The Functional Analysis of Granulins

2014

By Kate Young

School of Medicine

A thesis submitted to the University of Manchester for the degree of Master of Philosophy (MPhil) in the Faculty of Medical and Human Sciences

-1-

Table of Contents Declaration ...... 6 Copyright Statement ...... 6 Acknowledgements ...... 7 Abstract ...... 8 Abbreviations ...... 9 Chapter 1 Introduction ...... 10 1.1 Frontotemporal Lobar Degeneration...... 11 1.2 Clinical Syndromes...... 11 1.3 Neuropathology...... 13 1.4 FTLD Proteinopathies...... 14 1.4.1 FTLD Tau/FTLD-17 ...... 14 1.4.2 FTLD-TDP43 ...... 15 1.4.3 Less Common types of FTLD ...... 16 1.5 Gene Mutations Associated with FTLD...... 16 1.5.1 MAPT linked FTLD ...... 16 1.5.2 GRN Linked FTLD ...... 17 1.5.3 Chromosome 9 linked FTLD ...... 18 1.5.4 Rare Mutations...... 19 1.6 Progranulin...... 21 1.6.1 Progranulin Function ...... 22 1.6.2 Progranulin Receptors...... 23 1.6.3 Inflammation and Wound Healing ...... 23 1.6.4 Cell Signalling ...... 24 1.6.5 Tumorigenesis ...... 24 1.6.6 Progranulin and Neurodegeneration ...... 24 1.7 The Functional Analysis of Granulins...... 25 1.8 Aims and Objectives...... 25 Chapter 2 Progranulin and Granulin Cloning ...... 26 2.1 Introduction ...... 27 2.2 Construct DNA Amplification ...... 29 -2-

2.2.1 Transformation and Clonal Selection ...... 30 2.2.2 Restriction Digestion ...... 30 2.2.3 Agarose Gel Electrophoresis ...... 30 2.2.4 Ligation ...... 31 2.2.3 Transient Transfection ...... 32 2.2.4 Gel Electrophoresis ...... 32 2.2.5 Western Blotting ...... 33 2.3.1 PCR Protocol ...... 35 2.3.2 Sequencing...... 36 Chapter 3 Protein Expression and Purification ...... 40 3.1 Introduction ...... 41 3.2 Stable Transfection ...... 41 3.3 Protein Purification...... 42 3.3.1 Cell harvesting ...... 42 3.3.2 Halotag resin purification ...... 42 3.3.3 Coomassie Staining ...... 44 3.3.4 High-Performance Liquid Chromatography (HPLC) ...... 45 3.4 Gel Electrophoresis ...... 46 3.5 Silver Staining...... 46 Chapter 4 Cell Treatment and Next Generation Sequencing ...... 49 4.1 Introduction...... 50 4.2 Differentiating SH-SY5Y cells...... 50 4.3 Cell Treatment ...... 50 4.4 RNA Extraction and Homogenisation...... 51 4.5 Library Preparation for Next Generation Sequencing ...... 52 4.5.1 Ribosomal RNA Depletion ...... 53 4.5.2 RNA Fragmentation ...... 54 4.5.3 Library Preparation ...... 56 4.5.4 Library Amplification ...... 57 4.5.5 Library Sequencing ...... 59 4.6 Next Generation Sequencing Data Analysis ...... 60

-3-

4.6.1 Alignment to the Reference Genome...... 60 4.7 Validation by Real Time PCR ...... 62 4.7.1 Introduction ...... 62 4.7.2 RNA Extraction and cDNA Synthesis ...... 63 4.7.3 Real Time PCR ...... 64 4.8 Gene Expression Results ...... 66 Chapter 5 Discussion ...... 74 RNA Seq Discussion ...... 77 Possible Future Work...... 82 Supplementary information ...... 84 Materials ...... 84 References ...... 86

Tables

Table 1.1 Clinical, diagnostic and pathological features associated with the different sub- types of FTLD...... 13 Table 4.1 Cell Treatments were performed as shown in the table below...... 51 Table 4.2 Overview of sequencing statistics. The ERRC Spike in control counts show between 60 and 70 transcripts detected out of a possible 92...... 62 Table 4.3 Genes selected using the P=0.01 cut off by class ...... 67 Table 4.4 Pathway analysis carried out using DAVID for the GRN and PGRN treatments. 69 Table 4.5 shows the highest log2fold change up and down- regulated genes for each granulin and PGRN at the 30 minute time point...... 71

Figures

Figure 1.1. Isoforms of Tau...... 14 Figure 1.2 Schematic diagram of the classification of the different FTLD sub-types...... 20 Figure 1.3 Alignment of granulin amino acid sequences as described by Bhandari et al [64]...... 21 Figure 1.4 Schematic diagram of progranulin showing granulin domains...... 21 Figure 2.1 pCR®2.1 TOPO® Cloning Vector...... 28 Figure 2.2 pEGFP-N1 Expression Vector...... 31 Figure 2.3 Halotag® CMV-neo Vector...... 34 -4-

Figure 2.4 Diagram of one cycle of dye terminator (di-deoxy dNTP) cycle sequencing. .... 36 Figure 2.5. Examples of agarose gels...... 38 Figure 3.1 Western blot of over-expressed Halo®-tagged granulins pre and post binding to Halotag resin...... 43 Figure 3.2 TEV Cleaved Granulin ...... 44 Figure 3.3 HPLC trace of gel filtration column fractions...... 45 Figure 3.4 Example of Granulin HPLC trace...... 46 Figure 3.5 Silver–stained gel of purified granulins ...... 47 Figure 3.6 Purified Progranulin ...... 48 Figure 4.1 Example of Nano Chip Readout from Agilent 2100 Bioanalyzer...... 52 Figure 4.2 Pico Chip Readout...... 56 Figure 4.3 DNA Chip Readout...... 59 Figure 4.4 Taken from NGS raw data Phred score versus read position...... 60 Figure 4.5 Example SHSY-5Y Standard Curve for EMX2...... 65 Figure 4.6. Real Time PCR Validation of NGS Results...... 66 Figure 4.7 Principal Components analysis of NGS data from GRN and PGRN treatments . 68

Total word count 22,419

-5-

Declaration No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

Copyright Statement 1. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. 2. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. 3. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. 4. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectualproperty.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/ regulations) and in The University’s policy on presentation of Theses

-6-

Acknowledgements

I would like to thank my supervisor Stuart Pickering-Brown for giving me the opportunity to undertake this degree and also having faith in my ability to successfully complete it. My eternal grateful thanks goes to all my long-suffering colleagues; Janis Bennion- Callister, for your help and advice (and apologies for the stupid questions), Sarah Ryan thank you for making me laugh and Sara Rollinson whose patience I have stretched to the limit and without whose cajoling, encouragement and endless explanations this thesis would never have been written.

I also have to thank my Mum for unconditional love and believing that I could do anything and everything and finally Andrew, for his love and support and for giving me the courage to “go for it”

-7-

Abstract Progranulin (PGRN) is a 593 amino acid secreted glycoprotein which is encoded for by the GRN gene. It is ubiquitously expressed in a number of tissues and is associated with numerous cellular functions including angiogenesis, metabolism and acting as a neurotrophic growth factor. The full-length protein comprises 7.5 tandem granulin repeats containing highly conserved cysteine repeats. The full length PGRN can be cleaved by neutrophil elastase and other metalloproteases to produce a series of 6kDa peptides which, in some instances have been shown to have opposing effects on cell proliferation and inflammation to the parent molecule. Levels of PGRN have been shown to be raised in some cancers, inflammation, autoimmune disease and wound healing however it is PGRN haploinsufficiency caused by mutations in the GRN gene that has been implicated in Frontotemporal Lobar Degeneration (FTLD).

Although PGRN and the granulins obviously play a significant role in disease processes, it has not been possible to identify a specific measurable function or activity regulated by these proteins. For this reason we decided to investigate the differential gene expression produced by treating differentiated neural cells with exogenous PGRN and granulin peptides in order to try to elucidate specific biological processes which could lead to the identification of disease markers or therapeutic targets.

Pure PGRN and GRN peptides were produced by synthesising sequence-specific DNA, cloning into a Halo®-tagged expression vector and expressing in HEK293 cells. The resultant proteins were purified by HPLC and used to treat fully differentiated SH-SY5Y cells. Post-treatment cells were lysed and used to produce cDNA for RNA Seq analysis and real time PCR or protein for validation experiments.

The RNA Seq data produced identified a number of genes up and down-regulated in all treatments. PGRN treatment resulted in the down-regulation of several genes associated with both the proteosome and the spliceosome pathways. The fact that, in addition, there was also a trend towards significance in the down-regulation of the spliceosome by granulin treatments suggests that this may be a common reaction of both the parent molecule and its breakdown components. In addition the down-regulation of CHMP1A and CHMP5 may suggest a role for the ESCRTlll pathway in PGRN mediated FTLD.

The identification of these pathways may provide evidence for the selection of potential candidates for putative treatment options although further evaluation and validation of the data will be required.

-8-

Abbreviations

FTLD Frontotemporal Degeneration

AD Alzheimer’s disease bvFTD Behavioural variant frontotemporal dementia

SD Semantic Dementia

PNFA Progressive Non-Fluent Aphasia

ALS Amyotrophic Lateral Sclerosis

CBD Corticobasal Syndrome

PSP Progressive Supranuclear Palsy

MAPT Microtubule Associated Protein Tau

TDP Transactive Response DNA Binding Protein

FUS Fused in Sarcoma

UPS Ubiquitin Proteosome System

NCI Neuronal Cytoplasmic Inclusion

DN Dystrophic Neurite

NII Neuronal Intranuclear Inclusion

GCI Glial Cytoplasmic Inclusion

GRN Progranulin gene

C9orf72 Chromosome 9 Open Reading Frame 72 gene

CHMP2 Charged Multivesicular Body Protein gene

ESCRTlll Endosomal Sorting Complex Required for Transport

VCP Valosin Containing Protein

PGRN Progranulin protein

GRN Granulin peptide

-9-

Chapter 1 Introduction

-10-

1.1 Frontotemporal Lobar Degeneration.

Frontotemporal lobar degeneration (FTLD) was first recognised as a clinical syndrome by Arnold Pick in 1892 [1] and accounts 5-15% of all dementias [2]. It is the second most common form of pre-senile dementia after Alzheimers disease (AD). Disease onset usually occurs between the ages of 50 and 60 and affects both males and females equally. FTLD is characterised by progressive neurodegeneration of the frontal and temporal cortical areas of the brain and forty percent of cases have been shown to have a familial link with the remainder occurring sporadically [3-5].

Due to the heterogeneity of clinical phenotypes FTLD has been difficult to categorise and until fairly recently differences in nomenclature have caused confusion within the field.

1.2 Clinical Syndromes.

FTLD has now been classified into three clinically distinguishable subtypes: behavioural variant FTLD (bvFTD), semantic dementia (SD) and progressive non-fluent aphasia (PNFA). BvFTD is characterised by severe personality changes including inappropriate social and sexual responses, lack of emotion, alteration in eating habits, obsessive/compulsive behaviour and apathy, all of which can lead to social isolation. Memory and spatial awareness seem to be unaffected in the initial stages of the disease [6-9]. BvFTD accounts for approximately two thirds of diagnosed cases of FTLD with some sufferers also displaying Parkinson-like symptoms of uncontrolled limb movements and rigidity [10].

SD presents as loss of semantic memory and a gradual decline in language comprehension. Although patients are able to perceive objects and faces they often have trouble naming them; fluency of speech and episodic memory are, however, not impaired. PNFA exhibits as loss of fluency and grammar without apparent comprehension deficits although single word comprehension can become impaired as the disease progresses. Patients with both SD and PNFA have also been shown to exhibit symptoms

-11-

of bvFTD in the later stages of disease. However, these are less severe than in bvFTD with PNFA sufferers being the least affected [11].

In recent years there has been increasing awareness of a significant percentage of FTLD cases also presenting with symptoms of motor neuron disease. Amyotrophic lateral sclerosis (ALS) is the most common form of motor neuron disease and is characterised by progressive muscle weakness and atrophy affecting limb movements, speech, and, in the later stages, swallowing and breathing. Approximately 75% of cases are associated with difficulty walking, stumbling or tasks requiring manual dexterity (limb onset). In the region of 25% present with speech defects such as slurring or difficulty swallowing (bulbar onset). Respiratory onset patients where the muscles supporting breathing are affected are extremely rare.

Although cognitive deficits and behavioural abnormalities have been noted in ALS sufferers for over a century it has only relatively recently been possible to establish a definitive link between the two diseases. Up to 50% of ALS patients studied showed some degree of functional loss in frontal lobe tests [12]. This was sufficient in 15% of cases to support a diagnosis of FTLD [13]. Similarly, 40% of FTLD cases present with motor dysfunction with 10 - 15% able to be classified as ALS [14]. The familial inheritance of FTLD-ALS is thought to be extremely high (circa 50%) with a very poor survival rate, from disease onset, of between 2 and 3 years [11]. This evidence strongly supports the theory of a disease continuum with FTLD at one end and ALS at the other end of the spectrum [15].

In addition to the relationship between FTLD and ALS there is an overlap of clinical symptoms with the atypical Parkinsonisms, Corticobasal Degeneration (CBD) and Progressive Supranuclear Palsy (PSP) which can also be considered to be part of the disease spectrum [16, 17].

-12-

1.3 Neuropathology.

Pathological examination of FTLD patient brains has shown some correlation between clinical phenotype and pattern of atrophy. For example, bvFTD brains exhibit a more or less symmetrical atrophy of the frontal and anterior temporal lobes. SD patients show bilateral atrophy of the neocortex of the temporal lobe with left hemisphere atrophy associated with semantic deterioration and the right with visual agnosia. In PNFA the affected area is the perisylvian cortex of the left temporal lobe [11] (Table 1.1).

Table 1.1 Clinical, diagnostic and pathological features associated with the different sub-types of FTLD.

From Seltman, R.E. and B.R. Matthews, Frontotemporal lobar degeneration: epidemiology, pathology, diagnosis and management. CNS Drugs, 2012. 26(10):p.841- 70.

-13-

1.4 FTLD Proteinopathies.

FTLD is one of a number of diseases known as proteinopathies in which certain proteins become structurally abnormal causing the formation of intracellular inclusions which are thought, subsequently, to affect cell function.

1.4.1 FTLD Tau/FTLD-17

Tauopathy is a type of proteinopathy in which the protein tau forms aggregates within the brain leading to a variety of pathological and clinical presentations.

Tau is the microtubule associated protein expressed by the MAPT gene in the peripheral and central nervous systems. It is a heat-resistant phospho-protein and is responsible for microtubule assembly and stability within the axons of neurons. In normal adult brain tau exists in 6 different isoforms half of which have 3 microtubule binding domains and half with 4. The ratio between the two forms, referred to as 3R and 4R, is 1:1 (Figure 1.1).

Figure 1.1. Isoforms of Tau.

Alternate splicing of the MAPT gene results in 6 isoforms of the tau protein. The presence/absence of exon 10 denotes the 3R/4R forms whilst the inclusion of exons 2 and 3 are denoted by 0,1 and 2N nomenclature.

It is thought that a stem-loop structure at the exon 10/intron 10 boundary is responsible for the alternate splicing of exon 10 leading to alterations in the ratio of 3R and 4R tau

-14-

[18]. In disease states this can then lead to microtubule destabilisation and neurotoxic aggregation of tau protein within neurons and microglia [11, 19].

Picks disease (PiD) was first identified as a histologically distinct disease by Alois Alzheimer in 1911 when he showed spherical deposits within neurons as opposed to the characteristic neurofibrillary tangles of AD [20]. These deposits, known as Pick bodies, have been subsequently shown to consist of 3R tau. However, other forms of FTLD, including CBD and PSP are predominantly positive for 4R tau.

1.4.2 FTLD-TDP43

Immunohistological studies revealed that more than 50% of FTLD brains with inclusions were tau-negative but stained positive for ubiquitin. Initially this group was referred to as FTLD-U as the composition of the protein was unknown. However, the major protein subsequently identified in these ubiquitin-positive, tau-negative inclusions was TDP-43 (transactive response DNA binding protein) leading to the designation FTLD-TDP [11, 21]. It should also be noted that around 95% of ALS patients have TDP-43 nuclear inclusions further strengthening the evidence for a disease continuum [22].

FTLD-TDP has four subtypes which are associated with differing neuropathology and clinical phenotypes. Type A is the most common and occurs in 40-50% of cases with GRN or C9orf72 mutations. It is mainly associated with bvFTLD and PNFA and with numerous neuronal cytoplasmic inclusions (NCIs) and dystrophic neurites (DNs) in cortical layer II, neuronal intranuclear inclusions (NIIs) in superficial cortical layers and glial cytoplasmic inclusions (GCIs). Type B is associated with C9orf72 mutations and occurs in 28-34% of the disease population. The clinical phenotypes are bvFTLD and FTLD-ALS and ALS with NCIs throughout the cerebral cortex, hypoglossal nucleus and ventral horn of the spinal cord. The genetic cause of type C (17-25%) is unknown but is associated with SD and bvFTLD with long DNs in superficial cortical layers and a few NCIs and NIIs. Type D is very rare and has been attributed to valosin containing protein (VCP) mutation associated with inclusion body myopathy with Paget’s disease and FTD (IBMPFD). Neuropathologically, it can be identified by numerous NIIs and DNs and few NCIs throughout the entire cortex [11, 23].

-15-

TDP-43 is thought to function as a transcription repressor and regulator. It is expressed in the nucleus of neuronal cells with a small proportion present in the cytoplasm. However, TDP-43 becomes hyperphosphorylated,ubiquitinated and cleaved to produce C-terminal fragments in disease states which lead to protein mis-folding and intracytoplasmic mis- localisation [24].

1.4.3 Less Common types of FTLD

Although very rare, a number of ubiquitin – positive FTLD cases were found to be TDP-43 negative but found to contain FUS (fused-in-sarcoma) inclusions. These are now referred to as FTLD-FUS [25] [26]. FUS, in common with TDP-43, is a DNA/RNA binding protein involved in transcription regulation [27]. FUS inclusions also stain for other members of the FET family of proteins such as Ewing Sarcoma Protein and TATA-binding protein- associated factor 15 implicating that they may also be involved in the neurodegenerative process [28].

A small number of FTLD-U cases which were TDP-43 and FUS negative still contain protein inclusions of unknown composition. These have now been designated FTLD- ubiquitin proteasome system (FTLD-UPS) [21].

1.5 Gene Mutations Associated with FTLD.

As previously stated 40 % of FTLD cases are associated with a family history of disease and mutations in several genes have now been shown to be responsible.

