Comorbidity between temporal lobe epilepsy and depression: a [18F]MPPF PET study

Abstract

Brain and brainstem changes of serotoninergic 5-hydroxytryptophan (5-HT)(1A) receptor density have been reported in patients with major depressive disorder as well as in patients with temporal lobe epilepsy (TLE), using PET and the selective antagonist radiotracers [(11)C]WAY-100635 or [(18)F]FC-WAY. We used a distinct 5-HT(1A) antagonist, [(18)F]MPPF, whose binding potential depends on both receptor density and extracellular serotonin concentration, in 24 patients with drug-resistant TLE and MRI evidence of hippocampal sclerosis but without prior antidepressant exposure. Their Beck Depression Inventory (BDI-2) score ranged from 0 to 34, with nine patients having a score >11. We used a simplified reference tissue model, statistical parametric mapping and anatomical regions of interest (ROIs) to correlate parametric images of [(18)F]MPPF BP with the total BDI score and its four subclasses. The total BDI score, as well as symptoms of psychomotor anhedonia and negative cognition, correlated positively with [(18)F]MPPF BP in the raphe nuclei and in the insula contralateral to seizure onset, whereas somatic symptoms correlated positively with [(18)F]MPPF binding potential in the hippocampal/parahippocampal region ipsilateral to seizure onset, the left mid-cingulate gyrus and the inferior dorsolateral frontal cortex, bilaterally. We confirm an association of depressive symptoms in TLE patients with changes of the central serotoninergic pathways, in particular within the raphe nuclei, insula, cingulate gyrus and epileptogenic hippocampus. These changes are likely to reflect lower extracellular serotonin concentration in more depressed patients, with an upregulation of receptors a less likely alternative.

Fig. 1

(A) Illustration of the explicit mask of the grey matter used in this study (axial slices at the level of the hippocampus and the insula). R = right; L = left. (B) Average [¹⁸ F]MPPF-BP_ND PET volumes of our 24 patients, with the volume of the 8 patients with a left TLE being flipped so that all epileptogenic temporal lobes are right sided. A clear cut asymmetry is observed over the temporolimbic regions and the insula, ipsilateral to epileptogenic temporal lobe. R = right; L = left.

Fig. 2

(A) Positive correlation between total BDI score and MPPF BP_ND values of unflipped PET data. Height threshold P uncorrected at the voxel level P < 0.01, P corrected at the cluster level after correction for multiple comparisons P = 0.006. Only clusters showing significant correlation after correction for multiple comparisons are displayed. The blue cross shows the significant cluster localized in the left insula. Colour bar, t-scores. (B) Regression plot of the MPPF BP_ND values in the peak voxel of the significant cluster in the left insula plotted against the BDI scores.

Fig. 3

(A) Positive correlation between symptoms of negative cognition and MPPF BP_ND values of unflipped PET data. Height threshold P uncorrected at the voxel level P < 0.01, P corrected at the cluster level after correction for multiple comparisons <0.001. Only clusters showing significant correlation after correction for multiple comparisons are displayed. The blue cross shows the significant cluster localized in the left insula. Colour bar, t-scores. (B) Regression plot of the MPPF BP_ND values in the peak voxel of the significant cluster in the left insula plotted against the negative cognition.

Fig. 4

(A) Positive correlation between somatic symptoms and MPPF BP_ND values of unflipped PET data. Height threshold P uncorrected at the voxel level P < 0.01, P corrected at the cluster level after correction for multiple comparisons ≤0.002. Only clusters showing significant correlation after correction for multiple comparisons are displayed. The blue cross shows the significant cluster localized in the left cingulate gyrus. Colour bar, t-scores. (B) Positive correlation between somatic symptoms and MPPF BP_ND values of flipped PET data. Height threshold P uncorrected at the voxel level P < 0.01, P corrected at the cluster level after correction for multiple comparisons ≤0.002. Only clusters showing significant correlation after correction for multiple comparisons are displayed. The blue cross shows the significant cluster localized in the hippocampus ipsilateral to seizure onset. Colour bar, t-scores. (C) Regression plot of the MPPF BP_ND values in the peak voxel of the significant cluster in the hippocampus ipsilateral to seizure onset plotted against the somatic symptoms.

Fig. 5

(A) Positive correlation between symptoms of psychomotor-anhedonia symptoms and MPPF BP_ND values of flipped PET data. Height threshold P uncorrected at the voxel level P < 0.01, P corrected at the cluster level after correction for multiple comparisons <0.005. Only clusters showing a significant correlation after correction for multiple comparison are displayed in this figure. The blue cross shows the significant cluster localized in the insula contralateral to seizure onset. Colour bar, t-scores. (B) Regression plot of the MPPF BP_ND values in the peak voxel of the significant cluster in the insula contralateral to seizure onset plotted against the psychomotor-anhedonia symptoms.