PREMATURE TERMINATION CODON READTHROUGH

RESTORES PROGRANULIN EXPRESSION IN PRECLINICAL

MODELS OF FRONTOTEMPORAL DEMENTIA AND

NEURONAL CEROID LIPOFUSCINOSIS

by

Jonathan Frew

B.Sc., The University of British Columbia, 2016

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

in

THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES

(Experimental Medicine)

UNIVERSITY OF BRITISH COLUMBIA

(Vancouver)

April 2021

© Jonathan Frew, 2021

ii

The following individuals certify that they have read, and recommend to the Faculty of

Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:

Premature termination codon readthrough restores progranulin expression in preclinical models of frontotemporal dementia and neuronal ceroid lipofuscinosis

submitted by Jonathan Frew

in partial fulfillment of the requirements for

the degree of Doctor of Philosophy

in Experimental Medicine

Examining Committee:

Haakon B. Nygaard, Neurology

Supervisor

Robin G-Y. Hsiung, Neurology

Supervisory Committee Member

Stephanie Willerth, Biomedical Engineering

Supervisory Committee Member

Eric Jan, Biochemistry & Molecular Biology

University Examiner

Neil Cashman, Neurology

University Examiner

Additional Supervisory Committee Members:

Michel Roberge, Biochemistry & Molecular Biology

iii

Abstract

Frontotemporal dementia (FTD) is a devastating and progressive disorder and a

common form of early-onset dementia. There are currently no disease-modifying

therapies available, signifying the need for new therapeutic approaches. Progranulin

(PGRN) haploinsufficiency due to autosomal dominant mutations in the progranulin

gene (GRN) is a major cause of familial FTD (FTD-GRN), with nearly a quarter of these

genetic cases resulting from a nonsense mutation. Nonsense mutations introduce

premature termination codons (PTCs) that can be therapeutically targeted by

compounds allowing readthrough, and aminoglycoside antibiotics are known to be

potent PTC readthrough drugs. When this research project was initiated, restoring

PGRN through PTC readthrough had not been explored as a therapeutic intervention in

FTD-GRN.

We used human induced pluripotent cell (hiPSC) lines bearing clinical nonsense

mutations spanning the GRN coding region (S116X+/-, R418X+/-, R493X-/- KI) to evaluate

G418 and novel PTC readthrough enhancer (CDX-series) combination treatments. Our

aim was to demonstrate proof-of-concept GRN PTC readthrough and lower the required

dose of G418 to address the known toxicity of traditional aminoglycoside PTC

readthrough agents.

Screening in HEK293 cells expressing nonsense mutant (S116X, R418X, R493X) GRN

expression constructs found PTC readthrough combination treatment with G418 and

CDX5-288 enhancer most potently induced GRN readthrough. We demonstrated in vivo

iv

proof-of-concept GRN PTC readthrough by performing single intracerebroventricular

(ICV) injections of G418 with or without CDX5-288 enhancer in our GRN R493X adeno-

associated virus-based mouse model. Combination treatment in hiPSC-derived isogenic

R493X-/- KI cortical neurons significantly restored PGRN levels and normalized

overexpression of the mature form of the lysosomal enzyme cathepsin D.

We attempted to achieve in vivo PTC readthrough of GrnR493X through repeated ICV

administrations of G418 in GrnR493X/R493X mice. However, G418 doses capable of

eliciting PTC readthrough in these mice were associated with significant neurotoxicity.

We next conducted further neuropathological characterization of lysosomal dysfunction,

neuroinflammation, and neurodegeneration in GrnR493X/R493X mice, identifying several

phenotypes recently reported in Grn-/- mice, including decreased thalamic excitatory

neuronal density. Taken together, our findings suggest that PTC readthrough may be a

potential therapeutic strategy for FTD caused by GRN nonsense mutations and support

further investigations into novel readthrough drugs with improved tolerability.

v

Lay Summary

Neurodegenerative diseases represent a growing burden to Canadian society.

Frontotemporal dementia (FTD) is a neurodegenerative syndrome that is the second

most common cause of early-onset dementia after Alzheimer’s disease. Approximately

25% of FTD cases are caused by mutations in the progranulin gene (GRN), meaning

that patients only have 50% of the normal levels of the progranulin protein (PGRN). A

nonsense mutation is a type of loss-of-function mutation that causes a protein to be cut

short in length by a process known as premature termination. Nonsense mutations in

GRN cause approximately 7.5-10% of FTD cases (FTD-GRN). We have discovered

drugs that silence the effect of nonsense mutations by preventing premature

termination. The goal of this work was to test whether these drugs promote full-length

PGRN production in cells derived from FTD-GRN patients and in FTD-GRN mouse

models. This work is important because there are currently no effective treatments for

FTD.

vi

Preface

Parts of chapters 3 and 4 were published in the following articles:

1) Frew, J., Baradaran-Heravi, A., Balgi, A.D., Wu, X., Yan, T.D., Arns, S.,

Shidmoossavee, F.S., Tan, J., Jaquith, J.B., Jansen-West, K.R., Lynn, F.C., Gao, F.B.,

Petrucelli, L., Feldman, H.H., Mackenzie, I.R., Roberge, M., Nygaard, H.B. (2020).

Premature termination codon readthrough upregulates progranulin expression and

improves lysosomal function in preclinical models of GRN deficiency. Molecular

Neurodegeneration, 15(21). https://doi.org/10.1186/s13024-020-00369-5. This article

was minimally modified to fit formatting guidelines of this thesis and was reused under

the Spring Nature Creative Commons license (CC BY 4.0).

2) Frew, J., Wu, X., Hsiung, G.Y., Feldman H.H., Mackenzie, I.R., Nygaard, H.B. (2019).

Generation of an induced pluripotent stem cell line (UBCi001-A) from a presymptomatic

individual carrying the R418X progranulin gene mutation. Stem Cell Research,

41(101582). https://doi.org/10.1016/j.scr.2019.101582. This article was minimally

modified to fit formatting guidelines of this thesis and was reused under the Elsevier

B.V. Creative Commons license (CC BY-NC-ND 4.0).

The design of all research, data analysis, and manuscript preparation were an original

intellectual product of the author, myself, with the guidance and mentorship of Dr.

Haakon B. Nygaard. I performed all experiments myself with the following exceptions:

• GRN expression vector cloning and HEK293 experiments that generated data

presented in Figures 3-1, 3-2, and 3-4 were performed by A. Baradaran-Heravi

and A.D Balgi.

vii

• AAV-GRN-R493X-V5 vector packaging was performed by K.R. Jansen-Weston.

• AAV-GRN-R493X-V5 RIPA-soluble mouse brain lysate V5 automated capillary

electrophoresis western analysis presented in Figure 3-7 was performed by A.D

Balgi.

• hiPSC CRISPR/Cas9 gene editing guidance was provided by F.C. Lynn.

• X. Wu prepared DNA samples for Sanger sequencing.

• T.D Yan assisted with some hiPSC tissue culture work.

• S. Arns, F.S. Shidmoossavee, J. Tan, and J.B Jaquith were involved in the

synthesis of CDX-series compound derivatives.

Parts of chapter 5 were published in the following article:

1) Frew, J., Nygaard, H.B. (2021). Neuropathological and behavioural characterization

of aged Grn R493X progranulin-deficient frontotemporal dementia knock-in mice. Acta

Neuropathologica Communications (in press). This article was minimally modified to fit

formatting guidelines of this thesis and was reused under the Elsevier B.V. Creative

Commons license (CC BY 4.0).

The design of all research, data analysis, and manuscript preparation were an original

intellectual product of the author, myself, with the guidance and mentorship of Dr.

Haakon B. Nygaard. I performed all experiments myself with the following exception:

• Dr. Kunho Choi isolated and treated Grn+/+ and GrnR493X/R493X MEFs and

conducted the Pgrn linker-1 antibody PTC readthrough western blot assay

presented in Figure 5-16 (unpublished data).

viii

Animal studies were performed in compliance with the Canadian Council on Animal

Care and were approved by the University of British Columbia (UBC) Animal Care

Committee under protocols A19-0623 (FTD mouse model breeding), A19-0062

(generation of mouse embryonic fibroblasts), A17-0225 (AAV injection of P0 C57BL/6J

pups) and A16-0161 (CNS delivery of PTC readthrough drugs). hiPSC studies were in

compliance with the Tri-Council Policy Statement on ‘Ethical Conduct for Research

Involving Humans’, as well as ICH Good Clinical Practice Guidelines and were

approved by the UBC under protocol HO7-03022. I have successfully completed the

following Canadian Council on Animal Care courses: Rodent Biology and Husbandry,

Introduction to Rodent Anesthesia Inhalable and Injectable, Introduction to Rodent

Aseptic Surgery – Full Sterile and Sterile Tip (Online animal care training program

certificate number: 6870-14). I have also completed the UBC Laboratory Biological

Safety Course (certificate number: 2019-swyay).

ix

Table of Contents

Abstract ......................................................................................................................... iii

Lay Summary ................................................................................................................. v

Preface .......................................................................................................................... vi

Table of Contents ......................................................................................................... ix

List of Tables .............................................................................................................. xvi

List of Figures ........................................................................................................... xvii

List of Symbols .......................................................................................................... xxi

List of Abbreviations ................................................................................................. xxii

Acknowledgements ................................................................................................. xxvii

Dedication ................................................................................................................. xxix

Chapter 1: An introduction to frontotemporal lobar degeneration, progranulin

neurobiology, and premature termination codon readthrough ................................ 1

1.1 Frontotemporal lobar degeneration ................................................................... 1

1.1.1 History of FTLD ............................................................................................. 1

1.1.2 Clinical presentations of FTD ........................................................................ 2

1.1.3 Neuropathology of FTLD ............................................................................... 3

1.1.4 Genetics of FTLD .......................................................................................... 4

1.1.5 FTLD caused by PGRN haploinsufficiency .................................................... 6

1.1.6 Genetic modifiers of FTLD-GRN ................................................................... 8

1.1.7 Clinical and neuropathological phenotypes of FTLD-GRN and CLN11 ....... 12

1.2 PGRN neurobiology ........................................................................................ 14

x

1.2.1 PGRN structure ........................................................................................... 15

1.2.2 PGRN expression in the brain ..................................................................... 15

1.2.3 PGRN regulation in the brain ....................................................................... 18

1.2.4 Mouse models of Grn-deficiency ................................................................ 24

1.2.4.1 Neuroinflammation ............................................................................... 24

1.2.4.2 Lysosomal dysfunction ......................................................................... 28

1.2.4.3 Dysregulated lipid metabolism ............................................................. 31

1.2.4.4 Neural connectivity deficits and neurodegeneration ............................. 33

1.2.5 hiPSC-derived CNS models of GRN-deficiency .......................................... 37

1.2.6 Proposed mechanisms of neurodegeneration in FTLD-GRN ...................... 41

1.2.7 Therapeutic development for the treatment of FTLD-GRN .......................... 47

1.3 Premature termination codon readthrough ..................................................... 48

1.3.1 Aminoglycoside PTC readthrough ............................................................... 52

1.3.2 Aminoglycoside toxicity ............................................................................... 54

1.3.3 Promising PTC readthrough drugs .............................................................. 56

1.3.4 PTC readthrough enhancers ....................................................................... 59

1.3.5 Nonsense mediated mRNA decay .............................................................. 60

1.3.6 PTC readthrough in cellular models of FTLD-GRN ..................................... 62

1.4 Thesis research questions .............................................................................. 63

Chapter 2: Materials and methods ............................................................................. 67

2.1 Mice ................................................................................................................ 67

2.2 Cells ................................................................................................................ 67

2.2.1 HEK293 ....................................................................................................... 67

xi

2.2.2 Mouse embryonic fibroblasts ....................................................................... 68

2.3 Expression vectors .......................................................................................... 68

2.3.1 Vector mutagenesis ..................................................................................... 68

2.3.2 Transfection ................................................................................................. 69

2.3.3 AAV vector packaging ................................................................................. 69

2.3.4 In vitro transduction ..................................................................................... 70

2.4 Antibodies ....................................................................................................... 71

2.5 Drug treatments .............................................................................................. 73

2.5.1 In vitro.......................................................................................................... 73

2.5.2 AAV mouse model ....................................................................................... 73

2.5.3 GrnR493X/R493X mouse model ......................................................................... 74

2.6 Western blot .................................................................................................... 74

2.6.1 Conventional western blot ........................................................................... 74

2.6.2 ProteinSimple Wes ...................................................................................... 75

2.7 hiPSC models ................................................................................................. 76

2.7.1 hiPSC reprogramming ................................................................................. 76

2.7.2 hiPSC culture .............................................................................................. 76

2.7.3 hiPSC trilineage differentiation .................................................................... 77

2.7.4 hiPSC CNS lineage differentiation ............................................................... 77

2.7.5 Karyotyping analysis ................................................................................... 78

2.8 CRISPR/Cas9 gene editing............................................................................. 78

2.9 Multielectrode array electrophysiology ............................................................ 79

2.10 ELISA .............................................................................................................. 80

xii

2.11 qPCR .............................................................................................................. 81

2.12 Mouse genotyping ........................................................................................... 81

2.13 Histology ......................................................................................................... 82

2.13.1 Immunocytochemical staining and quantification ........................................ 82

2.13.2 Brain immunofluorescence microscopy and quantification .......................... 83

2.13.3 Brightfield microscopy ................................................................................. 87

2.14 Surgical procedures ........................................................................................ 87

2.14.1 Intracerebroventricular injection of AAV particles ........................................ 87

2.14.2 Bolus intracerebroventricular injection of drugs ........................................... 87

2.14.2.1 AAV-GRN-R493X-V5 mice ................................................................... 87

2.14.2.2 Grn+/+ and GrnR493X/R493X mice .............................................................. 88

2.14.3 Intracerebroventricular iPRECIO® pump implantation in Grn+/+ and

GrnR493X/R493X mice ................................................................................................. 89

2.15 Open-field behavioural assay .......................................................................... 90

2.16 Brain collection of processing ......................................................................... 90

2.16.1 AAV mouse model ....................................................................................... 90

2.16.2 GrnR493X/R493X baseline characterization ...................................................... 91

2.16.3 Vehicle/G418 treated Grn+/+ and GrnR493X/R493X mice ................................... 92

2.17 Statistical analysis ........................................................................................... 92

Chapter 3: Exogenous PTC readthrough of nonsense mutant GRN expression

constructs .................................................................................................................... 94

3.1 Introduction ..................................................................................................... 94

3.2 Results ............................................................................................................ 96

xiii

3.2.1 PTC readthrough in transiently transfected HEK293 cells ........................... 96

3.2.2 PTC readthrough in stably transfected HEK293 cells ................................ 100

3.2.3 PTC readthrough in HEK293 cells transduced with AAV-GRN-R493X-V5

viral particles ........................................................................................................ 102

3.2.4 Technical demonstration of CNS P0 delivery of AAV particles and bolus ICV

injection in adult mice ........................................................................................... 104

3.2.5 PTC readthrough in AAV-GRN-R493X-V5 mice ........................................ 107

3.3 Discussion..................................................................................................... 109

Chapter 4: Generation of FTLD-GRN hiPSC lines and endogenous PTC

readthrough of GRN nonsense mutant hiPSC-derived CNS cell types ................ 111

4.1 Introduction ................................................................................................... 111

4.2 Results .......................................................................................................... 114

4.2.1 Generation of WT and FTLD-GRN patient-derived hiPSC lines ................ 114

4.2.2 Generation of an isogenic GRN R493X homozygous knock-in hiPSC line

using CRISPR/Cas9 gene editing ........................................................................ 114

4.2.3 Characterization of hiPSC line pluripotency .............................................. 115

4.2.4 Differentiation of WT and GRN-deficient hiPSC lines into cortical neurons

and astrocytes ...................................................................................................... 121

4.2.5 PTC readthrough in GRN-deficient hiPSC-derived cortical neurons ......... 128

4.2.6 PTC readthrough rescues lysosomal dysfunction in R493X-/- KI hiPSC-

derived cortical neurons ....................................................................................... 137

4.2.7 PTC readthrough in R493X-/- KI hiPSC-derived astrocytes ........................ 141

xiv

4.2.8 Impact of G418 on PGRN expression in WT hiPSC-derived cortical neurons

and astrocytes ...................................................................................................... 149

4.3 Discussion..................................................................................................... 153

Chapter 5: Phenotypic characterization of the GrnR493X/R493X mouse model and

dose-limiting in vivo toxicity of G418 ...................................................................... 159

5.1 Introduction ................................................................................................... 159

5.2 Results .......................................................................................................... 160

5.2.1 Pgrn expression in the brains of aged GrnR493X/R493X mice ......................... 160

5.2.2 Lysosomal dysfunction in the brains of aged GrnR493X/R493X mice .............. 165

5.2.3 Neuroinflammation and astrogliosis in the ventral thalamus of aged

GrnR493X/R493X mice ................................................................................................ 177

5.2.4 Partial preservation of inhibitory synaptic density in the thalamus of aged

GrnR493X/R493X mice ............................................................................................... 185

5.2.5 Reduced thalamic excitatory neuron density in the brains of aged

GrnR493X/R493X mice ............................................................................................... 188

5.2.6 Aged male GrnR493X/R493X mice exhibit an increased anxiety phenotype .... 190

5.2.7 PTC readthrough in GrnR493X/R493X MEFs ................................................... 198

5.2.8 G418-induced toxicity in GrnR493X/R493X mice limits therapeutic potential ... 201

5.3 Discussion..................................................................................................... 216

Chapter 6: Conclusions ............................................................................................ 223

6.1 Contributions to the fields of FTLD-GRN and CLN11 ................................... 223

6.1.1 In vitro and in vivo PTC readthrough of exogenously expressed nonsense

mutant human GRN expression constructs .......................................................... 223

xv

6.1.2 In vitro and in vivo PTC readthrough in human and mouse models

endogenously expressing nonsense mutant GRN ............................................... 224

6.2 Final conclusions .......................................................................................... 226

6.3 Future directions ........................................................................................... 227

References ................................................................................................................. 231

xvi

List of Tables

Table 1-1 Primary antibodies ........................................................................................ 71

xvii

List of Figures

Figure 1-1 Brain regulation of extracellular PGRN proteolysis ...................................... 22

Figure 1-2 Nonsense mutation readthrough .................................................................. 51

Figure 1-3 Graphical overview of experimental models................................................. 66

Figure 3-1 Induction of PTC readthrough by G418 and CDX5 enhancers in cells stably

expressing nonsense mutant GRN-V5 .......................................................................... 97

Figure 3-2 Induction of PTC readthrough by G418 in transiently transfected cells

expressing nonsense mutant GRN-V5 .......................................................................... 99

Figure 3-3 Chemical structures of CDX5-196 and CDX5-288 ..................................... 101

Figure 3-4 Induction of PTC readthrough in cells transduced with increasing titers of

AAV-GRN-R493X-V5 .................................................................................................. 103

Figure 3-5 Technical demonstration of AAV P0 ICV injection method ........................ 105

Figure 3-6 Technical demonstration of bolus ICV injection method in adult mice ....... 106

Figure 3-7 Induction of PTC readthrough in AAV-R493X-GRN-V5 mice ..................... 108

Figure 4-1 Generation of isogenic CRISPR/Cas9 gene edited R493X-/- KI hiPSC line

from WT ...................................................................................................................... 116

Figure 4-2 PGRN immunofluorescence in WT and R493X-/- KI hiPSC-derived cortical

neurons ....................................................................................................................... 118

Figure 4-3 Characterization of WT and GRN-deficient hiPSC lines ............................ 119

Figure 4-4 Differentiation and characterization of cortical neurons and astrocytes

derived from WT and GRN-deficient hiPSCs .............................................................. 122

Figure 4-5 Characterization of hiPSC-derived cortical neuron synaptic development and

neural network formation ............................................................................................. 124

xviii

Figure 4-6 Co-culturing hiPSC-derived cortical neurons with hiPSC-derived astrocytes

accelerates electrophysiological maturation and neural network formation................. 126

Figure 4-7 Baseline expression and secretion of PGRN in WT hiPSC-derived cortical

neurons and astrocytes ............................................................................................... 130

Figure 4-8 Detecting PGRN in WT and R493X-/- KI hiPSC-derived cortical neurons by

western blot ................................................................................................................. 131

Figure 4-9 Induction of PTC readthrough by G418 and enhancers in hiPSC-derived

cortical neurons bearing FTLD-GRN nonsense mutations .......................................... 132

Figure 4-10 Quantification of G418-induced neuronal cytotoxicity .............................. 134

Figure 4-11 Selective increase of nonsense mutant GRN mRNA in R493X-/- KI hiPSC-

derived cortical neurons in response to PTC readthrough treatments ........................ 136

Figure 4-12 GRN PTC readthrough rescues FTLD-GRN/CLN11 lysosomal pathological

CTSD maturation phenotype in R493X-/- KI hiPSC-derived cortical neurons .............. 139

Figure 4-13 qPCR analysis of GRN mRNA levels in WT hiPSC-derived cortical neurons

and astrocytes ............................................................................................................. 143

Figure 4-14 Induction of PTC readthrough by G418 and enhancers in R493X-/- KI

hiPSC-derived astrocytes demonstrated by western blot ............................................ 144

Figure 4-15 Induction of PTC readthrough by G418 and enhancers in R493X-/- KI

hiPSC-derived astrocytes demonstrated by ELISA ..................................................... 147

Figure 4-16 Selective increase of nonsense mutant GRN mRNA in R493X-/- KI hiPSC-

derived astrocytes in response to PTC readthrough treatments ................................. 148

Figure 4-17 G418-mediated disruption of PGRN homeostasis in WT hiPSC-derived

astrocytes .................................................................................................................... 151

xix

Figure 5-1 GrnR493X/R493X genotyping............................................................................ 162

Figure 5-2 Pgrn expression in the brains of aged GrnR493X/R493X mice ......................... 163

Figure 5-3 Lysosomal dysfunction in the ventral thalamus and CA3 hippocampal region

of aged GrnR493X/R493X mice .......................................................................................... 167

Figure 5-4 Enlarged lysosomes in the ventral thalamus and CA3 hippocampal region of

aged GrnR493X/R493X mice .............................................................................................. 169

Figure 5-5 Global lysosomal dysfunction in the brains of aged GrnR493X/R493X mice ..... 171

Figure 5-6 Neuronal TDP-43 proteinopathy is localized to the ventral thalamus of aged

GrnR493X/R493X mice ....................................................................................................... 174

Figure 5-7 Absence of global p-TDP-43 proteinopathy in the brains of aged

GrnR493X/R493X mice ....................................................................................................... 176

Figure 5-8 Neuroinflammation in the ventral thalamus of aged GrnR493X/R493X mice ..... 180

Figure 5-9 Microglial skeletal analysis method ............................................................ 182

Figure 5-10 Severe astrogliosis in the ventral thalamus and CA3 hippocampal region of

aged GrnR493X/R493X mice .............................................................................................. 183

Figure 5-11 Inhibitory synaptic density is preserved in the thalamus of aged

GrnR493X/R493X mice ....................................................................................................... 186

Figure 5-12 Reduced thalamic Foxp2+ excitatory neuron density in the brains of aged

GrnR493X/R493X mice ....................................................................................................... 189

Figure 5-13 Aged male GrnR493X/R493X mice exhibit an increased anxiety phenotype ... 192

Figure 5-14 GrnR493X/R493X brain lysosomal dysfunction does not exhibit a sex-specific

phenotype ................................................................................................................... 194

xx

Figure 5-15 GrnR493X/R493X microgliosis and astrogliosis does not exhibit a sex-specific

phenotype ................................................................................................................... 195

Figure 5-16 GrnR493X/R493X decreased thalamic excitatory neuron density and inhibitory

synaptic pruning does not exhibit a sex-specific phenotype ........................................ 196

Figure 5-17 Induction of PTC readthrough by G418 and enhancers in GrnR493X/R493X

MEFs ........................................................................................................................... 199

Figure 5-18 A single ICV injection of G418 may induce increased Pgrn expression in the

brains of GrnR493X/R493X mice ........................................................................................ 203

Figure 5-19 Technical demonstration of bolus iPRECIO® pump-mediated ICV injections

targeting the lateral ventricles ..................................................................................... 205

Figure 5-20 G418’s PTC readthrough activity is preserved despite prolonged incubation

at physiologic temperature .......................................................................................... 206

Figure 5-21 Schematic describing iPRECIO® pump experiment infusion programs and

toxicity outcomes in mice ............................................................................................ 208

Figure 5-22 Quantification of G418 concentration in the brains of Grn+/+ and

GrnR493X/R493X mice treated ICV with G418 .................................................................. 209

Figure 5-23 Multiple iPRECIO® pump-mediated ICV infusions of 50 µg G418 may

induce increased Pgrn expression in the brains of GrnR493X/R493X mice ....................... 210

Figure 5-24 Grn+/+ tolerate iPRECIO® pump ICV infusion of saline ............................. 213

Figure 5-25 Increased Pgrn expression induced by 2X repeated ICV G418 injections in

the brains of GrnR493X/R493X mice .................................................................................. 214

xxi

List of Symbols

µ micron

α alpha

β beta

δ delta

γ gamma

κ kappa

xxii

List of Abbreviations

2-DOS 2-deoxystreptamine AAV Adeno-associated virus AAVS1 Adeno-associated virus integration site 1 AD Alzheimer’s disease ADAMTS7 A disintegrin and metalloproteinase with thrombospondin motifs 7 ADP Adenosine diphosphate ALP Autophagy-lysosomal pathway ALS Amyotrophic lateral sclerosis ANOVA Analysis of variance ATP Adenosine triphosphate ATRA Vitamin A1 metabolite all-trans retinoic acid BBB Blood-brain-barrier BDNF Brain-derived neurotrophic factor bvFTD Behavioural variant frontotemporal dementia C/EBPβ CCAAT-enhancer binding protein β C1qa Complement C1q subcomponent subunit A C3b Complement component C3b C9orf72 Chromosome 9 open reading frame 72 Cas9 CRISPR associated protein 9 CE Cholesterol ester CF Cystic fibrosis CLEAR Coordinated lysosomal expression and regulation CLN11 Ceroid lipofuscinosis neuronal 11 CNS Central nervous system CreER Cre-estrogen receptor CRISPR Clustered regularly interspaced short palindromic repeats crRNA trans-activating CRISPR RNA CSF Cerebrospinal fluid CSF1R Colony stimulating factor 1 receptor CTF C-terminal fragment CTIP2 COUP-TF-interacting protein 2 CTSB/D/L/S Cathepsin B/D/L/S CX3CR1 CX3 chemokine receptor 1 DAG Diacylglycerides DAM Disease-associated microglia DAPI 4′,6-diamidino-2-phenylindole DIV Day in vitro DLDH Dementia lacking distinctive histopathology DMD Duchenne muscular dystrophy DMEM Dulbecco's Modified Eagle Medium DMSO Dimethyl sulfoxide DN Dystrophic neurites

xxiii

DNMT DNA methyltransferase DPPII/7 Dipeptidyl peptidase II/7 DTT Dithiothreitol EDTA Ethylenediaminetetraacetic acid eGFP Enhanced green fluorescent protein EJC Exon junction complex ELISA Enzyme-linked immunosorbent assay EP Erythroid progenitor ER Endoplasmic reticulum eRF1/3 Eukaryotic release factor 1/3 EWS Ewing sarcoma FAM171A2 Family with sequence similarity 171 member A2 FBS Fetal bovine serum FOXG/O/P1 Forkhead box protein G/O/P1 FTD Frontotemporal dementia FTD-MND Frontotemporal dementia with motor neuron disease FTLD Frontotemporal lobar degeneration FUS Fused in Sarcoma GAD65 Glutamic acid decarboxylase 65-kilodalton isoform GAPDH Glyceraldehyde 3-phosphate dehydrogenase GCase Glucocerebrosidase GDNF Glia cell-derived neurotrophic factor GFAP Glial fibrillary acidic protein GNS N-acetylglucosamine-6-sulfatase GOF Gain-of-function GPNMB Glycoprotein NMB gRNA Guide RNA GRNs Granulin peptides GWAS Genome-wide association studies HEK293 Human embryonic kidney 293 cells HEXB Hexosaminidase subunit beta hiPSC Human induced pluripotent stem cells HPRT1 Hypoxanthine phosphoribosyltransferase 1 Hsp70 Heat shock protein 70 IBA1 Allograft inflammatory factor 1 ICV Intracerebroventricular IF Immunofluorescence IFNγ Interferon gamma IKBKB Inhibitor of NF-κB subunit beta IL-10 Interleukin 10 IL-1β Interleukin 1 beta kDa Kilodaltons KO Knockout LAMP1 Lysosomal-associated membrane LC Locus ceruleus

xxiv

LC3-I/II Microtubule-associated proteins I/II light chain 3B LD Lipid-droplet LOF Loss-of-function LPS Lipopolysaccharide LRP1 Low density lipoprotein receptor-related protein 1 M6PR Mannose 6-phosphate receptor MAC Membrane attack complex MAP2 Microtubule-associated protein 2 MAPK Mitogen-activated protein kinase MAPT Microtubule-associated protein tau MCM Microglia conditioned media MEA Multielectrode array MEF Mouse embryonic fibroblast miRNA Micro-RNA Mitoribosome Mitochondrial ribosome mPFC Medial prefrontal cortex MSN Medium spiny neurons mTOR C1 Mammalian target of rapamycin complex 1 N2a Neuro2a NA Nucleus accumbens NCI Neuronal cytoplasmic inclusions NCL Neuronal ceroid lipofuscinosis NF-κB Nuclear factor kappa beta nfvPPA Non-fluent variant primary progressive aphasia NII Neuronal intranuclear inclusions NMD Nonsense-mediated mRNA decay NPC Neuronal progenitor cells OCD Obsessive-compulsive disorder ORF Open reading frame PBS Phosphate-buffered saline PCR Polymerase chain reaction PFA Paraformaldehyde PFU Plaque-forming unit PGRN Progranulin PGRN/GRNs Full-length PGRN, truncated PGRN, and GRN peptides PI3K Phosphoinositide 3-kinase PKB Protein kinase B PKCδ Protein kinase C delta PLO/L Poly-L-ornithine / laminin Poly I:C Polyinosinic:polycytidylic acid PP2A Protein phosphatase 2 PSAP Prosaposin PTC Premature termination codon p-TDP-43 Phosphorylated TAR DNA-binding protein 43

xxv

qPCR Quantitative polymerase chain reaction rec. PGRN Recombinant PGRN RF Release factor RIPA Radioimmunoprecipitation assay RNAi RNA interference ROS Reactive oxygen species RT Room temperature S1BF Barrel field region of the rodent primary somatosensory cortex S6K2 S6 kinase beta SAP Saposin SCARB2 Scavenger receptor class B member 2 SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis SEM Standard error of mean siRNA Small interfering RNA SLPI Secretory leukocyte protease inhibitor SMAD Small mothers against decapentaplegic SNP Single nucleotide polymorphism snRNA-seq Single-nuclei RNA-sequencing SORT1 Sortilin SOX1/17 SRY-related HMG-box 1/17 ssODN Single-stranded oligonucleotide svPPA Semantic-variant primary progressive aphasia TAF15 TBP associated factor 15 TAG Triacylglycerides TARDBP TAR DNA-binding protein 43 gene TBR1 T-box brain transcription factor 1 TBS Tris-buffered saline TCA Tricarboxylic acid TDP-43 TAR DNA-binding protein 43 TFEB Transcription factor EB TGF-β Transforming growth factor beta TGFβR-1 Transforming growth factor beta receptor 1 TLR4 Toll-like receptor 4 TMEM106B Transmembrane protein 106B TNFR1/2 Tumor necrosis factor receptor 1/2 TNFα Tumor necrosis factor alpha TP53/p53 Tumor protein p53 TPP1 Tripeptidyl-peptidase 1 TUJ1 Neuron specific class III beta-tubulin UBC University of British Columbia UPF1/2/3 Up-frameshift suppressor 1/2/3 UTR Untranslated region V-ATPase Vacuolar-type ATPase VGAT Vesicular gamma-aminobutyric amino acid transporter VGLUT1 Vesicular glutamate transporter 1

xxvi

VPL Ventral posterolateral VPM Ventral posteromedial VT Vehicle-treated WB Western blot WHP Woodchuck hepatitis virus Wnt Wingless-related integration site WPRE WHP post-transcriptional regulatory element WT Healthy control hiPSC line

xxvii

Acknowledgements

I conducted this research and wrote this dissertation on the unceded traditional

ancestral territories of the Coast Salish peoples, including the Qayqayt (the only First

Nation without a land base), the xʷməθkʷəy̓əm (Musqueam), the skwxwú7mesh

(Squamish) and the səl̓ilwətaɁɬ (Tsleil-Waututh). As a Canadian decedent of white

colonial settlers, I acknowledge the historical and ongoing injustices against indigenous

people and am committed to the lifelong practices of decolonization and reconciliation.

I offer my tremendous gratitude to my supervisor Dr. Haakon Nygaard for his incredible

support, guidance, and generosity throughout my graduate studies. I am grateful for the

trust he had in me to establish his new lab's scientific foundation. His encouragement,

thoughtful feedback, and belief in my abilities helped me develop the resilience required

to face major challenges I encountered throughout my graduate studies. I thank my

committee member Dr. Michel Roberge, whose frequent collaborative and insightful

discussions greatly improved the work presented in this thesis. I would also like to thank

my current and former committee members, Dr. Stephanie Willerth, Dr. Robin Hsiung,

and Dr. Christian Naus, for their invaluable guidance.

I also thank the Nygaard lab members, past and present, who created a lab

environment that nurtured my scientific curiosity and contributed to my ultimate success

in accomplishing my research goals. In no particular order, thank you: Tyler Yan, Dr.

Brianne Kent, Dr. Xiajuan Wu, Julia Boyle, Katina Mak, Hannah Pae, Chris Lee, Ariel

Frame, Clara Dutton-Kneaves, Meghan Chen, Kiana Yau, and Adrienne Kinman. Many

xxviii

thanks also to past and present Roberge lab members, especially to Dr. Aruna Balgi,

Dr. Alireza Baradaran-Heravi, and Dr. Kunho Choi.

Thank you to the Weston Brain Institute, who provided funds for the research presented

in this thesis. I would also like to thank the Canadian Institute for Health Research for

providing salary support through the Canada Graduate Scholarships – Master’s

program.

Special thanks are owed to my family, especially to my parents Mark and Debbie Frew,

who have always believed in me. I was so lucky to have you both so close by for

emotional support and encouragement throughout all the successes and failures that

defined this journey. Last but certainly not least, I would like to also thank my wife,

Madison Galts, for her unconditional love and support, cheering me on every step of the

way.

xxix

Dedication

I dedicate this thesis to my wife Madison for the endless sacrifices she happily made to

support me throughout this challenging endeavor, and to my late grandpa Roy whose

battle with Alzheimer’s disease inspired me to pursue graduate studies in the field of

neurodegeneration.

1

Chapter 1: An introduction to frontotemporal lobar degeneration,

progranulin neurobiology, and premature termination codon

readthrough

1.1 Frontotemporal lobar degeneration

Frontotemporal lobar degeneration (FTLD) is the underlying pathology uniting a

clinically, genetically, and pathologically heterogeneous group of disorders that primarily

affect the frontal, temporal, and parietal lobes. FTLD causes a spectrum of clinical

presentations of frontotemporal dementia (FTD), including progressive changes in

behaviour, personality, executive function, and language. 75-80% of FTD patient

symptomatic onset occurs before age 65, and FTLD has an incidence of 10-30 per

100,000 individuals aged 45-65 years old (1). FTLD is the 2nd most common cause of

early-onset dementia (2). There are currently no approved therapies that can stop or

alter the progression of this disease. Symptomatic management with selective serotonin

reuptake inhibitors and other antipsychotics represents the only available form of

pharmacological intervention for FTD patients (3).

1.1.1 History of FTLD

Pathology resembling FTLD has been clinically described for over a century; however, it

was often misdiagnosed as Alzheimer’s disease (AD) or other psychiatric disorders (4).

Arnold Pick was first to characterize a now rare form of FTLD, Pick’s disease,

diagnosed post-mortem pathologically by the observation of neuronal ballooning (Pick

2

cells) and intraneuronal argentophilic inclusion bodies (Pick bodies) in the absence of

AD plaques/tangles (5-7). In the early 1980s, American neurologists believed it was

essentially impossible to distinguish between Pick’s disease and AD using either clinical

or radiological measures (8). Meanwhile, neurologists in Europe had observed a

behavioural phenotype not classically associated with AD and developed a clinical

rating scale to differentiate between AD and Pick’s disease / ‘frontal degeneration of

non-Alzheimer’s type’ (9,10). It was later discovered that many FTLD cases in America

had been misdiagnosed as AD, likely due to a lack of awareness of its distinct clinical

presentation (11). In 1994, European clinicians published a consensus report on the

clinical and neuropathological criteria for FTLD; thus, solidifying its recognition as a

separate condition from AD dementia (2). More recently, advances in DNA sequencing,

neuropathological indicators, and radiographic technologies have further revealed the

complex clinical heterogeneity underlying FTLD.

1.1.2 Clinical presentations of FTD

There are three main clinical subtypes of FTD: behavioural variant (bvFTD), semantic

variant primary progressive aphasia (svPPA), and non-fluent variant progressive non-

fluent aphasia (nfvPPA). Few cases present in a relatively pure form, with most patients

developing overlapping syndromes throughout disease progression. bvFTD is

characterized by atrophy of the frontal, anterior temporal, and lateral parietal lobes

resulting in various degrees of social disinhibition, loss of sympathy/empathy, and

increased apathy (12). svPPA is characterized by a consistent asymmetric atrophy

pattern in the anterior temporal lobe leading to an inability to access conceptual

3

knowledge (12). nfvPPA is characterized primarily by atrophy of the left hemisphere,

disrupting smooth speech articulation, and grammatical proficiency (12).

Fluorodeoxyglucose positron emission tomography imaging measuring glucose uptake

in FTLD patients’ brains has identified deficits in cerebral glucose metabolism as an

early marker of disease progression (13). Further, 15% of patients with FTLD express

comorbidity in the form of motor neuron disease (14,15), resembling amyotrophic lateral

sclerosis (ALS), highlighting the significant overlap of genetic, clinical, and pathological

characteristics shared by these two disorders (16). In general, the prognosis for FTLD

patients is poor, leading to death within 3-10 years following diagnosis (17).

1.1.3 Neuropathology of FTLD

Post mortem analysis of clinically diagnosed FTLD patients’ brains has led to the

identification of four major types of abnormal aggregated protein accumulation in

neurons and glia. Approximately 45% of clinically diagnosed FTLD cases develop

neuronal intracytoplasmic inclusions (NCI) that are composed of the microtubule-

associated protein tau (FTLD-tau) (18). FTLD-tau can be further divided into two

subtypes, with the hallmark either being round tau-positive Pick bodies observed in

Pick’s disease or neurofibrillary tangle-like structures observed in corticobasal

syndrome and progressive supranuclear palsy (19). Tau-negative FTLD cases were

initially classified as dementia lacking distinctive histopathology (DLDH) (20), though it

was later discovered that the majority of these DLDH cases possessed ubiquitin-

positive inclusions (21,22).

4

A few years later, the elusive ubiquitinated protein that constitutes the aggregates in

about 50% of FTLD cases was discovered to be the RNA/DNA binding protein TAR

DNA-binding protein 43 (TDP-43) (23,24). TDP-43 aggregates can appear as neuronal

cytoplasmic inclusions (NCI), neuronal intranuclear inclusions (NII), and dystrophic

neurites (DN) in the cortices of FTLD patients (FTLD-TDP). The ratio and prevalence of

these different TDP-43 aggregates distinguish FTLD-TDP type A-D sub-classifications

(25). Subtype A cases have a near-equal ratio of NCI and DN; in type B, NCI

predominates over DN; type C is applied when DN predominates over NCI, and type D

when NII is the most common form of TDP-43 pathology. The discovery that

chromosome 9 open reading frame 72 (C9orf72) hexanucleotide expansion mutations

cause the majority of familial FTLD cases has led to the identification of distinct

histopathological hallmarks. In addition to TDP-43 pathology, these patients develop

sense and antisense C9orf72 expanded repeat RNA foci (26), and translation of the

hexanucleotide expansion derived mRNA results in the cytoplasmic accumulation of

dipeptide repeat proteins (27). The remaining 5% of FTLD patient brains accumulate

protein inclusions (NCI and sometimes NII) consisting of Fused in Sarcoma (FUS)

protein and sometimes other members of the FET (FUS, EWS, and TAF15) family of

proteins (FTLD-FUS) (28).

1.1.4 Genetics of FTLD

An estimated 40% of FTLD cases are familial, with patients often presenting with a

history of similar diseases within their family (29). Furthermore, a clear autosomal

dominant inheritance pattern has been observed in ~10% of cases, thus indicating

5

genetics play a key role in disease pathogenesis (29). Certain FTD clinical phenotypes

exhibit differential inheritance patterns; for example, svPPA is rarely familial, while

bvFTD can be either familial or sporadic in nature (29). The three most prevalent genes

confirmed to cause familial FTLD are MAPT, GRN, and C9orf72. There are 44 known

pathogenic MAPT mutations, causing between 5-20% of familial FTLD cases (30,31).

There are two main types of MAPT mutations: 1) mutations that disrupt splicing of exon

10 that result in an increased ratio of 4R:3R tau isoforms (32) that lead to accumulation

of 4R insoluble tau aggregates forming neurofibrillary tangles; 2) missense mutations

throughout the gene that reduce the ability of all tau isoforms to bind to microtubules,

increasing the pool of unbound tau accelerating 3R/4R tau aggregation into Pick bodies

(33). After identifying MAPT mutations on chromosome 17q21, there were still

numerous FTLD cases that had been linked to 17q21 but were null for mutations in

MAPT (34). Subsequently, loss-of-function (LOF) mutations in GRN were discovered in

these patients within a 6.2 Mb segment of 17q21 that also contained the MAPT locus

(35,36). Approximately 5-20% of familial FTLD cases are caused by GRN

haploinsufficiency (FTLD-GRN) (37). Most heterozygous pathogenic GRN mutations

lead to a > 50% loss in progranulin protein (PGRN) levels/function leading to disease

through haploinsufficiency (38).

In the years following the identification of causal FTLD-GRN mutations, there was still a

significant proportion of FTLD patients that exhibited clear autosomal dominant pattern

of inheritance with unknown genetic origin. In 2011, researchers identified the

GGGGCC hexanucleotide repeat expansion in the non-coding region of the C9orf72

6

gene as a key driver of familial FTLD and ALS (39,40). The majority (~25%) of familial

FTLD cases are caused by C9orf72 hexanucleotide expansions (41), with patients

carrying 100-1000 repeats compared to 2-20 repeats in healthy control populations (38).

It has been demonstrated that transcripts bearing the GGGGCC repeat sequences can

be translated through an unconventional repeat-associated non-ATG initiated form of

translation, which results in the accumulation of dipeptide repeats (27). The mechanism

driving neurodegeneration in C9orf72 repeat expansion carriers is still up for debate,

with findings from numerous disease models supporting both LOF and gain-of-function

(GOF) pathogenic processes (42,43). Interestingly, mutations in TARDBP (TDP43) are

rarely associated with FTLD-TDP pathology (19). Meanwhile, several rare mutations in

genes associated with autophagolysosomal pathways (CHMP2B, VCP, SQSTM1,

TBK1) have been shown to cause a small minority of FTLD-TDP cases (44-47). This

suggests that disrupted proteolytic mechanisms play a major role in the establishment

of FTLD-TDP pathology. Lastly, most FTLD-FUS cases appear to lack clear genetic

origin and are believed to be sporadic in nature (19).

1.1.5 FTLD caused by PGRN haploinsufficiency

As discussed in section 1.1.4, heterozygous mutations in GRN on chromosome 17q21

cause a familial form of FTLD. Going forward, we will refer to symptomatic GRN

haploinsufficient mutation carriers as having FTLD-GRN and will refer to models of GRN

deficiency as FTLD-GRN models. All pathogenic GRN mutations have either been

confirmed or are predicted to produce a LOF allele (48). Of the known 79 known

pathogenic GRN mutations, ~84% are either nonsense or frameshift mutations that

7

introduce premature termination codons (PTCs) (49-51). However, less common

missense and deletion mutations have also been identified that can: disrupt mRNA

maturation leading to nuclear mRNA degradation (52); prevent initiation of protein

translation at the start codon (53); destroy the signal peptide sequence critical for

protein secretion (54); and, eliminate key cysteine residues that do not decrease total

protein/mRNA levels but cause impaired protein function and processing into granulin

peptides (GRNs) (55).

FTLD-GRN has a variable age of onset, ranging from 35-89 years of age. This presents

a significant challenge for designing clinical trials since even family members bearing

the same GRN mutation can develop symptoms at significantly different ages (56).

Extrapolating the incidence of FTLD (see section 1.1) using the proportion of FTLD

patients bearing GRN mutations (~4%) implies a global incidence of 0.5-1.5 FTLD-GRN

cases per 100,000 people (36,57). This finding, along with the lack of clinical correlation

between age of onset and PGRN plasma/cerebrospinal fluid (CSF) concentration (58),

implies that other factors such as disease modulating genetic variants or environmental

conditions may play a role in determining age of onset. The vast majority of FTLD-GRN

patients express 70-80% less serum PGRN than control subjects (59). These patient

PGRN levels are lower than expected, given the heterozygous LOF nature of FTLD-

GRN mutations. Reprogramming FTLD-GRN patient fibroblasts that express < 50%

GRN mRNA levels compared to healthy control cells into human induced pluripotent

stem cells (hiPSCs) increased GRN mRNA back to the expected 50% level (60).

Pluripotent reprogramming resets a differentiated cell’s epigenetic signature; therefore,

8

suggesting that epigenetic mechanisms may be responsible for the additional basal

suppression of PGRN expression in FTLD-GRN mutation carriers.

1.1.6 Genetic modifiers of FTLD-GRN

A host of genetic elements have been associated with the age of onset and disease

progression in FTLD-GRN. The majority of these factors are either known to influence

lysosomal function, are involved in the regulation of GRN expression, or participate in

direct protein-protein interactions with PGRN (61). Most of these genetic markers have

been identified through genome-wide association studies (GWAS); thus, it is important

to recognize the limitations of interpreting these disease-associated risk loci. The

original variant or single nucleotide polymorphism (SNP) identified by a GWAS is rarely

confirmed to be the true causal variant resulting in the molecular modification that

confers greater disease risk (62); this is because there are usually 10s-100s of variants

in strong linkage disequilibrium with the sentinel SNP at the specified genetic locus,

representing a set of co-inherited variants or haplotypes. There is the potential for any

of these co-inherited variants to be the true underlying cause of altered disease risk. It is

also important to consider that the risk-haplotype can often span multiple genes, further

complicating the identification of the gene to which a certain GWAS signal belongs.

TMEM106B is a single-pass integral membrane glycoprotein localized to late

endosomes and lysosomes and is important for regulating lysosomal size, motility, and

responsiveness to stress (63). TMEM106B and PGRN have both been shown to co-

localize in the endolysosomal compartment in various cell types (64).

9

SNPs in TMEM106B have been shown to have protective effects on FTLD-GRN

disease progression (65,66). The minor allele of the non-coding TMEM106B SNP

rs1990662 has been detected at a low frequency in FTLD-GRN patients compared to

healthy controls, suggesting GRN mutation carriers expressing the minor allele may

exhibit delayed symptomatic onset (65,66). Further, the frontal cortices of FTLD-TDP

patients and unaffected individuals carrying the major allele exhibit significantly elevated

TMEM106B mRNA levels compared to those with the minor allele (65). Remarkably, the

age of onset was delayed by an average of 13 years in FTLD-GRN patients

homozygous for the rs1990662 “protective” TMEM106B allele (67). These patients were

also found to possess significantly higher plasma PGRN levels than those homozygous

for the “risk” allele, possibly explaining their slower disease progression (67).

More recently, TMEM106B SNP rs1990620, an additional non-coding “risk” allele, has

been identified in complete linkage disequilibrium with the sentinel SNP (rs1990622)

that functionally increases recruitment of the chromatin-organizing protein CTCF to

downstream regulatory sequences of the TMEM106B locus, positively regulating

transcription levels (68). The “protective” allele of the lone TMEM106B coding region

SNP rs3173615 or p.T185S (also in linkage disequilibrium with rs1990662) has been

shown to impart faster protein turnover, resulting in an overall decrease in TMEM106B

expression (64). Together these findings suggest that these “risk” alleles confer genetic

susceptibility through increased TMEM106B gene expression or GOF mechanism.

Finally, independent of TMEM106B SNPs, FTLD-GRN patients exhibit reduced levels (<

50%) of the TMEM106B targeting miRNA-312 cluster members (miRNA-132, miRNA-

10

132*, miRNA-212) compared to controls, resulting in the derepression of TMEM106B

expression (69), further strengthening the hypothesis that reducing TMEM106B

expression may be beneficial in FTLD-GRN patients (70).

