View and Expressed As the Number of Mhspb1 Positive Cells Over Total Cells

Interrogating the Functional Consequences of Peripheral Neuropathy

Associated Mutations in Heat Shock Protein B1

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of

Philosophy in the Graduate School of The Ohio State University

Patrick Lawrence Heilman

Ohio State University Biochemistry Graduate Program

The Ohio State University

2017

Dissertation Committee

Dr. Stephen Kolb, Advisor

Dr. Juan Alfonzo

Dr. Sharon Amacher

Dr. Mark Parthun

Copyrighted by

Patrick Lawrence Heilman

2017

ABSTRACT

Heat shock protein beta-1 (HSPB1), is a ubiquitously expressed, multifunctional protein chaperone that plays an important role in protein quality control, inflammation, apoptosis, cell growth and cytoskeleton maintenance.

Mutations in HSPB1 result in the development of a late-onset, distal hereditary motor neuropathy type II (dHMN) and axonal Charcot-Marie-Tooth disease with sensory involvement (CMT2F). The functional consequences of HSPB1 mutations associated with hereditary neuropathy are unknown. HSPB1 also displays neuroprotective properties in many neuronal disease models, including the motor neuron disease amyotrophic lateral sclerosis (ALS). HSPB1 is upregulated in SOD1-ALS animal models during disease progression, predominately in glial cells. Glial cells are known to contribute to motor neuron loss in ALS through a non-cell autonomous mechanism. Here, I studied the non- cell autonomous role of wild-type and mutant HSPB1 in an astrocyte-motor neuron co-culture model system of ALS. Astrocyte-specific overexpression of wild-type HSPB1 was sufficient to attenuate SOD1(G93A) astrocyte-mediated toxicity in motor neurons, whereas, overexpression of mutHSPB1 failed to ameliorate motor neuron toxicity. Expression of a phosphomimetic HSPB1 mutant in SOD1(G93A) astrocytes also reduced toxicity to motor neurons,

i suggesting that phosphorylation may contribute to HSPB1 mediated- neuroprotection.

To elucidate a mechanism of HSPB1-mediated neuroprotection, I examined the role of wild-type and mutant HSPB1 in regulating inflammatory gene expression through activation of NF-kB signaling and mRNA decay apthways. Wild-type, but not mutant HSPB1 was able to reduce the activation of

NF-kB signaling, a major inflammatory signaling cascade, through an unidentified mechanism. Additionally, I found no evidence of HSPB1’s involvement in regulating AU-rich element mRNA decay. Together, these data suggest that

HSPB1 neuroprotection may be mediated in part by its regulation of inflammatory signaling.

DEDICATION

To my parents, Richard and Patricia Heilman

iii

ACKNOWLEDGMENTS

There are numerous people to whom I owe a tremendous amount of thanks and gratitude to. First and foremost, to my Ph.D. advisor, Dr. Stephen

Kolb, for his guidance, mentorship and patience. I am forever grateful for everything he has taught me, from how to isolate the hippocampus from a mouse to properly writing scientific literature. He has helped me to mature into the scientist I am today, and his mentorship will serve as an important foundation for the rest of my scientific career.

To my committee members, Dr. Juan Alfonzo, Dr. Sharon Amacher and

Dr. Mark Parthun. Their support and guidance has taught me to keep an open mind and always look at a problem from multiple angles. A particular thank you goes to Dr. Dan Battle, who spent countless hours helping me to develop and analyze data from my co-immunoprecipitation experiments.

This work would not have been possible without the help and support of all the alumni and current members of the Kolb Lab Group. To Samantha Renusch and Christopher Wier, thank you for being amazing friends and colleagues, who were always willing to help with an experiment, to lend an ear, or share a beer. I

iv hope that we can maintain our friendship with one another and help one another in the future.

A special thanks must also go to our collaborators in Dr. Brian Kaspar’s lab, Dr’s. Kathrin Meyer, SungWon Song, and Carlos Miranda, who were instrumental in teaching me about viral production and how to culture primary neuronal cell lines. Without them, this work would never have been possible.

A heart-felt thanks goes out to my family. To my girlfriend, Lauren Stober, who has kept me sane over the past 3 years, and been there to pick me up every time I stumbled. To my brothers, Paul and Raymond, for being supportive of my life choices and putting up with all my scientific rambling. To my parents, Richard and Patricia who have loved and supported me every step of the way, from deciding to major in Biochemistry after high school to the decision to enter a

Ph.D. program. No matter how hard things got, they were always there for me, and have taught me that hard work always pays off in the end. I cannot thank you enough for all the effort and sacrifice.

VITA

2011 ...... B.S. Biochemistry, Duquesne University of the Holy Spirit

2011-Present ...... Graduate Research/Teaching

Associate, The Ohio State Biochemistry Program, The Ohio State

University

PUBLICATIONS

Heilman, P.L., Song, S.W., Miranda, C., Meyer, K., Wier, C.G., Knapp, A.R.,

Kaspar, B.K., Kolb, S.J. Hereditary neuropathy-associated HSPB1 mutations disrupt non-cell autonomous protection of motor neurons. Experimental

Neurology 2017; 297:101-109

FIELDS OF STUDY

Major Field: Biochemistry

vii

TABLE OF CONTENTS

ABSTRACT ...... i DEDICATION ...... iii ACKNOWLEDGMENTS ...... iv VITA...... vi TABLE OF CONTENTS ...... viii LIST OF TABLES ...... xii LIST OF FIGURES ...... xiii CHAPTER 1: INTRODUCTION ...... 1 1.1 HEAT SHOCK PROTEINS ...... 1 1.1.1 THE HEAT SHOCK PROTEIN FAMILY ...... 1 1.2 HEAT SHOCK PROTEIN BETA 1 (HSPB1) ...... 7 1.2.1 STRUCTURE AND GENE REGULATION ...... 7 1.2.2 HSPB1 PHOSPHORYLATION ...... 11 1.2.3 HSPB1 OLIGOMERIZATION ...... 11 1.2.4 DIFFERENTIAL REGULATION AND CELLULAR LOCALIZATION ... 13 1.3 CELLULAR FUNCTIONS OF HSPB1 ...... 14 1.3.1 MOLECULAR CHAPERONE ...... 15 1.3.2 REGULATION OF CELL SIGNALING/APOPTOSIS ...... 19 1.3.3 REGULATION OF THE CYTOSKELETON...... 23 1.4 HSPB1 AND THE NERVOUS SYSTEM ...... 25 1.4.1 EXPRESSION OF HSPB1 IN THE NERVOUS SYSTEM ...... 25 1.4.2 HSPB1 AND NEURONAL INJURY ...... 26 1.4.3 HSPB1 AND ALS ...... 27 1.4.4 HSPB1 AND PARKINSON’S DISEASE ...... 28 1.4.5 HSPB1 and ALZHEIMER’S DISEASE ...... 29 1.5 MUTATIONS IN HSPB1 RESULT IN MOTOR NEURON DISEASE ...... 30 1.5.1 CLINICAL DIAGNOSIS OF HSPB1-ASSOCIATED CMT2/dHMNII .... 30

viii

1.5.2 POTENTIAL MECHANISMS OF HSPB1-LINKED CMT2 AND dHMNII ...... 34 CHAPTER 2: NON-CELL AUTONOMOUS NEUROPROTECTION OF HSPB1 IS DISRUPTED BY MUTATIONS THAT RESULT IN MOTOR NEUROPATHY ..... 39 2.1 INTRODUCTION ...... 39 2.2 EXPERIMENTAL METHODS ...... 42 2.2.1 MICE ...... 42 2.2.2 HUMAN TISSUE SAMPLES ...... 42 2.2.3 ES MOTOR NEURON DIFFERENTIATION ...... 43 2.2.4 MOUSE NPC ISOLATION AND ASTROCYTE DIFFERENTIATION .. 43 2.2.5 CONVERSION OF SKIN FIBROBLASTS TO iNPCs AND DIFFERENTIATION INTO ASTROCYTES ...... 44 2.2.6 LENTIVIRAL PRODUCTION ...... 45 2.2.7 ASTROCYTE-MOTOR NEURON CO-CULTURE ASSAY ...... 47 2.2.8 ASTROCYTE VIABILITY ASSAY ...... 48 2.2.9 WESTERN BLOTTING ...... 48 2.2.10 IMMUNOFLUORESCENCE ...... 49 2.2.11 HSPB1 ELISA ...... 50 2.2.12 PROTEIN SECRETION ASSAY ...... 51 2.2.13 INFLAMMATORY GENE MICROARRAY ...... 52 2.2.14 STATISTICS ...... 52 2.3 RESULTS ...... 53 2.3.1 NEURAL PROGENITOR CELL-DERIVED SOD1 (G93A) ASTROCYTES UP-REGULATE ENDOGENOUS HSPB1 ...... 53 2.3.2 IN VITRO NON-CELL AUTONOMOUS NEUROPROTECTION FROM HUMAN HSPB1 ...... 56 2.3.3 NON-CELL AUTONOMOUS NEUROPROTECTION MAY BE PHOSPHORYLATION DEPENDENT ...... 64 2.3.4 MUTATIONS IN HSPB1 DISRUPT ITS NEUROPROTECTIVE PROPERTIES IN VITRO ...... 65 2.3.5 HSPB1 OVEREXPRESSION IN PATIENT-DERIVED ASTROCYTES ...... 71 2.3.6 HSPB1-dHMN PATIENT-DERIVED ASTROCYTES ARE NOT TOXIC IN VITRO ...... 75 2.4 DISCUSSION ...... 78 ix

CHAPTER 3: MUTATIONS IN HSPB1 DISRUPT TNFa INDUCED NF-kB SIGNALING ...... 85 3.1 INTRODUCTION ...... 85 3.2 EXPERIMENTAL METHODS ...... 90 3.2.1 HELA CELL CULTURE ...... 90 3.2.2 PLASMIDS ...... 91 3.2.3 TRANSIENT TRANSFECTIONS ...... 91 3.2.4 NF-kB LUCIFERASE ASSAYS ...... 92 3.2.5 TNFa STIMULATION OF HELA CELLS ...... 92 3.2.6 WESTERN BLOTTING ...... 93 3.2.7 IMMUNOFLUORESCENCE ...... 95 3.2.8 CO-IMMUNOPRECIPITATIONS ...... 95 3.2.9 STATISTICS ...... 96 3.2.10 HEAT SHOCK EXPERIMENTS ...... 97 3.3 RESULTS ...... 97 3.3.1 TNFa STIMULATION RAPIDLY ACTIVATES NF-kB SIGNALING IN HELA CELLS ...... 97 3.3.2 HSPB1(WT) OVEREXPRESSION REDUCES TNFa-INDUCED NF-kB ACTIVITY ...... 100 3.3.3 OVEREXPRESSION OF HSPB1(WT), HSPB1(G84R), OR HSPB1(R136W) DOES NOT AFFECT p65 RELEASE AND NUCLEAR TRANSLOCATION ...... 104 3.3.4 HSPB1 IS NOT ASSOCIATED WITH THE IKK COMPLEX IN HELA CELLS ...... 108 3.4 DISCUSSION ...... 110 CHAPTER 4: HSPB1-ASSOCIATED ASTRC COMPLEX IS NOT PRESENT IN HELA CELLS DURING CELLULAR STRESS ...... 114 4.1 INTRODUCTION ...... 114 4.2 EXPERIMENTAL METHODS ...... 119 4.2.1 CELL CULTURE ...... 119 4.2.2 PLASMIDS ...... 119 4.2.3 TRANSIENT TRANSFECTIONS ...... 119 4.2.4 siHSPB1 TRANSFECTIONS ...... 120 4.2.5 CO-IMMUNOPRECIPITATIONS ...... 120 x

4.2.6 WESTERN BLOTTING ...... 121 4.2.7 mRNA DECAY ASSAY ...... 122 4.2.8 QPCR MEASUREMENTS...... 123 4.2.9 RNA CO-IMMUNOPRECIPITATION ASSAY ...... 124 4.2.10 STATISTICS ...... 125 4.3 RESULTS ...... 125 4.3.1 HSPB1 DOES NOT BIND mRNA IN HELA CELLS ...... 126 4.3.2 HSPB1 OVEREXPRESSION DOES NOT ALTER ARE-mRNA DECAY IN HELA CELLS ...... 128 4.3.3 ASTRC COMPLEX FAILS TO FORM IN HELA CELLS AFTER TNFa TREATMENT ...... 131 4.4 DISCUSSION ...... 133 CONCLUSIONS ...... 136 REFERENCES ...... 139

LIST OF TABLES

Table 1.1 Classification of human heat shock protein families ...... 5 Table 1.3 Properties of HSPB1 mutants in vitro ...... 37 Table 2.1 HSPB1 construct primer sequences ...... 47 Table 3.2 Antibodies used in this chapter ...... 94 Table 4.2 qPCR primers used in this chapter ...... 124

xii

LIST OF FIGURES

Figure 1.1 Cell stress conditions that induce the heat shock response...... 2 Figure 1.2 The protein quality control network...... 3 Figure 1.3 The small heat shock protein family...... 7 Figure 1.4 Gene structure and protein domain sequence of HSPB1...... 9 Figure 1.5 Crystal structure of B1a...... 10 Figure 1.6 Nuclear translocation of HSPB1 in response to heat shock...... 14 Figure 1.7 Proposed mechanism of Hsp27-substrate binding...... 18 Figure 1.9 Involvement of HSPB1 in apoptotic pathways...... 22 Figure 1.10 The effects demyelination and axonal damage on myelinated axons...... 32 Figure 1.11 Mutations in HSPB1 result in CMT2F or dHMNII...... 34 Figure 2.1 Knockdown of murine HSPB1 in SOD1(G93A) astrocytes has no effect on motor neuron survival...... 55 Figure 2.2 Overexpression of HSPB1(WT) in SOD1(G93A) astrocytes improves MN survival...... 57 Figure 2.3 Expression of SOD1(G93A) has no effect on astrocyte viability...... 59 Figure 2.4 Overexpression of HSPB1(WT) does not reduce SOD1(G93A) expression...... 61 Figure 2.5 Detection of SOD1 and HSPB1 in astrocyte conditioned medium. .... 63 Figure 2.6 Phosphomimetic HSPB1 expression in SOD1(G93A) astrocytes protects MNs from astrocyte-mediated toxicity...... 65 Figure 2.7 Mutations in HSPB1 attenuate non-cell autonomous motor neuron protection...... 68 Figure 2.8 Overexpression of HSPB1 constructs alters SOD1(G93A) astrocyte inflammatory profile...... 70 Figure 2.9 HSPB1 expression levels in ALS patient i-astrocytes...... 72 Figure 2.10 Overexpression of HSPB1(WT) in fALS i-astrocytes improves MN survival...... 76 Figure 2.12 dHMN-patient derived i-astrocytes are not toxic to MNs in co-culture...... 77 Figure 3.1 HSPB1 regulates classical NF-kB signaling at multiple levels...... 99 Figure 3.3 TNFa-stimulated NF-kB reporter assay...... 101 Figure 3.4 Overexpression of HSPB1(WT) reduces NF-kB activity in HeLa cells...... 103 Figure 3.5 Overexpression of HSPB1 does not alter IkBa degradation following TNFa stimulation...... 105 xiii

Figure 3.6 Overexpression of HSPB1 does not affect P65 nuclear translocation following TNFa stimulation...... 107 Figure 3.7 HSPB1 and IKKb do not co-precipitate in HeLa cells...... 109 Figure 4.1 Domain organization of AUF1 isoforms...... 116 Figure 4.2 RNA does not precipitate with HSPB1 in naive HeLa cells...... 127 Figure 4.3 RNA does not precipitate with HSPB1 in TNFa stimulated HeLa cells...... 128 Figure 4.4 HSPB1 expression levels do not alter mRNA stability in HeLa cells...... 130 Figure 4.5 Immunoprecipitation of ASTRC subunits from HeLa cells...... 133

xiv

CHAPTER 1: INTRODUCTION

1.1 HEAT SHOCK PROTEINS

1.1.1 THE HEAT SHOCK PROTEIN FAMILY

Heat shock proteins (HSPs), or heat stress proteins, are a large group of proteins first described in temperature-shocked Drosophilla melanoganster larva in the 1960s (Ritossa, 1962). Since this initial observation, heat shock proteins have been identified in plants, bacteria, yeast and mammals. HSPs are multifunctional, and are involved in almost every cellular process, from maintenance of the cell cycle and regulating cell proliferation, to organization of the cytoskeleton and regulation of apoptosis (Figure 1.1). HSPs act as molecular chaperones, participating in the synthesis, folding, and transport of proteins, as well as the prevention of protein aggregation (Figure 1.2). They work closely with the ubiquitin/proteasome and autophagic systems to ensure the integrity of the cellular proteome (Acunzo et al., 2012; McDonough and Patterson, 2003;

Wyttenbach et al., 2000). Under stressful conditions like hyperthermia, ischemia, hypoxia, and oxidative stress, heat shock protein expression is up-regulated, providing potent cytoprotection to cells. The induction of HSPs in response to stress, known as the heat shock response, is regulated by the interaction of heat

1 shock transcription factors (HSFs) with gene promoter elements called heat shock elements (HSEs) (Richter et al., 2010). Although HSPs were initially identified as stress-inducible proteins, it is now known that HSP families also contain constitutively expressed members.

Figure 1.1 Cell stress conditions that induce the heat shock response. Major categories of environmental and physiological stress inducers of the HSR include environmental stress, growth and development, pathophysiology, and protein conformational diseases. From (Morimoto, 2011).

Figure 1.2 The protein quality control network. The protein quality control network combines the actions of many chaperones to facilitate the folding of proteins into their native conformations, to prevent misfolding and to ‘reactivate’ misfolded and aggregated proteins. Chaperones aid in the co-translational folding of newly synthesized proteins (1), the folding of unfolded proteins recently released from the ribosome (2) and the remodelling of non-native proteins (3), which are often caused by heat stress and other stress conditions (4). Additionally, non-native proteins may be targeted to chaperone-associated proteases for degradation (5). Some molecular chaperones slow or prevent protein misfolding and aggregation (6). If the chaperone network becomes overwhelmed during stress, non-native proteins may form large, amorphous aggregates (7). However, other chaperones can extract and unfold polypeptides from aggregates (8), thus providing another chance for proper folding (3). From (Doyle et al., 2013).

HSPs are classified based on their molecular weights (Table 1.1). The larger HSPs (HSP60, HSP70, HSP90) are ubiquitously expressed, ATP- dependent chaperones. HSP70 is the most abundant molecular chaperone and is found within all compartments of eukaryotic cells (Mayer and Bukau, 2005;

Richter-Landsberg, 2007). HSP70 binds to nascent polypeptide chains and partially folded protein intermediates to prevent their misfolding and aggregation

(Mayer, 2013; Meimaridou et al., 2009). HSP70 interacts with HSP40, a co- chaperone, and a HSP110, a nucleotide exchange factor (NEF), to promote protein refolding and repair, and disassembly of protein aggregates (Kampinga and Craig, 2010; Misselwitz et al., 1998). HSP90 is also highly abundant in eukaryotic cells, comprising up to 2% of all soluble protein in an unstressed cell

(Pratt et al., 2010). HSP90 appears to be a more specialized chaperone, as its client proteins, the term for HSP substrates, are enriched in signaling molecules such as kinases and transcription factors (Citri et al., 2006; Saibil, 2013;

Theodoraki and Caplan, 2012). Less abundant are HSP60 chaperones, also known as chaperonins, which are localized to mitochondria, and aid in folding of mitochondrial proteins. Recent evidence suggests that HSP60 may leave the mitochondria in tumors to inhibit apoptosis and promote cell survival (Cappello et al., 2008).

Table 1.1 Classification of human heat shock protein families

The small heat shock protein (sHSP) family are ATP-independent chaperones (Jakob et al., 1993). In humans, this family consists of 10 proteins, which all contain a highly-conserved a-crystalline domain ~90 residues long, flanked by variable NH2- and COOH- terminal regions that are highly disordered

(Figure 1.3) (Kappe et al., 2003; Kriehuber et al., 2010). These proteins form

5 reversible, poly-disperse hetero- and homo- oligomeric complexes that can be regulated in part by phosphorylation (Haslbeck, 2002; Hayes et al., 2009). This ability to modulate their oligomeric state allows them to trap misfolded proteins and hold them in a folding-competent state, defined in the literature as a

“holdase” activity (Horwitz, 1992; Jakob et al., 1993). Bound proteins can then be transferred to the HSP70-HSP90 machinery for re-folding or degradation

(Bryantsev et al., 2007; Ehrnsperger et al., 1997; Haslbeck, 2002). In addition to their chaperone activity, sHSPs are involved in signal transduction, transcription, apoptosis regulation, cytoskeleton remodeling and proteolysis (Bakthisaran et al.,

2015; Haslbeck et al., 2005; Kamradt et al., 2005; Van Montfort et al., 2001). One of the 10 proteins, heat shock protein B1 (HSPB1), is the subject of this thesis.

Figure 1.3 The small heat shock protein family. A. Table of small heat shock proteins and some of their properties [Modified from (Mymrikov et al., 2011)]. B. General structure of a small heat shock protein. Note that there is a high amount of variance in the N-terminal and C-terminal regions between individual small heat shock proteins.

1.2 HEAT SHOCK PROTEIN BETA 1 (HSPB1)

1.2.1 STRUCTURE AND GENE REGULATION

The human HSPB1 gene is located on chromosome 7q11.23, and two

7 pseudogenes are found in the ninth and X chromosomes respectively (Stock et al., 2003). The HSPB1 gene contains a consensus HSE upstream of its transcriptional start site, which is the primary regulatory element (Figure 1.4A).

Interestingly, deletion of this HSE site in two breast cancer cell lines increased basal activity of the HSPB1 promoter by twofold (Oesterreich et al., 1996). This led to the discovery of non-HSE regulatory elements in the HSPB1 gene, notably an SP1 site and AP2 site (de Thonel et al., 2012; Hickey et al., 1986).

Furthermore, an additional HSE element was found within the first intron of

HSPB1 (i-HSE). In Vitro studies using a chloramphenicol acetyltransferase (CAT) reporter demonstrated that this i-HSE negatively regulates transcriptional activity of the reporter in the presence of the upstream promoter HSE (Cooper et al.,

2000).

Figure 1.4 Gene structure and protein domain sequence of HSPB1. A. Gene structure of HSPB1, including transcription factor binding sites and location of coding sequence exons. B. Sequence alignment of human (hHSPB1), rat (rHSPB1) and mouse (mHSPB1) protein sequences. An (*) indicates a fully conserved residue. An (:) indicates conservation between groups of strongly similar properties. A (.) indicates conservation between groups of weakly similar properties. Coloration guide: Red = small hydrophobic residue, Blue = acidic residue, Magenta = basic residue, Green = hydroxyl, sulfhydryl and amine containing residues. Sequence alignment from (Sievers et al., 2011).

HSPB1 encodes for HSPB1, a 205-amino acid polypeptide that is highly conserved across species (Figure 1.4B). Like all sHSPs, it bears the hallmark a- crystalline domain flanked by NH2- and COOH- terminal regions. The NH2-

9 terminal region contains 3 phosphorylation sites, and a hydrophobic WDPF-motif, while the COOH- terminal region contains a highly conserved (I/V/L)-X-(I/V/L) motif. The a-crystalline domain folds into a compact b-sandwich, comprised of 2 b-sheets with 3 b-strands per sheet (Figure 1.5). Crystallographic studies propose that anti-parallel interactions between the b7 strands of two a-crystalline domains give rise to sHSP dimers (Baranova et al., 2011).

Figure 1.5 Crystal structure of B1a. Superposition of the monomers of B1 a (green) and human aB-crystallin. a-crystalline domain fragment (PDB ID: 2WJ723) (violet) shown as ribbon diagrams. Secondary structure elements that are missing in the B1a structure but are present in aB-crystalline are colored cyan. From (Baranova et al., 2011).

1.2.2 HSPB1 PHOSPHORYLATION

HSPB1 was first identified as a phosphoprotein in embryonic rat fibroblasts grown under stressful conditions (heat shock) or when treated with phorbol ester or a calcium ionophore (Welch and Suhan, 1985). Phosphorylation occurs at 3 serine residues, Ser15, Ser78 and Ser82, located in the NH2-terminal region of the protein (Figure 1.4B). Numerous kinases are responsible for

HSPB1 phosphorylation, most notably the mitogen-activated protein kinases

MK2 and MK3, which are targets for MAP p38 protein kinase (Stokoe et al.,

1992a; Stokoe et al., 1992b). Other kinases shown to phosphorylate HSPB1 include the MK5 (New et al., 1998), Akt/protein kinase B (O'Shaughnessy et al.,

2007; Rane et al., 2003), protein kinase C (Takai et al., 2007), protein kinase D

(Doppler et al., 2005; Kostenko and Moens, 2009; Stetler et al., 2012), and cGMP-dependent protein kinase (Butt et al., 2001).

1.2.3 HSPB1 OLIGOMERIZATION

Another hallmark of sHSPs is the ability to form poly-disperse hetero- and homo-oligomers, and HSPB1 has been studied extensively in this regard. In cells, HSPB1 is in equilibrium between large oligomeric structures and smaller dimers (Bova et al., 2000; Hayes et al., 2009; Lelj-Garolla and Mauk, 2012). As stated above, HSPB1 dimerization is predicted to be the result of interactions between a-crystalline domains of HSPB1 monomers, while the formation of

11 higher-order oligomers is mediated by the NH2-terminal domain (Lambert et al.,

1999). This oligomerization is reversible, and HSPB1 dimers are able to rapidly exchange between oligomeric species (Bova et al., 2000; Stengel et al., 2010).

Dimer formation is dependent on the cysteine residue at position 137, and on inter-subunit contacts made by the b7 strand (Diaz-Latoud et al., 2005). In 2012,

Lelj-Garolla et al. demonstrated that the NH2-terminal region of HSPB1 is involved in the formation of higher order oligomers (Lelj-Garolla and Mauk,

2012). Using sedimentation velocity experiments, Lelj-Garolla et al. found that

D1-14 and D1-24 N-terminal truncated HSPB1 display decreased oligomeric size compared to wild-type HSPB1, while loss of the COOH-terminal region, D182-

205, had no effect on oligomeric size (Lelj-Garolla and Mauk, 2012).

Changes in the native size of HSPB1 oligomers can occur rapidly in response to temperature, or physiological alterations of the cellular environment.

Studies have shown the temperature dependence of HSPB1 self-association. At low temperatures, 10°C, the largest oligomers identified were 8-12 mers, while at

40°C HSPB1 forms oligomers up to 30-mer in size (Lelj-Garolla and Mauk,

2006). It is generally accepted that phosphorylation of HSPB1 in response to stress results in the rapid dissociation of larger oligomeric species into dimers

(Lelj-Garolla and Mauk, 2006; McDonald et al., 2012; Shashidharamurthy et al.,

2005). This phenomenon can also be observed when the 3-phosphorylatable serine residues are replaced with aspartic acids to mimic HSPB1 phosphorylation. There are exceptions to this rule however, as phosphomimetic 12

HSPB1 can also form large oligomers at high concentrations (Lelj-Garolla and

Mauk, 2006). Furthermore, in vivo and in vitro studies of HSPB1 phosphomimics in tumor cells found that cell-cell contacts can induce the formation of large oligomers, even with phosphomimetic constructs (Bruey et al., 2000b). Detailed analysis of HSPB1’s oliogmeric state in vivo is challenging, and more work is needed to fully understand the mechanisms that influence HSPB1 oligomerization, especially in neuronal tissues.

1.2.4 DIFFERENTIAL REGULATION AND CELLULAR LOCALIZATION

HSPB1 is expressed ubiquitously, and under non-stressed conditions

HSPB1 expression is highest in the eye lens and smooth, skeletal and cardiac muscle tissues (Kato et al., 1992; Sugiyama et al., 2000). Cellular stress, such as heat shock, hypoxia and reactive oxygen species, rapidly induces HSPB1 expression in most cells. In the nervous system, basal HSPB1 expression occurs primarily in the motor and sensory neurons of the brain stem and spinal cord.

