Objectives. Chronic neuropathic pain has been an enigma to physicians and researchers for decades. A better understanding of its pathophysiology has given us more insight into its various mechanisms and possible treatment options. We now have an understanding of the role of various ionic channels, biologically active molecules involved in pain, and also the intricate pain pathways where possible interventions might lead to substantial pain relief. The recent research on laboratory animals using virus-based vectors for gene transfer at targeted sites is very promising and may lead to additional human clinical trials. However, one needs to be aware that this “novel” approach is still in its infancy and that many of its details need to be further elucidated. The purpose of this article is to thoroughly review the current available literature and analyze the deficiencies in our current knowledge.
Design. Literature review.
Methods. After an extensive online literature search, a total of 133 articles were selected to synthesize a comprehensive review about chronic neuropathic pain and gene therapy in order to understand the concepts and mechanisms.
Results. Most of the studies have shown benefits of gene therapy in animal models, and recently, phase 1 human trials using herpes simplex virus vector have started for intractable cancer pain.
Conclusion. Although animal data have shown safety and efficacy, and initial human trials have been promising, additional studies in humans are required to more completely understand the actual benefits and risks of using gene therapy for the treatment of chronic neuropathic pain.
The International Association for the Study of Pain (IASP) has defined neuropathic pain as “pain initiated or caused by a primary lesion or dysfunction in the nervous system”. Recently, a task force in collaboration with the IASP Special Interest Group on Neuropathic Pain has proposed to replace the current definition of neuropathic pain by the following wording: “Pain arising as a direct consequence of a lesion or disease affecting the somatosensory system.” A grading system of certainty for presence of neuropathic pain has also been proposed according to which pain can be graded as being definite, probable, or possible neuropathic pain depending on four criteria based on clinical symptoms and signs .
Neuropathic pain is a fairly common condition with an estimated prevalence of 1.5% in the general population . However, precise estimation of its incidence and prevalence is extremely difficult because of the lack of a standard definition and cutoff for diagnosis [4,5]. Recent studies have found the prevalence of pain of predominantly neuropathic origin to be as high as 8.2% in the general population . It is a cause of significant disability, decreases productivity and quality of life, and increases the cost of health care by approximately threefold . Current pharmacological and interventional modalities available for the treatment of chronic neuropathic pain have variable efficacy. In addition, other problems with current therapy include intolerable adverse effects. Emerging modalities including gene therapy have shown considerable promise in this area; however, there are several aspects that need to be clarified regarding their use including mechanism of action, specific advantages and disadvantages, as well as their practicality for general clinical use. This article will review currently available literature to highlight the state of our current knowledge regarding the current and potential future role of gene therapy for chronic neuropathic pain.
An extensive literature search was completed by the authors involved in the study. Medline was used as the primary source of articles that were searched using Pubmed, the key words being various combinations of “neuralgia,”“chronic neuropathic pain,”“pathophysiology,”“mechanisms,”“treatment,” and “gene therapy.” Further information was obtained from Internet Websites of the IASP (http://www.iasp-pain.org) and Kimball's Biology Pages (http://biology-pages.info). Additional articles were identified by manual review of retrieved references. English language full-text articles ranging from 1980 to November 2010 were selected based on their quality, relevance to the subject with focus on pathophysiology of chronic neuropathic pain, and its application to gene therapy for treatment of neuropathic pain.
Target Sites and Molecules for Gene Therapy of Chronic Neuropathic Pain
Gene therapy for chronic neuropathic pain encompasses the introduction of specific genes into neurons, where they synthesize specific proteins that interact with their receptors and/or ionic channels involved in the pathogenesis of neuropathic pain. Hence, it is important to understand the role of key biologically active molecules and ionic channels in causation of chronic neuropathic pain. This will be reviewed below.
Role of Biologically Active Molecules
Ongoing research has resulted in the identification of several biologically active molecules that have been implicated in the causation of chronic neuropathic pain. They can be categorized as:
Excitatory amino acids (EAA), substance P (SP), neurokinin (NK) A: Persistent activation of afferent C fibers causes the release of EAAs like glutamate, SP, and neurokinin A. Acting via the N-methyl D-aspartate (NMDA), NK-1, and NK-2 receptors, respectively, they have been shown to rapidly induce hyperalgesia and spinal facilitation [8,9]. Several studies have shown that glutamate is involved in prolonged nociception [10–14]. Intrathecal administration of nonselective NMDA and non-NMDA (quisqualate/kainate) glutamate receptor antagonists in rats has been shown to effectively reduce the early as well as delayed peaks of chemical nociception induced by formalin. Selective NMDA receptor antagonism however produces a marked decrease in the second delayed peak with little effect on the initial peak . This suggests that activation of the NMDA receptor plays a significant role in induction of central spinal facilitation and, once induced, could provide the foundation for further modulation and persistence of pain [11,15]. NMDA receptors are also the key receptors in the “wind up” phenomenon of central sensitization whereby each successive stimulus of a pulse volley sensitizes the dorsal root neurons to generate more action potential [15,16]. Other studies have also suggested that EAA and SP do not actually play a role in pain transmission but in the central facilitation of the process, which, once activated, does not require their continued presence .
