The consequences of pain in early life: injury-induced plasticity in developing pain pathways
Department of Neuroscience, Physiology & Pharmacology, University College LondonLondon, UK
Correspondence: Dr M. Fitzgerald, as above., E-mail: m.fitzgerald@ucl.ac.ukAbstract
Pain in infancy influences pain reactivity in later life, but how and why this occurs is poorly understood. Here we review the evidence for developmental plasticity of nociceptive pathways in animal models and discuss the peripheral and central mechanisms that underlie this plasticity. Adults who have experienced neonatal injury display increased pain and injury-induced hyperalgesia in the affected region but mild injury can also induce widespread baseline hyposensitivity across the rest of the body surface, suggesting the involvement of several underlying mechanisms, depending upon the type of early life experience. Peripheral nerve sprouting and dorsal horn central sensitization, disinhibition and neuroimmune priming are discussed in relation to the increased pain and hyperalgesia, while altered descending pain control systems driven, in part, by changes in the stress/HPA axis are discussed in relation to the widespread hypoalgesia. Finally, it is proposed that the endocannabinoid system deserves further attention in the search for mechanisms underlying injury-induced changes in pain processing in infants and children.
Introduction
Pain and injury in early life can cause lasting changes to developing somatosensory and pain systems. The immature nervous system in both humans and rodents is highly responsive to tactile and noxious stimulation; neurophysiological recordings reveal strong spinal nociceptive reflex activity and distinct nociceptive cortical potentials in response to clinically required skin-breaking procedures in newborn human infants (Slater et al., 2010b; Fabrizi et al., 2011; Cornelissen et al., 2013). This noxious evoked activity is more prolonged in the youngest infants and rat pups and decreases in duration with postnatal age (Fitzgerald & Gibson, 1984; Cornelissen et al., 2013). Importantly, when noxious stimulation in infancy is repeated or persistent, such as in neonatal intensive care or neonatal surgery, the effects outlast the period of stimulation itself and can result in profound and long-lasting changes in nociceptive neural pathways. By the time preterm infants reach term, they are already displaying enhanced cortical activity to an acute noxious stimulation compared with age-matched term-born controls (Slater et al., 2010a) and later in childhood these children display considerable alterations in somatosensory and pain processing (Hohmeister et al., 2009; Walker et al., 2009a). Taken together, studies show that both preterm and full-term children with neonatal intensive care unit (NICU) experience display greater perceptual sensitization to tonic heat (Hohmeister et al., 2009) accompanied by a generalized decreased sensitivity to cutaneous thermal (Walker et al., 2009a) and mechanical (Schmelzle-Lubiecki et al., 2007) stimulation. Specific injuries, such as moderate or severe burns in infancy, also result in greater pain and perceptual sensitization to noxious stimulation combined with depressed general mechanical and thermal sensitivity (Wollgarten-Hadamek et al., 2009). Severe infant burns are also associated with an attenuated social stress-induced analgesia, suggesting reduced function in phasic endogenous pain inhibitory mechanisms later in childhood and adolescence (Wollgarten-Hadamek et al., 2011). Major surgery within the first 3 months of life increases pain sensitivity and analgesic requirements to subsequent surgery compared with infants with no prior surgery (Peters et al., 2005). The different patterns of altered sensory and pain sensitivity in these studies may reflect differences in experienced pain, stress and analgesic treatment in these children but nevertheless it is clear that early painful injuries can induce local and global, long-term alterations in sensory and pain processing.
Long-term changes in pain processing in ex-preterm and early injured infants are likely to have several underlying causes. The number of tissue-breaking, presumed painful, procedures in neonatal intensive care is directly correlated with reduced brain white matter and subcortical grey matter (Brummelte et al., 2012) as well as delayed corticospinal development (Zwicker et al., 2013) and lower postnatal growth (Vinall et al., 2012), all with wide-ranging complex effects upon central nervous system (CNS) function. The aim of this review is to focus on the specific neurobiological mechanisms underlying injury-induced plasticity in the developing pain system. We summarize the main animal models in this field and review the evidence for activity-dependent cellular and synaptic changes in nociceptive circuitry in these models. We discuss two ways in which these changes may be maintained, namely neuroimmune activation and altered descending pain control, driven from the hypothalamic/pituitary/adrenal (HPA) axis. Finally, we suggest that the endocannabinoid system may be an interesting target for the prevention of injury-induced plasticity in the developing pain pathways.
Rodent models of long-term effects of neonatal tissue injury
Understanding how pain and injury in early life can cause lasting changes to developing somatosensory and pain systems requires good animal models. A key requirement of such a model is that the effect is ‘age sensitive’ or has a ‘critical period’ such that the same injury applied in older or adult animals fails to have the same long-lasting effect on pain processing (Walker, 2013). For identification of neonatal critical periods in animal studies, appropriate controls in adulthood are crucial. The same degree and intensity of injury at different ages must be achievable to isolate specific developmental effects associated with injury or stress during a particular developmental period.
Table1 shows the different rodent models of long-term effects of neonatal injury upon pain processing. Many have been adapted from adult pain research and compare similar types and severities of injury in neonatal and adult rodents. These include: neonatal hindpaw plantar incision (Walker et al., 2009c), hindpaw inflammation with agents such as carrageenan (CAR) or complete Freund’s adjuvant (CFA) (Beland & Fitzgerald, 2001; Walker et al., 2003; Ren et al., 2004; Hohmann et al., 2005; LaPrairie & Murphy, 2007), full thickness skin wound (Reynolds & Fitzgerald, 1995; Beggs et al., 2012a), repeated needle prick (Anand et al., 1999; Knaepen et al., 2013), peripheral nerve injury (Howard et al., 2005; Vega-Avelaira et al., 2012) and visceral injury caused by distension or inflammation (Al-Chaer et al., 2000; Randich et al., 2006; Wang et al., 2008).
| Model | Neonatal injury | Effects on adult baseline sensitivity | Effects on adult response to re-injury | Sex differences | Controls |
|---|---|---|---|---|---|
| Surgical incision | Hindpaw plantar incision | Generalized baseline hyposensitivity (F. Schwaller & S.M. Walker, pers. comm.) | Re-incision: enhanced segmental mechanical and thermal hyperalgesia (Hathway et al., 2009b; Walker et al., 2009bc; Beggs et al., 2012b) | (Beggs et al., 2012b) | Brief neonatal handling & anaesthesia. Age-matched adults undergoing their first incision |
| Abdominal surgery | Laparotomy (P0) | Generalized baseline hyposensitivity (Sternberg et al., 2005) | N/A | ? | Brief maternal separation, anaesthesia and saline injection (also have baseline hypoalgesia). *Not compared with adult laparotomy |
| Skin wound | 1 mm × 1 mm hindpaw skin removal | Behavioural hypersensitivity and hyperinnervation at site of injured skin (Reynolds & Fitzgerald, 1995; Beggs et al., 2012a) | N/A | ? | Age-matched non-wounded animals *Hyperinnervation, but not behaviour, compared with adult skin wound |
| Needle | Four needle pricks daily between P0 and P7 | Thermal hyperalgesia (Anand et al., 1999) | Increased hypersensitivity after adult paw incision (Knaepen et al., 2013) | Hypersensivity to CFA injection in males only (Knaepen et al., 2013) | Neonatal tactile stimulation. *No comparison made to adult needle pricks (Anand et al., 1999). |
| Inflammation | Hindpaw carageenan (CAR) or complete Freunds adjuvant (CFA) injection | Generalized baseline hyposensitivity (detectable after P34) (Ren et al., 2004). | Enhanced hyperalgesia after re-inflammation (Ren et al., 2004) Increased capsaicin-evoked hyperalgesia (Hohmann et al., 2005) | Exacerbated hyposensitivity in females (LaPrairie & Murphy, 2007) | Neonatal saline injection Age-matched adults undergoing their first CAR/CFA injection |
| Visceral distension & inflammation | Colorectal distension (CRD) or mustard oil application daily Bladder inflammation | Increased sensitivity to distension. Increased noxious heat responses in abdominal and paw skin (Al-Chaer et al., 2000; Wang et al., 2008) No difference in thermal/mechanical threshold baseline response (Randich et al., 2006). Bladder hypersensitivity (DeBerry et al., 2010) | Increased abdominal hyperalgesia following bladder inflammation with zymosan (Randich et al., 2006) | ? | Brief neonatal handling. Age-matched adults undergoing their first bout of CRD |
| Nerve Injury | Spared nerve injury (SNI) | No initial neuropathic pain response (Howard et al., 2005). Hypersensitivity appears after P30 (Vega-Avelaira et al., 2012) | N/A | ? | Sham surgery at P10 (thigh incision) Compared with adult animals undergoing nerve injury |
A common feature of these models is that a tissue injury at a critical period of development has long-term effects, outlasting the injury itself, resulting in adults with altered pain sensitivity compared with controls (Fig.1). The difference in pain sensitivity is assessed when the animals are adults as (1) changes in baseline sensory and nociceptive sensitivity as compared with controls (handled in exactly the same way as neonates but with no injury) and (2) changes in pain sensitivity (hyperalgesia) to re-injury as adults, compared with controls (adults receiving their first injury). This approach reveals a dual long-term effect of mild injury to the hindpaw (CAR inflammation or surgical incision) administered in the first 7–10 days of life. Mild injuries are associated with a widespread whole body baseline depression in sensory and nociceptive thresholds, or hyposensitivity, that emerges only when the rat is adolescent, i.e. 4–5 weeks old (Ren et al., 2004; Sternberg et al., 2005; Laprairie & Murphy, 2009). However, the area in and around the site of the neonatal injury retains an enhanced sensitivity to pain, so that a new injury applied to the region results in enhanced hyperalgesia that is greater in amplitude and more prolonged than controls (Ren et al., 2004; Chu et al., 2007; Walker et al., 2009b; Beggs et al., 2012b). The enhanced pain sensitivity can be observed within days of the first injury, but importantly is also present in the adult, long after the original neonatal injury has resolved. Neither the basal hyposensitivity nor the enhanced re-injury-associated hyperalgesia subside with age and are still evident in 120–125-day-old rats (Ren et al., 2004). Importantly, none of these effects occurs if the early inflammation or skin incision is administered after the first 7–10 days of life (Fig.2).
More extensive skin and subcutaneous, joint or visceral tissue injuries in early life also have long-term effects but are not associated with widespread hyposensitivity. These are characterized by a long-lasting mechanical sensitivity in and around the site of the neonatal injury, long after it has healed (Reynolds & Fitzgerald, 1995; Beggs et al., 2012a). This is especially striking after neonatal visceral injury, where repeated distension or mild inflammation in the first 2 weeks of life results in a long-lasting baseline visceral hypersensitivity, which also spreads outside the viscera to include the overlying abdominal and nearby cutaneous tissues (Al-Chaer et al., 2000; Randich et al., 2006; Wang et al., 2008; Christianson et al., 2010). Furthermore, when the animals are adults, a new visceral insult causes a greatly enhanced hyperalgesia compared with controls (Randich et al., 2006; DeBerry et al., 2010). Unlike milder cutaneous injuries, the critical period for these visceral effects is not constrained to the very early postnatal period (Fig.1). Thus, different types of injury may trigger different long-term effects in the somatosensory and pain system.
Peripheral nerve injuries in early life have a very different long-term effect on pain from neonatal tissue injury or inflammation. In the first postnatal days, the neurotrophin dependency of primary sensory neurons means that nerve injury causes substantial sensory neuron death (Huang & Reichardt, 2001) but even after this period, for the first 3 weeks of life little or no neuropathic pain results from nerve injury, in marked contrast to the effect in adults (Howard et al., 2005). However, later in life, when the animal reaches adolescence, nerve injury causes a distinct and persistent mechanical hypersensitivity (Vega-Avelaira et al., 2012).
The existence of two long-lasting changes in pain sensitivity after early tissue damage, increased pain and sensitivity to repeat injury on the one hand and widespread hypoalgesia on the other, each with different onset times, suggests distinct underlying mechanisms, some of which may lie outside the pain system (see Box ).
Boxed Text
The influence of early life injury beyond pain pathways
Many models of early life pain reveal changes to the CNS beyond the pain system. The overlap between pain and reward pathways (Borsook et al., 2007) suggests that neonatal pain experience may influence reward-related pathways and, in support of this, repeated neonatal skin incision does cause alterations in adult brain motivational orexinergic pathways, known to modulate mesolimbic dopaminergic reward circuitry, in an animal model of novelty-induced hypophagia (Low & Fitzgerald, 2012). Complex changes in alcohol preference are also observed following neonatal formalin and heel stick (Anand et al., 1999; Bhutta et al., 2001) and reduced exploratory activity following early visceral distension (Wang et al., 2008).
Neonatal pain has also been shown to influence the generation of new neurons in the dentate gyrus region of the hippocampus, thought to be involved in memory formation. Rat pups receiving intraplantar injections of the painful inflammatory agent complete Freund’s adjuvant, on P8, had more BrdU-labelled cells and a higher density of cells expressing doublecortin, both measures of newborn neurons, in the subgranular zone of the dentate gyrus (Leslie et al., 2011). In addition, neonatal paw inflammation causes long-term alterations to acute stress responses in adulthood, depending on the type of injury. A single intraplantar injection of carrageenan in the first postnatal week reduces adrenocorticotrophic hormone and glucocorticoid levels and reduces anxiety-like behaviours during the forced swim test in adulthood (Anseloni et al., 2005; Victoria et al., 2013) while hindpaw inflammation with complete Freund’s adjuvant and formalin increases anxiety-like behaviour in adults during elevated plus maze testing and forced swim tests (Roizenblatt et al., 2010; Mohamad et al., 2011; Negrigo et al., 2011).
