Thermogenic adipocytes promote HDL turnover and reverse cholesterol transport
Part of the results reproduced; minor but material deviations remained.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- ✓No relevant deviation in data/preprocessing
- ✓Any deviation was negligible
- 🔴Could not use the authors’ exact input data
- 🔴Reported values were only indirectly comparable
- 🔴A deviation was attributed to the published material
- 🟡Reported values were not (fully) derivable from the shared data
- 🟡The central claim did not (fully) hold under reproduction
- 🟡Overall, the reproduction showed a material discrepancy
▸Reproduction agent’s raw note
DROP. Nature Communications 2017 paper (10.1038/ncomms15010, PMC5399294) on thermogenic adipocytes, HDL turnover and reverse cholesterol transport. Described well enough scientifically, but NOT a computational-pipeline reproduction unit: the work is in vivo mouse/human metabolic physiology (HDL kinetics, 3H-cholesterol reverse cholesterol transport, atherosclerosis, FPLC, qPCR) plus targeted lipidomics quantified in proprietary Bruker vendor software (DataAnalysis 4.0 / TargetAnalysis 1.3) with statistics in GraphPad Prism 5.0 / SPSS 20.0 (GUI t-tests + Benjamini-Hochberg FDR). There is NO RNA-seq/microarray/omics, NO public data accession of any kind (GEO/SRA/Metabolights/PRIDE/figshare/zenodo) - data are 'available from the corresponding authors upon reasonable request' (data_restricted) - and NO code repository or code-availability statement (no_code). Even re-running the published lipidomics statistics is impossible: the only supplementary object is a figures+tables PDF with no machine-readable per-sample values. Primary drop_reason: non_pipeline; compounding: data_restricted, no_code. No «our HPC» compute was submitted because there is no public dataset to fetch and no pipeline to run - fabricating a job would have produced no genuine reproduction. Not attempted: all wet-lab/in vivo claims (C1-C3, C5), the BH-corrected lipidomics (C4, no per-sample data), and any third-party-tool reanalysis (no obtainable input). Transparency: the signed-in operator («email») is co-author Schlein C; this does not change the independent verdict. Datasets profiled honestly in data/dataset_profile.json (n_observed=null everywhere - nothing public to count).
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.
Assessment versions
Every reproduction run is kept as an immutable version — anchored to the data as it stood, with a tamper-evident chain hash. A rerun (e.g. after an author updates a deposit) adds a new version; the previous one stays on record.
-
v1 current initial assessmentassessed: 2026-06-18 ⛓ bb72659746c2
✎ 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.
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
- 2026-06-18
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-18no human curator yet
- Last updated
- 2026-08-05
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: sonnetBased on human epidemiological associations linking thermogenic adipocyte activity to lipid profiles, the authors hypothesized that the metabolic activity of thermogenic adipocytes is mechanistically linked to HDL metabolism and reverse cholesterol transport.
- ★ Pharmacological (CL316,243) activation of thermogenic adipocytes increases HDL-cholesterol and reduces atherosclerosis in E3L.CETP mice, with HDL-cholesterol an independent predictor of atherosclerosis finding
- ★ Cold and CL-induced thermogenic activation remodels the HDL lipidome (PC, lyso-PC, CE species) in mice and in lean (but not obese) humans, largely independent of total HDL-cholesterol levels finding
- ★ Thermogenic adipocyte activation increases macrophage-to-faeces reverse cholesterol transport (RCT) in mice finding
- ★ Thermogenic activation accelerates total and selective HDL-cholesterol plasma clearance and hepatic uptake, especially in the postprandial (lipolytically active) state finding
