Postnatal Conditional Deletion of Bcl11b in Striatal Projection Neurons Mimics the Transcriptional Signature of Huntington’s Disease

The dysregulation of striatal gene expression and function is linked to multiple diseases, including Huntington’s disease (HD), Parkinson’s disease, X-linked dystonia-parkinsonism (XDP), addiction, autism, and schizophrenia. Striatal medium spiny neurons (MSNs) make up 90% of the neurons in the striatum and are critical to motor control. The transcription factor, Bcl11b (also known as Ctip2), is required for striatal development, but the function of Bcl11b in adult MSNs in vivo has not been investigated. We conditionally deleted Bcl11b specifically in postnatal MSNs and performed a transcriptomic and behavioral analysis on these mice. Multiple enrichment analyses showed that the D9-Cre-Bcl11btm1.1Leid transcriptional profile was similar to the HD gene expression in mouse and human data sets. A Gene Ontology enrichment analysis linked D9-Cre-Bcl11btm1.1Leid to calcium, synapse organization, specifically including the dopaminergic synapse, protein dephosphorylation, and HDAC-signaling, commonly dysregulated pathways in HD. D9-Cre-Bcl11btm1.1Leid mice had decreased DARPP-32/Ppp1r1b in MSNs and behavioral deficits, demonstrating the dysregulation of a subtype of the dopamine D2 receptor expressing MSNs. Finally, in human HD isogenic MSNs, the mislocalization of BCL11B into nuclear aggregates points to a mechanism for BCL11B loss of function in HD. Our results suggest that BCL11B is important for the function and maintenance of mature MSNs and Bcl11b loss of function drives, in part, the transcriptomic and functional changes in HD.


Introduction
The basal ganglia comprise interconnected subcortical nuclei that are responsible for motor learning and control, executive functions, and emotions. The striatum, composed of the caudate and putamen in humans, is the largest component of the basal ganglia. It receives and integrates glutamatergic and dopaminergic inputs from several brain regions, including the cortex, thalamus, hippocampus, and amygdala. These inputs target inhibitory γ-amino butyric acid (GABA)-ergic medium spiny neurons (MSNs), the principal output neurons of the striatum, and make up 90-95% of its total neurons. MSNs are morphologically homogeneous; however, they can be distinguished by their output targets and their specific gene expression. Direct MSNs project to the global pallidus internal or Behavioral Testing of Bcl11b-Deletion Mice Locomotor activity: Spontaneous locomotor activity was measured using the Digiscan D-Micropro automated activity monitoring system (Accuscan, Inc., Columbus, OH, USA). This system consists of transparent plastic boxes (45 × 20 × 20 inch) set inside metal frames that are equipped with 16 infrared light emitters and detectors with 16 parallel infrared photocell beams. Breaks were recorded using a computer interface in 5-min bins. Mice were habituated to the testing chamber for 2 days, and on the third day, their locomotor activity was recorded for 60 min prior to returning them to their home cages.
Balance beam test: Balance was assessed by measuring the ability of mice to traverse a narrow beam as described [33,34], with brief modifications. The beam consisted of an 85-cm-long wooden prism, divided into 5-cm frames, with a 1-cm face, placed 40 cm above the bench surface. During the training session, mice were allowed to walk on the beam for 2 min. After 4 h, mice were returned to the beam, and their latency to cover 30 frames and total distance traveled were measured.
Vertical pole test: Motor coordination and balance were assessed by measuring the ability of mice to turn and descend from a narrow pole, as described in [33]. The pole consisted of a 60-cm wooden cylinder (1-cm diameter) wrapped in tape to facilitate walking. Mice were trained for 2 consecutive days and tested on the third day. Mice were placed just below the top of the pole facing upwards. Time to completely orient the body downward (time to turn) and time to climb down (time to descend) the pole were measured. An average of three test trials is shown.
Haloperidol-induced catalepsy: Mice were intraperitoneally injected with 1 mg/kg Haloperidol (Sigma Aldrich, H-030) or saline vehicle (0.9% NaCl, Teknov, Hollister, CA, USA, S5824). After 30 min, mice were gently positioned in catalepsy position, placing their forelimbs on a 0.5 cm diameter steel rod, covered with non-slippery tape, that was 5 cm above the surface of the bench. A researcher measured the time to remove both front paws from the bar (catalepsy time). Catalepsy was measured every 30 min after the first trial.
