Aging entails distinct requirements for Rb at maintaining adult neurogenesis

Highlights

• •Rb is not a key player in maintaining neurogenesis in the aged SVZ.
• •Unlike in the young OB, Rb is required for early maturation of newborn neurons in the aged OB.
• •Rb suppresses cell cycle re-entry in post-mitotic adult-born neurons in the young and aged OB.
• •NestinCreERT2 line is efficient in studying acute KO effects in the mid-aged brain and to a lesser extent in the old-aged brain.

Abstract

Cell cycle proteins play essential roles in regulating embryonic and adult neurogenesis in the mammalian brain. A key example is the Retinoblastoma protein (Rb) whose loss disrupts the whole neurogenic program during brain development, but only results in increased progenitor proliferation in the adult subventricular zone (SVZ) and compromised long-term neuronal survival in the adult olfactory bulb (OB). Whether this holds true of neurogenesis in the aged brain remains unknown. In this study, we find no evidence of irregular proliferation or early commitment defects in the mid-aged (12-month-old) and old-aged (20-month-old) SVZ following tamoxifen-inducible Rb knockout (Rb iKO) in mice. However, we highlight a striking defect in early maturation of Rb-deficient migrating neuroblasts along the rostral migratory stream (RMS), followed by massive decline in neuronal generation inside the aged OB. In the absence of Rb, we also show evidence of incomplete cell cycle re-entry (CCE) along with DNA damage in the young OB, while we find a similar trend towards CCE but no clear signs of DNA damage or neurodegenerative signatures (pTau or Synuclein accumulation) in the aged OB. However, such phenotype could be masked by the severe maturation defect reported above in addition to the natural decline in adult neurogenesis with age. Overall, we show that Rb is required to prevent CCE and DNA damage in adult-born OB neurons, hence maintain neuronal survival. Moreover, while loss of Rb alone is insufficient to trigger seeding of neurotoxic species, this study reveals age-dependent non-monotonic dynamics in regulating neurogenesis by Rb.

1. Introduction

Aging is associated with physiological and cognitive deterioration of the mammalian brain, where it is thought to be the primary risk factor for most neurodegenerative diseases in humans [1]. To ameliorate or compensate for age-related threats to neuronal health, a growing effort aims to investigate the possibility of manipulating endogenous neurogenic regions within the adult brain – namely the subventricular zone (SVZ) and hippocampal dentate gyrus (DG) [2], [3], [4]. So far, however, the molecular profiling of adult neurogenesis in the aged brain remains incomplete, especially in terms of identifying the molecular mechanisms (including age-specific roles) for key proteins known to be essential during embryogenesis and young adulthood [5]. It is therefore critical to define how neurogenic regulation evolves with age and adopt the most appropriate tools to reflect its true physiological setting.

Adult-born neurons derived from the SVZ migrate along the rostral migratory stream (RMS) as immature neuroblasts and then fully differentiate in the olfactory bulb (OB) as inhibitory interneurons of the granule cell layer (GCL) and the periglomerular layer (GL) [6], [7]. With age, the SVZ witnesses an increase in neural stem cell (NSC) quiescence, permanent cell cycle exit (senescence) and defective lysosomal activity, along with a reduction in progenitor proliferation and subsequent neuroblast migration [8], [9], [10]. While the aged RMS remains intact [11], aging in the OB reveals a drop in overall neuronal turnover, potentially as a compensatory mechanism to the decline in neurogenesis levels [12], [13]. In addition, there is a decrease in afferent synaptic input and local modulatory synapses of OB glomeruli [14] as well as a deterioration in dopaminergic stimulation [15], which translates into several behavioral deficits in terms of fine olfactory stimulation, short-term olfactory memory and olfactory perceptual learning [15], [16], [17].

The Retinoblastoma protein, Rb, is a major regulator of the G1/S checkpoint in lineage specification of all cell types [18], [19]. In the developing central nervous system, the Rb-E2F pathway has been widely involved in the control of NSC activation, progenitor proliferation and neuronal differentiation [20], [21], [22], [23], [24]. More recent studies have looked at its role in the young adult brain (2–6 month-old), especially the SVZ and DG, where Rb was found to negatively regulate SVZ progenitor proliferation and is needed for the maintenance of long-term survival of adult-born OB neurons [25], [26]. These findings highlight distinct requirements for Rb and potentially other pockets proteins in the regulation of embryonic versus adult neurogenesis. Interestingly, some studies have also shown that Rb plays an essential role in maintaining the survival of postmitotic neurons [27], [28] such that the presymptomatic phase of neurodegenerative diseases e.g. Parkinson’s disease and Alzheimer’s disease was linked to cell cycle re-entry (CCE) and Rb deactivation in human postmortem tissues and mouse models [13], [29], [30], [31], [32], [33]. However, the role of Rb in the aging brain and its potential involvement in neuropathogenesis in vivo has not been established yet.

In this study, we investigated the role of Rb in the mid-aged and old-aged mouse brain, specifically along the neurogenic SVZ-OB axis, using two Rb tamoxifen-inducible knockout (Rb iKO) lines driven by the Nestin promoter. We report no apparent requirement for Rb in regulating NSC/progenitors’ proliferation or neuronal commitment in the aged SVZ. However, we delineate a more critical role for Rb in early maturation of adult-born neurons in the aged OB compared to the young OB. Moreover, in the absence of Rb, we show clear evidence of incomplete CCE and DNA damage in the young OB (and to a lesser extent in the old-aged OB), thus explaining the cell death phenotype of Rb-null newborn neurons reported earlier [25]. Of note, we also find that loss of Rb alone is insufficient to trigger seeding of neurotoxic species after long survival periods in the aged brain. However, such phenotype could be masked by the natural decline in adult neurogenesis with age and restricted to the Rb iKO line studied in this context. Overall, this study establishes that the requirement for Rb in regulating adult neurogenesis changes with age. Hence, Rb seems to play distinctive roles in the mid-aged and old-aged brain compared with the young brain, which highlights the dynamic variation in cell cycle control of NSC/progenitors and adult-born neurons with aging.

