Plasticity of thalamocortical axons is regulated by serotonin levels modulated by preterm birth

Abstract

Sensory inputs are conveyed to distinct primary areas of the neocortex through specific thalamocortical axons (TCA). While TCA have the ability to reorient postnatally to rescue embryonic mistargeting and target proper modality-specific areas, how this remarkable adaptive process is regulated remains largely unknown. Here, using a mutant mouse model with a shifted TCA trajectory during embryogenesis, we demonstrated that TCA rewiring occurs during a short postnatal time window, preceded by a prenatal apoptosis of thalamic neurons-two processes that together lead to the formation of properly innervated albeit reduced primary sensory areas. We furthermore showed that preterm birth, through serotonin modulation, impairs early postnatal TCA plasticity, as well as the subsequent delineation of cortical area boundary. Our study defines a birth and serotonin-sensitive period that enables concerted adaptations of TCA to primary cortical areas with major implications for our understanding of brain wiring in physiological and preterm conditions.

Keywords: plasticity; preterm birth; sensitive period; serotonin; thalamic axons.

Fig. 1.

Embryonic TCA mistargeting is rapidly corrected after birth. (A) NetG1 staining on the coronal section of E18.5 (n_control = 4; n_mutant = 4) showing the abnormal pathfinding of TCA in the subpallium (open arrowheads) in Ebf1cKO (Dlx5/6::Cre;Ebf1^fl/−) compared to control (solid arrowheads) embryos. Coronal section of the subpallium at E17.5 after DiI injections in the dLGN (n_control = 7; n_mutant = 6) indicating that dLGN TCA are misrouted in the ventral subpallium in Ebf1cKO (open arrowheads) compared to control (solid arrowheads) embryos. Coronal section of the thalamus after DiI injection into V1 (red) and DiA injection into caudal S1 (green) in control (n = 8) and Ebf1cKO (n = 7) mice, demonstrating that TCA target inappropriate cortical domains at P0. Schematic representation of DiI/DiA placements as well as aberrant TCA pathfinding through the mutant subpallium compared to controls. (B) Cleaved caspase-3 staining (cCasp3) on the coronal section, revealing massive apoptosis in the VP thalamic nucleus (solid arrowheads) and the dLGN at a more rostral level (Inset, open arrowheads) in E17.5 Ebf1cKO (n = 4) compared to control (n = 3) embryos, while limited staining was observed in both conditions at P0 (n_control = 4; n_mutant = 4). (C) TCA-GFP (n_control = 10; n_mutant = 9) staining and in situ hybridization of specific cortical markers Rorβ and Lmo4 (at least n = 5 for each condition) on the P7 flattened cortex showing a severe reduction of V1 and S1 areas (dotted lines delineation) at the expense of the intercalated HO PTLp domain in Ebf1cKO mice: −32% for S1; −36% for V1 and +41% for the PTLp domain. (D) Quantification of the size of the cortical surface (control = 100 ± 1.34%, n = 8; mutant = 98.03 ± 1.25%, n = 5), V1 area (control = 100 ± 2.20%, n = 16; mutant = 63.30 ± 3.64%, n = 11), S1 area (control = 114.7 ± 3.30%, n = 17; mutant = 82.49 ± 2.71%, n = 11) and PTLp width (control = 105.7 ± 3.86%, n = 17; mutant = 147.40 ± 8.40%, n = 11) at P7. (E) Coronal section of the thalamus at P1, P2, and P7 and of the subpallium at P2 after DiI injection into V1 (red) and DiA injection into caudal S1 (green) showing rapid TCA rewiring after birth with some ventrally misrouted dLGN TCA reaching V1 in Ebf1cKO (open arrowheads) compared to control (solid arrowheads) mice. (F) Quantification of the matching index (as defined in SI Appendix, Fig. S3B) by measuring the proportion of the red signal in visual thalamic nuclei (dLGN and LP) following DiI injection into V1. P0_control = 99.95 ± 0.03% (n = 8) vs. P0_mutant = 6.89 ± 2.95% (n = 7); P1_control = 99.89 ± 0.11% (n = 7) vs. P1_mutant = 24.05 ± 9.53% (n = 7); P2_control = 99.99 ± 0.01% (n = 8) vs. P2_mutant = 99.82 ± 0.17% (n = 4); P7_control = 99.99 ± 0.01% (n = 9) vs. P7_mutant = 99.97 ± 0.02% (n = 5). (G) Sequential events enabling postnatal TCA plasticity in Ebf1cKO mice and subsequent impact on the primary cortical areas. Values are presented as mean ± SEM; ***P < 0.001, Mann–Whitney U test for D and two-way ANOVA with Sidak’s correction for multiple comparisons for F (Scale bars, 250 μm). Ctx, cortex; dLGN, dorsal lateral geniculate nucleus; LP, lateral posterior nucleus; POm, posteromedial nucleus; PTLp, posterior parietal association area; S1, primary somatosensory cortex; SubP, subpallium; Thal, thalamus; V1, primary visual cortex; VP, ventroposterior nucleus. See also SI Appendix, Figs. S1–S3 and S8.