1.5.1 MAPT linked FTLD

Linkage analysis of several pedigrees beginning in 1994 [29] identified a locus on chromosome 17q21, 2 centimorgans in length, from which arose the disease designation FTDP-17 (frontotemporal dementia with Parkinsonism linked to chromosome 17) [30]. As these families all showed a similar clinical presentation it was thought that mutations in the same gene were responsible for disease. Pathological examination showed that a significant proportion of these individuals exhibited abnormal deposits of tau protein -16-

within cortical neurons. It was natural to assume, therefore, that the gene responsible was MAPT which is located in the same linkage region and encodes the microtubule associated protein tau. Mutations occurring in MAPT can either change an amino acid within the protein leading to inability of tau to bind microtubules or interfere with alternative splicing which can change the ratio of tau isoforms [31]. For example, mutations in exon 10 are linked to an increase in the expression of 4R tau which forms aggregates in the brains of affected individuals [32]. 44 MAPT mutations have been found in 134 FTLD families to date [11] and account for roughly 10% of all FTLD cases [33].

1.5.2 GRN Linked FTLD

As more genetic linkage studies were undertaken and more cases linked to chromosome 17q21 were identified it became evident that not all patients displaying clinical symptoms of FTD had MAPT mutations. Pathological examination also showed neuronal inclusions that did not contain tau protein but ubiquitin positive TDP-43 aggregates. The only possible explanation was the existence of another gene within the same 2 cM region. This indeed proved to be the case when in 2006 the gene encoding progranulin (GRN) was discovered 1.7 Mb centromeric to MAPT. Currently 69-70 different GRN mutations have been identified in 231 families [11, 34]. For example, missense mutations of the methionine codon in exon 2 disrupt the Kozak sequence affecting protein secretion whilst a single base change in the splice donor site of intron 1 produces a transcript that is thought to be retained and degraded in the nucleus leading to the loss of mRNA [35]. Although produced by different mechanisms all mutations are thought to result in reduced progranulin expression leading to haploinsufficiency [36]. In contrast to MAPT mutations leading to the accumulation of tau, mutations in GRN do not lead to the neuronal deposition of granulin protein. The pathological inclusions found in the brains of patients with GRN mutations contain TDP-43. Although GRN mutations have been found to account for 5 - 10% of inherited FTLD [37-39] they were not thought to cause ALS although recently a patient, clinically diagnosed with ALS was found to have a GRN A9D mutation [40].

-17-

Sporadic cases of FTLD without GRN mutations have been linked to hypermethylation of CpG units in the GRN promoter region blocking transcription. It has also been shown that DNA methyltransferase 3a (DNMT3a) is upregulated in FTLD patients resulting in a reduction of GRN promoter activity and expression levels [41].

1.5.3 Chromosome 9 linked FTLD

In 2006 genome wide linkage studies identified a region on the short arm of chromosome 9 linked to two families of ALS/FTLD sufferers [42, 43]. A subsequent study of ALS in Finland enabled the locus 9p21 to be identified [44]. Most recently, C9orf72, the gene located on chromosome 9, has been shown to contain a hexanucleotide expansion found in familial FTD/ALS [45]. This GGGGCC expansion, which is located between the first two five prime non-coding exons, is currently thought to be the major genetic cause of familial FTD/ALS worldwide accounting for approximately 34.2% of ALS and 24.9% of FTD cases worldwide with around a third of patients showing symptoms of both diseases. C9orf72 encodes a protein which is currently uncharacterised and of unknown function although in silico analysis has suggested that it shares a distant homology with Differentially Expressed in Normal and Neoplasia (DENN) proteins and as such could be involved in membrane trafficking [46].

Alternative splicing leads to the production of three different transcripts that are predicted to encode two protein isoforms. Isoform A is a 481 amino acid protein encoded by exons 2-11 and is expressed by transcripts 2 and 3 whilst isoform B, expressed by transcript 1, is a 222 amino acid protein encoded by exons 2-5 [47].

As previously described, C9orf72 mutation pathology cases show TDP-43 positive inclusions however, in addition, inclusions which stain for ubiquitin and p62 have been found in hippocampal pyramidal neurons and the cerebellar granule cell layer [48, 49]. Translation of the sense RNA transcripts of the hexanucleotide expansion in all three reading frames has been shown to produce a series of dipeptide repeat proteins, poly Gly-Arg, poly Gly-Pro and poly Gly-Ala. The mechanism by which this occurs is thought to be repeat non-ATG initiated translation (RAN) which was first described for CAG repeat expansions which produced protein aggregates containing polyglutamine, polyalanine

-18-

and polyserine in myotonic dystrophy type 1 and spinocerebellar ataxia type 8 patient tissues [50] [51]. Antibodies raised against the dipeptide repeats, anti-GR, anti-GP and anti-GA have been shown to stain aggregates within the granular layers of the cerebellum and dentate gyrus, hippocampus and molecular layer of the cerebellum and neocortex [51].

Although several groups have shown reduced levels of C9orf72 individual transcripts in mutation carriers leading to a possible loss of function mechanism [45, 47, 52], as is the case with other repeat expansion diseases, pathogenesis could also be caused by RNA toxic gain of function [47, 53-55]. One possible explanation for this could be the existence of G-quadruplexes.The formation of these multimers is influenced by the hexanucleotide repeat length and may lead to the accumulation of RNA transcript aggregates forming nuclear foci and the sequestration of binding proteins [47, 56, 57].

1.5.4 Rare Mutations

Rare pathogenic nonsense mutations in the CHMP2B gene on chromosome 3 which encodes charged multivesicular protein 2B part of the endosomal ESCRTlll complex, have been reported in a single Danish FTLD pedigree [58] and one Belgian FTLD patient [59]. The Danish family mutation occurs in the splice acceptor site within exon 6 forming two novel transcripts which are both present in the FTLD patient brains. This ultimately results in the loss of the final 36 amino acids. The mutation present in the Belgian pedigree changes a glutamine residue to a stop codon leading to the loss of the final 49 amino acids. Both mutations, therefore, share a common mechanism resulting in the deletion of the C-terminus of the protein[60].

Valosin containing protein (VCP) is a member of the AAA-ATPase superfamily of genes and is implicated in several cellular processes one of which is as part of the ubiquitin- proteasome system (UPS) where it is responsible for the degradation of unwanted proteins. Mutations within the gene can lead to loss of VCP activity and the formation ubiquitinated protein inclusions within the brain. It is associated with inclusion body myopathy with Paget’s disease and FTD (IBMPFD) and ALS [61, 62]. Mutations in MAPT,

-19-

GRN, and C9orf72 account for at least 17% of all familial FTLD cases with mutations in CHMP2B and VCP occurring in less than 1% each [11].

Previously unmentioned within the FTLD-tau disease group are white matter tauopathy with globular glial inclusions (WMT-GGI), argyrophilic grain disease, (AGD) and tangle only dementia (TOD). Other FTLD-FUS syndromes are neuronal intermediate filament inclusion disease (NFID) atypical FTLD with ubiquitin inclusions (aFTLD-U) and basophilic inclusion body disease (BIBD) (See Figure 1.2) [19, 63].

Figure 1.2 Schematic diagram of the classification of the different FTLD sub-types.

Classification of different FTLD sub-types according to the protein composition of the intracellular inclusions associated with them. MAPT- Microtubule-Associated Protein Tau gene, CBD – Corticobasal Degeneration, PSP – Progressive Supranuclear Palsy, WMT-GGI - White Matter Tauopathy with Globular Glial Inclusions, AGD – Argyrophilic Grain Disease, TOD – Tangle Only Dementia, TDP – Transactive Response DNA-binding Protein, U – Ubiquitin, GRN – Progranulin gene, TARDP - Transactive Response DNA-binding Protein gene, VCP- Valosin Containing Protein gene, C9orf72 – chromosome 9 open reading frame 72 gene, FUS – Fused in -20-

Sarcoma, NIFID - Neuronal Intermediate Filament Inclusion Disease, aFTLD-U – atypical FTLD with ubiquitin inclusions, BIBD – Basophilic Inclusion Body Disease, FTLD –UPS – FTLD –Ubiquitin Proteasome System, CHMP2B – Charged Multivesicular Body Protein gene.

From Cairns, N.J. and N. Ghoshal, FUS: A new actor on the frontotemporal lobar degeneration stage. Neurology, 2010. 74(5): p. 354-6

1.6 Progranulin.

PGRN is composed of 7.5 tandem repeats of a highly-conserved cysteine rich motif separated by linker regions [64]. It is a 68kDa protein with 5 putative glycosylation sites giving rise to a full length secreted protein of 88kDa[65] (Figure 1.3).

Figure 1.3 Alignment of granulin amino acid sequences as described by Bhandari et al [64].

Paragranulin CPDGQFCPVACCLDPGGASYSCCRPLLDKWPTTLSRHL Granulin G GGPCQVDAHCSAGHSCIFTVSGTSSCCPFPEAVACGDGHHCCPRGFHCSADGRSCF Granulin F AIQCPDSQFECPDFSTCCVMVDGSWGCCPMPQASCCEDRVHCCPHGAFCDLVHTRC Granulin B VMCPDARSRCPDGSTCCELPSGKYGCCPMPNATCCSDHLHCCPQDTVCDLIQSKCL Granulin A DVKCDMEVSCPDGYTCCRLQSGAWGCCPFTQAVCCEDHIHCCPAGFTCDTQKGTCE Granulin C VPCDNVSSCPSSDTCCQLTSGEWGCCPIPEAVCCSDHQHCCPQGYTCVAEGQCQ Granulin D IGCDQHTSCPVGQTCCPSLGGSWACCQLPHAVCCEDRQHCCPAGYTCNVKARSCE Granulin E DVECGEGHFCHDNQTCCRDNRQGWACCPYRQGVCCADRRHCCPAGFRCAARGTKCL

The figure shows the high degree of similarity between granulin peptides with conserved cysteine residues high-lighted in lilac.

PGRN can be cleaved at the linker regions by several different proteases including neutrophil elastase [66], proteinase 3 [67], matrix metalloproteinase-14 (MMP-14) [68], and ADAM metallopeptidase with thrombospondin type 1 motif 7 (ADAMTS 7) [69] to produce individual granulins. However, not all the linkage regions are susceptible to cleavage (Figure 1.4).

Figure 1.4 Schematic diagram of progranulin showing granulin domains.

N P G F B A C D E C -21-

Indicates neutrophil elastase cleavage sites. [66]

Depicts other proteolytic cleavage sites within the linker regions where the specific protease responsible has not yet been identified.

Incubating recombinant progranulin with these proteases does not just release 7.5 6kDa granulins as might be expected. Several intermediate peptides larger than 15kDa have been found. Specifically, digestion with MMP-12 a C-terminal antibody detected bands between 15 and 45kDa [66, 67, 70].

PGRN cleavage can be inhibited by secretory leucocyte protease inhibitor (SLPI) which binds to elastase enabling relative levels of full length GRN and granulins to be controlled [71].

1.6.1 Progranulin Function

The GRN gene is made up of 13 exons, the first of which is non-coding. It encodes a 593 amino acid multifunctional protein which, due to its diverse functions, has been named as proepithelin, acroganin, granulin-epithelin precursor (GEP), epithelial transforming growth factor, PC cell derived growth factor (PCDGF) and progranulin (PGRN).Expression can be detected in all human tissues but predominantly in epithelial and hemapoietic cells [72, 73].

PGRN has been shown to function as a neurotrophic growth factor and thus play a major role in neuronal morphology and connectivity [74]. Exogenous application of mature PGRN to cultured mouse cortical neurons promoted an increase in neurite outgrowth [75] [74]. Further evidence was provided by siRNA knockdown of PGRN in rat primary hippocampal neurons which caused decreased neuronal length and outgrowth and additionally a reduction in synaptic densities. Synaptic transmission is, however, increased and is thought to be due to higher pre-synaptic release in the remaining synapses [76].

-22-

1.6.2 Progranulin Receptors

Sortilin is a transmembrane protein of the VPS10 family and is responsible for intracellular protein trafficking. Binding of PGRN to sortilin has been shown to be similar to binding of neurotensin insofar as PGRN binds to the same beta propeller region via its C-terminus which is surprising given that only the last three amino acids show any similarity [77]. Deletion of these last three amino acids totally abolished binding of both mature PGRN and one of its proteolytic cleavage products, GRN E. Furthermore a truncated PGRN construct minus the C-terminal GRN E was unable to bind sortilin [78]. Co-expression of GRN and sortilin in HEK293 cells resulted in dramatically reduced PGRN secretion which is thought to be the result of endocytosis and trafficking of PGRN to the lysosome [79, 80]. Splice variants of SORT1 have been shown to abrogate PGRN binding by producing a non-functional receptor [81].

1.6.3 Inflammation and Wound Healing

One group has suggested that PGRN can inhibit the cytokine TNF-α by binding to its receptor, tumour necrosis factor receptor (TNFR) thereby acting as anti-inflammatory in a mouse model of arthritis [82]. However, another group have recently demonstrated a lack of any direct interaction between PGRN and either TNFR1 or 11 implying that loss of PGRN in the central nervous system (CNS) and its concomitant neuroinflammation are not mediated by the blocking of TNF binding but the perturbation of other receptors which are, as yet, unidentified [83].

Granulins A and B block neutrophil activation by TNF-α and produce a pro-inflammatory response via activation of IL-8 in epithelial cells [82, 84].

PGRN has also been shown to increase proliferation of neutrophils, endothelial cells and macrophages in damaged tissue thus indicating its importance in wound healing [85].

-23-

1.6.4 Cell Signalling

Elevated levels of PGRN have been shown to stimulate various growth factor signalling pathways, including phosphatydyl inositol-3 (PI3K) kinase, extracellular regulated kinase (ERK) and p70S6K [86-88]. Recently suggestions have been made that disruption of the Wnt signalling pathway caused by decreased levels of PGRN may be a contributory factor in FTLD –TDP [89].

1.6.5 Tumorigenesis

PGRN is important in several processes relating to tumour formation including proliferation, migration, invasiveness, anchorage independence and resistance to chemotherapy [88, 90, 91]. Significantly there seems to be a correlation between levels of GRN and upregulation of VEGF promoting angiogenesis in oesophageal squamous cell carcinoma [92].

High levels of PGRN have also been shown to be a good indicator of tumour severity as PGRN knockdown reduces tumour formation [93] .

1.6.6 Progranulin and Neurodegeneration

It has recently been suggested that there may be a link between progranulin and the lysosomal storage disease neuronal ceroid lipofucsinosis (NCL) [94].Two siblings having a homozygous GRN null mutation were clinically diagnosed with early-onset NCL although this has not yet been confirmed neuropathologically [95, 96].

The question as to whether the lack of PGRN leads to increased levels of lysosomal proteins Transmembrane Protein 106B (TMEM106B), Lysosomal –Associated Membrane Protein (LAMP1) and Cathepsin D (CTSD) in neurons remains to be answered [97].

It has also been suggested that GRN mRNA can be bound by TDP-43 causing destabilisation and a reduction in PGRN levels [98]. This stress further reduces GRN levels by causing TDP-43 to be mis-located to the cytoplasm by a positive feedback loop mechanism. This interactive relationship between GRN mRNA and TDP-43 may result in a PGRN deficient threshold being reached in neurons leading to neurodegeneration [98].

-24-

1.7 The Functional Analysis of Granulins.

Initial purification studies of various tissues and cellular extracts yielded several peptides 6kDa in length that were all encoded by the GRN gene which translated into the full length precursor protein PGRN [99-101] (Figure 1.4).

This would seem to indicate that the individual granulins have functions which differ from the mature secreted protein. As previously mentioned this has already been shown in relation to the pro-inflammatory activation of IL-8 by granulins A and B [82, 84].

Cloning of the individual granulins and transfection into mammalian cells will allow us to examine the effects of the proteins under physiologically relevant conditions. As there is no known read-out for granulin or PGRN activation, the biological effects of the peptides and the parent molecules will be initially studied using cell treatment followed by RNA expression as measured by RNASeq.

1.8 Aims and Objectives.

The overall aim of the project is to identify differences/similarities of the granulins and progranulin in terms of functionality within the cell.

This will be achieved by:-

 Cloning of progranulin and the granulins into an appropriate expression vector.  Expression of the individual proteins in mammalian cells.  Purification of pure, active recombinant proteins.  Treatment of differentiated neuroblastoma cells with the recombinant proteins.  Production of RNA for RNASeq analysis.  Analysis of the data thus produced to identify differentially expressed genes and possible disease pathways in which progranulin and the granulins they may be involved.

-25-

Chapter 2 Progranulin and Granulin Cloning

-26-

2.1 Introduction

In order to elucidate the specific functions of the individual granulins compared to the full length protein it was necessary to produce sufficient quantities of recombinant protein for each. In order to accomplish this a cloning strategy was designed which allowed the granulin/progranulin sequences to be incorporated into a cloning vector and subsequently sub-cloned into an expression vector.

We designed constructs for all granulins and paragranulin according to the following pattern:

BamH1 site-Kozak sequence-Start codon-GRANULIN-His tag-stop codon-Not1 site

The designs were then sent to Eurofins MWG/Operon to be synthesised, and amplified by PCR. They were then cloned into the pCR®2.1 vector using TOPO®-TA cloning and the sequences verified. When the plasmid constructs were received from MWG they were dissolved in TE Buffer to a concentration of 100ng/µl before use.

-27-

Figure 2.1 pCR®2.1 TOPO® Cloning Vector.

Map of the pCR®2.1-TOPO® cloning vector into which the individual granulin constructs were inserted via the multiple cloning site (MCS). The insert is high-lighted in black at the top of the diagram. The antibiotic resistance genes for ampicillin and kanamycin are down-stream of the MCS.

The sequences used, with the exception of paragranulin, were designed to include the whole granulin domain plus three amino acids preceding and one amino acid following from the linker regions between the granulins as defined by Bhandari et al [102] and with a start codon at the beginning of each. The paragranulin sequence used the signal peptide, paragranulin domain and the linker region up to the last three amino acids which form the beginning of the Grn G sequence.

