Abstract

Disruption of the blood–brain barrier (BBB) is critical to initiation and perpetuation of disease in multiple sclerosis (MS). We report an interaction between oligodendroglia and vasculature in MS that distinguishes human white matter injury from normal rodent demyelinating injury. We find perivascular clustering of oligodendrocyte precursor cells (OPCs) in certain active MS lesions, representing an inability to properly detach from vessels following perivascular migration. Perivascular OPCs can themselves disrupt the BBB, interfering with astrocyte endfeet and endothelial tight junction integrity, resulting in altered vascular permeability and an associated CNS inflammation. Aberrant Wnt tone in OPCs mediates their dysfunctional vascular detachment and also leads to OPC secretion of Wif1, which interferes with Wnt ligand function on endothelial tight junction integrity. Evidence for this defective oligodendroglial–vascular interaction in MS suggests that aberrant OPC perivascular migration not only impairs their lesion recruitment but can also act as a disease perpetuator via disruption of the BBB.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Data availability

The data that support the findings of this study are available from the corresponding author upon request. R code used for the mRNAseq analysis can be found on the following github page: https://github.com/baranzini-lab/RNAseq_QuantSeq_Fancy. Raw sequence data (fastq) for the mRNAseq data are available on DASH data share (https://doi.org/10.7272/Q63N21KB).

Additional information

Journal peer review information: Nature Neuroscience thanks Ken Arai, Sarah Kucenas, and other anonymous reviewer(s) for their contribution to the peer review of this work.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Compston, A. & Coles, A. Multiple sclerosis. Lancet 372, 1502–1517 (2008).

  2. 2.

    Maggi, P. et al. The formation of inflammatory demyelinated lesions in cerebral white matter. Ann. Neurol. 76, 594–608 (2014).

  3. 3.

    Alvarez, J. I., Cayrol, R. & Prat, A. Disruption of central nervous system barriers in multiple sclerosis. Biochim. Biophys. Acta 1812, 252–264 (2011).

  4. 4.

    Minagar, A. & Alexander, J. S. Blood–brain barrier disruption in multiple sclerosis. Mult. Scler. 9, 540–549 (2003).

  5. 5.

    Obermeier, B., Daneman, R. & Ransohoff, R. M. Development, maintenance and disruption of the blood–brain barrier. Nat. Med. 19, 1584–1596 (2013).

  6. 6.

    Franklin, R. J. M. & Ffrench-Constant, C. Regenerating CNS myelin - from mechanisms to experimental medicines. Nat. Rev. Neurosci. 18, 753–769 (2017).

  7. 7.

    Gallo, V. & Deneen, B. Glial development: the crossroads of regeneration and repair in the CNS. Neuron 83, 283–308 (2014).

  8. 8.

    Chang, A., Tourtellotte, W. W., Rudick, R. & Trapp, B. D. Premyelinating oligodendrocytes in chronic lesions of multiple sclerosis. N. Engl. J. Med. 346, 165–173 (2002).

  9. 9.

    Kuhlmann, T. et al. Differentiation block of oligodendroglial progenitor cells as a cause for remyelination failure in chronic multiple sclerosis. Brain 131, 1749–1758 (2008).

  10. 10.

    Boyd, A., Zhang, H. & Williams, A. Insufficient OPC migration into demyelinated lesions is a cause of poor remyelination in MS and mouse models. Acta Neuropathol. 125, 841–859 (2013).

  11. 11.

    Petersen, M. A. et al. Fibrinogen activates BMP signaling in oligodendrocyte progenitor cells and inhibits remyelination after vascular damage. Neuron 96, 1003–1012.e7 (2017).

  12. 12.

    Fancy, S. P. et al. Parallel states of pathological Wnt signaling in neonatal brain injury and colon cancer. Nat. Neurosci. 17, 506–512 (2014).

  13. 13.

    Tsai, H. H. et al. Oligodendrocyte precursors migrate along vasculature in the developing nervous system. Science 351, 379–384 (2016).

  14. 14.

    Hsieh, J. C. et al. A new secreted protein that binds to Wnt proteins and inhibits their activities. Nature 398, 431–436 (1999).

  15. 15.

    Liebner, S. et al. Wnt/beta-catenin signaling controls development of the blood–brain barrier. J. Cell Biol. 183, 409–417 (2008).

  16. 16.

    Daneman, R. et al. Wnt/beta-catenin signaling is required for CNS, but not non-CNS, angiogenesis. Proc. Natl Acad. Sci. USA 106, 641–646 (2009).

  17. 17.

    Zhou, Y. et al. Canonical WNT signaling components in vascular development and barrier formation. J. Clin. Invest. 124, 3825–3846 (2014).

  18. 18.

