BGPT: Paper Review: YTHDC2 suppresses bladder cancer by inhibiting SOX2-mediated tumor plasticity

Fuel Your Discoveries

Quick Explanation Copied

Bottom line

The paper proposes a YTHDC2 → m6A-SOX2 translation suppression axis that restrains bladder cancer stemness/EMT plasticity, supported by multi-omics plus functional perturbations and a SOX2 rescue. Core claims come primarily from the authors’ own datasets and experimental validation.

Long Explanation

Paper Review

"YTHDC2 suppresses bladder cancer by inhibiting SOX2-mediated tumor plasticity"

DOI: 10.1038/s41419-025-08079-w · Journal: Cell Death & Disease · Received 2024-10-03; Accepted 2025-09-15

Proposed mechanism

YTHDC2 binds m6A-modified SOX2 mRNA and inhibits SOX2 translation → reduced cancer stemness/EMT plasticity → smaller tumors and more favorable biology.

Mechanistic backbone is from the authors’ multi-omics + binding + polysome + reporter + SOX2-rescue data.

Evidence map (what supports what)

Diagram is a conceptual evidence graph built solely from study sections visible in the provided full-text excerpt.

Raw-data-style summaries (derived directly from the excerpt)

The paper reports several quantitative thresholds and counts for multi-omics discovery. Below are concise visual summaries of those reported numbers.

Reported counts/percentages above are taken from the provided full-text excerpt (RNA-seq: |log2FC| > 0.585, P < 0.05; Proteomics: |log2FC| > 0.263, P < 0.05; MeRIP-seq common peaks 55.15%).

Claims and how well they are supported (with skeptical checks)

1) Clinical association: YTHDC2 is down in bladder cancer; lower relates to worse survival

The paper states large-scale RNA-seq analyses across many tissues using TNMplot (56,938 samples) and additional datasets (TCGA-BLCA, GTEx, GSE13507, GSE31684), plus IHC on a small tissue microarray (60 tumor cores; 13 adjacent normal cores; with exclusions).

Supported by: (i) reported differential expression and (ii) Kaplan–Meier subgroup analyses and (iii) IHC staining differences including subtype/stage comparisons.

Skeptical checks / blind spots: (a) expression–survival associations can be confounded by stage, molecular subtype, and batch effects; the excerpt doesn’t show whether fully multivariable models were used. (b) the IHC cohort is small, so effect stability is uncertain without independent validation. (These are methodological cautions; the excerpt does not provide enough detail to score them as proven limitations.)

2) Phenotype: YTHDC2 perturbation alters proliferation, migration/invasion, EMT markers, tumor growth, and stemness-like traits

The paper uses cell line selection based on endogenous YTHDC2 level (lowest vs highest) among 5637, T24, UM-UC-3, J82 and then performs CRISPR knockout and lentiviral overexpression. It reports corresponding shifts in proliferation (colony formation, CCK-8), migration/invasion (transwell), EMT marker changes (E-cadherin up, N-cadherin down with YTHDC2 overexpression; opposite upon deletion), and in vivo xenograft growth restriction.

Supported by: multi-assay consistency and bidirectional perturbations.

For stemness, it reports tumorsphere formation changes, extreme limiting dilution assay (ELDA), and marker changes (ALDH1+, CD133+, and KRT14 by IF; plus CD133 IHC in xenografts).

Supported by: multiple stemness readouts.

Skeptical checks / blind spots: (a) stemness markers (ALDH1/CD133/KRT14) can reflect multiple states beyond true tumor-initiating capacity; ELDA helps, but the excerpt doesn’t include replicate counts or model fit details. (b) EMT marker shifts indicate epithelial–mesenchymal transition, but EMT/MET dynamics aren’t tracked over time; the authors explicitly mention limited EMT–MET tracking due to culture duration.

3) Mechanism: YTHDC2 suppresses SOX2 translation via m6A-dependent binding and translational regulation assays

The paper claims that transcript-level changes from YTHDC2 silencing are not well aligned with stemness phenotypes, while proteomics changes are more consistent; therefore it prioritizes candidates where mRNA is stable but protein changes. SOX2 is identified as one such candidate, with reported mRNA stability and protein-level shifts in KO vs overexpression conditions.

Supported by: integrative multi-omics and concordant protein-level directionality.

For direct mechanism, it reports (i) GSEA suggesting translational repression/m6A-related RNA binding signatures with high YTHDC2 expression, (ii) MeRIP-qPCR showing SOX2 m6A enrichment, (iii) RIP-qPCR showing YTHDC2 interaction with SOX2 mRNA, (iv) RNA pull-down showing preferential binding to m6A-modified SOX2 probes, (v) polysome profiling showing increased ribosome loading after YTHDC2 KO, and (vi) luciferase reporters where effects depend on three predicted m6A sites.

Supported by: multiple orthogonal assays targeting binding, m6A dependence, and translational control.

Finally, causality is supported by rescue: SOX2 overexpression in YTHDC2-overexpressing T24 cells is reported to restore invasion and tumorsphere/self-renewal capacity.

Supported by: phenotype rescue experiments.

Skeptical checks / blind spots: (a) translation assays (polysomes and reporters) are strong, but the excerpt does not include full details about normalization controls, replicates, and potential reporter-specific artifacts. (b) Even with triple m6A site dependence in reporters, other m6A sites/cis-elements or indirect pathways could contribute to the net phenotype; the paper does report nine candidates, but the excerpt only fully validates SOX2. (c) Binding assays show association and preference for m6A-modified probe RNA, but they do not, in the excerpt, provide atomic-resolution structural evidence or competition with other YTH-domain targets.

Reproducibility & data availability (from provided excerpt)

The excerpt states that all data needed are present in the paper and that uncropped western blots are provided as supplementary material ("uncropped_original_western_blots.pdf").

Potential reproducibility gaps not fully checkable from excerpt: (i) raw sequencing files accessions aren’t visible in the excerpt. (ii) full statistical test details (exact n per experiment) are not always present in this excerpt. (iii) IHC scoring methodology is described, but inter-pathologist agreement metrics aren’t visible here.

Critique: what would most likely disprove or weaken the main axis?

SOX2 independence test: demonstrate that SOX2 knockdown in YTHDC2-overexpressing cells does not fully remove the phenotype (would weaken the rescue-based causality).
m6A-site competition: show that disrupting YTHDC2–SOX2 binding (e.g., via alternative SOX2 mutants) changes translation without reproducing the EMT/stemness phenotype, or vice versa.
Off-target phenotype: show that YTHDC2 perturbation yields similar phenotypes even when SOX2 is translationally constrained but not necessarily when SOX2 mRNA is altered.
Context dependence: confirm axis performance across additional bladder cancer models and in primary tumor-derived systems (the excerpt acknowledges limited tracking of EMT–MET dynamics and relies heavily on cell lines + xenografts).

Further BGPT next queries (actionable)

Note: The full paper text excerpt you provided includes many bibliographic items but not DOIs for all of them. To keep citations strictly accurate per your requested format, I cite only sources where DOIs are explicitly available in the provided material (primarily the reviewed paper itself).

Run an independent Science AI agent (iterative checks)

Author reviews (open BGPT deep dives)

Feedback:

Updated: March 23, 2026