Extra-axial inflammatory signal and its relationship to peripheral and central immunity in depression

Abstract

Although both central and peripheral inflammation have been observed consistently in depression, the relationship between the two remains obscure. Extra-axial immune cells may play a role in mediating the connection between central and peripheral immunity. This study investigates the potential roles of calvarial bone marrow and parameningeal spaces in mediating interactions between central and peripheral immunity in depression. PET was used to measure regional TSPO expression in the skull and parameninges as a marker of inflammatory activity. This measure was correlated with brain TSPO expression and peripheral cytokine concentrations in a cohort enriched for heightened peripheral and central immunity comprising 51 individuals with depression and 25 healthy controls. The findings reveal a complex relationship between regional skull TSPO expression and both peripheral and central immunity. Facial and parietal skull bone TSPO expression showed significant associations with both peripheral and central immunity. TSPO expression in the confluence of sinuses was also linked to both central and peripheral immune markers. Group-dependent elevations in TSPO expression within the occipital skull bone marrow were also found to be significantly associated with central inflammation. Significant associations between immune activity within the skull, parameninges, parenchyma and periphery highlight the role of the skull bone marrow and venous sinuses as pivotal sites for peripheral and central immune interactions.

Keywords: inflammation; microglia; parameninges; positron emission tomography; skull; translocator protein.

Figure 1

Generation of skull mask from pseudo-CT images. (A) T₁ structural brain image. (B) Psuedo-CT (pCT) image synthesized from T₁ structural brain image. (C) pCT image with binary skull mask overlaid. Skull voxels were identified by applying a lower intensity threshold (600 Hounsfield units). Gaps and voids in the binarized image were filled to create a complete skull mask.

Figure 2

Confluence of sinuses region of interest generation process. (A) Binary mask delineating non-brain areas, overlaid on T₁ structural brain image. (B) Refinement of the mask to isolate the posterior cranial fossa. (C) Exclusion of small clusters and dilatation of retained clusters to ensure coverage of the confluence of sinuses. (D) Using PET data to identify voxels exhibiting ≥25% of the peak activity observed in the mode peak frame time, indicating a predominant presence of blood in those voxels. (E) Only voxels identified as predominantly containing blood are retained in the binary mask. (F) Removal of smaller noise clusters, refinement of mask edges through dilatation and erosion, and retention of the largest volume cluster to produce the final binary mask representing the confluence of sinuses.

Figure 3

Three-dimensional representation of skull regions of interest. See also Supplementary Fig. 1 for the three-dimensional rendering depicting specific regions of interest within the skull. Regions of interest were informed by the template provided by Hadjikhani et al. and pseudo-CT renderings derived from participants’ structural MRIs. The use of pseudo-CTs facilitated the refinement of these regions, ensuring that voxels represented bone marrow exclusively, without including adjacent dura mater. DLPFC = dorsolateral prefrontal cortex.

Figure 4

ACC SUV effect on regional extra-axial SUV. Partial regression plots presenting the relationship between ACC SUV and regional extra-axial SUV, accounting for group and logTNFα. ACC = anterior cingulate cortex; SUV = standardized uptake value.

Figure 5

LogTNFα effect on regional extra-axial SUV. Partial regression plots presenting the relationship between logTNFα and regional extra-axial SUV, accounting for group and ACC SUV. ACC = anterior cingulate cortex; SUV = standardized uptake value.

Figure 6

Group effect on occipital skull SUV. Box-and-whisker plot representing group differences in occipital skull SUV, adjusted for ACC SUV and logTNFα, with all data-points shown. The horizontal line within the boxes indicates the median, each box represents the values within the 25th–75th percentile, and whiskers show the range of SUVs. ACC = anterior cingulate cortex; DP = depressed individuals; HCv healthy controls; SUV = standardized uptake value.