Differential Binding Affinities and Kinetics of Transthyretin Stabilizers

Abstract:

Transthyretin amyloid cardiomyopathy (ATTR-CM) is a progressive, fatal disease. Dissociation of tetrameric transthyretin (TTR) is the triggering event in the pathogenic mechanism; destabilizing TTR mutations accelerate the process. The TTR stabilizers, tafamidis and acoramidis, are the only United States Food and Drug Administration (FDA) approved treatments for patients with ATTR-CM. By mimicking the stabilizing characteristics of the super-stabilizing, disease-protecting variant T119M, we hypothesize that acoramidis displays differential TTR binding, kinetic stability, and tetramer stabilization compared with other TTR stabilizers, such as tafamidis and diflunisal. The TTR binding affinity and thermodynamic stability of TTR interaction of acoramidis and tafamidis were assessed by surface plasmon resonance and microscale thermophoresis (MST). Tetrameric TTR stabilization by acoramidis, tafamidis, and diflunisal in the presence of plasma proteins against acidic denaturation was measured by immune blots. In kinetic studies, surface plasmon resonance demonstrated 4 times longer residence time for acoramidis bound to TTR than tafamidis. The dissociation constants were consistent with those determined by equilibrium measurements in MST. The affinity of acoramidis for purified TTR, as measured by MST, was 4 times higher than that of tafamidis. When tested at clinically relevant plasma concentrations, acoramidis stabilized TTR against acidic denaturation to a much higher extent (≥90%) than tafamidis or diflunisal. Of note, both tafamidis and diflunisal demonstrated partial stabilization of tetrameric TTR. Relative to other stabilizers, acoramidis is more potent as independently assessed by TTR binding affinity, kinetic stability, and acid-mediated denaturation. These properties may contribute to the ability of acoramidis to achieve near-complete stabilization of TTR in plasma samples.

INTRODUCTION

Transthyretin (TTR) amyloid cardiomyopathy (ATTR-CM) is a progressive, fatal disease characterized by destabilization of TTR tetramers resulting in dissociation into monomers.¹ The monomers misfold and aggregate into toxic amyloid precursors, which form fibrils that deposit in the myocardium causing progressive heart failure and death.¹ Dissociation of TTR into monomers can be prevented by binding of small molecules in the thyroxine binding sites of the intact TTR tetramer, thereby stabilizing the TTR tetramers.^2,3

Tafamidis (Vyndaqel, Vyndamax; Pfizer; New York, NY) and acoramidis (Attruby; BridgeBio; Palo Alto, CA) are 2 oral TTR stabilizers indicated for the treatment of ATTR-CM and have demonstrated clinical benefit in reducing death and cardiovascular hospitalization.^4–7 Even though the relationship between the molar ratio of tafamidis:TTR tetramer and TTR tetramer stabilization (based on a population pharmacokinetic–pharmacodynamic analysis) is adequately described by a sigmoid Emax model,^8,9 the relationship between dose, exposure, or extent of TTR stabilization of tafamidis and its clinical benefit could not be established. There are significant differences between the clinical^4,5 and pharmacologic^10,11 profiles of the 2 TTR stabilizers. Both agents are indicated for improving survival and decreasing hospitalization in patients with ATTR-CM.^4,5 Contemporary real-world evidence after long-term outcomes in patients treated with tafamidis shows a differential profile for probability of all-cause mortality from those treated with acoramidis.¹² As in vitro characterization reported under varied experimental conditions involves different methodologies (isothermal titration calorimetry, amyloid fibril formation, FLAG-tagged protein subunit exchange assay), it is difficult to perform a head-to-head comparison of the pharmacologic potency of tafamidis and acoramidis. Furthermore, the 2 compounds bind to plasma proteins to different extents, and unequal free fractions of each molecule are available for target interaction at fixed plasma concentrations.¹³

This research study was conducted to further explore the comparative activity of these agents beyond reported literature and investigate whether the extent and duration of TTR binding, potency of stabilization, and activity in the presence of plasma proteins may be indicative of differential therapeutic benefit seen in patients with ATTR-CM.

