Characterization and Stage-Dependent Lineage Analysis of Intermediate Progenitors of Cortical GABAergic Interneurons

Intermediate progenitors of both excitatory and inhibitory neurons, which can replenish neurons in the adult brain, were recently identified. However, the generation of intermediate progenitors of GABAergic inhibitory neurons (IPGNs) has not been studied in detail. Here, we characterized the spatiotemporal distribution of IPGNs in mouse cerebral cortex. IPGNs generated neurons during both embryonic and postnatal stages, but the embryonic IPGNs were more proliferative. Our lineage tracing analyses showed that the embryonically proliferating IPGNs tended to localize to the superficial layers rather than the deep cortical layers at 3 weeks after birth. We also found that embryonic IPGNs derived from the medial and caudal ganglionic eminence (CGE) but more than half of the embryonic IPGNs were derived from the CGE and broadly distributed in the cerebral cortex. Taken together, our data indicate that the broadly located IPGNs during embryonic and postnatal stages exhibit a different proliferative property and layer distribution.

Intermediate progenitors of both excitatory and inhibitory neurons, which can replenish neurons in the adult brain, were recently identified. However, the generation of intermediate progenitors of GABAergic inhibitory neurons (IPGNs) has not been studied in detail. Here, we characterized the spatiotemporal distribution of IPGNs in mouse cerebral cortex. IPGNs generated neurons during both embryonic and postnatal stages, but the embryonic IPGNs were more proliferative. Our lineage tracing analyses showed that the embryonically proliferating IPGNs tended to localize to the superficial layers rather than the deep cortical layers at 3 weeks after birth. We also found that embryonic IPGNs derived from the medial and caudal ganglionic eminence (CGE) but more than half of the embryonic IPGNs were derived from the CGE and broadly distributed in the cerebral cortex. Taken together, our data indicate that the broadly located IPGNs during embryonic and postnatal stages exhibit a different proliferative property and layer distribution.
Neural progenitors (also referred to as apical progenitors or radial glia) are located along the ventricles, and apical progenitors of the excitatory neurons produce intermediate progenitors, which divide at the subventricular zone (SVZ) of the dorsal telencephalon to generate two neurons (Noctor et al., 2004;Wu et al., 2005;Pilz et al., 2013). Recently, it was reported that apical progenitors of the inhibitory neurons also produce intermediate progenitors (Wu et al., 2011;Hansen et al., 2013;Petros et al., 2015). Lineage tracing studies have revealed that the Nkx2-1-lineage progenitor cells undergo symmetric and asymmetric division in the SVZ of the MGE (Brown et al., 2011;Ciceri et al., 2013;Sultan et al., 2014). However, the intermediate progenitors of IPGNs have not yet been well characterized. In primates, including humans, many IPGNs are observed at the cortical SVZ as well as the SVZ of the ganglionic eminence at embryonic stages (Letinic et al., 2002;Zecevic et al., 2005Zecevic et al., , 2011Petanjek et al., 2009;Jakovcevski et al., 2011;Yu and Zecevic, 2011), suggesting broad expansion of IPGNs in primates. These IPGNs in the cortical SVZ derive mainly from the outer SVZ of the ganglionic eminence (Hansen et al., 2013;Ma et al., 2013). The CGE, in particular, contributes to a large number of the cortical GABAergic neurons in humans (Hansen et al., 2013).
By contrast, we and others have observed a small number of IPGNs in the cortical-SVZ in embryonic and postnatal mouse brains (Inta et al., 2008;Wu et al., 2011). These IPGNs express neuronal and proliferative markers at the perinatal stage in glutamic acid decarboxylase 67 (GAD67)-GFP knock-in mice, which are widely used to visualize GABAergic neurons and IPGNs. Our previous study revealed that ∼1.5% of the perinatal cortical GAD67-GFP-positive cells self-renew and produce a small number of cortical GABAergic neurons (Wu et al., 2011). Furthermore, a small number of cortical GABAergic neurons are generated in the cortical-SVZ (Inta et al., 2008) and early postnatal dorsal white matter (Riccio et al., 2012). Thus, unlike the intermediate progenitors of the excitatory neurons, IPGNs are widely dispersed from the subpallium to the cerebral cortex. However, the distribution patterns of IPGNs and their fates in adult cerebral cortex are poorly understood.
In this study, we performed systematic analyses of spatiotemporal distribution patterns of IPGNs in mice. Our lineage tracing analyses indicated that the laminar distributions of the embryonic and postnatal IPGNs differ. Late embryonic IPGNs tended to differentiate into reelinpositive (Reln + ) or vasoactive intestinal peptide positive (Vip + ) GABAergic interneurons but were also able to differentiate into parvalbumin-positive (Pvalb + ) or somatostatin-positive (Sst + ) neurons, suggesting that the IPGNs observed at the late embryonic stage are mainly derived from the CGE, rather than the MGE. Consistently, IPGNs were found primarily in the CGE-SVZ but also the MGE-SVZ and, to lesser extent, in the mantle zone of the ganglionic eminence and cortical-SVZ/intermediate zone (IZ) at embryonic stages. These findings indicate that IPGNs exhibit a unique spatiotemporal distribution and significantly contribute to several subtypes of GABAergic interneurons in adult cerebral cortex.

