You are on page 1of 8


REVIEW QX-314 (8) completely abolished noxious heat

sensitivity, demonstrating that the fibers are
sufficient to account for heat pain, despite a
Neural circuits for pain: more limited contribution from each channel.
Given that TRPV1, TRPM3, and ANO1 show

Recent advances and current views considerable overlap in their distribution, analy-
ses of double and triple genetic deletion of the
channels in mice may be required to fully un-
Cedric Peirs1,2 and Rebecca P. Seal1,2* derstand at the cellular level how noxious heat
is transduced. Cold is also transduced through
The mammalian nervous system encodes many different forms of pain, from those multiple channels. Mice with a genetic deletion
that arise as a result of short-term low-grade interactions with noxious thermal, chemical, of the menthol-sensitive TRPM8, a critical trans-
or mechanical sources to more serious forms of pain induced by trauma and disease. In ducer of innocuous cooling, showed only par-
this Review, we highlight recent advances in our understanding of the neural circuits that tial avoidance of noxious cold temperatures,
encode these types of pain. Promising therapeutic strategies based on recent advances whereas selective ablation of TRPM8-expressing
are also highlighted. cells completely abolished noxious cold sensi-

tivity (9). The molecules that transduce me-
he ability to sense pain protects us from this Review is to provide an overview of pain chanical nociception remain stubbornly elusive.

Downloaded from on November 3, 2016

harm and is thus an essential aspect of circuitry, with an emphasis on recent impor- Deletion of the leading candidate molecules in
our well-being. Patients suffering from tant advances in our understanding of path- mice unexpectedly produced little to no change
channelopathies that eliminate the ability ological pain. Insights into neural circuits for in noxious mechanical pain sensitivity (10, 11).
to feel pain have very high rates of early pain that have therapeutic potential are also
mortality, largely due to self-mutilation and re- discussed. Transduction in the periphery:
petitive fractures (1). Pain, of course, is also a More than just neurons
source of substantial discomfort, and humans How is pain detected in the periphery? Epithelial cells such as keratinocytes and Merkel
have long sought ways to ameliorate pain, as Pain is produced through the activation of no- cells directly interact with peripheral axon ter-
exemplified by the Ebers Papyrus from ancient ciceptive primary sensory neurons categorized minals and have been implicated in the mod-
Egypt. As a concept, pain has evolved from ar- into the classes C, Ad, and to a lesser extent Ab, ulation of sensory transduction. Recently, several
chaic notions of demonic punishment to more depending on their axon caliber, degree of mye- elegant studies have demonstrated an active
contemporary views of biological circuit–based lination, and conductivity properties. Nociceptors role for epithelial cells in tuning the response of
origins for the sensation (2). The most recent innervate the skin, deep tissues, and internal sensory neurons. Merkel cells are the epider-
definition of pain proposed by the International organs and are tuned to detect a wide variety of mal end organ of slowly adapting type 1 (SA1)
Association for the Study of Pain, unchanged noxious mechanical, thermal, and chemical stim- mechanoreceptors and are involved in the
since its first publication in 1979, is “an unplea- uli through the activation of modality-specific detection of features such as edges and textures.
sant sensory and emotional experience associ- sensory transduction molecules. A number of Depolarization of the Merkel cells by optoge-
ated with actual or potential tissue damage, or netic stimulation or by touch, which depends
described in terms of such damage.” The full on the innocuous mechanotransduction chan-
biological complexity of pain and its underlying nel Piezo2, directly excites SA1 fibers through
circuitry is not wholly conveyed by this defini-
tion, however. Pain stems from a varied array of
“Pain stems from a varied still-unidentified neurotransmitter signaling
mechanisms (12). Epithelial cells have also been
peripheral sensors that detect nociceptive stim- array of peripheral sensors shown to actively contribute to nociceptive trans-
uli within internal tissues and those produced that detect nociceptive mission. Activation of channelrhodopsin or
by the external world. The information is sub- TRPV1 channels ectopically expressed by kera-
sequently distributed to a series of complex
stimuli within internal tinocytes in mice induced action potential firing
neural circuits in the spinal cord dorsal horn tissues and those produced in sensory neurons, neuronal activity in the
and then to numerous brain regions, producing by the external world.” spinal cord, and nocifensive behaviors (13, 14).
a diverse set of emotions, actions, and sensa- Conversely, inhibition of the cells blocked the re-
tions (Fig. 1). Pain also manifests with different sponse of nociceptors and other primary sensory
qualities, including stabbing, pricking, burning, these key molecules have been identified (3). neurons to cutaneous stimulation (14). Studies
or aching, further highlighting the heterogene- Most if not all modalities are conveyed through that identify the salient signaling molecules or
ity of the underlying neural circuitry. Longer- the activation of more than one transducer, as establish the extent to which epithelial cells con-
lasting pain states produced by nerve and tissue exemplified by heat transduction. Genetic de- tribute behaviorally to pain will be of interest.
damage, which are also heterogeneous in na- letion of the heat-sensing transient receptor
ture, provide ongoing awareness of the injured potential channel vanilloid 1 (TRPV1) in mice Is pain a labeled line?
area. Pain can become chronic and debilitating impaired the response to heat only at high nox- Most nociceptors express various combinations
when the tissue damage persists and, in some ious temperatures (4), pointing to the existence of sensory transducers that are specific for dif-
conditions, even after the wound has healed. of one or more additional heat sensors. Recent ferent modalities (often heat and mechanical)
Because of the high prevalence and lack of ade- efforts have now identified TRPM3 (5) and the and therefore are classified as “polymodal.”
quate treatment options, unraveling the bi- calcium-activated chloride channel ANO1 (6) as Large-scale efforts using single-cell RNA tran-
ological basis of persistent pain continues to important contributors to the sensation of heat. scriptome analyses have sought to better un-
be an area of intense study. The objective of As was observed with TRPV1, deletion of either derstand the molecular and functional logic of
channel alone strongly reduced, but did not primary sensory neurons by parsing the tran-
Departments of Neurobiology and Otolaryngology, University completely eliminate, noxious heat sensitivity. scriptome profiles of hundreds of individual
of Pittsburgh School of Medicine, 3501 Fifth Avenue, BST3, In contrast, ablation of TRPV1+ fibers by in- cells into groups defined by convergent gene ex-
Pittsburgh, PA 15213, USA. 2Pittsburgh Center for Pain
Research, University of Pittsburgh School of Medicine, 200
trathecal injection of the TRPV1 ligand capsai- pression patterns (15–17). These cluster analyses
Lothrop Street, Pittsburgh, PA 15213, USA. cin (7) or silencing the cells by selective uptake offer a richer picture of the diversity of noci-
*Corresponding author: Email: of the voltage-gated sodium channel blocker ceptors than previous molecular marker–based