2. Methods

2.1. Mice maintenance and genotyping

Two NestinCreER^T2-Rosa26^YFP-Rb^flox/flox lines were used in this study. The first (Line-1) is based on the NestinCreER^T2 line previously obtained from Dr. Amelia Eisch’s lab at the University of Texas Southwestern Medical Center, USA through material transfer agreement. Dr. Eisch developed the mice in collaboration with Dr. Pierre Chambon at the Institute of Genetics and Molecular and Cellular Biology (IGBMC), Strasbourg, France [34]. The second line (Line-2) is the one obtained from Dr. Suzanne Baker’s lab and showing significantly higher recombination efficiency [35]. All experiments and data described in this study were performed using in Line-1 except for the experiments shown in Fig. 5 that were conducted using Line-2. Both lines were previously established to successfully induce, upon tamoxifen (Tam) administration, the conditional knockout of Rb and YFP reporter gene expression during adult neurogenesis (Line-1: [25], Line-2: [26]). All animal experiments were performed in accordance with the approved guidelines of the Institutional Animal Care and Use Committee (IACUC) of the American University of Beirut (AUB). Both male and female mice were used in this study, with their results combined and averaged. Aged animals were only derived from Line-1 and split into two age groups: 12–14 months old mice (mid-aged or MA) and 20 months old mice (old-aged or OA). Young mice (2 months old, YA) were derived from both lines.

Fig. 5.

Rb-iKO triggers cell cycle re-entry (CCE) and DNA damage response (DDR) in adult-born neurons inside the young adult OB. (A) Bar graph showing normalized YFP cell counts in the GCL at 28dpt in two Nestin-CreERT2-YFP transgenic lines used here with significantly higher recombination efficiency in Line 2 compared with Line-1. (B) Triple immunohistochemistry performed on OB coronal sections showing YFP (green), PCNA (red) and Ki67 (blue) in the GCL in YA Rb^+/− and Rb^−/− mice, and bar graph comparing percentages of PCNA+, Ki67+ and PCNA+ Ki67+ cells in the total YFP population in the GCL between genotypes. Note the increase in all three populations upon loss of Rb as a sign of CCE. (C) YA mice were subjected to birthdating studies with BrdU (refer to methods for detail). Triple staining for YFP (green), PCNA (red) and BrdU (blue) at 28dpt in both genotypes, bar graphs showing percentage of BrdU+ cells in total GCL YFP+ cells, percentages of YFP+, PCNA+ and YFP+ PCNA+ in total GCL BrdU+ cells, and percentage of PCNA+ in (YFP+ BrdU+) cell population in the GCL, respectively. Note the sharp increase in the percentage of newborn neurons (YFP+ BrdU+) co-expressing PCNA, confirming the presence of CCE inside the OB. (D) Triple immunostaining for YFP (green), γH2A.X (red) and pH3 (blue) at 28dpt in the GCL at YA in both genotypes with bar graph showing percentages of γH2A.X+, pH3+ and γH2A.X+ pH3+ in the total GCL-YFP cell population. This data shows DDR in Rb-null newborn YFP+ neurons as indicated by the high % of (γH2A.X+ YFP+) cells upon loss of Rb. White arrows indicate co-labeled YFP+ cells. Independent Student’s t-test: *p < 0.05, **p < 0.01, ***p < 0.001. Error bars correspond to S.E.M. Scale bar: 50 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Before treatment and sacrifice, DNA extraction of mice earpieces was performed, followed by PCR genotyping of three genes involved: Cre (Fwd: 5′-ATTTGCCTGCATTACCGGTC-3′; Rev: 5′-ATCAACGTTTTCTTTTCGG-3′), Rosa26YFP (Fwd wt: 5′-AAAGTCGCTCTGA GTTGTTAT-3′; Fwd mut: 5′-GCGAAGAGTTTGTCCTCAACC-3′; Rev: 5′-GGAGCGGG AGAAATGGATATG-3′) and Rb (Fwd1: 5′-GGCGTGTGCCATCAATG-3′; Rev1: 5′-AACTCAAGGGAGACCTG-3′; or Fwd2: 5′-CTCATGGACTAGGTTAAGTTGTGG-3′; Rev2: 5′-GCATTTAATTGTCCCCTAATCC-3′). After tamoxifen treatment, mice of NestinCreER-Rosa26^YFP-Rb^+/flox genotypes were designated as Rb^+/− (or controls), while those of NestinCreER-Rosa26^YFP-Rb^flox/flox as Rb^−/− (or mutants).

2.2. Tamoxifen and BrdU treatments

Mice of all age cohorts (YA, MA and OA) received intraperitoneal injections of tamoxifen (Tam, Sigma T5648, 30 mg/mL dissolved in 90% sunflower seed oil and 10% ethanol absolute) which was prepared daily before treatment and administered for 4–5 consecutive days according to mouse body weight (180 mg/kg). Mice were then sacrificed at 28 days post-tamoxifen (dpt), 60dpt, 210dpt or 480dpt. For Line-2 YA animals, Tam was similarly prepared but administered at 150 mg/kg/day for 5 days. These mice were then subject to a BrdU birthdating study, whereby they received BrdU injections at 100 mg/kg/day for 3 consecutive days after end of Tam treatment, and were later sacrificed at 28dpt [26].

2.3. Tissue fixation and cryopreservation

Before sacrifice, mice were euthanized with 1.5 µl/g Ketamine and 0.25 µl Xylazine and then underwent intracardial perfusion with ∼20 mL 1xPBS followed by ∼20 mL 4% cold paraformaldehyde (PFA). This was followed by dissection of brain tissues, which were further fixed overnight in 4% PFA. Then, brains were washed in 1xPBS and subject to serial dehydration through sucrose gradient (3 days in 20% sucrose, followed by 4–5 days in 30% sucrose) to attain cryoprotection. Last, brains were frozen on dry ice using isopentane at −35 °C and, when set to be sectioned, were embedded with Tissue-Tek O.C.T. (Surgipath). Brains were sectioned by a cryostat (Leica, CM1850) with a thickness of 8–10 µm per section (Line-1: sagittal; Line-2: coronal), mounted on SuperFrost adhesion slides (Thermo Scientific, Fisher Scientific and Leica) and stored at −80 °C.

2.4. Immunohistochemistry

Cryopreserved tissues were thawed at room temperature (RT) for 30 min then washed in 1xPBS. They were then incubated for 1 h in blocking solution (1% BSA (Sigma 2153), 0.3% Triton® X-100 (Amresco M143), 5% Donkey Serum in 0.1 M PBS). Primary antibodies were then prepared in blocking solution and added to blocked tissues for incubation overnight. Sections were washed the next day in 1xPBS three times, then incubated in secondary antibodies (1:400) and counterstained for Hoechst (1:50000) for 1–2 h. Last, slides were washed 3 times in 1xPBS and mounted with 3:1 PBS:Glycerol solution.

