Given increased awareness and improved non-invasive diagnostic tools, cardiac amyloidosis has become an increasingly recognised aetiology of increased ventricular wall thickness and heart failure with preserved ejection fraction. Once considered a rare disease with no treatment options, translational research has harnessed novel pathways and led the way to promising treatment options. Gene variants that contribute to amyloid heart disease provide unique opportunities to explore potential disease-modifying therapeutic strategies. Amyloidosis has become the model disease through which gene therapy using small interfering RNAs and antisense oligonucleotides has evolved.
- Amyloid heart disease
- gene silencing
- restrictive cardiomyopathy
Statistics from Altmetric.com
Amyloid heart disease occurs when endogenous or clonal proteins undergo conformational changes that render them insoluble, leading to direct deposition into myocardial tissue, increased wall thickness and organ dysfunction. Cardiac amyloidosis is becoming an increasingly recognised aetiology of increased ventricular wall thickness and heart failure with preserved ejection fraction (HFpEF). Besides the myocardium, amyloid infiltrates can occur in the conduction system, valvular tissues, coronary and large arteries and autonomic or peripheral nerves, leading to a myriad of clinical manifestations. The precise mechanisms of such myocardial tissue damage are still not fully understood, although it is likely beyond alterations in tissue architecture caused by amyloid deposits. Indeed, the possibility of intrinsic toxicity from amyloidogenic precursors is evident when infusion of light chains purified from patients with severe amyloid cardiomyopathy can generate diastolic dysfunction in isolated mouse hearts.1
Although over 25 proteins have been described to form amyloid, there are predominantly two major forms which affect the heart—transthyretin (TTR) and immunoglobulin light chains (amyloid light chain, or AL)—both with novel genetic insights. Diagnosing amyloid cardiomyopathy is achieved with the gold standard endomyocardial biopsy using Congo red or Thioflavin stains; although the disease may be suspected non-invasively with a characteristic low voltage ECG,2 echocardiography with an apical sparing pattern of longitudinal strain3 and cardiac MRI with global subendocardial or transmural delayed gadolinium enhancement and kinetics.4 ATTR may be differentiated from AL on biopsy using immunohistochemistry or mass spectrometry, and non-invasively with nuclear bone tracers such as technetium pyrophosphate.5 The aforementioned diagnostic workup is essential in differentiating cardiac amyloidosis (particularly ATTR in older patients) from other forms of HFpEF and recognising it as a contributing factor in low-flow, low-gradient aortic stenosis.6
While the clinical presentation and organ involvement are diverse in amyloidoses, the importance of heredity in the expression of these diseases has been known for many years.7 Some precursor proteins appeared to be entirely heritable from specific genetic variants, whereas acquired amyloidoses may also be affected by factors associated with post-translational modifications. Herein, we explore our current understanding of the genetics and newer gene-based therapies for treating or preventing this increasingly recognised cause of HFpEF.
Genetic determinants of cardiac amyloidosis
Transthyretin amyloidosis (ATTR)
The TTR protein is synthesised and secreted by the liver and choroid plexus, and functions as a transporter of thyroxine and retinol-binding protein (responsible for vitamin A transport). Previously termed prealbumin, this protein is typically found in soluble tetramers in its native form. ATTR has become the most prevalent form of cardiac amyloidosis encountered in clinical practice with broader recognition by non-invasive diagnostic imaging tools.3 8 9 Cardiac involvement of ATTR most commonly presents in the sixth and seventh decades of life as HFpEF,10 with ‘wild-type’ or ‘senile systemic amyloidosis’ (lack of a known gene variant) still being the most common cause in the USA (over 50% in the Transthyretin Amyloidosis Outcomes Survey (THAOS) registry).11 The mean age of presentation of wild-type ATTR is in the mid-70s, and over 80% of patients with wild-type disease are male. The reason for such a male sex bias is currently unclear. The two most common organ systems involved are the heart and peripheral nervous system, but may involve ligaments and tenosynovium leading to carpal tunnel syndrome and spinal stenosis.
