The development of the neocortex requires co-ordination between proliferation and differentiation, as well as the precise orchestration of neuronal migration. Eph/ephrin signaling is crucial in guiding neurons and their projections during embryonic development. In adult ephrin-A2 knockout mice we consistently observed focal patches of disorganized neocortical laminar architecture, ranging in severity from reduced neuronal density to a complete lack of neurons. Loss of ephrin-A2 in the pre-optic area of the diencephalon reduced the migration of neocortex-bound interneurons from this region. Furthermore, ephrin-A2 participates in the creation of excitatory neurons by inhibiting apical progenitor proliferation in the ventricular zone, with the disruption of ephrin-A2 signaling in these cells recapitulating the abnormal neocortex observed in the knockout. The disturbance to the architecture of the neocortex observed following deletion of ephrin-A2 signaling shares many similarities with defects found in the neocortex of children diagnosed with autism spectrum disorder.
Development of the neocortex relies on a balance of progenitor cell proliferation and differentiation in the ventricular zone (VZ) to generate the diversity of neuronal populations. Maintenance of an adequate pool of mitotic progenitors is achieved through signaling pathways promoting symmetrical self-renewing division, while activation of proneural genes promotes the production of neurons through asymmetrical division generating a neuron and a single apical progenitor. During the later stages of corticogenesis, progenitor cells switch to terminal symmetric division producing two neurons, ultimately depleting the progenitor pool1,2,3. The equilibrium between symmetric and asymmetric division is controlled by sets of instructions present in the local environment mediated through cell-cell interactions. The Notch-Delta pathway is particularly important in regulating the transition from one division mode to the other4 but guidance molecules, identified for their ability to direct neuronal migration and axon navigation, can also act as fate determinants5. Disruption of neurogenesis has dramatic consequences on the establishment of the neocortex, leading to severe malformations and cognitive disorders, including Autism Spectrum Disorder (ASD)6.
Glutamatergic excitatory neurons, the largest fraction of neocortical neurons, are born locally in the neocortex, from Pax6+ apical progenitors in the VZ and Tbr2+ basal progenitors in the SVZ7,8,9. Following the upregulation of neuronal markers including NeuN and Tbr1, the newborn excitatory neurons follow a radial migratory route1,10,11 to their appropriate laminar position in the developing neocortex.
In contrast, interneurons originate from subcortical domains, combining tangential and radial migration to reach the neocortex3,12. In rodents, 90% of neocortical interneurons are born in the ventral telencephalon, including the ganglionic eminences13, with the remaining 10% emerging from the diencephalic preoptic area (POA)14,15,16. Interneurons arising from these distinct domains give rise to non-overlapping subtypes occupying specific compartments in the neocortex17,18,19.
Eph/ephrin signaling regulates many aspects of cell interaction in the developing neocortex, including migration and axonal target recognition20. Ephrins are membrane bound ligands, tethered to the cell membrane via a GPI link (ephrin-As) or transmembrane region (ephrin-Bs). They interact with Eph tyrosine kinase receptors triggering bi-directional responses: a forward signaling cascade in the receptor-bearing cell and a reverse signaling cascade in the ligand-bearing cell21,22,23. Receptor-ligand interactions principally elicits repulsion, contributing to the migration and segregation of neuronal populations including neocortical excitatory neurons24,25 and interneurons26,27,28. Eph/ephrin interactions also regulate other aspects of neocorticogenesis, including brain size through regulation of apoptosis29 and progenitor proliferation30.
Here we describe focal lamination defects in the neocortex of all adult ephrin-A2 homozygote knockout mice (efnA2 KO), forming discrete patches of reduced neuronal density. Their distribution is disparate and not restricted to specific neocortical areas, highly reminiscent of similar aberrations observed in the neocortex of children diagnosed with Autism Spectrum Disorder (ASD)31. Using in vivo and ex vivo knockdown approaches we demonstrate that ephrin-A2 contributes to the establishment of excitatory and inhibitory neocortical neurons, with distinct roles in each population. Specifically, ephrin-A2 promotes the migration of interneurons emerging from the POA and modulates the proliferation of apical progenitors in the neocortex. We propose that ephrin-A2 acts as a proneural cell fate determinant, with the loss of ephrin-A2 resulting in abnormal neocortical lamination akin to ASD-associated deficits.
Disruption of the neuronal cytoarchitecture in the neocortex of efnA2 KO mice
The neocortical cytoarchitecture of adult mice lacking efnA2 appeared normal when labeled with the DNA stain, Hoechst (Fig. 1a). However, labeling with the neuronal marker, NeuN, revealed discrete patches with reduced density of neuronal cell bodies (Fig. 1a’) and dendritic arborisation (Fig. S1a,b’). The patches established sharp boundaries between normal and disrupted zones and extended through the neocortex with the loss of neurons becoming more severe towards the center of the patch, as illustrated on the series of adjacent sections (Fig. 1a’). The phenotype varied in severity, yet occurred with a 100% penetrance, with all mutant brains exhibiting at least one disorganized patch per hemisphere. Their distribution, however, was highly variable and apparently random, with the disruption affecting all layers indiscriminately (Fig. 1b’, layers 2/3; Fig. 1c’, layer 5), confirmed by a signal reduction on the fluorescence intensity profile (Fig. 1b”, c”). The stability of the Hoechst nuclear labeling (Fig. 1b”, c”) and haematoxylin and eosin histology (Fig. S1c, d’) demonstrates that despite the reduction in the density of NeuN+ cells, the neocortical patches were uniformly populated with structurally normal cells. To characterize the identity of the cells localized within the patches we double-labeled efnA2 KO sections with NeuN to demarcate the abnormal regions, and either ALDH1L1, Olig2, or Iba1 to label astrocytes, oligodendroglia or microglia/macrophages respectively. Both astrocytes and oligodendrocytes were noticeably reduced within the regions lacking neurons, with only a few cells observed close to the border (Fig. 1d–d’), contrasting with the uniform distribution of microglia (Fig. 1d”). The expression profile of 10 additional markers is summarized in Supplementary Table 1. Outside the perturbed regions, the distribution of neurons appeared normal, and the cell density was not significantly different from wild type (number of neurons per 0.01 mm2; wild type, 39.2 ± 2.32; efnA2 KO, 38.01 ± 1.62, n = 3; Fig. 1e–e”).