The following primary antibodies were used: chicken anti-YFP (1:1000; Abcam ab13970), rabbit anti-Ki67 (1:400, Encor RPCA), goat anti-DCX (1:400, Santa Cruz sc-8066), goat anti-Sox2 (1:150, Santa Cruz sc-17320), rabbit anti-pH3 (1:1000, Millipore 06–570), rabbit anti-CR (1:4000, Swant), mouse anti-PCNA (1:1000, Vector), rat anti-BrdU (1:400, Abcam ab6326), mouse anti-γH2A.X (1:400, Millipore 05–636), rabbit anti-AC-3 (1:500, Cell Signaling 9664), rabbit anti-H3K9me3 (1:400, Millipore 07–442), rabbit anti-MAPT (1:100, Life Span Bioscience LS-C48043), rabbit anti-α-Synuclein (1:100, Santa Cruz sc-7011) and mouse anti-phospho-tau AT8 (1:500, Invitrogen MN1020). For secondary antibodies (Abcam, Alexa Fluor, Molecular Probes, Invitrogen and Jackson Immunoresearch), the following were used: donkey anti-chicken 488, donkey anti-rabbit 488, donkey anti-mouse Cy3, donkey anti-rabbit Cy3, donkey anti-goat Cy3, donkey anti-rabbit Cy5, donkey anti-rat Cy5 and donkey anti-goat Cy5.

For Sox2, pH3, PCNA and γH2A.X, tissues were subject to antigen retrieval before blocking and staining. Slides were treated with 10 mM Sodium Citrate (Fisher Scientific BP327) at pH = 6 for 15 min at 95 °C. For BrdU staining, slides were incubated in 1 N HCl for 20 min at 37 °C, after which they were neutralized in 0.1 M Sodium Borate (Fisher Scientific S-249) at pH = 8.5 for 10 min, followed by two washes in 1xPBS before blocking and staining.

2.5. Imaging, cell counts and statistical analysis

Fluorescent images were captured with Leica DM6B microscope with UV light and digital camera. SVZ, RMS and OB sections were matched at medial levels between genotypes and across timepoints with 3–5 non-consecutive sections counted per brain for each marker (n = 3 brains per genotype, and, n = 4 mutants in Fig. 3). In addition, raw counts were normalized either to perimeter in mm (for SVZ counts) or to surface area in mm² (for RMS and OB counts). Image overlapping and analyses were carried out using the Leica Application Suite (LAS X) software and ImageJ. All statistical analysis was based on independent Student’s t-test and univariate ANOVA (using SPSS 20) for genotype comparison, as well as 28dpt vs. 60 dpt and MA vs. OA comparisons; p value was assigned to < 0.05 to detect statistical significance. Data is presented as mean ± standard error of the mean (S.E.M.). All data were generated in at least 3 animals per genotype. For MA-28dpt Rb^+/− brain sections, only one brain tissue was available for YFP co-immunohistochemistry (e.g. with Ki67, DCX, etc.). Instead, age-matched control mice (lacking Cre genotype, n = 3) were used to fill the gaps of single labeled cells expressing these markers. Data visualization was performed using GraphPad Prism 8 and BioRender.

Fig. 3.

Defective neuronal maturation along the RMS in MA and OA mice following Rb iKO. Double immunostaining on brain sagittal sections with YFP (green)-DCX (red) and counterstained with Hoechst (blue), marking the three partitions of the RMS: VA, ventral arm; E, elbow; HA, horizontal arm in MA-28pt (A) Rb+/- control mice and (B) Rb-/- mutant mice. Insets in A and B show higher magnification images of the HA partitions, respectively. Bar graphs showing cell counts normalized to surface area in mm² in each RMS partition of (C) YFP+ cells and (D) (YFP+ DCX+) cells in both genotypes at timepoints indicated. (E) Graph illustrating % of (YFP+ DCX+) cells in total YFP population in each RMS partition in both genotypes at various timepoints in MA and OA mice. Compared with Rb+/- controls, Rb-iKO mice display a severe decline in the numbers of YFP+ cells and (YFP+ DCX+) cells detected in the VA and HA at MA and OA, primarily at 28dpt (A-D), with no difference in % of double positive cells in the total YFP population. This data highlights a defective neuronal maturation of migratory neuroblasts in the aged brain in the absence of Rb. Independent Student’s t-test were performed in C: *p < 0.05, **p < 0.01. Univariate ANOVA test was run in C and D: # p < 0.05, ## p < 0.01, ### p < 0.001. ‡ in D and E refers to data not assessed due to lack of matching levels in the RMS elbow at OA-28dpt, and, very low cell counts at OA-60dpt in both genotypes, respectively. Error bars correspond to S.E.M. Scale bar: 100 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3. Results

3.1. Maintenance of neurogenesis in the aged SVZ is not affected by loss of Rb

We previously used a tamoxifen inducible NestinCreER^T2/Rosa26^YFP/Rb^flox mouse line to induce an acute deletion of Rb in the SVZ of young adult mice (YA, ∼2 months old). This Rb iKO resulted in a controlled increase in proliferating neuronal progenitors (but not NSCs) in the absence of any major neuronal differentiation and maturation defects [25]. In this study, we relied on the same line (designated Line-1 here) to examine the role of Rb in the aging brain at two timepoints of treatment: mid-age (MA, 12–14 months old) and old-age (OA, 20 months old), whereby mice were sacrificed at 28-days and 60-days post-tamoxifen (dpt) for each group (refer to methods for detail). In both MA-SVZ and OA-SVZ, we found significant efficiency of Cre recombination in Line-1 in control mice as we recently established [36]. To test the effect of Rb iKO in the MA-SVZ and OA-SVZ, we co-stained YFP with stage-specific markers including Ki67 (a proliferation marker), DCX (an early commitment/differentiation marker), Sox2 (a late stem/early progenitor marker) and pH3 (a late mitotic marker) (Fig. 1, Fig. 2) [37].

Fig. 1.

Conditional loss of Rb has no major effect on SVZ-NSCs/progenitors’ proliferation in mid-aged (MA) and old-aged (OA) mice. Double immunostaining performed on brain sagittal sections in the aged SVZ with (A) YFP (green)-Ki67(red)-Hoechst (blue) in MA and OA Rb^+/− control and Rb^−/− mutant mice at 28dpt and 60dpt. Bar graphs showing normalized cells counts inside the SVZ of (B) YFP+ cells, (C) total Ki67+ cells including Ki67+ YFP+ cells and (D) % of Ki67+ cells in the total YFP population at various timepoints in both genotypes. Note the age-associated decrease in overall cell proliferation inside the SVZ between MA and OA mice as well as MA-28dpt and MA-60dpt in graphs C and D. The increase in % (YFP+ Ki67+) in total YFP population between OA-28dpt and OA-60dpt in D is likely due to the decrease in the size of YFP population associated with cell quiescence with age as seen in B. Asterisks indicate statistical significance of independent Student’s t-test: *p < 0.05, **p < 0.01, ***p < 0.001. Error bars correspond to standard error of the mean (SEM). Scale bar: 50 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2.

