Reactive astrocytes in pain neural circuit pathogenesis

Abstract

Reactive astrocytes are commonly activated in the spinal dorsal horn (SDH) of various animal models of pathological pain. Previous investigations suggest an association between astrogliosis and pain pathogenesis. However, our understanding of the mechanisms underlying astrogliosis activation and the contributions of reactive astrocytes to pain neural circuit malfunction is rudimentary. This short review highlights recent advances in these areas.

Introduction

Astrocytes are a major type of glia in the central nervous system (CNS) of mammals. Emerging evidence indicates that astrocytes modulate neuronal circuits involved in sensory and cognitive functions, including pain signal processing in the spinal dorsal horn (SDH) [1,2]. Astrocytes intimately interact with neurons to support neuronal functions via various mechanisms, including metabolic coupling [3]. Astrocytic processes often wrap neuronal synapses to form a complex known as the tripartite synapse, through which astrocytes modulate synaptic activity by removing excessive extracellular neurotransmitters (e.g., glutamate) and releasing gliotransmitters (e.g., glutamate, D-serine, and ATP) [4]. The bidirectional interaction between neurons and astrocytes is essential for maintaining the homeostatic functions of neural circuitry under physiological conditions [5].

Astrocyte–neuron interactions might be altered under pathological conditions. Various stresses or insults have been shown to induce astrogliosis, a process during which astrocytes undergo morphological and physiological changes and differentiate into activated or reactive astrocytes [6]. Although reactive astrocytes are commonly identified using specific molecular markers, such as increased glial fibrillary acidic protein (GFAP) expression, they are diverse in morphology and probably exert different biological functions [7]. For example, reactive astrocytes are thought to include a proinflammatory, neurotoxic A1 subtype and an anti-inflammatory, neuroprotective A2 subtype [8]. Reactive astrocytes likely have dysregulated astrocyte–neuron interactions that may cause dysfunction of neuronal circuits underlying neurological disorder [9,10].

Pathological pain is one of the most common neurological conditions associated with astrogliosis. In the last two decades, converging evidence from preclinical models suggests that reactive astrocytes critically contribute to pain pathogenesis. The objective of this short review is to provide an update on recent progress toward understanding the role of astrocyte–neuron interactions in pain pathogenesis rather than providing a comprehensive and balanced overview, which can be found in other reviews [11–13].

Reactive astrocytes in pain pathogenesis

Reactive astrocytes were initially identified in the spinal cord of a rat neuropathic pain model, created by sciatic nerve constriction injury [14] and have since been observed in various other pain models such as inflammatory pain [15] and cancer pain [16]. The relevance of astrogliosis in pain pathogenesis in humans is suggested by the observation of reactive astrocytes in the SDH autopsies specifically from HIV patients who developed chronic pain [17]. It was suggested that chronic pain is a gliopathy [18,19]. The pathogenic significance of reactive astrocytes was tested using glial inhibitors. However, the results from such studies were not conclusive because of potential non-specific effects of the inhibitors. Other studies speculate the contribution of reactive astrocytes based on genetic manipulations of astrocyte-specific genes. The interpretation of such results is complicated by its circumventive nature because they do not allow one to directly evaluate the role of reactive astrocytes, especially when the genes are also expressed in naïve astrocytes that are not undergo astrogliosis.

Recently Liu et al. used genetic approaches to directly evaluate the role of reactive astrocytes in pain pathogenesis in two pain models. Their previous work indicates that development of pathological pain in HIV patients is associated with astrogliosis in the SDH [17]. To determine the pathogenic contribution of reactive astrocytes, they used the GFAP-thymidine kinase (TK) transgenic mice [20] to generate mouse models of HIV-associated pain by intrathecal (i.t.) injection of HIV-1 envelope glycoprotein 120 (gp120) [21]. Reactive astrocytes were selectively ablated in the models by i.t. injection of ganciclovir. The authors found that reactive astrocyte ablation abolished gp120-induced pain [22]. In a separate study, they used GFAP-TK mice to determine the significance of reactive astrocytes in the mouse model of opioid-indued hyperalgesia (OIH), and found that ablation of reactive astrocytes during the OIH induction phase also blocked the expression of OIH [23]. To specify the temporal role of reactive astrocytes, the authors performed delayed ablation of reactive astrocytes when OIH had already established and showed that the delayed ablation was able to reverse the OIH at this stage. These results indicate that reactive astrocytes were critical for both OIH expression and maintenance. Hence, data from both pain models support a crucial role of reactive astrocytes in pain pathogenesis.