MWG verified sequences of His-tagged granulins were:-

Paragranulin:MWTLVSWVALTAGLVAGTRCPDGQFCPVACCLDPGGASYSCCRPLLDKWPTTLSRHLHHHHHH

TGGACCCTGGTGAGCTGGGTGGCCTTAACAGCAGGGCTGGTGGCTGGAACGCGGTGCCCAGATGGTCAGTTCTGCCC TGTGGCCTGCTGCCTGGACCCCGGAGGAGCCAGCTACAGCTGCTGCCGTCCCCTTCTGGACAAATGGCCCACAACACT GAGCAGGCATCTGCATCATCACCATCACCACTGA

GRN1/G : MGGPCQVDAHCSAGHSCIFTVSGTSSCCPFPEAVACGDGHHCCPRGFHCSADGRSCFHHHHH

-28-

GGTGGCCCCTGCCAGGTTGATGCCCACTGCTCTGCCGGCCACTCCTGCATCTTTACCGTCTCAGGGACTTCCAGTTGCT GCCCCTTCCCAGAGGCCGTGGCATGCGGGGATGGCCATCACTGCTGCCCACGGGGCTTCCACTGCAGTGCAGACGGG CGATCCTGCTTCCATCATCACCATCACCACTGA

GRN2/F : MAIQCPDSQFECPDFSTCCVMVDGSWGCCPMPQASCCEDRVHCCPHGAFCDLVHTRCIHHHHHH

GCCATCCAGTGCCCTGATAGTCAGTTCGAATGCCCGGACTTCTCCACGTGCTGTGTTATGGTCGATGGCTCCTGGGGG TGCTGCCCCATGCCCCAGGCTTCCTGCTGTGAAGACAGGGTGCACTGCTGTCCGCACGGTGCCTTCTGCGACCTGGTT CACACCCGCTGCATCCATCATCACCATCACCACTGA

GRN3/B : MVMCPDARSRCPDGSTCCELPSGKYGCCPMPNATCCSDHLHCCPQDTVCDLIQSKCLHHHHHH GTCATGTGTCCGGACGCACGGTCCCGGTGCCCTGATGGTTCTACCTGCTGTGAGCTGCCCAGTGGGAAGTATGGCTGC TGCCCAATGCCCAACGCCACCTGCTGCTCCGATCACCTGCACTGCTGCCCCCAAGACACTGTGTGTGACCTGATCCAGA GTAAGTGCCTCCATCATCACCATCACCACTGA

GRN4/A : MDVKCDMEVSCPDGYTCCRLQSGAWGCCPFTQAVCCEDHIHCCPAGFTCDTQKGTCEHHHHHH

GATGTGAAATGTGACATGGAGGTGAGCTGCCCAGATGGCTATACCTGCTGCCGTCTACAGTCGGGGGCCTGGGGCTG CTGCCCTTTTACCCAGGCTGTGTGCTGTGAGGACCACATACACTGCTGTCCCGCGGGGTTTACGTGTGACACGCAGAA GGGTACCTGTGAACATCATCACCATCACCACTGA

GRN5/C : MVPCDNVSSCPSSDTCCQLTSGEWGCCPIPEAVCCSDHQHCCPQGYTCVAEGQCQHHHHHH

GTCCCCTGTGATAATGTCAGCAGCTGTCCCTCCTCCGATACCTGCTGCCAACTCACGTCTGGGGAGTGGGGCTGCTGT CCAATCCCAGAGGCTGTCTGCTGCTCGGACCACCAGCACTGCTGCCCCCAGGGCTACACGTGTGTAGCTGAGGGGCA GTGTCAGCATCATCACCATCACCACTGA

GRN6/D : MIGCDQHTSCPVGQTCCPSLGGSWACCQLPHAVCCEDRQHCCPAGYTCNVKARSCEHHHHHH

ATCGGCTGTGACCAGCACACCAGCTGCCCGGTGGGGCAGACCTGCTGCCCGAGCCTGGGTGGGAGCTGGGCCTGCT GCCAGTTGCCCCATGCTGTGTGCTGCGAGGATCGCCAGCACTGCTGCCCGGCTGGCTACACCTGCAACGTGAAGGCTC GATCCTGCGAGCATCATCACCATCACCACTGA

GRN7/E : MDVECGEGHFCHDNQTCCRDNRQGWACCPYRQGVCCADRRHCCPAGFRCAARGTKCLHHHHHH

GACGTGGAGTGTGGGGAAGGACACTTCTGCCATGATAACCAGACCTGCTGCCGAGACAACCGACAGGGCTGGGCCT GCTGTCCCTACCGCCAGGGCGTCTGTTGTGCTGATCGGCGCCACTGCTGTCCTGCTGGCTTCCGCTGCGCAGCCAGGG GTACCAAGTGTTTGCATCATCACCATCACCACTGA

Key:- Abc denotes the signal peptide in the paragranulin sequence

Abc – His tag

2.2 Construct DNA Amplification

In order to amplify the individual constructs it was necessary to transform and select clones containing the inserted DNA.

-29-

2.2.1 Transformation and Clonal Selection

50ng of DNA was added to 50µl XL10 Gold ultracompetent cells (Stratagene) in a pre- chilled 1.5ml microfuge tube, mixed by gentle tapping and placed on ice for 30 minutes. The samples were then heat shocked by placing in a water bath at 42°C for exactly 30 seconds and immediately replacing on ice for 2 minutes. 250µl sterile super optimal broth containing glucose (SOC) medium was added to each tube before incubating in a shaking incubator for 1 hour at 37°C. 50µl of cell suspension was then spread on a pre-warmed LB Agar plate containing 50µg/ml Kanamycin. The plates were allowed to dry before inverting and placing in a 37°C incubator overnight. Colonies were then picked from each plate and used to inoculate 4mls of Luria Broth (LB) medium containing 50µg/ml Kanamycin. Cultures were grown overnight in a shaking incubator (37°C, 225rpm) and plasmid DNA purified from each using a Qiagen Miniprep kit. DNA concentrations were measured using a Nanodrop ND 1000 spectrophotometer (Thermo Scientific).

2.2.2 Restriction Digestion Approximately 1µg of the resulting purified DNA was digested by adding 1µl each of Not1 and BamH1 high fidelity restriction enzymes (20,000 units/ml) ,1µl of bovine serum albumin (BSA) (100µg/ml) and 5µl of 10x reaction buffer (New England Biolabs) and making up to 50µl with sterile water. The reaction mixture was incubated at 37°C for 1.5 hours followed by heat inactivation for 20 minutes at 65°C.

2.2.3 Agarose Gel Electrophoresis

The digested plasmid was then loaded onto a 1.5% agarose gel prepared by dissolving 1.5% w/v agarose in 40mM Tris Acetate, 1mM ethylenediaminetetraacetic acid (EDTA) buffer (TAE), to which 1ug/ml ethidium bromide (Sigma 10mg/ml solution) was added. The gel was run at 60 volts (Bio-Rad power pack) in TAE buffer for 60-90 minutes before being visualised under UV light. The digested bands were excised with a razor blade and removed to fresh 1.5 ml microfuge tubes for purification. The pEGFP-N1 vector (Figure 2.2) was similarly digested and loaded onto a 0.75% agarose gel. The linearised vector

-30-

was then excised from the gel and both inserts and vector DNA purified using a Qiagen Gel Extraction Kit according to the manufacturer’s protocol.

2.2.4 Ligation

The linearised vector was treated with Calf Intestinal Phosphatase (CIP) (NEB) as per the standard protocol in order to dephosphorylate the 5’ ends and prevent self-ligation. The granulin inserts and pEGFP-N1 purified DNAs were then ligated together in a 20µl reaction volume containing 50µg vector,1µl T4 Quick Ligase, 10µl 2x Reaction buffer (both NEB) and using molar ratios of 1:1 and 1:3, vector: insert.

Figure 2.2 pEGFP-N1 Expression Vector.

The pEGFP-N1 expression vector is under the control of the CMV promoter. The MCS contains the restriction sites BamHI and NotI. Digestion of the plasmid and vector with BamHI and NotI restriction enzymes and ligation as previously described (2.2.2) enabled the granulin inserts to be sub-cloned into the pEGFP-N1 vector and removal of the GFP tag from the vector as we planned to use the His tag contained within the granulin constructs for detection and purification.

-31-

2.2.3 Transient Transfection Human embryonic kidney (HEK 293) cells were plated out at a density of 1.5-2.5x105 per well of a 6-well plate (Corning) in 2mls of Dulbecco’s Modified Eagles Medium (DMEM) containing 10% heat-inactivated foetal bovine serum (FBS), 2mM L-glutamine and 20 units/ml penicillin and 20µg/ml streptomycin.

Transfection was carried out under standard conditions i.e. 1:2 DNA to jetPRIME ratio over a range of DNA concentrations (w/v) by adding granulin plasmid DNA to 200µl/well jetPRIME (Polyplus) transfection buffer followed by the addition of jetPRIME transfection reagent. Samples were vortexed for 10 seconds and centrifuged before the transfection reagent addition then vortexed and centrifuged again before incubation for 10 minutes at room-temperature. The transfection mix was then added to the wells at 60-80% confluency and incubated at 37°C for 4 hours at which point the growth medium was replaced with complete DMEM as above. The same protocol was followed using human neuroglioma (H4) cells.

Cells were harvested at 24 hour and 48 hour time points. Medium was aspirated from individual plates and the cells rinsed with ice-cold PBS. Cells were lysed by the addition of 100ul of RIPA buffer containing protease inhibitors (see Materials section) and transferred to 1.5 ml microfuge tubes which were then frozen at -80°C.

2.2.4 Gel Electrophoresis

Samples were thawed, 4 x Lithium Dodecyl Sulphate (LDS) buffer (Life Technologies) added and heated at 70°C for 5 minutes. After cooling they were loaded onto a 4-12% Bis Tris Novex NuPage gel (Life Technologies) with 12µl Novex Sharp pre-stained protein standard. Proteins were separated by electrophoresis using MES running buffer (20 x solution (Life Technologies) diluted with distilled water and a charge of 150 volts for approximately 1 hour.

-32-

2.2.5 Western Blotting

Four layers of filter paper and a layer of nitrocellulose were cut to approximately the same size as the gel and soaked in transfer buffer for >5 minutes. A gel sandwich was assembled by placing two layers of filter paper, the membrane, gel and then the remaining two layers of filter paper on top being careful to remove any air bubbles between layers. The gel sandwich was then placed in a Bio-Rad Semi-dry transfer apparatus and proteins transferred by electro-blotting from gel to membrane at 15 volts for 1 hour. The membrane was rinsed briefly in TBST (50mM Tris, 150mM NaCl, 0.05% Tween 20 pH7.6) before blocking for 1 hour at room temperature in 5% BSA (bovine serum albumin) dissolved in TBST followed by overnight incubation, with shaking, at 4°C in 10 mls 5% BSA/TBST containing mouse anti-His antibody (Agilent) diluted at 1:2000. After 3 x 10 minute washes in TBST, the membrane was incubated in horseradish peroxidase conjugated goat anti-mouse (Santa Cruz) antibody diluted at 1:4000 in 10mls 5% non-fat dried milk powder for 1 hour at room temperature. Blots were visualised on an ImageQuant 300 following 3 x 10 minute washes in TBST and chemiluminescent detection using ECL reagents (GE) as per the manufacturer’s instructions. Several experiments were done to try and optimise conditions however we were unable to detect any protein.

Following discussion with Dr Edward McKenzie (MIB Manchester) it was felt that this was probably due to problems with protein instability as a result of the small size of the expressed granulins (~6-7kDa). It was decided to use the vector pHTN HalotagR CMV-neo (Promega), (a gift from Dr Edward McKenzie), for protein expression as firstly, the size of the Halotag (34kDa) stabilised the expressed granulin and secondly it was easily removed during purification by covalent binding to a Halotag resin and cleavage using a TEV protease (both Promega) to produce a pure untagged protein.

By using the original constructs as templates and designing new primers it was possible to amplify the granulins by PCR, clone into the pCR2.1 vector (Figure 2.1) using the TA Cloning Kit (Invitrogen) and sub-clone into the pHTN Halotag® vector (Promega) (Figure 2.3) using the EcoRI and NotI sites (the pHTN vector does not contain a BamHI site). The

-33-

intermediate TA cloning step was used to enable larger amounts of granulin insert DNA to be produced for Halo-tag cloning and obviate the need for multiple PCR steps.

Figure 2.3 Halotag® CMV-neo Vector.

.

The figure shows the pHTN Halotag® CMV-neo vector which allows insertion of the granulin into the MCS down-stream of the Halotag® and TEV cleavage site and is controlled by the T7 promoter.

PRIMER DESIGN FOR GRANULIN CLONING

Pattern for constructs was:-

EcoRI site-Granulin-His Tag-Stop codon-NotI site with the following primer sequences designed to amplify the granulins.

Reverse Primer - same for all. 3’ GCGGCCGCTCAGTGGTGATG 5’

PARAGRANULIN:-Forward Primer 5’ GCAGAATTCGCTGGAACGCGGTGC 3’

GRN G:- Forward Primer 5’ GCAGAATTCGGTGGCCCCTGC 3’

GRN F:- Forward Primer 5’ GCAGAATTCGCCATCCAGTGC 3’

GRN B:- Forward Primer 5’ GCAGAATTCGTCATGTGTCCG 3’

GRN A:- Forward Primer 5’ GCAGAATTCGATGTGAAATGT 3’

GRN C:- Forward Primer 5’ GCAGAATTCGTCCCCTGTGAT 3’

GRN D:- Forward Primer 5’ GCAGAATTCATCGGCTGTGAC 3’

GRN E:- Forward Primer 5’ GCAGAATTCGACGTGGAGTGT 3’ -34-

All primers were 100pmol/µl in PCR grade water and 10pmol/µl stock solutions were made by adding 10µl of forward and 10µl of reverse primer and diluting to 100µl with PCR grade water.

2.3.1 PCR Protocol

The template constructs (100ng/µl) were all diluted 1:10 in PCR grade water.

The following were added to 0.2ml thin walled PCR tubes:- Template 10µl, 10x Reaction Buffer 5µl, 50mM dNTP’s 0.5µl, primers (10pmol/µl stock) 2µl, PCR grade water 31.5µl and finally 2.5 units of Qiagen Hotstart Taq polymerase making a total of 50µl. The tubes were placed in a thermal cycler and heated to 94°C for 10 minutes followed by 25 cycles of 94°C 1 minute, 55°C 1 minute and 72°C 1 minute, finishing with 10mins at 72 °C. 5µl of each PCR product was then run on a 1.5% agarose gel to check the size of the products. Since cloning into the pCR®2.1 vector (Figure 2.1) is dependent on 3’ A-overhangs which degrade over time the ligation reactions were set up immediately in order to achieve optimal ligation efficiencies.

For each ligation reaction the following were added to 0.2ml thin-walled PCR tubes and incubated at 14°C overnight (thermal cycler):- Fresh PCR product 2µl, 10x Ligation Buffer 1µl, pCR2.1 vector (25ng/µl) 2µl, PCR water 4µl and T4 DNA Ligase (4.0 Weiss units) 1µl. 2µl of each ligation reaction was used to transform 50µl TOP10 cells (TA Cloning Kit) as described previously. The LB Agar plates containing 50µg/ml Kanamycin were equilibrated at 37°C for 30 minutes, spread with 40µl of 40mg/ml X-Gal (for blue/white colony selection) and allowed to dry before spreading 50µl of each transformation vial per plate. Plates were allowed to dry, inverted and incubated at 37°C overnight before placing at 4°C for 2-3 hours to allow for proper colour development. Approximately ten white colonies from each plate were picked, grown in LB/Kanamycin overnight and DNA prepared using a Qiagen Miniprep kit.10µl of each eluted DNA sample was digested with EcoR1 and Not1 restriction enzymes and separated on a 1.5% agarose gel as previously described (2.2.2) for Not1/BamH1 digestion to check for the correct size insert. One clone for each granulin was then selected for sequencing using T7 forward and M13 reverse primers.

-35-

2.3.2 Sequencing.

In order to validate any clone it is necessary to confirm that the sequence is in frame and is correct in terms of its amino acid sequence as unexpected recombinations can occur spontaneously leading to the production of aberrant protein products. Fluorescent DNA sequencing uses a chemistry in which the dyes, one for each base, are attached to the di- deoxy dNTPs. DNA template, unlabelled primer, buffer, the four dNTP’s, the four fluorescently labelled di-deoxy dNTPs, and DNA polymerase are added to the reaction tube. Incorporation of the di-deoxy dNTP terminates the PCR reaction for that particular fragment leading to the production of a series of PCR fragments of differing lengths each with a fluorescently labelled amino acid at the 3’ end. This allows the nucleotide sequence to be determined from the colour of the fluorescence produced by the product fragments (Figure 2.4).

Figure 2.4 Diagram of one cycle of dye terminator (di-deoxy dNTP) cycle sequencing.

The PCR reaction of cycle sequencing results in the incorporation of di-deoxy dNTPs labelled with four different dyes. The resultant products are electrophoresed and separated by size allowing the fluorescence of the individual bases to be read by a laser. The sequencing software then converts the fluorescent peaks into bases allowing the sequence to be determined.

Reagents used for the sequencing reactions were as follows:-

100pmol/µl forward and reverse primer stocks diluted to a concentration of 3.2pmol/µl in PCR grade water, 250-500ng template plasmid DNA per reaction, Big Dye (vs3.1 Applied Biosystems) and 5x Sequencing Buffer (Agilent).

For each reaction:- 1µl Template DNA, 1µl Primer, 2µl 5x Sequencing Buffer, 1µl Big Dye and 5µl PCR water.

-36-

PCR programme

96 °C for 1 minute followed by 25 cycles of:-

96°C 10 seconds, 50 °C 10 seconds, 60 °C 4minutes

Followed by ramp to 4 °C.

To purify the DNA and remove any non-incorporated dNTP’s from the reaction the tubes were centrifuged briefly and the contents transferred to sterile 1.5ml microfuge tubes. Each PCR tube was rinsed with 10µl of PCR grade water which was also added to the microfuge tubes with 2µl 3M sodium acetate and 50µl 100% ethanol. The tube contents were then mixed by gentle vortexing and incubated at room temperature in the dark for 15 minutes.

After centrifugation for 30 minutes at 14000rpm to pellet the DNA the ethanol was removed from the tubes by inversion. The DNA pellets were then washed by the addition of 70µl 70% ethanol and centrifuging at 14000rpm for 15 minutes immediately after which the ethanol was carefully removed using a pipette and the pellets allowed to air dry.

The DNA samples were then run on an Applied Biosystems 3730 DNA Analyzer and sequence chromatograms examined using Sequencher v5.1 Demo version.

Having ascertained that the sequences were correct it was now possible to sub-clone the granulins into the pHTN Halotag® vector by digesting the pCR®2.1-Granulin plasmids and pHTN Halotag vector with EcoR1 and Not1, agarose gel separation and purification as described in section 2.2.2/3 and shown in Figure 2.5.

