|Home | About | Journals | Submit | Contact Us | Français|
Rapid Conversion of Fibroblasts into Functional Forebrain GABAergic Interneurons by Direct Genetic Reprogramming.
Colasante G, Lignani G, Rubio A, Medrihan L, Yekhlef L, Sessa A, Massimino L, Giannelli SG, Sacchetti S, Caiazzo M, Leo D, Alexopoulou D, Dell'Anno MT, Ciabatti E, Orlando M, Studer M, Dahl A, Gainetdinov RR, Taverna S, Benfenati F, Broccoli V. Cell Stem Cell 2015;17:719–734 [PubMed].
Transplantation of GABAergic interneurons (INs) can provide long-term functional benefits in animal models of epilepsy and other neurological disorders. Whereas GABAergic INs can be differentiated from embryonic stem cells, alternative sources of GABAergic INs may be more tractable for disease modeling and transplantation. We identified five factors—Foxg1, Sox2, Ascl1, Dlx5, and Lhx6—that convert mouse fibroblasts into induced GABAergic INs (iGABA-INs) possessing molecular signatures of telencephalic INs. Factor overexpression activates transcriptional networks required for GABAergic fate specification. iGABA-INs display progressively maturing firing patterns comparable to cortical INs, form functional synapses, and release GABA. Importantly, iGABA-INs survive and mature upon being grafted into mouse hippocampus. Optogenetic stimulation demonstrated functional integration of grafted iGABA-INs into host circuitry, triggering inhibition of host granule neuron activity. These five factors also converted human cells into functional GABAergic INs. These properties suggest that iGABA-INs have potential for disease modeling and cell-based therapeutic approaches to neurological disorders.
Epilepsy is characterized by uncontrolled recurrent seizures and is often associated with an increase in excitatory circuits or a decrease in inhibitory circuits (1). One approach to restore inhibition is to transplant GABAergic progenitor cells derived from the medial ganglionic eminence (MGE) that mature into GABAergic inhibitory neurons; however, obtaining MGE-derived progenitor cells from human embryonic tissue remains a source of ethical concern (2–4). Another strategy is to generate GABAergic neuronal precursors for transplantation from embryonic stem (ES) cells or induced pluripotent stem cells (iPSCs) (5); then again, these methods may be costly and time consuming; thus, alternative methods are warranted. In this study, Colasante et al. establish GABAergic interneurons (INs) by direct conversion of fibroblasts and conclude their protocol is more appropriate for making “renewable, safe, and standardized” GABAergic INs. Until now, many neuronal subtypes have been established by direct conversion, albeit most with problems that have limited widespread use of the approach; however, this is the first description of a method to directly convert GABAergic INs from fibroblasts.
The authors set out to convert fibroblasts to GABAergic INs by screening transcription factors (TFs) using mouse fibroblasts from GAD-GFP mouse embryos, which allows fate monitoring in real time. First, they introduced Ascl1 (A) plus one TF at a time from pools of 21 TFs to GAD67-GFP fibro-blasts and showed that a combination of Ascl1 (A) plus Foxg1 (F) or Sox2 (S) can convert mouse fibroblasts to GAD67-GFP–positive neuronal cells with immature neuronal morphology. To obtain more mature GABAergic INs, they performed further screening by adding one factor at a time from the pool to the AFS combination and discovered GAD67-GFP positive neurons with mature morphology are induced most efficiently by the combination of transiently expressed Sox2 and Foxg1 plus constitutively expressed Ascl1, Lhx6, and Dlx5. These newly derived cells were named induced GAB-Aergic INs (iGABA-INs). Using RNA-sequencing, the authors confirmed expression of cortical GABAergic IN markers and showed that markers of midbrain, cerebellum, or hindbrain GABAergic INs were not detected in iGABA-INs. Next, the authors attempted to analyze the molecular mechanism of direct conversion to iGABA-INs. Ascl1 is already known to induce upregulation of Dlx1/2, which activates GAD65/67, a well-known marker of GABAergic INs, while ectopic expression of Ascl1 alone in fibroblasts generates not GABAergic but glutamatergic neurons (6, 7). Considering their observation that Ascl1 converted fibroblasts to GABAergic INs only in combination with either Sox2 or Foxg1, the authors hypothesized that in the course of development, the interaction between Ascl1, Sox2, and Foxg1 is indispensable for induction of GABAergic INs. The authors forced expression of Ascl1 with and without Sox2/Foxg1 shRNA in developing cortex to strengthen this point.
