Laminar Expression of Ephrin-A2 in Primary Somatosensory Cortex of Postnatal Rats



Several Eph receptors, prominently EphA4 and EphA7, and their corresponding ligands are known to influence neocortical development, including topographic sorting of thalamocortical axons within primary somatosensory cortex (SI). This study investigated postnatal expression of a ligand that can bind to these receptors, ephrin-A2. Quantitative methods revealed that expression of ephrin-A2 mRNA in SI reached maximum levels on postnatal day (P) 4 and dropped thereafter to background by P18. Ephrin-A2 mRNA expression assessed by in situ hybridization qualitatively revealed a similar time course and localized the expression pattern primarily in two broad laminae in SI, comprising the supragranular and infragranular layers, and with additional expression in the subplate. This expression pattern was investigated in greater detail using immunohistochemistry for ephrin-A2 protein. Immunoreactivity generally showed the same laminar distribution as seen with in situ hybridization, except that it persisted longer, lasting to approximately P14. Expression in the cortical plate was low or absent within presumptive layer IV, and it remained so as cortical lamination progressed. Double-labeling immunohistochemistry with confocal microscopy revealed that cortical neurons were the principal elements expressing ephrin-A2 protein. These findings are consistent with possible involvement of ephrin-A2, in concert with one or more Eph receptors, in influencing arbor development of thalamocortical axons at cortical layer IV boundaries. Anat Rec, 2012. © 2011 Wiley Periodicals, Inc.


Neurons in rodent primary somatosensory cortex (SI) form a sensory map of the body surface. This representation is tangential to the cortical surface and parcellated into numerous discrete somatotopic regions, the “barrels,” which are especially evident within one major subdivision, the posteromedial barrel subfield (PMBSF; Woolsey and Van der Loos,1970). Orthogonal to this organization are columns of cell bodies, dendrites, and axons involved in layer-specific intrinsic signaling and input–output connections to and from barrels (Lund and Mustari,1977; Agmon et al.,1993). Cortical barrels develop in rats and mice close to the time of birth when thalamocortical axons (TCAs) from the ventrobasal (VB) thalamus emerge from the internal capsule, grow through the intermediate zone and subplate, and project into the cortical plate. During the first postnatal week, cortical layer IV differentiates from the cortical plate, and TCAs expand terminal arbors in a barrel pattern within this layer (reviewed by Killackey et al.,1995). The precise targeting of TCAs onto cortical neurons in certain layers, but not others, clearly requires multiple factors guiding their development.

Several members of the Eph family of receptor tyrosine kinases have been variously implicated in neocortical development (e.g., Flanagan and Vanderhaeghen,1998; Bolz et al.,2004; Price et al.,2006). Complementary gradients of ligand ephrin-A5 on cortical cells and EphA4 or EphA7 on TCA growth cones operate via mutually repulsive interactions to guide cortical pattern formation in SI as well as in primary visual cortex (Prakash et al.,2000; Vanderhaeghen et al.,2000; Sestan et al.,2001; Bishop et al.,2002; Dufour et al.,2003; Cang et al.,2005). However, additional factors may also be required. For tangential gradients of repulsive molecules to be effective, the ligand and receptor bearing processes might require some laminar confinement to assure that the fibers actually track within the plane of the gradients implicated in map formation. Otherwise, elongating axons may end-run above or below the crucial interaction zone. The importance of confining molecular boundaries is evident in axon sorting patterns made within in vitro systems, such as stripe assays (e.g., Castellani and Bolz,1997). Previous reports by Yamamoto and coworkers suggest that molecules with properties resembling ephrin-A ligands would be suitable for confining TCAs within cortical layer IV (Yamamoto et al.,1997,2000).

We report here qualitative and quantitative findings that cortical neurons in supragranular and infragranular layers of SI express ephrin-A2 mRNA as well as protein during the first two postnatal weeks. Over this period, ephrin-A2 expression is relatively low within layer IV where the TCAs are proliferating terminal branches. This molecule is thus a potential candidate to restrict radial growth of TCA terminals to their correct layer.

