Building a Case for the Astrocyte Contribution to Pain Signaling
Glial cells have recently emerged as important regulators of chronic pain, particularly in the spinal cord where they contribute to the neuronal sensitization thought to underlie neuropathic pain hypersensitivity. Among glial cells, microglia have garnered most of the attention so far, but astrocytes are increasingly recognized as major players in chronic pain (see PRF related news story). New work presented at the annual meeting of the Society for Neuroscience, held November 9-13, 2013, in San Diego, California, US, further supports a role for astrocytes in pain signaling.
Lead investigator Ru-Rong Ji at Duke University, Durham, North Carolina, US, and colleagues show that peripheral nerve injury in mice increases expression of the hemichannel protein connexin 43 (Cx43) in astrocytes. That results in the release of chemokines that sensitize spinal cord neurons and promote chronic pain.
Maiken Nedergaard, a senior collaborator on the study at Rockefeller University, New York, US, told PRF in an email, “This work shows that astrocytes, in response to acute pain, may upregulate signaling pathways that contribute to chronic pain.”
Astrocytes and pain hypersensitivity
The current study is part of an emerging body of work from Ji’s group and others showing that astrocytes have important pain signaling functions that may be distinct from those of microglia (for a review, see Ji et al., 2013). “Accumulating evidence shows that astrocyte activation [in the spinal cord following nerve injury] could be long-lasting and more persistent than the microglial reaction, indicating that astrocytes can play a key role in pain maintenance distinct from microglial mechanisms,” said Ji.
In an earlier study, Ji’s team showed that spinal cord astrocytes release the chemokine CCL2 (also known as monocyte chemoattractant protein 1 [MCP1]) following nerve injury in mice, which contributed to the development of pain hypersensitivity (Gao et al., 2009). In more recent work, they also showed that another chemokine, CXCL1, and its receptor CXCR2, were upregulated in astrocytes and neurons, respectively, and contributed to pain hypersensitivity in the spinal nerve ligation model of neuropathic pain in mice (Zhang et al., 2013).
Ji described chemokines as smaller versions of cytokines, the inflammatory signaling peptides released from many immune cells and from astrocytes. “We are sure these chemokines contribute to neuropathic pain,” Ji told PRF. “Now the question we have addressed here is how they are released and controlled in astrocytes.”
A role for a hemichannel protein
Gap junctions serve to fasten two neighboring astrocytes together by forming a pore that passes through the membrane of both cells, connecting their cytoplasm. To form the passageways, each of the two adjoining cells contributes one hemichannel protein; in astrocytes, that protein is Cx43. Gap junctions allow signaling molecules such as ATP, glutamate, and calcium to pass freely from cell to cell, endowing astrocytes with rapid communication capabilities. Individual Cx43 hemichannels can also open to the extracellular space, providing a portal for release of astrocytic mediators such as chemokines. Following nerve injury, astrocytes lose functional gap junctions, so Ji postulated that the cells might switch from fast gap-junction signaling to slower paracrine signaling through Cx43 hemichannels.
Nedergaard’s group, together with Ji’s team, had previously found that Cx43 contributes to the development of pain hypersensitivity in a mouse model of spinal cord injury (Chen et al., 2012). In the new work, Ji, presenting author Gang Chen (also at Duke), and colleagues extended that finding to peripheral nerve injury using the chronic constriction injury (CCI) model and uncovered more details about how astrocyte signaling sensitizes neurons and maintains neuropathic pain.
The researchers began by looking for Cx43 in the spinal cord after CCI and found that the protein was upregulated in spinal cord astrocytes. Importantly, the increased expression lasted three weeks, into the so-called late phase of neuropathic pain, suggesting that Cx43 was important for chronic pain maintenance.
