MEANS AND METHODS FOR TREATING PARKINSON'S DISEASE

Abstract

The present invention relates to a method for screening a compound for Parkinson's disease treatment, comprising contacting a cell having a DJ-1 gene containing a c.192G>C mutation with a compound of interest, and testing whether said compound of interest prevents skipping of exon 3 of said DJ-1 gene, thereby identifying said compound as candidate for Parkinson's disease treatment. The present invention further relates to a compound which prevents skipping of exon 3 of the DJ-1 gene for use in a method of treatment of Parkinson's disease. The present invention further relates to a pharmaceutical composition comprising one or more compounds which prevents skipping of exon 3 of the DJ-1 gene for use in a method of treatment of Parkinson's disease.

Claims

1. An in vitro method for screening a compound for Parkinson's disease treatment, comprising (a) contacting a cell having a DJ-1 gene containing a c.192G>C mutation with a compound of interest; (b) testing whether said compound of interest prevents skipping of exon 3 of said DJ-1 gene, thereby identifying said compound as candidate for Parkinson's disease treatment.

2. The method of claim 1, wherein said cell has a c.192G>C mutation in at least one allele of the DJ-1 gene.

3. The method of any one of the preceding claims, wherein said cell has a c.192G>C mutation in both alleles of the DJ-1 gene.

4. The method of any one of the preceding claims, wherein said c.192G>C mutation leads to a p.E64D amino acid exchange in the DJ-1 protein.

5. The method of claim 1, wherein said cell is a mammalian cell, preferably a human cell line.

6. The method of claim 5, wherein said cell contains a DJ-1 gene having a c.192G>C mutation.

7. The method of any one of the preceding claims, wherein said cell is a primary fibroblast cell, an immortalized fibroblast cell, an induced pluripotent stem cell (iPSC), a small-molecule-derived neuronal precursor cell (smNPC), or a midbrain-specific dopaminergic neuro (mDA) from a patient having at least one c.192G>C mutation in a DJ-1 gene.

8. The method of any one of the preceding claims, wherein the DJ-1 gene containing a c.192G>C is fused to a reporter gene.

9. The method of any one of the preceding claims, wherein said testing in step (b) is accomplished by western blot, qPCR, mass spectrometry, detection of reporter gene activity, microscopy-based assay, or FACS-based assay

10. A compound which prevents skipping of exon 3 of the DJ-1 gene for use in a method of treatment of Parkinson's disease.

11. The compound for the use of claim 10, which is obtainable by the method of any one of claims 1 to 9.

12. The compound for the use of claim 10 or 11, which is a mutated U1 snRNA allowing the U1 complex to bind to the mutant pre-mRNA.

13. The compound for the use of claim 12, wherein said mutated U1 snRNA has the nucleotide sequence shown in SEQ ID NO: 7.

14. The compound for the use of claim 10 or 11, which is an optionally substituted C1-C10 alkyl carboxylic acid.

15. The compound for the use of claim 14, which is butyric acid.

16. The compound for the use of claim 14, which comprises at least one aryl and/or at least one heteroaryl substituent.

17. The compound for the use of claim 16, which is a compound having the structural formula (I): ##STR00003## wherein R1 is optionally substituted aryl or optionally substituted heteroaryl; and n is an integer of from 0 to 8.

18. The compound for the use of claim 17, which is 4-phenyl butyric acid.

19. The compound for the use of claim 10 or 11, which is a compound having the structural formula (II): ##STR00004## wherein R.sup.2 is hydrogen or optionally substituted alkyl; and X is hydrogen or halogen.

20. The compound for the use of claim 19, which is N6-furfuryladenine.

21. The compound for the use of claim 19, which is 2-chloro-N-[(furan-2-yl)methyl]-7H-purin-6-amine or 2-chloro-N-[(furan-2-yl)methyl]-N-methyl-7H-purin-6-amine.

22. A pharmaceutical composition comprising one or more compounds which prevents skipping of exon 3 of the DJ-1 gene for use in a method of treatment of Parkinson's disease.

23. The pharmaceutical composition for the use of claim 22, wherein the one or more compound comprises a compound as defined in any one of claims 10 to 21.

24. The pharmaceutical composition for the use of claim 22 or 23, wherein the composition comprises at least two compounds as defined in any one of the claims 10 to 21.

25. The pharmaceutical composition for the use of any one of claims 22 to 24, wherein the composition comprises a compound as defined in any one of claims 14 to 18 and a compound as defined in any one of claims 19 to 21.

26. The pharmaceutical composition for the use of any one of claims 22 to 25, wherein the composition comprises 4-phenyl butyric acid and 2-chloro-N-[(furan-2-yl)methyl]-7H-purin-6-amine.

27. The method, compound or pharmaceutical composition of any one of the preceding claims, wherein the Parkinson's disease is early onset Parkinson's disease, idiopathic Parkinson's Disease, Parkinson's disease associated with aberrant splicing, Parkinson's disease associated with a U1-dependent splicing defect, PARK7-associated Parkinson's disease, or Parkinson's disease associated with a c.G192C mutation in the PARK7 gene.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0078] FIG. 1: cDNA sequence of human wild type (wt) PARK7 (DJ-1) (SEQ ID NO: 1) and human c.192G>C PARK7 (DJ-1) mutant (SEQ ID NO: 2). Exon 2 is denoted in standard letters, exon 3 is denoted in italic letters, exon 4 is denoted in underlined letters, exon 5 is denoted in bold letters, exon 6 is denoted in italic and underlined letters, exon 7 is denoted in bold italic letters. Exon 3 in the human c.192G>C PARK7 (DJ-1) mutant (SEQ ID NO: 2) is only present if the mRNA has been correctly spliced, e.g. after pharmacological rescue.

[0079] FIG. 2: amino acid sequence of human DJ-1 wild type (wt) protein (SEQ ID NO: 5) and of human DJ-1 ED64D mutant (SEQ ID NO: 6).

[0080] FIG. 3: mutated U1 snRNA. The mutated nucleic acid (c->g substitution) is underlined. U1 termination sequence is shown in italic.

[0081] FIG. 4: A) Western blot and B) densitometry of Western blot of cell lysates prepared from human control fibroblasts and c.192G>C DJ-1 mutation carriers. Samples are separated in a 4% acrylamide SDS-PAGE gel. Result shows a variation of DJ-1 level in controls and heterozygous c.192G>C DJ-1 mutation carriers and results show the undetectability of DJ-1 protein in homozygous c.192G>C DJ-1 mutation carriers. N=1.

[0082] FIG. 5: A) Western blot and B) densitometry of Western blot of DJ-1 protein level in immortalised fibroblasts. Cell lysates prepared from immortalised fibroblasts of a healthy control and indicated c.192G>C DJ-1 mutation carriers were prepared and separated in a 4% acrylamide SDS-PAGE gel and detected with a DJ-1 specific antibody. Result shows an appoximatly 65% reduced level of DJ-1 protein in the heterozygous c.192G>C DJ-1 mutation carrier and DJ-1 protein is not detectable homozygous c.192G>C DJ-1 mutation carriers. Ratios were normalised to healthy control female A. N=1.

[0083] FIG. 6: DJ-1 amplification of cDNA of indicated patient derived fibroblasts of c.192G>C DJ-1 mutation carriers and healthy control. Fibroblasts were pelleted, RNA was extracted from the cell pellet and cDNA was transcribed. DJ-1 was amplified by PCR and subsequently run on an agarose gel. DNA gel analysis showed reduced mRNA length in both homozygous c.192G>C DJ-1 mutation carriers. Bands corresponding to both reduced and normal length mRNA were visible in the heterozygous c.192G>C DJ-1 mutation carrier.

