Gene Therapy for Pain

Jay Yang & Christopher L Wu. American Scientist. Volume 89, Issue 2. Mar/Apr 2001.

Pain encompasses a cruel irony One of the most universal of human sensations, pain seems nevertheless to lie beyond ordinary comprehension. Its cause has been ascribed to the activities of otherworldly forces, such as demons, gremlins and, more recently, extraterrestrial aliens, as though the condition itself were otherworldly. It is difficult, in spite of the fact that we have all felt pain, to put the sensations into words. For how, asked German philosopher Theodor Adorno, can a reasonable person express suffering “in the medium of experience,” when to do so “would be irrational by reason s own standards”?

Since suffering seems so remote to healthy people, pain-especially chronic pain-often evokes skepticism rather than sympathy. In her book, The Body in Pain, Elaine Scarry notes “…for the person in pain, so incontestably and unnegotiably present is it that `having pain may come to be thought as the most vibrant example of what it is `to have certainty,’ while for the other person it is so elusive that ‘hearing about pain,’ may exist as the primary model of what it is ‘to have doubt’.”

What results is a kind of divide between healthy people and those in chronic pain, a divide that seems to prevent society from giving pain its due. People in chronic pain are often dismissed as complainers and hypochondriacs. And even when the pain is acknowledged and addressed, the remedy is generally insufficient. Recent articles in the popular press, as well as in the professional literature, suggest that health-care workers are undertreating pain, leaving large numbers of people to suffer. Almost 60 percent of hospitalized people experience severe acute pain, and approximately 90 percent of patients with advanced cancer report pain-often chronic. A study performed by the U.S. Department of Health and Human Services revealed that approximately half of all surgical patients receive inadequate pain relief.

Part of the problem lies outside the realm of science and medicine, as there exist a number of myths and misconceptions about pain remedies-particularly about the use of opioids-that make patients fearful to use them and health-care workers hesitant to administer them. But even when pharmacological agents are administered correctly, many simply fail to do the job. Almost all of the currently available pharmacological and nonpharmacological treatments have significant limitations and side effects, especially in the treatment of certain types of chronic pain. This clinical reality demands a novel approach to pain management.

As the biological mechanism of pain is understood in ever greater detail, scientists are coming to learn that pain is a dynamic phenomenon, where the physiological relation between the painful stimulus and the behavioral output may be significantly modified over time. A good deal of progress has been made in elucidating the molecular and cellular events that give rise to neuropathic pain-that caused by damaged or diseased nerves. A number of the receptors and neurotransmitters in pain pathways have been identified, and their relationships have been defined. At the same time, gene therapy has emerged as a potentially useful treatment for many diseases. If it could be applied to the problem of pain, gene therapy might provide just the sort of rational treatment that is currently lacking. Gene therapy can be targeted specifically to components of the pain– conduction pathways in the nervous system. And gene therapies may allow health-care providers to fine tune treatments in accord with the biochemical and physiological changes brought about by persistent pain.

Such treatments, if they are, as we believe, appropriate for pain management, are a long way from becoming clinical reality. It will take a number of years and many trials before the numerous technical problems are worked out and physicians know whether such an approach can be clinically useful. But our preliminary studies in cells and laboratory animals encourage us to think that someday we may be able to add gene therapy to existing pain—treatment regimes.


Running parallel to its long history of misconception and misinterpretation is an almost 400-year history of legitimate scientific inquiry into pain’s etiology and mechanism. Such considerations began with the 17th-century philosopher, mathematician and physiologist Rene Descartes, who first proposed a link between peripheral sensation and the brain. While contemplating the mind-body connection, Descartes suggested that sensations stimulated in the body are conveyed directly to the brain, where they are actually perceived. Although this view is now considered overly simplistic, that should not diminish Descartes’s insightful realization that sensory perception is in fact a function of the brain. The Cartesian model gave rise to the notion of a hard-wired system, where pain signals were carried by fixed connections within the nervous system. This idea was reinforced by anatomical studies conducted during the 19th century and has endured, with a few modifications, until fairly recently.

In the mid-1960s Ronald Melzack and Patrick Wall of McGill University challenged the notion of a hard-wired system with their view that sensory information undergoes dynamic integration and modulation. The current view of nociceptive pain derives from this idea. Neuroscientists now think of the nervous system as plastic. They no longer believe the relay of pain information to be based on an immutable relationship between a painful stimulus and the sensory output of pain. Rather, the perception of pain results from the integration of information from a variety of sources. Of course information is relayed from the injured tissue or organ in the periphery, but the strength of this signal can be modified by emotional and behavioral information coming down from the brain, as well as by inputs from other peripheral sensations. Furthermore, biologists now think that the integration of these signals actually takes place in the spinal cord, not in the brain, and that the integrated information is then carried up to the brain for further processing.