Initially, ATTR was thought to be a rare form of amyloidosis until researchers began noticing that amyloidogenic gene variants were present in about 10% of patients with suspected systemic AL amyloidosis.12 The TTR gene, located in chromosome 18, has over 130 known pathogenic variants that covers ~40% of its 127 amino acid residues, with different variants leading to different phenotypic presentations (table 1). Gene variants are passed down in largely an autosomal dominant fashion, often variable in penetrance, and the type of mutation is often concordant with clinical presentation, age of onset, organ involvement and disease progression. How gene variants lead to TTR fibrillisation is unknown, although it is conceivable that they may directly contribute to conformational change of the monomers and dissociation of the tetramer into monomers.
Predominantly in patients from African descent that almost exclusively leads to cardiac disease,13 Val122Ile is the most prevalent gene variant in the USA (over 45% in the THAOS registry).11 Neuropathy or prior carpal tunnel syndrome is rare. In the THAOS registry, Val122Ile subjects were younger and more often female and black than patients with wild-type disease and had similar cardiac phenotype. Carpal tunnel was slightly less common than in wild-type disease (29% vs 33% in THAOS), but there was a greater burden of neurological symptoms (pain, numbness, tingling and walking disability) and worse quality of life.11 First described in 1990,14 the epidemiology of this mutation has been extensively studied in large epidemiological cohorts from the Arteriosclerosis Risk in Communities (ARIC) Study and Cardiovascular Health Study (CHS).15 They described 5000 African American patients aged 41–93 who underwent genotyping for the Val122Ile mutation and found a prevalence of about 3%. After age 65 in CHS, there was a higher frequency of heart failure symptoms (38% vs 15%) and mortality (76% vs 53%) in those with Val122Ile as compared with the general population. However, no differences were noted in patients <65 years old as part of the ARIC study.15 Long-term propensity matched analysis from ARIC-evaluated cardiac structure, function and outcomes in carriers versus non-carriers—after 21.5 years of follow-up, patients carrying the Val122Ile mutation were more likely to develop incident heart failure without a difference in mortality. In addition, markers of systolic (global longitudinal strain) and diastolic function (by tissue Doppler imaging) were more impaired in Val122Ile carriers, along with greater left ventricular wall thickness and higher N-terminal pro b-type natriuretic peptide at 5 years.16 Val122Ile was also noted to be a common aetiology of heart failure in this population in Europe, where it constituted the fourth highest prevalence in Afro-Carribean patients in the United Kingdom (8.5%) and demonstrated the worst prognosis compared with other aetiologies of heart failure.17
Meanwhile, the most common variant in the United Kingdom leading to cardiomyopathy is Thr60Ala.18 This mutation was first described in Ireland and is particularly endemic in the Donegal region of north west Ireland and patients of Irish descent. In the USA, this mutation was found in 20% of THAOS patients.11 In a landmark epidemiological analysis of patients in the United Kingdom, median age of onset was 63 years and 93% presented with cardiac involvement by echocardiography. The majority also presented with autonomic or peripheral neuropathy. Median survival was 6.6 years after onset of symptoms and 3.4 years from diagnosis, with cardiac disease driving mortality.18
First described in Portugal in 1952,19 this gene variant remains endemic to certain areas of the country as well as pockets of Japan, Sweden and other countries. It is the most common mutation in the TTR gene worldwide.20 The mutation demonstrates various phenotypic expressions, especially with lower limb neuropathy (familial amyloidotic polyneuropathy). Broadly, there is a bimodal presentation with early onset disease (third and fourth decade of life) rarely causing cardiomyopathy and late onset disease (fifth and sixth decade of life) leading to cardiac involvement.21 All forms of disease typically have neuropathic symptoms. Penetrance is variable in patients with gene mutations and may be related to the particular endemic region as well as which parent provided the mutation (maternal inheritance with higher penetrance and earlier presentation).22 Mechanistically, 115–124 residues of Val30Met TTR are exposed on the molecular surface with a conformational change, and antibodies targeting these cryptic epitopes have shown to inhibit fibrillisation.23 24
AL amyloidosis is caused by deposition of immunoglobulin light chains secreted from the monoclonal proliferation of plasma cells. AL amyloidosis is now considered less common than ATTR (estimated incidence of 9 cases/million or ~3000 cases diagnosed per year in the USA).25 Cardiac diagnosis in patients with AL amyloidosis is often earlier (mean age of 65 years), and more commonly associated with female gender, lower left ventricular mass and lower electrocardiographic voltage than those with ATTR.10 Systemic and multiorgan involvement in AL is frequently encountered, with the liver, gastrointestinal tract, lung, kidney, nerve and soft tissue as potential sites of fibril deposition.26 While the level of free kappa light chains is generally higher in normal subjects, AL amyloidosis due to monoclonal lambda proliferation predominates in a 3:1 ratio.27 Interestingly, the particular subtype of lambda protein expressed may be associated with dominant organ involvement (typically heart or kidney) and prognosis.28 Recent studies have also observed that the presence of a single gene, IGVL1–44, in the variable region germline genes of lambda light chains is associated with a fivefold increase in the odds of dominant heart involvement.29 These observations support a genetic basis of cardiotropic mechanisms of amyloid deposits.