The purified granulin inserts and linearised vector were then ligated using a 1:2 vector: insert ratio and a vector only control as follows:- Vector 50ng, Insert 3.2ng, 2x ligation buffer 10µl, PCR water to a final volume of 20µl followed by the addition of 1µl of T4 ligase and incubation at room temperature for 2 hours. 2µl of each ligation reaction was transformed with XL10 Gold ultracompetent cells and 50µl spread on LB Agar plates containing 100µg/ml ampicillin followed by overnight incubation at 37°C.

-37-

Colonies were screened by Qiagen mini-prep, restriction digest and agarose gel electrophoresis (Figure 2.5). Clones which contained the correct size digestion products were then Sanger sequenced for verification.

Figure 2.5. Examples of agarose gels.

Grn C Grn D

200bp

Both gels show NotI/EcoRI restriction digests of PCR®2.1 clones of granulins. The arrows highlight the expected 200bp fragment used for insertion into Halotag™ vector.

Granulin inserts for sub-cloning into the 200bp Halotag® vector

1 2 3 4 5 6

Lanes 1 and 6 Bioline Hyperladder IV. Lane 2-Paragranulin, Lane 3-Grn A, Lane 4-Grn B, Lane 5- Grn F.

Unfortunately granulins A, B, F and paragranulin all contained extra sequence which was traced back to the template constructs used for TA cloning. In order to correct this new

-38-

forward primers were designed which used the Pvu1 restriction site not contained in the pCR®2.1 plasmid but present in the pHTN Halotag vector.

PARAGRANULIN :-Forward Primer 5’ GCGATCGCTGCTGGAACGCGGTGC 3’

GRN F:- Forward Primer 5’ GCGATCGCTGCCATCCAGTGCCCT 3’

GRN B:- Forward Primer 5’ GCGATCGCTGTCATGTGTCCGGAC 3’

GRN A:- Forward Primer 5’ GCGATCGCTGATGTGAAATGTGAC 3’

PCRs and cloning into pCR®2.1 and sub-cloning into pHTN Halotag® were all performed as previously described. Sequences of positive clones were verified by Sanger sequencing (2.3.2/3).

-39-

Chapter 3 Protein Expression and Purification

-40-

3.1 Introduction

The use of mammalian cell expression enables the production (usually) of active post- translationally modified protein. However, it is not a good method for the production of large quantities of material. As our initial aim was to produce sufficient amounts of the granulins and progranulin for cell treatments and next generation sequencing experiments as well as possible antibody production it was necessary to stably transfect cells in order to have a continuing supply of the proteins. Stable transfection allows positive colonies from one transient transfection to be maintained in culture by the use of a selective antibiotic thereby conserving plasmid DNA. The expressing cells can then be grown in bulk, harvested and the protein of interest purified.

3.2 Stable Transfection

HEK 293 cells were plated at a density of 2-2.5x106 in a T75cm2 flask (Corning) in 10mls of Dulbecco’s Modified Eagles Medium (DMEM) containing 10% heat-inactivated foetal bovine serum (FBS), 2mM L-glutamine and 20 units/ml penicillin and 20µg/ml streptomycin. Transfection was carried out by adding 5µg of granulin plasmid DNA to 500µl jetPRIME (Polyplus) transfection buffer followed by the addition of 10µl jetPRIME transfection reagent. Samples were vortexed for 10 seconds and centrifuged before the transfection reagent addition then vortexed and centrifuged again before incubation for 10 minutes at room temperature. The transfection mix was then added to a T75cm2 flask at 60-80% confluency and incubated at 37°C for 4 hours at which point the growth medium was replaced with complete DMEM as above. The following day the medium was again replaced with complete DMEM containing 400µg/ml G418 (Promega) to select for cells in which the granulin-halotag plasmid DNA had been incorporated.

This process was repeated until the cells numbers started to increase. This indicated that the only cells remaining and growing were G418 resistant and therefore contained the granulin-halotag™. Once the cells had reached 80-90% confluence they were expanded by removing the medium by aspiration, resuspending in 10mls complete DMEM containing 400µg/ml G418 and used to seed 5 x 150mm culture dishes per 75 cm2 flask.

-41-

3.3 Protein Purification.

3.3.1 Cell harvesting

Medium was aspirated from individual dishes and the cells rinsed with ice-cold PBS. Cells were gently scraped into a 50ml Falcon tube and centrifuged at 2000rpm to pellet. Surplus liquid was removed by aspiration and the cell pellet frozen at -80°C.

3.3.2 Halotag resin purification

Individual cell pellets from an amalgamation of 40 x 150mm culture dishes were lysed by resuspending in 10mls Mammalian Lysis Buffer to which was added 200µl 50 x Protease Inhibitors and 200µl RQ1 DNase (all Promega) followed by rotation at room temperature for 15 minutes. The lysed cells were diluted 1:3 with PBS containing 1mM dithiothreitol and 0.05% Nonidet P40 (protein purification buffer) and centrifuged for 30 minutes at 10,000 x g and 4°C. The resulting supernatant was then added to 1.25 mls Halotag™ Resin (Promega) which had been previously equilibrated by resuspending in protein purification buffer, rotating for 5 minutes at room temperature and centrifuged at 1500 x g for a total of 5 washes. The cell supernatant was rotated with the Halotag resin for 2 hours at room temperature then centrifuged at 1500 x g for 5 minutes and the supernatant removed to a fresh tube. The resin was washed for 3 x 30 minutes by resuspending and centrifuging in (1) PBS + 350mMNaCl (2) PBS diluted 1:2 with distilled water and (3) protein purification buffer. Cell supernatant was analysed before and after binding to the Halotag® resin by gel electrophoresis and Western blotting as described in section 2.2.5. The blot was incubated overnight at 4°C with a Halotag® polyclonal primary antibody (Promega) diluted at 1:2000 (Figure 3.1).

-42-

Figure 3.1 Western blot of over-expressed Halo®-tagged granulins pre and post binding to Halotag resin.

40kDa

1 2 3 4 5 6

1 – Grn A total lysate, 2 – Grn A lysate post-binding. 3 – Grn B total lysate, 4 – Grn B lysate post-binding. 5 – Grn C total lysate, 6 – Grn C lysate post-binding. The blot shows that majority of the expressed granulins have bound to the Halotag® resin

The resin was then resuspended in 2.5 mls of protein purification buffer to which was added 100µl TEV protease (Promega). The cleavage reaction was allowed to proceed by rotating at 4°C overnight. The resin was centrifuged at 1500 x g and the supernatant removed to a fresh tube being careful to minimise resin transfer (Eluate 1). A further 2.5mls of protein purification buffer was added to the resin and rotated at room temperature for 30 minutes before centrifuging and removing as previously (Eluate 2). The two eluates were then pooled, 50 x protease inhibitors added (Promega), centrifuged at 10,000 x g for 1 minute and the supernatant passed through a spin filter to remove any resin particles. The resulting solution was then concentrated in a centrifugal filter (Amicon ultra 4ml 3kDa MWCO) to a volume of approximately 250µl.

The mature full-length progranulin contains a signal peptide allowing the protein to be secreted into the medium. This was removed from the cells, centrifuged at 2000 rpm to remove any cell debris, concentrated in a centrifugal filter to approximately 15 mls (from 300 mls medium, Amicon Ultra 15ml 30kDA MWCO) and bound to the Halotag® resin as described for the granulin peptide cell lysates. A small sample of the TEV cleaved eluate

-43-

was loaded onto a 4-20% acrylamide gradient gel and electrophoresed as described in section 2.2.4.

3.3.3 Coomassie Staining

The gel was rinsed briefly in distilled water and fixed overnight in a solution containing 50% methanol and 10% glacial acetic acid. The solution was removed and the gel stained with Coomassie Blue R-250 solution (Bio-Rad) for between 30 minutes. The gel was then destained with several changes of 40% methanol, 10% glacial acetic acid until all background colour was removed and bands were clearly visible (Figure 3.2). All steps were performed with gentle agitation.

Figure 3.2 TEV Cleaved Granulin

40kDa Halo-tagged GRN D

10kDa GRN D

1 2 3

Coomassie-stained gel showing the over-expressed halo-tagged GRN in the whole cell lysate in lane 1, a sample of lysate after binding to the Halotag resin in Lane 2 and the TEV protease cleaved granulin in Lane 3. TEV cleavage releases the purified granulin from the resin but still produces some higher molecular weight contaminating bands.

-44-

3.3.4 High-Performance Liquid Chromatography (HPLC) In order to obtain pure protein it was necessary to further purify the resin eluates by gel filtration chromatography. Samples were passed through a column and the proteins separated according to size. This enabled the removal of any higher or lower molecular weight contaminants. The concentrated eluates were loaded onto a 200µl loop and injected onto a Superdex Peptide 10/300GL column. The column was run on an Ettan LC machine (GE) with PBS containing 1mM dithiothreitol and fractions collected (Figures 3.3 and 3.4).

Figure 3.3 HPLC trace of gel filtration column fractions.

mAu Fractions 800 C4,5 and 6

700

600

500

400

300

200

100

0

0.0 5.0 10.0 15.0 20.0 25.0 30.0 mls

0.0 0.0 .0 0.0 0 .0 C4,5 and 6 fractions correspond to the protein peak for progranulin high-lighted by the arrow

.

-45-

Granulin D 200ul 160613:10_UV1_214nm Granulin D 200ul 160613:10_UV2_280nm Granulin D 200ul 160613:10_UV3_0nm Granulin D 200ul 160613:10_Cond Granulin D 200ul 160613:10_Cond% Granulin D 200ul 160613:10_Conc Granulin D 200ul 160613:10_Pressure Granulin D 200ul 160613:10_Fractions Granulin D 200ul 160613:10_Inject Granulin D 200ul 160613:10_Logbook

mAU mS/cm

Figure400 3.4 Example of Granulin HPLC trace.

80 mAU 300300

60 Fraction A10

200200

40

100100

20

0

1A1 1A2 1A3 1A4 1A5 1A6 1A7 1A8 1A9 1A101A111A121B121B111B10 1B9 1B8 1B7 1B6 1B5 1B4 1B3 1B2 1B1 1C1 1C2 1C3 1C4 1C5 1C6 1C7 1C8 1C9 1C101C11 Waste 0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 ml 0.0 5.0 10.0 15.0 20.0 25.0 30.0 mls HPLC trace of gel filtration of granulin D. Fraction A10 corresponds to the granulin peak denoted by the black arrow.

HPLC fraction A10 for each granulin was concentrated and fractions C4, 5 and 6 for PGRN were pooled and concentrated (Amicon Ultra filters 0.5ml 3kDa MWCO). The protein concentration of each sample was then determined using the Qubit protein assay kit as per the manufacturer’s protocol.

3.4 Gel Electrophoresis

Equal amounts of each granulin were loaded onto a 4-20% acrylamide gradient gel and run under the same conditions as described in section 2.2.4.

3.5 Silver Staining.

The gel was rinsed briefly in distilled water and fixed overnight in 50% methanol, 12% acetic acid, 0.05% formalin (35% Formaldehyde).Following 3x20 minute washes in 35% ethanol the gel was sensitised by incubating for 2 minutes in 0.02% sodium thiosulphate solution. After washing for 3 x 5 minutes in distilled water it was then incubated for 20 minutes in silver nitrate solution: 0.2% AgNO3, 0.076% formalin (35% Formaldehyde)

-46-

before 2 further distilled washes of 1 minute each. The gel was developed in 6% sodium carbonate, 0.05% formalin (35% Formaldehyde) and 0.0004% sodium thiosulphate until bands were clearly visible. The reaction was stopped by the addition of 50% methanol, 12% acetic acid (Figure 3.5).

Figure 3.5 Silver–stained gel of purified granulins

15kDa 10kDa

3.5kDa 1 2 3 4 5 6 7 8

Lane 1- Paragranulin, 2 – Grn A, 3 – Grn B , 4 – Grn C, 5 – Grn D, 6 – Grn E, 7 – Grn F, 8 – Grn G

The figure shows that although ostensibly equal amounts of protein were loaded there is a difference in the band intensity on the gel. There is also a slight variation in the observed size of the granulins, notably Granulin C which runs at 15kDa.

Full-length progranulin including the signal peptide was cloned into the pHTN Halotag® vector by Dr Janis Bennion Callister. HEK 293 cells were transfected with the plasmid DNA and a stable cell line produced by treating with 400µg/ml G418 (Promega).The presence of the signal peptide allowed the mature halo-tagged progranulin to be secreted into the medium which was centrifuged to remove any cells and concentrated before being bound to the Halotag® resin. Subsequent cleavage by the TEV protease and purification by HPLC produced pure progranulin protein which had an apparent molecular weight of 68kDa when run on a 4-20% acrylamide gradient gel under reducing conditions (Figure 3.6).

-47-

Figure 3.6 Purified Progranulin

75kDa Progranulin

50kDa

Coomassie-stained gel of a sample of purified PGRN which has undergone HPLC purification and concentration.

-48-

Chapter 4 Cell Treatment and Next Generation Sequencing

-49-

4.1 Introduction.

The aim of the experiment was to use next generation sequencing technology to identify and compare global gene expression changes in differentiated neural cells when treated with exogenous granulin peptides and full length progranulin.

4.2 Differentiating SH-SY5Y cells

SH-SY5Y human neuroblastoma cells (ATTC® CRL-2266™) were cultured in minimum essential medium (Life Technologies 21090-022) containing 10% heat-inactivated foetal bovine serum (Life Technologies 10270-106), 2mM L-Glutamine (Sigma) 100 µg streptomycin/ml,100i Penicillin/ml (Sigma) and 1% Non-essential amino acids (Life Technologies 11140-035) and plated at 8-10 x103 per cm3 in T-75 flasks (Corning). After 24 hours growth in a gassing incubator at 37°C, 5% CO2 the medium was changed to N2 (Dulbecco’s Modified Essential Medium:F12 containing 15mM HEPES and L-Glutamine (Life Technologies 31330-038) containing 2mM L-Glutamine (Sigma), 100 µg streptomycin/ml,100i Penicillin/ml (Sigma) 1% N2 100x (Life Technologies 17502-048) (final concentration) and 1 µM all-trans retinoic acid (Sigma R2625) (1mM dissolved in 99% ethanol and used at 1:1000). After 72 hours, five confluent T75 cm2 flasks of semi- differentiated SH-SY5Y cells were pelleted and resuspended in 25mls of N2 medium containing 1 µM retinoic acid (as above). I.0ml of the resulting cell suspension was then added to each of sixteen 10cm2 cultures dishes (Corning) each containing a further 8mls of N2 medium. Further differentiation was facilitated by media changes every 3-4 days until the cells were terminally differentiated after 3 weeks in culture.

4.3 Cell Treatment

Previous experiments undertaken by Dr Janis Bennion Callister using different concentrations of progranulin had shown that treating cells with 0.5µg/ml progranulin for 30 minutes was optimal for detecting changes in some down-stream cell signalling pathways (data not shown). The granulins at 6 -7kDa are approximately 10-fold smaller in size than progranulin (68kDa) therefore in order for the treatments to be comparable a 10-fold lower amount of the granulins (i.e. 50ng/ml) was used. -50-

Cells were treated by replacing the medium in each dish with fresh medium (2.5mls/dish) containing either 0.5 µg/ml progranulin or 50ng/ml of the individual granulins as per table 4.1. Control dishes contained medium only.

Table 4.1 Cell Treatments were performed as shown in the table below.

Sample Treatment Time (mins) 1 Control 30 2 GRN A 30 3 GRN B 30 4 GRN C 30 5 GRN D 30 6 GRN E 30 7 GRN F 30 8 GRN G 30 9 Paragranulin 30 10 PGRN 30 11 PGRN 30 12 PGRN 90 13 PGRN 90 14 GRN A 90 15 Control 90 16 GRN C 90

The table shows the duration and type of cell treatments carried out.

4.4 RNA Extraction and Homogenisation.

After incubation with the appropriate treatment, cells were harvested at the specific time points by removing the medium, washing the cells briefly with ice-cold DPBS (Sigma) and lysing by adding 1.050mls RLT buffer (Qiagen RNeasy Mini Kit (Cat No 74104)). The lysate from each plate was then removed to separate 1.5ml microfuge tubes. Two thirds of each sample was then used to prepare RNA, the remainder being stored at -80°C for later use.

RNA was prepared from each sample by first homogenising the cell lysates using spin columns (Qiagen Qiashredder kit Cat No 79654) and purifying using the RNA Qiagen

-51-

RNeasy Mini Kit (Cat No 74104).The quantities for each sample were measured using a Nanodrop 1000 Spectrophotometer and the quality of the RNA produced assessed by running the samples on an Agilent 2100 Bioanalyzer using Agilent RNA 6000 Nano Kit chips (Figure 4.1). The RIN number of each sample is a measure of the quality of the RNA and must be 7 or above to proceed to the next stage of the protocol.

After confirming that all samples had an RIN of 7 or above a minimum of 5µg of each RNA sample was spiked with control RNAs. This was done by preparing a 1:10 dilution of ERCC RNA Spike-In Mix 1 in nuclease-free water and adding 1µl of this to each sample.

Figure 4.1 Example of Nano Chip Readout from Agilent 2100 Bioanalyzer.

28S 18S

The figure shows samples 1 to 8 demonstrating the lack of degradation and the intact ribosomal RNA bands 28S and 18S. 4.5 Library Preparation for Next Generation Sequencing

To prepare sample libraries for Ion Torrent analysis RNA was digested and ligated to linkers and cDNA prepared. Subsequent amplification by PCR produced sufficient material to run on the Ion Torrent Personal Genome Machine.

-52-

4.5.1 Ribosomal RNA Depletion

Ribosomal RNA (rRNA) accounts for around 80% of the RNA present in the cell. It is important for protein synthesis however it carries no biological information. In order to obtain relevant sequence data for the investigation of gene expression changes it is therefore necessary to remove it as 80% of the RNA Seq results would be uninformative and would overwhelm the potentially significant data.

Depletion of ribosomal RNA was carried out using the Ribominus™ Eukaryote Kit (Ambion A15026) which removes ribosomal RNA by hybridising it to specific probes which are then bound to nucleic acid beads removing rRNA and retaining the rRNA- depleted RNA in the supernatant.

Hybridisation of the Ribominus™ Eukaryote probe to the RNA was performed by adding the probe mix to the x2 hybridisation buffer and RNA and diluting to 100µl with nuclease- free water in a sterile RNase-free 1.5ml microfuge tube. The mixed samples were then placed in a PCR machine for 5mins at 70°C, the PCR machine switched off and the tubes allowed to cool in situ for 30 minutes to promote sequence–specific hybridisation.

