Intrinsic neuronal determinants locally regulate extrasynaptic and synaptic growth at the adult neuromuscular junction

Abstract

Long-term functional plasticity in the nervous system can involve structural changes in terminal arborization and synaptic connections. To determine whether the differential expression of intrinsic neuronal determinants affects structural plasticity, we produced and analyzed transgenic mice overexpressing the cytosolic proteins cortical cytoskeleton-associated protein 23 (CAP-23) and growth-associated protein 43 (GAP-43) in adult neurons. Like GAP-43, CAP-23 was downregulated in mouse motor nerves and neuromuscular junctions during the second postnatal week and reexpressed during regeneration. In transgenic mice, the expression of either protein in adult motoneurons induced spontaneous and greatly potentiated stimulus-induced nerve sprouting at the neuromuscular junction. This sprouting had transgene-specific features, with CAP-23 inducing longer, but less numerous sprouts than GAP-43. Crossing of the transgenic mice led to dramatic potentiation of the sprout-inducing activities of GAP-43 and CAP-23, indicating that these related proteins have complementary and synergistic activities. In addition to ultraterminal sprouting, substantial growth of synaptic structures was induced. Experiments with pre- and postsynaptic toxins revealed that in the presence of GAP-43 or CAP-23, sprouting was stimulated by a mechanism that responds to reduced transmitter release and may be independent of postsynaptic activation. These results demonstrate the importance of intrinsic determinants in structural plasticity and provide an experimental approach to study its role in nervous system function.

Figure 1

Expression of CAP-23 in mouse motor nerves and at the neuromuscular junction. (A) At the neuromuscular junction, CAP-23 immunoreactivity is downregulated during the second postnatal week. The figure shows double-labeling immunocytochemistry of cryostat sections from formaldehyde-fixed gluteus maximus muscle (equivalent photographic exposures). To visualize neuromuscular junctions, the sections were counterstained with RITC–α-bungarotoxin. CAP-23 immunoreactivity was well detectable at P8 and nearly undetectable at P14. In parallel experiments, a combination of α-bungarotoxin and antibody to neurofilament-160 labeled all CAP-23– positive structures (data not shown), suggesting that in postnatal muscles, CAP-23 expression was restricted to intramuscular nerves. (B) CAP-23 is reexpressed in regenerating intramuscular nerves in the adult. Double-labeling immunocytochemistry of gastrocnemius sections from a control (CON) and a mouse 10 d after mid-thigh level crush of the sciatic nerve (REG). Note that neurofilament-160 (NF-160)–positive nerves do not express CAP-23 in the adult but reexpress this protein during regeneration. These findings indicate that CAP-23 is a growth-associated protein of motor nerves. Bar, 23 μm.

Figure 2

A mouse Thy1.2-based expression cassette drives expression of transgenic chick CAP-23 in adult mouse neurons, including spinal motoneurons. (A) Neuronal expression of chick CAP-23 transgene. In situ hybridization of adult mouse (line CAP-23[C11]) brain and spinal cord cryostat sections with digoxigenin-labeled chick CAP-23 cRNA. (Left) Strong transgene expression was detected in several neuronal types in the hippocampal formation (e.g., dentate gyrus granule cells, hilar cells, CA1; low in CA3) and in thalamic nuclei. No signal was detected with a corresponding sense probe (center). (Right) Strong transgene expression in spinal cord neurons (lumbar level), including large ventral horn motoneurons (arrows). Note absence of signal in the white matter (oligodendrocytes, astrocytes). (B) Transgene expression levels in adult mouse brain were comparable to those detected for endogenous CAP-23 in E17 chick brain, when levels of this protein are maximal. The immunoblot of brain homogenate fractions (40 μg of protein) was probed with monoclonal antibody 15C1, which specifically detects chick, but not mouse (nontransgenic sample) CAP-23. The transgenic lines were CAP-23(C11), CAP-23(C13), and CAP-23(C17). (C) Detection of transgenic, but not endogenous CAP-23 at the neuromuscular junction of a CAP23(C11) transgenic mouse. Double-labeling immunocytochemistry for CAP-23 and α-bungarotoxin. (Left) Section reacted with monoclonal antibody 15C1. (Right) Section reacted with antiserum against carboxyl-terminal sequence from mouse CAP-23 (no crossreactivity with chick CAP-23). Bar, 45 μm.

