Fast motor axon loss in SMARD1 does not correspond to morphological and functional alterations of the NMJ

Fast motor axon loss in SMARD1 does not correspond to morphological and functional alterations of the NMJ Frank Krieger a , Nicole Elflein a , Rocio Ruiz c , Joana Guerra d , Antonio L. Serrano d , Esther Asan b , Lucia Tabares c , Sibylle Jablonka a, ⁎ a Institute for Clinical Neurobiology, University of Wuerzburg, Josef-Schneider-Str. 11, 97080 Wuerzburg, Germany b Institute for Anatomy and Cell Biology, University of Wuerzburg, Koellikerstr. 6, 97070 Wuerzburg, Germany c Department of Medical Physiology and Biophysics, School of Medicine, University of Seville, 41009 Seville, Spain d Cell Biology Group, Department of Experimental and Health Sciences, Pompeu Fabra University (UPF), CIBER on Neurodegenerative diseases (CIBERNED); 08003 Barcelona, Spain abstractarticle info Article history: Received 16 September 2012 Revised 30 November 2012 Accepted 21 December 2012 Available online 4 January 2013 Keywords: SMARD1 Nmd 2J mouse Motor axon Neuromuscular junction Neurotransmission Spinal muscular atrophy with respiratory distress type 1 (SMARD1) is a childhood motoneuron disease caused by mutations in the gene encoding for IGHMBP2, an ATPase/Helicase. Paralysis of the diaphragm is an early and prominent clinical sign resulting both from denervation and myopathy. In skeletal muscles, muscle atrophy mainly results from loss of motoneuron cell bodies and axonal degeneration. Although it is well known that loss of motoneurons at the lumbar spinal cord is an early event in the pathogenesis of the disease, it is not clear whether the corresponding proximal axons and NMJs are also early affected. In order to address this question, we have investigated the time course of the disease progression at the level of the motoneuron cell body, proximal axon (ventral root), distal axon (sciatic nerve), NMJ, and muscle fiber in Nmd 2J mice, a mouse model for SMARD1. Our results show an early and apparently parallel loss of motoneurons, proximal axons, and NMJs. In affected muscles, however, denervated fibers coexist with NMJs with normal morphology and unaltered neurotransmission. Furthermore, unaffected axons are able to sprout and reinnervate muscle fibers, suggesting selective vulnerability of neurons to Ighmbp2 deficiency. The preservation of the NMJ morphology and neurotransmission in the Nmd 2J mouse until motor axon loss takes place, differs from that observed in SMA mouse models in which NMJ impairment is an early and more general phenomenon in affected muscles. © 2013 Elsevier Inc. All rights reserved. Introduction Spinal muscular atrophy with respiratory distress type 1 (SMARD1) is the second most common form of motoneuron disease in children, after proximal spinal muscular atrophy (SMA) (Bertini et al., 1989; Grohmann et al., 1999, 2001; Guenther et al., 2007). The disease normally manifests in the first two years after birth with a characteristic paralysis of the diaphragm and distal limb muscle atrophy. These pathological features correspond to motoneuron loss and muscle fiber degeneration, including myopathic changes which are prominent in the diaphragm (Diers et al., 2005; Grohmann et al., 2001; RudnikSchoneborn et al., 2004). As a monogenetic disorder, SMARD1 is caused by mutations in the IGHMBP2 (Immunoglobulin μ-binding protein 2) gene on chromosome 11q13 which codes for an ATPase/Helicase that belongs to the SF1 superfamily (Grohmann et al., 2001; Guenther et al., 2009a; Jankowsky, 2011). However, little is known about the impact of the described cell specific abnormalities on the disease phenotype and its progression. In order to address this question we studied the neuromuscular degeneration mouse (Nmd 2J )asamodelforthejuvenile form of SMARD1. The Nmd 2J mouse carries a spontaneous point mutation in intron 4 of Ighmbp2, which leads majorly to an alternative spliced and truncated transcript (Cox et al., 1998). The mouse develops muscle weakness and atrophy, and also shows myopathic changes in the diaphragm at later disease stages (Cox et al., 1998; Grohmann et al., 2004). Former studies revealed that degeneration of muscle fibers in calf muscles appears relatively early in the second to third week after birth. In eight weeks old Nmd 2J mice, additional degenerative changes in the diaphragm were observed (Grohmann et al., 2004). However, motoneuron cell loss in Nmd 2J mice (~30%) is first detectable 10 days after birth, when these mice appear clinically healthy, accompanied by subtle axonopathic changes of the sciatic nerve. Loss of motoneurons remains within a constant level of about 30% until three weeks after birth, when Nmd 2J mice have developed hindlimb paralysis and a substantial number of degenerating axons in the sciatic nerve becomes detectable. At later disease stages motoneuron loss is mitigated (Grohmann et al., 2004). When Ighmbp2 expression is rescued selectively in neurons by transgenic methods, survival is prolonged but mice develop skeletal muscle degeneration and cardiomyopathy, indicating that disease pathology is complex and also directly affects striated muscles Neurobiology of Disease 54 (2013) 169–182 ⁎Corresponding author at: Institute for Clinical Neurobiology, Josef-Schneider-Str. 11, D-97080 Wuerzburg, Germany. Fax: +49 931 201 49788. E-mail address: [email protected] (S. Jablonka). Available online on ScienceDirect (www.sciencedirect.com). 