    Chang, J. et al. Gpr124 is essential for blood–brain barrier integrity in central nervous system disease. Nat. Med. 23, 450–460 (2017).

  19. 19.

    Daneman, R., Zhou, L., Kebede, A. A. & Barres, B. A. Pericytes are required for blood–brain barrier integrity during embryogenesis. Nature 468, 562–566 (2010).

  20. 20.

    De La Fuente, A. G. et al. Pericytes stimulate oligodendrocyte progenitor cell differentiation during CNS remyelination. Cell Rep. 20, 1755–1764 (2017).

  21. 21.

    Phoenix, T. N. et al. Medulloblastoma genotype dictates blood brain barrier phenotype. Cancer Cell 29, 508–522 (2016).

  22. 22.

    Lengfeld, J. E. et al. Endothelial Wnt/β-catenin signaling reduces immune cell infiltration in multiple sclerosis. Proc. Natl Acad. Sci. USA 114, E1168–E1177 (2017).

  23. 23.

    Zhu, X. et al. Age-dependent fate and lineage restriction of single NG2 cells. Development 138, 745–753 (2011).

  24. 24.

    Schüller, U. et al. Acquisition of granule neuron precursor identity is a critical determinant of progenitor cell competence to form Shh-induced medulloblastoma. Cancer Cell 14, 123–134 (2008).

  25. 25.

    Rivers, L. E. et al. PDGFRA/NG2 glia generate myelinating oligodendrocytes and piriform projection neurons in adult mice. Nat. Neurosci. 11, 1392–1401 (2008).

  26. 26.

    Robanus-Maandag, E. C. et al. A new conditional Apc-mutant mouse model for colorectal cancer. Carcinogenesis 31, 946–952 (2010).

  27. 27.

    Cahoy, J. D. et al. A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J. Neurosci. 28, 264–278 (2008).

  28. 28.

    Mizutani, M. et al. The fractalkine receptor but not CCR2 is present on microglia from embryonic development throughout adulthood. J. Immunol. 188, 29–36 (2012).

  29. 29.

    Arnett, H. A. et al. bHLH transcription factor Olig1 is required to repair demyelinated lesions in the CNS. Science 306, 2111–2115 (2004).

  30. 30.

    Lock, C. et al. Gene-microarray analysis of multiple sclerosis lesions yields new targets validated in autoimmune encephalomyelitis. Nat. Med. 8, 500–508 (2002).

  31. 31.

    Mei, F. et al. Micropillar arrays as a high-throughput screening platform for therapeutics in multiple sclerosis. Nat. Med. 20, 954–960 (2014).

  32. 32.

    Sauer, I., Dunay, I. R., Weisgraber, K., Bienert, M. & Dathe, M. An apolipoprotein E-derived peptide mediates uptake of sterically stabilized liposomes into brain capillary endothelial cells. Biochemistry 44, 2021–2029 (2005).

Download references

Acknowledgements

This work was supported by the National Natural Science Foundation of China (grant no. 31871045) and the National Science Foundation of Chongqing (grant no. cstc2018jcyjAX0702) to J.N., the US National Institutes of Health (grant no. R01-5R01NS088155) to S.E.B. (S.E.B. is the Heidrich Family and Friends Endowed Chair in Neurology at UCSF), and from the US National Institutes of Health (grant no. 1R01NS097551) to S.P.J.F. S.P.J.F. is a Harry Weaver Neuroscience Scholar of the National Multiple Sclerosis Society.

Author information

Affiliations

  1. Department of Neurology, University of California at San Francisco, San Francisco, CA, USA

    • Jianqin Niu
    • , Hui-Hsin Tsai
    • , Kimberly K. Hoi
    • , Kicheol Kim
    • , Sergio E. Baranzini
    • , Jonah R. Chan
    •  & Stephen P. J. Fancy
  2. Department of Histology and Embryology, Collaborative Innovation Center for Brain Research, Third Military Medical University, Chongqing, China

    • Jianqin Niu
    • , Nanxin Huang
    • , Guangdan Yu
    •  & Lan Xiao
  3. Department of Pediatrics, University of California at San Francisco, San Francisco, CA, USA

    • Stephen P. J. Fancy
  4. Division of Neuroimmunology and Glial Biology, University of California at San Francisco, San Francisco, CA, USA

    • Stephen P. J. Fancy
  5. Newborn Brain Research Institute, University of California at San Francisco, San Francisco, CA, USA

    • Stephen P. J. Fancy

Authors

  1. Search for Jianqin Niu in:

  2. Search for Hui-Hsin Tsai in:

  3. Search for Kimberly K. Hoi in:

  4. Search for Nanxin Huang in:

  5. Search for Guangdan Yu in:

  6. Search for Kicheol Kim in:

  7. Search for Sergio E. Baranzini in:

  8. Search for Lan Xiao in:

  9. Search for Jonah R. Chan in:

  10. Search for Stephen P. J. Fancy in:

Contributions

J.N. and S.P.J.F. conceived the study. J.N. and S.P.J.F. designed the experiments and analyzed the data. J.N. performed most of the experiments. S.P.J.F. performed experiments in human MS tissue. H.H.T. assisted with live imaging experiments. K.K.H. assisted with tissue processing and staining. N.H. and G.Y. assisted with 3D reconstruction of confocal images. S.E.B. and K.K. performed analysis of the mRNAseq data. L.X. contributed to discussion. J.R.C. helped design some of the experiments and contributed to discussion. J.N. and S.P.J.F. wrote the manuscript.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Stephen P. J. Fancy.

