Application of nucleoside analogue labelling to study the cell cycle of xenografted PDAC cell lines in the chorioallantoic membrane model

  1. Robin Colenbier
  2. Denisa Lodoaba Jurescu
  3. Jean-Pierre Timmermans
  4. Johannes Bogers

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 chorioallantoic membrane (CAM) model is an underutilised alternative animal model within the scope of cell cycle-related research of tumour xenografts. The usefulness of nucleoside labelling in standard rodent xenograft models is limited due to the extended labelling durations as a result of intraperitoneal route of administration. Due to its easy accessibility, the CAM allows xenograft S-phase nuclei to be labelled in as little as 30 minutes in vivo for a large number of biological replicates in parallel. We show that for the BxPC-3 and AsPC-1 cell lines, nucleoside labelling with 5-ethynyl-2’-deoxyuridine (EdU) can be multiplexed successfully with other cell-cycle markers such as cyclin B1 and Ki67, especially when combined with digital image analysis techniques. The latter also allows for accurate human versus chicken cell segmentation. Moreover, starting from embryonic day of development 14 (ED14), we observe the presence of a chicken embryonic cell type that appears to possess a high-quantity of extranuclear accumulation of EdU. Initial assessment of these cells showed that they are likely (non-proliferating) granulocytes which can be found in the embryonic liver of grafted and non-grafted embryos, as well as in xenograft sections. Importantly, these cells do not express chicken MHC II, in turn making it less likely that they represent professional antigen presenting cells. Our findings demonstrate that the CAM xenograft model can be used as a valuable tool to study the cell cycle of tumour cells in vivo . Lastly, we potentially revealed a novel biological phenomenon, namely that of extranuclear nucleoside accumulation in certain chicken embryonic cells.

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

    We thank the Referees for their valuable input and critical review of our manuscript. Please find below our responses to the provided feedback.

    Referee # 1

    1.This is a well-done study. However, I ask to the Authors to include also figures showing the macroscopic in ovo evidence

    Response: Two photographs, each depicting an in situ xenograft tumour of the BxPC-3 and PANC-1 cell line have been added as a new figure (new Figure 1).

    and to discuss the role of the CAM assay in the study of xenografts.

    Response: In the introduction, we briefly explained in more detail the rationale behind the use of the CAM model for xenografting and pointed out an important limitation of the use of nucleoside labeling within this model.

    Referee # 2

    Major comments

    1.The manuscript's clinical relevance is limited, and methodological flaws prevent proper statistical validation of the in vitro findings.

    Response: We would like to clarify that the aim of this paper is to share some important preclinical scientific findings highly relevant for future translational studies. First, we demonstrate that cell cycle labelling can be performed in CAM model xenografts via the application of nucleoside analogues. Additionally, we provide evidence of a novel biological phenomenon that on itself could entail an additional important limitation to the use of nucleoside labelling in the CAM model. Additionally, we have only provided two images concerning in vitro culture of cell lines which serve as an illustration for the similarity to the *in ovo *growth pattern and thus are unsure what statistical validation is required for this aspect of the manuscript.

    2.The entire article is based on the assertion that "... " However, even here, the authors do not provide any evidence that these are erythrocytes and not some other cells.

    Response:

    We completely agree with the reviewer that not all chicken (embryonic) cells produce IgY. However, we would like to point out that the majority (if not all) IgY that is present within the developing embryo at the time of the CAM assay is derived originally from the hen and passed through to the embryo via the egg yolk in order to provide passive immunity. In fact, IgY production by embryo/chick cells is commonly observed only several days post hatching.1-3 Thus, IgY detection in chicken embryonic tissues remains species-specific. Though, it appears that large quantities of IgY are sequestered in the embryonic connective tissues. As such, this helps in segregation of 'chicken/stromal' pixels versus 'human/xenograft' pixels. In the revised manuscript, Figure 6 now includes quantitative data regarding the cellular origins with the different strategies described to substantiate our claims.

    Since application of this principle is of course not exclusively linked to FITC as a fluorophore, we performed additional immunofluorescent labelling on xenograft sections using Cy3-labelled Donkey anti-chicken IgY-antibodies as antibody staining to demonstrate the principle and reproducibility of using anti-IgY immunofluorescence. Additionally, using an anti-goat IgG-directed secondary antibody we demonstrate that, in the latter case, a completely different and non-specific staining pattern is obtained compared to that which is observed following anti chicken IgY labelling. Lastly, we added an extra image demonstrating the differences in staining pattern and intensity between xenograft cells and embryonic tissues and embryonic CAM epithelial cells. These images are provided as supplementary Figure S1.

    We have further nuanced the use of the secondary anti-chicken IgY labelling to the identification of chicken stroma rather than generalising the detection method to include specific labelling of all chicken cells. Though, in our experience, human cancer cells do not exhibit substantial levels of autofluorescence in the orange to far red spectrum within the xenografts. In contrast, some autofluorescence in the blue-green spectrum is observed for chicken tissues and some cells in non-immunolabelled sections of non-grafted embryos). Therefore, we reserve the spectral regions with the least amount of autofluorescence (i.e., orange-far red) for subsequent (IF)-labelling for xenograft cells in order to maintain sufficient specificity. In other words, the FITC-staining aids in determining whether a pixel is chicken in identity rather than human, in our tissue preparations.

    1. Dias da Silva, W. & Tambourgi, D. V. IgY: A promising antibody for use in immunodiagnostic and in immunotherapy. Vet. Immunol. Immunopathol. 135, 173-180 (2010).
    2. Ulmer-Franco, A. M. Transfer of Chicken Immunoglobulin Y (IgY) from the Hen to the Chick. Avian Biol. Res. 5, 81-87 (2012).
    3. Carlander, D., Wilhelmson, M. & Larsson, A. Immunoglobulin Y Levels in Egg Yolk From Three Chicken Genotypes. Food Agric. Immunol. 15, 35-40 (2003). Since application of this principle is of course not exclusively linked to FITC as a fluorophore, we performed additional immunofluorescent labelling on xenograft sections using Cy3-labelled Donkey anti-chicken IgY-antibodies as antibody staining to demonstrate the principle and reproducibility of using anti-IgY immunofluorescence. Additionally, using an anti-goat IgG-directed secondary antibody we demonstrate that, in the latter case, a completely different and non-specific staining pattern is obtained compared to that which is observed following anti chicken IgY labelling. Lastly, we added an extra image demonstrating the differences in staining pattern and intensity between xenograft cells and embryonic tissues and embryonic CAM epithelial cells. These images are provided as supplementary figure S1.

    We have further nuanced the use of the secondary anti-chicken IgY labelling to the identification of chicken stroma rather than generalising the detection method to include specific labelling of all chicken cells. Though, in our experience, human cancer cells do not exhibit substantial levels of autofluorescence in the orange to far red spectrum within the xenografts. In contrast, some autofluorescence in the blue-green spectrum is observed for chicken tissues and some cells in non-immunolabelled sections of non-grafted embryos). Therefore, we reserve the spectral regions with the least amount of autofluorescence (i.e., orange-far red) for subsequent (IF)-labelling for xenograft cells in order to maintain sufficient specificity. In other words, the FITC-staining aids in determining whether a pixel is chicken in identity rather than human, in our tissue preparations.

    3.The latter is actually confirmed by the authors in Fig. 7 in the form of the following sentence: "Highly autofluorescent (nucleated) embryonic erythrocytes can be observed throughout the tissue (arrowheads)." However, even here, the authors do not provide any evidence that these are erythrocytes and not some other cells.

    Response: We agree with the reviewer that the use of the green emission spectra in fluorescence should be used with caution, especially when evaluating tissue samples. In fact, this is the reason why we reserved the higher wavelength channels for anti-human fluorescent detection or the (click-based) detection of nucleosides.

    Indeed, we did not present full identification of these cells, but their avian origin is undisputed as these cells are also observed in the same quantity in tissues on non-grafted embryos. However, we now adapted the wording in the manuscript to the more general term 'chicken blood cells'. Throughout the manuscript and discussion section, we now elaborate on the possibility that nucleated cells, observed within blood vessels of the CAM can be erythrocytes, leukocytes or thrombocytes (which also are nucleated in avians).

    *4. *Tissue-specific and species-specific monoclonal antibodies to avian red cell nuclear proteins. Proceedings of the National Academy of Sciences of the United States of America, 79(20), 6265-6269. https://doi.org/10.1073/pnas.79.20.6265], and I recommend the authors use these instead of the fluorescein-labeled donkey anti-chicken IgY antibodies, which were misused. On the same matter, the article doesn't clarify if the antibodies for Ki67 and cyclin B1 can differentiate between human and chicken antigens.

    Response: Other groups have also applied this anti Ki-67 primary antibody in CAM-model studies without clear evidence of cross-reactivity to chicken embryonic nuclei.4,5 We have not observed ourselves any substantial species cross-reactivity for both the applied primary anti-Ki-67 antibody nor the primary anti-cyclin B1 which is also illustrated by the images provided within the manuscript.

    Jarrosson, L. et al. An avian embryo patient-derived xenograft model for preclinical studies of human breast cancers. iScience 24, 103423 (2021). Javed, S., Soukhtehzari, S., Fernandes, N. & Williams, K. C. Longitudinal bioluminescence imaging to monitor breast tumor growth and treatment response using the chick chorioallantoic membrane model. Sci. Rep. 12, 17192 (2022).

    5. Why is there no staining with antibodies to human antigens for tumor cell identification in Figures 1-2?

    Response: The aim of this figure was to demonstrate the highly pleiotropic nuclear morphology of the PANC-1 cell line compared to BxPC-3 cells while growing *in vitro *monocultures in turn comparing the growth pattern to that observed in.

    *6. *What specific markers in Fig. 4 should lead the reader to conclude that the nuclei pointed out by the arrows represent embryonic epithelial nuclei (EE), nuclei of chicken embryonic blood cells (EB), and human AsPC-1 tumor xenograft nuclei (T)?

    Response: The identification of different nuclei within these images was performed on the basis of their morphological aspects (i.e. irregular shape, larger size...) in combination with the relative localisation of these nuclei within the tissue section and the typical nucleolar staining pattern. The latter is not observed in any of the human xenografts we routinely perform. Nevertheless, we agree that accurate discrimination between embryonic epithelial nuclei and embryonic blood cells cannot be guaranteed with absolute certainty in the absence of the use of additional markers. However, the morphological aspects combined with the anti-Ki-67 staining, strongly suggest the human identity of the annotated nuclei. We have modified Figure 4 (now Figure 3) to include Anti-Ki67 staining. Additionally, an annotation that was erroneously pointing out a mitotic figure was omitted. For the remaining mitotic nuclei, clear perichromosomal localisation of Ki-67 is observed which supports the claim that these are mitotic nuclei.

    7. The nuclei shown in this same figure, which supposedly display mitotic figures of dividing tumor nuclei and can also be clearly distinguished (M), actually more closely resemble giant multinucleated cells.

    Response:

    We have added the immunofluorescent staining for Ki-67 of these images (new Figure 3); these nuclei present with different patterns for the marker, which would be highly unlikely in a single multinucleated cell. Further, we have altered the description of embryonic epithelial cells to the more general term embryonic cells (E). Additionally, annotation of embryonic cells was omitted from the lower panels in order to draw the focus on the polymorphic nature of the human tumour nuclei rather than the embryonic surrounding cells.

    8. The statement "After immunolabeling for human Ki67, we confirmed that all EdU+ xenograft nuclei were also Ki67+, confirming the specificity and compatibility of both labeling strategies in CAM xenograft tumor cells (Figure 5)" is not supported by the image, which shows that the majority of Ki67-positive nuclei are EdU-negative.

    Response:

    During cell cycle progression, nuclear levels of Ki-67 gradually increase during S-phase and peak during mitosis. As a consequence, all S-phase cells (EdU+ nuclei) will present detectable nuclear levels of Ki-67. The reverse is not necessarily true: the nuclei of cells that are within G2 and mitosis during the nucleoside labelling will present with high levels of nuclear Ki-67 but will not show nuclear incorporation for EdU as these have already completed replication during the S-phase. Figure 4 now shows quantification of both markers across 7 different tumour BxPC-3 xenografts. With the applied classification strategy, 94% of the detected EdU+ nuclei were also Ki-67+. Irrespective of Ki-67 positivity, 44% of BxPC-3 nuclei were calculated to be EdU+ for the labelling duration of 1 hour.

    9. The caption for Figure 6 needs to be revised, particularly the statement "Combined, these markers allow in ovo segregation of proliferating tumor cells into early S-phase (ES, EdU+), late S or early G2 (LS, EdU+CB1+), and G2 (EdU-CB1+)".

    Response: We have altered the phrasing to stress the cytoplasmic presence of cyclin B1 as indicative for late S-phase and G2 phase.

    1. What do the pink cell nuclei in Figures 8 and 9 represent?

    Response: The orange immunofluorescent staining in these images demonstrates anti-Ki-67 labelling. Due to overlap with the nuclear (blue) staining the appearance may have a pink undertone but this does not alter the interpretability of the images.

    1. In Fig. 10, it is impossible to see the yellow line, which, according to the authors' statement ". yellow lines indicate the area classified as tumor", should indicate the tumor origin of the cells!

    Response: In the submitted PDF version of the manuscript, a yellow line is clearly visible in panels C and F. Perhaps due to compression-related quality loss, its presence is more difficult to see. We now submitted a higher resolution image.

    1. There is a distinct lack of evidence in Figures 11A-C suggesting that the S-phase nuclei are attributed to both embryonic liver epithelial cells (ES) and BxPC-3 cells (BS). Furthermore, there is no evidence that the apparent cytoplasmic EdU inclusions (arrowheads) belong specifically to chicken embryonic cells.

    Response:

    Concerning the identity of the cells in Figure 11 A-B (Now Figure 7). We would like to clarify that these images are taken from embryos which had not been subjected to tumour grafting. Therefore, the presence of any human tumour cells within these liver sections can 100% be excluded. This has now also been stressed in the text and the image caption. With respect to Figure 11C (now 7C), we have noticed that this image contained a tissue processing artefact which may lead some readers to question its authenticity. Therefore, we have replaced this panel (7C, now 7C) with another image taken from the same tissue section and included the anti-IgY staining channel to allow identification of CAM tissue.

    We have addressed the confusion regarding the apparent different magnifications in the figure legend: all images presented in Figure 7 were acquired using a 40x magnification objective. In order to focus on some select regions, different digital zoom levels are present for each panel. To account for this, each panel is now annotated with its own scale bar.

    1. As no macroscopic images depicting tumor nodules from the implantation of AsPC-1 and PANC-1 tumor cells into the CAM were provided.

    Response: A new figure (now Figure 1) has been added with macroscopic images of PANC-1 and BxPC-3 tumours in situ at ED14. Successful tumour grafting of AsPC-1 cells was demonstrated via the histological images provided throughout the manuscript.

    1. A newly identified biological phenomenon: non-nuclear EdU accumulation in chicken embryonic cells," is not well supported unless compelling evidence is presented to establish that these cells indeed belong to chicken embryos

    Response: We have clarified in the body of the text as well as in various figure legends that the apparent nucleoside presence within the cytoplasm of cells has been consistently observed also in liver sections taken from embryos that have not received tumour grafts.

    15. Lacking any demonstration of concurrent EdU accumulation alongside cytoplasmic and/or membrane staining within the same cells. The latter is quite feasible by staining the cells with a suitable agent (or fluorophore), such as...

    Response: We agree with the reviewer that we have not performed dedicated staining for cytoplasmic or membrane components in order to demonstrate the colocalisation of the EdU signal with the . Though we believe that the simultaneously acquired brightfield images are sufficiently convincing that these signals localise to the cytoplasm rather than to the extracellular space. Additional experiments were performed with F-ara-EdU and BrdU/IdU labelling in non-grafted embryos which demonstrated the robustness of these findings and further point towards cytoplasmic rather than extracellular signal. Multiple new figures and paragraphs were added into the revised manuscript which support our claim.

    1. Third, regarding "Extranuclear EdU staining": Extranuclear EdU staining is not a standard or typical...

    Response: We agree that this is not the intended use of nucleoside incorporation assays (including halogenated analogues). With this manuscript we aimed to illustrate the unexpected findings that could lead to misinterpretation of experimental data as a consequence of a new biological phenomenon.

    16.1* Cellular damage or death: Cells that are dying or have been damaged may release their DNA, causing it to be detected outside the nucleus.*

    Response: The cells with apparent extranuclear (F-ara-)EdU, and BrdU and IdU) show no signs of nuclear pyknosis nor DNA fragmentation. In addition, we also describe this phenomenon in healthy developing embryos at various stages of development, in livers that show no macroscopic signs of tissue damage. Moreover, if the extranuclear EdU detection does signify degraded DNA (with the incorporated analogue), the remaining nuclear DNA would also show signs of (replication-dependent) EdU incorporation.

    16.2* Apoptosis: During apoptosis, cells undergo DNA degradation, and some DNA fragments may be found in the extranuclear space*

    Response: We would like to refer the reviewer to the reply above. It seems unlikely that from ED14 onwards such massive apoptotic evens would be taking place in healthy embryos. Moreover, nucleoside incorporation and (caspase-dependent) DNA fragment generation would need to take place within one hour (= a typical nucleoside labelling period).

    16.3 Technical artifacts: Errors during cutting of formalin-fixed tissues, cell fixation, permeabilization, or the staining procedure itself can lead to mislocalization of the fluorescent signal.

    Response: We agree that several technical reasons can lead to altered subcellular localisation during the detection of some (protein) markers. Iduring detection.6,7 In our experiments, tissues were fixed in ice-cold 4% formalin for at least 12 hours. Therefore, it is highly unlikely that insufficient fixation would have caused this phenomenon.

    Regarding the possibility that the staining procedure itself can lead to mislocalisation, it is important to note that the copper-catalysed click detection of EdU is not susceptible to many of the possible causes for altered epitope detection through standard immuno-labelling. We can also not conceive how the click-reaction, performed on formalin-fixed tissues, could induce a complete shift in EdU from the nuclear to the extranuclear compartment of these select cells only without any other alterations in the nuclear or cellular morphology. We have performed several control staining procedures, such as performing the click reaction on tissue sections of embryos that had not received any alkyne-containing nucleoside labelling. This did not result in the detection of any signal. These results are referred to in the manuscript.

    Lastly, the simultaneous presence of correctly localised (nuclear) (F-ara-)EdU and 'incorrectly' localised (F-ara-)EdU within the same tissue section, and across different tissue types (liver versus CAM xenograft) demonstrates that it is unlikely that the observed phenomenon is the result of a technical artifact.

    Yoshida, S. R., Maity, B. K. & Chong, S. Visualizing Protein Localizations in Fixed Cells: Caveats and the Underlying Mechanisms. J. Phys. Chem. B 127, 4165-4173 (2023). Stadler, C., Skogs, M., Brismar, H., Uhlén, M. & Lundberg, E. A single fixation protocol for proteome-wide immunofluorescence localization studies. J. Proteomics 73, 1067-1078 (2010).

    16.4 Unique cell types: In some specific cell types, such as megakaryocytes, DNA or other nuclear components may be localized extranuclearly, leading to extranuclear staining with techniques like EdU.

    Response: Referring to the paper of Lan et al. (2019; doi: 10.1111/acel.12901) to which the reviewer alluded by citing the following statement "*Nicked DNA was strongly visible in old cells, prominently in the cytosol, but undetectable in young cells, and was more intense in old cells upon" *, we would like to stress that these authors investigated the presence of dsDNA in various human cell lines *in vitro *through OTHER assays than nucleoside labelling. Additionally, in the cited paper, also no specific staining for cytoplasm or cell membranes was performed to accurately segment the nucleus versus cytoplasm. Lastly, the intense focal-like nucleoside signal we observe within our tissue sections does not resemble the ambiguous and apparently random signal localisation within the cytoplasm of the cells presented in the cited paper. Lastly, the reviewer seems not to be aware of the fact that megakaryocytes dot NOT occur in avian species.

    16.5 Extranuclear accumulation of histones and nucleosomes is an early event of apoptosis in human lymphoblasts (https://doi.org/10.1136/ard.2003.011452)

    Response: The referred paper investigated cellular processes during cell death (apoptosis). Unfortunately, it does not present microscopic images to compare our data to; we therefore cannot assess its relevance to our findings.

    16.6 In some specific cell types, such as megakaryocytes, DNA or other nuclear components may be localized extranuclearly, leading to extranuclear staining with techniques like EdU [Frydman, G.H., Tessier, S.N., Wong, K.H.K. et al. Megakaryocytes

    Response: As already mentioned above, to our knowledge, megakaryocytes have NOT been described or identified in avian species so far. Therefore, we disagree that the cited phenomenon/reference would be highly relevant for nucleoside accumulation in the cytoplasm of non-megakaryocyte cells in in ovo model systems. Furthermore, the paper by Frydman et al., does not describe the use of nucleoside labelling in these cells.

    16.7 However, since the authors do not present any absolute markers proving that EdU-cyto+ cells are EdU+ chicken granulocytes, the authors' statement that "Given the high number of EdU+ granulocytes observed, it is more likely that these are neutrophils rather than eosinophils" appears highly speculative.

    Response: We agree with the reviewer that our claim that EdU-cyto+ cells represent granulocytes is still speculative. However, concerning the exact wording; brightfield images did show colocalisation of the nucleoside signal within the cytoplasm of granule-containing chicken cells. We have emphasized more clearly within the discussion section that the extranuclear nucleoside signal is highly unlikely to be nucleoside-containing DNA. Additional experiments were conducted to investigate whether these cells represent chicken thrombocytes, known to possess phagocytotic functionality. IF staining for CD41/CD61 as a marker for thrombocytes revealed that CD41/CD61-positive cells do not exhibit the alluded phenomenon. A new figure, Figure 10 illustrates this finding.

    Methodological ambiguities

    1. In the section "Nucleoside labeling of dividing cells in the CAM model": It is not clear when BrdU and EdU were given after tumor cell implantation, or how to standardize the distance from the tumor as suggested by "...as far away from the visible tumor as possible...". I question whether such a "precise" description of the application site would contribute "...to ensure low variability in labeling duration...", especially when it was performed by two independent researchers "...in tandem, ensuring that an experimental group of 16 embryos could be labeled in less than five minutes."

    Response:

    In this manuscript, we have described nucleoside labelling in grafted embryos as well as in non-grafted embryos. Alongside depictions of nucleoside labelling, the embryonic day at which the assay was performed is mentioned in the figure legend of all figures demonstrating nucleoside labelling. For example, BrdU labelling of the BxPC-3 xenograft presented in Figure 2 (revised manuscript) was performed at ED13 (which is 5 days following the grafting procedure at ED7). In the revised manuscript, we have ensured that for each labelling depicted, its timing in the developmental period is explicitly mentioned. Nucleoside labelling in non-grafted embryos was performed throughout various stages of development; the timing of the assay is now also mentioned in the corresponding figure legends.

    We fully agree that there is still room for improvement concerning the standardisation of the application of the labelling solution onto the CAM. The main aim of the present manuscript, however, is to demonstrate the feasibility of this labeling protocol. In addition, we report the unexpected but important finding of cytoplasmic accumulation due to a possible biological cause. In this respect we rephrased the text and pointed to a potential pitfall of the application in the CAM model so that future studies can anticipate to misinterpretations.

    1. In the section "Immunofluorescence labeling of cryosections: "...a blocking and permeabilizing solution in a humidified atmosphere."" ... : Absence of detergent is used during incubation with primary and secondary antibodies (even the widely recommended Tween20!). I wonder whether it might be the source of non-specific tissue labelling by donkey anti-chicken IgY that was considered as an evidence of chicken origin.

    Response: We interpreted this as that the reviewer concluded that we applied our primary and secondary antibody staining procedures in the presence of Triton X-100. We can assure that following the initial blocking incubation step and three rinses with PBS, all subsequent antibody labelling procedures (primary or secondary) are performed in the absence of detergent within the staining buffer. This was mentioned in the second paragraph under the heading "__Immunofluorescence ______labelling of cryosections" ____in the Materials & Methods section.

    Referee #3

    Major comments

    1. The current manuscript lacked figure panels to show the quantification results of the labeling to convincedly demonstrate the positive correlation between the nucleoside labeling versus Ki67. Bar charts with statistical analysis should be included. The study utilized 2 human cancer cell lines as an example. T

    Response: We have added quantification data of EdU and Ki-67 across several BxPC-3 xenografts and added statistical analysis in Figure 4E.

    1. The study would be improved if the authors could include cancer cell lines that are known to be highly proliferative vs slowly proliferative to demonstrate the robustness of this method.

    Response: We do not expect failure of nucleoside labelling due to tumour cell-line dependent characteristics for several reasons. For instance, when performing EdU labelling in vitro, initial dose-and duration titration are recommended to establish satisfactory labelling efficiency for the research hypothesis in question. In our research, we are predominantly focusing on relatively short-lived effects of treatments on S-phase progression and therefore prefer to apply labelling (in vitro and in ovo) for durations of less than 60 minutes. In contrast, extended labelling durations of several hours will allow one to capture replication even in very slowly proliferating (cancer) cells at the cost of temporal resolution. Importantly, the growth rates of individual cell lines within the CAM model may differ from their behaviour in vitro. Additionally, long duration nucleoside (EdU) labelling may induce additional toxicity. Alternative analogues such as F-ara-EdU with improved safety profiles for long labelling durations should then be considered. In the revised manuscript, labelling with BrdU/IdU as well as F-ara-EdU of non-grafted chicken tissues were added which demonstrate the robustness of the application of the technique in the CAM model. Longer-term labelling of lowly proliferative cells can be achieved with F-ara-EdU to minimise embryo (and tumour cell) toxicity which will extend the application potential of the described .

    In summary, labelling parameters should be optimised for each specific research question, but nucleoside incorporation remains a gold standard for identifying S-phase progression across diverse cell types. Given that we successfully established its feasibility in the CAM model, there is no biological reason to suggest it would not be equally valid for any proliferating cancer cell line.

  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

    This study explored the utilization of nucleoside labelling in the human cancer cell xenografts grown in the chick CAM. The authors tried 2 different reagents (BrdU and EdU) which would label S-phase proliferative cells in a species non-specific way. The authors concluded that EdU labeling reliably detected Ki67 positive cells compared to BrdU labeling which would detect non-Ki67 positive cells as well (might be due to dsDNA denaturing). In order to distinguish the human versus the chick cells, the authors further utilized chicken-specific antibody anti-IgY and the segmentation algorithm in QPath to distinguish cells of the two species. This allows the authors to develop (supervised) automated or manual annotation for individual cell detection, in this case the proliferating Ki67+ human cancer cells. The study also showed an unexpected finding of cytoplasmic positive EdU cells in the embryonic chicken liver, which the authors speculated to be non-antigen presenting granulocytes.

    Major comments:

    This study provided a good methodology path to analyze and to quantify proliferating human cancer cells inside the CAM xenograft. The current manuscript lacked figure panels to show the quantification results of the labeling to convincedly demonstrate the positive correlation between the nucleoside labeling versus Ki67. Bar charts with statistical analysis should be included. The study utilized 2 human cancer cell lines as an example. The study would be improved if the authors could include cancer cell lines that are known to be highly proliferative vs slowly proliferative to demonstrate the robustness of this method.

    Minor comments:

    The authors used parentheses for the subheadings inside the discussion section. This is not an usual practice unless it is required by the journal formatting requirement.

    Significance

    General assessment: The strength of this study is to develop a straight forward solution to detect human-specific proliferative cancer cells inside the chicken CAM xenograft. The data presented in the IF staining and segmentation results were clear. The limitation is that the study only tested 2 human cancer cell lines and it is unknown whether the current method is robust enough in a pan-cancer setting.

    Advance: This study showed good advances in the methodology for the CAM xenograft field.

    Audience: The audience of this study would include researchers in cancer biology, imaging processing, and 3R.

    My expertise: cancer biology, CAM, systems and spatial biology

  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

    An intriguing article, "Application of Nucleoside Analogue Labelling to Study the Cell Cycle of Xenografted PDAC Cell Lines in the Chorioallantoic Membrane Model," explores the overlooked potential of the chorioallantoic membrane (CAM) model as an innovative alternative in the realm of cell cycle research related to tumor xenografts. The authors were resolute in their pursuit of evidence to bolster their fascinating hypothesis, which posits that exposure to nucleosides enhances the labelling of nucleosides with 5-ethynyl-2'-deoxyuridine (EdU). This improvement allows effective multiplexing with cell-cycle markers like cyclin B1 and Ki67, especially when combined with advanced digital image analysis. They claimed that accurately separating human and chicken cells allowed them to identify a specific type of chicken embryonic cell with a high level of extranuclear EdU accumulation. Researchers propose that cells without chicken MHC II are likely to be non-proliferating granulocytes found in the embryonic liver of both grafted and non-grafted embryos, including xenograft tissues. According to the authors, the CAM xenograft model effectively helps in studying tumor cell cycles in live conditions. Behind, authors claim a novel biological phenomenon, namely that of extranuclear nucleoside accumulation in certain chicken embryonic cells. If the authors successfully prove their hypothesis, it will significantly confirm CAM as a unique in vivo tool for better immuno-/histological exams and more precise cell cycle assessments compared to standard rodent models. The research encourages greater use of this alternative animal model in cancer studies to help us understand the regulation of the cancer cell cycle. This, in turn, may improve the implementation of existing treatment methods or uncover potential vulnerabilities in the cancer cell cycle.

    Unfortunately, in this article, the authors do not provide sufficiently convincing evidence of the hypothesis postulated by them. The manuscript's clinical relevance is limited, and methodological flaws prevent proper statistical validation of the in vitro findings. My questions and remarks below are designed to find undeniable proof or opposing views that back the authors' key perspective.

    Major - Conceptual:

    Point 1 - In Result section: The entire article is based on the assertion that "... distinguishing between both species can already be achieved to a large extent in a relatively simple manner through the use of anti-chicken IgY fluorophore-conjugated antibodies, which are routinely used only as secondary antibodies. In doing so, cellular as well as acellular chicken embryonic components are stained (Figure 7)." This assumption seems off, as anti-chicken IgY staining is employed in research to detect chicken antibodies (IgY). This secondary antibody is designed to specifically bind to the IgY immunoglobulin in chickens, targeting both heavy and light chains. A donkey anti-chicken IgY antibody labels chicken tissues to identify cells that express chicken IgY immunoglobulins. It's clear that not every cell in a chicken embryo tissues produces IgY immunoglobulins. Hence, cells stained by donkey anti-chicken IgY FITC-conjugated antibody cannot represent all chicken tissue cells. Lastly, the use of the green fluorescence channel is often unfounded, considering that there are always cells with non-specific and considerably high fluorescence levels in this area of the spectrum. The latter is actually confirmed by the authors in Fig. 7 in the form of the following sentence: "Highly autofluorescent (nucleated) embryonic erythrocytes can be observed throughout the tissue (arrowheads)." However, even here, the authors do not provide any evidence that these are erythrocytes and not some other cells.

    Point 2 - In their article, the authors consistently emphasize how their method excels in differentiating human tumor cells from the abundant chicken embryonic cells that encase the tumor. Nonetheless, the authors fail to present direct evidence regarding the identity of the nuclei or cells in question, nor do they substantiate the validity of the algorithm selected for processing digital images of fluorescently labeled cells, which would be essential for discerning their origin, whether chicken or human. Given the importance of the authors' subsequent conclusions, such evidence should be provided, at least for initial validation. Antibodies for chicken nuclear antigens are well established [e.g., Kane, C. M., Cheng, P. F., Burch, J. B., & Weintraub, H. (1982). Tissue-specific and species-specific monoclonal antibodies to avian red cell nuclear proteins. Proceedings of the National Academy of Sciences of the United States of America, 79(20), 6265-6269. https://doi.org/10.1073/pnas.79.20.6265], and I recommend the authors use these instead of the fluorescein-labeled donkey anti-chicken IgY antibodies, which were misused. On the same matter, the article doesn't clarify if the antibodies for Ki67 and cyclin B1 can differentiate between human and chicken antigens. Why is there no staining with antibodies to human antigens for tumor cell identification in Figures 1-2? What specific markers in Fig. 4 should lead the reader to conclude that the nuclei pointed out by the arrows represent embryonic epithelial nuclei (EE), nuclei of chicken embryonic blood cells (EB), and human AsPC-1 tumor xenograft nuclei (T)? The nuclei shown in this same figure, which supposedly display mitotic figures of dividing tumor nuclei and can also be clearly distinguished (M), actually more closely resemble giant multinucleated cells. However, their affiliation with either chicken or human tumor cells is neither obvious nor proven. The statement "After immunolabeling for human Ki67, we confirmed that all EdU+ xenograft nuclei were also Ki67+, confirming the specificity and compatibility of both labeling strategies in CAM xenograft tumor cells (Figure 5)" is not supported by the image, which shows that the majority of Ki67-positive nuclei are EdU-negative. The caption for Figure 6 needs to be revised, particularly the statement "Combined, these markers allow in ovo segregation of proliferating tumor cells into early S-phase (ES, EdU+), late S or early G2 (LS, EdU+CB1+), and G2 (EdU-CB1+)". The authors' method for labeling cells should consider that the location of cyclin B1 is key to determining a cell's stage in the cell cycle: o Cytoplasmic: Associated with G2/M arrest. o Nuclear: Associated with the transition into mitosis and the G2 to M phase. The assertion that "Species-distinction can be further facilitated by combining anti-chicken IgY IF (Figure 8C) with additional anti-human IF, such as anti-human Ki67 (Figure 8D)" lacks clarity, as the images in Figure 8 do not support the authors' claim. What do the pink cell nuclei in Figures 8 and 9 represent? In Fig. 10, it is impossible to see the yellow line, which, according to the authors' statement "... yellow lines indicate the area classified as tumor", should indicate the tumor origin of the cells! There is a distinct lack of evidence in Figures 11A-C suggesting that the S-phase nuclei are attributed to both embryonic liver epithelial cells (ES) and BxPC-3 cells (BS). Furthermore, there is no evidence that the apparent cytoplasmic EdU inclusions (arrowheads) belong specifically to chicken embryonic cells. Furthermore, in Fig. 11 B and C, the equal magnification of 40x is apparently incorrectly written.

    Discussion section:

    The initial assertion made by the authors, "In this paper, we have demonstrated that our previously published protocol concerning the xenografting of the BxPC-3 cell line10 can be applied to the AsPC-1 and PANC-1 cell lines," seems inadequately supported, as no macroscopic images depicting tumor nodules from the implantation of AsPC-1 and PANC-1 tumor cells into the CAM were provided.

    The authors' second assertion, "A newly identified biological phenomenon: non-nuclear EdU accumulation in chicken embryonic cells," is not well supported unless compelling evidence is presented to establish that these cells indeed belong to chicken embryos. Moreover, the assertion regarding non-nuclear EdU accumulation seems to be speculative, lacking any demonstration of concurrent EdU accumulation alongside cytoplasmic and/or membrane staining within the same cells. The latter is quite feasible by staining the cells with a suitable agent (or fluorophore), such as fluorophore-conjugated Phalloidin for the cytoskeleton or PKH26 (https://www.sigmaaldrich.com/RU/en/product/sigma/pkh26gl?srsltid=AfmBOorEbKBTeYSCKZY6qs-pWjCZCg4lhOvNqE0YYByS2A545f-POa24) for membrane structures.

    Third, regarding "Extranuclear EdU staining": Extranuclear EdU staining is not a standard or typical application for the EdU (5-ethynyl-2′-deoxyuridine) assay, which is designed to label and detect newly synthesized DNA in the nucleus during the S-phase of the cell cycle. Extranuclear EdU staining likely refers to a misinterpretation or unusual experimental result where EdU or its detection reaction product is found outside the nucleus. This could be due to cell damage, processing issues, tissue cutting artefacts or a cell type with unique DNA localization, as seen in some contexts of apoptosis or extranuclear DNA accumulation. Potential reasons for extranuclear EdU staining may include, but not limited to:

    • Cellular damage or death: Cells that are dying or have been damaged may release their DNA, causing it to be detected outside the nucleus. • Apoptosis: During apoptosis, cells undergo DNA degradation, and some DNA fragments may be found in the extranuclear space.
    • Technical artifacts: Errors during cutting of formalin-fixed tissues, cell fixation, permeabilization, or the staining procedure itself can lead to mislocalization of the fluorescent signal.
    • Unique cell types: In some specific cell types, such as megakaryocytes, DNA or other nuclear components may be localized extranuclearly, leading to extranuclear staining with techniques like EdU. Several examples:
    • Nicked DNA was strongly visible in old cells, prominently in the cytosol, but undetectable in young cells, and was more intense in old cells upon induction of DNA damage by the DNA damaging agent cytarabine/Ara‐C which causes DSBs [Lan YY, Heather JM, Eisenhaure T, Garris CS, Lieb D, Raychowdhury R, Hacohen N. Extranuclear DNA accumulates in aged cells and contributes to senescence and inflammation. Aging Cell. 2019 Apr;18(2):e12901. doi: 10.1111/acel.12901].
    • Extranuclear accumulation of histones and nucleosomes is an early event of apoptosis in human lymphoblasts. [Gabler, C., Blank, N., Hieronymus, T., Schiller, M., Berden, J. H., Kalden, J. R., & Lorenz, H. M. (2004). Extranuclear detection of histones and nucleosomes in activated human lymphoblasts as an early event in apoptosis. Annals of the rheumatic diseases, 63(9), 1135-1144. https://doi.org/10.1136/ard.2003.011452]
    • In some specific cell types, such as megakaryocytes, DNA or other nuclear components may be localized extranuclearly, leading to extranuclear staining with techniques like EdU [Frydman, G.H., Tessier, S.N., Wong, K.H.K. et al. Megakaryocytes contain extranuclear histones and may be a source of platelet-associated histones during sepsis. Sci Rep 10, 4621 (2020). https://doi.org/10.1038/s41598-020-61309-3].

    I would, if I could, strongly concur with the statements made by the authors: "It is also evident that the presence of EdU-cyto+ cells is not confined to the liver, as they were also found within the CAM. These cells likely represent a ubiquitously distributed cell population present in both healthy and xenografted chicken embryos," and "...since EdU-cyto+ cells are consistently present in non-grafted embryos as well, the phenomenon is not triggered by xenografting or the presence of PDAC cells." However, since the authors do not present any absolute markers proving that EdU-cyto+ cells are EdU+ chicken granulocytes, the authors' statement that "Given the high number of EdU+ granulocytes observed, it is more likely that these are neutrophils rather than eosinophils" appears highly speculative.

    Fourth, regarding Ki67 staining: Careful consideration and robust validation are essential when drawing conclusions and interpretations about Ki67's role in cell proliferation and the cell cycle. Research shows that Ki-67, an important cell cycle marker, has two main splice variants, α and β, which are regulated differently in normal and cancer cells at mRNA and protein levels. Moreover, Ki-67 undergoes constant regulation and degradation through the proteasome system in both cell types, indicating a dynamic control mechanism for this protein. It was also observed a putative extranuclear elimination pathway of Ki-67, where it is transported to the Golgi apparatus. Furthermore, the unforeseen extranuclear removal of Ki-67 strongly indicates the necessity to examine this protein beyond the confines of the "nuclear box," a perspective that has been overlooked until now [see e.g., Chierico L, Rizzello L, Guan L, Joseph AS, Lewis A, Battaglia G (2017) The role of the two splice variants and extranuclear pathway on Ki-67 regulation in non-cancer and cancer cells. PLoS ONE 12(2): e0171815. https://doi.org/10.1371/journal.pone.0171815].

    Last but not least, regarding the authors' conclusion: "We report a novel phenomenon: the apparent cytoplasmic accumulation of EdU in nondividing chicken granulocytes." The phenomenon of DNA replication taking place in the cytoplasm is not a novel observation. In instances like replication stress, the cytosol can initiate a response pathway that encompasses the detection of cytosolic DNA and subsequent signaling processes focused on genome protection. Cytosolic DNA generated after replication stress activates a Ca2+-dependent pathway to protect stalled replication forks [Li, S., Lu, H. T., & You, Z. (2025). Cytosolic DNA and intracellular Ca2+: Maintaining genome stability during replication stress. DNA repair, 152, 103877. https://doi.org/10.1016/j.dnarep.2025.103877]. Damage to the DNA template caused by environmental pollutants, like radiation and genotoxic chemicals, can impede the replication process.Physiological stressors also affect fork dynamics, including metabolic byproducts like reactive oxygen species, conflicts in replication and transcription, repetitive DNA elements such as telomeres, sequences that form secondary structures, DNA-RNA hybrids, misincorporated ribonucleotides, and low availability of DNA precursors. In response to these challenges, cells have developed an intricate network of surveillance and repair mechanisms. They identify replication stress, support stalled forks, fix problems, and allow replication to proceed. These pathways are crucial for maintaining genome stability and ensuring proper cellular function. Several excellent reviews on the topic included [M.R. Higgs BOD1L is required to suppress deleterious resection of stressed replication forks Mol. Cell (2015) R. Kumar et al. RIF1: a novel regulatory factor for DNA replication and DNA damage response signaling DNA Repair(2014) W. Leung ATR protects ongoing and newly assembled DNA replication forks through distinct mechanisms Cell Rep.(2023) M.B. Adolph et al. Mechanisms and regulation of replication fork reversal DNA Repair (Amst. )(2024) Z. You et al. The role of single-stranded DNA and polymerase alpha in establishing the ATR, Hus1 DNA replication checkpoint J. Biol. Chem.(2002) A. Kumagai TopBP1 activates the ATR-ATRIP complex Cell(2006) J. Lee et al. The Rad9-Hus1-Rad1 checkpoint clamp regulates interaction of TopBP1 with ATR J. Biol. Chem.(2007)]. To validate the authors' conclusions regarding the presence or absence of the declared phenomenon, I would recommend modulating TREX1 expression. Overexpression of TREX1, a nuclease that degrades cytosolic DNA, suppresses TRPV2-mediated Ca2+ release under replication stress. TREX1 depletion, however, leads to cytosolic DNA accumulation.

    Methodological ambiguities: In the section "Nucleoside labeling of dividing cells in the CAM model": It is not clear when BrdU and EdU were given after tumor cell implantation, or how to standardize the distance from the tumor as suggested by "...as far away from the visible tumor as possible...". I question whether such a "precise" description of the application site would contribute "...to ensure low variability in labeling duration...", especially when it was performed by two independent researchers "...in tandem, ensuring that an experimental group of 16 embryos could be labeled in less than five minutes."

    In the section "Immunofluorescence labeling of cryosections": "...a blocking and permeabilizing solution in a humidified atmosphere." This solution is comprised of final concentrations of 0.1% Triton-X-100, 0.02% sodium azide, 5% horse serum (Sigma-Aldrich, Cat#H1270), 0.01% thimerosal, and 0.3% bovine serum albumin (Sigma-Aldrich, Cat#A7284) in PBS at pH 7.4." Absence of detergent is used during incubation with primary and secondary antibodies (even the widely recommended Tween20!). I wonder whether it might be the source of non-specific tissue labelling by donkey anti-chicken IgY that was considered as an evidence of chicken origin.

    Significance

    General asssesment:

    The research has technical significance, and encourages greater use of this alternative animal model in cancer studies to help us understand the regulation of the cancer cell cycle. This, in turn, may improve the implementation of existing treatment methods or uncover potential vulnerabilities in the cancer cell cycle. No clinical significance so far. Certan groups conducting cancer biology studies and modelling might be interested in implementation of the described method, although it will be challenging without macroscopic evidence of the expected tumor nodules.The data and the methods presented in such a way that reproducing them might be challenging.The experiments are not adequately replicated and statistical analysis inadequate.

  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

    By using the CAM assay, the Authors of this studyshow that for the BxPC-3 and AsPC-1 cell lines, nucleoside labelling with 5-ethynyl-2'-deoxyuridine (EdU) can be multiplexed successfully with other cell-cycle markers such as cyclin B1 and Ki67, especially when combined with digital image analysis techniques. Starting from ED14, they observe the presence of a chicken embryonic cell type that appears to possess a high-quantity of extranuclear accumulation of EdU. Initial assessment of these cells showed that they are likely granulocytes which can be found in the embryonic liver of grafted and non-grafted embryos, as well as in xenograft sections. These cells do not express chicken MHC II, in turn making it less likely that they represent professional antigen presenting cells.

    Significance

    Remarks. This is a well-done study. However, I ask to the Authors to include also figures showing the macroscopic in ovo evidence and to discuss the role of the CAM assay in the study of xenografts.

  5. Version published to 10.1101/2025.09.09.675135 on bioRxiv