Table 1.
List of primary and secondary antibodies (Abs) used in immunohistochemistry and immunoconfocal morphometry.
Fig 1.
A subpopulation of NG2+ OPCs resides on microvessels.
a-c The processes of NG2+juxtavascularOPCs (JV, arrows) are in contact with the vessel wall (V). A In particular, there is a JV OPC process crossing a GFAP+endfoot through a small hole (arrowhead). B Two NG2+ perivascular OPCs (PV; arrows) apposed on the abluminal side of microvessels (V) with their bodies and processes projecting around the vessel wall and in the surrounding neuropil. c A representative parenchymal OPC (Pa, arrowhead) virtually unassociated with vessels because none of its processes reaches vessel wall (V) in this image stack (20 μm-thick) is together with two JV OPCs (JV, arrows). Scale bars, 30 μm.
Fig 2.
PDGFRα/CD31-based confocal morphometry identifies vessel-associated OPCs.
a-d Representative confocal images (stacks of xy planes on the z-axis; projection images) of cerebral cortex sections double-immunolabelled for CD31, a marker for endothelial cells, and PDGFRα, used to identify and count juxtavascular (JV), perivascular (PV), and parenchymal (Pa) OPCs. e, f Representative xy single planes and their respective xz (bottom) and yz (right) planes, used to disclose single OPC/endothelial cell contacts (arrowheads). g, h Bar charts show the ratio between JV OPC and Pa OPC volumetric density (g) and the ratio between PV OPC and Pa OPC volumetric density (h); both the parameters demonstrate a significant increase of JV and PV OPCs in EAE WT mice at 20 dpi; no significant differences are observed in the JV OPC density between naïve and EAE NG2KO mice. Data for both panels g and h are presented as mean ± SD; *p< 0.05, §p< 0.01, #p< 0.001. n = 5. EAE-affected WT mice (at 20 dpi:clinical score range = 2.0–3.5, mean number of counted JV OPCs 270 per mouse. At 40 dpi: clinical score range = 2.0–3.0 mean number of counted JV OPCs = 230 per mouse), EAE-affected NG2KO mice (at 20 dpi: clinical score range = 1.5–2.5 mean number of counted JV OPCs = 108 per mouse. At 40 dpi: clinical score range = 1.0–2.0 mean number of counted JV OPCs = 203 per mouse). Nuclear counterstaining with TO-PRO-3. Scale bars, a-d 50 μm and e, f 10 μm.
Fig 3.
Increase in doublet OPCs in EAE suggests the proliferation of vascular OPCs.
a Representative confocal image from an EAE-affected WT mouse (clinical score = 2.5) of multiple immunostaining for NG2, PDGFRα and PDGFRβ shows doublet OPCs (D-OPC) reactive for NG2 and PDGFRα, which also contact the microvessel wall revealed by PDGFRβ+pericytes (V), together with NG2+/PDGFRα+processes of OPCs (OPC). Zoomed insets (3x) show details of the vessel-contacting D-OPC (right zoom), a parenchymal D-OPC (left lower zoom), and a parenchymal OPC (left upper zoom). b The bar chart shows that the total value of JV and PV D-OPCs, quantified by confocal morphometry on brain sections immunolabelled with anti-PDGFRα antibody, increases in EAE-affected WT mice at 20 dpi compared with all the other experimental groups. Data are presented as mean ± SD cell number/106 μm3; n = 5; *p<0.05, §p< 0.01, #p<0.001. EAE WT mice (at 20 dpi: clinical score range = 2.0–3.5, mean number of counted JV doublet OPCs = 25 per mouse. At 40 dpi: clinical score range = 2.0–3.0, mean number of counted JV doublet OPCs = 12 per mouse), EAE NG2KO (at 20 dpi: clinical score range = 1.5–2.5, mean number of counted JV doublet OPCs = 8 per mouse. At 40 dpi: clinical score range = 1.0–2.0, mean number of counted JV doublet OPCs = 16 per mouse). Nuclear counterstaining with TO-PRO-3. Scale bar, 50 μm.
Table 2.
Density of JV, PV, and Pa OPCs at P61 and adulthood2.
Table 3.
Density of JV+PV doublet OPCs and Pa doublet OPCs at P61 and adulthood2.
Fig 4.
CD13/NG2-based morphometry detects activated pericytes in EAE.
