Transthyretin (TTR) amyloidosis (ATTR amyloidosis) is an underdiagnosed and important type of cardiomyopathy and/or polyneuropathy that requires increased awareness within the medical community. Raising awareness among clinicians about this type of neuropathy and lethal form of heart disease is critical for improving earlier diagnosis and the identification of patients for treatment. The following review summarizes current criteria used to diagnose both hereditary and wild-type ATTR (ATTRwt) amyloidosis, tools available to clinicians to improve diagnostic accuracy, available and newly developing therapeutics, as well as a brief biochemical and biophysical background of TTR amyloidogenesis.
There is a growing awareness of the prevalence of protein-misfolding diseases in today's aging population. Under certain conditions, proteins that are normally folded into a thermodynamically low-energy state can misfold and aggregate, forming insoluble ‘amyloid’ fibrils. Amyloidosis describes a group of diseases characterized by the extracellular deposition of these toxic, insoluble, cross β-sheets within various locations in the body, causing the disruption and dysfunction of normal surrounding tissue [1,2]. The prevalence of these types of diseases in older generations may be the result of microenvironmental changes associated with aging, including changes in intra/extracellular pH, proteasomal functioning and the body's ability to eliminate aggregates [3–5]. Amyloid formation, however, is not restricted to old age and can occur at any decade of life as a result of mutations within the amyloid precursor protein's primary structure, chaperonopathies or altered protein metabolism [4,5].
There are more than 30 proteins known to form amyloid in vivo. All demonstrate the characteristic yellow to green birefringence upon exposure to Congo Red dye and polarized light [6–8]. The specific amyloid type and the clinical outcomes associated with misfolding are dictated by the nature of the amyloid precursor protein [7,9]. For nomenclature purposes, each type of amyloid is prefaced with the prefix A followed by a suffix that identifies the fibril precursor protein . For example, transthyretin (TTR) amyloidosis (ATTR amyloidosis) is a systemic disease caused by the deposition of transthyretin amyloid (ATTR) throughout the body. ATTR amyloidosis is a widely underdiagnosed and important disease that can be lethal if ATTR infiltrates the heart, a common site of ATTR deposition. Although originating from one protein, pathogenesis and manifestations in ATTR amyloidosis may vary and make the disease heterogeneous. Its underestimated prevalence is further supported by the fact that many patients are not diagnosed with ATTR amyloidosis until post-mortem examination [10–14]. The following review will outline the basic background behind this condition with the aim of helping clinicians understand the biophysics and pathophysiology behind ATTR amyloidosis, the heterogeneous array of signs and symptoms associated with the disease, current diagnostic tools available to clinicians to improve diagnostic accuracy, as well as current and developing treatment interventions available for those patients identified at early enough stages of the disease trajectory.
Transthyretin (earlier known as prealbumin due to its molecular mass causing it to characteristically migrate at a band prior to albumin on serum electrophoresis) is an abundant, soluble, serum protein that transports both vitamin A (via retinol-binding protein) and thyroxine throughout the body [15,16]. TTR is also involved in the binding and redistribution of β-amyloid in the choroid plexus as well as in the retention of T4 in the cerebral spinal fluid (CSF) . TTR concentration is 10-fold higher in plasma (3.6 μM) than in CSF (0.36 μM) . Although the liver is the primary site of synthesis, alternatives sites of production include the choroid plexus, α-cells in the islets of Langerhans, ocular sites that may include retinal pigment epithelia, and in rabbits, the ciliary pigment epithelia [18–21]. TTR is a β-strand rich 55 kDa homotetramer that can dissociate into its 127 amino acid monomeric subunits. These monomers can undergo aberrant changes and become amyloidogenic intermediates that can self-associate and eventually form amyloid fibrils that accumulate as amyloid deposits throughout the body, resulting in ATTR amyloidosis (Figure 1).
TTR fibril formation pathway
ATTR amyloidosis can be sub-classified as either wild-type (wt) or hereditary. In wild-type ATTR (ATTRwt) amyloidosis, formerly referred to as ‘senile systemic amyloidosis’, systemic amyloid deposition is the result of the misfolding of wt TTR. This is in contrast with hereditary ATTR amyloidosis, where point mutations within the TTR allele cause the deposition of mutant TTR amyloid throughout the body. Regardless of the type of ATTR diagnosis, post-mortem analysis reveals TTR amyloid deposition in almost all organs and tissues of affected individuals [22,23]. Despite characteristically wide systemic amyloid deposition, both hereditary and ATTRwt amyloidosis can result in cardiac complications, renal complications such as proteinuria due to glomerular involvement and ATTR deposition in the papillae of the renal medulla [24,25], as well as neuropathy due to deposition of TTR amyloid in the endoneurium and surrounding neuron vasculature . At initial clinical presentation, however, the most common neurological complaint from patients with hereditary ATTR amyloidosis is the loss of sensory function [26,27]. At early stages of hereditary ATTR amyloidosis, the effects of neurodegeneration are most pronounced in the small myelinated and unmyelinated nerve fibres, whereas thicker fibres and their respective sensory nerve action potentials are less affected . The sensorimotor effects caused by this neurodegeneration can result in a progressive reduction in pain and temperature sensation inwards from the distal extremities . When larger fibres become involved at later stages of disease, patients are at risk of developing muscle weakness and atrophy, which can be debilitating . Impotence, urinary and GI dysfunction are also common and can indicate autonomic nervous system involvement . There have been additional reports of vocal hemiparesis and dysarthria in patients with hereditary ATTR due to central nervous system involvement . However, in both ATTRwt amyloidosis and several forms of hereditary ATTR amyloidosis, cardiac manifestations strongly predominate.
