The amygdala between sensation and affect: a role in pain
© Veinante et al.; licensee BioMed Central Ltd. 2013
Received: 19 March 2013
Accepted: 11 May 2013
Published: 5 June 2013
The amygdala is a structure of the temporal lobe thought to be involved in assigning emotional significance to environmental information and triggering adapted physiological, behavioral and affective responses. A large body of literature in animals and human implicates the amygdala in fear. Pain having a strong affective and emotional dimension, the amygdala, especially its central nucleus (CeA), has also emerged in the last twenty years as key element of the pain matrix. The CeA receives multiple nociceptive information from the brainstem, as well as highly processed polymodal information from the thalamus and the cerebral cortex. It also possesses the connections that allow influencing most of the descending pain control systems as well as higher centers involved in emotional, affective and cognitive functions. Preclinical studies indicate that the integration of nociceptive inputs in the CeA only marginally contributes to sensory-discriminative components of pain, but rather contributes to associated behavior and affective responses. The CeA doesn’t have a major influence on responses to acute nociception in basal condition, but it induces hypoalgesia during aversive situation, such as stress or fear. On the contrary, during persistent pain states (inflammatory, visceral, neuropathic), a long-lasting functional plasticity of CeA activity contributes to an enhancement of the pain experience, including hyperalgesia, aversive behavioral reactions and affective anxiety-like states.
KeywordsAmygdala Hypoalgesia Hyperalgesia Affective/emotional component Persistent pain
Central amygdaloid nucleus
Lateral part of the central amygdaloid nucleus
Capsular part of the central amygdaloid nucleus
Medial part of the central amygdaloid nucleus
Calcitonin gene-related peptide
Extracellular signal-regulated kinase
Metabotropic glutamatergic receptor
Protein kinase A
Tumor necrosis factor α.
Many cerebral structures, constituting the pain matrix, participate in nociception and in the expression and modulation of pain. They transmit and decode nociceptive information, generate, amplify or reduce the pain sensation, allow expression of defensive or distress behavior, and influence the affect. Among these brain structures, the amygdaloid complex, or amygdala, appears as a key component of the pain matrix . Indeed, while this group of sub-cortical nuclei within the temporal lobe is well known for its critical role in the control of emotions , it has also been the target for a large number of fundamental research in the pain field during the two past decades. Overall, the amygdala is considered to provide an emotional value – either positive or negative – to sensory information, thus leading to adapted behavioral and affective responses and contributing to emotional memory. This role has been more particularly studied in the conditioned fear experimental paradigm [2–4]. Such conditioning is an adaptive process in response to potential danger. As nociceptive stimulations are also potentially harmful for body integrity, it is not surprising that the amygdala can integrate various nociceptive information to initiate or modulate the autonomic and behavioral responses according to the environmental context (internal and external) and to the affective state. This adaptation includes mechanisms of hypoalgesia/analgesia that allows reflex inhibition and facilitates fight or flight responses. Moreover, this brain structure participates in the anticipation/prediction of potential danger, based on salient sensory cues (sounds, images, odors…), to initiate the same adaptive responses as in direct presence of the danger. This is for example leading to conditioned or to stress-induced analgesia.
Pain includes a strong emotional component making it aversive and a reciprocal relation exists between pain and affective states. Stress can either inhibit or amplify pain; and anxiety as well as depressive states are often associated with more intense pain sensation, particularly when pain becomes chronic. In such chronic pain state, the amygdala could contribute to the hyperalgesia as well as to the anxio-depressive consequences of pain.
A few publications reviewed the role of the amygdala in persistent pain [5, 6]. The present review will consider the recent published data on the subject and address the key role of amygdala, especially its central nucleus, in various pain processes, either physiological or pathological. This will be done by reminding the anatomical context and presenting the morphological and functional evidences allowing insertion of the amygdala within the pain matrix. Then, we will show that the amygdala can exert either inhibitory or facilitating action on nociception and pain as well as on its affective component and consequences. Finally, we will address recent electrophysiological, neurochemical and biochemical data that are giving insights into underlying cellular and molecular mechanisms. While experimental data that are presented concern studies conducted in rodents, we will confront them with human brain imaging studies, thus stressing the potential role of the amygdala in human pain, especially with clinical relevance.
