Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Oct 5.
Published in final edited form as: Neuron. 2016 Oct 5;92(1):45–51. doi: 10.1016/j.neuron.2016.09.035

Clonally related interneurons are not constrained by functional or anatomical boundaries

Christian Mayer 1, Rachel C Bandler 1, Gord Fishell 1,*
PMCID: PMC5241137  NIHMSID: NIHMS819999  PMID: 27710788

Abstract

Clonally related excitatory neurons maintain a coherent relationship following their specification and migration. Whether cortical interneurons behave similarly is a fundamental question in developmental neuroscience. In Mayer et al. 2015, we reported that sibling interneurons disperse over several millimeters, across functional and anatomical boundaries. This finding demonstrated that clonality is not predictive of an interneuron’s ultimate circuit specificity. Comparing the distribution of interneurons published in Mayer et al. to a random computer simulation, Sultan et al. suggest that clonally related interneurons are “not randomly dispersed.” We argue that this comparison provides no insight into the influence of clonality on interneuron development because the entire population of cortical interneurons is “not randomly dispersed” in vivo. We find that the majority of cortical interneurons are similarly distributed whether or not they share a lineal relationship. Thus at present there is no compelling evidence that clonality influences the position or function of interneurons.

Introduction

The mammalian cortex is subdivided into areas devoted to vision, sensation, audition, and other functions. Each area can further be divided physiologically into smaller units or functional columns. Excitatory and inhibitory neurons, the two main cell types of the cortex and hippocampus, have distinct embryonic origins (Anderson et al., 1997), and have segregated into separate lineages by the time the primary prosencephalon has developed into the secondary prosencephalon (Rubenstein et al., 1998). Excitatory cells are derived from the dorsal telencephalon or pallium. Consecutive rounds of asymmetric cell division produce lineage-related sister excitatory neurons that migrate short distances towards the pia into the overlaying developing cortical plate. After migration, this results in spatially organized vertical clusters of excitatory sibling neurons (referred to as clonal units) that form functional columnar microcircuits in the neocortex (Li et al., 2012; Noctor et al., 2001). By contrast, inhibitory cells are entirely derived from the ventral telencephalon or subpallium (reviewed in Marin and Rubenstein, 2001 and Fishell and Rudy, 2011), most prominently from the medial and caudal ganglionic eminences (MGE and CGE, respectively), and migrate over large distances to integrate into the developing cortex, hippocampus, or other subcortical forebrain structures.

Conflicting results from four recent studies examining interneuron lineages

Despite the technical difficulties associated with fate mapping interneuronal lineages due to their complex migration patterns, four recent studies (Brown et al., 2011; Ciceri et al., 2013; Harwell et al., 2015; Mayer et al., 2015) have endeavored to explore whether, similar to what is observed in excitatory neurons, clonally related interneurons are selectively positioned within cortical units. If clonally related interneurons were confined to discrete anatomical brain units (i.e., columns of the cortex), this would support the idea that cell lineage is dictating the integration of interneurons into functional cortical networks.

All four groups agreed that before migration the majority of interneurons are generated from symmetric and asymmetric divisions of MGE progenitor cells, leading to radially aligned interneuron precursors in close proximity to each other (Brown et al., 2011; Ciceri et al., 2013; Harwell et al., 2015; Mayer et al., 2015). Postmitotic interneurons reach their final positions within the cortex through long-range tangential migration that requires them to travel 100 times farther than excitatory pyramidal neurons to reach the cortical plate.