Carpal tunnel syndrome (CTS), a well-studied median nerve neuropathy, is also a common finding in affected individuals due to the deposition of ATTR in the tenosynovial tissue [31,32]. CTS is one of the earliest signs of ATTR amyloidosis development, especially in ATTRwt amyloidosis as CTS can precede ATTRwt-related cardiac complications by 6 (± 4.6) years . Clinicians should therefore perform accurate neurological assessment in those patients who present with clinical signs of CTS and prior to referral for surgical relief interventions such as carpal tunnel release. TTR has also been shown to deposit in soft tissues, and there have been reports of ATTR deposits in the ligamentum flavum of the spinal canal, resulting in the development of lumbar spinal stenosis [33,34]. Additional complications exclusively associated with hereditary ATTR amyloidosis include ocular amyloid deposition resulting in impaired vision due to vitreous deposits, pupillary and lacrimal dysfunction, as well as glaucoma [35–37].
ATTR amyloidosis results in systemic amyloid deposition, but has lethal implications when it results in ATTR deposition within the heart. TTR-related cardiac amyloidosis is often a progressive infiltrative cardiomyopathy and the deposition of ATTR amyloid within cardiac tissue varies. Even if progressive infiltrative cardiomyopathy is not present, cardiac amyloid deposits may cause severe conduction disturbances. Because it mimics hypertensive and hypertrophic cardiomyopathies, TTR-related cardiac amyloidosis often goes undiagnosed and both earlier and accurate clinical recognition by physicians are critical .
ATTR amyloidosis: genetics and inheritance patterns
The first report of hereditary ATTR amyloidosis originated in Northern Portugal in 1952, followed by reports from ‘endemic’ areas in Japan (1968) and Sweden (1976) [39–41]. The most frequently reported cause of hereditary ATTR amyloidosis is the TTRVal30Met mutant, commonly affecting but not limited to patient populations in Portugal, Japan, Sweden and Brazil ; the lack of geographical proximity between these high areas of prevalence raised the question regarding the origin of the allele . Studies have shown that although the mutation is the same in Portugal and Brazil, the mutation affecting Swedish patient populations has occurred independently . In addition, the geographical origin and ethnic background of the patient can have implications for disease onset and initial symptoms for a given mutation . For example, in ATTRVal30Met amyloidosis, patients of Brazilian and Swedish descent are the youngest and oldest patient populations to start developing signs of disease respectively, despite having the same disease-causing mutation [26,44]. Although the age of onset for hereditary ATTR amyloidosis follows a bimodal distribution across a wide variety of mutations, the median age of onset is generally earlier in patients with hereditary ATTR amyloidosis in comparison with those of ATTRwt amyloidosis: 39.0 (25.9–64.5) compared with 71.4 (60.1–81.6) years of age respectively .
Although ATTRwt amyloidosis is the most common form of ATTR amyloidosis in the U.S.A. (43%) with ATTRVal122Ile being the second most prevalent (23%), ATTRVal30Met amyloidosis is the most common of the hereditary forms worldwide (76%) . In addition to the ATTRVal30Met and ATTRVal122Ile mutants, there have been >100 additional mutations reported worldwide, though their global distribution varies . Hereditary ATTR amyloidosis is an autosomal dominant disease with heterogeneous phenotypic expression. Signs and symptoms of this form of amyloidosis can be multisystemic, ranging from only neuropathic or cardiomyopathic complications to a mixture of both . Although homozygosity does not necessarily result in earlier disease onset, it has been suggested that those patients who are homozygous for mutant TTR alleles (i.e. TTRVal30Met) often develop more severe signs and symptoms due to increased central nervous involvement in addition to having increased ATTR infiltration of the leptomeninges and subarachnoid vasculature in comparison with heterozygotes [47,48]. These findings conflict with another study which found phenotypic severity, when compared with heterozygotes, to be unaffected by homozygosity . Phenotypic penetrance can vary between mutations though, and for example in Sweden, many carriers of the TTRVal30Met mutation never develop the disease . Cases of monozygotic twins in which one twin develops the disease and the other does not, have also been described . Compound heterozygotes, in which a patient has both a disease-causing and disease-inhibiting mutation, can also occur. The introduction of an anti-amyloidosis mutation (i.e. TTRThr119Met) in the otherwise ATTR-causing TTR mutated proteins inhibits aggregation by increasing the tetrameric dissociation activation barrier that thermodynamically stabilizes the TTR protein and impairs clinical manifestation of the disease .
