Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
INHIBITORS OF SRSF1 TO TREAT NEURODEGENERATIVE DISORDERS
Document Type and Number:
WIPO Patent Application WO/2017/207979
Kind Code:
A1
Abstract:
The present disclosure relates to antagonists that target Serine/Arginine-Rich Splicing Factor (SRSF1); expression vectors comprising SRSF1 antagonists;and the use of such antagonists in therapy for the treatment of neurodegenerative disorders and cancer and screening methods that identify agents that inhibit the expression or activity of SRSF1.

Inventors:
HAUTBERGUE GUILLAUME (GB)
AZZOUZ MIMOUN (GB)
WHITWORTH ALEXANDER (GB)
SHAW PAMELA (GB)
Application Number:
PCT/GB2017/051539
Publication Date:
December 07, 2017
Filing Date:
May 30, 2017
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV SHEFFIELD (GB)
International Classes:
C12N15/113; A61K38/45; A61K48/00; C12N9/12
Domestic Patent References:
WO2016077687A12016-05-19
Other References:
CLARIBEL D. WEE ET AL: "Targeting SR Proteins Improves SMN Expression in Spinal Muscular Atrophy Cells", PLOS ONE, vol. 9, no. 12, 15 December 2014 (2014-12-15), pages e115205, XP055398623, DOI: 10.1371/journal.pone.0115205
OLGA ANCZUKÓW ET AL: "The splicing factor SRSF1 regulates apoptosis and proliferation to promote mammary epithelial cell transformation", NAT. STRUCT. MOL. BIOL., vol. 19, no. 2, 15 January 2012 (2012-01-15), pages 220 - 228, XP055398714, ISSN: 1545-9993, DOI: 10.1038/nsmb.2207
S. DAS ET AL: "Emerging Functions of SRSF1, Splicing Factor and Oncoprotein, in RNA Metabolism and Cancer", MOLECULAR CANCER RESEARCH, vol. 12, no. 9, 7 May 2014 (2014-05-07), US, pages 1195 - 1204, XP055279178, ISSN: 1541-7786, DOI: 10.1158/1541-7786.MCR-14-0131
A MAVROU ET AL: "Serine-arginine protein kinase 1 (SRPK1) inhibition as a potential novel targeted therapeutic strategy in prostate cancer", ONCOGENE, vol. 34, no. 33, 10 November 2014 (2014-11-10), pages 4311 - 4319, XP055398710, ISSN: 0950-9232, DOI: 10.1038/onc.2014.360
SEBASTIAN OLTEAN ET AL: "SRPK1 inhibition in vivo : modulation of VEGF splicing and potential treatment for multiple diseases", BIOCHEMICAL SOCIETY TRANSACTIONS, vol. 40, no. 4, 1 August 2012 (2012-08-01), GB, pages 831 - 835, XP055398712, ISSN: 0300-5127, DOI: 10.1042/BST20120051
SARAH CRUNKHORN: "Neurodegenerative disease: Phosphatase inhibitor prevents protein-misfolding diseases", NATURE REVIEWS. DRUG DISCOVERY, vol. 14, no. 6, 15 May 2015 (2015-05-15), GB, pages 386 - 386, XP055398881, ISSN: 1474-1776, DOI: 10.1038/nrd4638
MAKRINA DANIILIDOU ET AL: "Detection of elevated antibodies against SR protein kinase 1 in the serum of Alzheimer's disease patients", JOURNAL OF NEUROIMMUNOLOGY, ELSEVIER SCIENCE PUBLISHERS BV, NL, vol. 238, no. 1, 24 June 2011 (2011-06-24), pages 67 - 72, XP028292973, ISSN: 0165-5728, [retrieved on 20110701], DOI: 10.1016/J.JNEUROIM.2011.06.013
GUILLAUME M. HAUTBERGUE ET AL: "SRSF1-dependent nuclear export inhibition of C9ORF72 repeat transcripts prevents neurodegeneration and associated motor deficits", NATURE COMMUNICATIONS, vol. 8, 5 July 2017 (2017-07-05), United Kingdom, pages 16063, XP055398627, ISSN: 2041-1723, DOI: 10.1038/ncomms16063
Attorney, Agent or Firm:
SYMBIOSIS IP LIMITED (GB)
Download PDF:
Claims:
Claims

1. An expression vector comprising a transcription cassette comprising: a nucleic acid molecule encoding an antagonistic agent wherein said agent inhibits the expression of a nucleic acid encoding a Serine/Arginine-Rich Splice Factor [SRSF1] or inhibits the activity of a SRSF1 protein wherein said nucleic acid molecule is operably linked to a promoter adapted to express said agent.

2. The vector according to claim 1 wherein said nucleic acid molecule encodes nucleic acid based agent.

3. The vector according to claim 1 wherein said nucleic acid molecule encodes an antisense nucleic acid. 4. The vector according to claim 2 or 3 wherein said nucleic acid based molecule is an inhibitory RNA.

5. The vector according to claim 4 wherein said inhibitory RNA is a siRNA or shRNA molecule. 6. The vector according to any one of claims 3 to 5 wherein said inhibitory RNA comprises or consists of a nucleotide sequence as set forth in SEQ ID NO: 5, 6, 7, 8, 9, 10, 1 1 , 12, 13 or 14 or 47, 48, 50, 99, 154, 155, 156, 157, 158, 60, 97, 98, 100, 101 , 102, 103, 104, 105, 106, 107, 109, 1 11 , 113, 115, 117, 119, 121 , 123, 125, 127, 129, 131 , 133, 135, 137, 139 or 141. 7. The vector according to claim 1 wherein said agent is a peptide.

8. The vector according to claim 7 wherein said peptide comprises or consists of an amino sequence as set forth in SEQ ID NO: 15. 9. The vector according to claim 1 wherein said agent is a dominant negative protein comprising a modification of the amino acid sequence set forth in SEQ ID NO: 2 or 4.

10. The vector according to claim 9 wherein said dominant negative protein comprises or consists of an amino acid sequence as set forth in SEQ ID NO: 2 or 4 wherein said amino acid sequence is modified by addition, deletion or substitution of one or more amino acid residues.

11. The vector according to claim 9 or 10 wherein said dominant negative protein comprises or consists of the amino acid sequence as set forth in SEQ ID NO: 16 or 17.

12 The vector according to claim 1 wherein said antagonistic agent is a protein kinase specific for SRSF1 to maintain the phosphorylation state of SRSF1.

13. The vector according to claim 12 wherein said protein kinase is SRPK1.

14. The vector according to claim 13 wherein SRPK1 is encoded by a nucleic acid molecule comprising a nucleotide sequence as set forth in SEQ ID NO: 18.

15. The vector according to claim 13 wherein SRPK1 comprises an amino acid sequence as set forth in SEQ ID NO: 19.

16. The vector according to claim 12 wherein said protein kinase is CDC like kinase 1.

17. The vector according to claim 16 wherein CDC like kinase 1 is encoded by a nucleic acid molecule comprising a nucleotide sequence as set forth in SEQ ID NO: 20.

18. The vector according to claim 16 wherein CDC like kinase 1 comprises an amino acid sequence as set forth in SEQ ID NO: 21.

19. The vector according to claim 1 wherein said antagonist agent is a phosphatase inhibitor specific for human protein phosphatase-1 and inhibits the expression of human protein phosphatase.

20. The vector according to claim 19 wherein human protein phosphatase-1 is encoded by a nucleotide sequence as set forth in SEQ ID NO: 22.

21. The vector according to claim 19 said antagonist is an inhibitory RNA.

22. The vector according to any one of claims 1 to 21 wherein said promoter is a constitutive promoter.

23. The vector according to any one of claims 1 to 21 wherein said promoter is a regulated promoter, for example an inducible or cell specific promoter.

24. The vector according to any one of claims 1 to 23 wherein said expression vector is a viral based expression vector. 25. The expression vector according to claim 24 wherein said viral based vector is an adeno-associated virus [AAV].

26. The expression vector according to claim 25 wherein said viral based vector is AAV9. 27. The expression vector according to claim 24 wherein said viral based vector is a lentiviral vector.

28. A pharmaceutical composition comprising an expression vector according to any one of claims 1 to 27 and an excipient or carrier. 29. An expression vector according to any one of claims 1 to 27, for use as a medicament.

30. An expression vector according to any one of claims 1 to 29 for use in the treatment of a neurodegenerative disease.

31. The vector according to claim 30 wherein said neurodegenerative disease is selected from the group consisting of: motor neurone disease, frontotemporal lobar dementia (FTLD), Huntington's like disorder, primary lateral sclerosis, progressive muscular atrophy, corticobasal syndrome, Alzheimer's disease and Dementia with Lewy Bodies.

32. The agent or vector according to claim 30 or 31 wherein said neurodegenerative disease is motor neurone disease.

33. An expression vector according to any one of claims 1 to 29 for use in the treatment of cancer.

34. The expression vector according to claim 33 wherein said cancer is metastatic cancer.

Description:
Gene Therapy

Field of the Disclosure

The present disclosure relates to antagonists that target Serine/Arginine-Rich Splicing Factor 1 (SRSF1); expression vectors comprising SRSF1 antagonists; and the use of such antagonists in therapy for the treatment of neurodegenerative disorders and cancer.

Background the Disclosure

Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are adult-onset neurodegenerative diseases with no effective treatment. FTD is a common cause of early onset dementia characterised by a progressive loss of neuronal cells in frontal and temporal lobe leading to alterations of cognitive function and personality features, leaving patients unable to care for themselves and resulting in death from between 2-15 years from disease onset. ALS is the premature degeneration of motor neurons leading to muscle atrophy and paralysis leading to death from respiratory failure within 2 to 3 years from symptom onset. MND is a degenerative disease of motor neurons of the nerve cells which control muscles. MND leads to progressive paralysis and death within 2-5 years from when symptoms appear. The disease prevalence is around 6-8/100,000 people.

Neuroprotective treatment options are currently extremely limited and the anti-glutamatergic agent riluzole prolongs survival in ALS patients by only approximately 3 months or around 6 months of life extension in MND patients. ALS and FTD show a substantial clinical and pathological overlap, 40-50% of ALS patients present FTD dysfunctions, whereas in about 5- 10% FTD cases patients develop ALS and are therefore proposed to constitute one disease spectrum. A number of genes have been discovered and are thought to be responsible for ALS such as SOD1 , TARDBP, FUS, OPTN and VCP. Moreover, both diseases are characterised by the presence of protein transactivation response DNA-binding protein (TDP-43) inclusions throughout the central nervous system. However the most commonly identified genetic cause of ALS and FTD involves polymorphic repeat expansions composed of hundreds to thousands of the GGGGCC hexanucleotide-repeat sequence (hereafter abbreviated G4C2) in the first intron of the C90RF72 gene, with autosomal dominant inheritance and incomplete penetrance ( 1, 2). The C90RF72 gene is located on the short arm of chromosome 9 open reading frame 72 and occurs in two isoforms. This protein is found in many regions of the brain in the cytoplasm of neurons as well as in presynaptic terminals, and is also found mutated in diseases such as frontotemporal lobar dementia (FTLD), Huntington's like l disorder, primary lateral sclerosis, progressive muscular atrophy, corticobasal syndrome, Alzheimer's disease and Dementia with Lewy Bodies. The pathophysiology potentially involves three extensively-studied mechanisms which may co-exist: (i) RNA toxic gain-of- function by sequestration of RNA-binding factors (5-9), (ii) protein toxic gain-of-function due to repeat associated non-ATG (RAN) translation that occurs in all sense and antisense reading frames to produce five dipeptide-repeat proteins (DPRs (10-14) and (iii) haplo- insufficiency due to decreased expression of the C90RF72 protein (1 , 15, 16).

Typically, therapeutic approaches have targeted C90RF72. WO2016/024205 discloses oligomers complimentary to the C90RF72 gene and in particular compositions for targeting RNA containing a pathological number of hexanucleotide repeats for use in the treatment of a neurological disorder selected from the group of ALS and FTD. WO2014/062691 discloses compositions and methods for reducing the expression of C90RF72 mRNA and protein in an animal for the treatment of ALS, FTD, corticalbasal degeneration syndrome, atypical Parkinson syndrome and olivopontocerebellar degeneration by the delivery of antisense RNA directed against C90RF27 nucleic acids. The nucleic acids are administered into the central nervous system intrathecal^ or intraventriculary. Only two antisense drugs have been approved; fomivirsen for the treatment of cytomegalovirus retinitis and mipomersen for the treatment of homozygous familial hypercholesterolemia. Viral vectors present an alternative form of delivery vehicle for genetic material. Viral vector systems, based on adeno-associated viruses and lentiviruses, are ideally suited to mediate RNAi because they can safely transduce a wide range of tissues and provide sustained levels of gene expression. Nuclear export of mRNA is mediated by NXF1 (nuclear export factor 1) protein and export adaptors such as SRSF1 (serine/arginine-rich splicing factor 1) and ALYREF (Aly/REF export factor) are thought to increase the affinity for mature mRNAs, preventing the export of unprocessed transcripts. Immunohistochemistry in central nervous system tissue from C9orf72+ patients with ALS demonstrated co-localization of GGGGCC repeat RNA with SRSF2, hnRNP H1/F, ALYREF and hnRNP A1 in cerebellar granule cells and with SRSF2, hnRNP H1/F and ALYREF in motor neurons, the primary target of pathology in ALS. Direct binding of proteins to GGGGCC repeat RNA was also confirmed.

The present disclosure has identified that excessive binding of nuclear export adaptor(s) onto G4C2-repeat transcripts forces interactions with NXF1 , overriding the normal nuclear retention mechanisms and teaches that depletion of factors such as SRSF (also known as SF2 or ASF) 1 confers neuroprotection. Statements of Invention

According to an aspect of the invention there is provided an antagonistic agent that inhibits the expression of a nucleic acid molecule encoding Serine/Arginine-Rich Splice Factor 1 [SRSF1] or inhibits the activity of a SRSF1 protein.

Inhibition of expression defines the reduction of expression ranging from 1 -100% when compared to expression of the nucleic acid/protein found in the wild type.

According to a further aspect of the invention there is provided a transcription cassette comprising: a nucleic acid molecule encoding an antagonistic agent wherein said agent inhibits the expression of a nucleic acid encoding a Serine/Arginine-Rich Splice Factor [SRSF1] or inhibits the activity of a SRSF1 protein wherein said nucleic acid molecule is operably linked to a promoter adapted to express said agent.

In a preferred embodiment of the invention SRSF1 is encoded by a nucleic acid molecule comprising a nucleotide sequence as set forth in SEQ ID NO: 1 , or polymorphic sequence variant that has 90%-99% sequence identity over the full length nucleotide sequence as set forth in SEQ ID NO: 1.

In a preferred embodiment of the invention SRSF1 is encoded by a nucleic acid molecule comprising a nucleotide sequence as set forth in SEQ ID NO: 3 or polymorphic sequence variant that has 90%-99% sequence identity over the full length nucleotide sequence as set forth in SEQ ID NO: 3.

In a preferred embodiment of the invention SRSF1 is represented by the amino acid sequence as set forth in SEQ ID NO:2, or a polymorphic sequence variant that has 90-99% sequence identity over the full length amino acid sequence as set forth in SEQ ID NO: 2.

In an preferred embodiment of the invention SRSF1 is represented by the amino acid sequence as set forth in SEQ ID NO: 4, or a polymorphic sequence variant that has 90-99% sequence identity over the full length amino acid sequence as set forth in SEQ ID NO: 4. In a preferred embodiment of the invention said nucleic acid molecule encodes nucleic acid based agent. In a preferred embodiment of the invention said nucleic acid molecule encodes an antisense nucleic acid.

In a preferred embodiment of the invention said nucleic acid based molecule is an inhibitory RNA.

In a preferred embodiment of the invention said inhibitory RNA is a siRNA or shRNA or miRNA molecule.

A technique to specifically ablate gene function is through the introduction of double stranded RNA, also referred to as small inhibitory or interfering RNA (siRNA), into a cell which results in the destruction of mRNA complementary to the sequence included in the siRNA molecule. The siRNA molecule comprises two complementary strands of RNA (a sense strand and an antisense strand) annealed to each other to form a double stranded RNA molecule. The siRNA molecule is typically derived from exons of the gene which is to be ablated. The mechanism of RNA interference is being elucidated. Many organisms respond to the presence of double stranded RNA by activating a cascade that leads to the formation of siRNA. The presence of double stranded RNA activates a protein complex comprising RNase III which processes the double stranded RNA into smaller fragments (siRNAs, approximately 21-29 nucleotides in length) which become part of a ribonucleoprotein complex. The siRNA acts as a guide for the RNase complex to cleave mRNA complementary to the antisense strand of the siRNA thereby resulting in destruction of the mRNA.

In a preferred embodiment of the invention said inhibitory RNA molecule is between 19 nucleotides [nt] and 29nt in length. More preferably still said inhibitory RNA molecule is between 21 nt and 27nt in length. Preferably said inhibitory RNA molecule is about 21 nt in length.

In a preferred embodiment of the invention said inhibitory RNA comprises or consists of a nucleotide sequences set forth in SEQ ID NO: 5, 6, 7, 8, 9, 10, 11 , 12, 13 or 14.

In a preferred embodiment of the invention said inhibitory RNA comprises or consists of a nucleotide sequences set forth in SEQ ID NO: 97, 98, 99, 100, 101 , 102, 103, 104, 105 or 106.

In a preferred embodiment of the invention said inhibitory RNA comprises or consists of a nucleotide sequences set forth in SEQ ID NO: 107, 109, 11 1 , 113, 115, 117, 119, 121 , 123 or 125. In a preferred embodiment of the invention said inhibitory RNA comprises or consists of a nucleotide sequences set forth in SEQ ID NO: 127, 129, 131 , 133, 135, 137, 139 or 141.

In a preferred embodiment of the invention said inhibitory RNA comprises or consists of a nucleotide sequences set forth in SEQ ID NO: 47, 48, 50, 99, 154, 155, 156, 157, 158 or 60. In a preferred embodiment of the invention said nucleic acid based agent comprises modified nucleotides.

In an alternative embodiment of the invention said agent is a peptide.

In a preferred embodiment of the invention said peptide comprises an amino acid sequence that is at least 32 amino acids in length and comprises the amino acid sequence set forth in SEQ ID NO: 15.

In a preferred embodiment of the invention said peptide is at least 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95 or at least 100 amino acids in length but less than the full length amino acid sequence set forth in SEQ ID NO: 2 or 4.

In a preferred embodiment of the invention said peptide consists of an amino sequence as set forth in SEQ ID NO: 15. In a preferred embodiment of the invention said peptide is modified, for example said peptide is cyclised.

In an alternative embodiment of the invention said protein is a dominant negative protein comprising a modification of the amino acid sequence set forth in SEQ ID NO: 2 or 4.

In a preferred embodiment of the invention said dominant negative protein comprises or consists of an amino acid sequence as set forth in SEQ ID NO: 2 or 4 wherein said amino acid sequence is modified by addition, deletion or substitution of one or more amino acid residues.

In a preferred embodiment of the invention said modified protein comprises or consists of the amino acid sequence as set forth in SEQ ID NO: 16 or 17.

In a preferred embodiment of the invention said antagonist agent is a protein kinase specific for SRSF1 to maintain the phosphorylation state of SRSF1. In a preferred embodiment of the invention said protein kinase is SRPK1.

In a preferred embodiment of the invention SRPK1 is encoded by a nucleic acid molecule comprising a nucleotide sequence as set forth in SEQ ID NO: 18.

In an alternative embodiment of the invention SRPK1 comprises an amino acid sequence as set forth in SEQ ID NO: 19. In an alternative embodiment of the invention said protein kinase is CDC like kinase 1.

In a preferred embodiment of the invention CDC like kinase 1 is encoded by a nucleic acid molecule comprising a nucleotide sequence as set forth in SEQ ID NO: 20.

In an alternative embodiment of the invention CDC like kinase 1 comprises an amino acid sequence as set forth in SEQ ID NO: 21.

In an alternative embodiment of the invention said antagonist agent is a phosphatase inhibitor specific for human protein phosphatase-1. In a preferred embodiment of the invention said antagonist inhibits expression or activity of human protein phosphatase-1.

In a preferred embodiment of the invention human protein phosphatase-1 is encoded by a nucleotide sequence as set forth in SEQ ID NO: 22.

In a preferred embodiment of the invention said antagonist is an inhibitory RNA.

In a preferred embodiment of the invention said promoter is a constitutive promoter. In an alternative embodiment of the invention said promoter is a regulated promoter, for example an inducible or cell specific promoter.

According to a further aspect of the invention there is provided an expression vector comprising a transcription cassette according to the invention. A number of viruses are commonly used as vectors for the delivery of exogenous genes. Commonly employed vectors include recombinantly modified enveloped or non-enveloped DNA and RNA viruses, for example baculoviridiae, parvoviridiae, picornoviridiae, herpesveridiae, poxviridae, adenoviridiae, picornnaviridiae or retroviridae. Chimeric vectors may also be employed which exploit advantageous elements of each of the parent vector properties (See e.g., Feng, et al (1997) Nature Biotechnology 15:866-870). Such viral vectors may be wild-type or may be modified by recombinant DNA techniques to be replication deficient, conditionally replicating or replication competent. Conditionally replicating viral vectors are used to achieve selective expression in particular cell types while avoiding untoward broad spectrum infection. Examples of conditionally replicating vectors are described in Pennisi, E. (1996) Science 274:342-343; Russell, and S.J. (1994) Eur. J. of Cancer 30A(8):1 165-1171.

