UFMTrack: Under-Flow Migration Tracker enabling analysis of the entire multi-step immune cell extravasation cascade across the blood-brain barrier in microfluidic devices
Curation statements for this article:-
Curated by eLife
eLife Assessment
This work is important because it elucidates how immune cells migrate across the blood brain barrier. In the revised version of this study, the authors present a convincing framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging using microfluidic devices. This work will be of interest to the cancer biology, immunology and medical therapeutics fields.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
The endothelial blood-brain barrier (BBB) strictly controls immune cell trafficking into the central nervous system (CNS). In neuroinflammatory diseases such as multiple sclerosis, this tight control is, however, disturbed, leading to immune cell infiltration into the CNS. The development of in vitro models of the BBB combined with microfluidic devices has advanced our understanding of the cellular and molecular mechanisms mediating the multi-step T-cell extravasation across the BBB. A major bottleneck of these in vitro studies is the absence of a robust and automated pipeline suitable for analyzing and quantifying the sequential interaction steps of different immune cell subsets with the BBB under physiological flow in vitro.Here we present the Under-Flow Migration Tracker ( UFM Track) framework and a pipeline built based on it to study the entire multi-step extravasation cascade of immune cells across brain microvascular endothelial cells under physiological flow in vitro. UFM Track achieves 90% track reconstruction efficiency and allows for scaling due to the reduction of the analysis cost and by eliminating experimenter bias. This allowed for an in-depth analysis of all behavioral regimes involved in the multi-step immune cell extravasation cascade. The study summarizes how UFM Track can be employed to delineate the interactions of CD4 + and CD8 + T cells with the BBB under physiological flow. We also demonstrate its applicability to the other BBB models, showcasing broader applicability of the developed framework to a range of immune cell-endothelial monolayer interaction studies. The UFM Track framework along with the generated datasets is publicly available in the corresponding repositories.
Article activity feed
-
-
-
eLife Assessment
This work is important because it elucidates how immune cells migrate across the blood brain barrier. In the revised version of this study, the authors present a convincing framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging using microfluidic devices. This work will be of interest to the cancer biology, immunology and medical therapeutics fields.
-
Reviewer #1 (Public review):
Summary:
It is evident that studying leukocyte extravasation in vitro is a challenge. One needs to include physiological flow, culture cells and isolate primary immune cells. Timing is of utmost importance and a reproducible setup essential. Extra challenges are met when extravasation kinetics in different vascular beds is required, e.g., across the blood-brain barrier. In this study, the authors describe a reliable and reproducible method to analyze leukocyte TEM under physiological flow conditions, including this analysis. That the software can also detect reverse TEM is a plus.
Strengths:
It is quite a challenge to get this assay reproducible and stable, in particular as there is flow included. Also for the analysis, there is currently no clear software analysis program, and many labs have their own …
Reviewer #1 (Public review):
Summary:
It is evident that studying leukocyte extravasation in vitro is a challenge. One needs to include physiological flow, culture cells and isolate primary immune cells. Timing is of utmost importance and a reproducible setup essential. Extra challenges are met when extravasation kinetics in different vascular beds is required, e.g., across the blood-brain barrier. In this study, the authors describe a reliable and reproducible method to analyze leukocyte TEM under physiological flow conditions, including this analysis. That the software can also detect reverse TEM is a plus.
Strengths:
It is quite a challenge to get this assay reproducible and stable, in particular as there is flow included. Also for the analysis, there is currently no clear software analysis program, and many labs have their own methods. This paper gives the opportunity to unify the data and results obtained with this assay under label-free conditions. This should eventually lead to more solid and reproducible results.
Also, the comparison between manual and software analysis is appreciated.
Weaknesses:
The authors stress that it can be done in BBB models, but I would argue that it is much more broadly applicable. This is not necessarily a weakness of the study but more an opportunity to strengthen the method. So I would encourage the authors to rewrite some parts and make it more broadly applicable.
-
Reviewer #2 (Public review):
Summary:
This paper develops an under-flow migration tracker to evaluate all the steps of the extravasation cascade of immune cells across the BBB. The algorithm is useful and has important applications.
Strengths:
The algorithm is almost as accurate as manual tracking and importantly saves time for researchers. The authors have discussed how their tool compares to other tracking methods.
Weaknesses:
Applicability can be questioned because the device used is 2D and physiological biology is in 3D. However, the authors have addressed this point in their revised manuscript.
-
Reviewer #3 (Public review):
Summary:
The authors aimed to establish a faster and more efficient method of tracking steps of T-cell extravasation across the blood brain barrier. The authors developed a framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging. The authors succinctly describe the basic requirements for tracking in the introduction followed by an in-depth account of the execution.
Weaknesses and Strengths:
Materials & methods and results
(1) The methods section also lacks details of the microfluidic device that the authors talk about in the paper. Under physiological sheer stress, the T-cells detach from the pMBMEC monolayer, and are hence unable to be detected; however, this …
Reviewer #3 (Public review):
Summary:
The authors aimed to establish a faster and more efficient method of tracking steps of T-cell extravasation across the blood brain barrier. The authors developed a framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging. The authors succinctly describe the basic requirements for tracking in the introduction followed by an in-depth account of the execution.
Weaknesses and Strengths:
Materials & methods and results
(1) The methods section also lacks details of the microfluidic device that the authors talk about in the paper. Under physiological sheer stress, the T-cells detach from the pMBMEC monolayer, and are hence unable to be detected; however, this observation requires an explanation pertaining to the reason of occurrence and potential solutions to circumvent it to ensure physiologically relevant experimental parameters.
(2) The author describes a method for debris exclusion using UFMTrack that eliminates objects of <30 pixels in size from analysis based on a mean pixel size of 400 for T lymphocytes. However, this mean pixel size appears to stem from in-vitro activated CD8 T cells, which rapidly grow and proliferate upon stimulation. In line with this, activated lymphocytes exhibit increased cytoplasmic area, making them appear less dense or "brighter" by phase microscopy compared to naïve lymphocytes, which are relatively compact and subsequently appear dimmer. Given this, it is not clear whether UFMTrack is sufficiently trained to identify naïve human lymphocytes in circulating blood, nor smaller, murine lymphocytes. Analysis of each lymphocyte subtype in terms of pixel size and intensity would be beneficial to strengthen the claim that UFMTrack can identify each of these populations. Additionally, demonstrating that UFMTrack can correctly characterize the behavior of naïve versus activated lymphocytes isolated from murine and human sources would strengthen the claim that UFMTrack can be broadly applied to study lymphocyte dynamics in diverse models without additional training
(3) Average precision was compared to the analysis of UFMTrack but it is unclear how average precision was calculated. This information should have been included in the methods section
(4) CD4 and CD8 T cells exhibit distinct biology and interaction kinetics driven in part by their MHC molecule affinity and distinct receptor expression profiles. Thus, it is unclear why two distinct mechanisms of endothelial cell activation are needed to see differences between the populations.
(5) The BMECs are barrier tissues but were cultured on µdishes in this study. To study the transmigration of T-cells across the endothelium, the model would have been more relevant on a semi-permeable membrane instead of a closed surface.
(6) Methods are provided for the isolation and expansion of human effector and memory CD4+ T cells. However, there is no mention of specific CD4+ T cell populations used for analysis with UFMTrack, nor a clear breakdown of tracking efficiency for each subpopulation. Further, there is no similar method for the isolation of CD8+ T cell compartments. A clear breakdown of the performance efficiency of UFMTrack with each cell population investigated in this study would provide greater insight into the software's performance with regard to tracking the behavior and movement of distinct immune populations.
(7) The results section is quite extensive and discusses details of establishment of the framework while highlighting both the pros and cons of the different aspects of the process, for example, the limitation of the two models, 2D and 2D+T were highlighted well. However, the results section includes details which may be more fitting in the methods section.
(8) A few statements in the results section lacked literary support, which was not provided in the discussion either, such as support for increased variance of T-cell instantaneous speed on stimulated vs non-stimulated pMBMECs. Another example is the enhancement of cytokine stimulation directed T-cell movement on the pMBMECs that the authors observed but failed to relay the physiological relevance of it. The authors don't provide enough references for developments in the field prior to their work which form the basis and need for this technology.
(9) The rationale for use of OT-1 and 2D2-derived murine lymphocytes is unclear here. The OT-1 model has been generated to study antigen-specific CD8+ T cell responses, while the 2D2 model has been generated to recapitulate CD4 T cell-specific myelin oligodendrocyte glycoprotein (MOG) responses.
Figures and text
(1) There are certain discrepancies and misarrangement of figures and text. For example, discussion of the effect of sheer flow on T cell attachment as part of the introduction in Figure 1 and then mentioning it in the text again in the results section as part of Figure 4 is repetitive.
(2) Section IV, subsection 1 of the results section, refers to 'data acquisition section above' in line 279, however the said section is part of materials and methods which is provided towards the end of the manuscript.
(3) There are figures in the manuscript that have not been referenced in the results section, for example, Figures 3A and B. Figure 1 hasn't been addressed until subsection 7 of materials and methods
(4) A lack of significance but an observed trend of increased variance of T cell instantaneous speed is reported in line 296-298; however, the graph (figure 4G) shows a significant change in instantaneous speed between non-stimulated and TNFα-stimulated systems. This is misleading to the readers.
(5) The authors talk about three beginner experimentors testing the manual T cell tracking process but figure 5 only showcases data from two experimentors without stating the reason for excluding experimentor 1.
Discussion
(1) While the discussion captures the major takeaways from the paper, it lacks relevant supporting references to relate the observation to physiological conditions and applicability.
(2) The discussion lacks connection to the results since the figures were not referenced while discussing an observed trend
(3) The authors briefly looked into mouse and human BMECs and their individual interaction with T-cells, but don't discuss the differences between the two, if any, that challenged their framework.
(4) Even though though the imaging tool relies on difference in appearance for detection, the authors talk about lack of feasibility in detecting transmigration of BMDMs due to their significantly different appearance. The statement lacks a problem solving approach to discuss how and why this was the case.
Relevance to the field:
Utilizing the framework provided by the authors, the application can be adapted and/or utilized for visualizing a range of different cell types, provided they are different in appearance. However, this would require extensive changes to the script and won't be adaptable in its current form.
-
Author response:
The following is the authors’ response to the original reviews.
Public Reviews:
Reviewer #1 (Public Review):
Summary:
It is evident that studying leukocyte extravasation in vitro is a challenge. One needs to include physiological flow, culture cells and isolate primary immune cells. Timing is of utmost Importance and a reproducible setup essential. Extra challenges are met when extravasation kinetics in different vascular beds is required, e.g., across the blood-brain barrier. In this study, the authors describe a reliable and reproducible method to analyze leukocyte TEM under physiological flow conditions, including this analysis. That the software can also detect reverse TEM is a plus.
Strengths:
It is quite a challenge to get this assay reproducible and stable, in particular as there is flow included. Also for the …
Author response:
The following is the authors’ response to the original reviews.
Public Reviews:
Reviewer #1 (Public Review):
Summary:
It is evident that studying leukocyte extravasation in vitro is a challenge. One needs to include physiological flow, culture cells and isolate primary immune cells. Timing is of utmost Importance and a reproducible setup essential. Extra challenges are met when extravasation kinetics in different vascular beds is required, e.g., across the blood-brain barrier. In this study, the authors describe a reliable and reproducible method to analyze leukocyte TEM under physiological flow conditions, including this analysis. That the software can also detect reverse TEM is a plus.
Strengths:
It is quite a challenge to get this assay reproducible and stable, in particular as there is flow included. Also for the analysis, there is currently no clear software analysis program, and many labs have their own methods. This paper gives the opportunity to unify the data and results obtained with this assay under label-free conditions. This should eventually lead to more solid and reproducible results.
Also, the comparison between manual and software analysis is appreciated.
We thank the Reviewer for their positive evaluation of our manuscript and highlighting the value of obtaining more reproducible and unbiases results, as well as detection of forward and reverse transmigration with UFMTrack.