0969-9961/$ –see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.nbd.2012.12.010 Contents lists available at SciVerse ScienceDirect Neurobiology of Disease journal homepage: www.elsevier.com/locate/ynbdi (Maddatu et al., 2004). Selective Ighmbp2 overexpression under a muscle-specific promoter erases the residing myopathy (Maddatu et al., 2005). Although these observations indicate cell autonomous disease mechanisms occurring both in muscles and nerves of the Nmd 2J mouse, the question which cellular components are mostly affected in skeletal muscle groups needs to be solved. Thus, we performed a morphometric and functional analysis of motor axons, NMJs, and skeletal muscle fibers focussing on the gastrocnemius, the diaphragm, and the levator auris longus muscle (LAL) from Nmd 2J mice. Material and methods Transgenic mouse lines The Nmd 2J mouse line (Cox et al., 1998) was housed in the central animal facilities of the University of Wuerzburg and the University of Seville. It was kept on a C57Bl/6 background by backcrossing every 3 generations. The thy1-YFP-H tg mice (YFP-12-H) originally described by (Feng et al., 2000) were obtained from Jackson Laboratories and also kept on a C57Bl/6 background in the animal facility in Wuerzburg. All described procedures and experiments were approved by the animal care and ethics committees of our institutions. In the thy1-YFP-H tg mouse only few, but not all motoneurons are YFP positive. This has an important advantage that the axonal arborization of individual motoneurons can be investigated in target muscles in comparison to conventional staining with axonal markers like phospho-tau, which gives such a dense network of axonal processes that it is impossible to follow the branching of all axonal terminals from one individual motoneuron. The thy1 promotor becomes active earliest 12 days after birth. Stable YFP expression is generally observed from P16 onwards. Therefore, we could detect both motoneuron cell bodies and axonal processes and their branching by thy1-YFP overexpression at symptomatic stages P16 and P21. Genotyping of Nmd 2J and the YFP-H tg mouse line was performed as described previously (Cox et al., 1998; Feng et al., 2000). Muscle and nerve histology Preparation of the gastrocnemius muscle for immunohistochemistry Mice were sacrificed by cervical dislocation, and the gastrocnemius muscle was dissected and immediately covered with 4% paraformaldehyde (PFA) and incubated for 20 min. After fixation single muscle fibers were teased. After washing with 0.1M Glycine in PBS for 30 min, the postsynaptic part of the neuromuscular junction was stained with α-Bungarotoxin Alexa Fluor 488 (1:500, Invitrogen) in PBS for 20 min. Subsequently, the muscle fibers were washed in PBS for 20 min and penetrated with ice-cold methanol for 2 min at −20 °C. After washing for 20 min with PBS a blocking solution containing 10% bovine serum albumin (BSA, Sigma) and 0.3% Triton X-100 was applied for 1 h. Rabbit polyclonal antibodies against neurofilament M (1:500, Millipore) and polyclonal guinea pig antibodies against Synaptophysin (1:500, Synaptic Systems) diluted in blocking solution were applied over night at 4 °C. Thereafter the muscle fibers were washed three times for 10 min in PBS. As secondary antibodies Cy3-conjugated polyclonal anti-guinea pig (1:400, Dianova) and Cy5-conjugated polyclonal anti-rabbit (1:400, Dianova) were diluted in the blocking solution and applied for 1 h at room temperature. Finally, the muscle fibers were washed three times in PBS for 10 min and mounted with a DABCO-Solution (9.97% PBS, 89.77% Glycerin and 0.26% DABCO, Merck) or Aqua PolyMount (Polysciences Inc.). For the morphological analysis of the neuromuscular endplate at P10, P16 and P21 three control and three Nmd 2J mice per time point were analyzed. The statistical difference between the genotypes was assessed using Student's unpaired two-tailed t-test (from GraphPad Prism Software). Results were considered statistically different when the p value was b0.05 and indicated by n.s. in case of non-significance. Exact p values, t values and degrees of freedom (df) are listed in the figure legend. Whole mount preparation of the gastrocnemius muscle The gastrocnemius muscles from adult Nmd 2J thy1-YFP-H tg and Nmd +/+ thy1-YFP-H tg mice were squeezed between two object slides and fixed with 4% PFA for 30 min. After fixation of the muscle tissue, it was washed in PBS two times for 30 min. The postsynaptic part of NMJs was stained with α-Bungarotoxin Alexa Fluor 555 or 647 (1:500, Invitrogen) for 30 min. The muscle tissue was then washed in PBS for 2 h. Finally, the squeezed muscles were mounted with a DABCO-Solution (9.97% PBS, 89.77% Glycerin and 0.26% DABCO). Axon course and branching studies were performed with 8 control and 3 Nmd 2J mice for P16, and 9 control and 6 Nmd 2J mice for P21. Quantification of muscle fiber size