Increased Calcium Influx through L-type Calcium Channels in Human and Mouse Neural Progenitors Lacking Fragile X Mental Retardation Protein

Summary

The absence of FMR1 protein (FMRP) causes fragile X syndrome (FXS) and disturbed FMRP function is implicated in several forms of human psychopathology. We show that intracellular calcium responses to depolarization are augmented in neural progenitors derived from human induced pluripotent stem cells and mouse brain with FXS. Increased calcium influx via nifedipine-sensitive voltage-gated calcium (Ca_v) channels contributes to the exaggerated responses to depolarization and type 1 metabotropic glutamate receptor activation. The ratio of L-type/T-type Ca_v channel expression is increased in FXS progenitors and correlates with enhanced progenitor differentiation to glutamate-responsive cells. Genetic reduction of brain-derived neurotrophic factor in FXS mouse progenitors diminishes the expression of Ca_v channels and activity-dependent responses, which are associated with increased phosphorylation of the phospholipase C-γ1 site within TrkB receptors and changes of differentiating progenitor subpopulations. Our results show developmental effects of increased calcium influx via L-type Ca_v channels in FXS neural progenitors.

Highlights

• •Responses to activity are augmented in neural progenitors in fragile X syndrome (FXS).
• •Increased Ca²⁺ influx contributes to the exaggerated FXS progenitor responses
• •L-type voltage-gated channels are abnormally activated in FXS progenitors
• •Reduced BDNF diminishes Ca²⁺ influx and modulates FXS progenitor differentiation

In this article, Maija Castrén and colleagues show contribution of increased Ca²⁺ influx through L-type voltage-gated calcium channels to augmented responses to depolarization and glutamate receptor activation in neural progenitors derived from human induced pluripotent stem cells (iPSCs) and mouse brain modeling fragile X syndrome.

Introduction

Monogenic disorders such as fragile X syndrome (FXS) are useful for studies that investigate defective molecular pathways in intellectual disability syndromes. FXS results from the absence of fragile X mental retardation protein (FMRP), which is caused by a CGG triplet repeat expansion leading to transcriptional silencing of the FMR1 gene (Verkerk et al., 1991). FMRP is an RNA-binding protein that is important for translational regulation of many brain mRNAs (Darnell et al., 2011). FMRP is necessary for normal brain development and formation of functional neuronal connections (Contractor et al., 2015). The behavioral phenotype of FXS includes cognitive impairment, defective communication, abnormal sensory reactivity, anxiety, hyperactivity, gaze aversion, and impulsivity (Lozano et al., 2014). The main phenotype and molecular findings in human FXS are recapitulated in Fmr1 knockout (KO) mice (Kooy et al., 1996). Altered neuronal differentiation and migration disrupt formation of cortical layers in the Fmr1 KO mouse brain and affect developmental processes involved in neuronal circuit formation and function (Gonçalves et al., 2013, La Fata et al., 2014, Tervonen et al., 2009).

Ca²⁺-mediated signals regulate many processes during neuronal development, including progenitor proliferation, neuronal migration and differentiation, axon guidance, and dendrite growth (Rosenberg and Spitzer, 2011, Zheng and Poo, 2007). Several forms of Ca²⁺ activity take place during cortical development and are mediated by metabotropic glutamate receptors (mGluRs), gap junctions, GABAa receptors, and ionotropic glutamate receptors (iGluRs). Voltage-gated calcium (Ca_v) channels are the major source of Ca²⁺ influx in electrically excitable cells and have a great impact on cell signaling (Rosenberg and Spitzer, 2011). Three subfamilies of Ca_v channels have been identified (Ca_v1-3); L-type calcium (Ca_v1) channels represent one of the major classes of Ca_v channels (Dolphin, 2016). Ca_v1 channels are subdivided into four isoforms (Ca_v1.1, Ca_v1.2, Ca_v1.3, and Ca_v1.4) based on the pore-forming α1 subunit that selectively conducts calcium ions. Ca_v1.2 and Ca_v1.3 (encoded by the Cacna1c and Cacna1d genes, respectively) are the predominant subunits of the L-type Ca_v channels in the brain and are expressed in neural progenitor cells (NPCs) (Louhivuori et al., 2013). Treatment with Ca_v1 blockers inhibits functional maturation of neurons (D'Ascenzo et al., 2006) and reduces dendritic outgrowth and synapse formation in murine NPC cultures (Lepski et al., 2013).

FXS neurons differentiated in vitro from human pluripotent stem cells, and cultured from Fmr1 KO mouse brain show abnormal maturation and functional deficits (Boland et al., 2017, Braun and Segal, 2000, Telias et al., 2015). In both human and mouse FXS NPCs, aberrant functional responses are detectable already at the very early stages of neuronal differentiation (Achuta et al., 2017, Achuta et al., 2018), and augmented activity-dependent intracellular calcium responses are consistent with increased neuronal excitability in FXS (Achuta et al., 2014, Louhivuori et al., 2011). Here we studied the contribution of L-type Ca_v channels to the differentiation defects of NPCs derived from human induced pluripotent stem cells (iPSCs) and mouse brain with FXS. Our results demonstrate a central role for increased Ca²⁺ influx through L-type Ca_v channels in augmented responses to activity, altered fate determination of neuronal cells, and migration defect of FXS NPCs.

