Research ArticleFragile X Syndrome

MDM2 inhibition rescues neurogenic and cognitive deficits in a mouse model of fragile X syndrome

See allHide authors and affiliations

Science Translational Medicine  27 Apr 2016:
Vol. 8, Issue 336, pp. 336ra61
DOI: 10.1126/scitranslmed.aad9370

MDM2 inhibitor rescues fragile X deficits

Mutation of the FMRP protein in humans leads to fragile X syndrome, the most common inherited intellectual disability. Li et al. now show that FMRP controls the activities of neural stem cells in the adult mouse brain, which is critical for production of new neurons and learning and cognition. They discovered that FMRP regulates neural stem cells through controlling the expression of the E3 ubiquitin ligase MDM2. They found that treatment with an inhibitor of MDM2 called Nutlin-3 rebalanced neural stem cell activities and rescued cognitive deficits in a mouse model of fragile X syndrome.

Abstract

Fragile X syndrome, the most common form of inherited intellectual disability, is caused by loss of the fragile X mental retardation protein (FMRP). However, the mechanism remains unclear, and effective treatment is lacking. We show that loss of FMRP leads to activation of adult mouse neural stem cells (NSCs) and a subsequent reduction in the production of neurons. We identified the ubiquitin ligase mouse double minute 2 homolog (MDM2) as a target of FMRP. FMRP regulates Mdm2 mRNA stability, and loss of FMRP resulted in elevated MDM2 mRNA and protein. Further, we found that increased MDM2 expression led to reduced P53 expression in adult mouse NSCs, leading to alterations in NSC proliferation and differentiation. Treatment with Nutlin-3, a small molecule undergoing clinical trials for treating cancer, specifically inhibited the interaction of MDM2 with P53, and rescued neurogenic and cognitive deficits in FMRP-deficient mice. Our data reveal a potential regulatory role for FMRP in the balance between adult NSC activation and quiescence, and identify a potential new treatment for fragile X syndrome.

INTRODUCTION

Fragile X syndrome affects 1 in 4000 males and 1 in 6000 females and is the most common form of inherited intellectual disability (1, 2). Fragile X syndrome arises largely because of mutations in the fragile X mental retardation (FMR1) gene, resulting in functional deficiency of fragile X mental retardation protein (FMRP), a brain-enriched RNA binding protein. Genome-wide analyses predict that FMRP regulates many mRNAs (3, 4). FMRP is highly expressed in neurons, and studies show that a loss of FMRP leads to altered expression of many neuronal genes. Therefore, a number of neurotransmitters and neuronal signaling pathways have been intensely investigated as potential therapeutic targets, including pathways associated with group 1 metabotropic glutamate receptor type 5 (mGluR5), N-methyl-d-aspartate receptor subunits, γ-aminobutyric acid type A (GABAA) receptor, mammalian target of rapamycin, tuberous sclerosis complex 2, and glycogen synthase kinase-3β (GSK3β) [reviewed by (1, 5)]. These studies led to a number of potential fragile X syndrome treatments (6). Unfortunately, disappointing results from recent clinical trials of inhibitors of mGluR5, the best known FMRP drug target, underscore a pressing need for innovation in terms of both target selection and consideration of cell types in the central nervous system beyond neurons alone (7). Despite the intense interest in fragile X syndrome, especially because of its potential as a gateway to understanding autism, the precise mechanisms behind this unique syndrome remain unclear, and there is still no U.S. Food and Drug Administration (FDA)–approved treatment. Exploring mechanisms to explain the etiology of fragile X syndrome is therefore essential if we are to develop new treatments.

Adult neurogenesis has been explored because of its potential to influence information processing associated with cognitive outcomes (8). Impaired adult neurogenesis is implicated in both neuropsychiatric disorders, such as depression and schizophrenia, and neurodegenerative diseases, such as Alzheimer’s and Huntington’s disease (9). Interventions aimed at regulating adult neurogenesis are thus being evaluated as potential therapeutic strategies. In many adult organs, both quiescent and activated (actively dividing) stem cells coexist, and the regulation of stem cell activation plays an essential role in tissue maintenance, regeneration, function, plasticity, aging, and disease (10). In the adult hippocampus, quiescent neural stem cells (NSCs) dynamically integrate both extrinsic and intrinsic signals to either maintain their dormant state or become activated and give rise to intermediate progenitor cells (IPCs), which subsequently differentiate into postmitotic dentate gyrus granule neurons or astrocytes (9). Many studies over the last 20 years have uncovered intracellular factors that regulate the activation of dentate gyrus NSCs. Although it is clear that maintaining a balance between quiescence and activation is important for the preservation of adult NSCs, whether this balance directly affects production of neurons has yet to be explored.

We have shown that FMRP deficiency in adult mouse NSCs leads to increased proliferation of NSCs but reduced differentiation into neurons and impaired hippocampus-dependent learning. Impaired hippocampus-dependent learning can be rescued by either restoring FMRP expression in adult NSCs or treating mice with a small molecule that promotes neuronal differentiation (1113). Several key questions remain unanswered. (i) Does FMRP play a role in regulating the activation of adult NSCs? (ii) Can manipulating NSC activation directly affect production of neurons? (iii) Will rebalancing adult NSC quiescence and activation rescue aberrant hippocampal neurogenesis and associated fragile X syndrome cognitive deficits?

Here, we tested the hypothesis that FMRP deficiency leads to activation of adult NSCs and that rebalancing NSC quiescence and activation could rescue cognitive deficits in a fragile X syndrome mouse model. We show that a loss of FMRP leads to greater activation of adult NSCs. We identified mouse double minute 2 homolog (MDM2), an E3 ubiquitin ligase, as a mediator of FMRP action in adult NSCs. FMRP regulates the expression of MDM2 by reducing the stability of its mRNA. In addition, the absence of FMRP also leads to elevated AKT signaling in NSCs, resulting in enhanced phosphorylation of MDM2. Thus, a lack of FMRP results in increased expression of phosphorylated (activated) MDM2 (p-MDM2) through two distinct mechanisms. p-MDM2 targets and degrades tumor suppressor protein P53. We found that specific inhibition of the MDM2-P53 interaction with a small molecule called Nutlin-3 rescued both the neurogenic and cognitive deficits in fragile X syndrome mice.

RESULTS

FMRP deficiency leads to increased activation of NSCs

To investigate the role of FMRP in regulating activation of NSCs, we crossed Fmr1 mutant [knockout (KO)] mice with mice expressing green fluorescent protein (GFP) driven by the promoter of a neural progenitor cell marker Nestin (Nestin-GFP). We thus created both Fmr1 KO (Fmr1−/y::Nestin-GFP) and wild-type (Fmr1+/y::Nestin-GFP) mice in which adult mouse NSCs and IPCs were labeled by GFP (Fig. 1A and fig. S1A). We found that the Fmr1 KO mice had an increased number (Fig. 1B) and density (Fig. 1C) of GFP+ cells in the dentate gyrus compared to their wild-type littermates but without significant changes in the overall volume of the dentate gyrus [fig. S1, C (P = 0.76, t test) and D (P = 0.47, t test); n = 5]. We next analyzed different populations of GFP+ cells in the subgranular zone using the radial glia marker glial fibrillary acidic protein (GFAP) for type 1 radial glia-like NSCs (GFP+GFAP+) and the marker TBR2 for IPCs (TBR2+ or GFP+GFAP). Fmr1 KO mice had a greater number of NSCs (Fig. 1, D and E; P < 0.01, t test; n = 5) and IPCs (Fig. 1, F to H; P < 0.0001, t test for both F and H; n = 5) compared to their wild-type littermates. We then used minichromosome maintenance complex component 2 (MCM2), a marker for cell cycle initiation and stem cell activation, to distinguish activated NSCs (GFP+GFAP+MCM2+) from quiescent NSCs (GFP+GFAP+MCM2) (Fig. 1D and fig. S1E). Fmr1 KO mice had a higher percentage of activated NSCs (Fig. 1I; P < 0.0001, t test; n = 5) and a higher number of MCM2+ cells (fig. S1F; P < 0.0001, t test; n = 5) in the adult dentate gyrus compared to their wild-type littermates. In addition, the total numbers of GFP+ cells, NSCs, and IPCs, as well as activated (MCM2+) cells were elevated in both dorsal and ventral dentate gyrus (fig. S1, G to O). As expected, both NSCs and progenitor cells exhibited greater proliferation as assessed by the cell cycle marker Ki67 in Fmr1 KO versus wild-type mice (Fig. 1, J to L; P < 0.0001 for K and P < 0.01 for L, t test; n = 5). Thus, a loss of FMRP leads to increased activation and proliferation of NSCs in the adult dentate gyrus.

