Volume 175, Issue 12 p. 2362-2374
REVIEW ARTICLE
Free Access

Calcium signalling through L-type calcium channels: role in pathophysiology of spinal nociceptive transmission

Olivier Roca-Lapirot

Olivier Roca-Lapirot

Interdisciplinary Institute for Neuroscience (IINS, CNRS UMR 5297), University of Bordeaux, Bordeaux Cedex, France

Search for more papers by this author
Houda Radwani

Houda Radwani

Interdisciplinary Institute for Neuroscience (IINS, CNRS UMR 5297), University of Bordeaux, Bordeaux Cedex, France

Search for more papers by this author
Franck Aby

Franck Aby

Interdisciplinary Institute for Neuroscience (IINS, CNRS UMR 5297), University of Bordeaux, Bordeaux Cedex, France

Search for more papers by this author
Frédéric Nagy

Frédéric Nagy

Interdisciplinary Institute for Neuroscience (IINS, CNRS UMR 5297), University of Bordeaux, Bordeaux Cedex, France

Search for more papers by this author
Marc Landry

Marc Landry

Interdisciplinary Institute for Neuroscience (IINS, CNRS UMR 5297), University of Bordeaux, Bordeaux Cedex, France

Search for more papers by this author
Pascal Fossat

Corresponding Author

Pascal Fossat

Interdisciplinary Institute for Neuroscience (IINS, CNRS UMR 5297), University of Bordeaux, Bordeaux Cedex, France

Correspondence Pascal Fossat, PhD, Interdisciplinary Institute for Neuroscience (IINS), University of Bordeaux, CNRS UMR5297. 146 rue Leo Saignat, 33076 Bordeaux Cedex, France. E-mail: [email protected]Search for more papers by this author
First published: 18 February 2017
Citations: 38

Abstract

L-type voltage-gated calcium channels are ubiquitous channels in the CNS. L-type calcium channels (LTCs) are mostly post-synaptic channels regulating neuronal firing and gene expression. They play a role in important physio-pathological processes such as learning and memory, Parkinson's disease, autism and, as recognized more recently, in the pathophysiology of pain processes. Classically, the fundamental role of these channels in cardiovascular functions has limited the use of classical molecules to treat LTC-dependent disorders. However, when applied locally in the dorsal horn of the spinal cord, the three families of LTC pharmacological blockers – dihydropyridines (nifedipine), phenylalkylamines (verapamil) and benzothiazepines (diltiazem) – proved effective in altering short-term sensitization to pain, inflammation-induced hyperexcitability and neuropathy-induced allodynia. Two subtypes of LTCs, Cav1.2 and Cav1.3, are expressed in the dorsal horn of the spinal cord, where Cav1.2 channels are localized mostly in the soma and proximal dendritic shafts, and Cav1.3 channels are more distally located in the somato-dendritic compartment. Together with their different kinetics and pharmacological properties, this spatial distribution contributes to their separate roles in shaping short- and long-term sensitization to pain. Cav1.3 channels sustain the expression of plateau potentials, an input/output amplification phenomenon that contributes to short-term sensitization to pain such as prolonged after-discharges, dynamic receptive fields and windup. The Cav1.2 channels support calcium influx that is crucial for the excitation-transcription coupling underlying nerve injury-induced dorsal horn hyperexcitability. These subtype-specific cellular mechanisms may have different consequences in the development and/or the maintenance of pathological pain. Recent progress in developing more specific compounds for each subunit will offer new opportunities to modulate LTCs for the treatment of pathological pain with reduced side-effects.

Linked Articles

This article is part of a themed section on Recent Advances in Targeting Ion Channels to Treat Chronic Pain. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v175.12/issuetoc

Abbreviations

  • BZT
  • benzothiazepine
  • CAN
  • calcium-activated non-specific cationic channels
  • cp8
  • compound 8
  • CREB
  • cAMP response element binding
  • DHNs
  • dorsal horn neurons
  • DHP
  • dihydropyridine
  • HVA
  • high-voltage-activated channels
  • KCa channels
  • calcium-dependent potassium channels
  • Kir channels
  • inwardly rectifying potassium channels
  • LTCs
  • L-type calcium channels
  • LTP
  • long-term plasticity
  • LVA
  • low-voltage-activated channels
  • mGlu receptors
  • metabotropic glutamate receptors
  • Phe
  • phenylalkylamine
  • VGCC
  • voltage-gated calcium channel
  • Tables of Links

    TARGETS
    Voltage-gated ion channelsa Enzymesb
    Cav1.1 channels COX2
    Cav1.2 channels GPCRsc
    Cav1.3 channels GABAB receptors
    Cav1.4 channels mGlu receptors
    KCa channels NK1 receptors
    Kir channels Ligand gated ion channelsd
    NMDA receptors
    LIGANDS
    Bay K8644 Nimodipine
    Diltiazem NS1619
    Gabapentin Pregabalin
    Nicardipine Verapamil
    • These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (a,b,c,dAlexander et al., 2015a,b,c,d).

    Introduction

    Voltage-gated calcium channels (VGCCs) are activated by changes in electrical membrane potentials and lead to calcium entry into the cells. In turn, calcium influx through these channels a wide range of cell functions such as muscle contraction (Reuter 1979; Tsien 1983; Catterall 1991; Tanabe et al. 1993; Bers 2002), hormone and neurotransmitter release (Tsien et al. 1988; Dunlap et al. 1995; Yang and Berggren, 2006; Catterall and Few, 2008) and the cell signalling cascade that leads to gene expression (Flavell and Greenberg, 2008).

