The remarkably selective response properties of individual neurons in visual cortex result from specific patterns of synaptic connections that link large numbers of cortical neurons. In some species, including primates and carnivores, cortical neurons with similar response properties (e.g. similar preferred orientation) and shared connectivity are grouped together into radial columns, forming orderly maps of stimulus features (Hubel and Wiesel, 2005
). In rodents, cortical neurons with different orientation preferences are intermingled in a ‘salt-and-pepper’ fashion (Ohki et al., 2005
). Nevertheless rodents exhibit fine scale specificity in the organization of synaptic connections (Yoshimura and Callaway, 2005
, Yoshimura et al., 2005
) and preferential connectivity among neurons with similar orientation tuning (Bock et al., 2011
, Ko et al., 2011
). Whether arranged as columns or not, functionally defined neural circuits are a fundamental feature of cortical organization, and the developmental mechanisms that are responsible for their construction remain an important and unresolved problem.
Several recent studies have shed new light on this issue, suggesting that cell lineage plays a key role in laying down the scaffold for building functionally distinct cortical circuits. By tracing the neurons that are derived from a single radial glia progenitor cell, Yu et al. (2012)
demonstrated that ‘sister neurons’ have a much higher probability of being electrically coupled via gap junctions than non-sister pairs and that sister neurons are more likely to be connected via chemical synapses later in development (Yu et al., 2009
). Li et al. (2012)
combined lineage tracing of single radial glia progenitors with in vivo
2-photon imaging of calcium signals to demonstrate that sister neuron pairs are more likely to have similar orientation preferences than non-sister pairs, and that this similarity depends on the presence of functioning gap junction communication during the first postnatal week. Taken together these results provide compelling support for cell lineage as a significant factor in determining the specificity of connections that underlies functionally defined cortical circuits in rodents (Li et al., 2012
, Mrsic-Flogel and Bonhoeffer, 2012
In this issue of Neuron, Ohtsuki and colleagues
have used a different approach to address the role of lineage in the specification of cortical circuits. While their study adds evidence supporting a role for lineage, it also suggests that the role of lineage is limited, and that other factors may play significant roles in specification of functionally defined cortical circuits. Previous studies have focused on a small number of progeny derived from a single radial glial cell at a relatively late stage in the generation of cortical neurons. The study by Ohtsuki et al. (Ohtsuki et al 2012
) was designed to examine the large number of neurons that are derived from a single progenitor cell at an earlier time point in development. The authors use a mouse driver line that expresses Cre-recombinase in a sparse subset of progenitor cells to label a population of 600–800 radially-dispersed neurons derived from a single progenitor (Magavi et al., 2012
). The authors then used in vivo
2-photon calcium imaging to examine the orientation tuning properties of both clonally-related neurons and surrounding cells derived from different progenitors. Orientation preferences among clonally-related cells were more similar than among unrelated neurons, and, in several cases, the tuning preference of the clone was significantly different from the surrounding population. However, there was also considerable diversity within the clonally-related population of neurons, with nearly half of all neuron pairs showing differences in orientation preference greater than 30 degrees, and a quarter exhibiting differences greater than 60 degrees.
Assuming that the approaches used in both studies are labeling the progeny of single progenitor cells, how does one explain the greater diversity in the functional properties of clonally derived neurons in the Ohtsuki et al
. study? At first glance, it is tempting to conclude that these results can be explained by the fact that neuronal fate simply becomes more restricted over the course of development. Daughters of a single radial glial cell formed relatively late in the generation of cortical neurons are more similar to each other than two neurons whose lineage diverged from a common progenitor a number of generations earlier; and this difference accounts for the increased similarity in functional properties. In this regard it would be interesting to determine whether the preferential patterns of connectivity that have been found for the daughters of a single radial glia cell late in development are missing for the larger multigenerational clones labeled in the study by Ohtsuki et al
But factors other than the number of generations represented within the clonal population may be relevant for understanding the broader range of orientation preferences that are found in the Ohtsuki et al
. study. Because their progenitors were labeled at a relatively late time point in the generation of cortical neurons, the neurons imaged by Li et al. (2012)
lie very close to each other; most are displaced a small distance axially and exhibit only modest lateral displacement (most less than 90 μm). In contrast, the progeny derived from a common ancestor in the Ohtsuki et al
. study occupy a lateral extent of several hundred microns. It may be the case that the impact of lineage on patterns of connectivity and functional properties declines as a function of distance, and the disparate tuning preferences observed for clonally-related cells by Ohtsuki et al
. reflect the inclusion of progeny that are located at separation distances not present in the Li et al
. study. Further analysis of the Ohtsuki et al
. data could evaluate the interaction between cell lineage and separation distance in determining shared connectivity and response properties.
If distance is a factor in the degree of clonal specificity, it raises the possibility that the shared micro-environment in which the neurons migrate and develop may contribute to the similarity in their connections and response properties. Indeed, the neurons whose properties were evaluated in the Li et al
. study are likely to have migrated along the same radial glial fiber, encountering many of the same features en route to their ultimate locations, while those in the Ohtsuki et al
. study had much more diverse routes to their final destinations. In principle, a shared developmental journey through the cortical neuropil could contribute as much to the similarity in connectivity and response properties as shared lineage. However, teasing apart the contribution of shared ancestry and developmental micro-environments is a challenging task.
Compounding the difficulty in reconciling the results from these two studies is the fact that Ohtsuki et al
. and Li et al
. also differ in the developmental time-point at which they assessed the orientation preferences of their clonally derived neurons. Li et al. (2012)
found great similarity in animals that were imaged shortly after eye-opening (P12–17), whereas Ohtsuki et al
. observed more diversity in preference in older animals (P49–62). Among sister cells derived from a single radial glia, gap junction coupling declines from P1–2 and is nearly absent by P6 (Yu et al., 2012
) with preferential chemical synaptic connectivity appearing by P10–17 (Yu et al., 2009
). It may be that this preferential clonal connectivity, along with the response similarity it helps convey, dominates early cortical networks, but is eroded with visual experience and the accompanying strengthening of connections from unrelated neurons through mechanisms of Hebbian synaptic plasticity. Alternatively, the similarity in connectivity and response properties among closely related sister neurons may be maintained throughout development, and this accounts for the degree of similarity in orientation preference that is seen in the Ohtsuki et al
. study. Additional experiments that explore the properties of early and late clonally derived populations at different postnatal ages would clarify the extent to which visual experience impacts the patterns of connections and response properties that are specified by cell lineage.
The current study by Ohtsuki et al
., along with that of Li et al
., establishes a clear link between cortical cell lineage and shared response properties. At the same time they emphasize how much we have yet to understand about how lineage combines with other mechanisms to specify the connectivity and response properties of cortical circuits.