During the cooling step the Ribominus™ Magnetic purification beads were prepared for use by vortexing to resuspend followed by the addition of 500µl of the bead suspension to separate 1.5ml sterile RNase-free microfuge tubes. The tubes were placed on a magnetic stand for one minute and the cleared supernatant removed by aspiration. The beads were washed twice by resuspending in 500µl nuclease–free water, placing on the magnet and aspirating the cleared supernatant. The beads were then resuspended in 200µl/tube x1 hybridisation buffer (made by diluting x2 with nuclease-free water) and placed in a 37°C heat block for a minimum of 5 minutes.

After the 30 minute cooling step the RNA/probe mixes were vortexed briefly, transferred to the tubes containing the washed magnetic beads and the solutions mixed well by pipetting up and down. The samples were heated at 37°C (heat block) for 5 minutes, centrifuged briefly and the solution cleared by placing on the magnetic stand for one minute. In order to clean and concentrate the rRNA depleted RNA the samples were subjected to a magnetic bead purification step as follows. The supernatants (~300µl) -53-

containing the rRNA-depleted RNA were transferred to fresh tubes. 400µl binding solution* (*8mls 100% ethanol added to bottle of binding concentrate provided) was added to 10µl Nucleic Acid binding beads and mixed by careful pipetting up and down. The 300µl of rRNA depleted RNA was then added and the samples mixed as before. 1ml of absolute ethanol was subsequently added to each sample and the tubes mixed well by inversion. After incubating at room temperature for 5 minutes the samples were briefly centrifuged and placed on the magnetic stand for 3 minutes until the solutions cleared. The solutions were then aspirated from each tube being careful not to disturb the beads. The tubes were removed from the stand and washed by dispensing 300µl of the prepared wash solution down the side of each tube, replacing on the stand and aspirating the cleared solution. The beads were allowed to dry at room temperature for 2 minutes. The purified RNA was then eluted from the beads by the addition of 12µl pre-heated (70°C) nuclease free water, incubation for 1 minute and replacing on the rack to allow removal of the supernatant containing the eluted RNA to a new 200µl sterile tube.

After performing this clean-up step to remove the probe with attached rRNA the resultant total RNA product was quantified by taking 1µl of the eluted RNA and measuring against known standards using the Qubit RNA Assay Kit and Qubit Fluorometer (Life Technologies).

4.5.2 RNA Fragmentation

The Ion Torrent Proton is only able to sequence RNA fragments of <400bps so in order to produce suitably sized fragments the RNA above must be treated with RNase followed by size exclusion and subsequent magnetic bead clean-up. The undigested RNA is removed by washing and the correctly sized fragments can then be eluted from the beads for preparation of the sample libraries.

Between 100 and 300ng of each sample RNA in a volume of 10µl was added to separate 0.2 ml PCR tubes (on ice) to which was added 1µl of 10x RNase III Reaction Buffer and 1µl RNase III (Ion Total RNA-Seq Kit v2 Life Technologies 4475936). After heating at 37°C for 3 minutes (PCR machine), the samples were immediately placed on ice and 20µl nuclease- free water added to each. The fragmented RNA was purified by binding to the nucleic

-54-

acid binding beads and concentrating. The beads were resuspended by gentle vortexing and 5µl added to fresh, sterile 1.5ml microfuge tubes. 90µl concentrated binding solution was added to each tube and mixed by pipetting up and down x10. 3µl of RNA digest was then added to the tubes and 150µl of 100% ethanol (using a tip which had been pre- wetted) added to each. The samples were then mixed by pipetting up and down x10 using a fresh tip for each. After 5 minutes incubation at room temperature the tubes were placed on the magnetic stand until the solution was clear. The supernatant was removed, without disturbing the beads, and discarded. The beads were washed on the stand by adding 150µl per tube of the wash solution* (*44mls 100% ethanol previously added to concentrate provided), incubating for 30 seconds and aspirating and discarding the wash solution. The beads were then allowed to air-dry on the stand for 1-2 minutes before removing from the stand and adding 12µl pre-warmed (37°C) nuclease-free water and mixing by pipetting up and down x10. After incubating at room temperature for 1 minute the tubes were replaced on the magnetic stand and the eluted RNA carefully pipetted to a fresh tube.

The yield of RNA for each sample was quantitated using the Qubit RNA Assay kit as previously. Fragment size was checked by diluting an aliquot of the sample to between 50 and 5000pg/µl with nuclease free water and running on an Agilent 2100 Bioanalyzer using Agilent Bioanalyser Picochips (Figure 4.2).

-55-

Figure 4.2 Pico Chip Readout.

1kbp 500bp 200bp

The figure shows that the RNA from samples 1 to 8 has been digested producing fragments of <500 base pairs. Note the lack of ribosomal bands seen in Figure 4.1 which have been removed during ribosomal depletion.

4.5.3 Library Preparation

Having established that the RNA fragments were in the right size range, they were then used to prepare the whole transcriptome libraries. The libraries are prepared by ligating two oligonucleotides to the RNA, the A adaptor to the 5’end and the P1 adaptor to the 3’end. This allows the first strand of cDNA to be synthesised using the P1 adaptor as a priming site.

In order to ligate the primers onto each end of the RNA hybridisation reactions were carried out by adding 5µl of a hybridisation master mix containing 2µl Ion Adaptor mix v2 and 3µl hybridisation solution per sample to 3µl of each RNA (>/= 100ng) in a 96 well sequencing plate. Each sample was pipetted up and down x10 to mix and the plate sealed and placed on a thermal cycler on at 65°C for 10 minutes followed by 5 minutes at 30°C. The plate was then immediately placed directly on ice and ligation reagents added as follows. To each hybridisation reaction 10µl of x2 Ligation Buffer and 2µl Ligation Enzyme Mix were added giving a final volume of 20µl. The contents were mixed by pipetting up and down x5 and centrifuged briefly to collect the contents in the bottom of the wells.

-56-

The plate was then incubated in a thermal cycler (heated lid off) for 30 minutes if the input depleted rRNA was less than 100ng or 60 minutes if it was greater than 100ng.

The Reverse Transcriptase (RT) master mix is comprised of the following:-Nuclease free water 2µl, X10 RT Buffer 4µl, 2.5mM dNTP mix 2µl and Ion RT Primerv2 8µl giving a total of 16µl per reaction.16µl of the RT master mix was added to each ligated RNA sample, gently vortexed and centrifuged followed by 10mins at 70°C on a thermal cycler (with heated lid) and snap cooling on ice.4µl 10x SuperScript III Enzyme mix was then added to each sample, vortexed and centrifuged as before incubating in a thermal cycler with a heated lid at 42°C for 30 minutes. Magnetic bead clean-up of the reactions to remove excess primers and reagents was carried out as previously described (4.5.2) except that 120µl of binding solution was used for bead resuspension. 60µl nuclease free water was added to each 40µl RT reaction and the whole 100µl transferred to a tube containing the prepared beads. 125µl per tube of 100% ethanol was added using a pre-wet tip as previously described and the resulting suspension mixed by pipetting up and down x10. The incubation, wash and elution steps were performed as previously described.

4.5.4 Library Amplification

As each library has undergone several size exclusion and purification steps the amount of material remaining is considerably diminished. It is therefore necessary to amplify the cDNA to produce a sufficient quantity for sequencing.

6µl of each cDNA sample was transferred to a 96 well sequencing plate and 45µl Platinum PCR Supermix High Fidelity master mix , 1µl Ion 5’ PCR primer V2 and 1µl Ion 3’ PCR primer V2 were added to each well. The plate was sealed, mixed and centrifuged and run in a thermal cycler on the following programme:

-57-

94°C 2 minutes

Followed by two cycles of

94°C 30 seconds ; 50°C 30 seconds ; 68°C 30 seconds.

Followed by 14 cycles if greater than 100ng RNA template or 16 cycles if less

94°C 30 seconds ; 62°C 30 seconds ; 68°C 30 seconds.

With a final hold of 68°C for 5 minutes.

To purify the amplified cDNA the beads were prepared by adding 180µl binding buffer to 5µl bead suspension and resuspending as previously described. The 53µl PCR reactions were then added to the bead suspensions and bound, washed and eluted as before except the volume of ethanol used was 130µl and the elution volume of nuclease free water 15µl. cDNA library concentrations were determined using the Qubit DNA assay kit and Qubit Fluorometer and DNA size analysed using the high sensitivity DNA chips followed by smear analysis on the Agilent Bioanalyzer 2100. This is to enable quantification of the percentage of DNA that is between 50 and 160bps and the molar concentrations of DNA between 50 and 1000bps to be calculated (Figure 4.3).

-58-

Figure 4.3 DNA Chip Readout.

The top panel shows the distribution of DNA fragment size per sample with most samples being below 300 bps. NB After construction of the DNA sequence libraries, sample 2 failed and was re-purified.

The bottom figure shows a graphical representation of fragment size for one sample. The peaks at 35 and 10380bp represent the upper and lower time markers for the ladder. The library peak (between 50 and 160bp) represents approximately 85% of the loaded sample.

4.5.5 Library Sequencing

The prepared libraries were sent to the Clinical Research Centre in Edinburgh where they were diluted to 100pM for template preparation.

Clonal libraries were made using the Ion PI Template OT2 200 Kit v3 (#4488318) and the sequencing reactions set up using the Ion PI Sequencing 200 Kit v3 (#4488315). These were then run on Ion PI Chip Kit v2 (#4482321) using the Ion Torrent Proton Personal Genome Machine (#2456290-0449). FastQ files were generated and all subsequent data and pathway analysis was performed by Dr Sara Rollinson. -59-

4.6 Next Generation Sequencing Data Analysis

Raw NGS data is a mixture of good and poor quality reads. In order to produce good quality data the poor reads needs to be identified so that they can be removed. This was achieved using a programme called Fastqc http://www.bioinformatics.babraham.ac.uk/projects/fastqc/.

The data was filtered using a Phred score of 17 i.e. the probability of the base being correctly read of 98%. Any sequences longer than 250 base pairs were discarded as Figure 4.4 shows a decline in the quality of the sequencing data after point. The first 5 base pairs of sequence were discarded to ensure full adapter trimming which may interfere with subsequent analysis. Reads shorter than 35 base pairs were also discarded as these were predominantly adapter and primer sequences.

Figure 4.4 Taken from NGS raw data Phred score versus read position.

The figure shows an example of the output from the Fastqc (displaying the quality score (Phred value) across all bases for library 1.

4.6.1 Alignment to the Reference Genome.

Having produced a set of reasonable sequences for each library it was then necessary to align them to a reference genome to identify what the sequences code for. The reference used was the human genome version hg19. By using specialist software it is possible to identify the chromosomal location of individual sequences.

-60-

All products from other next generation sequencing platforms (e.g. SOLiD ®, Roche) are paired i.e. the sequence is read from each end and alignment to the genome allows the missing section in the middle to be deduced. Also the products are of the order of 100bps in length therefore the data analysis tools are optimised accordingly. Ion Torrent data is not paired and the sequence is continuous. The products are also of differing lengths so need a different analytical approach. Life Technologies recommend a two-step alignment process for speed and sensitivity followed by merging of the two files. In this instance the first alignment was performed using a programme called STAR [103]. The resulting output consists of mapped and unmapped files. The unmapped files are then converted to FastQ and run on Bowtie 2, which is a slower, more sensitive programme and will align smaller pieces of DNA than the faster, less sensitive STAR. The two files produced by Star and Bowtie 2 were then merged and sorted to produce a list of sequences according to their chromosomal location i.e. chromosome 1 base pairs 1 to 1000, chromosome 1 base pairs 1450 to 2300, chromosome 2 base pairs 1 to…… etc. This was done using Picard Java tools (http://biowolf.nih.gov/apps/picard.html). The resultant list of sequences was then compared to a file containing the known locations of each, exon or feature mapped to the genome. Thus it was possible to obtain a list of the number of sequences (counts) for each exon location. A gene count was produced by summing the exon counts. This list of counts was produced using HTSeq [104].

Once a count list has been prepared, analysis of differential gene expression can be carried out. The first step is to normalise the libraries as the range of the number of reads varied considerably (see Table 4.2).

-61-

Table 4.2 Overview of sequencing statistics. The ERRC Spike in control counts show between 60 and 70 transcripts detected out of a possible 92.

Sample ID Total Mean Read ERCC % of ERCC analysis ERCC Number of length (bp) Total Reads plug-in. * number + Reads 1 89,052,778 103 1.28 0.92 66 2 83,009,226 101 1.09 0.89 61 3 86,901,123 122 1.57 0.94 63 4 40,785,227 114 1.07 0.91 60 5 70,707,455 107 1.01 0.87 65 6 87,617,557 114 1.15 0.92 64 7 87,768,540 125 1.02 0.86 65 8 88,769,177 115 0.99 0.89 65 9 77,518,453 82 2.45 0.94 69 10 77,751,774 89 1.78 0.92 70 11 77,745,396 95 1.79 0.91 66 12 72,271,131 94 1.66 0.88 63 13 77,448,177 105 1.17 0.91 64 14 74,210,693 101 1.80 0.92 65 15 80,335,627 84 1.92 0.92 69 16 80,283,369 83 1.87 0.92 67 The table above shows the total number of sequence reads for each library (sample) before quality control or alignment. Between 10 and 21% of reads were lost during the QC step and a further 59 to 82% after the sequence alignment steps.

*R2 value should be >0.8.

+ Number of ERCC transcripts detected out of 92 added pre-library construction.

Normalisation and subsequent gene count comparisons were carried out using DESeq [105].

4.7 Validation by Real Time PCR

4.7.1 Introduction

Real Time PCR is a method used to investigate single gene expression one gene at a time. It is a PCR-based method which allows data to be analysed as the reaction proceeds by utilising fluorescent dyes which combine with the double stranded DNA allowing the amount of product to be measured at each cycle.

-62-

4.7.2 RNA Extraction and cDNA Synthesis Untreated, undifferentiated SH-SY5Y cells were used to prepare standard cDNA curves. SH-SY5Y cells were grown in T75 flasks as previously described. Cells were washed with PBS and lysed in the flask by the addition of 1ml Trizol and repeated pipetting up and down. The lysed cell solution was then transferred to a 1.5ml Eppendorf tube and stored at -80°C until extraction. The Trizol lysed cells were thawed and 0.2mls chloroform added. This was shaken vigorously for 2 minutes and then allowed to sit at room temperature for 2-3 minutes before centrifuging at a speed of 14,000rpm and 4°C for 15 minutes. The upper aqueous phase was then carefully transferred to a fresh tube and the RNA precipitated by the addition of 0.5mls isopropanol. After 10 minutes at room temperature the solution was spun at 12,000g and 4°C for 10mins. The supernatant was carefully removed by pipette and the pellet washed by the addition of 1ml 75% ethanol. This was spun at 14,000rpm for 5 minutes at 40°C, the ethanol carefully removed and the pellet allowed to dry on ice for 5-30 minutes. The RNA pellet was dissolved in 20µl of RNase free water by flicking the tube and leaving on ice for 20-30 minutes. The concentration was measured using a Nanodrop 1000 Spectrophotometer.

RNA samples 1,5,10, and 11 (prepared as described in section 4.2) were thawed and cDNA prepared from the treated samples and SH-SY5Y cells as follows.

1.5µg of each RNA sample was made up to 12µl with RNase free water in separate 0.2ml PCR tubes. 1µl Random Hexamers was added to each tube and mixed by flicking. The tubes were then heated to 65°C for 5 minutes to denature the RNA using a thermal cycler and immediately cooled on ice for one minute (to quench the denaturation step) before adding 7µl per tube of reverse transciptase master mix made up as follows:-Per tube:- 4µl x5 Buffer, 1µl 10µM dNTPs, 1µl RNase Out and 1µl Transcriptor Reverse Transcriptase (All Roche) making a final volume of 20ul. After addition of the master mix, and mixing by flicking, the reaction tubes were replaced in the PCR machine, heated at 25°C for 10 minutes to anneal the random hexamers, followed by 60 minutes at 50°C for extension of the cDNA and 5minutes at 85°C for denaturation of any random transcripts. Samples were stored at -20°C until use.

-63-

4.7.3 Real Time PCR

The SH-SY5Y cDNA was used to set up a series of standard curves with concentrations of 100, 20, 1, 0.1 and 0.01 ng cDNA and a no template control each in quadruplicate wells in a volume of 3µl. A separate curve was used for each gene tested and GAPDH was used as a control. The samples from section 4.7.2 and earlier cDNA samples 13, 15 and 16 were diluted to 16ng/µl and 3µl added per well in quadruplicate. (For sample ID see table 4.1). Primers for each gene to be tested were designed by Dr Sara Rollinson and supplied as lyophilised stocks by Invitrogen. Sequences were as follows:-

PLXDC1_TM_F (5’-3’) GCTACTCCGACAACTCCACAGTT PLXDC1_TM_R (5’-3’) ACGTGGTCCCACTGAACCA MAP1B_TM_F (5’-3’) CACCGAGGTGCGCTTAATG MAP1B_TM_R (5’-3’) TCAGCACGAGCAGCTTGTGT EMX2_TM_F (5’-3’) GCTTCCAAGGGAACGACACTAG EMX_TM_R (5’-3’) GGGCCAGCGCGTTGT UBE2SP1_TM_F (5’-3’) CAGATCCCAGGAAACAGTGACA UBE2SP1_TM_R (5’-3’) CCCCACTCCTCTGGTGTAAAGT USP29_TM_F (5’-3’) GCTGCTCAGTCAGGTCACATG USP29_TM_R (5’-3’) CCCAAAAGGGCTCAATGGT SNX2_TM_F (5’-3’) ACTCCAATGGCCCAAAACC SNX2_TM_R (5’-3’) TGTGGCTTCTGCAAAAAGATCTT SERPINB6_TM_F (5’-3’) ATGGCCCAGATACTTTCTTTCAAT SERPINB6_TM_R (5’-3’) GGAAGCCCTGGTGGATGTC VPS72_TM_F (5’-3’) CGGCTCTGACAGTCGGAAGT VPS72_TM_R (5’-3’) AAGGAACGTTTGTCGTGTATG HFM1_TM_F (5’-3’) TCCTGCTCCATTGATTTCAGA HFM1_TM_R (5’-3’) TTTCCTGACCTAACAGTTTAT PID1_TM_F (5’-3’) GCTGATGAAGACAAGGACTCA PID1_TM_R (5’-3’) GGTGGAGACTTTGCCCAGGTA EPHA3_TM_F (5’-3’) AGGTTTTATGTGCCAAGCTTG EPHA3_TM_R (5’-3’) CTGTGAGGCGGGCACTTAG GRAP_TM_F (5’-3’) AGCCCCATCCGTGGTACTC GRAP_TM_R (5’-3’) TTCCGCTTCATCAGAATCTCT C2_TM_F (5’-3’) GCTGGCCCAGAAAGTAAAGAT C2_TM_R (5’-3’) CCAGATTGGCCTCCATCGT HOMER3_TM_F (5’-3’) TGGGCTTTGCCTCTGAACA HOMER3_TM_R (5’-3’) TTCCTTCACTTCCTGGAACTT GAPDH_F (5’-3’) CCTGTTCGACAGTCAGCCG GAPDH_R (5’-3’) CGACCAAATCCGTTGACTCC

Genes were chosen for validation based on the magnitude of the fold change resulting from cell treatment and their involvement in potentially interesting pathways.

