Distorted Coarse Axon Targeting and Reduced Dendrite Connectivity Underlie Dysosmia after Olfactory Axon Injury
Aya Murai, Ryo Iwata, Satoshi Fujimoto, Shuhei Aihara, Akio Tsuboi, Yuko Muroyama, Tetsuichiro Saito, Kazunori Nishizaki and Takeshi Imai
eNeuro 5 October 2016, 3 (5) ENEURO.0242-16.2016; DOI: https://doi.org/10.1523/ENEURO.0242-16.2016
Laboratory for Sensory Circuit Formation, RIKEN Center for Developmental Biology, Chuo-ku 650-0047, JapanDepartment of Otolaryngology–Head and Neck Surgery, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama 700-8558, Japan
Mistargeting of OSN axons after OSN lesion and recovery. A, Schematic representation of unilateral axotomy of dorsal-zone OSNs. Axon bundles of dorsal zone OSNs projecting to dorsolateral OB were transected near the cribriform plate. Only the right side was injured. B, The micro knife was coated with TMR-dextran dye to visualize the incision site. MOR29B-EYFP mice were used. Axotomy was limited to dorsal zone OSNs (broken line). C, OE and OB 7 d after the axotomy. D, OE and OB 14 d after the axotomy. An almost complete OSN lesion in the anterodorsal OE and OB was observed in eight of eight mice at day 14. E, Recovery of OSNs 42 d after axotomy. OSNs were regenerated, and the number of OSNs recovered to levels comparable to those of the controls on the noninjured side. However, OSN axons were scarce in the posterior part of the dorsal OB (right, white broken line). F, High-magnification images of E, middle, showing axotomized (orange) and control (gray) sides. OSN axons formed small glomerular-like structures in the anterior OB after axotomy. G, Quantification of the diameter for glomerular-like structures in the control and axotomized sides. OMP-GFP mice were used in C–G. ***p < 0.001 (Mann–Whitney U test). n = 559 (axotomy) and 136 (control) from 12 mice. A, Anterior; D, dorsal. Scale bars: C–E, 200 μm; F, 100 μm.
Anterior–posterior topography is disrupted after axotomy. A, Defective axonal projection of OSNs expressing MOR29B-EYFP. On the control side, a major glomerulus for MOR29B was observed in the posterolateral area of the dorsal OB. However, MOR29B-expressing OSNs projected to the anterior and/or medial region of the OB after OSN axon injury (42 d recovery; in six of six animals analyzed). A, Anterior; L, lateral. B, Mistargeting of Nrp1-positive OSN axons to the anterior OB. C, High-magnification confocal images of boxed areas in B. On the control side, Nrp1-positive axons were sorted to the superficial layer of the olfactory nerve layer and projected to the posterior OB. However, on the axotomized side, Nrp1-positive axons were found in the deeper portions of the olfactory nerve layer, and invaded glomeruli in the anterior OB. Nrp1-positive and Nrp1-negative axons formed small glomerular-like structures within glomeruli in the anterior OB (arrowheads). Glomerular layer is demarked by dotted lines. An OMP-GFP mouse was analyzed. Note that Nrp1-positive and Nrp1-negative OSN axons formed distinct glomerular-like structures within glomeruli in the high-magnification images. Scale bars: A, 500 μm; B, 200 μm; C, 100 μm.
Heterogeneous OSN axons converge onto a glomerulus. OB sections stained with anti-GFP (OMP-GFP) and anti-Kirrel2 antibodies were imaged with confocal microscopy. DAPI staining shows the distribution of periglomerular cells surrounding glomeruli. The layout of periglomerular cells remains unchanged. The glomerular layer is demarked by dotted lines. The anterior part of the dorsal OB is shown. In the control OB, Kirrel2 immunoreactivity was glomerulus specific and homogeneous within a glomerulus. However, on the axotomized side, distinct glomerular-like structures with differential Kirrel2 levels were found within each glomerulus (arrowheads). Similar results were obtained with 17 experiments. Scale bars, 100 μm.
Connectivity of mitral cell dendrites is reduced after axotomy. A, Stacked confocal images (20 μm thick) of OB slices from Thy1-YFP (G line) mice. Mitral cells are labeled with YFP in this mouse line. Sagittal OB slices were cleared with SeeDB2G, and anterodorsal part (red) was analyzed with confocal microscopy. Note that control OBs show thick primary dendrites extending to the glomerular layer (white arrowheads), whereas only faint signals were seen in the outer external plexiform layer under axotomized conditions. For unbiased quantification, brightly labeled mitral cells (top 20%) were traced for quantification. B, Representative Neurolucida tracing of mitral cell primary dendrites in the control and axotomized OBs. Note that only primary dendrites are shown. Primary dendrites of mitral cells partially deinnervated from glomeruli in the axotomized side. Innervated glomeruli are shown in blue, and other glomeruli are shown in gray. C, Quantification of primary dendrite innervation. Mitral cell dendrites were often thinned and became untraceable at 21 d after axotomy. The partial deinnervation of primary dendrites persisted for at least 84 d after axotomy. The numbers of mitral cells (n) and animals (N) are indicated. ***p < 0.001 (χ2 test for mitral cells with innervating dendrites, compared with the control side). Scale bars, 100 μm.
Atrophy of dendritic tufts in mitral cells. A, Mitral/tufted cells were labeled with tdTomato using in utero electroporation at E12. OMP-GFP mice were used to visualize OSNs. Axotomy was performed at postnatal day 56, and mitral cell morphology was analyzed 42 d after axotomy. The z-stacked confocal images (52 μm thick) are shown. High-magnification images show the tufted structures within a glomerulus in the control, but not after axotomy. B, Representative Neurolucida tracing of mitral cell dendrites, including tufted structures within glomeruli. We show five representative neurons, including the ones shown in A. Boxed areas are shown in high-magnification images in A. Scale bars, 100 μm.
Odor responses in mitral cells are reduced after axotomy. A, Two-photon Ca2+ imaging of the olfactory bulb using Thy1-GCaMP6f mice (GP5.11). In this mouse line, GCaMP6f is specifically expressed in mitral cells. The glomerular layer in the anterodorsal part of the OB (boxed area) was imaged using two-photon microscopy. Right OBs (demarcated by orange broken lines) were analyzed. A, Anterior; L, lateral. B, The odor map after axotomy was altered (day 42). Odor-evoked responses in the OB were analyzed by two-photon Ca2+ imaging. Average activation maps are shown. In the control OB, butyric acid activates the anteromedial part of the dorsal OB (class I), whereas acetophenone, guaiacol, and vanillin activated more of the posterolateral area (class II). However, the response amplitude was much reduced, and localization was not clear after the recovery after axotomy. C, Cumulative curves for odor-evoked responses (pixel-based ΔF/F0 during the first 5 s of odor exposure). After recovery from axotomy, odor-evoked responses in the OB were much reduced. *p < 0.05, ***p < 0.001 (Kolmogorov–Smirnov test). Scale bars, 1 mm.
Schematic representation of defective circuit recovery after axotomy. A, Embryonic development of the olfactory map. Target cues (e.g., Sema3A) guide axons to correct positions in the OB. A, Anterior; P, posterior. B, Normal OSN turnover process. We assume that axon–axon interactions (pioneer–follower interactions) are important to maintain the topographic olfactory map in the normal turnover process. C, Our proposed model for defective OSN projection and partial dendrite deinnervation after axotomy. Without positional cues and axon–axon interactions, regenerated OSN axons cannot find correct targets in the OB after axotomy. Reduced connectivity of primary dendrites may lead to reduced odor responses in mitral cells.