Ephrin-A2 modulates the transition of neocortical progenitors towards differentiation
The loss of neuronal cell labeling in the neocortex could indicate a developmental role for efnA2. We, therefore, examined the developmental expression of ephrin-A2 in the wildtype neocortex. At E14.5 ephrin-A2 was detected in the marginal zone (MZ) and in the VZ, comprised of apical progenitor cells (Fig. 2a–a”). Intense ephrin-A2 expression persisted in the VZ at E16.5, with expression extending to the adjacent SVZ (Fig. 2b’). Ephrin-A2+ cells in the cortical plate (CP) also expressed the neuronal marker NeuN (Fig. 2b”), revealing that ephrin-A2 is present in both progenitor cells and differentiated neurons in the developing neocortex.
To explore the role of ephrin-A2 in the establishment of neuronal populations in the neocortex, we performed loss- and gain-of-function in the apical progenitors lining the ventricular surface of the telencephalon. We electroporated bicistronic plasmid DNA encoding the GFP coding sequence and either ephrin-A2 (efnA2-GFP), shRNA against ephrin-A2 (sh_efnA2-GFP) or a scrambled shRNA control sequence (sh_ctrl-GFP) at E15.5. At 3 days post electroporation (dpe), GFP+ cells were present in all compartments of the embryonic neocortex, with the exception of the acellular MZ (Fig. 2c–c”’) in varying quantities. While the proportion of GFP+ cells transiting through the intermediate zone did not vary between conditions (scrambled control, 0.33 ± 0.02, n = 5; ephrin-A2, 0.40 ± 0.04, n = 4; sh_efnA2, 0.36 ± 0.01, n = 4; Fig. 2d), the fraction of cells entering the CP was comparable between control and ephrin-A2 overexpressing conditions but significantly reduced by ephrin-A2 knockdown (scrambled control, 0.57 ± 0.02, n = 5; ephrin-A2, 0.53 ± 0.05, n = 4; sh_efnA2 0.22 ± 0.02, n = 4; p = 0.01, Kruskal-Wallis test; Fig. 2d). Only a small fraction of control and ephrin-A2 overexpressing cells remained in the neurogenic ventricular and subventricular layers and this proportion was significantly increased following ephrin-A2 knockdown (scrambled control, 0.09 ± 0.02, n = 5; ephrin-A2, 0.07 ± 0.01, n = 4; sh_efnA2 0.42 ± 0.01, n = 4; p = 0.005, Kruskal-Wallis test; Fig. 2d).
Failure of cells to exit the neurogenic compartment following ephrin-A2 knockdown suggests that ephrin-A2 signaling is necessary for newly generated neurons to migrate out of the neurogenic compartment. Alternatively, ephrin-A2 may act at an earlier stage of neurogenesis, with the reduction in ephrin-A2 resulting in progenitors failing to commit to a differentiated neuronal fate and arrest in the VZ.
To determine the identity of the sh_efnA2-GFP cells, we tested for the expression of progenitor markers and demonstrated that efnA2 knockdown cells persisting in the neurogenic zone maintain the expression of Pax6 (Fig. 2e–e”). In the embryonic neocortex, Pax6 expression is restricted to apical progenitors dividing at the ventricular surface9. Pax6 controls neocortical cell numbers by regulating the balance between proliferation and cell cycle exit32. Altogether, our results suggest that loss of ephrin-A2 causes progenitors to remain in the VZ and maintain Pax6 expression, which ultimately would result in increased proliferation at the expense of neuronal differentiation.
Ephrin-A2 reverse signaling regulates neocortical progenitor proliferation
We further investigated the relationship between ephrin-A2 and cell cycle and whether this effect was mediated through forward or reverse signalling using a combination of in utero electroporation and in vitro approaches. To determine if ephrin-A2 knockdown at E15.5 affected the proliferation of apical progenitors, we quantified the proportion of proliferative cells within the GFP+ population in the VZ at 3dpe (Fig. 3a). The fraction of GFP+ cells expressing the proliferation marker Ki67 increased in sh_efnA2-GFP electroporated mice compared to controls (sh_ctrl-GFP, 0.37 ± 0.05; sh_efnA2, 0.53 ± 0.08, n = 5, p = 0.01, Mann-Whitney test; Fig. 3a’). This result confirms that reducing ephrin-A2 causes apical progenitors to remain in a proliferative undifferentiated progenitor state.
It is common for cells expressing ephrin ligands to co-express Eph receptors, segregated in separate microdomains to prevent cis-activation33. Therefore, the phenotype observed following loss of ephrin-A2 could result from a deficit in forward or reverse signaling. To discriminate between forward or reverse signaling in the control of apical progenitor differentiation, we dissociated the neocortices of E14.5 wild-type mice and plated the cells at low density to limit cell-cell interaction. Next, we added clustered recombinant frame chain (Fc) ephrin-A2 (ephrin-A2-Fc) to induce ephrin-A2 forward signaling (Fig. 3b’); reverse signaling by adding clustered EphA4-Fc (Fig. 3b”); with clustered Fc fragment used as a negative control (Fig. 3b). While clustered EphA4-Fc has been demonstrated to activate ephrin-A2 reverse signaling28, it could potentially bind ephrin-A5 as well. However, the absence of ephrin-A5 expression in the embryonic neocortex between E13.5 and E16.534 ensured that ephrin-A2 reverse signaling was selectively activated in our culture system. BrdU was used to tag cells in the S-phase of the cell cycle. We found that the number of BrdU+ cells did not vary between control and ephrin-A2-Fc conditions. However, activation of ephrin-A2 reverse signaling significantly reduced the number of BrdU+ cells (control-Fc, 16.9 ± 0.94; ephrin-A2-Fc, 17.1 ± 1.3; EphA4-Fc, 11.3 ± 0.55, n = 3, p < 0.001, Kruskal-Wallis test; Fig. 3b”’). Therefore, the increase in proliferation of neocortical progenitors following ephrin-A2 knockdown is likely to be mediated via a reduction in ephrin-A2 reverse signaling rather than a reduction in forward signaling on Eph receptor bearing cells.
