Skip to main content
European Commission logo print header

Dissecting the gene regulatory mechanisms that generate serotonergic neurons

Final Report Summary - SEROTONIN (Dissecting the gene regulatory mechanisms that generate serotonergic neurons.)


Serotonin (5HT) is one of the main neurotransmitters in the brain. In mammals, serotonergic neurons are a relatively small population of around 100.000 cells confined in the hindbrain. Despite its reduced number, they project to almost every area in the brain and to the spinal cord and they regulate a plethora of behaviors. Moreover, serotonergic dysfunction has been proposed to underlie several pathological conditions such as depression, bipolar disorder, schizophrenia, anorexia, etc.
5HT neurotransmission is very well conserved through evolution being present in all animal groups. Importantly, the enzymes and transporters used for the synthesis, release and re-uptake of 5HT (5HTergic pathway genes) are also phylogenetically conserved (Figure 1a). Little is known about the genetic program that regulates 5HT differentiation in any organism.
The main aim of this project was to take advantage of the phylogenetic conservation of 5HTergic systems and use the simple model organisms C.elegans to identify the transcription factors (TFs) required for the coordinated activation of the 5HTergic pathway genes (Aim1). Moreover, we wanted to test if the nematode 5HTergic program is phylogenetically conserved in mammals, using mouse models and embryonic stem cells (mESC) as our experimental model (Aim2).
At the end of the project, as will be detailed below, we have accomplished almost all sub-objectives of Aim1, and we have also expanded some of them. Regarding Aim2, part of the objectives are almost finished and have also been expanded, while objectives regarding mouse embryonic stem cells (mESC) are developing more slowly and might need some changes.

The main research achievements in this period are the following:
C.elegans nervous system is composed by 302 neurons, the serotonergic system consists of three different neuronal classes organized as three pairs of neurons: the NSM , the ADF and the HSN neurons (Figure 1b). Although all 5HT neuron subtypes express the 5HT pathway genes, the rest of their transcriptomes and their functions are very different (NSM is a neurosecretory neuron, ADF is a chemosensory neuron and HSN is a motor neuron). In the first set of experiments we explored if 5HT pathway genes expression in the NSM, ADF and HSN was regulated through the same cis regulatory elements or to the contrary, was regulated in a subtype specific manner. Our results show a modular organization in which, each neuron subtype activates independent cis-regulatory modules to start the expression of the same gene (Figure 1c). These results, that cover the originally proposed Aim1b, are in contrast to our previous findings on C.elegans dopaminergic system differentiation, as dopamine pathway gene expression is globally regulated in all dopamine neuron subtypes by the same cis-regulatory modules (Flames & Hobert, Nature 2009, DOI: 10.1038/nature07929; Doitsidou*, Flames* et al Genes and Dev. 2013, DOI: 10.1101/gad.217224.113) (Figure 2).
This apparently contradictory findings are explained by two reasons: 1) Contrary to the 5HTergic system, dopaminergic neuron subtypes in C.elegans have very similar transcriptomes and functions, 2) According to our results and results from other laboratories, it seems that the transcriptome of each neuron subtype is globally co-regulated by a limited combination of transcription factors (terminal selectors) that act directly on the regulatory regions of most of the genes composing that transcriptome (Figure 3). Thus, as all dopaminergic neurons subtypes share transcriptomes they also share their terminal selector code, while 5HT neuron subtypes have diverse transcriptomes, and thus are regulated by independent terminal selector codes that recognize different cis regulatory modules in each 5HT subtype.
Once determined that the regulatory logic of each 5HT neuron subtype was different, we aimed to determine the terminal selector codes responsible for the activation of the 5HT pathway genes in each 5HT neuron subclass. In collaboration with my postdoctoral supervisor, Dr Oliver Hobert, at Columbia University we identified a combination of two transcription factors, the Lim homeodomain TF TTX-3 and the POU homeodomain UNC-86, which are required for the correct differentiation of the NSM neuron. Two lab members, Miren Maicas and Carla Lloret contributed to this work that was published in a high impact journal (Zhang et al. Development 2013, doi 10.1242/dev.099721) (Figure 4).
As stated on Aim 1a, one of our aims was the analysis of HSN terminal differentiation program. Combining cis-regulatory analysis together with mutant analysis we have been able to identify a combination of seven different TFs that act directly on the regulatory modules of the 5HT pathway genes to regulate its coordinated expression in the HSN. As predicted by the terminal selector model, these TFs are required for the expression of most of the HSN transcriptome (Figure 4). These results show a complex regulation in which at least seven different TFs, from different TF families, are required to regulate the expression of the terminal features. To our knowledge there is little precedent of such detailed description of transcriptional regulation of a specific neuronal cell fate.
Moreover, to our surprise, mammalian homologs for five out of the seven C.elegans TFs were already known to be required for mouse 5HT differentiation (Figure 4). Thus, we hypothesized that mammalian homologs of UNC-86 (POU TF) and SEM-4 (Spalt TF) could also be required for correct 5HT differentiation. We have identified that Brn2 (mouse POU TF) and Sall2 (mouse Spalt TF), are expressed in the raphe 5HT neurons. We are currently analyzing Brn2 and Sall2 mutant mice 5ht system to see if these animals show any differentiation defect compared to controls (covering our initial Aim2a)
Finally, we were puzzled by the fact that it seems that there is deep homology between mouse raphe 5HT neurons and HSN C.elegans neurons. Why precisely HSN 5HT neurons and not NSM or ADF neurons? We have performed a bioinformatics analysis comparing the raphe 5HT transcriptome with the available data of the transcriptome of all 302 C.elegans neuron subtypes. Strikingly, from all C.elegans neuron HSN expression profile is the closest to raphe transcriptome. This finding shows an unprecedented degree of conservation of the genetic program regulating the differentiation of both populations and suggests they could have arisen from a unique ancestral cell type.

We are finishing the Brn2 and Sall2 mutant mice characterization to send the manuscript for publication (Lloret et al. MS in preparation). We think there are several important findings in this paper, including modular organization of serotonergic differentiation in C.elegans deep homology between HSN and Raphe neurons and possible identification of new mammalian factors required for 5HT specification.
We have also started to study ADF regulatory logic, we have identified ETS, RFX and CSL binding sites required for expression of 5HT pathway genes in this cell type. We are now analyzing different mutant strains to identify the TFs responsible for the regulation of ADF differentiation.
Finally, regarding dopaminergic differentiation, the initial results were performed by me and by Maria Doitsidou as postodoctoral fellows at Columbia University, however additional experiments were performed and are being performed in my laboratory at the Instituto de Biomedicina de Valencia. Part of the dopamine results where already published in a high impact journal, with me as Maria Doitsidou as first co-authors (Doitsidou, Flames et al. Genes and Dev 2013, doi 10.1101/gad.217224.113). In addition, we are currently working the phylogenetic conservation of dopaminergic fate and plan on sending a manuscript for publication by the end of 2015 (Remesal et al. MS in preparation).

In summary work supported from the CIG grant has resulted in two publications so far and we plan to send two other manuscripts in the next few months, in which I am senior author and corresponding (Figure 4).
Finally, the Marie Curie CIG grant has helped me in the transition from a postdoctoral fellow to an independent investigator. I started my own group four years ago with a five-year contract, currently I have a permanent position at the Instituto de Biomedicina de Valencia and my group, the developmental neurobiology unit, is composed by eleven members (Five graduate students, four postdocs and two technicians) all working in different aspects of neuronal differentiation both in worms and mice.

Updated information about our lab and our project can be found in our webpage:
http://www3.ibv.csic.es/index.php/es/investigacion/patologia/und