[0084] FIG. 7: Schematic overview of Minigene assay. wt DJ-1 exon 3 or mutant DJ-1 exon 3 carrying the c.192G>C DJ-1 mutation respectively including flanking intronic sequences were cloned in a vector (pSPL3) (Invitrogen GmbH, Karslruhe, Germany) between two given exons (A and B). After transfection of the final plasmids splicing of wt or mutant exon 3 was analysed by rtPCR and subsequent sequencing of cDNA. It was analysed wether the c.192G>C DJ-1 mutation leads to physiological splicing or missplicing with skipping of exon 3.

[0085] FIG. 8: rtPCR result of minigene assay of wt and c.192G>C DJ-1 mutation carrying exon. Minigene transient transfection assay showing the effect of wt exon 3 and mutant DJ-1 exon 3 carrying the c.192G>C mutation. Two prominent products are seen on agarose gel, with the shorter band representing a PCR product lacking DJ-1 exon 3 (lane 3 and 4) and the longer band a product including DJ-1 exon 3 (lane 2 and 5). No template control (NTC). Size standard: 100 bp ladder (Invitrogen GmbH, Karslruhe, Germany).

[0086] FIG. 9: qPCR of DJ-1 RNA expression level with and without exon 3 in immortalised fibroblasts. Cell pellets were collected, RNA was isolated and transcribed into cDNA. cDNA was amplified by qPCR using one primerpair that amplifies DJ-1 cDNA with and without exon 3 (A) and one primerpair amplifying DJ-1 cDNA only in the presence of exon 3 (B). A) DJ-1 RNA expression is about 40% reduced in c.192G>C DJ-1 mutation carriers. B) Correctly spliced full length DJ-1 mRNA is around 30% reduced in heterozygous c.192G>C DJ-1 mutation carrier and about 80% in homozygous c.192G>C DJ-1 mutation carriers. Ratios were normalised to healthy control. Values show mean+SEM. ** p0.01, **** p0.0001 by one-way ANOVA followed by Tukey's multiple comparisons test. N=3.

[0087] FIG. 10: qPCR of DJ-1 RNA expression level with and without exon 3 in iPSCs. Cell pellets of indicated iPSC clones were collected, RNA was isolated and transcribed into cDNA. cDNA was amplified by qPCR using one primerpair that amplifies DJ-1 cDNA with and without exon 3 (A) and one primerpair amplifying DJ-1 cDNA only in the presence of exon 3 (B). A) DJ-1 RNA expression is reduced in two indicated homozygous c.192G>C DJ-1 mutation carrying clones from affected and premotor diseased individuals. B) Correctly spliced full length DJ-1 gene expression is reduced to 50% in one heterozygous c.192G>C DJ-1 mutation carrying clone in comparison to healthy control and strongly reduced in all homozygous c.192G>C DJ-1 mutation carriers. Ratios were normalised to healthy control 2. N=1.

[0088] FIG. 11: Scheme of splicing of c.192G>C DJ-1 mutation carrying pre-mRNA. Upper lane shows DJ-1 exon 3 carrying the c.192G>C mutation, which is the last base of exon 3 (red) and therefore lies in the splice donor site. During the splicing process of the pre-mRNA, exon 3 does not get recognised as exon as the c.192G>C DJ-1 mutation changes the splice donor site which, no longer gets recognised as such. Due to that, the splice acceptor does not get recognised either. Finally, this leads to skipping of exon 3.

[0089] FIG. 12: Binding of U1 snRNP to wt DJ-1 exon 3 and c.192G>C mutant DJ-1 exon 3. U1 snRNP binds to the pre-mRNA by base pairing to initiate the formation of the spliceosome. A) Binding of U1 snRNP to wt DJ-1 exon 3. Binding takes place in the presence of 4 mismatches. B) Not binding of U1 snRNP to c.192G>C mutant DJ-1 exon 3. The c.192G>C DJ-1 mutation leads to the fifth mismatch between U1 snRNA and the pre-mRNA therefore U1 snRNP does not bind to c.192G>C mutant DJ-1 exon 3.

[0090] FIG. 13: Binding of C>G mutant U1 snRNP to c.192G>C mutant DJ-1 exon 3. By mutagenesis, a C>G base exchange (indicated in cyan) was introduced in the U1 snRNA so that it matches the c.192G>C DJ-1 mutation and hence binds to c.192G>C mutant DJ-1 exon 3.

[0091] FIG. 14: rtPCR result of minigene assay of wt and c.192G>C DJ-1 mutation carrying exon cotransfected with wt U1 snRNA or c.192G>C U1 snRNA. Minigene transient transfection assay showing the effect of wt exon 3 and mutant DJ-1 exon 3 carrying the c.192G>C mutation cotransfected with wt U1 snRNA or c.192G>C U1 snRNA. Wt exon 3 in all three indicated and tested conditions leads to the inclusion of exon 3, resulting in a long band in the rtPCR (lane 2, 3 and 4). rtPCR results of c.192G>C mutation carrying exon 3 mingene without cotransfection and after cotransfection with wt U1 snRNA lead to a short band, representing skipping of exon 3 (lane 5 and 6). Two prominent products, the short and the long band, are seen on agarose gel, when c.192G>C exon 3 minigene was cotransfected with c.192G>C U1 snRNA (lane 7). No template control (NTC). Size standard: 100 bp ladder (Invitrogen GmbH, Karslruhe, Germany).

[0092] FIG. 15: Transient transfection of wt and c.192G>C U1 snRNA in immortalised fibroblasts of the index patient. 6 g of empty vector, wt U1 snRNA or c.192G>C U1 snRNA respectively were electroporated into immortalised fibroblasts of the index patient carrying the c.192G>C DJ-1 mutation using the Amaxa nucleofector I (Lonza Group Ltd, Basel, CH). 24 h after electroporation cell pellets were collected, RNA was isolated and transcribed into cDNA. cDNA was amplified by qPCR using one primer pair that amplifies DJ-1 cDNA with and without exon 3 (A) and one primer pair amplifying DJ-1 cDNA only in the presence of exon 3 (B). A) A significantly higher RNA expression was measured after electroporation of wt U1 snRNA. Expression was significantly higher in comparison to RNA expression after electroporation with empty vector and c.192G>C U1 snRNA. B) An increase in RNA expression of DJ-1 RNA including exon 3 was measured after electroporation with wt U1 snRNA in comparison to electroporation with empty vector. The increase was even higher after electroporation with c.192G>C U1 snRNA. Ratios were normalised to empty vector treatment. Values show mean+SEM. * p0.05, ** p0.01 by one-way ANOVA followed by Tukey's multiple comparisons test. N=3. C) DJ-1 was amplified by rtPCR from cDNA from homozygous c.192G>C DJ-1 mutation carrier after transient electroporation with empty vector control (lane 2), wt U1 snRNA (lane 3) or c.192G>C U1 snRNA (lane 4). Lane 1 shows marker and lane 5 NTC. Rescue of splicing of DJ-1 exon 3 indicated by longer band can only be seen after electroporation with c.192G>C U1 snRNA.

[0093] FIG. 16: Expression of recombinant DJ-1 in patient-derived smNPC. A. Transduction strategy with lentiviral vectors expressing either Ex3- or full length DJ-1. B. Enrichment of stably transduced cells. C. DJ-1 protein levels in transduced cells (n=3).