and dystrophin staining Gastrocnemius and diaphragm muscles were prepared from mice immediately after cervical dislocation, embedded in OCT compound (Tissue Tek) and freshly frozen in nitrogen-cooled 2-methylbutane. Ten μm thick cryosections were prepared and stained with a standard HE protocol. For the analysis of the fiber membrane and the NMJ innervation, gastrocnemius and diaphragm cross-sections were treated for 1 h with a blocking solution containing 10% BSA, 0.3% Triton X-100 in PBS. Muscle sections were then incubated overnight in blocking solution at 4 °C with a mouse monoclonal antibody against dystrophin (1:100, Novocastra), a rabbit polyclonal antibody against neurofilament M (1:500, Millipore) and bungarotoxin coupled to Alexa Fluor 555. Sections were then washed three times with PBS and incubated for 1 h at room temperature with a Cy2-conjugated anti-mouse (1:400, Dianova) and a Cy5-conjugated anti-rabbit (1:400, Dianova) secondary antibody. After three washing steps with PBS the sections were mounted with a DABCO-Solution (9.97% PBS, 89.77% Glycerin and 0.26% DABCO). The statistical significance of differences between control and Nmd 2J mice which is not depicted in Fig. 2, was assessed using the Student's unpaired two-tailed t-test (from GraphPad Prism Software): at P10 gastrocnemius muscle (12.9±0.3 μm versus 13.1±0.7 μm, p=0.7823, t=0.2916, df=5 for 4 control and 3 Nmd 2J mice), at P21 gastrocnemius muscle (17.8±0.3 μm versus 14.1± 1.0 μm, *p=0.0131, t=3.484, df=6 for 4 control and 4 Nmd 2J mice), at P42 gastrocnemius muscle (33.6±1.0 μm versus 16.2±0.7 μm, ***pb0.0001, t=14.24, df=6 for 4 control and 4 Nmd 2J mice), at P10 diaphragm (13.5±0.1 μm versus 13.8±0.1 μm, p=0.1191, t=1.978, df=4 for 3 control and 3 Nmd 2J ), at P21 diaphragm (14.7±0.1 μm versus 14.6±0.2 μm, p=0.6743, t=0.4526, df=4 for 3 control and 3 Nmd 2J mice), and at P42 diaphragm (18.6±0.3 versus 14.0±0.4, ***p=0.0006, t=9.658, df=4 for 3 control and 3 Nmd 2J mice). Preparation and staining of cryostat slices of sciatic nerves, spinal cord and ventral roots Nmd 2J thy1-YFP-H tg and Nmd +/+ thy1-YFP-H tg mice were perfused with 4% PFA and the sciatic nerves, ventral roots (L3–L6) and the lumbar part of the spinal cord were separated. After postfixation of the tissue with 4% PFA overnight, it had been transferred into a 30% sucrose solution. The tissues were then embedded in OCT compound and frozen within 2-methylbutane, cooled with liquid nitrogen. Subsequently, 10 μm thick cross-sections of sciatic nerves and ventral roots were performed; the spinal cords were cut in 15 μm thick cross-sections, respectively. For immunostaining, the sections were blocked with 10% BSA and 0.5% Triton X-100 in PBS for 1 h. Axons of the sciatic nerves and ventral roots were labeled overnight at 4 °C with a rabbit polyclonal antibody against neurofilament M (1:500, Millipore) whereas the motoneurons in the ventral horn of the spinal cord were stained with a goat polyclonal antibody against Choline Acetyltransferase (1:100, Millipore). Thereafter the slices were washed three times with PBS. The secondary antibodies Cy5 170 F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 goat anti-rabbit and Cy5 donkey anti-goat were applied for 1 h at room temperature, respectively. Finally, the slices were washed three times with PBS for 10 min and mounted with a DABCOSolution (9.97% PBS, 89.77% Glycerin and 0.26% DABCO). The quantitative analysis of spinal cord morphology included 4 control and 4 Nmd 2J mice at P16 and 3 control and 3 Nmd 2J mice at P21. For the quantification of YFP positive axons in the ventral roots we analyzed 4 animals per each genotype at P16 and 3 animals per genotype at P21. At P16, YFP positive motor axons in the sciatic nerve were counted in 4 control and 3 Nmd 2J mice, and at P21 three animals per genotype were studied. The statistical significance of differences between the genotypes was assessed by Student's unpaired two-tailed t-test using the GraphPad Prism Software. Results were considered statistically different when the p value was b0.05. Exact p values, t values and degrees of freedom (df) are indicated in the figure legends. Imaging and confocal analysis Imaging was acquired with an Olympus FluoView™FV1000 confocal microscope. The final processing of all images was performed with MacBiophotonics ImageJ, Photoshop 7.0 (Adobe) and Illustrator CS5 (Adobe). Linear contrast enhancement was applied to images and all individual panels contained in these figures (Figs. 1A–C, 2D–I, 3A, B, 4A–I) were treated similarly. Electromicroscopic studies of gastrocnemius muscle Gastrocnemius muscles from 4 control and 3 Nmd 2J P21 mice were dissected as described above, and small tissue pieces were immediately fixed in 2% glutaraldehyde (GA) /2% PFA in PBS (12 h at 4 °C). After fixation, the muscle tissue pieces were thoroughly washed in PBS, postfixed in 1% OsO 4 for 1 h, and embedded in Epon (Serva, Heidelberg) following dehydration in ascending concentrations of ethanol including en-bloc contrasting using 2% uranylacetate (Serva, Heidelberg) in 70% ethanol for 1 h. Ultrathin sections of ~80 nm were cut on a Reichert-Jung ultramicrotome (Leica, Bensheim, Germany), mounted on formvar-coated nickel grids, contrasted with uranyl acetate and lead citrate (Reynolds, 1963), and observed using a Leo 912 AB transmission electron microscope (Carl Zeiss NTS, Oberkochen, Germany). The images shown in Fig. 5 were taken from one wildtype and one Nmd 2J mouse. Histological and morphometric analysis of levator auris longus muscles Thelevatoraurislongus(LAL)musclesfromNmd 2J and Nmd +/+ mice were quickly dissected after killing the animals by CO 2 inhalation. Muscles were embedded in OCT compound, frozen in 2-methylbutane pre-cooled in liquid nitrogen, and stored at −80 °C until histological analysis. Histology was performed on muscle cryosections (10-μm thickness). Samples were collected on glass slides and were stained with hematoxylin/eosin (HE) or with the following primary antibodies for immunohistochemistry on unfixed transversal sections: mouse monoclonal anti-dystrophin antibody (clone MANDYS8, 1:400, Sigma); mouse monoclonal anti-eMHC (clone F1.652, undiluted supernatant, Developmental Studies Hybridoma Bank). Secondary antibodies were, respectively, goat anti-mouse IgG Alexa Fluor 488 (1:400, Invitrogen) and biotinylated anti-mouse antibody from the commercially available M.O.M. fluorescein kit (Vector laboratories). Immunostaining was performed according to manufacturer instructions. Stainings of the nuclei were performed with TO-PRO-3 from Sigma. Neuromuscular junctions were labeled by incubation of the cryosections for 30 min at room temperature with α-bungarotoxin Alexa Fluor 488 conjugated (1:1500, Invitrogen) followed by sequential steps of PBS washing, 2% PFA fixation for 5 min, PBS washing, blocking with 5% horse serum and 0.1% Triton X-100 in PBS for 30 min and incubation over night with mouse monoclonal anti-Neurofilament H (SMI32, 1:1500, Covance). Primary antibody was revealed by incubation of the samples with Alexa Fluor 568goat anti-mouse IgG (1:400, Invitrogen) for 2 h at 4 °C. Control experiments without primary antibody demonstrated that the signals observed were specific. Digital images were acquired with a microscope DMR6000B (Leica) equipped with a DFC300FX camera for bright field color pictures and a Hamamatsu ORCA-ER camera for immunofluorescence. Morphometrical analyses were performed on digital pictures of muscle fibers that were outlined manually and their cross-sectional area was determined with the public domain computer-assisted image analysis software ImageJ (NIH, USA). GraphPad Prism software was used for statistical analyses of group comparisons and frequency distribution. The effect of the mutation on LAL myofiber size was assessed using the Wilcoxon– Mann–Whitney two-sample rank-sum test. A two-tailed analysis on 2160 myofibers from 6 different muscles (three mice per genotype) was performed and significance of differences was accepted at pb0.05. Electrophysiological studies of LAL muscle Muscle preparation Mice were anesthesized with tribromoethanol (2%, 0.15 ml/10 g body weight, i.p.) and sacrificed by exsanguination. The LAL muscles were dissected with their nerve branches intact and pinned to the bottom of a 2 ml chamber, over a bed of cured silicone rubber (Sylgard, Dow Corning). Preparations were continuously superfused with a solution of the following composition (in mM): 125 NaCl, 5 KCl, 2 CaCl 2 ,1 MgCl 2 , 25 NaHCO 3 and 15 glucose. The solution was continuously gassed with 95% O 2 and 5% CO 2 , which maintained the pH at 7.35. Recording was performed at room temperature (22–23 °C). Intracellular recording The nerve was stimulated by means of a suction electrode. The stimulation consisted of square-wave pulses of 0.2–0.5 ms duration and 2–40 V amplitude at variable frequencies (0.5–100 Hz). A glass microelectrode (10–20 MΩ)filled with 3M KCl was connected to an intracellular recording amplifier (Neuro Data IR283; Cygnus Technology) and used to impale single muscle fibers near the motor nerve endings. Evoked endplate potentials (EPPs) and miniature EPPs (mEPPs) were recorded from different NMJs within the muscle as described previously (Ruiz et al., 2008). Muscular contraction was prevented by including 3–4μMμ-conotoxin GIIIB (Alomone Laboratories) in the bath, a specific blocker of muscular voltage gated sodium channels. Data analysis The mean amplitudes of the EPPs and mEPPs recorded at each NMJ were linearly normalized to −70 mV resting membrane potential. EPP amplitudes were corrected for nonlinear summation (McLachlan and Martin, 1981). Quantal content (QC) was estimated by the direct method, which consists of recording mEPPs and EPPs (nerve stimulation 0.5 Hz) simultaneously and then calculating the ratio: QC=Average Peak EPP/ Average Peak mEPP. To evaluate the readily releasable pool (RRP) size the QC values during 1 s train of stimuli at 100 Hz were plotted against the cumulative number of quanta. The linear portion of the curve was back extrapolated. The x-intercept gives an estimate of RRP based upon the assumption of negligible mobilization of vesicles into the RRP (Elmqvist and Quastel, 1965) during the first 10–12 stimuli (Ruiz et al., 2011). All electrophysiological data are given as group mean values± SEM, with n being the number of muscles fibers per group. A total of 33 myofibers from 7 muscles (3 control and 3/4 Nmd 2J mice) were analyzed. Measurements from different muscle fibers were grouped by mouse genotype. Statistical comparisons between Nmd 2J and control measures were made using Student's unpaired two-tailed t-test. Results were considered statistically different when the p value was b0.05. Additionally, we also analyzed the data with ANOVA including the mouse as a covariate for the data in order to detect whether particularly influential mouse's fibers drove the effect. No significant differences were found among groups, indicating that the influence of the individual mouse of a covariate was not significant in these series of experiments. 