Like a thorny vine climbing along a wrought iron gate, pain and anxiety are inextricably intertwined. Anxiety disorders often develop in people with chronic pain, and being anxious worsens pain. Now, research led by Min Zhuo at the University of Toronto, Canada, indicates that a novel form of synaptic plasticity in the anterior cingulate cortex (ACC) may account for anxiety associated with chronic pain. The work was published online December 31 in Neuron.
“The study moves us one step closer to understanding the interaction between pain and anxiety in the brain,” write David Seminowicz, University of Maryland, Baltimore, US, and Marta Ceko, National Center for Complementary and Integrative Health, Bethesda, US, in a commentary for PRF (see complete comment below). Seminowicz and Ceko were not involved in the work.
The ACC plays a critical role in chronic pain, and human studies suggest that it might be a hub for anxiety as well. In people with chronic pain, anxiety levels predict the extent of pain, and in imaging studies, people with anxiety show increased activation in the ACC. Zhuo and colleagues had previously shown that changes in synaptic strength in the ACC contribute to the genesis of neuropathic pain, through a form of long-term potentiation (LTP), in which AMPA-type glutamate receptors are increased on the postsynaptic membrane (see PRF related news story).
In the new study, Zhuo and colleagues looked at a different kind of LTP, generated by increased glutamate release from presynaptic terminals onto pyramidal neurons of the ACC. Their results show that this presynaptic LTP (pre-LTP) is induced after nerve injury or inflammation, and contributes not only to pain, but also to anxiety in a mouse model of neuropathic pain.
Incoming LTP
To characterize presynaptic LTP in the ACC, first author Kohei Koga induced pre-LTP with a low-frequency stimulation protocol in brain slices from adult mice. As expected based on previous work in hippocampus and amygdala, pre-LTP did not require NMDA-type glutamate receptors, postsynaptic calcium, G protein-coupled receptors, or protein kinase M zeta (PKMζ), all of which participate in postsynaptic LTP. Pre-LTP did involve kainate-type glutamate receptors (KAR), the investigators showed. Adenylyl cyclase 1 (AC1), the enzyme that produces the second messenger cyclic AMP, and protein kinase A (PKA) were also required for pre-LTP, as they are for post-LTP in the ACC.
Pre-LTP in the hippocampus involves hyperpolarization-activated cyclic nucleotide-gated (HCN) ion channels. Using electron microscopy of gold-particle immunolabeling and electrophysiological recordings, the researchers found HCN channels and currents in both pre- and postsynaptic terminals in the ACC and, importantly, in thalamic neurons that project to the ACC. Block of HCN channels in the ACC prevented pre-LTP and reversed established pre-LTP, indicating the channels are required for induction and maintenance of plasticity. That result suggests that HCN antagonists might selectively dampen pre-LTP and the behaviors it underlies.
A cause for anxiety?
To better understand the functional physiological consequences of pre-LTP, the researchers turned to animal models of pain. Acute pain did not induce pre-LTP, but in mice with chronic inflammation or nerve injury, the synaptic amplification was saturated, the researchers found. They used transgenic mice with a fluorescent reporter protein to show which specific ACC neurons had been activated by nerve injury. “The ACC lights up like a Christmas tree—we see a lot of neurons activated by injury,” Zhuo said. In those injury-activated neurons, blocking HCN channels significantly decreased their activity, demonstrating that pre-LTP was present in the mice with chronic pain.
The ACC is known to play a role in providing emotional context to experiences, so the authors next investigated how anxiety impacted pre-LTP. Their experiments showed that pre-LTP was induced in normal, uninjured mice by exposure to a maze designed to amp up anxiety. In mice allowed to recover from the anxiety-provoking experience, however, pre-LTP had diminished, indicating that an anxiety-producing situation can evoke transient pre-LTP. Interestingly, classical fear conditioning did not produce pre-LTP in the mice, indicating that pre-LTP in the ACC contributes to anxiety but not fear.
The researchers next tested the idea that pre-LTP in the ACC underlies the anxiety that accompanies chronic pain. Compared to sham-operated mice, mice with nerve injury displayed significantly more anxiety-like behaviors, but those behaviors were dampened by bilateral injection of an HCN blocker into the ACC. Importantly, the blocker injection also attenuated mechanical allodynia in nerve-injured mice, suggesting that dampening pre-LTP could reduce neuropathic pain behaviors. Blocking activity of PKMζ to prevent post-LTP reduced allodynia but had no effect on anxiety behavior in nerve-injured mice, confirming that pre-LTP is specifically tethered to anxiety.
Studies have shown that ACC activity is heightened not only in people experiencing physical pain, but psychological pain such as divorce, the death of a loved one—basically with suffering, Zhuo said. Much like pain patients, “we know those people suffer tremendous anxiety. We think our study will really link pain of all sorts with anxiety at the level of the ACC.” If scientists could one day find a way to safely prevent pre-LTP in the ACC, perhaps by blocking HCN channels, it might alleviate these forms of anxiety as well as in chronic pain.