    VGCCs are classified according to their physiological and pharmacological properties and comprise several subunits. The main subunit is a Cav subunit of about 2000 amino acids that form the channel pore responsible for Ca2+ influx (Tanabe et al., 1987; Catterall, 2011), with other accessory subunits (β, α2δ and γ) modulating the properties of the Cav subunit, e.g. kinetics and subcellular targeting). There are two main families of VGCCs, a low-voltage-activated family (LVA) and a high-voltage-activated family (HVA). LVA are commonly called T-type channels and comprise the Cav3 subunit as the pore-forming channel. There are four subgroups of HVA channels according to their pharmacological sensitivity. Cav1 or L-type calcium channels (LTCs) are blocked by the dihydropyridine (DHPs, such as nifedipine), the phenylalkylamines (Phe, such as verapamil) and the benzothiazepines (BZT, such as diltiazem). The LTC blockers of these three families have similar modes of action, that is, by blocking Cav1 channels in open or inactivated states by binding the IIIS5–6 and IVS6 transmembrane segments of the Cav1 subunit (Hockerman et al., 1997). The Cav2 family comprises N-type calcium channels (Cav2.2 channels) that are blocked by ω-conotoxin GVIA (Tsien et al., 1988; Olivera et al., 1994), P/Q-type calcium channels (Cav2.1 channels) that are blocked by ω-agatoxin IVA (Mintz et al., 1992; Randall and Tsien, 1995), and R-type calcium channels (Cav2.3 channels) that are blocked by SNX-482 (Newcomb et al., 1998). The Cav1 family consists of four channel subtypes (Cav1.1 to 1.4 channels) which are expressed in a wide range of cells in mammals (Catterall, 2011). However, in the CNS, only Cav1.2 and Cav1.3 subunits are present. Their differences are due to specific gating kinetics, voltage-sensitivity, pharmacology and spatial distribution. The Cav1.3 channel activates at a lower voltage than the Cav1.2 channel (Koschak et al., 2001; Xu and Lipscombe, 2001; Helton et al., 2005; Bock et al., 2011) and is less sensitive to blockade by the DHPs (Xu and Lipscombe, 2001; Sinnegger-Brauns et al., 2009). In the CNS, Cav1.2 and Cav1.3 channelsare predominantly post-synaptic channels and are expressed in the soma and dendrites (Di Biase et al., 2008; Jenkins et al., 2010). In the spinal cord, Cav1.2 and Cav1.3 channels differ in their neuronal distribution, with Cav1.3 channels being expressed in the dendritic tree, in ventral horn neurons and in the whole somato-dendritic compartment up to the distal dendrites in the dorsal horn (DH) while Cav1.2 channels are mostly restricted to the soma and proximal dendrites (Westenbroek et al., 1998; Simon et al., 2003; Dobremez et al., 2005; Radwani et al., 2016). The lower voltage-activated Cav1.3 subunit carries inward currents at threshold electrical membrane potentials and shapes neuronal firing (Striessnig et al., 2006). Downstream of the depolarization induced by Ca2+ through both types of LTCs, the level of intracellular Ca2+ controls neuronal excitability by activating calcium-dependent channels (e.g. calcium-activated non-specific cationic channels (CAN) (Partridge et al., 1994) and calcium-dependent potassium channels, (KCa channels) (Maljevic and Lerche, 2013). Ca2+entry through Cav1.2 channels modulates calcium-dependent signalling pathways (Avery and Johnston, 1996; Magee et al., 1996; Dolmetsch et al., 2001; Li et al., 2004; Olson et al., 2005). LTCs are essential for the physiopathology of neurons and are involved in many CNS functions such as learning, memory and emotions (McKinney and Murphy, 2006; White et al., 2008; Busquet et al., 2010; Lee et al., 2012). They also contribute to pathologies such as Parkinson's disease (Striessnig et al., 2015) and developmental disorders such as autism syndromes (Striessnig et al., 2015). There is now strong evidence that LTCs actually subsume pain processing in the DH of the spinal cord (Schaible et al., 2000; Fossat et al., 2007; Fossat et al., 2010; Favereaux et al., 2011; Radwani et al., 2016).

    This review first describes the functional properties supported by Cav1.2 and Cav1.3 channels in the DH of the spinal cord. Then it documents their involvement and their modulation in information transfer, sensitization and in pathological states associated with pain. Finally, potential therapeutic avenues for chronic pain treatment are discussed.

