Fall 1996 Volume 5, Number 4
Spinal cord injury usually occurs when the spinal column, the bony encasement of the spinal cord, is fractured, causing the bones to misalign, thereby crushing the delicate spinal cord underneath. The disability that occurs is not so much because the neurons within the spinal cord are damaged, but rather that the axons that run through the area are severed so that the lower motor neurons, below the site of the injury, are disconnected from the brain.
It's quite striking that, when we look at cut or crush injuries of peripheral nerves in the arms or legs, over time we often see a reasonable degree of recovery of function. Why doesn't that happen in the spinal cord?
The reason is that axons in the peripheral nerves regenerate, but axons in the spinal cord don't. After an axon has been cut, the part detached from the rest of the nerve cell degenerates. In order for an axon to regenerate, three things must occur: The tip of the cut axon must sprout specialized structures called growth cones. The growth cones must then elongate the axon toward the target cells. After the target cell is reached, new synapses have to be formed.
A growth cone looks to me like a human hand, with the fingers being fine structures sticking out, called filopodia, that are responsible for pulling the axon along. Growth cones are very sensitive to molecules they encounter in their pathway. Some molecules promote the growth of axons and direct them toward the right target; others retard the growth of growth cones.
In peripheral nerve, sprouting of new growth cones, elongation, and synapse reformation occur very well. In the spinal cord, you see many growth cones at the site of injury, but they fail to elongate. There is good evidence that if they could elongate, they would actually reform synapses with the correct cells. Thus, the failure of axonal regeneration in the spinal cord is mainly due to a severe limitation of elongation.
Is this because of some intrinsic limitation of the axon within the spinal cord, or is something else perhaps going on outside of that axon that may inhibit the elongation process? Well, for many years, most neuroscientists believed that neurons of the spinal cord and brain were simply incapable of regenerating their axons. Pessimism ruled the day until about 1980, when Albert Aguayo and colleagues published the first in a series of papers that showed unequivocal evidence that axons could regenerate in the proper environment.
In one of their experiments (with rats) they removed a segment of spinal cord and replaced it with a piece of peripheral nerve from the sciatic nerve in the leg, one of the largest nerves in the body. Then they sewed the animal up and waited for a while. Normally after an incision, there would be a lot of sprouting of growth cones near the site of injury, but very little regrowth. But they found that the axons regenerated very well within the peripheral nerve graft, for many millimeters, even centimeters. However, when the axons reached the other end of the graft and encountered spinal cord again, they stopped.
So, experiments like these prove that the spinal cord axons do have the capacity to regenerate and elongate great distances, if they are given the right environment. And one such environment is that of the peripheral nerve.
What does a good environment for regeneration look like? If you look through a high-powered electron microscope at a section of nerve beyond the site of injury, you see that the axon's "insulation," called myelin, has disappeared. We are left with the "basal lamina," previously a nice round structure around the myelin, but now collapsed down a bit. However, it persists.
The basal lamina is continuous from the site of the injury all the way down to the muscle cell, and within the basal lamina are Schwann cells. So, the Schwann cells also remain, even though the myelin of the injured axon has degenerated, and the Schwann cells divide and align themselves within the tubes.
That is the micro environment that supports axonal regeneration. The axons regenerate within these tubes, never outside them. We now know why: Schwann cells make molecules that are very good for growth cone movement. Not only do they have molecules on their surface that encourage axonal elongation, they are responsible for secreting molecules into the basal lamina that also encourage movement. They also secrete molecules into the fluid of the tube, that encourage axonal elongation.
The spinal cord lacks Schwann cells and lacks the basal lamina, and seems to therefore lack growth-promoting molecules.
In the last few years, we've come to realize that, not only does the spinal cord lack these positive molecules, it also seems to contain molecules that inhibit the growth of axons. In the petri dish, investigators cultured regenerating axons, then added cells called "oligodendrocytes" --cells that make myelin in the spinal cord. Just after the growth cone came in contact with the oligodendrocyte, they saw that the growth cone had collapsed down into a little ball and the filopodia had retracted. The axon was unable to grow any farther.
Dr. Martin Schwab and colleagues at the University of Zurich, who did these experiments, then made an antibody --a molecule that can shut down another one --against the inhibitory molecule on oligodendrocytes. They found that if they treated the oligodendrocyte with this antibody, axonal regeneration would go right past the oligodendrocyte; it was no longer inhibitory if treated with this antibody.