578 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 SCIENCE

classifications, as well as a deeper understand- However, limits on the ability to predict the TRPV1, TRPM3, and to a lesser extent ANO1,
ing of the molecular constituents that give rise functional output of nociceptors on the basis of predictably produced a profound loss of heat but
to the functional properties of sensory neurons their molecular and biophysical properties were did not alter innocuous or noxious mechanical
(Fig. 2). Although the precise relationships also recently demonstrated in studies in which se- sensation (21). Loss of CGRP-expressing sensory
between the molecular constituents and the lect polymodal populations were ablated in mice. neurons, which are not responsive to cooling
response properties of the neurons will require Ablation of nociceptors that express the Mas- agents (22), greatly increased tonic and evoked
further study, the transcriptome data do show related G protein–coupled receptor subtype D activity in spinal neurons associated with cold.
some predictive capacity: For example, the ty- (MRGPRD) markedly reduced acute and persist- This latter finding, mechanistically resembling
rosine hydroxylase (TH+) cluster expresses high ent mechanical pain but had no effect on thermal the activation of dorsal horn itch circuits after
levels of Piezo2, but not the heat sensors TRPV1, sensitivity (7). Similarly, ablation of the calcitonin silencing pain afferents (pain normally inhibits
TRPM3, or ANO1, consistent with the response gene–related peptide (CGRP) neurons, a popula- itch at the level of the spinal cord), further
of the cells to light touch, but not to heat (18–20). tion that partially overlaps with the heat sensors highlights the complexity of primary afferent

Downloaded from on November 3, 2016

Fig. 1. Overall organization of somatosensory circuits. Cutaneous sen- represented): neurons transiently expressing VGLUT3 (tVGLUT3, orange) in
sory neurons (DRG) are activated by a variety of stimuli (bottom left) and laminae III and IV, PKCg (green) in lamina IIiv, calretinin (blue) in outer lamina II

project to the spinal cord dorsal horn (DH, middle left). In the DH (right), (IIo) and lamina IIid, vertical cells (yellow) in lamina IIo, and projection
the central terminals of high-threshold nociceptors (HT) are located in neurons (red) in laminae I, IV, and V. Projection neurons send information
the most superficial laminae [lamina I to the dorsal part of inner lamina II to the brainstem and thalamus and then on to several brain regions im-
(IIid)] and lamina V. Low-threshold mechanoreceptors (LTMR) preferen- plicated in sensory-discriminative (upper left, light brown) and emotional
tially end in the deep dorsal horn [ventral part of inner lamina II (IIiv) to (upper left, dark brown) sensory perception. ACC, anterior cingular cortex;
lamina V]. The spinal cord is divided into 10 laminae (the DH is I to VI) and is SI (II), primary (secondary) somatosensory cortex; PAG, periaqueductal gray
composed of numerous neuronal populations. Some identified popula- area; PB, parabrachial nucleus; AMY, amygdala; PFC, prefrontal cortex; BG,
tions are organized in longitudinal layers (only excitatory neurons are basal ganglia.

SCIENCE 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 579


coding, as well as an important role for central Recent studies of dorsal horn nociceptive circuits nocuous tactile sensitivity is primarily dependent
processing in shaping the sensory percept. have focused largely on persistent pain because on the cation channel Piezo2 (28). In primary
of its high prevalence and the need for better sensory neurons, the channel is almost exclu-
Integration in the spinal cord treatment options (23, 24). sively expressed by low-threshold mechanore-
Primary sensory neurons innervate the six Rexed ceptors. Conditional deletion of Piezo2 in all
laminae of the spinal cord dorsal horn, the major Mechanical allodynia: sensory neurons in mice resulted in markedly
locus for the integration of peripheral sensory Peripheral contributions reduced mechanical activation of low-threshold
input and descending supraspinal modulation. Nerve and tissue damage produces dramatic mechanoreceptors, as well as reduced behav-
The central terminals of sensory neurons are changes in peripheral and central somatosen- ioral touch sensation, with no change in acute
somatotopically organized, forming ventrodorsal- sory circuits. Although acute pain typically mechanical pain (29). Because low-threshold
oriented columns similar to the arrangement of follows the activation of nociceptors, light touch– mechanoreceptors are thought to convey some
the primary somatosensory cortex. Most C- and activated neurons can be recruited into the forms of mechanical allodynia, Piezo2 could have
Ad-nociceptive afferents form synaptic contacts nociceptive network to cause pain after injury. a role in this type of pain—although after an
in the superficial laminae (I and II), whereas This pathological condition, termed mechan- inflammatory insult, mice with a conditional
low-threshold Ad- and Ab-afferents generally ical allodynia, occurs in numerous peripheral deletion of Piezo2 in primary sensory neurons
project to deeper laminae (III to V) (Fig. 1). neuropathies and central pain disorders, pre- did not show evidence of a defect in their me-
Accordingly, in vivo electrophysiological record- senting in up to 50% of patients with neuro- chanical pain threshold under the conditions
ings in mammals have shown that laminae I pathic pain (25). Still unknown is the identity tested. In contrast, a separate study, which dem-

