Introduction

As fundamental knowledge of the molecules involved in nociception evolves, the utilization of this knowledge to formulate experimental molecular neurosurgical techniques also advances. The resulting dynamic combination is elucidating the role of specific receptors, neurons, and pathways in the generation and maintenance of pain states [1–3]. The numerous molecular targets that are identified provide an excellent mechanistic basis for clinical therapeutic intervention, either by molecular neurosurgical approaches or through subsequent adaptation to more conventional pharmacological approaches. The goal in using targeted cytotoxins in the clinical setting is to elicit profound and sustained pain relief through prolonged, even permanent, deactivation or deletion of subpopulations of primary afferent pain-sensing neurons via peripheral or central (spinal CSF) administration [4–7].

Translation of the mechanistic understanding of potential therapeutic interventions to clinically effective analgesic compounds has been a challenge in the field of chronic pain. Studies in laboratory animals, mainly rodents, with experimentally induced pain states have been only partially successful in predicting human clinical trial outcomes [8–15]. These experimentally induced conditions may not adequately model the natural disease process that leads to pain. Thus, supplementation of rodent models with additional models constitutes an informative transitional step for drug development and application of novel treatments to human patients. In this regard, large animal pain conditions or disorders that closely mimic their human counterparts and human models for demonstrating target engagement are important new adjuncts [6,16,17]. There is growing interest in using the diseases that spontaneously develop in companion dogs to investigate efficacy of new pharmacological agents and interventional administration approaches. In the evaluation of TRPV1 agonist efficacy for chronic pain, the naturally occurring companion dog models of bone cancer and osteoarthritis have been instrumental in documenting analgesic efficacy and potential side effects, as well as informing future clinical trial design, justifying the starting dose, and selection of outcome measures and primary endpoints.

The initial sections of this chapter examine several mechanistic neurobiological and pharmacological factors that determine optimal therapeutic actions of TRPV1 agonists. The mechanistic and practical insights gained from a succession of investigations involving in vitro ectopic expression systems, ³H-RTX binding, TRPV1 immunocytochemistry, in vivo preclinical studies in rodents, and the large animal veterinary models were key elements for understanding how to use TRPV1 agonists for the treatment of pain. TRPV1 agonists, unlike TRPV1 antagonists [16,17], are not administered systemically or orally; rather, they are used in an interventional fashion, being administered by injection into, or close to, the site that is generating the pain problem (reviewed in Refs. [18,19]). Although we understand much of the necessary underlying neurobiology and pharmacology, because resiniferatoxin (RTX) administration is an interventional technique, going forward into the multiple, potential human pain indications for which this approach may be appropriate will entail clinical experimentation to determine factors such as optimal administration protocols, optimal formulations, and development of expertise on the part of the specialist [20,21].

The second set of sections examines canine clinical pain conditions and the actions of RTX as a pain control agent in these disorders. Although such investigations establish RTX as a treatment in and of itself for veterinary patients, they also constitute a “transitional model” for human clinical trials. Control of pain from canine osteosarcoma by intrathecal RTX was one of the key results that provided incentive and determined progress to human clinical trials. We also present observations on intraarticular (IA) RTX for pain in canine osteoarthritis. Given the strong predictive information for human translation that emerged from the canine osteosarcoma pain control trials, we are optimistic about translating the osteoarthritis results into human and animal clinical practice.

Mechanisms of TRPV1 Agonist Therapeutic Action

First, it is important to acknowledge that RTX is an ultrapotent TRPV1 agonist. RTX causes extremely prolonged channel opening and calcium influx, which results in cytotoxicity to the TRPV1-positive pain fibers or cell bodies [22]. When applied peripherally, the nerve terminal is not just made unresponsive to further direct stimulation of the TRPV1 ion channel as in the case of antagonists, it is actually temporarily destroyed due to calcium cytotoxicity—making it incapable of being stimulated by any agonists of any algesic receptor targets normally expressed by these pain sensing nerve terminals [4,23,24]. Thus, the analgesia is far more complete than that attained by any single channel or receptor antagonist because the latter leave the nerve terminal intact and the nerve ending susceptible to stimulation via the other channels that are left unblocked [17,24].