    L-type calcium channels and properties of the dorsal horn neurons of the spinal cord

    L-type calcium channels and plateau potentials

    Cav 1.2 and Cav 1.3 LTCs are expressed in the spinal cord both in the dorsal and the ventral horn (Perrier et al., 2002; Simon et al., 2003; Dobremez et al., 2005; Fossat et al., 2010; Radwani et al., 2016). They carry slowly inactivating calcium currents allowing the expression of regenerative membrane properties in dorsal horn neurons (DHNs), called plateau potentials. Plateau potentials may be considered as the active depolarization of the neuronal electrical membrane potential during a constant stimulating pulse (Figure 1A), which is carried by L-type calcium current blocked by the DHP nifedipine and is enhanced by Bay K8644 (Morisset and Nagy, 1996; Russo and Hounsgaard, 1996; Morisset and Nagy, 1999). Once the action potential threshold is reached, the active depolarization induces a progressive increase in discharge frequency (Figure 1A). Subsequently, calcium influx through LTCs activates a calcium-dependent non-specific cationic current (CAN) that maintains the neuron in a depolarized state and sustains an after-discharge after termination of the stimulation (Figure 1A) (Morisset and Nagy, 1999). Calcium influx through LTCs also activates KCa channels that restrict the amplitude of the plateau potentials and play a role in the termination of the after-discharge (Morisset and Nagy, 1999). Therefore, the expression of plateau potentials correlates with a clear increase in the neuronal response in both the amplitude and time domains, and plateau potentials may be seen as an intrinsic mechanism of input/output amplification.

    Details are in the caption following the image
    (A) DHNs express an activity-dependent bi-stable state called plateau potentials. Plateaus can be artificially activated by a depolarizing current pulse. The pulse activates high-threshold voltage-dependent calcium channels that promote a non-linear response of the neuron and a progressive increase in discharge frequency. At the end of the pulse, DHNs can be maintained in a high state owing to the activation of depolarizing, non-specific cationic, calcium-dependent, channels (Ican). DHNs can switch from being silent to activated. (B1) Changes in peripheral receptive field. Repetitive stimulations of the hind paw in a low probability firing zone induces a switch of the zone into a firing zone. (B2) Switch in peripheral receptive field can also be elicited by activation of plateau potentials. (C1) Plateau potential induces in a computational model of DHNs when Cav1.2 and Cav1.3 channels are introduced. (C2) When Cav1.2 channels are removed, plateau potentials are still expressed. (C3) When Cav1.3 channels are removed, plateaus are not expressed so that the Cav1.3 channels are necessary for the expression of plateau potentials.

    Activation of plateau potentials profoundly alters the output properties of DHNs in response to sensory inputs. It promotes the information transfer in the spinal DH by significantly increasing the probability that an afferent spike will elicit a DHN response (Derjean et al., 2003). Nociceptive primary afferent stimulation may elicit intense and prolonged responses in plateau-generating DHNs, whereas brief bursts of spikes are evoked in other ways. Moreover, by increasing responsiveness to weaker stimuli, plateau potentials in DHNs modify the cutaneous mechanoreceptive field of these neurons. Expression of plateaus expands the excitatory receptive field where skin stimulation has a high probability to elicit firing, and contracts the low-probability firing fringe of the field, thereby increasing the nociceptive transmission through DHNs (1) (Reali et al., 2011; Reali and Russo, 2013). Plateaus may also be generated repetitively, leading to rhythmic bursting (Jiang et al. 1995; Derjean et al. 2003), a pattern associated with distortion of nociceptive processing in chronic pathological pain (Calvino et al., 1987a,b). In vitro, up to 90% of deep DHNs express regenerative properties under appropriate neuromodulation (Derjean et al. 2003), whereas in vivo, the proportion of plateau-generating neurons in lamina I and V increases in persistent pain conditions (Dougherty and Hochman, 2008; Reali et al. 2011), rising from 28% of the recorded neurons in the control condition to more than 80% in neuropathic pain (Reali et al., 2011). Taken together, these data indicate that, in the context of nociceptive transmission and pain, LTC-activated plateau potentials greatly contribute to the switch from the normal to the sensitized system. As the dendritic tree is the preferential target of synaptic inputs, Cav1.3 channels are strategically located to elicit plateau potentials and amplify nociceptive inputs. In fact, a model of DHNs strongly suggests that the Cav1.3 channels are the major contributors to plateau potentials (Radwani et al., 2016; Figure 1C).

    Finally, targeting plateau potentials with appropriate blockers should change the enhanced excitability of DHNs observed in the neuropathic pain model and may ultimately serve to reduce pain in chronic pain sufferers (see below).

    L-type calcium channels and short-term sensitization: windup of DHN discharges

    In the spinal cord, both DHNs and motor neurons express a form of short-term activity-dependent sensitization called windup of discharge. Windup is a progressive increase in neuronal discharge in response to low-frequency stimulations, which eventually ends up in a prolonged after-discharge (Mendell, 1966; Woolf and Wall, 1986; Thompson et al., 1990; Urban et al., 1994). Windup sensitization lasts several minutes after the cessation of the nociceptive stimulations. Windup and central sensitization to pain clearly depend on different mechanisms (Woolf, 1996). However, as the amplitude of windup may increase in models of persistent pain, it has long been used as an indicator of the level of central sensitization to pain (Baranauskas and Nistri, 1998; Herrero et al., 2000). In the ventral horn of the spinal cord, discharge of motor neurons can also wind up during repetitive stimulation. It is therefore possible to evaluate windup in vivo by recording nociceptive sensitization of the paw withdrawal reflex in animals, and the equivalent RIII reflex in human. Alteration of such sensitization is used as a cue for pathological pain (Gozariu et al., 1997). In humans, fibromyalgia patients have an exacerbated windup of second pain, a late nociceptive response expressed upon nociceptive stimulation (Price et al., 2002).

