Heterozygous loss-of-function mutations in the transcription factor FOXP1 are strongly associated with autism. Dopamine receptor 2 expressing (D2) striatal projection neurons (SPNs) in heterozygous Foxp1 (Foxp1+/−) mice have higher intrinsic excitability. To understand the mechanisms underlying this alteration, we examined SPNs with cell-type specific homozygous Foxp1 deletion to study cell-autonomous regulation by Foxp1. As in Foxp1+/− mice, D2 SPNs had increased intrinsic excitability with homozygous Foxp1 deletion. This effect involved postnatal mechanisms. The hyperexcitability was mainly due to down-regulation of two classes of potassium currents: inwardly rectifying (KIR) and leak (KLeak). Single-cell RNA sequencing data from D2 SPNs with Foxp1 deletion indicated the down-regulation of transcripts of candidate ion channels that may underlie these currents: Kcnj2 and Kcnj4 for KIR and Kcnk2 for KLeak. This Foxp1-dependent regulation was neuron-type specific since these same currents and transcripts were either unchanged, or very little changed, in D1 SPNs with cell-specific Foxp1 deletion. Our data are consistent with a model where FOXP1 negatively regulates the excitability of D2 SPNs through KIR and KLeak by transcriptionally activating their corresponding transcripts. This, in turn, provides a novel example of how a transcription factor may regulate multiple genes to impact neuronal electrophysiological function that depends on the integration of multiple current types – and do this in a cell-specific fashion. Our findings provide initial clues to altered neuronal function and possible therapeutic strategies not only for FOXP1-associated autism but also for other autism forms associated with transcription factor dysfunction.
Forkhead box P1 (FOXP1) is a transcription factor that is highly enriched in the developing embryonic and mature neocortex, hippocampus, and striatum . Heterozygous mutations and deletions in FOXP1 resulting in loss-of-function are causative for Intellectual Disabilities (ID) and Autism Spectrum Disorders (ASD) [2,3,4] and are among the most significant recurrent de novo mutations associated with ASD [5,6,7]. Characterization of whole brain Foxp1 conditional knockout mice (cKO) has shown that these mice exhibit ASD-relevant repetitive behavior and reduced social interaction . Moreover, heterozygous Foxp1 (Foxp1+/−) mice, which are a relevant genetic model for FOXP1-associated ASD, also show altered vocalizations .
ASD patients have altered striatal function and connectivity [10,11,12,13], and Foxp1 is among the few genes specifically enriched in the mouse striatum . Deletion of Foxp1 in the whole brain leads to a dramatic reduction in striatal size and is required for striatal projection neuron (SPN) differentiation in vitro, indicating its critical role in striatal development [8, 9, 15, 16]. Beginning at embryonic stages and continuing into adulthood, Foxp1 is expressed in dopamine receptor subtype 1 (D1) and dopamine receptor subtype 2 (D2) expressing SPNs [1, 9, 16, 17]. However, very little is known about the role of Foxp1 in the maturation of these striatal neurons.
Some mechanistic details, such as the role of SUMOylation, in FOXP1-dependent regulation of brain development have been reported in the past . However, none of the studies have made a compelling link between downstream targets of FOXP1 – either direct or indirect – with neuronal development. Moreover, previous studies have mostly employed embryonic Foxp1 deletion and examined early developmental processes [19,20,21,22], but as Foxp1 is expressed in adult brain as well, its involvement in postnatal development is also possible. These gaps in understanding apply to other members of the FOXP family, like FOXP2. Therefore, further examination of FOXP1 function is needed, not only for FOXP1 alone, but also for the FOXP family as a whole.
FOXP1 has been reported to be involved in the development of synaptic transmission and intrinsic excitability – the latter being the propensity to fire action potentials [8, 9, 23]. Brain-wide conditional Foxp1 deletion decreases intrinsic excitability of hippocampal pyramidal neurons . In Foxp1+/− mice, intrinsic excitability is increased in D2 SPNs, but not in D1 SPNs . These studies indicate that FOXP1 may have differing roles depending on neuron type, but this issue is confounded by the different mouse models used and has not been addressed so far. Furthermore, mechanistic details of these excitability alterations remain unknown.
Both D1 and D2 SPNs express several potassium (K+) channels, including two classes of subthreshold K+ channels – leak (KLeak) and inwardly rectifying (KIR) [24,25,26,27,28,29]. KLeak channels are the two-pore domain channels that are not gated by membrane potential . Evidence supports the existence of KLeak channels, particularly KCNK2 (TREK-1), in SPNs [28, 29, 31, 32]. KIR currents in SPNs are mediated by KCNJ2 (Kir2.1) and KCNJ4 (Kir2.3) channels [24, 25, 27, 33] and are a major source of conductance at hyperpolarized potentials. While some signaling pathways in SPNs have been reported to play a role in the maturation of intrinsic excitability through the regulation of KIR currents [27, 34], the transcriptional regulation of this maturation is not known.
Transcriptional regulation of neuronal function during development, and its impairment in ASD, is not well understood. The difficulty is due to the large number of genes impacted by transcriptionally related proteins and to differing roles in different cell-types. Among the top 50 most commonly studied autism-associated genes , 14 are involved in transcriptional regulation. Among these, which includes Foxp1, a connection between ion channel transcripts and function has only been made for Mef2c and Mecp2 [36, 37]. In addition, for both of these, a direct demonstration on how this link affects neuronal intrinsic excitability has not been reported. Moreover, for both, a description of a more multi-faceted regulation involving multiple downstream transcripts and corresponding channel function and how this contributes to neuronal excitability in different neuron-types is lacking.