Upon stress, HSPB1 expression is rapidly induced in astrocytes and oligodendrocytes. Interestingly, neurons do not appear to up-regulate HSPB1 after stress.

Cellular localization of HSPB1 is highly dependent on the cell type and the state of the cell (Hoch et al., 1996; Lutsch et al., 1997). In non-stressed conditions, most HSPB1 resides in the cytoplasm, while a small fraction localizes to the nucleus (Figure 1.6) (McClaren and Isseroff, 1994). Heat shock causes

HSPB1 to translocate into the nucleus, and also increases its association with the cytoskeleton (McClaren and Isseroff, 1994). It is hypothesized that the re- distribution of HSPB1 is partly responsible for cellular protection and survival in response to heat shock.

Figure 1.6 Nuclear translocation of HSPB1 in response to heat shock. HeLa cells were subjected to a 1 hour heat shock @ 45°C and then allowed to recover for 3 hours (HS +3R) or 24 hours (HS + 24R). Cells receiving no heat shock (No HS) were used as a control sample. Total (T), cytoplasmic (C) and nuclear (N) fractions were created from each sample and analyzed for the expression of HSPB1, Lamin B1 and Actin by semi-quantitative western blotting. See section 3.2.10 for methodology.

1.3 CELLULAR FUNCTIONS OF HSPB1

1.3.1 MOLECULAR CHAPERONE

Unlike the ATP-dependent HSP families, the sHSPs are incapable of refolding denatured polypeptides on their own. Instead, they exhibit a

“chaperone-like” activity, in which they bind to denatured or misfolded polypeptides and hold them in a folding competent state (Horwitz, 1992; Jakob et al., 1993). These trapped proteins can then be transferred to the larger HSP machinery for refolding, or degradation via the ubiquitin/proteasome or autophagic systems (Bryantsev et al., 2007; Ehrnsperger et al., 1997;

McDonough and Patterson, 2003; Nivon et al., 2009; Shen et al., 2016). A common assay used to measure HSPB1 chaperone activity involves denaturing a substrate protein by heat or disulfide bond reduction in the presence of HSPB1, and then monitoring either the formation of light scattering aggregates or protein solubility via centrifugation (Haslbeck and Buchner, 2002; Lelj-Garolla and Mauk,

2005; Rogalla et al., 1999; Shashidharamurthy et al., 2005). Despite extensive study, the exact mechanism by which HSPB1 recognizes client proteins and prevents their aggregation is unclear and remains an active area of research.

Two mechanisms of HSPB1 chaperone activity have been posited in the literature. The first proposes that large oligomers of HSPB1 are responsible for binding to denatured proteins, and phosphorylation of HSPB1 disrupts this activity by inducing oligomer dissociation (Rogalla et al., 1999). Both HSPB1 and

Hsp25 (murine HSPB1) prevented the aggregation of thermally denatured citrate synthase or DTT-induced insulin denaturation. Phosphomimetic mutants of 15

HSPB1 displayed reduced inhibition of substrate aggregation compared to wild- type HSPB1. Interestingly, the authors noted that phosphorylation did not completely inhibit the chaperone activity of HSPB1. Lelj-Garolla and Mauk confirmed this mechanism and further showed that chaperone activity was thermally induced (Lelj-Garolla and Mauk, 2006; Lelj-Garolla and Mauk, 2012).

Using light scattering assays, they examined DTT-induced aggregation of insulin.

Their data showed that wild-type HSPB1 forms large oligomers, prevents aggregation of insulin, and displays the highest chaperone activity above 40°C.

In agreement with Rogalla et al. (1999), the authors found that phosphomimetic

HSPB1 forms primarily dimers and tetramers, and has reduced chaperone activity compared to wild-type HSPB1. They observed that a small portion of phosphomimetic HSPB1 does form large oligomers of similar size to those of wild-type HSPB1. This suggests that large oligomers of HSPB1 are responsible for chaperone activity and dissociation of these large oligomers by phosphorylation disrupts this activity.

At the same time as Lelj-Garolla and Mauk. proposed their mechanism, a second group suggested a different model for HSPB1 chaperone activity (Figure

1.7). They found that large oligomers formed by wild-type HSPB1 do not prevent protein aggregation, but rather the small, phosphorylated dimers are responsible for recognition and binding substrates, and that once a substrate is bound

HSPB1 re-forms larger oligomers (McDonald et al., 2012; Shashidharamurthy et al., 2005). They find that phosphomimetic HSPB1 is predominately dimeric, and 16 binds to a model substrate, T4 Lysozyme (T4L), with a 4-fold greater affinity than wild-type HSPB1 (Shashidharamurthy et al., 2005). In 2009, Hayes and colleagues reached similar conclusions, by comparing the chaperone activity of full-length wild-type HSPB1 to the phosphomimetic variant, and to a fully phosphorylated wild-type HSPB1 (Hayes et al., 2009). They show that phosphorylated HSPB1 and the phosphomimetic variant predominately form dimers, while wild-type HSPB1 predominates in large oligomers. The authors induced the aggregation of 2 model substrates, insulin and a-lactalbumin, using

DTT. Phosphorylated wild-type HSPB1 demonstrated the greatest chaperone activity for both substrates, followed by the phosphomimetic mutant. The wild- type HSPB1 displayed the worst chaperone activity (Hayes et al., 2009).

Importantly, Hayes et al. also found that HSPB1 is unable to reverse protein aggregation of insulin, regardless of its phosphorylation state.

Figure 1.7 Proposed mechanism of Hsp27-substrate binding. A. Equation 1 is the unfolding equilibrium of the substrate where I and U are respectively partially and globally unfolded states. Equation 2 represents the activation equilibrium of the sHSP. For Hsp27 the equilibrium is modulated by concentration and phosphorylation. Equation 3 is the reassembly following binding of partially or globally unfolded substrates. C refers to the Hsp27/substrate complex. B. Model of substrate binding to Hsp27 highlighting the role of the N-terminal domain in the three equilibria of Scheme 2. The N-terminal domain is shown to be unfolded and solvent exposed in the dimer. The only contacts between Hsp27 monomer in the dimer is along the a-crystallin domain interface. Data show that upon reassembly the N-terminal domain and presumably the T4L are sequestered on the inside of the oligomer (dashed box). The low affinity binding which may involve T4L binding to the surface of the Hsp27 oligomer is hypothetical. Modified from (McDonald et al. 2012). 18

1.3.2 REGULATION OF CELL SIGNALING/APOPTOSIS

Apoptosis or programmed cell death, is a highly coordinated and energy dependent process used by multicellular organisms to eliminate cells that are damaged or no longer needed. There are 2 major pathways that initiate apoptosis, the extrinsic pathway and the intrinsic pathway (Figure 1.8) (Beesoo et al., 2014). The extrinsic pathway is activated by ligand binding to transmembrane receptors of the tumor necrosis factor (TNF) receptor superfamily (Locksley et al., 2001). Ligand binding to the receptor results in the formation of a death-inducing signaling complex (DISC), activating caspase-8, which triggers the execution phase of apoptosis (Takayama et al., 2003). The intrinsic pathway is activated by non-receptor mediated stimuli, including hypoxia, hyperthermia, free radicals, toxins, and cytokines, that result in the activation of pro-apoptotic Bax, and Bid proteins that stimulate the release of cytochrome c and second mitochondria-derived activator of caspases (Smac/Diablo) from mitochondria (Elmore, 2007; Takayama et al., 2003). Cytosolic cytochrome c interacts with apoptotic protease activating factor-1 (Apaf-1) and procaspase-9 to form the “apoptosome” (Hill et al., 2004). Formation of the apoptosome activates caspase-9, which in turn activates caspase-3 and caspase-7. Smac/Diablo promotes apoptosis by inhibiting the activity of inhibitors of apoptosis proteins

(IAP).

Figure 1.8 Extrinsic and intrinsic pathways involved in apoptosis. Apoptotic cell death can be induced through the extrinsic (also called receptor-mediated) or the intrinsic (also called mitochondria-mediated) signaling pathways. The extrinsic pathway involves ligation of death receptors with their ligands resulting in a sequential activation of caspase- 8, and -3, which cleaves target proteins leading to apoptosis. This pathway is negatively controlled by the anti-apoptotic proteins c-FLIP and XIAP that inhibit activation of caspase-8 and caspase-3, respectively. Intrinsic death 20 stimuli, e.g. ROS, DNA-damaging reagents, or Ca2+ mobilisation directly or indirectly activate the mitochondrial pathway by inducing release of cytochrome c and formation of the apoptosome, composed of Apaf-1 and caspase-9. Caspase-9 is activated at the apoptosome and, in turn, activates pro-caspase-3. This death pathway is largely controlled by the proapoptotic (e.g. Bax, Bak, Bid and Smac/DIABLO) and anti-apoptotic (e.g. Bcl-2, Bcl- xL, Mcl-1 and XIAP) proteins. Caspase-8 may also induce cleavage of Bid, which induces the translocation of Bax and/or Bak to the mitochondrial membrane and amplifies the mitochondrial apoptosis pathway. From (Beesoo et al., 2014).

HSPB1 displays potent anti-apoptotic properties, and can regulate both the intrinsic and extrinsic apoptotic pathways (Figure 1.9). In the intrinsic pathway, HSPB1 can act upstream of cytochrome c release from mitochondria, by preventing the translocation Bid (Paul et al., 2002), and Bax (Havasi et al.,

2008; Rane et al., 2003; Wu et al., 2007). HSPB1 can also interact with cytosolic cytochrome c to inhibit apoptosome formation (Bruey et al., 2000a), and prevent cleavage of pro-caspase 3 by the apoptosome (Pandey et al., 2000). The extrinsic pathway can be inhibited by interactions between HSPB1 and Daxx,

(Charette et al., 2000) or through the promotion of Akt activation (Kim et al.,

2013; Kim et al., 2012; Zhuang et al., 2010). HSPB1 can also regulate nuclear- factor kB (NF-kB) signaling cascades (discussed in chapter 3), which play a role in promoting cell survival, growth and proliferation.

Figure 1.9 Involvement of HSPB1 in apoptotic pathways. From (Mymrikov et al., 2011).

Where in the signaling pathway HSPB1 acts to inhibit apoptosis depends not only on the inducer of apoptosis, but also on the phosphorylation state and oligomeric state of HSPB1. Charette and colleagues demonstrated that phosphorylation of HSPB1 resulted in its translocation to the nucleus, where it prevents the activation of Daxx (Charette and Landry, 2000; Charette et al.,

2000). Small, phosphorylated HSPB1 oligomers can stabilize the cytoskeleton

(see below), inhibiting apoptosis induced by drugs that disrupt F-actin architecture (Paul et al., 2002; Paul et al., 2010), and also activate Akt in response to TRAIL-induced apoptosis (Kim et al., 2013; Kim et al., 2012).

1.3.3 REGULATION OF THE CYTOSKELETON

The eukaryotic cytoskeleton is composed of 3 major components, microtubules (MTs), microfilaments (MFs) and intermediate filaments (IFs). The cytoskeleton plays a vital role in maintaining the shape and integrity of a cell, spatial organization of subcellular organelles and cellular motility (Fletcher and

Mullins, 2010). MTs are polymers of tubulin ~25 nm in diameter, which provide structural strength and cell shape. Inside a cell, they act as highways on which kinesin and dynein motor proteins can transport specific cellular cargoes, including organelles (Hirokawa, 1998). This is particularly important in neurons, where axonal transport along MTs is crucial to cell survival (Goldstein and Yang,

2000). MFs, also referred to as actin filaments, are primarily responsible for cell motility. MFs are formed by the rapid polymerization and depolymerization of actin. Monomeric actin is termed globular actin or G-actin, while the polymerized 23 form of actin is denoted as filamentous actin or F-actin (Oda et al., 2009). One of the first signs of cellular stress is the disruption of the cytoskeleton and loss of actin fibers. IFs are the main component of the cytoskeleton that provides structural support to the cell. Over 70 genes encode for IFs in the human genome, and they are divided into six groups based on their structure (Helfand et al., 2004). Notably, the IF family contains the glial fibrillary acidic protein (GFAP) and the neurofilament light, medium and heavy chains, all of which play an important role in the health and maintenance of cells in the nervous system.

HSPB1 can interact with all 3 major cytoskeletal components. HSPB1 was first identified as an inhibitor of actin polymerization, able to cap the plus end of actin filaments (Miron et al., 1991; Miron et al., 1988). HSPB1’s interaction with actin is hypothesized to protect the cytoskeleton from stress. Overexpression of

HSPB1 protected cells from actin fragmentation induced by oxidative stress, while overexpression of a non-phosphorylatable HSPB1 mutant did not (Huot et al., 1996). Phosphorylated HSPB1 can bind to thermally denatured F-actin and prevent its aggregation, sequestering it in highly soluble complexes (Pivovarova et al., 2007; Pivovarova et al., 2005). Recent evidence indicates that non- phosphorylated HSPB1 interacts with F-actin in neurons, and heat shock causes not only the phosphorylation of HPSB1, but promotes its interaction with F-actin

(Clarke and Mearow, 2013). Less is known about the interactions of HSPB1 with

IFs and MTs. Alexander disease is linked to mutations in the intermediate filament GFAP, which cause the protein to aggregate into bundles termed 24

Rosenthal Fibers (Li et al., 2005). HSPB1 and HSPB5 have been identified as components of Rosenthal Fibers, although their role in fiber formation is unknown

(Iwaki et al., 1989). MT binding by HSPB1 was first identified in studies of neuropathy-associated HSPB1 mutants (Almeida-Souza et al., 2011) (discussed in 1.5). It has since been shown that wild-type HSPB1 also binds MTs, facilitating the formation of non-centrosomal MTs in HeLa cells (Almeida-Souza et al.,

2013).

1.4 HSPB1 AND THE NERVOUS SYSTEM

1.4.1 EXPRESSION OF HSPB1 IN THE NERVOUS SYSTEM

Systematic analysis of HSPB1 mRNA and protein expression levels has been carried out in both mouse and rat models (Armstrong et al., 2001; Chen and Brown, 2007; Quraishe et al., 2008). These studies demonstrated that under physiological conditions, HSPB1 mRNA and protein are constitutively expressed in the brain and spinal cord. Glial cells and motor neurons express the most

HSPB1 protein. In response to stress or injury, HSPB1 expression can be rapidly induced in both the brain and the spinal cord (Benn et al., 2002; Kato et al., 1994;

Krueger-Naug et al., 2000). Interestingly, in the central nervous system (CNS),

HSPB1 up-regulation is primarily observed in glial cells, not neurons. Despite the lack of endogenous HSPB1 in most neurons, numerous studies have demonstrated the neuroprotective effect of exogenous HSPB1 in various pathological conditions, discussed in detail below.

1.4.2 HSPB1 AND NEURONAL INJURY

Injury to the axon of the CNS and peripheral nervous system (PNS) can initiate a number of cellular mechanisms that promote nerve repair, re- myelination and muscle re-innervation. This includes the up-regulation of HSPB1

(Benn et al., 2002; Costigan et al., 1998; Hirata et al., 2003; Keeler et al., 2012;

Lewis et al., 1999), which may promote axonal re-growth through its interaction with actin, and increase cell survival by inhibiting apoptotic signaling pathways.

Costigan et al. were the first to describe the upregulation of HSPB1 following peripheral nerve injury. They found that transection of the sciatic nerve in adult rats resulted in up-regulation of HSPB1 mRNA and protein in primary motor and sensory neurons (Costigan et al., 1998). In contrast, a sciatic axotomy in neonatal rats resulted in up-regulation of HSPB1 in only a few sensory neurons

(Lewis et al., 1999). Additionally, Benn et al. showed that HSPB1 was induced and phosphorylated following sciatic nerve transection in adult rats but not neonatal rats. Induction of HSPB1 was crucial for survival of motor and sensory neurons following injury (Benn et al., 2002). Furthermore, they demonstrated that overexpression of human wild-type HSPB1 promoted neuronal survival in the neonatal rats, while overexpression of a non-phosphorylatable variant did not.

More recently, Ma et al. observed that HSPB1 transgenic mice exhibit accelerated axonal growth following a crush or transection of the sciatic nerve compared to non-transgenic mice (Ma et al., 2011).

HSPB1 also plays a neuroprotective role in cerebral ischemia (An et al., 26

2008; Badin et al., 2006; Stetler et al., 2008; Stetler et al., 2012). In vitro experiments showed that transduction of PEP-HSPB1 protein can protect astrocytes and primary neuronal cells from oxidative stress, and adeno- associated virus (AAV)-HSPB1 transduction can protect primary cortical neuron cultures in response to oxygen-glucose deprivation (An et al., 2008; Stetler et al.,

2012). In vivo studies have shown similar results, with HSPB1 transgenic mice exhibiting improved behavioral recovery following focal ischemia (Badin et al.,

2009; Stetler et al., 2008; Stetler et al., 2012; van der Weerd et al., 2010). Stetler et al. demonstrated that wild-type HPSB1 and phosphomimetic HSPB1 can protect against ischemia, but a non-phosphorylatable variant did not.

Furthermore, they showed that HSPB1 exerts its neuroprotective function through the inhibition of apoptosis signal-regulating kinase 1 (ASK1), and that phosphorylation of HSPB1 is crucial for this regulation (Stetler et al., 2012).

1.4.3 HSPB1 AND ALS

ALS is an adult-onset, progressive motor neuron disease that is characterized by degeneration of upper and lower motor neurons in the motor cortex, brain stem and spinal cord (Katsuno et al., 2012). 10% of ALS cases follow a dominant inheritance pattern, and are classified as familial ALS (fALS), with the other 90% being termed sporadic ALS (sALS). Approximately 20% of fALS cases are a result of mutations in the gene encoding Cu/Zn superoxide dismutase 1 (SOD1) (Rosen et al., 1993). The exact mechanism of SOD1- mediated toxicity remains unclear, however most evidence suggests that 27 mutations confer a toxic gain of function, due partly to the propensity of mutant

SOD1 to misfold and aggregate. In vitro studies have demonstrated that HSPB1 and HSPB5, can reduce the rate of mutant SOD1 aggregation in vitro, in a concentration-dependent manner (Yerbury et al., 2013). Overexpression of

HSPB1 in SOD1-astrocyte cell cultures subjected to oxidative stress resulted in a

2-fold increase in cell viability, and reduced mutant SOD1 aggregation (Jin An,

2008). Additionally, viral delivery of HSPB1 and HSP70 to ND7 and dorsal root ganglion cells expressing mutant SOD1 reduced cell death in response to serum deprivation and a variety of other lethal stimuli (Patel et al., 2005). To test if

HSPB1 is neuroprotective in a mouse model, two groups crossed SOD1(G93A)-

ALS mice with mice overexpressing HSPB1. These studies have yielded conflicting results. Krishnan et al. found no neuroprotective effect in

SOD1(G93A)-HSPB1 double transgenic mice (Krishnan et al., 2008), while

Sharp et al. found that overexpression of HSPB1 delayed disease progression and increased motor unit survival (Sharp et al., 2008). Recently, two HSPB1 variants were identified in individuals with apparent sALS, further indicating a potential link between ALS and HSPB1 malfunction (Capponi et al., 2016)

1.4.4 HSPB1 AND PARKINSON’S DISEASE

Parkinson’s Disease (PD) is a neurodegenerative disorder characterized by the selective loss of dopaminergic neurons within the substantia nigra (SN)

(Wirdefeldt et al., 2011), resulting in a depletion of the critical neurotransmitter dopamine. A major pathological hallmark of PD is the formation of intracellular 28 protein aggregates, termed Lewy bodies, in surviving neurons of PD patients

(Gibb and Lees, 1988; Wirdefeldt et al., 2011). Alpha-synuclein (a-synuclein) and ubiquitin are the primary components of these aggregates, but over 300 other proteins have been identified in Lewy bodies, including HSP70 and HSPB1

(Leverenz et al., 2007; McLean et al., 2004; Outeiro et al., 2006). Aggregation of a-synuclein has been linked to the progression of PD, and missense mutations in the gene encoding a-synuclein result in autosomal dominant PD

(Polymeropoulos et al., 1997; Zarranz et al., 2004). Recent evidence suggests that neuronal toxicity is associated with soluble oliogmeric aggregates of a- synuclein (Winner et al., 2011), and that Lewy bodies may in fact be a cellular mechanism to sequester potentially toxic species of a-synuclein. Studies using

ND7 cells demonstrated that overexpression of HSPB1 can protect cells from a- synuclein-induced apoptosis (Zourlidou et al., 2004). Further, Outeiro et al. showed that HSPB1 is up-regulated at both the mRNA and protein levels in PD patient samples, and that overexpression of HSPB1 reduced a-synuclein aggregation and toxicity in H4 neuroglioma cells (Outeiro et al., 2006).

1.4.5 HSPB1 and ALZHEIMER’S DISEASE

Alzheimer’s disease (AD), the most common neurodegenerative disorder, results in progressive mental, functional and behavioral decline, and is characterized by the loss of cortical neurons (Mann, 1996). Pathological hallmarks of AD include insoluble extracellular protein deposits of amyloid-beta

29 protein (Ab), termed senile plaques, and intracellular accumulation of hyperphosphorylated tau, which form neurofibrillary tangles. HSPB1 and HSPB5 are up-regulated in AD patient brains (Renkawek et al., 1994; Renkawek et al.,

1993), and both co-localize with senile plaques (Wilhelmus et al., 2006b). The role of HSPB1 in AD remains unclear, but studies have shown that HSBP1 binds to mutant Ab, decreasing its aggregation and associated neurotoxicity (Lee et al.,

2006; Wilhelmus et al., 2006a).

1.5 MUTATIONS IN HSPB1 RESULT IN MOTOR NEURON DISEASE

1.5.1 CLINICAL DIAGNOSIS OF HSPB1-ASSOCIATED CMT2/dHMNII

Charcot-Marie-Tooth disease (CMT), or hereditary sensory and motor neuropathy (HSMN), is the most common inherited peripheral neuropathy, with a prevalence of approximately 1 in 2,500 (Skre, 1974). Distal hereditary motor neuropathy (dHMN) is clinically similar to CMT, but lacks significant sensory involvement, and is sometimes called “spinal CMT”. These neuropathies are characterized by length-dependent axonal degeneration of peripheral nerve axons, resulting in sensory loss, distal muscle wasting, and weakness (Ikeda et al., 2009; Zuchner and Vance, 2006). CMT can be divided into two groups, a demyelinating form (CMT1) and an axonal form (CMT2) (Figure 1.10).

Demyelinating neuropathies are characterized by a loss of the myelin sheath surrounding axons, resulting in reduced nerve conduction velocity and eventually axonal degeneration. In axonal neuropathies, the myelin sheath surrounding the

30 axon remains unaffected, while the axon itself undergoes degeneration. CMT is clinically and genetically heterogeneous, and mutations in over 80 genes have been identified in patients, including mutations in HSPB1, HSPB3 and HSPB8

(Evgrafov et al., 2004; Kolb et al., 2010; Timmerman et al., 2014). The majority of

CMT mutations are inherited in an autosomal dominant fashion, though there are cases of autosomal recessive and X-linked CMT.

Figure 1.10 The effects demyelination and axonal damage on myelinated axons. A. A single neuron and its axon, which has five myelin internodes that are separated by nodes of Ranvier. B. In an inherited demyelinating neuropathy, two myelin internodes have been lost, leaving demyelinated segments that will usually be re-myelinated. Even though demyelination is the primary pathology, axonal degeneration (see C) is a common, long-term consequence. C. In an inherited axonal neuropathy, the distal part of the axon has degenerated (dashed region), and the myelin sheaths that formerly surrounded the degenerated region have disappeared as a secondary consequence of the axonal degeneration. From (Scherer, 2011).

Since their discovery in the early 2000s, over 31 mutations in HSPB1 have been identified in CMT2F or dHMNII patients (Figure 1.11). HSPB1 mutations are almost exclusively inherited in an autosomal dominant fashion, except for one allele, L99M, that displays an autosomal recessive inheritance pattern. Most patients carrying a mutation in HSPB1 usually present symptoms between 20 and 40 years of age, although cases of onset as early as 4 and as late as 70 years of age have been documented (Capponi et al., 2011; Evgrafov et al., 2004;

Kijima et al., 2005; Luigetti et al., 2016). Current data indicates that mutations in the C-terminal region of HSPB1 correlate to an earlier disease onset (Kijima et al., 2005; Luigetti et al., 2016; Rossor et al., 2012a; Rossor et al., 2016).

Mutations result in slowly progressive distal muscle atrophy and weakness, beginning in the legs. Over time, atrophy and weakness appear in the upper limbs in a length-dependent manner. Patients exhibit foot deformations, steppage gait and decreased or absent deep tendon reflexes (Rossor et al., 2012b).

CMT2-linked mutations also result in distal sensory loss (Tazir et al., 2014).

Figure 1.11 Mutations in HSPB1 result in CMT2F or dHMNII. Domain structure of HSPB1 showing the flexible NH2 and COOH-terminal regions, and the highly conserved a-crystallin domain. The NH2 domain contains the major sites of phosphorylation, indicated as S15, S78 and S82 respectively. It also contains the conserved WDPF motif, which plays a role in HSPB1 oligomerization. Mutations in HSPB1 that result in Charcot Marie Tooth Disease Type 2F (CMT2F) are indicated in green, while mutations that result in distal Hereditary Motor Neuropathy (dHMN) are indicated in blue. Mutations that give rise to both CMT2F and dHMN are indicated in orange.

1.5.2 POTENTIAL MECHANISMS OF HSPB1-LINKED CMT2 AND dHMNII

Investigations into the molecular mechanisms underlying HSPB1- neuropathy are ongoing. The biophysical properties of many HSPB1 mutants

(mutHSPB1) have been studied in vitro (Table 1.2). HSPB1 mutant variants exhibit decreased chaperone activity, decreased thermostability and increased oligomeric size compared to wild-type protein (Muranova et al., 2015; Nefedova et al., 2013a; Nefedova et al., 2013b), supporting the idea that dimers of HSPB1, not higher order oligomers, are the species responsible for HSPB1’s observed chaperone activity (McDonald et al., 2012; Shashidharamurthy et al., 2005).

Interestingly, studies in cell culture models have yielding opposing results

(Almeida-Souza et al., 2011; Almeida-Souza et al., 2010). In SH-SY5Y cells, mutant forms of HSPB1 display either no change or an increase in chaperone activity and a corresponding decrease in oliogomeric size (Table 1.3). Taken together, these data suggest a model in which mutations in the a-crystalline domain alter dimer formation, while mutations in the N-terminal and C-terminal regions alter higher order oligomeric formation. It is important to note however that none of the mutations studied using recombinant protein have been investigated in a cell culture model. This is particularly important when investigating effects on chaperone activity, since HSPB1 can form heteroligomeric complexes with HSPB5 that also exhibit chaperone activity in vitro (Skouri-Panet et al., 2012). How mutations in HSPB1 might affect these interactions is unknown. In addition to altering chaperone activity, overexpression of HSPB1(P182L) or HSPB1(S135F) disrupts neurofilament assembly (Ackerley et al., 2006), and causes motor neuron toxicity and degeneration in vitro (Zhai et al., 2007).

