Tumore der Nebennieren stellen häufige Tumore dar, welche bei mindestens 3 % der Population über 50-Jähriger vorkommen. Im Gegensatz dazu ist das Nebennierenrindenkarzinom mit einer Inzidenz von 1-2 Einwohner pro Million ein sehr seltener Tumor. Da seine Prognose allerdings ungünstig, und diese maßgeblich davon abhängt wie fortgeschritten der Tumor bei Diagnosestellung ist, ist es wichtig, dass die richtige Diagnose frühzeitig gestellt wird. Bis heute ist kein zuverlässiger immunhistochemischer Nebennierenrindenkarzinom-spezifischer Marker etabliert um das Nebennierenrindenkarzinom von anderen retroperitonealen Tumoren zu differenzieren. Sasano et al. schlug bereits 1995 erstmalig den Transkriptionsfaktor Steroidogenic Factor 1 (SF1) als Marker zur Differenzierung von Nebennierenrinden- und Nicht-Nebennierenrindentumoren vor. Allerdings wurde die diagnostische Wertigkeit bisher nur in sehr kleinen Fallserien mit insgesamt nur 17 Nebennierenrindenkarzinomen untersucht. In der vorliegenden Arbeit wurde die SF1 Protein-Expression bei 163 Nebennierenrindenkarzinomen, 52 Nebennierenrinden-Adenomen, 12 normalen steroidogenen Geweben (6 Nebennieren und 6 Ovare), sowie 73 Nicht-Steroidtumoren immunhistochemisch untersucht. Hierbei zeigte sich, das SF1 bei 158 von 161 evaluierbaren Nebennierenrindenkarzinomen und bei allen Proben von normalen und gutartigen Geweben (n=64) nachweisbar war. Im Gegensatz dazu war keine der 73 Nicht-Steroidgeweben SF1 positiv, so dass die diagnostische Genauigkeit extrem gut ist (Sensitivität: 98.6 %, Spezifität: 100 %, positive und negative predictive value jeweils 100 % und 97.3 %). In einem zweiten Schritt wurde untersucht ob die Protein-Expression von SF1 beim Nebennierenrindenkarzinom auch prognostische Bedeutung hat. Hierbei zeigte sich, dass Patienten mit Tumoren mit starker SF1 Färbung (30 %) ein deutlich schlechteres tumorstadium-adjustiertes Rezidiffreies- und Gesamt-Überleben haben als Patienten mit geringer SF1 Expression (hazard ratio: 2.45). Zusätzlich zu den immunhistochemischen Untersuchungen wurden FISH Analysen durchgeführt. Hierbei zeigte sich allerdings keine signifikante Korrelation zwischen SF1 Gendosis und der SF1 Protein-Expression, so dass zu vermuten ist, dass SF1 maßgeblich auf Transkriptions- und Translationsebene reguliert wird. In einem Versuch diese Frage zu beantworten wurden zwei mutmaßliche SF1 Interaktionspartner, FATE1 und DAX1, genauer immunhistochemisch untersucht. Hierbei wurde deutlich, dass FATE1 bei 62 von 141 evaluierbaren Nebenierenrindenkarzinomen und 12 von 62 normalen und gutartigen Geweben nachweisbar war. Im Gegensatz hierzu waren alle 9 Nicht-Steroidgewebe FATE1 negativ. Dies zeigt, das FATE1 nicht zur Diagnostik nutzbar ist (Sensitivität: 61 %, Spezifität: 100 %, positive und negative predictive value 100 % bzw. 14 %). Die DAX1 Analyse zeigte, dass alle 20 normalen und gutartigen Gewebe eine positive DAX1 Färbereaktion zeigten. Von 126 Nebennierenrindenkarzinomen waren 71 DAX1 positiv. Von den 8 untersuchten Nicht-Steroidgeweben waren 6 DAX1 positiv. Diese Ergebnisse belegen, dass auch DAX1 keine diagnostische Genauigkeit besitzt (Sensitivität: 56 %, Spezifität: 25 %, positive und negative predictive value 92 % bzw. 4 %). Die Untersuchung der prognostischen Fähigkeiten von FATE1 und DAX1 zeigte, dass Patienten mit Tumoren mit starker FATE1 Färbung (39 %) ein schlechteres tumorstadium-adjustiertes Gesamt- aber nicht Rezidiffreies-Überleben haben als Patienten mit niedriger FATE1 Protein-Expression (hazard ratio: 2.01). Weiterhin wurde deutlich, dass DAX1 keine deutlichen prognostischen Fähigkeiten besitzt. Zusammenfassend läßt sich aus der vorliegenden Arbeit folgern, das SF1 aktuell der beste diagnostische Marker zur Diagnose von Tumoren der Nebennierenrinde ist und damit Eingang in die