TTR mutations are associated with varying risk of cardiac involvement . Patient subpopulations of African American descent carrying the ATTRVal122Ile mutation are at particular risk for developing cardiac-related ATTR compared to many other groups. This is also true for the ATTRVal111Met mutation . There has been some discussion in the literature regarding the ATTRVal122Ile allele prevalence within the African American population [54–56]; its current estimate of 3–4% would translate to approximately 1.3 million allele carriers who are at an increased risk of developing cardiac amyloidosis in the U.S.A. alone [54–56]. This mutant allele has a phenotypic penetrance of 100% that results in deposition of ATTR in the heart, and is one of the major reasons why clinicians must be cautious in managing patients of African American descent who present with early signs of heart failure.
Regardless of the subtype of ATTR amyloidosis, all affected patient populations are at significant risk of disease progression due to misdiagnosis. ATTR amyloidosis is often masked by symptoms and signs that mimic similar, more common diagnoses that have shared symptomatology (such as obesity, diabetes or hypertension). This is compounded by the similarity of these patients’ profiles to those with ATTR amyloidosis, i.e. ethnic, older patient populations. In addition, correct diagnosis can be delayed further as the first cardiac sign of the disease, left ventricular hypertrophy (LVH) is one with an expansive differential diagnosis, with diseases much more prevalent and common than ATTR amyloidosis .
ATTR subtypes and pathology
The cardiac pathophysiology of ATTR amyloidosis is due to the deposition of amyloid in the heart. In Type A amyloidosis, ATTR deposition in cardiac tissue results in the development of a restrictive cardiomyopathy and can have lethal clinical outcomes. For ATTRVal30Met amyloidosis, there are two discrete and well-separated patterns, each with a very different clinical outcome. In Type A ATTR amyloidosis, amyloid appears patchy in the myocardium and these foci tend to coalesce into homogeneous amyloid masses, compressing cardiomyocytes . This form of ATTR amyloidosis is associated with progressive restrictive cardiomyopathy and can lead to severe cardiomegaly. In Type B ATTR amyloidosis, amyloid appears as thinner streaks interstitially and subendocardially . Although quite conspicuous amounts of amyloid can also build up in this type of ATTR amyloidosis, restrictive cardiomyopathy does not develop. However, conduction disturbances are common and are the reason many patients require the insertion of a pacemaker.
Interestingly, Type A and Type B fibrils are different in that Type A mainly contain C-terminal fragments of TTR whereas Type B fibrils only contain full-length TTR molecules. The resulting amyloids differ in that Type A fibrils are short and have weak affinity for Congo Red whereas Type B fibrils are long, slender, and strongly stain with Congo Red. The most important distinction, however, is that Type A fibrils tend to recruit more wtTTR to deposits than Type B fibrils do. Typically, Type A ATTR deposits continue to grow following liver transplantation in contrast to Type B amyloid . In addition, for at least Swedish patient populations with ATTRVal30Met amyloidosis, Type B amyloid is most common in early onset cases whereas Type A is seen in late onset (>50 years of age). Type A ATTR amyloidosis is by far the most common form found for all other TTR mutations and is also the single finding in ATTRwt amyloidosis . The differences between Type A and Type B fibrils make it reasonable to believe that there are (at least) two different ways by which TTR can form fibrils. Previous investigations have shown the importance of the C-terminal segment of TTR in fibrillogenesis, but for Type B amyloid other mechanisms may exist [61–63].
Cardiac MRI (CMRI) can be of great help for ATTR amyloidosis diagnosis. CMRI utilizes the relaxation times of protons within a magnetic field to generate images of tissue based on differential proton density or water content. CMRI signal intensities within amyloid-infiltrated tissue can be enhanced when paired with a contrast agent such as gadolinium and result in diffuse, global and subendocardial patterning . The unique patterning of late gadolinium enhancement in cardiac amyloidosis is useful for ruling out hypertrophic or hypertensive cardiomyopathies and diagnosing cardiac amyloidosis (80% sensitivity; 94% specificity) [64,65]. Contrast-enhanced CMRI does, however, have limitations in that it cannot distinguish between different subtypes of cardiac amyloidoses and does not permit quantification . Native myocardial T1 mapping, however, is a quantitative imaging modality that relies on the mapping of longitudinal relaxation times for a magnetized proton . Native T1 myocardial mapping has both diagnostic value for ATTR amyloidosis (both hereditary and wt) and allows for tracking of disease progression . In comparison with normal patients or those with hypertrophic cardiomyopathy, a common clinical mimicker, patients with ATTR amyloidosis (both hereditary and wt) have increased T1 relaxation times, potentially indicative of the amyloid-induced hydro-effects on ATTR-infiltrated cardiac tissue . Although native T1 myocardial mapping is limited in its ability to differentiate between different types of cardiac amyloidoses, lower T1 times have been reported for ATTR amyloidosis in comparison with AL amyloidosis . There have been additional reports on the diagnostic potential of magnetic resonance neurography (MRN), a type of MRI that can be used for the evaluation of polyneuropathy in ATTR amyloidosis. High-resolution MRN has been used for the detection and quantification of lower limb nerve damage in both symptomatic and presymptomatic gene-carriers of hereditary ATTR-related polyneuropathy .