Table 1.2 Physiochemical properties of HSPB1 mutants

Table 1.3 Properties of HSPB1 mutants in vitro

To study mutant HSPB1 in vivo we and others have developed mouse models of HSPB1-associated CMT2F. Overexpression of mutHSPB1 results in 37 the development of a very mild phenotype, with only electrophysiological deficits

(Srivastava et al., 2012), or a more severe phenotype with behavioral and electrophysiological deficits (d'Ydewalle et al., 2011; Lee et al., 2015). Peripheral nerves from the latter model display decreased abundance of acetylated a- tubulin, and DRG neurons cultured from these mice exhibit axonal degeneration and impaired axonal transport (d'Ydewalle et al., 2011). Treatment of these mice with histone deactylase 6 inhibitors increased acetylated a-tubulin levels, corrected the axonal transport deficits and rescued the CMT2F phenotype in mutant HSPB1 mice (d'Ydewalle et al., 2011). Importantly, mouse models expressing mutHSPB1 at endogenous levels fail to develop a neuropathy phenotype (Bouhy et al., 2016). This suggests that low or normal levels of mutHSPB1 expression may be insufficient to result in neuropathy, and disease onset may require an upregulation of the mutant protein, most likely through a cellular stress response.

CHAPTER 2: NON-CELL AUTONOMOUS NEUROPROTECTION OF HSPB1 IS DISRUPTED BY MUTATIONS THAT RESULT IN MOTOR NEUROPATHY

2.1 INTRODUCTION

While HSPB1 clearly demonstrates neuroprotective properties, the role of

HSPB1 in the nervous system remains unclear. Under physiological conditions, constitutive HSPB1 expression occurs primarily in glial cells and motor and sensory neurons of the brain stem and spinal cord (Plumier et al., 1997;

Wilhelmus et al., 2006b). In response to stress such as hyperthermia, ischemia or seizures, HSPB1 is rapidly induced (Akbar et al., 2001; Kato et al., 1995a;

Kato et al., 1995b; Kato et al., 1994; Krueger-Naug et al., 2000). Interestingly, this up-regulation occurs predominately in astrocytes, not neurons. HSPB1 up- regulation is also observed in many neurodegenerative disorders, including ALS,

AD, Multiple Sclerosis and PD (Maatkamp et al., 2004; Outeiro et al., 2006;

Peferoen et al., 2015; Wang et al., 2008). While these diseases result in the selective loss of neuronal cells, up-regulation of HSPB1 is not observed in neuronal cell populations, but in glial cells, suggesting that astrocytes may up- regulate HSPB1 as a protective mechanism in response to signaling from apoptotic neurons.

Neurons are extremely sensitive to changes in their microenvironment,

39 and studies have shown that astrocytes cooperate with neurons on multiple levels to help maintain neuronal homeostasis. Astrocytes protect neurons from excitotoxicity through the efficient removal of neurotransmitters from the synaptic cleft (Anderson and Swanson, 2000). Of particular note is the removal of glutamate, which can be extremely toxic to neurons (Sattler and Tymianski,

2001). Astrocytes recycle glutamate and transfer it back to neurons to maintain appropriate glutamate levels needed for neurotransmission (Bak et al., 2006).

Neurons also depend on astrocytes for protection from oxidative stress.

Glutathione (GSH) is an important anti-oxidant molecule in the brain, and its synthesis occurs in both neurons and astrocytes, however neuronal synthesis of

GSH highly depends on extracellular release of GSH by astrocytes (Dringen et al., 1999). Evidence indicates that depletion of GSH in astrocytes reduces their ability to protect neurons from oxidative stress (Chen et al., 2001), while increasing GSH synthesis in astrocytes promotes neuronal survival (Shih et al.,

2006). Furthermore, as immunocompetent cells, astrocytes can become activated through a process known as astrogliosis, and secrete a variety of cytokines, chemokines and growth factors to protect neurons from potentially deleterious threats (Belanger and Magistretti, 2009).

Due to the high level of cooperation between astrocytes and neurons, aberrant astrocyte activity can lead to neurotoxicity. Indeed, recent evidence has demonstrated the direct involvement of astrocytes in the progression of neurodegenerative disorders, including ALS. Pramatarova et al. were the first to 40 suggest the involvement of non-neuronal cells in SOD1-ALS pathogenesis after demonstrating that motor neuron-specific expression of mutant SOD1 was insufficient to induce neurodegeneration (Pramatarova et al., 2001). Later studies showed that selective reduction of mutant SOD1 in microglia or astrocytes drastically delayed disease progression in transgenic mice, but did not alter disease onset (Boillee et al., 2006; Yamanaka et al., 2008b). Since then, numerous in vitro astrocyte co-culture model systems have been established to model this non-cell autonomous neurodegeneration. These models propose that astrocytes may contribute to disease progression through multiple pathways, including disruption in glutamate uptake, release of toxic factors, or an inability of astrocytes to protect against reactive oxygen species (Fritz et al., 2013; Nagai et al., 2007; Rojas et al., 2014; Van Damme et al., 2007).

HSPB1 may be neuroprotective in SOD1-ALS. Endogenous levels of

Hsp25 (murine HSPB1) are up-regulated in transgenic mouse models of

SOD1(G93A)-ALS (Krishnan et al., 2008; Maatkamp et al., 2004; Vleminckx et al., 2002). This is also true in human post-mortem spinal cord from ALS patients

(Vleminckx et al., 2002). These HSPB1-positive cells were predominately glial in nature. Moreover, astrocytic inclusion bodies in ALS patient spinal cords co-stain for SOD1 and HSPB1 (Kato et al., 1997), and inclusion bodies isolated from various mutant SOD1 mouse models contain heat shock proteins (HSPs), including HSPB1. In this chapter I utilized an in vitro, astrocyte-motor neuron co- culture system (Dodge et al., 2008; Haidet-Phillips et al., 2011; Meyer et al., 41

2014; Papadeas et al., 2011) to study the consequences of wild-type or mutHSPB1 overexpression in murine-derived SOD1(G93A) astrocytes. I found that overexpression of wild-type HSPB1 in SOD1(G93A) astrocytes improves survival of co-cultured motor neurons through an unidentified mechanism, and that motor neuropathy-linked mutations in HSPB1 attenuate this neuroprotection.

My data further suggests that HSPB1 phosphorylation may be necessary for its non-cell autonomous neuroprotection. I also show that expression of mutHSPB1 in wild-type astrocytes has no effect on motor neuron survival. I further examined this non-cell autonomous neuroprotection in astrocytes derived from ALS-patient fibroblasts, and find that HSPB1 promoted MN survival in a single SOD1-ALS patient line. Finally, I characterized the non-cell autonomous neurotoxicity of astrocytes from a patient carrying the HSPB1(G84R) mutation.

2.2 EXPERIMENTAL METHODS

2.2.1 MICE

All procedures were performed in accordance with NIH Guidelines and approved by the Institutional Animal Care and Use Committee of the Research

Institute at Nationwide Children’s Hospital. Wild-type, B6SJL, and B6SJLTg

SOD1(G93A) mice were obtained from Jackson Laboratories (Bar Harbor, ME).

2.2.2 HUMAN TISSUE SAMPLES

Human postmortem neuronal progenitor cell (NPC)-derived astrocytes have been described previously (Ferraiuolo et al., 2011). Human skin fibroblast

42 samples were obtained from Stephen J. Kolb (ALS/MND Clinic, Department of

Neurology, The Ohio State University, Wexner Medical Center, Columbus, OH), as well as John Ravits (University of California, San Diego, School of Medicine) and Pamela J. Shaw. Informed consent forms were obtained from all subjects before sample collection.

2.2.3 ES MOTOR NEURON DIFFERENTIATION

Mouse embryonic stem cells expressing GFP under the MN- specific promoter HB9 (HBG3 cells; kind gift from Tom Jessell, Columbia University, New

York) were cultured on primary mouse embryonic fibroblasts (Millipore). For differentiation into MNs, cells were lifted with trypsin and resuspended in DFK10 culture medium consisting of knockout DMEM/F12, 10% knockout serum replacement, 1% N2, 0.5% L-glutamine, 0.5% glucose (30% in water), and

0.0016% 2-mercaptoethanol. The cells were plated on non-adherent Petri dishes to allow formation of embryoid bodies. After 1 day of recovery, 2 μM retinoic acid

(Sigma) and 2 μM purmorphamine (Calbiochem) were added freshly every day with new medium. After 5 days of differentiation, the embryoid bodies were dissociated and sorted for GFP on a BD FACSVantage/DiVa sorter.

2.2.4 MOUSE NPC ISOLATION AND ASTROCYTE DIFFERENTIATION

NPCs were isolated according to methods previously described (Miranda et al., 2012; Ray and Gage, 2006). Briefly, spinal cords were enzymatically dissociated in the same way as described for astrocytes. The cell suspension obtained was then mixed with an equal volume of isotonic Percoll (GE 43

Healthcare) and was centrifuged at 20,000 RCF for 30 minutes at room temperature. Cells from the low-buoyancy fraction (5-10 ml above the red blood cell layer) were harvested, washed thoroughly with D-PBS/PSF (Invitrogen) and plated on 60 mm uncoated plates. Cells were grown in growth medium

(DMEM/F12, Invitrogen) with 1% N2 supplement (Invitrogen), 20 ng/ml of fibroblast growth factor-2 (FGF-2, Peprotech, Rocky Hill, NJ) and 20 ng/ml of endothelial growth factor (EGF, Peprotech). Cells were first grown as neurospheres and then were placed on a polyornithine-laminin (P/L)-coated plates, in which they grow as monolayer cultures. NPC cultures were found to be devoid of astrocytes, microglia, and oligodendrocytes contaminants. Once cultures were established, NPCs from wild-type and SOD1(G93A) mice were used to generate astrocytes by withdrawing growth factors and supplementing the medium with 10% FBS. The media was changed every 2 days thereafter.

Astrocytes matured for 7 days prior to being used in the experiments described above. Highly enriched astrocyte cultures were obtained with no detectable levels of microglia, and oligodendrocytes.

2.2.5 CONVERSION OF SKIN FIBROBLASTS TO iNPCs AND DIFFERENTIATION INTO ASTROCYTES

Induced Neural Progenitor Cells (iNPCs) were derived from patient fibroblasts as previously described (Meyer et al., 2014). Briefly, a total of 10^4 fibroblasts were seeded into one well of a six-well plate. One day after seeding, a mixture consisting of the retroviral vectors Oct3/4, Sox2, Klf4, and c-Myc was

44 applied with a multiplicity of infection of 10 for each viral vector. The cells were incubated overnight in a final volume of 700 µL medium/viral vector. The following day, cells were washed 2x with PBS, and regular fibroblast medium

(DMEM plus 10% FBS) was applied for 3 days. After this recovery time, the cells were washed 1x with PBS before 2 mL of NPC conversion medium consisting of

DMEM/F12, 1% N2, 1% B27, 20 ng/mL FGF2, 20 ng/mL EGF, and heparin (5

μg/mL; Sigma-Aldrich) was added to the cells. This medium was changed every day thereafter. Once the cells change shape and form sphere-like structures, they were collected with a pipette and expanded in an individual well of a six-well plate previously coated with human fibronectin (5 μg/mL; Millipore). As soon as the NPC culture was established, the medium was switched to NPC medium consisting of DMEM/F12, 1% N2, 1% B27, and FGF2 (40 ng/mL) only. To differentiate into i-astrocytes, iNPCs were seeded in NPC medium at low density in a fibronectin-coated 10-cm dish. The day after, the medium was switched to

DMEM containing 10% FBS and 0.3% N2. Astrocytes then matured for at least 7 days, with growth medium being changed every 2 days.

2.2.6 LENTIVIRAL PRODUCTION

The wild-type HSPB1 viral vector was generated by digestion of M13-

HSPB1 plasmid (GeneCopoeia) with BstB1 and NotI, and ligated into the pcDNA4/TO vector (Invitrogen), in frame with an N-terminal FLAG sequence.

The HSPB1 cDNA was then PCR amplified to create a 5` BamH1 site and 3`

Xho1 site. This construct was then ligated into a pCSC-SP-PW (Addgene)

45 backbone. Site directed mutagenesis (Agilent) was used to generate

HSPB1(G84R), HSPB1(R136W), HSPB1(TriA) and HSPB1(TriD) vectors (see

Table 2.1 for primer sequences). Viral particles were produced as previously described (Tiscornia et al., 2006) and delivered to astrocytes with a multiplicity of infection of 20. Briefly, HEK293FT cells grown in IMDM supplemented with 10%

FBS (v/v) were transfected with VSV-G, REV, MDL and HSPB1 plasmids by calcium phosphate precipitation. 24 hours later, the transfection medium was removed and replaced with IMDM supplemented with 2% FBS (v/v). Viral supernatants were collected every 24 hours for 3 days, and subsequently centrifuged at 1000 RCF for 5 minutes and then filtered through a sterile 0.45 µM

PVDF membrane. Viral particles were then concentrated using ultracentrifugation and re-suspended in TNE buffer (50 mM Tris-HCl pH 7.4, 100 mM NaCl, 0.1 mM

EDTA). Astrocytes were cultured with viral particles for 48 hours prior to all experiments.

Table 2.1 HSPB1 construct primer sequences

PCR Primers HSPB1 BamH1 Up TTTTTTGGATCCATGACCGAGCGCCGC HSPB1 Xho1 Down TTTTTTCTCGAGTTACTTGGCGGCAGT

Site Directed Mutagenesis Primers Mutation Forward Primer Reverse Primer HSPB1 CGGCAACTCAGCAGCCG GGATCTCCGAGACCCGGCT G84R GGTCTCG GCTGA HSPB1 GCATGGCTACATCTCCTG CGTGTATTTCCGCGTGAAG R136W GTGCTTCA CACCA HSPB1 TGCGGGGCCCCGCCTGG AGGGGTCCCAGGCGGGGC S15A GACCCCT CCCGCA HSPB1 CTCCTGCGGGGCCCCGA CGGAAGGGGTCCCAGTCG S15D CTGGGACCCCT GGGCCCCGC HSPB1 AGCCGCGCGCTCGCCCG CTGAGTTGCCGGGCGAGC S78A GCAACTCAG GCGCGGCT HSPB1 GACCCGCGCGCTCGACC GCTGAGTTGCCGGTCGAGC S78D GGCAACTCAGC GCGCGGGTC HSPB1 CGCCCGGCAACTCGCCA AGACCCCGCTGGCGAGTTG S82A GCGGGGTCT CCGGGCG HSPB1 CAGCCGGCAACTCGACA CGAGACCCCGCTGTCGAGT S82D GCGGGGTCTCG TGCCGGCTG

2.2.7 ASTROCYTE-MOTOR NEURON CO-CULTURE ASSAY

Murine or human NPC-derived astrocytes were plated in 96-well plates coated with human fibronectin (2.5 μg/mL; Millipore) at a density of 35,000 cells per well. One day later, FACS-sorted GFP-positive MNs were re-suspended in

MN media consisting of DMEM/F12, 5% horse serum, 2% N2, 2% B27 plus

GDNF (Invitrogen; 10 ng/mL), BDNF (Invitrogen; 10 ng/ mL), and CNTF

(Invitrogen; 10 ng/mL) and added to the astrocytes at a density of 10,000 per well. Each plate was scanned every day for 6 days with the fully automated IN

CELL 6000 confocal plate reader to capture GFP-positive cells. The IN CELL developer and analyzer software were used to create whole-well pictures and to automatically count MNs.

2.2.8 ASTROCYTE VIABILITY ASSAY

Murine NPC-derived astrocytes were plated in 10-cm2 plates coated with human fibronectin (2.5 µg/mL; Millipore) at a density of 6X10^6 cells per dish.

Media was changed every 2 days. After 4 days, cells were washed 2X with PBS and incubated with Accutase (Gibco) to lift cells from the culture dish. Accutase was neutralized with culture media, and cells were centrifuged for 5 minutes at

500 RCF. The resulting cell pellets were re-suspended in 1 mL of 1X PBS and viability was determined using an NC-100 Nucleo-counter (Chemometec), according to the manufacturer’s protocol.

2.2.9 WESTERN BLOTTING

Astrocytes were lysed in a buffer containing 20 mM Tris (pH 7.4), 150 mM

NaCl, 2 mM EDTA, 1 mM DTT, 10% Glycerol, 1% Triton X-100, 2 mM sodium pyrophosphate, 25 mM b-glycerophosphate, 1 mM Na3VO4, 10 mM NaF, 10

μg/mL leupeptin, 10 μg/mL aprotinin, and 5 mM PMSF on ice for 5 minutes.

Lysates were centrifuged (16,000 RCF) at 4°C for 10 minutes. The resulting supernatant was then boiled at 95°C for 5 minutes in 2X Laemmli Buffer (20%

Glycerol, 2% b-Mercaptoethanol (v/v), 100 mM Tris-HCl pH 6.8, 2% SDS (v/v)).

10 µg of each sample was separated by SDS-PAGE and transferred to a PVDF

(Bio-Rad) membrane. Immunoblots were blocked in 5% nonfat dry milk in Tris- buffered saline (TBS) for 1 hour at room temperature. The blots were then incubated with primary antibodies against human HSPB1, mHSPB1, SOD1, misfolded SOD1 or Tubulin for 1 hour at room temperature. The antibody against misfolded SOD1 has been previously described to be specific to human SOD1 under denaturing conditions and specific for mutant SOD1 isoforms under non- denaturing conditions (Gros-Louis et al., 2010; Israelson et al., 2015; Leyton-

Jaimes et al., 2016; Nizzardo et al., 2016; Patel et al., 2015). Blots were then washed 3 times with TBS-T, and incubated with secondary antibody for 1 hour at room temperature, and then scanned using a Licor Odyssey Classic. Primary antibodies used for these experiments were: Anti-HSPB1 (Ab Cam, ab2790

[1:5,000]), Anti-Hsp25/27 (Millipore MAB3842 [1:1,000]), Anti-SOD1 (Santa Cruz sc-8637 [1:250]), Anti-Misfolded SOD1 (MediMabs, MM-0070 [1:1,000]), and

Anti-Tubulin (Ab Cam, ab7291 [1:10,000]). Secondary antibodies used were:

Anti-Mouse 800CW (Licor, 925-32210 [1:10,000]) All antibodies were diluted in blocking buffer + 0.1% Tween-20.

2.2.10 IMMUNOFLUORESCENCE

Cells were fixed with 4% paraformaldehyde for 15 min and washed 3 49 times with Phosphate-buffered saline (PBS) before the blocking solution consisting of PBS with 10% goat serum, 3% BSA, and 0.1% Triton X-100 was applied for 1 hour. Incubation with primary antibody was performed for 1 hour at room temperature. Cells were then washed 3 times in PBS + 0.1% Triton X-100 then incubated with secondary antibody and DAPI for 1 hour at room temperature. Coverslips were then washed 3 times with PBS + 0.1% Triton X-

100. Coverslips were mounted to glass slides using Fluoromount-G

(SouthernBiotech) and imaged on a Andor spinning disk confocal microscope.

Antibodies used for these experiments: Anti-GFAP (Ab Cam, ab4676 [1:250]),

Anti-Hsp25 (ENZO, ADI-SPA-801 [1:1000]), Anti-Misfolded SOD1 (MediMabs,

MM-0070 [1:250]), Anti-Mouse AlexaFluor 488 (Invitrogen, A-11001 [1:1,000]),

Anti-Rabbit AlexaFluor 488 (Invitrogen, A-11034 [1:1,000]) and Anti-Chicken

AlexaFluor 594 (Invitrogen, A-11042 [1:1,000]). All antibodies were diluted in blocking solution.

2.2.11 HSPB1 ELISA

HSPB1 levels were quantified using an enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s protocol (Ab Cam ab193757).

Briefly, recombinant HSPB1 protein, and SOD1(G93A) + HSPB1(WT) astrocyte cell lysate were resuspended in 1x cell extraction buffer PTR. 10-fold serial dilutions of the recombinant HSPB1 protein were made to generate a standard curve. SOD1(G93A) + HSPB1(WT) lysate was determined by Bradford assay and diluted to 7.5 µg/mL for use as a positive control. 50 µL of recombinant 50 protein, cell lysate or cultured medium from WT, SOD1(G93A) and SOD1(G93A)

+ HSPB1(WT) was pipetted into a 96-well plate pre-coated with anti-HSPB1 antibodies. 50 µL of antibody cocktail was added to each well and incubated for 1 hour at room temperature on a shaker at 250 rpm. The 96-well plate was then washed 3x with 350 µL of 1X wash buffer PT. 100 µL of TMB substrate was added to each well and incubated in the dark for 10 minutes at room temperature on a plate shaker at 250 rpm. 100 µL of Stop Solution was then added to each well and further incubated for 1 minute at room temperature on a plate shaker at

250 rpm. Optical density was then measured at 450 nm on a Tecan F200 microplate reader. OD’s from the recombinant protein were used to generate a standard curve. Unknown samples were compared to the standard curve to obtain absolute HSPB1 protein levels.

2.2.12 PROTEIN SECRETION ASSAY

Murine NPC-derived astrocytes were plated in 10-cm2 plates coated with human fibronectin (2.5 µg/mL; Millipore) at a density of 6X10^6 cells per dish.

Media was changed every 2 days. After 4 days, cells were washed 3X with PBS and placed in serum-free media. After 24 hours, the conditioned medium was removed from cells and secreted proteins were isolated as previously described

(Basso et al., 2013). Briefly, conditioned media was centrifuged at 12,000 RCF for 10 minutes to remove cellular debris. The supernatant was then mixed with 4 volumes of acetone and placed on a nutator for 2 hours at 4°C. Precipitated proteins were pelleted by centrifugation at 15,000 RCF for 15 minutes at 4°C. 51

Protein samples were washed in 1 mL of lysis buffer (see 2.2.9) and re- suspended in 200 µL of lysis buffer. 15 mg of each sample were analyzed for

SOD1 and HSPB1 expression by western blotting (see 2.2.9).

2.2.13 INFLAMMATORY GENE MICROARRAY

SOD1(G93A), SOD1(G93A) + HSPB1(WT) and SOD1(G93A) +

HSPB1(R136W) astrocytes were plated in 96-well plates coated with human fibronectin (2.5 μg/mL; Millipore) at a density of 35,000 cells per well. One day later, FACS-sorted GFP-positive MNs were re-suspended in MN media consisting of DMEM/F12, 5% horse serum, 2% N2, 2% B27 plus GDNF

(Invitrogen; 10 ng/mL), BDNF (Invitrogen; 10 ng/ mL), and CNTF (Invitrogen; 10 ng/mL) and added to the astrocytes at a density of 10,000 per well. Five days later, RNA was harvested using the RT2 q-PCR grade RNA isolation kit

(SABiosciences, Frederick, MD) and total RNA was reverse transcribed with RT2

First Strand Kit (SABiosciences) according to the manufacturer’s instructions.

Inflammatory gene expression was measured in astrocytes co-cultured with MNs using a cataloged RT2 Profiler PCR Array for murine Inflammatory Cytokines and

Receptors (SA Biosciences). Fold change analysis was performed using the

DDCT method (Schmittgen and Livak, 2008). Fold changes less than one are reported as the negative inverse value (1/fold change value).

2.2.14 STATISTICS

Statistical analysis was performed by one-way ANOVA unpaired t-test for

52 mean differences between the average of wild-type control lines versus

SOD1(G93A)-derived lines and/or HSPB1 mutant lines (GraphPad Prizm

Software). All experiments were performed at least in triplicate.

Disclaimer: The work presented in this chapter is part of a published manuscript, and was done in collaboration with Dr. Brian K. Kaspar’s lab. All motor neuron co-culture assays were performed by Dr. Sungwon Song and Dr.

Kathrin Meyer. Patient derived i-Astrocytes were generated by Dr. Kathrin Meyer.

The inflammatory gene panel was run by Dr. Sungwon Song. Data analysis and creation of figures was performed by Patrick Heilman.

2.3 RESULTS

2.3.1 NEURAL PROGENITOR CELL-DERIVED SOD1 (G93A) ASTROCYTES UP-REGULATE ENDOGENOUS HSPB1

To determine whether endogenous Hsp25 (murine HSPB1, herein referred to as mHSPB1) protein expression is increased in neural progenitor cell

(NPC) derived astrocytes, I prepared total cellular lysates from both wild-type and

SOD1(G93A) astrocytes. mHSPB1 protein levels were measured by semi- quantitative western blot using an antibody specific to murine HSPB1 (Figure

2.1C). Endogenous mHSPB1 levels were enriched in lysates from SOD1(G93A) astrocytes compared to wild-type astrocytes. Immunofluorescent staining of mHSPB1 similarly showed upregulation in SOD1(G93A) astrocytes. 13.25 ±

3.05% of wild-type astrocytes were positive for mHSPB1, whereas 37.40 ±

3.18% of SOD1(G93A) astrocytes were positive for mHSPB1 (Figure 2.1A,

2.1B). Co-staining, with antibodies specific to misfolded SOD1 (Gros-Louis et al.,

2010; Israelson et al., 2015; Leyton-Jaimes et al., 2016; Nizzardo et al., 2016;

Patel et al., 2015), revealed that SOD1 and mHSPB1 are visualized in the cytoplasm, and that 41.03 ± 3.25% of mHSPB1 positive astrocytes were also positive for misfolded SOD1.

Our group, and others, have shown that SOD1(G93A) astrocytes exhibit non-cell autonomous toxicity to motor neurons (MNs) in vitro (Dodge et al., 2008;

Fritz et al., 2013; Meyer et al., 2014). To determine if mHSPB1 expression in astrocytes affects MN survival, we used lentiviral-mediated knockdown to reduce mHSPB1 protein levels by approximately 50% in SOD1(G93A) astrocytes

(Figure 2.1C, 2.1D). These astrocytes were then co-cultured with wild-type MNs

(Figure 2.1E, 2.1F). There was no observable difference in survival at 6 days between MNs cultured with shHspb1 astrocytes, 36.44 ± 5.86%, or a scrambled control shRNA, 38.07 ± 3.75%, versus non-transduced SOD1(G93A) astrocytes,

32.81 ± 5.57%, suggesting that reduced levels of mHSPB1 protein do not affect

SOD1(G93A) astrocyte-mediated MN toxicity.

Figure 2.1 Knockdown of murine HSPB1 in SOD1(G93A) astrocytes has no effect on motor neuron survival. A. Immunofluorescent staining of murine HSPB1 (mHSPB1) and misfolded SOD1 in wild-type and SOD1(G93A) mouse astrocytes. Nuclei were visualized with DAPI. B. Percentage of wild- type or SOD1(G93A) astrocytes stained positive for mHSPB1. Cells were 55 counted in 5 random fields of view and expressed as the number of mHSPB1 positive cells over total cells. C. Representative immunoblot of mHSPB1 protein levels in wild-type and SOD1(G93A) mouse astrocytes (left panel) and SOD1(G93A) astrocytes infected with lentiviral shRNA against Hspb1. D. Quantification of mHSPB1 immuno-staining in SOD1(G93A) astrocytes and SOD1(G93A) astrocytes transduced with shHspb1. mHSPB1 signal intensity was normalized to mHSPB1 signal intensity from SOD1(G93A) astrocyte lysates. Error bars denote s.e.m. E. MN survival at day 6 of MN co-culture assay with wild-type, SOD1(G93A), SOD1(G93A) + shHspb1 RNA or SOD1(G93A) + shSCRM RNA astrocytes. MN survival was normalized to counts from MNs cultured with wild-type astrocytes (N = 3 for all groups, each n was run in triplicate). Error bars denote s.e.m. ****P<0.0001, ns, non-significant. F. Representative images of HB9-GFP expressing MNs (shown in black) after 6 days in co-culture with wild-type astrocytes, SOD1(G93A) astrocytes, and SOD1(G93A) astrocytes transduced with shSCRM or shHspb1.

