Widespread decrease of cerebral vimentin-immunoreactive astrocytes in depressed suicides

Abstract

1AbstractPostmortem investigations have implicated cerebral astrocytes immunoreactive (-IR) for glial fibrillary acidic protein (GFAP) in the etiopathology of depression and suicide. However, it remains unclear whether astrocytic subpopulations IR for other astrocytic markers are similarly affected. Astrocytes IR to vimentin (VIM) display different regional densities than GFAP-IR astrocytes in the healthy brain, and so may be differently altered in depression and suicide. To investigate this, we compared the densities of GFAP-IR astrocytes and VIM-IR astrocytes in postmortem brain samples from depressed suicides and matched non-psychiatric controls in three brain regions (dorsomedial prefrontal cortex, dorsal caudate nucleus and mediodorsal thalamus). A quantitative comparison of the fine morphology of VIM-IR astrocytes was also performed in the same regions and subjects. Finally, given the close association between astrocytes and blood vessels, we also assessed densities of CD31-IR blood vessels. Like for GFAP-IR astrocytes, VIM-IR astrocyte densities were found to be globally reduced in depressed suicide relative to controls. By contrast, CD31-IR blood vessel density and VIM-IR astrocyte morphometric features in these regions were similar between groups, except in prefrontal white matter, in which vascularization was increased and astrocytes displayed fewer primary processes. By revealing a widespread reduction of cerebral VIM-IR astrocytes in cases vs. controls, these findings further implicate astrocytic dysfunctions in depression and suicide.

Implication of cerebral astrocytes in major depression: A review of fine neuroanatomical evidence in humans

“…However, GFAP strongly labels astrocyte somas, and GFAP‐IR soma volume has been used to morphologically distinguish astrocytes in ventral prefrontal white matter from fibrous astrocytes (255 ± 34 μm 3 ) and smooth astrocytes (44 ± 2 μm 3 ) (Rajkowska et al, 2018). Using a two‐dimensional assessment, we found both GFAP‐IR and VIM‐IR fibrous astrocyte soma sizes in the dorsolateral prefrontal cortex are similar to these findings in the ventral prefrontal cortex (O'Leary et al, 2020; O'Leary et al, 2021). Another study used fresh resected cortical tissue, leading to brilliant visualization of astrocytes using a combination of GFAP immunostaining and diolistic labeling, and found human protoplasmic astrocytes have 38 ± 5 processes and a mean process length of 98 ± 5 μm (Oberheim et al, 2009).…”

“…This model combines cortical VIM‐IR astrocyte morphometry with regional densities of Giemsa‐stained astrocytes to reveal that tiling would be spatially impossible for the vast majority of astrocytes in cortical regions, as neighboring astrocyte domains would frequently and extensively overlap due to a lack of physical space. By contrast, ours and other reports of GFAP‐IR astrocyte density allow for the extent of overlap expected for tiling, based on a previous study of astrocyte domains in human tissue (Oberheim et al, 2009; O'Leary et al, 2020; O'Leary et al, 2021). This model indicates that if tiling is indeed a common feature of astrocyte organization in the human brain, it could not feature for an astrocyte population much larger than the GFAP‐IR astrocytes already known to display it, as shown in studies of both human and rodent brain (Bushong et al, 2002; Bushong et al, 2004; Oberheim et al, 2009).…”

“…A previous study from our group found Golgi‐stained astrocytes in healthy individuals have a larger soma and simpler branch complexity (process numbers, nodes, and terminals) than VIM‐IR astrocytes, however, fibrous astrocytes had a comparable total process length (O'Leary et al, 2021; Torres‐Platas et al, 2011). We suspect the discrepancies in soma diameter and branch complexity for Golgi‐stained astrocytes relative to other studies is due to distortion from overimpregnation, and that process length differences may be a regional feature seen in proisocortex but not isocortex.…”

“…Postmortem studies have generally reported that healthy individuals have more GFAP‐IR astrocytes in cortical white matter than in cortical grey matter, and vice‐versa for VIM‐IR astrocytes in the same subject and regions (McNeal et al, 2016; O'Leary et al, 2020; O'Leary et al, 2021). Remarkably, VIM‐IR astrocytes are almost absent in the mediodorsal thalamus, a region where GFAP‐IR astrocytes are especially abundant in samples from the same subjects (O'Leary et al, 2020; O'Leary et al, 2021).…”

“…Three lines of evidence suggest reduced astrocyte densities in depression are associated with deficits in the number or function of astrocytes. First, several research groups have observed reduced astrocyte densities in multiple brain regions of individuals with MDD (Altshuler et al, 2010; Bernstein et al, 2015; Cobb et al, 2016; Gos et al, 2013; Hamidi, Drevets, & Price, 2004; O'Leary et al, 2021; Rajkowska et al, 2018). Second, there is also a decrease in the area fraction of immunostaining for astrocyte markers—which is an indirect measure for a reduced number of astrocytes—in individuals with MDD (Miguel‐Hidalgo et al, 2000; Webster et al, 2001; Si, Miguel‐Hidalgo, O'Dwyer, Stockmeier, & Rajkowska, 2004; Miguel‐Hidalgo et al, 2014).…”