Preferred vectors are derived from the adenoviral, adeno-associated viral and retroviral genomes.

In a preferred embodiment of the invention said expression vector is a viral based expression vector. In a preferred embodiment of the invention said viral based vector is an adeno-associated virus [AAV].

In a preferred embodiment of the invention said viral based vector is AAV9. In an alternative preferred embodiment of the invention said viral based vector is a lentiviral vector.

According to a further aspect of the invention there is provided a pharmaceutical composition comprising an expression vector according to the invention and an excipient or carrier. The agent or expression vector compositions of the present invention are administered in pharmaceutically acceptable preparations. Such preparations may routinely contain pharmaceutically acceptable concentrations of salt, buffering agents, preservatives, compatible carriers and supplementary therapeutic agents'. The agent or expression vector compositions of the invention can be administered by any conventional route, including injection or by gradual infusion over time. The agent or expression vector compositions of the invention are administered in effective amounts. An "effective amount" is that amount of the agent or expression vector that alone, or together with further doses, produces the desired response. In the case of treating a disease, the desired response is inhibiting the progression of the disease. This may involve only slowing the progression of the disease temporarily, although more preferably, it involves halting the progression of the disease permanently. This can be monitored by routine methods. Such amounts will depend, of course, on the particular condition being treated, the severity of the condition, the individual patient parameters including age, physical condition, size and weight, the duration of the treatment, the nature of concurrent therapy (if any), the specific route of administration and like factors within the knowledge and expertise of the health practitioner. These factors are well known to those of ordinary skill in the art and can be addressed with no more than routine experimentation. It is generally preferred that a maximum dose of the individual components or combinations thereof be used, that is, the highest safe dose according to sound medical judgment. It will be understood by those of ordinary skill in the art, however, that a patient may insist upon a lower dose or tolerable dose for medical reasons, psychological reasons or for virtually any other reasons.

The agent or expression vector compositions used in the foregoing methods preferably are sterile and contain an effective amount of agent or expression vector according to the invention for producing the desired response in a unit of weight or volume suitable for administration to a patient. The doses of agent or vector administered to a subject can be chosen in accordance with different parameters, in particular in accordance with the mode of administration used and the state of the subject. Other factors include the desired period of treatment. In the event that a response in a subject is insufficient at the initial doses applied, higher doses (or effectively higher doses by a different, more localized delivery route) may be employed to the extent that patient tolerance permits. Other protocols for the administration of agent or vector compositions will be known to one of ordinary skill in the art, in which the dose amount, schedule of injections, sites of injections, mode of administration and the like vary from the foregoing. The administration of compositions to mammals other than humans, (e.g. for testing purposes or veterinary therapeutic purposes), is carried out under substantially the same conditions as described above. A subject, as used herein, is a mammal, preferably a human, and including a non-human primate, cow, horse, pig, sheep, goat, dog, cat or rodent. When administered, the agent or expression vector compositions of the invention are applied in pharmaceutically-acceptable amounts and in pharmaceutically-acceptable compositions. The term "pharmaceutically acceptable" means a non-toxic material that does not interfere with the effectiveness of the biological activity of the active agent. Such preparations may routinely contain salts, buffering agents, preservatives, compatible carriers, and optionally other therapeutic agents' (e.g. those typically used in the treatment of the specific disease indication). When used in medicine, the salts should be pharmaceutically acceptable, but non-pharmaceutically acceptable salts may conveniently be used to prepare pharmaceutically-acceptable salts thereof and are not excluded from the scope of the invention. Such pharmacologically and pharmaceutically-acceptable salts include, but are not limited to, those prepared from the following acids: hydrochloric, hydrobromic, sulfuric, nitric, phosphoric, maleic, acetic, salicylic, citric, formic, malonic, succinic, and the like. Also, pharmaceutically-acceptable salts can be prepared as alkaline metal or alkaline earth salts, such as sodium, potassium or calcium salts.

The pharmaceutical compositions containing the agents or expression vectors according to the invention may contain suitable buffering agents, including: acetic acid in a salt; citric acid in a salt; boric acid in a salt; and phosphoric acid in a salt. The pharmaceutical compositions also may contain, optionally, suitable preservatives, such as: benzalkonium chloride; chlorobutanol; parabens and thimerosal.

The agent or expression vector compositions may conveniently be presented in unit dosage form and may be prepared by any of the methods well-known in the art of pharmacy. All methods include the step of bringing the active agent into association with an agent or vector which constitutes one or more accessory ingredients. The preparation may be formulated according to known methods using suitable dispersing or wetting agents and suspending agents. The sterile injectable preparation also may be a sterile injectable solution or suspension in a non-toxic parenterally-acceptable diluent or solvent, for example, as a solution in 1 , 3-butanediol. Among the acceptable solvents that may be employed are water, Ringer's solution, and isotonic sodium chloride solution. In addition, sterile, fixed oils are conventionally employed as a solvent or suspending medium. For this purpose any bland fixed oil may be employed including synthetic mono-or di-glycerides. In addition, fatty acids such as oleic acid may be used in the preparation of injectables. Carrier formulation suitable for oral, subcutaneous, intravenous, intramuscular, etc. administrations can be found in Remington's Pharmaceutical Sciences, Mack Publishing Co., Easton, PA.

According to a further aspect of the invention there is provided an antagonistic agent according to the invention for use as a medicament. According to a further aspect of the invention there is provided an expression vector according to the invention for use as a medicament.

According to an aspect of the invention there is provided an antagonistic agent according to the invention for use in the treatment of a neurodegenerative disease.

According to a further aspect of the invention there is provided an expression vector according to the invention for use in the treatment of a neurodegenerative disease. In a preferred embodiment of the invention said neurodegenerative disease is selected from the group consisting of: motor neurone disease, frontotemporal lobar dementia (FTLD), Huntington's like disorder, primary lateral sclerosis, progressive muscular atrophy, corticobasal syndrome, Alzheimer's disease and Dementia with Lewy Bodies. In a preferred embodiment of the invention said neurodegenerative disease is motor neurone disease.

According to a further aspect of the invention there is provided a method for the treatment of a neurodegenerative disease comprising administering a therapeutically effective amount of an antagonistic agent according to the invention to prevent and/or treat said neurodegenerative disease.

According to a further aspect of the invention there is provided a method for the treatment of a neurodegenerative disease comprising administering a therapeutically effective amount of an expression vector according to the invention to prevent and/or treat said neurodegenerative disease.

According to an aspect of the invention there is provided an antagonistic agent according to the invention for use in the treatment of cancer, in particular metastatic cancer.

According to a further aspect of the invention there is provided an expression vector according to the invention for use in the treatment of cancer, in particular metastatic cancer.

Cancer includes reference to tumours. For example adenocarcinomas include malignancies such as most colon cancers, renal-cell carcinoma, prostate cancer, testicular tumours, non- small cell carcinoma of the lung, cancer of the small intestine and cancer of the esophagus. The term "carcinoma" is art recognized and refers to malignancies of epithelial or endocrine tissues including respiratory system carcinomas, gastrointestinal system carcinomas, genitourinary system carcinomas, testicular carcinomas, breast carcinomas, prostatic carcinomas, endocrine system carcinomas, and melanomas. Exemplary carcinomas include those forming from tissue of the cervix, lung, prostate, breast, head and neck, colon and ovary. The term "carcinoma" also includes carcinosarcomas, e.g., which include malignant tumours composed of carcinomatous and sarcomatous tissues. An "adenocarcinoma" refers to a carcinoma derived from glandular tissue or in which the tumor cells form recognizable glandular structures. In general, doses of antagonistic agents, for example, antisense oligonucleotide, siRNA or shRNA will be between 1 nM - 1 μΜ generally will be formulated and administered according to standard procedures. Preferably doses can range from 1 nM-500nM, 5nM-200nM, and 10nM-100nM. In addition dosages of peptides and/or protein antagonist are formulated and administered in doses between 1 ng and 1 mg, and preferably between 10 ng and 100 μg.

According to a further aspect of the invention there is provided the use of a nucleic acid molecule encoding SRSF1 , or a protein with the activity associated with SRSF1 in the identification of an agent that inhibits the expression or activity of SRSF1. According to a further aspect of the invention there is provided a screening method for the identification of an agent that inhibits the expression of SRSF1 or the activity of a protein with activity associated with SRSF1 comprising the steps:

i) providing a cell that expresses SRSF1 ;

ii) contacting the cell with one or more agents to be tested for inhibitory activity with respect to a nucleic acid encoding SRSF1 or a SRSF1 protein;

iii) monitoring the effect of said agent[s] on the expression or activity of SRSF1 compared to a cell that has not been contacted with said agent[s].

In a preferred method of the invention said cell is a nerve cell, for example an astrocyte.

In a preferred method of the invention said cell is modified to recombinantly express SRSF1.

In a preferred method of the invention said agent is a nucleic acid based agent. In a preferred method of the invention said nucleic acid based agent comprises modified nucleotides. The term "modified" as used herein describes a nucleic acid molecule in which: i) at least two of its nucleotides are covalently linked via a synthetic internucl eotide linkage (i.e., a linkage other than a phosphodiester linkage between the 5' end of one nucleotide and the 3' end of another nucleotide). Alternatively or preferably said linkage may be the 5' end of one nucleotide linked to the 5' end of another nucleotide or the 3' end of one nucleotide with the 3' end of another nucleotide; and/or ii) a chemical group, such as cholesterol, not normally associated with nucleic acids has been covalently attached to the single-stranded nucleic acid.

Preferred synthetic internucl eotide linkages are phosphorothioates, alkylphosphonates, phosphorodithioates, phosphate esters, alkylphosphonothioates, phosphoramidates, carbamates, phosphate triesters, acetamidates, peptides, and carboxymethyl esters.

The term "modified" also encompasses nucleotides with a covalently modified base and/or sugar. For example, modified nucleotides include nucleotides having sugars which are covalently attached to low molecular weight organic groups other than a hydroxyl group at the 3' position and other than a phosphate group at the 5' position. Thus modified nucleotides may also include 2' substituted sugars such as 2'-0-methyl-; 2-O-alkyl; 2-O-allyl; 2'-S-alkyl; 2'-S-allyl; 2'- fluoro-; 2'-halo or 2;azido-ribose, carbocyclic sugar analogues a- anomeric sugars; epimeric sugars such as arabinose, xyloses or lyxoses, pyranose sugars, furanose sugars, and sedoheptulose. Modified nucleotides are known in the art and include alkylated purines and/or pyrimidines; acylated purines and/or pyrimidines; or other heterocycles. These classes of pyrimidines and purines are known in the art and include, pseudoisocytosine; N4, N4-ethanocytosine; 8- hydroxy-N6-methyladenine; 4-acetylcytosine, 5-(carboxyhydroxylmethyl) uracil; 5- fluorouracil; 5-bromouracil;5-carboxymethylaminomethyl-2-thiouracil; 5 carboxymethylaminomethyl uracil; dihydrouracil; inosine; N6-isopentyl-adenine; I- methyladenine; 1-methylpseudouracil; 1-methylguanine; 2,2-dimethylguanine; 2- methyladenine; 2-methylguanine; 3-methylcytosine; 5-methylcytosine; N6-methyladenine; 7- methylguanine; 5- methylaminomethyl uracil; 5-methoxy amino methyl-2-thiouracil; beta-D- mannosylqueosine; 5-methoxycarbonylmethyluracil; 5-methoxyuracil; 2 methylthio-N6- isopentenyladenine; uracil-5-oxyacetic acid methyl ester; psuedouracil; 2-thiocytosine; 5- methyl-2 thiouracil, 2-thiouracil; 4-thiouracil; 5-methyluracil; N-uracil-5-oxyacetic acid methylester; uracil 5— oxyacetic acid; queosine; 2-thiocytosine; 5-propyluracil; 5- propylcytosine; 5-ethyluracil; 5-ethylcytosine; 5-butyluracil; 5-pentyluracil; 5-pentylcytosine; and 2,6,-diaminopurine; methylpsuedouracil; 1-methylguanine; 1-methylcytosine. Modified double stranded nucleic acids also can include base analogs such as C-5 propyne modified bases (see Wagner et al., Nature Biotechnology 14:840-844, 1996). The use of modified nucleotides confers, amongst other properties, resistance to nuclease digestion and improved stability. In an alternative method of the invention said agent is a peptide or protein agent.

In a preferred method of the invention said peptide agent comprises one or more modified amino acids or are cyclised. Cyclisation is known in the art, (see Scott et al Chem Biol (2001), 8:801 -815; Gellerman et al J. Peptide Res (2001), 57: 277-291 ; Dutta et al J. Peptide Res (2000), 8: 398-412; Ngoka and Gross J Amer Soc Mass Spec (1999), 10:360-363.

In a further method of the invention said agent is a small organic agent.

Throughout the description and claims of this specification, the words "comprise" and "contain" and variations of the words, for example "comprising" and "comprises", means "including but not limited to", and is not intended to (and does not) exclude other moieties, additives, components, integers or steps. "Consisting essentially" means having the essential integers but including integers which do not materially affect the function of the essential integers.

Throughout the description and claims of this specification, the singular encompasses the plural unless the context otherwise requires. In particular, where the indefinite article is used, the specification is to be understood as contemplating plurality as well as singularity, unless the context requires otherwise.

Features, integers, characteristics, compounds, chemical moieties or groups described in conjunction with a particular aspect, embodiment or example of the invention are to be understood to be applicable to any other aspect, embodiment or example described herein unless incompatible therewith.

An embodiment of the invention will now be described by example only and with reference to the following figures:

Figure 1 : Engineering human SRSF1-RNA\ depletion vectors. (A) Diagrams of pCDNA6.2- GW/EmGFP-SRSF7 miR1 , EmGFP-Sf?SF7 miR2 and EmGFP-Sf?SF7 miR1+2. The pCDNA6.2-GW/EmGFP-SRSF7 miR1 and EmGFP-SRSFI miR2 were built separately using the BLOCK-iT Pol II miR RNAi Expression Vector Kit with EmGFP (see online methods). The human SRSF1 pre-miR2 RNAi cassette was then chained by subcloning the BamHI/Xhol-cut fragment into the Bglll and Xhol sites of pcDNA6.2 GW/EmGFP-human SRSF1 miR1. (B) Plasmid diagram of EmGFP-human SRSF1 miR1 +2 cloned into the self- inactivating lentiviral (SIN-W-PGK) vector using the restriction sites Spel and Xhol;

Figure 2: Co-cultures of mouse HB9:GFP motor neurons and SRSF1-RNAi transduced patient-derived astrocytes. A) Evaluating the efficiency of lentiviral-mediated SRSF 7-RNAi depletion in HEK cells and iAstrocytes derived from control and C90RF72-ALS patients. SRSF1 transcript levels were quantified in transfected HEK cells and iAstrocytes transduced with increased MOI doses of LV-SRSF1-RNAi. snRNA U1 transcript levels were used for normalization in triplicate experiments (Mean ± SEM; two-way ANOVA, NS: non-significant, ****: p<0.001 ; N (qPCR reactions) = 6); (b) High content imaging pictures showing how the Columbus analysis software recognizes Hb9:GFP motor neurons and the axons sprouting from them over SRSF1-RNAi-transduced astrocyte background;

Figure 3: Depletion of SRSF1 leads to cytoplasmic reduction and nuclear accumulation of sense RNA foci. Representative images of sense RNA foci visualized using Cy3-CCCCGG Fluorescence In Situ Hybridization (red) by confocal microscopy in untransduced/transduced i-Astrocytes. The nuclei were stained in blue using DAPI. Arrows point to RNA foci. Cells with detectable RNA foci represent approximately 15-40% of the cell population depending on the individual patient-derived iAstrocyte line. Quantification was performed on 20-25 cells containing RNA foci (see Additional Data table 1 and table 4 for individual counts and Fig. 10E for bar chart);

Figure 4: Generation of RAN-dependent uninterrupted G4C2-sense and C4G2-antisense repeat constructs. (A) Agarose gels confirming both the annealing and concatemerisation of the G4C2x15 oligonucleotides. Arrows point to monomeric and multimeric forms of annealed oligonucleotides. (B) Trimeric oligonucleotides were treated with Mung Bean nuclease for blunt cloning. 8% acrylamide gel of Hindlll/Xhol digests of pcDNA3.1/RAN constructs containing uninterrupted G4C2-sense and C4G2-antisense repeats with 5' and 3' flanking regions. (C) Standard curve generated from the acrylamide gel analysis using the Gene Tools Image software. (D) Table showing theroretical and experimental base pair size information for Hindlll/Xhol inserts. The RAN constructs contain at least 38 G4C2-sense or 39 C4G2-antisense uninterrupted repeats based on the experimental size of the inserts (see sequences in E). (E) Sanger sequencing using betaine was also performed in the 5' and 3' directions using the T7 and T3 promoter sequencing primers respectively. Each sequence read covered the 5' or 3' flanking region and 9-16 G4C2 or C4G2 repeats prior to interruption. Sequencing traces are available on request. The RNA transcripts generated from these constructs are highlighted in blue (flanking regions) and red (sense or antisense repeats). Sequences highlighted in black/underlined correspond to the 3'end of the promoter sequence and in black/italics to the start of the terminator sequence. Note the absence of initiating codons (ATG) in both sense and antisense transcripts generated from the RAN- dependent DPR-expression constructs. Stop codons are shown in all frames in bold (SEQ ID NO 40 and 41). The constructs expressing 15 repeats in sense or antisense orientation are identical except that they only contain 15 repeats. Sequencing and size analysis further showed that the number of repeats remained stable over multiple rounds of transformation and replication in E. coli;

Figure 5: Engineering mouse SRSF1-RNA\ depletion vectors. Diagrams of the mouse EmGFP-SRSFI miR1 , EmGFP-SRSFI miR2 and EmGFP- SRSF1 miR1 +2. The pCDNA6.2- GW/EmGFP-SRSF7 miR1 and EmGFP-SRSFI miR2 were built separately using the BLOCK-iT Pol II miR RNAi Expression Vector Kit with EmGFP (see online methods). The mouse SRSF1 pre-miR2 RNAi cassette was then chained by subcloning the BamHI/Xhol-cut fragment into the Bglll and Xhol sites of pcDNA6.2 GW/EmGFP-mouse SRSF1 miR1 ;

Figure 6: SRSF1 depletion and inhibition of the SRSF1 :NXF1 interaction inhibit the nuclear export of hexanucleotide-repeat RAN-translated transcripts and the production of DPRs. Western blot quantification of DPRs. (A) Western blots from N2A cells co-transfected with either a Ctrl or SRSF1-RNAi vector and control backbone plasmid (no DPR Ctrl) or the same plasmid expressing 38 uninterrupted G4C2-sense-repeats (G4C2x38) or 39 uninterrupted C4G2-antisense-repeats (C4G2x39) were quantified in triplicate experiments (Mean ± SEM; two-way ANOVA, ****: p<0.001 ; N = 3). One replicate of western blots is presented in Fig. 11A.. (B and C) Western blots from N2A cells co-transfected with either G4C2x38 (B) or C4G2x39 (C) plasmids and control (FLAG Ctrl) or FLAG-tagged SRSF1 aa11-196 wild type (SRSF1), SRSF1-m2 or SRSF1-m4 were quantified in triplicate experiments (Mean ± SEM; two-way ANOVA, **: p<0.01 , ****: p<0.001 ; N = 3). One replicate of each western blots is presented in Fig. 11 C; (d, e,) Western blots shown in Fig 1 1A for the GP36 (d) and GA36 (e) panels were quantified in triplicate experiments (mean ± SEM; two-way ANOVA; N = 3). (f) Total extracts from N2A cells transfected with either FLAG control (FLAG Ctrl) and either FLAG-tagged SRSF1 aa11-196 wild type (SRSF1), SRSF1-m2 or SRSF1-m4 are subjected to anti-FLAG immunoprecipitation. Co-immunoprecipitation of endogenous NXF1 is assessed using anti-NXF1 antibodies; Figure 7: SRSF1 depletion in Drosophila models expressing DPRs independently of G4C2 repeat expansions. Neuronal expression of Gly-Arg (GR36) and Pro-Arg (PR36) dipeptide repeats causes adult climbing deficits that are not rescued by SRSF1 -depletion (mean ± SEM normalized to Control (GAL4/Control-RNAi); Kruskal-Wallis with Dunn's multiple comparison test, NS: non-significant, **: p<0.01 , ****: p<0.001 ; N = Control: 188, GR36+Ctrl- RNAi: 8, GR36+SRSF1-RNAi: 7, PR36+Ctrl-RNAi: 125, PR36+SRSF1-RNAi: 119);