Knockdown of ephrin-A2 in neocortical progenitors reduces the production of excitatory neurons
To investigate the long-term effect of ephrin-A2 knockdown on cortical progenitors, animals electroporated with sh_efnA2-GFP at E15.5 were analyzed at P12. The laminar distribution of electroporated neurons was similar between control and ephrin-A2 knockdown animals, with the majority of GFP+ cells present in the supergranular layers (Fig. 4a,a’). However, some GFP+ cells remained in the SVZ of ephrin-A2 knockdown mice (Fig. 4a’, arrowheads) whereas there were very few GFP+ cells in the SVZ of sh_ctrl-GFP mice. Consequently, the proportion of sh_efnA2-GFP+ cells in the supergranular layers was reduced compared to the control (layer 2/3: scrambled, 0.91 ± 0.0.3, n = 5; sh-efnA2, 0.74 ± 0.09, n = 4, p = 0.016, Mann-Whitney test; Fig. 4a”), and the proportion of sh_efnA2-GFP+ cells in the SVZ was increased relative to controls (scrambled, 5.7 × 10−5 ± 1.3 × 10−4, n = 5; sh-efnA2 0.115 ± 0.003, n = 4, p = 0.018, Mann-Whitney test; Fig. 4a”). We confirmed that the GFP + cells in the SVZ of sh_efnA2-GFP mice expressed the transcription factor Sox2, consistent with a progenitor identity (Fig. 4b–b’, arrowheads). Therefore, while the vast majority of cells electroporated with the control shRNA differentiated into neurons and migrated to supergranular layers, over 10% of ephrin-A2 knockdown cells maintained their progenitor identity for 2 weeks after electroporation.
Despite cellular density appearing normal in the neocortices of ephrin-A2 knockdown animals (Fig. 4c), we detected discontinuities in the laminar distribution of GFP+ neurons in the electroporated hemispheres (Fig. 4c’). The supergranular layers of ephrin-A2 knockdown neocortices exhibited discrete regions devoid of GFP + cells which were also devoid of NeuN + cells (Fig. 4c”), similar to the cortical defects (patches) we described in the neocortex of adult efnA2 KO (Fig. 1). Therefore, loss of ephrin-A2 in apical progenitors through targeted knockdown is sufficient to phenocopy the abnormal laminar architecture observed in the knockout animals.
Reduction of the calbindin-positive interneuron fraction in the supragranular layers of efnA2 KO mice
To determine if loss of ephrin-A2 also affected interneurons, we compared calbindin immunolabeling between efnA2 KO and wild type control. We identified a severe reduction in interneurons density within the patches of reduced NeuN+ labeling in efnA2 KO (Fig. 5a–a”, hatched region). Calbindin + interneurons were present in the rest of the neocortex, exhibiting a distribution profile comparable to the wild type control with the majority of the cells populating the supragranular layers 2/3 and fewer cells in the infragranular layers 5/6 (Fig. 5b–b’). However, the density of calbindin+ cells throughout the entire neocortex, excluding the abnormal patches, was significantly reduced (number of cells in layers 2 to 4 per 80,000 µm2; wild type, 135.50 ± 4.96; efnA2 KO, 61.59 ± 14.17, p = 0.029, n = 4, Mann-Whitney test; Fig. 5b”). This suggests that ephrin-A2 plays a major role in the establishment of this fraction of inhibitory neurons but, unlike the focal effect observed in the case of excitatory neurons, loss of ephrin-A2 affects interneurons throughout the entire neocortex suggesting that the ligand has distinct roles in excitatory and inhibitory neurons development.
Ephrin-A2 is expressed by superficially migrating interneurons
At E14.5, interneurons emerging from two transient telencephalic regions, the medial ganglionic eminence (MGE) and the POA (Fig. 6a) and characterized by the expression of GAD 65 & 67 also express ephrin-A2 (Fig. 6a’; boxed region magnified in 6a”’). Ephrin-A2+ cells were organized in a continuous stream extending from the POA to dorsolateral telencephalon, (Fig. 6a”; open arrowheads), correlating with the superficial migratory stream. Interneurons migrating along the superficial migratory stream reach the neocortex at the level of the MZ as illustrated by the dense GAD65 & 67 staining observed in the MZ at E18.5 (Fig. 6b’). Double-labeling revealed the presence of ephrin-A2+ cells in the MZ at this stage, in addition to the dense population in the VZ and scattered throughout the CP observed at earlier stages (Fig. 6b”). Ephrin-A2 expressing cells in the MZ co-express GAD 65 & 67 (Fig. 6c–c”) confirming their interneuronal identity. Altogether, this expression profile demonstrates that interneurons originating from the MGE and the POA upregulate ephrin-A2 expression after exiting the MGE and POA neurogenic zones, characterized by the high nuclei density (Fig. 6a) and the lack of GAD 65 & 67 expression (Fig. 6a’), as they join the superficial migratory stream to reach the neocortex. They maintain the ligand expression after completing the tangential phase of their migration to the MZ and dispersing radially in the CP. The lack of ephrin-A2 expression in the VZ of the 3rd ventricle suggests that the ligand does not participate in progenitor proliferation but might play a role in regulating tangential migration.
Ephrin-A2 promotes the migration of interneurons emerging from the POA
Interneurons emerging from the POA follow a superficial migratory stream along the ventral surface of the brain and are kept segregated from their MGE-born counterparts through EphA4-ephrinB3 mediated bidirectional repulsion27. To elucidate the function of ephrin-A2 in regulating the migration of POA-born interneurons we undertook an ex vivo electroporation approach, enabling us to specifically knockdown ephrin-A2 in the POA (Fig. 7a,b). After 2 days of in vitro culture (2div), we counted the number of GFP+ cells in 100µm-wide bins parallel to the medial edge of the VZ to quantify the extent of cell migration from the site of tissue electroporation (see schematic Fig. 7). Fewer than 50% of cells electroporated with control DNA remained within 100 µm of the medial edge of the VZ of the POA whereas more than 60% of cells electroporated with ephrin-A2 knockdown shRNA remained within this region (scrambled, 0.493 ± 0.054; sh-efnA2 0.629 ± 0.036, n = 15 sections/condition, p = 0.049, Student’s t-test; Fig. 7c). The reduced migration of sh_efnA2 electroporated cells was also reflected by a significant decrease in the proportion of cells that had migrated from 200–300 µm (scrambled, 0.128 ± 0.018; sh-efnA2 0.059 ± 0.01, n = 15, p = 0.0034, Student t-test; Fig. 7c). Overall, our data suggest that knockdown of ephrin-A2 within interneurons emerging form the POA reduces their rate of tangential migration, rather than completely inhibiting their ability to reach the neocortex.
