Developing a Potential Substrate Reduction Therapy for Six Mucopolysaccharidoses by Decreasing NDST1 Activity

Ilona Tkachyova

A thesis submitted in conformity with the requirements for the degree of Master’s Laboratory Medicine and Pathobiology Department University of Toronto

Ilona Tkachyova

Master’s Degree

Laboratory Medicine and Pathobiology Department University of Toronto

2013

Abstract

Mucopolysaccharidoses result from genetic mutations in lysosomal enzymes required for degradation of glycosaminoglycans. The deficiency in any of eight lysosomal enzymes needed to degrade heparan sulfate leads to an accumulation of both non-degraded and partially degraded polysaccharides within the lysosomes of many tissues. Interestingly, six of these deficient enzymes can be treated by a relatively new approach – substrate reduction therapy

(SRT), which aims to reduce the synthesis of the substrate for the deficient enzyme being targeted. I developed a cell-based high throughput screen assay for the identification of compounds that decrease the expression of the first modifying enzyme in HS biosynthesis,

N-deacetylase/N-sulfotransferase 1, by inhibiting the transcription of its mRNA. From the high throughput screen, I identified several compounds, with a previous history of use in humans, which significantly decreased the endogenous NDST1 expression and therefore, could be considered as potential SRT agents for up to six Mucopolysaccharidoses.

Acknowledgments

I would like to extend my sincere gratitude to my supervisor Dr. Don J. Mahuran for giving me the opportunity to be a part of his team. I am grateful for his support, guidance, and expertise in the field of lysosomal biology. He could always find meaningful answers to my confusing results. I appreciate the freedom that was given to me to work independently to acquire more confidence in my own research.

I would like to thank the people in Dr. Mahuran’s lab who have helped me throughout my studies: Dr. Michael Tropak for his fresh ideas and trouble-shooting advice, Dr. Brigitte Rigat for her tremendous support in helping me with my research writings and presentations, Dr.

Xiaolian Fan for sharing her knowledge on enzyme assay protocols.

I would also like to thank my committee members Dr. Berge Minassian and Dr. Gil Privé for providing thoughtful and invaluable feedback on my research project.

I especially thank Dr. Jessie Cameron and Valeriy Levandovskiy for advising on some aspects of my experiments.

I would like to acknowledge Dr. Harry Elsholtz for giving me the chance to enter the LMP program. Without this opportunity I will not be able to obtain this degree.

Thanks to my family and friends who have supported me and were so patient with my spending of endless hours in the lab and on my thesis.

iii

Table of Contents

Abstract ...... ii

Acknowledgments ...... iii

Table of Contents ...... iv

List of Tables ...... viii

List of Figures ...... ix

List of Appendices ...... xi

List of Abbreviations ...... xii

CHAPTER I Introduction ...... 1

1.1 Mucopolysaccharidoses ...... 1

1.1.1 Overview of Mucopolysaccharidoses ...... 1

1.1.1.1 MPS I ...... 1

1.1.1.2 MPS II ...... 2

1.1.1.3 MPS III ...... 2

1.1.1.4 MPS IVA and MPS IVB ...... 2

1.1.1.5 MPS VI ...... 4

1.1.1.6 MPS VII ...... 4

1.1.1.7 MPS IX ...... 4

1.1.2 Present MPS Therapies ...... 5

1.1.3 Emerging MPS Therapies ...... 6

1.2 HS and Related MPSs ...... 8

1.3 HS Proteoglycans ...... 10

1.3.1 Cell Surface HSPGs ...... 11

1.3.2 Extracellular HSPGs ...... 13

1.4 HS Structural and Functional Relationships ...... 13 iv

1.4.1 Growth factors ...... 13

1.4.2 Cytokines ...... 14

1.5 HS GAGs Synthesis ...... 15

1.6 Four N-Deacetylase/N-Sulfotransferases ...... 17

1.6.1 NDST Genes ...... 17

1.6.1.1 Translational and Transcriptional Regulation of NDSTs ...... 18

1.6.2 NDST Isozymes ...... 20

1.6.2.1 Structure of NDST isozymes ...... 20

1.6.2.2 Distribution of NDSTs ...... 22

1.7 SRT for HS Degrading Lysosomal Enzymes ...... 23

1.7.1 Decreasing Activity of EXTs ...... 23

1.7.2 Targeting NDSTs ...... 24

1.8 Hypothesis ...... 25

1.9 Aims and Objective ...... 26

CHAPTER II NDST1 Promoter Analysis ...... 28

2.1 Introduction ...... 28

2.2 Materials and Methods ...... 31

2.2.1 In silico analysis of the alternative NDST1 promoters...... 31

2.2.1.1 Transcription start site prediction ...... 31

2.2.1.2 Histone modifications, Pol II binding sites, CAGE analysis and CpG islands ...... 31

2.2.2 Reverse transcription PCR ...... 32

2.2.2.1 Complimentary DNA PCR ...... 32

2.2.2.2 Genomic DNA PCR ...... 34

2.2.3 Validation of the NDST1 promoter activity ...... 35

2.2.3.1 The NDST1, GAPDH, RO1 and RO2 renilla luciferase reporter constructs ...... 35 v

2.2.3.2 Cell cultures and transfections ...... 36

2.2.3.3 Renilla luciferase activity assay ...... 36

2.2.3.4 Hexosaminidase assay ...... 36

2.3 Results ...... 37

2.4 Conclusion ...... 48

CHAPTER III A High Throughput Screening ...... 49

3.1 Introduction ...... 49

3.2 Materials and Methods ...... 49

3.2.1 The NDST1 promoter firefly luciferase reporter construct ...... 49

3.2.2 Establishment of the NDST1 promoter-renilla luciferase expressing permanent cell line ...... 50

3.2.3 Firefly luciferase activity assay ...... 51

3.2.4 Cells treatment for the primary screen ...... 51

3.2.5 In vitro evaluation of the NDST1 inhibitors (hits from the primary screen) ...... 52

3.2.6 Generation of the dose-response curves for each hit ...... 52

3.2.7 Protein determinations through a Coomassie dot blot analysis ...... 52

3.2.8 Validation of the activity of each HTS hit on the endogenous NDST1 protein expression by SDS-PAGE and Western Blotting ...... 53

3.2.9 Quantitation of the endogenous NDST1 and GAPDH by densitometry analysis ...... 54

3.2.10 Statistical Significance ...... 54

3.2.11 Total NDST (1-4) enzyme activity by an ELISA assay ...... 55

3.3 Results ...... 55

3.3.1 The Screening of Prestwick Library ...... 55

3.3.2 The Analysis of Pyrimethamine Derivatives ...... 77

3.4 Conclusion ...... 82

3.4.1 Property of pyrimethamine (2,4 diamino 5-(4-chlorophenyl)-6-ethylpyrimidine) .... 83

3.4.2 Property of SAHA (Vorinostat, suberoylanilidehydroxamic acid) ...... 84 vi

3.5 Future Directions ...... 84

References ...... 87

Appendices ...... 96

vii

List of Tables

Table 1.1 Mucopolysaccharidoses classification...... 3

Table 1.2 Heparan sulfate proteoglycans ...... 12

Table 2.1 NDST1 cDNA clones submitted to GenBank...... 29

Table 2.2 Internal control primers ...... 32

Table 2.3 NDST1 mRNA primers...... 33

Table 2.4 Reaction setup for cDNA PCR...... 33

Table 2.5 Thermocycling conditions for cDNA PCR ...... 34

Table 2.6 NDST1 gDNA primers...... 35

Table 3.1 Reaction settings for ligation of NDST1 promoter fragment into pGL4.22 [luc2CP/Puro] vector ...... 50

Table 3.2 Potential inhibitors of NDST1 transcription...... 62

viii

List of Figures

Figure 1.1 Principle of EET and SRT...... 6

Figure 1.2 Lysosomal enzymes degrading HS...... 9

Figure 1.3 HSPG structures...... 10

Figure 1.4 Heparan sulfate biosynthesis...... 16

Figure 1.5 Basic structure of NDST isozymes...... 21

Figure 1.6 Mouse and human sulfotransferase domains of NDSTs...... 22

Figure 2.1 Alternative NDST1 mRNA transcripts...... 30

Figure 2.2 Prediction of TSS for NDST1 putative transcripts...... 37

Figure 2.3 Analysis of NDST1 genomic DNA in the vicinity of exon 1 for histone-associated features, RNA polymerase binding site, and 5’- cap sequencing...... 40

Figure 2.4 Presence of CpG islands in the promoter region of NDST1...... 42

Figure 2.5 Reverse transcription PCR for alternative NDST1 transcripts...... 44

Figure 2.6 Alignment of the AB209107 PCR product and the AB209107 mRNA sequence in the data base...... 45

Figure 2.7 Activity of NDST1 promoter renilla luciferase reporter construct...... 47

Figure 3.1 Restriction digestion of NDST1 in pLightSwitch_Prom and empty pGL4.22 [luc2CP/Puro] vector...... 57

Figure 3.2 Restriction digestion of NDST1 in pGL4.22[luc2CP/Puro]...... 58

Figure 3.3 Activity of HeLa clones stably expressing NDST1 in pGL4.22 [luc2CP/Puro]...... 59

Figure 3.4 Standard deviation from the mean of luciferase activity...... 60

Figure 3.5 A high throughput screen of 1,200 Prestwick Chemical Library compounds...... 61

Figure 3.6 The effect of potential NDST1 inhibitors on firefly luciferase activity...... 65

Figure 3.7 The Coomassie dot blot analysis...... 66

Figure 3.8 Dose-dependent inhibition of the NDST1 reporter...... 67

Figure 3.9 Effects of six compounds at their 1 x EC50 on the endogenous NDST1 expression. .. 72

Figure 3.10 Densitometric analysis of NDST1 protein expression compared to loading control

GAPDH after 1 x EC50 treatment...... 73

Figure 3.11 Effects of six compounds at their 2 x EC50 on the endogenous NDST1 expression. 74

Figure 3.12 Densitometric analysis of NDST1 protein expression compared to loading control

GAPDH after 2 x EC50 treatment...... 75

Figure 3.13 ELISA assay of total NDST activity in normal human fibroblast cells...... 77

Figure 3.14 Dose-response relationship of the pyrimethamine derivatives ...... 78

Figure 3.15 Derivatives of pyrimethamine ...... 81

List of Appendices

Supplementary 3.1 Compounds inhibiting the NDST1 promoter firefly reporter construct activity more than 75% ...... 96

Supplementary 3.2 HDAC Inhibitors ...... 101

List of Abbreviations

2-OST 2-O-sulfotransferase 3-OST 3-O-sulfotransferase 6-OST 6-O-sulfotransferase BBB Blood brain barrier BMP Bone morphogenic protein BMT Bone marrow transplantation CAGE Cap analysis gene expression cDNA Complimentary DNA CNS Central nervous system CS Chondroitin sulfate DMSO Dimethyl sulfoxide DS Dermatan sulfate ECM Extracellular matrix EET Enzyme enhancement therapy ER Endoplasmic reticulum ERAD Endoplasmic reticulum associated degradation ERT Enzyme replacement therapy ESTs Express sequence tags EXT1 Exostosin 1 EXT2 Exostosin 2 EXTL2 Exostosin-like 2 EXTL3 Exostosin-like 3 FGFs Fibroblast growth factors GABP GA binding protein GALT1 Galactosyltransferase 1 GALT2 Galactosyltransferase 2 gDNA Genomic DNA GlcA Glucuronic acid GLCAT1 Glucuronyltransferase 1

xii

GlcN Glucosamine GlcNAc N-acetylglucosamine GPI Glycophosphatidylinositol HA Hyaluronic acid Hex Hexosaminidase Hhs Hedgehogs HS Heparan sulfate HSPGs Heparan sulfate proteoglycans HTS High throughput screen IdoA Iduronic acid IFN interferon- IL-2, -3, -4, -6, -7, -8 interleukin-2, -3, -4, -6, -7, -8 IRES Internal ribosomal entry site kb Kilobase pairs KS Keratan sulfate LSDs Lysosomal storage disorders MEF Mouse embryonic fibroblast MITF mi-transcription factor MPS I Mucopolysaccharidosis I MPS II Mucopolysaccharidosis II MPS IIIA-E Mucopolysaccharidosis IIIA-E MPS IVA Mucopolysaccharidosis IVA MPS IVB Mucopolysaccharidosis IVB MPS IX Mucopolysaccharidosis IX MPS VI Mucopolysaccharidosis VI MPS VII Mucopolysaccharidosis VII MPSs Mucopolysaccharidoses NDST N-deacetylase/N-sulfotransferase NDST 1-4 N-deacetylase/N-sulfotranserases 1-4 PAPS 3’-phosphoadenosine 5’-phosphosulphate PDGF Platelet derived growth factor PDZ Postsynaptic density-95, disc large protein, zonula occludents-1

xiii

PG Proteoglycan Pol II RNA polymerase II SD Standard deviation SRT Substrate reduction therapy Sulf 1 Sulfatases 1 Sulf 2 Sulfatases 2 TCOF1 Treacher Collins-Franceschetti syndrome 1 TIGR The Institute of Genome Research TM Transmembrane TSS Transcription start site uAUG Upstream AUG UCSC University of California Santa Cruz UTR Untranslated region UTRs Untranslated regions VEGF Vascular endothelial growth factor XYLT1 Xylosyltransferase 1 XYLT2 Xylosyltransferase 2

xiv 1

CHAPTER I Introduction 1.1 Mucopolysaccharidoses 1.1.1 Overview of Mucopolysaccharidoses

The Mucopolysaccharidoses (MPSs) are a group of rare autosomal recessive lysosomal storage disorders (LSDs) in which glycosaminoglycans (GAGs) serve as the primary accumulated substrate within the lysosomes of many tissues. All MPS disorders result from a deficiency in any one of 12 enzymes involved in cellular GAGs degradation. Based on the clinical phenotypes and the specific enzyme deficiency, MPSs are subdivided into several different groups (Table

1.1).

1.1.1.1 MPS I

Mucopolysaccharidosis type I (MPS I) is caused by a deficiency of α-L-iduronidase activity due to mutations in the IDUA gene, leading to elevated levels of undegraded dermatan sulfate (DS) and heparan sulfate (HS) GAGs within lysosomes. Based on clinical variation in the disease severity observed in patients, MPS I is subdivided in three different phenotypes; Hurler,

Hurler/Scheie, and Scheie. The most severe outcome, Hurler, is generally characterized by the total loss of enzyme activity leading to progressive neurodegenaration, severe aberrations in bone and skeletal development, organomegaly, and cardiorespiratory problems. Hurler/Scheie disorders exhibit less severe phenotypes with almost no impact on cognitive development. The mildest form of MPS I, Scheie, progresses slowly with patients often being diagnosed only in adulthood [1]. The life expectancy ranges from ~8 years for Hurler patients to ~22 years for

Hurler/Scheie, with Scheie patients having a normal life expectancy [2].

1.1.1.2 MPS II

The deficiency of iduronate-2-sulphatase, encoded by the IDS gene, results in

Mucopolysaccharidosis type II (MPS II). To date, more than 330 different mutations in the IDS gene have been identified [3]. Due to this large number of genetic changes, MPS II patients display very heterogeneous phenotypes. Most affected children remain undiagnosed until they reach adolescence. The clinical symptoms of MPS II are characterized by the delay in cognitive and motor development, changes in cardio respiratory function and bone morphogenesis [4].

1.1.1.3 MPS III Mucopolysaccharidosis III is a group of five diseases, MPS IIIA-E, all of which result from abnormal catabolism of HS. Mutations in any of the genes SGSH, NAGLU, HGSNAT, GNS or

ARSG, which code for heparan-N-sulphatase, α-N-Acetylglucosaminidase, acetyl CoA:N- acetyltransferase, N-acetylglucosamine-6-sulphatase or arylsulfatase G, respectively, can result in the accumulation of undegraded HS within lysosomes. The clinical manifestations seen in this group of disorders are mainly attributed to motor and behavioral changes, leading to progressive mental retardation and irreversible neurocognitive decline. Life span of children with MPS III usually ranges from adolescence to early adulthood [5].

1.1.1.4 MPS IV

Mucopolysaccharidosis IV is composed of two diseases, MPS IVA and IVB. A deficiency of either of two enzymes, galactose 6-sulphatase and β-galactosidase, results in the lysosomal storage of keratan sulfate (KS). MPS IVA and IVB result from a genetic defect in the GALNS or

GLB1 gene, respectively. The most common clinical presentation of the diseases is the impairment in respiratory function and skeletal dysplasia; however patients with MPS IVB present with a less severe phenotype [6,7].

Table 1.1 Mucopolysaccharidoses classification.1

Disease Clinical Enzyme deficiency Stored Prevalence2 targets phenotypes substrate MPS I Hurler α-L-iduronidase DS, HS3 1 in 88,000 Scheie Hurler/Scheie MPS II Hunter Iduronate-2-sulphatase DS, HS 1 in 136,000 MPS IIIA Sanfilippo A Heparan-N-sulphatase HS 1 in 114,000 MPS IIIB Sanfilippo B α-N-Acetylglucosaminidase HS 1 in 211,000 MPS IIIC Sanfilippo C Acetyl CoA:N- HS 1 in 1,407,000 acetyltransferase MPS IIID Sanfilippo D N-acetylglucosamine- HS 1 in 1,056,000 6-sulphatase MPC IIIE4 Sanfilippo E Arylsulfatase G HS n/a MPS IVA Morquio A Galactose 6-sulphatase KS 1 in 169,000 MPS IVB Morquio B β-Galactosidase KS n/a MPS VI Maroteaux- N-Acetylgalactosamine DS 1 in 235,000 Lamy 4-sulfatase MPS VII Sly Β-Glucuronidase CS,DS,HS 1 in 2,111,000 MPS IX n/a Hyaluronoglucosaminidase-1 HA n/a

1. Neufeld, E. F. and Muenzer, J., The Mucopolysaccharidoses. In: Scriver, C.R., Beaudet, A.L., Sly, W.S., Valle, D. (ed.): The metabolic and molecular bases of inherited diseases. New York: McGraw-Hill Co., 2001, 3421-3452

2. The incidence of Mucopolysaccharidosis diseases was obtained from Lysosomal Diseases Australia (www.lda.org.au/about.html)

3. Abbreviations: DS – dermatan sulfate, HS – heparan sulfate, KS – keratan sulfate, CS – chondroitin sulfate, HA – hyaluronic acid, n/a – data not available

4. Kowalewski, et al. Proc Natl Acad Sci (USA) 109 (2012) 10310-10315

1.1.1.5 MPS VI

Mucopolysaccharidosis VI (MPS VI) is caused by a deficient of N-acetylgalactosamine

4-sulphatase (arylsalfatase B) activity, due to mutations in the ARSB gene, resulting in lysosomal storage of the DS GAGs. This disorder is characterized by the developmental disturbance of bones formation, organomegaly, and heart failure, but no cognitive involvement.

The severity of symptoms varies wildly among the patients. Severely affected children may survive into their teens, while less severely affected patients live into adulthood [8].

1.1.1.6 MPS VII

Mucopolysaccharidosis VII (MPS VII) is the rarest of all MPSs disorders. Genetic mutations in the GUSB gene, encoding β-glucuronidase, result in accumulation of the chondroitin sulfate

(CS), DS, and HS GAGs in different tissues [9]. The most severe manifestation of this disease is hydropsfetalis, which often result in the death of the premature child. Other MPS VII patients experience developmental delay leading to a progressive mental retardation, with some suffering from skeletal abnormalities and heart dysfunction [10].

