1.1 Dorsal diencephalon conduction system: the habenula-interpeduncular circuit
Habenula is one of the highly conserved structures in vertebrate brains, and is involved in pain, emotion, motivation, light-dark cycle, and reward process (Fakhoury and López, 2014; Hikosaka, 2010a; Shelton et al., 2012). It mainly transmits information or signals from the limbic system to the interpeduncular nucleus (IPN) in the midbrain (Fakhoury and López, 2014; Hikosaka, 2010b). In mammals, the habenular complex contains medial and lateral habenula. In zebrafish, the habenular complex contains dorsal and ventral parts that show homology to the medial and lateral habenula, respectively, in mammals (Amo et al., 2010). The medial hebunula mainly conducts signal to IPN, the lateral habenula mainly conducts signal to both IPN and raphe nuclei (Bianco and Wilson, 2009; Hikosaka, 2010b; Shelton et al., 2012). In zebrafish, habenula is involved in the determination of winner or loser in a fight (Chou et al., 2016), and the motivation of fear and anxiety (Jesuthasan, 2012). Disruption of habenula causes enhanced anxiety and defect in avoidance (Lee et al., 2010).
Utilizing zebrafish as the model, several genetic regulations of habenula development have been studied. Nodal signaling plays an important role in habenular asymmetry determination at 18 to 24hpf, and left-side bias Fgf8 is expressed at 24hpf (Aizawa, 2013; Taylor et al., 2010). Nodal signaling and Fgf8 guide the parapineal migration to left side and encourage neuron stem cells in habenula to start generating neural precursors (Aizawa, 2013). Asymmetric generation of habenula seems to be due to the activity of Notch signaling that represses the generation of neuron stem cells (Aizawa, 2013; Taylor et al., 2010). The identity and location of habenula progenitors
havebula neurons nevertheless are largely unknown. The habenula starts stretching its axons to IPN at as early as 32hpf and reaches the midbrain at 57hpf (Beretta et al., 2012), and its axon encircles IPN afterwards. In addition, genes like nrp1a (Kuan et al., 2007) and daam1a (Colombo et al., 2013) were found to show left-right asymmetric expression in or near habenula and regulate habenula axon development.
1.2 Molecules that involve in habenular axonal target recognition Previous study by Kuan et al. showed that Semaphorin 3D (Sema3D), one of the Class III Semaphorins, plays a role in attracting Neuropilin 1a (Nrp1a)-positive habenular axons to their targets in the IPN (Kuan et al., 2007).
Class III Semaphorins usually play roles in axon guidance as a repulsive signal and have been well studied. Semaphorin3A (Sema3A) is the most well-studied repulsive signal in axon guidance. When a Nrp1-positive neuron faces Sema3A, Sema3A binds to Neuropilin-1 and increases local intracellular cGMP concentration (Pasterkamp, 2012;
Shelly et al., 2011). cGMP activates phosphodiesterase 4 (PDE4) through activating protein kinase G (PKG) and decreases cAMP level (Pasterkamp, 2012; Shelly et al., 2011). Lower cAMP level causes lower protein kinase A (PKA) activity and decreases the level of PKA-dependent protein phosphorylation in serine/threonine-protein kinase 11(STK11) and glycogen synthase kinase 3β (GSK3β) that regulate cytoskeletal rearrangement. (Pasterkamp, 2012; Shelly et al., 2011)
As a Class III Semaphorin, Sema3D is thought to be involved in repulsive axon guidance in the same manner as other family members usually did (Song et al., 1998).
Indeed, previous studies showed that Sema3D is mainly involved in repulsive axon guidance events that are mediated by different Neuropilins (Liu et al., 2004; Wolman et al., 2004), but during HA-IPN circuit development in zebrafish larvae Sema3D was
found to function as attractants (Kuan et al., 2007). In brief, in the zebrafish larvae, Sema3D is expressed in dorsal IPN (dIPN), and Neuropilin-1a (Nrp1a) expressed in LHA at 2dpf to 4dpf. Loss of Nrp1a shows axon guidance defect of LHA. It is due to the loss of Nrp1a-mediated axon attraction to their target. (Kuan et al., 2007). This observation suggested that Neuropilins and their downstream molecules play important roles in mediatory of attractive or repulsive axon guidance of Sema3D. However how Nrp1a mediates attractive axon guidance is completely unknown.
1.3 Sema3D-Nrp1a signaling pathway in axon attraction
Sema3D plays a role in attractive axon guidance in the formation of other neural connection system in larval zebrafish brain as well (Kuan et al., 2007; Wolman et al., 2004). Previous study shows that Nrp1a mediates axon guidance that is repulsed by Sema3D in the nucleus of the anterior commissure, on the other hand, both Nrp1a and Nrp2b mediate axon guidance that is attracted by Sema3D in the medial longitudinal fasciculus (MLF) (Wolman et al., 2004). The attractive role of Sema3D in both HA-IPN circuit and MLF development in zebrafish larvae. In order to understand the mechanism of Class III Semaphorins as an attractive axon guidance cues, a systematic approach to uncover more players that involve in Sema3D-Nrp1a-mediated attractive axon guidance is needed.