Our study demonstrates that loss of ephrin-A2 during development leads to perturbations in the establishment of neocortical layers, resulting in the formation of discrete, random, patches depleted of neurons. Although Eph/ephrins are traditionally known for regulating migratory events, our results suggest that alteration of the neocortical architecture in the absence of efnA2 results from the abnormal production of neurons before migration to the cortical plate is initiated. Although the lamination defects in efnA2 KO neocortex were spatially restricted to patches, with the neuronal density comparable to control animals in unaffected regions, we uncovered a significant loss of calbindin-expressing neocortical interneurons in apparently ‘normal’ regions of the neocortex, suggesting that in addition to differentiation of pyramidal neurons, ephrin-A2 also contributes to the establishment of neocortical interneurons.
Comparable local disruptions of the neocortical layers observed in efnA2; A3; A5 triple knockout mice24 have been attributed to the abnormal lateral dispersion of excitatory neurons in the cortical plate, without specifying the role of each gene. The presence of ephrin-A2+ neurons in the cortical plate from E14.5 to E18.5, supports the hypothesis that the ligand participates in the regulation of neuronal migration. However, during these stages, ephrin-A2 is predominantly expressed in the VZ suggesting that it has additional functions. The VZ is comprised of Pax6+ apical progenitors9 undergoing mainly asymmetric division (78%), to generate a self-renewed apical progenitor and either a neuron (65.8%) or a basal progenitor. Only 10% of divisions are symmetrical with both daughter cells remaining in the VZ2,35. Accumulation of proliferative Pax6+ cells in the VZ following reduction of efnA2 suggests that a larger proportion of apical progenitors divided according to a symmetrical self-renewing mode and re-entered the cell cycle instead of generating excitatory neurons.
Consistent with these findings, fewer neurons deriving from progenitors in the VZ were present in the postnatal neocortical layers, with dramatic disruption of the laminar architecture observed in extreme cases, replicating the efnA2 KO phenotype. Our results reveal that ephrin-A2 acts as an early cell-fate determinant in apical progenitors, directing their progeny towards a differentiated neuron identity and repressing cell division. The lack of ephrin-A2 signaling promotes cell proliferation and maintains the apical progenitor pool leading to abnormal neuron production and alteration of the neocortical architecture (see Fig. 8).
The signaling elicited by the interaction of ephrin-A2 with an Eph receptor is bidirectional, activating transduction pathways in both ligand and receptor-bearing cells. Although activation of EphA4 forward signaling by intrinsic ephrin-B130 or extrinsic ephrin-A534 promotes apical progenitor proliferation and neuron production in the mouse embryonic neocortex, the addition of clustered ephrin-A2 to dissociated embryonic neocortex cultures had no effect on cell proliferation. Consistent with evidence that ephrin-A2 reverse signaling inhibits the proliferation of adult stem cells in the mouse SVZ36, we observed reduced proliferation following activation of ephrin-A2, demonstrating that ephrin-A2 function on apical progenitors is mediated through reverse signaling. Ephrin-A2 is tethered to the cell membrane via a GPI anchor, preventing it from directly activating intracellular transduction molecules. Instead, activated ephrin-A2 molecules form complexes with other transmembrane molecules and “highjack” their signaling pathway20. The neurotrophin receptors p75 and TrkB, critical for neural progenitors commitment to a neuronal fate by inhibiting cell division and promoting neuronal maturation37,38, are privileged co-receptors for ephrin-A239,40.
While ephrin-A2 reverse signaling is activated by EphA7 in the adult SVZ36, we were not able to identify a unique receptor responsible for eliciting this pathway in the embryonic neocortex with EphA3, A4, A5 and A7 all being strongly expressed in the neocortex. EphA3 is an interesting binding partner candidate for ephrin-A2 expressed on apical progenitors in the neocortical VZ. It has previously been demonstrated that ephrin-A2/EphA3 interaction elicits the recruitment of the metalloproteinase ADAM10 resulting in cleavage of ephrin-A2 from the cell surface41,42. This molecular interaction could also underlie the observed scarcity of ephrin-A2 beyond the VZ in the mouse neocortex.
The significant reduction of calbindin labeling in efnA2 KO neocortex suggests that ephrin-A2 participates in the establishment of neocortical interneurons. Unlike our observations in the neocortex, ephrin-A2 is not expressed in the germinal layers of the regions where interneurons are born, principally the MGE14,15 and the POA. Ephrin-A2 is detected in the superficially migrating interneurons, in agreement with recent demonstration that it promotes the motility of MGE-derived interneurons28. Our new evidence confirms that this role is conserved in POA-born interneurons. However, cell migration is not arrested in absence of ephrin-A2; it is only delayed with interneurons eventually reaching the neocortex, albeit later than normal. Therefore, the reduction in neocortical calbindin labeling is unlikely to correspond to an absence of interneurons due to abnormal migration.
Although the calbindin + population is reduced by 50% across the supergranular layers, the overall neuronal density is normal in efnA2 KO neocortex, beside the patches. Therefore, the decrease in calbindin + cells might not reflect an actual absence of supergranular interneurons but instead abnormal expression of the marker itself. As a member of the calcium-binding molecule family, Calbindin becomes upregulated in mature interneurons to buffer Ca2+ flux43. Interneuron maturation is dependent upon correct synaptogenesis and pruning, which are both under the control of Eph/ephrin signaling44,45. In particular, ephrin-A2 reverse signaling utilizes the TrkB transduction pathway40 known to modulates calbindin upregulation in maturing interneurons46,47. Similarly to efnA2 KO, TrkB deficient animals exhibit a reduction of calbindin + interneurons, consequence of the downregulation of the cell marker and not an actual loss of interneurons. In addition, delayed migration following loss of ephrin-A2 implies that interneurons are exposed to an aberrant, more mature, environment when they disperse through the neocortex. Therefore, the abnormal signals they are exposed to and the atypical network of connections they attempt to integrate are likely to affect the maturation of interneurons lacking ephrin-A2, ultimately affecting calbindin expression.