    Windup in the DH is mainly expressed by deep convergent neurons also called wide dynamic range neurons (Herrero et al., 2000). This depends in part on a synaptic component involving NMDA and neurokinin receptors (Baranauskas and Nistri, 1998; Herrero et al. 2000). However, as shown both by in vitro and in vivo experiments, windup of DHN discharges also critically depend on activation of the two main conductances of the plateau potential, IL (i.e. through LTCs) and Ican (through the CAN channels). Windup is suppressed by the DHP nifedipine, the Phe verapamil and the BZT diltiazem and is enhanced by BayK8644, both in vitro and in vivo (Table 1, see also Russo and Hounsgaard, 1994; Morisset and Nagy, 2000; Fossat et al., 2007; Yamamoto et al., 2012). Moreover, none of the synaptic components elicited by nociceptive fibre stimulation, including the NMDA-receptor-mediated response, is sufficient on its own to induce windup when conductances of the plateau potential are blocked pharmacologically. Conversely, when NMDA receptors are blocked, windup can still be evoked providing that the inhibitory inputs to DHNs are suppressed (Morisset and Nagy, 2000; Fossat et al. 2007). Therefore, windup primarily depends on the intrinsic amplification properties of spinal neurons, whereas, in contrast to previous hypotheses, NMDA receptors do not directly mediate the sensitization of spinal neurons during windup, but provide the appropriate excitation that ultimately activates LTCs within a dynamic balance of excitatory and inhibitory inputs. This demonstrates the co-operation between NMDA receptors and LTCs, necessary to this form of short-term sensitization related to pain.

    Table 1. Pharmacology of LTCs: role in pain states
    LTCs Blockers and pain states
    Family LTC Blockers Injection Functional consequences Pain state Ref
    Antisense Cav1.2 i.t. Ø Mechanical allodynia

    Neuropathy

    (Fossat et al., 2010)
    Cav1.3 i.t. Ø Windup Control (Radwani et al., 2016)
    Phenylalkylamines Verapamil i.t. Ø Windup Control (Fossat et al., 2007) (Yamamoto et al., 2012)
    Verapamil = Tactile allodynia Neuropathy (Chaplan et al., 1994)
    Verapamil i.t. = DHN discharge Formalin (Diaz and Dickenson, 1997)
    Verapamil i.t. = First response Control (Fossat et al., 2007)
    Benzothiazepines Diltiazem i.t. DHN discharge Inflammation (Neugebauer et al., 1996)
    Diltiazem Ø Windup Control (Yamamoto et al., 2012)
    Diltiazem p.o. Cold hyperalgesia Chemotherapy-induced neuropathy (Kawashiri et al., 2012)
    Diltiazem i.t. or i.v. = Tactile allodynia Neuropathy (Chaplan et al., 1994)
    Dihydropyridines

    Nifedipine

     

    Nifedipine

    i.t.

     

    i.t.

    = DHN discharge

     

    Ø Windup/plateaus

    Formalin

     

    Control

    (Diaz and Dickenson, 1997)

    (Fossat et al., 2007) (Morisset and Nagy, 1999) (Morisset and Nagy, 2000) (Yamamoto et al., 2012)

    Nifedipine i.t. = First response Control (Fossat et al., 2007)
    Nicardipine Osmotic pump, i.t. Tactile allodynia Neuropathy

    (Fossat et al., 2010)

    Nifedipine p.o. Cold hyperalgesia Chemotherapy-induced neuropathy (Kawashiri et al., 2012)
    Nimodipine topical i.t. or iontophoresis i.t. DHN discharge Inflammation (Neugebauer et al., 1996)
    Nimodipine i.t. or i.v. = Tactile allodynia Neuropathy (Chaplan et al., 1994)
    M4 i.p. or i.t. ↘ Tactile allodynia Neuropathy (Gadotti et al., 2015)
    M4 i.p. or i.t. ↘ Formalin test Control (Gadotti et al., 2015)
    PYT (cp8) Cav1.3 ? ? ? (Kang et al., 2012)
    • ↘, decrease; Ø, suppress; =, no change. PYT, pyrimidine-2,4,6-triones.

    Given that LTCs are molecular determinants of windup, the question arises of the degree of involvement of Cav1.2 and Cav1.3 channels. As described above, Cav1.3 channels are expressed in the dendritic tree and activates at electrical membrane potential around the threshold. They are therefore good candidates to mediate windup (Xu and Lipscombe, 2001; Simon et al., 2003; Radwani et al., 2016). The windup of a nociceptive flexion reflex is completely abolished when expression of Cav1.3 channels is blocked with an antisense strategy (Radwani et al., 2016). In agreement, a model of DHN neurons suggests that Cav1.3 channels are essential to elicit windup of a DHN discharge. By contrast, Cav1.2 channels are not necessary for the expression of windup, although blockade of the expression of Cav1.2 channels slightly decreases its amplitude (Radwani et al., 2016).