Abbreviations used: AChE = acetylcholinesterase; CO = cytochrome oxidase; DEPC = Diethyl Pyrocarbonate; GPI = glycosylphosphatidylinositol; GAPDH = glyceraldehyde 3-phosphate dehydrogenase; P = postnatal day with P0 = the date of birth; PMBSF = posteromedial barrel subfield; RTPCR = real time polymerase chain reaction; SERT = serotonin uptake transporter; SI = primary somatosensory cortex; TCA(s) = thalamocortical axons or their terminals; VB = ventrobasal thalamic nuclei (i.e., the ventromedial and ventrolateral posterior nuclei).



Sprague Dawley breeder rats were obtained from Taconic Farms (Hudson, New York) and bred at the University of Toledo. At predefined ages from P0 to P21, rat pups were euthanized by CO2 asphyxiation followed by decapitation. All experiments were conducted in accordance with guidelines set forth in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and with review by the Institutional Animal Care and Use Committee of the University of Toledo.


Ephrin-A2 protein expression was characterized using a polyclonal antibody (sc-912, Santa Cruz Biotechnology) developed in rabbit from the C-terminus of an ephrin-A2 peptide sequence of mouse origin. The sensitivity and specificity of this antibody in the peripheral and central nervous system of various species has been previously demonstrated (Davenport et al.,1998; Matsunaga et al.,2000; Lai et al.,2001; King et al.,2004; Lee and Warcho,2005; Rodger et al.,2005). As a positive control, olfactory bulb sections were processed in parallel with cortical sections (Supporting Information Fig. S1). The olfactory bulb was selected for evaluation due to its robust and precise expression of ephrin family proteins, including ephrin-A2, in neonatal rodents (Zhang et al.,1996; St John et al.,2002; Cutforth et al.,2003; Deschamps et al.,2010). Negative controls for each experiment included alternate cortical, thalamic, and olfactory bulb sections processed using either no primary antibody, a host-specific immunoglobulin in place of the primary antibody, or primary antibody adsorbed with the antigenic peptide.

For immunohistochemistry of ephrin-A2 protein, rats were euthanized and transcardially perfused with heparinized saline followed by 4% paraformaldehyde. Brains were removed and placed in 4% paraformaldehyde at 4°C overnight and then frozen and sectioned (50 μm) in the coronal plane or horizontally after flattening each hemisphere. Sections were treated with 3% H2O2, blocked with BSA (Sigma), permeabilized with Triton-X (Sigma) and incubated overnight in 1:1,000 ephrin-A2 primary antibody. They were next washed and then incubated for 2 hr in 1:450 biotinylated secondary antibody (Chemicon); washed and incubated with Vectastain avidin:biotin enzyme complex (Vector Labs) for two additional hours; washed again and then reacted with diaminobenzoate (Fluka).

For fluorescent, double-label immunohistochemistry, tissue was collected and treated as described earlier, except that 1:500 AlexaFluor -488 and -555 labeled-secondary antibodies (Invitrogen) were used in conjunction with the primary antibody for ephrin-A2 and with either an antibody against Hu RNA binding protein (1:40,000 dilution, a gift of Dr. Marthe Howard, University of Toledo) or an antibody for the serotonin transporter (SERT; 1:1,000, Millipore). Hu RNA binding protein is specific for neurons and ideal for localization of cortical neurons at all stages of maturation from early postmitotic neurons in the ventricular zone to mature neurons in the cortex (Okano and Darnell,1997). Antibodies to SERT label TCAs selectively from VB thalamus, and not those from the posterior nucleus or zona incerta (Lebrand et al.,1998). The double-labeled sections were imaged by a Leica SP5 Confocal microscope equipped with Argon-488 and diode pumped solid state-561 and -633 laser sources and providing continuously variable laser output gain and channel-specific spectral bandwidth collection settings. Crosstalk between channels was minimized or eliminated by empirically adjusting the laser output attenuation and the spectral bandwidth of emission filters so that no detectable signal was collected from the 488 line using the longer wavelength emission filter and no visible signal from the 561 line was collected using the shorter emission filter. The resultant bandpass settings were 500–535 and 570–700 nm, respectively. Thus, signal from Hu, labeled using Alexa Fluor 555-conjugated secondary antibody, contributed no more than a few percent to the total signal when using 488 nm excitation matched to the chromaphore for ephrin-A2. Resulting images were visualized using Leica LAS software and analyzed using Image J software (National Institutes of Health). A colocalization analysis of double-labeled immunoreactive structures was performed using a 63.0 × 1.40 N.A. oil immersion objective and imaged with 512 × 512 pixel spatial arrays and a resolution of 34 pixels/μm × 8-bits pixel depth for the two-excitation wavelengths and expressed as Pearson product correlations (R2).

