MACF1 Controls Migration and Positioning of Cortical GABAergic Interneurons in Mice

Abstract

GABAergic interneurons develop in the ganglionic eminence in the ventral telencephalon and tangentially migrate into the cortical plate during development. However, key molecules controlling interneuron migration remain poorly identified. Here, we show that microtubule-actin cross-linking factor 1 (MACF1) regulates GABAergic interneuron migration and positioning in the developing mouse brain. To investigate the role of MACF1 in developing interneurons, we conditionally deleted the MACF1 gene in mouse interneuron progenitors and their progeny using Dlx5/6-Cre-IRES-EGFP and Nkx2.1-Cre drivers. We found that MACF1 deletion results in a marked reduction and defective positioning of interneurons in the mouse cerebral cortex and hippocampus, suggesting abnormal interneuron migration. Indeed, the speed and mode of interneuron migration were abnormal in the MACF1-mutant brain, compared with controls. Additionally, MACF1-deleted interneurons showed a significant reduction in the length of their leading processes and dendrites in the mouse brain. Finally, loss of MACF1 decreased microtubule stability in cortical interneurons. Our findings suggest that MACF1 plays a critical role in cortical interneuron migration and positioning in the developing mouse brain.

Introduction

There are two primary modes of neuronal migration in the developing brain (Marin et al. 2003; Moffat et al. 2015). Cortical pyramidal neurons, which are mostly excitatory, are generated in the cortical ventricular (VZ) and subventricular zones (SVZ), and migrate radially to form cortical layers in the cerebral cortex (Tan et al. 1998; Noctor et al. 2001; Rakic 1972; Geschwind and Rakic 2013). Meanwhile, GABAergic inhibitory interneurons originate from the ganglionic eminence and migrate tangentially into the cerebral cortex (Gelman and Marin 2010). Cortical interneurons follow two tangentially oriented migratory streams in order to enter the cortex. Prior to E12, early-migrating interneurons traverse the marginal zone (MZ) en route to the cortex, whereas late-migrating interneurons move through the intermediate zone (IZ) and SVZ before reaching their cortical destinations. Cortical interneurons constitute ~20% of all cortical neurons; yet, these interneurons are crucial in regulating the balance, flexibility, and functional architecture of cortical circuits (Markram et al. 2004; Klausberger and Somogyi 2008). Dysfunctions of cortical GABAergic circuits have been implicated in neurodevelopmental disorders including autism, schizophrenia, intellectual disability, and epilepsy (Rossignol 2011). Several transcription factors regulate the development of GABAergic interneurons derived from the lateral ganglionic eminence (LGE), medial ganglionic eminence (MGE), and caudal ganglionic eminence (CGE). The DLX family of genes are expressed in the SVZ of the embryonic LGE, MGE, and CGE and are continuously expressed in GABAergic interneurons throughout migration and differentiation (Eisenstat et al. 1999; Long et al. 2009). The NKX2.1 transcription factor is expressed within the MGE by interneuron progenitors in both the VZ and SVZ (Butt et al. 2008).

Microtubule-actin cross-linking factor 1 (MACF1) is a member of the spectraplakin family of cytoskeletal linker proteins (Sonnenberg and Liem 2007). This protein interacts with both F-actin and microtubules with an actin-binding domain in the N-terminus and a microtubule-binding domain in the C-terminus. MACF1 aids in promoting microtubule capture in lamellipodia during cell migration (Zaoui et al. 2010; Wu et al. 2011). MACF1 is highly expressed in the developing brain (Ka et al. 2014b) and plays a role in neuronal differentiation. For example, mouse MACF1 and its Drosophila homolog Shot, promote the growth of axons and dendrites and maturation of synaptic terminals (Prokop et al. 1998; Lee et al. 2000; Sanchez-Soriano et al. 2009). In addition to the role in neurite differentiation, MACF1 is associated with nuclear and somal translocation of neural cells. During migration, neurons extend a leading process, toward which the soma elongates and translocates via alterations in the actin and microtubule cytoskeletons (Rivas and Hatten 1995; Tsai and Gleeson 2005). Centrosomes coordinate nuclear and somal translocation by maintaining appropriate tension between the leading process and soma. Centrosomes are positioned near the nucleus and move toward the distal tip of the leading process, and guide the nucleus in the same direction (Rakic 1971; Gregory et al. 1988; Solecki et al. 2004; Tanaka et al. 2004; Tsai and Gleeson 2005; Sakakibara et al. 2014). By stabilizing microtubules and promoting centrosome-mediated somal translocation, MACF1 regulates the radial migration of pyramidal neurons during brain development (Ka et al. 2014b). In this study, we investigated the loss of MACF1 function in tangential migration of cortical interneurons in the developing brain. Because complete MACF1 deletion is lethal during early embryonic development in mice (Chen et al. 2006), we conditionally deleted MACF1 in interneurons. Our results identified an essential role for MACF1 in interneuron migration and positioning in the developing cortex.

