Author's response to reviews Title: Wnt1 is epistatic to Id2 in modeling mammary gland development and in causing mammary tumors Authors: Susan Marino (smarino@cmgm.stanford.edu) Claire Romelfanger (clairefrog@yahoo.com) Yoshifumi Yokota (yyokota@fmsrsa.fukui-med.ac.jp) Roel Nusse (rnusse@cmgm.stanford.edu) Version: 2 Date: 13 November 2004 see over Author's response to reviews:
We have revised the paper according to the reviewer s comments. Below a point by point reply to the reviewers. Replies in BOLDFACE. Reviewer: Caroline Alexander Reviewer's report: 1. The number of mice in the key cohort, the Id-/-:Wnt1 mice, is very small (8). This is not usually considered enough animals to draw a tumor survival curve. The authors should demonstrate there is sufficient power to say that these 2 curves are not different. (It is not clear what strain background the transgenic Wnt mice are on). WHEN WE SET OUT TO DO THE EXPERIMENT, WE AIMED AT GETTING MANY MORE MICE IN THE CRITICAL COHORT (30) BUT DUE TO UNEXPECTED LETHALITY, WE ENDED UP WITH 8. SINCE ALL OF THESE MICE DEVELOPED TUMORS, THE ANALYSIS IS STILL MEANINGFUL. HOWEVER, IT IS TRUE THAT A PARTIAL RESCUE MIGHT NOT BE DETECTED WITH THESE NUMBERS. WE HAVE ADJUSTED THE CONCLUSIONS IN THE DISCUSSION TO REFLECT THIS POSSIBILITY. WE HAVE ADDED THE STRAIN INFORMATION. 2. Lobulo-alveolar is usually used to designate the differentiation and growth typical of pregnancy. The original paper that described the Id2-/-phenotype (cited several times) used this term in just this way. That paper says there is NO phenotype in the virgin gland. This paper, in contrast, is suggesting that there is a phenotype in the virgin ductal tree. It should be described as such and not used interchangeably with the effects observed in that previous study. THE SUGGESTION THAT THERE IS A PHENOTYPE IN THE NON-TRANSGENIC ID2-/- VIRGIN DUCTAL TREE WAS AN EDITORIAL OVERSIGHT ON OUR PART. THERE IS NO PHENOTYPE IN THE VIRGIN DUCTAL TREE IN NON-TRANSGENIC ANIMALS. WE HAVE CORRECTED THIS. THE TERM PRECOCIOUS LOBULO-ALVEOLAR DEVELOPMENT HAS BEEN USED IN THE PAST TO DESCRIBE THE WNT1 PHENOTYPE IN THE VIRGIN GLAND (REFS.7 AND 11) BUT WE HAVE UPDATED THE TERMINOLOGY TO THAT USED BY THE REVIEWERS. BOTH LOBULO ALVEOLAR DEVELOPMENT AND DUCTAL SIDE BRANCHING REQUIRE CELL PROLIFERATION. ID2 EXPRESSION OCCURS IN BOTH GLANDULAR AND DUCTAL EPITHELIUM AND THE ID2 LACTATION DEFECT HAS BEEN ATTRIBUTED TO A PROLIFERATION PROBLEM. DUCTAL SIDE BRANCHING IN THE PREGNANT ID2-/- MOUSE HAS ALSO BEEN SHOWN AS SPARSE(REF4). THEREFORE THE ISSUE IS HYPERPLASIA VERSUS A PROLIFERATION DEFECT. THE CHANGE IN TERMINOLOGY FROM LOBULAR-ALVEOLAR DEVELOPMENT TO DUCTAL SIDE BRANCHING DOES NOT AFFECT OUR PRIMARY CONCLUSION, THAT WNT1 CAN OVERCOME THE ID2 PROLIFERATION BLOCK.