171F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 Results Early loss of motoneurons in lumbar spinal cord and proximal motor axons in L3-6 ventral roots in Nmd 2J mutants Here, in order to study further the development of the pathophysiological cascade leading to muscular atrophy in Nmd 2J mice, we investigated the temporal relationship between motoneuron cell body loss in the spinal cord and the corresponding motor axon loss in the ventral roots. In addition, we quantified axon loss in the sciatic nerve. For this purpose, we used thy1-YFP-H tg Nmd 2J mice which allow easy tracking of motor axons and their terminal branches along the spinal cord-muscle axis. Thy1-YFP expression in motoneurons starts at P12 (Caroni, 1997), with robust expression from P16 onwards. Less than Fig. 1. Time course of motoneuron and motor axon loss in spinal cord, ventral roots, and sciatic nerve in Nmd 2J mice. (A) Number of YFP positive motoneurons is significantly reduced in P16 and P21 old Nmd 2J mice (at P16 1.7±0.1 versus 0.9±0.1, **p=0.0021, t=5,159, df=6 for 4 control and 4 Nmd 2J mice; and at P21 2.2±0.4 versus 0.8±0.1, *p=0.0224, t=3.618, df=4 for 3 control and 3 Nmd 2J mice). The reduction of the motor axon number keeps a comparable level over the ventral roots (at P16 16.6±0.4 versus 9.1±0.3, ***pb0.0001, t=16, df=6 for 4 control and 4 Nmd 2J mice; and at P21 17.9±0.7 versus 8.0±0.5, ***p=0.0004, t=11,21, df=4 for 3 control and 3 Nmd 2J mice) (B) up to the sciatic nerve (at P16 63.8±6.5 versus 43.6±2.0, *p=0.0499, t=2.572, df=5 for 4 control and 3 Nmd 2J mice and at P21 73.7±1.7 versus 48.3±3.8, **p=0.0038, t=6.015, df=4 for 3 control and 3 Nmd 2J mice) in (C). In all three neuronal compartments in Nmd 2J mice the relative reduction of motor fibers is constant. Representative pictures are shown for P21. Statistical analysis was performed using Student's unpaired two-tailed t-test. Bars represent mean±SEM, significance is indicated by stars (*pb0.05; **pb0.01; ***pb0.001). 172 F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 10% of all motor axons of the thy1-YFP-H tg mouse line are labeled (Feng et al., 2000), allowing reliable analysis of the arborisation of axonal processes from individual motoneurons. The number of labeled axons observed was consistent with the quantitative analysis of YFP positive motoneuron cell bodies in the spinal cord of the examined control mice (P16 and P21). On average, 1–2 cell bodies per 15 μmcryosection express YFP, thus the obtained data are not comparable to the data from (Grohmann et al., 2004), where the absolute number of spinal Fig. 2. Time course of fiber atrophy in gastrocnemius and diaphragm muscle from Nmd 2J mice. (A–C) Fiber caliber size distribution in the gastrocnemius muscle at three different time points (P10, P21, P42) in Nmd 2J and control mice. Neurofilament (magenta), dystrophin (green) and BTX (red) stainings are depicted in (D–F) for control mice and in (G–I) for Nmd 2J mice. The fiber caliber size distribution of the diaphragm from Nmd 2J and control mice at P10, P21 and P42 is shown in (J–L). The mean caliber size of control fibers is determined. First value of each graph is 50% less than the mean value of the control fibers. The final value is higher than 150% of the mean value of the control fibers. The values of the statistical differences are indicated in the material and method part. 173F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 motoneurons was determined. We included cryosections from L3 to L6 of the spinal cord into our counts and compared the absolute number of YFP positive cell bodies from control and Nmd 2J mice at P16 and P21. At both time points, YFP positive spinal motoneuron cell bodies are significantly reduced in Nmd 2J mice (Fig. 1A; at P16, the reduction was 1.7± 0.1 in control versus 0.9±0.1 in Nmd 2J mice, and at P21 2.2±0.4 in control versus 0.8±0.1 in Nmd 2J mice, with a representative image for controls). We subsequently determined the number of YFP positive axons in the ventral roots in control and Nmd 2J mice and found a reduction of YFP positive motor axons that parallels the loss of motoneurons (at P16 16.6±0.4 versus 9.1±0.3 and at P21 17.9±0.7 versus 8.0± 0.5) (Fig. 1B with representative images). We then investigated axons within the sciatic nerve. The sciatic nerve contains sensory and motor fibers which hardly can be discriminated. Quantitative analysis shows a Fig. 3. Axon arborization and sprouting at P16 and P21 in Nmd 2J gastrocnemius muscle. (A, B) YFP positive motor axons from controls at P21 show similar axon course (A) but thicker axon bundles in the medial part at the muscle entry in contrast to Nmd 2J (B). In Nmd 2J mice the number of YFP positive motor axons is declined (A2 and A5, B2 and B4; insets) and the arborization is less (A1, A6, B1, B5), but the percentage of the frequency of small additional sprouts (A1, A3, A4, A6, B1, B3, B5, yellow arrow heads) is increased (Table 1). For P21, 9 control and 6 Nmd 2J mice were analyzed; and for P16, muscles from 8 control and 3 Nmd 2J mice were taken. 