The first GWAS study to search for genetic modulators of CSF PGRN levels identified a

SNP (rs708384) in another gene encoding a transmembrane protein, FAM171A2, for

which the minor allele (A allele) was significantly linked to lower CSF PGRN levels in a

dose-dependent manner (71). FAM171A2 is highly expressed in cerebral vascular

endothelium and microglia in the mouse brain (71). HEK293 cells transfected with the

minor allele exhibited higher FAM171A2 expression levels compared to the major allele

(71). Overexpression of FAM171A2 in human umbilical vein endothelial cells decreased

PGRN expression and secretion, and siRNA knock-down of FAM171A2 significantly

increased PGRN expression (71). Little is known about the function of the single-pass

type I membrane protein FAM171A2; therefore, the precise mechanisms underlying its

regulation of PGRN expression are currently unclear (71).

The discovery of genetic variability within the GRN allele itself has led to the

identification of novel mechanisms that modify GRN expression levels. In particular, the

minor T-allele rs5848, a SNP in the 3’ untranslated region (UTR) of GRN, has been

shown to affect the predicted binding site of miRNA-659 (72). The presence of the minor

T-allele enhances the binding affinity of miRNA-659, thus expediting the rate of GRN

mRNA turnover (72). Homozygous carriers of rs5848 exhibit decreased serum and CSF

PGRN levels (58,72,73), though there is mixed evidence correlating this genotype with

an increased risk of developing non-GRN linked FTLD-TDP (74). Moreover, FTLD-GRN

11

patients with the common R493X GRN mutation carrying the A-allele of SNP rs9897526

found 21 bp downstream of the GRN intron 2 splice donor site of their intact allele

exhibit delayed disease progression (75). Though the precise mechanism for this delay

is currently unclear, its proximity to the splice site suggests it may alter the maturation

and stability of GRN mRNA. Additional non-coding regions of the GRN allele have also

been implicated in regulating GRN expression, including the methylation state of the

GRN promoter which is inversely correlated with GRN expression in lymphoblasts and

post-mortem brain tissue derived from FTLD patients (76,77). Expression of DNA

methyltransferase 3a (DNMT3a) mRNA is upregulated in the brain tissue of FTLD-TDP

patients compared to controls and thus may be responsible for hypermethylation of the

GRN locus (76). Also, overexpressing DNMT3a in lymphoblasts resulted in decreased

GRN expression, which could be reversed through treatment with a DNMT inhibitor

(76). Improving the understanding of the epigenetic regulation of GRN expression may

lead to the identification of novel therapeutic targets.

Several proteins are known to interact directly with PGRN, which mainly function to

regulate its trafficking and cellular localization. Sortilin (SORT1) is a type I membrane

glycoprotein and high-affinity neuronal receptor for PGRN, mediating PGRN

endocytosis and delivery to the lysosome, therefore regulating extracellular PGRN

concentrations (78). Sort1-/- knockout (KO) mice exhibit elevated (2.5-5X) brain and

serum levels of PGRN (78). The minor C-allele SORT1 SNP (rs646776) is associated

with decreased serum PGRN levels, with each allele copy generating a subsequent

~15% decrease, though it is unlikely that this SNP influences CNS PGRN levels since it

12

primarily alters hepatic SORT1 expression levels (79). Prosaposin (PSAP), another

protein that interacts with PGRN, is a precursor protein that is localized to the lysosome,

where it is processed into saposin (SAP) activators that are critical for glycosphingolipid

degradation (80). Further, PSAP facilitates the SORT1-independent delivery of

extracellular PGRN to lysosomes, made evident by the 5X increase in serum PGRN

levels in Psap-/- mice (81). At a genetic level, more copies of the PSAP minor T-allele

SNP (rs1867977) were correlated with decreased plasma PGRN levels in healthy

controls (82). Additional PGRN binding partners have been identified with roles

regulating extracellular PGRN levels, including members of the endoplasmic reticulum

(ER) chaperone network, such as calreticulin and disulfide isomerases ERp57 and

ERp5 (83). These factors likely play a key role in orchestrating the PGRN folding

process by forming several disulfide bonds throughout the cysteine-rich polypeptide.

1.1.7 Clinical and neuropathological phenotypes of FTLD-GRN and CLN11

FTLD-GRN includes a wide range of clinical presentations, including bvFTD, nfvPPA,

svPPA, Parkinson’s disease-like syndromes such as corticobasal syndrome, and an

amnestic disorder resembling AD (84). The most common clinical presentation in FTLD-

GRN patients is bvFTD, though there is also a high prevalence of nfvPPA (35,85).

Clinical symptoms of bvFTD in FTLD-GRN cases commonly include behavioural

disinhibition and increased apathy (35), while nfvPPA symptoms include effortful

inconsistent speech and agrammatism (86). Extrapyramidal motor dysfunction is rare in

FTLD-GRN but occurs with high prevalence in particular families beginning with

muscular rigidity and akinesia, eventually progressing to an inability to walk (35,87).

13

Generally, disease duration ranges from 3 to 12 years (84). The common

neuropathological hallmarks in FTLD-GRN include global atrophy of the frontal,

temporal, and, in some cases, the parietal lobes (85), hypometabolism of frontotemporal

cortices (88), significant white matter degeneration (89), FTLD-TDP type A pathology in

neocortical layer II (90,91), microgliosis (92), and pathological features associated

neuronal ceroid lipofuscinosis (NCL) including elevated expression of lysosomal

enzymes (93). Interestingly, only the unaffected FTLD-GRN post-mortem brain regions

show a clear reduction in PGRN protein levels compared to healthy control samples,

whereas atrophied regions show similar PGRN levels to controls (94). The elevated

PGRN in diseased brain regions may be derived from infiltrating microglial populations,

known to secrete high levels of PGRN.

The growing interest in understanding the contribution of lysosomal dysfunction to

FTLD-GRN disease progression stems from the recent discovery of a rare form of NCL

caused by homozygous GRN LOF mutations, CLN11 (95). Two GRN-null siblings were

the first identified cases of CLN11, presenting with early adulthood-onset NCL,

developing progressive retinal dystrophy, cognitive decline, cerebellar ataxia, and

recurrent myoclonic seizures. Eleven CLN11 cases have now been identified, exhibiting

a range of early- to late-onset phenotypes (96). The first neuropathological

characterization of an early-onset case (died age 27) found severe cerebellar atrophy,

numerous lipofuscin deposits in hippocampal and temporal cortical neurons, and early

signs of cytoplasmic translocation of TDP-43 with a lack of classic FTLD-GRN

inclusions (96). Therefore, these findings demonstrate that GRN-mutation dosage

14

strongly impacts the severity and age of disease onset and suggest that FTLD-GRN

may neuropathologically resemble a mild form of NCL (95).

1.2 PGRN neurobiology

In a quest to characterize novel granulocyte peptides with unknown immunoregulatory

functions, Bateman et al. discovered several GRNs in a minor subfraction of cysteine-

rich peptides purified from human leukocyte extracts (97). Key sequences of these

approximately 6 kDa cysteine-rich peptides are highly conserved between species

suggesting a long evolutionary history. Phylogenetic analysis of the GRN motif suggests

it evolved only once ~1.5 billion years ago in a primordial unicellular eukaryote predating

the divergence of the plant and animal kingdoms (98). These GRNs are believed to

have played a role in early cell signaling mechanisms, predating the evolution of

classical growth factor pathways such as Wnt, Hedgehog, and TGF-β (99,100). Given

PGRNs ancient evolutionary origin, it is not surprising that it is a key player in a diverse

array of biological phenomena, including proliferation (101), wound healing (102),

immunoregulation (103), embryogenesis (104,105), neurodegeneration (35,92), and

autophagy-lysosomal function (106,107). The identification of GRNs in such a variety of

biological contexts resulted in an equally divergent nomenclature; therefore, PGRN

(108) is known synonymously as granulin-epithelin precursor (109), proepithelin (110),

PC cell-derived growth factor (111), acrogranin (112), and epithelial transforming growth

factor (113).

15

1.2.1 PGRN structure

The human progranulin gene is located on chromosome 17q21 and consists of 12

coding exons (114). Full-length PGRN is an 88 kDa multifunctional secreted

glycoprotein made up of 7.5 cysteine-rich repeat units known as GRN modules (para-

GRN & GRN1-7) separated by short intervening spacer/linker sequences and an N-

terminal 17 amino acid signal peptide for secretion (111). Post-translational

glycosylation of PGRN contributes 20 kDa to the 68 kDa 593 amino acid polypeptide. It

is a member of the cysteine-rich mini protein family, which are generally known to

function as hormones, growth factors, ion channel modulators, or enzyme inhibitors. As

its name suggests, PGRN is a pro-protein, which cleavage gives rise to individual 6 kDa

GRNs. The highly conserved GRN modules consist of 4 pairs of cysteines plus 2

additional cysteine residues at each amino/carboxyl-terminal (CX5-6 CX5CCX8CCX6CC

XDXXHCCPX4CX5-6C) for a total of 12 cysteines, except for GRN-1 and para-GRN,

which possess 10 and 6 cysteine residues, respectively. Though no tertiary structures

have been established for full-length PGRN, some of the GRNs (GRN-2, -4, and -5)

have been crystallographically solved, identifying that these cysteine residues

significantly contribute to PGRNs rigid peptide backbone. The globular structure of

individual GRNs is composed of a core of 6 disulfide bridges forming a parallel stack of

4 beta-hairpins that twists through a left-handed super-helix (115,116).

1.2.2 PGRN expression in the brain

During early murine embryonic neurodevelopment (E15.5), Grn expression is limited to

small cell populations throughout the developing brain, often associated with

16

neurovasculature and epithelial populations (117). Neuronal Grn expression gradually

increases from E15.5 to P7, though it does not appear to be critical for neocortical

development since immature neurons migrating to the cortical plate (E15.5) lack Pgrn

immunoreactivity (117). Post-mitotic mature cortical neurons begin to show progranulin

positivity at P0, which increases into adulthood (117). Brightly Pgrn immunoreactive

Iba1+ amoeboid microglia are also scattered throughout the embryonic and postnatal

mouse brain from E15.5 to P14, often localized along developing white matter tracts

(117). In adult mice, Pgrn is expressed widely throughout the neocortex and

hippocampus, with immunoreactivity observed most prominently in large pyramidal cells

in cortical layers V/VI and the CA3 hippocampal region (117). Moderate Pgrn

expression is also observed in the thalamus, hypothalamus, amygdala, and midbrain,

with little to no expression in the striatum and brain stem (117). Additionally, only the

large Purkinje neurons of the cerebellum are positive for Pgrn staining. Interestingly, this

Pgrn expression is achieved despite a lack of a neuronal Grn expression observed

throughout the adult cerebellum, possibly indicating that Purkinje cell Pgrn may be

derived from distal brain regions (117).

Intracellularly, PGRN is primarily localized to lysosomal-associated membrane protein 1

(LAMP1) positive early lysosomes, the Golgi apparatus, and in vesicles immediately

adjacent to and partially overlapping with Ras-related protein Rab7+ late endosomes

(78,92). Recent findings suggest that once transported to the lysosome, full-length

PGRN undergoes rapid proteolysis into stable GRNs (118). Given that most anti-PGRN

antibodies also detect a subset of GRNs, it is likely that a significant proportion of

17

intracellular PGRN immunoreactivity likely represents cleaved GRNs.

Histological characterization of CNS cell type-specific Pgrn expression patterns has

shown that both neurons and microglia express Pgrn, though relative staining intensity

is much greater in microglia (117). GFAP-positive astrocytes, on the other hand, have

been shown to exhibit no Pgrn immunoreactivity in the adult mouse brain (117).

Nevertheless, subsequent western blot analysis of astrocyte- and neuron-enriched

whole mouse brain lysate fractions and single-nuclei RNA-sequencing (snRNA-seq) of

Grn+/+ mouse thalami have both observed similar baseline Pgrn/Grn expression levels in

both astrocyte and neuronal populations (119,120). Moreover, cultured primary human

astrocytes (121) and our own hiPSC-derived astrocytes (122) have both been shown to

express and secrete PGRN; thus, suggesting that astrocyte Pgrn expression in the adult

mouse brain may have been below the threshold of histological detection in the

previous study (117).

Conditional neuronal- and microglia-specific Grn KO mouse models have further

delineated individual cell types' contributions to total brain Pgrn expression levels. A

conditional neuronal floxed Grn KO mouse achieved through Nestin promoter-driven

Cre recombinase expression resulted in a 50% reduction in CNS Pgrn and no

detectable Pgrn in primary neuronal cultures, implying that approximately half of mouse

CNS Pgrn originates in neurons (123,124). The initial attempts to generate conditional-

microglia Grn KO mice, on the other hand, proved more difficult with primary microglia

cultures and isolated adult microglia only exhibiting a 50% decrease in Grn mRNA/Pgrn

18

expression (125-128). However, a subsequent study using the tamoxifen-inducible Cre-

estrogen receptor (CreER) fusion protein KO system driven by the microglial promoter

Cx3Cr1 achieved a 95% reduction in isolated adult microglial Grn mRNA expression

and a 26% decrease in overall brain Pgrn levels (129). Thus, suggesting that microglia

produce approximately a quarter of total CNS Pgrn, despite representing only 10% of

mouse brain cells (130); therefore, the remaining unaccounted for ~25% of CNS Pgrn

expression is likely collectively produced by astrocytes, endothelial cells, and

oligodendrocytes (120).

1.2.3 PGRN regulation in the brain

Mature neurons are post-mitotic and dependent on a high basal rate of autophagy

compared to non-neuronal cells since toxic metabolites and misfolded proteins cannot

be diluted through serial cell division. Transcription factor EB (TFEB) is a master

regulator of autophagy mediating the activation of the coordinated lysosomal expression

and regulation (CLEAR) gene network resulting in increased lysosomal biogenesis and

expression of lysosomal enzymes (131). The GRN gene possesses two TFEB binding

sites upstream of its coding sequence (131,132); therefore, GRN expression can be

upregulated by activation of TFEB (132,133). Inhibiting the key negative regulator of

autophagy, mTORC1, which functions to maintain TFEBs cytosolic localization (134),

increases PGRN expression in several cell lines, though importantly not in neuronal cell

types (135). Activating autophagy with the disaccharide trehalose through a mTORC1-

independent mechanism has been shown to increase GRN expression in hiPSC-

derived neurons in a partially TFEB-dependent manner (135).

19

Additionally, a genome-wide RNAi screen performed in the Neuro2A (N2a) mouse

neuroblastoma cell line identified that several genes regulating GRN expression were

enriched for autophagy-lysosomal pathways (ALPs) involved in both transcriptional and

post-transcriptional mechanisms (136). For example, when expression of the autophagy

regulator Foxo1 transcription factor was suppressed, Pgrn levels were increased by 1.7-

fold in N2a cells (136). Moreover, siRNA knockdown of Foxo1 in primary mouse cortical

neurons increased both Grn expression and Pgrn secretion, and simultaneous siRNA

knockdown of both Foxo1 and Grn prevented the accumulation of enlarged lysosomal

vesicles observed following siRNA-mediated Grn knockdown alone (136). The opposing

effects of TFEB and FOXO1 ALP transcription factor activation on the regulation of Grn

expression highlights the complex role PGRN plays in regulating lysosomal biology,

which will be explored further in the following sections.

Microglia constantly survey the brain parenchyma for damaged or unnecessary neurons

and synapses, infectious agents, and abnormal chemical stimuli. The GRN gene is

highly expressed in the myeloid lineage and is strongly upregulated in response to brain

and spinal cord injury (117,137). The GRN promoter region also contains several

CCAAT-enhancer-binding protein β (C/EBPβ) response elements (138). The C/EBPβ

transcription factor is commonly activated by the action of upstream pro-inflammatory

cytokines such as IL-6, known to induce GRN expression (139,140). The vitamin A1

metabolite all-trans retinoic acid (ATRA) has also been shown to induce upregulation of

GRN mRNA expression, specifically in cells of myeloid origin (141). Though the GRN

20

allele lacks a canonical retinoic acid response element in its promoter, ATRA may

activate downstream transcription factors capable of upregulating GRN transcription

(141). Interestingly, treating primary human fetal microglia cultures with pro-

inflammatory cytokines (IL-1β + IFNγ) and endotoxins (LPS, Poly I:C) was shown to

significantly suppress GRN/PGRN expression at the transcript and protein level (121).

Additionally, this study also demonstrated an opposing phenomenon in primary human

fetal astrocyte cultures, which despite lower baseline PGRN levels, significantly

upregulated GRN/PGRN expression in response to stimulation with either IL-1β + IFNγ

or Poly I:C (121). LPS treatment was omitted because human astrocytes lack TLR4 co-

receptor CD14 expression necessary for LPS-induced TLR4 activation (142).

Differential expression of alternative splice variants has also been implicated in the

regulation of GRN expression. Alternatively spliced GRN transcripts harboring a long

5’UTR sequence (219 nucleotides) containing two upstream open reading frames

(ORFs) exhibit translational repression and lower mRNA stability compared to

transcripts bearing short 5’UTRs (38-93 nucleotides) that lack upstream ORFs (143).

Upon GRN mRNA translation and translocation into the ER, PGRN is post-

translationally processed. Here, PGRN interacts with ER chaperones to ensure correct

protein folding through disulfide bond formation and undergoes regulated glycosylation

(144,145), which is essential for its sorting through the secretory pathway (146). After

transit through the trans-Golgi network, PGRN is either secreted into the extracellular

environment (147) or localized to the endolysosomal network via SORT1-dependent or

SORT1-independent transport mechanisms (78,81,148). Once secreted, full-length

21

PGRN can also undergo reuptake or be taken up by neighboring cells resulting in

lysosomal delivery; therefore, the relative abundance of endocytic PGRN receptors

(SORT1, LRP1, M6PR) in the plasma membrane plays a key role in regulating the

levels of intracellular PGRN/GRNs (78,81,148). Lysosomal PGRN can be cleaved into

individual GRNs by the lysosomal protease cathepsin L (CTSL) and at least one

additional unknown lysosomal enzyme that may be activated by cathepsin B (CTSB)

(149,150). As previously discussed in section 1.1.6, extracellular PGRN has been

shown to interact with another pro-protein, PSAP, which promotes PGRN homodimer

and oligomer formation and regulates the extracellular levels of PGRN (82). The ratio of

secreted full-length PGRN to individual GRNs is regulated by the secretory leukocyte

protease inhibitor (SLPI) (103). SLPI forms a complex with full-length PGRN preventing

its conversion into GRNs by extracellular proteases (e.g., elastase, metalloprotease

ADAMTS-7, proteinase 3), which functions to promote an anti-inflammatory and wound

healing response (103). SLPI is upregulated by neurons and astrocytes in response to

either neurological damage or stimulation with pro-inflammatory cytokines (121,151)

and, therefore, may function to regulate PGRN biology in the brains of FTLD-GRN

patients.

22

Figure 1-1 Brain regulation of extracellular PGRN proteolysis. Upon detecting

cellular damage/pathogen associated molecular patterns, microglia secrete

proinflammatory cytokines into the extracellular space. This induces upregulation and

secretion of SLPI by neurons and astrocytes. Microglia secrete the majority of

extracellular PGRN, which can be degraded into GRNs by extracellular proteases

(elastase, ADAMST-7, proteinase 3), however, neuron/astrocyte-derived SLPI can bind

to PGRN and protect it from proteolysis into GRNs, enhancing PGRNs anti-

inflammatory function.

23

As highlighted in section 1.1.6, genetic variance in TMEM106B has been identified as a

major modifier of FTLD-GRN disease progression (66). The TMEM106B type-II

transmembrane protein is highly expressed in neurons and is localized to the

endolysosomal compartment, where it controls the size, number, motility, and trafficking

of lysosomes in both neuronal and non-neuronal cell types (63,152,153).

Overexpression of TMEM106B in neurons has been shown to induce lysosomal

dysfunction resulting in enlarged vacuolar lysosomes with impaired acidification and a

reduced endocytic degradative capacity (63,152,154). Contrary to the GOF risk

hypothesis proposed based on SNP data, excess Tmem106b in N2a cells has been

shown to cause intra- and extracellular accumulation of PGRN (152). Further,

exogenously delivered TMEM106B expression in primary cortical mouse neurons

significantly increased TFEB nuclear localization activating expression of the CLEAR

gene network, providing a potential mechanism for TMEM106B-dependent positive

regulation of CNS PGRN expression (63).

Nevertheless, recent in vivo evidence from Tmem106b KO and overexpression studies

in the context of Grn-deficiency suggest that both resulted in detrimental consequences

to disease progression (155-158), therefore complicating exploration into the potential

therapeutic benefits of disrupting TMEM106B function. Another transmembrane protein,

FAM171A2, expressed in murine cerebral vascular endothelial cells and microglia, has

recently been demonstrated to regulate CSF PGRN levels (71). FAM171A2

overexpression and knockdown induced suppression and increased levels of PGRN,

respectively, indicating an inverse proportional relationship between FAM171A2 and

24

PGRN expression levels (71). Therefore, the mechanisms driving FAM171A2-mediated

regulation of CSF PGRN levels are currently under active investigation.

1.2.4 Mouse models of Grn-deficiency

Over the past 15 years, significant progress has been made characterizing the role of

PGRN in the CNS through the use of Grn KO mouse models. Although FTLD-GRN is

caused by GRN haploinsufficiency, Grn+/- mouse models have not replicated the

majority of the pathological hallmarks observed in this variant of FTLD-TDP (159).

Therefore, most studies have focused on models bearing homozygous deletion of Grn

expression (Grn-/-), which better recapitulate FTLD-GRN pathological progression

(92,93,106,107,159-164) but would be more accurately described as a model of the rare

form of NCL (CLN11) caused by homozygous LOF GRN mutations. The

characterization of Grn-/- mice has uncovered that PGRN is an essential regulator of

neuroinflammation, autophagy, and synaptic pruning in the CNS.

1.2.4.1 Neuroinflammation

Microgliosis is a hallmark of FTLD-GRN, in which the affected brain regions are

populated with increased numbers of hyperactivated microglia with thick processes and

an enlarged soma. Therefore, major research efforts have sought to better understand

the role of PGRN in regulating inflammatory processes. Cultured Grn-/- and

GrnR493X/R493X macrophages and microglia express and secrete elevated pro-

inflammatory cytokines in response to treatment with inflammatory mediators, likely due

to their suppressed levels of Il-10 expression and exaggerated NF-κB activation

25

(125,129,161). These excessive pro-inflammatory phenotypes can be acutely rescued

through exogenous delivery of murine Grn (125,165). Extraordinarily, Grn-/- microglia

stimulated with LPS + IFNγ were found to secrete elevated levels of Il-10 into the culture

medium, despite lower levels of Il-10 mRNA, reflecting differential regulation at the

transcriptional and secretory levels (125). Nevertheless, increased Il-10 secretion was

insufficient to reduce the increased pro-inflammatory state observed in Grn-/- microglia,

possibly indicating that Grn-/- microglia are less responsive to the anti-inflammatory

action of Il-10 signaling (125). Treatment of Grn+/+ rat primary cortical neurons with

human PGRN increased expression and secretion of cytokines/chemokines associated

with chemoattraction, including Il-10 (166). Grn-/- hippocampal slices are significantly

more sensitive to microglia-mediated neurotoxicity than Grn+/+; however, this phenotype

can be replicated in Grn+/+ slices by inhibiting Il-10 function (161). These findings

highlight PGRNs essential role in suppressing hyperinflammatory immune responses.

Pgrn-deficient microglia and macrophages overexpress and secrete excessive

quantities of TNFα (125,161). Several reports have suggested that Pgrn mediates its

anti-inflammatory effects by extracellularly disrupting TNFα signaling through direct

inhibition of its plasma membrane receptors (TNFR1/2) (167,168); however, the veracity

of this direct binding hypothesis has been challenged (169-171). Whether PGRN

directly modulates TNFα signaling through TNFR1/2 binding competition or an unknown

indirect mechanism, several studies have clearly established that PGRN and the multi-

GRN derivative Atsttrin negatively regulate TNFα expression and function (167,172-

174). Conditioned media obtained from activated (LPS + IFNγ) primary Grn-/- microglia

26

induced a greater increase in cell death in Grn+/+ primary cortical neurons than media

derived from Grn+/+ microglia, suggesting that activated Grn-deficient microglia may

secrete neurotoxic factors (125). A recent publication by the Huang lab further

elucidated this process, establishing that elevated levels of complement proteins (C1qa

and C3b) in primary Grn-/- microglia conditioned media (MCM) are at least partially

responsible for increased primary cortical neuronal death compared to Grn+/+ MCM

(120). Since inhibition of membrane attack complex (MAC) formation completely

blocked Grn-/- MCM-induced neurotoxicity, it is likely that the additional toxicity

promoting factors in Grn-/- MCM may be downstream complement pathway components

involved in MAC assembly (120).

Modelling complete loss of Pgrn in vivo consistently results in significant astro- and

microgliosis established as early as 7 months of age, which then rapidly accelerates

throughout the aging process (159,161,162,164,165). The most neuroinflammation-

prone regions in Grn-/- mouse brain include the cerebral cortex, CA3 hippocampal

region, and ventral posteromedial (VPM)/ventral posterolateral (VPL) thalamic nuclei.

Furthermore, a region of cortical layer IV in Grn-/- mice (S1BF) innervated by the highly

pathologically vulnerable VPM/VPL thalamic nuclei has also been shown to exhibit

increased microgliosis (164). Interestingly, a recent study examining differentially

expressed genes in primary Grn-/- microglia identified significant enrichment for the

overexpression of cell cycle-related genes, suggestive of an increased proliferative

capacity in line with the observed microgliosis phenotype (175). The VPM/VPL regions

of the thalamus in aged Grn-/- mice have also been shown to accumulate deposits of the

27

complement proteins C1qa and C3b, which when disrupted (Grn-/-;C1qa-/- and Grn-/-

;C1qa-/-;C3-/- mice) partially reduces the severity of thalamic microgliosis (92,120). Grn-/-

microglia also exhibit impaired motility as evidenced by their decreased chemotaxis

towards ADP/ATP gradients and limited process extensions/retractions observed at

baseline and post-laser-induced injury (129). The loss of Pgrn has also been shown to

increase the sensitivity of Grn-/- mice to both chemically- and physically-induced CNS

damage resulting in exaggerated neuroinflammation and increased neurotoxicity

compared to Grn+/+ (125,176). Grn-/- mice acutely infected with Listeria monocytogenes

(L. monocytogenes) mount a dysfunctional innate immune response that fails to

efficiently clear the bacterial burden in various organs (106,161).

The previously held notion that microglia only exist in the binary categorization of either

resting or activated (M1 vs. M2) has been disputed by transcriptomic analyses, which

have identified greater biological complexity in the form of several unique homeostatic

and disease-associated microglia (DAM) gene expression signatures (177). Several

studies have transcriptomically characterized aged Grn-/- brains identifying upregulated

genes related to innate immunity, including several cytotoxic factors (92,107,164).

Additionally, the aforementioned snRNA-seq study of microdissected Grn+/+ and Grn-/-

thalami identified drastic and dynamic microglia-specific transcriptional changes

throughout disease progression, characterized by downregulation of homeostatic genes

and overexpression of genes associated with inflammation, lysosomal function, and

neurodegeneration (120). Moreover, a similar 11 gene DAM signature emerged

following quantification of differentially expressed genes in Grn+/+ and Grn-/- primary

28

microglia cultures, including the downregulation of the homeostatic genes Tgfbr1 and

Csf1r (175). Intriguingly, conditionally depleting either neuronal or microglia Grn

expression fails to replicate any of the major neuropathological hallmarks observed in

Grn-/- mice. This may be because only partial microglial Grn KO (LysM promoter-driven)

was achieved compared to nearly complete penetrance in conditional neuronal KO

models (123,124,127,128). However, a recent effort achieved a more robust depletion

of Grn expression using the tamoxifen-inducible CreER system driven by the microglial-

specific Cx3Cr1 promoter and observed behavioural deficits in aged mice resembling

Grn-/- phenotypes (129). Cumulatively, these findings suggest that PGRN plays a major

role in suppressing aberrant microglial activation throughout the aging process.

1.2.4.2 Lysosomal dysfunction

Since the identification and characterization of CLN11, a rare form of NCL caused by

homozygous GRN mutations, there has been growing research interest in uncovering

the role of PGRN in lysosomal function. Lysosomal abnormalities are among the first

observable phenotypes in Grn-/-, developing months before chronic hyperinflammation

and microgliosis (107,119,120,178). 3 month old Grn-/- hippocampal neurons exhibit

increased lysosomal size and abundance, often presenting with lamellar

pseudomembranous inclusions characteristic of NCL (107). Transcriptomic and

proteomic analysis of 2 month old Grn-/- mice cerebral cortex identified significant

upregulation of several key lysosomal enzymes (Hexb, Tpp1, Ctsb, Ctsl, Dpp7) (179).

Lipofuscin accumulation, a hallmark of NCL, commonly observed in neuropathological

hot spots of aged Grn-/- brains (159,163-165), has been observed in the thalamus of 2

29

month old Grn-/- mice (179). Moreover, 2-3 month old Grn-/- mice exhibit major

deficiency in xenophagy (ability to clear intracellular pathogens) following intravenous

infection with L. monocytogenes, with increased bacterial burdens detected in the brain,

liver, and spleen in Grn-/- mice 1 week post-infection (106,161). Lipidomic analysis of 2-3

month old Grn-/- brains has detected decreased levels of both di-18:1 and di-22:6

bis(monoacylglycerol)phosphates, lipid species known to stimulate lysosomal hydrolase

activity important for lysosomal sphingolipid degradation (Denali Therapeutics,

Alzheimer’s Association International Conference 2020). However, these studies failed

to identify which specific cell populations are most responsible for driving lysosomal

dysfunction.

Following the discovery that aged Grn-/- mice brains exhibit elevated levels of NCL

related proteins, such as SapD (PSAP peptide product), cathepsin D (Ctsd), Lamp1,

Tmem106b, and subunit C of mitochondrial ATP synthase (93), the Capell lab sought to

determine whether a particular CNS cell type plays an outsized role in the establishment

of these lysosomal pathologies. Western blot analysis of whole-brain lysates derived

from aged 20 month old Grn-/- mice found an increase in the maturation rate of

cathepsin enzymes (Ctsd, Ctsb, and Ctsl), resulting in increased enzyme activity that is

absent in 3 month old Grn-/- mice (119). However, characterizing microglial enriched

lysate fractions from young Grn-/- brains revealed decreased cathepsin maturation

levels, indicated by the accumulation of the pro-forms of Ctsd, Ctsb, Ctsl, and Ctss, not

observed in Grn+/+ microglial lysates (119). Furthermore, these young Grn-/- derived

microglial enriched fractions also possessed increased lysosomal proteins Lamp1 and

30

SapD. On the other hand, the microglial depleted fractions produced from young Grn-/-

and Grn+/+ express similar levels of the pro- and mature-forms of cathepsins and exhibit

no increase in Lamp1 or SapD (119). Strikingly, throughout the aging process (3 months

to 12 months), dysregulation of Ctsd and Ctsb maturation patterns appeared to be

partially normalized in Grn-/- microglia, while non-microglial cells compensated for the

early decreased degradative capacity of microglia by increasing their respective Ctsd

enzyme activity (119). Therefore, providing evidence supporting microglia as the

primary drivers of the early lysosomal pathology in Grn-/- mice. These results are

supported by several CNS cell type-specific transcriptomic analyses suggesting that

microglia-specific transcriptomic signatures in Pgrn-deficient brains exhibit the highest

degree of differential gene expression, transitioning from a homeostatic to disease state

characterized by significant endolysosomal dysfunction (92,106,120).

The development of FTLD-TDP type A pathology in FTLD-GRN patients is thought to be

caused by an inability to proteolytically process nuclear and cytoplasmic aggregated

TDP-43 granules. Significant research efforts have been conducted to explore PGRNs

role in regulating the autophagolysosomal pathway to better understand this

pathological hallmark. Grn-deficient mice have been shown to accumulate neuronal

cytoplasmic inclusions of both TDP-43 and the ubiquitin-binding autophagosome cargo

protein p62, most notably in the thalamus and hippocampus (93,120,161,164,165). This

selective vulnerability was elegantly demonstrated by infecting both Grn+/+ and Grn-/-

primary cortical neurons with an adeno-associated virus (AAV) vector expressing the

aggregation-prone C-terminal fragment (CTF) of TDP-43. This resulted in an increase in

31

TDP-43 CTF levels in Grn-/- neurons that could be rescued by recombinant (rec.) Pgrn

treatment and phenotypically replicated in Grn+/+ neurons by inhibiting autophagosome

degradation (106). Activation of autophagy by serum/amino acid starvation in primary

Grn-/- microglia and cortical neuron cultures has been used to further highlight their

propensity for overexpression of lysosomal genes/proteins (180) and impaired

autophagosome clearance (106), both of which could also be reversed through

exogenous delivery of rec. Pgrn. Grn-/- primary cortical neurons and Grn-/- cortical tissue

have been found to express reduced levels of phosphorylated 5’ AMP-activated protein

kinase alpha, a key positive regulator of autophagy; therefore, PGRN may positively

regulate upstream autophagolysosomal signaling pathways (106). Together these

findings suggest that impaired autophagy caused by PGRN deficiency plays a central

role in TDP-43 pathology.

1.2.4.3 Dysregulated lipid metabolism

Lipidomic analysis of the frontal cortices of aged (12 month old) Grn-/- mice identified the

accumulation of polyunsaturated triacylglycerides (TAGs) and cholesterol esters (CEs),

accompanied by a corresponding reduction in diacylglycerides (DAGs) and

phosphatidylserines of increasing chain length and number of unsaturated bonds (107).

Thus, suggesting that PGRN/GRNs are involved in the regulation of long-chained

polyunsaturated fatty acid metabolism. This hypothesis is further supported by the

observed Grn-/- specific transcriptomic differential downregulation of DAG-specific lipid

hydrolases, providing a potential mechanism for PGRN-dependent positive regulation of

the enzymatic conversion of phosphatidic acid into DAGs (107). Long polyunsaturated

32

TAG-rich lipid-droplets (LDs) are produced in cells of myeloid origin in response to

inflammation and stress; intriguingly, a recent study identified GRN as a primary

negative regulator of LD formation (181). Additionally, they found that 9-10 month old

Grn-/- mice hippocampi contain increased levels of LD-accumulating microglia, a

process previously identified in a Drosophila model of neurodegeneration (182). Grn-/-

LD-rich microglia exhibited decreased phagocytic activity and secreted elevated levels

of pro-inflammatory cytokines following LPS stimulation compared to LD-low Grn-/-

microglia. Since lysosomal dysfunction precedes the establishment of pro-inflammatory

phenotypes in Grn-/- brains and LDs are known sites of cytokine production and storage

(183), altered lipid metabolism caused by GRN deficiency may prime microglia for

hyperinflammation throughout the aging process.

Huang et al. recently conducted whole-brain proteomic network analysis in 3 and 19

month old Grn+/+ and Grn-/- mice. This unbiased quantitative proteomic analysis

replicated previously described early lysosomal dysfunction phenotypes (179), detecting

overexpression of lysosomal proteins (Gns, Scarb2, Hexb, Ctss); however, they also

observed a substantial subset of lysosomal proteins that exhibited decreased

expression (178). A significant proportion of the most downregulated proteins identified

in 3 month old Grn-/- pooled brain lysates were enriched for proteins involved in

lysosomal lipid catabolism, with functions ranging from regulation of lipogenesis and

lipid homeostasis to enzymatic roles in lipid degradation pathways (178). The most

significantly elevated protein in the aged Grn-/- cohort was the transmembrane type I

glycoprotein Gpnmb (178), which exhibits diverse cellular functions and has been found

33

to exhibit increased expression in several neurodegenerative diseases (184). CNS

Gpnmb expression was primarily observed in Iba1+ microglia and was most prominent

in the thalamus of 19 month old Grn-/- mice (178). Though it is unclear whether Gpnmb

plays a direct role in Grn-/- pathophysiology, results demonstrating that Gpnmb is

upregulated in the murine CNS in response to direct inhibition of the lysosomal

hydrolase glucocerebrosidase (GCase) suggest that it may be elevated in Grn-/- brains

in response to lysosomal dysfunction (185). Moreover, this hypothesis is further

strengthened by findings that Pgrn directly regulates GCase lysosomal localization and

enzymatic activity levels (186,187). Interestingly, microglia expressing elevated levels of

Gpnmb have been implicated in the demyelination process in multiple sclerosis (188),

potentially explaining the significant decrease in myelin basic protein levels found in the

aged Grn-/- brain (178), providing another plausible link between lysosomal dysfunction

and the onset of chronic inflammation. Once again, the early establishment of lysosomal

dysfunction was shown to precede the onset of neuroinflammatory processes,

suggesting that disruptions to lysosomal lipid metabolism may represent a key

instigating factor driving neurodegenerative processes in FTLD-GRN.

1.2.4.4 Neural connectivity deficits and neurodegeneration

Microglia play a pivotal role in modulating neural networks during neurodevelopment

and aging through a process known as synaptic pruning, by which microglia trogocytose

(selective partial phagocytosis) the membranes of presynaptic boutons and axons

(189). The disruption of microglial immunoregulation in the CNS of mouse models of

Grn deficiency has profound downstream consequences on the neural activity of the

34

thalamocortical circuit, which may drive behavioural abnormalities (92).

Electrophysiological recordings of hippocampal slice cultures have demonstrated

reduced synaptic connectivity in Grn-/- mice, likely due to impaired synaptic plasticity

resulting in an overall depression of activity (163). Additionally, the Leavitt laboratory

also showed that the apical dendritic arbors of hippocampal CA1 region pyramidal

neurons showed significantly lower spine density in Grn-/- compared to Grn+/+ (163).

Grn-deficient primary microglia co-cultured with Grn+/+ primary cortical neurons engulfed

higher levels of C1qa-tagged synapses resulting in lower synaptic density surrounding

Grn-/- microglia (92). Grn-/- mice also exhibited decreased synaptic density in the ventral

thalamus, which was first detected at 7 months of age and continues to decrease

throughout the aging process (92). The rescue of synaptic loss in double Grn-/-;C1qa-/-

and triple Grn-/-;C1qa-/-;C3-/- KO mice demonstrated that complement-mediated

microglial synaptic pruning is responsible for this observed Grn-/--specific decrease in

synaptic density (92,120).

These neurological aberrancies drive behavioural defects in both Grn+/- and Grn-/- mice

(160,162,163,190). Two of the most well-characterized Grn-/- behavioural abnormalities

include a chronic grooming phenotype resembling a form of obsessive-compulsive

disorder (OCD) that results in an increased incidence of skin lesions and a male-specific

increase in anxiety (92,160,163,165). Remarkably, the C1qa-dependent synaptic

pruning in Grn-/- VPM/VPL thalamic nuclei has been shown to selectively target

inhibitory synapses for elimination (92). The loss of inhibitory synapses in the

thalamocortical circuit in Grn-/- mice results in electrophysiological hyperexcitability in

35

thalamic slice cultures, which may in part be responsible for their excessive grooming

phenotype (92). Incredibly, both the hyperexcitability and chronic grooming Grn-/-

phenotypes can be rescued by C1qa deletion (92).

Given the association of increased plasma TNFα levels in OCD patients (191) and the

growing consensus that PGRN negatively regulates TNFα signaling, researchers sought

to demonstrate the impact of Tnfα deficiency on the Grn-/- OCD phenotype (129). Since

the cortico-basal ganglia circuit in the nucleus accumbens (NA) of ventral striatum has

also been implicated in OCD patients (192), electrophysiological brain slice recordings

were performed on the medium spiny neurons (MSNs) in the core of the NA, identifying

a similar hyperexcitability phenotype (129). Extraordinarily, both hetero- and

homozygous deletion of Tnfα rescued both the chronic grooming and NA MSNs

hyperexcitability Grn-/- phenotypes (129). However, both Grn-/-;Tnfα+/- and Grn-/-;Tnfα-/-

mice still possess defects in social behavioural observed in Grn-/- mice, suggesting

TNFα may only play an important role in the regulation of the neural circuits associated

with compulsive behaviours (129). As mentioned in section 1.2.4.1, Krabbe et al. also

observed this OCD-like chronic grooming phenotype in 8-10 month old tamoxifen-

induced microglia-specific Grn KO mice, suggesting loss of microglial Pgrn may be

sufficient to disrupt neural circuitry (129). Ikbkb encodes the inhibitor of NF-κB subunit

beta required for NF-κB activation and therefore regulates the transcription factor

responsible for upregulating inflammatory cytokines, including TNFα. To further

evaluate the contribution of microglial-induced neuroinflammation to Grn-/-

pathophysiology, they also generated the less penetrant (LysM promoter-driven) Cre/lox

36

microglial-specific KO mouse model targeting both Grnfl/fl and Ikbkbfl/fl (129). Despite

only achieving an ~50% reduction in adult Cd11b+ microglia Grn mRNA levels, this

LysM-Grnfl/fl mouse model exhibited both OCD and social behavioural abnormalities,

that were rescued by simultaneous microglia-specific KO of Ikbkb (LysM-Grnfl/fl;Ikbkbfl/fl)

(129). Thus, suggesting that the increased propensity for NF-κB activation in Grn-/-

microglia (193) is at least partially responsible for their hyperinflammatory phenotype,

which results in the overexpression of pro-inflammatory cytokines, including TNFα that

drives the OCD phenotype.

Neurodegeneration in Grn-/- has been shown to occur primarily in the VPM/VPL thalamic

nuclei, and therefore may also contribute to the development of behavioural phenotypes

(120,164). Interestingly, 19 month old Grn-/- thalamic snRNA-seq transcriptomes

exhibited a strikingly lower quantity of nuclei expressing cell type-specific markers

associated with excitatory neurons and a modest increase in the number of inhibitory

neurons (120). Immunohistochemical verification of these findings confirmed that the

loss of excitatory neurons could be detected by Pkcδ and Foxp2 staining in Grn-/-

thalamus in both 12 and 19 month old Grn-/- mice (120). 19 month old Grn-/- thalamic

Foxp2+ neurons were found surrounded by microglia containing intracellular Foxp2+

debris suggestive of microglial-mediated neurotoxicity and phagocytic clearance of

dying excitatory neurons (120). The selective vulnerability of Grn-/- excitatory primary

cortical neurons to Grn-/- MCM compared to Grn-/- GABAergic primary cortical neurons

supports these findings (120). Extraordinarily, Grn-/-;C1qa-/-;C3-/- mice have near-

complete protection of thalamic Pkcδ+ excitatory neurons, further signifying the

37

importance of complement-mediated neurodegenerative mechanisms in FTLD-GRN.

Though the loss of VPM/VPL inhibitory synapses is apparent in 8,12, and 19 month old

Grn-/- mice, the electrophysiological hyperexcitability phenotype was only assessed at

12 months old (92); therefore, it is possible that the hyperexcitability phenotype may be

less prominent in 19 month old Grn-/- mice due to the continued neurodegeneration of

thalamic Pkcδ+ and Foxp2+ excitatory neurons between 12 and 19 months of age

(120). Furthermore, the increase in the number of thalamic nuclei expressing inhibitory

neuron markers in Grn-/- mice from 12 to 19 months of age may represent a

compensatory mechanism due to complement-mediated inhibitory synaptic pruning

(92,120).

1.2.5 hiPSC-derived CNS models of GRN-deficiency

Though rodent models have been essential for uncovering the CNS functions of Pgrn,

one must be careful when directly applying these findings to the clinical manifestation of

FTLD-GRN. The inability to translate vast amounts of data obtained in preclinical

therapeutic interventions in AD mouse models into viable disease-modifying agents

provides reason to be cautious projecting murine pathogenic mechanisms onto human

disease (194). Until recently, post-mortem analysis was the primary option for studying

human neurobiology; however, in 2007, the Yamanaka group developed a technology

that could achieve human adult somatic cell reprogramming into hiPSCs (195),

unlocking the potential to generate human cell types from any genetic background.

Subsequently, researchers around the world worked to establish hiPSC differentiation

protocols for a variety of cell types, including those present in CNS tissue such as

38

neurons, astrocytes, and microglia (196-198). To date, our group and several others

have generated hiPSC lines from symptomatic and presymptomatic carriers of

haploinsufficient GRN mutations and differentiated them into CNS cell types for FTLD-

GRN disease modelling and preclinical drug development (60,122,135,199,200).

The first FTLD-GRN hiPSC line was generated from fibroblasts obtained from a patient

carrying the heterozygous p.S116X mutation (designated S116X+/-) and was

successfully differentiated into neurons and microglia (60). S116X+/- hiPSC-derived

neurons expressed and secreted reduced levels of PGRN, exhibited increased

sensitivity to the broad-spectrum kinase inhibitor staurosporine, and expressed

decreased levels of the ribosomal Ser/Thr kinase S6K2 involved in PI3K/PKB and

MAPK signaling pathways. Importantly, these latter two phenotypes could be rescued

by lentiviral delivery of GRN expression (60). Another study generated hiPSC lines from

three different FTLD-GRN patients carrying the same GRN IVS1+5G>C splice donor

site mutation (199). Differentiation of healthy control and FTLD-GRN splice donor

mutant hiPSC lines into cortical neurons uncovered that these GRN-deficient lines

expressed significantly lower levels of cortical neuron markers (CTIP2, TBR1, FOXG1)

at days in vitro (DIV) 40 compared to healthy control neurons. This inefficient cortical

neuron maturation phenotype was corrected by genetically inserting GRN cDNA into the

AAVS1 safe harbor locus of one of the FTLD-GRN lines (199). However, more recent

studies, including those published by our lab using other cortical neuron differentiation

methods, have failed to observe this defect in several FTLD-GRN hiPSC lines, including

a homozygous knock-in GRN p.R493X mutant line (122,200,201). These findings

39

suggest that GRN haploinsufficiency does not disrupt all pathways required for inducing

robust cortical neuron differentiation and that a more efficient differentiation technique

may have been required to obtain similar differentiation efficiency across their healthy

control and FTLD-GRN lines.

Valdez et al. generated an FTLD-GRN patient hiPSC line bearing the p.A9D signal

peptide disrupting mutation that blocks PGRN post-translational glycosylation and

secretion in order to assess lysosomal defects in patient-derived cortical neurons

(54,200). Subsequently, they corrected this heterozygous GRN mutation using

CRISPR/Cas9 gene editing to produce an isogenic control hiPSC line in order to

eliminate any potential phenotypic contributions from disparate genetic backgrounds

between the isogenic control and FTLD-GRN line. Aged GRN+/- hiPSC-derived cortical

neurons exhibited decreased lysosomal proteolysis, resulting in the accumulation of

intracellular electron-dense lipofuscin deposits and increased TDP-43 expression (200).

This study also assessed CTSD enzyme maturation levels in both isogenic control and

mutant hiPSC-derived cortical neurons. Both 35 and 100 DIV GRN+/- neurons exhibited

increased levels of mature CTSD; notably, the ratio of mature to immature CTSD was

significantly increased in the aged GRN+/- neurons (200). However, CTSD enzyme

activity normalized to mature CTSD expression levels were significantly lower in both 35

and 100 DIV GRN+/- hiPSC-derived cortical neurons (200). Though not demonstrated in

hiPSC-derived neurons, this study also added to the growing consensus suggesting that

PGRN and CTSD are binding partners and that PGRN’s C-terminus and/or GRN-7 may

promote CTSD maturation/activity (202,203). This finding is counterintuitive given the

40

increased levels of CTSD maturation frequently observed in the context of GRN-

deficiency, although it is possible that the impact of general lysosomal dysfunction on

CTSD maturation outweighs the effects of limited PGRN chaperone activity.

FTLD-GRN patient hiPSC-derived CNS tissue provides an invaluable resource for the

preclinical development of new therapeutics aimed at elevating GRN/PGRN expression

levels. For example, neurons derived from the S116X+/- hiPSC line have been used for

the in vitro validation of the SORT1 antagonist 2-methyl-6-(phenylethynyl)-pyridine,

which induced accumulation of extracellular PGRN in cell culture medium (204). As we

will discuss in section 1.2.7, a SORT1 targeting approach using the AL001 anti-SORT1

antibody has recently begun phase 3 clinical trials for FTLD-GRN, highlighting the

importance of preclinical validation in hiPSC models. Additionally, treatment of S116X+/-

hiPSC-derived cortical neurons with suberoylanilide hydroxamic acid, a histone

deacetylase inhibitor previously shown to upregulate GRN expression through

epigenetic modification (205), has been shown to significantly increase intra- and

extracellular PGRN expression from the intact allele (201). Another group, studying the

therapeutic potential of the disaccharide trehalose, generated hiPSC lines from a

healthy control and FTLD-GRN patient carrying the GRN p.R198GfsX19 mutation (135).

They demonstrated that p.R198GfsX19 mutant hiPSC-derived neurons express slightly

elevated levels of the LC3-II autophagosome marker compared to control neurons.

Treatment of p.R198GfsX19 mutant hiPSC-derived neurons with the autophagy

activator trehalose increased intracellular PGRN expression but further increased the

levels of LC3-II (135). In chapter 4, we will describe studies we conducted in our lab

41

using novel GRN nonsense mutant hiPSC lines demonstrating the therapeutic potential

of nonsense mutation readthrough drugs for the treatment of FTLD-GRN (122).