Neuronal commitment in the SVZ does not seem to be affected by loss of Rb in MA and OA mice. Double immunostaining performed on brain sagittal sections of the aged SVZ with (A) YFP(green)-DCX(red), (D) YFP (green)-Sox2(red), and (E) YFP(green)-pH3(red) in MA and OA Rb^+/− control and Rb^−/− mutant mice at the timepoints indicated (Hoechst in blue). Bar graphs showing normalized cell counts inside the SVZ of (B) total DCX+ cells including YFP+ cells, (F) total Sox2+ cells including YFP+ cells, and (G) total PH3+ cells including YFP+ cells in both genotypes at indicated timepoints. (C) Graph showing % of DCX+ cells in total YFP population inside the SVZ. Independent Student’s t-tests showed no significant difference in the numbers of early stem/progenitors cells (YFP+ Sox2+) and neuroblasts (DCX+ YFP+) including dividing (Ki67+) and post-mitotic (pH3+) cells between genotypes in both age groups and timepoints studied. Error bars correspond to standard error of the mean (SEM). Scale bar: 50 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

We report no significant differences in the expressions of the above markers inside the SVZ between Rb^+/− (controls) and Rb^−/− (iKO mutants) mice at MA-28dpt, MA-60dpt, OA-28dpt and OA-60dpt timepoints, separately (Fig. 1A, Fig. 2A, D and E). For instance, at MA-28dpt, single cell counts were similar inside the SVZ between genotypes with respect to total Ki67+ cells (Fig. 1A and C, Rb^+/−: 17.7 cells/mm ± 7.2 S.E.M. vs. Rb^−/−: 16.3 cells/mm ± 2.0, p = 0.87, n = 3 per genotype), DCX+ cells (Fig. 2A and B, Rb^+/−: 18.6 cells/mm ± 4.0 vs. Rb^−/−: 19.5 cells/mm ± 3.25, p = 0.88), Sox2+ (Fig. 2D and F, Rb^+/−: 114 cells/mm ± 26.3 vs. Rb^−/−: 87.3 ± 5.19 cells/mm, p = 0.37) and pH3+ cell counts (Fig. 2E and G, Rb^+/−: 2.34 cells/mm ± 0.76 vs. Rb^−/−: 3.56 cells/mm ± 0.496, p = 0.37). In addition, we report no effect of Rb iKO on co-expression of the above markers with YFP+ cells as double positive cell counts normalized to the SVZ perimeter (Fig. 1C, Fig. 2B, F and G) or percentages (%) of the total YFP+ population. However, we detected a significant decline in Ki67+ cells and % of Ki67+ of total YFP population between MA and OA (Fig. 1C and D, 2.3 fold decline in % Ki67, p = 0.0001), consistent with a decline in cell proliferation with age as reported earlier [5], [8]. This decline was also detected in % of Ki67+ cells in the total YFP+ population between MA-28dpt and MA-60dpt (Fig. 1D). In contrast, we saw a significant increase in % Ki67+ among the total YFP population between OA-28dpt and OA-60dpt (Fig. 1D, 2.2 folds, p = 0.002), but this is likely explained by the drop of total YFP+ cell counts and not that of Ki67+ or YFP+ Ki67+ at OA-60dpt (compare Fig. 1B–D). As for DCX, we report no significant differences in normalized (YFP+;DCX+) counts or percentages across all four timepoints (Fig. 2B and C).

Hence, compared with young adult, we found that Rb alone does not seem to play a distinct role in maintaining the aged SVZ niche in terms of progenitor proliferation or neuronal commitment. Moreover, our data corroborates earlier findings on the overall decline in SVZ proliferation and NSC depletion with aging.

3.2. Rb deletion impairs early maturation of adult-born neurons in mid-aged and old-aged mice, leading to compromised neurogenesis in the OB

Beyond its role in NSC/progenitor regulation, Rb was found to control neuronal commitment and migration in the embryonic brain [22], [38] but not during adult SVZ neurogenesis in young-aged mice [25]. To address this role during aging, we examined the expressions of YFP and DCX in MA and OA mice following Rb iKO, along the RMS in its three partitions: the vertical arm (RMS-VA), the elbow (RMS-E) and the horizontal arm (RMS-HA) [11] (Fig. 3A and B and data not shown). We report a significant decrease in normalized YFP+ counts in the RMS-HA in Rb^−/− mice compared to matching Rb^+/− controls at MA-28dpt (Rb^+/−: 915 cells/mm² ± 160 vs. Rb^−/−: 75 cells/mm² ± 15, p = 0.035, n = 3 per genotype) as well as in the RMS-VA and RMS-HA at OA-28dpt (for RMS-VA, Rb^+/−: 1979 cells/mm² ± 446 vs. Rb^−/−: 700 cells/mm² ± 136, p = 0.025, n = 3 Ct and n = 4 mut, Fig. 3C). Of note, a similar trend towards decline in YFP+ cells is observed in the RMS partitions in both age groups at 60dpt, albeit not significant due to low/very low cell counts (Fig. 3C). The above phenotype is also affecting the DCX+ population, given that we detected a significant and gradual decrease in normalized cell counts of (YFP+ DCX+) cells in RMS-VA and RMS-HA in Rb iKO mice compared with Rb+/- controls in both age groups at 28dpt (compare Fig. 3D and C). This decrease was also noticeable in MA-60dpt but not OA-60dpt where the rate of neurogenesis is greatly reduced with age, thus restricting proper statistical analyses (Fig. 3D). In parallel, the effect of Rb iKO on the DCX+ population of migratory neuroblasts was proportional in all age groups and time points as indicated by similar % of (YFP+ DCX+) cells in the total YFP population between genotypes (Fig. 3E). To test whether the above defect in early commitment is associated with increased cell death, we stained the RMS for active-caspase 3 (AC-3; apoptotic marker) at all timepoints, but did not detect a significant difference in signal due to low level of expression along the RMS in both genotypes (data not shown). Overall, our data indicates that Rb is essential for early maturation of newly born neuroblasts migrating across the whole RMS in mid-aged and old-aged mice.