histopathologische Routine-Diagnostik von Nebennierentumoren finden wird. Zusätzlich ist die SF1 Expression ein sehr guter prognostischer Marker beim Nebennierenrindenkarzinom, wobei sich die prognostische Aussage durch zusätzliche Färbung von FATE1 und DAX1 nur unwesentlich verbessern läßt. / Adrenal tumors are common tumors which are present in at least 3 % in the human population over their 5th decade. However, adrenocortical carcinoma (ACC) is a rare malignancy which shows an approximate anual incidence of 1-2 per million. Prognosis of ACC is generally poor and depends strongly on the tumor stage. Thus, early and correct diagnosis is important. Until now, no reliable immunohistochemical ACC-specific marker has been established for its differentiation from other retroperitoneal tumors. Already in 1995, Sasano et al. suggested the transcription factor Steroidogenic Factor 1 (SF1) as useful marker for differentiation of adrenocortical and non-adrenocortical tumors. Up to now, SF1's value as diagnostic marker for ACC was investigated only in small series of in a total of 17 samples. In our work, SF1 expression was investigated by immunohistochemistry in 163 ACC, 52 adrenocortical adenomas, 12 normal steroidogenic tissues (6 adrenal glands and 6 ovaries), as well as 73 non-steroidogenic tumors. SF1 protein expression was shown in 158 of a total of 161 evaluable ACC, as well as all normal and benign steroidogenic tissues. In contrast, no SF1 protein was detectable in the non-steroidogenic tumors. Thus, SF1 protein expression is a highly specific diagnostic tool (sensitivity: 98.6 %, specificity: 100 %, positive and negative predictive value: 100 % and 97.3 %, respectively). In a second step, SF1 protein expression was investigated as a prognostic tool in ACC. As shown by us, ACCs presenting strong SF1 immunoreactivity (30 %) showed a strong correlation with overall and recurrence-free patients survival than ACCs presenting low SF1 protein expression (hazard ratio: 2.45). Moreover, FISH analyses were performed which revealed no significant correlation of SF1 gene dosis and SF1 protein expression, suggesting a regulatory mechanism at transcriptional and translational level. To investigate the hypothesis, we investigated two putative interaction partners of SF1, namely FATE1 and DAX1 protein, by immunohistochemistry. FATE1 protein was expressed in 62 of a total of 141 evaluable ACC as well as 12 of a total of 62 normal and benign steroidogenic tissues. In contrast, all non-steroidogenic tissues were FATE1 negative (n=9). Thus, FATE1 is no valuable diagnostic tool (sensitivity: 61 %, specificity: 100 %, positive and negative predictive value: 100 % and 14 %, respectively). DAX1 immunohistochemistry showed that all normal and benign steroidogenic tissues (n=20) were DAX1 positive as well as 71 of a total of 126 ACC samples. Furthermore, 6 out of a total of 8 non-steroidogenic tissues stained DAX1 positive, showing that DAX1 protein is no diagnostic tool (sensitivity: 50 %, specificity: 25 %, positive and negative predictive value: 92 % and 4 %, respectively). Investigation of the prognostic value of FATE1 and DAX1 revealed that patients with tumors characterized by strong FATE1 immunoreactivity (39 %) had a worse outcome in overall but not recurrence-free survival than patients showing low FATE1 expression (hazard ratio: 2.01). DAX1 protein expression has no prognostic value in ACC. In summary, we showed that SF1 is currently the best available diagnostic marker for differentiation of adrenocortical tumors from other retroperitoneal tumors, and that it will be suitable for histopathological diagnostic routine. Furthermore, SF1 expression is a well-suited prognostical tool in adrenocortical carcinoma which is only marginally enhanced by subsequent staining of FATE1 and DAX1 protein.