[0094] FIG. 17: Genetic rescue of DJ-1 by expression of mutant U1 snRNA. A. Transduction strategy with lentiviral vectors expressing either Ex3- or full length DJ-1 and enrichment of stably transduced cells. B. DJ-1 protein levels of stably transduced cells. C. Level of Dj-1 full length mRNA in transduced patients smNPC line M1 normalized to beta Actin.

[0095] FIG. 18: Genetic rescue of DJ-1 by expression of mutant U1 snRNA and full length DJ-1. A. DJ-1 protein levels in smNPC. B. DJ-1 protein levels neurons DIV 14. C. Level of full length DJ-1 mRNA in patient's smNPC line M1 normalized to beta Actin (left), DAG1 (middle) and HMBS (right).

[0096] FIG. 19: Chemical rescue of DJ-1 by expression in patient-derived smNPC. Cells were treated for 48 h with 4-Phenybutric acid or Sodium Butyrate at indicated concentrations and DJ-1 protein level was subsequently analyzed by western blot analysis. Dj-1 level is normalized to beta Actin and DJ-1 level of healthy control line K2 is set to 1.

[0097] FIG. 20: Chemical rescue of DJ-1 by expression in patient-derived smNPC. Cells were treated for 7 d with Rectas or its analogue 032.631.702 at indicated concentrations and DJ-1 protein level was subsequently analyzed by western blot analysis. Dj-1 level is normalized to beta Actin and DJ-1 level of healthy control line K2 is set to 1.

[0098] FIG. 21: The c.192G>C DJ-1 mutation leads to pronounced loss of DJ-1 protein in mutation carriers. (a) Pedigree structure of the family of c.192G>C mutation carriers. Black symbols denote homozygous carriers of the c.192G>C DJ-1 mutation. Symbols with a black dot denote heterozygous c.192G>C DJ-1 mutation carriers. White symbols denote healthy individuals and grey symbols containing a question mark denote untested individuals. Pedigree has been modified to prevent identification of individuals. (b-e) Immunoblots and densitometry of DJ-1 levels in fibroblasts (b), iPSCs (c), smNPCs (d) and mDA neurons (e) comparing controls (black bars), heterozygous (grey bars) and homozygous (white bars) c.192G>C mutation carriers. (f) Representative immunoblot of DJ-1 levels after blocking of proteasomal degradation in patient-derived smNPCs with indicated increasing concentrations of MG132 compared to untreated, DMSO treated and healthy control cells. -actin was used as loading control and detection of increasing polyubiquitination as treatment control. n=4 Western blots (g) Representative immunoblot of DJ-1 levels after blocking of degradation via autophagy-lysosomal pathway in patient-derived smNPCs with indicated increasing concentrations of bafilomycin A1 compared to untreated, DMSO treated and healthy control cells. -actin was used as loading control and detection of increasing LC3II levels as treatment control. n=4 Western blots, (H) Number of peptides with a sequence unique for DJ-1 (solid bars, right Y-axis) detected by mass spectrometry and signal intensity of this measurements calculated as label free quantification (LFQ) (striped bars, left Y-axis) in protein lysates of control and patient derived smNPCs.

[0099] FIG. 22: Loss of DJ-1 protein in homozygous c.192G>C mutation carrier. (a) Comparison of DJ-1 levels in fibroblasts by Western blot. (left) GAPDH and DJ-1 immunoblot of lysates from healthy control fibroblasts (left panels) and from fibroblasts of heterozygous and homozygous c.192G>C mutation carriers (right panels). (right) Quantification of the Western blot. (b) (top) Immunoblot of indicated smNPC lines comparing control lines with patient derived lines using three different anti-DJ-1 monoclonal antibodies (mAB) from Cell Signaling (mAB D29E5 XP, left panels), Abcam (mAB EP2816Y, middle panels) and Invitrogen (mAB E2.1, right panels). (bottom) Regions corresponding to the peptides that have been used to generate the three mAB are highlighted in the DJ-1 amino acid sequence with exon 3 highlighted in red letters.

[0100] FIG. 23: Characterization, transduction and differentiation of smNPC. (a) Transduction efficiency with lentiviral vectors encoding GFP and either U1 wt snRNA or G>C U1 snRNA was detected by flow cytometry after enrichment by FACS sorting. (b) Neuronal differentiation of smNPC. (left) Representative plots of flow cytometric analysis of DIV 30 neurons. Cells were stained with antibodies against CD200 and CD49f and gates were set to detect percentage of CD200+/CD49f-neurons. (right) Average percentage of neurons after in vitro differentiation from indicated smNPC lines. Bars show mean+s.e.m. n=6 (c) Transduction efficiency with lentiviral vectors encoding GFP and either wt or ex3 DJ-1 was detected by flow cytometry after enrichment by FACS sorting.

[0101] FIG. 24: Mechanism of loss of protein. (a) Quantification of DJ-1 immunoblots of indicated smNPC clones transduced with wt DJ-1 (dark grey bars) or ex3 DJ-1 (light grey bars) lentiviral constructs in comparison to untransduced (black bars) clones. DJ-1 levels were normalized to -actin. Values are normalized to control C2 and show mean+s.e.m. *** p0.001, **** p0.0001 by one-way ANOVA followed by Dunnett's multiple comparisons test. n5 (b) One-Step RT-qPCR results of the indicated smNPC clones. Levels of full-length DJ-1 mRNA (left graph) and ex3 DJ-1 mRNA (right graph) were detected by duplex One-step RT-qPCR using TaqMan probes for either DJ-1 variant and -actin mRNA as internal control. DJ-1 mRNA levels were normalized to expression levels of the -actin gene ACTB. Values show mean+s.e.m. **** p0.0001 by one-way ANOVA followed by Dunnett's multiple comparisons test (left graph) and Turkey's multiple comparison test (right graph). n3 (c) Quantification of DJ-1 immunoblots of neurons differentiated in vitro from indicated smNPC clones transduced with wt DJ-1 (dark grey bars) and ex3 DJ-1 (light grey bars) lentiviral constructs in comparison to untransduced clones (black bars). DJ-1 levels were normalized to -actin. Values are normalized to control C2 and show mean+s.e.m. ** p0.01 by one-way ANOVA followed by Dunnett's multiple com-parisons test. n3 (d-e) Secondary structure predictions for wt DJ-1 mRNA (d, SEQ ID NO: 12) and ex3 DJ-1 mRNA (e, SEQ ID NO: 13) were generated using the RNAfold software (f) (left) Northern blot of cytosolic (first panel) and nuclear (2nd panel) fraction of indicated smNPC lines. mRNA was visualized using a DJ-1 specific probe. (right) Western blot of the same cytosolic and nuclear fractions of smNPC C3 (3rd panel) and hom2-4 (4th panel) stained with anti-p84 and anti-Tubulin antibodies to confirm purity of the fractions. (g) Expression of recombinant DJ-1 in prokaryotic system by transformation of BL21 bacteria with wt DJ-1 and ex3 DJ-1. Representative figure of DJ-1 cDNA amplification from bacterial RNA without and upon induction of plasmid expression by Isopropyl -D-1-thiogalactopyranoside (IPTG) (upper panel) and immunoblot of DJ-1 protein from bacterial lysates without and upon IPTG induction (lower panel).