AL amyloidosis is almost exclusively an acquired condition without an obvious inherited genetic component, and often with poorer prognosis compared with ATTR (survival after onset of heart failure ~6 months if no treatment). Nevertheless, gene expression profiling has described a number of potentially significant genes and pathways found to be dysregulated in AL amyloidosis.30 Hereditary light chain amyloidosis remains a rare condition and should be considered in patients with systemic AL disease without evidence of plasma cell dyscrasia.31
Disease-modifying therapeutic approaches based on genetic insights
Pharmacological therapies for ATTR can be divided into three general strategies: blocking production of pathological TTR proteins by the liver, interfering with protein dissociation by stabilising the TTR tetramer and degrading formed amyloid fibrils (table 2). While there are no current Food and Drug Administration (FDA) approved treatments for ATTR in the USA, the most clinical data exist with the non-steroidal anti-inflammatory drug (NSAID) diflunisal and the non-NSAID benzoxazole derivative tafamidis which act as protein stabilisers, making TTR less amyloidogenic. Tafamidis is approved in Europe and is commonly used for this condition to slow disease progression.32 Other pharmaceutical agents such as epigallocatechin gallate and doxycycline-tauroursodeoxycholic acid have theoretical mechanistic benefits with limited clinical data. Monoclonal antibodies to conformation-specific TTR proteins23 33 34 and anti-serum amyloid P (SAP) component35 36 are in the pipeline, having experienced preclinical and early clinical success in clearing amyloid deposits from affected organs. In this review, we will focus on genetic suppression of amyloidogenic proteins.
Gene-silencing therapy for TTR
Gene variants that contribute to amyloid heart disease provide unique opportunities to explore potential disease-modifying therapeutic strategies, especially with ATTR. Gene-based therapy acts by blocking production of TTR by the liver; this is accomplished by harnessing small interfering ribonucleic acids (siRNAs) or ASOs.
Small interfering RNA
RNA is an essential component in the coding and expression of genes; yet, not all RNAs directly lead to protein expression. Non-coding RNA, that is, microRNA, siRNA and antisense RNA, are involved in the orchestration of gene expression, sometimes inhibiting its own function under normal biological settings through RNA interference. siRNAs are non-coding double-stranded RNAs; one strand is degraded and the other (typically the antisense RNA) acts to regulate subsequent translation and protein synthesis. Researchers have taken advantage of this normally occurring cellular process, creating siRNA molecules that inhibit pathological gene translation.
siRNAs were developed specific to TTR that target a conserved sequence in untranslated regions of non-mutant and mutant messenger RNA (mRNA). The main obstacle to creating a siRNA molecule in humans lies in the mechanism of deliver of this molecule to the liver cell where TTR is produced.37 The objective is to knock down hepatic mutant and wild-type ATTR production, thereby reducing unstable circulating ATTR tetramers and preventing organ deposition of ATTR monomers and amyloid fibrils. The ultimate goal would be to demonstrate disease regression; though, to date, this has only been shown using siRNA in a mouse model of hereditary ATTR. The extent of TTR tissue deposit regression appeared to be linearly correlated with the degree of RNA interference (RNAi)-mediated knockdown and serum TTR protein exposure.38
Ultimately, first- and second-generation lipid nanoparticles were used to encapsulate these siRNA for study in humans. A phase I trial to assess safety and effect of ALN-TTR01 (now termed revusiran, encapsulated with DLin-MC3-DMA-based lipid nanoparticle) on TTR levels with dose escalation was devised and demonstrated positive results as a proof of concept. This first-generation formulation was given to 32 patients with hereditary ATTR and demonstrated significant knockdown of TTR levels. The second-generation formulation ALN-TTR02 (now patisiran, encapsulated with GalNAc conjugate) (figure 1) was given to healthy volunteers and demonstrated a similar degree of knockdown.39 This discovery was truly ground breaking, as it was the first time that RNAi therapy was used in humans to successfully target mRNA transcribed from a disease-causing gene. This high-level publication has led the way for excitement about this potential therapy in amyloidosis and other disease processes.