In order to find the downstream effectors that are involved in Nrp1a-mediated HA axon attraction, previous members in our laboratory started to screen for several small GTPases that have been shown to regulate cytoskeleton dynamics during axon path finding and cell proliferation. The results showed that rasd1 and rab6bb are coexpressed with nrp1a in zebrafish LHA at 4dpf (Fig.1). To know whether they are involved in LHA axon guidance, morpholinos (MOs) knocking down rasd1or rab6bb
had been performed, their morphants showed similar habenula axon projection defects that was observed in nrp1a morphants (Kuan et al., 2007).
1.4 Function and expression control of rasd1
Reports have raised some phenocopy issues between the morphants and knockout mutants (Kok et al., 2015; Rossi et al., 2015), the generation of rasd1 and rab6bb knockout lines therefore were conducted to further verify the roles of rasd1 and rab6bb during HA axonal target finding process.
Rasd1 belongs to the Ras superfamily of small GTPase and was first uncovered being over-expressed in mouse AtT-20 cells, an adrenocorticotropic hormone (ACTH) secreting cell line, with dexamethasone (DEX) treatment (Kemppainen and Behrend, 1998). Further study shows that Rasd1 is involved in glutamate-NMDA neurotoxicity in primary cortical neurons (Fig.1-A) (Chen et al., 2013). When glutamate-NMDA binds to its receptor, calcium inflow triggers neuronal nitric oxide synthase (nNOS) through calmodulin and S-nitrosylation of cysteine-11of Rasd1 which leads to the binding of Rasd1 to the peripheral benzodiazepine receptor associated protein 7 (PAP7) (Cheah et al., 2006). PAP7-Rasd1 complex interacts with the divalent metal transporter 1 (DMT1) which causes ferrous (Fe2+) import and eventually the death of the primary cortical neurons (Cheah et al., 2006; Chen et al., 2013).
Rasd1 also was found to play a role in the development of ESR1-positivec breast cancer cells. Estrogen receptor 1 (ESR1) that is overexpressed in around 70% invasive breast cancer cells is downregulated by RASD1 through upregulating microRNA-375 (miR-375) which is shown to be able to inhibit the expression of RASD1 in ESR1-positive breast cancer cells (Fig.1-B) (Simonini et al., 2010). It forms a potential positive feedback loop that enhances ESR1 expression in ESR1-positive breast cancer
cells.
Rasd1 functions as a guanylyl nucleotide exchange factor (GEF) of Giα that inhibits adenylyl cyclase (AC) (Graham et al., 2004). Rasd1 is induced by dexamethasone treatment or hypertonic stress inhibits AC activity by activating Giα, and it causes lower cAMP level in Att-20 cells (Fig.1-C) (Greenwood et al., 2016).
Lower cAMP level causes lower protein kinase A (PKA) activity, and decreases phosphorylation level of cAMP response element binding protein (CREB) that plays a role in neuronal plasticity (Greenwood et al., 2016; Kandel, 2012). These evidences suggest that Rasd1 potentially regulates neuronal plasticity via inhibiting cAMP level.
However whether Rasd1 plays a role in Nrp1a-mediated HA axonal target recognition has never been explored before.
1-5 Rab6b is Rab6bb homology and regulates retrograde transport from Golgi to ER
Rab6bb is a member of the Rab subfamilly of small GTPases that is involved in cellular vesicle formation, trafficking, and membrane fusion (Stenmark, 2009; Zerial and McBride, 2001). Rab6b is Rab6bb homology has been proposed to regulate the retrograde transport from Golgi to ER in human neuronal cells (Wanschers et al., 2007), and loss of Rab6b causes growth retardation in mouse (MGI, http://www.informatics.jax.org/). Golgi is a major membrane donor of exocytic vehicles.
Endocytosis and exocytosis play a role in axon development and mediated by Rab subfamily GTPases (Tojima and Kamiguchi, 2015). Rab6b mediates membrane dynamics at Golgi suggests that Rab6bb possibly regulates exocytosis leading to axon growth in LHA.
In this thesis project, my goal is to examine the function of Rasd1 and Rab6bb in
HA development and explore the expression regulation of rasd1 upon dexamethasone (DEX) treatment. My analyses showed that rasd1 may not play a strong role in LHA axonal target recognition process, because the axonal project volumes showed no distinguishable difference in wildtype (WT) and rasd1 mutants. Interestingly, I found that in zebrafish embryos or larvae, tissue-specific ectopic rasd1 expression can be induced by DEX treatment, whereas the rasd1 expression in the habenular neurons was unaffected. I further confirmed that one of the ectopic rasd1 expression tissues located in the hindbrain is the choroid plexus (Jiao et al., 2011). To study the role of rab6bb in HA axonal development, I have started to generate rab6bb mutant by CRISPR/Cas9 system. One F0 fish is confirmed to carry mutation in the germline cells at the targeted location. The role of rab6bb in LHA axonal targeting process will be examined once the line is established.