Thus, the dramatic interneuronal phenotype we report in efnA2 KO results from an impaired execution of maturation programs leading to lower than normal calbindin expression and not an actual depletion of neocortical interneurons.
efnA2 KO mice exhibit an interesting behavioral phenotype characterized by perturbed social interactions and repetitive grooming which reflects aspects of autistic-like behaviors in humans48. Members of the Eph/ephrin family, including EphA6 and EphB2, have been associated with Autism Spectrum Disorder (ASD) in recent genome-wide association studies (GWAS)49,50 but no studies have yet linked ephrin-A2 with the disorder. It could be a consequence of technical limitations as sequences with high GC content (ephrin-A2 GC ratio = 67%) are poorly captured in exome sequencing studies51. The discrete and indiscriminate disruption of neuronal density we observed in efnA2 KO neocortex has, however, been reported in at least one systematic anatomical study of the neocortical architecture of children diagnosed with ASD31. The lamination defects they report in all ASD samples and absent from age-matched controls share many of the characteristics we observe in efnA2 KO mice, including the spatially restricted configuration with sharp borders and the stochastic distribution pattern. While the high degree of variability regarding neocortical areas and layers disrupted in ASD patients could be anticipated considering the large spectrum of symptoms observed in patients and the long list of genes associated with ASD, it is less expected in an animal model bearing a mutation in a single gene.
The phenotypic consistency observed between ASD samples and the efnA2 KO mice suggests a strong link between Eph/ephrin and ASD. Here, we demonstrate a novel mechanism by which ephrin-A2 regulates neurogenesis and determine that disruption of ephrin-A2 activity in the VZ is a potential contributor to ASD. Furthermore, our findings establish a new time-frame for the cellular mechanisms underpinning ASD, suggesting that they can occur as early as the initial stages of neocorticogenesis.
Materials and Methods
All experiments and procedures were conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes and were approved by Monash University and the University of Western Australia Animal Ethics Committees. Animals were housed in cages of 2–5 in a controlled environment with a 12 hour light/dark cycle. Pregnant mice were housed individually 48 h prior to surgery to allow habituation to their environment. Adult efnA2 homozygous knockout (background strain C57bl/6j) brains were obtained from the University of Western Australia. For in utero electroporation and tissue collection, time-mated 8 weeks old pregnant C57bl/6j mice were obtained from local colonies. The plug date was defined as E0.5, and the day of birth was defined as P0. A minimum sample size of n = 3 was used.
In utero electroporation
The enfA2 knockdown construct tagged with GFP was a gift from Professor Jürgen Bolz (see 28 for details). A GFP-tagged scrambled shRNA was used as a control. Mouse ephrin-A2 was amplified using the forward GGGCGGCAGATCTCCAAG and reverse AAAAGTCCACCCCACTCCC primers and the resulting 1564 bp fragment cloned into pCAGS-IRES-GFP. Pregnant mice were assigned a construct using a simple randomization method. Time-mated pregnant females were anesthetized with isoflurane (4% for induction and 2% to maintain) on a heated mat. The abdominal cavity was open and the uterine horns exposed. The plasmid suspension (0.5 µg.µL−1 in PBS) was injected in the lateral ventricle using a glass needle and an electric current (35 V, 50 ms, 950 ms interval ×5 pulses) was delivered using 1 mm platinum tweezertrodes (BTX) connected to EMX BTX (Harvard apparatus). Embryos were harvested 3 or 15 days post electroporation (dpe).
Ex vivo electroporation and slice assay
E14.5 embryonic brains were dissected in Complete HBSS medium (see 52, the two hemispheres separated and the randomly assigned plasmid solution injected directly in the POA followed by application of an electric current (25 V, 50 ms, 250 ms interval ×4 pulses) using 1 mm platinum tweezertrodes (BTX). Electroporated hemispheres were embedded in 4% low-melting point agarose (LMP, Sigma) and sectioned using a vibratome into 300 µm thick coronal slices which were transferred on the membrane of a cell culture insert (Millipore) pre-coated with poly-L-lysine (100 µg.mL−1) and laminin (10 µg.mL−1). Slices were cultured for 48 hours in slice culture medium52 at 35 °C and 5% CO2 after which they were fixed with 4% PFA overnight at 4 °C, rinsed in PBS and mounted on microscope slides using Fluoromount-G mounting medium (SouthernBiotech).
Embryonic neocortices E14.5 were dissociated as previously described53 and seeded (4.104 cells.mL−1) on glass coverslips pre-coated with polyornithine (0.5 mg.mL−1; Sigma) and laminin (20 µg.mL−1; Gibco) in serum-free medium (Neurobasal; 1% B27, KCl 25 mM, 1X Glutamax, Glucose 3 g.L−1, penicillin/streptomycin). Eph and ephrin recombinant protein containing a human Fc fragment or the recombinant human Fc fragment only for control (8 µg.mL−1; R&D) were clustered for 1hr at 37 °C with anti-human Fc antibody (20 µg.mL−1). Pre-clustered recombinant proteins were added to the culture medium with BrdU 0.2 µM. After 24 hours at 37 °C with 5% CO2, cultures were fixed with 4% PFA, stained with mouse anti-TuJ1 (1:1,000; MM3-435P, Covance), incubated with HCl 2 M for 15 min at 37 °C, rinsed with Borate buffer and labeled with rat anti-BrdU (1:20; OBT0030S, AbD Serotec) and nuclei were counterstained with Hoechst. Coverslips were mounted on microscope slides using Fluoromount-G mounting medium (SouthernBiotech). The coverslips were imaged (20 random fields/coverslip; operator blinded to the experiment) using a Zeiss Imager. Z1 microscope with an Axiocam HRm digital camera and Axiovision 4.8.2 software (Zeiss). The images were blinded and the number of BrdU+ cells counted, after applying a grid, using the CellCounter plugin in the software Fiji. The resulting values were unblinded and imported into GraphPad Prism 6.0 for Mac OSX for statistical analysis and graphing. A one-way non-parametric Kruskal-Wallis test was used to test for significance.
