062) amplification

UV-B repressed sequence, HUR 7 (X98307) no change

Gains of chromosomal regions

18q21.3-q22

Apolipoprotein A-I precursor (X02162)

CD48 antigen (B-cell membrane protein) (M37766)

1p31 Leptin receptor (U66497)

11q23

16p13.3. N-methylpurine-DNA glycosylase (M74905)

14q11.2-13.1

11q13.1-13.3 Cell cycle checkpoint control protein (U53174)

4p14 Ubiquitin carboxyl-terminal hydrolase L1 (X04741)

9q33.1 astrotactin 2 (BC010680)

1q23

amplification

19q13.3 OTK18 (D50419) no change

no change

2q35 Fibronectin 1 (X02761)

31

different templates: non-amplified as control, amplified with DOP-PCR and ampified with

our protocol: DOP-PCR in combination with QRT-PCR (Figure 3). In traditional DOP-PCR

protocol we applied 21 cycles for each reaction, however genomic DNA was PCR amplified

with our combined DOP-PCR-QRT-PCR method in 13-15 cycles, which were sufficient to

reach early saturation phase. The copy numbers of selected genes were normalised to the copy

number of the metallopeptidase 1 gene. This single copy gene did not show any alteration in

copy number in either melanoma sample and gave reproducible results in all cases. Relative

copy number change resulted being less than 0.5 meant loss of genetic material (heterozygous

loss). If this value was greater than 1.5, it was defined as gain or gene amplification at that

certain chromosome location (duplication of one allele). The results of the confirmatory QRT-

PCR are shown in Table 3.

Genomic changes (no change, loss or gain of copy number) were determined for all

the 44 genes in each method on both samples (primary melanoma and metastatic tumour) and

all the 88 values were compared to each other. In case of the QRT-PCR protocol 88.6% of the

copy number changes matched the non-amplified samples, while traditional DOP-PCR

resulted in only 63.6% accuracy. The correlation coefficient for the comparison of the non-

amplified and the DOP-PCR amplified genomic DNA was only 0.27, while this value was

significantly increased to 0.65 when the same protocol was used, but the amplification status

was followed by QRT-PCR.

To demonstrate the differences in relative copy numbers, which were obtained with

different amplification methods, a scatter plot was prepared (Figure 4). In case of our

combined method most of the genes had relative copy numbers similar to those of the non-

amplified method, while traditional DOP-PCR distorted the relative copy numbers detected in

the non-amplified samples.

To strengthen the extension of the preliminary observations a larger number of tumour

types were further analysed. Amplified samples from epidermoid, bronchiolo-alveolar lung

carcinoma, two renal cell carcinoma, two colorectal carcinoma, Conn and Cushing adenoma

samples were processed as described above. Copy number changes of eight different

chromosome regions were followed by QRT-PCR. Non-amplified and amplified samples

obtained from traditional DOP-PCR technique and from our improved protocol were

analysed. The results of copy number changes can be seen at http://160.114.60.33/Data2/.

32

Figure 4. Scatter plot of relative copy number of genes in primary and metastatic melanoma samples obtained with three different protocols. Values of copy number changes are in log2 scale. Only those genes were selected of which the values by any of the protocols were above +0.58 and below -1. The black squares and triangles mean those values which are different from the relative copy numbers of non-amplified samples.

In the case of the analysed eight tumour samples there was 84.4% concordance in

copy number alterations between the QRT-PCR protocol and the non-amplified samples,

while this value was only 64.0% for the traditional DOP-PCR (Figure 5).

Figure 5. Comparison of relative copy number of amplified samples genes to the original copy number

( : change in original copy number of genes : no change in original copy number of genes)

-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50

0 10 20 30 40Gene samples

Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s

Non-amplified DOP-PCR+QRT-PCR DOP-PCR

-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


-2,00

-1,50

-1,00

-0,50

0,00

0,50

1,00

1,50

2,00

2,50


Rel

ativ

eco

pynu

mbe

rofc

hrom

osom

alre

gion

s


0

20

40

60

80

100

DOP-PCR +QRT-PCR

TraditionalDOP-PCR

85.6%63.8%

14.4%36.2%

0

20

40

60

80

100

DOP-PCR +QRT-PCR

TraditionalDOP-PCR

85.6%63.8%

14.4%36.2%

33

In conclusion, our method improved the traditional DOP-PCR technique for whole

genome amplification. Improved reliability of the QRT-PCR-controlled DOP-PCR was

confirmed by the analysis of 152 copy number changes in ten tumour samples. We assume

that one of the most distorting effects of the DOP-PCR technique in the study of amplified

genomic DNA is the over-amplification of the products. Over-amplification has a dramatic

distorting effect on the original genomic rearrangements and can generate false result. This

effect can be decreased by adding QRT-PCR to the original protocol. Since DOP-PCR is a

widely used technique and the preservation of the original genome set is especially important

in genomic studies, therefore we propose to include QRT-PCR to follow and control the

amplification status.

7.2. A second field metachronous MCC of the lip and the palatine tonsil

MCCs are highly divergent when analysed for chromosome aberrations (84, 86, 87).

Reported recurrent chromosomal imbalances detected by CGH analysis were loss of 3p, 10q,

13q and 17p and gains of 1q, 3q, 5p and 8q (87). In the present study, we could also detect the

loss of 3p and 17p, but none of the above-mentioned amplified regions could be confirmed

(Table 4, Figure 6).

Table 4. Common and individual gains and losses of MCC and the tonsillar tumour detected by CGH.

Deletion Amplification

both tumour

1p36, 1p31, 1p13, 1q21-23, 1q32, 1q42.13, 2p13, 2p11.2, 2q35, 3p21, 3q13.2, 3q14.3, 3q26-28, 4p16, 4q21, 5q35,

6p21, 6q24, 7q21, 8p21, 8q24, 9p12, 9q34, 11p11.2, 11q13, 11q23.3, 12q24.1, 14q11.2, 15q23-24, 16p13-12,

16q24, 17p13, 17q12, 17q23, 18q12, 18q21, 19p13, 20p13, 21q22, 22q13.1, X monosomy

2p23, 10p15

Merkel cell carcinoma 11p11.2, 11q12.3, 12q13, 16p13.1, 17q21.1, 17q25 2p25, 12q12,

15q14

mesopharynx1p34.1-32.2, 1p21-22, 1q42, 2q37, 4p16.3, 4q28.3, 5q11.1, 5q22, 5q31-32, 7p14, 8p22-24.13, 10q21.1,

11pter-p15.5, 12q14.3, 14q11.2, 14q31-32, 15q11-13, 16q23-24, 17q21.3-q23, 20q13.1, 21q22.13, 22q11

__

In another study, only 3 of the 10 MCCs exhibited common gains and losses and they

shared a gain of 8q21-q22 and a loss of 4p15-pter (86). In the present study, the 4p16 region

was also found to be deleted in both the primary and secondary tumours. Unfortunately, to

date no known tumour suppressor gene has been mapped to this region (Table 5). Speleman’s

34

group detected losses involving the entire chromosome 10 or partial loss of the chromosome

10 long arm in one-third of the MCC cases they analysed with a possible loss of

heterozygosity (LOH) of 10q23, including the PTEN tumour-suppressor gene (85). In our

case, deletion of this region was not observed in either tumour sample.

Figure 6. Copy number aberrations found using microarray CGH analysis and real-time polymerase chain reaction of primary MCC and the tonsillar cancer. Boxes on the left side of each chromosome ideogram show regions of reduced copy number (losses of DNA in the tumour genome). Circles on the left side of each chromosome ideogram show regions of increased copy number (gains of DNA in the tumour genome). Filled boxes denote MCC and empty boxes tonsillar tumour

In previous observations, series of MCCs showed no evidence of high-level

amplification (87). Recurrent gains of chromosomes 1, 6, 18q and 20 were detected in 2 cases

(84). In our case, only 2 regions, 2p and 10p were commonly over-represented. In a very

recent study, 19 primary MCCs were analysed by CGH and in 13 samples extensive gains and

losses were detected (155). It was shown that a majority of the alterations were gains, while

only a few common losses were detected, mainly in regions 4q, 13q and 16q. Neither of our

gains was found in any of the 19 cases they analysed. Most of the losses were detected in at

least one case reported in the above-mentioned study, but the 13 MCCs exhibited a very

heterogenous pattern, with diverse regions with losses.

35

Our CGH results revealed several new and a few other previously known

chromosomal regions that have been presumed to be involved in the pathogenesis of MCC.

We found 2 common gains and 41 common losses in the 2 samples. However, in 31

chormosome locations we observed differences in gene copy numbers between the 2 tumours.

From the results obtained by CGH analysis, we believe that the mesopharynx cancer and the

MCC of the lip derived from the same, genetically altered field, thus the mesopharynx cancer

can be regarded as a second field tumour. From the results obtained by CGH analysis, we

hypothesize that Merkel cells in two adjacent anatomical sites, i.e. in the lip and

mesopharynx, underwent same precancerous genetic alteration, and both tumours arose from

a common cell clone. If our hypothesis is correct, the mesopharynx cancer can be regarded as

a second field tumour.

Table 5. Apoptotic genes and tumour supressors mapped to regions with common chromosomal aberrations. (Apoptotic and tumour suppressor genes in bold character. Apoptotic and tumour suppressor related genes in italics in brackets.)

Several oncogenes and tumour-suppressor and apoptotic genes were assumed to be

changed in their copy number, and many of them were mapped to the regions having changes

in their copy number in one or both of the tumours (Tables 5 and 6).

11p11.2 (MADD)

11q13 DPF2, BAD, FADD

(RELA, CCND1, RAD9A)

DOC-1R, MEN1 (ST3, CCND1, SHANK2)

11q23.3 THY1 12q24.1 (SCA2)

16p13-12 ASC TSC2

16q24 WFDC1, GAS8, MGC15419, CBFA2T3

17p13 TP53, HIC1, OVCA2, DPH2L1

17q12 (CCR7) (MPP3)

18q21 (BCL2) DCC, SERPINB5 (E2F5)

19p13 DIRAS1 (TJP3, SMARCA4, GIPC3)

21q22 (MX1) 22q13.1 (HMOX1) ST13 Xp22.1 (SH3KBP1) Xq12 ING2 Xq28 RPL10

1p36 DFFB, TP73, CASP9 (CDC42, MFN2)

UBE4B, TUSC3, PRDM2, C1orf1, ALPL

(E2F2, TP73)

1p31 CLCA2 1p13 ST7L

1q21-23 MCL1, DAP3,

TNFSF6 (MUC1, APCS, CRP, SELL)

2p13 MAD

3p21.3 (CCR3) RASSF1, BAP1 (SEMA3B)

3q26-28 (ETV5, OPA1) TSBF1 4p16 (GAK) 4q21 (SNCA) 5q35 (PDLIM7) 6p21 BAK1, TNF (HSPA) (CDKN1A) 6q24 SASH1 (PLAGL1)

8p21 BNIP3 (CLU) PDGFRL (CNOT7, PDLIM2)

8q24 (RNF139) 9q34 TRAF2 (ABL1)

Chromosome location Apoptotic genes Tumor supressors Chromosome

location Apoptotic genes Tumor supressors

In both tumor11p11.2 (MADD)




11q23.3 THY1 12q24.1 (SCA2)

16p13-12 ASC TSC2



17q12 (CCR7) (MPP3)






(E2F2, TP73)


1q21-23 MCL1, DAP3,


2p13 MAD




8q24 (RNF139) 9q34 TRAF2 (ABL1)


location Apoptotic genes Tumor supressors Chromosome location Apoptotic genes Tumor supressors Chromosome


In both tumorIn both tumour

Tumour suppressors Tumour suppressors

11p11.2 (MADD)




11q23.3 THY1 12q24.1 (SCA2)

16p13-12 ASC TSC2



17q12 (CCR7) (MPP3)






(E2F2, TP73)


1q21-23 MCL1, DAP3,


2p13 MAD




8q24 (RNF139) 9q34 TRAF2 (ABL1)



In both tumor11p11.2 (MADD)




11q23.3 THY1 12q24.1 (SCA2)

16p13-12 ASC TSC2



17q12 (CCR7) (MPP3)






(E2F2, TP73)


1q21-23 MCL1, DAP3,


2p13 MAD




8q24 (RNF139) 9q34 TRAF2 (ABL1)


location Apoptotic genes Tumor supressors Chromosome location Apoptotic genes Tumor supressors Chromosome


In both tumorIn both tumour

Tumour suppressors Tumour suppressors

36

Table 6. Apoptotic genes and tumour supressors mapped to regions with chromosomal aberrations specific to MCC or tonsillar tumour.

The limitation of this list is that in these cases the deletions of the genes themselves

were not proven in this study; only those regions were determined which were located to

certain regions or proximity to regions where DNA segments on the microarrays originated.

The micorarray-based CGH techniques used in this study could result in distorted data,

especially when paraffin-embedded tissue is the target. Another limitation of this study is that

the deleted regions were determined by using hybridizations based on complementary cDNA

- genomic DNA annealing events, which could generate more biases. As a consequence of the

above problems, we determined the reliability of the results obtained by the CGH microarray

by using QRT-PCR. 10 genes were selected to follow the copy number changes by QRT-

PCR. Out of the 10 genes, one exhibited no changes in the copy number found by CGH

microarray analysis in all cases: the dihydrofolate reductase gene. This was therefore used an

internal control. The copy numbers of all the other 9 genes (sequences and chromosome

locations are listed in Table 2, accesion numbers are listed in Table 7) were related to this

inner control. We used DNA purified from normal lymphoid tissue as control and determined

copy number changes of the other two tumours (Table 4, Figure 6).

We confirmed the deletions in 7 cases for both tumours and 2 were confirmed for only

MCC. Although the number of the confirmatory QRT-PCRs were limited and therefore does

not allow determination of the reliability of the CGH methods, the main purpose of the study

Chromosome location Apoptotic genes Tumor supressors

In Merkel cell carcinoma 11p11.2 (MADD)

12q13 (CDK2, KRT18, NR4A1, WNT1)

17q25 BIRC5 In Mesopharynx

1p34.1-32.2 FABP3 3p22-21 ADPRTL3 5q11.1 (LOC145908) 5q22 MCC

5q31-32 C5orf7 10q21.1 (CDC2)

11pter-p15.5 (CTSD, HRAS) MRVI1 14q31-32 SIVA 16q23-24 (LOC145908) CBFA2T3 21q22.13 RUNX1

22q11 BID MYO18B

Tumour suppressors Chromosome location Apoptotic genes Tumor supressors

In Merkel cell carcinoma 11p11.2 (MADD)

12q13 (CDK2, KRT18, NR4A1, WNT1)

17q25 BIRC5 In Mesopharynx

1p34.1-32.2 FABP3 3p22-21 ADPRTL3 5q11.1 (LOC145908) 5q22 MCC

5q31-32 C5orf7 10q21.1 (CDC2)

11pter-p15.5 (CTSD, HRAS) MRVI1 14q31-32 SIVA 16q23-24 (LOC145908) CBFA2T3 21q22.13 RUNX1

22q11 BID MYO18B

Tumour suppressors

37

was to determine the relation of the two tumours, not a precise determination of the deleted

regions.

Table 7. Confirmation of CGH data by QRT-PCR. Only two common losses could not be confirmed. Gains are indicated as grey background.

Microarray data

confirmed by QPCR Clone name Accesion No. Chrom. location Distance

Steroid 5-alpha-reductase 2 (SRD5A2) M74047 2p23 31724 Hepatocyte nuclear factor-3 alpha U39840 14q12-q13 36052 Zinc finger protein AF020591 19q13.43 63450

Androgen receptor (dihydrotestosterone receptor)

M23263 Xq12 63075

C8FW phosphoprotein AJ000480 8q24.13 126404 ESTs, Highly similar to Phosohatidylcholine transfer protein AA933627 17q23 54340

both tumour

Cytochrome P450, subfamily XIA M14565 15q23-24 70670 EST N22765 15q14 no data Merkel cell

carcinoma Cardiac myosin binding protein-C U91629 11p11.2 49310

We confirmed the deletions in 7 cases for both tumours and 2 were confirmed for only

MCC. Although the number of the confirmatory QRT-PCRs were limited and therefore does

not allow determine the reliability of the CGH methods, the main purpose of the study was to

determine the relation of the two tumours, not a precise determination of the deleted regions.

In summary, the partly different molecular patterns obtained by CGH, the similar

histological features, and the close proximity to the primary tumour indicate that the tonsillar

cancer was a second field tumour. The common origin was further confirmed, because out of

3 early markers (9p, 3p and 17p) two of them (9p and 17p) were common in both cancer

samples, one (17p) bearing the p53 tumour suppressor marker gene. The common copy

number changes do not support the possibility that the tonsillar cancer was a second primary

tumour. The tonsils are very rare and unusual sites of metastatic disease in MCC. There are

three reports in the literature describing oropharyngeal metastasis of a MCC (71, 156, 157). In

all these reports, the metastasis occurred within 24 months after resection of the primary

tumour. In our case, it was clinically very unlikely that the tonsillar tumour was a metastasis

of the MCC of the lip, because no local recurrence or regional lymph node metastasis or

distant haematogenous metastases were observed in the 7-year follow-up. In harmony with

38

the clinical situation, the molecular patterns of the 2 tumours were not similar, therefore the

possibility of a metastasis could be rejected. Although the concept of second primary tumours

and field cancerization has been formulated for oral and oropharyngeal squamous cell

carcinomas (78), our results indicate that Merkel cells in adjacent anatomic sites may also be

the targets of field cancerization.

7.3. Determination of several copy number changes in different genomic DNA of thyroid

carcinoma based on disease course

We obtained genomic DNA from paraffin-embedded samples of primary- and

terminal-stage APTC of a sixty-three-year-old male patient (by the latter, histology indicated

dedifferentiation), which genomic DNA was amplified by our improved DOP-PCR technique.

As a follow-up, relative DNA losses and gains were determined for each tumour sample by

normalizing the intensities to the values obtained after hybridization with labeled probes from

GPTC. The CGH-array analysis were performed on human cDNA microarray containing

3200 gene-specific cDNA samples in duplicate on glass slides.

Specific primers were designed for thirty chromosomal regions exhibiting changes

(deletion or amplification) in order to confirm our CGH results using QRT-PCR, which was

performed on the original, non-amplified genomic DNA. We found that 3 out of the 30

regions (FGF7, EIF4EBP3, KIAA0549) indicated the same changes (DNA losses) in both

tumours and corresponded to CGH-array results. In one case (Image EST-Acc. No. R78712),

the QRT-PCR results corresponded to the CGH-array data of primary APTC; and at further

examination, DNA gain was detected in the case of terminal-stage APTC when it was

compared to the copy number of normal thyroid tissue. These results demonstrated a

difference in chromosome copy numbers of primary and terminal-stage APTC.

Further on, two additional gene loci were identified with changes only within

terminal-stage APTC, namely Tre-2 oncogene and TB10, which were positive in each thyroid

carcinoma during immunohistochemical experiments – especially in the case of ATC (158).

Only one region – Tre-2 oncogene – out of six regions was confirmed as over-represented in

the case of GPTC.

Thirty two additional thyroid carcinomas examined as well. QRT-PCR analysis was

performed on the non-amplified genomic DNA of 32 PTCs as well as the first GPTC and the

first terminal-stage APTC, upon which the CGH-array analysis was carried out. The copy

39

number of the six regions described above (Table 8) was determined on 34 thyroid carcinoma

genomes altogether (10 GPTCs; 14 IPTCs, 10 APTCs); we compared it with two normal

thyroid samples, and the relative ratios were normalized to the copy numbers of the alpha1-

fetoprotein transcription factor and the H3 histone (family 3A) as internal control genes

separately.

Since we could not detect a significant difference in the examined six regions – in the

copy numbers of IPTC and APTC samples, these two groups were treated as one group

(I/APTC) (that is where the patient had a lymph node metastasis in the cervical areas or some

other distant metastasis /spread: lung, bone/, suffered from local recurrance, died because of

their PTC). Table 8. Copy number alterations in different disease course thyroid carcinoma. Dark gray denotes

DNA gains common amplification of TB10 and Image EST, while light gray denotes separately amplification.

Chrom. Loc.Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD)1.35 (0.20) N 3.28 (1.93) A 0.75 (0.07) N 0.58 (0.26) N 0.99 (0,10) N 0.39 (0.16) D1.31 (0.32) N 1.26 (0.32) N 2.62 (0.88) A 0.59 (0.19) N 0.86 (0.35) N 0.53 (0.15) D1.29 (0.17) N 0.46 (0.15) D 1.80 (0.31) A 0.54 (0.13) D 0,74 (0.18) N 0.31 (0.06) D1.12 (0.41) N 0.49 (0.25) D 2.63 (1.14) A 0.46 (0.18) D 0.52 (0.16) D 0.35 (0.10) D1.14 (0.30) N 0.98 (0.34) N 1.64 (0.28) A 1.10 (0.25) N 1.13 (0.24) N 0.69 (0.23) N0.95 (0.27) N 0.67 (0.26) N 2.74 (0.80) A 0.58 (0.22) N 0.45 (0.05) D 0.89 (0.39) N1.88 (0.13) A 1.27 (0.34) N 3.14 (0.95) A 0,00 D 1.11 (0.10) N 0.52 (0.08) D1.18 (0.35) N 0.64 (0.56) N 2.24 (0.83) A 0.48 (0.04) D 0.91 (0.31) N 0.40 (0.35) D0.70 (0.22) N 0.89 (0.22) N 0.69 (0.64) N 0,09 (0.07) D 1.65 (0.27) A 0.18 (0.07) D2.96 (1.18) A 0.81 (0.29) N 2.19 (0.12) A 0.64 (0.36) N 0.97 (0.32) N 1.20 (0.43) N

2.95 (1.10) A 0.90 (0.32) N 1.28 (0.46) N 0.52 (0.26) D 0.49 (0.15) D 0.63 (0.15) N2.02 (0.60) A 0.70 (0.34) N 3.82 (0.30) A 0.36 (0.09) D 0,47 (0.23) D 0.64 (0.23) N1.79 (0.63) A 0.54 (0.28) D 3.99 (1.16) A 0.40 (0.21) D 0.49 (0.23) D 0.55 (0.26) D3.03 (0.99) A 2,36 (0.94) A 1.51 (0.37) A 1.33 (0.26) N 0.65 (0.16) N 0.79 (0.23) N1.13 (0.29) N 0.66 (0.18) N 3.93 (1.96) A 0.39 (0.09) D 0.40 (0.14) D 0.20 (0.06) D0.83 (0.13) N 3.64 (1.64) A 2.82 (0.17) A 1.06 (0.26) N 5.26 (0.54) A 1.14 (0.33) N0.92 (0.41) N 1.09 (0.41) N 1.26 (0.21) N 0.82 (0.25) N 1.03 (0.21) N 0.79 (0.26) N4.87 (1.08) A 2.38 (1.07) A 5.00 (1.51) A 1.08 (0.19) N 8.14 (0.84) A 0.85 (0.31) N4.27 (1.19) A 1.86 (0.97) A 2.33 (0.72) A 1.14 (0.35) N 1.16 (0.12) N 0.37 (0.10) D1.32 (0.29) N 1.30 (0.73) N 3.24 (0.84) A 2.03 (0.54) A 1.18 (0.28) N 0.73 (0.17) N1.67 (0.34) A 2.25 (1.27) A 1.73 (0.43) A 3.39 (1.04) A 1.90 (0.39) A 1.10 (0.29) N1.02 (0.13) N 1.12 (0.46) N 1.97 (0.52) A 0.72 (0.17) N 1.18 (0.27) N 0.81 (0.16) N4.13 (1.85) A 0.67 (0.20) N 3.67 (0.50) A 0.25 (0.12) D 1.47 (0.18) N 0.42 (0.14) D2.00 (0.56) A 0.93 (0.29) N 3.92 (1.14) A 0,00 D 0.72 (0.07) N 0.58 (0.20) N

3.77 (1.25) A 0.53 (0.21) D 3.46 (1.59) A 0.65 (0.21) N 0.37 (0.22) D 0.69 (0.26) N4.16 (1.19) A 0.53 (0.28) D 2.29 (0.69) A 0.21 (0.11) D 0.66 (0.36) N 0.21 (0.06) D0.72 (0.13) N 0.68 (0.22) N 0.91 (0.18) N 0.38 (0.15) D 0.62 (0.19) N 0.51 (0.19) D3.27 (0.79) A 0.84 (0.40) N 2.95 (0.60) A 0.06 (0.04) D 0,.76 (0.30) N 0.28 (0.10) D1.50 (0.34) A 0.72 (0.19) N 1.15 (0.32) N 0.46 (0.20) D 0.83 (0.42) N 0.35 (0.11) D1.35 (0.34) N 0.53 (0.20) D 4.47 (1.65) A 0.58 (0.23) N 0.50 (0.18) D 0.70 (0.27) N2.07 (0.66) A 2.11 (0.66) A 1,15 (0.41) N 0.76 (0.15) N 0.85 (0.32) N 0.65 (0.20) N0.77 (0.18) N 7.34 (6.65) A 3.52 (0.90) A 0.71 (0.24) N 2.95 (1.84) A 0.78 (0.22) N0.80 (0.15) N 0.71 (0.31) N 1.75 (0.25) A 1.91 (0.52) A 0.97 (0.17) N 1.18 (0.35) N3.32 (2.52) A 2.06 (1.86) A 2.44 (2.10) A 0.26 (0.37) D 0.50 (0.43) D 0.62 (0.30) N

Fibroblast growth factor 7 EIF4EBP3Gene

5q31.3 15q25 17p13 2q33 15q15-q21.1 5q31.3

KIAA0549 protein

AP

TCs

('B' g

roup

)

Tre-2 oncogene

GP

TCs

('A' g

roup

)IP

TCs

('B' g

roup

)

Image EST Acc.No.R78712

Thymosin, beta 10

40

In the present study, Tre-2 oncogene was found to be over-represented in 80% of the

GPTC and I/APTC groups (Figure 7). Tre-2 is a recombinant gene isolated from NIH3T3

cells transfected with human Ewing’s sarcoma DNA. Genetic elements of Tre-2 originate

from chromosomes 5, 18 and 17. Our examined field of Tre-2 oncogene is mapped onto the

17p13 chromosome region. Many researchers have demonstrated that this oncogene is

consistently transcribed in various human cancer cells (159). Expression of Tre-2 in normal

somatic cells, however, has not yet been reported. Tre-2 oncogene seems to encode a

nonfunctional Rab (GAP). Regions flanking the TBC (Tre-2/Bub2/Cdc16) domain may be

crucial for catalytic activity. These domains are predicted to encode GTPase-activating

proteins (GAPs) for Rab family G proteins. Martinu et al. (160) have identified this oncogene

as a novel regulator of the Arf6-regulated plasma membrane recycling system. Masuda-

Robens et al. have demonstrated that Tre-2 coprecipitated specifically with the active forms of

Cdc42 and Rac1 in vivo, which play fundamental roles in transformation and actin

remodeling.

Figure 7. Copy number alterations of six chromosomal regions in two disease course thyroid carcinomas.

