Molecular, circuit, and anatomical changes in the prefrontal cortex in chronic pain

Summary:

The prefrontal cortex undergoes functional and structural reorganization in chronic pain conditions in both rodents and humans. We provide an illustrated overview of the molecular, functional, and connectivity pathology occurring in the prefrontal cortex in chronic pain states.

The brain integrates information about the intensity, quality and location of noxious inputs with other states such as attention, anxiety, fear and expectation. A key brain region where this integration occurs is the prefrontal cortex (PFC). Importantly, the PFC is also responsible for higher executive functioning [11]. While acute pain is critical for survival, chronic pain is a detrimental, pathological state that drives changes in the PFC, culminating in pain amplification and cognitive problems. Here, we summarize how chronic pain affects the PFC in patients and in preclinical rodent models and why this is an important area of research in pain neuroscience.

Decades of lesion studies in rodents have demonstrated that the medial portion of the PFC (mPFC) controls higher executive functioning [8; 14; 27; 32]. The mPFC’s subregions, the prelimbic (PrL; or dorsolateral PFC in humans) and the infralimbic (IL; ventromedial PFC in humans) cortices are highly interconnected with one another as well as with other regions such as the amygdala, hippocampus, nucleus accumbens, and striatum [20; 29; 46]. The PrL and IL are responsible for modulating goal directed behaviors by integrating thought, motivation and action to achieve a goal. As such, a cardinal sign of PFC dysfunction in chronic pain patients presents as cognitive impairment which occurs in a variety of chronic pain conditions [2; 5; 6; 9; 15–18; 30; 34; 37]. Interestingly, pain relief using currently prescribed analgesics is insufficient to reverse cognitive impairments as the deficits persist and even worsen after analgesic treatment [16; 17; 34; 37], indicating that PFC dysfunction is resistant to transient analgesia.

Another sign of PFC dysfunction in chronic pain patients is gray matter loss. Shrinkage of the frontal cortical gray matter has been identified consistently across a variety of pain conditions [3; 12; 19; 24–26; 38]. This anatomical abnormality is severe and has been equated to the cortical loss seen over 10–20 years of normal aging in a healthy individual [3]. As none of these neuroimaging studies excluded patients who were taking analgesics, it is unlikely that cortical thinning is reversed by currently prescribed pain therapeutics. In fact, to date, the only two methods shown to reverse cortical thinning are cognitive behavioral therapy [39] and effective interventional pain management [35; 36; 40]. Cortical gray matter restoration has been observed in patients with hip osteoarthritic pain which dissipated after undergoing total hip replacement surgery [35; 36] and in chronic low back pain after spinal surgery or facet joint injections [40]. There is a clear need to investigate treatment options that can target both pain and its PFC-driven comorbidities.

Importantly, restoration of frontal cortical gray matter in chronic pain indicates the disease is not neurodegenerative, suggesting that structural reorganization of resident neurons and/or glia in the PFC account for the abnormality. Indeed, PFC morphological plasticity has been identified in rodents with neuropathic pain. Layer 2/3 pyramidal neurons display increased spine density and basal dendritic branching in the PrL contralateral (right) to nerve injury [28], while the apical dendrites of layer 5 pyramidal neurons are shrunken and less complex [23]. Axon initial segments are shrunken in the bilateral IL in mice with neuropathic pain [41; 42]. Microglia also proliferate [7], activate [7], and appear to take on an M1 phenotype as proinflammatory cytokines such as interleukin-6 and interleukin-1β are markedly increased in the PFC of rodents with neuropathic pain [44].

Neuroinflammation or structural pruning are either the cause or consequence of physiological dysfunction in PFC neurons. Reductions in PFC glutamate levels have been detected in rats [23] and humans [31] with chronic pain. Correspondingly, layer 5 pyramidal neurons display a loss in spontaneous and evoked firing arising from enhanced peri-somatic inhibition by local GABAergic interneurons in the PrL [21–23]. This disruption in excitation-inhibition balance is driven by augmented monosynaptic connections from basolateral amygdala (BLA) projection neurons onto layer 5 PrL inhibitory interneurons, resulting in feed-forward inhibition [21; 22]. Interestingly, strengthened glutamatergic inputs onto PrL inhibitory interneurons is due to a loss of the Gi-coupled cannabinoid receptor 1, resulting in disinhibition of glutamatergic afferents into the PrL [21]. While detection of PFC activity changes in humans has been inconsistent [4; 13], the emerging picture is that the PFC is deactivated in chronic pain.

Additionally, rodents with chronic pain show a loss in activity of PrL L5 pyramidal neurons that signal to the periaqueductal gray (PAG) [10; 21], a region responsible for mediating endogenous analgesia. Restoration of PrL cortical activity or activation of PrL afferents to the PAG can attenuate nociceptive behaviors in rodents with neuropathic pain [21; 45]. The chronification of pain may result in part from disruption of this descending analgesic circuitry that stems from PFC deactivation. Human data shows there is a loss of fiber track density and reorganization of white matter connectivity from the PFC to the insula, basal ganglia and other regions involved in pain processing such as the anterior cingulate cortex [19; 43], indicating that PFC output may be disrupted in human chronic pain patients as well.

Transcriptome analysis of the mPFC using quantitative real‐time PCR or sequencing has also identified specific mRNA transcripts that are dysregulated in chronic pain. Rodents with neuropathic pain display an increase in brain-derived neurotrophic factor, prodynorphin, and κ-opioid receptor [33]. Interestingly, mRNA for glial fibrillary acidic protein (GFAP), a commonly-used marker of astrocytes, was shown to be down-regulated in the mPFC of mice that had neuropathic pain for 6 months, suggesting that astrocyte populations may be diminishing at these later time points [1].

The work summarized here shows that complementary human and rodent studies have led to important insight into how the PFC changes in chronic pain states. A reverse translational approach has clearly been embraced wherein human symptomology and neuroimaging is directing preclinical investigations. Although there is still much unknown, the current picture is that the PFC modulates the pain experience in critical ways and that many comorbidities of painful disease are driven by PFC changes. Continuously growing insight into this pathology has great promise for improving pain care.