Results

Augmented Intracellular Calcium Responses via L-type Ca_v Channels in Mouse FMRP-Deficient NPCs

Depolarization with high extracellular potassium ([K⁺]_e) results in augmented intracellular calcium ([Ca²⁺]_i) responses in differentiating NPCs generated from Fmr1 KO mice compared with wild-type (WT) controls (Achuta et al., 2014). Differentiating NPCs express functional Ca_v channels that mediate Ca²⁺ influx in response to membrane depolarization (D'Ascenzo et al., 2006). We studied the contribution of L-type Ca_v channels to the increased Ca²⁺ entry to mouse cortical progenitors lacking FMRP by investigating the effects of nifedipine (a blocker of L-type Ca_v channels) on [Ca²⁺]_i responses in Fmr1 KO neurospheres. We induced NPC differentiation by withdrawal of mitogens and monitored [Ca²⁺]_i responses in all or most differentiated cells migrated out from the neurosphere (Figures 1A and 1B). The migration area was divided into three zones to analyze the functional responses in differentially migrated cells (Figure 1B). Nifedipine blocked the [Ca²⁺]_i responses elicited by depolarization with 75 mM [K⁺]_e in FMRP-deficient NPCs to the levels of nifedipine-treated WT cells in the whole migration area (Figure 1C), indicating that Ca²⁺ influx into FMRP-deficient cells was abnormally increased through Ca_v1 channels.

Figure 1.

Intracellular Calcium Responses via L-type Voltage-Gated Calcium Channels and Expression of Channel Subunits in Mouse NPCs Lacking FMRP

(A) Representative images of WT and Fmr1 KO mouse neurospheres immunostained for GLAST (green) and MAP2 (red) at day 1 of differentiation. Scale bars, 100 μm.

(B) A bright-field image of a neurosphere and a Fura-2 fluorescence image showing the migration area that was categorized into three zones (zones 1–3). Scale bars, 100 μm.

(C) Average amplitude of [Ca²⁺]_i response in differentiated cells of WT and Fmr1 KO mouse neurospheres after stimulation with high (75 mM) [K⁺]_e and effects of nifedipine (nife) on [K⁺]_e-evoked responses. WT(n) = 8, Fmr1 KO(n) = 8, WT + nife(n) = 4, and Fmr1 KO + nife(n) = 4 neurospheres; responses of 75–100 cells were measured in each experiment.

(D and E) The relative mRNA expression of Cacna1c, Cacna1d, Cacna1g, Cacna1h, and Cacna1a in WT and Fmr1 KO neurospheres (D) and the L-type Ca_v channel/T-type Ca_v channel expression ratio in NPCs at day 1 of differentiation (E). Each group(n) = 4–5 in twice repeated experiment.

(F) Immunofluorescent images showing L-type Ca_v channel α_1c subunit expression (green) in WT and Fmr1 KO NPCs differentiated for 1 day.

Data are expressed as means ± SEM; ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05.

Ca_v channels are heteromultimers composed of a pore-forming α1 subunit and auxiliary β and α₂δ subunits (Dolphin, 2016). All Ca_v channel α1 subunits were expressed at low levels and showed high variation in NPCs grown in neurospheres. Detectable amounts of mRNAs allowed analysis of the expression of Cacna1c and Cacna1d for L-type-specific α1 subunits, Cacna1g and Cacna1h for T-type α subunits, and Cacna1a for P/Q-type α subunits. Although Cacna1c and Cacna1d expression showed a tendency to be increased in FMRP-deficient mouse NPCs compared with controls, we found no significant differences of any α1 subunit expression between WT and Fmr1 KO NPCs (Figure 1D). However, the ratio of the L-type/T-type α1 subunit mRNA expression was 5.1-fold higher in undifferentiated Fmr1 KO neurospheres and 2.3-fold higher in FMRP-deficient NPCs compared with WT controls at day 1 of differentiation (Figure 1E). As shown in Figure 1F, the protein expression of the L-type channel α1 subunits did not differ between Fmr1 KO and WT NPCs at the single-cell level, indicating changes in the differentiated subpopulations of cells and enhanced differentiation of NPCs expressing L-type channels in the absence of FMRP.

Augmented Single-Cell Responses to Membrane Depolarization in Human FXS iPSC-Derived NPCs

As in mouse FXS NPCs, responses to membrane depolarization by increased [K⁺]_e were higher (2.1-fold) in human FXS iPSC-derived NPCs than in controls (Figure 2A). We found that the relative CACNA1G expression was reduced (Figure 2B), while the CACNA1C expression was not affected (data not shown) in FXS NPCs at day 1 of differentiation relative to that of undifferentiated NPCs in the transcriptome analysis of NPCs derived from three control and three FXS iPSC lines. The expression of the Ca_v channel α1 subunits showed very high variability between cell lines and we found no significant differences in the expression of the α1 subunits between FXS and control NPCs (Figure 2C). The results of both human and mouse FXS NPCs suggested that the expression of T-type Ca_v channels was switched to L-type channels and correlated with enhanced NPC differentiation to glutamate-responsive cells in FXS NPCs.

Figure 2.

Augmented Intracellular Ca²⁺ Responses to Depolarization and the Expression of Ca_v Channels and a Splicing Defect of the CAST Gene in Human FXS NPCs

(A) Representative single-cell responses to high [K⁺]_e and average amplitude of [Ca²⁺]_i responses in human FXS and control (CTRL) NPCs. Data are from n(CTRL) = 5 and n(FXS) = 6 neurospheres. HEL46.11, HEL11.4, HEL100.1, HEL100.2, and HEL70.3 cell lines; responses of 50–100 cells were measured in each neurosphere.