Calcitonin gene-related peptide (CGRP): Data from several animal studies implicate CGRP in the generation and maintenance of hyperexcitability of spinal cord neurons in inflammatory models of hyperalgesia . Possible mechanisms may involve the release of SP and EAA in the dorsal horn neurons [18,19].
Inflammatory cytokines: Spinal inflammatory cytokines such as tumor necrosis factor-α (TNF-α) [20–24] and interleukin-1 (Il-1) [25–27] have also been shown to cause hyperalgesia and enhance nociception at the spinal level. Intrathecal administration of Il-1 has been shown to induce mechanical and thermal hyperalgesia , the effect being antagonized by intrathecally administered Il-1 receptor antagonists  and neutralizing antibodies against it . Intrathecal etanercept, a TNF-α inhibitor, administered before spinal nerve ligation decreases mechanical allodynia by almost 50% . A synergistic action between Il-1β and TNF-α, and Il-1 and other neurotransmitters [31,32] have also been suggested. Il-10, however, has been shown to inhibit the expression of Il-1 and TNF-α, and has antinociceptive properties [34,35]. Il-2 also has antinociceptive properties  mediated via its binding to opioid receptors [37,38] and reversed by naloxone . Il-4 is also a prototype of an anti-inflammatory cytokine, which has been shown to decrease the expression of Il-1  and inhibit the induction of nitric oxide (NO) synthase  and cycloxygenase-2 enzymes . Its antinociceptive properties have been well documented in prior studies [34,41]. The source of these cytokines (Il-1β and TNF-α) has been traced to spinal cord neurons, neuroglial cells, and vascular endothelial cells in rat spinal cord after acute spinal cord injury [20,42,43], and in spinal cord neurons and microglial cells in humans after traumatic spinal cord injury [44,45]. Studies have linked a chemokine called fractalkine to the activation of glial cells that cause the release of inflammatory mediators [46,47].
NO: Spinal NO has long been implicated in the central mechanisms of inflammatory hyperalgesia. The inducible NO synthase activity has been demonstrated within 20 minutes  to 1 hour  of the application of a nociceptive stimulus depending upon the type of stimulus. NO further causes the release of SP, CGRP, and inflammatory cytokines, and helps in central facilitation of synaptic transmission.
Prostaglandins: The fact that the prostaglandins stimulate the peripheral afferent fibers and cause pain in inflammatory conditions is well known. However, recent evidence has also implicated them in central sensitization . Membrane depolarization and activation of NMDA receptors increase the concentration of arachidonic acid in the neurons, which, when acted upon by enzyme cyclooxygenase-2, forms prostaglandins . Prostaglandins then diffuse into the extracellular space and cause the release of neurotransmitters such as EAA, SP, CGRP, and NO. These neurotransmitters can further increase the release of prostaglandins .
Melanocortin: The role of the melanocortin system in neuropathic pain is being actively explored. As early as 1981, Sandman and Kastin had demonstrated the hyperalgesic effect of intraventricular injections of α subtype of melanocyte-stimulating hormone (MSH) in rats in a double-blind study and suggested their role in the body as endogenous anti-opiates . It is well known that the precursor molecule pro-opiomelanocortin (POMC) is cleaved to the peptides adrenocorticotropic hormone (ACTH) and MSH, and is present in the pituitary. However, recent work has demonstrated the presence of mRNA of POMC in the rat spinal cord and also the presence of MSH receptors, specifically the subtype MS4, in the lamina I, II, and X of the spinal gray mater that are involved in nociception [53,54]. Intrathecal administration of a melanocortin receptor antagonist effectively decreases the sensitivity to cold and mechanical stimulation in animal models of chronic neuropathic pain. In contrast, a melanocortin receptor agonist has the opposite effect . Interestingly, γ subtype of MSH acts via γ-amino butyric acid (GABA) receptor pathway rather than melanocortin pathway and has analgesic activity . Even more intriguing is the interaction between melanocortins and opioids. Apart from producing MSH, cleavage of POMC also produces β-endorphin, which is an endogenous opioid. This opioid along with its µ- and δ-receptor subtypes are present in lamina X of the spinal gray mater involved in nociception. Whereas α-MSH acts via melanocortin G-protein-coupled receptor to increase the activity of adenylate cyclase, β-endorphin acts via δ-receptor to decrease its activity. Thus, adenylate cyclase is the intracellular integrator of both pathways and, depending upon which pathway predominates, can lead to nociception or analgesia . There is downregulation of µ-opioid receptors and increase in MS4 receptor gene expression in the spinal cord of rats tolerant to morphine . Kalange et al. found that chronic intracerebroventricular administration of an MS4 receptor antagonist along with morphine in rats delayed the development of tolerance and reduced the dependence on morphine .
Endogenous opioids: The central pain inhibitory mechanisms involve the endogenous opioids (β-endorphins, dynorphins, enkephalins), and noradrenergic and serotoninergic systems. Endogenous enkephalins are synthesized from a precursor molecule called preproenkephalin that, upon processing, produces met-enkephalin and leu-enkephalin . These have been demonstrated in a wide variety of central and peripheral neurons including those of the substantia gelatinosa (lamina II) and dorsal root ganglion . They also have been shown to be decreased in chronic inflammatory pain states . The endogenous opioids act on the µ-, δ-, or κ-receptors present mostly in the presynaptic terminals of the primary afferents. The receptors are coupled to G-proteins that, upon activation, inhibit calcium channels and adenylate cyclase, and also open up the potassium channels, leading to reduced neuronal excitability . Interestingly, studies have shown the presence of opioid receptors on peripheral nerve terminals as well. These receptors are of subtypes mu, lambda, and kappa , and assume an important antinociceptive role during inflammation when they bind to opioid peptides released from inflammatory cells . Perineural disruption due to inflammation improves the access of these peptides to the opioid receptors. New opioid receptors are also synthesized over the next few days and are transported to the nerve endings by axonal transport . This is important because targeted gene therapy can be used to increase the concentration of opioid receptors at peripheral nerve terminals as detailed later.