Materials and Methods

Materials

EMTB-3XGFP (Addgene) and blebbistatin (Calbiochem) were purchased.

Mice

Mice were handled according to our animal protocol approved by the University of Nebraska Medical Center. The MACF1 floxed mouse was described previously (Wu et al. 2011). GAD67-EGFP mouse (Oliva et al. 2000) and Nkx2.1-Cre mouse (Xu et al. 2008) were purchased from the Jackson Laboratory. Dlx5/6-CIE mouse (Stenman et al. 2003) was obtained from Dr. Kenneth Campbell.

Immunohistochemistry

Immunohistochemical staining of brain sections or dissociated neurons was performed as described previously (Kim et al. 2005; Ka et al. 2014b). The following primary antibodies were used: Rabbit anti-MACF1 (Wu et al. 2011), rabbit anti-MACF1 (Santa Cruz), rabbit anti-GABA (Sigma), mouse anti-parvalbumin (Millipore), rabbit anti-calbindin (Millipore), mouse anti-MAP2 (Covance), chicken anti-GFP (Invitrogen), and rabbit anti-GFP (Invitrogen). Appropriate secondary antibodies conjugated with Alexa Fluor dyes (Invitrogen) were used to detect primary antibodies.

Morphometry

For cell counts, numbers of neurons positive to GFP, parvalbumin, GABA, calbindin, MAP2, or DAPI were obtained as described previously (Ka et al. 2014a; Jung et al. 2016). Five mice for each condition (control or mutant) were used. Cell counts were described in figure legends. More than 20 different coronal brain sections alongside rostro-caudal axis from each brain were taken with Zeiss LSM510 and LSM710 confocal microscopes and a Nikon Eclipse epifluorescence microscope attached with a QImaging CCD camera. The images were analyzed using ZEN (Zeiss), LSM image browser (Zeiss), QCapture software (QImaging), and ImageJ (NIH). The calculated values were averaged, and some results were recalculated as relative changes versus control. For analyzing cultured cells, more than 20 fields scanned horizontally and vertically were analyzed in each condition. Cell numbers were described in figure legends. For the quantification of the length and number of leading processes and branches, images of 20 different brain sections at periodic distances along the rostro-caudal axis from 5 mice for each condition were examined. The numbers of cells examined are described in figure legends.

Primary Neuron Cultures

Primary neuronal culture was performed as described previously (Kim et al. 2004, 2009a, 2009b). Briefly, MGEs from E13.5-16.5 mice were isolated and dissociated with trituration after trypsin/EDTA treatment. Then, the cells were plated onto poly-D-lysine/laminin-coated coverslips and cultured in neurobasal medium containing 5% serum, B27, and N2 supplements.

Cell Transfection

Mouse primary neurons were transfected with EMTB-3XGFP as described in previous papers (Ka and Kim 2015; Ka et al. 2016). Briefly, embryonic MGEs were dissociated and suspended in 100 µL of Amaxa electroporation buffer (Lonza) with 1–10 µg of plasmid DNA. Then, suspended cells were transferred to an electroporation cuvette and electroporated with Amaxa Nucleofector (Lonza). Cells were then plated onto coated coverslips and the medium was changed 4 h later to remove the remnant transfection buffer.

Time-Lapse Imaging

Organotypic brain slices were prepared from E14.5 mice as described previously (Ka et al. 2014b). Briefly, control and MACF1^loxP/loxP; Dlx5/6-CIE brains were collected and sectioned at E14.5. The brains were embedded in 3% low melting point agarose, and coronal brain slices at 250 μm thickness were prepared using a LEICA VT1000S vibratome. The slices were then placed on poly-lysine/laminin-coated transwell inserts and cultured in neurobasal media organotypically using an air interface protocol until imaging.

For time-lapse imaging, a LSM710 inverted confocal microscope (Zeiss) equipped with a CO₂ incubator chamber (5% CO₂, 37 °C) was used. Multiple Z-stacks with the options of 10–20 successive “z” optical planes spanning 50–70 μm were acquired on preselected positions of electroporated slices. Repetitive imaging was performed every 15 min for up to 11 h. Mean velocity of migrating cells was obtained using the ImageJ plugin Manual tracking.

Statistical Analysis

Normal distribution was tested using the Kolmogorov–Smirnov test, and variance was compared. Unless otherwise stated, statistical significance was determined by two-tailed unpaired Student's t-test for two-population comparison and one-way analysis of variance with the Bonferoni correction test for multiple comparisons. Data were analyzed using GraphPad Prism and presented as mean ± SEM. P values were indicated in figure legends. Each experiment in this study was performed blind and randomized. Animals were assigned randomly to the various experimental groups, and data were collected and processed randomly. The allocation, treatment, and handling of animals were the same across study groups. Control animals were selected from the same litter as the test group. Pups weighing <20% of the intergroup and intragroup average were excluded from the experiments. The individuals conducting the experiments were blinded to group allocation and the allocation sequence.

