Lunatic fringe-mediated Notch signaling regulates adult hippocampal neural stem cell maintenance
- Baylor College of Medicine, United States
- Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, United States
- The Basque Science Foundation, Spain
- Rosalind Franklin University of Medicine and Science, United States
- Keck School of Medicine, University of Southern California, United States
Figures
Figure 1 with 1 supplement
Lfng-eGFP is expressed in the NSCs of dentate gyrus.
(A) Left panel: Confocal photomicrograph of the dentate gyrus in 2 month-old Lfng-eGFP mouse shows Lfng-eGFP expressing cells in the subgranular zone only. Right panel: Lfng-eGFP expressing cells have typical NSC morphology: triangular cell body in the subgranular zone, a single radial process spanning the granular zone, and fine terminal arborizations in the molecular layer. ML=molecular layer, GZ=granule cell zone, SGZ=subgranular zone. (B) In situ hybridization against the Lfng mRNA shows probe expression in the SGZ. Dotted lines indicate the borders of SGZ. (C) In a double transgenic LfngTm1Grid;Lfng-eGFP mouse, β-galactosidase (β–Gal) and eGFP are co-expressed, confirming that Lfng promoter guiding the eGFP expression is active in the same cells that express β–Gal. LfngTm1Grid=Lfngβ-Gal mouse that carries β-Gal insertion in the Lfng locus. (D) Lfng-eGFP colocalizes with other NSC markers, such as Nestin, Sox2, Vimentin, and GFAP. Confocal photomicrographs of the representative examples and relative quantitation is shown. (E) eGFP is briefly retained in the first progeny of NSC. Left panel: Lfng-eGFP expressing cell (thick arrow) and its immediate ANP progeny (arrowhead) are both BrdU+ and in cytoplasmic contact. BrdU+ ANP not in the cytoplasmic contact with the BrdU+ NSC has very low eGFP expression (thin arrow). Left graph: eGFP is cleared from BrdU+ ANPs between 1 to 3 days following a single dose of BrdU, corresponding to the half-life of GFP protein (N = 3–5 per group, p<0.001 or p<0.0001 for pairwise comparisons of all timepoints, ANOVA with Tukey post-hoc test). Right graph: Lfng-eGFP is expressed only in minority of Type2a ANPs (Sox2+ GFAP- cells in the SGZ; 10.58 ± 2.38, N = 3), unlike Nestin-GFP (94.7 ± 0.34, N = 4). Right panel: Late, Tbr2+ ANPs (arrowheads) do not express Lfng-eGFP. Arrow points to Lfng-eGFP NSC. (F) Lfng-eGFP does not co-localize with the markers of neuronal lineage (Dcx+ neuroblasts and immature neurons, and NeuN+ granule cells) nor S100β+ astrocytes. (G) While in the Nestin-GFP (N) mice GFP is expressed in NSCs, ANPs, and other cell types throughout the dentate gyrus in approximately equal proportions (30.16 ± 2.43 NSCs, 36.12 ± 3.73 ANPs, 33.71 ± 2.81 other cell types), in the Lfng-eGFP (L) mice it is expressed predominantly in NSCs (81.68 ± 2.62% NSCs, 16.37 ± 1.54% ANPs, 1.95 ± 1.14% other cell types; N = 4 per genotype). Bars represent mean±SEM. **p<0.001, ***p<0.0001. Scale bars = 100 µm (A left, (B), 20 µm (A right, (C–F). See Figure 1—figure supplement 1 for further details.
Figure 1—figure supplement 1
Lfng-eGFP labels a small number of non-NSC cell types in the dentate gyrus.
(A) In a double transgenic Lfng-eGFP; Nestin-CFPnuc mouse, NSCs express both CFP and eGFP, and have a radial GFAP+ process (arrow). Those that express only CFP have round morphology with no processes and are presumed ANPs (arrowheads). Orthogonal projection shows a NSC expressing both Lfng-eGFP and Nestin-CFPnuc (black arrowhead) and a neighboring ANP expressing only Nestin-CFPnuc (white arrowhead). Scale bar = 20 µm. (B) Lfng-eGFP is expressed in tip-like cells in close proximity to the endothelial cells immunolabeled with CD31 (arrowhead). Scale bar = 20 µm. (C) Random unidentified Lfng-eGFP+ cells are detected in the hilus (left panel; arrowhead) and molecular layer (right panel; arrowhead). Scale bar = 20 µm.
