Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Study of the development of pain pathways and mechanisms is fundamental to our understanding and treatment of the many infants and children around the world who suffer acute or chronic pain. Increasing evidence that exposure to tissue injury and excessive noxious input in early life can change the sensitivity of pain pathways in adulthood underlines the importance and impact of these studies. This chapter summarizes the key maturational processes that take place throughout development at every level of the pain system from peripheral nociceptor transduction to the cortical representation of pain. The cellular, synaptic, and circuit processes that underpin the emergence of pain are placed in the context of mammalian infant pain behavior and the pediatric clinical pain experience.
Pain mechanisms, pain plasticity, and pain experience change with age . Although even the youngest preterm infant is capable of displaying a robust physiological response to tissue injury, such nociceptive responses in early life are not simply a rudimentary form of the more complex adult pain experience; they have characteristic features that change as the child develops from infancy through to adolescence. Evidence is also emerging that modulation of the nociceptive system by tissue damage and inflammation is dependent on age, thus suggesting that the pattern and degree of hyperalgesia, allodynia, and other characteristics of pain states differ in young and adult subjects.
Historically, infant and childhood pain has been undertreated ( ), and even now, when pain management in children is benefiting from increased evidence-based data, many treatments for the youngest patients continue to be empirically rather than scientifically based ( , ). Understanding of the developmental neurobiology of pain processing is essential for appropriate measurement and treatment of children’s pain and for understanding how it may shape the future pain experience in adulthood.
Developmental changes, which are likely to have an impact on the experience and treatment of pain in infancy and childhood, are found at every level of the nociceptive pathway, from transduction of noxious stimuli at nociceptor terminals, transmission of neural activity along peripheral sensory nerves, integration and modulation by nociceptive circuits in the dorsal horn, descending supraspinal control arising in the brain stem, and cortical processing of centrally transmitted nociceptive activity. Since many of these processes are dependent on balanced sensory input at critical developmental ages, it is not surprising that tissue injury and excess noxious input in early life can modify the development of pain pathways.
Almost everything that we know about the biology of pain has been obtained from laboratory studies in animal models, and thus their importance in providing and continuing to provide fundamental knowledge in this field cannot be overestimated. This research is commonly undertaken in young rats and mice, both naïve and with local injuries, to model the experience of human infants and children. Examples of models that we have used are paw or ankle injections of carrageenan and complete Freund’s adjuvant (CFA) to produce short- and long-lasting inflammation, spared nerve injury and chronic constriction nerve injury in the sciatic nerve to produce neuropathic pain, and plantar skin incision to model surgery. Rats are altricial animals, like humans, but they develop very rapidly and are considered adult by 8 weeks of age, which makes age translation difficult. Figure 9-1 illustrates the developmental stages of rats over the postnatal days and weeks and how these stages can be broadly translated to human infants, children, and adolescents.
Pain development begins with the emergence of nociceptors as a distinct subpopulation of sensory neurons, separate from low-threshold mechanoreceptors and proprioceptors. The majority of sensory neurons originate from neural crest cells and become specified as nociceptors, thermoreceptors, mechanoreceptors, or proprioceptors in the dorsal root ganglia (DRGs) ( ). Both the survival and phenotype of sensory neurons depend on signaling from distinct neurotrophin receptors belonging to the Trk (tyrosine kinase) family ( ). Beginning at embryonic day 11 (E11) in the rat, axons from DRGs innervate the periphery in an organized proximal-to-distal manner such that by birth, sensory fibers have reached the most distal skin of the foot ( ). Sensory fibers grow to peripheral and central targets with remarkable precision, and from the outset each DRG innervates characteristic skin dermatomes ( ). Target innervation critically depends on both nerve growth factor (NGF) signaling and activation of the serum response factor (SRF) transcription factor ( ), as well as the interaction of Trk receptors with leucine-rich repeat and immunoglobulin (LIG) family proteins ( ). The overall number of DRG neurons does not change significantly after birth, but neurotrophin support continues to be required for the first postnatal week in rodents, as shown by the dramatic fall after neonatal axotomy ( ).
