Neuron, Vol. 8, 865-868, May, 1992, Copyright 0 1992 by Cell Press

Agrin Released by Motor Neurons Induces the Aggregation of Acetylcholine Receptors at Neuromuscular Junctions Noreen E. Reist,* Michael J. Werie, and U. j. McMahan Department of Neurobiology Stanford University School of Medicine Stanford, California 94305-5401

Summary

To test the hypothesis that agrin mediates motor neuron- induced aggregation of acetylcholine receptors (AChRs) in skeletal muscle fibers and to determine whether the agrin active in this process is released by motor neurons, we raised polyclonal antibodies to purified ray agrin that blocked its receptor aggregating activity. When the anti- bodies were applied to chick motor neuron-chick myo- tube cocultures, they inhibited the formation of AChR aggregates at and near neuromuscular contacts, demon- strating that agrin plays a role in the induction of the aggregates. Rat motor neurons, like chick motor neu- rons, induce AChR aggregates on chick myotubes. This effect was not inhibited by our antibodies, indicating that, although the antibodies inhibited the activity of chick agrin, they did not have a similar effect on rat agrin. We conclude that agrin released by rat motor neurons induced the chick myotubes to aggregate AChRs.

Introduction

The axon terminals of motor neurons induce myofi- bers to aggregate acetylcholine receptors (AChRs) and other proteins at neuromuscular junctions. Several lines of evidence (McMahan, 1990; McMahan and Wallace, 1989;Tsim et al., 1992) have led tothe hypoth- esis that the motor neuron’s inductive activity is medi- ated by agrin, a protein purified from the electric or- gan of Torpedo californica. For example, when agrin is applied to cultured myotubes, it induces the myo- tubes to form aggregates of AChRs and other synaptic proteins. Moreover, immunohistochemistry, activity assays, and the cellular localization of agrin tran- scripts indicate that agrin is synthesized in the cell bodies of motor neurons and is transported along their axons to muscle. There is additional evidence that muscle fibers produce agrin-like proteins, which also might play a role in AChR aggregation (Reist et al., 1987; Fallon and Gelfman, 1989). The aims of the experiments we describe here were 2-fold: to deter- mine whether agrin mediates the motor neuron-in- duced aggregation of AChRs on myotubes at devel- oping neuromuscular junctions and, if so, to learn whether agrin active in this process is derived from

* Present address: Department of Molecular and Cellular Physi- ology, Stanford University School of Medicine, Stanford, Califor- “la 94705.

the motor neurons. For these studies we raised poly- clonal antibodies against T. californica agrin that

blocked its AChR aggregating activity in myotube cul- tures and examined the antibodies’ effect on motor neuron-induced AChR aggregation in cocultures of motor neurons and myotubes from chick and rat.

Results

Two rabbits were injected with polypeptides isolated from extracts of electric organ of the marine ray T. californica by immunoprecipitation with anti-agrin monoclonal antibody (MAb) 6D4 (Reist et al., 1987). SDS-polyacrylamide gel electrophoresis separates the polypeptides into four major bands. Two, con- taining agrin, are active in AChR aggregation assays, and two are not (Nitkin et al., 1987). All are probably

proteolytic fragments of a larger protein (Smith et al., submitted; Tsim et al., 1992). Each rabbit generated antiserum that recognized all four bands on immu- noblots and inhibited the AChR aggregating activity of purified ray agrin on chick myotubes in culture (Reist, 1990). The serum with the highest titer was se- lected for experiments on motor neuron-myotube co- cultures. In initial tests on cocultures, we found that this anti-agrin serum inhibited neuron-induced AChR aggregation; however, in some experiments so did the preimmune serum. Accordingly, we isolated im- munoglobulin C (IgGs) from the anti-agrin,,,, preim- mune, and normal rabbit sera by protein A column chromatography(Eyet al., 1978). The purified IgGs did not display the nonspecific inhibitory activity present in the sera. The number of AChR aggregates that spon- taneously form at low levels on myotubes cultured without motor neurons (Sytkowski et al., 1973) was not affected by anti-agrin,,, IgGs. !n myotubecultures exposed for 3 days to 0.1 mg/ml anti-agrinray IgGs, as used in motor neuron-myotubecoculturesdescribed below, the number of spontaneously formed aggre- gates per myotube was 103% k 11.4% (mean k SEM, n = 6) of that in untreated cultures. Thus the IgGs did not alter the muscle fiber’s ability to aggregate AChRs.

