2.1.3 Localization and clinical presentation
Due to its neural crest origin, neuroblastoma may occur anywhere along the sympathetic chain. Hence, presenting signs and symptoms are highly variable and commonly a manifestation of tumour location. About half of all neuroblastomas arise in the adrenal medulla, and the remainder originate in paraspinal sympathetic ganglia in the chest, neck or abdomen, or in pelvic ganglia (Brodeur and Maris, 2006). The intraabdominal neuroblastoma, especially when retroperitoneal, tends to present as an incidental finding detected by parents or during routine clinical visits. Similarly, thoracic neuroblastoma presents often as an incidental mass on chest X-rays. Pelvic masses may produce urinary retention or constipation due to compression of the bladder or rectosigmoid colon. Primary or metastatic cervical neuroblastoma can occasionally compromise sympathetic innervations leading to Horners syndrome, i.e. miosis, ptosis and anhidrosis on the affected side of the face, or cause cervical lymphadenopathy. Around 40% of patients will present with localised disease, most of which have favourable biological features and are successfully treated with surgery alone (Matthay et al., 1989; Evans et al., 1996; Kushner et al., 1996; Alvarado et al., 2000; Perez et al.,
Review of the Literature 2000). About half of all patients present with evidence of haematogenous dissemination at diagnosis. Unlike patients with localised disease, who often have only few minor symptoms at diagnosis, patients with metastatic neuroblastoma have more extensive tumour burdens and are therefore in poorer health at presentation. Dissemination in the central nervous system is a rare event (Ahdevaara et al., 1977), but can occur especially with progression or relapse of the disease. In general, severe symptoms occur when the tumour has reached a critical size and/or developed metastases. The most common clinical symptoms are typically nonspecific symptoms, such as pain, fever, weight loss, and general malaise. If the tumour bleeds spontaneously, an acute onset of abdominal pain may occur. Bone metastases present usually with local bone pain and sometimes with sudden changes in the activity level of children. In the case of skull involvement, periorbital ecchymoses and/or proptosis may occur. Paraspinal tumours may extend through the vertebral foramina and compress the spinal cord, thereby generating motor or sensory deficits of spinal origin. During infancy, blueberry muffin syndrome, i.e. painless and bluish subcutaneous nodules, is a relatively common feature in stage 4s disease. Blueberry muffin syndrome patients have a favourable prognosis with frequent spontaneous tumour regression. Paraneoplastic syndromes are rare, but a neuroblastoma that secretes vasoactive intestinal peptide may typically cause intractable secretory diarrhoea with subsequent dehydration and hypokalemia (El Shafie et al., 1983). Opsoclonus-myoclonus syndrome, another neuroblastoma induced paraneoplastic syndrome, consists of involuntary jerking limb movements, ataxia, and rapid conjugated eye movements. These symptoms are probably caused by autoantibodies against the cerebellar and/or cerebral neural tissue. Symptoms of opsoclonus-myoclonus syndrome tend to persist despite successful treatment of neuroblastoma, and they may result in neurodevelopmental delay (Hiyama et al., 1994).

Table 1. International Neuroblastoma Risk Group (INRG) Consensus Pretreatment Classification schema (table from Cohn et al., 2009). GN = ganglioneuroma, GNB = ganglioneuroblastoma.

2.1.5.1 Stage

In an effort to facilitate comparison of the results of clinical trials performed throughout the world, the International Neuroblastoma Staging System (INSS) was published in 1993 (Brodeur et al., 1993), and it has been in use in numerous European and North American countries ever since. The INSS stratification is based partially on surgical result (Table 2). Stages 1, 2 and 4s are commonly considered to be favourable, but infants (under one year of age) belong to the favourable group regardless of the INSS group (Castleberry et al., 1994). Since surgical approaches differ from one institution to another, and not all neuroblastomas are treated with surgery, the INSS definitions for patients with locoregional disease vary substantially between institutes. Thus, the staging system is not highly reliable. The 4-year OS rate for infants with the INSS definitions 1, 2, 3 and 4s is 98.5%, whereas the OS for the INSS definition 4 patients is 73.1% (Ikeda et al., 2002). Interestingly, the 4-year OS rate for patients over one year of age is 100% in stage 1, 2, 3, and 4s tumours, but only 48.5% for stage 4 tumours (Ikeda et al., 2002). In the proposed INRG staging system (Monclair et al., 2009), the extent of neuroblastoma is defined by pretreatment imaging studies and bone marrow morphology. Radiological features distinguish locoregional tumours involving local structures (INRG stage L1) from locally invasive tumours (INRG stage L2), and stages M and MS are proposed to categorise neuroblastomas that are widely disseminated or have an INSS 4S pattern of disease (Table 2), respectively (Monclair et al., 2009). In terms of documenting stage and extent of the disease, the generally utilized method is still the INSS definition (Table 2) (Brodeur et al., 1993).
Table 2. International Neuroblastoma Staging System (INSS).