In Situ Hybridization

For in situ hybridization experiments, rats were euthanized, transcardially perfused using diethyl pyrocarbonate (DEPC)-treated equipment and solutions. The brains were removed and placed into 4% paraformaldehyde at 4°C overnight and cryoprotected in 30% sucrose for 48 hr; and then sectioned at 20 μm onto glass slides using a Leica CM3050S Cryostat. The sections were then dried, fixed, permeabilized with proteinase K, and hybridized with digoxigenin-labeled ephrin-A2 antisense at 60°C overnight (plasmids obtained as gift from Dr. David Feldheim, University of California Santa Cruz). The sections were washed, then blocked and incubated overnight in 1:2,000 antidigoxigenin antibody (Roche), and the following day reacted with NBT/BCIP. Slides were placed in fixative, then coverslipped, and imaged. Each antisense probe was tested for specificity by hybridization of dot blots including sense strands for either ephrin-A2, ephrin-A3, ephrin-A5, EphA4, or EphA7. No hybridization to probes other than ephrin-A2 was observed.

Quantitative Real-Time Polymerase Chain Reaction

Quantitative real-time polymerase chain reaction (RTPCR) experiments were performed to measure levels of ephrin-A2 mRNA in the PMBSF of SI cortex. After rats were decapitated, the brains were quickly removed and cortices were detached and flattened. From one hemisphere, a tissue block containing the PMBSF through the full cortical thickness was dissected and placed into RNAqueous (Ambion). As no visible markers are apparent in the fresh cortex to define the PMBSF, it was dissected blind, based on extensive experience, and the accuracy of this procedure was subsequently confirmed in every preparation used. To enable histologic verification of the sampling site, the remainder of the hemisphere, as well as the opposite, intact hemisphere, was placed into 4% paraformaldehyde for fixation and then reacted for cytochrome oxidase (CO), as described previously (Wong-Riley,1979). Histologically confirmed PMBSF tissue blocks were homogenized, mRNA isolated, and reverse transcribed. The resulting cDNAs were quantified on a NanoDrop and then frozen at −20°C. For RTPCR, cDNA for ephrin-A2 was analyzed simultaneously with that for glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as an endogenous control, using the Applied Biosystems 7500 fast system along with their ephrin-A2 and GAPDH primer-probe sets. Levels (measured as ΔΔCt values) of ephrin-A2 from RTPCR were first normalized to GAPDH levels and then expressed for each age relative to the value at P18.


Ephrin-A2 mRNA Expression in Sensory Cortex

The time course of ephrin-A2 mRNA expression within the PMBSF was assessed using quantitative RTPCR analyses. For these measurements, blocks of tissue containing the PMBSF were excised from the cortices of rats on postnatal day (P) 2, and also P4, P6, P8, and P18. The remaining portions of these cortices were then reacted for CO, which reveals barrel patterns, to confirm the localization accuracy of samples used for analysis (Fig. 1A,B). Ephrin-A2 mRNA levels were maximal early during the first postnatal week, followed by a rapid reduction, dropping 10–20-fold lower by P18 (Fig. 1C).

