Single-Nucleus Profiles and Lesion Geography in the MS Cortex
Multiple sclerosis (MS) damages cortical grey matter and subcortical white matter differently, and cortical pathology includes damage to the axon, the neurite and the neuron cell body, particularly in areas underlying meningeal inflammation. Whether that process reaches all cortical neurons or only a subset was not understood. Schirmer and colleagues answered it by counting. They ran single-nucleus RNA sequencing across cortical grey matter and neighbouring subcortical white matter lesion areas at various stages of inflammation and demyelination, alongside control tissue from unaffected individuals. Of 19 MS tissue samples screened from 17 individuals, 12 passed an RNA integrity threshold above 6.5, as did 9 of 16 control samples, with age, sex, postmortem interval and RNA integrity not significantly different between the groups. After quality control the run yielded 48,919 single-nucleus profiles across 22 cell clusters, at a median of 1,400 genes and 2,400 transcripts per nucleus.
The Tissue Comes From a Window That Has Closed
One feature of the cohort shapes everything read from it. These are high-quality archival samples from patients who did not receive modern immunomodulatory therapies, so they represent the end point of the natural disease course, with relatively early death at 30 to 50 years of age. The pipeline was built to keep spatial information rather than discard it: serial sectioning of entire tissue blocks covering lesion and non-lesion grey and white matter areas plus meningeal tissue, then unbiased nuclei isolation on a sucrose gradient, barcoding and cDNA sequencing. None of the 22 clusters was made of nuclei captured from only one MS or control sample. Cell types were annotated on lineage markers, with neuronal subtypes split by SLC17A7 for excitatory neurons, CUX2 for the upper layer, RORB for layer 4 and TLE4 for deep layers, and GAD2 for interneurons with PVALB, SST, VIP and SV2C marking subtypes.
The Loss Is Selective, Down to a Single Layer
Comparing normalized nuclei counts between MS and control samples produced a selective reduction in excitatory neurons of the upper cortical layer, layers 2 to 3, in MS samples with cortical demyelination. Everything around them held steady. Layer 4 excitatory neurons, deep-layer excitatory neurons of layers 5 and 6, pyramidal cells with high THY1 and NRGN expression, and interneurons expressing VIP, SST or PVALB were present in similar numbers in MS and control tissue. Differential gene expression ranked the same way: layer 2 to 3 excitatory neurons carried the highest number of dysregulated genes, followed by layer 4 excitatory neurons and myelinating oligodendrocytes, with the fewest in SST-expressing interneurons and oligodendrocyte precursor cells. The upper-layer VIP-expressing interneurons, which sit immediately beside the vulnerable population, showed much less gene dysregulation.
Ordering the Damage Along a Trajectory
Unsupervised pseudotime trajectories within the layer 2 to 3 excitatory neuron cluster separated control from MS cells, and position along the trajectory tracked both conventional inflammatory lesion staging and the degree of upper-layer cortical demyelination. Cells sitting at the trajectory end came predominantly from samples harbouring late chronic inactive lesions with extensive subpial demyelination, and the association was strongest for upper-cortical-layer demyelination rather than deep-cortical-layer or subcortical demyelination. The genes moving along that trajectory read as a stress programme. Upregulated were cell stress and death markers (FAIM2, ATF4, CLU, B2M), the heat-shock response (HSPH1, HSP90AA1), protein accumulation and axon degradation (APP, NEFL, UBB), energy metabolism and oxidative stress (COX7C, PKM, PPIA), and the long non-coding RNAs NORAD and BCYRN1. Downregulated were mitochondrial energy consumption (FARS2), glutamate signalling (GRIA4, GRM5), potassium and ion homeostasis (KCNB2, KCNN2, SLC22A10), neuronal signalling (NELL1), axon plasticity (ROBO1) and LINC01266.