Downloaded from on November 3, 2016

and II respond mainly to noxious peripheral of the light touch–sensitive primary sensory onstrated sensitization of Piezo2 by the cyclic
stimulation, whereas neurons in deep layers are population(s) that conveys mechanical allody- adenosine monophosphate sensor EPAC1 after
more sensitive to touch. The dorsal horn is com- nia. A large number of primary afferents re- neuropathic injury, pointed to a potential role
posed of a large number of excitatory (75%) spond to light mechanical touch and are thus for the mechanotranducer in mechanical al-
and inhibitory (25%) interneuron populations, potential candidates (26). lodynia (30).
as well as a smaller number of projection pop- Although the identity of the mechanosensory
ulations located in laminae I, III, IV, and V. The neurons engaged during mechanical allodynia Mechanical allodynia: The silent dorsal
projection neurons relay information to numer- remains unclear, progress has been made in our horn circuit
ous supraspinal sites to give rise to both qual- understanding of innocuous sensory mechano- Changes in the dorsal horn circuitry contribute
itative and affective aspects of the pain sensation. transduction (27). As mentioned previously, in- to the expression of mechanical allodynia. One

A Myelinated sensory neurons Unmyelinated sensory neurons






Proprioception Mechanical Touch Touch Touch Touch Itch Itch Heat Mechanical Cold






Fig. 2. Organization of primary sensory neurons. (A) The categories of and TRPM8) and immunohistological analyses (TRKA, TRPV1, TAC1). Mye-
myelinated and unmyelinated neurons and their respective functional roles linated neurons preferentially express neurofilament heavy chain (NEFH), and
are derived from large-scale transcriptional analyses, behavioral analyses, and unmyelinated neurons preferentially express the sodium channels Nav1.8
the literature. (B) The locations of the central terminations of the primary and -1.9. Laminae are indicated on the left. RET, ret proto-oncogene;
sensory neuron categories are shown. The schematic is based on analyses of CALB1, calbindin 1; TAC1, tachykinin 1; SST, somatostatin; NPPB; natri-
gene reporter mouse lines (PV, RET, TRKB, TH, MRGPRD, MRGPRA3, CGRP, uretic peptide type B.

580 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 SCIENCE

of the principal concepts to explain these changes
is based on the gate control theory in which
touch inhibits pain through a feed-forward in- NK1R projection neuron
hibitory circuit within the superficial layers of
the dorsal horn (31). In the case of mechanical I
allodynia, injury impairs the feed-forward in- Aδ/C peptidergic
hibitory circuit, allowing light touch to engage afferents
the nociceptive network. A seminal study, per-
formed in rodent spinal cord slices, showed
that low-threshold mechanosensory A-fiber in- Dynorphin
put is able to activate lamina I pain projection GABA/Gly
neurons through a polysynaptic network when central transient interneuron
inhibitory receptor antagonists are present IIo cell Glutamatergic
(mimicking the injury-induced decrease in in- vertical cell
hibition) (32). Importantly, the study established
that the mechanical allodynia circuit, which is islet cell
normally silent, is already in place under phys-
iological conditions.

Downloaded from on November 3, 2016

Dorsal horn mechanical allodynia
circuits: Excitatory neurons
The dorsal horn circuit for mechanical allodynia
was originally modeled as a dorsally directed IIid Calretinin
pathway that begins with excitatory interneu-
C non-peptidergic
rons at the border between laminae II and III afferents
that express the g isoform of protein kinase C PKCγ interneuron
(PKCg) and ends with lamina I pain projection
neurons (33) (Fig. 3). Two intermediary pop-
ulations in this pathway were recently identi- IIiv Glutamatergic
fied through paired recordings, the transient interneuron
central cells in inner lamina II and vertical cells
in outer lamina II (34). The dendrites of vertical
cells extend deep into lamina III and receive
input from most classes of primary afferents,
suggesting a potential role as integrators of PV GABA/Gly
the network (35). An expansion of the circuit interneuron
into deeper laminae was shown by studies of the III
vesicular glutamate transporter 3 (VGLUT3). A
population of excitatory interneurons was iden- tVGLUT3
tified in lamina III that transiently expresses interneuron
VGLUT3 and receives almost exclusively low- Myelinated
threshold input. Selective activation of the cells Aβ/Aδ
by using a designer receptor strategy (DREADD)
further demonstrated their important role in the
transmission of touch as painful (36). Calretinin-
expressing excitatory interneurons in inner
lamina II were also identified as an essential
element of the circuit. A functional study of the
calretinin and PKCg populations that used c-Fos Fig. 3. Organization of the dorsal horn circuit for pain. Peripheral nociceptors (blue) project
analysis revealed the existence of distinct ex- onto excitatory interneurons in lamina IIo (central cell, dark gray; vertical cells, yellow) and onto
citatory microcircuits for mechanical allodynia, neurokinin 1 receptor (NK1R) projection neurons (red) in lamina I. Nonpeptidergic afferents ex-
which are differentially engaged depending on pressing MRGPRD (green) project to lamina IIid, including to excitatory vertical cells with ventrally
whether the injury is inflammatory or neuro- directed elongated dendrites. Both primary afferents also contact inhibitory islet cells (horizontally
pathic in nature (36). elongated, pink). Stimulation of nociceptive afferents activates excitatory central cells, vertical
cells, and NK1R projection neurons to mediate noxious pain. Inhibitory islet cells modulate this
Mechanical allodynia circuits: activity. Innocuous afferents (orange) project onto excitatory interneurons expressing tVGLUT3 in
The “gates” lamina III (brown), PKCg in lamina IIiv (teal), and vertical cells in lamina IIo. Myelinated afferents also
Inhibitory neurons are a fundamental element contact PV inhibitory interneurons in lamina III (radial, pink) and dynorphin inhibitory vertical cells