Ablation of reactive astrocyte may disrupt the integrity of neural circuits; however, additional experiments demonstrated that astrogliosis inhibition via genetic manipulation also suppressed pain development in these models [22,23]. In these studies, the conditional knockout (CKO) of either Wnt5a in neurons or its receptor, receptor tyrosine kinase–like orphan receptor 2 (ROR2), in astrocytes inhibited both astrogliosis and hyperalgesia in both pain models. These findings provide compelling evidence that the reactive state of astrocytes is important for pain pathogenesis.

The results of cell ablation experiments indicate that proliferation of astrocytes is important for critical for the pain pathogenesis in both the gp120 and OIH models [22,23]. They also raise the interesting possibility that Wnt5a-ROR2 signaling, which is critical for not only the pain development but also astrogliosis in these models [22,23], may specifically contribute to the cell proliferation aspect of astrogliosis induced by gp120 or morphine. Although this hypothesis remains to be tested, it is consistent with the proposed role of this pathway in regulation of proliferation and morphogenesis of other cells [24,25]. In addition to the gp120 and OIH models, the role of Wnt5a signaling in pain pathogenesis was also proposed in multiple other models associated with astrogliosis, including multiple sclerosis-associated pain [26], nucleoside reverse transcriptase inhibitor (NRTI)-induced pain [27], chronic post-thoracotomy pain (CPTP) [28] and neuropathic pain [29], suggesting the Wnt5a-ROR2 pathway is a general mechanism underlying chronic pain as well as pain-associated astrogliosis.

Astrocyte activation during pain pathogenesis

How astrocytes are activated to assume reactive states in pain neural circuits remains largely unknown and may depend on the pain context [11]. Proinflammatory mediators are frequently thought to elicit astrogliosis. For examples, interleukin 1 beta (IL-1β), tumor necrosis factor-alpha (TNF-α), and chemokines have been reported to activate astrogliosis in the SDH of various pain models [11–13]. Because astrogliosis is often attenuated by inhibiting specific cytokines (e.g., IL-1β or TNF-α), these factors may function sequentially in a neuroinflammatory network that supports astrogliosis. These factors may activate specific intracellular signaling pathways, such as the c-Jun-N-terminal kinase (JNK) and signal transducer and activator of transcription 3 (STAT3) pathways, to induce astrogliosis during pain pathogenesis [30,31].

Reactive microglia are thought to play key roles in the regulation of astrogliosis by releasing cytokines or chemokines to stimulate astrocytes and promote pain pathogenesis [32]. Indeed, microglia and astrocytes are known to interact reciprocally [33,34]. According to a popular model, microglia are activated to support the initiation phase of pathological pain, followed by the later activation of astrocytes to support the maintenance phase of pathological pain [32]. This model is generally consistent with the temporal profiles of microgliosis and astrogliosis and the reported contribution of microglia to the early phase of pathological pain development [22,35]. However, it remains to be rigorously validated. A recent study showed that microglial ablation in an OIH model did not affect spinal astrogliosis [36], indicating microglia are dispensable for astrogliosis in this pain model. It will be interesting to determine the potential contribution of microglia to astrogliosis in other pain models.

Another possible mechanism for astrocyte activation during pain pathogenesis involves alterations in neuronal activity. Although neurons and astrocytes interact intimately, the potential role of neuronal activity in stimulating astrogliosis during pain pathogenesis is not well understood. Studies have shown that pain signals elicit transient astrocytic reactions or activation, indicated by increased intracellular Ca²⁺ levels [1,2]. Although interesting, whether such pain-induced astrocytic reactions are critical for initiating astrogliosis remains unclear.