-64-

The SH-SY5Y cDNA was used to generate a standard curve for each gene (Figure 4.5). Samples were then read off the curve to determine the amount of template thus ensuring that each sample is amplified within the optimal template concentration range for that assay. Thus both the reference and the gene of interest samples are amplified under the same conditions with equal efficiency making the results more reproducible when compared to the delta CT method [106].

Each primer was diluted to 100pmol/ul with RNase free water and 10pmol/µl stocks of each primer pair made by adding 10µl of forward 100pmol/µl stock to 10µl of reverse 100pmol/µl stock and making up to 100ul with RNase free water. A master mix for each primer pair was made by adding 5µl SYBR® green master mix (Life Technologies) to 1µl 10pmol/µl stock primer and 1µl RNase free water per well. The plate was run on a 7900HT Fast Real-Time PCR System (Applied Biosystems).

Figure 4.5 Example SHSY-5Y Standard Curve for EMX2.

Known dilutions of SH-SY5Y cDNA are amplified under the same conditions as the cDNA of the gene of interest allowing the amount of template to be determined from the graph.

Fourteen genes were chosen for validation of which one assay failed (PLXDC1), two failed validation (PID1 and SNX2) and eleven passed. This gives a validation rate of 78.6% (Figure 4.6).

-65-

Figure 4.6. Real Time PCR Validation of NGS Results.

Log2fold change above 1 indicates an up-regulation, below 1, a down-regulation of gene expression. The dark blue bars represent the NGS fold changes of expression levels whilst the light blue bars represent the Real Time PCR results.

GRAP, EPAH3, C2, HFM1, MAP1B and USP49 were all up-regulated in the NGS results and this was validated by Real time PCR. Conversely VPS72, SERPIN B6, UBE2SP1, HOMER 3 and EMX2 were down-regulated which again validated.

4.8 Gene Expression Results

Between 60 and 169 genes were affected by the individual 30 minute GRN treatments with 447 showing differential expression with PGRN treatment. The additional longer time point for GRNS A, C and PGRN showed fewer differentially expressed genes than the shorter time point samples. GRN A: N = 60 at 30 minutes versus N = 45 at 90 minutes, GRN C: N = 131 at 30 minutes and N = 92 at 90 minutes and PGRN: N = 447 at 30 minutes compared to N = 36 at 90 minutes (Table 4.3).

-66-

Table 4.3 Genes selected using the P=0.01 cut off by class.

RNA Micro Small Long Non- Treatment gene# RNA Nucleolar RNA Y RNA coding RNA Antisense Pseudogene Coding Total GRN B 30 34 4 7 0 3 3 5 46 102 GRN D 30 10 7 4 3 1 5 38 18 86 GRN E 30 11 3 2 4 4 2 33 22 81 GRN F 30 20 12 11 2 3 2 44 75 169 GRN G 30 7 7 2 0 1 1 18 44 80 ParaGRN 30 12 14 5 3 3 3 53 68 161 GRN A 30 22 1 4 2 4 0 13 14 60 GRN A 90 12 0 6 0 1 1 8 17 45 GRN C 30 25 8 6 1 2 0 14 75 131 GRN C 90 28 0 2 2 3 1 24 32 92 PGRN 30 43 13 7 5 18 10 56 295 447 PGRN 90 7 0 1 1 2 0 16 9 36 #RNA Genes include rRNA, tRNA, mtRNA. The figure shows the number of RNAs found listed by class for each cell treatment. The RNA genes were universally expressed in all treatments indicating that their expression is probably non-specific. 90 minute treatment times appear to reduce the number of genes differentially expressed particularly in the PGRN treated samples.

Despite having undergone rRNA depletion, ribosomal RNA genes were to be found in all treatments. In addition, transfer and mitochondrial RNAs were also found to be universally differentially expressed. As these are implicit in the transcription and translation processes they are probably non-specific for the cell treatments .Pseudogenes are genes often containing a number of mutations which destroy their ability to code protein thus rendering them functionally redundant. Pseudogenes can sometimes affect mRNA levels by interacting with regulatory factors and producing a product [107]. Many other RNA species were also differentially expressed. Micro RNAs are 22 nucleotide pieces of RNA that are able to bind to complementary strands of mRNA and silence the transcript by causing it to degrade. Small nucleolar RNAs chemically modify other RNA species e.g. small nuclear RNA, rRNA and tRNA, and can also function as micro RNA. Y RNAs interact with chromatin and are essential for DNA replication. Long intergenic non- coding or Linc RNAs are, as their name suggests, long (>200 nucleotides) and are multifactorial. They have been associated with many functions including targeting transcription factors and epigenetic regulation [108]. Antisense RNAs are transcripts that

-67-

may affect splicing or regulation of genes although their exact function is currently unknown.

While the GRN peptides have a conserved structure there is little if any information on the function of individual GRN peptides (other than the pro-inflammatory effects of GRNs A and B when compared the anti-inflammatory effect of PGRN [109]). As such, principle components analysis was carried out using the significant gene lists, enabling the samples to be grouped by similarity of expression pattern, for example replicates should cluster together. Principle components or Eigen values were generated using cluster version 2.11 [110] and plotted using excel (Dr Sara Rollinson) (Figure 4.7).

Figure 4.7 Principal Components analysis of NGS data from GRN and PGRN treatments

GRNF

GRNG

PC 2 PC GRNE GRNC GRND GRNB

PRGN30 PARA GRNA PC 1

. PCA analysis generated using Cluster.

A marked similarity between PGRN and paragranulin possibly indicates shared functionality. Similarly GRNs B, D, E and G cluster together. GRNs A, C and F appear to non-related to any of the other GRNs or to each other.

In order to try and understand the functions of the individual GRNs and PGRN, data analysis was carried out using specialist pathway analysis software: The Database for Annotation, Visualization and Integrated Discovery (DAVID) which enables a specific list of

-68-

genes to be compared to the whole human genome thereby identifying specific pathways that show enrichment in gene number by treatment (Table 4.4).

Table 4.4 Pathway analysis carried out using DAVID for the GRN and PGRN treatments.

-69-

Treatment Term Count Benjamini P- P - value Value* GRN B 30 Ribosome 4 1.60E-02 9.50E-04 Systemic lupus 4 1.20E-02 1.40E-03 erythematosus

GRN D 30 Systemic lupus 2 5.40E-01 9.40E-02 erythematosus

GRN E 30 Ribosome 4 2.50E-03 2.50E-04

GRN F 30 Ribosome 5 2.80E-02 7.70E-04 Glycine, serine and 3 1.70E-01 9.90E-03 threonine metabolism Systemic lupus 4 1.40E-01 1.20E-02 erythematosus

GRN G 30 Glycine, serine and 3 5.20E-02 2.70E-03 threonine metabolism

ParaGRN 30 Spliceosome 4 6.20E-01 2.10E-02

GRN A 30 Ribosome 3 1.20E-02 1.70E-03 GRN A 90 Ribosome 2 3.40E-01 6.70E-02

GRN C 30 Ribosome 4 1.60E-02 9.50E-04 Systemic lupus 4 1.20E-02 1.40E-03 erythematosus GRN C 90 Ribosome 7 4.30E-06 2.20E-07 Systemic lupus 4 3.70E-02 4.00E-03 erythematosus

PGRN 30 Ribosome 10 1.40E-02 1.20E-04 Systemic lupus 8 3.20E-01 6.40E-03 erythematosus Cysteine and methionine 5 2.30E-01 6.60E-03 metabolism Spliceosome 9 1.90E-01 7.00E-03 Pyruvate metabolism 5 2.40E-01 1.20E-02 Proteasome 5 3.30E-01 2.00E-02 Glycolysis / Gluconeogenesis 5 5.40E-01 4.50E-02

PGRN 90 N/A *Benjamini P-Value corrected as per the false discovery rate. The PGRN 90 treatment showed no pathway enrichment due to the low number of genes differentially expressed using a P-value cut-off of 0.01.

-70-

The gene lists were also visually examined, proteins being linked using the gene ontology biological process and molecular function fields, to examine individual proteins whose pathways did not reach significance in analysis due to the lack of sample replicates (Table 4.5).

Table 4.5.Differential Gene expression resulting from Cell Treatments.

GRN Name hgnc_ log2Fold Description symbol Change GRN A Sprouty-related, EVH1 domain SPRED3 -3.16 Regulates MAP kinase signalling containing 3 CD44 molecule CD44 -2.14 Receptor for hyaluronic acid (HA). Plexin domain containing 1 PLXDC1 2.58 Possible role in endothelial cell capillary morphogenesis NADH dehydrogenase (ubiquinone) 1 NDUFA4 2.19 Electron transport chain of alpha sub complex, 4 mitochondria. GRN B VGF nerve growth factor VGF -2.98 Hormone expressed in neuroendocrine cells Zinc finger protein 579 ZNF579 -2.87 Transcriptional regulation NADH dehydrogenase (ubiquinone) 1 NDUFA4 2.79 Electron transport chain of alpha sub complex, 4, mitochondria. Plexin domain containing 1 PLXDC1 2.53 Possible role in endothelial cell capillary morphogenesis GRN C Family with sequence similarity 129, FAM129 -1.75 Regulates phosphorylation of member A proteins involved in translation regulation Dynein, axonemal, heavy chain 17 DNAH17 -1.44 Microtubule associated motor protein complex Histone cluster 2, H2aa4 HIST2H2 2.26 Histone component of the AA4 nucleosome Aquaporin 1 AQP1 2.06 Membrane protein, water channel protein GRN D Empty spiracles homeobox 2 EMX2 -5.79 Transcription factor Proprotein convertase subtilisin/kexin PCSK1N -1.74 May function in the control of the type 1 inhibitor neuroendocrine secretory pathway Plexin domain containing 1 PLXDC1 1.96 Possible role in endothelial cell capillary morphogenesis HFM1, ATP-dependent DNA helicase HFM1 1.89 DNA helicase homolog, associated homolog with spastic hemiplegia

-71-

GRN E Zinc finger protein 579 ZNF579 -2.11 Transcriptional regulation VGF nerve growth factor VGF -1.90 Possible regulation of homeostasis ,metabolism and synaptic plasticity Adenosine A3 receptor ADORA3 2.97 G coupled protein receptor Oocyte expressed protein OOEP 2.77 Progression beyond first embryonic cell division GRN F Empty spiracles homeobox 2 EMX2 -8.02 Transcription factor Synaptic vesicle glycoprotein 2B SV2B -3.72 Histone cluster 2, H4b HIST2H4 3.06 Histone component of the B nucleosome Histone cluster 2, H2aa4 HIST2H2 2.57 Histone component of the AA4 nucleosome GRN G Empty spiracles homeobox 2 EMX2 -5.85 Transcription factor Inhibin, beta E INHBE -3.01 Pancreatic exocrine cell growth and proliferation Solute carrier organic anion SLCO1C1 3.25 Solute carrier organic anion transporter family, member 1C1 Solute carrier organic anion SLCO5A1 2.77 Solute carrier organic anion transporter family, member 5A1 Paragranulin Synaptotagmin VI SYT6 -4.19 Calcium dependent exocytosis of secretory vesicles Empty spiracles homeobox 2 EMX2 -2.33 Transcription factor Bone morphogenetic protein 5 BMP5 3.37 Bone morphogenic, may act as a signalling molecule Solute carrier organic anion SLCO5A1 1.43 Solute carrier organic anion transporter family, member 5A1 PGRN Empty spiracles homeobox 2 EMX2 -1.91 Transcription factor CDC28 protein kinase regulatory CKS1B -1.60 Regulation of cyclin-dependent subunit 1B kinases Protein tyrosine phosphatase, receptor PPFIA2 3.77 May regulate the disassembly for type, f polypeptide (PTPRF), interacting focal adhesions protein (liprin), alpha 2 ArfGAP with coiled-coil, ankyrin repeat ACAP1 3.46 Regulation of FRF GTPase activity, and PH domains 1 recycling endosome membrane

The table shows the highest log2fold change up and down- regulated genes for each granulin and PGRN at the 30 minute time point and the description of their activities.

-72-

GRNs A, B, and D all show up-regulation of PLXDC1 with log2 fold changes of 2.58, 2.53 and 1.96 respectively. The transcription factor EMX2 is down regulated in GRN D (-5.79 log2 fold change), GRN F (– 8.02 log2fold change), GRN G (-5.85 log2 fold change), Paragranulin (-2.33 log2 fold change) and PGRN (-1.91 log2 fold change). Histone up- regulation is common to GRN C (2.26 log2 fold change) and GRN F 3.06 and 2.57 log2 fold change) and solute carrier organic ion transporters were shown to be up-regulated in GRN G (3.25 and 2.77 log2 fold change) and Paragranulin (1.43 log2 fold change). Although there are similarities in the differential gene expression for most of the granulins e.g. non-expressed rRNA pseudogenes and transcription factors, there is insufficient data to identify potential pathways without further validation.

-73-

Chapter 5 Discussion

-74-

PGRN is a secreted glycoprotein which is widely expressed in the human body and has been shown to have several distinct functions. It is known to act as a neurotrophic growth factor [74] and is implicated in a variety of biological processes including angiogenesis, tumorigenesis, inflammation and wound healing. It has also recently been discovered to be involved in obesity and insulin resistance indicating a metabolic function [78].

Granulins A, B, C and D have been purified from human inflammatory cells [72] and F from urine [111] proving that they exist as physiological species. However there is no evidence for the separate existence of granulins E and G or paragranulin. Granulins A, B, C, D, E, F and G have been expressed using recombinant DNA in Escherichia coli (E.coli) and structural studies undertaken [112]. However, because E.coli is a prokaryote it does not always have the ability to produce eukaryotic post-translational modifications such as O and N glycosylation and phosphorylation. Protein is also likely to be expressed in insoluble aggregates called inclusion bodies requiring harsh denaturation and chemical re-folding to produce eukaryotic protein which is active and correctly folded. Therefore in order to replicate in vivo conditions as closely as possible, we felt that it was important to use a mammalian system of expression. The main drawback of using a mammalian system is the fact that expression levels are usually quite low but as high levels would also affect other cellular processes this could be considered an advantage. The initial constructs were designed with a His tag which would enable purification using standard Ni-NTA columns however we were unable to express detectable amounts of the individual granulins to attempt to purify. This was thought to be due to instability of the expressed peptides owing to their small size, (Personal communication from Dr Edward McKenzie).

In order to overcome this we re-cloned the granulins using the original constructs as templates and inserted an EcoR1 restriction site. This enabled us to sub-clone the granulins into the pHTN Halotag vector increasing the size of the expressed tagged peptides to 40kDa which improved protein stability and protected them from degradation. Unfortunately we then encountered a further problem when, on sequencing, it was discovered that some of the halo-tagged granulins had incorporated

-75-

some extra sequence. On consideration this may have been due to the fact that the original pCR2.1 plasmid contains two EcoR1 sites and the primers could have bound to either. To remove the possibility of this occurring again another site in the multiple cloning site of the pHTN vector was used that was not present in the pCR2.1 vector. New primers were designed containing the Pvu1 restriction site thus removing the extra sequence and allowing the granulins to be re-cloned with the correct amino acid composition.

Having finally constructed the DNA plasmid vectors and successfully transfected them into cells we found that it was necessary to produce a stable cell line for each in order to provide enough material for purification. Nevertheless it was still only possible to produce µg quantities of each granulin. This was enough for cell treatments but for antibody production, mg amounts were required. Initial experiments using a Pichia pastoris yeast expression system which also generates post-translationally modified proteins, failed to produce the quantities needed so it was decided to use synthetic peptides. As the granulins are very similar in structure it was only possible to make four antibodies.

Determining the concentration of the peptides also proved problematic as the standard measurement techniques; Bradford, which uses dye binding to quantify the protein by spectrophotometric readout, is dependent upon amino acid composition and BCA which relies on the reduction of copper ions by peptide bonds, both required a significant amount of purified protein and lacked accuracy for small peptides. The best estimation was obtained using the Qubit protein assay kit which uses a fluorescent dye that interchelates with peptides irrespective of amino acid composition and produces an absorbance reading that can then be quantified. However, when running what appeared to be the same amount of protein on a gel there was still a good deal of variation in the intensity of the silver stained bands. This is perhaps unsurprising as silver staining, although a very sensitive means of protein detection, can produce variable results depending upon the conditions used [113].

Cell treatments were therefore undertaken using the Qubit determined protein concentrations. -76-

RNA Seq Discussion

The RNA Seq experiment was designed with the aim of identifying differences in gene expression levels between progranulin treated, granulin treated and untreated differentiated neural cells. Ideally we would have been able to use biological replicates to increase the accuracy of analysis and interpretation. However, due to financial constraints, we were only able to perform single sample treatments at a single time point for most of the granulins. The lack of sample replicates meant that small differences in gene expression measurements from sample handling, library construction and analysis may have produced false positive results. Unfortunately the preparation of the samples for RNA Seq was not optimal as the protocol indicates that the libraries should contain more long reads than our experiment produced. This result could have been improved by performing the rRNA depletion twice and RNase treating for a shorter time period. (Personal communication Kelly Warrington –Life Technologies Field Technical specialist). It is possible that some of the granulins do not exist as separate entities in vivo and any variation in expression levels could simply be attributed to the addition of a synthetic peptide to the cells. In addition, the 90 minute time points showed a reduction in the number of genes differentially expressed which could be due to the fact that the progranulin/granulin treatments produced a transient response which had decayed by the longer time point. Interpretation of the results should therefore be approached with caution and any differential expression changes validated before any meaningful conclusions can be drawn.

All treatments, with the exceptions of paragranulin and GRN G were significantly enriched for ribosomal pathway genes. This could be due to the fact that although ribosomal depletion was carried out it was not of sufficient stringency as several ribosomal proteins were up-regulated in all treatments (>1.5log2 fold change). Two rounds of depletion may have helped to remove these transcripts (Personal communication Kelly Warrington Life Technologies). However, the fact that ribosomal genes were not similarly enriched in both the paragranulin and Grn G treatments may indicate that this could constitute a genuine effect on expression levels.

