RNA interference therapy for the Spinocerebellar ataxias

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Spring 2014 RNA interference therapy for the Spinocerebellar ataxias Pavitra Shyam Ramachandran University of Iowa Copyright 2014 Pavitra Ramachandran This dissertation is available at Iowa Research Online: Recommended Citation Ramachandran, Pavitra Shyam. "RNA interference therapy for the Spinocerebellar ataxias." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Genetics Commons

2 RNA INTERFERENCE THERAPY FOR THE SPINOCEREBELLAR ATAXIAS by Pavitra Shyam Ramachandran A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Genetics in the Graduate College of The University of Iowa May 2014 Thesis Supervisor: Professor Beverly L. Davidson

3 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL PH.D. THESIS This is to certify that the Ph.D. thesis of Pavitra Shyam Ramachandran has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Genetics at the May 2014 graduation. Thesis Committee: Beverly L. Davidson, Thesis Supervisor Paul B. McCray Arlene V. Drack Michael G. Anderson Pedro Gonzalez-Alegre

4 To my family ii

5 ACKNOWLEDGMENTS I would like to thank my mentor Beverly Davidson for giving me the opportunity to work in her laboratory. She has inspired and motivated me through the years and will continue to do as a role model. She has always encouraged and fully supported my ideas, helping me build confidence in my decisions. I will always be indebted to her for shaping my career as a scientist. I would like to thank Dr. Arlene Drack for her mentorship and collaboration on the retinal project. She has always taken time out of her busy schedule to accommodate our experiments and encourage our ideas. I would like to thank Sajag Bhattarai from the Drack laboratory who worked tirelessly on the retinal project with us. I am grateful to the Davidson laboratory personnel for their support and advice during lab meetings and otherwise. They make the Davidson lab a happy and fun place to work. I would especially like to thank Ryan, Megan, Alex and Luis for helping and guiding me with my projects. I would also like to thank Matt Sowada, Kellie and Matt Wilson for helping me with my experiments. I would like to thank my committee members Drs. McCray, Drack, Anderson, Gonzalez- Alegre for their encouragement and guidance over the years. Thank you to the Genetics program, Dr. Eberl, Anita Kafer, Linda and Isabelle for making my life as a graduate student smooth sailing. My family is my support system and backbone. They have always encouraged me to aim high and work hard. I owe this thesis to them. To my husband Shyam and our little miracle, Keshav, I couldn t have done this without you. iii

6 ABSTRACT The spinocerebellar ataxias are a group of diseases characterized by loss of motor coordination. Spinocerebellar ataxia types 2 and 7 are monogenic, autosomal dominant, lateonset neurodegenerative diseases characterized by ataxia with no effective treatments in the clinic. The most striking feature of these diseases is the degeneration of Purkinje neurons of the cerebellum. Spinocerebellar ataxia type 7 is also characterized by vision loss due to degeneration of the retinal photoreceptors. In this work, we tested the hypothesis that reducing mutant gene expression by RNAi would alleviate disease phenotypes in these two spinocerebellar ataxias. For spinocerebellar ataxia type 7 (SCA7), we designed and tested RNAi sequences that could reduce the expression of both wildtype and mutant ataxin-7, an approach that would be applicable to all SCA7 patients. We found that AAV1-mediated delivery of a candidate RNAi sequence to the Purkinje neurons of SCA7 mice resulted in long-term sustained reduction of both wildtype and mutant ataxin-7 and resulted in significant improvements in ataxic and neuropathological phenotypes. We also delivered the RNAi sequence (AAV1-mediated) to reduce the expression of both mutant and wildtype ataxin-7 in the SCA7 mouse retina and evaluated retinal function long-term. We observed a preservation of normal retinal function and no adverse toxicity with reduction of wildtype and mutant ataxin-7 alleles. These studies address an important safety concern regarding non-allele specific silencing of ataxin-7 for SCA7 therapy. To identify therapies for spinocerebellar ataxia type 2 (SCA2), we designed and tested several RNAi sequences to reduce the expression of both wildtype and mutant ataxin-2 in vitro and in vivo. We found that reduction of wildtype ataxin-2 expression in the mouse cerebellum was tolerated 4 months post injection without inducing behavioral deficits or cerebellar pathology. Additionally, we tested other sequences for improved silencing efficacy, and iv

7 identified a potent RNAi sequence that significantly reduced the expression of both mutant and wildtype ataxin-2 in the cerebellum of a SCA2 mouse model. Ongoing work will establish if long-term reduction of both mutant and wildtype ataxin-2 will provide therapeutic benefit in the SCA2 mouse setting, and the safety of this sequence in normal cerebella. v

8 TABLE OF CONTENTS LIST OF TABLES... ix LIST OF FIGURES... x LIST OF ABBREVIATIONS... xii CHAPTER I: INTRODUCTION... 1 Spinocerebellar ataxia type 7 (SCA7)... 2 Ataxin-7 function... 3 SCA7 Mouse Models... 4 Spinocerebellar ataxia type 2 (SCA2)... 6 Ataxin-2 function... 8 Sca2 mouse models... 8 Therapeutic approaches for SCA7 and SCA RNAi as a tool for directed gene knockdown Designing therapeutic RNAi molecules Delivery of therapeutic RNAi molecules Viral delivery to the CNS Objectives Contributions to work presented in this thesis CHAPTER II: NON-ALLELE SPECIFIC SILENCING OF ATAXIN-7 IMPROVES PHENOTYPES IN A MOUSE MODEL OF SPINOCEREBELLAR ATAXIA TYPE Abstract Introduction Results Silencing ataxin-7 expression mis4 improves motor phenotypes Reduction of ataxin-7 in the pcs does not induce toxicity mis4 has a therapeutic impact on the SCA7 cerebellum Discussion Materials and methods vi

9 Animals Viral vectors AAV injections and tissue harvesting Immunohistochemistry and western blotting RT-qPCR and stem-loop PCR Behavior analysis CHAPTER III: RNA INTERFERENCE-BASED THERAPY FOR SPINOCEREBELLAR ATAXIA TYPE 7 RETINAL DEGENERATION Abstract...45 Introduction Results Silencing ataxin-7 expression in the sca7 retina Sustained silencing of ataxin-7 in the sca7 retina Reducing ataxin-7 expression does not induce neuropathology in the SCA7 mouse retina Retinal function is not altered in mis4 treated retinas Discussion Materials and methods Animals Viral vectors AAV injections Tissue harvesting and histology In situ hybridization RT-qPCR and stem-loop PCR Retinal assays CHAPTER IV: IDENTIFICATION OF A POTENT RNAI SEQUENCE FOR SPINOCEREBELLAR ATAXIA TYPE 2 THERAPY Abstract Introduction Results Reducing ataxin-2 expression in vitro and in vivo Reduction of wildtype ataxin-2 expression in the cerebellum does vii

10 not affect motor coordination or memory Sustained reduction of wildtype ataxin-2 expression in the cerebellum does not induce neuropathology Screening additional RNAi sequences with higher potency Testing mis7 and mis12 sequences in a SCA2 mouse model Discussion Materials and methods Animals Viral vectors AAV injections Tissue harvesting and immunohistochemistry RT-qPCR Behavior analyses CHAPTER V: FUTURE DIRECTIONS SCA7 therapy Targeting the SCA7 brain Targeting the SCA7 retina Gene expression changes in mis4 treated SCA7 cerebellum and retina SCA2 therapy Moving RNAi therapy towards the clinic REFERENCES viii

11 LIST OF TABLES Table 1: SCA7 mouse models Table 2: Progress in RNAi therapeutics demonstrating therapeutic potential in cell and animal models Table 3: List of mirna sequences targeting ataxin ix

12 LIST OF FIGURES Figure 1: Co-opting the microrna pathway for delivery of RNAi triggers to the CNS and retina Figure 2: Reduction of ataxin-7 mrna in vitro and in vivo Figure 3: Sustained reduction of ataxin-7 expression in the SCA7 mouse cerebellum. 41 Figure 4: mis4 mice demonstrate significant improvement in ataxia phenotypes Figure 5: Histological and QPCR analysis of cerebellar tissue at 40 weeks Figure 6: Changes in cerebellar transcripts post-injection at 40 weeks Figure 7: Reduction of ataxin-7 mrna in the SCA7 mouse retina Figure 8: Sustained suppression of ataxin-7 expression Figure 9: Assessing toxicity post-injection at 30 weeks Figure 10: Retinal function 30 weeks post-injection Figure 11: In situ hybridization Figure 12: Reduction of mouse and human ataxin-2 expression in vitro Figure 13: Reduction of rhesus ataxin-2 mrna levels Figure 14: Microglia expression in rhesus cerebellum observed with Iba-1 antibodies Figure 15: Behavior assays at 4 and 6 months of age Figure 16: Reduction in mouse ataxin-2 mrna levels at 6 months of age Figure 17: Preservation of cerebellar morphology in mis4 and mis5 treated cerebella.84 x

13 Figure 18: Absence of gliosis in the transduced cerebellum Figure 19: Microglial expression in the cerebellum Figure 20: Screen of additional RNAi sequences in HeLa cells Figure 21: Reduction of mouse ataxin-2 mrna in wildtype cerebellum Figure 22: Reduction of mutant (human) and wildtype (mouse) ataxin-2 in the SCA2 mouse cerebellum Figure 23: Allele-specific silencing of ataxin Figure 24: SCA7 mice show deficits in cone-erg response Figure 25: Experimental design for long-term SCA2 studies proposed xi

14 LIST OF ABBREVIATIONS AAV: Adeno-associated virus Atxn7: ataxin-7 Atxn2: Ataxin-2 ALS: Amyotrophic lateral sclerosis BAC: Bacterial artificial chromosome BG: Bergmann glia CNS: Central nervous system DCN: Deep cerebellar nucleus ERG: Electroretinogram GFAP: Glial fibrillary acidic protein HAT: Histone acetyl transferase HD: Huntington disease IHC: Immunohistochemistry IO: Inferior olive LV: Lentivirus mirna: microrna NI: Nuclear inclusions OCT: Optical coherence tomography PC: Purkinje cell PolyQ: Polyglutamine Prp: Prion promoter RNAi: RNA interference xii

15 SCA: Spinocerebellar ataxia shrna: small hairpin RNA sirna: Small interfering RNA xiii

16 1 CHAPTER I: INTRODUCTION The spinocerebellar ataxias (SCA) are a heterogeneous group of inherited neurodegenerative diseases that are characterized primarily by ataxia (1). Ataxia, a loss of motor control, can be caused by degeneration of the cerebellum, spinocerebellar tracts and brainstem. SCA patient symptoms can include areflexia, oculomotor deficits, cognitive impairment and macular degeneration among many others (1). Clinical features of the SCAs can also include symptoms of other diseases such as Parkinsonism, amyotrophic lateral sclerosis, dystonia and depression (1). Currently, there are no disease-modifying treatments for the SCAs. The term SCA is used to describe the autosomal dominant nature of these diseases (1) and there are approximately 30 different SCAs described to date. The incidence varies, they have overlapping symptoms, and the causative gene(s) has only been identified for approximately half of these diseases, making clinical diagnosis difficult for the various SCAs (2, 3) (4). Six of the SCAs (SCA 1, 2, 3, 6, 7 and 17) are classified as polyglutamine (polyq) repeat diseases and these patients constitute the majority of the SCAs. These diseases are caused by an expansion of CAG in the underlying gene, which encodes an expanded glutamine repeat. The polyq expanded protein attains a toxic gain of function and accumulates as aggregates in affected and non-affected tissues. Here, we focus on two of the polyq diseases, spinocerebellar ataxia type 7 (SCA7) and spinocerebellar ataxia type 2 (SCA2).

17 2 Spinocerebellar ataxia type 7 (SCA7) Spinocerebellar ataxia type 7 (SCA7) is an autosomal dominant polyq expansion disease, where the expansion of CAG repeats is in the coding region of the ataxin-7 gene. SCA7 is characterized by cerebellar ataxia (lack of coordination, imbalance) and moderate to profound vision loss due to retinal degeneration, with vision loss often being the first reported symptom (5). Vision loss is unique to SCA7 and is not seen in other polyq diseases. With disease progression, symptoms may also extend to include ophthalmoplegia, dysarthria, dysphagia and abnormal reflexes (6, 7). Normal individuals have ~10 CAG repeats in the coding region of ataxin-7, while an expansion of greater than 37 repeats results in SCA7. Patients with an expansion of CAG repeats generally show cerebellar neurodegeneration initially, while those with an expansion of >59 CAG repeats often also present with retinal degeneration (5). SCA7 is typically a late-onset disease, however, anticipation is a feature of the disease with increasing polyq expansions causing an earlier onset of the disease within generations of an affected family (7). Early onset SCA7 is severe and progresses rapidly, causing death early in life (8). The National Ataxia Foundation has reported the incidence of SCA7 to be <1 in 100,000 people, with SCA7 representing 2% of all SCAs (9). SCA7 has a wide geographical distribution and in certain populations there is a higher incidence of the disease. In South African populations it is the second most prevalent ataxia (10). SCA7 brain pathology is often referred to as olivoponto cerebellar atrophy, as marked atrophy of the inferior olive, pons and cerebellum is seen in brain MRIs. Neuronal loss in the inferior olive, brainstem and Purkinje cells (PC) of the cerebellum

18 3 along with gliosis is observed, and there is also notable demyelination of pyramidal tracts (11). The retinal degeneration is a cone-rod dystrophy, which may manifest as a maculopathy when the onset is in adulthood, or as a diffuse retinopathy or geographic atrophy with childhood onset (12). Typically a decrease in visual acuity is noted, followed by loss of peripheral vision and ultimately complete blindness in many cases (13). By funduscopy, granular pigmentation of the macula is seen along with pigmentary atrophy. As the disease progresses, a loss of photoreceptors and ganglion cells along with a thinning of the outer and inner plexiform layers is observed (11). Nuclear inclusions are characteristic of SCA7, similar to other polyq diseases. The mutant polyq expanded protein accumulates as nuclear inclusions (NI) in PCs, nonaffected areas of the brain and in retinal cells (14). PolyQ proteins aggregate when the expansion of glutamines exceeds a certain threshold, in the case of SCA7, greater than 37. These polyq protein aggregates contain components of the ubiquitin proteasome system and transcription factors in addition to ataxin-7[chou, 2010 #7]. Ataxin-7 function Ataxin-7 is an 892 amino acid protein. Ataxin-7 is expressed in all neurons of the central nervous system (CNS) with high expression in the cerebellar Purkinje cells (PCs) in the brain and expression throughout the retina (14). Ataxin-7 is a core component of the transcriptional co-activator complexes STAGA (Spt3-Taf9-Ada-Gcn5- acetyltransferase), TFTC (TBP-free TAF containing complex) and PCAF (15). These coactivator complexes are recruited to promoter or enhancer regions and promote the

19 4 transcription of target genes via GCN5 - mediated histone acetylation (16). Gcn5 loss exacerbated SCA7 disease phenotypes as seen in the Gcn5 - / - ; SCA7 mice, suggesting that loss of GCN5 activity contributes to disease pathogenesis and toxicity (17). Mutant ataxin-7 disrupts the function of the STAGA complex and induces cerebellar and retinal dysfunction by affecting the transcription of many key genes (18, 19). The genes affected include transcription factors and glutamine transporters, such as CRX in the retina and GLAST in the cerebellum (20, 21). Ataxin-7 shuttles between the nucleus and cytoplasm and has distinct functions in both compartments. In the nucleus it activates gene transcription by associating with one of the three Gcn5 containing complexes, while in the cytoplasm it stabilizes the cytoskeleton by associating with the microtubule network (22). Ataxin-7 transcription can be regulated in cis by a convergently transcribed, non-coding RNA called SCAANT1, which in turn, is regulated by the transcriptional repressor CTCF (CCCTC binding factor) (23). In earlier work, La Spada and colleagues found that a loss of SCAANT1 expression resulted in a de-repression of sense ataxin-7 transcription. More recently, Duncan et al found that wildtype and polyq expanded ataxin-7 interacted with the histone deacetylase protein HDAC3, and also that HDAC3 expression was increased in SCA7 (24). Increased HDAC3 expression caused increased cellular toxicity in vitro, suggesting that HDAC3 may contribute to the neurotoxicity in SCA7. SCA7 mouse models A number of SCA7 mouse models have been generated using different promoters (Table 1). Early models that were made include the Purkinje cell specific P7E and rod

20 5 cell specific R7E models. While the R7E is a good retina specific model, the P7E model, which expresses mutant ataxin-7 (90Q) from a PC specific promoter, did not have a robust ataxia phenotype with late onset of disease at 11 months (25). In my work, I used the BAC-PrP-SCA7-92Q (BAC-SCA7) mouse model generated in the lab of Dr. Albert La Spada s laboratory, University of California, San Diego, a collaborator on my future work. This transgenic mouse expresses the human ataxin-7 cdna with 92 pathogenic repeats from the prion promoter (PrP) (26). The prion promoter promotes ubiquitous expression of the transgene. The transgene in this model is floxed, so as to allow excision with a Cre endonuclease. A number of features of the BAC-SCA7 model recapitulate the human SCA7 ataxia condition and pathology. The BAC-SCA7 mouse displays progressive ataxia with onset at approximately 20 weeks of age and is severe by weeks. Motor coordination, which can be scored using the rotarod, the ledge test, clasping analysis, and gait and activity assays are all affected in the BAC-SCA7 mice starting at ~20 weeks and progressively worsening with age (26). The characteristic NIs are seen throughout the brain by immunohistochemistry (IHC). In addition to thinning of the cerebellar molecular layers, the Bergmann glia (BG) processes are lost and aberrant climbing fiber-pc synapses are observed, reflecting the pathology of the inferior olive in this model. This model provides the unique ability to study the consequences of excising mutant ataxin-7 temporally and spatially. Albert La Spada s group generated mice that lacked the expression of mutant ataxin-7 specifically in the PCs and inferior olive (IO), the BG and PCs by crossing the BAC-SCA7 mice to specific Cre lines (26). Excision of mutant ataxin-7 from the PCs and IO resulted in partial amelioration of motor

