Asymmetric Notch activity by differential inheritance of lysosomes in human neural stem cells

Abstract

Symmetric and asymmetric cell divisions are conserved strategies for stem cell expansion and the generation of more committed progeny, respectively. Here, we demonstrate that in human neural stem cells (NSCs), lysosomes are asymmetrically inherited during mitosis. We show that lysosomes contain Notch receptors and that Notch activation occurs the acidic lysosome environment. The lysosome asymmetry correlates with the expression of the Notch target gene HES1 and the activity of Notch signaling in the daughter cells. Furthermore, an asymmetry of lysosomes and Notch receptors was also observed in a human organoid model of brain development with mitotic figures showing preferential inheritance of lysosomes and Notch receptor in that daughter cell remaining attached to the apical membrane. Thus, this study suggests a previously unknown function of lysosomes as a signaling hub to establish a bias in Notch signaling activity between daughter cells after an asymmetric cell division of human NSCs.

Fig. 1.. Human NSCs tend to differentiation upon FGF2 withdrawal accompanied by an increase in asymmetric LAMP1 segregation during mitosis.

(A) Representative brightfield images and immunostaining of rosette-type NSCs from two healthy control cell lines (Ctrl#1 and Ctrl#2) for NSC markers (Sox2 and Nestin) and neuron markers (Tubb3 and HuC/D). (B and C) Immunostaining for Sox2 and HuC/D and respective quantification of the amount of NSCs (Sox2⁺) and neurons (HuC/D⁺) after FGF2 withdrawal from Ctrl#1-NSCs over a time course of 7 days (n = 3, two-way ANOVA with Bonferroni’s multiple comparison test, means + SEM). (D and E) Representative images of symmetric and asymmetric segregation of LAMP1⁺ and CD63⁺ vesicles during mitosis of Ctrl#1-NSCs. Yellow lines indicate ROIs surrounding daughter cells, defined by DAPI and phalloidin staining (see also fig. S1E). (F and G) Quantification of LAMP1 and CD63 asymmetry based on sum intensity ratios between paired daughter cells (see fig. S1D) in conditions with and without FGF2 (N = 6 coverslips from three independent experiments with n = 42 to 58 cells per coverslip, two-way ANOVA with Bonferroni’s multiple comparison test, means + SEM). (H) Representative images of LAMP1 costained with Notch1 receptors in different mitotic phases of Ctrl#1-NSCs. (I) Correlation of asymmetry indices A of LAMP1 and Notch1 in Ctrl#1 and Ctrl#2-NSCs. Dots represent individual mitotic event, and linear regression is shown. Scale bars, 100 μm (A and B), 10 μm (D, E, and H), 5 μm [zoomed in (H)]. DNA was counterstained with DAPI. *P < 0.05 and ***P < 0.001. n.s., not significant. See also fig. S1.

Fig. 2.. Notch1 receptors are accumulated in lysosomes after clathrin-mediated endocytosis.

(A) Representative Western blot of sucrose gradient fractions collected from Ctrl#1-NSCs. (B and C) Protein quantification of vesicle markers (B), Notch1 receptors and γ-secretase subunit Presenilin1 (C) along the sucrose gradient [n = 3, means + SEM, mean of LAMP1 again in (C) as reference]. (D and E) Immunostaining for LAMP1 and Notch1 in Ctrl#2-NSCs treated for 1 hour with 0.1% (v/v) DMSO (D) or 50 μM dynasore (E). (F) Manders’ co-occurrence analysis of LAMP1 and Notch1 receptors with and without dynasore treatment (N = 179 cells from three independent experiments, Kruskal-Wallis test with Dunn’s multiple comparison test, box plot with dots for individual cells). (G) Representative images of Ctrl#2-NSCs stained with CTB-AF555 and immunostained for the N1ECD and clathrin. Scale bars, 20 μm (D, E, and G) and 5 μm [zoomed in (D), (E), and (G)]. Nuclei were counterstained with DAPI in (D and E). ***P < 0.001.

Fig. 3.. Notch internalization is triggered by ligand binding and vesicle acidification is necessary for receptor activation.