1.2.6 Proposed mechanisms of neurodegeneration in FTLD-GRN

The multifunctional nature of PGRN has complicated the elucidation of the key

biological mechanisms driving FTLD-GRN pathophysiology. The pathological cascade

resulting in FTLD-GRN symptomatic onset is multi-decade in nature, involving both

biochemical and cellular phases that precede the clinical manifestation of FTLD-GRN

(206,207). Drawing heavily upon the discoveries made in Grn-/- mice, it is likely that

microglial lysosomal dysfunction is a key initial step in the disease process, preceding

the establishment of neuroinflammation and gliosis (179). Götzl et al. elegantly

demonstrated that disrupted microglial lysosomal function in Grn-/- mice is one of the

earliest pathological events. Interestingly, the lysosomal dysfunction phenotype was not

observed in non-microglial populations until the latter stages of disease progression.

This may reflect both the heavy proteolytic strain associated with constantly sampling

their extracellular microenvironment and the fact that microglia express the most PGRN

in the brain on a per-cell basis (see section 1.2.2). The role of PGRN/GRNs in the

regulation of lysosomal biology is multifaceted, with functions including activation of

lysosomal proteases (119,202,208) and glycolipid hydrolases (186,187), promotion of

lysosomal acidification (180), and the homeostatic maintenance of lysosomal

morphology and biogenesis (93,107,178).

42

As discussed in section 1.2.4.3, GRN-deficiency has been increasingly linked to

dysfunctional microglial lipid metabolism, as evidenced by abnormal accumulation of

TAG and CE lipid species and LDs (107,181). These findings, combined with the

mounting evidence that PGRN is a key regulator of autophagy (106,131,209), provide a

potential mechanistic link to the eventual establishment of the pro-inflammatory DAM

phenotype through immunometabolic disruption. Immunometabolism is a cellular

process that regulates the way immune cells respond to environmental stimuli through

the flexible reprogramming of metabolic pathways (210). Microglial metabolic plasticity

is increasingly understood as a central node that controls the distinct states of microglial

homeostasis and activation. For example, homeostatic microglia preferentially

metabolize glucose through oxidative phosphorylation, while DAM are primarily

dependent on glycolysis to fuel their pro-inflammatory activities (reviewed here

(210,211)). Remarkably, transcriptomic analysis of LD-accumulating microglia isolated

from 9 month old Grn-/- mice hippocampi were found to be hyperglycolytic, exhibiting

significant disruption of metabolic pathways, including the TCA cycle and lipid

metabolism (181). The loss of microglial PGRN/GRNs may disrupt metabolic flexibility,

limiting the ability of microglia to revert back to a homeostatic state following initial pro-

inflammatory activation (212). However, it is still unclear whether disturbances to Grn-/-

microglial energetic pathways are a result of early lysosomal dysfunction or are an

additional independent phenotype caused by the loss of multifunctional PGRN/GRNs.

Regardless, the mechanisms driving dysfunctional microglial immunometabolism in

FTLD-GRN warrant further investigation.

43

Neuroinflammation is a prominent feature of FTLD-GRN, with severe microgliosis

observed in pathologically affected brain regions. Highly reactive pro-inflammatory

microglia with enlarged somas and thick processes are the central drivers of FTLD-GRN

pathogenesis. GRN-deficient DAM exhibit a transcriptomic signature unique to FTLD-

GRN neurodegeneration, characterized by suppression of homeostatic microglial genes

(Tgfbr1, Mef2c, P2ry12, etc.) and upregulation of microglial neurodegeneration-

associated genes (Apoe, Csf1, Clec7a, etc.) (120,212). These activated DAM secrete

elevated levels of pro-inflammatory mediators into the surrounding brain parenchyma,

including complement cascade proteins C1QA and C3. Complement proteins have

been shown to accumulate in the frontal cortices of FTLD-GRN patients and have also

been detected at heightened levels in their CSF. Interestingly, the levels of complement

proteins C1QA and C3B in FTLD-GRN patient CSF are negatively correlated with mini-

mental status exam score, in which lower scores indicate greater cognitive decline (92).

The Huang group’s research has strongly implicated these complement proteins in

FTLD-GRN pathogenesis, identifying several mechanisms linking neuroinflammation to

neurodegeneration (92,120). These include both selective targeting of inhibitory

synapses for complement-mediated microglial synaptic pruning (92) and the formation

of complement MAC on the excitatory neuronal cell surface that induces TDP-43

cytoplasmic translocation and neuronal cell death (120). Clinical validation of these

disease processes discovered in mouse models has been obtained through an fMRI

study on asymptomatic/prodromal GRN mutation carriers compared to healthy controls.

This study identified enhanced connectivity across four major neural networks (salience,

non-fluent variant primary progressive aphasia, perirolandic, and default mode), with the

44

unifying feature being thalamocortical hyperconnectivity (213), mirroring complement-

mediated phenotypic observations in aged Grn-/- mice.

Examining the relationship between PGRN and the FTLD-GRN risk factor TMEM106B

is essential for developing a comprehensive understanding of the basis of neuronal

lysosomal dysfunction in FTLD-GRN. Neuronal TMEM106B lysosomal functions include

maintaining lysosomal acidity through positive regulation of V-ATPase activity and

regulating axonal transport of lysosomes (179,214). Current evidence suggests that

TMEM106B risk alleles accelerate FTLD-GRN pathology through a GOF mechanism

(see section 1.1.6). Overexpression of TMEM106B driven by a neuron-specific promoter

in aged Grn+/+ mice identified tight regulation of TMEM106B protein levels, which

exhibited no increase despite a several-fold increase in TMEM106B mRNA (155).

However, TMEM106B overexpression in aged Grn-/- mice resulted in a 50% increase in

TMEM106B levels compared to normal aged Grn-/- mice, providing further evidence that

PGRN is involved in homeostatic negative regulation of TMEM106B levels (155).

Overexpression of TMEM106B in aged Grn-/- mice accelerated lysosomal dysfunction in

these mice (155), thus providing additional support for the GOF TMEM106B risk allele

hypothesis. On the contrary, several recent publications have demonstrated that

deletion of Tmem106b in a Grn KO background also exacerbates Grn-/- pathology

(156,157,215). A more clinically relevant double heterozygous KO model (Grn+/-

;Tmem106b+/-) was generated to assess whether a partial reduction in Tmem106b

levels may rescue Grn+/- phenotypes, but this only normalized the activity of a single

lysosomal enzyme (β-glucuronidase) (216). The current consensus suggests that

45

cortical neurons are particularly susceptible to excess or limited TMEM106B in the

context of GRN-deficiency, and that tight spatiotemporal regulation of TMEM106B in

neurons is required for neuronal lysosomal function.

As mentioned above, neuronal lysosomal dysfunction has been shown to develop in

aged Grn-/- mice further along the pathological progression (119). It has been

speculated that neurons may compensate for early decreased microglial autophagic

capacity; although, it is also possible that microglia-derived secreted neuroinflammatory

factors exacerbate this neuronal phenotype (120) or that neuronal autophagy is more

resilient to the effects of GRN haploinsufficiency. Ultimately, disrupted neuronal

proteolysis contributes to the major neuropathological hallmark of ubiquitin- and TDP-

43-positive NCI and DN subtype A pathology in neocortical layer II of the frontal and

temporal lobes of FTLD-GRN patients (90,91). As previously discussed in section 1.1.7,

the neuropathological hot spots in FTLD-GRN patient brains exhibit greater GRN

mRNA/PGRN levels than would be expected for a haploinsufficient condition (94). This

is likely a consequence of severe microgliosis in the affected regions, given that

activated microglia are known to upregulate GRN expression in response to

pathological insults (137,217). However, it appears that the levels of upregulated

microglial-derived GRN expression are insufficient to result in any meaningful impact on

FTLD-GRN neurodegeneration.

In aged Grn-/- mice, delivery of AAV-Grn expression vectors to the medial prefrontal

cortex (mPFC) increased neuronal Pgrn levels to 10X more than Grn+/+ baseline mPFC

46

expression levels (218). Despite only inducing a minor increase in neuronal Grn/Pgrn

expression in the epicenter of Grn-/- pathology (thalamic VPM/VPL), exogenous delivery

of Pgrn was capable of reducing thalamic microgliosis and lysosomal dysfunction 8-10

weeks post-viral injection (218). Importantly, this finding demonstrates that increasing

neuronal Pgrn expression, even in the latter stages of disease progression, can disrupt

neuropathology associated with Grn-deficiency. Given the multi-decade progression of

FTLD-GRN neuropathology, it is possible that increased microglial GRN expression

derived from the normal allele occurs too late in the disease process to provide

therapeutic benefit or that the microglia-derived PGRN fails to reach the threshold

concentration required for clinical improvement. It is currently unknown why particular

brain regions are sparred, and others are vulnerable to the neurodegenerative effects of

GRN haploinsufficiency, although neuroimaging studies have shed light on the earliest

functional and metabolic brain changes observed in the presymptomatic stages of

FTLD-GRN, which generally precedes grey/white matter atrophy (219). The pattern of

disrupted glucose metabolism and altered network activity expands topographically from

anterior to posterior as the disease progresses. The most consistent abnormalities are

initially observed in medial/dorsal prefrontal and fronto-insular regions mapping onto the

salience network, which plays key roles in social-emotional processing (219). Further

exploration of the complex immunometabolic regulatory interactions between microglia

and neurons/astrocytes may provide insights into the neuropathological basis for the

selective vulnerability observed in these brain regions (210).

47

1.2.7 Therapeutic development for the treatment of FTLD-GRN

There are currently no disease-modifying therapies that can slow down or reverse the

progression of FTLD-GRN. The therapeutic efforts explored to date have aimed to

increase GRN expression from the intact allele (135,201,205,220), increase PGRN’s

extracellular half-life (204), and restore PGRN through exogenous delivery (218,221).

Thus far, the proposed therapies have attempted to upregulate GRN/PGRN expression

levels by alkalizing lysosomes to decrease lysosomal turnover of PGRN (222), calcium

channel blockade (223), promoting the acetylation state of the GRN allele by modulating

epigenetic regulation through deacetylase inhibition (201,205,220), disrupting the

SORT1-PGRN axis to prevent the endocytosis and extend extracellular bioavailability of

PGRN (204), activating autophagy in an mTOR- and TFEB independent manner (135),

or directly delivering GRN to the CNS (218,221). Though several of these mechanisms

have been tested clinically, only the anti-SORT1 antibody AL001 (Alector, INFRONT-3)

and AAV GRN gene therapies PR006 (Prevail Therapeutics, PROCLAIM) and PBFT02

(Passage Bio Inc., upliFT-D) have currently demonstrated clinical potential (224-226).

Recent phase I/IIb clinical trial results presented at the 2020 Alzheimer’s Association

International Conference showed that short- (single dose 60 mg/kb, 12 days,

asymptomatic) and extended-treatment (3X 30 mg/kg, 4 weeks, symptomatic) of FTLD-

GRN patients with the Alector’s AL001 human monoclonal anti-SORT1 antibody was

well tolerated and significantly restored plasma and CSF PGRN levels to normal levels.

PGRN CSF levels in 8 symptomatic FTLD-GRN patients remained in the normal range

8 weeks after receiving the final of the three doses, and extended treatment rescued

48

several pathological CSF biomarkers of lysosomal dysfunction (CTSB),

neuroinflammation (osteopontin), and gliosis (chitotriosidase). These findings have

initiated the development of a phase III clinical study (INFRONT-3) to assess the impact

of AL001 on FTLD-GRN pathological progression. Symptom management is currently

the only clinical option available to FTLD-GRN patients. Some patients expressing the

behavioural variant of FTLD-GRN have been shown to respond positively to serotonin

reuptake inhibitors and antipsychotic medications. For example, bvFTD symptoms such

as irritability, agitation, and depression are improved by treatment with the serotonergic

compound trazodone in a double-blind placebo-controlled crossover trial (227). There is

a clear need for more research and preclinical development into novel therapies

capable of treating or preventing FTLD-GRN.

1.3 Premature termination codon readthrough

Nonsense mutations change an amino acid codon to a termination codon (UAG, UAA

(228), or UGA (229)), resulting in the production of a truncated protein and mRNA

destabilization (230). PTC readthrough is a process that suppresses nonsense

mutations through the pairing of a near-cognate aminoacyl-tRNA at a PTC, allowing for

the incorporation of an amino acid instead of termination, leading to translation of the

full-length protein and increased nonsense mutant mRNA stability (Figure 1-2) (230-

233). Under normal conditions, PTC readthrough occurs at a rate of ~1% of translation

events (234,235), while suppression of natural stop codons occurs at a frequency of ≤

0.1% (236,237). The rationale for the observed discrepancy between these rates is two-

fold; the surrounding mRNA sequence context has been under evolutionary pressure to

49

select for efficient termination at natural stop codons (238-240), and the stabilization

effect imparted by the proximity of stop codons to the poly(A) binding protein bound to

the 3’ poly(A) tail which enables interaction with release factor (RF) eRF3 to enhance

termination of translation (241,242).

Over the past several decades, significant efforts have been made to identify

compounds that are capable of increasing the rate of PTC readthrough as a means to

restore expression from nonsense mutant alleles back to normal levels for therapeutic

gain. To date, many PTC readthrough-inducing drugs has been identified, developed,

and tested clinically; the most common include aminoglycosides (natural (243,244) and

synthetic derivatives (245-247)), negamycin analogs (248,249), and non-

aminoglycoside compounds such as ataluren (formally known as PTC124) (250,251).

Still, common concerns often raised against research into PTC readthrough as a

disease-modifying therapy include the potential for negative impact on global translation

efficiency and whether these efforts would increase suppression of normal stop codons

at the end of ORFs. Failing to terminate translation at stop codons would significantly

increase misfolded proteins due to C-terminal extensions, inevitably inducing the

unfolded protein response. Such a response would be expected to cause upregulation

of the inducible form of heat shock protein 70 (Hsp70) (252), however, treating

fibroblasts with a high concentration of readthrough drugs only slightly increased Hsp70

expression (253) and had no impact on overall translation rates (254,255). On the

contrary, results from a recent global ribosomal profiling study treating HEK293 cells

with high concentrations (0.5 mg/mL) of aminoglycosides, observed increased ribosome

50

density within 3’UTRs in samples treated with G418 and paromomycin. This finding is

indicative of readthrough of natural stop codons, highlighting the need to identify

aminoglycoside readthrough enhancer compounds that may lower the concentration of

aminoglycosides required to induce PTC readthrough (256).

According to a recent analysis of the Orphanet database (257), 3.5-5.9% of the human

population is affected by rare genetic diseases, and of that, ~11% of all genetic

disorders are caused by single base substitutions leading to the introduction of an in-

frame PTC (258); thus, the potentially addressable patient population for PTC

readthrough treatment is substantial. In the past 20 years, there have been numerous

clinical trials conducted testing candidate PTC readthrough drugs in nonsense mutation

bearing Duchenne muscular dystrophy (DMD) (244,251,259) and cystic fibrosis (CF)

(260-263) patients. Though these trials have resulted in the approval of ataluren (known

commercially as Translarna™) to treat children with DMD by the European Medicines

Agency (264), the drug only provides marginal clinical benefit to a subset of nonsense

mutation carriers. The success of previous clinical studies testing PTC readthrough

drugs has been limited by low readthrough efficiencies (0.5-20% increase) failing to

reach the therapeutic threshold required for meaningful disease-modification (265) and

significant toxicity associated with the long-term treatment required to delay and/or

prevent further disease progression (266-268). There is growing interest in applying a

readthrough-based approach to neurodegenerative diseases that possess a significant

proportion of nonsense mutation bearing patients, and in some cases exhibit a lower

therapeutic threshold.

51

Figure 1-2 Nonsense mutation readthrough. A, Translation termination at a natural

mRNA stop codon. B, Premature translation termination at a nonsense codon. C, Drug-

induced PTC readthrough resulting in full-length translation of nonsense mutant mRNA.

52

1.3.1 Aminoglycoside PTC readthrough

Aminoglycosides are bactericidal antibiotics consisting of two to three amino sugar

groups joined to either a streptidine or 2-deoxystreptamine (2-DOS) ring by glycosidic

linkages and are potent inhibitors of protein synthesis at high doses and reduce

translational fidelity at low doses in Gram-negative bacteria (269,270). Aminoglycoside-

mediated PTC readthrough has been thoroughly demonstrated in several mammalian

cell lines and in vivo models expressing disease-associated nonsense mutations

(reviewed here (271-273)). In 2-DOS aminoglycosides, this is achieved through binding

to the 30S ribosomal subunit’s 16S rRNA decoding center, which is responsible for

proof-reading codon-anticodon base-pairing complementarity (274). The eukaryotic 18S

rRNA decoding center sequence differs from prokaryotic 16S rRNA at two key

nucleotides, weakening the binding affinity of 2-DOS aminoglycosides by 25-50X

(275,276). Therefore, eukaryotic ribosomes are less susceptible to translational

inhibition by aminoglycosides at low doses, supporting their clinical use as therapeutic

antibiotics targeting bacterial infections. The weaker affinity of aminoglycosides to the

eukaryotic ribosomal decoding center is fundamental to their propensity to suppress

nonsense mutations while minimizing the misincorporation of amino acids at sense

codons (277).

The differential biochemical interactions between the eukaryotic ribosomal A-site and

aminoacyl tRNAs / translational RFs in the presence and absence of aminoglycosides

have been thoroughly investigated (reviewed here (278,279)). Translational fidelity is

monitored by three universally conserved ribosomal rRNA nucleotides in helix 18

53

(G530) and 44 (A1492, A1493). For simplicity, rRNA residues will be numbered

according to the numbering used in E. coli 16S rRNA. As mentioned above, the

eukaryotic rRNA decoding center also contains non-conserved nucleotides; A1408G

and G1491A base substitutions in helix 44 are responsible for the diminished

aminoglycoside binding affinity (274,276). Upon forming a cognate codon-anticodon

pair, A1492 and A1493 are flipped out into the minor groove of the codon-anticodon

mini-helix core, and G530 transitions from a syn to anti conformation (280,281). This

conformational change, often referred to as the ON-state, brings these proof-reading

rRNA nucleotides in proximity to the codon-anticodon base-pairing event, allowing

hydrogen bonding via 2’ OH groups to assess complementarity.

In contrast, natural stop codons and PTCs are instead recognized through a nucleotide-

protein interaction with RFs, which trigger translational termination. The eukaryotic RF

(eRF1) is capable of recognizing all three stop codons (282), while prokaryotes possess

RF1 and RF2, which recognize UAG/UAA and UGA/UAA, respectively (283). These

RFs share 3D structural homology with aminoacyl tRNAs and can directly recognize the

three stop codons; however, there are subtle differences in how they interact with the

ribosomal A-site. Unlike sense codons, only one of the A1492 or A1493 position

nucleotides is flipped out of helix 44 (A1492 in prokaryotes vs. A1493 position in

eukaryotes) in response to RF stop codon/PTC recognition, leaving the remaining

adenine free to form a stacking interaction with A1913 of helix 69 on 23S rRNA which

signals the RF to activate peptidyl transferase activity (284,285). The ribosomal A-site

configuration where either one or both of the conserved adenines are directed towards

54

the helical core is known as the non-decoding or OFF-state, which has also been

associated with the absence of a cognate tRNA (286,287).

Aminoglycoside binding to the 16S rRNA decoding center is stabilized by hydrogen

bonds with G1408 and displaces both A1492 and A1493 from the helical core, similar to

the ON-state induced by cognate tRNAs, which may introduce steric clash and disrupt

potential base stacking interactions between A1493 and A1913, thus limiting the access

of RFs to stop codons/PTCs (288,289). Furthermore, aminoglycoside-mediated

displacement of A1493 has been shown to strongly shift its phosphate group, which

relaxes the decoding pocket on the codon side (290). These conformation changes

simultaneously increase the likelihood of near-cognate tRNA sampling in the ribosomal

A-site and the accommodation of a single-nucleotide anticodon mismatch at PTCs,

stimulating readthrough.

1.3.2 Aminoglycoside toxicity

To prevent or delay the progression of any nonsense mutation-linked diseases, patients

will likely require lifelong treatment with PTC readthrough-inducing drugs to maintain

adequate expression from their nonsense alleles. Therefore, long-term tolerability with

minimal toxicity and low rates of drug-associated adverse events are essential for the

success of nonsense suppression therapy as a treatment for genetic disorders.

Therefore, a major impediment to the clinical application of aminoglycosides as

translational readthrough-inducing drugs has been the association of their prolonged

use with significant nephro- and ototoxicity (reviewed here (267,291,292)).

55

Aminoglycoside-induced nephrotoxicity is a reversible process that results in both the

necrosis and apoptosis of renal proximal tubular epithelial cells (293,294). In contrast,

ototoxicity induced by aminoglycosides results in permanent bilateral sensorineural

hearing loss by triggering apoptosis of vestibular and cochlear hair cells (295,296).

Three main mechanisms contribute to aminoglycoside-induced toxicity: lysosomal

phospholipidosis, accumulation of intracellular reactive oxygen species (ROS), and

inhibition of mitochondrial protein synthesis.

Aminoglycoside antibiotics gain entry into cells through the endocytic membrane

receptor, megalin, which is highly expressed on the apical membrane of epithelial cells

in the renal proximal tubule (297) and stria vascularis marginal cells of the cochlear duct

(298,299). Following endocytosis, most aminoglycosides enter the endolysosomal

network, those destined for the lysosome become positively charged (300), allowing

them to bind to negatively charged lysosomal membrane phospholipids. This affinity for

the lysosomal membrane disrupts phospholipid turnover and inhibits phospholipase

function, resulting in lysosomal phospholipidosis (266,301). Eventually, membrane

damage becomes so severe that lysosomal membranes rupture (302), releasing

aminoglycosides into the cytosol, causing the formation of ROS and subsequent

oxidative damage (303,304) via iron-aminoglycoside complexes (305), and

aminoglycoside-mediated mitochondrial dysfunction (306,307). Mitochondrial ribosomal

(mitoribosomal) rRNA shares significant homology with prokaryotic rRNA; given that the

mitoribosome is a remnant from a previously symbiotic bacteria, the mitoribosomal

decoding site is more susceptible to aminoglycoside-mediated inhibition of protein

56

synthesis than the eukaryotic ribosome. Genetic analysis has revealed that individuals

carrying specific SNPs in their mitochondrial DNA encoding the helix 44 decoding

center that further enhance aminoglycoside-mitoribosomal affinity are at greater risk of

aminoglycoside-induced ototoxicity (268,308,309). Molecularly characterizing the

mechanisms of both PTC readthrough and aminoglycoside-induced toxicity have

provided researchers with a rational basis for designing novel synthetic

aminoglycosides with improved readthrough activity while minimizing toxicity.

1.3.3 Promising PTC readthrough drugs

As discussed in section 1.3.2, there are several known clinical limitations to using

aminoglycosides as potential PTC readthrough drugs (266-268). While natural

aminoglycosides are currently not suitable for long-term human administration, efforts

are underway to develop novel aminoglycoside derivatives with improved tolerability

and PTC readthrough potency. Medicinal chemistry approaches for designing synthetic

aminoglycosides have been largely focused on limiting their affinity for the prokaryotic

and mitochondrial ribosomal decoding center while maintaining a weak affinity for

eukaryotic ribosomes as a means to reduce toxicity. The aminoglycoside G418

(geneticin) is simultaneously the strongest readthrough inducer and most toxic of all

natural aminoglycosides tested to date (234,310,311). Over the past 15 years, Baasov

and co-workers have systematically and elegantly demonstrated which aminoglycoside

chemical moieties of G418 are essential for its PTC readthrough activity and those

responsible for its toxicity that may be expendable (312-316). In summary, these

exceptional molecular remodeling studies confirmed that ring I is essential for G418-

57

induced PTC readthrough activity, and its garosamine ring III is largely responsible for

its significant toxicity.

Their 4th generation aminoglycoside derivative NB124 only differed from G418 by

modifications to ring III, swapping the garosamine ring to a 5 amino ribose with an (S)-

5’’-methly group and its ring II attachment position (C6 instead of C5). Compared to

G418, these ring III alterations in NB124 dramatically reduced its affinity for both the

mitochondrial and prokaryotic ribosomes (> 100X) while maintaining similar PTC

readthrough efficiency (316). Following its initial synthesis, NB124 was acquired by

Eloxx Pharmaceuticals and became one of their lead investigational therapeutic

candidates (ELX-02). ELX-02 is a novel eukaryotic ribosomal selective glycoside that

exhibits potent PTC suppression activity in several nonsense mutant disease models,

comparable to G418 (317,318). Recent results of a phase I clinical trial have

established that ELX-02 is safe and tolerated in healthy human subjects (246), thus

prompting ongoing phase II clinical trials in CF and cystinosis patients harboring

nonsense mutations (319-321). Despite the early success of ELX-02, Baasov et al.

have continued their efforts to further structurally optimize aminoglycoside derivatives,

applying a rational design strategy to maximize readthrough efficacy while minimizing

nephro- and ototoxicity (317,322).

In parallel, significant research efforts have been made to identify PTC readthrough-

inducing drugs of non-aminoglycoside antibiotic origin. The poor long-term tolerability of

the aminoglycoside gentamicin highlighted the need for novel nonsense suppressors;

58

this search culminated in a high-throughput PTC readthrough screen of over 800,000

compounds identifying the small molecule oxadiazole, ataluren, as the most potent

stimulator of a firefly luciferase-based PTC readthrough reporter assay (250). However,

shortly following its discovery, evidence emerged demonstrating that ataluren binds to

and stabilizes the firefly luciferase enzyme, which suggested that the ataluren ‘hit’ may

have been the result of a PTC readthrough-independent mechanism (323,324). Several

subsequent PTC readthrough studies failed to observe ataluren-induced PTC

readthrough activity in various genetic contexts (325-327). On the other hand, multiple

preclinical studies have confirmed ataluren-mediated PTC readthrough activity and

observed lower toxicity levels compared to aminoglycosides (250,328-330). An

additional mechanistic study unambiguously demonstrated that ataluren promotes the

insertion of near cognate tRNAs at PTCs, similar to aminoglycosides (331). In further

support of this interaction, co-treatment of both nonsense mutant cell cultures and CF

patients with tobramycin, an aminoglycoside with a strong affinity for the eukaryotic

ribosomal A site (332), has been discovered to competitively inhibit ataluren-induced

readthrough (263,331). Confidence in ataluren’s therapeutic potential as a PTC

readthrough drug ultimately resulted in multiple phase III clinical trials treating nonsense

mutation carrying DMD and CF patients with ataluren but failed to significantly alter

disease progression compared to placebo groups (251,263,333). Nevertheless, these

varied results have led to the development of promising novel oxadiazole containing

ataluren analogs with improved readthrough performance, currently under preclinical

development (334-336).

59

1.3.4 PTC readthrough enhancers

Our collaborators in the Roberge lab previously identified several readthrough enhancer

compounds capable of synergistically increasing the potency of aminoglycoside-

mediated suppression of nonsense mutations. In 2016, they discovered a novel class of

aminoglycoside PTC readthrough enhancer compounds (CDX-series) using a high-

throughput screen for nonsense suppression in yeast (337). The CDX-series

compounds are N-substituted phthalimide derivatives that they further systematically

functionalized to increase their potency. They demonstrated that co-treatment with

CDX5-1 and G418 enhanced PTC readthrough in TP53 nonsense mutant HDQ-P1

cancer cells. At present, the mechanism underpinning the readthrough enhancement

achieved by CDX compounds is unclear. However, qualitative observations suggest that

combination treatment with G418 and CDX compounds inhibit proliferation in cell

cultures; therefore, toxicity may limit clinical potential. In a subsequent study, the

Roberge group also identified that the antimalaria drug mefloquine could potentiate

aminoglycoside-induced PTC readthrough in several TP53 nonsense mutant cancer cell

lines (338). Importantly, readthrough-derived full-length p53 retained biological function

as observed through increased Ser15 phosphorylation in response to ionizing radiation.

They demonstrated that mefloquine-induced readthrough enhancement is not achieved

through increased extracellular uptake of G418 or increased lysosomal permeability to

G418, suggesting that mefloquine may directly target translation machinery.

60

1.3.5 Nonsense mediated mRNA decay

The abundance of nonsense mutant mRNA transcripts is a determining factor of PTC

readthrough efficiency. mRNA stability is regulated by several biological processes,

including microRNA silencing, mRNA maturation, and nonsense-mediated mRNA decay

(NMD). mRNAs bearing nonsense mutations are targeted for degradation by NMD, and

PTC readthrough enables escape of nonsense mutant mRNA from degradation by

NMD. NMD efficiency can vary up to 4-fold in the general population, likely resulting in a

wide variance in basal PTC readthrough levels in individual nonsense mutation carriers

(339). For this discussion, only the exon-junction complex (EJC)-dependent model of

mammalian NMD will be presented, as EJC-independent NMD is not triggered by the

presence of a PTC (reviewed here (233,340,341)). For a PTC containing mRNA to

elude destruction by NMD, it must be translated through to at least 50-55 nucleotides

upstream of the final exon-exon junction during its ‘pioneer’ round of translation. This

degree of translation is sufficient because it removes al EJCs that are deposited 20-24

nucleotides upstream of every exon-exon junction during pre-mRNA splicing. During the

translation of a nonsense mutant mRNA transcript, the NMD factors UPF1 and SMG1

kinase interact with eRF1 and eRF3 upon PTC recognition forming a pretermination

complex. When a ribosome encounters a PTC with an intact downstream EJC (contains

NMD factors UPF2 and UPF3), UPF1 can interact with UPF2/UPF3, which induces

UPF1 phosphorylation by SMG1, thus marking the mRNA for degradation. This

phosphorylation recruits additional NMD factors (SMG5, SMG6, SMG7, PP2A), forming

an mRNA surveillance complex that facilitates the breakdown of the PTC-containing

61

mRNA. Concurrently, the PP2A phosphatase dephosphorylates UPF1, thus enabling

the detection of additional nonsense mutant mRNAs.

The elucidation of the NMD pathway and understanding its role in suppressing the

levels of nonsense mutant mRNAs has identified potential drug targets to inhibit NMD

as an adjunct to PTC readthrough agents. Because the phosphorylation and

subsequent dephosphorylation of UPF1 are the main rate-limiting steps driving NMD,

interfering with these processes leads to efficient NMD inhibition. Thus far, several

compounds have been identified that inhibit NMD by targeting the phosphorylation

status of UPF1. For example, the natural substances caffeine and wortmannin both

inhibit the SMG1 kinase, preventing the phosphorylation of UPF1, which rescues

nonsense mutant mRNA from being tagged for degradation (342). Another group

performed a virtual screen, capitalizing on the crystallographic resolution of the UPF1-

SMG7 complex to identify commercially available compounds capable of binding to the

SMG7 pocket critical for association with UPF1. This screen led to the identification of

NMDI-14, which prevents phosphorylated UPF1 from recruiting the SMG5-SMG7

heterodimer, which together serves as an adapter for phosphatase PP2A-mediated

UPF1 dephosphorylation (343).

Another compound, NMDI-1, inhibits NMD through a similar mechanism by disrupting

the association between SMG5 and UPF1, which leads to the sequestration of

hyperphosphorylated UPF1 (344); notably, NMDI-1 is a 1500X more potent NMD

inhibitor than caffeine. In vivo co-administration of the NMD attenuator, NMDI-1, and

62

gentamicin in the IduaW392X/W392X mucopolysaccharidosis type I mouse model

synergistically enhanced Idua enzyme activity and rescued glycosaminoglycan

accumulation and lysosomal dysfunction (345). Despite the synergistic benefits of

combining NMD inhibition and PTC readthrough treatment, it is important to consider

the potential negative consequences of inhibiting NMD factors. For example, NMD

factors have also been implicated in the transcriptomic regulation of ~10% of all

mammalian mRNA (non-PTC bearing) transcripts (265), playing important roles in B

and T cell maturation (346), as well as telomere maintenance (347). Therefore, efforts

exploring the clinical application of NMD inhibition should proceed with caution,

although, considering that a 4-fold variation in NMD efficiency has been observed in the

general population, it is likely that some degree of NMD disruption may be tolerated.

1.3.6 PTC readthrough in cellular models of FTLD-GRN

Of the 79 known GRN mutations, 18 (23%) are single-base substitutions introducing

nonsense mutations, and notably, the most common familial GRN mutation is the

p.R493X nonsense mutation. These disease characteristics make FTLD-GRN an ideal

candidate for PTC readthrough drug development. Following our abstract presented at

the 11th International Conference in Frontotemporal Dementias in 2018, Kuang et al.

corroborated parts of our findings, showing that aminoglycosides induced GRN

nonsense mutation readthrough in cells transfected with nonsense mutant expression

GRN constructs (348). They transfected N2a cells with C-terminally FLAG-tagged GRN

nonsense mutant expression constructs (p.Q125X, p.Y229X, p.R493X) and confirmed

gentamicin- and G418-mediated PTC readthrough of UGA nonsense codon (p.R493X),

63

while UAG/UAA codons (p.Q125X/p.Y229X) were unresponsive (348). Analysis of the

subcellular localization of G418-induced PTC readthrough-derived full-length PGRN

found co-localization with the lysosomal and Golgi apparatus similar to wild-type PGRN;

thus, suggesting the readthrough product may function similarly to wild-type PGRN

(348). Moreover, they reported evidence demonstrating escape of FLAG-tagged GRN

R493X mRNA from NMD, resulting in a 2X increase in mRNA levels following a 24 h

treatment with G418 (348). Two months later, we published our manuscript

demonstrating G418-mediated PTC readthrough in HEK293 cells transfected with

similar C-terminally tagged nonsense mutant GRN expression constructs, further

improving upon these efforts by co-treating with novel readthrough enhancer drugs. The

results from our studies demonstrating GRN PTC readthrough will be presented in detail

in the subsequent chapters.

1.4 Thesis research questions

The primary objectives of this thesis were to answer the following questions: 1) can we

demonstrate in vitro and proof-of-concept in vivo GRN PTC readthrough of exogeneous

nonsense mutant GRN expression constructs? 2) do FTLD-GRN and CLN11 hiPSC

models of GRN-deficiency caused by nonsense mutations exhibit mutant lysosomal

phenotypes? 3) can we demonstrate proof-of-concept PTC readthrough of endogenous

pathogenic GRN mutations in both in vitro and in vivo models, and reverse disease-

related phenotypes? 4) What behavioural and neuropathological phenotypes develop in

the recently developed GrnR494X/R493X knock-in mouse model? A graphical overview and

64

summary of our experimental approach is described below (Figure 1-3).

65

66

Figure 1-3 Graphical overview of experimental models. Exogenous PTC

readthrough was demonstrated in HEK293 cells transfected with several C-terminally

V5-tagged GRN expression constructs, each bearing a causal FTLD-GRN pathological

mutation. Stably transfected HEK293 lines provided an ideal model for screening PTC

readthrough CDX-series enhancer compounds treated in combination with G418. The

coding sequence of the R493X GRN-V5 expression construct was cloned into an AAV

expression vector, which was used to generate AAV-GRN-R493X-V5 mice. Top

readthrough enhancer compound CDX5-288 identified in HEK293 screening was used

in combination with G418 to induce in vivo CNS PTC readthrough of the V5-tagged

GRN R493X construct. To demonstrate PTC readthrough of endogenously expressed

GRN, we established novel FTLD-GRN patient- and isogenic CRISPR/Cas9 gene

editing-derived hiPSC models of GRN LOF caused by nonsense mutations. GRN

nonsense mutant hiPSC-derived neurons and astrocytes were treated with top

candidate PTC readthrough drug combinations and assessed for rescued PGRN

expression. We also obtained and similarly treated aged GrnR493X/R493X mice with G418

to further establish proof-of-concept in vivo readthrough; however, high levels of toxicity

prevented us from achieving long-term delivery of therapeutic dosages. Therefore, we

limited the scope to a novel phenotypic characterization of this mouse model, assessing

several neuropathological phenotypes previously reported in Grn-/- mice.

67

Chapter 2: Materials and methods

2.1 Mice

C57BL/6J mice for AAV studies were acquired from Dr. Ann Marie Craig (Department of

Psychiatry, Centre for Brain Health, UBC). GrnR493X/R493X mice on a C57BL/6J

background, and C57BL/6J Grn+/+ mice were obtained from Jackson Laboratories

(Stock# 029919 and #000664, respectively). Animal procedures were approved by the

UBC Animal Care and Biosafety Committees. Mice were housed at the UBC Centre for

Disease Modelling and Animal Research Unit facilities, where work outlined in

institutionally approved breeding and experimental protocols (A19-0623, A19-0062,

A16-0161, and A17-0225) was conducted. Mice were provided with a 12 h light/12 h

dark cycle and allowed food and water ad libitum. All efforts were made to minimize

suffering with minimally invasive procedures, and where necessary, isoflurane

anesthesia was used.

2.2 Cells

2.2.1 HEK293

HEK293 cell studies were conducted in Dr. Roberge’s lab (Life Science Center, UBC)

and cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS,

Thermo Fisher Scientific) + 1% Pen/Strep (Thermo Fisher Scientific).

68

2.2.2 Mouse embryonic fibroblasts

Grn+/+ and GrnR493X/R493X mouse embryonic fibroblasts (MEFs) were isolated from E14.5

embryos. Briefly, dissected embryos were finely minced and digested in 0.25% Trypsin-

EDTA (Thermo Fisher Scientific) at 37 °C for 10 min. This solution was centrifuged at

100 x g for 3.5 min, and the pellet was resuspended in 0.125% Trypsin-EDTA for further

digestion at 37 °C for 25 min. Following the addition of FBS to neutralize the trypsin, the

solution was vigorously pipetted with FBS and passed through a cell strainer for

collection of a single cell suspension. This resulting suspension was centrifuged at 300

x g for 5 min, and the supernatant was discarded and replaced with DMEM-high

glucose medium (Thermo Fisher Scientific) supplemented with 15% FBS + 1%

Pen/Strep. The cells were counted and plated at a density of 1.5E7 cells per 150 mm

plate and passaged once confluent. Early passage frozen stocks were prepared and

thawed as needed for experiments.

2.3 Expression vectors

2.3.1 Vector mutagenesis

Using GeneArt mutagenesis service (Thermo Fisher Scientific), the coding sequence of

GRN was synthesized and cloned into pDONR221 Entry vector. Three base

substitutions c.347C>A, c.1252C>T, and c.1477C>T were engineered into GRN using a

targeted PCR-based strategy to generate p.S116X (UAA), p.R418X (UGA) and

p.R493X (UGA) nonsense mutations. Finally, to generate expression clones, LR

recombination reaction was used to recombine the wild-type and mutated samples from

the Entry vectors into the pcDNA-6.2/V5-DEST vector (Thermo Fisher Scientific). These

69

C-terminal V5 tagged GRN expression constructs were used for HEK293 cell

transfections.

2.3.2 Transfection

HEK293 cells were transiently transfected with pcDNA6.2/V5-DEST vector expressing

C-terminally V5-tagged mutated GRN with nonsense mutations (p.S116X, p.R418X,

p.R493X) using Lipofectamine 2000 (Thermo Fisher Scientific). 24 h after transfection,

each sample was split into two wells and either left untreated or treated with G418 (100

µg/mL). After 72 h, cells were lysed and subjected to automated capillary

electrophoresis western analysis (Wes, ProteinSimple). To generate stable cell lines,

HEK293 cells were transfected with pcDNA-6.2/V5-DEST vectors expressing mutant

GRN as described above and subjected to blasticidin (Thermo Fisher Scientific)

selection. Individual clones that were resistant to 15 μg/mL blasticidin were selected.

2.3.3 AAV vector packaging

The packaging of AAV-GRN-R493X-V5 was performed by the Petrucelli group at the

Mayo Clinic in Jacksonville, Florida. The coding sequence for V5-tagged p.R493X

mutant GRN was cloned into the AAV expression vector pAM/CBA-EGFP-WPRE-BGH,

AAV particles were packaged into serotype 9 capsid (AAV9), and purified using

standard methods (349). Briefly, AAV was generated by co-transfection with the cis

plasmids pF Delta6 and pRepCap9 into HEK293T cells. Cells were harvested 72 h after

transfection, treated with 50 Units/ml Benzonase (Sigma-Aldrich), and lysed by freeze-

thaw. The virus (AAV-GRN-R493X-V5) was then purified using a discontinuous

70

iodixanol gradient and buffer exchanged to PBS using an Amicon Ultra 100

Centrifugation device (EMD Millipore). The genomic titer of each virus was determined

by quantitative PCR using the ABI 7700 (Applied Biosystems) and primers specific to

the WPRE. The viral DNA samples were prepared for quantification by treating the

virus with DNaseI (Invitrogen) and Proteinase K (Invitrogen), and samples were

compared against a standard curve of supercoiled plasmid. AAV9-eGFP-Cre viral

particles (AddGene, 105540-AAV9) were used to demonstrate technical proficiency of

P0 intracerebroventricular (ICV) AAV injection by assessing brain-wide human synapsin

promoter-driven eGFP expression in 10 week old AAV9-eGFP-Cre mice.

2.3.4 In vitro transduction

HEK293 cells were transduced for 5 days with 2 mL volumes of media containing

increasing concentrations of AAV-GRN-R493X-V5 expression vector (8.8E8, 8.8E9, and

8.8E10 PFU/mL) and 8 µg/mL polybrene for increased transduction efficiency. 48 h after

transduction, each transduced well was treated with G418 (100 µg/mL). Following 72 h

G418 treatment, supernatant was collected and cells were lysed. Cell lysates and

supernatant were both subjected to automated capillary electrophoresis western

analysis.

71

2.4 Antibodies

Table 1-1 Primary antibodies

Antibody Dilution Source Identifier

Polyclonal Rabbit Anti-Human ACTIN Antibody

1:10000 (WB / Wes)

Novus Biologicals NB600-532, RRID: AB_10002039

Monoclonal Rabbit Anti-Mouse C1q Antibody

1:500 (IF) Abcam ab182451, RRID: AB_2732849

Polyclonal Goat Anti-Human CTSD Antibody

1:1000 (WB) R&D Systems AF1014, RRID: AB_2087218

Polyclonal Goat Anti-Mouse Ctsd Antibody

1:500 (WB) R&D Systems AF1029, RRID: AB_2087094

Polyclonal Rabbit Anti-Human DESMIN Antibody

1:500 (IF) Invitrogen PA5-16705, RRID: AB_10977258

Polyclonal Goat Anti-Mouse DppII Antibody

1:500 (IF) R&D Systems AF3436, RRID: AB_2093882

Polyclonal Rabbit Anti-Human FOXG1 Antibody

1:500 (IF) Abcam ab18259, RRID: AB_732415

Polyclonal Rabbit Anti-Mouse Foxp2 Antibody

1:500 (IF) Abcam ab16046, RRID: AB_2107107

Polyclonal Rabbit Anti-Human GAD65/67 Antibody

1:500 (IF) Sigma-Aldrich G5163, RRID: AB_477019

Polyclonal Rabbit Anti-Human GFAP Antibody

1:500 (IF) STEMCELL Technologies

60128, RRID: AB_215513

Monoclonal Rat Anti-Mouse Lamp1 Antibody

1:500 (IF/WB)

BD Biosciences 553792, RRID: AB_2134499

Polyclonal Rabbit Anti-Mouse LC3-I/II Antibody

1:1000 (WB) Cell Signaling Technology

2775, RRID: AB_915950

Polycloncal Rabbit Anti-Human MAP2 Antibody

1:500 (IF) Proteintech 17490-1-AP, RRID: AB_2137880

72

Antibody Dilution Source Identifier

Polycloncal Goat Anti-Human NANOG Antibody

1:1000 (IF) R&D Systems AF1997, RRID: AB_355097

Monoclonal Mouse Anti-Human OCT3/4 Antibody

1:1000 (IF) STEMCELL Technologies

60093, RRID: AB_2801346

Polyclonal Goat Anti-Human PGRN Antibody

1:1000 (IF) R&D Systems AF2420, RRID: AB_2845434

Polyclonal Rabbit Anti-Human PGRN Antibody

1:1000 (WB) Sigma-Aldrich HPA008763, RRID: AB_1850339

Polyclonal Rabbit Anti-Mouse Pgrn Antibody

1:1000 (WB) Yu/Herz lab (82) Mouse Linker 1 Peptide (C)TLLKKF PAQKTNRAVSI

Polyclonal Sheep Anti-Mouse Pgrn Antibody

1:100 (WB) R&D Systems AF2557, RRID: AB_2114504

Monoclonal Rabbit Anti-Human SOX1 Antibody

1:1000 (IF) STEMCELL Technologies

60095, RRID: AB_2801347

Polyclonal Goat Anti-Human SOX17 Antibody

1:500 (IF) R&D Systems AF1924, RRID: AB_355060

Polyclonal Rabbit Anti-Human SYNAPSIN Antibody

1:500 (IF) EMD Millipore 5747777, RRID: AB_212517

Polyclonal Rabbit Anti-Human TBR1 Antibody

1:500 (IF) Abcam ab31940, RRID: AB_2200219

Polyclonal Rabbit Anti-Mouse N-terminal TDP-43 Antibody

1:1000 (WB), 1:500 (IF)

Proteintech 10782-2-AP, RRID: AB_615042

Polyclonal Rabbit Anti-Human phospho-Ser409 TDP-43 Antibody

1:500 (WB) Cosmo Bio USA CAC-TIP-PTD-P03, RRID: AB_1961901

Polyclonal Chicken Anti-Human TUJ1 Antibody

1:500 (IF) Neuromics CH23005, RRID: AB_2210684

73

Antibody Dilution Source Identifier

Monoclonal Mouse Anti-V5 Antibody

1:500 (Wes) Abcam ab27671, RRID: AB_471093

Monoclonal Mouse Anti-Mouse Vgat Antibody

1:300 (IF) Synaptic Systems 131011, RRID: AB_887872

Monoclonal Mouse Anti-Human VGLUT1 Antibody

1:200 (IF) Synaptic Systems 135311, RRID: AB_887880

2.5 Drug treatments

2.5.1 In vitro

G418 sulfate powder was reconstituted in sterile PBS to 50 mg/mL and CDX-series

compounds in dimethyl sulfoxide (DMSO) to 50 mM. These stock solutions were stored

at -20C. Upon addition of CDX compounds to cell culture media, gentle vortexing was

applied for 2 min to ensure complete solubilization. Vehicle solutions were prepared by

adding an equivalent volume of PBS and/or DMSO. Rec. PGRN (Adipogen, AG-40A-

0068Y-C010) was reconstituted to 0.1 mg/mL in sterile water. CTSL inhibitor (Z-Phe-

Phe-FMK, Abcam, ab141386) 10 mM stock in DMSO was first diluted to 100 and 300

µM, prior to 1:10 dilution in cell culture medium.

2.5.2 AAV mouse model

In vivo treatment solutions of vehicle (1.6% solutol), 16.6 mg/mL G418 (1.6% solutol),

0.5 mM CDX5-288 (1.6% solutol), and 16.6 mg/mL G418 + 0.5 mM CDX5-288 (1.6%

solutol) were prepared in sterile saline. A 50 mM solution of CDX5-288 in DMSO was

first diluted to 3 mM in 10% solutol (diluted in sterile saline) and further diluted 1:6 in

74

treatment solutions.

2.5.3 GrnR493X/R493X mouse model

In vivo treatment solutions of either vehicle (saline) and G418 reconstituted in saline (25

and 100 mg/mL) were used to treat Grn+/+ and GrnR493X/R493X mice. A range of G418

concentrations and injection volumes were used to achieve different doses of G418 in

the brain.

2.6 Western blot

2.6.1 Conventional western blot

Cortical neuron, astrocyte, and MEF supernatant were collected, and cell monolayers

were rinsed with 1 mL ice-cold PBS. Cells were lysed in 50 µL lysis buffer (20 mM Tris–

HCl pH 7.5, 150 mM NaCl, 1mM EDTA, 1 mM EGTA, 1% (v/v) Triton X-100, 2.5 mM

sodium pyrophosphate, 1 mM β-glycerophosphate) supplemented with fresh 1 mM

Na3VO4, and 1X complete protease inhibitor cocktail (Roche Molecular Biochemicals).

Lysates and supernatants were pre-cleared by centrifugation at 19,500 x g for 10 min at

4 °C. Neuronal, astrocyte, and MEF lysate protein concentrations were quantitated

using the Bradford assay and diluted in 4X SDS-PAGE loading buffer + 100 mM DTT.

Hemi-brain RIPA-soluble and -insoluble lysates were prepared and quantified according

to section 2.16.2 and were diluted in 4X SDS-PAGE loading buffer + 100 mM DTT. In

brief, 10-30 µg protein from each boiled (5 min at 95 °C) SDS lysate was separated on

4-15% gradient precast polyacrylamide gel (Bio-Rad), electroblotted onto a 0.45 µm

nitrocellulose membrane (Thermo Fisher Scientific) and blocked for 1 hour in 5% (w/v)

75

non-fat milk in TBS + 0.1% (v/v) Tween-20 (TBST). Samples from GrnR493X/R493X and

wildtype control mice (Grn+/+) were run on the same precast gel to allow for direct

comparison. Membranes were incubated with primary antibodies diluted in 5% non-fat

milk TBST blocking buffer overnight at 4 °C, washed 3X 5 min with TBST, incubated

with 1:10,000 HRP-conjugated secondary antibody diluted in 5% non-fat milk TBST

blocking buffer, washed again 3X 5 min with TBST, and incubated with enhanced

chemiluminescence substrate (EMD Millipore). Films were developed, scanned, and

analyzed using the Fiji processing package for ImageJ (National Institute of Health) for

densitometry analysis. To reprobe blots with the same species of primary antibody,

membranes were stripped with 0.1 N NaOH for 5 min, washed twice with deionized

H2O, and incubated for 30 min in 5% non-fat milk TBST blocking buffer before

incubation with an additional primary antibody.