Based on results of these phase I data, a phase II open-label extension trial of revusiran was initiated. Additionally, a phase II trial of patisiran was undertaken which confirmed a significant dose-dependent knockdown of the TTR protein in patients with hereditary ATTR. The dose of 0.3 mg/kg every 3 weeks was coalesced due to maximum effect and rare infusion-related side effects.39 40 The results of the phase I and II trials for siRNAs have led to phase III trials assessing longer term outcomes. The ENDEAVOUR phase III study of revusiran sought to enrol 200 patients with a primary outcome of 6 min walk distance and serum TTR levels. Secondary outcomes included cardiovascular mortality, hospitalisation and Kansas City Cardiomyopathy Questionnaire results (clinical trial identifier NCT02319005). APOLLO is a randomised, placebo-controlled study of patisiran in patients with hereditary ATTR. The primary end point is disease progression based on a modified neuropathy impairment score (mNIS +7) at 18 months. Secondary end points include quality of life and changes in other neuropathic measures (clinical trial identifier NCT01960348).
Unfortunately, the data monitoring committee recently suspended the ENDEAVOUR trial due to safety concerns. Thirteen of the 25 patients receiving revusiran in the phase II open-label extension study had developed new or worsening peripheral neuropathy, eight of which developed symptoms only after >10 months of therapy. Seven patients developed elevated blood lactate levels. The pharmaceutical company has required the cessation of revusiran in all patients in the clinical trial or extension study. Currently, the APOLLO trial using the second-generation lipid nanoparticle delivery system (patisiran) remains underway. The origin of these side effects remains unclear. Disruption of essential cellular mechanisms or the build-up of toxic precursors are possible aetiologies. An intriguing hypothesis relates mitochondrial toxicity. We know from the HIV literature that nucleoside reverse transcriptase inhibitors can interact off-target with other human DNA polymerases like mitochondrial DNA polymerase gamma. This leads to depletion of mitochondrial DNA resulting in organelle dysfunction and impairment of oxidative phosphorylation. Lactic acidosis and hepatic steatosis are feared side effects due to this process. A similar mechanism may be at work with the siRNA revusiran; further study into these troubling side effects is warranted. Not to be overlooked are the physiological effects of TTR, acting in thyroid hormone and retinol transport. It has also been found to be involved in behaviour, cognition, neuropeptide modification, neurogenesis, nerve regeneration and axonal growth.41
ASOs are another example of non-coding RNAs which coordinate gene expression and may be involved in protein silencing (figure 2). Early animal experimentation using ASOs was performed in the laboratory of Merrill Benson. A transgenic mouse model carrying the human Ile84Ser TTR mutation was developed, and animals were injected with different ASO preparations yielding varying effects on TTR levels.42 These data were used to develop an ASO for human use, branded as ISIS-TTR(Rx), targeting TTR mRNA. A phase I clinical trial utilised this second-generation antisense technology in both a human TTR transgenic mouse model and cynomolgus monkeys. Treatment in both species led to a dose-dependent reduction of >80% of both the mRNA and plasma protein levels.43 The same molecule, now termed IONIS-TTR(Rx) after the company’s change of name, was studied for its effects on both mutant and wild-type TTR in animal models as well as in healthy human volunteers. Robust reductions in protein levels were achieved in all three species. Effects were dose dependent, lasted for weeks after dosing and were well tolerated without any remarkable safety issues after 4 weeks. A once-weekly subcutaneous dosing regimen was coalesced on for future clinical trials using this agent.44
Results from this phase I study in healthy volunteers allowed for direct initiation of a phase III multicentre, international, double-blind, placebo controlled study. Currently underway, this trial has completed enrolment of patients with hereditary ATTR with neurological involvement, without symptomatic heart failure (New York Heart Association functional class ≥3 symptoms excluded). Completion of follow-up is expected in September 2017, with a primary end point of neuropathy impairment and quality of life scores (ClinicalTrials.gov Identifier: NCT01737398). Preliminary data in patients having completed dosing in this NEURO-TTR study demonstrated that both wild-type and mutant TTR levels were substantially reduced and that the reductions observed were approximately equal in magnitude regardless of the presence or location of the mutation.28 Data from a phase II open label study were also presented at the International Society of Amyloidosis conference in 2016, demonstrating MRI disease stabilisation and slight regression of LV mass in five subjects having been treated for at least 12 months.45 A study of patients with wild-type cardiac ATTR was fashioned; however, this was placed on hold by the FDA until safety results of the open-label phase II study and NEURO-TTR are available (ClinicalTrials.gov Identifier: NCT02627820).
