Author response:
The following is the authors’ response to the original reviews
Public Reviews:
Reviewer #1 (Public review):
Summary:
This manuscript assesses the utility of spatial image correlation spectroscopy (ICS) for measuring physiological responses to DNA damage. ICS is a long-established (~1993) method similar to fluorescence correlation spectroscopy, for deriving information about the fluorophore density that underlies the intensity distributions of images. The authors first provide a technical but fairly accessible background to the theory of ICS, then compare it with traditional spot-counting methods for its ability to analyze the characteristics of γH2AX staining. Based on the degree of aggregation (DA) value, the authors then survey other markers of DNA damage and uncover some novel findings, such as that RPA aggregation inversely tracks the sensitivity to PARP inhibitors of different cell lines.
The need for a more objective and standardized tool for analyzing DNA damage has long been felt in the field and the authors argue convincingly for this. The data in the manuscript are in general well-supported and of high quality, and show promise of being a robust alternative to traditional focus counting. However, there are a number of areas where I would suggest further controls and explanations to strengthen the authors' case for the robustness of their ICS method.
Strengths:
The spatial ICS method the authors describe and demonstrate is easy to perform and applicable to a wide variety of images. The DDR was well-chosen as an arena to showcase its utility due to its well-characterized dose-responsiveness and known variability between cell types. Their method should be readily useable by any cell biologist wanting to assess the degree of aggregation of fluorescent tags of interest.
Weaknesses:
The spatial ICS method, though of longstanding history, is not as intuitive or well-known as spot-based quantitation. While the Theory section gives a standard mathematical introduction, it is not as accessible as it could be. Additionally, the values of TNoP and DA shown in the Results are not discussed sufficiently with regard to their physical and physiological interpretation.
We agree that a major limitation in adaption of this approach is a deeper understanding of the theory and results. We have updated the theory section to include further discussion (Page 4 line 132)
The correlation of TNoP with γH2AX foci is high (Figure 2) and suggestive that the ICS method is suitable for measuring the strength of the DDR. The authors correctly mention that the number of spots found using traditional means can vary based on the parameters used for spot detection. They contrast this with their ICS detection method; however, the actual robustness of spatial ICS is not given equal consideration.
We found it difficult to give equal consideration of robustness to ICS. The major limitation of traditional approaches is proper selection of an intensity threshold that is necessary to define and separate foci from background intensity. However, ICS does not employ a threshold, therefore we could not test different thresholding applications in ICS as we did with traditional methods. In our view the absence of the need for a threshold is profoundly advantageous. The only inputs we employ in the ICS analysis are used to segment cell nuclei, yet these have no impact on the ICS calculation and are necessary for any analysis of the DDR.
Reviewer #2 (Public review):
Summary:
Immunostaining of chromatin-associated proteins and visualization of these factors through fluorescence microscopy is a powerful technique to study molecular processes such as DNA damage and repair, their timing, and their genetic dependencies. Nonetheless, it is well-established that this methodology (sometimes called "foci-ology") is subject to biases introduced during sample preparation, immunostaining, foci visualization, and scoring. This manuscript addresses several of the shortcomings associated with immunostaining by using image correlation spectroscopy (ICS) to quantify the recruitment of several DNA damage response-associated proteins following various types of DNA damage.
The study compares automated foci counting and fluorescence intensity to image correlation spectroscopy degree of aggregation study the recruitment of DNA repair proteins to chromatin following DNA damage. After validating image correlation spectroscopy as a reliable method to visualize the recruitment of γH2AX to chromatin following DNA damage in two separate cell lines, the study demonstrates that this new method can also be used to quantify RPA1 and Rad51 recruitment to chromatin following DNA damage. The study further shows that RPA1 signal as measured by this method correlates with cell sensitivity to Olaparib, a widely-used PARP inhibitor.
Strengths:
Multiple proof-of-concept experiments demonstrate that using image correlation spectroscopy degree of aggregation is typically more sensitive than foci counting or foci intensity as a measure of recruitment of a protein of interest to a site of DNA damage. The sensitivity of the SKOV3 and OVCA429 cell lines to MMS and the PARP inhibitors Olaparib and Veliparib as measured by cell viability in response to increasing amounts of each compound is a valuable correlate to the image correlation spectroscopy degree of aggregation measurements.
Weaknesses:
The subjectivity of foci counting has been well-recognized in the DNA repair field, and thus foci counts are usually interpreted relative to a set of technical and biological controls and across a meaningful time period. As such:
(1) A more detailed description of the numerous prior studies examining the immunostaining of proteins such as γH2AX, RAD51, and RPA is needed to give context to the findings presented herein.
We apologize for not providing enough detail. We have added further references and discussion. γH2AX foci counting, in particular, has been used in thousands of previous studies. (Pages 18 line 513 and 517)
(2) The benefits of adopting image correlation spectroscopy should be discussed in comparison to other methods, such as super-resolution microscopy, which may also offer enhanced sensitivity over traditional microscopy.
Thank you for raising this point. We have added this discussion (page 19 line 553). The limiting factor that ICS addresses is the partition coefficient of signal in a foci or cluster versus outside the cluster. Super-resolution will not necessarily improve this unless it is resolved down to single molecule counting. However, one would still need to evaluate how to define a cluster or foci in the background of non-cluster distribution.
(3) Additional controls demonstrating the specificity of their antibodies to detection of the proteins of interest should be added, or the appropriate citations validating these antibodies included.
We have added text stating that we only use validated antibodies (page 6 line 193). One thing to note is that we are measuring differences between treatment conditions, thus, if an antibody has non-specific labeling of proteins of cellular structures that do not change upon treatment, our approach would overcome this limitation.
Reviewer #3 (Public review):
Summary:
This paper described a new tool called "Image Correlation Spectroscopy; ICS) to detect clustering fluorescence signals such as foci in the nucleus (or any other cellular structures). The authors compared ICS DA (degree of aggregation) data with Imaris Spots data (and ImageJ Find Maxima data) and found a comparable result between the two analyses and that the ICS sometimes produced a better quantification than the Imaris. Moreover, the authors extended the application of ICS to detect cell-cycle stages by analyzing the DAPI image of cells. This is a useful tool without the subjective bias of researchers and provides novel quantitative values in cell biology.
Strengths:
The authors developed a new tool to detect and quantify the aggregates of immunofluorescent signals, which is a center of modern cell biology, such as the fields of DNA damage responses (DDR), including DNA repair. This new method could detect the "invisible" signal in cells without pre-extraction, which could prevent the effect of extracted materials on the pre-assembled ensembles, a target for the detection. This would be an alternative method for the quantification of fluorescent signals relative to conventional methods.
Recommendations for the authors:
Reviewer #1 (Recommendations for the authors):
Major comments:
(1) The ICS theory section is essential and based on an excellent review from one of the authors. It would benefit greatly from a diagram showing where the quantities 𝒈(𝟎, 𝟎), 𝝎𝟎, and 𝒈inf come from in the 2D Gaussian fit, ideally for two cases where these quantities differ (i.e., how they correspond to different DA or TNoP values). In my opinion, this addition would greatly increase the manuscript's accessibility for DDR researchers. The citation of the review at the beginning would also be a plus.
We have added the review citation at the front of the theory section (page 3 line 87).We have highlighted where g(0,0), the most critical measurement for determination of TNoP and DA, derives from in Figure 2D. However, it is difficult to describe all the curve fit parameters in an image as they have some interdependency on each other and thus labeling one in a single image would not independently capture how they might be observed in a different curve fit.
(2) The TNoP measured in Figure 2 is a quantity about 2000-3000 times greater than the number of "traditionally detected" foci by both methods and the linear relations have very low Y intercepts. Can the authors comment explicitly on the physical interpretation of this number - are 2 to 3 thousand independent particles present within each "focus" detected by traditional means? If so, then what might one "particle" correspond to? (a single secondary antibody or fluorophore? a nucleosome?). In a similar vein, the X intercepts lie at around 25 foci, meaning that in images with fewer than that number of foci detected by ImageJ or Imaris, the ICS method should detect zero TNoP - is this in line with the authors' predictions? Is it possible that a first-order line fit is not the most appropriate relation between the two methods?
We apologize for our brevity here. Since DA proved to be a more useful metric we did not spend much effort discussing TNoP. TNoP correlates to the number of clustered particles, or non-diffuse fluorophores. TNoP is the inverse of the number of individual particles per nucleus, but the value is not a direct measure of foci. If a sample had no clustering at all, the number of individual particles would be at a maximum and the TNoP would be at a minimum. However, as fluorophores cluster, the number of individual particles (i.e. non-clustered fluorophores) decreases, which increases the TNoP value. Therefore, TNoP has a correlation to the number of foci detected through traditional measurements, as we found here. Yet, TNoP is a relative measurement and cannot be compared across different conditions. Similar to foci counting, TNoP is unable to factor the size or intensity of each cluster, thus DA is a more appropriate quantification of the DNA damage response.
The value of TNoP is dependent on the fitted point spread function and the area of the nucleus. The y=0 intercept of TNoP is defined by the optical setup and is not expected to necessarily go through x=0. Intriguingly, other groups have found that some foci identified through traditional measurements are actually clusters of multiple smaller foci, thus the concept of what a foci represents is difficult to interpret. Thus, here we aimed to show a general correlation of TNoP with foci count through traditional methods to reflect how ICS is similar to foci counting, then employed DA to overcome the limitations of defining a foci.
We have tried to clarify this in the text (page 8, line 266)
(3) Some suggestions to address the robustness of ICS:
For a given sample (i.e. one segmented nucleus), the calculation of DA and TNoP should be similar between different images of that same nucleus taken at different times, similar to how the number of traditionally detected foci would be fairly invariant. In particular, it should be shown that these values are not just scaling with the higher normalized intensity seen in stronger DDR responses. In the same vein, the linear relationship between TNoP and "foci" should not change even if the confocal settings are slightly different (i.e., higher/lower illumination intensity) as long as the condition stipulated by the authors in the Discussion holds ("ICS can be implemented on any fluorescence image as long as the square relative fluorescence intensity fluctuations are detectable above noise fluctuations."). To show, as the title states, that spatial ICS is a robust tool, it would be desirable to demonstrate this with a series of images of the same cell at the same or varying excitation intensities.
Thank you for your suggestions. Indeed, the calculation will be the same over sequential images of the same cell. Observations of dose dependent DA that does not correlate with intensity for RPA1 and RAD51 results (Fig. S5) directly demonstrates that DA does not just scale with intensity.
We would not expect the TNoP to change with confocal setting, however we show in Figure 1 that the number of foci does indeed change with intensity settings as captured by thresholds. Therefore, any interpretation of TNoP vs. foci count would be very difficult to make at different microscope settings. To ensure we are fairly comparing ICS to existing analysis we keep the settings the same and measure changes between conditions.
(4) More information is needed on how intensity normalization was performed. The Methods states "Measurements across experiments were normalized by the control in each dataset." The DMSO (0mM drug) plots all appear to have a mean of 1.0, so it appears the values for each set of control nuclei were divided by their own mean, and then the values for each set of experimental nuclei were divided by the mean value of all 3 controls as an aggregate; is this correct?
We apologize for not being more clear. Thank you for raising this point. We normalized data to a control from each experimental group. Thus, in figures 3,4 and 5 data were collected over multiple experiments with one control per experiment and each treatment condition included in each experiment. Therefore, we normalized each result to the corresponding control from that imaging session. However, in Figure 8 we ran experiments at much higher throughput with multiple controls per experiment, thus the data were normalized to the overall average of the controls, which is why the control averages are not all at a value of 1. We have clarified this in the text. (Page 7 line 218).
(5) Some more information about the ICS analysis should be given if the full code is not provided - in particular, how the nucleus mask was implemented on the "signal" channel (were the edges abruptly set to zero or was a window function introduced to avoid edge effects in the discrete FFT?
Thank you for raising this point. We have added the code to GitHub - github.com/ dubachLab/ics. The signal region was established by simply applying the nuclear mask from the DAPI channel to the IF channel. Each region is padded with average intensity value at the edges for 2x the dimensions of the ROI to remove edge effects in the FFT.
Minor comments:
(1) Figure 3, 4, 5: I think it would aid figure readability if channels were labeled in the images themselves, not just in the legend.
Thank you for the suggestion, we tried doing this and struggle to fit a label with the layout of the images. We were also concerned about interpretation of data in each column and the potential to assign data to each figure if they were so prominently labeled.
(2) Supplemental Figures are mislabeled; the order given in the legends is S1, S2, S3, S2, S3. S4 is called out in the main text where it should be S5.
Thank you for catching this error. We have made the necessary corrections. S4 contains data on cellular response to the drugs, while S5 contains intensity data in response to MMS.
(3) It should be stated for each Figure what kind of microscopy was performed - I assume that it is confocal for everything except when widefield is explicitly stated, but for clarity please add this information.
Indeed, this is correct, we have indicated which microscopy was used for each figure.
(4) The MATLAB code and full (uncropped) Western blots should be provided as supplemental data if possible.
We have included a GitHub link for the code and un-cropped western blots.
(5) The p values from significance tests should indicate whether multiple comparisons correction was necessary (if suggested by Prism) and performed.
Apologies for a lack of clarity but this was not necessary, significance was calculated vs. the next lower dose (e.g. 10 micromolar vs. 1 micromolar). We have clarified this in the methods (page 7 line 221).
Reviewer #2 (Recommendations for the authors):
Major points:
In addition to the weaknesses noted above, to encourage widespread adoption of this method, the authors should make the tools that they used for their analysis publicly available. In a few instances (e.g., compare Figures 3J and 3L), other methods outperform DA. It would be meaningful to discuss when especially DA may be a better measure than others (such as intensity or number of foci).
We have made code available on Github. We expect results, such as those in Figures 3J and 3L where intensity is significantly higher at the highest concentration but DA is not are reflective of the underlying biology and this may be interpreted differently under different experimental conditions. Imaris spots (Fig. 3K) also does not capture a significant increase at the highest dose of olaparib, suggesting that intensity may raise but it doesn’t not generate more foci. These results are likely highly dependent on the mechanism of olaparib at such a high concentration and the DDR response. We are hesitant to draw biological conclusions from these results and instead would like to highlight the capacity of ICS to evaluate the DDR, therefore we don’t want to make any broad comments about different applications.
Minor points:
(1) Pg. 12: "We used MMS to induce DNA damage in SKOV3 and OVCA429 cells. As expected, normalized intensity for RPA1 and RAD51 values (Figure S5) did not display a dose dependence on MMS concentration."
Please provide a citation for the claim that RPA1 and RAD51 normalized intensities do not display a dose dependence on MMS concentration.
These were data that we generated. We were not expecting an intensity change as that would presumably require increased protein generation in response to MMS, compared to gH2AX where the phospho-specific H2AX is generated in the DDR.
(2) Pg. 12: "Similar to RPA1, RAD51 does not form distinguishable foci in the nuclei in cells without preextraction (Fig. 5)." Please provide a citation for this claim.
We did not do pre-extraction and our results don’t produce changes in distinguishable foci. We provided citations discussing how, without pre extraction, foci formation for these proteins is not obvious (REF 38 and 39).
(3) I noted that the authors cite one paper [38] apparently showing that RPA and Rad51 do not always form foci, however, this is in the C. elegans germline in response to micro irradiation, therefore I am not sure that it is applicable to human cells.
We apologize for referencing a paper on C elegans. Most papers looking at RPA and RAD51 in the DDR use pre-extraction as it seems necessary to observe foci. Therefore, there are not as many papers, that we could find, that do not use pre-extraction. Reference 39 is in Hela cells.
Reviewer #3 (Recommendations for the authors):
Major points:
(1) Page 8, the second paragraph: In the Result section, it is better to describe how the authors carried out immuno-staining (without pre-extract subtraction) and ICS briefly, although the method is described in detail in the Method section.
Thank you for the suggestion, we have added this description (page 8, line 259)
(2) In Figure 5K-P: The authors analyzed "invisible" RAD51 foci on the image (Fig. 5L, M, O, and P) without pre-extraction. As a control experiment, it is useful to check whether pre-extraction would provide "visible" RAD51 foci and to examine the similar MMS concentration dependency shown in Figure 5R (or 5T). This would strengthen the power of the ICS analysis.
Thank you for the suggestion. In our hands, pre-extraction is extremely subjective. We have tried performing pre-extraction but find highly variable results depending on conditions. Therefore, we did not include any pre-extraction here. We expect that performing these experiments may or may not agree with results in Figure 5 largely because we are unable to achieve repeatable pre-extraction foci counting.
(3) Figure 6D (and 6C) looks very interesting. It would be important to show the interpretation of this correlation shown in the graph. Although the authors argued that ICS analysis results shown in the graph could provide new insight into the DDR (page 14, last line 5), as shown in another part, it is important to carry out the same analysis by using Imaris Spots. Moreover, it is interesting to apply the analysis to RAD51 foci (shown in Figure 5), given that the PARPi effect is enhanced in the absence of RAD51mediated recombination.
We completely agree that this analysis may generate interesting results to help interpret the DDR response to PARP inhibition. These experiments are part of an ongoing follow up study where we extend the use of ICS to other parts of the DDR and investigate protein clustering across several proteins with impact on PARPi response. Therefore, since the focus of this manuscript is introducing ICS as a tool to study the DDR, we believe that omitting those data here does not deter from the central points of the manuscript. We including results in Figure 6 because we wanted to show how ICS could impact DDR research. Furthermore, combined with our advances shown in Figures 7 and 8, we are currently working on adapting ICS to be high-throughput and much simpler than Imaris spots for handling large datasets needed to generate results like those in Figure 6.
Minor points:
(1) Figure 1I, blue arrows: These showed an area with a higher background. Because of a low magnification, it is very hard to see the difference from the other areas of the background. It is better to show a magnified image of the representative region with a higher background.
We hope that readers can see the higher intensity in the diffuse area. We attempted to construct a zoomed in area, but that either blocked a significant portion of the nonzoomed image or added complexity to the figure. We have noted that images in Figure S1 are larger and more obviously capture an increase in background intensity.
(2) Figure 2 legend, line 5, the same as "A)": This should be "B".
Here, the number of independent particle clusters is intended to be the same as A, the difference is that the independent particles are clusters in C and individual fluorophores in A.
(3) Page 9, the first paragraph, last line, foci formation, and foci composition: These should be "focus formation and focus composition".
We have changed this.
(4) Page 15, the first paragraph, line 5, palbociclib, camptothecin, or etoposide: please explain what kinds of the drugs are.
We have added that these drugs cause cells to stall at different cell cycle stages. Explaining the drugs would take considerable room in the text.
(5) Page 16, the first paragraph, line 1, bleomycin: Please explain what this drug is.
Similar to above, we have stated that this drug causes DNA damage, going into detail would take several sentences.