We next looked for further alterations in neurogenesis following Rb iKO in the aged OB. Compared with Rb^+/− controls, we observed a sharp decline in total YFP+ cell counts inside the GCL and the GL of Rb iKO mice at MA-28dpt but not at other timepoints (Fig. 4A–C; MA-28dpt GCL Rb^+/−: 154 cells/mm² ± 22 vs. Rb^−/−: 8 ± 5 cells/mm², p = 0.0012; MA-28dpt GL Rb^+/−: 178 cells/mm² ± 9 vs. Rb^−/−: 22 ± 17 cells/mm², p = 0.0062, n = 3 per genotype). Moreover, Rb^+/− controls showed a significantly higher YFP expression in both the GCL and the GL at MA-28dpt compared with YA-28dpt and OA-28dpt (Fig. 4A–C). This suggests that, in contrast to OB neurogenesis in young adult mice, there is a significant boost in the number of immature adult-born neurons produced in the mid-aged OB. However, this increase does not seem to be sustained after longer survival periods e.g. 30 days later, in MA-60dpt, probably due to lack of further pro-survival signals and/or a higher neuronal turnover rate inside the OB (Fig. 4B and C). Across timepoints, a significant decline in total YFP+ cells is seen in both the GCL and the GL between MA-28dpt and OA-28dpt Rb^+/− controls (p = 0.001); yet it is partially recovered in OA-28dpt Rb^−/− inside the GL but not GCL (OA-28dpt GL Rb^+/−: 78 cells/mm² ± 9 vs. Rb^−/−: 116 cells/mm² ± 17, p = 0.14, n = 3 per genotype) (Fig. 4B and C). This data suggests the existence of a distinct requirement for Rb during neurogenesis inside the OB of mid-aged compared to old-aged mice, which could also be layer/cell type specific.

Fig. 4.

OB neurogenesis is severely reduced in the mid-aged brain following Rb iKO. (A) YFP immunostaining (green) counterstained with Hoechst (blue) on brain sagittal sections in the GCL (granule cell layer) and GL (periglomerular layer) of the OB in young aged YA, MA and OA Rb^+/− and Rb^−/− mice at 28dpt and 60dpt. Bar graphs designate YFP+ cell counts normalized to surface area in mm² in the (B) GCL and (C) GL at indicated timepoints in Rb^+/− and Rb^−/− mice. Note the sharp boost in OB neurogenesis (YFP+ cells) detected in Rb^+/− mice at MA-28dpt compared with YA-28dpt and OA-28dpt (A-C). Loss of Rb negatively affects this increase at MA-28dpt in both OB layers. (D) Scheme showing downregulation of DCX and upregulation of CR during normal neuronal maturation inside the OB. Panel in D shows double immunohistochemistry of DCX (green) and CR (red) in the GCL of Rb+/- mice. Bar graphs showing CR+ cell counts normalized to GCL surface area in mm², and % of DCX+ cells in total CR population in the GCL between age groups in Rb+/- control mice. Note the delayed neuronal maturation in MA-control brains as indicated by the high percentage of DCX+ co-labeled with CR (immature neurons) and the low number of CR+ cells only (mature neurons). Independent Student’s t-test; *p < 0.05, **p < 0.01, ***p < 0.001. Error bars correspond to S.E.M. Scale bar: 50 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

To confirm the presence of an age-dependent discrepancy in adult neurogenesis dynamics, we co-labeled for DCX and Calretinin (CR; an interneuron-specific marker) in the GCL of YA, MA and OA in Rb^+/− controls. As newborn OB interneurons mature, they downregulate DCX expression and a large subset of these neurons upregulate CR expression, thus marking a late stage of maturation. We observed a significantly higher percentage of DCX+ neurons among the total CR+ population from YA to MA which then drops at OA, suggesting the existence of a longer survival window for immature neurons and/or delayed maturation at this age (% DCX+ in CR+ population in YA: 6.7% ± 0.95 vs. MA: 14.6% ± 0.7 vs. OA: 2.0% ± 1.2, n = 3 Ct). Notably, this can also be confirmed by the lower count of CR+ cells/mm² in the MA GCL compared to YA GCL and OA GCL (Fig. 4D). Altogether, the above results further suggest the presence of noticeable spatio-temporal differences in neurogenesis dynamics inside the OB with age, however, this requires further investigation in future studies.

3.3. Rb suppresses cell cycle re-entry and DNA damage response in adult-born neurons in the young OB

Given the early maturation defects observed in the RMS-OB with age in the absence of Rb, we examined the role of Rb at later postmitotic stages. Cell cycle re-entry (CCE) was shown to precede neuronal loss in postmitotic Rb-deficient embryonic and adult cortical neurons and would so qualify as a candidate mechanism underlying the increase in long-term cell death following Rb iKO in adult-born neurons [27], [39], [40]. We therefore co-stained MA and OA OB sagittal sections with YFP, PCNA and Ki67 at 28dpt in both genotypes. Yet, we could not assess the phenotype in MA-28dpt mice due to the severe decline in OB neurogenesis associated with Rb loss (as reported above). On the other hand, compared with Rb+/- controls, we noticed a significant increase in % PCNA+ cells but not % Ki67+ or % Ki67+ PCNA+ cells out of total YFP population inside the GCL in OA-28dpt Rb iKO mice (Suppl. Fig. 1A). Once again, phenotypic analysis was hindered by the very low cells counts in the latter two cells populations at OA. Instead, we reverted to the YA OB, which showed a severe compromise in long-term survival of adult-born neurons following Rb iKO [25]. In addition, we made use of second NestinCreER^T2-Rosa26^YFP-Rb^flox/flox line (Line-2, [26]) that is characterized by a higher normalized YFP+ cell counts throughout the SZV-OB axis at YA-28dpt (higher Cre recombination efficiency than our line of choice Line-1, [25]) (Fig. 5A and data not shown). Interestingly, our results showed that, out of total YFP population, 16% and 2.5% co-labeled with PCNA and both (PCNA, Ki67) in the GCL of Rb iKO mice compared to near zero-levels in Rb controls, respectively (Fig. 5B), suggesting the presence of CCE entry in mature neurons following loss of Rb. In comparison, we detected somehow a similar trend in CCE profile using Line-1 but with no statistical difference due to lower cell counts as expected (Suppl. Fig. 1B).