Gene-silencing therapy for AL light chain amyloidosis
Treatment of AL amyloidosis has centred on suppression of clonal plasma cells using anti-plasma cell therapy. A combination of bortezomib, dexamethasone and an alkylating agent (typically cyclophosphamide) has emerged as the regimen of choice in AL amyloidosis with cardiac involvement, leading to decreased mortality.46 Autologous stem cell transplantation may also be beneficial, particularly after heart transplantation in patients with end-stage cardiomyopathy.47 Novel agents for AL are on the horizon, including antibodies against SAP and a cryptic epitope of the amyloid fibril.48
Gene-based therapies are in earlier stages of development in this subtype of amyloidosis as compared with ATTR. An ASO targeting the variable domain of human immunoglobulin lambda light chain was initially described in 2002. This portion of the light chain was selected, as it is found in much higher densities in amyloid fibrils than the constant region or the entire immunoglobulin.28 Application of this ASO led to inhibition of the expression of mRNA and decreased light chain levels in vitro and in vivo using a mouse model.49 Meanwhile, siRNA molecules targeting the mRNA of amyloidogenic light chains have also been developed. siRNAs to the V, J or C domains of light chains were tested in a mouse model. There was an average of 40% reduction in target mRNA and subsequent light chain synthesis.50 An in vitro/in vivo plasmacytoma mouse model of AL amyloidosis has similarly demonstrated that these siRNAs can significantly reduce local production and circulating levels of light chains.51 Subsequently, a siRNA molecule targeting the constant region of the lambda light chain successfully reduced its production in human myeloma cell lines. This led to endoplasmic reticulum (ER) stress and eventual apoptosis.52 These models highlight the therapeutic potential of siRNA for AL amyloidosis.
Although rarely affecting the heart, inflammatory amyloid A (AA) amyloidosis has also been a substrate for ASO development. After induction of AA amyloidosis by amyloid-enhancing factor and silver nitrate injection, injection of targeted ASO led to reduced peak serum AA levels. Examination of tissues by Congo red staining and serum amyloid A (SAA)/AA immunohistochemistry revealed consistently less amyloid in the organs of ASO-treated mice compared with saline-treated counterparts.30 53
Targeting the misfolded protein response
A novel therapeutic pathway involves utilising the cell’s own machinery to target misfolded proteins. When the ER finds misfolding errors, it downregulates protein synthesis and upregulates synthesis of chaperones and proteases to prevent misfolded proteins from reaching the bloodstream to form toxic amyloid deposits. The membrane-bound transcription factor activating transcription factor 6 alpha (ATF6α) plays a cytoprotective role in the unfolded protein response; these signalling pathways are responsible for regulating proteostasis within the secretory pathway in response to ER stress. ATF6 activation appears to preferentially reduce secretion of the Val122Ile variant but not wild-type TTR, thus enhancing the stringency of ER quality control.54 It is conceivable that strategies to fine tune the unfolded protein response to increase a cell’s ability to find errors in protein production could lead to enhanced treatments for human amyloid disease.
Given increased awareness and improved non-invasive diagnostic tools, cardiac amyloidosis has become an increasingly recognised aetiology of increased ventricular wall thickness and heart failure with preserved ejection fraction. With advancing mechanistic understanding from genetic and molecular technologies, novel therapeutic strategies are on the horizon. Amyloidosis has become the model disease through which gene therapy using siRNA and ASOs has evolved. While phase I and II clinical studies utilising these exciting new technologies have shown some promise, robust phase III trials with long-term evidence of clinical efficacy and robust safety data are still needed.
Contributors BWS and WHWT drafted and edited the manuscript together.
Funding WHWT is supported by grants from the National institutes of Health (R01HL103931) and Collins Family Fund.
Provenance and peer review Commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.