We found, as reported by others (Frank and Fisch- bath, 1979; Role et al., 1985), that chick motor neurons cultured with chick myotubes induced the myofibers to form AChR aggregates at and near the site of neuro- muscular contact. Many of the AChR aggregates in our chick cocultures were quite small (-1 pm in diam- eter) and were often arranged in clusters (Figure 1, top). Similar clusters of AChR aggregates were in- duced by chick motor neurons on rat myotubes and rat motor neurons on chick myotubes (Figure 1, bot- tom) as observed by De La Porte et al. (1986). These findings, taken together with the fact that agrin in- duces AChR aggregates on cultured myotubes and recent identification of similar agrin-like mRNAs in ray (Smith et al., submitted), chick (Tsim et al., 1992), and

Neuron 866

Figure 1. AChRAggregatesatand near NeuromuscularContacts in Culture

Motor neurons isolated from chick (top) or rat (bottom) splnal cords were plated onto chick myotubes. AChR aggregates (red/ orange) were labeled with rhodamine-a-bungarotoxin, and neu- rites were labeled with an anti-neurofilament MAb and fluo- rescein-conjugated goat anti-mouse IgG. Bar, IO bm.

rat (Rupp et al., 1991) motor neurons and muscles, indicate that aggregation of AChRs at neuromuscular

junctions is mediated by agrin in all three species. When we grew chick motor neurons together with

chick myotubes in the presence of anti-agrin,,, IgCs, the number of AChR aggregates at and near neuro- muscular contacts was reduced by >90% of that in cultures grown in preimmune and normal IgCs (Fig- ure 2). These findings indicate that the IgGs recog- nized chick agrin and that agrin mediates the motor neuron-induced formation of AChR aggregates on myotubes, although they do not provide any indica- tion as to whether the neurons or myotubes are the source of the active agrin.

Evidence as to the source of agrin active in AChR aggregation at neuromuscular contacts in culture came from experiments on hybrid cocultures. When we grew rat motor neurons with chick myotubes in the presence of anti-agrin,,, IgGs, the formation of AChR aggregates on the myotubes was not inhibited (Figure3, top). Thus theanti-agrin,,, IgGs did not block the AChR aggregating activity of agrin released by the rat motor neurons, nor did they prevent the chick myotubes from responding to it. This finding, coupled

Normal Preimmune

0 0.01 0.03 0.1

Amount of IgG Added (mg/ml)

Figure 2. Anti-Agrin IgGs inhibit the Motor Neuron-Induced Formation of AChR Aggregates tn Chick Cocultures

Chick motor neurons were cultured with chick myotubes tor 3 days in the presence or absence of normal rabbit, prelmmune, or anti-agrin,,, IgGs. AChRaggregates are expressed as a percent- age of control. Values represent the mean i SEM of 3-15 dishr\.

with the 90% inhibition by anti-agrin,,, IgGs in chick motor neuron-chick myotube cultures, indicates that agrin released by motor neurons induces the forma- tion of AChR aggregates at neuromuscular contacts. To test this conclusion, we grew chick motor neurons with rat myotubes in the presence of the anti-agrinra, IgGs. Consistent with the above evidence that motor neuron-released agrin mediates AChR aggregation at neuromuscular contacts, the number of AChR patches at and near these hybrid neuromuscular con- tacts was reduced by >70% of that in cultures grown in preimmune and normal rabbit IgGs (Figure 3, bottom).

