Neuroepithelial stem cell proliferation requires LIS1 for precise spindle orientation and symmetric division

Abstract

Mitotic spindle orientation and plane of cleavage in mammals is a determinant of whether division yields progenitor expansion and/or birth of new neurons during radial glial progenitor cell (RGPC) neurogenesis, but its role earlier in neuroepithelial stem cells is poorly understood. Here we report that Lis1 is essential for precise control of mitotic spindle orientation in both neuroepithelial stem cells and radial glial progenitor cells. Controlled gene deletion of Lis1 in vivo in neuroepithelial stem cells, where cleavage is uniformly vertical and symmetrical, provokes rapid apoptosis of those cells, while radial glial progenitors are less affected. Impaired cortical microtubule capture via loss of cortical dynein causes astral and cortical microtubules to be greatly reduced in Lis1-deficient cells. Increased expression of the LIS/dynein binding partner NDEL1 restores cortical microtubule and dynein localization in Lis1-deficient cells. Thus, control of symmetric division, essential for neuroepithelial stem cell proliferation, is mediated through spindle orientation determined via LIS1/NDEL1/dynein-mediated cortical microtubule capture.

Figure 1

(A–D) Heterozygous Lis1^+/hc; Pax2-Cre; Rosa26 and (E–H) homozygous Lis1^hc/hc; Pax2-Cre; Rosa26 embryos stained with X-gal at E9.5. Lateral (A, E) and top (B, F) views of whole mounts with blue stained pharyngeal arches (arrows) and midbrain (arrowheads) indicative of Cre expression. Mid-sagittal sections demonstrate the midbrain-hindbrain expression of Cre (C, G). An adjacent section (D, H) stained with DAPI (blue) and TUNEL (bright spots) demonstrates massive apoptosis in the Xgal-positive regions in homozygotes compared with heterozygous controls. (I–L) Heterozygous Lis1^+/hc; Wnt1-Cre; Rosa26 and (M–P) homozygous Lis1^hc/hc; Wnt1-Cre; Rosa26 embryos at E10.5 (I, J, M, N) and E12.5 (K, L, O, P), stained with X-gal. Arrow (N, P) denotes drastic loss of blue cells in the midbrain in the homozygote. (Q–T) Wild-type Lis1^+/+; hGFAP-Cre; Rosa26 and (U–X) homozygous Lis1^hc/hc; hGFAP-Cre; Rosa26 neonates (P3). (Q, U) Cresyl violet staining. (Q–T, V–X) Immunohistochemistry with anti-testis-1 (to mark CA1 of the hippocampus and layers 2 and 5 of the cortex). Note absence of the hippocampus in the homozygote (arrows in V, W) with severe disruption and underpopulation of cortical layers 2 and 5 of the cortex (arrow in X).

Figure 2

Left top panels: Single optical confocal sections of neural tubes from wild-type (A, C) and Lis1^hc/ko (B, D) embryos at E9.5 (A, B) and E14.5 (C, D) were stained with nestin (neural stem cell marker, green) and pericentrin (centrosomal marker, red) in the left panels, and pancadherin (apical marker, red) in the right panels to detect the cadherin hole (arrows). All sections were counterstained with DAPI. Dotted lines are examples of spindle cleavage plane measurements relative to the ventricular zone surface, and arrows the apical location of the cadherin hole. Right top panels: Single optical confocal sections of neural tubes from Lis1^hc/hc littermates either without (E, G) or with (F, H) Cre-ER^TM, a tamoxifen inducible Cre, after tamoxifen injection into pregnant females at E8.5, and embryo isolation 12 or 24 hr later. Neural tubes were stained with nestin (neural stem cell marker, green) and pericentrin (centrosomal marker, red) in the left panels, aPKCζ (apical marker, red) in the central panels and pan-cadherin (apical marker, red) in the right panels to detect the cadherin hole (arrows). All sections were counterstained with DAPI. (I) Mean (box) and standard deviation (lines) of angles of spindle cleavage of NE (E9.5) progenitors in WT (blue, 80.72° ± 9.56), Lis1^+/ko (red, 64.73° ± 24.5) and Lis1^hc/ko (green, 48.40° ± 22.0) embryos, as well as RG (E14.5) progenitors in WT (blue, 70.15° ± 18.3), Lis1^+/ko (red, 65.27° ± 24.5) and Lis1^hc/ko (green, 40.49° ± 25.9) embryos are shown. (J) Mean (box) and standard deviation (lines) of angles of spindle cleavage of littermate Lis1^hc/ko (−CRE, blue) or Lis1^hc/ko; Cre-ER^TM (+CRE, orange) embryos 12, 24 and 36 hr after pregnant females were injected intraperitoneally at E8.5 with tamoxifen. Average angles at 12, 24 and 36 hr of −CRE embryos were 80.32° ± 8.33, 78.52° ± 9.93 and 79.75° ± 8.19, respectively, while for +CRE embryos the angles were 49.58° ± 26.5, 44.36° ± 25.2 and 50.34° ± 16.4 at 12, 24 and 36 hrs, respectively. (K) Mean (bars) and standard deviation (lines) of percentage of apoptotic cells of littermate Lis1^hc/ko (−CRE, blue) or Lis1^hc/ko; Cre-ER^TM (+CRE, orange) embryos 12, 24 and 36 hr after pregnant females were injected intraperitoneally at E8.5 with tamoxifen. (L) Mean (bars) and standard deviation (lines) of angles of the number of mitotic cells measured by phopho-H3 histone of littermate Lis1^hc/ko (−CRE, blue) or Lis1^hc/ko; Cre-ER^TM (+CRE, orange) embryos 12, 24 and 36 hr after pregnant females were injected intraperitoneally at E8.5 with tamoxifen. Green areas at the top of each bar represents the portion of phospho-H3 postitive cells away (non-apical) from their normal position at the apical surface (brackets). Asterisks in I-L indicate * p <0.05, ** p <0.02, *** p <0.0001 by unpaired ANOVA with Bonferroni post-hoc test.