21 6 phenotypes, and prevented molecular layer thinning and BG degeneration. Excision from the BG alone did not prevent ataxia or pathology, while excision from the PCs, IO and BG delayed disease onset and had a robust affect on both motor phenotypes and pathology. A subsequent study from the La Spada group showed that excision of the mutant transgene by ~50% just after disease onset could still ameliorate motor phenotypes (27). Spinocerebellar ataxia type 2 Spinocerebellar ataxia type 2 (SCA2) was originally identified in India in 1971 where patients displayed cerebellar ataxia and slow saccadic eye movements (28). The saccades and smooth-pursuit eye movements are controlled by the vestibulocerebellum and the cerebellar vermis (29). The PCs in the lateral vestibulocerebellum project to the medial vestibular nucleus to control eye movements. Lesions in these regions result in eye movement deficits suggesting that these cerebellar regions control eye movement. SCA2 was later characterized as olivopontocerebellar atrophy. SCA2 is characterized by a variety of symptoms including gait ataxia, dysarthria, slow saccadic movements, dysphagia, peripheral neuropathy and abnormal reflexes (30, 31). Organ systems such as gastrointestinal, cardiac and olfactory can also be affected. Furthermore, mutations in the SCA2 gene have been associated with Parkinsonism and amyotrophic lateral sclerosis (ALS), which can complicate diagnosis (32, 33). SCA2 is caused by CAG repeat expansions of greater than 34 CAG repeats in the coding region of Ataxin-2. Onset is usually in adulthood, but similar to other polyq diseases, larger polyq tract expansions cause earlier disease onset of (34). SCA2 is the

22 7 second most prevalent ataxia (15%) after SCA3, and is the most prevalent ataxia in certain geographical regions such as Mexico and Cuba (35, 36). Interestingly, intermediate CAG repeat lengths of in Ataxin-2 are associated with an increased risk for ALS (37). Tar DNA binding protein (TDP-43) mutations are also associated with ALS, and TDP-43 cytoplasmic and nuclear aggregates are a hallmark of ALS. Intermediate length ataxin-2 interacts with TDP-43 and enhances TDP-43 toxicity in culture. Ataxin-2 expanded repeats are also found in Parkinson s and Alzheimer s disease patients, although not at as high a frequency as those with ALS (38). The N terminus of both wildtype and mutant ataxin-2 interact with parkin, an E3 ubiquitin ligase implicated in Parkinson s disease (39). SCA2 is characterized by atrophy of the cerebellum and brainstem and shows a similar pattern of atrophy as seen in SCA1 patients. In particular, dendritic thinning and marked degeneration of the PCs, atrophy of the pons and olivary nucleus are evident with disease progression (40). Degeneration of the thalamus and myelin loss has also been observed in SCA2 patient brains. Neuronal loss was also accompanied by gliosis as evidenced by increased GFAP expression. While many post-mortem SCA2 studies reported no observation of inclusions (41), cytoplasmic and nuclear inclusions were observed in two SCA2 patients with a widely used anti-cag repeat specific antibody, the IC2 antibody (40). In that work, neuronal inclusions were found in many regions of the CNS, and the inclusions were also immunoreactive with anti-ubiquitin antibodies. Cytoplasmic aggregates typically predominate in SCA2 patient brains, with no large inclusion bodies or nuclear inclusions (41, 42).

23 8 Ataxin-2 function Ataxin-2 is expressed widely in the nervous system with high expression in the PCs and interestingly, Ataxin-2 is upregulated in SCA2 brains (41). Ataxin-2 is a 140kDa protein and is highly basic in nature. It is a cytoplasmic protein and has been implicated in a wide range of functions including RNA metabolism. Evidence for this comes from studies showing that Ataxin-2 interacts with PABP via a PAM2 motif at the C terminal end, and with an RNA splicing regulator called ataxin-2 binding protein (A2BP) (43). Ataxin-2 has an LSm (like-sm) domain, which has been implicated in RNA splicing and mrna decay (44). Studies in C. elegans also support the notion that ataxin-2 functions in mrna translation (45). Ataxin-2 also plays a role in assembling stress granules and P bodies (46). In Drosophila, ataxin-2 functions in the mirna pathway to repress specific mrnas (47). The SCA2 knockout model shed light on yet another function of ataxin-2 in regulating body weight and metabolism. Ataxin-2 has also been implicated in actin filament formation, endocytosis and epidermal growth factor receptor trafficking(48). Mouse models of SCA2 SCA2-58Q transgenic mice were generated in the lab of Dr. Stephan Pulst, University of Utah, a collaborator on my future work. The Pulst lab expressed a human mutant ataxin-2 cdna with 58 CAG repeats under the control of a Purkinje cell specific promoter (Pcp2) (42). This mouse displayed late-onset progressive ataxia with rotarod and stride length deficits at 26 weeks. Progressive Purkinje cell loss, dendritic thinning and reduction in molecular layer thickness were also observed. Diffuse cytoplasmic staining of mutant ataxin-2 with obvious nuclear inclusions (42). An increase in calcium

24 9 release from ER stores was observed in PCs cultured from SCA2-58Q mice. When the mice were treated with the calcium stabilizer Dantrolene, motor phenotypes were alleviated along with a rescue of PC cell death (49). A second mouse model was recently generated by the Pulst lab to recapitulate early onset SCA2 (50). The mutant transgene is expressed by the Pcp2 promoter and display an early onset of the disease. SCA2-127Q transgenic mice displayed rotarod deficits starting at 4 weeks, which deteriorated further with age. Molecular layer thinning and PC loss along with perinuclear aggregates were observed later in the disease. In addition, there was a decline in the PC firing rate and a decline in cerebellar gene expression, which preceded the observed motor decline. As noted, both of the above described mouse models have PC specific mutant ataxin-2 transgene expression. A mouse model that recapitulates the human SCA2 early onset condition more closely expresses human mutant ataxin-2 75Q under the control of the endogenous SCA2 promoter (51). PC specific degeneration and early onset motor deficits are observed in this model. These mouse models provide a SCA2 setting in which to study the disease and test therapeutic molecules. Therapeutic approaches for SCA7 and SCA2 PolyQ expanded proteins can induce transcriptional dysregulation, oxidative damage, trafficking defects, excitotoxicity, mitochondrial dysfunction and apoptosis. Thus, current drug-based strategies to treat polyq expansion disorders include but are not limited to; chaperone delivery to target mutant protein misfolding; histone deacetylase

25 10 inhibitors to correct transcriptional dysregulation; inhibiting apoptosis to prevent cerebellar neuronal atrophy; reducing intranuclear aggregate formation (52). These drugbased strategies have shown promise in in vivo models of polyq diseases, for example, in Huntington s disease (HD) (53, 54). However, what is most desirable is to treat the disease directly by targeted silencing of the mutant gene using RNA interference (RNAi). As my work demonstrates proof-of-principle RNAi therapies for SCA2 and SCA7, I discuss strategies for RNAi therapy and delivery in this chapter. RNAi as a tool for directed gene knockdown RNAi is an innate essential mechanism to regulate gene expression. RNAi makes use of double stranded RNA molecules to silence the expression of an mrna molecule by complementary base pairing (55). The double stranded RNA molecules that occur naturally in the cell are micrornas (mirs). These are transcribed as stem loop structured primary mirs that are cleaved by the Drosha-DGCR8 complex in the nucleus to form precursor-mirs (pre-mirs). The pre-mirs are then exported out of the nucleus to be processed by the Dicer containing complex in the cytoplasm into short mature mir sequences (21 nucleotides) that are then loaded onto RISC (RNA induced silencing complex) to carry out translational repression. RNAi in recent years has rapidly evolved as a tool for directed gene silencing. Scientists can hijack the RNAi machinery in many ways to achieve targeted gene silencing, one of which is the viral delivery of short hairpin RNAs (shrnas) or artificial micrornas (mirs) (Figure 1). ShRNA delivery has recently been shown to be potent but toxic to the cell, while artificial mirs were found to be safer and less toxic (56, 57). Viral

26 11 delivery of these artificial mirs has shown therapeutic benefit in mouse models of SCA1 and HD (56, 58). This earlier work in our laboratory showed that RNAi-based gene therapy could provide substantial therapeutic benefit for autosomal dominant polyq diseases. Although non-allele specific silencing has shown therapeutic benefit in HD mouse models, whether patients will tolerate partially reduction of both alleles is a matter of debate (59). For SCA7, neither allele-specific nor non-allele specific approaches have been investigated in vivo. It is unclear how ataxin-7 functions, how poly-q expanded ataxin-7 causes disease, and importantly, if partial reduction of ataxin-7 would be tolerated (RNAi rarely reduces 100% of its target). As there is no knockout model for SCA7, the consequences of knocking down 100% ataxin-7 remain to be tested. On the other hand, SCA2 knockout mice are viable (60) suggesting that partial reduction of ataxin-2 may be well tolerated in SCA2 patients. Designing therapeutic RNAi molecules Designing an sirna sequence to reduce target gene expression with efficacy and specificity is one of the key factors to achieve successful RNAi. Important steps in designing an efficient sirna have been described by several groups (61-63). Two important criteria include designing sequences for proper strand biasing and minimizing off-targeting. To favor antisense strand incorporation into RISC, the antisense strand of the sirna should have strong G-C base pairing at the 3 end and weak base pairing (A-U or G-U) at the 5 end, as RISC loads the strand with the lowest 5 thermodynamic stability (64). There are many online tools and guidelines that help to design sirna sequences to

27 12 a gene target of interest (65-68). Newer online tools also incorporate the very important aspect of sirna off-targeting (69-71). Off-targeting is a phenomenon by which an sirna binds to and represses unintended targets due to complementarity with the sirna seed sequence. Fedorov et al. in 2006 reported toxic effects due to off-target effects of sirnas (72). As mirnas primarily target the 3 UTRs of genes, it was found that seed complementarity to hexamers in the 3 UTR s of genes is proportional to the number of off-target effects (73, 74). Limiting off-targeting is especially critical when designing sirnas for therapeutic purposes. Our lab has designed a web-based program that designs highly specific sirna sequences to a target of interest by taking into consideration the off-targets of an sirna seed sequence, and gives each sirna a score depending on the number of potential offtargets (POTS) (69). Researchers can thus pick sirna sequences with low POTS and screen them in vitro to identify candidate sequences for delivery in vivo. Alternatively, the sirna sequence can be incorporated into a mirna or shrna backbone for delivery in vitro and in vivo. The process of designing and screening of hairpin-based RNAi sequences (shrnas and artificial mirnas) has been described in detail previously (75). Delivery of therapeutic RNAi molecules RNAi delivery to the CNS is challenging. The blood-brain barrier (BBB) or the blood-retina barrier obstacle must be overcome for effective RNAi delivery via the blood, while for direct injection, steps must be taken to avoid toxic or inflammatory reactions. Ideally, delivery of an RNAi molecule to the CNS should be minimally immunogenic, non-toxic, target specific cells, knockdown the specific target mrna efficiently and be

28 13 easy to manufacture (76). Two major types of delivery systems differing in production, safety and efficacy are viral and non-viral delivery systems. While non-viral methods have been used in vivo, they are generally less efficacious and are by nature transient, requiring repeated delivery. This transient nature can be advantageous for therapies that do not require long-term treatment, such as in antitumor and anti-viral therapies. In addition, this provides an important safety measure; in the case of adverse side effects, treatment can simply be terminated. While methods to improve the efficacy of non-viral molecules are currently under development, viral vectors have been used successfully from mouse models to human studies in gene replacement strategies. Viral Delivery to the CNS There are a number of viral vectors that can be used for gene delivery to the CNS as discussed in a prior review (76). The two main viral vector systems that are used to transduce the CNS are lentiviruses (LV) and adeno-associated viruses (AAV). Both viruses are minimally immunogenic and can transduce a number of CNS cell types. Recombinant lentiviruses are pseudotyped with various glycoproteins which can impart different tropisms after directed delivery into brain (77, 78) and they have been used successfully in gain of function studies (79) and loss of function studies (80-83). One difference between lentivirus and AAV or Ad-based systems is the level of expression. This is due in part because lentivirus-mediated transduction often results in low copy numbers of transgene/cell. Also, the placement of the expression cassette in the lentivirus genome can affect expression levels (84). Most lentivirus vectors integrate

29 14 unless the integrase activity has been inactivated. As integrase deficient vectors often have low titers compared to their integrase competent counterparts, their production for use for therapeutic applications is impractical. Integration competency for CNS applications may be less of an issue than in the setting of stem cell transduction (most cells in the CNS are not dividing), where integration and activation of an oncogenic gene provides a growth advantage for the transformed cell (85, 86). AAV belongs to the genus Dependovirus and in its wild type state requires a helper virus, such as Adenovirus, to replicate. A number of factors make AAV suitable for gene delivery in vivo. AAV can be easily manufactured and it is scalable for human use (87, 88); particularly for the relatively low volumes needed for retina- and brainexpressed targets. Additionally, AAV rarely integrates into the genome. In general, AAVs are non-pathogenic and have low immunogenic properties, which make them ideal for gene delivery in vivo (76). AAVs confer robust expression, efficiently transduce neurons and other cell types, and in the absence of an immune response to what is being expressed, can afford long-term expression (89-91). Tissue tropism of AAV is dictated by the capsid serotype. AAV capsids with different cell/tissue tropisms have been identified and depending on the capsid serotype, AAV can transduce neurons, astrocytes, glia and ependymal cells of the brain and retina with high transduction efficiency (92-98). The AAV capsid can be modified to alter its tropism by several ways including directed evolution, capsid shuffling and incorporation of targeting peptides. Directed evolution involves mutagenesis of the capsid, which may alter tropism (99, 100). Capsid shuffling involves the assembly of variant capsid sequences to give rise to recombinant capsids with tropisms to different cell types (101-

30 15 103). AAV tropism can also be altered by the incorporation of a targeting ligand into the capsid, to mediate ligand specific receptor binding ( ). Gene transfer after direct delivery of AAV vectors by intraparenchymal, intraventricular or intrathecal injections to target cells of the brain or, sub-retinal or intravitreal injections to target retinal cells is used for RNAi delivery. Direct delivery to the brain by intra-parenchymal injections has proven effective for targeting neurons in various neurodegenerative diseases, and it limits transduction to those tissues most relevant to disease. Widespread exposure of a transgene product occurs after intraventricular or intrathecal delivery of AAVs, when the transgene product is a secreted molecule (107). Recently, intrathecal injection of AAV9 or AAV2.5 showed robust transduction of the brain and spinal cord in non-human primates (108). Peripheral delivery of AAVs for brain targeting has also been used ( , ). Concerns with this approach for clinical application are the high dose needed, the transduction of peripheral organs (which may not be desirable) and the induction of a robust anti-aav and likely anti-transgene response. Nonetheless, using vectors that can cross the BBB may be beneficial for some applications. Intravenous delivery of AAV9, a recently identified serotype, can cross the BBB after and transduce neurons in neonatal mice and astrocytes and scattered neurons in adult mice and rhesus macaques (109, 110). Also, variants of AAV9 transduce motor neurons and astrocytes after systemic delivery by intravenous injection to adult mice (110, 111). Delivery to the retina doesn t pose as many challenges as delivery to the brain for several reasons. One, the retina is easily accessible to allow delivery of molecules, two, therapeutic outcomes can be easily tested using minimally invasive techniques such as

31 16 electroretinography or optical coherence tomography and three, the blood-retina barrier helps avoid a systemic immune response. Subretinal injections of AAV deliver the viral particles into the subretinal space, between the RPE and photoreceptor cells, while intravitreal injections deposit the viral particles into the vitreous and transduce the inner retina (112). This is due to the inner limiting membrane that acts as a barrier between the retinal cell layers. Recently, newer serotypes have emerged from directed evolution that can transduce the outer retina from intravitreal injections (113). AAV2, AAV5 and AAV8 have successfully been used for gene delivery to the retina in animal models of various retinal diseases ( ) and in human patients with Lebers congenital amaurosis (LCA) (119). The LCA trials showed that AAV2 delivery to both retinas in human patients was efficient and safe without causing an immune response (120). Although AAVs have a small packaging capacity (~4.7kb), they are ideally suited to deliver the small RNAi expression cassette. RNAi sequences delivered to the brain after AAV injection have shown therapeutic promise in mouse models of dominantly inherited polyglutamine (polyq) diseases and other neurological disorders (Table 2)(121). AAV mediated RNAi delivery to the retina has shown therapeutic benefit in animal models of dominant retinal diseases such as retinitis pigmentosa and RDS (retinal degeneration slow) pattern dystrophy (122)(123, 124). Objectives To date there are no disease modifying therapies for SCA2 or SCA7. In Chapter 2, we provide evidence that long-term reduction of both wildtype and mutant ataxin-7 in the cerebellum by RNAi is well tolerated and significantly improves phenotypes in the

32 17 SCA7 mouse. In Chapter 3, we reduced the expression of both mutant and wildtype ataxin-7 in the SCA7 mouse retina by RNAi and observed a preservation of normal retinal function and no adverse toxicity. These studies in the mouse cerebellum and retina provide a potential therapeutic for SCA7 patients in the clinic. In Chapter 4, we found that the reduction of wildtype ataxin-2 (the disease causing gene in SCA2) by RNAi was well tolerated in the mouse cerebellum. We further demonstrate that we can reduce the expression of both mutant and wildtype ataxin-2 by RNAi in the SCA2 mouse cerebellum. This body of work represents proof-of-principle studies in animal models demonstrating the therapeutic efficacy of RNAi therapy for SCA7 while identifying a potential RNAi sequence for further testing for SCA2 therapy. Contributions to work presented in this thesis Therapeutic approaches for SCA2 and SCA7 discussed in Chapter 1 are adapted from Ramachandran et al., 2013 published in Neurotherapeutics. Chapter 2 represents work submitted for publication in Molecular Therapy and is currently under revision. Chapter 3 represents work submitted for publication in PLoS One and is currently under revision. Chapter 4 details preliminary work on SCA2 therapy with ongoing work to complete the research study for submission. Contributions by authors other than PSR are indicated in the figure legends.