Figure 2 with 1 supplement
Lfng-eGFP-expressing cells are functional NSCs.
(A) Most Lfng-eGFP+ NSCs are quiescent. Bar graphs represent the total number of GFP+ cells (left panel), GFP+ NSCs (middle panel) and GFP+ ANPs (right panel) in 3 month-old Nestin-GFP and Lfng-eGFP mice (N = 4 per genotype) treated with temozolomide (TMZ). The difference between the two mouse models is most notable with respect to ANPs: while Nestin-GFP labels a large number of ANPs, Lfng-eGFP does not - it labels primarily quiescent cells. (B) Electroconvulsive shock (ECS) and physical exercise (PE) both activate Lfng-eGFP+ NSCs (N = 4 per group in ECS and N = 6 per group in PE). (C) Lfng-eGFP NSCs produce neuronal progeny. The relative number of newborn, BrdU+ progeny was quantified over a 30 day period (NSCs: Nestin-GFP+ or Lfng-eGFP+ cells with GFAP+ radial processes; ANPs: GFAP- Dcx- NeuN-; NBs: Dcx+ neuroblasts and immature neurons; GCs: NeuN+; Other: BrdU+ Dcx- NeuN- cells outside the SGZ; N = 4 per genotype per timepoint). Cumulative BrdU paradigm (four 150 mg/kg injections given 2 hr apart) was used to increase the yield of labeled newborn cells. N=Nestin-GFP mice, L=Lfng-eGFP mice. (D) The number of Lfng-eGFP+ NSCs declines over an 18 month period comparably to the number of Nestin-GFP+ NSCs (left panel; N = 4 per timepoint per genotype). However, the contribution of GFP+ cell types in the Nestin-GFP and Lfng-eGFP mice differs at different age (right panel). While Lfng-eGFP remains selective for NSCs during aging (p>0.15 for all timepoints, Tukey post-hoc test), Nestin-GFP labels significantly more non-neuroprogenitors in older mice (p<0.002; Tukey post-hoc test). Bars represent mean±SEM. NS=non-significant, *p<0.05, **p<0.001. See Figure 2—figure supplement 1 for further details.
Figure 2—figure supplement 1
The specificity of Lfng-eGFP expression for NSCs does not change over time.
(A) The total number of BrdU+ cells (NSCs and their progeny) in Nestin-GFP and Lfng- eGFP mice over a 30 day period indicates that cell survival is similar in both mouse models. (B) Neither the total number BrdU+ cells (left panel) nor the relative number of BrdU+ NSCs (right panel) in Nestin-GFP and Lfng-eGFP mice significantly differs over the 18 month period investigated (N = 4 per genotype per timepoint), indicating that the proliferative capacity of NSCs in the two mouse models is similar. (C) Confocal photomicrographs are representative examples of the dentate gyri of Nestin-GFP and Lfng-eGFP mice at a given age. Note a decline in the number of GFP+ NSCs in both models. However, while the GFP+ non-progenitors (‘other cells’) prevail as the Nestin-GFP mice age, in the Lfng-eGFP mice eGFP remains expressed in NSCs and the contribution of the GFP+‘other cells’ is minimal. See Figure 2D for quantification. Bars represent mean ± SEM. Scale bar = 20 µm.
Figure 3 with 1 supplement
Lfng-expressing NSCs generate diverse progeny.