2.1.5.2 Age

Age has usually been analysed as a dichotomized variable in neuroblastoma risk stratifications. Despite extensive clinical research, there is currently no consensus on age grouping for stratifying neuroblastoma patients, in particular for INSS 4 neuroblastoma patients. There is emerging evidence that the suggested cut-off age of 18 months may be less predictive in prognostic evaluations than previously thought. Moreover, neuroblastomas in adolescents and adults indicate that age categories have a number of limitations that should be noted when designing therapeutic protocols for older patients. Analysis of data for 3666 patients (the same patients as in the INRG classification system analysis) with neuroblastoma suggests that a cut-off of 460 days (around 15 months) might be more reliable than the cut-off age of 18 months (London et al., 2005). The 4-year event free survival (EFS) rate for patients younger than 460 days was 82%, whereas the same EFS rate for patients older than 460 days was 42% (London et al., 2005). There are ongoing clinical trials which are testing the safety of therapy reduction in children between 12 and 18 months of age with INSS 3 or 4 neuroblastoma and favourable biological characteristics (no MYCN amplification).

Table 3. List of some tumour markers of neuroblastoma that were identified by the systematic review by Riley et al (table from Riley et al., 2004).
2.2.1 Neuroblastoma phenotype
A tumour phenotype describes observable tumour characteristics, such as histopathology and expression characteristics. Different phenotypes result from combined effects of acquired genetic code and environmental factors. A tumour phenotype, in general, is an important stratification factor in the evaluation and design of treatment modalities.

2.2.1.1 Histology

Evaluation of the degree of cellular differentiation within a neuroblastoma is the main role of histological analyses. Neuroblastoma is the commonly used term for all types of neuroblastic tumours, which are traditionally divided on the basis of differentiation and Schwannian stromal development into four different histopathological subgroups: neuroblastoma, intermixed ganglioneuroblastoma, nodular ganglioneuroblastoma and ganglioneuroma (Hicks and Mackay, 1995; Shimada et al., 1999a; Shimada et al., 1999b; Peuchmaur et al., 2003). These subgroups are further divided into more specific subtypes by the International Neuroblastoma Pathology Committee (INPC), which has published a widely accepted and used classification method of neuroblastomas (Shimada et al., 1999a; Shimada et al., 1999b). In brief, neuroblastomas (undifferentiated, poorly differentiated, differentiating) are rich in neuroblasts with little Schwannian stroma. Ganglioneuroblastomas (intermixed, nodular) belong to an intermediary category possessing features of both the immature neuroblastomas and differentiated ganglioneuromas. Ganglioneuromas (maturing, mature) are well-differentiated benign tumours with mature ganglion cells, increased stroma compartments and sparse neuroblasts. The International Classification system distinguishes a favourable and unfavourable histology group on the basis of age and mitosis-karyorrhexis index (MKI) (Shimada et al., 1999a; Shimada et al., 1999b). Tumours in the favourable group include e.g. poorly differentiated subtype of neuroblastoma with either a low (defined as fewer than 100 mitotic or karyorrhectic cells per 5000 cells) or an intermediate (100200 per 5000 cells) MKI (<18 months of age), differentiating subtype with a low MKI (<60 months of age), intermixed ganglioneuroma, and ganglioneuroma (Shimada et al., 1999a; Shimada et al., 1999b). In contrast, tumours in the unfavourable histology group include e.g. undifferentiated subtype (any age), poorly differentiated subtype (18 months), differentiating subtype (60 months), high (>200 per 5000 cells) MKI (any age), intermediate MKI (18 months), and low MKI (60 months) in the neuroblastoma category. According to the revised INPC system (Peuchmaur et al., 2003), the prognostic group for ganglioneuroblastoma is determined by evaluating the grade of neuroblastic differentiation and MKI of the neuroblastomatous nodule(s) using the same age-dependent criteria described above (Shimada et al., 1999a; Shimada et al., 1999b).