The four studies draw different conclusions about how lineage contributes to the final location of interneurons after long-range migration. Brown et al. and Ciceri et al. described clonal clusters in the cortex that were sufficiently compact to suggest that they are confined by functional boundaries. Specifically, Brown et al. suggested that presumptive clones were aligned into horizontal and radial columns, very similar to their excitatory counterparts (Brown et al., 2011; Yu et al., 2012) (Fig. 1A), raising the possibility of a lineage-dependent functional matching in the organization of inhibitory and excitatory neurons (Brown et al., 2011). Ciceri et al. primarily described laminar clusters. In contrast, Mayer et al. and Harwell et al. both concluded that clonally related interneurons can disperse across anatomical and functional boundaries within the forebrain, and are not restricted to narrow cortical columns or lamina (Fig. 1B). Importantly, Mayer et al. and Harwell et al. agree that sibling interneurons reside in a volume that far exceeds functional cortical units, such as the whisker barrels (Bruno et al., 2003) of the somatosensory cortex (the average distance between pairs of sibling neurons was more than 2mm in (Mayer et al., 2015); Fig. 1B). These data imply that integration of interneurons into functional units is unlikely to be determined by lineage.

Figure 1. The distribution of interneuron clones in Mayer et al. does not support that they are distributed as local radial or laminar clonal clusters.

Figure 1

Open in a new tab

A) Schematic based on Brown et al. Fig. S14 (Brown et al., 2011), illustrating the clonal organization of neocortical interneurons. The authors of the study concluded that “after arriving at their destination in the neocortex, inhibitory interneuron clones do not randomly disperse, but form spatially organized vertical or horizontal clusters.”

B) Reconstructions of cortical clones: Front, side and top view of 3D-reconstructions illustrating the distribution of 16 cortical clones in the forebrain. The average distance (AD) of each clone is indicated. From Mayer et al. (Mayer et al., 2015) Fig. S2.

Differences in experimental methods explain the conflicting results of the four studies

