Long QT Syndrome Type 2: Emerging Strategies for Correcting Class 2 KCNH2 (hERG) Mutations and Identifying New Patients

Abstract

Significant advances in our understanding of the molecular mechanisms that cause congenital long QT syndrome (LQTS) have been made. A wide variety of experimental approaches, including heterologous expression of mutant ion channel proteins and the use of inducible pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) from LQTS patients offer insights into etiology and new therapeutic strategies. This review briefly discusses the major molecular mechanisms underlying LQTS type 2 (LQT2), which is caused by loss-of-function (LOF) mutations in the KCNH2 gene (also known as the human ether-à-go-go-related gene or hERG). Almost half of suspected LQT2-causing mutations are missense mutations, and functional studies suggest that about 90% of these mutations disrupt the intracellular transport, or trafficking, of the KCNH2-encoded Kv11.1 channel protein to the cell surface membrane. In this review, we discuss emerging strategies that improve the trafficking and functional expression of trafficking-deficient LQT2 Kv11.1 channel proteins to the cell surface membrane and how new insights into the structure of the Kv11.1 channel protein will lead to computational approaches that identify which KCNH2 missense variants confer a high-risk for LQT2.

Keywords: KCNH2; hERG; ion channel; long QT syndrome; trafficking.

Figure 1

(A). This cartoon of an electrocardiogram recording shows normal sinus rhythm with a long QT interval that degenerates into Torsade de Pointes. (B). Loss-of-function mutations in KCNH2 that decrease the amplitude of I_Kr increase the ventricular action potential duration. The top panel shows a ventricular action potential with the different phases of the action potential labeled (0, 1, 2, 3, and 4) in control conditions (black traces) or after a loss in I_Kr (LQT2, red). The bottom panel shows the corresponding I_Kr. The dashed lines represent 0 mV or 0 pA. (C). The top pie chart shows the % of candidate LQT2 mutations that are nonsense (Class 1) or missense. To generate this graph, we defined candidate LQT2 missense mutations as ones that are reported to be “likely pathogenic” or “pathogenic” in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar). The bottom pie chart shows the relative percentage of LQT2 missense mutations that are expected to cause LQTS by a Class 2, 3, or 4 mechanism. These data are based on the functional studies of missense variants reported by Anderson and colleagues (2006 and 2014) [24,25]. A few mutations did not have an obvious dysfunctional phenotype (neutral) and might represent KCNH2 VUS misclassified as causing LQT2.

Figure 2

A CryoEM structure of Kv11.1a channel as viewed from the extracellular and intracellular side of the membrane (top left and right, each Kv11.1a α-subunit is denoted I-IV). A side-view of an individual Kv11.1a α-subunit highlighting the selectivity filter (G-F-G) and the secondary structures of the PAS, PORE, and CNB domain (middle and bottom). The location of LQT2 missense mutations that have been tested and disrupt Kv11.1a channel trafficking is shown in red. Green denotes the location of LQT2 missense mutations that have been tested and do not disrupt Kv11.1a trafficking.

Figure 3

A cartoon diagram of the Kv11.1 channel trafficking pathway. KCNH2 is transcribed and spliced to mRNA, which is then translated in the ER as a core glycosylated Kv11.1a α-subunit with a molecular mass of 135 kDa. It undergoes terminal glycosylation as it traffics through the Golgi apparatus and its molecular weight increases to 155 kDa. Kv11.1 channels recycle on and off the membrane in endosomes every few minutes for several hours before being degraded in the lysosome pathway. Representative Western blots of cells expressing WT-Kv11.1a α-subunit or a Class 2 LQT2 channel protein are shown below the cartoon. In control conditions, the immunoblot of cells expressing WT-Kv11.1a contains both core and terminally glycosylated Kv11.1a protein bands, whereas cells expressing the Class 2 LQT2 mutant channel protein show only the core glycosylated form. Incubating cells in I_Kr blockers (e.g., E-4031) increases terminal glycosylation and functional expression of Class 2 LQT2 channel proteins (pharmacological correction, dashed arrow).

Figure 4

(A). The American College of Medical Genetics and Genomics and the Association of Molecular Pathology provides a rubric that can be used to classify variants identified in Mendelian-linked disease alleles as pathogenic, likely pathogenic, uncertain significance, likely benign, and benign. (B). Shown is an example of how molecular dynamics (MD) simulations could be used to determine the structural impactof known and newly identified variants on Kv11.1a channel configurations. (C). MD simulations could be used to relate structural configurations to thermodynamic changes, like those controlling folding, that cause a LOF, gain of function (GOF) or neutral phenotype. Such predictions could be compiled for known variants into a descriptor matrix to train machine learning approaches for de novo prediction of the structural and functional impact of newly identified VUS in Kv11.1a channels. Over time, the sensitivity of these approaches might be fine-tuned to determine which VUS act as genetic modifiers.