1.1.1.7 MPS IX

The failure of hyaluronidase 1 to degrade hyaluronan in the lysosome results in

Mucopolysaccharidosis type IX (MPS IX). Six human genes encoding hyaluronidases have been identified. First three genes (HYAL1, HYAL2, and HYAL3) are located on chromosome 3p21.3, and other three (HYAL4, HYALP1, and SPAM1) are located on chromosome 7q31.3. Mutations in only one of these genes, HYAL1, encoding hyaluronidase 1, have been implicated in MPS IX

[11]. To date four such patients have been diagnosed. The common clinical presentation of the disease is limited to progressive joint degeneration [12].

1.1.2 Present MPS Therapies

Presently, there is no cure for any of the MPSs; however several treatments, such as bone marrow transplantation (BMT) and enzyme replacement therapy (ERT), can ease or slow the progression of some of these disorders [13]. The main goal of these therapies relies on supplying the fully functioning enzyme to compensate for its deficiency in the patient. In BMT the functional enzyme is supplied endogenously by the transplanted cells through secretion of the enzyme followed by its re-capture by receptors on deficient cells and transport to their lysosomes. BMT has been somewhat successful in treating the severe form of MPS I (Hurler) if performed early in the child’s life, before the blood-brain barrier (BBB) is fully formed [14].

Early studies showed that the majority of the severely affected MPS I patients who undergone

BMT had less decline in cognitive development and more improvement in symptoms; such as, clearance of upper airways, movement, joint stiffness, and hearing. Application of BMT for other

MPSs has shown very little efficacy, especially for those with neurological involvement.

Furthermore, despite the improvements in the BMT procedure, the rates of lethality for this treatment remain high [13]. In ERT the active enzyme is supplied exogenously through intravenous injections of the recombinant enzyme. Like the enzyme produced by BMT cells, the replacement enzyme is recognized by receptors on deficient cells, endocytosed and compartmentalized in their lysosomes where it can turnover stored substrate. ERT has been developed for patients having somatic forms of MPS I, MPS II, MPS VI, and it is also in development for MPS IV, and MPS VII [13,15]. Despite having shown some success in treating several of the MPSs, neither of BMT nor ERT can be used for those disorders with neurological complications, since access of the enzymes derived from bone marrow or the recombinant enzymes used in ERT is precluded by the BBB.

1.1.3 Emerging MPS Therapies

As alternatives or adjuncts to BMT and ERT, two small molecule based approaches, enzyme enhancement therapy (EET) and substrate reduction therapy (SRT), have been developed [16].

Unlike enzymes, small molecules have a greater potential to be able to cross BBB and act in the central nervous system (CNS). In EET, small molecules function in the endoplasmic reticulum

(ER) as pharmacological chaperons, assisting the mutant proteins to fold correctly, which protects them from degradation by the ER associated degradation system (ERAD) and facilitating their delivery to lysosomes. In SRT, small molecules act by directly inhibiting a key target enzyme involved in the biosynthesis of the stored substrate (Figure 1.1).

Enzyme Enhancement therapy Substrate Reduction Therapy

Figure 1.1 Principle of EET and SRT.

In EET, small molecules bind to and promote the proper folding of the mutant lysosomal enzyme in the ER, protecting it from degradation by ERAD and facilitating its delivery into the lysosome. In SRT, small molecules bind to and inhibit a target enzyme that is involved in the biosynthesis of the deficient enzyme’s substrate.

EET has been explored for LSDs such as Fabry [17], Gaucher [18], Pompe [19,20], and GM1 and

GM2 gangliosidoses [21,22,23]. Interestingly, EET is the only treatment option for GM1 and GM2 patients. SRT has been used in a GM2 (Sandhoff) mouse model with promising results. The SRT agent used inhibited an early step in glycolipid synthesis (glucosylceramide biosynthesis). In this mouse model GM2 ganglioside storage was reduced, and a delay in the onset and progression of the disease symptoms was observed [24,25]. However, the SRT agent failed to show significant efficacy in human patients [26]. On the other hand the same molecule and another

“improved” inhibitor of glucosylceramide biosynthesis have undergone successful clinical trial as an oral treatment for type I (non-neurological) Gaucher disease [27], but neither proved to be as effective as ERT. SRT has also been shown to be a promising strategy for the treatment of neuropathic forms of MPS. Treatment of a mouse model with Rhodamine B, an effective inhibitor of the elongation of the HS precursor chain, resulted in decreased GAG levels in brain, liver, and urine of MPS IIIA mice [28,29]. Unfortunately, in humans rhodamine B appears to be toxic [30]. Another compound, genistein, a tyrosine kinase inhibitor, has been shown to be effective in treating some symptoms, including behavior and cognitive abilities, in both MPS

IIIA and IIIB mouse models and human patients. However, the efficacy of this treatment is not high enough to reverse or even prevent the progression of the disease in patients [31,32,33,34].

Neither rhodamine B nor genistein are specific inhibitors of GAG synthesis and their mechanisms of action towards the MPSs are unclear [35,36].

1.2 HS and Related MPSs

In eight of the MPSs, the primary substrate stored within lysosomes is HS. The initial degradation of the heparan sulfate proteoglycans (HSPGs) involves extracellular heparanases that release the GAG chains from the core protein. The smaller GAG fragments are then transported into lysosomes and completely degraded [37]. HS is degraded by eight lysosomal enzymes; i.e. four sulfatases (iduronate-2-sulfatase, glucosamine-6-sulfatase, arylsalfatase G, and sulfamidase), three glycosidases (α-L-iduronidase, N-acetylglucosaminidase, and glucoronidase), and one acetyltransferase (heparan-α-glucosaminide N-acetyltransferase). Deficiencies in these enzymes correspond to MPS II, MPS IID, MPS IIE, MPS IIIA, MPS I, MPS IIB, MPS VII, and

MPS IIIC disorders, respectively (Figure 1.2) [38,39]. In addition several recent studies indicate that HS storage results in a positive feed-back to the enzymes involved in the production of the mature HS GAG chains. For instance, mice deficient in α-N-acetylglucosaminidase (MPS IIIB) showed increased expression of multiple HS synthesizing enzymes in the CNS, resulting in abnormal synthesis and elevated levels of membrane-bound cell surface HSPGs [40]. HS levels, composition, and localization are also altered in mice deficient in α-L-iduronidase (MPS I). The excess of HS in these mice resulted in increased activity of the HS-modifying N-deacetylase/N- sulfotransferase (NDST) enzymes [41]. This up-regulation was found to worsen the lysosomal storage pathology and hasten disease progression.

Figure 1.2 Lysosomal enzymes degrading HS.

HS is degraded in a step-wise manner. First, four sulfatases remove specific sulfate groups from a sugar residue located at the non-reducing end of the polysaccharide. Then, free of any sulfate groups, the sugar residue itself is hydrolyzed by a specific glycosidase. Terminal glucosamine residues cannot be directly hydrolyzed. They require an additional synthetic modification, which converts the amino group produced by the N-sulfatase, to an N-acetyl group. This reaction is carried out by N-acetyl transferase.

1.3 HS Proteoglycans

HS proteoglycans (HSPGs) are complex molecules with long unbranched GAG chains covalently attached to core proteins. Core proteins vary with respect to their amino acid composition and the number of GAG chain attachment sites; although usually, not all of these sites are fully occupied. The length of the attached GAG chains and the degrees of post-synthetic modification that occur to them, i.e. addition of sulfate groups, are also variable, with some modification being restricted to specific types of proteoglycans (PGs) [42]. There are three main groups of the HSPGs: i) cell surface transmembrane spanning PGs, ii) those linked to the cell surface through a glycophosphatidylinositol (GPI) domain, and iii) soluble PGs that are secreted into the extracellular matrix (ECM) (Figure 1.3 and Table1.2).

Figure 1.3 HSPG structures.

The integral-membrane syndecans and GPI-anchored glypicans are major HS cell-surface components. The soluble perlecan is the main proteoglycan of the ECM.

1.3.1 Cell Surface HSPGs

HSPGs that span the cell membrane are syndecans. Syndecans are type I glycoproteins having a short cytoplasmic domain, a transmembrane domain, and a large ectodomain. The cytoplasmic domain interacts with various proteins containing a postsynaptic density-95, disc large protein, zonula occludents-1 (PDZ) domain [43]. The extracellular domain is a site for GAGs attachment and interaction with extracellular ligands. Syndecans are ubiquitously expressed in cells of different tissues. Syndecan-1 is mainly expressed in epithelial cells, syndecans-2 is present on endothelial and fibroblasts, syndecan-3 can be found in neurons, and syndecan-4 is present in all cell types [44].

GPI-anchored HSPGs are glypicans. There are six isoforms of glypicans (glypicans 1-6) in human, and all of them specifically carry HS GAG chains. The glypicans core proteins structures are formed as globular domains that are supported by disulfide bonds through conserved cysteine residues [45]. The HS GAG chains are always attached to the base of the glypicans, close to the cell membrane. In general, glypicans are expressed in different tissues mainly during early development [46].

Table 1.2 Heparan sulfate proteoglycans HS proteoglycan Class Core HS GAG Tissue protein, chain MW kDa number Syndecan-1 TM1 33 5 Epithelial cells Syndecan-2 23 3 Fibroblast, endothelial cells Syndecan-3 43 5 Neurons Syndecan-4 22 3 All cell types Glypican-1 GPI- 60-70 1-3 CNS linked Glypican-2 CNS

Glypican-3 All cell types Glypican-4 Neurons, kidney Glypican-5 CNS, limb buds Glypican-6 Heart, liver, kidney Agrin ECM 220 2-3 Neuromuscular junction Collagen type 10-19 1-3 Vascular basement membrane XVIII Perlecan 400 1-4 Basement membrane, cartilage, bone marrow

1. Abbreviations: TM – transmembrane, GPI-linked – glycophosphatidylinositol-linked, ECM – extracellular matrix, CNS – central nervous system.

1.3.2 Extracellular HSPGs

Extracellular HSPGs are perlecan, agrin, and collagen type XVIII. Perlecan is the largest of all proteoglycans (PGs). It can be found in basement membrane, cartilage, bone marrow and in muscle tissues. The core protein of perlecan is approximately 400-450 kDa and together with its

HS GAGs attached, its mass can reach up to 800 kDa [47,48]. Agrin is a large glycoprotein located at the basal lamina of the neuromuscular junction. It presumably carries both HS and CS

GAGs [49]. Collagen type XVIII occupies vascular basement membrane. Its structure is highly complex containing collagenous and noncollagenous domains [50].

1.4 HS Structural and Functional Relationships

The complex structure of the HSPGs allows them to bind and modulate the function of a variety of molecules. Although some evidence suggests that HSPGs can bind to certain class of ligands via core proteins [51], the dominant binding areas within the HSPG are defined by the presence of N-sulfated (NS) domains within its GAG chains.

1.4.1 Growth factors

The best-characterized role of HSPG is ascribed to its ability to bind fibroblast growth factors

(FGFs) and their tyrosine kinase receptors [52]. Specific NS domains of HS are required for

FGF-1, FGF-2 and FGF-7 to respond to exogenous stimuli controlling cellar proliferation

[53,54]. Also, defined sequences of 2-O an 6-O sulfation patterns of HS have been shown to be essential for the activation of their corresponding FGF receptors [55]. The activity of another growth factor, vascular endothelial growth factor (VEGF), was found to be impaired in the absence of N- and 6-O-sulfation, whereas mild effects were observed in the absence of 2-O- sulfate groups within the HS NS domains [56]. Another study indicated that endothelial cell

14 proliferation was dependent on certain HS oligosaccharide structures containing trisulfated domains that exhibited higher binding affinity to VEGF [57]. The structural specificity of HS NS domains was also observed for hepatocyte growth factors (HGF), which preferably bind oligosaccharides with 6-O-sulfation of N-sulfated glucosamine residues [58]. The presence of

HSPGs has been shown to be important for signaling of other growth factors, such as platelet derived growth factor (PDGF), Wnt family, Hedgehogs (Hhs), and bone morphogenic protein

(BMP); however the exact mechanism of these interactions are not well defined [59].

1.4.2 Cytokines

The interaction of HSPGs with cytokines plays an important role in regulation of cell-to-cell communication. The number of cytokines containing binding domains for HSPGs has been identified. This includes interleukin-2, -3, -4, -6, -7, -8, and -10 (IL-2, -3, -4, -6, -7, -8), chemokines, and interferon- (IFN ). HS has been shown to bind interleukins in highly selective and differential manner. For example, the cytokine IL-4 displays high affinity for NS domains of

HSPG [60], whereas IL-5 requires only the presence of iduronic acid (IdoA) residues [61].

Chemokines, a group of 50 small molecules, form at least four distinctive complexes with HS using a small subset of HSPG NS domains [62]. The binding of IFN to its receptor is achieved by the interaction of the amino acid cluster KTKRKR of the cytokine with N-acetylated and glucuronic acid (GlcA)-rich N-sulfated domains flanked by hexa- or octasaccharide of HSPG

[63,64].

HSPGs structural motifs have also shown to be involved in regulation of the intracellular uptake of other multimeric ligands including lipoproteins [65,66] and viruses [67]. The exact

15 mechanism of their internalization is under investigation, but substantial evidence now suggests the sulfated HS GAGs are critical for these interactions [68].

1.5 HS GAGs Synthesis

HS GAG chains synthesis is initiated in the Golgi apparatus and involves a number of specific biosynthetic enzymes. First, the selected serine residues of the core protein are modified by four glycosyltransferases; i.e., xylosyltransferase 1 and 2 (XYLT1 and XYLT2), galactosyltransferase

1 (GALT1), galactosyltransferase 2 (GALT2), and glucuronyltransferase 1 (GLCAT1); to produce the tetrasaccharide linkage region consisting of glucuronic acid (GlcA), two galactose residues and xylose. Then, this linkage region is further extended by exostosin-like 2 and exostosin-like 3 (EXTL2 and EXTL3), which add the first molecule of N-acetyl glucosamine

(GlcNAc), followed by exostosin 1 (EXT1) and exostosin 2 (EXT2), which add numerous repeating disaccharide units of GlcNAc and GlcA (approximately 40-100 sugar residues) to form the HS polysaccharide backbone. The nascent HS polysaccharide is subsequently modified by glucosaminyl N-deacetylase/N-sulfotranserases 1-4 (NDST 1-4), which remove acetyl groups from the selected GlcNAc residues replacing them with N-sulfate groups, using the high-energy sulfate donor 3’-phosphoadenosine 5’-phosphosulphate (PAPS) molecule. After this modification, C5-epimerase can convert GlcA units to iduronic acid (IdoA) units and then, three sulfotransferases; i.e., 2-O-sulfotransferase (2-OST), 6-O-sulfotransferase (6-OST), and 3-O- sulfotransferase (3-OST); can fully complete the final modifications of the HS GAG by the addition of 2-O-sulfo groups to IdoA and GlcA residues, and 6-O-sulfo and 3-O-sulfo groups to

N-acetylglucosamine or glucosamine N-sulfate (GlcN) residues (Figure 1.4) [69]. The final length and number of the sulfated regions vary substantially. Also, HS GAG chains can undergo additional modifications by specific extracellular sulfatases, sulfatase 1 and 2 (Sulf 1 and Sulf 2),

16 which can cause the shedding of some 6-O-sulfo groups [70], and by heparanases (HPAs), which can cleave some HS chains into fragments [37].

Figure 1.4 Heparan sulfate biosynthesis.1 Polymerization of HS is initiated in Golgi apparatus by the synthesis of a tetrasaccharide linkage region on a serine residue in a core protein. Complexes of EXT polymerases extend a polysaccharide chain by alternating addition of N-acetylglucosamine (GlcNAc) and glucuronic acid (GlcA) residues. The nascent polysaccharide chain is subsequently modified by the N- deacetylase/N-sulfotransferase (NDST) isoenzymes, which transfer sulfate groups to deacetylated residues of GlcNAc molecules. This step is required for further modifications by the epimerase and the 2-O-, 6-O-, and 3-O-sulfotransferases that act only in the vicinity of N- sulfated groups. 1. Hereditary Cancer in Clinical Practice 2004; 2(4) pp. 161-173

1.6 Four N-Deacetylase/N-Sulfotransferases

The substitution of the acetyl for N-sulfate groups in GlcNAc is the tipping point in transformation of the precursor to the mature and fully functioning form of HSPG. This conversion step is believed to be activated by any of four glucosaminyl N-deacetylase/N- sulfotranserases (NDST) isozymes. These NDST isozymes are encoded by four evolutionarily related genes (NDST1-NDST4). DNA sequence analysis suggested that NDST3 and NDST4 were derived from NDST1 and NDST2 by duplication with NDST1 and 2 being derived from an ancestral NDST-like gene early in vertebrate evolution [71,72]. The ancestral origin of NDST genes is also supported by the presence of a single ortholog gene in Drosophila melanogaster and Caenorhabditiselegans.

1.6.1 NDST Genes

The first human NDST gene to be discovered, NDST1, came out of a screen of a placental cDNA library using a yeast artificial chromosome contig for Treacher Collins-Franceschetti syndrome 1

(TCOF1) [73]. The location of NDST1 was found downstream of TCOF1 on chromosome 5. The subsequent screening of a human endothelial cell cDNA library resulted in a discovery of another NDST homologous gene, NDST2 [72]. Unlike the NDST1 gene, NDST2 was mapped on chromosome 10 and was transcribed from a complementary strand. The structure of NDST2 exons and intron-exon junction distances were found to be similar to the ones present in NDST1, but lacked a homolog to exon 1. Human NDST3 and NDST4 genes were identified and mapped on chromosome 4 [71,74]. The genes are separated by ~300 kb and transcribed in opposite directions, i.e. NDST3 from the leading and NDST4 from the complementary strands. NDST3 was identified from the searches of EST database using homologous nucleotide sequences to

NDST1. The sequence of NDST3 gene was then used to examine The Institute of Genome

Research (TIGR) library of bacterial artificial chromosomes end sequencing database [75], which uncovered the last member of the NDST family, NDST4.

1.6.1.1 Translational and Transcriptional Regulation of NDSTs

Various aspects of both the characteristics of each gene’s genomic DNA and the structure of its mRNA can influence the rates of transcription and/or translation. The promoter strength and availability of transcription factors binding sites in the genomic DNA are obvious examples of the former. However, the length and complexity of the untranslated regions (UTR) in the transcribed mRNA, as well as the presence of regulatory elements within these regions also play an important role in determining the level of the protein that is ultimately translated.

The regulation of NDSTs translation has proven to be complex and unique for each NDST isozyme. Sequences of the 5’ UTRs in each of the four NDST mRNAs are long with a high complexity of secondary structures, suggesting their restricted 5’-cap translation [76]. In addition, these regions contained several abortive initiation codons, located upstream of the open reading frame (ORF) and preceding the actual translation initiation site. These abortive initiation codons, also known as upstream AUG (uAUG) sequences, have the capacity to restrain a cap- dependent translation through inhibition of ribosome scanning [77]. Furthermore, analysis of the

NDST 5’ UTRs reveal the presence of internal ribosomal entry site (IRES) elements in NDST2,

NDST3 and NDST4 mRNAs [76]. The ability of these IRES to drive the cap-independent translation was confirmed in vitro. These results suggested that differential activity of at least three NDST proteins is controlled on the translational level trough IRES elements. In addition to having long and structurally complex 5’ UTR sequences, two NDST mRNAs, NDST1 and

NDST2, have unusually extended (~5 kb for NDST1 and ~1 kb for NDST2) 3’ UTRs. Recent studies of 3’ end sequences underlined their pivotal role in the regulation of the transcriptional

19 mechanism. For example, several recognized motifs in 3’ UTR known as AU-rich elements

(AREs) have been implicated in mRNA stability through recognition of these elements by specific 3’-UTR-binding proteins, which can either initiate mRNA degradation or enhanced its translation [78]. Moreover, 3’ UTRs can contain specific binding sites recognized by microRNAs. These small non-coding RNAs are capable of gene silencing by either target mRNA cleavage or translational repression of the gene expression [79]. Other evidence of the long 3’

UTR involvement in translational regulation is its ability to interact with 5’ terminus by forming a loop that facilitates interaction of transcription regulating element [80,81]. Several studies have shown that this interaction is required for optimal mRNA translation [80,82]. Whether any of

NDST1 or NDST2 transcripts have any of the above mentioned 3’ UTR regulatory characteristics remain to be determined. However, one can speculate that their unusually long sequences might participate in mRNA loop formation, which can facilitate gene transcription. In fact, increased levels of NDST1 and NDST2 mRNAs compared to the levels of NDST3 and

NDST4 transcripts were observed in multiple RT-PCR experiments.