Four Main Determinants of TRPV1 Agonist Actions

Four aspects of the dorsal root ganglion (DRG) TRPV1 receptor system and TRPV1 agonist action are particularly relevant for optimal use of agonists as therapeutic agents. These involve (1) specificity of TRPV1 agonists, (2) TRPV1 agonist pharmacodynamics, and (3) selective cellular TRPV1 expression in DRG. All three of these attributes make important contributions to the profile of therapeutic action. A fourth element (4) is the requirement for localized administration. Placing the drug next to, or into, the peripheral site that is generating nociceptive activity adds a further level of specific action. Such peripheral generators can be from an identified source; these might include surgical incisions, trigger zones for abnormal activity in a chronic pain condition (e.g., a neuroma), or an arthritic joint. The four main criteria determining TRPV1 agonist actions are summarized in Table 6.1, and the basic research findings supporting the underlying mechanisms or properties are examined in the following sections.

Selectivity of RTX at the Cellular Level

The presence of TRPV1 is the major determinant of RTX action. Without expression of TRPV1, RTX at concentrations used to lesion expressing cells, nerve terminals, or axons does not appear to produce any negative effects at the cellular level [5,6]. In fact, concentrations well above the effective dose (~ 1000 ×) do not appear to affect non-TRPV1-expressing cells. Even when nonexpressing neurons are adjacent to TRPV1 expressing neurons that are undergoing damage from agonist activation, the nonexpressing cells appear to remain intact. These conclusions were reached using several techniques including live cell imaging of cultured DRG neurons, imaging of transiently, and stably transfected with TRPV1, and histology of ganglia injected with RTX [4–7,25,26].

An example of a live cell imaging study is shown in Figure 6.1 using cultured cells transiently transfected with a plasmid engineered to express TRPV1. This series of images, taken over a 45-min time period, show that only cells that take up the plasmid and express TRPV1 are susceptible to vanilloid agonists. Surrounding nontransfected cells that have not taken up the plasmid and do not express TRPV1 are not susceptible to RTX. The process of cell destruction is rapid: the TRPV1 expressing cells undergo intracellular remodeling of the endoplasmic reticulum (ER) and the mitochondria and eventually the nuclear and plasma membranes that can be seen starting as soon as 1 min after exposure to RTX. Eventually the cell is irrevocably damaged and dies. The rapid and massive fragmentation of the ER and mitochondria is a direct result of the prolonged influx of calcium across the cell membrane. This process is illustrated in Figure 6.1. After transient transfection with a TRPV1-expressing plasmid, only a subpopulation of cells takes up the plasmid. At 24 h after transfection, the cells are incubated in MitoTracker Red, a vital dye that stains mitochondria in living cells and with the cell membrane impermeant dye, propidium iodide. Once the cell is damaged, it will become permeable to the propidium iodide. Upon addition of RTX, both the ER lattice (which can’t be seen in these images) and the elongate mitochondria fragment into vesicular structures [5,25].