of the mechanical allodynia network in the in lamina IIo (vertical, pink). tVGLUT3 interneurons project onto excitatory vertical cell dendrites
spinal cord, forming “gates” that prevent the re- and intermediate excitatory interneurons in lamina III. Intermediate interneurons project onto PKCg
cruitment of low-threshold fibers into the no- and calretinin excitatory interneurons in lamina IIiv. Inhibitory interneurons prevent A-fiber–mediated
ciceptive network under physiological conditions activation of the nociceptive network through feed-forward circuits that act on PKCg interneurons,
(Fig. 3). Using paired recordings in spinal cord vertical cells, and NK1R projection neurons. After nerve injury, inhibition by PV and dynorphin in-
slices, a recent study showed that the PKCg terneurons is reduced, allowing A-fiber–mediated activation of a dorsally directed circuit that in-
interneurons, which receive Ab-fiber input, are cludes tVGLUT3 and PKCg neurons. After inflammatory injury, reduction in a still-unknown mechanism
normally under feed-forward glycinergic disy- of disinhibition allows A-fiber–mediated activation of a dorsally directed circuit that includes tVGLUT3
naptic inhibition (34). After neuropathic injury, and lamina IIiv calretinin neurons.

SCIENCE 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 581


which is known to impair inhibitory transmis- initial activation of the microglia, as well as for throughout the pain matrix, which would re-
sion in the dorsal horn, Ab-fibers were able to cell proliferation and self-renewal. Deletion of quire analysis by other methods and by exper-
drive action potentials in PKCg neurons and CSF1 specifically from primary sensory neurons imental designs that allow for causal inference.
presumably the rest of the nociceptive network. or deletion of its downstream effector, the mi- Machine-learning algorithms to assess fMRI
Consistent with this finding, selective ablation croglial membrane adaptor protein DAP12, patterns of activity across the pain matrix are
of lamina III parvalbumin (PV)–expressing in- was sufficient to markedly reduce microglial showing a high predictive capacity for distinct
terneurons, which form contacts with the PKCg activation, as well as the behavioral expres- forms of physical pain, as well as for social pain
neurons, induced mechanical allodynia, indi- sion of mechanical allodynia induced by nerve such as vicarious pain (49).
cating a role for PV interneurons in gating the injury (46). Complementing these recent findings from
allodynic pathway (37). Conversely, selective brain-imaging studies in humans, work in ro-
activation of the PV interneurons by the ex- Pain processing in the brain dents has focused on synaptic-level mechanisms
citatory DREADD strategy reversed mechanical Functional magnetic resonance imaging (fMRI) of pain processing and plasticity in particular
allodynia induced by nerve injury. In addition studies in humans have demonstrated the co- regions of the pain matrix. These studies pro-
to PKCg neurons, PV interneurons were also ordinated activation of several brain areas in vide further support for the idea that individ-
proposed to act directly on the Ab-afferents in- response to noxious somatic and visceral stim- ual brain areas have multiple functional roles.
nervating these neurons, suggesting alternative uli, including the thalamus, anterior cingulate Emotional aspects of pain such as anxiety and
mechanisms for inhibitory control of the circuit cortex (ACC), insular cortex, primary and sec- fear have been linked to the ACC. A study in
(38). The involvement of a second population ondary sensory cortices, prefrontal cortex, basal mice identified a form of presynaptic long-