The expression of genes depends largely on the rate of mRNA synthesis. This process is actively controlled by a complex of regulatory proteins that bind to specific motifs, promoters, in the region of the start site in exon 1. This enables RNA polymerase II (Pol II) to start transcription.

Very little information is available on the transcription regulation of the NDST genes. However, a recent study has demonstrated that for one of the four, NDST2, transcriptional activation is dependent on two mast cells specific transcription factors, the GA binding protein (GABP) and the mi-transcription factor (MITF) [83]. Mutation in the mi allele affected GABP binding to the

GGAA motif of the NDST2 5’ UTR and resulted in decreased expression of NDST2 protein in the skin mast cells.

1.6.2 NDST Isozymes 1.6.2.1 Structure of NDST isozymes

The NDSTs have an essential role in the biosynthesis of HS chains since their activity defines all the other modifications that occur to the HS precursor. All NDSTs are Golgi-residing type II membrane protein with a short cytoplasmic N-terminal tail and a helical transmembrane (TM) region. The large luminal C-terminus contains both the active site domains for N-deacetylation and N-sulfation [84]. The sulfotransferase part of the N-sulfation reaction requires the binding and use of a high-energy sulfate donor 3’-phosphoadenosine 5’-phosphosulphate (PAPS). The

NDST isozymes have 4-6 potential N-glycosylation (predicted by NetGlyc 1.0 Server, data not shown) and 5-8 phosphorylation sites (determined by PhosphoSitePlus, data not shown) (Figure

1.5). These post-translational modifications potentially provide additional control over NDSTs’ activity.

The first well characterized catalytic domain of NDSTs was the N-sulfotransferase domain of

NDST1. It has been crystallized in the presence of PAPS [85] and the structure used as a template to determine the structures of the N-sulfotransferase domains of NDST2-4. Structural modeling of these domains has revealed a high degree of similarity in the PAPS binding sites between the murine and human isozymes; although conformational variety, due to differences in several amino acids of these domains, was observed in both species (Figure 1.6) [71]. The same study also showed that the four murine NDSTs exhibited different kinetic properties, potentially due to the observed differences in their sulfotransferase active sites. The N-deacetylase domain structure of NDSTs has yet to be resolved; however based on in-vitro studies of NDST2 activity, its location has been mapped to the N-terminal portion of the large luminal domain of the protein

(Figure 1.5) [86].

Figure 1.5 Basic structure of NDST isozymes.

The NDSTs are bifunctional enzymes that catalyze N-deacetylation and N-sulfation of GlcNAc residues of HS precursor. All the NDST isozymes have similar structures and locations for their catalytic domains. Post-transcriptional modification however varies among these isozymes. NDST1 has least modifications (shown in here). It contains 4 N-linked glycosylation sites (of these, only 3 appear to be occupied [87]) and 5 phosphorylation sites (3 phosphate groups on serine 502, 503, and 552, and 2 phosphate groups on tyrosine 851, and 852). Abbreviation: TM – transmembrane domain

Figure 1.6 Mouse and human sulfotransferase domains of NDSTs.1 Sulfotransferase domains of NDST isozymes in human and in mouse models were found to be similar in their structures. However, the amino acid composition and surface charges of the active site differ among four NDST isoforms.

1. Aikawa, J., Grobe, K., Tsujimoto, M., and Esko, J.D., Multiple isozymes of heparan sulfate/heparin GlcNAc N-deacetylase/GlcN N-sulfotransferase. Structure and activity of the fourth member, NDST4. J BiolChem, 2001, 5876-82 (276).

1.6.2.2 Distribution of NDSTs

With the exception of skeletal muscle, NDST1 and NDST2 transcripts are ubiquitously expressed in all other fetal and adult, murine and human, tissues. However high levels of NDST2 transcription are restricted to mast cells (see section 1.6.1.1), making NDST1 the primary isozyme in most tissues. The highest NDST1 and 2 (outside of mast cells) mRNA levels were observed in brain, heart, kidney, liver, pancreas, and placenta. NDST3 and NDST4 are expressed at much lower levels, with NDST3 mainly being expressed during early embryogenesis in kidney, liver, lung, spleen and thymus. [74]. It should be noted, however, that levels of

23 transcription do not always correlate with the protein expression, e.g. the NDST isozymes have been shown to be post-transcriptionally and post-translationally modified [76].

1.7 SRT for HS Degrading Lysosomal Enzymes

As mentioned earlier in the section 1.1.3, the goal of SRT for the MPSs is to reduce the biosynthesis of the substrate that is being accumulated within lysosomes due to a deficiency in a glycosidase or sulfatase. In principal any enzyme primarily involved in HS polysaccharide polymerization or post polymerization modifications represents a potential target for SRT.

However, in practice one would prefer to target an enzyme whose inhibition would result in potential SRT for multiple MPS disorders. This goal requires targeting either an enzyme involved in the early synthesis of HS, e.g. EXT and EXTL glycosyltransferases, or one that initiates its post synthetic maturation, e.g. NDST.

1.7.1 Decreasing Activity of EXTs

Decreasing the activity of EXT and EXTL glycosyltransferases that are involved in the synthesis of the HS backbone could potentially be used as SRT for all deficient HS degrading enzymes.

Targeted disruption of EXT1 and EXT2 in mice completely abolished HS synthesis; although all mutants are neonatal lethal [88,89]. Recently, SRT for MPS has been explored by applying small interfering RNA (siRNA) and genetic approaches directed against EXTs. Gene silencing of

EXTL2 and EXTL3 resulted in significant reduction in HS biosynthesis and decreased levels of accumulated polysaccharides in MPS IIIA cell lines. Reduction in HS biosynthesis was also observed in MPS I cell lines, but at much lower levels [90]. Similarly, another genetic SRT

(gSRT) approach showed that crossbred MPS IIIA mice with those that are heterozygous for a deletion in either their EXT1 or EXT2 genes (to produce a homozygous MPS IIIA mouse that is

24 also heterozygous for a deletion in its EXT1 or EST2 genes) significantly reduced levels of HS in all major organs, including the brain. This study demonstrates that by reducing rather than eliminating the activity of a key biosynthetic enzyme, such as EXT, can diminish the disease pathology without harming the organism [91]. Reducing the levels of the EXT glycosyltransferases as a method for SRT remains a promising approach; however there is presently no straightforward method of assaying for their activity, making high throughput screening (HTS) for inhibitors or even secondary screening to confirm a reduction in their activity impractical.

1.7.2 Targeting NDSTs

Similar to the EXTs, SRT targeting NDSTs could be applied as a treatment for multiple MPSs.

The contribution of each individual NDST isozyme to the maturation of HS is not fully understood, but based on currently available data, NDST1 is likely the best candidate for this approach. Additionally given the similarity in structure and the development in our laboratory of a simple and robust ELISA-based assay for NDST activity, HTS for small molecules that directly inhibit all of the NDSTs is possible. Additionally the ELISA assay has allowed us to evaluate the effects of my candidate transcriptional inhibitors of NDST1 on the total NDST activity in human cells (see below).

NDST1 and NDST2 are the major isozymes in most tissues and the brain. They share a ~70% sequence identity [71,92]. To what extent NDST3 and particularly NDST4 are involved in HS biosynthesis remains to be determined, but their tissue restricted mRNA expression suggests they could be active only in the early stages of development (see section 1.6.2.2) [74].

Systemic inactivation of NDST1 in mice produces reduced levels of epimerization, and N- and

O- sulfation of HS [93,94]. These mice also display multiple defects in brain, lung and skeletal development. The majority of NDST1 knockout mice die neonatally [94,95]. However, the accumulation of partially sulfated HS in their tissues supports the idea that other NDST isozymes, i.e NDST2, also contribute to the maturation of the HS GAG chain [94]. On the other hand, mice lacking NDST2 [93] are viable, exhibiting almost no affect on HS biosynthesis and maturation. However, they exhibit selective defects in their connective tissue-type mast cells

[96]. These results imply that the loss of NDST2, but not NDST1 can be compensated for by the other NDST isozyme. Thus, NDST1 appears to be unique. Additionally, its absence does not result in an up-regulation of other HS synthesizing enzymes and therefore, abnormal changes in

HS structure are unlikely to occur if levels of NDST1 were to be reduced. Therefore, down- regulating NDST1 should only result in an overall reduction in the size and frequency of the NS domains in HS. Moreover, NDST activity has been reported to increase rapidly in response to the accumulation of undegraded substrate within lysosomes [40,41], as well as to changes in the expression of other HS modifying enzymes [97,98]. Increased activity of NDST results in a positive-feedback loop producing N-sulfated regions of greater length and frequency in the HS polysaccharide thereby increases the rate of MPS storage in patient cells and thus, increasing the rate of disease progression and exacerbating their clinical phenotype.

1.8 Hypothesis

I believe that by decreasing the levels of NDST1 activity through inhibiting the transcription of its mRNA, the size and frequency of the NS domains in HS will also be decreased. This will result in the reduction of stored, undegraded HS fragments within lysosomes of patients with anyone of 6 different MPS disorders. Decreased storage will slow, stop or even reverse (if

26 sufficient residual mutant enzyme activity remains) the clinical course of patients with either

MPS I, MPS II, MPS IIIA, MPS IIIC, MPS IIID, or MPS IIIE.

1.9 Aims and Objective

Based on the above previously published studies, it is evident that the NDSTs play a key role in defining the size and frequency of the NS-domains in HS. Of the four NDST isozymes, NDST1 is the most important to target (see section 1.7.2). Therefore reducing its activity directly by identifying inhibitors or indirectly by down regulating its transcription or translation will decrease the levels of N- and O-sulfation, and the number of IdoA residues in the mature HS polysaccharide. My objective was to identify small molecules with a previous history of use in humans that could also reduce NDST1 activity by inhibiting the transcription of its mRNA. To reach this objective I had the following aims:

Analyze the genomic DNA sequences submitted to the University of California

Santa Cruz (UCSC) genome database for a potential NDST1 promoter region

Assess the activity of the predicted NDST1 promoter in luciferase-based assay

and establish a permanent cell line stably expressing luciferase driven by the

NDST1 promoter to be used as a HTS assay

Develop 96-well cell-based HTS assay to find small molecules-inhibitors from the

Prestwick Library of Chemicals with a previous history of use in humans that

target NDST1 mRNA transcription and thus endogenous NDST1 enzyme activity

in normal human cells

Confirm the effect of the most potent NDST1 transcriptional inhibitors by treating

a normal human fibroblast cell line and examining the effects on the expression

levels of NDST1 protein by Western blot analysis, and on total NDST activity by

using our establish NDST enzyme activity assay.

This thesis is based on the results obtained from the HTS using my luciferase-NDST1 promoter construct. The small molecules “hits” from the HTS were then validated by examining their effect on the endogenous NDST1 promoter (protein levels) using normal human fibroblasts.

Their ability to decrease the endogenous levels of total NDST activity were then obtained using an ELISA-based enzyme assay developed by Dr. Xiaolian Fan.

CHAPTER II NDST1 Promoter Analysis

2.1 Introduction

The human NDST1 promoter has not been characterized experimentally. Consequently, in order to define which of the NDST1 genomic DNA (gDNA) region I should use in the HTS, I first analyzed the express sequence tags (ESTs) and complimentary DNA (cDNA) clones of NDST1 submitted into GenBank. I found that the submitted NDST1 cDNA sequences were complex.

They varied in length and potentially had multiple promoters acting on multiple initiation sites

(Table 2.1). In the UCSC database there are eight transcripts that are identical in their coding sequences, but differ in their 5’ non-coding regions. The first four (AB209107, AK293746,

AK292448, and U26600) mRNAs have exon 1 and a part of exon 2 encoding the 5’ untranslated region (UTR); whereas the last four (U17970, BC012888, AB590478, and U18918) mRNAs do not contain exon 1 and potentially utilize a part of exon 2 as 5’ UTR. Based on the sequence similarity of these clones, I selected four transcripts AB209107, AK293746, AK292448, and

U18918, which have distinctive exon 1 positions, for further analysis (Figure 2.1). Preliminary analysis of these putative exons positions indicated that they are ~10 kilobase pairs (kb) apart from each other, which excludes the possibility that they are overlapping. Therefore, only differential promoter usage could produce these four mRNA transcripts. To determine which of the gDNA regions in the proximity of these alternative exons most likely contain the promoter, I applied available computational methods to search for promoter-associated features; such as transcription start site, RNA polymerase II (Pol II) binding site, CpG islands, and histone acetylation or methylation. I also looked at the cap analysis gene expression (CAGE) data. In addition, in order to confirm the transcription of these alternative mRNAs, I performed reverse

29 transcription-PCR analysis using RNA and direct PCR using gDNA samples extracted from normal human fibroblast cell lines.

Table 2.1 NDST1 cDNA clones submitted to GenBank.

mRNA Accession Total Full Exon 1 5’ UTR, 3’ UTR, number exon transcript chromosomal bp bp number sequence, bp location

NDST1 AB209107 5,936 149,865,562 426 2,912

AK293746 15 3,449 149,877,340 589 267

AK292448 3,285 149,887,674 502 134

U26600 3,627 149,887,754 423 558

U17970 3,569 149,900,589 228 596

BC012888 14 2,198 149,900,594 223 249

AB590478 2,646 149,900,817 9 8

U18918 3,901 149,900,600 216 1,035

All NDST1 cDNA have similar exon composition in the coding region, whereas the chromosomal position of their exon1s, encoding part of the 5’ UTR, varies substantially.

Figure 2.1 Alternative NDST1 mRNA transcripts.

GenBank submitted sequences of NDST1 cDNA clones indicated four mRNA isoforms AB209107, AK293746, AK292448, and U18918 with distinctive 5’ regions of exon 1 potentially transcribed from four different promoters. E1, E2, E3, and E4 correspond to alternative exons and P1, P2, P3, and P4 to their corresponding promoters.

2.2 Materials and Methods 2.2.1 In silico analysis of the alternative NDST1 promoters 2.2.1.1 Transcription start site prediction

Chromosomal location of the TSS was identified by several computational methods that utilize different predictive models. To evaluate the likelihood of the submitted NDST1 mRNAs being transcribed from their related promoters on the genomic DNA, I analyzed 10 kb of the genomic

DNA sequences upstream and 1 kb downstream of each exon with several TSS prediction programs, using distinct algorithms and matrices; i.e., FPROM (Softberry, Inc.), NNPP (Neural Network

Promoter Prediction, version 2.2, Berkeley Drosophila Genome Project, University of Hohenhei),

TSSG (Softberry, Inc.), FirstEF (CSH, Cold Spring Harbor Laboratory, New York), WWW

Promoter Scan (Computational Bioscience and Engineering Lab, University of Minnesota), and EDP

(Eukaryotic Promoter Database, Swiss Institute of Bioinformatics).

2.2.1.2 Histone modifications, Pol II binding sites, CAGE analysis and CpG islands

To determine chromatin activity in the vicinity of each potential exon 1 of NDST1, two marks were selected as positive indicators, i.e. trimethylation of H3 on lysine 4 (H3K4me3) and acetylation of

H3 on lysine 9 (H3K9ac), and one mark, i.e. trimethylation of H3 on lysine 27 (H3K27me3) was used as a negative indicator of transcription. The values for the selected parameters were obtained from UCSC database.

Mapping of the Pol II binding site and 5’ cap sequencing (CAGE analysis) were also obtained from submitted data to UCSC database.

32 2.2.2 Reverse transcription PCR 2.2.2.1 Complimentary DNA PCR

A PCR-based approach was used to validate the presence of human NDST1 mRNA transcripts in normal human fibroblast cells. PCRs were performed using specific forward primers unique to each candidate exon 1 (non-coding region) and the same reverse primer to the conserved exon 2 (coding region) using cDNA prepared from the extracted total RNA. U18918 NDST1 mRNA was excluded from the analysis because its exon 1 does not exist. Total RNA was prepared from normal human fibroblast cells using RNeasy Mini Kit (Qiagen). To obtain cDNA, total RNA was reversely transcribed by using QuantiTect Reverse Transcription Kit (Qiagen). Total RNA extraction and cDNA preparation were accomplished using the protocols supplied by the manufacturer. PCR reactions were performed using OneTaq2 x Master Mix with GC buffer and OneTaq High GC

Enhancer (New England BioLabs). As an internal control for cDNA PCRs, 18S rRNA primers were used. Primers for cDNA PCR, reaction set up, and thermocycling conditions were designed as follows:

Table 2.2 Internal control primers

18S rRNA Primer Sequence Annealing PCR temperature product size, bp NR_003286.2 Forward CGAACGAGACTCTGGCATGCTAACT 59 0C 484 Reverse TCCTTCCGCAGGTTCACCTACG

33 Table 2.3 NDST1 mRNA primers.

NDST1 Primer Sequence Annealing PCR mRNA temperature product size, bp AB209107 Forward TAATA-TAGGCGCTGGGCCGAGGA 59 0C 637 Reverse CTGCCTGCACAGGCTTGAGT AK293746 Forward CCTCCTGTCTCTTGAGTATC 55 0C 688 Reverse CTGCCTGCACAGGCTTGAGT AK292448 Forward GGATTAGGTGCAGCCGTGTT 59 0C 700 Reverse CTGCCTGCACAGGCTTGAGT Forward and reverse primers flanked the 5’ UTR (exon 1) and the coding (exon 2) region of NDST1 mRNAs. The AB209107 forward primer was modified by addition of TAATA tail to the 5’ end to reduce its melting temperature. Internal control primers flanked the coding region of 18S rRNA.

Table 2.4 Reaction setup for cDNA PCR.

Component Final concentration 10 µM NDST1 Forward primer 0.2 µM 10 µM NDST1 Reverse primer 0.2 µM 2.5 µM 18S rRNA Forward primer 0.05 µM 2.5 µM 18S rRNA Forward primer 0.05 µM cDNA 1 µg OneTaq2 x Master Mix with GC buffer 1 x OneTaq High GC Enhancer 20% Nuclease-free water

Table 2.5 Thermocycling conditions for cDNA PCR

Temperature Time Initial denaturation 94 0C 4 minutes 40 cycles 94 0C 30 seconds 55 0C or 59 0C 60 seconds 68 0C 2 minutes Final extension 68 0C 5 minutes Hold 10 0C ∞

After completion of PCR runs, 22.5 µL of the amplified products were mixed with 2.5 µL of 10 x

DNA loading dye (20% glycerol, 0.1 M disodium EDTA pH 8, 1% SDS, 0.25% bromphenol blue,

0.25% xylene cyanol) and separated on a 1% agarose gel. Then PCR products were excised from the gel and purified with the Gel Extraction Kit (Qiagen), according to the manufacturer’s protocol. The identity of the purified PCR fragments was confirmed by the sequencing using the designed primers for cDNA PCRs.