Fig. 1. FMRP deficiency leads to increased NSC activation in the adult dentate gyrus.

(A) Schematic diagram showing stages of neurogenesis in the adult dentate gyrus and cell lineage–specific markers. qNSC, quiescent NSC; aNSC, activated NSC. (B and C) Quantitative comparison of the number of total GFP+ cells (B) and the density of GFP+ cells (C) in the dentate gyrus of adult Fmr1 KO mice and wild-type (WT) littermate controls. (D) Sample confocal images used in (E), (F), and (I) for identifying NSCs (GFP+GFAP+), progenitors (GFP+GFAP), and activated NSCs (GFP+GFAP+MCM2+) in the dentate gyrus of adult Fmr1 KO mice and WT mice bred onto a Nestin-GFP mouse background. Blue, DAPI (4′,6-diamidine-2-phenylindole dihydrochloride); green, GFP; red, MCM2; white, GFAP. Scale bars, 20 μm. (E and F) Quantitative comparison of the number of NSCs (E) and IPCs (F) in the mouse dentate gyrus. (G and H) Sample confocal images (G) and quantitative comparison of the number of IPCs (GFP+TBR2+) in the adult mouse dentate gyrus. (I) Comparison of the percentage of activated NSCs (GFP+GFAP+MCM2+) among the total NSCs (GFP+GFAP+) in the mouse dentate gyrus. (J) Sample confocal images used in (K) and (L) for identifying proliferating (Ki67+) NSCs (GFP+GFAP+Ki67+) and progenitors (GFP+GFAPKi67+). Green, GFP; red, Ki67; white, GFAP; white arrowheads, proliferating NSCs; white arrows, proliferating IPCs. Scale bar, 20 μm. (K) Comparison of the percentage of proliferating NSCs among total NSCs in the mouse dentate gyrus. (L) Comparison of the percentage of proliferating progenitors among total progenitors. (M) Sample confocal images from Fmr1-cKO mice used for identifying activated NSCs (tdT+GFAP+MCM2+). Green, MCM2; red, tdT; white, GFAP. Scale bar, 20 μm. (N) Quantitative comparison of the total numbers of NSCs in the dentate gyrus of cKO::Cre::Ai14 mice and Cre::Ai14 control mice. (O) Comparison of the percentage of activated NSCs among total NSCs in the dentate gyrus (n = 6 per genotype). Student’s t tests were used for all data. **P < 0.01; ***P < 0.001. From (A) to (N), n = 5 per genotype. Data are presented as means ± SEM.

To determine whether FMRP regulation of NSC activation is intrinsic to NSCs, we created a triple transgenic line (cKO::Cre::Ai14) by crossing Fmr1-floxed (cKO) mice with inducible Nestin promoter (Nes)–driven Cre transgenic mice (Nes-CreERT2) and Rosa26-STOP-tdTomato (Ai14) reporter mice as described previously (fig. S2A) (12). We injected adult triple transgenic mice (cKO::Cre::Ai14) and control littermates (Cre::Ai14) with tamoxifen to achieve targeted deletion of FMRP specifically in adult NSCs and their subsequent progenies and, 4 weeks later, analyzed the steady-state level of NSC activation by assessing the number of tdTomato-positive (tdT+) cells that both expressed GFAP and met the morphological criteria of NSCs. We found that the cKO::Cre::Ai14 mice had more NSCs (tdT+GFAP+) (Fig. 1, M and N, and fig. S2B) and elevated NSC activation compared to Cre::Ai14 littermates (Fig. 1O; P < 0.0001, t test; n = 6). Again, the overall volume of the dentate gyrus was unchanged (fig. S2C). Our data suggest that FMRP regulates NSC activation through an intrinsic mechanism within NSCs.

FMRP regulates MDM2 expression in adult neural stem/progenitor cells

FMRP is known to bind to a subset of specific mRNAs in the brain and regulate their translation (4, 14). Given that our goal was to discover FMRP targets that regulate NSC activation, we used published FMRP PAR-CLIP (photoactivatable ribonucleoside-enhanced crosslinking and immunoprecipitation) data on proliferating cells [human embryonic kidney (HEK) 293 cells] (3) to select candidate FMRP targets. Given that the number of FMRP-associated reads precipitated for each mRNA reflects the likelihood of interaction, we chose an arbitrary threshold of 1000 reads, yielding 417 prospective FMRP-regulated mRNAs (Fig. 2A and table S1). Meanwhile, we searched the Mammalian Adult Neurogenesis Gene Ontology (MANGO) database (15) and literature to curate a set of 16 “neurogenic regulators” with well-established roles in adult NSC quiescence and activation (Fig. 2A and table S2). Among them, only MDM2 also appeared as one of the 417 FMRP “candidate targets.” Because FMRP may influence neurogenesis by acting either directly on these 16 neurogenic regulators or on related proteins participating in their functional pathways, we used the Biological General Repository for Interaction Datasets (BioGRID) 3.4 protein interaction database and GeneMANIA to identify 322 proteins that have literature-supported physical interactions with the 16 neurogenic regulators, including some of the core neurogenic regulators themselves (table S3). Comparing these 335 “neurogenic interactors” to the 417 FMRP candidate targets revealed an intersection of 29 putative FMRP targets known to participate in neurogenic pathways. Candidates were ranked by the number of pathways in which they participated, and MDM2 ranked first with four interactions supported by 25 publications (table S4). The known interaction between MDM2 and its target P53 is an established mechanism for regulating NSC activation (16). Because MDM2 was among our core neurogenic regulators and exhibited the highest degree of interaction with strongly supported neurogenic pathways, we chose to focus on MDM2 as the mediator of FMRP-regulated NSC activation.

Fig. 2. FMRP regulates MDM2 expression in adult mouse NPCs.

(A) Workflow for selecting potential FMRP target genes that may regulate activation of adult NSCs. (B to E) FMRP RNA immunoprecipitation (IP) followed by quantitative real-time PCR analyses for Mdm2 (B), GSK3β (C), Map1b (D), and Gapdh (E, negative control) mRNAs in Fmr1 WT and KO mouse NPCs (n = 3). The amount of each mRNA in immunoprecipitated samples was normalized to the amount of mRNA in the input samples. (F and G) Quantification analyses of Mdm2 mRNA (F) and MDM2 protein (G) in WT and Fmr1 KO mouse NPCs (n = 3). Gapdh was used as the internal control for quantitative PCR analysis, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control for Western blot analysis. (H) Mdm2 mRNA stability in Fmr1 WT and KO mouse NPCs treated with actinomycin D to inhibit transcription. The percentage of Mdm2 mRNA remaining in NPCs was quantified using real-time PCR. Comparisons of the different decay rates were performed by two-way analysis of variance (ANOVA) (n = 3). Half-life of decay was calculated after log2, transformation. Except for Fig. 2H, Student’s t tests were used for data analyses. Data are presented as means ± SEM. *P < 0.05; **P < 0.01. n.s., no significant difference.