('A' group: PTC of good disease outcome; 'B' group: intermediate and aggressive type of PTC)

AmplificationNo changeDele tion




41

Our study showed amplified regions of TB10 in greater extent in I/APTC (62.5%),

whereas in the case of GPTC group the quantity of DNA gains could be detected in only 20%

of the cases. This is a significant difference between the two groups based on the outcome of

the disease. Chiappetta et al. (161) have shown that TB10 gene expression can function as a

possible tool in the diagnosis of thyroid neoplasias, since TB10 positive staining was found in

all human thyroid carcinoma, particularly in the anaplastic histotypes. Many researchers have

demonstrated that TB10 acts as an actin-mediated tumour suppressor (162). Some

observations suggest that TB10 plays a significant role in the cellular processes controlling

apoptosis (163). Other authors have shown that the TB10 gene is highly expressed in human

thyroid carcinoma cell lines and tissues, whereas its expression is almost undetectable in

normal thyroid; furthermore, TB10 gene expression can function as a useful marker for

diagnosis and prognosis of a large variety of human cancers (162). In a study by Takano et al.

(158), TB10 mRNA is highly over-expressed in human and experimental thyroid tumours and

its expression is abundant in undifferentiated thyroid carcinomas (158, 164). These results

seem to be closely related to our results in connection with genomic rearrangement. TB10

gene overexpression is a general event in human carcinogenesis (162). Overexpression of

TB10 was showed in thyroid, pancreatic, gastric, breast, ovarian, uterine, colon, esophageal

cancer and germ cell tumours. Lee et al. (165) have described that overexpressed TB10

significantly inhibited vascular endothelial growth factor-induced endothelial cell

proliferation, migration, invasion and tube formation in vitro.

The examined Image EST (Acc. No. R78712) is mapped to the middle of the AKAP13

gene. In the case of this region, our results indicated DNA gains in one case of the GPTC

group and in an ascending quantity in the I/APTC. We can postulate that the AKAP13 gene is

affected; however, further analysis is needed to provide this hypothesis. AKAP13 anchors

both protein kinase A and 14-3-3, and by this means inhibits the Rho-GEF activity of the

AKAP-Lbc signaling complex (166). Alternative splicing of this gene results in at least 3

transcript variants encoding different isoforms containing a dbl oncogene homology (DH)

domain and a pleckstrin homology (PH) domain. The DH domain is associated with guanine

nucleotide exchange activation for the Rho/Rac family of small GTP binding proteins,

resulting in the conversion of the inactive GTPase to the active form capable of transducing

signals.

42

In two regions we found DNA losses, in one case out of which – the EIF4EBP3 gene –

indicated a decreased copy number in many cases during our experiments. This protein is the

negative regulator of protein byosinthesis and the initiation of translation; in other words, it

represses translation. DNA losses were detected in 70% of the GPTC and only in 33% of

I/APTC. The KIAA0549 gene was under-represented in half of both groups. When examined,

the sixth gene, namely FGF7, demonstrated both DNA gains and losses in small quantity, but

this rearrangment was greater in the case of I/APTC. FGF family members possess broad

mitogenic and cell survival activities, and are involved in a variety of biological processes,

including embryonic development, cell growth, morphogenesis, tissue repair, tumour growth

and invasion. Unfortunately, we do not have knowledge on the role these rearrangements of

genes can play within the tumourigenesis progress.

Our results indicate that synchronous detection of over-representation of TB10 and

Image EST (Acc. No. R78712) can become a good diagnostic marker for aggressive thyroid

carcinoma.

7.4. QRT-PCR-based exponential amplification for microarray gene expression profiling

Global RNA expression analysis using microarray technology allows the identification

of genes that are differentially expressed in tumour versus normal tissues. In this case the

elimination of overamplified cDNA is also extremely important, for this reason we developed

an alternative exponential sample amplification method. Total RNA was reverse transcribed

and amplified with two PCR protocols resulting in DNA samples from early phase (13th–15th

cycles) and late exponential, early saturation phase (21st cycle), similarly to the proving of the

reliability of the improved DOP-PCR, which can amplify the whole genome. Reproducibility

and reliability of the methods were determined performing cDNA microarray and

confirmative QRT-PCR experiments on LPS-treated mouse macrophage cDNA prepared in a

different manner.

Our exponential sample amplification protocol preserves the original expression ratios

and allows unbiased gene expression analysis from minute amounts of starting material.

43

7.5. Analysis of the gene expression pattern of parallel expression of αIIbβ3 and αvβ3

integrins in human melanoma cells using microarray technology

In cancerous tissues new vessels are required to provide cancer cells with nutrients and

are targets for invading cancer cells themselves. At the beginning of its natural history, a

cancer is not vascularised, and it does not grow beyond 2 mm in size unless vascularisation

has occurred. The switch to the angiogenic phenotype is a critical point in tumour progression

(167) and depends on the additive effect of progressive genetic alterations (168).

The effect of αIIbβ3-transfection on gene expression was tested using a homemade

3200 cDNA microarray (43, 150). The fastest growing and most vascularized clone (19H)

was selected and compared to the mock cells, 3.1P to reveal the molecular pathomechanism

of large growth and cell survival. We defined those genes to be regulated by αIIbβ3

transfection in melanoma cell cultures, in the case of which the average-fold change (increase

or decrease) of the 4 data points was at least 2.0-fold.

Forty-eight genes in 19H cells corresponded to such strict criteria from the 3200

analysed; 36 genes were found to be upregulated and 12 have been downregulated. Among

the downregulated genes, there were no known angiogenic/endothelial ones represented;

however, among the 19 known genes upregulated in 19H cells compared to 3.1P, 3 could be

considered endothelial-specific, such as CD34 antigen, endothelin receptor type B and

prostaglandin I-2 synthase, the changes of which were also confirmed by QRT-PCR analysis

(data not shown).

Parallel expression of the two β3-integrins in human melanoma cells induced

modulation not only of bFGF but of other genes as well, which is demonstrated by microarray

analysis. Interestingly, 3 out of 19 known genes upregulated significantly in αIIbβ3-

transfected 19H cells are endothelial cell-specific genes, CD34, endothelin receptor B and

prostaglandin I-2 synthase. Recapitulation of an embryonic endothelial/angioblastic genetic

program, called vasculogenic mimicry (169), resulting in vivo formation of tumour cell lined

vascular channels, was recently described and documented in case of human melanoma and

other tumours (170). There have been several genes identified to be responsible for this

vasculogenic phenotype including VE-cadherin, CD34, TIE2, EphA2, LAMC2, endoglin,

EDG1, ESM1 and EDF1 (171).

44

VIII. CONCLUSIONS In many cases only a minute amount of partially degraded genomic DNA can be

extracted from archived clinical samples. Forensic analyses, tests on biopsy samples, prenatal

diagnosis and many other investigations have access to very little starting material.

Investigations of the genetic background of intratumour heterogenity require manipulation of

minute amounts of neoplastic tissue on a cellular scale, often below the amounts required for

routine CGH. A method to overcome the small quantity of the DNA available is to perform

whole genome amplification. One commonly used method is DOP-PCR, due to its simplicity

and degree of success. In our study a new QRT-PCR-based DOP-PCR amplification method

was developed, on the basis of which we have proved the preservation of the original copy

number of different chromosomal regions in amplified genomic DNA.

This improved genomic amplification technique was also applied in the case of two

clinical investigations, during which the molecular imbalances of tumours were detected by

microarray based CGH analysis. In the first case we analysed the MCC and the seven years

later appearing tonsillar tumour in order to assess the genetic relationship of the tumours. On

the basis of our results we have demonstrated that the tonsillar tumour can be pathogenetically

regarded as a second field tumour.

The improved DOP-PCR technique was also employed for detection of molecular

patterns of PTCs of different disease courses. The question arises as to whether genomic

differences (markers) are to be found at the appearance of the disease, or an unfavorable

outcome can be predicted that could allow a more effective treatment for the patient. Our

CGH-array and QRT-PCR results have shown rearrangements of six regions of the genome,

which can play a role in the tumourigenesis progress, and synchronous detection of some of

these genes can become a good diagnostic marker for the detection of malignant thyroid

carcinoma.

After reaching a certain tumour size, tumour cells need to be provided with a sufficient

amount of oxygen and nutriment. It was demonstrated that the transfection of integrin αIIbβ3

into human melanoma cells promoted their in vivo growth and cell survival. We have revealed

nineteen known genes upregulated in transfected melanoma cells; three could be considered

endothelial-specific by performing microarray experiments, which can provide explanation

for a part of underlying pathomechanism of increased cancer growth and angiogenesis.

45

IX. ACKNOWLEDGMENTS

I hereby thank Dr. László Puskás for the giving me the opportunity to work in his

group, and to express my thankfulness for introducing me the methods and techniques of

molecular biology and the ideas, help and instructions he provided during my work. I am

thankful for the staff of the Laboratory of Functional Genomics for the creative atmosphere of

the lab: I would like to thank Dr. Ágnes Zvara, Dr. László Hackler, Rozália Török, János

Zsigmond Kelemen, Dr. Zoltán Varga-Orvos, Laura Vass, Dr. Béla Iványi, Dr. István

Sonkodi, Dr. József Tímár, Dr. Erzsébet Rásó, Dr. Gábor Pocsay, Dr. László Krenács, Dr.

Zoltán Nemes, Dr. Judit Nagy, and Dr. Réka Pocsay for their help.

X. REFERENCES 1. Pardue ML, Gall JG. Molecular hybridization of radioactive DNA to the DNA of cytological preparations.

Proc Natl Acad Sci U S A. 1969;64:600-4. 2. John HA, Birnstiel ML, Jones KW. RNA-DNA hybrids at the cytological level. Nature. 1969; 223:582-7. 3. Schrock E, du Manoir S, Veldman T, et al. Multicolor spectral karyotyping of human chromosomes.

Science. 1996; 273:494-7. 4. Kallioniemi A, Kallioniemi OP, Sudar D, et al. Comparative genomic hybridization for molecular

cytogenetic analysis of solid tumours. Science. 1992; 258:818-21. 5. Abeln EC, van Kemenade FD, van Krieken JH, et al. Rapid identification of mixed up bladder biopsy

specimens using polymorphic microsatellite markers. Diagn. Mol. Pathol. 1995; 4:286–91. 6. Orita M, Iwahana H, Kanazawa H, et al. Detection of polymorphisms of human DNA by gel electrophoresis

as single-strand conformation polymorphisms. Proc Natl Acad Sci U S A. 1989; 86:2766-70. 7. Myers RM, Maniatis T, Lerman LS. Detection and localization of single base changes by denaturing

gradient gel electrophoresis. Methods Enzymol. 1987; 155:501-27. 8. Dracopoli, N.C. Ed. "Detection of Mutations by RNase Cleavage". Current Protocols in Human Genetics.

1998. John Wiley & Sons, Inc. 9. Roest PA, Roberts RG, Sugino S, et al. Protein truncation test (PTT) for rapid detection of translation-

terminating mutations. Hum Mol Genet. 1993; 2:1719-21. 10. Maxam AM, Gilbert W. A new method for sequencing DNA. 1977. Biotechnology. 1992; 24:99-103. 11. Frommer M, McDonald LE, Millar DS, et al. A genomic sequencing protocol that yields a positive display

of 5-methylcytosine residues in individual DNA strands. Proc Natl Acad Sci U S A. 1992; 89:1827-31. 12. Greider CW, Blackburn EH. Telomeres, telomerase and cancer. Sci. Am. 1996; 276:92–97. 13. Chuang SE, Daniels DL, Blattner FR. Global regulation of gene expression in Escherichia coli. J Bacteriol.

1993; 175:2026-36. 14. DeRisi JL, Iyer VR, Brown PO. Exploring the metabolic and genetic control of gene expression on a

genomic scale. Science. 1997; 278:680-6. 15. Nelson SF, McCusker JH, Sander MA, et al. Genomic mismatch scanning: a new approach to genetic

linkage mapping. Nat Genet. 1993; 4:11-8. 16. Haab BB, Dunham MJ, Brown PO. Protein microarrays for highly parallel detection and quantitation of

specific proteins and antibodies in complex solutions. Genome Biol 2001;. 2, RESEARCH0004. 17. Zhu H, Klemic JF, Chang S, Bertone P, et al. Analysis of yeast protein kinases using protein chips. Nat

Genet. 2000; 26:283-9. 18. Ziauddin J, Sabatini DM. Microarrays of cells expressing defined cDNAs. Nature. 2001; 411:107-10. 19. Kuruvilla FG, Shamji AF, Sternson SM, et al. Dissecting glucose signalling with diversity-oriented

synthesis and small-molecule microarrays. Nature. 2002; 416:653-7. 20. Schena M, Shalon D, Davis RW, et al. Quantitative monitoring of gene expression patterns with a

complementary DNA microarray. Science. 1995; 270:467-70. 21. Zvara A, Hackler LJr, Nagy ZB, et al. New molecular methods for classification, diagnosis and therapy

prediction of hematological malignancies. Pathol Oncol Res. 2002; 8:231-40.

46

22. Jenson SD, Robetorye RS, Bohling SD, et al. Validation of cDNA microarray gene expression data obtained from linearly amplified RNA. Mol Pathol. 2003; 56:307-12.

23. Saksela K, Stevens C, Rubinstein P, et al. Human immunodeficiency virus type 1 mRNA expression in peripheral blood cells predicts disease progression independently of the numbers of CD4+ lymphocytes. Proc Natl Acad Sci U S A. 1994; 91:1104-8.

24. Futscher BW, Blake LL, Gerlach JH, et al. Quantitative polymerase chain reaction analysis of mdr1 mRNA in multiple myeloma cell lines and clinical specimens. Anal Biochem. 1993; 213:414-21.

25. White TJ, Madej R, Persing DH. The polymerase chain reaction: clinical applications. Adv Clin Chem. 1992; 29:161-96.

26. Wackym PA, Simpson TA, Gantz BJ, et al. Polymerase chain reaction amplification of DNA from archival celloidin-embedded human temporal bone sections. Laryngoscope. 1993; 103:583-8.

27. O'Garra A, Vieira P. Polymerase chain reaction for detection of cytokine gene expression. Curr Opin Immunol. 1992; 4:211-5.

28. Kittler R, Stoneking M, Kayser M. A whole genome amplification method to generate long fragments from low quantities of genomic DNA. Anal Biochem. 2002; 300:237-44.

29. Barbaux S, Poirier O, Cambien F. Use of degenerate oligonucleotide primed PCR (DOP-PCR) for the genotyping of low-concentration DNA samples. J Mol Med. 2001; 79:329-32.

30. Simpson DJ, Bicknell EJ, Buch HN, et al. Genome-wide amplification and allelotyping of sporadic pituitary adenomas identify novel regions of genetic loss. Genes Chromosomes Cancer. 2003; 37:225-36.

31. Klein CA, Schmidt-Kittler O, Schardt JA, et al. Comparative genomic hybridization, loss of heterozygosity, and DNA sequence analysis of single cells. Proc Natl Acad Sci U S A;1999; 96:4494-9.

32. Ludecke HJ, Senger G, Claussen U, et al. Cloning defined regions of the human genome by microdissection of banded chromosomes and enzymatic amplification. Nature. 1999; 338:348-50.

33. Saunders RD, Glover DM, Ashburner M, et al. PCR amplification of DNA microdissected from a single polytene chromosome band: a comparison with conventional microcloning. Nucleic Acids Res. 1989; 17:9027-37.

34. Zhang L, Cui X, Schmitt K, et al. Whole genome amplification from a single cell: implications for genetic analysis. Proc Natl Acad Sci U S A. 1992; 89:5847-51.

35. Dietmaier W, Hartmann A, Wallinger S, et al. Multiple mutation analyses in single tumour cells with improved whole genome amplification. Am J Pathol. 1999. 154:83-95.

36. Telenius H, Carter NP, Bebb CE, et al. Degenerate oligonucleotide-primed PCR: general amplification of target DNA by a single degenerate primer. Genomics. 1992; 13:718-25.

37. Nagy J, Feher LZ, Sonkodi I, et al. A second field metachronous Merkel cell carcinoma of the lip and the palatine tonsil confirmed by microarray-based comparative genomic hybridisation. Virchows Arch. 2005; 446:278-86.

38. Nelson DL,. Ledbetter SA, Corbo L, et al. Alu polymerase chain reaction: a method for rapid isolation of human-specific sequences from complex DNA sources. Proc Natl Acad Sci U S A. 1989; 86:6686-90.

39. Dean FB, Hosono S, Fang L, et al. Comprehensive human genome amplification using multiple displacement amplification. Proc Natl Acad Sci U S A. 2002; 99:5261-6.

40. Phillips J, Eberwine JH. Antisense RNA Amplification: A Linear Amplification Method for Analyzing the mRNA Population from Single Living Cells. Methods. 1996; 10:283-8.

41. Dean FB, Nelson JR, Giesler TL, et al. Rapid amplification of plasmid and phage DNA using Phi 29 DNA polymerase and multiply-primed rolling circle amplification. Genome Res. 2001; 11:1095-9.

42. Lizardi PM, Huang X, Zhu Z, et al. Mutation detection and single-molecule counting using isothermal rolling-circle amplification. Nat Genet. 1998; 19:225-32.

43. Puskas LG, Zvara A, Hackler L Jr, et al. RNA amplification results in reproducible microarray data with slight ratio bias. Biotechniques. 2002; 32:1330-40.

44. Puskas LG, Fartmann B, Bottka S. Restricted PCR: amplification of an individual sequence flanked by a highly repetitive element from total human DNA. Nucleic Acids Res. 1994; 22:3251-2.

45. Hughes S, Arneson N, Done S, et al. The use of whole genome amplification in the study of human disease. Prog Biophys Mol Biol. 2005; 88:173-89.

46. Wang G, Brennan C, Rook M, et al. Balanced-PCR amplification allows unbiased identification of genomic copy changes in minute cell and tissue samples. Nucleic Acids Res. 2004; 32:e76.

47. Larsen J, Ottesen AM, Lundsteen C, et al. Optimization of DOP-PCR amplification of DNA for high-resolution comparative genomic hybridization analysis. Cytometry. 2001; 44:317-25.

48. Peng DF, Sugihara H, Mukaisho K, et al. Alterations of chromosomal copy number during progression of diffuse-type gastric carcinomas: metaphase- and array-based comparative genomic hybridization analyses of multiple samples from individual tumours. J Pathol. 2003; 201:439-50.

49. Grant SF, Steinlicht S, Nentwich U, et al. SNP genotyping on a genome-wide amplified DOP-PCR template. Nucleic Acids Res. 2002; 30:e125.

50. Lasko D, Cavenee W, Nordenskjold M. Loss of constitutional heterozygosity in human cancer. Annu Rev Genet. 1991; 25:281-314.

47

51. Innis MA, Gelfand DH, Sninsky JJ, et al. Optimization of PCRs. In: Innis MA ed. PCR Protocols: A guide to Methods and Applications. Academic Press, New York, NY, USA. 1990:3-12.

52. Sardelli A. Plateau effect – understanding PCR limitations. Amplifications. 1993; 9:1-5. 53. Morrison C, Gannon F. The impact of the PCR plateau phase on quantitative PCR. Biochim Biophys Acta.

1994; 1219:493-498. 54. Dennis Lo YM. Introduction to the Polymerase Chain Reaction. In: Dennis Lo YM ed. Methods Mol. Med:

Clinical Applications of PCR. Humana Press Inc, Totowa, NJ. 55. Chee M, Yang R, Hubbell E, et al. Accessing genetic information with high-density DNA arrays. Science.

1996; 274:610-4. 56. Wang DG, Fan JB, Siao CJ, et al. Large-scale identification, mapping, and genotyping of single-nucleotide

polymorphisms in the human genome. Science. 1998; 280:1077-82. 57. Cho RJ, Campbell MJ, Winzeler EA, et al. A genome-wide transcriptional analysis of the mitotic cell cycle.

Mol Cell. 1998; 2:65-73. 58. Lockhart DJ, Dong H, Byrne MC, et al. Expression monitoring by hybridization to high-density

oligonucleotide arrays. Nat Biotechnol. 1996; 14:1675-80. 59. Wodicka L, Dong H, Mittmann M, et al. Genome-wide expression monitoring in Saccharomyces cerevisiae.

Nat Biotechnol. 1997; 15:1359-67. 60. Solinas-Toldo S, Lampel S, Stilgenbauer S, et al. Matrix-based comparative genomic hybridization:

biochips to screen for genomic imbalances. Genes Chromosomes Cancer. 1997; 20:399-407. 61. Pinkel D, Segraves R, Sudar D, et al. High resolution analysis of DNA copy number variation using

comparative genomic hybridization to microarrays. Nat Genet. 1998; 20:207-11. 62. Pollack JR, Perou CM, Alizadeh AA, et al. Genome-wide analysis of DNA copy-number changes using

cDNA microarrays. Nat Genet. 1999; 23:41-6. 63. Halata Z, Grim M, Bauman KI. Friedrich Sigmund Merkel and his "Merkel cell", morphology,

development, and physiology: review and new results. Anat Rec A Discov Mol Cell Evol Biol. 2003; 271:225-39.

64. Mir R, Sciubba JJ, Bhuiya TA, et al. Merkel cell carcinoma arising in the oral mucosa. Oral Surg Oral Med Oral Pathol. 1988; 65:71-5.

65. Schmidt-Westhausen A, Reichart PA, Gross UM. Gingival metastasis of merkel cell carcinoma: a case report. J Oral Pathol Med. 1996; 25:44-7.

66. Antoniades K, Giannouli T, Kaisaridou D. Merkel cell carcinoma in a patient with Recklinghausen neurofibromatosis. Int J Oral Maxillofac Surg. 1998; 27:213-4.

67. Bellome J, Bays DR. Merkel cell carcinoma of the ear: report of a case. J Oral Maxillofac Surg. 1998; 56:984-8.

68. Longo F, Califano L, Mangone GM, et al. Neuroendocrine (Merkel cell) carcinoma of the oral mucosa: report of a case with immunohistochemical study and review of the literature. J Oral Pathol Med. 1999; 28:88-91.

69. Park GC, Shelton JB Jr, Ow RA, et al. Merkel cell carcinoma of the lower lip: a case report and histopathologic study. Arch Otolaryngol Head Neck Surg. 1999; 125:907-11.

70. Bickle K, Glass LF, Messina JL, et al. Merkel cell carcinoma: a clinical, histopathologic, and immunohistochemical review. Semin Cutan Med Surg. 2004; 23:46-53.

71. Reichel OA, Mayr D, Issing WJ. Oropharyngeal metastasis of a Merkel cell carcinoma of the skin. Eur Arch Otorhinolaryngol. 2003; 260:258-60.

72. Hashimoto K. Fine structure of Merkel cell in human oral mucosa. J Invest Dermatol. 1972; 58:381-7. 73. Vigneswaran N, Muller S, Lense E, et al. Merkel cell carcinoma of the labial mucosa. An

immunohistochemical and ultrastructural study with a review of the literature on oral Merkel cell carcinomas. Oral Surg Oral Med Oral Pathol. 1992; 74:193-200.

74. Koss LG, Spiro RH, Hajdu S. Small cell (oat cell) carcinoma of minor salivary gland origin. Cancer. 1972; 30:737-41.

75. Benning TL, Vollmer RT, Crain BJ, et al. Neuroendocrine carcinoma of the oral cavity. Mod Pathol. 1990; 3:631-4.

76. Slaughter DP, Southwick HW, Smejkal W. Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin. Cancer. 1953; 6:963-8.

77. Califano J, van der Riet P, Westra W et al. Genetic progression model for head and neck cancer: implications for field cancerization. Cancer Res. 1996; 56:2488-92.

78. Braakhuis BJ, Tabor MP, Leemans CR, et al. Second primary tumours and field cancerization in oral and oropharyngeal cancer: molecular techniques provide new insights and definitions. Head Neck. 2002; 24:198-206.

79. Braakhuis BJ, Tabor MP, Kummer JA, et al. A genetic explanation of Slaughter's concept of field cancerization: evidence and clinical implications. Cancer Res. 2003; 63:1727-30.

80. Ha PK, Califano JA. The molecular biology of mucosal field cancerization of the head and neck. Crit Rev Oral Biol Med. 2003; 14:363-9.

48

81. Hodgson G, Hager JH, Volik S, et al. Genome scanning with array CGH delineates regional alterations in mouse islet carcinomas. Nat Genet. 2001; 29:459-64.

82. Snijders AM, Nowak N, Segraves R, et al. ssembly of microarrays for genome-wide measurement of DNA copy number. Nat Genet. 2001; 29:263-4.

83. Cowell JK, Wang YD, Head K, et al. Identification and characterisation of constitutional chromosome abnormalities using arrays of bacterial artificial chromosomes. Br J Cancer. 2004; 90:860-5.

84. Harle M, Arens N, Moll I, et al. Comparative genomic hybridization (CGH) discloses chromosomal and subchromosomal copy number changes in Merkel cell carcinomas. J Cutan Pathol. 1996; 23:391-7.

85. Van Gele M, Leonard JH, Van Roy N, et al. Frequent allelic loss at 10q23 but low incidence of PTEN mutations in Merkel cell carcinoma. Int J Cancer. 2001; 92:409-13.

86. Popp S, Waltering S, Herbst C, et al. UV-B-type mutations and chromosomal imbalances indicate common pathways for the development of Merkel and skin squamous cell carcinomas. Int J Cancer. 2002; 99:352-60.

87. Van Gele M, Leonard JH, Van Roy N, et al. Combined karyotyping, CGH and M-FISH analysis allows detailed characterization of unidentified chromosomal rearrangements in Merkel cell carcinoma. Int J Cancer. 2002;101:137-45.

88. Siironen P, Louhimo J, Nordling S, et al. Prognostic factors in papillary thyroid cancer: an evaluation of 601 consecutive patients. Tumour Biol. 2005; 26:57-64.

89. Gerdes J, Lemke H, Baisch H, et al. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J Immunol. 1984; 133:1710-5.

90. Kitamura Y, Shimizu K, Tanaka S, et al. ssociation of allelic loss on 1q, 4p, 7q, 9p, 9q, and 16q with postoperative death in papillary thyroid carcinoma. Clin Cancer Res. 2000; 6:1819-25.

91. De Micco C, Zoro P, Garcia S, et al. Thyroid peroxidase immunodetection as a tool to assist diagnosis of thyroid nodules on fine-needle aspiration biopsy. Eur J Endocrinol. 1994; 131:474-9.

92. Christensen L, Blichert-Toft M, Brandt M, et al. Thyroperoxidase (TPO) immunostaining of the solitary cold thyroid nodule. Clin Endocrinol (Oxf). 2000; 53:161-9.

93. Wu G, Mambo E, Guo Z, et al. Uncommon mutation, but common amplifications, of the PIK3CA gene in thyroid tumours. J Clin Endocrinol Metab. 2005; 90:4688-4693.