- ★ Adipocyte lipoprotein lipase (LPL)-mediated intravascular lipolysis is required for thermogenesis-induced HDL remodelling and turnover, shown using adipocyte-specific LPL knockout (aLKO) mice mechanism
- ★ Hepatic scavenger receptor B-I (SR-BI) mediates cold-induced accelerated HDL-cholesterol clearance, shown using Scarb1-/- mice mechanism
- ★ HDL lipidomic remodelling does not alter ex vivo macrophage cholesterol efflux capacity or the plasma clearance/hepatic uptake capacity of isolated HDL particles themselves finding
- Lipolysis-driven transfer of TRL-derived surface remnants to HDL (Eisenberg-type mechanism) is proposed to explain accelerated HDL-cholesterol flux to the liver mechanism
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| FPLC lipoprotein cholesterol profiling | C57BL/6J wild-type, Apoa5-/-, and E3L.CETP mice | cold exposure or CL316,243 (7 days) | lipoprotein cholesterol distribution (TRL/HDL), plasma TG and cholesterol | FPLC |
| Lipidomics (high-resolution full-scan MS with collision-induced dissociation) | isolated HDL and TRL from Apoa5-/- and E3L.CETP mice | cold exposure or CL316,243 | PC, lyso-PC, and cholesteryl ester (CE) species composition | mass spectrometry |
| Lipidomics of isolated lipoprotein fractions | plasma from lean and obese human subjects | 2-day cold exposure at 16°C | HDL-associated PC, lyso-PC, CE species and lipoprotein profile | — |
| Cholesterol efflux assay | primary mouse peritoneal macrophages | ex vivo exposure to serum/HDL from cold- or CL-treated mice | macrophage cholesterol efflux capacity | — |
| Radiolabelled HDL plasma clearance/hepatic uptake assay | wild-type mice injected with 3H-cholesterylether-labelled HDL from mock- or CL-treated mice | ex vivo radiolabelled HDL from CL-treated donors | plasma clearance and hepatic uptake of HDL | — |
| In vivo macrophage-to-faeces reverse cholesterol transport (RCT) assay | wild-type and E3L.CETP mice injected with LDL/3H-cholesterol-loaded macrophages | cold exposure or CL316,243 | 3H-cholesterol in plasma, liver, and faeces | — |
| Double-radiolabelled HDL turnover assay (3H-cholesterylether + 125I-apolipoprotein) | wild-type, adipocyte LPL knockout (aLKO), and Scarb1-/- mice | cold adaptation (fasted and re-fed states) | total and selective plasma cholesterol clearance, liver/gallbladder/adipose uptake | — |
| Western blot | liver tissue from cold-adapted wild-type mice | cold exposure vs thermoneutrality | hepatic SR-BI protein levels | — |
- ▲ Chronic CL treatment increased HDL-cholesterol in E3L.CETP mice on Western-type diet, and HDL-cholesterol was an independent predictor of atherosclerosis by multiple regression
- ▲ In Apoa5-/- mice, cold and CL induced a shift of cholesterol from TRL to HDL fraction, increasing HDL-cholesterol; this shift was not seen in wild-type or E3L.CETP mice
- ▲ Cold and CL treatment increased faecal 3H-cholesterol excretion (RCT) in wild-type and E3L.CETP mice, with decreased plasma and liver 3H-cholesterol
- ▲ Cold adaptation increased total and selective cholesterol clearance from plasma in fasted mice without changing liver 3H-cholesterylether uptake; in re-fed mice both clearance and liver/gallbladder uptake were increased
- ▼ Adipocyte LPL deficiency (aLKO) reduced total and selective plasma cholesterol clearance and hepatic HDL-cholesterol uptake under cold adaptation, while 125I-HDL uptake was unaffected
- ▼ Scarb1-/- mice showed impaired total and selective cholesterol clearance and decreased 3H-cholesterylether uptake into liver and gallbladder under cold adaptation, while 125I-HDL uptake was unaltered
- – Cholesterol efflux capacity of serum and isolated HDL from cold- or CL-treated mice was not different from controls
- – In lean humans, cold exposure significantly reduced HDL-associated PC34:1 and Lyso-PC18:1 and increased Lyso-PC16:0, with no significant changes in obese individuals
Statistical methods review
Model: sonnetA neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.