BACKGROUND: Breast cancer is the most commonly diagnosed cancer and second leading cancer death cause among females in the U.S.A. About 1 in 8 women in U.S will develop invasive breast cancer over the course of her lifetime. In 2013, 234,580 new invasive breast cancer cases are expected to occur in women within the US and approximately 64,640 non-invasive carcinomas in situ were diagnosed in 2013, most of which were ductal carcinoma in situ (DCIS). Along with technological advances, a wide variety of candidate biomarkers have been proposed for cancer diagnosis and prognosis, including DNA content and non-coding RNA. Current techniques for detecting DNA content abnormalities in formalin-fixed, paraffin-embedded (FFPE) tissue samples by flow cytometric analysis have used cells recovered from ≥50µm whole tissue sections. Here, in our first study, a novel core punch sampling method was investigated for assessing DNA content abnormalities and intratumoral heterogeneity in FFPE specimens. Secondly, long non-coding RNAs (lncRNAs) has been examined. LncRNA participates in a broad spectrum of biological activities by diverse mechanisms and its dysregulation is associated with tumorgenesis. Some lncRNAs may function as oncogenes (O) and others as tumor suppressor genes (TSG). To date, lncRNA has been investigated primarily by qRT-PCR and RNA sequencing. This study has examined the relationship of lncRNA expression patterns to breast tumor pathology by chromogenic in situ hybridization (CISH). METHODS: Firstly, FFPE breast carcinoma specimens were selectively targeted using 1.0 mm diameter punch needles. Extracted cores were assayed by flow cytometry using a modified-Headley method. Secondly, the lncRNA expression levels of 6 lncRNAs: HOTAIR, H19, KCNQ1OT1, MEG3, MALAT11 and Zfas1, was examined by RNAscope® CISH using FFPE breast tissue microarrays (TMAs) comprising normal adjacent epithelia (NA), DCIS, and invasive carcinoma (IC) from 46 patients. LncRNA associate polycomb complex protein EZH2 was evaluated by immunohistochemistry (IHC). LncRNA data was also compared to standard breast tumor data including ER, PR, Her2 and Ki67 IHC. SYSTAT version 11 statistical package was used to perform for all the tests. RESULTS: Following optimization experiments of the core punch flow cytometric approach, DNA index and percent S-phase fraction intratumoral heterogeneities were detected in 10/23 (44%) and 11/23 (47%) specimens respectively. The lncRNA CISH study utilized a TMA that contained 36 spots of NA breast tissues, 34 DCIS spots and 43 IC spots. HOTAIR CISH staining was significantly stronger in IC than DCIS (p CONCLUSION: Core-punching is an effective alternative to whole specimen sectioning and shows that macro-level genomic heterogeneity is common even within a single FFPE block. The interrelationship of DNA content heterogeneity to other forms of heterogeneity requires further study. RNAscope CISH supports bright-field microscopy investigations of lncRNA expression in FFPE tissue specimens. HOTAIR, H19 and KCNQ1OT1 may be potential breast cancer biomarkers, both HOTAIR and H19 may be a marker for DCIS at increased risk of progression to invasive cancer. HOTAIR, in particular, may be a predictor for invasive cancer grade.