94. Suarez HG. Genetic alterations in human epithelial thyroid tumours. Clin Endocrinol (Oxf). 1998; 48:531-46. 95. Garcia-Rostan G, Zhao H, Camp RL, et al. ras mutations are associated with aggressive tumour phenotypes

and poor prognosis in thyroid cancer. J Clin Oncol. 2003; 21:3226-35. 96. Santoro M, Carlomagno F, Hay ID, et al. Ret oncogene activation in human thyroid neoplasms is restricted

to the papillary cancer subtype. J Clin Invest. 1992; 89:1517-22. 97. Nikiforov YE. RET/PTC rearragement in thyroid tumours. Endocr Pathol 2002; 13:3-16. 98. Santoro M, Papotti M, Chiappetta G, et al. RET activation and clinicopathologic features in poorly

differentiated thyroid tumours. J Clin Endocrinol Metab. 2002; 87:370-9. 99. Fagin JA. Challenging dogma in thyroid cancer molecular genetics-role of RET/PTC and BRAF in tumour

initiation. J Clin Endocrinol Metab 2004; 89:4264-4266. 100. Tallini G, Santoro M, Helie M, et al. RET/PTC oncogene activation defines a subset of papillary thyroid

carcinomas lacking evidence of progression to poorly differentiated or undifferentiated tumour phenotypes. Clin Cancer Res. 1998; 4:287-94.

101. Sugg SL, Ezzat S, Rosen IB, et al. istinct multiple RET/PTC gene rearrangements in multifocal papillary thyroid neoplasia. J Clin Endocrinol Metab. 1998; 83:4116-22.

102. Collins BJ, Chiappetta G, Schneider AB, et al. RET expression in papillary thyroid cancer from patients irradiated in childhood for benign conditions. J Clin Endocrinol Metab. 2002; 87:3941-6.

103. Bongarzone I, Vigneri P, Mariani L, et al. RET/NTRK1 rearrangements in thyroid gland tumours of the papillary carcinoma family: correlation with clinicopathological features. Clin Cancer Res. 1998; 4:223-8.

104. Xu X, Quiros RM, Gattuso P, et al. High prevalence of BRAF gene mutation in papillary thyroid carcinomas and thyroid tumour cell lines. Cancer Res. 2003; 63:4561-7.

105. Puxeddu E, Moretti S, Elisei R, et al. BRAF(V599E) mutation is the leading genetic event in adult sporadic papillary thyroid carcinomas. J Clin Endocrinol Metab. 2004; 89:2414-20.

106. Ciampi R, Knauf JA, Kerler R, et al. Oncogenic AKAP9-BRAF fusion is a novel mechanism of MAPK pathway activation in thyroid cancer. J Clin Invest. 2005; 115:94-101.

107. Fusco A, Viglietto G, Santoro M. A new mechanism of BRAF activation in human thyroid papillary carcinomas. J Clin Invest. 2005;115:20-3.

108. Vasko V, Hu S, Wu G, et al. High prevelance and possible de novo formation of BRAF mutation in metastasized papillary thyroid cancer in lymph nodes. J Clin Endocrinol Metab. 2005; 90:5265-5269.

109. Melillo RM, Castellone MD, Guarino V, et al. The RET/PTC-RAS-BRAF linear signaling cascade mediates the motile and mitogenic phenotipe of thyroid cancer cells. J Clin Invest. 2005; 115:1068-1081.

110. Xing M. BRAF mutation in thyroid cancer. Endocr Relat Cancer. 2005; 12:245-62. 111. Quiros RM, Ding HG, Gattuso P, et al. Evidence that one subset of anaplastic thyroid carcinomas are

derived from papillary carcinomas due to BRAF and p53 mutations. Cancer. 2005; 103:2261-8.

49

112. Elisei R, Romei C, Vorontsova T, et al. RET/PTC rearrangements in thyroid nodules: studies in irradiated and not irradiated, malignant and benign thyroid lesions in children and adults. J Clin Endocrinol Metab. 2001; 86:3211-6.

113. Fusco A, Chiappetta G, Hui P, et al. Assessment of RET/PTC oncogene activation and clonality in thyroid nodules with incomplete morphological evidence of papillary carcinoma: a search for the early precursors of papillary cancer. Am J Pathol. 2002; 160:2157-67.

114. Bongarzone I, Vigneri P, Mariani L, et al. RET/NTRK1 rearrangements in thyroid gland tumours of the papillary carcinoma family: correlation with clinicopathological features. Clin Cancer Res. 1998; 4:223-8.

115. Musholt TJ, Musholt PB, Khaladj N, et al. Prognostic signifi cance of RET and NTRK1 rearrangements in sporadic papillary thyroid carcinoma. Surgery. 2000; 128:984-93.

116. Donghi R, Longoni A, Pilotti S, et al. Gene p53 mutations are restricted to poorly differentiated and undifferentiated carcinomas of the thyroid gland. J Clin Invest. 1993; 91:1753-60.

117. Fagin JA, Matsuo K, Karmakar A, et al. High prevalence of mutations of the p53 gene in poorly differentiated human thyroid carcinomas. J Clin Invest. 1993; 91:179-84.

118. Ito T, Seyama T, Mizuno T, et al. Unique assotiation of p 53 mutations with undifferentiated but not with differentiated carcinomas of the thyroid gland. Cancer Research. 1992; 52:1369-1371.

119. Dobashi Y, Sakamoto A, Sugimura H, et al. Overexpression of p53 as a possible prognostic factor in human thyroid carcinoma. Am J Surg Pathol. 1993; 17:375-81.

120. Halaban R, Kwon BS, Ghosh S, et al. bFGF as an autocrine growth factor for human melanomas. Oncogene Res. 1988; 3:177-86.

121. Meier F, Nesbit M, Hsu MY, et al. Human melanoma progression in skin reconstructs: biological significance of bFGF. Am J Pathol. 2000; 156:193-200.

122. Varner JA, Cheresh DA. Integrins and cancer. Curr Opin Cell Biol. 1996; 8:724-30. 123. Brooks PC, Montgomery AM, Rosenfeld M, et al. Integrin alpha v beta 3 antagonists promote tumour

regression by inducing apoptosis of angiogenic blood vessels. Cell. 1994; 79:1157-64. 124. Brooks PC, Stromblad S, Klemke R, et al. Antiintegrin alpha v beta 3 blocks human breast cancer growth

and angiogenesis in human skin. J Clin Invest. 1995; 96:1815-22. 125. Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell. 2002; 110:673-87. 126. Piotrowicz RS, Orchekowski RP, Nugent DJ, et al. Glycoprotein Ic-IIa functions as an activation-

independent fibronectin receptor on human platelets. J Cell Biol. 1988; 106:1359-64. 127. Ill CR, Engvall E, Ruoslahti E. Adhesion of platelets to laminin in the absence of activation. J Cell Biol.

1984; 99:2140-5. 128. Sonnenberg A, Modderman PW, Hogervorst F. Laminin receptor on platelets is the integrin VLA-6.

Nature. 1988; 336:487-9. 129. Staatz WD, Rajpara SM, Wayner EA, et al. The membrane glycoprotein Ia-IIa (VLA-2) complex mediates

the Mg++-dependent adhesion of platelets to collagen. J Cell Biol. 1989; 108:1917-24. 130. Bennett JS, Chan C, Vilaire G, et al. gonist-activated alphavbeta3 on platelets and lymphocytes binds to the

matrix protein osteopontin. J Biol Chem. 1997; 272:8137-40. 131. Paul BZ, Vilaire G, Kunapuli SP, et al. Concurrent signaling from Galphaq- and Galphai-coupled pathways

is essential for agonist-induced alphavbeta3 activation on human platelets. 132. Duperray A, Berthier R, Chagnon E, et al. Biosynthesis and processing of platelet GPIIb-IIIa in human

megakaryocytes. J Cell Biol. 1987; 104:1665-73. 133. Kolodziej MA, Vilaire G, Gonder D, et al. Study of the endoproteolytic cleavage of platelet glycoprotein

IIb using oligonucleotide-mediated mutagenesis. J Biol Chem. 1991; 266:23499-504. 134. Bennett JS. Structural biology of glycoprotein IIb-IIIa. Trends Cardiovasc. Med. 1996; 6:31-37. 135. Diaz-Gonzalez F, Forsyth J, Steiner B, et al. Transdominant inhibition of integrin function. Mol Biol Cell.

1991; 12:1939-51. 136. Albelda SM, Mette SA, Elder DE, et al. Integrin distribution in malignant melanoma: association of the beta

3 subunit with tumour progression. Cancer Res. 1990; 50:6757-64. 137. Hieken TJ, Ronan SG, Farolan M, et al. Molecular prognostic markers in intermediate-thickness cutaneous

malignant melanoma. Cancer. 1999; 85:375-82. 138. Nierodzik ML, Kajumo F, Karpatkin S. Effect of thrombin treatment of tumour cells on adhesion of tumour

cells to platelets in vitro and tumour metastasis in vivo. Cancer Res. 1992; 52:3267-72. 139. Chen YQ, Gao X, Timar J, et al. Identification of the alpha IIb beta 3 integrin in murine tumour cells. J Biol

Chem. 1992; 267:17314-20. 140. Tang DG, Onoda JM, Steinert BW, et al. Phenotypic properties of cultured tumour cells: integrin alpha IIb

beta 3 expression, tumour-cell-induced platelet aggregation, and tumour-cell adhesion to endothelium as important parameters of experimental metastasis. Int J Cancer. 1993; 54:338-47.

141. Chiang HS, Peng HC, Huang TF. Characterization of integrin expression and regulation on SW-480 human colon adenocarcinoma cells and the effect of rhodostomin on basal and upregulated tumour cell adhesion. Biochim Biophys Acta. 1994; 1224:506-16.

50

142. Trikha M, Timar J, Lundy SK, et al. Human prostate carcinoma cells express functional alphaIIb(beta)3 integrin. Cancer Res. 1996; 56:5071-8.

143. Trikha M, Raso E, Cai Y, et al. Role of alphaII(b)beta3 integrin in prostate cancer metastasis. Prostate. 1998; 35:185-92.

144. Chen YQ, Trikha M, Gao X, et al. Ectopic expression of platelet integrin alphaIIb beta3 in tumour cells from various species and histological origin. Int J Cancer. 1997; 72:642-8.

145. Honn KV, Chen YQ, Timar J, et al. Alpha IIb beta 3 integrin expression and function in subpopulations of murine tumours. Exp Cell Res. 1992; 201:23-32.

146. Trikha M, Timar J, Lundy SK, et al. The high affinity alphaIIb beta3 integrin is involved in invasion of human melanoma cells. Cancer Res. 1997; 57:2522-8.

147. Trikha M, Timar J, Zacharek A, et al. Role for beta3 integrins in human melanoma growth and survival. Int J Cancer. 2002; 101:156-67.

148. Sambrook J, Fritsch EF, Maniatis T, Molecular cloning: A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 1989.

149. Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 2001; 29:e45.

150. Kitajka K, Puskas LG, Zvara A, et al. The role of n-3 polyunsaturated fatty acids in brain: modulation of rat brain gene expression by dietary n-3 fatty acids. Proc Natl Acad Sci U S A 2002; 99:2619–24.

151. Puskas LG, Zvara A, Hackler L Jr, et al. Production of bulk amounts of universal RNA for DNA microarrays. Biotechniques. 2002; 33:898-904.

152. Puskas LG, Hackler L Jr, Kovacs G, et al. Recovery of cyanine-dye nucleotide triphosphates. Anal Biochem. 2002; 305:279-81.

153. Yang YH, Dudoit S, Luu P, et al. Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res. 2002; 30:e15.

154. Balazs M, Adam Z, Treszl A, et al. Chromosomal imbalances in primary and metastatic melanomas revealed by comparative genomic hybridization. Cytometry. 2001; 46:222-32.

155. Larramendy ML, Koljonen V, Bohling T, et al. Recurrent DNA copy number changes revealed by comparative genomic hybridization in primary Merkel cell carcinomas. Mod Pathol. 2004; 17:561-7.

156. Lehnerdt KH, Broicher R, Tschubel K, et al. Der interessante Fall Nr.49. Laryngorhinootologie 2001; 80:687-689.

157. Tesei F, Farneti G, Cavicchi O, et al. A case of Merkel-cell carcinoma metastatic to the tonsil. J Laryngol Otol. 1992; 106:1100-2.

158. Takano T, Hasegawa Y, Miyauchi A, et al. Quantitative analysis of thymosin beta-10 messenger RNA in thyroid carcinomas. Jpn J Clin Oncol. 2002; 32:229-32.

159. Onno M, Nakamura T, Mariage-Samson R, et al. Human TRE17 oncogene is generated from a family of homologous polymorphic sequences by single-base changes. DNA Cell Biol. 1993; 12:107-18.

160. Martinu L, Masuda-Robens JM, Robertson SE, et al. The TBC (Tre-2/Bub2/Cdc16) domain protein TRE17 regulates plasma membrane-endosomal trafficking through activation of Arf6. Mol Cell Biol. 2004; 24:9752-62.

161. Chiappetta G, Pentimalli F, Monaco M, et al. Thymosin beta-10 gene expression as a possible tool in diagnosis of thyroid neoplasias. Oncol Rep. 2004; 12:239-43.

162. Santelli G, Califano D, Chiappetta G, et al. Thymosin beta-10 gene overexpression is a general event in human carcinogenesis. Am J Pathol. 1999; 155:799-804.

163. Hall AK. Thymosin beta-10 accelerates apoptosis. Cell Mol Biol Res. 1995; 41:167-80. 164. Califano D, Monaco C, Santelli G, et al. Thymosin beta-10 gene overexpression correlated with the highly

malignant neoplastic phenotype of transformed thyroid cells in vivo and in vitro. Cancer Res. 1998; 58:823-8. 165. Lee SH, Son MJ, Oh SH, et al. Thymosin {beta}(10) inhibits angiogenesis and tumour growth by

interfering with Ras function. Cancer Res. 2005; 65:137-48. 166. Diviani D, Abuin L, Cotecchia S, et al. Anchoring of both PKA and 14-3-3 inhibits the Rho-GEF activity of

the AKAP-Lbc signaling complex. EMBO J. 2004; 23:2811-20. 167. Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during

tumourigenesis. Cell. 1996; 86:353-64. 168. Semenza GL. Angiogenesis in ischemic and neoplastic disorders. Annu Rev Med. 2003; 54:17-28. 169. Maniotis AJ, Folberg R, Hess A, et al. Vascular channel formation by human melanoma cells in vivo and in

vitro: vasculogenic mimicry. Am J Pathol 1999; 155:739–75. 170. Hendrix MJC, Seftor EA, Hess AR, et al. Molecular plasticity of human melanoma cells. Oncogene 2003;

22:3070–5. 171. Hendrix MJC, Seftor EA, Hess AR, et al. Vasculogenic mimicry and tumour-cell plasticity: lesson from

melanoma. Nature Rev Cancer 2003; 3:411–21.

Improved DOP-PCR–Based RepresentationalWhole-Genome Amplification Using Quantitative

Real-Time PCR

Liliana Z. Feher, MSc,* Margit Balazs, PhD, DSc,w Janos Z. Kelemen, MSc,* Agnes Zvara, PhD,*Istvan Nemeth, MD,z Zoltan Varga-Orvos, MD,* and Laszlo G. Puskas, PhD*

Abstract: In many cases, only a minute amount of partially

degraded genomic DNA can be extracted from archived clinical

samples. Diverse whole-genome amplification methods are

applied to provide sufficient amount of DNA for comparative

genome hybridization, single-nucleotide polymorphism, and

microsatellite analyses. In these applications, the reliability of

the amplification techniques is particularly important. In PCR-

based approaches, the plateau effect can seriously alter the

original relative copy number of certain chromosomal regions.

To eliminate this distorting effect, we improved the standard

degenerate oligonucleotide-primed PCR (DOP-PCR) technique

by following the amplification status with quantitative real-time

PCR (QRT-PCR). With real-time detection of the products, we

could eliminate DNA overamplification. Probes were prepared

from 10 different tumor samples: primary and metastatic

melanoma tissues, epidermoid and bronchioloalveolar lung

carcinomas, 2 renal cell carcinomas, 2 colorectal carcinomas,

and a Conn and Cushing adenoma. Probes were generated by

using nonamplified and amplified genomic DNA with DOP-

PCR and DOP-PCR combined with QRT-PCR. To demon-

strate the reliability of the QRT-PCR based amplification

protocol, altogether 152 relative copy number changes of 44

regions were determined. There was 85.6% concordance in copy

number alterations between the QRT-PCR protocol and the

nonamplified samples, whereas this value was only 63.8% for

the traditional DOP-PCR. Our results demonstrate that our

protocol preserves the original copy number of different

chromosomal regions in amplified genomic DNA than standard

DOP-PCR techniques more accurately.

Key Words: degenerate oligonucleotide-primed PCR (DOP-

PCR), genome amplification, quantitative real-time PCR, copy

number changes

(Diagn Mol Pathol 2006;15:43–48)

H igh-throughput genomic analysis requires largeamounts of template; however, the typical yield of

DNA from individual samples can be often limited. Theprimary drawback of the use of laser capture microscopyin DNA analysis is that microdissections yield insufficientnucleic acids due to low genomic DNA recovered fromsmall sample sizes. With samples such as these, the abilityto conduct sample amplification becomes imperative, toensure that enough material is available for geneticanalysis. Not only the amount but also the quality ofgenomic DNA obtained from various samples is impedi-ment for further studies. For example, paraffin-embeddedtissues often provide partially degraded low-qualityDNA. Therefore, developing a whole-genome amplifica-tion method which is not entirely dependent on high-quality DNA is eminently important to provide anextensive supply of DNA for large-scale genetic studies1

such as metaphase comparative genomic hybridization(CGH),2 array CGH,3 single-nucleotide polymorphism(SNP) genotyping,4 microsatellite genotyping,5 single-stranded conformational polymorphism (SSCP),6 loss ofheterozygosity (LOH),7 and sequence analysis.8 Geneamplifications and deletions are common in tumors andcan influence tumorigenesis and metastatic potential.Traditional CGH can detect aneuploidies, deletions, andunbalanced translocations. There are 3 basic techniquesto amplify the entire genomic DNA: PCR-based ap-proaches,5–15 strand or multiple displacement amplifica-tion (SDA or MDA),16 and T7-based linear amplification(TLAD).17

The SDA and MDA isothermal methods employthe unique biochemical properties of Phi29 DNA poly-merase to amplify linear DNA. These techniques arebased on the rolling circle amplification by which circularDNA molecules such as plasmids or viruses frequentlyreplicate. These rolling circle amplification methods wereinitially adapted for the amplification of large circularDNA templates18,19 and more recently for the amplifica-tion of genomic DNA.16 The drawback of SDA is that theefficiency of amplification depends on the length of thetemplate. TLAD is applicable for the amplification ofgenomic DNA of varying fragment size and quality, andCopyright r 2006 by Lippincott Williams & Wilkins

From the *Laboratory of Functional Genomics, Biological ResearchCenter, Hungarian Academy of Sciences, Szeged, Hungary; wDepart-ment of Preventive Medicine, School of Public Health, Medical andHealth Science Center, University of Debrecen, Hungary; andzDepartment of Pathology, University of Szeged, Szeged, Hungary.

Supported by grants from the Hungarian Medical Research Council(ETT 400/2003) and a bilateral grant between the Hungarian PrimeMinister’s Office and Hungarian Academy of Sciences (40.0499/2004).

Reprints: Laszlo G. Puskas, Laboratory of Functional Genomics,Biological Research Center, Hungarian Academy of Sciences, P.O.Box 521, Szeged, H-6701, Hungary (e-mail: [email protected]).

ORIGINAL ARTICLE

Diagn Mol Pathol � Volume 15, Number 1, March 2006 43

it is based on a protocol devised by Phillips and Eberwinefor the amplification of mRNA for cDNA microarrays.17

The interspread repetitive sequence PCR15 was thefirst PCR-based genome amplification method which usedprimers designed to anneal to Alu repeats within thegenome. This strategy was further improved to amplifyspecific DNA flanked by Alu repeats by using the so-called restricted-PCR technology.20 An alternative meth-od,9 termed the linker adapter technique PCR (LA-PCR),involved restriction digestion of the DNA and ligation ofadaptors to serve as priming sites for subsequent PCR. Inprinciple, the ligation-mediated PCR (LMP) is verysimilar to the LA-PCR method.9,10 Nonetheless, thisapproach has not been widely applied, likely due totechnical difficulties involved. PCR-based genome ampli-fication methods, primer extension preamplification(PEP),11,12 and DOP-PCR13,14 techniques have beendeveloped for the amplification of genomic DNA startingfrom as little as a single cell. The original PEP protocolinvolves a 50-cycle PCR program using a 15 bp longoligonucleotide primer; however, via this approach onlyB78% of the genome from a single haploid cell can becopied.11 The DOP-PCR technique is increasingly beingapplied for simultaneous amplification of multiple loci intarget DNA using oligonucleotide primers of partiallydegenerate sequences13; however, the amplified DNAdoes not cover the genome completely. Contrary to otherPCR-based amplification methods (Alu-PCR, IRS-PCR),DOP-PCR is a species-independent technique. Theprotocol is adapted for the use of microdissected,formalin-fixed, and paraffin-embedded tissue in compre-hensive genetic analyses.21

DOP-PCR and PEP protocols amplify DNA in anexponential fashion, making them highly susceptible tobiases, and they have a tendency to specifically over-represent or downrepresent certain sequences. The majordisadvantage of PCR-based protocols is that this coulddistort the ratio of the starting genomic imbalances.22

Recently, Wang et al23 have described a modified DOP-PCR protocol for balanced amplification of genomicDNA; however, the amplification status has not beenfollowed during PCR. In their protocol, genomic DNAwas digested with a restriction enzyme, and tagged linkerswere ligated to the products before amplification. In ourprotocol, direct amplification of genomic DNA can beperformed without any pretreatment of the samples. Ourimproved protocol of DOP-PCR includes QRT-PCR tofollow the cycling status which results in a more reliableand reproducible sample amplification.

MATERIALS AND METHODS

Tissue SpecimensMelanoma tissues were obtained from the Depart-

ment of Dermatology, University of Debrecen, Hungary.Primary melanoma (location: face) and metastatic tumor(location: lymph node) were removed from the samepatient at the same time. The relative amount of tumorcells was a minimum of 70% in all lesions as confirmed

after hematoxylin and eosin staining by light microscopy.Epidermoid, bronchioloalveolar lung carcinoma, 2 renalcell carcinoma, and 2 colorectal carcinoma specimenswere obtained from the Institute of Pathology, SzegedUniversity. Conn and Cushing adenoma samples wereobtained from Semmelweis University, Budapest. High–molecular-weight DNA was extracted according to astandard procedure starting from 50mg of tissue.24,25

Chromosomal CGH and image analyses were performedas described earlier in details elsewhere.25

Preparation of DOP-PCR Amplified andOveramplified Genomic DNA

Genomic DNA (20 ng) from each sample wasamplified with a modified version of the DOP-PCRprotocol by using a RotorGene 3000 QRT-PCR instru-ment (Corbett Research, Mortlake, Australia) to followthe amplification.14,26 The DNA concentration wasassessed spectrophotometrically by NanoDrop (Rock-land, DE). Reactions were performed in a total volume of100 mL. The cycling parameters were as follows: heat startat 951C (15minutes), 8 cycles denaturation at 941C (50seconds), annealing for 2minutes from 451C to 721C witha 0.21C/s ramp, and extension at 721C (90 seconds). After8 cycles, the reaction mix was divided into two 50-mLaliquots and SybrGreen was added to both (1� finalconcentration, Molecular Probes, Eugene, USA). Thefollowing cycling protocol was performed in a real-timePCR instrument: denaturation at 951C (40 seconds),annealing at 581C (60 seconds), and extension at 721Cfor 80 seconds. To avoid overamplification of theproducts, cycling was terminated just before the reactionreached the plateau (ie, at 13–15 cycles); in contrast, theoveramplified genomic DNA was obtained from PCRwith 21 cycles.

The PCR products were purified with PCR-purifi-cation columns (Bioneer, Daejeon, Korea). In thereactions, UN primer (50-CCGACTCGAGNNNNN-NATGTGG-30) was used in 1 mmol/L concentration.13

The reactions were performed with 1� ABsolute QPCRMixes (ABgene, Surrey, UK).

Confirmatory Quantitative Real-Time PCRQRT-PCR was carried out in a final volume of

20 mL with 20 ng nonamplified or amplified genomicDNA in a RotorGene 3000 instrument. The reactionswere performed with 5 pmol/each forward and reversegene-specific primers in 1� ABsolute QPCR Mix(ABgene, Surrey, UK) with SybrGreen (1 mmol/L finalconcentration, Molecular Probes) with the followingprotocol: heat start for 15minutes at 951C, 45 cycles of25 seconds denaturation at 951C, 25 seconds annealing at591C, and 25 seconds extension at 721C. Fluorescentsignals were collected after each extension step at 721C.Curves were analyzed with the RotorGene software,using dynamic tube and slope correction methods,whereas ignoring data from cycles close to baseline.Primers were designed by using the PrimerExpress soft-ware (Applied Biosystems, Foster City, CA). Relative

Feher et al Diagn Mol Pathol � Volume 15, Number 1, March 2006

44 r 2006 Lippincott Williams & Wilkins

ratios were normalized to the Ct values of the metallo-peptidase 1 gene as an internal control and calculatedwith the Pfaffl method.27 The sequences of PCR primersused in this study were deposited to http://160.114.60.33/Data1/. All the PCRs were performed 3 times in separateruns.

RESULTS AND DISCUSSIONDOP-PCR is a simple method for whole-genome

amplification in comprehensive genetic studies (CGH,SSCP, SNP, and microsatellite genotyping) that isincreasingly being applied for simultaneous amplificationof multiple loci in target DNA. DOP-PCR utilizesoligonucleotide primers with partially degenerate se-quences close to their 30-end.13 DOP-PCR is a species-independent approach which produces sufficient amountsof DNA for analyzing clinical tumor samples, forprenatal diagnosis, and for many other investigations. Acomparison of several DOP-PCR methods is alsodescribed by Larsen et al.21 Unfortunately, PCR-basedprotocols often result in misrepresentation of the genome,with some regions being overrepresented or even lost. Thereliability of PCR-based amplification mostly depends onthe exponential phase during cycling. At the late cycles ofamplification, the original copy number of chromosomalregions can be changed and the method provides falseresults. The plateau effect28,29 is the result of a markedshift of the overall mass balance in favor of the reactionproduct. Different factors seem to be responsible for theattainment of the plateau rather than a single factor orparameter, as readdition of presumably exhausted re-agents at late cycles did not cause the reaction to proceedwith increased efficiency.30 The plateau effect is caused bythe accumulation of product molecules, usually theconsequence of the significant degree of annealingbetween complementary strands of the products, ratherthan between the primers and template. Furthermore, thefinite amount of enzyme molecules present will be unableto extend all the primer-template complex in the givenextension time.31

Here, we present a more reliable PCR-basedgenome amplification procedure, developed as an alter-native sample amplification method which follows theamplification status by QRT-PCR. With real-time detec-tion of the products, DNA overamplification could bediminished. After initial amplification of the genome,SybrGreen dye is added to the reaction mix and a secondamplification step is applied in a QRT-PCR instrument.The cycling is performed until the reaction reaches theplateau (Fig. 1).

