Silencers versus stabilizers in amyloid cardiomyopathy. Are we asking the wrong questions?

Transthyretin (TTR) is a small tetrameric molecule with a half-life of 2–2.5 days.¹ TTR cardiomyopathy is an increasingly recognized cause of progressive heart failure in the elderly and is associated with amyloid production derived from the breakdown of either wild-type TTR or, less commonly, mutant TTR. Two therapeutic mechanisms have been utilized to treat TTR cardiomyopathy, tetramer stabilization and tetramer silencing. The former is designed to lessen amyloidogenic monomers formed from TTR breakdown, which in turn can form dimers, oligomers and eventually amyloid fibrils, and the latter to significantly reduce hepatic TTR production, thereby markedly reducing the amount of breakdown products. The initial therapeutic mechanism to be addressed was TTR stabilization and in both of the two pivotal stabilizer trials of tafamidis and acoramidis respectively, treatment significantly reduced combined morbidity/mortality compared to placebo^{2, 3}. However, while both trials showed that therapy produced a significant blunting of N-terminal pro-B-type natriuretic peptide (NT-proBNP) rise and lesser deterioration of 6-min walk over time compared to placebo, neither showed an improvement in these parameters compared to baseline. In addition, some patients with congestive heart failure progressed to death despite treatment with active drug.

Failure to prevent disease progression in some patients, particularly those entering the trial as New York Heart Association (NYHA) class III, led to the postulate that TTR stabilizers may be suboptimal for producing complete cessation of amyloid production and raised the question whether a silencer, by significantly reducing TTR production, might be a more effective treatment. The possibility that this might be the case was underscored by the relative efficacy of TTR silencers compared to stabilizers in TTR familial amyloid polyneuropathy (FAP), a disease due exclusively to mutant TTR. While no direct comparison of stabilizers and silencers has been performed in this condition, post-hoc comparison of tafamidis and silencer efficacy suggests that silencers are superior in significantly slowing the relatively rapid progression of polyneuropathy associated with FAP.⁴ In addition, one study of the TTR silencer patisiran examining its efficacy in FAP showed a subtle improvement in longitudinal ventricular contraction in the subgroup with cardiomyopathy,⁵ and vutrisiran therapy in FAP showed a trend to similar results,⁶ raising the possibility that silencers might be associated with improvement of cardiac function.

The recently-published HELIOS-B trial of vutrisiran, a small-interfering RNA for the treatment of TTR cardiac amyloidosis, was the first such silencer trial to be completed and confirmed the benefit of lowering TTR in cardiac amyloidosis.⁷ Treated patients had a significantly lower combined rate of hospitalization or urgent visits for heart failure and death compared to those receiving placebo. However, contrary to the hopes that TTR silencers in amyloid cardiomyopathy may show superiority over the results of stabilizer trials, vutrisiran, although showing excellent outcome in treated patients compared to placebo, did not show improvement in biomarkers or 6-min walk in the treated group when compared to their own baseline, again suggesting no clinically significant cardiac improvement. Indeed, 43.9% of the vutrisiran-treated patients had outpatient worsening of heart failure over 36 months in the extended trial and the relative risk reduction of the primary endpoint evaluated at 30 months (the time frame of the stabilizer trials) was very similar to that seen with stabilizers⁸ (Table 1). This raises the possibility that stopping (or markedly reducing) amyloid production, whether by a stabilizer or silencer, results in a similar efficacy, with a persistent residual group of patients progressing to death or hospitalization. The failure of both stabilizers and silencers to prevent progressive heart failure and death suggests that progressive heart failure in TTR amyloid patients treated with either of these classes of drugs could be due to a mechanism other than incomplete suppression of amyloid production, and we believe that these results should lead to a critical rethinking of our concepts about this increasingly recognized disease.