2.2.2.2 Genomic DNA PCR

Genome integrity and appropriate selection of the primers for the unsuccessful cDNA PCR reactions were validated from the subsequent PCRs by using gDNA extracted from the same normal human fibroblast cell line. Genomic DNA was prepared by using Puregen Kit A (Quigen) following procedures recommended by the supplier. Genomic DNA PCR reactions were performed using same forward primers (for non-coding region of exon 1) as for the cDNA and the reverse primer for the intron 1 unique to each transcript. To increase binding capacity of the primers, the gDNA was

35 fragmentized prior to its addition to the PCR reaction (without affecting regions of interest) with

DpnI restriction enzyme overnight at 37 0C. Primers for gDNA PCRs were designed as follows:

Table 2.6 NDST1 gDNA primers.

NDST1 Primer Sequence Annealing PCR product mRNA temperature size, bp

AK292448 Forward GGATTAGGTGCAGCCGTGTT 56 0C 820

Reverse ATTGCCGCCTGTGAGTCTCT

AK293746 Forward CCTCCTGTCTCTTGAGTATC 52 0C 821

Reverse TGAGGAACAGGTCAGCTAAC Forward and reverse primers flanked the 5’ UTR (exon 1) and the non coding (intron 1) region of NDST1 gDNA. Forward primers were identical to those used in cDNA PCRs (all of which used the same reverse primer).

Thermocycling conditions and reaction setup for gDNA PCRs were identical to cDNA PCRs, except there was no internal control and amount of gDNA was used at 200 ng per reaction.

2.2.3 Validation of the NDST1 promoter activity 2.2.3.1 The NDST1, GAPDH, RO1 and RO2 renilla luciferase reporter constructs

The renilla luciferase reporter constructs containing either the NDST1 promoter, GAPDH promoter

(a positive control) or one of two random genomic DNA fragments, RO1 and RO2 (negative controls) in the pLightSwitch_Prom expression vector were purchased from SwitchGear Genomics

(switchgeargenomics.com). The actual NDST1 genomic DNA sequence cloned in front of the renilla luciferase consisted of 660 bp upstream and 401 bp downstream of exon 1 (mRNA sequence from the UCSC genome browser: AB209107). The NCBI sequences of the reference human genome for

36 the GAPDH promoter corresponded to NG_007073.2, for RO1 – NG_013083.1 and for RO2 –

NW_004078000.

2.2.3.2 Cell cultures and transfections

Cell lines were normally maintained in AMEM containing 10% fetal bovine serum (FBS), 100

0 units/mL penicillin, and 100 µg/mL streptomycin at 37 C in 5% CO2. All reagents were purchased from Wisent. To validate the activity of the NDST1 promoter-renilla luciferase reporter construct, I performed transient transfections using HeLa cells. HeLa cells were seeded into white 96-well plate

(BD Falcon) (4 x 104 cells per well) 24 hours before transfection. The following day, cells were incubated with 200 ng of NDST1, GAPDH, RO1 and RO2 renilla luciferase reporter constructs plus

0.5 µL of Lipofectamine 2000 (Invitrogen) in antibiotic-free culture medium. The pmaxGFP plasmid

(Lonza) was used as a positive control to determine the transfection efficiency.

2.2.3.3 Renilla luciferase activity assay

The renilla luciferase activity was determined eighteen hours post-transfection using the luciferase assay kit according to the protocol supplied by SwitchGear Genomics.

2.2.3.4 Hexosaminidase assay

The lysosomal hexosaminidase (Hex) assay was used as a control for variability in cell number in transient transfections. The luminescence values produced by renilla luciferase were normalized to the fluorescence generated by the activity of Hex. An assay was performed in clear 96-well plate

(BD Falcon). After measuring luciferase activity, 20 µL of same cell lysate was transferred into the well of 96-well plate containing 5 µL of CP buffer (0.1 M citric acid, 0.2 M Na2HPO4, pH 4.1) supplemented with 0.5% HSA. Plate was equilibrated at 37 0C for 10 minutes followed by the addition of 25 µL of 3.2 mM of the fluorogenic Hex substrate MUG (4-methylumbelliferyl-2-

37 acetamide-2-deoxy-b-D-glycopyranoside). The reaction was carried out at 37 0C for 30 minutes and then stopped by the addition of 200 µM MAP, pH 10.5 (0.1 M 2-amino-2-methyl-1-propanol, 40 mM HCl). Fluorescence readings were obtained at 365 nm excitation and 450 nm emission using

Molecular Devices Spectra Max M2.

2.3 Results

To evaluate the possibility that the NDST1 gene can use different promoters to produce mRNAs with different exon 1 sequences, I have performed in silico analysis of each potential exon 1 for a possible transcription start site. Genomic DNA analysis of 10 kb upstream and 1 kb downstream of the NDST1 exon 1 produced much higher scores for TSS in the AB209107 and U18918 transcripts than in the AK293746 and AK292448 transcripts. The overall highest score was accredited to

AB209107, as all five TSS prediction programs located the TSS at a very close proximity (-9,500

+10,500 bp region) to the chromosomal location of exon 1 submitted to GenBank (Figure 2.2).

Figure 2.2 Prediction of TSS for NDST1 putative transcripts. Genomic DNA, 10 kb upstream to 1 kb downstream from the first nucleotide of each alternative exon 1, was analyzed by TSS prediction programs: FPROM, NNPP, TSSG, FirstEF, WWW Promoter Scan. AB209107 exon 1 region obtained highest score of having TSS.

Figure 2.2 Prediction of TSS for NDST1 putative transcripts (continue)

Following the NDST1 TSS prediction analyses, I analyzed the gDNA sequences around each alternative exon 1 for the presence of promoter-associated features; such as histone acetylation, histone methylation, and Pol II binding sites. This information was retrieved from the data submitted to UCSC database. Furthermore, I have included the latest technology data on CAGE analysis for each putative NDST1 transcript. Promoter specific characteristics, such as histone modifications, indicated active chromatin mostly for the AB209107 transcript. Pol II binding site and 5’ cap sequencing was identified for AB209107, AK293746 and U18918 (Figure 2.3).

Figure 2.3 Analysis of NDST1 genomic DNA in the vicinity of exon 1 for histone-associated features, RNA polymerase binding site, and 5’- cap sequencing.

Genomic DNA region within ± 500 bp of exon 1 (red box) was examined for histone associated features. H3K4me3 and H3K9ac marks (pale and dark green bars) were selected as indicators of high chromatin activity, and H3K27me3 (purple bars) was used as low chromatin activity. RNA polymerase binding site is shown in blue. Sequencing of 5’- cap (ENCODE CAGE analysis) is indicated by black bars.

Figure 2.3 Analysis of NDST1 genomic DNA in the vicinity of exon 1 for histone-associated features, RNA polymerase binding site, and 5’- cap sequencing (continue).

To complete the in silico analysis of the NDST1 alternative promoters, I examined gDNA sequences for the presence of CpG islands using the MethPrimer application. Two genomic DNA regions

(AB209107 and U18918) contained CpG islands with AB209107 having the longest region of

~1,500 bp (Figure 2.4).

AB209107 TSS

AK293746 TSS

Figure 2.4 Presence of CpG islands in the promoter region of NDST1.

The presence of CpG islands was determined by MethPrimer. Minimum GC content was set at 50%. A long CpG Island (~ 1,500 bp) was present in the gDNA within exon 1 of AB209107, as were a short island (~120 bp) downstream of AK293746 TSS, and another short island (~ 200 bp) downstream distant from TSS in AK293746.

AK229448 TSS

U18918 TSS

Figure 2.4 Presence of CpG islands in the promoter region of NDST1 (continued).

The cumulative data of the different TSS prediction programs, promoter associated features and the 5’-cap sequencing gave the highest scores for the promoter in the vicinity of exon 1 for AB209107.

44 To identify expression levels of each alternative NDST1 transcripts in normal human fibroblast cells,

I performed reverse transcription PCRs using primers indicated in Table 2.3. The complementary sequences of these oligonucleotides were located in the 5’ UTR of each candidate for exon 1 and in the coding region common to all in exon 2, resulting in PCR products of 637 bp for AB209107, 700 bp for AK292448, and 688 bp for AK293746, respectively. PCRs were performed with the addition of 18S rRNA primers to amplify the region of 484 bp to serve as an internal positive control. The

PCR of the alternative NDST1 transcripts identified the efficient transcription of the AB209107 mRNA, whereas no PCR product was obtained for AK293746 and AK292448 mRNAs (Figure 2.5).

Figure 2.5 Reverse transcription PCR for alternative NDST1 transcripts. A) Complementary DNA reversely transcribed from total extracted RNA of normal human fibroblast was used in the PCR to identify expression levels of each of the possible NDST1 transcripts. Primers flanked the sequence of 5’ UTR (exon 1) and coding region (exon 2) should produce 637, 700, 688 bp amplicons for AB209107, AK292448, and AK293746, respectively. PCRs with these primers amplified only the predicted region in AB209107. The identity of this PCR product was confirmed by sequencing. 18S rRNA was used as an internal control. B) PCRs of genomic DNA PCR were used to confirm the integrity of the genome and validity of the selected primers. Primers flanked the sequence of 5’ UTR (exon 1) and non coding region (intron 1) produced the predicted 820 bp for AK292448 and 821 bp for AK293746.

Sequencing of the amplified PCR fragment resulted in 100% homology to published sequence for 209107 cDNA clone (Figure 2.6)

Exon 1

start Ssssss Start

Start

tart

Figure 2.6 Alignment of the AB209107 PCR product and the AB209107 mRNA sequence in the data base. AB209107 PCR fragment was aligned to AB20910 mRNA using ClustalW2 (EMBL_EBI application).

46 To confirm that inefficient PCRs for AK293746 and AK292448 were not due to inappropriate primers selection or genomic instability, I performed another set of PCR reactions using gDNA extracted from the same normal human fibroblast cell line. In the genomic PCR, forward primers were the same as those used in the cDNA PCRs, i.e. flanking the 5’ exon 1 region and the reverse primer flanking the region of intron 1. The genomic PCRs were successful for both AK293746 and

AK292448 transcripts (Figure 2.5). Reverse transcription PCRs indicated that AB209107 was transcribed at high levels. On the other hand, absence of PCR bands for AK293746 and AK292448 could be explained by either very low levels of expression, or selective expression in a specific cell type and/or in time dependent manner.

Base on overall predictions described above I chose to analyze the activity of the promoter in the vicinity of AB209107 exon 1 in gDNA by the luminescent reporter based assay. The plasmid expression constructs were purchased from SwitchGear Genomics containing a human genomic

DNA fragment corresponding to the sequence of the AB209107 promoter (660 bp upstream and 401 bp downstream of exon 1), two random genomic DNA regions (as negative controls), and GAPDH promoter (as a positive control). All four DNA fragments were cloned in front of renilla luciferase in pLightSwitch_Prom expression vector. After being transiently transfected into HeLa cells, the activity of each construct (the luminescent light produced) was measured accordingly instructions supplied by the company. To control for variability in cell number, luminescence was normalized to fluorescence produced by the enzymatic reaction of lysosomal Hex with its substrate MUG (Figure

2.7).

Figure 2.7 Activity of NDST1 promoter renilla luciferase reporter construct.

Renilla luciferase expression vector containing either; i) two random DNA sequences, RO1 and RO2 (as negative controls); ii) the predicted NDST1 promoter; or iii) the GAPDH promoter (as positive control) were transiently transfected into HeLa and their relative activities were measured.

48 The results obtained from HeLa cells transient transfections indicated that NDST1 promoter activity was approximately three fold higher than the activity of negative controls (RO1 and RO2). In comparison to positive control GAPDH, the NDST1 promoter activity was ~5-fold lower. It is clear from these results that NDST1 promoter is not as active as GAPDH; however it showed a sufficient amount of activity to be used for the HTS.

2.4 Conclusion

The comparative analysis of the gDNA in the vicinity of the first exons of the NDST1 cDNA sequences submitted to GenBank, predicted the most likely promoter region was around exon 1 of the AB 209107 transcript. The expression of this transcript was confirmed by the reverse transcription PCR and by the sequencing of the PCR product. The activity of the predicted NDST1 promoter was weaker than the GAPDH promoter, which is known to be a strong promoter.

Nonetheless the activity produced by the NDST1 promoter fragment was sufficient for its use in the

HTS. It must also be noted that it is possible that this promoter fragment lacks some additional regulatory elements present in the endogenous promoter. However this situation would only result in false negatives in the HTS, not false positives.

The possibility of NDST1 is also being transcribed from other alternative promoters cannot be completely excluded, because some of the transcription associated features were present for the other alternative transcripts. For example, gDNA around exon 1 of the AK293746 and U18918 transcripts contained some indication of Pol II binding. Additionally, both transcripts were validated by CAGE analysis. To confirm the existence of U18918 mRNA is virtually impossible as this transcript does not contain exon 1; its transcription is initiated from the non-coding portion of exon 2. The other alternative NDST1 transcripts are likely expressed at very low levels, if at all, and therefore are not detectable by reverse transcription PCR.

CHAPTER III A High Throughput Screening

3.1 Introduction

In order to develop a 96-well based HTS assay for the inhibitors of NDST1 transcription, I used the in silico analyzed and reverse transcription PCR confirmed NDST1 promoter defined in the previous chapter. The NDST1 promoter region that was cloned into the renilla luciferase pLightSwitch_Prom expression vector, consisted of 660 bp of genomic DNA upstream and 401 bp downstream of exon 1

(as defined by the AB209107 mRNA). Initial analysis showed that this promoter is functional and active, and therefore could be used in my HTS assay. In this chapter, I present the primary screen results obtained from the Prestwick Chemical library, which consists of 1,200 FDA-approved compounds.

3.2 Materials and Methods 3.2.1 The NDST1 promoter firefly luciferase reporter construct

The firefly luciferase reporter construct containing the NDST1 promoter was generated by sub- cloning of the promoter fragment from pLightSwitch_Prom vector into pGL4.22 [luc2CP/Puro] vector (Promega). The NDST1 promoter in pLightSwitch_Prom and the pGL4.22 [luc2CP/Puro] vectors were digested with XhoI and BglII restriction enzymes overnight at 37 0C. Digested products were separated on a 1% agarose gel in TBE (90 mM Tris, 90 mM Boric acid, 2 mM EDTA) and purified by using QIAquick Gel Extraction Kit (Qiagen). The ligation of the NDST1 promoter fragment into the digested pGL4.22 [luc2CP/Puro] was performed by using T4 DNA ligase (New

England BioLabs). The ligation reaction settings were as follows:

50 Table 3.1 Reaction settings for ligation of NDST1 promoter fragment into pGL4.22 [luc2CP/Puro] vector

Component Amount

pGL4.22 [luc2CP/Puro] vector 300 ng

NDST1 promoter fragment 200 ng

T4 DNA ligase reaction buffer 3 µL

dH2O 6.5 µL

10 x T4 DNA ligase 0.5 µL

Incubation conditions 16 0C overnight

Clones containing the NDST1 promoter insert were obtained by the transformation of the overnight ligation reaction into the DH5α competent cells (Invitrogen). Transformation procedure and selection of the NDST1-positive clones were accomplished by following the protocol supplied by

Invitrogen. The integrity of the NDST1 promoter in pGL4.22 [luc2CP/Puro] clones was confirmed by the restriction digestion with XhoI and BglII and by the sequencing, using specific primers for pGL4.22 vector (ACGT Corp).

3.2.2 Establishment of the NDST1 promoter-renilla luciferase expressing permanent cell line

To obtain a permanent cell line stably expressing the firefly luciferase under the control of the

NDST1 promoter, HeLa cells were transfected with 200 ng of the NDST1 promoter in pGL4.22

[luc2CP/Puro] vector. The pGL4.22-NDST1 promoter DNA was incubated with 0.5 µL of

Lipofectamine 2000 in antibiotic-free culture medium and added to the 4 x 104 cells for 24 hours.

The next day cells were passaged at 1:50 dilution into fresh AMEM containing 10% FBS, 100

51 units/mL penicillin, 100 µg/mL streptomycin and 3 µg/mL puromycin (for selection of transfected cells). At three weeks post-transfection, clonal populations of surviving cells were validated for the expression of firefly luciferase activity.

3.2.3 Firefly luciferase activity assay

HeLa cells stably expressing the NDST1 promoter construct were seeded into white 96-well plate (4

4 0 x 10 cells per well) for 16 hours at 37 C in 5% CO2. The next day, cells were washed 2 x with PBS and lysed with 20 µL of the lysis buffer (1% Triton-100, 20 mM Tricine). Following lysis, 100 µL of a home-made firefly luciferase assay reagent (20 mM Tricine pH 7.8, 4 mM MgSO4, 0.1 mM

EDTA, 33.3 mM DTT, 270 µM CoA litium salt, 530 µM ATP, 470 µM D-luciferin sodium salt) was added into each well. The luminescence produced by cells was measured using Envision 2102

Multilabel Reader (Perkin-Elmer).

3.2.4 Cells treatment for the primary screen

HeLa cells stably expressing the NDST1 promoter construct (8.75 x 103 cells per well) were seeded into a solid white 96-well plate for 24 hours. The next day, 2 µL of each Prestwick library compound

(final concentration 100 µM in 1% DMSO) was added into 80 wells (A2- G11) and 2 µL of DMSO

(as a blank) into 16 wells (A1-G1 and A12-G12 rows) of each plate. Following an overnight

0 treatment at 37 C in 5% CO2, the compounds-containing medium was aspirated and cells were washed 2 x with PBS. Lysis buffer (20 µL) was then added to each well, immediately followed by

30 second incubation at room temperature with rocking on a platform shaker. After lysis, 100 µL of the home-made firefly luciferase assay reagent was added. The luminescence readings were collected using Envision 2102 Multilabel plate reader (Perkin-Elmer). Drug treatments were performed in triplicates.

52 3.2.5 In vitro evaluation of the NDST1 inhibitors (hits from the primary screen)

In vitro testing of NDTS1 inhibitors was done to determine if any of the hits could have a nonspecific effect on firefly luciferase activity. One confluent 10 cm plate of HeLa cells stably expressing the NDST1 promoter construct (8.6 x 106 cells per well) was lysed in 1.7 mL of the lysis buffer. Then 20 µL of the cell lysate was mixed with 2 µL of 1 mM compound stock in a solid white

96-well plate and left for 10 minutes at room temperature with rocking on a platform shaker.

Following the incubation, 100 µL the home-made firefly luciferase assay reagent was added and the emitted light was measured using Envision 2102 Multilabel plate reader (Perkin-Elmer).

3.2.6 Generation of the dose-response curves for each hit

The selected compounds “hits” from the Prestwick Chemical library; i.e., R(-) apomorphine hydrochloride hemihydrates, fenbendazole, mifepristone, meclocycline sulfosalicylate, thiostrepton, pentamidine isetionate, entacapone, vorinostat (SAHA), anthralin, pyrimethamine, and griseofulvin; were diluted in DMSO to obtain 0.011, 0.033, 0.1, 0.3, 0.9, 2.7, and 8.1 mM stock concentrations.

Twenty-four hours prior to treatment, 8.75 x 103 of the NDST1 reporter transfected HeLa cells were seeded into the wells of a white 96-well plate. The following day compounds, at final concentrations of 0.11 µM, 0.33 µM, 1 µM, 3 µM, 9 µM, 27 µM, 81 µM, were added to each well and incubated overnight at 37 0C. After treatment, the luminescence response was detected as described above.