To compare the original gene copy number ofdifferent templates generated by using the traditionalDOP-PCR and with the protocol using QRT-PCR, 44genes were selected for confirmation by using QRT-PCRanalysis in cases of 2 melanoma samples. Genes wereselected on the basis of metaphase CGH data performedearlier.25 The QRT-PCR was performed on DNAsamples obtained from primary and metastatic melanoma

tissues resected from the same patient along with normalDNA prepared from peripheral blood of a healthyindividual as a control. Altogether 88 relative copynumbers were determined on 3 different templates:nonamplified as control, amplified with DOP-PCR, andamplified with our protocol: DOP-PCR in combinationwith QRT-PCR (Fig. 1). In traditional DOP-PCRprotocol, we applied 21 cycles for each reaction; however,genomic DNA was PCR amplified with our combinedDOP-PCR-QRT-PCR method in 13 to 15 cycles, whichwere sufficient to reach early saturation phase.26 The copynumbers of selected genes were normalized to the copynumber of the metallopeptidase 1 gene. This single-copygene did not show any alteration in copy number in eithermelanoma sample and gave reproducible results in allcases. Relative copy number change that resulted in lessthan 0.5 meant loss of genetic material (heterozygousloss). If this value was greater than 1.5, it was defined asgain or gene amplification at that certain chromosomelocation (duplication of one allele). The results of theconfirmatory QRT-PCR are shown in Table 1.

Genomic changes (no change, loss or gain of copynumber) were determined for all the 44 genes in eachmethod on both samples (primary melanoma and

FIGURE 1. Whole-genomic DNA amplification using QRT-PCR. Diagram of the primary, metastatic melanoma tissue,and normal peripheric lymphocyte. The amplification haltedat the 13th cycle, the amplified genomic DNA (2) wasgenerated in the exponential phase of the reaction; theoveramplified genomic DNA (1) was isolated from reactionshalted at the 21st cycle; number 3 denotes for thenontemplate control.

Diagn Mol Pathol � Volume 15, Number 1, March 2006 Whole-Genome Amplification

r 2006 Lippincott Williams & Wilkins 45

TABLE 1. Comparison of Copy Number Changes of Several Gene loci Obtained from Primary and Metastatic Tumour SamplesRelative to Normal Tissue by Using Non-Amplified and Different Amplified Genomic DNA as Templates. Light Gray:Altered Relative Copy Number at Certain Chromosomal Regions in the Amplified Sample Compared to Control,Non-Amplified Sample

Chrom.

Location

Gene Product

(Acession No.)

Type of

Melanoma

Non-Amplified

(SD)

DOP-PCR+QRT-PCR

(SD)

DOP-PCR

(SD)

Metaphase

CGH

Loss of chromosomal regions12q14.3 Chaperonin cont. t-complex

polypept. 1b (AF026293)Primary 0.43 (0.21) 0.27 (0.28) 2.20 (0.39)

DeletionMetastatic 0.44 (0.12) 0.27 (0.03) 0.31 (0.31)

17p13 Human clone 23665 (U90913) Primary 0.50 (0.07) 0.39 (0.03) 0.31 (0.05)Deletion

Metastatic 0.48 (0.01) 0.33 (0.03) 0.48 (0.07)17p13.1 15S-lipoxygenase (U78294) Primary 0.50 (0.08) 0.50 (0.06) 1.77 (0.10)

DeletionMetastatic 0.51 (0.02) 0.44 (0.12) 1.44 (0.00)

2p23 Anaplastic lymphoma kinase(Ki-1) (NM_004304)

Metastatic 0.57 (0.03) 0.56 (0.28) 0.28 (0.13) Deletion

10q26.13-q26.3 Dedicator of cytokinesis 1(NM_001380)

Primary 0.58 (0.34) 0.52 (0.06) 0.74 (0.05) Deletion

Genes with no change2p23 Anaplastic lymphoma kinase

(Ki-1) (NM 004304)Primary 1.11 (0.07) 0.68 (0.03) 2.77 (0.95) No change

2p25.3 Myelin transcription factor1-like (BC071612)

Primary 1.16 (0.09) 0.50 (0.17) 0.74 (0.16) No changeMetastatic 0.06 (0.01) 0.40 (0.09) 0.54 (0.06) Deletion

10p12.1 G protein-coupled receptor158 (AY528411)

Primary 0.87 (0.04) 0.52 (0.03) 1.47 (0.05) No changeMetastatic 0.64 (0.06) 0.59 (0.02) 1.25 (0.07) Deletion

9p23-p24.3 Prot. tyrosine phosphatase,receptor D (NM_130392)

Primary 0.83 (0.02) 0.53 (0.22) 1.63 (0.93)Deletion

Metastatic 0.85 (0.16) 0.58 (0.25) 0.69 (0.29)11p15.5 p27-BWR1B (AF037066) Primary 0.89 (0.06) 0.92 (0.03) 0.90 (0.12)

No changeMetastatic 0.82 (0.31) 1.13 (0.08) 0.72 (0.06)

11q23 Apolipoprotein A-Iprecursor (X02162)

Primary 0.76 (0.04) 0.86 (0.38) 0.90 (0.12)No change

Metastatic 0.90 (0.07) 0.82 (0.32) 0.75 (0.10)1q23 CD48 antigen (B-cell

memberane protein) (M37766)Primary 1.07 (0.11) 0.69 (0.14) 0.48 (0.12)


9q33.1 Astrotactin 2 (BC010680) Primary 1.01 (0.10) 0.49 (0.16) 0.66 (0.25)No change

Metastatic 1.17 (0.32) 0.59 (0.65) 0.65 (0.15)1p31 Leptin receptor (U66497) Primary 0.86 (0.01) 1.10 (0.02) 1.27 (0.23)


2q35 Fibronectin 1 (X02761) Primary 1.03 (0.11) 0.99 (0.16) 0.99 (0.31)No change

Metastatic 1.14 (0.17) 1.11 (0.06) 1.24 (1.03)5q11.2 Integrin, alpha 2 (M28249) Primary 1.20 (0.19) 0.97 (0.16) 0.48 (0.06)


7q31 Met proto-oncogene (J02958) Primary 1.05 (0.23) 0.82 (0.07) 0.87 (0.06)No change

Metastatic 1.31 (0.04) 1.10 (0.03) 0.41 (0.06)19q13.3 OTK18 (D50419) Primary 1.18 (0.12) 0.51 (0.02) 1.01 (0.55)


14q11.2-13.1 Ribonuclease k6 precursorgene (U64998)


Metastatic 1.31 (0.05) 1.11 (0.06) 0.93 (0.12)14q22.3-q23.1 Retinoblastoma-binding

protein 1 (S66427)Primary 1.24 (0.09) 0.82 (0.01) 1.34 (0.03)


14q23-24.1 RNA pol. transcriptionalregulation mediator (U78082)


Metastatic 1.14 (0.04) 0.89 (0.06) 1.17 (0.10)4p14 Ubiquitin carboxyl-terminal

hydrolase L1 (X04741)Primary 0.90 (0.14) 0.35 (0.27) 0.62 (0.25)

AmplificationMetastatic 0.98 (0.19) 0.60 (0.48) 0.80 (0.04)

16p12-13.2 Epithelial amiloride-sensitiveNa channel g (X87160)


Metastatic 1.30 (0.07) 1.05 (0.11) 1.14 (0.10)11q13.1-13.3 Cell cycle checkpoint control

protein (U53174)Primary 1.03 (0.13) 0.84 (0.20) 0.67 (0.13)


15q23-25 Neuromedin B precursor(M21551)


Metastatic 1.22 (0.10) 1.42 (0.05) 2.16 (1.91)15q15-q21.1 Fibroblast growth factor 7

(M60828)Primary 0.89 (0.12) 0.68 (0.11) 0.91 (0.02)


16p13.3 N-methylpurine-DNA glycosylase(M74905)


Metastatic 1.34 (0.12) 1.73 (0.03) 0.78 (0.15)19q13.3-4 Electron-transfer-flavoprotein b

(X71129)Primary 1.24 (0.06) 1.12 (0.10) 0.99 (0.41)


1p33-34.1 Mitotic centromere-associatedkinesin (U63743)


Metastatic 1.03 (0.11) 0.83 (0.07) 0.51 (0.02)2q32-33 Amyotrophic lateral sclerosis

2 (AB011121)Primary 1.10 (0.37) 1.37 (0.09) 1.41 (0.11)


14q31 Galactocerebrosidase (D86181) Primary 1.46 (0.12) 0.68 (0.10) 0.50 (0.01)No change

Metastatic 1.41 (0.04) 0.87 (0.08) 0.55 (0.08)



metastatic tumor), and all the 88 values were comparedwith each other. In case of the QRT-PCR protocol,88.6% of the copy number changes matched to thenonamplified samples, whereas traditional DOP-PCRresulted in only 63.6% accuracy. The correlation coeffi-cient for the comparison of the nonamplified and theDOP-PCR amplified genomic DNA was only 0.27,whereas this value was significantly increased to 0.65when the same protocol was used, but the amplificationstatus was followed by QRT-PCR.

To demonstrate the differences in relative copynumbers, which were obtained with different amplificationmethods, a scatter plot was prepared (Fig. 2). In case ofour combined method, most of the genes had relative copynumbers similar to those of the nonamplified method,whereas traditional DOP-PCR distorted the relative copynumbers detected in the non-amplified samples.

To strengthen the extension of the preliminaryobservations, a larger number of specimens were furtheranalyzed. Amplified samples from epidermoid, bronchio-loalveolar lung carcinoma, 2 renal cell carcinoma, 2colorectal carcinomas, and Conn and Cushing adenomasamples were processed as described earlier. Copynumber changes of 8 different chromosome regions were

followed by QRT-PCR. Nonamplified and amplifiedsamples obtained from traditional DOP-PCR techniqueand from our improved protocol were analyzed. Theresults of copy number changes can be seen at http://160.114.60.33/Data2/. In the cases of the analyzed 8tumor samples, there was 84.4% concordance in copynumber alterations between the QRT-PCR protocol andthe nonamplified samples, whereas this value was only64% for the traditional DOP-PCR.

In conclusion, our method improved the traditionalDOP-PCR technique for whole-genome amplification.Improved reliability of the QRT-PCR-controlled DOP-PCR was confirmed by the analysis of 152 copy numberchanges in ten tumor samples. We assume that one of themost distorting effects of the DOP-PCR technique in thestudy of amplified genomic DNA is the overamplificationof the products. Overamplification has a dramaticdistorting effect on the original genomic rearrangementsand can generate false result. This effect can be decreasedby adding QRT-PCR to the original protocol. BecauseDOP-PCR is a widely used technique and the preserva-tion of the original genome set is especially important ingenomic studies, we propose to include QRT-PCR tofollow and control the amplification status.

TABLE 1. (continued)

Chrom.

Location

Gene Product

(Acession No.)

Type of

Melanoma

Non-Amplified

(SD)

DOP-PCR+QRT-PCR

(SD)

DOP-PCR

(SD)

Metaphase

CGH

15q25 Image EST R78712 (R78712) Primary 1.22 (0.09) 0.76 (0.03) 0.68 (0.25)No change

Metastatic 1.22 (0.17) 0.91 (0.03) 0.65 (0.05)2q31.1 Recepin (U03644) Primary 0.96 (0.00) 0.60 (0.08) 0.82 (0.11)


20q13.12 Image EST 364782 (AA034412) Primary 1.25 (0.11) 0.49 (0.02) 1.11 (0.34)No change

Metastatic 1.40 (0.12) 0.50 (0.07) 1.31 (0.03)3p12 Roundabout, axon guidance

receptor (BC057243)Primary 1.20 (0.06) 1.31 (0.21) 0.80 (0.11) No changeMetastatic 1.13 (0.07) 0.97 (0.24) 2.89 (0.12) Amplification

7p15 CG1 (U97198) Primary 1.14 (0.13) 0.79 (0.05) 1.71 (0.30) No change4p14 Amyloid beta precursor prot.

binding B2Primary 1.03 (0.11) 1.49 (0.7) 0.55 (0.11) Amplification

3q26.31 N-acetylated a-linked acidicdipeptidase 2 (AF04099)

Primary 1.34 (0.09) 1.05 (0.17) 1.65 (0.40) No change

Gains of chromosomal regions7p15 CG1 (U9718) Metastatic 1.52 (0.08) 0.92 (0.00) 2.63 (0.78) No change4p14 Amyloid beta precursor

protein-binding B2 (BC02794)Metastatic 1.50 (0.13) 1.51 (0.08) 0.81 (0.03) Amplification

10q26.13-q26.3 Dedicator of cytokinesis 1(NM 001380)

Metastatic 1.52 (0.07) 0.43 (0.00) 0.86 (0.24) Deletion

3q26.31 N-acetylated a-linked acidicdipeptidase 2 (AF 04099)

Metastatic 1.56 (0.06) 1.67 (0.18) 1.09 (0.15) Amplification

1p36.21 Cardiodilatin-atrial natriureticfactor (AL021155)

Primary 1.80 (0.21) 0.80 (0.04) 1.14 (0.19)Amplification

Metastatic 2.08 (0.20) 0.86 (0.09) 1.48 (0.02)6p25.2 Cell death protein (RIP) (U50062) Primary 1.71 (0.05) 1.60 (0.32) 1.23 (0.16)


18q21.3-q22 UV-B repressed sequence,HUR 7 (X98307)


Metastatic 1.54 (0.12) 2.48 (0.65) 0.17 (0.38)6p12.2 Polycystic kidney, hepatic

disease 1 (AY074797)Primary 1.61 (0.21) 1.64 (0.17) 1.09 (0.10)


17q11.2-q12 Amiloride-sensitive cationchannel 1 (BC075043)


Metastatic 2.02 (0.40) 1.49 (0.12) 0.95 (0.20)17q23 Breast carcinoma amplified

sequence 3 (BC001250)Primary 1.69 (0.06) 1.74 (0.04) 4.23 (1.98)


5q21.3 F-box, leucine-rich repeatprotein 17 (BC063316)


Metastatic 2.64 (1.08) 2.46 (0.67) 1.73 (0.56)

Diagn Mol Pathol � Volume 15, Number 1, March 2006 Whole-Genome Amplification

r 2006 Lippincott Williams & Wilkins 47

REFERENCES1. Zvara A, Hackler L Jr, Nagy ZB, et al. New molecular methods for

classification, diagnosis and therapy prediction of hematologicalmalignancies. Pathol Oncol Res. 2002;8:231–240.

2. Ottesen AM, Skakkebaek NE, Lundsteen C, et al. High-resolutioncomparative genomic hybridization detects extra chromosome arm12p material in most cases of carcinoma in situ adjacent to overtgerm cell tumors, but not before the invasive tumor development.Genes Chromosomes Cancer. 2003;38:117–125.

3. Peng DF, Sugihara H, Mukaisho K, et al. Alterations ofchromosomal copy number during progression of diffuse-typegastric carcinomas: metaphase- and array-based comparativegenomic hybridization analyses of multiple samples from individualtumours. J Pathol. 2003;201:439–450.

4. Grant SF, Steinlicht S, Nentwich U, et al. SNP genotyping on agenome-wide amplified DOP-PCR template. Nucleic Acids Res.2002;30:e125.

5. Kittler R, Stoneking M, Kayser M. A whole genome amplificationmethod to generate long fragments from low quantities of genomicDNA. Anal Biochem. 2002;300:237–244.

6. Barbaux S, Poirier O, Cambien F. Use of degenerate oligonucleotideprimed PCR (DOP-PCR) for the genotyping of low-concentrationDNA samples. J Mol Med. 2001;79:329–332.

7. Simpson DJ, Bicknell EJ, Buch HN, et al. Genome-wide amplifica-tion and allelotyping of sporadic pituitary adenomas identify novelregions of genetic loss. Genes Chromosomes Cancer.2003;37:225–236.

8. Klein CA, Schmidt-Kittler O, Schardt JA, et al. Comparativegenomic hybridization, loss of heterozygosity, and DNA sequenceanalysis of single cells. Proc Natl Acad Sci USA. 1999;96:4494–4499.

9. Ludecke HJ, Senger G, Claussen U, et al. Cloning defined regions ofthe human genome by microdissection of banded chromosomes andenzymatic amplification. Nature. 1999;338:348–350.

10. Saunders RD, Glover DM, Ashburner M, et al. PCR amplificationof DNA microdissected from a single polytene chromosome band: a

comparison with conventional microcloning. Nucleic Acids Res.1989;17:9027–9037.

11. Zhang L, Cui X, Schmitt K, et al. Whole genome amplification froma single cell: implications for genetic analysis. Proc Natl Acad SciUSA. 1992;89:5847–5851.

12. Dietmaier W, Hartmann A, Wallinger S, et al. Multiple mutationanalyses in single tumor cells with improved whole genomeamplification. Am J Pathol. 1999;154:83–95.

13. Telenius H, Carter NP, Bebb CE, et al. Degenerate oligonucleotide-primed PCR: general amplification of target DNA by a singledegenerate primer. Genomics. 1992;13:718–725.

14. Nagy J, Feher LZ, Sonkodi I, et al. A second field metachronousMerkel cell carcinoma of the lip and the palatine tonsil confirmed bymicroarray-based comparative genomic hybridisation. VirchowsArch. 2005;446:278–286.

15. Nelson DL, Ledbetter SA, Corbo L, et al. Alu polymerase chainreaction: a method for rapid isolation of human-specific sequencesfrom complex DNA sources. Proc Natl Acad Sci USA.1989;86:6686–6690.

16. Dean FB, Hosono S, Fang L, et al. Comprehensive human genomeamplification using multiple displacement amplification. Proc NatlAcad Sci USA. 2002;99:5261–5266.

17. Phillips J, Eberwine JH. Antisense RNA amplification: a linearamplification method for analyzing the mRNA population fromsingle living cells. Methods. 1996;10:283–288.

18. Dean FB, Nelson JR, Giesler TL, et al. Rapid amplification ofplasmid and phage DNA using Phi 29 DNA polymerase andmultiply-primed rolling circle amplification. Genome Res.2001;11:1095–1099.

19. Lizardi PM, Huang X, Zhu Z, et al. Mutation detection and single-molecule counting using isothermal rolling-circle amplification. NatGenet. 1998;19:225–232.

20. Puskas LG, Fartmann B, Bottka S. Restricted PCR: amplificationof an individual sequence flanked by a highly repetitiveelement from total human DNA. Nucleic Acids Res. 1994;22:3251–3252.

21. Larsen J, Ottesen AM, Lundsteen C, et al. Optimization ofDOP-PCR amplification of DNA for high-resolutioncomparative genomic hybridization analysis. Cytometry. 2001;44:317–325.

22. Hughes S, Arneson N, Done S, et al. The use of whole genomeamplification in the study of human disease. Prog Biophys Mol Biol.2005;88:173–189.

23. Wang G, Brennan C, Rook M, et al. Balanced-PCR amplificationallows unbiased identification of genomic copy changes in minutecell and tissue samples. Nucleic Acids Res. 2004;32:e76.

24. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: ALaboratory Manual. Cold Spring Harbor, NY: Cold Spring HarborLaboratory Press; 1989.

25. Balazs M, Adam Z, Treszl A, et al. Chromosomal imbalances inprimary and metastatic melanomas revealed by comparativegenomic hybridization. Cytometry. 2001;46:222–232.

26. Nagy ZB, Kelemen JZ, Feher LZ, et al. Real-time polymerase chainreaction-based exponential sample amplification for microarraygene expression profiling. Anal Biochem. 2005;337:76–83.

27. Pfaffl MW. A new mathematical model for relative quantification inreal-time RT-PCR. Nucleic Acids Res. 2001;29:e45.

28. Innis MA, Gelfand DH, Sninsky JJ, et al. Optimization of PCRs. In:Innis MA, ed. PCR Protocols: A Guide to Methods and Applications.New York, NY: Academic Press; 1990:3–12.

29. Sardelli A. Plateau effect—Understanding PCR limitations. Ampli-fications. 1993;9:1–5.

30. Morrison C, Gannon F. The impact of the PCR plateau phase onquantitative PCR. Biochim Biophys Acta. 1994;1219:493–498.

31. Dennis Lo YM. Introduction to the Polymerase Chain Reaction. In:Dennis Lo YM, ed. Methods Mol. Med: Clinical Applications ofPCR. Vol. XVI. Totowa, NJ: Humana Press Inc; 1999:1–20.

FIGURE 2. Scatter plot of relative copy number of genes inprimary and metastatic melanoma samples obtained with 3different protocols. Values of copy number changes are in log2

scale. Only those genes were selected which values obtainedby the any one of protocols were above +0,58 and below �1.The black squares and triangles mean those values which aredifferent from the relative copy numbers of nonamplifiedsamples.



Virchows Arch (2005) 446: 278–286DOI 10.1007/s00428-004-1176-0

ORIGINAL ARTICLE

Judit Nagy . Liliána Z. Fehér . István Sonkodi .József Lesznyák . Béla Iványi . László G. Puskás

A second field metachronous Merkel cell carcinoma of the lipand the palatine tonsil confirmed by microarray-basedcomparative genomic hybridisation

Received: 9 September 2004 / Accepted: 8 November 2004 / Published online: 25 February 2005# Springer-Verlag 2005

Abstract Merkel cell carcinomawas diagnosed in a 79-year-old Caucasian woman. The tumour was localised to the upperlip and was in stage T2. After successful cryosurgery and a 7-year tumour-free period, a new tumour developed in herpalatine tonsil. Histologically and immunohistochemically,this resembled the tumour in the lip. The regional lymphnodes were devoid of metastasis. The paraffin-embeddedmaterial of the two tumours and the unaffected lymphatictissue were analysed with DNA microarrays for comparativegenomic hybridisation to assess the genetic relationship of thetumours. In both tumours, regions on 2p and 10p werecommonly over-represented, while 41 regions on chromo-somes 1–4, 6, 8–9, 11 and 14–22 were commonly under-represented. Chromosomes 1, 3, 4, 16–18 and X were mostfrequently involved in the DNA losses. In gene copy numbersin the two tumours, 31 chromosome locations were found tobe differently affected. The partly similar and partly differentmolecular patterns indicated a genetic relationship betweenthe tumours and excluded the possibility that the tonsillartumour was a metastasis. The findings suggest that a genet-ically altered field was the reason for the development of thetonsillar cancer; thus, it can be regarded pathogenetically as asecond field tumour.

Keywords Merkel cell carcinoma . Second field tumour .Comparative genomic hybridisation . DNA microarray .Chromosome imbalances

Introduction

Merkel cells are found in the skin and in those parts of themucosa derived from the ectoderm [11]. These cells are theorigin of a rare, malignant neuroendocrine tumour thatoccurs predominantly in the sun-exposed areas of the skincalled Merkel cell carcinoma (MCC) [1, 2, 4, 19, 20, 22,28]. Local recurrence and regional or distant metastases gen-erally develop within a short period of time. Oropharyngealmetastasis is very rare, and metastasis to the palatine tonsilhas been described in only one case [27]. A perioral or in-traoral localisation of the MCC is very infrequent: [13, 20] todate, 10 cases of MCC of the lip [35] and 14 cases of intraoralMCCs have been described [3, 16, 20]. During the pastdecade, we treated and followed up an elderly woman withMCC of the upper lip. After a long tumour-free period, ananaplastic carcinomawith neuroendocrine features developedin her palatine tonsil, raising the possibility of a late hae-matogenous metastasis, a second field tumour or a secondprimary tumour.

The term “secondary field tumour” was introduced in1953 by Slaughter et al., who examined slides from patientswith head and neck cancer [29]. It was observed that all ofthe epithelium beyond the margins of the tumours dis-played histological alterations, and 11% of the patientswere found to have more than one independent area ofmalignancy. The authors concluded that the mucosa of thehead and neck had undergone a change, perhaps due tocarcinogen exposure, and was, therefore, more susceptibleto the development of foci of malignant transformation.Organs in which field cancerisation has been describedsince then are the oral cavity, oropharynx, larynx, lung,oesophagus, colon, skin, vulva, cervix, breast, renal pelvis,ureter and bladder [5–7, 10].

The determination of molecular patterns of first and sec-ond tumours has become a valuable tool to explore the

J. Nagy . I. SonkodiDepartment of Oral Medicine, Department of Dentistry andOral Surgery, University of Szeged, Faculty of Medicine,Szeged, Hungary

L. Z. Fehér . L. G. Puskás (*)Laboratory of Functional Genomics, Biological ResearchCentre, Hungarian Academy of Sciences,P.O. Box 521, 6701 Szeged , Hungarye-mail: [email protected].: +36-62-599782Fax: +36-62-432576

J. LesznyákDepartment of Histopathology, Bács-Kiskun County Hospital,Kecskemét, Hungary

B. IványiDepartment of Pathology, University of Szeged,Szeged, Hungary

relationship between them: similar aberrations indicate ametastasis, partly different aberrations indicate a secondfield tumour and different aberrations denote a second pri-mary tumour [5, 6, 10]. In the present study, we identifiedthe molecular patterns with comparative genomic hybrid-isation (CGH) with microarray technology. Array-basedCGH consists of a series of mapped artificial bacterialchromosomes or sequenced cDNA clones on glass slides,to which DNA from test and control samples is hybridised[8, 19–21, 30]. This approach allows the determination ofcopy number changes of chromosome regions without aneed for tissue cultures and the preparation of metaphasespreads, as DNA extracted from a tumour specimen is used.Another major advantage of CGH is that archived speci-mens can be also studied. We applied human cDNA micro-arrays as a tool, DNA from paraffin-embedded tumourmaterials (from both the primary and secondary tumours)as test probes and DNA from normal lymph nodes as con-trol probes.

Previous CGH studies of MCCs have identified divergentregions that are affected, with a small number of similaritiesbetween the samples analysed. Recurrent chromosomal im-balances detected using CGH analysis were loss of 3p, 4p15-pter, 10q, 13q and 17p and gains of 1q, 3q, 5p, 6, 8q, 18 and20 [12, 24, 33, 34]. In the present study, we detected some ofthe previously found regions and also numerous novel re-gions with chromosomal imbalances. Copy number changesof nine chromosome regions detected by CGH were ana-lyzed using real-time quantitative polymerase chain reaction(QPCR). Potential oncogenes, tumour suppressor and apo-ptotic genes mapped to regions detected as changed werealso recorded. The results suggest that a genetically alteredfield was the reason for the development of the tonsillarcancer; thus, it can be regarded pathogenetically as a secondfield tumour.

Materials and methods

Case report

Between 1970 and 2003, 4418 patients with malignantorofacial tumours were treated in our clinic. One of thesepatients suffered from MCC. A 79-year-old Caucasianwoman presented with a tumour in her upper lip. She wasan outdoor worker and never smoked. She was disturbedonly about the aesthetics. The physical examination re-vealed a 2–3-cm firm, compact mass in the skin of theupper lip on the left side. It was sharply separated from itssurroundings and had a teleangiectatic surface (Fig. 1). Thetumour was classified as stage T2. No regional lymph-nodemetastasis was detected. Surgical resection of the tumourand removal of the regional lymph nodes were suggested,but she refused it. As an alternate therapy, a biopsy wastaken, and cryotherapy was applied according to the stan-dard protocol. Light microscopy revealed that the tumourwas confined to the dermis and was sharply separated fromthe epidermis (Fig. 2a). The tumour cells were arrangedin nests and cords. Cytologically, the tumour cells were

monomorphic, the nuclei were round and the chromatinpattern was finely granular and “dusty”. The nucleoli weresmall; often two or even three were detected. The cytoplasmwas scanty. The mitotic rate was high (six to ten figures/high-power field). The Grimelius staining was negative.The results of immunostainings are shown in Table 1 andFig. 2. The histological and immunohistochemical featurespointed to MCC of the lip. Following the histologicaldiagnosis, distant metastases in the lungs and bones weresearched for, with negative results.