Clifford Woolf of Harvard University and Michael Salter of the University of Toronto recently enumerated the three general levels at which neural information could be modified in response to chronic pain. They noted that the extent and duration of the response to the stimulus at the periphery could be modified. Alterations can also take place at a chemical level within any one or several of the neurons along the pain-conduction pathway. These include changes in the number or sensitivity of receptors, ion channels and internal signaling molecules. Finally, chronic pain can induce a modulation of the neurotransmitters that affect the flow of information from one neuron to the next, or it can even alter the anatomical features of these neurons and their interconnections. The set of alterations described by Woolf and Salter may lead to longterm changes in the connectivity and organization among nerve cells. This, in turn, may lead to a “pain memory,” not much different from ordinary memory in the brain.

The Pain-Conduction Pathway

Pain can be divided into several categories. Neuropathic pain is the result of direct damage of a component of the nervous system or spinal cord. Psychogenic pain has no discernable physical cause and is assumed to be psychological in origin. Nociceptive pain arises when a tissue or organ is damaged through either injury or disease. Pain can also be acute or chronic, and not infrequently, acute pain can become chronic if it is not managed correctly at its onset. In our laboratories, we are concerned primarily with chronic pain and its management.

Damage to an organ by mechanical forces, heat or chemical abrasion is registered by nerve cells that innervate the damaged organ. These are not ordinary nerve cells, however. Their endings are specialized to detect and respond to noxious stimuli. These nerve endings, called nociceptors, are also sensitive to inflammation-inducing chemicals, such as histamine, bradykinin or prostaglandins released by damaged tissues. In some cases, these inflammatory chemicals can actually cause the nociceptors to become overly sensitive to the painful stimulus and respond in an exaggerated fashion whenever the stimulus is present.

The nociceptor transmits the “pain” message to the central nervous system via an electrical impulse that travels the length of the nerve-cell fiber. This fiber carries the sensory information to relay stations within the spinal cord. Specifically, it contacts a second neuron located on the back side of the spine, a region called the dorsal horn. Contact between the two neurons is achieved chemically, through the release of neurotransmitters from the end terminal of the first neuron. These transmitter molecules traverse the space between the neurons, called the synaptic cleft and contact receptor molecules on the second neuron.

The transmitters can be pronociceptive, which means they induce pain, or they can be antinociceptive and inhibit the sensation of pain. Pronociceptive transmitters excite the postsynaptic neuron and activate a series of internal chemical events that translate the chemical signal into an electrical impulse that travels down the nerve cell fiber. When it reaches the terminal end of the nerve fiber, the electrical impulse stimulates the release of transmitter molecules that will either excite or inhibit the next neurons) in the circuit, and so on. Inhibitory transmitters either dampen the strength of an excitatory impulse or block its transmission altogether. In this way excitatory and inhibitory impulses together modulate the strength and transmission characteristics of a message carried from the periphery to the brain and vice versa.

Pronociceptive transmitters include the excitatory amino acids glutamate and aspartate, which act upon (alpha—amino-3—hydroxy-5-methyl-4-isoazoleproprionate (AMPA) receptors to bring about a relatively short-lived excitation of the postsynaptic neuron. In contrast, excitatory amino acids acting on the N-methyl-D-aspartate (NMDA) receptors, or other transmitters, such as substance P, working at the neurokinin-1 (NK-1) receptors produce a more prolonged period of excitation. Together with the excitatory amino acids, substance P mobilizes a sequence of second-message molecules within the postsynaptic nerve cell that help translate the received chemical signal into an electrical impulse. These second messengers activate the release of calcium ions from internal stores, which in turn can cause the cell to deploy other biochemicals that activate a membrane receptor, which facilitates the transmission of the pain signal. An additional consequence of this cascade is the activation of certain genes, which can alter the very structure of the nerve cell involved. Persistent pain can alter the quantity or the responsiveness of membrane receptors and transmitter molecules. It can also cause the cell to strengthen its connections with adjacent nerve cells and, in the long term, change the anatomical structure of the pain pathway. In this way, the individual becomes more sensitive to the same level of painful stimulation.

The intracellular biochemical relays that are activated when a neurotransmitter contacts a receptor on the postsynaptic neuron also activate ion channels embedded in the cell’s membrane. These expel or admit various ions to or from the extracellular fluid. The movement of these ions in and out of the cell and down the length of the nerve fiber is responsible for the electrical current that can be detected with probes and is termed an action potential. Thus excitatory impulses can be responsible for a whole flurry of chemical activity that results in an action potential and, if prolonged, can also result in long-term changes in receptor dynamics as well as in the very structure of neuronal connections.