FAPα ist ein Membranglykoprotein mit Dipeptidyl-Peptidase- und Typ I Kollagenase-Aktivität, das eine wichtige Rolle im Rahmen des Gewebeumbaus und der Angiogenese spielt. Die Expression von FAPα wurde für eine Reihe von Geweben gezeigt. So konnte bereits gezeigt werden, dass FAPα nach Myokardinfarkt auf humanen kardialen Fibroblasten (HCF) verstärkt exprimiert wird. Ausgehend von der Erkenntnis, dass eine im Blut zirkulierende Form von FAPα, APCE, gefunden wurde, sollte in der vorliegenden Arbeit eine mögliche Assoziation des in HCF exprimierten FAPα mit dem zirkulierenden APCE untersucht werden. Hierfür wurden ebenfalls die Signalwege, die zur Induktion von FAPα führen, näher untersucht. Im Rahmen der vorliegenden Arbeit konnte gezeigt werden, dass das von humanen kardialen Fibroblasten exprimierte FAPα ebenfalls im Zellkulturüberstand nachgewiesen werden kann. Wie bereits beschrieben, führte eine Stimulation mit TGFβ1 zur Expression von FAPα. Diese ist abhängig von Smad-3 und konnte durch Zugabe des Inhibitors SB431542 blockiert werden. Auch Smad-2 scheint die FAPα-Expression zu beeinflussen. Ein Effekt von EGR-1, CTGF und TNFα auf die FAPα-Expression konnte hingegen nicht nachgewiesen werden. Darüber hinaus wurde im Rahmen der vorliegenden Arbeit ein ELISA zur Messung von humanem FAPα im Plasma erstmalig etabliert. In der Evaluation des ELISA wurden die Grenzwerte für die Detektion und Quantifizierung bestimmt, die Intra- und Intervariabilität des ELISAs berechnet und die Linearität der Wiederfindung von humanem rekombinantem FAPα bewertet, sowie die Stabilität von FAPα nach mehrfachen Einfrier-Auftau-Zyklen und nach der Lagerung bei verschiedenen Temperaturen evaluiert. Da hier nur kommerziell erwerbliche Materialien verwendet wurden, ist eine Anwendung durch Dritte leicht durchführbar. Es wurde weiterhin eine klinische Studie zur Messung der FAPα-Plasmaspiegel in 101 gesunden Blutspendern und 407 Patienten mit akutem Koronarsyndrom (ACS) sowie von 25 Patienten mit stabiler koronarer Herzerkrankung (KHK) durchgeführt. Hierbei konnten erstmals Referenzwerte für zirkulierendes FAPα im Plasma beschrieben werden. Die FAPα-Plasmaspiegel der gesunden Kontrollgruppe unterschieden sich nicht von denen der Patienten mit stabiler KHK. Bei Patienten mit akutem Koronarsyndrom war die FAPα-Konzentration im Vergleich mit den FAPα-Plasmaspiegeln der Kontrollgruppe hingegen signifikant erniedrigt. Auch zeigte sich bei den Patienten mit FAPα-Spiegeln im kleinsten Quartil eine erhöhte Mortalität. Die biologische Funktion von FAPα bzw. dessen mögliche Pathophysiologie im Rahmen eines ACS sowie die mit niedrigen FAPα-Spiegeln assoziierte Mortalität sind noch unklar und bleiben Gegenstand der weiteren Forschung. / FAPα is a membrane protein with dipeptidyl-peptidase and type-I-collagenase activity involved in angiogenesis and tissue repair. A number of cells and tissues including human cardiac fibroblasts (HCF) after myocardial infarction express FAPα. Due to the recent discovery of a circulating form of FAPα, APCE, a possible association between the cell-bound form of FAPα and APCE was studied in HCF. Furthermore, the signaling pathways that lead to an induction of FAPα were analyzed. It could be demonstrated that HCF shed FAPα, which can be detected in the cell culture supernatant. Moreover, in accordance with earlier results, FAPα expression is induced by TGFβ1. This is dependent on Smad-3 and could be blocked by the inhibitor SB431542. Additionally, Smad-2, but not EGR-1, CTGF or TNFα are involved in FAPα expression. Within this work an ELISA for the quantification of FAPα in human plasma was established and the limits of detection and quantification, the intra- and interassay variability and the linearity of recovery of human FAPα were determined. Furthermore, the stability of FAPα after multiple thaw-freeze-cycles or different storage conditions was evaluated. Because of the exclusive use of commercially available products, others could easily repeat this ELISA. FAPα plasma concentrations in 101 healthy blood donors, 407 patients with acute coronary syndrome (ACS) and in 25 patients with stable coronary artery disease were quantified by ELISA. This is the first time reference values for FAPα plasma levels could be derived. There was no significant difference between healthy donors and patients with stable coronary artery disease. However, compared to the healthy control group, FAPα levels were significantly lower in patients with ACS. Importantly, ACS patients with FAPα levels in the lowest quartile showed an increased mortality. The biological function of FAPα and its possible pathophysiological role in ACS are still unclear and remain the object of future studies.

