Microglial activation has long been implicated in neuropathic pain, but the signal from injured sensory neurons that prompts this response has remained elusive—until now. Researchers led by Zhonghui Guan, Julia Kuhn, and Allan Basbaum, University of California, San Francisco, US, report that the cytokine colony stimulating factor 1 (CSF1), undetectable in healthy neurons, is induced de novo in both sensory and motor neurons following nerve injury and transported to the spinal cord, where it acts on the microglial CSF1 receptor (CSF1R). Furthermore, they find that CSF1 is both necessary and sufficient to induce microglial activation and mechanical hypersensitivity in the spared nerve injury (SNI) model of neuropathic pain. The researchers further report that CSF1 also triggers microglial proliferation, but that only the activation and pain were dependent on the downstream microglial adaptor protein DAP12.
“We have long known that microglia are involved in the development and maintenance of pain after nerve injury. This new study really elegantly and convincingly fills in the gap about how a nerve tells the spinal cord, and the microglia within it, that an injury has happened,” said Simon Beggs, University of Toronto, Canada, who was not involved with the new study.
The study was published online December 7 in Nature Neuroscience.
CSF1: The missing link
Microglia, the central nervous system’s resident macrophages, are activated after peripheral nerve damage and thought to play a role in the generation of neuropathic pain (see PRF related news story; Beggs et al., 2012; Ji et al., 2013). However, the exact signal(s) used by injured sensory neurons to trigger microglia activation continued to elude pain researchers. Several candidates such as ATP, matrix metallopeptidase 9 (MMP-9), and the chemokine CCL2 had been identified, though none seemed to be both necessary and sufficient to induce microglia activation, Basbaum told PRF.
In the new study, the researchers went after the mystery signal by using RNA sequencing (RNA-Seq) to perform an unbiased analysis of the genes expressed in the dorsal root ganglia (DRG) following nerve injury in mice. They identified a number of genes whose expression was altered in ipsilateral DRG neurons in the week following sciatic nerve ligation and transection, including CSF1. Levels of messenger RNA (mRNA) encoding CSF1R were also increased in the ipsilateral dorsal spinal cord following nerve injury. However, expression of interleukin 34 (IL-34), another CSF1R ligand, was unchanged following nerve injury. These findings were confirmed using quantitative RT-PCR.
“Hundreds and hundreds of genes changed [after nerve injury]. We were struck by the induction of CSF1 and CSF1R because they have long been known to be necessary for the development of microglia,” Basbaum said. For example, CSF1 induces microglia proliferation in vitro (Suzumura et al., 1990), and CSF1R is necessary for microglia development (Ginhoux et al., 2010). “We found that was worth pursuing,” he added.
The team next investigated the tissue location of CSF1 using in situ hybridization and immunostaining. Strikingly, the investigators did not detect CSF1 in healthy DRG neurons, but instead found that it was dramatically upregulated in damaged neurons within 18 hours of nerve injury in the SNI model of neuropathic pain. This effect persisted for at least three weeks. Additional experiments showed that CSF1 underwent intra-axonal transport to the spinal cord, where CSF1R was expressed exclusively in microglia and upregulated following nerve injury. Subsequent experiments revealed that CSF1 was also upregulated in nerve-injured motor neurons, and required for ventral horn microglial activation and proliferation.
“[Following peripheral nerve injury] CSF1 is turned on, goes to the spinal cord, and its target is CSF1R that is upregulated in microglia,” Basbaum said.
Necessary and sufficient
Guan, Kuhn, Basbaum, and colleagues next investigated the consequences of preventing CSF1 induction, using Cre recombinase-mediated deletion of CSF1 specifically in DRG sensory neurons. Microglial development did not differ from that of wild-type mice prior to nerve injury. However, as expected, microglia activation (as measured by expression of IBA1, a marker of activated microglia) was significantly reduced in the ipsilateral dorsal horn of nerve-injured CSF1-deleted mice compared to nerve-injured wild-type mice. Behavioral experiments showed that deletion of CSF1 also completely prevented mechanical hypersensitivity following SNI.
