On the Origin of Pontocerebellar Hypoplasia: finding genes for a rare disease
This thesis describes novel genes involved in pontocerebellar hypoplasia (PCH). The identification of TOE1 (Chapter 2) and CLP1 (Chapter 4) as PCH causing genes is an important finding for PCH research and is directly implementable in PCH diagnostics. Furthermore, we elucidated genotype-phenotype correlations in these patients and in patients with mutations in EXOSC3 (Chapter 3). With this knowledge, relatives of patients suffering from PCH can be better informed about the expected disease progress. Here, I will summarise and discuss our findings, elaborate on the underlying disease mechanism of PCH and propose ideas for further research.
Chapter 1 provides an introduction of the clinical and genetic aspects of pontocerebellar hypoplasia. Ten subtypes of the disease have been described so far (PCH1-10). Common features in all subtypes include hypoplasia of the pons and cerebellum and severe mental and motor disabilities. Onset of the disease is premature, and life expectancy ranges from few days to late twenties. The spectrum of PCH is expanding, which is partly a matter of nomenclature. In some subtypes hypoplasia of the pons and cerebellum is the main hallmark, whereas in other subtypes it solely one aspect of the disease manifestation. In around half of PCH and PCH-like patients, a genetic mutation is identified. In 2008, the first gene causing PCH was discovered by linkage analysis . Nowadays, novel techniques as next generation sequencing allow us to identify more genes for rare diseases as PCH. The majority of genes involved in PCH play a role in RNA processing, which seems to be a common pathomechanism in the disease.
In Chapter 2 we present the TOE1 gene as the locus for PCH7. We started our search for a gene by performing exome sequencing in a Turkish family with two children with PCH plus genital abnormalities. Selecting for rare non-synonymous variants under an autosomal recessive model revealed three candidate genes: SLC39A1, CLK2 and TOE1. The latter two genes have a potential role in RNA processing, an apparently common disease pathway in PCH, which made us decide to follow up both genes. Injecting an antisense morpholino (MO) against the ATG start site region of clk2 gave a brain phenotype in zebrafish embryos. Injected zebrafish embryos showed a small head, an abnormally shaped midhindbrain boundary and increased cell death. However, we were unable to prove the specificity of the ATG MO induced phenotype, thus we cannot conclude that the abnormal brain in these fish is due to CLK2 knockdown. On a biochemical level, we showed that the p.A390S variation in CLK2 does not abolish kinase activity of the CLK2 protein in vitro despite it being a highly conserved residue. CLK2 knockout mice did not show brain or genital abnormalities. Last but not least, we identified CLK2 variations in only a single family with PCH7. Based on these results we concluded that CLK2 is not the gene causing PCH7.
In the meantime, we collected a cohort of ten patients (eight families) with PCH and genital abnormalities. We identified rare homozygous and compound heterozygous missense and nonsense mutations in the TOE1 gene in all affected individuals. Mutations in this gene seem to be restricted to PCH7, since no potentially pathogenic TOE1 variations were identified in PCH patients with other subtypes. Analysis of the protein structure of TOE1 revealed that the identified mutations possibly interrupt enzymatic activity and/or protein-protein interactions of TOE1.
All 46,XY patients in our PCH7 cohort have some degree of undervirilisation of the internal and external genitals. The two patients with karyotype 46,XX had normal female external genitalia and in one 46,XX patient no ovaries could be identified by medical imaging. Endocrinological investigations in six patients revealed hypergonadotrophic hypogonadism. Neurological manifestations included axial hypotonia with increased tone in the limbs and seizures. Development was severely delayed with no or very limited ability to sit, walk, talk and interact. Three patients died at the age of 24 weeks, 2 years and 3 years. The patients alive vary in age from one year to late twenties. Brain MRI of the patients with TOE1 mutations showed pontocerebellar hypoplasia, ventriculomegaly and decreased white matter including a thin corpus callosum.