Next, we used primary neural stem/progenitor cells (NPCs) isolated from the adult dentate gyrus as a model system to interrogate the mechanism behind FMRP’s regulation of NSC activation. We confirmed that Fmr1 KO NPCs indeed exhibited increased proliferation and astroglial production but reduced neuronal differentiation (fig. S3), as described previously (1113). To determine whether FMRP regulates MDM2, we performed RNA binding protein immunoprecipitation and identification of coprecipitated RNA (RNA-IP) with an FMRP antibody coupled with real-time polymerase chain reaction (PCR) analysis of Mdm2 mRNA. We confirmed that FMRP was associated with Mdm2 mRNA in NPCs (Fig. 2B). We also detected the association of FMRP with Gsk3β and Map1b mRNAs, two known targets of FMRP (11, 17), but not with Gapdh mRNA (18) (Fig. 2, C to E), validating the specificity of our assay.

We next found that both Mdm2 mRNA and MDM2 protein expression were higher in Fmr1 KO NPCs compared to wild-type NPCs (Fig. 2, F and G, and figs. S4 and S5A). The higher Mdm2 mRNA expression could be due to either increased gene transcription or increased mRNA stability. Because FMRP is known to associate with target mRNAs and modulate mRNA stability (19), we evaluated Mdm2 mRNA decay kinetics over an 8-hour period in Fmr1 KO and wild-type NPCs in the presence of an inhibitor of transcription, actinomycin D. We found that Mdm2 mRNA exhibited fast decay kinetics, with most (>90%) RNA disappearing within 4 hours in both wild-type and KO NSCs (fig. S6A). We then repeated this experiment by analyzing mRNA over a 120-min period. We found that Mdm2 mRNA had a longer half-life in Fmr1 KO (T1/2 = 58.7 to 61.3 min) versus wild-type (T1/2 = 44.4 to 45.0 min) NPCs (Fig. 2H and fig. S6, B and C), whereas the decay of control Gapdh and Ywhaz mRNA exhibited no difference between genotypes (fig. S6, D and E). Therefore, FMRP controls MDM2 expression in NPCs at least in part by reducing the half-life of Mdm2 mRNA.

FMRP regulates MDM2 phosphorylation through the AKT pathway

The elevated MDM2 protein expression in Fmr1 KO NPCs prompted us to assess p-MDM2, the active form of MDM2 (Fig. 3A). MDM2 has multiple phosphorylation sites. Phosphorylation on SER395 and SER407 is known to regulate cell death (20), whereas phosphorylation of MDM2 on serine SER166 and SER186 is related to cell growth (21). Because we found no altered cell death or survival in Fmr1 KO NPCs compared to wild-type NPCs (fig. S7), we assessed the phosphorylation of Ser166, which is known to be simultaneously phosphorylated with Ser186 by phosphorylated AKT (p-AKT) in response to cell growth signals (22). We found that Fmr1 KO NPCs had more p-MDM2 (Fig. 3, B and C; P = 0.0015, t test; n = 3) and a higher p-MDM2 to total MDM2 ratio (Fig. 3D) compared to wild-type NPCs, suggesting that increased p-MDM2 in Fmr1 KO NPCs may result from both increased total MDM2 protein expression and elevated phosphorylation. Activated AKT (AKT phosphorylated at Ser473 or Thr308) is known to phosphorylate MDM2 at Ser166 and Ser186 (22), and FMRP deficiency results in increased AKT phosphorylation at Ser473 in neurons (23). Elevated phosphatidylinositol 3-kinase/AKT signaling in adult NSCs leads to increased NSC activation (24). We found that Fmr1 KO NPCs had a higher ratio of p-AKT to total AKT than wild-type NPCs (Fig. 3, E and F; P = 0.013, t test; n = 3). Thus, Fmr1 KO NPCs have increased p-MDM2 due to both increased total MDM2 expression and elevated phosphorylation, possibly by p-AKT.

Fig. 3. FMRP regulates MDM2 phosphorylation through the AKT pathway.

(A) Schematic drawing showing that, in addition to inhibiting Mdm2 mRNA expression, FMRP can also inhibit MDM2 phosphorylation at Ser166/Ser186 through AKT signaling. MK-2206 is a selective AKT inhibitor. (B to D) Western blot analyses (B) of total MDM2 (C) and phosphorylated MDM2 at Ser166 (p-MDM2) (D) in Fmr1 WT and KO mouse NPCs (n = 3). (E and F) Western blot analyses of total and phosphorylated AKT at Ser473 (p-AKT) (n = 3). (G and H) Western blot analyses of P53 showing that P53 was decreased in Fmr1 KO mouse NPCs compared with WT NPCs (n = 3). GAPDH was used as a loading control. (I to M) Representative Western blot image (I) and quantitative analyses of p-AKT (J), p-MDM2 (K), total MDM2 (L), and P53 (M) in Fmr1 KO and WT mouse NPCs treated with MK-2206 or vehicle. Total AKT was used as a loading control for p-AKT, and GAPDH was used as a loading control for p-MDM2, MDM2, and P53 (n = 3). *P < 0.05; **P < 0.01; ***P < 0.001; Student’s t test was used in (C) to (H); one-way ANOVA was used in (J) to (M). Data are presented as means ± SEM.

MDM2 is phosphorylated on Ser166 in response to growth factor and mitogenic signaling, leading to increased MDM2-mediated ubiquitylation and degradation of P53 (25, 26). P53 is known to repress NSC activation (27), and we found that Fmr1 KO NPCs had reduced expression of P53 (Fig. 3, G and H; P < 0.0001, t test; n = 3). Treatment with MK-2206, a specific inhibitor of AKT phosphorylation (28) (Fig. 3A), led to reduced p-AKT (Fig. 3, I and J) and reduced p-MDM2 (Fig. 3, I and K), without affecting total MDM2 (Fig. 3, I and L) in both KO and wild-type NPCs. Also, MK-2206 treatment enhanced P53 expression in Fmr1 KO NPCs (Fig. 3, I and M). Thus, in addition to controlling MDM2 expression through mRNA stability, FMRP also restricted the amount of p-MDM2 through the AKT pathway. The absence of FMRP led to increased p-MDM2 and subsequently reduced P53 in NPCs.

The amount of MDM2 and p-MDM2 directly affects the proliferation and differentiation of NPCs

To investigate the effect of elevated p-MDM2 on the proliferation and differentiation of NPCs, we acutely knocked down MDM2 expression in NPCs using lentivirus expressing short hairpin RNA (shRNA) against Mdm2 (shMdm2; Fig. 4A and fig. S8A). Lenti-shMdm2–treated NPCs had decreased total amounts of MDM2 and p-MDM2 (Fig. 4, B to E; P < 0.01 for C, P < 0.05 for D, and P < 0.05 for E, one-way ANOVA; n = 3) and increased P53 (Fig. 4, F and G; P < 0.01, one-way ANOVA; n = 3) compared to lenti-shNC (nonsilencing control shRNA)–treated NPCs. We next investigated the impact of MDM2 knockdown on NPC proliferation as assessed by BrdU (5-bromo-2′-deoxyuridine) pulse-labeling and differentiation as assessed by cell lineage–specific antibodies, β-tubulin III (Tuj1) for neurons and GFAP for astrocytes. Acute knockdown of MDM2 (shMdm2) led to reduced cell proliferation in both wild-type and Fmr1 KO NPCs (Fig. 4, H and I), increased neuronal differentiation (Fig. 4, J and K), and reduced astroglial differentiation (fig. S8, B and C) in both wild-type and Fmr1 KO NPCs. The shMdm2 treatment restored the proliferation and differentiation of Fmr1 KO NPCs to that of wild-type (lenti-shNC–treated) NPCs (Fig. 4, I and K). Thus, reducing MDM2 in NPCs rescued proliferation and differentiation deficits of FMRP-deficient NPCs.