Review of the Literature a better EFS prognosis than older patients (4-year EFS of 93% in a 12- to 18month-old subgroup versus 38% in a 19- to 24-month-old subgroup), whereas children with MYCN-amplified neuroblastomas and patients older than 24 months had a poor prognosis regardless of ploidy (George et al., 2005). Hyperdiploid and near triploid tumours in infants may have whole chromosome gains without structural rearrangements, whereas hyperdiploid and near triploid tumours in older patients most probably have several genetic aberrations. Taken together, flow cytometry and cytogenetic investigations suggest that hyperdiploid neuroblastomas are distinctly different from diploid neuroblastomas, and that ploidy is a significant prognostic factor in patients under 1824 months of age.
2.2.3 Summary of prognostic molecular characteristics of neuroblastoma
Prognostic risk group stratifications on the basis of cancer biology may improve the accuracy of prognostic prediction by replacing or supplementing clinical prognostic factors. Minimizing treatment-associated morbidity by reserving the most heavy and toxic treatments for neuroblastoma patients with the worst prognosis has become an important objective, given that permanent morbidity is the factor that has the most profound effect on the quality of life of a cured child. Thus far, prognostic risk groups have not been optimized to maximize treatment success while minimizing morbidity, and there is an unquestioned need to identify and define more precise prognostic factors to stratify treatments. To date, the number of published analyses including a considerable number of neuroblastoma patients (> 100 patients) is still limited, and only a relatively small number of studies have reported multivariate survival analyses of multiple clinical and molecular factors. Moreover, most of the prognostic molecular studies include patients treated with different protocols, which further causes a considerable confounding effect, in addition to the well-known heterogeneity of neuroblastoma itself. Therefore, the results of these analyses should be interpreted with caution. MYCN amplification status is the only molecular characteristic of neuroblastoma with a definite independent value in treatment stratifications. Despite very intensive research efforts, no single genetic aberration has been found to be present in all neuroblastomas, and none has been identified crucial for tumourigenesis. The significance of the most commonly reported prognostic molecular markers, VMA:HVA ratio, TrkA, TrkB, NSE, LDH, ferritin, MRP, CD44, LOH 1p36, 17q gains and DNA index (tumour ploidy), has been identified in a number of univariate analyses. As mentioned earlier, only a few of these remain significant and powerful enough after multivariate analysis and correction for the most important variables, which are age, stage and MYCN

Review of the Literature Kononen et al (Kononen et al., 1998). Up to 200 consecutive sections (48 m thick) can be cut from a single TMA block. Research on heterogeneous tissues or tumour samples requires numerous samples of various regions in order to be representative enough for molecular analyses. In this sense, TMA does not have an obvious advantage over conventional methods. For example, in liver research, it is often necessary to evaluate staining patterns of larger samples than an area of about 3 mm2 (2 mm diameter) or 0.27 mm2 (0.6 mm diameter), which cannot include the minimum requirement of one portal tract and one central vein in the same sample. Therefore, for the use of TMAs in the research of heterogeneous tumours, defining a strict methodology to select best representative spots to be punched from a wide area of a donor tissue block could further facilitate the usage of TMAs.
Figure 2. A tissue microarray microscopic slide showing numerous punched tissue samples on one microscopic slide. Thin sections of the paraffin-embedded tissue samples in a multitissue array block are cut and fixed on a microscopic slide for subsequent molecular analyses.
2.3.2 Gene amplification analyses
Two techniques predominate in the detection of MYCN amplification in clinical neuroblastoma samples: a.) time-consuming (12-week turnaround) Southern blotting (requires 510 g of DNA), and b.) genomic PCR (requires only 50-100 ng of DNA). Both of these techniques provide semiquantitative information, i.e. yield an approximation of the number of MYCN gene copies, but the results presented as an average of a whole tumour sample do not represent the small proportion of highly malignant cells, and therefore the results may be
Review of the Literature misleading. For example, in Southern blot analysis of MYCN copy number, the sample under evaluation may contain only a small subpopulation of MYCNamplified cells, and therefore nonamplified cells may diminish the average copy number of MYCN. Together with a high cut-off value, this may result in a significant misinterpretation. Similarly, PCR-based methods for the detection of genetic aberrations or gene copy numbers are extremely sensitive to the amount of normal DNA contaminating the tumour sample. An alternative approach for Southern blot analysis and PCR-based methods is FISH (Shapiro et al., 1993), which provides results that are more reliable than those from molecular analyses. Indeed, the FISH approach has several advantages. There is a minimal requirement for tumour material, since FISH is readily applicable to tumour touch imprint slides (Taylor et al., 1994), fine needle aspirates (Frostad et al., 1999), and paraffin wax embedded tissue sections (Leong et al., 1993). Most importantly, FISH enables the examination of tumour samples at an individual cell level, and hence it is clearly more specific and sensitive than Southern blot and PCR-based analyses. In other words, FISH can detect gene amplifications even in a few sample cells without difficulty (Lorenzana et al., 1997). The importance of this is especially evident in neuroblastoma, where a mixture of cells with copy numbers ranging from less than ten to many hundreds is commonly seen (Shapiro et al., 1993; Squire et al., 1996). The widespread use of FISH has been hampered due to its expensive (confocal microscopes and rapidly fading fluorescent probes) and laborious nature. To overcome these practical limitations, a chromogenic in situ hybridization (CISH) technique, in which the DNA probe is detected using a simple immunohistochemistry-like peroxidase reaction instead of fluorescent probes, has been developed (Tanner et al., 2000). However, even though CISH is routinely used in the detection of HER-2 oncogene amplification in breast carcinoma, its applicability in evaluating MYCN amplification in neuroblastoma samples has not been widely studied.