With the identification of TOE1 as PCH7-causing gene, we confirm the presence of PCH plus genital abnormalities as a monogenetic syndrome.
The brain anomalies seen in PCH7 patients are not restricted to the pons and the cerebellum. One could argue that this is not within the spectrum of PCH, and also the prominent disorders of sex development make PCH7 an outsider in the spectrum. It is interesting to investigate whether overlap exists in disease mechanism between PCH7 and the other PCH subtypes. Alike the other genes causing PCH, TOE1 is involved in several RNA processing events: it has deadenylating activity  and is involved in Cajal body maintenance, which might implicate a role in RNA splicing . It was recently shown that TOE1 can penetrate cells and impair transcription and replication of viruses . Possibly, TOE1 has an effect on transcription of other genes as well.
Many questions remain to be answered about the disease mechanism of PCH7. TOE1s function in mRNA splicing and deadenylation are interesting to follow up, as well as its role in transcription regulation. Another clue that might lead to elucidating the mechanism is the finding that both 46,XY and 46,XX seem to have aberrations in genital development. This suggests that the onset of deviations in gonadal development is already before the formation of the bipotential gonad. This bipotential gonad is formed in the fifth week of development and gives rise to either testes or ovaries . Comparing TOE1 with genes as the homeobox gene EMX2, which is involved in formation of the bipotential gonad, mRNA transport and is associated with schizencephaly  could be very worthwhile.
In Chapter 3 describes a cohort of patients with EXOSC3-related PCH. It has been previously shown that EXOSC3 mutations account for around half of all PCH1 patients [7-9]. We show that mutations in this gene are restricted to this subtype and we present the disease course of twelve additional families with a genetic mutation in this gene. With this we show that the specific genetic mutation in EXOSC3 can largely predict the disease course. Patients with a homozygous p.D132A mutation in this gene display a prolonged disease course compared to other PCH patients, with possible survival into puberty. These patients can achieve some milestones as crawling and sitting, which are often lost when disease progresses. The pons can be unaffected in patients with this mutation. The p.G31A mutation is prevalent amongst the Roma population and lead to death in infancy. Contractures are reported in these patients and pontine hypoplasia is more pronounced than in the p.D132A group. Patients with a p.D132A and a nonsense or p.Y109N mutation on the other allele, or with a homozygous p.G135E mutation, present a severe phenotype as in the homozygous p.G31A group. Our study helps in predicting the disease course of PCH1 patients, and broadens the phenotype of PCH1, since EXOSC3 mutations can be considered in case of an unaffected pons.
Mutations in either EXOSC3 or EXOSC8 lead to PCH1 . Both genes are components of the exosome complex, which main function is to process and degrade various species of small RNAs . A signal for decreased exosome function is the accumulation of unprocessed ribosomal RNA . This was not seen in fibroblasts of patients with an EXOSC3 mutation . Perhaps, we should look for another, or a more subtle cellular effect in patients, or shift to studies in neurons rather than fibroblasts. In view the other genes involved in PCH, it is interesting to further investigate the role of the exosome complex in tRNA processing and the amino acid response.
