Inhibition of human transthyretin aggregation by non-steroidal anti-inflammatory compounds: a structural and thermodynamic analysis

Abstract

Transthyretin (TTR) is a homotetrameric protein that circulates in plasma and cerebral spinal fluid (CSF) whose aggregation into amyloid fibrils has been associated with at least two different amyloid diseases: senile systemic amyloidosis (SSA) and familial amyloid polyneuropathy (FAP). In SSA aggregates are composed of WT-TTR, while in FAP more than 100 already-described variants have been found in deposits. Until now, TTR-related diseases have been untreatable, although a new drug called Tafamidis has been approved only in Europe to specifically treat V30M patients. Thus, new strategies are still necessary to treat FAP caused by other variants of TTR. TTR has two channels in the dimer interface that bind to the hormone thyroxin and that have been used to accommodate anti-amyloidogenic compounds. These compounds stabilize the tetramers, rendering TTR less amyloidogenic. Here, we investigated the effects of three non-steroidal anti-inflammatory compounds-sulindac (SUL), indomethacin (IND) and lumiracoxib (LUM)-as tetramer stabilizers and aggregation inhibitors. WT-TTR and the very aggressive TTR variant L55P were used as models. These compounds were able to stabilize TTR against high hydrostatic pressure (HHP), increasing the ΔGf by several kcal. They were also effective in inhibiting WT-TTR and L55P acid- or HHP-induced aggregation; in particular, LUM and IND were very effective, inhibiting almost 100% of the aggregation of both proteins under certain conditions. The species formed when aggregation was performed in the presence of these compounds were much less toxic to cells in culture. The crystal structures of WT-TTR bound to the three compounds were solved at high resolution, allowing the identification of the relevant protein:drug interactions. We discuss here the ligand-binding features of LUM, IND and SUL to TTR, emphasizing the critical interactions that render the protein more stable and less amyloidogenic.

Figure 1

Chemical structures of non-steroidal anti-inflammatory compounds investigated here as possible inhibitors of transthyretin (TTR) aggregation. The structure of diclofenac is also shown for comparison.

Figure 2

Evaluating the thermodynamic stability of WT-TTR and L55P against HHP in the absence or presence of SUL and IND. WT (panel A) or L55P (panel B) were incubated at 1 μM in the absence (circles) or presence of 10 μM SUL (triangles) or IND (diamonds) at pH 7.5, 1 °C. Samples were subjected to a stepwise increase in pressure for 10 min at each pressure value before collecting tryptophan emission spectra (excitation = 280 nm and emission = 300 to 400 nm). The insets show the linear regressions from which the thermodynamic parameters were calculated according to Equations 5 and 6 in Material and Methods.

Figure 3

LUM, SUL and IND inhibit HHP-induced aggregation. WT-TTR (panel A; 3.5 μM) and L55P (panel B; 1 μM) were compressed for 1 h at 3 kbar at 1 °C, pH 5, in the absence or presence of LUM, IND or SUL. After this time, pressure was removed, and the temperature was increased to 37 °C, at which aggregation is triggered. Then, the light scattering (LS) was recorded and normalized to the initial LS value. Aggregation in the absence of any addition is shown as circles, while LUM is shown as squares, IND as diamonds and SUL as triangles. Filled symbols refer to a TTR:compound molar ratio of 1:10, while hollowed symbols represent the 1:2 TTR:compound molar ratio. LS was collected by exciting the samples at 320 nm and colleting the scattered light from 315 to 325 nm. Immediately after the end of the aggregation kinetics, an aliquot of each sample was mounted onto mica and analyzed by AFM. WT-TTR and L55P (panels C and E, respectively) aggregated in the absence of any addition, and WT-TTR and L55P (panels D and F) aggregated in the presence of 8 and 2 μM LUM (1:2 molar ratio).

Figure 4

LUM completely inhibits the aggregation of T₄*-WT-TTR but only partially inhibits the aggregation of T₄*-L55P. Three-point-five micro molar per liter WT-TTR (panel A) or 1 μM L55P (panel B) were compressed at 3 kbar, 1 °C, pH 5, for 60 min in the absence of any addition. After this time, the pressure was released, the samples were kept on ice and LUM was added in a 1:10 TTR:LUM molar ratio. In the control samples, the same volume of buffer was added. Then, the temperature was raised to 37 °C to trigger aggregation, and the light scattering was recorded (panels A and B). Circles represent the aggregation of WT-TTR or L55P in the absence of any addition, while squares represent the aggregations after the addition of LUM. The oligomeric state of the species of WT-TTR (panel C) or L55P (panel D) formed after aggregation phases was evaluated by SEC-HPLC (dotted line). The continuous line shows the profiles of the samples before compression, where a homogenous population of tetramers was observed. For clarity, Panel E shows the experimental scheme adopted here.