Results

MACF1 Is Expressed in Tangential Migratory Streams During Cortical Development

To determine the role of MACF1 in interneuron migration, we first examined the expression pattern of MACF1 in interneuron migration routes using immunostaining. At E15.5, MACF1 was expressed in the ventral telencephalon, including the MGE and LGE (Fig. 1A). It is noted that MACF1 was highly expressed in tangential migratory streams, that is, the dorsolateral to subplate/MZ route (arrows in Fig. 1A) and MGE to SVZ/VZ route (arrow heads in Fig. 1A). Next, we examined the cellular localization of MACF1 in cultured interneurons. MACF1 was expressed in MAP2-positive interneurons and localized in the soma and in neurites (Fig. 1B).

Figure 1.

MACF1 is expressed in tangential migratory streams in the developing cortex. (A) Immunostaining of E15.5 mouse brain sections. MACF1 was expressed throughout the ventral and dorsal telencephalon. Notably, MACF1 was accumulated in tangential migratory streams, that is, marginal zone (arrows) and ventricular zone (arrow heads) routes in the developing brain. Scale bar, 200 μm. Right panel is a higher magnification image showing MACF1 expression in the ganglionic eminence. (B) MACF1 is expressed in neurites and somas of cultured interneurons. E14 primary neurons from the medial ganglionic eminence were cultured and immunostained with a MAP2 and a MACF1 antibodies. MACF1 was co-localized with MAP2 in somas and neurites. Scale bar, 25 μm.

Elimination of MACF1 Causes Abnormal Positioning of Cortical Interneurons in the Embryonic and Postnatal Brain

To investigate the role of MACF1 in cortical interneuron development, we deleted MACF1 in cortical interneurons using a conditional knockout strategy. MACF1 floxed alleles were crossed with the Dlx5/6-Cre-IRES-EGFP (Dlx5/6-CIE) mouse line (Stenman et al. 2003), which expresses Cre recombinase exclusively in GABAergic interneurons but not in dividing neural progenitors in the ganglionic eminence. Dlx5/6-CIE expresses EGFP within a separate open reading frame, which allows simple marking and tracing of interneurons expressing Cre recombinase. We generated control (MACF1^loxP/+; Dlx5/6-CIE) and knockout (MACF1^loxP/loxP; Dlx5/6-CIE) mice and examined the position and number of EGFP-positive interneurons in the dorsolateral cortex at E15.5 (Supplementary Fig. 1) and E19.5 (Fig. 2). Compared with control samples, MACF1-knockout brains showed a 64% increase in the number of EGFP-positive interneurons in the lateral cerebral cortex (Fig. 2A,B, defined region 1). In the dorsal cerebral cortex of MACF1-mutant mice, however, the number of EGFP-positive neurons was decreased by 25% (Fig. 2A,B, defined region 2). Additionally, we found that MACF1-deleted interneurons were aberrantly positioned in different cortical layers (Fig. 2A,C). In the dorsal cortex, control neurons were found mostly in the MZ, cortical plate (CP), and VZ, and were sparsely located in the IZ. In contrast, MACF1-deleted brains showed no distinct accumulation of interneurons in the MZ or CP. Instead, the highest number of EGFP-positive interneurons was measured in the IZ. A similar pattern was observed in the lateral cortex where MACF1-deleted brains showed a higher number of interneurons in the IZ, while control brains had a majority in the MZ and CP. These phenotypes of migration and positioning in MACF1^loxP/loxP; Dlx5/6-CIE brains appeared to be independent of cell death because there was no significant difference in the level of cleaved caspase-3 between E19.5 control and knockout ventral brains (Fig. 2D, E). We further examined interneurons in two major tangential migratory streams, the dorsomarginal (dotted arrow in Fig. 2F) and VZ (lined arrow in Fig. 2F) streams in the cerebral cortex. Both streams were thinner in MACF1-deleted brains compared with controls (Fig. 2F,G). The fluorescence intensities of the streams were also reduced in MACF1-mutant brains. These findings demonstrate that MACF1 governs the interneuron number and positioning in the developing dorsal telencephalon.

Figure 2.

