Blood–Brain Barrier Dynamic Device with Uniform Shear Stress Distribution for Microscopy and Permeability Measurements

Round 1

Reviewer 1 Report

This study addresses an important issue in a novel way; they have designed a new device in which D3 BBB endothelial cells can be cultured on a 0.4um membrane. The channel for culturing the D3s has been designed in such a way that a large proportion of the cells in the centre of the channel can be subjected to consistent shear stress conditions at their apical surface, thereby mimicking blood flow, whilst simultaneously exposing their basolateral side to another environment beneath the 0.4um membrane, which could potentially mimic cerebral conditions. Furthermore the 0.4um membrane allows for permeability studies to be conducted and importantly the devices transparency allows for cells to be visualised using a microscope. The availability of these types of devices is limited and as the authors discuss in their introduction & discussion, many of the commercial ones and others developed by labs have limitations.

 

Despite the importance of this new device there are crucial details missing from this manuscript that need to be addressed and also some misinterpretation of the endothelial cell phenotypic data. Furthermore the poor English makes some sections difficult to understand, hence would benefit form a thorough read and edit .

 

Points that need to be addressed:

 

Introduction:

Figure in intro would benefit from a legend and needs to be cited in text.

In the introduction authors state that they have produced a device that:

• Reproduces BBB architecture by providing cerebral and vascular compartments, however authors didn’t mimic the cerebral compartment as no cerebral cells were present, furthermore the culture media used was identical to that on the apical side of the endothelium, hence this is a bit of a stretch and should be rephrased.
• Authors also state that they monitored cells in real time, when they actually removed device at intervals to assess under a microscope, so this statement is misleading and needs to be amended.
• Assesses barrier integrity, however authors didn’t use recognised resistance/impedance sensing, they assessed two intracellular molecules that are involved in the formation of cell junctions (beta-catenin and ZO1), when ideally to infer barrier formation they should be looking at the molecules that actually form the extracellular junctions between neighbouring endothelial cells i.e. VE-cadherin/CD144 and Claudin 5 (Claudin 5 is an excellent one for assessing BBB junctions). Beta catenin and ZO1 are great to use in addition to the actual junction molecules, to validate them, but not alone. So the term barrier integrity is inaccurate and needs to be amended (this repeated misinterpretation re the molecules assessed needs to be amended throughout the entire manuscript especially in 3.4)

 

 

Materials & Methods/Results

 

Being able to run 6 or 12 channels/treatments in parallel under identical shear stress conditions would be fantastic and is certainly a desirable feature, however it is not clear how the authors have achieved this? They state that they use the Ibidi pump which can take 4 units, so how have the authors managed 6 & 12, which in order to provide identical shear stress would need to run from one pump? Please can the authors provide a diagram in the manuscript clarifying this important aspect of their design.

 

Can the authors please clarify in the manuscript why they correctly used the unidirectional setting on the pump, rather than oscillating or pulsate.

 

The Ibidi pump is an air displacement pump, so did the authors observe a slight interruption in flow rate when the syringes switched? If so please discuss this limitation in manuscript.

 

Can the authors provide a diagram in the manuscript/sup data showing how the pump, units and flow chip are setup relative to the incubator i.e. is all the apparatus in the incubator or is the pump outside and tubing run into the incubator? If setup as on Ibidi web page can refer to this instead.

 

Please clarify in manuscript how long after seeding cells took to become confluent and method used to confirm confluence, before shear stress applied.

 

Please provide additional detail in manuscript re how media was replenished every 24hrs for static channel beneath membrane; this is important as ideally wouldn’t want to stop/disturb shear stress conditions above.

 

Can authors also please clarify in manuscript how media was replenished for the flow channel and how frequently this was done.

 

Can the authors clarify why they used Ringer-HEPES for Permeability measurement, rather than sustaining culture conditions (avoiding cell shock) in phenol red free media? Also why wasn’t this attempted under shear stress? Can the authors please justify the approach they used and discuss its limitations in manuscript.

 

From what I understand from the manuscript it appears authors repeatedly removed device to visualise cells under and microscope during the multi day culture period? Presumably the fluidic device would need to be disconnected, thereby interrupting the shear stress conditions. If this is the case can the authors please state this in the manuscript and discuss the issues associated with this.

 

In the future could this system fit inside a live cell culture microscopy platform for real time imaging?

 

The authors state that 82% of the monolayer can be subjected to consistent shear stress conditions, which is high relative to other platforms. Can the authors please provide a table in the manuscript showing this data from other platforms in order to validate their statement.