Using a cell-autonomous deletion strategy in either D1 or D2 SPNs, we examine the electrophysiological mechanisms by which FOXP1 regulates the maturation of intrinsic excitability. In D2 SPNs, we find that Foxp1-dependent regulation involves multiple subthreshold currents and transcripts of underlying channels. However, a similar process is either absent or diminished in D1 SPNs. Specifically, we find that FOXP1 negatively regulates hyperexcitability by promoting the function of KIR and KLeak currents in D2 SPNs during development. Using striatal single-cell RNA sequencing data, we also identify specific candidate channels that underlie these currents: KCNJ2 and KCNJ4 for KIR and KCNK2 for KLeak.
Materials and methods
All mice used were of C57BL/6J background strain. Drd2 conditional Foxp1 knockout (D2 Foxp1cKO) mice were: Drd2-Cretg/−:Foxp1flox/flox:Drd2-eGFPtg/−, while Drd1α Foxp1cKO (D1 Foxp1cKO) were: Drd1α-Cretg/−:Foxp1flox/flox:Drd1α-tdTomatotg/−. Heterozygous version of D2 Foxp1cKO is referred to as D2 Foxp1cHet.
Brain slices and recordings
Unless stated otherwise, P14-23 mice were used for the electrophysiological recordings. Only D2 SPNs were recorded in D2 Foxp1cKO mice, where WT controls and Foxp1-deleted neurons were named Foxp1CTL and Foxp1cKO D2 SPNs, respectively. Similar nomenclature applies to D1 SPNs. D2 and D1 SPNs were identified with GFP and tdTomato expression respectively. For the slices where Foxp1 was deleted postnatally using virus, uninfected SPNs were named Foxp1CTL SPNs while virus-infected neurons were named Foxp1vcKO SPNs.
In current clamp, incremental current steps (500 ms duration) were applied at resting potential to measure the number of action potentials elicited at each step to make F-I curves.
Single −10 voltage step to measure input resistance, capacitance, and conductance
In voltage clamp, a single −10 mV voltage step (500 ms duration) was applied to measure input resistance, capacitance, normalized cell conductance, and the voltage-dependence of holding current at −85, −65, and −55 mV holding potentials.
Multi-step protocol using IV-plots
In voltage clamp, a multi-step protocol (500 ms duration) was applied while maintaining −60 mV holding potential (ranging from −30 to −120 mV voltages in −10 mV steps). The current measured was the average of a 200 ms window at the end of the step (see Fig. 2E).
Morphology and analysis
Dendritic morphology was measured by filling neurons with a fluorescent dye (AF488, 100 µM, Life Technologies, Inc.) and biocytin (2 mg/ml, Life Technologies, Inc.) during the whole-cell recording. Sholl analysis is performed using concentric radii step intervals of 20 µm and dendritic crossings at each radius are counted.
For all data, sample number is the number of neurons, and data are stated in this order: WT control followed by Foxp1 deletion groups. All data are plotted as mean ± standard error. Unless stated otherwise, a repeated measures (RM) 2-way ANOVA is performed with Geisser-Greenhouse correction followed by the Holm-Sidak test for the correction for multiple comparisons. When ANOVA analysis is not employed, we use a Mann-Whitney (MW) t-test.
Single-cell RNA sequencing
We used single-cell RNA sequencing data from our previous study and analyzed D1 and D2 SPNs .
Increased intrinsic excitability in D2 SPNs with cell autonomous Foxp1 deletion
We first examined how intrinsic excitability was altered in D2 SPNs with homozygous Foxp1 deletion. We performed selective Foxp1 deletion in D2 SPNs using D2 Foxp1cKO mice to isolate FOXP1-regulated cell-autonomous mechanisms. We targeted D2 SPNs and obtained “firing versus current injection” curves, or “F-I” curves (Fig. 1A,B). Foxp1cKO D2 SPNs exhibited a ~30–500% increase in the number of evoked action potentials when compared to the Foxp1CTL D2 SPNs, indicating that FOXP1 negatively regulates intrinsic excitability of D2 SPNs.
Subthreshold membrane properties of the D2 SPNs were also altered upon Foxp1 deletion as observed by higher resting membrane potential (Fig. 1C), higher input resistance (Fig. 1D), and altered holding current density (Fig. 1E). When the average holding current density in Foxp1cKO D2 SPNs was subtracted from that of Foxp1CTL D2 SPNs, a difference current density was obtained with a reversal potential near −85 mV (Fig. 1F). This is very close to the estimated reversal potential for K+ in our experiments (−88 mV). These changes indicate that a down-regulation of subthreshold K+ currents with Foxp1 deletion causes an increase in input resistance and a depolarized resting potential, thereby contributing to their hyperexcitability.
To examine how these subthreshold alterations contributed to increased intrinsic excitability, we re-analyzed the F-I data in Fig. 1B and plotted the number of spikes at different epochs during the current step (Fig. 1G). Action potential firing in the first 50 ms epoch is expected to depend most on the subthreshold resting state. For the subsequent epochs, depolarization-activated currents are engaged and the impact of the previous subthreshold properties is diminished. The greatest increase in excitability is observed in the first 50 ms epoch compared to all other subsequent epochs suggesting that subthreshold membrane properties are involved in the hyperexcitability of Foxp1cKO D2 SPNs.
To better isolate the role of input resistance and spike voltage-threshold changes underlying increased intrinsic excitability, we analyzed the current step data only for the neurons having resting potential in the range between −65 mV and −85 mV. The F-I curves showed similar hyperexcitability for Foxp1cKO D2 SPNs (Data not shown). In this analysis, the average resting potential and spike voltage-threshold were not different in Foxp1cKO D2 SPNs (Table 1). However, spike current-threshold and input resistance were both approximately doubled. Therefore, increased input resistance appears to be a key mediator for hyperexcitability.
To determine if D2 SPN hyperexcitability in Foxp1+/− mice is due to cell-autonomous deletion , we performed the same experiments in D2 SPNs that were selectively heterozygous for Foxp1 (Foxp1cHet). We found no change in excitability or subthreshold properties (Supplementary Fig. S1A–D) indicating that hyperexcitability in D2 SPNs of the Foxp1+/− mouse may not be due to a purely cell-autonomous process.