MACF1 determines the number and position of tangentially migrating interneurons in the developing cerebral cortex. (A) MACF1-deleted mutants show disrupted patterns of interneuron positioning. Coronal sections of E19.5 control (MACF1^loxP/+; Dlx5/6-CIE) and MACF1^loxP/loxP; Dlx5/6-CIE brains were immunostained with a GFP antibody. Two panels in the right represent the boxed regions in left panels. (1): lateral cortex; (2): dorsal cortex. Scale bar, 200 μm and 50 μm. (B) Quantification of GABAergic interneuron positions in the cortical plate (CP). The graph indicates the distribution of interneurons in the areas of the CP as indicated in (A) in each genotype. MACF1-knockout neurons were accumulated in the lateral cortex, while control interneurons were present at higher concentrations in the dorsomedial cortex. Control: MACF1^loxP/+; Dlx5/6-CIE. KO: MACF1^loxP/loxP; Dlx5/6-CIE. N = 5 mice for each condition; cell counts = 5020 cells for control and 5122 cells for KO. Statistical significance was determined by two-tailed Student's t-test. ***P <  0.001. (C) Quantification of neuron positions within different layers of the cerebral cortex. Top and bottom panels show the dorsal and lateral cortex, respectively. N = 5 mice for each condition; cell counts = 8017 cells for control and 8212 cells for KO. Statistical significance was determined by multiple t-tests with the Bonferoni correction test. Data shown are mean ± SEM. Asterisks indicate significant difference when compared with controls. **P <  0.01; ***P < 0.001. (D) MACF1^loxP/loxP; Dlx5/6-CIE ventral brains show no difference in cell death. Western blotting using lysates from E19.5 control and KO ventral brains was performed to measure the level of cleaved caspase-3. (E) Quantification of (D). No significant change in the cleaved caspase-3 level in KO samples compared with controls. N = 3 independent experiments using 3 mice for each condition. (F) Aberrant tangential migration routes in MACF1-deleted brains. Top panels show two distinct migration routes of developing interneurons (dotted arrow: MZ route; lined arrow: VZ route). Bottom panels are higher magnification images of the MZ and VZ containing migrating interneurons. Scale bar, 20 μm. (G) Quantification of GABAergic interneurons within the migratory routes. The number of interneurons in MACF1-knockout brains was decreased in both the MZ and VZ routes as assessed by the thickness and intensity of GFP bands. N = 5 mice for each condition. Statistical significance was determined by two-tailed Student's t-test. **P <  0.01; ***P < 0.001.

Next, we investigated whether the abnormal interneuron phenotype seen in embryonic MACF1^loxP/loxP; Dlx5/6-CIE brains persists into postnatal stages. Thus, we analyzed control and knockout (MACF1^loxP/loxP; Dlx5/6-CIE) brain tissues at postnatal day 14. The number of EGFP-positive interneurons in the dorsal cortex in MACF1-deleted mice was reduced by 42% compared with controls (Fig. 3A,B). Cortical interneurons are classified into subtypes according to their origin, morphology, and function. We assessed interneuron subtypes in MACF1^loxP/loxP; Dlx5/6-CIE brains by immunostaining with antibodies to different interneuron markers, GABA, parvalbumin, and calbindin. The numbers of GABA, parvalbumin, and calbindin neurons were drastically reduced by 46%, 72%, and 54%, respectively, in MACF1^loxP/loxP; Dlx5/6-CIE dorsal cortices compared with controls (Fig. 3A,B).

Figure 3.

MACF1 is required for the placement of postnatal cortical interneurons. (A) Postnatal positioning patterns of control and MACF1-mutant interneurons. Cortical sections of control (MACF1^loxP/+; Dlx5/6-CIE) and MACF1^loxP/loxP; Dlx5/6-CIE mice at P14 were immunostained with interneuron subtype markers: GABA, parvalbumin, or calbindin antibody. MACF1-knockout brains exhibited decreased numbers of different types of interneurons in the dorsal cortex compared with controls. Scale bar, 250 and 50 μm. (B) Quantification of interneuron numbers in control and MACF1-knockout brains. N = 3105 cells from 5 mice for control, and N = 2516 cells from 5 mice for MACF1-knockout brains. Statistical significance was determined by two-tailed Student's t-test. *P < 0.05, **P < 0.01, ***P < 0.001. (C) Postnatal interneuron positioning in control (MACF1^loxP/+; Nkx2.1-Cre; Gad67-EGFP) and MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP mice. Brain sections of control and MACF1-mutant mice at P14 were immunostained with GABA or parvalbumin antibody. MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP brains exhibited fewer interneurons in the cerebral cortex compared with controls. Scale bar, 1 mm, 250 μm, and 50 μm. (D) Quantification of interneuron numbers in control and MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP brains. N = 2114 cells from 3 mice for control, and N = 1959 cells from 5 mice for MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP brains. Statistical significance was determined by two-tailed Student's t-test. ***P < 0.001.