A common feature of all studies considered above was that interneuron progenitors in the MGE of mouse embryos were labeled through infection using fluorescently tagged retroviruses. What then explains the disparate results reported in these four investigations? Discrepancies almost certainly arose from the different methods used to assess and define interneuron clonality. Mayer et al. and Harwell et al. used a replication-defective retroviral library containing a highly diverse set of DNA barcodes, an approach pioneered by Walsh and Cepko in the early 90s (Walsh and Cepko, 1993), to determine lineal relationships between labeled interneurons. Recovering the barcodes from the mature progeny of infected progenitor cells enabled Mayer et al. and Harwell et al. to unambiguously determine the lineal relationship between clones regardless of their geometric distribution within the brain. Brown et al. and Ciceri et al. by contrast used a combination of approaches, including: (1) time-lapse imaging (before migration), (2) mixing of red and green retroviruses and (3) presumptive clonal labeling with low-titer retrovirus injections, followed by the use of geometrical criteria to infer lineal relationships among retrovirally labeled neurons. For the following reasons we believe that none of the aforementioned methods used by Brown et al. and Ciceri et al. reliably indicated lineal relationships among interneurons. First, while in principle time-lapse imaging could be used to determine lineal relationships, given both the distances involved and the protracted time over which interneurons migrate from their birth to their settling position, this approach is impractical. Second, the use of red and green retroviruses is confounded by technical difficulties which, when addressed by the Ciceri et al. study, revealed that assigned clusters of interneurons are polyclonal in nature. In brief, the authors report that when retroviruses encoding GFP and mCherry were mixed before ultracentrifugation, “most clusters were likely to include cells from a different progenitor (i.e., a different fluorescent protein), even at very limiting dilutions.” The authors conclude, “this strongly suggested that lineage relationships are not exclusive determinants of interneuron clustering.” Third, while low-titer retroviral injections, followed by the use of geometrical criteria can in principle be used to determine lineal relationships, in practice this proves untenable. Both Brown et al. and Ciceri et al. compared the distance from each labeled interneuron to its closest neighbor (nearest neighbor distance, or NND) to a randomly computer-simulated dataset to test whether the labeled interneurons were clustered. Ciceri et al. then calculated the number of clusters in the experiment using a threshold distance value that maximized the difference between the number of clusters observed in the experimental data set and the mean number of clusters in 100 simulated populations of randomly distributed neurons. Brown et al. used spatial parameters that picked up excitatory neuron clusters to predict clonally related inhibitory interneuron clusters. In principle, if a single progenitor could reliably label with a single injection, it would indeed be possible to trace sibling interneurons in the forebrain, even if individual siblings pursued drastically different migration paths. However, the labeling of a single progenitor cell cannot be guaranteed with current technology. Brown et al. attempted to use low-titer retroviral injections “to label dividing progenitor cells in the ventricular zone […] at clonal density.” Given this claim, we were surprised when we looked at the raw data provided by Brown et al. (Fig. 2A, B), to see that individual brains contained up to 538 labeled cells, which far exceeds what our analysis indicates would allow for “clonal labeling.” When local clusters are deemed to be clonal, lumping errors (clustered cells that are not clonal) and splitting errors (dispersed cells that are clonal but are not recognized as such) are a major confounding factor, particularly for datasets with a large number of total neurons (such as used in Brown et al., Fig. 2A, B). This is because as the number of labeled neurons increases in a dataset, the chance that a non-clonally related cell will be found nearby clonally related cells increases. Sultan et al. recognize this point, as the authors state, “a clone forming a local cluster does not preclude the presence of nearby non-clonally related interneurons…The more data points, the shorter the distance in general between them. Therefore, it is crucial to take into consideration the total number of data points in each dataset.” Even with lower rates of infections per brain (Fig. 2C, D), Mayer et al. and Harwell et al. reported a large number of interneurons that were nearest neighbors, but were not clonally related (i.e., they had different DNA barcodes, indicating that they originated from different progenitors; See dendrogram analysis in Mayer et al. 2015: Fig. 3D, E). Sultan et al. justify the high number of labeled cells in Brown et al. by stating “a significant fraction (~47.2%; 17 out of 36) of the brains in Brown et al. had a comparable labeling density in the cortex to those in Mayer et al.” This logic is flawed. As mentioned above, determining clonality with geometric means relies on low infection rates to avoid lumping and splitting errors (see dendrogram analysis), whereas if DNA barcodes are used, there is no such requirement for sparse labeling. Harwell et al., in this same issue, provide an additional detailed analysis, showing that the spatial parameters used in Brown et al. to cluster interneurons fail to identify lineal boundaries in either our dataset or that of Harwell et al.. Furthermore, because these methods require that any “clonal” group of cells be constrained to a specified geometric area, as a matter of principle these methods cannot be used to study dispersed clones that reside in different forebrain structures or distant locations within the neocortex.

Figure 2. Comparison of the distributions of retrovirally infected interneurons in Mayer et al. and Brown et al.

Figure 2

Open in a new tab

A+B) Two experimental datasets from Brown et al. are shown. Three-dimensional reconstructions of the distribution of cortical interneurons in a postnatal Nkx2.1Cre/+;R26LSL-TVAiLacZ/+ mouse infected with retroviruses expressing enhanced green fluorescence protein (EGFP). Datasets in Brown et al. contained up to 538 data-points per brain. To predict clonal relations of inhibitory interneurons, Brown et al. applied spatial parameters based on the observed distributions of excitatory neuron clusters (not shown).

C+D) Three-dimensional reconstructions of a representative dataset from Mayer et al., illustrating the distribution of cortical interneurons in a postnatal Nkx2.1Cre/+;R26LSL-TVAiLacZ/+ mouse that was infected with a retroviral library. The same dataset is shown, before (C) and after (D) determination of clonal relations based on retroviral barcodes. The dark-red symbols represent single-cell clones (i.e., neurons harboring a barcode that occurred only once in the dataset); Light-red symbols represent multi-cell clones, whereby symbols with the same shape indicate the location of sister-interneurons (i.e., neurons with the same barcode).

Figure 3. Interneuron clones within the cortex in Mayer et al. are not spatially segregated when compared to a biologically appropriate control group; Relates to Fig. 3F in Mayer et al.