Figure 8: Depleting SRSF1 or inhibiting its repeat-RNA sequestration and interaction with NXF1 inhibit the nuclear export of RAN-translated transcripts to alleviate neurotoxicity, (a, b) Protein: RNA UV crosslinking assays using purified recombinant proteins and 32P-end- radiolabeled G4C2x5 (a) and C4G2x5 (b) repeat RNA probes. Proteins are visualised on SDS-PAGE stained with Coomassie blue (left panels) and covalently linked RNA:protein complexes by autoradiography on Phospholmages (Right panels), (c, d) RNA immunoprecipitation (RIP) assays. Formaldehyde was added to the medium of live N2A cells co-transfected with G4C2x15, G4C2x38 (c), C4G2x15 or C4G2x39 (d) and either FLAG control (FLAG Ctrl), FLAG-tagged SRSF1 aa11-196 wild type (SRSF1) or SRSF1-m4 were subjected to anti-FLAG immunoprecipitation. Purified RNA was analysed by qRT-PCR following normalization to U1 snRNA levels in triplicate experiments (mean ± SEM; two-way ANOVA; N (qRT-PCR reactions) = 6). (e) Cellular fractionation quality check. N2A cells co- transfected with G4C2x38 and either Ctrl or SRSF1-RNAi plasmids or with G4C2x38 and either FLAG-tagged SRSF1 aa1 1-196 wild type (SRSF1) or SRSF1-m4 were subjected to cellular fractionation using hypotonic lysis conditions to yield cytoplasmic fractions. The quality of the cellular fractionation was checked using antibodies against the nuclear chromatin-remodeling SSRP1 factor and alpha-tubulin. This was repeated for each fractionation and RNA were kept for qRT-PCR analysis if the cytoplasmic fraction showed no nuclear contamination, (g) Western blots of Drosophila expressing G4C2x36 and either Ctrl or SRSF1-RNAi, subjected to cellular fractionation using hypotonic lysis to yield cytoplasmic fractions. Histone H3 is used to check for potential nuclear contamination, (h) Cytoplasmic and total G4C2-repeat sense transcript levels were normalized to Tub84b levels in G4C2x36+Ctrl-RNAi or G4C2x36+SRSF1 76 -RNAi Drosophila in triplicate experiments prior to plotting as a ratio to account for potential changes in mRNA transcription/stability (mean ± SEM; paired t-test; N (qRT-PCR reactions) = 3). In contrast to cytoplasmic levels, total levels of hexanucleotide repeat transcripts were not significantly altered upon expression of SRSF1-m4 or depletion of SRSF1 in cells or flies (Fig. 28B);

Figure 9: Depletion of SRSF1 prevents in vivo neurodegeneration and rescues locomotor function. (A) SRSF1 (red) and ALYREF (blue) transcript levels were quantified by qRT-PCR analysis in independent knockdown lines in G4C2x36 flies. Tub84b transcript levels were used for normalization in triplicate experiments (mean ± SEM; two-way ANOVA, NS: nonsignificant, ****: p<0.001 ; N (qRT-PCR reactions) = 6). (B) Targeted expression of G4C2x36 causes a rough eye phenotype which is fully rescued by Sf?SF7-knockdown. Normal eye phenotypes are presented for Ctrl-RNAi and control G4C2x3 flies. Representative scanning electron micrographs are presented below light microscopy images. Scale bar: 100 μηι. (C and D) Neuronal expression of G4C2x36 causes larval crawling (C) and adult climbing (D) deficits that are both rescued by SRSF 7-depletion (mean ± SEM normalized to control (Ctrl)- RNAi); Kruskal-Wallis with Dunn's multiple comparison test, NS: non-significant, **: p<0.01 , ***: p<0.005, ****: p<0.001 ; N (larvae) = 10; N (adults) = Control: 93, G4C2x3: 62, G4C2x36+Ctrl-RNAi: 51 , G4C2x36+SRSF1-RNAi: 50, G4C2x36+ALYREF-RNAi: 36)(e) SRSF1 depletion in Drosophila models expressing DPRs independently of G4C2 repeat expansions and RAN translation 16. (f) Neuronal expression of poly-Gly-Arg DPRs (GR36) and poly-Pro-Arg DPRs (PR36) causes adult climbing deficits that are not restored by SRSF1 depletion (mean ± 95% CI normalized to Control; N = Control (GAL4/luciferase- RNAi): 239, GR36 + Ctrl-RNAi: 12, GR36 + SRSF1-RNAi: 7, PR36 + Ctrl-RNAi: 125, PR36 + SRSF1-RNAi: 1 19);

Figure 10: Depletion of SRSF1 suppresses patient-derived C90RF72-mediated astrocytic toxicity and motor neuron death. (A) Untransfected HEK293T (UT), control (Ctrl)-RNAi, SRSF1-RNA\ or LV-Sf?S 7-RNAi and FLAG-tagged SRSF1 transfected cells were analyzed 72h post-transfection by immunoblotting using anti-FLAG, anti-ALYREF and loading control anti-a-Tubulin antibodies. (B) SRSF1 and ALYREF transcript levels were quantified by qRT- PCR analysis following normalization to U1 snRNA levels in triplicate experiments (mean ± SEM; two-way ANOVA, NS: non-significant, ****: p<0.001 ; N (qRT-PCR reactions) = 6). Three control (Ctrl-pat) and three C90RF72-ALS (C9-ALS-pat) color-coded patient lines were used. (C) Representative immunofluorescence microscopy images from iAstrocytes and motor neuron co-cultures. Arrows point to examples of axons of motor neurons. Scale bar: 100 μηι. (D) Same control (Ctrl-pat) and C90RF72-ALS (C9-ALS-pat) color-coded patient lines (panel B) are used for quantification of Hb9-GFP+ motor neuron (MN) counts in four replicate co-cultures of astrocytes and motor neurons (mean ± SEM; one-way ANOVA, NS: non-significant, ***: p<0.005, ****: p<0.001 ; N (cells) = Ctrl-pat154: 567/589/582/500, Ctrl-pat154+SRSF1-RNAi: 620/543/504/349, Ctrl-pat155: 602/610/553/571 , Ctrl- pat155+SRSF1-RNAi: 554/584/532/516, Ctrl-pat209: 519/599/584/535, Ctrl-pat209+SRSF1- RNAi: 617/486/425/572, C9-ALS-pat78: 352/279/294/258, C9-ALS-pat78+SRSF1-RNAi: 569/451/398/583, C9-ALS-pat183: 200/188/154/145, C9-ALS-pat183+SRSF1-RNAi: 480/420/380/399, C9-ALS-pat201 : 201/243/261/224, C9-ALS-pat201+SRSF1-RNAi: 486/463/444/485). (E) Quantification of nuclear and cytoplasmic sense RNA foci in SRSF1- RNAi-transduced iAstrocytes (mean ± SEM; two-way ANOVA, NS: non-significant, **: p<0.01 , ***: pO.005, ****: p<0.001 ; N (cells with 1-5 RNA foci) = C9-ALS-pat78+MOI0: 21 , C9-ALS-pat78+MOI5: 20, C9-ALS-pat78+MOI7: 22, C9-ALS-pat183+MOI0: 21 , C9-ALS- pat183+MOI5: 24, C9-ALS-pat183+MOI7: 22, C9-ALS-pat201 +MOI0: 24, C9-ALS- pat201 +MOI5: 23, C9-ALS-pat201 +MOI7: 22). >95% of cells with RNA foci presented a total of 5 or fewer foci;

Figure 1 1 : SRSF1 -depletion and blocked SRSF1 AP/NXF1 interaction both inhibit the nuclear export of C90RF72-repeat RAN-translated transcripts and DPR production. (A) Western blots from N2A cells co-transfected with either a Ctrl or SRSF1-RNA\ vector and control backbone plasmid (no DPR Ctrl) or the same plasmid expressing 38 uninterrupted G4C2-sense-repeats (G4C2x38) or 39 uninterrupted C4G2-antisense-repeats (C4G2x39). Sense/antisense poly-Gly-Pro and sense poly-Gly-Ala DPR proteins are produced by internal repeat RAN translation in the absence of an initiating start codon (fig. 4) (nucleotide sequences in Fig. 4e. A hexanucleotide-repeat specificity control was provided by co- transfection of plasmids expressing poly-Gly-Ala (GAx36) or poly-Gly-Pro (GPx36) independently of the G4C2/C4G2-repeat sequences nucleotide sequences in Fig. 26) and either Ctrl or SRSF 7-RNAi vectors. (B) N2A cells co-transfected with G4C2x38 and either Ctrl or SRSF 7-RNAi plasmids (left part) and either FLAG-tagged SRSF1 aa11-196 wild type (SRSF1) or SRSF1-m4 (right part) were subjected to cellular fractionation using hypotonic lysis to yield cytoplasmic fractions (fig. 8e). Cytoplasmic and whole-cell (total) G4C2-repeat sense transcript levels were normalized to U1 snRNA levels in triplicate experiments prior to plotting as a ratio to account for potential changes in mRNA transcription/stability (mean ± SEM; one-way ANOVA, NS: non-significant, **: p<0.01 ; N (qRT-PCR reactions) = 6). (C) Western blots from N2A cells co-transfected with either G4C2x38 or C4G2x39 plasmids and control (FLAG Ctrl) or FLAG-tagged SRSF1 aa11-196 wild type (SRSF1), SRSF1-m2 or SRSF1-m4. (D) MTT cell proliferation assay performed on N2A cells transfected with either G4C2x38 or C4G2x39 plasmids and Ctrl-RNAi, SRSF 7-RNAi, FLAG Ctrl, SRSF1 or SRSF1- m4 in triplicate experiments (mean ± SEM; one-way ANOVA, NS: non-significant, **: p<0.01 , ***: p<0.005, ****: p<0.001 ; N (OD 650 values) = 12);

Figure 12: DNA sequences used to generate recombinant gene therapy vectors; A Human sequences of SRSF1 ; A1) NCBI accession number NM_006924.4 (mRNA sequence, transcript variant 1) - Coding sequence CDS: 210-956; A2) Amino acid sequence encoded by NM_006924.4; A3) NCBI accession number NM_001078166.1 (mRNA sequence, transcript variant 2) - Coding sequence CDS: 210-815; A4) Amino acid sequence encoded by NM_001078166.1 ; A5) NCBI accession number NR_034041.1 (mRNA sequence, transcript variant 3) - Considered non-coding as potential substrate for non-sense mediated decay; B) Mouse sequences of SRSF1; B1)_NCB\ accession number BC046773.1 (mRNA sequence) - Coding sequence CDS: 99-845; B2) Amino acid sequence encoded by BC046773.1 :B3) NCBI accession number NM_173374.4 (mRNA sequence, transcript variant 1) - Coding sequence CDS: 467-1213. Encodes the same sequence of aminoacids as BC046773.1 ; B4) Amino acid sequence encoded by NM_173374.4: B5) NCBI accession number NM_001078167.2 (mRNA sequence, transcript variant 2) - Coding sequence CDS: 467-1072; B6) Amino acid sequence encoded by NM_001078167.2;

Figure 13 Expression or cell-permeable fusion of SRSF1 sequences encoding a dominant negative mutant which does not interact with NXF1. Use of any sequences encompassing the expression of human SRSF1 or orthologous sequences encoding the two RNA Recognition Motifs (amino acid 1 1-196) and mutations of arginines 90, 93, 117, 1 18 in the linker region between the two 2 RNA Recognition Motifs. These correspond to the following amino acids for human SRSF1 including mutations of arginines 90, 93, 1 17, 118 into alanine (labelled in bold): Other aminoacids can be used for substitution eg glycine, valine, etc .A) SRSF1 amino acids 1-196 R90, 93, 117, 11 , B) SRSF1 amino acids 11-196 R90,93, 117, 118A;C; SRSF1 amino acids 1-248 R90,93,1 17, 118A;

Figure 14 Expression or cell-permeable fusion of SRSF1 sequences encoding antagonistic peptides which interact with NXF1. Use of any orthologous SRSF1 sequences encoding peptides which bind to NXF1. These encompass amino acids 89-120 of human SRSF1 ;

Figure 15 Expression of SRPK1 amino-acids 1-655 to maintain phosphorylation state of SRSF1 for inhibiting interaction of SRSF1 with NXF1 A) NCBI accession number NM_003137.4 (mRNA sequence, transcript variant 1) - Coding sequence CDS: 125-2092; B) Amino acids sequence encoded by NM_003137.4;

Figure 16a nucleotide sequence of CLK1 or Clk/Sty (Homo sapiens CDC like kinase 1 (CLK1), transcript variant 1 , mRNA) - NCBI Reference Sequence: NM_004071.3; Figure 16b amino acid sequence of CLK1 ;

Figure 17a nucleotide sequence of PP1 (Human protein phosphatase-1 catalytic subunit mRNA, complete cds) - GenBank: M63960.1 ; amino acid sequence of human PP1 ; Figure 17b is the amino acid sequence of PP1 ; Figure 18 Promoters sequences used to generate recombinant gene therapy vectors; A) 6.1/ CMV early enhancer / hybrid Chicken b-actin promoter (CAG); B) CMV promoter; C) CBA promoter (Chicken b-actin) D) Neuronal specific promoter (Synapsin 1); E) Glial specific promoter (GFAP); F) H1 promoter; G) U6 promoter; and h)H1_pLVTHM promoter, j)H1 promoter in scAAV backbone, k) U6 promoter, I) Chicken^-actin promoter (CBA), m) CAG promoter (hybrid chicken b-actin promoter), n) Mouse PGK_promoter from LV-LacZ, o) Elongation factor short 1a (EFS1a), p) CMV enhancer or q) CMV promoter;

Figure 19 Adeno-associated virus serotype 1 , 2, 5, 6, 8, 9, 10 A) Self-complementary AAV- CMV-GFP; B) pAAV2/9 - For construction of scAAV9;

Figure 20 Purified ALYREF and SRSF1 proteins directly interact with hexanucleotide repeat sense and antisense RNA. (a, b) Protein:RNA UV crosslinking assays using purified recombinant proteins and 32P-end-radiolabeled G4C2x5 (a) and C4G2x5 (b) repeat RNA probes. Proteins are visualised on SDS-PAGE stained with Coomassie blue (left panels) and covalently linked RNA:protein complexes by autoradiography on Phospholmages (Right panels). UV exposure is indicated by +. (c) Fluorescence confocal microscopy images show co-localization of SRSF1 (labeled in green by immunofluorescence) and sense RNA foci (labeled in red using Fluorescence In Situ Hybridization) in motor neurons from post-mortem spinal cord tissues of two human C90RF72-ALS cases. Nuclei are stained in blue by DAPI. Scale bar: 3 μηι;

Figure 21 Generation of neuronal cell models recapitulating the RAN-dependent translation of sense and antisense DPRs. (a) diagrammatic representations of constructs, (b) Sense G4C2 RNA foci stained with Cy3-labelled antisense C4G2 probe. DAPI was used to stain nuclei of neuronal N2A cells in blue. Scale bar: 5 μηι. (c) Western blots from N2A cells transfected with a control backbone plasmid (no DPR Ctrl) or the same plasmid expressing either 15 uninterrupted G4C2-sense repeat (G4C2x15), 38 uninterrupted G4C2-sense- repeats (G4C2x38), 15 uninterrupted C4G2-antisense repeat (C4G2x15) or 39 uninterrupted C4G2-antisense-repeats (C4G2x39). Membranes were probed with antibodies against poly- Gly-Pro DPRs, poly-Gly-Ala DPRs or loading control D-tubulin. (d) MTT cell proliferation assay performed on N2A cells transfected with either a control backbone plasmid (no DPR Ctrl) or the same plasmid expressing various length of uninterrupted hexanucleotide repeat transcripts in triplicate experiments (mean ± SEM; one-way ANOVA; N (OD650 values) = 12);

Figure 22 Depleting SRSF1 or inhibiting its repeat-RNA sequestration and interaction with NXF1 inhibit the production of DPRs in primary neurons, (a) Immunofluorescence microscopy of cultured rat cortical neurons. DPRs were detected in the red channel using anti-V5 and anti-mouse ALEXA594 antibodies. The Ctrl-RNAi and SRSF1-RNAi constructs co-express GFP while the FLAG-tagged SRSF1 proteins were stained using an anti-FLAG antibody conjugated to FITC allowing detection and quantification of transfected neurons in the green channel. Scale bar: 5 μηι. (b, c) Statistical assessment of the cortical neuron counts was performed from approximately 100 transfected neurons for each group (Fisher's exact test; N (transfected neurons) = Ctrl-RNAi: 95, SRSF1-RNAi: 1 12, SRSF1 : 106, SRSF1-m4: 121);

Figure 23 Depletion of SRSF1 specifically inhibits the nuclear export of C90RF72 repeat transcripts retaining hexanucleotide repeat expansions in intron-1 in patient-derived neurons, (a) The axonal length of patient-derived iMNs treated or not with SRSF1-RNAi was assessed by high content imaging in triplicate experiments (mean ± SEM; one-way ANOVA, NS: non- significant; N (average axon length/well) = 9). (b) The cell body area of patient-derived iMNs treated or not with SRSF1-RNAi was assessed by high content imaging in triplicate experiments (mean ± SEM; one-way ANOVA; N (average cell body area/well) = 9). (c) The survival of patient-derived Ctrl or C90RF72-ALS iMNs treated or not with SRSF1-RNAi was quantified in co-cultures with patient-derived Ctrl or C90RF72-ALS iAstrocytes in six replicate experiments at day 4 (mean ± SEM; one-way ANOVA; N (iMNs) = Ctrl-pat209: 142/165/174/1 17/122/168, C9-ALS-pat78: 77/65/41/68/71/70; C9-ALS-pat78+SRSF1-RNAi: 106/84/81/97/1 13/83; C9-ALS-pat201 : 68/69/68/62/66/74; C9-ALS-pat201+SRSF1-RNAi: 131/104/96/82/11 1/104). (d) Western blots of iNeurons differentiated from control (Ctrl-pat 154, Ctrl-pat155) and C90RF72-ALS (C9-ALS-pat78, C9-ALS-pat183) patients treated or not with LV-SRSF1-RNAi (MOI 0 or 5) were subjected to cellular fractionation using hypotonic lysis to yield cytoplasmic fractions. The chromatin remodeling SSRP1 factor is used to check for potential nuclear contamination in cytoplasmic fractions. Depletion of actin in nuclear fractions was used to check for quality of the nuclear fractions. The efficacy of SRSF1-RNAi was also validated by qRT-PCR achieving approximately 80% SRSF1 mRNA knockdown ( Fig. 29). (e) Total, nuclear and cytoplasmic levels of intronl -spliced C90RF72 transcripts (as measured by the exon1-exon3 junction) were quantified in duplicate experiments by qRT-PCR following normalization to U1 snRNA levels and to 100% in control patients at MOI0 (mean ± SEM; one-way ANOVA, NS: non-significant; N (qRT-PCR reactions) = 8). (f) Intronl-spliced C90RF72 transcripts levels normalized to U1 snRNA levels (panel d) were plotted as a ratio SRSF1-RNAi MOI5 over MOI0 to evaluate the specific effect due to the SRSF1 -RNAi (mean ± SEM; one-way ANOVA; N (qRT-PCR reactions) = 8). (g) Total, nuclear and cytoplasmic levels of unspliced C90RF72 transcripts retaining intronl (as measured by the exonl -intronl junction) were quantified in duplicate experiments by qRT-PCR following normalization to U1 snRNA levels and to 100% for control patients at MOI0 (mean ± SEM; one-way ANOVA, NS: non-significant; N (qRT-PCR reactions) = 8). (h) Unspliced C90RF72 transcripts retaining intronl levels normalized to U 1 snRNA levels (panel f) were plotted as a ratio SRSFI -RNAi MOI5 over MOI0 to evaluate the specific effect due to the SRSFI -RNAi (mean ± SEM; one-way ANOVA; N (qRT-PCR reactions) = 8);

Figure 24 Model for the nuclear export of pathological C90RF72 hexanucleotide repeat transcripts and therapeutic manipulation, (a) The nuclear export of sense and antisense C90RF72 transcripts retaining expanded hexanucleotide repeats in introni specifically depend on the sequestration of SRSF1 and its interaction with the nuclear export receptor NXF1. In contrast, the nuclear export of introni -spliced C90RF72 transcripts required for the production of the C90RF72 protein does not involve the interaction of SRSF1 with NXF1 however the nuclear export adaptor(s) (NEA) remain to be identified, (b) The depletion of SRSF1 specifically inhibits the nuclear export of C90RF72 transcripts retaining expanded hexanucleotide repeats in introni , likely due to a reduction in the sequestration of endogenous SRSF1 onto the C90RF72 hexanucleotide repeats and failure to abnormally remodel NXF1 in a high RNA-binding mode, while it does not affect the expression levels and splicing/retention of introni Moreover, the depletion of SRSF1 does not affect the expression levels, the splicing of introni or the nuclear export of wild type introni -spliced C90RF72 transcripts required for the production of the C90RF72 protein, (c) Over- expression of the SRSF1-m4 protein, which fails to interact efficiently with NXF1 , competes endogenous SRSF1 for sequestration onto hexanucleotide repeats preventing in turn interactions with NXF1 and nuclear export of C90RF72 repeat transcripts;

Figure 25 Partial depletion of SRSF1 leads to prominent reduction of sense and antisense poly-GP DPRs in Drosophila and motor neurons derived from C90RF72-ALS patients, (a) Total protein extracts from G4C2x3 + Ctrl-RNAi, G4C2x36 + Ctrl RNAi and G4C2x36 + SRSF1-RNAi Drosophila larvae were analysed by dot blots using poly-GP and loading control a-tubulin antibodies, (b) Total protein extracts from HEK cells transfected with either Ctrl or GP36 plasmids and patient-derived iMNs transduced (MOI=5) or not (MOI=0) with LV-SRSF1-RNAi viruses are analysed by dot blots using poly-GP DPRs and loading control alpha-tubulin antibodies;

Figure 26 Generation of synthetic constructs expressing DPRs independently of RAN- translation and G4C2 repeat hexanucleotides. (a) Nucleotide sequence encoding poly-Gly- Pro x36 DPRs. (b) Nucleotide sequence encoding poly-Gly-Ala x36 DPRs. The DPR sequence is highligted in blue. The ATG start codon is highlighted in red while the TAA stop codon is highlighted in bold. A V5-tag is also present and highlighted in green. Figure 27 Partial loss of sbr/NXF1 restore locomotor deficits in G4C2x36 expressing flies. Neuronal expression of G4C2x36 causes larval crawling (a) and adult climbing (b) deficits that are both restored by sbr depletion (mean ± 95% CI normalized to Control; N (larvae) = 10; N (adults) = Control (GAL4/luciferase-RNAi): 105, G4C2x36 + Ctrl-RNAi: 70, G4C2x36 + sbr-RNAi: 72);

Figure 28 Depleting SRSF1 or inhibiting its sequestration and interaction with NXF1 alter the cytoplasmic levels of hexanucleotide repeat transcripts but not their total levels.

