Isogenic human pluripotent stem cell pairs reveal the role of a KCNH2 mutation in long-QT syndrome

Abstract

Patient-specific induced pluripotent stem cells (iPSCs) will assist research on genetic cardiac maladies if the disease phenotype is recapitulated in vitro. However, genetic background variations may confound disease traits, especially for disorders with incomplete penetrance, such as long-QT syndromes (LQTS). To study the LQT2-associated c.A2987T (N996I) KCNH2 mutation under genetically defined conditions, we derived iPSCs from a patient carrying this mutation and corrected it. Furthermore, we introduced the same point mutation in human embryonic stem cells (hESCs), generating two genetically distinct isogenic pairs of LQTS and control lines. Correction of the mutation normalized the current (IKr) conducted by the HERG channel and the action potential (AP) duration in iPSC-derived cardiomyocytes (CMs). Introduction of the same mutation reduced IKr and prolonged the AP duration in hESC-derived CMs. Further characterization of N996I-HERG pathogenesis revealed a trafficking defect. Our results demonstrated that the c.A2987T KCNH2 mutation is the primary cause of the LQTS phenotype. Precise genetic modification of pluripotent stem cells provided a physiologically and functionally relevant human cellular context to reveal the pathogenic mechanism underlying this specific disease phenotype.

Figure 1

Generation of hiPSCs from a patient with type-2 long-QT syndrome. (A) Genetic screening in the patient revealed the heterozygous single-nucleotide mutation A→T in exon 13 of the KCNH2 gene, in position 2987 of the coding sequence (CDS) (c.A2987T, NM_000238.3), resulting in the substitution of an asparagine with an isoleucine at position 996 of the protein (N996I, NP_000229.1). (B) The N996I mutation (red dot) is located in the C-terminal of the HERG protein, which is made of six trans-membrane domains (S1–S6), an amino (NH2) domain, a carboxyl (COOH) domain, and a pore (P) region. (C) Example of a hiPSC colony harbouring the c.A2987T (N996I) KCNH2 mutation (LQT2-hiPSCs^N996I). Scale bar: 400 μm. (D) Immunofluorescence analysis of pluripotency markers SSEA4 (green) and NANOG (red) in a representative LQT2-hiPSC^N996I clone, with nuclear staining (DNA, blue). The image on the right is a magnification of the area framed in the left image. Scale bars: 100 μm (left image); 50 μm (right image).

Figure 2

Gene targeting by homologous recombination in LQT2-hiPSCs^N996I and in NKX2.5^eGFP/w hESCs. (A) Schematic showing the project rationale. Patient-specific LQT2-hiPSCs harbouring the N996I mutation were corrected and NKX2.5^eGFP/w hESCs were mutated by gene targeting. Parental and genetically modified hPSC lines were differentiated into CMs and their electrophysiological phenotypes were analysed and compared. (B) The strategy for precise genomic modification of KCNH2. Top line, structure of the KCNH2 locus. Numbered black boxes indicate exons 6–15. Exon 13 is mutated (red T) in LQT2-hiPSCs^N996I and wild type (black A) in NKX2.5^eGFP/w hESCs. The gene targeting vector for correcting the mutation in LQT2-hiPSCs has the wild-type adenine nucleotide, whereas the gene targeting vector for introducing the mutation in NKX2.5^eGFP/w hESCs has the mutated thymine nucleotide. NeoR, the PGK-Neo cassette encoding G418 resistance flanked by loxP sequences (grey triangles), was inserted in the reverse direction. PCR primers (a, b) and (c, d) were used to identify the targeted clone. (C) PCR analysis using these primers generated specific bands of 4.8 kb (5′ homology arm) and 7.3 kb (3′ homology arm) from targeted clones (LQT2-hiPSCs^NeoR and NKX2.5^eGFP/w hESCs^NeoR). (D) Sequence analysis of PCR-amplified genomic DNA showing correction of the c.A2987T mutation in the LQT2-hiPSCs^corr line and mutation in the NKX2.5^eGFP/w hESCs^N996I line. The wild-type reference sequence (Ref) is shown in the top line. Source data for this figure is available on the online supplementary information page.

