Selection of Transthyretin Amyloid Inhibitors

Irina Iakovleva

Department of Medical Biochemistry and Biophysics Umeå 2016

Responsible publisher under swedish law: the Dean of the Medical Faculty This work is protected by the Swedish Copyright Legislation (Act 1960:729) ISBN: 978-91-7601-528-5 ISSN: 0346-6612 New series nr: 1826 Cover: TTR tetramer © Irina Iakovleva Elektronisk version tillgänglig på http://umu.diva-portal.org/ Tryck/Printed by: Print & Media, Umeå University Umeå, Sweden 2016

Till alla som jag älskar…

Table of Contents

Table of Contents i Abstract iii List of papers v Abbreviations vii Introduction 1 1.1 Amyloidosis 1 1.1.1 Protein folding and misfolding 1 1.1.2 Amyloid proteins 1 1.2.2. Suggested mechanisms of amyloid formation 3 2. Transthyretin, from gene to protein 4 2.1 History 4 2.2 Expression and metabolism 4 2.3 Structural features of TTR 5 2.4 The function and physiological significance of TTR 7 2.4.1 Transport of thyroxine and retinol 7 2.4.2 Protective role in Alzheimer’s disease 9 2.4.3 The proteolytic activity of TTR 10 2.4.4. Other functions of TTR 10 3. TTR as an amyloidogenic protein 12 3.1 What triggers TTR aggregation? 12 3.2 TTR-mediated cytotoxicity 15 3.3 Sporadic and hereditary amyloidosis: the clinical spectrum 16 3.3.1 Wild-type TTR amyloidosis 16 3.3.2 Hereditary TTR amyloidosis 17 Familial amyloid polyneuropathy 17 Familial amyloid cardiomyopathy (FAC) 18 3.4 Therapeutic strategies for TTR amyloidosis 18 3.4.1. Liver transplantation 19 3.4.2. RNA suppression therapies 20 ASO therapy 20 siRNA therapy 20 3.4.3. Stabilizing the structure of TTR tetramer 21 Tafamidis 22 Diflunisal 22 3.4.4. Clearance of the amyloid matrix 23 Aims 24 4. Methods 25 4.1 Monitoring drug selectivity in plasma (urea-mediated denaturation plasma assay) 25 5.1 The presence of plasma components strongly influences ligand selectivity 28

i 5.2 Enthalpy is a better predictor of selectivity 30 5.3 Using the method to estimate the suitable therapeutic dosage 33 5.4 Luteolin suppresses the TTR-mediated pathological effect in cells and in an animal model of FAP 34 5.5 Enzymatic modification inactivates the effect of luteolin on TTR 35 5.6 Substitution of 7-OH with 7-Cl and 7-MeO increases the resistance of luteolin to glucuronidation 37 5.7 Substitution of 7-OH decreases binding affinity and selectivity in plasma 38 5.8 TBBPA effectively prevents TTR aggregation and binds TTR with extremely high selectivity in human plasma 40 5.9 TTBPA from a drug perspective 42 Concluding remarks 44 Acknowledgements 46 References 49

ii Abstract

Amyloidosis is a group of clinical disorders caused by the aggregation of specific proteins into abnormal extracellular deposits. Today, 31 different proteins have been linked to amyloid diseases including transthyretin- related amyloidosis (ATTR). ATTR occurs through the aggregation of either wild-type plasma protein transthyretin (TTR) or a mutated form. TTR is a homotetramer that under normal circumstances functions as a carrier of thyroxine and retinol binding protein. The aggregation cascade requires dissociation of the tetramer into monomers, and preventing this dissociation represents a potential mode of intervention. Interestingly, small molecules, referred as kinetic stabilizers, can bind to TTR’s thyroxine-binding site (TBS) and such molecules are currently being used as a therapeutic approach to impair tetramer dissociation.

The efficacy of TTR stabilization is directly correlated to the binding affinity of the ligand to TBS. However, the binding of the ligand to TTR in vivo can be affected by other plasma components resulting in poor efficacy. Thus, the selectivity of ligands is an important parameter. We have designed an assay where the ability to stabilize TTR can be directly evaluated in plasma and we have investigated the stabilizing effect of nine potential TTR binders (Paper I). The results, surprisingly, revealed that the binding affinity of molecules has a poor correlation to its selectivity. However, the nature of protein-ligand complex formation can also be described by enthalpic (∆H) and entropic (∆S) energy contributions. ∆H represents the change in chemical bonds and frequently requires a higher order of orientation compared to the ∆S component, which mainly represents the hydrophobic effect via the exclusion of water. We hypothesized that ligands possessing high ΔH in binding to their co-partner would also be more specific in a complex environment such as plasma. By applying a thermodynamic analysis using isothermal titration calorimetry, we found that the selectivity in plasma correlates well with the ∆H contribution and might, therefore, be a better predictor for selectivity.

Luteolin was found to be a highly selective stabilizer of TTR and was investigated further (Paper II). The ligand displayed a significant rescuing effect in both cell culture and animal models. However, luteolin undergoes rapid enzymatic degradation in the liver and this impairs its use as a potential therapeutic drug. To attempt to circumvent this issue, we modified the most exposed hydroxyl group thus rendering the molecule inert towards glucuronidation (Paper III). The substitutions resulted in higher stability in

iii the face of hepatic degradation molecules, but they also affected the selectivity in a negative manner.

The screening for new TTR stabilizers resulted in the discovery of tetrabromobisphenol A, which displayed a very high selectivity (Paper IV). This study also included a comparison with the drug Vyndaqel™ which currently is in clinically use, and showed how the dosage could be altered to acquire a better level of saturation and possibly also a better clinical effect.

Taken together we present new molecules with the ability to stabilize TTR, and these can serve as scaffolds for the design of new drugs. We present a method to measure the efficacy of a TTR-stabilizing drugs in a complex matrix and as well as a way to adjust the dosage of existing drugs. We also show that the selectivity of a drug is affected by the relative proportion of ∆H and ∆S, and this is of interest for drug design in general.

iv List of papers

This thesis is based on the following papers

Iakovleva I, Brannstrom K, Nilsson L, Gharibyan AL, Begum A, Anan I, Walfridsson M, Sauer-Eriksson AE, Olofsson A. Enthalpic Forces Correlate with the Selectivity of Transthyretin-Stabilizing Ligands in Human Plasma. J Med Chem, 2015. 58(16): p. 6507-15.

II.

Iakovleva I, Begum A, Pokrzywa M, Walfridsson M, Sauer-Eriksson AE, Olofsson A. The flavonoid luteolin, but not luteolin-7-O-glucoside, prevents a transthyretin mediated toxic response. PLoS One, 2015. 10(5): p. e0128222.

III.

Nilsson L, Larsson A, Begum A, Iakovleva I, Carlsson M, Brannstrom K, Sauer-Eriksson AE, Olofsson A. Modifications of the 7-Hydroxyl Group of the Transthyretin Ligand Luteolin Provide Mechanistic Insights into Its Binding Properties and High Plasma Specificity. PLoS One, 2016. 11(4): p. e0153112.

IV.

Iakovleva I, Begum A, Brännström K, Wijsekera A, Nilsson L, Zhang J, Andersson PL, Sauer-Eriksson AE, Olofsson A. Tetrabromobisphenol A Is an Efficient Stabilizer of the Transthyretin Tetramer. PLoS One, 2016. 11(4): p. e0153529.

v Contributions to other papers, not included in this thesis

Brannstrom K, Lindhagen-Persson M, Gharibyan AL, Iakovleva I, Vestling M, Sellin ME, Brannstrom T, Morozova-Roche L, Forsgren L, Olofsson A. A generic method for design of oligomer-specific antibodies. PLoS One, 2014. 9(3): p. e90857.

vi Abbreviations

Aβ amyloid-β peptide AβPP Aβ precursor protein ACL aceclofenac AD Alzheimer’s disease API apigenin ASO antisense oligonucleotides ATTR transthyretin-related amyloidosis ATTRwt wild-type ATTR Cm midpoint for unfolding CSF cerebrospinal fluid ∆H enthalpy DIC diclofenac DIF diflunisal ∆S entropy EFSA European Food Safety Authority ER endoplasmic reticulum FAC familial amyloid cardiomyopathy FAB familial amyloid polyneuropathy GEM gemfibrozil HBP halogen-binding pocket HSA human serum albumin

IC50 50% inhibition concentration ITC isothermal titration calorimetry

KD binding affinity LUT luteolin MEA meclofenamic acid MST microscale thermophoresis NIF niflumic acid NSAID non-steroidal anti-inflammatory drug PD Parkinson’s disease RAGE receptor advanced glycation end RBP retinol binding protein retinol vitamin A siRNAs small interfering RNAs

vii TBBPA Tetrabromobisphenol A TBG T4-binding globulin TBS T4 binding site TTR transthyretin TOA tolfenamic acid TUDCA taurousodeoxycholic acid T4 thyroxine T3 triiodothyronine 7-Cl-Lut chloroluteolin 7-Met-Lut methoxyluteolin

viii

Introduction

1.1 Amyloidosis

1.1.1 Protein folding and misfolding

Proteins are involved in virtually all processes in living organisms. They possess different structures and, consequently, have different purposes and functions in an organism. Each protein is composed of one or more polypeptides, which are chains of amino acids joined together by peptide bonds. Once synthesized, a polypeptide chain adopts a unique conformation as the result of its gene-encoded amino acid sequence. The diffusional search for the most stable protein conformation is assisted by a large number of relatively weak interactions, including hydrogen bonds, dispersion forces, electrostatic interactions, and hydrophobic forces. These interactions ensure that the protein adopts a native fold with the lowest free energy in a specific environment.

After being produced, proteins undergo quality control by chaperones, folding enzymes, and metabolites. The role of these molecules is to enhance the properly folded state of other proteins and to ensure that misfolded proteins are sent to proteasomes for degradation [1]. Protein quality control plays an important role in the maintenance of protein homeostasis, or proteostasis, and is crucial for normal cell function. A decline in proteostatic control can lead to an imbalance between the soluble form of a certain protein and its aggregated form. This can contribute to the formation abnormal extracellular aggregates called amyloids, which is an underlying cause in several severe diseases.

1.1.2 Amyloid proteins

The term amyloid was first introduced by the German scientist Rudolph Virchow in 1854. The term emerged when Virchow misinterpreted the results of iodine staining of abnormal structures in the cerebral corpora amylacea. He concluded that the structures were made of cellulose and he named them after the Latin word amylum, which means starch-like. It took five years of debate and extensive investigations before it was confirmed that a ‘mass’ of amyloid is made up of proteins and not of carbohydrates, but the name remained unchanged [2].

Today, amyloid is used to describe the group of proteins that undergo amyloidosis to form insoluble extracellular depositions that lead to clinical

disorders. More than 30 different unrelated amyloid proteins have been identified and linked to approximately 50 diseases, including Alzheimer’s disease (AD), Parkinson’s disease (PD), and transthyretin-related amyloidosis (ATTR). Numerous different proteins can form amyloids, but all amyloids share a common β sheet core structure with the polypeptide chains running perpendicular to the long axis of the fibril.

The formation of amyloid is a complicated process that is initiated by the structural rearrangement of the soluble native state of a protein/peptide into a β-sheet conformation. Amyloid can generally form from either polypeptides that intrinsically have no defined structure [3-5] or from partially unfolded proteins [6, 7].

Polypeptides can exist in different native states. Some peptides or proteins possess no inherent structure under normal physiological conditions. Many of these peptides and proteins are involved in the most common amyloid diseases, for example, the amyloid-β peptide (Aβ) in AD [3, 4] and the α- synuclein in PD [5]. However, the majority of these naturally unfolded polypeptides is not prone to aggregation and is instead involved in many key cellular processes [7]. Thus the natural absence of structure in a protein is not necessarily correlated with the formation of amyloid.

The other suggested starting point for amyloid formation is partially unfolded proteins [6, 8]. In some cases, the proteins must undergo conformational changes in order to yield an amyloidogenic intermediate that is capable of self-assembling into amyloid. Transthyretin (TTR) is an example of such a protein, and amyloid formation by TTR requires the dissociation of the native tetrameric structure into natively folded monomers [9-11]. Once the TTR monomer is formed, it has to undergo partial denaturation in order to initiate the aggregation cascade [6].

Several diverse factors, including mutations, post-translational modifications, aging, stress, and disease can disrupt the balance in proteostasis and lead to aggregation. Aging is one of the most significant risk factors for the development of amyloidosis [12-14] due to increased protein oxidation and modifications that enhance the aggregation of protein that – together with deficient stress-responsive signaling – can significantly destabilize proteostasis and increase the accumulation of misfolded proteins [13]. Because amyloid diseases do not occur in all people as they age, other factors have to be considered. In addition, the presence of mutations within amyloid proteins, e.g. the various point mutations in TTR [15], is another factor that can predispose for the aggregation of protein and accelerate the onset of amyloidosis.

1.2.2. Suggested mechanisms of amyloid formation

Several mechanisms have been suggested for the aggregation process. Nucleation-dependent polymerization is the most accepted scenario of amyloid formation [16]. This mechanism resembles the process of crystal growth where the formation of a nucleus is the rate-limiting step. The nucleus formation is initiated by the association of two or more monomeric species. This event is unfavorable. Once the critical concentration of nuclei is achieved and the activation energy barrier is reached, the nuclei can then be used as templates for the sequential addition of monomers, ultimately forming a fibrillar structure. In vitro, the rate-limiting step of nucleus formation is characterized by a lag-phase, which is followed by a rapid growth phase corresponding to fibril formation.

The other proposed mechanism is nucleated conformational conversion [17]. In contrast to nucleation-dependent polymerization, the aggregation process is initiated by rapid oligomerization into micelle-like assemblies followed by a slow conversion into amyloid fibrils. The conversion of specific size oligomers into β-sheet structures is the rate-limiting step and is characterized by a high activation energy barrier. Both of these pathways have been proposed for Aβ aggregation [17, 18].

Downhill polymerization [19] starts after the conversion of the native structure into a partially unfolded intermediate. This mechanism does not require the formation of nuclei, and the association of unfolded species into aggregates is a favorable process due to the relatively low activation energy barrier during the formation of oligomers. This mechanism has been proposed for TTR amyloid formation in which the dissociation of the tetrameric structure is the rate-limiting step. Once the amyloidogenic monomer is formed, the process of downhill polymerization leads to rapid amyloid aggregation.

Despite many hypotheses, the mechanism of amyloid formation remains unclear due to the heterogeneity of the molecules involved and to the complexity of the aggregation steps. Several intermediates are formed during the amyloid formation process, including dimers, oligomers, and protofibrils, and the wide diversity of structures and morphologies makes mechanistic studies particularly difficult.

2. Transthyretin, from gene to protein

2.1 History

TTR was first discovered in cerebrospinal fluid (CSF) in 1942 [20], and later it was found in serum where it was identified as component X [21]. The protein was initially called prealbumin due its electrophoretic migration ahead of albumin. The binding of TTR to thyroxine was discovered in the 1950s, but subsequent studies showed that TTR is not the principle transporter of thyroxine in serum [22]. However, the high expression levels of TTR in the choroid plexus and the high concentration of TTR protein in CSF, which was discovered in 1980s, suggested that TTR is a major transporter of thyroxine in the brain [23-25]. The involvement of TTR in the transport of vitamin A (retinol) in complex together with retinol binding protein (RBP) was observed in the late 1960s [26]. In 1981, the protein was renamed by the International Union of Biochemists to transthyretin due to its primary functions as a transporter of thyroxine and retinol. The first crystallographic studies and structural analysis of TTR were reported in 1971 [27], and three years later the analysis of the amino acids sequences in the protein was completed [28]. In the 1980s, the TTR gene composition and its location on human chromosome 18 were identified [29]. These findings led to the discovery of several variants of TTR, and today more than 120 different amyloidogenic point mutations have been identified [30].

2.2 Expression and metabolism

TTR is a highly conserved protein found in several vertebrate species, including human, mouse, sheep, rat, fish, rabbit, lizard, and chicken [31]. The liver is the major source of TTR in the plasma [32] and it produces 90% of the plasma TTR in humans, which occurs at a concentration ranging from 200 mg/L to 400 mg/L (5-7 μM) [33-35]. TTR levels in human plasma change as one ages with lower TTR levels in healthy newborns compared to adults [36, 37] and decreasing TTR concentrations in people after 50 years of age. Interestingly, men tend to have higher TTR levels than women. Reduced TTR levels in plasma can also be observed during acute and chronic inflammation [38, 39], in patients with hepatic [33] and malignant diseases [40], and during preeclampsia [41]. Many studies have suggested that TTR is a good marker of energy malnutrition due to its decreased serum levels and decreased hepatic expression during protein/energy deprivation [42, 43].

The epithelial cells of the choroid plexus are the source of TTR production in the CSF [44]. TTR is the main protein produced by the choroid plexus, and it makes up approximately 25% of the total CSF protein content [45]. Its role is to distribute the thyroxine hormone (T4) throughout the CSF [46]. Other sources of TTR are the retinal pigment epithelium [47], the pancreatic islet of Langerhans [48, 49], and, to a smaller extent, the heart, skeletal muscles, and spleen [23], but the purpose of this expression is unclear. Recently, TTR expression and secretion have been reported in the human placenta [50].