Downloaded from on November 3, 2016

of inhibitory interneurons in the feed-forward ganglia, cerebellum, and amygdala. This net- term potentiation (pre-LTP) at thalamocorti-
gating of mechanical allodynia was also re- work of brain regions involved in both sensory- cal synapses in the ACC that was induced by
cently demonstrated. Ablation of dynorphin- discriminative and emotional-affective aspects either anxiety-evoking experiences or chronic
expressing interneurons in outer lamina II and of pain is termed the “pain matrix” (47). pain models (50). It was concluded that the
at the border of laminae II and III induced Several recent studies have questioned whether pre-LTP represents anxiety, which can be trig-
mechanical allodynia. The dynorphin gate was activation of the pain matrix specifically repre- gered by chronic pain. The coexistence of this
further demonstrated to control somatostatin- sents pain or is a more generalized system to form of pre-LTP with a previously identified
expressing interneurons, a major population of detect salient events and the state of the body. form of postsynaptic LTP required for the be-
excitatory neurons that partially overlaps with The posterior insular cortex encodes nociceptive havioral expression of chronic pain (51) suggests
PKCg neurons and potentially also vertical and an important role for these ACC synapses in the
transient central cells (39). Studies of the me- mutually enhancing interactions of anxiety and
chanical allodynia network are beginning to not chronic pain.
only reveal individual elements but also produce Mechanisms of plasticity affecting motiva-
conceptual advances, including the understand- “...pain-related information tion in models of persistent pain were also
ing that there are distinct excitatory micro- is likely encoded by specific investigated recently. The nucleus accumbens
circuits and multiple distinct gates related to has been implicated in subjective pain pro-
specific types of injuries.
subregional patterns of cessing in humans and in motivation in ro-
activity throughout the dents. The induction of persistent pain by
Mechanisms of disinhibition underlying
mechanical allodynia pain matrix...” either nerve injury or an inflammatory medi-
ator impaired motivational behavior in mice,
The central sensitization responsible for me- as measured using a progressive ratio oper-
chanical allodynia produces various forms of intensity, and lesions in this brain area disrupt ant task (52). The decrease in motivation was
disinhibition in the dorsal horn (40). One mech- pain, whereas direct stimulation of the area in- linked to enhanced release of the neuropep-
anism involves the injury-induced release of duces pain. However, the insular cortex also has tide galanin in the nucleus accumbens core,
brain-derived neurotrophic factor (BDNF), which a role in affective, cognitive, and homeostatic which induced a form of long-term depression
triggers down-regulation of the potassium chlo- functions for nonnociceptive sensory modalities. (LTD) in dopamine receptor 2–expressing me-
ride exchanger 2 (KCC2) in tropomyosin receptor Intracerebral recordings of the anterior and dium spiny neurons (D2-MSNs), thereby reduc-
kinase B (TRKB)–expressing pain-transmitting posterior insular cortex showed activity levels ing their activity. The injury-induced effects on
neurons, reducing the driving force for chloride that responded similarly to different sensory D2-MSNs and motivated behavior were reversed
and thus the strength of inhibitory transmis- modalities (visual, auditory, and somatosensory), by knockdown of the galanin receptor GALR1
sion (41, 42). The mechanism of injury-induced as well as to noxious and innocuous somato- and by interfering with intracellular signaling
BDNF release was recently shown to differ be- sensory stimuli, consistent with the idea that pathways required for the LTD. Synaptic-level
tween males and females. Males require the the insula is not specific for pain-related infor- investigations of pain-related mechanisms are
activation of microglia, whereas females require mation (2). Additionally, patients with loss-of- thus providing important insights into the role
adaptive immune cells, possibly T lymphocytes function mutations in the sodium channel Nav1.7 of individual brain areas in the pain experience,
(43). Sex differences are frequently observed in (SCN9) show congenital insensitivity to pain due including interactions with other brain func-
animal models of chronic pain and may provide to a loss of noxious peripheral sensory drive (1). tions. Such studies are also revealing potential
insight into the many clinically observed sex In a recent study, the response of patients and new therapeutic avenues.
differences in pain in humans (44). controls to normally noxious peripheral stimuli
Investigations into the mechanisms under- was assessed by fMRI. Both groups reported Blocking pain circuits at the periphery
lying BDNF release in the dorsal horn of male experiencing similar levels of sensation; al- Nerve growth factor (NGF) and its major recep-
rodents after injury have focused on adenosine though controls reported the stimuli as painful tor TRKA have essential roles in nociceptor func-
triphosphate signaling through purinergic P2X4 and patients reported no pain, the two groups tion both during development and in adults.
receptors, which requires microglia activation, nevertheless showed similar patterns and levels Specific mutations in the genes encoding NGF
and on the involvement of chemokine signaling of activity within the pain matrix (48). Gross or TRKA in humans produce phenotypes that
cascades that promote and maintain microglia activity measured in supraspinal structures thus include insensitivity to pain, largely resulting
activation (45). De novo production and release may not reliably predict the pain experience. from defects in the development of nocicep-
of colony-stimulating factor 1 (CSF1) by injured Rather, pain-related information is likely en- tors (53). In adults, NGF signaling through
primary sensory neurons is required for the coded by specific subregional patterns of activity TRKA receptors at the periphery contributes

582 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 SCIENCE

considerably to the pain associated with many Another innovative approach to inhibit no- peripheral nerve injury (66). Humans with a
types of injuries and diseases’ including bone ciceptors targets the voltage-gated sodium chan- mutation in the HSN2-containing splice vari-
cancer, lower back injury, diabetic neuropathy, nel blocker QX-314 to select populations, thereby ant of the kinase WNK1 suffer from congenital
and osteoarthritis. As a new class of analgesic blocking nerve conduction. Coapplication of cap- insensitivity to pain. WNK1/HSN2 plays a role
treatment, humanized antibodies raised against saicin, which permits entry of QX-314 through in nerve injury–induced mechanical allodynia
NGF, such as tanezumab (Pfizer), now in phase the cation channel TRPV1, to the mouse plantar by reducing KCC2 activity in lamina II neu-
III clinical trials for osteoarthritis pain, are showing hind paw markedly reduced the response to heat rons of the dorsal horn (67). Antagonizing WNK
promise. Inhibitors of the TRKA receptor are also and mechanical pressure (8). QX-314 has also activity in rodents restored GABA (g-aminobutyric
of therapeutic interest. been used to ameliorate persistent pain. The acid)–mediated inhibition of spinal cord lam-
Recent studies have focused also on the ther- Toll-like receptor TLR5 was identified as a re- ina II neurons and reversed the behavioral
apeutic potential of targeting voltage-gated so- ceptor specific to A-fibers (63). Coapplication evidence of mechanical allodynia induced by
dium channels in sensory neurons to relieve of the TLR5 ligand flagellin and QX-314 to peripheral nerve injury, but not by inflamma-
pain. In addition to the Nav1.7 loss-of-function the mouse plantar hind paw blocked A-fiber tory injury.
mutations that cause congenital insensitivity activity, as measured electrophysiologically in Using an entirely different strategy, based on
to pain in humans, studies of Nav1.7 and Nav1.8 vitro and in vivo. Although the precise mecha- the observation that GABAergic precursors from
in mice indicate a required role for the chan- nism by which QX-314 enters the cells is still the medial ganglion eminence grafted into the
nel in many forms of acute and pathological unclear, the approach successfully reversed me- forebrain correct neurological disorders of hy-
pain (54, 55). A highly specific Nav1.7 inhibitor chanical allodynia and ongoing pain in a num- perexcitability, such as epilepsy, Bráz et al. tested