Fig. 4. MDM2 and p-MDM2 expression directly affect the proliferation and differentiation of mouse NPCs.

(A) Schematic model showing that both acute knockdown of MDM2 and treatment with MK-2206 may repress MDM2, relieve P53 repression, and rescue proliferation and differentiation of Fmr1 KO mouse NPCs. (B to E) Western blot analyses (B) of MDM2 (C) and p-MDM2 (D) in NPCs with acute knockdown of MDM2 using shMdm2 (n = 3, normalized to GAPDH or total MDM2). (F and G) Western blot analyses (F) of P53 expression in NPCs with acute knockdown of MDM2 by shMdm2 (G) (n = 3). GAPDH was used as a loading control. (H) Representative images showing that both Fmr1 WT and KO mouse NPCs incorporated BrdU (red) under proliferating conditions, with or without shMdm2 treatment (green). Scale bar, 20 μm. (I) Quantitative analysis showing that acute knockdown of MDM2 by shMdm2 reduced the proliferation rate of both Fmr1 WT and KO mouse NPCs (n = 3). (J and K) Acute knockdown of MDM2 by shMdm2 rescued neuronal differentiation phenotypes of Fmr1 KO NPCs, as assessed using the neuronal marker Tuj1+ (J) (red for Tuj1; green for shMdm2 or shNC viral-infected cells); quantification is in (K) (n = 3). (L) Both WT and KO mouse NPCs incorporated BrdU (red) under proliferating conditions. Scale bar, 20 μm. (M) MK-2206 treatment reduced the proliferation rate of both Fmr1 WT and KO NPCs (n = 3). (N and O) MK-2206 treatment rescued neuronal differentiation phenotypes of Fmr1 KO NPCs (n = 3). One-way ANOVA was used for all data analyses. *P < 0.05; **P < 0.01; ***P < 0.001. Data are presented as means ± SEM.

We next investigated whether reducing MDM2 phosphorylation alone could also rescue the proliferation and differentiation of Fmr1 KO NPCs. We found that MK-2206 treatment resulted in decreased cell proliferation (Fig. 4, L and M) and astroglial differentiation (fig. S8, D and E) but increased neuronal differentiation (Fig. 4, N and O) in both wild-type and Fmr1 KO NPCs. MK-2206–treated Fmr1 KO NPCs exhibited similar proliferation and differentiation as vehicle-treated wild-type NPCs. Therefore, both reducing the expression of MDM2 and blocking MDM2 phosphorylation can rescue the proliferation and differentiation deficits of Fmr1 KO NPCs.

Nutlin-3 treatment rescues the proliferation and differentiation of NPCs

Although MK-2206 could correct the deficits of Fmr1 KO NPCs, its effects on wild-type NPCs raised concerns over potentially broad effects resulting from AKT pathway inhibition, with obvious potential therapeutic complications. We therefore searched for a specific inhibitor of MDM2 and found Nutlin-3, a potent MDM2 inhibitor that specifically interrupts MDM2 and P53 interactions (29, 30). Nutlin-3 has been evaluated extensively for its therapeutic potential and mechanism of action in human cancer and is currently in a phase 1 clinical trial for the treatment of retinoblastoma (31). We found that Nutlin-3 treatment did not affect p-MDM2 levels in either Fmr1 KO or wild-type cells (Fig. 5, A to C) but led to increased P53 in KO cells with no significant effect on wild-type cells (Fig. 5, D and E). Next, we found that Nutlin-3 treatment reduced the proliferation and differentiation of Fmr1 KO NPCs without a significant effect on wild-type NPCs (Fig. 5, F to K). Nutlin-3 also did not exhibit significant effects on cell death and survival of either KO or wild-type NPCs (fig. S7, D to F). Thus, Nutlin-3, a specific inhibitor of the MDM2-P53 interaction, can rescue the proliferation and differentiation of Fmr1 KO NPCs without affecting wild-type NPCs.

Fig. 5. Nutlin-3 treatment rescues the proliferation and differentiation of NPCs in vitro.

(A) Schematic showing that Nutlin-3 may inhibit the interaction between p-MDM2(Ser166/Ser186) and P53, relieving the repression of P53 and rescuing the proliferation and neuronal differentiation of NPCs. (B to E) Western blot analyses of p-MDM2 (B and C) and P53 (D and E) in Fmr1 WT and KO NPCs treated with Nutlin-3 (Nt3), showing that Nutlin-3 had no significant effect on p-MDM2 expression in Fmr1 KO and WT NPCs but specifically rescued P53 expression in Fmr1 KO NPCs without affecting WT cells (n = 3). GAPDH was used as a loading control in Western blot analyses. (F and G) Nutlin-3 treatment rescued the cell proliferation phenotype of Fmr1 KO NPCs, as demonstrated by immunostaining cells using the cell proliferation marker BrdU (F, red; scale bar, 20 μm), followed by quantitative analysis of BrdU+ cells (G) (n = 3). (H and I) Nutlin-3 treatment rescued neuronal differentiation phenotypes of Fmr1 KO NPCs, as assessed by a neuronal marker Tuj1+ (red; scale bar, 20 μm) (H); quantitative analysis in (I) (n = 3). (J and K) Nutlin-3 treatment specifically rescued astroglial differentiation phenotypes of Fmr1 KO NPCs, as assessed using the astroglial marker GFAP+ (green; scale bar, 20 μm) (J); quantitative analysis in (K) (n = 3). **P < 0.01; ***P < 0.001. One-way ANOVA was used for all data analyses. Data are presented as means ± SEM.

Nutlin-3 treatment rescues neurogenic deficits in Fmr1 KO mice

We next investigated whether Nutlin-3 could rescue the neurogenic deficits in Fmr1 KO mice. We found that in Fmr1 KO mice, the percentage of p-MDM2+ cells among GFP+ cells was significantly higher than that in Fmr1 wild-type mice (Fig. 6, A to C; P = 0.003, t test; n = 3). In addition, we found that in the brain of a female Fmr1+/−::Nestin-GFP mouse in which half of the cells expressed FMRP, whereas the other half of the cells did not because of X chromosome inactivation, FMRP and p-MDM2 expression exhibited an inverse relationship (fig. S9). Nutlin-3 has been used in mice at high dosages (50 to 200 mg/kg) for 2 weeks or longer to examine its ability to block cancer cell growth in vivo (3235). Because our goal was to correct rather than totally abolish NSC activation in the adult brain, we chose a dosage of 10 mg/kg, which is the lowest dosage with known effects on in vivo cell proliferation (36). We also chose a shortened injection period based on the literature (37) and our experience (13). We treated Fmr1 KO and littermate wild-type mice with either vehicle or Nutlin-3 (10 mg/kg) every other day for five injections over 9 days. We found that Nutlin-3 treatment specifically reduced NSC activation in Fmr1 KO mice to wild-type levels (P < 0.0001, one-way ANOVA; n = 3), with no significant effect on wild-type mice (n = 3, one-way ANOVA) (Fig. 6, E and F). Because FMRP deficiency impairs adult neurogenesis (11, 12), we next determined whether rebalancing NSC activation via Nutilin-3 treatment could rescue neuronal production (Fig. 6G). Mice received both Nutlin-3 treatment and BrdU injections to label dividing neural progenitors and were analyzed 4 weeks after the last Nutlin-3 treatment. Nutlin-3 treatment rescued both neuronal and astroglial differentiation in Fmr1 KO mice and brought them to levels similar to those of wild-type (vehicle-treated) mice, without any significant effect on wild-type mice (Fig. 6, H to K). These results suggest that Nutlin-3 treatment can correct NSC activation and rescue adult neurogenesis deficits in Fmr1 KO mice.