Figure 3

Open in a new tab

A) The Nearest Neighbor Distance (NND) decreases as the number of cells per dataset increases. In contrast, the Average Distance (AD) between pairs of neurons is not sensitive to the total number of data points per dataset. To illustrate this, NNDs and ADs were calculated for a high (200; blue dot) and a low (10; red five-pointed star) number of simulated data-points (200; blue dot) in a given volume; N=100 repeats;

B) Schematic illustration demonstrating how intra and inter-clonal distances were calculated. The “intra-clonal distance” was calculated as the average distance between all possible pairs of clonally related interneurons (Bi). The “inter-clonal distance” was calculated as the average distance between all possible pairs of unrelated interneurons (Bii-iv), which includes the distance between (1) ‘multi-cell clones’ (Bii), (2) ‘single-cell clones’ (Biii) and (3) ‘multi-cell clones’ and ‘single-cell clones’ (Biv).

C) Box and whiskers plot of the intra-clonal and inter-clonal distance. Whiskers indicate Min to Max values. All three datasets from Mayer et al. were included. The individual categories that made up all possible pairs of inter-clonal distance shown. The number above the boxes indicates the n-number of interneuron pairs.

D) Scatter dot plot of the intra-clonal and inter-clonal distance broken down by brain (experimental dataset). No significance was observed when the intra-clonal and inter-clonal distances were compared (Kruskal-Wallis test, multiple comparison; Mann-Whitney Nonparametric t-Test); Number above the boxes indicate the n-number of interneuron pairs.

Although cortical interneurons are clustered compared to a computer-simulated random distribution this finding lacks biological significance

By further analyzing the dataset published by Mayer et al. 2015, Sultan et al. conclude that clonally related interneurons are “not randomly” (or “not non-specifically”) dispersed. We agree with this conclusion. However, given the biological constraints placed on interneurons during their development, it should come as no surprise that interneurons (clonally related or not) are not randomly distributed. For example, it is known that interneurons’ ultimate location in the brain is heavily influenced by (1) their position and time of birth (Miyoshi et al., 2007) (2) prescribed paths of migration (Marin, 2013; Tanaka et al., 2006), and (3) stereotyped radial migration from the marginal and subventricular zone to the cortical plate (Miyoshi and Fishell, 2011). All of these factors indicate that while perhaps the dispersion of interneurons is stochastic, it is also tightly regulated, and therefore a random dispersion model will be grossly inaccurate. Recognizing this, we deliberately avoided making statements about random dispersion in our original study (Mayer et al., 2015).

Clonally related cortical interneurons are no more clustered than interneurons that are not lineally-related