Production of GAD67-CrePR Knock-in Mouse
Progesterone inducible Cre recombinase (CrePR) was generated to fuse CrePR and progesterone receptor ligand binding motif genes (Kellendonk et al., 1996;Kitayama et al., 2001). To produce GAD67-knock-in CrePR mice, we designed a targeting vector in which a CrePR recombinase gene was inserted into the translational initiation site of the GAD67 gene in frame (Supplementary Figure 1). A knock-in vector pGAD67CrePRTV contained a 3 kb fragment at the 5 side, a CrePR gene placed behind the GAD67 translational start, a Pgk-neo-p(A) cassette flanked by two Flp recognition target (frt) sites, a 7 kb fragment at the 3 side, and an MC1 promoter-driven diphtheria toxin gene. Culture of embryonic stem (ES) cells and generation of chimeric mice were performed as described previously (Kitayama et al., 2001;Higo et al., 2009). Briefly, a linearized pGAD67CrePRTV was introduced into C57BL/6 mouse ES cells (RENKA) and then, G418resistant clones were picked up. Homologous recombined ES clone was identified by Southern blotting. To produce germline chimera, the selected ES cells were microinjected into eight cell-stage embryos of the CD-1 mouse strain. The germline chimeras of GAD67-CrePR mice were crossed with C57BL/6 mice to generate the GAD67-CrePR mouse line. Because the knock-out of both GAD67 alleles is lethal at birth (Asada et al., 1997), mice heterozygous for the altered GAD67 allele were used for all the observations in this study. Genotypes were identified by Southern blot hybridization or PCR. Tail genomic DNA was digested with Spe-I or Afl-II and hybridized with a 5 probe or a 3 probe, respectively. PCR was performed with specific 3 . The sequence of each primer and the approximate length of the amplified DNA fragments are described as follows: Gad1CrePR, g67-2 (5 -TTCCGGAGGTACCACACCTT-3 ), g67-5 (5 -TAAGTCGACGCTAGCGAGCGCCTCCCCA-3 ), and CreR1 (5 -TTGCCCCTGTTTCACTATCC-3 ); wild type, 1.8 kbp; mutant: 1.4 kbp.

Animals
Mice were housed and treated in strict accordance with the rules for animal care and use for research and education of Kumamoto University. They were kept in cages and exposed to a 12 h light/dark cycle with food and water provided ad libitum. GAD67-CrePR mice express a mifepristone-inducible CrePR under the control of the GAD67 gene. To label GAD67 + cells, 0.05 mg/g body weight of mifepristone (Sigma-Aldrich, St. Louis, MO, United States) was administered to pregnant GAD67-CrePR mice, or 0.05 mg/g body weight mifepristone was administered to postnatal day 0 pups by intraperitoneal (i.p.) injection. Mifepristone was dissolved in 100% ethanol to 50 mg/mL and then diluted in corn oil (Wako, Japan) to 5 mg/mL. To prevent abortions in the pregnant mice, progesterone dissolved in corn oil (20 mg/mL, stored at 4 • C; Sigma-Aldrich, St. Louis, MO, United States) was administered at 0.15 mg/g body weight once a day between E12 and E18 (Rose et al., 2009). To detect Cre-mediated recombination in GAD67-CrePR knock-in mice, mice that express TdTomato reporter upon Cre-mediated recombination were used (Jackson stock no. 007909 with an ICR background) (Madisen et al., 2010). GAD67-CrePR knock-in male mice were mated with R26-TdTomato (Ai9) reporter female mice or ICR mice (Oriental Yeast, Japan). As a previous study showed that CreER-mediated reporter recombination is initiated 6 h after tamoxifen administration, intermediate progenitors of IPGNs and GABAergic neurons are sufficiently labeled with TdTomato 24 h after the injection (Zervas et al., 2004). RFP antibody were utilized for enhanced visualization of TdTomato expression. Newborn pups were allowed to develop up to 3 weeks. Pregnant mice and juvenile mice (3 weeks) were anesthetized as described below and perfused with a fixative for immunohistochemistry.