    L-type calcium channels in long-term plasticity of DHNs

    Unlike short-term plasticity such as windup, functional changes occurring in chronic pain may last for months or years. Many studies emphasize the possible role of long-term synaptic plasticity in the mechanisms of persistent central sensitization (Sandkuhler, 2007; Sandkuhler, 2009). In the DH of the spinal cord, long-term potentiation of the nociceptive pathway can be elicited with a large barrage of afferent inputs from nociceptive fibres (Liu and Sandkuhler, 1995). This long-term plasticity (LTP) of DHNs primarily depends on synaptic receptors, such as the NK1 or NMDA receptors (Liu and Sandkuhler, 1995; Liu and Sandkuhler, 1997). Their activation elicits calcium entry in post-synaptic DHNs that acts as a second messenger inducing mobilization of receptors and increases the depolarization induced during synaptic transmission. Thus, the post-synaptic response and hence nociceptive transmission are increased. Maintenance of LTP is also due to calcium entry that activates a signalling cascade leading to gene activation. LTP can also be triggered by chemical treatments that target metabotropic glutamate (mGlu) receptors (Azkue et al., 2003). Activation of mGlu receptors induces calcium waves that can be prevented by LTC blockers, suggesting a possible role of post-synaptic LTCs in LTP (Heinke and Sandkuhler, 2007). Moreover, it is well established that LTCs mediate LTP in many brain regions including the visual cortex, hippocampus or amygdala (Aroniadou and Teyler, 1992; Johnston et al., 1992; Weisskopf et al., 1999). LTCs are also involved in a presynaptic form of LTP in the lateral nucleus of the amygdala (Fourcaudot et al., 2009). In spinal cord slices, LTP can be induced by paired stimulations that couple a strong afferent barrage to a post-synaptic depolarization. Blockade of LTCs with the DHP nifedipine totally abolished this form of LTP, again suggesting the role of these channels in the LTP of DHNs (Naka et al., 2013).

    Therefore, LTCs are important post-synaptic components that participate in both the short- and long-term mechanisms of plasticity. However, their actual role in pain processing remains largely controversial.

    L-type calcium channels and pain transmission

    Acute pain

    In the DH, LTCs contribute to the electrical membrane potential of post-synaptic DHNs (Heinke et al., 2004). They activate slowly and at a rather high electrical membrane potential, especially the Cav1.2 channels. Thus, their activation requires a strong- or long-enough DHN depolarization and it is unlikely whether they participate in acute nociceptive transmission. Indeed, the acute nociceptive withdrawal reflex response was not modified by the DHP nifedipine (Fossat et al., 2007), and the Phe verapamil did not modify the response of DHNs to formalin injection (Diaz and Dickenson, 1997) (Table 1). The DHP nimodipine and the BZT diltiazem decreased the evoked response of WDR neurons to noxious and innocuous stimulation of the knee joint (Neugebauer et al., 1996; Schaible et al., 2000) (Table 1), but the stimulations used lasted 15 s, so they may have induced LTC-dependent sensitization. Moreover, in animals that had received a 4 day injection of antisense molecules targeting either Cav1.2 or Cav1.3 channels, the paw withdrawal threshold measured with the von Frey test remained unchanged with blockade of Cav1.3 channels, confirming the weak contribution of this subunit to acute nociceptive stimulation. In contrast, block of Cav1.2 channels slightly increased the paw withdrawal threshold, suggesting that these channels could play a role in acute pain transmission in naive animals (Fossat P, unpublished data).

    L-type calcium channels in model of persistent pain

    Inflammatory pain

    Inflammatory pain is a form of persistent pain generated by peripheral release of inflammatory molecules that modify the excitability of nociceptive afferent fibres and DHNs (Moalem and Tracey, 2006). Persistent increase in DHN hyperexcitability can exceed the time course of action of inflammatory molecules and generate persistent pain syndromes. Mustard oil, kaolin or complete Freund's adjuvant are commonly used as pro-inflammatory stimuli. Inflammation promotes long-term changes in afferent fibres and DHNs and durably induces the hyperexcitability of sensory neurons and central sensitization. This increased excitability promotes the activation of LTCs, which are likely to be involved in inflammation-induced central sensitization. Indeed, blockade of LTCs with an intrathecal application of the DHP nimodipine or the BZT diltiazem significantly reduced the evoked discharge of DHNs in a model of knee joint inflammation induced by mustard oil or kaolin (Neugebauer et al., 1996; Schaible et al., 2000) (Table 1). These results indicate that LTCs play a role in nociceptive transmission after inflammation. However, no data are available regarding the putative specific roles of Cav1.2 or Cav1.3 channels.

    Neuropathic pain

    Neuropathic pain is a chronic pain syndrome caused by lesions of the nervous system. It is the most widespread chronic pain syndrome and affects around 8% of the population (Bouhassira et al., 2008). To study the mechanisms leading to chronic pain after nerve injury, various animal models are commonly used that consist in cutting or ligating the peripheral sensory roots or the sciatic nerve (Bennett and Xie, 1988; Seltzer et al. 1990; Kim and Chung, 1992; Kim et al. 1997; Decosterd and Woolf, 2000). In all these models, chronic pain syndromes are characterized by two phenomena: allodynia and hyperalgesia (Sandkuhler, 2009). Allodynia is a decreased pain threshold that induces pain reaction to normally innocuous stimulations, while hyperalgesia is an increased/exaggerated pain response to noxious stimulations. It is now clear that VGCCs such as N- and T-type are involved in the development and maintenance of neuropathic pain syndromes (Seward et al., 1991; Bourinet et al., 1996; Bourinet et al., 2005; Gribkoff, 2006). These channels predominantly alter the presynaptic compartment and synaptic transmission. By contrast, LTCs are predominantly post-synaptic channels that control the excitability of DHNs. The role of LTCs has long been difficult to highlight and seems to depend on the mode of application of the agonists and antagonists. For instance, acute intrathecal application of the LTC blockers, nimodipine, verapamil or diltiazem, did not significantly influence tactile allodynia in the sciatic nerve ligation (SNL) model, whereas continuous infusion of the DHP nicardipine through osmotic pumps did (Table 1) (Chaplan et al., 1994; Fossat et al., 2010). Moreover, it has been recently demonstrated in a model of neuropathy induced by oxaliplatin (chemotherapeutic drug) that cold hyperalgesia is prevented by oral administration of the DHP nifedipine or the BZT diltiazem (Kawashiri et al., 2012) (Table 1). Finally, the major argument against the role of LTC in neuropathic pain processing is that the many treatments for cardiovascular diseases with oral or systemic DHP in humans yield no analgesic effects. However, according to animal studies on DHP bioavailability in brain ventricles after systemic injection (Tsukahara et al., 1989), there is approximately a difference of three orders of magnitude between intra-arterial and CSF concentrations. At therapeutic intravenous doses, the concentration of DHP in the CSF may be far too low to exert any anti-hyperalgesic effect (Porchet et al., 1992; Cheung et al., 1999). In any case, there is now strong evidence that LTCs are involved in neuropathic pain (Fossat et al., 2010; Favereaux et al., 2011; Radwani et al., 2016).