Review of the Literature Currently, the most commonly used platform contains short oligonucleotides that are synthesized on the microscopic slide itself. Microarray technology has multiple possibilities in genomic research, including highly multiplexed genotyping and polymorphism analyses, evolutionary studies and monitoring the binding of proteins to nucleic acids and other proteins. More sensitively detected genomic DNA alterations integrated with mRNA expression profiles may further facilitate important pathway discoveries, and transcriptional profiles can perhaps be used to define prognostic signatures. The most recent development in the array technology is a single nucleotide polymorphism (SNP) array that has greatly facilitated the detection of deletions.
Figure 3. The figure shows the main principle of aCGH using BACs. (A and B) BACs can be selected and prepared based on e.g. genome maps. (C) Clones are automatically spotted on a microarray slide. (D) After preparing reference and test DNA samples, (E) labelled DNA probes are hybridized onto the previously spotted microarray slide (C). (F) Image analysis is performed using automated fluorescence scanners and a special software, which converts the scanned information to graphical illustrations (G and H) (figure from Garnis et al., 2004).
2.3.4 In silico screening
In silico is a Latin expression that translates into "performed on computer or via computer simulation." Gene discovery and analysis using in silico approaches is becoming a rapidly expanding and powerful tool in bioinformatics. Public databases, such as dbEST, GenBank, UniGene and Ensembl, are used for the collection, distribution and identification of, for instance, new genes and gene sequences. The largest human expressed sequence tag (EST) database is UniGene, which contains 000 ESTs (http://www.ncbi.nlm.nih.gov/UniGene/UGOrg.cgi?TAXID=9606). Ensembl is a software system project producing genome databases for vertebrates and other eukaryotic species, and the European Molecular Biology Laboratory, European Bioinformatics Institute and the Wellcome Trust Sanger Institute govern the data. ESTs originate from cDNA libraries prepared from specific cell types, tissues, or organs, and are usually short (typically 400600 bp long) cDNA fragments representing a transcribed mRNA. If each cDNA library was a representation of all mRNA transcripts in that specific tissue, the frequency of each EST throughout the UniGene database, for example, would represent the level of that mRNA of a specific gene in the tissue studied. In other words, gene-specific expression could be directly quantified in relation to the total number of mRNAs in that tissue- or cell-specific library. Data mining of the human ESTs has been used to evaluate tissue-specific gene expression profiles (Vasmatzis et al., 1998; Hwang et al., 2000). Construction strategies of cDNA libraries together with techniques used when obtaining ESTs may strongly bias the representativeness of libraries. For example, ESTs obtained from a random-primed cDNA library contain most probably more than one EST per mRNA template, whereas subtracted or normalized cDNA libraries represent only differentially transcribed gene transcripts. In contrast, if ESTs were obtained using oligo(dT) priming, an equal representation of all polyA-tailed transcripts in the target tissue is probable. In addition, each sequencing method of ESTs has its benefits and disadvantages. By using, for example, the UniGene Digital Differential Display tool (http://www.ncbi.nlm.nih.gov/Tools/), the comparison of EST-based expression profiles between several libraries in UniGene becomes possible, thus enabling the identification of genes that differ in transcription frequency in libraries of different tissues. Other useful tools are SAGEmap and the GEO Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/Tools/), which are databases designed for comparison of gene expression data from different experiments. There is a great need for database integration. The enormous quantity of the data requires extensive use of bioinformatic tools to identify the genes with true tissue-specific expression patterns. For example, the use of only UniGene could provide data which is based on all available cDNA libraries. In this case, the prediction of tissue-specific gene expression and putative function of the

4.4 Chromogenic in situ hybridization
MYCN amplification status in neuroblastomas was determined by CISH analysis with a digoxigenin-labelled probe complementary to the MYCN gene (Spot-Light N-Myc Probe, Zymed, South San Francisco, CA, USA) as described (Rummukainen et al., 2001). Briefly, hotspot tumour tissue microarray slides and three bone marrow samples were hybridized after deparaffination, denaturation and dehydration with 10 ml of probe cocktail (2 ml of digoxigenin-labelled MYCN probe, 1 ml of 9.9mg/ml human placental DNA, and 1 ml of 1mg/ml Cot-1 DNA (Roche Molecular Biochemicals, Mannheim, Germany), and 6 ml of master mix (Rummukainen et al., 2001). Hybridization was carried out after codenaturation of the probe mixture in a humid chamber at 37oC for 1624 h. After hybridization, the slides were washed, and the MYCN probe detected as described (Rummukainen et al., 2001). Microscopy
Materials and Methods was performed after light counterstaining with haematoxylin with a transmitted light microscope (Zeiss). In every hotspot tumour tissue microarray sample, 100 nonoverlapping tumour cell nuclei were randomly scored to determine the number of MYCN signals. A tumour sample was considered to be MYCN amplified, when an average of 6.00 or more nuclear signals per cell were seen, or when tumour nuclei showed large clustered signals. No adjustment for potential hyperploidy was made.