Fixation
The mice were anesthetized with sodium pentobarbital (50 mg/kg body weight, i.p.) and perfused transcardially with 10 mL PBS [0.9% (w/v) saline buffered with 5 mM sodium phosphate, pH 7.4], followed by 50 mL PBS containing 4% (w/v) formaldehyde. The brains were removed, postfixed with the same fixative for 12 h, and then cryoprotected in 25% sucrose in PBS overnight. The brains were sliced into 50 µm-thick coronal sections on a cryostat.

Immunohistochemistry
Immunohistochemistry was performed on the brains of male and/or female mice. Immunohistochemistry was performed as described previously (Tamamaki et al., 2003;Miyamoto and Fukuda, 2015). Briefly, brain sections were rinsed in PBS several times and incubated overnight at room temperature with primary antibody diluted in incubation buffer (0.3% Triton X-100, 5% donkey serum, and 0.01% sodium azide in PBS). The primary antibodies used in this experiment are listed in

EdU Labeling
To detect proliferative IPGNs, EdU, an analog of the nucleoside thymidine that incorporates into replicating DNA was injected to the GAD67-CrePR;Ai9 mice 24 h after the mifepristone administration. This inducible genetic fate mapping method enables labeling of postnatally proliferative GAD67 + IPGNs and tracing of their lineage in the mature brain. Direct labeling of EdU was performed after incubations with primary antibody. Tissue sections were incubated with 1% bovine serum albumin in PBS for 1 h and then with a freshly prepared cocktail containing 50 mM Tris (pH 7.4), 150 mM NaCl 2 , 2 mM CuSO 4 , 10 µM Alexa Fluor 647-azide (from 1 mM stock in dimethyl sulfoxide, A10277; Thermo Fisher, Waltham, MA, United States), and 10 mM sodium ascorbate (added last from 0.1 M stock) in the dark for 2 h. Sections were washed in PBS three times (protocol modified from that described by Hua and Kearsey (2011)]. Then, sections were incubated with secondary antibodies.

Quantifications and Statistical Analysis
EdU-labeled and immunoreactive cells were counted under a confocal microscope (Nikon C2) individually by capturing threedimensional images at a 1-2 µm thickness. Total thickness of three-dimensional images in the depth direction were more than 15 µm. More than 8 hemisphere of cortical slices and more than 50 cells were analyzed from a brain in each experiment, respectively. Data are presented as means ± standard errors (SEs). Statistical significance was determined by two-sided unpaired Student's t-tests, with a P-value of <0.05 considered statistically significant.

A Small Proportion of GAD67 + Intermediate Progenitors of GABAergic Inhibitory Neurons (IPGNs) Proliferate in Perinatal Cerebral Cortex
We previously reported the presence of a small number of IPGNs in the SVZ and even in the IZ of the cerebral cortices of GAD67-GFP knock-in mice at the perinatal stage (Wu et al., 2011). However, the number of postnatally generated IPGNs in the cortical-SVZ (or cortical-IZ) in rodents is much lower than that in primates, and it is not known whether IPGNs survive and mature in adult rodent cortex.
To assess the distribution of cortical GABAergic neurons generated postnatally from IPGNs, we utilized GAD67-CrePR mice, which expressed CrePR, a mifepristone-inducible Cre, in GABAergic neurons and IPGNs ( Figure 1A). To detect the initial population of postnatally proliferating IPGNs in the cortex, we performed immunohistochemistry using GAD67-CrePR;Rosa26-floxed-TdTomato (Ai9 line) mice at postnatal day 1 (P1) that had been administered mifepristone at P0 and EdU 30 min before fixation ( Figure 1B). GAD67 + TdTomato + cells were broadly distributed throughout the brain, including the cerebral cortex. Although EdU + cells were mainly located in the neurogenic niches, they could be found throughout the brain, including the superficial layers of the cortex (Figures 1C-E).