33 Figure 1. Co-opting the microrna pathway for delivery of RNAi triggers to the CNS and retina. Primary mirnas are transcribed in the nucleus and processed by the Drosha-DGCR8 complex to give precursor mirnas (pre-mirnas). Pre-miRNAs are exported out of the nucleus by Exportin-5 and undergo further processing by Dicer in the cytoplasm to give rise to mature mirnas. The mature mirna is then loaded into RISC to carry out silencing by binding to complementary mrna sequences. Artificial mirnas or shrnas can be delivered via AAV and enter the mirna pathway at different stages of the mirna pathway. (Modified from Ramachandran et al., 2013) 18

34 19 Table 1. SCA7 mouse models Mouse model Expression Pathology Phenotypes PrP-SCA7-92Q (Transgenic) CNS except PCs PC degeneration, retinal degeneration Ataxia, cone-rod dystrophy (20, 21) Gfa2-SCA7-90Q Astrocytes, PC and BG Ataxia (125) (Transgenic) Bergmann Glia degeneration SCA7-266Q (Knockin) Ubiquitous SCA7-R7E-90Q (Transgenic) SCA7-P7E-90Q (Transgenic) PDGF- SCA7-52Q (Transgenic) BAC-PrP-SCA7-92Q (Transgenic) Retina (rods only) PCs Neurons, astrocytes, oligodendrocytes of cerebellum, brainstem and cerebral cortex Ubiquitous Decrease in cell size of PCs, retinal degeneration Retinal degeneration, glial activation PC and deep cerebellar nuclei degeneration PC degeneration, defective myelination PC and BG degeneration Ataxia, myoclonic seizures, cone-rod dystrophy leading to blindness (126) Photoreceptor dysfunction, decrease in ERG a- wave amplitude (25) Ataxia manifests late and is not severe (25) Ataxia (18) Ataxia (26)

35 20 Table 2. Progress in RNAi therapeutics demonstrating therapeutic potential in cell and animal models. (Adapted from Ramachandran et al., 2013). Disease Target Gene Approach used Delivery vehicle Study demonstrating therapeutic potential References Huntington s Disease (HD) HTT Allelespecific (AS) and Non-allele specific (NAS) AAV1, AAV2, chemically modified ss-sirnas Silencing of endogenous HTT by shrnas and artificial mirnas in rhesus is tolerated up to 6 months without toxicity. Potent allelespecific silencing of mutant HTT is demonstrated in HD mice using chemically modified sssirnas targeting expanded CAG repeats. (56, ) SCA1 ATXN1 NAS AAV1 Silencing of ATXN1 using shrnas and more recently artificial mirnas in SCA1 transgenic and knock-in mouse models show improvement of motor coordination without toxicity. (132, 133)

36 21 Table 2 continued SCA2 ATXN2 AAV1 Partial suppression of InsP3R in the cerebellum improved motor coordination, reduced Purkinje cell degeneration in SCA2 transgenic mice. SCA3 ATXN3 NAS, AS Lentivirus Non-allele specific silencing and allele-specific silencing of mutant ATXN3 was well tolerated and reduced neuropathology in a rat model of SCA3. SCA6 CACNA1A AS Splice-isoform specific RNAi using artificial mirnas demonstrated allele-specific silencing of mutant CACNA1A in vitro SCA7 ATXN7 AS Allele-specific silencing of mutant ATXN7 is demonstrated in vitro using shrnas. (131) (134, 135) (136) (10)

37 22 Table 2 continued Parkinson s Disease (PD) SNCA Lentivirus, AAV2 Allele-specific silencing of α- syn using shrnas was observed in the rat brain and ameliorated behavioral deficits, but was also toxic in dopamine neurons. (137, 138) Alzheimer s Disease LRRK2 AS Allele-specific silencing of mutant α-syn and LRRK2 was achieved in vitro using artificial mirtron mimics. BACE1 Lentivirus shrnas silence BACE1 to reduce amyloid production and behavioral deficits in a transgenic mouse model Tau AS Allele-specific silencing of mutant Tau demonstrated using shrnas in vitro. APP AS AAV5 Allele-specific silencing of APP in a transgenic AD mouse model mitigated phenotypic progression (139) (140) (141) (142)

38 23 Table 2 continued PS1 AS Allele-specific sirnas silence mutant PS1 in vitro and reduced amyloid β42 production. CDK5 Lentivirus shrnas targeting CDK5 reduced neurofibrillary tangles in a transgenic mouse model PLK1 Lentivirus RNAi silencing of Plk1 in vitro reduced Amyloid β- induced cell death. (143, 144) (145) (146) MSUT2 Silencing of MSUT2 using sirnas decreased tau aggregation in vitro (147) Dystonia TOR1A AS Lentivirus Allele specific silencing of TorsinA(ΔE) by shrnas worked well in vitro, but when moved into a mouse model, the shrnas proved to be toxic. (83, 148)

39 24 Table 2 Continued SBMA CELF2 AAV9 Overexpression of naturally occurring mir- 196a indirectly enhances decay of Androgen Receptor through silencing of CELF2 in vivo ALS SOD1 AS Lentivirus Silencing SOD1 slows progression and extends survival in rodent models of ALS (149) (81, 82, 132, 150)

40 25 CHAPTER II: NON-ALLELE SPECIFIC SILENCING OF ATAXIN-7 IMPROVES PHENOTYPES IN A MOUSE MODEL OF SPINOCEREBELLAR ATAXIA TYPE 7 Abstract Spinocerebellar ataxia type 7 (SCA7) is a late-onset neurodegenerative disease characterized by ataxia and vision loss with no effective treatments in the clinic. The most striking feature is the degeneration of Purkinje neurons of the cerebellum caused by the presence of polyglutamine expanded ataxin-7. Here, we designed RNAi sequences to reduce the expression of both wildtype and mutant ataxin-7 in the Purkinje neurons of SCA7 mice. We observed long-term (33 weeks post-injection) sustained reduction of both wildtype and mutant ataxin-7 as well as a significant improvement of ataxia phenotypes. Furthermore, we observed a reduction in cerebellar molecular layer thinning and a remarkable reduction in nuclear inclusions, a hallmark of SCA7. In addition, we observed recovery of cerebellar specific transcripts in the presence of reduced ataxin-7 levels. Together, we demonstrate in this proof-of principle study that reduction of both wildtype and mutant ataxin-7 by RNAi is well tolerated and efficacious in the SCA7 mouse providing a potential therapeutic for SCA7 patients in the clinic.

41 26 Introduction Spinocerebellar ataxia type 7 (SCA7) is an autosomal dominant neurodegenerative disease and one of nine known polyglutamine (polyq) diseases. SCA7 patients suffer from loss of vision and motor coordination including dysarthria, dysphagia and slower reflexes(11). Anticipation is a feature of this disease, reflected in earlier onset in offspring of affected individuals (8). Currently, there are no disease modifying treatments for SCA7. SCA7 is caused by an expansion of >37 CAG repeats in exon 3 of the ataxin-7 gene (ATXN7)(151). While patients with repeats <58 present with cerebellar and brainstem degeneration prior to exhibiting retinal disease, patients with >59 CAG repeats typically present with retinal degeneration and then develop cerebellar ataxia(152). Ataxin-7 is ubiquitously expressed, yet only specific neuronal populations are affected with cerebellar degeneration in SCA7 is caused by the loss of Purkinje cells (PCs) and the loss of brainstem neurons in the inferior olive that project to the molecular layer(5). A hallmark of SCA7 is the presence of nuclear inclusions (NI), which contain aggregates of mutant polyq expanded ATXN7 and other ubiquitin-proteasome components(153). ATXN7 is a core component of the Spt-Taf9-Gcn5 acetyltransferase (STAGA) complex(19), a transcriptional co-activator complex that confers histone acetyltransferase (HAT) activity to its target genes. Mutant polyq ATXN7 alters STAGA recruitment to target genes, and disrupts its activity(154). Thus, reducing the levels of mutant polyq ATXN7 could reduce the downstream effects induced by mutant ATXN7. Previous studies have shown that ~50% reduction of mutant ataxin-7 in a SCA7 mouse model using Cre-medicated excision early after disease onset, alleviates motor

42 27 symptoms(27). In human patients, targeting a disease associated SNP is one way to specifically reduce mutant ATXN7, leaving the wildtype protein intact. Taking advantage of a SNP on mutant ATXN7 found in 50% of South African SCA7 patients, Scholefield et al designed allele-specific short hairpin RNAs (shrnas) to reduce mutant ataxin-7 and accomplished allele-specific decreases in ATXN7 aggregate formation in vitro (10). However, targeting mutant alleles via SNPs is currently impractical from a drug development perspective. We therefore set out to test if partial reduction of both the mutant and normal alleles could be beneficial in the setting of SCA7. To test this, we silenced both alleles in a SCA7 mouse model and asked if phenotypes improved or worsened. The BAC Prp SCA7-92Q mouse model expresses human mutant ataxin-7 cdna containing 92 pathogenic CAG repeats(26). The onset of PC degeneration is ~23 weeks of age, after which motor abnormalities become apparent. Ataxia is progressive and becomes severe as seen by clasping, rotarod and gait deficits and worsening scores on the ledge test (27, 155). Here, we show in proof-of-concept studies that non-allele specific silencing of ataxin-7 by RNAi is safe and well tolerated long-term in the SCA7 mouse. We observed a delayed disease onset, significant improvement of the ataxic and molecular phenotypes, and dramatic reduction in nuclear inclusions in cerebellar PCs with recovery of cerebellar transcripts that are known to be dysregulated in this disease.

43 28 Results Silencing ataxin-7 expression Small interfering RNAs (sirnas) targeting ataxin-7 were generated using a low off-target prediction algorithm(156) and were cloned into an artificial mirna expression vector(75). Plasmids expressing the candidate mirnas were tested for their ability to reduce ataxin-7 expression in vitro. We found that mis4 significantly reduced mouse ataxin-7 mrna levels by ~50% (p<0.05) in mouse neuro2a cells and exogenously expressed mutant human ataxin-7 protein levels in HEK293 cells, relative to a scrambled RNAi control (mic) or an empty vector control (U6) (Figure 2a,b). To target the mouse cerebellar PCs, mis4 and mic were subsequently cloned into adeno-associated viral (AAV) shuttle vectors to generate AAV2/1 viruses expressing mis4 or mic. A second expression cassette within the vector genome encoded egfp for visualization of transduced cells (Figure 2c). AAV2/1 has been previously reported to transduce cerebellar PCs(58). We injected AAV2/1-miS4, AAV2/1-miC or saline into the deep cerebellar nucleus (DCN) of the BAC Prp SCA7-92Q mouse to target human mutant ataxin-7 expression in cerebellar PCs. AAV2/1 injections into the DCN resulted in trafficking of the viral particles to the PC cell bodies and dendrites by axonal retrograde transport(58, 157, 158). One-month post injection, we observed a significant reduction (~50%, p<0.001) in human mutant ataxin-7 mrna levels in the mis4 injected cerebella (Figure 2d) relative to controls (saline or mic). To assess the long-term effects of non-allele specific silencing of ataxin-7, we injected pre-symptomatic BAC Prp SCA7-92Q mice at 7 weeks and followed them until 40 weeks of age (SCA7 mice die at ~50 weeks of age)

44 29 (Figure 3a). For this, SCA7 mice were injected bilaterally into the DCN with saline, AAV2/1-miC or AAV2/1-miS4. Untreated wildtype littermates were used for assessing relative recovery of disease. At 33 weeks post injection (40 weeks of age), we observed sustained reduction in both human and mouse ataxin-7 transcript levels (50%, p<0.001) in the mis4 injected cerebella relative to the controls (Figs. 2b,c). The expressed mis4 sequence was detected by stem loop PCR (Figure 3d) and widespread transduction of the PCs and DCN was observed by egfp expression (Figure 3e). mis4 improves motor phenotypes We evaluated SCA7 mice motor phenotypes at baseline (6 weeks; data not shown) and post-injection at 25 and 40 weeks of age (Figure 3a). The polyq SCA-7 disease mice clasp their hindlimbs when suspended by the tail, unlike wildtype mice, which splay their hindlimbs(58, 159, 160). The BAC PrP SCA7-92Q mouse also exhibits hindlimb clasping relative to their wildtype littermates(27), but in mis4 treated animals, there was a significant improvement in this phenotype, while saline or mic injected SCA7 mice continued to exhibit severe clasping (p<0.001; Figure 4a). mis4-treated mice also improved on the ledge test, while untreated or control treated SCA7 mice showed progressive decline, as reported earlier(21, 27) (p<0.001; Figure 4b). Motor skills and gait assays were also assessed, as BAC Prp SCA7-92Q mice exhibit progressive deficits in these tasks(27). At 25 weeks, there was no difference in rotarod performance between the SCA7 and wildtype mice, but at 40 weeks, mis4 treated mice had a significantly improved rotarod phenotype compared to control treated mice and performed at levels closer to wildtype on day 4 (Figure 4c; p<0.05 or p<0.01). mis4

45 30 mice also demonstrated a significantly improved stride length relative to control treated SCA7 mice (Figure 4d; p<0.001). Wildtype mice did significantly better relative to all groups (Figure 4d; p<0.001). Reduction of ataxin-7 in the PCs does not induce toxicity As one measure of assessing the tolerability of reducing both wildtype and mutant ataxin-7 expression, we evaluated injected cerebella for microglial activation at 40 weeks. We performed immunohistochemistry using a microglia specific marker (Iba-1) and found no notable increase in microglia activation in mis4 tissue sections relative to controls (Figure 5a). We also quantified the expression of glial fibrillary acidic protein (GFAP), an astrocyte marker whose expression increases in the presence of injury or inflammation. Untreated and control treated SCA7 mice exhibit higher levels (~25%, p<0.05) of GFAP relative to wildtype(125), while GFAP levels in the mis4 cerebella were not different from wildtype levels (Figure 5b), suggesting that mis4 treatment improved the progressive SCA7 neuroinflammation phenotype. mis4 has a therapeutic impact on the SCA7 cerebellum We next assessed the effect of ataxin-7 gene silencing on NIs. NIs are a hallmark of SCA7 and other polyq diseases and cells laden with NIs increase in number with progressing disease(27). We scored the NIs in mic- and mis4-transduced PCs at 40 weeks, using egfp as a marker of transduction. We also scored the NIs in saline injected, tissue-matched sections, for comparison. mis4 transduced PCs had a dramatic ~80%

46 31 reduction in NIs relative to mic or saline treated PCs (p<0.001) supporting decreased levels of nuclear ataxin-7 protein (Figure 5c). BAC Prp SCA7-92Q mice show progressive cerebellar lobule thinning(27). To assess potential benefit from ataxin-7 silencing, the molecular layer width was measured in mis4 vs. control-treated cerebellar lobules at 40 weeks. mis4-treated mice showed improved cerebellar morphology with greater molecular layer widths (by ~10%, p<0.001) than those of controls (Figure 5d). Wildtype mice retained significantly larger molecular layer widths compared to all groups. Lastly, we assessed if the reduction of mutant ataxin-7 affected cerebellar specific gene transcript levels (Figure 6). Calbindin1 is a PC specific marker and is downregulated in various SCA models(50, 161). Calbindin1 mrna levels were significantly reduced by ~40% in the BAC PrP SCA7-92Q control treated mice relative to wildtype mice (p<0.001). mis4 treatment partially restored calbindin1 levels to wildtype levels (p<0.05) and were significantly (~20%) higher than those of control treated SCA7 mice (p<0.01) (Figure 6a). We also analyzed transcripts that are altered in in response to mutant ataxin-7 expression(125, ). For example, the insulin-like growth factor pathway is dysregulated in the SCA7 knock-in mouse model, with Igfbp5 mrna levels down-regulated relative to normal(162). We found that Igfbp5 transcript levels were decreased in BAC PrP SCA7-92Q mice relative to wildtype (Figure 6e; p<0.05). Igfbp5 transcript levels in mis4-treated mice trended upward, but were not significantly improved relative to control-treated SCA7 mice. Expression of Eaat4, a PC-expressed glutamate transporter, is reduced in SCA7-90Q mice(164), and mis4 treatment restored Eaat4 levels to normal in contrast to control treated BAC Prp SCA7-92Q mice (Figure