Integrated supplementary information

  1. Supplementary Figure 1 RNF43 and WIF1 in MS.

    (a) Frequency of total RNF43+ cells seen in NAGM, NAWM, and areas of chronic inactive and active lesions from the patients described in Supplementary Table 1. Data were analyzed by one-way ANOVA. Active vs. Chronic p = 0.007, Active vs. NAWM p = 5.06E-5, Active vs. NAGM p = 3.94E-5. (b) Frequency of RNF43+ cell clusters seen in active lesions (n=6) from the patients described in Supplementary Table 1. (c) Percentage of RNF43+ cells associated with vessels in active lesions (n=6) from the patients described in Supplementary Table 1. (d) Frequency of WIF1+ cells associated with vessels in active lesions (n=6) from the patients described in Supplementary Table 1. ** P < 0.01, **** P < 0.0001. The measure of center in all represents the mean.

  2. Supplementary Figure 2 OPC recruitment into murine lysolecithin spinal cord remyelinating lesions.

    (a) Double staining for PDGFRα+ OPCs and CD31+ vasculature shows OPC distribution before (0dpl ‘No lesion’) and after lysolecithin lesioning at 1day post lesioning (1dpl), 2dpl and 3dpl in dorsal funiculus of the spinal cord in mice. OPCs are eliminated from the lesion at 1dpl, with recruitment of small numbers back into the lesion by 2dpl. (b) Double staining for Ki67 and PDGFRα to mark proliferating OPCs in and around mouse spinal cord dorsal funiculus lesions at these same time points. (c) Quantification of OPC proliferation outside the lesion (dorsal funiculus outside the lesion) in unlesioned (0d) dorsal funiculus spinal cord and at 1, 2, 3, 5, 7 and 14dpl after lysolecithin lesioning, showing total number of PDGFRα cells per mm2 (black lines and black stars; n=6 animals at each time point; all statistical analyses compared with 0d, 1d vs. 0d ns p = 0.9705, 2d vs.0d ns p = 0.0691, 3d vs. 0d **p = 0.0054, 5d vs. 0d ns p = 0.0613, 7d vs. 0d ns p = 0.7910, 14d vs. 0d ns p = 0.5827) and those that are Ki67+ (red lines and red stars; n=6 animals at each time point; all statistical analyses compared with 0d, 1d vs. 0d ns p = 0.9314, 2d vs. 0d ****p = 3.04 E-9, 3d vs. 0d ****p = 2.14 E-8, 5d vs. 0d ****p = 3.70 E-6, 7d vs. 0d *p = 0.0141, 14d vs. 0d ns p = 0.9999). Data were analyzed by unpaired two-sided Student’s t test. (d) Quantification of OPC proliferation inside the lesion in dorsal funiculus spinal cord at 1, 2, 3, 5, 7 and 14dpl after lysolecithin lesioning, showing total number of PDGFRα cells per mm2 (black lines and black stars; n=6 animals at each time point; all statistical analyses compared with 0d, 1d vs. 0d *p = 0.0322, 2d vs. 0d ns p = 0.9195, 3d vs. 0d ****p = 3.96 E-8, 5d vs. 0d ****p = 2.25 E-8, 7d vs. 0d ****p = 2.39 E-5, 14d vs. 0d ****p = 1.74 E-4) and those that are Ki67+ (red lines and red stars; n=6 animals at each time point; all statistical analyses compared with 0d, 1d vs. 0d ns p = 0.9715, 2d vs. 0d *p = 0.0204, 3d vs. 0d ****p = 3.42 E-16, 5d vs. 0d ****p = 6.95 E-17, 7d vs.0d ****p = 1.48 E-7, 14d vs. 0d ns p = 0.0668). Data were analyzed by unpaired two-sided Student’s t test. (e) Analysis of OPCs with or without leading processes and their association with vasculature (‘On’ vessel or ‘Off’ vessel) in unlesioned mouse spinal cord dorsal funiculus or in dorsal funiculus outside a 2dpl lysolecithin lesion (n=6 animals; the percentage of total OPCs on blood vessel: Unlesioned On vs. 2d On, white bars and black stars, ****p = 1.36 E-5; the percentage of OPCs on blood vessel with leading process: Unlesioned On w/ leading process vs. 2d On w/ leading process, red bars and red stars, ****p = 2.11 E-9; the percentage of total OPCs NOT on the blood vessel: Unlesioned Off vs. 2d Off, white bars and black stars, ****p = 1.66 E-5; the percentage of OPCs NOT on the blood vessel with leading process: Unlesioned Off w/ leading process vs. 2d Off w/ leading process, red bars, ns p = 0.437). Data were analyzed by unpaired two-sided Student’s t test. Scale bars, 40 μm (a, b). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Values are mean ± s.d.