What might be learned by these observations? There are always pitfalls when attempting to compare outcomes across individual trials of a disease state, and TTR amyloid cardiomyopathy is no exception. The progression of the disease is non-linear and it is being diagnosed at an earlier stage than it was in patients in the earliest trial. Nevertheless, a consistent feature of the trials seems to be that the sicker patients at trial entry, whether defined by NT-proBNP levels or NYHA class, have a poorer response rate to therapy than do less sick patients. This and other apparent similarities in outcome among all three trials raise, and may help to answer, several questions. These include: (1) Are familial TTR amyloidosis and wild-type TTR amyloidosis two sides of one disease, or should they be considered separate diseases? (2) Does the similarity in outcome between the silencer trials and the vutrisiran trial imply that the drugs are of equal efficacy in all situations? (3) If progression of heart failure in some patients receiving the appropriate silencer or stabilizer therapy is associated with progressive disease, might the mechanism be unrelated to ongoing amyloid deposition but, rather, to the inexorability of progression due to mechanisms untouched by anti-amyloid therapy?

To answer the initial question, we should first ask whether it is valid to extrapolate from outcomes in TTR neuropathy trials to cardiomyopathy trials. We would argue not. By definition, all patients in FAP trials had variant TTR, whereas the minority of the patients in the cardiomyopathy trials had a TTR variant, ranging from 9.7% to 24% in the three trials (Table 1).^{2, 3, 7} Neuropathy trials utilize a more continuous variable of progressive sensorimotor loss as an endpoint, whereas cardiomyopathy trials have more discrete endpoints of hospitalization or death. And, perhaps most convincingly, mutant TTR monomers and dimers have been shown to be neurotoxic in vivo,⁹ suggesting a major toxic component of amyloid neuropathy whereas cardiomyopathy outcomes relate to infiltrative amyloid burden. Thus silencers, which reduce mutant TTR by 80–90%, may have a logical reason to be superior to stabilizers in FAP, since mutant TTR still remains after therapy with stabilizers, albeit in a more stable form.

It is easy to understand why a single amino acid substitution, by rendering the TTR molecule unstable, might lead to dysfunctional TTR breakdown with consequent amyloidogenic monomer and oligomer formation. The commonest TTR variant-associated cardiomyopathy in the United States and also in the clinical trials is the Val122Ile variant and it is noteworthy that Val122Ile has a more rapid course and poorer prognosis than wild-type TTR,^{10, 11} suggesting that tetramer instability plays a role in prognosis, an observation that suggests caution in ‘lumping’ the two conditions together. But why does wild-type TTR result in amyloid formation in some people but not in others? Despite extensive literature on TTR amyloidosis, the mechanism of amyloidogenesis in wild-type TTR amyloidosis is poorly understood. Among patients with wild-type TTR cardiomyopathy, the circulating TTR may be slightly more unstable than among subjects without TTR amyloidosis, but this difference is not striking compared to the instability of variant TTR.¹² This raises the question as to whether amyloid formation in the absence of a mutant protein is entirely a function of intrinsically unstable wild-type protein or whether other factors play a role.¹³ For example, it has been suggested that age-related oxidative modifications of TTR may be a contributing factor to amyloidogenesis in these patients.¹⁴

Given the probable differences in mechanisms of amyloidogenesis and the difference in clinical presentation (significant amyloid neuropathy in mutant TTR with minimal or no neuropathy in wild-type TTR), we would argue that wild-type and variant TTR amyloidosis should be considered two different diseases which, although responding to the same treatments, have different prognoses and which do not allow extrapolation of outcomes one to another. Such a concept should encourage the further exploration of subcellular mechanisms of TTR breakdown, perhaps leading to the development of novel downstream therapies to prevent monomer amyloidogenesis. These include novel drugs designed to prevent the cascade of amyloid seeding triggered by misfolded monomers.¹⁵

Despite the outcomes of the cardiomyopathy trials, the data that TTR silencers are more effective than stabilizers when treating FAP are compelling—why might this be? There is good evidence that pre-amyloid oligomers in FAP are neurotoxic, whereas the neurotoxicity of such oligomers derived from wild-type TTR manifests no or significantly diminished toxicity to the nerves.⁹ This would explain the lack of clinically significant neuropathy among patients with wild-type TTR amyloid cardiomyopathy. We have previously argued that FAP is predominantly a toxic condition with some degree of infiltration whereas amyloid cardiomyopathy is an infiltrative condition with little cardiac toxicity whether in familial or wild-type cardiac amyloidosis.¹⁶ It also seems that tafamidis and acoramidis, even if they differ in degree of TTR stabilization (a debatable and difficult parameter to measure since assays are usually done in vitro at a non-physiologic pH¹⁷) have a very similar clinical effect. This would occur if there was a threshold effect of protein stabilization adequate to prevent further amyloid formation and such that greater stabilization has no further effect on amyloid production.

