CD24 is expressed by myofiber synaptic nuclei and regulates synaptic transmission

Abstract

The genes encoding several synaptic proteins, including acetylcholine receptors, acetylcholinesterase, and the muscle-specific kinase, MuSK, are expressed selectively by a small number of myofiber nuclei positioned near the synaptic site. Genetic analysis of mutant mice suggests that additional genes, expressed selectively by synaptic nuclei, might encode muscle-derived retrograde signals that regulate the differentiation of motor axon terminals. To identify candidate retrograde signals, we used a microarray screen to identify genes that are preferentially expressed in the synaptic region of muscle, and we analyzed one such gene, CD24, further. We show that CD24, which encodes a small, variably and highly glycosylated, glycosylphosphatidylinositol (GPI)-linked protein, is expressed preferentially by myofiber synaptic nuclei in embryonic and adult muscle, and that CD24 expression is restricted to the central region of muscle independent of innervation. Moreover, we show that CD24 has a role in presynaptic differentiation, because synaptic transmission is depressed and fails entirely, in a cyclical manner, after repetitive stimulation of motor axons in CD24 mutant mice. These deficits in synaptic transmission, which are accompanied by aberrant stimulus-dependent uptake of AM1-43 from axons, indicate that CD24 is required for normal presynaptic maturation and function. Because CD24 is also expressed in some neurons, additional experiments will be required to determine whether pre- or postsynaptic CD24 mediates these effects on presynaptic development and function.

Fig. 1.

CD24 mRNA is concentrated in the central region of muscle. Whole mounts of diaphragm muscles from wild-type E18 embryos (A and C) and P16 mice (B, D, and E) were processed for in situ hybridization by using a probe for CD24 (A and B), AChR δ subunit (C and D), or myelin basic protein (MBP) (E), respectively. CD24 RNA, like AChR RNA, is concentrated in the central, synaptic region of embryonic and postnatal muscle. Although the level of AChR δ subunit expression decreases after birth, CD24 expression appears to increase postnatally. MBP is expressed in myelinating Schwann cells, associated with axons. Although MBP expression is enriched in the central region of the muscle, the pattern of MBP expression is distinctly different from the patterns of CD24 or AChR expression. (Scale bar: 200 μm.)

Fig. 2.

CD24 mRNA is concentrated in the myofiber synaptic region. Whole mounts of intercostal muscles from wild-type (A and C) and Isl1^cre;nrg-1^flox/− (B and D) E18 embryos were processed for in situ hybridization by using probes for CD24 (A and B) and AChR δ subunit (C and D), respectively. CD24, like AChR δ, is expressed preferentially in the central region of muscle from neuronal nrg-1 mutant embryos (B and D). As neuronal nrg-1 mutant embryos lack Schwann cells, these results indicate that CD24 is expressed by myofiber synaptic nuclei. The positions of the ribs (R) and muscle (M) are indicated in A. (Scale bar: 200 μm.)

Fig. 3.

CD24 expression is patterned in the absence of innervation. Whole mounts of intercostal muscles from E18.5 wild-type (A and C) and HB9^cre;Isl2^DTA (B and D) embryos were processed for in situ hybridization by using probes for CD24 (A and B) and AChR δ subunit (C and D), respectively. CD24 expression, as well as AChR δ expression, is patterned in the central region of muscle from wild-type and HB9^cre;Isl2^DTA mice, indicating that CD24 expression is patterned independent from innervation. The width of the CD24-expressing zone, like the AChR expressing zone, is narrower in wild-type embryos (A and C) than in embryos lacking motor neurons (B and D). (Scale bar: 200 μm.)

Fig. 4.

Neuromuscular synapses in CD24 mutant mice appear morphologically normal. Whole mounts of diaphragm muscles from wild-type (A) and CD24 mutant (B) P11 mice were stained with antibodies to neurofilament and synaptophysin to visualize motor axons and nerve terminals (green), Alexa 594–α-BGT to label postsynaptic AChRs (red), and Alexa 660–phalloidin to visualize muscle fibers (blue). In wild-type and CD24 mutant mice, the main intramuscular nerve is positioned in the middle of the muscle and oriented perpendicular to the muscle fibers. AChR clusters are concentrated at synaptic sites both in wild-type and CD24 mutant mice (A and B). The shape and size of synaptic AChR clusters are similar in wild-type and CD24 mutant mice (A and B). (Scale bar: 200 μm.)

Fig. 5.

Defects in synaptic transmission in CD24 mutant mice. (A) EPPs, which were recorded from CD24 mutant junctions in response to stimulus trains at different stimulus frequencies, reveal periods of total transmission failures at ≥20 Hz, as well as depression in EPP amplitude at the higher frequencies (calibration = 50 msec and 15 mV). No failures were detected in wild-type littermates at any frequency, as the 100-Hz example (A) illustrates. (B) Quantification of transmission parameters at CD24 mutant (black bars) and wild-type (white bars) junctions showing the number of failures within a 1-sec train at different stimulation frequencies (n = 28 cells), and the quantal content (calculated for four cells as the mean EPP amplitude/mean MEPP amplitude), MEPP frequency, and MEPP amplitude (each calculated from 12 cells with at least 500 MEPPs from each cell). Quantal content at CD24 mutant junctions was significantly depressed compared with wild-type (P < 0.005). Data are expressed as mean ± SEM. (C) The number of transmission failures during a 1 sec, 100 Hz stimulus train in CD24 mutant endplates and in CD24 mutant junctions after the introduction of an MLCK agonist peptide and a peptide that inhibits MLC (n = 28 cells). The MLCK agonist peptide completely rescued the transmission failures, whereas the MLC inhibitory peptide did not produce additional failures.

Fig. 6.

Alterations in AM1–43 uptake in CD24 mutant junctions. (A) Examples of wild-type and CD24 mutant terminals stained with rhodamine-conjugated α-BGT to visualize the endplate and AM1–43 to visualize synaptic vesicles that have been loaded after a 10-min stimulation of the muscle nerve at 10 Hz, a protocol previously shown to result in optimal labeling of the pool of cycling vesicles. Arrows indicate sites of aberrant uptake along the preterminal axon in the CD24 mutant terminals, whereas uptake at the wild-type junction is precisely confined to the junction as defined by α-BGT staining. (Scale bar: 40 μm.) (B) Quantification of AM1–43 uptake in CD24 mutant (n = 17; black bars) and wild-type (n = 17; white bars) junctions. Mean pixel intensity after a 10 sec, 10 Hz loading protocol is significantly reduced in CD24 mutant junctions (P < 0.005) (Upper). Loading for 10 min in the absence of electrical stimuli was minimal and did not differ between CD24 mutant and wild-type junctions. Data are expressed as mean ± SEM.