METHODS

Surface Plasmon Resonance

Purified TTR was coupled on CM-5 chips through N-hydroxysuccinimide/1-ethyl-3-(3-dimethylaminopropyl) carbodiimide treatment. Four independent experiments were performed for acoramidis and tafamidis using purified recombinant wild-type human TTR (AlexoTech AB; Västerbottens län, Sweden) or TTR purified from human plasma (Athens Research & Technology; Athens, GA). Acoramidis (Eidos Therapeutics; San Francisco, CA) and tafamidis (Cayman Chemical; Ann Arbor, MI) were prepared as 10 mM stock solutions in dimethyl sulfoxide (DMSO) and tested at concentrations ranging from 3.91 to 1000 nM by multicycle kinetics. Experiments were performed using PBS-P+ buffer (Cytiva; Marlborough, MA) and 2% DMSO in a Biacore T200 instrument. Flow cell temperature was 25°C, flow rate was 30 µL/min, the association phase was 150 seconds, and the dissociation phase was 600 seconds. Sensorgrams were double referenced to blank ligand surfaces and blank analyte buffer injections.

Microscale Thermophoresis

The free cysteine (p.Cys30/Cys10) in recombinant TTR (AlexoTech) was labeled with cysteine-reactive second-generation RED dye (NanoTemper Technologies; South San Francisco, CA) according to the manufacturer's instructions. Premium capillaries were filled with solutions containing 20 nM fluorescently labeled TTR and either acoramidis or tafamidis at concentrations of 0.3 nM–10 µM in PBS-P+ buffer (Cytiva) and 2% DMSO. The dissociation constants (K_D) of TTR to stabilizer interactions were determined using a NanoTemper Monolith NT.115 instrument. MST experiments were performed at 24°C using 20% excitation laser power and 40% infrared laser power. Four independent experiments were performed for each compound.

Western Blot Analysis

In vitro TTR stability in plasma against acidic denaturation was determined as described previously.¹¹ In brief, pooled plasma from healthy human donors (Innovative Research; Novi, MI) was incubated with acoramidis (0.5–10 µM), tafamidis (5–100 µM), or diflunisal (40–800 µM; Sigma) for 2 hours. Samples were acidified for 72 hours, cross linked by glutaraldehyde, and the proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Immune detection was performed using polyclonal rabbit anti-TTR antibody (Dako/Agilent; Santa Clara, CA).

Data Analysis

Results of surface plasmon resonance (SPR) experiments were analyzed using Biacore T200 software. The kinetic parameters (k_on and k_off) were calculated from a global fit to a single-site binding model. The equilibrium dissociation constant (K_D) was calculated by dividing k_off by k_on; residence time was the reciprocal of k_off. PALMIST software was used for analysis of MST results.¹⁴ The combined intensity of TTR bands (TTR tetramer and tetramer bound to retinol-binding protein) in Western blots was quantified using an Odyssey IR imaging system (LI-COR Biotech; Lincoln, NE) and reported as a percentage of native TTR tetramer relative to control. Tetrameric TTR remaining in the presence of stabilizer was plotted against a logarithmic scale of ligand concentration for comparison between acoramidis, tafamidis, and diflunisal.

RESULTS

Kinetic stability of the complex of TTR with tafamidis and acoramidis was assessed by SPR. The slope of the sensorgrams during the association phase of the experiment indicates how rapidly the tested molecule binds to TTR. In this study, the slopes of acoramidis sensorgrams were steeper than those of tafamidis sensorgrams at the same concentration. After the sensorgrams reached a saturating value, the formed TTR–acoramidis complex dissociated at a slower rate than the TTR–tafamidis complex (Fig. 1, A and B). Analysis of the kinetic data in a single-site model yielded the dissociation (k_off) and association rate constant (k_on), and the data are summarized in Table 1. The ratio of k_on and k_off for each curve provides an estimate for the dissociation constant (K_D) for the 2 stabilizers; averages for the 4 experiments per stabilizer are given in Table 1. The data demonstrate a 7-fold higher affinity of acoramidis than tafamidis for tetrameric TTR. Moreover, the lower dissociation rate constant of acoramidis relative to tafamidis indicates a potential for higher persistence of the complex, suggesting a longer lasting pharmacologic effect. This is clearly observed in the 4-fold longer residence time of acoramidis than tafamidis.

FIGURE 1.

SPR characterization of TTR-binding kinetics. Representative SPR sensorgrams from an immobilized purified human TTR surface probe with either (A) acoramidis or (B) tafamidis. Colored lines are traces for each compound at decreasing concentrations from top to bottom as follows: 1000, 500, 250, 125, 62.5, 32.3, 15.6, 7.8, and 3.9 nM. Black lines correspond to the fitted kinetic parameters. Four independent experiments were performed for each compound. RUs, resonance units.