[0102] FIG. 25: Loading controls of FIG. 24. (a) Formaldehyde gel of RNA from cytosolic and nuclear fractions of indicated smNPC served as loading control for the Northern blot from FIG. 4e. RNA was visualized by ethidium bromide. (b) Ponceau red staining of the nitrocellulose membrane used in FIG. 4f was used to confirm equal loading of protein amounts of all samples.

[0103] FIG. 26: Genetic and pharmacologic rescue of loss of DJ-1-associated cellular phenotypes. (a) (left) Representative images of TMRM (red) and DAPI (blue) stained immortalized fibroblasts from control C-im (left image) and patient-derived line hom2-im (right image). (right) MMP represented as mean fluorescence intensity (MFI) of TMRE signal normalized to control C-im. Images from 36 individual cells of each line were analyzed by an investigator blinded to the experimental design. Values show mean+s.e.m. * p0.05, ** p0.01 by two-way ANOVA followed by Dunnett's multiple comparisons test. (b) MMP quantification as shown in (a) of indicated lines after transfection with empty or wt DJ-1 expression vector. Values show mean+s.e.m. n=3. **** p0.0001 by one-way ANOVA followed by Tukey's multiple comparisons test. (c) Levels of correctly spliced full-length DJ-1 mRNA in in vitro differentiated neurons from indicated clones after 4 d treatment with solvent controls or indicated concentrations of sodium phenylbutyrate and Rectas (grey bars) compared to untreated neurons (black bars) determined by duplex TaqMan RT-qPCR. Full-length DJ-1 mRNA levels were normalized to expression levels of the -actin gene ACTB. Values are normalized to control C2 and show mean+s.e.m. * p0.05 by one-way ANOVA followed by Tukey's multiple comparisons test. n5 (d) (top) Representative immunoblot of DJ-1 and -actin of in vitro derived neurons from indicated clones treated for 14 d with solvent controls or indicated concentrations of sodium phenylbutyrate and Rectas (grey bars) compared to untreated neurons (black bars). (bottom) Quantification of DJ-1 levels normalized to -actin. Values are normalized to control C2 and show mean+s.e.m. * p0.05 by one-way ANOVA followed by Tukey's multiple comparisons test. n3 (e) Densitometry of DJ-1 immunoblots of indicated immortalized fibroblast lines after 4 d of treatment with solvent controls or indicated concentrations of sodium phenylbutyrate and Rectas (grey bars) compared to untreated lines (black bars). DJ-1 levels were normalized to -actin. Values are normalized to control c-im and show mean+s.e.m. * p0.05 by one-way ANOVA followed by Dunnett's multiple comparisons test. n=5 (f) Flowcytometric analysis of MMP by TMRE staining in indicated immortalized fibroblast lines after 4 d treatment with solvent controls or indicated concentrations of sodium phenylbutyrate and Rectas (grey bars) compared to untreated lines (black bars). Values are normalized to control C-im and show mean+s.e.m. * p0.05, ** p0.01 by one-way ANOVA followed by Dunnett's multiple comparisons test. n=7.

[0104] FIG. 27: Level of relative full length DJ-1 protein in patient-derived neurons after 4 or 8 d of treatment with solvent controls or indicated concentrations of sodium phenylbutyrate or Rectas (grey bars). DJ-1 levels were normalized to -actin.

EXAMPLES

[0105] The following examples illustrate the invention. These examples should not be construed as to limit the scope of this invention. The examples are included for purposes of illustration and the present invention is limited only by the claims.

Example 1: Establishment of an iPSC Model from Patient Derived Fibroblasts

[0106] Reprogramming of both homozygous c.192G>C DJ-1 mutation carriers and the heterozygous c.192G>C DJ-1 mutation carrier was performed using retroviral vectors expressing the classical four Yamanaka factors. A minimum of 20 clones per individual was picked, passaged and stocks were frozen down for subsequent characterization of the clones. Characterization included tests about pluripotency and stem cell characteristics of the clones. Therefore, qPCR was performed to ensure silencing of the four exogenous viral factors and endogenous expression of pluripotency markers, ICC was used to show the expression of stem cell markers. Furthermore, the picked clones were tested for major chromosomal aberrations and mutations via G-banding and a genotyping microarray analysis. Additionally, the mutation carrying exon 3 of DJ-1 was sequenced in all clones to verify the correct genotype. Likewise, the ability of the clones to differentiate in cell types of all three germ layers was tested by ICC after differentiating the clones. At least two different clones per one individual were correctly characterized and a sufficient number of stocks was frozen down.

Example 2: DJ-1 Protein Level Varies Between Different Human Control Fibroblasts and c.192G>C DJ-1 Mutation Carriers do not Express DJ-1 Protein

[0107] Loss of DJ-1 function is a rare cause of familial PD. To calculate the steady state DJ-1 protein levels in different healthy individuals in comparison to the c.192G>C DJ-1 mutation carriers, samples from the respective individuals were analyzed on a low percentage SDS-PAGE gel (4% acrylamide). We took cell lysates from three different female human control individuals and from two different male human control individuals. Lysates from fibroblasts of family members of the c.192G>C DJ-1 family were taken as well. Lysates were taken from individual IV.9 who is a female heterozygous c.192G>C DJ-1 mutation carrier, from individual III.7, who is a male heterozygous c.192G>C DJ-1 mutation carrier and from individual III.11 and III.13 who are both homozygous c.192G>C DJ-1 mutation carriers; III.11 being female and premotor diseased and III.13 being the male index patient with motor symptoms.

[0108] DJ-1 proteins were examined by immunoblotting using a DJ-1 specific antibody (FIG. 4). Whereas DJ-1 protein levels vary between control fibroblasts and heterozygous c.192G>C DJ-1 mutation carriers, DJ-1 protein is not detectable in both homozygous c.192G>C DJ-1 mutation carriers.

[0109] Primary fibroblasts from patients are a useful tool but at the same time they only grow very slow and particularly from older people they become senescent after a certain times of passaging. To not lose the ability to work with fibroblasts from the c.192G>C DJ-1 family we immortalized fibroblasts using a lentiviral vector expressing SV40. This opened the opportunity to work with cells from patients that grow faster and can be used at higher passages.

[0110] To test if the immortalization had an effect on DJ-1 protein level and to see if the c.192G>C DJ-1 mutation still results in undetectable DJ-1 protein in immortal cells we repeated the Western blot for DJ-1 in this cell type which has a higher metabolic rate. Indeed DJ-1 protein could not be detected in both homozygous c.192G>C DJ-1 mutation carriers. The DJ-1 protein level of the heterozygous c.192G>C DJ-1 mutation carrier was reduced to about 20% of the amount seen in the healthy control (FIG. 5).

Example 3: The c.192G>C DJ-1 Mutation Causes Skipping of DJ-1 Exon 3

[0111] The findings on DJ-1 protein level in c.192G>C DJ-1 mutation carriers raised interest in the underlying effect of the mutation. In a first experiment, RNA from patient derived fibroblasts was isolated, transcribed into cDNA and amplified DJ-1 by PCR. Agarose gel revealed a reduced length of c.192G>C DJ-1 carrying cDNA and a normal size band as well as the short band in the heterozygous c.192G>C carrying cDNA. cDNA from a healthy control served as positive control for correct length (FIG. 6).

[0112] Subsequently, these results led us to sequence the amplified cDNA to find out which sequence segment was missing in the cDNA of c.192G>C DJ-1 mutation carriers. Sequencing result shows that the end of DJ-1 exon 2 is directly followed by the beginning of exon 4. Complete sequence of DJ-1 exon 3 was missing or skipped.