2.6.2 ProteinSimple Wes

Automated capillary electrophoresis western analysis was carried out with

manufacturer’s reagents according to the user manual (Wes, ProteinSimple). HEK293

cell lysates were prepared and quantified following the methods used for

neurons/astrocytes/MEFs (see section 2.6.1), and AAV-GRN-R493X-V5 whole-brain

RIPA-soluble lysates were prepared and quantified as in section 2.16.1. Briefly, 5.6 µL

of 1 mg/mL HEK293 cell or whole-brain lysate was mixed with 1.4 µL fluorescent master

mix and heated at 95 °C for 5 min. The samples, blocking reagent, wash buffer, primary

antibody, secondary antibody, and chemiluminescent substrate were dispensed into the

76

microplate provided by the manufacturer, and the microplate was centrifuged at 1000 x

g for 5 min. The electrophoretic separation and immunodetection were performed

automatically using default settings. The data were analyzed with inbuilt Compass

software (ProteinSimple). The truncated and full-length V5 (PGRN) peak intensities

(area under the curve) were normalized to that of the actin peak, used as a loading

control. In Figure 3-1, Figure 3-2, Figure 3-4, and Figure 3-7, electropherograms are

represented as pseudo-blots, generated from the quantification of chemiluminescence

by the Compass software.

2.7 hiPSC models

2.7.1 hiPSC reprogramming

Erythroid progenitor (EP) cells were isolated from peripheral blood obtained from a

healthy control subject (WT) and a GRN p.R418X mutation carrier (R418X+/-) (350)

using the erythroid progenitor reprogramming kit (STEMCELL Technologies). Expanded

EPs were reprogrammed into hiPSCs with Epi5TM episomal reprogramming kit plasmids

(Invitrogen, Thermo Fisher Scientific) delivered in AmaxaTM Human CD34+ Cell

nucleofection buffer (Lonza) using the AmaxaTM Nucleofector II (Lonza) electroporation

device according to the erythroid progenitor reprogramming kit manufacturer’s

instructions.

2.7.2 hiPSC culture

hiPSCs were cultured in a feeder-independent manner on Matrigel (BD Biosciences)

coated plates and fed daily with mTeSR1 medium (STEMCELL Technologies). Every 4-

77

5 days, 70-80% confluence cultures were passaged as aggregates using ReLeSR

(STEMCELL Technologies) at ~1:5 split ratio. During the first 24 h post-plating,

mTeSR1 medium was supplemented with 10 µM rho-associated protein kinase inhibitor

(Y-27632, EMD Millipore). S116X+/- hiPSCs were obtained from existing stock and

previously described (60).

2.7.3 hiPSC trilineage differentiation

hiPSCs were differentiated into the three germ layers using the StemDiff™ trilineage

differentiation kit (STEMCELL Technologies) according to manufacturer's instructions.

After differentiation, cells were fixed in 4% PFA for 15 min for immunofluorescence

analysis with anti-SOX1 (ectoderm), anti-SOX17 (endoderm), and anti-DESMIN

(mesoderm) antibodies (see section 2.13.1).

2.7.4 hiPSC CNS lineage differentiation

WT and FTLD-GRN mutant hiPSCs were differentiated into neuronal progenitor cells

(NPCs) using the dual SMAD inhibition protocol (351). These NPCs were frozen and

thawed as needed onto Matrigel-coated plates in neural stem cell medium for week-long

expansion. Expanded NPCs were then plated onto poly-L-ornithine / laminin (PLO/L)

coated plates and further differentiated into cortical neurons using complete

BrainPhysTM media system (352) supplemented with 1X CultureOneTM (Gibco, Thermo

Fisher Scientific) for the first two weeks of neuron maturation. cAMP and ascorbic acid

were withdrawn from BrainPhysTM media after 3 weeks (from hiPSC stage, DIV 50) for

continued neuronal maturation until DIV 80 to allow lysosomal phenotypes to develop.

78

Astrocytes were differentiated from expanded cryopreserved NPCs according to the

STEMDiffTM astrocyte differentiation and maturation kit (STEMCELL Technologies)

protocol to produce DIV 60+ frozen stocks to be thawed and passaged as needed onto

Matrigel-coated plates in STEMDiffTM astrocyte maturation medium.

2.7.5 Karyotyping analysis

Karyotyping was performed on WT, R418X+/-, & R493X-/- KI by WiCell Cytogenetics, Inc.

(Madison, WI). The S116X+/- line was previously karyotyped by Cell Line Genetics

(Madison, WI) following the initial production of this line (60).

2.8 CRISPR/Cas9 gene editing

An isogenic hiPSC line homozygous for GRN R493X knock-in (R493X-/- KI) was

generated from WT using aspects of a previously established CRISPR/Cas9 gene

editing protocol (353). Briefly, guide RNA (gRNA) sequences were designed according

to the online Optimized CRISPR Design tool (http://crispr.mit.edu) to target a region for

double-strand breakage slightly upstream of the GRN R493 codon. A mutagenizing

single-stranded oligonucleotide (ssODN, Integrated DNA Technologies) was designed

to knock-in the R493X mutation and silently introduce a HindIII restriction enzyme site.

hiPSCs were co-transfected (Lipofectamine Stem Transfection Reagent, Invitrogen)

with knock-in ssODN and gRNA/Cas9 ribonucleoprotein (Integrated DNA Technologies)

complexes for 72 h. CRISPR/Cas9 edited hiPSCs were seeded at clonal density (25

cells/cm2). Individual colonies of adequate size and morphology, were picked and plated

for expansion. Genomic DNA was isolated using QuickExtractTM DNA Extraction

79

Solution (Lucigen) and PCR products containing the R493X target site were digested

with HindIII (New England Biolabs). Clones with positive digestion signal were selected

for Sanger sequencing to confirm the clean homozygous introduction of the R493X KI

mutation. Sanger sequencing was performed by the UBC Sequencing + Bioinformatics

Consortium.

2.9 Multielectrode array electrophysiology

WT and mutant hiPSC-derived NPCs (DIV 30) were co-cultured with hiPSC-derived WT

astrocytes (DIV 60+) at a 1:1 ratio and plated onto 48-well multielectrode array (MEA)

plates (Axion Biosystems) coated with PLO/L. For the 2 month time course MEA

experiment WT NPCs, WT astrocytes, and co-cultures of WT NPCs and WT astrocytes

were compared to evaluate the electrophysiological benefits of astrocyte inclusion. Cells

were plated using the drop seed method, where 15 µL of mixed cell suspension (60,000

cells in total) in a 10% dilution of Matrigel in BrainPhysTM media was pipetted onto the

electrode array in the center of each well. Cultures were fed by partial media

replacement every 3 days according to our cortical neuron differentiation protocol

described in section 2.7.4. Spontaneous electrophysiological activity of neuron-

astrocyte co-cultures was recorded for either 10 min at cortical neuron DIV 50

(WT/mutant NPC + WT astrocyte co-cultures) or 5 min weekly for 2 month time course

experiment using the Axion Biosystems Maestro MEA at 37 °C and 5% CO2. Data

analysis was performed using AxIs software (Axion BioSystems) to extract the number

of spikes and bursts from the recording file. Quality criteria for the assays were defined

as follows: an electrode having an average of more than 5 spikes/min and wells with

80

less than 30% of the total electrodes active were considered inactive and excluded from

the analysis. To assess synchronous network activity, at least 25% of the total 16

electrodes in each well were required to participate in a network event to be designated

as a network burst.

2.10 ELISA

PGRN levels in hiPSC-derived cortical neuronal and astrocyte whole cell lysates

(prepared according to section 2.5.1) and concentrated supernatants (Amicon Ultra 0.5

mL, 50 kDa, 25X concentrated) were determined by ELISA (Adipogen, human) using

the manufacturer’s protocol. Cultures were treated with an equal volume of media, and

an equal volume of supernatant was concentrated for each sample. Cortical neuron and

astrocyte lysates (1 mg/mL) and concentrated supernatant were diluted in ELISA buffer

at optimized dilutions. Pgrn levels in MEF whole cell lysates (prepared according to

section 2.5.1) and RIPA-soluble Grn+/+ and GrnR493X/R493X hemi-brain lysates (prepared

according to 2.16.2) were determined by ELISA (Adipogen, mouse) using the

manufacturer’s protocol. MEF lysates (1 mg/mL, Grn+/+ 1:20 and GrnR493X/R493X 1:5) and

RIPA-soluble hemi-brain lysates (10 mg/mL, Grn+/+ 1.5:10 and GrnR493X/R493X 1:5) were

diluted in ELISA buffer. G418 concentration levels in whole-brain lysates (prepared

according to 2.16.3) were determined by Gentamicin ELISA (Creative Diagnostics)

according to the manufacturer’s protocol using a G418 standard curve of 0.1, 0.3, 1, 3,

10, 30, and 100 ng/mL.

81

2.11 qPCR

Total RNA was extracted from hiPSCs using RNeasy Plus Mini Kit (Qiagen). cDNA was

produced by reverse transcription using the High-Capacity RNA-to-cDNA Kit (Applied

Biosystems, Thermo Fisher Scientific). To measure the endogenous gene expression of

pluripotency factors, qPCR analysis was performed using the 7900HT Fast Real-Time

PCR System (Applied Biosystems, Thermo Fisher Scientific). mRNA was detected with

TaqmanTM probes (Thermo Fisher Scientific, OCT4 Hs01895061_u1, LIN28

Hs00702808_s1, NANOG Hs02387400_g1, SOX2 Hs00602736_s1, GRN

Hs00963707_g1, GAPDH Hs03929097_g1, and HPRT1 Hs02800695_m1) in

combination with the TaqmanTM Universal PCR Master Mix (Applied Biosystems,

Thermo Fisher Scientific). Gene expression of hiPSC markers was normalized to

GAPDH housekeeping gene and compared to human fibroblast expression using the

ΔΔCt-method. GRN gene expression in cortical neurons and astrocytes was normalized

to the mean of GAPDH and HPRT1 housekeeping genes and compared to vehicle-

treated WT expression using the ΔΔCt-method.

2.12 Mouse genotyping

Mouse ear notch DNA was isolated in Chelex® 100. Briefly, 100 µL Chelex® 100 was

added to ear notch, vortexed for 10 sec, and pulse spun to ensure tissue was

submerged in the solution. Samples were then incubated at 95 C for 20 min. After 10

min, sample tubes were opened to release pressure and agitated to enhance tissue

digestion before completing the remaining 10 min of incubation. Samples were then

further agitated by running the bottom of the tubes across a metal tube storage rack and

82

centrifuged at 21,000 x g for 1.5 min. The PCR was run according to specifications

outlined on the Jackson Laboratories B6.129S4(SJL)-Grntm2.1Far/J strain data sheet with

2 µL of the resulting DNA-containing supernatant from each sample loaded into a 23 µL

PCR master mix solution. The resulting PCR products were resolved on a 2% agarose

gel and detected using SafeView Classic (Applied Biological Materials).

2.13 Histology

2.13.1 Immunocytochemical staining and quantification

Cells were fixed with 4% paraformaldehyde (PFA) for 15 min and washed three times

with Dulbecco PBS (D-PBS, Gibco, Thermo Fisher Scientific). Cells were blocked and

permeabilized with 10% donkey or goat serum (Sigma-Aldrich) in D-PBS containing

0.1% Triton X-100 (Abcam) for 1 h at room temperature (RT). Primary antibodies were

then diluted (Table 1-1) in 10% donkey or goat serum in D-PBS and applied to cells

overnight at 4 °C. All Alexa Fluor®-tagged secondary antibodies (Molecular Probes,

Thermo Fisher Scientific) were applied at a dilution of 1:500 at RT for 2 h. The

coverslips were then mounted in DAPI mounting medium (Vector Laboratories). Images

were captured with ZEN 2 software using a Zeiss 880 scanning laser confocal

microscope. Image quantification was performed using Fiji. The % of neurons

expressing FOXG1/TBR1 was determined by applying uniform thresholds and

converting images to binary using the watershed segmentation tool and quantifying the

number of individual DAPI+/neuronal marker+ cells with the analyze particles function.

The % neurons+ for MAP2 and % astrocytes+ for GFAP was determined by manually

counting nuclei of MAP2- and GFAP+ cells. Subtracting the number of MAP2- cells from

83

the total number of DAPI+ nuclei and dividing it by the total number of DAPI+ nuclei

yielded % MAP2+ and dividing the number of GFAP+ cells by the number of DAPI+

nuclei yielded % GFAP+. Cytotoxicity was assayed by staining fixed 96-well plate

neuronal cultures with Hoechst dye (1.5 µg/mL, Invitrogen). These plates were imaged

using a Cellomics ArrayScanTM plate scanner, and the number of cells/field were

automatically counted with acQuisitionTM software.

2.13.2 Brain immunofluorescence microscopy and quantification

Free-floating sagittal brain sections were blocked and permeabilized with 10% donkey

or goat serum (Sigma-Aldrich) in D-PBS containing 0.1% Triton X-100 (Abcam) for 2 h

at RT. Primary antibodies were then diluted (Table 1-1) in 10% donkey or goat serum in

D-PBS and applied to sections overnight at 4 °C with gentle agitation. All Alexa Fluor®-

tagged secondary antibodies were used at a dilution of 1:500 at RT for 2 h. The

sections were then mounted in DAPI mounting medium (Vector Laboratories). AAV9-

eGFP-Cre mouse brain sections were simply mounted in DAPI mounting medium to

assess eGFP expression. Z-stacks and tilescans were captured with ZEN 2 software

using a Zeiss 880 scanning laser confocal microscope using 20X and 40X objective

lenses. AAV9-eGFP-Cre whole mouse brain tilescans were captured with a 20X

objective lens, while all other brain section imaging was done with a 40X objective lens.

High magnification (100X, 2.5X digital zoom with 40X objective) Tuj1-TDP-43-DAPI co-

stained images were captured as single Z-plane images. All other images were

obtained through single field 30 µm thick (1 image / 5 µm step) z-stack and z-stack

tilescan acquisitions and were processed into maximum intensity projections using ZEN

84

2 software prior to image quantification performed using the Fiji processing package for

ImageJ (National Institutes of Health).

The total area (µm2) of lysosomal marker (Lamp1 and DppII) staining per CA3 or

VPM/VPL field was quantified using Fiji by applying a uniform threshold to all images

and watershed segmentation to distinguish individual lysosomal vesicles. The analyze

particle function was used to quantify areas ranging from 0.05-infinity µm2, providing

both total Lamp1+/DppII+ area and average lysosomal vesicle size. Total Pgrn staining

area was quantified by applying a uniform threshold across all images and analyzing

particle areas 0-infinity µm2. Microgliosis was quantified by assessing both total Iba1+

area and the number of Iba1+ cells; this was determined by applying a uniform

threshold and analyzing particles with an area of 0.25-infinity µm2 and 20-infinity µm2,

respectively. Astrogliosis was similarly quantified by measuring total Gfap+ area in

uniformly thresholded images using the analyze particles tool (0.25-infinity µm2).

Thalamic VPM/VPL total C1qa complement staining area was quantified by applying a

uniform threshold across all images and analyzing particle areas 0-infinity µm2. Ventral

thalamic density of Vgat+ synapses was determined by quantifying both the total Vgat+

area and the number of Vgat+ puncta. Vgat images were uniformly thresholded and

segmented using the watershed function, and particles limited to a 0.3-5 µm2 area were

analyzed. The % of Vgat+ synaptic area co-stained for C1qa was determined by

generating regions of interest (ROIs) of Vgat+ synapses using the above image

processing and analyze particle specifications. These Vgat ROIs were then applied to

corresponding co-stained C1qa images, and any C1qa+ staining outside of ROIs was

85

cleared using the clear outside function. The remaining intra-ROI C1qa staining was

uniformly thresholded and quantified by analyzing particle areas 0-infinity µm2. The total

C1qa area within Vgat ROIs was divided by total Vgat area and multiplied by 100 to

obtain % of Vgat+ synaptic area co-stained for C1qa. For data normalized on a per cell

basis, the number of DAPI+ nuclei were quantified per field by applying a uniform

threshold to all images and watershed segmentation to distinguish individual nuclei. The

analyze particle function was used to quantify areas ranging from 25-infinity µm2 to

determine the total number of cells per field.

Foxp2+ excitatory neuron and Vgat+ synaptic total thalamic density were obtained from

MIPs generated from z-stack tilescans containing both the hippocampal and thalamic

brain structures (40X magnification, 0.6 zoom, 8 x 12 tilescan). The polygon selection

tool was used to draw a boundary surrounding the thalamus to measure the total

thalamic area (mm2). Any Foxp2+ or Vgat+ staining outside of the thalamic boundary

was eliminated using the clear outside function, and a uniform threshold was applied to

each tilescan image. Following watershed segmentation to allow for clear detection of

individual Foxp2+ nuclei and Vgat+ synaptic puncta, the number of particles 50-200 µm2

(Foxp2) or 7.75-25 µm2 (Vgat) in size were analyzed and normalized to total thalamic

area.

Microglial skeletal analysis was conducted using the Fiji Analyze Skeleton (2D/3D)

plugin using a modified version of a previously described protocol (354). Pre-analysis

Iba1 image processing required adjusting image brightness maximum (-40 to 215),

86

applying unsharp mask (radius 3 and mask weight 0.6), and reducing background noise

using the despeckle tool. These processed images were then uniformly thresholded,

further despeckled, processed using the binary Close- tool, and background noise was

reduced further using the remove outliers function. The analyze particles function was

then used to generate ROIs masking Iba1+ staining of a 10-infinity µm2 area and add

them to the ROI manager. The original image was then reopened, and the initial Iba1

image processing steps were reapplied (brightness, unsharp mask, and despeckled). All

ROIs were then selected in the ROI manager and applied to this image using the OR

(combine) function. Any Iba1+ staining outside of the applied ROIs was eliminated using

the clear outside function, and a threshold of 1-255 was applied to completely fill in the

ROIs generating a clean binary image of microglial cell shapes. This image was then

skeletonized using the binary processing tool and analyzed with the Analyze Skeleton

(2D/3D) plugin (prune cycle: none, checked boxes = ‘show detailed info’ and ‘display

labelled skeletons’) to determine the total number of microglial branches per field. To

quantify microglial Pgrn fluorescent intensity (integrated density) from Iba1-Pgrn co-

staining microglial ROIs were generated from Iba1 images according to the above

skeletal analysis method. These ROIs were applied to their corresponding Pgrn co-

stained image, and any Pgrn+ staining outside of the applied microglial ROIs was

removed using the clear outside function. Pgrn integrated density in each microglial ROI

was then measured from the ROI manager. The calculated integrated densities for all

microglial ROIs per image were averaged, providing the Pgrn microglial fluorescent

intensity per field.

87

2.13.3 Brightfield microscopy

Live cultures were photographed with a digital camera mounted to a tissue culture

microscope (40X obj. lens) throughout the differentiation process from hiPSCs to

cortical neurons.

2.14 Surgical procedures

2.14.1 Intracerebroventricular injection of AAV particles

P0 C57BL/6J neonates were anesthetized using isoflurane and bilaterally injected ICV

with 0.5 µL of viral vector (AAV-GRN-R493X-V5, 8E13 PFU/mL and AAV9-eGFP-Cre,

1E13 PFU/mL) through a finely drawn glass micropipette as described previously (355).

PTC readthrough treatments were performed when AAV-GRN-R493X-V5 mice were 6

weeks old.

2.14.2 Bolus intracerebroventricular injection of drugs

2.14.2.1 AAV-GRN-R493X-V5 mice

A single bolus ICV injection of vehicle, CDX5-288, or G418 ± CDX5-288 was performed

in 6 week old AAV-GRN-R493X-V5 mice according to previously published works with

some modifications (356). First, to establish ICV targeting technical proficiency, a

healthy C57BL/6J control mouse right lateral ventricle was injected stereotactically

through a borehole (coordinates: -1.0 mm lateral/-0.3 mm posterior/-3.0 mm depth to

bregma) by loading a glass micropipette needle attached via tubing to a 5 µL glass

syringe (Hamilton, 600 series, model 65) with 5 µL 0.1% Fast Green FCF dye (Sigmal-

Aldrich) driven by microinjection syringe-pump (UMP3T-1, World Precision Instruments)

88

to deliver the dosage at a flow rate of 1 µL/min. The dye-injected mouse was

immediately euthanized, and its brain was collected to evaluate the pattern of CNS

delivery. Following similarly conducted 3 µL drug injections in AAV-GRN-R493X-V5

mice, the needle was left in place for 3 minutes and then removed at a rate of 1

mm/second while holding a cotton swab against the skull at the base of the needle. The

incision was then sutured with monocryl sutures (Ethicon, J303H) to close the skin, and

the mice were placed on the heating pad warmed half of a cage to ensure the mice

were kept at 36-38 °C for the first 1.5 h while they recovered from the anesthesia, with

the option for them to move away from the heat source. After treated AAV-GRN-R493X-

V5 mice had recovered fully from surgery, they were singly housed for 72 h. 5 mg/kg of

meloxicam was injected subcutaneously once per day for the first 2 days post-

operation.

2.14.2.2 Grn+/+ and GrnR493X/R493X mice

Bolus ICV injections of both vehicle and G418 were performed in Grn+/+ and

Grn493X/R493X mice to validate in vivo PTC readthrough in this model and optimize G418

dosing for a proposed recurrent injection model according to previously published works

with some modifications (122). Briefly, both 1X (100 µg, 72 h treatment) and 2X (50 µg

repeated 48 h treatments) ICV injections of G418 were tested. Mice lateral ventricles

were injected stereotactically (1-3 µL volume, coordinates: -1.0 mm lateral/-0.3 mm

posterior/-3.0 mm depth to bregma) with microinjection syringe-pump driven 10 µL glass

syringe (Hamilton, 701 LT) connected to a 51 mm Luer-locking needle (Hamilton, 31G,

89

K-Fel) loaded with either saline or G418 solutions delivered at a flow rate of 0.5 µL/min.

The incision was sutured, and post-surgical care was provided as in 2.14.2.1.

2.14.3 Intracerebroventricular iPRECIO® pump implantation in Grn+/+ and

GrnR493X/R493X mice

To establish iPRECIO® pump ICV targeting technical proficiency, a SMP-310R

iPRECIO® Programmable Infusion Pump (AZLET) was loaded with 0.1% Fast Green

FCF dye, programmed to immediately deliver dye the max rate of 8 µL/hour for 1 hour,

and implanted subcutaneously in a Grn+/+ mouse under isoflurane anesthesia. Pumps

were connected to a 30G 3 mm long Brain Infusion Kit 3 (AZLET) cannula. The base of

the cannula was coated in Loctite 454 (AZLET) glue and was stereotactically driven

through a borehole in mice right lateral ventricles (coordinates: -1.0 mm lateral/-0.3 mm

posterior/-3.0 mm depth to bregma). The physical pump and excess tubing were then

sterilely inserted into a subcutaneous pocket created caudally from the incision using a

blunt dissection tool. Following completion of iPRECIO® pump dye ICV perfusion, the

mouse was immediately euthanized, and its brain was collected to evaluate the pattern

of CNS delivery. iPRECIO® pumps loaded with saline or 25 mg/mL G418 were similarly

implanted, and the brain infusion cannula was inserted into the right lateral ventricles of

Grn+/+ and Grn493X/R493X mice. Saline/G418 loaded pumps were programmed to allow

mice to recover for 48-72 h and then dispense recurring bolus ICV doses of 1-2 µL

volumes (saline or 25/50 µg G418) at the max flow rate (8 µL/hr) every 48-72 h for the

experimental duration (see Figure 5-20 for more details). The incision was sutured, and

90

post-surgical care was provided as in 2.14.2.1.

2.15 Open-field behavioural assay

The open-field test was performed to evaluate anxiety in the GrnR493X/R493X mouse

model of frontotemporal dementia. The test consisted of a single 10 min trial in a white

opaque 40 cm x 40 cm x 30 cm arena (Maze Engineers). The center zone was defined

as a square covering 16% of the total area (16 cm x 16 cm central square). The mice

were moved to the experimental room at least 1 hour before starting the tests. To begin

each test, a mouse was introduced to the center of the square, and its behaviour was

captured on video for 10 min. On recording days, males were tested before any

females, and the area was cleaned with 50% ethanol and allowed to dry completely

between each test. The duration that each mouse stayed in either the peripheral or

central regions was quantified using JWatcher software (UCLA).

2.16 Brain collection and processing

2.16.1 AAV mouse model

72 h following ICV vehicle/drug injection, mice were anesthetized using isoflurane and

transcardially perfused with PBS before brain collection. Brains were immediately flash

frozen on dry ice and stored at -80 °C until lysate preparation. Whole-brains were later

thawed, homogenized in 500 µL TBS lysis buffer (150 mM NaCl, 1 mM EDTA, 1 mM

sodium orthovanadate, 1 mM NaF, 1 mM β-glycerophosphate, 2.5 mM sodium

pyrophosphate, 1 mM PMSF, PhosSTOP, cOmplete mini protease inhibitor), sonicated

for 10 sec at 20% amplitude and ultracentrifuged at 100,000 x g for 20 min at 4 °C. The

91

supernatants were collected as TBS-soluble extract, and the pellets were homogenized

in RIPA buffer (TBS lysis buffer + 1% NP-40), sonicated, and again ultracentrifuged at

100,000 x g for 20 min at 4 °C. The supernatants were collected as RIPA-soluble

extract. The protein concentration of RIPA-soluble fractions was measured by Bradford,

diluted to 1 mg/mL for automated capillary electrophoresis western analysis. 10-week

old AAV9-eGFP-Cre mice were perfused with PBS prior to brain collection. Brains were

immediately incubated in 4% PFA overnight and transferred to PBS + 0.05% azide at 4

°C for long term storage. 40 µm thick sagittal brain sections were prepared using a

vibratome (Leica, VT1000 S) and stored in PBS + 0.05% azide at 4 °C.

2.16.2 GrnR493X/R493X baseline characterization

80 week old Grn+/+ and GrnR493X/R493X mice were anesthetized using isoflurane and

transcardially perfused with PBS before brain collection. Brains were hemisected down

the midsagittal plane, one half was immediately flash frozen on dry ice, and the other

fixed overnight in 4% PFA. Frozen halves were stored at -80 °C until lysate preparation,

and fixed halves were transferred to PBS + 0.05% azide for long term storage. Floating

sections were prepared and stored according to section 2.16.1. Frozen brain halves

were thawed, homogenized in 250 µL RIPA buffer sonicated for 10 sec at 20%

amplitude and ultracentrifuged at 100,000 x g for 20 min at 4 °C. The supernatants were

collected as RIPA-soluble extract, and the insoluble protein pellet was further extracted

with urea buffer (30 mM Tris-HCl pH 8, 7 M urea, 2 M thiourea, 4% CHAPS,

PhosSTOP, cOmplete mini protease inhibitor) and centrifuged at 150,000 x g for 45 min

at 4 °C. The protein concentration of RIPA-soluble fractions was measured by Bradford,

92

diluted to 1 mg/mL for ELISA and western blots. The insoluble urea fractions were

diluted according to the protein concentration of corresponding RIPA-soluble fractions.

2.16.3 Vehicle/G418 treated Grn+/+ and GrnR493X/R493X mice

ICV injected (bolus/iPRECIO® pump) vehicle/G418 treated Grn+/+ and GrnR493X/R493X -

mice were anesthetized using isoflurane and transcardially perfused with PBS prior to

brain collection. The majority of brains were hemisected down the midsagittal plane,

one half was immediately flash frozen on dry ice, and the other fixed overnight in 4%

PFA, though a subset of experiments were conducted on whole flash frozen brains.

Frozen brain tissue was stored at -80 °C until lysate preparation, and fixed halves were

transferred to PBS + 0.05% azide for long term storage. Floating sections were

prepared and stored according to section 2.16.1. In a minority of cases, whole-brains

were dissected, flash frozen, and stored at -80 °C until lysate preparation. Frozen hemi-

brains were thawed, homogenized in 250 µL (500 µL for whole-brains) RIPA buffer

sonicated for 10 sec at 20% amplitude and ultracentrifuged at 100,000 x g for 20 min at

4 °C. The protein concentration of RIPA-soluble fractions was measured by Bradford,

diluted to 1 mg/mL for Pgrn and G418 quantification by ELISA/western blot.

2.17 Statistical analysis

All values are expressed as the mean ± standard error of mean (SEM). In experiments

where two groups were compared a standard unpaired two-tailed Student’s t-test was

performed. For comparisons of more than two groups, one-way analysis of variance

(ANOVA) was used followed by Tukey’s post hoc test. For comparisons of two or more

93

groups with multiple time points, two-way ANOVA was used followed by Bonferroni post

hoc test comparing each column to all other columns. For comparing the frequency of

lesion development and whisker barbering in our cohorts of Grn+/+ and GrnR493X/R493X

mice, Chi Square test was used. p-values less than 0.05 were considered significant.

Statistical analysis was performed using GraphPad Prism Software, Version 5.0.

94

Chapter 3: Exogenous PTC readthrough of nonsense mutant GRN

expression constructs

3.1 Introduction

Despite 23% of known causal FTLD-GRN variants introducing nonsense mutations,

PTC readthrough of GRN nonsense mutations had not yet been experimentally

demonstrated when this research project was initiated. Exogenous nonsense mutant

gene expression is a common approach used to validate whether a particular mutation

and surrounding sequence context are susceptible to PTC readthrough. Transfection is

a process that enables exogenous delivery of extracellular material (DNA expression

plasmids, mRNA, protein) into animal cells through a variety of methods. Typically,

transfection methods gain entry into the cell by creating transient pores in the plasma to

enable uptake of exogenous materials. Two of the most common transfection methods,

lipofection, and electroporation were employed in this chapter. Lipofection gains entry

into the cell through liposomal-membrane fusion, which releases encapsulated target

material into the cytoplasm. Electroporation involves a brief application of an electric

field to the cells, which temporarily increases the permeability of cell membranes

allowing for the introduction of foreign material. The highly transfectable HEK293 cell

line is believed to have originated from an embryonic adrenal precursor cell transformed

by incorporation of an ~4.5 kb segment of adenovirus 5 DNA into chromosome 19

(357,358). HEK293 cells are used widely for transfection experiments due to their short

doubling time, propensity to highly express exogenously delivered genes, and routine

culture methods.

95

AAVs are single-stranded DNA viruses commonly used as expression vectors in

mammalian systems. These viruses are particularly well-suited for transduction due to

their apparent lack of pathogenicity and ability to stably integrate into the host cell

genome at a specific AAVS1 locus in chromosome 19. These properties have

generated significant interest in utilizing AAVs for gene therapy applications. As

previously mentioned in section 1.2.7, an AAV GRN gene therapy-based approach has

been shown to restore GRN expression in preclinical models (218,221), leading to the

development of the AAV9 GRN expression vector, PR006, which is currently

undergoing phase I/II clinical validation in FTLD-GRN patients.

Generating complex genetic knock-in mouse models generally requires time-consuming

and expensive germ-line manipulations. ICV AAV injection into the murine neonatal

brain represents an alternative, more cost-effective way to manipulate mouse CNS

gene expression (359). When AAV are injected ICV on P0, mice develop stable long-

term brain-wide neuronal transgene expression (355). In this chapter, we sought to

determine whether we could induce PTC readthrough in HEK293 cells exogenously

expressing nonsense mutant GRN with the aminoglycoside G418. We also set out to

determine whether any of the recently generated CDX-series readthrough enhancer

compounds could synergistically increase GRN nonsense mutation readthrough.

Finally, we assessed whether we could induce in vivo brain PTC readthrough of

exogenously expressed nonsense mutant GRN in our AAV-GRN-R493X-V5 mice.

96

3.2 Results

3.2.1 PTC readthrough in transiently transfected HEK293 cells

PTC readthrough efficiency is affected by the specific nonsense codon sequence and

the flanking nucleotide sequences (234,240). Therefore, to survey the responsiveness

of GRN nonsense mutations to PTC readthrough, we designed several clinical GRN

nonsense mutation expression constructs. We first transiently transfected HEK293 cells

with S116X (UAA), R418X (UGA), and R493X (UGA) C-terminally V5-tagged human

PGRN expression constructs (Figure 3-1, A) and treated them for 72 h with the

aminoglycoside G418. Mock transfected HEK293 cells showed no detectable V5 signal

(Figure 3-2). Since the V5 tag was inserted at the C-terminus, any V5 detected

represented full-length PGRN generated by PTC readthrough. G418 induced GRN PTC

readthrough in cells with the R418X and R493X mutations, with V5 (full-length PGRN)

detected in both the cell lysate (intracellular) and supernatant (extracellular) fractions

(Figure 3-2). The accumulation of full-length PGRN in both the intra- and extracellular

fractions of transiently transfected HEK293 cells suggests that the PGRN readthrough

product retains its ability to be post-translationally processed via the secretory pathway.

The S116X mutant did not respond to G418 treatment (Figure 3-2), and this may be

because aminoglycoside-induced PTC readthrough is most efficient at UGA nonsense

codons and least efficient at UAA nonsense codons (UGA > UAG > UAA) (234).

97

98

Figure 3-1 Induction of PTC readthrough by G418 and CDX5 enhancers in cells

stably expressing nonsense mutant GRN-V5. (A) Schematic of full-length PGRN

highlighting the position of the S116X (UAA), R418X (UGA), and R493X (UGA)

nonsense mutations in relation to the position of individual GRNs and the C-terminal V5

tag. (B) HEK293 cells transiently transfected with GRN-V5 expression constructs

bearing the indicated nonsense mutations were treated with G418 for 72 h. (C) HEK293

cell lines stably expressing GRN-V5 with the indicated nonsense mutations were treated

with G418 and the indicated concentrations of CDX5-1, CDX5-196, and CDX5-288 for

72 h. Cell culture supernatants (extracellular) and cell lysates (intracellular) were

subjected to automated capillary electrophoresis western analysis. Full-length PGRN

was detected with a V5 antibody. Actin was measured in cell lysates as a loading

control. The readthrough enhancement ratios (normalized to G418 alone) are indicated

under the lanes. The proportion loaded was 15-20X lower for the extracellular samples

than for the intracellular samples.

99

Figure 3-2 Induction of PTC readthrough by G418 in transiently transfected cells

expressing nonsense mutant GRN-V5. HEK293 cells transiently transfected with

GRN-V5 expression constructs bearing the indicated nonsense mutations were treated

with G418 for 72 h. Cell culture supernatants (extracellular) and cell lysates

(intracellular) were subjected to automated capillary electrophoresis western analysis.

Full-length PGRN was detected with a V5 antibody. Actin was measured in cell lysates

as a loading control. The proportion loaded was 15-20X lower for the extracellular

samples than for the intracellular samples.

100

3.2.2 PTC readthrough in stably transfected HEK293 cells

To further screen and validate the GRN PTC readthrough observed in transiently

transfected cells, we generated HEK293 lines stably expressing GRN-V5 nonsense

mutant constructs. When treated with G418, these cell lines exhibited a similar GRN

PTC readthrough response to that observed in transiently transfected cells (Figure 3-1,

B). Stable GRN-V5 expressing HEK293 lines provided an ideal screening platform to

test which CDX-series compounds could most effectively enhance G418-induced GRN

PTC readthrough. We treated HEK293 cell lines expressing nonsense mutant GRN-V5

with G418 alone or in combination with readthrough enhancers CDX5-1 (337), CDX5-

196 (Figure 3-3, A), or CDX5-288 (Figure 3-3, B) for 72 h, and measured intra- and

extracellular V5 levels. CDX5 compounds may enhance G418-induced PTC

readthrough activity in both the R418X and R493X mutant lines, and full-length PGRN

was detected in both the intra- and extracellular fractions (Figure 3-1, B). Given that the

data presented in Figure 3-1 is n=1, we did not conduct statistical analysis and therefore

viewed this data as a preliminary low confidence observation. Since CDX5-288 may

possess the highest enhancement of readthrough activity (Figure 3-1, B), we selected

co-treatment with G418 and CDX5-288 for further validation in patient hiPSC-derived

cortical neurons and astrocytes and AAV-GRN-R493X-V5 mouse model.

101

Figure 3-3 Chemical structures of CDX5-196 and CDX5-288. N-substituted

phthalimide derivatives CDX5-196 (A) and CDX5-288 (B).

102

3.2.3 PTC readthrough in HEK293 cells transduced with AAV-GRN-R493X-V5

viral particles

Upon generating AAV-GRN-R493X-V5 viral particles, we wanted to assess whether

transduced cells expressing the GRN-R493X-V5 construct were responsive G418-

induced PTC readthrough before proceeding with in vivo studies. To this end, we

transduced HEK293 cells for a total of 5 days with increasing concentrations of AAV-

GRN-R493X-V5 and co-treated them with G418 for 72 h. G418-induced GRN-R493X-

V5 PTC readthrough was only observed in HEK293 cells transduced with the highest

viral titer of 8.8E10 PFU/mL, as indicated by both intra- and extracellular V5 detection

(Figure 3-4).

103

Figure 3-4 Induction of PTC readthrough in cells transduced with increasing titers

of AAV-GRN-R493X-V5. HEK293 cells transduced for 5 days with increasing titers of

AAV-GRN-R493X-V5 viral vector were treated with G418 48 h post-transduction for an

additional 72 h. Cell culture supernatants (extracellular) and cell lysates (intracellular)

were subjected to automated capillary electrophoresis western analysis. Full-length

PGRN was detected with a V5 antibody. The proportion loaded was 15-20X lower for

the extracellular samples than for the intracellular samples.

104

3.2.4 Technical demonstration of CNS P0 delivery of AAV particles and bolus

ICV injection in adult mice

Since the AAV-GRN-R493X-V5 plasmid does not contain a reporter gene, we first

wanted to establish that our AAV injection technique achieved an adequate percentage

of neuronal transduction. Before testing the AAV-GRN-R493X-V5 virus, we first

conducted P0 pup ICV injections of a serotype matched AAV9-eGFP-Cre expression

vector. We observed diffuse eGFP expression in the brains of 10 week old AAV9-eGFP-

Cre mice (Figure 3-5), thus validating that our AAV injection technique should be

capable of achieving robust CNS expression of GRN-R493X-V5. To confirm whether

the stereotactic coordinates (-1.0 mm lateral/-0.3 mm posterior/-3.0 mm depth to

bregma) we planned to use for targeted infusion of mice right lateral ventricles resulted

in successful ICV delivery; we conducted a test ICV injection of 0.1% Fast Green FCF

dye in a healthy control C57BL/6J mouse. Photographic observation of freshly

harvested dye-injected brain tissue demonstrated the expected ventricular distribution of

Fast Green FCF dye (Figure 3-6).

105

Figure 3-5 Technical demonstration of AAV P0 ICV injection method. P0 ICV

injection using an AAV9-eGFP-Cre viral vector with expression driven by the human

SYN1 promoter. Diffuse brain-wide eGFP (green) expression apparent 10 weeks after

bilateral P0 ICV injection. Cell nuclei were counterstained with DAPI (blue). Scale bar, 1

mm.

106

Figure 3-6 Technical demonstration of bolus ICV injection method in adult mice. A

C57BL/6J mouse was stereotactically injected with 5 µL 0.1% Fast Green FCF dye at

coordinates: -1.0 mm lateral/-0.3 mm posterior/-3.0 mm depth to bregma. Immediately

following dye injection, the brain was collected and photographed, demonstrating

ventricular delivery.

107

3.2.5 PTC readthrough in AAV-GRN-R493X-V5 mice

Having demonstrated PTC readthrough in exogenous GRN nonsense mutant

expression models in vitro, we sought to demonstrate initial proof-of-concept for this

approach in vivo. Brain PTC readthrough induced by ICV G418 has been previously

demonstrated in mice harboring an inducible gene-targeting system driven by the full-

length expression of a nonsense mutant Cre recombinase and in R294X Mecp2 knock-

in mice (360,361). Newborn pups (P0) were bilaterally injected ICV with virus solution

and aged to at least 6 weeks to allow for sufficient blood-brain-barrier (BBB) maturation

before treatment (362). RIPA-soluble brain lysates from each ICV treated AAV-GRN-

R493X-V5 mouse were probed for V5 expression to assess the extent of GRN R493X

PTC readthrough in vivo. No V5 signal was detected in either the vehicle (n=6) or

CDX5-288 (n=4) AAV-GRN-R493X-V5 brains 72 h after ICV injection (Figure 3-7, A-B).

By contrast, ICV injections with G418 alone (n=5) or G418 and CDX5-288 (n=4)

induced clear GRN-R493X-V5 PTC readthrough, as shown by the detection of V5 signal

at the expected molecular weight of PGRN, using automated capillary electrophoresis

western analysis (Figure 3-7, A-B). G418 significantly induced readthrough in AAV-

GRN-R493X-V5 mice compared to treatment with vehicle solution (Figure 3-7, A-B).

Though robust V5 levels were detected, co-treatment of AAV-GRN-R493X-V5 mice with

CDX5-288 did not increase full-length PGRN levels compared to G418 alone (Figure 3-

7, A-B). Our data indicate that short term administration of the aminoglycoside G418

can induce GRN PTC readthrough in vivo when delivered intraventricularly in an

exogenous AAV transgene expression model.

108

Figure 3-7 Induction of PTC readthrough in AAV-R493X-GRN-V5 mice. (A) 6-week-

old AAV-GRN-R493X-V5 mice were stereotactically injected ICV with either vehicle

solution, G418, CDX5-288, or a combination of G418 + CDX5-288 ICV and sacrificed

after 72 h. Whole-brain RIPA-soluble protein extracts were subjected to automated

capillary electrophoresis western analysis. Full-length PGRN was detected with a V5

antibody, and actin was measured in brain lysates as a loading control. (B)

Chemiluminescence quantification of automated capillary electrophoresis western

analysis V5 detection normalized to actin. Values are shown as mean ± SEM; * p <

0.05, one-way ANOVA with Tukey’s multiple comparison test.

109

3.3 Discussion

We generated several human expression constructs with known clinical GRN nonsense

mutations and observed a more robust PTC readthrough response to G418 and CDX5-

288 enhancer treatment in R418X and R493X mutant HEK293 cells compared to G418

alone or co-treatment with other enhancer compounds. To avoid confounding measures

of endogenous PGRN in HEK293 cells and ensure the measured PGRN was reflective

of full-length protein produced by readthrough, the GRN expression constructs bore a

C-terminal V5 tag. The readthrough enhancer compounds are N-substituted phthalimide

derivatives identified in a high-throughput screen in combination with sub-active

concentrations of the aminoglycoside paromomycin and further systematically

functionalized to increase their potency (337). These results provide a clear

demonstration that GRN UGA PTCs are most susceptible to readthrough, in alignment

with previously published reports evaluating the impact of nonsense codon sequence on

readthrough efficiency (234).

The production of AAV-GRN-R493X-V5 viral particles enabled us to validate whether

PTC readthrough could be achieved from exogenously delivered CNS nonsense mutant

GRN. Bilaterally injecting this virus into the developing P0 murine lateral ventricles

induced widespread CNS neuronal expression of V5-tagged R493X GRN. Since G418

and other aminoglycosides do not readily cross the BBB, we sought to determine

whether ICV administration of compounds could be a viable strategy for widespread

brain exposure and readthrough. Indeed, our results provide proof-of-concept that a

single ICV injection of G418 can induce PTC readthrough in an AAV-GRN-R493X-V5

110

mouse model in only 72 h. However, given the likely massive overexpression of the

GRN-R493X-V5 construct driven by the strong hybrid chicken β-actin promoter, it is

difficult to make meaningful conclusions regarding the optimal dosage of either G418 or

CDX5-288 enhancer compound. The recently generated GrnR493X/R493X nonsense

mutant knock-in mouse (165) is a superior model for further exploration of the optimal in

vivo doses of both G418 and enhancer compounds (see chapter 5).

111

Chapter 4: Generation of FTLD-GRN hiPSC lines and endogenous

PTC readthrough of GRN nonsense mutant hiPSC-derived CNS cell

types

4.1 Introduction

For decades ethical controversy and limited availability of human embryonic stem cells

limited their basic science research applications. As discussed in section 1.2.5, over a

decade ago, Dr. Shinya Yamanaka published an elegant study demonstrating the first

instance of adult somatic cell reprogramming into hiPSCs (195). This masterful work

has since sparked a revolution in our understanding of human developmental biology

and our ability to model human genetic disorders that originate in previously

unattainable tissues. The reprogramming of human somatic cells into hiPSCs is now

routine in labs worldwide, enabling rapid developments in personalized medicine and

improved investigation of pathological mechanisms driving complex disease processes.

Several methodologies have been established enabling somatic cell to hiPSC

reprogramming through the exogenous delivery of Yamanaka factors (OCT3/4, SOX2,

KLF4, MYC, and LIN28); the most common being Sendai viral transduction, non-

integrating episomal transfection, mRNA transfection, and protein transduction

(reviewed here (363,364)).

These technological advances have been particularly relevant to the fields of neurology

and neurodegeneration, as a multitude of hiPSC differentiation methods have been

112

developed that enable the derivation of distinct neuronal subtypes and glial cell types

(196-198). Decades of embryonic neurodevelopmental biology research provided the

foundational knowledge of growth factor patterning gradients required to guide the

establishment of hiPSC to neuroectoderm differentiation protocols. Chemical inhibition

of the anaplastic lymphoma kinase family has been shown to disrupt BMP4/TGF-β

signaling through the TGFβR-1 receptor and downstream activation of SMAD signal

transducers, promoting expression of the neuroectoderm lineage SOX1 transcription

factor while suppressing endoderm/mesoderm pathways (196). This approach, known

as dual-SMAD-inhibition, is widely used to derive pure populations NPCs for further

downstream differentiation/maturation into various CNS cell types.

hiPSC-derived NPCs neurodevelopmentally mirror the neuroepithelial cells that line the

apical surface of the ventricular zone and asymmetrically divide to produce radial glial

stem cells, which give rise to the majority of neuronal and glial cell types (365). Extrinsic

patterning cues in the form of recombinant protein growth factors (e.g., BDNF/GDNF)

are often provided to mature NPCs into the desired neuronal cell type. In 2015, Bardy et

al. generated a novel neuronal culture medium (BrainPhysTM) that more accurately

reflected neurophysiological ion concentrations and enhanced hiPSC-derived neuronal

electrophysiological maturation (352). Since obtaining human CNS neurologic tissue

primary cell cultures is severely limited by the risks and potential complications

associated with brain biopsies, the use of hiPSC-derived CNS tissue modelling has

accelerated research into human neurobiology.

113

The field of molecular biology was radically transformed following the landmark 2012

publication by Dr. Doudna and Dr. Charpentier, demonstrating that the bacterial

adaptive immune system (CRISPR/Cas9) could be co-opted for targeted genetic

engineering (366). Briefly, gRNA (dual complex of crRNA and tracrRNA) forms a

ribonucleoprotein complex with the Cas9 endonuclease, and the crRNA-guide

sequence localizes the complex to a complementary sequence of the genome triggering

double-strand break formation. Utilization of CRISPR/Cas9 gene editing is ubiquitous

throughout molecular biology research, with hiPSC modelling efforts being a major

benefactor. Dissecting precise phenotypic contributions of disease-associated SNPs

has been hampered by inadequate control hiPSC lines often derived from healthy

and/or genetically related individuals, which introduces background genetic variation

that may either silence or amplify observed phenotypes. CRISPR/Cas9 gene editing

solves this dilemma, enabling the generation of isogenic control/mutant hiPSC lines with

identical genetic backgrounds.

To expand on the results presented in chapter 3, we sought to determine whether PTC

readthrough treatments could restore endogenous nonsense mutant GRN expression in

hiPSC-derived cortical neurons and astrocytes. To evaluate readthrough efficiency in

GRN nonsense mutant hiPSC-derived brain cells, we measured both GRN mRNA and

intracellular/extracellular PGRN protein levels using qPCR and immunoassays. We also

assessed whether restoring PGRN expression in mutant neurons could rescue mutant

lysosomal dysfunction.

114

4.2 Results

4.2.1 Generation of WT and FTLD-GRN patient-derived hiPSC lines

We generated hiPSC lines from erythroid progenitors isolated from peripheral blood

obtained from a healthy control (WT) and an FTLD-GRN patient carrying the g.2923 C >

T (p.R418X) GRN nonsense mutation (R418X+/-) (350). Additionally, our collaborators

provided a stock of a previously published hiPSC line produced from an FTLD-GRN

patient bearing the g.585 C > A (p.S116X) GRN nonsense mutation (S116X+/-) (60).

These pluripotent cell lines enabled us to produce disease-relevant CNS cell types to

assess whether the compounds identified in chapter 3 could also restore endogenous

nonsense mutant GRN expression.