A Small Proportion of GAD67 + Intermediate Progenitors of GABAergic Inhibitory Neurons (IPGNs) Proliferate and Survive in Cerebral Cortex During Perinatal Development
To compare the fates of proliferative GABAergic neuron progenitors between late embryonic stage and the perinatal stage, we performed a fate mapping analysis. To detect the proliferative GABAergic neuron progenitors during late embryonic stages, EdU was injected into GAD67-CrePR;Ai9 mice from E14.5 to E18.5 (late embryonic stage) or from P1 to P5 (perinatal stage) with mifepristone administered at P0 to label embryonically or postnatally proliferative IPGNs, respectively (Figure 2A). At 3 weeks after birth, TdTomato + cells were broadly distributed throughout the brain, including the cerebral cortex (Figures 2B,C), hippocampus, striatum, thalamus, hypothalamus, midbrain, and cerebellar cortex (Supplementary Figure 2). TdTomato + cells were observed in each cortical layer but tended to distribute in the superficial layers (i.e., layers I and II/III) (Figures 2B-D).
To further characterize these EdU + and TdTomato + doublepositive GAD67-lineage cells in each layer, we performed immunohistochemical analyses with EdU. When labeled from P1 to P5, EdU + cells were broadly distributed throughout the cortex. By contrast, when labeled at the late embryonic stages (from E14.5 to E18.5), EdU + cells tended to populate the superficial layers (i.e., layers I and II/III) ( Figure 2C). Approximately 51% [51.2 ± 2.8% (total 336/657 cells from three brains)] of the TdTomato + GAD67-lineage cells (labeled at P0) were colabeled with EdU that was injected from E14.5 to E18.5 (Figures 2F-H). By contrast, approximately 1.5% [1.5 ± 1.2% (13/1,046 cells from six brains)] of the TdTomato + GAD67-lineage cells (labeled at P0) were colabeled with EdU that was injected at P1-P5 (Figures 2F-H and Table 2). These results demonstrate that a very small number of postnatal proliferative IPGNs were represented in perinatal cortex by our fate mapping study. Thus, the proliferative rates of the GABAergic neuron progenitors during late embryonic stage were nearly 34-fold higher than that of the postnatally proliferative IPGNs (labeled at P1-P5), indicating that significantly more GABAergic neurons are produced during embryonic stages. Moreover, these GAD67 + embryonic proliferating cells (EdU labeled at E14.5-E18.5) were more abundant in the superficial cortical layers (i.e., layers II/III) than in the deep layers (Figures 2F,G and Table 2). These results delineate the fate of later-born embryonic GABAergic neurons and postnatally proliferative IPGNs.