To confirm this and assess the level of OB neurogenesis at YA-28dpt in Line-2, mice were subject to BrdU birthdating (several BrdU injections 3 weeks before sacrifice; refer to methods for detail). As a result, we detected a 2.1 fold reduction in the percentage of BrdU+ cells out of the total YFP+ population in the GCL of Rb iKO mice, compared to matching controls (Rb^+/−: 5.8% ± 0.4 vs. Rb^−/−: 3.1% ± 0.2, p = 0.015, n = 3 per genotype), thus confirming the presence of a cell death phenotype as previously reported in Line-1[25] (Fig. 5C). While we did not detect a difference in the percentage of YFP+ cells out of total BrdU+ cells between genotypes, we observed a significant and similar increase in the percentages of PCNA+ and YFP+ PCNA+ cells in Rb iKO compared with Rb controls (∼6% increase in each population in the absence of Rb, p < 0.05). Importantly, ∼17% of the YFP+ BrdU+ population (postmitotic newborn neurons) co-expressed PCNA in Rb iKO mice compared to 1.4% in controls only (Fig. 5C), therefore confirming CCE and ruling out the presence of a cell cycle exit defect. Moreover, this increase matches the percentage of PCNA+ cells out of all YFP+ cells (16%) observed earlier inside the GCL. Taken together, the above data shows that, rather than causing a delay in cell cycle exit, loss of Rb triggers CCE in postmitotic adult-born neurons of the YA OB.

Another marker found to be associated with CCE following loss of Rb is the phosphorylation of H2A.X, an indicator of DNA damage response (DDR) [40], [41]. In apoptotic neurons, the expression of this signal does not overlap with phospho-histone 3 (pH3; G2-M phase marker) following CCE as shown earlier [42]. Therefore, we checked the expression of H2A.X (γH2A.X) and pH3 in YFP+ cells of the GCL of Line-2 mice. We found that Rb iKO mice showed a significant 10-fold increase in the proportion of γH2A.X+ of total YFP+ cells compared to Rb controls (p = 0.007), but not in the percentages of pH3+ or γH2A.X+ pH3+ in this population (Fig. 5D). We therefore suggest that incomplete DDR activation following CCE is likely the trigger for neuronal loss following Rb iKO. Notably, we also observed a trend towards an increase in DDR signature in Line-1 inside the GCL following Rb iKO at YA-28dpt (but not OA-28dpt). Yet, this did not reach statistical significance, which is likely due to the lower recombination efficiency in this line compared to Line-2 (Suppl. Fig. 1).

3.4. No apparent signs of neurotoxic species deposition after long survival periods in Rb iKO aged mice

Rb is a key regulator of cellular senescence [43], [44] which might also be the case for NSCs in the aging SVZ. Moreover, Rb dysregulation was linked to neuronal abnormalities and pathogenesis of neurodegenerative diseases in human postmortem tissues [29], [31], [33], [45]. To assess these outcomes in our Rb iKO model, we sacrificed mice at longer intervals post-tamoxifen treatments, specifically at MA-210dpt and YA-480dpt, expecting to capture a more optimal YFP signal and potentially neurotoxic signs in the OB. At MA-210dpt, we report no variation between Rb^+/− and Rb^−/− mice in SVZ YFP+ (Rb^+/−: 12.73 cells/mm ± 3.03 vs. Rb^−/−: 12.69 cells/mm ± 2.4, p = 0.99), Ki67+, YFP+ Ki67+, PCNA+ (Rb^+/−: 11.26 cells/mm ± 0.97 vs. Rb^−/−: 14.22 cells/mm ± 2.16, p = 0.28) and YFP+ PCNA+ cell counts (Fig. 6A, n = 3 per genotype). In addition, we detected similar expression patterns of YFP and H3K9me3 (a senescence marker) in the MA-210dpt SVZ in both genotypes, suggesting no evident change in NSC quiescence in the aging SVZ in the absence of Rb (Fig. 6B). At the same timepoint, we found that the total YFP+ cells are not significantly different in the GCL and the GL between genotypes, respectively (e.g. GCL YFP+ in Rb^+/−: 56 cells/mm² ± 16 vs. Rb^−/−: 96 cells/mm² ± 18, p = 0.17). Moreover, at MA-210dpt, no major effect of Rb iKO on CCE or DDR was noticed (Fig. 6C, n = 3 per genotype), nor an alteration in the co-expression of YFP+ with two key proteins implicated in early neuropathogenesis, i.e. MAPT (total Tau; Suppl. Fig. 2A), pTau (Fig. 6D) or Synuclein (Fig. 6E). Similarly, we did not detect any sign of CCE, DDR or Tau/Synuclein aggregation following Rb iKO at YA-480dpt (Suppl. Fig. 2B-C). This data indicates that loss of Rb alone is not associated with the development of long-term neurodegenerative signatures in the aging OB (at least in Line-1), however, such a phenotype could be masked by cell death associated with loss of Rb after longer survival periods (refer to discussion).

Fig. 6.

Rb-iKO mice show no signs of CCE, DDR or neurodegeneration along the SVZ-OB axis at MA-210dpt. (A) Bar graphs showing normalized cell counts inside the SVZ of total YFP+ cells, total Ki67+ cells including (YFP+ Ki67+), and total PCNA+ cells including (YFP+ PCNA+) cells in Rb^+/− and Rb^−/− mice at MA-210dpt. (B) Double immunostaining showing expressions of YFP (green) and H3K9me3 (red) with no significant difference in signal between genotypes inside the SVZ at MA-210dpt. (C) Bar graphs showing, left, YFP+ cell counts normalized to surface area in GCL and GL in Rb^+/− and Rb^−/− brains; middle, percentages of PCNA+, Ki67+ and (PCNA+ Ki67+) cells out of total GCL YFP+ cells; and, right, percentages of γH2A.X+, pH3+ and (γH2A.X+ pH3+) out of total GCL YFP+ cells with no significant difference between genotypes at MA-210dpt. (D) Double immunohistochemistry showing YFP (green) and pTau (red) expressions with normalized counts of pTau+ and (YFP+ pTau+) cells in the GCL of both genotypes at MA-210dpt. (E) Double immunohistochemistry showing YFP (green) and Synuclein (red) expression with normalized counts of (YFP+ Syn+) cells in the GCL of both genotypes at MA-210dpt. Data indicates no major change in the expression patterns of pTau and Syn (or accumulation of aggregates) between genotypes. Error bars correspond to standard error of the mean (SEM). Scale bar: 50 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

We investigated here the role of Rb along the neurogenic SVZ-OB axis of the aging mouse brain. We reported graded changes in adult neurogenesis across distinct age groups upon loss of Rb (young adult versus mid-aged versus old-aged) (Fig. 7). Unlike in the young adult brain, we found that Rb seems dispensable for the maintenance of SVZ neurogenesis, particularly the control of neuronal proliferation and commitment. In contrast, Rb is needed for early maturation of adult-born neuroblasts or immature neurons in the mid-aged and old-aged OB compared to the young OB (showing no major maturation defects in the absence of Rb). Moreover, Rb is critical to maintain survival of adult-born neurons by suppressing CCE and DNA damage in the young adult brain, and is likely doing a similar function in the aged brain. We also noted interesting dynamic changes during SVZ-OB neurogenesis across short time periods in control mice, particularly a noticeable yet transient increase in the production of immature neurons in mid-aged mice compared to young and old–aged mice. Overall, this study establishes distinctive requirement for Rb in regulating adult neurogenesis in the aged versus young brain, which highlights the dynamic variation in cell cycle control of NSC/progenitors and adult-born neurons with aging.