Next, we further confirmed these results with another deletion strategy of interneuron populations in the MGE using the Nkx2.1-Cre driver (Xu et al. 2008). The GAD67-EGFP mouse line (Oliva et al. 2000) was used to label interneurons. We generated control (MACF1^loxP/+; Nkx2.1-Cre; Gad67-EGFP) and knockout (MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP) mice. In control brains, EGFP-positive interneurons were positioned normally in the dorsal cortex at P14 (Fig. 3C,D). In striking contrast, MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP brains showed that EGFP-positive interneurons were mostly found in the region around the ventrolateral area and very few interneurons were located in the cerebral cortex (Fig. 3C). The number of EGFP-positive interneurons in the dorsal cortex was reduced by 89% in MACF1 mutants, compared with control mice (Fig. 3C,D). We also assessed GABA and parvalbumin-positive interneurons in MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP mice. The numbers of GABA- and parvalbumin-positive interneurons were reduced by 78% and 69%, respectively, in MACF1-mutant mice (Fig. 3C,D). There were less EGFP-positive (GAD67) neurons than GABA-positive neurons in control and knockout brains. The Gad67-EGFP mouse line used in this study is TgN(GadGFP)45 704Swn from the Jackson Laboratory. This mouse expresses EGFP under the control of the mouse Gad1 promoter in a subpopulation of cortical and hippocampal GABA neurons that express somatostatin. Thus, it is expected to have less GAD67 than GABA-positive neurons. Together, these results show that MACF1 is required for normal positioning of cortical interneurons and suggest that MACF1 plays an important role in tangential migration of interneurons during cortical development.

MACF1 Controls the Speed of Interneuron Migration

The reduced number of interneurons in the dorsal telencephalon of MACF1-mutant mice raises two possibilities. Individual MACF1-deleted interneurons may migrate slower, or alternatively, they may migrate at a normal rate, but exhibit a more back-and-forth movement pattern through the cortex. To directly test the migration rate of individual cells, we assessed the movement of migrating neurons using time-lapse imaging on control and MACF1^loxP/loxP; Dlx5/6-CIE cortical slices. We found that MACF1-deleted interneurons migrated much slower than control neurons (Fig. 4A,B). Control interneurons migrated at an average speed of 17 ± 3 μm/min, and MACF1-deleted interneurons migrated at an average of 5 ± 2 μm/min (Fig. 4B). This represents a 70% reduction in the speed of migration for MACF1-deleted interneurons, compared with controls.

Figure 4.

Aberrant migration speed of MACF1-knockout interneurons. (A) Time-lapse imaging of control and MACF1-knockout neurons. MACF1-mutant GABAergic interneurons displayed a reduced migration speed during tangential migration. Control (MACF1^loxP/+; Dlx5/6-CIE) and MACF1^loxP/loxP; Dlx5/6-CIE brains were sliced at E15. The tangential movement of migrating interneurons in the slice was traced using fluorescent confocal imaging. Asterisks, arrows, and arrowheads traced movement of same cells. Dotted lines were used for movement comparison. (B) Quantification of interneuron migration. Migration distances per hour were measured. N = 20 cells from 3 mice for control, and N = 20 cells from 3 mice for MACF1-knockout brains.

Abnormal Tangential to Radial Transition of MACF1-Deleted Interneurons

Cortical interneurons begin migrating tangentially from the MGE toward the CP during development. Once interneurons arrive in the cerebral cortex, they reorient themselves and begin migrating radially into more superficial cortical layers (Baudoin et al. 2012). By examining the orientation of leading processes, we tested whether MACF1 plays a role in this tangential to radial transition in migrating interneurons. We collected E15.5 control and MACF1^loxP/loxP; Dlx5/6-CIE brains and assessed the direction of leading processes in EGFP-positive interneurons. In control brains, a majority of interneurons were located within the tangential migratory stream with leading processes extending toward the medial cortex (59%) (Fig. 5). Some neurons in the tangential stream were directed toward the subpallium (19%). A minority of interneurons was radially oriented (13%), mostly angled toward the CP with leading processes extending toward the pial surface. However, in MACF1-mutant brains, interneurons appeared to be spread throughout the tangential and radial migratory streams (Fig. 5). The number of tangentially oriented interneurons in MACF1 mutants was decreased by 48% compared with controls, but the number of radially oriented interneurons was increased by 179% (Fig. 5). These findings show that MACF1 regulates the transition of migration direction from a tangential to a radial route during cortical development. Together, our data suggest that decreased migratory speed and abnormal targeting contribute to the reduced number of cortical interneurons in MACF1 mutants.

Figure 5.

Abnormal switching from tangential to radial migration of MACF1-knockout interneurons. (A) Brain sections from E15 control and MACF1^loxP/loxP; Dlx5/6-CIE mice were immunostained with an anti-GFP antibody. Orientation of interneuron movement was assessed by the direction of leading process extension. Control neurons showed mostly tangential orientation toward the medial cortex. In contrast, MACF1-deleted GABAergic interneurons were oriented more radially toward the dorsal cortex. Scale bar, 50 and 20 μm. (B) Quantification of the orientation of leading processes in control and MACF1-knockout interneurons. MACF1-mutant brains showed a shift of migration orientation from tangential to radial in the CP and MZ compared with controls. Within the intermediate zone and SVZ, the shift was still found but less severe than in the CP. N = 311 cells from 3 mice for control, and N = 307 cells from 3 mice for MACF1-knockout brains.