No significant morphological or anatomical changes with Foxp1 deletion in D2 SPNs
It is possible that Foxp1cKO D2 SPNs are smaller in size and the resulting reduction in membrane surface area could contribute to increased input resistance and increased excitability. To test this idea, we filled D2 SPNs with biocytin, and measured branching complexity (Sholl Analysis) and total dendritic length. We could not find any effect of Foxp1 deletion on these properties (Fig. 1H–J), which suggests that cell morphology is not affected. We also measured soma size in live, acutely prepared slices and found no difference (Fig. 1K). In summary, there was no anatomical evidence for change in membrane surface area.
Another parameter that crudely estimates the membrane surface area is the cell capacitance. When pooled over all experiments conducted under control conditions in our study, there was no detectable change in the cell capacitance after Foxp1 deletion (Supplementary Fig. S2, Control). However, under strong blockade of multiple channels, which may more accurately report membrane surface area (or “cell size”) due to better space clamp, Foxp1cKO D2 SPNs had a 19% reduction in capacitance (Supplementary Fig. S2, Cs+,Ba2+). In summary, while this change is difficult to interpret, there may be some contribution of decreased membrane surface area to the hyperexcitability of Foxp1cKO D2 SPNs.
Postnatal Foxp1 deletion in D2 SPNs also results in hyperexcitability
It is unclear whether hyperexcitability of D2 SPNs is due to deletion of Foxp1 during embryonic or postnatal development. When Foxp1 is deleted during embryonic development, indirect effects involving cell fate determination and migration may affect excitability. This could indeed be a factor since D2 SPN number is reduced in D2 Foxp1cKO mice (see Supplementary Methods). However, if hyperexcitability can be induced with Foxp1 deletion at a postnatal stage, then the underlying mechanism is likely independent of these early developmental processes. To test this, we deleted Foxp1 postnatally by infecting SPNs with AAV-Cre-GFP in P0-1 Foxp1flox/flox mice, resulting in Foxp1 deletion by P10 (see Supplementary Methods). Uninfected D2 SPNs in the same mice served as controls (Foxp1CTL).
The F-I curves showed an increased intrinsic excitability of D2 SPNs with Foxp1 deletion (Fig. 1L, M). There was a statistical trend for a depolarized resting potential (p = 0.06, Fig. 1N) and input resistance was higher (Fig. 1O) in Foxp1vcKO D2 SPNs, which was consistent with the down-regulation of a subthreshold K+ current. Dendritic complexity, total dendritic length, soma size, and capacitance were unaltered (Supplementary Fig. S3A–E). When AAV-Cre-GFP was injected into WT mice, none of the electrophysiological changes were observed (Supplementary Fig. S4A–C). These results indicate a postnatal role for Foxp1 in regulating excitability.
With postnatal deletion, there was no decrease observed in D2 SPN number as with embryonic deletion in D2 Foxp1cKO mice (see Supplementary Methods) while the mechanism inducing hyperexcitability is similar for both deletion strategies. Since embryonic deletion was more efficient, consistent, and less intrusive, we used D2 Foxp1cKO mice for the remainder of this study.
Down-regulation of Cs+-sensitive currents indicates that KIR is decreased in D2 SPNs with Foxp1 deletion
To determine the identity of the subthreshold K+ currents that are downregulated with Foxp1 deletion, we examined the effect of specific K+ channel blockers on D2 SPNs. KIR current (KIR) is a prime candidate since KIR channels are highly expressed in SPNs and are open only at hyperpolarized, subthreshold membrane potentials [24, 25, 27, 33]. To measure the effect of Foxp1 deletion on KIR, we used Cs+ to isolate KIR at subthreshold potentials [25, 38,39,40] (Supplementary Table S1, S2).
In nominal ACSF, we observed the similar approximate doubling of the input resistance in Foxp1cKO D2 SPNs as observed in Fig. 1D (Fig. 2A). We also calculated normalized conductance as a measure of membrane ion permeability that is more independent of membrane surface area and found this to be roughly halved (Fig. 2C). With application of CsCl (1 mM), the effects relevant to KIR were best observed at −85 mV membrane potential (see Supplementary Methods, Membrane potentials for measuring KIR and KLeak). For both Foxp1CTL and Foxp1cKO D2 SPNs, there was an increase in the input resistance and decrease in the normalized conductance at −85 mV (Fig. 2A, C). Consistent with the down-regulation of KIR with Foxp1 deletion, these Cs+-induced absolute changes at −85mV were less in Foxp1cKO D2 SPNs. However, the proportional change in both of these properties was not different between Foxp1CTL and Foxp1cKO D2 SPNs (Fig. 2B, D) indicating that Cs+-sensitive currents made equal proportional contributions to these properties in both groups. These results indicate that while Foxp1 deletion did indeed reduce KIR, it did not alter the proportion of KIR relative to other currents.
We used a multi-step voltage protocol to measure KIR current density (Fig. 2E). Current-voltage plots (IV-plots) derived from these measurements indicated that evoked current densities were consistently less in Foxp1cKO D2 SPNs compared to Foxp1CTL D2 SPNs and application of Cs+ affected these currents (Fig. 2E, F). For each neuron, the amount of Cs+-sensitive KIR was approximated by subtraction of currents (Fig. 2E, right). Average of these difference currents indicated a significant reduction in KIR with Foxp1 deletion as observed in the steps to −120 to −90 mV (Fig. 2G). With conditional heterozygous deletion, we observed no change in KIR (Supplementary Fig. S5A–C), which is consistent with the unaltered intrinsic excitability under these conditions (see Supplementary Fig. S1).