3.2.7 Protein determinations through a Coomassie dot blot analysis

To control for luminescence variability, the total concentration of the protein in each well was calculated. Following cell lysis, 2 µL of the cell lysates were spotted onto Mini Trans-Blot Filter

Paper (Bio-Rad). An optimal range of the protein amounts to be detected on the filter paper was

53 determined from the standard curve of the serially diluted BSA controls (5, 25, 50, 125, 250, and 500

µg/mL). After air-drying the samples for 30 minutes, the filter paper was immersed in Coomassie blue stain (50% methanol, 10% acetic acid, 0.05% brilliant blue R-250) for 20 minutes with slow shaking. Then it was extensively washed with dH2O until a white background was obtained. To calculate the protein concentration in each Coomassie stained dot, the filter paper was scanned on

Epson Perfection 4990 Photo Scanner and analyzed using ImageJ Dot blot analysis

(http://imagej.nih.gov/ij/).

EC50 for each tested compound was calculated using log (inhibitor) vs. response equation (built-in nonlinear regression analysis) in GraphPad Prism version 5.00 for Windows, GraphPad Software,

San Diego California USA, www.graphpad.com.

3.2.8 Validation of the activity of each HTS hit on the endogenous NDST1 protein expression by SDS-PAGE and Western Blotting

The effects of the selected compounds from the Prestwick library on endogenous NDST1 protein levels were validated by Western blot analysis. After 5 days treatment at either 1 x EC50 or 2 x EC50, cells from one 10 cm plate were harvested by scraping and then solubilized in 300 µL of the lysis buffer (1% Triton-100, 50 mM Tris pH 7.4, 150 mM NaCl) by vortexing. The insoluble cellular remains were separated from the lysates by centrifugation at 13,000 rpm for 1 minute. Before loading on to a SDS-PAGE gel, samples were prepared by adding 4 x sample buffer (62.5 mM Tris pH 6.8, 10% glycerol, 2% SDS, 0.002% bromophenol blue, 5% β-mercaptoethanol) to the 30 µg of the total lysate proteins and incubating for 15 minutes at 56 0C. Proteins were separated by SDS-

PAGE using Nu-PAGE 4-12% Bis-Tris gel (Invitrogen) and electrophoretically transferred to the polyvinylidene fluoride (PVDF) membrane (Bio-Rad) in transfer buffer (10 mM CAPS pH 11, 10%

54 methanol) for 1 hour at 4 0C. After transfer, the membrane was blocked with 5% skim milk in TBS-

T (25 mM Tris pH 7.4, 137 mM NaCl, 2.7 mM KCl, 0.1% Tween-20) overnight at 4 0C. Following blocking, the membrane was probed with the primary antibody for 1.5 hour at room temperature and the secondary antibodies for 1 hour at room temperature after washing with TBS-T. Signals were produces using enhanced chemiluminescence (ECL) (Amersham GE Healthcare). The primary antibodies used in Western blot analysis were: mouse monoclonal NDST1 (E-9) antibody (Santa

Cruz) at 1:200 dilutions, and rabbit polyclonal GAPDH antibody at 1:10,000 dilutions. The secondary antibodies were either horseradish peroxidase conjugated goat anti-mouse (Santa Cruz) or horseradish peroxidase conjugated donkey anti-rabbit (Jackson Immunology) at 1:10,000 dilutions.

Films were developed using SRX-101A Medical Film Processor (Konica Minolta Medical &

Graphic Inc.).

3.2.9 Quantitation of the endogenous NDST1 and GAPDH by densitometry analysis

The changes in NDST1 expression were assessed using lane profile plots generated by ImageJ one- dimensional electrophoretic gel analysis. Peak areas for NDST1 and GAPDH bands, obtained by the

ImageJ Western blot analysis, were converted into percentages and were plotted using relative ratios of samples treated with the compound to the sample treated with DMSO alone.

3.2.10 Statistical Significance

All compounds of the Prestwick library were screened in a 96-well format (80 compounds per plate) in triplicates. DMSO was used as a negative control. Variation within a sample set (one plate) was tested by calculating the standard deviation of the average of DMSO treatment.

55 3.2.11 Total NDST (1-4) enzyme activity by an ELISA assay

A streptavidin-coated plate was washed with TBS-T (10 mM Tris-HCl buffered saline with 0.05%

Tween-20, pH 7.4). Then 100 μl of K5 (a bacterial polysaccharide with a structure identical to the nascent, unsulfated HS chain)-biotin, which was diluted to 1.3 μg/ml with 0.2% gelatin in TBS-T, was added to each well and incubated at 37 0C for 30 minutes. After washing with TBS-T, 50 μl of a cell lysate from either yeast expressing mouse NDST1 (positive control) or treated and untreated normal human fibroblasts, were diluted with MES buffer (50 mM MES, 1% Triton X-100, 10 mM

MnCl2, pH 6.5) containing 0.2% gelatin, and added to each well. The enzyme reaction was carried out at 37°C for 30 min and stopped by adding 5μl 5M NaCl (see below) to each well. After completion of the enzyme reaction, the wells were washed and 100 μl of the monoclonal antibody

JM403 (Seikagaku Corporation), which recognizes GlcNH3 residues produced in the K5 polysaccharide by NDST activity, diluted to 1 μg/ml (1:1000 dilution) with 0.2% gelatin in TBS-T was then added. After a 1hr incubation at room temperature, the wells were washed with TBS-T, and

100 μl of an HRP-labeled goat anti-mouse immunoglobulin G+M antibody [82], which was diluted with 0.2% gelatin in TBS-T (1:20,000 dilution), was added to each well. After another 1 hr incubation at room temperature, the wells were washed again with TBS-T, and 100 μl/well of

3,5,3’,5’ tetramethylbenzidine (TMB, Thermo Scientific) substrate solution was added to develop the color. The reaction was stopped by adding 100 μl of 2M H2SO4 to each well, and the absorbance at 450 nm was measured (reference wavelength 630 nm).

3.3 Results 3.3.1 The Screening of Prestwick Library

To determine the effect of small molecules inhibitors (hits) on NDST1 promoter activity and to eliminate the potential inconsistency of transient transfections I have established a permanent cell

56 line stably expressing the firefly luciferase under the control of the NDST1 promoter. The initial

NDST1 promoter-renilla luciferase construct used in transient transfections could not be used to generate a permanent cell line as it did not contain an appropriate antibiotic selection marker. For this reason I cut the NDST1 promoter fragment from pLightSwitch_Prom with XhoI and BglII enzymes, and religated it into the pGL4.22 [luc2CP/Puro] expression vector (Figure 3.1). The pGL4.22 [luc2CP/Puro] vector contained puromycin resistant gene for selection of cells that incorporate the NDST1 promoter construct into their genome. In addition, this vector had a modified firefly luciferase gene with a double degredative sequence at its 3’ end for fast detection in response to any change induced by the drug treatment. The newly produced NDST1 promoter-firefly luciferase construct was validated by restriction digestion with XhoI and BglII (Figure 3.1 and 3.2) followed by sequencing.

Figure 3.1 Restriction digestion of NDST1 in pLightSwitch_Prom and empty pGL4.22 [luc2CP/Puro] vector.

XhoI and BglII digested NDST1 promoter in the pLightSwitch_Prom vector and the empty pGL4.22 [luc2CP/Puro] vector separated on a 1% agarose gel. The 1,080 bp DNA fragment corresponds to the NDST1 promoter construct. DNA fragments of 3,656 bp and 5,583 bp correspond to the pLightSwitch_Prom and the pGL4.22 [luc2CP/Puro] vectors, respectively. NDST1 and pGL4.22 [luc2CP/Puro] were gel purified and ligated to produce a new NDST1 promoter expression construct.

Figure 3.2 Restriction digestion of NDST1 in pGL4.22[luc2CP/Puro].

Religated NDST1 promoter in the pGL4.22 [luc2CP/Puro] vector digested with XhoI and BglII produced fragments of the expected size, corresponding to 1,080 bp for NDST1 and 5,583 bp for pGL4.22 [luc2CP/Puro].

Validated by restriction digestion and sequencing, the NDST1 promoter construct in pGL4.22

[luc2CP/Puro] was used to transfect HeLa cells and establish a stable cell line. I selected several antibiotic resistant clones positive for the expression of the NDST1 construct. The activity of these clones and the luminescent light generated were validated using firefly luciferase enzyme assay. The clone with the highest luminescence activity was selected for the HTS (Figure 3.3).

Figure 3.3 Activity of HeLa clones stably expressing NDST1 in pGL4.22 [luc2CP/Puro].

The level of luminescence produced by several clonal colonies stably expressing firefly luciferase under the control of the NDST1 promoter was examined. The clone with highest NDST1 promoter activity (clone 6) was selected for subsequent experiments.

HeLa cells stably expressing the NDST1 promoter construct were seeded into solid white 96-well plates and incubated for 24 hours at 37 0C. After the addition of 2 µL of the 10 mM stock of each compound from the library (final concentration of the drug in media was 100 µM), the cells were incubated for another 16 hours. Each plate also contained cells that were treated only with DMSO to be used as a negative control. We did not have a positive control at that time as there was no known drug that down-regulate the transcription of the NDST1 mRNA. After lysing the cells and adding the firefly luciferase substrate, luminescence was measured and compared to control wells that were treated with DMSO alone. To eliminate cytotoxic compounds that could potentially produce false positives, lysosomal Hex assays were performed in parallel with the luminescence screen. Cells with survival rate of less than 75% were excluded from further testing. To determine the distribution of the luminescence values produced by HeLa cells stably expressing the luciferase reporter prior to

60 HTS, one plate of cells was treated with DMSO alone. Based on the standard deviations from the mean of the DMSO-only treated cells we concluded that our established 96-well firefly luciferase- based assay meets the criteria of reliability and robustness (Figure 3.4).

Figure 3.4 Standard deviation from the mean of luciferase activity.

Standard deviation of HeLa cells stably expressing NDST1 promoter construct treated with DMSO were in a range of ±9% of the average mean luciferase activity. Luciferase activity is expressed in relative luminescence units (l.u.).

Upon completion of the primary screen, compounds that decreased the NDST1 reporter activity were divided into two groups: strong inhibitors, which decreased the reporter activity by 75-100%, and medium inhibitors, which decreased the reporter activity by 50-75%. Compounds increasing the luminescence were considered as NDST1 enhancers, with at a ≥30% increase above the mean of

DMSO control. From the primary screen I have identified 38 compounds as potentially strong inhibitors of NDST1 mRNA transcription (Figure 3.5 and Supplementary 3.1).

Figure 3.5 A high throughput screen of 1,200 Prestwick Chemical Library compounds.

From the primary screen, 38 hits (shown in red) decreased the activity of NDST1 reporter by >75% (values >> -3 standard deviations (stdev), i.e ≥25% of the DMSO mean activity). Additionally, 7 compounds (shown in green) increased the NDST1 reporter activity by at least 30%.

The high scoring candidate inhibitors of NDST1 transcription could be divided into three groups, based on their chemical structure similarities; i) 17 compounds containing a steroid back-bone structures, ii) 10 compounds containing pyridine or imidazole ring structures, and iii) 11 compounds containing unique structures. I selected nine compounds (hits), several from each group for further analysis. In addition, two other compounds, that resulted in >50% reduction in luciferase activity, were selected (Table 3.6). The first, pyrimethamine (64% reduction), was selected because; i) it is a well-studied anti-malarial drug that crosses BBB, ii) our lab has previously shown it to act as an enzyme enhancement agent for mutant forms of β-hexosaminidase and successfully tested it in a

Phase I/II clinical trials with adult GM2 gangliosidosis patients [22], and iii) as a part of our previous work, we obtained a set of novel derivatives of the molecule from Marco Ciufolini (UBC) that we

62 could also test. The second compound, griseofulvin (58% reduction), was picked as a random control to see whether compounds inhibiting NDST1 reporter activity by 50% were generally worthy of additional testing. Another set of 20 compounds had SDs over 3, increasing the activity of NDST1 reporter by more than 30%, suggesting possible enhancers of NDST1 gene expression (Figure 3.5, hits shown in green).

Table 3.2 Potential inhibitors of NDST1 transcription.

Plate Compound Structure % Inhibition PubChem CID CAS Number Mechanism Well

02C02 R(-) Apomorphine 82.66 6852399 41372-20-7 D1 agonist hydrochloride hemihydrate

03E11 Fenbendazole 79.34 3334 43210-67-9 Microtubule formation inhibitor

04F10 Mifepristone 81.24 55245 84371-65-3 Progesterone receptor antagonist

06F07 Meclocycline 94.85 5282520 73816-42-9 Ribosomal sulfosalicylate protein synthesis inhibitor

07E03 Thiostrepton 89.02 16220061 1393-48-2 n/a

Abbreviations: n/a – data not available

63 Table 3.2 Potential inhibitors of NDST1 transcription (continue). Plate Compound Structure % Inhibition PubChem CID CAS Number Mechanism Well

07H04 Pentamidine 75.87 6604151 140-64-7 n/a

isethionate

08A10 Entacapone 86.87 5281081 130929-57-6 catechol-O- methyl transferase inhibitor

14C11 Vorinostat (SAHA) 92.38 5311 149647-78-9 histone deacetylase (HDAC) inhibitor

15D10 Anthralin 78.3 10187 1143-38-0 antipsoriatic

01D08 Pyrimethamine 64.44 4993 58-14-0 Folic acid antagonist

03G07 Griseofulvin 57.9 441140 126-07-8 Enzymatic inductor

Abbreviations: n/a – data not available

Nine compounds from the Prestwick Chemical Library inhibiting the NDST1 reporter construct by more than 75% and two compounds inhibiting the NDST1 reporter construct by more than 50% were selected for a secondary screen.

64 The ability of firefly luciferase to produce light is a commonly used technique in HTS assays.

During the in vitro reaction, firefly luciferase catalyses the formation of a luciferyl-adenylate intermediate from the D-luciferin substrate and ATP. The luciferyl-adenylate is subsequently oxidized by molecular oxygen and converted into a high-energy oxyluciferin, which emits light on return to the ground state. While luciferase based assays are becoming wildly used in gene expression studies, it should be noted that reaction of the firefly luciferase with its substrate can be inhibited by ATP- or D-luciferin-analogues [99,100]. Also, one recent study indicated that 12% of

360,864 compounds from the NIH Molecular libraries could directly bind to the firefly luciferase resulting in decline of its activity [101]. Since the compounds of the Prestwick Chemical library could also potentially interact with the components of the firefly luciferase assay, i.e. false positives, it was necessary to confirm that our identified inhibitors were specifically acting through their interaction with the NDST1 promoter fragment. Thus, I performed in vitro luciferase assays in the presence of each hit. Equal volumes of cell lysates containing firefly luciferase were incubated with

100 µM (final concentration) of the selected compounds at room temperature followed by luminescence measurements. Fifty percent of tested compounds had almost no effect (7-15% decrease) and the other fifty had moderate effect (26-42% decrease) on bioluminescence readings

(Figure 3.6). However these values represent the maximal effect, as the concentration of each drug in cell lysates (diluted with assay buffer) analyzed in the primary screen would be much less than the

100 µM that was initially added to the cell media. Therefore, we concluded that none of the tested compounds were direct inhibitors of the luciferase assay.

Figure 3.6 The effect of potential NDST1 inhibitors on firefly luciferase activity.

In order to determine an optimal drug concentration for use in normal human fibroblast and MPS patient cell lines, I investigated the dose dependant relationship of the NDST1 reporter activity with serially diluted compounds selected from the primary screen (11 hits). Each compound was three- fold serially diluted, added to the permanently expressing NDST1 reporter HeLa cell line and incubated for 16 hours. To control for possible variations in luminescence reading, I determined the total protein concentration in each sample. Commonly used colorimetric assays to determine protein concentration, e.g. Lawry or BCA, were not acceptable for our luciferase-based assays, because the luciferase substrate buffer (a firefly luciferase assay reagent) contained a high concentration of DTT

(33.3 mM), which is absolutely required for optimal catalytic activity of the firefly luciferase.

Unfortunately, the concentration of DTT in colorimetric protein assay is limited to 1 mM. Therefore,

I had to develop an alternative method for measuring total protein levels in the samples. I decided to use a Coomassie blue stain dot blot analysis. Accuracy of this method was determined by use of serially diluted BSA standards spotted on a 3 mm Whatman paper. The relationship of the protein

66 concentration and the density of each dot was determined using ImageJ application. Standard BSA curve indicated that there was a good correlation between the total BSA protein applied on Whatman paper and the calculated integrated densities in a range of 25-125 µg/mL. I determined that 2 µL of

NDST1 reporter cell lysate would be sufficient to obtain readings within the linear range for each sample (Figure 3.7).

Figure 3.7 The Coomassie dot blot analysis.

Shown in here an inverted image of the Coomassie stained BSA standards. Serially diluted BSA standards were spotted onto 3 mm Whatman paper and stained with a Coomassie blue stain. The standard curve was produced by plotting the integrated densities against concentration of BSA standards. Linear detection range for protein concentration detected between 25-125 µg/mL.

After calculation of the protein concentrations in each individually treated sample, I determined EC50

(the concentration at which each drug produced 50% of its maximal inhibitory effect on the firefly luciferase activity) for each tested compound, using log (inhibitor) vs. response equation (built-in nonlinear regression analysis) in GraphPad Prism. Dose-response curves were produced for 10 of the

11 hits, i.e. fenbendazole did not produce a curve from which an EC50 could be calculated. Also, a reproducible, but non-standard response was obtained for thiostrepton, which suggested an activation of the NDST1 transcription at lower concentration and inhibition at higher concentration

(Figure 3.8).

meclocycline

1.5 EC50=5.11

M D

1.0

e v

i Total protein (µg/ml)

a l

e Luminescence (l.u.)

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

Figure 3.8 Dose-dependent inhibition of the NDST1 reporter.

Permanent HeLa cells stably expressing the NDST1 reporter construct were treated with various concentrations of the selected compounds. After lysis and luminescence readings, total protein concentration was measured by Coomassie dot blot analysis. Data are graphed as units of the DMSO-negative control, i.e. 1= no change. Each data point represents the mean of three separate experiments.

entacapone

EC =41.67

1.5 50

M D

1.0

e Total protein (µg/ml)

v i

t Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

thiostrepton

EC50=4.37

1.5

M D

1.0 o

t Total protein (µg/ml)

e v

i Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

mifepristone

EC =11.46

1.5 50

M D

1.0

o t

Total protein (µg/ml)

v i

t Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

Figure 3.8 Dose-dependent inhibition of the NDST1 reporter (continue).

SAHA

EC50=1.2

1.5

M D

1.0

e Total protein (µg/ml)

v i

t Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

griseofulvin

EC =14.51

1.5 50

M D

1.0

e Total protein (µg/ml)

i t

a Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

anthralin

EC50=2.2

1.5

M D

1.0

v Total protein (µg/ml)

t a

l Luminescence (l.u.) e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

Figure 3.8 Dose-dependent inhibition of the NDST1 reporter (continue).

penthamidine

EC50=6.58

1.5

M D

1.0

v Total protein (µg/ml)

t a

l Luminescence (l.u.) e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

pyrimethamine

EC50=5.63

1.5

M D

1.0

o t

Total protein (µg/ml)

v i

t Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

apomorphin

EC50=32.95

1.5

M D

1.0

e Total protein (µg/ml)

v i

t Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

Figure 3.8 Dose-dependent inhibition of the NDST1 reporter (continue).

fenbendazole

1.5

M D

1.0

o t

Total protein (µg/ml)

v i

t Luminescence (l.u.)

l e

r 0.5

n U 0.0 -1 0 1 2 Log [Concentration], µM

Figure 3.8 Dose-dependent inhibition of the NDST1 reporter (continue).

