A human Dravet syndrome model from patient induced pluripotent stem cells

Abstract

Background: Dravet syndrome is a devastating infantile-onset epilepsy syndrome with cognitive deficits and autistic traits caused by genetic alterations in SCN1A gene encoding the α-subunit of the voltage-gated sodium channel Na(v)1.1. Disease modeling using patient-derived induced pluripotent stem cells (iPSCs) can be a powerful tool to reproduce this syndrome's human pathology. However, no such effort has been reported to date. We here report a cellular model for DS that utilizes patient-derived iPSCs.

Results: We generated iPSCs from a Dravet syndrome patient with a c.4933C>T substitution in SCN1A, which is predicted to result in truncation in the fourth homologous domain of the protein (p.R1645*). Neurons derived from these iPSCs were primarily GABAergic (>50%), although glutamatergic neurons were observed as a minor population (<1%). Current-clamp analyses revealed significant impairment in action potential generation when strong depolarizing currents were injected.

Conclusions: Our results indicate a functional decline in Dravet neurons, especially in the GABAergic subtype, which supports previous findings in murine disease models, where loss-of-function in GABAergic inhibition appears to be a main driver in epileptogenesis. Our data indicate that patient-derived iPSCs may serve as a new and powerful research platform for genetic disorders, including the epilepsies.

Figure 1

Characterization of generated iPSCs and neuronal differentiation. (A) SCN1A sequencing of the indicated cell material. Solid arrowheads point to the c.4933C>T substitution. (B) iPSC morphology and immunostaining of pluripotency markers (Oct 4, Nanog, Tra-1-60, Tra-1-81, and SSEA4 without SSEA1). Scale bar, 500 μm. (C) iPSC-derived teratomas generated in NOD-SCID mouse testes comprised tissues from all three germ layers. Scale bar, 200 μm (neural rosettes and respiratory epithelium) and 400 μm (others). (D) Real-time PCR analysis showed suppressed expression of the four reprogramming factors in both patient and control iPSCs compared to patient fibroblasts transduced with the same four factors (D1-HDF-4F). (E) G-band karyotyping showed normal chromosome numbers (46,XX) in all tested colonies (N = 20 each). (F) Representative images of embryoid bodies (EB) and neurospheres (NS). Scale bar, 500 μm. (G) Expression of βIII-tubulin, a neuronal marker (green) and GFAP, an astrocyte marker (red) in iPSC-derived neural cells. Scale bar, 200 μm. Day numbers indicate the days of differentiation in adherent culture after neurosphere formation.

Figure 2

Na_v channel expression in iPSC-derived neurons. (A) Real-time PCR addressing neuronal Na_v expression at 30 days of differentiation (N = 3 in each cell line) Crossing point differences to β-actin (ΔCp = Cp^β-actin - Cp^Nav) closer to zero indicate higher expression. PCR efficiencies were nearly identical (Additional file 2). Asterisks indicate a significant difference to SCN1A (P < 0.5, one-way ANOVA). Expression strength of the indicated Na_v genes was constant across the cell lines (P = 0.92, two-way ANOVA) (B) Normalized expression levels for each Na_v gene (SCN1A + SCN2A + SCN3A + SCN8A)/4 = 1. Compared to the control, SCN1A expression tended to be higher in D1-1 (P = 0.0929, one-way ANOVA), and it was significantly higher in D1-6 (^*P = 0.0078). The distribution of Na_v genes expression ratios in each cell line was significantly different between the control and the patient lines (P =0.0086 and <0.0001 for D1-1 and D1-6, respectively, two-way ANOVA), but identical between D1-1 and D1-6 (P = 0.11). (C) Sequencing of SCN1A reverse transcribed mRNA isolated from iPSCs-derived neurons. Patient-neurons show a double peak at mutation site (solid arrowheads), confirming the heterozygous state of the cells (D) Immunocytochemical characterization of Na_v1.1 expression in control neurons: strong (solid arrows), moderate (open arrows), weak (solid arrowheads), and faint (open arrowhead). Despite weak staining in the cell body, neurite staining was often apparent (solid arrowheads). (E) Neurite co-localization of Na_v1.1 and the AIS marker ankyrin G (AnkG, solid arrowheads). (F) PAN-Na_v staining of SCN1A Venus-positive neurons (via anti-GFP, see Figure 3) in the AIS (arrowheads). (G) Co-localization of Na_v1.1 and GAD67 staining. (H) VGlut1-positive neuron with SCN1A Venus expression. Scale bars: 100 μm (D), 30 μm (F), 200 μm (G) and 50 μm (others).

Figure 3

Structure and characterization of the lentiviral SCN1A-reporter used in the electrophysiological analyses. (A) The reporter comprised (5^′ to 3^′) a 1.2-kb upstream sequence, a 5^′-untranslated exon, the 5^′-end of the first coding exon, and, following the ATG start codon, Venus cDNA. (B) & (C) 201B7 neurons labeled for Venus (using a GFP antibody) and Na_v1.1 (B) or GABA (C). (B) Open arrowheads indicate GFP-pseudopositive neurons lacking Na_v1.1 staining. (C) GFP-positive neurons with (arrowhead) and without GABA staining (open arrowhead). Scale bars: 200 μm.

Figure 4

Electrophysiological characteristics of mature iPSC-derived neurons. (A) Capacitance, resting membrane potential (RMP), action potential (AP) firing threshold, and voltage peak were identical across all neurons analyzed (P >0.05, Kruskal-Wallis test); error bars indicate S.E.M. (B) Representative traces of AP trains triggered by a 500-ms depolarizing current at the indicated intensities. Transverse dotted lines demark 0 mV membrane potential. Scale bars: 20 mV vs. 100 ms. (C) Action potential (AP) decrement at current intensities triggering >10 APs calculated as a percentage: 10^th/1^st AP amplitude. Control vs. D1-1 (P = 0.078) and D1-6 (^*P = 0.045, ANOVA); D1-1 vs. D1-6 (P = 0.839) (D) Total number of APs during the 500-ms stimulation period vs. current injection intensity. When exposed to strong current injections, both patient-derived cell lines produced significantly fewer APs compared to the control (the slope of AP numbers at ≥50 pA, P = 0.0102 and 0.0011 for D1-1 and D1-6, respectively, ANCOVA, ^*P <0.05 for D1-6 only, ^**P <0.05 for both D1-1 and D1-6, Wilcoxon rank-sum test).