Fig. 6. Nutlin-3 treatment can rescue neurogenesis deficits in Fmr1 KO mice.

(A to C) Sample confocal images of GFP+p-MDM2+ (A) and GFP+p-MDM2 (B) cells in the adult mouse dentate gyrus, and quantitative comparison of the percentage of p-MDM2+ cells among GFP+ cells in Fmr1 KO and WT mice. Quantitation in (C) (n = 3). **P < 0.01, Student’s t test. Blue, DAPI; green, GFP; red, p-MDM2 (Ser166). Scale bars, 20 μm. (D) Experimental scheme for assessing NSC activation in Fmr1 WT and KO mice treated with Nutlin-3 or vehicle. (E) Sample confocal images of activated NSCs (GFP+GFAP+MCM2+) in the dentate gyrus of adult Fmr1 WT/KO-Nestin-GFP mice. Blue, DAPI; green, GFP; red, MCM2. Scale bar, 20 μm. (F) Comparison of the percentage of activated NSCs among total NSCs in the dentate gyrus of Fmr1 KO mice and WT mice with or without Nutlin-3 treatment (n = 3 per group). (G) Experimental scheme for assessing NSC differentiation in Fmr1 WT and KO mice treated with Nutlin-3 or vehicle. (H and I) Nutlin-3 treatment rescued neuronal differentiation in Fmr1 KO mice, as assessed by immunostaining with the neuronal marker NeuN (green) and BrdU (red) (H). Scale bar, 20 μm. Quantification of percentage of neurons among BrdU+ cells in (I) (n = 3 to 5 per group). (J and K) Nutlin-3 treatment rescued astroglial differentiation specifically in Fmr1 KO mice, as assessed by the astroglial marker S100β (green) and BrdU (red) (J). Scale bar, 20 μm. Quantification of the percentages of astrocytes among BrdU+ cells in (K) (n = 3 to 5 per group). **P < 0.01; ****P < 0.0001, one-way ANOVA. Data are presented as means ± SEM. The boxes with dotted white lines in (A), (B), (E), (H), and (J) indicate regions where higher magnification images have been provided.

Nutlin-3 treatment rescues cognitive deficits in Fmr1 KO mice

We next investigated whether Nutlin-3 could reverse the behavioral deficits of Fmr1 KO mice. We showed previously that Fmr1 KO mice exhibited deficits in two hippocampus-dependent learning tasks: the delayed nonmatch to place radial arm maze (RAM) and trace conditioning (12, 13). However, these tests require stressful food restriction or electric shock, so we decided to use two cognitive tests with minimal stress on the mice: the novel object recognition test and the novel location test. First, using Fmr1 cKO::Cre::Ai14 triple transgenic mice (fig. S2), we confirmed that mice with selective deletion of FMRP from adult NSCs and their progenies exhibited deficits in the novel location test (fig. S10). This result corroborated our previous findings using the RAM test (12, 13) and validated the use of the novel location test for assessing spatial memory of FMRP-deficient mice. We then treated Fmr1 KO mice and their wild-type littermates with either vehicle or Nutlin-3 and analyzed their cognition 4 weeks later (Fig. 7A). Nutlin-3 had no significant effect on overall health and body weight (fig. S11, A and B) or locomotor activity and anxiety (fig. S11, C and D). Consistent with previous studies (18, 38, 39), Fmr1 KO mice exhibited both impaired spatial learning on the novel location test (Fig. 7, B and C) and defective learning on the novel object recognition test (Fig. 7, D and E) compared to wild-type mice. Administration of Nutlin-3 rescued performance on both the novel location test (Fig. 7C; P < 0.01, one-way ANOVA; n = 9 to 13 per group) and the novel object recognition test (Fig. 7E; P < 0.01, one-way ANOVA; n = 9 to 13 per group) in Fmr1 KO mice, with no significant effect on wild-type mice. Therefore, Nutlin-3 treatment could rescue cognitive deficits of Fmr1 KO mice on these two tests. We found that, similar to Fmr1 KO mice, postmortem hippocampal and cortical tissues from patients with fragile X syndrome also exhibited higher MDM2 mRNA expression and p-MDM2 protein levels compared to those from age- and gender-matched healthy controls (fig. S12).

Fig. 7. Nutlin-3 treatment rescues cognitive functions in Fmr1 KO mice.

(A) Experimental scheme for assessing cognitive functions in Fmr1 WT and KO mice treated with Nutlin-3. (B) Schematic of novel location test for assessing spatial learning. (C) Nutlin-3 treatment fully rescued spatial memory deficits in Fmr1 KO mice in the novel location test (n = 9 to 13 mice per group). (D) Schematic of the novel object recognition test. (E) Nutlin-3 treatment fully rescued deficits in the novel object recognition test in Fmr1 KO mice (n = 8 to 14 mice per group). **P < 0.01; ***P < 0.001. One-way ANOVA was used for all data analyses. Data are presented as means ± SEM.

DISCUSSION

Here, we show that FMRP plays an important role in adult NSC activation through MDM2 regulation (fig. S13). We further demonstrate that specific inhibition of the interaction of MDM2 with P53 using a small molecule undergoing clinical trials for cancer rescued both the neurogenic and cognitive deficits in a mouse model of fragile X syndrome. Our findings show that manipulating NSC activation can affect the production of mature neurons contributing to cognitive function. Because most adult brain cells are postmitotic, targeting a cell proliferation pathway in NSCs provides specificity for therapeutic application.