Representative confocal images of cerebral cortex sections double-immunolabelled for CD13 and NG2 (a-c, e-k) and for PDGFRβ and NG2 (d, h, l) to identify and quantify NG2+ activated pericytes in naïve (a-d) and EAE-affected (e-l) WT mice. A high number of hypertrophic, JV OPCs, which overexpress NG2 together with numerous CD13+/NG2+ activated pericytes are recognizable in cortical layers from 2 to 6 in EAE-affected (e-j), compared with naïve (a, b) WT mice. High magnification of single optical planes on ‘z’ axis (_z24 and _z04, _z21 in naïve and EAE-affected WT mice, respectively) localizes NG2 molecules on the plasma membrane around the cell nucleus, where the proteoglycan in part co-localizes with CD13 (c, g, k). PDGFRβ+/NG2+ pericytes are recognizible in cortical microvessels, NG2 molecules preferentially localize in cell bodies, whereas PDGFRβ in pericyte processes (d, h, l). m The bar chart shows a significant increase of activated pericytes in EAE-affected WT mice at 20 dpi, measured as the percentage of NG2+ pericytes over the total number of CD13+ pericytes ($p = 0.005; n = 4; unpaired Student’s t-test). n The percentageof NG2+ pericytes over the total number of PDGFRβ+ pericytes is similar in naïve and EAE-affected WT mice (p = 0.78; n = 4; unpaired Student’s t-test). EAE-affected WT mice (at 20 dpi: clinical score range = 2.0–3.5, mean number of counted PDGFRβ+ pericytes = 729 per mouse), EAE-affected NG2KO (at 20 dpi: clinical score range = 1.5–2.5, mean number of counted PDGFRβ+ pericytes = 459 per mouse). Nuclear counterstaining with TO-PRO-3. Scale bars, a-l 15 μm.
Fig 5.
Comparison between the mean linear density of PDGFRβ+ pericytes and PDGFRα+ JV and PV OPCs.
The PDGFRβ+ pericytes linear density (checkerboard pattern) and the PDGFRα+ JV and PV OPC linear density (smooth pattern), reported as number of identified cells on cumulative vessel length (multiple of 100 μm), are presented as mean ± SD. Direct comparison of linear density values calculated for pericytes and OPCs in WT, NG2KO, EAE-affected WT(20 dpi) and EAE-affected NG2KO (20 dpi) shows a consensual, significant increase at EAE WT (20 dpi). *p< 0.05, #p< 0.0001. n = 5. EAE-affected WT mice (at 20 dpi: clinical score range = 2.0–3.5, mean number of counted PDGFRβ+ pericytes = 729 and PDGFRα+JV and PV OPCs = 312 per mouse), EAE-affectedNG2KO (at 20 dpi: clinical score range = 1.5–2.5, mean number of counted PDGFRβ+ pericytes = 459 and JV and PV OPCs = 136 per mouse).
Fig 6.
Endothelial tight junctions are maintained in EAE-affected NG2KO mice.
a-l Representative confocal images of cerebral cortex sections immunolabelled for claudin-5 and occludin. a, b, e, f The cerebral cortex microvessels show continuous staining patterns for both claudin-5 and occludin in P6 WT and adult naïve WT mice. c, d, g, h Claudin-5 and occludin appear irregularly organized in P6 NG2KO mice, with these modified patterns persisting in adult, naïve NG2KO mice; note the typical chain-like claudin-5 pattern in (g). i, j In the cortex microvessels of EAE WT mice, claudin-5 and occludin are lost along junctional tracts, whereas in EAE-affected NG2KO mice TJs proteins staining appears reinforced (k, l) and occludin acquires a ‘perforated ribbon-like’ configuration, with full and empty tracts. Nuclear counterstaining with TO-PRO-3. Scale bars, 10 μm.
Fig 7.
Barrier tightness is maintained in EAE-affected NG2KO mice.
a-d Representative confocal images from mice injected with FITC-Dextran 70 kDa for BBB permeability assessment, then immunolabelled for PDGFRα to show vessel/OPC association. a, b The exogenous tracer stains the vessel lumen in BBB cerebrocortical microvessels of both naïve WT and naïve NG2KO mice: note in (a) a transversely cut microvessel with FITC-dextran in the vessel lumen (arrow) and in (b) a minimal leakage around a small venule (arrows). c, d The deep cerebral cortex microvessels (at the boundary with the subcortical white matter) of EAE-affected WT mice appear permeable to the exogenous tracer that forms a fluorescent halo in the surrounding neuropil (c), while little signs of increased permeability are shown by cortex microvessels in EAE-affected NG2KO mice (d). E The bar chart shows the percentage of JV OPCs associated with leaky microvessels, a value that is significantly increased in EAE-affected WT mice (20 dpi) compared with naïve WT mice. No significant differences are observed between naïve NG2KO and EAE NG2KO mice (mean ± SD; n = 5;*p <0.05). FITC-dextran injected EAE WT mice (20 dpi clinical score range: 1.5–2.5, mean number of counted JV OPCs contacting leaky vessel: 14 per mouse), FITC-dextran injected EAE NG2KO (20 dpi clinical score range: 1.5–2.5, mean number of counted JV OPCs contacting leaky vessel: 5 per mouse). Nuclear counterstaining with TO-PRO-3. Scale bars, 50 μm.