2.3.2 IN VITRO NON-CELL AUTONOMOUS NEUROPROTECTION FROM HUMAN HSPB1

Next, I determined whether human HSPB1(WT) overexpression in

SOD1(G93A) astrocytes could improve MN survival in an in vitro co-culture model. Using lentiviral transduction, our collaborators overexpressed a FLAG- tagged HSPB1(WT) construct in SOD1(G93A) astrocytes, which resulted in a

100% increase in MN survival compared to non-transduced SOD1(G93A) astrocytes, 26.02 ± 2.30% vs. 6.99 ± 0.57% respectively (P<0.0005) (Figure

2.2A). Previous in vitro studies found that HSPB1 expression can influence neurite maintenance and growth (Williams and Mearow, 2011; Williams et al.,

2006). Consistent with previous reports, we also found that motor neurons cultured on SOD1(G93A) astrocytes exhibit reduced soma size and neurite

56 length compared to motor neurons cultured on wild-type astrocytes (Figure 2.2B and 2.2C). Interestingly, the soma size of MNs cultured on SOD1(G93A) astrocytes overexpressing HSPB1(WT) was indistinguishable from astrocytes bearing the SOD1(G93A) mutation alone, but neurite length was increased by

75% when MNs were cultured in contact with SOD1(G93A) astrocytes overexpressing HSPB1(WT) (Figure 2.2B and 2.2C).

Figure 2.2 Overexpression of HSPB1(WT) in SOD1(G93A) astrocytes improves MN survival. A. MN survival at day 6 of MN co-culture assay with wild-type, SOD1(G93A), and SOD1(G93A) + HSPB1(WT) astrocytes. MN survival was normalized to counts from MNs cultured with wild-type astrocytes. (n = 5 for all groups, each n was run in triplicate.) B. Length of MN neurites in µm at day 6 of MN co-culture experiments. C. Measurement of soma size in μm2 at day 6 of MN co-culture experiments. (B. C. N = 100 motor neurons for all groups. Wild-type astrocytes = green line, SOD1(G93A) astrocytes = red line, SOD1(G93A) + HSPB1(WT) astrocytes = blue line.) Error bars denote s.e.m. ***P<0.0005, ****P<0.0001, ns, non- significant.

Previous studies have demonstrated that overexpression of SOD1(G93A) in primary cortical astrocyte cultures was toxic to motor neurons, but did not 57 affect the viability of the astrocytes (Benkler et al., 2013; Tortarolo et al., 2004).

To confirm this in my experiments, I compared the viability of wild-type astrocytes to non-transduced SOD1(G93A) astrocytes and SOD1(G93A) astrocytes transduced with HSPB1(WT) (Figure 2.3). I observed no difference in the viability between cell lines, with 97.13 ± 0.80% viable wild-type astrocytes and

96.74 ± 0.98% viable SOD1(G93A) astrocytes at day 9 in culture. Similarly, transduction of SOD1(G93A) astrocytes with HSPB1(WT) had no effect on astrocyte viability compared to a mock transduction, 92.22 ± 1.57% vs. 92.31 ±

0.68% respectively, suggesting that the observed increase in motor neuron survival in the co-culture assay is not from an alteration in astrocyte viability.

Figure 2.3 Expression of SOD1(G93A) has no effect on astrocyte viability. Percentage of viable cells from wild-type, SOD1(G93A), SOD1(G93A) + PBS and SOD1(G93A) + HSPB1(WT) astrocytes after 9 days in culture. Viability was determined using an NC-100 nucleo-counter (n = 2 for all groups, each n was run in duplicate.) Error bars denote s.e.m., n.s., non-significant.

I addressed the possibility that HSPB1 increases MN survival by altering

SOD1 localization and expression in astrocytes. I examined the immunohistochemical staining pattern for HSPB1 and misfolded SOD1 in both

SOD1(G93A) and SOD1(G93A) + HSPB1(WT) astrocyte lines. Staining for misfolded SOD1 was indistinguishable between the two cell lines (Figure 2.4A).

To examine SOD1 protein levels in a more quantitative fashion, I compared

SOD1 and misfolded-SOD1 protein expression in total cellular lysates from wild- type, SOD1(G93A) and SOD1(G93A) + HSPB1(WT) astrocytes by western blot.

There was no difference in the amount of misfolded SOD1, or SOD1 protein in astrocytes overexpressing HSPB1(WT) (Figure 2.4B, 2.4C), suggesting that overexpression of HSPB1 does not alter SOD1 protein levels.

Figure 2.4 Overexpression of HSPB1(WT) does not reduce SOD1(G93A) expression. A. Immunofluorescenct staining of FLAG and, SOD1(G93A) in non-transduced SOD1(G93A) astrocytes and, HSPB1(WT) transduced SOD1(G93A) astrocytes. Nuclei were visualized with DAPI. B. Representative immunoblot of HSPB1, SOD1 and misfolded SOD1 protein levels in wild-type, SOD1(G93A) and SOD1(G93A) + HSPB1(WT) cell lines. C. Quantification of misfolded human SOD1 staining in SOD1(G93A) and SOD1(G93A) + HSPB1(WT) cell lines. Band intensities were quantified and normalized to a tubulin loading control. Data represents the average of 3 independent western blots. Error bars denote (s.e.m.) n.s., not significant.

Published literature has demonstrated that SOD1 can be secreted by numerous cell types, including neuronal and glial cells (Basso et al., 2013; Lafon-

Cazal et al., 2003; Turner et al., 2005; Urushitani et al., 2006) In particular, Basso et al. (2013) provided evidence that astrocyte secreted SOD1 (wild-type or the

G93A variant) can be taken up by neurons, and secreted SOD1(G93A) is selectively toxic to motor neurons. I prepared protein samples from SOD1

(G93A) and SOD1 (G93A) + HSPB1 (WT) astrocyte-conditioned medium and examined the levels of secreted SOD1 by semi-quantitative western blotting

(Figure 2.5A). My data showed no difference in the amount of SOD1 secreted by cells expressing HSPB1(WT) compared to non-transduced astrocytes.

HSPs, including HSPB1, can also be secreted by astrocytes under both normal and stress conditions (Burut et al., 2010; Hecker and McGarvey, 2011;

Nafar et al., 2016). Extracellular HSPB1 can induce the release of cytokines and chemokines in macrophages (De et al., 2000; Jin et al., 2014; Salari et al., 2013), and has been proposed to protect mice from atherosclerosis (Rayner et al., 2008;

Rayner et al., 2010). I tested the possibility that HSPB1 secretion by transduced

SOD1(G93A) astrocytes may promote MN survival. Using an ELISA specific to human HSPB1, I determined the protein levels of HSPB1 in the medium from our co-culture assays (Figure 2.5B). I was unable to detect HSPB1 in media from

WT, SOD1 (G93A) or SOD1 (G93A) + HSPB1 (WT) astrocyte cultures,

62 suggesting that in this system astrocytes either do not release HSPB1 into the extracellular space or the amount of HSPB1 secreted is below the detection limit of the assays used here.

Figure 2.5 Detection of SOD1 and HSPB1 in astrocyte conditioned medium. A. Immunoblot of secreted proteins from SOD1(G93A) and SOD1(G93A) + HSPB1(WT) astrocytes. 40, 20, 10 and 5 µg of each sample were separated by SDS-PAGE. Immunoblots were incubated with antibodies against human HSPB1 and SOD1. B. Wild-type, SOD1(G93A) and SOD1(G93A) + HSPB1(WT) astrocytes were co-cultured with wild-type motor neurons for 6 days. The cultured medium was collected and analyzed for the presence of HSPB1 protein by ELISA. Absolute protein quantities were determined by comparing unknown samples to a standard curve. Lysate from 63

SOD1(G93A) astrocytes transduced with HSPB1(WT) lentivirus was ran as a positive control. (n = 2. Each n was run in triplicate).

2.3.3 NON-CELL AUTONOMOUS NEUROPROTECTION MAY BE PHOSPHORYLATION DEPENDENT

HSPB1 is a phosphoprotein, with 3 phosphorylation sites located in the N- terminal region (Figure 1.11). Previous studies have shown that the phosphorylation state of HSPB1 is critical to its neuroprotective properties (Benn et al., 2002; Stetler et al., 2012; Williams and Mearow, 2011). To determine whether phosphorylation was required for neuroprotection in this assay,

SOD1(G93A) astrocytes were transduced with one of two mutHSPB1 constructs, in which the 3 major phosphorylation sites had been mutated to alanine or aspartic acid respectively, to mimic a non-phosphorylated HSPB1, HSPB1(TriA), or a constitutively phosphorylated protein, HSPB1(TriD). When astrocytes expressing HSPB1(TriA) are co-cultured with MNs (Figure 2.6A), I observed no difference in MN survival as compared to non-transduced astrocytes, 50.48 ±

3.25% and 46.02 ± 1.84% respectively. In contrast, overexpression of

HSPB1(TriD) in astrocytes resulted in increased MN survival compared to non- transduced SOD1(G93A) astrocytes (Figure 2.6A), 63.01 ± 4.73% and 46.01 ±

1.84% respectively (P<0.001). Expression of HSPB1 transgenes was confirmed by western blot (Figure 2.6B). These data support the hypothesis that

64 phosphorylation of HSPB1 may be a necessary condition for its observed non- cell autonomous neuroprotection.

Figure 2.6 Phosphomimetic HSPB1 expression in SOD1(G93A) astrocytes protects MNs from astrocyte-mediated toxicity. A. MN survival at day 6 of MN co-culture assay with wild-type, SOD1(G93A), SOD1(G93A) + HSPB1(TriA) or SOD1(G93A) + HSPB1(TriD) astrocytes. MN survival was normalized to counts from MNs cultured with wild-type astrocytes. (n = 3 for all groups, each n was run in triplicate.) Error bars denote s.e.m. **P<0.001, ****P<0.0001, n.s., non-significant. B. Representative immunoblot confirming the expression of HSPB1(TriA) and HSPB1(TriD) in SOD1(G93A) astrocytes using an antibody specific to human HSPB1.

2.3.4 MUTATIONS IN HSPB1 DISRUPT ITS NEUROPROTECTIVE PROPERTIES IN VITRO

Since mutations in HSPB1 result in motor neuropathy and recent studies also indicate a potential role in sporadic ALS, I examined the consequences of mutHSPB1 overexpression on motor neuron survival in the SOD1(G93A) co- culture system. I selected two mutations to study, a glycine to arginine mutation near the phosphorylation sites in the N-terminal region, HSPB1(G84R) and an

65 arginine to tryptophan mutation in the α-crystallin domain of the protein,

HSPB1(R136W), (Figure 1.11) and overexpressed FLAG-tagged versions of these isoforms in astrocytes via lentiviral transduction. These mutations were selected for their divergent effects on HSPB1 chaperone activity, with

HSPB1(G84R) exhibiting reduced chaperone activity (Nefedova et al., 2013b) and HSPB1(R136W) exhibiting increased chaperone activity in vitro (Almeida-

Souza et al., 2010). In contrast to the overexpression of wild-type HSPB1 in

SOD1(G93A) astrocytes, overexpression of HSPB1(G84R) failed to alter MN survival compared to non-transduced astrocytes, 41.50 ± 3.19% vs. 34.66 ±

2.21%, (P>0.1), while overexpression of HSPB1(R136W) resulted in only a mild increase in MN survival, 46.38 ± 3.56%, vs. 34.66 ± 2.21%, (P<0.01) (Figure

2.7A, 2.7B, 2.7C), indicating that both mutations interfere with the protective effect of HSPB1 in this co-culture system.

To determine whether expression of the mutant HSPB1 isoforms in astrocytes had a negative effect on motor neurons, mutHSPB1 was overexpressed in wild-type astrocytes and MN survival was recorded. Wild-type astrocyte lines overexpressing RFP, HSPB1(WT), HSPB1(G84R) or

HSPB1(R136W) were generated and co-cultured with wild-type MNs. Expression of RFP has previously been shown to have no effect on MN survival (Frakes et al., 2014), and was used as a control for toxicity associated with infecting astrocytes with a viral vector. Similar to RFP-expressing astrocytes,

66 overexpression of HSPB1(WT) or mutHSPB1 constructs resulted in an ~10% loss of MN survival compared to non-transduced astrocytes (Figure 2.7D, 2.7E).

These results indicated that specific overexpression of mutHSPB1 in wild-type astrocytes had little effect on motor neuron survival in co-culture.

Figure 2.7 Mutations in HSPB1 attenuate non-cell autonomous motor neuron protection. A. MN survival at day 6 of MN co-culture assay with wild-type, SOD1(G93A), SOD1(G93A) + HSPB1(WT), SOD1(G93A) + HSPB1(G84R), SOD1(G93A) + HSPB1(R136W) astrocytes. MN survival was normalized to counts from MNs cultured with wild-type astrocytes. (n = 3 for all groups, each n was run in triplicate.) B. Representative immunoblot confirming the expression of human HSPB1 constructs in SOD1(G93A) astrocytes. Western blots were probed with an antibody specific to human HSPB1 C. Representative images of HB9-GFP expressing MNs (shown in black) after 6 days in co-culture with wild-type astrocytes, SOD1(G93A) astrocytes, and SOD1(G93A) astrocytes transduced with HSPB1(WT), HSPB1(G84R) or HSPB1(R136W). D. MN survival at day 6 of MN co-culture assay with wild-type, WT+RFP, WT + HSPB1 (WT), WT + HSPB1(G84R), WT 68

+ HSPB1(R136W) astrocytes. MN survival was normalized to counts from MNs cultured with wild-type astrocytes. (n = 3 for all groups, each n was run in triplicate.) E. Representative immunoblot confirming the expression of HSPB1 constructs in wild-type astrocytes using an antibody specific to human HSPB1. Error bars denote s.e.m. ***P<0.0005, *P<0.01, n.s., non- significant.

In conjunction with microglia, astrocytes play a central role in neuroinflammation, a pathological hallmark of ALS. Recent studies have demonstrated that astrocytes expressing SOD1(G93A) display an increased inflammatory state, and secrete a higher number of inflammatory cytokines than non-transgenic astrocytes (Hensley et al., 2006). It has been suggested that chronic activation of inflammation secretory pathways may contribute to non-cell autonomous neurodegeneration. I examined the possibility that overexpression of HSPB1(WT), a known regulator of inflammatory signaling, may reduce the inflammatory state of SOD1(G93A) astrocytes as a mechanism to promote motor neuron survival in vitro. Using an inflammatory gene microarray, our collaborators determined the expression of a select panel of cytokines, chemokines and other inflammatory genes in SOD1(G93A), SOD1(G93A) +

HSPB1(WT) and SOD1(G93A) + HSPB1(R136W) astrocytes (Figure 2.8). I found that HSPB1(WT) reduced the expression of a number of pro-inflammatory genes, and increased the expression of some select anti-inflammatory genes.

Some genes, such as the macrophage inhibitory factor (MIF) have been

69 previously reported to play a beneficial role in SOD1-ALS models (Israelson et al., 2015; Leyton-Jaimes et al., 2016). Future work will be required to validate these data, and determine the overall impact these changes in gene expression have on motor neuron survival.

Figure 2.8 Overexpression of HSPB1 constructs alters SOD1(G93A) astrocyte inflammatory profile. Inflammatory gene profile from SOD1(G93A), SOD1(G93A) + HSPB1(WT) and SOD1(G93A) + HSPB1(R136W) astrocytes. mRNA expression was determined using an RT2-Profiler Array and fold change analysis was performed using the – DDCT method. Gene expression in SOD1(G93A) astrocytes was set to a value of 1, and compared to gene expression values in SOD1(G93A) + HSPB1(WT) and SOD1(G93A) + HSPB1(R136W) cell lines. Blue indicates a decrease in gene expression, while red indicates an increase in gene expression. Undetected transcripts are shown in grey.

2.3.5 HSPB1 OVEREXPRESSION IN PATIENT-DERIVED ASTROCYTES

Our collaborators recently reported a technique for the reprogramming of adult human fibroblasts to induced neural progenitor stem cells (iNPCs) (Meyer et al., 2014). These iNPCs can be differentiated into astrocytes (i-astrocytes) and subsequently used in co-culture assays. I sought to determine if i-astrocytes from

ALS patients exhibit increased levels of HSPB1 expression. Using semi- quantitative western blotting, I compared HSPB1 protein levels in i-astrocytes from patients with sALS or fALS to non-ALS control i-astrocytes (Figure 2.9A,

2.9B). The data showed that HSPB1 protein levels vary from patient to patient, and unlike my murine data, HSPB1 is not upregulated in ALS patient astrocytes.

Interestingly, i-astrocytes from a fALS patient carrying the SOD1(D91A) mutation had very low levels of HSPB1 expression (Figure 2.9A, patient 295).

Figure 2.9 HSPB1 expression levels in ALS patient i-astrocytes. HSPB1 protein levels were quantified in healthy and ALS patient derived i- Astrocytes. Total cellular lysates were prepared from i-Astrocyte cell cultures and separated by gel electrophoresis for western blot analysis. A. Representative immunoblot of HSPB1 expression in ALS patient i- astrocytes. Immunoblots were stained with a-HSPB1 and a-Tubulin antibodies. B. Quantification of HSPB1 protein levels detected in immunoblot from A. HSPB1 levels were normalized to tubulin expression levels. HSPB1 protein levels varied from patient to patient, with no significant difference in protein expression between ALS patients and healthy controls. One patient carrying an SOD1 mutation, 295, had

72 markedly low levels of HSPB1, and was used in subsequent astrocyte- motor neuron co-culture experiments.

Our group previously demonstrated that astrocytes derived from ALS patient fibroblast cell lines are toxic to motor neurons in vitro (Meyer et al., 2014).

We investigated if overexpression of HSPB1(WT) in i-astrocytes could improve motor neuron survival in co-culture, similar to my observations in murine astrocytes (Figure 2.2A, Figure 2.7A). Overexpression of the HSPB1(WT) vector had no effect on MN survival in control i-astrocytes, and in i-astrocytes derived from an sALS fibroblast line (Figure 2.10A, 2.10B). However, we did observe a significant increase in MN survival when HSPB1(WT) was overexpressed in the low-expressing HSPB1 fALS cell line (Figure 2.10A,

2.10B).

Figure 2.10 Overexpression of HSPB1(WT) in fALS i-astrocytes improves MN survival. A. Representative immunoblot confirming transduction of i- astrocytes with HSPB1(WT) lenti virus. B. MN survival at day 4 of MN co- culture assay with sALS i-astrocytes, fALS i-astrocytes or healthy control i- astrocytes. MN survival was normalized to counts from MNs cultured with control i-astrocytes. (n = 2 for all groups, each n was run in triplicate.) Error bars denote s.e.m. *P<0.01.

2.3.6 HSPB1-dHMN PATIENT-DERIVED ASTROCYTES ARE NOT TOXIC IN VITRO

Fibroblast samples were obtained from a patient carrying the

HSPB1(G84R) mutation by skin biopsy and converted into i-astrocytes. Western blot analysis of fibroblast and i-astrocyte cell-cultures revealed no difference in

HSPB1 expression between healthy control and mutHSPB1 patient samples

(Figure 2.11B, Figure 2.12B). I also examined the immunostaining pattern of

HSPB1 in fibroblasts and i-astrocytes (Figure 2.11A, 2.12A). While there was no distinguishable difference in the staining of fibroblasts from healthy controls compared to the mutHSPB1 patient cells, a portion of the mutHSPB1 i-Astrocytes demonstrate an increased fluorescent intensity compared healthy controls.

However, it is unclear if this different staining pattern is due to protein aggregation or some other factor. We next investigated if mutHSPB1 patient samples are toxic to motor neurons in co-culture (Figure 2.12C). After 4 days,

MNs co-cultured with mutHSPB1 expressing i-astrocytes showed no difference compared to MNs co-cultured with control i-astrocytes, suggesting that astrocytic expression of mutHSPB1 alone is not sufficient to induce MN toxicity in vitro.

Figure 2.11 Expression and localization of HSPB1 in patient fibroblasts. A. Immunofluorescent staining of patient fibroblast samples with a-HSPB1 antibodies (Green). Nuclei were stained with DAPI (Blue). Images were captured at 60X magnification. B. Representative immunoblot of HSPB1 levels in patient fibroblast cellular lysates.

Figure 2.12 dHMN-patient derived i-astrocytes are not toxic to MNs in co- culture. A. Immunofluorescent staining of i-astrocytes with a-HSPB1 antibodies (Green). Nuclei were stained with DAPI (Blue). Images were captured at 20X magnification. B. Representative immunoblot of HSPB1 levels in i-astrocytes. C. MN survival at day 4 of MN co-culture assay with sALS i-astrocytes, dHMN i-astrocytes or healthy control i-astrocytes. MN survival was normalized to counts from MNs cultured with control i- astrocytes. 170 = healthy control, 105 = HSPB1(G84R) dHMN, 009 = sALS. (n = 2 for all groups, each n was run in triplicate.) Error bars denote s.e.m.

2.4 DISCUSSION

The molecular consequences of mutations in HSPB1 that result in the development of axonal neuropathies is interesting because of what it teaches us about the unique role that HSPB1 plays in the maintenance of motor neurons.

Previously characterized mouse models suggest a cell autonomous mechanism(s); neuronal overexpression of HSPB1 variants is sufficient to cause axonal neuropathy phenotypes, like the phenotype observed in dHMNII and

CMT2 patients (d'Ydewalle et al., 2011; Srivastava et al., 2012). Boughy et al. recently characterized two new transgenic mouse models of HSPB1 neuropathy that used the Cre-loxP system to express HSPB1 transgenes at physiological levels specifically in neurons (Bouhy et al., 2016). These mice failed to develop a phenotype, and one interpretation of this result and others, is that glial cell types may be involved in the development of axonal neuropathies.

Multiple studies have shed light on the roles of astrocytes, microglia and oliogodendrocytes in the progression of neurodegenerative disorders (Boillee et al., 2006; Ralph et al., 2005; Yamanaka et al., 2008a; Yamanaka et al., 2008b).

With help from our collaborators, I investigated the potential non-cell autonomous role for HSPB1, using an established astrocyte-motor neuron co-culture system.

These results demonstrate that expression of HSPB1(WT) in SOD1(G93A) astrocytes improves wild-type motor neuron survival in vitro and that expression of two different HSPB1 variants, HSPB1(G84R) and HSPB1(R136W), attenuate this effect. Furthermore, wild-type astrocytes expressing mutant HSPB1 were not found to be toxic to motor neurons. These data support the idea that neuropathy- associated mutations in HSPB1 may result in a loss of a supportive function in glial cells. I propose that this loss of non-cell autonomous function is paired with a toxic cell autonomous function in motor neurons that results in the axonal motor neuropathy.

HSPB1’s role in ALS has been studied almost exclusively in mouse cell culture or animal model systems. Using i-astrocyte cultures from sALS and fALS patients, I examined the expression pattern of HSPB1 and found no difference in protein levels between most healthy control and ALS cell lines. Interestingly, one fALS patient carrying the SOD1(D91A) mutation had drastically reduced HSPB1 expression. It is unclear if the lack of HSPB1 expression is due specifically to the presence of mutant SOD1, or some other genetic factor in this individual. Based on my murine cell culture data, I hypothesized that overexpression of

HSPB1(WT) in ALS patient-derived astrocytes would improve MN survival. We observed improved MN survival in the SOD1(D91A) patient cell line, but not in an sALS patient line. While a larger data set is needed, this data suggests that

HSPB1 may only be neuroprotective in mutant SOD1 ALS model systems.

Future co-culture studies using a battery of patient cell lines harboring various

ALS-linked genetic mutations, including SOD1 variants, will be crucial in determining the specificity of HSPB1 neuroprotection.

While the neuroprotective effect of HSPB1 in ALS has been studied, there is conflicting literature regarding the mechanistic basis for such protection. The function of HSPB1 is regulated by post-translational modifications and structural oligomerization. Chaperone activity of HSPB1 is associated with large, unphosphorylated oligomeric complexes (Rogalla et al., 1999). Phosphorylation of HSPB1 results in a shift from larger oligomeric species to smaller multimers, which can directly regulate cellular apoptotic pathways to promote cell survival

(Kim et al., 2013; Rogalla et al., 1999; Shen et al., 2016; Stetler et al., 2008;

Stetler et al., 2012). These smaller multimers also regulate a host of other cellular processes, including inflammatory response, mRNA decay and cytoskeleton stabilization (Knapinska et al., 2011; Lavoie et al., 1995; Liu et al.,

2010; Parcellier et al., 2003; Pivovarova et al., 2007; Sinsimer et al., 2008). My data implies that the smaller HSPB1 species are predominantly responsible for the observed neuroprotection in this SOD1-ALS model system. I report that overexpression of HSPB1(WT) or HSPB1(TriD), a phosphomimetic mutant, improved motor neuron survival, while a non-phosphorylatable HSPB1 mutant,

HSPB1(TriA), did not. Interestingly, the HSPB1(TriD) construct only demonstrated partial neuroprotection compared to the HSPB1(WT) construct.

One reason that HSPB1(TriD) may have partial neuroprotection, is that this mutant form disrupts the ability to fluctuate between phosphorylated, partially phosphorylated and dephosphorylated states, which may reduce the full range of interactions between HSPB1 and client proteins that are necessary for robust neuroprotection.

These findings agree with a recent study that showed expression of

HSPB1(WT) and HSPB1(TriD), but not HSPB1(TriA) was able to reduce cell death from oxygen and glucose deprivation in cortical neurons (Stetler et al.,

2012). They also found that transgenic mouse lines expressing HSPB1(WT) or

HSPB1(TriD) had decreased infarct volume, and performed better than control cohort mice in 4 neurofunctional tests after neuronal ischemia compared to

HSPB1(TriA) mice and non-transgenic controls (Stetler et al., 2012).

Furthermore, the authors demonstrate that HSPB1(WT) and HSPB1(TriD) confer this neuroprotection by directly inhibiting apoptosis signal-regulating kinase 1

(ASK1)-mediated apoptosis (Stetler et al., 2008; Stetler et al., 2012).

Characterization of these mice showed HSPB1 expression was primarily in neurons, but some glial cells were positive for the transgene (Stetler et al., 2008).

Thus, mounting evidence, including this data, suggests that HSPB1 neuroprotection is less linked to protein chaperone activity, and more expressed through interaction with client proteins via as yet undefined signal pathways.

Furthering the notion of non-chaperone mediated-cell survival, I observed that expression of HSPB1(WT) alters the inflammatory profile of SOD1(G93A) astrocytes. Some genes, such as MIF, have been previously reported to be beneficial in SOD1-ALS mouse models (Israelson et al., 2015; Leyton-Jaimes et al., 2016). Others, such as interleukin-4 (IL-4), may provide novel mechanisms for neuroprotection. Studies have shown that treatment of astrocytes with IL-4 promote the expression of the brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF) (Awatsuji et al., 1993; Brodie et al., 1998; Derecki et al., 2010), while reducing the production of superoxide radicals such as nitric oxide (Hu et al., 1995). Interestingly, mutations in HSPB1 that attenuate its observed neuroprotective effects fail to up-regulate these genes to similar levels as wild-type HSPB1. Examination of the inflammatory profile of SOD1(G93A) astrocytes expressing other HSPB1 mutants, including the HSPB1(TriA) and

HSPB1(TriD) variants, may identify key inflammatory genes that promote non- cell autonomous neuroprotection.

Therapies that target the induction of heat shock proteins (HSPs) have been developed as potential treatments for ALS patients. Resveratrol, Celastrol and Arimoclomol are 3 compounds that have been reported to delay disease onset and extend survival in SOD1(G93A) transgenic mice (Han et al., 2012;

Kiaei et al., 2005; Kieran et al., 2004). Resveratrol, a polyphenol found in grapes and red wine, activates SIRT1, a deacetylase that works upstream of heat shock

82 factor 1 (HSF1), the master regulator of HSP gene transcription. Spinal cord homogenates from mice treated with resveratrol display increased levels of

HSPB1 and HSP70, suggesting that induction of these HSPs may be beneficial to motor neuron survival (Han et al., 2012). Celastrol, a triterpene, is also a potent activator of HSF1. In vitro experiments demonstrate that Celastrol treatment is capable of inducing HSPB1 in human neuronal cell lines (Deane and

Brown, 2016). Arimoclomol, which is currently in phase 2/3 clinical trials for ALS patients, is a co-activator that prolongs the binding of HSF1 to heat shock elements of Hsp genes(Kalmar et al., 2008). Interestingly, a recent study by

Deane et al. showed that treatment of SH-SY5Y cells with Celastrol, but not

Arimoclomol induces the expression of HSPB1 and other HSPs. However, treatment with both Celastrol and Arimoclomol, produced an even higher level of

HSPB1 expression than Celastrol alone (Deane and Brown, 2016). Each of these treatments hold promise, yet the effect these compounds have on HSP regulation in non-neuronal cells has been poorly characterized. More attention may need to be paid to this aspect of HSP induction therapies given my results. It is likely that HSPB1-inducing therapies would be used as part of a combinatorial strategy to treat motor neuron disease. To maximize the potential benefits of

HSP neuroprotection, I suggest that future studies examine the role of HSPs in glial cells in addition to neuronal cell types. To determine if HSPs can be used as a broad class of therapies for ALS, studies to examine the efficacy of HSP therapies in ALS models other than mutant SOD1 must be performed.

