NPAS1 Represses the Generation of Specific Subtypes of Cortical Interneurons

NPAS1 and NPAS3 Expression During Interneuron Development

The subpallium generates neocortical interneurons (Flandin et al., 2011; Marin, 2012; Rudy et al., 2011). We examined NPAS1 and NPAS3 RNA expression by in situ hybridization (ISH) at E13.5, E15.5 and P5 (Figure 1), and assessed NPAS1 and NPAS3 expression by immunofluorescence at P0, P5, P15, and P30 (Figure 1 and Figures S1 and S2 and Tables S1 and S2). At E13.5, both had pallial and subpallial ventricular zone (VZ) expression. NPAS1 showed notable expression in the VZ and subventricular zone (SVZ) of the dorsal and ventral MGE, and CGE. By E15.5, NPAS3 was expressed in the MGE mantle zone; NPAS1 expression was prominent in the pallial and subpallial SVZ.

Previous studies have described co-expression of NPAS1 or NPAS3 with cortical interneurons using GABA, GAD-67, or calretinin antibodies in the adult mouse brain (Erbel-Sieler et al., 2004). We have extended co-expression analysis of NPAS1 or NPAS3 with various interneuron markers during cortical interneuron development and in the adult (Figure 1 and Figures S1 and S2 and Tables S1 and S2). At P0, NPAS1 was expressed in neocortical interneurons; ~100% of NPAS1⁺ cells express GAD67-GFP; ~30% had MGE-like properties (Lhx6-GFP⁺) (Figures 1D and 1F and Figures S1A and S1B). By P5, NPAS1 and NPAS3 were expressed in rostro-caudal gradients in neocortical interneurons; we are unaware of other TFs with this property. Virtually all neocortical NPAS1⁺ cells (99 ± 0.29%) and the majority of NPAS3⁺ cells (67 ± 2.94%) expressed GAD67-GFP at P5 (Figures S1C and S1D). By P15, NPAS1 and NPAS3 were expressed by a majority of reelin⁺ (NPAS1: 68 ± 2.78%; NPAS3: 79 ± 4.79%) and SST⁺ (NPAS1: 65 ± 1.95%; NPAS3: 75 ± 0.44%) interneurons. Both NPAS1 and NPAS3 were expressed in a small proportion of PV⁺ cells (NPAS1: 6 ± 0.85%) (NPAS3: 13 ± 1.29%) (Figures S1E, S1F, S1H, S1I, S1J, and S1L). At P30, NPAS1⁺ cells co-expressed reelin, SST, calretinin (CR), or neuropeptide Y (NPY), but rarely co-expressed PV (reelin: 42 ± 1.94%; SST: 36 ± 3.72%; CR: 28 ± 2.89%; NPY: 12 ± 0.70%; PV: 5 ± 1.24%). On the other hand, NPAS3⁺ was expressed in a large fraction of all interneuron subtypes assayed, including PV (reelin: 74 ± 2.31%; SST: 75 ± 3.52%; CR: 51 ± 1.23%; PV: 43 ± 1.40%) (Figures S2B–S2E and S2G–S2J and data not shown).

Increased Numbers of Neocortical Interneurons in NPAS1^−/− Mutants

We studied the effect of an NPAS1 null allele (NPAS1^−/−) (Erbel-Sieler et al., 2004) on neocortical interneuron development using a GAD67-GFP allele to label all of the interneurons (Tamamaki et al., 2003). By E15.5 there was an increased density of GFP⁺ interneurons through the intermediate zone and then throughout the cortical wall at E17.5 and P0 (26–41%; E15.5: 26 ± 4.67%, p = 0.00099; E17.5: 35 ± 2.08%, p = 9.08E-8; P0: 41 ± 4.46%, p = 2.48E-6) (Figures 2A–2C and 2E–2G). Even though there was ~2-fold increased interneuron cell death at P7 (activated caspase-3, Figures S3A and S3B), P30 mice maintained ~15% (15 ± 5.43%, p = 0.044) more interneurons (Figures 2D and 2H). Surprisingly, while SST⁺, VIP⁺, NPY⁺ and reelin⁺ interneuron subtypes were increased (28–44%; SST: 32 ± 5.07%, p = 0.001; VIP: 44 ± 7.99%, p = 0.014; NPY: 34 ± 7.42%, p = 0.0019; reelin: 28 ± 3.01%, p = 2.7E-5), PV⁺ interneuron density was normal (Figures 3A–3E, 3A′–3E′, and 3F–3J). MRI quantification showed that P30 cortical volume was increased 10% (10 ± 1.40%, p = 0.0055) (Figures S3F and S3G and Table S3). Nissl section analysis supported the MRI findings, and showed 11% increased (11 ± 3.13%, p = 0.03) rostral neocortical width (Figures S2A and S2B). Assessment of NeuN⁺ (neuronal marker) cells in the P21 somatosensory cortex suggests that the enhancement in cortical volume may be the result of increased neuron numbers. In NPAS1^−/− mutants, the density of NeuN⁺ cells was increased by 11% (cortex: 11 ± 2.21%, p = 0.0011; layer I: 22 ± 7.34%, p = 0.030; layer IV: 7 ± 2.09%, p = 0.025; layer V: 16 ± 3.32%, p = 0.0024; layer VI: 13 ± 2.66%, p = 0.0052) (Figures S4A and S4B).