The antihyperalgesic effect of endogenous opioids has been well substantiated. However, dynorphin has also been found to demonstrate significant non-opioid pro-nociceptive activity. It acts via kappa receptors to produce analgesia but, at the same time, increases the release of excitatory neurotransmitters, SP, and CGRP from primary afferent neurons. Studies have shown that prolonged opioid administration increases the spinal dynorphin expression . This has been postulated to play a major role in opioid-induced pain, possibly resulting in behavioral pattern of tolerance to these drugs.
Norepinephrine: Norepinephrine plays a very complex role in pain modulation. Its various receptors are widely distributed in the central nervous system and peripheral tissues. At the peripheral level, norepinephrine may have both pain facilitatory and inhibitory action depending on the type of receptors it acts upon [66–70]. At the spinal level, norepinephrine released from descending pathways cause inhibition of pain at presynaptic and postsynaptic level . At supraspinal levels, norepinephrine has variable pain modulatory effects depending on factors like the supraspinal site of action, the duration and chronicity of pain, and the type of adrenergic receptor present [72,73].
GABA: Various studies have shown that GABA, the principal neurotransmitter of the spinal interneurons, plays an important role in pain inhibition. In fact, partial nerve injury has been shown to promote a selective loss of GABAergic synaptic pathways in the dorsal horn of the spinal cord . GABA is synthesized from glutamate in the presence of a rate-limiting enzyme glutamic acid decarboxylase (GAD), which exists in two isoforms: 65 and 67. Most studies of gene therapy have been done with GAD67, which is expressed in interneurons and other neurons that fire tonically, and synthesizes cytosolic GABA involved in general metabolic activity .
Role of Ionic Channels
After a nerve injury, the regenerating sprouts of the primary afferent fiber have shown abnormal excitability with mechanical stimulation and spontaneous erratic discharge characteristics . These are in part mediated by the ionic channels [76,77]. Several subtypes of voltage-gated sodium channels have been characterized, each having different properties. Nerve injury triggers membrane remodeling, which leads to membrane hyperexcitability. This is in part attributed to expression of different subtypes of sodium channels in the membrane with abnormal inactivation kinetics, allowing for repetitive firing. Subtypes 1.3 and 1.7 are examples of such channels and hence can be a target for gene therapy . An interaction between opioid receptor activation and voltage-gated sodium channels has also been suggested .
Expression of calcium channels also allows for calcium entry into the neurons where it modulates the regulation of growth related proteins. It is shown to increase the release of SP and CGRP from rat nerve neuromas .
Gene Therapy—Current Insights Regarding Its Mechanism(s) of Action
Delivery of genes that encode proteins with antinociceptive properties or those that can antagonize nociceptive molecules into the neurons forms the basis of gene therapy of neuropathic pain. The delivery of genes however requires a vehicle, which could be a virus or a non-virus.
Viral vectors: The advantages of using recombinant viruses as vectors are their natural capacity to enter the cells and deliver their genome to the host cell. However, the main drawbacks are their natural infectivity and their limited capacity to carry the genome. Also, the lytic effect of the virus on the host cell combined with the immune response mounted against the virus or the newly expressed gene product may limit the duration for which such a therapy is effective .
Various viruses have been studied and modified to overcome these limitations. These include herpes simplex virus (HSV), adenovirus, adeno-associated virus (AAV), and retroviruses.
HSV: The genome of HSV has been completely sequenced. It is relatively large (152 kb) and has 72 genes organized as double-stranded DNA . Some of the important genes of the virus include the infected cell polypeptides (ICP) 4 gene, thymidine kinase (Tk) gene, and the latency-associated transcript (Lat). The ICP 4 gene is expressed immediately after the viral entry into the nucleus and is responsible for initiating and controlling the expression of genes vital for viral replication . The Tk gene is responsible for causing reactivation of the virus from the latent stage . The Lat gene is the only gene that is actively transcribed during latency, hence indicating its role in maintaining latency of the virus [60,84]. Because almost half of the HSV genes are nonessential and can be replaced by the transgene, it enables the virus to carry inserts of about 30–40 kb . One advantage of using HSV as a vector is that it is a neurotropic virus that can be applied directly on the abraded skin or given by intradermal injections. It multiplies in the skin epithelia for a few cycles, penetrates the peripheral nerve endings, and travels retrogradely to the cell bodies of the sensory neurons, where it establishes a latent infection in the form of episome, without integrating with the host genome. Another advantage with HSV is that the release of gene product is restricted to only those areas of spinal cord where the nociceptive neurons are projecting. Also, the axonal transport of gene product to the nerve endings may provide additional reduction in pain .
