Astrocytes in the Gli1 lineage are broadly distributed in the cortex
We first examined the expression of Gli1 in the early postnatal mouse cortex in Gli1nlacZ/+ mice carrying a nuclear lacZ in the Gli1 locus . At postnatal day 0 (P0), there was a pronounced population of Gli1-expressing cells in the ventral SVZ, corresponding to the population of SVZ precursors that generate deep granule neurons and periglomerular cells in the olfactory bulb [20, 22, 23] (Fig. 1). Gli1-expressing cells were also observed in the dorsolateral corner of the SVZ (dlSVZ), within the postnatal germinal zone of cortical astrocytes , and in a region inferior to the white matter overlying the ventricles (Fig. 1). Few cells were observed in the cortex at this age (Fig. 1). We quantified and mapped the distribution of βGal-labeled cells within the dlSVZ and cortex and found that the cortex harbored relatively few Gli1-expressing cells at P0 compared to the dlSVZ (Fig. 1).
Gli1-expressing progenitors residing in the subcallosal dorsolateral domain of the SVZ (dlSVZ) generate a large fraction of oligodendrocytes that populate the corpus callosum and overlying white matter tract . Retroviral labeling experiments demonstrate that this region also generates glial cells that populate the cortex . To determine whether Gli1-expressing progenitors also contribute glial cells to the cortex, we performed fate mapping in Gli1CreER/+ mice  carrying the Ai14 tdTomato (tdTom) reporter  (Gli1CreER/+;Ai14). Cre-mediated recombination promotes expression of tdTom that is both permanent and heritable. We marked Gli1-expressing precursors by administering tamoxifen to Gli1CreER/+;Ai14 mice at P0 and analyzed the distribution of tdTom at various postnatal ages. One day after tamoxifen (P1), there was a substantial population of marked cells in the cortex that were distributed across all layers (Fig. 1). At P3, there was a dramatic expansion of marked cells (Fig. 1), suggesting extensive proliferation between P1 and P3. We also observed many residual radial glial fibers and cells with transitional morphologies, consistent with radial glia undergoing transformation into multipolar astrocytes (Supplemental Fig. 1). These cells co-express vimentin, a marker of radial glia (Supplemental Fig. 1), suggesting that in addition to progenitors residing in the dlSVZ, Shh also signals to radial glial progenitors. There was a further expansion observed at P7 (Fig. 1), suggesting that Gli1-expressing cells marked at P0 correspond to actively dividing glial progenitor cells. There was no further expansion in the number of marked cells between P7 and P14, suggesting Gli1-expressing cells marked at birth proliferate largely during the first postnatal week. To confirm the observed expansion was due to proliferation, we administered BrdU to mice approximately 12 h after tamoxifen, to ensure sufficient Cre-mediated recombination prior to incorporation of BrdU. We analyzed tissues at P14 and found that 49% of marked cells in the cortex are co-labeled with BrdU (Fig. 1). Conversely, 54% of BrdU labeled cells co-expressed tdTom, indicating that only a fraction of proliferating glial precursors at P0 express Gli1 (Fig. 1). These data show that Gli1-expressing cells residing within the dlSVZ generate cells that migrate into the cortex and expand locally.
Gli1 progenitors generate a subpopulation of cortical astrocytes
We examined the morphologies of marked cells in the cortex and found that at early time points, marked cells showed simple morphologies, with one or two processes, consistent with an immature phenotype (Fig. 2). Over time, marked cells developed an increasingly complex morphology, extending multiple processes. By P14, cells exhibited several major primary branches together with many fine branchlets, conferring the typical bushy morphology of protoplasmic astrocytes (Fig. 2). We performed colocalization analysis with glial cell-type specific markers at P14. A small fraction of marked cells co-expressed the oligodendrocyte-specific marker, CC1 (Fig. 2), whereas the overwhelming majority were co-labeled with S100β, identifying them as astrocytes (Fig. 2). We also observed substantial expression of tdTom in the white matter overlying the ventricles in tissues marked as late as P7 (Figs. 1 and 4), consistent with Gli1-derived oligodendrocytes . These data demonstrate that neonatal progenitors in the dlSVZ correspond to astrocyte progenitor cells responsible for contributing cortical astrocytes. Although the fraction of marked cells corresponding to oligodendrocytes at P14 was small, this fraction increased at P28 and P60 (Fig. 2), consistent with oligodendrocyte production during late postnatal development . However the fraction of marked cells differentiating into oligodendrocytes remained small and never exceeded 14% (Fig. 2). Notably, marked oligodendrocytes in the cortex were only observed in mice that received tamoxifen at P0. Tamoxifen given to mice beyond P0 generated marked cells that were identified predominantly as astrocytes (Table 1). These data suggest that Gli1-expressing progenitor cells residing in the early postnatal dlSVZ correspond predominantly to astrocyte progenitors.
