The Ups and Downs of Endocannabinoid Signaling in Chronic Pain-Induced Depression

Individuals with chronic pain often experience depression, resulting in greater disability and a worse prognosis than either condition alone (Goesling et al., 2013). The prevalence of depression among individuals experiencing neuropathic pain, one type of chronic pain caused by injury of the nervous system, is estimated at 65% (Cherif et al., 2020). While there have been advances in identifying the neural circuits involved in both depression and the affective dimension of pain, it has been difficult to reconcile how these systems interact over time to produce chronic-pain-induced depression (Goesling et al., 2013). Understanding the relationship between chronic pain and depression, their temporal dynamics, and the mechanisms driving this relationship are critical steps toward the development of effective interventions for pain-induced depression.

A prevalent model for studying the neurobiological effects of chronic pain is spared nerve injury (SNI), a well-characterized neuropathic pain paradigm in which two of the terminal branches of the sciatic nerve are cut in a rodent's leg (Decosterd and Woolf, 2000; Pertin et al., 2012). SNI produces long-lasting sensitization of peripheral pain neurons (Baliki et al., 2006), mechanical and thermal hyperalgesia (Decosterd and Woolf, 2000; Pertin et al., 2012), and a delayed increase in depression-like behaviors (Xu et al., 2017). Given that SNI results in increased nociceptive input to cortical structures implicated in affective processing, including the mPFC (Ji and Woolf, 2001), it is an attractive model for studying the chronic pain-depression comorbidity.

Recently, Mecca et al. (2021) tested several hypotheses about why rates of depression are so high in patients with chronic pain. Because chronic stress can induce depression (Hammen, 2005), Mecca et al. (2021) first considered the possibility that chronic pain acts as a stressor that leads to depression. To investigate this, they induced SNI in rats and weighed the adrenal and thymus glands, two organs critical to stress response. There were no significant differences between SNI animals and controls, indicating it is unlikely that dysregulation of the stress response underlies the onset of depression. The authors then considered the possibility that concurrent pain was the source of depressive-like behaviors. Treatment with an analgesic for 2 weeks after SNI reversed hyperalgesia and depression-like behaviors, but acute administration of an analgesic did not, indicating that the presence of long-term pain, not concurrent pain, contributes to the development of depression. These results suggest that there is a more complex interaction between the long-term experience of pain and the delayed onset of depression-related behavior. To elucidate some of the mechanisms underlying this interaction, Mecca et al. (2021) used a combination of brain slice experiments, in vivo experiments, and molecular techniques.

To understand the role of pain processing in the delayed development of depression, the authors examined activity levels in the mPFC, a region implicated in pain integration (Ji and Woolf, 2001; Baliki et al., 2006). How mPFC activity changes with the transition from acute to chronic pain had not been well characterized previously. Using an implantable 16-channel multielectrode array, Mecca et al. (2021) found that, relative to sham-operated animals, the firing rate of mPFC pyramidal neurons in freely moving SNI rats was higher at 3 d but lower at 35 d after surgery. These results indicate that the mPFC is initially hyperactive after injury but hypoactive during sustained neuropathic pain. To investigate the extent to which excitatory and inhibitory neural activity contributed to this biphasic change in mPFC activity, the authors used the expression of the immediate early gene c-fos as an indirect measure of neural activity and costained with selective markers of glutamatergic and GABAergic neurons. SNI rats showed a higher percentage of active glutamatergic neurons in the mPFC 3 d after injury but no difference in GABAergic activity at any time point, suggesting that injury changes the excitatory/inhibitory balance in the mPFC.

To begin to understand this dynamic shift in the excitatory/inhibitory balance in the mPFC, Mecca et al. (2021) sought to reveal how synaptic activity was altered. To address this, the authors recorded evoked excitatory postsynaptic currents (eEPSCs) or evoked inhibitory postsynaptic currents (eIPSCs) in mPFC slices. At 3 d following surgery, only the eEPSC was elevated in the SNI group compared with shams, suggesting that increased evoked excitatory synaptic transmission could contribute to the transient hyperactivity of the mPFC. Conversely, at 35 d, only the eIPSC was higher in the SNI group, suggesting that increased evoked inhibitory synaptic transmission could contribute to hypoactivity as neuropathic pain becomes chronic. Notably, the frequency of spontaneous miniature IPSC was greater in the SNI group compared with the sham group 35 d after injury. Given that changes in spontaneous miniature IPSC frequency reflect changes in resting inhibitory presynaptic neurotransmitter release (del Castillo and Katz, 1954), this finding indicates that increased inhibitory presynaptic neurotransmitter release contributes to the late rise in inhibitory synaptic transmission and, in turn, mPFC hypoactivity.