4.2.2 Generation of an isogenic GRN R493X homozygous knock-in hiPSC line

using CRISPR/Cas9 gene editing

The development of isogenic hiPSC lines limits non-specific phenotypic differences from

background genetic variability when comparing two distinct hiPSC lines derived from

healthy control and mutation carriers. We utilized the CRISPR/Cas9 gene editing tool to

homozygously introduce the most common FTLD-GRN mutation, p.R493X, into our WT

hiPSC line (Figure 4-1, A-C). Clones positive for the successful introduction of the

HindIII restriction enzyme digest site (Figure, 4-1, B) were Sanger sequenced to confirm

seamless homologous recombination of the knock-in ssODN (Figure 4-1, C). The

resulting isogenic mutant clone, R493X-/- KI, is the first hiPSC model of the rare form of

NCL, CLN11. As expected, R493X-/- KI hiPSC-derived cortical neuronal cultures

115

immunofluorescently stained for PGRN exhibit limited expression levels compared to

the parental isogenic WT neurons (Figure 4-2).

4.2.3 Characterization of hiPSC line pluripotency

To characterize the hiPSC lines, we assessed their expression of pluripotent markers at

the protein and mRNA level, and their ability to differentiate into all three germ layers

(Figure 4-3, A-B). Colonies from the WT, S116X+/-, R418X+/-, and R493X-/- KI hiPSC

lines exhibited strong nuclear staining for the pluripotent transcription factors NANOG

and OCT4 (Figure 4-3, A). Moreover, these hiPSC lines possessed elevated expression

of pluripotent markers, LIN28, OCT3/4, SOX2, and NANOG mRNA transcripts (Figure

4-3, B). To functionally validate these hiPSC lines’ pluripotential, we conducted in vitro

differentiation into the primary germ layers. All four hiPSC lines could differentiate into

ectodermal, mesodermal, and endodermal tissue as indicated by SOX1, DESMIN, and

SOX17 expression, respectively (Figure 4-3, A). Since chromosome abnormalities are

known to occur during hiPSC reprogramming, we had the hiPSC lines that we

generated (WT, R418X+/-, and R493X-/- KI) karyotyped, confirming that they were

karyotypically normal (Figure 4-3, A). Karyotyping analysis of the S116X+/- hiPSC line

was normal as previously described (60).

116

117

Figure 4-1 Generation of isogenic CRISPR/Cas9 gene edited R493X-/- KI hiPSC line

from WT. (A) Schematic representation of the gene editing strategy using the

CRISPR/Cas9 system. (B) Clonal screening for the silent introduction of the AAGCTT

HindIII restriction enzyme site to identify potential R493X knock-in clones. (C)

Simultaneous Sanger sequencing of the R493 codon region in both GRN alleles in WT

and isogenic R493X-/- KI clone confirming homozygous introduction of the UGA

nonsense codon at codon 493 (red highlight) and silent deletion of the PAM site.

118

Figure 4-2 PGRN immunofluorescence in WT and R493X-/- KI hiPSC-derived

cortical neurons. R493X-/- KI hiPSC-derived cortical neurons (DIV 50) possess greatly

reduced and more diffuse PGRN expression (green) than the WT parental line. Cell

nuclei were counterstained with DAPI (blue). Scale bar, 10 µm.

119

Figure 4-3 Characterization of WT and GRN-deficient hiPSC lines. (A)

Immunofluorescence analysis of pluripotent markers in WT, S116X+/-, R418X+/-, and

R493X-/- KI hiPSC lines, and their respective normal karyotypes. In vitro trilineage

differentiation of WT, S116X+/-, R418X+/-, and R493X-/- KI hiPSC lines, cells were

immunostained for SOX1 (ectoderm), DESMIN (mesoderm), SOX17 (endoderm). Cell

nuclei were counterstained with DAPI (blue) except for SOX17. Scale bar, 50 µm. (B)

120

mRNA expression of pluripotent reprogramming factors in WT, S116X+/-, R418X+/-, and

R493X-/- KI hiPSC lines relative to the values in WT fibroblasts, as assessed by qPCR.

Values are shown as mean ± SEM.

121

4.2.4 Differentiation of WT and GRN-deficient hiPSC lines into cortical neurons

and astrocytes

We then differentiated the hiPSC lines into cortical neurons using a recently reported

version of the dual SMAD inhibition neuroectoderm induction protocol (351) in

combination with BrainPhysTM medium NPC neuronal maturation (352) (Figure 4-4, A).

The resulting neuronal cultures possessed high levels of purity, as indicated by the high

percentage of MAP2 positive (> 95%) and low percentage of GFAP positive cells (< 4%)

(Figure 4-4, B-D). This high degree of purity is likely attributable to the early addition of

the proprietary CultureOneTM supplement known to limit NPC proliferation during

neuronal maturation. Additionally, greater than 75% of cells expressed both FOXG1 and

TBR1, confirming these neurons express cortical layer VI markers (Figure 4-4, B, E,

and F). Synaptogenesis was demonstrated in these cultures by staining to analyze the

expression of synapsin and excitatory/inhibitory synaptic markers (Figure 4-5, A-B).

Interestingly, GRN mutant hiPSC-derived cortical neurons all exhibited significantly

lower VGLUT1+ excitatory synapse levels than WT (Figure 4-5, A-B). Astrocytes were

also differentiated from NPCs using an astrocyte differentiation/maturation kit protocol.

Mature astrocyte cultures (DIV 60+) derived from both WT and R493X-/- KI hiPSC lines

both expressed high levels (80%) of GFAP (Figure 4-4, G-H). Co-culturing cortical

neurons with WT astrocytes on MEA plates induced accelerated their

electrophysiological maturation and the formation of neural networks with frequent and

robust spontaneous action potentials measured by MEA (Figure 4-5, C and Figure 4-6,

A-B).

122

123

Figure 4-4 Differentiation and characterization of cortical neurons and astrocytes

derived from WT and GRN-deficient hiPSCs. (A) Brightfield images of the cells at

different stages of cortical neuronal differentiation. SB = SB 431542, LDN = LDN

193189, RI = Y-27632, BP = BrainPhysTM, N2 = N-2 supplement, B27 = B-27

supplement. (B) Representative immunofluorescence images of DIV 50 WT, S116X+/-,

R418X+/-, and R493X-/- KI hiPSC-derived cortical neurons stained for MAP2, GFAP,

TUJ1, FOXG1 (forebrain), and TBR1 (cortical layer VI). Cell nuclei were counterstained

with DAPI (blue). Scale bar, 50 µm. (C-F) Cells positive for MAP2, GFAP, FOXG1, and

TBR1 (neuronal, astrocytes, forebrain, and cortical layer VI markers, respectively) as a

percentage of DAPI+ cells. On average, ~900 cells were analyzed per replicate, in (C-F)

n = 3 independent cultures, 3 images per biological replicate; values are shown as

mean ± SEM. (G) Representative immunofluorescence images of DIV 60+ WT and

R493X-/- KI hiPSC-derived astrocytes stained for GFAP. Cell nuclei were counterstained

with DAPI (blue). Scale bar, 50 µm. (H) Cells positive for GFAP as a percentage of

DAPI+ cells. On average, ~400 cells were analyzed per replicate, n = 3 independent

cultures, 3 images per biological replicate; values are shown as mean ± SEM.

124

125

Figure 4-5 Characterization of hiPSC-derived cortical neuron synaptic

development and neural network formation. (A) Representative immunofluorescence

images of DIV 50 WT, S116X+/-, R418X+/-, and R493X-/- KI hiPSC-derived cortical

neurons stained for synapsin, VGLUT1 (excitatory), and GAD65 (inhibitory). Cell nuclei

were counterstained with DAPI (blue). Scale bar, 20 µm. (B) Quantification of the

synaptic density (excitatory/inhibitory) per 100 µm2 area of TUJ1+ staining in DIV 50

WT, S116X+/-, R418X+/-, and R493X-/- KI hiPSC-derived cortical neuron cultures. n = 3

independent cultures, 3 images per biological replicate; values are shown as mean ±

SEM; *** p < 0.0001, one-way ANOVA with Tukey’s multiple comparison test. (C)

Electrophysiological properties of DIV 50 WT, S116X+/-, R418X+/-, and R493X-/- KI

hiPSC-derived cortical neurons co-cultured with WT human astrocytes were recorded

via MEA to quantify the number of spikes and bursts detected over a 10 min interval. n

= 4 independent cultures; values are shown as mean ± SEM.

126

Figure 4-6 Co-culturing hiPSC-derived cortical neurons with hiPSC-derived

astrocytes accelerates electrophysiological maturation and neural network

formation. Weekly electrophysiological recordings were performed on WT astrocytes,

WT cortical neurons, and WT cortical neurons + WT astrocytes MEA plate cultures.

Recordings were obtained via MEA to quantify the number of spikes (A) and network

bursts (B) detected over a 5 min interval. WT astrocytes alone were excluded from

network burst analysis due to lack of electrical activity detected in (A). n = 4

127

independent cultures; values are shown as mean ± SEM, *** p < 0.001, ** p < 0.01, p *

< 0.05, two-way ANOVA with Bonferroni post-tests.

128

4.2.5 PTC readthrough in GRN-deficient hiPSC-derived cortical neurons

hiPSC-derived cortical neurons were matured to DIV 50 and exposed to PTC

readthrough compounds for 72 h. The levels of WT neuronal full-length PGRN

expression were not sufficiently high for detection by western blotting with available anti-

PGRN antibodies (Figure 4-7, A-B and Figure 4-8). As an alternative, intracellular and

concentrated extracellular fractions were assayed by ELISA using a PGRN polyclonal

antibody to multiple epitopes (GRN 1/3/5/7 peptides) that detects not only full-length

PGRN but also truncated PGRN as well as granulin (GRN) peptides, collectively

referred to here as PGRN/GRNs. Vehicle-treated S116X+/-, R418X+/-, and R493X-/- KI

neurons expressed 33.1% ± 1.5%, 49.9% ± 1.3%, and 74.5% ± 1.2% less intracellular

PGRN/GRNs than vehicle-treated WT neurons, as expected (Figure 4-9, A-D). In

general, treatment of FTLD-GRN mutant neurons with G418 without or with CDX5-288

produced a similar pattern of PGRN/GRNs expression to that observed in the stably

transfected HEK293 cell lines bearing the same mutations. Again, the S116X+/- neurons

exhibited little to no intra- or extracellular increase in the levels of PGRN/GRNs in

response to PTC readthrough treatment (Figure 4-9, A, E). R418X+/- neurons exposed

to G418 alone or to G418 and CDX5-288 showed significantly increased intracellular

PGRN/GRNs levels to 67.5% ± 1.2% and 75.8% ± 5.8% of vehicle-treated WT levels,

respectively (Figure 4-9, B). Treatment of R493X-/- KI neurons with G418 alone

significantly increased intracellular PGRN/GRNs levels, to 54.0% ± 5.5% of vehicle-

treated WT levels, and exposure to G418 and CDX5-288 further significantly increased

intracellular PGRN/GRNs levels compared G418 alone, to 83.9% ± 1.3% of vehicle-

treated WT (Figure 4-9, C). For both R418X+/- neurons and R493X-/- KI neurons,

129

extracellular PGRN/GRNs levels were not significantly increased in response to

treatment, although a trend toward a minor increase was observed (Figure 4-9, F-G).

Since R493X-/- KI neurons have no intact GRN allele, the observed increase in

intracellular PGRN/GRNs is derived from the mutant alleles. G418 did not show

significant cellular toxicity in iPSC-derived neurons (Figure 4-10).

130

Figure 4-7 Baseline expression and secretion of PGRN in WT hiPSC-derived

cortical neurons and astrocytes. PGRN expression is significantly greater in the intra-

(A) and extracellular (B) fractions of WT hiPSC-derived astrocytes compared to cortical

neurons. WT and R493X-/- KI hiPSC-derived cortical neurons and astrocytes were

cultured in fresh medium for 72 h. Intracellular (1 mg/mL lysate) and extracellular (25X

concentrated supernatant) samples were subjected to PGRN ELISA. Values are shown

as mean ± SEM; *** p < 0.0001, Student’s t-test.

131

Figure 4-8 Detecting PGRN in WT and R493X-/- KI hiPSC-derived cortical neurons

by western blot. Expression of intracellular PGRN was assessed in DIV 80 WT and

R493X-/- KI hiPSC-derived cortical neuron cultures treated with G418, combination with

CDX5-288, or rec. human PGRN. Expression of intracellular ~70 kDa PGRN was only

detected in WT and R493X-/- KI neurons treated with 1 µg/mL rec. human PGRN. GRN-

2,3 peptides were only detected in WT neurons treated with rec. human PGRN.

Neuronal lysates were analyzed by western blotting, using actin as the loading control.

ns = nonspecific.

132

133

Figure 4-9 Induction of PTC readthrough by G418 and enhancers in hiPSC-

derived cortical neurons bearing FTLD-GRN nonsense mutations. DIV 50 S116X+/-,

R418X+/-, R493X-/- KI, and WT hiPSC-derived cortical neurons were treated with G418

and CDX5-288 at the indicated concentrations for 72 h. Intracellular (A-D, 1 mg/mL

lysate) and extracellular (E-H, 25X concentrated supernatants) samples were subjected

to PGRN ELISA using a polyclonal anti-PGRN antibody that targets multiple epitopes

(GRN 1/3/5/7 peptides) and cannot differentiate between PGRN/GRNs. n = 3

independent cultures; values are shown as mean ± SEM; * p < 0.05, ** p < 0.01, *** p <

0.0001, one-way ANOVA with Tukey’s multiple comparison test.

134

Figure 4-10 Quantification of G418-induced neuronal cytotoxicity. R418X+/- hiPSC-

derived neurons treated with escalating doses of G418 for 120 h with fresh media/drug

solutions replaced after 72 h. Cultures were stained with Hoechst dye and counted

using a Cellomics ArrayScanTM device. n = 3 independent cultures, 12 images per

biological replicate; values are shown as mean ± SEM.

0

2000

4000

6000A

vg

. # o

f cells /

fie

ld

G418 (g/mL) - 10 30 100

135

qPCR analysis revealed that readthrough compounds significantly increased GRN

mRNA levels in R493X-/- KI neurons. This observation was anticipated as mRNAs

bearing nonsense mutations are targeted for degradation by nonsense-mediated mRNA

decay (NMD) (233), and PTC readthrough enables escape of nonsense mutant mRNA

from degradation by NMD. Vehicle-treated R493X-/- KI neurons expressed 5.1-fold less

GRN mRNA relative to vehicle-treated WT neurons (Figure 4-11), indicating

downregulation of mutant mRNA by NMD. When treated with G418 or combination,

R493X-/- KI neurons express increased GRN mRNA levels (Figure 4-11). The selective

stabilization of PTC-bearing GRN mRNA over WT GRN mRNA strongly supports

inhibition of NMD consequent to readthrough in R493X-/- KI neurons. Taken together,

these results show that the G418 - CDX5-288 combination induces PTC readthrough

and effectively restores intracellular levels of PGRN in both heterozygous and

homozygous GRN nonsense mutant neurons to levels approaching that of healthy

control cells.

136

Figure 4-11 Selective increase of nonsense mutant GRN mRNA in R493X-/- KI

hiPSC-derived cortical neurons in response to PTC readthrough treatments.

qPCR analysis of human GRN mRNA levels in DIV 50 WT and R493X-/- KI hiPSC-

derived cortical neurons treated with G418 and CDX5-288 at the indicated

concentrations for 72 h. Relative mRNA levels were normalized to the mean of GAPDH

and HPRT1 housekeeping genes and compared to vehicle-treated (VT) WT. n = 3

independent cultures; values are shown as mean ± SEM; * p < 0.05, ** p < 0.01, *** p <

0.0001, one-way ANOVA with Tukey’s multiple comparison test.

137

4.2.6 PTC readthrough rescues lysosomal dysfunction in R493X-/- KI hiPSC-

derived cortical neurons

The specific nonsense mutant codon (e.g., R493X-/- KI, Arg-CGA to X-UGA) restricts

which near-cognate aminoacyl-tRNA molecules (Trp UGG, Cys UGC, Arg CGA, etc.)

pair with the PTC in the ribosome A site and contribute their amino acid to the growing

polypeptide during a readthrough event. Previous studies have demonstrated that the

amino acids most commonly incorporated at UGA nonsense mutations during

readthrough were Trp, followed by Cys and Arg (367). Thus, it is likely that a proportion

of full-length PGRN derived from GRN R493X-/- KI PTC readthrough would possess a

missense mutation at codon 493. Therefore, it was important to assess the functionality

of PGRN produced by PTC readthrough through its ability to rescue known cellular

FTLD-GRN and NCL phenotypes.

As discussed in sections 1.2.4 and 1.2.6, lysosomal defects are among the earliest

disease phenotypes in Grn-/- mice. Grn-/- brains upregulate the expression, maturation

rate, and catalytic activity of cathepsin lysosomal proteases (cathepsin D, B, and L) in

an age-dependent manner (119), and this aberrant lysosomal phenotype has been

previously rescued in vivo by treating aged Grn-/- mice with AAV-mediated Grn gene

therapy (218). Moreover, hiPSC-derived neurons with heterozygous GRN mutations

have been shown to exhibit an increased CTSD maturation phenotype (200). We

quantified the CTSD levels in aged (DIV 80) WT and R493X-/- KI hiPSC-derived cortical

neurons and observed significantly increased expression of the mature form of CTSD in

vehicle-treated R493X-/- KI neurons compared with WT neurons (Figure 4-12, A-B).

138

Treating WT neurons with G418, G418 + CDX5-288, or rec. PGRN did not affect mature

CTSD expression levels (Figure 4-12, A-B). However, treating R493X-/- KI neurons with

G418 and CDX5-288 significantly reduced mature CTSD levels, thus rescuing their

dysregulated lysosomal enzyme phenotype. Importantly, this effect was also observed

in R493X-/- KI neurons treated with extracellular rec. PGRN with a trend towards

reduced mature CTSD levels (Figure 4-12, A-B). These findings provide evidence that

PGRN expression restored by PTC-readthrough in R493X-/- KI cortical neurons (Figure

4-8, C) is biologically active and functional in vitro.

139

Figure 4-12 GRN PTC readthrough rescues FTLD-GRN/CLN11 lysosomal

pathological CTSD maturation phenotype in R493X-/- KI hiPSC-derived cortical

neurons. (A) DIV 80 WT and R493X-/- KI hiPSC-derived cortical neurons were treated

with vehicle solution, G418 alone, G418 in combination with CDX5-288, and rec. human

PGRN at the indicated concentrations for 72 h. Expression of mature CTSD in treated

WT and R493X-/- KI cortical neuron lysates analyzed by western blotting, using actin as

140

the loading control. (B) Densitometric quantification of CTSD expression in the

aforementioned cortical neuron lysates normalized to vehicle-treated WT levels. n = 3-6

independent cultures; values are shown as mean ± SEM; * p < 0.05, ** p < 0.01, one-

way ANOVA with Tukey’s multiple comparison test.

141

4.2.7 PTC readthrough in R493X-/- KI hiPSC-derived astrocytes

Next, WT and R493X-/- KI hiPSC-derived astrocytes were matured to DIV 60+ to extend

testing of readthrough compounds to another relevant CNS cell type with endogenous

nonsense mutant GRN expression. WT hiPSC-derived astrocytes produced 17.9-fold

more intracellular PGRN/GRNs, secreted 25.0-fold more PGRN/GRNs, and expressed

3.7-fold higher GRN mRNA than WT hiPSC-derived neurons as measured by ELISA

and qPCR (Figure 4-6 and 4-13). This considerably higher expression level enabled

analysis of intracellular PGRN levels by western blotting in hiPSC-derived astrocyte

cultures (Figure 4-14). The PGRN antibody used in this western blot analysis has

previously been shown to detect full-length PGRN and GRNs (GRN-2,3) (118). Both

PGRN and GRN-2,3 peptides were detected in WT astrocytes, though the GRN-2,3

peptides were more abundant than full-length PGRN (Figure 4-14, A). R493X-/- KI

astrocytes expressed small amounts of a band at ~70 kDa that may be PGRN truncated

at R493, and no detectable GRN-2,3 peptides (Figure 4-14, A).

Treating R493X-/- KI astrocytes with either G418 alone or G418 and CDX5-288

significantly increased intracellular ~70 kDa PGRN by 3.3- and 4.6-fold, respectively

(Figure 4-14, A, B.i). This increase in intracellular ~70 kDa PGRN was accompanied by

a small increase in GRN-2,3 peptide signal, which indicates that at least some of the

protein entered lysosomes and was cleaved into individual GRNs (Figure 4-14, A, B.ii).

Exposing R493X-/- KI astrocytes to extracellular rec. PGRN showed that these cells

could take up and process PGRN, as the vast majority of endocytosed full-length PGRN

was converted into GRNs (Figure 4-14, A-B). Since the conversion rate of full-length

142

PGRN to GRN-2,3 peptides is so efficient in astrocytes, we hypothesized that full-length

PGRN generated by readthrough would be rapidly processed into GRNs, perhaps

explaining why we observe limited full-length PGRN in G418 and CDX5-288

combination-treated R493X-/- KI astrocytes. The lysosomal cysteine protease CTSL

cleaves PGRN into GRNs (150). Therefore, we tested whether inhibiting CTSL during

G418 and CDX5-288 induced PTC readthrough in R493X-/- KI astrocytes could enable

clear visualization of full-length PGRN by western blot. Co-treating R493X-/- KI

astrocytes with 30 µM Z-Phe-Phe-FMK (CTSL inhibitor), G418, and CDX5-288 led to

accumulation of PTC readthrough derived full-length PGRN (Figure 4-14, C-D),

consistent with our hypothesis.

143

Figure 4-13 qPCR analysis of GRN mRNA levels in WT hiPSC-derived cortical

neurons and astrocytes. Relative GRN mRNA levels were measured in vehicle-

treated (VT) DIV 50 WT hiPSC-derived cortical neurons and DIV 60+ WT hiPSC-derived

astrocytes and were normalized to the mean of GAPDH and HPRT1 housekeeping

genes and compared to VT cortical neurons. n = 3 independent cultures; values are

shown as mean ± SEM; *** p < 0.0001, Student’s t-test.

144

145

Figure 4-14 Induction of PTC readthrough by G418 and enhancers in R493X-/- KI

hiPSC-derived astrocytes demonstrated by western blot. (A) R493X-/- KI hiPSC-

derived astrocytes were treated with vehicle solution, G418 alone, G418 in combination

with CDX5-288, and rec. human PGRN at the indicated concentrations for 72 h.

Expression of PGRN and GRN-2,3 peptides in treated WT and R493X-/- KI astrocyte

samples were analyzed by western blotting, using actin as the loading control. (B)

Densitometric quantification of ~70 kDa PGRN (i) and GRN-2,3 peptide (ii) in astrocyte

lysates (A) normalized to vehicle-treated WT levels. (C) R493X-/- KI hiPSC-derived

astrocytes were treated with vehicle solution, G418 in combination with CDX5-288, and

G418 CDX5-288 combination with either 10 or 30 µM of Z-Phe-Phe-FMK for 72 h.

Again, expression of PGRN in WT and R493X-/- KI astrocyte lysates was also analyzed

by western blotting, using actin as the loading control. n = 3 independent cultures;

values are shown as mean ± SEM; p < 0.05, ** p < 0.01, *** p < 0.0001, one-way

ANOVA with Tukey’s multiple comparison test.

146

We next used ELISA to measure intra- and extracellular PGRN/GRNs levels in

astrocytes. Vehicle-treated R493X-/- KI astrocytes expressed 81.8% ± 1.6% less

intracellular PGRN/GRNs than vehicle-treated WT astrocytes (Figure 4-15, A-B).

R493X-/- KI astrocytes exposed to G418 alone or G418 and CDX5-288 showed

significantly increased intracellular PGRN/GRNs levels, to 75.7% ± 6.5% and 75.8% ±

3.2%, respectively, of vehicle-treated WT levels, only slightly less than PGRN/GRNs

restoration achieved through the application of exogenous rec. PGRN (Figure 4-15, B).

Vehicle-treated R493X-/- KI astrocytes secreted 94.9% ± 0.3% less PGRN/GRNs than

vehicle-treated WT astrocytes (Figure 4-15, C-D). Exposure to G418 alone or G418 and

CDX5-288 caused a considerable increase in secreted PGRN/GRNs, to 30.4% ± 1.7%

and 32.5% ± 1.2%, respectively, of vehicle-treated WT levels (Figure 4-15, C-D), unlike

R493X-/- KI cortical neurons, where treatment increased intracellular PGRN/GRNs but

only subtly increased PGRN/GRNs secretion.

As observed in cortical neurons, qPCR analysis found that PTC readthrough

compounds also significantly increased in GRN mRNA levels in R493X-/- KI astrocytes.

Vehicle-treated R493X-/- KI astrocytes expressed 3.3-fold less GRN mRNA than vehicle-

treated WT astrocytes (Figure 4-16). When treated with G418 or combination with

CDX5-288, R493X-/- KI astrocytes showed significantly increased GRN mRNA levels

compared to vehicle-treated control (Figure 4-16). These findings support of inhibition of

NMD consequent to PTC readthrough in R493X-/- KI astrocytes.

147

Figure 4-15 Induction of PTC readthrough by G418 and enhancers in R493X-/- KI

hiPSC-derived astrocytes demonstrated by ELISA. DIV 60+ WT and R493X-/- KI

hiPSC-derived astrocytes were treated with G418, CDX5-288, and rec. human PGRN at

the indicated concentrations for 72 h. Intracellular (A-B, 1 mg/mL lysate) and

extracellular (C-D, 25X concentrated supernatant) samples were subjected to PGRN

ELISA using a polyclonal anti-PGRN antibody that targets multiple epitopes (GRN

1/3/5/7 peptides). Therefore, the antibody cannot differentiate between PGRN/GRNs. n

= 3 independent cultures; values are shown as mean ± SEM; * p < 0.05, ** p < 0.01, ***

p < 0.0001, one-way ANOVA with Tukey’s multiple comparison test.

148

Figure 4-16 Selective increase of nonsense mutant GRN mRNA in R493X-/- KI

hiPSC-derived astrocytes in response to PTC readthrough treatments. qPCR

analysis of human GRN mRNA levels in DIV60+ WT and R493X-/- KI hiPSC-derived

astrocytes treated with G418 and CDX5-288 at the indicated concentrations for 72 h.

Relative GRN mRNA levels were normalized to the mean of GAPDH and HPRT1

housekeeping genes and compared to vehicle-treated (VT) WT. n = 3 independent

cultures; values are shown as mean ± SEM; * p < 0.05, ** p < 0.01, *** p < 0.0001, one-

way ANOVA with Tukey’s multiple comparison test.

149

4.2.8 Impact of G418 on PGRN expression in WT hiPSC-derived cortical neurons

and astrocytes

Contrary to expectation, we observed that WT cortical neurons treated with G418 alone

and in combination with CDX5-288 showed significantly increased intracellular

PGRN/GRNs levels relative to vehicle-treated cells, by 37.9% ± 3.2% and 42.9% ± 3.7,

respectively (Figure 4-8, D). This intracellular increase in WT neuronal PGRN/GRNs

levels was not due to PTC readthrough as both GRN alleles are WT. CDX5-288 alone

did not increase intracellular WT neuronal PGRN/GRNs, and its combination with G418

did not increase PGRN/GRNs over G418 treatment alone, indicating the observed

increase is mediated by G418 (Figure 4-8, D). The G418-mediated increase in

intracellular WT neuronal PGRN/GRNs expression was not accompanied by a

corresponding increase in PGRN/GRNs secretion (Figure 4-8, H). These findings in WT

neurons suggest that G418 may disrupt PGRN exocytosis. The potential mechanisms

driving this phenomenon are addressed in more detail in the context of WT astrocytes

(Figure 4-15 and Figure 4-16).

G418 increased intracellular PGRN/GRNs in WT astrocytes by 31.7% ± 5.6%, without a

corresponding increase in secreted PGRN/GRNs (Figure 4-15, A, C), as was observed

with WT cortical neurons. Western blotting revealed that G418 significantly increased

the intracellular levels of PGRN while decreasing the levels of GRN-2,3 peptides in WT

astrocytes (Figure 4-17). Therefore, G418 may disrupt the normally highly efficient

lysosomal processing of full-length PGRN to GRNs in astrocytes. To further probe the

mechanism of this unanticipated readthrough-independent increase in WT intracellular

150

PGRN, we conducted GRN qPCR analysis in WT hiPSC-derived neurons and

astrocytes. Treating WT neurons or astrocytes with G418 alone or in combination with

CDX5-288 did not significantly increase GRN mRNA levels (Figure 4-10 and Figure 4-

16). We speculate that G418 affects PGRN homeostasis at the protein level resulting in

reduced lysosomal processing of full-length PGRN into individual GRNs.

151

Figure 4-17 G418-mediated disruption of PGRN homeostasis in WT hiPSC-derived

astrocytes. G418 disrupts PGRN homeostasis in WT hiPSC-derived DIV 60+ astrocyte

cultures. (A) Expression of intracellular full-length PGRN and GRN-2,3 peptides in

G418, combination, and rec. human PGRN treated WT astrocyte lysates analyzed by

western blotting, using actin as the loading control. Densitometric quantification of

PGRN (B) and the ratio of PGRN:GRN-2,3 peptide (C) expression in the

152

aforementioned astrocyte lysates normalized to vehicle-treated WT levels. n = 3

independent cultures; values are shown as mean ± SEM; * p < 0.05, ** p < 0.01, one-

way ANOVA with Tukey’s multiple comparison test.

153

4.3 Discussion

As the founding graduate student researcher in the Nygaard laboratory, it was my

responsibility to establish novel cellular models for the study of FTLD-GRN. A critical

aim of my thesis was to assess the potential of PTC readthrough drugs to restore

endogenous nonsense mutant GRN expression in disease-relevant CNS cell types;

therefore, it was essential to generate and characterize healthy control- and FTLD-GRN

patient-derived hiPSC lines. WT and R418X+/- hiPSC lines were produced from healthy

control and a presymptomatic member of a Canadian family of GRN haploinsufficient

mutation carriers (UBC15), respectively. Our lab resource publication of the R418X+/-

hiPSC line in the journal of Stem Cell Research represents the contribution of a novel

research tool to the field of FTLD (350). Since the phenotypes reported in past studies

of GRN+/- hiPSC-derived neurons are not particularly robust (see section 1.2.5), we

used CRISPR/Cas9 to generate a WT hiPSC-derived isogenic homozygous GRN

R493X hiPSC line (R493X-/- KI). Additionally, the isogenicity of the WT and R493X-/- KI

hiPSCs provides greater confidence in attributing phenotypic differences to a lack of

GRN expression. Moreover, the homozygous nature of this knock-in hiPSC line ensured

that any increase in full-length PGRN expression detected in response to PTC

readthrough treatment was derived from a nonsense allele and not the intact allele

present in heterozygous hiPSC lines (S116X+/- or R418X+/-).

Considerable efforts were undertaken to optimize and identify an hiPSC neuroectoderm

induction method that reproducibly generated high-quality homogenous NPC cultures

from various hiPSC lines derived from either fibroblasts or erythroid progenitor cells.

154

This specific variant of dual-SMAD inhibition was adapted from a 2018 Rose et al. study

(351) and has been validated on at least eight different hiPSC lines in our lab. This

protocol’s key benefit was its ability to rapidly (3.5 weeks) produce 1E8+ NPCs starting

from a single 6-well plate well of confluent hiPSCs that could be frozen down for long-

term storage and thawed for future experiments requiring neuronal/astrocyte

differentiation. Thawed NPCs plated in BrainPhysTM supplemented with CultureOneTM

immediately initiated neurite sprouting and neural network formation, and within just 3

weeks, cultures were positive for forebrain and cortical neuron markers FOXG1 and

TBR1. Furthermore, the inclusion of CultureOneTM likely contributed to the minimal

accumulation of GFAP+ cells. This may be achieved through a mechanism similar to

previous differentiation efforts utilizing DAPT (γ-secretase and Notch pathway inhibitor)

to inhibit NPC proliferation and increase neuronal culture purity (368). A secondary

benefit of generating large quantities of frozen NPC stocks from each hiPSC line was

the ability to differentiate astrocytes from the same cells used to produce cortical

neurons. Astrocytes differentiated using the commercial STEMCELL Technologies

differentiation/maturation kit expressed high levels of GFAP. Importantly, we also

observed that the inclusion of WT astrocytes with WT cortical neurons in MEA co-

cultures significantly accelerated their electrophysiological development (Figure 4-5, A-

B); these hiPSC-derived astrocytes functionally support neuronal maturation as

expected. The generation of these hiPSC lines and optimization of the neural lineage

differentiation method was integral, enabling further assessment of their responsiveness

to nonsense suppression therapy.

155

Here, we showed that PGRN insufficiency caused by GRN nonsense mutations can be

ameliorated through PTC readthrough in disease-relevant cell types and that the

resultant increase in PGRN reverses FTLD-GRN related lysosomal dysfunction. We

used a commercially available ELISA kit that detects PGRN at multiple epitopes (GRN-

1, 3, 5, & 7) to measure intra- and extracellular of cortical neuron PGRN expression

levels, precluding discrimination between truncated and full-length forms of PGRN or

GRNs. This is because we could not detect full-length PGRN in hiPSC-derived cortical

neurons by western blot, likely due to their low levels of basal PGRN expression or

rapid processing into smaller GRNs. In contrast, we showed that full-length and

truncated PGRN is measurable by western blot in human astrocytes, allowing more

precise measurements of PGRN protein of various sizes, including full-length, truncated

forms, and GRNs. In response to G418 and CDX5-288, there was a significant increase

in truncated and likely full-length PGRN in R493X-/- KI astrocytes, as well as an increase

in GRN-2,3. This increase reflects a greater than 6-fold increase in GRN mRNA in

response to G418 and CDX5-288 combination treatment. PTC readthrough treatment

only restored R493X-/- KI neuronal GRN mRNA to WT levels, while astrocytes exhibited

> 2-fold increase in R493X GRN mRNA over WT levels, perhaps reflecting their overall

elevated basal GRN expression level. This increase in mRNA is characteristic of escape

from NMD, in which nonsense mutant mRNA undergoes its pioneering round of full-

length translation without stalling at its PTC.

This process is thought to be regulated by the degree of PTC readthrough, and only a

minor increase in translation beyond a proposed 0.5% threshold would be sufficient for

156

substantial NMD inhibition and increased translation, as shown in yeast (369). G418

has been previously demonstrated to inhibit NMD through this readthrough-dependent

mechanism (370). Stabilization of R493X mutant GRN mRNA leads to accumulation of

truncated protein, which was evident in R493X-/- KI astrocytes treated with both G418

alone and CDX5-288 combination. Moreover, the increase in ~70 kDa PGRN signal

detected in the R493X-/- KI astrocytes was accompanied by a faint, slightly higher MW

band (Figure 4-14, A), which likely represents full-length PGRN. Given that both WT

and R493X-/- KI astrocytes rapidly and efficiently process PGRN into individual GRNs,

likely, the majority of PTC readthrough-derived full-length PGRN would also be cleaved

into GRNs, thus making full-length PGRN particularly difficult to detect. This hypothesis

is supported by the observed increase in full-length PGRN in R493X-/- KI astrocytes in

response to readthrough treatment when the major lysosomal PGRN processing

enzyme CTSL is inhibited. Thus, suggesting that the increase in intracellular

PGRN/GRNs in R493X-/- KI cortical neurons likely reflects a mixture of truncated and

full-length PGRN. These findings highlight the value of conducting drug screening in

multiple cell types to evaluate differential drug responses, informing cell type-specific

therapeutic targeting methods.

Interestingly, treating R493X-/- KI cortical neurons with G418 and CDX5-288

combination rescued aberrant lysosomal function in these cells, reducing mature CTSD

levels back to WT levels. Recent studies have discovered a physical interaction

between PGRNs C-terminus and CTSD (200,202,203). Despite some conflicting reports

(119), the growing consensus is that PGRN binds to the pro-form of CTSD, promoting

157

its maturation, thus increasing its enzymatic activity (200,208). However, this hypothesis

is contradicted by the observation that mature CTSD is upregulated in Grn-/- mice

brains/microglia and can be rescued with exogenous mouse rec. Pgrn (180,218).

Nevertheless, our results suggest that PGRN generated by PTC RT can perform

biological functions and provides supplementary validation for at least a portion of the

readthrough derived neuronal PGRN being full-length. It has also been recently

demonstrated that the truncated PGRN R493X protein retains some of its biological

properties, such as lysosomal localization and an ability to suppress a proinflammatory

immune response in Grn-/- mice bone marrow-derived macrophage cultures (165).

Therefore, we cannot rule out a contribution of truncated PGRN in the correction of this

excess mature CTSD phenotype.

Unexpectedly, G418 induced accumulation of WT endogenous intracellular PGRN in all

of our in vitro models. Given that G418 increases WT PGRN levels, it was difficult to

prove that PTC readthrough was responsible for the elevated PGRN expression

detected in heterozygous R418X+/- cortical neurons, as any increase in PGRN could be

attributable to either increased expression from the WT allele or PTC-readthrough of the

nonsense mutant allele. This provided the rationale for developing a homozygous

R493X-/- KI line to address this issue and establish that at least a portion of the

increased PGRN expression we observed in R418X+/- neurons were likely reflective of

PTC readthrough. GRN expression analysis in WT cortical neurons and astrocytes

treated with G418 confirmed that the increase in intracellular PGRN is not due to an

accumulation of GRN mRNA, suggesting the mechanism is post-translational in nature.

158

The G418-mediated intracellular PGRN increase in WT neurons and astrocytes does

not result in corresponding accumulation of PGRN in culture media, implying that G418

may interfere with the PGRN secretory pathway. Further studies are needed to better

understand the possible post-translational role of G418 in PGRN processing.

159

Chapter 5: Phenotypic characterization of the GrnR493X/R493X mouse

model and dose-limiting in vivo toxicity of G418

5.1 Introduction

The neuropathology observed in patients bearing GRN LOF mutations is dictated by a

gene dosage-dependent effect, with haploinsufficient individuals developing FTLD-GRN

(35,36) and those lacking intact GRN alleles suffering from a rare form of NCL, CLN11

(95,96). The discovery that GRN-null individuals develop a lysosomal storage disease

has encouraged investigation into the role of PGRN and GRNs in regulating lysosomal

function (see section 1.2.4.2). The majority of the known neurobiological functions of

PGRN have been uncovered using Grn-/- mouse models, partially because models of

Grn haploinsufficiency do not replicate many of the neuropathological hallmarks

observed in either FTLD-GRN or CLN11. Microglial lysosomal dysfunction and neuronal

lipofuscin accumulation are the earliest pathological phenotypes observed in the Grn-/-

mouse brain (119,179), preceding well-established neuroinflammatory and

neurodegenerative phenotypes (see sections 1.2.4.1 and 1.2.4.4).

Previous in vivo FTLD-GRN therapeutic intervention studies have aimed to identify

drugs capable of upregulating the intact allele and thus had to be performed in Grn+/-

mice (135). The GrnR493X mouse model was recently generated to more accurately

model FTLD-GRN by introducing one of the most frequent GRN mutation (R493X) at

the analogous mouse Grn codon (R504X), providing an ideal model for preclinical

160

validation of novel PTC readthrough therapies. Characterization of this nonsense

mutant Grn model identified several previously observed Grn-/- disease phenotypes,

including microgliosis, lipofuscin accumulation, synaptic loss, and TDP-43 pathology

(165). We sought to comprehensively characterize behavioural and neuropathological

phenotypes in aged GrnR493X/R493X mice and attempt to restore CNS Pgrn expression

through PTC readthrough treatment. These findings provide additional phenotypic

markers of pathogenesis in aged GrnR493X/R493X mice that will contribute to better

defining mechanisms underlying FTLD-GRN, and offer relevant outcome measures for

preclinical efficacy testing of novel therapeutics that target nonsense mutations leading

to this devastating disease.

5.2 Results

5.2.1 Pgrn expression in the brains of aged GrnR493X/R493X mice

We obtained the previously generated FTLD-GRN/CLN11 mouse model bearing

homozygous Grn R504X mutations analogous to the human GRN R493X mutation for

further neuropathological and behavioural characterization (165). Homozygous

introduction of the GrnR493X PTC in these mice was confirmed by PCR (Figure 5-1).

Brain Pgrn expression in 18 month old Grn+/+ and GrnR493X/R493X mice was assayed using

multiple immunological detection methods, including western blot (Figure 5-2, A), ELISA

(Figure 5-2, B), and immunofluorescence microscopy (Figure 5-2, C-D). These results

demonstrate that nonsense mutant Pgrn expression is significantly reduced, detecting

GrnR493X/R493X global Pgrn brain expression levels of 14.7% ± 1.7% (ELISA) and 21.7% ±

161

2.5% (western blot) relative to Grn+/+ expression levels. Immunofluorescent

quantification of Pgrn expression in hippocampal CA3 and thalamic VPM/VPL

GrnR493X/R493X pathological hotspots identified even lower levels of Pgrn, with total Pgrn

area per cell of 4.3% ± 0.7% (CA3) and 4.6% ± 0.7% (VPM/VPL) of Grn+/+ tissue levels.

162

Figure 5-1 GrnR493X/R493X genotyping. PCR genotyping results confirm the presence of

the homozygous GrnR493X knock-in alleles in GrnR493X/R493X mice. nt = no template, bp =

base pair.

163

Figure 5-2 Pgrn expression in the brains of aged GrnR493X/R493X mice. 18 month old

Grn+/+ and GrnR493X/R493X hemibrain RIPA-soluble lysate samples were subjected to both

Pgrn western blot (A) and Pgrn ELISA (B). Densitometric quantification of Pgrn western

blot signal was normalized to actin loading control. (C) Pgrn expression was also

detected in the hippocampal CA3 and thalamic VPM/VPL regions by

164

immunofluorescence staining of 18 month old Grn+/+ and GrnR493X/R493X brain sections

(scale bar, 20 µm). (D) Total Pgrn+ area in CA3 (i) and VPM/VPL (ii) was quantified and

normalized to the total number of cells (DAPI, blue). For western blot/ELISA n = 6 mice

were used per sex/genotype; for immunofluorescence staining n=10 mice were used

per sex/genotype (except male GrnR493X/R493X n=8); values are shown as mean ± SEM;

*** p < 0.0001, Student’s t-test.

165

5.2.2 Lysosomal dysfunction in the brains of aged GrnR493X/R493X mice

NCL pathophysiology is generally driven by disrupted lysosomal catabolic functions

resulting from lysosomal enzyme deficiencies. As discussed in section 1.2.4, a growing

consensus in the literature suggests that loss of PGRN directly and indirectly disrupts

both protein and lipid recycling pathways. Studies of Grn-deficient mouse models have

helped identify PGRN’s roles in maintaining autophagic homeostasis in the CNS

(107,119,120,178). To assess the potential efficacy of therapeutic interventions in our

GrnR493X/R493X mice, it was critical to first characterize the various disease phenotypes

that develop in this model. Additionally, since GrnR493X/R493X mice express low levels of a

semi-functional truncated form of Pgrn (Pgrn-R493X) (165), it was also important to

assess whether these mice pathologically mirror Grn-/- mice disease progression.

Previous studies in Grn-/- mice have found CNS lipofuscin accumulation occurs as early

as 2-3 months of age (179). We replicated the thalamic/hippocampal autofluorescent

lipofuscin phenotype reported in the seminal GrnR493X/R493X model characterization paper

(165), observing extensive lipofuscin in 18 month old mice in both the CA3 hippocampal

region and thalamus (Figure 5-3, A). Further, aged R493X knock-in mice exhibit

increased expression of lysosomal proteins Lamp1 and DppII, which were also shown

to accumulate in these brain regions (Figure 5-3, B-C). We quantified Lamp1+ and

DppII+ areas in the thalamus and hippocampus of aged GrnR493X/R493X mice and

observed a significant increase in the expression of these lysosomal markers in the CA3

and VPM/VPL brain regions (Figure 5-3, D-G). Increased lysosomal vesicle size has

been previously reported in several models of Grn-deficiency, including both Grn-/-

mouse hippocampal neurons (107) and in GrnR493X/+ primary cortical neurons (136). We

166

also found that GrnR493X/R493X CA3 and VPM/VPL regions accumulate significantly larger

Lamp1+ lysosomal vesicles than Grn+/+ control mice (Figure 5-4, A-B).

167

168

Figure 5-3 Lysosomal dysfunction in the ventral thalamus and CA3 hippocampal

region of aged GrnR493X/R493X mice. Representative hippocampal/thalamic tilescans

highlighting autofluorescent lipofuscin (A, green channel), Lamp1 (B), and DppII (C)

accumulation in the CNS of 18 month old GrnR493X/R493X mice (scale bar, 500 µm). (D-E)

Representative Lamp1 and DppII immunofluorescence images of hippocampal CA3 and

thalamic VPM/VPL regions in 18 month old Grn+/+ and GrnR493X/R493X brain sections

(scale bar, 20 µm). (F-G), Total Lamp1+ and DppII+ area/cell in CA3 (i) and VPM/VPL

(ii) images were quantified and normalized to the total number of cells (DAPI, blue). For

immunofluorescence staining n=10 mice were used per sex/genotype (except male

GrnR493X/R493X n=8); values are shown as mean ± SEM; ** p < 0.01, *** p < 0.0001 was

determined by Student’s t-test.

169

Figure 5-4 Enlarged lysosomes in the ventral thalamus and CA3 hippocampal

region of aged GrnR493X/R493X mice. Average Lamp1+ vesicle size was quantified in

images of Grn+/+ and GrnR493X/R493X CA3 (A) and VPM/VPL (B). For immunofluorescence

staining n=10 mice were used per sex/genotype (except male GrnR493X/R493X n=8);

values are shown as mean ± SEM; *** p < 0.0001 was determined by Student’s t-test.

170

Aged Grn-/- mice develop global CNS lysosomal dysfunction, including disrupted

autophagosome clearance (106), overexpression of lysosomal enzymes (119), and

cytoplasmic accumulation of insoluble TDP-43 (93). To evaluate CNS-wide lysosomal

dysfunction in this model, we conducted a series of western blot assays on RIPA-

soluble and -insoluble hemi-brain lysates. Unlike previous findings in Grn-/- mice (218),

Lamp1 protein levels were not found to be significantly increased in the GrnR493X/R493X

brain, although the data trended toward elevated Lamp1 expression (Figure 5-5, A-B).

Increased expression and enzymatic maturation of the lysosomal protease Ctsd has

been reproducibly observed in whole Grn-/- brain tissue (119). Expression of both the

pro- and mature forms of Ctsd were significantly increased in GrnR493X/R493X brain

compared to Grn+/+ (Figure 5-5, A, C). Lysosomal dysfunction can lead to functional

impairments in autophagy, an essential process in maintaining cellular health. Chang et

al. identified impaired ALP signaling in whole cortical tissue obtained from 18 month old

Grn-/- mice, including an approximate 35% increase in the LC3-II:LC3-I ratio, which

indicates autophagosome accumulation (106). Our similarly aged GrnR493X/R493X mice

exhibited a nearly 2-fold increase in LC3-II:LC3-I over Grn+/+ (Figure 5-5, A, D),

suggesting a dysregulated autophagy phenotype persists despite the low levels of

lysosomal accumulation of partially functional Pgrn-R493X that may occur in these mice

(165).

171

Figure 5-5 Global lysosomal dysfunction in the brains of aged GrnR493X/R493X mice.

Representative western blots of hemibrain RIPA-soluble lysates from 18 month old

Grn+/+ and GrnR493X/R493X mice probed with the indicated antibodies (A). Expression of

Lamp1, total/pro/mat-Ctsd, and LC3-I/II in Grn+/+ and GrnR493X/R493X hemibrains lysates

was analyzed by western blotting, using actin as the loading control. Densitometric

quantification of brain-wide Lamp1 (B) and Ctsd (C) expression was normalized to actin

and Grn+/+ levels. The LC3-II:LC3-I densitometric expression ratio (D) was normalized to

Grn+/+ levels. n=6 mice were used per sex/genotype; values are shown as mean ± SEM;

ns = not significant, ** p < 0.01, *** p < 0.0001, Student’s t-test.