AAV: The recombinant AAV vectors are nonpathogenic and replication defective, and integrate into the host genome, hence allowing long-term expression of the gene . When AAV was used as a vector to transfer opioid gene, it was observed that it causes the expression of gene only in dorsal root ganglion (DRG) and peripheral afferents sparing the spinal cord. This may help prevent stimulation of central compensatory mechanisms mediated via dynorphin, which can lead to opioid tolerance over time, as discussed above.
One disadvantage of using AAV as vector is the fact that it is not neurotropic. For this reason, delivery of the virus to the target site poses a difficulty. Animal studies have experimented with direct injections of the recombinant virus into the DRGs and also with intrathecal and intraneuronal routes of administration. For practical purposes, the intrathecal route seems to be the most appropriate choice. In a series of experiments, Storek et al. tried to identify the efficacy of intrathecal delivery of recombinant AAV and ways to make it more effective and durable [88,89]. They found that upon intrathecal injection of the commonly employed recombinant AAV serotype into rats, the neuronal tissues did not show an expression of the marker protein green fluorescent marker protein (GFP) at the end of 4 weeks. They postulated that this could be secondary to the limited capacity of the capsid proteins to mediate transfer of genome to the target cells or because of limited capacity of the target cell machinery to convert AAV's single-stranded DNA to double stranded. To overcome these obstacles, the authors modified the vector by pseudotyping it (a process by which different viral serotypes are phenotypically mixed, and capsid glycoproteins are exchanged) to generate a new capsid and, at the same time, modifying the genome to make it double stranded. This self-complementary recombinant AAV serotype 8 (scAAV8) was shown to be effective in causing marker protein expression for at least 12 weeks upon intrathecal injection. None of their studies reported any cytotoxicity or activation of immune response in the host. However, a few queries regarding the rostrocaudal extent of transgene expression in the DRG, effect on brain and other organs outside the nervous system, and toxicity of single injection vs two sequential injections into cerebrospinal fluid (CSF) are still unanswered, and further studies are needed in this respect .
Retrovirus: A novel vector called human foamy virus (HFV) has also been recently studied. It is a retrovirus that can transduce different types of cells including neurons across various mammalian species [91,92]. It is nonpathogenic, not inactivated by human serum , can transduce nondividing cells effectively , has a large genome, and, unlike other retroviruses, has a second promoter called internal promoter at the end of env gene providing for additional regulation .
Nonviral vectors: Some studies have experimented with injecting naked plasmid, liposomal-complexed DNA, or polymer-complexed DNA intrathecally, each with their own merits and demerits [95,96]. These nonviral vectors have the advantage that they do not carry the inherent risk of infection, which does exist with any viral vector; in addition, they can be used for genes that are too large to be carried by viruses. While these vectors have been shown to be safe end extremely effective in transfecting the cells in vitro, their in vivo efficiency has been found to be rather low . Although research designed to improve their cellular uptake is ongoing, and few studies have found some success by addition of a pH-sensitive fusogenic segment to the liposomal complex , as of now, their limitations still preclude their use for therapeutic purposes.
Basics of Gene Therapy Using Viral Vectors
The gene to be inserted into the viral genome (called transgene) is first placed under the control of a promoter, which will regulate its expression. Some studies have used a modified HSV latency promoter consisting of a fusion between the region upstream from the HSV Lat core promoter and elements of Moloney murine leukemia virus (MLV) long terminal repeat (LTR) . Others have used the human cytomegaloviral immediate-early promoter/enhancer, which has been shown to drive transgene expression both acutely and after establishment of latency (Figure 1). It is then recombined with the viral genome by use of site-specific recombination (e.g., the recombination occurring at a specific site called LoxP mediated via Cre recombinase enzyme) .
Sometimes, the transgene may be the one that encodes protein, which is actually pro-nociceptive (e.g., CGRP). However, it is inserted into the viral genome in such a way that upon transcription, the mRNA synthesized has a sequence opposite to that of mRNA synthesized in the host cell from native DNA (“antisense sequence”). Hence, the virally encoded mRNA hybridizes the native host mRNA by simple Watson–Crick interaction and effectively blocks its translation into the biologically active protein with pro-nociceptive properties, thus reducing its concentration within the cell (Figure 2).
Clinical Trials of Gene Therapy for Treatment of Neuropathic Pain
Several clinical trials have been carried out in animal models of neuropathic pain and have demonstrated the efficacy and safety of gene therapy in them. A few human clinical trials are also underway. These trials will be reviewed based on the molecules they target with gene therapy.
Studies Targeting the Opioid Pathways Using Gene Therapy
Genes that direct the synthesis of opioid and their receptors have been introduced into the neurons of various animals using viral vector, and the effects were studied.