In humans the half-life of TTR is approximately two days [51]. The catabolic pathways of TTR are poorly understood, but based on investigations in rats it is known that the majority of TTR degradation occurs in the liver (35–40%). The muscles and skin are responsible for approximately 25% of the total body TTR catabolism, and 6% occurs in the kidneys. The highest rates of degradation have been detected in the liver and kidneys [52]. The uptake of TTR by the liver and kidneys occurs through a receptor-mediated mechanism. The presence of natural ligands T4 and RBP in complex with TTR affects the hepatic uptake, resulting in lower uptake of TTR/RBP complexes (70% decrease) and higher uptake of TTR-T4 complexes (only 20% decrease) [53]. The renal internalization of TTR is mediated by the endocytic receptor megalin, and compared to the liver no difference was observed in the uptake of TTR by the kidneys for different TTR/ligand complexes [54].

2.3 Structural features of TTR

TTR circulates in plasma as a 55 kDa tetramer composed of four identical non-covalently attached subunits consisting of 127 amino acid residues each (Fig. 1A). A subunit (or monomer) is composed of eight antiparallel β- strands (A-H) (Fig. 1B) that are arranged into two four-stranded sheets and one α-helix situated on β-strand E [27, 28]. The association of monomers into a dimer is achieved by antiparallel hydrophobic interactions between equivalent sheets of the monomers, and the tetrameric structure is accomplished by interactions between loops on strands A and B and G and H of the two dimers (Fig. 1B) [27, 28]. The monomer-monomer interactions and dimer-dimer interactions are crucial for TTR’s stability.

The protein has two hormone-binding sites that are structurally identical and are buried deep inside the tetramer on two distinct dimer-dimer interfaces [55]. TTR also contains four binding sites for RBP on its outer surface [56, 57]. The hydrophobic binding pockets of TTR are divided into

two parts: a small inner subsite and a larger outer subsite (Fig. 1C) [55, 58]. These sites are made of three symmetrical hydrophobic depressions referred to as halogen-binding pockets (HBPs) due to their binding to the iodine atoms of T4 [59]. The outer subsite is composed of positively charged Lys15 and negatively charged Glu54 residues, which complement the zwitterionic structure of T4 [58, 60]. The two hormone-binding sites of TTR possess negative cooperativity [61], and thus virtually the protein is only able to bind one ligand at a time. In vitro, TTR can only bind two out of four RBPs due to steric hindrance. However, in vivo, only one TTR-RBP complex is formed due to the low concentration of RBP in plasma [57]. The binding of RBP and T4 to TTR has a stabilizing effect on the tetrameric structure.

Figure 1. TTR structure. (A) TTR tetramer (PDB ID: 4DEW). The red rectangles show the hormone-binding sites (B) TTR dimer with the antiparallel β-strands

denoted A–H (PDB ID: 1F41). (C) The close-up view of the T4 binding site. The pocket is shown together with the natural ligand T4 (PDB ID: 2ROX).

2.4 The function and physiological significance of TTR

The precise function of TTR is unknown, but it is believed that its main task is to transport the T4 hormone and retinol (also known as vitamin A) in complex with RBP. In addition to its transport function, TTR possesses proteolytic and chaperone-like activity [62-64], which might play an important role in the prevention of AD [65, 66]. Moreover, the protein is important in thermoregulation, it acts as an indicator of malnutrition, and it is involved in other physiological processes that will be described briefly below.

2.4.1 Transport of thyroxine and retinol

The thyroid hormones: thyroxine (T4), biologically active triiodothyronine (T3), and biologically inactive reverse T3 (rT3) are synthesized by the thyroid glands and are essential for brain development and function, maintaining metabolic balance, and controlling many other vital processes in the human body. T4, the inactive form of thyroid hormone, is secreted by the thyroid glands and is distributed to its target tissues in plasma by T4-binding globulin (TBG), TTR, and serum albumin. The properties of transported proteins are indicated in Table. 1.

In human plasma, TTR is not the main transporter of T4, and it is only responsible for transporting 10–15 % of the total thyroid hormone fraction. This means that only about 1% of the TTR proteins in the plasma are occupied with transporting T4. Due to negative cooperativity, TTR binds only one hormone molecule at time [67]. TTR has a low affinity for T3 and can only bind the inactive form of the thyroid hormone, T4.

The importance of TTR in the transport of T4 to target tissues is still a matter of debate. The evaluations of TTR-null mice revealed that despite an absence of TTR the animals still remained euthyroid [68], suggesting that TTR does not have a crucial role in T4 delivery to tissues. On the other hand, TTR transports 80% of the T4 in the CSF [69]. It also appears that TTR is significant in the delivery of the maternal T4 to the developing fetus [70]. Based on these different observations, it is clear that the role of TTR as a transporter of T4 requires further study.

Table 1. Properties of thyroxine-binding proteins in human serum a.

a Human serum albumin, HSA; transthyretin, TTR; and T4-binding globulin, TBG.

Retinol is also transported by TTR as a complex with RBP [71]. Vitamin A is an essential molecule for several physiological processes, and it is involved in vision, the immune system, development, reproduction, and bone growth. It has also debated if retinol has a protective effect in AD and whether retinol supplementation might be a possible treatment [72]. Retinol is transported in complex with RPB [26], which is mainly produced in the liver and secreted into the plasma as a holo-complex together with retinol [73]. In the plasma, approximately 50% of TTR proteins are involved in the transport of retinol-RBP [74]. The association of TTR with retinol-RBP is essential for stabilizing the holo-complex during the transport and to prevent RBP from being filtered out and degraded in the kidneys [75, 76]. The binding of RBP to TTR does not interfere with the association of T4 [73].

The role of TTR in the distribution of RBP-retinol was investigated in TTR- null mice. No symptoms of vitamin A deficiency were observed [77], suggesting that TTR’s role in transporting RBP-retinol is not essential. The plasma levels of both RBP and retinol were greatly reduced in TTR-null mice, whereas no difference in total retinol was detected between tissue levels [78]. The decrease in RBP and retinol in plasma was explained by

filtering of the complex by the kidneys, which is where TTR most likely plays its main role.

2.4.2 Protective role in Alzheimer’s disease

The first association between TTR and AD was reported in 1982 where the authors showed that TTR is a common constituent of plaques from AD patients [79]. However, because these results could not be confirmed in larger sets of AD brains, the interest in this topic waned until 1993, when a new report revealed that CSF could inhibit Aβ fibril formation in vitro [80]. In 1994, TTR in human CSF was first suggested as a major Aβ ‘sequester’ or binding protein by Schwarzman and et al. [81]. This opened up a new approach in the amyloid research field, and many in vivo and in vitro studies have since investigated the interaction between TTR and Aβ.

Several in vitro studies have confirmed that interactions between Aβ and TTR interfere with Aβ aggregation [81-83], and it appears that monomeric TTR better suppresses Aβ fibril formation than tetrameric TTR [84, 85]. Moreover, TTR has a higher affinity for the Aβ1-42 monomer than the Aβ1- 40 monomer [85]. It has also been shown that TTR interacts preferentially with Aβ aggregates rather than with Aβ monomers [82, 84, 85]. Recently obtained NMR data have revealed that the interactions between the Aβ1-40 monomer and TTR occur via the T4 binding pocket of the tetramer [66]. Moreover, newly reported NMR analyses have also shown interactions between TTR and the C99 fragment of the Aβ precursor protein (AβPP), possibly leading to the inhibition of Aβ production [86]. In vivo, however, interactions between Aβ and TTR have only been observed in human kidneys [87] and muscles [88], but the importance of this association is not clear due to the absence of clinical AD pathology in the subjects in that study.

The critical role of TTR in AD was established in several AD transgenic mouse models. The ameliorative effect of TTR on the AD phenotype in vivo was indicated in an APP23 murine AD model designed with tissue-specific overexpression of human TTR [85]. Moreover, studies from several laboratories have also reported that knockout of the TTR gene in AD mouse models results in distinct behavioral deficiencies and neuropathological changes [85, 89], suggesting that TTR is important in AD. However, the results from TTR gene silencing experiments in transgenic mouse AD models have not been uniform. Some studies have demonstrated no effect [90] or have shown the opposite effect [91] of TTR gene knockout in AD mouse models.

Interestingly, some but not all studies of CSF from AD patients have reported a reduction in TTR levels [92-95]. It has been proposed that the decline in TTR might be the result of TTR-Aβ interactions or that it might be related to the neuronal loss in AD that eliminates the neuronal source of TTR synthesis [96]. The cerebral injection of anti-TTR antibodies in Tg2576 transgenic mice could enhance Aβ-associated pathology, and the observed cognitive decline in these mice was explained as a result of decreased TTR levels [97].

The mechanism by which TTR might play a protective role in AD is not fully understood. To summarize the existing in vivo and in vitro data, TTR appears to act as a multimodal suppressor of AD pathogenesis by inhibiting AβPP processing [86, 96], by binding to the Aβ peptide, and, perhaps, by proteolytically cleaving it [65], thus preventing the formation of cytotoxic oligomers and fibrils. TTR might also play a role by transporting Aβ from the brain to the liver [98].

2.4.3 The proteolytic activity of TTR

In addition to its transport functions, TTR can act as a protease. The proteolytic activity of TTR was first revealed during an investigation of the TTR-Apolipoprotein A-I interaction [64]. Apolipoprotein A-I is a major protein component of high-density lipoproteins, and it binds to about 1–2% of the total TTR molecules [99]. In vitro studies showed that the nature of this interaction has a proteolytic character that leads to the cleavage of Apolipoprotein A-I at its C-terminus [64]. Moreover, other in vitro studies demonstrated that TTR is able to cleave the Aβ peptide, which suggests a protective role of TTR in AD [65]. However, no evidence for this mechanism has been found in vivo. In addition, Neuropeptide Y was also found to be a substrate of TTR proteolytic activity, but the nature of this action is unclear [100]. Thus, the impact of TTR’s proteolytic activity and the physiological role of such activity require further investigations.

2.4.4. Other functions of TTR

It has been debated whether our knowledge about TTR’s function is complete [101, 102]. TTR has been highly conserved throughout evolution and so far no human with a TTR gene deletion has been reported [31]. According to previous reports, the transport of T4 in plasma is most likely not a main function of TTR [68]. Extensive in vitro studies have shown that over one thousand different aromatic small molecules have structural complementary to the T4 binding site (TBS) of TTR [103]. The wide array of

molecules that can potentially bind to TTR suggests that TTR might have ‘hidden’ functions.

According to Buxbaum and Reixach [102], TTR might possess a detoxifying function. Many metabolites that circulate in human blood are naturally occurring degradation products from either nutrition intake or other environmental exposures. Binding of these molecules to TTR might allow them to be transported to the liver or kidneys for further degradation. The interactions between hydroxy-metabolites and TTR have been reported earlier in toxicological reports [104, 105]. The role of such binding is still a matter of controversy, and it might be toxicity-mediating or detoxifying. Moreover, TTR has been suggested to bind to the degradation products of noradrenaline [106], but the physiological significance of this is not known. It is likely that TTR circulating in plasma can bind to many different ligands, and this might explain why the amyloid depositions of some very unstable mutant TTRs assemble in the kidneys (where no TTR synthesis occurs). The binding of ligands to the tetramer has a stabilizing effect on the protein [107]. The excretion of ligands into the kidney would decrease the stability of TTR variants that are highly prone to aggregation, and this might contribute to the aggregation, and this might contribute to the accumulation of aggregates in the kidney [102].

3. TTR as an amyloidogenic protein

3.1 What triggers TTR aggregation?

In addition to the physiological roles of TTR described above, it is also a representative amyloidogenic protein in humans. The amyloidogenic pathway of TTR aggregation is shown in Figure 2. In order to initiate the aggregation, the tetrameric TTR must dissociate into monomers [9-11]. This event is the rate-limiting step in the amyloid cascade and depends on the kinetic stability of the protein. In order to start the downhill polymerization, the monomer has to become partially unfolded [6]. This conversion depends on the thermodynamic stability of TTR. The equilibria between tetramer and monomer and between monomer and unfolded monomer in TTR are strongly linked, and consequently the destabilization of either the tetramer or the monomer will enhance the aggregation [108].

Figure 2. The TTR aggregation cascade. The dissociation of TTR is the rate- limiting step. Once the monomeric TTR is formed, it has to undergo a partial unfolding in order to initiate the downhill polymerization cascade. The downhill polymerization of TTR results in several morphologically different species, including oligomeric and fibrillar structures.

There are several factors, as discussed above, that can predispose for aggregation in amyloidogenic proteins. In the case of TTR, more than 120 different amyloidogenic point mutations have been identified [30]. Both the wild-type TTR tetramer alone and the tetramer mixed with different variant TTR subunits can dissociate, misfold, and aggregate into insoluble fibrils leading to amyloidosis. Mutations in the protein can significantly increase

the efficiency of TTR amyloidosis by affecting the rate of tetramer dissociation and/or the misfolding of the monomer. Several TTR mutants have been mapped according their kinetic and thermodynamic stability and are established in Figure 3 [15].

Figure 3. The relationship between the kinetic and thermodynamic stabilities of TTR mutants. The thermodynamic stability (x Axis) is expressed in term of Cm (urea denaturation midpoints) and the kinetic stability is established on y Axis as t1/2 of tetramer dissociation evaluated in vitro. Nonpathogenic TTR variants (white circles), suppressor mutants (black circles), variants associated with CNS depositions (blue diamonds), non-CNS associated (red diamonds) and mutants that establish sometimes CNS symptoms (green diamonds). The figure is adopted with permission from Sekijima, et.al [15].

However, the instability of the mutant TTR is not always directly correlated with its pathogenicity in vivo. For example, the most unstable TTR variants Asp18Gly and Ala25Thr are not the most pathogenic mutants [109, 110]. The lack of strong pathogenic potential in the most highly unstable mutants can be explained by the efficient work of the endoplasmic reticulum (ER) degradation machinery that results in a reduced concentration of secreted amyloidogenic precursors [15]. Thus, the pathogenic potential of the TTR variant depends on a combination of two factors: the stability of its tetrameric form and the efficiency with which it is excreted from the cell into the plasma [15].

Posttranslational modifications, including partial proteolysis, have also been suggested to play a key role in the aggregation of TTR. The presence of truncated TTR in cardiac deposits and adipose tissue has been indicated previously, but the importance of those truncated form has not been clarified [111, 112]. Moreover, the truncated form that corresponds to the polypeptide of residues 49–127 has also been shown to be a main fragment in ex vivo TTR amyloid fibrils [113]. The responsible protease for this truncation has not yet been identified, but the highly specific cleavage between Lys48 and Thr49 suggests that it is likely to be a trypsin-like serine protease. Recently, a study also reported that the C-terminal fragment (49–127 a.a.) is highly prone to self-aggregate in vitro, indicating that the proteolytic activity targeting TTR plays the crucial role in amyloid formation [114]. However, the mechanism underlying TTR fragmentation is not known. The proteolytic site at position 48 in TTR is fairly hidden, and the release of the N-terminal fragment requires sheering forces [114]. In addition, it seems that only aggregated TTR is prone to proteolytic activity, and soluble forms of TTR are inert [115]. It is possible that the appearance of truncated TTR in cardiac deposits might be the result of a combination of several posttranslational modifications that destabilize the protein that lead to partially aggregated TTR and subsequent proteolytic cleavage.

TTR exists in several isoforms in human plasma. The formation of the isoforms is a result of the modification of a single reactive cysteine at position 10 (Cys10) in the monomer [116]. Under normal physiological conditions Cys10 can be found in an oxidized (Cys10-SO3–), reduced (Cys10- SH), or sulfonated (Cys10-S-SO3–) form [117]. The formation of isoforms in vivo commonly occurs through cysteinylation, glutathionylation, or sulfonation. Depending on the structural modification of Cys10, the stability of the tetramer can be influenced in different ways; sulfonation yields a more stable form of TTR, whereas cysteinylation and glutathionylation cause destabilization of the TTR tetramer [118]. However, the changes in the native tetrameric structure of TTR do not trigger amyloid formation themselves but only lead to a less protected state of TTR against factors that can initiate aggregation [119].

Additionally, the formation of amyloid can also be promoted by extracellular factors. The interaction between soluble TTR oligomers and sulfated glycosaminoglycans, especially heparin, has been reported to accelerate the self-assembly of small aggregates into higher molecular weight fibrils by quaternary structural conversion [120].

3.2 TTR-mediated cytotoxicity

Historically, most attention has been focused on the structure of fibrils due to the assumption that amyloid deposits are the primary cause of the tissue damage associated with amyloid diseases. However, a number of recent studies have provided strong evidence that oligomeric or low molecular mass soluble species are more aggressive and have stronger cytotoxic potential than mature fibrils [121, 122]. As mentioned above, the aggregation of protein in TTR amyloidosis is a downhill polymerization that starts with a monomer and continues through to the formation of different misfolded aggregates on the way to becoming mature fibrils.

Different cell-based studies have investigated each of the downstream species of TTR aggregation. The results have shown that both monomers and low molecular weight oligomers, but not species larger than 100 kDa, are able to trigger cell death [123-125]. The toxicity of mature fibrils, however, is still under debate. Several studies have shown that TTR fibrils possess no toxicity [123-125], while some reports suggest that the heavy deposition of TTR fibrils might induce tissue damage through direct compression and blockage of the median nerve. For example, carpal tunnel syndrome can be relieved by surgical removal of amyloid deposits [126].