Studies on neuronal Wnt5a strongly suggest a key role for neurons in controlling astrogliosis during HIV-1 gp120-induced pain pathogenesis [22]. Wnts are secreted signaling proteins that control cellular proliferation and differentiation during development and oncogenesis [37]. In the CNS of adult animals, Wnts, including Wnt5a, are predominantly expressed in neurons, concentrated at synaptic compartments [38,39], and secreted in a synaptic activity–regulated manner [38,40]. Wnt signaling is implicated in the pathogenesis of pain associated with various diseases [26,41], and the CKO of Wnt5a in neurons or its receptor, ROR2, in astrocytes blocks gp120-induced astrogliosis in the SDH [22]. The neuronal Wnt5a–astrocytic ROR2 intercellular signaling pathway is also essential for astrogliosis in the SDH of OIH models [23]. These results suggest that Wnt5a secreted from neurons, likely elicited by neuronal activity evoked under different pain conditions, stimulates astrogliosis by binding to ROR2 on astrocytes (Figure 1). Consistent with this model, ROR2 was recently proposed to regulate the A1 subtype differentiation of reactive astrocytes [42]. Wnt5a activates multiple downstream noncanonical Wnt signaling pathways, including JNK, Ca²⁺ signaling [43], and STAT3 pathways [44]. The JNK and STAT3 pathways may play critical roles in spinal astrogliosis in pain models [30,31], and Ca²⁺ is a hallmark of transient astrocyte activation induced by pain signals [1,2]. Whether Wnt5a–ROR2 signaling regulates these astrocytic intracellular signaling pathways to alter reactive astrocytic phenotypes during pain pathogenesis remains unknown. In addition to the Wnt5a-ROR2 pathway, studies by Jiang et al. suggest that C–X–C motif ligand 13 (CXCL13) secreted by neurons may also stimulate astrogliosis via binding with astrocytic C–X–C motif receptor 5 (CXCR5) in neuropathic pain models [45]. Thus, multiple neuron-to-astrocyte signals may be involved in the regulation of astrogliosis during pain pathogenesis. Whether Wnt5a- and CXCL13-regulated astrogliosis are pain context-dependent remains to be determined.

Figure 1. A mechanism of neuron-astrocyte bidirectional interaction during pain pathogenesis.

Previous work suggests that the synthesis and secretion of Wnt protein in neurons is controlled by neuronal activity elicited by various stimuli [38,40,52]. Recent studies from Liu et al. indicate that Wnt5a is a neuron-to-astrocyte signal that is essential for astrogliosis induced by HIV-1 gp120 or morphine and that astrocytic Wnt5a receptor ROR2 mediates this activity of Wnt5a [22,23]. The studies also suggest that IL-1β is a potential reactive astrocyte-to-neuron signal that disturbs the homeostasis of pain neuronal circuits in the spinal dorsal horn (SDH), as suggested by the expression of neural circuit polarization (NCP). Ablation and inhibition of astrogliosis or antagonizing IL-1 receptor blocks NCP and hyperalgesia induced by gp120 or chronic morphine. NCP may provide a circuitry mechanism for the expression of pathological pain.

Reactive astrocyte–mediated mechanisms of pain neural circuitry pathogenesis

How reactive astrocytes affect neural circuits during pain pathogenesis is elusive. Normal astrocytes contribute to the maintenance of synaptic homeostasis through multiple mechanisms, including the uptake of extracellular neurotransmitters through glutamate transporters, the release of gliotransmitters that may stimulate synaptic receptors, and the maintenance of ion homeostasis [46]. Astrogliosis is expected to disturb these biological processes of normal astrocytes and consequently dysregulate the activity of pain neuronal circuits. However, few studies have been performed to directly test how astrogliosis might contribute to pain circuitry abnormalities underlying pain pathogenesis.

Recent studies by Xu et al. suggest that spinal astrocytes may regulate nociceptive signal–induced synaptic plasticity in SDH pain neural circuits [1]. The authors found that activation of spinal astrocytes in naïve mice gates the expression of long-term depression (LTD) in neurokinin 1 receptor (NK1R)-positive projection neurons when stimulated by large-diameter afferent fibers (Aβ-fibers) [1]. Future studies to determine if and how reactive astrocytes might dysregulate LTD may provide important insights into astrogliosis–mediated circuitry malplasticity during pain pathogenesis.

Work from Liu et al. revealed an essential role for reactive astrocytes in a novel form of neural circuitry malplasticity in the SDH of mouse models of HIV-associated pain [22]. They showed that HIV-1 gp120 induced increase of excitatory postsynaptic currents (EPSCs) in excitatory neurons and concomitantly decrease of EPSCs in inhibitory neurons, a phenomenon they named neural circuit polarization (NCP) [22]. Importantly, ablation of reactive astrocytes blocked both NCP and mechanical allodynia induced by gp120. Similar inhibitory effects on NCP and pain expression were also achieved by blocking the astrogliosis via CKO of neuronal Wnt5a or astrocytic ROR2. These findings provide strong evidence for the critical role of reactive astrocytes in disruption of SDH pain neural circuit homeostasis during the expression of gp120-induced mechanical allodynia. In a separate study, the authors showed that Wnt5a–ROR2-regulated reactive astrocytes are also required for NCP in the OIH model [23], indicating a general role for reactive astrocytes in disruption of neural circuitry homeostasis underlying the development of different types of pathological pain.