(a) N2A cells co-transfected with G4C2x38 and either Ctrl or SRSF1-RNAi plasmids (left part) and either FLAG-tagged SRSF1 aa1 1-196 wild type (SRSF1) or SRSF1-m4 (right part) were subjected to cellular fractionation using hypotonic lysis to yield cytoplasmic fractions (Fig. 6e). Total and cytoplasmic G4C2-repeat sense transcript levels were normalized to U1 snRNA levels in triplicate experiments (mean ± SEM; one-way ANOVA; N (qRT-PCR reactions) = 6). (b) Drosophila expressing G4C2x36 and either control (Ctrl)-RNAi or SRSF1-RNAL Whole flies were subjected to cellular fractionation using hypotonic lysis to yield cytoplasmic fractions (Fig. 6g). Total cytoplasmic G4C2-repeat sense transcript levels were normalized to Tub84b levels in triplicate experiments (mean ± SEM; paired t-test; N (qRT-PCR reactions) = 3);

Figure 29 Transcript sequence. Boxes represent the Hindi 11 (AAGCTT) and Xhol (CTCGAG) cloning sites used to clone the G4C2x38 annealed oligonucleotides (Fig. 3). A synthetic construct encoding for the 3x V5 tags (sequences highlighted in orange, green and violet) with 3 stop codons (TAA, underlined/bold) were cloned in a second step using the Notl (GCGGCCGC) and Xbal (TCTAGA) sites. The RNA transcript generated from this construct is highlighted in blue (flanking regions) and red (38 G4C2-sense repeats) and orange, green and violet (3x V5 tags). Sequences highligted in black/underlined correspond to the 3'end of the promoter sequence and in black/italics to the start of the terminator sequence. Note the absence of initiating codons (ATG) in the transcript generated from the RAN-dependent DPR-expression construct. Sequencing and size analysis further showed that the number of repeats remained stable over multiple rounds of transformation and replication in NEB® 10- beta E. coli (New England Biolabs);

Figure 30 iNPC-differentiation of neurons derived from patient fibroblasts. Tuj1 immunofluorescence microscopy was performed on neurons differentiated from induced- Neural Progenitor Cells (iNPCs) derived from control (Ctrl-pat154) or C90RF72-ALS (C9- ALS-pat78) patient fibroblasts using the red channel. DAPI was used to stain nuclei in blue; Figure 31 Evaluating the efficiency of lentiviral-mediated SRSFI-RNAi depletion in iNeurons derived from control and C90RF72-ALS patients. SRSF1 transcript levels were quantified in transfected HEK cells and iAstrocytes transduced with increased MOI doses of LV-SRSF1- RNAi. snRNA U1 transcript levels were used for normalization in two comtrol (pat154, pat 155) or C9-ALS (pat78, pat183) cell lines in duplicates (mean ± SEM; two-way ANOVA; N (qPCR reactions) = 8); and

Figure 32 An SRSF1 inhibitory cell permeable peptide inhibits the production of DPRs expressed in a disease relevant RAN-dependent manner in human HEK cells transfected with a G4C2x38 repeat construct that expresses 3xV5 tags in all frames. Sequence of the peptide: PRSGRGTGRGGGGGGGGGAPRGRYGPPSRRSE GG GKPIPNPLLGLDST GG YGRKKRRQRRR.

Table 1

MOI=0 M0I=5 M0I=7

RNA Foci/cell iC9 Pat. 78 C9 Pat. 183 ! C9 Pat. 201 C9 Pat. 78 C9 Pat. 183 ! C9 Pat. 201 C9 Pat. 78 C9 Pat. 183 ! C9 Pat. 201

Nuclear RNA foci 0.86 0.95: 0.88! 1.20 1.08: 1.48! 1.82 1.68: 1.95!

Cyto RNA foci 0.71 0.71 0.58: 0.35 0.50 0.35! 0.23 0.27 0.23!

SEM

Nuclear RNA foci 0.14 0.18: 0.15 0.17 0.18: 0.18! 0.20 0.26! 0.23!

Cyto RNA foci 0.16 0.12: 0.13 0.13 0.13: 0.10! 0.09 0.10! 0.09!

Table 2 Examples of human SRSF1 miRNA sequences

These involve expression of pre-miRNA cassettes which are processed in the cell using the physiological RNAi machinery (Dicer cleavage to generate -22 nucleotides mature miRNA and incorporation into RISC (RNA-induced silencing complex).

10 sequences of human SRSF1 mRNA targeted by mature miRNA:

10 6 3'UTR 1436 AATGGTATGACTCCAAGTGCT

11 7 3'UTR 1695 GCTAATTTGTCACAGTGCTTA

12 8 3'UTR 1725 GTTAATGTGTGACCTGCTGTT

13 9 3'UTR 1804 ACTGCTAAATCTGCATGTCCT

14 10 3'UTR 1834 TGATAGAGCGTTGCTATTTCA

RRM2: RNA recognition motif 2; RS: Arginine-Serine rich; 3'UTR: 3' untranslated region

Table 3 Examples of mouse SRSF1 miRNA sequences

10 sequences of mouse SRSF1 mRNA targeted by mature miRNA:

RRM2: RNA recognition motif 2; RS: Arginine-Serine rich; 3'UTR: 3' untranslated region

Table 4: Number and cellular distribution of G4C2 RNA foci in iNPCs-derived astrocytes treated with increasing MOI of LV-SRSF1-RNAi

Table 5 I List and characteristics of patient-derived cells used in this study.

Patien t Ethnicity Gender Cell type Age at biopsy

Sampk collection

78 Caucasian Male » ) 0 11 72 66

183 Caucasian Male Ci ) 0 11 72 49

201 Caucasian Female CS !i 11 72 66

154 Caucasian Female Control 55

155 Caucasian Male Control 40

209 Caucasian Female Control 69

Table 6 Examples of shRNAs that target human SRSF1 (NM 006924.4)

Materials and Methods

Drosophila inducible short-hairpin RNAi lines. Lines (UAS-RNAi) were obtained from the

Vienna Drosophila Resource Centre: http://stockcenter.vdrc.at/contro!/main

SRSF1 (SF2/ASF) - RNAi lines:

V27775: FlyBase ID = FBst04571 17

V27776: FlyBase ID = FBst04571 18

Independent insertion lines, both lines carry the following inverted repeat sequence: (SEQ ID NO 57)

ATGCCGACGA TGCGGTGAAG GCGCGCGACG GCTACGACTA CGATGGGTAT CGTCTGCGCG TGGAGTTCCC GCGGGGCGGT GGTCCTGGAA GCTACCGCGG CGGCAACCGC AATGACCGAA GCCGCGACGG TGGGGGACGG ATGGGCGGAC GCGGACCGCC AGCCAAGCGC TCGCAGTACC GCGTCATGGT TACTGGACTG CCCGCCTCCG GATCGTGGCA AGATCTCAAG GATCACATGC GCGAGGCCGG CGACGTCTGC TTCGCGGACA CTTACAAGGA TGGTTCCGGC GTCGTTGAGT TCCTGCGCCA CGAGGACATG AAGTACGCAA TCAAAAAATT GGACGACTCT CGCTTCCGA ALYREF (Ref1) - RNAi lines:

v12301 (GD): FlyBase ID = FBst0450381 - the line carry the following inverted repeat sequence (SEQ ID NO 58)

GGTCCGATAA AGAAGGCGGC AGTGCACTAC GATCGCTCCG GTCGCTCGTT

GGGCACCGCT GACGTGATTT TCGAACGTCG CGCCGACGCC TTGAAGGCCA TTAAACAGTA CCATGGCGTA CCTTTGGACG GACGCCCTAT GACCATTCAG

CTGGCCGTCT CAGACGTGGC CGTGTTGACC CGTCCCGTAG CCGCCACCGA

TGTCAAGCGT CGCGTGGGTG GTACTGCACC AACTTCATTC AAGCGTGGTG

GTGGCCAAGC TGGTGGCACG GCGCGTCGCG GCTTCAAACG TCCGGTCGGT

GGCAAGCCGG CGGCAGGCGG CCAGCGACGG GAGCGCAAGG CCCCGCCCAC TGCTGAGGAG CTGGACGCCG AACTGGACTC A v104471 (KK): FlyBase ID = FBst0476329 - - the line carry the following inverted repeat sequence (SEQ ID NO 59)

GTCGAACTTG ATAAAGCGCA TTTCTAAATA CAATAAATAC AGCATCAAAT GTATTTCAGT TATCTTAACA TCCGCCGCAT TGGCAAAACT AACAATTAAT

GGATAAATGC GCAAGTGGTT GATTGATTTG ATGTCCGATG CTTTCAAAGA

TCTGCTCCTG GGCGCGGCGT TGTCGATGCG TTTGCATTTA TGTACCATGC

GGGGGGTGTC CATATGGTAG GCTTAAAACT ATAGATTGGG CTGCTCTTCT

ATTCTTGTTA GACTAATTCA GACTATTCAC TATTTAGATC TTCATGTCGT TGATGTATGA GTCCAGTTCG GCGT

AAV vectors

Suitable AAV vectors are self-complementary scAAV and AAV vectors of varied serotypes (for examples, serotypes 2, 5, 6, 9 and AAV9 derivatives such as AAV-PHP.A and AAV- PHP. Further alternatives are AAV9: (Valori et al. Sci TransI Med. 2010; 2:35ra42) and AAV- PHP-A and B) Deverman et al. Nature Biotech 2016: 34:204-9) (see also Figure 19)

Promoter elements for driving expression of the SRSF1 antagonists:

SRSF1-RNAi cassettes would be driven using H 1 or U6 promoters; Expression of SRSF1 proteins or peptides using CBA, CAG, PGK, EFS1 a or CMV; further examples are listed in figure 18. Drosophila husbandry and locomotor assays

Drosophila were raised under standard conditions on food consisting of agar, cornmeal, molasses and yeast unless otherwise stated. All C9orf72 related transgenic lines (14) were a gift from Adrian Isaacs and Linda Partridge (University College London). GMR-GAL4 (#1104), D42-GAL4 (#8816), nSyb-GAL4 (#51635) da-GAL4, UAS-sbr-RNAi [P{TRiP.HM05135}attP2] and UAS-/tyc/7erase-RNAi (#31603), used as the control RNAi, were obtained from the Bloomington Drosophila Stock Centre (Bloomington, IN).

Eye phenotypes were analyzed by induction of transgene expression by GMR-GAL4 raised at 25°C. For larval crawling assays, transgenes were expressed by nSyb-GAL4 and animals were grown at 29°C. Wandering third instar larvae were collected from vials, briefly rinsed with distilled water, and placed in petri dishes with a 1 % agarose matrix. Larvae were observed directly for 2 minutes and the number of peristaltic waves recorded. Climbing assays were performed as previously described (32) with transgenic expression induced by D42-GAL4. Crosses were started at 25°C for 3 days and transferred to 29°C for a further 7 days. Adult flies were tested 1-3 days after eclosion.

For all the experiments only male flies were used. Drosophila were housed in 12:12 light:dark cycle.

Drosophila light microscopy imaging and scanning electron microscopy

For light microscopy of Drosophila eyes, stacks of images were collected on a Nikon motorized SMZ stereo zoom microscope fitted with 1x Apo lens. Extended focus images were then generated using Nikon Elements software. Scanning electron microscopy (SEM) was performed according to a standard protocol (33) and images were captured using a Philips XL-20 SEM microscope. All animals of a given genotype displayed essentially identical phenotypes and randomly selected representative images are shown. Plasmids

The FLAG-tagged SRSF1/SF2/ASF plasmids were generated as in reference (23) by cloning of a full-length human SRSF1/SF2/ASF PCR fragment into p3XFLAG-myc-CMV26 (Sigma). miRNA oligonucleotide sequences were designed using the "miR RNAi" Block-IT RNAi designer tool (ThermoFisher) at: https://maidesigner.thermofisher.com/rnaiexpress/

Two miRNA hairpins were designed against human SRSF1 (GenBank: NM_006924.4, mRNA) and mouse SRSF1 (GenBank: BC046773.1 , mRNA NCBI Reference Sequence: NM_173374.4,) and rat SRSF1 (NCBI Reference Sequence: NM_001109552.2, mRNA). The SRSF1 sequence targeted by miRNA hairpin 1 is identical in human, mouse and rat SRSF1. The bold regions in sequences below represent the mature miR RNAi sequences which targets the complementary sense sequences on SRSF1 (italics):

Targeted human and mouse and rat SRSF1 miR1 sequence (TTAAAGTTGATGGGCCCAGAA; SEQ ID NO 6) starts at 784 nt (NCBI RefSeq NM_006924.4 - RRM2 region), and 673 nt respectively (RRM2 region), 1 ,041 nt (NCBI RefSeq NM_ 173374.4 - RRM2 region) and 699 nt (NCBI RefSeq: NM_001 109552.2 - RRM2 region):

human/mouse SRSF1 - miR1 -Top strand (SEQ ID NO 61)

5'- TGCTGTTCTGGGCCCATCAACTTTAAGTTTTGGCCACTGACTGACTTAAAG

TTTGGGCCCAGAA -3'

human/mouse/rat SRSF1 - miR1 - Bottom strand: (SEQ ID NO 62)

5'- CCTGTTCTGGGCCCAAACTTTAAGTCAGTCAGTGGCCAAAACTTAA/AGTTG

ATGGGCCCAGAAC -3'

Targeted human SRSF1 miR2 sequence (AATGGTATGACTCCAAGTGCT) (SEQ ID NO 10) starts at 1436 nt (3'UTR region):

human SRSF1 - miR2 - Top strand: (SEQ ID NO 63)

5'- TGCTGAGCACTTGGAGTCATACCATTGTTTTGGCCACTGACTGACAATGGT

ATCTCCAAGTGCT -3'

human SRSF1 - miR2 - Bottom strand: (SEQ ID NO 64)

5'- CCTGAGCACTTGGAGATACCATTGTCAGTCAGTGGCCAAAACAATGGLAT

GACTCCAAGTGCTC -3'

Targeted mouse SRSF1 miR2 sequence (AATGTCTATTCTGCTCTGGTT) (SEQ ID NO 65) starts at 1 , 105 nt (3'UTR region):

Targeted mouse SRSF1 miR2 sequence (AATGTCTATTCTGCTCTGGTT) (SEQ ID NO 66) starts at 1 ,473 nt (NCBI RefSeq NM_173374.4 - 3'UTR region):

mouse SRSF1 - miR2 - Top strand: (SEQ ID NO 67)

5'- TGCTGAACCAGAGCAGAATAGACATTGTTTTGGCCACTGACTGACAATGT

CTACTGCTCTGGTT -3'

mouse SRSF1 - miR2 - Bottom strand: (SEQ ID NO 68)

5'- CCTGAACCAGAGCAGTAGACATTGTCAGTCAGTGGCCAAAAC/W TGTCTA

TTCTGCTCTGGTTC -3' Synthesised oligonucleotides (Sigma) were annealed and ligated into pcDNA6.2 GW/EmGFP using the BLOCK-iT Pol I I miR RNAi Expression Vector Kit with EmGFP (ThermoFisher, catalogue number K4936-00). Furthermore, the pre-miR2 RNAi cassettes were each chained by subcloning BamHI/Xhol-cut fragments into the Bglll and Xhol sites of pcDNA6.2 GW/EmGFP-SRSF7 miR1. The PCR fragments encompassing EmGFP and the chained SRSF1 pre-miRNA cassette were additionally cloned into the lentiviral expression plasmid SI N-PGK-cPPT-GDNF-WHV(9)(5) using the restriction enzymes Spel and Xhol . Restriction of the SI N-PGK-cPPT-GDNF-WHV(9) by Spel/Xhol allows removal of the GDN F insert and cloning of the human or mouse EmGFP-SRSF1 RNAi cassette (fig. 1 and 5).

Uninterrupted hexanucleotide sense GGGGCCx38 (SEQ I D NO 69) and antisense CCCCGGx39 (SEQ ID NO 70) C90RF72 repeats were built using the synthetic oligonucleotides 5'-(GGGGCC)i 5 -3' and 5'-CCCC-(GGCCCC)i 4 -GG-3' (SEQ I D NO 71 ). Oligonucleotides were annealed by heating to 99°C for 30 minutes and cooling 0.5°C/min to ambient with incubation at 70°C for 10 minutes. Oligonucleotides were phosphorylated with T4-Polynucleotide Kinase (New England Biolabs), ligated using T4 DNA Ligase (ThernoFisher) and treated with Mung Bean nuclease (New England Biolabs) for blunt ligation. Oligomeric forms of annealed oligonucleotides were confirmed following analysis on agarose gel (fig. 4). The band corresponding to trimeric oligonucleotides was then excised, gel purified and ligated into pcDNA3.1 (Invitrogen) with blunted EcoRI ends to allow cloning in both sense and antisense orientation. Sequences are presented in fig. 4.

Synthetic sequences encoding poly-Gly-Pro and poly-Gly-Ala x36 DPRs independently of G4C2 repeats were first cloned into pcDNA3.1 (Invitrogen) using the EcoRI and Notl sites. Synthetic sequences encoding poly-Gly-Pro and poly-Gly-Ala x36 were then subcloned using BamHI/Notl into pCI-neo-V5-N using Bcll/Notl. Bell restriction site was previously introduced into pCI-neo-V5-N by site directed mutagenesis using forward actctagaggtaccacgtgatcattctcgagggtgctatccaggc (SEQ I D NO 72) and reverse gcctggatagcaccctcgagaatgatcacgtggtacctctagagt (SEQ I D NO: 73) primers (QuikChange Lightning Site-Directed Mutagenesis Kit, Agilent).

Lentivirus production

The miRNA construct was sub-cloned into a self-inactivating lentiviral (SI N-W-PGK) vector using standard cloning methods. Lentiviruses were propagated in HEK293T cells using calcium phosphate transfection (34). 13 μg pCMVAR8.92, 3.75 μg pM2G, 3 g pRSV and 13ig SI N-CMV-miRNA were transfected into HEK293T cells. Twenty 10 cm dishes seeded with 3x10 6 HEK293T cells/dish were each transfected with 13 μg pCMVAR8.92, 3.75 \ g pM2G, 3 pRSV and 13 μg SIN-CMV-miRNA using calcium phosphate transfection Cells were allowed to produce virus for 72 hours, then the supernatant was collected, filtered using a 0.45 μηι filter and centrifuged at 24,000 rpm for 90 minutes at 4 ° C. The supernatant was discarded and the viral pellet was resuspended in 1 % BSA in PBS and stored at -80'C. The biological titre of the viral vector was determined by transducing HeLa cells with 10 "2 , 10 " 3 and 10 "4 dilutions of the vector. 72 hours post transduction cells were fixed in 4% paraformaldehyde, washed in PBS, and the percentage of GFP positive cells was measured with a Fluorescent-Activated cell sorter (FACS, LSRII). The biological titer is expressed as the number of transducing units per ml_ (TU/ml) and can be calculated with the following formula: Vector titer= [(% positive cells x number of cells during transduction) x dilution factor x 2] TU/ml.

Tissue culture and transfection or transduction

HEK293T cells were cultured in Dulbecco's Modified Eagle Medium (Sigma) supplemented with 10% fetal bovine serum [FBS] (Gibed) and 5U/ml Penstrep (Lonza). Cells were maintained in a 37°C incubator with 5% C0 2 . For qRT-PCR analysis 50,000 HEK cells were split into each well of 24-well plates and transfected for 72h with 700 ng pcDNA6.2- GW/EmGFP-Control or human SRSF1 miR RNAi constructs using 3.5 μg PEI/ml media and one tenth medium volume OptiMEM (ThermoFisher). For Western blot analysis, HEK cells were transfected for 72h with 650 ng pcDNA6.2-GW/EmGFP-Control miRNA, pCDNA6.2- GW/EmGFP-human SRSF1 miR1 +2 RNAi or LV-EmGFP-human SRSF1 miR1 +2 RNAi constructs and 50 ng p3XFLAG/human-SRSF1 using 3.5 μg PEI/ml media and one tenth medium volume of OptiMEM. Neuro-2a (ATCC) cells were cultured in Dulbecco's Modified Eagle Medium (Sigma) supplemented with 10% fetal bovine serum [FBS] (Gibco), 5U/ml Penstrep (Lonza) and 5 mM sodium pyruvate. Cells were maintained in a 37°C incubator with 5% C0 2 . 75,000 Neuro-2a cells were split into each well of 24-well plates and transfected for 72h with 500 ng pcDNA6.2-GW/EmGFP-Control miR RNAi (ThermoFisher) or pcDNA6.2-GW/EmGFP- mouse SRSF1 miR1+2 RNAi and 200 ng pcDNA 3.1/RAN-G4C2x38-sense or RAN- C4G2x39-antisense using 3 μg PEI / 1 μg DNA and one tenth medium volume OptiMEM (ThermoFisher).