The molecular mechanism that underlies TTR-mediated toxicity remains unclear, but some possible pathways have been suggested. One hypothesis is that Ca2+ homeostasis can be disrupted by the formation of highly toxic ion channels in lipid membranes by pore-like annular oligomers and that this leads to cell induction of apoptosis [127]. The formation of annular oligomers appears to be a common feature in the large range of amyloidogenic proteins in vitro [128-130]. Moreover, the presence of annular oligomers has been observed in AD brains, suggesting the importance of annular species in AD pathogenesis [131]. Recently, the formation of annular oligomers was also reported to be a component in the TTR amyloidogenic pathway, and this suggests that the same mechanism might be responsible for TTR-mediated cytotoxicity [132].

Interestingly, the involvement of calcium homeostasis as a contributing factor in TTR-mediated cytotoxicity has been discussed previously. According to Hou et al. [123], low molecular weight TTR species can effectively induce an influx of calcium ions via L- and N-type voltage-gated calcium channels. The mechanism of TTR’s action on Ca2+ influx, however, is still not clear. These authors have hypothesized that the disruption of Ca2+ homeostasis might be a result of plasma membrane fluidity mediated by TTR

binding [133], which modifies the activity of proteins, including ion channels.

The activation of the receptor advanced glycation end (RAGE) – which leads to ER stress, the activation of ERK1/2, and caspase-dependent apoptosis – has also been suggested as a possible mechanism underlying TTR-mediated toxicity [134-136]. The expression levels of RAGE have been shown to be increased in tissues affected by familial amyloid polyneuropathy (FAP). Moreover, the binding of fibrillar TTR to RAGE leads to the activation of nuclear factor (NF)-κB expression and its downstream factors. Thus, the interactions between TTR and RAGE trigger cell stress, inflammation, the activation of caspases, and eventually cell death [136]. Interestingly, the expression levels of RAGE, nitric oxide synthases and cytokines were found to be higher in patients at early stages of the disease before the amyloid depositions had appeared. These observations led to the suggestion that up- regulation of expression levels might be a result of interactions between soluble TTR aggregates and RAGE [136].

Another suggested mechanism in TTR-mediated cytotoxicity is age-related oxidation of TTR. According to Reixach et al. [137], the oxidized form of TTR exerts cytotoxic effects in a human cardiomyocyte cell line. This mechanism was suggested to be behind late-onset ATTR, including wild-type ATTR (ATTRwt) previously known as senile systemic amyloidosis.

3.3 Sporadic and hereditary amyloidosis: the clinical spectrum

There are two forms of transthyretin-related amyloidosis: hereditary and sporadic.

3.3.1 Wild-type TTR amyloidosis

ATTRwt is caused by the deposition of wild-type TTR and affects predominantly men over the age of 80 [138-140]. The depositions are mostly found in the heart [141, 142] but can also be manifested in other organs such as soft tissues, the bladder, and the gastrointestinal tract [143-146]. The gender-related incidence is poorly understood. In some FAP (will be discussed later) populations, the onset of disease in males has been recorded to be 2–11 years earlier compared to the onset in females [147]. In addition, the recovery from liver transplantation (which will be discussed below) is

also less successful in males compared to females within the late-onset disease group [148]. The involvement of sex hormones as a possible effector of TTR levels has been suggested. Studies in mice have shown that both male and female sex hormones are able to up-regulate mRNA levels of TTR, although the effect of testosterone on TTR expression was stronger than the effect of estrogen [147]. These observations suggest that the higher levels of TTR expression and the increased incidence of ATTRwt in elderly men might be caused by a slow decrease in testosterone production [147].

3.3.2 Hereditary TTR amyloidosis

Mutation of the TTR gene can lead to hereditary ATTR, and more than 120 different point mutations in the TTR gene have been identified as pathogenic [149]. Depending on the site of deposition, hereditary ATTR has been divided into two phenotypes: FAP, which predominantly affects the nervous system, and familial amyloid cardiomyopathy (FAC), a form of amyloidosis in which aggregates form in the heart. However, the course of disease, including symptoms, progression, and age of onset can vary within the group and even between individuals who carry the same mutation [150].

Familial amyloid polyneuropathy

FAP was first described by the Portuguese neurologist Corino Andrade more than 50 years ago. He discovered a group of patients with a fatal form of inherited amyloidosis characterized by progressive sensory-motor and autonomic polyneuropathy [151]. In 1984, amyloidogenic TTR with a mutation at position 30 (Val30Met) was identified as a common underlying genetic variant of FAP [152].

Historically, FAP was considered to be an endemic disease with high prevalence in Portugal, Japan, and Sweden [153-155]. However, in the last 20 years, large numbers of sporadic cases have also been reported in other European countries and in the US [156-159]. Depending on the age of onset of the disease, patients with the Val30Met mutation are classified into two groups: early onset and late onset. The clinical characteristics in both groups differ greatly. For patients with early onset (symptoms of the disease appear before the age of 50), the disease has a high penetrance rate, gives rise to loss of superficial sensation, impairs thermal sensibility, causes distinct autonomic dysfunctions affecting the heart and gastrointestinal tract, and is characteristic for populations in Japan and Portugal [160]. In contrast to early onset, the late-onset disease (when symptoms appear after the age of 50) has a low penetrance rate, a male prevalence, leads to the loss of all sensory sensitivities, and causes relatively mild autonomic dysfunctions and

amyloid cardiomyopathy. The area of prevalence of late-onset FAP is in northern Sweden [161].

Familial amyloid cardiomyopathy (FAC)

The mutation at position 122 (Val122Ile) in TTR is a common mutation in FAC, and the deposition of these mutant TTR aggregates occurs predominantly in the heart. This variant of TTR has been reported to target 3.9% of African-Americans in the US and is a late-onset disease [162, 163]. Patients with FAC exhibit congestive heart failure, cardiac arrhythmia, and obstructed blood flow conduction blocks [164].

3.4 Therapeutic strategies for TTR amyloidosis

The therapeutic approaches to ameliorate TTR amyloidosis can be classified depending on their mode of action (Fig. 4). Gene therapy is used to inhibit the production of TTR, kinetic stabilizers prevent the dissociation of tetrameric TTR, and clearance strategies seek to degrade the TTR amyloid matrix.

Figure 4. Therapeutic strategies for TTR amyloidosis. The inhibition of mutant TTR production is achieved by liver transplantation. Antisense oligonucleotides (ASO) and small interfering RNAs (siRNAs) are RNA-targeting therapies that suppress both mutant and wild-type TTR production. Binding of kinetic stabilizers (red circles) to TBS decreases the rate of dissociation (red arrow) of TTR tetramer into monomers and prevents its aggregation. Degradation of TTR aggregates and amyloid fibrils is achieved by the combined use of the antibiotic

doxycycline and the biliary acid taurousodeoxycholic acid (TUDCA). The immune therapy has not yet been tested and requires further investigations.

3.4.1. Liver transplantation

The first liver transplantation as a treatment for ATTR was performed in 1990, and three years later the treatment was finally accepted due to its encouraging outcomes [165, 166]. Previous to this, ATTR had been considered an incurable disease. The main goal of the transplantation is to replace the liver that contains the mutant TTR gene with a liver that only contains the wild-type TTR gene, hence reducing the levels of mutant TTR in plasma and replacing it with wild-type TTR [167, 168]. The follow up in long- term studies in transplanted patients has shown a clear regression of amyloid deposits in abdominal fat tissue [169]. Interestingly, the survival rate was significantly better in the group of patients with the Val30Met mutation (82%) than in those with the non-Val30Met mutation (59%) at 5 years after surgery [170]. However, the survival rate for the transplanted patients in the early onset group did not differ significantly compared to the non-transplanted patients in the late-onset group [148, 171].

Currently, liver transplantation is the only accepted treatment to ameliorate ATTR. Despite its beneficial impact on the disease, it also has some limitations. First, the treatment only targets a specific group of ATTR patients, and it is not applicable for ATTRwt and FAC patients [138, 172]. Second, it will only replace mutant TTR produced by the liver, and mutated TTR will still be produced by other organs such as the choroid plexus. Third, the progression of wild-type TTR cardiac deposition has been observed in some patients after surgery [173]. Some reports explain this phenomenon by suggesting that pre-existing fibrils will have a seeding effect on wild-type TTR [174, 175]. This is, however, controversial due to fact that TTR aggregation has been shown to not be dependent on seeding [103]. Interestingly, depositions of posttranslationally modified TTR are more progressive in patients with fibrils of type A than of type B [111]. Type B fibrils are composed of full-length form of TTR. Type A fibrils consist mostly of the truncated form of TTR, which might be prone to seeding. However, the seeding ability of those fibrils has not been investigated yet. Fourth, an extended surgical procedure with the requirement for long-term post- transplantation immunosuppressive therapy limits the group of potential transplant patients with regard to age and the stage of the disease.

3.4.2. RNA suppression therapies

Antisense oligonucleotides (ASO) and small interfering RNAs (siRNAs) are RNA-targeting therapies that are currently being tested in clinical trials [176, 177]. Despite their different chemical structures, delivery methods, and intracellular mechanisms, they have some common features. They are both designed to bind to the 3’ untranslated region (3’UTR) of TTR mRNA and to degrade it. In addition, they will also inhibit the production of both wild-type TTR and mutated TTR [178, 179]. Studies of these RNA suppressor therapies have shown a promising effect with significant reductions of both wild-type TTR and variant TTR proteins in plasma [180, 181]. However, one theoretical side effect of the therapies is the impairment of night vision due to reduced transport of retinol, and it is still not established if these therapies have the ability to reduce cardiac depositions.

ASO therapy

ISIS-TTRRx (ISIS 420915) is a single-stranded oligomer consisting of 20 nucleotides that is complementary to the 3’UTR region of TTR mRNA. The binding (hybridization) to mRNA results in RNase H-mediated degradation of the mRNA [182]. However, the ASO treatment does not affect the production of TTR in the CSF due to its inability to cross the blood-brain barrier. This fact might be important due to the potential neuroprotective role of TTR in AD (as described above). A clinical phase I study where healthy voluntaries were administered a 300 mg dose of ISIS-TTRRx over the course of four weeks demonstrated a 75% reduction of TTR protein levels without showing any adverse side effects [181]. Currently, ISIS-TTRRx is being evaluated in a long-term phase II/III study to estimate the efficacy and safety of the drug in FAP patients. This clinical trial is planned to end in 2017. siRNA therapy siRNA is another RNA-targeting therapy that aims to suppress the production of the TTR protein. In this therapy, double-stranded RNA molecules are synthesized that target and cleave complementary mRNA. Currently, there are two siRNAs in clinical development: patisiran (ALN- TTR02) for the treatment of FAP, and revusiran (ALN-TTRsc) for the treatment of FAC [177, 180]. Patisiran is administered by intravenous infusion with a second-generation formulation of lipid nanoparticles targeting the liver. Revusiran consists of siRNAs conjugated to GalNAc, which is recognized by asialoglycoprotein receptors on hepatocytes. The effect of patisiran has been analyzed in healthy volunteers in a phase I study

and in a short-term phase II study in FAP patients. Revusiran has been evaluated in healthy volunteers in a phase I study and is currently being investigated in a phase II study with FAC and ATTRwt patients. Both drugs effectively reduce TTR levels in the serum by more than 80% [180, 181]. A phase III study for revusiran is ongoing and will be completed in 2017.

3.4.3. Stabilizing the structure of TTR tetramer

In 1993, Coelho et al. [183] reported on a family in Portugal that carried the VAL30MET mutation, which is associated with highly penetrant FAP, but did not develop polyneuropathy [184]. Further investigations revealed that the suppression of the disease occurred due to the expression of a Thr119Met variant instead of wild-type TTR from their second allele [185, 186]. The Thr119Met variant was later shown to be kinetically much more stable than wild type. According to biophysical measurements, the dissociation rate of the Thr119Met homotetramer is 25-fold slower than the dissociation rate of the wild-type homotetramer. Thus, the inclusion of the Thr119Met subunit into a tetramer containing disease-associated subunits results in the formation of a more kinetically stable heterozygotic tetramer, and this protects these individuals from amyloidogenesis and disease [185, 186]. Hence, another strategy to ameliorate ATTR focuses on the possibility of stabilizing the native TTR structure. This can be done by introducing small molecules – referred to as kinetic stabilizers – that can bind to the TBS of TTR. The idea of TTR kinetic stabilizers as a therapeutic approach in ATTR amyloidosis was suggested and developed in the lab of Jeffrey F. Kelly [58, 186]. In the early 1990s, they showed that the amyloid cascade in TTR amyloidosis is a downhill polymerization that requires TTR dissociation [19]. They could also show that the binding of the T4 and 2,4,6-triiodophenol could prevent TTR amyloid formation [187]. These observations led to the development of the kinetic stabilization strategy that resulted in the discovery of tafamidis (trade name Vyndaqel®) [188], which is currently the only approved drug for preventing TTR amyloidosis.

This strategy is considered to be most conservative for treating ATTR. The dissociation of TTR into monomers is the first and rate-limiting step in the aggregation cascade [186]. Once the misfolded monomer is formed, TTR aggregation becomes very efficient and results in a variety of aggregated species, including amyloid fibrils and oligomers [19]. Thus, preventing aggregation after TTR dissociation seems to be an inefficient solution. The binding of drugs to TTR will not significantly inhibit the transport of T4 because only 1% of the total TTR is involved in such transport.

Several ligands have now been shown to stabilize the tetrameric structure of TTR and prevent its aggregation [58]. Some of these ligands are already in clinical use, and promising findings have recently been reported for the non- steroidal anti-inflammatory ligands diflunisal [189] and tafamidis [190-192] for the treatment of FAP.

Tafamidis

Tafamidis (2-(3,5-dichloro-phenyl)-benzoxazole-6-carboxylic acid) is currently the only approved drug on the market for the treatment of FAP. Tafamidis was initially approved for commercial use in Europe in 2011 and later on in Japan, Mexico, and Argentina where it is used as a first option in the treatment of patients with early-stage FAP [193]. Tafamidis binds with high selectivity to the TBS of TTR in a negatively cooperative manner with a

KD1 of 2 nM and a KD2 of 154 nM. The drug kinetically stabilizes both wild- type native TTR and mutant TTRs [188]. The European Medical Agency reports that a plasma concentration corresponding to 3.5–8.5 µM of tafamidis is clinically beneficial and that a corresponding steady state level of the drug can be obtained after 14 days of daily administration at a subclinical dosage [194].

Two clinical trials have shown that treatment with a 20 mg daily dose of tafamidis effectively reduces neurological progression and maintains overall quality of life and nutrition status in FAP patients [190, 195]. According to several reports, tafamidis appears to be a generally safe drug [190, 195]. Moreover, TTR stabilization was achieved in more than 90% of the patients [195]. The drug has a greater effect at the early stage of the disease, and was not able to stop disease progression in the most advanced FAP patients [196]. An extended open-label study in non-Val30Met variants is ongoing and will be complete in August 2018.

Diflunisal

The non-steroidal anti-inflammatory FAD ligand diflunisal also binds to TTR with negative cooperativity (Kd1 of 75 nM and a Kd2 of 1100 nM) [197]. Compared to tafamidis, diflunisal only displays a modest binding selectivity for TTR. However, due to its high oral bioavailability and high plasma concentration (100–200 µM) after administration of a 250 mg dose, diflunisal showed beneficial results in stabilizing TTR [198, 199]. Recently published results of a diflunisal trial showed that the drug significantly decreased the rate of neuropathy and maintained the quality of life in all stages of FAP over a two-year period [189]. However, non-steroidal anti- inflammatory drugs (NSAIDs) like diflunisal are frequently associated with

gastrointestinal side effects due to their inhibitory effect on the cyclooxygenase 1 system, and prolonged treatment might therefore be problematic [200].

3.4.4. Clearance of the amyloid matrix

The reduction of TTR deposits is achieved by the combined use of the antibiotic doxycycline and the biliary acid taurousodeoxycholic acid (TUDCA). Both drugs are under clinical investigation for treating ATTR. Studies in a FAP transgenic mouse model have shown that doxycycline acts as a disrupter of TTR amyloid formation, while TUDCA acts to remove nonfibrillar TTR aggregates [201]. An open-label phase II trial in 12 patients with ATTR showed that the combined TUDCA/doxycycline treatment over 12 months had a stabilizing effect on the disease in the majority of patients by impairing the progression of cardiac involvement and stabilizing the neuropathy scores. The preliminary data from that trial have also shown that the combined treatment with doxycycline (200 mg/daily) and TUDCA (750 mg/daily) has an acceptable toxicity profile [202].

Aims

• to select and evaluate new TTR tetramer stabilazers by repurposing already FDA-approved drugs

• to elucidate the properties that influence efficacy of ligands to stabilaze TTR tetramer in human plasma

4. Methods

The work presented here is mostly focused on the kinetic stabilization of TTR. To investigate important aspects regarding ligand binding to TTR, we used different biochemical and biophysical methods. The screening for potential TTR stabilizers was performed with a well-established assay based on monitoring TTR stability upon ligand addition under acidic conditions. Isothermal titration calorimetry (ITC) was used in our studies to identify the thermodynamic signature of TTR-ligand complex formation, which includes binding affinity (KD), entropy (∆S), and enthalpy (∆H). Moreover, to monitor the stabilizing effect of ligands on TTR in a cell model, we used the Alamar Blue cell viability assay based on resazurin reduction. In addition, the fruit fly Drosophila melanogaster model for FAP was also used to investigate the effects of different ligands on TTR variants, and we used x-ray crystallography to acquire a better understanding of the mechanisms behind the interactions between ligands and TTR. Finally, to characterize the effects of ligands on TTR in human plasma, which is a natural environment for TTR, we developed a new assay as described below.