Increased Synaptic Inhibition onto Neocortical Pyramidal Neurons in NPAS1^−/−

To test whether the increase in cortical interneurons altered the physiology of the NPAS1^−/− cortex we performed patch-clamp recordings of layer II/III somatosensory pyramidal neurons in slices from P21-30 NPAS1^−/− mice and wildtype (WT) littermates (Figures 4A–4C). Compared to WT littermates, pyramidal neurons from NPAS1^−/− mice showed an increase in the frequency of spontaneous IPSCs (sIPSCs) (Wildtype: 8.17 ± 1.29 Hz; NPAS1^−/−: 11.18 ± 1.26 Hz; p = 0.042). However, the amplitudes and kinetics of inhibitory events were not significantly changed (Figures 4A and 4B). To further characterize the increase in GABA-mediated synaptic inhibition, tetrodotoxin (TTX) was added to the ACSF to isolate miniature IPSCs (mIPSCs). Similar to spontaneous events, an increase in mIPSC frequencies was observed in NPAS1^−/− mice (Wildtype: 6.2 ± 1.24 Hz; NPAS1^−/−: 8.74 ± 1.02 Hz; p = 0.036); amplitudes and kinetics of mIPSCs were not significantly changed (Figure 4C).

Next, we evaluated the ratio of interneuron numbers to total neuron numbers by quantifying the percentage of NeuN⁺ cells (total neurons) that express GAD67-GFP in each layer of P21 somatosensory cortex (Figures S4B and S4C). At P21, all NeuN⁺ cells were GAD67-GFP⁺ in layer I somatosensory cortex. NPAS1^−/− mutant mice displayed a 19% increase (19 ± 5.94%, p = 0.025) in GAD67-GFP⁺ interneurons within layer I (Figure S4C). In layer II/III, the percentage of NeuN⁺ cells that express GAD67-GFP was increased by 18% in NPAS1^−/− mutants (NPAS1^+/+: 27 ± 1.13%; NPAS1^−/−: 32 ± 0.96%, p = 0.0024) (Figures 4D–4I). We also estimated the excitatory to inhibitory cell ratio (E/I) in each cortical layer at P21. The E/I ratio, determined as the ratio of NeuN⁺ and GAD67⁻ cells to GAD67⁺ cells, was decreased by 23% in layer II/III of the NPAS1^−/− somatosensory cortex (NPAS1^+/+: 2.74 ± 0.17; NPAS1^−/−: 2.12 ± 0.091, p = 0.0040) (Figure S4D). No significant change in the E/I cell ratio was found in deep cortical layers (IV, V, or VI) (Figure S4D). These findings were supported by stereological quantification of NeuN⁺ and GAD67⁺ cells within the cerebral cortex. The NPAS1^−/− mutant cerebral cortex exhibited a 35% increase (35 ± 5.96%, p = 0.0053) in NeuN⁺ cells and a 57% increase (57 ± 10.49%, p = 0.0012) in GAD67⁺ cells at 3 months of age (Figures S4E–S4G). Cortical pyramidal cell density was unchanged in P21 NPAS1^−/− mutants as indicated by Cux1⁺, Ctip2⁺, and Tbr1⁺ immunofluorescence staining which labels layer II/III, layer V, and layer VI, respectively (Figures S4H–S4K). Thus, the enhanced level of synaptic inhibition onto layer II/III neocortical pyramidal neurons in the NPAS1^−/− mutant was associated with an increased fraction of inhibitory neurons.