 

At end of section 3.2, need some discussion acknowledging the irregular surface of endothelial cells and their likely variable height, which will influence laminar flow.

 

In Fig 3 legend can authors please state that ‘these images are representative of ? independent experiments’. It would also be useful to conduct some image analysis on these images to robustly demonstrate the change in morphology when exposed to shear stress.

 

In Fig 4 legend can authors please state that ‘these images are representative of ? independent experiments’.

 

Section 3.4.: need to discuss limitations re using beta-catenin and ZO1, as discussed earlier in this report. Also there is punctate beta catenin staining (likely endosomal) and the relevance of this needs to be discussed in manuscript.

 

Discussion

Authors reviewed current devices and their limitations in discussion section, to justify their design approach, when maybe this info should be in the introduction? (But is ok to leave in discussion, just a note for the future).

 

You state that flow conditions can be individually adjusted for both channels, however you didn’t show this in study, amend text accordingly.

 

Amend text re cell junctions, as discussed earlier in report.

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper entitled “Blood-Brain Barrier dynamic device with uniform shear stress repartition for microscopy and permeability measurements” deals with  an innovative device specifically designed for cell culture under shear stress. The authors finely describe the device in term of design, materials, dimension and characterize it in term of flow profile, cell culture and barrier integrity and function.

The article is well written, and clearly exposed. Results are properly documented and conclusion supported.

I have no remarks.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

While the importance of devising cellular microfluidic models that mimic the blood-brain-barrier (BBB) can be considered very relevant to the study of disease processes and for the design of new drugs that can circumvent the barrier, it is equally important that these models mimic the cellular architecture and more importantly the cellular phenotypes of the real in vivo situation.

Given that astrocytes have a crucial role in maintaining the BBB through the release of GDNF, IL, BDNF etc that maintain the health of the tight junctions, one main drawback of this model is that these particular glial cells were excluded from this structure. It is well known that even mutations in GFAP have a huge impact on the functionality and regulation of the BBB properties (Glia. 22 (4), 390-400 (1998).)

Another drawback in the characterization of this model is the limited experimental studies that pertain to the permeability studies.

The model for the calculation of shear forces in my opinion is well devised. However, this could have been strengthened further by monitoring the release of NO that is known to be generated by endothelial cells during transmural changes in arterial pressure (as a function of flow).

The conclusion then fairly outlines some of the disadvantages of this model and proposes the way forward.

 

Others:

Since the method for the construction of the system is rather tedious and complex, it is in my opinion that a video that shows the different stages of each process would be more valuable than the figures presented. Moreover, the figures show some of the main stages in the fabrication process and are of good quality and are well described.

Fig 4.0. The ZO-1 in (B) are aligned vertically in relation to F-actin and B-catenin. I do not see why this should be the case.

The presence of B-catenin and ZO-1 ALONE, does not proof the existence of the restrictive function of the BBB. The integrity of the barrier assessment is in my opinion very limited. Lipophilic ad hydrophilic compounds and testing of the P-glycoprotein efflux transporters etc should have been conducted along with the use of a variety of dextran dyes with different molecular weights. Transendothelial electrical resistance measurements at junctions would have garnished further credibility. Same goes with testing for receptors involved in transcytosis.

Line 125 was coated FOR one hour

In the methods section, there is no mention of the purity of the endothelial cell line or the reproducibility between different experiments.

Minor - The verb tense should be changed in most cases. Example line 187: Results were presented should be changed to 'results are presented'. Thereafter, the Results section is riddled with inappropriate verb tenses (will be S-shaped, will be straight!) . These should be changed.

 

In conclusion, I endorse this study in my understanding that the authors address this critique and acknowledge some of the limitations of this study.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The authors have addressed most of the controversial issues raised by myself and have adequately strengthened their manuscript by including additional sections to substantiate further their findings. Further details of the research methodology were provided so that it is in my opinion that these details are now more appropriate to the scientific background that is applied.

The manuscript is technically correct and the methods are used correctly so that it is in my opinion that the data is sufficient to corroborate the claims. The reporting appears to also be  sufficiently transparent to repeat the experiments for those not familiar with these types of experiments.

New sections were added to address some of the technical and scientific issues based on proper citations.  Amendments were made for clarification of certain statements including a revamped discussion with clear supporting references. The figures are appropriate and the supporting evidence is reliable and properly applied. Based on my critique, the authors have acknowledged the general limitations of this study (and model) and have appropriately addressed these issues in a convincing manner.

In conclusion,the paper is now easy to follow and free from any grammatical errors. I endorse for publication as is and I hope that this work will be a useful resource in this field of science.