MeCP2 deficiency disrupts axonal guidance, fasciculation, and targeting by altering Semaphorin 3F function

Abstract

Rett syndrome (RTT) is an autism spectrum disorder that results from mutations in the transcriptional regulator methyl-CpG binding protein 2 (MECP2). In the present work, we demonstrate that MeCP2 deficiency disrupts the establishment of neural connections before synaptogenesis. Using both in vitro and in vivo approaches, we identify dynamic alterations in the expression of class 3 semaphorins that are accompanied by defects in axonal fasciculation, guidance, and targeting with MeCP2 deficiency. Olfactory axons from Mecp2 mutant mice display aberrant repulsion when co-cultured with mutant olfactory bulb explants. This defect is restored when mutant olfactory axons are co-cultured with wild type olfactory bulbs. Thus, a non-cell autonomous mechanism involving Semaphorin 3F function may underlie abnormalities in the establishment of connectivity with Mecp2 mutation. These findings have broad implications for the role of MECP2 in neurodevelopment and RTT, given the critical role of the semaphorins in the formation of neural circuits.

Fig. 1. Mecp2-null mice show defective MOB laminar targeting

(A-C) Coronal MOB sections obtained from WT (A) or Mecp2-null mice (B) were immunostained with anti-OMP antibodies to visualize mature olfactory axons that innervate the MOB. Olfactory nerve layer (ONL); Glomerular layer (GL); External plexiform layer (EPL). (B) Arrowheads indicate axons overshooting the EPL in Null mice, which are shown in more detail in B’ (inset). (C) Percentage of glomeruli that show aberrant projections in MOB from WT and Mecp2-null mice at P14 and P28. For quantification, 3-4 mice per genotype per age were used. From each mouse, four MOB sections at similar rostro-caudal levels were analyzed blind for genotype and a total of 200 glomeruli per mice were assessed (**p< 0.01, Mann-Whitney t test). Scale Bar: 50 μm.

Fig. 2. Axonal convergence of M72 axons is preserved in Mecp2-null mice

(A and D) Whole-mount staining of MOB at P14 showing representative M72 glomeruli. (A) Whole–mount view of lateral M72 glomeruli showing the method of quantification. Scale Bar: 400μm. (B) Ratio between the distances of lateral M72 glomeruli from the rostral tip of the OB (μm) divided by the total distance from the rostral tip to the caudal tip (μm). (C) Ratio between the distances of lateral M72 glomeruli from the lateral tip of the OB (μm) divided by the total distances from the lateral border to the medial border (μm). (D) Whole– mount view of a medial M72 glomerulus. Scale Bar: 400μm (E) Ratio between the distances of medial M72 glomeruli from the rostral tip of the OB (μm) divided by the total distances from the rostral tip to the caudal tip (μm). (F) Ratio between the distances of medial M72 glomeruli from the dorsal tip of the OB (μm) divided by the total distances from the dorsal tip to the ventral tip (μm). Each bar is the average of 16 glomeruli obtained from 4 animals per genotype. No significant changes were founded in the relative position of M72 glomeruli from WT and Mecp2-null mice (Mann-Whitney t-test).

Fig. 3. Mecp2 expression in the accessory olfactory system

(A-B) Coronal sections of VNO from P14 (A-B) WT mice were incubated with polyclonal anti-Mecp2 (red) as previously indicated (Cohen et al., 2003). DAPI (blue) was used for nuclei labelling. Mecp2 was detected in the cell body of all receptor neurons (RC) from the VNO but not in the supporting cell layer (SC). (C-D) Similarly, sagittal sections from P14 AOB were stained for Mecp2 and the expression was restricted to neuronal bodies located in the glomerular layer (GL), periglomerular layer (PL) and granular layer (GrL), and absent from the nerve layer (NL). A similar expression pattern was detected in accessory olfactory tissues from P28 mice (not shown). Top: dorsal. Scale Bar: 100μm.

Fig. 4. Mecp2-null mice show defective vomeronasal nerve (VNN) fasciculation

(A-D) Coronal and sagittal sections of MOB were obtained from WT and Mecp2-null mice at P14 and stained with EC lectin to label VNO projections. WT mice VNO axons are organized into large fascicles (A, arrowheads), while they form several smaller bundles in Mecp2-null mice (B, arrowheads). Top: dorsal. Scale bar: 100 μm. On a sagittal view, WT VNO fibers coalesce on the medial surface of the OB as they extend caudally to the AOB (C), whereas Mecp2-null projections show distorted trajectories and some axons terminate erratically in the OB (D, arrow). Top: caudal. Scale bar: 200 μm. n=5 mice per genotype, per age.