(B) Relative CACNA1G expression in NPCs at day 1 of differentiation compared with that in undifferentiated FXS and control neurospheres (NS).

(C) mRNA expression of the pore-forming α1 subunits of Ca_v channels in human FXS and control NPCs. n(CTRL) = 3 cell lines (HEL11.4, HEL23.3, and HEL46.11) and n(FXS) = 3 cell lines (HEL69.5, HEL70.3, and HEL10.2).

(D) Schematic presentation of the alternative splicing of the CAST gene in FXS NPCs. Splicing index of the exon and junction probe set is shown; splicing index 18.35 correlates with a 21.06-fold change; event pointer p < 0.05.

Data are expressed as means ± SEM; ^∗∗∗p < 0.001, ^∗p < 0.05 by Student's t test.

We analyzed the transcriptome data of human iPSC-derived NPCs to identify Ca²⁺regulated pathways that associate with the augmented Ca_v1 responses in FXS NPCs. In this respect, only one significant change was identified; in the splicing-specific transcriptome analysis the CAST gene expression was altered in FXS NPCs. The CAST gene encodes calpastatin, which can regulate L-type Ca_v function directly or as a natural inhibitor of calpains. We found an alternative splicing event leading to exon skipping in the CAST gene in FXS NPCs (Figure 2D). The probe set, including both exon and exon junction probes, detected an alternative splicing event (p = 0.026) in chromosome 5 at the genome location 96740745–96740783 (ENSE00003628178). Exon 19 encoding 39 base pairs of the CAST gene was removed in FXS NPCs. This exon encodes part of the calpastatin protein and its differential splicing has not previously been reported.

Genetic Reduction of BDNF Expression Prevents Augmentation of Activity-Dependent Responses in FMRP-Deficient NPCs

Ca_v channel-mediated elevation of Ca²⁺ levels triggers transcription of the Bdnf gene (Zheng et al., 2011) and the increased activation of Ca_v1 channels in FMRP-deficient NPCs correlates with the upregulated expression of brain-derived neurotrophic factor (BDNF) mRNAs shown previously in undifferentiated FXS mouse NPCs (Uutela et al., 2014). BDNF mRNA expression is also increased and regulated in an activity-dependent manner in mouse FMRP-deficient neurons in vitro and in vivo, but in Fmr1 KO neurospheres, reduced differentiation of cells expressing BDNF modulates total BDNF mRNA and protein expression (Louhivuori et al., 2011). In human iPSC-derived NPCs, BDNF mRNA expression showed high variation and did not differ significantly between FXS and control NPCs at the early stage of differentiation (Figure 3A). To explore effects of BDNF on aberrances of FXS NPC differentiation we investigated the differentiation of BDNF-deficient FXS NPCs derived from dMT mice (Uutela et al., 2012). BDNF protein expression was reduced by 50% in dMT NPCs compared with that in WT and Fmr1 KO NPCs (Figure 3B). We found that the amplitude of [Ca²⁺]_i responses to high [K⁺]_e was reduced in dMT NPCs compared with that in Fmr1 KO NPCs at days 1 (Figures 3C–3E) and 7 of differentiation (data not shown). Expression of both L- and T-type Ca_v channels was reduced in dMT NPCs; the expression of Cacna1h was reduced versus Fmr1 KO and control NPCs and the expression of Cacna1d versus controls (Figure 3F).

Figure 3.

Effects of Reduced BDNF Expression on Augmented Responses to Activity in FXS NPCs

(A) BDNF mRNA expression in human iPSC-derived progenitors at day 1 of differentiation. CTRL(n) = 3 cell lines (HEL11.4, HEL23.3, and HEL46.11) and FXS(n) = 3 cell lines (HEL69.5, HEL70.3, and HEL10.2).

(B) BDNF protein expression in neurospheres derived from WT, Fmr1 KO, and dMT mouse brain. Data are from three cultures of each experimental group.

(C) Fura-2 images showing migration area of WT, Fmr1 KO, and dMT neurospheres. Scale bars, 100 μm.

(D) Representative single-cell response after stimulation with 75 mM [K⁺]_e in WT, Fmr1 KO, and dMT NPCs during the period indicated.

(E) Average amplitude of [Ca²⁺]_i in cells in three zones of the migration area of WT, Fmr1 KO, and dMT neurospheres after depolarization with 75 mM [K⁺]_e. WT(n) = 8, Fmr1 KO(n) = 8, and dMT(n) = 7 neurospheres; responses of 75–100 cells were measured in each experiment.

(F) The relative mRNA expression of Cacna1c, Cacna1d, Cacna1g, and Cacna1h in dMT NPCs compared with WT and Fmr1 KO mouse NPCs at day 1 of differentiation. Each group(n) = 4–5 in twice repeated experiments.

Data are expressed as means ± SEM; ^∗∗∗p < 0.001, ^∗∗p < 0.01, and ^∗p < 0.05, one-way ANOVA with Tukey post hoc test.