NPAS1^−/− and NPAS3^−/− Had Opposite Effects on Cortical Interneuron Numbers

NPAS3^−/− mice display reduced proliferation and size of the adult hippocampal dentate gyrus (Pieper et al., 2005). Given that NPAS1^−/− mutants have increased cortical volume (Figures S3F and S3G) and interneuron numbers (Figures 3A–3C, 3E, 3A′–3C′, 3E′, 3F–3H, and 3J), we hypothesized that NPAS1 and NPAS3 may have opposing functions on interneuron development. Therefore, we compared interneuron numbers in NPAS1^−/−, NPAS3^−/− and NPAS1/3^−/− P30 neocortices. While NPAS1^−/− mutants had increased density of cortical interneurons (Figures 3A–3C, 3E, 3A′–3C′, 3E′, 3F–3H, and 3J), NPAS3^−/− mutants had reduced density of SST⁺, VIP⁺, NPY⁺ and reelin⁺ interneurons (Figures 3A–3C, 3E, 3A″–C″, 3E″, 3F–3H, and 3J). Like NPAS1^−/− mutants, there was not a change in the density of PV⁺ interneurons (Figures 3D––3D″ and 3I).

MRI quantification revealed that P30 cortical volume was increased in NPAS1^−/− mutant mice, as opposed to a 21% decrease (21 ± 1.47%, p = 0.0039) in NPAS3^−/− mutants. Also, we found differential changes in P30 NPAS1^−/− and NPAS3^−/− mutant basal ganglia volumes by MRI. NPAS1^−/− mutant mice displayed a 12% increase (12% ± 1.74%, p = 0.042) and NPAS3^−/− mutants a 25% decrease (25% ± 2.24%, p = 0.0087) in basal ganglia volume (Figures S3F–S3H and Table S3).

NPAS1/3^−/− and NPAS3^−/− mutants had similar interneuron phenotypes (Figures 3A–E, 3A″–3E″, 3A‴–3E‴, and 3F–3J). In addition, we observed probable NPAS1^−/−, NPAS3^−/−, and NPAS1/3^−/− mutant basal ganglia phenotypes (N=1), which include subtle cholinergic defects (NPAS1 and NPAS3 are expressed in the pallidum) (Figure 1 and data not shown) (Flandin et al., 2010; Nobrega-Pereira et al., 2010). The distribution of choline acetyltransferase (ChAT)⁺ cells around the globus pallidus (GP) was abnormal in NPAS1/3^−/− and NPAS3^−/− mutants. We detected increased ChAT⁺ cells inside the GP and reduced numbers in the ventral pallidum. While parvalbumin (PV)⁺ expression in the GP was not grossly abnormal in the mutants, the GP may be enlarged in the NPAS1^−/− mutant, and smaller in the NPAS3^−/− mutant. PV⁺ interneuron numbers in the striatum and basolateral nucleus of the amygdala appeared grossly normal in all three mutant genotypes (data not shown).

NPAS1^−/− and NPAS3^−/− Ganglionic Eminences Had Opposite Proliferation and ERK-Signaling Phenotypes

Given that NPAS1 and NPAS3 are expressed in MGE and CGE progenitors (Figures 1A and 1B), we explored whether changes in interneuron numbers in NPAS1^−/− and NPAS3^−/− mutants were due to changes in proliferation using a marker of M-phase cells (phospho-histone H3, PH3). We found that NPAS1^−/− mutants had a 1.6-fold increase (1.61 ± 0.13, p = 0.0007) in PH3⁺ cells in the VZ, and a 1.3-fold increase (1.31 ± 0.039, p = 0.0033) in the SVZ of the CGE at E13.5; at E15.5 NPAS1^−/− mutants had a 1.3-fold increase (VZ: 1.32 ± 0.042, p = 1.09E-5; SVZ: 1.30 ± 0.035, p = 0.022) in PH3⁺ cells in the VZ and SVZ of the MGE (Figures 5A, 5B, 5E, 5F, 5I, and 5K). In contrast, NPAS3^−/− progenitors displayed reduced proliferation in the E13.5 MGE, including a 1.3-fold decrease (1.26 ± 0.043, p = 0.010) in PH3⁺ cells in the VZ and a 1.4-fold decrease (1.44 ± 0.031, p = 6.54E-5) in the SVZ (Figures 5Q, 5R, and 5W).

BrdU labeling analysis of E13.5 MGE and CGE progenitors, utilized to detect cells in S-phase, supported the PH3 results. 1 hour BrdU pulse-labeling showed increased proliferation in the SVZ of the NPAS1^−/− mutant MGE (total number: 31 ± 1.26%, p = 0.0007; density: 48 ± 2.15%, p = 0.0004) and CGE (total number: 19 ± 3.31%, p = 0.039; density: 34 ± 5.24%, p = 0.022), while NPAS3^−/− mutant mice exhibited decreased proliferation in the MGE SVZ (total number: 26 ± 2.05%, p = 0.0018; density: 21 ± 3.08%, p = 0.023) (Figures S5A–S5E). In addition, examination of proliferation in the hippocampal dentate gyrus following intraperitoneal injection of BrdU once daily for 12 days revealed a 31% increase (31 ± 7.89%, p = 0.003) in BrdU⁺ cells in 5–10 week old NPAS1^−/− mice (Figures S5F and S5G).

