Role of Gigaxonin in the Regulation of Intermediate Filaments: a Study Using Giant Axonal Neuropathy Patient-Derived Induced Pluripotent Stem Cell-Motor Neurons Bethany Johnson-Kerner Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy under the Executive Committee of the Graduate School of Arts and Sciences COLUMBIA UNIVERSITY 2013 © 2012 Bethany Johnson-Kerner All rights reserved Abstract Role of Gigaxonin in the Regulation of Intermediate Filaments: a Study Using Giant Axonal Neuropathy Patient-Derived Induced Pluripotent Stem Cell-Motor Neurons Bethany Johnson-Kerner Patients with giant axonal neuropathy (GAN) exhibit loss of motor and sensory function and typically live for less than 30 years. GAN is caused by autosomal recessive mutations leading to low levels of gigaxonin, a ubiquitously-expressed cytoplasmic protein whose cellular roles are poorly understood. GAN pathology is characterized by aggregates of intermediate filaments (IFs) in multiple tissues. Disorganization of the neuronal intermediate filament (nIF) network is a feature of several neurodegenerative disorders, including amyotrophic lateral sclerosis, Parkinson’s disease and axonal Charcot-Marie-Tooth disease. In GAN such changes are often striking: peripheral nerve biopsies show enlarged axons with accumulations of neurofilaments; so called “giant axons.” Interestingly, IFs also accumulate in other cell types in patients. These include desmin in muscle fibers, GFAP (glial fibrillary acidic protein) in astrocytes, and vimentin in multiple cell types including primary cultures of biopsied fibroblasts. These findings suggest that gigaxonin may be a master regulator of IFs, and understanding its function(s) could shed light on GAN as well as the numerous other diseases in which IFs accumulate. However, an interaction between gigaxonin and IFs has not been detected and how IF accumulation is triggered in the absence of functional gigaxonin has not been determined. To address these questions I undertook a proteomic screen to identify the normal binding partners of gigaxonin. Prominent among them were several classes of IFs, including the neurofilament subunits whose accumulation leads to the axonal swellings for which GAN is named. Strikingly, human motor neurons (MNs) differentiated from GAN iPSCs recapitulate this key phenotype. Accumulation of nIFs can be rescued by reintroduction of gigaxonin, by viral delivery or genetic correction. GAN iPS-MNs do not display survival vulnerability in the presence of trophic factors, but do display increased cell death in the presence of oxidative stress. Preliminary experiments suggest that in iPS-MNs nIFs are degraded by contributions from both the proteasome and lysosome. Gigaxonin interacts with the autophagy protein p62 which has been implicated in the clearance of ubiquitin aggregates by the lysosome, and this interaction is greatly enhanced in conditions of oxidative stress. My data provide the first direct link between gigaxonin loss and IF aggregation, and suggest that gigaxonin may be a substrate adaptor for the degradation of IFs by autophagy, pointing to future approaches for reversing the phenotype in human patients. Table of contents Chapter 1. General Introduction 1 Part I. Giant axonal neuropathy in humans 3 Part II. Intermediate filaments in normal neuronal function and disease 21 Part III. Protein turnover: UPS and autophagy 46 Part IV. Existing cellular and mouse models of giant axonal neuropathy 61 Part V. Proposed functions of gigaxonin 73 Part VI. Use of ES and iPS-derived neurons for disease modeling 86 Chapter 2. Mass spectrometry screen to identify binding partners of gigaxonin 96 Chapter 3. Studies of iPS-derived motor neurons from GAN patients 124 Chapter 4. Viral and genetic rescue experiments in GAN iPS-derived motor neurons 146 Chapter 5. Survival studies on HBG1 ES- and GAN iPS-derived motor neurons 169 Chapter 6. Preliminary studies on novel binding partners and functions of gigaxonin 198 Chapter 7. General Discussion 220 Chapter 8. Experimental Procedures 240 Chapter 9. References 248 i Figure List Figure 1.1. IF changes in GAN. Figure 1.2. Human GAN pathology. Table 1.3. Differential diagnoses of GAN. Figure 1.4. Reported gigaxonin mutations. Table 1.5. Classification of intermediate filaments and their expression patterns. Figure 1.6. Schematic of IF structure. Figure 1.7. Individual steps of IF assembly. Figure 1.8. IF interactions with the cytoskeletal network. Table 1.9. Disease-specific alterations of intermediate filaments. Figure 1.10. Lysosomal pathways. Figure 1.11. The ubiquitin-proteasome system. Table 1.12. Mutations in protein degradation components linked to neurological disease. Table 1.13. Summary of existing Gan-/- mouse models. Figure 1.14. Gan-/- mouse pathology. Figure 1.15. Gigaxonin sequence and homology with other BTB/Kelch proteins. Figure 1.16. Regulation of Nrf2 by Keap1. Table 1.17. iPS models of motor diseases. Figure 2.1. Design of the proteomic screen to determine gigaxonin binding partners. Table 