Tissue preparation for immunolabeling
Animals received a lethal dose of sodium pentobarbital (100 mg.kg−1) and were transcardially perfused with 0.1 M heparinised sodium phosphate buffer (PBS) containing 0.1% sodium nitrite to optimise vasodilation, followed by 4% paraformaldehyde (PFA) in PBS. Mouse embryos harvested from time-mated females were anesthetized on, quickly decapitated, and the brains dissected. All brains were postfixed in 4% PFA for 24 h at 4 °C. Electroporated embryonic brains were embedded in 3% LMP and 100 µm sections cut using a Vibratome (Leica). Following post-fixation, wild type (non-electroporated) embryonic brains and all postnatal brains were cryoprotected in PBS- 20% sucrose, embedded in optimal cutting temperature (OCT) medium, frozen in −45 °C isopentane and stored at −80 °C. Coronal cryostat sections of embryonic brains (18 µm) were collected on Superfrost slides and stored at −20 °C. Free floating sections (40 µm; coronal plane) were collected in serial order and stored at −20 °C in a cryoprotectant solution (50% 0.05 M phosphate buffer, 30% ethylene glycol, 20% glycerol).
Each staining was repeated at least 3 times independently. The representative images included represent results that were consistently observed. Free-floating sections were washed in PBS, while slide-mounted sections were rehydrated in PBS, before being incubated for 1 hour at room temperature with the blocking solution (10% Normal Goat Serum, 0.3% Triton X-100 in PBS). Due to low sensitivity, sections labeled with rabbit Anti-ALDH1L1 were incubated for 2 min in citrate buffer at 90 °C for antigen retrieval. The following primary antibodies were incubated overnight at 4 °C in the blocking solution: rabbit anti-ALDH1L1 (1:500; ab87117, Abcam), rat anti-BrdU (1:20; OBT0030S, AbD Serotec), mouse anti-calbindin (1:1,000; CB38, Swant), goat anti-Doublecortin (1:200; ab18723, Abcam), rabbit anti-ephrin-A2 (1:300; SC-912, Santa Cruz), mouse anti-GAD65 (1:500; ab26113, Abcam), mouse anti-GAD67 (1:500; MAB5406, Merck-Millipore), chicken anti-GFP (1: 1,000; ab13970, Abcam), rabbit anti-Iba1 (1:500; 019–19741, Wako), rabbit anti-Ki67 (1:1,000; ab16667, Abcam), rabbit anti-Nestin (1:200, ab27952, Abcam), mouse anti-NeuN (1:500; MAB377, Merck-Millipore), rabbit anti-NeuN (1:1,000; ABN78, Merck/Millipore), mouse anti-NNF (1:1,000; SMI32R, Covance), mouse anti-Olig2 (1:500; MABN50, Merck/Millipore), mouse anti-Parvalbumin (1:1,000; PV235, Swant), rabbit anti-Pax6 (1:200; 901301, Covance), mouse anti-Satb2 (1:200; ab51502, Abcam), mouse anti-Sox2 (1:200; 3579, Cell Signaling), rabbit anti-Tbr1 (1:200; ab31940, Abcam), mouse anti-TuJ1 (1:1,000; MM3–435P, Covance), chicken anti-Vimentin (1:200; AB_528505, DSHB). Sections were washed in 0.1% Tween-20 in PBS before incubation with the appropriate fluorescently conjugated antibodies Alexa Fluor 594 and 488 (1:1,000; Molecular Probes). Fluorescently labeled sections were subsequently incubated with Hoechst solution (Dako) to stain cell nuclei, washed in PBS, free floating sections were mounted onto Superfrost slides and coverslipped with Fluoromount-G mounting medium (SouthernBiotech).
In situ hybridisation
Digoxygenin-labelled riboprobe for murine efnA2 (NM_007909.3; nucleotides 158–1002) cloned into pGEMT was used as previously described54.
efnA2 KO cell quantification
The experimenter was blinded to the genotype of the animals until after all the counting were completed. Low-magnification photomicrograph (1300 × 1030 dpi) of calbindin and NeuN stained sections of adult wild type and efnA2 −/− brains were acquired using a Zeiss Imager. Z1 microscope with an Axiocam HRm digital camera and Axiovision 4.8.2 software (Zeiss). Labeled cells were counted using the Fiji CellCounter plugin after applying a 10,000 µm2 grid. 4 squares spanning the neocortical thickness were counted, 3 rows/images, 160 µm interval between sections, n = 3 for NeuN counting and n = 4 for calbindin. Statistical analysis and graphing of the data was performed using GraphPad Prism 6.0 for Mac OSX. The one-way non-parametric Mann-Whitney U test was used to compare two experimental conditions, for which the normal distribution could not be assumed.
In utero electroporation cell quantification
Coronal sections from electroporated brains collected at E18.5 (3 dpe) were incubated in Hoechst, mounted on glass microscope slides and coverslipped using Fluoromount mounting medium. Electroporated animals collected at P12 (15 dpe) were processed for immunolabeling, including GFP, as described above. For quantification of the number of GFP+ cells per compartment, low magnification images were captured to include the electroporated region from inner to outer surface, using a Zeiss Imager. Z1 microscope with an Axiocam HRm digital camera and Axiovision 4.8.2 software (Zeiss). Images were blinded using a script available in the free image analysis software Fiji. The entire electroporated region was counted, using a grid to facilitate the process. An average of 7.5 images/animal (between 5 and 10) was analyzed for E15.5+ 3 dpe experiments; 4 to 5 images were analyzed for E15.5+ 15 dpe. For the quantification of proliferative cells at E15.5+ 3 dpe, sections were imaged at higher magnification, Z-stacks were captured using an ApoTome and converted to maximum intensity projection. The resulting images were then processed as described above; with an average of 4 images/animal analyzed. The data were unblinded prior to statistical analysis and graphing using GraphPad Prism 6.0 for Mac OSX. The one-way non-parametric Mann-Whitney and Kruskal-Wallis tests were used to compare two experimental conditions and more than two experimental conditions, respectively.