Down-regulation of Ba2+-sensitive/ Cs+-insensitive currents indicates that KLeak is decreased in D2 SPNs with Foxp1 deletion
Both KIR and KLeak are thought to be prominent subthreshold currents in SPNs. Because differences in subthreshold currents and conductances remained between Foxp1CTL and Foxp1cKO D2 SPNs after blocking KIR with Cs+, we next determined if the remaining difference was due to a putative KLeak (using the term “putative” to acknowledge the lack of effective, specific KLeak blockers in neurons). We examined KLeak through application of BaCl2 (4 mM) since barium blocks most KLeak channel subtypes – including KCNK2 (Supplementary Table S1, S2). CsCl (1 mM) was always present to examine membrane properties independent of KIR. This experiment directly measures FOXP1-regulation of Ba2+-sensitive/Cs+-insensitive currents, and many KLeak subtypes fit this description - including KCNK2. Because Cs+ does not completely block KIR [39, 40], currents collected in control conditions at and below −75 mV membrane potential may have included some residual KIR, and therefore, currents at and above −65 mV membrane potential were analyzed for putative KLeak (see Supplementary Methods, Membrane potentials for measuring KIR and KLeak). It is possible that a change in KCNQ currents, known to be activated at these potentials [41, 42], would confound our KLeak measurement, but direct measurements of KCNQ currents revealed no change with Foxp1 deletion (Supplementary Fig. S6).
Application of Ba2+ led to a decrease in conductance (Fig. 3A) and an increase in input resistance (Supplementary Fig. S7A, B) across all membrane potentials in both Foxp1CTL and Foxp1cKO D2 SPNs. Absolute changes in these properties were smaller in Foxp1cKO D2 SPNs, which was consistent with reduced putative KLeak. If FOXP1 preferentially regulates KLeak, we would predict a decreased proportional effect on conductance and input resistance by Ba2+ application in Foxp1cKO D2 SPNs. Indeed, this was the case (Fig. 3B, Supplementary Fig. S7C). From this, we conclude that Foxp1 deletion preferentially downregulated putative KLeak compared to other subthreshold currents.
Data from IV-plots (Fig. 3C, Supplementary Fig. S8) further supported KLeak being reduced with Foxp1 deletion. As expected for incomplete KIR blockade, plots obtained in control ACSF at membrane potentials below −70 mV displayed inflections indicating that KLeak was not isolated at those potentials (Fig. 3D, left). Application of Ba2+ strongly reduced current densities at all membrane potentials by blocking most of the KLeak and residual KIR and induced a linearization of the IV-plots (Fig. 3D, right). As depicted in difference IV-plots (Fig. 3E, left), the reduction in current density in Foxp1cKO D2 SPNs was less compared to Foxp1CTL D2 SPNs at depolarized membrane potentials, particularly at −70 and −50 mV, confirming that putative KLeak was downregulated with Foxp1 deletion. The same data replotted into genotypic difference IV-plots (Fig. 3E, right) not only confirmed our finding, but also validated that the downregulated current was a K+ current as it either crossed at or was not detectably different from the x-axis at the K+ reversal potential (−88 mV).
We also examined the down-regulation of Ba2+- and Cs+-sensitive currents combined. In the earlier Cs+-application experiment, the disparity in conductance between Foxp1CTL and Foxp1cKO D2 SPNs in control conditions was 215 pS/pF (Fig. 2C, left, at −85 mV). When we block KIR and KLeak with both Ba2+ and Cs+, this disparity was reduced to 6 pS/pF (Fig. 3A, right, at −85 mV). In terms of percentage, the conductance disparity decreased from 53 to 17% of control. Therefore, KLeak and KIR account for a large portion of the conductance that is downregulated with Foxp1 deletion.
Targeting KCNK2 channels further implicates KLeak down-regulation with Foxp1 deletion
Pharmacological blockade in the experiments above did not provide a complete accounting for currents downregulated with Foxp1 deletion since differences remained between Foxp1CTL and Foxp1cKO SPNs (Fig. 3A, right, 3D, right). One possible reason could be an incomplete block of KLeak by Ba2+. KCNK2 is the most expressed KLeak channel in the striatum [28, 29] which is only 90% blocked by Ba2+ .
We examined whether additional blockade of KCNK2 channels could further reduce the disparity in currents between Foxp1CTL and Foxp1cKO D2 SPNs, which would provide additional evidence that KLeak is regulated by FOXP1. We first performed recordings in control ACSF and then in the presence of acidification (pH6.2) and Spadin (500 nM, referred to by pH6.2/Spadin-application) to specifically block KCNK2 channels [44,45,46,47]. Because we wanted to determine the effect of blocking KCNK2 channels relative to the Cs+ and Ba2+ blockade examined above, extracellular Cs+ and Ba2+ were always present.
As observed in similar conditions above (see Fig. 3A, right), normalized conductance was significantly lower in Foxp1cKO D2 SPNs than Foxp1CTL SPNs in control ACSF (Fig. 3F, left). With pH6.2/Spadin-application, this disparity was no longer observed (Fig. 3F, right). Furthermore, the percent reduction in the normalized conductance was higher in Foxp1CTL D2 SPNs as compared to Foxp1cKO D2 SPNs indicating a greater contribution of pH6.2/Spadin-sensitive currents in control SPNs (Fig. 3G). These results indicated that KCNK2 was preferentially decreased with Foxp1 deletion relative to other subthreshold currents.
In IV-plots, we observed effects of pH6.2/Spadin-application (Fig. 3H, I, Supplementary Fig. S9), and the difference IV-plot (Fig. 3J, left) indicated less pH6.2/Spadin-sensitive currents in Foxp1cKO D2 SPNs compared to Foxp1CTL D2 SPNs. These difference plots also indicated that pH6.2/Spadin-sensitive currents were not purely KCNK2-mediated since they did not cross the x-axis at the expected reversal potential for K+ and displayed a downward deflection at depolarized potentials. We attribute these effects to a non-specific effect of acidification on membrane currents. However, when we replot the data into genotypic difference IV-plots (Fig. 3J, right), the non-specific effects are subtracted away and the difference currents at −70 mV and above illustrate down-regulation of KCNK2 in Foxp1cKO D2 SPNs (see Supplementary Methods, Membrane potentials….).
