Article Figures & Data
Figures
- Figure 1.
Conditional elimination of EphA4 in the forebrain but not the cervical spinal cord. Immunoblotting analysis of motor cortex lysates shows decreased levels of EphA4 expression between wild-type and the two EphA4 conditional mutants (A), but minimal reductions in the cervical spinal cord (B). Lanes show data from individual mice (A, B). EphA4 was localized at 130 kDa and actin was localized between molecular weight markers at 55 and 35 kDa. Exposure times for gels were as follows: motor cortex EphA4, 20 s; actin, 6 s; cervical spinal cord, 60 s, actin, 6 s. C–E, Bars plot ratio of mean (±SE) EphA4 and Actin protein levels for each group (n = 4–5 mice/group). There is a significant reduction of EphA4 protein level in the motor cortex of homozygous Emx1-Cre;EphA4tm2Kldr compared with WT (C; p = 0.029, Mann–Whitney test). This reduction is more pronounced in the Emx1-Cre;EphA4tm2Kldr than the Emx1-Cre;EphA4tm1Bzh mutant (D; p = 0.028, Mann–Whitney test). Immunoblotting analysis from cervical spinal cord lysates showed similar protein level in wild-type and two EphA4 conditional mutants (p = 0.232, Kruskal–Wallis test, Dunn's posttest: p > 0.05; n = 4/group).
- Figure 2.
Laterality of spinal interneurons and brainstem neurons in Emx1-Cre;EphA4tm2Kldr mice. A, Overlapping dark-field and fluorescence images of the cervical spinal cord for a representative WT mouse highlighting the site of unilateral RDA injection and the labeling observed in the ipsilateral and contralateral C7/C8 segments. B, C, Similar to A, but for retrogradely labeled neurons in the red nucleus and medullary reticular formation. Rubrospinal neurons (B) are located contralateral to the spinal injection site, whereas reticulospinal neurons are located bilaterally (C). White lines mark the midline (A, C) and the border of the red nuclei (B); calibrations: 500 μm. D–F, Mean percentage (±SE) of labeled ipsilateral and contralateral neurons (n = 4 mice/group) revealed no significant differences between Emx1-Cre;EphA4tm2Kldr and WT mice at all three levels (Mann–Whitney test, cervical cord: p = 0.23; rubrospinal neurons: p = 0.2; reticulospinal neurons, contralateral and ipsilateral: p = 0.63).
- Figure 3.
Effects of conditional elimination of EphA4 in the forebrain on forelimb movements during locomotion and exploratory reaching. A, B, Stacked histograms of forelimb locomotor behavior. During overground locomotion (A), both mutant mice (n = 6–10) and WT mice (n = 10) step with their forelimbs nearly 100% of the time (p = 0.30, F = 1.27, one-way ANOVA, Bonferroni's posttest p > 0.05). During unobstructed treadmill locomotion (B), Emx1-Cre;EphA4tm2Kldr mice (n = 15) only showed a slight but significant percentage of hopping compared with WT (n = 10) and Emx1-Cre;EphA4tm1Bzh mice (n = 6; one-way ANOVA, p = 0.003, F = 7.4; Bonferroni's posttest p < 0.05). C, D, Mean percent (±SE) forelimb hopping during adaptive locomotion. Data for three 3 treadmill speeds (6 cm/s; 10 cm/s; 17 cm/s) and one obstacle height (1.0 cm) are plotted (EphA4tm1Bzh, n = 18; Emx1-Cre;EphA4tm1Bzh, n = 18; WT, n = 10; Emx1-Cre;EphA4tm2Kldr, n = 15). There was a significant increase in hopping between conditional mutant groups and their respective controls: Emx1-Cre;EphA4tm1Bzh,F = 8.8, p = 0.004, no differences between treadmill velocities (F = 0.08, p = 0.092); Emx1-Cre;EphA4tm2Kldr, F = 117, p < 0.0001, no differences between treadmill velocities, F = 0.02, p = 0.098. In addition, the Emx1-Cre;EphA4tm2Kldr mice showed a significant increase in hopping over EphA4tm2Kldr controls (17 cm/s; 1 cm; forelimb: WT mean = 1.5%±0.74%; EphA4tm2Kldr mean = 2 ± 2%, Emx1-Cre;EphA4tm2Kldr: mean = 47.2%±5.9%, p < 0.0001, one-way ANOVA, Bonferroni's posttest revealed no difference between WT and EphA4tm2Kldr, p > 0.05). E, Histograms of forelimb reaching exploratory behavior. There was no statistically difference between WT and Emx1-Cre;EphA4tm1Bzh mice (n = 6; p > 0.05). In contrast, Emx1-Cre;EphA4tm2Kldr mice (n = 10) showed a significant use of both forelimbs when reaching the wall of the cylinder compared with WT (n = 11; one-way ANOVA, p < 0.0001, F = 66.9; Bonferroni posttest, p < 0.05). The schematic shows a diagram of a mouse in the reaching cylinder.