Figure 2

A mouse Thy1.2-based expression cassette drives expression of transgenic chick CAP-23 in adult mouse neurons, including spinal motoneurons. (A) Neuronal expression of chick CAP-23 transgene. In situ hybridization of adult mouse (line CAP-23[C11]) brain and spinal cord cryostat sections with digoxigenin-labeled chick CAP-23 cRNA. (Left) Strong transgene expression was detected in several neuronal types in the hippocampal formation (e.g., dentate gyrus granule cells, hilar cells, CA1; low in CA3) and in thalamic nuclei. No signal was detected with a corresponding sense probe (center). (Right) Strong transgene expression in spinal cord neurons (lumbar level), including large ventral horn motoneurons (arrows). Note absence of signal in the white matter (oligodendrocytes, astrocytes). (B) Transgene expression levels in adult mouse brain were comparable to those detected for endogenous CAP-23 in E17 chick brain, when levels of this protein are maximal. The immunoblot of brain homogenate fractions (40 μg of protein) was probed with monoclonal antibody 15C1, which specifically detects chick, but not mouse (nontransgenic sample) CAP-23. The transgenic lines were CAP-23(C11), CAP-23(C13), and CAP-23(C17). (C) Detection of transgenic, but not endogenous CAP-23 at the neuromuscular junction of a CAP23(C11) transgenic mouse. Double-labeling immunocytochemistry for CAP-23 and α-bungarotoxin. (Left) Section reacted with monoclonal antibody 15C1. (Right) Section reacted with antiserum against carboxyl-terminal sequence from mouse CAP-23 (no crossreactivity with chick CAP-23). Bar, 45 μm.

Figure 3

Neuromuscular junctions of CAP-23–overexpressing mice exhibit spontaneous nerve sprouting, with features distinct from those induced by GAP-43. Data are from 4–6-wk-old mice; gluteus maximus muscle. (A) Ultraterminal nerve sprouting in a CAP-23(C11) and a GAP-43(wt3) mouse. The combined silver-esterase reaction visualizes nerves (black) and acetylcholine esterase reaction product (blue; delimiting synaptic area). Note long ultraterminal sprouts in the presence of CAP-23 (arrows) and shorter sprouts (arrows; sprout at the bottom with growth cone structures) in the presence of GAP-43. (B) Contents of denervation-sensitive mRNA (γ-subunit of AChR) in skeletal muscle of nontransgenic and CAP-23(C11) mouse. As in nontransgenic animals, the gluteus of CAP-23(C11) mice did not express genes induced by the absence of electrical activation (e.g., BotA–induced paralysis [8 d; paralyzed]). (C) Quantitative analysis of ultraterminal sprouting in nontransgenic and CAP-23(C11) mice; comparison with the sprouting reaction induced in the same type of muscle by local paralysis (8 d) with Bot-A. N = 8. (D) Quantitative analysis of nerve sprouting patterns at the neuromuscular junction of nontransgenic (Con), CAP-23(C11), CAP-23(C2), GAP-43(wt3), and GAP-43(wt2) mice. Branch points per endplate: silver-stained processes; sprouting and nonsprouting endplates included. For a description of the sampling and analysis procedures, see Materials and Methods. Note distinct features of sprouting in the presence of CAP-23 (longer, less numerous sprouts) and GAP-43 (high number of branching points per endplate). Also note that “% of endplates with sprouts,” “total length of sprouts,” “mean sprout length,” and “number of sprouts per sprouting endplate” refer to extrasynaptic branches, whereas “branch points per endplate” is a measure for branching within the endplate region, thus including intra- and ultraterminal growth. N = 5. Bar, 57 μm.