Hb9GFP mouse stem cells were cultured as previously described (35) and differentiated into motor neurons with 2 μΜ retinoic acid (Sigma) and 1 μΜ Smoothened Agonist (SAG) (Millipore) for 5 days. Embryoid bodies were then dissociated with papain and sorted using the FACSAria™ III (BD Biosciences). Cells were maintained in a 37°C incubator with 5% C0 2 .

Human patient-derived differentiated astrocytes (iAstrocytes) were differentiated from induced Neural Progenitor Cells (iNPCs) as previously described (29) and cultured in DMEM Glutamax (Gibco) with 10% FBS (Sigma) and 0.02% N2 (Invitrogen). Cells were maintained in a 37°C incubator with 5% C0 2 .

For Hb9GFP+ motor neuron and patient-derived iAstrocytres co-cultures, 20,000 induced neural progenitor cells (iNPCs) were plated in 6-well plates in astrocyte medium. The day after plating, iAstrocytes were transduced with adenovirus expressing red fluorescent protein (RFP) since the co-culture experiments were performed using GFP+ motor neurons from Hb9GFP+ mouse stem cells and with lentivirus co-expressing human SRSF 7-RNAi and GFP at an MOI of 5, 7 or 10. GFP expression in iAstrocytes was used to monitor transduction efficiency of the SRSF 7-RNAi. Cells were maintained in a 37°C incubator with 5% C0 2 .

For patient-derived cell cultures, informed consent was obtained from all subjects before sample collection (Study number STH16573, Research Ethics Committee reference 12/YH/0330). Human patient-derived differentiated astrocytes (iAstrocytes) were differentiated from induced Neural Progenitor Cells (iNPCs) as previously described52 and cultured in DMEM Glutamax (Gibco) with 10% FBS (Sigma) and 0.02% N2 (Invitrogen). Cells were maintained in a 37°C incubator with 5% C02.

Human patient and control-derived neurons (iNeurons) were differentiated from the previously established iNPCs. iNPCs were then differentiated into neurons using a modified version of the protocol described in reference 40 . Briefly, 30,000 iNPCs were plated in a 6-well plate coated with fibronectin (Millipore) and expanded to 70-80% confluence. Once they reached this confluence, iNPC medium was replaced with neuron differentiation medium (DMEM/F-12 with Glutamax supplemented with 1 %N2, 2%B27 (Gibco)). On day one of differentiation the cells were treated with 2.5 μΜ of DAPT (Tocris) to promote differentiation towards neuronal lineage. On day three the neuron differentiation medium is supplemented with 1 μΜ retinoic acid (Sigma), 0.5 μΜ Smoothened Agonist (SAG) (Millipore) and 2.5 μΜ Forskolin (Sigma) for 7 days. This protocol leads to typical yields of 70% μ-ΙΙΙ tubulin (Tuj1) positive cells. To obtain iMotor Neurons, iNeurons were re-plated on fibronectin and cultured in retinoic acid, SAG and Forskolin for 14 more days with addition of BDNF, CNTF and GDNF (all at 20 ng/ml) for the last 10 days of differentiation. For SRSF1 knockdown, cells were transduced with lentivirus expressing control GFP or human SRSF1-RNAi co-expressing GFP at an MOI of 5 at day 14 along with the HB9:RFP adenovirus.

Co-cultures of human patient-derived iAstrocytes and mouse or human motor neurons

For Hb9GFP motor neuron and patient-derived iAstrocyte co-cultures, 20,000 induced neural progenitor cells (iNPCs) were plated in 6-well plates in astrocyte medium. The day after plating, cells were transduced with lentivirus expressing control GFO or human SRSF 7-RNAi at an MOI of 5, 7 or 10. The human SRSF1-RNA\ virus also co-expressed GFP to allow evaluation of the transduction efficiency. Since the co-culture experiments were performed using GFP+ motor neurons from Hb9GFP+ mouse stem cells, on the same day cells were also transduced with an adenovirus expressing red fluorescent protein (RFP). Cells were maintained in a 37°C incubator with 5% C0 2 .

Seven days post-transduction with Ad-RFP and L\/-SRSF1-RNA\, iAstrocytes were plated at a density of 10,000 cell/well. The day after, Hb9GFP embryoid bodies were dissociated and sorted for GFP+ cells. 10,000 GFP+ motor neurons were plated onto the astrocytes in motor neuron medium consisting of DMEM/F12, 2% horse serum (Invitrogen), 2% N2, 2% B27 plus GDNF {Invitrogen; 10 ng/ml), BDNF {Invitrogen; 10 ng/ml), CNTF {Invitrogen; 10 ng/ml) and IGF-1 {Invitrogen; 10 ng/ml). Nine 10x images/well to cover the whole well surface were acquired daily for 3 days using the high content imaging system InCell 2000 {GE Healthcare), gathering data on neuronal cell size and number, axonal length and neurite branching. Data analysis was performed using the InCell Developer software. Data analysis was performed using the Columbus software {PerkinElmer). Data are presented for 3 days of co-culture. The programme designed for co-culture analysis only takes into account GFP+ cells with at least one projection to exclude counts of cell debris. For iMN on iAstrocyte cultures, iAstrocytes were plated in 384-well plates 24h before plating 1 ,000 FACS-sorted iMNs. Cultures were maintained for 4 days. Data are presented for 4 days of co-culture. Cytoplasmic fractionation

Cytoplasmic fractionation was performed 72h post-transfection. 300,000 cells from one well of a 6-well plate were removed from the plate using DEPC PBS and pelleted by centrifugation at 800xg for 5 minutes. Cell pellets were quickly washed with hypotonic lysis buffer (10 mM HEPES pH 7.9, 1.5 mM MgCI 2 , 10 mM KCI, 0.5 mM DTT). Cell pellets were then lysed in hypotonic lysis buffer containing 0.16 ΙΙ/μΙ_ Ribosafe RNase inhibitors (Bioline), 2 mM PMSF (Sigma) and SIGMAFAST™ Protease Inhibitor Cocktail tablets, EDTA free (Sigma) according to manufacturer's instructions. Cells were lysed gently using a cut P1000 tip ensuring no physical force is exerted on the cell pellet. The lysate then underwent differential centrifugation (1500g, 3 min, 4 °C then 3500g rpm, 8 min, 4 °C and then 17 OOOg, 1 min, 4 °C) transferring the supernatant to a fresh tube after each centrifugation. The supernatant was then added to PureZOL (Biorad) to extract RNA. The nuclear pellet obtained after centrifugation at 1500g for 3 min was lysed in Reporter lysis buffer (Promega) for 10 minutes on ice before centrifugation at 17000g, 5 min, 4 °C.

Total fractions were collected directly in Reporter lysis buffer (Promega) containing 0.16 U/μΙ- Ribosafe RNase inhibitors (Bioline), 2 mM PMSF (Sigma) and SIGMAFAST™ Protease Inhibitor Cocktail tablets, EDTA free (Sigma) according to manufacturer's instructions and lysed for 10 min on ice before centrifugation at 17 OOOg, 5 min, 4 °C. The supernatant was then added to PureZOL (Biorad) to extract RNA.

Equal volumes of total, nuclear and cytoplasmic lysates were resolved using SDS-PAGE, electroblotted onto nitrocellulose membranes and probed using anti-SSRP1 and anti-a- tubulin antibodies (see western blot analysis section below for details on antibodies).

Western blot analysis

HEK cells were transfected for 72h with 650 ng pcDNA6.2-GW/EmGFP-Control miRNA, pCDNA6.2-GW/EmGFP-human SRSF1 miR1 +2 RNAi or LV-EmGFP-human SRSF1 miR1+2 RNAi constructs and 50 ng p3XFLAG/human-SRSF1 using 3.5 μg PEI/ ml media and one tenth medium volume of OptiMEM. Neuro-2a cells were split into each well of 24- well plates (75,000 cells/well) and transfected for 72h with 350 ng pcDNA6.2-GW/EmGFP- Control miR RNAi (ThermoFisher), pcDNA6.2-GW/EmGFP-mouse SRSF1 miR1+2 RNAi, p3xFLAG, p3xFLAG/SRSF1 (1 1-196), p3xFLAG/SRSF1 (1 1-196)-m2 or p3xFLAG/SRSF1 (1 1-196)-m4 and 350 ng pcDNA 3.1/RAN-G4C2x38-sense or RAN-C4G2x39-antisense using 3 μg PEI / 1 μg DNA and one tenth medium volume OptiMEM.

Proteins were extracted from HEK or Neuro-2a cells 72 hours after transfection. Cells were briefly washed in ice-cold phosphate-buffered saline (PBS) and then scraped into ice-cold lysis buffer (50mM Hepes pH7.5, 150 mM NaCI, 10% glycerol, 0.5% triton X-100, 1 mM EDTA, 1 mM DTT, protease inhibitor cocktail (Sigma)). Cells were left to lyse on ice for 10 minutes followed by centrifugation at maximum speed at 4°C for five minutes. Protein extracts were quantified using Bradford Reagent (BioRAD), resolved by SDS-PAGE, electro blotted onto nitrocellulose membrane and probed using the relevant primary antibody. Human/mouse SRSF1/SF2 [1 : 1000 dilution] (Cell Signaling #8241) and poly-Gly-Pro [1 : 10,000 dilution] (kindly received from Prof Stuart Pickering Brown) primary antibodies were detected with horseradish peroxidase (HRP)-conjugated rabbit secondary antibody [1 :5000 dilution] (Promega), while a-tubulin [1 : 10000 dilution] {Sigma, clone DM1 A), FLAG [1 :2000 dilution] (Sigma F1804, clone M2), ALYREF [1 :2000 dilution] (Sigma A9979, clone 11 G5), SSRP1 [1 :500 dilution] (Abeam 26212, clone 10D7) and poly-Gly-Ala [1 :500 dilution] (kindly provided from Prof Dieter Edbauer) primary antibodies were detected using HRP- conjugated mouse secondary antibody [1 :5000 dilution] (Promega). For dot blot analysis, 50 μg total protein extracts prepared in ice-cold lysis buffer were loaded onto a nitrocellulose membrane using a microfiltration apparatus (Biorad) and analysed by western immunoblotting as previously described.

Cytoplasmic fractionation

Neuro-2a cells were split into each well of 6 well plates (2 x 10 6 cells/well) and transfected for 72 hours with 1 μg pcDNA6.2-GW/EmGFP-Control miR RNAi, pcDNA6.2-GW/EmGFP- mouse SRSF1 miR1+2 RNAi, p3xFLAG/SRSF1 (1 1-196), or p3xFLAG/SRSF1 (1 1-196)-m4 and 1 μg pcDNA 3.1/RAN-G4C2x38-sense or RAN-C4G2x39-antisense using 3 μg PEI / 1 μg DNA and one tenth medium volume OptiMEM. iNeurons were cultured in 6-well plates and transduced with 5MOI human L\/-SRSF1-RNA lentivirus for 5 days.

Cytoplasmic fractionation was performed as follows. Cells were removed from the plate using DEPC PBS and pelleted by centrifugation at 800xg for 5 minutes. Cell pellets were quickly washed with hypotonic lysis buffer (10 mM HEPES pH 7.9, 1.5 mM MgCI 2 , 10 mM KCI, 0.5 mM DTT). Cell pellets were then lysed in hypotonic lysis buffer containing 0.16 ΙΙ/μΙ_ Ribosafe RNase inhibitors (Bioline), 2 mM PMSF (Sigma) and SIGMAFAST™ Protease Inhibitor Cocktail tablets, EDTA free (Sigma) according to manufacturer's instructions. Cells were lysed gently using a cut P1000 tip ensuring no physical force was exerted on the cell pellet. For fly tissue, da-GAL4 was used to drive transgene expression in all tissues, and 10 third instar larvae were homogenised and lysed using the same buffer and a dounce homogenizer. The lysate then underwent differential centrifugation (1 ,500g, 3 min, 4°C then 3,500g rpm, 8 min, 4 °C and then 17,000g, 1 min, 4 °C) transferring the supernatant to a fresh tube after each centrifugation. The supernatant was then added to PureZOL (Biorad) to extract RNA. The nuclear pellet obtained after centrifugation at 1 ,500g for 3 min was lysed in Reporter lysis buffer (Promega) for 10 minutes on ice before centrifugation at 17,000g, 5 minutes, 4°C.

Total fractions were collected directly in Reporter lysis buffer (Promega) containing 0.16 U/μΙ- Ribosafe RNase inhibitors (Bioline), 2 mM PMSF (Sigma) and SIGMAFAST™ Protease Inhibitor Cocktail tablets, EDTA free (Sigma) according to manufacturer's instructions and lysed for 10 min on ice before centrifugation at 17,000g, 5 minutes, 4°C. The supernatant was then added to PureZOL (Biorad) to extract RNA. Equal volumes of total, nuclear and cytoplasmic lysates were resolved using SDS-PAGE, electroblotted onto nitrocellulose or PVDF membranes and probed using SSRP1 and a- tubulin (Neuro-2a), SSRP1 and Actin (iNeurons), or a-tubulin and a-Histone H3 (Drosophila).

Quantitative RT-PCR

Total RNA was extracted from 5-10 crushed larvae or adult Drosophila using 800 μΙ PureZOL (BioRAD) and a syringe with a 21G gauge needle for homogenization. Lysate was cleared by centrifugation for 10 minutes at 12,000 g at 4 ° C. 200 nl of chloroform was added to the supernatant and vigorously shaken for 15 seconds. After 10 minutes of incubation at room temperature, tubes were centrifuged at 12,000 g for 10 minutes at 4°C and supernatants (400 μΙ) were collected. RNA was precipitated for 30 minutes at room temperature with 2 μΙ Glycogen (5 mg/ml, Ambion) and 500 μΙ isopropanol and pelleted at 12,000 g for 20 minutes at 4°C. Pellets were washed with 70% DEPC EtOH and re- suspended in 40 μΙ DEPC water. Total RNA was then treated with DNasel (Roche) and quantified using a Nanodrop (NanoDrop ® Technologies).

For HEK cells, 50,000 cells were split into each well of 24-well plates and transfected with 700 ng pcDNA6.2-GW/EmGFP-Control or human SRSF1 miR RNAi constructs using 3.5 μg PEI/ml media and one tenth medium volume OptiMEM (ThermoFisher). For iAstrocytes, 20,000 induced neural progenitor cells (iNPCs) were plated in 6-well plates in astrocyte medium. The day after plating, 3 wells were transduced with lentivirus expressing human SRSF1-RNA at an MOI of 5.

Total RNA was extracted from HEK cells 72 hours after transfection or iAstrocytes 5 days after transduction and RNA extracted using the EZ Total RNA Isolation Kit (Geneflow). Briefly, cells were washed in DEPC-treated PBS before lysis directly in the culture dish at room temperature using the denaturing solution. Lysed cells were scraped and removed from the culture dish and equal volume extraction buffer added, vigorously shaken, incubated at room temperature for 10 minutes and then centrifuged for 15 minutes at 4°C and 12,000 g. RNA was subsequently precipitated from the upper phase using equal volume isopropanol overnight at -20°C, pelleted at 12,000 g, 4°C for 15 minutes, washed with 70% DEPC EtOH and re-suspended in 22.5 μΙ DEPC water. RNA was then treated with DNasel (Roche) and quantified using a Nanodrop (NanoDrop ® Technologies).

Following RNA quantification, 2 μg RNA was converted to cDNA using BioScript™ Reverse Transcriptase (Bioline). qRT-PCR primers were designed using Primer-BLAST (NCBI) and validated using a 1 in 4 serial template dilution series (standard curve with R 2 >0.97). qRT- PCR reactions were performed in duplicate for 3 independent biological repeats using the Brilliant III Ultra-Fast SYBR® Green QPCR Master Mix {Agilent Technologies) on a MX3000P QPCR system {Statagene). qRT-PCR data was analysed using MxPro (Stratagene) and GraphPad Prism (Version 6). The following qPCR primers were used:

Drosophila SF2 (designed using Primer-BLAST) (SEQ ID NO 74/75)

Fwd: 5'-TACCGCGTCATGGTTACTGG-3'

Rev: 5'-GTACGCGAATGTAGGCAACC-3'

Drosophila Ref1 (designed using Primer-BLAST) (SEQ ID NO 76/77)

Fwd: 5'- CGATATGTACGACGGACCGAA-3'

Rev: 5'- CGGACCAAAGTCGTTGAAGAG-3'

Drosophila Tub84b (described in reference (36)) (SEQ ID NO 78/79)

Fwd: 5'-TGGGCCCGTCTGGACCACAA-3'

Rev: 5'-TCGCCGTCACCGGAGTCCAT-3'

Drosophila C9 3'UTR (described in reference 69 ) (SEQ ID NO 80/81)

Fwd: 5'-JTCCAACCTATGGAACTGATGA-3'

Rev: 5'- GGTTTTCCTCATTAAAGGCATTC-3'

Human SRSF1 (designed using Primer-BLAST) (SEQ ID NO 82/83)

Fwd: 5'-CCGCATCTACGTGGGTAACT-3'

Rev: 5'- TCGAACTCAACGAAGGCGAA-3'

Human ALYREF/ THOC4 (designed using Primer-BLAST) (SEQ ID NO 84/85)

Fwd: 5'- TCTGGTCGCAGCTTAGGAAC-3'

Rev: 5'- CCACCTCTGTTTACGCTCTGT-3'

Human U1 snRNA (designed using Primer-BLAST) (SEQ ID NO 86/87)

Fwd: 5'-CCATGATCACGAAGGTGGTT-3'

Rev: 5'-ATGCAGTCGAGTTTCCCACA-3' Human SMN (described in reference 70 ) (SEQ ID NO 88/89)

Fwd 5'-CTTGTGAAACAAAATGCTTTTTAACATCCAT-3'

Rev 5'-GAATGTGAGCACCTTCCTTCTTTTT-3' Human JUN (designed using Primer BLAST) (SEQ ID NO 90/91)

Fwd 5'-GAACTGCACABCCAGAACAC-3'

Rev 5TGGGTTGAAGTTGCTGAGG-3'

C9RAN (designed using Primer-BLAST). Primers anneal downstream of the G4C2 or C4G2 repeat sequences in the 3'UTR of mRNA transcribed from pcDNA3.1 constructs. (SEQ ID NO 92/93)

Fwd 5'-GGGCCCTTCGAACCCCCGTC-3'

Rev: 5'GGGAGGGGCAAACAACAGAT-3'

Human C90RF72 Exon-1 Forward (designed using Primer BLAST) (SEQ ID NO 94) 5'-TCAAACAGCGACAAGTTCCG-3'

Human C90RF72 Exon-3 Reverse (designed using Primer BLAST) (SEQ ID NO 95) 5'-GTCGACATGACTGCATTCCA-3'

Human C90RF72 lntron-1 Reverse (designed using Primer BLAST) (SEQ ID NO 96) 5'-GGAGAGAGGGTGGGAAAAAC-3'

MTT cell proliferation assay

Neuro-2a (N2A) cells were split into each well of a 24 well plate (30,000 cells / well). Each plate contained 4 wells with only media to serve as a blank and 4 wells / treatment. Cells were transfected for 72 hours with either 500ng pcDNA 3.1 , pcDNA 3.1/RAN-G4C2x15 RAN-G4C2x38-sense, RAN-C4G2x15 or RAN-C4G2x39-antisense; or 250 ng pcDNA6.2- GW/EmGFP-Control miR RNAi, pcDNA6.2-GW/EmGFP-mouse SRSF1 miR1 +2 RNAi, p3xFLAG/SRSF1 (1 1-196), or p3xFLAG/SRSF1 (1 1-196)-m4 and 250 ng pcDNA 3.1 , pcDNA 3.1/RAN-G4C2x38-sense or RAN-C4G2x39-antisense. 250 mg Thiazolyl Blue Tetrazolum Bromide reagent (MTT) was added to each well and incubated in the dark at 37°C for 1 hour. Cells were subsequently lysed with equal volume MTT lysis buffer (20%SDS, 50% Dimethylformamide (DMF)) and incubated, shaking, at room temperature for 1 hour. Absorbance at 595 was then assessed with a PHERAstar FS (BMG Labtech). Experiments were performed in triplicate for each treatment. Absorbance data was analysed using PHERAstar MARS (BMG Labtech) and GraphPad Prism (version 6). Immunofluorescence of rat cortical neurons

Cortical neurons were isolated, cultured and transfected as described previously 36 . Briefly, neurons were transfected using Lipofectamine LTX with PLUS reagent according to the manufacturer's instructions (Thermofisher; DNA: PLUS: LTX ratio of 1 :0.5:0.5 with 2 μg DNA/100,000 cells/cm 2 ). After 6 hours, the transfection mix was replaced with conditioned medium.