Table 4.
JV+PV OPCs associated with non-leaky vs leaky microvessels.
Fig 8.
NG2KO mice show reduced laminin and collagens VI and IV in the vessel basal lamina.
a-l Representative confocal images of cerebral cortex sections immunolabelled for laminin, collagen VI and IV. Compared with naïve WT mice (a, d, g) the staining of the examined VBL molecular components appears reduced in both naïve and EAE-affected NG2KO mice (b, c, e, f, h, i). j-l The differences in laminin immunostaining between naïve WT and naïve or EAE-affected NG2KO mice are better seen at a higher magnification, which also served for confocal morphometry. Nuclear counterstaining with TO-PRO-3. Scale bars, a-i 30 μm and j-l 10 μm.
Fig 9.
VEGF-A and VEGFR2 as candidate pathway in NVU OPC recruitment.
a-i Representative confocal images of cerebral cortex sections from EAE-affected WT mice, immunolabelled for growth factors and their receptors: VEGF-A and its receptor VEGFR2 (a-c), FGF2 and its receptor FGFR1 (d-f), PDGF-AA/PDGFRα (g, h) and TGFβ (i), in combination with GFAP, PDGFRα, and PDGFRβ, as cell-specific markers. a VEGF-A co-localizes with GFAP in astrocyte endfeet (arrowheads) on vessel wall (V). b A typical PDGFRα+ JV OPC (arrow) and PDGFRβ+pericytes (asterisks) do not show VEGF-A reactivity; note VEGF-A+ astrocyte-like endfeet (arrowheads). c VEGFR2 co-localize with PDGFRα+on a JV OPC (arrowhead). d, e FGF2 is localized on endothelial cells and stains neither PDGFRβ+ pericytes (d) nor PDGFRα+ OPCs (e). f FGFR1 is localized on endothelial cells but not on pericytes or OPCs. g, h PDGF-AA stains endothelial cells of cortex microvessels surrounded by PDGFRα+ JV and PV OPCs (g) and by PDGFRβ+ pericytes (h). i The endothelial cells of a cortex microvessel contacted by PDGFRα+ processes (arrows) are immunoreactive for TGFβ. Nuclear counterstaining with TO-PRO-3. V, vessels. Scale bars, a-d and f-i 30 μm, e 15μm.
Fig 10.
Expression of growth factors on isolated brain microvessels and cerebral cortex sections.
a-c Representative confocal images of cerebral cortex isolated microvessels from EAE-affected mice, immunolabelled for CD31 (a, b) and for CD31/GFAP (c). a, b CD31 identifies the endothelial cells and appears localized at the junctional endothelial contacts (b, arrowheads). c GFAP+ astrocyte endfeet are recognizable in the close vicinity of the CD31-stained vessel wall. d-f Representative confocal images of cerebral cortex of EAE-affected WT mice processed for the detection of GFAP antibody/Vegfa mRNA and PDGFRβ antibody/Vegfr2 mRNA by dual RNAscope IHC/ISH. d GFAP+ perivascular astrocyte close to the vessel wall (V), and (e) perivascular astrocyte processes on a microvessel (V) show red fluorescent puncta corresponding to Vegfa ISH signal (arrows); note fluorescent puncta also in nuclei of neuron-like cells (arrowheads); f PDFGRα+ perivascular (PV), juxtavascular (JV) and parenchymal (Pa) OPCs show Vegfr2 ISH signal in the cell body (arrows). g, h Phase contrast microscopy images of the pelleted microvessels. i Real-time PCR analysis of mRNAs expression of Fgf2, Pdgfa and Tgfb in cerebral cortex of naïve and EAE-affected WT mice. The analysis by real-time PCR shows a significant increase in Fgf2 and Pdgfa levels in the isolated vessels of cerebral cortex from EAE-affected compared with WT mice. Instead, Tgfb expression shows no statistically significant difference. Nuclear counterstaining with TO-PRO-3 (a-c) and Sytox Green (d-f). Scale bars, a-f 25 μm and g, h 20 μm.