Amygdala and extended amygdala
The central nucleus of the amygdala (CeA), main component of the central group of amygdala nuclei, is considered as the output nuclei of the amygdaloid complex. It integrates information treated by the corticobasal group and it influences effector centers. The CeA is also reciprocally connected to more rostral forebrain structures, such as the lateral part of the bed nucleus of the stria terminalis (BSTL) and the dorsal part of the substantia innominata. This group of interconnected structures presents strong morphofunctional homogeneity and is designed as the central extended amygdala [8, 13]. In this review, we will more specifically detail data on the CeA, which is the most studied amygdaloid structure for its role in nociception and pain, and also refer to findings on the extended amygdala, as CeA functions are closely related to this forebrain macrostructure.
The CeA receives multiple nociceptive informations
Among the various CeA afferents, two main pathways are preferentially conveying nociceptive information (Figure 2). A first pathway originates from the BLA, and conveys highly integrated polymodal information, including nociceptive, from the thalamus and the cerebral cortex, allowing for example fear conditioning [2, 4, 7]. A cascade of projections, originating in the ventroposterior, posterior, triangular and posterior intralaminar nuclei of the thalamus, the second somatosensory area and the insular cortex, brings nociceptive information to the BLA which in turn transmits it to the CeA [7, 17, 18]. More precisely, the lateral nucleus of the amygdala projects lightly to the CeLC and the CeL and massively to the basolateral nucleus which targets preferentially to the CeM [4, 7, 11]. Additionally, lateral and basolateral projections to the CeA can be relayed in the intercalated cell masses, providing an inhibitory interface in the BLA-CeA pathway [4, 12]. Moreover, the CeA also receives less dense but direct projection from these thalamic and cortical areas [17, 18]. These largely polysynaptic efferents thus allow the CeA to integrate a nociceptive information which already acquired affective and cognitive significance in cortical circuits.
A second pathway conveys to the CeA more direct and raw nociceptive information. While spinal cord only sparsely sends direct projection to the amygdala, spinal nociceptive information is largely transmitted to the CeA via the parabrachial nucleus (PB). This pontine integrative center is a major target for superficial layers of the spinal cord, but also for deeper layers, for the trigeminal complex and for the nucleus of the solitary tract [19, 20]. The nociceptive aspect of these afferents was clearly established by electrophysiology . The PB thus gathers nociceptive information from all cutaneous, deep tissue and visceral territories. Three main ascending pathways are then originating from the PB, toward the medial thalamus, the medial hypothalamus and the central extended amygdala . The parabrachio-amygdaloid pathway is highly organized, with a specific topography connecting various PB subdivisions to the ones of the CeA and the BSTL [19, 20]. Moreover, unitary analysis of parabrachio-amygdaloid axonal branching showed that the different components of the central extended amygdala are innervated by axonal collaterals from a same PB neuronal population that never send projection to the medial thalamus or the medial hypothalamus .
Reported changes in the rodent CeA in different pain situations
Changes in the CeA
Acute somatic stimulations
Changes (mostly excitation) in electrical activity
ip acetic acid
- Increased c-fos mRNA expression
Esophageal acetic acid
- Increased c-Fos immunoreactivity
- Increased neuronal excitability
- Enhanced the PB-CeA, but not the BLA-CeA transmission
- Increased crf mRNA expression
- Increased c-Fos and Krox-24 imunoreactivities
- Increased crf mRNA expression
- Induced ERK activation in the right CeA
Acid-induced muscle pain
- Increased ERK activation
- Enhanced the PB-CeA transmission
Knee joint arthritis
- Increased spontaneous activity in the right CeA
- Increased neuronal excitability in the right CeA
- Enhanced the PB-CeA and the BLA-CeA transmission
- Increased mGluR1 and mGluR5 expression
- Increased phosphorylation of NR1 subunit
Sciatic nerve ligation or section
- Increased spontaneous and evoked activity differentially in the left and right CeA
- Enhanced the PB-CeA transmission
- Increased crf mRNA expression and CRF immunoreactivity
- Increased glucocorticoid receptor mRNA expression
- Increased cell proliferation
The recruitment of the CeA in visceral and inflammatory pain has also been evidenced using the expression of immediate early genes such as cfos (Table 1). For example, intraperitoneal or esophageal acetic acid injection [24, 25], colorectal distension  or experimental cystitis  induce cFos in the CeA and the BSTL. Interestingly, this cellular recruitment in the CeA is not observed after intradermal formalin injection in the hindpaw. In this classical inflammatory pain model, cFos recruitment rather appears in the basolateral nucleus of the amygdala . However, extracellular signal-regulated kinase (ERK) phosphorylation is observed in the CeA after intraplantar formalin  or in acid-induced muscle pain . Last, in visceral pain models but also in more persistent pain models, such as knee intraarticular injection of kaolin-carrageenan, or neuropathic models based on sciatic nerve ligation, major changes can be observed in the CeA (Table 1), such as an increase in spontaneous and evoked activity [33, 34, 37], an enhancement of synaptic transmission at the PB-CeA and the BLA-CeA synapses [35, 38], an increase in the phosphorylation of NR1 glutamatergic subunit , an increased expression of CRF, of group I metabotropic glutamatergic receptors (mGluRs) and of corticosterone receptor [26, 29, 35, 39, 40] and an increased cell proliferation .
