The difficulty to model Huntington's disease in vitro using striatal medium spiny neurons differentiated from human induced pluripotent stem cells

Abstract

Huntington's disease (HD) is an autosomal dominant neurodegenerative disorder caused by an expanded polyglutamine repeat in the huntingtin gene. The neuropathology of HD is characterized by the decline of a specific neuronal population within the brain, the striatal medium spiny neurons (MSNs). The origins of this extreme vulnerability remain unknown. Human induced pluripotent stem cell (hiPS cell)-derived MSNs represent a powerful tool to study this genetic disease. However, the differentiation protocols published so far show a high heterogeneity of neuronal populations in vitro. Here, we compared two previously published protocols to obtain hiPS cell-derived striatal neurons from both healthy donors and HD patients. Patch-clamp experiments, immunostaining and RT-qPCR were performed to characterize the neurons in culture. While the neurons were mature enough to fire action potentials, a majority failed to express markers typical for MSNs. Voltage-clamp experiments on voltage-gated sodium (Nav) channels revealed a large variability between the two differentiation protocols. Action potential analysis did not reveal changes induced by the HD mutation. This study attempts to demonstrate the current challenges in reproducing data of previously published differentiation protocols and in generating hiPS cell-derived striatal MSNs to model a genetic neurodegenerative disorder in vitro.

Figure 1

Schematic representation of the Stanslowsky protocol. (A) Timeline of the differentiation protocol. DM, SB and IWP2 induce the neural pathway driving the formation of embryoid bodies. IWP2 associated with PMA induces the regional patterning towards the subpallium that comprises large parts of the basal ganglia including striatum and globus pallidus. Y27632 inhibits apoptosis to prevent cell decline at the beginning of the differentiation. The neuronal growth factors BDNF, GDNF and TGF-β3 terminate the differentiation into MSNs and drive their maturation, while dbcAMP increases the intracellular Ca²⁺ levels and contributes to the maturation process. (B) hiPS cells grow in basal medium until DIV0. DIV0 to DIV12-14 initiates the formation of embryoid bodies. (C) From DIV12-14 to DIV55 ± 3, the neurons are exposed to growth factors to induce neurite outgrowth (D) and to establish neuronal networks (E). Scale bar: 1,000 µm (B, C, E) and 200 µm (D). The dotted lines delineate the different developmental stages during differentiation of MSNs from hiPS cells.

Figure 2

Schematic representation of the Fjodorova protocol. (A) Timeline of the differentiation protocol. DM, LDN and SB favor neural induction by inhibiting the dual SMAD signaling. Activin A induces a regional patterning towards the lateral ganglionic eminence that originates from the subpallium. The neuronal growth factors BDNF and GDNF terminate the differentiation into MSNs and mature the neurons. (B) hiPS cells grow in basal medium until DIV0. (C) Between DIV0 and DIV9, hiPS cells grow into neuroectodermal cells. (D) From DIV9 to DIV20, Activin A application induces the formation of lateral ganglionic eminence (LGE) progenitors. (E) Between DIV20 and DIV35 ± 3, the LGE progenitors mature under the influence of neuronal growth factors. Scale bar: 1000 µm (B, C, D) and 200 µm (E). The dotted lines delineate the different developmental stages during differentiation of MSNs from hiPS cells.

Figure 3

The percentage of active neurons and neurons firing more than 2 APs is increased following a longer time in culture for the Stanslowsky protocol. (A) The percentage of active control and HD neurons firing at least one AP at DIV40 ± 2 (n = 19 Ctrl1, 18 Ctrl2, 6 Ctrl3 and 20 HD72 neurons) and the same neurons with HD109 at DIV55 ± 3 (n = 60 Ctrl1, 51 Ctrl2, 28 Ctrl3, 38 HD72 and 37 HD109 neurons) for the Stanslowsky protocol. (B) The percentage of active control neurons firing at least 1 AP at DIV35 for the Fjodorova protocol (n = 34 Ctrl1, 73 Ctrl2, 49 HD72 and 20 HD109 neurons). (C) The percentage of phasic firing neurons (generating only 1 AP) and of tonic firing neurons (generating two and more APs) at DIV40 and DIV55 for the Stanslowsky protocol. (D) The percentage of phasic firing and of tonic firing neurons at DIV35 for the Fjodorova protocol. There was no extended time in culture for the hiPS cell-derived neurons of this protocol. See Supplementary Table S5 for details.

Figure 4

Both differentiation protocols produce only a small percentage of MSNs. (A,B) Representative immunostaining of Ctrl2 (bottom) and HD72 neurons (top) from the Stanslowsky protocol (A) and the Fjodorova protocol (B). DAPI staining is represented in blue, Tuj1 and DARPP32 in red and GAD67 in green. Scale: 50 µm. (C) Ctrl2 neurons of the Fjodorova protocol stained with Tuj1 in green and TH in red (top) or with GAD67 in green and TH in red (bottom). (D) Percentage of Tuj1⁺ neurons in each culture compared to the number of DAPI-stained nuclei. (E) Percentage of GAD67⁺ neurons (GABAergic neurons) in each culture compared to the total number of DAPI-stained nuclei. (F) Percentage of DARPP32⁺ neurons (striatal MSNs) in each culture compared to the number of GAD67⁺ neurons. See Supplementary Table S6 for quantitative data. Data show mean (1 technical replicate).