D1 SPNs become hyperexcitable with Foxp1 deletion, but through different mechanisms compared to D2 SPNs
In Foxp1+/− mice, D2 SPNs were hyperexcitable, but not D1 SPNs . Therefore, either Foxp1 is not involved in maturation of intrinsic excitability or a single copy of the Foxp1 allele is sufficient to maintain normal excitability in D1 SPNs. To investigate this further, we examined the cell-autonomous effects of Foxp1 deletion in D1 SPNs using D1 Foxp1cKO mice. In F-I curves, we observed increased intrinsic excitability in Foxp1cKO D1 SPNs compared with Foxp1CTL D1 SPNs (Fig. 4A, B), indicating a cell-autonomous role for FOXP1 in the proper maturation of intrinsic excitability. However, subthreshold current changes in Foxp1cKO D1 SPNs were much less than that observed in Foxp1cKO D2 SPNs. The changes in resting potential (Fig. 4C) and input resistance (Fig. 4D) were either small or not significantly different. IV-plots were unchanged – including the difference plot depicting isolated KIR (Fig. 4E, F). Postnatal Foxp1 deletion, using AAV-Cre-GFP infection in D1 SPNs, also did not induce any changes in excitability or subthreshold membrane properties (Supplementary Fig. S10A–C).
As described for D2 SPNs, we performed two additional analyses of F-I curves to determine the relative roles of subthreshold and suprathreshold changes underlying increased intrinsic excitability in Foxp1cKO D1 SPNs. First, we replotted the F-I curve data (in Fig. 4B) at different 50 ms epochs during the current step. Unlike D2 SPNs, we found that the relative differences between Foxp1CTL and Foxp1cKO D1 SPNs were similar in both early and later 50 ms epochs (Fig. 4G) suggesting a decreased role for subthreshold membrane properties in the hyperexcitability of Foxp1cKO D1 SPNs compared to Foxp1cKO D2 SPNs. (see Fig. 1G).
Next, we replotted F-I curves only for D1 SPNs falling within a limited resting potential range (between −75 and −90 mV) and found a similar 2-fold increase in the spike numbers (data not shown). However, the average resting membrane potential and input resistance were not detectably different (Table 1). Like D2 SPNs, there was no effect of Foxp1 deletion on the threshold potential. The threshold current step to evoke a spike was decreased by 17% in Foxp1cKO D1 SPNs compared to controls (there was a 44% decrease in Foxp1cKO D2 SPNs, Table 1). Again, these data indicate that subthreshold alterations play a less prominent role in the hyperexcitability of Foxp1cKO D1 SPNs when compared with Foxp1cKO D2 SPNs. Consistent with this possibility, changes in action potential waveform suggest a greater role for suprathreshold currents in D1 SPNs as compared to D2 SPNs (Supplementary Table S3).
Single-cell RNA sequencing data indicate that KIR and KLeak channel transcripts are downregulated with Foxp1 deletion in D2 SPNs
We analyzed our single-cell RNA-seq (scRNA-seq) dataset  from D1 and D2 SPNs to determine whether expression of subthreshold potassium channels are altered when Foxp1 is deleted. Among KIR channels, KCNJ2 and KCNJ4 are the most highly expressed channels in SPNs [24, 25, 33]. Our analysis also showed high expression of their transcripts in Foxp1CTL D1 and D2 SPNs compared to other Kcnj genes (Fig. 5A). However, Foxp1 deletion significantly reduced their expression in D2 SPNs, consistent with the reduced KIR currents observed in Foxp1cKO D2 SPNs (Fig. 5B, C). In contrast, only the expression of Kcnj4 was found to be reduced in Foxp1cKO D1 SPNs, and this reduction appeared slightly less than that observed in D2 SPNs.
The KLeak currents are mediated by the KCNK family of potassium channels. Kcnk2 is highly expressed in the striatum [28, 29, 31, 32], and consistent with this, we found higher Kcnk2 transcript expression in Foxp1CTL D1 and D2 SPNs compared to other Kcnk genes (Fig. 5A). Upon Foxp1 deletion, Kcnk2 transcript levels were significantly reduced in D2 SPNs (Fig. 5D), suggesting an underlying molecular mechanism for decreased KLeak currents. Kcnk2 expression was decreased in Foxp1cKO D1 SPNs as well; however, this reduction was less compared to Foxp1cKO D2 SPNs.
Our study provides the most thorough assessment to date of a role for FOXP1 in the normal physiological development of neurons. We report that FOXP1 negatively regulates intrinsic excitability of D2 SPNs during development through positive regulation of two subthreshold K+ currents - KIR and KLeak. This regulation is cell-autonomous and likely involves FOXP1 function during both embryonic and postnatal developmental stages. Our pharmacological and single-cell RNA sequencing data indicate that FOXP1 promotes the expression of KCNJ2, KCNJ4, and KCNK2 channels, thereby providing specific channel candidates for FOXP1-mediated regulation of KIR and KLeak. FOXP1 also appears to negatively regulate intrinsic excitability in D1 SPNs, but unlike D2 SPNs, subthreshold currents play a much smaller role.
A larger role of KLeak in D2 SPN intrinsic excitability and its regulation by FOXP1
Little is known about the developmental and functional roles of KLeak in SPNs. This partly stems from the lack of pharmacological reagents that can specifically and effectively block the KCNK channels in neurons. Nevertheless, the combination of pharmacological manipulations, Foxp1 deletion, and single-cell transcriptomics enabled us to examine KLeak in SPNs. In our experiments, reduction in the normalized conductance at −65 to −55 mV membrane potential due to Foxp1 deletion most likely reflected changes in KLeak. Consistent with Cs+ not blocking KLeak [42, 48], we observed minor effects on the conductance of D2 SPNs at these potentials with Cs+-application (<10%, Fig. 2C). However, we observed large effects in the presence of Ba2+, which is known to block KLeak  (Fig. 3A).