Human brain imaging confirms a role of the amygdala in pain processes . However, the spatial resolution only allows a crude discrimination of the various nuclei of the amygdaloid complex. Globally, while older studies failed to detect pain-related changes in the amygdala, more recent studies revealed unilateral and bilateral activation or inhibition. These changes were observed after nociceptive thermal laser stimulation [43, 44] or colorectal distention . The amygdala response is usually correlated to the intensity of the nociceptive stimulus, however the importance of attention and emotional factors has also been evidenced. Indeed, repeated thermal nociceptive stimulations of increased intensity lead to an activation of the amygdala that is correlated to the pain perception . However, when sub-threshold stimulus is interpreted as painful by the subject, this situation may activate the amygdala similar to a nociceptive painful stimulus reflecting a state of anxiety or the anticipation of a potentially aversive event. In another study , the subjects were informed that they were about to be expose to a cold painful stimulus lasting 1 minute, or 2 minutes. Functional imaging during the first minute of stimulation revealed that the anticipation of longer stimulus duration led to the deactivation of the amygdala. This suggests that the impact of nociceptive information on the amygdala depends on the context, and not simply of the intrinsic properties of the stimulus.
The CeA influences the nociceptive centers
The CeA receives and integrates nociceptive information, but it also influences the main pain centers. The CeA massively projects to other components of the central extended amygdala, to the lateral hypothalamus and to the brainstem . Anatomical data indicate that the intrinsic and the extrinsic projections to the central extended amygdala arise from distinct neuronal populations . Indeed, studies of unitary axonal reconstruction  show that medium size spiny neurons of the CeLC and the CeL mostly project to the BSTL and the dorsal substantia innominata, while axons of CeM projection neurons innervate the various brainstem structure through collateral branches. Since the nociceptive inputs from the PB preferentially target the CeLC and the CeL, such organization suggest that nociceptive information are integrated through an interneuron network intrinsic to the CeA and to the central extended amygdala before being delivered to output neurons influencing effector centers of the brainstem. Some of these centers are part of the nociceptive descending controls. Partially opioidergic CeA outputs are for example innervating the PAG that is a key element of descending controls of nociception through its influence on the ventromedial reticular formation . The PB is also part of descending controls and it receives dense inputs from the CeA . Beside these projections toward integrative centers, the CeA also directly projects to the ventromedial reticular formation , the dorsal reticular nucleus  and monoaminergic centers such as the substantia nigra and the rostral ventral tegmental area (dopamine), the locus coeruleus (noradrenaline) and the raphe nuclei (serotonin) [48, 50–52]. These structures are known to influence the nociceptive message. Moreover, monoaminergic centers projecting to the forebrain modulate striatal and cortical activities, potentially influencing affective, emotional and motivational aspects of pain. Finally, within the central extended amygdala, the dorsal substantia innominata that receives strong inputs from the CeA – and particularly from the CeLC – includes cholinergic neurons that innervate the prefrontal and insular cortices [8, 20].
Thus, the CeA has connections allowing the modulation of both the sensory and the affective, emotional and cognitive aspects of pain.
