Synaptic Long-Term Potentiation

Open in new tab

Figure 2. LTP can be induced in vivo and in vitro at spinal synapses of primary afferent C-fibers by various conditioning stimuli, such as direct electrical nerve stimulation at C-fiber intensity at a low (A, LFS, 2 Hz in vitro) or high (B, HFS100 Hz, in vivo) frequency, by natural noxious stimulation with intraplantar injection of transient receptor potential vanilloid type 1 agonist capsaicin (C), or by the abrupt withdrawal from the µ-opioid receptor agonist remifentanil after an intravenous infusion that lasted one hour (black horizontal bar, D). Modified from Sandkühler and Gruber-Schoffnegger (2012).

Long-term potentiation (LTP) between nociceptive nerve fibers and spinal dorsal horn neurons is considered a fundamental mechanism contributing to long-lasting pain amplification (for reviews, see Sandkühler, 2009; Luo, Kuner, & Kuner, 2014; Zhuo, 2014). In spinal cord dorsal horn, LTP can be induced at C-fiber synapses in vivo and in vitro by diverse conditioning stimuli such as natural noxious stimuli, peripheral inflammation, nerve crush, direct electrical nerve stimulation at C-fiber intensity at high (~100 Hz) or low (~2 Hz) frequencies and by the abrupt withdrawal from opioids (see Figure 2 and Sandkühler, 2009). LTP can further be induced by spinal application of agonists at the dopamine D1 receptor, BDNF or ATP (Liu & Zhou, 2015) as well as by cytokines such as IL-1β‎ and tumor necrosis factor (Gruber-Schoffnegger et al., 2013). Most, if not all, of these LTP-inducing stimuli also lead to hyperalgesia in behaving animals (Sandkühler, 2013) and activate glial cells in the spinal cord (Ji, Berta, & Nedergaard, 2013; McMahon & Malcangio, 2009). Glial cell activation has been found to be indispensable both for the induction of LTP at spinal C-fiber synapses (Gruber-Schoffnegger et al., 2013; Zhong et al., 2010; Park et al., 2011) and for many forms of hyperalgesia tested so far (Old, Clark, & Malcangio, 2015). In fact, selective activation of glial cells is not only necessary but can also be sufficient to induce LTP at C-fiber synapses (Kronschläger et al., 2016). This gliogenic LTP may underlie LTP by conditioning stimuli that activate glial cells (Kronschläger et al., 2016). The underlying mechanisms of LTP in nociceptive are currently evolving (see reviews in Sandkühler, 2013; Liu & Zhou, 2015; Luo et al., 2014; Sandkühler, 2009) and involve both pre- (Luo et al., 2012) and postsynaptic signaling (Ikeda, Heinke, Ruscheweyh, & Sandkühler, 2003; Ikeda et al., 2006). See Sandkühler and Gruber-Schoffnegger (2012) for a review.

Neuronal Excitability

Activation of glial cells also directly impacts neuronal excitability. For example, astrocytes regulate extracellular potassium and glutamate concentrations via Kir4.1 potassium channels and glial specific glutamate transporter 1, respectively. The expression of both of these proteins is reduced by about 80% after crush injury of the spinal cord (Olsen, Campbell, McFerrin, Floyd, & Sontheimer, 2010). The resulting enhanced potassium and glutamate levels facilitate neuronal excitation. Furthermore, non-neuronal cells, including glial cells, may release substances that directly enhance neuronal excitability. This includes glutamate that is released not only from neurons but also from glial cells and that triggers synchronous neuronal excitation (Tian et al., 2005). D-serine is also released by glial cells and enhances neuronal excitability by acting as a co-agonist at the essential glycine binding site of the N-Methyl-D-aspartate (NMDA) receptor. IL-1β‎ likewise enhances NMDA receptor-mediated currents by binding to type 1 IL-1β‎ receptor that is associated with the GluN2B/NR2B subunit of the NMDA receptor complex. Another prototypic pro-inflammatory cytokine, tumor necrosis factor, poentiates α‎-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor-mediated synaptic currents in the spinal cord by enhancing their membrane insertion (Ferguson et al., 2008).