We next investigated the extent to which astrocytes in the Gli1 lineage contribute to the total astrocyte population in the mature cortex. We administered tamoxifen at P0 and analyzed the fraction of cells labeled with the pan-astrocytic marker, S100β, that co-express tdTom at P60. We found that 46% of astrocytes were co-labeled with tdTom (Fig. 3), indicating that astrocytes within the Gli1 lineage make up nearly half of the total cortical astrocyte population. This suggests that the cortex harbors a mixed population of astrocytes from different lineages. These data further suggest that Gli1-expressing astrocyte progenitor cells comprise a subpopulation of the total pool of astrocyte progenitors. Alternatively, these data could reflect limitations of tamoxifen-dependent Cre-mediated recombination. To rule out the possibility that recombination was inefficient within the Gli1-expressing progenitor population, we examined the dlSVZ of Gli1nlacZ/+ mice, enabling us to identify cells actively expressing Gli1 independent of the requirement for recombination. We identified the pool of astrocyte progenitors at P0 by Sox9 expression and found that 31% co-expressed Gli1 (Fig. 3), consistent with the idea that progenitor cells expressing Gli1 comprise a subpopulation of astrocyte progenitors in the P0 dlSVZ. This was confirmed with a second astrocyte progenitor marker, BLBP, in which we found that many BLBP labeled cells were Gli1 negative (Fig. 3). Taken together, these data suggest that the astrocyte progenitor pool in the early postnatal dlSVZ is comprised of a molecularly distinct subpopulation that can be defined by Shh signaling.
Shh signaling recurs in a subpopulation of differentiated astrocytes
Our fate mapping studies showed that astrocytes within the Gli1 lineage are distributed ubiquitously throughout all layers of the neocortex. In contrast, Shh activity in the mature cortex is found predominantly in astrocytes localized to layers IV and V [3, 25]. To determine the age at which the distinctive laminar pattern of Shh signaling in the cortex emerges, we administered tamoxifen over three consecutive days at various ages during postnatal development. Tissues were analyzed 2 weeks after the initial tamoxifen dose, with the exception of P7 animals which were analyzed at one and 3 weeks after tamoxifen. There was no difference between these chase periods and the data were pooled for P7. In mice that received tamoxifen at P3, marked cells were observed throughout all cortical layers, showing a similar distribution as that seen at P0 (Fig. 4). However by P7, fewer marked cells are found in the upper cortical layers than in tissues marked at younger ages (Fig. 4). At P14, marked cells are found predominantly in deep layers and are largely absent from superficial layers, a pattern that was consistent with that observed at P28 and beyond (Fig. 4). Single cell analysis with cell-type specific markers showed that the vast majority of marked cells labeled at all ages were subsequently identified as astrocytes (Table 1). Quantitative analysis of the fraction of marked cells in each layer shows that tamoxifen labeling at P0 produces marked cells throughout all layers, with a moderate bias towards layer 5, which harbors 30% of marked cells, whereas layer 1 harbors only 12%. In contrast, the vast majority of marked cells are found in deep layers when tamoxifen is administered to adult mice (Fig. 4). This shift in marked cell distribution across development was associated with a concomitant reduction in the fraction of astrocytes, identified by S100β, that are marked at these ages (Fig. 4). At P0, 49% of astrocytes across all layers are marked, but by P3, that proportion declined to 32%, though this was not significant (Fig. 4). At P7, the fraction of marked astrocytes declined significantly from P0 to 23%. This fraction remained steady at P14 and P28 (21 and 24%, respectively, Fig. 4). There was a further modest reduction to 15% in adults, though this was not significant (Fig. 4). Double labeling with BrdU administered 12 h after tamoxifen shows that the fraction of marked cells that are dividing declines dramatically over the first postnatal week. At P0, the fraction of marked cells double labeled with BrdU was 49%, whereas at P3, that fraction was significantly reduced to 33%. By P7, nearly all marked cells were postmitotic as only 3% of tdTom cells co-labeled with BrdU (Fig. 4). This suggests that astrocytes within the Gli1 lineage are generated predominantly during the first few days after birth.