Elevated plus maze: Anxiety-related behavior was tested by an elevated plus-maze as described in [35].
Differentiation into MSNs: MSNs were prepared as described in [38]. Briefly, 96-well plates were coated with a 50 µg/mL solution of Matrigel (Corning, Corning, NY, USA, CB-40234) for 24 h. Passage 13 NSCs were plated in NPM medium at a concentration of 90,000 cells per well. To start differentiation, NPM medium was replaced with Synaptojuice A. Half-medium changes were done every other day for 7 days. On day eight of differentiation, Synaptojuice A was replaced with Synaptojuice B for 10 days with half-medium changes every other day. Both Synaptojuice A and Synaptojuice B were supplemented with 25 ng/mL of Activin A (Peprotech, 120-14P).
BCL11B and DARPP-32 immunostaining in MSNs: Cells were washed with PBS and fixed with 4% paraformaldehyde for 12 min at room temperature. Cells were incubated in block buffer containing 1% normal donkey serum and 0.1% Triton-X-100 in PBS for 1 h. Primary antibodies were diluted at 1:100 in blocking buffer, and the cells were incubated with it overnight. The cells were then washed with buffer containing 0.1% Triton-X-100 in PBS for 5 min, three times. Secondary antibodies were diluted at 1:500 in blocking buffer with 300 nM DAPI, and the cells were incubated in it for 2 h. The cells were then washed three times with 0.1% Triton-X-100 in PBS for 5 min and imaged using a Cytation 5 instrument (Biotek). The antibodies used were rabbit anti-BCL11B (Novus Biological Littleton, CO, USA, NB100-79809) paired with donkey-anti-rabbit Alexa-488 (ThermoFisher, A-21206), and mouse anti-DARPP-32 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA, sc-271111) paired with donkey-anti-mouse Alexa-647 (ThermoFisher, A-31571).
Confocal microscopy of nuclear aggregates of BCL11B: MSNs grown in plasticbottomed microplates immunostained for BCL11B were imaged using a Zeiss LSM980/ Airyscan2 laser scanning confocal microscope, using an LD LCI Plan-Apochromat 40×/ 1.2 Imm Korr objective lens with glycerol immersion, and Airyscan super resolution mode (41nm/pixel resolution). Images were analyzed in Image Analyst MKII (Version 4.1.3, Image Analyst Software, Novato, CA, USA) using the "Nuclear foci area measurement" standard pipeline, providing counts of BCL11B foci per nucleus, and the area of each nucleus based on the DAPI staining.
Although many markers specifically associated with MSNs were downregulated, there were some notable exceptions ( Figure 1D,E). One of the top upregulated genes was latent transforming factor beta binding protein (Ltbp2). It is critical in the TGFβ signaling and regulation of this pathway that it has been linked to MSN developmental processes and HD neuropathogenesis [1]. Correspondingly, activin A receptor like type 1 (Acvrl1), is upregulated. During the development of the lateral ganglionic eminence (LGE), the ligand for Acvr1, Activin A, plays a critical role in the specification of striatal fate [39,40], and both activin receptors and activated activin are expressed in the developing LGE [39,40]. Wnt8b is involved in the caudalization of a regional identity [41]. Upon the reduction of Bcl11b, ras-specific guanine-nucleotide releasing factor 2 (RasGRF2) was also reduced. This calciumregulated exchange factor [42] alters the ERK-dependent cocaine reward in mice [43]. Some of the top upregulated genes are involved in cell death signaling pathways (e.g., Clec12a, a uric acid receptor that potentiates type I interferon responses) [44]. Sstr3 is a G-protein-coupled receptor (GPCR) whose signaling affects neuronal cilia and apoptosis and is upregulated after heroin exposure [45,46]. In addition, glutamate metabotropic receptor 2 (Grm2) is increased in the Bcl11b MSN-deletion mice. A group of top dysregulated genes are not known to be specifically associated with MSNs and therefore warrant further investigation ( Figure 1D,E, Table S1). Although many markers specifically associated with MSNs were downregulated, there were some notable exceptions ( Figure 1D,E). One of the top upregulated genes was and Cre-negative mice (4-months-old) were analyzed using immunohistochemistry (IHC) with a BCL11B antibody. Scale bar is 500 mm. (B) PCA plots using the rlog-transformed values indicate a significant difference in the transcriptome Bcl11b deleting MSNs, and controls. (C) Scatter plot shows that Bcl11b gene expression is much less in D9-Cre-Bcl11b tm1.1Leid mice than Crecontrol mice.