4.1 Monitoring drug selectivity in plasma (urea-mediated denaturation plasma assay)

To estimate the kinetics of TTR dissociation in the presence of plasma, an indirect method of detection is required. The plasma assay we developed (for more details, see [203]) is based on urea-mediated denaturation similar to the method of Miller et al. [197] but without the cross-linking step. The denaturation by urea causes the reversible conversion of tetrameric TTR into monomeric TTR, and in order to dissociate the tetramer into monomer in plasma we used a urea concentration of 4.0 M, which is in the range of the midpoint for unfolding of wild-type TTR (Cm, unfold ~3.8 M) [108]. The addition of a ligand at different concentrations prior to the denaturation step shifts the rate of dissociation toward the native form of TTR. This, in turn, results in reduced levels of monomeric species in a concentration-dependent manner, which can be correlated to TTR stability as a function of ligand concentration. The separation of monomeric TTR species from whole plasma samples is achieved by gel electrophoresis. However, to prevent the renaturation of TTR without additional denaturation, we optimized the concentration of SDS in both the loading (0.2%) and running (0.025 %) buffers. The level of dissociation can then be monitored by the density of the monomeric bands using a Western blot approach. The estimated areas for

each band in turn can be plotted as a function of the corresponding ligand concentrations in order to determine the 50% inhibition concentration (IC50) for each ligand.

5. Results and Discussion

It is well known that the formation of TTR amyloid requires dissociation of its tetrameric structure [9-11]. Binding of ligands to the TBS on TTR reduces the dissociation rate of the protein and represents a potential avenue for intervention [204]. Several ligands have been discovered to inhibit aggregation of TTR in vitro [197, 199, 205-217], but only two of them – diflunisal [189] and tafamidis [190-192] – are in clinical trials. Although diflunisal has shown promising outcomes in clinical trials, it might be not a suitable candidate for prolonged treatment due to its gastrointestinal side effects. Tafamidis is the only drug that is currently approved for FAP treatment, and thus the identification of new TTR stabilizers is of significant interest.

The design of new drugs and their subsequent approval is a highly expensive and a time-consuming process. It takes approximately 13 years for a successful candidate to be approved for the market after being discovered, and repurposing already FDA-approved drugs for new applications can offer a ‘short cut’ for the identification of new candidates. This approach avoids many steps in drug development, including safety approval and pharmacokinetic studies, and thus reduces the time and cost of the process. Applying this idea, we screened substances from the Prestwick Chemical Library® based on FDA-approved drugs against TTR. The evaluation of hits resulted in the selection of nine potential TTR tetramer-stabilizing ligands. These included diflunisal (DIF), diclofenac (DIC), niflumic acid (NIF), aceclofenac (ACL), gemfibrozil (GEM), luteolin (LUT), apigenin (API), meclofenamic acid (MEA), and tolfenamic acid (TOA), which are shown in Figure 5. All compounds except GEM, ACL, and TOA have previously been demonstrated to impart kinetic stabilization on the TTR native state [214, 218, 219].

An effective TTR stabilizer must show high affinity for the TBS and also display a high selectivity for its target in the human body. Moreover, it should have a low incidence of side effects and should be resistant to enzymatic degradation in vivo. Targeting a specific protein in human plasma is a challenging task in any drug design due to nonspecific binding. Focusing on the selectivity factor of TTR-stabilizing ligands, we characterized the nine ligands listed above both in the presence and the absence of plasma with the goal of determining the properties that influence their efficacy.

Figure 5. Structures of selected TTR ligands. Luteolin (LUT), apigenin (API), diflunisal (DIF), diclofenac (DIC), aceclofenac (ACL), meclofenamic acid (MEA), niflumic acid (NIF), gemfibrozil (GEM), and tolfenamic acid (TOA).

5.1 The presence of plasma components strongly influences ligand selectivity

One of the main requirements for a ligand to stabilize TTR in vivo is that it must be continuously present at an effective concentration in plasma to saturate the majority of the TTR molecules. Most TTR ligands, including the drugs tafamidis and diflunisal, bind TTR with negative cooperativity (KD2 is higher than KD1). Regarding the natural TTR ligand T4, the binding affinities between the first and the second ligands differ by a factor of 100 [220]. This means that the second ligand will contribute only 1% to the total stabilizing energy, thus the dominating stabilizing effect will be provided from binding of the first ligand. Despite the fact that some differences in cooperativity can be observed between different ligands, this will only have a minor impact on the overall stabilizing effect. However, the TTR concentration in human plasma is approximately 4 µM, and in order to saturate the binding sites on TTR, the ligand must be present at least at this concentration [221]. Based on

this, a 90% saturation of TTR with ligand at a concentration of 4 µM would require KD values less than 200 nM according to law of mass action.

Assuming that ligands bind to TTR with equal affinities, we determined the affinity constants for nine different TTR ligands by ITC (Table 2). The estimated KD values ranged from 70 nM to 1200 nM. Among of these ligands, LUT, TOA, and GEM displayed significantly low affinity constants (below 200 nM), and thus, according to our estimation, would require concentrations lower than 4 µM to saturate 90% of the TTR molecules in plasma. However, although the efficacy of a ligand in stabilizing TTR is directly correlated to its binding affinity, in vivo efficacy is dependent on several other factors. One of these is nonspecific binding, which adversely affects selectivity and decreases the effective concentration of the ligand in plasma.

Table 2. The properties of the TTR ligandsa.

aThe table lists the IC50 in plasma for each compound as well as chemical properties of the corresponding ligands with respect to molecular weight, number of hydrogen- bond donors/acceptors, XLogP3, number of rotatable bonds, topological polar surface area, and the specific charge in water at pH 7. Luteolin (LUT), diflunisal (DIF), apigenine (API), niflumic acid (NIF), aceclofenac (ACL), meclofenamic acid (MEA), diclofenac (DIC), gemfibrozil (GEM), and tolfenamic acid (TOA).

The urea-mediated denaturation plasma assay (see Methods section) allowed the selectivity factor to be estimated for each ligand in the presence of plasma, and the half maximal inhibitory concentration (IC50) values were used as a measure to quantify the efficacy of the ligand in plasma (Table 2). Our results revealed that the ability of a ligand to bind TTR was significantly

decreased due to nonspecific binding to other plasma components. Moreover, the efficacy was not only reduced for all ligands in the presence of plasma, and the effect of non-specific components in plasma also varied greatly between the ligands. For instance, the ligands LUT and GEM both display the highest affinities (70 nM and 100 nM, respectively) among the investigated ligands, but they differ dramatically in their ability to stabilize

TTR in plasma (Fig. 6). The highly selective LUT shows an IC50 corresponding to only 7 µM, which is very close to a 1:1 stoichiometric ratio between one ligand and a TTR tetramer in plasma, while GEM requires a concentration of 180 µM to inhibit TTR dissociation, which is approximately a 40-fold excess of the ligand relative to the TTR concentration in plasma to reach the IC50.

Figure 6. Selectivity of LUT and GEM in plasma. The selectivity of ligands in human plasma is reported as the half maximal inhibitory concentration (IC50) obtained from the urea-denaturation plasma assay. The IC50 on the graph is indicated by dashed lines. The dissociation of the tetramer is shown as a monomeric band (bottom panel) on a western blot and quantified through densitometric analysis with the ImageJ 1.48 program. The percent denaturation is plotted against the corresponding ligand concentration.

5.2 Enthalpy is a better predictor of selectivity

Nonspecific binding of drugs is a common problem in drug design, and targeting the factors that determine selectivity in a complex environment is of significant interest. To elucidate the mechanism and binding properties

that determine selectivity, we investigated the relationship between the binding affinity of a ligand and its selectivity in plasma. From the plot in

Figure 7, it is clear that the selectivity factor (IC50) correlates poorly with the binding affinity (KD). Although there is a slight correlation between KD and IC50 in plasma, some ligands show a strong deviation from linearity. Based on these results, we concluded that KD is a poor parameter for predicting selectivity in a complex system and that it should not be the only factor used for this purpose.

Figure. 7. Correlation between KD and IC50 in plasma. A plot between KD estimated by ITC in PBS and IC50 in plasma. The IC50 values are presented as the mean ± standard deviation (SD) of three independent experiments.

Because the KD of a protein-ligand complex can be expressed by the Gibbs free energy (∆G), as shown in Equation 1, the nature of complex formation can be described in terms of the enthalpic and entropic components in Equation 2.

Eq. 1 ∆G = −RT * ln KD where R is the gas constant and T the temperature in Kelvin

Eq. 2 ∆G = ∆H − T∆S

Based on the fact that different ligands with identical KD values can differ in their relative contribution of enthalpic and entropic forces, this might also have different effects on the binding properties.

The enthalpic component, ∆H, is predominantly associated with the formation of new chemical bonds upon protein-ligand complex formation. Because all chemical bonds have a distance requirement, and because hydrogen bonds are also position dependent, ∆H is highly dependent on the orientation of the molecules. The energy change in the entropic component (∆S) upon complex formation is predominantly linked to the restriction of free motions as well as desolvation of bound water, and it is mainly the result of hydrophobic interactions. The ∆S component is therefore less dependent on the specific orientation of molecules [222].

Based on this, we hypothesized that a ligand with a high contribution of enthalpic forces (∆H) when binding to its co-partner would be more selective in a complex environment than a ligand with similar affinity but with a higher ∆S component. The relative contributions of ∆S and ∆H between TTR and the ligands were estimated by ITC (Table 2). Interestingly, by plotting the ΔH versus IC50 for all ligands (Fig. 8A), we found that, in contrast to the binding affinity, the ΔH correlates better with the plasma selectivity.

Moreover, despite the similar KD values of LUT and GEM, LUT has a large negative ΔH and has proportionally higher selectivity as compared to GEM, which has a lower contribution of ΔH. In addition, to confirm just the relative contribution of ∆H to the total ∆G, the contribution of the affinity factor was eliminated for each ligand by estimating ∆H/∆G. A plot of ∆H/∆G versus IC50 for all ligands also correlates well with selectivity (Fig. 8B).

We also examined how the ligands’ properties as determined from the de novo analysis of their structures (listed in Table 2) correlated with their selectivity. Almost all parameters failed to correlate with selectivity except for the number of rotatable bonds. The most selective ligands – LUT, API, and DIF – have a low number of rotatable bonds. This correlation, however, was expected due to the fact that the more flexible a ligand molecule is, the more likely it will interact with more than one target, resulting in a loss of selectivity.

Figure 8. Correlation between enthalpy and IC50 in plasma. (A) A correlation plot between enthalpy (∆H) and IC50 in plasma. (B) A correlation plot between the relative contribution of enthalpy and the total Gibbs free energy (∆H/∆G) and IC50 estimated in plasma.

Taken together, our results show that the ΔH of the binding interaction might be a better predictor of selectivity than the KD due to its better correlation with IC50. This finding is significant for the design and selection of more specific TTR-stabilizing drugs, which is of high therapeutic importance. The screening for a high enthalpic component, however, could also be a general approach in all drug design to identify highly specific ligands in a crowded environment, which is an almost universally desired feature in drug design.

5.3 Using the method to estimate the suitable therapeutic dosage

The evaluation of a suitable therapeutic dosage is necessary to achieve the desired effect of a drug on TTR stabilization. By applying the estimated IC50 values from our assay, we investigated the effect of currently used dosages of tafamidis and diflunisal. For tafamidis, the administration of 20 mg daily results in a plasma concentration of 3.5–8.5 µM, which is clinically beneficial [194]. Moreover, the corresponding steady state level of tafamidis is acquired after 14 days of daily administration at subclinical dosage [194]. Based on the urea-mediated denaturation plasma assay results, the IC50 value for tafamidis was determined to be 4.6 µM, and this concentration corresponds to saturation levels of 43–65%. Interestingly, even though the saturation of

TTR might, based on our results, not be complete, tafamidis still exerts a clinical effect in vivo. An increased dosage might have a more profound effect in vivo.

Another TTR-stabilizing drug, diflunisal, which has a clinical effect in vivo

[189], has an IC50 of 25 µM as estimated in our assay. Despite modest binding selectivity to TTR, diflunisal has high oral bioviability resulting in a high plasma concentration [223]. From reported data, the administration of diflunisal with dosages of 250 mg or 500 mg twice daily over seven days results in serum concentrations of 145 µM and 420 µM, respectively [198]. Applying our assay, this corresponds to a saturation level between 85% and 94%. These high saturation levels might explain the reason for its clinical efficacy. However, it should be noted here that although our assay provides a good means to compare the effect of different drugs on plasma TTR stabilization, it is performed under denaturing conditions that might slightly shift the observed IC50 value and subsequently the estimated saturation level in plasma. Therefore, an accurate figure of the saturation levels in plasma should be determined by an independent method.

5.4 Luteolin suppresses the TTR-mediated pathological effect in cells and in an animal model of FAP

As discussed above, luteolin is a highly selective ligand that stabilizes TTR in plasma. The estimation of IC50 for luteolin in plasma gave a value of only 7 µM, which is at stoichiometric level. From a therapeutic point of view, the extremely high selectivity of luteolin is of particular interest. Luteolin (3´,4´,5,7-tetrahydroxyflavone) is an abundant flavonoid synthesized in plants that possesses several health benefits, including neuroprotective and antioxidant properties [224, 225]. Moreover, it is widely used within traditional Chinese medicine and has few side effects even upon prolonged intake at high doses [226]. Although a group of naturally derived flavonoids was previously characterized as potential TTR stabilizers [107], the effect of luteolin in TTR stabilization was only recently introduced by Trivella et al. [214].

It is well-established that preservation of the TTR tetramer impairs the ability of TTR to exert a cytotoxic response [227], and the effect of luteolin on TTR stabilization was investigated both in cell and animal models. We tested luteolin against TTR-induced toxicity on a human neuroblastoma cell line and found that it efficiently prevented the cytotoxic effects of TTR (Fig.

9A). The effect of luteolin was also examined in the D. melanogaster model of FAP. This animal model expresses the Val30Met TTR variant and displays an impaired neurological phenotype that resembles the clinical manifestations of TTR Val30Met-related amyloidosis in humans [228]. Administration of luteolin to D. melanogaster completely rescued the impaired phenotype as determined by a climbing assay [228] (Fig. 9B).

Figure 9. Effect of luteolin on (A) TTR-induced cytotoxicity and (B) on the fruit fly D. Melanogaster model for FAP. Luteolin-7-O-glucoside shows no effect on TTR- induced cytotoxicity.

5.5 Enzymatic modification inactivates the effect of luteolin on TTR

To obtain the desired effect on stabilization of TTR in the human body, the ligand must be constantly present at its effective concentration. Thus, in order to preserve the required concentration, which in the case of TTR is above 4 µM, the ligand has to be resistant to enzymatic degradation. Luteolin, however, has rather low metabolic stability in vivo, and it is modified by enzymatic degradation pathways in the liver and intestines. The

preferred site of modification in humans is position 7 of the A-ring, which is particularly susceptible to glucuronidation [229]. According to a previous structural analysis of TTR in complex with luteolin, position 7 of the A-ring is facing into the solvent [214]. Based on this, we hypothesized that conjugation with glucuronide at this position would not interfere with luteolin binding to TTR. The effect of enzymatic conjugation at position 7 of the A-ring in luteolin was investigated by using the luteolin-7-O-glucoside analog, also known as cynaroside (Fig. 10)

Figure 10. Structures of luteolin analogs.

We used a well-established assay based on turbidity measurements when TTR amyloid formation is induced at low pH [197], and we could clearly show that a glycoside group in position 7 of the A-ring abolished its ability to prevent aggregation (data not shown). Moreover, the cell viability assay in the neuroblastoma cell line revealed that luteolin-7-O-glucoside is unable to

rescue the cells from the cytotoxic effects of TTR (Fig. 9B). These results clearly indicated that a large modification at position 7 of the A-ring renders the luteolin molecule inactive in stabilizing TTR.

To more closely investigate the molecular interaction between luteolin and TTR, we solved an atomic-resolution crystal structure of TTR in complex with luteolin (Fig. 11A). Interestingly, according to our structure, the A-ring of luteolin is actually buried within the core of the TBS and is not located on the outer surface of the protein as was shown previously (Fig. 11B). This orientation explains the abolished effect of luteolin-7-O-glucoside on TTR both in the low pH assay and in neuroblastoma cells. Based on our structure, we conclude that any bulky substitutions at position 7 of the A-ring, including glucuronidation, would inactivate the luteolin as a TTR stabilizer due to steric exclusion of luteolin from the TBS. We also suggested that a similar orientation would be seen for all other flavonoids that are reported to bind TTR.

Figure 11. The crystal structures of the TTR-luteolin complex. (A) The structure obtained in this study, which shows that the A-ring of luteolin is buried within the core of the TBS (PDB ID: 1F41). (B) The structural analysis of Trivella et al. (PDB ID: 4DEW, REF).

5.6 Substitution of 7-OH with 7-Cl and 7-MeO increases the resistance of luteolin to glucuronidation

Luteolin, as mentioned above, has a low persistence against metabolic activity in the human body and thus would require very high dosages to inhibit TTR in vivo, which impedes its potential therapeutic use. Nevertheless, due to its high selectivity for TTR in plasma, luteolin presents an interesting platform for the development of novel drugs targeting TTR- associated disorders. Focusing on improving the stability of luteolin in the face of metabolic degradation, we decided to modify position 7 of aromatic ring A, which is the most susceptible to glucuronidation. Based on our previous structural and biochemical analyses, this position is important for luteolin’s inhibitory effect on TTR and cannot be exchanged with any bulky modifications. Moreover, the substituent must be inert to glucuronidation, thus, the modifications with commonly used carboxyls, thiols, and amines are not suitable [230-232]. These factors resulted in two options: chloro (7- Cl-Lut) and methoxy (7-MeO-Lut) substitutions.