Amygdala, acute pain and analgesia
The potential role of the amygdala in the modulation of pain has been suggested for a long time. It has been clearly shown that the anti- and pro-nociceptive effects are dependent on (1) the type of pain (acute, inflammatory or chronic); (2) the measured parameters (threshold or latency of reflex withdrawal, vocalizations, emotional component); and (3) the emotional state of the subjects (stress, anxiety, fear and expectation).
Effects of CeA manipulation on pain-related outcomes in different pain models
Pain related outcomeb
1. CeA lesion
- Reduced morphine-induced, stress-induced and conditioned hypoalgesia
- Reduced morphine-induced and conditioned hypoalgesia
- Decreased pain-induced CPA
- Decreased pain-induced CPA
2. Injection of muscimol
- Reduced mechanical hyperalgesia
- Decreased escape/avoidance
3. Injection of NMDA antagonist
- Decreased pain-induced CPA
4. Injection of group I mGluRs ligands
- Agonist induced visceral and mechanical hypersensitivity
- Antagonist reduced visceral sensitivity
- Antagonist reduced mechanical hypersensitivity
- Antagonist reduced mechanical hypersensitivity
- Antagonist decreased vocalizations
- Agonist increased, and antagonist decreased, pain-induced CPA
5. Injection of group III mGluRs agonists
- Decreased mechanical sensitivity (mGluR7)
- Decreased vocalizations and anxiety
- Increased mechanical sensitivity (mGluR8)
- Increased vocalizations and anxiety
6. Injection of cholinergic agonists
- Decreased thermal sensitivity, reduced jaw opening reflex
- Decreased vocalizations
7. Injection of noradrenergic α 2 ligands
- Agonist induced mechanical and thermal hypoalgesia
- Antagonist reduced stress-induced thermal hypoalgesia
- Agonist decreased pain-induced CPA
8. Injection of noradrenergic β antagonists
- Decreased pain-induced CPA
9. Injection of CGRP receptor ligands
- CGRP decreased mechanical and thermal reflexes
- CGRP increased mechanical reflexes
- CGRP increased vocalizations
- CGRP1 antagonist inhibited the enhanced reflex to mechanical stimulus
- CGRP1 antagonist decreased vocalizations
10. Injection of CRF receptor ligands
- CRF decreased mechanical and thermal sensitivity
- CRF increases mechanical sensitivity
- CRF increased vocalizations
- CRF1 antagonist reduced mechanical hypersensitivity
- CRF1 antagonist decreased vocalizations and anxiety
11. Injection of oxytocin, vasopressin, neurotensin, galanin
- Decreased mechanical and/or thermal sensitivity
12. Injection of opioid receptors ligands
- Morphine and β-endorphin induced mechanical and thermal hypoalgesia
- Morphine decreased vocalizations
13. Corticosterone implants
- Sensitized visceromotor reflexes to colorectal and urinary bladder distension and to somatic mechanical sensitivity
- Increased anxiety
14. BDNF gene deletion in the PB-CeA pathway
- Decreased morphine-induced mechanical and thermal hypoalgesia
15. Intracellular effectors
- ERK activator induced mechanical hypersensitivity
- ERK activation inhibitor decreased mechanical hypersensitivity
- ERK activation inhibitor and PKA inhibitor decreased mechanical hypersensitivity
- ERK activation inhibitor and PKA inhibitor decreased vocalizations
These data therefore show that even if the CeA is not directly involved in the modulation of basal nociceptive thresholds in normal situations, it strongly influences analgesia processes (Table 2).
Stressful situations (restraint or noise) and fear, especially the experimental model of fear conditioning, cause analgesia as measured by the tail flick or the hot-plate tests. Several studies demonstrated that bilateral lesion or inactivation of the CeA diminishes or abolishes this analgesic effect [55, 69, 70, 80, 92] (Table 2). In parallel, clinical studies showed that stressful stimuli, either painful (electric shocks) or painless (unexpected noises), and innocuous stimuli, previously associated with electric shocks (fear conditioned), cause hypoalgesia in a finger withdrawal test following thermal stimulation . In contrast, anxiety caused by electric shock induces hyperalgesia . Although the direct role of the amygdala in these effects has not yet been shown in humans, the results in rodents are consistent with clearly defined functions of the amygdala in emotional responses to different aversive situations .
