Role of Hox Genes in Sub-circuit Diversification During Cortico-Ponto-Cerebellar Map Formation

Cortical input is relayed to the cerebellum mainly via precerebellar pontine nuclei. The pontine nuclei (PN), which include pontine gray and reticulotegmental nuclei (RTN), are largest of the precerebellar nuclei providing principal input to the cerebellum. PN are hypothesized to serve as an integrator of cortical information before sending these signals to the cerebellum (Schwarz and Their, 1999). The pathway from the cerebral cortex to pontine to the cerebellum is crucial for cerebellar function and learning. It is due to their critical role in the cortico-cerebellar pathway, the pontine nuclei have received much attention. Even though the projection of cortical afferents in the PN is well established (reviewed in Kratochwil et al., 2017), the molecular mechanisms underlying this complex circuitry are poorly understood. During my Ph.D., I studied the role of Hoxa5 transcription factor in the development of pontine neurons and in formation of their input-output projections.
Cortical afferents, arising in layer V of the cerebral cortex, are mapped onto the pontine nuclei in a topographic manner (Leergaard and Bjaalie, 2007). It has been proposed that cortical afferents in the PN are organized in an inside-out fashion, matching the birthdate of PN neurons (Altman and Bayer, 1987, 1997). Earliest arriving cortico-pontine fibers grow in the core of PN where the earliest born PN neurons have settled (reviewed in Kratochwil et al., 2017). A rostro-caudal organization of cortico-pontine afferents is also suggested such that afferents arising in the visual cortical area project to the anterior, while afferents arising in the somatosensory areas are mapped to the posterior pontine nuclei (Leergaard and Bjaalie, 2007). A previous study from our lab has shown that the PN neurons born from the lower rhombic lip (lRL) of rhombomere 6 (r6) settle in anterior PN while neurons born from rhombomere 8 settle in posterior PN (Di Meglio et al., 2013). As a result, PN neurons can be sub-divided in clusters based on their Hox expression pattern. Anterior PN neurons express Hox2-3 genes while Hox5 genes are expressed only in posterior PN neurons. This suggests presence of an intrinsic topographic organization in the PN based on the rostro-caudal origin or Hox expression of PN neurons. Hox genes are known to influence the topographic organization as well as input-output connectivity of several nuclei in the hindbrain and spinal cord (Bechara et al., 2016; Karmakar et al., 2017; Philippidou and Dasen, 2013). In my thesis, I therefore focused on investigating the role of Hox5 genes, expressed in posterior PN neurons, in the formation of topographic cortico-pontine circuits.
We have identified role of Hoxa5 gene in defining the position of PN neurons and orchestrating somatosensory specific input connectivity of the PN neurons. Using mouse genetics and in-utero electroporation as a tool for embryonic gene manipulation, we show that Hoxa5 overexpression leads to change in position of PN neuron towards posterior PN. This is a result of downregulation of Unc5B, a repulsive cue to Netrin, upon ectopic expression of Hoxa5 in migrating PN neurons. The positioning of PN neurons toward posterior PN enables them to receive somatosensory specific cortical input. By using trans-synaptic rabies virus tracing technique, we could also show that Hoxa5 enables PN neuron to receive or attract somatosensory input and avoid visual input from the cerebral cortex irrespective of the PN neuron position.
In this thesis, I have also investigated the molecular basis of PN connectivity with the cerebellum. While cortical inputs are mapped onto PN in a topographic manner, the projections from pontine to the cerebellum are present in a fractured map (Leergaard et al., 2006). How continuous cortical maps are transformed to fractured maps in the cerebellum remains unanswered. We hypothesize that the combinatorial expression of Hox genes in PN neurons underlies their ability to project to different parts of the cerebellum. A PN neuron is more likely to express a combination of several Hox genes as we move from the rostral to the caudal part of the PN. To address this question, we used mouse genetics to identify neurons in anterior PN neuron subsets and found that these neurons primarily project to the paraflocculus, a lobule known for its role in the visual system (reviewed in Kheradmand and Zee, 2011), while Hoxa5 positive posterior PN neurons project to several lobules of cerebellum concerned with processing of somatosensory information (Leergaard et al., 2006). Thus, the output connectivity of PN neurons also matches their input connectivity. However, we still do not understand the role of individual Hox genes in shaping the ponto-cerebellar projections. The findings presented in this thesis will serve as a basis to understand involvement of Hox genes in fracturing of the information between cortex and the cerebellum.
As a whole, this thesis highlights the role of Hoxa5 gene in orchestrating the topographic input connectivity of pontine nuclei by defining the position of pontine neurons as well as by providing cues to somatosensory cortical afferents for targeting the PN. It also provides basis to understand the role of Hox genes in ponto-cerebellar connectivity.