Next, Mecca et al. (2021) asked how alterations in synaptic transmission in the mPFC may be associated with the development of depression-like behaviors. Previous studies have implicated the endocannabinoid (eCB) system in the pain-depression comorbidity (Hu et al., 2009; Vinod et al., 2012; Porta et al., 2015). eCBs in the CNS, including 2-arachidonoylglycerol (2-AG), modulate neural activity by activating Type 1 and Type 2 cannabinoid receptors (CB1Rs/CB2Rs) (Lutz, 2020). Notably, CB1Rs are primarily located in inhibitory interneurons in the mPFC (Hill et al., 2011), meaning that activation of CB1Rs suppresses GABA release and decreases inhibitory input to the mPFC. Given that eCB synthesis is activity-dependent (Farrell et al., 2021), the authors hypothesized that increased thalamic input to the mPFC after injury can induce eCB release in the mPFC. In turn, increased eCB/CB1R-mediated disinhibition of the mPFC might contribute to transient hyperactivity after injury. Consistent with this hypothesis, tandem mass spectrometry revealed elevated 2-AG concentration in mPFC slices only within 1 week of SNI. Given that levels of 2-AG in the mPFC are correlated with the affective symptoms of chronic pain (Vinod et al., 2012; Porta et al., 2015), the authors proposed a working model wherein eCB-mediated changes in neural activity following injury modulate the development of depression-like behaviors.

Next, Mecca et al. (2021) assessed whether this initial upregulation of eCBs contributes to the development of depression in a chronic pain model. Rats that received daily injections of 2-AG directly into the mPFC for 2 weeks after a sham surgery demonstrated elevated depression-like behaviors at 28 d. In contrast, administration of a CB1R antagonist to the mPFC for 3 weeks after surgery blocked the development of depression-like behaviors in SNI rats. These results suggest that long-term activation of CB1Rs in the mPFC is necessary and sufficient to induce the development of depression during chronic pain states.

Considering the critical role of long-term activation of CB1Rs in the mPFC, Mecca et al. (2021) asked how eCB signaling might account for the biphasic shift in mPFC activity and, in turn, the development of pain-induced depression. Systemic administration of a CB1R antagonist for 3 weeks after SNI prevented hypoactivity of the mPFC, indicating that interrupting long-term activation of CB1Rs after injury prevents mPFC hypoactivity. Given that inhibitory presynaptic neurotransmitter release was increased at 35 d after SNI, the authors hypothesized that CB1R-mediated disinhibition of the mPFC becomes functionally impaired in the weeks after SNI, leading to mPFC hypoactivity. Notably, SNI and sham-operated rats showed comparable 2-AG concentration in the mPFC after 1 week after surgery, meaning that changes in eCB release do not account for the impaired CB1R-mediated disinhibition of the mPFC. Instead, previous studies have demonstrated that prolonged CB1R activation causes receptor desensitization and/or internalization (Wu et al., 2008; Lutz, 2020), either of which could account for CB1R loss of function and, in turn, elevated inhibitory input to the mPFC during chronic pain states.

To test for the possibility that loss of function of CB1Rs is involved in the shift to mPFC hypoactivity during chronic pain states, Mecca et al. (2021) recorded from mPFC slices in the presence of a CB1R agonist. The agonist suppressed eIPSC amplitude in the SNI and control groups 3 d after surgery but had no effect at 35 d after SNI. The effects of the CB1R agonist on eIPSC amplitude were mimicked in the sham-operated group by injecting 2-AG into the mPFC for 14 d, indicating that prolonged CB1R activation may lead to impaired disinhibition of the mPFC. The authors further establish that SNI results in a loss of function of CB1Rs by demonstrating that depolarization-induced suppression of inhibition (DSI) is impaired 35 d after SNI. DSI is a type of short-term plasticity that occurs following bouts of strong activation or depolarization. In the mPFC, DSI is mediated by the release of 2-AG, and in sham rats, the induction of DSI suppressed eIPSC amplitude by ∼50%, whereas the SNI group showed <20% suppression. To isolate the role of CB1R-expressing neurons from that of other inhibitory interneurons, the authors compared the input–output curves of eIPSC before and 20 min after perfusion of the CB1R agonist into mPFC slices at 35 d after surgery. The resulting agonist-sensitive eIPSC was significantly decreased in the SNI group relative to shams, indicating that the agonist was less effective at inhibiting GABA release 35 d after SNI because of functional loss of CB1Rs. Based on these results, Mecca et al. (2021) proposed that the development of chronic pain-induced depression relies on the activity-dependent loss of eCB/CB1R signaling in the mPFC. Specifically, acute pain increases excitatory input, which induces 2-AG release, CB1R-mediated suppression of inhibitory input to the mPFC, and mPFC hyperactivity. Over time, CB1R-dependent disinhibition of the mPFC is lost, leading to mPFC hypoactivity and the manifestation of depression-like behaviors.