Discussion

Previous studies have demonstrated that the cell bod- ies of motor neurons contain molecules recognized by anti-agrin antibodies and that such molecules are transported along their axons to muscle (Magill-Sole and McMahan, 1988,199O). Moreover, extracts of frac- tions of the spinal cord enriched in motor neurons induce the formation of AChR aggregates on cultured myotubes, and the active proteins are immunoprecip- itated by anti-agrin antibodies (Magill-Solcand McMa- han, 1988). mRNA extracts of these motor neuron- enriched fractions are enriched in transcripts that encode protein both homologous to agrin and active in AChR aggregation assays (Tsim et al., 1992). At the neuromuscular junction in vivo, anti-agrin antibodies stain the synaptic basal lamina, which induces the formation of AChR aggregates on regenerating myo- fibers, as do motor axons (Burden et al., 1979; McMa- han and Slater, 1984; Reist et al., 1987). Together these results have led to the hypothesis that agrin is synthe- sized in the cell bodies of motor neurons and trans-

Agrin Induces AChR Aggregates at NMJs 067

Rat MN&hick Myotubes

I

L

normal anti-

J-

- pre-

agrin immune rabbit

Chick MNs-Rat Myotubes

I

r- normal anti-

I

Ll pre-

-

-

rabbit agrin immune

IgG (0.1 mg/ml)

Figure 3. Anti-Agrin l&s Inhibit Motor Neuron-Induced Aggre- gation of AChRs in Chick Motor Neuron-Rat Myotube Cultures but Not Rat Motor Neuron-Chick Myotube Cultures

Cultures were grown for 3 days in the presence or absence of IgCs purified from normal, anti-agrin,,,, or preimmune rabbit sera. Values represent the mean k SEM of 6 dishes.

ported along their axons to muscle where it is released into the synaptic cleft to bind both to the basal lamina and to the myofibers, causing the myofibers to form aggregates of AChRs. Our finding that antibodies that inhibit agrin activity also inhibit the activity of the inductive molecule provided by motor neurons leaves little doubt that agrin is that molecule.

We have demonstrated in a previous report (Reist et al., 1987) that the extrajunctional basal lamina of certain muscle fibers stains with anti-agrin antibodies, and others have shown that the AChR aggregates that form spontaneouslyon muscle fibers in aneural chick Iimbsarefrequentlycoextensivewith material stained with anti-agrin antibodies (Fallon and Gelfman, 1989). These observations have raised the possibility that agrin is also produced by muscle fibers and that mus-

cle agrin might induce AChR aggregation. However, the only agrin-like mRNAs detected in muscle, thus far, code for proteins lacking regions in agrin required

for its AChR aggregating activity (Ruegg et al., 1992). These observations together with our finding that agrin produced by motor neurons induces myotubes to aggregate AChRs lead to the conclusion that if

agrin-like molecules produced by muscle play any in- ductive role in the aggregation of AChRs at neuromus- cular junctions, it is secondary to that of agrin pro- duced by motor neurons.

Experimental Procedures

Immunization Protocol Rabbits were injected with immunopunfied agrin (Nitkin et al., 1987) according to the procedures of Ribi ImmunoChem Re- search, Inc. Ribi TCM + MPL adjuvant vials were warmed to 44OC for 10 min, and 200,000 U of agrin in a volume of 2 ml was added and emulsified. Emulsion (0.25 ml) was injected subcutaneously in the neck at each of two sites and intramuscularly in the hind limbs at each of two sites. This immunization protocol was re- peated 4-5 weeks later. Immune sera were collected 7-8 days after the second injection. Preimmune serum was collected prior to the initial injections.

Purification of l&s IgCs were purified from rabbit sera by affinity chromatography (Ey et al., 1978). In brief, sera were incubated with protein A-Sepharose beads (Pharmacia) in phosphate-buffered saline (PBS) for 4-16 hr at room temperature, and IgCs were eluted with 0.1 M glycine buffer (pH 2.5) and neutralized with a l/IO solution of 1 M Tris (pH 9.0). Peak fractions (determined by OD& were pooled, loaded into Centricon 30 microconcentrators (Amicon) along with 1 ml of PBS, and spun at 5000 x g for 1 hr. After the addition of approximately 1.5 ml of PBS, they were respun, and 3-5 mg of IgC (estimated by OD,,, and ELISA titration curves) ended up in a total volume of 80-100 1.11 of PBS. IgCs were stored at -2OOC.

