11 February 2011. Far from being sick, motor neurons in a mouse model of spinal muscular atrophy are actually raring to fire. Unfortunately, they do not receive the signals telling them to do so. So conclude the authors of a paper in the February 10 Neuron. The work helps answer a longstanding puzzle in spinal muscular atrophy (SMA): How do people and mice get so weak, even though their motor neurons remain relatively plentiful? In fact, the study showed, those motor neurons are hyperexcitable but lack the normal complement of synapses with sensory neurons. The researchers used dissociated spinal cords to examine a reflex circuit between sensory and motor neurons. The synaptic dysfunction they found closely paralleled some of the symptoms of people with SMA.
George Mentis led the work in the laboratory of Michael O’Donovan at the National Institute of Neurological Disorders and Stroke in Bethesda, Maryland, and then in his own laboratory at Columbia University in New York. Mentis and O’Donovan, who study spinal cord development, teamed with SMA expert Charlotte Sumner, formerly of the National Institutes of Health in Bethesda and now at Johns Hopkins University in Baltimore.
SMA is caused by recessive mutations in the gene, survival of motor neuron (SMN). The disease generally arises early in infants and children, progressively weakening muscles until the child dies. Logically enough, neuroscientists have long focused on the motor neurons in SMA. But even when an SMA mouse dies, Sumner said, it has plenty of motor neurons still hooked into neuromuscular junctions. Estimates of motor neuron death, depending on the area examined, range from 60 percent (Passini et al., 2010) to none at all (Kariya et al., 2008). Motor neuron degeneration cannot account for the severe muscle symptoms, she said. Thus, the authors looked beyond the motor neuron to a start-to-finish neural circuit. “The neuromuscular junction is really just the output,” Mentis said. “The way neurons, particularly motor neurons, communicate with each other is critical.”
To analyze how that happens, the researchers examined a common SMA mouse model which misses the SMN seventh exon (Le et al., 2005). The researchers chose the 1a reflex, a simple, complete circuit they could analyze in the isolated spinal cords. This is the category of circuit that fires, for example, when a doctor taps your knee and you kick involuntarily. The 1a sensory neurons in the dorsal root ganglion (DRG) collect proprioceptive (i.e., sensory) information from muscles, which they then transmit to motor neurons in the spinal cord. Those motor neurons tell the muscles to kick, reflexively, before the brain even knows what’s going on.
Mentis examined an entire neural circuit, artificially stimulating the sensory input (akin to the knee tap) and measuring the output of all the motor neurons (the kick). He first analyzed the L1 nerves, which infiltrate the muscles affected early in SMA. Malfunction was evident in spinal cords taken from mice as young as four days old. Circuits from the sick mice produced only 15 percent of the peak amplitude of cords from wild-type mice. The motor neurons were fully capable of carrying a membrane potential, however. In fact, they were hyperexcitable compared to motor neurons from wild-type mice.
Sensory neurons were another matter. When Mentis labeled sensory cell axons that project toward motor neurons in the spinal cord, he found they were shorter than normal. Further, there were fewer synapses between the sensory and motor neurons in the SMA spinal circuits than in wild-type ones. “Finally, this gives us an explanation of why [SMA muscles] are so weak,” said Mark Rich of Wright State University in Dayton, Ohio, who was not involved in the study.
The authors reasoned that SMA’s muscle weakness stems from events upstream, because afferent neurons pull back from motor neurons. But the situation turned out to be more complex. Signal transmission fell 70 percent in four-day-old SMA mouse spines compared to normal, even though the synapse numbers dropped by only half. The researchers suspect that the remaining synapses are transmitting signals poorly, although they do not know why.
The findings indicate that loss of sensory circuits precede motor neuron loss, but it is not clear yet what starts the process. Since SMN is ubiquitous, any cell could suffer from its absence and set off a degenerative cascade. The authors are planning experiments to limit SMN expression to specific cell types in the hopes of answering this chicken-or-egg quandary. The egg in this equation would be the site of disease onset and thus the best drug target, Mentis said.
“Clinically, we do not think of sensory involvement as a prominent part of the disease,” Sumner said. But she noted that people with SMA often have reduced reflexes. Doctors generally attribute the problems to muscle weakness, but abnormal reflexes can occur in people with SMA who are not very weak. Perhaps, the study suggests, poor reflexes actually reflect deficiencies in proprioceptive or other sensory inputs.
The researchers wondered if the reflex defects in their mice would mimic a pattern seen in people with SMA: Symptoms begin in proximal muscles near the trunk and move into more distal areas (Swoboda et al., 2005). Mentis compared the L1 circuit—innervating proximal muscles—with the L5 circuit, which controls distal muscles.
In spinal cords from four-day-old animals, transmission in the L1 circuit was again strikingly affected compared to wild-type cords. However, the L5 circuit activity was normal. But at the 13-day mark, both circuits from mutant animals were affected. Reductions in the motor neuron population followed the same trend. By four days, Mentis estimated that the L1 circuit had lost 65 percent of the normal number of sensory-motor synapses, and the percentage rose to 80 percent loss at 13 days. In contrast, the L5 synapse counts were normal at four days, but dropped by half at 13 days. The results suggest that pathology in the mouse begins in proximal circuits and later progresses to distal ones, nicely mirroring human symptoms.
The study provides a thorough tour of SMA physiology, which should form a useful baseline for future research, said Arthur Burghes of Ohio Sate University in Columbus, who developed the SMN mouse model that the researchers used.
Are sensory signals the major malfunction in mice, and people, with SMA? “I certainly think it is a big problem,” Burghes said. “I am not quite willing to say it is the absolute main problem.” As the authors note, they examined only one type out of many, many circuits in the spinal cord. It is not yet known if similar problems arise in other circuits, but Sumner expects they do.
“This dysfunction of synapses at multiple places…this may become a theme in motor neuron disease,” Rich said. One study suggested that the excitotoxicity suffered by motor neurons in ALS is due to an overdose of input from spinal interneurons (see ARF related news story on Jiang et al., 2009). Unfortunately, since Mentis’s technique only works with neonatal animals (older cords cannot handle the lack of oxygen in culture), it is impossible to do the same analysis with ALS models that sicken as they age.
With failure occurring upstream of motor neurons, any future SMA treatment must address that malfunction, the authors assert. They examined spinal circuits under a drug therapy that was known to improve symptoms in mice, trichostatin A (Narver et al., 2008). As expected, many of the treated mice gained weight and were better able to right themselves. When Mentis analyzed the spines 13 days in, he found the treated animals had better reflex responses—even at L1—and more sensory-motor synapses, indicating that a potential SMA treatment does indeed address the sensory part of the circuit.—Amber Dance.
Mentis GZ, Blivis D, Liu W, Drobac E, Crowder ME, Kong L, Alvarez FJ, Sumner CJ, O’Donovan M J. Early functional impairment of sensory-motor connectivity in a mouse model of spinal muscular atrophy. Neuron. 2011 Feb 10;69:453-67. Abstract