The vertebrate brain has evolved, in large part, to integrate and process multiple types of sensory information (“input”), and to produce and execute a response, or coordinated program for action (“output”). The thalamus develops from the embryonic brain region called the diencephalon, which also gives rise to the retina of the eye, and to the hypothalamus, epithalamus (pineal body), and subthalamic nucleus. The thalamus is a key relay center required to connect different brain structures, such as the cerebral cortex, basal nuclei, cerebellum, and brainstem nuclei, in functional networks that allow an animal to respond to its changing environment. Although applying computing systems analogies to the mammalian brain is an oversimplification, it might be useful to think of the thalamus as a router, which forwards information from the thousands of brain CPUs required for complex parallel processing. Undoubtedly, the thalamus has its own CPUs, but a major function of this brain structure is to receive information from sensory pathways (pain, touch, position sense, vision, hearing, taste) and from independent supercomputers like the cerebellum and basal nuclei, and to pass this information to the cerebral cortex in a very organized manner. Of course, the orderly connections between neurons in the thalamus and neurons in the cerebral cortex (thalamocortical projections) must be established during development, and this embryonic wiring process is dependent on accurate pathfinding by the processes (axons) that neurons send out from their “homes” in the various nuclei and cortical layers of the brain.
Figure 1: Topographic organization of thalamocortical projections. a) in mouse, ventrolateral thalamic nucleus projects to primary motor cortex (M1); ventrobasal nucleus (VB) projects to primary somatosensory cortex (S1); lateral geniculate nucleus (LGN) projects to visual cortex (V1) b) axons from the lateral VB nucleus project to medial S1 cortex (barrel field) d) same topographic organization in human cerebral cortex (from Vanderhaeghen and Polleux, 2004)
This is Anatomy Week here at Guadalupe Storm-Petrel, and the general question addressed by Powell and colleagues is a neuroanatomical one: what cues do thalamic axons use to navigate to the correct regions of the cerebral cortex (e.g. visual, auditory, motor, somatosensory)? Not only must the axons of thalamic neurons project to the correct cortical areas, but they must also maintain spatial relationships, or topography. This precise topographic organization of axons, which exists throughout the central nervous system, allows the brain to determine precisely where in the body sensory information came from, and where motor information must be sent. Dufour and colleagues (2003) reported that ephrins and their Eph receptors act as guidance cues for thalamocortical axons, as they undergo topographic sorting in an intermediate target, the ventral telencephalon. In mice that lack the adhesion molecule, close homolog of L1 (CHL1), axons from the VB nucleus of the thalamus are misrouted from the somatosensory cortex (their normal target), to the visual cortex (Wright et al., 2007). In the PLoS Biology article to be discussed in detail here, Powell and colleagues (doi:10.1371/journal.pbio.0060016) add three more molecules to the list of those involved in the patterning of thalamocortical projections: Netrin-1, DCC (deleted in colorectal cancer), and Unc5. As summarized in Figure 2, Netrin-1 attracts axons from the rostral thalamus, and repels those emerging from the caudal thalamus.
Figure 2. Diagrams of horizontal brain sections A. Previous work from Dufour et al. (2003) showed that Ephrin-A5 and Eph receptors are involved in patterning thalamocortical connections. B. and C. Netrin-1 attracts some thalamic axons, and repels others (from Powell et al., 2008 )
To perform the studies, Powell and colleagues used a “whole-mount telencephalic assay”, which allowed them to label thalamic neurons, trace their axons growing through the ventral telencephalon, and map projections to the cerebral cortex, in an accessible organ culture preparation that maintains anatomical relationships. In particular, the researchers used the landmark of the corticostriatal boundary (CSB), between ventral and dorsal telencephalon, to monitor the topographic organization of emerging thalamic axons, and to establish that this pattern exists well before axons reach the cortex. Low molecular weight biotinylated dextran amine (BDA) was injected into the thalamus, to fill thalamocortical axons rapidly and anterogradely, such that developing connections could be mapped and counted. As thalamic axons grow to the cerebral cortex, they form a structure called the internal capsule, and as a final descriptive step, Powell and colleagues used in situ hybridization and a gene trap mouse line to show that Netrin-1 is expressed in a gradient across this region of the ventral telencephalon: high levels at the rostral (nose) end, and low levels at the caudal (tail) end. The remaining experiments address the roles of Netrin-1, and its attractive (DCC) and repulsive (Unc5) receptors, in establishing the topographic organization of thalamic axons as they project to the cortex.