There have been recent advances in the development of alternative imaging strategies that can be used to improve the diagnostic accuracy of cardiac amyloidosis. Imaging studies have demonstrated promise for differentiating and diagnosing cardiac amyloidoses (including ATTR amyloidosis) from hypertensive heart disease using positron emission tomography (PET) with 11C-PIB or 18F-florbetaben labelling [70–72]. Technetium-labelled bone scintigraphy, a technique that measures a radiolabelled isotope to characterize disease, is commonly used for the imaging and diagnosis of bone pathologies. Interestingly, this method has been reported as a technique to specifically identify and characterize TTR-related cardiac amyloidosis, as affected cardiac tissues display an increased uptake of isotopes not seen in other cardiomyopathies, even amyloidogenic ones [73,74]. This technique has been suggested to provide complete diagnostic information in ATTR amyloidosis potentially making biopsy unnecessary . However, limitations of the technique include the risk of the tracer binding cardiac AL amyloid deposits  and its inability to visualize amyloid deposits composed of full-length ATTR (Type B amyloid) ; bone markers can visualize amyloid composed of C-terminal TTR fragments (Type A amyloid) possibly due to micro calcifications in the amyloid deposit . There are other amyloid tracers that are less fibril type-specific, such as Pittsburgh compound B (PiB), a radioactive thioflavin T derivative which in conjunction with PET can recognize both Type A and B amyloid . Nevertheless, while technetium-labelled bone scintigraphy has demonstrated value for characterizing ATTR-related cardiac amyloidosis, whether it can be expanded beyond the heart and used for the visualization of ATTR deposits to characterize ATTR infiltration of the nervous system or other organs would be worth investigation.
ATTR cardiomyopathy leads to decreased diastolic relaxation and consequently results in poor ventricular filling during diastole. ATTR amyloidosis results in the non-specific infiltration of both the left and right ventricles and valves . Although echocardiography (ECHO) is a common ultrasound imaging modality with limited specificity that can help recognize ATTR amyloidosis patients, thickening of the left ventricular free wall or septal thickness in ATTR amyloidosis can mimic hypertrophic or hypertensive cardiomyopathy and result in misdiagnosis [77,79]. A ‘sparkling’ or ‘granular’ pattern on ECHO can be characteristic of ATTR amyloidosis; however, this pattern cannot differentiate between different forms of cardiac amyloidosis . It is important to note that the diagnostic value of ‘myocardial sparkling’ for cardiac amyloidosis is insufficient for diagnosis when found via harmonic imaging . Although harmonic imaging is a type of ECHO that achieves better visualization of endo- and myo-cardial tissue [82,83] than conventional fundamental frequency ECHO, evidence of myocardial sparkling in hearts not infiltrated with amyloid remains a common finding of this technique .
Strain ECHO imaging, a method of determining the amount of deformation of the myocardial tissue throughout the cardiac cycle, has been demonstrated to be a useful technique to help characterize TTR-related cardiac amyloidosis from hypertrophic cardiomyopathy. Although both conditions are characterized by impaired atrial reservoirs and contractile function, the impairment is more severe for TTR-related cardiac amyloidosis due to the greater effects of amyloid deposition in comparison with fibrosis .
Low-voltage QRS complex analysis via electrocardiography (ECG) may be another potential strategy used for the detection of cardiac amyloidosis; however, its clinical usage is limited as the presence of low-voltage QRS complexes in infiltrative cardiomyopathies is inconsistent [78,85]. This holds true for other classic ECG findings such as the ‘pseudoinfarct’ pattern of Q waves, left atrial enlargement and atrial fibrillation, as well as left bundle branch block (LBBB), as these are not ATTR-specific nor sensitive findings . As such, the absence of low-voltage, restrictive pattern or other classic ECG findings should not be used to rule-out ATTR amyloidosis, especially in those patients with increased left atrial diameter and evidence of increased thickening of the right ventricular free wall, valves and interventricular septum . Because basic cardiac imaging or tracings are of limited value in ATTR amyloidosis diagnosis, tissue biopsy or alternative diagnostic tests are required. For diagnosis of any type of systemic amyloidosis, a tissue biopsy followed by biochemical characterization with immunological methods or mass spectrometry is warranted.
Currently, the diagnostic gold standard for cardiac involvement in ATTR amyloidosis is endomyocardial biopsy, which involves obtaining multiple cardiac biopsy samples via central venous access. Although of high diagnostic value for ATTR amyloidosis, there are multiple risks associated with this invasive procedure, including myocardial perforation, arrhythmia, haemothorax, pneumothorax, pericardial tamponade and death [87,88]. It is most common, however, to determine diagnosis using a biopsy obtained from alternative sites. Sites most commonly used are subcutaneous abdominal fat tissue and labial salivary glands. In most cases, both these sites provide diagnosis if correctly performed; however, experience in interpretation is crucial.