Requirement of MACF1 for the Formation and Differentiation of Leading Processes in Cortical Interneurons

Cortical interneurons tangentially migrate by extending a dynamic leading process, and abnormal leading processes has been repeatedly linked to migration deficits (Marin et al. 2010; Valiente and Marin 2010; Wang et al. 2011). Thus, we examined the morphology of interneuron leading processes at E19. Most control interneurons formed a single leading process with a typical single branch (Fig. 6A,B). However, MACF1-deleted neurons often developed multiple short leading processes. The length of leading processes was drastically reduced in MACF1-deleted interneurons by 92%, compared with control cells, while the number of leading processes was increased by 140% (Fig. 6C). These findings showed the requirement of MACF1 for the formation of normal leading process in cortical interneurons and suggest that MACF1 regulates interneuron migration in part via leading processes development.

Figure 6.

MACF1 is required for leading process growth and differentiation in GABAergic interneurons. (A) MACF1 is necessary for leading process formation. Brain sections from E15 control and MACF1^loxP/loxP; Dlx5/6-CIE mice were immunostained with an anti-GFP antibody. MACF1-knockout mice developed abnormal leading processes in migrating interneurons. Scale bar, 20 μm. (B) Representative morphologies of control and MACF1-knockout GABAergic interneurons with their leading processes. (C) Quantification of lengths and numbers of leading processes in control and MACF1-knockout GABAergic interneurons. MACF1-knockout GABAergic interneurons exhibited a decreased length, but an increased number of leading processes. N = 162 cells from 3 mice for control, and N = 183 cells from 3 mice for MACF1-knockout brains. Statistical significance was determined by two-tailed Student's t-test. ***P < 0.001. (D) MACF1-knockout inhibits neurite outgrowth and arborization of GABAergic interneurons in the developing cortex. Brain sections of control (MACF1^loxP/+; Nkx2.1-Cre; Gad67-EGFP) and MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP mice at P14 were immunostained with an anti-GFP antibody. MACF1-knockout interneurons exhibited a decreased number and length of neurites compared with controls. Scale bar, 20 μm. (E) Representative morphologies of control and MACF1-knockout interneurons. (F) Quantification of numbers and lengths of neurites in control and MACF1-mutant interneurons. N = 68 cells from 3 mice for control, and N = 81 cells from 5 mice for MACF1-knockout brains. Statistical significance was determined by two-tailed Student's t-test. *P < 0.05; ***P < 0.001.

Leading processes of migrating neurons differentiate into dendrites following the completion of migration during development (Marin et al. 2010). The abnormal leading processes of MACF1-deleted interneurons led us to examine the neurites of postnatal interneurons in control (MACF1^loxP/+; Nkx2.1-Cre; Gad67-EGFP) and MACF1 mutants (MACF1^loxP/loxP; Nkx2.1-Cre; Gad67-EGFP). Control interneurons developed multiple branched neurites, whereas MACF1-deleted neurons showed unbranched short neurites (Fig. 6D,E). The number of primary neurites was decreased by 30% in MACF1-deleted neurons compared with control neurons (Fig. 6F). Also, the length of neurites was decreased by 64% in MACF1-mutant neurons. These results indicate that MACF1 plays an important role in the formation of leading processes during embryonic development and their differentiation at postnatal stages.

Loss of MACF1 Decreases the Number of Interneurons in the Hippocampus

Proper hippocampal circuitry is required for maintaining the relationship between pyramidal neurons and interneurons. Almost all hippocampal interneurons originate from the ganglionic eminences and migrate tangentially to the hippocampus (Pleasure et al. 2000). The function of MACF1 in cortical interneuron migration strongly suggests that the protein might play a similar role during hippocampal development. Thus, we examined interneurons in control (MACF1^loxP/+; Dlx5/6-CIE) and knockout (MACF1^loxP/loxP; Dlx5/6-CIE) hippocampi at P14. EGFP-positive interneurons were severely reduced in the MACF1-mutant hippocampus (Fig. 7A). The numbers of interneurons in the CA1, CA2, and CA3 regions in MACF1-mutant mice were decreased by 45%, 51%, and 82%, respectively, compared with controls (Fig. 7B). Likewise, the number of interneurons in the dentate gyrus was reduced by 94% in MACF1 mutants, compared with controls (Fig. 7A,B).

Figure 7.

MACF1 regulates the number and positioning of hippocampal GABAergic interneurons. (A) Interneuron positioning patterns of control and MACF1^loxP/loxP; Dlx5/6-CIE mice. Brain sections of control (MACF1^loxP/+; Dlx5/6-CIE) and MACF1^loxP/loxP; Dlx5/6-CIE mice at P14 were immunostained with an anti-GFP antibody. MACF1-mutant brains exhibited decreased interneurons in the hippocampus compared with controls. Scale bar, 250 and 20 μm. (B) Quantification of the number of interneurons in different areas of control and MACF1-mutant hippocampi. N = 1887 cells from 5 control mice, and N = 1589 cells from 5 MACF1-mutant mice. Statistical significance was determined by two-tailed Student's t-test. ***P < 0.001.