We conclude that the reduction in Bcl11b affects multiple genes involved in MSN maintenance and identity and signal transduction.
Bcl11b reduction results in differentially expressed genes that correlate with pathways dysregulated in HD. We used Enrichr [47] to identify the co-expression networks that most overlap with the transcriptomics of the Bcl11b reduction in MSNs ( Figure 2). Strikingly, the expression changes for the D9-Cre-Bcl11b tm1.1Leid mice overlapped with expression profiles of HD mouse models, postmortem HD tissue and/or knockout mice (KO), including the genes Pde10a, Sirt1, Htra2, Npc1, and Ppargc1a (Figure 2A,B). Many of the overlapping genes are markers of MSNs, and the loss of striatal MSN identity overlaps with the D9-Cre-Bcl11b tm1.1Leid transcriptomics. The D9-Cre-Bcl11b tm1.1Leid mice transcriptional profile had drug perturbations from the GEO database that overlapped with soman, morphine, resveratrol, heroin, dexamethasone, coenzyme Q, creatine, levetiracetam, methamphetamine, and nicotinamide riboside ( Figure 2C). These drug perturbations correlate with striatal function or known drug targets in HD. We also evaluated the overlap of the gene expression profiles of the 10-month-old Q175 knockin mouse model [21] and the conditional D9-Cre-Bcl11b tm1.1Leid mice. There were 683 genes shared when comparing the transcriptomic data sets (zQ175, 2795 genes) with a p-value of 3.75E-76, when using the Fisher exact test. The top KEGG pathway (2021 human) is the dopaminergic synapse, and the protein-protein interaction hub protein is GRIN1. The shared genes have ontologies for the regulation of neurotransmitter receptor activity, calcium signaling, potassium channel regulation, protein/threonine kinase activity, activin receptor activity, glutamate receptor activity, and postsynaptic density. Kinase regulation includes CAMK4 and the regulation of a glutamate receptor by CK1 and CDK5. As expected, our data enriches to constitutively deleted Bcl11b mice. Table S2 summarizes the overlap with the majority of known mouse HD transcriptomics data sets and the overlap with the D9-Cre-Bcl11b tm1.1Leid mice.
Next, we carried out a term enrichment analysis for GO or KEGG processes or functions associated with the DEGs for the conditional D9-Cre-Bcl11b tm1.1Leid mice, compared to the controls ( Figure 3, Table S1, Supplemental Figure S1). The KEGG term enrichment analysis for gene signatures altered by D9-Cre-Bcl11b tm1.1Leid highlighted the axon guidance, dopaminergic synapses, adrenergic, estrogen, cAMP, MAPK, insulin, oocytes, and glutamatergic signaling ( Figure 3A). An IPA analysis summarizes the critical pathway for dopamine DARPP-32 feedback cAMP signaling that is enriched in the D9-Cre-Bcl11b tm1.1Leid mice ( Figure 3B). This is a key pathway disrupted in HD. Interestingly, many of the genes with a log fold-change >1.0 correlated with genes involved in calcium homeostasis (Table S1, Supplemental Figure S2). The top canonical pathways identified by a complementary analysis with an Ingenuity Pathway Analysis (IPA) were the opioid signaling pathway (p-value 6.22E-14), cAMPmediated signaling (p-value 3.60E-08) via which dopamine signals are transduced, the synaptogenesis signaling pathway (p-value 4.07E-8), protein kinase A signaling (p-value 4.11 × 10 −8 ), and calcium signaling (p-value 1.10E-07). The top upstream regulators were levodopa, CREB1, amino-5-phosphonovaleric acid, and huntingtin (HTT).