Increased Ca²⁺ Influx Augments Responses to Type I mGluR Activation in FMRP-Deficient NPCs

Overactive or inappropriate signaling via the type I mGluRs is implicated as the key mechanism in the pathophysiology of FXS (Huber et al., 2002). Our previous studies showed that [Ca²⁺]_i responses to (S)-3,5-dihydroxyphenylglycine (DHPG), a specific agonist of type I mGluRs, are augmented in FXS NPCs (Achuta et al., 2017) and that FXS NPCs give rise to abnormally high amounts of cells expressing Ca²⁺-permeable AMPA receptors (Achuta et al., 2018), which could facilitate mGluR5-signaling via L-type Ca_v channels (Kim et al., 2015). DHPG stimulation induced robust sustained [Ca²⁺]_i responses in cells correlating with radial glial cells in differentiating mouse NPCs (Figure 4A). The amplitude of DHPG responses was decreased in dMT NPCs below WT levels at days 1 and 7 of differentiation in the presence of extracellular calcium ([Ca²⁺]_e) (Figures 4A and 4B). When [Ca²⁺]_e was removed, DHPG induced only a transient [Ca²⁺]_i rise (Figure 4C), and the average amplitude of DHPG responses did not differ in Fmr1 KO, WT, and dMT NPCs (Figures 4C and 4D). The data indicated that the DHPG-induced [Ca²⁺]_i increase in Fmr1 KO NPCs was primarily caused by increased Ca²⁺ influx and reduced BDNF dampened DHPG responses by affecting Ca²⁺ entry into cells. The effects were specific to mGluR signaling, and the magnitude of [Ca²⁺]_i responses to kainic acid (KA) or N-methyl-D-aspartate (NMDA) did not differ in WT, Fmr1 KO, and dMT progenitors (Figures 4E and 4F).

Figure 4.

Effects of Extracellular Ca²⁺ and Reduced BDNF on Intracellular Ca²⁺ Responses in FMRP-Deficient NPCs

(A and B) Representative single-cell response (A) and average amplitude of [Ca²⁺]_i responses to DHPG in the presence of [Ca²⁺]_e in WT, Fmr1 KO, and dMT NPCs at day 1 and 7 of differentiation (B).

(C–F) Representative single-cell response (C) and average amplitude of [Ca²⁺]_i responses to DHPG in the absence of [Ca²⁺]_e in WT, Fmr1 KO, and dMT neurospheres at day 7 of differentiation (D). Average Ca²⁺amplitude of responses in WT, Fmr1 KO, and dMT NPCs (E) to kainate (KA) and (F) to NMDA at day 7 of differentiation. WT(n) = 6, Fmr1 KO(n) = 8, and dMT(n) = 5 neurospheres at day 1; WT(n) = 3–5, Fmr1 KO(n) = 3–5, and dMT(n) = 3–5 neurospheres at day 7; responses of 75–100 cells were measured in each experiment.

Data are expressed as means ± SEM. ^∗∗∗p < 0.001, one-way ANOVA with Tukey post hoc test.

Reduced BDNF Deceases Ca_v Channel Expression and Alters Progenitor Mobility and Differentiation in FMRP-Deficient Neurospheres

Having established a role for BDNF in increased Ca_v1 channel activity in FMRP-deficient NPCs, we examined the effects of reduced BDNF on the previously reported enhanced differentiation and motility of FXS NPCs (Achuta et al., 2017). We found that the enhanced progenitor differentiation to glutamate-responsive cells characteristic of FXS NPCs was normalized in dMT neurospheres. The proportion of cells responsive to glutamate agonists (DHPG/KA/NMDA) was reduced in dMT NPCs to WT levels at days 1 and 7 of differentiation (Figure 5A). We also investigated the migration of NPCs by monitoring the motility of freely moving neuron-like cells that appear at the outer border of the migration area during the first 24 hr of differentiation by time-lapse imaging (Figure 5B). Images of cells migrating out from the neurospheres were taken every 15 min. We observed that the freely moving cells migrated with a higher velocity in Fmr1 KO neurospheres than in WT neurospheres during 12–24 hr of differentiation and that the mean velocity was normalized in dMT NPCs (Figure 5C). A detailed analysis revealed that the velocity was particularly affected in FMRP-deficient NPCs during 12–18 hr (Figure 5D), when reduced BDNF expression normalized the motility of cells by decreasing the ratio of fast (surges) and slow (stalling) motility phases in dMT neurospheres to WT levels (Figure 5E). The results are consistent with the previous report showing that BDNF modulates the phasic motility of differentiating neuron-like cells during neurosphere differentiation by promoting the initiation and maintenance of phases of cell motility (Jansson et al., 2012).

Figure 5.

Effects of Reduced BDNF on the Expression of the Ca_v Channels and Differentiation and Motility of Progenitors in FMRP-Deficient Neurospheres

(A) The proportion of cells responsive to glutamate receptor agonists (DHPG/KA/NMDA) in WT, Fmr1 KO, and dMT NPCs at day 1 and 7 of differentiation (represented as number of cells responded/total number of cells). Each experimental group(n) = 5–8 and 3–5 neurospheres at days 1 and 7, respectively; responses of 75–100 cells were measured in each experiment.

(B) Representative images showing cells migrating out from the neurosphere at 0- and 24-hr time points. Neuron-like cells that moved freely outside the layer of radial glia are denoted as freely moving (fm) cells. Scale bar, 100 μm.

(C and D) Average speed of movement of fm cells at 12–24 hr (C) and 12–18 and 18–24 hr in neurospheres derived from WT, Fmr1 KO, and dMT mice (D).

(E) The time each cell moved fast (rate >30 μm/hr, surges) divided by the time it moved slowly (rate <30 μm/hr, stalling) in WT, Fmr1 KO, and dMT neurospheres at 12–18 hr of differentiation (represented as percent of control). WT(n) = 10, Fmr1 KO(n) = 10, and dMT(n) = 8 neurospheres.