[0113] After obtaining these results, we wanted to validate the finding of skipping of DJ-1 exon 3 in c.192G>C DJ-1 mutation carrying cells. Therefore, minigene assay to assess the effect of DJ-1 exon 3 on splicing was performed.

[0114] Minigene assay is used to analyze splicing of an exon of interest by cloning it into a vector (pSPL3) between two given exons (Exon A and Exon B) and subsequent analysis of the cDNA by rtPCR and sequencing (FIG. 7). In brief, wt DJ-1 exon 3 or mutant DJ-1 exon 3 carrying the c.192G>C DJ-1 mutation respectively including flanking intronic sequences were cloned in a vector (pSPL3) (Invitrogen GmbH, Karslruhe, Germany) between two given exons (A and B). After transfection of the final plasmids splicing of wt or mutant exon 3 was analyzed by rtPCR and subsequent sequencing of cDNA. It was analyzed whether the c.192G>C DJ-1 mutation leads to physiological splicing or missplicing with skipping of exon 3.

[0115] In the minigene assay we tested wt exon 3 of the control and mutant exon 3 (carrying the c.192G>C DJ-1 mutation) from the index patient. At the same time minigene constructs were tested where the wt exon 3 was mutated to carry the c.192G>C mutation and the corrected exon from the index patient now carrying the wt allele. Agarose gel electrophoresis (FIG. 8) revealed that cultures transfected with pSPL3 constructs with the mutant c.192 C-allele yielded rtPCR products that were clearly smaller than products from cultures transfected with the wildtype c.192 G-allele, indicating abnormal splicing/skipping of the mutant minigene construct. Subsequent sequencing showed that rtPCR products derived from the transfection with the mutant c.192 C-allele lacked exon 3 of the DJ1 gene.

[0116] The c.192G>C DJ-1 mutation in exon 3 which does not lead to a premature stop codon and which does not cause a frameshift in the RNA caused the skipping of exon 3 in the minigene assay. In addition to the correction of the mutation in the patient exon 3 and the mutating of exon 3 to carry the c.192G>C DJ-1 mutation from a healthy control in the minigene, sequencing of the entire DJ-1 locus of the patient's DNA was performed. This was done in order to have a second confirmation that the c.192G>C DJ-1 mutation suffices to cause skipping of exon 3 in c.192G>C DJ-1 mutation carriers and other sequence variants can be excluded.

[0117] The minigene assay showed, in an artificial system, that the c.192G>C DJ-1 mutation which lies at the end of exon 3 leads to skipping of the respective exon. Skipping of this exon does not lead to a frameshift or a premature stop codon.

[0118] Next, we wanted to verify our minigene assay results in the different cell types available from the c.192G>C DJ-1 mutation carriers. Therefore, cells were grown under standard conditions, harvested, RNA was extracted and transcribed into cDNA. Quantitative qPCR was performed using cDNA and two different pairs of primers for amplification. One primer pair to detect correctly spliced DJ-1 as well as cDNA lacking exon 3 and another primer pair where one primer is complementary with exon 2 and exon 3, leading to only correctly spliced (with exon 3) DJ-1 cDNA detected by qPCR. qPCR was performed with cDNA from immortalized fibroblasts (FIG. 9), iPSCs, mDA neurons and smNPCs.

[0119] qPCR results generated with primers that detect normal and misspliced DJ-1 cDNA show a reduction which is not significant of DJ-1 RNA expression in all c.192G>C DJ-1 mutation carriers (FIG. 9A). Reduction is even higher when looking at normal spliced RNA where we see a significant reduction of DJ-1 RNA expression between the healthy control and the heterozygous c.192G>C DJ-1 mutation carrier. Even lower levels in both indicated homozygous c.192G>C DJ-1 mutation carriers were seen. RNA expression of DJ-1 is significantly reduced in both homozygous c.192G>C DJ-1 mutation carriers when compared to healthy control or heterozygous carrier of the same mutation (FIG. 9B).

[0120] When looking at the qPCR results in iPSCs derived from these fibroblasts one sees a reduction in RNA expression of DJ-1 with and without skipping of exon 3 of the c.192G>C DJ-1 mutation carriers. With two clones (homozygous premotor diseased c4 and homozygous affected c4) being particularly low (FIG. 10A). The results generated with primers that only detect DJ-1 cDNA when exon 3 is present all cDNAs from homozygous carriers are practically not detectable. The two indicated heterozygous clones vary with one showing the same RNA expression level as the healthy control and the other one showing a 50% decrease in correctly spliced DJ-1 RNA (FIG. 10B).

Example 4: Genetically Modified U1 Rescues Skipping of Exon 3

[0121] Further experiments aimed at analyzing why the c.192G>C base exchange leads to missplicing of DJ-1 pre-mRNA and skipping of exon 3 during the splicing process. Hypothesized mechanism of splicing of pre-mRNA carrying the c.192G>C DJ-1 mutation is shown (FIG. 11). The c.192G>C DJ-1 mutation lies at the end of exon 3 and hence in the splice donor recognition site. If the splice donor sequence does not get recognized as such due to the mutation, the splice acceptor site will also not be recognized. During the splicing process, U1 snRNP and U2 snRNP will recognize the splice donor sequence of exon 2 and the splice acceptor sequence of exon 4 respectively, surrounding exon 3, but not exon 3 itself. This will finally lead to skipping of DJ-1 exon 3 and a shortened DJ-1 mRNA (FIG. 11).

[0122] As the c.192G>C DJ-1 mutation lies in the splice donor site normally it would get recognized by the U1 snRNP which is initiating the formation of the spliceosome. U1 snRNP recognizes the 5 splice site through base pairing at the 5 end. Of note, a hundred percent match is not necessary. In case of DJ-1 exon 3 in the wt situation there are already four mismatches between U1 snRNA and the splice donor site (FIG. 12A). The presence of the c.192G>C DJ-1 mutation leads to five mismatches between U1 and the splice donor site (FIG. 11B). The aim of this part of the project was to test whether this additional mismatch causes U1 snRNP to not recognize the splice donor site hence leading to skipping of exon 3 (FIG. 11).

[0123] To test this hypothesis, one base of the U1 snRNA was modified in the sense that it matches the c.192G>C mutation in DJ-1 exon 3, namely C>G were exchanged as indicated in cyan (FIG. 13).

[0124] In a splicing reaction where C>G mutant U1 snRNP is present and recognizes the splice donor sequence of c.192G>C mutant DJ-1 exon 3 skipping of this exon does not take place.

[0125] In order to test the hypothesis, the previously performed minigene experiment was repeated and a plasmid expressing c.192G>C U1 snRNA was co-transfected.

[0126] The minigene construct expressing wt exon 3 resulted in the inclusion of the exon 3 when the minigene assay was performed with the minigene construct alone, when wt U1 snRNP was co-transfected and when c.192G>C U1 snRNA was co-transfected, this is shown by the long prominent bands on the agarose gel picture in (FIG. 14) lane 2-4. When the experiment was performed with a minigene construct with the c.192G>C mutation carrying DJ-1 exon 3 alone or co-transfected with wt U1 snRNA, a short prominent band can be seen on lane 5 and 6 of the agarose gel as rtPCR result. This means that in these cases mutant exon 3 was skipped. When this minigene construct was co-transfected together with c.192G>C U1 snRNA in the minigene assay, however, two products were amplified by rtPCR, displayed by the short band, indicating skipping of exon 3, and the long band indicating the presence of exon 3. This is shown in lane 7 on the agarose gel picture in (FIG. 14).