Despite the progressive decline in the fraction of astrocytes that are marked during the first postnatal week, analysis of active Shh signaling in the cortex in Gli1nlacZ/+ mice shows a progressive increase over the first two postnatal weeks (Fig. 4). Stereological quantification of the number of βGal labeled cells in the cortex showed a significant increase in the number of cells at P14 compared to P3 (Fig. 4). There was no difference between P14 and the adult (>P90) cortex (Fig. 4), suggesting that Shh signaling in the cortex stabilizes by the second postnatal week. Despite the increase in the number of cells exhibiting Shh activity between P7 and adulthood, few cells marked at P7 are proliferating (Fig. 4), arguing against the possibility that the increase in Gli1-expressing cells between P7 and adulthood is due to proliferation of immature progenitor cells. These data suggest that Shh signaling in astrocyte progenitor cells residing in the SVZ is transient and is lost as cells migrate into the cortex and undergo maturation. In parallel, Shh activity in the cortex is low at birth, but increases during postnatal development in a subpopulation of mature, postmitotic astrocytes found mostly in deep layers, coincident with the localization of Shh-expressing neurons in layer V [26, 27]. Taken together, these data suggest that Shh activity in neonatal astrocyte progenitor cells declines in immature astrocytes, but recurs in a subpopulation of postmitotic astrocytes localized primarily in layers IV and V.
Recurrence of Shh signaling is independent of lineage
We next examined whether Shh signaling in differentiated astrocytes is restricted to those within the Gli1 lineage. We crossed Gli1CreER/+;Ai14 mice with Gli1nlacZ/+ mice (Gli1CreER/nlacZ;Ai14) to generate mice in which we could distinguish between temporally distinct populations of cells expressing Gli1. We reasoned that because tamoxifen administered at P0 will indelibly mark progenitor cells expressing Gli1 and their progeny, whereas βGal-expressing cells would reflect Gli1 activity at the conclusion of the experiment, this approach would enable us to identify individual cells showing differential Gli1 activity at two different time points within a single mouse. While this effectively produces a Gli1 null mouse, Gli1 is not required for Shh signaling during development and Gli1 null mice show no developmental or behavioral deficits . We administered tamoxifen to Gli1CreER/nlacZ;Ai14 mice at P0 and analyzed tissues at P60 for colocalization of the tdTom and βGal reporter proteins. Although cells marked at P0 are found throughout all cortical layers, this analysis was restricted to deeper layers where βGal-expressing cells are mostly found. At P60, a large proportion (67%) of tdTom-labeled cells did not co-express βGal (Fig. 5), suggesting that in a substantial fraction of astrocytes in the Gli1 lineage, Shh signaling is downregulated sometime after P0. The remaining fraction of tdTom-labeled cells were double labeled with βGal (Fig. 5), suggesting either persistent Shh activity since birth or recurrence of Shh signaling in the same cell. Interestingly, 40% of βGal-labeled cells did not co-express tdTom (Fig. 5), suggesting Shh activity in differentiated astrocytes is not restricted to cells within the Gli1 lineage. Consistent with this, a greater number of cells express Ptc than Gli1 in the adult cortex , indicating that Gli1 activity reflects a fraction of cells capable of transducing Shh signal. Taken together, these data suggest that Shh signaling in differentiated astrocytes reflects recurrence of Shh activity in mature cells that is independent of developmental lineage.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.