Cultures Motor neurons were isolated from chick or rat spinal cords as described by Dohrmann et al. (19861 and Schnaar and Schaffner (1981). Briefly, cervical and lumbar regions of spinal cords from embryonic day6 chick or embryonic day 13 rat spinal cords were removed and incubated in 0.05% trypsin for 30 mm. After gentle trituration in versene (Cibco), cells were pelleted at 100 x g through a 3.5% bovine serum albumin layer. Cells were resus- pended in versene and layered on top of a 6.8% metriramide solution and spun for 15 mtn at 400 :< g. The motor neuron fraction was collected from the interface between the versene and the metrizamide. Motor neurons were plated (75,000 per dish) onto chick or rat myotubes that had been cultured for 5 days according to standard techniques, which included inacti- vating serum complement by heat (Fischbach, 1972; Miles et al., 1987). Rat myotubes were not treated with mitotic inhibitors be- cause they remained more stably attached to the culture dishes when fibroblasts were present (Rubin, ‘1985). Myotubes were al- lowed to condition the medium for 1 day prior to the addition of neurons in order to enhance neuron survival. Neurons and myotubes were cocultured for 3 days In either the presence or absenceof IgGs. Therewas no discernibledifference in the num- ber of nerve cell bodies and their processes between cultures grown with or without the IgCs.

Assay AChR aggregates were labeled with rhodamine-a-bungarotoxin f10~8 M, Molecular Probes) for 2 hr, washed with PBS, and fixed in cold ethanol for 5 min. To label neurites, the ethanol-fixed cultures were rehydrated with PBS, Incubated with an anti-

neurofilament MAb (RT97, provided by Dr. John Wood, Sandoz Institute for Medical Research, London) for 1 hr, washed in PBS for 30 min, incubated with fluorescein-conjugated goat anti- mouse IgC (Cappel) for 1 hr, washed overnight in PBS, and dehy- drated with cold ethanol. Coverslips were mounted with Citi- fluor AF-1.

Under fluorescein optics, a field (diameter 600 Pm) containing both healthy muscle fibers and robust neurite outgrowth was located, and then, under rhodamine optics, all of the AChR ag- gregates within the field were counted. The level of AChR aggre- gation for a given culture was expressed as the mean number of AChR aggregates in ten such fields. To compare results from different platings, the mean number of AChR aggregates in IgG- containing dishes was divided by the mean number of aggre- gates in dishes grown without IgGs and expressed as a percent- age of control.

Acknowledgments

Wethank Dr. Karl Tsim for providing the immunopurified agrin, Robert Marshall and Yiran Wang-Zhu for technical assistance, and Dr. B. C. Wallace (University of Colorado School of Medi- cine) for his comments on the manuscript. This work was sup- ported by grants from the National Institutes of Health and the McKnight Foundation, gifts from Keith and Alice Linden and from Robert Sackman, and a Muscular Dystrophy Association Postdoctoral Fellowship (M. J. W.).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisemenf” in accordance with 18 USC Sec- tion 1734 solely to indicate this fact.

Received January 23, 1992; revised March 2, 1992.

References

Burden, S. J., Sargent, P. B.,and McMahan, U. J.(1979).Acetylcho- line receptors in regenerating muscleaccumulateat original syn- aptic sites in the absence of the nerve. J. Cell Biol. 82, 412-425.

De La Porte, S., Vallette, F. M., Crassi, J., Vigny, M., and Koenig, J. (1986). Pre-synaptic or post-synaptic origin of acetylcholinester- ase at neuromuscular junctions? An immunological study in het- erologous nerve-muscle cocultures. Dev. Biol. 776, 69-77.

Dohrmann, U., Edgar, D., Sendtner, M., and Thoenen, H. (1986). Muscle-derived factors that support survival and promote fiber outgrowth from embryonic chick spinal motor neurons in cul- ture. Dev. Biol. 778, 209-221.

Ey, P. L., Prowse, S. J., and Jenkin, C. R. (1978). Isolation of pure IgG,, IgG,, and lgCzb immunoglobulins from mouse serum using protein A-sepharose. Immunochemistry 75, 429-436.