We calculated the approximate contribution of a putative KLeak to the total conductance of Foxp1CTL D2 SPNs as well as the amount that is regulated by Foxp1 (Supplementary Table S4). Putative KLeak accounted for 51% of the conductance in Foxp1CTL D2 SPNs, which decreased to 42% in Foxp1cKO D2 SPNs. This decrease is consistent with our finding of preferential regulation of KLeak by FOXP1 compared to other subthreshold currents (see Fig. 3B, G). Of the downregulated conductance due to Foxp1 deletion, KLeak contributed 62%. Even with the possibility that our calculations involve minor contamination by other currents, these calculations suggest a significant impact by KLeak on D2 SPN excitability and its regulation by FOXP1.
Mechanisms of FOXP1-mediated regulation of KIR and KLeak
Our data suggest that FOXP1 negatively regulates the excitability of D2 SPNs through KIR and KLeak by transcriptionally activating their corresponding transcripts - Kcnj2 and Kcnj4 for KIR and Kcnk2 for KLeak. Consistent with this, other investigators found that Kcnk2 and Kncj4 are up-regulated when human FOXP1 is overexpressed in mouse striatum . Previously, using ChIP-seq data from human neural progenitor cells, we have shown that FOXP1 binds to Kcnj2 and Kcnk2 promoters  (Kcnk2 regulation is unpublished). Therefore, changes in their expression may involve a direct FOXP1-mediated regulation. Taken together with previous studies, our data indicate that FOXP1-dependent regulation of these channels, either directly or indirectly, may occur at the transcriptional rather than post-transcriptional level. Finally, this involves postnatal processes, and we have made additional observations suggesting that KIR changes emerge earlier in development compared to KLeak changes (see Supplementary Fig. S11).
It is important to note that mRNA transcript and protein levels may not always have a direct correspondence [50, 51], and a recent study of KIR regulation in SPNs also suggests a lack of this correlation . However, there is evidence for RNA transcript levels of ion channels in the brain having a greater correlation with their derived protein products’ levels than transcripts of other cellular components . Nonetheless, future studies should include assessment of protein levels and localization of the potassium channels implicated in this study.
Links with alterations in the Foxp1 +/− mouse
The Foxp1+/− mouse is a good model for FOXP1-associated ASD, which is primarily due to heterozygous FOXP1 mutations. In a previous study, using Foxp1+/− mice, we observed changes that closely matched our observations with cell-autonomous homozygous deletion in D2 SPNs - increased intrinsic excitability in D2 SPNs and decreased striatal levels of two of the transcripts described above - Kcnk2 and Kcnj2 . However, in the current study, hyperexcitability was not observed with cell-autonomous heterozygous Foxp1 deletion in D2 SPNs (Supplementary Fig. S1). There are two possibilities that may explain this discrepancy: (1) hyperexcitability in Foxp1+/− mice is not a cell-autonomous process and involves alterations induced by heterozygosity in other cell types or (2) hyperexcitability is a cell-autonomous process in Foxp1+/− mice, but due to delayed deletion in D2 Foxp1cHet mice (Cre expression initiates at E14-15 ) this process is not induced. Alternatively, both possibilities could be occurring. In this scenario, because gene dosage is reduced in Foxp1+/− mice, the same cell-autonomous mechanism is only induced by additional non-cell-autonomous processes.
While we did not observe any change in the excitability of D1 SPNs in Foxp1+/− mice  or with postnatal Foxp1 deletion, these neurons were hyperexcitable in D1 Foxp1cKO mice. A few reasons may explain this discrepancy. First, FOXP1 reduction in Foxp1+/− mice is insufficient to induce hyperexcitability. Second, loss of FOXP1 function may be compensated by FOXP2 [15, 53]. Thirdly, D1 Foxp1cKO mice have significant Foxp1 deletion in the deep layers of neocortex as well, which include cell populations projecting to SPNs. Therefore, it is possible that hyperexcitability of D1 SPNs in D1 Foxp1cKO mice is not purely cell-autonomous and is influenced by the homozygous Foxp1 deletion in cortical neurons. Further experiments are required to resolve this issue.
The D2 Foxp1 cKO mouse model
Although there was a reduction in the number of D2 SPNs in the D2 Foxp1cKO (see Supplementary Methods), the changes in the excitability and subthreshold conductance were very similar to that observed in Foxp1+/− mice  and with postnatal Foxp1 deletion – models where no D2 SPN decrease was observed. Moreover, changes in the relevant KIR and KLeak transcripts were comparable to that observed in Foxp1+/− mice . These common changes among different deletion strategies indicate that our findings in the D2 Foxp1cKO mice are relevant to pathological processes occurring in Foxp1+/− mice – the model for ASD.
The functional changes that we describe may be involved in the communication deficits observed both in FOXP1-linked ASD patients and in the corresponding Foxp1+/− mouse model [2, 9]. Moreover, the specific proteins that have been identified as FOXP1 targets could be further studied as potential therapeutic targets for FOXP1-linked ASD.
Ferland RJ, Cherry TJ, Preware PO, Morrisey EE, Walsh CA. Characterization of Foxp2 and Foxp1 mRNA and protein in the developing and mature brain. J Comp Neurol. 2003;460:266–79.
Meerschaut I, Rochefort D, Revencu N, Petre J, Corsello C, Rouleau GA, et al. FOXP1-related intellectual disability syndrome: a recognisable entity. J Med Genet. 2017;54:613–23.
Siper PM, De Rubeis S, Trelles MDP, Durkin A, Di Marino D, Muratet F, et al. Prospective investigation of FOXP1 syndrome. Mol Autism. 2017;8:57.