In the 7-year follow-up, the patient was in a tumour-freecondition. After 7 years, she was admitted to the CountyHospital with a left palatine tonsillar tumour. This tumourwas resected in part, and the regional lymph nodes wereremoved. Histologically, the tumour displayed features ofanaplastic carcinoma (Fig. 3). The results of immunostain-ings are summarised in Table 1. The histological andimmunohistochemical features pointed to MCC of the lip.Following the histological diagnosis, distant metastases inthe lungs and bones were searched for, with negativeresults.

The morphological appearance of the tumour resembledthat of the MCC of the lip. A further excision was per-formed 3 months later. The morphological and immu-nohistochemical features of the tonsillar tumour wereessentially the same as observed 3 months earlier. Thepatient died at home on the 20th postoperative day. Au-topsy was not performed.

Sample preparation, labelling

Paraffin-embedded tissues (200–500 mg) were deparaffin-ised in hexane and washed with ethanol. DNAwas purified

Fig. 1 Tumourous mass sharply separated from the surroundings inthe skin of the upper lip

279

by using the DNA purification kit from Macherey-Nagel(Düren, Germany) according to the manufacturer’s instruc-tions. First, 100 ng genomic DNA from each sample wasamplified with a modified version of the DOP (degenerateoligonucleotide primed) PCR protocol using a real-timePCR instrument to follow the amplification [9, 15]. Re-actions were performed in a total volume of 100 μl. Thecycling parameters were as follows: heat start at 95°C for4 min, 8 cycles of 94°C denaturation for 50 s, 45°C an-nealing for 2 min to 72°C with 0.2°C/s and extension at72°C for 90 s. After the 8 cycles, the reaction mix wasseparated into two 50-μl portions, and SybrGreen wasadded to the mix (1× final concentration, Molecular

Table 1 Immunohistochemical findings

Immunohistochemical markers Lip tumour Tonsillar tumour

AE1/AE3 +++ +++Cytokeratin 20 20Paranuclear dots +++ ++Chromogranin ++ +Synaptophysin – –TTF – –HMB–45 – –CEA – –Lymphoid markers – –

Fig. 2 Merkel cell carcinomaof the lip. A Small, monomor-phic tumour cells fill the dermisarranged in nests and cords. Thenuclei exhibit moulding. Hae-matoxylin-eosin ×200. B Chro-mogranin positivity of tumourcells ×400. C Paranuclear dotsby cytokeratin-20 staining ×400

280

Probes, Eugene, USA). The following cycling protocol wasperformed in a real-time PCR instrument, RotorGene 2000(RotorGene, Sydney, Australia): denaturation at 95°C for40 s, annealing at 58°C for 1 min and extension at 72°C for80 s. The cycling was performed just before the reactionreached the plateau to avoid over-amplification of theproducts (usually in 18–22 cycles). The PCR products werepurified on PCR-purification columns (Bioneer, Daejeon,Korea). Next, 100 ng purified DNA was labelled withanother round of PCR in the presence of Cy3-dUCTP (0.05mM) in a total volume of 50 μl with the same parametersfor 20 cycles as in the second PCR. In all these reactions,UN primer (5′-CCGACTCGAGNNNNNNATGTGG-3′)was used in a 1-μM concentration [31]. The reactionswere performed with ExTaq DNA polymerase in 1× buffer(Takara, Tokyo, Japan). The labelled PCR products werepurified with the PCR purification kit (Bioneer). The elutedDNA was dried in a Speed-Vacuum.

Hybridisation, scanning, data processing

The labelled DNA was reconstituted in ChipHyb hybrid-isation buffer (Ventana Discovery, Tuchon, USA) contain-ing 20 μg human Cot DNA and 5 μg salmon sperm DNA(Invitrogen). Hybridisation was performed on humancDNA microarrays having 3200 gene-specific samples induplicate in 200 μl using the Ventana hybridisation station(Ventana) at 42°C for 8 h [25, 26]. After hybridisation, theslides were washed twice in 0.2×sodium saline citrate atroom temperature for 10 min and then dried. Scanning anddata analysis were performed as described previously [21].Briefly, data presented in this study were calculated fromthe results of four data points obtained from two separatelabelling and hybridisation protocols. The background-corrected intensity data was filtered for flagged spots andweak signal. Technical replicates on the same array wereaveraged. Data were excluded in cases where technicalreplicates were significantly different. Normalisation was

Fig. 3 A Anaplastic carcinomaof the tonsil. Hematoxylin-eosin×200. B Chromogranin positiv-ity in the tumour cells ×400. CSome tumour cells displayparanuclear dots by cytokeratin-20 staining ×400

281

performed using the print-tip LOWESS method [36]. Next,we used the one-sample t-test to determine the genes to beregarded as changed in copy number. Logarithm was takenfrom each ratio to fulfil the t-test’s requirement for a normaldistribution. Genes for which the mean of log-ratios acrossthe biological replicates was equal to zero at a significancelevel α=0.05 were considered to have an unchanged copynumber. However, genes with a P value smaller than α andwhere the average-fold change (over- or under-representa-tion) of the four data points were at least twofold wereconsidered as changes in DNA copy number.

Real-time QPCR

The confirmatory real-time QPCR was performed on aRotorGene 2000 instrument with gene-specific primers andthe SybrGreen protocol on non-amplified genomic DNA.Reactions were performed in a total volume of 20 μl (50 nggenomic DNA from each sample, 5 pmol/each forward andreverse primer, 1× BioRad SYBRGreen buffer, BioRad,Hungary) with the following protocol: denaturation for 10min at 95°C, and 45 cycles of 25 s denaturation at 95°C, 25s annealing at 59°C and 25 s extension at 72°C. Fluorescentsignals were collected after each extension step at 72°C.Curves were analysed with the RotorGene software, usingdynamic tube and slope correction methods, ignoring datafrom cycles close to baseline. Primers were designed byusing the ArrayExpress software (Applied Biosystems).Relative ratios were normalised to the Ct values obtainedwith the dihydrofolate reductase gene probe and calculatedwith the Pfaffl method [23]. The PCR primers used in thisstudy are listed in Table 2. All the PCRs were performedfour times in separate runs.

Results

CGH analysis

We analysed the primary MCC and the tumour of the tonsil,applying CGH with DNA from paraffin-embedded tumourmaterial as the probes. Relative DNA losses and gains weredetermined for each tumour sample by normalising theintensities to the values obtained after hybridisation withlabelled probes from normal lymphoid tissue.

Using microarray-based CGH, numerous changes inchromosome copy numbers were observed in both tumoursinvestigated. Both tumours showed complex DNA copynumber changes, with an abundance of DNA losses and afew DNA gains. Of the regions, 41 were detected with aDNA loss and 4 regions with a DNA gain (Table 3). CGHrevealed regions on 2p and 10p to be commonly over-represented and regions on 1p, 1q, 2p, 2q, 3p, 3q, 4q, 5q,6p, 6q, 7q, 8p, 9p, 9q, 11p, 11q, 12q, 14q, 15q, 16p, 16q,17p, 17q, 18q, 19p, 20p, 21q, 22q and X to be commonlyunder-represented. Chromosomes 1, 2, 3, 17 and 18 weremost frequently involved in the DNA losses. According tothe intensity ratios, monosomy of the X chromosome ispostulated. Besides the commonly affected regions, 31 addi-tional chromosome locations were found to be differentlyaffected in gene copy numbers in the two tumours. Common,primaryMCC-specificandsecondarytumour-specificchangesare listed in Table 3.While 22 regions could be observed withthe secondary tonsil cancer-specific DNA loss, only 6 regionswere specifically under-represented in the case of the primaryMCC. All other DNA losses were common in both tumours.Only2 regions, 2p23and10p15, exhibitedcommongains, and3regions,2p25,12q12and15q14,showedover-representationin the case ofMCC.

Potential oncogenes and tumour suppressor and apo-ptotic genes mapped to regions detected as changed were

Table 2 Sequences of the oligonucleotides (5′-3′) used in this study

Gene Chromosomelocation

Forward primer Reverse primer Productsize (bp)

Steroid 5-alpha-reductase 2 2p23 CAGAAGCCCCAAGCAACTTT CCTTCTTGAACAGGTCCTGAGAA 69Hepatocyte nuclearfactor-3 alpha

14q12-q13 CTCAAGAGTTGCTTGACCGAAA GGGCCATCTGTGGGTAGAGA 67

Zinc finger protein 19q13.43 GAAATTTCCCTGCCAGACCTT CATGAAGCCCCACTCTGAGAGT 73Androgen receptor(dihydro-testosterone receptor)

Xq12 CGGAAATGATGGCAGAGATCA TGGGCTTGACTTTCCCAGAA 66

C8FW phosphoprotein 8q24.13 GGACGATACCCCTTCCATGA GTCCACGCCGAATTTTGG 61Cardiac myosin bindingprotein-C

11p11.2 GCCTAAATCCGAGCATCTGTTT TGCACTCTCAGGGAATTTGAGA 70

EST 15q14 CCTGATTCCAGAAAAGCAAGTGT CACTGAGATTACCGGGCATGA 76Cytochrome P450,subfamily XIA

15q22 CTGGTGCAAGTGGCCATCTA AATTTTCCGGGTCGAAGAAGA 64

EST similar tophosphatidylcholintransfer protein

17q23 CACAAGGCTATGCACAAAGCA GGAAACTGAGGCGTCAAGATG 68

Dihydrofolate reductase 5q14 CTGTCATGGTTGGTTCGCTAAA TGCCGATGCCCATGTTC 60

282

also analysed. The genes mapped to regions that werechanged in both tumour samples are listed in Table 4,while changes characteristic of only one tumour are pre-sented in Table 5.

Confirmation of CGH data

The copy number changes of several chromosome regionsdetected using CGH were submitted for QPCR analysis.Specific primers were designed (Table 2) and used toamplify affected DNA regions of the genome using non-amplified genomic DNA as template. Relative ratios werenormalised to the copy numbers of the dihydrofolate reduc-tase gene, because this did not show any alteration in copynumber in either tumour and gave reproducible results in allcases. Nine regions exhibiting DNA losses or gains in bothtumours were selected for QPCR. In seven cases, the com-mon alterations were confirmed, while in two cases, onlyMCC-specific changes could be detected (Table 6).

Discussion

MCCs are highly divergent when analysed for chromosomeaberrations [12, 24, 34]. Reported recurrent chromosomalimbalances detected using CGH analysis were loss of 3p,10q, 13q and 17p and gains of 1q, 3q, 5p and 8q [34]. In thepresent study, we could also detect the loss of 3p and 17p,but none of the above-mentioned amplified regions couldbe confirmed (Table 3, Fig. 4). In another study, only 3 ofthe 10 MCCs exhibited common gains and losses, and they

shared a gain of 8q21-q22 and a loss of 4p15-pter [24]. Inthe present study, the 4p16 region was also found to bedeleted in both the primary and secondary tumours.Unfortunately, to date, no known tumour suppressor genehas been mapped to this region (Table 4). Speleman’s groupdetected losses involving the entire chromosome 10 or

Table 3 Common and individual gains and losses of Merkel cellcarcinoma and the tonsillar tumour detected using comparativegenomic hybridization

Deletion Amplification

Both tumours 1p36, 1p31, 1p13, 1q21–23, 1q32,1q42.13, 2p13, 2p11.2,2q35, 3p21, 3q13.2, 3q14.3, 3q26–28, 4p16, 4q21, 5q35,6p21, 6q24, 7q21, 8p21, 8q24, 9p12,9q34, 11p11.2, 11q13,11q23.3, 12q24.1, 14q11.2, 15q23–24, 16p13–12, 16q24,17p13, 17q12, 17q23, 18q12, 18q21,19p13, 20p13, 21q22,22q13.1, X monosomy

2p23, 10p15

Merkel cellcarcinoma

11p11.2, 11q12.3, 12q13, 16p13.1,17q21.1, 17q25

2p25, 12q12,15q14

Mesopharynx 1p34.1–32.2, 1p21–22, 1q42, 2q37,4p16.3, 4q28.3, 5q11.1,5q22, 5q31–32, 7p14, 8p22–24.13,10q21.1, 11pter-p15.5,12q14.3, 14q11.2, 14q31–32, 15q11–13, 16q23–24, 17q21.3-q23,20q13.1, 21q22.13, 22q11

–

Table 4 Apoptotic genes and tumour suppressors mapped toregions with common chromosomal aberrations. Apoptotic andtumour suppressor genes in bold character. Apoptotic and tumoursuppressor-related genes in italics in brackets

Chromosomelocation

Apoptotic genes Tumour suppressors

In bothtumours

1p36 DFFB, TP73, CASP9(CDC42, MFN2)

UBE4B, TUSC3, PRDM 2,C1orf1, ALPL (E2F2,TP73)

1p31 – CLCA21p13 – ST7L1q21–23 MCL1, DAP3,

TNFSF6 (MUC1,APCS, CRP, SELL)

2p13 MAD3p21.3 (CCR3) RASSF1, BAP1 (SEMA3B)3q26–28 (ETV5, OPA1) TSBF14p16 (GAK)4q21 (SNCA)5q35 (PDLIM7)6p21 BAK1, TNF (HSPA) (CDKN1A)6q24 SASH1 (PLAGL1)8p21 BNIP3 (CLU) PDGFRL (CNOT7,

PDLIM2)8q24 (RNF139)9q34 TRAF2 (ABL1)11p11.2 (MADD)11q13 DPF2, BAD, FADD

(RELA, CCND1,RAD9A)

DOC-1R, MEN1 (ST3,CCND1, SHANK2)

11q23.3 THY112q24.1 (SCA2)16p13–12 ASC TSC216q24 WFDC1, GAS8,

MGC15419, CBFA2T317p13 TP53, HIC1, OVCA2,

DPH2L117q12 (CCR7) (MPP3)18q21 (BCL2) DCC, SERPINB5 (E2F5)19p13 DIRAS1 (TJP3, SMARCA4,

GIPC3)21q22 (MX1)22q13.1 (HMOX1) ST13Xp22.1 (SH3KBP1)Xq12 ING2Xq28 RPL10

283

partial loss of the chromosome 10 long arm in one-third ofthe MCC cases they analysed with a possible loss ofheterozygosity of 10q23, including the PTEN tumour-suppressor gene [33]. In our case, deletion of this regionwas not observed in either tumour sample.

In previous observations, series of MCCs showed noevidence of high-level amplification [34]. Recurrent gainsof chromosomes 1, 6, 18q and 20 were detected in 2 cases[12]. In our case, only two regions, 2p and 10p were com-monly over-represented. In a very recent study, 19 primaryMCCs were analysed by CGH, and, in 13 samples, ex-tensive gains and losses were detected [17]. It was shownthat a majority of the alterations were gains, while only a

few common losses were detected, mainly in regions 4q,13q and 16q. Neither of our gains was found in any of the19 cases they analysed. Most of the losses were detected inat least 1 case reported in the above-mentioned study, butthe 13 MCCs exhibited a very heterogeneous pattern, withdiverse regions with losses.

Our CGH results revealed several new and a few otherpreviously known chromosomal regions that have beenpresumed to be involved in the pathogenesis of MCC. Wefound 2 common gains and 41 common losses in the twosamples. However, in 31 chromosome locations, we ob-served differences in gene copy numbers between the twotumours. From the results obtained by CGH analysis, webelieve that the mesopharynx cancer and the MCC of thelip derived from the same, genetically altered field; thus,the mesopharynx cancer can be regarded as a second fieldtumour. From the results obtained using CGH analysis, wehypothesise that Merkel cells in two adjacent anatomicalsites, i.e. in the lip and mesopharynx, underwent sameprecancerous genetic alteration, and both tumours arosefrom a common cell clone. If our hypothesis is correct, themesopharynx cancer can be regarded as a second fieldtumour.

Several oncogenes and tumour-suppressor and apoptoticgenes were assumed to be changed in their copy number, asmany of them were mapped to the regions having changesin their copy number in one or both of the tumours (Table 4and Table 5). The limitation of this list is that, in thesecases, the deletions of the genes themselves were notproved in this study; only those regions were determinedthat were located to certain regions or proximity to regionswhere DNA segments on the microarrays originated.

The microarray-based CGH techniques used in this studycould result in distorted data, especially when paraffin-embedded tissue is the target. Another limitation of thisstudy is that the deleted regions were determined usinghybridisations based on complementary cDNA—genomicDNA annealing events, which could generate more biases.In consequence of the above problems, we determined thereliability of the results obtained by the CGH microarrayusing real-time QPCR. Ten genes were selected to follow

Table 5 Apoptotic genes and tumour suppressors mapped to regionswith chromosomal aberrations specific to Merkel cell carcinoma ortonsillar tumour. Apoptotic and tumour suppressor genes in boldcharacter. Apoptotic and tumour suppressor related genes in italics inbrackets

Chromosomelocation

Apoptotic genes Tumoursuppressors

In Merkel cell carcinoma11p11.2 (MADD)12q13 (CDK2, KRT18, NR4A1,

WNT1)17q25 BIRC5In mesopharynx1p34.1–32.2 FABP33p22–21 ADPRTL35q11.1 (LOC145908)5q22 MCC5q31–32 C5orf710q21.1 (CDC2)11pter-p15.5 (CTSD, HRAS) MRVI114q31–32 SIVA16q23–24 (LOC145908) CBFA2T321q22.13 RUNX122q11 BID MYO18B

Table 6 Confirmation of comparative genomic hybridisation data using quantitative real-time polymerase chain reaction (QPCR). Onlytwo common losses could not be confirmed. Gains are indicated as grey background

Microarray data confirmed usingQPCR

Clone name Accessionno.

Chromosomelocation

Distance

Both tumours Steroid 5-alpha-reductase 2 (SRD5A2) M74047 2p23 31724Hepatocyte nuclear factor-3 alpha U39840 14q12-q13 36052Zinc finger protein AF020591 19q13.43 63450Androgen receptor (dihydrotestosterone receptor) M23263 Xq12 63075C8FW phosphoprotein AJ000480 8q24.13 126404ESTs, Highly similar to Phosohatidylcholine transferprotein

AA933627 17q23 54340

Cytochrome P450, subfamily XIA M14565 15q23–24 70670Merkel cell carcinoma EST N22765 15q14 no data

Cardiac myosin binding protein-C U91629 11p11.2 49310

284

the copy number changes by QPCR. Of the ten genes, oneexhibited no changes in the copy number found using CGHmicroarray analysis in all cases: the dihydrofolate reductasegene. Therefore, it was used as an internal control. Thecopy numbers of all the other nine genes (sequences andchromosome locations are listed in Table 2, accessionnumbers are listed in Table 6) were related to this innercontrol. We used DNA purified from normal lymphoidtissue as control and determined copy number changes ofthe other two tumours (Table 3, Fig. 4). We confirmed thedeletions in seven cases for both tumours, and two wereconfirmed for only MCC. Although the number of theconfirmatory real-time QPCRs was limited and, therefore,does not allow determine the reliability of the CGHmethods, the main purpose of the study was to determinethe relation of the two tumours, not a precise determinationof the deleted regions.

In summary, the partly different molecular patternsobtained using CGH, the similar histological features andthe close proximity to the primary tumour indicate that thetonsillar cancer was a second field tumour. The commonorigin was further confirmed, because of three early markers(9p, 3p and 17p) two of them (9p and 17p) were common inboth cancer samples, one (17p) bearing the tp53 tumoursuppressor marker gene. The common copy number changesdo not support the possibility that the tonsillar cancer was asecond primary tumour. The tonsils are very rare and un-usual sites of metastatic disease in MCC. There are threereports in the literature describing oropharyngeal metastasisof a MCC [18, 27, 31]. In all these reports, the metastasisoccurred within 24 months after resection of the primarytumour. In our case, it was clinically very unlikely that thetonsillar tumour was a metastasis of the MCC of the lip,because no local recurrence, regional lymph-node metastasisor distant haematogenous metastases were observed in the7-year follow-up. In harmony with the clinical situation,the molecular patterns of the two tumours were not similar,therefore, the possibility of a metastasis could be rejected.Although the concept of second primary tumours and field

cancerisation has been formulated for oral and oropharyn-geal squamous cell carcinomas [5], our results indicate thatMerkel cells in adjacent anatomic sites may also be thetargets of field cancerisation.

Acknowledgements This work was supported by a grant from theHungarian Ministry of Health (ETT 400/2003). LGP is supported bythe János Bolyai fellowship of the Hungarian Academy of Sciences.

References

1. Antoniades K, Giannouli TH, Kaisaridou D (1988) Merkel cellcarcinoma in a patient with Recklinghausen neurofibromatosis.Int J Oral Maxillofac Surg 27:213–214

2. Bellome J, Bays DR (1998) Merkel cell carcinoma of the ear:report of case. J Oral Maxillofac Surg 56:984–988

3. Benning TL, Vollmer RT, Crain BJ, Shelburne JD (1990)Neuroendocrine carcinoma of the oral cavity. Mod Pathol 3:631

4. Bickle K, Glass LF, Messina JL, Fenske NA, Siegrist K (2004)Merkel cell carcinoma: a clinical, histopathologic, and immu-nohistochemical review. Semin Cutan Med Surg 23:46–53

5. Braakhuis BJ, Tabor MP, Leemans CR, van der Waal I, SnowGB, Brakenhoff RH (2002) Second primary tumors and fieldcancerization in oral and oropharyngeal cancer: moleculartechniques provide new insights and definitions. Head Neck24:198–206

6. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brake-nhoff RH (2004) A genetic explanation of Slughter’s concept offield cancerization. Cancer Res 63:1727–1730

7. Califano J, van der Riet P, Westra W, Nawroz H, Clayman G,Piantadosi S, Corio R, Lee D, Greenberg B, Koch W, SidranskyD (1996) Genetic progression model for head and neck cancer:implications for field cancerization. Cancer Res 56:2488–2492

8. Cowell JK, Wang YD, Head K, Conroy J, McQuaid D, NowakNJ (2004) Identification and characterisation of constitutionalchromosome abnormalities using arrays of bacterial artificialchromosomes. Br J Cancer 90:860–865

9. Furák J, Troján I, Szőke T, Tiszlavicz L, Bodosi M, Kahán ZS,Micsik T, Puskás LG, Balogh Á (2002) Development of brainmetastasis 5 years before the appearance of the primary lungcancer: messenger metachronous metastasis. Ann Thorac Surg75:1016–1017

Fig. 4 Copy number aberra-tions found using microarraycomparative genomic hybridisa-tion analysis and real-timepolymerase chain reaction ofprimary Merkel cell carcinoma(MCC) and the tonsillar cancer.Boxes on the left side of eachchromosome ideogram showregions of reduced copy number(losses of DNA in the tumourgenome). Circles on the left sideof each chromosome ideogramshow regions of increased copynumber (gains of DNA in thetumour genome). Filled boxesdenote MCC and empty boxestonsillar tumour

285

10. Ha PK, Califano JA (2003) The molecular biology of mucosalfield cancerization of the head and neck. Crit Rev Oral BiolMed 14:363–369

11. Halata Z, Grim M, Baumann KI (2003) Friedrich SigmundMerkel and his “Merkel cell”, morphology, development, andphysiology: review and new results. Anat Rec 271:225–239

12. Harle M, Arens N, Moll I, Back W, Schulz T, Scherthan H(1996) Comparative genomic hybridization (CGH) discloseschromosomal and subchromosomal copy number changes inMerkel cell carcinomas. J Cutan Pathol 23:391–397

13. Hashimoto K (1972) Fine structure of Merkel cell in humanoral mucosa. J Invest Dermatol 58:381–387

14. Hodgson G, Hager JH, Volik S, Hariono S, Wernick M, MooreD, Nowak N, Albertson DG, Pinkel D, Collins C, Hanahan D,Gray JW (2001) Genome scanning with array CGH delineatesregional alterations in mouse islet carcinomas. Nat Genet29:459–464

15. Huang Q, Schantz SP, Rao PH, Mo J, McCormick SA,Chaganti RS (2000) Improving degenerate oligonucleotideprimed PCR-comparative genomic hybridization for analysis ofDNA copy number changes in tumours. Genes ChromosomesCancer 28:395–403

16. Koss LG, Spiro RH, Hajdu S (1972) Small cell (oat cell)carcinoma of minor salivary gland origin. Cancer 30:737

17. Larramendy ML, Koljonen V, Bohling T, Tukiainen E, KnuutilaS (2004) Recurrent DNA copy number changes revealed bycomparative genomic hybridization in primary Merkel cellcarcinomas. Mod Pathol 17:561–567

18. Lehnerdt KH, Broicher R, Tschubel K, Walther E (2001) Derinteressante Fall Nr.49. Laryngorhinootologie 80:687–689

19. Longo F, Califano L, Mangone GM, Errico ME (1999)Neuroendocrine (Merkel cell) carcinoma of the oral mucosa:report of case with immunohistochemical study and review ofthe literature. J Oral Pathol Med 28:88–91

20. Mir R, Sciubba JJ, Bhuiya TA, Blomquist K, Zelig D, FriedmanE (1988) Merkel cell carcinoma arising in the oral mucosa. OralSurg Oral Med Oral Pathol 65:71–75

21. Ónody A, Zvara Á, Hackler Jr L, Vígh L, Ferdinandy P, PuskásLG (2003) Effect of classic preconditioning on the geneexpression pattern of rat hearts: a DNA microarray study. FEBSLetters 536:35–40

22. Park GC, Shelton JB, Ow RA, Todd DH (1999) Merkel cellcarcinoma of the lower lip. Arch Otolaryngol Head Neck Surg125:907–911

23. Pfaffl MW (2001) A new mathematical model for relativequantification in real-time RT-PCR. Nucleic Acids Res 29:e45

24. Popp S, Waltering S, Herbst C, Moll I, Boukamp P (2002) UV-B-type mutations and chromosomal imbalances indicatecommon pathways for the development of Merkel and skinsquamous cell carcinomas. Int J Cancer 99:352–360

25. Puskás LG, Hackler L Jr, Kovács G, Kupihár Z, Zvara Á,Micsik T, van Hummelen P (2002) Recovery of cyanine-dyenucleotide triphosphates. Anal Biochem 305:279–281

26. Puskás LG, Zvara Á, Hackler L Jr, Micsik T, van Hummelen P(2002) Production of bulk amounts of universal RNA for DNAmicroarrays. Biotechniques 33:898–904

27. Reichel OA, Mayr D, Issing WJ (2003) Oropharyngealmetastasis of a Merkel cell carcinoma of the skin. Eur ArchOtorhinolaryngol 260:258–60

28. Schmidt-Westhausen A, Reichert PA, Gross UM (1996)Gingival metastasis of Merkel cell carcinoma: a case report. JOral Pathol Med 25:44–47

