Corpus 1,272 assessed · 1,173 scored · 643 reproduced ≥75 · 168 flagged ·∅ 74.1/100
← New search

Cancer-cell-secreted miR-204-5p induces leptin signalling pathway in white adipose tissue to promote cancer-associated cachexia.

Nat Commun · 2023
L1 85/100 PQI 95
Why this verdict

Part of the results reproduced; minor but material deviations remained.

Reproduced on the brainbox compute brainarbeit.com
Scoring basis — itemised

Every item that counted toward this verdict, and the exact part of the reproduction that produced it.

Supporting (toward a concern)
Content-critical question only partially held
+2 pts
From: Q7 · Core claim 🟡
Content-critical question only partially held
+2 pts
From: Q8 · Severity of the miss (overall human judgment) 🟡
Minor / cosmetic deviation
+1 pts
From: Q3 · Location of the main deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q6 · Severity of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +6
✓ What held up
  • Same input data as the authors
  • Reported values are derivable from the shared data
What did not (or only partly)
  • 🟡Reported values were only indirectly comparable
  • 🟡A deviation arose in the data or preprocessing
  • 🟡A deviation was attributed to the published material
  • 🟡The deviation was non-trivial in magnitude
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
85/100
Reproducibility score
0.6 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 67% of all assessed papers rank 348 of 1173 scored

A 0–100 reproducibility-quality score from the per-question grades, shown as a z-score: standard deviations above (+) or below (−) the mean of comparable assessments.

Reproduction agent’s raw note

salvaged by watchdog from agreement.json (agent omitted ROOM_RESULT.json)

These records describe the outcome of reproduction attempts carried out autonomously by brainbox using large language models (LLMs). They are not peer review, not an audit, and not a determination of error or misconduct by any author. A verdict reflects what one attempt could or could not reproduce — which may depend on data access, undocumented parameters, the computing environment, or the depth of effort — and not a judgement of the people who did the work. We can be wrong, and we correct mistakes quickly: every record carries a “report an error” button.

✎ I am an author of this paper

Updated or fixed a deposit, or is there an erratum? Ask us to re-run the metrics. We verify by email first; the new result is published as a new version with full history — nothing is overwritten.

Reason for the rerun

We email you a confirmation link first. The rerun is an objective re-measurement — it cannot change the verdict in your favour, only ask us to look again.

Provenance — full disclosure

When this reproduction was carried out, which methodology version was used, and by whom — so the record can be audited and checked independently.

Reproduced
2026-06-16
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-16
no human curator yet
Last updated
2026-07-29

Provisional, curator- or AI-assessed, and independently checkable. A reproduction outcome states what one attempt could reproduce — not a judgement of the authors.

Deep full-text extraction

Model: opus
Founding hypothesis

The study tests whether and how breast cancer-derived small extracellular vesicles (sEVs), specifically exosomal miR-204-5p, regulate white adipose tissue browning and fat loss to promote cancer-associated cachexia.