(A) The map of bacterial artificial chromosome (BAC) construct used to generate the Lfng-CreERT2 mouse. (B) In Lfng-CreERT2; RCL-tdT; Lfng-eGFP triple transgenic mouse, tdTomato+ and eGFP+ co-expressing cells demonstrate the specificity of the Lfng-CreERT2 line to NSCs. Left panel: Confocal photomicrograph of the dentate gyrus of a 6 month-old mouse shows the overlapping expression of eGFP and CreERT2-controlled tdTomato one day following tamoxifen injection (TMX; 120 mg/kg). Right panel: Quantification of the co-expression of tdTomato+ and eGFP+ in induced Lfng-CreERT2; RCL-tdT mice (N = 3). Bars represent mean±SEM. (C) TMX-induced cells have NSC morphology and divide both asymmetrically and symmetrically. Left panel: tdTomato+ NSC co-expresses GFAP and Sox2. Middle panel: tdTomato+ NSC (arrow) divides asymmetrically to produce ANP (arrowhead). Right panel: NSC divides symmetrically to produce two cells (arrows) with prominent GFAP+ radial processes. Scale bars = 20 µm. (D) tdTomato+ NSCs produce new neurons through established cascade of cell types, from Tbr2+ late ANPs (Type 2b cells), through Dcx+ immature neurons, to NeuN+ granule cells. Scale bars = 20 µm. (E) tdTomato+ NSCs also produce S100β astrocytes (or astrocyte-like cells) within the granule cell layer (left panel), as well as stellar Sox2+ GFAP+ cells in the hilus (right panel) Scale bars = 20 µm. (F) Fate mapping of tdTomato+ cells following TMX induction in Lfng-CreERT2; RCL-tdT mice reveals that NSCs form the majority (85.7% ± 0.9) of the induced cells at 1dpi, but progressively decline in ratio as they give rise to different progeny over the course of 2 months (N = 3–5 per timepoint). See Figure 3—figure supplement 1 for further details.
Figure 3—figure supplement 1
Lfng-CreERT2 expressing cells are NSCs giving rise to newborn neurons.
(A) tdTomato expression in Lfng-CreERT2; RCL-tdT animals 3 months following vehicle treatment. No tdTomato is observed, indicating no leakage. Scale bar = 50 µm. (B) Both the number of NSCs (N = 4, p=0.039) and the number of cycling Ki67+ cells (N = 4, p<0.0001) significantly decline following diphtheria toxin (DTX) injection in Lfng-CreERT2; iDTR mice, further confirming the stemness of Lfng-expressing NSCs. DTR expression in Lfng-CreERT2; iDTR mice was induced by tamoxifen (TMX (day 0), followed by four injections of DTX (16 µg/kg) two days apart to kill Lfng-CreERT2 derivatives. Mice were sacrificed at 15 days post-induction. (C) A confocal photomicrograph of a newborn neuron with prominent dendritic spines and axonal projections to CA3 region, observed at 30 days post-induction in Lfng-CreERT2; RCL-tdT. Scale bar = 100 μm. (D) Occasionally, tdTomato+ tip cells were observed following induction in Lfng-CreERT2; RCL-tdT mice. Scale bar = 20 μm.
Figure 4
Notch pathway elements are present in the SGZ NSC niche.
(A) Notch1 is expressed in Lfng-eGFP NSCs, facing granule cell layer, where late ANPs (type 2b cells) and granule cells are located. Jag1 is expressed in adjacent ANP, and Dll1 on adjacent granule cell-NSC boundary. Hes5, downstream target of canonical Notch signaling pathway, is present in some (arrowhead) but not all (empty arrowhead) NSCs. Scale bar = 10 µm (Notch1, Jag1, Dll1); 20 µm (Hes5). (B) Venus is expressed in some (arrowhead) but not all (empty arrowhead) NSCs (Sox2+ cell body and GFAP+ process) in CBF:H2B-Venus mice. Type 2b cells (Tbr2) and most of the neuroblasts and immature neurons are Venus -, whereas almost all granule cells (NeuN+) are Venus+ with various degrees of intensity. Some of the mature astrocytes (S100β+) have active Notch signaling (arrowheads). Quantification of Venus signal among various cell types reveals that most of the differentiating cells are devoid of active Notch signaling. As soon as the neurogenic differentiation finishes, Notch signaling is turned on again in the granule cells. Scale bar = 10 µm upper panels and 20 µm middle and lower panels. (C) NICD1 and Venus colocalize in CBF:H2B-Venus mice. NICD1 almost completely overlaps with the Venus signal in the SGZ (left inset), while it is lacking in the molecular layer (right inset). Scale bar = 20 µm. Quantification of Venus+ cells among NICD1+ cells in GZ and SGZ verifies high degree of colocalization (N = 3; 90.79 ± 0.77% and 96.95 ± 0.57%, respectively).