Organelle fragmentation begins to occur within minutes and is coincident with the sharp, agonist-induced rise in intracellular calcium ion concentration ([Ca⁺⁺]_i) [25]. In addition to the rapid actions on cell viability, Figure 6.1 also illustrates the specificity of the RTX-TRPV1 interaction. There is a notable lack of effect on non-TRPV1-expressing cells, even when in close proximity to RTX-compromised, TRPV1-expressing cells. At time zero, all the cell nuclei in the field of view are dark and exclude propidium iodide; also, all mitochondria in all the cells have a similar appearance. At 20 min after exposure to 1.6 nM RTX the mitochondria of several cells in the field of view have fragmented. The fragmentation is not so easy to appreciate at this magnification and can begin even earlier, within minutes of exposure [25]. By 25 min the TRPV1-expressing cells have become permeable to propidium iodide (the nuclei begin to turn red or brighter in the black-and-white image), and by 45 min the cells are irrevocably damaged (there are six cells with red/bright nuclei in the field of view, three in the upper half and three in the lower half of the frame; arrows indicate four of them). Small plasma-membrane-enclosed fragments of cytoplasm are also seen in the field at 45 min, which are visible as small red/bright droplets. The loss of plasma membrane can be detected using electrophysiological recording from DRG neurons and is measured as a decrease in membrane capacitance [26] reflecting the decreased ability of the membrane to store charge. Despite this level of cellular carnage, the nearby, nontransfected cells ostensibly appear undamaged: Their mitochondria are intact, and they persist in excluding propidium iodide. Thus, the expression of TRPV1 is the major determinant of the cellular specificity of RTX and is the mechanism for TRPV1-dependent, RTX-triggered cytotoxicity.

Experiments with TRPV1 ion channel blockers such as ruthenium red, orthosteric capsaicin antagonists, and RTX or capsaicin activation of TRPV1 in calcium free media demonstrate that the entire process of organelle fragmentation and cell death is attributable to calcium entering the cell through the pore of the TRPV1 ion channel using the complement of TRPV1 receptors located in the plasma membrane [25]. The exact same process occurs when cells are exposed to the calcium ionophore ionomycin [25,27].

Interestingly, DRG neurons and cells stably or transiently transfected with TRPV1-expressing plasmids also contain a population of TRPV1 protein located in the ER [22,25,28]. The ER localization was seen when we expressed a TRPV1eGFP fusion protein. Live cell imaging showed that the ER contained a prominent amount of fluorescently tagged TRPV1eGFP [25]. We were able to examine the activation kinetics of the ER TRPV1 population by incubating cells in calcium-free medium and performing RTX or capsaicin activation. Activation of the intracellular population of TRPV1 was possible because both agonist compounds are hydrophobic and readily cross the lipid bilayer of the plasma membrane. These studies showed that the ER contains enough TRPV1 and enough calcium that, when the ER TRPV1 is selectively activated (e.g., in zero extracellular calcium or when the plasma membrane TRPV1 channel is blocked with the membrane impermeant channel blocker ruthenium red), organelle fragmentation and cell death occur in a fashion identical to that achieved on activation of TRPV1 in the plasma membrane. However, the dose required to activate ER TRPV1 is about 10 times that needed to activate the plasma membrane TRPV1 population [22]. The reason for the difference in dose is not clear. However, based on the quantitative dose differential between TRPV1 in plasma membrane and ER, it appears that the ER TRPV1 does not directly influence the therapeutic pharmacodynamics because nearly all the calcium enters the cytoplasm via plasma membrane TRPV1 [22]. We have proposed that the ER TRPV1 may be used by cells to regulate intracellular calcium dynamics, and the dual localization has been explored in a variety of cell types and preparations, with cancer cells being the main cellular targets [29–32].

This same process of selective cell removal occurs in vivo in the sensory ganglia when RTX is microinjected directly into the sensory ganglia. With standard histological stains, some of the cells can be seen to be undergoing neuronophagia with the locations of the deleted cell bodies marked by hypercellular accumulation of lymphocytes called nodules of Nageotte. Nodules of Nageotte are frequently seen with infectious diseases such as leprosy that destroy sensory ganglion neurons [5,6]. The Nageotte nodules are eventually replaced by eosinophilic material surrounded by satellite cells [6]. During this process of cell death and reabsorption, adjacent cells remain intact, a situation similar to that seen in vitro in Figure 6.1. The very precise excision of pain-sensing neurons from the DRG reinforces the analogy of RTX to a chemical “molecular scalpel” for use in molecular neurosurgery [33].