Downloaded from on November 3, 2016

isolated from centipede venom, m-SLPTX-Ssm6a, ber of chronic neuropathic pain models in mice whether injecting the telencephalic GABAergic
when injected intraperitoneally in mice, showed without producing obvious deficits in touch precursors directly into the dorsal horn of
potent analgesic effects on chemical-induced sensation. nerve-injured mice could restore inhibitory tone
pain with an efficacy equivalent to or exceed- Lastly, the use of gene therapy to restore the and reverse mechanical allodynia (68). Although
ing that of morphine (56). Similarly, monoclo- normal expression of dysregulated proteins in the precursor cells maintained their cortical
nal antibodies against Nav1.7, when injected sensory neurons has become another attract- identity, they integrated successfully into the
into mice, effectively reduced mechanical al- ive strategy to treat pain. The expression of dorsal horn, forming local synaptic connections
lodynia and thermal hyperalgesia induced by the microRNA miR-7a is dramatically reduced with primary sensory and spinal cord neurons.
inflammation and neuropathic injury (57). Given after neuropathic injuries. Increasing its ex- By 3 weeks after transplantation, mechanical
the promising therapeutic benefit of targeting thresholds were elevated to pre–nerve injury
Nav1.7 channels to relieve pain, several selective levels. In contrast, inflammation-induced hyper-
inhibitors are now in phase II clinical trials. sensitivity was not altered by cell transplanta-
Animal venoms that elicit pain have been
particularly useful for identifying and charac-
“...gene therapy to restore tion. This work demonstrates that strategies to
reverse or compensate for injury-induced dis-
terizing therapeutically relevant nociceptive the normal expression inhibition in the dorsal horn show promise for
sensory molecules. MitTx, from the venom of
the Texas coral snake Micrurus tener tener, is
of dysregulated proteins the treatment of pain.

a highly specific and strong potentiator of the in sensory neurons has Conclusions and the road ahead
acid-sensing channel ASIC1. Injection of this become another attractive Recent advances at the molecular, cellular, and
peptide into the hind paw of mice induced pro- systems level are transforming our understand-
nounced nocifensive behaviors (58). Conversely,
strategy to treat pain.” ing of how the complex sensation of pain is
central and peripheral delivery of the isopep- encoded in the nervous system and providing
tides mambalgin 1 and 2, from the black mam- pression by intraganglionic injection of AAV6- exciting new avenues for therapeutic inter-
ba snake, reversed mechanical allodynia and miR-7a abolished mechanical allodynia and vention. Several key aspects of the nociceptive
heat hyperalgesia induced by inflammatory thermal hyperalgesia induced by nerve injury, network and the sensation of pain, however,
agents by inhibiting ASIC channels, highlight- although not by inflammation (64). In contrast, remain to be addressed. With respect to phys-
ing the therapeutic potential of this family of nerve injury substantially up-regulates the so- iological sensory processing, major outstand-
ion channels (59). dium channel Nav1.3. Intraganglionic injec- ing questions include: What is the identity of
Recent advances in the in vivo application tion of AAV5 expressing a small hairpin RNA the noxious mechanotransducer(s)? What are
of optogenetic methods also provide attractive against Nav1.3 reduced its expression and par- the specific contributions of molecularly dis-
therapeutic opportunities for pain. One of the tially reversed nerve injury–induced mechanical tinct primary sensory neurons to modality
first studies to report the use of optogenetics allodynia (65). coding? What is the logic of nociceptor coding
to evaluate pain circuits in vivo targeted chan- of pain? How do spinal cord circuits contrib-
nelrhodopsin to Nav1.8-expressing primary sen- Blocking pain circuits centrally ute to modality coding? What is the logic of
sory neurons in mice. Cutaneous light-mediated Because mechanisms of disinhibition in the the projection neuron populations? Which neu-
activation of the neurons elicited robust no- dorsal horn have a critical role in the expression ronal populations in the brain process the dif-
cifensive behaviors and conditioned place aver- of mechanical allodynia and spontaneous pain, ferent aspects of pain? In terms of pathological
sion (60). In testing of opsins for therapeutic therapeutic efforts have been directed toward pain, major outstanding questions include: What
purposes, the inhibitory opsin NpHR was tar- the restoration of inhibitory tone. Administra- are the neural networks for spontaneous pain?
geted to small-diameter primary sensory neu- tion of a small-molecule enhancer of KCC2 What are the peripheral and central components
rons by injection of AAV6-NpHR into the sciatic activity (CLP290) to rodents restored inhibitory of the neural networks that underlie mechanical
nerve. Cutaneous light-mediated inhibition of drive in dorsal horn neurons, as measured elec- allodynia induced by different types of injuries,
the infected primary sensory neurons alleviated trophysiologically in spinal cord slices, and such as inflammatory and neuropathic? How do
both inflammatory and neuropathic pain (61). reversed behavioral evidence of mechanical the neural circuits for pain differ as a function of
The low penetration of light through the skin allodynia induced by nerve injury. Similarly, chronicity? How does chronic pain manifest in
may be technically limiting; however, implant- overexpression of the cotransporter KCC2 in the brain? Understanding the precise contribu-
able miniaturized optoelectronic systems for the spinal dorsal horn and primary sensory tions of the neurons and molecules that underlie
wireless control of primary sensory and spinal neurons of rodents reversed behavioral evi- each form of pain will greatly facilitate therapeutic
cord neurons are now being developed (62). dence of mechanical allodynia induced by drug development.