3. The pictures of the Carmine stains (or the Carmine stains themselves) are not clear, and since they comprise an important piece of data in this paper, they should be improved so that they are more informative. It is not clear whether this data is derived from one mouse or is representative. The inter-mouse scatter in phenotypes is usually substantial, and here (in an F2 cohort) is likely to be even higher. It would be much clearer to present a picture of the whole gland, together with close-ups of each important phenotype. How does the Wnt1 phenotype in the purebred strain compare to these backcrosses? WE HAVE ASSEMBLED A NEW FIGURE 2, SHOWING BETTER PICTURES OF THE GLANDS. WE HAVE ADDED THE FACT THAT THE DUCTAL PHENOTYPE AS SHOWN IN THE PANELS WAS OBSERVED IN ALL MICE EXAMINED (24 MICE, 5 GLANDS PER MOUSE). THERE IS NO INFORMATION TO BE GAINED FROM SHOWING THE WHOLE GLAND AND THIS HAS NOT BEEN DONE IN THE CITED PAPERS EITHER. IT IS STATED IN OUR TEXT THAT THE WNT 1 PHENOTYPE IN THE BACKCROSSES WAS THE PREVIOUSLY DESCRIBED PHENOTYPE, BUT, AS STATED ABOVE, WE USED THE ORIGINAL TERMINOLOGY TO DESCRIBE IT. THIS HAS BEEN CHANGED NOW. ACCORDING TO THE VARMUS LABORATORY FROM WHICH WE OBTAINED THE TRANSGENIC MICE, STRAIN BACKGROUND DOES NOT AFFECT THE WNT1 TRANSGENIC PHENOTYPE. THE PHENOTYPE WE OBSERVED MATCHED THAT DESCRIBED IN THE LITERATURE. 4. The description of Fig. 2 refers to 6 types of glands, and all the discussion of this Figure is inaccurate because there are only 4 shown in this version. The last sentence of the second to last paragraph reads lead to a mammary gland that develops almost normally. This is confusing and incorrect. WE HAVE ADDED THE MISSING GLANDS AND REWORDED THE STATEMENT AS FOLLOWS: THEREFORE, IT APPEARS THAT FORCED EXPRESSION OF WNT1 IN VIRGIN MAMMARY GLANDS CAN OVERCOME THE ABSENCE OF ID2 AND LEAD TO A HIGHLY BRANCHED DUCTAL TREE RESEMBLING THE TREE ACHIEVED NORMALLY DURING PREGNANCY. 5. The implications of this study need to be better described. The abstract has a few problems, the principal one lies with the final sentence, placing the Wnt1 targets downstream of Id2. This I think is an editorial mistake by the authors. I think they mean that if Id2 is one of the Wnt1 target genes, it is not the one that is responsible for the preneoplastic hyperplasia or tumor development. There are of course several other interpretations that could bear Discussion. THE FINAL SENTENCE WAS AN EDITORIAL MISTAKE THAT HAS BEEN CORRECTED. WE HAVE BEEN CAREFUL NOT TO OVER-INTERPRET OUR RESULTS AND THIS IS WHY THE DISCUSSION IS BRIEF.
6. The end sentence of the Results section says We concluded that Wnt1 is epistatic to Id2. This use of epistatic (here and in the Title) is not correct to my definition; it is too easily misunderstood and should be rephrased. WE HAVE NOT CHANGED THIS TERM BECAUSE WE HAVE SHOWN THAT ID2 IS NOT REQUIRED FOR THE WNT1 PHENOTYPE. THIS IS THE CORRECT TERM. 7. The end paragraph of the Discussion is misleading. The Id2 phenotype is not rescued by Wnt1. The principal Id2-/-phenotype is a lack of terminal differentiation, the minor one described here is hypomorphic ductal branching in the virgin. In contrast, here, Wnt1-induced hyperplasia (a process that is not the same as lobuloalveolar development during pregnancy) proceeds despite the absence of Id2 (see modification of Abstract above). The last sentence Wnt1 signaling is independent of Again, I think the authors mean that Wnt1-induced tumor development is not affected by the absence of Id2. There is no direct assay of Wnt1 signaling. THE REVIEWER IS CORRECT IN THAT WE DID NOT TEST THE PHENOTYPE IN PREGNANT MICE AND THEREFORE COULD NOT ASSESS TERMINAL DEVELOPMENT. WE CONCLUDED THAT WNT1 IS EPISTATIC TO ID2 IN TUMORIGENESIS, JUST AS IT IS IN PROMOTING HYPERPLASIA AND SIDE BRANCHING OF THE MAMMARY GLAND (FIGURE 2). Reviewer: Gertraud Robinson Reviewer's report: General The manuscript by Marino et al. investigates the functional relationship between Id2 and over expression of Wnt1 in mammary gland development