[0127] As c.192G>C U1 snRNA was able to rescue skipping of c.192G>C DJ-1 mutation carrying exon 3 in the minigene assay in HEK cells, the next step was to test the effect of c.192G>C U1 snRNA in patient-derived cells.

[0128] Immortalised fibroblasts of the index patient carrying the c.192G>C DJ-1 mutation were electroporated with empty vector, wt U1 snRNA or c.192G>C U1 snRNA, respectively. Electroporation with wt U1 snRNA caused a significant increase in DJ-1 RNA expression of normal RNA and RNA lacking DJ-1 exon 3 in comparison to electroporation with empty vector or c.192G>C U1 snRNA. c.192G>C U1 snRNA also caused higher RNA levels in comparison to empty vector (FIG. 15A). When only looking at correctly spliced DJ-1 RNA (including exon 3), more RNA expression was seen after electroporation with wt U1 snRNA and even more expression after electroporation with c.192G>C U1 snRNA, compared to electroporation with empty vector (FIG. 15B).

Example 5: Rescue of DJ-1 Protein in Patient Based Cellular Models

[0129] The c.192G>C mutation causes mis-splicing of the DJ-1 pre mRNA upon which Exon 3 is spliced out. Although the resulting shorter Ex3-mRNA is present in the patient derived small molecule-derived neuronal precursor cells (smNPC), the protein levels of DJ-1 are reduced to an almost undetectable level. The Ex3-mRNA is, although stably expressed in the cells, not translated into protein. Even overexpression of a recombinant Ex3-mRNA that doesn't undergo splicing does not translate into expression of a truncated Ex3-DJ-1 protein. However, overexpression of recombinant full length DJ-1 restores DJ-1 full length mRNA and protein levels in patient derived.

[0130] Both constructs (Ex3 and full length DJ-1) were delivered to cells by stable transduction using lentiviral vectors. The vectors co-express GFP and population of successfully transduced cells were enriched by FACS sorting for GFP+ cells to an purity of over 80%. (FIG. 16).

[0131] As the DJ-1 Ex3-mRNA does not get translated into protein we aimed to rescue the pathological skipping of exon 3 genetically in the patient-derived cell model. By stably transducing the cells with a specifically designed mutated U1 snRNA (U1mut) that will allow the U1 complex to bind again to the mutant pre-mRNA, the amount of correctly spliced DJ-1 RNA was predicted to increase in the cell model. A lentiviral vector encoding wild type U1 snRNA (U1wt) was used as control to account for effects on protein levels that occur due to general increase of U1 levels in the cells. These lentiviral vectors were co-expressing GFP for later enrichment of transduced cells.

[0132] The enriched population of U1 transduced cells indeed showed a rescue of DJ-1 protein only when transduced with the mutant U1 snRNA (FIG. 17) The overexpression of wild type U1 was not sufficient to restore DJ-1 protein levels.

[0133] Both rescue strategies, expression of DJ-1 full length and U1mut, increase level of full length RNA and protein levels not only in smNPC but also in DIV14 neurons differentiated from smNPC (FIG. 18).

[0134] The genetic rescue shows that the DJ-1 protein can be successfully restored in patient cells after identification of the underlying mechanism by correcting the splicing defect.

[0135] This opens the opportunity for a second strategy to rescue DJ-1 protein level, the use of small chemical compounds. The patient-derived smNPC were used to screen for active compounds. Here, our experiments identified 4-Phenylbutric acid as active in rescuing DJ-1 protein expression in concentrations of 150 M and 300 M as shown by Western blot (FIG. 19). Similar tendency of rescue effect was observed in preliminary experiments in which smNPC were treated for 7 d with Rectas or with a Rectas analogue called 032.631.702 (FIG. 20).

Example 6: Impaired Translation of c.192G>C Mutant mRNA Causing Loss of DJ-1 Protein

[0136] Since exon 3 skipping is in frame, a shorter DJ-1 peptide would be expected, that is identical to the wt protein except for the deleted 34 amino acids encoded by exon 3. Strikingly, no truncated protein can be detected (FIG. 21, FIG. 22a-b). To exclude the possibility that expression of DJ-1 protein in our patient-derived cell model is impaired by defects of the translation machinery that are unrelated to the mutation itself, we overexpressed wt DJ-1 by lentiviral transduction of patient-derived smNPC. Moreover, we transduced cells with a ex3 DJ-1 cDNA vector whose transcript requires no splicing and is expected to translate into a truncated protein (FIG. 23c). Overexpression of wt DJ-1 restored protein levels in clones hom2-1 and hom2-4 (FIG. 24a). However, the overexpression of the ex3 vector failed to increase DJ-1 protein levels (FIG. 24a). In line with this observation, the overexpression of wt DJ-1 resulted in a significant increase of full-length mRNA (FIG. 24b, left graph). Interestingly, although levels of ex3 mRNA were significantly increased in transduced clones compared to non-transduced clones (FIG. 24b, right graph), truncated protein was still not detected (FIG. 24a). Neurons derived from smNPC transduced with wt DJ-1 also revealed significantly increased DJ-1 levels. DJ-1 levels in neurons derived from ex3 DJ-1 transduced smNPC were not rescued (FIG. 24c).

[0137] Next, we investigated the predicted centroid secondary structure of mRNA, where 47.1% of the bases were predicted to be unpaired for wt DJ-1, while only 40.3% were unpaired for the ex3 variant. This difference in the number of unpaired bases in the native mRNA structure is also reflected by a slightly higher predicted minimum free energy (271.1 kcal/mol) as compared to the ex3 variant (273.1 kcal/mol). These energy predictions can only provide indicative estimates, but structure visualizations (FIGS. 24d and e) pointed to systematic differences between the variants across several bases in one specific local branch, between the bases at position 209 and 486 in the native structure and bases 214 and 381 in the mutant. In this secondary structure branch, the majority of bases in the native structure are unpaired, whereas the corresponding branch in the mutant is dominated by base pairs forming extended stem-loop structures. However, the predicted differences of mRNA structure do not result in inhibition of mRNA transport from the nucleus to the cytoplasm where translation takes place. Northern blot analyses of cytosolic and nuclear fractions of smNPC showed no difference in DJ-1 mRNA levels between control lines and patient-derived cells (FIG. 24f, FIG. 25a). Next we explored, whether impaired ex3 DJ-1 translation is only related to interference with the eukaryotic translation machinery. When expressed in a prokaryotic organism, ex3 DJ-1 cDNA also failed to be translated (FIG. 24g). E. coli transformed with either ex3 DJ-1 or wt DJ-1 cDNA constructs expressed DJ-1 mRNA (FIG. 24g, upper panel), however, only wt DJ-1 mRNA was translated into protein (FIG. 24g, lower panel, FIG. 25b). Together, our results show that the lack of DJ-1 protein in homozygous c.192G>C carriers is not caused by mRNA instability or mis-localization, but rather interference with the translation machinery.