To confirm the stability against enzymatic degradation, we tested the modified luteolin molecules in an in vitro model based on human liver microsomes. Both substitutions improved the stability against glucuronidation and resulted in a significant increase in the half-lives compared to the parent molecule. Modification of luteolin with a chlorine atom increased resistance to glucuronidation by approximately 3-fold (t1/2 52 ± 7 min) relative to luteolin (t1/2 16 ± 1 min), while modification with a methoxy group (t1/2 36 ± 8 min) increased the resistance to glucuronidation 2-fold. However, it should be noted here that hydroxyl groups at positions 3’ and 4’ in luteolin are also subject to glucuronidation, though not to same extent as at position 7, and this will also affect the efficacy of luteolin on TTR stabilization.

5.7 Substitution of 7-OH decreases binding affinity and selectivity in plasma

The substitution of position 7 of luteolin with chloro and methoxy groups clearly improved the resistance of luteolin to enzymatic degradation. However, the main reason for attempting to improve the stability of luteolin in the first place was its remarkable selectivity in plasma. Thus the selectivity of the luteolin analogs in plasma was investigated with the plasma assay described above (Fig. 12). The estimated IC50 values show that both 7-Cl-Lut and 7-MeO-Lut have significantly reduced selectivity compared to the parent luteolin structure. The effect of nonspecific binding was more pronounced with 7-MeO-Lut (IC50 of ∼105 µM) resulting in an increase in the required

concentration up to 21-fold compared to luteolin (IC50 of ∼5 µM).

Substitution with a chlorine atom (IC50 of ∼65 µM) decreased the selectivity by about 13-fold relative to luteolin. To investigate the binding affinity of modified luteolin structures, we used microscale thermophoresis (MST). The MST results showed that both of the modified luteolin structures also had reduced binding affinity to TTR compared to luteolin, but they were still in the range (150–620 nM) of other high-affinity binders to TTR.

Figure 12. Stabilization of TTR in plasma. (A) luteolin, (B) luteolin-Cl, and (C) luteolin-MeO.

In addition, we solved crystal structures of TTR in complex with the luteolin analogs (data not shown). The structural analysis showed that 7-MeO-Lut binds to the TBS of TTR in the opposite orientation of luteolin with the methoxy group facing into the solvent, while 7-Cl-Lut binds in a similar orientation as luteolin. Interestingly, the structural analysis showed that both structures lost the ability to form hydrogen bonds between the 7-OH of the A-ring in luteolin and Ser117 and Thr119 of TTR, which is likely to be the reason for the decreased selectivity of these analogs. These observations are in good agreement with our hypothesis that enthalpic forces have a large effect on selectivity. However, we have investigated only two different modifications in this study, and it would be interesting to analyze other possibilities. The choice of modifications, as mentioned before, is size- restricted due to the confined space within the TBS, and useful substitutions must be inert to glucuronidation. In this context, the halogens are the most appropriate candidates. The introduction of bromine (–Br) or iodine (–I) atoms, however, is not a good option due to the large size of bromine and, in the case of iodine, there is a risk of becoming a T4 agonist. Fluorine is

interesting because the introduction of a fluorine atom at position 7 in luteolin would allow for hydrogen bonding between Ser117 and Thr119, which is desirable based on our hypothesis in order to preserve the high selectivity of luteolin. However, C–F is a very weak hydrogen bond acceptor [233, 234]. A structural study on complexes between TTR and ligands containing fluorine atoms within the hydrogen-bonding distance of Ser117 and Thr119 demonstrated that the fluorine is not involved in hydrogen bonding but rather is outcompeted in support of hydrogen bonds between the side chains of Ser117 and Thr119 [235, 236]. Based on this information, the substitution of 7-OH in luteolin with 7-F might not be able to preserve the high selectivity of luteolin.

Taking together, our results suggest that luteolin is a very interesting candidate from a therapeutic point of view due to its extremely highly selectivity for TTR in plasma. We have also shown that luteolin effectively attenuates the cytotoxic effect of TTR in human neuroblastoma cells and completely reverses the pathological phenotype in a D. melanogaster model of FAP. Unfortunately, luteolin has rather low stability in vivo and undergoes rapid degradation in the liver, which hampers its potential therapeutic use. The resistance to glucuronidation in the liver might be achieved by introducing relatively small substituents at position 7 of the A- ring in luteolin. To study this, we substituted the hydroxyl group at position 7 with a chlorine atom or a methoxy group. The modifications resulted in greater resistance to hepatic degradation, and both of the new luteolin structures showed high affinity to TTR, but in contrast to luteolin the selectivity was adversely affected in the presence of other plasma components.

5.8 TBBPA effectively prevents TTR aggregation and binds TTR with extremely high selectivity in human plasma

Another kinetic stabilizer, 2,2’,6,6’-tetrabromo-4,4’-isopropylidentediphenol (Tetrabromobisphenol A, TBBPA) has also been investigated in our studies. TBBPA is produced in large quantities (170,000 metric tons annually worldwide) for use as a reactive flame retardant in the laminate coating of printed circuit boards and as an additive flame retardant in acrylonitrile butadiene styrene plastics [237-240]. A number of organohalogen substances, including TBBPA, have previously been reported to bind to human TTR in vitro [237, 241].

We investigated the binding of TBBPA to TTR in both low and neutral pH environments. For comparison purposes, we also included diflunisal and tafamidis, which have been shown to have beneficial clinical effects in humans [189, 190, 195]. Turbidity measurements of the TTR dissociation rate under low pH showed that TBBPA, similar to diflunisal and tafamidis, efficiently impairs the aggregation of TTR (data not shown). We also estimated the binding affinity of TBBPA to TTR using ITC and found that

TBBPA binds to TTR with a KD1 of 6 nM and a KD2 of 975 nM, which is within the same range as tafamidis, which has previously been shown to bind

TTR with a KD1 of 2 nM and a KD2 of 154 nM [188].

However, the binding efficacy of ligands to TTR in plasma, as described above, is greatly reduced by other components that are present in plasma. Our plasma assay, as discussed in the methods section, showed that TBBPA exhibits an extraordinarily high selectivity in the presence of plasma and effectively stabilizes TTR at a stoichiometric ratio in plasma. The drugs tafamidis and diflunisal were also included for comparison (Fig. 13). The estimated IC50 shows that TBBPA is a significantly better ligand than diflunisal, and it has a binding selectivity within the same range as the recently approved drug tafamidis.

Figure 13. Stabilization of TTR in plasma. (A) TBBPA, (B) tafamidis, (C) diflunisal.

5.9 TTBPA from a drug perspective

As discussed above, the extraordinary selectivity of TBBPA raised the question of whether it could be used for therapeutic purposes. From the existing data on TBBPA, it is evident that the flame retardant has never been previously evaluated as a therapeutic drug. However, TBBPA is an abundant halogenated substance in the environment [242] and indoors [243], and due to its high usage and the almost constant human exposure to the substance, it has been extensively investigated in many toxicological studies. The reported data have revealed that TBBPA is not genotoxic and that there are no indications that TBBPA is carcinogenic for humans. The European Food Safety Authority (EFSA) recently concluded that the current dietary exposure to TBBPA does not raise a health concern [244]. Modification of thyroid hormone homeostasis, however, is the most sensitive endpoint with a critical reference point of response of 16 mg/kg body weight according to the EFSA panel’s conclusion [245].

A recent investigation in rats demonstrated that prolonged daily oral administration of 100, 300, or 1000 mg·kg-1·day-1 of TBBPA did not result in increased mortality [244]. Moreover, the reported studies in the literature revealed that there are no reproductive or teratogenic effects of TBBPA [245]. The existing data on TBBPA neurotoxicity, however, are very limited, and there are contradictory results with regard to the developmental neurotoxic potential of TBBPA. According to most reports, there are no neurobehavioral effects from either single doses (0.75 or 11.5 mg/kg body weight) in mice [246] or from repeated prolonged daily oral administration of 100, 300, or 1000 mg·kg-1·day-1 in rats [244]. However, some vague alterations of behavior in open field and contextual fear conditioning were reported in mice after administration of low doses of 0.1 and 5.0 mg/kg [247]. With regard to the effects of TBBPA on the liver, it was shown that TBBPA exhibits some signs of hepatotoxicity in rats and mice at high doses, but no sign of hepatic changes were observed at doses below 350 mg/kg body weight [248-250]. Adverse effects of high TBBPA doses were reported in a two-year analysis performed on Wistar Han rats and B6C3F1/N mice [245, 251]. The administration of TBBPA at dosages of 500 and 1000 mg/kg for 5 days/week increased the incidence of adenoma and adenocarcinoma in the uterus in the rats by 50% [251]. There was also some indication of increased incidence of testicular tumors in rats and hepatic tumors, hemangiosarcomas, and intestinal tumors in male mice [251]. Summarizing the existing data on TBBPA safety, it has an overall low acute toxicity, and adverse side effects are only seen at very high dosages over a prolonged time period [239].

Regarding the bioviability and the kinetics of TBBPA excretion in rodents and human subjects, the data clearly show that TBBPA undergoes extensive first-pass metabolism in the gastrointestinal tract and liver resulting in low bioviability of the parent TBBPA molecule [252, 253]. Similar to other bisphenols, the major metabolic pathways for TBBPA are glucuronidation and sulfonation [253], and the resulting metabolites are effectively eliminated via the bile and are found in the feces. Summarizing the quantitative measurements performed in serum, the oral intake of both single and repeated dosages in animals results in low levels of parent TBBPA. The kinetics data of TBBP, however, are not consistent. According to Schauer et al., after an oral dose of 300 mg/kg body weight of TBBPA in rats, a peak concentration of TBBPA of 103 µM was reached within 3 h, and the elimination half-life was estimated to be around 13 h [253]. Interestingly, a similar study based on a single oral dose corresponding to 250 mg/kg body weight monitored by isotope-labeled 14C-TBBPA indicated both a significantly faster decay and a lower uptake. According to that study, the maximal plasma concentration of TBBPA was only 5.4 µM. This was reached within 127 min, and the elimination half-life was around 4.5 h [254]. In a third study, maximal TBBPA plasma concentrations of 12, 18, and 31 µM were reached within 4–5 hours after oral administration of 200, 500, and 1000 mg/kg doses, respectively, in rats. The terminal half-lives were between 7.5 and 9.5 hours [255]. According to the same study, the daily administration of high oral doses of TBBPA for 14 days resulted in a cumulative effect of TBBPA in serum [255]. Based on the IC50 value from our assay, a plasma concentration of TBBPA of 5 µM would result in 48% saturation of the high-affinity binding site of TTR. Thus, inhibiting 90% of TTR would require approximately 10 µM TBBPA. Although a steady-state level can be reached in order to maintain the therapeutically relevant plasma concentration, this would still require repeated and very high dosages daily, which hampers the therapeutic potential of TBBPA in humans.

Taken together, our results show that TBBPA is a potent stabilizer of the TTR tetramer and that it displays exceptionally good selectivity in human plasma. Extensive studies in animals have confirmed that TBBPA is not linked to adverse cytotoxic effects unless very high doses are given. However, low bioviability impairs its use as a potential therapeutic drug.

Concluding remarks

ATTR affects about 50,000 people worldwide. The disease is progressive and lethal within 5 to 15 years after onset. Until the 1990s, ATTR was considered as incurable disease, and today the only approved treatment is liver transplantation. In the last two decades, several new therapies have been proposed, including kinetic stabilization of TTR by small ligands. The stabilization of TTR in plasma is challenged by the nonspecific binding of ligands to other components that are present in plasma. To determine the factors that influence the binding efficacy of TTR-stabilizing ligands, we investigated nine selected ligands both in the presence and absence of plasma. Focusing on the selectivity factor, we could show that the ΔH of the binding is a better predictor of selectivity than the KD due to its better correlation with IC50. We also believe that screening for a high enthalpic component could be a general approach in all drug design to identify ligands that are highly specific in a crowded environment.

Of our selected ligands, the flavonoid luteolin was shown to be the most efficient kinetic stabilizer of TTR in human plasma and was therefore used in further investigations. Luteolin was tested both in cell and animal models, and it showed a significant rescue effect in both cases. However, the bioavailability data showed that the compound undergoes rapid enzymatic degradation in the liver, which unfortunately impairs its potential use as a therapeutic drug. In humans, the enzymatic modification of luteolin preferentially occurs on the hydroxyl group at position 7 of the A-ring by UDP-glucuronosyltransferases. Based on the analysis of high-resolution crystal structures of luteolin in complex with TTR, we conclude that this position is crucial for binding of luteolin to the TBS. Thus, enhancing the ability of luteolin to withstand enzymatic modification in vivo while still maintaining its high selectivity might be achieved by introducing relatively small substituents to this position that can resist glucuronidation. To examine this, we created luteolin analogs containing a chlorine (7-Cl-Lut) or a methoxy group (7-MeO-Lut) at the 7 position of the A-ring. The substitutions resulted in molecules that were more stable to hepatic degradation. Moreover, both 7-Cl-Lut and 7-MeOLut displayed high affinity to TTR, but in contrast to luteolin, the binding selectivity was negatively affected in the presence of other plasma components.

Another potential kinetic stabilizer, the fire-retardant TBBPA, was also investigated in our studies. TBBPA has previously been shown to bind TTR with high affinity, and we could demonstrate that TBBPA is able to prevent TTR aggregation and to bind TTR with extremely high selectivity in human

plasma. The estimated effect for TBBPA was in the same range as the recently approved drug tafamidis and was better than diflunisal, both of which have shown therapeutic effects against FAP. We also investigated the possible role of TBBPA as a drug. Animal studies have confirmed that TBBPA is not linked to adverse effects unless very high doses are given, but its low bioviability impairs its use as a potential therapeutic drug.

Acknowledgements

Vart tog vägen 4,5 år? Är det dags att avsluta detta kapitel och börja ett nytt? Tiden gick för snabbt, och det gör den när man har det roligt.

Jag vill först och främst tacka härtligt min handledare och min mentor Anders. Stor tack Anders att du har inspirerat mig under min doktorand tid. Dina idéer som aldrig kunde ta slut och ditt charismatiska sätt att berätta dem som kunde smitta vem som helst (vi skulle jobba på att klona mig så jag skulle hinna med allt). Att jobba med dig var oerhört spännande och utvecklande på alla tänkbara sätt. Tack att du alltid hade tid och tålamod att svara på alla diverse frågor (verkligen diverse) och hjälpa med olika labbarbeten. Jag kommer verkligen att sakna ditt driv och din positiva inställning. Jag önskar dig lycka till och stor framgång i framtider.

Jag vill också tacka mina medarbetare. Det var kul att dela labb och arbetsrum med så många härliga personer. Kristoffer, vår mästerkock och drinkmaster, tack för dina idéer och kunskap inom Aβ fältet. Tack att du är en otroligt hjälpsam person och att du alltid ställde upp för mig. Tack för de fantastiska middagarna med så många otrolig goda saker och drycker. Tack för att du har lärt mig hur en riktig kräfta smakar.

Lina, min dans kompis, det var väldigt roligt att du har hittat till oss. Tack för ditt samarbete inom de TTR projekten. Jag kommer att sakna våra intressanta diskussioner både inom forskning och allmänt. Tack för din hjälp med bilderna och förståelse för min preferens för rosa färg (VS). Tack för din humor, det var mycket skratt i vårt rum. Lycka till med din resterande doktorand tid och ses på din disputation (no pressure!).

Det var mycket roligt att du kom till vår grupp Annelie, synd att du kom så ”sent” till oss för att jag skulle vilja lära mycket mer från dig. Stor tack för din hjälp med skrivningen och för de fantastiska rådgivningarna. Tack också att du tog hand om antikroppsprojektet och mina blommor under tiden när jag var borta.

Tohid, the well-organaised man, kul att du kom till vårt labb. Tack för ditt jobb inom gemensamma projekt och jag önskar dig lycka till med din doktorand tid. Enjoy!

Tack till mina briljanta studenter, mina extra händer, Alex, Sofia och Evelina det var mycket intressant, lärorikt och produktivt att jobba med er. Jag är så glad att ni har valt vårt labb. Alex, glöm inte regel nummer ett på

labbet: musiken skall vara alltid på, finns det rytmen hittar man flow. Jag ger en stor eloge för din insats i TBBPA projekt, otroligt bra jobbat. Evelina, tack för alla proteinreningar och din hjälp med diverse saker. Ni är grymma tjejer, jag hoppas att ni hittar era vägar i framtiden och lycka till med studierna!

Jag vill också tacka Pelle H för att sprida glädje i vårt labb, saknar verkligen din sjungande och ditt varmt bemötande.

Mathias P, jag skulle vilja tacka dig för din hjälp och idéer som du delade med mig. Tack Malin för att du tog hand om mig när jag kom till labbet och jag tackar dig också för alla råd som du gav till mig. Det var en stor ära att lära känna dig.

Tack till Liz och Afshan för alla X-ray data analyser och idéer, det var mycket produktivt och intressant att samarbeta med er. Liz, jag tackar dig också för hjälp med bilderna. Jag vill också tacka Anders Ö för ett intressant samarbete i TTR fält. Tack Miles T för de spännande diskussionerna och hjälp med de engelska korrektionerna.