Despite great efforts, there is no FDA-approved treatment for fragile X syndrome. Because FMRP is highly expressed in neurons, extensive research has naturally focused on neurons and has identified a number of neurotransmitter receptors and signaling pathways that are either targets or regulators of FMRP (2, 5). These discoveries led to the exciting development of potential drugs for fragile X syndrome. For example, lithium attenuates GSK3β activity, which reduces group 1 mGluR-dependent activation of protein translation (40). GSK3β inhibitors have been shown to improve cognition in a mouse model of fragile X syndrome (13, 41). GABAA receptor agonists reduce anxiety and audiogenic seizures in Fmr1 KO mice (42). Several mGluR5 antagonists, including fenobam, RG7090 (Roche), AFQ056 (Novartis), CTEP, MPEP, and STX107 (Seaside Therapeutics), reverse multiple phenotypes in mouse models of fragile X syndrome (6). However, human clinical trials have yielded disappointing results. Although earlier open-label studies (for example, fenobam) showed promising effects, recent double-blind clinical trials for inhibitors of mGluR5 (RG7090, AFQ056, and STX107), the best known fragile X drug target, yielded inconclusive results [summarized in (7)], highlighting the importance of searching for other cellular mechanisms and pathways. The contributions to fragile X syndrome of other cell types in the brain, such as glia and immune cells, are being actively pursued (1, 43). We have focused on the involvement of adult neurogenesis in fragile X syndrome. Although NSCs constitute a small number of cells in the adult brain, FMRP deficiency in adult-born new neurons directly affects cognition, and restoration of FMRP in adult new neurons rescues these deficits (fig. S8B) (12). We further showed that inhibition of GSK3β, an FMRP target, rescues adult neurogenesis and learning deficits (13). These data suggest that targeting adult neurogenesis might be an effective treatment for certain cognition deficits in adult fragile X syndrome. However, GSK3β is involved in diverse biological processes in many types of brain cells, making drug specificity a major concern. A critical question whether treatment aimed at targeting adult NSCs could be an effective therapy for fragile X syndrome has been raised. Here, we discovered that administration of a small molecule designed for cancer treatment, but used at a much lower dosage, could rebalance NSC activation and rescue the cognitive deficits in a mouse model of fragile X syndrome with no apparent effect on wild-type mice. Adult NSC activation is highly regulated; studies over the past two decades have identified a number of intrinsic regulators within NSCs and extrinsic factors provided by the stem cell niche (8, 9). Most studies have focused on the role of NSC activation in the maintenance of NSCs, whereas the impact of NSC activation on mature neuron production has gone unexplored. We found that rebalancing NSC activation during the initial neurogenic phase could change terminal differentiation of NSCs into neurons and astrocytes (analyzed 30 days later). Therefore, correcting the imbalance between NSC activation and quiescence during the early stage of neurogenesis could have long-lasting effects on cell fate. Because most adult brain cells are postmitotic, targeting a cell proliferation pathway may restrict drug action largely to NSCs without significantly affecting postmitotic cells, yielding the desired specificity for therapy. There are, however, several limitations to our study. Complex social and language deficits in fragile X syndrome are ambiguously modeled in mice, and the effect of Nutlin-3 on these behaviors has not been evaluated. Although the Fmr1 KO mice used in our study recapitulate several key deficits in human fragile X syndrome, the neurocognitive impairment in mouse models of this disease is much milder than those seen in human patients, making it difficult to predict the therapeutic efficacy of Nutlin-3 treatment in preclinical trials.

MDM2 is an E3 ubiquitin ligase with many biological functions, from wound healing and carcinogenesis to tissue regeneration. MDM2 is activated through phosphorylation at multiple sites, and the sites of phosphorylation are linked to distinct functions (44). Phosphorylation of MDM2 at Tyr394, Ser395, and Ser407 by ataxia telangiectasia mutated (ATM)/ATM and Rad3-related kinase (ATR) is stimulated by DNA damage, leading to inhibition of MDM2, which results in apoptosis. We found no significant difference in cell death between Fmr1 wild-type and KO NPCs. Therefore, we decided that MDM2 phosphorylation by ATM/ATR may have no significant role in Fmr1 KO NPCs. On the other hand, growth factor and mitogenic signaling trigger AKT phosphorylation of MDM2 at Ser166 and Ser186, leading to MDM2 activation and subsequent P53 ubiquitylation and degradation (25, 44). AKT activity is elevated in FMRP-deficient mouse brains and Drosophila neuroblasts (4547), as well as in lymphoblastoid cells derived from patients with fragile X syndrome (46). Here, we found a dysregulated AKT-MDM2-P53 pathway in Fmr1 KO NPCs, and that inhibiting either AKT or the interaction of MDM2 with P53 restored P53 expression and rescued NPC proliferation and differentiation deficits in Fmr1 KO NPCs. AKT signaling is frequently dysregulated in human cancers, and MK-2206, a potent allosteric inhibitor of all AKT isoforms, has been used to treat patients with tumors in several clinical trials (48, 49). Although our study might extend the reach of MK-2206 to fragile X syndrome, one concern is the molecule’s broad effects on many cellular processes in many types of cells. This concern is substantiated by our own observation that, although MK-2206 is highly effective in rescuing Fmr1 KO NPCs, it also has potent effects on NPCs in wild-type mice.

The MDM2-P53 interaction is not only necessary for early embryonic development (50, 51) but is also crucial for stem cell maintenance in adults (20). Particularly relevant to the present study, MDM2 is known to inhibit P53 activity and regulate adult NSC activation (16). Besides its well-described proapoptotic effect, P53 is implicated in nervous system development (52), and P53 null mice show deficits in learning and memory and behavioral alterations (53). We selected Nutlin-3 because it is a potent inhibitor specifically designed to inhibit MDM2-P53 interactions (29), has been evaluated extensively for its therapeutic potential and mechanism of action in human cancer (54), and is currently in phase 1 clinical trials for treating retinoblastoma (31). To limit the effect of Nutlin-3 on proliferating cells in mice, we chose a significantly lower dosage (10× lower) and shorter treatment scheme compared to those used for cancer treatment; our mice showed no obvious changes in health and behavior. We also observed that, unlike with MK-2206, Nutlin-3 treatment exhibited no effect on the NPCs of wild-type mice. Although both Fmr1 KO mice and patients with fragile X syndrome exhibited elevated p-MDM2 in hippocampal and cortical brain tissues, similar to what we have seen in NSCs, P53 expression was unchanged in Fmr1 KO mouse cortex and hippocampal tissues (p53 expression was too low to be reliably assessed in human tissues) (fig. S14). Therefore, it is likely that P53 is not a major target for MDM2 in mature neurons, which comprise a large proportion of cells in the brain tissue. It is not surprising that we found very low levels of P53 in cortex and hippocampus, especially in humans, because it has been shown that P53 expression is low in mature neurons and that elevated P53 in response to stress can lead to neuronal death (55). Therefore, our discovery of the widespread up-regulation of p-MDM2 in fragile X syndrome human and mouse tissues with no apparent impact on P53 expression in neurons suggested that targeting p-MDM2 might be a promising new therapeutic method for treating fragile X syndrome. MDM2 has other substrates (56), such as PSD95 in neurons (57). Comprehensive identification of p-MDM2 targets in neurons and specific inhibitors for p-MDM2’s interaction with these targets will be a fruitful area for translational research in fragile X syndrome. We do not rule out that in treated mice, Nutlin-3 may inhibit p-MDM2’s interaction with targets other than P53 and may exert effects on other types of cells besides NSCs, including mature neurons. However, Nutlin-3 was designed specifically to inhibit MDM2-P53 interactions, and much higher dosages would be needed before there will be an effect on other MDM2 targets (56). Because we used a low dose of Nutlin-3 in our study, the effects we observed are most likely to be through P53 inhibition in NSCs compared to other targets or in mature cell types. Cell type–specific targeting of Nutlin-3 action will be required to validate the direct link between restoration of NSC activity and cognitive rescue; however, such methods are not yet available. The high specificity of Nutlin-3 toward FMRP-deficient NPCs without apparent effect on wild-type NPCs, along with the fact that most brain cells are postmitotic and, therefore, likely to be less sensitive to Nutlin-3, makes it an attractive potential candidate for treating fragile X syndrome.

MATERIALS AND METHODS

Study design

The purpose of this study was to investigate the role of FMRP in adult NSC activation and whether rebalancing NSC quiescence and activation could rescue both neurogenesis and cognitive deficits in a mouse model of fragile X syndrome. On the basis of our publications and power analysis, at least three to six biological replicates were used for each in vitro or in vivo biochemical and histological analysis, whereas a sample size of 9 to 21 per group was used for behavioral testing. The NPCs used for in vitro analyses were isolated independently from different Fmr1 KO and wild-type littermates and were biological replicates. Data collection was carried out for a predetermined period of time, as dictated by literature- or core facility–based standards, and no exclusion criteria were applied. For drug treatment, animals were randomly assigned to treatment arms with approximately equivalent numbers in each group. All cell counting and behavioral analyses were performed by experimenters who were blind to the identity of the samples. All analyses were performed by scientists blinded to genotype and/or treatment arm.