The NPAS Drosophila homolog Trachealess regulates FGF signaling (Ohshiro and Saigo, 1997); thus we examined activation of the MAP kinase pathway by measuring the level of phospho-ERK immunofluorescence. NPAS1 mutants had a 1.3-fold increase (E13.5 CGE: 1.32 ± 0.043, p = 0.0016; E15.5 MGE: 1.35 ± 0.063, p = 0.00017) in phospho-ERK expression in the E13.5 CGE and E15.5 MGE, whereas NPAS3 mutants exhibited a 1.3-fold decrease (1.26 ± 0.060, p = 0.0079) in phospho-ERK expression in the E13.5 CGE and a 1.2-fold decrease (1.17 ± 0.024, p = 0.023) in the E13.5 MGE (Figures 5C, 5D, 5G, 5H, 5J, 5L, 5O, 5P, 5S, 5T, 5V, and 5X).

NPAS1^−/− Ganglionic Eminences Displayed Increased Arx Expression Which Mediated an Increase in Subpallial MAP Kinase Activity and Proliferation

We focused on identifying molecular mechanisms underlying the increased proliferation and phospho-ERK expression in the MGE and CGE of NPAS1 mutants by ISH screening of the MGE and CGE for expression of genes that are implicated in these processes (Arx, COUP-TFI, Cyclin D1, Cyclin D2, Dlx1, ErbB4, FGFR1, FGFR3, Lhx2, Lhx6, NPAS3, Olig1, Sp8, and Sprouty2) (Figure 6 and Figure S5 and data not shown). While we did not find robust changes for most of these, including FGF-signaling components, there were some interesting changes in Arx and Sp8 expression. In accordance with the increase in CGE proliferation, there was a subtle increase in Sp8⁺ (LGE/CGE marker) migrating interneurons at E13.5 (Figure 6B) (Ma et al., 2012). We also observed a clear increase of Arx TF RNA at E13.5 and E15.5 in regions of the MGE and CGE where NPAS1 is normally expressed (Figures 6A and 6C). Arx function is necessary for proper interneuron migration and maturation (Colombo et al., 2007; Kitamura et al., 2002; Marsh et al., 2009), as well as neocortical neuron progenitor proliferation (Friocourt et al., 2008; Kitamura et al., 2002).

To test whether NPAS1 and its co-factor, ARNT, can directly repress Arx expression through a known Arx subpallial enhancer element (Colasante et al., 2008; Visel et al., 2013), we performed luciferase assays in CGE primary cultures (Flandin et al., 2011). We found that this enhancer had two putative NPAS1/ARNT sites, designated A and B (Figure S7D) (Hogenesch et al., 1997). Consistent with the aforementioned phenotype, NPAS1/ARNT induced a ~2-fold repression of the Arx enhancer (0.58 ± 0.035, p = 6.69E-5). Mutation of the Arx enhancer at site A, but not site B, resulted in a rescue of the repression induced by NPAS1/ARNT (p = 0.034) (Figures 7C and 7D).

Next, to investigate whether the increase in Arx expression (Figures 6A and 6C) could alter subpallial development we used ultrasound to guide injection into the E13.5 CGE of a lentivirus encoding either Arx and GFP, or GFP alone. At E15.5, we compared the number of PH3⁺/GFP⁺ cells and observed that the Arx-encoding virus increased PH3⁺/GFP⁺ cells by ~2-fold (2.18 ± 0.039, p = 0.042) (Figures 7A and 7B and Figure S7A). We also assessed the level of phospho-ERK expression in GFP⁺ regions of the CGE ventricular zone and found that Arx overexpression increased phospho-ERK expression by 1.4-fold (1.39 ± 0.12, p = 0.046) (Figures S7B and S7C). These findings provide evidence that NPAS1 repression of Arx can regulate CGE proliferation and MAP kinase activity.