4.5 Immunohistochemistry

Paraffin-embedded hotspot tumour tissue microarray slides, bone marrow core biopsy samples on slides, and bone marrow aspiration smears were used for conventional immunostainings. After the slides were deparaffinized, microwave pretreatment was performed in 0.01 mol/l citrate buffer (pH 6.0, 2x5 min, 750 W) for Id2, nestin and NCAM antigen retrievals, in Tris-EDTA buffer (pH 9.0, 3x7 min, 850 W) for c-kit protein, and also in Tris-EDTA (pH 9.0, 2x7 min, 850 W) for Ki-67 protein. For polySia-NCAM and simultaneous double-labelling with anti-NCAM antibodies, slides were deparaffinized, rehydrated in a descending ethanol series and stained conventionally. Rabbit anti-human nestin IgG antibody at a concentration of 1 g/ml (ImmunoBiological Laboratories, Japan), and rabbit anti-Id2 polyclonal antibody at a concentration of 24 g/ml (Zymed Laboratories, San Francisco, CA, USA) were used as primary antibodies. PolySia-binding fluorescent fusion protein (EndoNA2-GFP) at a concentration of 10 g/ml was used for polySia detection (Jokilammi et al., 2004). Mouse anti-human NCAM antibody (123C3) at a concentration of 4 g/ml (Santa Cruz Biotechnology, Santa Cruz, CA) was used as a primary antibody. All antibody incubations were carried out overnight at 4oC. In immunohistochemistry, antibody detection was carried out using the anti-rabbit HRP polymer (PowerVision; ImmunoVision Technologies, Daly City, CA, USA). The immunoreaction was visualised with 3-aminoethyl-carbazole (AEC) for 10 min (ready-to-use, LabVision, Fremont, CA, USA), or DAB as chromogen. Slides were slightly counterstained with haematoxylin, and mounted with Faramount (DakoCytomation). In immunofluorescence, the Alexa Fluor 594 chicken anti-rabbit secondary antibody and Alexa Fluor 594 chicken anti-mouse secondary antibody (Molecular Probes, Eugene, OR, USA) were used, and slides were mounted with Immu-Mount (Shandon, USA). Expression of the c-kit protein was detected using the polyclonal rabbit antihuman c-kit antibody (CD117; 1:250 dilution; DakoCytomation, Glostrup, Denmark). Immunostaining was performed in an automated immunostainer (TechMateTM 500 Plus; DakoCytomation) using the avidinbiotin complex method with diaminobenzidine as the chromogen. A GIST tumour served as a positive control. Mast cells served as internal positive controls and as a reference for positive immunolabelling. Immunostaining in more than 30% of the cells was considered to be a positive reaction (Vitali et al., 2003). A cut-off

Figure 4. Fluorescence staining patterns of a primary neuroblastoma. (A) Expression of polySia NCAM (stained with an EndoNA2-GFP fusion protein), and (B) NCAM (stained with Alexa Fluor 594) at the same tumour site. (C) An overlay image of A and B identifying colocalized expression of polySia-NCAM and NCAM. Scale bar, 40 m.
5.6 Detection of polySia-NCAM-positive neuroblastic cells in bone marrow with immunofluorescence and flow cytometry (III)
To investigate whether metastatic neuroblastoma cells in bone marrow express polySia-NCAM, as do their polySia-NCAM-positive primary tumour foci, 19 paraffin-embedded bone marrow biopsies from six different patients were labelled with the fluorescent fusion protein. PolySia-NCAM-positive tumour cell clusters were found in two different bone marrow samples, taken at different time points from the same patient with a polySia-NCAM positive primary tumour. Contrary to the primary tumours, polySia-NCAM-positive bone marrow neuroblastoma cells appeared to be in a nonproliferative state, when samples were double-labelled with anti-Ki-67 antibody. Fresh bone marrow cells and in vitro cultured polySia-NCAM-positive SHSY5Y neuroblastic cells were mixed together in different ratios, and the mixed cell suspensions were labelled with the polySia-binding fluorescent fusion protein. Interestingly, smears with added polySia-NCAM-positive SH-SY5Y cells appeared as cell clusters, as in fixed bone marrow biopsy samples, whereas smears without added SH-SY5Y cells did not show any cell clustering. Normal bone marrow samples were considered polySia-NCAMnegative, even though they contained a few single cells (approximately 1-3 individual cells per microscopic field), which were polySia-NCAM-positive, but significantly smaller in size than polySia-NCAM-positive tumour cell clusters. In evaluating the possible applicability of flow cytometry to differentiate bone marrow metastases (polySia-NCAM-positive neuroblastoma cells or cell clusters) from normal bone marrow cells, our results revealed that polySiaNCAM-positive tumour cells produce distinct fluorescence emission wavelengths when compared to fusion protein-labelled normal bone marrow cells. As we had no permission to take fresh bone marrow samples from
Results neuroblastoma patients for research purposes, we had to use the abovedescribed artificially created stage 4 bone marrow samples to test the applicability of polySia-based detection of neuroblastoma cells in fresh samples.
5.7 Association of polySia-NCAM and NCAM expression with clinical parameters (age and stage) (III)
Patients age did not associate with polySia-NCAM or NCAM expression, not even when groups were dichotomized at the ages of 12 or 18 months. However, polySia-NCAM expression in neuroblastomas did associate with clinical stage, as patients with polySia-NCAM-expressing primary tumours had advanced (i.e. metastatic) disease at diagnosis (P = 0.047). The association between NCAM expression and clinical stage was nearly significant (P = 0.053).