47 32 6d; p<0.05). Other cerebellar specific transcripts such as Pcp2, Reln, Itpr1, Glast, and Grid2, which are downregulated in other SCA7 models(125, ), were not reduced in the BAC PrP SCA7-92Q mice, and there were no effects of treatment on their expression levels (Figure 6b,c,f-h). Discussion In this study, we provide evidence that non-allele specific silencing of ataxin-7 by RNAi is well tolerated in a SCA7 mouse model. Earlier studies in SCA7 mice demonstrated that ~50% excision of mutant ataxin-7 expression via Cre endonuclease immediate to disease onset partially alleviated ataxic symptoms(27). Here, we observed significant and robust improvements in the ataxic and neuropathological phenotypes as well as a delayed disease onset in SCA7 mice upon sustained reduction of both mutant and wildtype ataxin-7 expression in the cerebellar PCs. SCA7 is a core component of the STAGA co-activator complex, which confers Gcn5-mediated HAT activity when recruited to its target genes(15). Mutant ATXN7 causes a de-regulation of HAT activity and decondensation of chromatin, interfering with target gene activation(154). While other alleles causative of SCA result in viable mice upon gene knock-out(60, 166, 167), it is not known if reducing wildtype ataxin-7 expression is detrimental for survival. Thus it was important to understand the consequences of reducing the levels of ataxin-7 expression in vivo. Here, we found that reducing both mutant and wildtype ataxin-7 expression decreased neuroinflammation. We also noted an improvement in neuropathology including up-regulation of calbindin1 transcripts, among others. In addition, expression of our therapeutic RNAi sequence did

48 33 not induce microglial activation. Together, these data suggest that partial reduction of ataxin-7 expression by RNAi was tolerated and beneficial in the context of SCA7. NIs are a hallmark of SCA7 and consist of the truncated mutant ATXN7 protein along with ubiquitin components that increasingly accumulate with age(20, 126, 168). In primary rat embryonic neuronal cultures, ATXN7-100Q expression caused toxicity and neuronal cell death (169). Although it is debated as to whether nuclear inclusions per se are protective or toxic in vivo in SCA7, reducing levels of misfolded polyq-containing proteins should result in decreased nuclear inclusions and aggregation of mutant polyq proteins in the nucleus and therefore has been beneficial in polyq disease models. Vos et al demonstrated that overexpressing the heat shock protein HSPB7, a potent suppressor of polyq aggregation, decreased polyq induced toxicity and rescued retinal degeneration in a Drosophila polyq disease model(170). Another study in mice showed that reducing nuclear accumulation of mutant ATXN3 using calpain inhibitors reduced neuronal degeneration and dysfunction(171). Similarly, induction of PGC-1a expression in HD mice virtually eliminated huntingtin protein aggregation in the CNS and yielded a marked rescue of motor function(172). Here, we observed a dramatic ~80% reduction in the number of NIs in the mis4-treated SCA7 mice, confirming a decrease in mutant protein aggregation in PC nuclei. Thus, reducing nuclear-resident mutant ATXN7 likely contributes to the beneficial effects that we and others have noted(158, 173). RNAi therapy in other polyq diseases have shown that sustained reduction of mutant polyq alleles can alleviate disease phenotypes. RNAi-induced knockdown of the mutant SCA1 allele in PCs resulted in a complete rescue of motor and neuropathological phenotypes in a PC specific model of SCA1(58). On the other hand, RNAi-induced

49 34 knockdown of the mutant SCA3 allele in PCs resulted in a rescue of the neuropathological phenotypes but no rescue of motor phenotypes in a ubiquitously expressing transgenic model(158, 174). In our study, we used a BAC transgenic model with ubiquitous mutant ATXN7 expression. Upon RNAi treatment, there was significant improvement in several motor phenotypes and a delay in disease onset. In addition, we observed partial protection from the noted cerebellar molecular layer thinning and normalization of transcripts altered in SCA7. Our main target was cerebellar PCs, but mis4 injected mice also expressed mis4 in the brainstem with modest reduction of ataxin-7 transcripts in that region (data not shown). Recent work by Furrer et al. showed that the Bergman glia (BG), PCs and the inferior olive (IO) together contribute to the pathology of this mouse model(26). We did not observe complete rescue of motor phenotypes and neuropathology upon pre-symptomatic RNAi expression, possibly due to the fact that the PCs were the major site of knock-down. In prior work, crossing SCA7 mice with a floxed allele to a Cre-Pcp2 line for Cre-mediated excision in PCs and the IO, behavior phenotypes partially improved(26). However, when the transgene was excised from the BG, PCs and IO, a delayed disease onset and more robust improvements in behavior were seen(26). This suggests that cell types apart from the PCs contribute to SCA7 cerebellar pathology and it may be required to target the sirna to these different cell types as well for complete rescue. SCA7 is unique among the polyq diseases as mutant ataxin-7 causes vision loss and ataxia. While other studies have demonstrated partial improvement of some SCA7 motor phenotypes by overexpression of HGF(175) and IFN-β(173), here, we chose to target the causative gene directly and assess therapeutic efficacy. We found that reducing

50 35 both wildtype and mutant ataxin-7 expression by RNAi to be a viable therapeutic strategy in the SCA7 mouse with significant improvement in pathology and motor phenotypes. Materials and Methods Animals The University of Iowa Animal Care and Use Committee (IACUC) approved all animal protocols. PrP-floxed-SCA7-92Q BAC transgenic mice were generated in the La Spada lab and were maintained on the C57BL/6J background. Mice were genotyped using primers specific for the mutant human ataxin-7 transgene (27). Hemizygous and age-matched wild type littermates were used for the experiments. In the therapeutic trials, the treatment groups comprised approximately equal numbers of male and female mice. Mice were housed in a controlled temperature environment on a 12-hour light/dark cycle. Food and water were provided ad libitum. Viral vectors The plasmid expressing mouse U6-driven artificial mirna mis4, HS5 and mic was cloned as previously described using DNA oligonucleotides(75) using the following primers S4 forward primer: AAAACTCGAGTGAGCGGGGCTCAGGAAAGAAACGCAAACTGTAAAGCCACA GATGGG, S4 Reverse Primer: AAAAACTAGTAGGCGCGGCTCAGGAAAGAAACGCAAACCCATCTGTGGCTTT ACAG). Artificial mirna expression cassettes were cloned into paavmcscmvegfp plasmids which co-expressed CMV-driven egfp (75). Recombinant AAV serotype 2/1

51 36 vectors (AAV.miC.eGFP and AAV.miS4.eGFP) were generated by the University of Iowa Vector Core facility, as previously described (176). AAV vectors were dialyzed and resuspended in Formulation Buffer 18 (University of Iowa Gene Transfer Vector Core, Iowa City, IA) and titers (viral genomes/ml) were determined by RT-qPCR. AAV injections and tissue harvesting SCA7 transgenic mice were injected with AAV vectors as previously reported(177). For all cerebellar studies transgenic mice were injected bilaterally into the deep cerebellar nuclei (coordinates 6.0 mm caudal to bregma, ± 2.0 mm from midline, and 2.2 mm deep from cerebellar surface) with 4 µl of AAV1 virus (at viral genomes/ml) or saline. Following the injections, a topical antibiotic ointment was applied and the mice were allowed to recover according to the University of Iowa Animal Care and Use Committee s (IACUC) guidelines for Post-Anesthesia Monitoring, including monitoring of breathing and muscle tone. To harvest the cerebella, mice were anesthetized with a ketamine/xylazine mix and transcardially perfused with 20 ml of cold saline. Mice were decapitated, and for histological analyses, brains were removed and post-fixed overnight in 4% paraformaldehyde. Brains were stored in a 30% sucrose/0.05% azide solution at 4 C and cut on a sliding knife microtome at 45 µm thickness and stored at 20 C in a cryoprotectant solution. For RT-QPCR analyses, brains were removed, sectioned into 1 mm thick coronal slices using a brain matrix (Roboz, Gaithersburg, MD) and egfp expression was verified. Total RNA was isolated from whole cerebellum using the TRIzol (Life Technologies, Grand Island, NY)

52 37 extraction. RNA quantity and quality were measured using a NanoDrop ND-1000 (NanoDrop, Wilmington, DE). Immunohistochemistry and western blotting Free-floating sagittal cerebellar sections (45 µm thick) were washed in 1XPBS at room temperature and blocked for 1 h in 5% serum, 0.03% TritonX100 in 1XPBS. Sections were incubated with primary antibody in 3% serum and 0.03% TritonX in 1XPBS overnight at 4 C. Primary antibodies used were polyclonal rabbit anti-iba1 (1:1000; WAKO, Richmond, VA), polyclonal rabbit anti-calbindin (1:2000; Cell Signaling Technology, Danvers, MA) and polyclonal rabbit anti-atxn7 (1:1000; Thermo Fisher Scientific #PA1-749). For fluorescent IHC, sections were incubated with goat anti-rabbit Alexa Fluor 568 (1:1000; Life Technologies, Grand Island, NY) in 3% serum and 0.03% Triton-100 in 1XPBS for 1 hour at room temperature. For DAB IHC, sections were incubated in goat anti-rabbit biotin-labeled secondary antibody (1:200; Jackson Immunoresearch) in 3% serum and 0.03% Triton-X at room temperature for 1 hour. Tissues were developed with VECTASTAIN ABC Elite Kit (Vector Laboratories, Burlingame, CA), according to the manufacturer's instructions. All sections were mounted onto Superfrost Plus slides (Fischer Scientific, Pittsburgh, PA) and cover slipped with Fluoro-Gel (Electron Microscopy Sciences, Hatfield, PA). Images were captured on Leica Leitz DMR fluorescence microscope connected to an Olympus DP72 camera using the Olympus DP2-BSW software (Olympus, Melville, NY). For western blotting, anti-myc antibody was used (1:1000, Cell Signaling Technology).

53 38 RT-qPCR and stem loop PCR First-strand cdna synthesis was performed using 1 µg total RNA (High Capacity cdna Reverse Transcription Kit; Life Technologies, Grand Island, NY) as per manufacturer s instructions. RT-qPCR assays were performed on a sequence detection system using primers/probe sets specific for human or mouse ataxin-7, mouse GFAP, mouse Calbindin, Pcp2, Reln, Itpr1, Glast, Grid2, Eaat4, Igfbp5 or mouse β-actin (ABI Prism 7900 HT, TaqMan 2Xuniversal master mix and power SYBR green PCR master mix, Life Technologies, Grand Island, NY). RT-qPCR values were normalized to mouse β-actin. Stem loop PCR was performed as describer earlier (178). Briefly, PCR primers were designed to identify mis4. Reverse transcription was performed with RT specific primers (S4:GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACGGCTCA ) using the High Capacity cdna Reverse Transcription Kit; Life Technologies, Grand Island, NY and cdna obtained was subject to PCR using a specific forward primer (S4 Fwd: GCCCTTTGCGTTTCTTTCC) and a reverse primer (5 GTGCAGGGTCCGAGGT). Behavior Analysis All behavior assays were performed at 6, 25 and 40 weeks of age and 25 and 40 week data is presented here as means ± SEM unless otherwise specified. For all studies, p values were obtained by using one-way analysis of variance followed by Bonferroni posthoc analysis to assess for significant differences between individual groups, unless indicated otherwise. In all statistical analyses, p < 0.05 was considered significant.

54 39 Ledge test and hindlimb clasping assays and their scoring parameters are detailed previously(155). To take stride length measurements, mice were allowed to walk across a paper-lined chamber (100 cm 10 cm with 10 cm walls) and into an enclosed recess. Non-toxic red and blue paint was applied to their fore- and hind paws, respectively. Mice were then tested three times to produce three separate footprint tracings. Stride lengths were measured from the middle of each paw print between the same paws for steps taken during their gait. Steps were discarded in instances where a mouse stopped walking or turned around. Measurements were averaged for all four paws. Rotarod Mice were tested on an accelerated rotarod apparatus (model 47600; Ugo Basile, Comerio, Italy). Baseline testing was conducted at 6 weeks of age to separate SCA7 mice equally into treatment groups. No difference between SCA7 and wild type mice was seen at 4 weeks of age (data not shown). Mice were first habituated on the rotarod for 4 min. Mice were then tested three trials per day (with at least 30 min of rest between trials) for four consecutive days. For each trial, acceleration was from 4 to 40 rpm over 5 min, and then speed maintained at 40 rpm. Latency to fall (or if mice hung on for two consecutive rotations without running) was recorded for each mouse per trial. The trials were stopped at 400 s, and mice remaining on the rotarod at that time were scored as 400 s. Two-way analysis of variance followed by Bonferroni post-hoc analysis was used to assess for significant differences. Variables were time and treatment. Figure preparation All photographs were formatted with Adobe Photoshop software, all graphs were made with Prism GraphPad software, and all figures were constructed with Adobe Illustrator software.

55 Relative mouse ataxin-7 mrna levels 40 a U6 C S1 S2 S3 S4 HS5 mm unt b 1.5 myc 1.0 β-actin U6 mic mis4 * c d ITR ITR U6 RNAi TTTTT CMV egfp polya Relative human ataxin-7 mrna levels *** 0.0 Saline mic mis4 SCA7 Figure 2. Reduction of ataxin-7 mrna in vitro and in vivo. (a) HEK293 cells cotransfected with myc tagged human mutant ataxin-7 and several artificial mirna plasmids (C, mm- scrambled controls, S1-S5) or U6 (empty vector control) and untransfected (unt) cells were used as a negative control. 24 hours post transfection, protein was harvested and ataxin-7 protein levels were analyzed by western blot using a myc antibody (n=3). A representative western blot is shown. (b) Neuro2a cells were transfected with U6, mic or mis4 expression constructs. 24 hours post transfection RNA was harvested for RT-qPCR analysis. Results are represented as mean ±SEM (n=3), *p<0.05. (c) Cartoon of the designed AAV construct with the U6 promoter driving expression of the artificial mirna, followed by a CMV promoter driving expression of egfp. (d) AAV2/1 viruses expressing mis4 or mic or saline were injected into the SCA7 mouse DCN and one month post injection RNA was harvested from cerebellar tissue for RT-qPCR analysis. Results are represented as mean ±SEM (n=4), ***p<0.001.

56 Relative human ataxin-7 mrna levels 41 a Birth Cerebellar injections Behavior assays Behavior assays Week 0: b *** c Relative mouse ataxin-7 mrna levels *** *** Saline mic mis4 Saline mic mis4 Wt SCA7 SCA7 d mic mis4 e Figure 3. Sustained reduction of ataxin-7 expression in the SCA7 mouse cerebellum. (a) Experimental scheme for long-term cerebellar and retinal studies. (b) Relative human ataxin-7 mrna levels at 40 weeks from cerebellar tissue analyzed by RT-qPCR. Results are represented as mean ±SEM (n=8 saline; n=4 mic, n=8 mis4; n=7 wildtype), *** p < (c) Relative mouse ataxin-7 mrna levels at 40 weeks from cerebellar tissue analyzed by RT-qPCR. Results are represented as mean ±SEM (n=8 saline; n=4 mic, n=8 mis4; n=7 wildtype), *** p < (d) mis4 expression validated by stem-loop PCR after RT reaction with mis4 specific primers in cerebellar extracts at 40 weeks. mic was used as a negative control. (e) Sagittal section of mis4 SCA7 cerebellum showing transduction of the Purkinje cells (PC) and deep cerebellar nuclei (DCN).

57 42 a Hindlimb clasping score (0-3) ns *** *** Saline mic mis4 Wt SCA7 b Ledge score (0-3) ns *** *** Saline mic mis4 Wt SCA7 c Latency to fall (s) weeks 40 weeks ns * ns ns *** *** *** ** *** mic Saline misca7.s4 Wildtype 0 Day 1 Day 2 Day 3 Day 4 Day 1 Day 2 Day 3 Day 4 d 15 Stride Length (cm) 10 5 *** *** 0 Saline mic mis4 Wt SCA7 Figure 4. mis4 mice demonstrate significant improvement in ataxia phenotypes. (a) Hindlimb clasping score at 40 weeks, where 0 represents no clasping and 3 represents severe clasping. Results are represented as mean ±SEM (n=10 saline; n=13 mic, n=14 mis4; n=15 wildtype), *** p < (b) Ledge score at 40 weeks, where 0 represents good balance and coordination and 3 represents very poor or no balance and coordination. Results are represented as mean ±SEM (n=10 saline; n=13 mic, n=14 mis4; n=17 wildtype), *** p < (c) Rotarod analysis at 25 and 40 weeks with the Y- axis representing the latency to fall. Results are represented as mean ±SEM (n=10 saline; n=13 mic, n=14 mis4; n=15 wildtype), *** p <0.001, **p<0.01 two way ANOVA with Bonferroni post-hoc tests. (d) Box plots representing stride length at 40 weeks measured by footprint analysis. The median steps per mouse were included in the analysis (n=10-13 mice per group), *** p <0.001.