Although clonally related cells disperse several millimeters within the cortex, one might ask whether sibling interneurons reside closer to each other than non-related interneurons. As mentioned above, we argue that this question cannot be addressed by comparing the allocation of sibling interneurons to a computer-simulated random distribution. This is because all interneurons might form clusters independent of their clonal relationship when compared to a random distribution. Therefore, in Mayer et al. we instead compared clonal dispersion (i.e. “intra-clonal distance”) to non-clonal interneuron dispersion (i.e. “inter-clonal distance”). Inter-clonal distance is a biologically relevant control for clonal dispersion because all cells used in the analysis were labeled with the same retroviral approach and therefore have similar birthdates and paths of migration into the cortex. Similar to the analysis done in Mayer et al., but for cortical interneurons only, we examined whether the average distance between pairs of clonally related interneurons was preferentially reduced compared to unrelated interneurons. The “intra-clonal distance” was calculated as the average distance between all possible pairs of clonally related interneurons (Fig. 3Bi). The “inter-clonal distance” was calculated as the average distance between all possible pairs of unrelated interneurons (Fig. 3Bii-iv), which includes the distance between (1) ‘multi-cell clones’ (Fig. 3Bii), (2) ‘single-cell clones’ (Fig. 3Biii) and (3) ‘multi-cell clones’ and ‘single-cell clones’ (Fig. 3Biv). The average distance between 40 pairs of clonally related interneurons in the cortex of P16 mice (AD = 2134 ± 213, SEM, N = 40,) was not statistically different from 926 pairs of clonally unrelated interneurons (AD = 2145 ± 34, SEM, N = 926; P > 0.9, Kruskal-Wallis test, multiple comparison; P = 0.6, Mann-Whitney Nonparametric t-Test; Fig. 3C). Sultan et al. claim “it appears inappropriate to combine them [single cell clones] together with the bona fide clones [multi-cell clones] for the distance and clustering analysis.” We conducted a separate analysis excluding single cell clones and still did not find statistical differences between clonally related interneurons and clonally unrelated interneurons (Fig. 3C). When we broke down the analysis by dataset (i.e., each retrovirally infected brain), we did not detect a statistical difference despite the low numbers of clonally related pairs in each analysis (Fig. 3D; P > 0.1 in all three datasets, Mann-Whitney Nonparametric t-Test). Taken together, our results indicate that the majority of clonally related interneurons are not located closer to each other than a biologically similar group of non-related interneurons.

Sultan et al. also compared intra and inter-clonal distances, but the authors came to different conclusions. Their analysis was performed as following: “For “intra-clonal” and “inter-clonal” distance analysis, all single barcoded cells in individual datasets were considered as a “clone”. “Intra-clonal” distances were computed between all single barcoded interneurons and “inter-clonal” distances were computed between the single barcoded interneuron “clone” and other true interneuron clones in the same dataset”.

Single barcoded cells by definition are not clonal, so we do not know why Sultan et al. computed their intra-clonal distance using single barcoded cells. Any further analysis (e.g. computing “inter-clonal distance”) using this intra-clonal distance will not be accurate.

Problems with Sultan et al.’s interpretation of the nearest neighbor distance analysis

Re-analyzing the data published in Mayer et al. 2015, Sultan et al. highlight the following conclusions: “~60% of local interneuron clusters are clonally related siblings.”, and “At least ~50% of interneuron clones form local clusters.” This is consistent with our original article. We reported that 52% of the clones formed “local clusters” and agree that ~ 60% of “local interneuron clusters” are clonally related siblings. However, one needs to appreciate that cells being scored as “clustered” versus “not-clustered” in the underlying type of analysis (NND, dendrogram analysis), depends on the total number of data points (Fig. 3A). For example, in a dataset containing only one clone consisting of two sibling cells, 100% of “local interneuron clusters” are clonal. Hence, limiting the number of virally infected cells will always result in more clonal cells being nearest neighbors and vice versa (Fig. 3A; and Mayer et al. 2015 Fig S4; Mayer et al., 2015). Therefore, the percentage of “local interneuron clusters” or the percentage of clones that form “local clusters” in the dendrogram analysis has no biological significance. It is simply an epiphenomenon of how many cells were labeled by viral injection.

In Mayer et al., we in fact included this analysis as a cautionary note. We performed the dendrogram analysis to illustrate that lumping errors (clustered cells that are not clonal) and splitting errors (dispersed cells that are clonal but are not recognized as such) are inevitable if geometrical criteria are used to determine the clonality of retrovirally labeled cells, because these criteria implicitly assume that neighboring cells are clonally related (Brown et al., 2011; Ciceri et al., 2013).

We would like to address the “error corrections” made by Sultan et al.. In particular, Sultan et al. state that we failed to add clone #32 to our dendrogram. Calling this an “error correction” is inaccurate because we deliberately excluded this clone from our analysis. Clone #32 was located within the olfactory bulb, and the dendrogram analysis in Mayer et al. “focused on cortical, hippocampal, and striatal clones only.” In another such “error correction,” Sultan et al. note that clone #12 contained 3 cells in the cortex and 3 cells in the hippocampus, requiring them to “add all six clones to the dendrogram.” This implies that we incorrectly excluded these cells from our dendrogram, which again is inaccurate. We deliberately divided clones that crossed anatomical boundaries for analysis within brain structures.