Although there are no known diagnostic biomarkers for ATTRwt amyloidosis , protocadherin-10 (Pcdh10) expression has been linked to the progression of ATTR peripheral neuropathy. Current work has been focusing on Pcdh10 up-regulation being a potential biomarker of hereditary ATTR progression; however, this potentially promising area of research requires further investigation . Abnormal serum levels of N-terminal fragment of pro-brain natriuretic peptide (NT-proBNP), a cardiac dysfunction biomarker, has also been linked to both hereditary and ATTRwt amyloidosis. Its usage as a biomarker for ATTR amyloidosis, however, is limited since abnormal NT-proBNP levels are also found in patients with immunoglobulin light-chain (AL) cardiac amyloidosis .
With the exception of specific centres, clinicians rarely consider ATTR amyloidosis as a cause for their patients’ symptomatology, even when indicated. ATTR amyloidosis should be considered for patients who present with axonal polyneuropathy, especially if they have autonomic involvement , or in patients presenting with diastolic heart failure. Given the non-specific symptoms of the disease, and the difficulties in its differential diagnoses, most patients with evidence of heart failure and preserved ejection fraction ideally should be tested for cardiac amyloidosis . Because amyloidosis, especially with cardiac involvement, has poor prognostic outcomes for patients, improved diagnostic accuracy would result in better outcomes for patients due to early detection and more targeted therapeutic interventions .
Surgically invasive interventions
Depending on the type of ATTR amyloidosis, either hereditary or wt, the main therapeutic intervention for clinical ATTR amyloidosis management is to remove the organ synthesizing amyloid precursor protein and/or affect its rate of TTR synthesis. Accumulated ATTR amyloid deposits can have lethal implications, therefore early intervention is required. The gold standard for hereditary ATTR amyloidosis management is orthotopic liver transplantation, a procedure first performed in Sweden for two patients with ATTRVal30Met [92,93]. Since 1990, this procedure has been associated with positive long-term prognosis for those hereditary ATTR amyloidosis patients who are identified at early enough stages of the disease because it removes the origin of amyloid precursor protein and replaces circulating mutant TTR with a normal wtTTR . The long-term survival rates, however, appear to be strongly dependent on the amyloid fibril type (Type A compared with Type B) since those patients with Type A amyloid progress in cardiomyopathy following liver transplantation due to the addition of wtTTR molecules [59,94]. There are additional factors of importance, such as age at disease onset, patient nutritional status and type of disease-causing TTR mutation .
Despite the promising prognosis for these patients, the main concern and long-term complication of liver transplantation in patients with hereditary ATTR amyloidosis is that previously existing hereditary ATTR deposits can seed amyloid growth using the newly circulating normal wtTTR and result in disease progression [46,95]. Progression of cardiomyopathy in patients who undergo orthotopic liver transplant is therefore not uncommon, especially in those patients with previous cardiac amyloid infiltration or evidence of cardiac involvement, such as posterior wall and/or septal ventricular wall thickening [96–98]. Additional concerns for orthotopic liver transplantation include the failure to address additional sites of TTR production, such as the eye, where amyloid deposition can continue along the pupil and increase the risk of post-operative hereditary ATTR patients developing glaucoma . The procedure also neglects to address the issue of continuity between the subarachnoid and endoneural space. The connection between these two contiguous structures can potentially result in the post-operative accumulation of amyloidogenic TTR precursors in the peripheral nervous system [23,99].
In addition to continued ocular ATTR deposition, failure to address continued post-operative production of variant TTR production by the choroid plexus is also a concern. Even after orthotopic liver transplantation, there have been reports of patients with hereditary ATTR amyloidosis experiencing focal neurologic episodes due to ATTR-related cerebral amyloid angiopathy (CAA) as well as post-mortem evidence of CSF-derived variant ATTR infiltration of the leptomeninges and leptomeningeal vasculature . The risk of developing CNS symptoms attributed to undisrupted variant ATTR production in the CNS varies from 11.3 to 31%, a difference most likely due to the length of post-operative follow-up, with an average onset of 14.6 to 16.8 years from initial ATTR amyloidosis disease onset [100,101].
An alternative treatment option for patients with severe ATTR cardiac amyloidosis and heart failure is orthotopic heart transplantation, either alone in the case of ATTRwt amyloidosis or in combination with liver transplantation, for those patients with hereditary ATTR amyloidosis. First performed in England in 1984, heart transplantation for ATTR amyloidosis patients can be controversial due to the limited supply of hearts available compounded by the risk of disease recurrence in the cardiac allograft post-surgery [102,103]. The variable post-operative survival rates for those patients who do undergo orthotopic heart transplantation vary, with reported 1- and 5-year post-operative survival rates ranging from 60–74.6% and 30–54%, respectively, thereby suggesting this procedure can be used as a last resort for patients with extreme cardiac amyloidosis [103–104].