Figure 3. Co-culture of wild-type rostral thalamus with wild-type (B) or Netrin-1-/- ventral telencephalon (C). In these photos, the nose end (R) of the brain is to the right, and the lateral edge of the brain slice is toward the top of the figure. Note how the DTR axons wander caudally (towards the tail) in C.
One of the advantages of working with mouse embryos is the availability of lines with targeted (“knockout”) mutations in identified genes. Although normal numbers of thalamocortical axons are present in the brains of Netrin-1-/- mouse embryos, axons emerging from the rostral part of the thalamus project more caudally into the ventral telencephalon, than they do in the brain of a wild-type embryo. Moreover, Figure 3 shows that when wild-type rostral thalamus (DTR) is combined with Netrin-1-/- ventral telencephalon (VTel), many of the axons extend more caudally than they normally would. As one might expect, axons from the rostral thalamus express DCC receptors for Netrin-1, which allow them to respond to the attractive properties of this protein. When the function of this DCC receptor was blocked with specific antibodies, the DTR axons became disorganized, and extended more caudally into the VTel, just as in the experiment with the Netrin-deficient telencephalon.
Figure 4. Co-culture of caudal thalamus with ventral telencephalon, in the presence of a control (F) or Unc5-blocking (G) antibody. Note how the DTC axons extend rostrally in G (red arrow).
So the DCC: Netrin-1 interaction can get DTR axons to the correct rostral cortical regions, but now you might be wondering how the caudal thalamic (DTC) axons are kept out of the rostral VTel, and guided instead to their more caudal cortical targets. The answer lies with the Unc5 receptor, which also binds Netrin-1, but unlike DCC, causes axons to be repelled by this protein. Unc5 is expressed in the thalamus, in a pattern complementary to DCC, such that axons from the caudal thalamus (DTC) express lots of Unc5, and only a small amount of DCC. Figure 4 shows that when the repulsive function of the Unc5: Netrin-1 interaction was blocked with Unc5 antibodies, DTC axons inappropriately invaded rostral areas of the ventral telencephalon. Finally, to wrap up this Netrin-1/DCC/Unc5 story and tie it with an elegant bow, Powell and colleagues overexpressed the Unc5 receptor in DTR neurons, causing their axons to be repelled, rather than attracted, by the high levels of Netrin-1 in the rostral VTel, and to extend more caudally than normal as a consequence.
I’ve included only a small fraction of the beautiful photomicrographs of labeled axons in the whole mount telencephalon preps-go take a look at this Open Access article to see the others!
Dufour, A., Seibt, J., Passante, L., et al. (2003) Area specificity and topography of thalamocortical projections are controlled by ephrin/Eph genes. Neuron 39, 453-465.
Vanderhaeghen, P., and Polleux, F. (2004) Developmental mechanisms patterning thalamocortical projections: intrinsic, extrinsic and in between. Trends in Neurosciences 27(7), 384-391.
Wright, A.G., Demyanenko, G.P., Powell, A. (2007) Close homolog of L1 and neuropilin 1 mediate guidance of thalamocortical axons at the ventral telecephalon. J. Neuroscience 27(50), 13667-13679.
Powell, A.W., Sassa, T., Wu, Y., Tessier-Lavigne, M., Polleux, F., Ghosh, A. (2008). Topography of Thalamic Projections Requires Attractive and Repulsive Functions of Netrin-1 in the Ventral Telencephalon. PLoS Biology, 6(5), e116. DOI: 10.1371/journal.pbio.0060116