Immunofluorescence staining of rat cortical neurons was performed 72 hours after transfection as described previously 36 . Briefly, cells cultured on glass coverslips were fixed with 3.7% formaldehyde in PBS at room temperature for 20 minutes. Cells were washed once with PBS and residual formaldehyde was quenched with 50mM NH 4 CI in PBS at room temperature for 10 minutes. Cells were subsequently washed in PBS and permeabilised with 0.2% Triton X-100 in PBS at room temperature for 5 minutes. Cells were washed in PBS to remove excess Triton X-100, blocked with 4% goat serum in PBS for 2 hours at room temperature and then incubated with the V5 antibody [1 :1000 dilution] (ThermoFisher Scientific #R96025) in PBS containing 4% goat serum overnight at 4°C. Cells were washed 3 times with PBS containing 4% goat serum and incubated for 1 hour with PBS containing 4% goat serum & goat anti-mouse secondary antibody, Alexa Fluor 594 [1 :1000 dilution] (ThermoFisher Scientific). Cell transfected with pcDNA6.2-GW/EmGFP-Control or human SRSF1 miR RNAi constructs were subsequently stained with Hoechst 33342 for 10 minutes at room temperature, washed with 3 times PBS and mounted in fluorescence mounting medium (Dako). After incubation in the secondary antibody, cells transfected with p3xFLAG/SRSF1 (1 1-196) or p3xFLAG/SRSF1-m4 were washed 3 times with PBS containing 4% goat serum, incubated at room temperature for one hour with PBS containing 4% goat serum & anti-FLAG ® M2-FITC antibody [^g/ml] (Sigma-Aldrich #F4049) and subsequently stained with Hoechst 33342. Cells were then washed with 3 times PBS and mounted in fluorescence mounting medium (Dako).

Co-lmmunoprecipitation

Cells were split into 1 x 10 cm plates / treatment (1.5 x 10 6 cells/plate) and transfected with 15 μg p3xFLAG, p3xFLAG/SRSF1 (1 1-196), p3xFLAG/SRSF1 (1 1-196)-m2 or p3xFLAG/SRSF1 (1 1-196)-m4 using 3 μg PEI/1 μg DNA and one tenth medium volume OptiMEM. Proteins were extracted from Neuro-2a cells 48 hours after transfection. Cells were briefly washed in ice cold PBS, scraped into 500 μΙ ice cold lysis buffer, passed through a 21 G gauge needle 10 times and left to lyse on ice for 10 minutes. Lysed cells were cleared by centrifugation at maximum speed at 4°C for five minutes and protein extracts were quantified using Bradford Reagent. 2 mg of total protein in 1 ml lysis buffer was incubated with 30μΙ anti-FLAG ® M2 affinity resin slurry (Sigma A2220) (which had been blocked overnight with 1 % BSA in IP lysis buffer) for 2 hours at 4°C on a rotating wheel. The anti-FLAG ® M2 affinity resin captured protein complexes were washed 5 times with ice-cold lysis buffer and eluted in 50 μΙ IP lysis buffer supplemented with 100 μg/vΓ^\ 3xFLAG peptide {Sigma #F4799) for 30 minutes at 4°C on a rotating wheel. 30 μg total protein and 15 μΙ captured protein complexes were resolved by SDS-PAGE, electroblotted onto nitrocellulose membrane and probed using FLAG, NXF1 clone 53H8 [1 :2000] (Abeam ab50609) and D-tubulin.

RNA:protein UV crosslinking assays

Recombinant proteins expressed in 1.51 of E. coli BL21 (DE3)-RP (Novagen) cell cultures were purified by I MAC chromatography on TALON/Cobalt beads (Clontech) in 1 M NaCI containing buffers to prevent the potential co-purification of E. coli RNA (Lysis buffer: 50 mM TRIS-HCI pH8.0, 1 M NaCI, 0.5% Triton X-100; Wash buffer: 50 mM TRIS-HCI pH8.0, 1 M NaCI, 0.5% Triton X-100, 5 mM imidazole). Elution was achieved in step in buffer containing 200 mM imidazole (50 mM TRIS-HCI pH8.0, 500 mM NaCI, 200 mM imidazole) and 50 mM L-Arg and L-Glu to prevent protein precipitation while retaining interaction with RNA and NXF 1 38,39 32 p. ra diolabelled probes (synthetic G4C2x5 or C4G2x5 RNA oligonucleotides purchased from Dharmacon) for 10 min at room temperature and 10 min on ice prior to UV- crosslinking or not (10 min, 1.5J/cm 2 ). Binding reactions were resolved on SDS-PAGE prior to analysis by Coomassie staining and Phosphoimaging.

RNA immune-precipitation (RIP) assays

Cells were split into 1 x T-175 flasks / treatment (5 x 10 6 cells/plate) and transfected with 30 μg p3xFLAG, p3xFLAG/SRSF1 (1 1-196) or p3xFLAG/SRSF1 (1 1-196)-m4 and 10 μg pcDNA 3.1/RAN-G4C2x15-sense, RAN-G4C2x38-sense, RAN-G4C2x15-antisense or RAN- C4G2x39-antisense using 3 μg PEI/ ^g PEI and one tenth volume OptiMEM. Protein-RNA complexes were extracted from Neuro-2a cells 48 hours after transfection. Protein-RNA complexes were cross-linked using 1 % formaldehyde for 10 minutes, shaking at room temperature. Residual formaldehyde was quenched with 250mM Glycine at room temperature for 5 minutes, shaking. Cross-linked cells were subsequently washed in ice cold DEPC treated PBS and scraped into ice cold RNase free lysis buffer (DEPC treated water containing 50 mM Hepes pH7.5, 150 mM NaCI, 10% glycerol, 0.5% triton X-100, 1 mM EDTA, 1 mM DTT, 1 μΙ RNase inhibitor, protease inhibitor cocktail). Cells were passed through a 21 G gauge needle 10 times and left to lyse on ice for 10 minutes, followed by centrifugation at maximum speed at 4°C for five minutes and quantification using Bradford Reagent. 2.5 mg of total protein at a 1 mg/ml concentration was incubated with 40 μΙ anti- FLAG ® M2 affinity resin slurry (which had been blocked overnight with 1 % BSA and 5 μΙ/ml ssDNA) overnight at 4°C on a rotating wheel. One 15 th of the total protein extract was retained for an input sample.

The anti-FLAG ® M2 affinity resin captured protein-RNA complexes were washed 5 times with ice-cold RNase free lysis buffer. Complexes were subsequently reverse-crosslinked and eluted from the resin in EZ RNA extraction denaturing buffer for 1 hour at 70°C, re- suspending the resin every 10 minutes. The formaldehyde crosslinks were reversed by heating the samples for 1 hour at 70°C and RNA was extracted using PureZOL (for total samples) or the EZ Total RNA Isolation Kit (for eluted complexes) as described in the qRT- PCR section. Extracted RNA samples were re-suspended in 25μΙ RNase-free water. Extracted RNA samples were DNase treated and 10 μΙ input or eluate RNA was converted to cDNA as described previously in the quantitative RT-PCR section. qRT-PCR primers were designed using Primer-BLAST and validated using a 1 in 4 serial template dilution series (standard curve with R 2 >0.97). qRT-PCR reactions were performed in duplicate for 3 independent biological repeats using the Brilliant III Ultra-Fast SYBR® Green QPCR Master Mix {Agilent Technologies) on a CFX 96™ Real-Time System (BIO-RAD). qRT-PCR data was analysed using CFX Manager 3.1 (BIO-RAD) and GraphPad Prism (Version 6). The relative concentration of cDNA was calculated for each sample and the input samples multiplied by 15 to obtain a total input value. Eluate samples were expressed as a percentage of the input total input values. qPCR primers were described above in the qRT- PCR section.

RNA fluorescence in situ hybridisation (FISH) and visualization of RNA foci

To visualize sense RNA foci, RNA fluorescence in situ hybridisation (FISH) was performed as described previously (9). Briefly, iAstrocyte cells were grown on 13 mm coverslips for 5 days following LV-SRSF1-RNAi transduction, before being fixed and permeabilised in 4% paraformaldehyde & 0.2% Tween-20 (Sigma) for 10 minutes. Fixed cells were subsequently washed in DEPC-treated PBS and were blocked with hybridization solution [50% formamide, 2 x saline sodium citrate (SSC), 100 mg/ml dextran sulphate & 50 mM sodium phosphate pH 7.0] for 1 h at 66°C. Cells were then incubated with 400 ng/ml of denatured probe (a 5' TYE- 563-labelled LNA (16-mer fluorescent)-incorporated DNA probe against the sense C9orf72 RNA hexanucleotide repeat (Exiqon, Inc.)) in hybridization solution overnight at 66°C. After hybridization, cells were washed once in 2xSSC & 0.1 % Tween-20 at room temperature for 15 minutes, followed by three 15 minutes washes in 0.1 x SSC at 65°C. Coverslips were mounted using a DAPI-containing mounting medium (Vector Labs, Inc.). All solutions were made with DEPC-treated water. RNA foci were visualized using a Leica SP5 confocal microscope system and a 63/1.4 oil immersion objective lens. The presence of foci was assessed within a high resolution (848 μηι 2 per image, 393 x 393 pixels) using 0.9 μηι z-stacks through the entire volume of the cell.

Confocal immunofluorescence microscopy in human post mortem spinal cord tissue from C90RF72-ALS patients

This study was approved by the South Sheffield Research Ethics Committee and informed consent was obtained for all samples. Brain and spinal cord tissues were donated to the Sheffield Brain Tissue Bank for research, with the consent of the next of kin. IHC and RNA fluorescence in situ hybridisation (FISH) were performed on formalin fixed paraffin- embedded (FFPE) tissues from four C90RF72+ patients with ALS. After wax removal, antigen retrieval was performed by microwaving for 10 min in 0.8 mM EDTA pH9.5. A 5' TYE-563-labelled LNA (16-mer fluorescent)-incorporated DNA probe was used against the sense (Exiqon, Inc.; batch number 607323) RNA hexanucleotide repeat as described in previous section. Slides were then immediately transferred to PBS/5 % BSA for protein staining using anti-SRSF1 antibody (Cell Signaling #8241) at a dilution of 1 :200. After incubation with the primary antibodies overnight at 4 °C in DEPC-treated PBS/5 % BSA, slides were washed in DEPC PBS and incubated with fluorescent secondary anti-rabbit Alexa Fluor 488 antibodies (Abeam ab150077). Mounted slides were visualised by confocal microscopy as described in previous section.

Statistical analysis of data

We have used one-way or two-way ANOVA (analysis of variance) to statistically evaluate and plot our data using the GraphPad Prism version 6 (GraphPad Software Inc., San Diego, CA). All files and data points are available on request. RNA foci were counted in a blinded manner and several researchers carried out the analysis. Several researchers were also involved in producing the qRT-PCR and western blot data showing that depleting SRSF1 or inhibiting the interaction of SRSF1 with NXF1 lead to nuclear export alteration of C0ORF72 repeat transcripts and both sense and antisense DPRs. Either one-way or two-way ANOVA (analysis of variance) with Tukey's correction for multiple comparisons was used for most experiments with the following exceptions: DPR analysis in primary neurons used Fisher's exact test; adult fly climbing ability was analysed by Kruskal- Wallis non-parametric test with Dunn's correction for multiple comparisons; and the analysis of G4C2x36 transcripts in Drosophila used paired two-tailed f-test. No randomization was used in the animal studies. Data were plotted using GraphPad Prism 6. Significance is indicated as follows; NS: non-significant, *: p<0.05, **: p<0.01 , ***: p<0.001 , ****: p<0.0001 ; All files and data points are available on request. RNA foci, DPR-positive neurons, crawling and climbing assays were analysed in a blinded manner and several investigators carried out the analysis. Several researchers were also involved in producing the qRT-PCR and western blot data showing that depleting SRSF1 or inhibiting the interaction of SRSF1 with NXF1 lead to nuclear export alteration of C90RF72 repeat transcripts and reduction of both sense and antisense DPRs.

SRSF1 expression and RNAi gene therapy vectors

shRNA design for SRSF1 knockdown

Initial shRNA design was conducted using the Invitrogen Block-it RNAi designer (https://rnaidesigner.thermofisher.com/rnaiexpress/). Input sequence was the NCBI accession number: NM_001078166.1 - corresponding to short isoform transcript variant 2 of the SRSF1 mRNA, ensuring both SRSF1 isoforms are targeted. The ORF was selected as the region for target design and GC content was set at the default values of 35-55%. Output sequence strand orientation was sense-loop-antisense. The selected loop sequence was CGAA but this can be modified to incorporate different loop sequences if necessary.

The top 10 output sequences ranked by likelihood of target knockdown are shown in Table 7.

Table 7 - Invitrogen RNAi design output

These 10 highest ranked sequences were interrogated for their off-target potential by in silico analysis of the antisense strand using the siSPOTR online tool (https://sispotr.icts.uiowa.edu/sispotr/tools/lookup/evaluat e.html). In siSPOTR, shRNA sequences are given a potential of off-targeting score (POTS) where a score of ≤30 is considered ideal. A higher score suggests a high likelihood of off-targeting. The results of this analysis are shown in Table 8.

Table 8 - siSPOTR analysis of Invitrogen designed shRNA

siSPOTR also contains a function enabling the design of minimal off-targeting shRNA sequences irrespective of potential efficacy. Using the CDS of the short SRSF1 isoform as an input, 8 sequences with a POTS score lower than 30 were generated and are shown in Table 9. Table 9 - shRNA sequences with low off-targeting potential designed by siSPOTR

As the siSPOTR algorithm does not rank sequences based on their ability to knockdown the target transcript it would be sensible to choose a number of shRNA sequences from both design tools and validate their ability to silence SRSF1 mRNA in vitro. siSPOTR indicates that several of the shRNAs from the Invitrogen design tool have an extremely high POTS score far in excess of the recommended score of≤30, perhaps excluding them from an initial knockdown screen in favour of lower scored sequences. siSPOTR reference: Boudreau, R.L. et al., 2013. siSPOTR: a tool for designing highly specific and potent siRNAs for human and mouse. Nucleic Acids Research, 41 (1), p.e9.

Example 1

To gain functional insight into the nuclear export dependence of G4C2 repeat transcripts, we tested whether reducing the expression levels of the conserved nuclear export adaptors SRSF1 (26) and ALYREF (27) might rescue neurodegeneration in C90RF72-ALS Drosophila which exhibits both RNA and DPR-mediated neurotoxicity ( 14). Flies expressing 36 uninterrupted repeats (pure G4C2x36) were crossed with two independent transgenic RNAi lines each targeting SRSF1 or REFALY (28), overall achieving 70-80% reduction in mRNA expression levels (Fig. 9A). Targeted expression of G4C2x36 disrupts the compound eye and is minimally altered by co-expression of a control RNAi ( 14) (Fig. 9B). In contrast, co-expression of two different SRSF1-RNA\ (also called SF2/ASF) sequences completely prevented neurodegeneration, while two knockdown lines of REFALY (also called Refl or ALYREF) only showed a modest rescue of the neurodegenerative phenotype. We validated the successful knockdown of SRSF1 and ALYREF in the corresponding flies showing 70- 80% reduction in mRNA expression levels (Fig. 9b). Moreover, neuronal expression of G4C2x36 causes both larval and adult locomotor deficits in this model 44 . Consistent with the previous results, SRSF 7-depletion restored locomotor function while Ref1-RNA\ provided no rescue of the behavioural phenotype (Fig. 9, C and D).

This effect is specific of SRSF1 since depletion of ALYREF showed no effect, which is in agreement with the rough eye phenotypes. The neurotoxicity effects observed in the G4C2x36 C90RF72-ALS model of Drosophila were primarily attributed to the expression of DPRs16. Accordingly, we now show that the depletion of SRSF1 leads to prominent reduction in the production of both sense and antisense poly-GP DPRs (Fig. 25a).

Example 2

To test for the hexanucleotide-repeat expansion specificity of the SRSF1-RNAi rescued phenotypes, we used the previously established GR36 and PR36 fliesl 6 which respectively express 36-repeat poly-Gly-Arg and poly-Pro-Arg DPRs using alternative codons. As reported in the original study16, the GR36 flies have a high rate of mortality and only a few GR36 flies crossed with Ctrl or SRSF1-RNAi survived to adulthood. Nonetheless, the partial depletion of SRSF1 did not significantly ameliorate the rough eye phenotypes (Fig. 9e) or the locomotor deficits (Fig. 9f) induced by the G4C2-independent expression of DPRs in both GR36 and PR36-expressing flies. These results indicate that partial depletion of SRSF1 exerts neuroprotection through direct effects on the C90RF72 hexanucleotide repeat expansion rather than indirect effect on gene expression alteration or downstream accumulation of DPRs.

Partial depletion of SRSF1 prevents patient-derived astrocytic neurotoxicity and motor neuron death. To apply our in vivo findings to human C90RF72-related ALS, we sought to deplete SRSF1 in patient-derived cell models. Human Sf?SF7-knockdown plasmids co- expressing a GFP reporter and a pre-miRNA cassette were engineered to produce recombinant SRSF1-RNA\ lentivirus (LV-Sf?S 7-RNAi) (fig. 1). HEK293T cells co- transfected with SRSF 7-RNAi constructs and a FLAG-tagged SRSF1 expression plasmid showed efficient and specific depletion of SRSF1 (Fig. 10A). The survival of C90RF72-ALS patient-derived neurons is not impaired in vitro (6, 7) but altered microglial function was recently reported to contribute in vivo to C90RF72-related neurodegeneration (16). The survival and morphology of C90RF72-ALS patient-derived neurons is indistinguishable from control neurons 45, 46.

We thus assessed motor neuron survival in co-cultures with patient-derived astrocytes using our recently developed assay that recapitulates the astrocyte-mediated neurotoxicity observed in ALS for both primary mouse and human derived neurons (29). Transduction of human iN PC-derived astrocytes (iAstrocytes) using a viral multiplicity of infection (MOI) of 5 led to efficient transcript knockdown comparable to levels achieved in vivo in neuro-protected G4C2x36+ SRSF1-RNA\ Drosophila (fig. 2). Mouse GFP-Hb9+ motor neurons were plated onto LV- SRSF 1- RNAi transduced astrocytes derived from three controls and three C90RF72-ALS patient fibroblast lines (table 5) and survival was quantified using a high-throughput imaging system (Online methods, Fig 2b, methods).

Quantification of SRSF1 and ALYREF mRNA levels confirmed the specific and partial knockdown of SRSF1 transcripts in both control and C90RF72-ALS iAstrocytes (Fig. 10B). Control iAstrocytes transduced with an RFP-adenovirus (red) efficiently support the growth of GFP-Hb9+ motor neurons (green) while fewer motor neurons survived when co-cultured with astrocytes derived from C90RF72-ALS patient fibroblasts (29) (Fig. 10C). Quantification of surviving motor neurons from four replicate experiments showed that while depletion of SRSF1 is not detrimental to control co-cultures, motor neuron death was prevented by depletion of SRSF1 in co-cultures derived from three separate C90RF72-ALS cases (Fig. 10D). Example 3

To investigate potential nuclear export alteration of G4C2 repeat transcripts, we quantified nuclear and cytoplasmic sense RNA foci in the C90RF72-ALS iAstrocytes. Representative images (fig. 3) and individual counts are reported in Additional Data table 1. Upon depletion of SRSF1 , the number of cytoplasmic RNA foci decreased, while nuclear RNA foci concomitantly increased (Fig. 11 E) consistent with the predicted nuclear export inhibition of G4C2-repeat transcripts.

Due to poor detection of DPRs in flies or patient-derived iAstrocytes and iNeurons (not shown), we engineered plasmids expressing 38 uninterrupted G4C2-sense or 39 uninterrupted C4G2-antisense repeat transcripts that are substrates for specific RAN translation of DPRs (fig. 4). Transfection of mouse neuronal N2A cells with sense or antisense repeat constructs led to specific DPR production of poly-Gly-Pro (expressed from both sense and antisense transcripts and poly-Gly-Ala (expressed from sense transcripts) (Fig. 11 A). However, co-transfection of a mouse SRSF1-RNA\ plasmid (fig. 5) led to a marked reduction in the RAN translation of both sense and antisense DPRs (Fig. 11 A, left panels; quantification in fig. 6A). This does not depend on reduced splicing activity of SRSF1 since the RAN constructs were engineered without splicing sites. Moreover, it is specific for the hexanucleotide-repeat sequences since expression of synthetic poly-Gly-Ala x36 or poly- Gly-Pro x36 using alternatives codons (GGA/T, GCA/T, CCA/T) is not altered upon SRSF1 depletion (Fig. 11 A, right panels). Supporting this, SRSF1-RNA did not ameliorate locomotor deficits conferred by DPR expression Drosophila (fig. 7). These data are consistent with our previous reported findings that showed specific and direct binding of SRSF1 onto G4C2- repeat transcripts (9).