Modes of Disinhibition

The main inhibitory neurotransmitters in spinal nociceptive pathways are γ‎-aminobutyric acid (GABA), glycine, and the endogenous opioids enkephalin and dynorphin. Some of these neurotransmitters are co-released from spinal interneurons. In addition, the monoamines 5-hydroxtryptamine and noradrenaline are released from long, descending tract neurons largely originating from the rostral ventromedial medulla or the locus coeruleus, respectively (Todd, 2015; Zeilhofer, Wildner, & Yévenes, 2012). The properties and functions of the inhibitory systems can be impaired at multiple levels, as illustrated in Figure 3. Known mechanisms of altered inhibition include changes in the synthesis, storage, and/or release of inhibitory neurotransmitters, cell death or altered excitability of inhibitory interneurons, altered excitatory drive, or spontaneous activity of inhibitory interneurons. Impaired inhibition can also result from changes that take place in the postsynaptic neuron. For example, altered expression, functions, or trafficking of inhibitory neurotransmitter receptors or downstream signaling can have major impacts on the efficacy of postsynaptic inhibition. Even if all of the above parameters remained the same, inhibition could still be reduced, abolished, or converted into excitation by changes in the driving forces for ions across the postsynaptic plasma membrane, as this may drastically change the ion flux across neurotransmitter receptor channels (Price, Cervero, & De Koninck, 2005).

Disinhibition of Polysynaptic Pathways

Open in new tab

Figure 3. Sites (in blue) where inhibitory systems may be impaired. An inhibitory spinal interneuron presynaptically depresses input from nociceptive primary afferents and postsynaptically inhibits a spinal nociceptive neuron, here shown as a principal pain neuron that, by definition, contributes to the sensation of pain when excited. The inhibitory interneuron is activated either by input from sensory afferents (some of which may be nociceptors), by excitatory interneurons, and/or by descending excitatory pathways (not shown). Modified from Sandkühler (2013).

Disinhibition of polysynaptic excitatory pathways from deep to superficial spinal dorsal horn appears to be functionally important for an abnormal crosstalk between low-threshold Aβ‎-fiber afferents and the nociceptive system (Baba et al., 2003; Schoffnegger, Ruscheweyh, & Sandkühler, 2008). The activation of low-threshold mechanoreceptive primary afferent Aβ‎-fibers leads to the excitation of low-threshold neurons in deep layers of the spinal cord dorsal horn. Some of these low-threshold dorsal horn neurons also respond to high-intensity, noxious stimuli and are then referred to as “wide-dynamic range” neurons. In contrast, neurons with preferential excitatory input from high-threshold Aδ‎- and/or C-fibers are located in superficial laminae I and II (Todd, 2010). When GABAergic inhibition in the spinal cord is blocked, the activation of Aβ‎-fibers now leads to excitation of superficial dorsal horn neurons. The direct excitation of deep dorsal horn neurons by microinjections of glutamate likewise excites superficial dorsal horn neurons, but only when the normal GABAergic inhibition in the dorsal horn is blocked (Schoffnegger et al., 2008). The disinhibition of excitatory, polysynaptic pathways between deep and superficial laminae of the spinal cord dorsal horn may well explain how tactile stimuli evoke withdrawal responses and are perceived as painful in peripheral neuropathy and during an inflammation (Baba et al., 2003; Schoffnegger et al., 2008).

Neuroinflammation and Disinhibition

The inhibitory systems in the spinal dorsal horn interact at multiple levels with mechanisms of neuroinflammation. Receptors for all of the relevant inhibitory neurotransmitters are expressed both by neurons and by non-neuronal cells involved in neuroinflammation. The activity of inhibitory systems in the spinal cord can be sensed by microglia, astrocytes, T-cells, and/or vascular cells. For example, GABA receptors are expressed on microglia, astrocytes, vascular cells, and perhaps mast cells (Haenisch, Huber, Wilhelm, Steffens, & Molderings, 2013). Likewise, monoamines modulate not only the functions of neurons but also those of glial cells, T-cells, and vascular cells.