172

Impaired autophagy has been linked to increases in pathological forms of TDP-43, a

pathologic hallmark of FTLD-GRN. Although Grn-/- mice develop a limited form

cytoplasmic/nuclear TDP-43 aggregation slightly resembling histopathology observed in

FTLD-GRN patients (371), biochemical analyses have identified increased full-length

and phosphorylated TDP-43 (p-TDP-43) expression in whole-brain RIPA-insoluble

fractions (93,371). Full-length TDP-43 can be cleaved into aggregation-prone CTFs that

form the major protein component of TDP-43+ inclusions (23,24). 12 month old

GrnR493X/R493X brains were found to contain diffuse cytoplasmic TDP-43/p-TDP-43

positivity similar to Grn-/- mice, that was absent in Grn+/+ mice (165). We reasoned that

older mice might display a more robust TDP-43 phenotype. Using a polyclonal N-

terminal TDP-43 antibody known to detect multiple forms of TDP-43, including full-

length and several truncated CTFs (372), we probed RIPA-soluble and -insoluble hemi-

brain lysates for TDP-43 expression. Full-length TDP-43 expression in the soluble

fraction was significantly decreased in GrnR493X/R493X mice (Figure 5-6, A-B), with a

similar trend observed for CTFs (Figure 5-6, A, arrows). Surprisingly, decreased soluble

TDP-43 expression did not correspond with an increase in insoluble TDP-43 levels

(Figure 5-6, A, C). We further probed these lysate fractions for Ser409 p-TDP-43

expression and observed a similar phenotype to that observed with TDP-43 total (Figure

5-7, A-C). As previously observed in Grn+/+ and Grn-/- brains (371), the insoluble

fractions of both Grn+/+ and GrnR493X/R493X exhibited higher levels of full-length p-TDP-43

compared to the soluble fraction (Figure 5-6, A and Figure 5-7, A). Since TDP-43

pathology in this model has previously been demonstrated using TDP-43

immunofluorescent staining (165,373), we co-stained Grn+/+ and GrnR493X/R493X brain

173

sections for Tuj1 and TDP-43 to assess neuronal TDP-43 cellular localization (Figure 5-

6, D-E). Tuj1-TDP-43 GrnR493X/R493X hippocampal/thalamic tilescans showed a major

increase in neuronal TDP-43 expression localized to the thalamic VPM/VPL regions

compared to Grn+/+. Similar to previous reports, high magnification micrographs of the

GrnR493X/R493X VPM/VPL revealed considerable cytoplasmic TDP-43 accumulation in

ventral thalamic neurons (Figure 5-6, E).

174

175

Figure 5-6 Neuronal TDP-43 proteinopathy is localized to the ventral thalamus of

aged GrnR493X/R493X mice. Representative western blots of hemibrain RIPA-soluble and

-insoluble lysates from 18 month old Grn+/+ and GrnR493X/R493X mice probed for TDP-43

expression. (A) Expression of full-length TDP-43 in Grn+/+ and GrnR493X/R493X hemibrains

in soluble and insoluble lysates was analyzed by western blotting, using RIPA-soluble

actin as the loading control (no actin detected in insoluble urea fraction). Arrows indicate

TDP-43 CTFs, and the * demarks a remnant TDP-43 signal observed upon reprobing

stripped RIPA-soluble TDP-43 blot with actin antibody. Densitometric quantification of

brain-wide full-length TDP-43 expression in soluble (B) and insoluble (C) lysate fractions

were normalized to RIPA-soluble actin and Grn+/+ levels. (D) Representative

hippocampal/thalamic tilescans highlighting the neuronal (TUJ1, green) TDP-43 (red)

proteinopathy phenotype in the ventral thalamus of 18 month old GrnR493X/R493X mice

(scale bar, 500 µm). (E) High magnification images from the thalamic VPM/VPL regions

demonstrating neuronal cytoplasmic accumulation of TDP-43 in 18 month old

GrnR493X/R493X mice (DAPI, blue; scale bar, 10 µm). For western analysis n=6 mice were

used per sex/genotype; values are shown as mean ± SEM; ns = not significant, ** p <

0.01, Student’s t-test.

176

Figure 5-7 Absence of global p-TDP-43 proteinopathy in the brains of aged

GrnR493X/R493X mice. Representative western blots of hemibrain RIPA-soluble and -

insoluble lysates from 18 month old Grn+/+ and GrnR493X/R493X mice probed for TDP-43

expression. (A) Expression of full-length p-TDP-43 in Grn+/+ and GrnR493X/R493X

hemibrains in soluble and insoluble lysates was analyzed by western blotting, using

RIPA-soluble actin as the loading control (no actin detected in insoluble urea fraction).

Arrows indicate TDP-43 CTFs, and the * demarks a remnant TDP-43 signal observed

upon reprobing stripped RIPA-soluble TDP-43 blot with actin antibody. Densitometric

quantification of brain-wide full-length p-TDP-43 expression in soluble (B) and insoluble

(C) lysate fractions were normalized to RIPA-soluble actin and Grn+/+ levels. n=6 mice

were used per sex/genotype; values are shown as mean ± SEM; ns = not significant, **

p < 0.01, Student’s t-test.

177

5.2.3 Neuroinflammation and astrogliosis in the ventral thalamus of aged

GrnR493X/R493X mice

Pathologic increases of both microglia and astrocytes in the CA3 hippocampal region

and ventral thalamus are well established phenotypes in Grn-deficient mice (see section

1.2.4.1). We sought to characterize neuroinflammation and astrogliosis in aged

GrnR493X/R493X mice. Nguyen et al. previously described a temporal increase in thalamic

microglial density in GrnR493X/R493X mice (165), and this was also observed in our aged

GrnR493X/R493X mice (Figure 5-8, A). The number of microglia per thalamic VPM/VPL field

was significantly greater in mutant mice (Figure 5-8, B, C.ii); however, no microgliosis

was observed in the hippocampal CA3 region (Figure 5-8, B, C.i). A common feature of

pro-inflammatory activated microglia is a morphological transition from a highly ramified

state to an amoeboid shape with enlarged soma (374). To quantify microglia

morphology, we conducted skeletal analysis to measure the number of Iba1+ branches

and normalized that to the number of microglia present in a given field to obtain the

average branches/microglia (Figure 5-9, A-B) (354). Decreased microglial branching

was observed in both the CA3 and VPM/VPL regions of the GrnR493X/R493X brain,

although the phenotype was more pronounced in the VPM/VPL, which exhibited 56.3%

± 6.6% of the branching levels observed in wildtype mice (Figure 5-8, B, C.iii-iv).

Counterintuitively, increased GRN expression has been detected in the frontal and

temporal cortices of FTLD-GRN patients (94). This observation is believed to result from

upregulation of the intact GRN allele in hyperactivated, proliferating microglia in

178

degenerating brain regions. In GrnR493X/R493X mice, upregulation of the mutated alleles

could also result in increased basal PTC readthrough which would result in increased

full length Pgrn expression. To test this hypothesis, we conducted Pgrn-Iba1 co-staining

in Grn+/+ and GrnR493X/R493X mice to assess whether diseased brain regions in the knock-

in mice display upregulated microglial Pgrn-R493X expression. We found that microglial

Pgrn fluorescent intensity was significantly lower in the CA3 and VPM/VPL of

GrnR493X/R493X mice, suggesting that microglial activation in these regions does not result

in substantial basal PTC readthrough or accumulation of truncated Pgrn-R493X (Figure

5-8, B, D).

As discussed in section 1.2.4.4, the innate immune defense system has been previously

implicated in the pathology of FTLD-GRN (92,120). Complement-driven synaptic

pruning is a critical microglial-mediated neurodevelopmental mechanism (189) that is

hyperactivated in the context of Grn-deficiency, resulting in selective depletion of

thalamic inhibitory synapses in Grn-/- mice (92). Complement protein C1qa was

significantly increased in the thalamic VPM/VPL of GrnR493X/R493X mice, similar to that

observed in Grn-/- mice (Figure 5-8, E-F). Astrogliosis is another hallmark of FTLD-GRN

pathogenesis demonstrated to develop in the same brain regions as microgliosis, as

observed by Gfap staining in Grn-/- mice (163,371). We present the first characterization

of GrnR493X/R493X astroglial pathology (Figure 5-10), observing that Gfap+ staining was

significantly increased in the CA3 and VPM/VPL regions in GrnR493X/R493X mice (Figure

5-10, B-C). The astrogliosis phenotype was more striking than microgliosis, spanning

179

both the CA3 and VPM/VPL compared to microgliosis primarily observed in the

thalamus.

180

181

Figure 5-8 Neuroinflammation in the ventral thalamus of aged GrnR493X/R493X mice.

(A) Representative hippocampal/thalamic tilescans co-stained for Iba1/Pgrn show

severe microgliosis in the brains of 18 month old GrnR493X/R493X mice (scale bar, 500

µm). (B) Representative Iba1/Pgrn co-stained immunofluorescence images of

hippocampal CA3 and thalamic VPM/VPL regions in 18 month old Grn+/+ and

GrnR493X/R493X brain sections (scale bar, 20 µm). (C) Quantification of microglial density

(i-ii) and branching morphology (iii-iv) in the CA3 and thalamic VPM/VPL regions. (D)

Microglial Pgrn fluorescence intensity in the CA3 (i) and VPM/VPL (ii). (E-F)

Representative images and quantification of C1qa staining in the VPM/VPL. n=10 mice

were used per sex/genotype (except male GrnR493X/R493X n=8 and male Grn+/+ Iba1-Pgrn

staining of VPM/VPL n=9); values are shown as mean ± SEM; ns not significant, * p <

0.05, *** p < 0.0001, Student’s t-test.

182

Figure 5-9 Microglial skeletal analysis method. Iba1 immunofluorescent images were

processed using Fiji as described in section 2.12.2 (A) and skeletonized using the Fiji

Analyze Skeleton (2D/3D) plugin (B).

183

Figure 5-10 Severe astrogliosis in the ventral thalamus and CA3 hippocampal

region of aged GrnR493X/R493X mice. (A) Representative hippocampal/thalamic tilescans

stained for Gfap show severe astrogliosis in 18 month old Grn+/+ and GrnR493X/R493X brain

(scale bar, 500 µm). (B) Representative Gfap immunofluorescence images of

hippocampal CA3 and thalamic VPM/VPL regions in 18 month old Grn+/+ and

GrnR493X/R493X brain sections (scale bar, 20 µm). (C) Total Gfap+ area in CA3 (i) and

VPM/VPL (ii) images were quantified and normalized to the total number of cells (DAPI,

Gfap

GrnR493X/R493XGrn+/+

A.

Grn+/+

GrnR493X/R493X

0

20

40

60

80

100

Gfa

p A

rea

(

m2)

/ C

ellGrn

+/+Grn

R493X/R493X

0

20

40

60

Gfa

p A

rea

(

m2)

/ C

ell

CA

3V

PM

/VP

L

GrnR493X/R493XGrn+/+

B.Gfap

C.

VP

M/V

PL

CA

3

i.

ii.***

***

184

blue). n=10 mice were used per sex/genotype (except male GrnR493X/R493X n=8); values

are shown as mean ± SEM; ns not significant, * p < 0.05, *** p < 0.0001, Student’s t-

test.

185

5.2.4 Partial preservation of inhibitory synaptic density in the thalamus of aged

GrnR493X/R493X mice

As discussed in sections 1.2.4.4, 1.2.6, and 5.2.3, there is growing support in the

literature for a critical role of the complement innate immune signaling cascade in

driving aberrant microglial synaptic pruning in FTLD-GRN and CLN11 (92,120). Given

the robust increase in complement C1qa protein deposition observed in the

GrnR493X/R493X ventral thalamus (Figure 5-8, E-F), we assessed whether this produced a

corresponding decrease in inhibitory synaptic density as previously observed in Grn-/-

mice (92). First, we characterized the whole thalamus Vgat synaptic density from

hippocampal/thalamic tilescans, and observed a non-significant trend towards lower

inhibitory synaptic density in GrnR493X/R493X mice (Figure 5-11, A-C). Since the Grn-/-

synaptic phenotype was localized to the ventral thalamus, we further analyzed C1qa-

Vgat images of Grn+/+ and GrnR493X/R493X VPM/VPL regions and found a similar non-

significant trend towards lower Vgat synaptic area and number of puncta in

GrnR493X/R493X mice (Figure 5-11, A’-B’, D-E). Interestingly, this weak trend towards a

reduction of thalamic Vgat synaptic density was observed despite 26.9% ± 6.0% of total

Vgat+ synaptic area being co-localized with C1qa in the GrnR493X/R493X VPM/VPL

thalamic regions (Figure 5-11, F).

186

187

Figure 5-11 Inhibitory synaptic density is preserved in the thalamus of aged

GrnR493X/R493X mice. (A-B) Representative hippocampal/thalamic tilescans co-stained for

C1qa and Vgat in 18 month old Grn+/+ and GrnR493X/R493X mice, the dashed outline

depicts area quantified in (C) (scale bar, 500 µm). (A’-B’) C1qa-Vgat tilescan inset

immunofluorescence images (from A-B) of thalamic VPM/VPL region in 18 month old

Grn+/+ and GrnR493X/R493X brain sections (scale bar, 20 µm). (C) Quantification of thalamic

Vgat+ synaptic area within hippocampal/thalamic tilescans normalized to thalamic area

(white dashed outline). Vgat+ area (D), the number of Vgat+ synaptic puncta (E), and

the proportion of Vgat+ synaptic area that is positive for C1qa (F) was quantified in high-

resolution C1qa-Vgat VPM/VPL images (A’-B’). n=10 mice were used per sex/genotype

(except male GrnR493X/R493X n=8); values are shown as mean ± SEM; ns not significant,

*** p < 0.0001, Student’s t-test.

188

5.2.5 Reduced thalamic excitatory neuron density in the brains of aged

GrnR493X/R493X mice

As outlined in section 1.2.4.4, a recent snRNA-seq study by Zhang et al. characterized

disease progression in Grn-/- mice from 2-19 months of age discovered a selective

decrease in the abundance of excitatory neuron markers from 12-19 months of age

(120). They validated these findings immunohistochemically by staining for excitatory

neuron markers Pkcδ and Foxp2 and identifying Grn-/- microglia surrounding Foxp2+

neurons contain Foxp2+ debris in their cytoplasm, suggestive of phagocytosis of dying

excitatory neurons (120). We attempted to replicate these findings in GrnR493X/R493X mice

by performing Foxp2 immunofluorescence staining and quantifying the number of

Foxp2+ thalamic neurons. 18 month old GrnR493X/R493X thalami exhibited a significant

reduction in the number of Foxp2+ excitatory neurons compared to Grn+/+ (21.0% ±

4.3%) mice, consistent with a possible neurodegenerative phenotype in this model

(Figure 5-12, A-B).

189

Figure 5-12 Reduced thalamic Foxp2+ excitatory neuron density in the brains of

aged GrnR493X/R493X mice. (A) Representative hippocampal/thalamic tilescans from 18

month old Grn+/+ and GrnR493X/R493X mice brain sections stained for Foxp2, the dashed

outline depicts area quantified in (B) (scale bar, 500 µm). (B) Quantification of total

number of thalamic Foxp2+ nuclei. n=9 male Grn+/+ mice, n=7 female Grn+/+ mice, n=7

male GrnR493X/R493X, and n=10 female GrnR493X/R493X mice were used; values are shown

as mean ± SEM; *** p < 0.0001, Student’s t-test.

190

5.2.6 Aged male GrnR493X/R493X mice exhibit an increased anxiety phenotype

Several behavioural phenotypes have been identified in Grn+/- and Grn-/- mice, including

deficits in social dominance, excessive grooming, and increased anxiety

(92,124,160,163,165). The initial GrnR493X/R493X characterization found nearly identical

onset and progression of OCD-like grooming behaviour in Grn-/- and GrnR493X/R493X mice

resulting in severe skin lesions, which likely contributed to their 30% lower median

survival rate (165). Though we did not specifically aim to quantify these phenotypes in

our mice, we did observe significantly more animal facility health updates reporting

lesions and grade 4 whisker barbering in GrnR493X/R493X mice (Grn+/+ 5/199 mice, mean

age: 66.2 ± 1.8 weeks vs. GrnR493X/R493X 16/233 mice, mean age: 49.4 ± 1.8 weeks, Chi

Square test, p = 0.0359) often requiring euthanasia likely attributable to excessive

grooming behaviour.

The open-field test has been used to establish the increased male-specific anxiety

phenotype in Grn-/- mice by quantifying the time mice were in the central or peripheral

regions of the open-field (160,163). GrnR493X/R493X mice spent significantly less time in

the central region of the open-field compared to Grn+/+ mice (Figure 5-13, A). We

conducted a sex-specific analysis to further evaluate whether this anxiety phenotype

was limited to male GrnR493X/R493X mice and found that male knock-in mice spent

significantly less time in the central zone than Grn+/+ males while female

Grn+/+/GrnR493X/R493X mice spent similar amounts of time in each region (Figure 5-13, B-

C). However, there is a non-significant trend towards increased anxiety in female

GrnR493X/R493X (Figure 5-13, C), suggesting that a larger sample size may have detected

191

a female-specific anxiety phenotype. The lack of significant difference observed

between the amount of time male and female GrnR493X/R493X mice spent in the central

zone (Figure 5-13, E), suggests that the stronger male-specific anxiety phenotype is

likely reflective of the trend towards lower baseline anxiety levels in male vs. female

Grn+/+ mice (Figure 5-13, D). To probe whether any of the previously presented

neuropathological phenotypes show any sex-dependent effects, we conducted further

sex-specific analyses of these data and failed to observe any statistically significant

differences between sexes (Figure 5-14, Figure 5-15, and Figure 5-16).

192

Figure

Figure 5-13 Aged male GrnR493X/R493X mice exhibit an increased anxiety phenotype.

Proportion of time male/female (A), male (B), and female (C) Grn+/+ and GrnR493X/R493X

mice spent inside the center region of an open-field arena over a 10 min trial. (D-E)

193

Comparison of intra-genotypic sex-differences in the % of time spent inside the center

region of an open-field arena. Male Grn+/+ n=12, female Grn+/+ n=15, male GrnR493X/R493X

n=10, and female GrnR493X/R493X n=10; values are shown as mean ± SEM; ns = not

significant, * p < 0.05, Student’s t-test.

194

Figure 5-14 GrnR493X/R493X brain lysosomal dysfunction does not exhibit a sex-

specific phenotype. DppII+ (A-B) and Lamp1+ (C-D) area/cell in both male and female

Grn+/+ and GrnR493X/R493X CA3 and VPM/VPL images were quantified and normalized to

the total number of cells (DAPI). For immunofluorescence staining n=10 mice were used

per sex/genotype (except male GrnR493X/R493X n=8); values are shown as mean ± SEM,

ns = not significant, one-way ANOVA with Tukey’s multiple comparison test.

195

Figure 5-15 GrnR493X/R493X microgliosis and astrogliosis does not exhibit a sex-

specific phenotype. Iba1+ microglial density was quantified in both male and female

Grn+/+ and GrnR493X/R493X CA3 (A) and VPM/VPL (B). Gfap+ astroglial density was

quantified in both male and female Grn+/+ and GrnR493X/R493X CA3 (C) and VPM/VPL (D).

n=10 mice were used per sex/genotype (except male GrnR493X/R493X n=8 male and Grn+/+

Iba1 staining of VPM/VPL n=9); values are shown as mean ± SEM, ns = not significant,

one-way ANOVA with Tukey’s multiple comparison test.

196

Figure 5-16 GrnR493X/R493X decreased thalamic excitatory neuron density and

inhibitory synaptic pruning does not exhibit a sex-specific phenotype. (A)

Quantification of total number of thalamic Foxp2+ nuclei in male and female Grn+/+ and

GrnR493X/R493X. Quantification of total thalamic (B) and VPM/VPL (C) Vgat+ inhibitory

synaptic area in male and female Grn+/+ and GrnR493X/R493X. Total thalamic Foxp2 (A)

n=9 male Grn+/+ mice, n=7 female Grn+/+ mice, n=7 male GrnR493X/R493X, and n=10

female GrnR493X/R493X mice, total thalamic Vgat (B) n=9 male Grn+/+ mice, n=9 female

Grn+/+ mice, n=7 male GrnR493X/R493X, and n=10 female GrnR493X/R493X mice, and

VPM/VPL Vgat (C) n=10 mice were used per sex/genotype (except male GrnR493X/R493X

197

n=8 male); values are shown as mean ± SEM, ns = not significant, one-way ANOVA

with Tukey’s multiple comparison test.

198

5.2.7 PTC readthrough in GrnR493X/R493X MEFs

Since the ultimate goal of obtaining the GrnR493X/R493X mouse model was to demonstrate

in vivo PTC readthrough of endogenous nonsense mutant Grn expression, we

established Grn+/+ and GrnR493X/R493X MEF lines for initial in vitro validation of treatment

efficacy. GrnR493X/R493X MEFs were treated with either escalating doses of G418 alone or

in combination with 5 µM CDX5-288 PTC readthrough enhancer. Western blot analysis

of G418-treated GrnR493X/R493X MEF lysates using the anti-mouse Pgrn Linker-1 antibody

(82) detected a dose-dependent increase in both full-length and R493X truncated Pgrn

(Figure 5-17, A). Furthermore, combination treatment of GrnR493X/R493X MEF treatment

with 25 µg/mL G418 and 5 µM CDX5-288 appeared to increase full-length Pgrn levels

compared to 25 µg/mL G418, which may be indicative of PTC readthrough

enhancement (Figure 5-17, A). 5-fold less protein was loaded in the Grn+/+ sample lane

to reduce overexposure (Figure 5-17, A). A subset of these MEF samples was

assessed for Pgrn levels by mouse Pgrn ELISA (Figure 5-17, B). Pgrn levels in

GrnR493X/R493X MEFs treated with 200 µg/mL G418 and combination treatment of 25

µg/mL G418 and 5 µM CDX5-288 were increased by 8.3-fold and 6.8-fold, respectively

(Figure 5-17, B). Given the low n-values presented in Figure 5-17, we did not conduct

statistical analysis and therefore viewed this data as a preliminary low confidence

observation. These results suggested that the GrnR493X loci is susceptible to PTC

readthrough and prompted further in vivo preclinical investigations.

199

Figure 5-17 Induction of PTC readthrough by G418 and enhancers in GrnR493X/R493X

MEFs. Expression of intracellular Pgrn was assessed in Grn+/+ and GrnR493X/R493X MEF

cultures treated with escalating concentrations of G418 alone or in combination with 5

µM CDX5-288 enhancer compound. (A) Expression of full-length Pgrn and Pgrn R493X

in treated Grn+/+ (5-fold less protein loaded) and GrnR493X/R493X MEF lysates were

analyzed by western blotting. (B) Quantification of Pgrn expression in a subset of these

200

Grn+/+ and GrnR493X/R493X MEF lysates by mouse Pgrn ELISA (n=1 per treatment group);

values are shown as mean of 2 technical replicates.

201

5.2.8 G418-induced toxicity in GrnR493X/R493X mice limits therapeutic potential

In chapter 3, we presented data demonstrating in vivo G418-induced PTC readthrough

of a human GRN R493X mutant expression construct delivered to the CNS by P0 ICV

AAV-GRN-R493X-V5 injection. At the time of treatment, AAV-GRN-R493X-V5 mice

were beginning to exhibit minor signs of declining health commonly associated with P0

AAV injection; therefore, G418-induced toxicity was more difficult to assess. Single ICV

injections with 50 µg G418 in these mice were tolerated, though, by the end of the 72 h

treatment period, mice were beginning to show signs of toxicity, including increased fur

piloerection, slow movement, and 10% body weight loss. Here, we aimed to translate

this knowledge into CNS targeted PTC readthrough treatments of GrnR493X/R493X mice

intending to delay or reverse disease progression by at least partially restoring Pgrn

expression.

In vivo G418-induced neurotoxicity was expected given the established CNS ototoxicity

associated with aminoglycoside treatment discussed in section 1.3.2. Although previous

studies have found 2X repeated 80 µg ICV injections (at T=0 h and T=48 h) of G418

over a 96 h time point were capable of inducing CNS PTC readthrough with minimal

toxicity (360,361). We first aimed to replicate the R493X GRN PTC readthrough findings

we observed following short term treatment of AAV-GRN-R493X-V5 mice with 100 µg

G418 in young GrnR493X/R493X mice. We conducted single ICV treatments of either saline

or 100 µg G418 in Grn+/+ and GrnR493X/R493X mice for 72 h. At the time of euthanasia, two

out of the three G418 treated GrnR493X/R493X mice exhibited signs of significant toxicity,

including limited mobility and > 10-15% body weight loss. Measuring Pgrn expression in

202

whole-brain RIPA-soluble lysates indicated that a single 100 µg G418-treatment may

have induced a 2X increase in CNS Pgrn levels (Figure 5-18), though it was unclear

whether this increase was driven by PTC-readthrough or aminoglycoside-mediated

toxicity. Given the low n-values presented in Figure 5-18, we did not conduct statistical

analysis and therefore viewed this data as a preliminary low confidence observation.

203

Figure 5-18 A single ICV injection of G418 may induce increased Pgrn expression

in the brains of GrnR493X/R493X mice. 19-21 week old Grn+/+ and GrnR493X/R493X mice

were used to test the ability for a single ICV injection of G418 to induce CNS GrnR493X

PTC readthrough. Control Grn+/+ and GrnR493X/R493X mice received a single ICV injection

with saline, and G418 treated GrnR493X/R493X mice received a single ICV injection with

100 µg G418 dose. Whole-brains were collected after 72 h. Untreated and G418-treated

whole-brain RIPA-soluble protein extracts were subjected to mouse Pgrn ELISA. n=1

mouse for saline Grn+/+, n=1 mouse for saline GrnR493X/R493X, and n=3 mice for G418

treated GrnR493X/R493X; values are shown as mean ± SEM.

0

2

4

6

8

Grn+/+ GrnR493X/R493X

n=1 n=3

1.00

1.99

5.73

n=1

G418 (100 g): - +-

Pg

rn (

ng

/mL

) in

wh

ole

-bra

in lysate

s

204

In 2018, Arrant et al. bilaterally infused the medial prefrontal cortices of 10-12 month old

Grn-/- mice with AAV-Grn and achieved striking rescue of several neuropathological

phenotypes in this model at 4-6 weeks post-transduction (218). Therefore, we

hypothesized that recurrent long-term (~1 month) induction of in vivo Grn R493X PTC

readthrough might be necessary to influence pathological progression in GrnR493X/R493X

mice. Studies comparing daily bolus injections with sustained osmotic pump-mediated

release of a related aminoglycoside, gentamicin, have identified that recurrent daily

doses are more favorable for promoting aminoglycoside-induced PTC readthrough

(243,375). To explore whether chronic CNS delivery of PTC readthrough-inducing

doses of G418 could be tolerated in knock-in mice, we obtained implantable and

programmable iPRECIO® pumps capable of automated chronic ICV dosing for up to 2

months. Before conducting PTC readthrough treatments, we first confirmed ICV

targeting by implanting an iPRECIO® pump loaded with 0.1% Fast Green FCF dye

programmed to deliver 10 µL of dye through the cranially implanted brain infusion

cannula during the first hour post-operation. Photographic observation of freshly

harvested dye-infused brain tissue demonstrated the expected ventricular distribution of

Fast Green FCF dye (Figure 5-19). Additionally, we also validated G418’s

thermostability up to 28 days at 37 °C, observing that it retains in vitro PTC readthrough

activity after incubation in an environment thermally analogous to subcutaneous

implantation (Figure 5-20).

205

Figure 5-19 Technical demonstration of bolus iPRECIO® pump-mediated ICV

injections targeting the lateral ventricles. A Grn+/+ mouse received implantation of a

0.1% Fast Green FCF dye loaded iPRECIO® pump inserted cranially at ICV

coordinates: -1.0 mm lateral/-0.3 mm posterior/-3.0 mm depth to bregma. Immediately

following 1 h infusion of 10 µL dye, the brain was collected and photographed,

demonstrating ventricular delivery.

206

Figure 5-20 G418’s PTC readthrough activity is preserved despite prolonged

incubation at physiologic temperature. R493X-/- KI hiPSC-derived astrocytes were

treated with either vehicle solution or 50 µg/mL G418 for 72 h. G418 working stock

solutions were incubated for either 0, 7, 14, 21, or 28 days at 37 °C. Expression of ~70

kDa PGRN in treated R493X-/- KI astrocyte samples were analyzed by western blotting,

using actin as the loading control. n = 2 independent cultures per treatment condition.

207

Since single 100 µg ICV doses of G418 were poorly tolerated, we chose to assess

whether recurrent doses of either 25 µg or 50 µg G418 are tolerated and capable of

inducing Grn R493X PTC readthrough. Please refer to Figure 5-21 for a schematic

description of iPRECIO® pump experiments #1-3. In experiment #1, we programmed

pumps to constantly infuse G418 solution at the lowest flow rate (0.1 µL/h) to maintain

brain infusion cannula patency and prevent clogging. This chronic low basal G418

dosing scheme resulted in rapid deterioration in the health of both 25 µg and 50 µg

treated mice, likely due to major G418 accumulation in the brain to neurotoxic levels

(Figure 5-22). In experiment #2, we programmed a similar G418 dosing schedule but

set the flow rate to 0 µL/h during the recovery phase and between bolus 25 µg and 50

µg infusions. These mice appeared to tolerate this regimen marginally better than those

in experiment #1; 4/5 mice were either found dead or required humane endpoint

euthanasia by day 6. However, one extremely obese (70 g) GrnR493X/R493X mouse (blue

star, Figure 5-21) tolerated five recurrent (every 48-72 h) 25 µg G418 bolus doses

exhibiting initial weight loss that stabilized by day 10. Unfortunately, this experimental

mouse had to be prematurely euthanized due to UBC’s COVID-19 shutdown

requirements. Regardless, experiment #2 hemi-brain lysate Pgrn quantification found

that 5X 25 µg G418 doses failed to increase Pgrn levels compared to untreated

GrnR493X/R493X (Figure 5-23). Importantly, we did observe a modest Pgrn increase in

mice treated with 2X 50 µg G418 doses. Again, given the low number of samples

assessed in Figure 5-23, we did not conduct statistical analysis and therefore viewed

this data as a preliminary low confidence observation.

208

Figure 5-21 Schematic describing iPRECIO® pump experiment infusion programs

and toxicity outcomes in mice. iPRECIO® pump experiments #1-3 were conducted on

Grn+/+ and GrnR493X/R493X mice. Red lines indicate infusion events (25/50 µg G418 or 2

µL saline) and mice were euthanized at humane endpoints except for Exp#3 saline

infused mice which were healthy at experimental endpoint and 5X 25 µg G418 treated

mouse (*) which had to be euthanized due to UBC COVID-19 shutdown.

209

Figure 5-22 Quantification of G418 concentration in the brains of Grn+/+ and

GrnR493X/R493X mice treated ICV with G418. Whole-brain lysates obtained from Grn+/+

mice treated with 100 µg G418 for 2 h (n=2), GrnR493X/R493X mice treated with 100 µg

G418 for 72 h (n=3), and an iPRECIO® Exp#1 50 µg G418 / 72 h treated mouse that

was treated for ~100 h (blue triangle in Figure 5-20, n=1) were assayed for G418

concentration levels using a commercial Gentamicin ELISA kit; values are shown as

mean ± SEM.

0

20

40

60

Grn+/+ GrnR493X/R493X

n=3 n=1n=2

Time-post 100 g G418 (h): 72 -2

iPRECIO Exp#1 (50 g / 72 h): - +-

G418 (g

/mL

) in

wh

ole

bra

in lysate

210

Figure 5-23 Multiple iPRECIO® pump-mediated ICV infusions of 50 µg G418 may

induce increased Pgrn expression in the brains of GrnR493X/R493X mice. 76 week old

Grn+/+ and GrnR493X/R493X hemi-brain lysates derived from untreated controls and

iPRECIO® pump Exp#2 treated mice were assayed for Pgrn expression to assess

whether 25 µg G418 dose can induce PTC readthrough. Untreated and G418-treated

hemi-brain RIPA-soluble protein extracts were subjected to mouse Pgrn ELISA. n=1

mouse for saline Grn+/+, n=1 mouse for saline GrnR493X/R493X, and n=1 mouse for 2X 25

µg G418, n=2 mice for 2X 50 µg G418, and n=1 for 5X 50 µg G418; values are shown

as mean ± SEM.

0

5

10

15

Grn+/+ GrnR493X/R493X

iPRECIO Exp#2 G418 (g): --

--

252

502

255No. of infusions:

n=1 n=1 n=1 n=2 n=1

Pg

rn (

ng

/mL

) in

hem

i-b

rain

lysate

s

211

Finally, for experiment #3, we implanted saline loaded pumps programmed to deliver 2

µL ICV infusions every 72 h into the brains of aged Grn+/+ mice. These mice received

and tolerated 4X 2 µL saline bolus injections and were sacrificed on day 12 to assess

brain Pgrn levels. We found that pump implanted mice did not exhibit increased Pgrn

expression (Figure 5-24), suggesting that implanted cannula-associated

neuroinflammation may not be responsible for upregulation of Pgrn expression in G418

pump implanted GrnR493X/R493X mice. This also suggests that technical surgical errors

linked to pump implantation were not responsible for the severe morbidity observed in

GrnR493X/R943X implanted with G418 ICV perfusing pumps. To further validate the

increased Pgrn expression observed in 2X 50 µg G418 filled pump implanted mice in

Exp#2 and completely rule out the possibility that pump-related non-specific

phenomenon was responsible for this increase, we conducted 2X ICV syringe-pump

driven 50 µg G418 injections in GrnR493X/R493X mice.

As expected, these repeated ICV injections (at t=0 h and t=48 h over 96 h) of 50 µg

G418 resulted in significant toxicity, which became particularly apparent ~24 h post the

2nd G418 injection. Though all mice survived until the experimental endpoint, their

weight decline and limited mobility at 96 h post-injection #1 would have soon required

humane endpoint euthanasia. To definitively demonstrate GrnR493X/R493X G418-induced

PTC readthrough, we pooled iPRECIO® 2X 50 µg G418 (n=2) and these syringe-pump

2X 50 µg G418 injected mice (n=3) into a single treatment group (2X 50 µg G418) and

evaluated Pgrn expression compared to untreated controls. 2X 50 µg G418 injections

significantly increased Pgrn expression when assayed by both western blot and Pgrn

212

ELISA (Figure 5-25). The increased levels of truncated Pgrn-R493X indicate potential in

vivo G418-induced PTC readthrough of an endogenously expressed nonsense mutant

Grn allele (Figure 5-25, A-B). Observing that repeated ICV treatments with 50 µg G418

increased brain Pgrn expression using two independent injection methods provides

additional evidence that G418 was responsible for increased Pgrn expression.

213

Figure 5-24 Grn+/+ tolerate iPRECIO® pump ICV infusion of saline. 77-88 week old

Grn+/+ hemi-brain lysates derived from untreated controls and iPRECIO® saline pump

Exp#3 treated mice were assayed for Pgrn expression levels. Untreated and saline

pump treated hemi-brain RIPA-soluble protein extracts were subjected to mouse Pgrn

ELISA. n=7 untreated Grn+/+ mice and n=2 iPRECIO® saline pump treated Grn+/+ mice;

values are shown as mean ± SEM; ns = non-significant, Student’s t-test.

214

Figure 5-25 Increased Pgrn expression induced by 2X repeated ICV G418

injections in the brains of GrnR493X/R493X mice. 52-86 week old Grn+/+ and

GrnR493X/R493X mice were used to test the ability for G418 to induce CNS GrnR493X PTC

readthrough. Control Grn+/+ and GrnR493X/R493X mice were untreated or received 2X ICV

saline injections t=0 h and t=48 h over 96 h. G418 treated GrnR493X/R493X mice received

either 2X syringe driven bolus ICV doses of 50 µg G418 at t=0 h and t=48 h over 96 h

(n=3) or ICV iPRECIO® pump injected 2X 50 µg G418 doses delivered at t=48 h and

t=120 h over 124 h (n=2). Untreated and G418-treated hemi-brain RIPA-soluble protein

215

extracts were subjected to Pgrn western blot (A-B) and mouse Pgrn ELISA (C).

Cropped Grn+/+ and GrnR493X/R493X western blot lanes presented were detected on the

same membrane, and the loading control signal was obtained by reprobing the same

membrane with actin for normalization. The first two G418-treated lanes are from ICV

iPRECIO® pump injected mice, and the last three are from syringe-pump injected mice.

Both Pgrn western blot and ELISA data are presented relative to untreated

GrnR493X/R493X. Values are shown as mean ± SEM; * p < 0.05, *** p < 0.0001, one-way

ANOVA with Tukey’s multiple comparison test.

216

5.3 Discussion

In this chapter, we provide the first detailed analysis of lysosomal dysfunction and

selective loss of thalamic excitatory neurons in the brains of aged GrnR493X/R493X mice.

First, we established that Pgrn expression is suppressed in the brains of aged

GrnR493X/R493X mice, and pathologically active brain regions do not show upregulated

Pgrn expression via basal PTC readthrough. This limited expression is driven by NMD

(165); therefore, the majority of Pgrn detected in the GrnR493X/R493X brain likely consist of

the truncated Pgrn-R493X form. Since disrupted lysosomal homeostasis is a

pathological hallmark of Grn-/- mice, we chose to evaluate several previously

established brain lysosomal phenotypes in our aged GrnR493X/R493X cohort. The CA3

hippocampal and thalamic VPM/VPL brain regions of aged GrnR493X/R493X mice

displayed striking expansions of their lysosomal compartments. We further evaluated

aged knock-in mice for global changes in brain lysosomal function and impairments in

autophagy; identifying overexpression of both the pro- and mature-forms of lysosomal

protease Ctsd and impaired clearance of autophagolysosomes as indicated by an

increased LC3-II:LC3-I ratio, both of which have been previously observed in aged Grn-/-

mice (106,119). Notably, evidence of lysosomal dysfunction has been demonstrated in

Grn-/- mice as young as 2 months of age (179). Future efforts may seek to understand

whether the presence of a semi-functional, truncated Pgrn-R493X might delay the onset

of this early lysosomal phenotype.

As discussed in sections 1.1.3 and 1.1.7, GRN deficiency is associated with nuclear to

cytoplasmic translocation of TDP-43, ultimately resulting in the formation of insoluble

217

neuronal inclusions. The redistribution of TDP-43 from the nucleus to the cytoplasm has

also been observed in Grn-/- mice, a process which may involve excessive

neuroinflammation and dysfunctional autophagolysosomal and ubiquitin-proteasome

systems (120,165,371,373). Similar to these findings, we found TDP-43 pathology

limited to ventral thalamic neurons in GrnR493X/R493X mice, which exhibited intense

nuclear to cytoplasmic TDP-43 translocation. We further observed that GrnR493X/R493X

soluble hemi-brain lysates contained decreased TDP-43, potentially indicating that a

proportion of the soluble TDP-43 pool had transitioned into an insoluble form. This

phenomenon was observed in a mutant TDP-43 mouse model where soluble TDP-43

increased while insoluble TDP-43 decreased upon overexpression of PGRN (376).

However, we failed to observe a corresponding increase in the insoluble levels of TDP-

43 and were unable to detect p-TDP-43 in either soluble or insoluble lysates, both of

which have been previously found in aged Grn-/- mice (93,371). The former most likely

relates to our finding that TDP-43 pathology is largely limited to ventral thalamic

neurons, and the phenotype may thus be lost through dilution when assessing hemi-

brain lysates. Other discrepancies are less clear and may relate to inherent differences

in mouse models as well as the particular age chosen for analysis.

GrnR493X/R493X mice exhibit age-dependent microgliosis that begins around 6 months of

age, reaching a peak at 12 months, and is maintained until 18 months of age. We

observed significant neuroinflammation in the ventral thalamus of GrnR493X/R493X mice,

including astro- and microglial expansion and morphological transition into a

proinflammatory state. This observation is important because we show that the

218

chronically inflamed ventral thalamic brain region also develops TDP-43 proteinopathy,

providing additional support to the growing evidence demonstrating that factors

secreted by Grn-deficient microglia directly induce cytoplasmic accumulation of TDP-43

(165). These inflammatory mediators include the innate immune system complement

system (C1qa, C3, etc.) which have been implicated in Grn-deficient microglial

mediated neurotoxicity. Although most of the literature has focused on examining

microglial-driven neuropathological mechanisms, a recent study by Guttikonda et al.

conducted using hiPSC-derived neuron-astrocyte-microglia tricultures suggested that

reciprocal C3 signaling between microglia and astrocytes is critical to driving excessive

microglial C1qa complement protein expression and secretion (377). Therefore, future

studies assessing the role of Grn-deficient astrocytes in driving disease pathophysiology

may be critical to identifying novel therapeutic strategies.

As discussed in sections 1.2.4 and 1.2.6, microglial-mediated innate immune system

complement pathway activation has been directly and indirectly implicated in driving

thalamic neurodegeneration in Grn-/- mice through selective targeting of both inhibitory

synapses and excitatory neurons for elimination (92,120). Despite observing robust

C1qa tagging of Vgat+ synapses as seen in Grn-/- mice, neither GrnR493X/R493X whole

thalami or the VPM/VPL region showed a significant decrease in inhibitory synaptic

density. However, there was a trend towards lower levels of Vgat+ synapses in these

regions. It is possible that low basal Pgrn-R493X expression in GrnR493X/R493X microglial

limited their voracity for inhibitory synapses, but this is not known. A recent snRNA-seq

study found that selective loss of excitatory neurons in the ventral thalamus of Grn-/-

219

mice could be rescued by simultaneous deletion of both C1qa and C3 complement

genes (120). The authors proposed that elevated C1qa and C3 in this brain region

result in increased MAC formation on neuronal surfaces, which permeabilizes their

plasma membranes, triggering apoptosis (120). Similarly, our aged GrnR493X/R493X

displayed loss of thalamic Foxp2+ excitatory neurons, perhaps suggesting that a loss of

excitatory neurons precedes the synaptic pruning phenotype in this model.

Nevertheless, it is still unclear how these complement-mediated neurodegenerative

mechanisms selectively target inhibitory synapses and excitatory neurons.

Grn-deficient mice develop behavioural abnormalities impacting social interactions,

grooming frequency, and anxiety levels (92,124,160,163,165). GrnR493X/R493X mice were

no different, exhibiting an increased anxiety phenotype similar to that previously

reported in Grn-/- mice, observing a stronger anxiety phenotype in males (160,163). In

section 1.2.4.4, we discussed neuropathological correlates of behavioural disturbances

in Grn-/-, connecting decreased inhibition of the thalamocortical circuit to their OCD-like

grooming phenotype (92). Since we observed significant thalamic pathology in aged

GrnR493X/R493X mice, we analyzed the major neuropathological phenotypes presented

here for sexual dimorphism to assess whether increased pathology might explain the

male-predominant anxiety phenotype. While we found that levels of inhibitory synaptic

density in the thalamic VPM/VPL regions exhibited a strong trend towards a sex-

dependent phenotype in both Grn+/+ and GrnR493X/R493X mice (Figure 5-16, C), with males

displaying elevated inhibitory synaptic density compared to females, we did not observe

sexually dimorphic FTLD-related pathology in GrnR493X/R493X mice. Similar efforts have

220

been made to identify the neurological basis for the increased susceptibility of male

Grn-deficient mice to the development of an increased anxiety phenotype (378,379).

These studies found that Grn expression is upregulated in the ventromedial

hypothalamic nucleus in response to androgen and estrogen sex-hormones and that

Pgrn is an essential mediator of male sexual differentiation in the developing brain.

Since Grn+/+ female mice generally exhibit elevated anxiety compared to Grn+/+ males, it

has been proposed that a lack of Grn expression during sexual differentiation results in

at least partial fulfillment of the default female neurodevelopmental program (160).

Evidence supporting this hypothesis includes the observation that Grn-/- mice lack the

sexually dimorphic trait of differential locus ceruleus (LC) volume, which is normally

larger in Grn+/+ females (380). Because the LC is an important regulator of stress-

induced anxiety responses, it is possible that a relatively enlarged LC in male Grn-/-

compared to Grn+/+ mice could predispose them to increased anxiety (381). There

studies suggest profound developmental changes in the brain as a result of Grn

deficiency. Improved preclinical methodology, including the use of hiPSCs, may allow

future studies to probe these mechanisms which could reveal process far upstream of

known FTLD pathology that may prove central to FTLD-GRN pathophysiology.

Taken together, a striking finding in both aged GrnR493X/R493X mice and other Grn-/-

models is the pronounced involvement of select thalamic regions. For the clinician this

might be curious as neurodegeneration in the frontal, temporal, and parietal lobes is

widely recognized in driving clinical symptoms across FTLD syndromes. However, a

recent study in preclinical GRN carriers found prominent hyperconnectivity between

221

cortical hub regions and the thalamus in several intrinsic connectivity networks

assessed by functional magnetic resonance imaging (213). The implications of these

findings are not fully clear, but abnormal thalamic physiology, which is robustly

demonstrated in FTLD-GRN mouse models either through histology or

electrophysiology, may have important implications for the earliest changes in human

FTLD-GRN. As such, interventions to reverse these pathologic changes in mice may

have important translational value.

Aminoglycoside-induced toxicity (reviewed in section 1.3.2) has been a major deterrent

to the clinical translation of PTC readthrough therapy. Through the discovery of our

CDX-series of PTC readthrough enhancer compounds, we aimed to lower the required

concentration of G418 to a more tolerable level while still achieving readthrough. As

shown in GrnR493X/R493X derived MEFs, we were able to accomplish this goal in vitro. In

section 5.2.8, we chronicled our efforts to identify a dose of G418 capable of inducing

PTC readthrough that would be tolerated for at least a month-long recurrent treatment

regimen. These data represent the first demonstration of CNS delivery of a PTC

readthrough drug using an implantable programmable pump capable of infusing bolus

doses. We found that short term ICV doses of 50-100 µg G418 could induce GrnR493X

PTC readthrough, though were accompanied by significant and often fatal neurotoxicity.

Though not presented here, we attempted single syringe-pump ICV injections of 200 µg

and 400 µg G418 doses, which resulted in rapid-onset muscular paralysis requiring

euthanasia and sudden death, respectively. Given that 2/3 GrnR493X/R493X mice treated

with 25 µg G418 doses for iPRECIO® Exp#2 were either found dead or required

222

humane endpoint euthanasia and that even five recurrent 25 µg G418 doses over 15

days failed to increase Pgrn expression; we conclude that there is likely no tolerated

ICV dose of G418 that is capable sufficient CNS exposure to induce PTC readthrough

in the GrnR493X/R493X mouse model. Unable to establish a baseline tolerated dose of

G418 capable of inducing PTC readthrough, we chose to abandon efforts to attempt

combination therapy with CDX5-288, as we would not have been able to assess the

degree of PTC readthrough enhancement.

In conclusion, our aged cohort of GrnR493X/R493X mice displayed several pathologic

phenotypes, including lysosomal dysfunction and select thalamic synaptic degeneration

not previously described in this model, but in line with observations in other Grn-/- model

mice. Our characterization of aged GrnR493X/R493X mice provides the field with further

insight into neuropathological phenotypes that may be used to better define the

mechanisms underlying FTLD-GRN, and evaluate the preclinical efficacy of novel

therapeutics to target relevant nonsense mutations that cause FTLD-GRN.

223

Chapter 6: Conclusion and future directions

6.1 Contributions to the fields of FTLD-GRN and CLN11

FTLD-GRN is a devastating neurodegenerative disease, yet there are currently no

disease-modifying therapies available in the clinic. The majority of drug development

efforts for the treatment of FTLD-GRN have focused on restoring normal PGRN levels

by increasing expression from the intact allele or exogenous delivery of PGRN to the

CNS. The viability of preclinical investigation exploring novel nonsense mutation

suppression therapies is dependent on the prevalence of nonsense mutations in a given

genetic condition. Approximately 25% of causal FTLD-GRN mutations introduce in-

frame PTCs, providing a substantial patient population that may benefit from PTC

readthrough therapy. Despite the known toxicity of aminoglycosides, PTC readthrough

has not been pursued as a therapy for FTLD-GRN. The work here describes proof-of-

concept demonstration of PTC readthrough in several in vitro and in vivo models

bearing clinical GRN nonsense mutations, and represents a major contribution to the

field of FTLD-GRN incentivizing further development of this therapeutic approach.

6.1.1 In vitro and in vivo PTC readthrough of exogenously expressed nonsense

mutant human GRN expression constructs

Early in 2020, we showed that aminoglycoside-mediated GRN nonsense mutation

readthrough in cells transfected with C-terminally tagged nonsense mutant expression

GRN constructs (122). In chapter 3, we generated HEK293 cell lines stably expressing

human GRN bearing several clinically relevant FTLD-GRN nonsense mutations for

224

screening our lead PTC readthrough enhancer compounds in combination with the

aminoglycoside G418. These screening efforts identified that the UGA GRN R418X and

R493X mutations were most susceptible to PTC readthrough and that combination

treatment with CDX5-288 had the most synergistic effect on readthrough efficiency.

Based on these findings, we generated and transduced neonatal C57BL/6J mice with

AAV vectors expressing tagged R493X mutant GRN. We found that a single ICV

injection of G418-induced CNS wide PTC readthrough of human tagged R493X GRN

expression constructs in AAV-GRN-R493X-V5 mice. CDX5-288 co-treatment with G418

in these mice did not induce PTC readthrough enhancement, suggesting that the CDX5-

288 dosage we selected based on in vitro efficacy may have been too low.

Complications involving poor viability of AAV-GRN-R493X-V5 mice limited our ability to

conduct additional G418 and CDX5-288 combination treatment dose optimizations.

Furthermore, the artificial nature of GRN-R493X-V5 overexpression precluded

quantification of disease-relevant PTC readthrough efficiency in this model; therefore,

we view these results as proof-of-concept demonstration of in vivo PTC readthrough.