  3. Supplementary Figure 3 OPCs utilize single-cell perivascular migration for recruitment into murine remyelinating lesions.

    (a) Double staining for PDGFRα+ cells and CD31+ vasculature before lysolecithin lesioning (‘No lesion’), and at lesion edges at 3 days post lesioning (dpl) in mouse spinal cord dorsal funiculus. (b) OPC leading process and blood vessel correlation analysis in remyelination. The angles between blood vessels and leading processes of 295 PDGFRα+ OPCs (n=4 animals) at 2-day lesion edges were measured. (c) Vector map shows percentage of cells and leading process angle from vessel (0-90o). Red line indicates mean angle. (d) Time lapse imaging in slice cultures of adult spinal cord dorsal funiculus from 1.5dpl lysolecithin lesioned NG2creERT:TdTomato mice, following intracardiac infusion of fluorescein-lectin for vessel labeling. A TdTomato-expressing cell (red) first moves towards and engages a vessel, before migrating along it at the edge of a 1.5dpl lesion (Supplementary Video 2). Imaging times shown in top left corner. (e) Orientation of Dorsal (D), Ventral (V), Lateral (L) and Medial (M) relative to dorsal funiculus lesions for vector mapping quantifications used in (f) and (g). (f) OPC leading process direction analysis relative to lesion. Vector mapping shows orientation of OPC leading process relative to the lesion as explained in (e). Leading processes of 295 PDGFRα+ OPCs (n=4 animals) at 2-day lesion edges were measured (red line indicates the mean direction). (g) Direction and speed analysis of OPC migration from time lapse imaging of 1.5dpl lesioned NG2creERT:TdTomato mice above (n=4 animals). Vector mapping shows direction of migration with orientation relative to the lesion as explained in (e). Red line indicates the mean direction of OPC movement. Values in circles show the speed of OPC movement in μm/min. (h) Quantification of OPC ‘jumping’ between or ‘crawling’ along vessel movements in slices of lesioned NG2creERT:TdTomato above. (i) A subset of OLIG2+ OPCs in human HIE white matter injury express CXCR4 mRNA (unfilled arrows) and have cell body association with vessels (outline dotted line), in contrast to OLIG2+ cells which are negative for CXCR4 expression (filled arrows). Scale bars, 20 μm (a, d, i), 10 μm (b), 100 μm (e). For all staining results, experiments were repeated at least three times independently with similar results.

  4. Supplementary Figure 4 Characterization of labeled cells in lesioned NG2creERT: tdTomato mice.

    (a, b) Acute slice cultures of 1.5 day lesioned adult NG2creERT: TdTomato mice spinal cord dorsal funiculus were used for live imaging. PDGFRα (enlarged in b) staining shows that TdTomato labelled cells following tamoxifen treatment were predominantly OPCs at lesion edges. Blood vessels were labeled by tail vein injection of the Fluorescent-lectin dye. Arrows show close cell body association of OPCs with vasculature at lesion edges in these mice. (c) Quantification of TdTomato labelled cells in acute slice cultures of 1.5 day lesioned adult NG2creERT: TdTomato after tamoxifen treatment. The majority of cells were OPCs (n=6 animals, 86.13±9.17 % were Olig2 positive, 80.21±10.70 % cells were PDGFRα positive) with a minority of pericytes labelled (9.04±4.65% cells were PDGFRβ positive). (d) Recombination efficiency following tamoxifen treatment of NG2creERT:TdTomato in pericytes was significantly (n=6 animals, **** P =1.39 E-7) lower (10.90±1.75 %) compared to recombination in OPCs (71.19±4.14 %). Data were analyzed by unpaired two-sided Student’s t test. Scale bars, 30 μm (a, b). **** P < 0.0001. Values are mean ± s.d. For all staining results, experiments were repeated at least three times independently with similar results.

