Statistical analysis of the associations of hereditary factors with the risk of Type 1 diabetes and Diabetic Nephropathy based on the familial data collected through ascertaiment from population-based registers

In genetic epidemiology, population-based disease registries are commonly used to collect genotype or other risk factor information concerning affected subjects and their relatives. This work presents two new approaches for the statistical inference of ascertained data: a conditional and full likelihood approaches for the disease with variable age at onset phenotype using familial data obtained from population-based registry of incident cases. The aim is to obtain statistically reliable estimates of the general population parameters. The statistical analysis of familial data with variable age at onset becomes more complicated when some of the study subjects are non-susceptible, that is to say these subjects never get the disease. A statistical model for a variable age at onset with long-term survivors is proposed for studies of familial aggregation, using latent variable approach, as well as for prospective studies of genetic association studies with candidate genes. In addition, we explore the possibility of a genetic explanation of the observed increase in the incidence of Type 1 diabetes (T1D) in Finland in recent decades and the hypothesis of non-Mendelian transmission of T1D associated genes. Both classical and Bayesian statistical inference were used in the modelling and estimation. Despite the fact that this work contains five studies with different statistical models, they all concern data obtained from nationwide registries of T1D and genetics of T1D. In the analyses of T1D data, non-Mendelian transmission of T1D susceptibility alleles was not observed. In addition, non-Mendelian transmission of T1D susceptibility genes did not make a plausible explanation for the increase in T1D incidence in Finland. Instead, the Human Leucocyte Antigen associations with T1D were confirmed in the population-based analysis, which combines T1D registry information, reference sample of healthy subjects and birth cohort information of the Finnish population. Finally, a substantial familial variation in the susceptibility of T1D nephropathy was observed. The presented studies show the benefits of sophisticated statistical modelling to explore risk factors for complex diseases.

Molecular Background of Common Dyslipidemias

The leading cause of death in the Western world continues to be coronary heart disease (CHD). At the root of the disease process is dyslipidemia an aberration in the relevant amounts of circulating blood lipids. Cholesterol builds up in the arterial wall and following rupture of these plaques, myocardial infarction or stroke can occur. Heart disease runs in families and a number of hereditary forms are known. The leading cause of adult dyslipidemia presently however is overweight and obesity. This thesis work presents an investigation of the molecular genetics of common, hereditary dyslipidemia and the tightly related condition of obesity. Familial combined hyperlipidemia (FCHL) is the most common hereditary dyslipidemia in man with an estimated population prevalence of 1-6%. This complex disease is characterized by elevated levels of serum total cholesterol, triglycerides or both and is observed in about 20% of individuals with premature CHD. Our group identified the disease to be associated with genetic variation in the USF1 transcription factor gene. USF1 has a key role in regulating other genes that control lipid and glucose metabolism as well as the inflammatory response all central processes in the progression of atherosclerosis and CHD. The first two works of this thesis aimed at understanding how these USF1 variants result in increased disease risk. Among the many, non-coding single-nucleotide polymorphisms (SNPs) that associated with the disease, one was found to have a functional effect. The risk-enhancing allele of this SNP seems to eradicate the ability of the important hormone insulin to induce the expression of USF1 in peripheral tissues. The resultant changes in the expression of numerous USF1 target genes over time probably enhance and accelerate the atherogenic processes. Dyslipidemias often represent an outcome of obesity and in the final work of this thesis we wanted to address the metabolic pathways related to acquired obesity. It is recognized that active processes in adipose tissue play an important role in the development of dyslipidemia, insulin resistance and other pathological conditions associated with obesity. To minimize the confounding effects of genetic differences present in most human studies, we investigated a rare collection of identical twins that differed significantly in the amount of body fat. In the obese, but otherwise healthy young adults, several notable changes were observed. In addition to chronic inflammation, the adipose tissue of the obese co-twins was characterized by a marked (47%) decrease in amount of mitochondrial DNA (mtDNA) a change associated with mitochondrial dysfunction. The catabolism of branched chain amino acids (BCAAs) was identified as the most down-regulated process in the obese co-twins. A concordant increase in the serum level of these insulin secretagogues was identified. This hyperaminoacidemia may provide the feed-back signal from insulin resistant adipose tissue to the pancreas to ensure an appropriately augmented secretory response. The down regulation of BCAA catabolism correlated closely with liver fat accumulation and insulin. The single most up-regulated gene (5.9 fold) in the obese co-twins was osteopontin (SPP1) a cytokine involved in macrophage recruitment to adipose tissue. SPP1 is here implicated as an important player in the development of insulin resistance. These studies of exceptional study samples provide better understanding of the underlying pathology in common dyslipidemias and other obesity associated diseases important for future improvement of intervention strategies and treatments to combat atherosclerosis and coronary heart disease.