58 43 a Saline mic mis4 Wt Cortex DCN b d Relative GFAP mrna levels * ns Saline mic mis4 Wt SCA7 c % Inclusions in transduced Purkinje cells *** Saline mic mis4 SCA7 Molecular layer thickness normalized to Saline *** Saline mic mis4 Wt * SCA7 Figure 5. Histological and QPCR analysis of cerebellar tissue at 40 weeks. (a) Microglia expression with the Iba-1 antibody in the DCN and cortex of the cerebellum. n=3 mouse tissue sections were analyzed per group (n=3 mice per group). (b) Relative GFAP mrna levels in the injected cerebella by QPCR analysis. Results are represented as mean ±SEM (n=3 per group), * p <0.05. (c) Percentage of nuclear inclusions in the transduced PCs counted using image J software. Results are represented as mean ±SEM (n=3 per group), *** p < (d) Molecular layer thickness of the indicated lobule (right) in the various groups. Image J was used to measure the length of the lobule and all groups were normalized to saline. Results are represented as mean ±SEM (n=3 per group), *** p <0.001, * p <0.05.

59 Relative mouse Glast mrna levels Relative mouse Grid2 mrna levels Relative mouse Igfbp5 mrna levels Relative mouse Itpr1 mrna levels Relative mouse Reln mrna levels Relative mouse Eaat4 mrna levels Relative mouse Calb1 mrna levels Relative mouse Pcp2 mrna levels 44 a 2.0 *** * b 1.5 ns c Saline mic mis4 Wt SCA7 ns d Saline mic mis4 Wt SCA7 * ns e Saline mic mis4 Wt SCA7 * ns f Saline mic mis4 Wt SCA7 ns g Saline mic mis4 Wt SCA7 ns h Saline mic mis4 Wt SCA7 ns Saline mic mis4 Wt SCA7 0.0 Saline mic mis4 Wt SCA7 Figure 6. Changes in cerebellar transcripts post-injection at 40 weeks. RT-qPCR experiments representing the relative mrna expression levels of mouse (a) Calb1, (b) Pcp2, (c) Reln, (d) Eaat4, (e) Igfbp5, (f) Itpr1, (g) Glast and (h) Grid2. Results are represented as mean ±SEM (n=6 saline; n=6 mic, n=6 mis4; n=4 wildtype), *p<0.05.

60 45 CHAPTER III: RNA INTERFERENCE-BASED THERAPY FOR SPINOCEREBELLAR ATAXIA TYPE 7 RETINAL DEGENERATION Abstract Spinocerebellar ataxia type 7 (SCA7) is an autosomal dominant neurodegenerative disease characterized by loss of motor coordination and retinal degeneration with no current therapies in the clinic. The causative mutation is an expanded CAG repeat in the ataxin-7 gene whose mutant protein product causes cerebellar and brainstem degeneration and retinal cone-rod dystrophy. Here, we reduced the expression of both mutant and wildtype ataxin-7 in the SCA7 mouse retina by RNA interference and evaluated retinal function 23 weeks post injection. We observed a preservation of normal retinal function and no adverse toxicity with 50% reduction of wildtype and mutant ataxin-7 alleles. These studies address an important safety concern regarding non-allele specific silencing of ataxin-7 for SCA7 retinal therapy.

61 46 Introduction Spinocerebellar ataxia type 7 (SCA7) is a neurodegenerative disease characterized by ataxia, which manifests as loss of motor coordination, dysarthria, slower reflexes, and retinal degeneration leading to vision loss. Anticipation is a key feature of this disease and for SCA7 patients with early onset disease, vision loss can occur early in life, followed by ataxia (152). The retinal degeneration is a cone-rod dystrophy, which may manifest as a maculopathy when the onset is in adulthood, or as a diffuse retinopathy or geographic atrophy with childhood onset (12). Typically a decrease in visual acuity is noted, followed by loss of peripheral vision and ultimately complete blindness in many cases (13). SCA7 is caused by the expansion of CAG repeats in the ataxin-7 gene (ATXN7), which translates into a polyglutamine expanded protein. Studies in SCA7 mouse models and in vitro studies have demonstrated that polyglutamine expanded ataxin-7 disrupts the transcription of CRX, the cone-rod homeobox protein, in turn affecting the transcription of CRX-regulated genes resulting in a cone-rod dystrophy (19, 20). There is no known treatment for SCA7 retinal degeneration. Previous studies have demonstrated that reduction of mutant ataxin-7 in the brain by Cre-Lox excision in a SCA7 mouse model can alleviate motor phenotypes (27). However, from a clinical perspective, targeting the mutant allele alone will require targeting a patient specific polymorphism within the disease allele to discriminate it from wildtype. At the current time, the personalized development of RNA interference (RNAi) therapy on a per family basis is impractical. We therefore designed and tested an RNAi sequence that could

62 47 reduce the expression of both mutant and wildtype ataxin-7 (non-allele specific silencing) and assessed feasibility in vivo. The BAC-Prp-SCA7-92Q mouse model used in this work expresses the human mutant ataxin-7 cdna in the central nervous system (CNS), including the retina, and displays ataxic phenotypes (26). This model does not demonstrate retinal degeneration although the transgene is expressed in the retina as seen by RT-qPCR. Nonetheless, the model gives us a unique opportunity to test if reducing the wildtype and mutant ataxin-7 levels long-term in the SCA7 mouse retina is safe with maintenance of normal retinal function. We demonstrate for the first time that RNAi-based reduction of mutant and wildtype ataxin-7 expression is well tolerated in the SCA7 mouse retina and the introduction of an ataxin-7-targeted RNAi sequence does not affect normal retinal function. Results Silencing ataxin-7 expression in the SCA7 retina We designed small interfering RNAs (sirnas) targeting ataxin-7 using a low offtarget prediction algorithm (156). These sirnas were cloned into artificial mirna expression vectors (75) and were tested in vitro (Ramachandran et al., unpublished data). We identified one sequence (mis4) that could reduce human mutant ATXN7 and mouse ataxin-7 mrna expression by ~50% in vitro relative to controls (Ramachandran et al., unpublished data). To target the mouse photoreceptors, mis4 and a scrambled RNAi control sequence (mic) were cloned into adeno-associated viral (AAV) shuttle vectors and AAV2/1 viruses were generated expressing mis4 or mic. The reporter gene egfp

63 48 was expressed from a second cassette on the same vector for visualization of transduced cells. To target the photoreceptor cells in SCA7 mice (as SCA7 is a cone-rod dystrophy), AAV2/1.miS4 was subretinally injected into one eye and the contra-lateral eye was subretinally injected with AAV2/1.miC. Although subretinal injection of AAV2/1 has been previously reported to transduce primarily the retinal pigment epithelium (RPE) (179), we noted extensive transduction of the RPE, inner segments of the photoreceptors, cells in the outer nuclear layer and some inner nuclear layer cells (Figure 8d). One month post injection we observed a significant reduction of both human mutant ataxin-7 (~60%, p<0.001) and mouse ataxin-7 expression (~60%, p<0.01) in the mis4-injected retinas relative to mic controls (Figures 7a,b). Sustained silencing of ataxin-7 in the SCA7 retina To assess the long-term effects of non-allele specific silencing of ataxin-7 in the retina, we subretinally injected SCA7 mice at 7 weeks of age and followed them for 23 weeks (Figure 8a). We included mice injected with subretinal saline as an additional control group in our long-term study. RT-qPCR assays on retinal RNA harvested at 30 weeks of age revealed a significant reduction in ataxin-7 mouse (~50%, p<0.001) and human mutant (~65%, p<0.001) transcript levels in the mis4 injected retina relative to saline and mic injected groups (Figures 8b,c). Histological evaluation at 30 weeks of age showed extensive transduction of the RPE, photoreceptor cells and a few cells within the inner nuclear layer, and funduscopy demonstrated effective virus distribution across a

64 49 large area of the retina (Figures 8d,e). In addition, we detected the injected mis4 sequence at 30 weeks by stem loop PCR in the retina (Figure 8f). Reducing ataxin-7 expression does not induce neuropathology in the SCA7 mouse retina To assess if mis4 treatment induced toxicity in the retina, we examined retinal morphology by hematoxylin and eosin staining and observed no gross changes in treated vs. normal, untreated eyes (Figure 9a). We also measured the retinal thickness by optical coherence tomography (OCT) at 30 weeks before euthanizing the mice. Retinal thickness was not different among the injected groups (Figures 9b,c). To assess if gliosis was induced, we quantified the expression of the glial fibrillary acidic protein (GFAP) by RTqPCR and observed no mis4-induced increase in GFAP expression (Figure 9d) relative to saline injected eyes. Interestingly, eyes injected with the control mirna expressing vectors mic did show elevated GFAP levels. Retinal function is not altered in mis4 treated retinas At 30 weeks of age we tested retinal and visual function in SCA7 mice to determine the physiological effects of reducing wildtype and mutant ataxin-7 levels, and the general impact of RNAi on retinal function. This was assessed by electroretinogram (ERG), and optokinetic tracking for visual acuity (180). Full field ERG recordings showed that the mixed rod-cone response and the isolated cone response were not significantly different in the mis4 injected groups relative to the saline injected controls (Figures 10a-c). We did observe a significant decline in the mixed rod-cone b-wave

65 50 response in the mic injected retinas (p<0.05) relative to the saline treated eyes, and hence ERG values were normalized to saline injected controls. There was no difference relative to wildtype mice in the optokinetic tracking response (181) following sustained silencing of ataxin-7 expression (Figure 10d). Discussion Previous SCA7 mouse models displayed strong retinal degeneration phenotypes with thinning of retinal layers and progressively reduced ERG responses consistent with vision loss (20, 126). Although the transgene is expressed in the retina in the BAC-Prp- SCA7-92Q mouse, we were surprised to observe no notable retinal phenotype (tested by ERG recordings, funduscopy, histology; this work and data not shown). This may be due to inadequate expression of the transgene in the retinal cells, the background strain of the mice, or alternatively, the mice succumb to cerebellar disease prior to retinal pathology and dysfunction. This is in contrast to the SCA7 phenotype in patients, in which retinal disease precedes the ataxic phenotype in most cases. Ataxin-7 mrna is expressed in the outer nuclear layer, inner nuclear layer and ganglion cell layers in the mouse retina and has a similar pattern of expression to the human retina as evidenced by in situ hybridization (Figure 11) (182). With subretinal injections of AAV2/1, we transduced the outer retina efficiently but not the inner retina. However, since SCA7 is a cone-rod dystrophy, targeting the photoreceptor cells in the outer retina may be sufficient to produce a therapeutic effect. With our therapeutic RNAi sequence targeting mostly the outer retina, as seen by egfp expression, we were able to

66 51 achieve significant and sustained reduction in both mutant and wildtype ataxin-7 expression. There is no SCA7 knock-out model so at the time this study was undertaken it was unknown if reducing ataxin-7 expression would be tolerated. We showed that there was no gliosis in mis4 treated retinas and their histological appearance was similar to the normal retina, suggesting that neither the RNAi sequence nor ataxin-7 reduction induced toxicity. Confirming this, ERG recordings and optokinetic tracking demonstrated that retinal function was not altered in the mis4-injected retinas relative to controls. Therefore, non-allele specific silencing of ataxin-7 has no identifiable long-term negative effect on retinal structure or function. The toxic buildup of mutant protein in gain-of-function autosomal dominant diseases such as SCA7 makes them amenable to test gene knockdown therapies. Early gene knockdown strategies using ribozymes, RNA enzymes that cleave specific mrna molecules, demonstrated delayed photoreceptor degeneration in a rat model of autosomal dominant retinitis pigmentosa (183, 184). Antisense oligonucleotide and sirna delivery to the retina can also be used to efficiently silence a transcript and have shown therapeutic benefit ( ), however, they are transient and require repeated intraocular delivery. Long-term correction can be achieved by viral mediated delivery of the transgene with a minimal immune response. AAV mediated shrna delivery to the retina is effective (190, 191), however, shrnas are toxic in the nervous system and saturate the RNAi pathway (137, 148, 192). On the other hand, AAV mediated delivery of artificial mirnas or mir-based RNAi sequences have proven to be safe long term and do not saturate the mirna pathway (177, 193). Artificial mirna sequences were used

67 52 in a study to target peripherin-2 in the retina and 5 weeks post injection, efficient knockdown was observed while toxicity was not evaluated (194). Here, we analyzed the long term safety of AAV mediated artificial mirna expression in the retina 23 weeks post injection and found that our artificial mirna does not induce toxicity while effectively reducing ataxin-7 levels. Currently, there is no effective way of restoring vision to SCA7 patients. Therapeutic methods that are being considered in other cone-rod dystrophies and retinal degenerative disease where photoreceptor cells are lost early include stem cell derived photoreceptor cell replacement, electronic retinal implants and optogenetics (195, 196). RNAi based therapy for SCA7 is a disease modifying therapy that can be applied when photoreceptor cells are still present in the patient. Cumulatively our data suggest that non-allele specific silencing of ataxin-7 in the retina by RNAi may be a viable therapeutic strategy in patients with SCA7 retinal degeneration. In the future, it will be important to assess rescue of retinal function and prevention of retinal degeneration in a SCA7 retinal specific model such as the R7E (25) or the knock-in model (126), upon reduction of ataxin-7 transcripts by RNAi. Materials and Methods Animals The University of Iowa Animal Care and Use Committee (IACUC) approved all animal protocols. BAC-PrP-SCA7-92Q transgenic mice were generated in the La Spada lab and were maintained on the C57BL/6J background. Mice were genotyped using primers specific for the mutant human ataxin-7 transgene (27). Hemizygous and age-

68 53 matched wild type littermates were used for the experiments. Mice were housed in a controlled temperature environment on a 12-hour light/dark cycle. Food and water were provided ad libitum. Viral vectors The plasmids expressing mouse U6-driven artificial mirna mis4 and mic was cloned as previously described using DNA oligonucleotides (75) using the following primers S4 forward primer: AAAACTCGAGTGAGCGGGGCTCAGGAAAGAAACGCAAACTGTAAAGCCACA GATGGG, S4 Reverse Primer: AAAAACTAGTAGGCGCGGCTCAGGAAAGAAACGCAAACCCATCTGTGGCTTT ACAG Artificial mirna expression cassettes were cloned into paavmcscmvegfp plasmids which co-expressed CMV-driven egfp (75). Recombinant AAV serotype 2/1 vectors (AAV.miC.eGFP and AAV.miS4.eGFP) were generated by the University of Iowa Vector Core facility, as previously described (176). AAV vectors were dialyzed and resuspended in Formulation Buffer 18 (University of Iowa Gene Transfer Vector Core, Iowa City, IA) and titers (viral genomes/ml) were determined by RT-qPCR. AAV injections SCA7 transgenic mice were injected subretinally with 2 µl of AAV1 virus (at viral genomes/ml) or saline. Briefly, mice were anesthetized using a ketamine/xylazine mix and one drop of 50% betadine solution and topical 1% proparacaine was applied to anesthetize the eye. A sharp 30 gauge needle was inserted

69 54 through the posterior sclera of the eye followed by a blunt 33 gauge needle inserted into the subretinal space and 2 µl of virus was injected under direct visualization using an operating microscope. A retinal bleb could be seen following successful placement of the injections. Following the injections, a topical antibiotic/steroid ointment was applied and the mice were allowed to recover according to the University of Iowa Animal Care and Use Committee s (IACUC) guidelines for Post-Anesthesia Monitoring, including monitoring of breathing and muscle tone. Tissue harvesting and histology To harvest the retinas, mice were anesthetized with a ketamine/xylazine mix and sacrificed and the eyes removed. The eyes were then fixed in 4% PFA for histological analysis or retinas were dissected and put in TRIzol (Life Technologies, Grand Island, NY) for RNA isolation. RNA quantity and quality were measured using a NanoDrop ND-1000 (NanoDrop, Wilmington, DE). For histological analysis, the fixed eyes were enucleated and the eye cup was infiltrated in acrylamide embedding solution (1.2 M acrylamide, 0.9 mm bisacrylamide, 0.7% N,N,N,N -tetramethylethylenediamine, 1 mm MgCl 2, 1 mm CaCl 2 ) overnight and polymerized the next day. The acrylamide was trimmed to the eye cup and placed in OCT (Tissue-Tek). 7 µm thick retinal sections were cut using the Leica Microm Cryostat and used for histological analysis. Standard hematoxylin and eosin staining was performed on 7 µm sections mounted onto Superfrost Plus slides (Fischer Scientific, Pittsburgh, PA). Images were captured on Leica DMR microscope connected to an Olympus DP72 camera using the Olympus DP2-BSW software (Olympus, Melville, NY).