174 F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 significant reduction of YFP positive axons both at P16 (63.8±6.5 versus 43.6±2.0) and P21 (73.7±1.7 versus 48.3±3.8) (Fig. 1C with representative images). These findings indicate that motoneuron cell body loss corresponds to motor axon loss along the whole spinal cord-motor nerve axis even at early disease stages (P16 and P21). Fig. 4. Coexistence of intact and denervated neuromuscular endplates in the gastrocnemius muscle from Nmd 2J mice between P10 and P21. α-Bungarotoxin, synaptophysin and neurofilament are similar distributed in P10 control (A1–A4) and Nmd 2J motor endplates (B1–B4). The bar diagram in (J) depicts no markedly size difference of intact presynapses between Nmd 2J and control NMJs (31.9±1.1 μm versus 30.7±1.3 μm, p=0.5188, t=0.7065, df=4 for 3 control and 3 Nmd 2J mice). (C1–D4) shows that Bungarotoxin, synaptophysin and neurofilament are equally distributed in control NMJs and intact Nmd 2J NMJs at P16. The presynaptic NMJ size of Nmd 2J muscles is similar to that analyzed from controls (52.4±1.5 μm versus 54.9±3.6 μm, p=0.5559, t=0.6418, df=4 for 3 control and 3 Nmd 2J mice) (K). Only some of NMJs exhibit “degenerative signs”such as degenerating axons (E3, double arrow) and dotted synaptophysin distribution (E2) in contrast to intact axon innervating the NMJ (E3, arrow). (F1–F4) displays the intracellular distribution in P21 NMJs of acetylcholine receptors (BTX in green) postsynaptically and neurofilament (magenta) as well as the presence of synaptophysin (red) presynaptically. (G1–I4) depicts the α-Bungarotoxin (green), the neurofilament (magenta) and the synaptophysin (red) localization in Nmd 2J NMJs. Nearly half of all analyzed junctions in Nmd 2J mice at P21 show a pretzel-like structure (G1–G4) and a similar distribution of acetylcholine receptors (BTX in green), neurofilament (magenta) and synaptophysin (red). (H1–H4) Some NMJs at P21 in Nmd 2J exhibit a denervating process: NMJs lose their pretzel-like structure (H1, H4), the synaptophysin stain becomes weaker (H2, H4) and axons start to degenerate (double arrows in H3). Nearly half of all examined junctions in Nmd 2J mice at P21 are denervated (I1–I4). Neurofilament as well as synaptophysin stainings are completely abolished (I2–4). The quantification of the NMJ presynapse size of Nmd 2J and control endplates in (L) discloses no reduction (76.6±8.0 μm versus 75.1±2.7 μm, p=0.8681, t=0.1770, df=4 for 3 control and 3 Nmd 2J mice). The residual α-Bungarotoxin staining (green) (I1, I4) at the postsynapse, shows a clear plaque-like structure. The single values were obtained from at least 60 endplates of 3 different mice per genotype and postnatal stage. The number of NMJs analyzed per phenotype is indicated in each bar. Statistical analysis was performed using Student's unpaired two-tailed t-test. Bars represent mean±SEM, non-significance is indicated by n.s. 175F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 Atrophy in the gastrocnemius muscle occurs at P21 in Nmd 2J mutant mice and correlates with the development of hindlimb paralysis In order to explore the correlation between motor neuron/axon loss and muscle fiber degeneration, we measured the caliber size of gastrocnemius muscle fibers at three different time points of disease progression. At P10, when Nmd 2J mice are still clinically asymptomatic, at P21, when all the Nmd 2J mice show clear disease features and motor axon loss is already prominent (see Fig. 1), and at P42, a late disease stage. The morphological studies were done mainly with the gastrocnemius muscle, since we know from previous experiments, that this muscle is made up of mostly fast-twitch muscle fibers and it is highly affected. Thus subtle morphological alterations during disease progression are well quantifiable. We also compared the data with those obtained from the diaphragm which is not affected by motor axon loss in the phrenic nerve (unpublished data and Grohmann et al., 2004). We measured the distribution of caliber sizes of muscle fibers (Fig. 2). At P10, neither the diaphragm nor the gastrocnemius muscle showed any fiber caliber size reduction (Figs. 2A, D, G, J). At P21, when the Nmd 2J mice are phenotypically affected, the fiber caliber size of the skeletal muscle shows a shift to smaller fibers (Figs. 2B, E, H) and a significantly reduced mean value (17.8±0.3 μm in control versus 14.1±1.0 μminNmd 2J mice, see also material and method part), whereas the diaphragmatic