The generation of small-diameter nociceptive neurons is dependent on the interaction between intrinsic transcription factors expressed by the developing neurons and external, target-derived signals, which together specify functionally distinct sensory neurons for pain, temperature, and itch (Lui and Ma 2011). The initial specification of the broad category of small-diameter C-fiber nociceptors depends on the expression of neurogenin 1 (Ngn1) transcription factors in early embryonic life ( ). These neurons initially all express TrkA and require NGF for survival but, later in embryonic development, are divided into peptidergic and non-peptidergic classes under the influence of the transcription factor Runx1 ( Fig. 9-2 ). Thus, a subset of these neurons down-regulate TrkA during early postnatal life and become dependent on glial cell line–derived neurotrophic factor (GDNF) via expression of the GDNF receptor Ret, which can modulate the mechanical sensitivity of these nociceptors ( , ). In this way, the Ret-expressing, isolectin B4 (IB4)-positive, non-peptidergic phenotype emerges as a distinct group of nociceptors only postnatally. This switch to Ret signaling is driven by NGF itself, which also initiates expression of the transcription factor Runx1 ( ) and thereby leads to the extinction of calcitonin gene–related peptide (CGRP) expression and promotion of the non-peptidergic phenotype ( ). Meanwhile, Runx1 activity is suppressed in peptidergic neurons via hepatocyte growth factor (HGF)-Met signaling ( ). Recent evidence suggests that brain-derived neurotrophic factor (BDNF) signaling during the neonatal period is also required for the survival of both peptidergic and non-peptidergic nociceptors ( ). Figure 9-2 summarizes these events.
The generation of TrkC-expressing proprioceptors and cutaneous mechanoreceptors (expressing TrkB) can be distinguished from nociceptors by their dependence on the Ngn2 transcription factor ( ). Within the Ngn2 population, the activity of transcription factor Runx3 promotes the proprioceptive identity ( ) by decreasing TrkB expression in a subset of TrkC-expressing neurons ( ). Although many functional subtypes of low-threshold mechanoreceptors have been characterized, recent evidence shows that afferents innervating Merkel cells, hair follicles, and Meissner corpuscles all originate from a subpopulation of embryonic DRG neurons that co-express the transcription factor MafA with the GDNF receptors Ret and glial cell line-derived neruotrophic factor family receptor (GFR)-α2 shortly after genesis within the DRG ( ). These “early” Ret-expressing neurons constitute a separate population from the non-peptidergic nociceptors that express Ret later in life ( ).
Nociception begins with the transduction of noxious mechanical, thermal, or chemical stimuli through a set of specific transient receptor potential (TRP) channels expressed in nociceptor terminals. TRP expression is detected in embryonic life ( ), but with some notable subtype differences. The capsaicin receptor TRPV1, which responds to heat, protons, toxins, and the chilli pepper ingredient capsaicin, is expressed by E12.5 in the mouse, whereas TRPA1 channels, which respond to chemical irritants and noxious cold, are not expressed in nociceptors in the first weeks of life ( ), which explains the weak nociceptive reflex responses to the TRPA1 agonist mustard oil (allyl isothiocyanate) in young animals ( ). Conversely, expression of adenosine triphosphate (ATP)-gated P2X 3 ion channels undergoes a postnatal switch from all DRG cells to IB4-positive neurons only, which explains the postnatal reduction in electrophysiological and behavioral sensitivity to P2X receptor (P2XR) agonists (Chen et al 2012). Control of the expression patterns of sensory receptors and channels during development is an area of rapidly changing research. Runx1, which is initially expressed in most TrkA-positive neurons, together with two other transcription factors, islet1 and Brn3a, coordinates the expression of a variety of sensory channels and receptors, including P2X 3 , TRPA1, TRPM8, the Mrgpr family of G protein–coupled receptors (GPCRs), the sodium channel Na v 1.9, and the high TRPV1 expression observed in a small subset of DRG neurons ( ).
Functional polymodal nociceptors can be identified in fetal and neonatal life ( ) by their distinctive patterns of spike activity in response to a range of cutaneous noxious stimulation, including some activity that is not normally observed in adults ( , ). Nociceptors can be clearly distinguished from low-threshold mechanoreceptors in the fetal rat ( ), but the stimulus–response characteristics of the latter continue to mature over postnatal life ( , , ). The ability of unmyelinated nociceptors to respond to noxious heat depends on NGF during the early postnatal period inasmuch as treatment with NGF antisera between postnatal day 2 (P2) and P14 converts C-fiber mechano-heat receptors to a novel type of low-threshold mechanical pressure receptor ( ) whereas systemic neonatal NGF treatment leads to an increase in the proportion of C-fiber heat nociceptors ( ). Meanwhile, the number of cutaneous A-fiber mechano-heat nociceptors decreases during the first 10 postnatal days ( ), although whether this relates to postnatal changes in neurotrophin signaling remains unclear. Conduction velocities in immature afferent sensory nerves are lower than in mature nerves because the axons are smaller and incompletely myelinated. As a result, nociceptive transmission through Aδ- and C-fiber axons and innocuous sensory transmission through Aβ axons is slower in young mammals, and the frequency of action potential (AP) firing is also lower because of immature ion channel properties, which leads to less reliable stimulus frequency coding. The tetrodotoxin-resistant sodium channel Na v 1.8 (SNS/PN3) is expressed in developing C-fiber neurons by E17, with adult levels observed by P7 ( ).
