Diflunisal versus tafamidis on neuropathy and cardiomyopathy in hereditary transthyretin amyloidosis

Abstract

Objectives: Hereditary transthyretin (TTR) amyloidosis (ATTRv) is frequently complicated by polyneuropathy (ATTRv-PN) and cardiomyopathy (ATTRv-CM). The long-term efficacy of diflunisal on both polyneuropathy and cardiomyopathy in ATTRv patients, especially those with non-V30M genotypes, has not been fully investigated and compared with that of tafamidis.

Methods: We compared the structural and biochemical characteristics of A97S-TTR complexed with tafamidis with those of diflunisal, and prospectively followed up and compared the progression of polyneuropathy and cardiomyopathy between ATTRv-PN patients taking diflunisal and those taking tafamidis.

Results: Both diflunisal and tafamidis effectively bind to the two thyroxine-binding sites at the A97S-TTR dimer-dimer interface and equally and almost sufficiently reduce amyloid fibril formation. Thirty-five ATTRv-PN patients receiving diflunisal and 22 patients receiving tafamidis were enrolled. Compared with no treatment, diflunisal treatment significantly delayed the transition of FAP Stage 1 to 2 and Stage 2 to 3 and decreased the deterioration in parameters of the ulnar nerve conduction study (NCS). The progression of FAP stage or NCS parameters did not differ between patients treated with diflunisal and those treated with tafamidis. Both diflunisal and tafamidis treatments significantly decreased radiotracer uptake on ^99mTc-PYP SPECT and stabilized cardiac wall thickness and blood pro-B-type natriuretic peptide levels. No significant adverse events occurred during diflunisal or tafamidis treatment.

Interpretations: The binding patterns of both tafamidis and diflunisal to A97S-TTR closely resembled those observed in the wild type. Diflunisal can effectively delay the progression of polyneuropathy and cardiomyopathy with similar efficacy to tafamidis and may become a cost-effective alternative treatment for late-onset ATTRv-PN.

Figure 1

Structural analyses of stabilizer binding sites in A97S‐TTR. (A) NMR studies on tafamidis and diflunisal binding to A97S‐TTR. Comparison of 2D TROSY‐HSQC spectra: ¹⁵N‐labeled A97S‐TTR with (purple) and without (black) tafamidis. Intensity fluctuations observed in individual residues of A97S‐TTR upon tafamidis binding. (B) 2D TROSY HSQC spectra of 15N‐labeled A97S‐TTR with (orange) and without (black) the presence of diflunisal. Changes in the intensity of each residue of A97S‐TTR due to diflunisal binding. The ratio of NMR signal intensities, denoted as I/I₀, refers to the signal intensity of stabilizer‐bound A97S‐TTR. (C) Crystal structures of the A97S‐TTR ligand complexes. Global view of TTR‐A97S bound to tafamidis (PDB ID: 8YQD). The overall structure is represented as a white cartoon, and tafamidis is shown as purple sticks. (D) Close‐up view of one of the tafamidis binding sites in the TTR‐A97S/tafamidis complex structure. Two different binding modes of tafamidis are shown as purple sticks and lines. (E) Two different binding modes of diflunisal are shown as orange sticks and lines. The side chains of TTR‐interacting residues are labelled and represented by white sticks. (F) Both tafamidis and diflunisal effectively inhibited the acid‐mediated aggregation of A97S‐TTR. The A97S‐TTR samples were incubated in the absence or presence of tafamidis or diflunisal at a molar ratio of 1:2 (A97S:drug) at pH 4.0 for 6 days, and the acid‐induced fibrils of A97S‐TTR were quantified using Congo red (CR) dye.

Figure 2

The transition from the FAP stage and the changes in ulnar motor nerve neurophysiology between ATTRv‐PN patients treated with diflunisal and those not treated. (A) The Kaplan–Meier failure curve for the cumulative portion and time interval of transition from FAP Stage 1 to 2 showed that the use of diflunisal significantly delayed the transition from FAP Stage 1 to 2 (hazard ratio [HR], 0.43; 95% confidence interval [CI], 0.23–0.79; p = 0.007). (B) Similarly, the use of diflunisal significantly delayed the transition from FAP Stage 2 to 3 in ATTRv‐PN patients (hazard ratio [HR], 0.18; 95% confidence interval [CI], 0.08–0.43; p < 0.001). (C) The reduction in the amplitude of the ulnar compound muscle action potential (CMAP) per month (%) from baseline was significantly lower in patients treated with diflunisal than in treatment‐naive patients (p < 0.001). (D) The decrease in the ulnar motor conduction velocity (%) per month from baseline was significantly lower in patients treated with diflunisal than in treatment‐naive patients (p = 0.027).

Figure 3

The transition from the FAP stage and the changes in ulnar motor nerve neurophysiology between ATTRv‐PN patients treated with diflunisal and those treated with tafamidis. (A) The Kaplan–Meier failure curve for the cumulative portion and time interval of transition from FAP Stage 1 to 2 showed no difference in progression between patients treated with diflunisal and those treated with tafamidis (p = 0.332). (B) Similarly, the progression from FAP Stage 2 to 3 did not differ between patients receiving diflunisal and those receiving tafamidis (p = 0.993). (C) The reduction in the amplitude of the ulnar compound muscle action potential (CMAP) per month (%) from baseline (p = 0.889) and (D) the decrease in the ulnar motor conduction velocity per month from baseline (p = 0.623) were similar in patients treated with diflunisal and tafamidis.

Figure 4

Changes in the volumetric H/CL ratio on 99mTc‐PYP SPECT imaging in ATTRv‐PN patients treated with diflunisal and tafamidis. The volumetric H/CL ratio on 99mTc‐PYP SPECT imaging decreased significantly between baseline and follow‐up after the diflunisal (A, p = 0.007) and tafamidis (B, p < 0.001) treatments.