Tack till mina rysktalande vänner Anna G, Anna S, Olena, Rafil och Eugen som alltid har varit nära under glada och svåra stunder. Stort tack för allt stöd och råd som ni gav till mig!

Tack till alla er som spelade innebandy med mig. Nasim, Andrei, Rob, Jagan, Lionel, Dominika, Istvan, det var verkligen roliga och svettiga fredagarnas morgnar. Tack till alla ”beer club medlemmar” som gjorde det också roligt under fredagarnas kvällar. Dominika (twiny), Istvan, Maite, Verena, Mikkel, Ximena, Mahsa, Jano, Martin, Parham ni kommer alltid att finnas i mitt hjärta. Jeanette, min springkompis, det var kul att lära känna dig och jag tackar dig för alla råd och att du alltid hade tid att lyssna och reflektera. Ulf, tackar dig för ditt stöd, och att du är en fantastisk lyssnare.

Anna-Karin, tack för alla skidundervisningar det var otroligt roligt att lära mig att åka skidor på rätt sätt.

Stor tack till våra administratörer, Ingrid, Clas, Anna och Jenny. Ni vet allt om hur man skall hitta sig i den administrativa djungeln, jag skulle vara totalt ”lost” om ni inte skulle finnas på institutionen.

Stor tack till Elisabeth, som tar hand om all disk och ser till att vi alltid har rena saker på platsen och att autoklaveringen är klar när det behövs. Man

saknar dig högt när du är på semester. Tack att du alltid glad när man kommer till jobbet och har stund att prata.

Jag vill också tacka mina lärare/handledare från Stockholm Universitet. Tack Lena Mäler, Jan Nedergaard, Scarlett och Natalja Petrovic för alla kunskap och träning som ni gav mig och som hjälpte mig otrolig mycket under min doktorand tid.

Jag vill också tacka hjärtligt mina närmaste vänner Jess, Frida, Olga, Dasha, Ali, Nikolaj, Peter, Katja, Anna L. för att ni gav mig otroligt mycket stöd och glädje under alla år. Jag är så glad att jag har en sådan fantastisk familj.

Och jag vill också tacka alla på MedChem för alla dagar som ni har varit i mitt liv.

References

1. Wiseman, R.L., et al., An adaptable standard for protein export from the endoplasmic reticulum. Cell, 2007. 131(4): p. 809-21.

2. Sipe, J.D. and A.S. Cohen, Review: history of the amyloid fibril. J Struct Biol, 2000. 130(2-3): p. 88-98.

3. Haass, C. and D.J. Selkoe, Soluble protein oligomers in neurodegeneration: lessons from the Alzheimer's amyloid beta- peptide. Nat Rev Mol Cell Biol, 2007. 8(2): p. 101-12.

4. Hardy, J. and D.J. Selkoe, The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science, 2002. 297(5580): p. 353-6.

5. Polymeropoulos, M.H., et al., Mutation in the alpha-synuclein gene identified in families with Parkinson's disease. Science, 1997. 276(5321): p. 2045-7.

6. Colon, W. and J.W. Kelly, Partial denaturation of transthyretin is sufficient for amyloid fibril formation in vitro. Biochemistry, 1992. 31(36): p. 8654-60.

7. Dyson, H.J. and P.E. Wright, Intrinsically unstructured proteins and their functions. Nat Rev Mol Cell Biol, 2005. 6(3): p. 197-208.

8. Booth, D.R., et al., Instability, unfolding and aggregation of human lysozyme variants underlying amyloid fibrillogenesis. Nature, 1997. 385(6619): p. 787-93.

9. Quintas, A., et al., Tetramer dissociation and monomer partial unfolding precedes protofibril formation in amyloidogenic transthyretin variants. J Biol Chem, 2001. 276(29): p. 27207-13.

10. Cardoso, I., et al., Transthyretin fibrillogenesis entails the assembly of monomers: a molecular model for in vitro assembled transthyretin amyloid-like fibrils. J Mol Biol, 2002. 317(5): p. 683- 95.

11. Redondo, C., A.M. Damas, and M.J. Saraiva, Designing transthyretin mutants affecting tetrameric structure: implications in amyloidogenicity. Biochem J, 2000. 348 Pt 1: p. 167-72.

12. Amaducci, L. and G. Tesco, Aging as a major risk for degenerative diseases of the central nervous system. Curr Opin Neurol, 1994. 7(4): p. 283-6.

13. Balch, W.E., et al., Adapting proteostasis for disease intervention. Science, 2008. 319(5865): p. 916-9.

14. Morimoto, R.I., Stress, aging, and neurodegenerative disease. N Engl J Med, 2006. 355(21): p. 2254-5.

15. Sekijima, Y., et al., The biological and chemical basis for tissue- selective amyloid disease. Cell, 2005. 121(1): p. 73-85.

16. Ferrone, F., Analysis of protein aggregation kinetics. Methods Enzymol, 1999. 309: p. 256-74.

17. Lee, J., et al., Amyloid-beta forms fibrils by nucleated conformational conversion of oligomers. Nat Chem Biol, 2011. 7(9): p. 602-9.

18. Hortschansky, P., et al., The aggregation kinetics of Alzheimer's beta-amyloid peptide is controlled by stochastic nucleation. Protein Sci, 2005. 14(7): p. 1753-9.

19. Hurshman, A.R., et al., Transthyretin aggregation under partially denaturing conditions is a downhill polymerization. Biochemistry, 2004. 43(23): p. 7365-81.

20. Kabat, E.A., D.H. Moore, and H. Landow, An Electrophoretic Study of the Protein Components in Cerebrospinal Fluid and Their Relationship to the Serum Proteins. J Clin Invest, 1942. 21(5): p. 571-7.

21. Robbins, J., J.E. Rall, and R.W. Rawson, A new serum iodine component in patients with functional carcinoma of the thyroid. J Clin Endocrinol Metab, 1955. 15(11): p. 1315-31.

22. Robbins, J. and J.E. Rall, Proteins associated with the thyroid hormones. Physiol Rev, 1960. 40: p. 415-89.

23. Soprano, D.R., et al., Demonstration of transthyretin mRNA in the brain and other extrahepatic tissues in the rat. J Biol Chem, 1985. 260(21): p. 11793-8.

24. Dickson, P.W., et al., High prealbumin and transferrin mRNA levels in the choroid plexus of rat brain. Biochem Biophys Res Commun, 1985. 127(3): p. 890-5.

25. Dickson, P.W., et al., Thyroxine transport in choroid plexus. J Biol Chem, 1987. 262(29): p. 13907-15.

26. Kanai, M., A. Raz, and D.S. Goodman, Retinol-binding protein: the transport protein for vitamin A in human plasma. J Clin Invest, 1968. 47(9): p. 2025-44.

27. Blake, C.C., et al., An x-ray study of the subunit structure of prealbumin. J Mol Biol, 1971. 61(1): p. 217-24.

28. Blake, C.C., et al., Strjcture of human plasma prealbumin at 2-5 A resolution. A preliminary report on the polypeptide chain conformation, quaternary structure and thyroxine binding. J Mol Biol, 1974. 88(1): p. 1-12.

29. Sparkes, R.S., et al., Assignment of the myelin basic protein gene to human chromosome 18q22-qter. Hum Genet, 1987. 75(2): p. 147- 50.

30. Connors, L.H., et al., Tabulation of human transthyretin (TTR) variants, 2003. Amyloid, 2003. 10(3): p. 160-84.

31. Power, D.M., et al., Evolution of the thyroid hormone-binding protein, transthyretin. Gen Comp Endocrinol, 2000. 119(3): p. 241- 55.

32. Felding, P. and G. Fex, Cellular origin of prealbumin in the rat. Biochim Biophys Acta, 1982. 716(3): p. 446-9.

33. Smith, F.R. and D.S. Goodman, The effects of diseases of the liver, thyroid, and kidneys on the transport of vitamin A in human plasma. J Clin Invest, 1971. 50(11): p. 2426-36.

34. Oppenheimer, J.H., M. Martinez, and G. Bernstein, Determination of the maximal binding capacity and protein concentration of thyroxine-binding prealbumin in human serum. J Lab Clin Med, 1966. 67(3): p. 500-9.

35. Vergani, C., R. Stabilini, and A. Agostoni, Thyroxine-binding prealbumin in thyrotoxicosis and hypothyroidism. Clin Chem, 1969. 15(3): p. 216-8.

36. Stabilini, R., et al., Influence of age and sex on prealbumin levels. Clin Chim Acta, 1968. 20(2): p. 358-9.

37. Vahlquist, A., et al., The concentrations of retinol-binding protein, prealbumin, and transferrin in the sera of newly delivered mothers and children of various ages. Scand J Clin Lab Invest, 1975. 35(6): p. 569-75.

38. Oppenheimer, J.H., Role of plasma proteins in the binding, distribution and metabolism of the thyroid hormones. N Engl J Med, 1968. 278(21): p. 1153-62.

39. Oppenheimer, J.H., et al., Binding of Thyroxine by Serum Proteins Evaluated by Equilibrum Dialysis and Electrophoretic Techniques. Alterations in Nonthyroidal Illness. J Clin Invest, 1963. 42: p. 1769- 82.

40. Bellabarba, D., et al., Thyroxine transport and turnover in major nonthyroidal illness. J Clin Endocrinol Metab, 1968. 28(7): p. 1023- 30.

41. Zhu, L., et al., Transthyretin as a novel candidate biomarker for preeclampsia. Exp Ther Med, 2014. 7(5): p. 1332-1336.

42. Beck, F.K. and T.C. Rosenthal, Prealbumin: a marker for nutritional evaluation. Am Fam Physician, 2002. 65(8): p. 1575-8.

43. Haider, M. and S.Q. Haider, Assessment of protein-calorie malnutrition. Clin Chem, 1984. 30(8): p. 1286-99.

44. Aleshire, S.L., et al., Localization of human prealbumin in choroid plexus epithelium. J Histochem Cytochem, 1983. 31(5): p. 608-12.

45. Aldred, A.R., C.M. Brack, and G. Schreiber, The cerebral expression of plasma protein genes in different species. Comp Biochem Physiol B Biochem Mol Biol, 1995. 111(1): p. 1-15.

46. Herbert, J., et al., Transthyretin: a choroid plexus-specific transport protein in human brain. The 1986 S. Weir Mitchell award. Neurology, 1986. 36(7): p. 900-11.

47. Pfeffer, B.A., et al., Expression of transthyretin and retinol binding protein mRNAs and secretion of transthyretin by cultured monkey retinal pigment epithelium. Mol Vis, 2004. 10: p. 23-30.

48. Kato, M., et al., Plasma and cellular retinoid-binding proteins and transthyretin (prealbumin) are all localized in the islets of Langerhans in the rat. Proc Natl Acad Sci U S A, 1985. 82(8): p. 2488-92.

49. Jacobsson, B., et al., Transthyretin immunoreactivity in human and porcine liver, choroid plexus, and pancreatic islets. J Histochem Cytochem, 1989. 37(1): p. 31-7.

50. McKinnon, B., et al., Synthesis of thyroid hormone binding proteins transthyretin and albumin by human trophoblast. J Clin Endocrinol Metab, 2005. 90(12): p. 6714-20.

51. Socolow, E.L., et al., Preparation of I-131-labeled human serum prealbumin and its metabolism in normal and sick patients. J Clin Invest, 1965. 44(10): p. 1600-9.

52. Makover, A., et al., Plasma transthyretin. Tissue sites of degradation and turnover in the rat. J Biol Chem, 1988. 263(18): p. 8598-603.

53. Sousa, M.M. and M.J. Saraiva, Internalization of transthyretin. Evidence of a novel yet unidentified receptor-associated protein (RAP)-sensitive receptor. J Biol Chem, 2001. 276(17): p. 14420-5.

54. Sousa, M.M., et al., Evidence for the role of megalin in renal uptake of transthyretin. J Biol Chem, 2000. 275(49): p. 38176-81.

55. Blake, C.C., et al., Structure of prealbumin: secondary, tertiary and quaternary interactions determined by Fourier refinement at 1.8 A. J Mol Biol, 1978. 121(3): p. 339-56.

56. Monaco, H.L., M. Rizzi, and A. Coda, Structure of a complex of two plasma proteins: transthyretin and retinol-binding protein. Science, 1995. 268(5213): p. 1039-41.

57. Naylor, H.M. and M.E. Newcomer, The structure of human retinol- binding protein (RBP) with its carrier protein transthyretin reveals an interaction with the carboxy terminus of RBP. Biochemistry, 1999. 38(9): p. 2647-53.

58. Johnson, S.M., et al., Native state kinetic stabilization as a strategy to ameliorate protein misfolding diseases: a focus on the transthyretin amyloidoses. Acc Chem Res, 2005. 38(12): p. 911-21.

59. Wojtczak, A., et al., Structures of human transthyretin complexed with thyroxine at 2.0 A resolution and 3',5'-dinitro-N-acetyl-L- thyronine at 2.2 A resolution. Acta Crystallogr D Biol Crystallogr, 1996. 52(Pt 4): p. 758-65.

60. Hornberg, A., et al., A comparative analysis of 23 structures of the amyloidogenic protein transthyretin. J Mol Biol, 2000. 302(3): p. 649-69.

61. Ferguson, R.N., et al., Negative cooperativity in the binding of thyroxine to human serum prealbumin. Preparation of tritium- labeled 8-anilino-1-naphthalenesulfonic acid. Biochemistry, 1975. 14(2): p. 282-9.

62. Liz, M.A., et al., Transthyretin is a metallopeptidase with an inducible active site. Biochem J, 2012. 443(3): p. 769-78.

63. Cascella, R., et al., Transthyretin suppresses the toxicity of oligomers formed by misfolded proteins in vitro. Biochim Biophys Acta, 2013. 1832(12): p. 2302-14.

64. Liz, M.A., et al., Transthyretin, a new cryptic protease. J Biol Chem, 2004. 279(20): p. 21431-8.

65. Costa, R., et al., Transthyretin protects against A-beta peptide toxicity by proteolytic cleavage of the peptide: a mechanism sensitive to the Kunitz protease inhibitor. PLoS One, 2008. 3(8): p. e2899.

66. Li, X., et al., Mechanisms of transthyretin inhibition of beta- amyloid aggregation in vitro. J Neurosci, 2013. 33(50): p. 19423- 33.

67. Robbins, J., et al., Thyroxine transport proteins of plasma. Molecular properties and biosynthesis. Recent Prog Horm Res, 1978. 34: p. 477-519.

68. Palha, J.A., et al., Thyroid hormone metabolism in a transthyretin- null mouse strain. J Biol Chem, 1994. 269(52): p. 33135-9.

69. Chanoine, J.P. and L.E. Braverman, The role of transthyretin in the transport of thyroid hormone to cerebrospinal fluid and brain. Acta Med Austriaca, 1992. 19 Suppl 1: p. 25-8.

70. Landers, K.A., R.H. Mortimer, and K. Richard, Transthyretin and the human placenta. Placenta, 2013. 34(7): p. 513-7.

71. Gudas, L.J., Emerging roles for retinoids in regeneration and differentiation in normal and disease states. Biochim Biophys Acta, 2012. 1821(1): p. 213-21.

72. Ono, K. and M. Yamada, Vitamin A and Alzheimer's disease. Geriatr Gerontol Int, 2012. 12(2): p. 180-8.

73. Raz, A. and D.S. Goodman, The interaction of thyroxine with human plasma prealbumin and with the prealbumin-retinol- binding protein complex. J Biol Chem, 1969. 244(12): p. 3230-7.

74. Raz, A., T. Shiratori, and D.S. Goodman, Studies on the protein- protein and protein-ligand interactions involved in retinol transport in plasma. J Biol Chem, 1970. 245(8): p. 1903-12.

75. Goodman, D.S., Vitamin A and retinoids in health and disease. N Engl J Med, 1984. 310(16): p. 1023-31.

76. Noy, N., E. Slosberg, and S. Scarlata, Interactions of retinol with binding proteins: studies with retinol-binding protein and with transthyretin. Biochemistry, 1992. 31(45): p. 11118-24.

77. van Bennekum, A.M., et al., Biochemical basis for depressed serum retinol levels in transthyretin-deficient mice. J Biol Chem, 2001. 276(2): p. 1107-13.

78. Episkopou, V., et al., Disruption of the transthyretin gene results in mice with depressed levels of plasma retinol and thyroid hormone. Proc Natl Acad Sci U S A, 1993. 90(6): p. 2375-9.

79. Shirahama, T., et al., Senile cerebral amyloid. Prealbumin as a common constituent in the neuritic plaque, in the neurofibrillary tangle, and in the microangiopathic lesion. Am J Pathol, 1982. 107(1): p. 41-50.

80. Wisniewski, T., et al., Cerebrospinal fluid inhibits Alzheimer beta- amyloid fibril formation in vitro. Ann Neurol, 1993. 34(4): p. 631-3.

81. Schwarzman, A.L., et al., Transthyretin sequesters amyloid beta protein and prevents amyloid formation. Proc Natl Acad Sci U S A, 1994. 91(18): p. 8368-72.

82. Liu, L. and R.M. Murphy, Kinetics of inhibition of beta-amyloid aggregation by transthyretin. Biochemistry, 2006. 45(51): p. 15702-9.

83. Costa, R., et al., Transthyretin binding to A-Beta peptide--impact on A-Beta fibrillogenesis and toxicity. FEBS Lett, 2008. 582(6): p. 936-42.