(F and G) The proportion of cells responsive to DHPG in the presence of [Ca²⁺]_e (F) and after withdrawal of [Ca²⁺]_e in WT, Fmr1 KO, and dMT neurospheres (G).

(H–J) Representative single-cell responses to kainate (KA) and NMDA (H). The proportion of cells responsive to (I) KA and to (J) NMDA in WT, Fmr1 KO, and dMT NPCs at day 7 of differentiation. Each group(n) = 3–5; responses of 75–100 cells were measured in each experiment.

Data are expressed as means ± SEM. ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05, one-way ANOVA with Tukey post hoc test (A, F, G, I, and J), chi-square statistics followed by Bonferroni post hoc test (C–E).

The proportion of all DHPG-responsive progenitors was smaller in dMT neurospheres than that in Fmr1 KO and WT neurospheres in the presence of [Ca²⁺]_e (Figure 5F). When [Ca²⁺]_e was removed, the increased proportion of DHPG-responsive NPCs in Fmr1 KO neurospheres (Castrén et al., 2005) was normalized to WT levels in dMT neurospheres (Figure 5G). Finally, the proportion of cells responsive to KA or NMDA without DHPG responses was further increased from the abnormally high Fmr1 KO levels in dMT NPCs (Figures 5H–5J). Altogether, the data are in agreement with a special role of BDNF-mediated calcium dynamics in the differentiation of distinct subsets of NPCs (Louhivuori et al., 2011).

Increased Phosphorylation of TrkB Receptors in Fmr1 KO NPCs with the Genetic Deletion of BDNF

We found that the phosphorylation of the phospholipase-Cγ1 (PLCγ1)-binding tyrosine within TrkB receptors (TrkB^y816) was increased in undifferentiated dMT NPCs when normalized with total TrkB expression and compared with that of Fmr1 KO NPCs (Figures 6A–6C). We did not detect any increase in the TrkB phosphorylation within the other phosphorylation sites (Figure S1), suggesting that reduced BDNF expression regulated specifically the PLC-γ1 signaling in FXS NPCs. Indeed, Ca_v1 channel antagonist nimodipine has been shown to specially increase neuronal TrkB phosphorylation of the PLCγ1 domain in the mouse prefrontal cortex and hippocampus (Koskimäki et al., 2015). Thus, studies of dMT NPCs revealed the link between dysregulated dihydropyridine-sensitive Ca_v channels and TrkB^y816 phosphorylation in NPCs lacking FMRP.

Figure 6.

TrkB Phosphorylation in FMRP-Deficient NPCs

(A) Western blot image showing total TrkB (tTrkB), phosphorylated TrkB^y816 (pTrkB), and GAPDH expression in WT, Fmr1 KO, and dMT mouse neurospheres. A bar graph presenting optical density values showing (B) TrkB expression and (C) the increased phosphorylation of TrkB^y816 in WT, Fmr1 KO, and dMT NPCs. (n) = 4 per group. Data are expressed as means ± SEM. ^∗p < 0.05 one-way ANOVA with Tukey post hoc test (B) and Student's t test (C). See also Figure S1.

Discussion

Calcium signaling is implicated in the regulation of neuronal differentiation, migration, and survival (Platel et al., 2008, Rosenberg and Spitzer, 2011). Ca²⁺ influx triggers a wide range of intracellular mechanisms, such as gene transcription and neurotransmitter release (Simms and Zamponi, 2014, Wheeler et al., 2012). L-type Ca_v channels are the key regulators of [Ca²⁺]_i and depolarization-induced Ca²⁺ influx into neurons is primarily mediated via these channels. Here we demonstrated a central role for Ca²⁺-dependent mechanisms and increased Ca²⁺ influx through L-type Ca_v channels in altered differentiation of FXS NPCs. We showed that [Ca²⁺]_i responses to membrane depolarization were substantially augmented in both human and mouse FXS NPCs at the early stage of progenitor differentiation. The increased Ca²⁺ entry into NPCs via nifedipine-sensitive channels correlated with enhanced NPC differentiation and increased expression of the pore-forming α1 subunits of the L-type channels versus those of the T-type channels. Since Ca_v channel activity leads to increased BDNF expression, we investigated the effects of genetic reduction of BDNF expression on the altered differentiation of FXS NPCs. Reduced BDNF decreased the expression of the α subunits of both L- and T-type Ca_v channels and diminished [Ca²⁺]_i responses to depolarization. Both removal of [Ca²⁺]_e and reduced BDNF dampened enhanced responses to type I mGluR activation in FMRP-deficient mouse NPCs, suggesting that increased Ca²⁺ entry via L-type Ca_v channels facilitated mGluR5-signaling. BDNF-dependent mechanisms were involved in the regulation of calcium dynamics during differentiation, migration, and fate determination of FXS NPCs.

Inappropriate expression or dysregulation of Ca_v channels have been linked with various neurological disorders, including autism spectrum disorder (ASD), epilepsy, chronic pain, and migraine (Cain and Snutch, 2011, Pinggera et al., 2015, Strom et al., 2010). Two gain-of-function mutations in the genes encoding the Ca_v1.3 are associated with ASD (Pinggera et al., 2015). A gain-of-function phenotype of Ca_v1.2 is found in Timothy syndrome, which is characterized by lethal cardiac arrhythmia, immune deficiency, cognitive abnormalities, and autism (Splawski et al., 2004). Our present observations revealed that augmented Ca_v1 channel signaling in FMRP-deficient NPCs contributes to the pathophysiology of FXS that represents a clinically well-characterized variant of ASD. Most FXS individuals show at least some behavioral features of autism, and up to 60% of FXS males meet the diagnostic criteria of autism (Lozano et al., 2014). Our data suggest that mechanisms associated with activation of Ca_v channels could provide common molecular mechanisms leading to the autism phenotype in intellectual disability syndrome.