Fig. 2.

Preterm birth affects the efficiency of TCA rewiring. (A) Experimental procedure to induce preterm birth using mifepristone injection at E17.5. (B) TCA-GFP immunostaining on the flattened cortex of P2 to P3 untreated (full-term) and P3 to P4 preterm pups. (C) Barrel filed formation, quantified as the Barrel Field Intensity (BFI) index using the TCA-GFP line, increases in the postnatal stages: Untreated_P2 = 0.047 ± 0.011% (n = 6) vs. Untreated_P3 = 0.210 ± 0.023% (n = 12) or Preterm_P3 = 0.11 ± 0.013% (n = 6) and Preterm_P3 vs. Preterm_P4 = 0.25 ± 0.024% (n = 7). (D) Coronal section of the thalamus after DiI injection into V1 (red) and DiA injection into caudal S1 (green) in control and Ebf1cKO (Dlx5/6::Cre;Ebf1^fl−/) mice in the P7 untreated, P7 preterm, and P1 preterm conditions. In preterm Ebf1cKO at P7 and P1, V1 injection (red) simultaneously led to a retrolabeling in the VP and dLGN thalamic nuclei, indicating that some TCA of the VP nucleus targeted V1 instead of S1 (empty arrowheads). (E) Coronal section of the thalamus after CTB-555 injection into rostral V1 (red) and CTB-488 injection into caudal V1 (green) in control and Ebf1cKO mice at P7 in the untreated and preterm conditions. V1 double injections in preterm Ebf1cKO mice reveal the defective plasticity of TCA as shown by retrolabeling in the somatosensory-VP nuclei (arrowheads). (F) Quantification of the matching index by measuring the proportion of the red signal in visual thalamic nuclei (dLGN and LP) following DiI injection into V1 for D. At P7: Untreated_control = 99.99 ± 0.002% (n = 9) vs. Untreated_mutant = 99.97 ± 0.02% (n = 5); Preterm_control = 99.73 ± 0.12% (n = 12) vs. Preterm_mutant = 22.32 ± 13.23% (n = 5). At P1: Untreated_control = 99.89 ± 0.11% (n = 7) vs. Untreated_mutant = 24.05 ± 9.53% (n = 7); Preterm_control = 98.96 ± 1.03% (n = 8) vs. Preterm_mutant = 24.74 ± 8.80% (n = 8). (G) Quantification of the matching index at P7 by measuring the proportion of the signal in visual thalamic nuclei (dLGN and LP) following CTB injection into V1 for E. Untreated_control = 95.51 ± 2.18% (n = 6) vs. Untreated_mutant = 87.22 ± 6.12% (n = 5); Preterm_control = 94.85 ± 1.61% (n = 10) vs. Preterm_mutant = 42.25 ± 7.39% (n = 10). (H) Defective TCA plasticity in preterm Ebf1cKO at P7 as V1 was targeted by dLGN and some VP axons. Values are presented as mean ± SEM; **P < 0.01, ***P < 0.001, Mann–Whitney U test for C and two-way ANOVA with Sidak’s correction for multiple comparisons for F and G (Scale bars, 250 μm). Ctx, cortex; dLGN, dorsal lateral geniculate nucleus; LP, lateral posterior nucleus; POm, posteromedial nucleus; S1, primary somatosensory cortex; SubP, subpallium; V1, primary visual cortex; VP, ventroposterior nucleus. See also SI Appendix, Figs. S4, S5, and S8.

Fig. 3.