MACF1 Controls Microtubule Stability and Dynamics in Cortical Interneurons

We assessed microtubule stability and dynamics by tracing intracellular microtubules. We first transfected control (MACF1^loxP/loxP; Nkx2.1-Cre) and MACF1^loxP/loxP; Nkx2.1-Cre neurons with a plasmid encoding EMTB-3XGFP. Polymerized microtubules can be visualized by overexpression of the EMTB-3XGFP construct (Miller and Bement 2009). We measured the extent of microtubule rearrangement and stability using time-lapse imaging to comparing the microtubule cytoskeleton at adjacent time points. The images from adjacent time frames with 3-min intervals were superimposed to measure microtubule stability. In control neurons, polymerized microtubules were almost completely superimposed between neighbor frames indicating stability within the defined time period (Fig. 8A). However, MACF1-deleted neurons exhibited multiple time points without overlap between two superimposed images within short time frames (arrows in Fig. 8A). We further investigated the role of MACF1 on microtubule stability by treating cultured interneurons with blebbistatin. This pharmacological agent enhances neurite outgrowth by stabilizing microtubules (Rosner et al. 2007). Interneuron progenitors from MACF1^loxP/loxP; Dlx5/6-CIE MGEs at E14 were cultured in the absence or presence of blebbistatin for 3 days, and then dendrite number and length were assessed using immunostaining with a MAP2 antibody. Blebbistatin did not alter the number of primary dendrites, but it increased the length of dendrites by 197% in MACF1-deleted neurons, compared with untreated neurons (Fig. 8B, C). These results suggest that MACF1-mediated stabilization of microtubules may contribute to the migration and differentiation of cortical interneurons.

Figure 8.

Deletion of MACF1 disrupts microtubule stability and severing in developing GABAergic interneurons. (A) MACF1-knockout neurons showed faster dynamics of microtubule structures and increased microtubule severing. A plasmid encoding EMTB-3XGFP was transfected into control and MACF1^loxP/loxP; Nkx2.1-Cre neurons. Then, microtubule structures labeled by EMTB-3XGFP were traced by live cell imaging at 3-min intervals. Images from adjacent time intervals were superimposed to assess the structural changes of the microtubule cytoskeleton between different time points of observation. Any overlap of the microtubule cytoskeleton at different time points would indicate greater microtubule stability and fewer overall dynamic changes in microtubule structures. Arrows indicate microtubule severing. (B) Cortical neurons from MACF1^loxP/loxP; Dlx5/6-CIE brains were cultured at E14 and treated with 25 µM blebbistatin at 0 day in vitro (DIV). GFP-positive interneurons were observed at 3 DIV. Scale bar, 25 μm. (C) Quantification of the number and length of dendrites in MACF1-knockout interneurons. N = 3 independent primary cortical cultures using 3 mice for each condition. Statistical significance was determined by two-tailed Student's t-test. ***P < 0.001.

Discussion

Using in vivo mouse genetics, we have demonstrated a critical role for MACF1 in the tangential migration of cortical interneurons in the developing brain. Elimination of MACF1 in GE-derived interneurons leads to a decreased number of interneurons in the cerebral cortex and hippocampus. MACF1-deleted interneurons undergo an abnormal transition from tangential to radial migration after entering the cerebral cortex. The defective tangential migration in MACF1-deleted neurons appears to be caused by abnormal leading process morphology and microtubule instability. Our findings reveal a novel function of a cytoskeleton modulator in interneuron migration and suggest a potential pathogenic mechanism for neurodevelopmental disorders associated with interneuron positioning abnormalities, including schizophrenia and autism.

MACF1 is highly expressed in the nervous system during development (Chen et al. 2006; Ka et al. 2014b). Previous studies have shown that MACF1 regulates the radial migration of pyramidal neurons in the developing brain (Goryunov et al. 2010; Ka et al. 2014b). In the current study, we observed an enrichment of MACF1 in tangential migratory streams. This led us to question whether radial migration and tangential migration could both be regulated by similar mechanisms involving MACF1. Our data demonstrate that MACF1 indeed regulates the tangential migration of cortical interneurons, strongly suggesting that both modes of neuronal migration share common molecules and mechanisms. MACF1^loxP/loxP; Dlx5/6-CIE and MACF1^loxP/loxP; Nkx2.1-Cre brains have a decreased number of interneurons in the dorsomedial cortex, but show no significant difference in the total number of interneurons in the cerebral cortex, compared with controls. Interestingly, many MACF1-deleted interneurons appear to stall in the lateral portion of the cerebral cortex. These observations may exclude the involvement of MACF1 in the generation and survival of interneurons.