On the other hand, the properties and functions of inhibitory systems in the spinal cord dorsal horn are strongly modulated by mediators of neuroinflammation. For example, BDNF can be released from microglia, astrocytes, and potentially from CD4+ T-cells (Xin et al., 2012), as well as from neurons. BDNF may reduce the inhibition that is mediated by the activation of ionotropic GABA_A or glycine receptors. The inhibitory postsynaptic action of these receptors relies on the proper anion gradient across the plasma membrane of postsynaptic neurons. This gradient much depends on the action of the potassium-chloride co-transporter-2, which is depressed by BDNF (Coull et al., 2005). Thereby inhibition may be diminished or even converted into paradoxical excitation (Coull et al., 2005).

“Central Sensitization”

The literature on pain often uses the phrase “central sensitization” but does not define it well. Some authors use it as an umbrella term for all kinds of central nervous system mechanisms contributing to pain hypersensitivity. This concept was first proposed by Sturge in 1883 and was scientifically elaborated by Hardy and colleagues (1950). Hardy et al. (1950, p. 139) concluded that “the secondary hyperalgesia occurring in undamaged tissue having an unaltered pain threshold but increased pain sensibility is the result of a central excitatory state. It is suggested that this excitability is to be found in a network of internuncial neurons which intercalate the noxious impulses from visceral, deep somatic, and cutaneous tissues.” In 1977, Lynn further elaborated on this concept. The International Association for the Study of Pain (IASP) now defines “central sensitization” as the “increased responsiveness of nociceptive neurons in the central nervous system.” Nociceptive neurons, however, comprise a large and very heterogeneous group of neurons that serve distinct and sometimes opposing functions, as highlighted earlier. Only a subset of nociceptive neurons, termed principal pain neurons (Sandkühler, 2009), is actually involved in the sensation of pain. In fact, an “increased responsiveness of nociceptive neurons” as used by the IASP to define “central sensitization” may cause more intense pain (in case of principal pain neurons), may lead to analgesia (in case of some inhibitory nociceptive neurons), or may simply be unrelated to the sensation of pain (in case of motoneurons). Thus, the phrase “central sensitization” is not suggested for future use as a technical term in any scientific literature on pain. The IASP also suggests that the term central sensitization should normally not be used in any clinical context.

Wind-Up

Ongoing discharges in C-fibers at rates around 0.5 and 5 impulses per second may trigger progressively increasing action potential firing in some spinal dorsal horn neurons. This “wind-up” phenomenon is due to temporal summation of excitatory postsynaptic potentials and typically lasts a few seconds only. The discharges then level off or decrease again. The winding up of discharges during the first seconds of an ongoing noxious stimulus may be useful to enforce nocifensive responses. Wind-up is a completely normal aspect of the coding properties of some spinal dorsal horn neurons and per se is neither a form of neuronal plasticity nor a mechanism of hyperalgesia or even chronic pain. Long-lasting changes in wind-up properties can be used as markers—rather than as a mechanism—of plasticity in spinal nociceptive pathways (Herrero, Laird, & Lopez-Garcia, 2000; Sandkühler, 2009).

Expansion of Receptive Fields (to Low-Threshold Input)

In the somatosensory system, large receptive fields are generally associated with a low spatial resolution for the respective sensation. In the nociceptive system, injury may lead to enlarged receptive fields of some spinal dorsal horn neurons (McMahon & Wall, 1984; Cook, Woolf, Wall, & McMahon, 1987). The phenomenon of enlarged receptive fields may result from an array of mechanisms, including enhanced excitability, reduced inhibition, or LTP at synapses converging onto the neuron under study or elsewhere in the neuronal circuitry. A potential correlate of enlarged receptive fields has been reported in humans after induction of hyperalgesia with capsaicin. Capsaicin impaired the two-point discrimination ability for non-noxious, tactile stimuli and reduced the ability to detect differences of the roughness of various stimulation surfaces only within the capsaicin-treated area (area of primary hyperalgesia) (Kauppila, Mohammadian, Nielsen, Andersen, & Arendt-Nielsen, 1998). The clinical relevance of this finding is presently unknown.