The use of Euclidian distance for distance measurements

Sultan et al. noted that the use of Euclidian distances in structures such as the cortex is problematic, because in many if not most cases, migration along straight lines (for instance in cases where such trajectories would cross ventricles or sulci) is not biologically tenable. Nonetheless, all distances between pairs of neurons described in Mayer et al., as well as Sultan et al. and Brown et al., were calculated as Euclidian distances. Given the impossibility of determining more realistic trajectories, this approach is at least systematic and by its nature chronically underestimates the real distances between neurons. This only strengthens our conclusion that clonal dispersion does not respect functional boundaries, as properly corrected measurements of the distance between clones would only be larger rather than smaller.

Do interneuron clones populate specific brain structures in a stochastic manner?

Sultan et al. emphasize that “~70% of interneuron clones [in the dataset of Mayer et al. 2016] are restricted in one forebrain structure, predominantly the cortex.” This precisely reproduces the findings that we published in our original article (Mayer et al., 2015). We concluded that the remaining 30% of clones spanning more than one brain structure provide clear cases in which sibling neurons are neither confined to one brain structure nor to a functional cortical unit. Clones spanning across brain structures was a novel finding that had not been reported by previous studies. While the results of Mayer et al. and Harwell et al. demonstrate that interneuron clones are not obliged to populate particular anatomical structures, this does not rule out the possibility that they are predetermined to particular brain regions. Sultan et al. discuss this point as follows:

“1) Should lineage relationship have no influence on interneuron distribution, the relatively total interneuron output to different forebrain structures and the small clone size dictate that virtually all clones must be located in the cortex, the cortex and hippocampus, or the cortex and striatum. Interestingly, a significant fraction of clones was observed to be restricted to the hippocampus or striatum (Mayer et al., 2015), suggesting that some MGE/PoA progenitors specifically produce interneurons destined for these two brain structures.”

“2) While it is evident that the majority (~66% in Mayer et al. and 80% in Harwell et al.) of clones are located within one brain structure, i.e. the cortex, some are dispersed in more than one brain structure. However, this clonal dispersion largely occurs between the cortex and hippocampus, the two highly related forebrain structures emerging side-by-side in the dorsal telencephalon. The same tangential migration routes are responsible for interneuron distribution in the cortex and hippocampus (Ayala et al., 2007; Marin and Rubenstein, 2001, 2003). In comparison, only a small fraction (~12.5% in Mayer et al. and 20% in Harwell et al.) of clones is dispersed between developmentally unrelated brain structures such as the cortex and striatum, or globus pallidus, or olfactory bulb.“

While these statements are factually correct, understanding their implications requires a more nuanced analysis. Both the absolute size of the cortex, hippocampus and striatum, as well as the density of interneurons within these structures differs dramatically. For example, twenty percent of the cells within the cortex and hippocampus are interneurons (Fishell and Rudy, 2011), whereas the percentage of interneurons within the striatum is only three percent (Marin et al., 2000; Tepper et al., 2010). These facts demonstrate that even if interneurons were randomly distributed to different structures, probabilistically, they would be preferentially found in the cortex. That said, we again reiterate that we do not believe the distribution of interneurons is random. But what rules then underlie the distribution of discrete interneuron lineages? Our results definitively indicate that if interneuron lineages do have a covert logic as to how they populate different structures, clearly the rules of allocation are not as simple as a lineage being earmarked for cortex or hippocampus per se. Further examination of interneuron lineages will be required to address whether there is a degree of predetermination in the positioning of sibling neurons derived from a common lineage.

