The Keystone Lodge in scenic Colorado, US, was the setting for a wide-ranging pain meeting from June 15-20, 2014. In the shadow of the Rocky Mountains, organizers David Borsook of Children's Hospital Boston, US; David Dodick, Mayo Clinic, Scottsdale, Arizona, US; and Frank Porreca, University of Arizona, Tucson, US, assembled an international group for five days of lectures, discussion, and the occasional outdoor activity. In his opening remarks, Porreca said the organizers sought to bring together different communities of researchers that remain separate, despite their related interests. The agenda included talks on chronic pain, headache, and migraine, with a focus on changes in the brain that sustain chronic pain. This is Part 1 of a series covering selected presentations from the meeting.
But first, a word about the spinal cord
Michael Salter, Hospital for Sick Children and University of Toronto, Canada, opened the meeting with a keynote talk that set the tone for things to come by being comprehensive, thought provoking, and featuring some new data.
Salter studies the involvement of spinal cord microglia in the generation of chronic pain. The work of Salter and others over the past decade has uncovered a critical role of microglia in fostering chronic pain after nerve injury. After injury to a peripheral nerve, microglia in the spinal cord are activated to produce BDNF (brain-derived neurotrophic factor), which then acts on GABAergic spinal cord lamina 1 inhibitory interneurons to alter chloride balance in the cells. As a result, the interneurons lose their ability to inhibit pain signals in the spinal cord. This leads to changes in the activity of nociceptive output neurons, which may account for symptoms of hyperalgesia, mechanical allodynia, and spontaneous pain after nerve injury. 
In mouse models of nerve injury, hyperalgesia starts days after nerve injury and lasts indefinitely. Knocking out BDNF in microglia prevents the development of hyperalgesia after nerve injury. But the question remained whether BDNF was only required to induce the hyperalgesia or if it was needed to maintain the pain as well. To answer that question, Salter used an inducible, central nervous system microglia-specific BDNF knockout that allowed the researchers to disable the BDNF gene after the mice developed hyperalgesia. When the BDNF knockout was induced 30 days after nerve injury, the researchers observed reversal of hypersensitivity. “This says that ongoing sensitization can be reversed by blocking BDNF,” Salter explained. Recently, in looking at mice that had pain for three months, Salter and colleagues saw a similar reversal of pain hypersensitivity. “There is one idea out there that microglia are involved early and astroglia later, but our data say that out to three months, microglia are still needed,” Salter said. The group will soon have six-month data, he added.
Both sides now
Salter has been working with Jeffrey Mogil of McGill University, Montreal, Canada, to look at sex differences in spinal cord inhibitory pathways. The results have been surprising, and perhaps a bit humbling. Salter presented unpublished data that suggest fundamental differences in the role of microglia, and the effects of the microglia inhibitor minocycline, in male versus female mice. The results reinforce the rationale for including both sexes in animal research (see PRF related news story).
Could sex differences have contributed to the failure of propentofylline, a microglia-targeted treatment that disappointed in human trials for diabetic neuropathy? The failure of that trial caused some to question the applicability of the microglia work, all of which was done in animals, to human pain (see PRF related news story). “They took animal data from nerve injury and asked about post-herpetic neuralgia,” said Salter. “Knowing that some of the molecular events are different, and now the possibility that there are sex differences, may have set the trial up for failure, even if the underlying mechanism is correct.”
Microglia in inflammatory pain
The role of microglia in inflammatory pain has not been as well studied compared to nerve injury. Ru-Rong Ji, Duke University, Durham, US, presented recently published data on the role of spinal cord microglia in inflammatory pain. Ji showed that in two inflammatory pain models (formalin or bradykinin injection into the mouse paw), caspase 6 (a proteinase previously implicated in neuronal apoptosis and axonal degeneration) is induced in the terminals of primary afferent C-fibers in the dorsal horn, and that activation of neuronal caspase 6 is required to produce inflammatory pain and central sensitization (Berta et al., 2014). Ji and colleagues hypothesized that caspase 6 released by primary afferents might cleave and release TNFα (tumor necrosis factor alpha) from spinal cord microglia. TNF is well known to be involved in both peripheral and central sensitization, and they showed that adding recombinant caspase 6 to microglia in culture caused release of TNFα. Furthermore, injecting the enzyme spinally resulted in TNF release and microglia-dependent pain. Caspase 6 also caused synaptic potentiation in spinal cord slides via a TNF-dependent mechanism.
The data led Ji to propose a novel role for caspase 6. In his model, the normally intracellular enzyme is released from active nociceptors and signals to microglia to produce TNFα, that then feeds back to enhance synaptic transmission and inflammatory pain. This novel pathway occurs rapidly, and in the absence of gliosis, to enhance inflammatory pain.
Ji also presented his work on the role of the other glia, astrocytes, in persistent neuropathic pain after nerve injury. Ji and colleagues recently showed that the release of the inflammatory cytokine CXCL1 from astrocytes plays a key role in sustaining neuropathic pain. That release, they showed, is stimulated by TNF and is controlled by the gap junction protein CX43 (see previous PRF conference coverage and Chen et al., 2014). Now, Ji also showed that gap junction inhibitors can block CXCL1 release and reverse neuropathic pain in the animals.