Chapter 4 describes PCH10 caused by mutations in the CLP1 gene. Exome sequencing revealed nine families with a homozygous p.R140H mutation, all from the same region in Turkey [14,15]. Patients present pontocerebellar hypoplasia, a thin corpus callosum and both central and peripheral nervous system abnormalities. We mimicked the disease in zebrafish with a homozygous nonsense allele in Clp1. The fish die in early larval stage, and show brain atrophy and disorganised motor neurons. Injecting wild-type CLP1 mRNA rescues this phenotype, while injecting mRNA with the p.R140H mutation does not, demonstrating the deleterious effect of the mutation. CLP1 is physically and functionally connected to the TSEN complex, fulfilling a role in tRNA splicing. After tRNA intron cleavage by the TSEN complex, CLP1 phosphorylates the 5’ end of the 3’ tRNA exon. Hereafter, the two exon halves can be ligated to form a functional tRNA molecule. We show that the p.R140H mutation affects the function of CLP1 on various levels. Due to a disturbed interaction with the TSEN complex, tRNA cleavage is decreased in mutant cells. Furthermore, the kinase activity of CLP1 is reduced and the nuclear localisation of the CLP1 protein is impaired. Our results suggest that the mutation in CLP1 has a cell specific effect. In fibroblasts, no difference in levels of intron-containing pre-tRNA or mature tRNA was seen in patient versus control samples. In contrast, iNeurons show higher levels of pre-tRNA and lower levels of mature tRNA in patients cells versus control cells. This specific effect in neuronal cells is in line with the neurospecificity seen in PCH. Furthermore, tRNA fragments could play a role in CLP1-associated PCH. Firstly, accumulation of tRNA introns has been found , which possibly have a deleterious effect on the cell. Secondly, our experiments show that phosphorylation of the 3’ tRNA exon by CLP1 has a protective effect on stress-induced cell death. It remains elusive to which extent each of the aberrant functions - tRNA cleavage, nuclear localisation, kinase activity and tRNA fragments - account for the development of PCH10.
An interesting finding is that in PCH1 and PCH10, in contrast to other PCH subtypes, motor neurons are affected. A possible explanation is that the mechanism in PCH1 and PCH10 works two-fold: on one hand there is disturbed RNA processing potientially leading to reduced protein synthesis. On the other hand there are RNA fragments that become toxic for motor neurons in particular. In case of EXOSC3 and EXOSC8 these fragments could arise from defective clean-up by the exosome, in case of CLP1 they could arise from accumulating tRNA introns or tRNA exons. RNA foci composed of RNA fragments are often seen in motor neurons diseases as amyotrophic lateral sclerosis  and spinal cerebellar ataxia [17,18].
Chapter 5 discusses potential mechanisms underlying PCH. In the last few years, a start has been made in elucidating the pathomechanism of PCH. Most of the genes involved in PCH point towards a defect in (t)RNA processing and protein synthesis. Reduced protein synthesis can have various origins, for example from a lack of functional (t)RNAs. However, a global inhibition of translation can also arise via the intergrated stress response or due to effects of tRNA fragments. Future research might bring more clarity in the disease mechanism behind PCH and similar diseases.
As a results of the developments in next generation sequencing, enormous progress is made in finding new genes for rare diseases as PCH. This is very useful for diagnostics and prenatal counselling, but does not provide a treatment for a disease as PCH. The next challenging step is to elucidate the pathomechanism. As described in this thesis, the first steps are made in understanding the disease on cellular and molecular level, but many questions remain. Investigating whether the hypotheses about reduced protein synthesis and RNA metabolism are indeed the main bottleneck in PCH is the most important step in future research. When we discover more about what processes are affected in PCH, we can proceed with finding ways to intervene these processes. Nevertheless, finding a treatment for PCH will be challenging, considering the early - in most cases prenatal - onset of the disease.
This thesis describes several new genetic mutations causative for pontocerebellar hypoplasia and associated genotype-phenotype relations. It thereby provides directly implacable tools for improving diagnostics and genetic counselling of PCH. Furthermore, the studies in this thesis provide new clues for the disease mechanism of PCH and other neurodegenerative diseases.
1. Budde BS, Namavar Y, Barth PG, Poll-The BT, Nurnberg G, Becker C et al.: tRNA splicing endonuclease mutations cause pontocerebellar hypoplasia. Nat Genet 2008, 40: 1113-1118.
2. Wagner E, Clement SL, Lykke-Andersen J: An unconventional human Ccr4-Caf1 deadenylase complex in nuclear cajal bodies. Mol Cell Biol 2007, 27: 1686-1695.
3. Fong KW, Li Y, Wang W, Ma W, Li K, Qi RZ et al.: Whole-genome screening identifies proteins localized to distinct nuclear bodies. J Cell Biol 2013, 203: 149-164.
