Brain imaging of chronic pain : from the association with immune mechanisms to brain networks

Abstract

Musculoskeletal chronic pain encompasses the two mechanistically distinct categories of nociplastic pain and nociceptive pain. The former is classically typified by fibromyalgia whereas the latter is characteristic of rheumatoid arthritis. Both conditions are heterogeneous in their clinical manifestation with a multitude of factors potentially contributing to pain.

This thesis aims to provide a well-rounded investigation of the pain-related mechanisms in fibromyalgia and rheumatoid arthritis from a clinical and methodological perspective. The primary focus of the works in this thesis was to study the function and biochemistry of the human brain in the presence of pain through the use of brain imaging techniques such as functional magnetic resonance imaging (fMRI) and proton magnetic resonance spectroscopy (MRS). In fibromyalgia, pain-related neural activity (assessed via fMRI) and metabolite content of pain-related brain areas (assessed via proton MRS) were examined in relation to a combination of immune-related mechanisms and behavioral/clinical pain measures. Conversely, in rheumatoid arthritis, fMRI was used to explore the temporal changes in brain network organization across different spatial scales (brain communities and areas) during painful stimulation at an inflamed body site (joint) and a neutral body site (thumbnail).

In Study I, the Ala147Thr polymorphism of the gene encoding the translocator protein (TSPO, a biomarker of glial activation) was found to be involved in fundamental aspects of human pain regulation and metabolic content in thalamus and rostral anterior cingulate cortex, but not in the cerebral processing of evoked pain. Fibromyalgia patients and healthy subjects that were genetically inferred as high-affinity TSPO binders presented with a less efficient descending pain regulation including reduced conditioned pain modulation and expectancy-induced reduction of pain. The same subjects were also found to be featured by elevated thalamic glutamate concentrations and positive associations between glutamate and GABA in rostral anterior cingulate cortex. Altogether, these findings indicate that a less efficient endogenous pain modulation and brain-region specific changes in metabolite content might be ascribable to subjects determined as TSPO high-affinity binders regardless of baseline pain levels.

In Study II, fibromyalgia patients had higher levels of anti-satellite glia cells immunoglobulin G antibodies (anti-SGC IgG, proposed to be pathogenic autoantibodies in fibromyalgia) than healthy subjects. Patients with elevated anti-SGC IgG levels presented with high ongoing pain intensity and disease severity, and anti-SGC IgG levels correlated positively with these clinical measures. An inverse relation was found between the levels of these antibodies and baseline thalamic concentrations of metabolites such as scylloinositol, total choline and macromolecule 12. These findings and the fact that anti-SGC IgG levels were not found to relate to pressure pain sensitivity and cerebral pain processing (both assessing evoked pain) support the potential clinical relevance of these antibodies in ongoing pain and propose their functional relation with the central nervous system in fibromyalgia patients.

In Study III, the degree of interaction among brain communities was generally higher during and after painful pressure on the joint in rheumatoid arthritis patients as opposed to healthy subjects. The cerebral high demand that is often associated with a constant elevated interaction of brain communities might possibly be contributing to the maintenance of pain and the fatigue seen in these patients. However, in rheumatoid arthritis patients, the network organization was not found to reconfigure differently depending on the stimulation site. This might indicate that the cerebral processing of pain in these patients is unspecific to the clinical relevance of the area being stimulated. At the level of single pain-related brain areas and not communities, six tested brain areas were not found to bring significant contribution to the temporal changes of the network architecture.

The explorative nature of Study III profited from the visualization of brain communities and the selected pain-related brain areas by means of NetPlotBrain, a tool which we developed in Python for visualizing brain networks and viewing brain anatomy.

All the works in this thesis concern efforts to increase the understanding regarding potential parameters contributing to multiple aspects of pain in two mechanistically distinct chronic pain conditions.