Animal studyindependent cross-checkfull text checked I-INTB-04 in-depth review
Chemotherapy can activate tumour 'exit doors' in mice
Neoadjuvant chemotherapy induces breast cancer metastasis through a TMEM-mediated mechanism.
Overview
Karagiannis GS et al.. Sci Transl Med 2017 · PMID 28679654 · DOI 10.1126/scitranslmed.aan0026
In mice and in residual human tumours, chemotherapy increased markers of cancer-cell dissemination.
Does the study support the specific claim? Partly · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Test whether chemotherapy promotes metastasis
- How was the study done?
Mouse models and human residual tumour tissue
- How many people, animals or samples were studied?
Mice (about 11 per group) and 20 patients with residual tumour after chemotherapy
- What was actually found?
More TMEM activity and metastasis in mice; higher TMEM scores in humans
- How large was the observed effect?
2-3-fold higher TMEM score and about 2-fold more circulating tumour cells in mice; TMEM score increased in all 20 patients (no outcome data)
- What do the results mean in everyday life?
Chemotherapy can have unwanted biological effects, which researchers try to block with added drugs.
- What are the study's limitations?
Mainly mice; human part small, selected and without outcome data; authors hold a related patent
- What does it say about Elawady's specific claim?
Partly supports 'chemo makes cancer more aggressive' as a biological possibility; does not show worse survival.
- Do the results apply to Germany and Europe?
Moderate
- What can explicitly not be concluded from this study?
Cannot conclude that chemotherapy shortens life or should be avoided.
Scientific in-depth review
1. Full reference
- Authors
- Karagiannis GS et al.
- Journal
- Sci Transl Med
- Year
- 2017
- Identifier
- PMID 28679654 DOI 10.1126/scitranslmed.aan0026
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
Erratum published (PubMed ErratumIn PMID 28724576)
- Extent checked
- full text
- Conflicts of interest and funding
NIH grants (CA100324, CA150344, 1T32CA200561-01), SIG 1S10OD019961-01, institutional programmes; rebastinib provided by Deciphera Pharmaceuticals. Competing interests: two senior authors (J.S.C., M.H.O.) are inventors on a patent application covering methods of detecting and reducing chemotherapy-induced prometastatic changes in breast tumours.
2. Research question
Does chemotherapy change tumour microenvironment sites of cancer cell dissemination (TMEM) and promote metastasis?
3. Study design and population
- Study design
- Animal study – Preclinical experimental study (MMTV-PyMT transgenic and syngeneic transplant mice, two patient-derived xenograft models; randomised allocation, blinded readouts, intravital imaging) plus small descriptive pre/post analysis of human breast cancer tissue
- Population
Mice with early-stage mammary carcinomas (~11 per group planned; up to 25% removed per panel); 20 patients with ER+/HER2- breast cancer and residual disease (RCB 2-3) after neoadjuvant weekly paclitaxel followed by 4 cycles of doxorubicin plus cyclophosphamide; separate fine-needle aspirates from 5 patients before and 1 week after the second paclitaxel dose
- Country
USA (Albert Einstein College of Medicine, New York)
- Period
not reported in full text (human cohort had 5 years of follow-up, insufficient for recurrence analysis)
- Sample size
Mice: sample size 22 per two-arm experiment (11/group) planned; humans: 20 paired pre/post-NAC samples, plus 5 FNA patients
- Inclusion and exclusion criteria
Mice entered at tumour diameter ~2-3 mm; excluded necrotic/cystic tumours, inadequate tissue, <0.5 ml blood, obese (>33 g) or emaciated (<17 g) mice. Patients: ER+/HER2-, residual cancer burden 2-3 after NAC, no preoperative tamoxifen.
4. Methods and endpoints
- Exposure / intervention
Paclitaxel (10 mg/kg every 5 days x3) or doxorubicin/cyclophosphamide in mice; with or without TIE2 inhibitor rebastinib; MENA knockout tumours. Human: neoadjuvant paclitaxel then doxorubicin plus cyclophosphamide
- Comparator
Vehicle-treated mice; pre-NAC biopsies in the same patients
- Endpoints
TMEM score, perivascular TIE2hi/VEGFhi macrophages, TMEM-associated vascular 'bursting', circulating tumour cells (CTCs), lung micrometastases, MENA isoform expression; human TMEM score and MENAINV
- Statistical methods
Mann-Whitney U, Wilcoxon signed-rank, Pearson correlation, chi-square; sample size for 1.2 SD difference (power 0.8, alpha 0.025); P values <0.1 reported, <0.05 described as significant.