Identification of the Meckel syndrome gene (MKS1) exposes a novel ciliopathy

Meckel syndrome (MKS, MIM 249000) is an autosomal recessive developmental disorder causing death in utero or shortly after birth. The hallmarks of the disease are cystic kidney dysplasia and fibrotic changes of the liver, occipital encephalocele with or without hydrocephalus and polydactyly. Other anomalies frequently seen in the patients are incomplete development of the male genitalia, club feet and cleft lip or palate. The clinical picture has been well characterized in the literature while the molecular pathology underlying the disease has remained unclear until now. In this study we identified the first MKS gene by utilizing the disease haplotypes in Finnish MKS families linked to the MKS1 locus on chromosome 17q23 (MKS1) locus. Subsequently, the genetic heterogeneity of MKS was established in the Finnish families. Mutations in at least four different genes can cause MKS. These genes have been mapped to the chromosomes 17q23 (MKS1), 11q13 (MKS2), 8q22 (MKS3) and 9q33 (MKS4). Two of these genes have been identified so far: The MKS1 gene (this work) and the MKS3 gene. The identified MKS1 gene was initially a novel human gene which is conserved among species. We found three different MKS mutations, one of them being the Finnish founder mutation. The information available from MKS1 orthologs in other species convinced us that the MKS1 gene is required for normal ciliogenesis. Defects of the cilial system in other human diseases and model organisms actually cause phenotypic features similar to those seen in MKS patients. The MKS3 (TMEM67) gene encodes a transmembrane protein and the gene maps to the syntenic Wpk locus in the rat, which is a model with polycystic kidney disease, agenesis of the corpus callosum and hydrocephalus. The available information from these two genes suggest that MKS1 would encode a structural component of the centriole required for normal ciliary functions, and MKS3 would be a transmembrane component most likely required for normal ciliary sensory signaling. The MKS4 locus was localized to chromosme 9q32-33 in this study by using an inbred Finnish family with two affected and two healthy children. This fourth locus contains TRIM32 gene, which is associated to another well characterized human ciliopathy, Bardet Biedl syndrome (BBS). Future studies should identify the MKS4 gene on chromosome 9q and confirm if there are more than two genes causing MKS Finnish families. The research on critical signaling pathways in organogenesis have shown that both Wnt and Hedgehog pathways are dependent on functional cilia. The MKS gene products will serve as excellent model molecules for more detailed studies of the functional role of cilia in organogenesis in more detail.

Examination of Genetic Components Affecting Human Obesity-Related Quantitative Traits

Obesity increases the risk for several conditions, including type 2 diabetes mellitus, cardiovascular disease, hypertension, osteoarthirits and certain types of cancer. Twin- and family studies have shown that there is a major genetic component in the determination of body mass. In recent years several technological and scientific advance have been made in obesity research. For instance, novel replicated loci have been revealed by a number of genome wide association studies. This thesis aimed to investigate the association of genetic factors and obesity-related quantitative traits. The first study investigated the role of the lactase gene in anthropometric traits. We genetically defined lactose persistence by genotyping 31 720 individuals of European descent. We found that lactase persistence was significantly correlated with weight and body mass index but not with height. In the second study we performed the largest whole genome linkage scan for body mass index to date. The sample consisted of 4401 twin families and 10 535 individuals from six European countries. We found supporting evidence for two loci (3q29 and 7q36). We observed that the heritability estimate increased substantially when additional family members were removed from the analyses, which suggests reduced environmental variance in the twin sample. In the third study we assessed metabonomic, transcriptomic and genomic variation in a Finnish population cohort of 518 individuals. We formed gene expression networks to portray pathways and showed that a set of highly correlated genes of an inflammatory pathway associated with 80 serum metabolites (of 134 quantified measures). Strong association was found, for example, with several lipoprotein subclasses. We inferred causality by using genetic variation as anchors. The expression of the network genes was found to be dependent on the circulatory metabolite concentrations.