Both high-voltage activated (HVA) and low-voltage activated (LVA) Ca²⁺ channels are shown to be expressed in NPCs (D'Ascenzo et al., 2006, Lepski et al., 2013, Louhivuori et al., 2013). The mRNA expression of the pore-forming α1 subunits of all Ca_v channels was low in both human and mouse neurospheres at day 1 of differentiation. In mouse NPCs, we detected expression of the Cacna1, Cacna1d, and Cacna1a genes encoding HVA channels Cav1.2, Cav1.3, and Cav2.1, respectively, and the Cacna1g and Cacna1h genes encoding LVA channels Cav3.1 and Cav3.2, respectively. The subunit expression showed high variation, particularly in human NPCs. Although there was a tendency for an increase in Cacna1c and Cacna1d transcripts and a decrease in Cacna1g transcripts in mouse FXS NPCs, no significant differences in the expression of single subunits between FXS and their respective control NPCs were found. However, the ratio of the dihydropyridine-sensitive HVA Ca_v1/LVA Ca_v3 expression was increased. The FXS-specific change of Ca_v channel expression also existed in human iPSC-derived NPCs, which started to express less Ca_v3.2 during the first day of differentiation resulting in the increased ratio of Ca_v1/Ca_v3. There is evidence that Ca_v1 channels become more important as the neural cells mature; these results are consistent with enhanced differentiation of FXS NPCs. Reduced BDNF expression in FMRP-deficient NPCs decreased expression of both L- and T-type Ca_v channels and correlated with dramatically diminished [Ca²⁺]_i responses to depolarization.

According to the present data, augmentation of DHPG responses in FXS NPCs requires [Ca²⁺]_e and is facilitated by increased Ca²⁺ entry. Exaggerated group 1 mGluR function is a fundamental abnormality in FXS, and in Fmr1-KO mice mGluR-mediated long-term depression (LTD) is enhanced (Huber et al., 2002) and epileptiform discharges are prolonged (Chuang et al., 2005). Elevated responsiveness of FXS neurons associates with hyperarousal, sensory hypersensitivity, and susceptibility to epilepsy in the FXS phenotype (Lozano et al., 2014). By coupling Gq-type G-proteins, mGluR5 activates the PLC/inositol triphosphate receptor signaling pathway to mobilize [Ca²⁺]_i (Conn and Pin, 1997). The studies of Kim et al. (2018) showed that DHPG is not sufficient to activate PLC alone and local AMPA-mediated depolarization and Ca²⁺ influx via Ca_v1 channels are required to facilitate mGluR5-PLC cascades that underlie hippocampal mGluR-LTD. Both Ca_v1.2 and Ca_v1.3 are involved in glutamate-induced PLC activation, but Ca_v1.2 is the major isoform that interacts with mGluR5 and mediates L-type currents (Kato et al., 2012). We observed that both isoforms contributed to the increased expression of L-type channels in FXS NPCs. Reduced BDNF decreased the expression of these Ca_v1 isoforms and substantially decreased Ca²⁺ influx in response to depolarization. However, given that mGluR5 signaling dictates the differentiation of NMDA-responsive cells from FXR NPCs (Achuta et al., 2017) and reduced BDNF promoted further differentiation of NMDA-responsive cells, reduced Ca²⁺ entry through L-type Ca_v channels in BDNF-deficient FXS NPCs was not sufficient to normalize the mGluR5-mediated alterations of FXS NPC differentiation.

Ca_v1.2 and Ca_v1.3 channels are often expressed in the same cell, but they differ in subcellular localization and coupling to the transcription machinery (Hasreiter et al., 2014). Their biophysical properties also differ and specific interacting proteins may selectively modulate each isoform (Wheeler et al., 2012). We found abnormal alternative splicing in the gene encoding calpastatin in human FXS NPCs. Calpastatin is an endogenous calpain inhibitor that can directly interact with the calmodulin binding site of the Ca_v1.2 isoform and reprime the channel function (Minobe et al., 2006, Sun et al., 2014). Calpain-calpastatin and calmodulin provide irreversible and reversible Ca²⁺-activated systems, respectively, to control development, growth, and maintenance of cellular homeostasis (Murachi, 1983). We found that augmented responses to L-type Ca_v channel activation in FXS NPCs also correlated with enhanced motility of progenitors. There is evidence that L-type Ca_v channels regulate cell migration by promoting filopodia stability and maturation of focal adhesions through spatially restricted regulation of Ca²⁺ entry and subsequent activation of the protease calpain-1 (Jacquement et al., 2016). How alternative splicing of calpastatin affects the calpain-calpastatin system in FXS NPCs remains to be studied.