6.1.2 In vitro and in vivo PTC readthrough in human and mouse models

endogenously expressing nonsense mutant GRN

To our knowledge, this is the first and only report demonstrating nonsense mutation

readthrough to enhance PGRN expression in FTLD-GRN patient-derived cortical

neurons and astrocytes. In chapter 4, we reviewed our efforts reprogramming healthy

control and FTLD-GRN UBC15 patient somatic cells into hiPSCs and CRISPR/Cas9

225

gene editing the first isogenic hiPSC model of CLN11. These hiPSC models provided

excellent sources of disease-relevant cell types bearing common pathogenic GRN

nonsense mutations. We also examined the potency of our PTC readthrough

combination therapy in S116X+/-, R418X+/-, and R493X-/- KI hiPSC-derived cortical

neurons and astrocytes. Full-length translation of PGRN was detected in R493X-/- KI

hiPSC-derived astrocytes treated with G418 and CDX5-288 by co-treatment with a

CTSL inhibitor to prevent the rapid processing of readthrough-derived full-length PGRN

into GRNs observed in this cell type. This is an important observation for other

researchers in the field, aiming to pursue readthrough based approaches for the

treatment of FTLD-GRN. We further demonstrated that increased PGRN expression

induced by G418 and CDX5-288 in R493X-/- KI cortical neurons rescued an elevated

mature CTSD mutant lysosomal phenotype, indicating that the combination of full-length

and truncated PTC readthrough-derived PGRN had a functional benefit. These cell lines

represent valuable research tools for the field of FTLD.

In chapter 5, we characterized the GrnR493X/R493X knock-in mouse model bearing a

murine analog to the pathogenic R493X mutation. Our aged GrnR493X/R493X knock-in mice

cohort displayed several neuropathological features previously observed in Grn-/- mice.

Notably, we provide the first evidence that these mice develop striking lysosomal

dysfunction and neurodegeneration in the form of selective loss of thalamic excitatory

neurons. Moreover, we observed a male-specific increased anxiety phenotype in aged

GrnR493X/R493X mice previously reported in Grn-/- mice. We provided preliminary results

from a pilot study demonstrating in vivo PTC readthrough of endogenously expressed

226

GrnR493X through repeated ICV administrations of G418. However, the doses of G418

required to achieve readthrough caused severe neurotoxicity and were not suitable for

long term therapeutic intervention. Future studies are needed to further explore whether

in vivo PTC readthrough induced PGRN rescue is sufficient to surpass the therapeutic

threshold required to disrupt or slow down FTLD-GRN pathogenesis.

6.2 Final conclusions

In summary, we show that G418 increases PGRN through PTC readthrough in cellular

and mouse models of FTLD due to pathogenic progranulin nonsense mutations. The

enhancer compound CDX5-288 potentiates the readthrough effects of G418 in vitro and

increases PGRN over G418 treatment alone. We further show that increased

progranulin through PTC readthrough ameliorates lysosomal dysfunction in R493X-/- KI

cortical neurons, showing a functional impact of raised PGRN. A single ICV

administration of G418 in mice expressing tagged human GRN R493X nonsense

mutation achieves proof-of-concept in vivo PTC readthrough. Our characterization of

aged GrnR493X/R493X mice provides the field with insight into neuropathological

phenotypes that may be used to evaluate the preclinical efficacy of novel therapeutics in

this model. While G418 is not suitable for long-term human consumption, our work

highlights the importance of ongoing and future efforts to find novel readthrough

compounds more suitable for human use. Taken together, our findings suggest that

PTC readthrough may be a potential therapeutic strategy for a subset of patients with

FTLD bearing a GRN nonsense mutation, but highlight the need for better drugs with

improved tolerability.

227

6.3 Future directions

Our observations of severe dose-limiting G418-mediated toxicity following ICV delivery

in GrnR493X/R493X mice suggest that aminoglycoside readthrough compounds are unlikely

to have a therapeutic window for FTLD-GRN or CLN11 patients carrying nonsense

mutant GRN alleles. As discussed in section 1.3.3, systematic examination of the

chemical moieties responsible for aminoglycoside-induced toxicity and PTC

readthrough by Dr. Timor Baasov’s group resulted in the synthesis of the novel

eukaryotic ribosomal selective glycoside, ELX-02. Developing an academia-industry

partnership with Eloxx Pharmaceuticals may enable us to conduct similar in vivo brain-

targeting PTC readthrough studies in our GrnR493X/R493X mouse cohort testing ELX-02

and other synthetic aminoglycosides under their development. Additionally, we are

currently in the process of conducting in vitro validation of novel PTC readthrough drugs

through our collaborations with the Roberge lab, which may support further exploration

of in vivo nonsense mutation readthrough therapy in GrnR493X/R493X knock-in mice.

As discussed in section 1.2.7, several ongoing clinical investigations are evaluating the

efficacy of therapies aimed at restoring PGRN expression levels in FTLD-GRN patients.

Alector is currently recruiting participants for a phase III clinical trial (INFRONT-3)

testing the anti-SORT1 antibody AL001, which aims to disrupt SORT1-mediated

extracellular uptake of PGRN to increase its extracellular bioavailability. Since brain

SORT1 expression is primary localized to neurons, inhibition of SORT1-mediated

PGRN endocytosis likely results in greater microglial endocytosis of extracellular PGRN

through SORT1-independent transport mechanisms (described in section 1.2.3).

228

Microglia are increasingly recognized as primary drivers of FTLD-GRN pathophysiology;

therefore, it is possible that recently reported improvements in CSF FTLD-GRN

biomarkers following 4-week treatment with AL001 may be due to restored microglial

PGRN levels. Exogenous delivery of GRN expression through gene therapy or direct

delivery of intact rec. PGRN represents another promising approach to delaying or

reversing pathological progression in FTLD-GRN and CLN11 patients. Prevail

Therapeutics is actively recruiting symptomatic FTLD-GRN patients for a phase I/II

clinical trial (PROCLAIM) to validate the safety, tolerability, and immunogenicity of a

single dose (low, medium, or high) of their AAV9 GRN expression vector (PR006)

delivered intraventricularly via the cisterna magna. Additionally, Passage Bio Inc. will

soon begin recruiting for its own phase I/II clinical trial (upliFT-D) to test the safety,

tolerability, and pharmacodynamic effects of a single dose (3.3E10 or 1.1E11 genome

copies / gram of estimated brain weight) of their AAV1 GRN expression vector

(PBFT02) which will also be delivered by intra-cisterna magna injection. These trials

build upon several preclinical studies demonstrating viral CNS delivery of GRN

expression (218,221) and may provide insight into whether restoring GRN expression in

the brain can disrupt the pathogenic processes associated with FTLD-GRN. Importantly,

PGRN overexpression is known to be associated with oncogenicity in multiple organ

systems (382,383), therefore it will be critical to ensure GRN expression is limited to

normal healthy control levels.

Drugs targeting the pathophysiological mechanisms responsible for FTLD-GRN disease

pathogenesis may represent an alternative therapeutic modality for the treatment of

229

GRN deficiency. Our laboratory’s expertise and unique repository of FTLD-GRN patient-

and isogenic CRISPR/Cas9-derived hiPSC models are well suited for further elucidation

of disease-associated molecular pathways. Recent advances in cerebral organoid

technology have enabled the production of 3D iPSC-derived spheroids that

neurodevelopmentally replicate elements of the human embryonic brain (384). Although

these neural constructs lack resident microglia, hiPSC-derived microglia can be co-

cultured with maturing cerebral organoids, resulting in microglia chemotactic migration

and integration into the mini-brains. Microglia tile the cerebral organoid tissue adopting

a characteristic ramified morphology that upon neural injury become activated,

undergoing classic amoeboid morphological transition (198). hiPSC-derived microglia-

cerebral organoid co-cultures represent a valuable research tool for the study of FTLD-

GRN and CLN11 pathology. A hypothetical experiment could involve co-culturing WT or

R493X-/- KI hiPSC-derived microglia with WT or R493X-/- KI hiPSC-derived cerebral

organoids to assess whether mutant microglial induce any characteristic FTLD-GRN

phenotypes, such as TDP-43 cytoplasmic translocation or excessive secretion of

complement (C1QA, C3, etc.) proteins. Assuming we were to observe R493X-/- KI

microglia-specific phenotypes, this may represent an ideal model for high-throughput

screening to identify drugs capable of reversing the hyperinflammatory phenotype in

GRN-deficient microglia. It would then be important to test the top compounds in

Grn+/R493X and GrnR493X/R493X mice to assess whether treatment can decrease microglia-

induced neuropathology in the absence of increased PGRN expression.

The current slate of clinical trials actively recruiting FTLD-GRN patients represents the

230

culmination of 15 years of extensive research uncovering the pathophysiology of this

rare genetic condition and provide a reason for individuals carrying pathogenic GRN

mutations to be optimistic that the first disease-modifying therapies may receive clinical

approval within the next decade.

231

References

(1) Ratnavalli E, Brayne C, Dawson K, Hodges JR. The prevalence of frontotemporal dementia. Neurology 2002;58(11):1615-1621.

(2) Englund B, Brun A, Gustafson L, Passant U, Mann D, Neary D, et al. Clinical and neuropathological criteria for frontotemporal dementia. J Neurol Neurosurg Psychiatr 1994;57(4):416-418.

(3) Tsai RM, Boxer AL. Therapy and clinical trials in frontotemporal dementia: past, present, and future. J Neurochem 2016;138:211-221.

(4) Pasquier F, Petit H. Frontotemporal dementia: its rediscovery. Eur Neurol 1997;38(1):1-6.

(5) Pick A. Uber die Beziehungen der senilen Hirnatrophie zur Aphasie. Prag Med Wochenschr 1892;17:165-167.

(6) Pick A, Girling DM, Marková IS. Senile Hirnatrophie als Grundlage von Herderscheinungen: (Senile cerebral atrophy as the origin of focal symptoms). History of Psychiatry 1995;6(24):533-537.

(7) Pick A. Über einen weiteren Symptomenkomplex im Rahmen der Dementia senilis, bedingt durch umschriebene stärkere Hirnatrophie (gemischte Apraxie). Eur Neurol 1906;19(2):97-108.

(8) Wechsler AF, Verity MA, Rosenschein S, Fried I, Scheibel AB. Pick's disease: a clinical, computed tomographic, and histologic study with Golgi impregnation observations. Arch Neurol 1982;39(5):287-290.

(9) Gustafson L, Nilsson L. Differential diagnosis of presenile dementia on clinical grounds. Acta Psychiatr Scand 1982;65(3):194-209.

(10) Gustafson L, Brun A, Risberg J. Frontal lobe dementia of non-Alzheimer type. Adv Neurol 1990;51:65-71.

(11) Mendez MF, Selwood A, Mastri AR, Frey W2. Pick's disease versus Alzheimer's disease: a comparison of clinical characteristics. Neurology 1993;43(2):289-292.

(12) Neary D, Snowden JS, Gustafson L, Passant U, Stuss D, Black S, et al. Frontotemporal lobar degeneration: a consensus on clinical diagnostic criteria. Neurology 1998;51(6):1546-1554.

232

(13) Diehl-Schmid J, Grimmer T, Drzezga A, Bornschein S, Riemenschneider M, Förstl H, et al. Decline of cerebral glucose metabolism in frontotemporal dementia: a longitudinal 18F-FDG-PET-study. Neurobiol Aging 2007;28(1):42-50.

(14) Neary D, Snowden JS, Mann DM, Northen B, Goulding PJ, Macdermott N. Frontal lobe dementia and motor neuron disease. Journal of Neurology, Neurosurgery & Psychiatry 1990;53(1):23-32.

(15) Burrell JR, Halliday GM, Kril JJ, Ittner LM, Götz J, Kiernan MC, et al. The frontotemporal dementia-motor neuron disease continuum. The Lancet 2016;388(10047):919-931.

(16) Riedl L, Mackenzie IR, Förstl H, Kurz A, Diehl-Schmid J. Frontotemporal lobar degeneration: current perspectives. Neuropsychiatric disease and treatment 2014;10:297-310.

(17) Hodges JR, Davies R, Xuereb J, Kril J, Halliday G. Survival in frontotemporal dementia. Neurology 2003;61(3):349-354.

(18) Shi J, Shaw CL, Du Plessis D, Richardson AM, Bailey KL, Julien C, et al. Histopathological changes underlying frontotemporal lobar degeneration with clinicopathological correlation. Acta Neuropathol 2005;110(5):501-512.

(19) Mann DM, Snowden JS. Frontotemporal lobar degeneration: Pathogenesis, pathology and pathways to phenotype. Brain Pathology 2017;27(6):723-736.

(20) Knopman DS, Mastri AR, Frey WH2, Sung JH, Rustan T. Dementia lacking distinctive histologic features: a common non‐Alzheimer degenerative dementia. Neurology 1990;40(2):251-256.

(21) Jackson M, Lennox G, Lowe J. Motor neurone disease-inclusion dementia. Neurodegeneration 1996;5(4):339-350.

(22) Josephs KA, Holton JL, Rossor MN, Godbolt AK, Ozawa T, Strand K, et al. Frontotemporal lobar degeneration and ubiquitin immunohistochemistry. Neuropathol Appl Neurobiol 2004;30(4):369-373.

(23) Arai T, Hasegawa M, Akiyama H, Ikeda K, Nonaka T, Mori H, et al. TDP-43 is a component of ubiquitin-positive tau-negative inclusions in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Biochem Biophys Res Commun 2006;351(3):602-611.

(24) Davidson Y, Kelley T, Mackenzie IR, Pickering-Brown S, Du Plessis D, Neary D, et al. Ubiquitinated pathological lesions in frontotemporal lobar degeneration contain the TAR DNA-binding protein, TDP-43. Acta Neuropathol 2007;113(5):521-533.

233

(25) Mackenzie IR, Neumann M, Baborie A, Sampathu DM, Du Plessis D, Jaros E, et al. A harmonized classification system for FTLD-TDP pathology. Acta Neuropathol 2011;122(1):111-113.

(26) Mizielinska S, Lashley T, Norona FE, Clayton EL, Ridler CE, Fratta P, et al. C9orf72 frontotemporal lobar degeneration is characterised by frequent neuronal sense and antisense RNA foci. Acta Neuropathol 2013;126(6):845-857.

(27) Davidson YS, Barker H, Robinson AC, Thompson JC, Harris J, Troakes C, et al. Brain distribution of dipeptide repeat proteins in frontotemporal lobar degeneration and motor neurone disease associated with expansions in C9ORF72. Acta neuropathologica communications 2014;2(70).

(28) Davidson YS, Robinson AC, Hu Q, Mishra M, Baborie A, Jaros E, et al. Nuclear carrier and RNA‐binding proteins in frontotemporal lobar degeneration associated with fused in sarcoma (FUS) pathological changes. Neuropathol Appl Neurobiol 2013;39(2):157-165.

(29) Rohrer JD, Guerreiro R, Vandrovcova J, Uphill J, Reiman D, Beck J, et al. The heritability and genetics of frontotemporal lobar degeneration. Neurology 2009;73(18):1451-1456.

(30) Pickering-Brown SM, Rollinson S, Du Plessis D, Morrison KE, Varma A, Richardson AM, et al. Frequency and clinical characteristics of progranulin mutation carriers in the Manchester frontotemporal lobar degeneration cohort: comparison with patients with MAPT and no known mutations. Brain 2008;131(3):721-731.

(31) Seelaar H, Kamphorst W, Rosso SM, Azmani A, Masdjedi R, de Koning I, et al. Distinct genetic forms of frontotemporal dementia. Neurology 2008;71(16):1220-1226.

(32) Grover A, Houlden H, Baker M, Adamson J, Lewis J, Prihar G, et al. 5′ splice site mutations in tau associated with the inherited dementia FTDP-17 affect a stem-loop structure that regulates alternative splicing of exon 10. J Biol Chem 1999;274(21):15134-15143.

(33) Hong M, Zhukareva V, Vogelsberg-Ragaglia V, Wszolek Z, Reed L, Miller BI, et al. Mutation-specific functional impairments in distinct tau isoforms of hereditary FTDP-17. Science 1998;282(5395):1914-1917.

(34) Cruts M, Kumar-Singh S, Van Broeckhoven C. Progranulin mutations in ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Current Alzheimer Research 2006;3(5):485-491.

234

(35) Gass J, Cannon A, Mackenzie IR, Boeve B, Baker M, Adamson J, et al. Mutations in progranulin are a major cause of ubiquitin-positive frontotemporal lobar degeneration. Hum Mol Genet 2006;15(20):2988-3001.

(36) Baker M, Mackenzie IR, Pickering-Brown SM, Gass J, Rademakers R, Lindholm C, et al. Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature 2006;442(7105):916.

(37) Rademakers R, Neumann M, Mackenzie IR. Advances in understanding the molecular basis of frontotemporal dementia. Nature Reviews Neurology 2012;8(8):423-434.

(38) Pottier C, Ravenscroft TA, Sanchez‐Contreras M, Rademakers R. Genetics of FTLD: overview and what else we can expect from genetic studies. J Neurochem 2016;138:32-53.

(39) DeJesus-Hernandez M, Mackenzie IR, Boeve BF, Boxer AL, Baker M, Rutherford NJ, et al. Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron 2011;72(2):245-256.

(40) Renton AE, Majounie E, Waite A, Simón-Sánchez J, Rollinson S, Gibbs JR, et al. A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD. Neuron 2011;72(2):257-268.

(41) Majounie E, Renton AE, Mok K, Dopper EG, Waite A, Rollinson S, et al. Frequency of the C9orf72 hexanucleotide repeat expansion in patients with amyotrophic lateral sclerosis and frontotemporal dementia: a cross-sectional study. The Lancet Neurology 2012;11(4):323-330.

(42) Mizielinska S, Isaacs AM. C9orf72 amyotrophic lateral sclerosis and frontotemporal dementia: gain or loss of function? Curr Opin Neurol 2014;27(5):515-523.

(43) Gendron TF, Belzil VV, Zhang Y, Petrucelli L. Mechanisms of toxicity in C9FTLD/ALS. Acta Neuropathol 2014;127(3):359-376.

(44) Skibinski G, Parkinson NJ, Brown JM, Chakrabarti L, Lloyd SL, Hummerich H, et al. Mutations in the endosomal ESCRTIII-complex subunit CHMP2B in frontotemporal dementia. Nat Genet 2005;37(8):806-808.

(45) Watts GD, Wymer J, Kovach MJ, Mehta SG, Mumm S, Darvish D, et al. Inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia is caused by mutant valosin-containing protein. Nat Genet 2004;36(4):377.

235

(46) Rubino E, Rainero I, Chiò A, Rogaeva E, Galimberti D, Fenoglio P, et al. SQSTM1 mutations in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Neurology 2012;79(15):1556-1562.

(47) Gijselinck I, Van Mossevelde S, van der Zee J, Sieben A, Philtjens S, Heeman B, et al. Loss of TBK1 is a frequent cause of frontotemporal dementia in a Belgian cohort. Neurology 2015;85(24):2116-2125.

(48) Yu C, Bird TD, Bekris LM, Montine TJ, Leverenz JB, Steinbart E, et al. The spectrum of mutations in progranulin: a collaborative study screening 545 cases of neurodegeneration. Arch Neurol 2010;67(2):161-170.

(49) Cruts M, Theuns J, Van Broeckhoven C. Locus‐specific mutation databases for neurodegenerative brain diseases. Hum Mutat 2012;33(9):1340-1344.

(50) Mackenzie IR, Baker M, Pickering-Brown S, Hsiung GR, Lindholm C, Dwosh E, et al. The neuropathology of frontotemporal lobar degeneration caused by mutations in the progranulin gene. Brain 2006;129(11):3081-3090.

(51) Bronner IF, Rizzu P, Seelaar H, van Mil SE, Anar B, Azmani A, et al. Progranulin mutations in Dutch familial frontotemporal lobar degeneration. European Journal of Human Genetics 2007;15(3):369.

(52) Skoglund L, Matsui T, Freeman SH, Wallin A, Blom ES, Frosch MP, et al. Novel progranulin mutation detected in 2 patients with FTLD. Alzheimer Dis Assoc Disord 2011;25(2):173.

(53) Cruts M, Gijselinck I, Van Der Zee J, Engelborghs S, Wils H, Pirici D, et al. Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Nature 2006;442(7105):920-924.

(54) Mukherjee O, Wang J, Gitcho M, Chakraverty S, Taylor‐Reinwald L, Shears S, et al. Molecular characterization of novel progranulin (GRN) mutations in frontotemporal dementia. Hum Mutat 2008;29(4):512-521.

(55) Wang J, Van Damme P, Cruchaga C, Gitcho MA, Vidal JM, Seijo‐Martínez M, et al. Pathogenic cysteine mutations affect progranulin function and production of mature granulins. J Neurochem 2010;112(5):1305-1315.

(56) Seelaar H, Rohrer JD, Pijnenburg YA, Fox NC, Van Swieten JC. Clinical, genetic and pathological heterogeneity of frontotemporal dementia: a review. Journal of Neurology, Neurosurgery & Psychiatry 2011;82(5):476-486.

236

(57) Nguyen AD, Nguyen TA, Martens LH, Mitic LL, Farese Jr RV. Progranulin: at the interface of neurodegenerative and metabolic diseases. Trends in Endocrinology & Metabolism 2013;24(12):597-606.

(58) Meeter LH, Patzke H, Loewen G, Dopper EG, Pijnenburg YA, van Minkelen R, et al. Progranulin levels in plasma and cerebrospinal fluid in granulin mutation carriers. Dementia and geriatric cognitive disorders extra 2016;6(2):330-340.

(59) Finch N, Baker M, Crook R, Swanson K, Kuntz K, Surtees R, et al. Plasma progranulin levels predict progranulin mutation status in frontotemporal dementia patients and asymptomatic family members. Brain 2009;132(3):583-591.

(60) Almeida S, Zhang Z, Coppola G, Mao W, Futai K, Karydas A, et al. Induced pluripotent stem cell models of progranulin-deficient frontotemporal dementia uncover specific reversible neuronal defects. Cell reports 2012;2(4):789-798.

(61) Wauters E, Van Mossevelde S, Van der Zee J, Cruts M, Van Broeckhoven C. Modifiers of GRN-Associated Frontotemporal Lobar Degeneration. Trends Mol Med 2017;23(10):962-979.

(62) Edwards SL, Beesley J, French JD, Dunning AM. Beyond GWASs: illuminating the dark road from association to function. The American Journal of Human Genetics 2013;93(5):779-797.

(63) Stagi M, Klein ZA, Gould TJ, Bewersdorf J, Strittmatter SM. Lysosome size, motility and stress response regulated by fronto-temporal dementia modifier TMEM106B. Molecular and Cellular Neuroscience 2014;61:226-240.

(64) Nicholson AM, Finch NA, Wojtas A, Baker MC, Perkerson III RB, Castanedes-Casey M, et al. TMEM106B p. T185S regulates TMEM106B protein levels: implications for frontotemporal dementia. J Neurochem 2013;126(6):781-791.

(65) Van Deerlin VM, Sleiman PM, Martinez-Lage M, Chen-Plotkin A, Wang L, Graff-Radford NR, et al. Common variants at 7p21 are associated with frontotemporal lobar degeneration with TDP-43 inclusions. Nat Genet 2010;42(3):234.

(66) Finch N, Carrasquillo MM, Baker M, Rutherford NJ, Coppola G, Dejesus-Hernandez M, et al. TMEM106B regulates progranulin levels and the penetrance of FTLD in GRN mutation carriers. Neurology 2011;76(5):467-474.

(67) Cruchaga C, Graff C, Chiang H, Wang J, Hinrichs AL, Spiegel N, et al. Association of TMEM106B gene polymorphism with age at onset in granulin mutation carriers and plasma granulin protein levels. Arch Neurol 2011;68(5):581-586.

237

(68) Gallagher MD, Posavi M, Huang P, Unger TL, Berlyand Y, Gruenewald AL, et al. A dementia-associated risk variant near TMEM106B alters chromatin architecture and gene expression. The American Journal of Human Genetics 2017;101(5):643-663.

(69) Chen-Plotkin AS, Unger TL, Gallagher MD, Bill E, Kwong LK, Volpicelli-Daley L, et al. TMEM106B, the risk gene for frontotemporal dementia, is regulated by the microRNA-132/212 cluster and affects progranulin pathways. Journal of Neuroscience 2012;32(33):11213-11227.

(70) Nicholson AM, Rademakers R. What we know about TMEM106B in neurodegeneration. Acta Neuropathol 2016;132(5):639-651.

(71) Xu W, Han S, Zhang C, Li J, Wang Y, Tan C, et al. The FAM171A2 gene is a key regulator of progranulin expression and modifies the risk of multiple neurodegenerative diseases. Science Advances 2020;6(43):eabb3063.

(72) Rademakers R, Eriksen JL, Baker M, Robinson T, Ahmed Z, Lincoln SJ, et al. Common variation in the miR-659 binding-site of GRN is a major risk factor for TDP43-positive frontotemporal dementia. Hum Mol Genet 2008;17(23):3631-3642.

(73) Hsiung GR, Fok A, Feldman HH, Rademakers R, Mackenzie IR. rs5848 polymorphism and serum progranulin level. J Neurol Sci 2011;300(1-2):28-32.

(74) Rollinson S, Rohrer JD, van der Zee J, Sleegers K, Mead S, Engelborghs S, et al. No association of PGRN 3′ UTR rs5848 in frontotemporal lobar degeneration. Neurobiol Aging 2011;32(4):754-755.

(75) Rademakers R, Baker M, Gass J, Adamson J, Huey ED, Momeni P, et al. Phenotypic variability associated with progranulin haploinsufficiency in patients with the common 1477C→ T (Arg493X) mutation: an international initiative. The Lancet Neurology 2007;6(10):857-868.

(76) Banzhaf-Strathmann J, Claus R, Mücke O, Rentzsch K, van der Zee J, Engelborghs S, et al. Promoter DNA methylation regulates progranulin expression and is altered in FTLD. Acta neuropathologica communications 2013;1(1):16.

(77) Galimberti D, D’Addario C, Dell’Osso B, Fenoglio C, Marcone A, Cerami C, et al. Progranulin gene (GRN) promoter methylation is increased in patients with sporadic frontotemporal lobar degeneration. Neurological Sciences 2013;34(6):899-903.

(78) Hu F, Padukkavidana T, Vgter CB, Brady OA, Zheng Y, Mackenzie IR, et al. Sortilin-mediated endocytosis determines levels of the frontotemporal dementia protein, progranulin. Neuron 2010;68(4):654-667.

238

(79) Carrasquillo MM, Nicholson AM, Finch N, Gibbs JR, Baker M, Rutherford NJ, et al. Genome-wide screen identifies rs646776 near sortilin as a regulator of progranulin levels in human plasma. The American Journal of Human Genetics 2010;87(6):890-897.

(80) Hiraiwa M, Obrien JS, Kishimoto Y, Galdzicka M, Fluharty AL, Ginns EI, et al. Isolation, characterization, and proteolysis of human prosaposin, the precursor of saposins (sphingolipid activator proteins). Arch Biochem Biophys 1993;304(1):110-116.

(81) Zhou X, Sun L, de Oliveira FB, Qi X, Brown WJ, Smolka MB, et al. Prosaposin facilitates sortilin-independent lysosomal trafficking of progranulin. J Cell Biol 2015;210(6):991-1002.

(82) Nicholson AM, Finch NA, Almeida M, Perkerson RB, Van Blitterswijk M, Wojtas A, et al. Prosaposin is a regulator of progranulin levels and oligomerization. Nature communications 2016;7:11992.

(83) Almeida S, Zhou L, Gao F. Progranulin, a glycoprotein deficient in frontotemporal dementia, is a novel substrate of several protein disulfide isomerase family proteins. PloS One 2011;6(10):e26454.

(84) Hsiung GR, Feldman HH. GRN-related frontotemporal dementia. GeneReviews®[Internet]: University of Washington, Seattle; 2013.

(85) Josephs KA, Ahmed Z, Katsuse O, Parisi JF, Boeve BF, Knopman DS, et al. Neuropathologic features of frontotemporal lobar degeneration with ubiquitin-positive inclusions with progranulin gene (PGRN) mutations. Journal of Neuropathology & Experimental Neurology 2007;66(2):142-151.

(86) Mesulam M, Johnson N, Krefft TA, Gass JM, Cannon AD, Adamson JL, et al. Progranulin mutations in primary progressive aphasia: the PPA1 and PPA3 families. Arch Neurol 2007;64(1):43-47.

(87) Masellis M, Momeni P, Meschino W, Heffner Jr R, Elder J, Sato C, et al. Novel splicing mutation in the progranulin gene causing familial corticobasal syndrome. Brain 2006;129(11):3115-3123.

(88) Jacova C, Hsiung GR, Tawankanjanachot I, Dinelle K, McCormick S, Gonzalez M, et al. Anterior brain glucose hypometabolism predates dementia in progranulin mutation carriers. Neurology 2013;81(15):1322-1331.

(89) Caroppo P, Le Ber I, Camuzat A, Clot F, Naccache L, Lamari F, et al. Extensive white matter involvement in patients with frontotemporal lobar degeneration: think progranulin. JAMA neurology 2014;71(12):1562-1566.

239

(90) Mackenzie IR. The neuropathology and clinical phenotype of FTD with progranulin mutations. Acta Neuropathol 2007;114(1):49-54.

(91) Cairns NJ, Neumann M, Bigio EH, Holm IE, Troost D, Hatanpaa KJ, et al. TDP-43 in familial and sporadic frontotemporal lobar degeneration with ubiquitin inclusions. The American journal of pathology 2007;171(1):227-240.

(92) Lui H, Zhang J, Makinson SR, Cahill MK, Kelley KW, Huang H, et al. Progranulin deficiency promotes circuit-specific synaptic pruning by microglia via complement activation. Cell 2016;165(4):921-935.

(93) Götzl JK, Mori K, Damme M, Fellerer K, Tahirovic S, Kleinberger G, et al. Common pathobiochemical hallmarks of progranulin-associated frontotemporal lobar degeneration and neuronal ceroid lipofuscinosis. Acta Neuropathol 2014;127(6):845-860.

(94) Chen-Plotkin AS, Xiao J, Geser F, Martinez-Lage M, Grossman M, Unger T, et al. Brain progranulin expression in GRN-associated frontotemporal lobar degeneration. Acta Neuropathol 2010;119(1):111-122.

(95) Smith KR, Damiano J, Franceschetti S, Carpenter S, Canafoglia L, Morbin M, et al. Strikingly different clinicopathological phenotypes determined by progranulin-mutation dosage. The American Journal of Human Genetics 2012;90(6):1102-1107.

(96) Huin V, Barbier M, Bottani A, Lobrinus JA, Clot F, Lamari F, et al. Homozygous GRN mutations: new phenotypes and new insights into pathological and molecular mechanisms. Brain 2020;143(1):303-319.

(97) Bateman A, Belcourt D, Bennett H, Lazure C, Solomon S. Granulins, a novel class of peptide from leukocytes. Biochem Biophys Res Commun 1990;173(3):1161-1168.

(98) Eichinger L, Pachebat JA, Glöckner G, Rajandream M, Sucgang R, Berriman M, et al. The genome of the social amoeba Dictyostelium discoideum. Nature 2005;435(7038):43-57.

(99) Nichols SA, Dirks W, Pearse JS, King N. Early evolution of animal cell signaling and adhesion genes. Proceedings of the National Academy of Sciences 2006;103(33):12451-12456.

(100) King N, Westbrook MJ, Young SL, Kuo A, Abedin M, Chapman J, et al. The genome of the choanoflagellate Monosiga brevicollis and the origin of metazoans. Nature 2008;451(7180):783-788.

240

(101) Zanocco-Marani T, Bateman A, Romano G, Valentinis B, He Z, Baserga R. Biological activities and signaling pathways of the granulin/epithelin precursor. Cancer Res 1999;59(20):5331-5340.

(102) He Z, Ong CH, Halper J, Bateman A. Progranulin is a mediator of the wound response. Nat Med 2003;9(2):225-229.

(103) Zhu J, Nathan C, Jin W, Sim D, Ashcroft GS, Wahl SM, et al. Conversion of proepithelin to epithelins: roles of SLPI and elastase in host defense and wound repair. Cell 2002;111(6):867-878.

(104) Qin J, Díaz-Cueto L, Schwarze J, Takahashi Y, Imai M, Isuzugawa K, et al. Effects of progranulin on blastocyst hatching and subsequent adhesion and outgrowth in the mouse. Biol Reprod 2005;73(3):434-442.

(105) Desmarais JA, Cao M, Bateman A, Murphy BD. Spatiotemporal expression pattern of progranulin in embryo implantation and placenta formation suggests a role in cell proliferation, remodeling, and angiogenesis. Reproduction 2008;136(2):247-258.

(106) Chang MC, Srinivasan K, Friedman BA, Suto E, Modrusan Z, Lee WP, et al. Progranulin deficiency causes impairment of autophagy and TDP-43 accumulation. J Exp Med 2017;214(9):2611-2628.

(107) Evers BM, Rodriguez-Navas C, Tesla RJ, Prange-Kiel J, Wasser CR, Yoo KS, et al. Lipidomic and Transcriptomic Basis of Lysosomal Dysfunction in Progranulin Deficiency. Cell Reports 2017;20(11):2565-2574.

(108) He Z, Bateman A. Progranulin gene expression regulates epithelial cell growth and promotes tumor growth in vivo. Cancer Res 1999;59(13):3222-3229.

(109) Xu S, Tang D, Chamberlain S, Pronk G, Masiarz FR, Kaur S, et al. The granulin/epithelin precursor abrogates the requirement for the insulin-like growth factor 1 receptor for growth in vitro. J Biol Chem 1998;273(32):20078-20083.

(110) Plowman GD, Green JM, Neubauer MG, Buckley SD, McDonald VL, Todaro GJ, et al. The epithelin precursor encodes two proteins with opposing activities on epithelial cell growth. Journal of Biological Chemistry 1992;267(18):13073-13078.

(111) Zhou J, Gao G, Crabb JW, Serrero G. Purification of an autocrine growth factor homologous with mouse epithelin precursor from a highly tumorigenic cell line. Journal of Biological Chemistry 1993;268(15):10863-10869.

(112) Baba T, Hoff III HB, Nemoto H, Lee H, Orth J, Arai Y, et al. Acrogranin, an acrosomal cysteine‐rich glycoprotein, is the precursor of the growth‐modulating

241

peptides, granulins, and epithelins, and is expressed in somatic as well as male germ cells. Mol Reprod Dev 1993;34(3):233-243.

(113) Parnell PG, Wunderlich J, Carter B, Halper J. Transforming growth factor e: amino acid analysis and partial amino acid sequence. Growth Factors 1992;7(1):65-72.

(114) Bhandari V, Giaid A, Bateman A. The complementary deoxyribonucleic acid sequence, tissue distribution, and cellular localization of the rat granulin precursor. Endocrinology 1993;133(6):2682-2689.

(115) Hrabal R, Chen Z, James S, Bennett HP, Ni F. The hairpin stack fold, a novel protein architecture for a new family of protein growth factors. Nat Struct Biol 1996;3(9):747-752.

(116) Belcourt DR, Lazure C, Bennett HP. Isolation and primary structure of the three major forms of granulin-like peptides from hematopoietic tissues of a teleost fish (Cyprinus carpio). J Biol Chem 1993;268(13):9230-9237.

(117) Petkau TL, Neal SJ, Orban PC, MacDonald JL, Hill AM, Lu G, et al. Progranulin expression in the developing and adult murine brain. J Comp Neurol 2010;518(19):3931-3947.

(118) Holler CJ, Taylor G, Deng Q, Kukar T. Intracellular Proteolysis of Progranulin Generates Stable, Lysosomal Granulins that Are Haploinsufficient in Patients with Frontotemporal Dementia Caused by GRN Mutations. eNeuro 2017;4(4):ENEURO. 0100-17.2017.

(119) Götzl JK, Colombo A, Fellerer K, Reifschneider A, Werner G, Tahirovic S, et al. Early lysosomal maturation deficits in microglia triggers enhanced lysosomal activity in other brain cells of progranulin knockout mice. Molecular neurodegeneration 2018;13(48).

(120) Zhang J, Velmeshev D, Hashimoto K, Huang Y, Hofmann JW, Shi X, et al. Neurotoxic microglia promote TDP-43 proteinopathy in progranulin deficiency. Nature 2020;588(7838):459-465.

(121) Suh H, Choi N, Tarassishin L, Lee SC. Regulation of progranulin expression in human microglia and proteolysis of progranulin by matrix metalloproteinase-12 (MMP-12). PloS One 2012;7(4):e35115.

(122) Frew J, Baradaran-Heravi A, Balgi AD, Wu X, Yan TD, Arns S, et al. Premature termination codon readthrough upregulates progranulin expression and improves lysosomal function in preclinical models of GRN deficiency. Molecular Neurodegeneration 2020;15(21).

242

(123) Petkau TL, Blanco J, Leavitt BR. Conditional loss of progranulin in neurons is not sufficient to cause neuronal ceroid lipofuscinosis-like neuropathology in mice. Neurobiol Dis 2017;106:14-22.

(124) Arrant AE, Filiano AJ, Unger DE, Young AH, Roberson ED. Restoring neuronal progranulin reverses deficits in a mouse model of frontotemporal dementia. Brain 2017;140(5):1447-1465.

(125) Martens LH, Zhang J, Barmada SJ, Zhou P, Kamiya S, Sun B, et al. Progranulin deficiency promotes neuroinflammation and neuron loss following toxin-induced injury. J Clin Invest 2012;122(11):3955-3959.

(126) Minami SS, Min S, Krabbe G, Wang C, Zhou Y, Asgarov R, et al. Progranulin protects against amyloid β deposition and toxicity in Alzheimer's disease mouse models. Nat Med 2014;20(10):1157-1164.

(127) Petkau TL, Kosior N, de Asis K, Connolly C, Leavitt BR. Selective depletion of microglial progranulin in mice is not sufficient to cause neuronal ceroid lipofuscinosis or neuroinflammation. Journal of neuroinflammation 2017;14(255).

(128) Arrant AE, Filiano AJ, Patel AR, Hoffmann MQ, Boyle NR, Kashyap SN, et al. Reduction of microglial progranulin does not exacerbate pathology or behavioral deficits in neuronal progranulin-insufficient mice. Neurobiol Dis 2019;124:152-162.

(129) Krabbe G, Minami SS, Etchegaray JI, Taneja P, Djukic B, Davalos D, et al. Microglial NFκB-TNFα hyperactivation induces obsessive–compulsive behavior in mouse models of progranulin-deficient frontotemporal dementia. Proceedings of the National Academy of Sciences 2017;114(19):5029-5034.

(130) Erö C, Gewaltig M, Keller D, Markram H. A cell atlas for the mouse brain. Frontiers in neuroinformatics 2018;12:84.

(131) Tanaka Y, Matsuwaki T, Yamanouchi K, Nishihara M. Increased lysosomal biogenesis in activated microglia and exacerbated neuronal damage after traumatic brain injury in progranulin-deficient mice. Neuroscience 2013;250:8-19.

(132) Belcastro V, Siciliano V, Gregoretti F, Mithbaokar P, Dharmalingam G, Berlingieri S, et al. Transcriptional gene network inference from a massive dataset elucidates transcriptome organization and gene function. Nucleic Acids Res 2011;39(20):8677-8688.

(133) Sardiello M, Palmieri M, di Ronza A, Medina DL, Valenza M, Gennarino VA, et al. A gene network regulating lysosomal biogenesis and function. Science 2009;325(5939):473-477.

243

(134) Roczniak-Ferguson A, Petit CS, Froehlich F, Qian S, Ky J, Angarola B, et al. The transcription factor TFEB links mTORC1 signaling to transcriptional control of lysosome homeostasis. Science signaling 2012;5(228):ra42.

(135) Holler CJ, Taylor G, McEachin ZT, Deng Q, Watkins WJ, Hudson K, et al. Trehalose upregulates progranulin expression in human and mouse models of GRN haploinsufficiency: a novel therapeutic lead to treat frontotemporal dementia. Molecular neurodegeneration 2016;11(1):46.

(136) Elia LP, Mason AR, Alijagic A, Finkbeiner S. Genetic regulation of neuronal progranulin reveals a critical role for the autophagy-lysosome pathway. Journal of Neuroscience 2019;39(17):3332-3344.

(137) Naphade SB, Kigerl KA, Jakeman LB, Kostyk SK, Popovich PG, Kuret J. Progranulin expression is upregulated after spinal contusion in mice. Acta Neuropathol 2010;119(1):123-133.

(138) Bhandari V, Daniel R, Lim PS, Bateman A. Structural and functional analysis of a promoter of the human granulin/epithelin gene. Biochem J 1996;319(2):441-447.

(139) Frampton G, Invernizzi P, Bernuzzi F, Pae HY, Quinn M, Horvat D, et al. Interleukin-6-driven progranulin expression increases cholangiocarcinoma growth by an Akt-dependent mechanism. Gut 2012;61(2):268-277.

(140) Liu F, Zhang W, Yang F, Feng T, Zhou M, Yu Y, et al. Interleukin-6-stimulated progranulin expression contributes to the malignancy of hepatocellular carcinoma cells by activating mTOR signaling. Scientific reports 2016;6:21260.

(141) Ong CH, He Z, Kriazhev L, Shan X, Palfree RG, Bateman A. Regulation of progranulin expression in myeloid cells. American Journal of Physiology-Regulatory, Integrative and Comparative Physiology 2006;291(6):1602.

(142) Tarassishin L, Suh H, Lee SC. LPS and IL‐1 differentially activate mouse and human astrocytes: Role of CD14. Glia 2014;62(6):999-1013.

(143) Capell A, Fellerer K, Haass C. Progranulin transcripts with short and long 5′ untranslated regions (UTRs) are differentially expressed via posttranscriptional and translational repression. J Biol Chem 2014;289(37):25879-25889.

(144) Kanazawa M, Kawamura K, Takahashi T, Miura M, Tanaka Y, Koyama M, et al. Multiple therapeutic effects of progranulin on experimental acute ischaemic stroke. Brain 2015;138(7):1932-1948.

244

(145) Songsrirote K, Li Z, Ashford D, Bateman A, Thomas-Oates J. Development and application of mass spectrometric methods for the analysis of progranulin N-glycosylation. Journal of proteomics 2010;73(8):1479-1490.

(146) Jia Y. Analysis of the neuroprotective function and the N-glycosylation progranulin (PGRN) (University of British Columbia). 2009.

(147) Petoukhov E, Fernando S, Mills F, Shivji F, Hunter D, Krieger C, et al. Activity-dependent secretion of progranulin from synapses. J Cell Sci 2013;126(23):5412-5421.

(148) Zheng Y, Brady OA, Meng PS, Mao Y, Hu F. C-terminus of progranulin interacts with the beta-propeller region of sortilin to regulate progranulin trafficking. PloS One 2011;6(6):e21023.

(149) Zhou X, Paushter DH, Feng T, Sun L, Reinheckel T, Hu F. Lysosomal processing of progranulin. Molecular neurodegeneration 2017;12(62).

(150) Lee CW, Stankowski JN, Chew J, Cook CN, Lam Y, Almeida S, et al. The lysosomal protein cathepsin L is a progranulin protease. Molecular neurodegeneration 2017;12(55).

(151) Wang X, Li X, Xu L, Zhan Y, Yaish-Ohad S, Erhardt JA, et al. Up-regulation of secretory leukocyte protease inhibitor (SLPI) in the brain after ischemic stroke: adenoviral expression of SLPI protects brain from ischemic injury. Mol Pharmacol 2003;64(4):833-840.

(152) Brady OA, Zheng Y, Murphy K, Huang M, Hu F. The frontotemporal lobar degeneration risk factor, TMEM106B, regulates lysosomal morphology and function. Hum Mol Genet 2013;22(4):685-695.

(153) Schwenk BM, Lang CM, Hogl S, Tahirovic S, Orozco D, Rentzsch K, et al. The FTLD risk factor TMEM106B and MAP6 control dendritic trafficking of lysosomes. EMBO J 2014;33(5):450-467.

(154) Busch JI, Unger TL, Jain N, Tyler Skrinak R, Charan RA, Chen-Plotkin AS. Increased expression of the frontotemporal dementia risk factor TMEM106B causes C9orf72-dependent alterations in lysosomes. Hum Mol Genet 2016;25(13):2681-2697.

(155) Zhou X, Sun L, Brady OA, Murphy KA, Hu F. Elevated TMEM106B levels exaggerate lipofuscin accumulation and lysosomal dysfunction in aged mice with progranulin deficiency. Acta neuropathologica communications 2017;5(9).

(156) Feng T, Mai S, Roscoe JM, Sheng RR, Ullah M, Zhang J, et al. Loss of TMEM106B and PGRN leads to severe lysosomal abnormalities and neurodegeneration in mice. EMBO Rep 2020;21(10):e50219.

245

(157) Werner G, Damme M, Schludi M, Gnörich J, Wind K, Fellerer K, et al. Loss of TMEM106B potentiates lysosomal and FTLD‐like pathology in progranulin‐deficient mice. EMBO Rep 2020;21(10):e50241.

(158) Zhou X, Nicholson AM, Ren Y, Brooks M, Jiang P, Zuberi A, et al. Loss of TMEM106B leads to myelination deficits: implications for frontotemporal dementia treatment strategies. Brain 2020;146(6):1905-1919.

(159) Ahmed Z, Sheng H, Xu Y, Lin W, Innes AE, Gass J, et al. Accelerated lipofuscinosis and ubiquitination in granulin knockout mice suggest a role for progranulin in successful aging. The American journal of pathology 2010;177(1):311-324.

(160) Kayasuga Y, Chiba S, Suzuki M, Kikusui T, Matsuwaki T, Yamanouchi K, et al. Alteration of behavioural phenotype in mice by targeted disruption of the progranulin gene. Behav Brain Res 2007;185(2):110-118.

(161) Yin F, Banerjee R, Thomas B, Zhou P, Qian L, Jia T, et al. Exaggerated inflammation, impaired host defense, and neuropathology in progranulin-deficient mice. J Exp Med 2010;207(1):117-128.

(162) Yin F, Dumont M, Banerjee R, Ma Y, Li H, Lin MT, et al. Behavioral deficits and progressive neuropathology in progranulin-deficient mice: a mouse model of frontotemporal dementia. The FASEB Journal 2010;24(12):4639-4647.

(163) Petkau TL, Neal SJ, Milnerwood A, Mew A, Hill AM, Orban P, et al. Synaptic dysfunction in progranulin-deficient mice. Neurobiol Dis 2012;45(2):711-722.

(164) Tanaka Y, Chambers JK, Matsuwaki T, Yamanouchi K, Nishihara M. Possible involvement of lysosomal dysfunction in pathological changes of the brain in aged progranulin-deficient mice. Acta neuropathologica communications 2014;2(78).

(165) Nguyen AD, Nguyen TA, Zhang J, Devireddy S, Zhou P, Karydas AM, et al. Murine knockin model for progranulin-deficient frontotemporal dementia with nonsense-mediated mRNA decay. Proceedings of the National Academy of Sciences 2018;115(12):E2849-E2858.

(166) Pickford F, Marcus J, Camargo LM, Xiao Q, Graham D, Mo J, et al. Progranulin is a chemoattractant for microglia and stimulates their endocytic activity. The American journal of pathology 2011;178(1):284-295.

(167) Tang W, Lu Y, Tian Q, Zhang Y, Guo F, Liu G, et al. The growth factor progranulin binds to TNF receptors and is therapeutic against inflammatory arthritis in mice. Science 2011;332(6028):478-484.

246

(168) Jian J, Zhao S, Tian Q, Gonzalez-Gugel E, Mundra JJ, Uddin SM, et al. Progranulin directly binds to the CRD2 and CRD3 of TNFR extracellular domains. FEBS Lett 2013;587(21):3428-3436.

(169) Chen X, Chang J, Deng Q, Xu J, Nguyen TA, Martens LH, et al. Progranulin does not bind tumor necrosis factor (TNF) receptors and is not a direct regulator of TNF-dependent signaling or bioactivity in immune or neuronal cells. Journal of neuroscience 2013;33(21):9202-9213.

(170) Etemadi N, Webb A, Bankovacki A, Silke J, Nachbur U. Progranulin does not inhibit TNF and lymphotoxin‐α signalling through TNF receptor 1. Immunol Cell Biol 2013;91(10):661-664.

(171) Lang I, Füllsack S, Wajant H. Lack of evidence for a direct interaction of progranulin and tumor necrosis factor receptor-1 and tumor necrosis factor receptor-2 from cellular binding studies. Frontiers in immunology 2018;9:793.

(172) Wei F, Zhang Y, Jian J, Mundra JJ, Tian Q, Lin J, et al. PGRN protects against colitis progression in mice in an IL-10 and TNFR2 dependent manner. Scientific reports 2014;4(7023).

(173) Chen Q, Cai J, Li X, Song A, Guo H, Sun Q, et al. Progranulin promotes regeneration of inflammatory periodontal bone defect in rats via anti-inflammation, osteoclastogenic inhibition, and osteogenic promotion. Inflammation 2019;42(1):221-234.

(174) Liu L, Qu Y, Liu Y, Zhao H, Ma H, Noor AF, et al. Atsttrin reduces lipopolysaccharide-induced neuroinflammation by inhibiting the nuclear factor kappa B signaling pathway. Neural regeneration research 2019;14(11):1994-2002.