Wilson et al. introduced the complementary DNA (cDNA) for human preproenkephalin into the sensory neurons of mice using recombinant HSV1 as the vehicle . The transgene was placed under the control of cytomegaloviral immediate-early promoter/enhancer, and the mice were infected by application on abraded skin. Immunohistochemistry successfully demonstrated the synthesis and processing of the opioid precursor in the mouse sensory neurons. Study of the animal behavioral responses suggested reduced responsiveness of both C and Aδ fibers to nociceptive stimuli. The biological effect of therapy was first seen 4–5 days post-infection with its magnitude increasing in the next 2 weeks and remaining robust until at least 7 weeks of observation, despite the fact that by this time, the transgene expression had been reduced by 75%. This discrepancy was postulated to be because of the ability of low levels of proenkephalin to modulate sensory neurons and their effect on neuronal plasticity. Another mechanism could be that the active enkephalins are synthesized and stored in secretory vesicles in the nerve endings to be released only with robust activation of peripheral nerve afferents. This analgesic effect was blocked by intrathecal naloxone only when the fibers were pre-sensitized by capsaicin and dimethyl sulfoxide with no effect on basal withdrawal latencies, indicating that endogenous opioids are released only with substantial activation of the primary afferents and baseline nociception is not altered. The study overall suggests that the targeted application of this virus could result in antihyperalgesia mediated via opioids without their antecedent systemic toxicity and longer duration of action without disrupting normal sensory transmission in chronic neuropathic pain.
Similar results were substantiated by another study, which used the rat proenkephalin gene placed under the regulation of binary promoter  (HSV Lat core promoter and MLV LTR, as described above). This recombinant HSV1 KOS strain (known for its extremely low neurovirulence ) was applied to abraded skin of rats. Three weeks post-infection, they observed a 7-fold increase in the number of proenkephalin mRNA-expressing neurons, a 24-fold increase in the levels of proenkephalin mRNA, and a 160% increase in concentration of met-enkephalin-like material in DRG. Reversed-phase high-performance liquid chromatography and immunohistochemistry demonstrated that the overexpressed proenkephalin was completely processed into active peptides. These active peptides were preferentially transported to the peripheral processes of primary afferent neurons, where they accumulated in secretory vesicles to be released by electrical stimulation of dorsal roots or peripheral nerve. Like the previous study, this study also showed that basal nociception was not affected by this therapy. However, the authors went one step ahead to demonstrate its therapeutic application in rats . They applied this recombinant virus to increase the expression of proenkephalin gene in DRG of rats with adjuvant-induced polyarthritis and studied the effect of this therapy on their functional disability. They found that the locomotor activity of polyarthritic rats reached up to 70% of that of healthy rats, which was highly significant. The effect was also long standing and persisted until at least 8 weeks of observation.
Finegold et al. used an adenovirus as a vector to introduce the artificial fusion gene encoding the secreted form of β-endorphin (preprobeta-endorphin) into the meningeal cells . The gene was under the regulatory control of a cytomegaloviral promoter, and the recombinant virus was injected intrathecally. The CSF concentration of β-endorphin increased fourfold 1 week after infection but declined to control levels within 15 days because of an immune response against the vector. Like other studies, an antihyperalgesic effect to inflammatory stimulus was observed without effect on basal nociception. Yeomans et al. tested the efficacy of this therapy in monkeys . They found that upon topical application of recombinant HSV carrying the human preproenkephalin on the dorsal surface of monkeys' feet, there was a significant analgesia in response to periodic sensitization and activation of C nociceptors. The effect was sustained for 20 weeks. Immunohistochemistry and radioimmunoassay confirmed the production of enkephalin peptides by the DRG cells.
Other experiments have also been done to introduce a µ-opioid receptor gene into neurons. Zhang et al.  used two recombinant viruses, one with the µ-receptor gene (SGMOR) and another with the antisense sequence of gene (SGAMOR), placed under the cytomegaloviral promoter. They also inserted cDNA for GFP flanked by internal ribosome entry site, which could allow initiation of translation of GFP from the middle. The co-expression of GFP with the µ-receptor helped to deduce whether the changes in receptor expression were due to infection with their HSV constructs. The study demonstrated a significant increase in µ-receptor immunoreactivity in the skin where the virus was applied, in the dorsal root ganglion, as well as in the dorsal horn of the spinal cord with SGMOR and the opposite with SGAMOR. This showed that the µ-receptors not only were synthesized in the DRG, but also transported anterograde and retrograde, and thus can modulate pain and analgesia at both spinal and peripheral sites. With respect to DRG, the increase in µ-receptor immunoreactivity related to SGMOR was seen mostly in large-diameter neurons (which normally do not express endogenous µ-receptor), whereas the decrease associated with SGAMOR was seen in small- and medium-diameter neurons (which normally express endogenous receptors and are presumably nociceptive). The analysis of GFP expression revealed that 75% of the increase in µ-receptor expression in the peripheral afferents could be attributed to the viral constructs. The authors also demonstrated the effect of increasing doses of subcutaneous loperamide, which is a selective µ-receptor agonist, on paw withdrawal latencies in the SGMOR-infected, SGAMOR-infected, and the control rats, and the resultant dose–response curve was analyzed. As expected, the half-maximal effective concentration of loperamide (EC50) significantly decreased in rats that were infected with SGMOR as compared with that in rats infected with SGAMOR, and the changes were equivalent albeit opposite. They concluded that the effect of µ-receptor gene insertion and its knockdown is not only molecular but also functional, thus paving a way for its therapeutic application.