Non-Synonymous Mutations in NPAS1 and NPAS3 Are Found in Humans with ASD

NPAS3 mutations are observed in patients with schizophrenia (Kamnasaran et al., 2003; Macintyre et al., 2010). Given that autism spectrum disorders (ASD) and schizophrenia share some genetic risk factors (Murdoch and State, 2013) and, like NPAS1 mutant mice, enlarged brain size is frequent in ASD children (Butler et al., 2005; Hazlett et al., 2011), we screened 947 ASD probands for mutations in NPAS1 and NPAS3 exons, introns, and 5′ and 3′ untranslated regions, and found 10 non-synonymous variants in NPAS1 and seven in NPAS3 (See the Supplemental Text). We assayed 13 of those in 190 control subjects (unscreened for autism) and found that five of seven in NPAS1 and all six in NPAS3 were absent in the controls. Although the data lacks statistical power, and does not unambiguously implicate these variants in genetic susceptibility to ASD, these results support the possibility that NPAS1 and NPAS3 mutations are related to ASD risk.

NPAS1 and NPAS3 Regulate the Proliferation of Ganglionic Eminence Progenitors

Proliferation and ERK signaling in progenitor domains of the MGE and CGE were increased and decreased in NPAS1^−/− and NPAS3^−/− mice, respectively (Figure 5). Vertebrate NPAS function in embryonic neural precursors may be related to the function of Trachealess, the NPAS Drosophila homolog. Trachealess promotes FGF-signaling by increasing expression of the FGF receptor (Ohshiro and Saigo, 1997). This program is evolutionary conserved based on the observation that NPAS3^−/− mice have reduced FGFR1 expression in the adult hippocampus (Pieper et al., 2005). However, NPAS3^−/− mutant mice did not show this phenotype in the prenatal telencephalon (data not shown). Also, NPAS1^−/− mutants did not exhibit decreased FGFR1 expression in the adult hippocampus (Pieper et al., 2005), or in the prenatal telencephalon (Figure S6C). Thus, there is no evidence that changes in phospho-ERK levels, in the subpallium of NPAS1^−/− and NPAS3^−/− mutants (Figures 5C, 5D, 5G, 5H, 5J, 5L, 5O, 5P, 5S, 5T, 5V, and 5X), is related to altered FGF-signaling.

Our study reveals new insights into NPAS1 regulation of cell proliferation. Here, we have found that the increase in proliferation and MAPK signaling in NPAS1^−/− mutant CGE and MGE progenitors was associated with increased Arx expression. Previous studies of Arx mutant mice show that lack of this TF resulted in reduced cortical proliferation (Friocourt et al., 2008; Kitamura et al., 2002). We demonstrated that increased Arx expression was sufficient to enhance subpallial proliferation and MAP kinase activity, using in utero transduction of Arx into the CGE (Figures 7A and 7B and Figures S7B and S7C). Furthermore, we showed that NPAS1 could directly inhibit a bone fide Arx subpallial enhancer (Figures 7C and 7D and Figure S7D) (Colasante et al., 2008).

Given that NPAS1 and NPAS3 have opposite effects on proliferation, an imbalance of either NPAS1 or NPAS3 expression and/or function could potentially be compensated by changes in the expression of the other gene. However, we did not observe changes of NPAS3 expression in the NPAS1^−/− E13.5 and P0 forebrain (data not shown).

NPAS1 and NPAS3 Modified the Production of Specific Subtypes of Cortical Interneurons

The increase in Arx and phospho-ERK expression and the increase in proliferation in the MGE and CGE of NPAS1^−/− mutants were associated with an increase in subpallially-derived cortical interneurons (Figure 2). As a result, NPAS1 mutants had excessive SST⁺ and VIP⁺ interneurons, products of the MGE and CGE, respectively (Figures 3A, 3A′, 3E, 3E′, 3F, and 3J) (Rudy et al., 2011). In contrast, decreased phospho-ERK expression and cell proliferation in the MGE and CGE of NPAS3^−/− mutant mice resulted in reduced numbers of SST⁺ and VIP⁺ interneurons (Figures 3A, 3A″, 3E, 3E″, 3F, and 3J).

Both NPAS1 and NPAS3 mutants had a normal density of PV⁺ interneurons (Figures 3D––3D″, and 3I). PV and SST are MGE-derived (Rudy et al., 2011). NPAS1, along with NPAS3, COUP-TFI, Dlx1 and SatB1 (Cobos et al., 2005; Denaxa et al., 2012; Lodato et al., 2011) define a set of TFs that differentially regulate the numbers of interneurons that express PV and SST (Figures 3D––3D″, 3E–3E″, 3I, and 3J). At least one PV⁺ subtype, the Chandelier neuron, is generated late in gestation (Inan et al., 2012; Taniguchi et al., 2013); perhaps NPAS1 and NPAS3 regulation of MGE proliferation is time-dependent, and thus preferentially affects early-born SST⁺ interneurons. Based on the embryonic phenotypes of NPAS1^−/− and NPAS3^−/− mutants, we propose that NPAS1 and NPAS3 control the SST/PV ratio via their functions in MGE progenitor cells. Future studies are needed to examine the roles of NPAS1 and NPAS3 in regulating interneurons via their expression in maturing and mature interneurons (Figure 1C).