What then, if the stabilizers stop amyloid production, would account for progressive heart failure in some patients in both the stabilizer trials and the silencer trials? Examination of subgroups of patients in these trials seems consistently to show that patients with more extensive disease are more likely to progress than those with less extensive disease. Histologic studies from autopsy or endomyocardial biopsy in patients with TTR amyloidosis show that the extracellular space (the site of amyloid deposition) not only contains amyloid fibrils but that, in the advanced cases, demonstrates considerable fibrosis and collagen formation.¹⁸ This would not be expected to respond to anti-amyloid production agents and, if present in large enough amounts, would account for the inexorable progression of more advanced patients with cardiac amyloid infiltration who presumably have extensive cardiac disease including extracellular fibrosis. It would also account for the apparent lack of superiority of silencers over stabilizers in amyloid cardiomyopathy no matter how low one can get TTR levels, since tackling amyloid production alone in advanced patients is only addressing a portion of the pathology.

Where does this lead us? It may well be that we have achieved the maximum that can be done with drugs designed to reduce amyloid production. To this end, it would be unlikely that the gene-editing therapy in clinical trials will show any additional benefit over what we can already achieve, since the effect on lowering TTR levels is similar to the current generation of silencers.¹⁹ In contrast, antibodies derived against the amyloid or its precursors (termed ‘depleters’) offer an opportunity to decrease the extracellular space and potentially improve cardiac function, even if there will still inevitably be residual non-amyloid pathology remaining in some subjects.²⁰ Potentially adding to our understanding of the disease are new imaging agents such as florbetapir that can quantify the amyloid burden, (as opposed to ‘simply’ measuring cardiac extracellular volume) which may help identify patients at increased risk of progressive disease despite therapy with stabilizer or silencer therapies.^{21, 22} Additional information may come from the use of ⁶⁸Ga-labelled fibroblast activation protein inhibitor (FAPI) for imaging of active fibrosis, currently being studied in light-chain amyloid cardiomyopathy, in which condition the presence of a positive scan (indicative of active extracellular fibrosis deposition) is correlated with a poorer prognosis.²³ Hopefully, if these imaging agents are incorporated into ongoing large clinical trials, they may give us a greater understanding of the pathophysiology of heart failure in cardiac amyloidosis, and can potentially predict, early after diagnosis, which patients are more or less likely to progress despite anti-amyloid therapies (Figure 1).

The above argument, that fibrosis or other non-amyloid changes in more clinically advanced TTR cardiac amyloidosis is a major reason for progressive heart failure and death remains an hypothesis, but one that we believe fits the emerging data. Since FAPI (fibroblast) imaging is still in an early stage of investigation in all types of cardiac amyloidosis, and since it requires specialized equipment and expertise, it is not yet feasible to incorporate it into large multicentre clinical trials for all patients, but evaluation by participating centres with high recruitment and appropriate imaging experience would be of great value. Should depleters prove to be effective, such a substudy may help determine whether there is equal, greater, or lesser benefit in those patients deemed to have substantial active fibrosis at study entry. In addition, such imaging may help to determine the relationship between markedly increased NT-proBNP or troponin and the presence of non-amyloid cardiac pathology, elucidating the role of biomarkers as a possible surrogate for such pathology, thus allowing for better stratification in future trials.

While it seems fairly likely that we have reached our maximum ability to decrease mortality and morbidity in TTR amyloid cardiomyopathy by using agents that stop amyloid production, the ‘new frontier’ will hopefully not only be the addition of depleter agents to remove existing amyloid deposits,²⁰ but also an intensive reconsideration of why patients with the wild-type protein develop this disease, a recognition that heart failure progression does not always mean (and maybe rarely means) ongoing amyloid deposition, and an understanding of what precisely accounts for the apparent non-responders to current and investigational therapies. It behooves the pharmaceutical industry sponsoring future clinical trials in TTR amyloidosis to not only design trials to determine the efficacy of the novel agents studied, but also to spend the extra effort to utilize the trials to further elucidate the mechanisms of cardiac damage and dysfunction in this increasingly recognized disease. Such effort, in the end, will benefit patients, industry and clinicians alike.

Conflict of interest: none declared.