Small molecule drugs: TTR stabilizers
Not all patients diagnosed with ATTR amyloidosis, either wt or hereditary, are eligible for surgically invasive interventions such as orthotopic liver and/or heart transplantation. Failure to meet surgical eligibility criteria leaves these patients at risk of premature death, and therefore there is a critical need for the development of alternative, non-invasive treatment strategies. A number of small molecule drugs have been identified as potential pharmacotherapies, but none of these drugs have FDA-approval.
One drug that has gained approval for the treatment of hereditary ATTR in multiple countries, including Japan, Mexico, Argentina and the European Union, is tafamidis (Vyndaqel®) . Tafamidis is a non-NSAID (non-steroidal anti-inflammatory drug) benzoxazole derivative that functions as a TTR stabilizer [106,107]. TTR contains two pairs of dimer–dimer interfaces that together form two tetrameric hydrophobic-binding pockets normally occupied by thyroxine. Small molecule drugs such as tafamidis have been designed to occupy these T4-binding sites via negative co-operativity, kinetically stabilize the TTR tetramer, and result in the consequent decrease in the rate of tetrameric dissociation rates and fibril formation in vitro [107,108]. For patients diagnosed with early signs/mild hereditary ATTR amyloidosis, tafamidis can potentially delay neuropathic progression of ATTR amyloidosis for up to 5.5 years; however, a direct control group to assess whether this effect is due to the drug alone or natural disease progression would help clarify this . Although one concern of this drug is that occupation of these hormone-binding sites may have metabolic side effects by affecting thyroxine delivery throughout the body, clinical trials have found minimal evidence of this . This clinical finding supports the paradigm that thyroxine-binding globulin, rather than TTR, transports approximately 99% of the body's circulating thyroxine [110,111].
Other than tafamidis, there have been multiple tetrameric stabilizers that have gained scientific interest due to their reported effects on TTR amyloidogenesis. These include epigallcatechin-3-gallate (EGCG), curcumin, diflunisal, resveratrol, AG10 as well as tolcapone, which has gained more clinical interest and is currently undergoing clinical trials. There have been reports of dietary compounds such as EGCG and curcumin, major components of green tea and turmeric respectively, to have effects of TTR fibrillogenesis. At high concentrations, EGCG is able to stabilize the TTR tetramer and exhibit both fibrillogenesis inhibition and disruption of mature amyloid deposits in transgenic mice [112–114]. Transgenic mouse model studies have also shown curcumin (also known as diferuloylmethane) to decrease both ATTR amyloid deposition and ATTR-related cytoxicity, as well as result in amyloid remodelling in tissue .
Resveratrol has been shown to induce tetrameric stabilization via the thyroxine-binding site mechanism described above. Resveratrol has also gained recent popularity for its reported benefits for the treatment of Alzheimer's disease via its ability to stabilize TTR and thereby decrease Aβ levels in transgenic mice . In humans, however, the bioavailability of orally administrated resveratrol is limited to <5% due to rapid intestinal conjugation, limiting its clinical use in treating ATTR amyloidosis .
Administration of diflunisal to patients diagnosed with hereditary ATTR amyloidosis results in improvement of autonomic symptoms but its long-term side effects include induction of impaired renal function and thrombocytopenia [118,119]. Because diflunisal is an NSAID, chronic management of ATTR amyloidosis using this drug is often contraindicated or requires cautious monitoring in both elderly patients and those diagnosed with cardiovascular disease .
High-throughput screening has resulted in the development of the TTR kinetic stabilizer, AG10, a small molecule drug that occupies TTR's T4-binding pocket via negative co-operativity [108,121]. In comparison with tafamidis, AG10 can significantly both inhibit wt and V122I TTR fibrillogenesis in vitro at substoichometric concentrations and stabilize wt and V122I TTR in human serum .
Tolcapone, an anti-Parkinson agent, has been recently repurposed as a potential treatment for ATTR amyloidosis due to its ability to stabilize the TTR tetramer by occupying its T4-binding site, demonstrating higher binding affinity and potency than tafamidis in vitro . Phase I/II clinical trials to evaluate its TTR stabilization effects in patients with ATTR amyloidosis have been ongoing since July 2014.
Small molecule drugs: TTR aggregate disrupters
Drugs that have recently gained more clinical interest for the treatment of ATTR amyloidosis include doxycycline and tauroursodeoxycholic acid (TUDCA). In transgenic mouse models, TUDCA, a hydrophilic biliary acid, has been shown to decrease the deposition of toxic pre-fibrillular TTR oligomers and levels of cellular stress biomarkers normally associated with ATTR amyloid deposition, but showed no effect on mature amyloid fibrils [123,124]. Studies using transgenic mice have also reported additional effects of TUDCA, as well as curcumin, on ATTR amyloidosis progression via modulation of cellular autophagy processes . Doxycycline, a member of the tetracycline family, has exhibited amyloid fibril disruption effects both in vitro and in vivo, but has not demonstrated any effect on toxic pre-fibrillular TTR oligomers [123,126]. Recent advances include the development of polyglutamate–doxycycline conjugates, which have been reported to have enhanced fibril elimination effects in comparison with parental doxycycline-only controls . Phase I/II clinical trials investigating the effect of combining these latter two treatments (doxycycline and TUDCA) and evaluating their effects on the progression of ATTR cardiac amyloidosis have been underway since June 2013.