Animal studies

All animal procedures were performed according to protocols approved by the University of Wisconsin-Madison Care and Use Committee. All mice were on C57B/L6 genetic background. The Fmr1 KO/Nestin-GFP mice (Fmr1−/y::Nestin-GFP) were created by crossing female Fmr1 heterozygous KO mice (Fmr1+/−) (58) with homozygous Nestin-GFP transgenic males (59). Generation of FMRP inducible conditional mutant mice (Fmr1loxP/y::Nestin-CreERT2::Rosa26-tdT) and tamoxifen injections to induce recombination were performed as described (12). To induce recombination, mice (6 weeks old) received tamoxifen (160 mg/kg; Sigma-Aldrich) daily for 5 days as described (12). Nutlin-3 (10 mg/kg) was given to 7- to 8-week-old mice through intraperitoneal injections every other day for five injections and sacrificed at either 24 hours or 4 weeks after the last injection. For NSC differentiation analysis, mice also received four BrdU injections (100 mg/kg) between days 2 and 3 of Nutlin-3 injections.

Tissue preparation and immunohistochemistry

Brain tissue processing and histological analysis of mouse brains were performed as described in our publications (11, 12, 60). Briefly mice were euthanized by intraperitoneal injection of sodium pentobarbital and then transcardially perfused with saline followed by 4% paraformaldehyde (PFA). Brains were dissected out, postfixed overnight in 4% PFA, and equilibrated in 30% sucrose. Brain sections (40 μm) were generated using a sliding microtome and stored as floating sections in cryoprotectant solution [glycerol, ethylene glycol, and 0.1 M phosphate buffer (pH 7.4) (1:1:2 by volume)]. We performed immunohistological analysis on 1-in-6 serial floating brain sections (240 μm apart). After staining with primary and fluorescent secondary antibodies (see Supplementary Materials and Methods), sections were counterstained with DAPI (1:2000; Roche Applied Science) and then mounted, coverslipped, and maintained at 4°C in the dark until analysis.

In vivo cell quantification

The total numbers of GFP+ or tdT+ cells in the dentate gyrus of each animal were counted on a Zeiss ApoTome microscope using unbiased stereology (Stereo Investigator software, MBF Biosciences Inc.). A 5-μm guard zone and exhaustive counting method were used (61). Dentate gyrus volume was measured using Stereo Investigator as described (11, 12). Dorsal (bregma, −0.94 to −2.30 mm) and ventral (bregma, −2.30 to −3.80 mm) hippocampus regions were defined on the basis of literature (62). Phenotypic analysis was performed as described (12). Briefly, at least 50 GFP+ or BrdU+ cells in the dentate gyrus were randomly selected, and their colocalization with cell lineage markers was determined using a Nikon A1 confocal microscope and quantified using ImageJ software [National Institutes of Health (NIH)].

Bioinformatics analysis

To select candidate FMRP targets, we used FMRP CLIP data from Ascano et al. (3). mRNA transcripts were ranked by the number of reads precipitated in the experiment, and an arbitrary threshold of 1000 reads was chosen to determine the most likely targets of FMRP. This cutoff yielded 417 top genes or about 2% of all probes in the experiment. Sixteen genes with known roles in adult NSC activation or proliferation were selected from the MANGO (http://mango.adult-neurogenesis.de) database and literature and were named neurogenic regulators. For each of these neurogenic regulators, all known physical interactions were mapped by BioGRID 3.4 (http://thebiogrid.org) and GeneMANIA (http://www.genemania.org), producing 16 networks of 335 unique proteins that interact with neurogenic regulators (neurogenic interactors). Additionally, the number of articles supporting each interaction as listed by BioGRID 3.4 was used as an indicator of the strength of a given interaction. Overlap was then assessed between candidate FMRP targets and both neurogenic regulators and neurogenic interactors to find proteins likely to be regulated by FMRP that influence adult NSC activation.

Adult NPC analyses

NPCs were isolated from the dentate gyrus of 8- to 10-week-old male Fmr1 KO mice and wild-type littermate controls based on our published methods (12, 63). Independently isolated cells serve as biological replicates. Proliferation and differentiation of NPCs were analyzed as described (11, 12, 60). We used only early passage cells (between passages 4 and 10) and only the same passage numbers of wild-type and Fmr1 KO cells. For each experiment, triplicate wells of cells were analyzed, and results were averaged as one data point (n = 1). At least three independent biological replicates were used (n = 3) for statistical analyses.

Chemical treatment

For proliferation assay in vitro, 1 μM MK-2206 (Selleckchem, S1078) (AKT phosphorylation inhibitor) or 2 μM Nutlin-3 (Selleckchem, S1061) (Mdm2 inhibitor) was added to proliferating NPCs at 6 hours after plating. Cells were then pulse-labeled 18 hours later with 5 μM BrdU (Sigma-Aldrich, B5002) and incubated for 6 hours in the presence of inhibitor, followed by fixation, staining, and quantification. For differentiation assay in vitro, 1 μM MK-2206 and 2 μM Nutlin-3 were added to NPCs upon initiation of differentiation with retinoic acid/forskolin for 4 days, followed by fixation, staining, and stereological quantification.

RNA immunoprecipitation

RNA-IP was performed as described (11, 60). Briefly, wild-type and Fmr1 KO NPCs were harvested and homogenized in 1 ml of ice-cold lysis buffer [10 mM Hepes (pH 7.4), 200 mM NaCl, 30 mM EDTA, and 0.5% Triton X-100] with 2× complete protease inhibitors (Boehringer-Mannheim). Nuclei and debris were pelleted at 3000g for 10 min; the supernatant was collected and raised to 300 mM NaCl and clarified at 14,000g for 30 min. The resulting supernatant was precleared for 1 hour with 100 μl of recombinant protein G–agarose (Invitrogen) (prewashed with lysis buffer). An aliquot of precleared input was saved for RNA extraction (200 μl) and protein analysis (100 μl). A monoclonal antibody against FMRP (7G1-1, Developmental Studies Hybridoma Bank) or a monoclonal antibody against immunoglobulin G (IgG) (5415S, Cell Signaling) was incubated with recombinant protein G dynabeads at 4°C for 2 hours and washed three times with lysis buffer. Ribonuclease inhibitors (Roche) were added to the remaining lysates. The precleared lysates were immunoprecipitated with antibody-coated recombinant protein G–agarose at 4°C for 2 hours. After three washes with lysis buffer, 10% of immunoprecipitate was saved for protein analysis. The remainder was washed one more time, and the immunoprecipitate was resuspended into TRIzol (Invitrogen) for RNA isolation.

Real-time PCR assay

Real-time PCR was performed using standard methods as described (60). The first-strand complementary DNA (cDNA) was generated by reverse transcription with oligo(dT) primer (Roche). To quantify the mRNA expression using real-time PCR, aliquots of first-stranded cDNA were amplified with gene-specific primers and Power SYBR Green PCR Master Mix (Bio-Rad) using a StepOne Real-Time PCR System (Applied Biosystems). The PCR reactions contained 1 μg of cDNA (except the cDNA for immunoprecipitation, for which 5% of the cDNA was used for each gene examined), Universal Master Mix (Applied Biosystems), and 10 μM of forward and reverse primers in a final reaction volume of 20 μl. The mRNA expression of different samples was calculated by the data analysis software built in with the 7300 Real-Time PCR System. For RNA immunoprecipitation/real-time PCR, cDNA from immunoprecipitation and input was used, and immunoprecipitated samples were normalized to input samples. The sequences of primers are provided in Supplementary Materials and Methods.