In conclusion, HSPB1(WT) expression in SOD1(G93A) astrocytes improves motor neuron survival in vitro through the regulation of an unidentified signal pathway, and likely not via protein chaperone activity. Overexpression of neuropathy-associated HSPB1 mutants mitigated this protective effect and failed to induce non-cell autonomous motor neuron toxicity. These data suggest that neuropathy-associated mutations in HSPB1 may result in loss of a neuroprotective function in non-neuronal cells contributing to the pathogenesis of axonal neuropathy, and highlight the important role that HSPB1 plays in motor neuron health and maintenance.

CHAPTER 3: MUTATIONS IN HSPB1 DISRUPT TNFa INDUCED NF-kB SIGNALING

3.1 INTRODUCTION

Nuclear Factor kB (NF-kB) transcription factors bind to discrete DNA elements, kB elements, to promote or suppress gene transcription (Hayden and

Ghosh, 2004). NF-kB plays a key role in regulating an array of biological processes, including inflammatory and immune responses, cell growth, proliferation and apoptosis (Bonizzi and Karin, 2004; Pahl, 1999; Pasparakis et al., 2006). There are 5 mammalian NF-kB family members, p65/RelA, RelB and c-Rel, p50/p105 and p52/p100 that form homo- and heterodimeric complexes.

Under non-stressed conditions, NF-kB is sequestered in the cytoplasm through interactions with the inhibitory IkB family of proteins. Activation of NF-kB is initiated by phosphorylation of IkB by the multiprotein IkB kinase (IKK) complex, targeting IkB for degradation via the ubiquitin/proteasome pathway. After degradation of IkB, unbound NF-kB rapidly translocates to the nucleus (Hayden and Ghosh, 2004).

There are two main NF-kB pathways in mammalian cells. The canonical pathway is activated by most physiological stimuli, including tumor necrosis

85 factor alpha (TNFa), and interleukin-1 (IL-1). This pathway results in activation of the IKK complex, consisting of IKKa, IKKb, and IKKg subunits. IKKb is the main kinase responsible for phosphorylation of IkBa (Gerondakis et al., 1999).

Degradation of IkBa frees heterodimers of p50/p65 (NF-kB), which translocate to the nucleus and activate gene transcription (Henkel et al., 1992).

The cytoprotective properties of HSPB1 may be related to its regulation of canonical NF-kB signaling (Figure 3.1), particularly in response to TNFa and IL-

1. Stimulation of cells with TNFa results in the rapid phosphorylation of HSPB1, and overexpression of HSPB1 in L929 fibroblasts reduced cell death in response to this cytokine (Mehlen et al., 1995a; Mehlen et al., 1995b). A link between

HSPB1 and the NF-kB pathway was first described in the U937 human leukemic and mouse embryonic fibroblast (MEF) cell lines (Parcellier et al., 2003).

Overexpression of HSPB1 enhanced the nuclear translocation and transcriptional activity of NF-kB in response to TNFa, IL-1b and etoposide. HSPB1 was found in complex with phosphorylated IkBa and the 26S proteasome, and promoted

IkBa degradation via the ubiquitin/proteasome pathway. Shortly thereafter, Park et al. reported on the involvement of HSPB1 in TNFa-induced NF-kB activation in

HeLa cells, finding that overexpression of HSPB1 negatively regulated NF-kB signaling through its interaction with the IKK complex, which reduced IKK kinase activity (Park et al., 2003). Further studies identified yet another level of NF-kB regulation by HSPB1 (Wu et al., 2009). The authors demonstrated that HSPB1

86 directly interacts with TNF Receptor Associated Factor 6 (TRAF6), and overexpression of HSPB1 promoted IL-1b induced TRAF6 ubiquitination, leading to IKK activation. In contrast to TNFa activated NF-kB, phosphorylation of

HSPB1 by IL-1b stimulation leads to a decrease in NF-kB activity, by reducing

TRAF6 ubiquitination.

Figure 3.1 HSPB1 regulates classical NF-kB signaling at multiple levels. IL- 1b and TNFa activated NF- kB signaling cascades. Studies have shown that HSPB1 can regulate IL-1b mediated NF- kB signaling through its interactions with TRAF6. HSPB1 regulates TNFa-induced NF-kB signaling through interactions with the IKK complex, or IkBa and the 26S proteasome. Figure based on references (Dodd et al., 2009; Guo et al., 2009; Parcellier et al., 2003; Park et al., 2003; Salari et al., 2013; Sur et al., 2008; Wei et al., 2011; Wu et al., 2009)

However, regulation of NF-kB by HSPB1 remains poorly characterized, and multiple studies indicate that HSPB1 can activate or suppress NF-kB signaling, depending on the cell type, method of stimulation and phosphorylation state of HSPB1 (Table 3.1). In this chapter, I investigated the effect of HSPB1 overexpression on TNFa-induced NF-kB signaling in HeLa cells. I found that transient overexpression of HSPB1(WT) reduces expression of a NF-kB controlled luciferase reporter in response to TNFa treatment, but does not alter the degradation of IkBa or the translocation of the NF-kB subunit, p65. Further, I show that HSPB1(WT) does not interact with the IKK complex, or IkBa.

Interestingly, HSPB1(R136W) does not reduce the expression of the luciferase reporter in response to TNFa stimulation. Overexpression of HSPB1(R136W) also had no effect on IkBa degradation, or p65 translocation. Together, my data suggests that HSPB1(WT) may regulate NF-kB signaling through a yet unidentified mechanism, which is perturbed by neuropathy-associated mutations in this protein.

Table 3.1 Cell specific NF-kB regulation by HSPB1

3.2 EXPERIMENTAL METHODS

3.2.1 HELA CELL CULTURE

HeLa cells were purchased from ATCC (Manassas VA), and maintained in

DMEM media (Thermo Fisher) supplemented with 10% FBS (v/v) and 1%

Penicillin/Streptomycin cocktail, PS, (v/v) (Cell-Gro). HeLa cells were maintained

90 at 37°C and 5% CO2 in a water-jacketed incubator (Fisher).

3.2.2 PLASMIDS

The wild-type HSPB1 vector was generated by digestion of M13-HSPB1 plasmid (GeneCopoeia, Rockville MD) with BstB1 and NotI, and subsequently ligated into the pcDNA4/TO vector (Invitrogen), in frame with an N-terminal FLAG sequence. HSPB1(G84R) and HSPB1(R136W) vectors were generated using site directed mutagenesis (Agilent) (see Table 2.1). NF-kB Reporter plasmids were purchased from Qiagen.

3.2.3 TRANSIENT TRANSFECTIONS

Prior to transfection, HeLa cells were plated in 6 well plates for TNFa stimulation experiments, or onto glass coverslips in 24-well plates for immunofluorescent studies. HeLa cells were transiently transfected using

TransIT HeLa Monster (Mirus). Briefly, DNA was combined with HeLa reagent and Monster reagent in serum free media and incubated at room temperature for

20 minutes to allow DNA complexes to form. DMEM supplemented with 10%

FBS (v/v) and 1% PS (v/v) was then added to the transfection complexes and immediately added to the cells. Twenty-four hours after transfection, the media was replaced with fresh DMEM supplemented with 10% FBS (v/v) and 1% PS

(v/v). Transfected cell lines were given an additional 24 hours to recover before being used in experiments.

3.2.4 NF-kB LUCIFERASE ASSAYS

HeLa cells were plated in a 96-well plate at a density of 1.5X10^4 cells per well. Twenty-four hours post-plating, cells were transfected with Attractene

(Qiagen) according to the manufacturer’s protocol. Briefly, Attractene reagent was diluted in OPTI-MEM serum free medium (Gibco) and incubated at room temperature for 5 minutes. 100 ng of NF-kB reporter plasmids and 100 ng of

HSPB1 construct or empty vector control were diluted in OPTI-MEM, added to the Attractene reagent, and incubated at room temperature for 20 minutes. The transfection complexes were then added to HeLa cells in OPTI-MEM supplemented with 5% FBS (v/v), 1% Non-Essential Amino Acids (NEAA), (v/v) and 1% PS (v/v). 16 hours later, medium containing the transfection complexes was removed and replaced with OPTI-MEM supplemented with 0.5% FBS (v/v),

1% NEAA (v/v) and 1% PS (v/v). Twenty-four hours post-transfection, cells were stimulated with TNFa (20 ng/mL) for 6 hours. The Dual-Glo Luciferase Assay Kit

(Promega) was used to measure luciferase activity. Briefly, Dual-Glo Luciferase

Reagent was added to the cells in a 1:1 ratio (reagent:medium) and incubated for

15 minutes at room temperature. Lumincescence was measured using a Tecan

F-200 plate reader. An equal volume of Dual-Glo Stop and Go reagent was then added, and the cells were incubated in the dark for 15 minutes at room temperature. Luminescence was again measured on a Tecan F-200 plate reader.

3.2.5 TNFa STIMULATION OF HELA CELLS

For TNFa stimulation time course experiments, HeLa cells were plated in 92 a 6-well plate at a density of 3.0X10^5 cells per well or on glass coverslips in a

24-well plate at a density of 3.0X10^4 cells per well. Forty-eight hours after plating, cells were treated with 20 ng/mL of TNFa for the indicated times.

Experiments using transfected HeLa cells followed the protocol in 3.2.3 prior to stimulation with recombinant TNFa. Post stimulation, cells were either lysed for western blotting (see protocol 3.2.6) or stained for immunofluorescence (see protocol 3.2.7).

3.2.6 WESTERN BLOTTING

HeLa cells were lysed in a buffer containing 20 mM Tris (pH 7.4), 150 mM

NaCl, 2 mM EDTA, 1 mM DTT, 10% Glycerol, 1% Triton X-100, 2 mM sodium pyrophosphate, 25 mM b-glycerophosphate, 1 mM Na3VO4, 10 mM NaF, 10

μg/mL leupeptin, 10 μg/mL aprotinin, and 5 mM PMSF on ice for 5 minutes.

Lysates were centrifuged (16,000 RCF) at 4°C for 10 minutes. The resulting supernatant was then boiled at 95°C for 5 minutes in 2X Laemmli Buffer (20%

Glycerol, 2% B-Mercaptoethanol (v/v), 100 mM Tris-HCl pH 6.8, 2% SDS (v/v)).

15 µg of each sample was separated by SDS-PAGE and transferred to a PVDF

(Bio-Rad) membrane. Immunoblots were blocked in 5% nonfat dry milk in TBS for 1 hour at room temperature. The blots were then incubated with primary antibodies for 1 hour at room temperature (See table 3.2 for antibody list).

Blots were then washed 3 times with TBS-T, and incubated with secondary antibody for 1 hour at room temperature, and washed 3 times with TBS-T, then once with TBS. Immunoblots were scanned using a Licor Odyssey Classic. 93

Table 3.2 Antibodies used in this chapter

3.2.7 IMMUNOFLUORESCENCE

Cells grown on coverslips were fixed with 4% PFA for 10 minutes, and washed twice with ice-cold PBS before the blocking solution consisting of PBS +

10% Goat Serum and 0.1% Triton X-100 was applied for 1 hour. Coverslips were then incubated with primary antibody against HSPB1 (Ab Cam ab2790 [1:1,000]),

IkBa (CST [1:1,000]), and p65 (CST [1:500]) for 1 hour. Cells were then washed

3 times in PBS + 0.1% Triton X-100, and subsequently incubated with secondary antibodies, Goat anti-Mouse 594 (Invitrogen [1:10,000]) or Goat anti-Rabbit 488

(Invitrogen [1:10,000]) and DAPI for 1 hr at room temperature. Secondary antibody and DAPI were then removed by washing coverslips with PBS + 0.1%

Triton X-100 3 times. Coverslips were mounted to glass slides using

Fluoromount-G (SouthernBiotech). Samples were imaged using an Andor spinning disk confocal microscope.

3.2.8 CO-IMMUNOPRECIPITATIONS

HeLa cells were grown to confluency in 15 cm2 dishes and then treated with TNFa (20 ng/mL or 100 ng/mL) for 0 or 30 minutes. After stimulation, cells were washed twice with ice cold PBS and harvested using a cell scraper. Cells were pelleted by centrifugation (1000 RCF for 5 minutes at 4°C). Following centrifugation, the supernatant was removed and cell pellets were left untreated or crosslinked with formaldehyde and subsequently lysed in a buffer (B1 IP

Buffer) containing 20 mM Tris (pH 7.4), 150 mM NaCl, 2 mM EDTA, 1 mM DTT,

10% Glycerol, 1% Triton X-100, 2 mM sodium pyrophosphate, 25 mM b- 95 glycerophosphate, 1 mM Na3VO4, 10 mM NaF, 10 μg/mL leupeptin, 10 μg/mL aprotinin, and 5 mM PMSF on ice for 5 minutes. Input lysates were prepared by mixing 100 µg of cell lysate with 2X Laemmli Buffer (20% Glycerol, 2% b-

Mercaptoethanol (v/v), 100 mM Tris-HCl pH 6.8, 2% SDS (v/v)), and boiled for 5 minutes at 95°C. 1 mg of cell lysate was then incubated with 10 µg of antibodies against HSPB1 (Ab Cam ab2790), IKKb (CST 8943), Mouse IgG (Ab Cam ab18448) overnight at 4°C on a rotating arm. 10 µg of Protein G Sepharose bead slurry was used for each IP and the beads were prepared by washing for 10 minutes in B1 IP Buffer 3 times. The beads were then added directly to the lysate:antibody complexes and incubated for an additional 2 hours at 4°C on a rotating arm. Unbound cellular lysate was removed by washing the bead:antibody complexes with B1 IP Buffer 3 times (10 minutes per wash), and the protein samples were eluted from the bead by boiling in 2X Laemmli Buffer for 5 minutes at 95°C. The presence of HSPB1, IKKa, and IKKb proteins was then determined by western blotting, following the protocol in 3.2.6.

3.2.9 STATISTICS

Statistical analysis was performed by two-way ANOVA for mean differences between the average of naïve cell lines, HSPB1(WT) cell lines and

HSPB1(R136W) cell lines across treatments (GraphPad Prizm Software). All experiments were performed at least in triplicate.

3.2.10 HEAT SHOCK EXPERIMENTS

HeLa cells were plated in 10-cm2 dishes and grown to 80% confluency prior to heat shock assays. HeLa cells were then subjected to a 1 hour heat shock at 45°C and then allowed to recover at 37°C for 3 hours or 48 hours. Non- heat shocked cells were used as a control cell line. Following heat shock, HeLa cells were lysed in a buffer containing 20 mM Tris (pH 7.4), 150 mM NaCl, 2 mM

EDTA, 1 mM DTT, 10% Glycerol, 1% Triton X-100, 2 mM sodium pyrophosphate,

25 mM b-glycerophosphate, 1 mM Na3VO4, 10 mM NaF, 10 μg/mL leupeptin, 10

μg/mL aprotinin, and 5 mM PMSF on ice for 5 minutes to produce a total lysate.

Lysates were centrifuged (16,000 RCF) at 4°C for 10 minutes. The resulting supernatant was removed and used as the cytosolic protein fraction. The pellet was considered the nuclear fraction, and re-suspended in lysis buffer. Samples were then used in western blotting as described in 3.2.6.

3.3 RESULTS

3.3.1 TNFa STIMULATION RAPIDLY ACTIVATES NF-kB SIGNALING IN HELA CELLS

Treatment of HeLa-cells with TNFa causes the rapid phosphorylation of

HSPB1, but does not increase HSPB1 protein expression (Mehlen et al., 1995a).

I monitored the total levels of HSPB1 expression and phosphorylation in HeLa cells treated with TNFa for various times. Total cellular lysates were prepared from TNFa-stimulated cells, and phosphorylation of HSPB1 at S15, S78 and S82 was determined using semi-quantitative western blotting with phospho-specific 97 antibodies against HSPB1 (Figure 3.2A). I observed a major increase in phosphorylation of S78 after 5 minutes of treatment. Phosphorylation peaked at

10 minutes, and began to decrease after 60 minutes. I detected weak phosphorylation of S82, with the strongest signal following 10 minutes of TNFa stimulation, and no phosphorylation of S15. As expected, treatment with TNFa did not increase HSPB1 expression (Figure 3.2A). Next, I compared the rate of

HSPB1 phosphorylation to the rate of IkBa degradation, a key step in NF-kB signaling. My data show that a majority of IkBa degradation occurs after 10 minutes, coinciding with the phosphorylation of HSBP1 (Figure 3.2A). After 60 minutes, IkBa levels increased, concomitant with a decrease in HSPB1 phosphorylation.

I further characterized activation of the NF- kB signaling pathway through immunofluorescent staining of HeLa cells with antibodies against IkBa and p65.

Staining revealed a strongly reduced IkBa signal after 10 minutes, which correlates with my western blot data (Figure 3.2B). Similarly, I observed a strong nuclear translocation of p65 at 10 minutes, confirming that TNFa treatment rapidly activates the NF-kB pathway (Figure 3.2B). Together, these data indicate that TNFa induces the rapid phosphorylation of HSPB1 and activation of NF-kB signaling. Furthermore, the phosphorylation of HSPB1 coincides with the degradation of IkBa, suggesting a possible link between these two events.

Figure 3.2 IkBa degradation and HSPB1 phosphorylation occur concomitantly in response to TNFa-stimulation. A. Representative immunoblot from Hela cells stimulated with TNFa (20 ng/mL) for 0, 5, 10, 30, 60 and 120 minutes. HSPB1 phosphorylation peaks at 10 minutes, the same time that IkBa degradation occurs. B. Representative immunofluorescence of HeLa cells stimulated with TNFa (20 ng/mL) for 0,

5, 10, and 20 minutes. Cells were stained with antibodies against IkBa and the P65 subunit of the NF-kB heterodimer. Both proteins show strong cytoplasmic localization in un-stimulated cells. After 10 minutes of stimulation, there is a significant loss of IkBa signal, and P65 translocates from the cytoplasm to the nucleus.

3.3.2 HSPB1(WT) OVEREXPRESSION REDUCES TNFa-INDUCED NF-kB ACTIVITY

Next, I measured the activation of NF-kB reporter expression following

TNFa-stimulation. HeLa cells were transiently transfected with a NF-kB responsive luciferase reporter or control luciferase reporters (Figure 3.3A) and

24 hours later, stimulated with TNFa for 6 hours, after which luciferase activity was assayed. In cells transfected with control luciferase vectors TNFa stimulation had no effect on luciferase activity (Figure 3.3B). However, in cells expressing the NF-kB reporter, we observe a 3-fold increase in luciferase activity following

TNFa treatment (Figure 3.3C).

100

Figure 3.3 TNFa-stimulated NF-kB reporter assay. A. Diagram of NF-kB luciferase reporters used for my stimulation assays. The NF- kB responsive plasmid expresses firefly luciferase under the control of 6 kB binding sites. The negative control plasmid lacks a functional promoter upstream of the firefly luciferase gene, while the positive control plasmid expresses firefly luciferase under the control of the CMV promoter, making gene expression constitutively active. The normalization control plasmid was included in all experiments and was used to normalize the signal intensity from the firefly luciferase. B. C. Ratio of firefly/renilla luciferase signal intensity from HeLa cells treated with TNFa or cells that were left un-treated. HeLa cells were transfected with the positive control or negative control luciferase plasmids and with the normalization control plasmid (B.) or with the NF- kB responsive plasmid with the normalization control plasmid (C.). 48 hours following transfection, the cells were placed in serum free media and stimulated with 20 ng/mL of TNFa or PBS for 6 hours. Following stimulation, luciferase was measured according to the manufacturer’s

101 protocol and detected on a Tecan F-200 plate reader. Each sample was run in duplicate. Error bars represent s.e.m.

HSPB1 has previously been reported to negatively regulate TNFa-induced

NF-kB activity in HeLa cells. To confirm these results, I co-transfected HeLa cells with vectors encoding HSPB1(WT) and the NF-kB responsive luciferase reporter.

An empty vector was used as a transfection control. Overexpression of

HSPB1(WT) reduced luciferase activity by ~50% following TNFa-stimulation compared to cells transfected with an empty vector, 1.60 ± 0.11 vs. 2.78 ± 0.11

(Figure 3.4A).

In chapter 2, my data indicated that two neuropathy-associated mutations in HSPB1, p.G84R and p.R136W attenuate its non-cell autonomous neuroprotection, through an unidentified mechanism. Thus, I examined if overexpression of mutHSPB1 would alter TNFa-induced NF-kB activity in HeLa cells. My data indicates that overexpression of HSPB1(G84R) and

HSPB1(R136W) had no effect on the activity of the luciferase reporter, 2.70 ±

0.16 and 2.52 ± 0.19 respectively (Figure 3.4A). Using semi-quantitative western blotting, I determined that each transgene was expressed to similar levels

(Figure 3.4B). Together these data suggest that HSPB1 negatively regulates

TNFa-induced NF-kB signaling, and this negative regulation is abrogated by at

102 least two neuropathy-associated HSPB1 variants.

Figure 3.4 Overexpression of HSPB1(WT) reduces NF-kB activity in HeLa cells. A. Ratio of firefly/renilla luciferase signal intensity from HeLa cells treated with TNFa or cells that were left un-treated. HeLa cells were co-

103 transfected with 3 vectors: 1) the NF-kB responsive plasmid, 2) the normalization control plasmid and 3) an empty vector or HSPB1 construct. 48 hours following transfection, the cells were placed in serum free media and stimulated with 20 ng/mL of TNFa or PBS for 6 hours. Following stimulation, luciferase was measured according to the manufacturer’s protocol using a Tecan F-200 plate reader. Each sample was run in duplicate. Error bars represent s.e.m. ***P<0.0005, n.s., non-significant. B. Representative immunoblot showing expression of exogenous and endogenous HSPB1 in HeLa cells. HeLa cells were transiently transfected with vectors expressing FLAG tagged wild-type or mutant HSPB1 for 48 hours and then harvested for protein analysis. Due to the FLAG tag, exogenous HSPB1 constructs migrate slower than untagged endogenous HSPB1, allowing for the two protein species to separate on an SDS-PAGE gel.

3.3.3 OVEREXPRESSION OF HSPB1(WT), HSPB1(G84R), OR HSPB1(R136W) DOES NOT AFFECT p65 RELEASE AND NUCLEAR TRANSLOCATION

I sought to determine how HSPB1 regulates NF-kB signaling. Parcellier and colleagues demonstrated that overexpression of HSPB1 enhanced degradation of IkBa by the 26S proteasome (Parcellier et al., 2003), therefore I examined IkBa degradation in TNFa-stimulated cells expressing HSPB1(WT),

HSPB1(R136W) or an empty vector control. My data indicates there is no difference in the rate of IkBa degradation between cells expressing HSPB1(WT) or the empty vector control, nor the cells expressing HSPB1(R136W) (Figure

3.5A). Immunofluorescent staining for IkBa degradation and p65 translocation confirmed these data (Figure 3.6A, 3.6B). I also examined the phosphorylation of HSPB1 at S78 in HeLa cells overexpressing HSPB1(WT) or HSPB1(R136W). 104

The data shows that in all cell lines, HSPB1 phosphorylation peaks after 10 minutes of TNFa-stimulation (Figure 3.5B). Interestingly, while TNFa treatment increased the phosphorylation of exogenous HSPB1(WT) and HSPB1(R136W), I also observed phosphorylation of exogenous HSPB1 prior to TNFa treatment.

Since HSPB1 is phosphorylated in response to stress, this data suggests that transiently transfecting HeLa cells with plasmid DNA may elicit a stress response from the cells.

Figure 3.5 Overexpression of HSPB1 does not alter IkBa degradation following TNFa stimulation. A. Representative immunoblot of IkBa and 105

HSPB1 expression levels in HeLa cells stimulated with TNFa (20 ng/mL) for 0, 5, 10 or 20 minutes. HeLa cells were transfected 48 hours prior to TNFa stimulation with one of three vectors: 1) an empty control vector, 2) a wild- type HSPB1 vector, HSPB1(WT), or 3) a mutant HSPB1 vector, HSPB1(R136W). B. Representative immunoblot of IkBa and phosphorylated HSPB1 expression levels in HeLa cells stimulated with TNFa (20 ng/mL) for 0, 5, 10 or 20 minutes. HeLa cells were transfected 48 hours prior to TNFa stimulation with one of three vectors: 1) an empty control vector, 2) a wild- type HSPB1 vector, HSPB1(WT), or 3) a mutant HSPB1 vector, HSPB1(R136W).

106

Figure 3.6 Overexpression of HSPB1 does not affect P65 nuclear translocation following TNFa stimulation. Immunofluorescent staining of 107

IkBa and P65 from HeLa cells treated with TNFa (20 ng/mL) for 0, 5, 10 or 20 minutes. HeLa cells were plated onto glass coverslips and transfected with an empty vector (a-d and aa-dd), wild-type HSPB1 (e-h and ee-hh) or mutant HSPB1 (i-l and ii-ll) 48 hours prior to TNFa stimulation. Cells were stained for IkBa and P65, and were imaged using a spinning-disk confocal microscope.

3.3.4 HSPB1 IS NOT ASSOCIATED WITH THE IKK COMPLEX IN HELA CELLS

I examined the possibility that HSPB1 regulates NF-kB activity upstream of IkBa. Numerous studies have shown direct interaction between HSPB1 and the IKKa/b subunits of the IKK complex (Dodd et al., 2009; Kammanadiminti and

Chadee, 2006; Park et al., 2003; Sur et al., 2008). To identify if HSPB1 interacts with the IKK complex, I performed co-immunoprecipitation assays in HeLa cell lysates using antibodies against HSPB1 or IKKb. While IKKa precipitated with

IKKb, I did not detect any interaction between HSPB1 and IKKb (Figure 3.7A and 3.7B). TNFa-stimulation is reported to increase the interaction between

HSPB1 and IKKb, so I repeated the co-immunoprecipitation experiments in lysates from TNFa-stimulated HeLa cells. Similar to experiments using lysate from non-stimulated cells, I failed to detect an interaction between to two proteins

(Figure 3.7C). To maximize the chance of detecting an interaction, I fixed protein-protein interactions with formaldehyde crosslinking prior to isolating

HSPB1 and IKKb complexes. Again, I failed to detect any interaction (Figure

108

3.7D).

Figure 3.7 HSPB1 and IKKb do not co-precipitate in HeLa cells. A. Immunoblot of control IgG and HSPB1 immunoprecipitations from un- stimulated HeLa cell lysates. 50% of eluted proteins, 5% of unbound proteins and 5% of the first wash step were separated by SDS-PAGE and 109 probed for HSPB1, IKKb and Tubulin. B. Immunoblot of control IgG and IKKb immunoprecipitations from un-stimulated HeLa cell lysates. 50% of eluted proteins, 5% of unbound proteins and 5% of the first wash step were separated by SDS-PAGE and probed for HSPB1, IKKa, IKKb and Tubulin. C. Immunoblot of control IgG, IKKb and HSPB1 immunoprecipitations from HeLa cell lysates treated with 20 ng/mL of recombinant human TNFa for 0 and 30 minutes. 25% of eluted proteins were separated by SDS-PAGE and probed for HSPB1 and IKKb. D. Immunoblot of control IgG, IKKb and HSPB1 immunoprecipitations from formaldehyde cross-linked HeLa cell lysates treated with 20 ng/mL of recombinant human TNFa for 0 and 30 minutes. 25% of eluted proteins were separated by SDS-PAGE and probed for HSPB1 and IKKb.