Stephani Sutherland, PhD, is a neuroscientist, yogi, and freelance writer in Southern California.
Comments
David Seminowicz, University of Maryland
This comment was co-authored
This comment was co-authored by Marta Ceko, National Center for Complementary and Integrative Health, Bethesda, US.
Anxiety and pain interact in different ways: acute, prolonged, and inescapable pain is likely to induce anxiety; acute anxiety can be analgesic; chronic anxiety seems to exacerbate pain; and chronic pain and anxiety might be part of a vicious feedback cycle, with each intensifying the other. In chronic pain and anxiety, it has been difficult to determine which came first: are anxious people more likely to develop chronic pain, or does chronic pain lead to anxiety?
In their recent study, Koga and colleagues moved us one step closer to understanding the mechanism for anxiety and pain interaction in the brain. They determined that two types of long-term potentiation—one occurring presynaptically (pre-LTP), the other postsynaptically (post-LTP)—occur in the anterior cingulate cortex (ACC) in mice. While the existence of pre-LTP and post-LTP has been characterized in the hippocampus and amygdala, cortical LTP has been thought to consist mainly or entirely of post-LTP, and the blockage of post-LTP through inhibition of AMPA receptors has been shown to produce analgesia. The current study investigated pre-LTP in the ACC, which involves kainate receptors (KARs) and a pathway involving AC, PKA, and HCN channels (the authors performed many careful experiments to confirm the involvement of KARs and the intracellular pathways, and the exclusion of post-LTP mechanisms). HCN channels were more highly expressed in thalamocortical (thalamo-ACC) neurons than other neurons in either ACC or thalamus, and are required for the development and maintenance of pre-LTP (but not post-LTP). This indicated that HCN could be important in the regulation of thalamocortical modulation.
Once the characteristics of pre-LTP in the ACC were established, the authors went on to test the role of pre-LTP in multiple pain conditions (acute thermal, inflammation, nerve injury) as well as in fear (classic fear conditioning) and anxiety-like behaviors (elevated plus maze [EPM], forced open arm exposure). Pre-LTP was abolished in chronic pain models, but not in acute pain. Similarly, anxiety, but not fear, attenuated pre-LTP. Microinjections of an HCN blocker in the ACC did not reduce anxiety-like behavior in non-injured mice. However, in mice with chronic pain (nerve injury model), anxiety-like behaviors were enhanced compared to shams, and this enhanced anxiety (measured on both the EPM and open field tests) was reduced by blocking HCNs in the ACC. The authors suggest that the pre-LTP mechanism could explain the co-development of chronic pain and anxiety, that HCN blockers relieve anxiety, and that HCNs could be future pharmacological targets.
This study represents an important advance in our understanding of synaptic mechanisms linking anxiety and pain. And, like any good study, it leads to more questions:
What other brain regions are involved? The current study looked at thalamocortical connections, but only in terms of HCN expression. There are thus remaining questions about how these pathways might be involved in the expression of pain and anxiety. Furthermore, neuroimaging studies in humans and rodents have continuously pointed to an important role of the ACC in various aspects of acute and chronic pain, as well as in pain modulation. It has also been shown that anterior cingulotomy is usually effective in relieving intractable pain. But what is also becoming rapidly clearer is that the ACC is only one region or node of multiple relevant brain networks. Further work will have to be done to look beyond the ACC.
How strong is the link between anxiety and pain in this pre-LTP mechanism, and is it really specific to pain? Pain only affected pre-LTP in the presence of anxiety, and thus it was assumed that this is a mechanism through which pain and anxiety interact. The authors state that this pre-LTP could be adding a “salience factor.” That’s an intriguing thought, and the ACC is considered a major node in the salience network (which also includes bilateral anterior insula in humans). However, since pain enhances anxiety, there remains a question of whether this really reflects an interaction or if the effect simply depends on a high level of anxiety. In other words, does something other than pain that also increases anxiety also affect pre-LTP in the same way that pain did?
How well will this finding translate to human chronic pain conditions? The anxiety exposure in this study was brief, and the chronic pain models only persisted for a few days. The timescales in the current study were between two and eight days, which might not be relevant for human conditions, which are thought to take weeks, months, or even years for the full effects of the disease to be established. These chronic pain models of such short duration might be better conceptualized as prolonged acute pain. Also, it is unclear if the pre-LTP effect that was dependent on chronic pain was actually more related to pain intensity—i.e., that the chronic pain models were simply more painful than the acute stimuli. Future work including models of long-term anxiety and pain will be needed to understand how the current findings might relate to chronic pain in humans.