Mdm2 mRNA stability assay

Dentate gyrus NPCs from wild-type and Fmr1 KO mice were grown in proliferating condition (see above). Actinomycin D (10 μg/ml; Sigma-Aldrich) was added on the basis of published protocol (60), and total RNA was isolated at various time intervals. The Mdm2 mRNA expression was normalized to Gapdh or Ywhaz mRNA as measured by real-time PCR. RNA decay kinetics and half-life were analyzed using published methods (61). Briefly, we used the exponential function M = M0e−ƛt (M: amount of mRNA at t time; M0: amount of mRNA at t time). ƛ = (ln)/T1/2 (T1/2 is the half-life of the mRNA).

Western blotting analyses

Protein samples were separated on SDS–polyacrylamide gel electrophoresis gels (Bio-Rad), transferred to polyvinylidene difluoride membranes (Millipore), and incubated with primary antibodies (see Supplementary Materials and Methods). After incubation with fluorescence-labeled secondary antibodies (LI-COR), the membranes were imaged using LI-COR, and quantification was performed using ImageJ software. At least n = 3 independent blots were used for statistical analysis.

Human tissues

Human tissue was obtained from the University of Maryland Brain and Tissue Bank, a Brain and Tissue Repository of the NIH NeuroBioBank. Frozen hippocampal and cortical tissues from two age-, gender-, and postmortem time–matched patients with fragile X syndrome (80 and 85 years old) and two control male Caucasian individuals (82 and 84 years old) were used for MDM2 mRNA and p-MDM2 (S166) analysis.

Novel location test

This test measures spatial memory through an evaluation of the ability of mice to recognize the new location of a familiar object with respect to spatial cues. The experimental procedure was performed as previously described (64). The mice were exposed to two identical objects for 6 min during the acquisition phase three times and tested at 3 min later. Object preference was evaluated during this session. During the trial session, one of the objects was moved to a novel location, the mice were allowed to explore the objects for 6 min, and the total time spent exploring each object was measured. Exploration was considered as any investigative behavior (that is, head orientation and sniffing occurring within <1.0 cm) or deliberate contact occurring with each object. The discrimination index was calculated as the percentage of time spent investigating the object in the new location minus the percentage of time spent investigating the object in the old location: discrimination index = (novel location exploration time/total exploration time × 100) − (old location exploration time/total exploration time × 100). All experiments were videotaped and scored by scientists who were blinded to experimental conditions.

Novel object recognition test

This test is based on the natural propensity of rodents to preferentially explore novel objects over familiar ones. The experimental procedure was performed as previously described (38). On the first day, mice were habituated for 10 min to the maze in which the task was performed. On the second day, mice were put back in the maze for 10 min, and two identical objects were presented. Twenty-four hours later, one of the familiar objects was replaced with a novel object, the mice were again placed in the maze and were allowed to explore for 10 min, and the total time spent exploring each of the two objects (novel and familiar) was measured. The discrimination index was calculated as the difference between the percentages of time spent investigating the novel object and the time spent investigating the familiar objects: discrimination index = (novel object exploration time/total exploration time × 100) – (familiar object exploration time/total exploration time × 100). All experiments were videotaped and scored by scientists who were blinded to experimental conditions.

Statistical analysis

All experiments were randomized and blinded to scientists who performed quantification. Statistical analysis was performed using ANOVA and Student’s t test, unless specified, with the GraphPad software. Two-tailed and unpaired t test was used to compare two conditions. One-way ANOVA was used for comparison among multiple experimental conditions. Bonferroni post hoc test was used when comparing among each condition. For Mdm2 mRNA stability analysis, two-way ANOVA was used for comparison of the different decay rates. All data were shown as means ± SEM. Probabilities of P < 0.05 were considered as significant.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/8/336/336ra61/DC1

Materials and Methods

Fig. S1. FMRP deficiency does not affect dentate gyrus volume but leads to increased number of activated cells in the adult dentate gyrus and increased numbers of GFP+ cells, NSCs, and IPCs in both dorsal and ventral dentate gyrus.

Fig. S2. Conditional deletion of FMRP leads to increased NSC numbers in the adult dentate gyrus without affecting the overall dentate gyrus volume.

Fig. S3. Adult NPCs derived from Fmr1 KO mice exhibit increased proliferation, reduced neuronal differentiation, but increased astroglial differentiation.

Fig. S4. Assessment of optimal protein loading amount for protein quantification using Western blotting.

Fig. S5. Quantitative tables for p-MDM2 assessment using Western blot analysis.

Fig. S6. Mdm2 mRNA stability analysis in wild-type and Fmr1 KO NPCs.

Fig. S7. Fmr1 KO NPCs do not exhibit increased cell death.

Fig. S8. Both acute knockdown of MDM2 and treatment with MK-2206 rescue astroglial differentiation of Fmr1 KO NPCs.

Fig. S9 Inverse correlation of FMRP and p-MDM2 expression in Fmr1+/−::Nestin-GFP female brains.

Fig. S10. Selective deletion of FMRP from Nestin-positive NSCs and their progenies resulted in cognitive deficits, without affecting locomotor activities.

Fig. S11. Nutlin-3 treatment does not affect general health and activities of mice.

Fig. S12. Increased MDM2/Mdm2 mRNA and p-MDM2 protein expression in the hippocampal and cortical tissues from Fmr1 KO mice and postmortem brains of fragile X syndrome patients.

Fig. S13. Models for FMRP regulation of the MDM2 and P53 pathway, which affects adult neurogenesis and cognition.

Fig. S14. P53 protein levels did not show significant changes in the hippocampal and cortical tissues from Fmr1 KO mice.

Table S1. FMRP candidate targets (provided as a separate Excel file).

Table S2. Neurogenic regulators that regulate adult NSC quiescence and activation.

Table S3. All known physical interactions with proteins known to be important in NSC activation/proliferation (provided as a separate Excel file).

Table S4. Number of physical interactions between FMRP candidate targets identified by PAR-CLIP-seq (table S1) and neurogenic regulators with role in NSC activation obtained from MANGO database and literature search (table S2).

References (6581)

REFERENCES AND NOTES

Acknowledgments: We thank C. T. Strauss for editing; Y. Xing, E. Berndt, E. M. Jobe, and N. E. Patzlaff for technical assistance; and the Zhao laboratory members for discussion. We thank J. Cottrell, R. Johnson, and the National Institute of Child Health and Human Development Brain and Tissue Bank at the University of Maryland for providing human postmortem tissues. We thank A. Ozaita (Universitat Pompeu Fabra) for helping us optimize the novel object recognition test. Funding: Supported by grants from the NIH (R01MH080434, R01MH078972, and R21NS095632 to X.Z.) and a Center grant from the NIH to the Waisman Center (P30HD03352). Author contributions: Y.L. and X.Z. conceived the concept, designed the experiments, analyzed data, and wrote the manuscript. Y.L. performed all in vitro and in vivo experiments. M.E.S. performed behavioral and cell fate counting. I.B. and J.L.M. performed cell fate counting. B.E.E. performed bioinformatics analysis. Y.G. created the lentiviral vectors. A.B. participated in human tissue analysis. Competing interests: X.Z. and Y.L. are co-inventors on a U.S. provisional patent related to this work: “Methods for treating cognitive deficits associated with fragile X syndrome,” no. 62/222267. The other authors declare that they have no competing interests. Data and materials availability: All materials generated are available for distribution.
View Abstract

Navigate This Article