Direct Deviations in Astrocyte Free Ca²⁺ Concentration Control Multiple Arteriole Tone States

Round 1

Reviewer 1 Report

The authors investigated the relationship between astrocytic [Ca²⁺] and the bidirectional regulation of vascular diameter in acute cortical slices. They found that clamping astrocytic [Ca²⁺] to 100 nM did not affect blood vessel diameter. In contrast, clamping [Ca²⁺] to 250 nM dilated and to 750 nM constricted blood vessels. Further pharmacological experiments revealed that vasodilation was mediated by COX-1 whereas vasoconstriction involved CYP450. These are interesting and clearly described results that directly show how astrocytic [Ca²⁺] bidirectionally regulates vascular diameter. Overall, this is a manuscript that focusses on a single important question and provides interesting new insights. I have only relatively minor comments.

1) It is stated that experiments were performed at room temperature. It should be briefly discussed if and how that affects the interpretation of the data and the comparison to other (in vivo) studies.

2) The authors very briefly discuss under which circumstances steady-state changes of astrocytic ‘resting’ [Ca²⁺] to 250 nM and 750 nM could occur. The other question is how long [Ca²⁺] would need to stay at that level. It appears unlikely that during a short and sufficiently large astrocytic [Ca²⁺] transient the vascular response rapidly switches from dilation to constriction and back to dilation. Do the authors think that the vascular tone represents the temporal average of astrocytic [Ca²⁺] and, if so, on what timescale?

Author Response

Reviewer 1

1) It is stated that experiments were performed at room temperature. It should be briefly discussed if and how that affects the interpretation of the data and the comparison to other (in vivo) studies.

We acknowledge that the slice experiments were conducted at room temperature rather than a temperature closer to physiological.  This was to be consistent with our prior work (Rosenegger et al., 2015)(Haidey J et al. Cell Reports 2021) and the necessity came about due to use of bulk loaded Rhod-2/AM.  Rhod-2 is rapidly extruded from cells at higher temperatures, preventing us from measuring stable baseline Ca²⁺ and making good measures during manipulations that change Ca²⁺, such as Ca²⁺ clamp. We load Rhod-2/AM at 37^oC to optimize dye loading, but subsequently allow brain slices to adjust to room temperature for 45 minutes prior to performing experiments as dye extrusion is much slower. 

The room temperature condition, along with other factors including lack of blood flow and perfusion pressure, are limitations of the brain slice preparation, which is now clearly stated in the discussion on page 7 line 363.  Our objective was to use calcium clamping with patch pipettes to test how a direct manipulation of free Ca2+ in astrocytes impacted arteriole tone, which is tractable in brain slices.  In vivo whole-cell patch clamp of peri-arteriole astrocytes has scarcely between reported in the literature as it is a very challenging technique.  Nevertheless, our findings in slice are consistent with other in vivo work showing a similar relationship between endfoot Ca2+ to the constrictions and dilations measured in a model of epilepsy using quantitative Ca2+ imaging (Zhang C. et al. 2017 JCBFM).  Ca2+ uncaging in astrocytes in vivo in healthy conditions evokes vasodilation (Takano et al. 2006 Nat Neurosci).  Furthermore, 20-HETE-dependent vasoconstrictions are observed in response to cortical spreading depression (CSD)(Fordsman J. et al 2013 J Neurosci), when a large Ca2+ signal is detected in astrocyte endfeet (Chuquet J. 2007 J Neurosci), both of which are consistent with our work here in slices.  These points and citations are already covered in the existing introduction and discussion, so we have added only a brief statement to the discussion regarding the reviewer’s request:

“An important limitation of our work is the use of acute brain slices, maintained at room temperature and containing arterioles that lack blood flow and perfusion pressure.  Though an unrealistic preparation, our data showing vasodilation at modest levels of astrocyte free Ca2+ and vasoconstriction at higher levels, is consistent with in vivo observations [11]-[12], [25], one of which employed quantitative Ca²⁺ measurements with Fluorescence Lifetime Imaging Microscopy [11].”   

2) The authors very briefly discuss under which circumstances steady-state changes of astrocytic ‘resting’ [Ca²⁺] to 250 nM and 750 nM could occur. The other question is how long [Ca²⁺] would need to stay at that level. It appears unlikely that during a short and sufficiently large astrocytic [Ca²⁺] transient the vascular response rapidly switches from dilation to constriction and back to dilation. Do the authors think that the vascular tone represents the temporal average of astrocytic [Ca²⁺] and, if so, on what timescale?