  5. Supplementary Figure 5 Excessive Wnt tone mediates OPC perivascular clustering in remyelination.

    (a-b) OPC clusters have excessive Wnt tone evidenced by significantly increased Axin2 mRNA expression (arrows in b) in PDGFRα-creERT2:APC fl/fl at 10dpl lesion edges, which is not seen in APCfl/fl controls (a). (c-d) Axin2 expressing OPC clusters (arrows) accumulate on vasculature (CD31+) at 10dpl lesion edges in PDGFRα-creERT2:APC fl/fl (c), and co-localize with expression of Olig2 protein (d, arrow). (e-f) OPC clusters (labelled with PDGFRα) are labelled with GFP expression and can be seen around vasculature (CD31+) at 10dpl lesion edges in PDGFRα-creERT2:APCfl/fl:Rosa-GFP (arrows in f), but are not seen in PDGFRα-creERT2:APCfl/+:Rosa-GFP (e). Scale bars, 20 μm (a, d, e). For all staining results, experiments were repeated at least three times independently with similar results.

  6. Supplementary Figure 6 OPC perivascular clustering in remyelination impairs their recruitment into lesions.

    (a) Double staining for PDGFRα+ OPCs and CD31+ vasculature shows OPC distribution before (0dpl ‘No lesion’ top row) and after lysolecithin lesioning at 3 days post lesioning (3dpl) in dorsal funiculus of the spinal cord in APCfl/fl control, Olig2cre:APCfl/fl and PDGFRα-creERT2:APCfl/fl mice. There are reduced OPC numbers (stars) at early 3dpl time points in spinal cord lesion centers in both PDGFRα-creERT2:APCfl/fl and Olig2cre:APCfl/fl mice. (b) Quantification of number of PDGFRα+ OPCs in lesion centers in unlesioned (0 days) and 3 days post lesioning (3dpl) spinal cord dorsal funiculus lesions in APCfl/fl (white bars) and Olig2cre:APCfl/fl (grey bars) mice (n=4 animals; 3d APC fl/fl vs. 3d Olig2 cre:APC fl/fl **p = 0.0013). Data were analyzed by unpaired two-sided Student’s t test. (c) Quantification of number of PDGFRα+ and Olig2+ cells in lesion centers in unlesioned (0 days) and 3 days post lesioning (3dpl) spinal cord dorsal funiculus lesions in APCfl/fl (white bars) and PDGFRα-creERT2:APCfl/fl (grey bars) mice (n=4 animals; PDGFRα: 3d APCfl/fl vs. 3d PDGFRα-creERT2:APC fl/fl ***p = 0.0003; Olig2: 3d APCfl/fl vs. 3d PDGFRα-creERT2:APC fl/fl **p = 0.0031). Data were analyzed by unpaired two-sided Student’s t test. (d) Quantification of number of PDGFRα+ OPCs on blood vessels at lesion edges in unlesioned (0 days) and 3 days post lesioning (3dpl) spinal cord dorsal funiculus lesions in APCfl/fl (white bars) and PDGFRα-creERT2:APCfl/fl (grey bars) mice (n=4 animals; 3d APCfl/fl vs. 3d PDGFRα-creERT2:APCfl/fl *p = 0.0215). Data were analyzed by unpaired two-sided Student’s t test. (e-h) Recruitment deficits into lesions in mice with Wnt-hyperactive OPCs do not seem to be due to differences in OPC proliferation. (e) Double staining for PDGFRα+ OPCs and Ki67 at 3 days post lesioning (3dpl) in dorsal funiculus of the spinal cord in APCfl/fl and PDGFRα-creERT2:APCfl/fl mice, quantified in (f) as percentage of PDGFRα+ cells that are also Ki67+ in lesion centers at 3dpl (n=4 animals; 0d APCfl/fl vs. 0d PDGFRα-creERT2:APCfl/fl, ns p = 0.3234; 3d APCfl/fl vs. 3d PDGFRα-creERT2:APCfl/fl, ns p = 0.9845). Data were analyzed by unpaired two-sided Student’s t test. (g-h) OPCs in vitro from APCfl/fl and Olig2cre:APCfl/fl proliferate to a similar extent (g), quantified in (h) as percentage of PDGFRα+ cells that are also Ki67+ (n=4 independent experiments; APCfl/fl vs. Olig2-cre:APCfl/fl, ns p = 0.9235). Data were analyzed by unpaired two-sided Student’s t test. Scale bars, 60 μm (a, e), 30 μm (g). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Values are mean ± s.d.

  7. Supplementary Figure 7 Morphological similarity of OPC perivascular clusters in MS and Olig2cre:APCfl/fl.

    Side by side example comparisons of morphological similarity of OPC perivascular clusters (vasculature marked with CD31, green) marked by OLIG2+ (red) in MS active lesions (a) and with TdTomato (red) in postnatal brain of Olig2cre:APCfl/fl:TdTomato (b). Arrows in each panel indicate OPC clusters. Scale bars, 20 μm in all panels. For all staining results, experiments were repeated at least three times independently with similar results.

