BGPT: Paper Review: Deubiquitinase USP18 regulates reactive astrogliosis by stabilizing SOX9

Explore by Goal

Quick Answer Copied

Core finding (mechanism → function)

After spinal cord injury (SCI), USP18 is upregulated in astrocytes and promotes reactive astrogliosis by binding SOX9 and stabilizing SOX9 protein via reduced SOX9 Lys48-linked ubiquitination ().

The paper further links upstream regulation: Hedgehog (Hh)/Smo→Gli2 transcriptionally activates USP18; pharmacologic Smo activation with SAG increases USP18/reactive astrogliosis and improves SCI recovery ().

Skeptical takeaway: the causal chain (USP18↔SOX9 protein stability↔reactive phenotype↔behavior) is well-supported in mice and in vitro, but translation is limited by reliance on markers (e.g., GFAP) and restricted data availability (“on request”), and by broad, non-specific pathway manipulation using SAG.

Long Answer

Paper Review (science-focused): USP18–SOX9 axis controls reactive astrogliosis after SCI

Glia • DOI: 10.1002/glia.23992

1) Reported temporal pattern (qualitative): USP18 rises after SCI

Qualitative reconstruction from the paper’s description that USP18 mRNA peaks at Day 7 and USP18 protein is increased at multiple timepoints post-injury ().

2) Proposed causal chain (USP18 → SOX9 stability → astrogliosis phenotype)

The chain is drawn directly from the paper’s narrative: Hh/Gli2 induces USP18, USP18 binds/deubiquitinates SOX9 stabilizing SOX9 protein, which increases reactive astrogliosis and limits inflammatory infiltration with improved functional recovery ().

3) Perturbation map (loss vs gain)

Manipulation	Core molecular lever	Reported astrocyte phenotype	Inflammation (CD11b+)	Functional outcomes
Astrocyte-specific USP18 conditional knockout (USP18 CKO)	↓ USP18 in astrocytes	↓ reactive astrogliosis / reduced GFAP area	↑ widespread infiltration of CD11b+ cells	Worse: lower BMS; poorer rotarod/footprints; altered MEPs
In vitro USP18 knockdown (astrocyte scratch injury)	↓ USP18	↓ proliferation + ↓ migration	Not directly measured	Not assessed
AAV9 astrocyte USP18 overexpression (AAV-USP18)	↑ USP18	↑ GFAP/reactive astrogliosis	↓ inflammatory infiltration	Better: higher BMS/rotarod; improved footprints and MEP measures
Hedgehog pathway activation (SAG; Smo agonist)	↑ Hh/Smo→Gli2→USP18 transcription (as proposed)	↑ reactive astrogliosis	↓ infiltration of inflammatory cells	Improved recovery/lesion outcomes and neuronal viability

Perturbation-direction claims are taken from the paper’s results narrative ().

4) Evidence layering: binding → ubiquitination → stability → phenotype

The underlying mechanistic steps are all described as experimental results in the paper: USP18-SOX9 binding via IP/MS + Co-IP; reduced SOX9 ubiquitination with WT USP18 and increased ubiquitination with USP18 knockdown; catalytic requirement evidenced by WT vs C64S failing to stabilize SOX9; SOX9 half-life reduction upon USP18 knockdown; and in vivo reversal by SOX9 depletion of USP18 overexpression effects ().

Methods & experimental logic (what they did)

Gene-expression discovery used microarrays of injured vs sham spinal cord, identifying 47 deubiquitinase genes and emphasizing USP18 upregulation ().
Astrocyte specificity was argued using immunofluorescence: “almost exclusively expressed by astrocytes” and absent from CD11b+, NeuN+, and CD31+ populations ().
Causality: loss-of-function used astrocyte-specific conditional knockout (GFAP-Cre; USP18 flox/flox) and in vitro USP18 siRNA; gain-of-function used AAV9 driven USP18 expression and pharmacologic pathway activation using SAG ().
Mechanism: binding (IP/MS + Co-IP), ubiquitination (IP-based ubiquitination assays; chain-type focus), protein stability (cycloheximide chase/half-life), and functional dependence (SOX9 knockdown reverses USP18 phenotypes) ().

Critical appraisal (skeptical / blind spots)

1) Marker-based “reactivity” may oversimplify astrocyte heterogeneity

The paper uses GFAP (and Ki67+/GFAP) as core indicators of reactive astrogliosis (). Because astrocytes span multiple reactive programs, GFAP-based scoring can miss program-specific or time-specific states.

2) SAG is a broad pathway perturbation

SAG activates Smo/Hh signaling, which can influence multiple cell types and transcriptional programs. The paper directly reports that SAG increases USP18 expression and reactive astrogliosis and improves outcomes, but causally attributing all SAG phenotypes solely to USP18 requires additional epistasis controls (e.g., SAG effects in USP18 CKO) that the authors themselves suggest would strengthen the Hh→USP18 link ().

3) Data availability is restricted (“on request”)

Reproducibility depends on access to underlying data. The paper states supporting data are available on request and not publicly available due to privacy/ethical restrictions (). This limits independent reanalysis.

4) USP18 has known non-SOX9 functions (IFN signaling), which could confound interpretation

USP18 is widely recognized as a multifaceted regulator, including type I interferon signaling regulation that can be independent of catalytic activity (). The paper acknowledges this as a possibility for the broader SCI phenotype and cannot fully rule out alternative or parallel mechanisms beyond SOX9 stabilization ().

Contextual grounding (why astrocyte “reactivity” can be protective or harmful)

Reactive astrocytes are frequently described as having context-dependent dual roles; mechanistically, glial scar formation can both limit inhibitory spread and shape inflammatory trajectories, yet may also impede axon regeneration. The broader literature emphasizes this nuance (; ). This matters because the paper’s finding that dampening astrocyte reactivity (USP18 CKO) worsens functional recovery is consistent with a protective barrier model, but it also highlights why therapy might need mechanism- and timing-specific modulation rather than blanket suppression.

Explore author-specific expert reviews on BGPT

Feedback:

Updated: March 25, 2026