Organization
The history of the University Medical Centre Groningen (UMCG) starts in 1797 with the opening of the Nosocomium Academicum. Now 228 years later, several things are changed, the size of the hospital has been increased and patients are treated in the hospital whether they are used for research or not. To date the UMCG is busy expanding and improving the way of work the UMCG. The current concept of the UMCG of the current decennia is ‘Healthy Aging’. This concept is used as a red thread in the research, care and education. The UMCG works closely together with the RUG when it comes to research and education. The UMCG has a couple of sectors within the company, these sectors can be seen on figure 1.1 (the image is in Dutch).
The department neuroscience is divided into sector F, this sector contains different departments that are responsible for the development of scientific knowledge and transmission of this knowledge to students and professionals. The goal of the department is using and improving the usage of the devices and techniques. The department neuroscience is subdivided in different four research groups. The group neuroimmunology performs research on neural -glial signaling of neuroinflammation [51][53][54][41].
Figure 0.1: Organogram of the UMCG [33].
1. Introduction
Alzheimer’s Disease (AD) is discovered in 1906 by dr. Alois Alzheimer due to an obsession with one of his patients. He met Auguste Deter when she was 51 years of age in 1901, the patient records and her brain where send over to the Kreapelin’s lab when she died in 1906. Alzheimer identified several pathological conditions; a shrunken cortex, presence of neurotic plaques (amyloid plaques) and neurofibrillary tangles (formed by collapsing tau proteins) during the research on Auguste’s brain [2][4][9][21][34].
To date AD is known as the most common form of dementia. AD is a neurodegenerative disease; this means that cell death of the brain cells occur within a certain time course. The characteristic of AD are the build up plaques among the cells in the brain and tangles in the dead/dying cells, this results to the loss of connections among nerve cells [3][5][31]. Amyloid Beta (Aß) pieces clump together and form plaque, small clumps that block the synapses. This can result into activating the inflammatory response that will devour the disabled cells. The normal function of tau is to make sure that the transportation goes flawless in healthy tissue, but when tangles are formed, tau collapses and nutrients cannot travel across the cell anymore [3].
AD can be categorized into two age groups; early onset Alzheimer’s disease (EOAD) and late onset Alzheimer’s disease (LOAD). EOAD is the category for people with an age of 30-65 years. This group contain about 5% of all of the Alzheimer’s patients and it contains a strong inheritance. Three genes are identified in the EOAD; Presenilin 1 (PSEN1), Presenilin 2 (PSEN2) and Amyloid Precursor protein (APP). The PSEN genes are transmembrane proteins and can be found in the gamma secretes protease that is used to cleave APP [7][20][30]. LOAD categorizes people with an age older than 65 years, this type is sporadic. APOE is one of the main genes, this gene is involved in transporting cholesterol to the neurons [30][32].
The APP23 transgenic mice are a commonly used model in AD research and also used in this project. These mice contain a 7-fold over-expression of the human APP gene with a double Swedish mutation. APP is a transmembrane protein and contains a high level expression in the brain [10][39]. The analysis of the transcriptome of APP23 mice brain involve the determination of the genes, pathways and the gene ontology classes that are affected and to correlate this data to obtained data. The age of the mice are 6-8 weeks, 6 months, 18 months and 24 months. The complete brains of these mice are Ribonucleic Acid (RNA) sequenced using the Illumina Hiseq2500 platform at Genetics facility of the UMCG.
APP is cleaved by alpha-secretase, beta-secretase or gamma-secretase. These secretase pathways can be divided into two other pathways, the non-amyloidogenic pathway and the amyloidogenic pathway. The non-amyloidogenic pathway uses the alpha-secretase and the gamma-secretase and is used 90% of the time when APP is cleaved. The amyloidogenic pathway uses the beta-secretase and the gamma-secretase and accounts for the other 10% when the APP is cleaved. These numbers can differ due to age, environmental factors or mutations[7][22][25][42][47].
When the APP is cleaved in the non-amyloidogenic pathway for the first time with alpha-secretase, sAPP-alpha is formed outside the membrane of the cell and CTF-alpha remains left, bound to the cell membrane. This step makes sure that the Aß region of APP is cleaved, as a result that the Aß isn’t formed. sAPP-beta and CTF-beta are formed when the APP is cleaved in the amyloidogenic pathway with beta-secretase. CTF-beta will be cleaved with gamma-secretase and creates Aß and AICD [22][25][47]. The length of the Aß peptide van vary among the 38 and 43 residues, the most common Aß is Aß40 and the least common Aß is the Aß42 (10% of the created Aß residues). The Aß42 variant is more likely to form Aß plaques than the other residues. But not all the Aß are bad, only the insoluble amyloid plaques disrupt the neuronal functions. Mutations in the APP can result in production of more APP that can be cleave in the amyloidogenic pathway. Also mutations in the PSEN1 and PSEN2 can lead to in increase of the Aß42 pathway [7][25].
Searching differentially expressed genes (DEG) was done with the use of several R packages; EdgeR, Limma, DEseq2, NOISeq and LFCSeq. These five packages can be subdivided into two different methods, the parametrical method and the non-parametrical method. The parametrical method needs normally distributed data and a large sample size, this is not the case for the non-parametrical method. The non-parametrical methods checks the data with the use of a ranking system where the most significant sample is ranked first. The DEG found in each package were used to check the pathway and the gene ontology. One of the above packages was ranked best, by comparing the results of the DEG found in each method.
The APP23 dataset can be compared to two other datasets (one human and one mice). These other two datasets needed to be preprocessed before they can use the R packages. The first step was to download the data, this was done with the use of the sra toolkit and the SRR numbers that are unique for the given sample of the datasets. These files needed to be written into a fastq files where they were used with the program FastQC to determine the quality of the dataset. The sequence needed to be trimmed when the quality was poor. The next step was to align the dataset with the use of STAR. The quality of this alignment was checked, with the use of RNA_SeQC. The last step was to write the bam outcome of RNA_SeQC to a text file, with the use of HTSeq, so that these files could be used the same way as the APP23 dataset in R.
To summarize the three main question that will be answered in this project:
- Which AD genes are found in the APP23 model.
- Which R package gives the best visualization for the RNA sequencing analysis. The package that is used most for the DEG analysis is EdgeR, the answer to this question will determine if this package will be used further or that this package will be replaced.
- How well is AD described by the APP23 mouse model.