It is interesting that in most instances of neuronal cell reprogramming—either through iPSCs or direct conversion of fibroblasts—often a cocktail of multiple TFs is needed. Next, the authors attempted to work out the molecular circuitry by which Ascl1, Sox2, and Foxg1 promoted GABAergic neuronal fate. Considering the fact that ASCL1 binds to the I12b inter-genic enhancer region of Dlx1/2 and activates the expression of these genes (8), the authors assessed whether SOX2, FOXG1, and ASCL1 have the capacity to bind to this same region. Interestingly, in chromatin immunoprecipation and co-immunoprecipitation experiments, the authors found that SOX2, but not FOXG1, forms a complex with ASCL1 and binds to this enhancer region. This finding led them to conclude that FOXG1 does not bind to I12b but enables the SOX2-ASCL1 complex to interact with this region and induce Dlx1/2 genes, critical for differentiation into GABAergic INs.
The authors went on to confirm that the converted iGABA-INs were functional GABAergic INs by electrophysiological analysis and by transplantation into the hippocampus of immune deficient mice. The authors also showed that converted iGABA-INs could functionally integrate into synaptic circuits in vitro and in vivo by employing an optogenetic strategy. In co-culture experiments with cortical neurons, converted iGABA-INs, infected with a channelrhodopsin 2 (ChR2)-encoding lentivirus and stimulated with blue light, were able to inhibit cortical neurons, confirming functional synaptic connections between iGABA-INs and host neuronal networks. Similarly, optogenetic stimulation of transplanted iGABA-INs in adult mouse hippocampi exhibited an inhibitory effect on granule neurons, supporting their ability to integrate into existing hippocampal neuronal circuits. Finally, the authors revealed that their five-factor combination of Ascl1, Sox2, Foxg1, Lhx6, and Dlx5 was also able to convert human ES cells and fibroblasts into physiologically functional GABAergic INs.
Colasante et al. have established a novel method for directly converting somatic cells to GABAergic INs and also discovered a previously unknown pathway through which ASCL1, SOX2, and FOXG1 control GABAergic IN fate during development and direct conversion. While these studies represent an important first step regarding the clinical utility of iGABA-INs, there are still remaining issues to be addressed. The authors state that directly induced GABAergic INs are safer than those induced from ES/iPSCs and serve as more appropriate graft material for transplantation. However, Ascl1, a critical TF for direct conversion into iGABA-INs, has been shown to be involved in many tumors, such as lung cancer, thyroid cancer, and astrocytoma (9). Thus, it remains an open question whether constitutive or residual expression of Ascl1 in iGABA-INs will have tumorigenic potential. While the authors transplanted iGABA-INs into mouse brain and observed their survival for 6 weeks, they have not excluded the possibility of tumor formation in this study. Another emerging strategy is the pharmacologic induction of GABAergic INs from endogenous neural progenitors or reactive astrocytes emerging after neuronal injury, which avoids the overexpression of potential oncogenic factors or complications of transplantation surgery and might be a safer alternative to treat epilepsy. Currently, several groups are working on cell-fate conversion and successfully established iPSCs and neural stem cells by using small molecules without virally introduced TFs (10). The molecular mechanism highlighted in this research in which iGABA-INs are generated during development and direct conversion will contribute to this goal as well.
Editor's Note: Authors have a Conflict of Interest disclosure which is posted under the Supplemental Materials link.