In order to study the effect of the selected compounds on the activity of the endogenous NDST1 promoter, I performed a five-day treatment of normal human fibroblast cells with each compound at their calculated EC50. During the treatment, four compounds; fenbendazole, anthralin, apomorphine hydrochloride hemihydrates and penthamidine isetionate; were found to be toxic toward this cell type. The cells treated with the remaining six compounds; mifepristone, meclocycline sulfosalicylate, entacapone, vorinostat (SAHA), pyrimethamine, and griseofulvin; were harvested by scraping and lysed with 1% Triton-100 in TBS. After separating equal protein amounts of the cell lysates by SDS-PAGE, endogenous NDST1 levels were detected using a specific NDST1 antibody

(Figure 3.9). The changes in NDST1 expression were validated by densitometric analysis using

NDST1 protein levels normalized to GAPDH levels (Figure 3.10).

Figure 3.9 Effects of six compounds at their 1 x EC50 on the endogenous NDST1 expression.

The normal human fibroblast cells were treated with the selected compounds at their calculated EC50 for five days. Total cell lysates were separated by SDS-PAGE and analyzed by Western blot using a NDST1 antibody (top panel). A GAPDH antibody was used to visualize a loading control band (bottom panel).

Figure 3.10 Densitometric analysis of NDST1 protein expression compared to loading control

GAPDH after 1 x EC50 treatment.

The effect of the compounds at their EC50 on the NDST1 expression was validated by determining the relative density of each NDST1 band relative to the GAPDH loading control band visualized by Western blotting. Compared to DMSO treatment, significant reduction of NDST1 expression (>50%) was observed for enthacapone, SAHA, and pyrimethamine.

To further evaluate the effectiveness of the selected hits on the expression of NDST1, I repeated experiments using increased concentrations of the drugs. The cells were treated using compounds at their twice EC50 for five days. NDST1 expression was analyzed by Western blot (Figure 3.11). As with the previous analyses at1 x EC50, the density of each of the NDST1 bands was normalized to those of the corresponding GAPDH loading control (Figure 3.12).

Figure 3.11 Effects of six compounds at their 2 x EC50 on the endogenous NDST1 expression.

The normal human fibroblast cells were treated with selected compounds at their 2 x EC50 for 5 days. Equal amount of cell lysates was assessed by Western blot using NDST1 antibody (top panel). GAPDH was used as the loading control (bottom panel).

Figure 3.12 Densitometric analysis of NDST1 protein expression compared to loading control

GAPDH after 2 x EC50 treatment.

The effect of each compound at 2 x its EC50 on the NDST1 expression was evaluated using densitometric analysis of eachNDST1 band normalized to its corresponding GAPDH band. Compared to DMSO treatment, significant reduction of NDST1 expression (> 50%) was observed for enthacapone, mifepriston, SAHA, and pyrimethamine.

The results of treating normal human fibroblast cell line with each hit closely mimicked those obtained from the primary NDST1 reporter HTS and those from the dose-response experiments.

Unfortunately, five-day treatments with thiostrepton, anthralin, pentamidine, and apomorphin were found to be toxic to normal human cells and had to be excluded from the subsequent analysis. When cells were treated with the remaining compounds at their calculated 1 x EC50, enthacapone, SAHA, and pyrimethamine showed significant reduction (> 50%) in NDST1 expression. By contrast, there was no response observed when cells were treated with griseofulvin, our randomly selected low-

76 inhibition control. Only moderate differences in NDST1 expression were observed in cells treated with mifepriston and meclocyclin (~30%). It must be noted that results derived from the 1 x EC50 treatments might not represent the full inhibitory capacity of these drugs. As seen in 2 x EC50 treatments, mifepriston was able to decrease NDST1 expression by additional 20% bringing total reduction to 50%. The effects of enthacapone, meclocycline and pyrimethamine remained unchanged suggesting that these compounds reached their saturation point at 1 x EC50. It would be interestingly to see whether reduction of these concentrations could preserve the same inhibitory effect on NDST1 expression. An unexpected result was obtained from cells treatment with SAHA; at its 2 x EC50 (2.4 µM) SAHA reduced expression of NDST1 nearly to baseline levels. This result suggested that SAHA might be a highly potent drug having a broad inhibitory capacity over NDST1 expression.

To evaluate if the observed effect of SAHA on NDST1 expression in normal human fibroblast cells correlates with the decrease of endogenous total NDST(1-4) activity, we performed our NDST

ELISA assay (developed by Dr. Xiaolian Fan). After treatment of cells with 1.5 µM of SAHA for 5 days, cell lysates were incubated with immobilized, biotin labeled K5 (N-acetyl heparosan) or with

DMSO alone for 30 minutes at 37 C. The N-deacetylase activities of NDST1-4 were determined by an antibody specific for deacetylated form of heparosan (PAPS was not present in the assay, thus N- sulfate could not occur). The data obtained from this treatment demonstrated that SAHA was able to decrease the total activity of NDST by ~ 40% (Figure 3.13).

Figure 3.13 ELISA assay of total NDST activity in normal human fibroblast cells.

The whole lysates of cells treated with either SAHA or DMSO were assessed for total NDST activity. Treatment with 1.5 µM of SAHA resulted in a ~ 40% decrease in NDST enzyme activity.

3.3.2 The Analysis of Pyrimethamine Derivatives

The modifications to the chemical structures are often employed to improve the property of the existing compounds. As a part of our laboratory’s study of EET for late GM2 gangliosidosis, we obtained a series of pyrimethamine (PYR) derivatives. I have tested the effects of six of these derivatives on the NDST1 reporter activity. Each compound was serially diluted and used to determine a dose-response relationship, as previously described. The response of the NDST1 reporter indicated that the lowest EC50 value was produced by jts 22, ~4-fold lower than PYR

(Figure 3.14). All the other derivatives produced higher EC50 and all, but vsa 8, exhibited poorer maximal inhibitory potency. The structures of the PYR derivatives jts 22 and zjm 1-91 differ only slightly in the length of their pyrimidine side arm; i.e..methyl (jts 22), ethyl (PYR) and butyl (zjm 1-

78 91) groups (Figure 3.15). However, as can be seen from the dose-response curves, a shorter arm produces a more potent drug.

Figure 3.14 Dose-response relationship of the pyrimethamine derivatives

The serially diluted derivatives were tested on NDST1 reporter stable cell line. The depicted EC50 for pyrimethamine - 5.63 µM, jts 22 – 1.42 µM, zjm 1-91 – 49.75 µM, ksh 3-10 - 33.45 µM, edp – 33.31 µM, vsa 8 – 17.85 µM, and zjm 7-69 – 51.08 µM.

Figure 3.14 Dose-response relationship of the pyrimethamine derivatives (continue).

Figure 3.15 Derivatives of pyrimethamine

The calculated EC50 for pyrimethamine derivatives were related to their chemical structures. The derivative with the shortest pyrimidine side arm jts 22 was identified as the most potent inhibitor of the NDST1 reporter construct.

3.4 Conclusion

My objective was to develop a cell-based high throughput screening (HTS) assay to identify compounds from the 1,200 compounds of the Prestwick Chemical Library that decrease the activity of NDST1 through inhibiting the transcription of its mRNA. For this purpose, I established a luciferase reporter based assay, while others in the laboratory developed an enzyme-linked immuno sorbent assay (ELISA) to directly evaluate NDST enzymatic activity. The genomic DNA fragment corresponding to the NDST1 promoter was identified and cloned in front of the firefly luciferase reporter gene. I used this construct in 96-well based HTS of the Prestwick Chemical Library. From the primary screen I have identified 38 “hits” that reduced NDST1 reporter activity by more than

75%. Two compounds were selected from those that produced good dose-dependent response curves in cells expressing the luciferase reporter for validation in normal human fibroblast cells. The effects of these compounds on endogenous NDST1 expression were analyzed by Western blotting and their effects on total NDST enzyme activity by our ELISA assay. One of the most potent inhibitors of the

NDST1 promoter reporter construct, SAHA, was able to decrease the endogenous NDST1expression levels almost to baseline levels at 2.4 µM and to reduce the total NDST activity by 40% at 1.5 µM.

Further studies are needed to determine the effect of the reduced NDST1 activity on the levels of HS stored in patient cells.

In summary, I demonstrated that the 96-well NDST1 promoter firefly luciferase reporter assay could be used to identify small molecules, with a previous history of use in humans, which also could act as potent inhibitors of the endogenous NDST1 expression. Furthermore, our established ELISA based enzyme assay has proven to be useful in confirming the efficacy of these drugs on the total

NDST activity in normal human cell lines. From the analysis of 11 out of 38 potential inhibitors of

83 NDST1, two compounds, pyrimethamine and SAHA, could be considered for further testing in MPS patient cell lines and in MPS animal models.

3.4.1 Property of pyrimethamine (2,4 diamino 5-(4-chlorophenyl)- 6-ethylpyrimidine)

Pyrimethamine (PYR) was originally developed as a dihydrofolate reductase inhibitor, which is used for treatment of parasitic diseases, including chloroquine-resistant malaria and toxoplasmosis [102].

PYR is an orally administered drug, with a well-studied pharmacokinetic profile [103]. Studies have shown that 12-26% of serum levels cross the BBB [104]. However, it has also been reported in mice that the PYR concentration in the brain reaches 3-time that found in the plasma. PYR also has a long half-life in humans of ~85 hrs [105]. In 2006 our lab identified PYR as a competitive inhibitor of β- hexosaminidase and a potential pharmacological chaperone for late onset GM2 gangliosidosis [22].

We also completed a phase I/II clinical trial of the drug for this purpose [106]. We found that a dose of 25-50 mg/day was well tolerated and resulted in a plasma level of 4-6 µM (PYR had an EC50 in my assay of 5.6 µM). PYR has also been identified as being able to produce a dose-dependent reduction in super oxide dismutase 1 (SOD1) protein levels in human cells in vitro, and it was tested in a phase I trial as a treatment for amyotrophic lateral sclerosis (ALS) [107]. Finally PYR has been shown to be an inhibitor of a signal transducer and activator of transcription 3 (STAT3) factor, and is suggested as a potential treatment for adult polycystic kidney disease [108].

Of the set of PYR derivative prepared for us by Marco Ciufolini (UBC), I found two that had better maximal inhibitory activity and/or lower EC50s. Interestingly these two derivatives were much poorer inhibitors of hexosaminidase than the parent compound, indicating that in these two instances, PYR is not acting through a common mechanism.

84 3.4.2 Property of SAHA (Vorinostat, suberoylanilidehydroxamic acid)

SAHA is a histone deacetylase inhibitor (HDAC inhibitor). Oral SAHA was reported to have linear pharmacokinetics in humans from 200 to 600 mg/day and could be safely administered chronically.

It had an apparent half-life ranging from 91 to 127 minutes and was 43% oral bioavailability [109].

SAHA was tested in the R6/2 mouse model of Huntington's disease and found to be brain penetrant, with a brain:plasma ratio of 0.12 [110]. In another study of patients with advanced cutaneous T-cell lymphoma, it was reported that at steady state in the fed-state, oral administration of multiple 400- mg/d doses of vorinostat (SAHA) resulted in a mean area under the concentration–time curve and maximum concentration and a median maximum time of 6.0 ± 2.0µM* hour, 1.2 ± 0.53µM, and 4

(0.5–14) hours, respectively. The median number of days on protocol treatment was 97.5 (range, 2–

480+ days). Seventeen (19.8%) patients received treatment for >24 weeks and eight (9.3%) patients were treated beyond 1 year. Adverse events resulted in discontinuation of therapy in eight patients

(9.3%). An additional nine patients (10.5%) required a dose reduction [111]. It is interesting that

HDAC inhibitors generally increase the transcription of genes by preventing the tight coiling of

DNA around histones. In the case of the NDST1 promoter SAHA acts to inhibit its transcription.

This seeming contradiction could point to SAHA possessing other previously unknown biological activities; e.g., similar to PYR, which now has 4 independent targets (see above).

3.5 Future Directions

The positive results obtained from the analysis of the first set of selected compounds (11 out of 38 hits) suggest that the remaining untested NDST1 transcription inhibitors could also be effective toward inhibiting the expression of endogenous NDST1.The remaining compounds from the primary screen will be analyzed by the dose-response experiments and by the Western blotting as previously

85 described. My overall results suggest that NST1 expression is regulated by several different mechanisms. In case of PYR it is most likely achieved through the STAT pathway [108], whereas in case of SAHA the expression is likely inhibited by a mechanism yet to be identified. To confirm that

SAHA is not acting through its known HDAC inhibitor activity, several other HDAC inhibitors should be tested for activity against the NDST1 reporter assay system I developed. Similarly other compounds that have shown activity towards the STAT3 pathway should be tested.

To explore the inhibitory mechanism of NDST1 by PYR and SAHA, I will first determine if PYR acts on NDST1 through STAT3 pathway. I will treat normal human fibroblast cells for 5 days with

PYR, as previously described in the methods, and confirm whether STAT3 is down-regulated by

PYR [108] by using Stat3 and Phospho-Stat3 antibodies (Cell Signaling). Then, I will use any of the commercially available STAT3 inhibitors, for example Stat3 Inhibitor III, WP1066 (Santa Cruz

Biothechnology Inc.) which is a cell permeable STAT3 and JAK2 inhibitor. I will treat the cells with this inhibitor for 5 days to determine if it can mimic the effect of PYR on NDST1 expression. To determine the mechanism of SAHA, I will assess if its activity is structurally or functionally related.

Presently, there are 31 HDAC inhibitors commercially available (Supplementary 3.2), and some of these inhibitors are structurally or biochemically similar to SAHA. I will test the response of NDST1 expression in normal human fibroblast cells after treatment with selected SAHA analogs. Another possible function of SAHA is the activation of the transcription of other proteins that negatively regulates the expression of NDST1. The NDST1 promoter has a long stretch of CpG islands, which acts as a transcriptional regulatory mechanism. It is possible that SAHA activates the DNA methyltransferase that hypermethylate CpG island resulting in the silencing of the NDST1 transcription. To explore this possibility, I will treat cells with SAHA in the presence of the potent

DNA methyltransferase inhibitor 5-azacytidine (Sigma). If the promoter of NDST1 is methylated,

86 the presence of 5-azacytidine will restore the expression of NDST1. Another experiment to confirm that PYR and SAHA target different pathways would be to perform an additive effect experiment. I will treat the normal human fibroblast with combination of SAHA and PYR at 50% of their calculated EC50 to see whether these drugs act additively. If the effect is proven to be additive, it would mean that compounds act on a different target. If this proves to be true it might be beneficial to treat patients with both in order to reduce the dosage of each individual drug and thus, potentially decrease their known side effects.

References

1. Campos D, Monaga M (2012) Mucopolysaccharidosis type I: current knowledge on its pathophysiological mechanisms. Metab Brain Dis 27: 121-129.

2. Moore D, Connock MJ, Wraith E, Lavery C (2008) The prevalence of and survival in Mucopolysaccharidosis I: Hurler, Hurler-Scheie and Scheie syndromes in the UK. Orphanet J Rare Dis 3: 24.

3. Froissart R, Da Silva IM, Maire I (2007) Mucopolysaccharidosis type II: an update on mutation spectrum. Acta Paediatr Suppl 96: 71-77.

4. Beck M (2011) Mucopolysaccharidosis Type II (Hunter Syndrome): clinical picture and treatment. Curr Pharm Biotechnol 12: 861-866.

5. Valstar MJ, Ruijter GJ, van Diggelen OP, Poorthuis BJ, Wijburg FA (2008) Sanfilippo syndrome: a mini-review. J Inherit Metab Dis 31: 240-252.

6. Matsushita Y, Kuroiwa Y (2000) [Morquio syndrome, beta-galactosidase deficiency]. Ryoikibetsu Shokogun Shirizu: 466-467.

7. Tomatsu S, Montano AM, Oikawa H, Smith M, Barrera L, et al. (2011) Mucopolysaccharidosis type IVA (Morquio A disease): clinical review and current treatment. Curr Pharm Biotechnol 12: 931-945.

8. Valayannopoulos V, Nicely H, Harmatz P, Turbeville S (2010) Mucopolysaccharidosis VI. Orphanet J Rare Dis 5: 5.

9. Csoka AB, Frost GI, Stern R (2001) The six hyaluronidase-like genes in the human and mouse genomes. Matrix Biol 20: 499-508.

10. Huang YL, Li SY, Zhao XY, Liu HS, Ou XB, et al. (2011) [Mucopolysaccharidosis VII: report of a case and review of the literature]. Zhonghua Er Ke Za Zhi 49: 455-458.

11. Triggs-Raine B, Salo TJ, Zhang H, Wicklow BA, Natowicz MR (1999) Mutations in HYAL1, a member of a tandemly distributed multigene family encoding disparate hyaluronidase activities, cause a newly described lysosomal disorder, mucopolysaccharidosis IX. Proc Natl Acad Sci U S A 96: 6296-6300.

12. Imundo L, Leduc CA, Guha S, Brown M, Perino G, et al. (2011) A complete deficiency of Hyaluronoglucosaminidase 1 (HYAL1) presenting as familial juvenile idiopathic arthritis. J Inherit Metab Dis 34: 1013-1022.

13. Valayannopoulos V, Wijburg FA (2011) Therapy for the mucopolysaccharidoses. Rheumatology (Oxford) 50 Suppl 5: v49-59.

88 14. Aldenhoven M, Boelens JJ, de Koning TJ (2008) The clinical outcome of Hurler syndrome after stem cell transplantation. Biol Blood Marrow Transplant 14: 485-498.

15. Wraith JE (2009) Enzyme replacement therapy for the management of the mucopolysaccharidoses. Int J Clin Pharmacol Ther 47 Suppl 1: S63-65.

16. de Ruijter J, Valstar MJ, Wijburg FA (2011) Mucopolysaccharidosis type III (Sanfilippo Syndrome): emerging treatment strategies. Curr Pharm Biotechnol 12: 923-930.

17. Fan JQ, Ishii S, Asano N, Suzuki Y (1999) Accelerated transport and maturation of lysosomal alpha-galactosidase A in Fabry lymphoblasts by an enzyme inhibitor. Nat Med 5: 112-115.

18. Yu Z, Sawkar AR, Kelly JW (2007) Pharmacologic chaperoning as a strategy to treat Gaucher disease. FEBS J 274: 4944-4950.

19. Okumiya T, Kroos MA, Vliet LV, Takeuchi H, Van der Ploeg AT, et al. (2007) Chemical chaperones improve transport and enhance stability of mutant alpha-glucosidases in glycogen storage disease type II. Mol Genet Metab 90: 49-57.

20. Parenti G, Zuppaldi A, Gabriela Pittis M, Rosaria Tuzzi M, Annunziata I, et al. (2007) Pharmacological enhancement of mutated alpha-glucosidase activity in fibroblasts from patients with Pompe disease. Mol Ther 15: 508-514.

21. Matsuda J, Suzuki O, Oshima A, Yamamoto Y, Noguchi A, et al. (2003) Chemical chaperone therapy for brain pathology in G(M1)-gangliosidosis. Proc Natl Acad Sci U S A 100: 15912- 15917.

22. Maegawa GH, Tropak M, Buttner J, Stockley T, Kok F, et al. (2007) Pyrimethamine as a potential pharmacological chaperone for late-onset forms of GM2 gangliosidosis. J Biol Chem 282: 9150-9161.