Figure 3

Differentiation of hPSCs into the cardiac lineage. (A) Confocal immunofluorescence images of cardiac sarcomeric proteins TNNI (green) and α-actinin (red) in human CMs generated from LQT2-hiPSCs^N996I, LQT2-hiPSCs^corr, NKX2.5^eGFP/w hESCs, and NKX2.5^eGFP/w hESCs^N996I. Nuclei are stained in blue. Bottom panels are a magnification of the area framed in the upper images. Top panels, scale bar: 25 μm; bottom panels, scale bar: 10 μm. (B) Transcriptional profile of human CMs generated from mutated and corrected hiPSCs (LQT2-hiPSCs^N996I and LQT2-hiPSCs^corr, respectively) and from wild-type and mutated hESCs (NKX2.5^eGFP/w hESCs and NKX2.5^eGFP/w hESCs^N996I, respectively). Undifferentiated cells from each hPSC line are also shown (Undiff.). Quantitative RT–PCR analysis was performed on the cardiac troponin gene (TNNT2), to show enrichment for the cardiomyocyte population, on the HERG channel gene (KCNH2), and on other key genes encoding for ion channels involved in the generation of the action potential in cardiac cells. All values are normalized to GAPDH and are relative to undifferentiated NKX2.5^eGFP/w hESCs. Raw minimum (min) and raw maximum (max) values were taken as a reference for heatmap representation.

Figure 4

I_Kr densities in mutated and corrected LQT2-hiPSC-derived CMs. (A) Representative current traces elicited upon 4 s depolarizing voltage steps to −30, −20, and −10 mV from a holding potential of −40 mV, before and after the application of 1 μM E-4031. Inset: voltage protocol. (B) Typical tail currents measured after the depolarizing step to −10 mV in corrected (black) and mutated (red) LQT2-hiPSC-CMs, showing a bi-exponential decay. (C) Average current–voltage (I–V) relationships for I_Kr, measured at the end of the test pulses, in mutated (red) and corrected (black) LQT2-hiPSC-derived CMs. Inset: voltage protocol. * indicates statistical significance (P=0.046, two-way rmANOVA; Holm–Sidak test post hoc analysis: −20 mV: P=0.015, −10 mV: P=0.007, 0 mV: P=0.018, 10 mV: P=0.011, 20 mV: P=0.013). (D) Average I–V relationships for peak tail currents in mutated (red) and corrected (black) LQT2-hiPSC-derived CMs. * indicates statistical significance (P=0.039, two-way rmANOVA; Holm–Sidak test post-hoc analysis: −20 mV: P=0.008, −10 mV: P=0.004, 0 mV: P=0.005, 10 mV: P=0.004, 20 mV: P=0.006). Inset: voltage protocol.

Figure 5

I_Kr densities in wild-type and mutated hESC-derived CMs. (A) Representative current traces elicited upon 4 s depolarizing voltage steps to −30, −20, and −10 mV from a holding potential of −40 mV, before and after the application of 1 μM E-4031. Inset: voltage protocol. (B) Typical tail currents measured after the depolarizing step to −10 mV in wild-type (black) and mutated (red) hESC-CMs, showing a bi-exponential decay. (C) Average I–V relationships for I_Kr, measured at the end of the test pulses, in wild-type (black) and mutated (red) hESC-derived CMs. Inset: voltage protocol. * indicates statistical significance (P=0.026, two-way rmANOVA; Holm–SIdak test post hoc analysis: −30 mV: P=0.004; −20 mV: P=0.000, −10 mV: P=0.002, 0 mV: P=0.009). (D) Average I–V relationships for peak tail currents in wild-type (black) and mutated (red) hESC-derived CMs. * indicates statistical significance (P=0.014, two-way rmANOVA; Holm–Sidak test post-hoc analysis: −30 mV: P=0.022, −20 mV: P=0.001, −10 mV: P=0.002, 0 mV: P=0.001, 10 mV: P=0.001, 20 mV: P=0.003). Inset: voltage protocol.