    Cav1.2 channels mediate calcium transients in persistent pain conditions

    Cav1.2 channels are overexpressed in the DH of the spinal cord in the SNL model of neuropathy (Dobremez et al., 2005; Fossat et al., 2010; Radwani et al., 2016). Moreover, when Cav1.2 channels are knocked-down by using an antisense approach in vivo, SNL-induced tactile allodynia and DHN hyperexcitability are abolished (Fossat et al., 2010) (Table 1). The effect of Cav1.2 channels may be due to calcium influx-induced gene expression through activation of the cAMP response element binding (CREB) transcription factor. This effect is mediated by an intranuclear calcium rise and is likely to be independent of classical intracellular pathways involving CamKinase II and PKA (Fossat et al., 2010). In addition, SNL animals show enhanced transcription of the CREB/CRE-dependent gene COX-2 and increased expression of COX-2, an eicosanoid enzyme whose activity is induced by inflammation or nerve injury. This over-expression is also reversed by knockdown of Cav1.2 channels (Fossat et al., 2010). All these results were obtained after the development of SNL-induced neuropathic syndromes, thus showing that Cav1.2 channels play an important role in the maintenance of neuropathic states.

    These data are in agreement with data from a previous study indicating that the Cav1.2 protein has a unique dual function as an ion pore and as a transcription factor (Gomez-Ospina et al., 2006). The authors reported that the C-terminal fragment of the Cav1.2 channel translocates to the nucleus upon channel activation and regulates transcription. Moreover, our seminal study reported Cav1.2 channel-mediated calcium transients in the nucleus of spinal neurons, and suggested that nuclear calcium serves as a signalling endpoint in pain-triggered synapse-to-nucleus communication (Fossat et al., 2010). As an important regulator of neuronal gene expression, nuclear calcium is involved in the conversion of synaptic stimuli into functional and structural changes of neurons (see Bading, 2013). Pain-associated calcium transients result in the regulation of a distinct spinal gene pool including both pain-related and plasticity-related genes (Simonetti et al., 2013). The same study also demonstrated that inhibition of nuclear calcium signalling attenuates the development of chronic inflammatory pain. Importantly, the authors demonstrated that blocking nuclear calcium transients also blocked activation of CREB, dependent on nociceptive activity, while ERK1/2 activation remained unchanged. This is in line with our results showing that the nuclear calcium influx mediated by Cav1.2 channels triggered CREB, but not ERK1/2, activation in neuropathic pain conditions (Fossat et al., 2010).

    Changes in Cav1.3 channels in neuropathic pain alter neuronal discharge

    By contrast, the role of Cav1.3 channels in persistent pain is less clear. First, expression of Cav 1.3 channels is decreased together with the amplitude of windup in the SNL model of neuropathy (Dobremez et al., 2005; Fossat et al., 2010; Radwani et al., 2016). Second, blockade of the expression of Cav1.3 channels had no effect on tactile allodynia or hyperexcitability of DHNs in the same model (Fossat et al., 2010). Considering that Cav1.3 currents are one of the main conductances responsible for the expression of plateau potentials, the latter cannot be involved in the hyperexcitability of DHNs after nerve lesion. However, in the SNL model of neuropathy, the number of DHNs expressing plateau potentials is significantly increased (Reali et al., 2011). Moreover, exaggerated after-discharges similar to those produced by plateau-generating DHNs are observed in vivo in models of persistent pain (Laird and Bennett, 1993; Sotgiu et al., 1995). Finally, computational approaches suggest that the decreased expression of Cav1.3 channels in the SNL model is sufficient to decrease the windup amplitude, whereas plateau potentials remain unaffected (Radwani et al., 2016). One possibility is that the low density of Cav1.3 channels could be sufficient to activate CAN but not KCa channels . Alternatively, Cav1.2 channels, whose expression is increased in the whole dendritic compartment in SNL animals (Dobremez et al., 2005), could compensate for the down-regulation of Cav1.3 channels and thus play a role in plateau potentials. Taken together, these results suggest that the hyperexcitability of DHNs after SNL facilitates the expression of plateau potentials in DHNs, which in turn would be involved in exaggerated pain sensations. Finally, an increase in the population of DHNs expressing a plateau together with a decrease in windup amplitude of individual DHNs are likely to induce an alteration of the signal-to-noise ratio that makes it possible to discriminate between pertinent nociceptive inputs and irrelevant stimulations (Derjean et al., 2003).