  8. Supplementary Figure 8 OPC perivascular clustering disrupts astrocyte Aqp4 expression around vasculature.

    Expression intensity profiles of Aqp4, PDGFRα, and CD31 along cross sections of blood vessels in P9 corpus callosum of APCfl/fl control (a, b) and Olig2-cre: APCfl/fl (c, d) mice. Fluorescence intensities were measured using ZEN software. Yellow lines in each image indicate the cross section lines through vasculature along which intensities were measured. In the Olig2 cre:APCfl/fl mouse brain (c, d), Aqp4 intensity levels are reduced on vasculature when PDGFRα positive OPCs are clustered on blood vessels. (Graph Y axis= Fluorescence intensities; X axis= distance along vessel cross section in pixels). (e) Quantification of loss of AQP4+ astrocytic endfeet on blood vessels in P9 SC Olig2-cre:APCfl/fl compared to controls (n=6; APCfl/fl vs. Olig2-cre:APCfl/fl ****p = 1.47 E-6). Data were analyzed by unpaired two-sided Student’s t test. (f) Quantification of percentage of blood vessels with evidence of OPC clusters in P9 SC Olig2-cre:APCfl/fl compared to controls (n=6 animals; APCfl/fl vs. Olig2-cre:APCfl/fl ****p = 9.11 E-8). Data were analyzed by unpaired two-sided Student’s t test. Scale bars, 10 μm (a). **** P < 0.0001. Values are mean ± s.d. For all staining results, experiments were repeated at least three times independently with similar results.

  9. Supplementary Figure 9 OPC clusters displace astrocyte end feet from vessels in PDGFRα-creERT2:APCfl/fl remyelination.

    (a) Transverse section of vessels at 10dpl lesion edges following lysolecithin demyelination in spinal cord dorsal funiculus of PDGFRα-creERT2:APCfl/fl mice. Vessels are shown in blue (CD31+), astrocytes and processes in green (GFAP+) and processes of OPC perivascular clusters in red (PDGFRα+). (b-d) 3D reconstruction of image in (a) showing vessels in blue (CD31+), astrocytes and processes in green (GFAP+) and processes of OPC perivascular clusters in red (PDGFRα+). OPC perivascular clusters displace astrocyte endfeet during remyelination, and the processes of perivascular clustered OPCs appear to exist between endothelial cells and overlying astrocyte endfeet (arrows b-d). Scale bar, 10 μm. For all staining results, experiments were repeated at least three times independently with similar results.

  10. Supplementary Figure 10 Pericyte coverage of vessels appears unaltered by OPC perivascular clustering.

    (a) Pdgfrβ staining (marking pericytes) outlines a vessel (in longitudinal section) in P9 corpus callosum of Olig2-cre:APCfl/fl:TdTomato:Aldh1l1-GFP, and pericyte vascular coverage appears unaffected (arrow in a) by the presence of an OPC perivascular cluster (TdTomato+). (b) Whilst astrocyte endfeet (green, filled arrow) are displaced from vessels with OPC perivascular clusters (red, TdTomato+) in P9 CC of Olig2-cre:APCfl/fl:TdTomato:Aldh1l1-GFP, pericytes labelled with Pdgfrβ (white) seem unaltered, and outline a blood vessel (unfilled arrow). (c) Examples of transverse sections through vessels in P9 corpus callosum of Olig2-cre:APCfl/fl:TdTomato showing intact Pdgfrβ+ staining marking pericytes in the presence of OPC perivascular clusters (TdTomato+). (d) There is no significant difference in the Pdgfrβ+ pericyte coverage of CD31+ vessels in P9 corpus callosum of Olig2cre:APCfl/fl versus control APCfl/fl mice (n=5 animals; APC fl/fl vs. Olig2 cre:APC fl/fl, ns p = 0.7158). Data were analyzed by unpaired two-sided Student’s t test. Scale bars, 20 μm in all panels. Values are mean ± s.d. For all staining results, experiments were repeated at least three times independently with similar results.

  11. Supplementary Figure 11 OPC perivascular clusters trigger local microglial activation.

    Staining for Iba1, CD11c, and iNOS/Arg1 in P9 corpus callosum of APCfl/fl control (a) and Olig2-cre:APCfl/fl (b) mice. There is marked upregulation of CD11c expression (which marks both activated microglia and infiltrated macrophages) and iNOS expression in P9 Olig2-cre:APCfl/fl brain around OPC perivascular clusters (stained with dapi) compared to APCfl/fl brain. (c-d) Use of CX3CR1-GFP: CCR2-RFP mice (which label microglia green, and macrophages red) crossed into Olig2cre:APC fl/fl identifies these cells as predominantly activated microglia around OPC clusters. (e) Quantification of CCR2-RFP+ cells in P9 corpus callosum of APCfl/fl control and around OPC clusters in P9 CC of Olig2-cre:APCfl/fl (n=6 animals; APCfl/fl:CCR2-RFP vs.Olig2-cre:APCfl/fl:CCR2-RFP ***p = 0.0002). Data were analyzed by unpaired two-sided Student’s t test. Scale bars, 40 μm (a, b), 20 μm (c, d). *** P < 0.001. Values are mean ± s.d. For all staining results, experiments were repeated at least three times independently with similar results.