Core claims
  • Breast cancer cell-secreted exosomal miR-204-5p induces HIF1A in white adipose tissue by targeting the VHL gene. mechanism
  • Elevated HIF1A protein induces the leptin signalling pathway and enhances lipolysis/browning in WAT. mechanism
  • Breast cancer-derived sEVs are taken up by WAT and directly drive fat loss and hypermetabolism during cancer-associated cachexia. finding
  • Exogenous VHL expression blocks the effect of exosomal miR-204-5p on WAT browning. mechanism
  • miR-204 directly binds the 3'UTR of human and mouse VHL/Vhl genes. finding
  • Mice lacking cancer-derived miR-204-5p secretion show reduced plasma phosphatidyl ethanolamine levels. finding
  • Rab27a knockout/knockdown impairs exosome secretion and attenuates cachexia phenotypes. method
  • Circulating miR-204 is closely related to hypermetabolism and energy consumption in vivo. finding
Experimental setups
Assay System Perturbation Readout Platform
Orthotopic xenograft tumour model with Lck-GFP tracing BALB/c mice (4T1, 4T1/Rab27a KO) and NOD/SCID/IL2Rγ-null (NSG) mice (MDA-MB-231, 231/Rab27a KD) Rab27a KO/KD; tumour xenograft GFP signal in iWAT/eWAT, tissue/body weight, fat morphology
Indirect calorimetry (metabolic cage) BALB/c and NSG tumour-bearing mice; sEV-injected mice tumour bearing / sEV intravenous injection Oxygen consumption (VO2) and heat production over 48 h
sEV intravenous injection NSG mice (MCF-10A/MDA-MB-231 sEVs) and BALB/c mice (4T1 sEVs) ~10 μg sEVs twice weekly for 5 weeks sEV uptake, body weight, fat deposition, food intake IVIS Spectrum system; micro-CT
Bulk RNA-seq with GSEA eWAT from MDA-MB-231 tumour-bearing/sEV-treated mice vs controls tumour/sEV exposure Hypoxia (HIF1) and leptin pathway enrichment (NES, FDR q)
qRT-PCR iWAT and eWAT from tumour-bearing and sEV-injected mice tumour/sEV/Rab27a KD mRNA abundance of Vhl, Hif1α, leptin
Western blot / immunoblot iWAT, eWAT, primary adipocytes, 3T3-L1, hypothalamus sEVs, miR-204 mimic, Vhl cDNA overexpression VHL, HIF1A, LEPTIN protein; STAT3 pTyr705
Luciferase reporter assay MCF-10A cells transfected with WT or mutated VHL 3'UTR constructs miR-204 expression Reporter responsiveness confirming direct 3'UTR targeting
ELISA Serum, iWAT, and SVF cell culture medium sEVs (MDA-MB-231/10A-miR-204), tumour bearing Leptin concentration ELISA kit
Key results
  • 4T1/Ctrl and 231/Ctrl mice showed elevated oxygen consumption and heat production vs tumour-free or Rab27a-impaired mice
  • 4T1/Ctrl and 231/Ctrl mice lost weight after five weeks ~6.41% (4T1) and ~7.57% (231)
  • Body weight change in NSG mice receiving MDA-MB-231 sEVs vs PBS/MCF-10A sEVs -7.48±6.95% (231 sEVs) vs 8.72±4.37% (PBS), 7.03±0.63% (MCF-10A)
  • Body weight change in BALB/c mice receiving 4T1 sEVs vs PBS -2.35±1.43% (4T1 sEVs) vs 16.32±4.52% (PBS)
  • VHL protein suppressed and HIF1A induced in eWAT/iWAT of tumour-bearing and sEV-treated mice
  • miR-204 overexpression yielded upregulated miR-204 in 10A/miR-204 sEVs vs MCF-10A sEVs 15-fold
  • Plasma miR-204 enrichment in 10A/miR-204 sEV mice and tumour-bearing mice vs controls 10-fold (10A/miR-204) and 6-fold (tumour-bearing)
  • Leptin (mRNA, protein, serum/tissue) elevated in WAT of miR-204 sEV and tumour-bearing mice
Key statistics
  • fold_change 15-fold upregulation of miR-204 in 10A/miR-204 cells derived sEVs (10A/miR-204 sEVs vs MCF-10A sEVs)
  • fold_change 10-fold plasma miR-204 enrichment in 10A/miR-204 sEVs mice; 6-fold in tumour-bearing mice (plasma vs control mice)
  • fold_change 5-10 fold upregulation of miR-204 in iWAT/eWAT (high miR-204 sEVs or 4T1/Ctrl mice vs controls)
  • mean -7.48±6.95% (MDA-MB-231 sEVs), 7.03±0.63% (MCF-10A sEVs), 8.72±4.37% (PBS) (body weight change in NSG mice)
  • mean -2.35±1.43% (4T1 sEVs) vs 16.32±4.52% (PBS) (body weight change in BALB/c mice)
  • other ~6.41% and ~7.57% weight loss after five weeks (4T1/Ctrl and 231/Ctrl mice)
  • count n=7 BALB/c per group; n=5 NSG per group (metabolic measurements)
  • count n=3 mice Control and n=3 mice MDA-MB-231 tumour (RNA-seq/GSEA groups)

Statistical methods review

Model: opus

A neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.