Figure 5 with 1 supplement
Lfng preserves NSCs by controlling their cell cycle.
(A) Left panel: Lfng-eGFP NSCs express GFAP+ radial processes originating from the Sox2+ cell nuclei located in the SGZ. Right panel: The total number of GFAP+ Sox2+ NSCs in 4-month-old wild-type mice does not differ from the total number of eGFP+ NSCs in 4 month-old Lfng-eGFP (N = 3 per group; Student’s t-test, p=0.99). Scale bar = 10 µm. (B) The total number of NSCs in 2-month-old wild-type, Lfng heterozygote (Lfng−/+), and homozygote (Lfng−/−) knockout mice shows Lfng dose-dependent decrease in the NSC population (N = 3–4 per group; One-way ANOVA p<0.00001, Tukey HSD post-hoc test: p<0.0001 for wild-type vs Lfng−/+ or Lfng−/−, p=0.0484 for Lfng−/+ vs Lfng−/−). (C) Lack of Lfng promotes increased division of NSCs. Left panel: NSCs lacking Lfng have a higher ratio of cycling Ki67+ NSCs compared to wild-type (N = 4 per group; Student’s t-test p=0.0221). Middle panel: NSCs lacking Lfng have a higher ratio of actively dividing, BrdU+ NSCs compared to wild-type (N = 4 per group; Student’s t-test p=0.0017). Right panel: Lfng acts cell-autonomously and in a dose-dependent manner to control the NSC division (N = 4 per group). The ratio of BrdU+ NSCs was compared between iLfngfl/fl mutant clones (tdTomato+), iLfngfl/fl wild type clones (tdTomato- GFAP+ Sox2+), and Lfng-CreERT2; RCL-tdT control mice (tdTomato+ GFAP+ Sox2+). N = 4; p=0.9978 for control vs iLfngfl/fl tdTom- clones; p=0.0049 for control vs iLfngfl/fl tdTom+ clones; p=0.0045 for iLfngfl/fl tdTom- clones vs iLfngfl/fl tdTom+ clones. (D) NSCs lacking Lfng are hyper-activated in response to ECS treatment (N = 3–4 per group; Student’s t-test, p=0.0178). (E) Lack of Lfng decreases Notch signal intensity in mutant NSCs. Relative intensities of NICD1 staining are significantly lower in iLfngfl/fl mutant NSCs compared to control (N = 4 for control; N = 3 for iLfngfl/fl; Student’s t-test, p=0.0004). (F) NSCs lacking Lfng spend less time in the active state than wild-type NSCs. Lfng absence is associated with decreased S-phase re-entry 3 (left panel) and 7 (middle panel) days following the initial division compared to the wild-type NSCs (N = 4 per group; Student’s t-test, p=0.0024 for 3d, p=0.0003 for 7d). CldU+ IdU+ cells represent NSCs that underwent first division at the time of CldU injection (day 0) and were in S-phase at the time of IdU injection (day 3 or day 7). Right panel: In iLfngfl/fl mice, no NSCs were found that re-entered S-phase 7 days after the initial division. CldU+ IdU+ cells represent NSCs that were induced at day 0, underwent first division 1 day post-induction (CldU+) and were in S-phase 7 days post-induction (IdU+; N = 3–4 per group; Student’s t-test, p=0.0014). (G) NSCs lacking Lfng mostly exit cell cycle within a week of first division. CldU+ Ki67+ NSCs represent NSCs that are actively cycling 7 days following the CldU injection (N = 4; Student’s t-test, p=0.0105). (H) NSCs lacking Lfng give rise to astrocyte-like cells. Left panel: The number of BrdU-retaining NSCs 7 days after the BrdU injection is significantly reduced in Lfng−/+ mice compared to wild-type (N = 4 per timepoint; Student’s t-test, p=0.0003). Middle panel: In iLfngfl/fl mice, significantly more tdTomato+ astrocyte-like cells accumulate 7 and 30 days following induction compared to controls, while the number of tdTomato+ ANPs significantly decreases (N = 4 per group, p=0.0003 and p<0.0001 for astrocyte-like cells, p=0.0024 and p=0.0049 for ANPs). Right panel: The number of BrdU-retaining astrocyte-like cells 7 days after the BrdU injection is significantly higher in iLfngfl/fl mice compared to controls, but the difference is lost at 30 days (N = 4 per group; Student’s t-test, p=0.0302 for 7 days, p=0.4412 for 30 days). Bars represent mean ± SEM. *p<0.05, **p<0.001, ***p<0.0001. See Figure 5—figure supplement 1 for further details.