  12. Supplementary Figure 12 OPC perivascular clusters resolve over time.

    OPCs perivascular clusters resolve over time with reinvestment of astrocyte endfeet on blood vessels and reduction in microglial activation. (a) Use of Olig2-cre:APCfl/fl:TdTomato mice, to label OPC perivascular clusters in red (arrows), shows a decrease in cluster size around CD31+ vessels in corpus callosum through postnatal times P9, P12, P16 and P30. (b) Use of Olig2-cre:APCfl/fl:Aldh1l1-GFP at P9, P12, P16 and P30 in the corpus callosum shows a reduction in perivascular clusters over time (labelled with dapi, arrows) and reinvestment of GFP-labelled astrocyte endfeet on CD31+ vasculature. (c) Quantification of OPC cluster frequencies during development in Olig2-cre:APCfl/fl mouse CC at P9, P12, P16, P20 and P30 (n=6 animals at each time point; all statistical analyses compared to P9, P12 vs. P9 ns p = 0.1948, P16 vs. P9 * p = 0.141, P20 vs. P9 **** p = 2.25 E-5, P30 vs. P9 **** p = 2.19 E-6). Data were analyzed by unpaired two-sided Student’s t test. (d) Use of Olig2-cre:APCfl/fl corpus callosum at P9, P12, P16 and P30 shows a reduction in activated microglia (labelled with F4/80) over time around perivascular clusters (labelled with dapi). Scale bars, 80 μm (a, d), 30 μm (b). * P < 0.05, **** P < 0.0001. Values are mean ± s.d. For all staining results, experiments were repeated at least three times independently with similar results.

  13. Supplementary Figure 13 Physical presence of OPC clusters on vessels is required for effects on vascular and BBB integrity.

    (ac) PDGFRα+ OPCs aggregate on blood vessels to form clusters in Olig2-cre:APC fl/fl mice P10 spinal cord (b, -AMD, arrow) but are not seen in APCfl/fl controls (a). Treatment of Olig2-cre:APCfl/fl mice with AMD3100 (+AMD) between P3-P10 leads to a reversal of OPC clustering on blood vessels (b), quantified in (c)(n=6 animals; APCfl/fl vs. Olig2-cre:APCfl/fl ****p = 6.29 E-8; Olig2-cre:APCfl/fl vs. Olig2-cre:APCfl/fl+AMD3100 ####p = 1.04 E-5). Data were analyzed by one-way ANOVA. (df) Astrocytic endfeet are displaced from blood vessels in P10 spinal cord of Olig2cre:APCfl/fl:Aldh1L1-GFP mice (e, -AMD, arrow), but AMD3100 treatment (+AMD) significantly reduces this loss of Aldh1L1-GFP+ astrocytic endfeet, quantified in (f)(n=6 animals; APCfl/fl vs. Olig2-cre:APCfl/fl ****p = 4.81 E-8; Olig2-cre:APCfl/fl vs. Olig2-cre:APCfl/fl+AMD3100 ###p = 0.0008). Data were analyzed by one-way ANOVA. (gi) OPC perivascular clustering in P10 spinal cord of Olig2-cre:APCfl/fl leads to deficits in astrocyte endfoot AQP4 staining at cluster sights (h, -AMD, arrows), but AMD3100 treatment (+AMD) significantly reverses this phenotype, quantified in (i)(n=6 animals; APCfl/fl vs. Olig2-cre:APCfl/fl ****p = 7.85 E-6; Olig2-cre:APCfl/fl vs. Olig2-cre:APCfl/fl+AMD3100 ###p = 0.0001). Data were analyzed by one-way ANOVA. (j-l) Staining for fibrinogen in P10 APCfl/fl, Olig2-cre:APCfl/fl (-AMD) and Olig2-cre:APC fl/fl+AMD3100 spinal cord shows that there are fibrinogen positive areas outside blood vessels in Olig2-cre:APCfl/fl mice (arrow, k, -AMD) but this extravasation is significantly reversed by AMD3100 treatment (l; n=6 animals; APCfl/fl vs. Olig2-cre:APCfl/fl ***p = 0.0008; Olig2-cre:APCfl/fl vs. Olig2-cre:APCfl/fl+AMD3100 ##p = 0.0023). Data were analyzed by one-way ANOVA. (m-o) Staining for CD11c positive activated microglial cells (green) in P10 APCfl/fl, Olig2-cre:APCfl/fl (-AMD) and Olig2-cre:APCfl/fl+AMD3100 spinal cord shows that there are significantly increased activated microglia around clusters in Olig2-cre:APCfl/fl mouse (-AMD, n), but this microglia activation is reduced by AMD3100 (+AMD), quantified in (o)(n=6 animals; APCfl/fl vs. Olig2-cre:APCfl/fl **p = 0.0032; Olig2 cre:APCfl/fl vs. Olig2-cre:APCfl/fl+AMD3100 ##p = 0.0044). Data were analyzed by one-way ANOVA. Scale bars, 20 μm in all panels. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Values are mean ± s.d. For all staining results, experiments were repeated at least three times independently with similar results.