The study is a series of controlled in vivo (mouse xenograft and sEV-injection) and in vitro experiments comparing tumour/sEV-treated groups against control groups, with most multi-group comparisons analysed by one-way ANOVA followed by Dunnett's multiple-comparison test against a reference control. Transcriptomic differences were assessed by RNA-seq with GSEA reporting normalized enrichment scores, nominal P-values and FDR q-values, while a luciferase reporter assay used an unpaired two-tailed t-test. Quantitative data are reported as mean ± s.e.m. with significance thresholds (*P<0.05 to ****P<0.0001), and exact P-values plus source data are provided in a Source data file.

Replicationbiological Sample sizePer-group sample sizes stated for individual figures (e.g., n=3–8 mice per group); no power/sample-size calculation described GroupsTumour-bearing or sEV-treated mice vs tumour-free/PBS/control-sEV groups; Rab27a KO/KD, miR-204 KO and rescue conditions Pairingunpaired Randomization/blindingnot stated DispersionSEM Exact p-valuesyes Effect sizesno Confidence intervalsno Multiplicity correctionDunnett's multiple-comparison test (post-hoc to one-way ANOVA); Benjamini-Hochberg-type FDR q-values for GSEA
Statistical tests used
Test Applied to n Assumptions
One-way ANOVA followed by Dunnett's multiple-comparison test Oxygen consumption/heat production and body-weight/fat-distribution comparisons across groups (Fig. 1d,e,j,k,l) and qRT-PCR of Vhl/Hif1α/leptin (Fig. 2b,d) n=7 (BALB/c) and n=5 (NSG) for metabolic cages; n=3 per group for micro-CT; n=5–6 mice per group for qRT-PCR not stated
Unpaired two-tailed t-test Luciferase reporter responsiveness of WT vs mutated VHL/Vhl 3'UTR to miR-204 (Fig. 2h) n=3 biological replicates not stated
GSEA (gene set enrichment) reporting NES, nominal P-value and FDR q-value RNA-seq of eWAT for hypoxia/HIF1 and leptin signalling pathways (Fig. 2a, Fig. 3a) n=3 mice per group na
Approaches that could also have been used
  • Multi-group comparisons used one-way ANOVA followed by Dunnett's test comparing each group to a single control.
    Could also: Tukey's HSD (or Sidak) post-hoc could also be used when all pairwise comparisons among groups are of interest. — Tukey/Sidak would additionally provide every group-to-group contrast with family-wise error control, whereas Dunnett focuses power on comparisons against the reference control.
  • Variability was summarized as mean ± s.e.m.
    Could also: Standard deviation or a 95% confidence interval could also be reported, alongside plotting individual data points. — SD/CI convey the spread of the data and estimate precision directly, which is often favoured for the small per-group n used here.
  • Several comparisons rely on small sample sizes (e.g., n=3) analysed with parametric tests.
    Could also: Non-parametric tests (e.g., Mann-Whitney U or Kruskal-Wallis with Dunn's post-hoc) could also be applied. — Non-parametric approaches do not assume normality, which can be useful when group sizes are small and the distribution is hard to verify.
  • The reporter assay (Fig. 2h) used an unpaired two-tailed t-test for WT vs mutant constructs across multiple binding sites.
    Could also: A two-way ANOVA (construct × site) or a t-test with multiplicity adjustment across the sites could also be used. — A single model would jointly account for multiple constructs/sites and control the error rate across the related comparisons.
  • Sample-size/power was described by stating per-group n without a power calculation, and randomization/blinding were not described.
    Could also: An a priori power analysis and explicit reporting of randomization/blinding could also be included. — These additions would document how n was chosen and how allocation/assessment bias was managed, aiding reproducibility.
  • Significance was emphasized with starred thresholds and exact P-values.
    Could also: Reporting estimated effect sizes with confidence intervals could also accompany the P-values. — Effect sizes with CIs communicate the magnitude and precision of differences, complementing significance testing.
Software: GSEA (gene set enrichment analysis) · miRNA target-prediction bioinformatics tools (for miR-204 targeting VHL)