3.4 DISCUSSION

It is now recognized that astrocytes play an active and critical role regulating the inflammatory response within the CNS (Colombo and Farina,

2016; Dong and Benveniste, 2001; Farina et al., 2007). Reactive astrocytes are immune-competent cells, capable of secreting a variety of chemokines and cytokines (Choi et al., 2014). In ALS, there is a growing appreciation for the role of neuroinflammation in disease pathogenesis. A potential role for reactive astrocytes in ALS was first reported in patient spinal cord samples, in which activation of inflammatory pathways, including NF-kB, was observed in astrocytes (Migheli et al., 1997). Recent evidence has also demonstrated that inflammatory and apoptotic-related genes are up-regulated in spinal cords from

SOD1(G93A) transgenic (Yoshihara et al., 2002), and that primary astrocyte

110 cultures from SOD1(G93A) transgenic mice present a neuroinflammatory phenotype (Hensley et al., 2006).

In chapter 2 I observed that overexpression of HSPB1 altered the expression level of numerous inflammatory genes, many of which are regulated through the NF-kB signaling pathway. Previous studies have reported on

HSPB1’s regulation of NF-kB signaling in various cell lines (see Table 3.1), suggesting that HSPB1’s cytoprotective properties may be due at least in part to its ability to modulate this inflammatory signaling pathway. Therefore, I examined the effects of wild-type and mutant HSPB1 overexpression on NF-kB activation via TNFa stimulation. Using a luciferase reporter under the control of kB binding sites, I found that overexpression of wild-type HSPB1 reduced luciferase expression after TNFa stimulation compared to cells expressing basal levels of

HSPB1. This reduction in reporter activity was independent of NF-kB activation however, as I detected no changes in IkBa degradation or nuclear translocation of P65. Additionally, co-immunoprecipitation experiments failed to show any interaction between HSPB1 and the IKK complex. Interestingly, overexpression of two neuropathy-associated HSPB1 variants did not reduce reporter activity in response to TNFa under similar conditions.

How HSPB1 may regulate NF-kB signaling in this model system remains unclear. My co-immunoprecipitation data conflicts with multiple published studies 111 that demonstrate a physical interaction between HSPB1 and the IKK complex

(Dodd et al., 2009; Guo et al., 2009; Park et al., 2003; Sur et al., 2008). It is unclear why I was unable to detect an interaction between HSPB1 and IKKb, as I could readily co-precipitate IKKb with IKKa. It is possible that the antibodies used in my experiments disrupted HSPB1-IKKb interactions. This seems unlikely though, since I also failed to detect an interaction from formaldehyde-crosslinked lysates. Additionally, I was unable to use the same antibodies as those used in previous studies, as they are no longer commercially available. HSPB1-IKKb interactions may be to transient in HeLa cells to detect reliably. Future work will require examining the other cell types where this interaction has been described, such as skeletal muscle (Dodd et al., 2009), or liver (Guo et al., 2009). It may also be easier to study HSPB1-IKKb interactions in immune-competent cell types, where such interactions may occur more frequently, or more stably.

Numerous studies have examined HSPB1’s role on the activation of NF- kB signaling in response to stimuli such as TNFa and IL-1b. In contrast with these studies, I found that overexpression of wild-type HSPB1 had no effect on the rate of IkBa degradation, or the translocation of P65 from the cytoplasm to the nucleus. However, these studies, including my own, often only examine activation of NF-kB signaling after a short stimulation period, usually 15 to 30 minutes. It is possible that HSPB1 may modulate not only the activation of this signaling cascade, but also regulate the termination of NF-kB signaling and 112 subsequent re-activation. IkBa acts as a negative feedback regulator of NF-kB signaling, by controlling the cellular localization of P65 (Chiao et al., 1994;

Fagerlund et al., 2015). Consistent with this notion, I observed IkBa protein levels increase after an hour of TNFa stimulation in naïve HeLa cells. It is conceivable that overexpression of wild-type HSPB1 could indirectly increase the rate of IkBa protein synthesis, resulting in an overall decrease in NF-kB activity.

This could explain why I detect a reduction in the luciferase reporter assays, which involved a 6-hour stimulation with TNFa. Additional studies on IkBa degradation and synthesis over the full 6-hour time period used for my reporter assays will be necessary to determine if HSPB1 expression alters the negative feedback loop of IkBa.

113

CHAPTER 4: HSPB1-ASSOCIATED ASTRC COMPLEX IS NOT PRESENT IN HELA CELLS DURING CELLULAR STRESS

4.1 INTRODUCTION

In response to stress, cells not only regulate the expression of certain genes through the activation or suppression of transcription factors, but also by altering the stability and decay of select mRNAs. Microarray data indicates that

40-50% of the changes in gene expression in response to stress arose due to altered mRNA stability (Cheadle et al., 2005a; Cheadle et al., 2005b; Fan et al.,

2002). AU-rich element (ARE) mediated mRNA decay (AMD), is one of the best characterized stress-response RNA surveillance pathway. Previous studies have demonstrated that HSPB1 plays a role in AMD, either through direct interactions with ARE-containing mRNAs (Sinsimer et al., 2008), or as part of a multisubunit complex termed the AUF1 Signal and Transduction Regulation Complex, or

ASTRC (Knapinska et al., 2011; Sinsimer et al., 2008).

AMD occurs due to the presence of AU-rich elements located in the 3` untranslated region (UTR) of mRNA transcripts (Chen and Shyu, 1995; Shaw and Kamen, 1986). AREs sequences are rich in adenylate and uridylate nucleotides, and are classified based on the number of AUUUA motifs they contain. So far, 3 major motifs have been identified (Gratacos and Brewer, 2010), 114 however a consensus sequence has not been determined. These ARE motifs make transcripts inherently unstable and short-lived (Chen and Shyu, 1995;

Shaw and Kamen, 1986). In humans, ~4,000 transcripts harbor an ARE, the majority of which encode for cytokines, growth factors and proto-oncogenes

(Halees et al., 2008). RNA-binding proteins (AUBPs) and microRNAs

(Palanisamy et al., 2012) bind AREs to regulate mRNA stability. Some AUBPs, such as AUF1, KSRP and TTP recruit factors that enhance degradation of ARE- containing mRNAs (Gherzi et al., 2004; Lykke-Andersen and Wagner, 2005;

Sarkar et al., 2003). Others, such as HuR have a stabilizing effect on ARE- mRNAs (Brennan and Steitz, 2001; Ma et al., 1997).

ARE/Poly(U)-binding/degradation factor 1 (AUF), is one of the most well studied AUBPs to date. It was identified as an AUBP by its ability to accelerate the decay of c-myc, an ARE-containing mRNA, in an in vitro decay system

(Brewer, 1991; Zhang et al., 1993). Currently, 4 different isoforms of AUF1 have been identified with differing molecular weights, p37(AUF1), p40(AUF1), p42(AUF1), and p45(AUF1). These isoforms are produced from a single transcript through alternative splicing (Figure 4.1) (DeMaria and Brewer, 1996;

Wagner et al., 1998; White et al., 2013). The p42(AUF1) and p45(AUF1) isoforms are predominately located in the nucleus, while the smaller two, p37(AUF1) and p42(AUF1) are distributed through both the nucleus and the cytoplasm (Arao et al., 2000; Wilson et al., 2003). In vitro experiments performed in NIH-3T3, COS1, HeLa and HEK293T cell lines indicate that the p37(AUF1) 115 isoform is primarily responsible for AUF1’s mRNA decay activity, though the p40(AUF1) isoform displays some activity (Sarkar et al., 2003).

Figure 4.1 Domain organization of AUF1 isoforms. The locations of peptide sequences encoded by alternatively spliced exons and the glutamine-rich (Q-rich) domain are shown flanking the tandem RNA Recognition Motifs (RRMs) common to all AUF1 isoforms. From (White et al. 2013)

A role for HSPB1 in mRNA stabilization was first reported in HeLa cells.

HSPB1(TriD) overexpression partially stabilized the mRNA transcript of cyclooxygenase 2 (Cox-2), which contains 2 ARE-motifs (Lasa et al., 2000).

Using a series of truncated Cox-2 reporters, the authors identified a single, 123- nt long portion of the 3` UTR was necessary for stabilization of the mRNA by

116

HSPB1. This same region was bound by AUF1. The authors speculated that

HSPB1-mediated Cox-2 regulation occurs by phosphorylation of HSPB1 by the p38-MK2 signaling pathway, which would stabilize Cox-2 mRNA by displacing

AUF1. HSPB1 and AUF1 were later identified in a complex together on a U-rich motif from the 3` UTR of cell death-inhibiting RNA (Shchors et al., 2002). This complex was eventually termed the ASTRC, and found to be comprised not only of AUF1 and HSPB1, but also the eukaryotic translation initiation factor 4G

(eIF4G), poly(A)-binding protein (PABP), and Hsp/Hsc70. Further, Knapinska and colleagues showed that overexpression of HSPB1(TriD) stabilized both

TNFa and IL-1b mRNAs in THP-1 cells (Knapinska et al., 2011). The authors proposed a mechanism by which phospphorylation of HSPB1 controls the ubiquitination and degradation of AUF1, thereby increasing mRNA stability through a loss of AUF1 binding to ARE-motifs. Surprisingly, siRNA-mediated knockdown of HSPB1 in THP-1 cells stabilized both reporter mRNA containing the core ARE from TNFa and the endogenous TNFa gene (Sinsimer et al.,

2008), while knockdown of HSPB1 in primary monocytes also stabilized the IL-1b mRNA (Hadadi et al., 2016).

HSPB1 may also regulate RNA stability independent of the ASTRC complex. Electrophorectic mobility shift assays using recombinant HSPB1 and radio-labeled 38nt-long RNAs demonstrated HSPB1 selectively bound the core

ARE from TNFa and showed no binding to a non-ARE fragment of the rabbit b-

117 globin gene (Knapinska et al., 2011; Sinsimer et al., 2008). HSPB1’s affinity for

AU-rich RNA was found to be comparable to that of other AUBPs, with dissociation constants in the nanomolar range.

Regulation of ARE-mRNA stability may be a novel mechanism of HSPB1 neuroprotection. In cerebellar granule neurons (CGNs), overexpression of

HSPB1 represses the translation of the proapoptotic BH3-only protein, Bim, promoting cell survival in response to H2O2-induced oxidative stress (Davila et al., 2014). The authors found that HSPB1 binds to the 3` UTR of bim, which contains multiple ARE-motifs, and that knockdown of HSPB1 in these neurons promoted bim translation.

Due to HSPB1’s ability to regulate ARE-mRNA stability, and the enrichment of ARE motifs within inflammatory genes, I investigated the role of

HSPB1 in ARE-mRNA decay as a possible mechanism for the non-cell autonomous neuroprotection in chapter 2. I found that under non-stressed and stressed conditions, HSPB1 failed to directly bind to endogenous mRNA in HeLa cells, and did not associate with identified ASTRC proteins. Further, I found that overexpression or knockdown of wild-type HSPB1 had no effect on the stability of

Cox-2 or CXCL1. These findings suggest that ASTRC formation may be cell type specific, and does not occur in HeLa cells.

118

4.2 EXPERIMENTAL METHODS

4.2.1 CELL CULTURE

HeLa cells were purchased from ATCC (Manassas VA), and maintained in

DMEM media (Thermo Fisher) supplemented with 10% FBS (v/v) and 1%

Penicillin/Streptomycin cocktail, PS, (v/v) (Cell-Gro). HeLa cells were maintained at 37°C and 5% CO2 in a water-jacketed incubator (Fisher).

4.2.2 PLASMIDS

The wild-type HSPB1 vector was generated by digestion of M13-HSPB1 plasmid (GeneCopoeia, Rockville MD) with BstB1 and NotI, and subsequently ligated into the pcDNA4/TO vector (Invitrogen), in frame with an N-terminal FLAG sequence. The HSPB1(TriD) vector was generated using site directed mutagenesis (Agilent) (See Table 2.1 for a list of primers used). siRNAs against

HSPB1 were purchased from Qiagen.

4.2.3 TRANSIENT TRANSFECTIONS

Prior to transfection, HeLa cells were plated at 50% confluency in 6 well plates. 24 hours post-plating, cells were transfected with Attractene (Qiagen) according to the manufacturer’s protocol. Briefly, Attractene reagent was diluted in OPTI-MEM serum free medium (Gibco) and incubated at room temperature for

5 minutes. Plasmids were diluted in OPTI-MEM, added to the Attractene reagent, and incubated at room temperature for 20 minutes. The transfection complexes were then added to HeLa cells in OPTI-MEM supplemented with 5% FBS (v/v),

119

1%, NEAA, (v/v) and 1% PS (v/v). 24 hours later, medium containing the transfection complexes was removed and replaced with DMEM supplemented with 10% FBS (v/v) and 1% PS (v/v). Downstream experiments were performed

48 hours post-transfection.

4.2.4 siHSPB1 TRANSFECTIONS

HiPerFect transfection reagent (Qiagen) was used to transiently transfect

HeLa cells with siRNAs against HSPB1 using the manufacturer’s fast forward

(reverse transfection) protocol. Briefly, HeLa cells were plated in a 6-well plate at a density of 4.0X10^5 cells/well in DMEM supplemented with 10% FBS (v/v) and

1% PS (v/v). 150 ng of siRNA was diluted in DMEM lacking FBS and PS.

HiPerFect transfection reagent was added to the DNA solution, vortexed for 5 seconds and then incubated at room temperature for 10 minutes. Transfection complexes were then added directly to the cells. 72 hours later, the media was changed and cells were harvested for western blotting or used for RNA decay assays.

4.2.5 CO-IMMUNOPRECIPITATIONS

HeLa cells were grown to confluency in 15-cm2 dishes and then treated with TNFa (20 ng/mL) for 0 or 30 minutes. After stimulation, cells were washed twice with ice cold PBS and harvested using a cell scraper. Cells were pelleted by centrifugation (1000 x g for 5 minutes at 4°C). Following centrifugation, the supernatant was removed and cell pellets were immediately lysed in a buffer (B1

120

IP Buffer) containing 20 mM Tris (pH 7.4), 150 mM NaCl, 2 mM EDTA, 1 mM

DTT, 10% Glycerol, 1% Triton X-100, 2 mM sodium pyrophosphate, 25 mM à- glycerophosphate, 1 mM Na3VO4, 10 mM NaF, 10 μg/mL leupeptin, 10 μg/mL aprotinin, and 5 mM PMSF on ice for 5 minutes. Input lysates were prepared by mixing 50 ug of cell lysate with 2X Laemmli Buffer (20% Glycerol, 2% B-

Mercaptoethanol (v/v), 100 mM Tris-HCl pH 6.8, 2% SDS (v/v)), and boiled for 5 minutes at 95°C. 1 mg of cell lysate was then incubated with 10 µg antibodies against HSPB1 (Ab Cam ab2790), the FLAG epitope (Sigma A2220), the Myc epitope (Sigma SAB4301136) overnight or a mouse IgG control (Ab Cam ab18448) at 4°C on a rotating arm. 10 µg of Protein G Sepharose bead slurry was used for each IP and were prepared by washing for 10 minutes in B1 IP

Buffer 3 times. Beads were then added directly to the lysate:antibody complexes and incubated for an additional 2 hours at 4°C on a rotating arm. Unbound cellular lysate was removed by washing the bead:antibody complexes with B1 IP

Buffer 3 times (10 minutes per wash), and the protein samples were eluted from the bead by boiling in 2X Laemmli Buffer for 5 minutes at 95°C. The presence of

HSPB1, AUF1, and PABP proteins was then determined by western blotting, following the protocol in 4.2.6.

4.2.6 WESTERN BLOTTING

For analysis of transgene expression and HSPB1 knockdown, 15 µg of each sample was separated by SDS-PAGE and transferred to a PVDF (Bio-Rad) membrane. Immunoblots were blocked in 5% nonfat dry milk in TBS for 1 hour at 121 room temperature. The blots were then incubated with primary antibodies for 1 hour at room temperature (See table 4.1 for antibody list). Blots were then washed 3 times with TBS-T, and incubated with secondary antibody for 1 hour at room temperature, and then scanned using a Licor Odyssey Classic.

Table 4.1 Antibodies used in this chapter

4.2.7 mRNA DECAY ASSAY

Transfected HeLa cells were stimulated with TNFa for 1 hour to promote cytokine gene expression, and then transcription was inhibited with the addition

122 of actinomycin D (5 µg/mL) for 0, 30, 60, 120 and 180 minutes. Cells were subsequently harvested and re-suspended in TRIzol. RNA was purified through a phenol:chloroform extraction and used immediately for qPCR experiments or stored at -80°C.

4.2.8 QPCR MEASUREMENTS

cDNA was generated from 1 µg of purified RNA with Oligo d(T) primers

(Invitrogen) using SMART MMLV reverse transcriptase (Clonetech) according to the manufacturer’s instructions. cDNA was then diluted 10-fold to create a stock cDNA solution. qPCR was then performed with the Bio-Rad iTaq Universal SYBR

Green Supermix system, and primers specific to CXCL1, Cox-2, Erg-1, eIF2B2,

RPL37A, and b-Actin (See Table 4.2). CT values of CXCL1, Cox-2 and Erg-1 were normalized to the geometric mean of eIF2B2, RPL37A and b-Actin CT values. Changes in gene expression were calculated using the DDCT method and plotted as a percentage of the time zero data point. The data was analyzed using non-linear regression in PRISM software, and the half-life was calculated from the first-order decay constant, k, where mRNA half-life was equal to: ln2/k.

123

Table 4.2 qPCR primers used in this chapter

4.2.9 RNA CO-IMMUNOPRECIPITATION ASSAY

HeLa cells were grown to confluency and treated with 20 ng/mL of recombinant TNFa for 30 minutes or left untreated. Cells were formaldehyde crosslinked, and then lysed in RSB-100 + 1% empigen, and sonicated on ice 3 times for 10 seconds. Lysates were then centrifuged at 16,000 RCF for 10 minutes at 4°C and the supernatant was precleared through a 0.45 micron PDVF filter. Protein:RNA complexes were immunoprecipitated using 10 µg antibodies against HSPB1 (Ab Cam ab2790), HuR (Santa Cruz) and Mouse IgG (Ab Cam ab18448) immobilized on Protein G Sepharose beads. Briefly, Protein G

Sepharose beads were washed 3 times in appropriate lysis buffer and then

124 incubated with antibodies overnight at 4°C on a rotating arm. Excess antibody was then removed by washing 3 times in appropriate lysis buffer. 500 ug of HeLa cell lysates were added to the bead:antibody complexes and incubated for 3 hours at 4°C on a rotating arm. After incubation, samples were washed 6 times with lysis buffer, and subsequently incubated with proteinase K to degrade all protein. RNA was isolated and purified using TRIzol (Thermo Fisher) following the manufacturer’s protocol and stored at -80°C.

Isolated RNA was used in RT-PCR reactions using a radio-labeled oligo d(T) primers. Briefly, oligo d(T) was radiolabeled with [g-32P]ATP using T4 polynucleotide kinase (NEB). The radiolabeled probe was then used to generate cDNA from isolated RNA samples with the SMART MMLV reverse transcriptase kit (Clontech), according to the manufacturer’s instructions. cDNA was loaded on an 8% acrylaminde-urea gel and run at 10W for 20 minutes. The gel was dried for 1 hour and then exposed to autoradiograph film for 18 hours.

4.2.10 STATISTICS

Statistical analysis was performed by two-way ANOVA for mean differences between the average half-life of each mRNA in naïve, HSPB1(WT),

HSPB1(TriD) and siHSPB1 cell lines (GraphPad Prizm Software). All experiments were performed in triplicate.

4.3 RESULTS

125

4.3.1 HSPB1 DOES NOT BIND mRNA IN HELA CELLS

HSPB1 has been reported to bind to the 3` UTR of ARE-containing mRNAs in neurons and monocytes (Davila et al., 2014; Knapinska et al., 2011;

Sinsimer et al., 2008), so I examined if HSPB1 displays RNA binding activity in

HeLa cells. I performed RNA co-immunoprecipitations on cytosolic lysates from non-treated HeLa cells with antibodies against HuR, a known RNA binding protein, and HSPB1, then performed RT-PCR to amplify precipitated RNA. As expected, HuR bound a large amount of RNA, while an IgG control did not

(Figure 4.2). I failed to detect any RNA in the HSPB1-precipitated sample

(Figure 4.2), suggesting that under non-stimulating conditions, HSPB1 does not exhibit RNA binding in HeLa cells. In cerebellar granule neurons (CGNs), treatment with hydrogen peroxide increased the association of HSPB1 with Bim mRNA (Davila et al., 2014). Since treatment of cells with hydrogen peroxide is known to induce HSPB1-phosphorylation, and the phosphomimetic variant of

HSPB1 binds specifically to ARE-containing RNAs, we decided to induce HSPB1 phosphorylation in HeLa cells. I previously observed that HSPB1 phosphorylation peaks following 30 minutes of stimulation with TNFa (Figure 3.2), therefore I repeated my RNA co-immunoprecipitations from cytosolic cellular lysates after 30 minutes of TNFa-stimulation and then analyzed bound RNA through RT-PCR.

Again, I found that HuR co-precipitated RNA, while an IgG control and HSPB1 did not (Figure 4.3). Together, my data indicate that HSPB1 does not display

RNA-binding activity in HeLa cell lysates.

126

Figure 4.2 RNA does not precipitate with HSPB1 in naive HeLa cells. RNA co-immunoprecipitations from non-stimulated HeLa cell lysates. Prior to immunoprecipitation, cells were crosslinked with formaldehyde to preserve RNA-protein interactions. RT-PCR was performed with P32-radiolabeled oligo d(T) primers. Reactions lacking reverse transcriptase were performed to control for genomic DNA contamination.

127

Figure 4.3 RNA does not precipitate with HSPB1 in TNFa stimulated HeLa cells. RNA co-immunoprecipitations from stimulated HeLa cell lysates, crosslinked with formaldehyde. HeLa cells were incubated with 20 ng/mL of recombinant TNFa for 30 minutes prior to formaldehyde crosslinking and immunoprecipitation. RT-PCR was performed with P32-radiolabeled oligo d(T) primers. Reactions lacking reverse transcriptase were performed to control for genomic DNA contamination.

4.3.2 HSPB1 OVEREXPRESSION DOES NOT ALTER ARE-mRNA DECAY IN HELA CELLS

While I did not detect a direct interaction between HSPB1 and mRNA in vitro, HSPB1 may still regulate ARE-mRNA stability through the formation of the

128

ASTRC. Lasa and colleagues showed that overexpression of phosphomimetic

HSPB1 in HeLa cells destabilized the expression of a b-globin reporter mRNA containing the 3` UTR core of cyclooxygenase 2 (Cox-2), which contains an

ARE-motif bound by AUF1 (Lasa et al., 2000). To determine role of HSPB1 in regulating mRNA stability, I generated HeLa cell lines overexpressing

HSPB1(WT), or HSPB1(TriD) (Figure 4.4A). I also used siRNAs to generate

HSPB1-deficient HeLa cells (Figure 4.4B). The decay of Cox-2 and another

ARE-containing mRNA, CXCL1, was monitored in these cell lines using actinomycin D time course assays. I also monitored the decay of Erg-1, whose mRNA does not contain an ARE. I observed no change in the half-life of all 3 mRNAs in the HSPB1-deficient cells and those overexpressing HSPB1(TriD)

(Figure 4.4C,D,E and Table 4.3). Interestingly, I observed an increase in the average mRNA half-life for CXCL1 (100.3 ± 15.56 vs. 56.48 ± 9.97, P<0.05) and

Erg-1 (39.32 ± 9.05 vs. 23.31 ± 3.87, P<0.05) mRNAs in HeLa cells overexpressing wild-type HSPB1. This result was unexpected, as published data indicate the overexpression of wild-type HSPB1 has no effect on mRNA stability, and further, the 3` UTR of Erg-1 does not contain ARE-motifs.

129

Figure 4.4 HSPB1 expression levels do not alter mRNA stability in HeLa cells. A. Representative immunoblot of HeLa cell lysates confirming the overexpression of HSPB1(WT) and HSPB1(TriD) following transient transfection. B. Representative immunoblot confirming knockdown of HSPB1 in HeLa cells using siRNAs. C. D. E. Average mRNA decay curves of ARE-containing mRNAs, CXCL1 (C.), Cox-2 (D.) and the non-ARE mRNA, Erg-1 (E.) in HeLa cells transfected with HSPB1(WT), HSPB1(TriD) plasmids or an siRNA targeted against human HSPB1. HeLa cells were stimulated with TNFa (20 ng/mL) for 1 hour and then treated with actinomycin D (5 µg/mL) for 0, 30, 60, 120 and 180 minutes to block transcription. RNA was purified and assessed by qPCR. The data was then analyzed using non- linear regression to determine mRNA half-life, reported in Table 4.1. (N = 3 for each cell line tested, and each N was run in triplicate.)

130

Table 4.3 Calculated half-life of endogenous mRNAs

4.3.3 ASTRC COMPLEX FAILS TO FORM IN HELA CELLS AFTER TNFa TREATMENT

Previous studies have shown that HSPB1 and AUF1 physically interact in

HeLa cells to modulate mRNA stability. I sought to identify an interaction between HSPB1 and AUF1 in HeLa cells and examine the effects of TNFa stimulation on HSPB1/AUF1 complex formation. HSPB1 was immunoprecipitated from naïve and TNFa treated HeLa cells and probed for the presence of AUF1 by semi-quantitative western blotting (Figure 4.5A). I failed to detect an interaction between HSPB1 and all 4 AUF1 isoforms. I also tested for the presence of another ASTRC component, PABP, that reportedly interacts with HSPB1.

Similarly, I found no evidence of an interaction between HSPB1 and PABP under naïve or stimulating conditions (Figure 4.5A). I hypothesized that interactions between HSPB1 with AUF1 may be masked by binding of my HSPB1 antibody to its target epitope, disrupting HSPB1:AUF1 complexes. Sinsimer and colleagues identified the interaction between AUF1 and HSPB1 by performing co-

131 immunoprecipitations with an antibody against AUF1 that was generated in their lab. They did not perform any precipitations with an antibody against HSPB1, so it is possible that the interaction is only detectable when isolating the complex with AUF1 antibodies. To address this, I transiently overexpressed differentially tagged HSPB1 and AUF1 constructs in HeLa cells and performed immunoprecipitation against the protein tags on each construct (Figure 4.5B).

Neither construct co-precipitated with the other, suggesting that HSPB1 and

AUF1 may not directly interact in HeLa cells.

132

Figure 4.5 Immunoprecipitation of ASTRC subunits from HeLa cells. A. Representative immunoblot HSPB1 co-immunoprecipitations. HeLa cells were stimulated with 20 ng/mL of recombinant TNFa for 15 minutes, or left untreated prior to immunoprecipitation. Samples were probed for the presence of HSPB1, AUF1 and PABP. B. Immunoprecipitation of FLAG- AUF1 and Myc-HSPB1 from HeLa cells. Immunoprecipitation from naïve HeLa cells served as a negative control for FLAG and Myc pulldowns. Immunoblots were probed for the presence of the FLAG epitope and the Myc epitope.