70 55 In situ hybridization Human donor eyes were obtained through the Iowa Lions eye bank (Iowa City, IA) with informed consent in accordance with the declaration of Helsinki. Tissues were processed as described previously (197) and 7 µm thick retinal sections were cut using the Leica Microm Cryostat. A modified 2 OMe ZEN ataxin-7 probe (198) (Integrated DNA Technologies, Coralville, IA) was used to hybridize to ataxin-7 mrna in the retina. Retinal sections from C57Bl/6J mice and human retina were used for in situ hybridization using previously described methods (199). Scrambled probes (mumgmu mamamc mamcmg mumcmu mamuma mcmgmc mcmc) were used as controls to the ataxin-7 probe (mcmcmumcmcmumcmamcmumgmgmamumamamcmcmgmamgmamamg mcmumgmgmcmumcmamgmumg). RT-qPCR and stem loop PCR First-strand cdna synthesis was performed using total RNA (High Capacity cdna Reverse Transcription Kit; Life Technologies, Grand Island, NY) as per manufacturer s instructions. RT-qPCR assays were performed on a sequence detection system using primers/probe sets specific for human or mouse ataxin-7, mouse GFAP or mouse β-actin (ABI Prism 7900 HT, TaqMan 2Xuniversal master mix and power SYBR green PCR master mix, Life Technologies, Grand Island, NY). RT-qPCR values were normalized to mouse β-actin. Stem loop PCR was performed as describer earlier (178). Briefly, PCR primers were designed to identify mis4. Reverse transcription was

71 56 performed with RT specific primers (S4:GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACGGCTCA ) using the High Capacity cdna Reverse Transcription Kit; Life Technologies, Grand Island, NY and cdna obtained was subject to PCR using specific forward primers (S4 Fwd: GCCCTTTGCGTTTCTTTCC) and a reverse primer (5 GTGCAGGGTCCGAGGT). Retinal assays Electroretinogram (ERG) Full-field ERG was performed using the Espion V5 Diagnosys system (Diagnosys LLC, Lowell, MA). Mice were dark adapted and were anesthetized with ketamine/xylazine mix. ERGs were recorded simultaneously from the corneal surface of each eye after pupil dilation (with 1% tropicamide), using gold ring electrodes (Diagnosys) referenced to a needle electrode (Roland Consult, Brandenburg an der Havel, Germany; LKC Technologies Inc., Gaithersburg, MD) placed on the back of the head. Another needle electrode placed in the tail served as the ground. A drop of 2.5% methylcellulose was placed on the corneal surface to ensure electrical contact and to maintain corneal integrity. Body temperature was maintained at a constant temperature of 38 C using a regulated heating pad. All stimuli were presented in a ColorDome (Diagnosys) ganzfeld bowl, and a camera monitored mouse head and electrode positions. Dim red light was used for room illumination until dark-adapted testing was completed. A modified International Society for Clinical Electrophysiology of Vision (ISCEV) protocol (200, 201) was used, including a dark-adapted dim flash of 0.01 cd.s/m 2, maximal combined response (standard combined response or SCR) to bright flash of 3

72 57 cd.s/m 2, light-adapted bright flash of 3 cd.s/m 2, and 5-Hz flicker stimuli at 3 cd.s/m 2. The a-wave was measured from the baseline to the trough of the first negative wave. The b- wave was measured from the trough of the a-wave to the peak of the first positive wave, or from the baseline to the peak of the first positive wave if no a-wave was present. Optokinetic tracking The optokinetic response was measured using an OptoMotry system according to previously described methods (Cerebral Mechanics, Lethbridge, Alberta, Canada)(181). Briefly, mice were placed onto an elevated platform in the test chamber, and tested after ~60 seconds of brief acclimation. Mice were presented with a vertically oriented sine wave grating rotating at 12 /s. The between stimulus blank was an equal luminance gray homogenous surround (152 cd.sm 2 ). The level of contrast was fixed at 100%, and the spatial frequency was varied to determine thresholds for stimulus detection using a standard protocol. Spatial frequency was expressed in cycles per degree (c/d) for one sine wave (paired vertical black and white bars). Tracking head movements were scored by an experienced operator blind to genotype. Test sessions were during the mid 4 hours of the light phase of a daily cycle. The test chamber was thoroughly cleaned between animals. Where no response was observed, an animal was tested at least 10 times. Optical coherence tomography (OCT) (Bioptigen, Research Triangle Park, NC) was performed after placing the animals under ketamine/xylazine anesthesia as described above. 1% tropicamide was used to dilate pupils, and the retinas were scanned. Methylcellulose lubricant was placed on the corneas, and the noncontact probe was

73 58 positioned near the cornea until the retinal image could be seen on the screen. This was then focused and oriented with the optic nerve (ON) in the middle of the scan as a landmark. Retinal OCT was performed using rectangular volume scan (volumetric acquisition made up of a series of B-scans) with a length of 1.40 mm at a width of 1.40 mm at a rate of 1000 A-scan/B-scan. An average of four repeated B scans (of the same region) centered on the ON was used for analysis. The total retinal thickness was measured on the OCT image at two locations, each 0.3 mm from the edge of the ON, on either side using the measuring tool provided (Bioptigen). Eyes were excluded from further analysis if retinal disruption more than four times the area of a typical injection site due to hemorrhage or chronic retinal detachment (greater than 1 week after injection) was noted on OCT. Figure preparation All photographs were formatted with Adobe Photoshop software, all graphs were made with Graphpad Prism software, and all figures were constructed with Adobe Illustrator software.

74 59 a Relative human ataxin-7 mrna levels 1.5 *** Relative mouse ataxin-7 mrna levels ** 0.0 mic mis4 SCA7 Figure 7. Reduction of ataxin-7 mrna in the SCA7 mouse retina. (a) Relative levels of human ataxin-7 one month post-injection with AAV2/1 mic or mis4. ***p< (b) Relative levels of mouse ataxin-7 one month post-injection with AAV2/1 mic or mis4. For both (a) and (b) results are represented as mean ±SEM (n=3), ***p<0.001, *p< mic mis4 SCA7

75 60 a Birth Subretinal injections Retina assays Week 0: 7 30 b c Relative mouse ataxin-7 mrna levels *** *** Saline mic mis4 Wt SCA7 Relative human ataxin-7 mrna levels *** Saline mic mis4 SCA7 d e f RPE 60 OS/IS ONL 40 mis4 C INL GCL Figure 8. Sustained suppression of ataxin-7 expression. (a) Experimental scheme. (b) Relative mouse ataxin-7 mrna levels in the retina. (c) Relative human ataxin-7 mrna levels in the retina. For both (b) and (c), results are represented as mean ±SEM (n=3 per group), *** p < (d) Transduction of the RPE, photoreceptors, outer nuclear layer and few inner nuclear layer cells in the SCA7 retina at 30 weeks. (e) egfp fluorescence as observed by funduscopy (Sajag Bhattarai). (f) mis4 expression validated by stem-loop PCR in retinal extracts. mic was used as the negative control.

76 61 a Saline mic mis4 Wt RPE OS/IS ONL INL GCL b Saline mic mis4 Wt c Retinal thickness in mm ns Saline mic mis4 Wt SCA7 Figure 9. Assessing toxicity post-injection at 30 weeks. (a) Hematoxylin and eosin staining of retinas at 30 weeks. (b) Optical coherence tomography (OCT) images of the retinas at the region of the optic nerve (Sajag Bhattarai). (c) Retinal thickness measured using the OCT images (Sajag Bhattarai). (d) Relative GFAP mrna levels in the injected retinas by RT-qPCR analysis. For (c; n 6 per group) and (d; n=3 per group), results are represented as mean ±SEM. d Relative GFAP mrna levels ns Saline mic mis4 Wt SCA7

77 62 a Amplitude of mixed rod-cone response b-wave (µv) ns * Saline mic mis4 b Amplitude of mixed rod-cone response a-wave (µv) ns Saline mic mis4 SCA7 SCA7 c 50 ns d 0.4 ns Amplitude of 5Hz flicker response (µv) Saline mic mis4 Spatial Frequency (c/d) Saline mic mis4 Wt SCA7 SCA7 Figure 10. Retinal function 30 weeks post-injection. (a) ERG b-wave of the mixed rodcone response, n 4 (Sajag Bhattarai) (b) ERG a-wave of the mixed rod-cone response, n 4 (Sajag Bhattarai) (c) 5 Hz flicker (cone) response, n 4 (Sajag Bhattarai) (d) Optokinetic response measured by the spatial frequency, n 3. (Pratibha Singh) Results are represented as mean ±SEM, *p<0.05.

78 63 Mouse Human control ataxin-7 control ataxin-7 RPE RPE ONL ONL INL GCL INL GCL Figure 11. In situ hybridization experiments showing ataxin-7 mrna expression in mouse and human retina sections relative to a scrambled control probe.

79 64 CHAPTER IV: IDENTIFICATION OF A POTENT RNAI SEQUENCE FOR SPINOCEREBELLAR ATAXIA TYPE 2 THERAPY Abstract Spinocerebellar ataxia type 2 (SCA2) is a late-onset neurodegenerative disease characterized primarily by loss of motor coordination and balance, difficulty in speech and swallowing and ophthalmoplegia. To identify therapies for SCA2, we designed and tested several RNAi sequences to reduce the expression of both wildtype and mutant ataxin-2 in vitro and in vivo. After preliminary testing of candidate sequences in vitro, we tested two RNAi sequences (mis2 and mis3) in non-human primates and found that mis2 was toxic and mis3 did not reduce ataxin-2 levels significantly. We next tested additional RNAi sequences with two emerging for further testing, this time with the preliminary step of long safety assessment in mice. As SCA2 mouse models were not available at the time of our work, these sequences were tested in wildtype mouse cerebellum. We found that reduction of wildtype ataxin-2 expression in the mouse cerebellum was tolerated 4 months post injection without inducing behavioral deficits or cerebellar pathology. We additionally tested other sequences for improved silencing efficacy, and identified one potent RNAi sequence (S12) that significantly reduced the expression of both mutant and wildtype ataxin-2 in the cerebellum of a SCA2 mouse model. Ongoing work will establish if long-term reduction of both mutant and wildtype ataxin-2 will provide therapeutic benefit in the SCA2 mouse setting, and the safety of this sequence in normal cerebella

80 65 Introduction Spinocerebellar ataxia type 2 (SCA2) is a late-onset autosomal dominant disease characterized predominantly by gait ataxia, slow reflexes and occulomotor deficits (28). SCA2 is the second most prevalent ataxia (15%) after SCA3 and is the most prevalent ataxia in certain geographical regions such as Mexico and Cuba (35, 36). It is one of nine known polyglutamine (polyq) diseases, which include Huntington s disease (HD), SCA 1-3,6,7,17, spinal bulbar muscular atrophy (SBMA) and dentatorubral-pallidoluysian atrophy (DRPLA). SCA2 is caused by an expansion of greater than 34 CAG repeats in the ataxin-2 that translates into a polyq tract expanded protein. Similar to other polyq diseases, disease onset correlates with repeat length (34). In SCA2, progressive neuropathology of the cerebellum and brainstem occurs with dendritic thinning, degeneration of the PCs and atrophy of the pons and olivary nucleus. Previous studies have shown that reducing the expression of mutant polyq proteins can reduce their toxic gain of function effects and alleviate disease phenotypes (56, 58, 135, 157). For polyq diseases, targeting the mutant allele specifically has been challenging. For SCA6, an allele specific approach is possible by targeting the mutant splice form (136). For other polyq diseases, like HD, SCA3 and SCA7, the mutant allele can be specifically targeted for a portion of affected individuals via SNPs (10, 134, 202). However, SNPs differ among various patient populations and thus mutant allele specific targeting may not be practical for drug development. On the other hand, non-allele specific silencing has shown promise in animal models of various polyq diseases. For example, pre-clinical studies have found that reducing wildtype Htt alleles is well tolerated in mouse models of HD and rhesus macaques (56, 129, 130). Similarly, non-

81 66 allele specific silencing is well tolerated in a SCA3 model (135). For SCA2, gene silencing of mutant and/or wildtype ataxin-2 has not been tested previously. It is known that the SCA2 knockout mouse is viable but has reduced fertility, obesity, hyperactivity and reduced learning (60, 203, 204). If non-allele specific silencing is to be used for SCA2 therapy, it is important to test if reducing wildtype ataxin-2 levels in the cerebellum has detrimental effects. Here, we assessed if reducing wildtype ataxin-2 levels in the mouse cerebellum using RNA interference (RNAi) based approaches is safe and tolerated in the mouse. We identified RNAi sequences that were potent in vivo and found that reducing wildtype ataxin-2 expression in the cerebellum was not detrimental at the behavioral or tissue levels. Additionally, we confirmed the potency of RNAi triggers in preliminary studies in a SCA2 mouse model using RNAi sequences that reduced wildtype and mutant ataxin-2 mrna levels significantly. Future and ongoing work will assess if reducing ataxin-2 levels in a SCA2 setting will yield therapeutic benefit. Results Reducing ataxin-2 expression in vitro and in vivo To target both mutant and wildtype ataxin-2 we screened sirna sequences incorporated into the mir-30 backbone (75) (known as artificial mirnas) for their ability to reduce ataxin-2 expression in vitro. The RNAi sequences were designed taking into consideration their off-targeting potential (156) and their ability to target the coding region of human, rhesus and mouse ataxin-2 mrna. We initially screened four RNAi sequences (mis2-mis5) to reduce human and mouse ataxin-2 expression in Hela and

82 67 mouse neuro2a cells, respectively. We identified that all sequences demonstrated significant knockdown in vitro (p<0.05 or p<0.01; Figure 12a,b). mis2 and mis3 were then cloned into AAV shuttle vectors and AAV2/1 mis2 and AAV2/1 mis3 were generated. In an unusual step, we skipped testing in mice and first tested our constructs for short term safety in nonhuman primates, reasoning that long term efficacy and safety in mice is irrelevant if the sequences are toxic in nonhuman primate brain; nonhuman primate studies are often requested by regulatory agencies prior to human use. Therefore, the sequences were first tested in the non-human primate cerebella for knockdown efficacy. In mice, AAV2/1 injection into the deep cerebellar nuclei (DCN) transduces PCs by retrograde transport (58, 131, 157, 205). In monkey, it was unknown if similar trafficking occurs following DCN injections. AAV2/1 -mis2 or -mis3 were stereotaxically injected bilaterally into the DCN of rhesus brains (n=2 injected sides per sequence) by Dr. Jeff Kordower, Rush University, Chicago. One month post injection the animals were euthanized and the DCN and cerebellar lobules were harvested. Untransduced cerebellar tissue was also harvested to use as controls. RNA was extracted from the harvested tissues and subjected to RT-qPCR. Tissue was also harvested for immunohistochemistry to assess toxicity. AAV2/1 mis2 demonstrated significant knockdown of rhesus ataxin-2 mrna levels (p<0.05) in the cerebellar lobules and DCN, while mis3 did not yield significant knockdown in most cerebellar lobules harvested (Figure 13). While ataxin-2 knockdown was significant with mis2, immunohistochemical analyses indicated that mis2 was toxic in vivo as indicated by the presence of increased activation of the microglia marker Iba-1 (Figure 14).

83 68 Given these data, we subsequently screened additional RNAi triggers, and tested two, mis4 or mis5, in mice. Even though data in nonhuman primate is the penultimate step in humans, preliminary testing given this unexpected toxicity was warranted. mis4, mis5 and mic were subcloned into adeno-associated viral (AAV) shuttle vectors and AAV2/1 viruses expressing mis4, mis5 or mic from a U6 promoter and egfp from a second promoter (CMV) were generated. We targeted the PCs of wildtype C57Bl/6J mice by performing stereotaxic injections of AAV2/1-miS4, -ms5 or -mic into the DCN of wildtype C57Bl/6J mice (n=3 per group). One month post injection, we euthanized the mice and harvested whole cerebellar tissue. We observed effective transduction of AAV2/1 to the PCs by histology and a significant reduction in mouse ataxin-2 levels with mis4 (49.4%, p<0.05) and mis5 (37%, p<0.05) relative to mic by RT-qPCR. To us, this was a reasonable level of knockdown considering that ataxin-2 is expressed throughout the cerebellum and we were only targeting the PCs. Reduction of wildtype ataxin-2 expression in the cerebellum does not affect motor coordination or memory As mis4 and mis5 showed significant knockdown of ataxin-2 in vivo, we pursued long-term studies to determine if reducing wildtype ataxin-2 levels in the PCs was safe and tolerated. We thus repeated the DCN injections in a larger cohort (n=8) of wildtype C57BL/6J mice (2 months of age) with mis4, mis5, mic or saline and followed them using a series of behavioral studies to 6 months of age. We performed a number of motor tests to assess the effects of RNAi on phenotype. We tested the mice on an accelerating rotarod to test motor coordination and

84 69 strength. Mice from all groups performed similarly at 4 and 6 months suggesting that there were no rotarod deficits in the mis4 and mis5 injected mice relative to controls (15a). We performed the hindlimb clasping assay, a classic test done in ataxic mice to assess the progression of cerebellar ataxia (155). We found that none of the injected groups (saline, mis4, mis5 or mic) displayed the hindlimb clasping phenotype exhibited by ataxic animals (data not shown). The ledge test (155) was performed at 6 months and a minimal score was recorded suggesting good motor coordination and balance on the ledge relative to controls (Figure 15b). As SCA2 knockout mice demonstrated learning and memory deficits (204), we tested if reducing wildtype ataxin-2 levels in the cerebellar PCs could induce learning and memory deficits. We performed the fear conditioning assay and found that mis4 and mis5 treated mice displayed normal learning and memory relative to the controls (Figure 15c-e). Sustained reduction of wildtype ataxin-2 expression in the cerebellum does not induce neuropathology Mice were euthanized 3.5 mo. after injection, at 6 months of age, to assess ataxin- 2 mrna reduction and the effects of that reduction on histological readouts. We observed significant sustained reduction of ataxin-2 mrna at 6 months with mis4 (29%, p<0.05) and mis5 (35%, p<0.05) relative to controls mic and saline (Figure 16). In addition, histological assessment demonstrated that the PCs and molecular layer widths were unaffected, and there was no increase in GFAP expression in transduced cerebella (Figure 17,18). We assessed microglial activation using the microglia marker Iba-1, and found no notable increase in expression of the microglial marker Iba-1 in the mis4 and

85 70 mis5 injected cerebella relative to mic or saline (Figure 19), suggesting that reduction in wildtype ataxin-2 expression was well tolerated in the mouse cerebellum. Screening additional RNAi sequences with higher potency We were concerned that our RNAi triggers may not be potent enough for therapeutic efficacy when moving into the SCA2 mouse model, in collaboration with Stefan Pulst. We therefore designed 10 additional sequences using our safe seed design (156) and tested their activity; 3 of 10 showed potent knockdown in vitro (Figure 20). These sequences are designed to target human, rhesus and mouse ataxin-2 mrna. Two sequences, mis7 and mis12 were packaged into AAV2/1 to test knockdown in wildtype mice to assess potency in vivo. mis7 was designed to preferably target human ataxin-2 mrna. AAV2/1 misca2 -S7, -S12 or mic were injected into the DCN of wildtype C57Bl/6J mice (n=3 per group). One month post injection, the mice were euthanized and transduced cerebella collected for RNA analysis. RT-qPCR analysis of mouse ataxin-2 levels showed that both mis7 and mis12 significantly reduced mouse ataxin-2 mrna levels in the cerebellum (Figure 21, p<0.001). Testing mis7 and mis12 sequences in a SCA2 mouse model In collaboration with Dr. Stefan Pulst and Dr. Daniel Scoles at the University of Utah, we used SCA2 transgenic mice expressing 127 pathogenic glutamine repeats in the coding region of human ataxin-2 (50) to test silencing activity. In this model, the transgene is expressed exclusively in the PCs of the cerebellum. The SCA2-Q127 mice display early onset ataxia with progressive rotarod deficits starting at 4 weeks of age.