and carcinogenesis. Id2 is a downstream target of canonical Wnt signaling in MEFs and colon cancer. To address the question whether Id2 also mediates wnt signals in mammary epithelial cells the authors crossed a mouse line than over expresses Wnt1 in mammary epithelial cells with a line in which Id2 is deleted. Over expression of Wnt1 in mammary epithelial cells causes increased side branching in virgin mice and leads to the development of tumors. Id2 deficient mammary epithelial cells display reduced proliferation due to an induction of the CDK inhibitor p21 and fail to differentiate during pregnancy. The results reported here show that the effects of Wnt over expression also occur in Id2 -/-animals, they present increased side branching in virgin mice and tumors develop at the same rate as in control mice. The entire analysis of the development of the gland consists of whole mount images, which is rather superficial -I would have been nice to show at least histological sections of the rescued Wnt1 Tg/ Id2-/-epithelia -but is sufficient to make the point that Wnt1 transgenics and Wnt1 transgenics on an Id2-/-background look similar and there is obviously no interaction in this system. In summary, these experiments are based on sound scientific reasoning. Even though the analysis of the results is executed with minimal effort they should be published. However, the results should not be over interpreted and need to be discussed in more depth. Major Compulsory Revisions (that the author must respond to before a decision on publication can be reached)
Although these data are in a sense negative data they should be published. However, the discussion is poorly written and needs substantial rewriting. For example the authors use different and rather vague formulations to interpret their findings: epistatic in the title placing Wnt1 targets downstream of Id2 in the abstract independent of Id2 in the discussion Since there is no evidence of an intersection of the genetic pathways to be gained from the data it seems appropriate to stick to independent and not to interpret the findings in genetic terms. It also cannot be excluded that in an artificial situation such as the over expression of a signaling molecule such as Wnt1, compensation through another member of the bhlh family could take place. These considerations need to be addressed in the discussion. WE THINK EPISTATIC IS THE CORRECT TERM AND HAVE NOT CHANGED IT. HOWEVER, WE HAVE BEEN CAREFUL TO SAY EPISTATIC IN WNT1 SIGNALING RATHER THAN IN LOBULO-ALVEOLAR DEVELOPMENT. PLACING WNT1 TARGETS DOWNSTREAM OF ID2 WAS AN EDITORIAL ERROR AND HAS BEEN CORRECTED. THE REVIEW SUGGESTS WE RETAIN THE TERM INDEPENDENT, AND WE HAVE. THE ISSUE OF COMPENSATION IS ADDRESSED BY REWORDING THE CONCLUSION TO SAY OUR RESULTS DEMONSTRATE THAT WNT1 IS NOT OPERATING SOLELY THROUGH ID2 OR THAT IT IS NOT OPERATING THROUGH ID2 AT ALL. Furthermore, the terminology the authors use to describe mammary gland development is incorrect. Remodeling is conventionally used to describe the events occurring during involution not as is the case here for the process of side branching. Likewise alveolar development takes place during pregnancy, a phase that is not investigated in this study. The development that is observed here should be referred to as side branching. WE HAVE OMITTED THE TERM REMODELING AND WE HAVE UPDATED OUR DESCRIPTION FROM THE PREVIOUSLY ACCEPTED PRECOCIOUS LOBULO ALVEOLAR DEVELOPMENT (REFS7,11) TO THE MORE PRECISE CURRENT DESCRIPTION SUGGESTED BY THE REVIEWER. THE PHENOTYPE OBSERVED IN THE TRANSGENIC MICE IS THE CLASSIC WNT1 PHENOTYPE IT IS NOT A NEW PHENOTYPE. Minor Essential Revisions (such as missing labels on figures, or the wrong use of a term, which the author can be trusted to correct) It is not stated clearly how the rate of tumorigenesis was measured (last paragraph of Results). What was the endpoint? What sizes were the tumors? BOTH ITEMS HAVE BEEN ADDED. The 2nd sentence of the Abstract is incomplete. THIS EDITORIAL OVERSIGHT HAS BEEN CORRECTED.