Modeling Fragile X syndrome in neurogenesis: An unexpected phenotype and a novel tool for future therapies

ABSTRACT

FMRP is an RNA-binding protein involved in synaptic translation. Its absence causes a form of intellectual disability, the Fragile X syndrome (FXS). Small neuroanatomical abnormalities, present both in human and mouse FMRP-deficient brains, suggest a subtle critical role of this protein in neurogenesis. Stable depletion of FMRP has been obtained in a mouse embryonic stem cell line Fmr1 (shFmr1 ES) that does not display morphological alterations, but an abnormal expression of a subset of genes mainly involved in neuronal differentiation and maturation. Inducing the differentiation of shFmr1 ES cells into the neuronal lineage results in an accelerated generation of neural progenitors and neurons during the first steps of neurogenesis. This transient phenotype is due to an elevated level of the Amyloid Precursor Protein (APP), whose mRNA is a target of FMRP. APP is processed by the BACE-1 enzyme, producing the β-amyloid (Aβ) peptide accelerating neurogenesis by activating the expression of Ascll. Inhibition of the BACE-1 enzyme rescues the phenotype of shFmr1 ES cells.

Here we discuss the importance of the shFmr1 ES line not only to understand the physiopathology of FXS but also as a tool to screen biomolecules for new FXS therapies.

Silencing of the Fragile X Mental Retardation gene (FMR1) causes the Fragile X Syndrome (FXS), the most common form of inherited intellectual disability (ID). FMR1 encodes the Fragile X mental retardation protein (FMRP), an RNA-binding protein involved in different steps of RNA metabolism, such as translational control, RNA transport along neurites and RNA export from the nucleus to the cytoplasm.¹ All FXS patients are affected by cognitive impairment and they may display attention deficit-hyperactivity disorder (ADHD), autistic behavior, seizures, anxiety and language delay.² Examination of brains from FXS patients has shown an increased density of long and tortuous dendritic spines. This abnormality is considered the cellular alteration underpinning FXS ID.³ The Fmr1 null mouse exhibits a phenotype with similarities to humans including abnormal dendrite morphology.^4,5 In mice, it has been possible to associate the altered dendritic spine morphology to some abnormal forms of synaptic plasticity (e.g., increased hippocampal LTD, reduced cortical LTP and epileptogenesis).^6,7 The functional defects of these neuronal structures characterizing Fmr1-null neurons have been linked to the role of FMRP in the regulation of translation of a subset of synaptic proteins (1,7; see below). In addition, children with FXS (aged between 18 and 42 months) have larger brain volumes and display enlargement in the temporal lobe white matter, cerebellar gray matter and caudate nucleus, but have a smaller amygdala.⁸ In another study analyzing boys aged 1 to 3 y the authors found, as in the adult patients, increased caudate, fusiform gyrus, and thalamus gray matter volume (GMV) as well as reduced superior temporal gyrus, hippocampus, and orbitofrontal cortex GMV. Specifically, a reduced GMV in the hypothalamus, insula, and medial and lateral prefrontal cortices was also described.⁹ In C57BL6/J Fmr1-null mice, differences in brain volumes were only observed in two deep cerebellar nuclei (fastigial nucleus and nucleus interpositus).¹⁰ On the other hand, FVB Fmr1 knockout mice display different brain phenotypes compared with the same model in the C57B6/J background. In fact, in FVB Fmr1 knockout mice significantly larger relative volume differences were found in major white matter structures throughout the brain. Moreover, a smaller striatum and a larger parieto-temporal lobe volume were observed.¹¹ These neuroanatomical abnormalities are likely to be generated early during development and may be associated to defects in proliferation and/or differentiation of neural progenitors, suggesting a critical role of FMRP in neurogenesis.¹²

Recent studies on the Down Syndrome, another form of ID, showed the possibility to treat young adults,¹³ but it was also underlined the importance to start treatments as early as possible. Consequently, some preclinical therapeutic approaches are targeting not only neonatal but also prenatal life¹⁴ focusing on those molecules that can act on neurogenesis defects.^15,16 Starting from these considerations and from the fact that an effective specific therapy for FXS is not available yet, a mouse embryonic stem cell line displaying a reduced expression of Fmr1 by stable transfection of a specific shRNA directed against Fmr1 (shFmr1 ES) has been generated.¹⁷ These cells do not display morphological abnormalities and cell cycle variations, however altered expression of a subset of genes mainly involved in neuronal differentiation and maturation determines a subjacent molecular pathology. Indeed, stimulating the differentiation of shFmr1 ES cells into the neuronal lineage results in an accelerated generation of neural progenitors and neurons during the first steps of differentiation. This phenotype is transient, as the final number of neurons is not affected at late phases of in vitro neurogenesis. Interestingly, neurogenesis is also accelerated in the embryonic brains of Fmr1 KO mice, indicating that the shFmr1 ES cell model recapitulates the molecular and cellular alterations present in vivo.¹⁷ This phenotype in shFmr1 ES cells is likely due to an elevated level of the Amyloid Precursor Protein (APP), whose mRNA is a known target of FMRP.¹⁸ APP is processed by the BACE-1 enzyme, producing the β-amyloid (Aβ) peptide that is known to accelerate neurogenesis by activating the expression of achaete-scute family bHLH transcription factor 1 (Ascll).^19,20 It is interesting to point out that the increased level of Aβ peptide in Fmr1-depleted ES cells induces the expression of Ascll.¹⁷ This latter factor has a pivotal role in neuronal differentiation²¹ and its induction in shFmr1 ES cells represents a surprising event and the key point to explain the subsequent altered neuronal differentiation.¹⁷ Consistently, the cell phenotype is rescued not only by re-expressing human FMR1, but also by reducing the processing of APP by the specific BACE-1 inhibitor LY2811376. The importance of the Aβ peptide in the physiopathology of FXS, as well as in other forms of autism and ID, has been extensively studied.²²