In one of the most encouraging results to date, they infused the antibody into the spinal cord of adult rats. They also added growth-promoting molecules and found that some of the axons actually did go past the site of injury and elongate for many millimeters.
We are greatly encouraged by experiments such as these. They clearly show that damaged spinal cord axons can regenerate if growth-promoting molecules and cells, such as those in peripheral nerve, are added to the spinal cord, and if growth inhibitory factors in the spinal cord are blocked. New therapies for the treatment of paralysis caused by spinal cord injuries are surely on the horizon.
DR. BROWN: The cell body of a spinal motor neuron lives in the spinal
cord. It has a long projection with an axon and a surrounding myelin sheath
that ends by plugging into muscle. The motor nerve is more than just an
electrical stimulus to fire a muscle cell. It actually endows the muscle cell
with certain nutritive or "trophic" properties, so that if you cut a motor
nerve, the wasting of the muscle cell goes well beyond what you would expect if
you just didn't fire it. That's important, when we come back to one of the
diseases I'm going to talk about.Diseases of the spinal motor neuron can, to some degree, be grouped according to which part of the motor nerve cell they attack. I'd like to talk about two diseases that affect the cell body. One is poliomyelitis; the other is Lou Gehrig's disease.
Everyone knows that poliomyelitis is an infectious disease caused by a small viral agent, a poliovirus, that has an extraordinary capacity to get into the blood stream through the gut, and fight its way up backwards to the cell body of the motor neuron. There it multiplies thousands of times, to essentially burst the motor neuron. In so doing, it injures the cell enough that the axon dies back, and the muscle loses innervation. This viral infection has been one of the major infectious challenges of the human species, for as far back as we can tell. If one looks as far back as the Egyptian hieroglyphics, one can find evidence there of polio infection.
I start with polio to make the point that one of the major achievements of medical science in this century has been the virtual eradication of poliovirus infections, at least in western society.
We're not quite there yet, however, with Lou Gehrig's disease. This is another disease in which the motor neuron cell body is afflicted and dies. The motor neuron undergoes a process of shrinkage or involution followed by death --without any evidence of viral infections or other obvious causes of this shrinking process. In fact, one has the sense, looking at the cell in Lou Gehrig's disease, that a master metabolic switch has at some level just been turned off, and the cell has thereby died. That is one of the issues that we want to focus on.
What happens clinically when this occurs? As in polio, there is withdrawal of the axon, loss of muscle mass, and paralysis. In polio, there can be some recovery once the acute viral infection has passed; but in Lou Gehrig's disease, all the motor neurons are afflicted. It may be focused in a few neurons at the start, but inevitably it spreads. All of them are afflicted, all of the muscles become paralyzed in the late stages, and ultimately there is death from paralysis of breathing.
As we attempt to think about what could affect the metabolism of a cell like the motor neuron, it is useful to come back and consider the very special anatomy of this cell. The motor neuron is the Hercules of the nervous system. Although it has a fairly small cell body, maybe a tenth or a twentieth of a millimeter in diameter, it has an enormously long axon.
For example, a low back motor neuron that sends an axon out to a muscle might have an axon length that's ten to twenty thousand times the width of the cell. The way I like to translate that is to say that, if I were a motor neuron and my right arm were my axon, then the muscle that I innervate with this right arm would be somewhere over in Cambridge, three or four miles away.
That emphasizes the terrific metabolic demand that must exist on this relatively small unit to keep the integrity of the axon maintained, over time. That is important, because we should emphasize that no neurons are able to divide. The motor neurons that we have to get us through life, hopefully 80 or 90 years, are those that we have at the time of birth. So, clearly the metabolic demands on this cell type must be daunting.
Do we lose some motor neurons with age?
I spent most of the last week looking at the marathon records of three of the best ever Boston Marathoners, Johnny Kelly, Sr., Johnny Kelly, Jr., and Clarence Demar, and plotted their marathon times as a function of their ages. These runners all show some dropout or loss of speed over the first three or four decades of running. For you who are aficionados, that's on the order of three seconds, per mile, per year. However, particularly important is that, about the age of 65, there is an accelerated rate of deterioration of function.