fibers are still unaffected at P21 (Fig. 2K). At a later disease stage (P42), both muscle types exhibit a clear and highly significant fiber size reduction with a shift to smaller fibers (Figs. 2C, F, I, L). Staining with antibodies against dystrophin revealed that muscle fiber membranes of Nmd 2J mutant fibers in the gastrocnemius muscle are intact at all time points (Figs. 2G–I) similar to the control situation (Figs. 2D–F). The same result was obtained for muscle fibers in the diaphragm (data not shown). Muscle fibers also exhibit intact postsynaptic structures, as revealed by BTX stainings which in manycasescorrelateswithmotorfiber presynaptic innervation, as Fig. 5. Ultrastructural alterations in the gastrocnemius muscle of NMJs from P21 Nmd 2J mouse and controls. (A, B) NMJs in control mice show vesicleand mitochondria-filled terminals (T) opposed to postsynaptic membrane areas displaying well-developed (large arrows in A, B) and, less frequently, irregular and/or shallow junctional folds (double arrow in A) and terminal Schwann cells (SC) covering the terminals with intertwined thin processes (small arrows in A). (C–I; box in D shown at higher magnification in E; H, I: serial sections) Some NMJs in P21 Nmd 2J mice show comparatively normal morphology (C–E) with somewhat less elaborate junctional folds than observed in the controls (double arrows in C, arrows in D). Frequently, postsynaptic areas without associated terminals and junctional folds displaying differential degrees of elaboration are observed (arrows in F, G). Occasionally, NMJs with signs of preand postsynaptic degeneration are found (H, I: serial sections of one NMJ) with subsynaptic membrane (arrowheads) showing only individual normal junctional folds (large arrows) and vacuolized subsynaptic sarcoplasm, and Schwann cells extending narrow processes (small arrows) into the widened synaptic cleft (double arrows) and completely enveloping a terminal (T) or parts of the pre-terminal axon (Schwann cell inclusion, SCI; I). Intramuscular nerve fascicles in Nmd 2J mice (K, L, M), which in control (J) mice contain numerous well-myelinated large-caliber axons, show individual normal appearing myelinated axons (K) in addition to signs of degeneration with Schwann cells containing myelin debris (arrow in L) or devoid of axons and myelin resembling Büngner bands (SC/arrow in M, SC) and abundant endoneurial collagen (EN; K, M). Gastrocnemius muscles from 4 control and 3 Nmd 2J mice were analyzed, images A, B and C–I are from specimen of one control and one Nmd 2J mouse, respectively. Bars in A, C, D, H, and I: 2 μm, in B, E, and G: 500 nm, PN: perineurium. Bars in: K, L, M: 1 μm, in E: 5 μm, in J, F: 2 μm. 176 F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182 shown by Neurofilament staining. At later disease stages, denervated postsynapses (arrows in Fig. 2I) could be monitored (see also below). Altered motor axon terminal arborization in the gastrocnemius muscle of Nmd 2J mice To follow the YFP-positive motor axons to their terminals at the gastrocnemius muscle, we investigated axonal telodendron elaboration between P16 and P21, the time points at which loss of YFP positive motoneuron cell bodies and proximal axons is ~50% (see Figs. 1A, B). We studied the medial part of the gastrocnemius muscle at P16 and P21. Images from P21 are depicted in Fig. 3, quantification data from P16 and P21 are shown in Table 1. The presynaptic compartments and axon bundles from representative mice at P21 are shown in Fig. 3A (control), and in Fig. 3B(Nmd 2J ). The motor axon course in Nmd 2J mutant muscles was not altered at P16 or P21, but the number of YFP positive axons seemed to be reduced (Figs. 3B2, B4) in comparison to controls (Figs. 3A2, A5). However, quantification of terminal arborization/sprouting in both genotypes (Figs. 3A3, A6 for controls and B1, B3, B5 for Nmd 2J mice, indicated by yellow arrow heads) revealed an increasing number of YFP positive axons which end up in more than 3 NMJs in the mutant situation between P16 and P21 (Table 1). This finding is in agreement with the report by Ruiz et al., who monitored a significant enlargement of motor units in Nmd 2J mice (Ruiz et al., 2005), and indicates that compensatory mechanisms occur in Nmd 2J mice, similar as in mouse models of mild forms of spinal muscular atrophy such as SMA type III models (Simon et al., 2010). Denervated and intact motor endplates between P10 and P21 in Nmd 2J mice The capability for sprouting requires an intact nerve ending. The presence of presynaptic terminals does not necessarily imply that the synapse functions normally. As a first step, we therefore started a morphological analysis of the NMJs of the gastrocnemius muscle from Nmd 2J and control mice at P10, before paralysis becomes apparent, and at P16 and P21, when first pathological signs have been manifested. The shape of the NMJs from control and Nmd 2J mice at P10, a stage where motoneuron cell loss in Nmd 2J mice is already detectable, was more plaque-like (Figs. 4A, B), while the shape in control NMJs and in intact NMJs from Nmd 2J mice at P16 and P21 was almost pretzel-like (Figs. 4C, D, F, G). We used