Genetic Basis of Pituitary Adenoma Predisposition

Much of the global cancer research is focused on the most prevalent tumors; yet, less common tumor types warrant investigation, since A rare disorder is not necessarily an unimportant one . The present work discusses a rare tumor type, the benign adenomas of the pituitary gland, and presents the advances which, during the course of this thesis work, contributed to the elucidation of a fraction of their genetic background. Pituitary adenomas are benign neoplasms of the anterior pituitary lobe, accounting for approximately 15% of all intracranial tumors. Pituitary adenoma cells hypersecrete the hormones normally produced by the anterior pituitary tissue, such as growth hormone (GH) and prolactin (PRL). Despite their non-metastasizing nature, these adenomas can cause significant morbidity and have to be adequately treated; otherwise, they can compromise the patient s quality of life, due to conditions provoked by hormonal hypersecretion, such as acromegaly in the case of GH-secreting adenomas, or due to compressive effects to surrounding tissues. The vast majority of pituitary adenomas arise sporadically, whereas a small subset occur as component of familial endocrine-related tumor syndromes, such as Multiple Endocrine Neoplasia type 1 (MEN1) and Carney complex (CNC). MEN1 is caused by germline mutations in the MEN1 tumor suppressor gene (11q13), whereas the majority of CNC cases carry germline mutations in the PRKAR1A gene (17q24). Pituitary adenomas are also encountered in familial settings outside the context of MEN1 and CNC, but unlike in the latter syndromes, their genetic background until recently remained elusive. Evidence in previous literature supported the notion that a tumor suppressor gene on 11q13, residing very close to but still distinct from MEN1, causes genetic susceptibility to pituitary tumors. The aim of the study was to identify the genetic cause of a low penetrance form of Pituitary Adenoma Predisposition (PAP) in families from Northern Finland. The present work describes the methodological approach that led to the identification of aryl hydrocarbon receptor interacting protein (AIP) as the gene causing PAP. Combining chip-based technologies (SNP and gene expression arrays) with traditional gene mapping methods and genealogy data, we showed that germline AIP mutations cause PAP in familial and sporadic settings. PAP patients were diagnosed with mostly adenomas of the GH/PRL-secreting cell lineage. In Finland, two AIP mutations accounted for 16% of all patients diagnosed with GH-secreting adenomas, and for 40% of patients being younger than 35 years of age at diagnosis. AIP is suggested to act as a tumor suppressor gene, a notion supported by the nature of the identified mutations (most are truncating) and the biallelic inactivation of AIP in the tumors studied. AIP has been best characterized as a cytoplasmic interaction partner of aryl hydrocarbon receptor (AHR), also known as dioxin receptor, but it has other partners as well. The mechanisms that underlie AIP-mediated pituitary tumorigenesis are to date largely unknown and warrant further investigation. Because AIP was identified in the genetically homogeneous Finnish population, it was relevant to examine its contribution to PAP in other, more heterogeneous, populations. Analysis of pituitary adenoma patient series of various ethnic origins and differing clinical settings revealed germline AIP mutations in all cohorts studied, albeit with low frequencies (range 0.8-7.4%). Overall, PAP patients were typically diagnosed at a young age (range 8-41 years), mainly with GH-secreting adenomas, without strong family history of endocrine disease. Because many PAP patients did not display family history of pituitary adenomas, detection of the condition appeared challenging. AIP immunohistochemistry was tested as a molecular pre-screening tool on mutation-positive versus mutation-negative tumors, and proved to be a potentially useful predictor of PAP. Mutation screening of a large cohort of colorectal, breast, and prostate tumors did not reveal somatic AIP mutations. These tumors, apart from being the most prevalent among men and women worldwide, have been associated with acromegaly, particularly colorectal neoplasia. In this material, AIP did not appear to contribute to the pathogenesis of these common tumor types and other genes seem likely to play a role in such tumorigenesis. Finally, the contribution of AIP in pediatric onset pituitary adenomas was examined in a unique population-based cohort of sporadic pituitary adenoma patients from Italy. Germline AIP mutations may account for a subset of pediatric onset GH-secreting adenomas (in this study one of seven GH-secreting adenoma cases or 14.3%), and appear to be enriched among young (≤25 years old) patients. In summary, this work reveals a novel tumor susceptibility gene, namely AIP, which causes genetic predisposition to pituitary adenomas, in particular GH-secreting adenomas. Moreover, it provides molecular tools for identification of individuals predisposed for PAP. Further elaborate studies addressing the functional role of AIP in normal and tumor cells will hopefully expand our knowledge on endocrine neoplasia and reveal novel cellular mechanisms of pituitary tumorigenesis, including potential drug targets.