Slice assay cell quantification
Sections exhibiting extensive pyknosis, suggestive of cell death, revealed by Hoechst staining and sections where the electroporation was not restricted to the POA were systematically excluded from the analysis. Z-stacks of the electroporated regions were acquired with a Leica SP5 multichannel confocal microscope running the software LAS AF Lite. Fiji was used to generate maximum intensity projections. Using the freehand tracing tool in Illustrator (CS3), the outline of the electroporated region of the ventricular zone was drawn, defining the origin of migration (0 µm) and parallel lines were pasted at 100 µm interval. The images were saved in a tiff format. The number of GFP+ cells in each bin was counted using the CellCounter plugin in Fiji after applying a grid. The resulting counts were imported into Excel, unblinded and imported into GraphPad Prism 6.0 for Mac OSX for statistical analysis (unpaired Student’s t-test) and graphing.
All relevant data are available from the authors.
Marín, O. & Rubenstein, J. L. R. Cell migration in the forebrain. Annu Rev Neurosci 26, 441–483 (2003).
Noctor, S. C., Martínez-Cerdeño, V., Ivic, L. & Kriegstein, A. R. Cortical neurons arise in symmetric and asymmetric division zones and migrate through specific phases. Nat Neurosci 7, 136–144 (2004).
Marín, O. & Rubenstein, J. L. A long, remarkable journey: tangential migration in the telencephalon. Nat. Rev. Neurosci. 2, 780–790 (2001).
Egger, B., Gold, K. S. & Brand, A. H. Notch regulates the switch from symmetric to asymmetric neural stem cell division in the Drosophila optic lobe. Development 137, 2981–2987 (2010).
Lah, G. J. & Key, B. Dual roles of the chemorepellent axon guidance molecule RGMa in establishing pioneering axon tracts and neural fate decisions in embryonic vertebrate forebrain. Devel Neurobio 72, 1458–1470 (2012).
Kaushik, G. & Zarbalis, K. S. Prenatal Neurogenesis in Autism Spectrum Disorders. Frontiers in Chemistry 4, 12 (2016).
Arnold, S. J. et al. The T-box transcription factor Eomes/Tbr2 regulates neurogenesis in the cortical subventricular zone. Genes Dev 22, 2479–2484 (2008).
Vasistha, N. A. et al. Cortical and Clonal Contribution of Tbr2 Expressing Progenitors in the Developing Mouse Brain. Cerebral Cortex bhu125, doi:https://doi.org/10.1093/cercor/bhu125 (2014).
Englund, C. et al. Pax6, Tbr2, and Tbr1 are expressed sequentially by radial glia, intermediate progenitor cells, and postmitotic neurons in developing neocortex. J Neurosci 25, 247–251 (2005).
Kriegstein, A. R. & Noctor, S. C. Patterns of neuronal migration in the embryonic cortex. Trends Neurosci 27, 392–399 (2004).
Rakic, P. Specification of cerebral cortical areas. Science 241, 170–176 (1988).
Marín, O. Cellular and molecular mechanisms controlling the migration of neocortical interneurons. Eur J Neurosci 38, 2019–2029 (2013).
Anderson, S. A., Eisenstat, D. D., Shi, L. & Rubenstein, J. L. Interneuron migration from basal forebrain to neocortex: dependence on Dlx genes. Science 278, 474–476 (1997).
Gelman, D. M. et al. The embryonic preoptic area is a novel source of cortical GABAergic interneurons. J Neurosci 29, 9380–9389 (2009).
Gelman, D. et al. A wide diversity of cortical GABAergic interneurons derives from the embryonic preoptic area. J Neurosci 31, 16570–16580 (2011).
Kanatani, S. et al. The COUP-TFII/Neuropilin-2 is a molecular switch steering diencephalon-derived GABAergic neurons in the developing mouse brain. Proc Natl Acad Sci USA 112, E4985–94 (2015).
Butt, S. J. B. et al. The temporal and spatial origins of cortical interneurons predict their physiological subtype. Neuron 48, 591–604 (2005).
Rudy, B., Fishell, G., Lee, S. & Hjerling-Leffler, J. Three groups of interneurons account for nearly 100% of neocortical GABAergic neurons. Devel Neurobio 71, 45–61 (2011).
Miyoshi, G. et al. Prox1 Regulates the Subtype-Specific Development of Caudal Ganglionic Eminence-Derived GABAergic Cortical Interneurons. J Neurosci 35, 12869–12889 (2015).
Klein, R. Eph/ephrin signalling during development. Development 139, 4105–4109 (2012).
Batlle, E. & Wilkinson, D. G. Molecular mechanisms of cell segregation and boundary formation in development and tumorigenesis. Cold Spring Harb Perspect Biol 4, a008227–a008227 (2012).
Egea, J. & Klein, R. Bidirectional Eph-ephrin signaling during axon guidance. Trends Cell Biol. 17, 230–238 (2007).
Klein, R. Bidirectional modulation of synaptic functions by Eph/ephrin signaling. Nature Publishing Group 12, 15–20 (2009).
Torii, M., Hashimoto-Torii, K., Levitt, P. & Rakic, P. Integration of neuronal clones in the radial cortical columns by EphA and ephrin-A signalling. Nature 461, 524–528 (2009).
Dimidschstein, J. et al. Ephrin-b1 controls the columnar distribution of cortical pyramidal neurons by restricting their tangential migration. Neuron 79, 1123–1135 (2013).
Zimmer, G. et al. Ephrin-A5 acts as a repulsive cue for migrating cortical interneurons. Eur J Neurosci 28, 62–73 (2008).
Zimmer, G. et al. Bidirectional ephrinB3/EphA4 signaling mediates the segregation of medial ganglionic eminence- and preoptic area-derived interneurons in the deep and superficial migratory stream. J Neurosci 31, 18364–18380 (2011).