Figure 5—figure supplement 1
Jag1 and Lfng affect cell survival in the SGZ.
(A) ApopTag immunostaining in control, iLfngfl/fl, and iJag1fl/fl mice. Regions outlined in left images are presented in high magnification on the right. Scale bars = 100 µm (low magnification); 20 µm (high magnification). (B) Cell death is significantly decreased in iLfngfl/fl and iJag1fl/fl SGZ in comparison to control both by ApopTag-based (left) and activated caspase 3 (act-casp3)-based (right) detection methods. The average number of ApopTag or act-casp3+ cells in control mice was used for normalization (N = 3 per group; One-way ANOVA p<0.00001, Tukey HSD post-hoc test: i) ApopTag: p=0.006 for control vs iLfngfl/fl, p=0.0375 for iLfngfl/fl vs iJag1fl/fl, p<0.00001 for control vs iJag1fl/fl; ii) act-casp3: p=0.0089 for control vs iLfngfl/fl, p=0.9261 for iLfngfl/fl vs iJag1fl/fl, p=0.0224 for control vs iJag1fl/fl). Bars represent mean ± SEM. *p<0.05, **p<0.001.
Figure 6 with 1 supplement
Notch ligands Jag1 and Dll1 regulate NSC cell cycle.
(A) The number of NSCs is diminished in mice lacking Dll1 compared to both wild-type and mice lacking Jag1 (N = 4 per group; One-way ANOVA p<0.00001, Tukey HSD post-hoc test p<0.0001 for Dll1 vs wild-type or Jag1−/+, p=0.1663 for wild-type vs Jag1−/+). (B) Lack of either Dll1 or Jag1 promotes increased division of NSCs. Left panel: Absence of Dll1 or Jag1 is associated with significantly higher proportion of cycling, Ki67+ NSCs (N = 4 per group, One-way ANOVA p<0.00001, Tukey HSD post-hoc test p<0.0001 for wild-type vs Dll1−/+ or Jag1−/+, p=0.046 for Dll1−/+ vs Jag1−/+). Middle panel: Mice lacking Dll1 have significantly higher proportion of dividing, BrdU+ NSCs compared to both wild-type and Jag1−/+ mice (N = 4 per group; One-way ANOVA p<0.0005, Tukey HSD post-hoc test p=0.0007 for wild-type vs Dll1−/+, p=0.0015 for Jag1−/+ vs Dll1−/+, p=0.8393 for wild-type vs Jag1−/+). Right panel: Absence of both copies of Jag1 in iJag1fl/fl mice is associated with significantly higher proportion of actively dividing, BrdU+ NSCs compared to controls (N = 4 per group; Student’s t-test, p=0.0028). (C) Lack of Dll1 has an opposite effect on NSC S-phase re-entry compared to lack of Jag1. Relative number of NSCs that re-enter S-phase 3 (left panel) and 7 (middle panel) days following the initial division is significantly lower in mice lacking Dll1 at 7 days, while it is significantly higher at both timepoints in mice lacking Jag1 compared to wild-type mice (for 3d: N = 4 per group; One-way ANOVA p=0.0074, Tukey HSD post-hoc test p=0.9624 for wild-type vs Dll1−/+, p=0.011 for Jag1−/+ vs Dll1−/+, p=0.0164 for wild-type vs Jag1−/+; for 7d: N = 4 per group; One-way ANOVA p<0.0001, Tukey HSD post-hoc test p=0.0016 for wild-type vs Dll1−/+, p<0.0001 for Jag1−/+ vs Dll1−/+, p=0.0002 for wild-type vs Jag1−/+). CldU+ IdU+ cells represent NSCs that underwent first division at the time of CldU injection (day 0) and were in S-phase at the