The preceding paragraphs provide a fairly detailed account of the cellular mechanisms underlying the therapeutic actions of RTX. The behavioral concomitants are that (a) heat pain sensation is lost in the dermatomes innervated by axons or ganglia exposed to RTX as well as (b) thermal hyperalgesia as measured in experimental peripheral inflammation, and (c) sensations of clinical pain are obtunded as seen in canine osteosarcoma [5,6]. More recently, RTX has been evaluated in a wider variety of rodent models of clinical pain using various routes of administration [34–40]. Another set of studies has shown that a long-duration hyperalgesic state can be instated in mice following systemic exposure to RTX by intraperitoneal administration [41,42]. The effects of parenteral administration probably do not have a bearing on the local administration protocols discussed in this review. Recent use in humans includes evaluation of topical RTX for controlling premature ejaculation [43], and the long-investigated urological applications of RTX, among the earliest uses of RTX in humans, has been recently reviewed [44].

RTX Produces a Prolonged Channel Open Time

Two types of physiological experiments demonstrate the prolonged channel opening that is produced by RTX. First, using calcium imaging and DRG neurons in primary culture, it was possible to directly compare dynamic alterations in intracellular calcium induced by capsaicin to those induced by RTX [25]. When cells are exposed to a pulse of RTX, a large and prolonged increase in intracellular calcium occurs. This prolonged elevation occurs in both rat and human DGR neurons in primary culture [5]. A second line of cell-based evidence came from manipulating the extracellular calcium concentration with and without RTX and with and without ruthenium red, the membrane impermeant channel blocker [22]. Although we did not determine the full duration for which the channel can remain open, the duration was on the order of 20 min. This was determined by adding RTX to cultured DRG neurons in calcium-free medium, washing the RTX out for various times, and then switching back to calcium-containing buffer. The chamber was subjected to prolonged perfusion to remove any RTX. Nonetheless, on reinstating normal extracellular calcium concentration, a rapid and massive increase in intracellular calcium occurred. The calcium entry was attributed to the ability of RTX to keep the plasma membrane-localized TRPV1 ion channel in the open state because the calcium entry could be blocked by inclusion of ruthenium red, a TRPV1 channel blocker, in the perfusion media. This kind of dynamic assessment is not as readily apparent from an examination of in vitro binding kinetics; nonetheless, the binding studies show that [³H]-RTX is a high affinity ligand for TRPV1, which is consistent with the in vivo and in vitro potency of RTX. The affinity of [³H]-RTX for TRPV1-containing membrane preparations from various animals or HEK293 ectopically expressing rat TRPV1 appears to vary with the type of preparation, but the K_i is generally in the single-digit nanomolar range [45,46].

TRPV1 Is Selectively Expressed in Subpopulations of DRG Neurons

Immunocytochemical staining of sections from TG or DRG shows that TRPV1 is expressed in a variety of small to medium-sized DRG neurons and exhibits a varying degree of expression (Figure 6.2). The small, very darkly stained neurons (midsize arrows) present a strong contrast to the medium and lightly stained neurons and especially to the large, unstained proprioceptive-somatosensory neurons illustrated in this section. Thus, although the degree of expression likely does not affect RTX specificity, it may be more of a determinant of RTX efficacy. A gradient of susceptibility to RTX can be hypothesized based on the amount of TRPV1 produced by any particular neuron such that the high expressors are the most susceptible to RTX, whereas lower expressing neurons are less susceptible. Evidence in support of this can be gleaned retrospectively from an earlier publication [5]. Panels C and D of Figure 6.2 in that report show that, after an intratrigeminal injection of RTX, most TRPV1⁺ neurons are eliminated, but the remaining TRPV1⁺ neurons all appear to meet the “lightly stained” criterion and would be classified as low expressors. This hypothesis needs more formal testing, but a gradient of expression among TRPV1⁺ DRG neurons seems to be a plausible explanation for the ability of some neurons to resist the actions of RTX. In these neurons, the expression of TRPV1 is low enough to allow them to sequester the incoming calcium and survive transient exposure to RTX.