DNA damage burden causes selective CUX2 neuron loss in neuroinflammation | Nature

Download PDF

Subjects

- Experimental models of disease

- Multiple sclerosis

Abstract

Neurodegeneration shows regional and cell-type-specific patterns in ageing and disease1, but the underlying mechanisms for cell-type-specific neuronal losses remain poorly understood. Previous studies have shown that upper cortical layer thinning occurs in progressive human multiple sclerosis (MS) and that cortical layer 2 and layer 3 (L2/3) excitatory neurons (L2/3ENs) that express CUT-like homeobox 2 (CUX2) are selectively vulnerable to degeneration2. Here we report that L2/3ENs within MS cortical lesions have an elevated DNA damage burden. DNA damage and selective loss of L2/3ENs were recapitulated in diverse mouse models of demyelination and pan-cortical inflammation, confirming their intrinsic vulnerability. Functions of Cux2 and activating transcription factor 4 (Atf4) were essential for resilience of L2/3ENs during postnatal neuroinflammation, acting in neurons to enhance DNA double-strand break repair. Interferon-γ, a cytokine implicated in MS pathogenesis3,4, was sufficient to elevate levels of reactive oxygen species, leading to DNA damage-mediated neuronal death in vitro, and caused selective depletion of L2/3 neurons in mice. These findings indicate that DNA damage burden and inadequate repair in CUX2+ L2/3ENs contributes to selective vulnerability in neuroinflammatory injury.

Main

Ageing and neurodegeneration are linked to the buildup of DNA damage in both glial cells and neurons5,6,7. Such damage can originate from internal sources like stress induced by metabolic by-products (for example, reactive oxygen species (ROS), advanced glycation end products and alkylation species) or normal cellular activities (such as transcription and replication)5. Moreover, DNA damage can arise from external factors such as cosmic ionizing radiation, diet, pollution or chemotherapy5. If left unrepaired, primary forms of DNA damage, including single-stranded breaks (SSBs), double-stranded breaks (DSBs) and oxidative base lesions, compromise neuronal function and lead to cell death. Neurons mount a DNA damage response (DDR), which involves repair pathways such as base-excision repair (BER) for SSBs and oxidative damage, transcription-coupled repair for DNA lesions blocking transcription, mismatch repair for correcting single base-pair mismatches or insertion–deletion loops, and error-prone non-homologous end joining (NHEJ) for DSBs that are critical for maintaining genomic stability7.

MS, an autoimmune disease targeting myelinating oligodendrocytes of the central nervous system (CNS), is the most common cause of neurological disability in young adults8. MS shows an initial relapsing–remitting course followed by progression, characterized by clinical deterioration and brain atrophy with upper cortical layer thinning9. Although the outcomes of individuals with relapsing–remitting MS have improved with therapies targeting T and B cells, patients still show disease progression and brain atrophy10,11, highlighting the need to better understand mechanisms of neurodegeneration in MS. Previously we reported selective loss of L2/3ENs in MS2, but whether this was indicative of their intrinsic vulnerability to direct neuroinflammatory injury remained unclear. L2/3ENs are marked by expression of CUT homeodomain proteins (such as CUX1, CUX2 and SATB2; Fig. 1a), which are known to promote DNA BER after oxidative stress12, yet their cortical DDR function remains largely unexplored. We previously reported significant upregulation of the genome stability factor non-coding RNA activated by DNA damage (NORAD) in MS L2/3ENs2,13, suggesting increased DNA damage burden. Here we investigated the mechanisms of vulnerability of L2/3ENs in relation to DNA damage and repair in the setting of neuroinflammatory stress.

Fig. 1: DNA damage and response in human MS and DTA mice.Full size image

a, Schematic of cortical neuron layers in control (Ctrl) and demyelinated grey matter (DMGM) in MS. b–f, Immunohistochemistry (b) and quantification (c–f) of pan-nuclear 53BP1 (c), γH2AX and 53BP1 foci (d), γH2AX foci (e) and pan-nuclear γH2AX (f) in NeuN+ L2/3 neurons. n = 5 (control) and n = 4 (MS). g, Schematic of the single-nucleus sequencing method. h, Pseudotime trajectory of DDR and apoptosis genes in MS L2/3ENs. n = 9 (control), n = 12 (MS). i, All DEGs (grey dots) overlaid with upregulated (red) or downregulated (blue) DDR DEGs in select cortical cell types. n = 9 (control), n = 12 (MS). j, Schematic of the disease course of DTA mice along weeks (W) after induction. k–r, Immunohistochemistry analysis of NeuN (k), CTIP2 (l), CUX1 (n,o), 53BP1, γH2AX and NeuN (p), OHdG and NeuN (q) and 8-OHG, 8-OHdG and 8-OHGua with NeuN (r), counterstained with DAPI (k,l) with cell counts (n = 5 per group (44 weeks), n = 3 per group (27 weeks)), PDDF or foci counts (n = 4 (5 weeks, control), n = 6 (5 weeks, DTA); n = 4 per group (10 weeks and 17 weeks)) and integrated den