Satterstrom FK, Kosmicki JA, Wang J, Breen MS, De Rubeis S, An JY, et al. Large-scale exome sequencing study implicates both developmental and functional changes in the neurobiology of autism. Cell. 2020;180:568–84.e23.
Iossifov I, O’Roak BJ, Sanders SJ, Ronemus M, Krumm N, Levy D, et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature. 2014;515:216–21.
Sanders SJ, He X, Willsey AJ, Ercan-Sencicek AG, Samocha KE, Cicek AE, et al. Insights into autism spectrum disorder genomic architecture and biology from 71 risk loci. Neuron. 2015;87:1215–33.
Stessman HA, Xiong B, Coe BP, Wang T, Hoekzema K, Fenckova M, et al. Targeted sequencing identifies 91 neurodevelopmental-disorder risk genes with autism and developmental-disability biases. Nat Genet. 2017;49:515–26.
Bacon C, Schneider M, Le Magueresse C, Froehlich H, Sticht C, Gluch C, et al. Brain-specific Foxp1 deletion impairs neuronal development and causes autistic-like behaviour. Mol Psychiatry. 2015;20:632–9.
Araujo DJ, Anderson AG, Berto S, Runnels W, Harper M, Ammanuel S, et al. FoxP1 orchestration of ASD-relevant signaling pathways in the striatum. Genes Dev. 2015;29:2081–96.
Scott-Van Zeeland AA, McNealy K, Wang AT, Sigman M, Bookheimer SY, Dapretto M. No neural evidence of statistical learning during exposure to artificial languages in children with autism spectrum disorders. Biol Psychiatry. 2010;68:345–51.
Di Martino A, Kelly C, Grzadzinski R, Zuo XN, Mennes M, Mairena MA, et al. Aberrant striatal functional connectivity in children with autism. Biol Psychiatry. 2011;69:847–56.
Langen M, Leemans A, Johnston P, Ecker C, Daly E, Murphy CM, et al. Fronto-striatal circuitry and inhibitory control in autism: findings from diffusion tensor imaging tractography. Cortex. 2012;48:183–93.
Radulescu E, Minati L, Ganeshan B, Harrison NA, Gray MA, Beacher FD, et al. Abnormalities in fronto-striatal connectivity within language networks relate to differences in grey-matter heterogeneity in Asperger syndrome. Neuroimage Clin. 2013;2:716–26.
Heiman M, Schaefer A, Gong S, Peterson JD, Day M, Ramsey KE, et al. A translational profiling approach for the molecular characterization of CNS cell types. Cell. 2008;135:738–48.
Anderson AG, Kulkarni A, Harper M, Konopka G. Single-cell analysis of Foxp1-driven mechanisms essential for striatal development. Cell Rep. 2020;30:3051–66.e7.
Precious SV, Kelly CM, Reddington AE, Vinh NN, Stickland RC, Pekarik V, et al. FoxP1 marks medium spiny neurons from precursors to maturity and is required for their differentiation. Exp Neurol. 2016;282:9–18.
Fong WL, Kuo HY, Wu HL, Chen SY, Liu FC. Differential and overlapping pattern of Foxp1 and Foxp2 expression in the striatum of adult mouse brain. Neuroscience. 2018;388:214–23.
Usui N, Araujo DJ, Kulkarni A, Co M, Ellegood J, Harper M, et al. Foxp1 regulation of neonatal vocalizations via cortical development. Genes Dev. 2017;31:2039–55.
Dasen JS, De Camilli A, Wang B, Tucker PW, Jessell TM. Hox repertoires for motor neuron diversity and connectivity gated by a single accessory factor, FoxP1. Cell. 2008;134:304–16.
Rousso DL, Gaber ZB, Wellik D, Morrisey EE, Novitch BG. Coordinated actions of the forkhead protein Foxp1 and Hox proteins in the columnar organization of spinal motor neurons. Neuron. 2008;59:226–40.
Palmesino E, Rousso DL, Kao TJ, Klar A, Laufer E, Uemura O, et al. Foxp1 and lhx1 coordinate motor neuron migration with axon trajectory choice by gating Reelin signalling. PLoS Biol. 2010;8:e1000446.
Li X, Han X, Tu X, Zhu D, Feng Y, Jiang T, et al. An autism-related, nonsense foxp1 mutant induces autophagy and delays radial migration of the cortical neurons. Cereb Cortex. 2019;29:3193–208.
Araujo DJ, Toriumi K, Escamilla CO, Kulkarni A, Anderson AG, Harper M, et al. Foxp1 in forebrain pyramidal neurons controls gene expression required for spatial learning and synaptic plasticity. J Neurosci. 2017;37:10917–31.
Shen W, Tian X, Day M, Ulrich S, Tkatch T, Nathanson NM, et al. Cholinergic modulation of Kir2 channels selectively elevates dendritic excitability in striatopallidal neurons. Nat Neurosci. 2007;10:1458–66.
Cazorla M, Shegda M, Ramesh B, Harrison NL, Kellendonk C. Striatal D2 receptors regulate dendritic morphology of medium spiny neurons via Kir2 channels. J Neurosci. 2012;32:2398–409.
Nisenbaum ES, Wilson CJ. Potassium currents responsible for inward and outward rectification in rat neostriatal spiny projection neurons. J Neurosci. 1995;15:4449–63.
Lieberman OJ, McGuirt AF, Mosharov EV, Pigulevskiy I, Hobson BD, Choi S, et al. Dopamine triggers the maturation of striatal spiny projection neuron excitability during a critical period. Neuron. 2018;99:540–54.e4.
Fink M, Duprat F, Lesage F, Reyes R, Romey G, Heurteaux C, et al. Cloning, functional expression and brain localization of a novel unconventional outward rectifier K+ channel. EMBO J. 1996;15:6854–62.
Talley EM, Solorzano G, Lei Q, Kim D, Bayliss DA. Cns distribution of members of the two-pore-domain (KCNK) potassium channel family. J Neurosci. 2001;21:7491–505.