confocal microscopy to investigate distribution of synaptic vesicles at the NMJs visualized by α-synaptophysin staining. In parallel, preterminal and terminal axon course was investigated by α-neurofilament staining, and the postsynaptic compartment by staining of acetylcholine receptors with α-Bungarotoxin. Quantification of presynaptic areas of innervated NMJs from P10 Nmd 2J mice and controls revealed no significant differences (Fig. 4J). The NMJ shape and the subcellular distribution of axonal neurofilament, presynaptic synaptophysin, and acetylcholine receptors at postsynaptic sites were also unchanged (Figs. 4A1-B4), confirmed by an unchanged transcript level of the acetylcholine receptor at P10 in Nmd 2J mice (data not shown). At P10, a time point when polysynaptic innervation is still detectable (Brook and Duchen, 1990), we quantified the relation of innervated, denervated and “degenerative”NMJs in Nmd 2J mutant muscles and found that 34% of the NMJs were denervated and 2% exhibit degenerative signs (Table 2). At day 16 after birth, NMJs in control mice showed an approximately pretzel-like shape (Figs. 4C1-C4); and at three weeks after birth, NMJs are normally fully developed and show a presynaptic and postsynaptic pretzellike structure (Figs. 4F1-F4). The dotted synaptophysin distribution in the presynaptic compartment shown in Figs. 4C2,C4andF2,F4 represents the vesicle localization relative to the postsynapse (Figs. 4C1, C4 and F1, F4). At both developmental stages (P16 and P21) about half of the NMJs in Nmd 2J mice presented normal structural traits (Table 2). First and surprisingly, numerous innervated NMJs were observed which display continuous axonal neurofilament staining and normal pretzel-like synaptophysin and acetylcholine receptor distribution, thus representing intact terminal axons, presynaptic terminals, and postsynapses (Figs. 4D1-D4 and G1-G4). The size of these intact presynapses did not vary between mutant and control NMJs at P16 and P21 (Figs. 4KandL); however some of the documented Nmd 2J mutant NMJs (3–4%) showed signs of ongoing axonal degeneration (Table 2), indicated by fragmented magenta neurofilament staining (Figs. 4E3 and H3, double arrows) and altered synaptophysin distribution (Figs. 4E2 and H2). Other NMJs completely lacked innervation and displayed a plaque-like morphology (Figs. 4I1-I4). Quantification of the relative number of NMJs in each subgroup between P16 and P21 in Nmd 2J mice showed that the percentage of intact endplates was between 50% and 55%, completely denervated NMJs between 42% and 46%, while the NMJs displaying “degenerative signs”represented a minority of 3–4% of all examined endplates (Table 2). The slight but statistically not significant increase in number of intact endplates (50% at P16 and 55% at P21 in Table 2) is probably due to increased sprouting (Table 1). The results from this morphological analysis are summarized in Table 2.FromP10toP21NMJsofNmd 2J mice showed an increase of denervated NMJs (34% to 42%) and a decrease of intact NMJs (64% to 55%). NMJs with obvious “degenerative”signssuchasa“dotted”synaptophysin staining and fragmented axons at the presynapse keep a constant level from 2–4% along the time course. These results indicate that the degenerative process at Nmd 2J mutant NMJs is hard to visualize, since only a minority of NMJs displays degenerative presynaptic features whereas the majority of the endplates is intact or completely denervated. In controls no NMJs with degenerative signs are detectable, only a minority of 3–6% which are not α-neurofilament positive are observed (data not shown). We also performed ultrastructural analyses of the gastrocnemius muscle at P21 when muscle weakness is manifest. In control muscle, NMJs generally showed presynaptic terminals filled with vesicles and mitochondria, postsynaptic areas with deep basal lamina-lined junctional folds, and terminal Schwann cells (SC) extending numerous thin processes to cover the terminals (Figs. 5A, B). Junctional fold areas not associated with terminals or containing few and/or shallow folds indicative of synaptic rearrangement (Cardasis and Padykula, 1981) or immaturity (Noakes et al., 1995) were rare (Fig. 5A, double arrow). In Nmd 2J mutant gastrocnemius muscle, fiber degeneration was observed and few NMJs or NMJ remnants were encountered in ultrathin sections. In specimen derived from one Nmd 2J mouse, NMJs could be classified into roughly three morphological categories. Firstly, comparatively normal NMJs were found displaying vesicle-filled Table 1 Quantification of axonal sprouting in control and Nmd 2J mice. NMJs/axon P 16 P21 Control Nmd 2J Control Nmd 2J 1 49 (79%) 10 (56%) 246 (85%) 9 (13%) 2 11 (18%) 4 (22%) 42 (15%) 24 (35%) 3 2 (3%) 2 (11%) 0 16 (23%) 4 0 2 (11%) 0 7 (10%) ≥5 0 0 0 13 (19%) Table 2 Quantification of NMJ morphology in control and Nmd 2J mice. NMJ morphology P10 P16 P21 Control Nmd 2J Control Nmd 2J Control Nmd 2J Intact 178 (97%) 82 (64%) 97 (94%) 60 (50%) 62 (95%) 33 (55%) Denervated 6 (3%) 44 (34%) 6 (6%) 55 (46%) 3 (5%) 25 (42%) Degenerating 0 3 (2%) 0 5 (4%) 0 2 (3%) 177F. Krieger et al. / Neurobiology of Disease 54 (2013) 169–182