Result convergence & founder nodes

Findings this paper shares with others that ran a comparable experiment. A node’s strength is how many independent papers report it (replication breadth) — not how often it is cited, so a heavily-replicated but under-cited founder still stands out.

Citation network

Where this publication sits in the reproducibility-weighted citation graph — what it is built on, and what is built on it. Citation data from OpenAlex.

Citations
61
Impact: high
Foundation confidence
None of its references are in our reproducibility record yet — its foundation cannot be assessed.
Topics

No assessed neighbours yet — the network grows as more papers are assessed.

Data lineage

The datasets this paper uses (text-mined from the full text via Europe PMC), and which other assessed papers stand on the same data. A shared dataset is a factual link — not a judgement.

GSE222380 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet

What was reproduced

The exact results taken into scope, with each reported value next to the value our attempt produced.

Scope — pmid-37620316

Paper: Hu et al. 2023, Nat Commun — "Cancer-cell-secreted miR-204-5p induces leptin signalling pathway in white adipose tissue to promote cancer-associated cachexia." DOI 10.1038/s41467-023-40571-9 · PMCID PMC10449837.

Data / code artifacts

  • GEO GSE222380 — bulk RNA-seq of mouse epididymal white adipose tissue (eWAT), Illumina NovaSeq 6000, Mus musculus, 9 samples in 3 groups (n=3 each):
    • Control / untreated eWAT (GSM6921889–891)
    • MDA-MB-231 sEV (EV)-treated eWAT (GSM6921892–894)
    • Tumor-bearing ("231/Ctrl") eWAT (GSM6921895–897)
    • Raw reads: SRA PRJNA921843. Processed: GSE222380_Raw_gene_counts_TPM.xlsx (5.4 MB).
  • Code link (per registry): https://github.com/jstjohn/SeqPrep — this is the generic FASTQ adapter-trim/merge tool used as the first pipeline step, not an authors' analysis repo. Per brief rule 2 (P16) applying a third-party tool to the paper's own data is a valid reproduction. The full analysis pipeline is described in the Methods (Majorbio-style boilerplate).

Pipeline (from Methods)

raw PE reads → SeqPrep + Sickle (adapter+quality trim, default params) → HISAT2 (align to reference, orientation mode) → RSEM (gene abundance, TPM) → DESeq2 (differential expression) → GSEA (pathway enrichment).

In scope (pipeline-derived → attempt to reproduce)

# Reported result Location Pipeline step
C1 eWAT gene-level quantification (raw gene counts / TPM per sample) GSE222380 processed file SeqPrep+Sickle→HISAT2→RSEM
C2 GSEA: "signalling by leptin" upregulated in 231/sEV and 231/Ctrl eWAT vs control Fig 3a DESeq2 + GSEA
C3 GSEA: HIF1 & hypoxia metagene pathway upregulated in 231/Ctrl eWAT vs control Fig 2a DESeq2 + GSEA

Out of scope (wet-lab / not pipeline-derived → NOT attempted)

  • qRT-PCR / Western / IHC / luciferase reporter / ELISA leptin measurements.
  • miR-204-5p sEV loading, in-vivo cachexia phenotyping (body/muscle weight), cell assays.
  • Any result with no shipped data + pipeline to regenerate it.