Figure 1.

Quantitative determination and laminar distribution of neonatal ephrin-A2 mRNA expression within SI. A section of cortex, stained with CO is shown after excision of the tissue sample used for the analysis by RTPCR (A). The contralateral cortical hemisphere (B) is shown for comparison, where the white box outlines the area of CO-labeled barrels in the PMBSF. The graph shows expression levels of ephrin-A2 mRNA in the PMBSF tissue blocks (C). Values are shown as fold changes relative to P18, when ephrin-A2 expression was negligible, and have been normalized to the expression levels of the stably expressed “housekeeping gene” for GAPDH. Error bars denote standard deviations. In situ hybridization revealed expression of ephrin-A2 mRNA in separate bands (white arrows) within the developing cortical plate (CP) at P0 (D). At P4, expression was seen in layers II/III and V as well as in the subplate (SP; white brackets) but remained at background levels in layer IV (E). By P8, expression of ephrin-A2 mRNA was down-regulated in SP and all cortical layers (F). As a control observation, no message for ephrin-A2 was evident in a section hybridized with sense primer (G). Scale bars in B and G represent 1 mm. Abbreviations: CO, cytochrome oxidase; CP, cortical plate; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; P, Postnatal day (P0 = the day of birth); PMBSF, posteromedial barrel subfield; RTPCR, real time polymerase chain reaction; SI, primary somatosensory cortex; SP, subplate.

Expression of ephrin-A2 mRNA in SI was also investigated using in situ hybridization. At P0, ephrin-A2 mRNA expression was evident in the subplate and developing cortical plate (Fig. 1D). By P4, the hybridization signal for ephrin-A2 had further increased in the subplate and in two separate bands in the nascent supragranular and infragranular layers of the cortical plate (Fig. 1E). Consistent with the RTPCR analysis, ephrin-A2 throughout all these structures had down-regulated by P8 (Fig. 1F)

Laminar Expression of Ephrin-A2 Protein

Immunohistochemistry was used to detect whether ephrin-A2 protein was expressed in a similar laminar pattern and timing sequence as its mRNA counterpart. As early as P0, laminar expression of ephrin-A2 was observed in SI, with maximal immunoreactivity seen in the subplate and in the developing cortical plate (Fig. 2A). With further differentiation of the cortical plate, two bands of ephrin-A2 expression could be localized respectively to the supragranular and infragranular layers (Fig. 2B,C). Peak labeling density in the supragranular layers appeared to be attained at P8–P10 (Fig. 2F). Throughout the first two postnatal weeks, ephrin-A2 protein expression stayed at background levels within layer IV itself, as identified by the barrel pattern revealed in adjacent sections reacted for CO (Fig. 2E) or stained for Nissl-substance (not shown). Ephrin-A2 protein expression in the cortex began to decline by P12 and was at or near background levels by P14 (Fig. 2G,H).

Figure 2.

Cortical expression of ephrin-A2 protein shown in photomicrographs of immunoreacted sections over postnatal weeks 1 and 2, with higher magnification views in insets. Ephrin-A2 immunoreactivity was observed in SI as early as P0 in the SP (arrows) and in the CP (A, bracket). Over the first postnatal week, ephrin-A2 expression was qualitatively elevated in developing layers II/III, in layer V, and in the SP (B and C). For laminar localization, a section adjacent to that taken at P8 (F) was reacted for CO to demonstrate the extent of layer IV (E, several barrels are indicated by white under-brackets). Controls for testing specificity of antibody labeling included sections processed identically to those in other panels but with the primary antibody to ephrin-A2 replaced with normal serum immunoglobulins (NS; D and left inset) or with the primary antibody adsorbed with its antigenic peptide (Ab + Ag; right inset of D). Scale bar in H represents 1 mm for all low-power views and 0.5 mm for insets A–C and 0.7 mm for insets D–F, G, and H. Abbreviations, as in Fig. 1 plus: insets within D Ab + Ag, antibody plus blocking peptide; NS, normal serum immunoglobulins.