Figure 3

(A) Western Blot of LIS1 and actin levels in wild-type (wt) and Lis1^hc/ko MEFs, as well as in Lis1^hc/hc; Cre-ER^TM; Rosa26 MEFs after 0-72h tamoxifen treatment. (B) Growth curve of wt, Lis1^+/ko and Lis1^hc/ko MEFS. (C) Growth curve of Lis1^cko/cko; Cre-ER^TM without (control) and with 1 µm tamoxifen treatment (tamoxifen). (D) Quantitation of mitotic index using antiphospho H3 Histone. Doubling time and mitotic index were significantly different among groups (p < 0.05) by one-way ANOVA with a Bonferroni post-hoc test. Astral MTs concentration measured by EB1 (green) and β-tubulin (red) display in wt (E), Lis1^hc/ko (F) and Lis1^48h (G) MEFs. White lines indicate the edge of the cell, and blue is Hoechst. Prometaphase wt (H), Lis1^hc/ko (I) and Lis1^48h (J) MEFs after immunohistochemistry with pericentrin (red) and α-tubulin (green). (K–N) MG-132 metaphase-arrested wt (K, L) and Lis1^hc/ko (M, N) MEFs were untreated (K, M) or cold-treated on ice (L, N) for an hour to depolymerize unattached MTs within the spindle, and stained for α-tubulin (green) and Hoechst (blue). Spindles from wt (O), Lis1^hc/ko (P), Lis1^24h (Q) and Lis1^48h (R) MG132-arrested MEFs stained with α-tubulin (green), SH-CREST (red) and Hoechst (blue). Lagging chromosomes (arrows) were observed in Lis1^48h MEFs. Anaphase structures of wt (S, T) and Lis1^24h (U, V) MEFs stained with α-tubulin (green) and Hoechst (blue). Arrows point to lagging chromosomes in Lis1^48h MEFs. Scale bars: H–J 5 µm, K–N 4 µm and O–R 5 µm.

Figure 4

Dynein components were examined by immunocytochemistry in wild-type (wt, A) and Lis1^hc/ko MEFs (hc/ko, B) with DIC (blue), LIS1 (red), and α-tubulin (TUB, green). White arrows and lines indicate areas at the edge of the cell. Yellow arrows in hc/ko cells indicate an area inside of the cell with increased LIS1 and DIC localization and MT stability. (C) Acetylated tubulin (ac-TUB, green,compare arrows and arrowheads in similar cells) and (E) western blot analysis. (D) α-tubulin (TUB, green) was depolymerized by nocodazole treatment in wild-type and Lis1^hc/ko MEFs. MEFs were then washed and allowed to recover for 10s. White lines outline the area of MT recovery. (F) NDEL1 (aqua) was mislocalized to the perinuclear region in MEFs with complete loss of LIS1 (RED-CRE), compared with wild-type MEFs, which displayed diffuse NDEL1 localization out to the cortex. The edge of the cell with RED-CRE is outlined. DAPI (blue) and γ-tubulin (green) define the nucleus and centrosome. (G) Cells with abnormal NDEL1 (defined as away from the periphery and more perinuclear in localization) were quantified in CRE-negative (white bar) and CRE-positive (black bar). (H–I) Dynein and CLIP-170 were examined at the cell cortex with decreasing doses of LIS1. (H, J) wild-type; (I, K) Lis1^hc/ko. All panels: LIS1, red; α-tubulin, blue; green: (K, L) DIC; (M, N) CLIP-170.