Reproduction strategy (two jobs)

  1. Downstream (quick): use authors' published count matrix → DESeq2 + GSEA → reproduce C2/C3 pathway directions (reproduces DE+GSEA steps).
  2. Full pipeline (heavy): download the 9 SRR, run SeqPrep+Sickle→HISAT2→ featureCounts/RSEM → reconstruct a gene-count matrix → correlate vs authors' published counts (C1), then DESeq2+GSEA on reconstructed counts (C2/C3).

Note: the paper reports the GSEA results qualitatively (direction of enrichment in figure panels); it does not print discrete DEG counts, per-gene fold-changes, or exact thresholds. Grading of C2/C3 is therefore direction/sign agreement, not a numeric tolerance.

Figures / tables: Fig 3aFig 2a
C1
Reported
Reproduced
m.public.grade.uncheckable
C2
Reported
leptin signalling UP (qualitative, Fig 3a)
Reproduced
tumor_vs_control NES=+1.909 padj=0.0481 (UP, sig); EV_vs_control NES=+0.995 padj=0.674 (UP direction, ns)
within tolerance
C3
Reported
hypoxia/HIF1 UP (qualitative, Fig 2a)
Reproduced
tumor_vs_control HALLMARK_HYPOXIA NES=+1.800 padj=2.88e-4 (UP, sig)
within tolerance

Assessments & scoring basis

Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.

🤖 AI curator · claude (ai-curator room) · v1.0 L1 85/100

An automated assessment. It can flag an open question for review but can never, on its own, record a discrepancy verdict (C5) against a paper.

🟢1. Data identity
🟡2. Endpoint comparability
🟡3. Location of the main deviation
🟡4. Cause of the deviation
🟢5. Derivability / plausibility
🟡6. Severity of the deviation
🟡7. Core claim
🟡8. Severity of the miss (overall human judgment)
Scoring basis — itemised

Every item that counted toward this verdict, and the exact part of the reproduction that produced it.

Supporting (toward a concern)
Content-critical question only partially held
+2 pts
From: Q7 · Core claim 🟡
Content-critical question only partially held
+2 pts
From: Q8 · Severity of the miss (overall human judgment) 🟡
Minor / cosmetic deviation
+1 pts
From: Q3 · Location of the main deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q6 · Severity of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +6

Reproduction ran on the authors' own deposited count matrix (GSE222380), so input data is identical and the reported qualitative directions are fully derivable — no fabrication concern. Both figure claims reproduce by sign of GSEA enrichment: hypoxia/HIF1 up (NES=+1.80, padj=2.9e-4) and leptin up in the tumor-bearing arm (NES=+1.91, padj=0.048). The one explainable shortfall is the sEV-vs-control leptin enrichment, which holds in direction but not significance (NES=+0.995, padj=0.67), attributable to our side — a proxy hypoxia set, an unspecified ranking, and a low-power 10-gene leptin set — rather than an authors' defect. Overall a solid, sign-agreement reproduction with method-dependent caveats; C1 full-pipeline quantification is still pending.

🤝
Reproduced automatically — and fairly

Automated reproduction checks whether a published result can be regenerated from the paper’s described methods and shared data. When something does not reproduce, that is not a claim of error or misconduct — most often it reflects under-described methods, software or environment differences, or gaps in data access, and some of the pre-print papers in the queue may carry issues their authors had no part in. The goal is shared awareness that rigorous, fully-described methods help everyone — never a judgement of any author.

Are you an author? We would genuinely like to hear from you — to clarify the record, add data or code, re-run the pipeline after an accession update, and publish your response right next to the assessment. Everything here is open and auditable.

🚩 Report an error in this record

Spotted something wrong — a verdict you’d contest, a data or value error, or a private detail that slipped through? Tell us, with a short justification. Authors and readers are equally welcome to write in; we review every report.

Prefer email, or the form below not working? Contact us at support@doesitreproduce.com.

Reproduction footprint

claude-opus-4-8

Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.

199.1 k
tokens (I/O) · 13.7 M incl. cache
76 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.