Inhibitory signals are also part of the pain-conduction pathway. Studies have shown that these come primarily from the brain. Pain-inhibiting, or antinociceptive, signals generally originate in the cortex or the brainstem in areas where opiates, noradrenaline or serotonin are the primary transmitters. These neurons synapse on neurons that release the antinociceptive neurotransmitters gamma-aminobutyric acid (GABA), serotonin or acetylcholine. Alternatively, neurons descending from the brain may block pain by inhibiting the release of pronociceptive neurotransmitters from the incoming sensory neurons.

Antinociceptive signals from the brain and pronociceptive signals from the periphery converge in the dorsal horn of the spinal cord. In this way, the signal descending from the brain can alter the strength of the signal coming from the periphery. Contact in the dorsal horn between descending fibers and ascending fibers can be direct. That is, fibers coming down from the brain may terminate directly on fibers ascending from the periphery in an area very close to the nerve terminal of the ascending fiber. Descending fibers may also terminate on small neurons—termed interneurons—whose cell bodies and nerve terminals both lie within the dorsal horn. The interneurons thus contacted may be excitatory or inhibitory, which means that inhibiting them may bring about either inhibition or excitation in the neurons) on which they terminate. The summation of all of these signals is then carried from the dorsal horn to the brain for further processing, the outcome of which forms a person’s perception of pain.

Thus the mechanism by which a peripheral noxious stimulus becomes encoded and transmitted to the central nervous system is not immutable, but instead involves intricate interconnections of various pronociceptive and antinociceptive neurotransmitters and receptors. These in turn may also be modulated at several levels within the nervous system.

Although scientists do not yet know the entire circuit in full detail, many of the likely biochemical suspects involved in pain have been identified. This has spurred many investigators—ourselves included—to search for specific target-oriented pain treatments.

Current Therapies for Pain Control

None of the currently available pain– treatment options is adequate for the control of chronic pain. Some options are nonspecific. They therefore may deliver some relief from pain, but their nonspecific activities give rise to undesirable and sometimes painful side effects. For example, the nonsteroidal anti-inflammatory drug ibuprofen diminishes pain reasonably well. But in high enough doses, or when taken over a long period of time, it can irritate the stomach lining and can cause internal bleeding and kidney problems. Tricyclic antidepressants and anticonvulsants are particularly effective against neuropathic pain. On the other hand, these drugs are somewhat slow to work, and their side effects, which include a decrease in the number of white blood cells, sedation, anemia and liver dysfunction, can be severe.

Of all the analgesics, however, the opioids seem to arouse the greatest fear and meet the greatest resistance. Opioids, which are extremely effective in treating nociceptive pain (but not so effective for neuropathic pain), can cause a number of side effects, including nausea, vomiting, sedation, constipation and, in rare instances, significant respiratory depression. Patients who continue on opioids can develop tolerance, which means they require increasingly higher doses of medication to bring about the same level of analgesia. It is not uncommon for a patient in chronic pain to require doses that would be lethal to an opioid-naive patient. It is perhaps this phenomenon that has led to the widespread misconception that opioid use for pain will lead to addiction, a notion that makes many health-care workers reluctant to administer the drugs and many patients reluctant to take them. Thus mythology and fear contribute greatly to the inadequate treatment of pain.

Gene Therapy to the Rescue?

Advances in molecular biology give us hope that a whole new armamentarium may be developed to combat pain. Our laboratories are among those considering the use of gene therapy for this purpose. In its very simplest application, gene therapy would allow us to enhance some current pain-management strategies.

Gene therapy permits us to insert into a cell a therapeutic gene, which could code for an antinociceptive receptor or enzymes required for neurotransmitter synthesis. In this way, we can induce a cell to express, or overexpress (if it already expresses the gene in question), antinociceptive molecules.

For example, cells that are normally found in the adrenal glands secrete antinociceptive molecules—opioids among them. Several preclinical and clinical trials conducted by Y. Lazorthes and colleagues at the Rangueil Hospital Medical School in France have demonstrated that chromaffin cells, when transplanted to an area just outside of the spinal cord, can diminish the severe pain endured by cancer patients. Presumbably these patients benefit from the antinociceptive molecules secreted by the implanted cells.