The paper combines pharmacological and cold-exposure interventions in multiple mouse models (C57BL/6J, Apoa5−/−, E3L.CETP, aLKO, Scarb1−/−) with a small human cold-exposure cohort to investigate how thermogenic adipocyte activation affects HDL metabolism, lipidomics, and reverse cholesterol transport. Key outcomes include plasma lipoprotein profiling by FPLC, high-resolution mass-spectrometry lipidomics of isolated lipoprotein fractions, radiolabelled HDL turnover assays, and a macrophage-to-faeces RCT assay. Multiple regression analysis was used to assess HDL-cholesterol as an independent predictor of atherosclerosis, and between-group differences in lipid species and tracer kinetics were described as statistically significant, though the specific inferential tests and dispersion measures applied to most comparisons are not named in the available text.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| Multiple regression analysis | HDL-cholesterol as independent predictor of atherosclerosis development alongside non-HDL-cholesterol (Fig. 1b) | — | not stated |
| Unspecified significance test(s) for between-group comparisons | HDL-associated lipid species (PC, Lyso-PC, CE) after cold or CL treatment vs. control in Apoa5−/− and E3L.CETP mice (Figs. 3, Supplementary Figs. 10–11) | — | not stated |
| Unspecified significance test(s) for between-group comparisons | HDL-associated PC34:1, Lyso-PC18:1, Lyso-PC16:0 in lean vs. obese humans before and after cold exposure (Fig. 4c–e) | — | not stated |
| Unspecified significance test(s) for between-group comparisons | Cholesterol efflux capacity from primary peritoneal macrophages ex vivo (Fig. 5a,b) | — | not stated |
| Unspecified significance test(s) for between-group comparisons | Radiolabelled (3H-cholesterol) plasma, liver, and faecal concentrations in RCT assay across thermoneutral, cold-adapted, and CL-treated mice (Figs. 5e–j) | — | not stated |
| Unspecified significance test(s) for between-group comparisons | Total and selective HDL-cholesterol plasma clearance and hepatic uptake in cold-adapted vs. thermoneutral controls, aLKO mice, and Scarb1−/− mice (Figs. 6a–l) | — | not stated |
-
Multiple regression was used to identify HDL-cholesterol as an independent predictor of atherosclerosis alongside non-HDL-cholesterol↳ Could also: A linear mixed model or ANCOVA accounting for within-animal variance, or a penalised regression approach, could also have been applied — Mixed models handle repeated-measures or correlated observations within the same animal and allow explicit modelling of random effects, which can improve precision of the predictor estimates when data have a hierarchical structure
-
Lipidomic comparisons of individual PC, Lyso-PC, and CE species across treatment groups were reported as significant without a named multiplicity correction↳ Could also: A false discovery rate (FDR) correction such as Benjamini-Hochberg, applied across all lipid species tested within each experiment, is a standard approach in lipidomic studies — When many lipid species are tested simultaneously, an FDR correction limits the expected proportion of false discoveries among declared positives, and its use is widely expected in mass-spectrometry-based lipidomics to aid interpretation of which species changes are most likely real
-
Radiolabelled HDL clearance curves (3H and 125I) were compared between groups (cold-adapted, aLKO, Scarb1−/−) using unspecified between-group tests↳ Could also: Compartmental pharmacokinetic modelling (e.g., two-compartment model fitted with non-linear least squares) or area-under-the-curve (AUC) analysis followed by a parametric or non-parametric group comparison could also quantify HDL turnover — Compartmental modelling provides explicit kinetic parameters (fractional catabolic rate, production rate) with standard errors, enabling more mechanistic interpretation and direct comparison of turnover rates across genotypes and conditions
-
The human cold-exposure experiment compared lean and obese subgroups before and after cold, with significance evaluated in lean subjects but not obese↳ Could also: A two-way repeated-measures design (group × time interaction) analysed with a mixed ANOVA or linear mixed model, with correction for the two lipid-species families tested, could also have been applied — A formal interaction term directly tests whether the cold-exposure effect differs between lean and obese participants, which is the central biological question, rather than requiring separate within-subgroup tests that do not directly compare the two groups' responses
-
Multiple mouse models and two genetic knockouts were compared to wild-type controls using separate between-group tests for each outcome↳ Could also: A one-way ANOVA (or Kruskal-Wallis for non-normal data) with a post-hoc correction such as Tukey HSD or Dunnett's test could also encompass all group comparisons within a single experiment — A single omnibus test with post-hoc correction controls the family-wise error rate across all pairwise comparisons within an experiment and is a widely used alternative when three or more groups are compared on the same outcome
-
The RCT assay compared 3H-cholesterol in plasma, liver, and faeces as three separate outcomes per experiment↳ Could also: A multivariate approach (e.g., MANOVA) or a mixed model with compartment as a within-subject factor could also jointly model the plasma-liver-faeces distribution of tracer — Because the three compartments are not independent (cholesterol excreted to faeces is necessarily depleted from plasma and liver), a joint model accounts for their correlation and can reduce the risk of contradictory conclusions from separately analysed outcomes
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-28422089
Paper: Bartelt A, John C, Schaltenberg N, … Schlein C, … Scheja L, Rensen PCN, Heeren J. "Thermogenic adipocytes promote HDL turnover and reverse cholesterol transport." Nature Communications 8, 15010 (2017). DOI: 10.1038/ncomms15010 · PMCID: PMC5399294 · PMID: 28422089.