Discussion al., 2006), between DIABLO expression and cancer progression. In this sense, the carcinogenetic role of DIABLO is ambiguous, and needs to be examined more thoroughly, in neuroblastoma, too. In summary, in addition to the previously identified oncogenes in neuroblastoma, our analyses led to the identification of the high frequency of gain of 12q24.31, which associated with intermediate prognosis in neuroblastoma patients. This low-level alteration may serve as a novel additional biomarker to assess neuroblastoma progression and prognosis.
6.2 Novel methods and strategies in molecular neuroblastoma research
Rapidly developing molecular techniques have revealed many potentially important genetic aberrations of neuroblastomas in recent years. Despite the technical progress, there is still no uniform method to assess the only prognostically significant example of oncogene activation, MYCN amplification. The classical technique for analysing MYCN amplification has been Southern blot hybridization. In order to assess the MYCN amplification status with Southern blot, it is necessary to extract DNA from a sufficient amount of fresh tumour sample. To analyse archival tumour samples, FISH has been used (Shapiro et al., 1993). The poor availability of fluorescence microscopes in clinical diagnostics in many countries has been one of the major obstacles for the widespread use of the FISH technique, and FISH has remained a complementary method in studies of MYCN amplification.
6.2.1 Hotspot-CISH technique in the analysis of MYCN
We have formulated a new MYCN copy number assessing strategy that utilizes the CISH method and the hotspot selection of neuroblastomas. Currently, an overall estimation of various tumour progression-associated parameters in one randomly selected spot may lead to faulty reasoning. In addition, results presented as an average of a molecular analysis of a whole tumour sample do not represent the small proportion of highly malignant cells, and therefore the results may be misleading. For example, in Southern blot hybridization analysis of the MYCN copy number, the sample under evaluation may contain only a small subpopulation of MYCN-amplified cells, and therefore non-cancer cells may diminish the actual copy number. Together with a high border value this may result in a statistical bias. In accordance with the idea that malignant changes in protein as well as in genomic level associate with accelerated proliferation, we measured the proliferation index of the whole neuroblastoma sample with the MIB-1 antibody, and selected the focus ( 2 mm) containing the highest fraction of proliferative cells for the TMA block and for further analysis of various molecular markers. The hotspot analysis of MYCN amplification revealed that the selection of a single focus appears to be reliable

Summary and Conclusions buried in the new genomic amplicon, 12q24.31, which appears to be associated with varying tumour behaviour and patient outcome, is an example of the powerful utilization of bioinformatics and biotechnology. The array-based methods provided could already be used in diagnostic settings. Although we have made some progress in identifying neuroblastomaspecific molecular targets for diagnostic applications and novel therapeutics, much work needs to be done for thorough standardisation and validation of the reported findings. Regardless, our results warrant additional investigation.