23. Tropak MB, Reid SP, Guiral M, Withers SG, Mahuran D (2004) Pharmacological enhancement of beta-hexosaminidase activity in fibroblasts from adult Tay-Sachs and Sandhoff Patients. J Biol Chem 279: 13478-13487.

24. Jeyakumar M, Butters TD, Cortina-Borja M, Hunnam V, Proia RL, et al. (1999) Delayed symptom onset and increased life expectancy in Sandhoff disease mice treated with N- butyldeoxynojirimycin. Proc Natl Acad Sci U S A 96: 6388-6393.

25. Platt FM, Neises GR, Reinkensmeier G, Townsend MJ, Perry VH, et al. (1997) Prevention of lysosomal storage in Tay-Sachs mice treated with N-butyldeoxynojirimycin. Science 276: 428-431.

26. Maegawa GH, van Giersbergen PL, Yang S, Banwell B, Morgan CP, et al. (2009) Pharmacokinetics, safety and tolerability of miglustat in the treatment of pediatric patients with GM2 gangliosidosis. Mol Genet Metab 97: 284-291.

89 27. McEachern KA, Fung J, Komarnitsky S, Siegel CS, Chuang WL, et al. (2007) A specific and potent inhibitor of glucosylceramide synthase for substrate inhibition therapy of Gaucher disease. Mol Genet Metab 91: 259-267.

28. Roberts AL, Rees MH, Klebe S, Fletcher JM, Byers S (2007) Improvement in behaviour after substrate deprivation therapy with rhodamine B in a mouse model of MPS IIIA. Mol Genet Metab 92: 115-121.

29. Roberts AL, Thomas BJ, Wilkinson AS, Fletcher JM, Byers S (2006) Inhibition of glycosaminoglycan synthesis using rhodamine B in a mouse model of mucopolysaccharidosis type IIIA. Pediatr Res 60: 309-314.

30. Dire DJ, Wilkinson JA (1987) Acute exposure to rhodamine B. J Toxicol Clin Toxicol 25: 603- 607.

31. Friso A, Tomanin R, Salvalaio M, Scarpa M (2010) Genistein reduces glycosaminoglycan levels in a mouse model of mucopolysaccharidosis type II. Br J Pharmacol 159: 1082-1091.

32. Jakobkiewicz-Banecka J, Wegrzyn A, Wegrzyn G (2007) Substrate deprivation therapy: a new hope for patients suffering from neuronopathic forms of inherited lysosomal storage diseases. J Appl Genet 48: 383-388.

33. Malinowska M, Wilkinson FL, Bennett W, Langford-Smith KJ, O'Leary HA, et al. (2009) Genistein reduces lysosomal storage in peripheral tissues of mucopolysaccharide IIIB mice. Mol Genet Metab 98: 235-242.

34. Malinowska M, Wilkinson FL, Langford-Smith KJ, Langford-Smith A, Brown JR, et al. (2010) Genistein improves neuropathology and corrects behaviour in a mouse model of neurodegenerative metabolic disease. PLoS One 5: e14192.

35. Snare MJ, Treloar, F.E., Ghiggino K.P., and Thistlethwaite (1982) The photophysics of rhodamine B. Photochem 18: 335-346.

36. Yamaguchi M (2012) Nutritional factors and bone homeostasis: synergistic effect with zinc and genistein in osteogenesis. Mol Cell Biochem 366: 201-221.

37. Bame KJ (2001) Heparanases: endoglycosidases that degrade heparan sulfate proteoglycans. Glycobiology 11: 91R-98R.

38. Freeman C, Hopwood J (1992) Lysosomal degradation of heparin and heparan sulphate. Adv Exp Med Biol 313: 121-134.

39. Kowalewski B, Lamanna WC, Lawrence R, Damme M, Stroobants S, et al. (2012) Arylsulfatase G inactivation causes loss of heparan sulfate 3-O-sulfatase activity and mucopolysaccharidosis in mice. Proc Natl Acad Sci U S A 109: 10310-10315.

40. McCarty DM, DiRosario J, Gulaid K, Killedar S, Oosterhof A, et al. (2011) Differential distribution of heparan sulfate glycoforms and elevated expression of heparan sulfate

90 biosynthetic enzyme genes in the brain of mucopolysaccharidosis IIIB mice. Metab Brain Dis 26: 9-19.

41. Holley RJ, Deligny A, Wei W, Watson HA, Ninonuevo MR, et al. (2011) Mucopolysaccharidosis type I, unique structure of accumulated heparan sulfate and increased N-sulfotransferase activity in mice lacking alpha-l-iduronidase. J Biol Chem 286: 37515- 37524.

42. Yan D, Lin X (2009) Shaping morphogen gradients by proteoglycans. Cold Spring Harb Perspect Biol 1: a002493.

43. Jemth P, Gianni S (2007) PDZ domains: folding and binding. Biochemistry 46: 8701-8708.

44. Bernfield M, Gotte M, Park PW, Reizes O, Fitzgerald ML, et al. (1999) Functions of cell surface heparan sulfate proteoglycans. Annu Rev Biochem 68: 729-777.

45. Fransson LA (2003) Glypicans. Int J Biochem Cell Biol 35: 125-129.

46. Veugelers M, De Cat B, Ceulemans H, Bruystens AM, Coomans C, et al. (1999) Glypican-6, a new member of the glypican family of cell surface heparan sulfate proteoglycans. J Biol Chem 274: 26968-26977.

47. Melrose J, Roughley P, Knox S, Smith S, Lord M, et al. (2006) The structure, location, and function of perlecan, a prominent pericellular proteoglycan of fetal, postnatal, and mature hyaline cartilages. J Biol Chem 281: 36905-36914.

48. Knox SM, Whitelock JM (2006) Perlecan: how does one molecule do so many things? Cell Mol Life Sci 63: 2435-2445.

49. Winzen U, Cole GJ, Halfter W (2003) Agrin is a chimeric proteoglycan with the attachment sites for heparan sulfate/chondroitin sulfate located in two multiple serine-glycine clusters. J Biol Chem 278: 30106-30114.

50. Marneros AG, Olsen BR (2005) Physiological role of collagen XVIII and endostatin. FASEB J 19: 716-728.

51. Hayashi K, Madri JA, Yurchenco PD (1992) Endothelial cells interact with the core protein of basement membrane perlecan through beta 1 and beta 3 integrins: an adhesion modulated by glycosaminoglycan. J Cell Biol 119: 945-959.

52. Yayon A, Klagsbrun M, Esko JD, Leder P, Ornitz DM (1991) Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64: 841-848.

53. Pye DA, Vives RR, Hyde P, Gallagher JT (2000) Regulation of FGF-1 mitogenic activity by heparan sulfate oligosaccharides is dependent on specific structural features: differential requirements for the modulation of FGF-1 and FGF-2. Glycobiology 10: 1183-1192.

91 54. Ye S, Luo Y, Lu W, Jones RB, Linhardt RJ, et al. (2001) Structural basis for interaction of FGF- 1, FGF-2, and FGF-7 with different heparan sulfate motifs. Biochemistry 40: 14429-14439.

55. Ornitz DM (2000) FGFs, heparan sulfate and FGFRs: complex interactions essential for development. Bioessays 22: 108-112.

56. Robinson CJ, Mulloy B, Gallagher JT, Stringer SE (2006) VEGF165-binding sites within heparan sulfate encompass two highly sulfated domains and can be liberated by K5 lyase. J Biol Chem 281: 1731-1740.

57. Zhao W, McCallum SA, Xiao Z, Zhang F, Linhardt RJ (2012) Binding affinities of vascular endothelial growth factor (VEGF) for heparin-derived oligosaccharides. Biosci Rep 32: 71- 81.

58. Lyon M, Deakin JA, Mizuno K, Nakamura T, Gallagher JT (1994) Interaction of hepatocyte growth factor with heparan sulfate. Elucidation of the major heparan sulfate structural determinants. J Biol Chem 269: 11216-11223.

59. Zhang L (2010) Glycosaminoglycan (GAG) biosynthesis and GAG-binding proteins. Prog Mol Biol Transl Sci 93: 1-17.

60. Lortat-Jacob H, Garrone P, Banchereau J, Grimaud JA (1997) Human interleukin 4 is a glycosaminoglycan-binding protein. Cytokine 9: 101-105.

61. Lipscombe RJ, Nakhoul AM, Sanderson CJ, Coombe DR (1998) Interleukin-5 binds to heparin/heparan sulfate. A model for an interaction with extracellular matrix. J Leukoc Biol 63: 342-350.

62. Lortat-Jacob H, Grosdidier A, Imberty A (2002) Structural diversity of heparan sulfate binding domains in chemokines. Proc Natl Acad Sci U S A 99: 1229-1234.

63. Lortat-Jacob H, Turnbull JE, Grimaud JA (1995) Molecular organization of the interferon gamma-binding domain in heparan sulphate. Biochem J 310 ( Pt 2): 497-505.

64. Sadir R, Forest E, Lortat-Jacob H (1998) The heparan sulfate binding sequence of interferon- gamma increased the on rate of the interferon-gamma-interferon-gamma receptor complex formation. J Biol Chem 273: 10919-10925.

65. Mahley RW, Ji ZS (1999) Remnant lipoprotein metabolism: key pathways involving cell-surface heparan sulfate proteoglycans and apolipoprotein E. J Lipid Res 40: 1-16.

66. Fuki II, Iozzo RV, Williams KJ (2000) Perlecan heparan sulfate proteoglycan. A novel receptor that mediates a distinct pathway for ligand catabolism. J Biol Chem 275: 31554.

67. Spillmann D (2001) Heparan sulfate: anchor for viral intruders? Biochimie 83: 811-817.

68. Belting M (2003) Heparan sulfate proteoglycan as a plasma membrane carrier. Trends Biochem Sci 28: 145-151.

92 69. Bishop JR, Schuksz M, Esko JD (2007) Heparan sulphate proteoglycans fine-tune mammalian physiology. Nature 446: 1030-1037.

70. Morimoto-Tomita M, Uchimura K, Werb Z, Hemmerich S, Rosen SD (2002) Cloning and characterization of two extracellular heparin-degrading endosulfatases in mice and humans. J Biol Chem 277: 49175-49185.

71. Aikawa J, Grobe K, Tsujimoto M, Esko JD (2001) Multiple isozymes of heparan sulfate/heparin GlcNAc N-deacetylase/GlcN N-sulfotransferase. Structure and activity of the fourth member, NDST4. J Biol Chem 276: 5876-5882.

72. Humphries DE, Lanciotti J, Karlinsky JB (1998) cDNA cloning, genomic organization and chromosomal localization of human heparan glucosaminyl N-deacetylase/N- sulphotransferase-2. Biochem J 332 ( Pt 2): 303-307.

73. Dixon J, Loftus SK, Gladwin AJ, Scambler PJ, Wasmuth JJ, et al. (1995) Cloning of the human heparan sulfate-N-deacetylase/N-sulfotransferase gene from the Treacher Collins syndrome candidate region at 5q32-q33.1. Genomics 26: 239-244.

74. Aikawa J, Esko JD (1999) Molecular cloning and expression of a third member of the heparan sulfate/heparin GlcNAc N-deacetylase/ N-sulfotransferase family. J Biol Chem 274: 2690- 2695.

75. Eriksson I, Sandback D, Ek B, Lindahl U, Kjellen L (1994) cDNA cloning and sequencing of mouse mastocytoma glucosaminyl N-deacetylase/N-sulfotransferase, an enzyme involved in the biosynthesis of heparin. J Biol Chem 269: 10438-10443.

76. Grobe K, Esko JD (2002) Regulated translation of heparan sulfate N-acetylglucosamine N- deacetylase/n-sulfotransferase isozymes by structured 5'-untranslated regions and internal ribosome entry sites. J Biol Chem 277: 30699-30706.

77. Kozak M (1991) Structural features in eukaryotic mRNAs that modulate the initiation of translation. J Biol Chem 266: 19867-19870.

78. Mitchell P, Tollervey D (2001) mRNA turnover. Curr Opin Cell Biol 13: 320-325.

79. Bushati N, Cohen SM (2007) microRNA functions. Annu Rev Cell Dev Biol 23: 175-205.

80. Gallie DR (1998) A tale of two termini: a functional interaction between the termini of an mRNA is a prerequisite for efficient translation initiation. Gene 216: 1-11.

81. Varani G (2001) Delivering messages from the 3' end. Proc Natl Acad Sci U S A 98: 4288-4289.

82. Wickens M, Anderson P, Jackson RJ (1997) Life and death in the cytoplasm: messages from the 3' end. Curr Opin Genet Dev 7: 220-232.

93 83. Morii E, Ogihara H, Oboki K, Sawa C, Sakuma T, et al. (2001) Inhibitory effect of the mi transcription factor encoded by the mutant mi allele on GA binding protein-mediated transcript expression in mouse mast cells. Blood 97: 3032-3039.

84. Grobe K, Ledin J, Ringvall M, Holmborn K, Forsberg E, et al. (2002) Heparan sulfate and development: differential roles of the N-acetylglucosamine N-deacetylase/N-sulfotransferase isozymes. Biochim Biophys Acta 1573: 209-215.

85. Kakuta Y, Sueyoshi T, Negishi M, Pedersen LC (1999) Crystal structure of the sulfotransferase domain of human heparan sulfate N-deacetylase/ N-sulfotransferase 1. J Biol Chem 274: 10673-10676.

86. Duncan MB, Liu M, Fox C, Liu J (2006) Characterization of the N-deacetylase domain from the heparan sulfate N-deacetylase/N-sulfotransferase 2. Biochem Biophys Res Commun 339: 1232-1237.

87. Presto J, Thuveson M, Carlsson P, Busse M, Wilen M, et al. (2008) Heparan sulfate biosynthesis enzymes EXT1 and EXT2 affect NDST1 expression and heparan sulfate sulfation. Proc Natl Acad Sci U S A 105: 4751-4756.

88. Lin X, Wei G, Shi Z, Dryer L, Esko JD, et al. (2000) Disruption of gastrulation and heparan sulfate biosynthesis in EXT1-deficient mice. Dev Biol 224: 299-311.

89. Stickens D, Zak BM, Rougier N, Esko JD, Werb Z (2005) Mice deficient in Ext2 lack heparan sulfate and develop exostoses. Development 132: 5055-5068.

90. Kaidonis X, Liaw WC, Roberts AD, Ly M, Anson D, et al. (2010) Gene silencing of EXTL2 and EXTL3 as a substrate deprivation therapy for heparan sulphate storing mucopolysaccharidoses. Eur J Hum Genet 18: 194-199.

91. Lamanna WC, Lawrence R, Sarrazin S, Lameda-Diaz C, Gordts PL, et al. (2012) A genetic model of substrate reduction therapy for mucopolysaccharidosis. J Biol Chem 287: 36283- 36290.

92. Ford-Perriss M, Guimond SE, Greferath U, Kita M, Grobe K, et al. (2002) Variant heparan sulfates synthesized in developing mouse brain differentially regulate FGF signaling. Glycobiology 12: 721-727.

93. Forsberg E, Kjellen L (2001) Heparan sulfate: lessons from knockout mice. J Clin Invest 108: 175-180.

94. Ringvall M, Ledin J, Holmborn K, van Kuppevelt T, Ellin F, et al. (2000) Defective heparan sulfate biosynthesis and neonatal lethality in mice lacking N-deacetylase/N-sulfotransferase- 1. J Biol Chem 275: 25926-25930.

95. Fan G, Xiao L, Cheng L, Wang X, Sun B, et al. (2000) Targeted disruption of NDST-1 gene leads to pulmonary hypoplasia and neonatal respiratory distress in mice. FEBS Lett 467: 7- 11.

94 96 Garner OB, Yamaguchi Y, Esko JD, Videm V (2008) Small changes in lymphocyte development and activation in mice through tissue-specific alteration of heparan sulphate. Immunology 125: 420-429

97. Li JP, Gong F, Hagner-McWhirter A, Forsberg E, Abrink M, et al. (2003) Targeted disruption of a murine glucuronyl C5-epimerase gene results in heparan sulfate lacking L-iduronic acid and in neonatal lethality. J Biol Chem 278: 28363-28366.

98. Qu X, Carbe C, Tao C, Powers A, Lawrence R, et al. (2011) Lacrimal gland development and Fgf10-Fgfr2b signaling are controlled by 2-O- and 6-O-sulfated heparan sulfate. J Biol Chem 286: 14435-14444.

99. Denburg JL, Lee RT, McElroy WD (1969) Substrate-binding properties of firefly luciferase. I. Luciferin-binding site. Arch Biochem Biophys 134: 381-394.

100. Lee RT, Denburg JL, McElroy WD (1970) Substrate-binding properties of firefly luciferase. II. ATP-binding site. Arch Biochem Biophys 141: 38-52.

101. Thorne N, Shen M, Lea WA, Simeonov A, Lovell S, et al. (2012) Firefly luciferase in chemical biology: a compendium of inhibitors, mechanistic evaluation of chemotypes, and suggested use as a reporter. Chem Biol 19: 1060-1072.

102. Leport C, Chene G, Morlat P, Luft BJ, Rousseau F, et al. (1996) Pyrimethamine for primary prophylaxis of toxoplasmic encephalitis in patients with human immunodeficiency virus infection: a double-blind, randomized trial. ANRS 005-ACTG 154 Group Members. Agence Nationale de Recherche sur le SIDA. AIDS Clinical Trial Group. J Infect Dis 173: 91-97.

103. Buxton I (2006) Pharmacokinetics and pharmacodynamics: the dynamics of drug absorption, distribution, action, and elimination. Goodman & Gilman's The Pharmacological Basis of Therapeutics 11 ed New York: McGraw-Hill Books.

104. Weiss LM, Harris C, Berger M, Tanowitz HB, Wittner M (1988) Pyrimethamine concentrations in serum and cerebrospinal fluid during treatment of acute Toxoplasma encephalitis in patients with AIDS. J Infect Dis 157: 580-583.

105. Cavallito JC, Nichol CA, Brenckman WD, Jr., Deangelis RL, Stickney DR, et al. (1978) Lipid- soluble inhibitors of dihydrofolate reductase. I. Kinetics, tissue distribution, and extent of metabolism of pyrimethamine, metoprine, and etoprine in the rat, dog, and man. Drug Metab Dispos 6: 329-337.

106. Clarke JT, Mahuran DJ, Sathe S, Kolodny EH, Rigat BA, et al. (2011) An open-label Phase I/II clinical trial of pyrimethamine for the treatment of patients affected with chronic GM2 gangliosidosis (Tay-Sachs or Sandhoff variants). Mol Genet Metab 102: 6-12.

107. Lange DJ, Andersen PM, Remanan R, Marklund S, Benjamin D (2013) Pyrimethamine decreases levels of SOD1 in leukocytes and cerebrospinal fluid of ALS patients: a phase I pilot study. Amyotroph Lateral Scler Frontotemporal Degener 14: 199-204.

95 108. Takakura A, Nelson EA, Haque N, Humphreys BD, Zandi-Nejad K, et al. (2011) Pyrimethamine inhibits adult polycystic kidney disease by modulating STAT signaling pathways. Hum Mol Genet 20: 4143-4154.

109. Kelly WK, O'Connor OA, Krug LM, Chiao JH, Heaney M, et al. (2005) Phase I study of an oral histone deacetylase inhibitor, suberoylanilide hydroxamic acid, in patients with advanced cancer. J Clin Oncol 23: 3923-3931.

110. Mielcarek M, Benn CL, Franklin SA, Smith DL, Woodman B, et al. (2011) SAHA decreases HDAC 2 and 4 levels in vivo and improves molecular phenotypes in the R6/2 mouse model of Huntington's disease. PLoS One 6: e27746.