(175) Telpoukhovskaia MA, Liu K, Sayed FA, Etchegaray JI, Xie M, Zhan L, et al. Discovery of small molecules that normalize the transcriptome and enhance cysteine cathepsin activity in progranulin-deficient microglia. Scientific reports 2020;10(13688).

(176) Tanaka Y, Matsuwaki T, Yamanouchi K, Nishihara M. Exacerbated inflammatory responses related to activated microglia after traumatic brain injury in progranulin-deficient mice. Neuroscience 2013;231:49-60.

(177) Sarlus H, Heneka MT. Microglia in Alzheimer’s disease. J Clin Invest 2017;127(9):3240-3249.

(178) Huang M, Modeste E, Dammer E, Merino P, Taylor G, Duong DM, et al. Network analysis of the progranulin-deficient mouse brain proteome reveals pathogenic mechanisms shared in human frontotemporal dementia caused by GRN mutations. Acta Neuropathologica Communications 2020;8(163).

247

(179) Klein ZA, Takahashi H, Ma M, Stagi M, Zhou M, Lam TT, et al. Loss of TMEM106B ameliorates lysosomal and frontotemporal dementia-related phenotypes in progranulin-deficient mice. Neuron 2017;95(2):281-296.

(180) Tanaka Y, Suzuki G, Matsuwaki T, Hosokawa M, Serrano G, Beach TG, et al. Progranulin regulates lysosomal function and biogenesis through acidification of lysosomes. Hum Mol Genet 2017;26(5):969-988.

(181) Marschallinger J, Iram T, Zardeneta M, Lee SE, Lehallier B, Haney MS, et al. Lipid-droplet-accumulating microglia represent a dysfunctional and proinflammatory state in the aging brain. Nat Neurosci 2020;23:194–208.

(182) Liu L, Zhang K, Sandoval H, Yamamoto S, Jaiswal M, Sanz E, et al. Glial lipid droplets and ROS induced by mitochondrial defects promote neurodegeneration. Cell 2015;160(1-2):177-190.

(183) den Brok MH, Raaijmakers TK, Collado-Camps E, Adema GJ. Lipid droplets as immune modulators in myeloid cells. Trends Immunol 2018;39(5):380-392.

(184) Budge KM, Neal ML, Richardson JR, Safadi FF. Glycoprotein NMB: an emerging role in neurodegenerative disease. Mol Neurobiol 2018;55(6):5167-5176.

(185) Moloney EB, Moskites A, Ferrari EJ, Isacson O, Hallett PJ. The glycoprotein GPNMB is selectively elevated in the substantia nigra of Parkinson's disease patients and increases after lysosomal stress. Neurobiol Dis 2018;120:1-11.

(186) Jian J, Tian Q, Hettinghouse A, Zhao S, Liu H, Wei J, et al. Progranulin recruits HSP70 to β-glucocerebrosidase and is therapeutic against Gaucher disease. EBioMedicine 2016;13:212-224.

(187) Zhou X, Paushter DH, Pagan MD, Kim D, Santos MN, Lieberman RL, et al. Progranulin deficiency leads to reduced glucocerebrosidase activity. PloS One 2019;14(7):e0212382.

(188) Hendrickx DA, Van Scheppingen J, van der Poel M, Bossers K, Schuurman KG, van Eden CG, et al. Gene expression profiling of multiple sclerosis pathology identifies early patterns of demyelination surrounding chronic active lesions. Frontiers in immunology 2017;8:1810.

(189) Weinhard L, di Bartolomei G, Bolasco G, Machado P, Schieber NL, Neniskyte U, et al. Microglia remodel synapses by presynaptic trogocytosis and spine head filopodia induction. Nature communications 2018;9(1):1228.

248

(190) Arrant AE, Filiano AJ, Warmus BA, Hall AM, Roberson ED. Progranulin haploinsufficiency causes biphasic social dominance abnormalities in the tube test. Genes, Brain and Behavior 2016;15(6):588-603.

(191) Konuk N, Tekın IO, Ozturk U, Atik L, Atasoy N, Bektas S, et al. Plasma levels of tumor necrosis factor-alpha and interleukin-6 in obsessive compulsive disorder. Mediators Inflamm 2007;2007(65704).

(192) Figee M, Luigjes J, Smolders R, Valencia-Alfonso C, Van Wingen G, De Kwaasteniet B, et al. Deep brain stimulation restores frontostriatal network activity in obsessive-compulsive disorder. Nat Neurosci 2013;16(4):386-387.

(193) Liu L, Guo H, Song A, Huang J, Zhang Y, Jin S, et al. Progranulin inhibits LPS-induced macrophage M1 polarization via NF-кB and MAPK pathways. BMC Immunology 2020;21(1):32.

(194) Windisch M. We can treat Alzheimer's disease successfully in mice but not in men: failure in translation? A perspective. Neurodegenerative Diseases 2014;13(2-3):147-150.

(195) Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 2007;131(5):861-872.

(196) Chambers SM, Fasano CA, Papapetrou EP, Tomishima M, Sadelain M, Studer L. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat Biotechnol 2009;27(3):275-280.

(197) Julia T, Wang M, Pimenova AA, Bowles KR, Hartley BJ, Lacin E, et al. An efficient platform for astrocyte differentiation from human induced pluripotent stem cells. Stem cell reports 2017;9(2):600-614.

(198) Abud EM, Ramirez RN, Martinez ES, Healy LM, Nguyen CH, Newman SA, et al. iPSC-derived human microglia-like cells to study neurological diseases. Neuron 2017;94(2):278-293.

(199) Raitano S, Ordovàs L, De Muynck L, Guo W, Espuny-Camacho I, Geraerts M, et al. Restoration of progranulin expression rescues cortical neuron generation in an induced pluripotent stem cell model of frontotemporal dementia. Stem cell reports 2015;4(1):16-24.

(200) Valdez C, Wong YC, Schwake M, Bu G, Wszolek ZK, Krainc D. Progranulin-mediated deficiency of cathepsin D results in FTD and NCL-like phenotypes in neurons derived from FTD patients. Hum Mol Genet 2017;26(24):4861-4872.

249

(201) Almeida S, Gao F, Coppola G, Gao F. Suberoylanilide hydroxamic acid increases progranulin production in iPSC-derived cortical neurons of frontotemporal dementia patients. Neurobiol Aging 2016;42:35-40.

(202) Beel S, Moisse M, Damme M, De Muynck L, Robberecht W, Van Den Bosch L, et al. Progranulin functions as a cathepsin D chaperone to stimulate axonal outgrowth in vivo. Hum Mol Genet 2017;26(15):2850-2863.

(203) Zhou X, Paushter DH, Feng T, Pardon CM, Mendoza CS, Hu F. Regulation of cathepsin D activity by the FTLD protein progranulin. Acta Neuropathol 2017;134(1):151-153.

(204) Lee WC, Almeida S, Prudencio M, Caulfield TR, Zhang Y, Tay WM, et al. Targeted manipulation of the sortilin–progranulin axis rescues progranulin haploinsufficiency. Hum Mol Genet 2013;23(6):1467-1478.

(205) Cenik B, Sephton CF, Dewey CM, Xian X, Wei S, Yu K, et al. Suberoylanilide hydroxamic acid (vorinostat) up-regulates progranulin transcription: rational therapeutic approach to frontotemporal dementia. J Biol Chem 2011;286(18):16101-16108.

(206) Rohrer JD, Warren JD, Barnes J, Mead S, Beck J, Pepple T, et al. Mapping the progression of progranulin-associated frontotemporal lobar degeneration. Nature Clinical Practice Neurology 2008;4(8):455-460.

(207) Borroni B, Alberici A, Cercignani M, Premi E, Serra L, Cerini C, et al. Granulin mutation drives brain damage and reorganization from preclinical to symptomatic FTLD. Neurobiol Aging 2012;33(10):2506-2520.

(208) Butler VJ, Cortopassi WA, Argouarch AR, Ivry SL, Craik CS, Jacobson MP, et al. Progranulin stimulates the in vitro maturation of pro-cathepsin D at acidic pH. J Mol Biol 2019;431(5):1038-1047.

(209) Altmann C, Vasic V, Hardt S, Heidler J, Häussler A, Wittig I, et al. Progranulin promotes peripheral nerve regeneration and reinnervation: role of notch signaling. Molecular neurodegeneration 2016;11(1):69.

(210) Bernier L, York EM, MacVicar BA. Immunometabolism in the brain: how metabolism shapes microglial function. Trends Neurosci 2020;43(11):854-869.

(211) Afridi R, Lee W, Suk K. Microglia gone awry: linking immunometabolism to neurodegeneration. Frontiers in Cellular Neuroscience 2020;14(246).

(212) Götzl JK, Brendel M, Werner G, Parhizkar S, Monasor LS, Kleinberger G, et al. Opposite microglial activation stages upon loss of PGRN or TREM2 result in reduced cerebral glucose metabolism. EMBO molecular medicine 2019;11(6):e9711.

250

(213) Lee SE, Sias AC, Kosik EL, Flagan TM, Deng J, Chu SA, et al. Thalamo-cortical network hyperconnectivity in preclinical progranulin mutation carriers. NeuroImage: Clinical 2019;22(101751).

(214) Lüningschrör P, Werner G, Stroobants S, Kakuta S, Dombert B, Sinske D, et al. The FTLD risk factor TMEM106B regulates the transport of lysosomes at the axon initial segment of motoneurons. Cell Reports 2020;30(10):3506-3519. e6.

(215) Zhou X, Brooks M, Jiang P, Koga S, Zuberi AR, Baker MC, et al. Loss of Tmem106b exacerbates FTLD pathologies and causes motor deficits in progranulin‐deficient mice. EMBO Rep 2020;21(10):e50197.

(216) Arrant AE, Nicholson AM, Zhou X, Rademakers R, Roberson ED. Partial Tmem106b reduction does not correct abnormalities due to progranulin haploinsufficiency. Molecular neurodegeneration 2018;13(32).

(217) Philips T, De Muynck L, Nguyen Thi Thu H, Weynants B, Vanacker P, Dhondt J, et al. Microglial upregulation of progranulin as a marker of motor neuron degeneration. Journal of Neuropathology & Experimental Neurology 2010;69(12):1191-1200.

(218) Arrant AE, Onyilo VC, Unger DE, Roberson ED. Progranulin gene therapy improves lysosomal dysfunction and microglial pathology associated with frontotemporal dementia and neuronal ceroid lipofuscinosis. Journal of Neuroscience 2018;38(9):2341-2358.

(219) Alexander C, Pisner D, Jacova C. Predementia brain changes in progranulin mutation: a systematic review of neuroimaging evidence. Dement Geriatr Cogn Disord 2019;47(1-2):1-18.

(220) She A, Kurtser I, Reis SA, Hennig K, Lai J, Lang A, et al. Selectivity and kinetic requirements of HDAC inhibitors as progranulin enhancers for treating frontotemporal dementia. Cell Chemical Biology 2017;24(7):892-906.

(221) Hinderer C, Miller R, Dyer C, Johansson J, Bell P, Buza E, et al. Adeno‐associated virus serotype 1‐based gene therapy for FTD caused by GRN mutations. Annals of Clinical and Translational Neurology 2020;7(10):1843-1853.

(222) Alberici A, Archetti S, Pilotto A, Premi E, Cosseddu M, Bianchetti A, et al. Results from a pilot study on amiodarone administration in monogenic frontotemporal dementia with granulin mutation. Neurological Sciences 2014;35(8):1215-1219.

(223) Sharon JS, Miller ZA, Min S, Zhou Y, Brown J, Mitic LL, et al. An 8-week, open-label, dose-finding study of nimodipine for the treatment of progranulin insufficiency from GRN gene mutations. Alzheimer's & Dementia: Translational Research & Clinical Interventions 2017;3(4):507-512.

251

(224) A Phase 3 Study to Evaluate Efficacy and Safety of AL001 in Frontotemporal Dementia (INFRONT-3). Available at: https://ClinicalTrials.gov/show/NCT04374136.

(225) Phase 1/2 Clinical Trial of PR006 in Patients with Frontotemporal Dementia with Progranulin Mutations (FTD-GRN) (PROCLAIM). Available at: https://ClinicalTrials.gov/show/NCT04408625.

(226) A Study of PBFT02 in Patients with Frontotemporal Dementia and Progranulin Mutations (FTD-GRN) (upliFT-D). Available at: https://ClinicalTrials.gov/show/NCT04747431.

(227) Lebert F, Stekke W, Hasenbroekx C, Pasquier F. Frontotemporal dementia: a randomised, controlled trial with trazodone. Dement Geriatr Cogn Disord 2004;17(4):355-359.

(228) Brenner S, Stretton A, Kaplan S. Genetic code: the ‘nonsense’ triplets for chain termination and their suppression. Nature 1965;206(4988):994-998.

(229) Brenner S, Barnett L, Katz ER, Crick F. UGA: a third nonsense triplet in the genetic code. Nature 1967;213(5075):449-450.

(230) Losson R, Lacroute F. Interference of nonsense mutations with eukaryotic messenger RNA stability. Proceedings of the National Academy of Sciences 1979;76(10):5134-5137.

(231) Wilhelm JM, Jessop JJ, Pettitt SE. Aminoglycoside antibiotics and eukaryotic protein synthesis: stimulation of errors in the translation of natural messengers in extracts of cultured human cells. Biochemistry 1978;17(7):1149-1153.

(232) Palmer E, Wilhelm JM, Sherman F. Phenotypic suppression of nonsense mutants in yeast by aminoglycoside antibiotics. Nature 1979;277(5692):148-150.

(233) Nicholson P, Yepiskoposyan H, Metze S, Orozco RZ, Kleinschmidt N, Mühlemann O. Nonsense-mediated mRNA decay in human cells: mechanistic insights, functions beyond quality control and the double-life of NMD factors. Cellular and molecular life sciences 2010;67(5):677-700.

(234) Manuvakhova M, Keeling K, Bedwell DM. Aminoglycoside antibiotics mediate context-dependent suppression of termination codons in a mammalian translation system. RNA 2000;6(7):1044-1055.

(235) Cassan M, Rousset J. UAG readthrough in mammalian cells: effect of upstream and downstream stop codon contexts reveal different signals. BMC molecular biology 2001;2(3).

252

(236) McCaughan KK, Brown CM, Dalphin ME, Berry MJ, Tate WP. Translational termination efficiency in mammals is influenced by the base following the stop codon. Proceedings of the National Academy of Sciences 1995;92(12):5431-5435.

(237) Tate WP, Poole ES, Horsfield JA, Mannering SA, Brown CM, Moffat JG, et al. Translational termination efficiency in both bacteria and mammals is regulated by the base following the stop codon. Biochemistry and cell biology 1995;73(11-12):1095-1103.

(238) Pavlov MY, Freistroffer DV, Dincbas V, MacDougall J, Buckingham RH, Ehrenberg M. A direct estimation of the context effect on the efficiency of termination. J Mol Biol 1998;284(3):579-590.

(239) Namy O, Hatin I, Rousset J. Impact of the six nucleotides downstream of the stop codon on translation termination. EMBO Rep 2001;2(9):787-793.

(240) Bonetti B, Fu L, Moon J, Bedwell DM. The efficiency of translation termination is determined by a synergistic interplay between upstream and downstream sequences in Saccharomyces cerevisiae. J Mol Biol 1995;251(3):334-345.

(241) Hoshino S, Imai M, Kobayashi T, Uchida N, Katada T. The eukaryotic polypeptide chain releasing factor (eRF3/GSPT) carrying the translation termination signal to the 3'-Poly(A) tail of mRNA. Direct association of erf3/GSPT with polyadenylate-binding protein. J Biol Chem 1999;274(24):16677-16680.

(242) Ivanov PV, Gehring NH, Kunz JB, Hentze MW, Kulozik AE. Interactions between UPF1, eRFs, PABP and the exon junction complex suggest an integrated model for mammalian NMD pathways. EMBO J 2008;27(5):736-747.

(243) Barton-Davis ER, Cordier L, Shoturma DI, Leland SE, Sweeney HL. Aminoglycoside antibiotics restore dystrophin function to skeletal muscles of mdx mice. J Clin Invest 1999;104(4):375-381.

(244) Wagner KR, Hamed S, Hadley DW, Gropman AL, Burstein AH, Escolar DM, et al. Gentamicin treatment of Duchenne and Becker muscular dystrophy due to nonsense mutations. Annals of Neurology: Official Journal of the American Neurological Association and the Child Neurology Society 2001;49(6):706-711.

(245) Shulman E, Belakhov V, Wei G, Kendall A, Meyron-Holtz EG, Ben-Shachar D, et al. Designer aminoglycosides that selectively inhibit cytoplasmic rather than mitochondrial ribosomes show decreased ototoxicity a strategy for the treatment of genetic diseases. J Biol Chem 2014;289(4):2318-2330.

(246) Leubitz A, Frydman‐Marom A, Sharpe N, van Duzer J, Campbell KC, Vanhoutte

F. Safety, tolerability, and pharmacokinetics of single ascending doses of ELX‐02, a

253

potential treatment for genetic disorders caused by nonsense mutations, in healthy volunteers. Clinical pharmacology in drug development 2019;8(8):984-994.

(247) Brasell EJ, Chu LL, Akpa MM, Eshkar-Oren I, Alroy I, Corsini R, et al. The novel aminoglycoside, ELX-02, permits CTNSW138X translational read-through and restores lysosomal cystine efflux in cystinosis. PloS One 2019;14(12):e0223954.

(248) Taguchi A, Nishiguchi S, Shiozuka M, Nomoto T, Ina M, Nojima S, et al. Negamycin analogue with readthrough-promoting activity as a potential drug candidate for duchenne muscular dystrophy. ACS medicinal chemistry letters 2012;3(2):118-122.

(249) Hamada K, Omura N, Taguchi A, Baradaran-Heravi A, Kotake M, Arai M, et al. New negamycin-based potent readthrough derivative effective against TGA-type nonsense mutations. ACS medicinal chemistry letters 2019;10(10):1450-1456.

(250) Welch EM, Barton ER, Zhuo J, Tomizawa Y, Friesen WJ, Trifillis P, et al. PTC124 targets genetic disorders caused by nonsense mutations. Nature 2007;447(7140):87-91.

(251) McDonald CM, Campbell C, Torricelli RE, Finkel RS, Flanigan KM, Goemans N, et al. Ataluren in patients with nonsense mutation Duchenne muscular dystrophy (ACT DMD): a multicentre, randomised, double-blind, placebo-controlled, phase 3 trial. The Lancet 2017;390(10101):1489-1498.

(252) Roles of molecular chaperones in cytoplasmic protein folding. Seminars in cell & developmental biology: Elsevier; 2000.

(253) Keeling KM, Brooks DA, Hopwood JJ, Li P, Thompson JN, Bedwell DM. Gentamicin-mediated suppression of Hurler syndrome stop mutations restores a low level of α-L-iduronidase activity and reduces lysosomal glycosaminoglycan accumulation. Hum Mol Genet 2001;10(3):291-300.

(254) Keeling KM, Bedwell DM. Clinically relevant aminoglycosides can suppress disease-associated premature stop mutations in the IDUA and P53 cDNAs in a mammalian translation system. Journal of molecular medicine 2002;80(6):367-376.

(255) Chernikov VG, Terekhov SM, Krokhina TB, Shishkin SS, Smirnova TD, Kalashnikova EA, et al. Comparison of cytotoxicity of aminoglycoside antibiotics using a panel cellular biotest system. Bull Exp Biol Med 2003;135(1):103-105.

(256) Wangen JR, Green R. Stop codon context influences genome-wide stimulation of termination codon readthrough by aminoglycosides. Elife 2020;9:e52611.

254

(257) Wakap SN, Lambert DM, Olry A, Rodwell C, Gueydan C, Lanneau V, et al. Estimating cumulative point prevalence of rare diseases: analysis of the Orphanet database. European Journal of Human Genetics 2020;28(2):165-173.

(258) Mort M, Ivanov D, Cooper DN, Chuzhanova NA. A meta‐analysis of nonsense mutations causing human genetic disease. Hum Mutat 2008;29(8):1037-1047.

(259) Bushby K, Finkel R, Wong B, Barohn R, Campbell C, Comi GP, et al. Ataluren treatment of patients with nonsense mutation dystrophinopathy. Muscle Nerve 2014;50(4):477-487.

(260) Wilschanski M, Famini C, Blau H, Rivlin J, Augarten A, Avital A, et al. A pilot study of the effect of gentamicin on nasal potential difference measurements in cystic fibrosis patients carrying stop mutations. American journal of respiratory and critical care medicine 2000;161(3):860-865.

(261) Clancy JP, Bebok Z, Ruiz F, King C, Jones J, Walker L, et al. Evidence that systemic gentamicin suppresses premature stop mutations in patients with cystic fibrosis. American journal of respiratory and critical care medicine 2001;163(7):1683-1692.

(262) Wilschanski M, Yahav Y, Yaacov Y, Blau H, Bentur L, Rivlin J, et al. Gentamicin-induced correction of CFTR function in patients with cystic fibrosis and CFTR stop mutations. N Engl J Med 2003;349(15):1433-1441.

(263) Kerem E, Konstan MW, De Boeck K, Accurso FJ, Sermet-Gaudelus I, Wilschanski M, et al. Ataluren for the treatment of nonsense-mutation cystic fibrosis: a randomised, double-blind, placebo-controlled phase 3 trial. The Lancet Respiratory Medicine 2014;2(7):539-547.

(264) Haas M, Vlcek V, Balabanov P, Salmonson T, Bakchine S, Markey G, et al. European Medicines Agency review of ataluren for the treatment of ambulant patients aged 5 years and older with Duchenne muscular dystrophy resulting from a nonsense mutation in the dystrophin gene. Neuromuscular disorders 2015;25(1):5-13.

(265) Keeling KM, Xue X, Gunn G, Bedwell DM. Therapeutics based on stop codon readthrough. Annual review of genomics and human genetics 2014;15:371-394.

(266) Carlier M, Rollman B, Van Hoof F, Tulkens P. Mechanism of aminoglycoside-induced lysosomal phospholipidosis: in vitro and in vivo studies with gentamicin and amikacin. Biochem Pharmacol 1982;31(23):3861-3870.

(267) O’neil WG. Aminoglycoside induced ototoxicity. Toxicology 2008;249(2-3):91-96.

255

(268) Hobbie SN, Akshay S, Kalapala SK, Bruell CM, Shcherbakov D, Böttger EC. Genetic analysis of interactions with eukaryotic rRNA identify the mitoribosome as target in aminoglycoside ototoxicity. Proceedings of the National Academy of Sciences 2008;105(52):20888-20893.

(269) Davies J, Gorini L, Davis BD. Misreading of RNA codewords induced by aminoglycoside antibiotics. Mol Pharmacol 1965;1(1):93-106.

(270) Edelmann P, Gallant J. Mistranslation in E. coli. Cell 1977;10(1):131-137.

(271) Dietz HC. New therapeutic approaches to mendelian disorders. N Engl J Med 2010;363(9):852-863.

(272) Linde L, Kerem B. Introducing sense into nonsense in treatments of human genetic diseases. Trends in genetics 2008;24(11):552-563.

(273) Keeling KM, Bedwell DM. Recoding therapies for genetic diseases. Recoding: Expansion of Decoding Rules Enriches Gene Expression: Springer; 2010. p. 123-146.

(274) Recht MI, Douthwaite S, Puglisi JD. Basis for prokaryotic specificity of action of aminoglycoside antibiotics. EMBO J 1999;18(11):3133-3138.

(275) Lynch SR, Puglisi JD. Structural origins of aminoglycoside specificity for prokaryotic ribosomes. J Mol Biol 2001;306(5):1037-1058.

(276) Fan-Minogue H, Bedwell DM. Eukaryotic ribosomal RNA determinants of aminoglycoside resistance and their role in translational fidelity. RNA 2008;14(1):148-157.

(277) Salas-Marco J, Bedwell DM. Discrimination between defects in elongation fidelity and termination efficiency provides mechanistic insights into translational readthrough. J Mol Biol 2005;348(4):801-815.

(278) Korostelev A, Ermolenko DN, Noller HF. Structural dynamics of the ribosome. Curr Opin Chem Biol 2008;12(6):674-683.

(279) Shalev M, Baasov T. When proteins start to make sense: fine-tuning of aminoglycosides for PTC suppression therapy. MedChemComm 2014;5(8):1092-1105.

(280) Carter AP, Clemons WM, Brodersen DE, Morgan-Warren RJ, Wimberly BT, Ramakrishnan V. Functional insights from the structure of the 30S ribosomal subunit and its interactions with antibiotics. Nature 2000;407(6802):340-348.

256

(281) Ogle JM, Brodersen DE, Clemons WM, Tarry MJ, Carter AP, Ramakrishnan V. Recognition of cognate transfer RNA by the 30S ribosomal subunit. Science 2001;292(5518):897-902.

(282) Frolova L, Le Goff X, Rasmussen HH, Cheperegin S, Drugeon G, Kress M, et al. A highly conserved eukaryotic protein family possessing properties of polypeptide chain release factor. Nature 1994;372(6507):701-703.

(283) Scolnick EM, Caskey CT. Peptide chain termination, V. The role of release factors in mRNA terminator codon recognition. Proceedings of the National Academy of Sciences 1969;64(4):1235-1241.

(284) Laurberg M, Asahara H, Korostelev A, Zhu J, Trakhanov S, Noller HF. Structural basis for translation termination on the 70S ribosome. Nature 2008;454(7206):852-857.

(285) Brown A, Shao S, Murray J, Hegde RS, Ramakrishnan V. Structural basis for stop codon recognition in eukaryotes. Nature 2015;524(7566):493-496.

(286) Kondo J, Urzhumtsev A, Westhof E. Two conformational states in the crystal structure of the Homo sapiens cytoplasmic ribosomal decoding A site. Nucleic Acids Res 2006;34(2):676-685.

(287) Kondo J, Westhof E. The bacterial and mitochondrial ribosomal A-site molecular switches possess different conformational substates. Nucleic Acids Res 2008;36(8):2654-2666.

(288) Youngman EM, He SL, Nikstad LJ, Green R. Stop codon recognition by release factors induces structural rearrangement of the ribosomal decoding center that is productive for peptide release. Mol Cell 2007;28(4):533-543.

(289) Shalev M, Kondo J, Kopelyanskiy D, Jaffe CL, Adir N, Baasov T. Identification of the molecular attributes required for aminoglycoside activity against Leishmania. Proceedings of the National Academy of Sciences 2013;110(33):13333-13338.

(290) Demeshkina N, Jenner L, Westhof E, Yusupov M, Yusupova G. A new understanding of the decoding principle on the ribosome. Nature 2012;484(7393):256-259.

(291) Mingeot-Leclercq M, Tulkens PM. Aminoglycosides: nephrotoxicity. Antimicrob Agents Chemother 1999;43(5):1003-1012.

(292) Selimoglu E. Aminoglycoside-induced ototoxicity. Curr Pharm Des 2007;13(1):119-126.

257

(293) Lopez-Novoa JM, Quiros Y, Vicente L, Morales AI, Lopez-Hernandez FJ. New insights into the mechanism of aminoglycoside nephrotoxicity: an integrative point of view. Kidney Int 2011;79(1):33-45.

(294) El Mouedden M, Laurent G, Mingeot-Leclercq M, Taper HS, Cumps J, Tulkens PM. Apoptosis in renal proximal tubules of rats treated with low doses of aminoglycosides. Antimicrob Agents Chemother 2000;44(3):665-675.

(295) Huth ME, Ricci AJ, Cheng AG. Mechanisms of aminoglycoside ototoxicity and targets of hair cell protection. International journal of otolaryngology 2011;2011(937861).

(296) N Abi-Hachem R, Zine A, R Van De Water, Thomas. The injured cochlea as a target for inflammatory processes, initiation of cell death pathways and application of related otoprotective strategies. Recent patents on CNS drug discovery 2010;5(2):147-163.

(297) llsø Christensen E, Birn H, Verroust P, Moestrup SK. Membrane receptors for endocytosis in the renal proximal tubule. International review of cytology: Elsevier; 1998. p. 237-284.

(298) Mizuta K, Saito A, Watanabe T, Nagura M, Arakawa M, Shimizu F, et al. Ultrastructural localization of megalin in the rat cochlear duct. Hear Res 1999;129(1-2):83-91.

(299) Konig O, Ruttiger L, Muller M, Zimmermann U, Erdmann B, Kalbacher H, et al. Estrogen and the inner ear: megalin knockout mice suffer progressive hearing loss. The FASEB Journal 2008;22(2):410-417.

(300) Lesniak W, Mc Laren J, Harris WR, Pecoraro VL, Schacht J. An isocratic separation of underivatized gentamicin components, 1H NMR assignment and protonation pattern. Carbohydr Res 2003;338(24):2853-2862.

(301) Giuliano RA, Paulus GJ, Verpooten GA, Pattyn VM, Pollet DE, Nouwen EJ, et al. Recovery of cortical phospholipidosis and necrosis after acute gentamicin loading in rats. Kidney Int 1984;26(6):838-847.

(302) Regec AL, Trump BF, Trifilis AL. Effect of gentamicin on the lysosomal system of cultured human proximal tubular cells: Endocytotic activity, lysosomal pH and membrane fragility. Biochem Pharmacol 1989;38(15):2527-2534.

(303) Schacht J. Antioxidant therapy attenuates aminoglycoside-induced hearing loss. Ann N Y Acad Sci 1999;884:125-130.

258

(304) Wu W, Sha S, Schacht J. Recent advances in understanding aminoglycoside ototoxicity and its prevention. Audiology and Neurotology 2002;7(3):171-174.

(305) Priuska EM, Clark-Baldwin K, Pecoraro VL, Schacht J. NMR studies of iron-gentamicin complexes and the implications for aminoglycoside toxicity. Inorg Chim Acta 1998;273(1-2):85-91.

(306) Simmons CF, Bogusky RT, Humes HD. Inhibitory effects of gentamicin on renal mitochondrial oxidative phosphorylation. J Pharmacol Exp Ther 1980;214(3):709-15.

(307) Walker PD, Shah SV. Gentamicin enhanced production of hydrogen peroxide by renal cortical mitochondria. American Journal of Physiology-Cell Physiology 1987;253(4):C495-C499.

(308) Estivill X, Govea N, Barceló A, Perelló E, Badenas C, Romero E, et al. Familial progressive sensorineural deafness is mainly due to the mtDNA A1555G mutation and is enhanced by treatment with aminoglycosides. The American Journal of Human Genetics 1998;62(1):27-35.

(309) Zhao H, Li R, Wang Q, Yan Q, Deng J, Han D, et al. Maternally inherited aminoglycoside-induced and nonsyndromic deafness is associated with the novel C1494T mutation in the mitochondrial 12S rRNA gene in a large Chinese family. The American Journal of Human Genetics 2004;74(1):139-152.

(310) Howard M, Frizzell RA, Bedwell DM. Aminoglycoside antibiotics restore CFTR function by overcoming premature stop mutations. Nat Med 1996;2(4):467-469.

(311) Bedwell DM, Kaenjak A, Benos DJ, Bebok Z, Bubien JK, Hong J, et al. Suppression of a CFTR premature stop mutation in a bronchial epithelial cell line. Nat Med 1997;3(11):1280-1284.

(312) Nudelman I, Rebibo-Sabbah A, Shallom-Shezifi D, Hainrichson M, Stahl I, Ben-Yosef T, et al. Redesign of aminoglycosides for treatment of human genetic diseases caused by premature stop mutations. Bioorg Med Chem Lett 2006;16(24):6310-6315.

(313) Nudelman I, Rebibo-Sabbah A, Cherniavsky M, Belakhov V, Hainrichson M, Chen F, et al. Development of novel aminoglycoside (NB54) with reduced toxicity and enhanced suppression of disease-causing premature stop mutations. J Med Chem 2009;52(9):2836-2845.

(314) Nudelman I, Glikin D, Smolkin B, Hainrichson M, Belakhov V, Baasov T. Repairing faulty genes by aminoglycosides: development of new derivatives of geneticin (G418) with enhanced suppression of diseases-causing nonsense mutations. Bioorg Med Chem 2010;18(11):3735-3746.

259

(315) Kandasamy J, Atia-Glikin D, Belakhov V, Baasov T. Repairing faulty genes by aminoglycosides: Identification of new pharmacophore with enhanced suppression of disease-causing nonsense mutations. MedChemComm 2011;2(3):165-171.

(316) Kandasamy J, Atia-Glikin D, Shulman E, Shapira K, Shavit M, Belakhov V, et al. Increased selectivity toward cytoplasmic versus mitochondrial ribosome confers improved efficiency of synthetic aminoglycosides in fixing damaged genes: a strategy for treatment of genetic diseases caused by nonsense mutations. J Med Chem 2012;55(23):10630-10643.

(317) Sabbavarapu NM, Shavit M, Degani Y, Smolkin B, Belakhov V, Baasov T. Design of novel aminoglycoside derivatives with enhanced suppression of diseases-causing nonsense mutations. ACS medicinal chemistry letters 2016;7(4):418-423.

(318) Bidou L, Bugaud O, Belakhov V, Baasov T, Namy O. Characterization of new-generation aminoglycoside promoting premature termination codon readthrough in cancer cells. RNA biology 2017;14(3):378-388.

(319) A Phase 2 Study to Evaluate the Safety, Tolerability, PK and PD of ELX-02 in Cystic Fibrosis Patients With G542X Allele. Available at: https://ClinicalTrials.gov/show/NCT04135495.

(320) A Phase 2 Study of ELX-02 in Patients With Nephropathic Cystinosis. Available at: https://ClinicalTrials.gov/show/NCT04069260.

(321) A Phase 2 Study to Evaluate the Safety, Tolerability, PK and PD in Cystic Fibrosis Patients With at Least 1 G542X Allele. Available at: https://ClinicalTrials.gov/show/NCT04126473.

(322) Sabbavarapu NM, Pieńko T, Zalman B, Trylska J, Baasov T. Exploring eukaryotic versus prokaryotic ribosomal RNA recognition with aminoglycoside derivatives. MedChemComm 2018;9(3):503-508.

(323) Auld DS, Thorne N, Maguire WF, Inglese J. Mechanism of PTC124 activity in cell-based luciferase assays of nonsense codon suppression. Proceedings of the National Academy of Sciences 2009;106(9):3585-3590.

(324) Auld DS, Lovell S, Thorne N, Lea WA, Maloney DJ, Shen M, et al. Molecular basis for the high-affinity binding and stabilization of firefly luciferase by PTC124. Proceedings of the National Academy of Sciences 2010;107(11):4878-4883.

(325) Dranchak PK, Di Pietro E, Snowden A, Oesch N, Braverman NE, Steinberg SJ, et al. Nonsense suppressor therapies rescue peroxisome lipid metabolism and assembly in cells from patients with specific PEX gene mutations. J Cell Biochem 2011;112(5):1250-1258.

260

(326) Brumm H, Mühlhaus J, Bolze F, Scherag S, Hinney A, Hebebrand J, et al. Rescue of melanocortin 4 receptor (MC4R) nonsense mutations by aminoglycoside‐mediated read‐through. Obesity 2012;20(5):1074-1081.

(327) McElroy SP, Nomura T, Torrie LS, Warbrick E, Gartner U, Wood G, et al. A lack of premature termination codon read-through efficacy of PTC124 (Ataluren) in a diverse array of reporter assays. PLoS Biol 2013;11(6):e1001593.

(328) Du M, Liu X, Welch EM, Hirawat S, Peltz SW, Bedwell DM. PTC124 is an orally bioavailable compound that promotes suppression of the human CFTR-G542X nonsense allele in a CF mouse model. Proceedings of the National Academy of Sciences 2008;105(6):2064-2069.

(329) Goldmann T, Overlack N, Möller F, Belakhov V, van Wyk M, Baasov T, et al. A comparative evaluation of NB30, NB54 and PTC124 in translational read‐through efficacy for treatment of an USH1C nonsense mutation. EMBO molecular medicine 2012;4(11):1186-1199.

(330) Sarkar C, Zhang Z, Mukherjee AB. Stop codon read-through with PTC124 induces palmitoyl-protein thioesterase-1 activity, reduces thioester load and suppresses apoptosis in cultured cells from INCL patients. Mol Genet Metab 2011;104(3):338-345.

(331) Roy B, Friesen WJ, Tomizawa Y, Leszyk JD, Zhuo J, Johnson B, et al. Ataluren stimulates ribosomal selection of near-cognate tRNAs to promote nonsense suppression. Proceedings of the National Academy of Sciences 2016;113(44):12508-12513.

(332) Salian S, Matt T, Akbergenov R, Harish S, Meyer M, Duscha S, et al. Structure-activity relationships among the kanamycin aminoglycosides: role of ring I hydroxyl and amino groups. Antimicrob Agents Chemother 2012;56(12):6104-6108.

(333) Konstan MW, VanDevanter DR, Rowe SM, Wilschanski M, Kerem E, Sermet-Gaudelus I, et al. Efficacy and safety of ataluren in patients with nonsense-mutation cystic fibrosis not receiving chronic inhaled aminoglycosides: the international, randomized, double-blind, placebo-controlled Ataluren Confirmatory Trial in Cystic Fibrosis (ACT CF). Journal of Cystic Fibrosis 2020;19(4):595-601.

(334) Pibiri I, Lentini L, Tutone M, Melfi R, Pace A, Di Leonardo A. Exploring the readthrough of nonsense mutations by non-acidic Ataluren analogues selected by ligand-based virtual screening. Eur J Med Chem 2016;122:429-435.

(335) Pibiri I, Lentini L, Melfi R, Tutone M, Baldassano S, Galluzzo PR, et al. Rescuing the CFTR protein function: Introducing 1, 3, 4-oxadiazoles as translational readthrough inducing drugs. Eur J Med Chem 2018;159:126-142.

261

(336) Campofelice A, Lentini L, Di Leonardo A, Melfi R, Tutone M, Pace A, et al. Strategies against nonsense: oxadiazoles as translational readthrough-inducing drugs (TRIDs). International journal of molecular sciences 2019;20(13):3329.

(337) Baradaran-Heravi A, Balgi AD, Zimmerman C, Choi K, Shidmoossavee FS, Tan JS, et al. Novel small molecules potentiate premature termination codon readthrough by aminoglycosides. Nucleic Acids Res 2016;44(14):6583-6598.

(338) Ferguson MW, Gerak CA, Chow CC, Rastelli EJ, Elmore KE, Stahl F, et al. The antimalarial drug mefloquine enhances TP53 premature termination codon readthrough by aminoglycoside G418. PloS One 2019;14(5):e0216423.

(339) Seoighe C, Gehring C. Heritability in the efficiency of nonsense-mediated mRNA decay in humans. PLoS One 2010;5(7):e11657.

(340) Maquat LE, Tarn W, Isken O. The pioneer round of translation: features and functions. Cell 2010;142(3):368-374.

(341) Kurosaki T, Popp MW, Maquat LE. Quality and quantity control of gene expression by nonsense-mediated mRNA decay. Nature reviews Molecular cell biology 2019;20(7):406-420.

(342) Usuki F, Yamashita A, Higuchi I, Ohnishi T, Shiraishi T, Osame M, et al. Inhibition of nonsense-mediated mRNA decay rescues the phenotype in Ullrich's disease. Annals of Neurology: Official Journal of the American Neurological Association and the Child Neurology Society 2004;55(5):740-744.

(343) Martin L, Grigoryan A, Wang D, Wang J, Breda L, Rivella S, et al. Identification and characterization of small molecules that inhibit nonsense-mediated RNA decay and suppress nonsense p53 mutations. Cancer Res 2014;74(11):3104-3113.

(344) Durand S, Cougot N, Mahuteau-Betzer F, Nguyen C, Grierson DS, Bertrand E, et al. Inhibition of nonsense-mediated mRNA decay (NMD) by a new chemical molecule reveals the dynamic of NMD factors in P-bodies. J Cell Biol 2007;178(7):1145-1160.

(345) Keeling KM, Wang D, Dai Y, Murugesan S, Chenna B, Clark J, et al. Attenuation of nonsense-mediated mRNA decay enhances in vivo nonsense suppression. PloS One 2013;8(4):e60478.

(346) Li S, Wilkinson MF. Nonsense surveillance in lymphocytes? Immunity 1998;8(2):135-141.

(347) Lew JE, Enomoto S, Berman J. Telomere length regulation and telomeric chromatin require the nonsense-mediated mRNA decay pathway. Mol Cell Biol 1998;18(10):6121-6130.

262

(348) Kuang L, Hashimoto K, Huang EJ, Gentry MS, Zhu H. Frontotemporal dementia nonsense mutation of progranulin rescued by aminoglycosides. Hum Mol Genet 2020;29(4):624-634.

(349) Zolotukhin S, Byrne BJ, Mason E, Zolotukhin I, Potter M, Chesnut K, et al. Recombinant adeno-associated virus purification using novel methods improves infectious titer and yield. Gene Ther 1999;6(6):973-985.

(350) Frew J, Wu X, Hsiung GY, Feldman HH, Mackenzie IR, Nygaard HB. Generation of an induced pluripotent stem cell line (UBCi001-A) from a presymptomatic individual carrying the R418X progranulin gene mutation. Stem Cell Research 2019;41:101582.

(351) Rose SE, Frankowski H, Knupp A, Berry BJ, Martinez R, Dinh SQ, et al. Leptomeninges-derived induced pluripotent stem cells and directly converted neurons from autopsy cases with varying neuropathologic backgrounds. Journal of Neuropathology & Experimental Neurology 2018;77(5):353-360.

(352) Bardy C, Van Den Hurk M, Eames T, Marchand C, Hernandez RV, Kellogg M, et al. Neuronal medium that supports basic synaptic functions and activity of human neurons in vitro. Proceedings of the National Academy of Sciences 2015;112(20):2725.

(353) Jacobi AM, Rettig GR, Turk R, Collingwood MA, Zeiner SA, Quadros RM, et al. Simplified CRISPR tools for efficient genome editing and streamlined protocols for their delivery into mammalian cells and mouse zygotes. Methods 2017;121-122:16-28.

(354) Young K, Morrison H. Quantifying microglia morphology from photomicrographs of immunohistochemistry prepared tissue using ImageJ. JoVE (Journal of Visualized Experiments) 2018(136):e57648.

(355) Kim J, Grunke SD, Levites Y, Golde TE, Jankowsky JL. Intracerebroventricular viral injection of the neonatal mouse brain for persistent and widespread neuronal transduction. JoVE (Journal of Visualized Experiments) 2014(91):e51863.

(356) DeVos SL, Miller TM. Direct intraventricular delivery of drugs to the rodent central nervous system. JoVE (Journal of Visualized Experiments) 2013(75):e50326.

(357) Louis N, Evelegh C, Graham FL. Cloning and sequencing of the cellular–viral junctions from the human adenovirus type 5 transformed 293 cell line. Virology 1997;233(2):423-429.

(358) Lin Y, Boone M, Meuris L, Lemmens I, Van Roy N, Soete A, et al. Genome dynamics of the human embryonic kidney 293 lineage in response to cell biology manipulations. Nature communications 2014;5(1):1-12.

263

(359) Passini MA, Wolfe JH. Widespread gene delivery and structure-specific patterns of expression in the brain after intraventricular injections of neonatal mice with an adeno-associated virus vector. J Virol 2001;75(24):12382-12392.

(360) Meng F, Han Y, Srisai D, Belakhov V, Farias M, Xu Y, et al. New inducible genetic method reveals critical roles of GABA in the control of feeding and metabolism. Proceedings of the National Academy of Sciences 2016;113(13):3645-3650.

(361) Merritt JK, Collins BE, Erickson KR, Dong H, Neul JL. Pharmacological readthrough of R294X Mecp2 in a novel mouse model of Rett Syndrome. Hum Mol Genet 2020;29(15):2461–2470.

(362) Vorbrodt AW, Lossinsky AS, Wisniewski HM. Localization of alkaline phosphatase activity in endothelia of developing and mature mouse blood-brain barrier. Dev Neurosci 1986;8(1):1-13.

(363) Hu K. All roads lead to induced pluripotent stem cells: the technologies of iPSC generation. Stem cells and development 2014;23(12):1285-1300.

(364) Borgohain MP, Haridhasapavalan KK, Dey C, Adhikari P, Thummer RP. An insight into DNA-free reprogramming approaches to generate integration-free induced pluripotent stem cells for prospective biomedical applications. Stem Cell Reviews and Reports 2019;15(2):286-313.

(365) Götz M, Huttner WB. The cell biology of neurogenesis. Nature reviews Molecular cell biology 2005;6(10):777-788.

(366) Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A programmable dual-RNA–guided DNA endonuclease in adaptive bacterial immunity. Science 2012;337(6096):816-821.

(367) Roy B, Leszyk JD, Mangus DA, Jacobson A. Nonsense suppression by near-cognate tRNAs employs alternative base pairing at codon positions 1 and 3. Proceedings of the National Academy of Sciences 2015;112(10):3038-3043.

(368) Theocharatos S, Wilkinson DJ, Darling S, Wilm B, Kenny SE, Edgar D. Regulation of progenitor cell proliferation and neuronal differentiation in enteric nervous system neurospheres. PloS One 2013;8(1):e54809.

(369) Floquet C, Deforges J, Rousset J, Bidou L. Rescue of non-sense mutated p53 tumor suppressor gene by aminoglycosides. Nucleic Acids Res 2010;39(8):3350-3362.

(370) Keeling KM, Lanier J, Du M, Salas-Marco J, Gao L, Kaenjak-Angeletti A, et al. Leaky termination at premature stop codons antagonizes nonsense-mediated mRNA decay in S. cerevisiae. RNA 2004;10(4):691-703.

264

(371) Wils H, Kleinberger G, Pereson S, Janssens J, Capell A, Van Dam D, et al. Cellular ageing, increased mortality and FTLD‐TDP‐associated neuropathology in progranulin knockout mice. J Pathol 2012;228(1):67-76.

(372) Tsuji H, Nonaka T, Yamashita M, Masuda-Suzukake M, Kametani F, Akiyama H, et al. Epitope mapping of antibodies against TDP-43 and detection of protease-resistant fragments of pathological TDP-43 in amyotrophic lateral sclerosis and frontotemporal lobar degeneration. Biochem Biophys Res Commun 2012;417(1):116-121.

(373) Fujita K, Chen X, Homma H, Tagawa K, Amano M, Saito A, et al. Targeting Tyro3 ameliorates a model of PGRN-mutant FTLD-TDP via tau-mediated synaptic pathology. Nature communications 2018;9(1):433.

(374) Damani MR, Zhao L, Fontainhas AM, Amaral J, Fariss RN, Wong WT. Age‐related alterations in the dynamic behavior of microglia. Aging cell 2011;10(2):263-276.

(375) Guerin K, Gregory-Evans CY, Hodges MD, Moosajee M, Mackay DS, Gregory-Evans K, et al. Systemic aminoglycoside treatment in rodent models of retinitis pigmentosa. Exp Eye Res 2008;87(3):197-207.

(376) Beel S, Herdewyn S, Fazal R, De Decker M, Moisse M, Robberecht W, et al. Progranulin reduces insoluble TDP-43 levels, slows down axonal degeneration and prolongs survival in mutant TDP-43 mice. Molecular neurodegeneration 2018;13(1):55.

(377) Guttikonda SR, Sikkema L, Tchieu J, Saurat N, Walsh RM, Harschnitz O, et al. Fully defined human pluripotent stem cell-derived microglia and tri-culture system model C3 production in Alzheimer’s disease. Nat Neurosci 2021:1-12.

(378) Chiba S, Matsuwaki T, Yamanouchi K, Nishihara M. Alteration in anxiety with relation to the volume of the locus ceruleus in progranulin-deficient mice. Journal of reproduction and development 2009;55(5):518-522.

(379) Suzuki M, Lee H, Kayasuga Y, Chiba S, Nedachi T, Matsuwaki T, et al. Roles of progranulin in sexual differentiation of the developing brain and adult neurogenesis. Journal of Reproduction and Development 2009;55(4):351-355.

(380) Guillamón A, de Blas MR, Segovia S. Effects of sex steroids on the development of the locus coeruleus in the rat. Dev Brain Res 1988;40(2):306-310.

(381) McCall JG, Al-Hasani R, Siuda ER, Hong DY, Norris AJ, Ford CP, et al. CRH engagement of the locus coeruleus noradrenergic system mediates stress-induced anxiety. Neuron 2015;87(3):605-620.

(382) Monami G, Emiliozzi V, Bitto A, Lovat F, Xu S, Goldoni S, et al. Proepithelin regulates prostate cancer cell biology by promoting cell growth, migration, and

265

anchorage-independent growth. The American journal of pathology 2009;174(3):1037-1047.

(383) Bandey I, Chiou SH, Huang AP, Tsai JC, Tu PH. Progranulin promotes Temozolomide resistance of glioblastoma by orchestrating DNA repair and tumor stemness. Oncogene 2015;34(14):1853-1864.

(384) Birey F, Andersen J, Makinson CD, Islam S, Wei W, Huber N, et al. Assembly of functionally integrated human forebrain spheroids. Nature 2017;545(7652):54-59.