TABLE 1.

Kinetic Parameters of Interaction of Acoramidis and Tafamidis With Purified Human TTR (Mean ± SD)

Compound	kon (1/Ms)	koff (1/s)	Residence Time (s)	KD (nM)
Acoramidis	1.27 × 106 ± 2.58 × 105	1.99 × 10−2 ± 2.21 × 10−3	50.70 ± 5.39	15.95 ± 2.46
Tafamidis	6.90 × 105 ± 1.32 × 105	8.56 × 10−2 ± 2.64 × 10−2	12.49 ± 3.54	123.75 ± 26.85

A thermodynamic binding affinity measurement based on MST was used to complement the kinetic SPR method. The free cysteine residue in TTR enables selective labeling of TTR without effect on the binding site of stabilizers. In addition, unlike in SPR where TTR is tethered to a chip, MST allows for solution phase measurement without immobilization of either ligand. Figure 2A, B shows the thermophoretic movement and corresponding changes in fluorescence over time of labeled TTR in the presence of acoramidis and tafamidis. The fluorescence changes observed in MST suggest different conformation of ligand–TTR complex for the 2 compounds. Binding of acoramidis to TTR results in a conformational change causing a change in signal output. This effect increases with higher acoramidis concentrations and was monitored over time. The smooth curves demonstrate the absence of potential artifacts by solubility problems. The normalized fluorescence changes in the marked sections were analyzed by regression to log-logistic curves, shown in Figure 2C. Similarly, the binding of tafamidis to TTR resulted in a smaller alteration of the fluorescence. Table 2 confirms 4-fold higher binding affinity of acoramidis than tafamidis for TTR when measured under equilibrium conditions.

FIGURE 2.

Determination of TTR binding affinity by MST quantitation. MST traces and fluorescence change caused by (A) acoramidis or (B) tafamidis binding to TTR. Traces are colored by compound concentration with the lowest (0.3 nM) in blue and the highest (10 µM) in red. The blue vertical bar indicates baseline, and the red vertical bar indicates the data analysis window. C, Relative fluorescence changes of acoramidis and tafamidis bound to TTR. Mean ± SD from 4 measurements is plotted. The dissociation constants (K_D) were determined by regression to the log-logistic curves. RFUs, relative fluorescence units.

TABLE 2.

Binding Affinity of Acoramidis and Tafamidis to Purified Human TTR (Mean ± SD)

Compound	KD (nM)
Acoramidis	26 ± 7
Tafamidis	105 ± 21

The ability of tafamidis, acoramidis, and diflunisal to stabilize TTR in human plasma was characterized by Western blot assay in plasma samples. The stabilizing effect of the compounds in these assays results from their binding affinity for TTR and the partitioning of the compounds between TTR and albumin. Each compound exhibited a concentration-dependent increase in TTR stabilization (Fig. 3A). A schematic of this assay is shown in Figure 3B and a representative Western blot is shown in Figure 3C. Acoramidis exhibited greater TTR stabilization at lower concentrations than tafamidis and diflunisal. Tafamidis did not achieve near-complete stabilization at 100 µM, well above the peak circulating concentration (26 µM) of its approved clinical dose.

FIGURE 3.

In vitro Western blot quantitation of tetrameric TTR incubated with stabilizers. A, Each point represents the percentage of tetrameric TTR remaining after 72 hours incubation at pH 3.8, (acoramidis, n = 7; tafamidis, n = 4; diflunisal, n = 3). Concentration is represented on a logarithmic scale. Mean (labeled data points) and SD (bars) shown. B, Schematic representation of the Western blot assay. C, Representative Western blots of human pooled plasma samples with acoramidis, tafamidis, or diflunisal added at the indicated concentrations. Brackets denote the molecular weight ranges for tetrameric TTR and TTR bound to retinal binding protein. Duration of sample acidification, either 72 or 0 hours, is indicated above. Plasma samples with DMSO or PBS added were included as controls. kDa, kilodaltons; PBS, phosphate buffered saline; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