The amygdala thus can participate in adaptive processes leading to alleviation of pain sensation but its role is not limited to antinociception. While CeA lesion can block the antinociception induced in the tail-flick test by electrical shock pre-exposure, the affective hyperalgesia, measured by the latency to vocalize after the shock is also reduced by the same CeA lesion . Interestingly, similar results are observed following lesion of the BSTL . Pharmacological manipulation of the CeA can also increase visceromotor and/or somatomotor reflexes in naïve animals, as shown by intra-CeA injection of agonists to group I mGluRs [76, 77]. Similarly, CGRP and CRF administration in the CeA can induce mechanical hypersensitivity [82, 84], which is in contradiction with older studies reporting antinociceptive effects of these peptides [58, 59] (see below). Finally, corticosterone implants in the CeA sensitize the visceromotor reflexes to colorectal and urinary bladder distension [87, 88] and increases mechanical sensitivity . The amygdala is thus likely to contribute to both analgesia and hyperalgesia.
Amygdala, sustained pain, hyperalgesia and affective component
Persistent pain, such as inflammatory or neuropathic pain, has a different profile than acute pain. While animal studies often don’t address all the symptoms described in humans, the models that are used can display spontaneous nociceptive behaviors, hyperalgesia and/or allodynia, aversion as well as the emotional consequences of pain such as anxiety and depression .
The role of the CeA in sustained pain has been examined after intraplantar injection of formalin (somatic inflammatory pain), intraperitoneal injection of acetic acid (visceral inflammatory pain), intraarticular injection of kaolin and carrageenan (somatic inflammatory pain), and ligation/compression of the sciatic nerve (neuropathic pain) (Table 2). Generally, the manipulation of CeA activity, either by activation or by inactivation, did not modify the spontaneous nociceptive behaviors [30, 71, 72]. However, persistent pain leads to increased neuronal activity and synaptic transmission in the CeA (Table 1) which may be involved in the induction and/or maintenance of hypersensitivity observed in these models [5, 6, 33, 38]. Thus, inhibiting the extracellular signal-regulated kinase (ERK) activation in the CeA decreases mechanical, but not thermal, hyperalgesia in formalin and arthritis models, while the direct activation of ERK in the CeA is sufficient to produce mechanical hyperalgesia in naïve animals [30, 91]. In addition, intra-CeA injections of mGluRs antagonist (groups I and III) or of CGRP1 or CRF1 antagonists alleviate the increased withdrawal reflexes as well as audible and ultrasonic vocalizations in monoarthritic animals [78, 79, 83, 85, 86]. In neuropathic rats, activation of GABA-A receptors in the CeA also diminishes mechanical hypersensitivity . These data suggest that changes in the activity and neurochemistry of the CeA contribute to the exacerbation of nociceptive responses in persistent pain.
The affective and emotional dimensions of pain can also be under the influence of the amygdala. The effects of pharmacological manipulations of the CeA on vocalizations reported above (see also Table 1) are pertinent in this context. Indeed, while audible vocalizations induced by a nociceptive stimulus are believed to reflect nocifensive responses, ultrasonic vocalizations would reflect an affective pain response [78, 95]. A widely used test to assess the affective/emotional component of pain, is the pain-induced conditioned place aversion. This behavior is suppressed in the intraplantar formalin and intraperitoneal acetic acid models by bilateral lesions of the CeA [72, 73] or by injection of β-adrenoceptor antagonist . Similarly, intra-CeA administration of a GABA-A receptor agonist, a NMDA antagonist or a group I mGluR antagonist reduces the place avoidance behavior in neuropathic rats [74, 75].
Finally, the CeA may be a part of the mechanism linking sustained pain to anxiety and depression-like states. It has been shown that the models of monoarthritic pain or of sciatic nerve ligation induce anxiety that is correlated with the activation of the CRF system in the CeA. The expression of this neuropeptide, which has a well established role in anxiety, is increased in the CeA of neuropathic animals [39, 40], while the intra-CeA injection of CRF1 antagonist reduces anxious behaviors produced by knee monoarthritis [85, 86]. In addition, the pronociceptive effects of corticosterone implants in the CeA are accompanied by anxiety-like behaviors .