111. Mann BS, Johnson JR, Cohen MH, Justice R, Pazdur R. (2007) FDA approval summary: vorinostat for treatment of advanced primary cutaneous T-cell lymphoma. Oncol 10:1247- 52.

96 Appendices

Supplementary 3.1 Compounds inhibiting the NDST1 promoter firefly reporter construct activity more than 75%

No Plate Well Compound Structure % Inhibition

Chiral

R(-) Apomorphine hydrochloride N CH 1 02C02 N 3 82.66 CH H 3 HO hemihydrate H HO

HO HO

O Chiral

O H3C 2 02E09 Spironolactone 88.19 CH3 H O H H

O S CH3

H3C CH3 Chiral Cl HO CH N H3 H 3 02G06 Chlortetracycline hydrochloride OH 84.98

NH2 OH OH O OH O O

S 4 03E11 Fenbendazole N O 79.34

CH3 N N O

HC Chiral

CH3 OH

5 04B04 Norethindrone H 81.5 H

HH O

CH3 Chiral N CH OH 3 H3C H CH3 6 04F10 Mifepristone H 81.24

H O

N 7 04F11 Diperodon hydrochloride O O 97.00 O O N

No Plate Well Compound Structure % Inhibition

8 05C04 Desipramine hydrochloride N 76.96

CH3

CH 3 N

9 05C11 Clozapine N 77.38 N

Cl N

OHChiral O

CH3 10 06D03 Corticosterone HO 80.55 CH3 H

H H O

O OChiral

CH 11 06D07 Digitoxigenin 3 94.22 H3C H

H OH HO H

O O Chiral OH CH3

H CH3

H OH O 12 06D08 Digoxin H 95.27 O

OH OH H3C O OH OH

O O O CH3

CH3

O Chiral CH3

13 06D11 Epiandrosterone CH3 H 79.08

H H

HO H

OHChiral CH3 14 06E02 Estradiol-17 beta H 81.64

H H HO

CH H C 3 Chiral Cl CH OH3 N 2 H 15 06F07 Meclocyclinesulfosalicylate OH 94.85

NH2 OH OH O OH O O

No Plate Well Compound Structure % Inhibition

H C S 3 Chiral O N HO N N O HO CH3 HO N CH N 3S O CHO3 S N N S N

OH H C O O 16 07E03 Thiostrepton 3 N 89.02

H3C O N N N CH3

H2C S N OH O N O O H3C N H2C N N O N O CH3 H3C O

H2N CH2

OH 17 07G02 Ciclopirox ethanolamine O N 80.22

CH3

HN NH 2

18 07H04 Pentamidineisethionate O 75.87

H2N

O O + N O N CH 19 08A10 Entacapone 3 86.87 HO H C N 3 OH

O O Chiral OH CH3

CH3 H

H OH H C O O 20 09B07 Lanatoside C 3 H 95.92 H3C O O OH

H3C O O OH O O HO O CH3 HO OH O OH

O CH3Chiral

CH3

21 09D02 Dydrogesterone CH3 H 79.54

H H O

H3C O CH HO 3 CH H 22 10C04 Medrysone 3 89.00 H H O

CH3

H C 3 OH CH

23 10F04 Norgestrel-(-)-D H H 75.33

H H O

No Plate Well Compound Structure % Inhibition

Cl O O CH O 3 CH HO 3 CH 24 10G02 Clobetasol propionate 3 83.79

CH3 H

F H O

CH3 Chiral

O O O 25 11F06 Beclomethasonedipropionate CH 78.05 3 O HO CH3 CH3 CH H 3 O Cl H O

26 11F10 Fluvastatin sodium salt OH OH O 81.83

O N

CH3 H3C

O O OH CH 27 12A04 Digoxigenin 3 95.78

CH3 H

H OH HO H

O Chiral CH3 28 12F08 Dehydroisoandosterone 3-acetate 92.74 CH3 H O H H

H3C O

O Chiral OH CH3 29 12H08 Deoxycorticosterone 80.47 CH3 H

H H O

N 30 13B08 Zardaverine F N 77.54 F O O CH3

OHChiral CH3 31 13C03 Nandrolone H H 81.65

H H O

100

No Plate Well Compound Structure % Inhibition

O O

CH3 CH 32 13C07 Proscillaridin A 3 96.32 CH3 H HO O H OH HO O H OH

O CH3 Chiral

CH3 O CH3 33 13H04 Nomegestrol acetate H H O 75.00 H H O

CH3

O Chiral H C O 3 O CH O O CH 34 14A05 Prednicarbate HO 3 3 78.14 CH3 H O

H H O

O 35 14C11 Vorinostat (SAHA) N OH 92.38 N O

HO CH 3 N H3C O O HO 36 15F03 Atorvastatin N 87.31 HO

HO CHChiral CH3

37 15E05 Ethinylestradiol H 90.7

H H HO

OH O OH 38 15D10 Anthralin 78.3

101

Supplementary 3.2 HDAC Inhibitors

Name Chemical formula Function SAHA inhibits class I and class II HDACs at nanomolar concentrations and arrests cell growth in a wide variety of transformed cells in culture at 2.5-5.0 µM. This compound efficiently suppressed MES-SA cell growth at a low dosage (3 µM) already after 24 hours treatment. Decrease of cell SAHA survival was even more pronounced after prolonged treatment and reached 9% and 2% after 48 and 72 hours of treatment, respectively. Colony forming capability of MES-SA cells treated with 3 µM vorinostat for 24 and 48 hours was significantly diminished and blocked after 72 hours. Inhibitor of HDAC 1,2,3,6 and 8 (DrugBank). Entinostat (MS-275) preferentially inhibits HDAC1 (IC50=300nM) over HDAC3 (IC50=8µM) and has no inhibitory activity towards HDAC8 (IC50>100µM). MS-275 induces cyclin-dependent kinase inhibitor 1A Entinostat (MS-275) (p21/CIP1/WAF1), slowing cell growth, differentiation, and tumor development in vivo. Recent studies suggest that MS-275 may be particularly useful as an antineoplastic agent when combined with other drugs, like adriamycin. LBH589, a novel hydroxamate analog HDAC inhibitor, has been shown to induce acetylation of histone H3 and H4, increase p21 levels, disrupt the chaperone function of hsp90, and induce cell-cycle G1 phase Panobinostat (LBH-589) accumulation and apoptosis of K562 cells and acute leukemia MV4-11 cells. The anti-tumor effect by Panobinostat was also demonstrated in multiple myeloma, NSCLC as well as castrate-resistant prostate cancer cell lines. An antifungal antibiotic with cytostatic and differentiating properties in mammalian cell culture. TSA inhibited proliferation of eight breast carcinoma cell lines with mean ± SD IC50 of 124.4 ± 120.4 nm (range, Trichostatin A (TSA) 26.4–308.1 nm). HDAC inhibitory activity of TSA was similar in all cell lines with mean ± SD IC50 of 2.4 ± 0.5 nm (range, 1.5–2.9 nm), and TSA treatment resulted in pronouncing histone H4 hyperacetylation.

102

Name Chemical formula Function PXD101 (Belinostat) is a novel hydroxamate-type inhibitor of histone deacetylase activity. It inhibits HDAC activity in HeLa cell extracts with an IC50 of 27 nM. PXD101 is cytotoxic in vitro in a number of tumor cell lines Belinostat with IC50s in the range 0.2-3.4 µM as determined by a clonogenic assay and induces apoptosis. This hdac inhibitor is currently under phase I/II testing in lymphoma, ovarian cancer and other solid tumors. MGCD0103 inhibits class 1 isoforms of HDAC, specifically HDAC1 (IC50=0.15 µM), HDAC2 (IC50=0.29 µM) and HDAC3 (IC50=1.66 µM), which may result in epigenetic changes in tumor cells and so tumor cell death. It exhibited potent and selective antiproliferative activities against a Mocetinostat broad spectrum of human cancer cell lines (IC50 from 0.09-20 µM) in vitro, and HDAC inhibitory activity was required for these effects. In vivo, MGCD0103 significantly inhibited growth of human tumor xenografts in nude mice in a dose-dependent manner and the antitumor activity correlated with induction of histone acetylation in tumors. MC1568 is a potent selective class II (IIa) histone deacetylas (HDAC II) inhibitor (IC50 of a maize deacetylase HD2 at 22.0 µM). MC1568 had no or MC1568 a weak effect on the class I HDACs. MC1568 shows no inhibitory activity against HDAC1 but was able to inhibit HDAC4.

Rocilinostat Rocilinostat (ACY-1215) is a selective HDAC6 inhibitor with IC50 of 5 nM

M344 is an inhibitor of histone deacetylases, inhibiting maize HDAC (IC50 = 100 nM) as well as human HDAC1 (IC50 = 46 nM). It shows a three-fold selectivity for inhibition of HDAC6 over HDAC1. This compound enhances M344 the sensitivity of human squamous carcinoma cells to radiation and promotes cell cycle arrest and apoptosis in human endometrial cancer and ovarian cancer cells (ED50 = 2.3 µM).

103

Name Chemical formula Function

PCI-34051 is a potent and specific histone deacetylase 8 (HDAC8) inhibitor with an IC50 of 10 nM. PCI-34051 inhibits HDAC6, HDAC1 and HDAC10 with IC50 of 2.9, 4 and 13μM. PCI-34051 possesses promising potency for HDAC8 with a Ki of 10 nM. PCI-34051 has high selectivity (approximately PCI-34051 fivefold) for HDAC8 relative to the other class I HDACs including HDAC1. PCI-34051 reveals greater than 200-fold selectivity over HDAC1 and HDAC6, and greater than 1000-fold selectivity over HDAC2, HDAC3 and HDAC10. PCI-34051 inhibits ovarian tumor line OVCAR-3 with a GI50 of 6 μM and 15% cell death.

Tacedinaline An anti-cancer drug which inhibits HDAC1 with IC50 of 0.57 μM.

Tubastatin A is a potent and selective HDAC6 inhibitor. Demonstrates 1093-fold selectivity over HDAC1 (IC50 values of 15 nM for HDAC6 vs 16.4 µM for HDAC1). Tubastatin A was substantially more selective than the known HDAC6 inhibitor Tubacin at all isozymes except HDAC8. In Tubastatin A addition, it displayed over 1000-fold selectivity against all HDAC isoforms excluding HDAC8, where it displayed 54-fold selectivity. This compound displayed dose-dependent protection against HCA induced neuronal cell death starting at 5 µM with near complete protection at 10 µM.

AR-42 is a novel HDAC inhibitor with an IC50 of 0.61 μM for acute lymphoblastic leukemia (697) cell linesAR-42 has in vitro and in vivo AR-42 efficacy at tolerable doses. In additon, AR-42 promoted hyperacetylation of H3, H4, and alpha-tubulin, and up-regulation of p21. Down-regulation of Kit occurred after AR-42 treatment via inhibition of Kit transcription.

104

Name Chemical formula Function

ITF2357 is an inhibitor of HDAC activity (class I and class II HDAC) but also is an anti-inflammatory agent in vitro and in vivo. Enzymatically, Givinostat ITF2357 inhibits different histone deacetylases from maize, and biochemical analysis revealed that, similar to SAHA, ITF2357 hyperacetylates histones in human PBMCs.

SB939 (Pracinostat) is a novel potent and orally active histone deacetylase inhibitor with an IC50 of 52 nM for HDAC1 and a GI50 of 0.56 µM for Colo205. Its IC50 for HCT-116 colon cancer cell line and HL-60 acute myeloid leukemia cell line is 0.48 μM and 70 nM, respectively. In vitro, SB939 (Pracinostat) inhibits class I, II, and IV HDACs, with no effects on other zinc binding enzymes, and shows significant antiproliferative activity Prasinostat against a wide variety of tumor cell lines. It has very favorable pharmacokinetic properties after oral dosing in mice, with >4-fold increased bioavailability and 3.3-fold increased half-life over suberoylanilidehydroxamic acid (SAHA). In contrast to SAHA, SB939 (Pracinostat) accumulates in tumor tissue and induces a sustained inhibition of histone acetylation in tumor tissue. Droxinostat is a novel HDAC inhibitor. It selectively inhibits HDAC3, HDAC6, and HDAC8 with an IC50 of 16.9, 2.47 and 1.46 µM/L, Droxinostat respectively, but did not inhibit HDAC1, HDAC2, HDAC4, HDAC5, HDAC7, HDAC9, and HDAC10, with an IC50 of >20 µM/L.

Largazole is derived from cyanobacteria that grow on coral reefs, which may have potentials as an anti-cancer drug. Largazole is a potent histone deacetylase inhibitor. Largazole has its potent antiproliferative activity Largazole against a number of cancer cell-lines including MDA-MB-231 mammary cells (GI50 7.7 nM), U2OS fibroblastic osteosarcoma cells (GI50 55 nM), HT29 colon cells (GI50 12 nM), and IMR-32 neuroblastoma cells (GI50 16 nM).

105

Name Chemical formula Function CUDC-101 is a potent multitargeted HDAC, EGFR and HER2 inhibitor with IC50 of 4.4, 2.4, and 15.7 nM, respectively. CUDC-101 has novel structure incorporating HDAC inhibitory functionality into the pharmacophore of the EGFR and HER2 inhibitors. In most tumor cell lines tested, CUDC-101 exhibits efficient antiproliferative activity with greater CUDC-101 potency than vorinostat (SAHA), erlotinib, lapatinib, and combinations of vorinostat/erlotinib and vorinostat/lapatinib. In vivo, CUDC-101 promotes tumor regression or inhibition in various cancer xenograft models including nonsmall cell lung cancer (NSCLC), liver, breast, head and neck, colon, and pancreatic cancer.

JNJ-26481585 (Quisinostat) is an HDAC inhibitor for HDAC1, HDAC2, Quisinostat HDAC4, HDAC10 and HDAC11 with IC50 of 0.11 nM, 0.33 nM, 0.64 nM, 0.46 nM and 0.37 nM, respectively.

LAQ824 (NVP-LAQ824, Dacinostat) (a cinnamichydroxamic acid), a histone deacetylase inhibitor, inhibited in vitro enzymatic activities(IC50= Dacinostat 0.032uM)and transcriptionally activated the p21 promoter in reporter gene assays(IC50= 0.3 μM).

PCI-24781 (CRA-024781) is a novel broad spectrum HDAC inhibitor targeting HDAC1, HDAC2, HDAC3, HDAC6, HDAC8 and HDAC10 with Ki of 7 nM, 19 nM, 8.2 nM, 17 nM, 280 nM, 24 nM, respectively. PCI- 24781 is a broad-spectrum phenyl hydroxamic acid HDAC inhibitor PCI-24781 currently being evaluated in phase I clinical trials in patients with neoplastic disease. The substance is a specific inhibitor of multiple HDAC isoforms (HDAC2, HDAC3/SMRT, HDAC6, HDAC8, HDAC10) that potently inhibits tumor growth in vivo with acceptable toxicity.

106

Name Chemical formula Function

BML-210 is a small molecule inhibitor of HDAC with an IC50 value of 30 µM when tested in HeLa cell nuclear extracts using 200 µM acetylated BML-210 fluorometric substrate. This compound also inhibits the deacetylation of the transcription factor FOXO3 by mammalian SIRT1 in cells oxidatively stressed by hydrogen peroxide.

Scriptaid is a novel histone deacetylase inhibitor that belongs to an existing class of hydroxamic acid-containing HDAC inhibitors. Scriptaid has a general property of transcriptional facilitation that applies to stably integrated or transiently transfected exogenous constructs, to promoters derived from viruses or an endogenous gene, to multiple reporter genes, and to different cell lines. Scriptaid does not interfere with the ability of a Scriptaid reporter construct to measure the positive activation of a transcription factor in response to a known signal transduction stimulus. In relation to other members of its class, the optimal concentration of scriptaid (6–8 µM) is similar to those reported for CBHA (4 µM) and SuberoylanilideHydroxamic Acid (2 µM), higher than Trichostatin A (1 µM), and much lower than those reported for EMBA (400 µM) and HMBA (5000 µM). NSC3852 has cell differentiation and antiproliferative activity in human breast cancer cells in tissue culture and antitumor activity in mice bearing P388 and L1210 leukemic cells. NSC3852 inhibited histone deacetylase (HDAC) activity in vitro and caused DNA damage and apoptosis in MCF-7 NSC3852 cells, consistent with their differentiation and antiproliferative activities. This compound could inhibit HDAC activity by 80% at 190 µM, it also inhibited an unrelated protease enzyme assay by 40%, indicating some nonspecific activity at this concentration.

107

Name Chemical formula Function

NCH-51 is a novel non-hydroxamic acid histone deacetylase (HDAC) inhibitor that could inhibit the cell growth of a variety of lymphoid malignant cells through apoptosis induction, more effectively than SAHA. NCH-51 Activation of caspase-3, -8 and -9, but not -7 was detected after the treatment with NCH-51. Gene expression profiles showed that NCH-51 and SAHA similarly upregulated p21 and downregulated anti-apoptotic molecules including survivin, bcl-w and c-FLIP. HNHA (N-hydroxy-7-(2-naphthylthio) heptanomide) is a novel HDAC inhibitor that shows better pharmacological properties than a known HNHA HDACIs present in the human fibrosarcoma cell - SAHA (SuberoylanilideHydroxamic Acid).

BML-281 is a potent and selective HDAC3 and HDAC6 inhibitor. Potent inhibitor of the proliferation of a variety of pancreatic cancer cell lines BML-281 (IC50 in the 0.1-0.6 µM range). This compound, for example, is active against both the Mia Paca-2 and Panc04.03 cell lines at the 100 nM level and at the 200-300 nM level against HupT3.

CBHA (m-carboxycinnamic acid bis-hydroxamide) is a member of a recently synthesized family of hybrid polar compounds that have been shown to be inhibitors of HDAC and potent inducers of transformed cell CBHA growth arrest and terminal differentiation at micromolar (LD50 range, 1–4 µM) concentrations. The effective dose of CBHA that inhibited 50% clonal growth (ED50) of the endometrial cancer cell lines was calculated and ranged between 1.8 x 10-6M and 2.5 x 10-6M for CBHA.

108

Name Chemical formula Function

The sirtuins (SIRTs) are a family of NAD+-dependent histone deacetylases involved in gene regulation. Salermide is an inhibitor of SIRT1 and SIRT2, causing tumor-specific apoptotic cell death. In MOLT4 leukemia cells, Salermide salermide causes 90% apoptosis within 72 hours (IC50 ~ 20 µM) by reactivating proapototic genes that are repressed by SIRT1. Salermide is a stronger Sirtuin inhibitor than sirtinol.

Pimelicdiphenylamide 106 is a class I HDAC inhibitor, demonstrating no activity against class II HDACs. This substance is a slow, tight-binding inhibitor of HDACs 1, 2, and 3, although inhibition for these enzymes Pimelic Diphenylamide occurs through different mechanisms. Pimelicdiphenylamide and related benzamide HDAC inhibitors may have therapeutic value in Friedrich's ataxia and Huntington's disease, in part due to their low animal toxicity.

APHA Compound 8 is a potent novel hydroxamate HDAC inhibitor, is in the same structural class as SAHA, an investigational drug standard currently in clinical trial. The IC50 for mouse HDAC1 is 0.5 µM. In a APHA Compound 8 separate study, significant inhibition was found in the proliferation of MEL cells at 24 µM and differentiation, assayed as the accumulation of hemoglobin, at 5 µM.

Compounds in red are selected for further analysis Sources: http://www.exchemistry.com/histone-deacetylase-inhibitors/SB939.html http://www.inhibitor2.com/hdac-histone-deacetylase-inhibitors.html