Sprouting of Low-Threshold Myelinated Afferents

Low-threshold mechanosensitive Aß-fiber afferents terminate in laminae II inner, III, and IV. It has been proposed that after nerve injury low-threshold mechanosensitive Aß-fibers would extensively sprout into more superficial laminae I and II outer and make synaptic contact with nociceptive specific neurons. This would then lead to Aß-fiber-mediated allodynia (Woolf, Shortland, & Coggeshall, 1992). Nerve injury does, however, change the phenotype of C-fibers, so that the marker used to label Aß-fibers (the B unit of the choleratoxin) now also labels C-fibers. This phenotypic switch in C-fibers, rather than sprouting of Aß-fibers, largely explains that the label was found extensively in superficial layers after nerve injury (Bao et al., 2002; Hughes, Scott, Todd, & Riddell, 2003). Labeling of individual Aß-fiber afferents suggested, however, that sprouting of Aß-fibers into superficial lamina does occur but to a rather minor extent. Opening of preexisting polysynaptic pathways between deep and superficial spinal dorsal horn neuron appears, as outlined earlier, to be functionally more important than sprouting of Aβ‎-fibers (Baba, Doubell, & Woolf, 1999; Schoffnegger et al., 2008).

Preemptive Strategies for Neuroinflammation

Recent studies have addressed the question of whether preemptive tackling of neuroinflammation has any beneficial effects on the incidence of neurodegenerative or neurovascular diseases, traumatic head or spinal cord injuries, or infections. Epidemiologic studies revealed mixed effects of preemptively depressing neuroinflammation on the risk of developing Alzheimer’s disease (Heneka et al., 2015), Parkinson’s disease (Leith, Wilson, You, Lumb, & Donaldson, 2014), other neurodegenerative diseases, or cardiovascular diseases (Awan & Genest, 2015). Neurogenic neuroinflammation directly results from enhanced neuronal activity (Xanthos & Sandkühler, 2014). It can thus be tackled by blocking or normalizing the underlying neuronal activity, very similar to preventing neurogenic neuroplasticity (as discussed in the next paragraphs). In the context of nociception, only few studies have addressed the potential effects of preemptive depression of neuroinflammation. Pretreatment with preferential cyclooxygenase-1 or -2 inhibitors before swim-stress in rats reduces both the increase in spinal prostaglandin E2 levels and thermal hypersensitivity. This suggests that preempting spinal neuroinflammation can prevent some forms of hyperalgesia (Guevara, Fernandez, Cardenas, & Suarez-Roca, 2015).