We agree with the reviewer that a large Ca²⁺ transient is unlikely to recruit these opposing pathways as Ca2+ rises and falls through these different concentrations.  It is much more likely that physiological Ca2+ levels in astrocytes that dilate blood vessels do not rise well above 250nM.  Only in pathological conditions, or perhaps when neuromodulators need to increase vascular tone, would higher Ca2+ concentrations be reached to evoke the role of 20-HETE.  This idea is conveyed in the introduction.  Another important thing to keep in mind is that we are indiscriminately raising free Ca2+ throughout the entire cytosol of the cell, and neighbouring cells.  Thus, these two pathways may not be simultaneous recruited when evoked in a more realistic manner.

The temporal average comment is interesting.  In a physiological context, endfoot Ca2+ may not experience long-lasting shifts as high as 250nM or 750 nM to provide sustained changes in arteriole tone, but perhaps an increased frequency of astrocyte Ca2+ transients towards these concentrations may bias the ambient engagement of COX-1 vs the 20-HETE. Whereas we explore the output of these competing vasoactive pathways ex vivo as sustained tone changes, perhaps in vivo, the degree to which these two pathways are engaged fine-tunes arteriole tone or the amplitude of vasomotion.  This is somewhat speculative, so we have not added any of these points to the discussion.

Reviewer 2 Report

Using Ca2+ imaging techniques,this study demonstrate that direct changes in astrocyte free Ca2+ can control multiple arteriole tone states through different mediators. Please explain how precise free Ca2+ concentration was obtained using rhod2! Please explain how a cox antagonist can alter free Ca2+ can cause opposite effect on vascular muscle for dilation and constriction 

Author Response

Reviewer 2

1) Please explain how precise free Ca2+ concentration was obtained using rhod2 

Free Ca²⁺ concentration was controlled using a BAPTA-based internal solution, amounts calculated by using the Max Chelator calculator, as described in our methods: (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcS.htm).  This solution buffers against increases or decreases in endfoot free Ca2+ around the target concentration of a particular solution, and has been successfully employed by us and others (Henneberger C. 2010 Nature)(Haidey J et al 2021 Cell Reports).  Rhod-2/AM only provided a relative measure of free Ca2+ changes and helped us confirm an expected increase in astrocyte free Ca2+ (250nM and 750nM) or little expected change (100nM).  Previously, we showed that the Rhod-2 signal decreases when we use a Ca2+-clamp solution targeted to 25nM, below resting Ca2+ (Haidey J. 2021 Cell Reports). Our measurements were consistent with what we expected from the proposed approach and with the literature.        

2) Please explain how a cox antagonist can alter free Ca2+ can cause opposite effect on vascular muscle for dilation and constriction 

We do not anticipate that the COX-1 antagonist used in this study directly altered astrocyte free Ca2+ because COX-1 is downstream of the Ca2+ signal in astrocytes important in arteriole diameter control (Takano et al. 2006 Nature Neuroscience).  We show that elevating endfoot Ca2+ to a target concentration of 250 nM elicits vasodilation, which was blocked by COX-1 antagonism. When we increased endfoot Ca2+ to an even higher concentration, 750 nM, we found that the vascular response reversed to a constriction, suggesting that in addition to COX-1, another vasoactive pathway was being engaged. Indeed, we found that a combination of COX-1 blockade and 20-HETE antagonism eliminated the 750 nM Ca2+ clamp effect. 

Reviewer 3 Report

This manuscript by Haidey and Gordon uses Ca2+ clamp whole-cell patching of peri-arteriole astrocytes to alter astrocyte free calcium and examine the resulting vascular response. They show very nicely that small/moderate astrocyte Ca2+ increases produce COX1-dependent vasodilations, likely via prostaglandins, while large increases in astrocyte Ca2+ produce vasoconstrictions mediated by 20-HETE. The dependence of the vascular response polarity on the magnitude of astrocytic Ca2+ was previously shown by several papers in an indirect way. By directly testing the effect of astrocyte Ca2+, this manuscript makes a significant, although small, step forward in understanding how astrocytes regulate cerebral blood flow in a complex manner. They further show that smaller Ca2+ signals induce arteriole constriction when oxygen level is high, supporting past findings.  The paper is a brief, well written report that would be an important addition to the field. I suggest a few improvements:

1. The authors mention that their data support past claims of 500 nM Ca²⁺ in astrocytes being a ‘cross-over point’ in the resulting vascular response. Experiments actually testing this idea, esp. whether COX1 and 20-HETE dependent pathway are ‘in balance’ at such a point (whether the predicted lack of response would convert to dilations in presence of HET0016, or constrictions in presence of SC-560) would really strengthen this study.
2. Rhod-2 signals (the indicator for Ca2+) is only shown as traces - Images of the Rhod-2 signal in the various example images should be included for Fig 1 A, D, G and Fig 2B, E, H. The authors show the Alexa 488 dye entering the endfeet and analyse t for that, but it would be nicer to see these data for the Rhod-2 signal, which is the critical signal for this manuscript.
3. Some of the Rhod-2 peak fluorescence signals in the 750 nM astrocyte Ca2+ condition are still very small (in the 0-20% range) and overlap with the size of the relative signals in 250 nM condition. Is there a correlation between the size of the measured Ca2+ signal and the resulting vessel response? This would be useful to see for all conditions (diameter change plotted against Ca2+ change) and would really strengthen the paper.
4. The defined concentration of Ca2+ is only exact in the patched astrocyte itself and, by diffusion alone, it will have degraded to a smaller concentration by the time the internal solution arrives at the endfeet. Therefore, the concentration at the endfeet is likely to be much lower than that introduced into the patched cell. This is important to discuss.
5. Inclusion of the high oxygen traces in Fig 1 J and M would be nice.
6. At which time point was the peak Ca2+ and peak diameter quantified? Were they correlated (diameter measurement was quantified at same time as peak Ca2+)? Fig 1 C, for example, shows that the average peak Ca2+ signal in 100 nM condition was approximately -18%. According to the average trace shown in Fig 1M, this % change is not reached until the 600s timepoint. Was this the ‘end’ timepoint where all diameter and Ca2+ signals were quantified? The methods say “Baseline Rhod-2/AM signal and arteriole diameter were compared to peak values at a single time point determined from averaged traces.” which is vague and unclear.

Author Response

Reviewer 3

1. The authors mention that their data support past claims of 500 nM Ca²⁺ in astrocytes being a ‘cross-over point’ in the resulting vascular response. Experiments actually testing this idea, esp. whether COX1 and 20-HETE dependent pathway are ‘in balance’ at such a point (whether the predicted lack of response would convert to dilations in presence of HET0016, or constrictions in presence of SC-560) would really strengthen this study.

We can appreciate the reviewer’s desire for us to run these interesting tests, but we believe that these experiments are unnecessary because 1) there must be a cross over point in the vascular response somewhere between 250nM and 750nM of free Ca2+ in astrocytes, 2) to get a clear answer from such a test would require a large number of extra experiments, and 3) it is not clear that such a cross over point is physiologically or pathologically relevant.  Because our 250nM Ca2+ experiment shows the contribution of just COX1, whereas the 750nM Ca2+ experiment shows the contribution of both COX1 and 20-HETE, it is reasonable to hypothesize that there is such a “balance point” where both pathways provide equal counteracting contributions to diameter.  However, determining this is not straight forward as we expect that every experiment will show a slightly different Ca2+ concentration where this crossover point occurs because astrocytes and vessels have variable physical and molecular characteristics that will govern their own unique responses.  It is expected that we would hit the balance point on very few experiments if we choose a test of 500nM for example.  Here, some vessels would likely exhibit a constriction whereas others would exhibit a dilation and thus we would need a relatively large sample.  Averaging all the data could produce a flat line, or a curve that is slightly dilated or slightly constricted.  Though this would point us in the right direction to where the cross over point is on average, we would have to guess again with the Ca2+ concentration and repeat the process. Better might be to titrate a range of Ca2+ concentrations into one astrocyte and see where the cross over point is, however, we do not have the equipment to swap out internal solutions, and this would require very long and difficult recordings.  Here, we would learn the cross over point for just one vessel at a time, and others would likely be a little different.  Adding blockers into either or these experiments only makes it more laborious.  Given this, we would prefer not to travel this long road with potentially little knowledge to be gained and with the relevance of this putative balance point uncertain. Because of this, we removed the sentence saying “Presumably, Ca²⁺-clamping astrocyte Ca²⁺ at ~500nM, would produce no arteriole response due to equally competing pathways.” as there is no reason to presume this in the discussion.    

2) Rhod-2 signals (the indicator for Ca2+) is only shown as traces - Images of the Rhod-2 signal in the various example images should be included for Fig 1 A, D, G and Fig 2B, E, H. The authors show the Alexa 488 dye entering the endfeet and analyse t for that, but it would be nicer to see these data for the Rhod-2 signal, which is the critical signal for this manuscript.

As requested by the reviewer, we have added Rhod-2 image data of representative endfeet to figure 1.  The point – that we can measure relative and expected changes in free Ca2+ with Rhod-2 – is shown nicely in figure 1 and we feel it unnecessary to add more data to figure 2.  In figure 1, we show a raw image, and two pseudo coloured images (16 colour LUT) for each condition to show the relative increase or no change in Ca2+ signal.  

3) Some of the Rhod-2 peak fluorescence signals in the 750 nM astrocyte Ca2+ condition are still very small (in the 0-20% range) and overlap with the size of the relative signals in 250 nM condition. Is there a correlation between the size of the measured Ca2+ signal and the resulting vessel response? This would be useful to see for all conditions (diameter change plotted against Ca2+ change) and would really strengthen the paper.

The reviewer offers an interesting suggestion regarding assessing the correlation between Ca2+ signals and the vessel response.  However, Rhod-2 provides only a relative measure from baseline values and though we find Rhod-2 useful for showing us the general direction of the Ca2+ change, unfortunately the magnitude of the change fails to provide an informative correlation to the magnitude/direction of the diameter change.  Nevertheless, the Ca2+ clamp internal solutions provided robust effects on vessel diameter, and to further support our use of Rhod-2 in this way, in our previous work we showed that the Rhod-2 signal does indeed decrease when we clamp astrocyte free Ca2+ at 25nM, below the resting value of 100nM (Haidey J. et al. 2021 Cell Reports).  Thus, Rhod-2 works for our intended purpose, but future investigations should use quantitative calcium imaging to determine the correlation between precise nM Ca2+ changes and the magnitude and direction of vascular responses.  Though not the same preparation as ours, this has been done previously in an animal model of epilepsy using FLIM (Zhang C. 2017 JCBFM).

4) The defined concentration of Ca2+ is only exact in the patched astrocyte itself and, by diffusion alone, it will have degraded to a smaller concentration by the time the internal solution arrives at the endfeet. Therefore, the concentration at the endfeet is likely to be much lower than that introduced into the patched cell. This is important to discuss.

The reviewer raises a valid point and it is for this reason that we patched astrocytes in close proximity to arterioles.  Note that the volume of internal solution being dialyzed is orders of magnitude greater than the volume of cytosol within the patched astrocyte; thus, we anticipate that the endfoot in question is largely saturated with Ca2+ clamp solution.  There is also a relatively large diameter process directly connecting the astrocyte soma to the endfeet, further lessening this concern.  Nevertheless, as the reviewer indicates, it is undoubtedly true that both the concentration of BAPTA and free Ca2+ in the calcium-clamp solution will be somewhat lower in the endfeet compared to the patch pipette.  If we can assume equal reductions of both species in the patched cell, then the clamp will still be at roughly the same intended free Ca2+ concentration, but the overall buffering capacity of the solution will be less (i.e. less able to deal with large increases or large decreases away from this equilibrium).  We have added these ideas to the discussion at the top of page 8.             

5) Inclusion of the high oxygen traces in Fig 1 J and M would be nice.

These traces have been added to figure 1 panel S

6) At which time point was the peak Ca2+ and peak diameter quantified? Were they correlated (diameter measurement was quantified at same time as peak Ca2+)? Fig 1 C, for example, shows that the average peak Ca2+ signal in 100 nM condition was approximately -18%. According to the average trace shown in Fig 1M, this % change is not reached until the 600s timepoint. Was this the ‘end’ timepoint where all diameter and Ca2+ signals were quantified? The methods say “Baseline Rhod-2/AM signal and arteriole diameter were compared to peak values at a single time point determined from averaged traces.” which is vague and unclear.

Thank you for spotting this lack of clarity in the methods.  We now say on page 3 line 107:

“Baseline Rhod-2/AM signal and arteriole diameter were compared to peak values, the latter of which were determined by examining the average trace data to find the time in which the peak occurred.  This time point was then used to extract a measurement from each of the individual experiments for statistical comparison.”

From this it follows that the Ca2+ peak and the diameter peak exhibited their own unique time point.  There is no reason to use the same time point for these different measurements.

Yes, sometimes the largest value occurred at the end of the experiment at 600 sec.  The small decrease in Rhod-2 fluorescence in the 100nM group at the end of the trace that the reviewer refers to is likely due to photobleaching.  Additionally, it is possible that the dilation peaks to 250nM may be slightly larger if we ran the experiment for longer (still climbing slightly by 10min) but larger dilations would just make the statistics even more significantly different then they already are.  

Round 2

Reviewer 3 Report

The addition of the Rhod2 images are very useful. One minor comment: I notice in the baseline pseudocolored image in each Rhod2 panel, there is a very fuzzy label saying ‘Rhod2’ with a scale bar. It would be better if this fuzzy label is removed and perhaps a scale bar provided separately.