Complete understanding of nociceptive processing in the immature nervous system requires characterization of the ontogeny of spinal networks in the dorsal horn. The spinal cord develops along a ventrodorsal gradient such that deep dorsal horn neurons are born after motoneurons and substantia gelatinosa (SG) cells are the last to mature ( ). Within the SG, supraspinal lamina I projection neuron generation is complete before that of propriospinal neurons and local circuit interneurons ( ). The specification of neurotransmitter phenotype within these interneurons is under tight regulatory control from an extensive complement of homeodomain (HD) and basic helix–loop–helix (bHLH) transcription factors.
From E11.5, two late-born populations of dorsal horn neurons populate the superficial laminae and are distinguished by their complementary expression of the HD genes Lmx1b and Pax2 ( ). Pax2 , along with Lbx1 , is required to induce the differentiation of both GABAergic ( ) and glycinergic ( ) interneurons within the developing dorsal horn. The specification of glutamatergic interneurons in the embryonic dorsal horn results from active repression of the GABAergic phenotype by the HD transcription factors Tlx3 and Tlx1 ( ). Recent work also suggests a role of the transcription factor DRG11 (i.e., Prrxl1 ) in promoting the glutamatergic phenotype within the dorsal horn ( ). Thus, the final proportions of GABAergic and glutamatergic interneurons in the embryonic dorsal spinal cord appear to be determined by the balance between Tlx3 and Lbx1/Pax2 expression ( ). This balance is also instrumental in the generation of dorsal horn neurons that express neuropeptides since cholecystokinin- and substance P (SP)-expressing neurons depend on Tlx3 and Tlx1 ( ) whereas cells containing enkephalin, neuropeptide Y, or galanin appear to require the expression of Ptf1a , Lbx1 , and Pax2 ( ).
Interneurons within the superficial dorsal horn (SDH) undergo significant morphological changes during the early postnatal period ( ) that depend in part on calcium-calmodulin–dependent kinase IIα (CAMKIIα) signaling ( ). Despite this growth of dendritic and axonal processes in the neonate, the thickness of laminae I and II outer remains relatively constant from birth to adulthood ( ). SDH neurons also exhibit developmental changes in their intrinsic membrane properties during early life, including up-regulation of A-type K + currents and an increase in AP amplitude ( ). The patterns of AP discharge seen in SDH neurons shift with age, with a greater proportion of cells firing repetitively in response to intracellular current injection ( ). In addition, the neonatal SDH is distinguished by the presence of “pacemaker” neurons within lamina I, which are defined by their intrinsic ability to generate spontaneous, rhythmic burst firing ( ). Although this pacemaker activity appears to be confined to interneurons during early life, both spino-parabrachial and spino–periaqueductal gray (PAG) lamina I projection neurons exhibit spontaneous activity throughout postnatal development ( ).
Large-diameter myelinated A-fiber afferents are the first to grow into the dorsal horn at E15 in the rat and from 11 weeks in humans ( , , ). The initial A-fiber terminal fields are diffusely distributed in the dorsoventral and rostrocaudal dimensions, gradually reorganize during the early postnatal period, and exhibit a nearly adult pattern by P21 ( , ). The more dorsal termination of A fibers in the first postnatal weeks means that they overlap in the SDH with C-fiber terminals. Not all A-fiber terminal arbors do this, but those that do appear to form synaptic contacts in lamina II ( , , ). The functional class of afferent that gives rise to these exuberant terminals is not clear, but they may arise partly from myelinated high-threshold mechanoreceptors ( ). Nonetheless, a high incidence of Aβ-evoked monosynaptic responses has been observed in patch-clamp studies of immature SG neurons ( , ), and polysynaptic Aβ-fiber synapses onto GABAergic neurons in laminae I–II are more prevalent in neonates ( ). In addition, activation of Aβ fibers evokes c-fos expression in the SG at P3 ( ), and repetitive A-fiber stimulation can produce sensitization in dorsal horn neurons in neonates but not in adults ( ).
Therefore, neurons in the SDH receive a greater degree of Aβ-mediated input during the neonatal period, either from direct Aβ-fiber projections to the region or from low-threshold, polysynaptic connections.