Search for genetic variants influencing human height

Human growth and attained height are determined by a combination of genetic and environmental effects and in modern Western societies > 80% of the observed variation in height is determined by genetic factors. Height is a fundamental human trait that is associated with many socioeconomic and psychosocial factors and health measures, however little is known of the identity of the specific genes that influence height variation in the general population. This thesis work aimed to identify the genetic variants that influence height in the general population by genome-wide linkage analysis utilizing large family samples. The study focused on analysis of three separate sets of families consisting of: 1) 1,417 individuals from 277 Finnish families (FinnHeight), 2) 8,450 individuals from 3,817 families from Australia and Europe (EUHeight) and 3) 9,306 individuals from 3,302 families from the United States (USHeight). The most significant finding in this study was found in the Finnish family sample where we a locus in the chromosomal region 1p21 was linked to adult height. Several regions showed evidence for linkage in the Australian, European and US families with 8q21 and 15q25 being the most significant. The region on 1p21 was followed up with further studies and we were able to show that the collagen 11-alpha-1 gene (COL11A1) residing at this location was associated with adult height. This association was also confirmed in an independent Finnish population cohort (Health 2000) consisting of 6,542 individuals. From this population sample, we estimated that homozygous males and females for this gene variant were 1.1 and 0.6 cm taller than the respective controls. In this thesis work we identified a gene variant in the COL11A1 gene that influences human height, although this variant alone explains only 0.1% of height variation in the Finnish population. We also demonstrated in this study that special stratification strategies such as performing sex-limited analyses, focusing on dizygous twin pairs, analyzing ethnic groups within a population separately and utilizing homogenous populations such as the Finns can improve the statistical power of finding QTL significantly. Also, we concluded from the results of this study that even though genetic effects explain a great proportion of height variance, it is likely that there are tens or even hundreds of genes with small individual effects underlying the genetic architecture of height.

The DISC1 Pathway in the Genetic Etiology of Schizophrenia

Schizophrenia is a severe psychotic disorder affecting 0.5-1 % of the population. The disorder is characterized by hallucinations; delusions; disorganized behavior and speech; avolition; anhedonia; flattened affect and cognitive deficits. The etiology of the disorder is complex with evidence for multiple genes contributing to the onset of the disorder along with environmental factors. DISC1 is one of the most promising candidate genes for schizophrenia. It codes for a protein which takes part in numerous molecular interactions along several pathways. This network, termed as the DISC1 pathway, is evidently important for the development and maturation of the central nervous system from the embryo until young adulthood. Disruption at these pathways is thought to predispose schizophrenia. In the present study, we have studied the DISC1 pathway in the etiology of schizophrenia in the Finnish population. We have utilized large Finnish samples; the schizophrenia family sample where DISC1 was originally shown to associate with schizophrenia and the Northern Finland birth cohort 1966 (NFBC66). Several DISC1 binding partners displayed evidence for association in the family sample along with DISC1. Through a genome-wide linkage study, we found a significant linkage signal to a locus where a DISC1 binding partner NDE1 is located at the carriers of a certain DISC1 risk variant. In a follow-up study, genetic markers in NDE1 displayed significant evidence for association with schizophrenia. Further exploration of association between 11 genes of the DISC1 pathway and schizophrenia led to recognition of novel variants in NDEL1, PDE4B and PDE4D that significantly either increased or decreased the risk for schizophrenia. Further, we found evidence that DISC1 itself has a significant role in the human mental functioning even in the healthy population. Variants in DISC1 had a significant effect on anhedonia which is a trait present at everybody but is in its severe form one of the main symptoms of schizophrenia and correlates with the risk of developing the disorder. Further, utilizing genome-wide marker data, we recognized three genes; MIR620; CCDC141 and LCT; that are closely related to the DISC1 pathway but which effects on anhedonia were observable only at the individuals who carried these specific DISC1 variants. Our findings significantly add up to the previous evidence for the involvement of DISC1 and the DISC1 pathway in the etiology of schizophrenia and psychosis. Our results support the concept of a number of DISC1 pathway related genes contributing in the etiology of schizophrenia along with DISC1 and provide new candidates for the studies of schizophrenia. Our findings also significantly increase the importance of DISC1 itself as having a role in psychological functioning in the general population.