Figure 3

Neuromuscular junctions of CAP-23–overexpressing mice exhibit spontaneous nerve sprouting, with features distinct from those induced by GAP-43. Data are from 4–6-wk-old mice; gluteus maximus muscle. (A) Ultraterminal nerve sprouting in a CAP-23(C11) and a GAP-43(wt3) mouse. The combined silver-esterase reaction visualizes nerves (black) and acetylcholine esterase reaction product (blue; delimiting synaptic area). Note long ultraterminal sprouts in the presence of CAP-23 (arrows) and shorter sprouts (arrows; sprout at the bottom with growth cone structures) in the presence of GAP-43. (B) Contents of denervation-sensitive mRNA (γ-subunit of AChR) in skeletal muscle of nontransgenic and CAP-23(C11) mouse. As in nontransgenic animals, the gluteus of CAP-23(C11) mice did not express genes induced by the absence of electrical activation (e.g., BotA–induced paralysis [8 d; paralyzed]). (C) Quantitative analysis of ultraterminal sprouting in nontransgenic and CAP-23(C11) mice; comparison with the sprouting reaction induced in the same type of muscle by local paralysis (8 d) with Bot-A. N = 8. (D) Quantitative analysis of nerve sprouting patterns at the neuromuscular junction of nontransgenic (Con), CAP-23(C11), CAP-23(C2), GAP-43(wt3), and GAP-43(wt2) mice. Branch points per endplate: silver-stained processes; sprouting and nonsprouting endplates included. For a description of the sampling and analysis procedures, see Materials and Methods. Note distinct features of sprouting in the presence of CAP-23 (longer, less numerous sprouts) and GAP-43 (high number of branching points per endplate). Also note that “% of endplates with sprouts,” “total length of sprouts,” “mean sprout length,” and “number of sprouts per sprouting endplate” refer to extrasynaptic branches, whereas “branch points per endplate” is a measure for branching within the endplate region, thus including intra- and ultraterminal growth. N = 5. Bar, 57 μm.

Figure 3

Neuromuscular junctions of CAP-23–overexpressing mice exhibit spontaneous nerve sprouting, with features distinct from those induced by GAP-43. Data are from 4–6-wk-old mice; gluteus maximus muscle. (A) Ultraterminal nerve sprouting in a CAP-23(C11) and a GAP-43(wt3) mouse. The combined silver-esterase reaction visualizes nerves (black) and acetylcholine esterase reaction product (blue; delimiting synaptic area). Note long ultraterminal sprouts in the presence of CAP-23 (arrows) and shorter sprouts (arrows; sprout at the bottom with growth cone structures) in the presence of GAP-43. (B) Contents of denervation-sensitive mRNA (γ-subunit of AChR) in skeletal muscle of nontransgenic and CAP-23(C11) mouse. As in nontransgenic animals, the gluteus of CAP-23(C11) mice did not express genes induced by the absence of electrical activation (e.g., BotA–induced paralysis [8 d; paralyzed]). (C) Quantitative analysis of ultraterminal sprouting in nontransgenic and CAP-23(C11) mice; comparison with the sprouting reaction induced in the same type of muscle by local paralysis (8 d) with Bot-A. N = 8. (D) Quantitative analysis of nerve sprouting patterns at the neuromuscular junction of nontransgenic (Con), CAP-23(C11), CAP-23(C2), GAP-43(wt3), and GAP-43(wt2) mice. Branch points per endplate: silver-stained processes; sprouting and nonsprouting endplates included. For a description of the sampling and analysis procedures, see Materials and Methods. Note distinct features of sprouting in the presence of CAP-23 (longer, less numerous sprouts) and GAP-43 (high number of branching points per endplate). Also note that “% of endplates with sprouts,” “total length of sprouts,” “mean sprout length,” and “number of sprouts per sprouting endplate” refer to extrasynaptic branches, whereas “branch points per endplate” is a measure for branching within the endplate region, thus including intra- and ultraterminal growth. N = 5. Bar, 57 μm.