DISCUSSION

The binding of tafamidis and acoramidis to TTR was characterized by kinetic and equilibrium measurements (SPR and MST). In addition, the stabilization of TTR tetramers by these compounds against acidic denaturation was determined and compared with diflunisal. Based on results reported for diflunisal as a less potent TTR stabilizer than tafamidis and acoramidis,^11,15 we did not include diflunisal in the purified protein experiments. Previous studies have characterized the kinetics of the interaction of tafamidis and related compounds with TTR.^3,11,16 However, diverse conditions, protein preparations, and data analysis complicate interstudy comparisons. For example, tafamidis SPR data were fit to a 2-site binding model by Corazza et al,¹⁶ preventing comparison with SPR data in Alhamadsheh et al,³ where TTR stabilizers such as diclofenac were fit to a 1-site model. Similarly, when diflunisal was tested by isothermal calorimetry, TTR dissociation constants varied between 75 and 900 nM depending on model assumptions (1 site vs. 2 sites).^15,17 This report compares the agents under consistent conditions. The most differentiating factor in our results is the 4-fold increased residence time of acoramidis compared with tafamidis. This difference in dissociation rate constant translates to 7 times higher affinity of acoramidis for TTR relative to tafamidis as measured by SPR. The increased residence time of acoramidis may be favorable to clinical efficacy relative to tafamidis.¹⁸ For nonsteroidal anti-inflammatory drugs, which include diflunisal, very short residence times have been reported,^3,19 and are consistent with suboptimal efficacy of diflunisal for treatment of ATTR-CM.²⁰ The slower dissociation of acoramidis than tafamidis from TTR is also consistent with molecular dynamics simulations, which suggest that tafamidis has a lower free energy barrier than acoramidis for dissociating from TTR.²¹ The ability of acoramidis, and inability of tafamidis, to form hydrogen bonds with TTR at both the inner and outer regions of the thyroxine binding pocket may contribute to the higher TTR binding affinity and longer residence time of acoramidis compared with tafamidis.^11,22 The dissociation constants determined kinetically by SPR were confirmed by equilibrium measurements in MST experiments. The difference in the relative fluorescence changes after binding to TTR might relate to the different conformation of the formed complex.

The primary limitation of this study is that the analyses were performed in vitro. Second, analyses in the presence of plasma proteins were performed with wild-type pooled plasma and not in samples from patients with ATTR-CM samples. The differences in stabilization of TTR in plasma by the 3 compounds against acidic denaturation seems to reflect the difference in the affinities to target. However, it also reflects the different partitioning of the compounds between albumin and TTR as reflected by the ratio of the dissociation constants of compound with TTR and albumin (acoramidis, 1980; tafamidis, 580; diflunisal, 16).¹³ Data from the Western blot analysis are consistent with those from Nelson et al,¹³ demonstrating more favorable partitioning of acoramidis between albumin and TTR; acoramidis is 4 times more potent than tafamidis at a fixed plasma concentration (10 µM). Results in this report also challenge the reported hypothesis that tafamidis reaches near-maximum effect on TTR stabilization,²³ and instead confirm that clinical doses of tafamidis result in only partial stabilization of the protein. For continuous TTR stabilization, steady-state trough may be more clinically impactful than peak pharmacologic concentrations. The difference in TTR stabilization between ∼10 µM (acoramidis trough at indicated dose) versus 16 µM (tafamidis trough at indicated dose) supports the hypothesis that greater target specificity leads to a lower circulating concentration requirement for a near complete (≥90%) TTR stabilization. Further data supporting this in vitro comparison are seen in a post hoc analysis of the phase 3 clinical trial ATTRibute-CM. Acoramidis achieved near-complete (≥90%) TTR stabilization in participants with ATTR-CM, which was substantially greater than TTR stabilization with tafamidis as seen in participants in the placebo group who received tafamidis as a concomitant medication.²⁴ In ATTRibute-CM, TTR stabilization with acoramidis translated into an increase in serum TTR levels that resulted in a significant difference over placebo in a 4-component hierarchical analysis of all-cause mortality, cumulative frequency of cardiovascular-related hospitalization, change from baseline in NT-proBNP, and 6-minute walking distance (P < 0.0001).⁴ These clinical results are consistent with the design principle of the acoramidis molecule. Acoramidis binds to tetrameric TTR and induces a conformation of TTR similar to that seen in individuals harboring the super-stabilizing, disease-protecting variant T119M.²⁵ General population-based genetics studies have revealed elevated concentrations of serum TTR in individuals carrying T119M variants, associating relative increases in serum TTR levels with improved survival.^26,27

CONCLUSIONS

In summary, the results of this in vitro study provide support for clinical results that differentiate between the efficacy of acoramidis and tafamidis in patients with ATTR-CM.