Fig. 5. Zonal targeting of vomeronasal axons is altered in Mecp2-null mice

(A-C) Sagittal sections of MOB were stained with anti-G_i2-α(green) to visualize VNO fibers that project to the anterior half of the AOB. The anterior-posterior border of the AOB (arrows) was defined using BS lectin (red), which binds to the posterior AOB. (A-A’) In WT mice, all G_i2-α⁺ projections innervate the anterior region of the AOB. (B-B’) Several improperly targeted G_i2-α⁺ axon bundles were detected deep in the posterior half of the AOB in Mecp2-null mice (arrowheads). Top: caudal. The number of G_i2-α⁺ bundles (per section) that innervate the posterior AOB is represented graphically in (C). Quantification was performed from 4-5 mice per genotype per age. From each animal, 6 sagittal sections showing clear anterior/posterior AOB were examined (6 AOB sections per side). Therefore, each bar represents the average of 12 sections per animal, analyzed in 4-5 animals, per genotype per age (**p < 0.01, Mann-Whitney t test). Scale Bars: 250 μm. (D-G) Coronal sections of VNO from P14 day old WT and Mecp2-null mice were incubated with anti-Gi2-α (D, E) or anti-Goα (F, G). The sections were contrasted with DAPI to visualize all the cells in the VNO. Cell bodies of Gi2α-expressing neurons are located in the apical VNO in both WT and Mecp2-null mice (D, E). Similarly, Go-expressing cell bodies are restricted to the basal region of the VNO in both WT and Mecp2-null mice (F, G). The pattern of staining for these markers is complementary, and the boundaries display an undulating shape as previously reported (Jia and Halpern, 1996). Scale bar: 100μm.

Figure 6. Mecp2-null mice show differential expression of guidance molecules

mRNA levels of Sema3A, Neuropilin-1, PlexinA4, CRMP2, Sema3F, PlexinA3, and Neuropilin-2 were determined in WT and Mecp2-null mice by real-time RT-PCR. (A) OE and OB or (B) VNO and AOB were obtained from 5-6 animals per group at P1, P7 and P14, and cDNA was prepared individually from each mouse. White bars represent WT mice and black bars Mecp2-null mice. Results are represented as a ratio between the relative amounts of target gene and GAPDH. WT results were normalized to 1. Real time-PCR reactions were run separately in triplicates for each mouse and data analyzed by unpaired Mann– Whitney t test with 95% confidence intervals. *p < 0.05; **p < 0.01, ***p < 0.001.

Figure 7. Mecp2-null olfactory axons respond normally to Semaphorin 3F in vitro

VNO explants (E16) from WT (A,B) or Mecp2-null mice (C,D) were co-cultured next to aggregates of 293 cells transfected with AP-Tag-4 vector (left panels) or Sema3F-AP (right panels). Explants were stained with TUJ1 (red), scored blindly using the criteria presented in E, and the percentage of explants assigned to each score was represented graphically (F). Data was analyzed by Kruskal-Wallis followed by Dunn’s multiple comparison tests with 95% confidence intervals. (a) p <0.001 with respect to WT-AP and Null-AP and (b) p <0.001 with respect to WT-AP and Null-AP. Scale Bar: 100 μm.

Figure 8. Sema3A-repellent effect and axonal outgrowth are not impaired in OE explants from Mecp2-null mice

(A-D) E16 OE explants from WT or Mecp2-null mice were cultured in the proximity of aggregates of 293 cells transfected with AP-Tag-4 vector or Sema3A-AP. Explants were stained for βIII-tubulin (TUJ1, green) and the extent of repulsion was scored blindly using the criteria shown in A-D. Scale Bar: 100μm. E) The percentage of explants assigned to each score represented graphically. Data was analyzed by Kruskal-Wallis followed by Dunn’s multiple comparison tests with 95% confidence intervals. (a) p <0.001 with respect to WT-AP and Null-AP and (b) p <0.001 with respect to WT-AP and Null-AP; F) Axonal outgrowth was determined in embryonic OE explants cultured alone by measuring the ratio between the area of axonal growth (determined by TUJ1 staining) and the area of the explant (p>0.05, Mann-Whitney t test).

Figure 9. Mecp2-null OB fails to repel VNO axons

(A-D) E16 VNO explants from WT or Mecp2-null mice were co-cultured in the proximity of WT or Mecp2-null MOB explants (P5-P7) as indicated in the methods and labeled for βIII-tubulin. Panels A-D show representative images for each condition. Scale Bar: 100 μm. (E). The repulsive response is reduced only when VNO axons are cultured in the proximity of Mecp2-null MOB. Data was analyzed by Kruskal-Wallis followed by Dunn’s multiple comparison tests with 95% confidence intervals. (a) p <0.01 with respect to WT_OB-WT_VNO and WT_OB-KO_VNO; (b) p <0.05 with respect to WT_OB-WT_VNO and WT_OB-KO_VNO. F) Similarly, WT VNO explants were co-cultured with WT MOB explants (P5-P7) in the presence of 100μg of anti–Npn-2 or rabbit IgG as a control. The MOB repellent effect is reduced by the addition of anti–Npn-2 (*p < 0.05, Mann-Whitney t test).