Next, we carried out a term enrichment analysis for GO or KEGG processes or functions associated with the DEGs for the conditional D9-Cre-Bcl11b tm1.1Leid mice, compared to the controls (Figure 3, Table S1, Supplemental Figure S1). The KEGG term enrichment analysis for gene signatures altered by D9-Cre-Bcl11b tm1.1Leid highlighted the axon guidance, dopaminergic synapses, adrenergic, estrogen, cAMP, MAPK, insulin, oocytes, and glutamatergic signaling ( Figure 3A). An IPA analysis summarizes the critical pathway for dopamine DARPP-32 feedback cAMP signaling that is enriched in the D9-Cre-Bcl11b tm1.1Leid mice ( Figure 3B). This is a key pathway disrupted in HD. Interestingly, many of the genes with a log fold-change >1.0 correlated with genes involved in calcium homeostasis (Table S1, Supplemental Figure S2). Enriched BPs included synapse organization, functions, or molecules related with transmembrane activities, such as a transmembrane transporter or ion channels (Supplemental Figure S1A). Correspondingly, the GO terms enriched for BPs in the downregulated genes were protein localization, dephosphorylation, the rhythmic process, and the postsynapse. The cellular components and molecular functions are also shown in Supplemental Figure S1B-G. Top upstream terms and network from the IPA analysis were HTT, NR4A1, CNTF, epilepsy, dyskinesia, synaptic depression, the organization of cells, and catalepsy (Supplemental Figure S3). The identification of NR4A1 as a top upstream regulator is interesting. Although Bcl11b appeared to regulate gene expression in both striosomes and in the matrix in our current study, Nr4a1 is highly enriched in the striosomes and is required for their development [3,31], and striosomes are altered in HD [48][49][50].
Thus, a decrease in Bcl11b alters the general neuronal and MSN-specific processes, including in synapse organization and functions, or in molecules related with transmembrane activities (e.g., transmembrane transporter and ion channels). Enriched BPs included synapse organization, functions, or molecules related with transmembrane activities, such as a transmembrane transporter or ion channels (Supplemental Figure S1A). Correspondingly, the GO terms enriched for BPs in the downregulated genes were protein localization, dephosphorylation, the rhythmic process, and the postsynapse. The cellular components and molecular functions are also shown in Supplemental Figure S1B-G. Top upstream terms and network from the IPA analysis were HTT, NR4A1, CNTF, epilepsy, dyskinesia, synaptic depression, the organization of cells, and catalepsy (Supplemental Figure S3). The identification of NR4A1 as a top upstream regulator is interesting. Although Bcl11b appeared to regulate gene expression in both striosomes and in the matrix in our current study, Nr4a1 is highly enriched in the striosomes and is required for their development [3,31], and striosomes are altered in HD [48][49][50]. Transcriptional network impacted by Bcl11b deletion: We evaluated the TFs and networks impacted by the deletion of Bcl11b. Using the mouse TF database, a total of 287 differentially expressed TFs were altered [51]. Among these, 109 were upregulated, and 178 were downregulated (Table S1). These differentially expressed TFs were used as the input for the gene regulatory network analysis to determine the key upstream regulators in the D9-Cre-Bcl11b tm1.1Leid MSNs (Figure 4). From the inferred networks, we identified the hub gene, Egr1 ( Figure 4A), which plays a key role in the induction of DARPP-32 expression in MSNs [52]. The enrichment of Foxo3, Foxj2, and Foxj3 showed that the decrease of Bcl11b alters the forkhead pathway ( Figure 4B), which is important in adult human neurogenesis and cell-cycle inhibition [53,54].
The decrease of Bcl11b in differentiated MSNs resulted in a decrease in NeuN+/DARPP-32+ cells, motor deficits, and a decreased response to haloperidol. Selective loss of striatal MSNs is a major hallmark in HD but is poorly recapitulated in mouse models [63]. To determine if Bcl11b deletion compromises neuronal viability, we counted the striatal neurons, and specifically MSNs, using a NeuN and DARPP-32 immunofluorescence. Fewer NeuN+ and DARPP-32+ cells were detected in the striatum of D9-Cre-Bcl11b tm1.1Leid mice than in the wildtype (WT) mice ( Figure 5A,B). The decreased neuronal numbers were not accompanied by increased numbers of microglia, as revealed by Iba1 immunostaining ( Figure 5C). The decrease in NeuN+ and DARPP-32+ cells may suggest a loss of neurons, de-differentiation, or a lack of differentiation.