29. Slaughter DP, Southwick HW, Smejkal W (1953) “Fieldcancerization” in oral stratified squamous epithelium. Cancer6:963–968

30. Snijders AM, Nowak N, Segraves R, Blackwood S, Brown N,Conroy J, Hamilton G, Hindle AK, Huey B, Kimura K, Law S,Myambo K, Palmer J, Ylstra B, Yue JP, Gray JW, Jain AN,Pinkel D, Albertson DG (2001) Assembly of microarrays forgenome-wide measurement of DNA copy number. Nat Genet29:263–264

31. Telenius H, Carter NP, Bebb CE, Nordenskjold M, Ponder BA,Tunnacliffe A (1992) Degenerate oligonucleotide-primed PCR:general amplification of target DNA by a single degenerateprimer. Genomics 13:718–725

32. Tesei F, Farneti G, Cavicchi O, Antonelli P, Zanetti G, Leone O(1992) View from beneath: pathology in focus. A case ofMerkel cell carcinoma metastatic to tonsil. J Laryngol Otol106:1100–1102

33. Van Gele M, Leonard JH, Van Roy N, Cook AL, De Paepe A,Speleman F (2001) Frequent allelic loss at 10q23 but lowincidence of PTEN mutations in Merkel cell carcinoma. Int JCancer 92:409–413

34. Van Gele M, Leonard JH, Van Roy N, Van Limbergen H, VanBelle S, Cocquyt V, Salwen H, De Paepe A, Speleman F (2002)Combined karyotyping, CGH and M-FISH analysis allowsdetailed characterization of unidentified chromosomal rearran-gements in Merkel cell carcinoma. Int J Cancer 101:137–145

35. Vigneswaren N, Müller S, Lense E, Stacey B, Hewan-Lowe K,Weathers DR (1992) Merkel cell carcinoma of the labialmucosa. Oral Surg Oral Med Oral Pathol 74:193–200

36. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, Speed TP(2002) Normalization for cDNA microarray data: a robustcomposite method addressing single and multiple slide sys-tematic variation. Nucleic Acids Res 30:e15

286

Parallel expression of aIIbb3 and avb3 integrins in human melanoma cells

upregulates bFGF expression and promotes their angiogenic phenotype

Balazs Dome1, Erzsebet Raso1, Judit Dobos1, Livia Meszaros1, Norbert Varga1, Laszlo G. Puskas2, Laszlo Z. Feher2,Tamar Lorincz1, Andrea Ladanyi1, Mohit Trikha3, Kenneth V. Honn3 and Jozsef Tımar1*

1Department of Tumor Progression, National Institute of Oncology, Budapest, Hungary2Laboratory of Functional Genomics, Biological Research Center, Hungarian Academy of Sciences, Szeged, Hungary3Department of Radiation Oncology, Wayne State University, Detroit, MI, USA

Previous studies indicated that transfection of the platelet integrinaIIbb3 into human melanoma cells expressing integrin avb3 pro-moted their in vivo (but not in vitro) growth and cell survival. Toreveal the underlying pathomechanism, we have analyzed theangiogenic phenotype of aIIbb3 integrin-transduced human mela-noma cells expressing integrin avb3. Upon heterotopic or ortho-topic (intracutaneous) injections into SCID mice, the aIIbb3integrin-overexpressing clones, ESL, ESH, 19L and 19H, grewmore rapidly than the mock transfectant (avb3 expressing) clone,3.1P. Morphometry demonstrated an increased intratumoralmicrovessel density in 19L and 19H tumors compared to 3.1P.Immunocytochemistry and flow cytometry indicated that vascularendothelial growth factor (VEGF) is constitutively expressed inthe majority of the cells of both the mock and the aIIbb3 integrin-transfected clones. However, the mock transfectant clone, 3.1P,did not express basic fibroblast growth factor (bFGF) at proteinlevel (<1%), unlike the aIIbb3 integrin-transfected clones, 19Land 19H, (33.9 and 84.1%, respectively). Quantitative PCR analy-sis of 6 related human melanoma clones with various levels ofaIIbb3 integrin expressions revealed a correlation between theaIIb protein and bFGF mRNA expressions. Furthermore, cDNAmicroarray analysis of the 19H cells revealed 12 downregulatedand 36 upregulated genes [among them 3 upregulated vasculo-genic mimicry-genes (CD34, endothelin receptor B, ProstaglandinI-2 synthase)] when compared to 3.1P cells. The altered bFGFexpression may be influenced by integrin-linked signaling, sinceb3-endonexin is upregulated in aIIbb3-transfected cells and tyro-sine kinase inhibitors downregulate bFGF both at mRNA and pro-tein levels. We propose here that the illegitimate expression ofaIIbb3 integrin in human melanoma cells already expressingavb3 integrin may alter their in vivo growth properties due to themodulation of their angiogenic phenotype.' 2005 Wiley-Liss, Inc.

Key words: �3 integrins; bFGF; vascularization; human melanoma;xenograft

Progression of human melanoma is a highly efficient processcompared to other malignant tumors since a few millimeters thicktumor (a very small primary in the case of other malignancies) cancolonize regional lymph nodes or visceral organs. It is thereforeextremely important to understand the molecular mechanismsbehind this highly aggressive behavior. Previous studies indicatedthat the autocrine growth regulation of melanoma acquired at theswitch from a less invasive radial to a more invasive verticalgrowth phase involves bFGF expression.1,2 On the other hand,overexpression of �v�3 integrin is also associated with the inva-sive growth3,4 and metastasis5,6 of melanoma. �IIb�3 integrin(GpIIbIIIa) is the predominant adhesion receptor of platelets andthe expression of the integrin �IIb chain is megakariocyte-spe-cific. It is involved primarily in platelet activation, since it isexpressed in a low affinity state and binds fibrinogen only uponactivation. We have previously shown the illegitimate expressionof �IIb�3 integrin in human melanoma,7,8 involved in the samephases of tumor progression as �v�3. Recently we found that thetransduction of �IIb�3 into �v�3-expressing human melanomacells did not affect the in vitro but promoted the in vivo growth oftumor cells due to decreased apoptosis.8 We have postulated thatone possible factor behind the increased in vivo growth is vascula-rization. Therefore, we have analyzed the gene expression changes

of �IIb�3-transfected human melanoma clones with special atten-tion to their angiogenic phenotype.

Material and methods

Human melanoma cells

The WM983B cell line that does not express �IIb�3 on the cellsurface was a kind gift of M. Herlyn (The Wistar Institute, Phila-delphia, PA). Mock transfected WM983B cells (3.1P) and �IIband �3 transfected WM983B cells (19L and 19H) have beendescribed earlier.8 New �IIb and �3 transfected WM983B cellclones, ESL and ESH, have been isolated as described before.8

19H transfected cell line was subcloned by limited dilution toobtain 7D7 and 8F3 subclones. The parental cell line was grownin RPMI-1640 medium supplemented with 5% FBS and antibiot-ics (Sigma Chemical Co., St. Louis, MO) in a 5% CO2 humidifiedatmosphere at 378C. Transfected cells were maintained in mediumcontaining G418 (Gibco BRL, Gaithersburg, MD).

Tumor growth in SCID mice: Orthotopic growth

Five SCID mice/group were anaesthetized by intraperitonealadministration of Nembutal (70 mg/kg), the thighs of the animalswere shaved and 105 tumor cells in 5 �l Hank’s balanced salt solu-tion (HBSS) were inoculated intradermally, under a dissectingmicroscope by using a precision syringe (diameter of 0.15 mm,Hamilton). Volume measurements of palpable tumors were madeon day 10 and 13 by using a precision dermatometer (length �width2 � p/6).

Lung colonization

Tumor cells (106/200 �l) were injected into the tail vein ofSCID mice and after 60 days mice were terminated by Nembutaloverdose and lungs were isolated, fixed and surface colonies werecounted under a stereomicroscope.

Liver colonization

Tumor cells (106/200 �l) were injected into the spleen of SCIDmice and after 30 days mice were terminated by Nembutal over-dose and the liver was isolated, fixed and surface colonies werecounted under a stereomicroscope.

Grant sponsor: Hungarian National Science Fund; Grant numbers:OTKA T38128; Grant sponsor: Ministry of Education; Grant number: 1/48/2001; Grant sponsor: T. Fox Fund; Grant sponsor: Hungarian PrimeMinister Office and Hungarian Academy of Sciences; Grant number: 4676/2003; Grant sponsor: NIH; Grant number: CA47115; Grant sponsor:NATO; Grant number: CLG 975187.*Correspondence to: Department of Tumor Progression, National Insti-

tute of Oncology, Rath Gy. u. 7-9, Budapest, H-1122 Hungary.Fax: þ36-1-224-8706. E-mail: [email protected] 9 February 2004; Accepted after revision 20 December 2004DOI 10.1002/ijc.20991Published online 10March 2005 inWiley InterScience (www.interscience.

wiley.com).

Int. J. Cancer: 116, 27–35 (2005)

' 2005 Wiley-Liss, Inc.

Publication of the International Union Against Cancer

Intracardiac injection

Tumor cells (106/200 �l) were injected into the left cardiac ven-tricle of SCID mice under anesthesia and after 60 days mice wereterminated by Nembutal overdose and visceral organs (lung, liverand brain as well as bones of the lower extremities) were isolated,fixed and surface colonies (visible colonies in case of bones) werecounted under a stereomicroscope. To confirm metastases, speci-mens were embedded into paraffin and H&E stained sections werealso analyzed microscopically.

Immunohistochemistry

On day 13, 3 tumors of each group were removed and frozen inisopentane cooled in liquid nitrogen. Five micrometer frozen sec-

tions were fixed in methanol (10 min, �208C), blocked in 0.1%bovine serum albumin (BSA) and immunostained with a rat mono-clonal antibody against mouse CD31 (60 min, 378C, 1:100 dilu-tion, Pharmingen, San Diego, CA), a goat anti-human VEGFantibody (60 min, 378C, 10 (�g/ml), (R&D Systems, Wiesbaden-

FIGURE 2 – Microvascular density of the primary tumors of trans-fected human melanoma clones. (a,b) Immunohistochemistry ofintratumoral vessels in primary tumors of human melanoma clones(confocal microscopy). Cryosections of 13-day-old tumors have beenlabeled for microvessels using a rat monoclonal anti-CD31 antibodyand a rabbit anti-rat-FITC conjugate. (a) ¼ 3.1P (mock transfectedclone); (b) ¼19H (�IIb�3-transfected high expressor). (c) Morphom-teric determination of microvascular density of the primary tumorsof transfected human melanoma clones. Vessels identified by immu-nohistochemistry (as in a,b) were counted on 5 different 200x fieldsof 3 parallel specimens. Data are expressed as mean 6 SEM (n ¼15). *p < 0.01 compared to 3.1P, determined by ANOVA singlefactor analysis.

FIGURE 1 – Growth of melanoma clones in SCID mice. (a) 106

tumor cells were injected into the spleen of SCID mice and the size ofthe primary was determined on the 30th day when animals were termi-nated by Nembutal overdose. Data are expressed in g (mean 6 SD,n ¼ 5). *p < 0.05. (b) One hundred thousand tumor cells wereinjected intradermally into the animals, tumor diameters were meas-ured and the tumor volume was calculated on the 10th and 13th post-injection days as described. Data are expressed in mm3 (means 6SEM, n ¼ 5). *p < 0.05, **p < 0.001.

TABLE II – INCIDENCE OF METASTASIS IN VARIOUS ORGANS OF �3INTEGRIN TRANSFECTED HUMAN MELANOMA CLONES1

Lung (i.v.) Liver (i.s.) Brain (i.c.) Bone (i.c.)

clone3.1P 100 100 100 83ESL 83 100 50 80ESH 100 100 75 7519L 86 100 60 4019H 100 80 100 75

1Tumor cells (106/animals) have been injected i.v. for lung colonyassay, intrasplenically for liver colonization and intracardially forbrain and bone colonization. Experiments have been terminated fol-lowing various periods as described in Materials and Methods, andmacroscopic colonies have been caunted under stereo microscope.Each group contained at least 5 SCID mice. Data are incidence ofmetastatic colonies expressed in % of animals.

TABLE I – SURFACE �3 INTEGRIN EXPRESSION IN HUMAN MELANOMACLONES (FLOW CYTOMETRY)1

Clone �IIb �3 �v�3

3.1 mock 0.66 6 0.7 80.0 6 3.2 44.0ESL 6.7 6 3.4 64.0 6 3.0 31.08F3 8.0 79.0 nd19L 12.3 6 4.9 67.0 6 5.6 10.5ESH 22.5 6 7.7 69.3 6 2.3 13.019H 27.2 6 3.9 78.0 6 6.2 28.4

1Data are in % of positive cells (�IIb and �3: mean 6 SEM, n ¼3–6, �v�3: average of 2 samples), nd ¼ not determined.

28 DOME ET AL.

Nordenstadt, Germany) or a goat anti-human bFGF antibody(60 min, 378C, 10 (�g/ml), (R&D Systems). FITC-conjugated rabbitanti-rat and anti-goat IgG antibodies (45 min, 378C, 1:50, JacksonImmunoresearch Lab, Inc., West Grove, PA) were used as secondary

antibodies. Following nuclear staining with 5 �M TOTO-3 in PBS(20 min, 208C, Molecular Probes, Inc.) slides were mounted andviewed by confocal laser scanning microscopy (Bio-Rad MRC1024, Munich, Germany). Counts of intratumoral blood vessels

FIGURE 3 – Immunocytochemical expression of angiogenic cytokines in transfected human melanoma clones (flow cytometry). 106 cellsfrom each studied clone were fixed in MetOH and labeled with anti-VEGF or anti-bFGF antibodies. The specifically bound primary antibod-ies were detected by an appropriate FITC-labeled secondary antibody. As negative control, isotype-matched nonimmune IgG was usedinstead of the primary. At each experimental point a minimum of 104 cells were measured for fluorescence and expressed on logarithmicscale. (a) Original printouts: 1–4 ¼ bFGF expression, 5–8 ¼ VEGF expression. 1,5 ¼ negative control, 2,6 ¼ 3.1P mock transfected clone,3,7 ¼ 19L �IIb�3-transfected low expressor clone and 4,8 ¼ 19H �IIb�3-transfected high expressor clone. (b) Graphical visualization ofVEGF and bFGF protein expression in human melanoma clones as determined by flow cytometry. Data are expressed in % of positive cells.

29PARALLEL EXPRESSION OF �IIB�3 AND �V�3 INTEGRINS

were determined visually by fluorescence microscopy in �200fields within the entire tumor mass.

Flow cytometry

For immunofluorescence labeling suspended cells were fixedwith 1% buffered paraformaldehyde (PFA) for 15 min at roomtemperature permeabilized with 0.2% saponin for 10 min. Nonspe-cific antibody binding was blocked using PBS containing 1%BSA. The following mouse monoclonal antibodies were used:50 �g/ml of anti-CD41 (anti-�IIb), anti-CD61 (anti-�3) (bothfrom Dako, Glostrup, Denmark) and 12.5 �g/ml of anti-�v�3complex (Chemicon, Hofheim, Germany). Cells incubated withthe appropriate isotype control IgG were used as negative controls.The bound primaries were detected with a biotinylated anti-mouseIg (1:100; Vector Laboratories, Inc., Burlingame, CA) followed byStreptavidin-FITC (1:100; Amersham, Little Chalfont, England).

To detect the expression of pro-angiogenic molecules, 106 tumorcells/sample were fixed in methanol (10 min, �208C) incubatedwith goat anti-human VEGF antibody (60 min, 208C, 10 �g/ml,R&D Systems) or goat anti-human bFGF antibody (60 min, 208C,10 (�g/ml), (R&D Systems) revealed by a FITC-conjugated sec-ondary anti-goat antibody (Jackson ImmunoResearc Labs, WestGrove, PA).

The percentage of positive cells was determined by a FACStarflow cytometer (Becton Dickinson, Sunnyvale, CA). Cells wereconsidered positive when the fluorescence intensity was above thatof 95% of the control cells labeled without the primary.

Inhibition of signal transduction in vitro

To inhibit protein kinase C activity, in vitro cultured tumor cellswere pretreated with Calphostin C (Kamiya Biomedical Co.,Thousand Oaks, CA) for 25 hr at a concentration of 5 �M usinglight activation. For the downregulation of PKC, we have exposedtumor cells to 8 ng/ml PMA (phosbol-12-myristate-13-acetate;Sigma Chemical Co.) for the same period of time. To inhibit pro-tein tyrosine kinases, tumor cell cultures were pretreated withGenistein or Staurosporine (Calbiochem) for 25 hr at a concentra-tion of 100 �g/ml.

Quantitative PCR

Total RNA were extracted from the cells growing in cell cultureusing Tri-Reagent (Sigma Aldrich Co, Pool Dorset, UK) accord-ing to the manufacturer’s protocol. Purity of the nucleic acid tem-plates was elevated by extra-purification with DNA-freeTM

(Ambion, Austin TX). One microgram of total RNA were reversetranscribed from each sample using deoxy (d)-NTPs (0.5 mmol/Leach), random primer (final concentration 3 �M), RNasin1 ribo-nuclease inhibitor (Promega) (0.4 U/�l), DTT (final concentration10 mM), reverse transcriptase buffer (containing 250 mM Tris-HCl, pH 8.3, 375 mM KCl and 15 mM MgCl2) and M-MLVreverse transcriptase (200 U/reaction; Sigma Chemical Co.).Twenty �l samples were incubated for 50 min at 378C and then at858C for 10 min. The sequences of bFGF primers were the follow-ing: 50-GCC ACA TCT AAT CTC ATT TCA CA-30 and 50-CTGGGT AAC AGC AGA TGC AA-30. The sequences of endothelinreceptor type B primers were 50-GCT CTT AAC AAC TTC CAGGAT ATT C-30 and 50- CAA CAA CTA AAC TGC TCT CTC

ATT T-30. The sequences of (3-endonexin primers were 50-CACTGA AGT TGG ATG GTC TGT TA-30 and 50-TTC ATT TGATAG TCC ATT TCT GTG C-30. (The primers were designed forthe 111 amino acid coding sequence of the �3-endonexin cDNA,which is common in the 3 different splice isoforms of the �3-endonexin).

The real-time PCR analysis for mRNA expression of bFGF,CD34, endothelin receptor typeB, prostacyclin synthase and �3-endonexin standardized by coamplifying these genes with thehousekeeping gene �-actin (primers: 50-GTG GGG CGC CCCAGG CAC CCA-30 and 50-GTC CTT AAT GTC ACG CAC GATTTC-30). The real-time PCR reaction was run on the iCycler iQTM

(Bio-Rad) using standard conditions, namely, optimized concen-tration of primers (final concentration 200 nM), IQTMSYBR1

Green Supermix (containing 100 mM KCL, 40 mM Tris-HCL, pH8.4, 0.4 nM of each dNTP, 6 mM MgCl2, 50 U/ml iTaq DNA poli-merase, SYBR Green I and 20 nM fluorescein) and 2 �l cDNA. Ano template control (containing water) was used as negative con-trol for every different primer-pair. The cycling parameters of thebFGF real-time PCR analysis were 958C (3 min), 40 cycles of958C (30 sec), 618C (30 sec) and 728C (1 min). The cyclingparameters of the �3-endonexin real-time PCR analysis were958C (3 min), 40 cycles of 958C (30 sec), 608C (30 sec) and 728C(1 min). The starting quantity of gene expression in the samplewas determined by comparison of unknown to a standard curvegenerated from a dilution series of template DNA of known con-centration. CT values were converted into relative concentrationvalues by scaling from 1 to 1 between the expression control andthe highest value of the samples using the autothreshold fit.

Construction of microarrays

Construction and use of microarrays were performed as previ-ously described.9,10 Briefly, 3,200 cDNA inserts from humancDNA libraries (melanoma, lymphocytes, heart and mixed tissuelibraries) were amplified and purified with MultiScreen-PCR plate(Millipore), resuspended in 50% dimethylsulfoxide/water andarrayed on FMB cDNA slides (Full Moon BioSystems, Sunnyvale,CA) by using a MicroGrid Total Array System (BioRobotics, Cam-bridge, UK) spotter with 16 pins in a 4 � 4 format. DNA elementswere deposited in duplicate. The diameter of each spot was approxi-mately 200 �m. After printing, DNA was UV crosslinked to theslides (Stratagene, Stratalinker, 700 mJ). Postprocessing and block-ing of the microarrays were performed as described previously.11

Microarray probe preparation and hybridization

RNA isolation from 5 � 106 cells was carried out with the RNAisolation kit of Macherey-Nagel (Macherey-Nagel, Duren, Ger-many) according the manufacturer’s instruction. RNA was elutedfrom the silica membrane and stored at �808C in the presence 30 Uof Prime RNAse inhibitor (Fermentas, Vilnius, Lithuania). Thequality and quantity of isolated RNA was checked by electrophore-sis and spectrophotometry (NanoDrop, Rockland, DE) respectively.

For probe preparation, 4 �g of total RNA was reverse tran-scribed using poly-dT primed Genisphere Expression Array 350Detection system (Genisphrere, Hatfield, PA) in 20 �l total vol-ume using 20 Unit RNAsin (Fermentas), 1� first strand buffer and200 Units of RNAse H (�) point mutant M-MLV reverse tran-scriptase (Fermentas). All the other probe preparation steps were

FIGURE 4 – Immunohistochemical detection of bFGF in transfected melanoma cells and in their primary tumors. (a,b) Fluorescence micro-scopy. In vitro cultured human melanoma clones, 3.1P (a) and 19H (b) have been fixed in paraformaldehyde, permeabilized with 0.5%TritonX-100 for 1 min and labeled for bFGF with goat anti-human bFGF and visualized by FITC-conjugated anti-goat antibody, where the nuclei havebeen labeled with propidium iodine (red). Note the rare cytoplasmic green signal in 3.1P cell population (a), compared to the strong uniform labelin 19H cells (b). (c,d) Confocal microscopy. Frozen sections of 13 day old orthotopic tumors were fixed and labeled with goat anti-human bFGFas above but the nuclei were labeled by TOTO-3 (blue). (c)¼3.1P, (d)¼19H clones. Note the high frequency of cytoplasmic reaction for bFGF inthe majority of tumor cells in transfected 19H tumor (d) as compared to the rare positivity in the case of the mock transfectant 3.1P clone (c).

30 DOME ET AL.

FIGURE 4.


done according the manufacturer’s instruction (Genisphere).cDNA was hybridized onto human cDNA microarrays in a Ven-tana hybridization station (Ventana Discovery, Tuchon, State) byusing the ‘‘antibody’’ protocol. First hybridization was performedat 428C for 6 hr in ‘‘Chiphyb’’ hybridization buffer (Ventana) andthen 2.5 �l of each Cy5 and Cy3 capture reagents were added tothe slides in 200 �l Ribohyb hybridization buffer (Ventana) andincubated at 428C for 2 hr. After hybridization, the slides werewashed in 0.2� SSC twice at RT for 10 min and then dried andscanned.

Microarray analysis

Each array was scanned under a green laser (543 nm for Cy3labeling) or a red laser (633 nm for Cy5 labeling) using a ScanAr-ray Lite (GSI Lumonics, Billerica, MA) scanning confocal fluores-cent scanner with 10 �m resolution (laser power: 85% for Cy5 and90% for Cy3, Gain: 80% for Cy5 and 75% for Cy3). Scanned out-put files were analyzed using the GenePix Pro3.0 software (AxonInstruments, Inc., Foster City, CA). Each spot was defined byautomatic positioning of a grid of circles over the image. Theaverage and median pixel intensity ratios calculated from bothchannels and the local background of each spot were determined.An average expression ratio (MeanR, denotes the average of localbackground corrected pixel intensity ratios) was determined foreach spot. Normalization was performed by the global Lowessmethod. Those data were flagged and excluded where the replicatespots from a different site of the same array were significantly dif-ferent. We defined those genes to be regarded in melanoma cellcultures, where the average-fold change (increase or decrease) ofthe 4 data points were at least 2.0-fold.

Results

Parallel expression of �IIb�3 and �v�3 integrins promotesortho- and hetero-topic growth of human melanoma cell lines

The 5 �IIb�3 transfected human melanoma clones expressedsurface �IIb at different levels (6.7–27.2%) compared to the 3.1Pmock transfected clone (0.7%) as determined by flow cytometry(Table I), while the �3 expression was similarly high in all theclones studied (64–80%). �v�3 expression was high in the mockcells and was maintained, although at a lower levels in most of thetransfected clones (Table I).

�IIb�3 transfected human melanoma clones were injected het-erotopically into the spleen of SCID mice. All the clones were

tumorigenic but the �IIb�3-transfected clones (ESL, ESH, 19Land 19H) grown significantly more rapidly than the mock cells,3.1P (Fig. 1a), as was observed in case of s.c. growth reportedbefore.8 Next, selected �IIb�3-transfected human melanomaclones (3.1P, 19L and 19H) were injected orthotopically (intracu-taneously) into SCID mice. All these clones started to grow, whichwas detectable from days 7–10. However, it became apparentagain that the �IIb�3-transfected clones, 19L and 19H, grew sig-nificantly faster at orthotopic site than the mock transfectant in theearly phase of tumor development (Fig.1b, p < 0.05 and p <0.001, respectively).

Next we have tested whether the �IIb�3-transfection affectedorgan colonization potential of human melanoma clones. We haveused 3 different models (intravenous, intrasplenic and intracardiacinjections), but we have not seen significant alterations in themetastatic potential of transfected clones compared to the mockcells, 3.1P (Table II).

Parallel expression of �IIb�3 and �v�3 integrins promotesvascularization and the angiogenic phenotype of humanmelanoma cells

Intracutaneously growing human melanoma xenografts have beenremoved on day 15, snap frozen, sectioned and labeled for mouseCD31 to visualize intratumoral blood vessels (Fig. 2a,b). Tumors ofthe transfected clones, 19L and 19H, contained more intratumoralblood vessels than that of the mock transfected 3.1P cells, and thedifference was significant in case of the 19H tumor (Fig. 2c).

Next we analyzed the immunocytochemical expression of themajor angiogenic cytokines of melanoma, VEGF and bFGF, in the�IIb�3-transfected human melanoma clones, 19L and 19H, byflow cytometry. Data indicated that VEGF is constitutivelyexpressed in the majority of the cells of both the mock and the�IIb�3-transfected clones (Fig. 3a,b). However, the mock trans-fected clone, 3.1P, did not express bFGF at protein level (<1% ofthe cell population, Fig. 3a,b), unlike the �IIb�3-transfectedclones, 19L (33.9%) and 19H (84.1%), where a significant propor-tion of the tumor cell population was positive (Fig. 3a,b). In vitrocultured tumor cells were also stained for bFGF protein and thedata supported the observations made by flow cytometry that thecells of the �IIb�3-transfected 19H clone are heavily positive forbFGF protein (Fig. 4a,b). We have questioned if this difference inbFGF expression was maintained in the intracutaneously growingtumors; therefore cryosections have been labeled for bFGF proteinand analyzed by confocal microscope. The 3.1P tumor barely con-tained bFGF positive cells, but the �IIb�3-transfected tumor(19H) was identified by a predominating population of tumor cellswith strong cytoplasmic bFGF labeling (Fig. 4c,d).