Preterm birth impairs TCA plasticity partly through serotonin levels. (A) Experimental procedure to induce preterm birth by mifepristone injection at E17.5 and reduced or increased 5-HT levels since birth with administration of daily injections of PCPA or Fluoxetine (Fx) respectively. (B) Coronal section of the thalamus after DiI injection into V1 (red) and DiA injection into caudal S1 (green) in control and Ebf1cKO (Dlx5/6::Cre;Ebf1^fl/−) mice at P7 in the untreated, preterm, PCPA, or Preterm+Fx conditions. In preterm or PCPA Ebf1cKO at P7, V1 injection (red) simultaneously led to retrolabeling in VP and dLGN thalamic nuclei, indicating that some TCA of the VP targeted V1 instead of S1 (arrowheads). Increased 5-HT levels induced by Fx injections in preterm Ebf1cKO restored TCA plasticity. (C) Quantification of the matching index by measuring the proportion of the red signal in visual thalamic nuclei (dLGN and LP) at P7 following DiI injection into V1. Untreated_control = 99.99 ± 0.01% (n = 9) vs. Untreated_mutant = 99.97 ± 0.02% (n = 5); Preterm_control = 99.73 ± 0.11% (n = 12) vs. Preterm_mutant = 22.32 ± 13.23% (n = 5); PCPA_control = 100 ± 0.01% (n = 6) vs. PCPA_mutant = 61.78 ± 7.03% (n = 5); Preterm+Fx_control = 100 ± 0.01% (n = 10) vs. Preterm+Fx_mutant = 99.72 ± 0.27% (n = 6). (D) Coronal section of the thalamus after CTB-555 injection into rostral V1 (red) and CTB-488 injection into caudal V1 (green) in control and Ebf1cKO mice at P7 in the untreated and Preterm+Fx conditions, demonstrating that increased 5-HT levels in preterm Ebf1cKO induced by Fx injections restore TCA plasticity. (E) Quantification of the matching index by measuring the proportion of the signal in visual thalamic nuclei (dLGN and LP) at P7 following the injections of CTB into V1. Untreated_control = 95.51 ± 2.18% (n = 6) vs. Untreated_mutant = 87.22 ± 6.12% (n = 5); Preterm+Fx_control = 92.67 ± 2.52% (n = 14) vs. Preterm+Fx_mutant = 85.92 ± 7.25% (n = 14). (F) Decreased 5-HT levels induced by PCPA injections since birth mimicked the defect of TCA plasticity observed in preterm Ebf1cKO at P7. Conversely, increased 5-HT levels induced by Fx injections in preterm Ebf1cKO restored TCA plasticity. Values are presented as mean ±SEM; ***P < 0.001, two-way ANOVA with Sidak’s correction for multiple comparisons (Scale bars, 250 μm). Ctx, cortex; dLGN, dorsal lateral geniculate nucleus; LP, lateral posterior nucleus; POm, posteromedial nucleus; S1, primary somatosensory cortex; SubP, subpallium; V1, primary visual cortex; VP, ventroposterior nucleus. See also SI Appendix, Figs. S5, S6, and S8.

Fig. 4.

TCA plasticity impairment impacts molecular identity of the V1 border. (A) In situ hybridization flattened cortex at P7 using specific cortical markers Rorβ and Lmo4 in control and Ebf1cKO (Dlx5/6::Cre;Ebf1^fl/−) mice in the untreated (n_control = 11; n_mutant = 8), preterm (n_control = 9; n_mutant = 6), PCPA (n_control = 8; n_mutant = 10) or Preterm+Fx (n_control = 16; n_mutant = 6) conditions. In the preterm and PCPA Ebf1cKO conditions, the molecular identity of the V1 border was affected, according to the blurry expression of Rorβ and Lmo4 (pink arrowheads) compared to its sharp delineation in controls or untreated Ebf1cKO. Fx injections in the preterm Ebf1cKO condition restored the well-defined molecular identity. (B) Quantification of the V1 boundary gradient observed with in situ hybridization. Control_Untreated = 1.005 ± 0.037 (n = 11) vs. Control_Preterm = 1.023 ± 0.05 (n = 9) vs. Control_PCPA = 0.909 ± 0.08 (n = 8) vs. Control_Preterm+Fx = 0.870 ± 0.07 (n = 16) and Mutant_Untreated = 0.833 ± 0.07 (n = 8) vs. Mutant_Preterm = 0.38 ± 0.04 (n = 6) vs. Mutant_PCPA = 0.43 ± 0.03 (n = 10) vs. Mutant_Preterm+Fx = 0.74 ± 0.08 (n = 6). (C) Defective V1 molecular identity in the preterm and PCPA Ebf1cKO conditions while Fx injections in preterm Ebf1cKO rescued the sharpness of the V1 border at a level comparable to that in the untreated control condition. Values are presented as mean ± SEM; *P < 0.05, ***P < 0.001, two-way ANOVA with Sidak’s correction for multiple comparisons (Scale bars, 250 μm)=; Ctx, cortex; dLGN, dorsal lateral geniculate nucleus; LP, lateral posterior nucleus; POm, posteromedial nucleus; PTLp, posterior parietal association area; S1, primary somatosensory cortex; SubP, subpallium; V1, primary visual cortex; VP, ventroposterior nucleus. See also SI Appendix, Fig. S7.