Acknowledgements

8 ACKNOWLEDGEMENTS
This work was carried out at the Department of Medical Biochemistry and Genetics, University of Turku. I owe my deepest gratitude and respect to my supervisor professor Jukka Finne, the head of the Department of Medical Biochemistry and Genetics, who has an exemplary mentality comprised of endless patience, wisdom and strength when it comes to supervising egos like me. In addition to visionary research ideas, Jukka has been ahead of his time with a strict policy of using Macintosh computers only. Therefore, he is also responsible for my expensive macaddiction, for which I will always be grateful. My second supervisor, Docent Hannu Haapasalo, is to be thanked a thousand times for his ceaseless enthusiasm, optimism and scientific productivity. Without Hannu and his extensive scientific networking, this thesis would not have been finished before my seventies. My current master, Professor Juha Hernesniemi, from the Department of Neurosurgery, Helsinki, your fatherly whipping was an absolute necessity to guide me through the lures of new ongoing projects, to finish the thesis. My sincere compliments to you. Professor Mika Niemel, from the Department of Neurosurgery, Helsinki, I thank you for giving an example of utmost efficacy, of which I will try to learn a bit. I am grateful to Professor Raija Tammi and Docent Matias Rytt, the official reviewers of the thesis, for their constructive and positive criticism. I owe a debt of numerous thanks to the whole staff at the Department of Medical Biochemistry and Genetics for the pleasant atmosphere at work. Professors Eero Vuorio, Risto Penttinen and Mikael Skurnik are acknowledged for providing the opportunity to use all the facilities and equipments of their own research groups. Of the many former and current workers of the Department, I would like to mention a few. A very special mention goes to Anne Jokilammi, who has contributed enormously to my thesis work. Jukka Hytnen and Arto Pulliainen are thanked for giving an encouraging example of graduation. No other than Rakel Mattsson is to thank for helping me in all possible matters. Without Marja Nyknen, Terttu Jompero, Marita Potila, Jukka Karhu and Liisa Peltonen this work would have taken an even longer time. I am also thankful to the rest of the staff: Harri Hirvonen, Sauli Haataja, Elina Jakobson, Vuokko Loimaranta, Annika Salminen, Pauli Ollikka, Tomi Pudas, Klaus Elenius, Hannu Jrvelinen, Outi Hirvonen, Erika Ekholm, Riku Kiviranta, Jukka Morko, Mikko Savontaus, Lassi Nelimarkka, Mirva Sderstrm, Heli SalminenMankonen, Anna-Marja Smnen, Marja-Leena Ruuskanen, Raili Salonen, Tuula Oivanen, Merja Lakkisto and Heidi Pakarinen. I wish to thank my collaborators, Professors Hannu Kalimo, Toivo T. Salmi, Jorma Isola and Olli Kallioniemi, who have opened up many possibilities

al. Different karyotypic patterns in early and advanced stage neuroblastomas. Cancer Res. 1987;47(1):311-8. 138. Kang JH, Rychahou PG, Ishola TA, Qiao J, Evers BM, and Chung DH. MYCN silencing induces differentiation and apoptosis in human neuroblastoma cells. Biochem Biophys Res Commun. 2006;351(1):192-7. Kang TI, Brophy P, Hickeson M, Heyman S, Evans AE, Charron M, et al. Targeted radiotherapy with submyeloablative doses of 131I-MIBG is effective for disease palliation in highly refractory neuroblastoma. J Pediatr Hematol Oncol. 2003;25(10):769-73. Kashkar H, Haefs C, Shin H, HamiltonDutoit SJ, Salvesen GS, Kronke M, et al. XIAP-mediated caspase inhibition in Hodgkin's lymphoma-derived B cells. J Exp Med. 2003;198(2):341-7. Kashkar H, Seeger JM, Hombach A, Deggerich A, Yazdanpanah B, Utermohlen O, et al. XIAP targeting sensitizes Hodgkin lymphoma cells for cytolytic T-cell attack. Blood. 2006;108(10):3434-40. Kiss JZ and Rougon G. Cell biology of polysialic acid. Curr Opin Neurobiol. 1997;7(5):640-6. 152. 143. Kleene R and Schachner M. Glycans and neural cell interactions. Nat Rev Neurosci. 2004;5(3):195-208. Kohl NE, Kanda N, Schreck RR, Bruns G, Latt SA, Gilbert F, et al. Transposition and amplification of oncogene-related sequences in human neuroblastomas. Cell. 1983;35(2 Pt 1):359-67. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, et al. Tissue microarrays for highthroughput molecular profiling of tumor specimens. Nat Med. 1998;4(7):844-7. Kramer K, Cheung NK, Gerald WL, LaQuaglia M, Kushner BH, LeClerc JM, et al. Correlation of MYCN amplification, Trk-A and CD44 expression with clinical stage in 250 patients with neuroblastoma. Eur J Cancer. 1997;33(12):2098-100. 147. Krams M, Parwaresch R, Sipos B, Heidorn K, Harms D, and Rudolph P. Expression of the c-kit receptor characterizes a subset of neuroblastomas with favorable prognosis. Oncogene. 2004;23(2):58895. Krushel LA, Tai MH, Cunningham BA, Edelman GM, and Crossin KL. Neural cell adhesion molecule (N-CAM) domains and intracellular signaling pathways involved in the inhibition of astrocyte proliferation. Proc Natl Acad Sci U S A. 1998;95(5):2592-6. Kushner BH, Cheung NK, LaQuaglia MP, Ambros PF, Ambros IM, Bonilla MA, et al. International neuroblastoma staging system stage 1 neuroblastoma: a prospective study and literature review. J Clin Oncol. 1996;14(7):2174-80. Kushner BH and Helson L. Monozygotic siblings discordant for neuroblastoma: etiologic implications. J Pediatr. 1985;107(3):405-9. Kushner BH, Kramer K, and Cheung NK. Phase II trial of the anti-G(D2) monoclonal antibody 3F8 and granulocyte-macrophage colonystimulating factor for neuroblastoma. J Clin Oncol. 2001a;19(22):4189-94. Kushner BH, Wolden S, LaQuaglia MP, Kramer K, Verbel D, Heller G, et al. Hyperfractionated low-dose radiotherapy for high-risk neuroblastoma after intensive chemotherapy and surgery. J Clin Oncol. 2001b;19(11):2821-8. Kushner BH, Yeh SD, Kramer K, Larson SM, and Cheung NK. Impact of metaiodobenzylguanidine scintigraphy on assessing response of high-risk neuroblastoma to dose-intensive induction chemotherapy. J Clin Oncol. 2003;21(6):1082-6. LaBrosse EH, Com-Nougue C, Zucker JM, Comoy E, Bohuon C, Lemerle J, et al. Urinary excretion of 3-methoxy-4hydroxymandelic acid and 3-methoxy4-hydroxyphenylacetic acid by 288 patients with neuroblastoma and related neural crest tumors. Cancer Res. 1980;40(6):1995-2001.