Signaling via Ca_v1 channels provides a mechanism to induce transcription of specific genes through changes of cytosolic and nuclear Ca²⁺. Membrane depolarization induces via Ca_v1 channels transcriptional activation of the BDNF gene in cortical neurons (Tao et al., 1998). In particular, the Ca_v1.2 is implicated in transcriptional control of the BDNF gene. We showed previously that BDNF mRNA expression is increased in undifferentiated mouse FMRP-deficient NPCs and neurons (Louhivuori et al., 2011, Uutela et al., 2014). We did not observe any significant difference in the total BDNF mRNA expression between Fmr1 KO and WT NPCs at day 1 of differentiation (data not shown) when activity-dependent responses were found to be increased in FXS NPCs. Reduced differentiation of cells expressing BDNF in FMRP-deficient neurospheres (Louhivuori et al., 2011) likely affected BDNF expression. In human NPCs, BDNF mRNA expression showed high variation and did not differ significantly between FXS and control NPCs. BDNF-mediated effects on fate determination of NPCs may contribute to diverse effects of reduced BDNF on the Fmr1 KO mouse phenotype in vivo (Uutela et al., 2014). The phenotype of dMT differs substantially from that of the single deletion mutants of the Bdnf and Fmr1 gene (Spencer et al., 2011). Altogether, the data suggest that the absence of FMRP may activate homeostatic mechanisms to compensate altered BDNF/TrkB signaling. In the brain of Bdnf^+/– mice, TrkB autophosphorylation is reduced during early development (Di Lieto et al., 2012). The present study showed that the phosphorylation of the PLCγ1-binding tyrosine within the TrkB was increased to WT levels in dMT NPCs expressing less L-type channel subunits. The results are consistent with increased phosphorylation of TrkB, especially within the PLCγ1 site in the mouse prefrontal cortex and hippocampus after treatment with the blocker of the L-type channels (Koskimäki et al., 2015). Since the PLCγ docking site of TrkB is essential for hippocampal synaptic plasticity and acquisition of associative and spatial learning (Minichiello, 2009), reduced phosphorylation of the PLCγ site in FXS progenitors is in line with the impaired learning and plasticity defects in FXS. However, associative learning is highly impaired in dMT mice deficient of both FMRP and BDNF (Uutela et al., 2014), and the assessment of the consequences of altered responses of dihydropyridine-sensitive Ca_v channels in dMT NPCs for function of mature neurons and neuronal networks is complicated by probable compensatory mechanisms activated by reduced BDNF.

FXS is defined as a channelopathy, and direct interaction between FMRP and Ca_v2 channels indicates the role of N- and P/Q-type Ca_v channels in FXS (Ferron et al., 2014). The present studies demonstrated the important contribution of L-type Ca_v channels to the altered differentiation of FXS NPCs modeling molecular mechanisms disrupted in autism. Similar changes of Ca_v channel function and expression in human and mouse FXS neural progenitors confirmed the FXS specificity of our findings and highlighted the value of human iPSCs for further studies to explore correlations between functional in vitro findings and phenotypic features in FXS. Improved understanding of disordered developmental processes during early NPC differentiation may open possibilities for stem cell-based discovery of novel treatment options.

Experimental Procedures

Mice

Fmr1 KO mice (B6.129P2-Fmr1^tml/Cgr/J) and Bdnf^+/– mice (C57BL/6JOlaHsd) were purchased from Jackson Laboratory (Bar Harbor, ME) and maintained on the C57BL/6JOlaHsd substrain. Fmr1^+/− female mice were crossed with male Bdnf^+/– mice to generate dMT mice, FXS, and WT littermates as described previously (Uutela et al., 2012). All animal experiments were performed in accordance with the guidelines of the Guide for the Care and Use of Laboratory Animals and European Economic Community Council Directive. The studies were approved by the Experimental Animal Ethics Committee of the National Laboratory Animal Center, Finland.

Mouse Cell Culture

NPCs were prepared from the WT, Fmr1 KO, and dMT mice at E14 as described previously (Castrén et al., 2005). To induce neuronal differentiation, neurospheres (mid-sized 200–250-μm spheres) were plated on poly-DL-ornithine (Sigma)-coated cover glasses in culture medium without mitogens and differentiated for the indicated periods of time. The mitogens were added to cell cultures always a day before initiation of the differentiation.

Human iPSC Cultures

The research using human iPSCs was approved by the Ethical Committee of the Hospital District of Helsinki and Uusimaa. Written informed consent was provided to obtain human samples. We used four FXS iPSC lines (HEL100.1, HEL100.2, HEL69.5, and HEL70.3) and three control cell lines (HEL46.11, HEL23.3, and HEL11.4) and the protocol for their differentiation to neurospheres described previously (Achuta et al., 2018). Neurospheres (average size 200–250 μm) were differentiated for 1 day on poly-D-lysine/laminin (Sigma-Aldrich)-coated cover glasses in culture medium without mitogens.

Immunochemistry

Differentiated neurospheres were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 10 min at room temperature (RT). Unspecific staining was blocked with 10% normal goat serum and 0.1% Triton X-100 in PBS for 45 min at RT, and followed by incubation with primary antibody recognizing microtubule-associated protein 2 and glutamate-aspartate transporter (both from Millipore; 1:500) at 4°C overnight. Secondary antibodies Alexa Fluor 488 (1:2,000, A11008; Invitrogen) and Alexa 568 (1:5,000, A11004; Invitrogen) were applied in PBS containing 1% BSA for 45 min at RT. After final washes, the cell nuclei were counterstained with Vectashield mounting media containing DAPI (Vector Laboratories). The fluorescent-stained differentiated NPCs were viewed and images were obtained with an LSM 5 Pascal laser scanning confocal microscope (Zeiss) and analyzed with the image analysis program ImageJ.