Plusieurs travaux s’accordent sur le fait que l’analyse de l’ADN circulant (ADNcir) est un outil à fort potentiel diagnostic, pronostic, théranostic et de suivi en oncologie clinique. Cependant, le manque de standardisation des procédures pré-analytiques, des connaissances approfondies des structures des ADNcir ainsi que les facteurs influençant la dynamique de son relargage constituent des obstacles à son transfert en pratique clinique. C’est dans ce contexte que nous avons choisi d’approfondir l’étude des structures des ADNcir d’origines nucléaire et mitochondriale, évaluer leur utilité clinique en oncologie et étudier l’effet de l’hypoxie sur leur libération. A partir d’une technique de qPCR optimisée et validée cliniquement pour l’analyse des ADN circulants, nous avons quantifié les ADN extracellulaires d’origines nucléaire et mitochondriale dans des milieux de culture cellulaire et dans des plasmas humains et murins.Nos données ont révélé l’influence des procédures pré-analytiques sur la quantification de l’ADN circulant en fonction de l’origine. En effet, nous montrons que la préparation du plasma par séparation en ficoll diminue de manière significative la quantification de l’ADNcir d’origine mitochondriale sans influencer la quantification de l’ADNcir d’origine nucléaire, ce qui suggère la présence de structures de densité différente contenant de l’ADN mitochondrial dans le plasma. D’autre part, j’ai contribué à l’évaluation d’un test de dépistage du cancer basé sur l’analyse de l’ADNcir nucléaire et mitochondrial mise au point au laboratoire. Nos résultats préliminaires montrent de façon significative la capacité diagnostic de ce test sur des cohortes de patients atteints de cancers colorectal (n = 127) versus des individus sains (n = 91) (AUC = 0,8657 ; p < 0,0001).Nous montrons in vitro que l’hypoxie module de manière différente le relargage des ADN extracellulaires en fonction de leur origine, et notamment que l’ADN extracellulaire d’origine mitochondriale semble être régulé négativement en condition hypoxique. In vivo, l’hypoxie entraîne une augmentation de la libération de l’ADNcir d’origine nucléaire (p=0,002), mais pas d’origine mitochondriale, dans le plasma de souris greffées avec des cellules tumorales de cancer de poumon.Pour conclure, les travaux réalisés au cours de cette thèse mettent en lumière : (i) l’intérêt de la mise en place de procédures pré-analytiques standardisées pour la compréhension des origines et structures des ADNcir ; (ii) le fort potentiel diagnostic de l’analyse de l’ADNcir en oncologie ; (iii) et l’influence de l’hypoxie sur le relargage de l’ADN circulant. / Different studies converge on the diagnostic, theragnostic and prognostic properties of circulating DNA analysis (cirDNA) in clinical oncology. There remain various obstacles, however, to its transfer to clinical practice. These include the lack of standardization of pre-analytical procedures, and limited knowledge of cirDNA structures and of the factors influencing the dynamics of their release. In this context, we studied the structures of cirDNA of nuclear and mitochondrial origins, evaluated their diagnostic potential, and then evaluated the effect of hypoxia on their release. Using an optimized qPCR technique previously validated in clinical studies for the analysis of cirDNA, we quantified extracellular DNA of nuclear and mitochondrial origins in cell culture medium and in human and mouse plasma.Our data also revealed the influence of pre-analytical procedures on the quantification of cirDNA, depending on its origin. Indeed, we showed that plasma preparation with Ficoll separation significantly reduces the quantification of mitochondrial cirDNA without influencing the quantification of nuclear cirDNA.In addition, we evaluated the value of nuclear and mitochondrial cirDNA analysis in a cancer-screening test developed in the laboratory. Our preliminary results demonstrated a significant discrimination between colorectal cancer patients (n=127) and healthy individuals (n=91) (AUC=0.8657; p<0.0001).We demonstrated that in vitro hypoxia modulates the release of extracellular DNA in different ways, depending on its origin, and showed that extracellular DNA of mitochondrial origin is negatively regulated. By contrast, we demonstrated in vivo that hypoxia leads to a greater release of nuclear cirDNA (p=0.002), but not of mitochondrial cirDNA, as compared to normoxia. These experiments were performed using the plasma of mice grafted with lung cancer tumor cells.In conclusion, the work carried out for this thesis highlights: (i) the importance of setting up standardized pre-analytical procedures when investigating the origins and structures of cirDNA; (ii) the strong diagnostic potential of cirDNA analysis in oncology and (iii) the influence of hypoxia on the release of cirDNA.