155. Langer I, Vertongen P, Perret J, Fontaine J, Atassi G, and Robberecht P. Expression of vascular endothelial growth factor (VEGF) and VEGF receptors in human neuroblastomas. Med Pediatr Oncol. 2000;34(6):386-93. Lantuejoul S, Moro D, Michalides RJ, Brambilla C, and Brambilla E. Neural cell adhesion molecules (NCAM) and NCAM-PSA expression in neuroendocrine lung tumors. Am J Surg Pathol. 1998;22(10):1267-76. 165. 157. Lasorella A, Boldrini R, Dominici C, Donfrancesco A, Yokota Y, Inserra A, et al. Id2 is critical for cellular proliferation and is the oncogenic effector of N-myc in human neuroblastoma. Cancer Res. 2002;62(1):301-6. 166. 158. Lastowska M, Nacheva E, McGuckin A, Curtis A, Grace C, Pearson A, et al. Comparative genomic hybridization study of primary neuroblastoma tumors. United Kingdom Children's Cancer Study Group. Genes Chromosomes Cancer. 1997a;18(3):162-9. Lastowska M, Roberts P, Pearson AD, Lewis I, Wolstenholme J, and Bown N. Promiscuous translocations of chromosome arm 17q in human neuroblastomas. Genes Chromosomes Cancer. 1997b;19(3):143-9. Laug WE, Siegel SE, Shaw KN, Landing B, Baptista J, and Gutenstein M. Initial urinary catecholamine metabolite concentrations and prognosis in neuroblastoma. Pediatrics. 1978;62(1):77-83. 169. 161. Lee SM, Lee KE, Chang HJ, Choi MY, Cho MS, Min SK, et al. Prognostic significance of CD44s expression in biliary tract cancers. Ann Surg Oncol. 2008;15(4):1155-60. 170. 162. Leong PK, Thorner P, Yeger H, Ng K, Zhang Z, and Squire J. Detection of MYCN gene amplification and deletions of chromosome 1p in neuroblastoma by in situ hybridization using routine histologic sections. Lab Invest. 1993;69(1):43-50. 163. Lesley J, Hyman R, and Kincade PW. CD44 and its interaction with extracellular matrix. Adv Immunol. 1993;54(271-335. Lipinski M, Hirsch MR, DeagostiniBazin H, Yamada O, Tursz T, and Goridis C. Characterization of neural cell adhesion molecules (NCAM) expressed by Ewing and neuroblastoma cell lines. Int J Cancer. 1987;40(1):81-6. London WB, Castleberry RP, Matthay KK, Look AT, Seeger RC, Shimada H, et al. Evidence for an age cutoff greater than 365 days for neuroblastoma risk group stratification in the Children's Oncology Group. J Clin Oncol. 2005;23(27):6459-65. Look AT, Hayes FA, Nitschke R, McWilliams NB, and Green AA. Cellular DNA content as a predictor of response to chemotherapy in infants with unresectable neuroblastoma. N Engl J Med. 1984;311(4):231-5. Look AT, Hayes FA, Shuster JJ, Douglass EC, Castleberry RP, Bowman LC, et al. Clinical relevance of tumor cell ploidy and N-myc gene amplification in childhood neuroblastoma: a Pediatric Oncology Group study. J Clin Oncol. 1991;9(4):581-91. Lorenzana AN, Zielenska M, Thorner P, Gerrie B, Weitzman S, and Squire J. Heterogeneity of MYCN amplification in a child with stroma-rich neuroblastoma (ganglioneuroblastoma). Pediatr Pathol Lab Med. 1997;17(6):875-83. Luscher B and Larsson LG. The basic region/helix-loop-helix/leucine zipper domain of Myc proto-oncoproteins: function and regulation. Oncogene. 1999;18(19):2955-66. Lux ML, Rubin BP, Biase TL, Chen CJ, Maclure T, Demetri G, et al. KIT extracellular and kinase domain mutations in gastrointestinal stromal tumors. Am J Pathol. 2000;156(3):7915. Mairs RJ, Fullerton NE, Cosimo E, and Boyd M. Gene manipulation to enhance MIBG-targeted radionuclide