84. Du, J. and R.M. Murphy, Characterization of the interaction of beta-amyloid with transthyretin monomers and tetramers. Biochemistry, 2010. 49(38): p. 8276-89.

85. Buxbaum, J.N., et al., Transthyretin protects Alzheimer's mice from the behavioral and biochemical effects of Abeta toxicity. Proc Natl Acad Sci U S A, 2008. 105(7): p. 2681-6.

86. Li, X., et al., Transthyretin Suppresses Amyloid-beta Secretion by Interfering with Processing of the Amyloid-beta Protein Precursor. J Alzheimers Dis, 2016.

87. Tsuzuki, K., et al., Transthyretin binds amyloid beta peptides, Abeta1-42 and Abeta1-40 to form complex in the autopsied human kidney - possible role of transthyretin for abeta sequestration. Neurosci Lett, 2000. 281(2-3): p. 171-4.

88. Askanas, V., et al., Inclusion body myositis, muscle blood vessel and cardiac amyloidosis, and transthyretin Val122Ile allele. Ann Neurol, 2000. 47(4): p. 544-9.

89. Choi, S.H., et al., Accelerated Abeta deposition in APPswe/PS1deltaE9 mice with hemizygous deletions of TTR (transthyretin). J Neurosci, 2007. 27(26): p. 7006-10.

90. Doggui, S., et al., Possible involvement of transthyretin in hippocampal beta-amyloid burden and learning behaviors in a mouse model of Alzheimer's disease (TgCRND8). Neurodegener Dis, 2010. 7(1-3): p. 88-95.

91. Wati, H., et al., Transthyretin accelerates vascular Abeta deposition in a mouse model of Alzheimer's disease. Brain Pathol, 2009. 19(1): p. 48-57.

92. Castano, E.M., et al., Comparative proteomics of cerebrospinal fluid in neuropathologically-confirmed Alzheimer's disease and non- demented elderly subjects. Neurol Res, 2006. 28(2): p. 155-63.

93. Gloeckner, S.F., et al., Quantitative analysis of transthyretin, tau and amyloid-beta in patients with dementia. J Alzheimers Dis, 2008. 14(1): p. 17-25.

94. Hansson, S.F., et al., Reduced levels of amyloid-beta-binding proteins in cerebrospinal fluid from Alzheimer's disease patients. J Alzheimers Dis, 2009. 16(2): p. 389-97.

95. Serot, J.M., et al., Cerebrospinal fluid transthyretin: aging and late onset Alzheimer's disease. J Neurol Neurosurg Psychiatry, 1997. 63(4): p. 506-8.

96. Li, X., et al., Neuronal production of transthyretin in human and murine Alzheimer's disease: is it protective? J Neurosci, 2011. 31(35): p. 12483-90.

97. Stein, T.D., et al., Neutralization of transthyretin reverses the neuroprotective effects of secreted amyloid precursor protein (APP) in APPSW mice resulting in tau phosphorylation and loss of hippocampal neurons: support for the amyloid hypothesis. J Neurosci, 2004. 24(35): p. 7707-17.

98. Alemi, M., et al., Transthyretin participates in beta-amyloid transport from the brain to the liver- involvement of the low- density lipoprotein receptor-related protein 1? Sci Rep, 2016. 6: p. 20164.

99. Sousa, M.M., L. Berglund, and M.J. Saraiva, Transthyretin in high density lipoproteins: association with apolipoprotein A-I. J Lipid Res, 2000. 41(1): p. 58-65.

100. Liz, M.A., et al., Substrate specificity of transthyretin: identification of natural substrates in the nervous system. Biochem J, 2009. 419(2): p. 467-74.

101. Palha, J.A., Transthyretin as a thyroid hormone carrier: function revisited. Clin Chem Lab Med, 2002. 40(12): p. 1292-300.

102. Buxbaum, J.N. and N. Reixach, Transthyretin: the servant of many masters. Cell Mol Life Sci, 2009. 66(19): p. 3095-101.

103. Johnson, S.M., et al., The transthyretin amyloidoses: from delineating the molecular mechanism of aggregation linked to pathology to a regulatory-agency-approved drug. J Mol Biol, 2012. 421(2-3): p. 185-203.

104. Darnerud, P.O., et al., Binding of a 3,3', 4,4'-tetrachlorobiphenyl (CB-77) metabolite to fetal transthyretin and effects on fetal thyroid hormone levels in mice. Toxicology, 1996. 106(1-3): p. 105-14.

105. Brouwer, A. and K.J. van den Berg, Binding of a metabolite of 3,4,3',4'-tetrachlorobiphenyl to transthyretin reduces serum vitamin A transport by inhibiting the formation of the protein complex carrying both retinol and thyroxin. Toxicol Appl Pharmacol, 1986. 85(3): p. 301-12.

106. Boomsma, F., A.J. Man in 't Veld, and M.A. Schalekamp, Not norepinephrine but its oxidation products bind specifically to plasma proteins. J Pharmacol Exp Ther, 1991. 259(2): p. 551-7.

107. Baures, P.W., S.A. Peterson, and J.W. Kelly, Discovering transthyretin amyloid fibril inhibitors by limited screening. Bioorg Med Chem, 1998. 6(8): p. 1389-401.

108. Hurshman Babbes, A.R., E.T. Powers, and J.W. Kelly, Quantification of the thermodynamically linked quaternary and tertiary structural stabilities of transthyretin and its disease- associated variants: the relationship between stability and amyloidosis. Biochemistry, 2008. 47(26): p. 6969-84.

109. Hammarstrom, P., et al., D18G transthyretin is monomeric, aggregation prone, and not detectable in plasma and cerebrospinal fluid: a prescription for central nervous system amyloidosis? Biochemistry, 2003. 42(22): p. 6656-63.

110. Sekijima, Y., et al., Energetic characteristics of the new transthyretin variant A25T may explain its atypical central nervous system pathology. Lab Invest, 2003. 83(3): p. 409-17.

111. Bergstrom, J., et al., Amyloid deposits in transthyretin-derived amyloidosis: cleaved transthyretin is associated with distinct amyloid morphology. J Pathol, 2005. 206(2): p. 224-32.

112. Ihse, E., et al., Amyloid fibrils containing fragmented ATTR may be the standard fibril composition in ATTR amyloidosis. Amyloid, 2013. 20(3): p. 142-50.

113. Thylen, C., et al., Modifications of transthyretin in amyloid fibrils: analysis of amyloid from homozygous and heterozygous individuals with the Met30 mutation. EMBO J, 1993. 12(2): p. 743- 8.

114. Mangione, P.P., et al., Proteolytic cleavage of Ser52Pro variant transthyretin triggers its amyloid fibrillogenesis. Proc Natl Acad Sci U S A, 2014. 111(4): p. 1539-44.

115. Ueda, M., et al. Fragmentations of TTR in cultured cells (XIV International Symposium on Amyloidosis, OP-57, Indianapolis, Indiana, USA). 2014; Available from: http://www.amyloidosis.nl/indpls/XIVth International Symposium on Amyloidosis Book.pdf.

116. Felding, P. and G. Fex, The increased negative charge of prealbumin in cerebrospinal fluid is acquired in vitro by oxidation of the cysteinyl residues without formation of disulphides. Scand J Clin Lab Invest, 1984. 44(3): p. 231-8.

117. Altland, K., et al., Sulfite and base for the treatment of familial amyloidotic polyneuropathy: two additive approaches to stabilize the conformation of human amyloidogenic transthyretin. Neurogenetics, 2004. 5(1): p. 61-7.

118. Zhang, Q. and J.W. Kelly, Cys10 mixed disulfides make transthyretin more amyloidogenic under mildly acidic conditions. Biochemistry, 2003. 42(29): p. 8756-61.

119. Jiang, X., et al., An engineered transthyretin monomer that is nonamyloidogenic, unless it is partially denatured. Biochemistry, 2001. 40(38): p. 11442-52.

120. Bourgault, S., et al., Sulfated glycosaminoglycans accelerate transthyretin amyloidogenesis by quaternary structural conversion. Biochemistry, 2011. 50(6): p. 1001-15.

121. Lambert, M.P., et al., Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proc Natl Acad Sci U S A, 1998. 95(11): p. 6448-53.

122. Walsh, D.M., et al., The role of cell-derived oligomers of Abeta in Alzheimer's disease and avenues for therapeutic intervention. Biochem Soc Trans, 2005. 33(Pt 5): p. 1087-90.

123. Hou, X., et al., Transthyretin oligomers induce calcium influx via voltage-gated calcium channels. J Neurochem, 2007. 100(2): p. 446-57.

124. Reixach, N., et al., Tissue damage in the amyloidoses: Transthyretin monomers and nonnative oligomers are the major cytotoxic species in tissue culture. Proc Natl Acad Sci U S A, 2004. 101(9): p. 2817-22.

125. Sorgjerd, K., et al., Prefibrillar transthyretin oligomers and cold stored native tetrameric transthyretin are cytotoxic in cell culture. Biochem Biophys Res Commun, 2008. 377(4): p. 1072-8.

126. Sekijima, Y., Transthyretin (ATTR) amyloidosis: clinical spectrum, molecular pathogenesis and disease-modifying treatments. J Neurol Neurosurg Psychiatry, 2015. 86(9): p. 1036-43.

127. Demuro, A., M. Smith, and I. Parker, Single-channel Ca(2+) imaging implicates Abeta1-42 amyloid pores in Alzheimer's disease pathology. J Cell Biol, 2011. 195(3): p. 515-24.

128. Chen, M., et al., An investigation into the formation of annular aggregates of human islet amyloid polypeptide on tantalum oxide surfaces. Chemistry, 2012. 18(9): p. 2493-7.

129. Lashuel, H.A., et al., Alpha-synuclein, especially the Parkinson's disease-associated mutants, forms pore-like annular and tubular protofibrils. J Mol Biol, 2002. 322(5): p. 1089-102.

130. Wacker, J.L., et al., Hsp70 and Hsp40 attenuate formation of spherical and annular polyglutamine oligomers by partitioning monomer. Nat Struct Mol Biol, 2004. 11(12): p. 1215-22.

131. Lasagna-Reeves, C.A., C.G. Glabe, and R. Kayed, Amyloid-beta annular protofibrils evade fibrillar fate in Alzheimer disease brain. J Biol Chem, 2011. 286(25): p. 22122-30.

132. Pires, R.H., et al., Distinct annular oligomers captured along the assembly and disassembly pathways of transthyretin amyloid protofibrils. PLoS One, 2012. 7(9): p. e44992.

133. Hou, X., et al., Binding of amyloidogenic transthyretin to the plasma membrane alters membrane fluidity and induces neurotoxicity. Biochemistry, 2005. 44(34): p. 11618-27.

134. Monteiro, F.A., et al., In vitro inhibition of transthyretin aggregate- induced cytotoxicity by full and peptide derived forms of the soluble receptor for advanced glycation end products (RAGE). FEBS Lett, 2006. 580(14): p. 3451-6.

135. Monteiro, F.A., et al., Activation of ERK1/2 MAP kinases in familial amyloidotic polyneuropathy. J Neurochem, 2006. 97(1): p. 151-61.

136. Sousa, M.M., et al., Familial amyloid polyneuropathy: receptor for advanced glycation end products-dependent triggering of neuronal inflammatory and apoptotic pathways. J Neurosci, 2001. 21(19): p. 7576-86.

137. Zhao, L., J.N. Buxbaum, and N. Reixach, Age-related oxidative modifications of transthyretin modulate its amyloidogenicity. Biochemistry, 2013. 52(11): p. 1913-26.

138. Cornwell, G.G., 3rd, et al., Frequency and distribution of senile cardiovascular amyloid. A clinicopathologic correlation. Am J Med, 1983. 75(4): p. 618-23.

139. Kyle, R.A., et al., The premortem recognition of systemic senile amyloidosis with cardiac involvement. Am J Med, 1996. 101(4): p. 395-400.

140. Ikeda, S., et al., Familial transthyretin-type amyloid polyneuropathy in Japan: clinical and genetic heterogeneity. Neurology, 2002. 58(7): p. 1001-7.

141. Ng, B., et al., Senile systemic amyloidosis presenting with heart failure: a comparison with light chain-associated amyloidosis. Arch Intern Med, 2005. 165(12): p. 1425-9.

142. Westermark, P., et al., Transthyretin-derived senile systemic amyloidosis: clinicopathologic and structural considerations. Amyloid, 2003. 10 Suppl 1: p. 48-54.

143. Cowan, A.J., et al., Amyloidosis of the gastrointestinal tract: a 13- year, single-center, referral experience. Haematologica, 2013. 98(1): p. 141-6.

144. Pitkanen, P., P. Westermark, and G.G. Cornwell, 3rd, Senile systemic amyloidosis. Am J Pathol, 1984. 117(3): p. 391-9.

145. Sueyoshi, T., et al., Spinal multifocal amyloidosis derived from wild-type transthyretin. Amyloid, 2011. 18(3): p. 165-8.

146. Ueda, M., et al., Clinicopathological features of senile systemic amyloidosis: an ante- and post-mortem study. Mod Pathol, 2011. 24(12): p. 1533-44.

147. Goncalves, I., et al., Transthyretin is up-regulated by sex hormones in mice liver. Mol Cell Biochem, 2008. 317(1-2): p. 137-42.

148. Okamoto, S., et al., Liver transplantation for familial amyloidotic polyneuropathy: impact on Swedish patients' survival. Liver Transpl, 2009. 15(10): p. 1229-35.

149. Rowczenio D, W.A. Mutations in hereditary amyloidosis. 2010; Available from: http://amyloidosismutations.com/mut-attr.php.

150. Kristen, A., et al. Cardiac findings in TTR amyloidosis: A report from the Transthyretin Amyloid Outcome

Survey (THAOS). XIVth International Symposium on Amyloidosis. PC-27 Indianapolis, Indiana, USA 2014; Available from: http://www.amyloidosis.nl/indpls/XIVth International Symposium on Amyloidosis Book.pdf.

151. Andrade, C., A peculiar form of peripheral neuropathy; familiar atypical generalized amyloidosis with special involvement of the peripheral nerves. Brain, 1952. 75(3): p. 408-27.

152. Saraiva, M.J., et al., Family studies of the genetic abnormality in transthyretin (prealbumin) in Portuguese patients with familial amyloidotic polyneuropathy. Ann N Y Acad Sci, 1984. 435: p. 86- 100.

153. Ando, Y., S. Araki, and M. Ando, Transthyretin and familial amyloidotic polyneuropathy. Intern Med, 1993. 32(12): p. 920-2.

154. Saraiva, M.J., et al., Amyloid fibril protein in familial amyloidotic polyneuropathy, Portuguese type. Definition of molecular abnormality in transthyretin (prealbumin). J Clin Invest, 1984. 74(1): p. 104-19.

155. Suhr, O.B., et al., Hereditary transthyretin amyloidosis from a Scandinavian perspective. J Intern Med, 2003. 254(3): p. 225-35.

156. Holt, I.J., et al., Molecular genetics of amyloid neuropathy in Europe. Lancet, 1989. 1(8637): p. 524-6.

157. Reilly, M.M., et al., Transthyretin gene analysis in European patients with suspected familial amyloid polyneuropathy. Brain, 1995. 118 ( Pt 4): p. 849-56.

158. Benson, M.D., et al., Hereditary amyloidosis: description of a new American kindred with late onset cardiomyopathy. Appalachian amyloid. Arthritis Rheum, 1987. 30(2): p. 195-200.

159. Palacios, S.A., et al., Familial amyloidotic polyneuropathy type 1 in Brazil is associated with the transthyretin Val30Met variant. Amyloid, 1999. 6(4): p. 289-91.

160. Sousa, A., et al., Genetic epidemiology of familial amyloidotic polyneuropathy (FAP)-type I in Povoa do Varzim and Vila do Conde (north of Portugal). Am J Med Genet, 1995. 60(6): p. 512-21.

161. Koike, H., et al., Natural history of transthyretin Val30Met familial amyloid polyneuropathy: analysis of late-onset cases from non- endemic areas. J Neurol Neurosurg Psychiatry, 2012. 83(2): p. 152- 8.

162. Jacobson, D.R., et al., Variant-sequence transthyretin (isoleucine 122) in late-onset cardiac amyloidosis in black Americans. N Engl J Med, 1997. 336(7): p. 466-73.

163. Yamashita, T., et al., A prospective evaluation of the transthyretin Ile122 allele frequency in an African-American population. Amyloid, 2005. 12(2): p. 127-30.

164. Jiang, X., J.N. Buxbaum, and J.W. Kelly, The V122I cardiomyopathy variant of transthyretin increases the velocity of rate-limiting tetramer dissociation, resulting in accelerated amyloidosis. Proc Natl Acad Sci U S A, 2001. 98(26): p. 14943-8.

165. Holmgren, G., et al., Clinical improvement and amyloid regression after liver transplantation in hereditary transthyretin amyloidosis. Lancet, 1993. 341(8853): p. 1113-6.

166. Holmgren, G., et al., Biochemical effect of liver transplantation in two Swedish patients with familial amyloidotic polyneuropathy (FAP-met30). Clin Genet, 1991. 40(3): p. 242-6.

167. Benson, M.D., Liver transplantation and transthyretin amyloidosis. Muscle Nerve, 2013. 47(2): p. 157-62.

168. Suhr, O.B., et al., Liver transplantation for hereditary transthyretin amyloidosis. Liver Transpl, 2000. 6(3): p. 263-76.