Xu et al.  used recombinant AAV to transfer µ-receptor gene placed under the control of rat neuron-specific enolase promoter  into the DRG of rats. They found that the viral system induced an effective expression of µ-receptors not only in small- and medium-sized neurons, but also in large neurons. This intervention thus increases the number of neurons that are sensitive to opioids as well as increase the sensitivity of individual neurons to opioids. The study also found that the therapy sensitizes the neurons to opioid-mediated calcium channel inhibition. The µ-receptor expression was long lasting (6 months after the initial infection) and was restricted to DRG cells and their presynaptic terminals, unlike studies using HSV that have found gene expression even in the spinal cord, thus having the potential advantage of not invoking mechanism causing opioid tolerance. However, as discussed above, one disadvantage of using AAV as vector is that it is not neurotropic, hence making the delivery of virus to the neurons difficult. For this study, the researchers had to remove a small piece of vertebra to inject the virus directly into the DRG. As was discussed, intrathecally injected recombinant scAAV8 was used by Storek et al. and found to be effective in transducing nociceptive neurons in DRG to produce β-endorphin and Il-10. It caused significant analgesia in chronic neuropathic pain model of rats. The effect was sustained for at least 3 months .
Studies Targeting the Interplay Between Inflammatory Cytokines Using Gene Therapy
Studies have also been conducted on the gene transfer of immune mediators of chronic neuropathic pain. Because these mediators of nociception have a very short half-life, their therapeutic application requires their continuous intrathecal infusion, which causes a lot of inconvenience to the patients, is costly, and therefore is not very practical. Induction of meningeal or DRG cells to produce these cytokines by gene therapy has therefore attracted a lot of attention. Hao et al. transduced the DRG neurons of rats to produce p55 TNF receptor (TNFR) by using recombinant HSV as a vector . They placed this gene under the control of an HSV immediate/early promoter and delivered the virus subcutaneously. They found that release of TNFR from the neurons reduced the expression of membrane TNF and Il-1, presumably by binding to membrane TNF and inducing reverse signaling  into the cells. It manifested as reduced thermal hyperalgesia in the infected animals. Another study used HSV-based vectors to deliver soluble TNFR and Il-1 receptor antagonist to nerve fibers in the joint cavity in the arthritis model of rats, and successfully demonstrated the expression of TNFR in significant concentration in the joint fluid and reported significant reduction of leukocytosis and synovitis. However, the effect on functional status of the joints was not studied . Several studies have tried to utilize anti-inflammatory properties of Il-10 to combat pain. Intrathecal injections of Il-10 have prevented and reversed Il-1-induced pain in different animal models, but its short half-life (2 hours) limits clinical application. Il-10 gene transfer using viral vectors show considerable promise but again are effective only for approximately 2 weeks due to activation of host immune system or inefficient promoter activity. Intrathecal injections of naked plasmid DNA encoding Il-10 have found considerable promise. Whereas anti-allodynic effect of a single injection may not be long lasting, a second injection at an interval of 2–3 days potentiates the effect of the first, and the intervention remains effective even after 40 days . The study found its efficacy in both acute and chronic neuropathic pain. Intrathecal injection of plasmid DNA encoding Il-10 was also found to be effective in preventing and reversing allodynia in a paclitaxel-induced peripheral neuropathy model of pain in rat . The authors found increased levels of Il-10 mRNA in the lumbar DRG and meninges, and decreased expression of Il-1 and TNF-α mRNA, 2 weeks after initial therapy. Because the Il-10 receptors are only expressed by glial cells and not the neurons, this therapy does not interfere with neuronal function and acts only to alter the local cytokine milieu in favor of anti-inflammatory mediators that prevent/reduce acute as well as chronic neuropathic pain. Gene therapy has also been used to transduce neuronal cells to produce Il-2, another inhibitory cytokine. Yao et al. injected recombinant adenovirus carrying Il-2 transgene under the control of cytomegaloviral immediate/early promoter enhancer into the lumbar subarachnoid space of rats and detected the mRNA in the cells of pia mater as well as dorsal horn of spinal cord . Behavioral experiments showed that the infection of the rats with recombinant adenovirus increased the basal nociceptive pain threshold and also had analgesic effect on chronic neuropathic pain; the response was maintained for 4 weeks. In another study, naked plasmid DNA encoding Il-2 was injected into the subarachnoid space, and the researchers found that the antinociceptive effect could be maintained for up to 6 days . Hao et al. subcutaneously injected recombinant HSV carrying Il-4 gene under the regulation of HSV ICP 4 promoter and observed that it had a significant reduction in mechanical allodynia and thermal hyperalgesia, the effect peaking at 2 weeks and lasting for 5 weeks . A second inoculation was equally effective in producing anti-allodynia, which also lasted for a similar duration, indicating no immune response to the first inoculation. The researchers also established that the intervention had resulted in reduced levels of PGE2 and Il-1 in the dorsal horn. Studies have also been done to increase the expression of glial-derived neurotrophic factor (GDNF) in the neurons, using recombinant HSV [113,114]. One study found that this intervention had an anti-allodynic effect that was maintained for 3–4 weeks and also upon subsequent inoculation .
Studies Targeting the GABA Pathway Using Gene Therapy
Transduction of DRG neurons to produce GAD67 for GABA synthesis has also been effective in reducing pain. The synthesized GABA is released from interneurons via reversal of GABA transporter and is independent of electrical depolarization . Studies have used HSV as a vehicle to deliver gene encoding GAD67 and demonstrated decreased mechanical allodynia as well as thermal hyperalgesia in rats [116,117]. Another study compared the efficacy of analgesia between HSV-transduced production of GAD and proenkephalin, and found that reduction in pain-related behavior (mechanical allodynia) was more profound and prolonged with GAD than with proenkephalin . In one of the studies, researchers were able to transduce DRG neurons of rats by subcutaneous inoculation of recombinant HFV containing GAD67 gene . They could detect GAD expression in the DRG that manifested as an increased synthesis of GABA. This intervention significantly reduced the mechanical allodynia as well as thermal hyperalgesia in the spinal cord injury model of rats. The therapeutic effect lasted for 5 weeks, and a second inoculation produced effects of similar magnitude, indicating lack of immune response and tolerance to first inoculation.