Figure 5

Synergism between CAP-23 and GAP-43 in nerve sprouting induction. All data from 4–6-wk-old mouse gluteus muscle. (A) Spontaneous ultraterminal nerve sprouting in GAP-43(wt3)-plus-CAP23(C11) double-transgenic mice. Combined silver-esterase reaction. Arrows point to neuromuscular junctions (thick) and sprouts (thin). (B) Quantitative analysis of spontaneous ultraterminal sprouting in heterozygous GAP-43(wt3) and CAP-23(C11), homozygous GAP-43(wt3), and double-transgenic GAP-43(wt3)- plus-CAP-23(C11) mice. N = 5. (C) Contents of denervationsensitive γ-AChR mRNA in gluteus muscle of doubletransgenic mice. Note absence of denervation signs, in spite of extensive nerve sprouting. (D) Comparison of Bot-A–induced nerve sprouting in nontransgenic, transgenic, and doubletransgenic mice. Combined silver-esterase reaction. Note dramatic growth at and behind neuromuscular junctions in double-transgenic mice. In nontransgenic mice, Bot-A induced elongation of most neuromuscular junctions and detectable nerve sprouting only at a subset of them. Some of the sprouts are indicated by arrows. Bar, 50 μm.

Figure 5

Synergism between CAP-23 and GAP-43 in nerve sprouting induction. All data from 4–6-wk-old mouse gluteus muscle. (A) Spontaneous ultraterminal nerve sprouting in GAP-43(wt3)-plus-CAP23(C11) double-transgenic mice. Combined silver-esterase reaction. Arrows point to neuromuscular junctions (thick) and sprouts (thin). (B) Quantitative analysis of spontaneous ultraterminal sprouting in heterozygous GAP-43(wt3) and CAP-23(C11), homozygous GAP-43(wt3), and double-transgenic GAP-43(wt3)- plus-CAP-23(C11) mice. N = 5. (C) Contents of denervationsensitive γ-AChR mRNA in gluteus muscle of doubletransgenic mice. Note absence of denervation signs, in spite of extensive nerve sprouting. (D) Comparison of Bot-A–induced nerve sprouting in nontransgenic, transgenic, and doubletransgenic mice. Combined silver-esterase reaction. Note dramatic growth at and behind neuromuscular junctions in double-transgenic mice. In nontransgenic mice, Bot-A induced elongation of most neuromuscular junctions and detectable nerve sprouting only at a subset of them. Some of the sprouts are indicated by arrows. Bar, 50 μm.

Figure 5

Synergism between CAP-23 and GAP-43 in nerve sprouting induction. All data from 4–6-wk-old mouse gluteus muscle. (A) Spontaneous ultraterminal nerve sprouting in GAP-43(wt3)-plus-CAP23(C11) double-transgenic mice. Combined silver-esterase reaction. Arrows point to neuromuscular junctions (thick) and sprouts (thin). (B) Quantitative analysis of spontaneous ultraterminal sprouting in heterozygous GAP-43(wt3) and CAP-23(C11), homozygous GAP-43(wt3), and double-transgenic GAP-43(wt3)- plus-CAP-23(C11) mice. N = 5. (C) Contents of denervationsensitive γ-AChR mRNA in gluteus muscle of doubletransgenic mice. Note absence of denervation signs, in spite of extensive nerve sprouting. (D) Comparison of Bot-A–induced nerve sprouting in nontransgenic, transgenic, and doubletransgenic mice. Combined silver-esterase reaction. Note dramatic growth at and behind neuromuscular junctions in double-transgenic mice. In nontransgenic mice, Bot-A induced elongation of most neuromuscular junctions and detectable nerve sprouting only at a subset of them. Some of the sprouts are indicated by arrows. Bar, 50 μm.