Fig. 7.

Summary of the distinct roles played by Rb during adult neurogenesis with age. Aging entails distinct requirement(s) for Rb in the regulation of adult neurogenesis inside the OB. Loss of Rb in the young adult brain leads to enhanced progenitor proliferation with no major defects in neuronal differentiation or maturation, but a severe compromise in long-term survival of newborn OB neurons [25], which is due to high predisposition for CCE and DDR. In contrast, in the mid-aged brain, conditional deletion of Rb does not influence progenitor proliferation or neuronal commitment but negatively affects maturation of neuroblasts, and potentially their survival (sharp decrease in OB neurogenesis). In the old aged brain, loss of Rb seems to be associated with a similar phenotype as in the mid-aged brain, with signs of CCE in Rb-null newborn OB neurons. Of note, phenotypic assessment in OA mice could be masked by the overall decline in the level of neurogenesis with age.

4.1. NestinCreER^T2/Rosa26^YFP Line is efficient in measuring SVZ-OB dynamics in Mid-aged brain but is more restricted at old age

We first emphasize our reliance on the transgenic Nestin-CreER^T2 mouse model to study adult neurogenesis during aging despite the presence of some technical limitations. Several Nestin-driven transgenic lines have been generated and widely used to trace the acute outcomes of genetic knock-in and knock-out models in young adult mice [46], [47]. One recent study used a Nestin line to trace stem/progenitor populations in the SVZ from neonatal up to 12 months of age [48]. However, no study to our knowledge has used this model to address the same effects in the mid-aged or old-aged SVZ/DG. Instead, previous reports have repeatedly relied on germline transgenic mice e.g. Btg1^−/− [49], p16^−/− [50] and p21^−/− [51] mice to name a few, despite a potential carryover KO effect from embryonic and young adult precursors to the aging brain. To assess the efficiency of genetic recombination in Nestin-CreER^T2-Rosa26-YFP in mid-aged and old-aged mice (Line-1, [25]), we recently examined the co-localization of the reporter YFP protein expression with Cre mRNA, and Nestin expression in the SVZ in wild-type control mice. We reported high overlap of YFP expression with both signals in mid-aged and old-aged brains, which marks specific and efficient recombination of Cre/loxP system in this line (Line-1) [36]. However, we did not assess further the YFP reporter signal along the RMS or inside the OB.

Several studies have reported a decline in SVZ stem/progenitor proliferation due to aging [8], [16], [52], [53], [54], which we replicated in our model here by comparing MA to OA brains but also MA-28dpt to MA-60dpt (Fig. 1D). In addition, we noted an increase in the percentage of YFP proliferating precursors between OA-28dpt and OA-60dpt (Fig. 1D), yet this is likely due to a drop in total YFP labeled cells in the SVZ at the latter timepoint (Fig. 1B). The decline hence seems limited to the (YFP+ Ki67-) cell population. While this could speak of a survival advantage of (YFP+ Ki67+) cells over the previous subpopulation, it is likely associated with a depletion-like effect of quiescent non-proliferating NSCs in the OA SVZ. Indeed, this was recently portrayed as a non-monotonic ‘reversal of trend’ effect before and after 18 months of age in SVZ dynamics, marked by a recovery of the MASH1+ population in number and proliferation rate in vivo at 22 months – along with faster division of activated type B/type C clones in vitro [55]. Meanwhile, we found no major age-dependent changes to SVZ neuronal commitment based on DCX expression between MA and OA timepoints (Fig. 2A–C), which is in line with some studies findings [52], [54] but in contrast to other reports showing a monotonic decline of DCX expression from 2 to 22 months-old SVZ [55]. The discrepancy in these studies could be due to technical limitations associated with the models used e.g. lower cell counts and progressive decline in the level of neurogenesis with age (which hinders proper statistical analysis), as well as time-points variations and or differences in strain background.

In the mid-aged OB, we identified a transient boost in the numbers of immature adult-born neurons marked by YFP at MA-28dpt compared to the YA and OA OBs in both the GCL and GL in Rb heterozygote controls. However, this effect was diminished following loss of Rb in MA-28dpt where YFP cell counts were comparable to the other timepoints (Fig. 4A and B). Moreover, the above transient effect was not detected in MA-60dpt (28 days later), which may underline a highly dynamic regulation in neuronal production or turnover even across short survival periods e.g. an effect restricted to immature neurons. Previous studies have relied on thymidine analog birthdating assays to measure neuronal turnover in the aging OB, where Bouab et al. showed a 77% decline in BrdU-retaining cells in the GCL of mid-aged compared to young adult brains [53]. Similarly, Shook et al. reported a 75% drop in EdU+ cells in the whole OB (and the GL only) between 3-months old and 20-months old mice [54]. While this approach is indicative of major changes occurring across different age groups, it assumes no effects of aging on SVZ neurogenesis dynamics between distinct time-points (e.g. as reported in our study) such that thymidine analogs are equally captured to label or mark the birthdate of adult-born neurons [56]. In fact, since this is not the case, Bouab et al. also reported a −67% drop- in BrdU+ SVZ cells at MA compared to YA (a lower drop compared to 77% in the OB) [53]. Therefore, results from birthdating assays should be interpreted with caution and may not be a very accurate metric for measuring survival dynamics of adult-born neurons particularly between different age groups. Instead, a more accurate approach would be to run a time-course birthdating analyses at each age separately to deduce the turnover rate in the OB and compare it accurately to YA mice [57]. BrdU/EdU-retaining cells are also born 2–3 weeks before mice sacrifice and therefore mark only a subset of postmitotic adult-born neurons arriving and undergoing integration in the OB circuit. Alternatively, a reporter gene expression e.g. YFP in this study can capture better graded changes in the spectrum of this lineage and could be more reliable to monitor cell dynamics across distinct timepoints, or when comparing short survival times as seen in the case of immature neurons dynamics reported here. Nonetheless, this method also shows some limitations attributed to incomplete recombination and/or lower recombination efficiency in some lines compared to others (as seen in Line-1 compared to Line 2). Moreover, it becomes more restricted in term of analytical depth in old-age mice compared with MA due to a major decline in neurogenesis, which limits statistical analyses (compare data in MA versus OA in Fig. 1B and C, Fig. 2B, Fig. 3C and D).