The significance of transcriptional repression in the regulation of cortical interneuron number is becoming apparent. A recent study demonstrates that Olig1 inhibits the generation of GABAergic interneurons produced in the MGE by repression of Dlx expression through the Dlx1/2 I12b intergenic enhancer (Silbereis et al., 2014). Likewise, NPAS1 inhibits the generation of cortical interneurons by repression of Arx expression through the subpallial Arx enhancer, which mediates CGE progenitor cell proliferation (Figure 7) (Colasante et al., 2008).

NPAS1 Function in Interneuron Generation Controls Cortical Excitation/Inhibition

NPAS1^−/− mutant mice display increased numbers of GAD67-GFP⁺ interneurons in layer I (Figure S4C). It is likely that an excess of GABAergic neurons in layer I would have an effect on the inhibitory tone of pyramidal cell dendrites within this layer. NPAS1 mutants also exhibit excessive SST⁺ and VIP⁺ interneurons. VIP⁺ interneurons are enriched in layers II/III (Kawaguchi and Kubota, 1997; Miyoshi et al., 2010), and form synapses with pyramidal cell somata and dendrites (Kawaguchi and Kubota, 1996) and other interneurons (Acsády et al., 1996; Dávid et al., 2007; Hajos et al., 1996). Thus, with an increase in both SST⁺ and VIP⁺ interneurons, it is difficult to deduce the net physiological effect a priori. We found that NPAS1 mutants had enhanced cortical inhibition in layer II/III of the somatosensory cortex (Figure 4), implying that the increase in interneurons resulted in a net increase in inhibition in superficial cortical layers. The enhanced cortical inhibition in layer II/III somatosensory cortex of NPAS1^−/− mutants suggests a decreased E/I ratio in layer II/III somatosensory cortex. This is supported by estimation of the cellular E/I ratio within this cortical region of NPAS1^−/− mutant mice (Figure S4D). While increased cortical inhibition is likely to reduce noise in the cortex, and it could be protective for epilepsy, it may be detrimental to cortical function by dampening cortical signals and thereby predispose to neuropsychiatric disorders.

NPAS1 and NPAS3 in Human Neuropsychiatric Disorders

Our examination of NPAS1 and NPAS3 functions in mice provides novel mechanistic insights into human neuropsychiatric disorders. NPAS3^−/− mutant mice display behavioral and neuroanatomical deficits associated with human schizophrenia (Erbel-Sieler et al., 2004) and NPAS3 mutations are observed in patients with schizophrenia (Kamnasaran et al., 2003; Macintyre et al., 2010). Furthermore, NPAS1 mutant mice exhibit an increase in brain size (Figures S3D–S3G), similar to the enlargement seen in some children with autism spectrum disorders (ASD) (Butler et al., 2005; Hazlett et al., 2011). Notably, we have found sporadic non-synonymous mutations in NPAS1 and NPAS3 in individuals with ASD, which suggests that NPAS1 and NPAS3 mutations may be associated with ASD risk. However, our analyses lack the statistical power to conclude that these alleles contribute to ASD. Nonetheless, these data, and the experimental analyses of the mouse mutants, strongly suggest that further genetic analyses of NPAS1 and NPAS3 are warranted in human neuropsychiatric disorders.

Mice

NPAS1^+/− and NPAS3^+/− mice were obtained from Steven McKnight (University of Texas Southwestern Medical Center) and genotyped as described by Erbel-Sieler and colleagues (Erbel-Sieler et al., 2004). NPAS1^+/− mice were maintained in a C57BL/6J strain background. NPAS1^+/−; NPAS3^+/− females mated to NPAS1^+/−; NPAS3^+/− males also produced C57Bl/6J strain litters. NPAS3^−/− mutant mice obtained from this cross were utilized for BrdU analysis of the ganglionic eminences, assessment of cortical layer-specific markers, and in the examination of interneuron markers in the P30 somatosensory cortex. However, due to increased mortality of NPAS3^−/− mutants maintained on a C57BL/6J strain background, subsequent experiments were performed with NPAS3^−/− mutant mice maintained on a CD-1 strain background. These experiments include assessment of apoptosis within the somatosensory cortex, examination of PH3 and phospho-ERK expression in the ganglionic eminences, and MRI analysis.