Studies Targeting CGRP Using Gene Therapy
Pohl and Braz placed an antisense sequence to CGRP under the control of modified HSV latency promoter consisting of a fusion between the region upstream from the HSV Lat core promoter and elements of Moloney MLV LTR . The antisense sequence was the complete sequence of the exon 5 of the rat CGRP gene in an antisense orientation. This transcriptional unit was then recombined with the HSV genome at the glycoprotein C locus using homologous recombination. This recombinant virus was then inoculated on the slightly scarified footpad surface of rat hind limbs. The study demonstrated that the concentration of CGRP in ipsilateral lumbar DRG of rats inoculated with the recombinant virus was significantly lower than the control rats after induction of inflammation achieved by subcutaneous injection of Freund's adjuvant into the hind paw of rats. They also observed that though the biological efficacy of the therapy (as measured by paw volumes indicating local edema) was sustained for all of the 16 days' period of observation, the biochemical efficacy (measured by reduction in the levels of CGRP in the DRG) lasted for only 12 days after induction of inflammation. They suggested that limited variations in the level of endogenous neuropeptides were sufficient to sustain a functional effect.
A similar study was carried out by Tzabazis et al.  using a slightly different viral construct. They placed the transgene with antisense sequence to CGRP under the control of human cytomegaloviral immediate-early promoter/enhancer and then recombined it with the ICP 4-negative HSV genome at the Tk locus, thus inactivating it, in order to disrupt the capacity of the virus to replicate in nondividing cells. The recombinant virus was thus isolated, purified, and rescued by introduction of ICP 4 DNA. The ICP 4 protein represses the latency-associated promoter, hence helping in viral replication. This virus was applied to the abraded skin of mice. The study demonstrated that such mice showed reduced thermal C fiber hyperalgesia after capsaicin application as compared with controls, comparable with the effect of intrathecal CGRP antagonist but longer lasting. They quantified the extent of CGRP knockdown to be approximately 80%, which was comparable to another study that used antisense sequence to NMDA receptor NR1 subunit delivered to cultured neurons via recombinant HSV . The effect of a single application of the virus lasted for 14 weeks. Both the studies established the effectiveness of transdermal route of viral delivery, which is highly desirable.
CGRP appears to be an attractive target for gene therapy because it has been shown to prevent the development of tolerance to antinociceptive effects of morphine .
Studies Targeting Norepinephrine Pathway Using Gene Therapy
Inhibition of norepinephrine synthesis by gene therapy in areas of brain involved in pain facilitation is a novel approach to treatment of pain. In a recent study, Martins et al.  used HSV-1 as a vector for delivery of tyrosine hydroxylase transgene inserted in an antisense orientation directly to dorsal reticular nucleus of the rat brain. Tyrosine hydroxylase is an enzyme that catalyzes the rate-limiting step in norepinephrine synthesis. Dorsal reticular nucleus is involved in supraspinal pain facilitation and receives a strong input from noradrenergic cell groups in the brain stem. Neuropathic pain was induced in these rats using spared nerve injury. The authors successfully demonstrated not only a significant decrease in norepinephrine associated with gene expression in this area, but also a significant inhibition of mechanical allodynia, cold allodynia, and mechanical hyperalgesia. The ability of the vector to transduce only the noradrenergic neurons in the supraspinal areas and to have sustained analgesic effect for almost 2 months was remarkable. Although still experimental, the study is very promising and has opened new potential mechanisms to target noradrenergic modulation of neuropathic pain.
Studies Targeting Ionic Channels Using Gene Therapy
Very few studies have examined the effects of targeting ionic channels with gene therapy. One such study used HSV as a vector to transfer the antisense sequence of gene encoding subtype 1.7 of voltage-gated sodium channels in mice. Application of the recombinant virus on the skin of the animals before treatment with complete Freund's adjuvant blocked the upregulation of this channel in DRG and also effectively prevented the development of thermal hyperalgesia after inflammation .
Chattopadhyay and colleagues established that transgene-mediated expression of enkephalin in rodent models of diabetic neuropathy prevented the increase in expression of neuronal sodium channels, subtype 1.7 by inhibition of protein kinase C, and p38 mitogen-activated protein kinase .