in the first 5 min ( Figure 6A). There was a trend towards subtle balance alterations in the balance beam test, in that D9-Cre-Bcl11b tm1.1Leid mice and WT mice crossed a similar number of frames, but the KO mice appeared to require more time to cross 30 frames than WT mice (p-value = 0.096; t-test) ( Figure 6B). Importantly, D9-Cre-Bcl11b tm1.1Leid mice displayed poor performance in the vertical pole test, requiring more time to turn and descend than WT mice (unpaired t-test, * p < 0.05; ** p < 0.01) ( Figure 6C). Bcl11b deletion did not alter anxiety-like behavior in the elevated plus maze (Supplemental Figure S4). These results suggest that motor abnormalities after Bcl11b deletion in adult MSNs overlap to some extent with HD mouse models.  The role of the striatum in movement and the overlap of the gene expression changes with HD prompted us to evaluate the motor behavior of striatal D9-Cre-Bcl11b tm1.1Leid mice. We found that spontaneous locomotor activity was reduced in D9-Cre-Bcl11b tm1.1Leid mice in the first 5 min ( Figure 6A). There was a trend towards subtle balance alterations in the balance beam test, in that D9-Cre-Bcl11b tm1.1Leid mice and WT mice crossed a similar number of frames, but the KO mice appeared to require more time to cross 30 frames than WT mice (p-value = 0.096; t-test) ( Figure 6B). Importantly, D9-Cre-Bcl11b tm1.1Leid mice displayed poor performance in the vertical pole test, requiring more time to turn and descend than WT mice (unpaired t-test, * p < 0.05; ** p < 0.01) ( Figure 6C). Bcl11b deletion did not alter anxiety-like behavior in the elevated plus maze (Supplemental Figure S4). These results suggest that motor abnormalities after Bcl11b deletion in adult MSNs overlap to some extent with HD mouse models.
Next, RNA-seq analysis pointed to specific alterations of genes involved in the dopaminergic synapse pathway (adjusted p-value = 3.27E10−7). A cell-type-specific enrichment analysis showed that Bcl11b deletion caused a downregulation of genes that were enriched in both D1 and D2-MSN subtypes (Supplemental Figure S4). Notably, Drd2 (log2fold = −0.60) was reduced more than Drd1 (log2fold = −0.26). To functionally explore D2R-mediated behavior, we performed the haloperidol-induced catalepsy test. Haloperidol Next, RNA-seq analysis pointed to specific alterations of genes involved in the dopaminergic synapse pathway (adjusted p-value = 3.27E10−7). A cell-type-specific enrichment analysis showed that Bcl11b deletion caused a downregulation of genes that were enriched in both D1 and D2-MSN subtypes (Supplemental Figure S4). Notably, Drd2 (log2fold = −0.60) was reduced more than Drd1 (log2fold = −0.26). To functionally explore D2R-mediated behavior, we performed the haloperidol-induced catalepsy test. Haloperidol treatment in mice produces a behavioral state in which the mice fail to correct externally imposed postures (i.e., catalepsy). Integrity of postsynaptic dopamine receptors is required to observe this phenotype [64]. Haloperidol (1 mg/kg) was injected intraperitoneally into WT and D9-Cre Bcl11b-deletion mice. Catalepsy was measured 30 min after the injection and every 30 min up to 2 h. Catalepsy time was lower in Bcl11b-deletion mice than WT littermates ( Figure 6D, two-way ANOVA, with Bonferroni post-hoc test, * p < 0.05).

Disruption of BCL11B function in a human HD MSN model.