ShFmr1 ES cells also present altered expression of other genes that could explain the FXS pathology at the molecular level. Indeed, a reduced expression of Tropomyosin Receptor Kinase B (TrkB) is present. Interestingly, treatment of adult Fmr1 knockout mice with 7, 8-dihydroxyflavone (7, 8-DHF), an agonist of TrkB, improves their spatial and fear memory.²³ Furthermore, expression of the small GTPase RhoA is reduced in shFmr1 ES cells. The mRNA encoding RhoA was already shown to be a target of FMRP²⁴ and Rho GTPases and actin remodelling have been already described as having a critical role in the physiopathology of FXS.^1,24

Collectively, these data underline the fact that the absence of FMRP modulates the expression of proteins and their related pathways spanning the earliest steps of embryonic life to adult. Depletion of FMRP alters the normal kinetics of neuronal differentiation.¹⁷ We can speculate that this event uncoordinates different brain maturation pathways and programs, leading to subtle architectural abnormalities of several brain regions and, ultimately, to intellectual deficit.

The phenotype of shFmr1 neural progenitors appears surprising since cell models of neural precursors for genes involved in other forms of ID/autism rather display a delay of neuronal differentiation or a disruption of neurogenesis.^25-27 Indeed, premature neurogenesis has been associated to gross brain abnormalities in α thalassemia/mental retardation syndrome X-linked (ATRX), consistently with the microcephaly observed in patients affected by this disorder.^28,29 Similarly, accelerated cell cycle and overproduction of GABAergic inhibitory neurons were described in iPSC-derived brain organoids of Autism Spectrum Disorder (ASD) patients characterized by macrocephalia. Molecularly, this abnormal neurogenesis was due to the overexpression of the transcription factor FOXG1.³⁰ However, it is worth noticing that depletion of Phosphatase and TENsin homolog (PTEN) in postnatal/young neural stem cells produced an altered neurogenesis characterized in a first step by an increased proliferation and differentiation rate of these cells.³¹ However, an early loss of Neural Stem/progenitor Cells (NSCs) was observed in these animals after some months. Interestingly, these mice displayed macrocephaly and, similarly to FXS mice, an impaired social interaction and an activation of AKT, S6 and GSK3b in hippocampal neurons.³⁰ Also in this case, it is possible to describe an altered kinetics of neurogenesis even if, due to the severity of the cellular alterations, the morphological brain abnormality appears more evident than in FXS brains. These considerations underline the importance to study embryonic neurogenesis in ID/autism animal models to decipher the physiopathology of these disorders and to identify helpful biomarkers for translational studies.

As it has been carefully detailed,¹⁷ the genetic heterogeneity of human FXS iPS and FXS human embryonic stem cells^17,32-35 complicates the use of these cell lines to study FXS. For this reason, the shFmr1 cell model can be considered as a very useful tool to study neurogenesis in the absence of FMRP.¹⁷ In addition, these cells can be very useful to search for novel therapies for FXS. Indeed, they can be used for screening of bioactive molecules, including libraries of small molecules that are already approved for clinical use. This screening can be feasible considering that, as shown above, the phenotype of the FXS cell model can be quantified and is reversible by pharmacological tools. Thus, the molecules that will reveal to be able to actively revert the phenotype of this model could be useful during all developmental ages of patients, since, as we have discussed, some FMRP-dependent pathways are conserved throughout life (Fig. 1). Collectively, these findings will contribute to improve the understanding of the molecular pathology of FXS and to provide a better stratification of FXS patients, which is as a weak aspect in the characterization of this syndrome in view of personalized therapies. New drugs identified by this approach will also highlight new pathways involved in the physiopathology of FXS and specific biomarkers for this syndrome.

Figure 1.

A schema of the different steps of research we plan to carry out to define a treatment starting from the correction of the abnormal morphology of shFmr1 neural progenitors. First, starting from shFmr1 neural progenitors, a high throughput screening will be performed to identify those small molecules (of a selection of already approved drugs) that are able to revert the shFMR1 phenotype. These molecules will be tested in the Fmr1-null mice, the murine model of FXS, to define their ability to rescue some behavioral phenotypes of this mouse at different ages. In particular the impact of these molecules on social interaction and cognition deficits, and in hyperactivity of FXS mice will be evaluated. Molecules able to correct these phenotypes will be directly tested on FXS patients in Phase II clinical trials.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors are grateful to Franck Aguila for graphical work.

Funding

This study was supported by INSERM, CNRS, CNRS LIA NEOGENEX, BB and EL are supported by Agence Nationale de la Recherche: ANR-11-LABX-0028–01. BB is supported by ANR-SVSE4–2012, FRM Team 2014, Monaco Against Autism Foundation, Fondation Jérôme Lejeune.