Cellular Localization of Ephrin-A2 Protein

To demonstrate which cellular elements in the cortex were involved in expression of ephrin-A2 protein, cortical sections through SI from P8 rats were double-labeled via immunohistochemistry for ephrin-A2 protein and Hu RNA binding protein, an age-independent, and neuron-specific cell marker. Confocal microscopic images, centered on the infragranular layers, revealed that ephrin-A2 protein and Hu were significantly colocalized (pixel R2 = 0.85–0.88) within neurons of varied morphology (Fig. 3). Using the same strategy, double-labeled brain sections for ephrin-A2 together with SERT, a marker for early postnatal TCAs originating in the VB thalamic nuclei, failed to demonstrate significant colocalization of these proteins (R2 = 0.2–0.3, data not shown). These results are consistent with expression of ephrin-A2 protein by resident cortical neurons, although they cannot completely exclude the possibility of expression on cortically projecting axons or glia.

Figure 3.

Confocal microscopy was used to determine whether ephrin-A2 protein expression localized within neurons. Cells from layer V in PMBSF of P8 cortices are shown following double immunolabeling for ephrin-A2 protein (green fluorescing secondary antibody, panels A and D) and for a neuron-specific cell marker, Hu RNA binding protein (red secondary antibody, B and E), and in merged images of both labels (C and F, respectively). Scale bar represents 50 μm (A–C) and 5 μm (D–F).


This study provides three lines of evidence that ephrin-A2, a glycosylphosphatidylinositol (GPI)-linked guidance molecule, is transiently expressed within cortical neurons of SI in early postnatal rats. Specifically, immunohistochemical localization demonstrated that the ephrin-A2 protein was expressed during the first two postnatal weeks in neurons within the supragranular and infragranular layers as well as in the subplate, but not within layer IV or its anlage in the cortical plate. Ephrin-A2 mRNA was also expressed in the same strata, indicating that the cortical neurons mainly in layers II/III and V, as well as in the subplate, are the source of the protein expression. No spatial gradients parallel to the cortical surface were noted for ephrin-A2 protein or mRNA. Ephrin-A2 mRNA, extracted from PMBSF samples and analyzed by RTPCR, increased during the first four postnatal days and then sharply declined, as much as 20-fold by P18. Thus, the specific temporal and laminar expression patterns of ephrin-A2 are consistent with the suggestion that this molecule may play a transient role in cortical development during the period of TCA arbor formation.

Comparison With Previous Studies

In situ hybridization for cortical ephrin-A2 has previously been performed in postnatal mice (Cang et al.,2005; Peuckert et al.,2008), and our results show both similarities and differences. The similarities are that ephrin-A2 expression in mouse visual cortex and in SI up as a band in the supragranular layers, with little hybridization signal seen in layer IV, other than at its superficial border. Also, SI in mouse showed no obvious tangential gradients of ephrin-A2 expression (Peuckert et al.,2008), although this differed from visual cortex in the same species (Cang et al.,2005). The features present in rat, but lacking in mouse sensory cortex, are the bands found in the infragranular layers and in the subplate. Interestingly, in ephrin-A5 knockout mice on C57BL6 background, there was up-regulation of ephrin-A2 in layers II/III and uppermost layer IV and, in addition, expression in the subplate, although none appeared in layers V/VI (Peuckert et al.,2008). Analysis of protein expression was not undertaken in these two studies.

Although many developmental features of barrel cortex are similar in rats and mice, the two species are known to differ slightly in laminar organization and timing. In mice, for example, acetylcholinesterase (AChE) activity associated with TCA ingrowth and arborization in SI forms a continuous band in upper layer V preceding its segmented appearance in layer IV as the TCA arbors develop. In contrast, rat cortex reveals little AChE activity in layer V until well after its appearance in the layer IV barrels (Sendemir et al.,1996). The molecular factors that differentially regulate AChE expression by TCAs at the layer IV/V boundary in SI of rats and mice are not presently known, but these findings provide an example in which molecular involvement at this border may differ across rodent species.