The strong correlation of the D9-Cre-Bcl11b tm1.1Leid mice transcriptome with HD models prompted us to determine how the HTT mutation mimics lower levels of Bcl11b. We differentiated isogenic human patient HD72 iPSCs (CAG repeat size 72) into MSNs ( Figure 7A). As expected, the HD72-MSNs had lower levels of DARPP-32 than isogenic control C116-MSNs. Top genes dysregulated in D9-Cre-Bcl11b tm1.1Leid mice follow similar trends in expression as measured by RT-PCR in human HD72-MSNs ( Figure 7B). KCNC3 and WNT10A were upregulated in HD72-MSNs, compared to control C116. Like the D9-Cre-Bcl11b tm1.1Leid mice transcriptomics, SLIT3 was downregulated in HD72-MSNs. BCL11B was modestly upregulated but not statistically significant (data not shown), and may represent the fact that the iPSC-derived MSNs are relatively immature, compared to mouse adult MSNs in vivo. As our current studies show a loss of Bcl11b in MSNs correlates with the HD transcriptome, we investigated the mechanism for how this might occur in HD. BCL11B expression was characterized by immunofluorescence in HD72-MSNs compared to control. In HD72-MSNs, BCL11B was concentrated in internuclear aggregates, but showed a diffuse pattern in C116-MSNs ( Figure 7C). Many more internuclear aggregates were noted in the HD72-MSNs than in controls ( Figure 7D). We conclude that the sequestration of BCL11B into nuclear aggregates may lead to loss of transcriptional activity of BCL11B in HD even in the presence of normal level of expression. treatment in mice produces a behavioral state in which the mice fail to correct externally imposed postures (i.e., catalepsy). Integrity of postsynaptic dopamine receptors is required to observe this phenotype [64]. Haloperidol (1 mg/kg) was injected intraperitoneally into WT and D9-Cre Bcl11b-deletion mice. Catalepsy was measured 30 min after the injection and every 30 min up to 2 h. Catalepsy time was lower in Bcl11b-deletion mice than WT littermates ( Figure 6D, two-way ANOVA, with Bonferroni post-hoc test, * p < 0.05).
Disruption of BCL11B function in a human HD MSN model. The strong correlation of the D9-Cre-Bcl11b tm1.1Leid mice transcriptome with HD models prompted us to determine how the HTT mutation mimics lower levels of Bcl11b. We differentiated isogenic human patient HD72 iPSCs (CAG repeat size 72) into MSNs ( Figure 7A). As expected, the HD72-MSNs had lower levels of DARPP-32 than isogenic control C116-MSNs. Top genes dysregulated in D9-Cre-Bcl11b tm1.1Leid mice follow similar trends in expression as measured by RT-PCR in human HD72-MSNs ( Figure 7B). KCNC3 and WNT10A were upregulated in HD72-MSNs, compared to control C116. Like the D9-Cre-Bcl11b tm1.1Leid mice transcriptomics, SLIT3 was downregulated in HD72-MSNs. BCL11B was modestly upregulated but not statistically significant (data not shown), and may represent the fact that the iPSC-derived MSNs are relatively immature, compared to mouse adult MSNs in vivo. As our current studies show a loss of Bcl11b in MSNs correlates with the HD transcriptome, we investigated the mechanism for how this might occur in HD. BCL11B expression was characterized by immunofluorescence in HD72-MSNs compared to control. In HD72-MSNs, BCL11B was concentrated in internuclear aggregates, but showed a diffuse pattern in C116-MSNs ( Figure 7C). Many more internuclear aggregates were noted in the HD72-MSNs than in controls ( Figure 7D). We conclude that the sequestration of BCL11B into nuclear aggregates may lead to loss of transcriptional activity of BCL11B in HD even in the presence of normal level of expression.

Discussion
We report that selective Cre-mediated deletion of the transcription factor Blc11b/Ctip2 in differentiated striatal MSNs leads to a transcriptional signature similar to HD and supports the notion that Bcl11b has a critical role in maintaining key pathways in the biological function and identity of MSNs in adult mice. Reduction in Bcl11b results in the lower expression of MSN differentiation markers, including FoxP1, DARPP-32 (also known as Ppp1r1b), Arpp21, Penk, Htr1b, Htr1D, Ryr1, Gabrd, Gabra4, Hrh3, Drd1, Drd2 and Grm1. Previous studies have shown that loss of Bcl11b during development results in deficits in MSN birth, migration and differentiation [15]. Our results, therefore, are consistent with a continued role of Bcl11b in MSN differentiation and/or maintenance of identity in adult mice.
A cell-type-specific enrichment analysis showed Bcl11b deletion caused a downregulation of genes that were enriched in both D1 and D2-MSN subtypes. Notably, Drd2 was reduced more than Drd1. We also found alterations in MSN gene expression in both the patch/striosome and matrix compartments of the striatum, without enrichment for either compartment. This includes the striosome markers Oprm1, Tac1, Spon1, Lydp1, Kremmen1, Tshz1, and Pdyn for patch and the matrix marker, Epha4. We functionally validated that the gene expression changes were large enough to compromise dopamine neurotransmission, as evidenced by an abnormal haloperidol-induced catalepsy test.
