Coexistence of cerebral hypometabolism and neuroinflammation in the thalamo-limbic-brainstem region in young women with functional somatic syndrome

Functional somatic syndrome (FSS) is a disorder characterized by clusters of medically unexplained symptoms even after extensive examinations [1]. Most well-known FSS conditions are fibromyalgia, chronic fatigue syndrome (CFS), and irritable bowel syndrome, which often exhibit several common symptoms and share similar treatments [1, 2]. Recent studies from Japan and Denmark have reported that some women exhibit various long-lasting physical, psychological, cognitive, and behavioral symptoms following human papillomavirus (HPV) vaccination (a prophylactic measure against future development of cervical cancer) [3, 4]. In Japan, the post-vaccination symptoms have finally regarded as various somatic symptoms without the causal relationship with HPV vaccination [5]; however, the pathophysiology of FSS itself remains elusive.

FSS symptoms, such as widespread pain, psychiatric symptoms, and cognitive dysfunction, suggest a disturbance of the somatosensory system based on the thalamus, a relay region for almost all sensory information, and the limbic system, a part of the brain involved in memory, emotion, and the default-mode network [6]. Thus, we hypothesized that the etiology of FSS may be associated with the disruption of the brain milieu in these regions.

Positron emission tomography (PET) with [¹⁸F]FDG has been extensively used to detect cerebral glucose metabolic alterations, reflecting cerebral function (neuronal and synaptic activity) and/or cellular abnormalities [7]. Recent studies indicated that a disturbance of the brain milieu accompanies neuroinflammation [8, 9], which can be visualized by PET with radiotracers for the 18 kDa translocator protein (TSPO) developed in glial cells such as microglia and astrocytes [8, 10‐12]. Among radiotracers for TSPO, [¹¹C]DPA713 is considered a second-generation ligand for the TSPO. Several studies on FSS implicated changes at the immunomolecular level, suggesting the involvement of neuroinflammation in the pathophysiology of FSS [13‐15]. While previous PET studies revealed regional cerebral abnormalities in fibromyalgia and CFS [16‐18], there are no pathophysiological studies on FSS regarding symptom-related brain regions with simultaneous cerebral glucose metabolism and neuroinflammation.

Therefore, we aimed to identify the brain regions showing altered cerebral metabolism and neuroinflammation in patients with FSS and to correlate the results of PET with clinical data to better understand the pathophysiology of FSS.

To the best of our knowledge, this is the first study to clarify the concurrent changes in cerebral glucose hypometabolism and neuroinflammation in patients with FSS. In this study, we showed that young women with FSS had significantly decreased [¹⁸F]FDG SUVR and increased [¹¹C]DPA-713 BP_ND in unique brain regions covering the thalamus, mesial temporal region, and tegmentum of the brainstem. We also showed that [¹⁸F]FDG SUVR and [¹¹C]DPA-713 BP_ND values in these regions were associated with clinical scores in WAIS-III and WMS-R tests. Thus, cerebral glucose hypometabolism and elevated neuroinflammatory responses in the confined brain regions together with their clinical relevance reflect the pathophysiology of FSS.

Cerebral glucose metabolism is related to the cerebral blood flow (CBF) [23], and CBF measurement by single-photon emission computed tomography (SPECT) might suggest the pathological mechanism of FSS. Indeed, our previous study using ¹²³IMP-SPECT in patients with various symptoms following HPV vaccination showed significant reduction in CBF in the regions associated with the right limbic system [24]. Our present study identified unique brain regions with glucose hypometabolism, some of which matched the regions with decreased CBF, namely the mesial temporal region and the thalamus [24].

Several other PET studies have reported changes in cerebral glucose metabolism in FSS spectrum disorders. Patients with CFS manifest a significant reduction in cerebral glucose metabolism in the medial–frontal cortex and brainstem [16] or a small reduction nonspecifically in the brain [25]. In patients with fibromyalgia, an elevation of glucose metabolism in the limbic system is observed post-treatment [17]. A series of studies on glucose metabolism in these FSS spectrum disorders indicated that the limbic region might be a core region for the development of FSS symptoms.

This one-stop measurement of glucose metabolism may be used to estimate the steady-state level of cerebral function. Combining metabolic steady-state and inflammatory state measurements could allow additional information on the brain milieu. Beside the limbic system, we detected significant glucose hypometabolism in the dorsal part of the brainstem and thalamus. As discussed below, the current cerebral hypometabolism regions in FSS might be linked with increased neuroinflammation because of co-occurrence with the decreased [¹⁸F]FDG and increased [¹¹C]DPA713, suggesting that glucose hypometabolism could be associated with neuroinflammation. At least in the current study, brain volume was not responsible for this hypometabolism because a VBM analysis showed no difference between FSS and age-matched control groups. Thus, cerebral glucose hypometabolism in these regions might be pathophysiologically implicated.

PET imaging of TSPO is widely used to monitor neuroinflammation in neurological and psychiatric disorders [9, 12]. A recent PET study revealed significantly higher [¹¹C](R)-PK11195 BP_ND in patients with chronic fatigue syndrome than in healthy controls in the thalamus, midbrain, and pons [18]. The pattern of high [¹¹C](R)-PK11195 BP_ND in these regions is consistent with our findings. Moreover, in our study using [¹¹C]DPA713, which is more specific than [¹¹C](R)-PK11195 [22], increased [¹¹C]DPA713 binding was found in broader regions covering the mesial temporal region, thalamus, and brainstem in FSS.