The chromaffin cells could be genetically modified so they increase their production of endogenous antinociceptive opioids and catecholamines (the class of compound that includes noradrenaline and serotonin) and even manufacture and secrete additional antinociceptive agents. This approach could be used with cells other than chromaffin cells, but could be some other type culled from the patient. As long as the cell source is the patient, the engineered and reimplanted cells will likely be perceived as “self” rather than “foreign” and so skirt attack from the immune system. But this approach also has a number of drawbacks. For one fi-thing, this approach is limited by the necessity to engineer cells that produce diffusible therapeutic molecules, since the cell graft can only be deposited outside the spinal cord. In addition, the grafted cells may still be able to divide, in which case the size of the mass they eventually produce limits the time they may remain in the body Finally, this approach, when used to deliver opioids, faces the same limitations of opioids delivered by conventional means; that is, the body develops tolerance-no matter how the opioids are administered-and the dosage needs to be increased.

Gene therapy can be used in other ways that exploit its most powerful advantage over conventional therapies: It allows us to get at targets inside the cell, where so much of the biochemical action takes place. Gene therapy potentially can be used to implant antinociceptive receptors in nerve cells, making them less sensitive to incoming pain signals.

Therapeutic gene delivery can be facilitated with a biological agent that slips easily inside specific cells. Such a Trojan horse exists in nature in the form of viruses, which enter cells via cell-surface receptors. The viruses can be genetically engineered so that therapeutic genes replace harmful ones that cause disease and that direct viral infection and replication. Thus, the replication-deficient viral vector retains the ability to attach to a cell and deposit its genetic cargo inside. However, this cargo includes the therapeutic gene instead of the normal viral genes.

Several different viral vectors can be used, each with a distinct biological profile. In general, the choice of which viral vector to use depends on the desired therapeutic goal. One of the most useful viruses for the purpose of gene therapy is the adenovirus. Adenovirus naturally infects a variety of cell types, including nerve cells. This virus, then, can be used to transfer antinociceptive receptors or neurotransmitters, for example, into a nerve cell along the pain-conduction pathway and dampen the strength of incoming pain signals. Therapeutic genes that could be introduced this way include those for receptors for opioids, adrenaline or acetylcholine. Gene therapy can even be used to block the production of pronociceptive molecules. This can be accomplished with what are known as antisense oligonucleotides. Antisense oligonucleotides interrupt the flow of events called gene expression.

First, the gene, which resides in the cell’s nucleus in the form of DNA, is copied into a molecule of RNA, which carries the genetic message out of the nucleus and into the cell’s cytoplasm. Once in the cytoplasm this so-called messenger RNA (mRNA) molecule directs the synthesis of the protein coded for in the gene.

Antisense oligonucleotides are short stretches of nucleic acid-either RNA or DNA-artificially constructed to recognize and deactivate specific mRNA molecules. Antisense molecules are therefore highly selective and, when inserted into a cell, bind up particular mRNA molecules and prevent them from being translated into protein. The protein encoded by the mRNA is therefore never made. Antisense oligonucleotides can be physically or chemically inserted into cells, or they can be carried by viral vectors in the same way that other therapeutic genes can be.

Since they in essence block gene expression, antisense oligonucleotides can be used to inhibit the production of neurotransmitters and receptors that promote pain. Two groups have used antisense oligonucleotides to “knock down” the production of the NK-1 (Tony Yaksh and colleagues at the University of California, San Diego) and brain NMDA receptors (Gerald Gebhart and colleagues at the University of Iowa) in rats and have shown that this treatment does indeed diminish the animals’ sensation of pain. We have also constructed antisense oligonucleotides to the NRl subunit of the NMDA receptor in the spinal cord, which is important for the development of chronic nociception and pain, and have shown that this is sufficient to reduce pain sensation in rats. Since antisense oligonucleotides have already been extensively studied in clinical human trials, it may not take as long for these drugs to come into use for neuropathic pain as it might if the approach were completely untested. Antisense oligonucleotides may become available for this use in the near future.

The Next Generation of Gene Therapy

Although it holds a great deal of promise for pain management, gene therapy is by no means a potential panacea. One of the major obstacles is the actual delivery of the gene-therapy product to the site of action. Many of the molecules involved in nociceptionthe would-be targets of gene therapy—are located within the central nervous system, a system impermeable to all but the smallest of molecules.

The central nervous system-the brain and spinal cord-are protected by a series of tightly networked endothelial cells called the blood-brain barrier. This barrier excludes all molecules larger than 600 daltons-approximately the size of a small organic molecule such as ATP-from entering the central nervous system. Antisense oligonucleotides and viral vectors are much too large to penetrate this barrier.

In order to introduce these agents into the nervous system, the oligonucleotide or the viral vector must be injected directly into the spinal fluid or into the nervous tissue itself. But even this might not be sufficient. Some data suggest that additional anatomical barriers may prevent virus injected into spinal fluid from entering the spinal cord.