    Therapeutic potential of LTC regulation

    LTCs play a role in the central sensitization to pain with a dendritic Cav1.3 channel involved in short-term sensitization and a somatic Cav1.2 channel in the long-term phenomenon related to pathological pain processes (Figure 2). In this scenario, the Cav1.2 channels allow a somatic calcium entry leading to changes in gene expression that are involved in the maintenance of neuropathic syndromes (Fossat et al., 2010). For therapeutic purposes, LTCs can be targeted directly with pharmacological agents or indirectly through their modulators.

    Details are in the caption following the image
    Overall scheme of the respective role played by Cav subunits in short- and long-term sensitization of pain.

    Direct pharmacological targeting of LTCs

    Targeting Cav1.2 channels seems to be more appropriate because these channels are directly involved in nerve lesion-induced neuropathy and Cav1.2 channels are more sensitive to classical LTC blockers. Unfortunately, owing to their role in cardiovascular physiology, it will be very difficult to develop drugs that specifically target spinal Cav1.2 subunits (Striessnig et al., 2015). Local delivery or indirect targeting (see below) should be considered to circumvent this difficulty.

    Because of their distribution and effects on the electrical membrane potential through generation of plateau potentials, Cav1.3 channels could indirectly play a role in long-term changes, providing that the depolarizing conductance is large and sustained. The recent development of more specific modulators of Cav1.3 channels such as the pyrimidine 2,4,6-trione and particularly the compound 8 (cp8) could help in deciphering the exact role of Cav1.3 channels in the patho-physiology of nociceptive transmission (Kang et al., 2012; Ortner et al., 2014; Zamponi et al., 2015). For instance, it will be interesting to assess the effect of cp8 on acute nociceptive transmission as well as on windup expression or plateau potentials in both control and neuropathic pain models. Ca2+ influx through Cav1.3 channels promotes activation of calcium-dependent channels. On the one hand, Cav1.3 channels activate CAN and exaggerate after-discharge. Therefore, therapeutic approaches blocking CAN (e.g. with flufenamic acids) could be of interest. On the other hand, Cav1.3 channels control KCa channels, which are important in limiting neuronal hyperexcitability. Interestingly, recent studies revealed the plasticity of KCa channels in various models of neuropathy, though mainly in dorsal root ganglion (Chen et al., 2009; Liu et al., 2015). Moreover, opening KCa channels with the channel opener NS1619 reduces allodynia and hyperalgesia in the SNL model of neuropathy (Chen et al., 2009). Therefore, activating KCa channels should also contribute to decreasing DHN hyperexcitability.

    Finally, it would also be interesting to improve the specificity of LTC blockers so as to avoid undesirable effects. For instance, the Phe and BZT compounds can also block other channels at higher concentrations (Hockerman et al., 1997). All LTC blockers used so far are specific to the open/inactivated state of LTC channels, meaning that LTCs need to be activated first. In heart physiology, rhythmic activity promotes the binding of LTCs. In the pain pathway, new drugs targeting the close state of LTCs could be an interesting approach. LTCs can also be blocked by peptides whose potential role in nociception and pain remain to be elucidated. Among them, calciseptine from the black mamba (Dendroaspis polylepis) or calcicludine from the green mamba (Dendroaspis angusticeps) are also good candidates to modulate nociception through LTCs, although the latter also binds to Na+ channels (Zamponi et al., 2015). Finally, developing new drugs jointly targeting several VGCCs specifically involved in pain disorders would be a good alternative. For instance, the new 1,4 DHP compounds that target Cav3.2 and Cav1.2 channels would be excellent candidates in the treatment of neuropathic pain (Gadotti et al., 2015).

    Besides the direct pharmacological regulation of LTCs, another possibility would be to target the modulators of these channels, as proposed below.

    Indirect modulation of LTCs

    The therapeutic benefits of pharmacological treatment are greatly limited by the ubiquitous distribution of LTCs and especially by their expression in the cardiovascular system where they participate in heart contraction and arterial resistance. Cav1.2 channels are the most prominent subunit in the cardiovascular system and one can hardly develop therapeutic strategies strictly restricted to the CNS. Indeed, only local application would be relevant to avoid collateral damage to the heart and vessels. Another approach would be to target Cav channels indirectly by interfering with their modulators.

    The LTCs are complex voltage-gated, pore-forming, channels that comprise several accessory subunits (β, α2δ and γ) that regulate the properties of the channels. The α2δ subunits increase maximal currents by increasing trafficking of the Cav1 subunits towards the plasma membrane (Hoppa et al., 2012). Gabapentin and pregabalin are two compounds promoting pain relief in neuropathic patients and both are modulators of the α2δ subunit (Field et al., 2006; Li et al., 2006). The β subunits located in the cytosol also increase maximal currents by enhancing trafficking of the Cav1 subunits and increasing the probability of opening (Lee, 2013). Mice deficient in β3 subunits have a decreased behavioural response to noxious stimuli (Murakami et al., 2002; Murakami et al., 2007). However, these accessory subunits are comprised in the various VGCCs and their modulation would influence other VGCCs involved in the pain pathway such as Cav2.1, Cav2.2 and Cav3.2 channels. Therefore, targeting these subunits would not offer specific control of LTCs.