Parallel expression of �IIb�3 and �v�3 integrins correlates withaltered gene expression profile of human melanoma cells

Next we questioned whether the �IIb�3 transduction into�v�3-positive human melanoma clones affected bFGF expressionat transcriptional level. We collected 6 human melanoma cloneswith various levels of �IIb�3 and �v�3 integrin expressions(Table I) and tested for bFGF mRNA levels. The quantitative-PCR analysis revealed that bFGF mRNA level increased parallelwith �IIb�3 protein expression in the analyzed clones, where thehighest bFGF expressions detected in clones with the highest pro-portion of �IIb�3 positive cells, 19L, ESH and 19H, respectively(Fig. 5). However, after a log increase of both �IIb�3 or bFGFexpressions, a further increase in integrin expression (clones ESHand 19H) does not necessarily result in further upregulation ofbFGF (19H), suggesting a more complex association between theregulation of the 2 genes.

Next we tested the effect of �IIb�3-transfection on geneexpression using a homemade 3.2K cDNA microarray9,10 that didnot contain a bFGF probe. For this purpose, we used the fastestgrowing and most vascularized clone, 19H, and compared it to themock cells, 3.1P. We defined those genes to be regulated by

FIGURE 5 – bFGF mRNA levels in transfected melanoma clonesdetermined by quantitative PCR. Parallel demonstration of �IIb-pro-tein expression (derived from Table I) and bFGF mRNA expressionsin transfected clones calculated from CT values which were convertedinto relative concentration values by scaling from 1 to 1 between theexpression control and the highest value of the samples using the auto-threshold fit (Bio-Rad). Left axis: �IIb protein expressed in % of posi-tive cells. Right axis: bFGF expression in Q-PCR relative units.

32 DOME ET AL.

�IIb�3-transfection in melanoma cell cultures, in the case ofwhich the average-fold change (increase or decrease) of the 4 datapoints was at least 2.0-fold. Forty-eight genes in 19H cells corre-sponded to such strict criteria from the 3,200 analyzed; 36 geneswere found to be upregulated and 12 have been downregulated(Table III). Among the downregulated genes, there were no knownangiogenic/endothelial ones represented; however, among the 19known genes upregulated in 19H cells compared to 3.1P, 3 couldbe considered endothelial-specific, such as CD34 antigen, endo-thelin receptor type B and prostaglandin I-2 synthase, the changesof which were also confirmed by q-PCR analysis (data notshown).

Based on this information, we have postulated that an alteredsignaling activity in the �IIb�3-transfected clones may play a rolein the changes in gene expressions, including bFGF. Q-PCR anal-ysis of the expression of a �3 integrin-associated signaling mole-cule, �3-endonexin, indicated a 2.5–5-fold increase in theexpression of this gene in various transfected clones compared to3.1P (Fig. 6a). Again, similarly to bFGF expression (Fig. 5), therewas no linear correlation between the �IIb (Fig. 5) and �3-endo-nexin expression levels in transfected clones. Next we used tyro-sine kinase (Genistein, Staurosporine) and PKC (Calphostin C)inhibitors to modulate cellular signaling in the 19H �IIb�3-trans-fected clone and measured the consequences on bFGF mRNA andprotein expressions as determined by q-PCR and flow cytometry.These studies indicated that 25 hr exposure to tyrosine kinaseinhibitors resulted in a decrease, while PKC-inhibition resulted inan increase in bFGF message (Fig. 6b). Measurement of the cyto-

plasmic bFGF protein by immunocytochemistry confirmed theseobservations, since both tyrosine kinase inhibitors (Genistein andStaurosporine) significantly inhibited the expression in transfected19H cells (Fig. 6c).

Discussion

Recently we have shown that transfection of �IIb�3 integrininto human melanoma cells constitutively expressing �v�3 pro-motes in vivo growth due to decreased apoptosis.8 Our analysis ofthe possible explanation for this phenomenon indicated that the�IIb�3 transfected clones are more vascularized than the cloneexpressing �v�3 alone providing an explanation for the increasedin vivo growth. Based on our results, the studied clones constitu-tively express VEGF, but bFGF mRNA and protein expressionsare induced only in the �IIb�3 transduced clones. In agreementwith previous studies,6 we found that the expression of �IIb�3parallel with �v�3 integrin does not modify the organ selectivityof the metastatic potential of human melanoma cells, suggesting aspecific modulatory effect of �IIb�3 integrin on the angiogenicphenotype.

bFGF was shown to be an autocrine growth factor of humanmelanoma which is involved in proliferation, survival and inva-siveness of transformed melanocytes.2,11 The biological signifi-cance of its expression is demonstrated by experiments wherebFGF or FGFR1 were used successfully as molecular targets toprevent in vivo growth of melanoma.12 Meanwhile bFGF is one of

TABLE III – GENES OVEREXPRESSED IN TRANSFERRED CLONE 19H VS. MOCK CONTROL, 3.1P

Fold Ac.No Name

2,002048 D84124 Prostaglandin I-2 (prostacyclin) synthase2,013939 AA933627 ESTs, Highly similar to PHOSPHATIDYLOCHOLINE

TRANSFER PROTEIN [Bos Taurus]2,014675 D21235 Human mRNA for HHR23A protein, complete cds2,035672 AA195261 EST2,045783 R06710 EST2,053254 AA524013 EST2,065131 AA251134 EST2,108728 A1024091 EST2,132049 AA526945 EST2,13392 D83780 Human mRNA for KIA0196 gene, complete cds2,156721 L76927 Galactokinase 12,173375 D131168 Endothelin receptor type B2,177483 AA731863 EST2,23487 AA648933 ESTs, Weakly similar to homologous to

mouse Rsu-1 (H.sapiens)2,302858 AC004262 Homo spiens chromosome 19, cosmid R293682,307286 L34408 Homo Sapiens (clone B3B3E13)

chromosome 4p16.3 DNA fragment2,332845 AA483426 EST2,366137 AA843532 EST2,387688 AA748199 EST2,387747 X62535 Diacylglycerol kinase, alpha (80kD)2.403949 M10942 Human metallothionein-le gene (hMT-le)2,418433 Y00971 Phosphoribosyl pyrophosphate synthetase 22,463528 N64780 EST2,613085 AA927480 EST2,624854 AA521146 EST2,650112 AF001862 Human SLP-76 associated protein

mRNA, complete cds2,660733 R17461 IMAGE EST2,684187 AA576731 EST2,735851 AA707308 EST2,77783 U07857 Signal recognition particle 14 kD protein2,785849 X54232 Glypican 12,821762 AA741117 EST2,950237 A1055892 EST2,987556 X52943 TRANSCRIPTION FACTOR ATF-A AND ATF-A-DELTA3,07509 S53911 CD34 antigen (hemopoietic progenitor cell antigen)4,460516 L03411 Radin blood group


the most ubiquitous angiogenic cytokine that plays a significantrole in vasculogenesis, as well as in physiological and pathologicalneoangiogenesis.13 It was repeatedly shown in melanoma that sev-eral angiogenic cytokines including VEGF are responsible for theangiogenic phenotype,14,15 but these studies pointed to the signifi-cant contribution of bFGF. This assumption was further corrobo-rated by the effects of downmodulation of FGF using antisensetherapy,16 resulting in decreased vascularization. Our observationin our study on the upregulated bFGF expression in a constitutiveVEGF background in the more angiogenic human melanomaclones further supports these data. Upregulation of bFGF duringthe progression of human melanoma was reported at the transitionfrom the radial to the invasive vertical growth phase.2 Studies onhuman samples demonstrated that �v�3 integrin expression isconstitutive in primary melanomas, while the �IIb�3 expressionincreased with tumor thickness.8,17 We suggest that in human mel-anoma, overexpression of �v�3 integrin may correlate with VEGFexpression, while the emergence of the illegitimate expression ofthe �IIb�3 integrin in a later stage of the disease (vertical growthphase) may contribute to the upregulation of bFGF expression,providing both another autocrine growth factor as well as anotherangiogenic cytokine for melanoma cells.

Parallel expression of the two �3-integrins in human melanomacells induced modulation not only of bFGF but other genes aswell, which is demonstrated by microarray analysis. Interestingly,3 out of 19 known genes upregulated significantly in �IIb�3-trans-fected 19H cells are endothelial cell-specific genes, CD34, endo-thelin receptor B and PGI-2 synthase. Recapitulation of anembryonic endothelial/angioblastic genetic program, called vascu-logenic mimicry,18 resulting in in vivo formation of tumor cell-lined vascular channels, was recently described and documentedin human melanoma and other tumors.19 There have been severalgenes identified to be responsible for this vasculogenic phenotypeincluding VE-cadherin, CD34, TIE2, Eph2A, LAMC2, endoglin,EDG1, ESM1 and EDF1.20

AlphaIIb�3 transfection into �v�3-expressing human mela-noma cells, demonstrated here, correlated with the overexpres-sion of certain vasculogenic genes (bFGF, CD34, ENDRB andPGI-2 synthase), suggesting that �3 integrin signaling could beone of the potential molecular mechanisms behind this geneticreprogramming. �v�3 integrin is involved in survival signalingin melanoma cells21 mediated either through FAK or Shc usingthe Ras-Raf-MAPK pathway22,23 as well as PI3K.24 On the otherhand, it was shown in murine melanoma cells, spontaneouslyexpressing �IIb�3 (but not �v) integrin, that �3 signaling is con-stitutively active and the ligation induced further activation ofPKC.25 Based on the observations that tyrosine kinase inhibitorsdownregulated bFGF expression in �IIb�3-transduced 19H cellswhile PKC inhibitor upregulated it, we suggest that the integrin-linked classical tyrosine signaling cascade might be involved inthe upregulation of bFGF and other gene expressions in �IIb�3-transfected human melanoma cells. �3 integrin is unique amongintegrins, since its cytoplasmic domain is associated with a sig-naling mediator that is a transcription factor: the �3-endonexinprotein.26,27 The fact that �3-endonexin is also upregulated inthe �IIb�3-transfected clones suggests that this integrin-linkedsignaling protein may also be involved in the observed altera-tions in the angiogenic geno- and pheno-types of human mela-noma cells.

Acknowledgements

This work was funded by Hungarian National Science fund(OTKA T38128 to JT), by the Ministry of Education (1/48/2001,TJ), by the T. Fox Fund, by a bilateral grant of the HungarianPrime Minister Office and Hungarian Academy of Sciences (4676/2003), NIH grant CA47115 (KVH) and NATO (CLG 975187,KVH&JT).

FIGURE 6 – Modulation of signal transduction in �IIb�3- trans-fected human melanoma clones. (a) q-PCR analysis of the expressionof �3-endonexin mRNA in �IIb�3-transfected ESL, 19L and 19Hcells compared to the mock-transfected 3.1P clone. Data are normal-ized for the expression of �-actin and are in relative units (mean of 2parallel samples). (b) q-PCR analysis of bFGF mRNA in 19H cellstreated with Staurosporine (Stau), Genistein (Gen, 100 �g/ml) or Cal-phostin C (Cal, 5 �M) for 25 hr at 378C. Data are normalized for theexpression of �-actin and are in relative units (mean of 3 parallel sam-ples). (c) Effect of protein kinase inhibitor pretreatments on theexpression of bFGF protein in melanoma cell lines as detected byimmunocytochemistry and flow cytometry. Adherent cells were pre-treated with the inhibitors for 25 hr at 378C (Genistein, Gen and Staur-osporin, Stau: 100 �g/ml). Following incubation, cells weresuspended, fixed in MetOH and have been labeled for bFGF andmeasured by flow cytometry as in case of Figure 3. Data are expressedas mean fluorescence intensity 6SD (n ¼ 3). *p < 0.05 determined byANOVA single factor analysis.

34 DOME ET AL.

References

1. Halaban R, Kwon BS, Ghosh S, Delli Bovi P, Baird A. bFGF as anautocrine growth factor for human melanomas. Oncogene Res1988;3:177–86.

2. Meier F, Nesbit M, Hsu MY, Martin B, Van Belle P, Elder DE,Schaumburg-Lever G, Garbe C, Walz TM, Donatien Ph, Cromble-holme TM, Herlyn M. Human melanoma progression in skin recon-struct: biological significance of bFGF. Am J Pathol 2000;156:193–200.

3. Albelda SM, Mette SA, Elder DE, Stewart RM, Damjanovich S, Her-lyn M, Buch CA. Integrin distribution in malignant melanoma: Asso-ciation of the �3 subunit with tumor progression. Cancer Res1990;50:6757–74.

4. Hsu MY, Shih DT, Meier FE, Van Belle P, Hsu JY, Elder DE, BuckCA, Herlyn M. Adenoviral gene transfer of �3 integrin subunit indu-ces conversion from radial to vertical growth phase in primary humanmelanoma. Am Pathol 1998;153:1435–42.

5. Hieken TJ, Ronan SG, Farolan M, Shikaitis AL, Das Gupta TK.Molecular prognostic markers in intermediate thickness cutaneousmalignant melanoma. Cancer 1999;85:375–92.

6. Felding-Habermann B, Fransvea E, O’Toole TE, Manzuk L, Faha B,Hensler M. Involvement of tumor cell integrin �v�3 in hematogenousmetastasis of human melanoma cells. Clin Exp Metast 2002;19:427–36.

7. Trikha M, Tımar J, Lundy SK, Szekeres K, Cai Y, Porter AT, Honn KV,The high affinity �IIb�b integrin is in involved in invasion of humanmelanoma cells Cancer Res 1997;57:2522–8.

8. Trikha M, Tımar J, Zacharek M, Nemeth JA, Cai Y, Dome B,Somlai B, Raso E, Ladanyi A, Honn KV. Role for �IIb�3 integrinin human melanoma growth and metastasis. Int J Cancer 2002;101:156–67.

9. Puskas LG, Zvara A, Hackler L Jr, van Hummelen P. RNA amplifica-tion results in reproducible microarray data with slight ratio bias. Bio-techniques 2002;32:1330–40.

10. Kitajka K, Puskas LG, Zvara A, Hackler L Jr, Barcelo-Coblijn G,Yeo YK, Farkas T. The role of n-3 polyunsaturated fatty acids inbrain: modulation of rat brain gene expression by dietary n-3 fattyacids. Proc Natl Acad Sci U S A 2002;99:2619–24.

11. Nesbit M, Nesbit HK, Bennett J, Andl T, Hsu MY, Dejesus E,McBrian M, Gupta AR, Eck SL, Herlyn M. Basic fibroblast growthfactor induces a transformed phenotype in normal human melano-cytes. Oncogene 1999;18:6469–76.

12. Kato J, Wanebo H, Calabresi P, Clark JW. Basic fibroblast growthfactor production and growth factor receptors as potential targets formelanoma therapy. Melanoma Res 1992;2:13–23.

13. Hanahan D, Folkman J. Patterns and emerging mechanisms of theangiogenic switch during tumorigenesis. Cell 1996;86:353–64.

14. Rofstad EK, Halsor EF. Vascular endothelial growth factor, interleu-kin 8, plateletderived endothelial cell growth factor, and basic fibro-blast growth factor promote angiogenesis and metastasis in humanmelanoma xenografts. Cancer Res 2000;60:4932–8.

15. Westphal JR, Vanc Hullenaar R, Peek R, Willems RW, Crickard K,Crickard U, Askaa J, Clemmensen I, Rutter DJ, de Wall RM. Angio-genic balance in human melanoma: expression of VEGF, bFGF, IL-8, PDGF and angiostatin in relation to vascular density of xenograftsin vivo. Int J Cancer 2000;86:768–76.

16. Wang Y, Becker D. Antisense targeting of basic fibroblast growth fac-tor and fibroblast growth factor receptor-1 in human melanomasblocks intratumoral angiogenesis and tumor growth. Nat Med1997;3:887–93.

17. Puerschel WCH, Gawaz M, Worret WL, Fring J, Immunoreactivityof glycoprotein �IIb is present in metastatized but not in non-meta-statized primary malignant melanoma Br J Dermatol 1996;135:883–7.

18. Maniotis AJ, Folberg R, Hess A, Seftor EA, Gardner LM, Pe’er J,Trent JM, Meltzer PS, Hendrix MJC. Vascular channel formation byhuman melanoma cells in vivo and in vitro: vasculogenic mimicry.Am J Pathol 1999;155:739–75.

19. Hendrix MJC, Seftor EA, Hess AR, Seftor REB. Molecular plasticityof human melanoma cells. Oncogene 2003;22:3070–5.

20. Hendrix MJC, Seftor EA, Hess AR, Seftor REB. Vasculogenic mimi-cry and tumor-cell plasticity: lesson from melanoma. Nature Rev Can-cer 2003;3:411–21.

21. Varner JA, Cheresh DA. Integrins and cancer. Curr Opin Cell Biol1996;8:724–30.

22. Eliceiri BP, Klemke R, Stromblad S, Cheresh DA. Integrin �v�3requirement for sustained mitogen-activated protein kinase activityduring angiogenesis. J Cell Biol 1998;140:1255–63.

23. Schwartz MA, Shattil SJ. Signaling networks linking integrins andRho family GTPases. Trends Biochem. Sci 2000;25:388–91.

24. Giancotti FG, Ruoslahti E. Integrin signaling. Science 1999;285:1028–32.

25. Raso E, Tovari J, Toth K, Paku S, Trikha M, Honn KV, Tımar J.Ectopic �IIb�3 integrin signaling involves 12-lipoxygenase- andPKC-mediated serine phosphorylation events in melanoma cells.Thromb Haemost 2001;85:1037–42.

26. Shattil SJ, O’Toole T, Eigenthaler M, Thon V, Williams M, BabiorBM, Ginsberg MH. Beta 3-endonexin, a novel polypeptide that inter-acts specifically with the cytoplasmic tail of the integrin beta 3 subu-nit. J Cell Biol 1995;131:807–16.

27. Ohtoshi A, Maeda T, Higashi H, Ashizawa S, Yamada M, Hata-keyama M. Beta3-endonexin as a novel inhibitor of cyclin Associatedkinase. Biochem Biophys Res Commun 2000;267:947–52.


ANALYTICALBIOCHEMISTRY

Analytical Biochemistry 337 (2005) 76–83

www.elsevier.com/locate/yabio

0003-2697/$ - see front matter 2004 Elsevier Inc. All rights reserved.doi:10.1016/j.ab.2004.09.044

Real-time polymerase chain reaction-based exponential sample ampliWcation for microarray gene expression proWling

Zsolt B. Nagya, János Z. Kelemena, Liliána Z. Fehéra, Ágnes Zvaraa,Kata Juhászb, László G. Puskása,¤

a Laboratory of Functional Genomics, Biological Research Centre, Hungarian Academy of Sciences, P.O. Box 521, Szeged H-6701, Hungaryb Institute of Biochemistry, Biological Research Centre, Hungarian Academy of Sciences, P.O. Box 521, Szeged H-6701, Hungary

Received 2 August 2004Available online 15 December 2004

Abstract

Conventional approaches to target labeling for gene expression analysis using microarray technology typically require relativelylarge amounts of RNA, a serious limitation when the available sample is limited. Here we describe an alternative exponential sampleampliWcation method by using quantitative real-time polymerase chain reaction (QRT-PCR) to follow the ampliWcation and elimi-nate the overampliWed cDNA which could distort the quantitative ratio of the starting mRNA population. Probes generated fromnonampliWed, PCR-ampliWed, and real-time-PCR-ampliWed cDNA samples were generated from lipopolysaccharide-treated andnontreated mouse macrophages and hybridized to mouse cDNA microarrays. Signals obtained from the three protocols were com-pared. Reproducibility and reliability of the methods were determined. The Pearson correlation coeYcients for replica experimentswere r D 0.927 and r D 0.687 for QRT-PCR-ampliWcation and PCR-overampliWcation protocols, respectively. �2 test showed thatoverampliWcation resulted in major biases in expression ratios, while these alterations could be eliminated by following the cyclingstatus with QRT-PCR. Our exponential sample ampliWcation protocol preserves the original expression ratios and allows unbiasedgene expression analysis from minute amounts of starting material. 2004 Elsevier Inc. All rights reserved.

Keywords: Exponential sample ampliWcation; Quantitative real-time PCR; Microarray; Gene expression proWling

Microarray technology has led to powerful advancesin identifying genetic proWles and to the ability to screenthousands of individual genes that may be diVerentiallyexpressed in diVerent samples [1]. cDNA microarrayanalysis requires relatively large amounts of total RNA(10–20�g). However, some sources of RNA have limitedyield, in addition, the quality of RNA can vary substan-tially, depending on the method of isolation, the pres-ence of endogenous RNases, and other factors that maybe introduced during an experiment. The need for repli-

cate hybridization increases the demand for RNA evenfurther [2].

To circumvent this issue, signiWcant eVort has beenfocused on approaches to reduce the RNA requirementper hybridization through the use of in vitro-transcrip-tion (IVT)1-based [3–6,17] or PCR-based [2,7,8,17,18]ampliWcation technologies. With respect to RNA ampli-Wcation strategies, T7 polymerase IVT is currently themost commonly used and well documented ampliWca-tion protocol [3,17]. Although successful strategies for

* Corresponding author. Fax: +36 62 432576.E-mail address: [email protected] (L.G. Puskás).

1 Abbreviations used: IVT, in vitro transcription; QRT-PCR, quan-titative real-time PCR; LPS, lipopolysaccharide; RT, reverse transcrip-tion; SSC, standard saline citrate.

mailto: [email protected]

mailto: [email protected]

PCR for microassay gene expression proWling / Z.B. Nagy et al. / Anal. Biochem. 337 (2005) 76–83 77

signiWcantly reducing RNA requirements have beendeveloped, IVT-based ampliWcation procedures haveseveral notable shortcomings. Methods require multiplesteps and are time consuming, labor intensive, andcostly. Moreover, if RNA is very limited, consecutiverounds of IVT ampliWcation are required to generatesuYcient material for hybridization [9,10], which addsfurther complexity to the procedure and reduces the line-arity of the ampliWcation.

PCR has a number of potential advantages: it is fasterand more cost eVective and oVers an almost unlimiteddegree of ampliWcation. A number of variations on PCRlabeling have been described [11–14], including thoseusing template switching PCR [15–19], commercializedby Clontech in their SMART PCR cDNA synthesis kit[5]. While these methods demonstrated the potential ofthe approach, they focused little attention on the rela-tionship between the degree of ampliWcation and theWdelity of the ampliWed material with respect to thestarting mRNA population. However understandableconcerns about the faithfulness of PCR ampliWcationhave held back the widespread adoption of the approachby the microarray community [2,20,21]. Because the reli-ability of PCR-based ampliWcation depends mostly onthe exponential phase during cycling, we therefore devel-oped an alternative sample ampliWcation method, whichfollows the ampliWcation status by quantitative real-timePCR (QRT-PCR). With real-time detection of the prod-ucts we could eliminate DNA overampliWcation which isof great risk when using traditional cycling and in-gelanalysis of the products.

Materials and methods

Cell culture, RNA isolation, and quality assessment

Mouse macrophage cell lines were cultured in Dul-becco medium with 10% bovine serum albumin. Cellswithout and with induction with 1 �g/ml LPS for 2 hwere harvested. Total RNA was extracted using Nucleo-spin RNA II kit (Macherey-Nagel, Düren, Germany).The RNA concentration was assessed spectrophotomet-rically by NanoDrop (Rockland, DE, USA) and thequality was monitored by Agilent 2100 Bioanalyzer(Agilent Technologies, Waldbronn, Germany). TotalRNA was used for microarray analysis and for QRT-PCR.

Preparation of cDNA targets for signal ampliWcation

For probe preparation, 1 �g of total RNA wasreverse-transcribed using poly(dT)-primed GenisphereExpression Array 350 Detection system (Genisphere,HatWeld, PA, USA) according to the manufacturer’sinstruction. cDNA with Cy5 capture sequence was

hybridized onto mouse cDNA microarray containing3200 mouse cDNA samples in duplicate. Postprocessingand blocking of the microarrays were performed asdescribed previously [17].

Preparation of PCR-ampliWed and PCR-overampliWed labeled probes

Reverse transcription (RT) reactions were carried outin a Wnal volume of 20 �l with 1�g mouse macrophagetotal RNA, 75 pmol universal T7T25V primer (5�-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG(T)25 V-3�), 75 pmol FOR primer (5�-TGTCTGCAGTGGTAACAACGCAGAGTACGCGGG-3�),30 U Prime RNase Inhibitor (Eppendorf, Hamburg,Germany), 2.5 mM each dNTP, and 200 U RevertAid HMinus M-MuLV Reverse Transciptase (MBI Fermen-tas, Vilnius, Lithuania) in 1£ reaction buVer (MBI Fer-mentas) at 42 °C for 2 h in a total volume of 20�l. RNAand the primer mixture were denatured at 70 °C for5 min before RT reaction. The reaction was terminatedby heat inactivation at 70 °C for 15 min. The RT reactionmixture was diluted three times with nuclease-free water;1 �l of the diluted reaction mix containing RNA fromapproximately 15 ng total RNA was used for QRT-PCRusing iQ SYBR Green Supermix (Bio-Rad, Hercules,CA, USA). QRT-PCR analysis for cDNA ampliWcationwas carried out in a total volume of 50 �l with 1�l of thediluted RT product, 50 pmol T7T primer (5�-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGGTTT-3�), 50 pmol FOR primer in 1£ iQ SYBRGreen Supermix (Bio-Rad). QRT-PCR was performedin a Rotor-Gene 2000 (Corbett Research, Mortlake,Australia) with the following protocol: 21 cycles of 25 sat 95 °C for denaturation, 25 s at 60 °C for annealing,and 4 min at 72 °C for extension. The reaction was haltedin the exponential phase of the ampliWcation at 13–15cycles depending on the ampliWcation proWle. Overam-pliWed cDNA was obtained from PCR with 21 cycles.

The PCR mixtures were puriWed with the MinElutePCR PuriWcation Kit (Qiagen, Courtaboeuf, France)according to the manufacturer’s instructions; 1 �g puri-Wed PCR-ampliWed DNA was transcribed in a total vol-ume of 20�l using RiboMAX T7 Express System(Promega, Madison, WI, USA) according to the manu-facturer’s instructions. The ampliWed aRNA was puri-Wed using Nucleospin RNA II kit (Macherey-Nagel) and50�g ampliWed aRNA was labeled with Cy5 during RTusing 75 pmol random heptamer primer, 0.1 mM d(G/T/A)TPs, 0.06 mM dCTP, 0.04 mM Cy5-dCTP (AmershamBiosciences, Uppsala, Sweden), 30 U Prime RNaseInhibitor (Eppendorf), 200 U RevertAidTM H MinusM-MuLV Reverse Transciptase (MBI Fermentas) in 1£reaction buVer (MBI Fermentas). The aRNA, primer,and RNase inhibitor were denatured at 70 °C for 5 minand cooled on ice before adding the remaining reaction

78 PCR for microassay gene expression proWling / Z.B. Nagy et al. / Anal. Biochem. 337 (2005) 76–83

components. After 2 h incubation at 37 °C, the aRNAwas hydrolyzed with NaOH (250 mM Wnal concentra-tion). The labeled cDNA was puriWed with the MinElutePCR PuriWcation Kit (Qiagen) according to the manu-facturer’s instructions.