86 71 These mice also recapitulate the SCA2 pathology seen in patients, with molecular layer thinning and loss of PCs and accumulation of mutant ataxin-2 aggregates (50). We injected AAV2/1 mic into the left DCN and either AAV2/1 -mis7 or -mis12 into the right DCN of the SCA2-127Q mice. One-month post injection, we harvested the transduced cerebellar tissue and analyzed the mrna by RT-qPCR and protein by western blot analysis. Wildtype and mutant ataxin-2 mrna and protein levels were significantly reduced with treatment with mis12 (p<0.001), while mis7 treatment did not result in significant knockdown (Figure 22). Ongoing and future work includes testing if mis12 will provide therapeutic benefit in the SCA2 mouse. Discussion In this study, we screened several RNAi sequences to identify a therapeutic sequence for SCA2 therapy. Silencing the dominant negative mutant gene has proven to be effective in other polyq disease animal models. While allele-specific targeting has been challenging, non-allele specific silencing of the causative gene may be an effective therapeutic strategy for some polyq diseases. However, it is important to understand the consequences of reducing the levels of the wildtype protein in vivo. Studies from the Pulst lab have demonstrated that SCA2 knock-out mice are viable suggesting that wildtype ataxin-2 function is not essential for survival (60). Hence, non-allele specific silencing of ataxin-2 may be a potential therapy for SCA2. We found that reducing wildtype ataxin-2 expression did not cause neuroinflammation or toxicity 4 months post injection in mice. In addition, we noted a preservation of cerebellar morphology and no adverse effects on motor coordination and

87 72 learning suggesting that reducing wildtype ataxin-2 levels in the cerebellum with our RNAi sequences did not cause any toxicity. We found that mis2 was toxic in rhesus macaques. This toxicity could have resulted from a number of factors: a) off-target effects of the mis2 RNAi sequence; b) an immune response resulting from the injections; c) knockdown of wildtype ataxin-2 in the cerebellum was not well tolerated. We noted an increase in microglial activation in the cerebellar tissue as compared to uninjected cerebellar tissue. The off-target effects of mis2 can be analyzed by RT-qPCR and western blot analysis of genes that mis2 could potentially target in the cerebellum. These off-target genes can be identified by using software programs such as sispotr (156) or TargetScan that will identify potential targets of mis2 by blasting the seed sequence of mis2 against the mrna sequence of the respective genome. Sustained reduction of mutant polyq alleles can alleviate disease phenotypes. RNAi-induced knockdown of the mutant SCA1 allele in PCs resulted in a complete rescue of motor and neuropathological phenotypes in a PC specific model of SCA1(58). On the other hand, RNAi-induced knockdown of the mutant SCA3 allele in PCs resulted in a rescue of the neuropathological phenotypes but no rescue of motor phenotypes in a ubiquitously expressing transgenic model (174, 205). Here, we screened RNAi sequences in a SCA2 setting and identified a potent sequence that was able to reduce both mutant and wildtype ataxin-2 levels in vivo. Studies from the Bezprozvanny lab have shown that there is a dysregulation in the calcium signaling pathway in SCA2 and that a calcium stabilizer Dantrolene alleviates PC loss and motor phenotypes in SCA2 mice (206). Further elucidating this mechanism,

88 73 InsP3R suppression was shown to be therapeutic in a SCA2 mouse model just as Dantrolene (131). Here, we chose to directly inhibit mutant protein production and study its effects. Ongoing and future studies in the SCA2 mouse will determine if reduction in mutant and wildtype ataxin-2 in the cerebellum will provide therapeutic benefit. Materials and Methods Animals The University of Iowa Animal Care and Use Committee (IACUC) approved all animal protocols. Wildtype C57BL/6J mice were obtained from The Jackson Laboratory. Mice were housed in a controlled temperature environment on a 12-hour light/dark cycle. Food and water were provided ad libitum. Viral vectors The plasmid expressing mouse U6-driven artificial mirnas were cloned as previously described using DNA oligonucleotides (75). The RNAi sequences are listed in Table 3. Artificial mirna expression cassettes were cloned into paavmcscmvegfp plasmids which co-expressed CMV-driven egfp (ref). Recombinant AAV serotype 2/1 vectors (AAV.miC.eGFP, AAV.miS4.eGFP and AAV.miHS5.eGFP) were generated by the University of Iowa Vector Core facility, as previously described (ref). AAV vectors were dialyzed and resuspended in Formulation Buffer 18 (University of Iowa Gene Transfer Vector Core, Iowa City, IA) and titers (viral genomes/ml) were determined by RT-qPCR.

89 74 Table 3. List of mirna sequences targeting ataxin-2 mir mis2 mis3 mis4 mis5 mis7 mis8 mis9 mis10 mis11 mis12 mis13 mis14 RNAi sequence AAAACTCGAGTGAGCGCTCCAACTGCCCATGCGCCAATCTGTAAAG CCACAGATGGGATTGGCGCATGGGCAGTTGGATCGCCTACTAGTAA AA AAAACTCGAGTGAGCGAGCCAATGATGCTAATGATGATCDGTAAAG CCACAGATGGGGTCGTCATTAGCATCATTGGCGCGCCTACTAGTAAA A AAAACTCGAGTGAGCGAAGCCCATTCCAGTCTCGATAACTGTAAAG CCACAGATGGGTTGTCGAGACTGGAATGGGCTGCGCCTACTAGTAA AA AAAACTCGAGTGAGCGTCCCCACATGGCCCACGTATTTCTGTAAAGC CACAGATGGGAGGTACGTGGGCCATGTGGGGTCGCCTACTAGTAAA A AAAACUCGAGUGAGCGCCCCAAUGAUAUGUUUCGAUAUCUGUAAA GCCACAGAUGGGAUAUCGAAACAUAUCAUUGGGACGCCUACUAGU AAAA AAAACUCGAGUGAGCGCCCCAAUGACAUGUUUCGAUAUCUGUAAA GCCACAGAUGGGAUAUCGAAACAUGUCAUUGGGACGCCUACUAGU AAAA AAAACTCGAGTGAGCGCTCGCCCACCTTCTCGCTATTACTGTAAAGC CACAGATGGGTGGTAGCGAGAAGGTGGGCGAACGCCTACTAGTAAA A AAACTCGAGTGAGCGCCCCACATGGCCCACGTATTTCCTGTAAAGCC ACAGATGGGGAGGTACGTGGGCCATGTGGGACGCCTACTAGTAAAA AAAACTCGAGTGAGCGCCAGTCTCGACAACAGCGCATTCTGTAAAG CCACAGATGGGAATGCGCTGTTGTCGAGACTGACGCCTACTAGTAA AA AAAACTCGAGTGAGCGCAGTACAGAATCCAGTTCGGGGCTGTAAAG CCACAGATGGG CTCCGAACTGGATTCTGTACTTCGCCTACTAGTAAAA AAAACTCGAGTGAGCGGGTGTTCCCTGGCCATCGCTTTCTGTAAAGC CACAGATGGGAAGGCGATGGCCAGGGAACACACGCCTACTAGTAAA A AAAACTCGAGTGAGCGACCCACATGGCCCACGTATTTCCTGTAAAGC CACAGATGGGGGGGTACGTGGGCCATGTGGGGCGCCTACTAGTAAA A

90 75 Table 3 continued mir mis15 mis16 RNAi sequence AAAACTCGAGTGAGCGAAGGTGTTCCCTGGCCATTGTCCTGTAAAGC CACAGATGGGGGCGATGGCCAGGGAACACCTCCGCCTACTAGTAAA A AAAACTCGAGTGAGCGCGGTGTTCCCTGGCCATCGTTTCTGTAAA GCCACAGATGGGAGGCGATGGCCAGGGAACACCTCGCCTACTAGTA AAA AAV injections Wildtype C57Bl/6J mice were injected with AAV vectors as previously reported. For all cerebellar studies transgenic mice were injected bilaterally into the deep cerebellar nuclei (coordinates 6.0 mm caudal to bregma, ± 2.0 mm from midline, and 2.2 mm deep from cerebellar surface) with 4 µl of AAV2/1 virus ( viral genomes/ml) or saline. Following the injections, a topical antibiotic ointment was applied and the mice were allowed to recover according to the University of Iowa Animal Care and Use Committee s (IACUC) guidelines for Post-Anesthesia Monitoring, including monitoring of breathing and muscle tone. Tissue harvesting and immunohistochemistry (IHC) To harvest the cerebella, mice were anesthetized with a ketamine/xylazine mix and transcardially perfused with 20 ml of cold saline. Mice were decapitated, and for histological analyses, brains were removed and post-fixed overnight in 4% paraformaldehyde. Brains were stored in a 30% sucrose/0.05% azide solution at 4 C and

91 76 cut on a sliding knife microtome at 45 µm thickness and stored at 20 C in a cryoprotectant solution. For RT-QPCR analyses, brains were removed, sectioned into 1 mm thick coronal slices using a brain matrix (Roboz, Gaithersburg, MD) and egfp expression was verified. Total RNA was isolated from whole cerebellum using the TRIzol (Life Technologies, Grand Island, NY) extraction. RNA quantity and quality were measured using a NanoDrop ND-1000 (NanoDrop, Wilmington, DE). For IHC, Free-floating sagittal cerebellar sections (45 µm thick) were washed in 1XPBS at room temperature and blocked for 1 h in 5% serum, 0.03% TritonX100 in 1XPBS. Sections were incubated with primary antibody in 3% serum and 0.03% TritonX in 1XPBS overnight at 4 C. Primary antibodies used were polyclonal rabbit anti-iba1 (1:1000; WAKO, Richmond, VA), polyclonal rabbit anti-calbindin (1:2000; Cell Signaling Technology, Danvers, MA) and polyclonal rabbit anti-gfap (1:200; Dako, USA). For fluorescent IHC, sections were incubated with goat anti-rabbit Alexa Fluor 568 (1:1000; Life Technologies, Grand Island, NY) in 3% serum and 0.03% Triton-100 in 1XPBS for 1 hour at room temperature. For DAB IHC, sections were incubated in goat anti-rabbit biotin-labeled secondary antibody (1:200; Jackson Immunoresearch) in 3% serum and 0.03% Triton-X at room temperature for 1 hour. Tissues were developed with VECTASTAIN ABC Elite Kit (Vector Laboratories, Burlingame, CA), according to the manufacturer's instructions. All sections were mounted onto Superfrost Plus slides (Fischer Scientific, Pittsburgh, PA) and cover slipped with Fluoro-Gel (Electron Microscopy Sciences, Hatfield, PA). Images were captured on Leica DMR fluorescence microscope connected to an Olympus DP72 camera using the Olympus DP2-BSW software (Olympus, Melville, NY).

92 77 RT-qPCR First-strand cdna synthesis was performed using 1 µg total RNA (High Capacity cdna Reverse Transcription Kit; Life Technologies, Grand Island, NY) as per manufacturer s instructions. RT-qPCR assays were performed on a sequence detection system using primers/probe sets specific for human ataxin-2, mouse ataxin-2 or mouse β- actin (ABI Prism 7900 HT, TaqMan 2Xuniversal master mix and power SYBR green PCR master mix, Life Technologies, Grand Island, NY). RT-qPCR values were normalized to mouse β-actin. Behavior Analyses All behavior assays were performed at 4 and 6 months post-injection and data is presented here as means ± SEM unless otherwise specified. For all studies, p values were obtained by using one-way analysis of variance followed by Bonferroni post-hoc analysis to assess for significant differences between individual groups, unless indicated otherwise. In all statistical analyses, p < 0.05 was considered significant. Ledge test and hindlimb clasping assays and their scoring parameters are detailed previously (155). Rotarod: Mice were tested on an accelerated rotarod apparatus (model 47600; Ugo Basile, Comerio, Italy). Mice were first habituated on the rotarod for 4 min. Mice were then tested three trials per day (with at least 30 min of rest between trials) for four consecutive days. For each trial, acceleration was from 4 to 40 rpm over 5 min, and then speed maintained at 40 rpm. Latency to fall (or if mice hung on for two consecutive rotations without running) was recorded for each mouse per trial. The trials were stopped

93 78 at 400 s, and mice remaining on the rotarod at that time were scored as 400 s. Two-way analysis of variance followed by Bonferroni post-hoc analysis was used to assess for significant differences. Variables were time and treatment. Fear Conditioning Assay: Mice were placed in a near-infrared fear-conditioning chamber (Med Associates, Inc., St Albuns, VT, USA). Context training consisted of an 8-min protocol with 3 min of exploration (recorded as baseline freezing ), followed by a 1-s shock (0.75 ma) administered at the end of 3 7 min. 24h later (day 2), the mice were placed in the same chamber and the assay was repeated. Increase in freezing was calculated as the freezing time in the first three minutes on day 2 subtracted from baseline freezing. Freezing was defined as an absence of movement other than respiration and scored with Video Freeze software (Med Associates Inc., St. Albans, VT, USA). To assess memory, the context was changed: smooth flooring was placed over the shock grid, a plastic triangular roof was inserted, peppermint scented extract was added to change odor and lighting was dimmed. Mice were left in this new environment and the first three minutes of freezing were taken into consideration. Reduction in freezing was calculated by subtracting freezing time in the first three minutes in the new context on Day 2 from the increase in freezing time.

94 ! (* Figure 12. Reduction of mouse and human ataxin-2 expression in vitro. a) Mouse neuro2a cells or b) HeLa cells were transfected with artificial mirna plasmids (micscrambled control, mis2-s5) or U6 (empty vector control). 24 hours post transfection RNA was harvested and analyzed by RT-qPCR. Results are represented as mean ±SEM (n=3), *p<0.05 or **p<0.01.!

95 ! )+ Figure 13. Reduction of rhesus ataxin-2 mrna levels. AAV2/1 viruses expressing mis2 or mis3 were injected into the rhesus macaque DCN and one month post injection RNA was harvested from several left and right cerebellar lobules, the DCN and the cerebellar cortex (CBC). CBC served as control tissue. Results are represented as mean ±SD, *p<0.05.!

96 ! )" Figure 14. Microglial expression in rhesus cerebellum observed with Iba-1 antibodies. Top panel: Coronal sections showing mis2 injected a) and b) DCN, c) cerebellar lobules Bottom panel: Sagittal sections showing mis3 injected a) and b) DCN and c) cerebellar lobules!

97 ! )# A Latency to fall (s) months 6 months!"#!"#!"#!"#!"#!"#!"#!"# Saline mic mis4 mis5 0 Day 1 Day 2 Day 3 Day 4 Day 1 Day 2 Day 3 Day 4 B Ledge Score !"# Saline mic mis4 mis5 C Baseline freezing (s) !"# Saline mic mis4 mis5 D Increase in freezing (s) !"# Saline mic mis4 mis5 E Decrease in freezing (s) !"# Saline mic mis4 mis5 Figure 15. Behavior assays at 4 and 6 months of age. a) Rotarod at 4 and 6 months of age b) Ledge score c), d) and e) Fear conditioning assay showing no change in learning and memory. Results are represented as mean ± SEM (n=8).!

98 83 Relative mouse ataxin-2 levels miscr *# misca2.s4 *# misca2.s5 Figure 16. Reduction in mouse ataxin-2 mrna levels at 6 months of age. Cerebellar tissue from wildtype mice injected with miscr (mic), mis4 or mis5 was harvested at 6 months of age and subject to RT-qPCR analysis. Results are represented as mean ± SEM (n=4), *p<0.05.

99 ! )% Figure 17. Preservation of cerebellar morphology in mis4 and mis5 treated wildtype cerebella. IHC showing transduction of the PCs (green) and calbindin (red) expression in a) mic b) mis4 and c) mis5 cerebellar sections.!

100 ! )& Figure 18. Absence of gliosis in the transduced cerebellum. Left panels: GFAP expression (red) in the cerebellum treated with mic, mis4 or mis5 Right panels: Corresponding transduced tissue (green)!

101 ! )' Figure 19. Microglial expression in the cerebellum. IHC using anti-iba-1 antibodies showing microglial expression in mic, mis4 or mis5 injected cerebellar tissue.!

102 ! )( Figure 20. Screen of additional RNAi sequences in HeLa cells. 24 hours post transfection of the mirna plasmids into HeLa cells, RNA was extracted and subject to RT-qPCR. Results are represented as mean ± SEM (n=3), ***p<0.001.!

