Structure-activity relationships of pentamidine-affected ion channel trafficking and dofetilide mediated rescue

Abstract

Background and purpose: Drug interference with normal hERG protein trafficking substantially reduces the channel density in the plasma membrane and thereby poses an arrhythmic threat. The chemical substructures important for hERG trafficking inhibition were investigated using pentamidine as a model drug. Furthermore, the relationship between acute ion channel block and correction of trafficking by dofetilide was studied.

Experimental approach: hERG and K(IR)2.1 trafficking in HEK293 cells was evaluated by Western blot and immunofluorescence microscopy after treatment with pentamidine and six pentamidine analogues, and correction with dofetilide and four dofetilide analogues that displayed different abilities to inhibit IKr . Molecular dynamics simulations were used to address mode, number and type of interactions between hERG and dofetilide analogues.

Key results: Structural modifications of pentamidine differentially affected plasma membrane levels of hERG and K(IR)2.1. Modification of the phenyl ring or substituents directly attached to it had the largest effect, affirming the importance of these chemical residues in ion channel binding. PA-4 had the mildest effects on both ion channels. Dofetilide corrected pentamidine-induced hERG, but not K(IR)2.1 trafficking defects. Dofetilide analogues that displayed high channel affinity, mediated by pi-pi stacks and hydrophobic interactions, also restored hERG protein levels, whereas analogues with low affinity were ineffective.

Conclusions and implications: Drug-induced trafficking defects can be minimized if certain chemical features are avoided or 'synthesized out'; this could influence the design and development of future drugs. Further analysis of such features in hERG trafficking correctors may facilitate the design of a non-blocking corrector for trafficking defective hERG proteins in both congenital and acquired LQTS.

Figure 1

Electrophysiological effects of 10 mg·kg⁻¹ pentamidine over 60 min in AV-blocked dogs paced from the high septum. A delayed effect of pentamidine on the QT interval was observed. QT prolongation became apparent 4 days after i.v. pentamidine administration that persisted for 17 days, after which the prolongation dissipated gradually. Data are presented as mean ± SD (n = 3).

Figure 2

Pentamidine and its analogues differentially affect hERG and K_IR2.1 protein levels. (A) Western blot showing the effects of pentamidine or its structural analogues (10 μM, 48 h) on hERG forward trafficking (n = 12 for control and pentamidine, n = 3 for PA-4, n = 4 for all other analogues). (B) Western blot showing the effects of pentamidine or its structural analogues (10 μM, 48 h) on K_IR2.1 protein expression (n = 3 for PA-4 and 7, n = 4 for control, pentamidine, and all other analogues). Total protein staining (Ponceau) was used as a loading control. Control protein levels (untreated cells) were designated as 100% after correction; *indicates P < 0.05 versus control, ‡indicates P < 0.05 versus pentamidine treatment.

Figure 3

Dofetilide corrects pentamidine-induced hERG trafficking defects. (A) Western blot showing the concentration-dependent correction of hERG protein levels in the presence of pentamidine. Dofetilide 1 μM completely restored mature hERG protein levels, higher concentrations increased mature hERG protein levels even further (n = 24 for control and pentamidine, n = 6 for all dofetilide concentrations). (B) Dofetilide time-dependently corrects mature hERG protein levels. Treatment of pentamidine-exposed (10 μM, 48 h) HEK-hERG cells with 1 μM dofetilide restored mature hERG levels within 4–6 h (n = 24 for control and pentamidine, n = 3 dofetilide). (C) In contrast to the beneficial effect on hERG trafficking, pentamidine-mediated downregulation of K_IR2.1 could not be corrected by application of dofetilide (n = 3 for all conditions). Total protein staining (Ponceau) was used as a loading control. Control protein levels (untreated cells) were designated as 100% after correction; *indicates P < 0.05 versus control, ‡indicates P < 0.05 versus pentamidine treatment.

Figure 4

Correction of pentamidine-induced hERG trafficking defects is influenced by structural modifications of dofetilide (A). Compared with dofetilide, the high affinity analogue DA-1 was an equally effective corrector, while the other analogues were significantly less effective when tested at 1 μM (n = 24 for control and pentamidine, n = 6 for dofetilide, n = 3 for DA-1 and 2, n = 5 for DA-3 and 4). Total protein staining (Ponceau) was used as a loading control. Control protein levels (untreated cells) were designated as 100% after correction; *indicates P < 0.05 versus control, ‡indicates P < 0.05 versus pentamidine treatment. (B) In control conditions, the hERG protein is present in the cytoplasm and at the plasma membrane. A similar staining pattern was seen after only dofetilide treatment. After pentamidine treatment, only cytoplasmic staining was apparent. Staining patterns after correction of these effects with dofetilide or its structural analogues were in agreement with the Western blot results; effective correction showed hERG staining both at the plasma membrane and intracellularly, while ineffective correctors showed only intracellular staining. The hERG protein is shown in green.

Figure 5

MD poses and 2D interaction profiles of dofetilide (A,B), DA-1 (C,D), DA-2 (E,F), DA-3 (G,H) and DA-4 (I,J). Snapshots after 50 ns are shown. Binding determinants Y652 and F656 are shown as raspberry sticks. Drug molecules are coloured orange, with nitrogen atoms coloured blue, oxygen atoms coloured red and sulfur groups coloured yellow. 2D interaction profiles were generated with PoseView (poseview.zbh.uni-hamburg.de). Green dotted lines indicate pi-pi interactions between aromatic rings, green solid lines represent hydrophobic interactions. Hydrogen bonds are shown as black dotted lines.