Figure 5

Representative (of 15 cells of each genotype) still images from live movies of tubulin-YFP from wild-type (A–D) and Lis1^hc/ko (E–G) MEFs. (A, B) Low-magnification views of two wild-type images. (C, D) MTs in wild-type MEFs display constant curling behavior, with very dynamic activity. (E) Low-magnification view of a representative Lis1^hc/ko MEF cell. (F) Regions of poor MT stability and (G) increased MT stability were seen in the same Lis1^hc/ko cell. Arrowheads and chevrons track the movement of individual MTs over indicated periods of time in seconds. The arrowhead at 180s is not present because that MT was completely destabilized back to the centrosomes. (G) MTs in certain regions of Lis1^hc/ko MEFs extend and bend at the cell cortex, exhibiting some curving movements and stabilization.

Figure 6

NDEL1 rescue of MEF defects in β-tubulin (A) and DIC (B) localization. Lis1^hc/ko MEFs were transfected with no DNA (−), or with expression constructs for GFP, GFP-LIS1, GFP-NDEL1, GFP-NDEL1 (tripleS/T-A) with all three CDK1 sites mutated to alanine, and GFP-NDEL1 (S251A), with the Aurora-A kinase site mutated to alanine, in the presence or absence of dsRed-Cre recombinase (Toyo-oka et al. 2005). Three days after transfection, β-tubulin (A) and DIC (B) localization was determined. Upper panels, immunofluorescence staining with anti-β-tubulin (A) and anti-DIC. Bottom panels, statistical analysis of distribution of β-tubulin (A) and DIC (B) localization in Lis1^hc/ko MEFs without (white) or with (black) Cre. Staining patterns of MEFs were classified as having peripheral MTs (A) or for perinuclear DIC (B). More than 100 cells were phenotyped for each measurement in three separate experiments. Arrows indicate centrosomal distribution of transfected proteins. Asterisks denote *** p <0.0001 by unpaired ANOVA with Bonferroni post-hoc test.

Figure 6

NDEL1 rescue of MEF defects in β-tubulin (A) and DIC (B) localization. Lis1^hc/ko MEFs were transfected with no DNA (−), or with expression constructs for GFP, GFP-LIS1, GFP-NDEL1, GFP-NDEL1 (tripleS/T-A) with all three CDK1 sites mutated to alanine, and GFP-NDEL1 (S251A), with the Aurora-A kinase site mutated to alanine, in the presence or absence of dsRed-Cre recombinase (Toyo-oka et al. 2005). Three days after transfection, β-tubulin (A) and DIC (B) localization was determined. Upper panels, immunofluorescence staining with anti-β-tubulin (A) and anti-DIC. Bottom panels, statistical analysis of distribution of β-tubulin (A) and DIC (B) localization in Lis1^hc/ko MEFs without (white) or with (black) Cre. Staining patterns of MEFs were classified as having peripheral MTs (A) or for perinuclear DIC (B). More than 100 cells were phenotyped for each measurement in three separate experiments. Arrows indicate centrosomal distribution of transfected proteins. Asterisks denote *** p <0.0001 by unpaired ANOVA with Bonferroni post-hoc test.

Figure 7

(A) Model for the role of LIS1 in neuroepithelial stem cells (NESCs) and radial glial progenitor cells (RGPCs or RGs) in developing cortex from wild-type and Lis1 mutants. M, G1, S, G2 refer to mitotic stages, and arrows next to soma reflect the direction of interkinetic nuclear migration. Spindle orientation in M-phase cells is indicated. Wild-type NESCs divide symmetrically to generate NESCs. Lis1 mutant NESCs divide asymmetrically, resulting in apoptosis. In RGPCs, interkinetic nuclear migration occurs only in the ventricular zone (VZ), but not in the intermediate zone (IZ) or cortical plate (CP), and cells divide symmetrically (Sym M) or asymmetrically (Asy M). In Lis1 mutants, the predominant asymmetric mode of division leads to progenitor depletion. (B) Model for the role of cortical dynein (o^o) in cortical MT capture. In wild-type cells, cellular contents can be moved by pulling MTs via cortical dynein, while in metaphase, the spindle can be rotated (arrows) by cortical dynein. In Lis1 mutant cells, cortical dynein is reduced, resulting in reduction in cortical MTs in interphase, as well as weakened spindle and astral MTs and an inability to rotate the mitotic spindle. See text for further details.