Vom Deletionssyndrom 22q11.2 Betroffene sind einem überdurchschnittlich hohen Risiko ausgesetzt im Entwicklungsverlauf psychisch zu erkranken. Häufige Störungsbilder sind unter anderem ADHS, Angsterkrankungen, affektive Störungen, Erkrankungen aus dem schizophrenen Formenkreis und Morbus Parkinson. Ziel der Studie war es, phänotypische Auffälligkeiten beim DS22q11 zu identifizieren, die dabei helfen könnten, Hochrisikogruppen innerhalb des Syndroms frühzeitig identifizieren zu können und in Form von Biomarkern messbar sind. Hierzu wurden die bereits in Forschung und teilweise auch in der Klinik etablierten Verfahren der transkraniellen Sonographie und der standardisierten Riechtestung eingesetzt. / Individuals affected with 22q11.2 Deletion Syndrome (22q11.2DS) are at an above average risk to develop neuropsychiatric disorders (such as attention-deficit/hyperactivity disorder, anxiety disorders, affective disorders, schizophrenia and Parkinson's disease). The aim of this study is to identify phenotypical characteristics in 22q11.2DS to help point out high-risk groups within the syndrome and then be measured as biomarkers. To achieve this, we performed transcranial sonography and olfactory testing which are known to be established clinical and research methods.

Friedreich’s ataxia (FRDA) is an inherited recessive disorder characterized by progressive neurological disability and heart disease. It is caused by a pathological intronic hyperexpansion of a GAA repeat in the FXN gene, encoding the essential mitochondrial protein frataxin. At the homozygous state, the GAA expansion induces a heterochromatin state with decreased histone acetylation and increased methylation, resulting in a partial deficiency of frataxin expression. This was established in cells from FRDA patients. We showed that the same chromatin changes exist in a GAA based mouse model, KIKI, generated in our laboratory. Furthermore, treatment of KIKI mice with a novel Histone Deacetylase Inhibitor (HDACi), 106, a pimelic diphenylamide that increases frataxin levels in FRDA cell culture, restored frataxin levels in the nervous system and heart of KIKI mice and induced histone hyperacetylation near the GAA repeat. As shown by microarrays, most of the differentially expressed genes in KIKI were corrected towards wild type. In an effort to improve the pharmacological profile of compound 106, we synthesized more compounds based on its structure and specificity. We characterized two of these compounds in FRDA patients’ peripheral blood lymphocytes and in the KIKI mouse model. We observed a sustained frataxin upregulation in both systems, and, by following the time course of the events, we concluded that the effects of these compounds last longer than the time of direct exposure to HDACi. Our results support the pre-clinical development of a therapeutic approach based on pimelic diphenylamide HDACis for FRDA. Laboratory tools to follow disease progression and assess drug efficacy are needed in a slowly progressive neurodegenerative disease such as FRDA. We used microarrays to characterize the gene expression profile in peripheral lymphocytes from FRDA patients, carriers and controls. We identified gene expression changes in heterozygous, clinically unaffected GAA expansion carriers, suggesting that they present a biochemical phenotype, consistent with data from animal models of frataxin deficiency. We identified a subset of genes changing in patients as a result of pathological frataxin deficiency establishing robust gene expression changes in peripheral lymphocytes. These changes can be used as a biomarker to monitor disease progression and potentially assess drug efficacy. To this end, we used he same methodology to characterize the gene expression profiles in peripheral lymphocytes after treatment with pimelic diphenylamide HDACi. This treatment had relevant effects on gene expression on peripheral patients’ blood lymphocytes. It increased frataxin levels in a dose-dependent manner, and partially rescued the gene expression phenotype associated with frataxin deficiency in the tested cell model, thus providing the first application of a biomarker gene set in FRDA.