Figure 5

RT-qPCR of the nine Nav α and the four Nav β subunits from each iPSC-derived neuron lines of the Stanslowsky (blue) or the Fjodorova (purple) protocol. (A-B) RT-qPCR of Ctrl1 (A) and Ctrl2 (B) neurons. (C-E) RT-qPCR of HD72 (C), HD109 (D) and HD180 (E) neurons. Error bars denote 95% confidence interval. Number of differentiations for the Stanslowsky protocol: n = 2 for Ctrl1, n = 2 for Ctrl2, n = 3 or HD72, n = 2 for HD109, n = 2 for HD180 and for the Fjodorova protocol: n = 1 for Ctrl1, n = 3 for Ctrl2, n = 1 or HD72, n = 2 for HD109, n = 2 for HD180.

Figure 6

The two protocols reveal large differences in the gating properties of Nav channels expressed in the differentiated neurons. (A) Pre-pulse protocol (top) designed to accurately record voltage dependence of activation in mature hiPS cell-derived neurons without space clamp artefacts. Pre-pulse (purple) and inter-pulse (green) are adjusted to each neuron individually. Example trace (bottom) is from Ctrl2 neurons of the Fjodorova protocol. (B) Representative recording of HD72 hiPS cell-derived central neurons of the Stanslowsky protocol before 500 nM TTX application (top) and following 500 nM TTX application (bottom). Scale bars are the same as in (A). (C) Voltage dependence of activation and steady-state fast inactivation of control and HD neurons of the Stanslowsky protocol (blue) and of the Fjodorova protocol (purple) (see Table S8 for results of each individual cell line). (D-E) Values for half-maximal (V_half) voltage-dependent activation (D) and steady-state fast inactivation (E) of Ctrl1 (filled circles, n = 24–26), Ctrl2 (filled squares, n = 13–23), HD72 (open circles, n = 19–30) and HD109 neurons (open squares, n = 20–22) for the Stanslowsky protocol and Ctrl1 (filled circles, n = 12–16), Ctrl2 (filled squares, n = 21–27), HD72 (open circles, n = 12–30) and HD109 neurons (open squares, n = 20–26) for the Fjodorova protocol. Data show mean ± 95% CI. (F–H) Plot displays the difference between V_half values of voltage dependence of activation (D) and steady-state fast inactivation (E) obtained from the two differentiation protocols when control/HD neurons of each protocol are pooled. A value of 1 (dotted line) represents equal V_half values in both protocols. Values < 1 mean hyperpolarized values in the Fjodorova protocol as compared to the Stanslowsky protocol. While both protocols provide comparable values for voltage dependence of activation (F), neurons from the Fjodorova protocol consistently display a hyperpolarized fast inactivation (G). (H) Plot displays the difference between V_half values of voltage dependence of activation and steady-state fast inactivation obtained from the two genotypes for each protocol. The genotype does not seem to influence the Nav channel gating of hiPS cell-derived neurons.

Figure 7

hiPS cell-derived neurons display distinct AP features within the two protocols. (A) Representative AP traces (black) of Ctrl2 neurons from Stanslowsky (left) and from Fjodorova (right) protocols. The grey trace represents the subthreshold current injection step. AP threshold for each individual neuron is indicated with an arrow. The red trace indicates the current injection step. Dotted black lines represent 0 mV. (B) Resting membrane potential (RMP). (C,E) AP features of control/HD hiPS cell-derived neurons from the Stanslowsky protocol (blue, n = 27 Ctrl1, 46 Ctrl2, 30 HD72 and 37 HD109 neurons) and the Fjodorova protocol (purple, n = 29 Ctrl1, 65 Ctrl2, 57 HD72 and 18 HD109 neurons): AP threshold (C), AP amplitude (D) and AP time-to-peak (E). Two dotted lines indicate the cut-off of neurons with a time-to-peak longer than 200 or 300 ms. (F) Maximal number of APs fired by control and HD neurons. Data show mean ± 95% CI. See Supplementary Table S9 for statistical results.

Figure 8

Changing internal Ca²⁺ does not affect steady-state fast inactivation or AP properties of hiPS cell-derived neurons from the Stanslowsky protocol. (A, B) Steady-state fast inactivation with and without 500 nM [Ca²⁺]_i in Ctrl1 (A) or in HD72 hiPS cell-induced neurons of the Stanslowsky protocol (B) (n = 8–10 neurons). (C, F) Effect of 500 nM [Ca²⁺]_i on AP characteristics in hiPS cell-derived neurons of Ctrl3 and HD72 of the Stanslowsky protocol (n = 9–22 neurons). (C) Resting Membrane Potential, (D) AP threshold, (E) AP amplitude, (F) AP time-to-peak. All data shown as mean ± 95% CI. See Supplementary Table S12 for values.