Embryonic Proliferation of GAD67 + Intermediate Progenitors of GABAergic Inhibitory Neurons (IPGNs)
Our data revealed a small number of postnatally proliferative IPGNs (Figures 2, 3). Moreover, the cortical GABAergic neurons that derived from postnatally proliferative IPGNs survived into later postnatal development. On the other hand, significantly more GABAergic neurons that were derived from the GABAergic neuron progenitors were produced during embryonic stages. However, the proportion of the GABAergic neurons derived from the IPGNs in each embryonic stage was unclear. To define the distribution patterns of embryonically proliferating IPGNs, we analyzed E14.5, E16.5, and E18.5 GAD67-CrePR;Ai9 mice that were administered mifepristone at E13.5 to label GAD67 + cells, which were sacrificed 30 min after EdU injection (Figure 4). At E14.5, 3.5 ± 0.4% of the TdTomato + cells were EdU + (27/793 cells from three brains). At E16.5, 0.6 ± 0.3% of the TdTomato + cells were EdU + (4/672 cells from three brains). At E18.5, 0.8 ± 0.2% of the TdTomato + cells were EdU + (5/584 cells from three brains). We found that approximately half of the GAD67-lineage TdTomato + cells migrated to the cortex and the other half were in the M/CGE-mantle zone and M/CGE-SVZ at E14.5 (Figures 4A,B). These cells migrated tangentially toward the cerebral cortex via two major migratory streams, a superficial route through the cortical-marginal zone and a deeper route through the cortical-SVZ/IZ, consistent with previous reports (Lavdas et al., 1999;Wichterle et al., 2001;Tanaka et al., 2006Tanaka et al., , 2009Miyoshi and Fishell, 2011;Lim et al., 2018). Furthermore, 3.5% of EdU + GAD67-lineage TdTomato + double-positive cells were observed in the MGE-SVZ/IZ or M/CGE-mantle zone at E14.5 (Figures 4A,B). These results suggest that ∼3.5% of the embryonic proliferating IPGNs mainly populated the M/CGE-SVZ/IZ or M/CGE-mantle zone at E14.5. However, no GAD67-lineage TdTomato + cells were found in the M/CGE at E18.5. Therefore, migration of the E13.5labeled GAD67-lineage-TdTomato + cells from the M/CGE to the cerebral cortex had already finished by E18.5. In addition, a very small proportion (<1%) of the GAD67-lineage TdTomato + cells colabeled with EdU in the cortex in the cortical plate and cortical-SVZ/IZ at E16.5 and E18.5 (Figures 4C-F). These results suggest that embryonic brains have multiple regions for embryonic IPGN proliferation. Furthermore, a substantial proportion of the GAD67 + IPGNs actively proliferate at around E14.5, and their proliferative activities decline with tangential migration to the cortex.
Cortical GABAergic neurons can be classified as Pvalb + , Sst + , Reln + , and Vip + (Miyoshi et al., 2010;Brandão and Romcy-Pereira, 2015). Whereas Sst + neurons are produced during the early stage of neurogenesis, a large proportion of the Pvalb + neurons are generated in the MGE during the late stages (Rymar and Sadikot, 2007). On the other hand, the CGEderived GABAergic neurons are produced in late embryonic stages and mainly express either Reln or Vip (Nery et al., 2002;Miyoshi et al., 2010;Rudy et al., 2011;Vitalis and Rossier, 2011). Reln + neurons are derived from both the CGE and MGE, as subpopulations of the MGE-derived Sst + neurons coexpress Reln (Miyoshi et al., 2010;Miyoshi and Fishell, 2011). Recently, a number of the CGE-derived GABAergic neurons, which express Reln and Vip, were found to populate superficial cortical layers independently of their birthdate (Miyoshi et al., 2010), and their peak production was during late embryonic stages (Miyoshi and Fishell, 2011). In primates, the IPGNs (basal/non-epithelial progenitors) derived from CGE contribute to a large number of cortical GABAergic neurons (Hansen et al., 2013). These previous reports indicate that the specification of the GABAergic neuron subtypes depends on their birthdates and localization.
Our results indicate that more than half of cortical IPGNs during late embryonic stage are derived from the CGE. This tendency may correspond to primate cortical GABAergic neuron development.
Several lines of evidence indicate that GABAergic neuron progenitors in the MGE produce both Pvalb + and Sst + cortical GABAergic neurons (Brown et al., 2011;Ciceri et al., 2013;Harwell et al., 2015;Mayer et al., 2015). Sst + GABAergic neurons are principally produced from early-stage GABAergic neuron progenitors (∼E12.5), whereas Pvalb + GABAergic neurons are continuously produced throughout development (Miyoshi et al., 2007;Wonders et al., 2008;Inan et al., 2012). However, these reports did not distinguish apical (epithelial) progenitors that divide in the ventricular zone (VZ) from IPGNs (basal/nonepithelial progenitors) that divide in the SVZ (Pilz et al., 2013). Basal/non-epithelial progenitors in the SVZ of the MGE mainly produce Pvalb + rather than Sst + cortical GABAergic neurons (Petros et al., 2015). Accordingly, our immunohistochemical analyses showed that GAD67 + IPGNs that derive from MGE (Pvalb + or Sst + ) and travel along the migratory stream to the cortex in late embryonic stages (E14-E18) tend to become Pvalb + neurons, although a small proportion become Sst + neurons. Thus, GAD67 + IPGNs in embryonic stages derived from the MGE are most likely to have and retain characteristics of latestage basal progenitors.
Cortical GABAergic neurons are produced from cortical layer 1 GABAergic neuron progenitor cells (L1-INP cells) (Ohira et al., 2010), and the generation of L1-INP cells increases under ischemic conditions and after treatment with fluoxetine, a selective serotonin reuptake inhibitor (Ohira et al., 2013). We observed that a small number of postnatal proliferative GAD67 + IPGNs that were labeled after mifepristone administration at E13.5 and P0 survived throughout postnatal development and tended to populate the superficial cortical layers (Figures 2G, 5G). These results suggest that GAD67 + IPGNs are maintained in the cortex and may act as L1-INP cells in adult brain.
In conclusion, the present study demonstrates that GAD67 + IPGNs, which are considered late-stage basal progenitors of cortical GABAergic neurons, substantially contribute to the GABAergic neuron population in adult cerebral cortex. In addition, the proliferative rates of IPGNs and the laminar distributions of their progenies change during cortical development. Thus, the characteristic features of IPGNs are spatiotemporally regulated during brain development.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

ETHICS STATEMENT
The animal study was reviewed and approved by the rules for animal care and use for research and education of Kumamoto University.