Nevertheless, Michael Iadarola and colleagues at the National Institutes of Health recently demonstrated that when adenovirus engineered to produce P-endorphin was directly injected into the spinal fluid of rats, it produced analgesia. Beta-endorphin is a small, naturally produced opioid. Once produced by the virus-infected cell outside the spinal cord, the opioid protein presumably travels into the spinal cord unimpeded by the anatomical barriers.

An alternative means of delivery mimics the natural route whereby viruses, such as a herpes virus, enter the central nervous system following peripheral infection. Viral vectors can be injected into a limb and travel through peripheral nerves into the central nervous system. In a recent study, Stephen Wilson and coworkers at University of South Carolina School of Medicine demonstrated the feasibility of this strategy. They injected mice with a herpes virus engineered to carry an analgesic opioid. The virus did indeed enter the central nervous system when injected into the animals’ hind paws and produced a long-lasting analgesic effect.

The second obstacle ahead for gene therapy has to do with specificity. In spite of its promise of specificity, the currently available gene-delivery systems-including both viral vectors and antisense oligonucleotides-are simply unable to target a specific subset of cells or genes. Viral vectors currently in use are “promiscuous” and infect all cells carrying the viral receptor. Thus far, delivery of viral vectors can be guided only by the judicious selection of injection sites near the target cell populations. Most of the time this will require the vector to be injected into the spinal cord, where nociceptive signals are processed.

Viral targeting can be artificially enhanced, again through genetic manipulation. A “smart virus” can be made to express a particular protein on its outer surface. This protein would be one that recognizes and binds to a receptor uniquely expressed by the target-cell population. For example, expression of a substance P-like peptide on the viral surface may allow specific targeting of the virus to nerve cells expressing substance P receptors, the precise subpopulation of cells mediating pain. But a virus can be made smarter still.

Not only might it be possible to make viruses adhere to and enter specific subsets of cells, it may also be possible to make the genes they carry respond to specific cues. One could, for instance, include a regulatory sequence that turns the gene on in the presence of certain substances. So, for example, one might include a control region that responds to the increase in intracellular calcium. Such an increase is observed when substance P contacts its receptor. In that case, the therapeutic gene would be activated each time substance P stimulates the cell. The value of such a control is twofold. First, although the therapeutic gene might enter any number of cells, it would only be activated in those cells that respond to substance P-a very small and specific subset of the total. In addition, if the therapeutic gene were, say, an antisense oligonucleotide against the substance P receptor, then gene activation would shut down the synthesis of this receptor and would in theory dampen the sensation of pain.

Yet another means to gain control over therapeutic genes is to include another type of regulatory region with the therapeutic gene-for example, by adding a region responsive to some kind of drug the individual takes. This regulatory region could, for instance, program the gene to become activated when the individual takes tetracycline, a common antibiotic. That way, the individual can control the expression of the therapeutic gene. The two types of control can be combined, so that the therapeutic gene is expressed in only a subset of relevant cells when the individual chooses to express it.

Even if all the above problems can be addressed, there are still several more issues that need to be controlled before gene therapy can become a practical solution to pain management. A problem with gene therapy-especially for chronic pain management-is its duration.

The viruses used as vectors for gene delivery elicit an immune response in the host. This immune response usually becomes stronger and swifter with repeated administration of the virus, which limits the number of times the virus can be given to the patient. Since the genes so delivered remain in the host only transiently, gene therapy can only be of limited use.

Again, however, a new generation of viral vectors is being developed to minimize this problem. Already, a “gutless adenovirus” has been constructed, from which the entire viral complement of genes has been removed. This reduces the amount of foreign material introduced into the host and limits the number of targets to which the immune system can respond.

In addition, investigators are studying other viruses that naturally escape immune detection. Of these, the adeno-associated virus has emerged as a useful vector. We are certain that other vectors will also be developed that combine all of the best features of existing vectors and eliminate many of the drawbacks.

Pain, wrote Emily Dickinson, “has no future but itself. Its infinite realms contain its past, enlightened to perceive new periods of pain.” But the tremendous growth in our understanding of the neurobiology of pain and the parallel growth in molecular biological expertise may change that. The potential combination of these two fields of knowledge may provide clinicians with unique opportunities for the treatment of pain. Animal data suggest that gene therapy can effectively produce antinociception. However, the drawbacks of current tools and gaps in our understanding of the biology of pain limit gene therapy’s uses right now. As our knowledge and molecular biological skill increase, the use of gene therapy to manage pain may indeed someday become a clinical reality.