4.2. Distinct requirements for Rb in maintaining the aged SVZ-OB axis

From embryonic to young adult neurogenesis, we recently made the case for a significant restriction in the roles played by Rb, such that it specifically controls progenitor proliferation in the YA SVZ without having a distinct role (direct or distinct) in NSC quiescence, lineage commitment and neuronal differentiation or maturation as seen in the developing brain [20], [21], [22], [25], [38]. This could be associated with contextual changes affecting the SVZ environment with age and entailing differential requirement for Rb in the control of adult neurogenesis. In fact, as reported here in the MA and OA SVZ, Rb becomes even more dispensable to SVZ NSCs/progenitors’ development with age in terms of quiescence, proliferation and fate determination (Fig. 1, Fig. 2). It might even be the case that Rb gets gradually inactivated with age, as indirectly suggested by increased E2f1 expression in cortical lysates of mice older than 9 months compared to younger mice [58]. Alternatively, the restricted roles attributed to Rb in the aging brain could be justified by compensatory roles played by other pocket proteins such as p107 and/or p130. In fact, we recently found that, in the young postnatal brain, loss of Rb becomes increasingly compensated for by the other two pocket proteins, p107 and p130, where a triple KO of Rb, p107 and p130 (TKO) in the YA SVZ recapitulates most developmental defects seen with single Rb iKO brains during embryogenesis (Bejjani A and Ghanem N, unpublished data). Indeed, besides Rb, p107 plays a critical role in this context as it directly controls NSC self-renewal and neuronal commitment simultaneously in the YA SVZ-RMS region [59], [60]. Whether this compensatory mechanism is more prominent in the aging brain remains to be clearly established.

Along the RMS, we remarkably found that Rb iKO compromises early maturation of neuroblasts in the MA and OA brain, an effect that is more prominent at 28dpt compared with 60dpt at both ages (Fig. 3A-D). This defect is also reflected at the level of neurogenesis in both the GCL and the GL at MA-28dpt (Fig. 4A–C). In fact, several studies established a requirement for Rb in neuronal migration and differentiation during embryogenesis, with one study underlining its role in suppressing E2F-mediated neogenin expression, known to be required for proper migration of ventrally-derived cortical interneurons [20], [22], [38], [61]. However, this was not the case during adult neurogenesis along the YA SVZ-OB axis [25]. Moreover, despite the significant drop of YFP+ cells and (YFP+ DCX+) cells in the VA-RMS and HA-RMS of MA-28dpt and OA-28dpt Rb iKO brains, we did not detect an associated increase in AC-3+ cells (data not shown). This can be explained by a relatively low number of YFP+ cells along the RMS as to detect a significant variation in the much fewer AC-3+ cells. Yet, Svoboda et al. reported that the role of Rb and p107 in neuronal apoptosis is independent of their role in migration during embryonic cortical lamination; rather, Rb/p107 double null neuroblasts accumulate in the neurogenic intermediate zone and fail to migrate [24]. Therefore, a loss of defective immature neurons along the RMS and/or their accumulation in the SVZ could explain the drop in YFP+ cells reported above. Of note, we detected a unique effect of Rb iKO that seems to prolong the ‘survival’ of GL YFP+ cells at OA-28dpt and OA-60dpt more than in matching control mice although the difference between genotypes did not reach statistical significance (Fig. 4C). We believe that this phenotype is likely to be true in the absence of Rb given that we and other have previously noticed differences in cellular and physiological properties attributed to the adult-born neurons populations in GL versus GCL [17], [25], [62]. For instance, previous studies have shown major differences in susceptibility to cell death (rate of neuronal survival) between the two layers under normal conditions (for comprehensive review, check [63] – Table 1). A similar observation was previously reported for the OA (24-month-old) OB which witnessed decreased cell death compared to the YA (2 month-old) OB through downregulation of procaspase-3 [12]. Yet, it is not clear whether this is the case here or if this could be linked to a lower turnover rate in aged mice. Alternatively, this above phenotype could be linked to non-monotonic roles for Rb in the OA but not MA GL, which warrants further investigation in the future.

In the YA-OB, we uncovered that loss of Rb triggers CCE and DNA damage in postmitotic adult-born neurons, which in turn lead to a higher susceptibility to neuronal death (Fig. 5), and a similar trend in Line 1 albeit not significant (Suppl. Fig. 1B). Interestingly, we report a similar phenotype in OA-28dpt OB despite the severe reduction in the total number of YFP+ cells because of the maturation defect reported earlier and the progressive decline in adult neurogenesis with age (Suppl. Fig. 1A). Interestingly, this effect seems consistent in the developing and adult rodent cortex [41], and was also reported as a result of Rb inactivation following miR-26b overexpression in human cortical neurons in vitro [29]. Moreover, Oshikawa et al. reported a stronger CCE effect in Rb-p107-p130 TKO cortical neurons of pMAP2-Cre electroporated mouse embryos than in any combination of double KO genotype of pocket proteins, suggesting that p130 and p107 act partially to compensate for Rb deficiency in suppressing CCE in postmitotic neurons [40]. Future studies should address whether such compensatory roles exist among pocket proteins in the young versus aged brain. In the MA-210dpt and YA-480dpt OB, we did not observe signs of CCE or DNA damage, nor did we find evidence of abnormal Tau accumulation or α-synuclein expression in YFP+ cells following Rb iKO (Fig. 6 and Suppl. Fig. 2). Nonetheless, the YFP signal at prolonged sacrifice timepoints is likely biased towards accumulation of incompletely recombined e.g. YFP+ Rb^−/flox cells, as we previously reported in the YA OB 4 months following loss of Rb [25]. Moreover, the absence of a clear phenotype could be attributed to cell death in the absence of Rb (phenotype observed in YA mice after long-survival periods) or another limitation of the NestinCreER^T2 line (Line-1) used here e.g. not high enough recombination efficiency – more efficient lines (e.g. Line-2, Fig. 5) are nonetheless likely to resolve this confounder.

In summary, this study establishes unique requirements for Rb during neurogenesis in the mid-aged and old-aged SVZ-OB axis (Fig. 7), some of which are consistent with known functions of Rb during embryonic and early adult neurogenesis while others are age- but also timepoint-specific.

CRediT authorship contribution statement

Saad Omais: Conceptualization, Investigation, Formal analysis, Methodology, Visualization, Writing – original draft. Rouba N. Hilal: Investigation, Formal analysis, Visualization, Writing – review & editing. Nour N. Halaby: Investigation, Resources. Carine Jaafar: Investigation, Resources. Noël Ghanem: Conceptualization, Methodology, Supervision, Project administration, Resources, Funding acquisition, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary figure 1.

Supplementary figure 2.