Final remarks

We show here and in Mayer et al. (Mayer et al., 2015), that clonally related interneurons are not more closely clustered than non-lineally related interneurons (proximally generated brethren). This of course neither should nor does end the debate as to whether lineage contributes to the development, subtype differentiation or connectivity of interneurons. Our own results are limited by the fact that the lineages we assembled were only partially reconstructed and we can say nothing regarding the fate of those sibling cells that we failed to recover. Furthermore, we know startling little about the phenotypic identity and nothing about the connectivity of clonally related siblings, both of which would be fascinating to explore. We would however implore any further examination of lineage to confine themselves to methods that provide high confidence of the lineage relationship of cells designated as clones.

Methods

Datasets published by Mayer et al. 2015 were used to calculate intra-clonal and inter-clonal distances of cortical interneurons. Cartesian coordinates of every neuron located in the cortex from which we recovered a barcode were exported from Neurolucida software (MBF Bioscience) to Matlab software (Mathworks), to calculate Euclidian distances between pairs of neurons. The pairwise distance was only calculated for cells located within the cortex and within one hemisphere. Multi-cellular cortical clones with sibling cells located outside the cortex were also included in the analysis, however, siblings positioned outside the cortex were excluded from the analysis. The “intra-clonal distance” was calculated as the average distance (AD) between all possible pairs of clonally related interneurons. The “inter-clonal distance” was calculated as the average distance between all possible pairs of unrelated interneurons, which includes the distance between (1) ‘multi-cell clones’, (2) ‘single-cell clones’ and (3) ‘multi-cell clones’ and ‘single-cell clones’ (Fig. 3B). Data are presented as mean ± SEM. Kruskal-Wallis multiple comparison test and Mann-Whitney Nonparametric t-Test were used for statistical significance estimations in Prism software (Graphpad). To illustrate that the NND of data-points in a given volume depends on the total number of data points, NNDs and ADs were calculated in Matlab for a high (200) and a low (10) number of simulated data-points (using ‘rand’ and ‘pdist’ functions; Fig. 3A). Each simulation was repeated 100 times and averaged.

Acknowledgments

We are grateful to Robert Machold, Xavier Jaglin, and Timothy Petros for comments on the manuscript. Research in the Fishell laboratory is supported by the NIH (NS081297, MH095147, MH071679, P01NS074972) and the Simons Foundation.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Author Contributions

C.M. conceptualized, designed, performed, and analyzed experiments and co-wrote the paper. R.C.B. and G.F. conceptualized and designed experiments and co-wrote the paper.