Antisense nucleotides and siRNAs
Alternatives to tetramer stabilizing pharmacotherapies are siRNAs and antisense nucleotides (ASOs), which interfere with and silence TTR gene transcription. Hepatic gene silencing of TTR allows for decreased systematic TTR production and decreased progression of ATTR amyloidosis without invasive, surgical intervention. Some of these drugs have had better success at affecting ATTR amyloidosis in vivo than popular TTR stabilizers. TTR specific siRNAs (siTTR) tested in murine models of ATTR amyloidosis have been reported to decrease ATTR amyloid deposition and induce ATTR amyloid regression better than the aforementioned TTR tetramer stabilizer tafamidis . siRNAs with similar mechanisms of action include patisiran and revusiran [129,130]. Alnylam Pharmaceuticals had both drugs in Phase III clinical trials until the latter was discontinued in October 2016 due to compromised patient safety; patisiran studies have reportedly been unaffected by this discontinuation. A second-generation antisense oligonucleotide, IONIS–TTRRx, currently in Phase II/III clinical trials, has also been reported to decrease TTR plasma levels by >80% in both monkeys and I84SATTR transgenic mice. This therapy was also tested in healthy humans, demonstrating that it can decrease wtTTR plasma concentrations in a dose-dependent manner . Phase III clinical trials of IONIS–TTRRx were initially delayed when the U.S. FDA put the trials on hold because of concerns for patient safety due to the development of thrombocytopenia. IONIS Pharmaceuticals has since reported positive results from ongoing trials with IONIS–TTRRx .
Despite the widespread use of monoclonal antibodies in a range of scientific and medical applications for the past few decades, their use to target pathologically specific conformations of a protein and to treat or prevent protein-misfolding diseases is less common. In immunoglobulin light chain (AL) amyloidosis, monoclonal antibodies that promote the clearance of amyloid in amyloidoma mouse models and that demonstrate positive biomarker responses in clinical trials have been developed [132,133]. Further monoclonal antibodies for AL amyloidosis treatment are under development . A similar approach is now being investigated for the treatment of ATTR amyloidosis.
Structure-based immunotherapeutic strategies have been recently reported for the potential development of non-surgically invasive treatment interventions for ATTR amyloidosis. Both polyclonal and murine monoclonal antibodies have been developed that target the residues TTR89-97 of human TTR, an epitope or ‘cryptotope’ deeply buried in the TTR tetramer but exposed upon tetrameric dissociation [135,136]. The concept behind this strategy is that, as mentioned previously, normal TTR remains a tetramer but under conditions that result in misfolding, it can dissociate into its monomeric subunits and expose residues normally inaccessible when in the native tetrameric form. This allows for the specific targeting of the pre-fibrillar and non-native conformations of the TTR protein, but not TTR's natively folded and non-pathological tetramer. These monoclonal antibodies have successfully demonstrated fibril inhibition, specific ATTR amyloid labelling in TTR amyloid-positive tissue, as well as the ability to induce phagocytosis by human monocytes upon exposure to non-native TTR in vitro . Murine and humanized monoclonal antibodies that target a cryptotope downstream of that targeted by the TTR89-97 antibodies have been recently reported with similar pre-clinical success in their ability to specifically recognize non-native forms of TTR, inhibit TTR fibrillogenesis in vitro, and induce phagocytosis of TTR fibrils by macrophages . These antibodies target residues TTR115-124 of human TTR and have also been shown to affect ATTR deposition in transgenic rats expressing human ATTRVal30Met .
Alternative immunotherapeutic approaches for ATTR amyloidosis include the development of therapeutic anti-serum amyloid P component (SAP) antibodies. SAP is a non-fibrillar plasma glycoprotein found in all types of human amyloid deposits, included ATTR. A phase I trial involving 15 patients with systemic amyloidosis, but not cardiac amyloidosis, studied the effects of CPHPC ((R)-1-[6-[(R)-2-carboxy-pyrrolidin-1-yl]-6-oxo-hexanoyl]pyrrolidine-2-carboxylic acid), a small molecule drug that depletes levels of circulating SAP, followed by a single dose of humanized IgG anti-SAP antibodies . These anti-SAP antibodies target residual SAP in amyloid deposits and have been shown to initiate macrophage-mediated clearance of SAP-infiltrated amyloid in mouse models . This small trial reported reduction in hepatic, renal and lymphatic amyloid load and no serious adverse side effects . Although none of the 15 amyloidosis patients involved had ATTR amyloidosis, expansion of this specific study  to include and evaluate the effects of treating patients with ATTR amyloidosis would be of value.