Steinecke, A., Gampe, C., Zimmer, G., Rudolph, J. & Bolz, J. EphA/ephrin A reverse signaling promotes the migration of cortical interneurons from the medial ganglionic eminence. Development 141, 460–471 (2014).
Depaepe, V. et al. Ephrin signalling controls brain size by regulating apoptosis of neural progenitors. Nature 435, 1244–1250 (2005).
North, H. A. et al. Promotion of proliferation in the developing cerebral cortex by EphA4 forward signaling. Development 136, 2467–2476 (2009).
Stoner, R. et al. Patches of disorganization in the neocortex of children with autism. N. Engl. J. Med. 370, 1209–1219 (2014).
Quinn, J. C. et al. Pax6 controls cerebral cortical cell number by regulating exit from the cell cycle and specifies cortical cell identity by a cell autonomous mechanism. Developmental Biology 302, 50–65 (2007).
Bonanomi, D. et al. Ret Is a Multifunctional Coreceptor that Integrates Diffusible- and Contact-Axon Guidance Signals. Cell 148, 568–582 (2012).
Gerstmann, K. et al. Thalamic afferents influence cortical progenitors via ephrin A5-EphA4 interactions. Development 142, 140–150 (2015).
Götz, M. & Huttner, W. B. The cell biology of neurogenesis. Nat Rev Mol Cell Biol 6, 777–788 (2005).
Holmberg, J. et al. Ephrin-A2 reverse signaling negatively regulates neural progenitor proliferation and neurogenesis. Genes Dev 19, 462–471 (2005).
Verdi, J. M. et al. p75LNGFR regulates Trk signal transduction and NGF-induced neuronal differentiation in MAH cells. Neuron 12, 733–745 (1994).
Hosomi, S., Yamashita, T., Aoki, M. & Tohyama, M. The p75 receptor is required for BDNF-induced differentiation of neural precursor cells. Biochem. Biophys. Res. Commun. 301, 1011–1015 (2003).
Lim, Y.-S. et al. p75(NTR) mediates ephrin-A reverse signaling required for axon repulsion and mapping. Neuron 59, 746–758 (2008).
Marler, K. J. M. et al. A TrkB/EphrinA interaction controls retinal axon branching and synaptogenesis. J Neurosci 28, 12700–12712 (2008).
Hattori, M., Osterfield, M. & Flanagan, J. G. Regulated cleavage of a contact-mediated axon repellent. Science 289, 1360–1365 (2000).
Murai, K. K. & Pasquale, E. B. ‘Eph’ective signaling: forward, reverse and crosstalk. J Cell Sci 116, 2823–2832 (2003).
Letinic, K. & Kostović, I. Postnatal development of calcium-binding proteins calbindin and parvalbumin in human visual cortex. Cereb Cortex 8, 660–669 (1998).
Brennaman, L. H. et al. Polysialylated NCAM and EphrinA/EphA Regulate Synaptic Development of GABAergic Interneurons in Prefrontal Cortex. Cereb Cortex, doi:https://doi.org/10.1093/cercor/bhr392 (2012).
Yu, X. et al. Accelerated experience-dependent pruning of cortical synapses in ephrin-A2 knockout mice. Neuron 80, 64–71 (2013).
Jones, K. R., Fariñas, I., Backus, C. & Reichardt, L. F. Targeted disruption of the BDNF gene perturbs brain and sensory neuron development but not motor neuron development. Cell 76, 989–999 (1994).
Fiumelli, H., Kiraly, M., Ambrus, A., Magistretti, P. J. & Martin, J. L. Opposite regulation of calbindin and calretinin expression by brain-derived neurotrophic factor in cortical neurons. J Neurochem 74, 1870–1877 (2000).
Wurzman, R., Forcelli, P. A., Griffey, C. J. & Kromer, L. F. Repetitive grooming and sensorimotor abnormalities in an ephrin-A knockout model for Autism Spectrum Disorders. Behavioural brain research 278C, 115–128 (2014).
Pinto, D. et al. Functional impact of global rare copy number variation in autism spectrum disorders. Nature 466, 368–372 (2010).
Krishnan, A. et al. Genome-wide prediction and functional characterization of the genetic basis of autism spectrum disorder. Nat Neurosci, doi:https://doi.org/10.1038/nn.4353 (2016).
Huguet, G., Ey, E. & Bourgeron, T. The genetic landscapes of autism spectrum disorders. Annu Rev Genomics Hum Genet 14, 191–213 (2013).
Polleux, F. & Ghosh, A. The slice overlay assay: a versatile tool to study the influence of extracellular signals on neuronal development. Sci. STKE 2002, pl9 (2002).
Homman-Ludiye, J., Merson, T. D. & Bourne, J. A. The early postnatal nonhuman primate neocortex contains self-renewing multipotent neural progenitor cells. PLoS ONE 7, e34383 (2012).
Homman-Ludiye, J. & Bourne, J. A. The guidance molecule Semaphorin3A is differentially involved in the arealization of the mouse and primate neocortex. Cerebral Cortex 24, 2884–2898 (2014).
The authors wish to acknowledge and thank Tobias Merson and Cornelius Gross for their useful comments on earlier versions of the article, Robert Bryson-Richardson for help with image analysis and Frederico Calhabeu for help with confocal imaging. National Health and Medical Research Council (NHMRC) Senior Research Fellowship support JAB (APP1077677) and JR (APP1002258) and Project Grant (APP1002049). Stem Cells Australia supports JHL. The Australian Regenerative Medicine Institute is supported by grants from the State Government of Victoria and the Australian Government.
The authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
About this article
Cite this article
Homman-Ludiye, J., Kwan, W.C., de Souza, M.J. et al. Ephrin-A2 regulates excitatory neuron differentiation and interneuron migration in the developing neocortex. Sci Rep 7, 11813 (2017). https://doi.org/10.1038/s41598-017-12185-x
Autism and Williams syndrome: Dissimilar socio-cognitive profiles with similar patterns of abnormal gene expression in the blood
Biomedicine & Pharmacotherapy (2020)
Pediatric Research (2019)
Identification of genes associated with cortical malformation using a transposon-mediated somatic mutagenesis screen in mice
Nature Communications (2018)