Determination: DROP — no reproducible computational pipeline + no public data
This is an in vivo metabolic-physiology paper (mouse + human cold-exposure / pharmacological-thermogenesis studies of HDL turnover, reverse cholesterol transport, and atherosclerosis). After reading the abstract, Methods, and the Data-Availability statement (PMC full text + Nature article page + supplementary listing), no pipeline-derived computational result with obtainable inputs could be identified. The candidate was harvested by text-mining but is not, in fact, a computational-pipeline reproduction unit.
Out of scope — wet-lab / in vivo / instrument assays (not attempted)
| Result class | Why out of scope |
|---|---|
| HDL turnover (³H/¹²⁵I-HDL kinetics), reverse cholesterol transport (³H-cholesterol macrophage→faeces tracer) | In vivo tracer physiology in mice/humans; not computational |
| Atherosclerosis lesion quantification (E3L.CETP) | Histology/imaging morphometry; manual |
| FPLC plasma lipoprotein profiling | Instrument chromatography; manual peak reading |
| qPCR gene expression (TaqMan: Ucp1, Ppargc1a, …) | Wet-lab; no deposited expression matrix |
| Plasma/lipid clinical chemistry | Enzymatic kit assays |
"Computational" content that exists — but is NOT a reproducible pipeline
| Component | Software | Why not reproducible here |
|---|---|---|
| Lipidomics (CE, LPC, PC species) by UHR-Q-TOF MS | Proprietary vendor software: Bruker DataAnalysis 4.0, TargetAnalysis 1.3; LIPID MAPS DB | Closed-source, vendor-locked; raw .d spectra not deposited (data on request only) |
| Statistics: Student's t-test, Benjamini–Hochberg FDR on lipid species | GraphPad Prism 5.0 / SPSS 20.0 (GUI) | Not a scripted pipeline; and re-running requires per-sample values, which are not deposited |
Data availability (verbatim)
"The datasets generated during and/or analysed during the current study are available from the corresponding authors upon reasonable request."
No public accession of any kind: no GEO / SRA / ArrayExpress / Metabolights /
PRIDE / figshare / zenodo. No RNA-seq, no microarray, no high-throughput omics.
The single supplementary object is ncomms15010-s1.pdf (Supplementary Information:
figures + summary tables) — no machine-readable per-sample data file and no
"Supplementary Data" spreadsheet.
Code availability
None. No GitHub/GitLab/Zenodo code link, no "Code availability" statement. (P16 — applying a third-party tool would be equally valid, but there is no obtainable input dataset to apply any tool to.)
Drop reasons (controlled vocabulary, SCREENING.md)
non_pipeline(primary) — not actually a computational-pipeline reproduction; the quantitative analyses are vendor-software lipidomics + GUI statistics over wet-lab readouts.data_restricted(compounding) — all data "available from the corresponding authors upon reasonable request"; no public accession.no_code(compounding) — no analysis-code repository or statement.
No «our HPC» compute was submitted: there is no public dataset to fetch and no pipeline to run. Submitting a job would have produced no genuine reproduction.
Auditor note
The signed-in operator («email») is co-author Schlein C on this paper. This is recorded for transparency; it does not change the independent reproducibility verdict — for any external party the data remain on-request and no pipeline/code is shipped.
Assessments & scoring basis
Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.
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.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
This is a correctly-classified DROP: Bartelt et al. 2017 (Nat Commun, 10.1038/ncomms15010) is in vivo metabolic physiology plus vendor-software targeted lipidomics with no scripted pipeline (non_pipeline), no public data accession (data_restricted, data on-request), and no code (no_code). Nothing was reproduced and no value mismatch was observed — the failure is one of feasibility/data-availability, not an authors' defect or fabrication, so per the rubric q1/q2/q4 are red (data not obtainable, no comparable endpoint) while derivability, core-claim and overall stay yellow (uncheckable) rather than red. The SI is only a figures+tables PDF with no machine-readable per-sample values, so even the published t-test + BH-FDR statistics (C4) cannot be re-run. Net: legitimate drop with no evidence for or against the central HDL-turnover conclusion.
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-4-8Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.