Although Mecca et al. (2021) propose that loss of function of CB1Rs contributes to mPFC hypoactivity during chronic pain, a minor consideration on this study is whether the results are replicable in vivo. While experiments using brain slices indicate that chronic pain impairs CB1R-mediated inhibition of the mPFC, a “rescue” experiment presented in this article provides some conflicting results. The authors report that intravenous injection of a CB1R agonist 35 d after SNI was able to reverse hypoactivity in mPFC based on in vivo electrophysiological recordings, suggesting that SNI does not necessarily cause CB1R desensitization and/or internalization. While this result is inconsistent with the other data presented, it may be explained by previous studies which show that CB1R agonists activate G-proteins in multiple brain regions in the absence of CB1Rs, raising the possibility that CB2Rs or other unknown eCB receptors expressed on excitatory afferents to the mPFC might contribute to the discrepancy between slice and in vivo recordings (Breivogel et al., 2001; Pertwee et al., 2010). Additionally, the authors did not directly compare in vivo neural activity of the mPFC in SNI rats that received the CB1R agonist with sham-operated rats; thus, it remains unclear whether the agonist returned mPFC activity to control levels.

The work by Mecca et al. (2021) provides insight into one of the neural substrates of chronic pain-induced depression. While previous studies have demonstrated an association between the eCB system and the pain-depression comorbidity (Hu et al., 2009; Vinod et al., 2012; Porta et al., 2015), Mecca et al. (2021) have localized alterations in eCB signaling to the frontal cortex. The finding that mPFC activity is dynamically altered via the eCB system after pain onset is significant, as it suggests that pain-induced depression could be treated by targeting eCB signaling differentially at pain onset or after chronic pain.

Importantly, this study is not the first to identify a neural structure associated with comorbid depression and chronic pain. The dorsal raphe nucleus, lateral habenula, central amygdala, anterior and posterior regions of the BLA, ACC, and NAc have all been shown to independently modulate depressive behaviors during chronic pain (Zhou et al., 2019; Serafini et al., 2020). Many of these structures comprise the mesolimbic dopaminergic circuit, typically referred to as the reward system, and have been shown to regulate motivation in concert (Serafini et al., 2020). A recent study showed that pain inhibits the mesolimbic dopaminergic system and results in anhedonia, a decrease in motivation to perform reward-driven behaviors (Markovic et al., 2021). Given that depression is often referred to as a motivation-deficit disorder and anhedonia is a symptom of depression (Austin et al., 2001), these results suggest that it is plausible these mesolimbic structures are also jointly involved in the development of depression following pain. The mesolimbic dopaminergic system and the PFC are functionally connected (Austin et al., 2001), making it likely the mPFC and these structures are collectively critical to chronic pain-induced depression. Future studies will need to explore how the activity of these structures and their connectivity is altered during the transition from acute to chronic pain, when the risk of depression is high.

Notably, while much research focuses on pain-induced depression, clinical evidence suggests that depression can also be the antecedent for pain. Patients with depression often experience unexplained comorbid pain symptoms (Bair et al., 2003; Goesling et al., 2013) and depressive episodes have been shown to predict vulnerability to chronic musculoskeletal pain (Bair et al., 2003). While Mecca et al. (2021) identify a dynamic neuroanatomical change involved in chronic pain-induced depression, these previous findings suggest that the path from pain to depression is not unidirectional.

In conclusion, the study by Mecca et al. (2021) highlights the potential importance of the loss of function of eCB signaling in the mPFC in mediating depression responses to chronic pain. The authors' work demonstrates that the path from chronic pain to depression does not entail static alterations in neural function. Instead, pain-induced depression relies on complex upregulation and downregulation of regions, such as the mPFC. Subsequent studies will need to reconcile this eCB regulatory loss with the multiple neural circuits implicated in the pain-depression comorbidity.