4. Sperandio S, Barat C, Cabrita MA, Gargaun A, Berezovski MV, Tremblay MJ et al.: TOE1 is an inhibitor of HIV-1 replication with cell-penetrating capability. Proc Natl Acad Sci U S A 2015, 112: E3392-E3401.
5. Eggers S, Ohnesorg T, Sinclair A: Genetic regulation of mammalian gonad development. Nat Rev Endocrinol 2014, 10: 673-683.
6. Brunelli S, Faiella A, Capra V, Nigro V, Simeone A, Cama A et al.: Germline mutations in the homeobox gene EMX2 in patients with severe schizencephaly. Nat Genet 1996, 12: 94-96.
7. Rudnik-Schoneborn S, Senderek J, Jen JC, Houge G, Seeman P, Puchmajerova A et al.: Pontocerebellar hypoplasia type 1: clinical spectrum and relevance of EXOSC3 mutations. Neurology 2013, 80: 438-446.
8. Biancheri R, Cassandrini D, Pinto F, Trovato R, Di RM, Mirabelli-Badenier M et al.: EXOSC3 mutations in isolated cerebellar hypoplasia and spinal anterior horn involvement. J Neurol 2013, 260: 1866-1870.
9. Schwabova J, Brozkova DS, Petrak B, Mojzisova M, Pavlickova K, Haberlova J et al.: Homozygous EXOSC3 mutation c.92G-->C, p.G31A is a founder mutation causing severe pontocerebellar hypoplasia type 1 among the Czech Roma. J Neurogenet 2013, 27: 163-169.
10. Boczonadi V, Muller JS, Pyle A, Munkley J, Dor T, Quartararo J et al.: EXOSC8 mutations alter mRNA metabolism and cause hypomyelination with spinal muscular atrophy and cerebellar hypoplasia. Nat Commun 2014, 5: 4287.
11. Liu Q, Greimann JC, Lima CD: Reconstitution, activities, and structure of the eukaryotic RNA exosome. Cell 2006, 127: 1223-1237.
12. Mitchell P, Petfalski E, Shevchenko A, Mann M, Tollervey D: The exosome: a conserved eukaryotic RNA processing complex containing multiple 3'-->5' exoribonucleases. Cell 1997, 91: 457-466.
13. Wan J, Yourshaw M, Mamsa H, Rudnik-Schoneborn S, Menezes MP, Hong JE et al.: Mutations in the RNA exosome component gene EXOSC3 cause pontocerebellar hypoplasia and spinal motor neuron degeneration. Nat Genet 2012, 44: 704-708.
14. Karaca E, Weitzer S, Pehlivan D, Shiraishi H, Gogakos T, Hanada T et al.: Human CLP1 mutations alter tRNA biogenesis, affecting both peripheral and central nervous system function. Cell 2014, 157: 636-650.
15. Schaffer AE, Eggens VR, Caglayan AO, Reuter MS, Scott E, Coufal NG et al.: CLP1 founder mutation links tRNA splicing and maturation to cerebellar development and neurodegeneration. Cell 2014, 157: 651-663.
16. Dejesus-Hernandez M, Mackenzie IR, Boeve BF, Boxer AL, Baker M, Rutherford NJ et al.: Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron 2011, 72: 245-256.
17. Chen IC, Lin HY, Lee GC, Kao SH, Chen CM, Wu YR et al.: Spinocerebellar ataxia type 8 larger triplet expansion alters histone modification and induces RNA foci. BMC Mol Biol 2009, 10: 9.
18. White MC, Gao R, Xu W, Mandal SM, Lim JG, Hazra TK et al.: Inactivation of hnRNP K by expanded intronic AUUCU repeat induces apoptosis via translocation of PKCdelta to mitochondria in spinocerebellar ataxia 10. PLoS Genet 2010, 6: e1000984.