At last, into the brain
For a meeting on the brain in pain, the first session seemed to linger in the spinal cord, but that did not last long. One outstanding question about microglia is whether activation occurs in supraspinal structures in response to peripheral nerve injury. In a short presentation, Magdalena Luciuk, University of Sydney, Australia, showed the results of a whole-brain analysis of microglia phenotypes after nerve injury in rats. She used spinal nerve ligation in adult male rats, an injury that results in the development of mechanical allodynia over the course of three weeks.
Luciuk took 20 μm-thick sequential brain sections from rats three weeks after injury, at the peak of mechanical allodynia, and labeled them with a monoclonal antibody to the glia activation marker CD11b. She found that reactive microglia were rare, and were limited to the brainstem, in the medullary reticular region, ipsilateral to the nerve injury. This area has many functions, including descending pain modulation. The morphology of the cells revealed processes with club-like endings that resemble frog feet.
The meaning of the findings is not clear, but Luciuk speculated that the changes are likely to represent functional changes in the cells, which may play a role in descending analgesia. Microglia are hypothesized to modulate neurotransmission via nodes of Ranvier, and Luciuk showed that the microglia processes co-localized with nodal protein neurofascin.
Luciuk stressed that, although she looked at the whole brain, she found these specific morphological changes only in this one location, around the rostral ventral medulla. A preliminary time course suggested the changes were strongest at three weeks. Michael Salter commented that his lab had done a similar analysis two weeks after injury and saw changes only in the gracilis nucleus region of the medulla, an area that receives sensory input from the lower body. The work suggests that changes in microglia in the brain after nerve injury are limited and may be subtle.
The meaning of inflammation
Manuel Graeber, a neuropathologist from the University of Sydney, Australia, gave a discussion-provoking talk on the use of the term neuroinflammation, which he called “a term in need of definition.” This popular term is used to describe many conditions regardless of whether there is demonstrable neuropathological evidence of inflammation in the affected tissues, he said. In some diseases, there is true neuroinflammation in the brain or CNS—for example, the visible infiltration of immune cells that occurs in multiple sclerosis. But in many cases, Graeber said, this increasingly popular term is also used to refer to microglial activation. But microglia are resident, on-demand immune cells, and microglia activation is not inflammation, said Graeber. He argued that the terminology needs some cleaning up. In pain research, Graeber said, “Let’s look for a term more specific than neuroinflammation. Let’s stop talking about ‘inflamed’ microglia.” (See a very recent and more extensive explication of this topic in Graeber, 2014, and Estes and McAllister, 2014).
Does terminology matter? Graeber pointed to his own recently published work where his group performed a cluster analysis of gene expression data from brain tissue from people with Alzheimer's disease (AD), Parkinson's disease (PD), schizophrenia, and multiple sclerosis (Filiou et al., 2014). The group found that the patterns of gene expression observed in multiple sclerosis tissue matched that seen in classical inflammatory diseases, but the gene expression patterns in AD, PD, and schizophrenia brains were different. “You cannot compare Alzheimer's disease, schizophrenia, and Parkinson's disease to inflammatory conditions, because they are not,” Graeber said. Yet, “patients get treated on the basis of semantics.”
A few final points
Graeber echoed remarks by both Salter and Ji that the function, and functional states, of microglia in the CNS are complex. People often speak of “activated” microglia, suggesting that microglia are either on or off. That is far too simplistic.
In fact, an emerging understanding holds that the primary function of microglia is not to respond to injury or pathology. Instead, evidence shows that microglia function in the healthy brain to promote synaptic function and plasticity, and to help sculpt the connections that make up brain networks and circuitry (see the review by Graeber, 2010, and more recent reviews by Salter and Beggs, 2014, and Svahn et al., 2014).
The reality is that glia have many different states that remain to be defined at the morphological, molecular, and functional levels. For morphological studies, the pain field may do well to look to the work of speaker Giorgio Ascoli, George Mason University, Fairfax, Virginia, US. Ascoli is not a pain researcher, but the creator and curator of NeuroMorpho.org, a collection of three-dimensional digital reconstructions of neurons. Ascoli calls it “a gene bank for neuronal morphology.” Ascoli’s team watches the literature for papers with morphological data, requests the data from the authors, and then incorporates the new neurons into the collection.
“This used to be done lab by lab. Now we can pool these so we can have tens of thousands, rather than hundreds, of neurons to look at,” Ascoli said. The publicly accessible resource now contains reconstructions of 11,335 neurons and metadata contributed from 144 labs and has been used by researchers to generate more than 200 new publications. As of now, experimental conditions do not involve any chronic pain models, he said, but that is something that could be added.
What about a resource for microglia? It should be possible—Ascoli said that, to his knowledge, only a few labs are doing microglia morphology, but some have reconstructions of many cells.
For other coverage of the meeting, see Part 2 and Part 3.
Image: © iStockphoto.com/raisbeckfoto