169. Tsuchiya, A., et al., Marked regression of abdominal fat amyloid in patients with familial amyloid polyneuropathy during long-term follow-up after liver transplantation. Liver Transpl, 2008. 14(4): p. 563-70.

170. Wilczek, H.E., et al., Long-term data from the Familial Amyloidotic Polyneuropathy World Transplant Registry (FAPWTR). Amyloid, 2011. 18 Suppl 1: p. 193-5.

171. Herlenius, G., et al., Ten years of international experience with liver transplantation for familial amyloidotic polyneuropathy: results from the Familial Amyloidotic Polyneuropathy World Transplant Registry. Transplantation, 2004. 77(1): p. 64-71.

172. Westermark, P., et al., Fibril in senile systemic amyloidosis is derived from normal transthyretin. Proc Natl Acad Sci U S A, 1990. 87(7): p. 2843-5.

173. Yazaki, M., et al., Cardiac amyloid in patients with familial amyloid polyneuropathy consists of abundant wild-type transthyretin. Biochem Biophys Res Commun, 2000. 274(3): p. 702-6.

174. Yazaki, M., et al., Progressive wild-type transthyretin deposition after liver transplantation preferentially occurs onto myocardium in FAP patients. Am J Transplant, 2007. 7(1): p. 235-42.

175. Carvalho, A., A. Rocha, and L. Lobato, Liver transplantation in transthyretin amyloidosis: issues and challenges. Liver Transpl, 2015. 21(3): p. 282-92.

176. Bennett, C.F. and E.E. Swayze, RNA targeting therapeutics: molecular mechanisms of antisense oligonucleotides as a therapeutic platform. Annu Rev Pharmacol Toxicol, 2010. 50: p. 259-93.

177. Coelho, T., et al., Safety and efficacy of RNAi therapy for transthyretin amyloidosis. N Engl J Med, 2013. 369(9): p. 819-29.

178. Benson, M.D., et al., Targeted suppression of an amyloidogenic transthyretin with antisense oligonucleotides. Muscle Nerve, 2006. 33(5): p. 609-18.

179. Kurosawa, T., et al., Selective silencing of a mutant transthyretin allele by small interfering RNAs. Biochem Biophys Res Commun, 2005. 337(3): p. 1012-8.

180. Suhr, O.B., et al., Efficacy and safety of patisiran for familial amyloidotic polyneuropathy: a phase II multi-dose study. Orphanet J Rare Dis, 2015. 10: p. 109.

181. Dubrey, S., E. Ackermann, and J. Gillmore, The transthyretin amyloidoses: advances in therapy. Postgrad Med J, 2015. 91(1078): p. 439-48.

182. Ackermann, E.J., et al., Clinical development of an antisense therapy for the treatment of transthyretin-associated polyneuropathy. Amyloid, 2012. 19 Suppl 1: p. 43-4.

183. Coehlo, T., et al., A strikingly benign evolution of FAP in an individual found to be a compound heterozygote for two TTR mutations: TTR MET 30 and TTR MET 119. J. Rheumatol, 1993. 20: p. 179.

184. Coehlo, T., et al., Compound heterozygotes of transthyretin Met30 and transthyretin Met119 are protected from the devastating effects of familial amyloid polyneuropathy. Neuromusc. Disord, 1996. 6: p. 27.

185. Hammarstrom, P., F. Schneider, and J.W. Kelly, Trans-suppression of misfolding in an amyloid disease. Science, 2001. 293(5539): p. 2459-62.

186. Hammarstrom, P., et al., Prevention of transthyretin amyloid disease by changing protein misfolding energetics. Science, 2003. 299(5607): p. 713-6.

187. Miroy, G.J., et al., Inhibiting transthyretin amyloid fibril formation via protein stabilization. Proc Natl Acad Sci U S A, 1996. 93(26): p. 15051-6.

188. Bulawa, C.E., et al., Tafamidis, a potent and selective transthyretin kinetic stabilizer that inhibits the amyloid cascade. Proc Natl Acad Sci U S A, 2012. 109(24): p. 9629-34.

189. Berk, J.L., et al., Repurposing diflunisal for familial amyloid polyneuropathy: a randomized clinical trial. JAMA, 2013. 310(24): p. 2658-67.

190. Coelho, T., et al., Tafamidis for transthyretin familial amyloid polyneuropathy: a randomized, controlled trial. Neurology, 2012. 79(8): p. 785-92.

191. Rappley, I., et al., Quantification of transthyretin kinetic stability in human plasma using subunit exchange. Biochemistry, 2014. 53(12): p. 1993-2006.

192. Nencetti, S., A. Rossello, and E. Orlandini, Tafamidis (Vyndaqel): A Light for FAP Patients. ChemMedChem, 2013.

193. Waddington Cruz, M. and M.D. Benson, A Review of Tafamidis for the Treatment of Transthyretin-Related Amyloidosis. Neurol Ther, 2015. 4(2): p. 61-79.

194. European Medicines Agency, Vyndaqel, Assessment report. 2011.

195. Coelho, T., et al., Long-term effects of tafamidis for the treatment of transthyretin familial amyloid polyneuropathy. J Neurol, 2013. 260(11): p. 2802-14.

196. Lozeron, P., et al., Effect on disability and safety of Tafamidis in late onset of Met30 transthyretin familial amyloid polyneuropathy. Eur J Neurol, 2013. 20(12): p. 1539-45.

197. Miller, S.R., Y. Sekijima, and J.W. Kelly, Native state stabilization by NSAIDs inhibits transthyretin amyloidogenesis from the most common familial disease variants. Lab Invest, 2004. 84(5): p. 545- 52.

198. Sekijima, Y., M.A. Dendle, and J.W. Kelly, Orally administered diflunisal stabilizes transthyretin against dissociation required for amyloidogenesis. Amyloid, 2006. 13(4): p. 236-49.

199. Tojo, K., et al., Diflunisal stabilizes familial amyloid polyneuropathy-associated transthyretin variant tetramers in serum against dissociation required for amyloidogenesis. Neurosci Res, 2006. 56(4): p. 441-9.

200. Turner, R.A., R.W. Shackleford, and J.P. Whipple, Comparison of diflunisal and aspirin in long-term treatment of patients with rheumatoid arthritis. Clin Ther, 1986. 9 Suppl C: p. 37-46.

201. Cardoso, I. and M.J. Saraiva, Doxycycline disrupts transthyretin amyloid: evidence from studies in a FAP transgenic mice model. FASEB J, 2006. 20(2): p. 234-9.

202. Obici, L., et al., Doxycycline plus tauroursodeoxycholic acid for transthyretin amyloidosis: a phase II study. Amyloid, 2012. 19 Suppl 1: p. 34-6.

203. Iakovleva, I., et al., Enthalpic Forces Correlate with the Selectivity of Transthyretin-Stabilizing Ligands in Human Plasma. J Med Chem, 2015. 58(16): p. 6507-15.

204. Kelly, J.W., et al., Transthyretin quaternary and tertiary structural changes facilitate misassembly into amyloid. Adv Protein Chem, 1997. 50: p. 161-81.

205. Klabunde, T., et al., Rational design of potent human transthyretin amyloid disease inhibitors. Nat Struct Biol, 2000. 7(4): p. 312-21.

206. Mairal, T., et al., Iodine atoms: a new molecular feature for the design of potent transthyretin fibrillogenesis inhibitors. PLoS One, 2009. 4(1): p. e4124.

207. Ferreira, N., M.J. Saraiva, and M.R. Almeida, Natural polyphenols inhibit different steps of the process of transthyretin (TTR) amyloid fibril formation. FEBS Lett, 2011. 585(15): p. 2424-30.

208. Ferreira, N., M.J. Saraiva, and M.R. Almeida, Natural polyphenols as modulators of TTR amyloidogenesis: in vitro and in vivo evidences towards therapy. Amyloid, 2012. 19 Suppl 1: p. 39-42.

209. Ferreira, N., et al., Dietary curcumin counteracts extracellular transthyretin deposition: insights on the mechanism of amyloid inhibition. Biochim Biophys Acta, 2013. 1832(1): p. 39-45.

210. Sant, et al., Inhibition of Human Transthyretin Aggregation by Non-Steroidal Anti-Inflammatory Compounds: A Structural and Thermodynamic Analysis. International Journal of Molecular Sciences, 2013. 14(3): p. 5284-5311.

211. Cotrina, E.Y., et al., Modulation of the fibrillogenesis inhibition properties of two transthyretin ligands by halogenation. J Med Chem, 2013. 56(22): p. 9110-21.

212. Vilaro, M., et al., Tuning transthyretin amyloidosis inhibition properties of iododiflunisal by combinatorial engineering of the nonsalicylic ring substitutions. ACS Comb Sci, 2015. 17(1): p. 32-8.

213. Sant'anna, R.O., et al., Inhibition of human transthyretin aggregation by non-steroidal anti-inflammatory compounds: a structural and thermodynamic analysis. Int J Mol Sci, 2013. 14(3): p. 5284-311.

214. Trivella, D.B., et al., Flavonoid interactions with human transthyretin: combined structural and thermodynamic analysis. J Struct Biol, 2012. 180(1): p. 143-53.

215. Trivella, D.B., et al., The binding of synthetic triiodo l-thyronine analogs to human transthyretin: molecular basis of cooperative and non-cooperative ligand recognition. J Struct Biol, 2011. 173(2): p. 323-32.

216. Alhamadsheh, M.M., et al., Potent kinetic stabilizers that prevent transthyretin-mediated cardiomyocyte proteotoxicity. Sci Transl Med, 2011. 3(97): p. 97ra81.

217. Penchala, S.C., et al., AG10 inhibits amyloidogenesis and cellular toxicity of the familial amyloid cardiomyopathy-associated V122I transthyretin. Proc Natl Acad Sci U S A, 2013. 110(24): p. 9992-7.

218. Munro, S.L., et al., Drug competition for thyroxine binding to transthyretin (prealbumin): comparison with effects on thyroxine- binding globulin. J Clin Endocrinol Metab, 1989. 68(6): p. 1141-7.

219. Baures, P.W., et al., Synthesis and evaluation of inhibitors of transthyretin amyloid formation based on the non-steroidal anti- inflammatory drug, flufenamic acid. Bioorg Med Chem, 1999. 7(7): p. 1339-47.

220. Cheng, S.Y., et al., Analysis of thyroid hormone binding to human serum prealbumin by 8-anilinonaphthalene-1-sulfonate fluorescence. Biochemistry, 1977. 16(16): p. 3707-13.

221. Maetzler, W., et al., Serum and cerebrospinal fluid levels of transthyretin in Lewy body disorders with and without dementia. PLoS One, 2012. 7(10): p. e48042.

222. Kawasaki, Y. and E. Freire, Finding a better path to drug selectivity. Drug Discov Today, 2011. 16(21-22): p. 985-90.

223. Addison, R.S., et al., Steady-state dispositions of valproate and diflunisal alone and coadministered to healthy volunteers. Eur J Clin Pharmacol, 2000. 56(9-10): p. 715-21.

224. Bastianetto, S., et al., Possible involvement of programmed cell death pathways in the neuroprotective action of polyphenols. Curr Alzheimer Res, 2011. 8(5): p. 445-51.

225. Dajas, F., et al., Neuroprotective actions of flavones and flavonols: mechanisms and relationship to flavonoid structural features. Cent Nerv Syst Agents Med Chem, 2013. 13(1): p. 30-5.

226. Taliou, A., et al., An open-label pilot study of a formulation containing the anti-inflammatory flavonoid luteolin and its effects on behavior in children with autism spectrum disorders. Clin Ther, 2013. 35(5): p. 592-602.

227. Reixach, N., et al., Cell based screening of inhibitors of transthyretin aggregation. Biochem Biophys Res Commun, 2006. 348(3): p. 889-97.

228. Pokrzywa, M., et al., Misfolded transthyretin causes behavioral changes in a Drosophila model for transthyretin-associated amyloidosis. Eur J Neurosci, 2007. 26(4): p. 913-24.

229. Lu, X., et al., Evaluation of hepatic clearance and drug-drug interactions of luteolin and apigenin by using primary cultured rat hepatocytes. Pharmazie, 2011. 66(8): p. 600-5.

230. Buchheit, D., et al., S-Glucuronidation of 7-mercapto-4- methylcoumarin by human UDP glycosyltransferases in genetically engineered fission yeast cells. Biol Chem, 2011. 392(12): p. 1089-95.

231. Kaivosaari, S., M. Finel, and M. Koskinen, N-glucuronidation of drugs and other xenobiotics by human and animal UDP- glucuronosyltransferases. Xenobiotica, 2011. 41(8): p. 652-69.

232. Tukey, R.H. and C.P. Strassburg, Human UDP- glucuronosyltransferases: metabolism, expression, and disease. Annu Rev Pharmacol Toxicol, 2000. 40: p. 581-616.

233. Harrell, S.A. and D.H. McDaniel, Strong hydrogen bonds. II. The hydrogen difluoride ion [14]. Journal of the American Chemical Society, 1964. 86(20): p. 4497.

234. Dunitz, J.D. and R. Taylor, Organic fluorine hardly ever accepts hydrogen bonds. Chemistry - A European Journal, 1997. 3(1): p. 89- 98.

235. Mairal, T., et al., Iodine atoms: a new molecular feature for the design of potent transthyretin fibrillogenesis inhibitors. PLoS One, 2009. 4(1): p. 6.

236. Gales, L., et al., Human transthyretin in complex with iododiflunisal: structural features associated with a potent amyloid inhibitor. Biochem J, 2005. 388(Pt 2): p. 615-21.

237. Meerts, I.A., et al., Potent competitive interactions of some brominated flame retardants and related compounds with human transthyretin in vitro. Toxicol Sci, 2000. 56(1): p. 95-104.

238. Marchesini, G.R., et al., Biosensor recognition of thyroid-disrupting chemicals using transport proteins. Anal Chem, 2006. 78(4): p. 1107-14.

239. Lai, D.Y., S. Kacew, and W. Dekant, Tetrabromobisphenol A (TBBPA): Possible modes of action of toxicity and carcinogenicity in rodents. Food Chem Toxicol, 2015. 80: p. 206-14.

240. Bromine Science and Environmental Forum (BSEF). BSEF poster on TBBPA. [cited 2015; Available from: http://www.bsef.com.

241. Hamers, T., et al., In vitro profiling of the endocrine-disrupting potency of brominated flame retardants. Toxicol Sci, 2006. 92(1): p. 157-73.

242. de Wit, C.A., D. Herzke, and K. Vorkamp, Brominated flame retardants in the Arctic environment - trends and new candidates. Science of the Total Environment, 2010. 408(15): p. 2885-2918.

243. Fromme, H., et al., Polybrominated diphenyl ethers (PBDEs), hexabromocyclododecane (HBCD) and "novel" brominated flame retardants in house dust in Germany. Environ Int, 2014. 64: p. 61- 8.

244. Cope, R.B., S. Kacew, and M. Dourson, A reproductive, developmental and neurobehavioral study following oral exposure of tetrabromobisphenol A on Sprague-Dawley rats. Toxicology, 2015. 329: p. 49-59.

245. National Toxicology Program. Technical Report on the toxicology studies of tetrabromobisphenol A (CAS NO. 79–94-7) in F344/NTac rats and B6C3F1/N mice and toxicology and carcinogenesis study of tetrabromobisphenol A inWistar Han [Crl:WI(Han)] rats and

B6C3F1/N mice (gavage studies). NTP Technical Report 587. 2014; Available from: http://ntp.niehs.nih.gov/results/pubs/longterm/reports/longterm/i ndex.html.

246. Eriksson, P., E. Jakobsson, and A. Fredriksson, Brominated flame retardants: a novel class of developmental neurotoxicants in our environment? Environ Health Perspect, 2001. 109(9): p. 903-8.

247. Nakajima, A., et al., Neurobehavioral effects of tetrabromobisphenol A, a brominated flame retardant, in mice. Toxicol Lett, 2009. 189(1): p. 78-83.

248. Szymanska, J.A., J.K. Piotrowski, and B. Frydrych, Hepatotoxicity of tetrabromobisphenol-A: effects of repeated dosage in rats. Toxicology, 2000. 142(2): p. 87-95.

249. Tada, Y., et al., Flame retardant tetrabromobisphenol A induced hepatic changes in ICR male mice. Environ Toxicol Pharmacol, 2007. 23(2): p. 174-8.

250. Tada, Y., et al., Effects of tetrabromobisphenol A, brominated flame retardant, in ICR mice after prenatal and postnatal exposure. Food Chem Toxicol, 2006. 44(8): p. 1408-13.

251. Dunnick, J.K., et al., Environmental chemical exposure may contribute to uterine cancer development: studies with tetrabromobisphenol A. Toxicol Pathol, 2015. 43(4): p. 464-73.

252. Kuester, R.K., et al., The effects of dose, route, and repeated dosing on the disposition and kinetics of tetrabromobisphenol A in male F- 344 rats. Toxicol Sci, 2007. 96(2): p. 237-45.

253. Schauer, U.M., W. Volkel, and W. Dekant, Toxicokinetics of tetrabromobisphenol a in humans and rats after oral administration. Toxicol Sci, 2006. 91(1): p. 49-58.

254. Knudsen, G.A., et al., TITLE Disposition and kinetics of Tetrabromobisphenol A in female Wistar Han rats. Toxicol Rep, 2014. 1: p. 214-223.

255. Kang, M.J., et al., Nephrotoxic potential and toxicokinetics of tetrabromobisphenol A in rat for risk assessment. J Toxicol Environ Health A, 2009. 72(21-22): p. 1439-45.