Pyramidal neurons commonly extend a single leading process oriented towards the direction of their final destination in the pia (Marin et al. 2010). By contrast, tangentially migrating interneurons exhibit leading processes with a single branch (Bellion et al. 2005; Martini et al. 2009). Our data indicate the MACF1 regulates the extension of these single, branched leading processes in tangentially migrating interneurons. We observed that MACF1 deletion causes interneurons to develop multiple short leading processes instead of the usual single, branched leading process. This is distinct from what we previously observed in MACF1-deleted pyramidal neurons, which predominantly extend a single, short leading process without branches (Ka et al. 2014b). Unusual morphology of aberrant leading process has been repeatedly linked to migration deficits (Marin et al. 2010; Valiente and Marin 2010; Wang et al. 2011), and centrosomal movement along the leading process is important to the migration of neurons (Higginbotham and Gleeson 2007; Ka et al. 2014b). The short leading processes in MACF1-deleted interneurons may prevent centrosomal movement from creating enough tension to pull soma along the migratory pathway. We have also shown that MACF1 controls switching from tangential to radial migration after entering the dorsal cerebral cortex. A lower percentage of MACF1-deleted interneurons maintained a tangential orientation in the cortex, and thus a higher percentage of mutant interneurons exhibited a radial orientation, compared with control neurons. This result suggests that MACF1-deleted interneurons lack proper temporal or spatial regulation over switching from tangential to radial migration. Thus, the mutant interneurons may prematurely exit the tangential stream before reaching their final destination, resulting in depleted interneuron numbers in the dorsomedial cortex.

Cortical interneurons play multiple roles in the establishment of structural and functional networks during development. They are also important in postnatal neural plasticity. Aberrant development and migration of GABAergic neural circuits has been implicated in various neurodevelopmental and psychiatric disorders such as schizophrenia and autism (Belmonte et al. 2004; Dani et al. 2005; Lewis et al. 2005). As an interacting partner of Disrupted in Schizophrenia (DISC1), MACF1 has been widely associated with schizophrenia and autism (Camargo et al. 2007; Costas et al. 2013; An et al. 2014; Kenny et al. 2014). Studies have shown that patients with these developmental disorders show abnormal numbers of parvalbumin (PV)-positive interneurons in the cerebral cortex (Kenny et al. 2014; Stansfield et al. 2015). Measurements of gamma oscillation have also revealed aberrant activity of PV-positive interneurons in those patients compared with normal individuals (Bartos et al. 2007). Given that MACF1 determines PV-positive interneuron number and positioning in the developing cortex, it will be important to examine whether MACF1 plays a role in the pathogenesis of schizophrenia and autism. Dendritic abnormality in interneurons is another feature of the most consistent pathologic correlates in neurodevelopmental disorders including autism and schizophrenia (Levitt et al. 2004; Yip et al. 2008; Rossignol 2011). Our data show that MACF1 is required for neurite outgrowth and branching in interneurons, supporting a potential pathogenic role of MACF1 in these neurological conditions. We found that MACF1 deletion dramatically reduces the number of interneurons in the hippocampus. The hippocampus is a major participant in learning and memory, and abnormalities of hippocampal function have been related to schizophrenia (Uhlhaas 2013; Heckers and Konradi 2015). For instance, a decreased number of interneurons in the CA2 region of the hippocampus has been associated with schizophrenia and manic depression (Benes et al. 1998). In mice, cortical interneurons are produced in the ventral telencephalon and tangentially migrate to the cerebral cortex. However, studies suggest that an additional subset of cortical interneurons is generated locally in human and nonhuman primate cortices (Letinic et al. 2002; Petanjek et al. 2009; Jakovcevski et al. 2011; Radonjic et al. 2014; Clowry 2015). This subset of interneurons may form evolutionary novelty to higher cognitive functioning in human and primates. It remains to be elucidated if MACF1 also regulates local interneurons in the human and primate cerebral cortex.

In conclusion, our work suggests a critical role for MACF1 in cortical interneuron migration, positioning, and differentiation in the developing brain. We have previously shown that MACF1 is required for radial migration of pyramidal neurons in the developing mouse brain (Ka et al. 2014b). Therefore, at least some aspects of different types of neuron migration appear to be regulated by the same molecules and same mechanisms. The differences in taking migratory directions and selecting pathways by migrating neurons may be regulated by surface-mediated interactions that specify either a radial (gliogenic) or tangential (neurophilic) mode of migration. The role in interneuron development and association with DISC1 imply that MACF1 may participate in the underlying pathogenic mechanism of neurodevelopmental and psychiatric disorders.

Author Contributions

M.K. and W-Y.K. conceived, designed, performed, and analyzed the study. J.J.M. performed experiments. W-Y..K. supervised the work. M.K. and W-Y.K. wrote the paper.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oxfordjournals.org/

Funding

Research reported in this publication was supported by an award from the National Institute of Neurological Disorders and Stroke of the National Institute of Health under award number R01NS091220 and an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number P20GM103471 to W.Y.K.