Example 4

To identify potential impairment in the nuclear export of C90RF72 repeat transcripts, we measured the cytoplasmic expression levels of G4C2 transcripts in the presence of Ctrl or SRSF1-RNA\ constructs. Cytoplasmic repeat transcript levels normalized to total levels are markedly reduced upon exposure to SRSF1-RNAi (Fig. 11 B, left image) indicating that the nuclear export of C90RF72-repeat transcripts is inhibited by the depletion of SRSF1. The quality of cellular fractionation was checked by immunoblotting using antibodies against the chromatin-remodeling SSRP1 factor, showing absence of nuclear contamination in the cytoplasmic fractions (fig. 8). SRSF1 mediates mRNA nuclear export through binding to NXF1 (26, 30). We previously showed that four arginine residues lying in the unstructured linker between the two RNA recognition motifs of SRSF1 (amino-acids 11-196) are required both for RNA nuclear export and interaction with NXF1 , while mutations of only two arginine residues lead to reduced binding to NXF1 (23). Co-transfection of the quadruple SRSF1 point mutant (SRSF1-m4) which does not interact with TAP/NXF1 (23) led to marked reduction in the normalized cytoplasmic repeat transcript levels (Fig. 10B, right image) and production of sense and antisense DPRs while the wild type region or the same sequence bearing only two arginine mutations (SRSF1-m2) respectively had no or little effect (Fig. 11 C, quantification in fig. 6 B and C) demonstrating that C90RF72-repeat transcripts are exported from the nucleus via a mechanism that requires the interaction of SRSF1 and NXF1. Accordingly, both the depletion of SRSF1 and the expression of the dominant negative mutant SRSF1-m4 suppress the neurotoxicity mediated by expression of the C90RF72-repeat transcripts in neuronal N2A cells (Fig. 11 D).

In this report, we show for the first time that the nuclear export of C90RF72 repeat transcripts and subsequent RAN translation depends on the interaction of the nuclear export adaptor SRSF1 with the NXF1 export receptor. Inhibiting this interaction or depleting SRSF1 lead to marked inhibition of the nuclear export of C90RF72 repeat transcripts and reduced RAN translation of both sense and antisense DPRs. Moreover, partial depletion of SRSF1 prevents in vivo neurodegeneration and rescues locomotor deficits in a Drosophila model of C90RF72-ALS, consistent with previous work which showed that DPRs cause neurodegeneration in Drosophila ( 14). Depletion of SRSF1 also suppresses neurotoxicity in co-cultures of motor neurons and human-derived C90RF72-ALS astrocytes and in neuronal N2A cells. Hence, this represents a promising prospect for the development of an effective neuroprotective strategy in C90RF72-related ALS. It is noteworthy that SRSF1 -RNA\ depletion was used as a therapeutic approach to prevent oncogenic transformation both in vitro and in vivo in mice (31). The cellular pathways causing neurodegeneration following nuclear export of C90RF72 G4C2 repeat sequences and the precise mechanism(s) of neuroprotection conferred by the partial depletion of SRSF1 remain to be elucidated in future studies.

Example 5

ALYREF and SRSF1 directly bind G4C2 and C4G2 repeat RNA. We performed in vitro UV- crosslinking assays using purified recombinant proteins and synthetic G4C2x5 and C4G2x5 RNA probes to investigate direct protein:RNA interactions. Recombinant hexa-histidine- tagged human ALYREF, SRSF1 amino-acids 1 1-196 which retains wild type ability to bind RNA and NXF136, and MAGOH, a control protein which does not bind RNA49, were purified by ion metal affinity chromatography in high salt to disrupt potential interactions with bacterial RNA. Purified proteins were incubated with 5'-end 32P-radiolabelled G4C2x5 (Fig. 20a) or C4G2x5 (Fig. 20b) RNA probes prior to irradiation with UV where indicated (+) and resolved by SDS-PAGE. As shown on the Phosphoimages, covalently-bound RNA molecules remained associated with ALYREF and SRSF1 visualised on the Coomassie- stained panels during the denaturing electrophoresis. These data demonstrate direct interactions of ALYREF and SRSF1 with both sense and antisense repeat RNA in agreement with our previous studies 42, 43 The interactions are specific since no binding of RNA was detected in absence of UV-irradiation or with the negative control protein MAGOH. These direct interactions are also consistent with our previously reported co-localization of ALYREF with RNA foci in motor neurons from C90RF72-ALS patients 42 . Furthermore, we show that SRSF1 co-localizes with RNA foci in motor neurons from human post mortem spinal cord tissue of C90RF72-ALS cases (Fig. 20c).

Example 6

Generation of neuronal cell models recapitulating the RAN-dependent translation of sense and antisense DPRs. To investigate whether the binding of SRSF1 to G4C2-sense and C4G2-antisense repeat RNA sequences has the ability to trigger the nuclear export of repeat transcripts, we generated synthetic mammalian expression constructs bearing increasing lengths of pure repeat sequences in the absence of ATG or Kozak elements to specifically investigate RAN-dependent translation of dipeptide repeat proteins. Following annealing of synthetic G4C2 or C4G2 repeat oligonucleotides as described in Methods and Fig. 4a-d, we engineered plasmids expressing transcripts containing 15 or 38 uninterrupted sense repeats (G4C2x15 or G4C2x38) and 15 or 39 uninterrupted antisense repeats (G4C2x15 or C4G2x39) with 3'-end stop codons in each of the three frames (Fig. 21a). The lengths of repeats were confirmed by sequencing and poly-acrylamide gel electrophoresis. The nucleotide sequences are presented in Supplementary Fig. 4e. In mammals, the bulk nuclear export of mRNA is predominantly coupled to the recruitment of the TREX complex during splicing 47 . Three C90RF72 transcripts, each containing 4 or 10 introns, are transcribed from the C90RF72 gene. The synthetic repeat constructs were engineered without splicing elements or intronic sequences to investigate the nuclear export potential of G4C2 and C4G2 repeat RNA sequences in repeat transcripts independently of functional coupling to pre-mRNA splicing. Transfections of mouse neuronal N2A cells with sense or antisense repeat constructs led to the formation of RNA foci for all repeat transcripts (Fig. 21 b, data shown for the G4C2x38 construct) and to specific DPR production of poly-Gly-Pro (expressed from both sense and antisense transcripts) and poly-Gly-Ala (expressed from sense transcripts) for repeat transcripts bearing 38 sense or 39 antisense repeats (Fig. 21 c). Interestingly, the expression of DPRs correlated with neurotoxic effects in MTT cell proliferation assays, while the control plasmid or the constructs expressing 15 sense or antisense repeats but no DPRs did not exhibit significant cytotoxicity in the neuronal cells (Fig. 21 d). We conclude that the minimal G4C2x38 sense and C4G2x39 antisense repeat RNA transcripts can be exported into the cytoplasm independently of functional coupling to pre-mRNA splicing and are substrates for subsequent RAN translation of DPRs. Example 7

Depletion of SRSF1 inhibits the production of sense and antisense DPRs in neuronal cells. A mouse SRSF1 -knockdown plasmid co-expressing a GFP reporter and a pre-miRNA cassette was engineered similarly to the previously described human SRSF1-RNAi (Fig. 5). Transfection of mouse neuronal N2A cells with G4C2x38 or C4G2x39 repeat constructs and the mouse SRSF1-RNAi plasmid led to a marked reduction in the RAN translation of both sense and antisense DPRs (Fig. 1 1 a, left panels and Fig. 6b for quantification). The SRSF1- RNAi dependent inhibition of DPR production does not depend on the splicing activity of SRSF1 since the RAN constructs are devoid of splicing sites. Moreover, the SRSF1 depletion is specific to the hexanucleotide-repeat sequences since expression of synthetic 36-repeat poly-Gly-Pro (GP36) or 36-repeat poly-Gly-Ala (GA36) DPRs using alternative codons (sequences available in Fig. 27) is not altered upon SRSF1 depletion (Fig. 1 1 a, right panels and Fig. 6e-f for respective quantifications of GP36 and GA36). These results support our previous findings that the depletion of SRSF1 did not ameliorate the rough eye phenotypes (Fig. 9e) or the locomotor deficits (Fig. 9f) conferred by G4C2-independent GR36 and PR36 DPRs expression in Drosophila.

SRSF1 mediates mRNA nuclear export through binding to NXF134,54. We previously showed that four arginine residues lying in the unstructured linker between the two RNA recognition motifs of SRSF1 (amino-acids 1 1-196) are required both for RNA nuclear export and interaction with NXF1 , while mutations of only two arginine residues lead to slightly reduced binding to NXF1 in human embryonic kidney cells36. Similarly, endogenous NXF1 is specifically immunoprecipitated in neuronal N2A cells transfected with FLAG-tagged SRSF1 1 1-196 wild type or double R1 17.1 18A mutant (SRSF1-m2). In contrast, the co- immunoprecipitation of NXF1 is severely impaired by the quadruple R93,94, 117, 118A mutations of SRSF1 (SRSF1-m4) (Fig. 6g). Co-transfection of the quadruple SRSF1-m4 dominant mutant further led to a marked reduction in the production of both sense and antisense DPRs while the wild type sequence or the variant bearing two arginine mutations (SRSF1-m2) respectively had no or little effect (Fig. 1 1 c). This was statistically assessed for both poly-GP and poly-GA DPRs produced by sense repeat transcripts (Fig. 6c) and poly- GP DPRs generated from antisense repeat transcripts (Fig. 56b. Taken together our data demonstrate that the expression of the C90RF72 repeat transcripts is dependent on a mechanism that requires the interaction of SRSF1 with the nuclear export receptor NXF1. Accordingly, both the depletion of SRSF1 and the expression of the dominant negative mutant SRSF1-m4 suppress the neurotoxicity mediated by expression of the C90RF72 repeat transcripts in neuronal N2A cells (Fig. 11 d). Supporting this, RNAi-mediated depletion of the Drosophila NXF1 homologue, sbr, could rescue the locomotor deficits in larvae and adult flies expressing G4C2x36 (Fig. 28).

Example 8

Sequestration of SRSF1 triggers the NXF1 -dependent nuclear export of hexanucleotide repeat transcripts in neuronal cells. Our result showing that expression of the SRSF1-m4 mutant protein acts as a dominant negative mutant for DPR production suggests that the SRSF1-m4 protein is sequestered onto the hexanucleotide repeat transcripts instead of the endogenous SRSF1 protein, preventing in turn interactions of repeat transcripts with NXF1 and nuclear export. Using in vitro UV cross-linking assays, we confirmed that the purified recombinant hexa-histidine-tagged SRSF1-m4 protein retains the ability to directly interact with synthetic 5'-end 32P-radiolabelled sense G4C2x5 (Fig. 8a) and antisense C4G2x5 (Fig. 8b) repeat RNA. These interactions are specific since no binding of RNA was detected in absence of UV-irradiation or with the negative control protein MAGOH. We next sought to investigate whether this was also true in live N2A cells using RNA immunoprecipitation (RIP) assays. N2A cells were transfected with FLAG control, FLAG-tagged SRSF1 or FLAG- tagged SRSF1-m4 and various lengths of sense or antisense repeat transcript constructs prior to fixing of ribonucleoprotein complexes. Cell extracts were then subjected to anti- FLAG immunoprecipitation under the same conditions used in the co-immunoprecipitation of NXF1 (Fig. 6g) prior to the qRT-PCR analysis of SRSF1-crosslinked RNA molecules. Validating our RIP assay, immunoprecipitation of both SRSF1 or SRSF1-m4 led to specific co-precipitation of SMN, a known SRSF1-binder55, but not of JUN, an intronless control transcript not expected to be bound by SRSF156, in N2A cells expressing either sense (Fig. 8c) or antisense (Fig. 8d) repeat transcripts. In sharp contrast, the levels of immunoprecipitated G4C2-sense or C4G2-antisense repeat transcripts significantly increased with the number of hexanucleotide repeats (Fig. 8c, d) showing length-dependent repeat-RNA sequestration of SRSF1 and SRSF1-m4 in neuronal cells.

To evaluate the effects of SRSF1 depletion or SRSF1-m4 expression on the nuclear export of C90RF72 repeat transcripts, we measured the total and cytoplasmic levels of G4C2x38 transcripts in the presence of Ctrl or SRSF1-RNAi and SRSF1 or SRSF1-m4 in transfected N2A cells. The quality of the cellular fractionation was checked by immunoblotting using antibodies against the chromatin-remodeling SSRP1 factor (Fig. 8e) showing absence of significant nuclear contamination in the cytoplasmic fractions. Total levels of G4C2x38 transcripts are not significantly altered upon SRSFI-RNAi or expression of SRSF1-m4 while the cytoplasmic levels of G4C2x38 transcripts are markedly reduced in both conditions (Fig. 29a). Cytoplasmic repeat transcript levels were also normalized to total levels to specifically assess the nuclear export process as in our previous studies 48, 49 . The cytoplasm ic/total mRNA level ratios are markedly reduced upon exposure to SRSF1 -RNAi (Fig. 8f, Ctrl-RNAi vs. SRSF1-RNAi). Similarly, the co-transfection of the SRSF1-m4 mutant which fails to interact with NXF1 led to marked reduction in the normalized cytoplasmic repeat transcript levels (Fig. 8f, SRSF1 vs. SRSF1-m4). To extend this analysis in vivo, Drosophila larvae expressing G4C2x36 and either Ctrl or SRSF1-RNAi were subjected to the same fractionation (Fig. 8g) and transcript analysis. As with cells, total levels of G4C2x36 transcripts are not significantly altered upon SRSF1-RNAi while the cytoplasmic levels of G4C2x36 transcripts are markedly reduced (Fig. 29b). Consequently, the cytoplasm ic/total mRNA level ratio is significantly reduced upon exposure to SRSF1-RNAi (Fig. 8h). Together these data demonstrate that depleting SRSF1 or preventing its repeat RNA-sequestration and interaction with NXF1 specifically inhibit the nuclear export of hexanucleotide repeat transcripts in vitro and in vivo.

Example 9

Depleting SRSF1 or inhibiting its RNA-repeat sequestration and interaction with NXF1 prevents the production of DPRs in primary neurons. We next sought to validate our findings in primary neurons. Due to high background staining obtained with poly-GP and poly-GA antibodies, we expressed V5 tags in all three frames downstream of the G4C2x38 repeat sequence (Fig. 30) to simultaneously detect all DPR species using the more specific and sensitive anti-V5 antibody. Cultured rat cortical neurons were transfected with the G4C2x38- 3xV5 construct and either the Ctrl-RNAi, SRSF1-RNAi, SRSF1 or SRSF1-m4 expression plasmids prior to immunofluorescence studies. The nucleotide sequence targeted by the SRSF1-RNAi miRNA hairpin-1 is identical in human, mouse and rat SRSF1 (Methods, Fig. 5). Microscopy image examples of DPR-negative and DPR-positive neurons are presented in Fig. 22a. The proportion of DPR-positive neurons in approximately 100 successfully transfected neurons from two independent experiments was quantified in each group and all counts were performed blinded. Depletion of SRSF1 led to a significant reduction (25%) in the proportion of neurons with RAN-translated DPR-staining compared to neurons transfected with Ctrl-RNAi (Fig. 22b). Inhibiting the sequestration of endogenous SRSF1 and the interaction with NXF1 by recruitment of the SRSF1 -m4 dominant mutant also led to a similar and significant reduction of neurons expressing DPRs compared to neurons transfected with wild type SRSF1 (Fig. 22c). Only cells showing absence of DPRs were counted DPR-negative. It is very likely that the effects of the SRSF1-RNAi or SRSF1-m4 expression have been under-estimated since neurons expressing reduced amounts of DPRs would still have been scored as DPR-positive. We concluded that our findings in N2A cells were corroborated in cultures of primary neurons. Example 10

SRSF1 depletion specifically inhibits the nuclear export of pathological transcripts retaining intron-1 hexanucleotide repeat expansions in neurons-derived from C90RF72-ALS patients. In order to investigate the nuclear export of C90RF72 transcripts in the context of wild type and repeat-expanded C90RF72 genes, we differentiated motor neurons from established induced-neural progenitor cells (iNPCs) derived from sex/age matched control and C90RF72-ALS patient fibroblasts52. Both control and C90RF72-ALS induced iNeurons express the neuronal lineage marker Tuj1 and exhibit the propensity to form complex branching (Fig. 31). High content imaging analysis of axonal length (Fig. 23a) and soma cell size (Fig. 23b) did not show any significant differences between control and C90RF72-ALS differentiated iNeurons under basal culture conditions, in agreement with previous reports 45,46 . To test a potential neuroprotective effect of SRSF1 depletion in disease relevant cells, we next differentiated iNPCs into motor neurons (iMNs), transduced them with an adenoviral vector expressing RFP under the Hb9 promoter and cultured them either in monoculture or in co-culture with control or ALS-derived iAstrocytes. We did not observe increased cell death or morphological abnormalities when the C90RF72-ALS iMNs were cultured without astrocytes (data not shown). Remarkably, however, the transduction of SRSF1-RNAi lentivirus in iMNs prior to co-cultures with C90RF72-ALS iAstrocytes resulted in significantly higher survival of iMNs against the astrocytic-derived toxicity (Fig. 23c) indicating a neuroprotective effect of SRSF1 depletion in motor neurons derived from C90RF72-ALS patients. Consistent with our previous data presented in Drosophila and neuronal cells models, we also found that depletion of SRSF1 in C90RF72-ALS patient- derived iMNs leads to specific reduction in the expression levels of poly-GP DPRs (Fig. 25b). We next quantified the total, nuclear and cytoplasmic levels of intron-1 -spliced C90RF72 transcripts (using qPCR primers annealing in exon-1 and exon-3) 50,51 and unspliced C90RF72 transcripts retaining intron-1 (using qPCR primers annealing in exon-1 and in intron-1 upstream of the hexanucleotide repeat expansion) 50, 51 to evaluate the potential impact of SRSF1-RNAi on the splicing of intron-1 and on the nuclear export of wild type and pathological C90RF72 transcripts. Depletion of the chromatin-remodeling factor SSRP1 in the cytoplasmic fractions and of actin in the nuclear fractions was used to validate the quality of the subcellular fractionation (Fig. 23d). The relative expression levels of SRSF1 mRNA were down regulated by approximately 80% upon SRSF1-RNAi transduction of iNeurons differentiated from either two control or two sex/age-matched C90RF72-ALS patient lines ( Fig. 32). No significant changes in the total, nuclear or cytoplasmic levels of intron-1 -spliced transcripts were measured between control and C90RF72-ALS iNeurons transduced or not with SRSFI-RNAi lentivirus (Fig. 23e). The mRNA level ratios of SRSFI-RNAi (MOI5) over untreated (MOI0) were further plotted to assess the net impact of the SRSFI -RNAi on each cellular compartment and control or C90RF72-ALS iNeurons (Fig. 23f). These data show that the proportion of exon1-exon3 spliced transcripts is not altered in C90RF72-ALS neurons and that the presence of the hexanucleotide repeat expansion does not affect the splicing of intron-1 , in full agreement with a recent study 50 . In addition, we also show here that the SRSF1-RNAi has no effect on the nuclear export or the splicing of intron-1 in the spliced transcripts that lead to the production of the wild type C90RF72 protein. The same experimental analysis was carried out for the C90RF72 transcripts retaining intron-1 (Fig. 23g and 23h). We were unable to detect significant levels (above non template control qRT- PCR reactions) of intron-1 -retaining C90RF72 transcripts in the cytoplasm of control iNeurons consistent with the nuclear retention of unspliced transcripts. In striking contrast, the presence of the hexanucleotide repeat expansion in C90RF72-ALS patients triggers the nuclear export of C90RF72 repeat transcripts retaining intron-1 (Fig. 23g) consistent with our previous data showing that the sequestration of SRSF1 on synthetic hexanucleotide repeat expansions promotes nuclear mRNA export through the interaction with NXF1 (Fig. 8). Whilst the depletion of SRSF1 did not affect the total level and biogenesis/stability of intron-1 -retaining transcripts in control or C90RF72-ALS iNeurons, it specifically triggers a cytoplasmic reduction and nuclear accumulation of pathological C90RF72 repeat transcripts in C90RF72-ALS iNeurons (Fig. 23g, h). These data demonstrate that the depletion of SRSF1 specifically inhibits the nuclear export of expanded C90RF72 repeat transcripts. Taken together, our data show that the SRSF1 depletion has no effect on the expression level, splicing or nuclear export of wild type spliced exon1-exon3 C90RF72 transcripts while it specifically inhibits the nuclear export of pathological C90RF72 transcripts retaining the hexanucleotide repeats in intron-1.

Example 11

Microsatellite expansions of 3-6 nucleotides in coding and non-coding regions of genes cause neurodegeneration through complex mechanisms involving protein loss-of-function and protein/RNA toxic gain-of-function mechanisms 52 . The production of toxic polymeric repeat proteins by RAN translation has now been characterised in multiple neurodegenerative disorders caused by microsatellite expansions including spinocerebellar ataxia type 8 (SCA8)58, myotonic dystrophy type 1 (DM 1) 53 , Fragile X-associated tremor and ataxia syndrome (FXTAS)59, C90RF72-ALS6, 13-16,27 and Huntington disease (HD)60. However, the mechanisms involved in the nuclear export of these disease-related repeat transcripts are currently unknown.

Example 12

We previously suggested that the sequestration of nuclear export adaptors onto C90RF72 repeat transcripts might trigger the abnormal nuclear export of C90RF72 repeat transcripts and the subsequent RAN translation of DPRs in the cytoplasm12. In this study, we identified for the first time the molecular mechanism driving the nuclear export of pathological C90RF72 repeat transcripts. We investigated whether the partial depletion of two evolutionarily conserved nuclear export adaptors which avidly interact with the hexanucleotide repeat transcripts'! 2, ALYREF and SRSF1 , would mitigate DPR-mediated neurotoxicity in an established Drosophila model of C90RF72-ALS16. We discovered that the partial depletion of SRSF1 prevents in vivo neurodegeneration and suppresses the associated locomotor deficits while the depletion of ALYREF only had marginal effects. The depletion of SRSF1 in C90RF72-ALS patient-derived motor neurons also conferred neuroprotection of motor neurons in co-culture with C90RF72-ALS astrocytes. Moreover, we also showed that this intervention does not affect the morphology or the growth of control and C90RF72-ALS patient-derived motor neurons. On the other hand, the depletion of SRSF1 in patient-derived C90RF72-ALS astrocytes significantly suppressed motor neuron death in a co-culture system. The mechanisms for suppression of astrocyte-mediated neurotoxicity remain however to be determined. They might involve a modification of the RNA or protein composition in the extra-cellular exosomes released by astrocytes.