SCIENCE 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 583



1. D. L. Bennett, C. G. Woods, Lancet Neurol. 13, 587–599 (2014).
2. R. Sabatowski, D. Schäfer, S. M. Kasper, H. Brunsch,
L. Radbruch, Curr. Pharm. Des. 10, 701–716 (2004).
3. D. Julius, J. R. Carlson, Curr. Opin. Neurobiol. 34, v–vi (2015).
4. M. J. Caterina et al., Science 288, 306–313 (2000).
Deconstructing the sensation of pain:
5. J. Vriens et al., Neuron 70, 482–494 (2011).
6. H. Cho et al., Nat. Neurosci. 15, 1015–1021 (2012).
7. D. J. Cavanaugh et al., Proc. Natl. Acad. Sci. U.S.A. 106,
The influence of cognitive processes
9075–9080 (2009).
8. C. Brenneis et al., J. Neurosci. 33, 315–326 (2013).
9. W. M. Knowlton et al., J. Neurosci. 33, 2837–2848 (2013).
on pain perception
10. Y. Roudaut et al., Channels 6, 234–245 (2012).
11. R. Sharif-Naeini, Prog. Mol. Biol. Transl. Sci. 131, 53–71 (2015). Katja Wiech1,2*
12. S. Maksimovic et al., Nature 509, 617–621 (2014).
13. Z. Pang et al., Pain 156, 656–665 (2015). Phenomena such as placebo analgesia or pain relief through distraction highlight the
14. K. M. Baumbauer et al., eLife 4, e09674 (2015). powerful influence cognitive processes and learning mechanisms have on the way
15. D. Usoskin et al., Nat. Neurosci. 18, 145–153 (2015).
16. I. M. Chiu et al., eLife 3, e04660 (2014).
we perceive pain. Although contemporary models of pain acknowledge that pain is not
17. C. L. Li et al., Cell Res. 26, 83–102 (2016). a direct readout of nociceptive input, the neuronal processes underlying cognitive
18. M. C. Delfini et al., Cell Rep. 5, 378–388 (2013). modulation are not yet fully understood. Modern concepts of perception—which include
19. L. Li et al., Cell 147, 1615–1627 (2011). computational modeling to quantify the influence of cognitive processes—suggest that

Downloaded from on November 3, 2016

20. R. P. Seal et al., Nature 462, 651–655 (2009).
21. E. S. McCoy et al., Neuron 78, 138–151 (2013).
perception is critically determined by expectations and their modification through learning.
22. E. S. McCoy, B. Taylor-Blake, M. J. Zylka, PLOS ONE 7, e36355 Research on pain has just begun to embrace this view. Insights into these processes
(2012). promise to open up new avenues to pain prevention and treatment by harnessing the
23. A. J. Todd, Nat. Rev. Neurosci. 11, 823–836 (2010). power of the mind.
24. J. Braz, C. Solorzano, X. Wang, A. I. Basbaum, Neuron 82,
522–536 (2014).