time of IdU injection (day 3 or day 7). Right panel: In iJag1fl/fl mice, significantly higher proportion of NSCs re-enter S-phase 6 days following the initial division compared to controls (N = 4 per group; Student’s t-test, p=0.0003). CldU+ IdU+ cells represent NSCs that were induced at day 0, underwent first division at 1 day post-induction (CldU+) and were in S-phase (IdU+) at 7 days post-induction. (D) Lack of Dll1 has an opposite effect on NSC cycling time compared to lack of Jag1 (N = 4 per group; One-way ANOVA p<0.00001, Tukey HSD post-hoc test p=0.0049 for wild-type vs Dll1−/+, p=0.005 for wild-type vs Jag1−/+ and p<0.0001 for Dll1−/+ vs Jag1−/+). CldU+ Ki67+ NSCs represent NSCs that are cycling 7 days following the CldU injection. (E) Lack of Jag1 increases Notch signal intensity in NSCs. Relative intensities of NICD1 staining are significantly higher in Jag1 mutant NSC clones compared to control NSCs (N = 3 for iJag1fl/fl, N = 4 for control; Student’s t-test p<0.0255). (F) In Dll1−/+ mice, no NSCs that retained BrdU 7 days after the BrdU injection were detected. In Jag1−/+ mice, the absolute number of BrdU-retaining NSCs is significantly higher at both timepoints compared to wild-type mice, suggesting increased self-renewal of NSCs lacking Jag1 (For 7d: N = 4 per group; One-way ANOVA p<0.00001, Tukey HSD post-hoc test p<0.0001 for wild-type vs Jag1−/+ and Jag1−/+ vs Dll1−/+, and p=0.0035 for wild-type vs Dll1−/+. For 2hr-7d comparisons: p=0.0028 for Jag1−/+ and p=0.001 for Dll1−/+ mice). (G) Left graph: Mice lacking Jag1 have significantly more Sox2+ ANPs compared to both wild-type and mice lacking Dll1 (N = 4; One-way ANOVA p<0.00001; p<0.00001 for wild-type vs Jag1−/+ and Dll1−/+ vs Jag1−/+; p=0.2828 for wild-type vs Dll1−/+). Right graph: Average size of Ki67+ clusters around Ki67+ NSCs is larger in Jag1−/+ mice compared to wild-type and Dll1−/+ mice (N = 4; One-way ANOVA p<0.00001; p=0.0001 for wild-type vs Jag1−/+; p=0.0015 for wild-type vs Dll1−/+, p<0.0001 for Dll1−/+ vs Jag1−/+). (H) One month following BrdU injection, Dll1−/+ mice have significantly higher ratio of S100β+ progeny among newly generated cells compared to wild-type (N = 4; Student’s t-test, p=0.016). (I) In iJag1fl/fl mice, significantly more tdTomato+ NSCs accumulate at 7 and 30 days post-induction compared to controls, suggesting increased self-renewal in Jag1 mutant NSCs. This is accompanied by increased number of tdTomato+ ANPs (N = 4 per group; p=0.0263 and p<0.0001 for NSCs at 7d and 30d, respectively; p=0.0033 and p=0.069 for ANPs at 7d and 30d, respectively). Please note that all wild-type and control mice presented here are same as in Figure 5, as the experiments using the knockout lines (Lfng−/+, Dll1−/+, Jag1−/+, iLfngfl/fl, iJag1fl/fl) were done at the same time. The results are presented in two figures for clarity. Bars represent mean ± SEM* p<0.05, **p<0.001, ***p<0.0001. See Figure 6—figure supplement 1 for further details.