In humans, brain imaging studies  demonstrated changes in the activity of the amygdala of patients with irritable bowel syndrome , arthritis  or mononeuropathy , suggesting that the amygdala may be involved in the emotional aspects related to these pathologies. Interestingly, a brain imaging study showed the participation of an amygdala - anterior cingulate cortex circuit in the higher subjective perception of pain in healthy subjects experiencing sadness . Conversely, viewing pictures of a romantic partner reduced self-reported pain, in association with activation of the amygdala . These data reinforce the role of amygdala in emotional impact on pain.
In summary, the amygdala, particularly the CeA, may have a mainly antinociceptive influence in acute/phasic pain conditions associated with situations of stress or fear, and a mainly pronociceptive influence in persistent/tonic pain conditions which concerns the sensory as well as the emotional and affective dimensions.
Potential mechanisms with several actors
Some information is available on the cellular and molecular mechanisms by which the amygdala can exert its bidirectional effects on pain parameters. Projections from the PB to the CeA give direct information about the perceived nociceptive stimulation through glutamatergic and peptidergic synapses [100–102], while the glutamatergic projections from the BLA to the CeA can bring polymodal information with an affective valence . This latter input can be turned into an inhibitory control through the intercalated cell masses, functionally placed between the BLA and the CeA [4, 12]. The influence of the CeA on its targets will depend on the confrontation of these inputs with other afferent information. The internal network of the CeA and of the central extended amygdala is based on GABAergic inhibitory neurons , and the output of this system can either be inhibitory or disinhibitory on target structures [4, 104, 105]. Changes in the level of pain perception seem to involve the projection of the CeA to the centers of descending pain control, especially the PAG and the ventromedial reticular formation , while the influence of the amygdala on the emotional and affective dimensions of pain arises from a larger network that involves indirect connections with cortical regions such as the insular and the cingulate cortex [5, 20].
It is now clearly established that persistent pain causes long-lasting changes in the activity of the CeA that could account for its pronociceptive influence on sensory and affective components of pain [5, 6]. In arthritic, visceral and neuropathic pain models, subpopulations of nociceptive neurons of the CeA exhibit increased membrane excitability, leading to higher spontaneous activity, as well as potentiation of synaptic transmission [27, 33, 35, 37, 38]. This synaptic plasticity, observed at both the PB-CeA and the BLA-CeA synapses can potentiate nociceptive transmission. Some specificity can be observed concerning pain-related plasticity in the CeA. Indeed, in arthritis model, the population of multireceptive neurons, responding to innocuous stimuli but preferentially to nociceptive stimuli, presents increased basal activity and increased responses to mechanical but not thermal nociceptive stimuli. A second neuronal population, normally insensitive to somatosensory stimulation, develops responses to mechanical, but not thermal nociceptive stimuli. In contrast, arthritis modifies neither the basal activity nor the responses of the nociceptive-specific neurons . In addition, while the PB and the BLA can synapse onto the same neurons, but on different dendritic compartments , specific plastic changes can occurs with different pain models. While arthritic and neuropathic pain models potentiates both the PB-CeA and the BLA-CeA synapses [33, 38], a model of visceral pain only potentiates the PB-CeA synapses .
An intriguing observation is that the right CeA appears largely more implicated in persistent pain than the left CeA. Indeed, the activation of ERK after intraplantar formalin is restricted to the right CeA, and the blockade of ERK activation in the right CeA decreased mechanical hyperalgesia at both hindpaws, irrespective to the side of formalin injection [30, 31]. In the arthritis model, the enhanced background activity and evoked responses are observed only in the right CeA, as well as the decreased activity following injection of a PKA inhibitor . Finally, in neuropathic rats, spontaneous activity and evoked responses increased in the left CeA 2 and 6 days after sciatic nerve ligation, but declined afterward, whereas these electrophysiological parameters increased in the right CeA at 14 days post-ligation . The different implication of left and right CeA in pain processes could partially account for the conflicting reports indicating antinociceptive [58, 59] and pronociceptive [82, 84] effects of CRF and CGRP injected into the CeA of naive rats. The anatomical and functional basis of this lateralization remains unclear, but it suggests a strong relation between right amygdala and affective/emotional component of persistent pain.
Beyond the central amygdala
While we focused this review on the CeA, it is necessary to remind the contribution of BSTL and BLA to pain process.