Advantages and Disadvantages Peculiar to Gene Therapy for Chronic Neuropathic Pain
The existing treatment modalities for the treatment of chronic neuropathic pain are limited by adverse effects and, for certain treatments, tolerance. Too often we feel, and most pain practitioners would agree, that even optimal utilization of current treatment modalities does not result in effective pain relief. Gene therapy conceptually offers the advantages of being localized, hence minimizing the adverse effects. Also, it aims to target the biomolecules that are directly involved in pathophysiology of chronic neuropathic pain and hence should be more effective. Several animal studies have shown biochemical evidence of efficacy and safety; however, human trials are just beginning to be completed. Some of the prior clinical trials on humans for conditions other than neuropathic pain have raised concerns about safety, and more clinical trials are needed before gene therapy can be deemed safe and effective for the use of humans, both for terminal and nonterminal causes of neuropathic pain. Another concern with gene therapy is the regulation of gene expression. The transduced cells constitutively synthesize the respective protein that is beyond any physiological control. Continuous production of TNFR, for example, may cause immunosuppresion and increased susceptibility to infections . A physiologically responsive gene therapy, which targets the specific tissue, is inducible by a specific stimulus only, produces the effector protein in therapeutic concentration, and has a rapid responsiveness that would be an ideal system . With respect to chronic neuropathic pain, however, this concern may not be very relevant because we actually require a constitutive expression of the gene. But this may be relevant for inflammatory pain conditions like rheumatoid arthritis, which has intermittent flares of inflammation .
Gene Therapy of Neuropathic Pain: Where Do We Stand Today?
Gene therapy with all its unique advantages seems to be an attractive modality for treatment of chronic neuropathic pain. However, most of the data regarding its safety and efficacy available to us today are from studies conducted on animals, mostly mice and rats. None of the studies have been conducted to assess these effects of gene therapy in humans with chronic neuropathic pain. However, a trial of intrathecally injected plasmid encoding Il-10 in patients with chronic neuropathic pain has been proposed . Wolfe et al. have recently begun a phase 1 human trial (began enrolling in December 2008) with replication-defective HSV encoding enkephalin for treatment of moderate to severe intractable pain due to malignancy anatomically located below the angle of the jaw. It is primarily a safety study, and per protocol, the eligible patients will be injected transdermally with escalating dose of the vector. The study is still ongoing and currently recruiting patients [128,129]. An efficacy study with HSV encoding GAD67 is being planned .
Evans et al.  conducted a study on postmenopausal women who were undergoing joint replacement surgeries for advanced rheumatoid arthritis. One week before the surgery, they injected those joints with autologous synovial fibroblasts transduced with retrovirus containing the Il-1 receptor antagonist gene. This retrovirus was derived from MLV. At surgery, results showed that the synovia of these joints produced more Il-1 receptor antagonist, and less of Il-6 and prostaglandin E2 than those from control joints. This was an indirect indication of the biological activity of the transgene product. They observed these patients for 5 years and did not find any evidence of delayed adverse effects related to the virus itself. This study does show the ability of gene transfer to induce production of gene product; however, its functional efficacy was not assessed. Also, in this study, the joints were removed after 1 week of injections; hence, concerns of safety if these joints were not removed, which would be the case more often in clinical practice, could not be addressed. A similar study was carried out by Wehling et al.  in two patients with rheumatoid arthritis. They used the same virus, transgene, and procedure to inject joints that were to undergo surgical synovectomy. However, they assessed the functional effect of this therapy on the joints by considering reduction in pain and joint swelling; also, they followed the patients for 4 weeks instead of 1 week. The study showed reduction in both parameters, which was maintained for 4 weeks with no adverse effects reported for the duration. Recently, however, one of the clinical trials reported death in a young patient with rheumatoid arthritis who received genetically engineered AAV with the cDNA for TNFR. The cause of death was reported to be disseminated histoplasmosis, and the patient was simultaneously taking adalimumab (a TNF-α inhibitor). The direct role of gene therapy in causation of this death is difficult to determine; however, Food and Drug Administration gave a go-ahead to this study after some modifications in the protocol . Other clinical trials done on humans have shown adverse effects related to the virus, for example, the development of acute lymphocytic leukemia in a patient with severe combined immunodeficiency injected with a transduced retrovirus .
Summary and Future Perspectives
Management of chronic neuropathic pain has been challenging for treatment providers and patients alike. Existing therapies pose problems with their adverse effects; possibility of treatment tolerance; short duration of action; in many instances, suboptimal pain reduction; and need for repeated administration often via invasive routes, thus adding to the patients' burden. Elucidation of the pain pathways and understanding of the pathophysiology at molecular level have enriched our knowledge of chronic pain and led to development of novel treatment strategies that can overcome most of the shortcomings of standard treatment. In the current scenario, gene therapy may be a very promising therapeutic option for chronic neuropathic pain. HSV and HFV hold considerable promise as vectors for gene transfer because of their ease of administration and efficacy. However, scAAV8 offers the advantage of long-term transduction of target cells albeit at the cost of taking intrathecal injections. As reviewed, studies have shown the efficacy of gene transfer in neurons to produce endogenous opioids, µ-opioid receptor TNFR, Il-2, Il-4, Il-10, GABA, and GDNF, or to decrease the production of CGRP in alleviating mechanical allodynia and thermal hyperalgesia in response to various experimental pain syndromes. Moreover, enhanced synthesis of opioids or their receptors does not affect basal nociception; hence, the protective role of acute pain in response to acute injury is still retained. µ-Receptor expression has been shown to increase the potency of intrathecal morphine and also to decrease the EC50 of its agonist. Inhibition of CGRP synthesis additionally prevents the development of tolerance to opioids. So far, results from animal studies have been very encouraging. Initial human trials with regard to gene therapy have shown efficacy, but acceptable safety is yet to be established. Further studies are needed to establish the safety and adverse effects of such a therapy, and to elucidate the dosing and monitoring patterns in humans.