One possible explanation for the presence of neuroinflammation may be related to the inflammatory cytokines in circulation. One review about patients with CFS described that circulating cytokines, such as tumor necrosis factor-α, interferon-γ, interleukin (IL)-6, and IL-1, are associated with the initiation of inflammation [26]. Patients with fibromyalgia were reported to show higher serum levels of IL-1 receptor antagonist, IL-6, and IL-8 [27]. Hence, inflammatory cytokines may affect the brain regions described above due to cytokine infiltration. Another possible mechanism of neuroinflammation may be stress-induced neuronal deactivation. Indeed, stress sensitizes peripheral as well as central cytokine responses, leading to neuroinflammation and hypothalamic–pituitary–adrenal axis activation [28]. Endogenous cytokine expression and associated cytokine receptor responses are found throughout the brain, including the circumventricular thalamic regions, brainstem nuclei, hypothalamus, and basal ganglia [28]. In any possibility, neuroinflammation in the unique brain regions is an underlying phenomenon responsible for FSS symptoms.

We found unique brain regions with cerebral glucose hypometabolism and neuroinflammation that were associated with key clinical features in patients with FSS. Significant correlations of psycho-behavioral scores with levels of [¹⁸F]FDG SUVR and [¹¹C]DPA713 BP_ND were found mainly in the thalamus and hippocampus (Table 3). Although there was a right–left lateralization in regions with significance, the levels of [¹⁸F]FDG SUVR and [¹¹C]DPA713 BP_ND in the thalamus and hippocampus on both sides correlated considerably well with psycho-behavioral scores (Table 3). This suggested that thecurrently observed lateralization is not essential in the clinico-bioindex relevance. It is known that the thalamus is important in working memory and processing speed [29], acting as a central monitor for language processing and cortical activity [30, 31]. The circumventricular regions that were highlighted in the present study are associated with neuroendocrine regulation, stress response, immunomodulation, and cognitive regulation [32]. The hippocampus is also associated with language comprehension and processing speed [33, 34]. In the brainstem, the dorsal part, including the locus coeruleus and raphe nucleus, were highlighted in the present study. These nuclei are origins of the noradrenergic and serotonergic systems, subserving autonomic and emotional control, respectively. Taken together, the neuroinflammation-induced cerebral glucose hypometabolism in these peculiar regions may lead to the development of FSS symptoms.

Interestingly, our direct comparison between [¹⁸F]FDG SUVR and [¹¹C]DPA713 BP_ND showed a positive correlation in the right hippocampus and amygdala. Since it was reported that TSPO-based PET findings in chronic, but not early, brain disease reflect a proinflammatory state [35, 36], it seems theoretically possible that lower cerebral metabolism couples with higher proinflammatory responses, i.e., [¹⁸F]FDG SUVR could negatively correlate with [¹¹C]DPA713 BP_ND. Our present result contradicts this hypothesis. Because [¹⁸F]FDG measurement does not differentiate anaerobic glycolysis, which is predominant in microglia [37], from aerobic glycolysis, the current result suggested that the higher [¹⁸F]FDG uptake accompanied greater glycolysis in the hippocampus and amygdala, which subserve memory and emotion in patients with FSS. In the present study, a significant reduction in glucose metabolism was found in patients with FSS compared with healthy subjects. Hence, one possible interpretation is that the true amount of glucose metabolism in neurons might be much lower than the observable level that contained glial glucose consumption. One solution may be the use of a new PET tracer for oxidative metabolism in mitochondria. Oxidative metabolism is higher in neurons than in glial cells. Therefore, the new PET tracer could be used to distinguish the neuron-dominant glucose metabolism from neuroinflammation-related glucose consumption [38].

The present study has several limitations. First, the etiology of FSS is unknown. The patients in the current study developed FSS symptoms after different time intervals following HPV vaccination. HPV vaccination might contribute partly to FSS: HPV-treated mice manifest altered behavioral responses and microglial activation in the hippocampus CA1 area [39]. However, our results showing a greater extent of neuroinflammation outside the hippocampus and the development of FSS symptoms irrespective of vaccination [40] suggest that HPV vaccination may not be involved in the development of FSS, rather a need to prepare for development of FSS if one has altered metabolism or neuroinflammation in the thalamo-limbo-brainstem region in the brain. Second, the rs6971 polymorphism poses an issue for second-generation TSPO tracers such as [¹¹C]DPA713 [41]. As most Asian people are moderate affinity binders with moderately homogenous pattern of polymorphism [41], our findings in Japanese women may not have been affected by the polymorphism. Third, there was a significant age difference in the [¹⁸F]FDG study between patients and controls (patients were younger than controls). While there was no significant change in brain volume in the current study from prior data of younger individuals, the previous study indicated that cerebral glucose metabolism is subject to age-related changes, especially in the dorsolateral prefrontal, inferior temporal/fusiform, and primary somatosensory cortices [42, 43]. This suggests that lower glucose metabolism found in the FSS group would have been more remarkable if a strictly age-matched control group was used. Fourth, our sample size was relatively small. Despite this limitation, we were able to show significant differences among groups under a rigorous statistical level.

Our results show that both cerebral glucose hypometabolism and neuroinflammation were concurrently present in unique brain areas covering the thalamo-limbo-brainstem region in patients with FSS. A close relationship of cognitive impairment with metabolic and inflammatory abnormality in these regions provides better understanding of the clinicopathophysiology of FSS.