This dissertation is concerned with the application of mathematical modeling and statistical signal processing into the rapidly expanding fields of proteomics and genomics. The research is guided by a translational goal which drives the problem formalization and experimental design, and leads to optimization, prediction and control of the underlying system. The dissertation is comprised of three interconnected subjects. In the first part of the dissertation, two Bayesian peptide detection algorithms are proposed to optimize the feature extraction step, which is the most fundamental step in mass spectrometry-based proteomics. The algorithms are designed to tackle data processing challenges that are not satisfactorily addressed by existing methods. In contrast to most existing methods, the proposed algorithms perform deisotoping and deconvolution of mass spectra simultaneously, which enables better identification of weak peptide signals. Unlike greedy template-matching algorithms, the proposed methods have the capability to handle complex spectra where features overlap. The proposed methods achieve better sensitivity and accuracy compared to many popular software packages such as msInspect. In the second part of the dissertation, we consider modeling and assessing the entire mass spectrometry-based proteomic data analysis pipeline. Different modules are identified and analyzed, resulting in a framework that captures key factors in system performance. The effects of various model parameters on protein identification rates and quantification errors, differential expression results, and classification performance are examined. The proposed pipeline model can be used to aid experimental design, pinpoint critical bottlenecks, optimize the work flow, and predict biomarker discovery results. Finally, the same system methodology is extended to analyze the work flow in DNA microarray experiments. A model-based approach is developed to explore the relationship among microarray data properties, missing value imputation, and sample classification in a complicated data analysis pipeline. The situations when it is suitable to apply missing value imputation are identified and recommendations regarding imputation are provided. In addition, a missing value rate-related peaking phenomenon is uncovered.

Non-small cell lung cancer (NSCLC) constitutes a clinically, histologically, and genetically heterogeneous disease entity that represents a major cause of cancer-related death. Early-stage patients, who undergo surgery with curative intent, experience high recurrence rates and the effect of adjuvant treatment is modest. Prognostic biomarkers would be of particular relevance to guide intensified treatment depending on expected outcome and moreover often infer a biological role in tumourigenesis. This thesis presents a translational study approach to establish a well-characterised NSCLC frozen-tissue cohort and to obtain a profile of each specimen with regard to genome-wide copy number alterations, global gene expression levels and somatic mutations in selected cancer-related genes. Furthermore, the generation of a formalin-fixed, paraffin-embedded tissue microarray enabled validation of findings on the protein level using immunohistochemistry. The comprehensive molecular characterisation, combined with data on clinical parameters, enabled the analysis of biomarkers linked to disease outcome. In Paper I, single nucleotide polymorphism arrays were applied to assess copy number alterations in NSCLC and associations with overall survival in adenocarcinoma and squamous cell carcinoma were described. In Paper II, we evaluated expression levels of selected stromal proteins in NSCLC using immunohistochemistry and the adhesion molecule CD99 was identified as an outcome-related biomarker in two independent cohorts. Paper III presents a strategy for prognostic biomarker discovery based on gene expression profiling, meta-analysis, and validation of protein expression on tissue microarrays, and suggests the putative tumour suppressor CADM1 as a candidate biomarker. In Paper IV, we propose a prognostic role for tumour-infiltrating IGKC-expressing plasma cells in the local tumour microenvironment, indicating an involvement of the humoral immune response in anti-tumor activity. In Paper V, we combined next-generation deep sequencing with statistical analysis of the TP53 database to define novel parameters for database curation. In summary, this thesis exemplifies the benefits of a translational study approach, based on a comprehensive tumour characterisation, and describes molecular markers associated with clinical outcome in NSCLC.