103 ! )) Figure 21. Reduction of mouse ataxin-2 mrna in wildtype cerebellum. One month post injection of AAV2/1 expressing mic, mis7 or mis12, cerebellar RNA was extracted and subject to RT-qPCR. Results are represented as mean ± SEM (n=3), ***p<0.001.!

104 ! )* Figure 22. Reduction of mutant (human) and wildtype (mouse) ataxin-2 in the SCA2 mouse cerebellum. One month post injection of AAV2/1 expressing mic, mis7 or mis12 into the DCN of SCA2 transgenic mice, cerebellar RNA was extracted and subject to RT-qPCR. Results are represented as mean ± SEM (n=3), ***p<0.001.!

105 90 CHAPTER V: FUTURE DIRECTIONS RNAi therapy is ideal for diseases where the mutant protein acquires a toxic gain of function. RNAi therapy has been effective in silencing such mutant genes and shown therapeutic efficacy in animal models of a number of neurological and retinal diseases such as the polyq diseases, Alzheimer s disease, Parkinson s disease, autosomal dominant retinitis pigmentosa, among others. My thesis work has focused on identifying effective RNAi therapies for SCA2 and SCA7. Future work described in this chapter is written with the aim to move these RNAi therapies to the clinic. SCA7 therapy Targeting the SCA7 brain This work identified that non-allele specific silencing of ataxin-7 by targeting the cerebellar Purkinje cells (PC) with AAV2/1 mis4 significantly improved a number of phenotypes in the BAC-SCA7-92Q mouse (Chapter 2). However, complete rescue of these phenotypes was not observed. We hypothesize that this could be due to two important factors. One, other cell types in the cerebellum such as the inferior olive (IO) and Bergmann glia (BG) can contribute to the pathogenesis of the disease, and a more robust rescue will require targeting these cell types as well. The cerebellar PCs form an intricate network with the IO and BG. The PCs receive input from the IO via the climbing fibers while the BGs eliminate excess glutamate from the PC synapses. Albert La Spada s group showed that Cre-mediated excision of mutant ataxin-7 in the PCs, IO and BG significantly rescued ataxic phenotypes in SCA7 mice as compared to excision

106 91 from the PCs and IO alone or BG alone indicating that in this transgenic model, all these cell types contribute to SCA7 disease pathogenesis (26). In a PC specific SCA7 model (25), it is possible that a more robust rescue would be seen as we are mainly targeting the PCs with our RNAi trigger. However, in the human disease, the mutant protein is expressed ubiquitously and the BAC-SCA7-92Q mouse recapitulates this important aspect. Thus similar to the human disease, the mutant protein is expressed in all neurons and glia of the CNS, nuclear inclusions are present in affected and unaffected regions of the brain, the IO contributes to pathogenesis (in human patients, there is a loss of IO neurons) and gliosis is observed in the cerebellum (153). Hence, testing RNAi in this model provides a better understanding of the extent of therapeutic efficacy achievable in patients. Two, although we delivered the mis4 RNAi sequence pre-symptomatically, it is possible that irreversible neurological damage occurs earlier. The mature cerebellum is fully developed only at ~P15 in the mouse. Although it is not fully understood if and how polyq proteins affect cerebellar development, there is evidence that polyq proteins disrupt PC function prior to symptom onset and observable neuropathology. In SCA2 mice, deficits in the PC firing rate and gene expression changes are observed prior to the onset of ataxic phenotypes (50). Studies in SCA3 mice also showed that changes in PC function occurred prior to neurodegeneration (207), and in SCA1 mice, there is a deficit in PC firing frequency at 5 weeks although there is no PC loss or a rotarod phenotype at this early time point (208). In other ataxia disease mice, the PC morphology appears normal even though PC function is disrupted (209). In the BAC-SCA7-92Q mice, PC histology appears normal at 10 weeks of age, but PC firing and gene expression changes

107 92 have not been analyzed at this and earlier time points. Thus, it is possible that like in the other polyq SCAs, neurological deficits occur during cerebellar development or prior to treatment in the BAC-SCA7-92Q mice. To address the first hypothesis, we will target the BG, IO and PCs with an AAV serotype that has tropism for all these cell types. The small size of AAV is ideal for gene transfer of RNAi sequences in vivo. AAV9 is a newly discovered serotype that can efficiently transduce neurons and glia in the brain and spinal cord as well as peripheral tissues by intravenous or intravascular injections (110). Delivery of AAV9 to neonatal rhesus macaques and mice by intravascular or intravenous injections can efficiently transduce the CNS including astrocytes and cerebellar neurons (110, ). AAV9- mediated systemic delivery of RNAi rescued neuropathological features and weight loss in a Huntington disease mouse (213). An advantage of delivering AAV9 mis4 to neonatal mice is that we can target all of the affected cells possibly prior to the onset of pathogenesis and not just prior to symptom onset. We will deliver AAV9 mis4 or AAV9 mic or saline intravenously to neonatal SCA7 mice as described previously (110). We will first confirm expression in the PCs, BG and IO in the cerebellum at 4 weeks of age when the cerebellar PCs are physiologically mature (214). We will re-inject a larger cohort of mice and then perform behavior testing as described in Chapter 2 (rotarod, ledge, clasping and gait assays) at 10-week intervals. We will also perform a survival study of the mis4 treated mice versus controls. Histological studies will be performed once the mice are euthanized at either ~ 2 years (life span of normal mice) or at end stage disease to assess the nuclear inclusions, GFAP expression, microglial activation and PC morphology as described in Chapter 2. Although La Spada s group observed a significant

108 93 rescue in the mice with 50% less mutant ataxin-7 expression, complete rescue was not observed in their study either (26). These proposed experiments will address if other cell types apart from the BGs, PCs and IO are important to target for treatment and if neonatal delivery can result in a more robust rescue than observed earlier. To address the second hypothesis, we will perform microarray experiments to analyze gene changes occurring during postnatal development and prior to 7 weeks of age. We will collect cerebellar tissue from SCA7 and wildtype mice at different time points in postnatal development and adulthood, harvest RNA and subject the cdna to microarray analysis. Gene expression changes will be analyzed and expression at the various time points will be compared to assess what gene expression changes occur in the SCA7 cerebellum prior to 7 weeks (when we delivered AAV mis4 in our studies). We will also perform electrophysiological studies to study the PC firing rate at early time points (2, 4, 6, 8 weeks) in the SCA7 mice. These experiments will determine if postnatal developmental deficits occur in SCA7 mice and possibly in SCA7 patients as well. In addition to identifying mis4 in our initial screen (described in Chapter 2), we also tested an RNAi sequence (micag9) that specifically targets the expanded CAG repeat (made in the Davidson lab by Dr. Mas Monteys). AAV2/1 -micag9, -mis4 or - mic were injected stereotaxically into the DCN of BAC-SCA7-92Q mice and one month later RNA was harvested for RT-qPCR analysis. We found that micag9 was allelespecific and targeted mutant ataxin-7 while leaving the wildtype ataxin-7 expression intact (Figure 23). We will test this RNAi sequence long-term as done previously with mis4 in the SCA7 mice to determine its therapeutic efficacy and compare it with mis4 mediated rescue. Gene expression changes performed from micag9 tissue will be

109 94 compared with mis4 gene expression changes to shed light on the function of wildtype ataxin-7. Targeting the SCA7 retina SCA7 patients with >52 CAG repeats in ataxin-7 often experience vision loss due to macular degeneration, a degeneration of the cone cells followed by rod cell death. Thus, we tested the mis4 RNAi sequence in the BAC-SCA7-92Q mouse retina and demonstrated for the first time that reducing the levels of mutant and wildtype ataxin-7 expression by 50% in the photoreceptors of these mice does not cause any adverse effects long-term (Chapter 3). However, the BAC-SCA7-92Q mouse does not exhibit impairments using a standard electroretinogram (ERG) nor are there histological phenotypes despite transgene expression in the retina. Recently we (with Dr. Arlene Drack and Sajag Bhattarai) reassessed the BAC-SCA7-92Q mouse retina with a different ERG protocol (escalating flicker ERG) to assess cone function alone; SCA7 is classically a cone-rod dystrophy. With this ERG protocol, we tested end stage mice and identified a significant cone ERG decline at ~40 weeks of age (Figure 24). Ongoing studies include repeating the RNAi injections (in collaboration with Dr. Drack and Sajag Bhattarai) in new animals at 7 weeks of age in the SCA7 retina to assess if reducing ataxin-7 expression will prevent the onset of this cone-erg phenotype in the BAC-SCA7-92Q mice. Cone-ERGs will be performed prior to the RNAi injections to get a baseline and then repeated at 10, 20, 30 and 40 weeks of age, as we do not know the time of onset of this phenotype. We expect to see a significant improvement in the cone-erg phenotype in the mis4 treated SCA7 mice compared to mic or saline treated mice with time.

110 95 Our in situ hybridization experiments indicate that Atxn7 mrna is expressed broadly in the retina (Figure 11). We however focused on targeting the photoreceptors with AAV2/1 as the disease is a cone-rod dystrophy. In SCA7 mouse models with a robust retinal phenotype such as the SCA7 knockin (126) and R7E models (25), thinning of the outer, inner nuclear cell and ganglion cell layers is observed. Thus, apart from the outer nuclear layer, it may be required to target the inner cell and ganglion cell nuclei as well to observe a therapeutic effect with mis4. To achieve pan-retinal transduction of all retinal cells, we will test AAV serotypes with tropism for the entire retina. There are a number of AAV serotypes generated recently by a number of groups for better transduction of the murine and rhesus retinas. Tyrosine mutated AAV2 reduces proteasome targeting and degradation of AAV and shows a transduction pattern that may be optimal for SCA7 retina therapy (112). Recently, Dalkara et al., demonstrated that an AAV serotype (7m8) derived by directed evolution could transduce all murine retinal cells by intravitreal delivery (113). We also have in hand AAV2/1 engineered with lysine-arginine mutations that also help avoid proteasome degradation (215, 216). We will test the lysine-arginine mutant AAV2/1, the tyrosine mutated AAV2, and compare these with the AAV2-7m8 variant by both intravitreal and subretinal injections. As with the tyrosine mutations in AAV2, we expect the lysine-arginine mutations to provide for a broader transduction pattern. We will evaluate the AAVs for optimal transduction in the SCA7 retina because disease processes may alter AAV tropism as noted earlier (217). We will make use of the most optimal AAV serotype to express mis4 in all cells of the SCA7 retina. We will confirm expression in all cell types of the retina and assess safety of the serotype long-term as described in Chapter 3.

111 96 Gene expression changes in mis4 treated SCA7 cerebellum and retina The La Spada group also demonstrated that Cre-mediated excision of mutant ataxin-7 by 50% just after disease onset significantly delayed disease progression, providing evidence that SCA7 patients could benefit from receiving treatment even after disease onset (27). To determine the full impact of non-allele specific silencing with mis4, we will first analyze global gene expression changes in the cerebellum and retina collected from mis4 and control treated mice. We will perform RNA sequencing using RNA harvested from a) wildtype uninjected b) wildtype injected with mis4 c) SCA7 mis4 injected and d) control SCA7 cerebella / retinas. This data will help to: i) determine the extent of baseline differences in gene expression, mrna splicing, and noncoding RNA expression that exists between normal and SCA7 uninjected mice. Pathway analysis of both normalized data and differentially expressed genes will shed light on the global impact of mutant Atxn7 expression. ii) We will test if reducing the levels of wildtype Atxn7 broadly induces gene expression changes compared to the uninjected animal. Of specific interest will be segregation profiles of the three groups. Will mis4 injected SCA7 tissue be more similar to normal brains than their uninjected counterparts? iii) We hope to gain insight into the function of mutant Atxn7 in the SCA7 cerebellum and retina. RNAi treatment will reduce both the mutant and wildtype protein, allowing us to study in greater detail the function of SCA7 protein in the brain and retina, iv) determine genes or non-coding RNAs whose expression changes as a function of the seed sequence of the RNAi trigger, v) determine if reduction of Atxn7 alleviates known SCA7-induced transcriptional dysregulation.

112 97 SCA2 Therapy In our studies (described in Chapter 4), we determined that a modest but significant reduction of wildtype ataxin-2 in the mouse cerebellum is well tolerated. Additionally, we identified a potent RNAi sequence, mis12, that could reduce the levels of mutant and wildtype ataxin-2 in a SCA2 mouse. To assess the therapeutic efficacy of mis12, ongoing studies include testing mis12 in the SCA2-127Q transgenic mouse in collaboration with Dr. Stefan Pulst. The SCA2-127Q mice have a rotarod phenotype observed after 4 weeks of age and declining progressively with age (50) SCA2 mice at 5 weeks of age will be injected bilaterally into the DCN with either AAV2/1 mis12, AAV2/1 mic or saline to target the PCs as done before. We will follow these mice until 40 weeks of age. Prior to doing the injections, a baseline rotarod phenotype will be established. Similar to our SCA7 study, at 20 and 40 weeks of age, the SCA2 mice and their wildtype littermates (equal numbers in each group) will be subjected to the rotarod, ledge, clasping and gait assays (Figure 25). In addition, we will also assess their learning and memory to determine if reducing wildtype ataxin-2 levels affects this phenotype, as done previously in Chapter 4, as ataxin-2 knockout mice have learning and memory deficits. A decline in the PC firing rate begins at 6 weeks of age in the SCA2-127Q mice. Hence, we will also perform electrophysiological studies in collaboration with Dr. Pulst s lab at 20 and 40 weeks to determine if we can resolve this decline with long-term expression of mis12. At 40 weeks, we will harvest RNA and protein from the cerebellum to confirm knockdown of ataxin-2 expression. We will also harvest tissue to assess toxicity and PC morphology as done previously. Aggregate formation will also be

113 98 assessed in the mis12 versus controls, as perinuclear aggregates have been observed to increase in size and number with age in these mice (50). The RNA harvested will be subjected to RNA sequencing to determine gene expression changes to evaluate off-target effects of mis12, the effect of wildtype ataxin-2 reduction and assess therapeutic efficacy as proposed previously for SCA7. Once we determine that mis12 is therapeutically effective in terms of improving the motor and molecular phenotypes, does not cause toxicity and is well tolerated, we will then move towards safety testing in non-human primates as described below. Moving RNAi therapy towards the clinic As a first step towards moving these therapies into the clinic we will evaluate the safety, efficacy, biodistribution and appropriate dose of the AAV2/1 virus in a large species with a brain size closer to humans, such as the rhesus macaque. Rhesus macaques have been used to test the safety of AAV-mediated RNAi therapy for Huntington s disease and 6 month suppression of wildtype Huntingtin was found to be safe and well tolerated (129, 130). We will test both AAV2/1 misca2.s12 and AAV2/1 misca7.s4 for both SCA2 and SCA7 therapy, respectively. Three different doses (a low, medium and high) of AAV2/1 RNAi will be delivered by stereotaxic injections to one cerebellar hemisphere and the contralateral side will be injected with AAV2/1-miC. Three different sites (DCN, fastigial and interposed nuclei) will be injected per hemisphere to ensure greater spread of the virus. One month post injection, the animals will be euthanized and tissue from the IO, DCN and cerebellar cortex will be collected for RNA, protein and histology. RNA and protein will be used to assess knockdown by RT-qPCR and western

114 99 blots, respectively. Tissue sections will be cut using the cryostat and used to assess toxicity by immunohistochemical staining for IBA-1 (microglia marker) and GFAP (glial marker). Biodistribution of the virus will be assessed by knockdown in the tissue harvested, egfp expression from the virus and semi-quantitative PCR for the RNAi sequence. Once we determine the dose, distribution and safety of the RNAi sequence short-term, we will then propose to do long-term safety testing in the rhesus macaques. We will also test the safety of mis4 in the rhesus retina for SCA7 therapy. AAV2 and AAV2-7m8 have been tested in the rhesus macaque by intravitreal delivery. We will perform subretinal and intravitreal injections in the rhesus macaque retina (based on the tropism of the AAV serotype determined from experiments above). We will perform knockdown analysis from the RNA harvested and assess toxicity and transduction efficiency by histology. Together these studies will determine if the RNAi sequences we designed are safe for testing in SCA2 and SCA7 patients, an important preclinical step towards therapy.

115 Figure 23. Allele-specific silencing of ataxin-7. AAV2/1 expressing artificial mirnas targeting the mutant human allele (micag9) or targeting both human and mouse alleles (misca7.s4) were injected stereotaxically into the deep cerebellar nuclei of SCA7 mice (n=3; 2E^12vg/ml) in our work targeting the ataxic phenotypes in this model. RNA was harvested 4 weeks later and analyzed by RT-qPCR. Significant non-allele specific silencing is obtained with misca7.s4 (upper & lower panels) and allele-specific silencing of ataxin-7 is seen with micag9 (lower panel). Data normalized to beta-actin (*, P < 0.05, Students t-test). 100

116 Figure 24. SCA7 mice show deficits in cone-erg response week old SCA7 mice (n>3) were subjected to the escalating ERG protocol. Error bars represent SEM. ANOVA analysis with Bonferroni post-hoc tests * p<0.05, *** p<

117 102 Figure 25. Experimental design for long-term SCA2 studies proposed SCA2 mice will be injected at 5 weeks of age and followed to 40 weeks. Behavior testing will be done at 20 and 40 weeks and euthanized at 40 weeks to harvest tissue for RNA, protein and immunohistochemistry.