4.4 DISCUSSION

HSPB1’s role in the regulation of AMD remains unclear. Numerous lines of evidence indicate that HSPB1 destabilizes ARE-mRNAs in a variety of cell lines through the formation of a large multisubunit complex termed the ASTRC. I 133 attempted to isolate the ASTRC from HeLa cells, and could find no interaction between HSPB1 and the core component of the ASTRC, AUF1. Additionally, I observed no change in the half-life of ARE-mRNAs in HSPB1-deficient HeLa cells. Further studies identified HSPB1 as an RNA-binding protein, with an affinity for ARE-motifs. I investigated the ability of HSPB1 to interact with RNA in vitro.

Using RNA co-immunoprecipitation assays, I found no evidence of HSBP1 directly binding to RNA in HeLa cells under non-stimulated and stimulated conditions. Taken together, my data suggests that in HeLa cells, HSPB1 does not participate in AMD.

Currently, the only evidence of HSPB1 being an AUBP is from EMSA experiments, which utilized recombinant HSPB1 and a 38-nucleotide core of the

TNFa ARE (Knapinska et al., 2011; Sinsimer et al., 2008). Most experimental data points towards HSPB1 forming complexes with AUF1 to facilitate ARE- binding. My results indicate that HSPB1 does not directly bind to any mRNA in

HeLa cells, nor does it interact with AUF1. It is possible HSPB1-RNA interactions occur in HeLa cells, but the assay was simply unable to detect the interaction. Such an explanation would suggest that HSPB1-RNA interactions are weak, that they occur extremely infrequently, or that target mRNAs bound by

HSPB1 are not expressed in HeLa cells. Indeed, Hsp25, the mouse homolog of

HSPB1, was found to bind to the 3` UTR of Bim, which contains ARE-motifs, in mouse CGNs (Davila et al., 2014). The human genome encodes 3 isoforms of

134

BIM; BIMEL, BIML and BIMg, however, their expression in HeLa cells is extremely low under basal conditions (Yamaguchi and Wang, 2002). Thus, future work in non-HeLa cell lines will be crucial in determining the binding capacity of

HSPB1 for ARE-mRNAs.

Regulation of mRNA stability by HSPB1 may present a novel mechanism for the pathogenesis of HSPB1-associated neuropathies. A recent study using dHMNII patient-derived lymphoblastoid cells identified a novel interaction between HSPB1 and poly(C)binding protein 1 (PCBP1), an RNA binding protein that recognizes C-rich elements or CREs. The authors report that mutant

HSPB1, specifically the HSPB1(P182L) variant, displayed an increased binding to PCBP1, leading to a loss of translational repression of genes associated with hereditary neuropathy (Geuens et al., 2017). Interestingly, PCBP1 also interacts with AUF1, and facilitates AUF1 binding to CREs (Hwang et al., 2017). Perhaps mutations in HSPB1 disrupt the normal interactions between HSPB1, PCBP1 and AUF1, causing a dysregulation of RNAs crucial in maintaining neuronal homeostasis. It will be interesting to see how these proteins function and interact with one another in neurons and the supporting cast of astrocytes, microglia and oligodendrocytes.

135

CONCLUSIONS

Here I examined the functional consequences of hereditary neuropathy associated mutations in the small heat shock protein, HSPB1. I reported a non- cell autonomous neuroprotective role for HSPB1 in an ALS co-culture system.

This neuroprotective effect was attenuated by mutations in HSPB1 that are associated with the development of late-onset distal hereditary motor neuropathy, implying that mutations in HSPB1 result in a novel loss of function in non-neuronal cells. My data suggests that the mechanism of HSPB1-mediated neuroprotection appears to be independent of HSPB1’s chaperone activity, as both wild-type and HSPB1(R136W), a variant that displays hyperactive chaperone activity, failed to alter the levels of intercellular and extracellular levels of SOD1. Microarray data generating using this model system indicate that

HSPB1’s neuroprotection may be mediated by its role in regulating inflammatory signaling. This is supported by the fact that a phosphomimetic variant of HSPB1, which exhibits little chaperone activity, but regulates inflammatory and apoptotic signaling, also demonstrated non-cell autonomous neuroprotection.

To this end, I investigated two possible mechanisms of inflammatory gene regulation, activation of the NF-kB signaling pathway, and alterations of mRNA stability through AU-rich element mRNA decay. Wild-type

136

HSPB1 reduced the activation of NF-kB signaling in HeLa cells, while mutant

HSPB1 was not. I was unable to identify the specific mechanism behind this reduction in NF-kB activity, but hypothesize it may be due to regulation of IkBa synthesis. I was unsuccessful at identifying a role for HSPB1 in AU-rich element mRNA decay, and therefore could not study the functional consequences of these mutations on this pathway. To gain better understanding of these pathways, I suggest that future studies be done in neuronal cell types, to identify relevant interactions involving HSPB1. Additionally, I recommend that studying the functional consequence of HSPB1 mutations should be done in patient- derived cell lines, to better model interactions that lead to the development of hereditary neuropathy.

I propose a mechanism by which mutations in HSPB1 alter both cell autonomous and non-cell autonomous processes, contributing to neuronal degeneration. mutHSPB1 disrupts axonal transport and ER stress response in neurons, making them more susceptible to damaging external stimuli. In non- neuronal cells, such as astrocytes, microglia and Schwann cells, mutHSPB1 disrupts the regulation of inflammatory signaling, which may convert a population of these cells from a resting non-inflammatory state to a more neurotoxic reactive state. Overtime, the chronically active glial cells wear down the already weakened neurons through a lack of neurotrophic support, and increased release of pro-inflammatory factors, leading to progressive cell loss. Further study is required however, to fully understand the role of both wild-type and mutHSPB1 in

137 inflammatory signaling, and how it may affect neurons in a non-cell autonomous fashion.

The discovery of novel functions for HSPB1 is just beginning, and future work in identifying all of roles HSPB1 plays will be crucial to understanding how mutations in this multifunctional protein give rise to the pathogenesis and progression of hereditary neuropathies. So little is known about HSPB1’s function in the nervous system, particularly in non-neuronal cells. Further understanding of HSPB1’s non-chaperone activities may also yield valuable knowledge for the development of new therapies targeting injury and disease of both the central and peripheral nervous system.

138

REFERENCES

Ackerley, S., et al., 2006. A mutation in the small heat-shock protein HSPB1 leading to distal hereditary motor neuronopathy disrupts neurofilament assembly and the axonal transport of specific cellular cargoes. Hum Mol Genet. 15, 347-54. Acunzo, J., et al., 2012. Small heat shock proteins HSP27 (HspB1), alphaB- crystallin (HspB5) and HSP22 (HspB8) as regulators of cell death. Int J Biochem Cell Biol. 44, 1622-31. Akbar, M. T., et al., 2001. Heat shock protein 27 shows a distinctive widespread spatial and temporal pattern of induction in CNS glial and neuronal cells compared to heat shock protein 70 and caspase 3 following kainate administration. Brain Res Mol Brain Res. 93, 148-63. Almeida-Souza, L., et al., 2011. Small heat-shock protein HSPB1 mutants stabilize microtubules in Charcot-Marie-Tooth neuropathy. J Neurosci. 31, 15320-8. Almeida-Souza, L., et al., 2013. HSPB1 facilitates the formation of non- centrosomal microtubules. PLoS One. 8, e66541. Almeida-Souza, L., et al., 2010. Increased monomerization of mutant HSPB1 leads to protein hyperactivity in Charcot-Marie-Tooth neuropathy. J Biol Chem. 285, 12778-86. An, J. J., et al., 2008. Transduced human PEP-1-heat shock protein 27 efficiently protects against brain ischemic insult. FEBS J. 275, 1296-308. Anderson, C. M., Swanson, R. A., 2000. Astrocyte glutamate transport: review of properties, regulation, and physiological functions. Glia. 32, 1-14. Arao, Y., et al., 2000. A nuclear matrix-associated factor, SAF-B, interacts with specific isoforms of AUF1/hnRNP D. Arch Biochem Biophys. 380, 228-36. Armstrong, C. L., et al., 2001. Constitutive expression of heat shock protein HSP25 in the central nervous system of the developing and adult mouse. J Comp Neurol. 434, 262-74. Awatsuji, H., et al., 1993. Interleukin-4 and -5 as modulators of nerve growth factor synthesis/secretion in astrocytes. J Neurosci Res. 34, 539-45. Badin, R. A., et al., 2006. Neuroprotective effects of virally delivered HSPs in experimental stroke. J Cereb Blood Flow Metab. 26, 371-81. Badin, R. A., et al., 2009. Protective effect of post-ischaemic viral delivery of heat shock proteins in vivo. J Cereb Blood Flow Metab. 29, 254-63. 139

Bak, L. K., et al., 2006. The glutamate/GABA-glutamine cycle: aspects of transport, neurotransmitter homeostasis and ammonia transfer. J Neurochem. 98, 641-53. Bakthisaran, R., et al., 2015. Small heat shock proteins: Role in cellular functions and pathology. Biochim Biophys Acta. 1854, 291-319. Baranova, E. V., et al., 2011. Three-dimensional structure of alpha-crystallin domain dimers of human small heat shock proteins HSPB1 and HSPB6. J Mol Biol. 411, 110-22. Basso, M., et al., 2013. Mutant copper-zinc superoxide dismutase (SOD1) induces protein secretion pathway alterations and exosome release in astrocytes: implications for disease spreading and motor neuron pathology in amyotrophic lateral sclerosis. J Biol Chem. 288, 15699-711. Beesoo, R., et al., 2014. Apoptosis inducing lead compounds isolated from marine organisms of potential relevance in cancer treatment. Mutat Res. 768, 84-97. Belanger, M., Magistretti, P. J., 2009. The role of astroglia in neuroprotection. Dialogues Clin Neurosci. 11, 281-95. Benkler, C., et al., 2013. Altered astrocytic response to activation in SOD1(G93A) mice and its implications on amyotrophic lateral sclerosis pathogenesis. Glia. 61, 312-26. Benn, S. C., et al., 2002. Hsp27 upregulation and phosphorylation is required for injured sensory and motor neuron survival. Neuron. 36, 45-56. Boillee, S., et al., 2006. Onset and progression in inherited ALS determined by motor neurons and microglia. Science. 312, 1389-92. Bonizzi, G., Karin, M., 2004. The two NF-kappaB activation pathways and their role in innate and adaptive immunity. Trends Immunol. 25, 280-8. Bouhy, D., et al., 2016. Characterization of New Transgenic Mouse Models for Two Charcot-Marie-Tooth-Causing HspB1 Mutations using the Rosa26 Locus. J Neuromuscul Dis. 3, 183-200. Bova, M. P., et al., 2000. Subunit exchange of small heat shock proteins. Analysis of oligomer formation of alphaA-crystallin and Hsp27 by fluorescence resonance energy transfer and site-directed truncations. J Biol Chem. 275, 1035-42. Brennan, C. M., Steitz, J. A., 2001. HuR and mRNA stability. Cell Mol Life Sci. 58, 266-77. Brewer, G., 1991. An A + U-rich element RNA-binding factor regulates c-myc mRNA stability in vitro. Mol Cell Biol. 11, 2460-6. Brodie, C., et al., 1998. Functional IL-4 receptors on mouse astrocytes: IL-4 inhibits astrocyte activation and induces NGF secretion. J Neuroimmunol. 81, 20-30. Bruey, J. M., et al., 2000a. Hsp27 negatively regulates cell death by interacting with cytochrome c. Nat Cell Biol. 2, 645-52. Bruey, J. M., et al., 2000b. Differential regulation of HSP27 oligomerization in tumor cells grown in vitro and in vivo. Oncogene. 19, 4855-63. 140

Bryantsev, A. L., et al., 2007. Regulation of stress-induced intracellular sorting and chaperone function of Hsp27 (HspB1) in mammalian cells. Biochem J. 407, 407-17. Burut, D. F., et al., 2010. Serum heat shock protein 27 antigen and antibody levels appear to be related to the macrovascular complications associated with insulin resistance: a pilot study. Cell Stress Chaperones. 15, 379-86. Butt, E., et al., 2001. Heat shock protein 27 is a substrate of cGMP-dependent protein kinase in intact human platelets: phosphorylation-induced actin polymerization caused by HSP27 mutants. J Biol Chem. 276, 7108-13. Cappello, F., et al., 2008. Hsp60 expression, new locations, functions and perspectives for cancer diagnosis and therapy. Cancer Biol Ther. 7, 801-9. Capponi, S., et al., 2011. HSPB1 and HSPB8 in inherited neuropathies: study of an Italian cohort of dHMN and CMT2 patients. J Peripher Nerv Syst. 16, 287-94. Capponi, S., et al., 2016. Molecular Chaperones in the Pathogenesis of Amyotrophic Lateral Sclerosis: The Role of HSPB1. Hum Mutat. 37, 1202- 1208. Charette, S. J., Landry, J., 2000. The interaction of HSP27 with Daxx identifies a potential regulatory role of HSP27 in Fas-induced apoptosis. Ann N Y Acad Sci. 926, 126-31. Charette, S. J., et al., 2000. Inhibition of Daxx-mediated apoptosis by heat shock protein 27. Mol Cell Biol. 20, 7602-12. Cheadle, C., et al., 2005a. Control of gene expression during T cell activation: alternate regulation of mRNA transcription and mRNA stability. BMC Genomics. 6, 75. Cheadle, C., et al., 2005b. Stability regulation of mRNA and the control of gene expression. Ann N Y Acad Sci. 1058, 196-204. Chen, C. Y., Shyu, A. B., 1995. AU-rich elements: characterization and importance in mRNA degradation. Trends Biochem Sci. 20, 465-70. Chen, S., Brown, I. R., 2007. Neuronal expression of constitutive heat shock proteins: implications for neurodegenerative diseases. Cell Stress & Chaperones. 12, 51. Chen, Y., et al., 2001. Astrocytes protect neurons from nitric oxide toxicity by a glutathione-dependent mechanism. J Neurochem. 77, 1601-10. Chiao, P. J., et al., 1994. Autoregulation of I kappa B alpha activity. Proc Natl Acad Sci U S A. 91, 28-32. Choi, S. S., et al., 2014. Human astrocytes: secretome profiles of cytokines and chemokines. PLoS One. 9, e92325. Citri, A., et al., 2006. Hsp90 recognizes a common surface on client kinases. J Biol Chem. 281, 14361-9. Clarke, J. P., Mearow, K. M., 2013. Cell stress promotes the association of phosphorylated HspB1 with F-actin. PLoS One. 8, e68978. Colombo, E., Farina, C., 2016. Astrocytes: Key Regulators of Neuroinflammation. Trends Immunol. 37, 608-20. 141

Cooper, L. F., et al., 2000. Transcriptional regulation involving the intronic heat shock element of the rat hsp27 gene. Biochim Biophys Acta. 1490, 348- 54. Costigan, M., et al., 1998. Heat shock protein 27: developmental regulation and expression after peripheral nerve injury. J Neurosci. 18, 5891-900. d'Ydewalle, C., et al., 2011. HDAC6 inhibitors reverse axonal loss in a mouse model of mutant HSPB1-induced Charcot-Marie-Tooth disease. Nat Med. 17, 968-74. Davila, D., et al., 2014. Hsp27 binding to the 3'UTR of bim mRNA prevents neuronal death during oxidative stress-induced injury: a novel cytoprotective mechanism. Mol Biol Cell. 25, 3413-23. De, A. K., et al., 2000. Exaggerated human monocyte IL-10 concomitant to minimal TNF-alpha induction by heat-shock protein 27 (Hsp27) suggests Hsp27 is primarily an antiinflammatory stimulus. J Immunol. 165, 3951-8. de Thonel, A., et al., 2012. Transcriptional regulation of small HSP-HSF1 and beyond. Int J Biochem Cell Biol. 44, 1593-612. Deane, C. A., Brown, I. R., 2016. Induction of heat shock proteins in differentiated human neuronal cells following co-application of celastrol and arimoclomol. Cell Stress Chaperones. 21, 837-48. DeMaria, C. T., Brewer, G., 1996. AUF1 binding affinity to A+U-rich elements correlates with rapid mRNA degradation. J Biol Chem. 271, 12179-84. Derecki, N. C., et al., 2010. Regulation of learning and memory by meningeal immunity: a key role for IL-4. J Exp Med. 207, 1067-80. Diaz-Latoud, C., et al., 2005. Substitution of the unique cysteine residue of murine Hsp25 interferes with the protective activity of this stress protein through inhibition of dimer formation. Antioxid Redox Signal. 7, 436-45. Dodd, S. L., et al., 2009. Hsp27 inhibits IKKbeta-induced NF-kappaB activity and skeletal muscle atrophy. FASEB J. 23, 3415-23. Dodge, J. C., et al., 2008. Delivery of AAV-IGF-1 to the CNS extends survival in ALS mice through modification of aberrant glial cell activity. Mol Ther. 16, 1056-64. Dong, Y., Benveniste, E. N., 2001. Immune function of astrocytes. Glia. 36, 180- 90. Doppler, H., et al., 2005. A phosphorylation state-specific antibody recognizes Hsp27, a novel substrate of protein kinase D. J Biol Chem. 280, 15013-9. Doyle, S. M., et al., 2013. Protein rescue from aggregates by powerful molecular chaperone machines. Nat Rev Mol Cell Biol. 14, 617-29. Dringen, R., et al., 1999. Synthesis of the antioxidant glutathione in neurons: supply by astrocytes of CysGly as precursor for neuronal glutathione. J Neurosci. 19, 562-9. Ehrnsperger, M., et al., 1997. Binding of non-native protein to Hsp25 during heat shock creates a reservoir of folding intermediates for reactivation. EMBO J. 16, 221-9.

142

Elmore, S., 2007. Apoptosis: a review of programmed cell death. Toxicol Pathol. 35, 495-516. Evgrafov, O. V., et al., 2004. Mutant small heat-shock protein 27 causes axonal Charcot-Marie-Tooth disease and distal hereditary motor neuropathy. Nat Genet. 36, 602-6. Fagerlund, R., et al., 2015. Anatomy of a negative feedback loop: the case of IkappaBalpha. J R Soc Interface. 12, 0262. Fan, J., et al., 2002. Global analysis of stress-regulated mRNA turnover by using cDNA arrays. Proc Natl Acad Sci U S A. 99, 10611-6. Farina, C., et al., 2007. Astrocytes are active players in cerebral innate immunity. Trends Immunol. 28, 138-45. Ferraiuolo, L., et al., 2011. Molecular pathways of motor neuron injury in amyotrophic lateral sclerosis. Nat Rev Neurol. 7, 616-30. Fletcher, D. A., Mullins, R. D., 2010. Cell mechanics and the cytoskeleton. Nature. 463, 485-92. Frakes, A. E., et al., 2014. Microglia induce motor neuron death via the classical NF-kappaB pathway in amyotrophic lateral sclerosis. Neuron. 81, 1009- 23. Fritz, E., et al., 2013. Mutant SOD1-expressing astrocytes release toxic factors that trigger motoneuron death by inducing hyperexcitability. J Neurophysiol. 109, 2803-14. Gerondakis, S., et al., 1999. Genetic approaches in mice to understand Rel/NF- kappaB and IkappaB function: transgenics and knockouts. Oncogene. 18, 6888-95. Geuens, T., et al., 2017. Mutant HSPB1 causes loss of translational repression by binding to PCBP1, an RNA binding protein with a possible role in neurodegenerative disease. Acta Neuropathol Commun. 5, 5. Gherzi, R., et al., 2004. A KH domain RNA binding protein, KSRP, promotes ARE-directed mRNA turnover by recruiting the degradation machinery. Mol Cell. 14, 571-83. Gibb, W. R., Lees, A. J., 1988. The relevance of the Lewy body to the pathogenesis of idiopathic Parkinson's disease. J Neurol Neurosurg Psychiatry. 51, 745-52. Goldstein, L. S., Yang, Z., 2000. Microtubule-based transport systems in neurons: the roles of kinesins and dyneins. Annu Rev Neurosci. 23, 39-71. Gratacos, F. M., Brewer, G., 2010. The role of AUF1 in regulated mRNA decay. Wiley Interdiscip Rev RNA. 1, 457-73. Gros-Louis, F., et al., 2010. Intracerebroventricular infusion of monoclonal antibody or its derived Fab fragment against misfolded forms of SOD1 mutant delays mortality in a mouse model of ALS. J Neurochem. 113, 1188-99. Guo, K., et al., 2009. Regulation of HSP27 on NF-kappaB pathway activation may be involved in metastatic hepatocellular carcinoma cells apoptosis. BMC Cancer. 9, 100. 143

Hadadi, E., et al., 2016. Differential IL-1beta secretion by monocyte subsets is regulated by Hsp27 through modulating mRNA stability. Sci Rep. 6, 39035. Haidet-Phillips, A. M., et al., 2011. Astrocytes From Familial and Sporadic ALS Patients are Toxic to Motor Neurons. Nature Biotechnology. 29. Halees, A. S., et al., 2008. ARED Organism: expansion of ARED reveals AU-rich element cluster variations between human and mouse. Nucleic Acids Research. 36, D137-D140. Han, S., et al., 2012. Resveratrol upregulated heat shock proteins and extended the survival of G93A-SOD1 mice. Brain Res. 1483, 112-7. Haslbeck, M., 2002. sHsps and their role in the chaperone network. Cell Mol Life Sci. 59, 1649-57. Haslbeck, M., Buchner, J., 2002. Chaperone function of sHsps. Prog Mol Subcell Biol. 28, 37-59. Haslbeck, M., et al., 2005. Some like it hot: the structure and function of small heat-shock proteins. Nat Struct Mol Biol. 12, 842-6. Havasi, A., et al., 2008. Hsp27 inhibits Bax activation and apoptosis via a phosphatidylinositol 3-kinase-dependent mechanism. J Biol Chem. 283, 12305-13. Hayden, M. S., Ghosh, S., 2004. Signaling to NF-kappaB. Genes Dev. 18, 2195- 224. Hayes, D., et al., 2009. Phosphorylation dependence of hsp27 multimeric size and molecular chaperone function. J Biol Chem. 284, 18801-7. Hecker, J. G., McGarvey, M., 2011. Heat shock proteins as biomarkers for the rapid detection of brain and spinal cord ischemia: a review and comparison to other methods of detection in thoracic aneurysm repair. Cell Stress Chaperones. 16, 119-31. Helfand, B. T., et al., 2004. Intermediate filaments are dynamic and motile elements of cellular architecture. J Cell Sci. 117, 133-41. Henkel, T., et al., 1992. Intramolecular masking of the nuclear location signal and dimerization domain in the precursor for the p50 NF-kappa B subunit. Cell. 68, 1121-33. Hensley, K., et al., 2006. Primary glia expressing the G93A-SOD1 mutation present a neuroinflammatory phenotype and provide a cellular system for studies of glial inflammation. J Neuroinflammation. 3, 2. Hickey, E., et al., 1986. Sequence and organization of genes encoding the human 27 kDa heat shock protein. Nucleic Acids Res. 14, 4127-45. Hill, M. M., et al., 2004. Analysis of the composition, assembly kinetics and activity of native Apaf-1 apoptosomes. EMBO J. 23, 2134-45. Hirata, K., et al., 2003. HSP27 is markedly induced in Schwann cell columns and associated regenerating axons. Glia. 42, 1-11. Hirokawa, N., 1998. Kinesin and dynein superfamily proteins and the mechanism of organelle transport. Science. 279, 519-26.

144

Hoch, B., et al., 1996. HSP25 in isolated perfused rat hearts: localization and response to hyperthermia. Mol Cell Biochem. 160-161, 231-9. Horwitz, J., 1992. Alpha-crystallin can function as a molecular chaperone. Proc Natl Acad Sci U S A. 89, 10449-53. Hu, S., et al., 1995. Differential regulation by cytokines of human astrocyte nitric oxide production. Glia. 15, 491-4. Huot, J., et al., 1996. HSP27 phosphorylation-mediated resistance against actin fragmentation and cell death induced by oxidative stress. Cancer Res. 56, 273-9. Hwang, C. K., et al., 2017. Phosphorylation of poly(rC) binding protein 1 (PCBP1) contributes to stabilization of mu opioid receptor (MOR) mRNA via interaction with AU-rich element RNA-binding protein 1 (AUF1) and poly A binding protein (PABP). Gene. 598, 113-130. Ikeda, Y., et al., 2009. A clinical phenotype of distal hereditary motor neuronopathy type II with a novel HSPB1 mutation. J Neurol Sci. 277, 9- 12. Israelson, A., et al., 2015. Macrophage migration inhibitory factor as a chaperone inhibiting accumulation of misfolded SOD1. Neuron. 86, 218-32. Iwaki, T., et al., 1989. Alpha B-crystallin is expressed in non-lenticular tissues and accumulates in Alexander's disease brain. Cell. 57, 71-8. Jakob, U., et al., 1993. Small heat shock proteins are molecular chaperones. J Biol Chem. 268, 1517-20. Jin An, J. e. a., 2008. Transduced HSP27 Protein Protects Neuronal Cell Death by Enhancing FALS-associated SOD1 Mutant Activity. Jin, C., et al., 2014. Human myocardium releases heat shock protein 27 (HSP27) after global ischemia: the proinflammatory effect of extracellular HSP27 through toll-like receptor (TLR)-2 and TLR4. Mol Med. 20, 280-9. Kalmar, B., et al., 2008. Late stage treatment with arimoclomol delays disease progression and prevents protein aggregation in the SOD1 mouse model of ALS. J Neurochem. 107, 339-50. Kammanadiminti, S. J., Chadee, K., 2006. Suppression of NF-kappaB activation by Entamoeba histolytica in intestinal epithelial cells is mediated by heat shock protein 27. J Biol Chem. 281, 26112-20. Kampinga, H. H., Craig, E. A., 2010. The HSP70 chaperone machinery: J proteins as drivers of functional specificity. Nat Rev Mol Cell Biol. 11, 579- 92. Kamradt, M. C., et al., 2005. The small heat shock protein alpha B-crystallin is a novel inhibitor of TRAIL-induced apoptosis that suppresses the activation of caspase-3. J Biol Chem. 280, 11059-66. Kappe, G., et al., 2003. The human genome encodes 10 alpha-crystallin-related small heat shock proteins: HspB1-10. Cell Stress Chaperones. 8, 53-61. Kato, H., et al., 1995a. An immunohistochemical study of heat shock protein-27 in the hippocampus in a gerbil model of cerebral ischemia and ischemic tolerance. Neuroscience. 68, 65-71. 145

Kato, H., et al., 1995b. Immunohistochemical localization of the low molecular weight stress protein HSP27 following focal cerebral ischemia in the rat. Brain Res. 679, 1-7. Kato, H., et al., 1994. Induction of 27-kDa heat shock protein following cerebral ischemia in a rat model of ischemic tolerance. Brain Res. 634, 235-44. Kato, K., et al., 1992. Copurification of small heat shock protein with alpha B crystallin from human skeletal muscle. J Biol Chem. 267, 7718-25. Kato, S., et al., 1997. Pathological characterization of astrocytic hyaline inclusions in familial amyotrophic lateral sclerosis. American Journal of Pathology. 151, 611-620. Katsuno, M., et al., 2012. Perspectives on molecular targeted therapies and clinical trials for neurodegenerative diseases. J Neurol Neurosurg Psychiatry. 83, 329-35. Keeler, B. E., et al., 2012. Acute and prolonged hindlimb exercise elicits different gene expression in motoneurons than sensory neurons after spinal cord injury. Brain Res. 1438, 8-21. Kiaei, M., et al., 2005. Celastrol blocks neuronal cell death and extends life in transgenic mouse model of amyotrophic lateral sclerosis. Neurodegener Dis. 2, 246-54. Kieran, D., et al., 2004. Treatment with arimoclomol, a coinducer of heat shock proteins, delays disease progression in ALS mice. Nat Med. 10, 402-5. Kijima, K., et al., 2005. Small heat shock protein 27 mutation in a Japanese patient with distal hereditary motor neuropathy. J Hum Genet. 50, 473-6. Kim, J., et al., 2013. TRAIL/MEKK4/p38/HSP27/Akt survival network is biphasically modulated by the Src/CIN85/c-Cbl complex. Cell Signal. 25, 372-9. Kim, J., et al., 2012. HSP27 modulates survival signaling networks in cells treated with curcumin and TRAIL. Cell Signal. 24, 1444-52. Knapinska, A. M., et al., 2011. Chaperone Hsp27 modulates AUF1 proteolysis and AU-rich element-mediated mRNA degradation. Mol Cell Biol. 31, 1419-31. Kolb, S. J., et al., 2010. Mutant small heat shock protein B3 causes motor neuropathy: utility of a candidate gene approach. Neurology. 74, 502-6. Kostenko, S., Moens, U., 2009. Heat shock protein 27 phosphorylation: kinases, phosphatases, functions and pathology. Cell Mol Life Sci. 66, 3289-307. Kriehuber, T., et al., 2010. Independent evolution of the core domain and its flanking sequences in small heat shock proteins. FASEB J. 24, 3633-42. Krishnan, J., et al., 2008. Over-expression of Hsp27 does not influence disease in the mutant SOD1(G93A) mouse model of amyotrophic lateral sclerosis. J Neurochem. 106, 2170-83. Krueger-Naug, A. M., et al., 2000. Hyperthermic induction of the 27-kDa heat shock protein (Hsp27) in neuroglia and neurons of the rat central nervous system. J Comp Neurol. 428, 495-510.

