The definitive feature of MNs is their ability to form functional neuromuscular junctions and thereby drive the contraction of skeletal muscle cells. Our study provides critical evidence that ESC-derived MNs can exhibit robust synaptic communication with muscle cells under simplified in vitro culture conditions. Empirically, this is a significant observation, as the use of stem cell-derived MNs for regenerative purposes or disease modeling requires that the cells faithfully mimic their natural counterparts in both molecular and functional properties. Our data show that ESC-derived MNs express several proteins, including nicotinic ACh receptors, Slc18a3 (VAChT), the high affinity choline transporter Slc5a7, and SNARE proteins found at native neuromuscular junctions, and exhibit both spontaneous and action potential-dependent, multi-quantal secretion of ACh to trigger post-synaptic potentials and muscle contraction. These results further provide an important extension to previous studies that have used bath application of glutamate to evoke post-synaptic potentials and muscle contraction in high-density MN-muscle cell cultures 
. Moreover, the ability to quantify the functional properties of individual nerve-muscle contacts offers the opportunity to rigorously assess the impact of a variety of experimental manipulations on these synaptic events.
To define the synaptic activity of MN-muscle pairs, our investigation was intentionally restricted to the differentiated progeny of mESCs and C2C12 muscle cells. Our data indicate that MN-muscle synapses formed under these simplified conditions recapitulate many features of neuromuscular communication seen in vivo. For example, the frequency of spontaneous mEPCs (0.04–0.3 Hz) is within the range reported for muscle fibers of late embryonic and early post-natal rodents 
. Similarly, the rise and decay times of EPCs are within the range observed for developing rat neuromuscular junctions 
. It is important to note that functional MN-muscle synapses formed under these conditions typically survived for a maximum of 8–9 days in culture. This limitation might reflect a lack of support provided in vivo by other cells including astrocytes and Schwann cells or presynaptic inputs to the MNs from spinal interneurons. The low-density culture conditions established in this study provide a suitable platform for evaluating the influence of different cell types in future work. Nevertheless, we have found that similar co-culture of human ESC and IPSC-derived motor neurons with muscle cells results in synaptic contacts that persist for several weeks (JAU, KLA, and BGN, unpublished data), suggesting that at least some aspects of synaptic stability are inherent to the MNs themselves and highly variable between species.
During embryonic development, different classes of MNs exhibit a high degree of selectivity in their choice of muscle targets 
. However, we infer from the present results that the programs that dictate motor innervation patterns are sufficiently malleable such that ESC-derived MNs can form functional synapses on C2C12 cells. Although this observation is not surprising given the promiscuity of mammalian MNs for forming neuromuscular junctions in vitro 
, precise matching of MN and muscle subtypes might nevertheless be crucial for ensuring full synaptic activity and stability 
. Progress has recently been made in understanding the mechanisms underlying MN fate selection 
, and it should be fruitful to determine whether this information can be harnessed to bias the differentiation of mESC-derived MNs to favor the innervation of specific classes of muscle cells both in vitro and in vivo.
Another important use for this co-culture system will be for modeling neuromuscular disorders. There is abundant evidence for an early and profound impairment of neuromuscular transmission in amyotrophic lateral sclerosis 
, and we showed previously that mutant forms of superoxide dismutase 1 (SOD1) alter the morphology and survival of hESC-derived MNs in vitro 
. Consequently, conditional expression of mutant SOD1 in MN-muscle co-cultures is likely to provide an informative system for clarifying the impact of SOD1 mutant alleles on nerve-muscle communication. Similarly, recent data suggest that proprioceptive circuits may be particularly vulnerable in spinal muscular atrophy 
. The in vitro system developed here might accordingly be expanded to assess the underlying cellular and molecular mechanisms that contribute to this decline in synaptic input to MNs. Thus, in addition to their utility for helping to answer fundamental biological questions, these co-cultures have clear applications in addressing problems of medical significance.