Figure 6

I_Kr activation and deactivation properties in mutated and corrected LQT2-hiPSC-derived CMs and in wild-type and mutated hESC-derived CMs. (A, B) Average peak tail current normalized to the maximal current following repolarization to −40 mV in mutated (red) and corrected (black) LQT2-hiPSC-derived CMs (A) and in wild-type (black) and mutated (red) hESC-derived CMs (B). Inset: voltage protocol. (C, D) Average time constants of I_Kr activation (τ) in mutated (red) and corrected (black) LQT2-hiPSC-derived CMs (C) and in wild-type (black) and mutated (red) hESC-derived CMs (D). (E, F) Average slow and fast time constants of I_Kr deactivation (τ_s and τ_f, respectively) in mutated (red) and corrected (black) LQT2-hiPSC-derived CMs (E) and in wild-type (black) and mutated (red) hESC-derived CMs (F).

Figure 7

AP characteristics in mutated and corrected LQT2-hiPSC-derived CMs and in wild-type and mutated hESC-derived CMs. (A, B) Representative examples of AP measured at 1 Hz (A) and average APD₅₀, APD₉₀, V_max, APA, and MDP (B) in mutated (red) and corrected (black) LQT2-hiPSC-derived CMs. * indicates statistical significance (APD₅₀: P=0.045, APD₉₀: P=0.026; t-test). (C, D) Representative examples of AP measured at 1 Hz (C) and average APD₅₀, APD₉₀, V_max, APA, and MDP (D) in wild-type (black) and mutated (red) hESC-derived CMs. * indicates statistical significance (APD₅₀: P=0.049, t-test).

Figure 8

Trafficking defect in CMs harbouring the c.A2987T (N996I) KCNH2 mutation. (A) lmmunofluorescence images of HERG channel (green) and actin (red) in representative human CMs derived from mutated and corrected LQT2-hiPSCs, and from wild-type and mutated hESCs. Nuclei are stained in blue. Bottom panels are a magnification of the area framed in the upper corresponding images. Top panels, scale bar: 25 μm; bottom panels, scale bar: 5 μm. (B) Flow cytometry purification of NKX2.5 eGFP⁺ and NKX2.5 eGFP⁻ hESC population from embryoid bodies differentiated from wild-type (left) and mutated (right) hESCs. The pictures show an individual eGFP-expressing (green) embryoid body; BF: bright field; scale bars: 400 μm. The representative dot plots show flow cytometric isolation of eGFP⁺ (green) and eGFP⁻ (orange) cell populations. (C) Representative western blot analysis of HERG protein in eGFP⁺ and eGFP⁻ cell populations purified from differentiated wild-type and mutated hESCs. Core- and complex-glycosylated HERG (135 and 155 kDa, respectively) are indicated. Actin is shown as a loading control. (D) Densitometric quantification of the 155-kDa and 135-kDa bands corresponding to the complex- and core-glycosylated HERG channel, respectively; ADU: arbitrary densitometric units; values are presented as mean±s.e.m., n=4. (E) Representative western blot analysis of HERG protein in LQT2-hiPSCs^corr- and LQT2-hiPSCs^N996I-derived CMs. Core- and complex-glycosylated HERG (135 and 155 kDa, respectively) are indicated. Actin is shown as a loading control. (F) Densitometric quantification of the 155-kDa and 135-kDa bands corresponding to the complex- and core-glycosylated HERG channel, respectively; ADU: arbitrary densitometric units; values are presented as mean±s.e.m., n=2. Source data for this figure is available on the online supplementary information page.

Figure 9

UPR pathway analysis in wild-type and mutated hESC-CMs. (A–C) Western blot analysis of ATF6 (A), Calnexin (B), and Calreticulin (C) in eGFP⁺ and eGFP⁻ cell populations purified from differentiated wild-type and mutated hESCs. Actin is shown as a loading control. (D, E) Western blot analysis of HERG channel, calnexin, and calreticulin (D) and quantification of trafficking efficiency (E) under basal conditions (CTR) and upon proteasome (LACT) or lysosome (LEUP) inhibition in eGFP⁺ cell populations purified from differentiated wild-type and mutated hESCs. Actin is shown as a loading control. Trafficking efficiency=fg/(fg+cg), where fg=fully-glycosylated 155 kDa band and cg=core-glycosylated 135 kDa band. Source data for this figure is available on the online supplementary information page.