Putting the Sequencing Result Back Into Tissue
Rather than rest on the sequencing, the authors validated the cell-type-specific changes in situ with large area spatial transcriptomic mapping, using chromogenic and multiplex single-molecule fluorescence in situ hybridization protocols optimized against the high background autofluorescence of frozen human brain. Layer-associated expression of RORB, THY1, TLE4, VIP and SST was confirmed directly. Counting the two neighbouring upper-layer populations settled the selectivity question: CUX2-expressing neurons were significantly reduced in completely and incompletely demyelinated cortical areas, while the numbers of abutting VIP-expressing interneurons were maintained. Meningeal infiltration of plasma cells expressing IGHG1 and MZB1, which predominated over SKAP1-positive T cells, was a common finding in sulci with underlying upper-layer demyelination and loss of CUX2-expressing neurons. The stress markers held too, with PPIA transcripts raised in neurons from demyelinated areas and from neighbouring cortical areas of normal appearance, and NORAD upregulation confirmed with cytoplasmic accumulation in lesions.
Three Glial Lineages, Each With Its Own Geography
The glial results are spatial rather than global. Astrocytes downregulated SLC1A2 and GPC5 and upregulated GFAP and CD44, with the GFAP signature confined to demyelinated subcortical white matter and not crossing into the demyelinated cortex; cortical grey matter astrocytes lost regulators of glutamate (SLC1A2, GLUL) and potassium (KCNJ10) homeostasis, while reactive astrocytes at chronically active lesion rims expressed the transcription factors BCL6 and FOS alongside EDNRB and LINC01088. Myelinating oligodendrocytes, marked by ST18, carried the third-highest number of differentially expressed genes, with FTL and FTH1 raised in PLP1-expressing cells at iron-rich lesion rims along with B2M and HLA-C, and with markers of differentiation and myelin synthesis (BCAS1, SGMS1), cell interaction (SEMA6A) and node of Ranvier formation (GLDN) falling. Microglia kept a homeostatic signature of P2RY12, RUNX1 and CSF1R in both groups, with MS-specific subclusters enriched for activation markers, complement factors, MHC class II-associated proteins and the lipid degradation genes ASAH1, ACSL1 and DPYD, while the synapse-remodelling transcript SYNDIG1 and the potassium channel KCNQ3 fell. Reactivity markers CD68, CD74, FTL and MSR1 co-localized with RUNX1 and mapped to the chronic active boundaries of subcortical lesions.
An Apparent Artifact, Tested Rather Than Dismissed
One microglial cluster was characterized by phagocytosis and enrichment for the oligodendrocyte markers PLP1, MBP and ST18, which raises an obvious worry that ingested myelin transcripts had co-purified with the nuclei of phagocytosing cells. The authors went and tested it. Human and mouse microglia cultured with purified myelin from rat brain, which contains myelin transcripts, showed Plp1 and Mbp transcripts inside the cell, around the nucleus and within the nucleus itself one day after exposure to pHrodo-labelled myelin, with ingested Mbp mRNA still detectable up to four days after feeding. Phagocytosing mouse microglia also upregulated the activation marker Cd163 and downregulated the homeostatic marker P2ry12, changes mirrored in human MS microglia by sequencing. What they claim and what they leave open are both stated. The number of MS samples studied could have resulted in under-reporting of certain lineages. Whether the transport of ingested myelin transcripts is beneficial or detrimental to disease course, for example by worsening inflammation, needs future work. And while CUX2-expressing projection neurons may be damaged by sustained meningeal inflammation and by retrograde axon pathology from juxtacortical white matter lesions, the authors suggest additional intrinsic factors might account for their lack of resilience, given how little cell loss the neighbouring inhibitory and excitatory neurons showed. The abnormal gene expression they found in cortical areas of normal appearance points, in their reading, to a gradient of pathology rather than a boundary.
Disclaimer: This blog post is based on the cited study and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.
Reference:
Schirmer, L., Velmeshev, D., Holmqvist, S., Kaufmann, M., Werneburg, S., Jung, D., Vistnes, S., Stockley, J. H., Young, A., Steindel, M., Tung, B., Goyal, N., Bhaduri, A., Mayer, S., Engler, J. B., Bayraktar, O. A., Franklin, R. J. M., Haeussler, M., Reynolds, R., Schafer, D. P., Friese, M. A., Shiow, L. R., Kriegstein, A. R., & Rowitch, D. H. (2019). Neuronal vulnerability and multilineage diversity in multiple sclerosis. Nature, 573, 75–82. https://doi.org/10.1038/s41586-019-1404-z