Urinary bladder cancer is a heterogeneous disease appearing in different forms, e.g. non-muscle invasive and muscle invasive. For all variants, the expression of proteins is interesting to analyze for diagnostic, predictive, prognostic and drug targeting purposes, since it reflects the altered gene expression causing the cancer. Since urothelial cells of the bladder are in direct contact with urine it is likely that this body fluid contains cancer-related proteins. In Paper I, unbiased analysis of proteins in urine from urinary bladder cancer patients and controls, using label-free quantification by mass spectrometry, was applied and four interesting proteins APOE, FGB, LRG and SERPINA1 were selected and further analyzed with western and dot blot. In Paper II, two more proteins, POLR1E and TOP2A, were validated as relevant proteins in bladder cancer urine. In Paper III and IV, the proteins GAL1 and STMN1 were investigated for their prognostic and therapeutic target potential in bladder cancer. In Paper II, III and IV, the expression of seven of the proteins were analyzed on tissue microarrays representing tumour tissue from 360 patients with different tumour stages. For the proteins identified by the urine screening approach, their protein expressions were confirmed in bladder cancer tissue. The expression level in tissue of five of the proteins, APOE, FGB, POLR1E (Paper II), GAL1 (Paper III) and STMN1 (Paper IV), increased with tumour stage, showing diagnostic relevance and three of the proteins, SERPINA1 (Paper II), STMN1 (Paper IV) and GAL1 (Paper III) had prognostic potential in urinary bladder cancer. In addition, GAL1 and STMN1 were demonstrated to be highly expressed in metastatic disease and inhibition of STMN1 reduced cell growth (Paper III and IV), indicating that these proteins are promising drug targets in urinary bladder cancer. In conclusion, the approach of this thesis has generated several candidate protein biomarkers in urine and tissue, validated with independent methods, which have the potential to improve the care for bladder cancer patients.

<p>Noble metal nanoparticles exhibit unique optical properties that are beneficial to a variety of applications, including molecular imaging. The large scattering cross sections of nanoparticles provide high contrast necessary for biomarkers. Unlike alternative contrast agents, nanoparticles provide refractive index sensitivity revealing information regarding the local cellular environment. Altering the shape and composition of the nanoparticle shifts the peak resonant wavelength of scattered light, allowing for implementation of multiple spectrally distinct tags. In this project, nanoparticles that scatter in different spectral windows are functionalized with various antibodies recognizing extra-cellular receptors integral to cancer progression. A hyperspectral imaging system is developed, allowing for visualization and spectral characterization of cells labeled with these conjugates. Various molecular imaging and microspectroscopy applications of plasmonic nanoparticles are then investigated. First, anti-EGFR gold nanospheres are shown to quantitatively measure receptor expression with similar performance to fluorescence assays. Second, anti-EGFR gold nanorods and novel anti-IGF-1R silver nanospheres are implemented to indicate local cellular refractive indices. Third, because biosensing capabilities of nanoparticle tags may be limited by plasmonic coupling, polarization mapping is investigated as a method to discern these effects. Fourth, plasmonic coupling is tested to monitor HER-2 dimerization. Experiments reveal the interparticle conformation of proximal HER-2 bound labels, required for plasmonic coupling-enhanced dielectric sensing. Fifth, all three functionalized plasmonic tags are implemented simultaneously to indicate clinically relevant cell immunophenotype information and changes in the cellular dielectric environment. Finally, flow cytometry experiments are conducted utilizing the anti-EGFR nanorod tag to demonstrate profiling of receptor expression distribution and potential increased multiplexing capability.</p> / Dissertation