(A) Representative frames from live-cell imaging of LysoTracker-stained Ctrl#2-NSCs starting after addition of DLL1-pHrodo at time point 0. LysoTracker- and pHrodo single-positive vesicular structures are indicated by red and green arrows, respectively, double-positive structures with yellow arrows. (B) Representative kymographs from confocal live-cell imaging of Ctrl#2-NSCs treated with LysoTracker and DLL1-pHrodo. LysoTracker- and pHrodo single-positive vesicular structures are indicated by red and green arrows, respectively, and double-positive structures are indicated by yellow arrows. (C) Immunostaining for Notch1, LAMP1, and 6xHis after incubation of Ctrl#2-NSCs with DLL1-6xHis for 30 min. Arrows indicate puncta colocalizing for all three proteins. (D to F) Western blot against cleaved N1ICD, total Notch1, and actin and respective quantifications of Ctrl#1-NSCs treated with 0.1% (v/v) DMSO, 20 μM DAPT, 100 nM BafA, or 200 μM Leu for 2 hours (n = 4, one-way ANOVA with Bonferroni’s multiple comparison test, means ± SEM). (G) Fold changes of gene expression of Notch downstream targets HES1, HES5, and HEY1 after treatment of Ctrl#2-NSC with 0.1% (v/v) DMSO, 20 μM DAPT, 100 nM BafA, or 200 μM Leu for 2 hours (n = 3, one-way ANOVA with Bonferroni’s multiple comparison test, means ± SEM). Scale bars, 20 μm (A and C), 5 μm [zoomed in (C)], and 10 μm (B). Time scale, hh:minmin:ss (A). *P < 0.05. See also fig. S2.

Fig. 4.. HES1 reporter tracks Notch signaling activity in NSCs.

(A) Targeting strategy for CRISPR-Cas9–mediated generation of HES1 reporter cell line using Ctrl#2-iPSCs (for characterization see fig. S3). (B) Immunostaining for NSC (Nestin and Sox2) and neuronal markers (Tubb3 and HuC/HuD) in HES1 reporter NSCs. (C) Representative bright-field images of CRISPR-Cas9–edited rosette-type NSCs with tdTomato expression indicating active Notch signaling. (D and E) Representative frames from live-cell imaging and respective intensity sum quantification using HES1 reporter NSCs treated with 0.1% (v/v) DMSO, 20 μM DAPT, 100 nM BafA, or 200 μM Leu for up to 12 hours (N = 27 cells from three independent experiments, two-way ANOVA with Bonferroni’s multiple comparison test, means ± SEM). Also shown are the exponential regression curves for DAPT and BafA treatment and the linear regression for DMSO and Leu. Scale bars, 100 μm (B and C) and 50 μm (D). Nuclei were counterstained with DAPI in (B). Time scale hh:minmin:ss. *P < 0.05. See also fig. S3.

Fig. 5.. Lysosome asymmetry is correlated with differential HES1 expression in NSC-derived daughter cells.

(A) Representative time frames from live-cell imaging of a mitotic HES1 reporter NSC stained with LysoTracker. On the basis of quantification of the LysoTracker intensity sum, daughter cell1 was defined as the cell receiving more LysoTracker signal at the end of mitosis (time point 0). (B) Distribution of LysoTracker symmetric and asymmetric cell division (n = 10 independent experiments). (C) Representative time frames of the tdTomato signal in daughter cells 1 and 2 starting 1.5 to 6 hours after mitosis. (D) Differences in z-normalized intensity sums between daughter cell 1 and cell 2 in LysoTracker symmetric and asymmetric divisions (N = 124/31 for symmetric/asymmetric cell divisions from 10 independent experiments, means + SEM). (E) Heatmap and clustering of HES1 expression dynamics in daughter cell pairs into four groups: (i) symmetric, (ii) opposed asymmetric, (iii) slightly, and (iv) highly asymmetric. (F) Dynamics of tdTomato fluorescence intensities in the different groups (N = 70, 27, 35, and 23 for groups 1 to 4, respectively, from 10 independent experiments, means ± SEM). (G) HES1 expression dynamics and cluster frequency after symmetric and asymmetric inheritance of LysoTracker during cell division of HES1 reporter NSCs (N = 124/31 of symmetric/asymmetric cell divisions from 10 independent experiments). 20 μm (A and C). Time scale hh:minmin:ss (A and C). See also fig. S3.

Fig. 6.. Neurogenic cell divisions in forebrain organoids show biased localization of LAMP1 and Notch1 in the apical daughter cell.

(A) Representative image of 20-day-old forebrain organoid immunostained for NSC marker Sox2 and neuronal marker Tubb3 with cortical loops indicated by white dashed line. (B) Illustration of forebrain organoids with cortical loops and NSCs dividing in a planar, intermediate, or delaminating manner at the apical membrane of the loops. (C) Immunostaining of forebrain organoids for phalloidin, LAMP1, and Notch1. Dividing cells at the apical membrane (yellow dashed line) were identified by phalloidin staining and grouped into planar, intermediate, or delaminating cell division. LAMP1 and Notch1 signal within the area of the ROIs (yellow) is shown. (D) Distribution of cell division modes as defined in (B) (n = 3 independent experiments, means + SEM). (E) Quantification of LAMP1 and Notch1 asymmetry based on sum intensity ratios (see also fig. S1) between paired daughter cells (N = 60/28/19 of planar/intermediate/delaminating cell divisions from three independent experiments, Kruskal-Wallis test with Dunn’s multiple comparison test, box plot with dots for individual mitotic events). Scale bars, 300 μm (A) and 5 μm (C). Nuclei were counterstained with DAPI in (A). **P < 0.01.