Ephrin-A2 as a Laminar Guidance Cue?

In rat and mouse SI, TCAs form terminal arbors in cortical layer IV without undergoing a phase of initial overgrowth and back-pruning (Killackey et al.,1995). Molnár and Blakemore (1991) have previously hypothesized that a repulsive signal within the cortex could be necessary for confining TCA termination to layer IV during cortical development. In vitro, coculture models involving TCAs from VB forming projections into cortical explants suggested that the molecular signals guiding this process were associated with neuronal membranes or the extracellular matrix (Yamamoto et al.,1992). The ephrin-A family of guidance molecules were viewed as ideal candidates for TCA arbor confinement to layer IV, that is, a “stop signal,” by Yamamoto and coworkers who showed that TCAs would continue their growth past layer IV and into supragranular layers following cleavage of GPI-anchored proteins (Yamamoto et al.,2000). Other findings from cocultures of embryonic thalamus with postnatal visual cortex suggested further that the molecular stop signal for TCA termination did not become well established until approximately P3 (Molnár and Blakemore,1999). Thus, a GPI-anchored molecule, such as an ephrin-A protein family member that is characterized by expression beginning about the time of TCA penetration into the cortical plate and then temporarily upregulated during the period of peak TCA arbor elaboration, would be a suitable candidate for an interlaminar stop signal.

Previous studies of ephrin-A family contributions to TCA pathfinding and connectivity in SI have largely focused on perinatal expression of ephrin-A5 (Castellani et al.,1998; Gao et al.,1998; Mackarehtschian et al.,1999; Vanderhaeghen et al.,2000; Yabuta et al.,2000; Mann et al.,2002; Dufour et al.,2003). At P3, ephrin-A5 mRNA has a weak medio-lateral gradient in barrels within layer IV (Vanderhaeghen et al.,2000; Dufour et al.,2003) and, while still present at later ages (P6 and P8), shows no marked density gradation (Castellani et al.,1998; Bolz et al.,2004). Moreover, ephrin-A5 is not expressed in supragranular or infragranular cortical layers (Yabuta et al.,2000). Thus, expression timing and laminar distribution of ephrinA5 are not well suited to constrain the termination of TCAs within layer IV. Consistent with this suggestion, results from ephrin-A5 deletion mutants show tangential distortions of the PMBSF (Prakash et al.,2000; Vanderhaeghen et al.,2000; Uziel et al.,2002; Dufour et al.,2003), but reveal no deficits in laminar targeting precision (Yabuta et al.,2000; Uziel et al.,2008). Taken together, the results from these studies indicate that ephrin-A5 may be involved in directing TCA growth relevant to topography, but is insufficient for restricting axonal growth to the correct layers.

Our results suggest that during normal development, transient ephrin-A2 expression at upper and lower layer IV borders may constrain the laminar distribution of TCAs until P12–P14 when arbor formation is well advanced. Within a different context, a similar role for ephrin-A2 has been proposed in the hippocampus. Several days following axotomy of the perforant pathway in adult mouse, ephrin-A2 undergoes long term upregulation in a band spanning the denervated terminal zone and may thus serve to limit local axonal sprouting while stimulating synaptogensis (Wang et al.,2005).

Several EphA receptors (EphA3, EphA4, and EphA7) capable of binding with ephrin-A2 (Gale et al.,1996) are expressed in VB neurons during the late embryonic and early postnatal period (Mackarehtschian et al.,1999; Vanderhaeghen et al.,2000; Sestan et al.,2001; Dufour et al.,2003; Bolz et al.,2004), a time when VB TCAs reach the cortex, make connections in layer IV, and refine those connections (Catalano et al,1996). We suggest that ephrin-A2 expression in cortical neurons acting in concert with EphA receptors on TCA terminals may serve a distinct role in regulating TCA arbor development in the appropriate lamina of SI after their topographic location has been specified by additional molecules.