Example 13

Using neuronal N2A cells, we demonstrated that the nuclear export of C90RF72 repeat transcripts and subsequent RAN translation depends on the interaction of SRSF1 with the nuclear export receptor NXF1. Depleting SRSF1 or inhibiting its endogenous RNA-repeat sequestration and interaction with NXF1 lead to a marked inhibition of the nuclear export of C90RF72 repeat transcripts and RAN translation of sense and antisense DPRs to prevent C90RF72 repeat-mediated neurotoxicity. We also showed that the SRSF1 -dependent inhibition of the nuclear export of C90RF72 repeat transcripts leads to altered production of DPRs in Drosophila and patient-derived motor neurons. Importantly, the depletion of SRSF1 in control or C90RF72-ALS patient-derived neurons does not affect the expression levels or the nuclear export of intron-1 -spliced transcripts required for the translation of the wild type C90RF72 protein. This also indicates that the nuclear export of non-repeat C90RF72 transcripts does not depend on the nuclear export adaptor SRSF1. In sharp contrast to control neurons, the presence of the hexanucleotide repeat expansion in intron-1 of C90RF72 transcripts led to SRSF1 -dependent mRNA nuclear export, while depletion of SRSF1 specifically inhibits the nuclear export but not the levels or splicing of C90RF72 transcripts retaining expanded hexanucleotide repeats in intron-1. Taking these data together, we show that sequestration of SRSF1 onto C90RF72 hexanucleotide repeats is able to license the NXF1 -dependent nuclear export of pathological C90RF72 repeat transcripts without functional coupling of the nuclear export process to pre-mRNA splicing. This explains in turn why the depletion of SRSF1 has no effect on the level, intron-1-splicing or nuclear export of wild type C90RF72 transcripts. Example 14

In conclusion, we have elucidated for the first time the molecular mechanism driving the nuclear export of pathological C90RF72 repeat transcripts which allows for RAN translation of dipeptide repeat proteins in the cytoplasm (Fig. 24a). The depletion of SRSF1 specifically inhibits the nuclear export of the pathologically expanded C90RF72 transcripts without interfering with biogenesis/processing of the wild type C90RF72 transcripts (Fig. 24b). The expression of the engineered SRSF1-m4 protein, which retains specific ability to be sequestered on repeat transcripts but fails to effectively interact with NXF1 , also inhibits the nuclear export of repeat transcripts and the production of DPRs. (Fig. 24c). Both these interventions represent promising prospects for the development of an effective neuroprotective strategy in C90RF72-related ALS. The effects of antagonizing SRSF1 in the vertebrate brain remain to be elucidated in wild-type mice as well as in murine C90RF72- ALS models. Interestingly, it was recently shown that whilst SRSF1 directly binds thousands of transcripts, the depletion of SRSF1 in isolation affects the nuclear export of only 225 transcripts (<1 % transcribed coding genes) due to the presence of 6 additional SRSF factors (SRSF2-7) which act as redundant NXF1-dependent nuclear export adaptors46. The cellular pathways causing C90RF72 repeat-mediated neurodegeneration and the precise mechanism(s) of neuroprotection conferred by the targeting of SRSF1 remain to be elucidated in future studies. Inhibiting the nuclear export of repeat transcripts might also confer neuroprotection in other microsatellite expansion disorders. However, it will remain essential to determine pathophysiological contributions between polymeric repeat protein production and RNA- mediated toxicity by nuclear retention of transcripts and/or sequestration of RNA-processing factors on repeat transcripts. While expression of repeat proteins can kill cells in vitro, it is difficult to evaluate the levels of RAN-translation in patients and the thresholds required for triggering neurotoxicity which will differ depending on the nature of the repeat expansions, the disease in question and the cell types. There is however growing evidence for a pathogenic role of RAN-translation and the data presented here fully support this. For example, FXTAS was initially thought to be caused by intranuclear retention of transcripts and sequestration of splicing factors 54,55 . However, the discovery of RAN translation in the same model challenged this view 56 . Similarly, in C90RF72-ALS, a 10-fold increase in the number of intranuclear RNA foci does not significantly alter survival or global RNA processing, while expression of DPRs caused neurodegeneration 50 in full agreement with the data presented here. Partial depletion of individual nuclear export adaptors does not appear to be detrimental to the functioning of higher eukaryotic cells. Therefore, they might constitute viable therapeutic targets for inhibiting the nuclear export of repeat transcripts and the production of toxic repeat proteins, particularly in neurodegenerative diseases where RAN-translation appears to have a prominent pathological role.

Example 15

Expression of cell permeable or AAV-encoded antagonistic SRSF1 peptides that interact with NXF1

SRSF1 aminoacids 89-120 (SEQ ID NO 15):

PRSGRGTGRGGGGGGGGGAPRGRYGPPSRRSE

We have used the following peptide and showed reduction in the production of DPRs expressed in a disease relevant RAN-dependent manner in human HEK cells transfected with a G4C2x38 repeat constructs that express 3xV5 tags in all frames (Figure 33).

SEQ ID NO 153

SRSF1 sequence (TAP/NXF1 -binding site) V5 tag TAT PTD

PRSGRGTGRGGGGGGGGGAPRGRYGPPSRRSE GG GKPIPNPLLGLDST GG YGRKKRRQRRR

Example 16

Expression of SRPK1 amino-acids 1-655 (SEQ ID NO 19) to maintain phosphorylation state of SRSF1 for inhibiting interaction of SRSF1 with NXF1. Example 17

Mouse models of C90RF72 related ALS/FTD

Ranum's mice (Liu Y et al. Neuron 2016; 90:521-34)

BAC mouse model of C9orf72 ALS/FTD that shows decreased survival, paralysis, muscle denervation, motor neuron loss, anxiety-like behavior, and cortical and hippocampal neurodegeneration. These mice express C90RF72 sense transcripts and upregulated antisense transcripts. This is our first model of choice to test our gene therapy vectors. Cleveland and Lagier-Tourenne's mice (Jiang et al. Neuron 2016; 90:535-50)

Mice expressing C90RF72 RNAs with up to 450 GGGGCC repeats or with one or both

C9orf72 alleles inactivated.

Virally-delivered models (Chew J. Science 2015; 348:1 151-4)

Expression of (G4C2)66 throughout the murine central nervous system by means of somatic brain transgenesis mediated by adeno-associated virus.

Example 18 Routes of administration

Cerebrospinal fluid (CSF) using cisterna magna or intracerebroventricular delivery of AAV vectors AAV-mediated gene delivery will be applied at 2 different time points; pre-onset and onset of symptoms. Transgenic mice (Liu Y et al. Neuron 2016; 90:521-34) will be divided into 2 groups to be treated either therapeutic vector as described above or control viruses [12 μΙ (2 μΙ iodixanol); 1.2 x 1013/ml], via cisterna magna or intracerebroventricular (ICV). 5 mice per group will be sacrificed 4 weeks post-injection and CNS tissue collected to assess viral biodistribution (in spinal cord) and SRF1 mRNA and protein levels (in spinal cord, striatum, brainstem, cerebellum and cortex). The remaining 15 mice per group underwent behavioural testing to assess effects on disease progression including weekly rotarod analysis after three consecutive days of training and gait analysis. In all studies, onset and progression of disease will be assessed by neurological scoring 3 times per week from 60 days of age and mice were also weighed weekly. Mice will be scored for tremor, hind-limb splay and overall neurological deficit using a previously reported scoring system (Mead RJ et al. PLoS One 201 1 ; 6:e23244). All mice continued to end-stage disease and the time to reach this stage will be recorded. All animals will be perfused under terminal anaesthesia and CNS tissue was collected to assess viral biodistribution, SRF1 mRNA and protein levels.

IV delivery

Mice will be injected in the facial or tail vein under isoflurane anaesthesia with 1-5x101 1 vector genome of either therapeutic vectors as described above or control viruses. The mice were then allowed to recover, rolled in the sawdust from their original cage and immediately returned to their cage. Behavioral analysis will be carried using the design described above.

References

1. M. DeJesus-Hernandez et al., Expanded GGGGCC hexanucleotide repeat in noncoding region of C90RF72 causes chromosome 9p-linked FTD and ALS. Neuron. 72, 245-256 (2011).

2. A. E. Renton et al., A hexanucleotide repeat expansion in C90RF72 is the cause of chromosome 9p21 -linked ALS-FTD. Neuron. 72, 257-268 (2011).

3. J. Cooper-Knock et al. , C90RF72 GGGGCC expanded repeats produce splicing dysregulation which correlates with disease severity in amyotrophic lateral sclerosis. PLoS ONE. 10, e0127376 (2015).

4. J. Cooper-Knock et al. , Antisense RNA foci in the motor neurons of. Acta

Neuropathol. 130, 63-75 (2015).

5. K. Mori et al., hnRNP A3 binds to GGGGCC repeats and is a constituent of p62- positive/TDP43-negative inclusions in the hippocampus of patients with C9orf72 mutations. Acta Neuropathol. 125, 413-423 (2013).

6. C. J. Donnelly et al., RNA Toxicity from the ALS/FTD C90RF72 Expansion Is Mitigated by Antisense Intervention. Neuron. 80, 415-428 (2013).

7. D. Sareen et al., Targeting RNA Foci in iPSC-Derived Motor Neurons from ALS Patients with a C90RF72 Repeat Expansion. Sci Transl Med. 5, 208ra149 (2013).

8. Y.-B. Lee et al., Hexanucleotide Repeats in ALS/FTD Form Length-Dependent RNA Foci, Sequester RNA Binding Proteins, and Are Neurotoxic. Cell Rep. 5, 1178-1186 (2013).

9. J. Cooper-Knock et al., Sequestration of multiple RNA recognition motif-containing proteins by C9orf72 repeat expansions. Brain. 137, 2040-2051 (2014).

10. K. Mori et al., The C9orf72 GGGGCC repeat is translated into aggregating dipeptide- repeat proteins in FTLD/ALS. Science. 339, 1335-1338 (2013).

11. K. Mori et al., Bidirectional transcripts of the expanded C9orf72 hexanucleotide repeat are translated into aggregating dipeptide repeat proteins. Acta Neuropathol. 126, 881-893 (2013).

12. P. E. A. Ash et al., Unconventional translation of C90RF72 GGGGCC expansion generates insoluble polypeptides specific to C9FTD/ALS. Neuron. 77, 639-646 (2013).

13. T. Zu et al., RAN proteins and RNA foci from antisense transcripts in C90RF72 ALS and frontotemporal dementia. Proc. Natl. Acad. Sci. U.S.A. 110, E4968-77 (2013).

14. S. Mizielinska et al., C9orf72 repeat expansions cause neurodegeneration in

Drosophila through arginine-rich proteins. Science. 345, 1 192-1194 (2014).

15. S. Ciura et al., Loss of function of C9orf72 causes motor deficits in a zebrafish model of Amyotrophic Lateral Sclerosis. Ann. Neurol. 74:180-187 (2013), doi: 10.1002/ana.23946.

16. J. G. O'Rourke et al. , C9orf72 is required for proper macrophage and microglial

function in mice. Science. 351 , 1324-1329 (2016).

17. M. J. Walsh et al., Invited review: decoding the pathophysiological mechanisms that underlie RNA dysregulation in neurodegenerative disorders: a review of the current state of the art. Neuropathol. Appl. Neurobiol. 41 , 109-134 (2015).

18. J. D. Rohrer et al., C9orf72 expansions in frontotemporal dementia and amyotrophic lateral sclerosis. Lancet Neurol. 14, 291-301 (2015).

19. T. W. Todd, L. Petrucelli, Insights into the pathogenic mechanisms of Chromosome 9 open reading frame 72 (C9orf72) repeat expansions. J Neurochem (2016),

doi: 10.11 11 /jnc.13623.

20. D. Edbauer, C. Haass, An amyloid-like cascade hypothesis for C9orf72 ALS/FTD.

Curr. Opin. Neurobiol. 36, 99-106 (2016).

21. B. D. Freibaum et al., GGGGCC repeat expansion in C9orf72 compromises

nucleocytoplasmic transport. Nature. 525, 129-133 (2015).

22. A. P. Golovanov, G. M. Hautbergue, A. M. Tintaru, L.-Y. Lian, S. A. Wilson, The

solution structure of REF2-I reveals interdomain interactions and regions involved in binding mRNA export factors and RNA. RNA. 12, 1933-1948 (2006).

23. A. M. Tintaru et al., Structural and functional analysis of RNA and TAP binding to SF2/ASF. EMBO Rep. 8, 756-762 (2007).

24. G. M. Hautbergue, M.-L. Hung, A. P. Golovanov, L.-Y. Lian, S. A. Wilson, Mutually exclusive interactions drive handover of mRNA from export adaptors to TAP. Proc. Natl. Acad. Sci. U.S.A. 105, 5154-5159 (2008).

25. N. Viphakone et al., TREX exposes the RNA-binding domain of Nxf1 to enable mRNA export. Nat Commun. 3, 1006 (2012).

26. Y. Huang, R. Gattoni, J. Stevenin, J. A. Steitz, SR splicing factors serve as adapter proteins for TAP-dependent mRNA export. Molecular Cell. 11 , 837-843 (2003).

27. F. Stutz et al., REF, an evolutionarily conserved family of hnRNP-like proteins,

interacts with TAP/Mex67p and participates in mRNA nuclear export. RNA. 6, 638- 650 (2000).

28. G. Dietzl et al., A genome-wide transgenic RNAi library for conditional gene

inactivation in Drosophila. Nature. 448, 151-156 (2007).

29. K. Meyer et al., Direct conversion of patient fibroblasts demonstrates non-cell

autonomous toxicity of astrocytes to motor neurons in familial and sporadic ALS.

Proc. Natl. Acad. Sci. U.S.A. 111 , 829-832 (2014).

30. Y. Huang, T. A. Yario, J. A. Steitz, A molecular link between SR protein

dephosphorylation and mRNA export. Proc. Natl. Acad. Sci. U.S.A. 101 , 9666-9670 (2004).

31. R. Kami et al., The gene encoding the splicing factor SF2/ASF is a proto-oncogene.

Nat. Struct. Mol. Biol. 14, 185-193 (2007).

32. J. C. Greene et al., Mitochondrial pathology and apoptotic muscle degeneration in Drosophila parkin mutants. Proc. Natl. Acad. Sci. U.S.A. 100, 4078-4083 (2003).

33. W. Sullivan, M. Ashburner, R. S. Hawley, Drosophila protocols. (Cold Spring Harbor, NY, Cold Spring Harbor Laboratory Press, 2000).

34. N. Deglon, J. L. Tseng, J. C. Bensadoun, Self-inactivating lentiviral vectors with

enhanced transgene expression as potential gene transfer system in Parkinson's disease. Human Gene Ther. 11 , 179-190 (2000).

35. H. Wichterle, I. Lieberam, J. A. Porter, T. M. Jessell, Directed differentiation of

embryonic stem cells into motor neurons. Cell. 110, 385-397 (2002).

36. D. Ling, P. M. Salvaterra, Robust RT-qPCR data normalization: validation and selection of internal reference genes during post-experimental data analysis. PLoS ONE. 6, e17762 (2011).

36. Webster, C. P. et al. The C9orf72 protein interacts with Rabl a and the ULK1 complex to regulate initiation of autophagy. EMBO J. 35, 1656-1676 (2016).

37. Jovicic, A. et al. Modifiers of C9orf72 dipeptide repeat toxicity connect nucleocytoplasmic transport defects to FTD/ALS. Nat. Neuroses. 18, 1226-1229 (2015).

38. Golovanov, A. P., Hautbergue, G. M., Tintaru, A. M., Lian, L.-Y. & Wilson, S. A. The solution structure of REF2-I reveals interdomain interactions and regions involved in binding mRNA export factors and RNA. RNA 12, 1933-1948 (2006).

39. Tintaru, A. M. et al. Structural and functional analysis of RNA and TAP binding to SF2/ASF. EMBO Rep. 8, 756-762 (2007).

40. Meyer, K. et al. Direct conversion of patient fibroblasts demonstrates non-cell autonomous toxicity of astrocytes to motor neurons in familial and sporadic ALS. Proc. Natl. Acad. Sci. U. S.A. 111, 829-832 (2014).

41. Deglon, N., Tseng, J. L. & Bensadoun, J. C. Self-inactivating lentiviral vectors with

enhanced transgene expression as potential gene transfer system in Parkinson's disease. Human Gene Ther. 11 , 179-190 (2000).

42. Cooper-Knock, J. et al. Sequestration of multiple RNA recognition motif-containing proteins by C9orf72 repeat expansions. Brain 137, 2040-2051 (2014).

43. Cooper-Knock, J. et al. Antisense RNA foci in the motor neurons of. Acta Neuropathol 130, 63-75 (2015).

44 Mizielinska, S. et al. C9orf72 repeat expansions cause neurodegeneration in Drosophila through arginine-rich proteins. Science 345, 1192-1194 (2014).

45. Donnelly, C. J. et al. RNA Toxicity from the ALS/FTD C90RF72 Expansion Is

Mitigated by Antisense Intervention. Neuron 80, 415-428 (2013).

46. Sareen, D. et al. Targeting RNA Foci in iPSC-Derived Motor Neurons from ALS

Patients with a C90RF72 Repeat Expansion. Sci Transl Med 5, 208ra149 (2013).

47. Masuda, S. et al. Recruitment of the human TREX complex to mRNA during splicing.

Genes & Development 19, 1512-1517 (2005).

48. Chang, C.-T. et al. Chtop is a component of the dynamic TREX mRNA export

complex. EMBO J. 32, 473-486 (2013).

49. Hautbergue, G. M. et al. UIF, a New mRNA export adaptor that works together with REF/ALY, requires FACT for recruitment to mRNA. Curr. Biol. 19, 1918-1924 (2009). 50. Tran, H. et al. Differential Toxicity of Nuclear RNA Foci versus Dipeptide Repeat Proteins in a Drosophila Model of C90RF72 FTD/ALS. Neuron 87, 1207-1214 (2015).

Niblock, M. et al. Retention of hexanucleotide repeat-containing intron in C9orf72 mRNA: implications for the pathogenesis of ALS/FTD. Acta Neuropathol Commun 4, 18 (2016).

Loureiro, J. R., Oliveira, C. L. & Silveira, I. Unstable repeat expansions in

neurodegenerative diseases: nucleocytoplasmic transport emerges on the scene. Neurobiol. Aging 39, 174-183 (2016).

Zu, T. et al. Non-ATG-initiated translation directed by microsatellite expansions. Proc. Natl. Acad. Sci. U.S.A. 108, 260-265 (2011).

Jin, P. et al. Pur alpha binds to rCGG repeats and modulates repeat-mediated neurodegeneration in a Drosophila model of fragile X tremor/ataxia syndrome. Neuron 55, 556-564 (2007).

Sofola, O. A. et al. RNA-binding proteins hnRNP A2/B1 and CUGBP1 suppress fragile X CGG premutation repeat-induced neurodegeneration in a Drosophila model of FXTAS. Neuron 55, 565-571 (2007).

Todd, P. K. et al. CGG repeat-associated translation mediates neurodegeneration in fragile X tremor ataxia syndrome. Neuron 78, 440-455 (2013).

Greene, J. C. et al. Mitochondrial pathology and apoptotic muscle degeneration in Drosophila parkin mutants. Proc. Natl. Acad. Sci. U.S.A. 100, 4078-4083 (2003). Sullivan, W., Ashburner, M. & Hawley, R. S. Drosophila protocols. (Cold Spring Harbor, NY, Cold Spring Harbor Laboratory Press., 2000).

Deglon, N., Tseng, J. L. & Bensadoun, R. S. Self-inactivating lentiviral vectors with enhanced transgene expression as potential gene transfer system in Parkinson's disease. Human Gene The. 11, 179-190 (2000).

Wichterle, H., Lieberam, I., Porter, J. A. & Jessell, T. M. Directed differentiation of embryonic stem cells into motor neurons. Cell 110, 385-397 (2002).

Ye, J. et al. Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics 13, 134 (2012).

Ling, D. & Salvaterra, P. M. Robust RT-qPCR data normalization: validation and selection of internal reference genes during post-experimental data analysis. PLoS ONE 6, e17762 (2011).

Yang, D. et al. FTD/ALS-associated poly(GR) protein impairs the Notch pathway and is recruited by poly(GA) into cytoplasmic inclusions. Acta Neuropathol (2015).

doi: 10.1007/s00401 -015-1448-6

Maranda, B., Fan, L, Soucy, J.-F., Simard, L. & Mitchell, G. A. Spinal muscular atrophy: clinical validation of a single-tube multiplex real time PCR assay for determination of SMN1 and SMN2 copy numbers. Clin. Biochem. 45, 88-91 (2012).