Figure 5

LUM, SUL and IND inhibit the acid-induced aggregation of TTR in a dose-dependent manner. WT-TTR (panel A, 3,5 μM) and L55P (panel B, 1 μM) were incubated in the absence (circles) or presence of LUM (squares), IND (diamonds) or SUL (triangles) at a molar ratio of 1:2 TTR:compound at pH 4.4 and 37 °C, and turbidity (400 nm) was evaluated over time. One hundred percent aggregation was assigned to the turbidity of the samples aggregated in the absence of any compound after 80 h under aggregating conditions. Panels C (WT-TTR) and D (L55P) show the extent of aggregation as measured by turbidity in the presence of increasing concentrations of the compounds (dose-dependent curves) LUM (squares), IND (diamonds) and SUL (triangles). The insets of panels C and D show the percent of Congo red binding to the samples aggregated in the absence or presence of each compound after 72 h under aggregating conditions. The morphology of WT-TTR aggregates after 72 h under acidic conditions was analyzed by TEM; Control (panel E) and in the presence of 8 μM SUL (panel F), 8 μM IND (panel G) or 8 μM LUM (panel H). Note the presence of amyloid fibrils in the control sample and their absence when aggregation was performed in the presence of the three compounds.

Figure 6

The inhibitory activity of LUM depends on the presence of tetramers in solution. 1 μM WT-TTR (panel A) and L55P (panel B) were incubated at pH 4.4 and 37 °C in the absence (circles) or presence of 8 μM LUM (triangles). In parallel, solutions with the same protein concentrations were left to aggregate in the absence of LUM. After 2 h (crosses), 10 h (squares) or 20 h (diamonds), 8 μM LUM was added. Aggregation was monitored by turbidity and was expressed as the percent of aggregation, setting as 100% the aggregation of the solutions in the absence of LUM at 72 h. Panel C shows the percentage of tetramers present during the aggregation kinetics experiment for WT-TTR (hollowed symbols) and L55P (filled symbols) in the absence (circles) or presence of 8 μM LUM (squares). It is possible to see that peak decreases corresponding to the elution of tetramers as aggregation proceeds in the chromatograms (samples aggregated in the absence of LUM) shown on the right.

Figure 7

The species formed in the aggregation pathway in the presence of LUM, IND or SUL are innocuous to N2a cells. Aggregates composed of WT-TTR and L55P were produced either under acidic conditions (pH 4.4, 37 °C, 72 h; panel A) or by HHP-treatment (3 kbar at pH 5 for 1 h; panel B), either in the absence or in the presence of LUM, SUL and IND in a molar ratio of 1:2. Then, N2a cells were exposed for 24 h to 4 μM of each aggregate, and cell viability was evaluated by MTT reduction assays. The percentages of viable cells were calculated, setting the cells treated only with culture medium as 100% viable. No toxic effect was observed when cells were treated with each of the three compounds at up to 35 μM. Statistical analysis was performed in three independent experiments using one-way ANOVA with Tukey’s test, and *** p < 0.001; ** p < 0.01 and * p < 0.05.

Figure 8

Overall view of the TTR crystal structures. The TTR crystal structure was solved in the apo form (red) and in the presence of LUM (green), IND (blue) and SUL (black), and the structures of the dimer from the asymmetric unit were superposed for purposes of comparison. (A) Backbone; (B) Backbone and side chains.

Figure 9

Ligand binding to TTR. The crystal structure of the TTR tetramer was generated by the application of symmetry operations and superposed for evaluating the conformation and ligand binding. (A–C) Backbone representations of the superposition of apo TTR (green) and TTR in complex with (A) LUM (yellow), (B) IND (magenta) and (C) SUL (cyan). (D–F) Detailed views of one of the two HBS of the apo TTR (green) and in complex with (D) LUM (yellow), (E) IND (magenta) and (F) SUL (cyan).

Figure 10

Details of the interactions and positioning of LUM, IND and SUL into binding pockets. Backbone representations of HBS interactions with LUM (A,B), IND (C,D) and SUL (E,F). The residues involved in more important interactions are enumerated. Distances between side chains and ligand interactions are depicted with blue dotted lines, and distances are measured in angstroms.