Fallon, J. R., and Gelfman, C. E. (1989). Agrin-related molecules are concentrated at acetylcholine receptor clusters in normal and aneural developing muscle. J. Cell Biol. 708, 1527-1535.

Fischbach, G. D. (1972). Synapse formation between dissociated nerve and muscle cells in low density cell cultures. Dev. Biol. 28, 407-429.

Frank, E., and Fischbach, G. D. (1979). Early events in neuromus- cular junction formation in vitro. Induction of acetylcholine re- ceptor clusters in the postsynaptic membrane and morphology of newly formed nerve-muscle synapses. J. Cell Biol. 83,143-158.

Magill-Sole, C., and McMahan, U. J. (1988). Motor neurons con- tain agrin-like molecules. I. Cell Biol. 707, 1825-1833.

Magill-Sole, C., and McMahan, U. J. (1990). Synthesis and trans- port of agrin-like molecules in motor neurons. J. Exp. Biol. 753, I-IO.

McMahan, U. J. (1990). Theagrin hypothesis. Cold Spring Harbor Symp. Quant. Biol. 50, 407-418.

McMahan, U. J., and Slater, C. R. (1984). The influence of basal lamina on the accumulation of acetylcholine receptors at synap- tic sites in regenerating muscle. J. Cell Biol. 98, 1453-1473.

McMahan, U. J., and Wallace, B. C. (1989). Molecules in basal lamina that direct the formation of synaptic specializations at neuromuscular junctions. Dev. Neurosci. 71, 227-247.

Miles, K., Anthony, D. T., Rubin, L. L., Greengard, P.. and Hu- ganir, R. L. (1987). Regulation of nicotinic acetylcholine receptor phosphorylation in rat myotubes by forskolin and CAMP. Proc. Natl. Acad. Sci. USA 84, 6591-6595.

Nitkin, R. M., Smith, M. A., Magill, C., Fallon, J. R., Yau, Y.-M., Wallace, B. G., and McMahan, U. J. (1987). Identification of agrin, a synaptic organizing protein from Torpedoelectric organ. J. Cell Biol. 705, 2471-2478.

Reist, N. E. (1990). Molecules that direct the neuron-induced for- mation of postsynaptic specializations at the neuromuscular junction. Ph.D. thesis, Stanford University, Stanford, California.

Reist, N. E., Magill, C., and McMahan, U. J. (1987). Agrin-like molecules at synaptic sites in normal, denervated, and damaged skeletal muscles. J. Cell Biol. 705, 2457-2469.

Role, L. W., Matossian, V. R., O’Brien, R. J., and Fischbach, C. D. (1985). On the mechanism of acetylcholine receptor accumula- tion at newly formed synapses on chick myotubes. J. Neuroscl. 5, 2197-2204.

Rubin, L. L. (1985). Increases in muscle CaLi mediate changes tn acetylcholinesterase and acetylcholine receptor caused by mus- cle contraction. Proc. Natl. Acad. Sci. USA 82, 7121-7125.

Ruegg, M. A., Tsim, K. W. K., Horton, S. E., Kroger, S., Escher, C., Gensch, E. M., and McMahan, U. J. (1992). The agrin gene codes for a family of basal lamina proteins that differ in function and distribution. Neuron 8, 691-699.

Rupp, F., Payan, D. C., Magill-Sole, C., Cowan, D. M., and Scheller, R. H. (1991). Structure and expression of a rat agrin. Neuron 6, 811-823.

Schnaar, R. L., and Schaffner, A. E. (1981). Separation of cell types from embryonic chicken and rat spinal cords: characterization of motoneuron-enriched fractions. J. Neurosci. 7, 204-217.

Sytkowski, A. J., Vogel, Z., and Nirenberg, M. W. (1973). Develop- ment of acetylcholine receptor clusters on cultured muscle cells. Proc. Natl. Acad. Sci. USA 70, 270-274.

Tsim, K. W. K., Ruegg, M. A., Escher, C., Kroger, S., and McMa- han, U. J. (1992). cDNA that encodes activeagrin. Neuron 8,677- 689.