Biophysics of ATTR
TTR forms a stable orthorhombic crystal structure (P21212 symmetry), with each of its four subunits consisting of eight β-strands (ABCDEFGH) assembled into a sandwich of four β-stranded anti-parallel sheets (DAGH and CBEF) . Two subunits/monomers can self-associate into a symmetric dimer via F–F′ and H–H′ strand interactions, and these dimers can in turn associate to form a homotetramer. Tetrameric dissociation is the first step of TTR fibril formation, followed by the partial unfolding of its monomeric subunits, resulting in amyloidogenic precursors with high propensity to aggregate into amyloid [141,142].
The initiating step of tetrameric dissociation can be initiated by low pH , TTR point mutations or age-associated protein modifications. X-ray crystal structures of wtTTR in both low and physiological pH conditions reveal that low pH results in large conformational changes in the TTR's E and F strand loop–helix regions, regions involved in the dimer–dimer interfaces. These changes result in large quaternary changes and the consequent unfolding of the tetramer into its amyloidogenic monomers . It is believed that amyloidogenic point mutations in the TTR protein instigate tetrameric dissociation by affecting the tetrameric dissociation kinetics rather than the tetramer's thermodynamic stability [145,146]. In addition to point mutations and low pH environments, age-related effects on both tetrameric wt and mutant TTR (TTRVal30Met), either oxidative modifications of sulfur-containing residues (i.e. Met and Cys) or age-associated protein carbonylation, have been reported as potential contributors to TTR amyloidogenicity . These age-associated protein modifications, more specifically carbonylation, also had negative in vitro side effects on the TTR stabilizer resveratrol and its ability to inhibit fibril formation due to carbonylation of residues at the resveratrol's T4-binding pocket . Whether there is association between accumulation of age-associated protein modifications and efficacy of other TTR stabilizing drugs in aging patient populations remains a relationship to be further explored.
There are multiple reports describing regions of high aggregation propensity within the TTR protein. Some studies report that isolated TTR segments within the thyroxine-binding pocket, TTR10–20 and TTR105–115 (corresponding to strands A and G), undergo amyloid fibril formation in vitro  and suggest the effect of TTR stabilizer drugs that target the T4-binding site may be the result of their interaction with the TTR105–115 segment . In silico studies have instead predicted F and H to be responsible for monomeric aggregation and have validated these findings in vitro using peptide inhibitors targeting these regions thereby inhibiting aggregation . Several other reports also show the importance of the C-terminal part of TTR in fibrillogenesis [62,63]. Although some studies report the C and D strands to be responsible for aggregation , others contradict these findings due to the presence of ‘gatekeeping residues’ such as Lys35 . Despite many amyloidogenic mutations reported in the B, C and D strands of TTR (encompassing residues TTR26–57), this region demonstrates poor amyloidogenicity in vitro due to the presence of the Lys35 residue, which decreases TTR's ability to form fibrils due to its charge and flexible side chain . Heparan sulfate, however, is able to bypass the protective effects of the Lys35 residue and bind to this region (TTR26-57); these findings support herparan sulfate's reputation as a TTR fibrillogenesis promoter and provide justification for its co-localization with ATTR deposits in cardiomyopathic heart tissue [151,152].
The mechanism of ATTR amyloid formation is a topic of evolving discussion. Studies have reported amyloid fibrils to contain tight association between F strands and significant solvent protection for multiple residues in both the F and E strands [150,153]. Others have postulated that if exposure of the natively buried F and H strands at the dimer–dimer interface is required for fibrillogenesis, then tetramer to monomer dissociation is required to allow fibril self-assembly. If true, this theory of TTR amyloidogenesis contradicts the one of amyloid formation via dimer scission [61,154].
ATTR amyloidosis is an under-recognized disease that requires increased awareness among clinicians so that patients are correctly diagnosed and those who are eligible for treatment receive it before it is too late. If left untreated, ATTR amyloidosis can result in debilitating conditions that affect the activities of daily life as well as lead to heart failure and death. Clinicians must be especially cautious before ruling out cardiac amyloidosis in patients presenting with signs and symptoms of hypertrophic or hypertensive cardiomyopathy, and be aware of diagnostic tools available to improve their diagnostic accuracy. There is urgent need for the development of non-surgically invasive treatment interventions for patients who suffer for ATTR amyloidosis, and multiple, novel treatment strategies are currently in development.
We thank Dr Vincent J.M. Brienza for his comments and constructive input into the article.
This work was supported by the Ted Rogers Centre for Heart Research; the Canadian Institutes of Health Research to A.C; Prothena Inc.; and the FAMY, FAMY Norrbotten and Amyl, and Erik, Karin and Gösta Selander's Foundation (to P.W.).
- ATTR amyloidosis
cerebral amyloid angiopathy
cerebral spinal fluid
carpal tunnel syndrome
magnetic resonance neurography
non-steroidal anti-inflammatory drug
pro-brain natriuretic peptide
positron emission tomography
serum amyloid P component