Example 7: Combination of Compounds

[0138] We hypothesized a synergistic effect of RECTAS when combined with PB to rescue DJ-1 in c.192G>C carriers. Therefore, DIV 30 neurons differentiated from the two patient-derived smNPC clones were treated for 4 days with 1 mM PB and indicated concentrations of RECTAS. mRNA levels of full-length DJ-1 were measured by qPCR. Compared to the untreated and the solvent control treated neurons correctly spliced full-length DJ-1 mRNA was significantly increased after treatment with 1 mM PB and 50 M RECTAS. Four days treatment with 1 mM PB and RECTAS concentrations of 10 M and 25 M increased the amount of correctly spliced mRNA in a dose-dependent manner (FIG. 26c). Therefore, we chose the concentration of 25 M and the most effective concentration of 50 M for longer treatments of 14 days to rescue DJ-1 protein. DJ-1 protein levels increased up to 9.841% and 17.77% upon treatment with 1 mM PB and 25 M or 50 M RECTAS, respectively, while solvent control treatment showed no increase (FIG. 26d). This rescued DJ-1 protein in patient-derived cells carrying the p.E64D mutation revealed no toxic effect in previous studies. Therefore, we hypothesized that the rescue of E64D DJ-1 in patient-derived cells would also rescue cellular phenotypes. Indeed, after 4 days treatment of hom2-im fibroblasts with PB 1 mM+RECTAS 10 or 25 M we observed a dose dependent increase of DJ-1 protein levels that reached statistical significance at the higher concentration (FIG. 26e). This increase was of physiological relevance as it led to a significant increase of MMP. While untreated and solvent control treated hom2-im fibroblasts showed decreased MMP compared to healthy control C-im it was significantly increased in fiboblasts treated with 1 mM PB+10 or 25 M RECTAS (FIG. 26f).

[0139] We also tested the effect of RECTAS and PB as a single treatment for rescuing DJ-1 in c.192G>C carriers. The experiments were carried out with patient-derived neurons as described above. As shown in FIG. 27, Rectas alone moderately enhances the portion of correctly processed mRNA, while such an increase can only be seen for PB at a concentration of 1 mM after 8 days of treatment. However, since PB generally enhances expression of DJ-1 mRNA rather than effecting correct splicing, an increase of DJ-1 protein level following PB single treatment, as shown in FIG. 19 may be explained by the overall increase of both correctly and incorrectly processed mRNA.

Example 8: Treatment of Parkinson's Disease Associated with Aberrant Splicing

[0140] Here we describe a novel mechanistic concept for the pathogenic role of the PD-associated c.192G>C mutation in PARK7. In contrast to a missense mutation that was predicted to cause instability of E64D mutant DJ-1 protein, we identified a drastic reduction of DJ-1 protein levels in patient-based cellular models due to U1-dependent mis-splicing of pre-mRNA. The demonstration of the disease-causing ex3 DJ-1 mRNA splice variant as cause of the drastically reduced expression of DJ-1 protein reported here underscores the relevance of access to patient material for functional studies in order to validate predictions and define the underlying molecular pathology as basis for precision medicine approaches.

[0141] The observed discrepancy between the overall amount of mutant pre-mRNA and the drastically reduced levels of DJ-1 protein in homozygous c.192G>C carriers indicated the involvement of pathological mRNA processing and/or translation and argued against nonsense-mediated decay (NMD) as underlying mechanism. In our case, the c.192G>C mutation in the DJ-1 gene leads to an in-frame deletion of exon 3 without a PTC and therefore mutant mRNA was predicted to encode a smaller protein lacking the 34 amino acids encoded by exon 3. However, neither mass spectroscopy nor Western blot with different antibodies directed against different epitopes of DJ-1 revealed a truncated protein encoded by ex3 DJ-1 pre-mRNA (FIG. 21, 22 FIG. 21b-h, FIG. 22b). Therefore, we speculate that significant changes of the secondary structure related to ex3 DJ-1 mRNA confer increased stability due to extended stem-loop structures and thus interfere with ribosomal function (FIG. 24d). Indeed similar changes in mRNA secondary structure created by minor shifts of hair pins by 9 base pairs were shown to reduce the translation of the encoded protein more than 50-fold (Babendure, J. R. Control of mammalian translation by mRNA structure near caps. RNA 12, 851-861 (2006)). Moreover, certain stable RNA structures like extended stem loops can stall the ribosome and lead to translational arrest (Doma, M. K. & Parker, R. Endonucleolytic cleavage of eukaryotic mRNAs with stalls in translation elongation. Nature 440, 561-564 (2006)).

[0142] In order to prove a U1-mediated pathogenic splicing mechanism underlying loss of DJ-1 function in PD, we performed genetic rescue experiments. This may open avenues for future gene transfer strategies to the central nervous system, as options for viral vectors delivered to the brain are becoming safer and more effective.

[0143] Here, we successfully applied an advanced literature mining approach for prioritization of candidate drugs to revert molecular and cellular phenotypes associated with mis-splicing of DJ-1 in homozygous carriers of the c.192G>C mutation. The novel combination of a RECTAS with PB significantly increased mRNA and protein levels of DJ-1 in patient-based cells across different tissues (FIGS. 26d and 26e). Interestingly, no complete rescue of protein levels to control level was required for restoring mitochondrial function. This is in line with the high variability of physiological DJ-1 levels observed in cells from healthy individuals, that overlapped with levels of DJ-1 in asymptomatic heterozygous carriers of the c.192G>C mutation (FIG. 22a).

[0144] Our findings on the combination of RECTAS and PB as causative treatment for DJ-1 deficiency have an immediate impact for homozygous carriers of the c.192G>C mutation, and may qualify for treatment at the prodromal stage of PD for addressing neuroprotective strategies in PD. As a first drug candidate for correction of aberrant splicing, recently kinetin underwent clinical assessments for pharmacokinetics, safety and effectiveness in vivo (Axelrod, F. B. et al. Kinetin improves IKBKAP mRNA splicing in patients with familial dysautonomia. Pediatr. Res. 70, 480-3 (2011); Shetty, R. S. et al. Specific correction of a splice defect in brain by nutritional supplementation. Hum. Mol. Genet. 20, 4093-101 (2011)). Within a clinical trial kinetin was well-tolerated, safe, and resulted in increased levels of correctly spliced mRNA in vivo. This indicates that drugs targeting U1-mediated splicing defects have a high potential to become a therapeutic tool and supports future clinical trials.

[0145] The therapeutic potential of combined RECTAS and PB treatment for targeting defective splicing is further substantiated by additional mutations in monogenic PD. Homozygous mutations in PARK7 affecting the 5 consensus splice site for U1 were described as a c.91-2AG mutation in an Iranian family and a c.317-322del mutation in a Turkish family. Moreover, a novel c.1488+1 G>A mutation in the PINK1 gene was shown to affect a U1 binding site and cause an in-frame deletion of exon 7 (Samaranch, L. et al. PINK1-linked parkinsonism is associated with Lewy body pathology. Brain 133, 1128-42 (2010)). Interestingly, as for the c.192G>C mutation in PARK7, no NMD of the mutant PINK1 mRNA was observed in affected carriers.