-77-

Most treatments also showed significant enrichment in genes involved in the Systemic lupus erythematosus (SLE) pathway, (maximum P = 0.0014). The majority of these were histones which form the core of the nucleosome. Histones can be post-translationally modified and depending on the modification, can affect DNA replication, transcription, alternate splicing, DNA repair and the relaxation or condensation of chromatin. The down-regulation of HIST2H3A (-1.52 log2fold) by progranulin and the up-regulation by the majority of the granulins e.g. GRN G (1.82 log2fold) may be linked to cell growth as histones have been shown to affect proliferation and tumour formation in a number of cancers [71, 78]. GRN peptides F and G and PGRN all showed enrichment in genes involved in amino acid metabolism. In the case of the granulins, glycine, serine and threonine (P=0.09) whilst PGRN treatment was found to increase the differential expression of cysteine and methionine metabolism genes, (P=0.006). Cysteine, histidine and glycine have been shown to significantly reduce NF-κB activation and the degradation of its inhibitor IκBα in cardiac endothelial cells stimulated with TNFα. E-selectin and IL-6 expression are also inhibited by all three amino acids. This could indicate that PGRN’s anti – inflammatory effect and link to metabolic function via increased insulin levels (as mentioned previously) may be, at least in part, a consequence of the modulation of amino acid levels [78, 114].

PGRN serum levels have been shown to be elevated in SLE and are thought to enhance Toll-like receptor 9 signalling which is involved in the innate immune response [115]. SLE is an autoimmune disease in which the immune system attacks normal bodily components producing inflammation and tissue damage. A recent paper also suggests a strong link between inflammation signalling and autoimmune disease in relation to PGRN mutation carriers and patients suffering from semantic variant primary progressive aphasia (svPPA) [116].

Both paragranulin and PGRN were both found to be enriched for genes involved in the spliceosome (maximum P =0.007). The spliceosome, as its name suggests, removes introns from transcribed pre-mRNA, this process is known as “splicing” and forms an integral part of the maturation of mRNA. Nine genes were found to be down-regulated by PGRN at the 30 minute time point. These included the common component the

-78-

heterogeneous nuclear riboprotein C (-0.88 log2 fold), and the CWC15 spliceosome associated protein homology (S.cerevisiae) (-0.82 log2 fold). It should be noted that most of the treatments showed some down-regulation of the spliceosome which may indicate a common effect of PGRN and its peptides.

Five genes associated with the ubiquitin proteosome system (UPS) pathway were found to be down-regulated after PGRN treatment (P=0.02). These were PSMC3, PSMD4, PSMA4, PSMA5, and PSMB6 proteasome components (-0.77 to -0.96 log2 fold change). The proteosome is a protein complex found in the nucleus and cytoplasm of eukaryotic cells. It is responsible for the regulation of cellular protein concentrations and the degradation of mis-folded proteins and transcription factors. One of the hallmarks of neurodegenerative disease is the accumulation of protein aggregates in the brain e.g. tau and β-amyloid in AD and TDP-43 in FTLD. In this study PGRN treatment appears to down- regulate genes which are part of the UPS pathway responsible for the degradation of aberrant protein. If PGRN protein deficiency increases aggregation of TDP-43 and this is responsible for destruction of neuronal tissue then one would expect PGRN treatment to up-regulate these genes. There is currently a great deal of controversy as to whether aggregation of TDP-43 is toxic, neuroprotective or simply associated with other disease- causing proteins [117]. The fact that PGRN down-regulates proteasomal genes indicates that the formation and/or mis-localisation of TDP-43 as a disease-causing mechanism is too simplistic an explanation and it may be the relationship between the UPS and autophagic removal of TDP-43 that is important [117].

Down-regulation of pyruvate metabolism (P = 0.01) and glycolysis/gluconeogenesis (P=0.04) pathways were also caused by PGRN treatment which may be of functional significance in metabolic disease [118]. It is tempting to suggest that PGRN treatment may be decreasing cellular activity possibly being anti-inflammatory by reducing the reactive oxygen species produced as a by-product of metabolism. Meanwhile the cells respond to inflammation by means of signalling pathways such as mTOR and Wnt [89, 119].

GRN A treatment resulted in the down-regulation of Sprouty-related EVH1 domain containing 3 gene (-3.16 log 2 fold change), which is thought to negatively regulate -79-

mitogen-activated protein kinase (MAPK) signalling by inhibiting the activation of the Ras/Erk pathway [120], and the proteoglycan versican (-1.11 log 2 fold change) which is up-regulated in tumour growth and involved in cell adhesion and proliferation [121].

Both VGF and chromogranin were down-regulated by GRN B treatment (-2.98 and 2.07 log 2 fold change respectively). They are both thought be involved in stress response and have been associated with synaptic degeneration in Alzheimer's disease [122]. Several genes known to regulate AKT/MAP kinase pathways are also down-regulated by GRN B e.g. insulin-like growth factor binding protein 5 (-2.12 log 2 fold change), ring finger protein 187 (-1.89 log 2 fold change) and major vault protein (-2.24 log 2 fold change) indicating a possible cell signalling involvement.

GRN C treatment down-regulates several proteins involved in folate metabolism including methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 2 (-0.87 log 2 fold change) and serine hydroxymethyltransferase 2 (mitochondria) (-1.23 log 2 fold change). Folate deficiency has been linked to an increase in the inflammatory response and has been associated, in macrophages, with the increased expression of IL1β, IL6 and TNFα [123] suggesting GRN C may be responding indirectly to inflammation.

GRN D and E treatments were most notable for their predominance of RNA genes in particular non-expressed rRNA pseduogenes. GRN F treatment produced a down- regulation in genes affecting transcriptional regulation either as transcription factors, activators or repressors. For example, empty spiracles homeobox 2 (-8.02 log2 fold change) which is also down-regulated by GRNS D, E, G paragranulin and PGRN. Conversely C1D nuclear receptor co-repressor was up-regulated (1.91 log2 fold change).

Additionally five solute carrier genes, which are involved in membrane transport were found to be differentially expressed including two; family 1 member 4 (-1.96 log 2 fold change) and family 7 member 11 (-1.72 log 2 fold change) which are associated with glutamate transport.

There are a lot of similarities in the differential expression patterns of the granulin peptides which is probably to be expected given the fact that they share a conserved structure. The main commonalities appear to be related to transcriptional regulation, -80-

stress response and genes involved in the maintenance of the cytoskeleton, actin and collagen.

PGRN treatment resulted in the differential expression of a much greater number of genes of which 295 had known protein expression. Down-regulation of several genes involved in innate immune response were identified, for example, prothymosin alpha (- 1.28), plasminogen receptor (-0.83) and canopy FGF signalling regulator 4 (-0.93)), myeloid differentiation primary response 88 (-0.90), plasminogen receptor (C terminal lysine transmembrane protein) (-0.83), prothymosin α (-1.28), and G patch domain and ankyrin repeats 1 (-0.92). The innate immune system provides an immediate defence against infection by the production of cytokines, activation of the complement cascade and activation of the adaptive (antibody) immune system. These results appear to confirm the anti-inflammatory effect of PGRN [124].

Other pathways such as membrane transport, endoplasmic reticulum, Golgi and endosomal sorting contained 16 down-regulated genes. These included charged multivesicular body proteins 1A and 5 (CHMP1A and CHMP5) part of the endosomal sorting complex required for transport (ESCRTlll) (-0.81 and -1.17 log2 fold change), developmentally regulated GTP binding protein 1 (-1.06), acyl-CoA dehydrogenase, very long chain (-0.77) and ER membrane protein complex subunit 3 (-1.35).The ESCRTlll pathway also includes the CHMP2B gene which, as has been previously mentioned, (1.5.4) has been linked to chromosome 3 related FTLD with mutations causing a C- terminal deletion in the protein. Interestingly CHMP1, 2 and 5 are all necessary for the correct assembly of the ESCRTlll complex [125]. The down-regulation of the CHMP genes suggests a role for the ESCRTIII pathway in PGRN mediated FTLD as haploinsufficiency would remove this control. Another notable gene involved in recycle of the endosome membrane, ArfGAP with coiled-coil, ankyrin repeat and PH domains 1 (ACAP1), was up- regulated by 3.46 log2 fold change. PGRN treatment also affected genes regulating splicing, most of which were down-regulated; CWC15 spliceosome-associated protein homolog (Saccharomyces cerevisiae) (-0.82), elongation factor Tu GTP binding domain containing 2 (-0.86). However, ubiquitin specific peptidase 49 which is involved in

-81-

epigenetic transcriptional activation and regulation of mRNA splicing was up-regulated 0.92 log2 fold.

Loss of function mutations of PGRN causing haploinsufficiency appears to activate certain cell signalling pathways. In particular the non-canonical Wingless type 5a (Wnt5a) is activated by PGRN deficiency and this effect can be modulated by the addition of exogenous PGRN [126]. Our data confirms that PGRN treatment affects multiple cell signalling pathways and could identify other significant pathway genes that are able to be regulated in this way. Having established that the effects of progranulin and the granulins are multifactorial and often acting in a completely opposing manner it is obvious that there is much work to be done to try and identify specific targets for putative treatment options. It is possible that the cleavage of PGRN into granulins or granulin complexes could be in itself a means of regulating the action of down-stream signalling or cellular levels of PGRN being regulated by some kind of feed-back loop dependant on the amount or specificity of cleavage. PGRN haploinsufficiency could also lead to fewer cleavage products i.e. granulins, thus the effects of granulin B on proteins associated with TGF-β signalling, AKT and MAPK pathways would also be affected. The differential gene expression produced by PGRN and granulin treatments has shown us that there are many potential targets for intervention in the treatment of FTLD-TDP resulting from PGRN haploinsufficiency however these will need further investigation before any meaningful conclusions can be drawn.

Possible Future Work.

Financial constraints meant that we were unable to study additional time points or replicates for treatment in the RNA Seq study. Lack of time also precluded testing of other larger granulin constructs (i.e. uncleaved fragments of PGRN) and antibody validation. Therefore there is scope to increase our understanding by further experimentation.

-82-

Novel transcripts could be identified by further interrogation of the RNA Seq data allowing us expand the genetic understanding of the effects of the granulin treatments. It will also be necessary to repeat the cell treatments and validate any findings by Real Time PCR. In addition, a time course with a greater spread of times could provide information as to the longevity of the effects of the treatments. This will enable further validation at the protein level to be undertaken by Western blotting of the treated cells with appropriate antibodies.

Having established down-stream effects of PGRN and granulin treatment of cells, Western blotting of brain tissue from patients with known PGRN mutations and control brains with no mutations would then provide a definitive link with disease. siRNA knockdown of the differentially expressed genes validated by Real Time PCR and Western blotting and consequent effects on PGRN levels could also be investigated.

As it is still uncertain as to whether all of the granulins exist as separate entities and may act in tandem, cell treatments using combinations of granulins or constructs containing more than one granulin motif for example CD,DE or CDE could provide further clarification as to the mode of action.

Localisation studies using the antibodies generated by our recombinant peptides will also require further study.

Additionally PGRN treatment of another neuronal cell type would provide information as to whether the affected genes were cell specific.

-83-

Supplementary information

Materials

Bacterial Cell Culture

SOC Medium (Invitrogen)

Luria Broth (LB) (Sigma)

Agar

Kanamycin

Ampicillin

Sequencing Reagents

Big Dye (Agilent)

Sequencing Buffer (x5) (Agilent)

10mM dNTP’s

Mammalian Cell Culture

Dulbecco’s Modified Eagle Medium (Sigma)

Heat-inactivated Foetal Bovine Serum (Gibco)

200mM L-Glutamine (Sigma)

Penicillin 10,000units/ml /Streptamycin 10mg/ml in 0.9% sodium chloride (Sigma)

G418 50mg/ml Promega

DNA Cloning jetPRIME transfection reagent (Polyplus) jetPRIME transfection buffer (Polyplus)

Buffers

TE buffer:- 10mM Tris pH7.5, 1mM EDTA

PBS:- Dulbecco’s Phosphate Buffered Saline without calcium or magnesium (Sigma)

TAE buffer:- 40mM Tris Acetate,1mM ethylenediaminetetraacetic acid -84-

TBST:- 50mM Tris, 150mM NaCl, 0.05% Tween 20 pH7.6

RIPA Buffer :- 50mM Tris pH7.5, 150mM NaCl, 1% Nonidet P40,0.5% sodium deoxycholate, one Complete Protease Inhibitor Cocktail Tablet (Roche)/50ml buffer

Protein Purification Reagents

Mammalian Lysis Buffer (Promega)

50x Protease Inhibitors (Promega)

Halotag® resin (Promega)

RQ1 DNase (Promega)

Gel Electrophoresis/Western Blotting Reagents

4-12% Bis Tris Novex NuPage gel (Life Technologies)

4 x LDS buffer (Life Technologies)

Novex Sharp pre-stained protein standard

Coomassie Blue R250 stain (BioRad)

Gel blotting paper (Whatman/GE GB005)

Protran BA83 0.2µm nitrocellulose blotting membrane (Whatman/GE)

Antibodies

Halotag® Rabbit Polyclonal Antibody (Promega)

HRP conjugated Goat anti Rabbit secondary antibody (Santa Cruz)

Real Time PCR

Transcriptor Reverse Transcriptase (Roche)

Transcriptor RT Reaction Buffer (5x) (Roche)

10mM dNTP’s (Roche)

Random Hexamer (Invitrogen)

RNase Out (Invitrogen)

-85-

References

1. Pick, A., Uber die Beziehungen der senilen Hirnantropie zur aphasi. Prager Medizinishe Wochenscrift, 1892. 17: p. 65-167. 2. Rademakers, R., M. Neumann, and I.R. Mackenzie, Advances in understanding the molecular basis of frontotemporal dementia. Nat Rev Neurol, 2012. 8(8): p. 423- 34. 3. Chow, T.W., et al., Inheritance of frontotemporal dementia. Arch Neurol, 1999. 56(7): p. 817-22. 4. Rizzu, P., et al., High prevalence of mutations in the microtubule-associated protein tau in a population study of frontotemporal dementia in the Netherlands. Am J Hum Genet, 1999. 64(2): p. 414-21. 5. Pickering-Brown, S.M., et al., Frequency and clinical characteristics of progranulin mutation carriers in the Manchester frontotemporal lobar degeneration cohort: comparison with patients with MAPT and no known mutations. Brain, 2008. 131(Pt 3): p. 721-31. 6. Gorno-Tempini, M.L., et al., Classification of primary progressive aphasia and its variants. Neurology, 2011. 76(11): p. 1006-14. 7. Mesulam, M.M., Primary progressive aphasia. Ann Neurol, 2001. 49(4): p. 425-32. 8. Neary, D., et al., Frontotemporal lobar degeneration: a consensus on clinical diagnostic criteria. Neurology, 1998. 51(6): p. 1546-54. 9. Rascovsky, K., et al., Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia. Brain, 2011. 134(Pt 9): p. 2456-77. 10. Wszolek, Z.K., et al., Frontotemporal dementia and parkinsonism linked to chromosome 17. Folia Neuropathol, 2005. 43(4): p. 258-70. 11. Sieben, A., et al., The genetics and neuropathology of frontotemporal lobar degeneration. Acta Neuropathol, 2012. 124(3): p. 353-72. 12. Witgert, M., et al., Frontal-lobe mediated behavioral dysfunction in amyotrophic lateral sclerosis. Eur J Neurol, 2010. 17(1): p. 103-10. 13. Ringholz, G.M., et al., Prevalence and patterns of cognitive impairment in sporadic ALS. Neurology, 2005. 65(4): p. 586-90. 14. Burrell, J.R., et al., Motor neuron dysfunction in frontotemporal dementia. Brain, 2011. 134(Pt 9): p. 2582-94. 15. Van Langenhove, T., J. van der Zee, and C. Van Broeckhoven, The molecular basis of the frontotemporal lobar degeneration-amyotrophic lateral sclerosis spectrum. Ann Med, 2012. 44(8): p. 817-28. 16. Kertesz, A. and P. McMonagle, Behavior and cognition in corticobasal degeneration and progressive supranuclear palsy. J Neurol Sci, 2010. 289(1-2): p. 138-43. 17. Kertesz, A. and D. Munoz, Relationship between frontotemporal dementia and corticobasal degeneration/progressive supranuclear palsy. Dement Geriatr Cogn Disord, 2004. 17(4): p. 282-6. 18. Grover, A., et al., 5' splice site mutations in tau associated with the inherited dementia FTDP-17 affect a stem-loop structure that regulates alternative splicing of exon 10. J Biol Chem, 1999. 274(21): p. 15134-43.

-86-

19. Halliday, G., et al., Mechanisms of disease in frontotemporal lobar degeneration: gain of function versus loss of function effects. Acta neuropathologica, 2012. 124(3): p. 373-82. 20. Alzheimer, A., Über eigenartige Krankheitsfälle des späteren Alters. Arch Psychiatr Nervenkr Z Gesamte Neurol Psychiatr 1911, 1911. 1911(4): p. 356-85.