Neonatal pain processing requires the formation of central connections between nociceptive primary afferents and spinal cord neurons. Nociceptive projections to the dorsal horn are not observed until E19 in the rodent ( ) and are dependent on activation of the paired HD transcription factor DRG11 ( ). From the outset, these fibers terminate in a somatotopically precise manner in laminae I–II of the dorsal horn ( ) and include SP, CGRP, galanin, and somatostatin-expressing fibers ( , ). The timing of C-fiber growth into the human fetal spinal cord remains unclear, but no afferent terminals were observed in lamina II at 19 weeks by DiI labeling ( ). Despite the anatomical presence of C fibers in the rodent spinal cord during the embryonic period, evidence points to delayed maturation of C-fiber synaptic connections in the SDH during the early postnatal period. Extracellular single-unit recordings in vivo have demonstrated that although low-intensity electrical stimulation of A fibers evokes APs in dorsal horn neurons from P3, long-latency C-fiber–evoked activity is not apparent until P10 ( ). This is not due to an absence of functional synapses before birth since electron microscopic studies have demonstrated the existence of synaptic terminals corresponding to small-diameter sensory afferents from birth ( ). In addition, capsaicin application increased release of glutamate in the SDH from P0, thus suggesting that nociceptive afferents expressing TRPV1 begin forming functional synapses before birth. During the neonatal period, the number of C-fiber synapses may be insufficient to effectively drive suprathreshold membrane depolarization in SDH neurons, whereas a subsequent proliferation of C-fiber synaptic input between P5 and P10 ( ) may underlie the appearance of C-fiber–evoked spiking at later ages.
Since the role of glutamate in the excitatory pathways related to pain is well established, the expression of both ionotropic receptors (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors [AMPARs]/kainate receptors, N -methyl- d -aspartate receptors [NMDARs]) and metabotropic glutamate receptors (mGluRs) within the developing dorsal horn is of key importance in neonatal pain ( ). The subunit composition of AMPARs appears to shift during the early postnatal period. Neonatal rat SG neurons exhibit higher levels of mRNA for the GluR1, 2, and 4 subunits than adult SG neurons do, although the ratio of GluR2 to GluR1, 3, and 4 is lower during early life ( ). Since the presence of GluR2 reduces the Ca 2+ permeability of AMPARs, this suggests the possibility of greater AMPAR-dependent Ca 2+ influx in neonatal spinal neurons. Ca 2+ -permeable AMPARs are expressed in both inhibitory (GABAergic) and excitatory (neurokinin 1 [NK1] positive) embryonic dorsal horn neurons in culture ( ), and Ca 2+ -permeable and -impermeable AMPARs contribute to primary afferent synaptic transmission throughout the first 3 postnatal weeks ( ). High levels of NMDAR expression are seen in the developing spinal cord of both rodents ( ) and humans, and immunoreactivity for the NMDAR subunits NR1 and NR2A–2D has been detected as early as 4.5 weeks’ gestation ( ). There is evidence of an age-dependent alteration in NMDAR subunit composition within the dorsal horn. NMDA excitatory post-synaptic currents (EPSCs) in SG neurons have rapid decay kinetics from birth, thus suggesting high levels of NR2A ( ). However, the unusually high Mg 2+ sensitivity of NMDARs in neonatal rat SG neurons predicts a novel subunit stoichiometry (Green and Gibb 2001), and NR2B subunits contribute to large, spontaneous EPSCs in neonatal lamina I neurons ( ). Alterations in the subunit composition of NMDARs may explain the observation that both receptor affinity for NMDA and NMDA-evoked calcium influx decline with postnatal age in rat SG neurons ( ). Interestingly, these changes were delayed following neonatal capsaicin treatment, which suggests that C-fiber synaptic input is important for the maturation of NMDAR function within the SDH. Finally, recent work suggests that the composition of NMDARs may vary across cell types in the developing dorsal horn since inhibitory interneurons exhibit greater expression of the NR2C/D subunits ( ).
Synapses possessing only NMDARs (“silent synapses”) are found in neonatal dorsal horn neurons and are developmentally down-regulated ( , , ), but they are unlikely to play a prominent role in excitatory synaptic transmission because 93–94% of glutamatergic synapses in laminae I–II of the neonatal rat express AMPARs ( ).
Since SP and the NK1 receptor (NK1R) play a key role in pain and hyperalgesia, the development of this excitatory transmitter system is also important. In rats, SP receptor density in the spinal cord as a whole is maximal in the first 2 postnatal weeks and decreases six-fold between P11 and P60. A postnatal shift in the laminar distribution of NK1R expression also occurs, with very few SP receptors in the neonatal SG gradually increasing to high levels by the second week of life ( ), although this is less clear in the newborn human cord ( ). Lamina I neurons express NK1R by the second postnatal week ( , ). The ontogeny of NK1R expression in the immature SDH coincides with the strengthening of C-fiber synaptic input during early life ( ), and the presence of C-fiber terminals appears to be required for normal NK1R expression in the adult dorsal horn ( ). Because C fibers directly synapse onto NK1R-positive SDH neurons ( ), one possible mechanism might involve the co-release of CGRP, which can increase NK1R mRNA levels and 125 I-SP binding in cultured neonatal dorsal horn neurons ( ).
Nociceptive processing requires effective inhibitory as well as excitatory transmission within the dorsal horn through the two major neurotransmitters, γ-aminobutyric acid (GABA) and glycine ( Fig. 9-3 ).
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here