- Figure 4.
Bilateral CST terminations into cervical spinal cord in Emx1-Cre;EphA4tm2Kldr conditional mutant mice. A, B, BDA-labeled axons in the medullary pyramid of WT (A) and Emx1-Cre;EphA4tm2Kldr (B) mice (calibration, 500 μm). Arrows point to corticoreticular projections to the ipsilateral and contralateral reticular formation in both mice (WT and Emx1-Cre;EphA4tm2Kldr). C, D, Confocal images of BDA-labeled axons in the cervical enlargement (C7/C8) after unilateral injections of anterograde tracer BDA into the M1 (C, WT; D, Emx1-Cre;EphA4tm2Kldr). Each image is a z-stack of 13 images (1 μm optical slices; calibration 50 μm). Note the extensive ipsilateral CST terminations in the Emx1-Cre;EphA4tm2Kldr mouse (D) and very sparse ipsilateral CST terminations in WT (C). The insets in C and D show similar CST labeling in the ventral dorsal column of each mouse. E, F, Mediolateral distributions of mean axon density from the left to the right borders of the contralateral and ipsilateral gray matter in the WT (E) and Emx1-Cre;EphA4tm2Kldr mice (F). Each graph plots the average CST density of axon terminations in Emx1-Cre;EphA4tm2Kldr (F) and WT mice (E); light-shading plots ± SEM. The arrows indicate the midline. There was a significant difference in the amount of ipsilateral, but not contralateral, labeling in the two groups (Mann–Whitney test, ipsilateral, p = 0.016; contralateral, p = 0.31).
- Figure 5.
M1 stimulation-evoked bilateral forelimb response in EphA4 conditional mutant mice. A–D, Ensemble EMG averages from contralateral and ipsilateral biceps muscles in response to threshold cortical stimulation in Emx1-Cre;EphA4tm2Kldr (A, C) and EphA4tm2Kldr controls (B, D). Each histogram plots the average of rectified EMG responses in each group. E, Average laterality index (integrated ipsilateral EMG value divided by integrated contralateral EMG value) for each group. The EMG data show a robust bilateral motor response in Emx1-Cre;EphA4tm2Kldr and contralateral responses in the controls (Mann–Whitney test, p = 0.0012). F, Cumulative distribution of laterality index for each mutant mouse. The figure plots cumulative distribution functions of laterality indices across our sample of 26 recorded motor cortex sites in the Emx1-Cre;EphA4tm2Kldr (solid line) and 20 recorded sites in the Emx1-Cre;EphA4tm1Bzh mice (dashed line). The arrows indicate the distribution means, which are significantly different (Emx1-Cre;EphA4tm1Bzh: mean = 0.7 ± 0.13, Emx1-Cre;EphA4tm2Kldr: mean = 1.2 ± 0.17; t test, p = 0.032; data reanalyzed from Paixão et al., 2013) and the distribution for the Emx1-Cre;EphA4tm2Kldr mice is shifted to the right.
- Figure 6.
Mirror site representations at threshold in EphA4 conditional knock-out mice. Color maps plot the occurrence of evoked mirror movements at each of the motor cortex stimulation sites in Emx1-Cre;EphA4tm1Bzh (A) and Emx1-Cre;EphA4tm2Kldr mice (B). The color represents a spectrum, from a high (maroon and red) to a low (blue) percentage of mirror sites. No mirror sites were found in the WT at the threshold. Whereas nearly all sites in Emx1-Cre;EphA4tm2Kldr mice evoked mirror movements, Emx1-Cre;EphA4tm1Bzh mice showed more varied responses. Bar graphs in C plot the average (n = 4–5 mice/group; n = 25 motor cortex sites within each of 17 hemispheres) of the percentage of sites from which the microstimulation evoked a mirror response. There was a significantly higher percentage of mirror sites in Emx1-Cre;EphA4tm2Kldr mice (dark gray bar) than in Emx1-Cre;EphA4tm1Bzh mice (light gray bar; one-way ANOVA, p < 0.0001, F = 94; p < 0.05, Bonferroni post hoc test).