Array hybridization and data analysis

Hybridization was done by Ventana Discoveryhybridization station (Ventana Discovery, Tucson, AZ,USA). The probes were dissolved in 200�l hybridizationbuVer (Ventana). Denaturation was performed at 75 °Cfor 6 min, hybridization was done at 42 °C for 6 h, andwashing was done at 37 °C with 2£ SSC. After hybrid-ization the arrays were washed with 0.2£ SSC for 1 min.The slides were scanned with confocal laser scanner(ScanArray Lite, GSI Lumonics, Billerica, MA, USA).Image Wles were analyzed with the GenePix Pro 3.0.5program (Axon, Union City, CA, USA). After auto-matic Xagging, manual Xagging was performed toexclude spots having irregularities such as scratches anddust particles. Data obtained by the median of featurepixels were accepted only if 30% of pixels were above 2£standard deviation of the background intensity.

Quantitative real-time PCR

Real-time PCR was carried out in a Wnal volume of20�l with 1�l 1:5 dilution of diluted cDNA mixture,10 pmol gene-speciWc Forward and 10 pmol Reverseprimer in 1£ iQ SYBR Green Supermix (Bio-Rad) withthe following protocol: 25 s at 95 °C for denaturation, 25 sat 60 °C for annealing, and 15 s at 72 °C for extension.The primers were designed using Primer Express software(Applied Biosystems, Foster City, CA, USA) (sequencesare listed in Table 1). The relative expression of each genewas normalized to that of mouse �-actin. Nontemplate

control sample was used for each PCR to check the geno-mic DNA contaminations of cDNA template. Analysis ofthe results was done using the PfaZ [22] method.

Results and discussion

cDNA ampliWcation with QRT-PCR

One of the primary concerns of researchers conduct-ing PCR-based cDNA ampliWcation for microarrayanalysis is its weaker reproducibility. To address thisissue we intended to eliminate biases introduced byoverampliWcation and thus to improve the reliability ofPCR-based cDNA ampliWcation for microarray experi-ments by using real-time quantitative PCR. Experi-ments were designed to investigate the reliability andreproducibility of the traditional and the real-timePCR-based exponential ampliWcation protocols. Forcontrol nonampliWed samples, the dendrimer-based sig-nal ampliWcation method (Genisphere) was used. Fourhybridizations were performed using LPS-treated andnontreated mouse macrophages with the ampliWed andnonampliWed probes to compare the reliability and thereproducibility of the methods used. Total RNA (1 �g)was reverse transcribed and 15-ng aliquots were PCRampliWed with two protocols resulting in DNA samplesfrom early phase (13th–15th cycles) and late exponen-tial, early saturation phase (21st cycle) (Fig. 1A). By fol-lowing the ampliWcation status with SYBR greenlabeling in the real-time PCR instrument, reactions werehalted at the appropriate cycle. After DNA isolation thequality and quantity of ampliWed cDNA were assessedby gel electrophoresis and spectrophotometry (Fig. 1B).For generating microarray probes with sample ampliW-

cation, PCR and a subsequent in vitro transcriptionwere used to obtain ampliWed RNA as described earlier

Table 1Primers used in QRT-PCR

Gene product Forward primer Reverse primer

�-Actin TTCAACACCCCAGCCATGT GCATACAGGGACAGCACAGCCAlcohol dehydrogenase 5 AGTGCATTGGCAACGTGAAG TGACACCCCAGCCTTTGTGAMP activated protein kinase GTGGCTCTGGGCATCTTTGT CGCCCTTTCTCATCCACTACAAngiogenin-3 precursor ACAATGAGCCCATGTCCTTTG GCCAGAGTTGGAGGAATCACAAnti-oxidant protein 2 TCCAGTCGAGAAGGACGATAACA GTCAGGGCCAAAAATGAACACAspartyl aminopeptidase GGGTCAGAGAGTGCACAAGGA GGCTGAGATGCGCCGTAGTBasigin AGCCGCATGGCAGCC ATGGTAACCAACACCAGGACCTLy-6 alloantigen CAGGAAGACCTCTGCAATGCA TAGCGGTCCCAGAGTGCCTMetallothionein CAAATGTACTTCCTGCAAGAAAAG GCAGCCCTGGGAGCACTTOsteoprotegerin GAAAGCAGCGTGCAGCG TCAAGGCAAGAAGCTGCTCTGN-myc downstream regulated 4 GGTGCCCAATGCCAAGAC TCCTCAGCAGGTGCATTATCTCNon muscle tropomyosin 5 ACTTGGAACGCACGGAGG TCTGATCTGTTCATCCATCTCTCPeroxisomal t-2-enoyl-Coa reductase TCCCGGAAGCTGGACAGA AGCTGGAGGGAGGCAGAGAProcollagen, type I, � 1 GACAAGGGTGAGACAGGCGA CCCTGGAGACCAGAGAAGCCSerine hydroxymethyl transferase 2 GGCCCCCGCGTTGA AAAATCGACGACTCTACGGAAGTCSerine/threonine kinase 10 GCGGCTACCCAAGATCCA TGTGCAGGCTCTTCTTGTACATG


[17]. PCR-ampliWed cDNA (1 �g) was used forin vitrotranscription. On average 81.2 �g (§6.6 �g) aRNA wasgenerated. The presence of SYBR Green had no eVecton the eYciency of the reaction (data not shown).AmpliWed RNA was labeled using direct Cy5 incorpo-ration during RT. Probes were hybridized onto mousecDNA microarrays having 3200 features in duplicate.To determine the reproducibility of the sample ampliW-

cation protocols hybridizations derived from four inde-pendent reactions were analyzed. The Pearsoncorrelation coeYcients for the replica experiments werer D 0.927 and r D 0.687 for the exponential and for thesaturated samples, respectively. The Pearson correlationcoeYcient for the control method, where no sampleampliWcation was performed, was r D 0.941. The corre-lation coeYcients of the exponentially ampliWed sam-ples were comparable with the nonampliWed controlcoeYcients, which conWrms our hypothesis that theQRT-PCR protocol improved the reliability of thePCR-based ampliWcation.

Gene expression changes of 3200 genes of mousemacrophages in response to LPS treatment were fol-lowed by mouse-speciWc cDNA microarrays. To demon-strate diVerences in results obtained with the twodiVerent ampliWcation protocols, 10 genes from each ofthe up-regulated, down-regulated, and nonchangedgenes having diVerent expression ratios with the expo-nential and overampliWcation techniques are shown inTable 2. Among the selected genes there were few withnonsignifcant changes (marked with * in Table 2) ineither method. From the relative expression values of the30 selected genes it can be concluded that overampliWca-tion resulted in biased expression ratios, while in the caseof QRT-PCR-based exponential ampliWcation the start-ing mRNA ratios remained unchanged; 13 of 30 genesexhibited diVerent expression ratios in the case of over-ampliWed samples, while only 4 genes had diVerent

expression ratios when the exponential ampliWcationprotocol was applied in comparison to the results of thenonampliWed samples.

Among the 3200 genes examined in the control, sig-nal-ampliWcation-based microarray experiments anaverage of 1045 showed signiWcant intensity ratio rela-tive to background and 9.09% (95 genes) showed alteredexpression: 56 genes exhibited signiWcant down-regula-tion and 39 were up-regulated due to LPS. In the case ofthe exponential ampliWcation QRT-PCR protocol 1127genes showed signiWcant intensity ratio and 10.04% (118genes) showed altered expression: 62 genes exhibitedsigniWcant down-regulation and 56 were up-regulated.In the PCR-overampliWed-based microarray experi-ments an average of 924 genes showed signiWcant inten-sity ratio relative to background and 7.68% (71 genes)showed altered expression: 38 genes exhibited signiW-

cant down-regulation and 33 were up-regulated. Therewere numerous genes with altered expression in thecases of the exponential ampliWcation and the nonam-pliWed control methods, which exhibit no change whenoverampliWed samples were tested (see examples inTable 2). This discrepancy might be due to the satura-tion eVect of the nonfollowed ampliWcation protocol inlate cycles.

ConWrmation of gene expression changes with QRT-PCR

To verify exact individual gene expression changesobtained with the protocols tested QRT-PCR analysiswas performed on 15 genes. Genes were selected fromthose shown in Table 2. �-Actin gene was used as a con-trol. Ten independent samples were generated fromampliWcation reactions for each protocol using the sametotal RNA as starting material to determine the repro-ducibility of cDNA ampliWcation. The results are shownin Table 3. One-sample Student t test was performed on

Fig. 1. cDNA ampliWcation with QRT-PCR. (A) QRT-PCR diagram of the LPS-treated mouse macrophage. With the QRT-PCR halted at the 14thcycle, the ampliWed cDNA (a2) was generated in the exponential phase of the reaction; the overampliWed cDNA (a1) was isolated from reactionshalted at the 21st cycle; a3 denotes the nontemplate control. (B) Electrophoretic assessment of cDNA-ampliWcation-based QRT-PCR from LPS-treated mouse macrophage. Lanes 2 and 3 contained ampliWed cDNAs obtained from the exponential phase; lanes 4–6 were loaded with overampli-Wed samples; lane 1 is a 100-bp DNA ladder (Bioneer, Daejeon, Korea).


each of the 15 genes to assess signiWcant expressionchanges. The resulting p values were interpreted withregard to categorical data: down-regulated, up-regu-lated, and not changed. We referred to a gene as regu-lated when the corresponding p value was below 0.05. InTable 3, it can be seen that relative gene expressionratios obtained with the overampliWcation protocolresulted in higher p values and thus nonsigniWcant ratiosin 10 of 15 cases, while in the case of exponential ampliW-

cation only 5 values were above 0.05 and in the case ofnonampliWcation protocol all of the p values were below0.05.

Using expression data, inferences on the inXuence ofthe cDNA ampliWcation method on expression ratioscould be made with the use of the �2 test for count data[23]. In this test gene expression changes (no change, up-or down-regulation) were determined for all the 15genes in each method; 15 values from each method werecompared to each other. The p value was resulted in

p D 0.8807 for the comparison of the nonampliWed andthe exponentially ampliWed cDNA. This p valuesuggests good correlation with the results obtainedby the two methods. In case of overampliWed cDNAthe p value was 0.0291. This low value indicates signiW-

cant diVerences between gene expression ratiosobtained by the overampliWed and the nonampliWedprotocols.

To demonstrate diVerences in ratios obtained withdiVerent ampliWcation protocols, a scatter plot, for thosegenes whose expressions were changed in response toLPS treatment in the case of the nonampliWed protocol,was constructed Fig. 2. Expression ratios for these genesobtained by the exponential and overampliWed protocolsare also presented. In the case of the exponential proto-col most of the genes had expression ratios similar tothose of the nonampliWed protocol, while most of theratios obtained with the overampliWed method lie in the“nonchanged +1 to ¡1 zone” and had diVerent values.

Table 2Selected genes with altered expression in response to LPS treatment in mouse macrophages compared to non-treated cells

Light gray: diVerent Log ratio in the nonampliWed sample compared to the exponentially ampliWed cDNA sample; dark gray: similar Log ratio in thenonampliWed and exponentially ampliWed samples but diVerent with the overampliWed sample.

* Values above and below +1 and ¡1, respectively, however, with nonsigniWcant changes.

Gene product (Accession No.) NonampliWed SD Exponential SD OverampliWed SD

Down-regulated genesAMP-activated protein kinase (AF036535) ¡1.43 0.27 ¡2.18 0.48 ¡0.95 0.63Angiogenin-3 precursor (NM_177544) ¡1.08* 0.36 ¡1.68 0.22 0.09 0.78Arginase II (BC023349) ¡1.23* 0.34 ¡1.02 0.42 ¡1.54 0.39B-cell translocation gene 2 (NM_007570) ¡1.35* 0.48 ¡1.98 0.35 ¡0.91 0.26Ly-6 alloantigen (X04653) ¡2.12 0.23 ¡2 0.73 ¡0.01 0.54Memb. metallo endopeptidase (BC034092) ¡1.56 0.47 0.23 0.52 0.98 0.36Peroxisomal t-2-enoyl-Coa red. (BC013530) ¡1.87 0.14 ¡2.18 0.41 ¡0.07 0.72Retinol dehydrogenase 7 (BC024603) ¡1.06 0.16 ¡1.35 0.14 ¡1.68 0.57Serine hydroxymethyl trans. 2 (NM_028230) ¡1.94 0.36 ¡1.28 0.23 1.07* 0.39Vacuolar proton tr. ATPase d (AY137757) ¡1.25 0.31 ¡2.12 0.83 ¡1.01* 0.85

Genes with no changeAlcohol dehydrogenase 5 (BC062879) ¡0.42 0.26 ¡0.38 0.12 1.1* 0.79� Tubulin 7 (NM_009449) 0.23 0.43 ¡0.02 0.22 0.72 0.35Basigin (D00611) 0.84 0.17 0.62 0.65 0.68 0.59� Tubulin 5 (NM_011655) 0.53 0.21 1.62 0.44 0.85 0.66Cathelin-like protein (X94353) 0.67 0.35 ¡0.53 0.29 0.78 0.43Heat shock protein 86 (M57673) 0.37 0.84 0.86 1.09 ¡0.04 0.67NADH dehydrogenase a 1 (BC052817) ¡0.18 0.72 0.04 0.36 ¡0.23 0.55Nonmuscle tropomyosin 5 (NT_039260) 0.51 0.64 0.57 0.45 ¡0.17 0.37Osteoprotegerin (BC049782) 0.76 0.47 0.47 1.04 0.88 0.12Procollagen, type I, � 1 (UO3419) 0.19 0.27 0.32 0.36 0.86 0.49

Up-regulated genesAmyloid beta precursor prot. (AY011334) 1.25* 0.73 1.33* 0.87 0.95 0.56Anti-oxidant protein 2 (AF004670) 1.18 0.25 1.24* 1.2 0.76 1.04Aspartyl aminopeptidase (AF005051) 1.25* 0.93 0.86 0.57 0.37 0.53Interleukin 12 � (BC057212) 1.04 0.27 1.05* 0.62 1.18* 0.82Matrix metalloproteinase 2 (NM_008610) 1.25* 0.47 1.03* 0.65 0.84 0.23Matrix metalloproteinase 9 (NM_013599) 1.02 0.14 1.25* 0.37 1.34* 0.49Metallothionein (K02236) 1.13* 0.39 1.21* 0.46 0.65 0.83N-myc downstream regul. 4 (NM_145602) 1.47* 0.59 1.1* 0.62 ¡0.18 0.89Pentaxin related (BC022176) 1.08* 0.93 1.34 0.38 1.52 0.51Serine/threonine kinase 10 (NM_009288) 1.51 0.17 0.84 0.47 1.16* 0.53


Tab

le 3

Con

Wrm

atio

n of

exp

ress

ion

rati

os o

btai

ned

by d

iVer

ent

prot

ocol

s w

ith

RT

-QP

CR

ana

lysi

s

�2 tes

t: p

D0.

8807

(no

nam

pliW

ed t

empl

ate

com

pare

d to

exp

onen

tial

ly a

mpl

iWed

tem

plat

e),

pD

0.02

91 (

expo

nent

ially

am

pliW

ed t

empl

ate

com

pare

d to

ove

r-am

pliW

ed t

empl

ate)

and

pD

0.04

26(n

onam

pliW

ed t

empl

ate

com

pare

d to

ove

r-am

pliW

ed t

empl

ate)

. One

-sam

pled

Stu

dent

’s t

tes

t (t

val

ue, p

val

ue; �

D0.

05).

Lig

ht g

ray:

diV

eren

t ex

pres

sion

rat

io f

rom

exp

onen

tial

ly a

mpl

iWed

cD

NA

inco

mpa

riso

n to

exp

ress

ion

rati

o fr

om t

he n

on-a

mpl

iWed

cD

NA

; Dar

k gr

ay: d

iVer

ent

expr

essi

on r

atio

fro

m o

ver-

ampl

iWed

cD

NA

in c

ompa

riso

n to

exp

ress

ion

rati

o fr

om n

on-a

mpl

iWed

cD

NA

.

Gen

e P

rodu

ctN

onam

pliW

edE

xpon

enti

alO

vera

mpl

iWed

Ct v

alue

sre

late

d to

�-a

ctin

Log

rat

io (

§SD

)t

valu

ep

valu

eL

og r

atio

(§

SD)

t va

lue

p va

lue

Log

rat

iot

valu

ep

valu

e

Ly-

6 al

loan

tige

n¡

2.03

§0.

13¡

6.29

00.

003

¡1.

97§

0.03

¡6.

626

>0.

001

¡0.

31§

0.96

0.90

70.

387

4.21

§0.

18P

erox

isom

al t

-2-e

noyl

-Coa

redu

ctas

e¡

1.91

§0.

06¡

9.31

8>

0.00

1¡

2.41

§0.

01¡

6.45

5>

0.00

1¡

0.23

§0.

66¡

1.34

20.

212

3.14

§0.

15

Seri

ne h

ydro

xym

ethy

ltr

ansf

eras

e 2

¡1.

51§

0.26

¡4.

695

0.00

9¡

1.37

§0.

02¡

5.24

3>

0.00

11.

23§

0.29

¡3.

177

0.01

13.

57§

0.19

AM

P a

ctiv

ated

pro

tein

kina

se¡

1.21

§0.

04¡

6.80

40.

002

¡1.

56§

0.04

¡8.

379

>0.

001

0.18

§0.

811.

008

0.33

92.

73§

0.09

Ang

ioge

nin-

3 pr

ecur

sor

¡1.

11§

0.02

¡10

.794

>0.

001

¡1.

42§

0.01

¡9.

118

>0.

001

¡0.

56§

0.89

0.74

60.

474

6.7

§0.

15P

roco

llage

n, t

ype

I, a

lpha

10.

19§

0.03

¡9.

471

>0.

001

¡0.

59§

0.84

0.94

20.

370

0.15

§0.

541.

481

0.17

25.

58§

0.08

Alc

ohol

deh

ydro

gena

se 5

0.35

§0.

03¡

10.4

58>

0.00

1¡

0.28

§0.

72¡

1.17

60.

269

1.1

§0.

052.

808

0.02

1.52

§0.

05O

steo

prot

eger

in0.

48§

0.25

¡4.

633

0.00

90.

36§

0.37

3.04

90.

013

0.71

§0.

026.

417

>0.

001

3.97

§0.

11B

asig

in0.

66§

0.01

¡9.

574

>0.

001

¡1.

01§

0.69

1.31

00.

222

0.84

§0.

641.

554

0.15

42.

76§

0.25

Non

mus

cle

trop

omyo

sin

50.

67§

0.17

¡5.

023

0.00

70.

23§

0.83

¡7.

676

0.37

3¡

0.74

§0.

97¡

0.78

80.

458.

61§

0.07

N-m

yc d

owns

trea

m r

egul

ated

41.

14§

0.02

16.4

9>

0.00

11.

07§

0.04

11.7

94>

0.00

10.

52§

0.02

5.75

8>

0.00

10.

29§

0.16

Seri

ne/t

hreo

nine

kin

ase

101.

17§

0.03

6.83

70.

002

1.67

§0.

41¡

1.38

30.

019

1.58

§1.

270.

230.

822

4.85

§0.

12A

spar

tyl a

min

opep

tida

se1.

24§

0.31

¡3.

886

0.01

71.

22§

0.26

¡3.

271

0.00

90.

6§

0.42

1.73

50.

116

1.63

§0.

43A

ntio

xida

nt p

rote

in 2

1.32

§0.

0216

.141

>0.

001

1.27

§0.

0210

.569

>0.

001

0.86

§1.

180.

378

0.71

32.

17§

0.14

Met

allo

thio

nein

1.54

§0.

028.

985

>0.

001

1.76

§0.

69¡

1.27

70.

233

0.89

§0.

034.

707

>0.

001

¡1.

54§

0.2


This means that in most cases overampliWcationchanged the original expression ratio observed in thecase of the nonampliWed sample.

To determine whether the overampliWcation protocoldistorts ratios more for highly abundant genes than forless abundant genes we determined relative Ct valuesfrom QRT-PCR experiments, which reXect the relativemRNA copy number of the corresponding gene to �-actin in the sample. As seen in Table 3, genes having dis-torted ratios in the overampliWcation method possesseddiVerent relative Ct values. This suggests that no speciWcdistortion eVect could be determined, although only 10genes could be analyzed in our case. It would be interest-ing to follow more genes with distorting ratios to studythis hypothesis in detail.

In conclusion, overampliWcation has a major distort-ing eVect on the true expression ratios, while the methodusing quantitative real-time PCR described here pre-serves the original expression ratios and can be used inmicroarray experiments.

Acknowledgments

This work was supported by grants from the Hungar-ian Medical Research Council (ETT-400/2003) and theHungarian National ScientiWc Research Foundation(OTKA) Grant F034637. A.Z. was supported by a Post-doctoral Fellowship of the Hungarian ScientiWcResearch Fund (OTKA D042197).

References

[1] P.O. Brown, D. Botstein, Exploring the new world of the genomewith DNA microarrays, Nat. Genet. 21 (1999) 33–37.

[2] L. Petalidis, S. Bhattacharyya, G.A. Morris, V.P. Collins, T.C.Freeman, P.A. Lyons, Global ampliWcation of mRNA by tem-

plate-switching PCR: linearity and application to microarrayanalysis, Nucleic Acids Res. 22 (2003) e142.

[3] R.N. Van Gelder, M.E. von Zastrow, A. Yool, W.C. Dement, J.D.Barchas, J.H. Eberwine, AmpliWed RNA synthesized from limitedquantities of heterogeneous cDNA, Proc. Natl. Acad. Sci. USA 87(1990) 1663–1667.

[4] M. Mahadevappa, J. Warrington, A high-density probe arraysample preparation method using 10- to 100-fold fewer cells, Nat.Biotechnol. 17 (1990) 1134–1136.

[5] E. Wang, L.D. Miller, G.A. Ohnmacht, E.T. Liu, F.M. Marincola,High-Wdelity mRNA ampliWcation for gene proWling, Nat. Bio-technol. 18 (2000) 457–459.

[6] J. Phillips, J.H. Eberwine, Antisense RNA ampliWcation: a linearampliWcation method for analyzing the mRNA population fromsingle living cells, Methods 10 (1996) 283–288.

[7] A.K. Dixon, P.J. Richardson, K. Lee, N.P. Carter, T. Freeman,Expression proWling of single cells using 3 prime end ampliWcation(TPEA) PCR, Nucleic Acids Res. 26 (1998) 4426–4431.

[8] A. Wang, A. Pierce, K. Judson-Kremer, S. Gaddis, C.M. Aldaz,D.G. Johnson, M.C. MacLeod, Rapid analysis of gene expression(RAGE) facilitates universal expression proWling, Nucleic AcidsRes. 27 (1999) 4609–4618.

[9] L.R. Baugh, A.A. Hill, E.L. Brown, C.P. Hunter, Quantitativeanalysis of mRNA ampliWcation by in vitro transcription, NucleicAcids Res. 29 (2001) e29.

[10] L.R. Baugh, A.A. Hill, D.K. Slonim, E.L. Brown, C.P. Hunter,Composition and dynamics of the Caenorhabditis elegansearly embryonic transcriptome, Development 130 (2003) 889–900.

[11] G.M. Makrigiorgos, S. Chakrabarti, Y. Zhang, M. Kaur, B.D.Price, A PCR-based ampliWcation method retaining the quantita-tive diVerence between two complex genomes, Nat. Biotechnol. 20(2002) 936–939.

[12] N.N. Iscove, M. Barbara, M. Gu, M. Gibson, C. Modi, N. Wine-garden, Representation is faithfully preserved in global cDNAampliWed exponentially from sub-picogram quantities of mRNA,Nat. Biotechnol. 20 (2002) 940–943.

[13] K. Aoyagi, T. Tatsuta, M. Nishigaki, S. Akimoto, C. Tanabe, Y.Omoto, S. Hayashi, M. Sakamoto, T. Yoshida, et al., A faithfulmethod for PCR-mediated global mRNA ampliWcation and itsintegration into microarray analysis on laser-captured cells, Bio-chem. Biophys. Res. Commun. 300 (2003) 915–920.

[14] L. Smith, P. Underhill, C. Pritchard, Z. Tymowska-Lalanne, S.Abdul-Hussein, H. Hilton, L. Winchester, D. Williams, et al., Sin-gle primer ampliWcation (SPA) of cDNA for microarray expres-sion analysis, Nucleic Acids Res. 31 (2003) e9.

[15] A. Chenchik, Y.Y. Zhu, L. Diatchenko, R. Li, J. Hill, P.D. Siebert,Generation and use of high-quality cDNA from small amounts oftotal RNA by SMART PCR, in: P. Siebert, J. Larrick (Eds.), GeneCloning and Analysis by RT-PCR, Biotechniques Books, Natick,MA, 1998, pp. 305–319.

[16] M. Matz, D. Shagin, E. Bogdanova, O. Britanova, S. Lukyanov, L.Diatchenko, A. Chenchik, AmpliWcation of cDNA ends based ontemplate-switching eVect and step-out PCR, Nucleic Acids Res. 27(1999) 1558–1560.

[17] L.G. Puskás, Á. Zvara, L. Hackler Jr., T. Micsik, P. van Humme-len, Production of bulk amounts of universal RNA for DNAmicroarrays, Biotechniques 33 (2002) 898–900 902, 904.

[18] L.G. Puskás, Á. Zvara, L. Hackler Jr., P. van Hummelen,RNA ampliWcation results in reproducible microarray datawith slight ratio bias, Biotechniques 32 (2002) 1330–1340.

[19] D. Seth, M.D. Gorrel, P.H. McGuinness, M.A. Leo, C.S. Lieber,G.W. McCaughan, P.S. Haber, SMART ampliWcation maintainsrepresentation of relative gene expression: quantitative valida-tion by real time PCR and application to studies of alcoholicliver disease in primates, J. Biochem. Biophys. Methods 55 (2003)53–66.

Fig. 2. Scatter plot of gene expression ratios after LPS treatmentobtained with three diVerent protocols. Only those values which wereabove +1 and below ¡1 obtained by the nonampliWcation protocolwere selected. Gene expression ratios corresponding to these spotlabels gathered from exponential and overampliWed sample ampliWca-tion protocols are also shown.


[20] T.C. Freeman, K. Lee, P.J. Richardson, Analysis of gene expres-sion in single cells, Curr. Opin. Biotechnol. 10 (1999) 579–582.

[21] G. Brady, M. Barbara, N.N. Iscove, Representative in vitro cDNAampliWcation from individual hemopoietic cells and colonies,Methods Mol. Cell. Biol. 2 (1990) 17–25.

[22] M.W. PfaZ, A new mathematical model for relative quantiWcationin real-time RT-PCR, Nucleic Acids Res. 29 (2001) e45.

[23] W.M. PateWeld, Algorithm AS159. An eYcient method of generat-ing r £ c tables with given row and column totals, Appl. Stat. 30(1981) 91–97.

previous phd theses: application of microarrayalbert szent-györgyi ph.d. library 1 liliána z....

Documents