Gut Microbiome Influences Efficacy of PD-1-Based Immunotherapy against Epithelial Tumors
2018 · paper · Biomedicine
Finds that gut microbiome composition affects response to PD-1 immunotherapy. Antibiotics reduced clinical benefit in cancer patients, transplants from responders improved treatment in mice, and the bacterium Akkermansia muciniphila was linked to responses and restored efficacy in mice.
4 findings
Reference or label Fig. 2 legend (A): "Data are means ± SEM of counts for patients experiencing PFS shorter or longer than 6 months", but both Fig. 2A panels are notched box-and-whisker plots.
- p. 3 Fig. 2 legend, (A)
Data are means ± SEM of counts for patients experiencing PFS shorter or longer than 6 months.
- p. 3 Fig. 2A, Gene Count and MGS count panels
Gene Count
- p. 3 Fig. 2 legend, panel A
Data are means ± SEM of counts for patients experiencing PFS
Judge's note: Looked at the Fig. 2A crops and the p3 page image. Both the gene-count and MGS-count panels show boxes with a median line and notches, dashed whiskers and separate grey outlier points; no mean marker or SEM bar is drawn. The legend describes a different kind of plot from the one printed.
Second check: another model agreed
Report a problem with this finding
Reference or label Fig. 3 legend: "(C) Means ± SEM of tumor sizes after Iso Ctrl mAb" and "Data are means ± SEM", but the error bars in the pooled Fig. 3C panels span roughly the middle of the scatter (about 105 to 200 mm² for Iso Ctrl, NR, with several dozen mice).
- p. 4 Fig. 3 legend, (C)
Means ± SEM of tumor sizes after Iso Ctrl mAb
- p. 4 Fig. 3 legend, last lines
Data are means ± SEM.
- p. 4 Fig. 3C, Iso Ctrl and αPD-1 pooled panels
Iso Ctrl
Only the authors can say what is right here.
Judge's note: Looked at Fig. 3C on the p4 page image; the listed crop C009-c shows Fig. 3A and 3B, not panel C, so the page image was used. The pooled Iso Ctrl NR group has several dozen points from about 20 to 300 mm² and bars from about 105 to 200 mm² around a mean near 150; the αPD-1 NR bars run from about 55 to 160. Bars of about ±50 mm² with this many animals are the size of a standard deviation, not a standard error, which would be several times narrower. Fig. 3D and 3E bars are similarly wide relative to their scatter, whereas Fig. 4C on p5 has narrow bars for similar scatter. Values are read off the plot and are approximate, but the gap is far larger than reading error. Only the authors can say what the bars are (SD or another range).
Second check: another model agreed
Report a problem with this finding
Reference or label The scale bar in the lower-right micrograph of Fig. 4F is labelled "100μM" (micromolar) where a length, 100 μm, is meant.
- p. 5 Fig. 4F, scale bar in lower-right micrograph (Foxp3, αPD-1 & Akkermansia)
100μΜ
- p. 5 Fig. 4F, scale bar in the lower-right Foxp3 micrograph
100μM
Judge's note: Looked at the Fig. 4F crops and an enlargement of the page image: the red label above the scale bar ends in a capital M, and the text layer also carries a capital letter after the μ. A scale bar on a micrograph is a length, so the unit as printed is wrong. The lettering is small, but the capital is legible in the crops.
Second check: another model agreed
Report a problem with this finding
Wording Fig. 4 legend (H): "each symbol represents one tumor and each group comprises five mice. One representative experiment out of four is shown", but the right panel (FMT NR L1) shows about 10 or 11, 6, 5 or 6 and 5 symbols in its four groups.
- p. 5 Fig. 4 legend, (H)
each symbol represents one tumor and each group comprises five mice. One representative experiment out of four is shown.
- p. 5 Fig. 4H, right panel (FMT NR L1)
NR L1
- p. 5 Fig. 4 legend, panel H
represents one tumor and each group comprises five mice
- p. 5 Fig. 4H, right panel (FMT NR L1)
p = 0.06
Only the authors can say what is right here.
Judge's note: Counted symbols on the Fig. 4H crop. Left (SPF) panel: five symbols in each of the four groups, as the legend says. Right panel: about 10 or 11 circles (αPD-1), 6 squares (αPD-1 + αIL-12), 5 or 6 triangles (αPD-1 + Akkermansia), 5 diamonds. The first two groups plainly exceed five. No statement in the article explains this: pooling is excluded by "one representative experiment", and two tumors per mouse is not described. The authors would need to say whether the right panel pools experiments or the group sizes differ.
Second check: another model agreed
Report a problem with this finding
7 minor (typos and the like; most checked by the judge only)
Wording Abstract: "Oral supplementation with A. muciniphila ... restored the efficacy of PD-1 blockade in an interleukin-12–dependent manner by increasing the recruitment of CCR9+CXCR3+CD4+ T lymphocytes into mouse tumor beds", but the tumor-bed panel (Fig. 4E) and the text ("elicited by oral gavage with a combination of A. muciniphila and E. hirae") show this only for the two-species mixture.
- p. 1 Abstract
Oral supplementation with A. muciniphila after FMT with nonresponder feces restored the efficacy of PD-1 blockade in an interleukin-12–dependent manner by increasing the recruitment of CCR9+CXCR3+CD4+ T lymphocytes
- p. 6 Results text, column 2
elicited by oral gavage with a combination of A. muciniphila and E. hirae
- p. 5 Fig. 4E x-axis label
Akkermansia
Only the authors can say what is right here.
Judge's note: Checked the abstract (p1 text), the results text (p6) and Fig. 4E (p5). The only treated group in Fig. 4E is labelled "Akkermansia & E.hirae"; the legend says animals were "compensated with a mixture of A. muciniphila and E. hirae". Fig. 4D (mesenteric lymph node, CCR9 only) has an Akkermansia-alone group, and Fig. 4H (IL-12) uses Akkermansia alone. The text also cites fig. S14, B and C for tumor beds; the supplement is not among the files, so an A. muciniphila-only tumor-bed panel there cannot be ruled out. Within the article, the abstract's attribution is wider than the data shown.
Second check: another model agreed
Report a problem with this finding
Typo The same French term is spelled "Equipe Labellisée–Ligue Nationale contre le Cancer" in affiliations 2 and 10 (p1) and "Ligue contre le Cancer (équipe labelisée)" in the acknowledgments (p7).
- p. 1 Affiliations 2 and 10
Equipe Labellisée–Ligue Nationale contre le Cancer
- p. 7 Acknowledgments
Ligue contre le Cancer (équipe labelisée)
- p. 7 Acknowledgments
(équipe labelisée)
Judge's note: Both spellings read on the crops of p1 and p7: double l in the affiliations (twice), single l in the acknowledgments. An inconsistency inside the paper; the affiliation spelling is the one used twice. The missing accent in "Medicale" mentioned in the same note is handled under C003.
Second check: none (judge only)
Report a problem with this finding
Typo Affiliation 2 prints "Institut National de la Santé et de la Recherche Medicale (INSERM)" without the accent, while the acknowledgments print "Fondation pour la Recherche Médicale".
- p. 1 Affiliation 2
Institut National de la Santé et de la Recherche Medicale (INSERM)
Judge's note: Read on the p1 affiliation crop: "Santé" is accented and "Medicale" is not, in the same line. The p7 acknowledgments crop shows "Médicale". Accents are kept elsewhere in the affiliations (Université, Expérimentale, Cité), so this is not a house style.
Second check: none (judge only)
Report a problem with this finding
Typo "RECIST1.1" is printed without a space in the results ("as assessed by RECIST1.1", p2) and in the Fig. 2 legend ("RECIST1.1 criteria", p3), against "RECIST 1.1" elsewhere on p2.
- p. 2 Results, 2nd paragraph
as assessed by RECIST1.1
- p. 3 Fig. 2 legend (B and C)
RECIST1.1 criteria
- p. 2 Results, 1st paragraph
based on RECIST 1.1 criteria
Judge's note: Checked on the p2 crop and the p3 page image: no space in either instance, while the p2 text has "based on RECIST 1.1 criteria" and "according to RECIST 1.1" with a space. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo The same MGS is labelled "CAG00168, 2534, Clostridiales bacterium VE202-14" in Fig. 2B and "CAG00168, 2534, Clostridiales VE202-14" in Fig. 2C.
- p. 3 Fig. 2B
CAG00168, 2534, Clostridiales bacterium VE202-14
- p. 3 Fig. 2C
CAG00168, 2534, Clostridiales VE202-14
Judge's note: Both labels read on the Fig. 2B and 2C crops; identical CAG identifier and gene count, with "bacterium" present in B and absent in C. Other labels shared by the two panels (for example CAG00211, CAG00141) match. A small labelling inconsistency; the entity is not in doubt.
Second check: none (judge only)
Report a problem with this finding
Typo Fig. 3 legend (C) prints "after PD1 mAb (center and right)" and "before PD1 mAb" without the hyphen, against "PD-1 mAb" in parts (A) and (B) of the same legend.
- p. 4 Fig. 3 legend, (C)
after PD1 mAb (center and right)
- p. 4 Fig. 3 legend, (C)
before PD1 mAb
- p. 4 Fig. 3 legend, (A)
PD-1 mAb was injected intraperitoneally
Judge's note: Read on the p4 legend crop: "PD1 mAb" twice in part (C), "PD-1 mAb" in parts (A) and (B) and in the rest of the paper. Cosmetic inconsistency.
Second check: none (judge only)
Report a problem with this finding
Typo The same marker is written "Foxp3" in the results text ("CD4/Foxp3 ratios") and on the Fig. 4F micrograph labels, and "FoxP3" in the Fig. 4 legend ("CD4 and FoxP3 infiltrates") and on the Fig. 4G axis ("CD4+/FoxP3+ ratio").
- p. 6 Results text, column 2
CD4/Foxp3 ratios in tumors
- p. 5 Fig. 4 legend, (F and G)
CD4 and FoxP3 infiltrates
- p. 5 Fig. 4F micrograph label and Fig. 4G axis
Foxp3
- p. 6 Results text
increased CD4/Foxp3 ratios
Judge's note: Read on the p5 figure crop (micrograph labels "Foxp3", axis "CD4+/FoxP3+ ratio"), the p5 legend crop and the p6 page image. Both capitalisations appear for one marker in one experiment. Both forms are in use in the literature, so this is a consistency point only.
Second check: none (judge only)