Functional genomics studies of PINK1

Abstract

Functional genomics has become an important and established research discipline during the last 10 years, mainly as a consequence of the completion of large-scale genome sequencing projects. The human genome is now predicted to transcribe 20,000-25,000 protein coding genes, which is only a quarter of the number suggested a few years ago. Instead, a dynamic RNA universe seems to provide diversity to mammalian cells. Around 60-70% of protein coding genes are predicted to generate two or more transcripts by alternative splicing, while non-protein coding RNA are also directly transcribed from the genome. One category of such non-protein coding RNA is cistranscribed natural antisense (NAT). Cis-NATs are transcribed from a gene s antisense DNA strand and are suggested to have regulatory functions through direct interaction. The work described in this thesis includes functional genomics studies of the PINK1 locus.

We undertook a study to discover novel candidate genes associated with physical inactivity, a known risk factor for type 2 diabetes. Gene expression profiling of skeletal muscle from subjects before and after 5 weeks of inactivity by quantitative real-time PCR demonstrated a co-ordinated reduction in mitochondrial gene expression. We thus established a human in vivo model for mitochondrial dysfunction. Microarray analyses of the same sample-set suggested that PINK1, a novel mitochondrial kinase, was down regulated during inactivity. PINK1 is transcribed from a complex locus, alternatively spliced and with an annotated cis-NAT. Mutations at this locus had also been linked to Parkinson s disease and we thus selected this locus for subsequent functional genomics studies. We utilized human in vivo models, gene expression and genomic association analysis and RNA interference (RNAi) to study the regulation of PINK1.

We demonstrated dynamic expression from the PINK1 locus during modulation of mitochondria in vivo in human skeletal muscle. PINK1 was down regulated in our mitochondrial dysfunction model, while a shorter splice variant of PINK1 (svPINK1) and the NAT (naPINK1) were concordantly up regulated. The opposite expression pattern was obtained in a human in vivo model for increased mitochondrial activity, suggesting a direct association between svPINK1 and naPINK1. Knockdown of naPINK1 utilizing siRNAs targeting two different sites of naPINK1 reduced the level of svPINK1. This directly supports a role for naPINK1 in promoting the abundance of svPINK1, a novel mechanism for regulation by natural antisense.

In contrast, all transcripts from the PINK1 locus were less abundant in muscle tissue from diabetics, compared to healthy controls. To investigate whether PINK1 transcript levels could affect metabolic fitness or if the lower expression rather was a secondary effect of diabetes, we measured PINK1 tagging single nucleotide polymorphisms (SNPs). Several SNPs associated with PINK1 transcripts levels. The genotypes associating with higher expression of PINK1 also associated with lower plasma levels of non-esterified fatty acid levels (NEFA) and glucose. Two sets of RNA interference studies provided support for these clinical associations. Firstly, knockdown of PINK1 in human neuroblastoma cells resulted in impaired basal glucose uptake. Secondly, FABP4, a lipid transport protein, was selectively down regulated following PINK1 knockdown in adipocytes. However, mitochondrial genes were not altered when PINK1 expression was ablated, despite the in vivo association between such genes.

Taken together, our data suggest a role of PINK1 in lipid and glucose metabolism while PINK1 does not appear to be essential for mitochondrial biogenesis in mammalian cells.