Feliciangeli S, Chatelain FC, Bichet D, Lesage F. The family of K2P channels: salient structural and functional properties. J Physiol. 2015;593:2587–603.
Heurteaux C, Guy N, Laigle C, Blondeau N, Duprat F, Mazzuca M, et al. TREK-1, a K+ channel involved in neuroprotection and general anesthesia. EMBO J. 2004;23:2684–95.
Maze I, Chaudhury D, Dietz DM, Von Schimmelmann M, Kennedy PJ, Lobo MK, et al. G9a influences neuronal subtype specification in striatum. Nat Neurosci. 2014;17:533–9.
Karschin C, Dissmann E, Stuhmer W, Karschin A. IRK(1-3) and GIRK(1-4) inwardly rectifying K+ channel mRNAs are differentially expressed in the adult rat brain. J Neurosci. 1996;16:3559–70.
Lieberman OJ, Frier MD, McGuirt AF, Griffey CJ, Rafikian E, Yang M, et al. Cell-type specific regulation of neuronal intrinsic excitability by macroautophagy. Elife. 2020;9:e50843.
Abrahams, B.S., Arking, D.E., Campbell, D.B. et al. SFARI Gene 2.0: a community-driven knowledgebase for the autism spectrum disorders (ASDs). Molecular Autism. 2013;4:36.
Harrington AJ, Raissi A, Rajkovich K, Berto S, Kumar J, Molinaro G, et al. MEF2C regulates cortical inhibitory and excitatory synapses and behaviors relevant to neurodevelopmental disorders. Elife. 2016;5:e20059.
Kahanovitch U, Cuddapah VA, Pacheco NL, Holt LM, Mulkey DK, Percy AK, et al. MeCP2 Deficiency Leads to Loss of Glial Kir4.1. eNeuro. 2018;5:ENEURO.0194-17.2018.
Mermelstein PG, Song WJ, Tkatch T, Yan Z, Surmeier DJ. Inwardly rectifying potassium (IRK) currents are correlated with IRK subunit expression in rat nucleus accumbens medium spiny neurons. J Neurosci. 1998;18:6650–61.
Hagiwara S, Miyazaki S, Rosenthal NP. Potassium current and the effect of cesium on this current during anomalous rectification of the egg cell membrane of a starfish. J Gen Physiol. 1976;67:621–38.
Hibino H, Inanobe A, Furutani K, Murakami S, Findlay I, Kurachi Y. Inwardly rectifying potassium channels: their structure, function, and physiological roles. Physiol Rev. 2010;90:291–366.
Shen W, Hamilton SE, Nathanson NM, Surmeier DJ. Cholinergic suppression of KCNQ channel currents enhances excitability of striatal medium spiny neurons. J Neurosci. 2005;25:7449–58.
Coetzee WA, Amarillo Y, Chiu J, Chow A, Lau D, McCormack T, et al. Molecular diversity of K+ channels. Ann NY Acad Sci. 1999;868:233–85.
Ma XY, Yu JM, Zhang SZ, Liu XY, Wu BH, Wei XL, et al. External Ba2+ block of the two-pore domain potassium channel TREK-1 defines conformational transition in its selectivity filter. J Biol Chem. 2011;286:39813–22.
Sandoz G, Douguet D, Chatelain F, Lazdunski M, Lesage F. Extracellular acidification exerts opposite actions on TREK1 and TREK2 potassium channels via a single conserved histidine residue. Proc Natl Acad Sci USA. 2009;106:14628–33.
Mazella J, Petrault O, Lucas G, Deval E, Beraud-Dufour S, Gandin C, et al. Spadin, a sortilin-derived peptide, targeting rodent TREK-1 channels: a new concept in the antidepressant drug design. PLoS Biol. 2010;8:e1000355.
Djillani A, Pietri M, Mazella J, Heurteaux C, Borsotto M. Fighting against depression with TREK-1 blockers: past and future. A focus on spadin. Pharm Ther. 2019;194:185–98.
Djillani A, Pietri M, Moreno S, Heurteaux C, Mazella J, Borsotto M. Shortened spadin analogs display better TREK-1 inhibition, in vivo stability and antidepressant activity. Front Pharm. 2017;8:643.
Lesage F, Lazdunski M. Molecular and functional properties of two-pore-domain potassium channels. Am J Physiol Ren Physiol. 2000;279:F793–801.
Tang B, Becanovic K, Desplats PA, Spencer B, Hill AM, Connolly C, et al. Forkhead box protein p1 is a transcriptional repressor of immune signaling in the CNS: implications for transcriptional dysregulation in Huntington disease. Hum Mol Genet. 2012;21:3097–111.
Ghazalpour A, Bennett B, Petyuk VA, Orozco L, Hagopian R, Mungrue IN, et al. Comparative analysis of proteome and transcriptome variation in mouse. PLoS Genet. 2011;7:e1001393.
Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet. 2012;13:227–32.
Bauernfeind AL, Babbitt CC. The predictive nature of transcript expression levels on protein expression in adult human brain. BMC Genomics. 2017;18:322.
Vernes SC, Oliver PL, Spiteri E, Lockstone HE, Puliyadi R, Taylor JM, et al. Foxp2 regulates gene networks implicated in neurite outgrowth in the developing brain. PLoS Genet. 2011;7:e1002145.
Special thanks to Mathew Harper for technical help. Funding provided by the Simons Foundation (SFARI 573689 and 401220) and NIH (MH102603) to GK and JG.
Conflict of interest
The authors declare that they have no conflict of interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Khandelwal, N., Cavalier, S., Rybalchenko, V. et al. FOXP1 negatively regulates intrinsic excitability in D2 striatal projection neurons by promoting inwardly rectifying and leak potassium currents. Mol Psychiatry 26, 1761–1774 (2021). https://doi.org/10.1038/s41380-020-00995-x