Intracellular Ca²⁺ Imaging

For calcium imaging experiments, 15 to 20 neurospheres were plated on coated cover glasses and differentiated for 1 and 7 days. The intracellular calcium recordings were performed as described previously (Achuta et al., 2018). Cells were challenged sequentially with glutamate receptor agonists at the following concentrations: DHPG (10 μM, Abcam Biochemicals); kainate (50 μM, Tocris Bioscience). Nifedipine (10 μM, Sigma) was used to block Ca_v1 channels.

Time-Lapse Imaging

For time-lapse imaging, a cell-culturing instrument combined with phase-contrast microscopy, automation, and environmental control (Cell-IQ, Chip-Man Technologies) was used. Neurospheres were plated on poly-DL-ornithine-coated six-well plates without mitogens. Cellular movement of adherent cells were monitored with an automated optics module controlled by machine vision-base firmware and analysis software during the first 24 hr of differentiation. The images were analyzed by an onboard Cell-IQ analyzer software package.

ELISA

BDNF expression was determined using BDNF ELISA (Quantikine human BDNF kit, R&D Systems) following the manufacturer's protocol. The cortical progenitors were homogenized in lysis buffer containing 137 mM NaCl, 20 mM Tris (pH 8.0), 1% (vol/vol) NP-40, 10% (vol/vol) glycerol, 50 mM sodium fluoride, 2× Complete Mini (Roche Diagnostics), and 2 mM sodium vanadate. Following homogenization, samples were incubated at 4°C for 20 min and centrifuged at 13,000 rpm for 15 min at 4°C. The total protein concentration was measured using a Bio-Rad DC protein assay (Bio-Rad), and the BDNF concentration was determined as pg/mL from the standard curve. To confirm the specificity of the ELISA kit, cell culture medium and lysis buffer without protein extract were used as negative controls.

Western Blotting

Proliferating mouse NPCs were collected and lysed in RIPA buffer supplemented with 1× protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific) for western analysis. The total protein concentration was determined using a DC protein assay (Bio-Rad). Equal amounts of protein extracts (40 μg) were separated by SDS-PAGE under reducing conditions and transferred onto a polyvinylidene fluoride membrane (300 mA for 1 hr at 4°C) using a Trans-Blot Turbo System (Bio-Rad). The membranes were blocked in 3% BSA in Tris-buffered saline with Tween 20 (TBST) (pH 7.4) (0.1 M Tris, 0.15 M NaCl, and 0.1% Tween 20) at RT for 1 hr and incubated with the following primary antibodies: anti-TrkB (1:2,000; Transduction Laboratories) and anti-TrkB^Y816 (1:1,000; kind gift from Dr. M. Chao, Skirball Institute, NY, USA) at 4°C overnight. The membrane was washed three times with TBST and then incubated with horseradish-peroxidase-conjugated secondary antibody (1:10,000; GE Healthcare). Detection of immunoreactive proteins was performed using Pierce ECL western blotting substrates (Thermo Fisher Scientific) and visualized using a G:BOX Chemi XX6 imaging system (Syngene).

RNA Extraction and qRT-PCR

Total RNA was extracted from neural progenitors and mouse prefrontal cortex tissue using the RNA Purification Kit (Norgen Biotek) according to the manufacturer's instructions. cDNA was synthesized from 1 μg of RNA using the Transcriptor First Strand cDNA Synthesis kit (Roche). qRT-PCR was performed using SYBR Green I kit and a LightCycler 480 II Real-Time PCR System (Roche Diagnostics). The primers used to amplify the specific cDNA regions of transcripts are appended (Table S1). Each sample was run in triplicate with melting curve analysis to detect primer dimer artifacts. Data were analyzed using LightCycler 480 software (Roche) by the second derivative max method determining the threshold cycle (C_t). Relative quantities of the expression of the α Ca_v subunits were calculated with the ΔΔC_t method (Livak and Schmittgen, 2001). Expression of GAPDH was used to normalize the data.

Microarray Analysis

The data for the transcriptome analysis using an Affymetrix Clariom D Human array (Thermo Fisher Scientific) have been deposited in NCBI's Gene Expression Omnibus (Achuta et al., 2018) and are accessible through accession number GEO: GSE103965). For the gene-level analysis, the raw expression data were normalized using the Expression Console program (v.1.3) (https://www.thermofisher.com/). The signal space transformation robust multi-array average method was used to normalize the data (Li and Wong, 2001). To identify differentially expressed genes, an empirical Bayes moderated t test was applied using the “limma” package (v.3.34.9) (Smyth, 2005). To address the problem of multiple testing, the p values were adjusted (Benjamini and Hochberg, 1995). A gene was considered differentially expressed if the adjusted p value was less than 0.05.

Alternative splicing analysis was performed using the transcriptome analysis console (TAC) (v.4.0) (https://www.thermofisher.com/). TAC was used with the default setting and only exon-level false discovery rate p values <0.05 were considered differentially expressed. The EventPointer (Romero et al., 2016) algorithm was used to detect splicing events and to test for significant differential splicing events.

Statistical Analysis

Data are expressed as means ± SEM. Statistical analysis was performed using Student's t test or one-way ANOVA with Tukey's HSD post hoc test and chi-square statistics with Bonferroni post hoc test. A p value <0.05 was considered statistically significant.

Author Contributions

M.L.C. conceived and designed the experiments. C.D., U.-K.P., V.S.A., G.T., N.M., I.A., and V.R. performed the experiments. C.D., M.L.C., V.S.A., G.T., A.I., and V.A. analyzed the data. M.L.C., C.D., and V.S.A. wrote the manuscript.

Published: November 29, 2018