The Lhx6-GFP BAC transgenic mouse line was obtained from The Gene Expression Nervous System Atlas Project (GENSAT) at The Rockefeller University (New York, NY). GAD67-GFP mice were genotyped as described by Tamamaki and colleagues (Tamamaki et al., 2003). As a note on nomenclature, experimental samples designated as NPAS1^+/+ indicates that the NPAS1 allele is unaltered; however data from these mice are not referred to as wildtype because they have either the Lhx6-GFP or GAD67-GFP alleles. CD-1 wildtype mice were obtained from Charles River Laboratories. Wildtype littermate mice were used as controls for all experiments. For staging of embryos, midday of the vaginal plug was calculated as embryonic day 0.5 (E0.5). Mouse colonies were maintained at the University of California, San Francisco and University of Texas Southwestern Medical Center, in accordance with National Institutes of Health, UCSF, and UT Southwestern guidelines.

Histology

See the Supplemental Experimental Procedures for details on immunohistochemistry, in situ hybridization, BrdU labeling, and Nissl stain analysis.

Magnetic Resonance Imaging (MRI)

P30 4% PFA-fixed brains were washed twice in 20 mL PBS for a total of 24 hours and imaged in Fluorinert FC-40 (Sigma Aldrich) for null background signal. The imaging was done on an 600 MHz NMR spectrometer (Agilent Technologies Inc.) with imaging gradients and the following parameters: 3D gradient echo, TE/TR 15/75 ms, 8 averages, field of view (FOV) 12.8 mm isotropic, resolution of 50 μm × 50 μm × 100 μm, and a total scan time of 5.5 hours. The acquired images were converted on the Varian console to the DICOM format and reformatted into the same orientation as the histology sections using OsiriX, an open source image viewer. Volumetric measurements were made using custom-built software in MATLAB. Cortical and basal ganglia volumes were statistically analyzed in three mice of each genotype. Results are presented as mean ± SEM. Statistical differences between experimental groups were assessed with the Student’s t-test using SPSS 15 software (IBM). See Table S3 in the Supplemental Experimental Procedures for details on cortical and basal ganglia volumes in control, NPAS1^−/−, and NPAS3^−/− mutants.

Electrophysiology

Coronal brain slices (300 μm) were prepared from P21-30 wildtype and NPAS1^−/− mice. Slices were submerged in the recording chamber and continuously perfused with oxygenated ACSF (32–34°C) containing (in mM): 124 NaCl, 3 KCl, 1.25 NaH₂PO₄-H₂O, 2MgSO₄-7H₂O, 26 NaHCO₃, 10 dextrose, and 2 CaCl₂ (pH 7.2–7.4, 300–305 mOsm/kg). Whole-cell patch-clamp recordings from layer II/III pyramidal cells in somatosensory neocortex were performed at 40X using an upright, fixed-stage microscope (Olympus BX50WI) equipped with infrared, differential interference contrast (IR-DIC). Patch pipettes (2–4 MΩ) were filled with an internal solution, containing (in mM): 140 CsCl, 1 MgCl₂, 10 HEPES, 11 EGTA, 2 NaATP, 0.5 Na₂GTP and 1.25 QX-314, pH 7.28. Recordings were obtained with an Axopatch 1D amplifier, filtered at 5 kHz, and recorded to pClamp 10.2 software (Clampfit, Axon Instruments). Spontaneous (s) and miniature (m) IPSCs were examined at a holding potential of −70 mV. GABAergic currents were isolated by adding 1mM kynurenic acid to the ACSF to block glutamate receptors, and tetrodotoxin (TTX, 2μM) was added to isolate mIPSCs. Series resistance was typically <15 MΩ and was monitored throughout the recordings. Data were only used for analysis if the series resistance remained <20 MΩ and changed by ≤20% during the recordings. Data analysis was performed using pClamp 10.2 (Clampfit, Axon Instruments), MiniAnalysis 6.0 (Synaptosoft), Microsoft excel, and Sigmaplot 12.3 programs. Events characterized by a typical fast rising phase and exponential decay phase were manually detected using MiniAnalysis. A 2 min sample recording per cell was used for measuring IPSC frequency, amplitude, 10–90% rise time, and decay time constant. Kinetic analysis of the IPSCs was performed with a single-exponential function. The threshold for event detection was currents with amplitudes greater than three times the root mean square (RMS) noise level. Spontaneous IPSC measurements were recorded from 18 wildtype and 20 NPAS1^−/− pyramidal cells. Miniature IPSC measurements were recorded from 14 wildtype and 16 NPAS1^−/− pyramidal cells. Seven wildtype and nine NPAS1^−/− animals were used for sIPSC and mIPSC measurements. Results are expressed as mean ± SEM. Statistical differences between experimental groups were assessed with the Mann–Whitney U test using SigmaPlot 12.3 software (SYSTAT).