146

Lafon-Cazal, M., et al., 2003. Proteomic analysis of astrocytic secretion in the mouse. Comparison with the cerebrospinal fluid proteome. J Biol Chem. 278, 24438-48. Lambert, H., et al., 1999. HSP27 multimerization mediated by phosphorylation- sensitive intermolecular interactions at the amino terminus. J Biol Chem. 274, 9378-85. Lasa, M., et al., 2000. Regulation of cyclooxygenase 2 mRNA stability by the mitogen-activated protein kinase p38 signaling cascade. Mol Cell Biol. 20, 4265-74. Lavoie, J. N., et al., 1995. Modulation of cellular thermoresistance and actin filament stability accompanies phosphorylation-induced changes in the oligomeric structure of heat shock protein 27. Mol Cell Biol. 15, 505-16. Lee, J., et al., 2015. Overexpression of mutant HSP27 causes axonal neuropathy in mice. J Biomed Sci. 22, 43. Lee, S., et al., 2006. Small heat shock proteins differentially affect Abeta aggregation and toxicity. Biochem Biophys Res Commun. 347, 527-33. Lelj-Garolla, B., Mauk, A. G., 2005. Self-association of a small heat shock protein. J Mol Biol. 345, 631-42. Lelj-Garolla, B., Mauk, A. G., 2006. Self-association and chaperone activity of Hsp27 are thermally activated. J Biol Chem. 281, 8169-74. Lelj-Garolla, B., Mauk, A. G., 2012. Roles of the N- and C-terminal sequences in Hsp27 self-association and chaperone activity. Protein Sci. 21, 122-33. Leverenz, J. B., et al., 2007. Proteomic identification of novel proteins in cortical lewy bodies. Brain Pathol. 17, 139-45. Lewis, S. E., et al., 1999. A role for HSP27 in sensory neuron survival. J Neurosci. 19, 8945-53. Leyton-Jaimes, M. F., et al., 2016. Endogenous macrophage migration inhibitory factor reduces the accumulation and toxicity of misfolded SOD1 in a mouse model of ALS. Proc Natl Acad Sci U S A. 113, 10198-203. Li, R., et al., 2005. Glial fibrillary acidic protein mutations in infantile, juvenile, and adult forms of Alexander disease. Ann Neurol. 57, 310-26. Liu, J., et al., 2010. Regulation of lipopolysaccharide-induced inflammatory response by heat shock protein 27 in THP-1 cells. Cell Immunol. 264, 127- 34. Locksley, R. M., et al., 2001. The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell. 104, 487-501. Luigetti, M., et al., 2016. Charcot-Marie-Tooth type 2 and distal hereditary motor neuropathy: Clinical, neurophysiological and genetic findings from a single-centre experience. Clin Neurol Neurosurg. 144, 67-71. Lutsch, G., et al., 1997. Abundance and location of the small heat shock proteins HSP25 and alphaB-crystallin in rat and human heart. Circulation. 96, 3466-76.

147

Lykke-Andersen, J., Wagner, E., 2005. Recruitment and activation of mRNA decay enzymes by two ARE-mediated decay activation domains in the proteins TTP and BRF-1. Genes Dev. 19, 351-61. Ma, C. H., et al., 2011. Accelerating axonal growth promotes motor recovery after peripheral nerve injury in mice. J Clin Invest. 121, 4332-47. Ma, W. J., et al., 1997. The Elav-like proteins bind to AU-rich elements and to the poly(A) tail of mRNA. Nucleic Acids Res. 25, 3564-9. Maatkamp, A., et al., 2004. Decrease of Hsp25 protein expression precedes degeneration of motoneurons in ALS-SOD1 mice. Eur J Neurosci. 20, 14- 28. Mann, D. M., 1996. Pyramidal nerve cell loss in Alzheimer's disease. Neurodegeneration. 5, 423-7. Mayer, M. P., 2013. Hsp70 chaperone dynamics and molecular mechanism. Trends Biochem Sci. 38, 507-14. Mayer, M. P., Bukau, B., 2005. Hsp70 chaperones: cellular functions and molecular mechanism. Cell Mol Life Sci. 62, 670-84. McClaren, M., Isseroff, R. R., 1994. Dynamic changes in intracellular localization and isoforms of the 27-kD stress protein in human keratinocytes. J Invest Dermatol. 102, 375-81. McDonald, E. T., et al., 2012. Sequence, structure, and dynamic determinants of Hsp27 (HspB1) equilibrium dissociation are encoded by the N-terminal domain. Biochemistry. 51, 1257-68. McDonough, H., Patterson, C., 2003. CHIP: a link between the chaperone and proteasome systems. Cell Stress Chaperones. 8, 303-8. McLean, P. J., et al., 2004. Geldanamycin induces Hsp70 and prevents alpha- synuclein aggregation and toxicity in vitro. Biochem Biophys Res Commun. 321, 665-9. Mehlen, P., et al., 1995a. Tumor necrosis factor-alpha induces changes in the phosphorylation, cellular localization, and oligomerization of human hsp27, a stress protein that confers cellular resistance to this cytokine. J Cell Biochem. 58, 248-59. Mehlen, P., et al., 1995b. Constitutive expression of human hsp27, Drosophila hsp27, or human alpha B-crystallin confers resistance to TNF- and oxidative stress-induced cytotoxicity in stably transfected murine L929 fibroblasts. J Immunol. 154, 363-74. Meimaridou, E., et al., 2009. From hatching to dispatching: the multiple cellular roles of the Hsp70 molecular chaperone machinery. J Mol Endocrinol. 42, 1-9. Meyer, K., et al., 2014. Direct conversion of patient fibroblasts demonstrates non- cell autonomous toxicity of astrocytes to motor neurons in familial and sporadic ALS. PNAS. 111. Migheli, A., et al., 1997. c-Jun, JNK/SAPK kinases and transcription factor NF- kappa B are selectively activated in astrocytes, but not motor neurons, in amyotrophic lateral sclerosis. J Neuropathol Exp Neurol. 56, 1314-22. 148

Miranda, C. J., et al., 2012. Aging brain microenvironment decreases hippocampal neurogenesis through Wnt-mediated survivin signaling. Aging Cell. 11, 542-52. Miron, T., et al., 1991. A 25-kD inhibitor of actin polymerization is a low molecular mass heat shock protein. J Cell Biol. 114, 255-61. Miron, T., et al., 1988. Characterization of an inhibitor of actin polymerization in vinculin-rich fraction of turkey gizzard smooth muscle. Eur J Biochem. 178, 543-53. Misselwitz, B., et al., 1998. J proteins catalytically activate Hsp70 molecules to trap a wide range of peptide sequences. Mol Cell. 2, 593-603. Morimoto, R. I., 2011. The heat shock response: systems biology of proteotoxic stress in aging and disease. Cold Spring Harb Symp Quant Biol. 76, 91-9. Muranova, L. K., et al., 2015. Characterization of Mutants of Human Small Heat Shock Protein HspB1 Carrying Replacements in the N-Terminal Domain and Associated with Hereditary Motor Neuron Diseases. PLoS One. 10, e0126248. Mymrikov, E. V., et al., 2011. Large potentials of small heat shock proteins. Physiol Rev. 91, 1123-59. Nafar, F., et al., 2016. Astrocytes release HspB1 in response to amyloid-beta exposure in vitro. J Alzheimers Dis. 49, 251-63. Nagai, M., et al., 2007. Astrocytes expressing ALS-linked mutated SOD1 release factors selectively toxic to motor neurons. Nat Neurosci. 10, 615-22. Nefedova, V. V., et al., 2013a. Physico-chemical properties of R140G and K141Q mutants of human small heat shock protein HspB1 associated with hereditary peripheral neuropathies. Biochimie. 95, 1582-92. Nefedova, V. V., et al., 2013b. Structure and properties of G84R and L99M mutants of human small heat shock protein HspB1 correlating with motor neuropathy. Arch Biochem Biophys. 538, 16-24. New, L., et al., 1998. PRAK, a novel protein kinase regulated by the p38 MAP kinase. EMBO J. 17, 3372-84. Nivon, M., et al., 2009. Autophagy activation by NFkappaB is essential for cell survival after heat shock. Autophagy. 5, 766-83. Nizzardo, M., et al., 2016. Morpholino-mediated SOD1 reduction ameliorates an amyotrophic lateral sclerosis disease phenotype. Sci Rep. 6, 21301. O'Shaughnessy, R. F., et al., 2007. AKT-dependent HspB1 (Hsp27) activity in epidermal differentiation. J Biol Chem. 282, 17297-305. Oda, T., et al., 2009. The nature of the globular- to fibrous-actin transition. Nature. 457, 441-5. Oesterreich, S., et al., 1996. Basal regulatory promoter elements of the hsp27 gene in human breast cancer cells. Biochem Biophys Res Commun. 222, 155-63. Outeiro, T. F., et al., 2006. Small heat shock proteins protect against alpha- synuclein-induced toxicity and aggregation. Biochem Biophys Res Commun. 351, 631-8. 149

Pahl, H. L., 1999. Activators and target genes of Rel/NF-kappaB transcription factors. Oncogene. 18, 6853-66. Palanisamy, V., et al., 2012. Control of cytokine mRNA expression by RNA- binding proteins and microRNAs. J Dent Res. 91, 651-8. Pandey, P., et al., 2000. Hsp27 functions as a negative regulator of cytochrome c-dependent activation of procaspase-3. Oncogene. 19, 1975-81. Papadeas, S. T., et al., 2011. Astrocytes carrying the superoxide dismutase 1 (SOD1G93A) mutation induce wild-type motor neuron degeneration in vivo. PNAS. 108. Parcellier, A., et al., 2003. HSP27 Is a Ubiquitin-Binding Protein Involved in I- B Proteasomal Degradation. Molecular and Cellular Biology. 23, 5790-5802. Park, K. J., et al., 2003. Heat shock protein 27 association with the I kappa B kinase complex regulates tumor necrosis factor alpha-induced NF-kappa B activation. J Biol Chem. 278, 35272-8. Pasparakis, M., et al., 2006. Dissection of the NF-kappaB signalling cascade in transgenic and knockout mice. Cell Death Differ. 13, 861-72. Patel, P., et al., 2015. Early-stage treatment with Withaferin A reduces levels of misfolded superoxide dismutase 1 and extends lifespan in a mouse model of amyotrophic lateral sclerosis. Neurotherapeutics. 12, 217-33. Patel, Y. J., et al., 2005. Hsp27 and Hsp70 administered in combination have a potent protective effect against FALS-associated SOD1-mutant-induced cell death in mammalian neuronal cells. Brain Res Mol Brain Res. 134, 256-74. Paul, C., et al., 2002. Hsp27 as a negative regulator of cytochrome C release. Mol Cell Biol. 22, 816-34. Paul, C., et al., 2010. Dynamic processes that reflect anti-apoptotic strategies set up by HspB1 (Hsp27). Exp Cell Res. 316, 1535-52. Peferoen, L. A., et al., 2015. Small heat shock proteins are induced during multiple sclerosis lesion development in white but not grey matter. Acta Neuropathol Commun. 3, 87. Pivovarova, A. V., et al., 2007. Small heat shock protein Hsp27 prevents heat- induced aggregation of F-actin by forming soluble complexes with denatured actin. FEBS J. 274, 5937-48. Pivovarova, A. V., et al., 2005. Effects of small heat shock proteins on the thermal denaturation and aggregation of F-actin. Biochem Biophys Res Commun. 331, 1548-53. Plumier, J. C., et al., 1997. Constitutive expression of the 27-kDa heat shock protein (Hsp27) in sensory and motor neurons of the rat nervous system. J Comp Neurol. 384, 409-28. Polymeropoulos, M. H., et al., 1997. Mutation in the alpha-synuclein gene identified in families with Parkinson's disease. Science. 276, 2045-7. Pramatarova, A., et al., 2001. Neuron-specific expression of mutant superoxide dismutase 1 in transgenic mice does not lead to motor impairment. J Neurosci. 21, 3369-74. 150

Pratt, W. B., et al., 2010. Proposal for a role of the Hsp90/Hsp70-based chaperone machinery in making triage decisions when proteins undergo oxidative and toxic damage. Exp Biol Med (Maywood). 235, 278-89. Quraishe, S., et al., 2008. Expression of the small heat shock protein family in the mouse CNS: differential anatomical and biochemical compartmentalization. Neuroscience. 153, 483-91. Ralph, G. S., et al., 2005. Silencing mutant SOD1 using RNAi protects against neurodegeneration and extends survival in an ALS model. Nat Med. 11, 429-33. Rane, M. J., et al., 2003. Heat shock protein 27 controls apoptosis by regulating Akt activation. J Biol Chem. 278, 27828-35. Ray, J., Gage, F. H., 2006. Differential properties of adult rat and mouse brain- derived neural stem/progenitor cells. Mol Cell Neurosci. 31, 560-73. Rayner, K., et al., 2008. Extracellular release of the atheroprotective heat shock protein 27 is mediated by estrogen and competitively inhibits acLDL binding to scavenger receptor-A. Circ Res. 103, 133-41. Rayner, K., et al., 2010. Heat shock protein 27: clue to understanding estrogen- mediated atheroprotection? Trends Cardiovasc Med. 20, 54-8. Renkawek, K., et al., 1994. Expression of small heat-shock protein hsp 27 in reactive gliosis in Alzheimer disease and other types of dementia. Acta Neuropathol. 87, 511-9. Renkawek, K., et al., 1993. Increased expression of heat-shock protein 27 kDa in Alzheimer disease: a preliminary study. Neuroreport. 5, 14-6. Richter, K., et al., 2010. The heat shock response: life on the verge of death. Mol Cell. 40, 253-66. Richter-Landsberg, C., 2007. Heat Shock Proteins in Neural Cells. Ritossa, F., 1962. A new puffing pattern induced by temperature shock and DNP in Drosophilia. Experientia. 18, 571-573. Rogalla, T., et al., 1999. Regulation of Hsp27 oligomerization, chaperone function, and protective activity against oxidative stress/tumor necrosis factor alpha by phosphorylation. J Biol Chem. 274, 18947-56. Rojas, F., et al., 2014. Astrocytes expressing mutant SOD1 and TDP43 trigger motoneuron death that is mediated via sodium channels and nitroxidative stress. Front Cell Neurosci. 8, 24. Rosen, D. R., et al., 1993. Mutations in Cu/Zn superoxide dismutase gene are associated with familial amyotrophic lateral sclerosis. Nature. 362, 59-62. Rossor, A. M., et al., 2012a. A novel p.Gln175X [corrected] premature stop mutation in the C-terminal end of HSP27 is a cause of CMT2. J Peripher Nerv Syst. 17, 201-5. Rossor, A. M., et al., 2012b. The distal hereditary motor neuropathies. J Neurol Neurosurg Psychiatry. 83, 6-14. Rossor, A. M., et al., 2016. Pilot phenotype and natural history study of hereditary neuropathies caused by mutations in the HSPB1 gene. Neuromuscul Disord. 151

Saibil, H., 2013. Chaperone machines for protein folding, unfolding and disaggregation. Nat Rev Mol Cell Biol. 14, 630-42. Salari, S., et al., 2013. Extracellular HSP27 acts as a signaling molecule to activate NF-kappaB in macrophages. Cell Stress Chaperones. 18, 53-63. Sarkar, B., et al., 2003. Selective degradation of AU-rich mRNAs promoted by the p37 AUF1 protein isoform. Mol Cell Biol. 23, 6685-93. Sattler, R., Tymianski, M., 2001. Molecular mechanisms of glutamate receptor- mediated excitotoxic neuronal cell death. Mol Neurobiol. 24, 107-29. Scherer, S. S., 2011. The debut of a rational treatment for an inherited neuropathy? J Clin Invest. 121, 4624-7. Schmittgen, T. D., Livak, K. J., 2008. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc. 3, 1101-8. Sharp, P. S., et al., 2008. Protective effects of heat shock protein 27 in a model of ALS occur in the early stages of disease progression. Neurobiol Dis. 30, 42-55. Shashidharamurthy, R., et al., 2005. Mechanism of chaperone function in small heat shock proteins: dissociation of the HSP27 oligomer is required for recognition and binding of destabilized T4 lysozyme. J Biol Chem. 280, 5281-9. Shaw, G., Kamen, R., 1986. A conserved AU sequence from the 3' untranslated region of GM-CSF mRNA mediates selective mRNA degradation. Cell. 46, 659-67. Shchors, K., et al., 2002. Cell death inhibiting RNA (CDIR) derived from a 3'- untranslated region binds AUF1 and heat shock protein 27. J Biol Chem. 277, 47061-72. Shen, L., et al., 2016. Phosphorylated heat shock protein 27 promotes lipid clearance in hepatic cells through interacting with STAT3 and activating autophagy. Cell Signal. 28, 1086-98. Shih, A. Y., et al., 2006. Cystine/glutamate exchange modulates glutathione supply for neuroprotection from oxidative stress and cell proliferation. J Neurosci. 26, 10514-23. Sievers, F., et al., 2011. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol. 7, 539. Sinsimer, K. S., et al., 2008. Chaperone Hsp27, a novel subunit of AUF1 protein complexes, functions in AU-rich element-mediated mRNA decay. Mol Cell Biol. 28, 5223-37. Skouri-Panet, F., et al., 2012. Structural and functional specificity of small heat shock protein HspB1 and HspB4, two cellular partners of HspB5: role of the in vitro hetero-complex formation in chaperone activity. Biochimie. 94, 975-84. Skre, H., 1974. Genetic and clinical aspects of Charcot-Marie-Tooth's disease. Clin Genet. 6, 98-118.

152

Srivastava, A. K., et al., 2012. Mutant HSPB1 overexpression in neurons is sufficient to cause age-related motor neuronopathy in mice. Neurobiol Dis. 47, 163-73. Stengel, F., et al., 2010. Quaternary dynamics and plasticity underlie small heat shock protein chaperone function. Proc Natl Acad Sci U S A. 107, 2007- 12. Stetler, R. A., et al., 2008. Hsp27 protects against ischemic brain injury via attenuation of a novel stress-response cascade upstream of mitochondrial cell death signaling. J Neurosci. 28, 13038-55. Stetler, R. A., et al., 2012. Phosphorylation of HSP27 by protein kinase D is essential for mediating neuroprotection against ischemic neuronal injury. J Neurosci. 32, 2667-82. Stock, A. D., et al., 2003. Heat shock protein 27 gene: chromosomal and molecular location and relationship to Williams syndrome. Am J Med Genet A. 120A, 320-5. Stokoe, D., et al., 1992a. MAPKAP kinase-2; a novel protein kinase activated by mitogen-activated protein kinase. EMBO J. 11, 3985-94. Stokoe, D., et al., 1992b. Identification of MAPKAP kinase 2 as a major enzyme responsible for the phosphorylation of the small mammalian heat shock proteins. FEBS Lett. 313, 307-13. Sugiyama, Y., et al., 2000. Muscle develops a specific form of small heat shock protein complex composed of MKBP/HSPB2 and HSPB3 during myogenic differentiation. J Biol Chem. 275, 1095-104. Sur, R., et al., 2008. Hsp27 regulates pro-inflammatory mediator release in keratinocytes by modulating NF-kappaB signaling. J Invest Dermatol. 128, 1116-22. Takai, S., et al., 2007. Protein kinase C delta regulates the phosphorylation of heat shock protein 27 in human hepatocellular carcinoma. Life Sci. 81, 585-91. Takayama, S., et al., 2003. Heat-shock proteins as regulators of apoptosis. Oncogene. 22, 9041-7. Tazir, M., et al., 2014. Hereditary motor and sensory neuropathies or Charcot- Marie-Tooth diseases: an update. J Neurol Sci. 347, 14-22. Theodoraki, M. A., Caplan, A. J., 2012. Quality control and fate determination of Hsp90 client proteins. Biochim Biophys Acta. 1823, 683-8. Timmerman, V., et al., 2014. Genetics of Charcot-Marie-Tooth (CMT) Disease within the Frame of the Human Genome Project Success. Genes (Basel). 5, 13-32. Tiscornia, G., et al., 2006. Production and purification of lentiviral vectors. Nat Protoc. 1, 241-5. Tortarolo, M., et al., 2004. Expression of SOD1 G93A or wild-type SOD1 in primary cultures of astrocytes down-regulates the glutamate transporter GLT-1: lack of involvement of oxidative stress. J Neurochem. 88, 481-93.

153

Turner, B. J., et al., 2005. Impaired extracellular secretion of mutant superoxide dismutase 1 associates with neurotoxicity in familial amyotrophic lateral sclerosis. J Neurosci. 25, 108-17. Urushitani, M., et al., 2006. Chromogranin-mediated secretion of mutant superoxide dismutase proteins linked to amyotrophic lateral sclerosis. Nat Neurosci. 9, 108-18. Van Damme, P., et al., 2007. Astrocytes regulate GluR2 expression in motor neurons and their vulnerability to excitotoxicity. Proc Natl Acad Sci U S A. 104, 14825-30. van der Weerd, L., et al., 2010. Overexpression of heat shock protein 27 reduces cortical damage after cerebral ischemia. J Cereb Blood Flow Metab. 30, 849-56. Van Montfort, R., et al., 2001. Structure and function of the small heat shock protein/alpha-crystallin family of molecular chaperones. Adv Protein Chem. 59, 105-56. Vleminckx, V., et al., 2002. Upregulation of HSP27 in a transgenic model of ALS. J Neuropathol Exp Neurol. 61, 968-74. Wagner, B. J., et al., 1998. Structure and genomic organization of the human AUF1 gene: alternative pre-mRNA splicing generates four protein isoforms. Genomics. 48, 195-202. Wang, J., et al., 2008. Differential regulation of small heat shock proteins in transgenic mouse models of neurodegenerative diseases. Neurobiology of Aging. 29, 586-597. Wei, L., et al., 2011. Hsp27 participates in the maintenance of breast cancer stem cells through regulation of epithelial-mesenchymal transition and nuclear factor-kappaB. Breast Cancer Res. 13, R101. Welch, W. J., Suhan, J. P., 1985. Morphological study of the mammalian stress response: characterization of changes in cytoplasmic organelles, cytoskeleton, and nucleoli, and appearance of intranuclear actin filaments in rat fibroblasts after heat-shock treatment. J Cell Biol. 101, 1198-211. White, E. J., et al., 2013. Post-transcriptional control of gene expression by AUF1: mechanisms, physiological targets, and regulation. Biochim Biophys Acta. 1829, 680-8. Wilhelmus, M. M., et al., 2006a. Small heat shock proteins inhibit amyloid-beta protein aggregation and cerebrovascular amyloid-beta protein toxicity. Brain Res. 1089, 67-78. Wilhelmus, M. M., et al., 2006b. Specific association of small heat shock proteins with the pathological hallmarks of Alzheimer's disease brains. Neuropathol Appl Neurobiol. 32, 119-30. Williams, K. L., Mearow, K. M., 2011. Phosphorylation status of heat shock protein 27 influences neurite growth in adult dorsal root ganglion sensory neurons in vitro. J Neurosci Res. 89, 1160-72.

154

Williams, K. L., et al., 2006. Heat shock protein 27 is involved in neurite extension and branching of dorsal root ganglion neurons in vitro. J Neurosci Res. 84, 716-23. Wilson, G. M., et al., 2003. Regulation of A + U-rich element-directed mRNA turnover involving reversible phosphorylation of AUF1. J Biol Chem. 278, 33029-38. Winner, B., et al., 2011. In vivo demonstration that alpha-synuclein oligomers are toxic. Proc Natl Acad Sci U S A. 108, 4194-9. Wirdefeldt, K., et al., 2011. Epidemiology and etiology of Parkinson's disease: a review of the evidence. Eur J Epidemiol. 26 Suppl 1, S1-58. Wu, R., et al., 2007. Hsp27 regulates Akt activation and polymorphonuclear leukocyte apoptosis by scaffolding MK2 to Akt signal complex. J Biol Chem. 282, 21598-608. Wu, Y., et al., 2009. HSP27 regulates IL-1 stimulated IKK activation through interacting with TRAF6 and affecting its ubiquitination. Cell Signal. 21, 143-50. Wyttenbach, A., et al., 2000. Effects of heat shock, heat shock protein 40 (HDJ- 2), and proteasome inhibition on protein aggregation in cellular models of Huntington's disease. Proc Natl Acad Sci U S A. 97, 2898-903. Yamaguchi, H., Wang, H. G., 2002. Bcl-XL protects BimEL-induced Bax conformational change and cytochrome C release independent of interacting with Bax or BimEL. J Biol Chem. 277, 41604-12. Yamanaka, K., et al., 2008a. Mutant SOD1 in cell types other than motor neurons and oligodendrocytes accelerates onset of disease in ALS mice. Proc Natl Acad Sci U S A. 105, 7594-9. Yamanaka, K., et al., 2008b. Astrocytes as determinants of disease progression in inherited amyotrophic lateral sclerosis. Nat Neurosci. 11, 251-3. Yerbury, J. J., et al., 2013. The small heat shock proteins alphaB-crystallin and Hsp27 suppress SOD1 aggregation in vitro. Cell Stress Chaperones. 18, 251-7. Yoshihara, T., et al., 2002. Differential expression of inflammation- and apoptosis-related genes in spinal cords of a mutant SOD1 transgenic mouse model of familial amyotrophic lateral sclerosis. J Neurochem. 80, 158-67. Zarranz, J. J., et al., 2004. The new mutation, E46K, of alpha-synuclein causes Parkinson and Lewy body dementia. Ann Neurol. 55, 164-73. Zhai, J., et al., 2007. Disruption of neurofilament network with aggregation of light neurofilament protein: a common pathway leading to motor neuron degeneration due to Charcot-Marie-Tooth disease-linked mutations in NFL and HSPB1. Hum Mol Genet. 16, 3103-16. Zhang, W., et al., 1993. Purification, characterization, and cDNA cloning of an AU-rich element RNA-binding protein, AUF1. Mol Cell Biol. 13, 7652-65. Zhuang, H., et al., 2010. Down-regulation of HSP27 sensitizes TRAIL-resistant tumor cell to TRAIL-induced apoptosis. Lung Cancer. 68, 27-38. 155

Zourlidou, A., et al., 2004. HSP27 but not HSP70 has a potent protective effect against alpha-synuclein-induced cell death in mammalian neuronal cells. J Neurochem. 88, 1439-48. Zuchner, S., Vance, J. M., 2006. Mechanisms of disease: a molecular genetic update on hereditary axonal neuropathies. Nat Clin Pract Neurol. 2, 45-53.

156