Figure 4

Potentiation of induced nerve sprouting in CAP-23–transgenic mice. All data from 5–8-wk-old mice; gluteus muscle or sciatic nerve. (A) Sprouting in untreated and Bot-A–paralyzed (8 d) muscle. Analysis as in Fig. 3 C (from which part of the data were replotted). N = 5. (B) Regenerative sprouting (at 64 h) distal from sciatic nerve crush. N = 4. (C) Ultraterminal nerve sprouting in chronically paralyzed (3 wk) muscle. Combined silver-esterase reaction; the plane of focus for the transgenic sample was selected to maximize visualization of sprouts (arrows, position of neuromuscular junctions; in part out of focus). Note dramatic extent of the sprouting reaction in the CAP-23(C11) mouse and very limited reaction in the nontransgenic mouse. These findings further support the conclusion that the presence of CAP-23 in motor nerves produces a true gain-of-function phenotype, leading to sustained potentiation of intramuscular nerve sprouting. Bar, 70 μm.

Figure 4

Potentiation of induced nerve sprouting in CAP-23–transgenic mice. All data from 5–8-wk-old mice; gluteus muscle or sciatic nerve. (A) Sprouting in untreated and Bot-A–paralyzed (8 d) muscle. Analysis as in Fig. 3 C (from which part of the data were replotted). N = 5. (B) Regenerative sprouting (at 64 h) distal from sciatic nerve crush. N = 4. (C) Ultraterminal nerve sprouting in chronically paralyzed (3 wk) muscle. Combined silver-esterase reaction; the plane of focus for the transgenic sample was selected to maximize visualization of sprouts (arrows, position of neuromuscular junctions; in part out of focus). Note dramatic extent of the sprouting reaction in the CAP-23(C11) mouse and very limited reaction in the nontransgenic mouse. These findings further support the conclusion that the presence of CAP-23 in motor nerves produces a true gain-of-function phenotype, leading to sustained potentiation of intramuscular nerve sprouting. Bar, 70 μm.

Figure 6

Increased neuromuscular synaptic areas in GAP43– and CAP-23–overexpressing mice. RITC–α-bungarotoxin labeling of gluteus muscle fiber whole mounts (4–6-wk-old mice). (A) Representative examples of neuromuscular junction configurations in nontransgenic and CAP-23(C11) mice. Part of the nontransgenic endplate is out of focus, but the entire long-axis is in focus. Note dramatic increase in branching index and synaptic area in the transgenic mouse. In contrast, the external dimensions of the synaptic region did not differ significantly between transgenic and nontransgenic mice. (B) Representative examples of neuromuscular junction configurations in nontransgenic and GAP-43(wt3) mice, with and without Bot-A–induced paralysis. Note that paralysis did not induce features of transgenic endplates (e.g., high branching index) in nontransgenic mice. Bar: (A) 15 μm; (B) 29 μm.

Figure 7

GAP-43– and CAP-23–transgenic mice reveal a sproutinducing mechanism sensitive to reduced transmitter release that may be independent of postsynaptic activation. Data from gluteus (A and B) or gastrocnemius (C) muscle of 4–7-wk-old mice. (A) Subparalyzing doses of Bot-A induce ultraterminal nerve sprouting and elevated contents of neurofilament-positive branch points at the neuromuscular junction. Nontransgenic (Con), GAP-43(wt3), and CAP-23(C11) mice were analyzed. 1 pg of Bot-A induced local paralysis in all mice, 0.5 pg failed to induce detectable paralysis signs in >80% of the mice, and 0.1 pg induced no detectable paralysis. Note that 0.05 pg of toxin was already highly effective in inducing sprouting and endplate branching (these two parameters were correlated) in GAP-43– and CAP-23–overexpressing mice. N = 6. (B) Time-course of Bot-A– induced nerve growth at the neuromuscular junction of GAP43(wt3) mice. Sprout length: average length of the longest sprout per sprouting endplate. Note that subparalyzing doses of the toxin rapidly induced endplate branching (and the emergence of short ultraterminal sprouts; data not shown), whereas paralysis promoted sprout elongation with comparatively slow kinetics. N = 4. (C) Endplate nerve branching is induced by blockade of transmitter release in the presence of either Bot-A (1 pg) or tetrodotoxin (TTX), whereas paralyzing doses of α-bungarotoxin (αBgtx) failed to induce this reaction. In contrast, sprout elongation was induced by all paralysis-inducing protocols. These experiments were carried out in GAP-43(wt3) mice. N = 4.