2.2. Select Kelch domain interacting partners. Figure 2.3. MAP1B, MAP8 and TBCB expression in HEK293 cells. Figure 2.4. Vimentin and peripherin interact with the Kelch domain of gigaxonin. Figure 2.5. Evidence for interaction between gigaxonin and Cul3. Table 2.6. Full list of hits from GAN proteomic screen. Table 3.1. GAN patient demographics. Figure 3.2. Characterization of GAN patient iPS lines. Figure 3.3. Levels of viral and pluripotency gene expression in GAN iPS lines measured by qRT-PCR. ii Figure 3.4. Gigaxonin levels in GAN fibroblasts, iPSCs and MNs are severely reduced as compared with control. Figure 3.5 Generation of iPS-derived motor neurons from GAN patients. Figure 3.6. Calcium imaging of GAN and control neurons. Figure 3.7. GAN motor neurons exhibit selectively elevated levels of IFs. Figure 3.8. GAN motor neurons have selectively elevated levels of peripherin. Figure 3.9. qPCR of IF and non-IF proteins in iPS-MNs. Table 4.1. AAV-mediated gene delivery approaches for the treatment of CNS disorders. Figure 4.2. Lentiviral construct design and titer. Figure 4.3. Gigaxonin overexpression rescues elevated peripherin levels in GAN iPS-MN cultures. Figure 4.4. Gigaxonin overexpression is not toxic to neurons. Figure 4.5. Testing of different AAV serotypes and promoters on motor neurons. Figure 4.6. Preliminary evidence for AAV9-GAN rescue of elevated peripherin levels in GAN iPS-MN cultures. Figure 4.7. Strategy for making rescue GAN lines stably expressing gigaxonin. Figure 4.8. Verification of FLAG-GAN positive iPS clones. Figure 4.9. Genetic replacement of gigaxonin rescues elevated levels of peripherin and NF-L. Figure 4.10. Differentiation efficiency of rescue lines. Table 5.1. Neurotrophic factor families and the major members. Figure 5.2. Trophic factor receptor signaling promotes cell survival. Table 5.3. Neurotrophic factors as modifiers of motor neuron disease. Figure 5.4. Two approaches to a human MN survival assay. Figure 5.5. Optimization of survival assay media. Figure 5.6. UFdU survival assay is reproducible and gives a ~5-fold difference between the NTF and basal condition. Figure 5.7. Individual neurotrophic factor titrations. Figure 5.8. GAN neurons do not display reduced survival. Figure 5.9. GAN Hb9::GFP reporters. iii Figure 6.1. nIFs are degraded over a 24 hr period in iPS-MN cultures. Figure 6.2. Peripherin is predominantly degraded by the lysosome whereas NF-L is degraded by the proteasome and lysosome. Figure 6.3. NF-L is monoubiquitinated. Figure 6.4. Evidence for an interaction between gigaxonin and p62. Table 6.5. MS hits that may play a role in GAN pathophysiology. Figure 6.6. Evidence for complex formation between gigaxonin, vimentin and Hsp90. Figure 6.7. GAN neurons show increased sensitivity to oxidative stress. Figure 7.1. Working model for the role of gigaxonin in nIF turnover. iv Acknowledgements This project would not be possible but for the incredible assistance and input from numerous individuals. First, my mentors Christopher Henderson and Hynek Wichterle have provided encouragement, insight and support throughout this project. I am very grateful for the opportunity to do my doctoral training in their laboratories. Second, this project would not be possible without the incredible cooperation of the GAN patients and their families, primarily organized by Lori Sames of Hannah’s Hope Fund (HHF). HHF connected me with tissue samples and reagents, as well as to a network of scientists studying GAN whose input was invaluable. Drs. Wendy Chung, Rudy Van Coster, and Paul Maertens generously provided GAN patient fibroblasts, and Dr. Doug Sproule was very generous in allowing me to see GAN patients in the clinic with him. I would also like to thank Valerie Estess and Meredith Hulbert of Project ALS for their incredible support over the years. Many scientists provided help along the way. Scott Noggle, Faizzan Ahmad, David Kahler and all the members of the New York Stem Cell Foundation provided critical assistance. Lukas Cermak of the Michele Pagano laboratory at NYU provided much help with the mass spectrometry portion of my thesis. Jean-Pierre Julien, Ron Liem, and Mark Hannink generously provided constructs, and Pascale Bomont, Ron Liem, Howard Worman, Saul Silverstein and Ai Yamamoto provided antibodies, for which I am very grateful. All the members of the Project ALS, Henderson and Wichterle labs have provided input and support over the years. In particular Alejandro Garcia Diaz provided technical help with the molecular biology in my thesis, and J Palmer Greene, a NYSTEM undergraduate, helped by compiling a GAN patient database as well as with Western blot quantification. Steven Gray v provided AAV reagents. Many members of Livio Pellizzoni’s laboratory provided biochemistry input, including Francesco Lotti and Darrick Li. Ai Yamamoto and Leora Fox provided input on the autophagy experiments. A special thanks to those who gave time when they really had none to spare, especially Alan Tenney and Kevin Kanning in the Henderson lab.