    GPCRs are ubiquitous modulators in the CNS that control many of the neuronal functions including plasticity phenomena. In the DH of the spinal cord, mGlu and GABAb receptors are well characterized modulators of nociceptive transmission (Goudet et al., 2009; Landry and Nagy, 2009; Chiechio and Nicoletti, 2012). Group I mGlu receptor agonists are pro-algesic molecules that increase DHN responsiveness as well as LTP (Azkue et al., 2003; Chiechio and Nicoletti, 2012; Laffray et al., 2012). This effect is mediated by calcium entry into DHN (Heinke and Sandkuhler, 2007). Moreover, plateau potentials are strongly potentiated by application of group I mGlu receptor agonists in the DH (Morisset and Nagy, 1996; Russo et al., 1997; Derjean et al., 2003), even though there is no direct modulation of L-type VGCCs by these agonists (Voisin and Nagy, 2001). Induction and promotion of plateau potentials is mediated by down-regulation of inwardly rectifying potassium channels (Kir3.x), which at threshold potential prevents the activation of the plateau (Derjean et al., 2003). Baclofen, a GABAB receptor agonist, has anti-nociceptive effects in both acute and chronic pain models (Goudet et al., 2009; Landry and Nagy, 2009). Post-synaptic effects of activated GABAB receptors (Perez-Garci et al., 2006) may be exerted through LTCs. Baclofen directly blocks L-type calcium currents of DHN (Voisin and Nagy, 2001). In addition, it activates the Kir3 channel (Derjean et al., 2003) and, as expected, both effects contribute to suppressing plateau potentials (Russo et al., 1998; Derjean et al., 2003). Therefore, by directly and indirectly controlling L-type VGCC activation, a dynamic balance between antagonistic modulatory inputs determines the capabilities of DHNs to transfer nociceptive information at any moment. Interestingly, in the SNL model of neuropathy, GABAB receptor signalling in the DH is impaired by overexpression of a chaperone protein of the GABAB receptor (Laffray et al., 2012). Blockade of the interaction with this protein restores the functionality of the GABAB receptor and enhances the anti-nociceptive effect of baclofen. Such impairment of GABAB receptor signalling is likely to result in disinhibition of LTCs and hence contribute to central sensitization.

    Other neuromodulators also alter LTCs in the DH of the spinal cord as they do for motor neurons in the ventral horn. For instance, 5-HT (serotonin) is well known to modulate L-type calcium currents and plateau potentials (Hounsgaard and Kiehn, 1989; Perrier and Hounsgaard, 2003). However, although they act on LTCs, agonists or antagonists of metabotropic receptors are not specific to these channels and may target many other molecular pathways, thus limiting their utilization for therapeutic purposes.

    Epigenetic modulators are likely to contribute much more specifically to the regulation of LTC signalling. Such is the case for miRNA, which have been known for more than a decade to modulate pain in the CNS (Elramah et al., 2014). miRNA control gene expression by targeting mRNA and inducing their degradation (Elramah et al., 2014). The miRNA miR-103 represses the expression of all three Cav1.2 subunits by targeting a 3′-UTR miRNA-binding sequence on Cacna1c, Cacna2d1 and Cacnab1 genes, coding for Cav1.2, α2δ and β1 subunits respectively (Favereaux et al., 2011). In the rat SNL model of neuropathy, miR-103 is down-regulated in the spinal cord, inducing an up-regulation of the three LTC subunits. Normalizing miR-103 expression in SNL rats also restores basal levels of all subunit transcripts, and more importantly alleviates both mechanical and cold allodynia in SNL rats. Screening miRNA specific to the CNS and targeting Cav1.2 subunits could be an interesting approach to developing therapeutic approaches that would avoid side-effects on the cardiovascular system.

    Concluding remarks

    In the DH of the spinal cord, Cav1.2 and Cav1.3 channels are mainly involved in sensitization to pain with little participation in nociceptive transmission per se. This makes them interesting targets to act specifically on pain-associated alterations of nociceptive transmission. The subcellular distribution and the kinetics of activation support the involvement of Cav1.3 channels in short plasticity phenomena such as expression of plateau potential and windup, while the Cav1.2 channels are responsible for calcium influx that elicits alteration of gene expression and the long-term changes associated with persistent pathological pain. However, it cannot be ruled out that Cav1.3 channel-mediated depolarization of DHNs could enhance activation of Cav1.2 channels after nerve injury and indirectly play a role in the establishment of chronic pain syndromes. Unfortunately, the ubiquitous presence and the fundamental role of LTCs in cardiovascular functions underline the need to develop tissue-specific molecules that are able to target neuronal LTCs. Finally, it has become increasingly clear that VGCCs are key elements in the spinal DH that act both at presynaptic and post-synaptic sites to integrate and transform the strong barrage of peripheral nociceptive inputs following nerve injury or inflammation into long-term changes of pain networks that durably modify pain sensation.

    Acknowledgements

    This work was supported by Conseil Général d'Aquitaine (N°11004369/11004370), Agence Nationale pour la Recherche (ANR MirPain), Labex ‘BRAIN’ and Erasmus Mundus Green It.

      Conflict of interest

      The authors declare no conflicts of interest.