5. Actual results
- Results
Mice: paclitaxel delayed tumour growth but increased TMEM score 2-3-fold (P<0.001), perivascular TIE2hi/VEGFhi macrophages, TMEM bursting, extravascular dextran ~3-fold (P<0.05), CTCs ~2-fold (P<0.05) and lung micrometastases; doxorubicin/cyclophosphamide increased TMEM (P<0.05) and CTCs ~1.3-fold (P=0.027). MENA knockout abolished CTCs and lung dissemination; rebastinib reduced paclitaxel-induced CTCs to control levels and abolished bursting. Humans (n=20): mean TMEM score increased after NAC (Wilcoxon P<0.0001); 50% moved from low/intermediate to high-risk TMEM category; no patient decreased; MENAINV-positive area increased (P<0.01); MENAINV expression rose in 4 of 5 FNA patients after two paclitaxel doses.
- Absolute risks
not applicable; no clinical outcome (metastasis or survival) data in humans
- Authors' conclusion
Chemotherapy increases TMEM assembly/activity and invasive MENA isoform expression and promotes dissemination in mouse models, with similar marker changes in residual human tumours; TMEM score and MENA isoforms may predict prometastatic changes, and TMEM inhibitors (e.g. TIE2 inhibitors) may improve clinical benefits of chemotherapy.
6. Strengths
- Multiple mouse models incl. patient-derived xenografts and two chemotherapy regimens
- Randomised allocation of animals; blinded pathology, imaging, CTC and expression analyses
- Mechanistic confirmation (MENA knockout, TIE2 inhibition)
- Paired human pre/post tissue
7. Weaknesses
- Primarily animal data; dosing schedules differ from clinical practice
- Human analysis: only 20 selected patients with residual disease (non-responders), unblinded TMEM scoring, surrogate markers only, no recurrence or survival analysis (follow-up insufficient)
- Does not address clinically detectable metastases (authors' limitation)
- Up to 25% of animals removed post hoc per panel by predefined criteria
- Many comparisons; P values <0.1 reported
- Authors hold a patent related to the tested diagnostic/therapeutic concept; drug supplied by manufacturer
8. Risk of bias
- Orienting tool
- No standard tool fits (preclinical animal study with small human tissue series); structured validity judgement (SYRCLE-oriented for the animal part)
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Overall unchanged. Internally reasonably rigorous preclinical work, but very indirect for clinical decisions: it does not show that chemotherapy worsens survival, which randomised trials of adjuvant/neoadjuvant chemotherapy contradict. Simplified single-reviewer judgement; no formal tool applied.
| Domain | Judgement | Rationale |
|---|---|---|
| Allocation and blinding (animal experiments) | low | Random allocation to treatment groups; outcome assessors blinded. |
| Attrition/exclusions (animal experiments) | some concerns | Predefined exclusions but up to 25% of animals removed per panel. |
| Selection and measurement (human tissue) | high | Only patients with residual disease after NAC (n=20); TMEM scoring could not be blinded for pre vs post samples. |
| Outcome relevance (directness) | high | Surrogate markers (TMEM, CTCs, micrometastases in mice); no human clinical outcomes. |
| Selective reporting / conflicts | some concerns | Multiple endpoints and models; patent interest of senior authors in the concept. |
Full-text review (25 October 2026): Added country, sample sizes (11 mice/group; 20 patients with residual disease; 5 FNA), regimen sequence (paclitaxel then AC), quantitative results, blinding/randomisation details, exclusions, absence of human outcome data, funding and patent-related COI. Bias overall unchanged with structured domains; directness set to very_indirect.
9. Assessment of the statistical interpretation
Effect sizes are fold-changes in surrogate markers in mice and a paired change in TMEM score in 20 selected patients; no estimate of metastasis or mortality risk in humans.
10. Relevance to the specific claim
Gives a real kernel to the idea that chemotherapy can have pro-metastatic side effects.
11. Independent contrary findings
Randomised trials show chemotherapy reduces breast-cancer mortality by about one-third (I-CHM-01).
12. Transferability to Germany and Europe
Mechanism applies to breast cancer biology generally.
13. Overall methodological judgement
Mechanistic preclinical signal that chemotherapy can induce pro-metastatic changes in the tumour microenvironment, proposed to be countered by adding TMEM/TIE2 inhibitors or adapting treatment; not evidence that chemotherapy should be avoided.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | not applicable |
| Directness | very indirect |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | unknown |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Whether NAC-induced TMEM increases translate into more distant recurrence or lower survival; trials of TIE2 inhibitors (e.g. rebastinib) with chemotherapy.