The expression changes for the D9-Cre-Bcl11b tm1.1Leid mice overlapped with the gene expression profiles of HD mouse models and the postmortem HD tissue. Interestingly, some gene expression changes that overlapped with D9-Cre-Bcl11b tm1.1Leid mice are genes involved in HD pathophysiology including Pde10a, Sirt1, Htra2, Npc1, and Ppargc1a. Transcriptional dysregulation has long been described as an important pathological change in HD. Many of the downregulated genes in HD striatum are enriched for genes that define MSN identity and function [21,[65][66][67][68][69][70]. Further, as in HD, genes whose expression are altered after the depletion of Bcl11b in MSNs were enriched in calcium and HDAC signaling. The mechanism behind mutant HTT-induced transcriptional effects is unclear. Our studies using human HD-MSNs suggests that the sequestration of BCL11B into nuclear aggregates may lead to loss of function of BCL11B in HD and the loss of MSN identity and function. This is consistent with a physical interaction of BCL11B with mHTT [19] and altered levels in HD mouse models [22].
Our results highlight a cascade of TFs that are impacted when Bcl11b is deleted in striatal MSNs. As discussed above, we identified that Egr1, required for DARPP-32 expression, is a hub gene [52]. Th enrichment of Foxo3, Foxj2, and Foxj3 showed that the deficiency of Bcl11b alters the forkhead pathway which is important in adult human neurogenesis and cell-cycle inhibition [53,54]. Stat1/3 are enriched in the differentially expressed TFs that are upregulated in the case of Bcl11b deletion MSNs. We recently identified that Stat1/3 is a TF that is required for striosome development [31]. Our results suggest that this may be an important pathway in adult MSNs as well for striosome identity maintenance.
Recent studies have used CRISPR/Cas9 to deplete human embryonic stem cells of BCL11B. The reduction of BCL11B in human MSNs leads to neuronal vulnerability and dysfunction. In the human model of MSNs where BCL11B is depleted, cAMP-Ca 2+ signaling, which integrates the PKA pathway, was identified as dysregulated [20]. These same pathways were identified in the current study. BCL11B knockdown likely leads to common alterations in signaling in both mice and human models.

Conclusions
In conclusion, dissecting the role of Bcl11b in adult striatal MSNs has provided valuable information on its function as well as supporting its role in basal ganglia diseases, such as HD. The postnatal deletion of Bcl11b in MSNs mimics aspects of the phenotype identified in genetic HD mouse models. However, the D9-Cre-Bcl11b tm1.1Leid mouse does not constitute an actual model of human HD, and certainly there are many other, multicellular contribu-tions to the HD phenotype. Finally, Bcl11b has a critical role in the maintenance of mature MSN phenotype and function, with a very distinct overlap with the HD transcriptome, via which its decrease may contribute to HD pathogenesis.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/biomedicines10102377/s1. Figure S1: GO enrichment of MSN Bcl11b deficiency; Figure S2: Calcium signaling pathways enriched in Bcl11b deficiency; Figure S3: IPA analysis of MSN mouse Bcl11b deficiency; Figure S4: Bcl11b deficiency does not induce anxiety-like behaviors. FigureS5: Cell-type enrichment analysis.; Table S1: Transcriptomics of Bcl11b deficiency with functional analysis; Table S2:HD mouse transcriptomics overlaps with Bcl11b data set. Funding: This work was supported by the National Institutes of Health awards: R01-NS100529 (Ellerby, Ehrlich). Support was also provided by "The Taube Family Program in Regenerative Medicine Genome Editing for Huntington's Disease" to LME. KTT was supported by a fellowship from the Collaborative Center for X-linked Dystonia Parkinsonism (CCXDP).
Institutional Review Board Statement: Experimental procedures were carried out according to the Institutional Animal and Care and Use Committee at Icahn School of Medicine at Mount Sinai (LA09-00272, 16-0847 PRYR1).

Informed Consent Statement: Not applicable.
Data Availability Statement: The deposited raw data for the transcriptomics are GSE185476.