[0146] These findings already indicate that the pathogenic relevance of exonic splicing mutations may be underestimated in PD, as it is known to represent a common cause for human diseases with approximately 15% of all mutations caused by aberrant spliceosome function, that translates into defective pre-mRNA processing (Nissim-Rafinia, M. & Kerem, B. Splicing regulation as a potential genetic modifier. Trends Genet. 18, 123-7 (2002)). To further assess the potential role of splicing-related mutations in PD, we performed an exploratory analysis of exonic mutations affecting U1-mediated splicing using publicly available databases including whole exome sequencing results from the PPMI study (Parkinson Progression Marker Initiative. The Parkinson Progression Marker Initiative (PPMI). Prog. Neurobiol. 95, 629-35 (2011).). We found a significantly higher burden of rare exonic variants affecting U1 snRNA binding sites in 380 PD patients compared to 162 control samples (n=109, P-value=0.015, OR=1.55, CI=1.05-2.29), with the majority of the burden defined by brain expressed genes (P-value=0.047, OR=1.50, CI=0.95-2.38). These results are in line with a recent study showing that approximately 10% of pathogenic missense variants predicted to alter protein code are essentially disrupting splicing (Soemedi, R. et al. Pathogenic variants that alter protein code often disrupt splicing. Nat. Genet. (2017). doi:10.1038/ng.3837). This supports our findings of an enrichment of missense and synonymous disease-associated variants in the exon boundary and underscores the therapeutic potential of compounds acting on pathological splicing.

[0147] Our study illustrates the promise for treatment concepts in precision medicine in PD that focus on genetic and molecular stratification. In order to account for the increasingly recognized heterogeneity in PD and other neurodegenerative disorders, new strategies have to be developed for the stratification of patients along shared pathogenic mechanisms.

[0148] Variant Annotation and Filtering:

[0149] In the current study we only focused on the intronic 5 splice single nucleotide variants (SNVs). Multi-allellic variants were decomposed by using variant-tests (Tan, A., Abecasis, G. R. & Kang, H. M. Unified representation of genetic variants. Bioinformatics 31, 2202-2204 (2015)) and left normalized by bcftools (Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078-2079 (2009)). Variants were annotated by using ANNOVAR (Wang, K., Li, M. & Hakonarson, H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 38, e164-e164 (2010)) version 2016December05 using RefSeq gene annotations and the dbNSFP v3.0 (Liu, X., Wu, C., Li, C. & Boerwinkle, E. dbNSFP v3.0: A One-Stop Database of Functional Predictions and Annotations for Human Nonsynonymous and Splice-Site SNVs. Hum. Mutat. 37, 235-241 (2016)) prediction scores. Only rare variants, as defined by variants with a MAF<0.05 in the current dataset and also European population of public databases such as 1000 genomes (The 1000 Genomes Project Consortium. A global reference for human genetic variation. Nature 526, 68-74 (2015)), ExAC (release 0.3), and the Exome variant server (http://evs.gs.washington.edu/EVS/) were selected. In order to prioritize the 5 splice variants based on their deleteriousness, we used three different scores. The first score is generated by using the MaxEntScan method (Yeo, G. & Burge, C. B. Maximum entropy modeling of short sequence motifs with applications to RNA splicing signals. J. Comput. Biol. J. Comput. Mol. Cell Biol. 11, 377-394 (2004).) which is based on the maximum entropy principle. The other two scores were ensemble scores (dbscSNV_ADA and dbscSNV_RF) generated from multiple splice site prediction tools (Jian, X., Boerwinkle, E. & Liu, X. In silico prediction of splice-altering single nucleotide variants in the human genome. Nucleic Acids Res. 42, 13534-13544 (2014)) which are available as part of dbNSFP database (Liu, X., Wu, C., Li, C. & Boerwinkle, E. dbNSFP v3.0: A One-Stop Database of Functional Predictions and Annotations for Human Nonsynonymous and Splice-Site SNVs. Hum. Mutat. 37, 235-241 (2016)).

[0150] Generation of MaxEntScan score:

[0151] In the current study we only focused on the variants present in the consensus splice site positions i.e., +3 to 6 from the exon/intron boundary. To prioritize such variants using MaxEntScan method, for each SNV that lies in the consensus splice site region (SNV.sub.splice) a wild type 9 mer (WT) was extracted from the reference genome (hg19). Then the variant was introduced within the WT by using the python module pyfaidx (Shirley, M. D., Ma, Z., Pedersen, B. S. & Wheelan, S. J. Efficient pythonic access to FASTA files using pyfaidx. (PeerJ PrePrints, 2015).), hence creating a mutated consensus splice site (MUT) sequence for each variant. In the next steps the scores were calculated for both WT and MUT sequences by using the scripts provided in the MaxEntScan website (http://genes.mit.edu/burgelab/maxent/Xmaxentscan_scoreseq.html). The relative percentage change (maxentscan_change) was calculated by using the formula below and all the variant sites were annotated with the WT.sub.score and maxentscan_change.

[00001]$\mathrm{maxentscan\_change}=\left(\frac{\mathrm{WTscore}-\mathrm{MUTscore}}{\mathrm{WTscore}}\right)*100$

[0152] Deleterious splice site variants (SNV.sub.splicedef) were defined as SNVs with the following criteria: WT.sub.score>5 and maxentscan_change>70 and dbscSNV_ADA score>0.9 and dbscSNV_RF score>0.9. If the ensemble scores are not available for any particular variant only MaxEntScan methods (WT.sub.score>5 and maxentscan_change>70) was used.

[0153] Burden Tests:

[0154] To test the genome-wide differential burden of SNV.sub.splicedef variants in cases compared to the controls CMCFisherExact test available part of rvtests (Zhan, X., Hu, Y., Li, B., Abecasis, G. R. & Liu, D. J. RVTESTS: an efficient and comprehensive tool for rare variant association analysis using sequence data. Bioinformatics 32, 1423-1426 (2016)) with default parameters was used and one sided P-values were reported. CMCFisherExact test collapses the variants in the given regions (gene or gene-set) and performs the fisher exact test. We used a list of brain expressed genes to see if there is an increased burden in the brain expressed genes (n=14,177) compared to the non-brain expressed genes (n=6,428).

[0155] Results:

[0156] Burden Analysis:

[0157] After the filtering based on ethnicity, cryptic relatedness and quality parameters the final dataset comprised of 380 PD and 162 control samples. A total of 128 SNV.sub.splicedef variants were included in the final analysis. We observed a genome-wide significant burden in cases compared to the controls (P-value=0.021, OR=1.49, CI=1.02-2.18). The signal is coming mainly from the exonic SNV.sub.splicedef variants (n=109, P-value=0.015, OR=1.55, CI=1.05-2.29) rather than the intronic ones (n=19, P-value=0.24, OR=1.34, CI=0.68-2.65). In the exonic SNV.sub.splicedef variants majority of the burden is caused by the brain expressed genes (P-value 0.047, OR=1.50, CI=0.95-2.38), compared to the genes that are not expressed in brain (P-value 0.12, OR=1.37, CI=0.84-2.23). The SNV.sub.splicedef variants are provided in the supplemental table 1, along with their respective case-control counts.

[0158] Research co-funded by the Fonds National de la Recherche Luxembourg (FNR) under grant number FNR/P13/6682797/Krger

[0159] The invention illustratively described herein may suitably be practiced in the absence of any element or elements, limitation or limitations, not specifically disclosed herein. Additionally, the terms and expressions employed herein have been used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed. Thus, it should be understood that although the present invention has been specifically disclosed by exemplary embodiments and optional features, modification and variation of the inventions embodied therein herein disclosed may be resorted to by those skilled in the art, and that such modifications and variations are considered to be within the scope of this invention.

[0160] The invention has been described broadly and generically herein. Each of the narrower species and subgeneric groupings falling within the generic disclosure also form part of the invention. This includes the generic description of the invention with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.

[0161] Other embodiments are within the following claims. In addition, where features or aspects of the invention are described in terms of Markush groups, those skilled in the art will recognize that the invention is also thereby described in terms of any individual member or subgroup of members of the Markush group.