Figure 6—figure supplement 1
Lack of both Lfng−/+ and Jag1−/+ has additive effect on NSCs.
(A) The number of NSCs in 6 month old Lfng−/+ (L-/+) and Lfng−/+; Jag1−/+ (L-/+;J−/+) mice is significantly lower compared to wild-type mice (N = 3 for Lfng−/+ and Lfng−/+; Jag1−/+ and N = 4 wild-type; One-way ANOVA p<0.00001, Tukey HSD post-hoc test: p<0.0001 for wild-type vs Lfng−/+ and Lfng−/+; Jag1−/+; and p=0.6869 for Lfng−/+ vs Lfng−/+; Jag1−/+). (B) As in Lfng−/+ mice, NSCs in Lfng−/+; Jag1−/+ mice have higher ratio of actively dividing BrdU+ NSCs compared to wild-type mice (N = 3; One-way ANOVA p<0.0002, Tukey HSD post-hoc test: p=0.0001 for wild-type vs Lfng−/+, p=0.0003 for wild-type vs Lfng−/+; Jag1−/+, p=0.6724 for Lfng−/+ vs Lfng−/+; Jag1−/+). (C) Lfng−/+; Jag1−/+ double knockout mice have significantly higher proportion of NSCs re-entering S-phase 1 week following the initial division compared to wild-type (N = 3; One-way ANOVA p<0.0001, Tukey HSD post-hoc test: p<0.00001 for wild-type vs Lfng−/+; Jag1−/+ and Lfng−/+ vs Lfng−/+; Jag1−/+). Bars represent mean ± SEM. **p<0.001, ***p<0.0001.
Figure 7
Proposed model for Lfng-mediated regulation of NSC maintenance by progeny.
(A) In the resting state, Lfng-expressing NSC (green) is surrounded mostly by Delta1 (Dll)-expressing granule cells (GC, grey). Dll1 binds to the Lfng-modified Notch receptor on the NSC, which boosts Dll1-mediated Notch signaling by producing more NICDs and the NSC is kept quiescent but ready to undergo cell cycle if stimulated. Once activated, NSC (red) starts to produce ANPs. The first progeny (Type 2a, small dark yellow cells) do not express Notch ligands and the NSC continues to divide. As ANPs mature into late ANPs (Type 2b, yellow cells), they start to express Jag1. Jag1 binding to the Lfng-modified Notch receptor on the NSC does not generate NICD and thus the Notch signaling strength in the ‘mother’ NSC decreases. Eventually, the NSC is surrounded by mostly Jag1-expressing ANPs, and it exits active state (green NSC). Thus, in the resting state, the granule cell progeny prevents overt activation of NSCs, while in the active state, the ANP progeny prevents overt division of the NSC. These feedback signaling from the progeny both act to preserve the NSC population and the integrity of the niche. (B) The summary of the loss-of-function data, focusing on the number of activated NSCs (red) and the final outcomes of their division. Mice lacking Lfng and Dll1 have similar phenotypes: NSCs are recruited in bulk, they divide less and faster, and eventually lead to depletion of NSC population. In Lfng mutant mice, there is an increased transformation into astrocytes (blue). Mice lacking Jag1 have the opposite phenotype: while NSCs are activated as in the wild-type, they divide more and longer, and produce large clusters of ANPs as well as self-renew.
Author response image 1
Jag1 signal is mostly eliminated from SGZ in iJag1fl/fl mice.
Immunostaining for Jag1 expression using anti-Jag1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) shows that the signal is localized to both SGZ and hilus in control (Ctrl) samples, whereas it is present mostly in hilus in the conditional knockout mice and not in the SGZ. Method is described in Materials and methods section. Scale bar=20µm.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Lunatic fringe-mediated Notch signaling regulates adult hippocampal neural stem cell maintenance
eLife 6:e24660.
https://doi.org/10.7554/eLife.24660