“Cutting a peripheral nerve without inducing CSF1 dramatically reduced the activation of microglia and mechanical hypersensitivity. That says that CSF1 is necessary for the induction of microglia [and pain], but it doesn’t say that it’s sufficient,” said Basbaum.
To address the sufficiency question, the team injected CSF1 intrathecally once a day for three days into uninjured, wild-type mice. This injection significantly increased IBA1 expression in the dorsal horn and produced mechanical hypersensitivity comparable to that produced by nerve injury. The hypersensitivity produced by intrathecal CSF1 was prevented by the microglial inhibitor minocycline, but somewhat unexpectedly persisted in mice in which the P2X4 purinergic receptor was deleted. The latter result indicates that CSF1 can exert its effects independent of ATP, a molecule previously shown to be involved in microglia activation following nerve injury. In addition, intrathecal CSF1 induced the expression of several downstream microglial genes, including brain-derived neurotrophic factor and cathepsin S, that have previously been implicated in the development of neuropathic pain (Coull et al., 2005).
“These results show that intrathecal administration of CSF1 is [not only necessary but] also sufficient for the induction of microglia and mechanical hypersensitivity,” Basbaum said.
To identify what the activated microglia released that triggered mechanical hypersensitivity, the researchers then returned to their RNA-Seq data, zeroing in on the adaptor protein DAP12, which was also upregulated following nerve injury. DAP12 is exclusively expressed in microglia and is involved in microglial development and proliferation, Basbaum said.
Consistent with a role for DAP12 in mediating CSF1-induced microglial activation and neuropathic pain, intrathecal administration of CSF1 into uninjured wild-type animals also induced DAP12 expression, and deletion of DAP12 completely prevented the mechanical hypersensitivity induced by both nerve injury and CSF1 administration.
In a final set of experiments, the investigators tackled another puzzling question about the role of microglia following peripheral nerve injury—the origin of increased numbers of these cells in the spinal cord. Whether this increase stems from the infiltration of circulating monocytes, which differentiate into microglia, or from local microglia self-renewal is controversial, Basbaum said (Priller et al., 2001; Ajami et al., 2007). To answer this question, the researchers again turned to their RNA-Seq analysis. They found that several microglial genes were upregulated following nerve injury, but levels of monocyte-specific gene expression remain unchanged, suggesting that the increased numbers of microglia came from local microglia self-renewal.
Subsequent experiments using incorporation of the thymidine analog bromodeoxyuridine (BrdU) as a measure of microglia proliferation confirmed local self-renewal in the dorsal horn and showed that de novo CSF1 expression was also necessary to induce microglial proliferation after nerve injury. Moreover, intrathecal CSF1 induced microglial proliferation, again demonstrating sufficiency. Surprisingly, however, this effect was not DAP12 dependent, said Basbaum. “It looks like CSF1 triggers both the DAP12-dependent neuropathic pain phenotype and a microglial proliferation phenotype that is DAP12 independent,” he added.
Overall, the results provide a better understanding of how nerve injury leads to microglia activation and pain. And, “the new findings potentially provide several new targets—CSF1, CSF1R, and DAP12—that have huge therapeutic potential for neuropathic pain,” said Beggs. The fact that CSF1 is not normally expressed in neurons and is only induced following nerve injury makes it an especially attractive drug target, as any CSF1-targeted treatment would be less likely to produce side effects, he added. Basbaum also noted the potential benefit of the persistence of IL-34 after CSF1 deletion—microglia themselves will be preserved and thus able to continue their phagocytosis and other helpful effects.
Beggs noted that he and others have recently shown there are substantial sex differences in how neuropathic pain is generated and maintained. While microglia are necessary for the mechanical allodynia following nerve injury in male mice, the story appears to be different in females, whose pain instead depended on T cells, according to recent research (see PRF related news story). “The current study only looked at male mice, so it will be interesting to see if CSF1 has the same effect in female mice,” Beggs said.
Allison Marin, PhD, is a neuroscientist-turned-science writer who resides in Pittsburgh, Pennsylvania, US.
Image credit: Guan et al., 2015, with permission from Macmillan Publishers Ltd.