Cannabinoids and long-term plasticity of pain pathways
Agents aiming to reduce the long-term effects of early injury in infants and children would ideally combine analgesia with protection against physiological stress and immune activation. The cannabinoids are primarily analgesic, via CB1 receptors expressed in peripheral and central pain pathways modulating GABAergic and glutamatergic neurotransmission (Rea et al., 2007) and potentiating GlyRs (Xiong et al., 2011). They have strong immunomodulatory properties mainly mediated by CB2 receptors localized on immune cells (Guindon & Hohmann, 2009).
The endocannabinoid system is well developed at birth and high levels of CB1 expression in the human infant cortex suggest strong endocannabinoid regulation of presynaptic neurotransmission in the first few years of human life (Long et al., 2012). The system has a number of important functional roles in the neonatal brain, but from the perspective of this discussion the participation of endocannabinoids in the rapid suppression of neonatal physiological stress by glucocorticoids should be noted (Buwembo et al., 2012). Endocannabinoids are neuroprotective against neonatal sciatic nerve injury-induced cell death in neonatal rats (Perez et al., 2013) but it is not known if they could also protect against injury-induced plasticity in developing pain pathways. Importantly, the developmental trajectory of this system is sensitive to early life experience. Maternal deprivation reduces endocannabinoid ligand and receptor expression in the brain (Suárez et al., 2010) and neonatal LPS selectively down-regulates CB1 receptor expression in the amygdala (Zavitsanou et al., 2013). Conversely, CB1 receptor expression in the brain is significantly up-regulated by neonatal capsaicin, a C nociceptor toxin (Zavitsanou et al., 2010). Further exploration of the role of endocannabinoids in the developmental plasticity of pain pathways would be of great interest.
Boxed Text
Neonatal environmental, endocrine and immune stressors and the developing HPA axis
Stress activates the HPA axis: secretion of corticotropin-releasing hormone (CRH) from the hypothalamic paraventricular nucleus (PVN) controls adrenocorticotrophic hormone (ACTH) release from the pituitary which, in turn, regulates release of adrenal glucocorticoids (GCs), such as cortisol in humans or corticotropin (CORT) in rodents, which modulate physiological stress responses in a transient and reversible manner. Exposure to stress in neonatal rodents ‘programmes’ the HPA axis and other areas of the brain (Lupien et al., 2009). A critical developmental period of reduced responsiveness to stressful stimuli between P3 and P14 in rodents is maintained by maternal presence, which keeps levels of GCs low and is essential for normal neural and behavioural maturation (Sapolsky & Meaney, 1986; Walker & Vrana, 1993; Levine, 2002). Increasing GC levels during this period either by exogenous administration of GCs or by abnormal early life stress causes long-term changes to neuroendocrine and behavioural responses to future stressors (Plotsky & Meaney, 1993; Levine, 2002). More active maternal care during neonatal life is associated with reduced HPA-axis activation and behavioural responses to stressors in adulthood (Liu et al., 1997; Macrì et al., 2008) while maternal deprivation is associated with HPA-axis hyperactivity caused by increased CRH mRNA levels and CRH transcription in the PVN (Plotsky & Meaney, 1993; Chen et al., 2012). In addition to maternal grooming, other factors such as altered feeding patterns and temperature homeostasis during maternal separation alter HPA-axis responses later in life (Rüedi-Bettschen et al., 2004; Spencer, 2013). Neonatal endocrine and immune system interactions are also important for establishing stress responses later in life. Neonatal exposure to LPS increases HPA-axis responses to stress and to subsequent immune challenges in adulthood and are associated with an anxiety-like phenotype in adult rats which can be inherited by untreated offspring of the neonatally LPS-treated rats (Spencer et al., 2006; Walker et al., 2009a, 2012).
Future perspectives
There is much still to understand in this important field of pain research. Future efforts should focus upon clarifying how the type, severity and site of injury affects the adult phenotype and the relative importance of stress vs. pain in this process. Equally important is to move beyond the effect itself and focus upon the peripheral and central mechanisms underlying injury-induced plasticity in developing pain pathways.
Glossary
- CAR
- carrageenan
- CNS
- central nervous system
- GlyR
- glycine receptor
- HPA
- hypothalamic/pituitary/adrenal
- LPS
- lipopolysaccharide
- PAG
- periaqueductal grey
- RVM
- rostroventral medulla