Primary Cell Culture and Luciferase Assays

CGE tissue was dissected from E12.5 WT embryos and mechanically dissociated with a P1000 pipette tip. 200,000 cells were seeded into tissue culture dishes precoated with poly-L-lysine 10 μg/ml, Sigma) then laminin (5 μg/ml, Sigma), and grown in N5 media (DMEM-F-12 with glutamax, N2 supplement, 35 μg/ml bovine pituitary extract, 20 ng/ml bFGF, and 20 ng/ml EGF) as previously described for MGE cells (Flandin et al., 2011).

CGE primary cultures were transfected using Fugene® 6 (Promega) with DNA expression vectors and reporters comprising four conditions: 1) empty firefly-luciferase reporter (PGL4.23) and a control DNA expression vector only expressing GFP (CMV-GFP), 2) empty PGL4.23 and CMV-NPAS1, 3) Arx enhancer (UAS3-PGL4.23) luciferase reporter and CMV-GFP, and 4) UAS3-PGL4.23 and CMV-NPAS1. All conditions also included the same amounts of CMV-ARNT (NPAS co-factor) and Renilla-luciferase (normalization control). Cell lysates were collected at two days post-transfection and procedures were performed according to the manufacturer’s protocol (Promega, dual luciferase assay system). Enhancer activity for each condition was determined from three independent experiments, each performed in triplicate. Results were calculated as mean ± SEM and presented as fold change relative to the basal condition containing CMV-GFP and the PGL4.23 empty luciferase reporter vector. Statistical differences between experimental groups were determined with the Student’s t-test using SPSS 15 software (IBM). See the Supplemental Experimental Procedures for details on expression and luciferase reporter vectors.

Virus Production and In Utero Injection

HEK293T cells grown in DMEM H21 with 10% FBS were transfected with four plasmids to generate lentivirus particles, (pVSV-g, pRSVr, pMDLg-pRRE and the lentiviral vector). Cells were transfected at ~70% confluency and media was completely replaced four hours after transfection, then cultured for four days before harvesting. On day four of culture, all the media (~35 mls) was collected and filtered through a 0.45 low protein binding membrane to remove cells and large debris. The filtered media was pooled and ultracentrifuged at 100,000 × g for 2.5 hours at 4°C. After the ultracentrifuge step, supernatant was removed and the pellet was resuspended overnight in sterile PBS then stored at −80°C until use. For infection, recipient pregnant females (E13.5) were anesthetized with isoflurane, their uterine horns exposed, and mounted under an ultrasound microscope (Vevo 770, VisualSonics). A beveled glass micropipette (~50 μm diameter) was front loaded with CMV-GFP or CMV-GFP-T2a-Arx lentivirus and inserted into the ventricular zone of the CGE under real-time ultrasound guidance imaging. ~1 μl of solution was injected into the ventricular zone of the CGE. Three successfully injected animals were analyzed two days later (E15.5) for each experimental condition. Image analysis of CMV-GFP⁺/PH3⁺, CMV-GFP⁺/phospho-ERK⁺, CMV-GFP-T2a-Arx⁺/PH3⁺ or CMV-GFP-T2a-Arx⁺/phospho-ERK⁺ co-expression was performed on images acquired on a confocal microscope (LSM 510 META NLO, Carl Zeiss International) with a 20X objective. The percentage of CMV-GFP⁺ or CMV-GFP-T2a-Arx⁺ that express PH3 was calculated in a 10,000 μm² area of the CGE ventricular zone. Results are presented as mean ± SEM. Statistical differences between experimental groups were assessed with the chi-squared test using SPSS 15 software (IBM). Phospho-ERK expression levels were measured in GFP⁺ regions of the CGE ventricular zone with the aid of Adobe Photoshop CS4 software. Expression levels were determined as a ratio of the integrated density of phospho-ERK⁺ CGE ventricular zone cells in a 4,000 μm² area to the integrated density of phospho-ERK⁺ cortical ventricular zone cells in a 4,000 μm² area. Integrated density is equal to the sum of the pixel values in a selected area. Results were calculated as mean ± SEM and presented as fold change relative to control. Statistical differences between experimental groups were determined with the Student’s t-test using SPSS 15 software (IBM). See the Supplemental Experimental Procedures for details on lentiviral vectors.

Human DNA Sequencing

See the Supplemental Experimental Procedures for details on human DNA sequencing.