References

  1. Anderson SA, Eisenstat DD, Shi L, Rubenstein JL. Interneuron migration from basal forebrain to neocortex: dependence on Dlx genes. Science. 1997;278:474–476. doi: 10.1126/science.278.5337.474. [DOI] [PubMed] [Google Scholar]
  2. Brown KN, Chen S, Han Z, Lu C-H, Tan X, Zhang X-J, Ding L, Lopez-Cruz A, Saur D, Anderson SA, Huang K, Shi S-H. Clonal production and organization of inhibitory interneurons in the neocortex. Science. 2011;334:480–486. doi: 10.1126/science.1208884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bruno RM, Khatri V, Land PW, Simons DJ. Thalamocortical angular tuning domains within individual barrels of rat somatosensory cortex. J. Neurosci. 2003;23:9565–9574. doi: 10.1523/JNEUROSCI.23-29-09565.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Ciceri G, Dehorter N, Sols I, Huang ZJ, Maravall M, Marin O. Lineage-specific laminar organization of cortical GABAergic interneurons. Nat. Neurosci. 2013;16:1199–1210. doi: 10.1038/nn.3485. [DOI] [PubMed] [Google Scholar]
  5. Fishell G, Rudy B. Mechanisms of inhibition within the telencephalon: "where the wild things are". Annu. Rev. Neurosci. 2011;34:535–567. doi: 10.1146/annurev-neuro-061010-113717. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Harwell CC, Fuentealba LC, Gonzalez-Cerrillo A, Parker PRL, Gertz CC, Mazzola E, Garcia MT, Alvarez-Buylla A, Cepko CL, Kriegstein AR. Wide Dispersion and Diversity of Clonally Related Inhibitory Interneurons. Neuron. 2015;87:999–1007. doi: 10.1016/j.neuron.2015.07.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Li Y, Lu H, Cheng P-L, Ge S, Huuatai X, Shi S-H, Dan Y. Clonally related visual cortical neurons show similar stimulus feature selectivity. Nature. 2012;486:118–121. doi: 10.1038/nature11110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Marin O. Cellular and molecular mechanisms controlling the migration of neocortical interneurons. Eur. J. Neurosci. 2013;38:2019–2029. doi: 10.1111/ejn.12225. [DOI] [PubMed] [Google Scholar]
  9. Marin O, Anderson SA, Rubenstein JL. Origin and molecular specification of striatal interneurons. J. Neurosci. 2000;20:6063–6076. doi: 10.1523/JNEUROSCI.20-16-06063.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Marin O, Rubenstein JLR. A long, remarkable journey: Tangential migration in the telencephalon. Nat. Rev. Neurosci. 2001;2:780–790. doi: 10.1038/35097509. [DOI] [PubMed] [Google Scholar]
  11. Mayer C, Jaglin XH, Cobbs LV, Bandler RC, Streicher C, Cepko CL, Hippenmeyer S, Fishell G. Clonally Related Forebrain Interneurons Disperse Broadly across Both Functional Areas and Structural Boundaries. Neuron. 2015;87:989–998. doi: 10.1016/j.neuron.2015.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Miyoshi G, Butt SJB, Takebayashi H, Fishell G. Physiologically distinct temporal cohorts of cortical interneurons arise from telencephalic Olig2-expressing precursors. J. Neurosci. 2007;27:7786–7798. doi: 10.1523/JNEUROSCI.1807-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Miyoshi G, Fishell G. GABAergic interneuron lineages selectively sort into specific cortical layers during early postnatal development. Cereb. Cortex. 2011;21:845–852. doi: 10.1093/cercor/bhq155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Noctor SC, Flint AC, Weissman TA, Dammerman RS, Kriegstein AR. Neurons derived from radial glial cells establish radial units in neocortex. Nature. 2001;409:714–720. doi: 10.1038/35055553. [DOI] [PubMed] [Google Scholar]
  15. Rubenstein JL, Shimamura K, Martinez S, Puelles L. Regionalization of the prosencephalic neural plate. Annu. Rev. Neurosci. 1998;21:445–477. doi: 10.1146/annurev.neuro.21.1.445. [DOI] [PubMed] [Google Scholar]
  16. Tanaka DH, Maekawa K, Yanagawa Y, Obata K, Murakami F. Multidirectional and multizonal tangential migration of GABAergic interneurons in the developing cerebral cortex. Development. 2006;133:2167–2176. doi: 10.1242/dev.02382. [DOI] [PubMed] [Google Scholar]
  17. Tepper JM, Tecuapetla F, Koos T, Ibanez-Sandoval O. Heterogeneity and diversity of striatal GABAergic interneurons. Frontiers in Neuroanatomy. 2010;4:150. doi: 10.3389/fnana.2010.00150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Walsh C, Cepko CL. Clonal dispersion in proliferative layers of developing cerebral cortex. Nature. 1993;362:632–635. doi: 10.1038/362632a0. [DOI] [PubMed] [Google Scholar]
  19. Yu Y-C, He S, Chen S, Fu Y, Brown KN, Yao X-H, Ma J, Gao KP, Sosinsky GE, Huang K, Shi S-H. Preferential electrical coupling regulates neocortical lineage-dependent microcircuit assembly. Nature. 2012;486:113–117. doi: 10.1038/nature10958. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES