Depletion of S100A4+ stromal cells results in abnormal nipple development and nursing failure

  1. Denisa Jaros Belisova
  2. Ema Grofova
  3. Viacheslav Zemlianski
  4. Zuzana Sumbalova Koledova

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

The nipple and mammary gland are essential for the survival of mammalian offspring, providing postnatal nourishment. Their development, like that of epidermal appendages, depends on instructive mesenchymal signals. S100A4 (fibroblast-specific protein 1) is expressed by mesenchymal cells and has been implicated in the development of eccrine glands, hair follicles, and mammary branching morphogenesis. However, the identity and functional contribution of S100A4-positive (S100A4+) cells to nipple and mammary development remain unclear. Here, we used a cell-depletion mouse model, S100a4-Cre;DTA , to investigate their role during lactation. S100a4-Cre;DTA dams exhibited a severe nursing defect leading to complete litter loss within the first day postpartum. Immunofluorescence and oxytocin stimulation assays revealed no abnormalities in mammary morphology, milk production, or alveolar contractility, but defective nipple development was observed. Bulk RNA sequencing of nipple tissue indicated inflammatory signatures. Lineage tracing and immunofluorescence identified S100A4+ cells as fibroblasts and immune cells in the nipple, while only immune cells expressed S100A4 in the mammary gland. Our study uncovers a previously unrecognized role of S100A4+ cells in nipple development, highlighting their importance for successful lactation and offering new insights relevant to breastfeeding medicine.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    1. General Statements [optional]

    We thank to all reviewers on their careful consideration of our manuscript. We highly appreciate their thoughtful comments and suggestions, that helped us to improve the quality of our work. We address each comment point-by-point below.

    2. Description of the planned revisions

    __Reviewer #1 __

    Minor comments:

    Figure 5 would be more informative if it included more higher magnification images that would reveal the staining at the cellular level.

    To fulfil the suggestion, we will perform a new round of immunostaining followed by high-resolution confocal imaging. This requires additional time for laboratory work.

    __Reviewer #2: __

    Major comments

    1d. The authors tried to attribute the minor phenotype to the incomplete depletion of S100A4+ cells. However, it is possible that if the S100A4+ cells only represented a minor population, their function may be compensated by other populations. This might be confirmed by quantification of S100A4+ cells or S100A4-Cre; GFP+ cells in fibroblast or CD45 populations from images showed in Figure 5.

    We will address this comment by performing required quantifications.

    Moreover, we have now included data on the presence of S100A4+ cells in S100a4-Cre;DTA mice (Figure for Reviewers 5a,b; Supplementary Figure 7a,b in the revised manuscript), which demonstrate incomplete depletion of the S100A4+ cells in the nipple and the mammary gland. This is likely due to ongoing tissue remodeling and continuous S100A4+ replenishment/ supply. Another study using the same S100a4-Cre;DTA mouse model reported an efficient S100A4+ cell depletion in mandibular condyle (Tuwatnawanit et al., 2025), which suggests that the presence of S100A4+ cells in the S100a4-Cre;DTA mammary gland and nipple is due to tissue-specific dynamics rather than lack of depletion efficiency.

            We have included in Discussion: “Notably, we observed incomplete depletion of S100A4+ cells in the mammary gland and nipple. Interestingly, a study using the same S100a4-Cre;DTA mouse model reported complete S100A4+ cell depletion in the superficial layer of mandibular condyle46. This suggests that incomplete depletion of S100A4+ cells in nipple and mammary gland is due to tissue-specific dynamics, rather than lack of depletion efficiency, indicating a compensatory mechanism that can balance the cell loss.”

    The images in Figure 5 and Figure S4 are difficult to confirm colocalization. A higher magnification image would be required for each panel. Furthermore, a precise quantification based on the current images would be more supportive of the conclusion regarding the discrepancy of the composition of S100A4 lineage between epidermis and mammary gland (lines 163-165).

    To address this comment, we will perform a new round of immunostaining and high-resolution confocal imaging and quantifications and include the results in the fully revised manuscript.

    Line 163, the author hypothesis the Langerhans cells due to morphology. Those cells should be able to be confirmed by a co-staining with F4/80 in addition to the current form of Fig 5h.

    To address this comment, we will perform co-staining of GFP and F4/80 (or, eventually, AIF1, depending on antibody availability) and include the results in the fully revised manuscript.

    Reviewer #3

    Minor comments

    Figure 2c: The H&E images are not fully convincing. Immunofluorescence analysis of epithelial architecture would support the authors' interpretation and should be feasible if tissues are already available.

    We will perform immunostaining for epithelial markers, such as keratins, and include the results in the fully revised manuscript.

    Figure 4f: The proliferation data are compelling, but the authors could extend this by examining how cell differentiation and epithelial organisation are affected.

    We will perform immunostaining for epithelial markers (keratins, αSMA) and include the results in the fully revised manuscript.

    Figure 5b: To more convincingly show that GFP+ cells contact endothelial cells, co-labelling with an endothelial marker such as CD31 would be helpful.

    We will perform the requested co-labeling of GFP and CD31 and include the results in the fully revised manuscript.

    Figure 5f-h: The structures referenced in the text (lines 159-163) should be clearly indicated on the immunofluorescence images.

    We will incorporate these explanations into the new, high-resolution/detailed Figure 5 in the fully revised manuscript.

    3. Description of the revisions that have already been incorporated in the transferred manuscript

    Reviewer #1:

    Major comments

    It is rather difficult to conclude whether the observed nipple phenotype reflects an early embryonic/prepubertal defect in establishing the nipple stroma, is caused by a constitutive response to ongoing cell death, or a response to continuous DTA expression (or a combination of some of these).

    The data raise a couple of additional questions: Is there a nipple phenotype at 3 wk of age? It would not be totally unsurprising if ablation of a major fraction of dermal fibroblasts in the nipple area would lead to an early embryonic/prepubertal phenotype but there is no data on this. Hence, is there a "congenital" nipple deformity, as concluded by the authors (line 191)?

    We appreciate the reviewer’s insightful comments. We have now included data on embryonic nipple development. These data demonstrate abundant S100A4-lineage cells in E15.5 and E18.5 skin of *S100a4-Cre;mT/mG *embryos (Figure for Reviewers 1a, corresponding to Figure S3a in the revised manuscript) and normal appearance of nipple sheath in *S100a4-Cre;DTA *embryos at E18.5 (Figure for Reviewers 1b, corresponding to Figure S3b in the revised manuscript), suggesting no embryonic defect.

    Unfortunately, we cannot provide data on 3-weeks old mice (we have not collected this timepoint previously and currently we do not have this mouse line alive). Instead, however, we provide in situ pictures of DTA and *S100a4-Cre;DTA *nipples at 7 weeks of age (Figure for Reviewers 1c; Figure S3c in the revised manuscript), which demonstrate that the phenotype of defective nipple is fully established at this timepoint. Because the late embryonic data did not support the “congenital” establishment of the nipple deformity and we could not provide any more data from early postnatal development, we have corrected the statement “we describe a congenital nipple deformity” in the discussion to “we describe a nipple deformity”.

    Are there S100a4+ cells in the nipple area of pubertal S100a4-Cre/DTA mice? I.e. is there a continuous supply of new S100a4+ cells and thereby continuous cell death and DTA expression as one might expect based on the RNA-seq data?

    The S100A4+ cells are present in the nipple area of *S100a4-Cre;DTA *mice, suggesting a continuous supply of new S100A4+ cells (Figure for Reviewers 1b, corresponding to Figure S3b in the revised manuscript; and Figure for Reviewers 5a,b, corresponding to Figure S7a,b in the revised manuscript). In the revised manuscript, we comment on this in Discussion: “Notably, we observed incomplete depletion of S100A4+ cells in the mammary gland and nipple. Interestingly, a study using the same S100a4-Cre;DTA mouse model reported complete S100A4+ cell depletion in the superficial layer of mandibular condyle46. This suggests that incomplete depletion of S100A4+ cells in nipple and mammary gland is due to tissue-specific dynamics, rather than lack of depletion efficiency, indicating a compensatory mechanism that can balance the cell loss.”

    Figure for Reviewers 1 (Figure S3 in the revised manuscript): Embryonic and pubertal nipple phenotype. (a) Representative images of cleared whole-mount S100a4-Cre;mT/mG nipple tissue at embryonic developmental time-points: E15.5 and E18.5. Scale bar = 100 µm. (b) Immunofluorescent labeling for S100A4 on embryonic* DTA *and S100a4-Cre;DTA whole-mount skin (E18.5). Scale bar = 100 µm. (c) Representative in situ photographs of nipples from *DTA *and S100a4-Cre;DTA pubertal (7-weeks old) mice. Scale bar = 1 mm.

    The subtitle on line 54 implies that that S100a4-Cre/DTA mice display a branching phenotype. However, it looks to me as if there is a pubertal outgrowth defect (as is also written in the body text, line 64) rather than a branching phenotype, potentially reflecting the much smaller size of S100a4-Cre/DTA mice (Fig. 2a). Unless there is a change in branch point frequency, I suggest rephrasing the title and discussion. Instead, I suggest the authors discuss the observed outgrowth delay considering the gross overall growth defect (Fig. 2a). If ductal outgrowth was normalized to the overall growth defect, would one still observe 'a delay in branching morphogenesis'?

    We apologize for the section title confusion. We have analyzed branching frequency in 7-weeks-old females and observed reduced total number of branching points in *S100a4-Cre;DTA *mice (Figure for Reviewers 2a, corresponding to Figure 2f in the revised manuscript). A significant difference in number of branching points remained also after their normalization to body weight, (Figure for Reviewers 2c, corresponding to Figure 2h in the revised manuscript). We have now added the new quantifications to the revised manuscript with accompanying descriptions in the main text “Analysis of mammary epithelial development using whole-mount carmine staining revealed no significant differences in the prenatal establishment of the mammary epithelial tree but did reveal significantly delayed epithelial outgrowth and reduced branching in pubertal (7 weeks old) S100a4-Cre;DTA mice (Figure 2e,f). Normalization of epithelial outgrowth and branching to body weight indicates that the observed defect represents a mammary-specific impairment rather than a consequence of reduced body growth (Figure 2g,h).”.

    __Figure for Reviewers 2 (Figure 2 in the revised manuscript): __Pubertal branching morphogenesis is delayed in S100a4-Cre;DTA. (a-c) The plots show total number of branching points (a), epithelial outgrowth [mm] normalized to body weight [g] (b), and total number of the branching points normalized to body weight [g] (c) in 7 weeks old DTA and S100a4-Cre;DTA mice. All plots show the mean ± SD, *p

    Fig. 4e shows Masson's Trichrome and Picrosirius Red staining and the authors report the findings as follows (lines 120-124): "collagen fibers were loosened in the DTA nipples and more densely packed in the S100a4-Cre;DTA nipples". Perhaps the authors could help non-specialists to observe the loosened fibers and if they wish to make quantitative statements ("more densely packed"), such statements should be backed-up by quantifications.

    Picrosirius Red staining viewed under polarized light is a classic way to assess collagen organization, thickness, and packing. Red / orange / yellow color typically marks thicker, more mature, and more tightly packed collagen fibers (often associated with type I collagen), while green color usually marks thinner, less organized, or less densely packed fibers (often associated with type III collagen or immature collagen). We had included this explanation in the Figure legend of the submitted manuscript already: “Typically, thicker collagen fibers exhibit stronger birefringence and appear red or orange, while thinner fibers exhibit weaker birefringence and appear green or yellow.” To help with the quantification, we have extracted the red channel and quantified color intensity. The results are shown in Figure for Reviewers 3, corresponding to Figure S4 in the revised manuscript. Moreover, we will also quantify the differences in pattern of the collagen fibers. The fibers in* DTA* nipples look shorter and more curved, while the fibers in *S100a4-Cre;DTA *nipples look longer and straighter, more aligned. The results will be included in the fully revised manuscript.

    Figure for Reviewers 3 (Figure S4 in the revised manuscript): Collagen fibers are densely packed in S100a4-Cre;DTA nipples contain more . (a) Representative pictures of histological sections of DTA and S100a4-Cre;DTA stained for collagen by Picrosirius red. Polarized light images and the red channel (mature/densely packed collagen) are shown alongside detail pictures of selected regions A and B. Scale bar = 200 µm and 100 µm (in detail pictures). (b) Quantification of Intensity Mean Value for the red channel (densely packed collagen), showing statistically non-significant difference. The plot shows the mean ± SD, ns p > 0.05 (Mann-Whitney test), n = 3 DTA / 4 S100a4-Cre;DTA.

    I found the Discussion on the various mouse models somewhat problematic. Overall, the paper is written is a way that it often remains unclear whether it refers to studies addressing the role of S100a4 itself, studies addressing the function of S100a4+ cells via ablation approaches (S100a4-Cre or S10 0a4-CreERT2 crossed with floxed DTA), or those where S100a4-Cre has been used to delete gene X/Y/Z. These are all very different experimental approaches where one approach is not necessarily informative when trying to understand the results from another one. The authors should make these points clear and consider whether all their discussion points are relevant.

    We apologize for the confusion. We have carefully reviewed the references and their interpretations, and corrected them as necessary.

    The abstract states S100a4 (fibroblast-specific protein 1) is "expressed by mesenchymal cells and has been implicated in the development of eccrine glands, hair follicles, and mammary branching morphogenesis". However, the study on eccrine glands (ref. 19) shows that S100A4+ cells play a role in eccrine gland development but it does not address the role of S100a4 itself, while the study on hair follicles (ref.20) in turn reports the expression pattern of S100a4 in hair follicles but does not address its function, nor the role of S100a4+ cells. Finally, I failed to find references in the paper to studies addressing the role of S100a4, or S100a4+ cells in the mammary gland.

    Instead, the paper had references to studies where S100A4-Cre had been used to delete different genes and these mice had various mammary phenotypes - which, as indicated above, is a very different approach compared to deleting S100a4 or ablating S100a4+ cells.

    Thank you for your comment. We addressed the concern in the Abstract and further in the Discussion. We revisited the present the cited studies more carefully, clearly distinguishing the different approaches and particular findings.

    In our literature review, we also considered studies that used S100a4-Cre mouse model, to manipulate gene expression within S100A4+ cells. We believe that these studies bring indirect evidence of S100A4+ cell involvement in development and/or homeostasis of a tissue, such as mammary gland. Please, find the rephrased part of Abstract in the text, and below:

    “S100A4 (S100 calcium binding protein A4, also known as fibroblast-specific protein 1) is expressed by mesenchymal cells and has been associated with hair follicle regeneration. S100A4-expressing cells have been implicated in the development of eccrine glands, and studies using S100a4-Cre to manipulate gene function have suggested that S100A4-expressing cells may contribute to mammary branching morphogenesis.”

    __In Discussion (lines 197-200), __the authors write: "We described significant delay in mammary branching morphogenesis in puberty, confirming an important role for S100A4+ cells in mammary development, as it was previously described (refs 37-39)."

    It should be noted that none of these studies addressed the role of S100A4+ cells:

    • Ref 37 used S100a4-Cre to delete sharpin
    • Ref 38 used the same Cre line to delete Ptch1, did not address the role of S100a4 or S100a4 expressing cells
    • Likewise ref 39 deleted another gene using S100a4-Cre

    Later on in Discussion, the authors compare the reported phenotype to previous studies (lines 248-255): "...targeting S100A4+ cells through knockout experiments can result in severe phenotypes, such as a reduction in adipose tissue (ref 26), skin phenotypes, a disrupted estrous cycle, reduced fertility (ref. 38), and complete infertility, hypogonadism and defects in pituitary endocrine function (ref. 28).

    Of these, Ref. 26 used the same approach as the current study (S100a4-Cre; DTA) (Fig. 7A in the paper)

    • these mice were significantly lean, with markedly reduced fat compared with the control mice - also the mice in the current study are very small, so perhaps they could also be described as 'lean'. Yet ref. 26 reports that female mice had comparable food uptake, respiratory exchange ratio and physical activity, and slightly increased energy expenditure

    Ref. 38 (as mentioned above) reports deletion of Ptch1 using S100a4-Cre lines and these mice "displayed a disrupted estrous cycle and dramatically reduced fertility over 6.5 weeks". However, this has nothing to do with the approaches where Fsp1/S100a4+ cells are depleted with DTA. Likewise, reference 28 analyzed the phenotype of S00a4-Cre;Ptch1fl/fl mice. Obviously, deleting Ptch1 using S100a4-Cre mice is quite a different approach than "targeting S100A4+ cells" through knockout experiments". Ptch1 deletion leads to a combination of gain-of-function (of Hedgehog activation) and loss-of-function (loss of Hh-independent functions of Ptch1) and hence comparisons with these phenotypes is rather challenging. I suggest the authors focus their phenotype comparisons to ref. 26 where S100a4/Fsp1+ cells were ablated with DTA, i.e. the same approach as in the current study.

    Please, find the rephrased part of Discussion in the text (lines 236-256), and below:

    “A key consideration when interpreting studies involving S100A4 is that fundamentally different experimental approaches have been used to investigate its role. These include descriptive analyses of S100A4 expression, functional studies targeting the S100A4 protein itself, genetic models using S100a4-Cre to manipulate unrelated genes in S100A4-expressing cells, and ablation models such as S100a4-Cre;DTA, which deplete S100A4⁺ cells. These approaches are not equivalent and provide distinct types of information. In the present study, we specifically assess the consequences of ablating S100A4-expressing cells, and comparisons to other studies should therefore be interpreted within this context.

    Studies using S100a4-Cre to manipulate specific signaling pathways (e.g. Wnt or Hedgehog signaling via gene deletion) in S100A4-expressing cells have reported diverse phenotypes, including effects on fertility and endocrine function28,34. However, these phenotypes primarily reflect the consequences of pathway perturbations within S100A4-expressing cells rather than the role of S100A4⁺ cells themselves. This is fundamentally different from the ablation approach used here, which removes the S100A4⁺ cell population.

    In contrast, studies employing S100a4-Cre–driven DTA–mediated ablation represent a directly comparable approach. Such studies have reported systemic phenotypes, including reduced adipose tissue and altered metabolic parameters26, indicating that S100A4-expressing cells contribute to multiple aspects of tissue homeostasis. Consistent with these previous reports, S100a4-Cre;DTA mice used in our study were significantly smaller than their littermates. Our findings extend these observations by identifying a specific and previously unrecognized role for this cell population in nipple morphogenesis.”

    I find the Discussion is somewhat off the topic by starting with WHO recommendations on breastfeeding and linking this to observed mouse phenotype. Overall, the discussion is rather long and from time-to-time more like a literature review. I would recommend keeping the Discussion more succinct and focused.

    To improve the conciseness and focus of Discussion, we have deleted this part of text.

    **Referee cross-comenting**

    I agree with the comments of other reviewers. However, to me it seems that the analysis of S100a4 knockout mice would not be feasible within a reasonable timeframe and would represent a study of its own. My understanding was that the authors were not interested in S100a4 itself. Rather, S100a4-Cre was used as a tool to understand the importance of a certain (fibroblast) cell population for mammary gland morphogenesis.

    Indeed, our goal was to study the role of a specific cell population (S100A4+ cells) in mammary gland morphogenesis, not to study the role of S100A4 protein per se.

    Reviewer #1 (Significance (Required)): General assessment:

    This study reveals the importance of the S100a4+ cell lineage for nipple formation while showing the same cells are dispensable for mammary gland morphogenesis. The main limitation is that it remains unclear whether the observed nipple phenotype is derived from an early embryonic/prepubertal defect in establishing the nipple stroma, is caused by a constitutive response to ongoing cell death, or a response to continuous DTA expression (or a combination of some of these). Hence its relevance as a model of human inverted nipple condition remains rather speculative.

    Thank you for consideration of our work and valuable feedback. We did not intend to claim that S100a4-Cre;DTA mouse represents a model of human inverted nipple condition. However, considering morphological features, it might resemble it. We now rephrased the Discussion so it is clearer and more concise.

    Reviewer #2

    Major comments:

    My key concern is the discussion part. I think the authors need to re-organize/re-phrase the discussion part, it confused me a bit in terms of logic, phrases and interpretation of literatures.

    We have significantly re-organized and re-phrased the Discussion.

    Here are few examples:

    The lines 195-199 contain lot of repeated information

    We have rephrased the paragraph and removed repeated information. The new text can be found in lines 201-206 in the revised manuscript.

    The authors mentioned the studies in ref 26,28 and 38 using "targeting S100A4+ cells through knockout experiment can result in sever phenotypes". This is very misleading. Those studies using the same (or similar if the origin is different) S100A4-Cre line as the current study but induced the activation of Wnt and sHH signalling pathways, respectively. The observed phenotypes are largely due to the pathway function, rather than the S100A4 gene or normal S100A4+ cell itself. This is significantly differed from the current study.

    We apologize for the confusion; we have now rephrased our claims (lines 236-256):

    “A key consideration when interpreting studies involving S100A4 is that fundamentally different experimental approaches have been used to investigate its role. These include descriptive analyses of S100A4 expression, functional studies targeting the S100A4 protein itself, genetic models using S100a4-Cre to manipulate unrelated genes in S100A4-expressing cells, and ablation models such as S100a4-Cre;DTA, which deplete S100A4⁺ cells. These approaches are not equivalent and provide distinct types of information. In the present study, we specifically assess the consequences of ablating S100A4-expressing cells, and comparisons to other studies should therefore be interpreted within this context.

    Studies using S100a4-Cre to manipulate specific signaling pathways (e.g. Wnt or Hedgehog signaling via gene deletion) in S100A4-expressing cells have reported diverse phenotypes, including effects on fertility and endocrine function28,34. However, these phenotypes primarily reflect the consequences of pathway perturbations within S100A4-expressing cells rather than the role of S100A4⁺ cells themselves. This is fundamentally different from the ablation approach used here, which removes the S100A4⁺ cell population.

    In contrast, studies employing S100a4-Cre–driven DTA–mediated ablation represent a directly comparable approach. Such studies have reported systemic phenotypes, including reduced adipose tissue and altered metabolic parameters26, indicating that S100A4-expressing cells contribute to multiple aspects of tissue homeostasis. Consistent with these previous reports, S100a4-Cre;DTA mice used in our study were significantly smaller than their littermates. Our findings extend these observations by identifying a specific and previously unrecognized role for this cell population in nipple morphogenesis.”

    In the lines 253-255, why the author believe complete S100A4+ depletion would leads to the fatal of mouse? Is there study suggest that? Or have authors checked the expression of S100A4 in the S100A4-Cre;DTA model to confirm the efficiency?

    We have now included, also in response to other Reviewers’ comments, data on S100A4 expression in the S100A4-Cre;DTA model (Figure for Reviewers 5, corresponding to Figure S7 in the revised manuscript), and commented on these results in lines 257-262: “Notably, we observed incomplete depletion of S100A4+ cells in the mammary gland and nipple. Interestingly, a study using the same S100a4-Cre;DTA mouse model reported complete S100A4+ cell depletion in the superficial layer of mandibular condyle48. This suggests that incomplete depletion of S100A4+ cells in nipple and mammary gland is due to tissue-specific dynamics, rather than lack of depletion efficiency, indicating a compensatory mechanism that can balance the cell loss.”

    In Fig. 1, the authors described the impaired nursing capacity of S100A4-Cre;DTA dam. However, it seems the little size is also smaller (Fig 1a). Do authors have any explanation or hypothesis?

    Thank you for this insightful observation. It is well established that metabolic and nutritional condition directly affect female reproductive functions. Adult *S100A4-Cre;DTA *mice are generally smaller compared to their litter counterparts, potentially because of lower body fat content or other anatomic/metabolic condition that might negatively influence fecundity, for instance, lowering ovulation rate and/or embryonic survival. In support of this, earlier studies have reported a positive correlation between growth rate/body condition and litter size (Eisen & Durrant, 1980). Unfortunately, in the case of *S100A4-Cre;DTA *mice, we can only speculate about the possible explanations, as we do not have supporting data which could confirm it.

    In lines 181-184, the authors states "the results showed that the tissue reacted to a foreign chemical or an endogenous compound....." , which results are referring here? I could not find any inflammation related GO terms in figure 6b. It would be more accurate to specify them in lines 179-181, which appears to be a technical statement rather than a result in current form.

    Thank you for this comment. Indeed, there are no GO terms explicitly labeled as “inflammation” and “repair”; however, several GO terms are functionally related to these processes. Our interpretation was based on broader biological context rather the explicit annotation. To clarify this, we revisited the text and included GO terms that reflect the tissue response (lines 187-193).

    “The GO terms indicated that the tissue reacted to a foreign chemical or an endogenous compound (xenobiotic metabolic process, cellular response to xenobiotic stimulus, response to xenobiotic stimulus, epoxygenase P450 pathway), and responded to inflammation and repair (actin filament-based process, actin cytoskeleton organization; eicosanoid and lipid metabolic processes) (Figure 6b).”

    The lines 182-184 was not clear. Does the author refer the "nipple tissue response" in general as malfunction of development or inflammation and tissue repair as mentioned in the previous sentence? If the later cases, the authors should consider the failure of lactation might mimic the involution, which may cause the apoptosis and inflammation as well. This might be independent of the DTA expression.

    Thank you for raising this point. Indeed, in this line, we refer to ongoing tissue inflammation and repair. We also considered the hypothesis that the ejection incapability (and consecutive milk stasis) triggers involution. However, tissues were collected within a few hours after parturition, when only very early signs of involution, if any, would be detectable; therefore, we expect minimal influence of involution. To reflect this comment, we added new text to the Discussion (lines 272– 277). “The observed tissue response can be also associated with hallmarks of mammary involution, the process which is triggered by the milk stasis. However, the tissues were collected within few hours after parturition, when the effect of involution should be minimal53. Rather, we hypothesize that immune cell recruitment, and the upregulation of the lipid skin barrier might be caused in response to the continuous apoptosis of S100A4+ cells and their replacement.”

    Minor comments:

    The authors demonstrated in Figure S1 and lines 92-96 that no significant differences were observed in pituitary glands and ovaries in S100a4-Cre:DTA and DTA mice. Have the authors checked the S100A4 expression or lineage cells in these organs, or have been reported by others?

    Yes, we checked the S100A4-lineage cells in the pituitary gland and ovary and have now included the results here (Figure for Reviewer 4a,b corresponding to Figure S1a,b in the revised manuscript), along with relevant text description (lines 94-95 in the revised manuscript). “We observed S100A4-lineage traced cells in pituitary gland and ovaries using S100a4-Cre;mT/mG model (Figure S1a,b).” The presence of S100A4+ cells in these organs was also reported previously (Ren et al., 2019).

    Figure for Reviewers 4 (Figure S1 in the revised manuscript): S100A4-lineage cells are abundant in the pituitary gland and ovary. (a) Representative images of a cleared whole-mount pituitary gland from a S100a4-Cre;mT/mG mouse. (b) Representative images of a cleared whole-mount ovary from a S100a4-Cre;mT/mG mouse. Scale bar = 100 µm.

    The authors have performed live imaging to evaluate the contraction of alveoli. It would be better to include a video together with the snapshots showed in Figure S2.

    We have included the videos as supplementary movies, Movie S1 (DTA) and Movie S2 (S100a-Cre;DTA).

    Since the study is mainly using S100a4, it would be better to avoid using FSP1 in the results, for example Fig 5h.

    We apologize for this oversight; it has now been corrected.

    What does L1 stand for? Lactation Day 1? It should be spelt out in the first instance.

    Yes, indeed, L1 is lactation day 1. Please note that it was already spelled out in the first version of the manuscript, now in line 48.

    Line 150. Figure S4 should be Figure S4a.

    (Please note, that by adding new Supplementary figures, this comment is referring to Figure S6 in the new version of manuscript.) Thank you for this comment. In the text, we state “GFP+ cells were spread throughout the fat pad but were also localized in the periepithelial stroma and infiltrated the epithelium”. This we show in Figure S6a and in S6b; therefore, we now changed the reference accordingly, as it might be more accurate.

    **Referee cross-comenting**

    I agree with the other reviewers, as well as the Consultation Comments. The manuscript would benefit greatly from a thoroughly optimised Discussion section to address issues raised by all reviewers.

    __ Reviewer #2__ (Significance (Required)):

    • Overall, this study is well designed and the key findings are valid, especially the role of S100A4 during nipple development is novel and interesting.

    -One limitation of the study is that RNA-seq was performed using a mixture of all cell types present in the nipple. While this approach is reasonable-given that depletion of the S100A4+ lineage may exert both direct and indirect effects contributing to nipple dysfunction-it should be more clearly acknowledged and discussed in the manuscript. Additionally, this experimental design may limit the utility of the dataset for other researchers interested in nipple development and the specific functions of S100A4.

    Reviewer #3

    Major comments:

    1. The differential systemic versus mammary-specific effects of DTA-mediated S100A4 cell ablation are intriguing. The authors should address why the mammary fat pad appears unaffected.

    Thank you for this comment. The role of S100A4+ cells in adipose tissue was previously reported (Zhang et al., 2018). Authors reported significantly smaller adipose tissue of S100a4-Cre;DTA mice (males and females), measured as the weight of the dissected fat pad. In our work, we measured the* in-situ* area of the fat pad, which appeared to be unaffected. It is possible that the volume (weight) of the fat pad would be different, however we do not have data to confirm / reject this hypothesis.

    Are S100A4 expressing cells present during embryonic mammary development, or are they mainly postnatal? Would an inducible S100A4CreERT model lead to similar phenotypes, or might the timing of depletion influence the outcome? Discussing these points would reinforce the conclusions regarding the contribution of S100A4-expressing cells to mammary and nipple development and could also clarify the transient nature of the ductal branching phenotype.

    S100A4-expressing cells are present during embryonic mammary development, too. Please, refer to the embryonic lineage-tracing time-points incorporated in the first version of the manuscript (Figure 5a and Figure S6a). Now, we have added Figure for Reviewers 1 corresponding to Figure S3 in the revised manuscript), which focuses on the embryonic nipple phenotype but also provides information on the presence of S100A4+ cells.

    We agree that the use of inducible S100a4-CreERT model could potentially bring new insights toward developmental stage-specific roles of S100A4+ cells, and thus would be interesting to use in a follow-up study. Currently, such experiments are beyond our capacity.

    Therefore, we have included a new subsection on Limitations of the study, where we comment:

    “A major limitation of this study is that the timing of DTA-mediated cell depletion cannot be precisely defined in the constitutive mouse model employing S100a4-Cre because recombination may occur continuously following the initial expression of S100a4 (E8.518). This limitation could be overcome by usage of inducible *S100a4-CreERT *instead. With this approach, it could be more feasible to determine if the nipple deformity arises as a defect of embryonic development or postnatal morphogenesis.”

    1. Although the authors attribute lactation failure primarily to defects in nipple architecture, the RNA seq data reveal downregulation of key milk production genes and luminal differentiation keratins, strongly suggesting impaired secretory activation. The authors should more explicitly discuss the relative contributions of epithelial functional maturation defects versus nipple structural abnormalities to the lactation failure observed upon S100A4+ cell depletion. Thank you for this comment. We believe that performing an immunofluorescence labeling of epithelial architecture (requested in the Minor comment 2) could bring more light into this. However, we deduce that secretory activation is not impaired, as the presence of the milk observed on in situ wholemounts, and H&E-stained alveoli (Figure 3d) implies luminal secretion of milk components. The observed phenotype of the lactating mammary gland strongly suggests there is a structural abnormality inhibiting the milk ejection.

    The downregulation of key milk production genes and luminal keratins in the bulk RNA-seq data may be influenced by differences in tissue composition between samples. In control mice, more fully developed nipples and an extended ductal network likely contribute to a greater representation of differentiated luminal epithelial cells, thereby increasing the expression of these markers.

    Minor comments:

    Figure 1: Including an immunohistochemistry or immunofluorescence control confirming depletion of S100A4 expressing cells would strengthen the conclusions.

    We have now included Figure for Reviewers 5 that corresponds to Figure S7 in the revised manuscript and comment on the results in sections Results (lines 169-171) and Discussion (lines 257-262).

    In Results: “Interestingly, S100A4 antibody labeling revealed presence of S100A4+ cells in S100a4-Cre;DTA tissues (Figure S3b, Figure S7a,b).”

    In Discussion: “Notably, we observed incomplete depletion of S100A4+ cells in the mammary gland and nipple. Interestingly, a study using the same S100a4-Cre;DTA mouse model reported complete S100A4+ cell depletion in the superficial layer of mandibular condyle48. This suggests that incomplete depletion of S100A4+ cells in nipple and mammary gland is due to tissue-specific dynamics, rather than lack of depletion efficiency, indicating a compensatory mechanism that can balance the cell loss.”

    Figure for Reviewers 5 (Figure S7 in the revised manuscript): S100A4+ cells are found in S100a4-Cre;DTA nipple and mammary tissues. (a) Immunofluorescent labeling for S100A4 and vimentin on FFPE sections of DTA and* S100a4-Cre;DTA* L1 nipples. (b) Immunofluorescent labeling for S100A4 and smooth muscle actin on FFPE sections of DTA and S100a4-Cre;DTA L1 mammary gland. Scale bar = 100 µm.

    Figure 3c: The histological defects more accurately reflect failure of secretory activation rather than "lactation failure" per se. The terminology should be refined to reflect this more precisely.

    Thank you for this comment. As explained in the response to your major comment 3, we believe our results show that the secretory activation is conserved in S100a4-Cre;DTA lactating mice. We understand that “lactation failure” might be misleading terminology, as the production of the milk is conserved as well. We therefore change the phrasing into “nursing defect” (line 51, 73, 83), as this could reflect the phenotype most precisely.

    **Referee cross-comenting**

    I agree with the Reviewer, the authors do not need to do knockout experiments in the revised manuscript. However, it would be great if they could address my comment in the discussion.

    Reviewer #3 (Significance (Required)):

    This is an important study for mammary developmental biology, addressing the relatively understudied mechanisms that govern nipple development at the stromal-epithelial interface, and the determinants of lactational performance. A major strength is the elegant integration of DTA-mediated cell ablation, advanced imaging, lineage tracing, and transcriptomics to uncover previously uncharacterised roles for S100A4-expressing stromal populations in shaping nipple morphology and function. The work lays a foundation for future studies into nipple biology and pathologies and mechanisms underlying successful lactation.

    Although the study is already mature, it could be further strengthened by incorporating more specific genetic models, such as inducible S100A4CreERT or S100A4 gene knockout/knockdown approaches.

    Thank you for appreciation of our work.

    4. Description of analyses that authors prefer not to carry out

    Reviewer #1

    Major Comment 1.

    It is rather difficult to conclude whether the observed nipple phenotype reflects an early embryonic/prepubertal defect in establishing the nipple stroma, is caused by a constitutive response to ongoing cell death, or a response to continuous DTA expression (or a combination of some of these). The data raise a couple of additional questions: Is there a nipple phenotype at 3 wk of age?...

    Unfortunately, we cannot provide data on 3 weeks old mice because we did not collect such samples before and we had to terminate our mouse colony due to an infection in the animal house (mouse line reanimation is possible because we had stored sperm of the mouse line but it would take a lot of time and resources). Nevertheless, we tried to address this comment by providing other relevant available data (see Figure for Reviewers 1).

    Reviewer #2

    Major Comment 3.

    In Fig S1c, d and lines 93-96, the authors investigated the estrus cycles to determine the potential cause of lactation failure. The data was presented as the number of mice in each stage. A more intuitive approach would be to follow the same mice for two to three cycles and observe the duration of each stage.

    We agree that the suggested approach would be more accurate in determining truly cycling females. Unfortunately, we cannot perform this experiment currently because we do not have these mice alive anymore. Nevertheless, because the S100a4-Cre;DTA females bore pups, they had cycled and were fertile.

    Reviewer #3

    Major comment 1.

    While the S100A4Cre::DTA model is powerful for evaluating the roles of S100A4 expressing cells, the authors should discuss the potential outcomes of using S100A4 knockout or knockdown approaches. If the authors have such data available, this could help distinguish phenotypes caused by loss of S100A4 function itself from those arising due to ablation of S100A4 expressing cell populations and would add mechanistic depth to the study.

    We thank the Reviewer for this insightful suggestion. We agree that genetic approaches targeting S100A4 function (e.g., knockout or knockdown) could, in principle, help disentangle cell-autonomous effects of S100A4 from those resulting from the loss of S100A4-expressing cell populations. However, we would like to clarify that the primary objective of our study is to investigate the functional contribution of S100A4⁺ stromal cells at the population level, rather than to dissect the molecular function of S100A4 protein per se. In this context, the S100A4-Cre;DTA model provides a well-established and appropriate strategy to ablate this cell population and assess its role in tissue development. Importantly, S100A4 is not only a functional protein but also a widely used marker of a heterogeneous stromal cell population. Genetic ablation of S100A4 itself would not eliminate these cells, and may result in relatively subtle or compensable phenotypes due to functional redundancy within the S100 protein family or context-dependent roles of S100A4. Therefore, such approaches would address a distinct biological question and may not directly recapitulate the phenotypes observed upon cell ablation.

    References

    Eisen, E. J., & Durrant, B. S. (1980). Genetic and Maternal Environmental Factors Influencing Litter Size and Reproductive Efficiency in Mice. Journal of Animal Science, 50(3), 428–441. https://doi.org/10.2527/jas1980.503428x

    Ren, Y. A., Monkkonen, T., Lewis, M. T., Bernard, D. J., Christian, H. C., Jorgez, C. J., Moore, J. A., Landua, J. D., Chin, H. M., Chen, W., Singh, S., Kim, I. S., Zhang, X. H. F., Xia, Y., Phillips, K. J., MacKay, H., Waterland, R. A., Cecilia Ljungberg, M., Saha, P. K., … Richards, J. A. S. (2019). S100a4-Cre–mediated deletion of Ptch1 causes hypogonadotropic hypogonadism: Role of pituitary hematopoietic cells in endocrine regulation. JCI Insight, 4(14). https://doi.org/10.1172/jci.insight.126325

    Tuwatnawanit, T., Wessman, W., Belisova, D., Sumbalova Koledova, Z., Tucker, A. S., & Anthwal, N. (2025). FSP1/S100A4-Expressing Stem/Progenitor Cells Are Essential for Temporomandibular Joint Growth and Homeostasis. Journal of Dental Research, 104(5), 551–560. https://doi.org/10.1177/00220345251313795

    Zhang, R., Gao, Y., Zhao, X., Gao, M., Wu, Y., Han, Y., Qiao, Y., Luo, Z., Yang, L., Chen, J., & Ge, G. (2018). FSP1-positive fibroblasts are adipogenic niche and regulate adipose homeostasis. PLoS Biology, 16(8). https://doi.org/10.1371/journal.pbio.2001493

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary:

    In this pre-print, Belisova et al. investigate the under-explored mechanisms regulating nipple development and its essential role in offspring nourishment, focusing on the contribution of S100A4‑expressing cells in the mouse mammary gland. The authors use an elegant combination of Cre::DTA-mediated cell depletion, lineage tracing, imaging, RNA-seq, and functional assays to reveal roles for S100A4‑expressing fibroblasts and immune cells in nipple morphogenesis and lactation. The manuscript is generally well written, and the experimental design is strong, with appropriate controls supporting the overall conclusions. However, I have several comments and suggestions to improve this initial manuscript.

    Major comments:

    1. While the S100A4Cre::DTA model is powerful for evaluating the roles of S100A4 expressing cells, the authors should discuss the potential outcomes of using S100A4 knockout or knockdown approaches. If the authors have such data available, this could help distinguish phenotypes caused by loss of S100A4 function itself from those arising due to ablation of S100A4 expressing cell populations and would add mechanistic depth to the study.

    2. The differential systemic versus mammary-specific effects of DTA-mediated S100A4 cell ablation are intriguing. The authors should address why the mammary fat pad appears unaffected. Are S100A4 expressing cells present during embryonic mammary development, or are they mainly postnatal? Would an inducible S100A4CreERT model lead to similar phenotypes, or might the timing of depletion influence the outcome? Discussing these points would reinforce the conclusions regarding the contribution of S100A4-expressing cells to mammary and nipple development and could also clarify the transient nature of the ductal branching phenotype.

    3. Although the authors attribute lactation failure primarily to defects in nipple architecture, the RNA seq data reveal downregulation of key milk production genes and luminal differentiation keratins, strongly suggesting impaired secretory activation. The authors should more explicitly discuss the relative contributions of epithelial functional maturation defects versus nipple structural abnormalities to the lactation failure observed upon S100A4+ cell depletion.

    Minor comments:

    1. Figure 1: Including an immunohistochemistry or immunofluorescence control confirming depletion of S100A4 expressing cells would strengthen the conclusions.

    2. Figure 2c: The H&E images are not fully convincing. Immunofluorescence analysis of epithelial architecture would support the authors' interpretation and should be feasible if tissues are already available.

    3. Figure 3c: The histological defects more accurately reflect failure of secretory activation rather than "lactation failure" per se. The terminology should be refined to reflect this more precisely.

    4. Figure 4f: The proliferation data are compelling, but the authors could extend this by examining how cell differentiation and epithelial organisation are affected.

    5. Figure 5b: To more convincingly show that GFP+ cells contact endothelial cells, co-labelling with an endothelial marker such as CD31 would be helpful.

    6. Figure 5f-h: The structures referenced in the text (lines 159-163) should be clearly indicated on the immunofluorescence images.

    Referee cross-comenting

    I agree with the Reviewer, the authors do not need to do knockout experiments in the revised manuscript. However, it would be great if they could address my comment in the discussion.

    Significance

    This is an important study for mammary developmental biology, addressing the relatively understudied mechanisms that govern nipple development at the stromal-epithelial interface, and the determinants of lactational performance. A major strength is the elegant integration of DTA-mediated cell ablation, advanced imaging, lineage tracing, and transcriptomics to uncover previously uncharacterised roles for S100A4-expressing stromal populations in shaping nipple morphology and function. The work lays a foundation for future studies into nipple biology and pathologies and mechanisms underlying successful lactation.

    Although the study is already mature, it could be further strengthened by incorporating more specific genetic models, such as inducible S100A4CreERT or S100A4 gene knockout/knockdown approaches.

    I have expertise in mammary epithelial biology.

    I estimate that revisions would require 3-6 months if new experiments are performed, and 1-3 months if revisions focus on clarifying claims and strengthening the discussion.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    In this study, Jaros Belisova et al. systematically investigated the composition and roles of S100A4+ cells during mammary gland development and identified a novel role for S100A4 for nipple development. Depletion of S100A4+ lineage using the S100a1-Cre;DTA model resulted in delayed pubertal mammary gland development but with normal morphology and milk production capacity during lactation. The authors further investigated the milk-ejection function of the alveoli using an ex vivo culture method combined with live imaging. This revealed that depletion of S100A4+ cells does not interfere with the normal function of alveoli. However, the abnormal development of the nipple, characterised by smaller size, shorter length, lacking protrusion, increased collagen composition and decreased cell proliferation at the onset of lactation, results in milk delivery failure which is responsible for the lethality of the pups. To further understand the consequences of S100A4+ cell depletion, the author utilised the S100a4-Cre;mTmG model to trace the cell types depleted in the DTA model across various developmental stages. Immunofluorescent staining revealed that S100A4 lineage cells comprised both fibroblasts and immune cells, consistent with previous studies. Interestingly, some S100A4 lineage (GFP+) retain the expression of S100A4. In addition, the RNAseq data comparing the nipple cells from S100a4-Cre;DTA and DTA lactation mice confirmed their observations in transcription level. Overall, the experiments are well designed and the key findings are valid, especially about the role of S100A4 during nipple development is novel and interesting.

    Major comments:

    1. My key concern is the discussion part. I think the authors need to re-organize/re-phrase the discussion part, it confused me a bit in terms of logic, phrases and interpretation of literatures. Here are few examples:

    a. The lines 195-199 contain lot of repeated information

    b. The authors mentioned the studies in ref 26,28 and 38 using "targeting S100A4+ cells through knockout experiment can result in sever phenotypes". This is very misleading. Those studies using the same (or similar if the origin is different) S100A4-Cre line as the current study but induced the activation of Wnt and sHH signalling pathways, respectively. The observed phenotypes are largely due to the pathway function, rather than the S100A4 gene or normal S100A4+ cell itself. This is significantly differed from the current study.

    c. In the lines 253-255, why the author believe complete S100A4+ depletion would leads to the fatal of mouse? Is there study suggest that? Or have authors checked the expression of S100A4 in the S100A4-Cre;DTA model to confirm the efficiency?

    d. The authors tried to attribute the minor phenotype to the incomplete depletion of S100A4+ cells. However, it is possible that if the S100A4+ cells only represented a minor population, their function may be compensated by other populations. This might be confirmed by quantification of S100A4+ cells or S100A4-Cre; GFP+ cells in fibroblast or CD45 populations from images showed in Figure 5.

    1. In Fig. 1, the authors described the impaired nursing capacity of S100A4-Cre;DTA dam. However, it seems the little size is also smaller (Fig 1a). Do authors have any explanation or hypothesis?
    2. In Fig S1c, d and lines 93-96, the authors investigated the estrus cycles to determine the potential cause of lactation failure. The data was presented as the number of mice in each stage. A more intuitive approach would be to follow the same mice for two to three cycles and observe the duration of each stage.
    3. The images in Figure 5 and Figure S4 are difficult to confirm colocalization. A higher magnification image would be required for each panel. Furthermore, a precise quantification based on the current images would be more supportive of the conclusion regarding the discrepancy of the composition of S100A4 lineage between epidermis and mammary gland (lines 163-165).
    4. Line 163, the author hypothesis the Langerhans cells due to morphology. Those cells should be able to be confirmed by a co-staining with F4/80 in addition to the current form of Fig 5h.
    5. In lines 181-184, the authors states "the results showed that the tissue reacted to a foreign chemical or an endogenous compound....." , which results are referring here? I could not find any inflammation related GO terms in figure 6b. It would be more accurate to specify them in lines 179-181, which appears to be a technical statement rather than a result in current form.
    6. The lines 182-184 was not clear. Does the author refer the "nipple tissue response" in general as malfunction of development or inflammation and tissue repair as mentioned in the previous sentence? If the later cases, the authors should consider the failure of lactation might mimic the involution, which may cause the apoptosis and inflammation as well. This might be independent of the DTA expression.

    Minor comments:

    1. The authors demonstrated in Figure S1 and lines 92-96 that no significant differences were observed in pituitary glands and ovaries in S100a4-Cre:DTA and DTA mice. Have the authors checked the S100A4 expression or lineage cells in these organs, or have been reported by others?
    2. The authors have performed live imaging to evaluate the contraction of alveoli. It would be better to include a video together with the snapshots showed in Figure S2.
    3. Since the study is mainly using S100a4, it would be better to avoid using FSP1 in the results, for example Fig 5h.
    4. What does L1 stand for? Lactation Day 1? It should be spelt out in the first instance.
    5. Line 150. Figure S4 should be Figure S4a.

    Referee cross-comenting

    I agree with the other reviewers, as well as the Consultation Comments. The manuscript would benefit greatly from a thoroughly optimised Discussion section to address issues raised by all reviewers.

    Significance

    • Overall, this study is well designed and the key findings are valid, especially the role of S100A4 during nipple development is novel and interesting.
    • One limitation of the study is that RNA-seq was performed using a mixture of all cell types present in the nipple. While this approach is reasonable-given that depletion of the S100A4+ lineage may exert both direct and indirect effects contributing to nipple dysfunction-it should be more clearly acknowledged and discussed in the manuscript. Additionally, this experimental design may limit the utility of the dataset for other researchers interested in nipple development and the specific functions of S100A4.

    My expertise:

    mammary gland development and breast cancer

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary

    In this study, Belisova et al. investigate the function of S100a4+ (a.k.a. Fsp1) cells in the mammary gland. S100a4 expressing cells were constitutively ablated using the DTA system by crossing S100a4-Cre mice with ROSA26-eGFP-DTA mice. Female mice exhibited a severe nursing defect, leading to whole-litter mortality within 1-2 days postpartum. However, no abnormalities were detected in the morphology of the mammary ductal tree, milk production, or alveolar contractility of S100a4-Cre;DTA mice. Instead, nipples were malformed, likely prevent normal suckling. Analysis of the lineage of S100a4 expressing cells in the mammary gland using the S100a4-Cre mouse in combination with a fluorescent Cre reporter identified S100a4+ cells as fibroblasts and immune cells in the nipple region, while only immune cells were labelled in the mammary gland stroma, findings that agree with previous studies.