I'm not going to claim that all of these changes are due to loss of motor neurons alone, but very qualified neuroscientists have actually counted motor neurons as a function of age and shown that, indeed, there is a dropout, and that the rate of decline is "biphasic" --that is, for the first six decades, a low rate of decline, and then a rapid rate of decline
Another sports parameter also allows one to gauge the effect of aging on muscular function: the batting average of a professional baseball player. That brings us back to Lou Gehrig himself.
Looking at the batting averages of Lou Gehrig and several age-matched Yankee controls over several years before Gehrig died, what you see is stunningly like what we saw with the marathoners. These athletes all showed some deterioration of function over time. And although Lou Gehrig was a stellar athlete, typically batting well in the .340's and .350's, his average declined over time, as did his teammates'.
But what is important is that, at a point in time, some process caused malignant demise of his motor neurons and was associated with a rapid deterioration in his athletic performance and dropoff in his batting average, such that he was forced to retire and died of respiratory paralysis within about one year.
What triggers the death process of motor neurons is the sixty-four thousand dollar question. I might add another question: What happens to start the process of motor neuron loss? We then might go on to ask, what happens after that to sustain the process and determine the slope or rate at which it occurs?
I don't know, for example, that the processes that trigger the disease are the same as those that maintain it. And of course, the related question is, if we understand those processes, can we in some way slow them down, and treat the illness?
We do have a bit of a handle on the issue of what triggers the process. It has been possible to do careful studies of individuals who inherit single gene defects that cause familial forms of Lou Gehrig's disease. In some of these families, there are mutations in a gene that is responsible for making a particular protein, called superoxide dismutase, or SOD for short.
SOD is a molecule in great abundance in every cell, in one form or another, in every organism that has been studied. Biologically, it is so abundant, that it has to be of fundamental importance to sustain life. It acts to scavenge, or detoxify, or sponge up, a free radical (a molecule of oxygen, with an extra, or free, electron). Free radicals are substances that, by virtue of having a free electron, are highly reactive. Like little molecular "Pacmen," they can interact with and destroy almost any constituent within the cell. Indeed, some believe that free radicals are the primary element that determines the aging process itself.
The SOD molecule helps sop up or scavenge this molecule, oxygen with an extra electron, by turning it into water. What happens when this molecule is mutated? Probably at least two things. One is slightly less scavenging of the free radical. But much, much, more important, a new reaction appears to be associated with the mutant molecule, a reaction that is itself toxic, which leads to some product that is murderous for the motor neuron. That is, of course, the subject of intense inquiry.
I want to conclude with one other experiment. If you take this mutant SOD molecule and express it at high levels in a mouse, extraordinarily enough, one can produce a form of Lou Gehrig's disease in a mouse. What one sees is an illness that begins in one limb in the mouse and progresses over time to involve two and then four limbs, and to go on to produce paralysis of breathing and death.
This is a model that we hope will be important. Number one, it will allow us to make molecular tools to dissect out what that abnormal chemical reaction is; and, number two, even if we don't know the chemical reaction, it ought to provide a way to test drugs that might treat the illness, independently of understanding the cause.
We've tried to present to you some of the problems inherent in understanding why spinal cord injury and motor neuron death are sometimes very difficult to treat. To be optimistic, I hope that the kinds of research that we've shown you, may lead us to the point that we will look back on these diseases and say that they are as rare as, in fact, poliomyelitis now is.
DR. SANDROCK: That's an interesting question. Some lower vertebrates do regenerate very well in the central nervous system, so if you take a goldfish optic nerve, which is also central nervous system tissue and cut it, the axons regenerate quite nicely and synapse back on the exact same cells that they had contacted before.
One clue is that some of these molecules are known to affect turning and guidance of axons, so that growth inhibitory molecules are present during development to direct axons away from certain regions. That's thought to be important in wiring up a nervous system. But why the adult mammalian nervous system actively discourages regeneration is actually not known.
Q: Is there anything to say about peripheral neuropathy and HIV or AIDS?
DR. SANDROCK: One type of neuropathy that occurs in HIV is due to problems with the myelin made by the Schwann cells. For some reason, perhaps by the direct infection of the Schwann cell, the myelin actually gets destroyed. Also the axons themselves may degenerate.
DR. BROWN: One question that I think has not been adequately explored is whether or not HIV can actually infect and reside in a nerve cell body. It's taught that HIV is a virus that usually needs to operate and multiply in a dividing cell, so one would guess that it might not be able to reside permanently in a nerve cell. I think there are now some data to suggest, in fact, that it can --it might be direct infection of the nerve cell body.