The CeA belongs to the central extended amygdala; it is largely interconnected with the BSTL and shares similar afferents, including polymodal and nociceptive inputs from the BLA and the PB [8, 21]. While there is no report of an influence of the BSTL on nociceptive sensitivity, a few studies addressed its role in pain-induced conditioned place aversion. Bilateral lesion of the BSTL , as well as intra-BSTL injection of a β-adrenoceptor antagonist [120, 121], α2-adrenoceptor agonist  or CRF1 antagonist  decrease place aversion in the intraplantar formalin and/or intraperitoneal acetic acid models. Moreover, noradrenaline and CRF release in the BSTL is enhanced in these pain models [120–122] and CRF mRNA is upregulated in the BSTL of neuropathic rats . These data suggest that the BSTL can be involved in affective/emotional component of pain, maybe in a complementary manner with the CeA, especially in view of BSTL implication in anxiety .
The BLA, beyond its role as a mere input provider to the CeA and BSTL, also appears as an important actor in pain processes. Some observations suggest that the BLA and the CeA can have a parallel role in nocifensive responses. Stress- and fear-induced analgesia are reduced by intra-BLA infusion of cannabinoid CB1 receptors antagonists, of muscimol or of diazepam [124–127]; mu opioid agonist, including morphine, injected in the BLA decreases vocalizations to tail shocks  and thermal nociceptive sensitivity [129, 130]. Sustained pain also impacts the BLA as intraplantar formalin induces c-fos mRNA , arthritis increases the expression of the pro-nociceptive cytokine tumor necrosis factor α (TNF-α)  and neuropathy increases cell proliferation in the BLA . A functional plasticity of BLA neurons is also observed in arthritis model, characterized by an increased in spontaneous and evoked activity and enhanced synaptic transmission. This pain-induced plasticity can be reversed by CRF1 antagonism .
The affective component of pain can also be modulated by the BLA. The formalin-induced conditioned place aversion is reduced by lesion of the BLA , as it is the case for the CeA and the BSTL, but also by BLA injection of NMDA antagonist or of morphine . In addition, neutralizing BLA TNF-α with Infliximab antibodies reduced anxiety-like behaviors associated to arthritis . Finally, CRF1 antagonism in BLA of arthritic rats decreased mechanical hypersensitivity, but also pain-related vocalizations and anxiety .
These influences of BLA on sensory and affective/emotional aspects of pain can, at least in part, involve its outputs to the CeA and/or BSTL. However, the BLA also appears to be implicated in pain-related cognitive impairment. Indeed, in arthritic rats, blockade of pain-induced plasticity in the BLA with CRF1 antagonism reversed the deficit in decision making evaluated in a gambling task. This deficit has been shown to be dependent upon a deactivation of the medial prefrontal cortex driven by the hyperactivity in the BLA. Importantly, this mechanism appears independent of the CeA . Thus, the connections of the BLA with the cerebral cortex, especially prefrontal, cingulate and insular cortices, could allow the amygdala to influence both the affective/emotional and the cognitive aspects of pain. In addition, the dense BLA input to the ventral striatum [9, 10] could also participates in pain-induced changes in motivation and goal-directed behaviors.
The data from the literature show that the amygdala, and in particular the CeA, contributes to pain processes. Its anatomical and functional relations with pain ascending systems, with pain facilitating or inhibitory descending systems, and with affective and cognitive centers are placing the CeA in a critical position within the pain matrix. Acute pain reactions only require CeA if they occur in specific emotional or adaptive contexts. On the other side, persistent and chronic pains alter CeA activity, which influences the pain experience and the related emotional, affective and motivational states. Reciprocally, these states can modify the amygdala processing of information and its impact on pain.
The amygdala has been mostly considered for its roles in emotion, especially in fear [2, 4]. While most studies focused on the BLA, recent researches on the fear circuit disclosed a part of the CeA microcircuitry involved in fear conditioning [104, 105]. It is very likely that a similar, if not the same, circuit underlies CeA functions in pain. Indeed, a recent study showed for the first time that the “nociceptive” amygdala (CeLC) undergoes synaptic plasticity at the PB-CeA and the BLA-CeA synapses during fear conditioning . This result emphasizes the need to consider the strong relation between pain and emotion. The contribution to the pain matrix of the CeA, of other components of the amygdala and of the central extended amygdala, should foster researches in the context of chronic pain and the associated anxio-depressive disorders.
This work was supported by Centre National de la Recherche Scientifique (UPR3212) and University of Strasbourg.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.