25. T. S. Jensen, N. B. Finnerup, Lancet Neurol. 13, 924–935 (2014).
26. V. E. Abraira, D. D. Ginty, Neuron 79, 618–639 (2013). atching a captivating film while a den- changes in pain reports indeed reflect a change
27. A. Zimmerman, L. Bai, D. D. Ginty, Science 346, 950–954 (2014). tist is fixing your tooth can help you in pain perception and neural pain processing—
28. B. Coste et al., Science 330, 55–60 (2010). endure the much-dreaded visit with sur- or just a report bias? Answering this question is
29. S. S. Ranade et al., Nature 516, 121–125 (2014).
30. N. Eijkelkamp et al., Nat. Commun. 4, 1682 (2013).
prisingly little pain. Cognitive processes, notoriously difficult given the highly subjective
31. R. Melzack, P. D. Wall, Science 150, 971–979 (1965). such as distraction, have the potential to nature of pain. Attempts to identify an objective
32. C. Torsney, A. B. MacDermott, J. Neurosci. 26, 1833–1843 (2006). change the way we perceive pain—for better or readout of pain based on brain imaging data
33. L. S. Miraucourt, R. Dallel, D. L. Voisin, PLOS ONE 2, e1116 (2007). for worse, as I will show below. Based on a rich are controversially discussed (2). Brain activation
34. Y. Lu et al., J. Clin. Invest. 123, 4050–4062 (2013).
35. T. Yasaka et al., Mol. Pain 10, 3 (2014).
psychological literature, brain imaging studies in induced by noxious input might in fact not be
36. C. Peirs et al., Neuron 87, 797–812 (2015). humans have sought to describe and character- pain-specific but also reflect processes that are
37. H. Petitjean et al., Cell Rep. 13, 1246–1257 (2015). ize the influence of cognitive factors on the neu- (inherently) linked to pain, such as the detec-
38. D. I. Hughes et al., J. Physiol. 590, 3927–3951 (2012). ral processing and perception of pain since the tion of salient events (3). Historically, the network
39. B. Duan et al., Cell 159, 1417–1432 (2014).
40. H. U. Zeilhofer, H. Wildner, G. E. Yévenes, Physiol. Rev. 92,
1980s. I will first give an overview of our current of brain regions involved in pain processing
193–235 (2012). understanding of mechanisms and neural path- (“pain matrix”) has been divided into sensory-
41. J. A. Coull et al., Nature 438, 1017–1021 (2005). ways to cognitive pain modulation and highlight discriminative and cognitive-affective systems.
42. J. A. M. Coull et al., Nature 424, 938–942 (2003). the most important recent strands of research The sensory-discriminative system, which in-
43. R. E. Sorge et al., Nat. Neurosci. 18, 1081–1083 (2015).
in this field, with an emphasis on experimental cludes the lateral thalamus and primary and
44. J. S. Mogil, Nat. Rev. Neurosci. 13, 859–866 (2012).
45. S. Beggs, T. Trang, M. W. Salter, Nat. Neurosci. 15, 1068–1073 studies in healthy individuals. In the second part, secondary somatosensory cortex (SI and SII, re-
(2012). I will outline how these findings may be inte- spectively), was thought to process nociceptive
46. Z. Guan et al., Nat. Neurosci. 19, 94–101 (2016). grated with modern concepts of perception by input, including its intensity, localization, and
47. P. Schweinhardt, M. C. Bushnell, J. Clin. Invest. 120,
3788–3797 (2010).
using computational models to explore the in- quality. In contrast, the cognitive-affective sys-
48. T. V. Salomons, G. D. Iannetti, M. Liang, J. N. Wood, JAMA fluence of cognition on pain at a more funda- tem, comprising regions such as the anterior
Neurol. 73, 755–756 (2016). mental level. insula and anterior cingulate cortex, was im-
49. A. Krishnan et al., eLife 5, e15166 (2016). plicated in psychological aspects of pain. How-
50. K. Koga et al., Neuron 85, 377–389 (2015). The search for a “signature of pain ever, this strict dichotomy turned out to be an
51. H. Xu et al., J. Neurosci. 28, 7445–7453 (2008).
52. N. Schwartz et al., Science 345, 535–542 (2014). in the brain” oversimplification, as sensory-discriminative brain
53. O. P. Carvalho et al., J. Med. Genet. 48, 131–135 (2011).
A modulatory influence of cognitive factors on regions, for instance, can also be sensitive to
54. M. S. Minett et al., Cell Rep. 6, 301–312 (2014).
55. M. A. Nassar et al., Proc. Natl. Acad. Sci. U.S.A. 101, the perception of pain has been documented cognitive processes. Moreover, studies using a
12706–12711 (2004). for a number of processes including attention, decoding approach (i.e., the prediction of pain
56. S. Yang et al., Proc. Natl. Acad. Sci. U.S.A. 110, 17534–17539 anticipation, catastrophizing, (re-)appraisal, and perception based on activation patterns in the
57. J. H. Lee et al., Cell 157, 1393–1404 (2014). perceived control over pain (1). Undoubtedly, the brain) demonstrated that the prediction is sig-
58. C. J. Bohlen et al., Nature 479, 410–414 (2011). most impressive and most extensively studied nificantly improved when the different brain re-
59. S. Diochot et al., Nature 490, 552–555 (2012). example is a placebo analgesic response. Patients gions are considered together. In the so-far most
60. I. Daou et al., eNeuro 10.1523/ENEURO.0140-15.2016 (2016).
61. S. M. Iyer et al., Nat. Biotechnol. 32, 274–278 (2014). with agonizing levels of pain can report com- rigorous attempt to characterize the “neurolog-
62. S. I. Park et al., Nat. Biotechnol. 33, 1280–1286 (2015). plete pain relief after administration of a sugar ical pain signature” (NPS), Wager and colleagues
63. Z. Z. Xu et al., Nat. Med. 21, 1326–1331 (2015). pill they think is a powerful painkiller. But do such (4) used a machine-learning algorithm to predict
64. A. Sakai et al., Brain 136, 2738–2750 (2013).
65. O. A. Samad et al., Mol. Ther. 21, 49–56 (2013). the perceived intensity of experimentally induced
66. L. Li et al., Cell Rep. 15, 1376–1383 (2016). 1
Oxford Centre for Functional Magnetic Resonance Imaging
heat pain in healthy volunteers. The identified
67. K. T. Kahle et al., Sci. Signal. 9, ra32 (2016). network—comprising brain regions such as thal-
68. J. M. Bráz et al., Neuron 74, 663–675 (2012). of the Brain, Nuffield Department of Clinical Neurosciences,
University of Oxford, John Radcliffe Hospital, Headley Way, amus, SI, SII, anterior insula, and anterior cingulate
ACKN OW LEDG MEN TS Oxford OX3 9DU, UK. 2Nuffield Department of Clinical cortex—afforded a specificity of about 90% in dis-
Neurosciences, Nuffield Division Anaesthetics, University of
Funding was provided by the Rita Allen Foundation and the
Oxford, John Radcliffe Hospital, Headley Way, Oxford OX3
criminating physical pain from related phenome-
American Pain Society to R.P.S.
9DU, UK. na (e.g., “pain” due to social exclusion). Although
10.1126/science.aaf8933 *Corresponding author. Email: these first findings are encouraging, further

584 4 NOVEMBER 2016 • VOL 354 ISSUE 6312 SCIENCE

Neural circuits for pain: Recent advances and current views
Cedric Peirs and Rebecca P. Seal (November 3, 2016)
Science 354 (6312), 578-584. [doi: 10.1126/science.aaf8933]

Editor's Summary

Downloaded from on November 3, 2016

This copy is for your personal, non-commercial use only.

Article Tools Visit the online version of this article to access the personalization and
article tools:

Permissions Obtain information about reproducing this article:

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week
in December, by the American Association for the Advancement of Science, 1200 New York
Avenue NW, Washington, DC 20005. Copyright 2016 by the American Association for the
Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.