Confinement Strategies for Neuroinflammation

Recent studies have identified a number of targets that limit or reverse neuroinflammation in animal models of chronic pain. Minocycline, a semisynthetic tetracycline, not only exerts antibiotic properties but also inhibits the release of pro-inflammatory cytokines from activated microglia. A large number of animal studies suggest that minocycline is beneficial for various forms of neuropathic, traumatic, and inflammatory pain (for a review, see Bastos, de Oliveira, Watkins, Moraes, & Coelho, 2012). Minocycline has been a clinically approved drug for over 30 years, and thus in principle it is also available for off-label use in human pain patients. The potential usefulness of other approved drugs that interfere with neuroinflammation may well be worth testing in patients with chronic pain. Examples are drugs that act on the progesterone receptor. In spinal cord injured rats, progesterone reduces overexpression of the mRNAs of interleukin 1β‎ and interleukin 6, tumor necrosis factor, inducible nitric oxide synthase and cyclooxygenase 2, and thus neuroinflammation (Labombarda et al., 2015). There are also more experimental approaches to block neuroinflammation. For example, resolvins are anti-inflammatory lipid mediators generated during the resolution phase of an inflammation. Resolvins E1 and D1 can effectively dampen neuroinflammation and postoperative pain (Xu et al., 2010; Ji, Xu, Strichartz, & Serhan, 2011). Another example is the intrathecal injection of bone marrow stromal cells. This alleviates mechanical and thermal hypersensitivity in rats with a chronic constriction injury by secreting transforming growth factor-β‎1 into the cerebrospinal fluid (Chen, Park, Xie, & Ji, 2015). Transforming growth factor-β‎1 is an anti-inflammatory cytokine that potently inhibits the activation and proliferation of microglia and astrocytes. It thereby reduces the expression of proinflammatory cytokines under various conditions, including neuropathic pain (Echeverry et al., 2009; Chen et al., 2013). CD200 is a membrane glycoprotein of the immunoglobulin superfamily with immune suppression effects via its receptor CD200R that dampen microglia activation. Intrathecal application of CD200 fusion protein reduces mechanical and thermal hypersensitivity in animals with a neuropathy (Bennett, 1994). Finally, i.t. injections of curcumin, an anti-inflammatory component of turmeric, reduces the activation of glial cells and the production of inflammatory mediators, including interleukin-1β‎, monocyte chemoattractant protein-1, and monocyte inflammatory protein-1 in the spinal cord after inflammation of a hindpaw. This treatment also reverses hyperalgesia in arthritic rats (Chen et al., 2015).

Preemptive Strategies for Neuroplasticity

In principle, neuronal plasticity that is triggered by neuronal activity (i.e., neurogenic neuroplasticity) can be prevented by any means that reduces or abolishes the causative activity. In case of nociceptor-driven neuroplasticity, local anaesthetic infiltrations of injured tissues, sensory nerves, plexuses, or spinal cord appear straightforward for preventing neuroplasticity. Their usefulness is, however, limited by incomplete blocks and by their notoriously short duration of action. The release of neuromediators from the central terminals of nociceptive nerve fibers can be inhibited presynaptically, for example, with agonists acting at µ-opioid receptors (Terman, Eastman, & Chavkin, 2001) or α‎₂-adrenergic receptors (Ge et al., 2006), thereby blocking the induction of activity-dependent plasticity in nociceptive pathways. Enhancing endogenous antinociceptive and/or reducing pro-nociceptive systems can likewise prevent or reduce the development of neurogenic neuroplasticity (You et al., 2016; Drake, Hulse, Lumb, & Donaldson, 2014). Long-term potentiation at C-fiber synapses and hyperalgesia can further be blocked by preventing postsynaptic rise in Ca²⁺, for example, by NMDA receptor antagonists (Ikeda et al., 2003, 2006) or ryanodine receptors (Cheng, Lu, Zhang, & Zhao, 2010; Ohsawa & Kamei, 1999). Deep surgical anaesthesia under the volatile anaesthetics isoflurane or sevoflurane is, however, insufficient to prevent LTP induction at C-fiber synapses (Benrath, Brechtel, Martin, & Sandkühler, 2004; Ikeda et al., 2006). Interestingly, the anaesthetic noble gas Xenon is highly effective in preventing LTP induction (Benrath, Kempf, Georgieff, & Sandkühler, 2007), probably because of its NMDA receptor blocking effect. Up to now the translation of these findings into clinical practice has largely failed. The reasons for this failure are presently unknown.

Confinement Strategies for Neuroplasticity

Synaptic LTP is long-lasting but not necessarily permanent. Many forms of LTP vanish over time periods of hours or days, whereas some may persist for longer. From a therapeutical point of view, reversal of LTP can be important in cases where LTP contributes to pain amplification or chronicity. LTP at C-fiber synapses has been reversed by conditions such as stimulation at Aδ‎-fiber intensity (Liu, Morton, Azkue, Zimmermann, & Sandkühler, 1998), by a brief, systemic application of a very high dose of a µ-opioid receptor agonist (Drdla-Schutting, Benrath, Wunderbaldinger, & Sandkühler, 2012) or by diazepam (Hu et al., 2006); for a review, see Sandkühler and Lee (2013).