It appears that reactive astrocytes produce IL-1β to support the expression of NCP and mechanical allodynia induced by gp120 and chronic morphine. IL-1β neutralization blocks NCP in models of gp120-induced pain and OIH [22,23]. Interestingly, reactive astrocytes produce IL-1β via a matrix metalloproteinase 2 (MMP2)-dependent pathway in the gp120 pain model, whereas IL-1β is produced by an inflammasome-mediated pathway in the OIH model [22,23]. These findings indicate that reactive astrocytes may use IL-1β as a feedback signal that promotes gp120-or morphine-induced NCP. Many interesting questions remain, including the synaptic loci (pre-vs. postsynaptic) of IL-1β action and the molecular mechanism by which IL-1β oppositely modulates EPSCs in excitatory and inhibitory neurons.

In addition to IL-1β, reactive astrocytes may release other factors that disturb pain neuronal circuit homeostasis. For example, monocyte chemoattractant protein-1 (MCP-1) produced by astrocytes has been suggested to support central sensitization in neuropathic pain models [47]. More recently, Luo et al. proposed that IL-17 produced by astrocytes induced neuronal hyperexcitability in the SDH of chemotherapy-induced neuropathy models [48].

Work from Romanos et al. revealed another mechanism through which dysfunctional astrocytes disrupt pain neural circuits [49]. Using a genetic mouse model of familial hemiplegic migraine type 2, the authors showed that impaired astrocytic glutamate uptake due to reduced expression of astrocytic Na⁺/K⁺-ATPases resulted in enhanced cingulate cortex pyramidal neuron output by promoting N-methyl-D-aspartate (NMDA) spikes. Their studies show that impaired astrocytic function altered neuronal activity in the cingulate cortex, which is implicated in cranial pain pathogenesis [49].

Conclusion

Recently, considerable progress has been made toward understanding the role of reactive astrocytes in pain pathogenesis. The new findings solidify a firm foundation for further mechanistic and translational investigations of astrocyte-mediated pathogenic processes. An emerging theme is the critical contribution of dysregulated astrocyte–neuron interactions to the pathogenesis of pain neural circuits. Previous studies have focused on the role of potential astrocytic factors in disrupting pain neural circuit homeostasis, whereas more recent studies have extended the mechanistic understanding of pain neural circuit pathogenesis by identifying neuron-to-astrocyte signals that control reactive astrocyte activation. Similar to the expanding array of astrocyte-to-neuron signals, the spectrum of neuron-to-astrocyte signals is likely to expand, although the current list remains short. As illustrated by Wnt5a, a potential common property of neuron-to-astrocyte signals is their activity-regulated secretion, which may provide a mechanism through which pain-induced neuronal activity is coupled with astrogliosis. The coordinated action of neuronal activity–regulated astrogliosis and astrocyte-mediated neuronal hyperactivation may create a bidirectional, self-perpetuating astrocyte–neuron interaction that sustains reactive astrocyte activity and neuronal hyperexcitability to maintain pathological pain (Figure 1). Establishing this self-perpetuating feedback loop is also likely to be a critical pathogenic event in initiating and maintaining chronic pain, although this model requires further testing in future studies. Disruption of the astrocyte–neuron intercellular signals may inhibit the self-perpetuating mechanisms of pathological pain initiation and maintenance and potentially serve as an attractive therapeutic approach for preventing and treating pathological pain. Emerging data from single-cell transcriptomic analyses of animal models have identified subtypes of reactive astrocytes in various CNS diseases [50,51], but the specific contribution of different reactive astrocytic subtypes to pain pathogenesis remain to be defined. Although current studies focus on the pathogenic role of reactive astrocytes, some astrocytic subtypes (e.g., A2 subtype) may play protective roles in pain neural circuits. Identifying protective reactive astrocytes may offer insights into designing interventions targeted at preventing pathological pain development and promoting the resolution of chronic pain states.