Novel Multiple Sclerosis Predisposing Genetic Variants Outside the HLA Region

Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system (CNS). Both environmental factors and several predisposing genes are required to generate MS. Despite intensive research these risk factors are still largely unknown, the pathogenesis of MS demyelination is poorly understood, and no curative treatment exists. Both prevalence and familial occurrence of MS are exceptionally high in a Finnish population subisolate, Southern Ostrobothnia, presumably due to enrichment of predisposing genetic variants within this region. Previous linkage scan on MS pedigrees from Southern Ostrobothnia detected three main MS loci on chromosomes 5p, 6p (HLA) and 17q. Linkage studies in other populations have also provided independent evidence for the location of MS susceptibility genes in these regions. Further, these loci are syntenic to the experimental autoimmune encephalomyelitis (EAE) susceptibility loci of rodents. In this thesis work an effort was made to localize MS predisposing alleles of the linked loci outside the HLA region by studying familial MS cases from the Southern Ostrobothnia isolate. Analysis of the 5p locus revealed one region, flanking the complement component 7 (C7) gene. The identified relatively rare haplotype seems to have a fairly large effect on genetic susceptibility of MS (frequency MS 12%, controls 4%; p=0.000003, OR=2.73). Evidence for association with alleles of the region and MS was seen also in more heterogeneous populations. Convincingly, plasma C7 protein levels and complement activity correlated with the risk haplotype identified. The finding stimulated us to study other complement cascade genes in MS. No evidence for association could be observed with the complement component coding genes outside 5p. A scan of the 17q locus provided evidence for association with variants of the protein kinase C alpha (PRKCA) gene (p=0.0001). Modest evidence for association with PRKCA was observed also in Canadian MS families. Finally we used a candidate gene based approach to identify potential MS loci. Mutations of DAP12 and TREM2 cause a recessively inherited CNS white matter disease PLOSL. Interestingly, DAP12 and TREM2 are located in MS regions on 6p and 19q, and we tested them as potential candidate genes in the Finnish MS sample. No evidence for association with MS was observed. This thesis provides an example of how extended families from special populations can be utilized in fine-mapping of the linked loci. A first relatively rare MS variant was identified utilizing the strength of a Finnish population subisolate. This variant seems to have an effect on activity of the complement system, which has previously been suggested to have an important role in the pathogenesis of MS.

Molecular Genetic Background of Tumours in Hereditary Leiomyomatosis and Renal Cell Cancer Syndrome

Hereditary Leiomyomatosis and Renal Cell Cancer (HLRCC) is a hereditary tumour predisposition syndrome. Its phenotype includes benign cutaneous and uterine leiomyomas (CLM, ULM) with high penetrance and rarer renal cell cancer (RCC), most commonly of papillary type 2 subtype. Over 130 HLRCC families have been identified world-wide but the RCC phenotype seems to concentrate in families from Finland and North America for unknown reasons. HLRCC is caused by heterozygous germline mutations in the fumarate hydratase (FH) gene. FH encodes the enzyme fumarase from mitochondrial citric acid cycle. Fumarase enzyme activity or type or site of the FH mutation are unassociated with disease phenotype. The strongest evidence for tumourigenesis mechanism in HLRCC supports a hypoxia inducible factor driven process called pseudohypoxia resulting from accumulation of the fumarase substrate fumarate. In this study, to assess the importance of gene- or exon-level deletions or amplifications of FH in patients with HLRCC-associated phenotypes, multiplex ligation-dependent probe amplification (MLPA) method was used. One novel FH mutation, deletion of exon 1, was found in a Swedish male patient with an evident HLRCC phenotype with CLM, RCC, and a family history of ULM and RCC. Six other patients with CLM and 12 patients with only RCC or uterine leiomyosarcoma (ULMS) remained FH mutation-negative. These results suggest that copy number aberrations of FH or its exons are an infrequent cause of HLRCC and that only co-occurrence of benign tumour types justifies FH-mutation screening in RCC or ULMS patients. Determination of the genomic profile of 11 HLRCC-associated RCCs from Finnish patients was performed by array comparative genomic hybridization. The most common copy number aberrations were gains of 2, 7, and 17 and losses of 13q12.3-q21.1, 14, 18, and X. When compared to aberrations of sporadic papillary RCCs, HLRCC-associated RCCs harboured a distinct DNA copy number profile and lacked many of the changes characterizing the sporadic RCCs. The findings suggest a divergent molecular pathway for tumourigenesis of papillary RCCs in HLRCC. In order to find a genetic modifier of RCC risk in HLRCC, genome-wide linkage and identical by descent (IBD) analysis studies were performed in Finnish HLRCC families with microsatellite marker mapping and SNP-array platforms. The linkage analysis identified only one locus of interest, the FH gene locus in 1q43, but no mutations were found in the genes of the region. IBD analysis yielded no convincing haplotypes shared by RCC patients. Although these results do not exclude the existence of a genetic modifier for RCC risk in HLRCC, they emphasize the role of FH mutations in the malignant tumourigenesis of HLRCC. To study the benign tumours in HLRCC, genome-wide DNA copy number and gene expression profiles of sporadic and HLRCC ULMs were defined with modern SNP- and gene-expression array platforms. The gene expression array suggests novel genes involved in FH-deficient ULM tumourigenesis and novel genes with putative roles in propagation of sporadic ULM. Both the gene expression and copy number profiles of HLRCC ULMs differed from those of sporadic ULMs indicating distinct molecular basis of the FH-deficient HLRCC tumours.