SOX4 Helps Specify Intermediate Progenitor Cells
Chao Chen, Garrett A. Lee, Ariel Pourmorady, Elisabeth Sock, and Maria J. Donoghue
(see pages 10629–10642)
The brain acquires its intricate structure through a tightly regulated developmental sequence orchestrated by numerous transcription factors. In the developing neocortex, activation and suppression of different transcriptional programs drives transitions from neuroepithelial cell to radial glia to intermediate progenitor to postmitotic neuron. Transcriptional programs that define different developmental stages and different cell types often include overlapping sets of transcription factors, which can acquire different roles depending on what other proteins are expressed. This is particularly evident for the SOX family of transcription factors. SOX proteins bind to DNA, but they cannot alter transcription by themselves. Instead, they interact with partners to either enhance or repress transcription. Changing either the SOX protein or the partner can alter the transcriptional program. In the developing nervous system, expression of proteins in the SOXB1 subfamily maintains the proliferative state of neural progenitors, SOXB2 subfamily members repress the genes activated by SOXB1 proteins, and expression of SOXC proteins accompanies the transition from progenitor to neuron (reviewed in Kamachi and Kondoh, 2013, Development 140:4129).
Each SOX subfamily has multiple members that are thought to act largely redundantly. Thus, whereas knocking out either SOX4 or SOX11 (members of the SOXC subfamily) appears to have little effect on mouse nervous system development, knocking out both has drastic effects. Nevertheless, Chen et al. have discovered that these two proteins have partially unique functions in neocortical development. Expression of SOX11 and SOX4 was elevated during cortical neurogenesis, but the expression patterns only partially overlapped. Specifically, SOX11 expression was skewed toward deep cortical layers, whereas SOX4 expression was skewed toward superficial layers. Unexpectedly, SOX4 was also expressed in cells that expressed neurogenin2, a marker of intermediate progenitor cells (IPCs). Furthermore, most cells that expressed the IPC transcription factor Tbr2 also expressed SOX4. Notably, overexpressing Sox4 increased, whereas knocking out SOX4 reduced, the number of Tbr2-expressing cells.
Based on these and previous results, the authors hypothesize that SOX4 partners with neurogenin2 to drive expression of Tbr2, thus specifying IPCs. If this is true, loss of SOX4 may have profound effects in primates, in which evolutionary expansion of the IPC population is thought to have enabled cortical enlargement.
Tectogeniculate Afferents May Be Drivers, Not Modulators
Martha E. Bickford, Na Zhou, Thomas E. Krahe, Gubbi Govindaiah, and William Guido
(see pages 10523–10534)
Most retinal ganglion cells (RGCs) innervate neurons in the thalamic lateral geniculate nucleus (LGN), which relay visual information to primary visual cortex. But RGCs provide only ∼10% of the synaptic input to LGN neurons: the remainder comes from cortex, subcortical structures, the thalamic reticular nucleus, and local interneurons. Inputs to thalamocortical relay neurons are often classified as drivers or modulators based on how much they influence the neurons' receptive field properties. Driver afferents typically have large synaptic terminals and evoke large EPSPs by activating ionotropic glutamate receptors on the proximal dendrites of thalamocortical neurons. These synapses exhibit strong paired-pulse depression, suggesting they have a high release probability. Modulator afferents, in contrast, have small synaptic terminals that release glutamate, GABA, or neuromodulators. They produce small PSPs, they can activate metabotropic receptors, and they exhibit paired-pulse facilitation (reviewed in Sherman, 2007, Curr Opin Neurobiol 17:417).
Tectogeniculate (dark reaction product) and retinogeniculate (red) afferents contact (arrows) the same dendrite (green) of a thalamocortical neuron in the dorsolateral shell of the dorsal LGN. The retinogeniculate afferent also contacts a GABAergic neuron (purple overlain with gold particles). See Bickford et al. for details.
RGCs have long been considered the sole drivers of thalamocortical LGN neurons. But Bickford et al. report that synapses formed by tectogeniculate projections from the superior colliculus (SC) to the dorsal LGN′s dorsolateral shell exhibit many of the properties of driver synapses. Nearly all tectogeniculate inputs were glutamatergic and innervated the proximal dendrites of LGN neurons. Furthermore, optogenetic activation of tectogeniculate terminals evoked large EPSPs in LGN neurons, and these EPSPs were fully dependent on ionotropic glutamate receptors. Finally, tectogeniculate synapses exhibited paired-pulse depression. Neurons that responded to tectogeniculate stimulation also responded to retinogeniculate stimulation, showing that these two inputs converge on individual cells.
These data indicate that, contrary to conventional models of thalamic circuitry, the receptive fields of thalamocortical relay neurons in the dorsolateral shell of the LGN may be determined by inputs from two sources: the retina and the SC. Previous studies have shown that neurons in the dorsolateral shell are direction-sensitive and are innervated by direction-sensitive RGCs. If tectogeniculate afferents are truly drivers, the direction selectivity of postsynaptic LGN neurons should be determined by these inputs in conjunction with those from the retina. This may help animals account for their own eye movements while visually tracking the movements of external stimuli.