    Major comments:

    1. It is rather difficult to conclude whether the observed nipple phenotype reflects an early embryonic/prepubertal defect in establishing the nipple stroma, is caused by a constitutive response to ongoing cell death, or a response to continuous DTA expression (or a combination of some of these). The data raise a couple of additional questions: Is there a nipple phenotype at 3 wk of age? It would not be totally unsurprising if ablation of a major fraction of dermal fibroblasts in the nipple area would lead to an early embryonic/prepubertal phenotype but there is no data on this. Hence, is there a "congenital" nipple deformity, as concluded by the authors (line 191)? Are there S100a4+ cells in the nipple area of pubertal S100a4-Cre/DTA mice? I.e. is there a continuous supply of new S100a4+ cells and thereby continuous cell death and DTA expression as one might expect based on the RNA-seq data?
    2. The subtitle on line 54 implies that that S100a4-Cre/DTA mice display a branching phenotype. However, it looks to me as if there is a pubertal outgrowth defect (as is also written in the body text, line 64) rather than a branching phenotype, potentially reflecting the much smaller size of S100a4-Cre/DTA mice (Fig. 2a). Unless there is a change in branch point frequency, I suggest rephrasing the title and discussion. Instead, I suggest the authors discuss the observed outgrowth delay considering the gross overall growth defect (Fig. 2a). If ductal outgrowth was normalized to the overall growth defect, would one still observe 'a delay in branching morphogenesis'?
    3. Fig. 4e shows Masson's Trichrome and Picrosirius Red staining and the authors report the findings as follows (lines 120-124): "collagen fibers were loosened in the DTA nipples and more densely packed in the S100a4-Cre;DTA nipples". Perhaps the authors could help non-specialists to observe the loosened fibers and if they wish to make quantitative statements ("more densely packed"), such statements should be backed-up by quantifications.
    4. I found the Discussion on the various mouse models somewhat problematic. Overall, the paper is written is a way that it often remains unclear whether it refers to studies addressing the role of S100a4 itself, studies addressing the function of S100a4+ cells via ablation approaches (S100a4-Cre or S100a4-CreERT2 crossed with floxed DTA), or those where S100a4-Cre has been used to delete gene X/Y/Z. These are all very different experimental approaches where one approach is not necessarily informative when trying to understand the results from another one. The authors should make these points clear and consider whether all their discussion points are relevant. The abstract states S100a4 (fibroblast-specific protein 1) is "expressed by mesenchymal cells and has been implicated in the development of eccrine glands, hair follicles, and mammary branching morphogenesis". However, the study on eccrine glands (ref. 19) shows that S100A4+ cells play a role in eccrine gland development but it does not address the role of S100a4 itself, while the study on hair follicles (ref.20) in turn reports the expression pattern of S100a4 in hair follicles but does not address its function, nor the role of S100a4+ cells. Finally, I failed to find references in the paper to studies addressing the role of S100a4, or S100a4+ cells in the mammary gland. Instead, the paper had references to studies where S100A4-Cre had been used to delete different genes and these mice had various mammary phenotypes - which, as indicated above, is a very different approach compared to deleting S100a4 or ablating S100a4+ cells.

    In Discussion (lines 197-200), the authors write: "We described significant delay in mammary branching morphogenesis in puberty, confirming an important role for S100A4+ cells in mammary development, as it was previously described (refs 37-39)." It should be noted that none of these studies addressed the role of S100A4+ cells:

    • Ref 37 used S100a4-Cre to delete sharpin
    • Ref 38 used the same Cre line to delete Ptch1, did not address the role of S100a4 or S100a4 expressing cells
    • Likewise ref 39 deleted another gene using S100a4-Cre

    Later on in Discussion, the authors compare the reported phenotype to previous studies (lines 248-255): "...targeting S100A4+ cells through knockout experiments can result in severe phenotypes, such as a reduction in adipose tissue (ref 26), skin phenotypes, a disrupted estrous cycle, reduced fertility (ref. 38), and complete infertility, hypogonadism and defects in pituitary endocrine function (ref. 28). Of these, Ref. 26 used the same approach as the current study (S100a4-Cre; DTA) (Fig. 7A in the paper)

    • these mice were significantly lean, with markedly reduced fat compared with the control mice - also the mice in the current study are very small, so perhaps they could also be described as 'lean'. Yet ref. 26 reports that female mice had comparable food uptake, respiratory exchange ratio and physical activity, and slightly increased energy expenditure

    Ref. 38 (as mentioned above) reports deletion of Ptch1 using S100a4-Cre lines and these mice "displayed a disrupted estrous cycle and dramatically reduced fertility over 6.5 weeks". However, this has nothing to do with the approaches where Fsp1/S100a4+ cells are depleted with DTA. Likewise, reference 28 analyzed the phenotype of S00a4-Cre;Ptch1fl/fl mice. Obviously, deleting Ptch1 using S100a4-Cre mice is quite a different approach than "targeting S100A4+ cells" through knockout experiments". Ptch1 deletion leads to a combination of gain-of-function (of Hedgehog activation) and loss-of-function (loss of Hh-independent functions of Ptch1) and hence comparisons with these phenotypes is rather challenging. I suggest the authors focus their phenotype comparisons to ref. 26 where S100a4/Fsp1+ cells were ablated with DTA, i.e. the same approach as in the current study.

    1. I find the Discussion is somewhat off the topic by starting with WHO recommendations on breastfeeding and linking this to observed mouse phenotype. Overall, the discussion is rather long and from time-to-time more like a literature review. I would recommend keeping the Discussion more succinct and focused.

    Minor comments:

    Figure 5 would be more informative if it included more higher magnification images that would reveal the staining at the cellular level.

    Referee cross-comenting

    I agree with the comments of other reviewers. However, to me it seems that the analysis of S100a4 knockout mice would not be feasible within a reasonable timeframe and would represent a study of its own. My understanding was that the authors were not interested in S100a4 itself. Rather, S100a4-Cre was used as a tool to understand the importance of a certain (fibroblast) cell population for mammary gland morphogenesis.

    Significance

    General assessment:

    This study reveals the importance of the S100a4+ cell lineage for nipple formation while showing the same cells are dispensable for mammary gland morphogenesis. The main limitation is that it remains unclear whether the observed nipple phenotype is derived from an early embryonic/prepubertal defect in establishing the nipple stroma, is caused by a constitutive response to ongoing cell death, or a response to continuous DTA expression (or a combination of some of these). Hence its relevance as a model of human inverted nipple condition remains rather speculative.

    Advance:

    This study provides novel information on nipple morphogenesis, with potential (though with reservation) relevance to the congenital human inverted nipple condition affecting 3-5% of women.

    Audience:

    This work should appeal to mammary gland biologists interested in mammary gland development and nipple formation; those with interest on fibroblasts biology given that S100a4 was once thought to be a broad marker of fibroblasts, as well as those with interest in the inverted nipple condition.

    My expertise:

    Mammary gland morphogenesis, developmental biology, cell signaling

  5. Version published to 10.64898/2026.01.26.701676 on bioRxiv