  14. Supplementary Figure 14 In vivo mRNA-seq analysis of Olig2cre:APCfl/fl versus APCfl/fl spinal cord.

    Whole spinal cord was dissected from three Olig2cre:APCfl/fl mice and three APCfl/fl littermate control mice at both postnatal day 4 (P4) and P9 time points. An RNAseq library was prepared and a differential expression analysis was performed as described in methods section. (a) Principal component analysis plot for all samples and outliers (arrowed). We removed one outlier sample (arrows) from APCfl/fl at P4 and one from Olig2cre:APCfl/fl at P9 since those samples are not clustered with the same experimental group. Finally, we performed differential expression analysis between two of APCfl/fl and three of Olig2cre:APCfl/fl at P4 and three of APCfl/fl and two of Olig2cre:APCfl/fl at P9 based on negative binomial distribution test using DESeq2 package. Significant level was calculated by two-sided Wald tests in DESeq2. (b) Heatmap for top 200 significant genes in the comparison of differential expressed genes between Olig2cre:APCfl/fl and APCfl/fl (b-1) in P4 and (b-2) in P9. Total 270 genes in P4 (43 up-regulated and 227 down-regulated genes) and 4,505 genes in P9 (2,092 up and 2,413 down) were significant in the comparison between Olig2cre:APCfl/fl and APCfl/fl (FDR < 0.05). (c) Heatmap of Wif1 gene (FDR < 2.13E-63 of the comparison at P4, FDR < 3.79E-89 at P9) and most significant genes (FDR < 1.6E-16 at P4, FDR < 2.0E-4 at P9) which were clustered with Wif1 by hierarchical clustering.

  15. Supplementary Figure 15 Western blot of Wif1 and Cldn5 protein levels in endothelial culture treatments.

    (a-b) Wif1 protein levels were measured in primary cultured OPCs from APCfl/fl control and Olig2-cre:APCfl/fl littermate mice using Western blot (a) and quantified in (b)(n=4 independent experiments; APCfl/fl vs. Olig2-cre:APCfl/fl *p = 0.0286). Data were analyzed by unpaired two-sided Student’s t test. (c-d) Primary cultured endothelial cells were treated with recombinant Wif1 protein, Wnt3a or Wif1+Wnt3a. Claudin5 (Cldn5) protein levels in treated endothelial cells were measured using Western blot (c). Quantification of these Cldn5 protein levels in treatment groups are shown in (d)(n=4 independent experiments; PBS vs. Wif1 **p = 0.0027; PBS vs. Wnt3a ****p = 4.19 E-7; Wnt3a vs. Wif1+Wnt3a ####p = 6.71 E-8). Data were analyzed by one-way ANOVA. (e-f) Cldn5 protein levels were measured by western blot in primary endothelial cultures treated for 2 days with conditioned medium from APCfl/fl or Olig2cre:APC fl/fl OPCs (labelled ‘WT-CM+BSA’ and ‘Olig2cre:APC fl/fl-CM+BSA’ respectively) or conditioned medium from the same OPCs depleted of Wif1 protein by overnight anti-Wif1 antibody (+Ab) treatment and agarose bead pull down. Quantification of Cldn5 protein levels in these treatment groups are shown in (f)(n = 4 independent experiments; WT-CM+BSA vs. APC-CM+BSA ***p = 0.0005; APC-CM+BSA vs. APC-CM+Ab #p = 0.0158). Data were analyzed by one-way ANOVA. (g) Wif1+ mRNA expressing cell density per mm2 in P9 Olig2cre:APCfl/fl mouse spinal cord versus APCfl/fl controls (n = 4 animals; APCfl/fl vs. Olig2-cre:APCfl/fl ****p = 2.67 E-5). Data were analyzed by unpaired two-sided Student’s t test. Values are mean ± s.d.

Supplementary information

About this article

Publication history

Received

Accepted

Published

Issue Date

DOI

https://doi.org/10.1038/s41593-019-0369-4