Weaknesses:
The authors stress that it can be done in BBB models, but I would argue that it is much more broadly applicable. This is not necessarily a weakness of the study but more an opportunity to strengthen the method. So I would encourage the authors to rewrite some parts and make it more broadly applicable.
We thank the Reviewer for this suggestion. In the revised version of our manuscript, we have now emphasized the broader applicability of UFMTrack to analyze the interaction of immune cells with 2dimensional endothelial monolayers in various contexts in the abstract, introduction, and discussion sections.
Reviewer #2 (Public Review):
Summary:
This paper develops an under-flow migration tracker to evaluate all the steps of the extravasation cascade of immune cells across the BBB. The algorithm is useful and has important applications.
Strengths:
Algorithm is almost as accurate as manual tracking and importantly saves time for researchers.
We thank the Reviewer for this positive evaluation of our work.
Weaknesses:
Applicability can be questioned because the device used is 2D and physiological biology is in 3D. Comparisons to other automated tools was not performed by the authors.
We thank the Reviewer for pointing our attention to these weaknesses in our manuscript.
We have clarified in the revised manuscript that using 2D endothelial monolayer models in parallel laminar flow chambers is still a state-of-the-art methodology for studying the multi-step extravasation process of immune cells across endothelial monolayers under physiological flow by in vitro live cell imaging. These models provide excellent optical quality that is not yet achieved in 3D models. We have extended the introduction to emphasize the limitations of existing tools that motivated us to establish UFMTrack. We have furthermore extended the discussion section to highlight the features unique to our UFMTrack framework.
Reviewer #3 (Public Review):
Summary:
The authors aimed to establish a faster and more efficient method of tracking steps of T-cell extravasation across the blood brain barrier. The authors developed a framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging. The authors succinctly describe the basic requirements for tracking in the introduction followed by an in-depth account of the execution.
We thank the Reviewer for their positive evaluation of our manuscript and highlighting the value of label-free analysis of the multistep immune cell extravasation cascade with UFMTrack.
Weaknesses and Strengths:
Materials & methods and results:
(1) The methods section also lacks details of the microfluidic device that the authors talk about in the paper. Under physiological sheer stress, the T-cells detach from the pMBMEC monolayer, and are hence unable to be detected; however, this observation requires an explanation pertaining to the reason of occurrence and potential solutions to circumvent it to ensure physiologically relevant experimental parameters.
We thank the Reviewer for pointing out this oversight. We have used a custom-made microfluidic device that has been published and described in detail before. This information has now been included in the Methods Section under Point 7, and the two references describing the flow chamber in depth are mentioned below and have been included in the manuscript.
Coisne Caroline, Ruth Lyck and Britta Engelhardt. 2013. Live cell imaging techniques to study T cell trafficking across the blood-brain barrier in vitro and in vivo. Fluids and Barriers of the CNS 10:7 doi:10.1186/20458118-10-7; 21 January 2013
Lyck R, Hideaki Nishihara, Sidar Aydin, Sasha Soldati and Britta Engelhardt. 2022. Modeling brain vasculature immune interactions in vitro. Angogenesis, 2nd edition. Editors PatriciaD’Amore and Diane Bielenberg Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a041185
T cell detachment is a physiologically relevant parameter besides T cell arrest, polarization, crawling, probing, and transmigration during the interaction with an endothelial monolayer. T cell detachment means that post-arrest, the T cell cannot engage adhesion molecules required for subsequent polarization and, eventually, transmigration.
(2) The author describes a method for debris exclusion using UFMTrack that eliminates objects of <30 pixels in size from analysis based on a mean pixel size of 400 for T lymphocytes. However, this mean pixel size appears to stem from in-vitro activated CD8 T cells, which rapidly grow and proliferate upon stimulation. In line with this, activated lymphocytes exhibit increased cytoplasmic area, making them appear less dense or “brighter” by phase microscopy compared to naïve lymphocytes, which are relatively compact and subsequently appear dimmer. Given this, it is not clear whether UFMTrack is sufficiently trained to identify naïve human lymphocytes in circulating blood, nor smaller, murine lymphocytes. Analysis of each lymphocyte subtype in terms of pixel size and intensity would be beneficial to strengthen the claim that UFMTrack can identify each of these populations. Additionally, demonstrating that UFMTrack can correctly characterize the behavior of naïve versus activated lymphocytes isolated from murine and human sources would strengthen the claim that UFMTrack can be broadly applied to study lymphocyte dynamics in diverse models without additional training
We thank the Reviewer for the suggestion to more precisely evaluate the range of cell sizes that can be analyzed by our framework. We have included a visualization of crawling cell sizes successfully analyzed by the UFMTrack in Supplementary Figure 7. It demonstrates that the human peripheral blood mononuclear cells, that are almost twice as small as the activated mouse CD4 T cells used in these assays, can be successfully segmented, tracked, and analyzed with the UFMTrack framework. Thus, our UFMTrack framework is suitable for a broad application to differentially sized immune cells during their interaction with the endothelial cell monolayer under flow.
(3) Average precision was compared to the analysis of UFMTrack but it is unclear how average precision was calculated. This information should have been included in the methods section
We thank the Reviewer for pointing our attention to the missing information. We have added a subsection, “Performance Analysis”, to the Materials and Methods section, where we describe the statistical methods and the performance metrics used to evaluate the UFMTrack framework.
(4) CD4 and CD8 T cells exhibit distinct biology and interaction kinetics driven in part by their MHC molecule affinity and distinct receptor expression profiles. Thus, it is unclear why two distinct mechanisms of endothelial cell activation are needed to see differences between the populations.
We thank the Reviewer for pointing out that different cytokine stimulations of endothelial cells were used in the assays used here to test our UFMTrack to analyze CD4 and CD8 T cell interactions with the endothelial monolayer. While the Reviewer is correct that CD4 and CD8 T cells use different mechanism to cross the pMBMEC monolayer as show by us (doi: 10.1002/eji.201546251.) and others and that recognition of cognate antigen on MHC class I on pMBMECs will arrest CD8 T cells and lead to CD8 T-cell mediated apoptosis ( doi: 10.1038/s41467-023-38703-2.) the focus of the present study was not on comparing CD4 and CD8 T cell interactions with the pMBMEC monolayer but rather to test suitability of UFMTrack to study the different multi-step transmigration of these T cell subsets across the endothelial monolayer.
(5) The BMECs are barrier tissues but were cultured on µdishes in this study. To study the transmigration of T-cells across the endothelium, the model would have been more relevant on a semi-permeable membrane instead of a closed surface.
We understand the critique of the Reviewer, but laminar flow chambers with endothelial monolayers still provide a state-of-the-art and established methodology to study immune cell migration across endothelial monolayers by in vitro live cell imaging including endothelial cells forming the blood-brain barrier.
(6) Methods are provided for the isolation and expansion of human effector and memory CD4+ T cells. However, there is no mention of specific CD4+ T cell populations used for analysis with UFMTrack, nor a clear breakdown of tracking efficiency for each subpopulation. Further, there is no similar method for the isolation of CD8+ T cell compartments. A clear breakdown of the performance efficiency of UFMTrack with each cell population investigated in this study would provide greater insight into the software’s performance with regard to tracking the behavior and movement of distinct immune populations.
We thank the Reviewer for this comment. Since a fair performance evaluation requires collecting reliable and consistent manual annotations, in this work we have performed such analysis only for the mouse CD8 T-cell population migrating on the pMBMEC monolayer. We have chosen this as a reference since it is a different cell population than the one the segmentation model was trained on. This provides an insight into how high performance is expected when other immune cell types are studied than the ones used for model development.
(7) The results section is quite extensive and discusses details of establishment of the framework while highlighting both the pros and cons of the different aspects of the process, for example the limitation of the two models, 2D and 2D+T were highlighted well. However, the results section includes details which may be more fitting in the methods section.
We thank the Reviewer for highlighting the extensive work carried out in the development of our UFMTrack framework. We decided to include in the results section only the description of key elements and design decisions taken when developing the framework, such as the need to include a time series of images for successful segmentation of the transmigrated cells. At the same time, the majority of implementational details can be found in the Supplementary Material.
(8) A few statements in the results section lacked literary support, which was not provided in the discussion either, such as support for increased variance of T-cell instantaneous speed on stimulated vs non-stimulated pMBMECs. Another example is the enhancement of cytokine stimulation directed T-cell movement on the pMBMECs that the authors observed but failed to relay the physiological relevance of it. The authors don’t provide enough references for developments in the field prior to their work which form the basis and need for this technology.
We thank the Reviewer for this comment and for asking for literature references. However, we cannot provide such references as these are original observations we made by employing the UFMTrack framework. This shows that UFMTrack observes T-cell behaviors that have previously been overlooked. Their physiological relevance will have to be explored in separate studies. We have extended the introduction section to include the details on the existing methods developed in the field, as well as their weaknesses that motivated the development of the UFMTrack framework.
(9) The rationale for use of OT-1 and 2D2-derived murine lymphocytes is unclear here. The OT-1 model has been generated to study antigen-specific CD8+ T cell responses, while the 2D2 model has been generated to recapitulate CD4 T cell-specific myelin oligodendrocyte glycoprotein (MOG) responses.
To establish and test the UFMTrack framework, we have made use of the specific T-cell subsets and endothelial cell models we generally use within our research context. Especially for animal work, this is according to the 3R rules requesting to reduce animal experimentation.
Figures and text:
(1) There are certain discrepancies and misarrangement of figures and text. For example, discussion of the effect of sheer flow on T cell attachment as part of the introduction in figure 1 and then mentioning it in the text again in the results section as part of figure 4 is repetitive.
We thank the Reviewer for pointing our attention to this misarrangement. We have adjusted the label of Figure 4 to emphasize that this effect is correctly captured by the UFMTrack.
(2) Section IV, subsection 1 of the results section, refers to ‘data acquisition section above’ in line 279, however the said section is part of materials and methods which is provided towards the end of the manuscript.
We thank the Reviewer for pointing our attention to this misarrangement. We have adjusted the text to reflect the correct chapter order.
(3) There are figures in the manuscript that have not been referenced in the results section, for example, figure 3A and B. Figure 1 hasn’t been addressed until subsection 7 of materials and methods
We thank the Reviewer for pointing our attention to this misarrangement. We have adjusted the text to refer to all figure panels and the clarification of the cell multiplicity estimation in the supplementary information section. References to Figure 1 were added in the introduction section to illustrate the in vitro under flow imaging setup as well as the typical T cell behaviors in such experiments.
(4) A lack of significance but an observed trend of increased variance of T cell instantaneous speed is reported in line 296-298; however, the graph (figure 4G) shows a significant change in instantaneous speed between non-stimulated and TNFα-stimulated systems. This is misleading to the readers.
We thank the Reviewer for pointing our attention to this discrepancy. We have expanded the text to indicate a low statistical significance for the TNF and no significance but just a trend for the IL1-beta conditions.
(5) The authors talk about three beginner experimentors testing the manual T cell tracking process but figure 5 only showcases data from two experimentors without stating the reason for excluding experimentor 1.
We thank the Reviewer for pointing our attention to this ambiguity. While both the migration analysis and the manual cell tracking were performed by all three beginner experimenters, the cell tracking data for the first one was unfortunately lost due to a hardware failure.
Discussion:
(1) While the discussion captures the major takeaways from the paper, it lacks relevant supporting references to relate the observation to physiological conditions and applicability.
This study is not about the physiological relevance of the microfluidic devices and immune cells used but rather about advancing methodology to analyze dynamic immune cell behavior on endothelial monolayers under physiological flow. Therefore, the discussion does not extend to comparing the physiological relevance of the specific in vitro models employed in this study.
(2) The discussion lacks connection to the results since the figures were not referenced while discussing an observed trend
We thank the Reviewer for pointing our attention to this misarrangement. We have included the references to the relevant figures as well as supporting references.
(3) The authors briefly looked into mouse and human BMECs and their individual interaction with Tcells, but don’t discuss the differences between the two, if any, that challenged their framework.
We thank the Reviewer for pointing our attention to this weakness. We have added to the discussion section clarifications on the challenges of analyzing the T cell interactions with the HBMEC and the BMDM interactions with the pMBMEC monolayer.
(4) Even though though the imaging tool relies on difference in appearance for detection, the authors talk about lack of feasibility in detecting transmigration of BMDMs due to their significantly different appearance. The statement lacks a problem solving approach to discuss how and why this was the case.
We thank the Reviewer for pointing our attention to this weakness and apologize for the misleading explanation of the problem of analyzing the BMDM sample. Since the transmigrated part of the macrophages differs in appearance from a transmigrated part of a T cell, its detection by a Deep Neural Network trained on the T cell data is worse than that for the T cells. At the same time, the detection performance before the transmigration is sufficient for the BMDM migration analysis. The potential approaches to alleviate this are added to the discussion section.
Relevance to the field:
Utilizing the framework provided by the authors, the application can be adapted and/or utilized for visualizing a range of different cell types, provided they are different in appearance. However, this would require extensive changes to the script and won’t be adaptable in its current form.
Recommendations for the authors:
Reviewer #1 (Recommendations For The Authors):
The authors should announce in the abstract that the software analysis Track is downloadable and free to use for all researchers. They may consider providing some sort of helpdesk, although I realize that that may run into too much time.
As said above, they stress that it can be done in BBB models, but I would argue that it is much more broadly applicable.
We thank the Reviewer for these suggestions. We have emphasized the broader applicability of UFMTrack in the abstract and pointed out the public availability of the code and data.
Can they add an experiment that shows that it also works for neutrophils for example? I understand that on paper yes it should work, but the neutrophils are of course faster etc.
This is an excellent suggestion, but we tested UFMTrack within the current framework of ongoing research, which does not include the investigation of neutrophil transmigration across endothelial monolayers.
Also, the combination of different leukocytes in one TEM assay would really be a step forward. If the software can detect different-sized leukocytes, then this should be possible.
We thank the Reviewer for this suggestion. We have added Supplementary Figure 7, demonstrating the range of cell sizes that were successfully analyzed by the UFMTrack framework throughout our manuscript. We also added a statement to the discussion that according to this data, “simply by discriminating cells by size, it is possible to extend UFMTrack to study the interaction of several types of immune cells migrating on top of a cellular monolayer under flow.”
Extra challenges: can the method also discriminate between paracellular and transcellular migration modes? In particular for T-cells this is known to happen.
We thank the Reviewer for this suggestion. We have added this to the potential applications of UFMTrack in the discussion section. While this differentiation is not feasible relying solely on the phasecontrast imaging data, UFMTrack can simplify this analysis by providing automatically the predictions of the transmigration locations, for analysis of the fluorescent data of the junctional labels.
Reviewer #2 (Recommendations For The Authors):
This paper develops an under-flow migration tracker to evaluate all the steps of the extravasation cascade of immune cells across the BBB. The algorithm is useful and has important applications. There are several points that need to be addressed, particularly about the claims made by the authors.
Please see the comments below for more details:
• Lines 88-92: Add a citation for the characteristics of the BBB as a barrier
We have added two references accordingly.
• Lines 94-95: Can the authors indicate what models were used for these studies and how those compare to their in vitro model? In addition, can the authors say whether T cells were manually tracked in this study to translate results to the clinic and whether the results were successful when translated to the clinic? This may enhance the argument that automatic trackers are needed if the translation was not 100% successful
This introductory paragraph summarizes in vivo and in vitro observations from several laboratories. Although these studies include manual tracking of T cells, they do not necessarily distinguish all sequential steps of the multi-step T cell transmigration cascade. Thus, automated tracking may provide additional insights, allowing for increased translation of findings to the clinic.
• Lines 96-98: Citing the work of Roger Kamm and Noo Li Jeon would be helpful here as they pioneered these BBB microfluidic models and have protocol papers on how to build them and how to use them for cancer cell extravasation studies. Roger Kamm has also worked on several extravasation studies with neutrophils, monocytes, and PBMCs from 3D vasculatures in microfluidic devices, under flow using pressurized fluid or recirculating pumps. Mentioning those would be helpful as they are directly related to what the authors are presenting in their paper.
We thank the Reviewer for this comment, and we consider the work of Roger Kamm and Noo Li Jeon as very valuable for the field. However, these authors have focused on developing functional 3D microfluidic devices, including, e.g., all cells of the neurovascular unit which is not the focus of this present study that solely employed parallel flow chamber devices and endothelial monolayers.
• Lines 110-116: Can the authors comment on the use of ImageJ or similar automatic tracking tools and how these compare to the under-flow migration tracker developed in this paper? Several groups use ImageJ to track cellular migration successfully and in an automatic manner with short intervals between each frame. One paper that comes to mind is Chen et al: DOI: 10.1073/pnas.1715932115 where neutrophil migration in 3D was assessed with ImageJ in microfluidic devices of the vasculature. If the authors can highlight differences between their tool and what is currently available and used for automatic tracking (e.g. ImageJ), this would help in understanding the advantages of the migration tracker developed in this paper.
• Lines 118-121: Add citations for the current state of the art for T cell extravasation tracking
We thank the Reviewer for these suggestions. We have extended the introduction to add more details on the available tools for tracking migrating immune cells and their limitations, as well as the discussion section to emphasize the features unique to the developed UFMTrack framework.
• Figure 1: The device used by the authors is considered to be a 2D microfluidic device with a monolayer of mouse brain endothelial cells. I would recommend the authors to carefully revise the claims made in the paper to mention that this is a 2D device as opposed to a 3D device, in order to not mislead readers who may be expecting these analyses to be performed in 3D vasculatures.
We thank the Reviewer for this suggestion. We have included in the summary the mention of the 2dimensional nature of the employed BBB model.
• Figure 1: The T cells used in this study are not fluorescently-labeled but the authors mention that this is an issue from current state-of-the-art tools. I would recommend that the authors remove this point as being an issue because it is not addressed in their paper. The T cells are also not labeled in this study so this limitation of other systems is not addressed in this paper.
We apologize to the Reviewer as we do not understand this question. There will be many experimental conditions not allowing to study fluorescently tagged T cells. Therefore, UFMTrack is tailored to follow and analyze T cells and other immune cells during their interaction with endothelial monolayers independent of a fluorescence tag.
• Figure 1: Was the shear stress controlled manually with a syringe? Or with the use of a pressure controller? I would clarify this aspect and discuss human errors that can be introduced from manually controlling the pressure applied to the monolayer.
We thank the Reviewer for pointing our attention to this ambiguity. We have added a mention of the automated syringe pump used to control the shear stress in the text where the values of shear stress applied to the sample are first mentioned.
• Figure 1: Does T cell attachment occur within the first 5 minutes? Can the authors comment on how they chose this timeline and the percentage of T cells that are washed off at the second step at 1.5 dynes/cm^2? Is 30 seconds enough to ensure all the non-adhered T cells are washed off with 1.5 dyns/cm^2?
Superfusion of the T cells over the endothelial monolayer is performed under 0.5 dynes/cm2 to allow the T cells to settle on the endothelial cell monolayer under flow. After increasing to physiological, flow non adherent T cells detach within 30 seconds, as described by the Reviewer. We have included in the Methods Section Point 7 the references describing in depth the design of the flow chamber device and methods used here.
• Line 154: How many images were used in the training vs. testing dataset for T cell migrations?
We thank the Reviewer for pointing our attention to this missing information. We have added the sizes of the training and validation datasets. Specifically, the 226MPix of available imaging data was split into 154Mpix training and 37 MPix validation sets. The gap in between was introduced to avoid a correlation between validation and training set that would compromise the performance evaluation.
• Are the supplementary videos at real speed or accelerated?
We thank the Reviewer for pointing our attention to this missing information. The videos are sped up by a factor of 96. We have added this information to the Supplementary video descriptions.
• Lines 208 216: Can the authors comment on how their initial adhesion timeframe of 30sec before starting the recording at 5.5min affects the number of T cells with rapid displacement? 30 seconds may not be enough to ensure T cells have adhered to the endothelium
Please see our comment above. The methodology used in the present assays has been set up and validated in numerous publications. We have included in the Methods Section under Point 7 the references describing in depth the design of the flow chamber device and the methods used here.
• Lines 275-277: Was the number of testing images 18? Can the authors comment on how this compares to training dataset size and whether these numbers are enough to achieve robust results?
We apologize for this ambiguity in our manuscript. The framework was evaluated on 18 imaging datasets, each corresponding to 32 minutes of recording, not 18 images. We have added this clarification to the “CD4+ T cell analysis” subsection. The total size of these datasets is 18 datasets * 191 timeframe/dataset * 9.9MPix/frame = 34MPix
• Figure 4B: Can the authors add statistics here? Individual datapoints on the error bars would be helpful too.
We thank the Reviewer for pointing our attention to this weakness. The data corresponds to the statistical errors as evaluated based on all cells in the 18 datasets. We have added the total number of cells in each of the endothelium stimulation conditions to the text.
• Figure 4C-J: Can the authors put individual datapoints here as well and explain whether they considered each T cell to be one datapoint or each endothelium (averaging all T cells) to be one datapoint?
We thank the Reviewer for this suggestion. However, adding about one thousand points corresponding to each cell would be impractical. We thus present the distributions of the evaluated from the data metrics as a histogram on the violin plot instead of the swarm plot.
• Figure 4: Did the authors wash the monolayers before introducing T cells? Soluble unbound cytokines may still be present and there are two different questions that would be studied here: “Is the inflamed endothelium affecting T cell migration?” (if washing was performed) or “Is T cell and microenvironmental inflammation affecting T cell migration?” (if no washing was performed)
The endothelial monolayers are “washed” by starting the flow in the flow chamber device and this is before superfusing the T cells over the endothelial monolayer. We agree that our flow chamber device combined with UFMTrack will allow to address all these questions.
• Figure 4I: Are all the T cells decelerating? (negative AM speed)
We thank the Reviewer for this question. The cells are moving along the flow, which, in our experiments, is from left to right. The vector of speed is thus pointing against the x-axis, and thus the AM speed is negative.
• Lines 302 306: Please explain how this compares to ImageJ or similar trackers that can achieve similar outputs.
We thank the Reviewer for this question. We have added a statement in the “T-cell tracking” section emphasizing that standard trackers are incapable of correctly capturing large displacements.
• Lines 306-309: It is not lower for TNF stimulation though. How do the authors address this? TNF is also a pro-inflammatory cytokine.
We have previously shown that stimulation of pMBMECs with IL-1 and TNF-a induces different cell surface levels of ICAM-1 and VCAM-1, which will influence T cell behavior on the pMBMEC monolayer.
• Lines 313-315: Could this be because the monolayer was not washed and soluble cytokines affected T cell response directly?
Please see our answer to lines 306-309.
• Lines 319: Please cite Roger Kamm and Noo Li Jeon’s papers on BBB models with human BMECs, pericytes and astrocytes in 3D microfluidic devices.
We thank the Reviewer again for pointing out these studies. As mentioned above, as our present study does not explore 3D models of the BBB, we think it does not fit into the framework of our study to elaborate on 3D models of the BBB. In addition, this would require the inclusion of a discussion of the work of others like, e.g., Peter Searson and others.
• Figure 5: Several statistics are missing from parts of the figure. Please add those.
We apologize – but we do not understand which statistical analysis the Reviewer is missing from this Figure.
• Can the authors comment on the number of T cells perfused over the monolayer and if this ratio of T cells to endothelial cells makes physiological sense? Too many T cells may result in endothelium inflammation and increased diapedesis.
The number of T cells used to suprerfuse over the endothelial monolayer is tested to avoid aggregation of T cells in suspension and thus artificial interactions with the endothelial monolayer. T cell behavior on the pMBMEC monolayer remains the same over the dilution of factor 10.
• Lines 381 383: How does this compare to analyses that look at the cross-section of the endothelium? It is difficult to assess transmigration looking at the top view of the endothelium. Perhaps, cross-section assessments will identify differences in manual vs. automatic tracking.
There is, to the best of our knowledge, no microscopic device that would allow for in vitro live cell imaging of a live endothelial monolayer – this is in the presence of tissue culture medium – from the side at a resolution that would allow to define transmigration. Our current study rather shows the UFMTrack can distinguish cells moving above or below the endothelial monolayer.
• Figure 5J: This is probably the most important argument of the paper. If the authors can show statistical differences in their graph, this would greatly help convince readers that this tool is necessary and actually computationally efficient compared to manual work by researchers.
We thank the Reviewer for this suggestion. However, comparing a single data point for automated measurement with four manual experimenter analysts is not a statistically sound comparison. We believe that Figure 5K is clearly showing the factor 5 difference in analysis speed as compared to manual analysis. More importantly, though, the automated analysis is taking the machine time, lifting the need for the experimenter to invest even 1/5th of the original analysis time.
• Figure 6: Did the authors use autologous immune cells and endothelial cells? This is particularly relevant with the use of human-derived T cells (line 436) on the BMEC monolayer. Can the authors comment on non-self reactivity by the T cells encountering BMEC from another human subject?
Autologous T cell interaction with BMECs would only be possible when using hiPSC-derived EECM-BMECs and the T cells from the same individual. All other experimental frameworks will not include autologous interactions. This is the experimental framework used by most authors studying immune cell interactions with commercially available donors. We have not studied alloreactive interactions in our assays and thus cannot further comment.
• Figure 6M,N,O: How does this compare to ImageJ for tracking of fluorescent cells? I recommend the authors to try that, at least for this section, as this may enhance their argument for their tool vs. standard tools like ImageJ if success rates are higher for their tool.
We thank the Reviewer for this suggestion. We included a note on the analysis of the fluorescent datasets using the TrackMate plugin for imageJ performed previously in our lab in the “Human T cells on immobilized recombinant BBB adhesion molecules” subsection.
• Figure 6: Please put individual datapoints on the bar or violin plots where they are missing.
We thank the Reviewer for this suggestion. However, adding about one thousand points corresponding to each cell would be impractical. We thus present the distributions of the evaluated from the data metrics as a histogram on the violin plot instead of the swarm plot.
• Lines 467-471: This argument is important and should be mentioned earlier in the introduction.
Another point that can be mentioned is the application of this platform to imaging modalities in vivo (mouse or human) given that there is no fluorescent staining in these cases. This review may be relevant: https://doi.org/10.1002/jcb.10454
We thank the Reviewer for this suggestion. We have clarified in the introduction that UFMTrack does not require fluorescent labels of the imaged migrating cells and relies solely on the phase contrast imaging data.
• Discussion: Please address a few more potential applications to this study. One can be cancer and immune infiltration.
We thank the Reviewer for this suggestion. We have elaborated on additional potential applications to the discussion section.
Reviewer #3 (Recommendations For The Authors):
(1) Line 327-328: The authors talk about ‘As we have previously shown…pMBMEC monolayers differs between CD4+ and CD8+ cells…’. Where was this shown? If it was in a previously published article, please provide a reference.
We have added these missing references.
(2) Line 353: Please provide clear location on where to find the associated information instead of stating ‘see below’.
We thank the Reviewer for pointing our attention to this ambiguity. We have corrected the phrase to “see next paragraph”
(3) Line 439: Please correct the acronym to BMECs
We thank the Reviewer for pointing our attention to this typo. We have corrected it.
-
-
-
eLife assessment
This work is important because it attempts to elucidate how immune cells migrate across the blood brain barrier. The authors developed a convincing framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging using microfluidic devices. The data gathered are solid, and this work will be of interest to the cancer biology, immunology and medical therapeutics fields.
-
Reviewer #1 (Public Review):
Summary:
It is evident that studying leukocyte extravasation in vitro is a challenge. One needs to include physiological flow, culture cells and isolate primary immune cells. Timing is of utmost importance and a reproducible setup essential. Extra challenges are met when extravasation kinetics in different vascular beds is required, e.g., across the blood-brain barrier. In this study, the authors describe a reliable and reproducible method to analyze leukocyte TEM under physiological flow conditions, including this analysis. That the software can also detect reverse TEM is a plus.
Strengths:
It is quite a challenge to get this assay reproducible and stable, in particular as there is flow included. Also for the analysis, there is currently no clear software analysis program, and many labs have their own …
Reviewer #1 (Public Review):
Summary:
It is evident that studying leukocyte extravasation in vitro is a challenge. One needs to include physiological flow, culture cells and isolate primary immune cells. Timing is of utmost importance and a reproducible setup essential. Extra challenges are met when extravasation kinetics in different vascular beds is required, e.g., across the blood-brain barrier. In this study, the authors describe a reliable and reproducible method to analyze leukocyte TEM under physiological flow conditions, including this analysis. That the software can also detect reverse TEM is a plus.
Strengths:
It is quite a challenge to get this assay reproducible and stable, in particular as there is flow included. Also for the analysis, there is currently no clear software analysis program, and many labs have their own methods. This paper gives the opportunity to unify the data and results obtained with this assay under label-free conditions. This should eventually lead to more solid and reproducible results.
Also, the comparison between manual and software analysis is appreciated.
Weaknesses:
The authors stress that it can be done in BBB models, but I would argue that it is much more broadly applicable. This is not necessarily a weakness of the study but more an opportunity to strengthen the method. So I would encourage the authors to rewrite some parts and make it more broadly applicable.
-
Reviewer #2 (Public Review):
Summary:
This paper develops an under-flow migration tracker to evaluate all the steps of the extravasation cascade of immune cells across the BBB. The algorithm is useful and has important applications.
Strengths:
Algorithm is almost as accurate as manual tracking and importantly saves time for researchers.
Weaknesses:
Applicability can be questioned because the device used is 2D and physiological biology is in 3D. Comparisons to other automated tools was not performed by the authors.
-
Reviewer #3 (Public Review):
Summary:
The authors aimed to establish a faster and more efficient method of tracking steps of T-cell extravasation across the blood brain barrier. The authors developed a framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging. The authors succinctly describe the basic requirements for tracking in the introduction followed by an in-depth account of the execution.
Weaknesses and Strengths:
Materials & methods and results:
(1) The methods section also lacks details of the microfluidic device that the authors talk about in the paper. Under physiological sheer stress, the T-cells detach from the pMBMEC monolayer, and are hence unable to be detected; however, this …
Reviewer #3 (Public Review):
Summary:
The authors aimed to establish a faster and more efficient method of tracking steps of T-cell extravasation across the blood brain barrier. The authors developed a framework to visualize, recognize and track the movement of different immune cells across primary human and mouse brain microvascular endothelial cells without the need for fluorescence-based imaging. The authors succinctly describe the basic requirements for tracking in the introduction followed by an in-depth account of the execution.
Weaknesses and Strengths:
Materials & methods and results:
(1) The methods section also lacks details of the microfluidic device that the authors talk about in the paper. Under physiological sheer stress, the T-cells detach from the pMBMEC monolayer, and are hence unable to be detected; however, this observation requires an explanation pertaining to the reason of occurrence and potential solutions to circumvent it to ensure physiologically relevant experimental parameters.
(2) The author describes a method for debris exclusion using UFMTrack that eliminates objects of <30 pixels in size from analysis based on a mean pixel size of 400 for T lymphocytes. However, this mean pixel size appears to stem from in-vitro activated CD8 T cells, which rapidly grow and proliferate upon stimulation. In line with this, activated lymphocytes exhibit increased cytoplasmic area, making them appear less dense or "brighter" by phase microscopy compared to naïve lymphocytes, which are relatively compact and subsequently appear dimmer. Given this, it is not clear whether UFMTrack is sufficiently trained to identify naïve human lymphocytes in circulating blood, nor smaller, murine lymphocytes. Analysis of each lymphocyte subtype in terms of pixel size and intensity would be beneficial to strengthen the claim that UFMTrack can identify each of these populations. Additionally, demonstrating that UFMTrack can correctly characterize the behavior of naïve versus activated lymphocytes isolated from murine and human sources would strengthen the claim that UFMTrack can be broadly applied to study lymphocyte dynamics in diverse models without additional training
(3) Average precision was compared to the analysis of UFMTrack but it is unclear how average precision was calculated. This information should have been included in the methods section
(4) CD4 and CD8 T cells exhibit distinct biology and interaction kinetics driven in part by their MHC molecule affinity and distinct receptor expression profiles. Thus, it is unclear why two distinct mechanisms of endothelial cell activation are needed to see differences between the populations.
(5) The BMECs are barrier tissues but were cultured on µdishes in this study. To study the transmigration of T-cells across the endothelium, the model would have been more relevant on a semi-permeable membrane instead of a closed surface.
(6) Methods are provided for the isolation and expansion of human effector and memory CD4+ T cells. However, there is no mention of specific CD4+ T cell populations used for analysis with UFMTrack, nor a clear breakdown of tracking efficiency for each subpopulation. Further, there is no similar method for the isolation of CD8+ T cell compartments. A clear breakdown of the performance efficiency of UFMTrack with each cell population investigated in this study would provide greater insight into the software's performance with regard to tracking the behavior and movement of distinct immune populations.
(7) The results section is quite extensive and discusses details of establishment of the framework while highlighting both the pros and cons of the different aspects of the process, for example the limitation of the two models, 2D and 2D+T were highlighted well. However, the results section includes details which may be more fitting in the methods section.
(8) A few statements in the results section lacked literary support, which was not provided in the discussion either, such as support for increased variance of T-cell instantaneous speed on stimulated vs non-stimulated pMBMECs. Another example is the enhancement of cytokine stimulation directed T-cell movement on the pMBMECs that the authors observed but failed to relay the physiological relevance of it. The authors don't provide enough references for developments in the field prior to their work which form the basis and need for this technology.
(9) The rationale for use of OT-1 and 2D2-derived murine lymphocytes is unclear here. The OT-1 model has been generated to study antigen-specific CD8+ T cell responses, while the 2D2 model has been generated to recapitulate CD4 T cell-specific myelin oligodendrocyte glycoprotein (MOG) responses.
Figures and text:
(1) There are certain discrepancies and misarrangement of figures and text. For example, discussion of the effect of sheer flow on T cell attachment as part of the introduction in figure 1 and then mentioning it in the text again in the results section as part of figure 4 is repetitive.
(2) Section IV, subsection 1 of the results section, refers to 'data acquisition section above' in line 279, however the said section is part of materials and methods which is provided towards the end of the manuscript.
(3) There are figures in the manuscript that have not been referenced in the results section, for example, figure 3A and B. Figure 1 hasn't been addressed until subsection 7 of materials and methods
(4) A lack of significance but an observed trend of increased variance of T cell instantaneous speed is reported in line 296-298; however, the graph (figure 4G) shows a significant change in instantaneous speed between non-stimulated and TNFα-stimulated systems. This is misleading to the readers.
(5) The authors talk about three beginner experimentors testing the manual T cell tracking process but figure 5 only showcases data from two experimentors without stating the reason for excluding experimentor 1.
Discussion:
(1) While the discussion captures the major takeaways from the paper, it lacks relevant supporting references to relate the observation to physiological conditions and applicability.
(2) The discussion lacks connection to the results since the figures were not referenced while discussing an observed trend
(3) The authors briefly looked into mouse and human BMECs and their individual interaction with T-cells, but don't discuss the differences between the two, if any, that challenged their framework.
(4) Even though though the imaging tool relies on difference in appearance for detection, the authors talk about lack of feasibility in detecting transmigration of BMDMs due to their significantly different appearance. The statement lacks a problem solving approach to discuss how and why this was the case.
Relevance to the field:
Utilizing the framework provided by the authors, the application can be adapted and/or utilized for visualizing a range of different cell types, provided they are different in appearance. However, this would require extensive changes to the script and won't be adaptable in its current form.
-
-