Next, the team set out to understand the behavioral consequences of Cx43 upregulation in the animals, using von Frey hairs to stimulate the hindpaw three weeks after CCI and then measuring mechanical pain thresholds. As expected, CCI animals’ withdrawal thresholds had fallen after injury, but that mechanical hypersensitivity was attenuated by intrathecal injection of the Cx43 blocker carbenoxolone (CBX) or a Cx43 mimetic peptide that blocked the pore-forming hemichannels, indicating that functioning Cx43 pores were required for mechanical hypersensitivity.
Electrophysiological experiments further indicated a role for Cx43 in mediating long-term pain hypersensitivity. Recordings from lamina II neurons in acute spinal cord slices taken from CCI animals three weeks after injury showed that spontaneous excitatory postsynaptic currents (EPSCs) were increased compared to control animals, an effect that was dampened by CBX or the mimetic peptide.
Signaling via chemokines
The group then wanted to better understand the mechanism behind the Cx43 pain signaling observed in the CCI animals by studying cultured astrocytes. It has been well established that astrocytes treated with the inflammatory cytokine tumor necrosis factor alpha (TNFα) enter an activated state, which includes the release of chemokines (Gao et al., 2009). In the new research, astrocytes cultured in a dish and treated with TNFα released more of the CXCL1 chemokine, a change that was prevented by CBX. The cultured cells also responded to TNFα treatment with increased Cx43 hemichannel activity, indicating flux through the pores.
The researchers next took the experiment in vivo. After treating cultured astrocytes with TNFα (and then washing the cells to remove any trace of the cytokine, which would otherwise have many direct effects in the spinal cord), they injected the activated cells intrathecally to uninjured mice. Within two hours of injection, the mice developed mechanical hypersensitivity similar to nerve-injured animals that was attenuated by a CXCL1-blocking antibody. “The activated cells released something that causes pain, and we believe they release chemokines because the pain could be reversed by the chemokine antibody,” Ji said. “Hyperactivity of Cx43-mediated hemichannels in neuropathic pain makes astrocytes ‘leaky,’ so they release large molecules such as chemokines,” he added.
In another experiment, the investigators pretreated astrocytes with small interfering RNA (siRNA) to block expression of Cx43 before activating the cells with TNFα. “These astrocytes no longer induced pain,” said Ji, showing that Cx43 was required for chemokine release and pain sensitivity.
Finally, electrophysiological recordings from spinal cord neurons indicated that CXCL1 increased spontaneous EPSCs when applied to slice preparations from naïve animals. Slices taken from nerve-injured mice three weeks after CCI already displayed elevated EPSCs. In both cases, “when we added an antagonist specific to the CXCL1 receptor CXCR2—which we knew was expressed on neurons—we saw the EPSCs diminish,” Chen told PRF.
“Those experiments really demonstrate that chemokines in the spinal cord can directly increase synaptic transmission,” Ji said. “Taken together, the results show that hemichannels formed by Cx43 can release chemokines such as CCL2 and CXCL1, further demonstrating that their release from astrocytes is sufficient to induce neuropathic pain,” he added.
Ji pointed out that a key difference underlying astrocyte versus microglial signaling is that, at synapses, astrocytes reside much closer to neurons than microglia do. “In terms of structural connectivity between cells, astrocytes have tighter connections with synapses—they directly cover them. So they may be the first cells to detect changes in synaptic activity,” he explained.
That is not to diminish the roles of microglia in the spinal cord, Ji noted. “It is not that microglia are not important for neuropathic pain maintenance; it’s not black and white,” he said. Many questions still need to be tackled for both microglia and astrocytes, including how glia-neuron signaling changes neuronal activity and how those changes can last so long. “How do glial cells maintain chronic pain? We see this work as a starting point,” he said.
Stephani Sutherland, PhD, is a neuroscientist, yogi, and freelance writer in Southern California, US.
Image: Spinal cord astrocytes labeled for Cx43 (green) and glial fibrillary acidic protein (red), an astrocyte marker. Credit: Gang Chen, Duke University.
Article Contributors:
