Reponses to Brian D Lehmann’s comments:
Comments: While a link between wnt signaling and TNBC is interesting, the authors do not
sufficiently demonstrate this with the data included in this manuscript.
Response: The revised MS provides the following data in the support of the conclusion (newly
added data are underlined):
 Micro-array based data from different cohorts (Fig 1)
 Transcript data from different cohorts (QC, GA, MSKCC-99, UNCCH-186, Stockholm159) (Fig.2)
 Cell line based phenotype (integrin directed wound healing assay and invasion assay)
data (Fig. 3A)
 Integrin-directed transwell migration assay, time course of protein expression of betacatenin following beta-catenin SiRNA (Fig. 3B)
 Real time movie to demonstrate the inhibition of the movement of live cells following wnt
modulators (Supplementary AVI files)
 Cytoskeletal rearrangement of actin by confocal microscopy (Fig. 3C i, ii, & iii) following
wnt-pathway modulator, sulindac sulfide in two TNBC cell lines; Fig. 3Ciii is a new data
 Effect of genetic (SiRNA) and pharmacological modulators (Sulindac sulfide) of wnt
pathway on beta-catenin protein (Western blot), its downstream effecter (active beta –
catenin), and transcriptional activity (TOP Flash) in TNBC cell (Fig. 3D); Fig. 3D is a
new data
 Survival data, relapse data, metastasis data including metastasis data from lung and
brain (Fig.4)
 Survival, recurrence and metastasis data (Fig. 6A-E)
 DASL assay reproducibility and concordance data (Fig. S1 A-L)
 Canonical pathways overrepresented in TN/basal like subtypes in QC, GA, MSKCC-99,
UNCCH-186, Stockholm-159 (Fig. S2 A-F)
 Canonical pathway by subtypes in QC, GA, MSKCC-99, UNCCH-186, Stockholm-159
(Fig. S3 A-F)
 KEGG Wnt pathway expression by subtypes in QC, GA, MSKCC-99, UNCCH-186,
Stockholm-159 (Fig. S4 A-F)
 Experimental pathway regulation by subtype in QC, GA, MSKCC-99, UNCCH-186,
Stockholm-159 (Fig. S5 A-F)
 Experimental wnt-beta-catenin expression by subtype in QC, GA, MSKCC-99, UNCCH186, Stockholm-159 (Fig. S6 A-F)
 Wnt-beta-catenin expression by classifier type in QC, GA, MSKCC-99, UNCCH-186,
Stockholm-159, EMC-192, EMC-286, GGI-327, UNCCH-73, UNCCH-67, Uppsala-249
(Fig. S7 A-K)
 Wnt+ patient meta-Analysis (Fig. S8)
Comments: They provide minimal evidence for their findings using a mixture of two cell lines
(MDA-MB-231 and HCC38) for functional studies evaluating migration and invasion in WntC59
treated cells and B-catenin knockdown
1
Response: In the revised MS, we have added separate independent data for both cell lines (Fig
3Ciii and Fig 3D) demonstrating the effect of genetic (SiRNA) and pharmacological modulators
(Sulindac sulfide) of wnt pathway on beta-catenin protein (Western blot), its downstream
effecter (active beta –catenin), and transcriptional activity (TOP Flash) in two TNBC cells.
Comments: There are formatting and organizational errors that detract from the science
reported.
Response: We are sorry for the confusion. In the revised MS, we have corrected the figures.
Specific comments
Comments (1) The use of wnt (+) is not clearly or uniformly used throughout the manuscript
and is essential to understanding any conclusions. Sometimes wnt(+) refers to the LWS-81,
sometimes it is the shrunken centroid of the WNT/b-catenin classifier. This use becomes
increasingly important in understanding the kaplan-meier graphs, as figure 1 shows there are
more than just two phenotypes of wnt signaling and these differ between datasets (figure 1A
and B).
Response: In the revised MS, we have corrected the legend to figure 1 (as shown below):
Figure 1: Unsupervised hierarchical clustering of Wnt3A induced genes from the LWS-81
signature (shown in the figure as Wnt+) (Nguyen, Chiang et al. 2009) common to the Breast
Cancer DASL panel segregated triple negative tumors (shown in black) from other subtypes in
both A) Quebec (QC-BCP) and B) Georgia (GA-BCP) cohorts. Each row represents a transcript,
and each column represents a tumor sample, with triple negative denoted in black, HER2+ in
grey, and HR+ in white. Yellow indicates genes upregulated in response to Wnt3A and blue
indicates genes down regulated. High expression is represented by red and low expression by
green.
Also, in order to maintain uniformity throughout the MS, we have replaced Wnt(+) with “Wnt
classifier signature”.
In the revised MS legend to figures 4 and 6 are changed and are presented below:
Figure 4: Kaplan-Meier survival curves of Wnt/β-catenin positive (Wnt+: Wnt classifier signature)
and Wnt/β- catenin negative (Wnt-) patients with respect to A) overall survival, B) recurrencefree survival, C) metastasis-free survival, D) lung metastasis-free survival, E) brain metastasisfree survival, and F) lung metastasis-free survival in TNBC patients. P-values were calculated
by the log-rank test.
Figure 6: Kaplan-Meier survival analyses for i) overall, ii) recurrence-free, iii)
Metastasis-free, iv) lung metastasis-free, v) brain metastasis-free, and vi) bone metastasis-free
survival for patients classified as Wnt+ (Wnt classifier signature) and Wnt- for A) all patients and
patients stratified by B) pathological determined subtype, C) intrinsic subtype, D) histological
grade, and E) lymph node status. Only significant (p<0.05) p-values are shown and are colorcoded to correspond to the respective stratum.
Comments (2) Discretionary Revisions-Throughout the manuscript TNBC and basal-like are
used interchangeable, however these are not synonymous and describe two separate patient
cohorts. It is true that 80% of TNBC display basal-like gene expression, so this grouping is at
best 80% correct.
Response: We are sorry for the confusion. In the revised MS, we have corrected the issues
appropriately. The QC and GA patient populations are indeed triple negative. In case of cell
lines, they are now referred to as “TNBC cell lines”.
2
Comments (3) Methods RNA, preparation, quality control and DASL assay. Minor Essential
Revisions-“we have added a supplementary table with the gene-list… to the revised-MS S10
and S11 for the convenience of the reader” should be removed and included in the response to
reviewer’s comments in the previous submission. I think that supplemental S3 and S4 are these
figures, as there are no S10 and S11 tables.
Response: We are sorry for the confusion. We have removed the text “we have added a
supplementary table with the gene-list… to the revised-MS S10 and S11 for the convenience of
the reader” following the reviewer’s suggestion from the MS. They now appear as
supplementary table 1 and 2 in the revised MS.
Comments (4) Minor Essential Revisions: Figure 1: The names of the genes should be
included so reviewers/readers could make their own minds up whether they are wnt inducible
genes. This gene list came from lung cells and there may be tissue-specific gene expression.
Response: Unsupervised hierarchical clustering of Wnt3A induced genes from the LWS-81
signature (Nguyen, Chiang et al. 2009) common to the Breast Cancer DASL panel segregated
TN tumors (shown in black) from other subtypes in both A) Quebec (QC-BCP) and B) Georgia
(GA-BCP) cohorts.
Comments (5) Major Compulsory Revisions -Overall it is very unclear as two which population
has activated wnt. Figure 1 demonstrates that there two populations of WNT (–) in the QC-BCP
group and two populations of WNT (+) in the GA BCP group. These clusters appear to display a
mix of wnt activated and repressed genes.
Response: In order to clear the confusion we have added the following text in the revised MS
(page 12)…. “This Wnt classifier signature (mentioned in figure as Wnt+) is composed of 81
genes identified by treating lung cancer cell lines with the Wnt3A ligand (hereafter LWS-81)”
Also, in the revised MS, we have added a separate section on the details of analysis and
statistical methods (Supplementary analysis and statistical methods) which addresses the
reviewer’s concern. We have added the following text…. “The details of the analyses and
statistical methods are presented in the supplementary section as “Supplementary analysis and
statistical methods” in our Method section of the revised MS.
Comments (6) Major Compulsory Revisions- It is unclear on how the Wnt expression is
quantified. Is this just the aggregate expression of all 81 genes in the LWS dataset? This could
be biased based on the number of genes activated vs. repressed.
Response: Lung cancer Wnt induced (LWS-81) genes (Nguyen, Chiang et al. 2009) used to
assess Wnt transcriptional regulation in the Quebec cohort profiled on the A) human cancer
panel (QC-HCP) and B) breast cancer panel (QC-BCP) as well as C) Georgia (GABCP), D)
MSKCC-99 (Minn, Gupta et al. 2005), E) UNCCH-186 (Parker, Mullins et al. 2009), and F)
Stockholm-159 (STH-159) cohorts (Pawitan, Bjohle et al. 2005). Significant p-values after
Bonferroni’s correction are shown relative to the TN or basal-like subtypes.
In the revised MS, we have added a separate section on the details of analysis and
statistical methods (Supplementary analysis and statistical methods) which addresses the
reviewer’s concern. We have added the following text…. “The details of the analyses and
statistical methods are presented in the supplementary section as “Supplementary analysis and
statistical methods” in our Method section of the revised MS.
Comments (7)The finding that the Wnt pathway is overrepresented in the DASL cohorts may
be a reflection of the biased selection of cancer genes, as the panels are quite small,
representing approximately 500 genes.
Response: The details about the 500 genes are already published in our article (PMID:
18361796, PMID: 22067903).
3
Comments Figure 3. (8) Minor Essential Revisions- Figures need to be compressed into one
page.
Response: In the revised MS, wherever possible we have combined the sub-figures into one
page.
Comments Major Compulsory Revisions (9) Appears that majority of experiments were
performed in MDA-MD-231 cells, however the authors switched to HCC38 cells for the final
panel (Ci and Cii) to evaluate cytoskeletal changes in F-actin.
Response: In the revised MS, we have provided cytoskeletal changes in F-actin in both
HCC38 (Fig 3Ci & ii) and MDA-MB231 cells (Fig 3Ciii). Fig 3Ciii is a new figure.
Comments (10)The comparison between Ci and Cii appears to be at different magnifications
and have few representative cells (2 to 4) to describe large phenotypic changes.
Response: We have provided “bars” to indicate specific magnification for each picture since we
have different magnifications for them. Within that small field at that high magnification, it is not
possible to accommodate many cells. We chose to use higher magnification and hence a lesser
number of cells only to demonstrate the cytoskeletal changes with an individual cell.
Comments (11) The use of additional cell lines treated to all of the conditions (#-catenin
knockdown, WNtC59 and sulindac sulfide treatment) would provide evidence for the
conclusions drawn.
Response: In the revised MS, following the reviewer’s suggestions, we have added a new
figure to demonstrate the effect of genetic (SiRNA) and pharmacological modulators (Sulindac
sulfide) of wnt pathway on beta-catenin protein (Western blot), and its downstream effecter
(active beta –catenin) in TNBC cell (Fig. 3D left panel).
Comments (12) The authors should consider showing that the wnt pathway was successfully
inhibited under all conditions by using a wnt luciferase reporter (TOPflash) or RNA/protein
analysis of downstream targets.
Response: We totally agree with the reviewer. In the revised MS, we have added a new figure
to demonstrate the effect of genetic (SiRNA) and pharmacological modulators (Sulindac sulfide)
of wnt pathway on wnt –transcriptional activity in TNBC cell (Fig. 3D right panel).
Compiling the new results, we have added the following text in the discussion of the MS (page
18):
We have studied the cause-effect relationship of Wnt-β-catenin pathway with metastasis in
TNBC cell line models using three tools, genetic (beta-catenin SiRNA), pharmacological (Wnt-βcatenin pathway modulator; Wnt-C59), and functional (sulindac sulfide). Our data show that
perturbation of the Wnt-β-catenin pathway abrogated metastasis-associated phenotypes in
TNBC cells following attenuation of beta-catenin transcriptional activity, proving a direct
mechanism based relationship between Wnt-β-catenin pathway and metastasis in TNBC.
Comments (13) Major Compulsory Revisions What are wnt (+) and wnt (-) in the analysis? This
analysis is entirely dependent of that classification and it is clear from figure 1 that there are
multiple clusters with mixed wnt activated and repressed genes.
Response: To clear the confusion, we have added a separate section on the details of analysis
and statistical methods (Supplementary analysis and statistical methods) in the revised MS,
which addresses the reviewer’s concern. We have added the following text…. “The details of
4
the analyses and statistical methods are presented in the supplementary section as
“Supplementary analysis and statistical methods” in our Method section of the revised MS.
Comments (14) Minor Essential Revisions- This figure does not add significantly to any of the
findings and should not be included.
Response: We partially agree with reviewer. However, respecting the previous reviewer’s as
well as other reviewer’s comments and keeping in mind how a diagrammatic representation
sometimes convey the story in a more lucid way for the new readers in this field, we have kept
the figure in the MS.
However please note that if the editor of the journal feels strongly about the removal of the
figure, we will be glad to remove it from the MS.
Comments (15) Minor Essential Revisions- This figure appears to be labeled S10. Figure
should also be formatted as to not include graphs with NA, as they add no information and are
distracting.
Response: In order to maintain an internal consistency and to provide a complete picture
coming out of the data we decided to keep the figure in its original form.
Comments (16), (17), (18), (19) and (20)
Response: To clear the confusion, we have added a separate section on the details of analysis
and statistical methods (Supplementary analysis and statistical methods) in the revised MS,
which addresses the reviewer’s concern. We have added the following text…. “The details of
the analyses and statistical methods are presented in the supplementary section as
“Supplementary analysis and statistical methods” in our Method section of the revised MS.
Comments (21) Major Compulsory Revisions-This figure does not appear to add anything and
actually demonstrates that the wnt (+) signature extends across all subtypes of breast cancer
contrasts with your overall conclusion that this TNBC specific.
Response: Following reviewer’s comments, we have removed the figure from the revised MS.
5
Reponses to Caterina Marchiò’s comments:
We highly appreciate reviewer’s comments.
Comments: It is true that the authors have added functional work, however please note
that the functional analysis was performed on one cell line only, without control cell lines
(such as not triple negative to compare the effect). Therefore I would suggest something
like “The results of the study MAY have implications for therapeutic target
identification….” And possibly mention that the functional data would benefit from
validation in other in vitro models. It is a minor amendment that can depict the scenario
in a more realistic way according to my opinion.
Response: We totally agree with the reviewer. We have embellished the revised MS now with
independent data from two TNBC cell lines in the place of one, using both genetic (SiRNA) and
pharmacological modulators (Sulindac sulfide) of wnt pathway and testing their effects on betacatenin protein (Western blot), its downstream effecter (active beta –catenin), and
transcriptional activity (TOP Flash) in TNBC cell (Fig. 3D) to substantiate our phenotypic data.
Following reviewer’s suggestion, we have added the text…….“The results of the study
MAY have implications for therapeutic target identification….” & also mentioned that
“the functional data would benefit from validation in other in vitro models” in the
discussion of the revised MS.
6