21. Mackenzie, I.R., et al., Nomenclature and nosology for neuropathologic subtypes of frontotemporal lobar degeneration: an update. Acta Neuropathol, 2010. 119(1): p. 1-4. 22. Xu, Z.S., Does a loss of TDP-43 function cause neurodegeneration? Mol Neurodegener, 2012. 7: p. 27. 23. Mackenzie, I.R., et al., A harmonized classification system for FTLD-TDP pathology. Acta Neuropathol, 2011. 122(1): p. 111-3. 24. Ayala, Y.M., et al., Structural determinants of the cellular localization and shuttling of TDP-43. J Cell Sci, 2008. 121(Pt 22): p. 3778-85. 25. Neumann, M., et al., Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science, 2006. 314(5796): p. 130-3. 26. Urwin, H., et al., FUS pathology defines the majority of tau- and TDP-43-negative frontotemporal lobar degeneration. Acta Neuropathol, 2010. 120(1): p. 33-41. 27. Wang, J.H., et al., [Effects of adipose-derived stem cell transplantation on the angiogenesis and the expression of bFGF and VEGF in the brain post focal cerebral ischemia in rats]. Xi Bao Yu Fen Zi Mian Yi Xue Za Zhi, 2008. 24(10): p. 958-61. 28. Neumann, M., et al., FET proteins TAF15 and EWS are selective markers that distinguish FTLD with FUS pathology from amyotrophic lateral sclerosis with FUS mutations. Brain, 2011. 134(Pt 9): p. 2595-609. 29. Wilhelmsen, K.C., et al., Localization of disinhibition-dementia-parkinsonism- amyotrophy complex to 17q21-22. Am J Hum Genet, 1994. 55(6): p. 1159-65. 30. Foster, N.L., et al., Frontotemporal dementia and parkinsonism linked to chromosome 17: a consensus conference. Conference Participants. Ann Neurol, 1997. 41(6): p. 706-15. 31. van Swieten, J.C., et al., Phenotypic variation in hereditary frontotemporal dementia with tau mutations. Ann Neurol, 1999. 46(4): p. 617-26. 32. van Swieten, J. and M.G. Spillantini, Hereditary frontotemporal dementia caused by Tau gene mutations. Brain Pathol, 2007. 17(1): p. 63-73. 33. Seelaar, H., et al., Distinct genetic forms of frontotemporal dementia. Neurology, 2008. 71(16): p. 1220-6. 34. Gass, J., et al., Progranulin: an emerging target for FTLD therapies. Brain Res, 2012. 1462: p. 118-28. 35. Gijselinck, I., C. Van Broeckhoven, and M. Cruts, Granulin mutations associated with frontotemporal lobar degeneration and related disorders: an update. Human mutation, 2008. 29(12): p. 1373-86. 36. Finch, N., et al., Plasma progranulin levels predict progranulin mutation status in frontotemporal dementia patients and asymptomatic family members. Brain, 2009. 132(Pt 3): p. 583-91. 37. Baker, M., et al., Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature, 2006. 442(7105): p. 916-9.

-87-

38. Cruts, M., et al., Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Nature, 2006. 442(7105): p. 920-4. 39. Pickering-Brown, S. and M. Hutton, The genetics of frontotemporal dementia. Handb Clin Neurol, 2008. 89: p. 383-92. 40. Cannon, A., et al., Clinicopathologic variability of the GRN A9D mutation, including amyotrophic lateral sclerosis. Neurology, 2013. 80(19): p. 1771-7. 41. Banzhaf-Strathmann, J., et al., Promoter DNA methylation regulates progranulin expression and is altered in FTLD. Acta Neuropathol Commun, 2013. 1(1): p. 16. 42. Morita, M., et al., A locus on chromosome 9p confers susceptibility to ALS and frontotemporal dementia. Neurology, 2006. 66(6): p. 839-44. 43. Vance, C., et al., Familial amyotrophic lateral sclerosis with frontotemporal dementia is linked to a locus on chromosome 9p13.2-21.3. Brain, 2006. 129(Pt 4): p. 868-76. 44. Traynor, B.J., Road to the chromosome 9p-linked ALS/FTD locus. J Neurol Neurosurg Psychiatry, 2012. 83(4): p. 356-7. 45. Renton, A.E., et al., A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD. Neuron, 2011. 72(2): p. 257-68. 46. Levine, T.P., et al., The product of C9orf72, a gene strongly implicated in neurodegeneration, is structurally related to DENN Rab-GEFs. Bioinformatics, 2013. 29(4): p. 499-503. 47. DeJesus-Hernandez, M., et al., Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron, 2011. 72(2): p. 245-56. 48. Bigio, E.H., et al., Frontotemporal lobar degeneration with TDP-43 proteinopathy and chromosome 9p repeat expansion in C9ORF72: clinicopathologic correlation. Neuropathology, 2013. 33(2): p. 122-33. 49. Mori, K., et al., hnRNP A3 binds to GGGGCC repeats and is a constituent of p62- positive/TDP43-negative inclusions in the hippocampus of patients with C9orf72 mutations. Acta Neuropathol, 2013. 125(3): p. 413-23. 50. Zu, T., et al., Non-ATG-initiated translation directed by microsatellite expansions. Proc Natl Acad Sci U S A, 2011. 108(1): p. 260-5. 51. Mori, K., et al., The C9orf72 GGGGCC repeat is translated into aggregating dipeptide-repeat proteins in FTLD/ALS. Science, 2013. 339(6125): p. 1335-8. 52. Gijselinck, I., et al., A C9orf72 promoter repeat expansion in a Flanders-Belgian cohort with disorders of the frontotemporal lobar degeneration-amyotrophic lateral sclerosis spectrum: a gene identification study. Lancet Neurol, 2012. 11(1): p. 54-65. 53. Duyao, M., et al., Trinucleotide repeat length instability and age of onset in Huntington's disease. Nature genetics, 1993. 4(4): p. 387-92. 54. Harley, H.G., et al., Unstable DNA sequence in myotonic dystrophy. Lancet, 1992. 339(8802): p. 1125-8. 55. Cossee, M., et al., ARX polyalanine expansions are highly implicated in familial cases of mental retardation with infantile epilepsy and/or hand dystonia. American journal of medical genetics. Part A, 2011. 155A(1): p. 98-105.

-88-

56. Reddy, K., et al., The Disease-associated r(GGGGCC)n Repeat from the C9orf72 Gene Forms Tract Length-dependent Uni- and Multimolecular RNA G-quadruplex Structures. J Biol Chem, 2013. 288(14): p. 9860-6. 57. Fratta, P., et al., C9orf72 hexanucleotide repeat associated with amyotrophic lateral sclerosis and frontotemporal dementia forms RNA G-quadruplexes. Sci Rep, 2012. 2: p. 1016. 58. Skibinski, G., et al., Mutations in the endosomal ESCRTIII-complex subunit CHMP2B in frontotemporal dementia. Nat Genet, 2005. 37(8): p. 806-8. 59. van der Zee, J., et al., CHMP2B C-truncating mutations in frontotemporal lobar degeneration are associated with an aberrant endosomal phenotype in vitro. Hum Mol Genet, 2008. 17(2): p. 313-22. 60. Isaacs, A.M., et al., Frontotemporal dementia caused by CHMP2B mutations. Curr Alzheimer Res, 2011. 8(3): p. 246-51. 61. Watts, G.D., et al., Inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia is caused by mutant valosin-containing protein. Nat Genet, 2004. 36(4): p. 377-81. 62. Johnson, J.O., et al., Exome sequencing reveals VCP mutations as a cause of familial ALS. Neuron, 2010. 68(5): p. 857-64. 63. Cairns, N.J. and N. Ghoshal, FUS: A new actor on the frontotemporal lobar degeneration stage. Neurology, 2010. 74(5): p. 354-6. 64. Bhandari, V. and A. Bateman, Structure and chromosomal location of the human granulin gene. Biochem Biophys Res Commun, 1992. 188(1): p. 57-63. 65. Songsrirote, K., et al., Development and application of mass spectrometric methods for the analysis of progranulin N-glycosylation. J Proteomics, 2010. 73(8): p. 1479-90. 66. Zhu, J., et al., Conversion of proepithelin to epithelins: roles of SLPI and elastase in host defense and wound repair. Cell, 2002. 111(6): p. 867-78. 67. Kessenbrock, K., et al., Proteinase 3 and neutrophil elastase enhance inflammation in mice by inactivating antiinflammatory progranulin. J Clin Invest, 2008. 118(7): p. 2438-47. 68. Butler, G.S., et al., Pharmacoproteomics of a metalloproteinase hydroxamate inhibitor in breast cancer cells: dynamics of membrane type 1 matrix metalloproteinase-mediated membrane protein shedding. Mol Cell Biol, 2008. 28(15): p. 4896-914. 69. Bai, X.H., et al., ADAMTS-7, a direct target of PTHrP, adversely regulates endochondral bone growth by associating with and inactivating GEP growth factor. Mol Cell Biol, 2009. 29(15): p. 4201-19. 70. Suh, H.S., et al., Regulation of progranulin expression in human microglia and proteolysis of progranulin by matrix metalloproteinase-12 (MMP-12). PLoS One. 7(4): p. e35115. 71. Ghidoni, R., et al., Secretory leukocyte protease inhibitor protein regulates the penetrance of frontotemporal lobar degeneration in progranulin mutation carriers. J Alzheimers Dis, 2014. 38(3): p. 533-9. 72. Bateman, A. and H.P. Bennett, Granulins: the structure and function of an emerging family of growth factors. J Endocrinol, 1998. 158(2): p. 145-51.

-89-

73. Bhandari, V., A. Giaid, and A. Bateman, The complementary deoxyribonucleic acid sequence, tissue distribution, and cellular localization of the rat granulin precursor. Endocrinology, 1993. 133(6): p. 2682-9. 74. Ryan, C.L., et al., Progranulin is expressed within motor neurons and promotes neuronal cell survival. BMC Neurosci, 2009. 10: p. 130. 75. Van Damme, P., et al., Progranulin functions as a neurotrophic factor to regulate neurite outgrowth and enhance neuronal survival. J Cell Biol, 2008. 181(1): p. 37- 41. 76. Tapia, L., et al., Progranulin deficiency decreases gross neural connectivity but enhances transmission at individual synapses. J Neurosci, 2011. 31(31): p. 11126- 32. 77. Zheng, Y., et al., C-terminus of progranulin interacts with the beta-propeller region of sortilin to regulate progranulin trafficking. PLoS One, 2011. 6(6): p. e21023. 78. De Muynck, L., et al., The neurotrophic properties of progranulin depend on the granulin E domain but do not require sortilin binding. Neurobiol Aging, 2013. 34(11): p. 2541-7. 79. Hu, F., et al., Sortilin-mediated endocytosis determines levels of the frontotemporal dementia protein, progranulin. Neuron, 2010. 68(4): p. 654-67. 80. Carrasquillo, M.M., et al., Genome-wide screen identifies rs646776 near sortilin as a regulator of progranulin levels in human plasma. Am J Hum Genet, 2010. 87(6): p. 890-7. 81. Prudencio, M., et al., Misregulation of human sortilin splicing leads to the generation of a nonfunctional progranulin receptor. Proc Natl Acad Sci U S A, 2012. 109(52): p. 21510-5. 82. Tang, W., et al., The growth factor progranulin binds to TNF receptors and is therapeutic against inflammatory arthritis in mice. Science, 2011. 332(6028): p. 478-84. 83. Chen, X., et al., Progranulin Does Not Bind Tumor Necrosis Factor (TNF) Receptors and Is Not a Direct Regulator of TNF-Dependent Signaling or Bioactivity in Immune or Neuronal Cells. J Neurosci, 2013. 33(21): p. 9202-9213. 84. Kojima, Y., et al., Progranulin expression in advanced human atherosclerotic plaque. Atherosclerosis, 2009. 206(1): p. 102-8. 85. He, Z., et al., Progranulin is a mediator of the wound response. Nat Med, 2003. 9(2): p. 225-9. 86. He, Z., et al., Progranulin (PC-cell-derived growth factor/acrogranin) regulates invasion and cell survival. Cancer Res, 2002. 62(19): p. 5590-6. 87. Kamrava, M., et al., Lysophosphatidic acid and endothelin-induced proliferation of ovarian cancer cell lines is mitigated by neutralization of granulin-epithelin precursor (GEP), a prosurvival factor for ovarian cancer. Oncogene, 2005. 24(47): p. 7084-93. 88. Monami, G., et al., Proepithelin promotes migration and invasion of 5637 bladder cancer cells through the activation of ERK1/2 and the formation of a paxillin/FAK/ERK complex. Cancer Res, 2006. 66(14): p. 7103-10. 89. Rosen, E.Y., et al., Functional genomic analyses identify pathways dysregulated by progranulin deficiency, implicating Wnt signaling. Neuron, 2011. 71(6): p. 1030- 42. -90-

90. He, Z. and A. Bateman, Progranulin gene expression regulates epithelial cell growth and promotes tumor growth in vivo. Cancer Res, 1999. 59(13): p. 3222-9. 91. Monami, G., et al., Proepithelin regulates prostate cancer cell biology by promoting cell growth, migration, and anchorage-independent growth. Am J Pathol, 2009. 174(3): p. 1037-47. 92. Chen, X.Y., et al., Expression of PC cell-derived growth factor and vascular endothelial growth factor in esophageal squamous cell carcinoma and their clinicopathologic significance. Chin Med J (Engl), 2008. 121(10): p. 881-6. 93. Zhang, H. and G. Serrero, Inhibition of tumorigenicity of the teratoma PC cell line by transfection with antisense cDNA for PC cell-derived growth factor (PCDGF, epithelin/granulin precursor). Proc Natl Acad Sci U S A, 1998. 95(24): p. 14202-7. 94. Kollmann, K., et al., Cell biology and function of neuronal ceroid lipofuscinosis- related proteins. Biochim Biophys Acta, 2013. 1832(11): p. 1866-81. 95. Smith, K.R., et al., Strikingly different clinicopathological phenotypes determined by progranulin-mutation dosage. Am J Hum Genet, 2012. 90(6): p. 1102-7. 96. Canafoglia, L., et al., Recurrent generalized seizures, visual loss, and palinopsia as phenotypic features of neuronal ceroid lipofuscinosis due to progranulin gene mutation. Epilepsia, 2014. 97. Gotzl, J.K., et al., Common pathobiochemical hallmarks of progranulin-associated frontotemporal lobar degeneration and neuronal ceroid lipofuscinosis. Acta Neuropathol, 2014. 127(6): p. 845-60. 98. Colombrita, C., et al., TDP-43 and FUS RNA-binding proteins bind distinct sets of cytoplasmic messenger RNAs and differently regulate their post-transcriptional fate in motoneuron-like cells. J Biol Chem, 2012. 287(19): p. 15635-47. 99. Bateman, A., et al., Granulins, a novel class of peptide from leukocytes. Biochem Biophys Res Commun, 1990. 173(3): p. 1161-8. 100. Shoyab, M., et al., Epithelins 1 and 2: isolation and characterization of two cysteine-rich growth-modulating proteins. Proc Natl Acad Sci U S A, 1990. 87(20): p. 7912-6. 101. Belcourt, D.R., C. Lazure, and H.P. Bennett, Isolation and primary structure of the three major forms of granulin-like peptides from hematopoietic tissues of a teleost fish (Cyprinus carpio). J Biol Chem, 1993. 268(13): p. 9230-7. 102. Bhandari, V., R.G. Palfree, and A. Bateman, Isolation and sequence of the granulin precursor cDNA from human bone marrow reveals tandem cysteine-rich granulin domains. Proc Natl Acad Sci U S A, 1992. 89(5): p. 1715-9. 103. Dobin, A., et al., STAR: ultrafast universal RNA-seq aligner. Bioinformatics, 2013. 29(1): p. 15-21. 104. Simon Anders, P.T.P., Wolfgang Huber, HTSeq - A Python framework to work with high throughput sequencing data bioRxiv preprint, 2014. 105. Anders, S. and W. Huber, Differential expression analysis for sequence count data. Genome Biol, 2010. 11(10): p. R106. 106. Ruijter, J.M., et al., Evaluation of qPCR curve analysis methods for reliable biomarker discovery: bias, resolution, precision, and implications. Methods, 2013. 59(1): p. 32-46. 107. Poliseno, L., et al., A coding-independent function of gene and pseudogene mRNAs regulates tumour biology. Nature, 2010. 465(7301): p. 1033-8. -91-

108. Marin-Bejar, O., et al., Pint lincRNA connects the p53 pathway with epigenetic silencing by the Polycomb repressive complex 2. Genome biology, 2013. 14(9): p. R104. 109. Tolkatchev, D., et al., Structure dissection of human progranulin identifies well- folded granulin/epithelin modules with unique functional activities. Protein science : a publication of the Protein Society, 2008. 17(4): p. 711-24. 110. Eisen, M.B., et al., Cluster analysis and display of genome-wide expression patterns. Proceedings of the National Academy of Sciences of the United States of America, 1998. 95(25): p. 14863-8. 111. Sparro, G., et al., Isolation and N-terminal sequence of multiple forms of granulins in human urine. Protein Expr Purif, 1997. 10(2): p. 169-74. 112. Tolkatchev, D., et al., Structure dissection of human progranulin identifies well- folded granulin/epithelin modules with unique functional activities. Protein Sci, 2008. 17(4): p. 711-24. 113. Chevallet, M., S. Luche, and T. Rabilloud, Silver staining of proteins in polyacrylamide gels. Nat Protoc, 2006. 1(4): p. 1852-8. 114. Hasegawa, S., et al., Cysteine, histidine and glycine exhibit anti-inflammatory effects in human coronary arterial endothelial cells. Clinical and experimental immunology, 2012. 167(2): p. 269-74. 115. Tanaka, A., et al., Serum progranulin levels are elevated in patients with systemic lupus erythematosus, reflecting disease activity. Arthritis research & therapy, 2012. 14(6): p. R244. 116. Frampton, G., et al., Interleukin-6-driven progranulin expression increases cholangiocarcinoma growth by an Akt-dependent mechanism. Gut, 2012. 61(2): p. 268-77. 117. Scotter, E.L., et al., Differential roles of the ubiquitin proteasome system and autophagy in the clearance of soluble and aggregated TDP-43 species. J Cell Sci, 2014. 127(Pt 6): p. 1263-78. 118. Nguyen, A.D., et al., Progranulin: at the interface of neurodegenerative and metabolic diseases. Trends Endocrinol Metab, 2013. 24(12): p. 597-606. 119. Dazert, E. and M.N. Hall, mTOR signaling in disease. Curr Opin Cell Biol, 2011. 23(6): p. 744-55. 120. Guy, G.R., et al., Sprouty proteins: modified modulators, matchmakers or missing links? J Endocrinol, 2009. 203(2): p. 191-202. 121. Ricciardelli, C., et al., The biological role and regulation of versican levels in cancer. Cancer metastasis reviews, 2009. 28(1-2): p. 233-45. 122. Kranenburg, O., et al., Amyloid-beta-stimulated plasminogen activation by tissue- type plasminogen activator results in processing of neuroendocrine factors. Neuroscience, 2005. 131(4): p. 877-86. 123. Kolb, A.F. and L. Petrie, Folate deficiency enhances the inflammatory response of macrophages. Molecular immunology, 2013. 54(2): p. 164-72. 124. Jian, J., J. Konopka, and C. Liu, Insights into the role of progranulin in immunity, infection, and inflammation. J Leukoc Biol, 2013. 93(2): p. 199-208. 125. Schmidt, O. and D. Teis, The ESCRT machinery. Current biology : CB, 2012. 22(4): p. R116-20.

-92-

126. Alquezar, C., et al., PGRN haploinsufficiency increased Wnt5a signaling in peripheral cells from frontotemporal lobar degeneration-progranulin mutation carriers. Neurobiology of aging, 2014. 35(4): p. 886-98.

-93-