Article activity feed

  1. Author Response

    Reviewer #1 (Public Review):

    This manuscript by O'Herron et al. describes an all-optical method combining optogenetic stimulation and 2-photon microscopy imaging to simultaneously manipulate and monitor brain microvasculature contractility in three dimensions. The method itself, which represents a microvasculature-targeted variation on a theme previously elaborated for simultaneous stimulation and monitoring of ensembles of neurons, employs a spatial light modulator (SLM) to create three-dimensional activation patterns in the brains of cranial window-model transgenic mice expressing the excitatory opsin, ReaChR, in mural cells (smooth muscle cells and pericytes) under control of the PDGFRβ promoter. The authors demonstrated that, by splitting a single 1040-nm stimulating beam into multiple beamlets using an SLM, this system is capable of optogenetically activating ReaChR at discrete depths in the neocortex, depolarizing mural cells and producing highly localized constrictions in targeted, individual microvessels. Using this system to investigate the kinetics of optogenetic-induced contraction and sensory-evoked dilation, the authors found that the onset of optogenetically evoked contraction was much more rapid than that of sensory-evoked dilation, concluding that the observed lag between sensory stimulation and vascular response does not reflect intrinsic limitations of mural cell contractile mechanisms but is instead attributable to the time course of neurovascular coupling mechanisms. They further found that by titrating the stimulation duration they could completely negate the vasodilatory response to a concurrent sensory stimulus.

    1. The red-shifted opsin, ReaChR, represents an improvement over opsins used in previously described 3D neuronal activation/monitoring systems. In particular, brief single-photon stimulation (100 ms) of ReaChR led to rapid, robust arteriole constrictions throughout the activation volume, whereas a previous generation ChR2 opsin required stimulation for seconds to achieve slowly appearing constrictions.

    Thank you for pointing out this key takeaway from our manuscript. In Figure 9 of the revised manuscript, we provide a comparison of ReaChR-induced vasoconstriction, with data previously collected across microvascular zones using line-scanning in ChR2-expressing mice. These data show how ReaChR produces faster and more potent vasoconstriction in alpha-SMA expressing SMCs and ensheathing pericytes, but has similar effects on the slow contraction with capillary pericytes.

    1. Single-photon stimulation was capable of completing stopping blood flow in a "first order pre-capillary branch". (Not clear what is meant by the phrase "pre-capillary branch"; anatomically, penetrating arterioles feed capillary branches.) While this speaks to the effectiveness of the method, it also highlights potential supraphysiological effects of stimulation and the importance of titrating stimulus intensity/duration to achieve physiologically meaningful responses.

    We have removed the term “pre-capillary” to avoid causing confusion, and now use the term arteriole-capillary transition to denote the alpha-SMA positive segment that lies between the penetrating arteriole (0th order) and the alpha-SMA low/negative capillaries (>4th order). The rationale for this terminology is provided in our new review (PMID: 34672718), which explains why the transitional zone should be considered a separate vessel type that is not arteriole and not capillary.

    We agree with the reviewer that titration of stimulation power/duration will be important and will depend on the application. We addressed this point by performing measurements of arteriole diameter with graded laser powers (Figures 5 & 7). There are many parameters to explore, but for the purposes of this manuscript, we clarify that the effect is titratable and that users should define physiological ranges in their specific circumstances, which may differ based on the experimental goals, age of mice, arteriolar size and vascular zone, and other factors.

    We also note that some applications may want to mimic pathophysiological levels of constriction, for example to mimic the effects of arterial vasospasm after subarachnoid hemorrhage, or ensheathing pericyte contraction with MCAo stroke (PMID: 26119027), or to examine the neural consequences of transient small vessel occlusion.

    1. In assessing effects of laser power, the authors assert that "increasing the laser power only slightly expanded the range of constriction". This seems a bit of an overstatement, given that increasing power (30-fold) had a greater effect on the spread (3x) than the magnitude (2x) of the response.

    Thank you for pointing this out. We have re-worded this section to avoid the overstatement and to emphasize the results more clearly on the spatial spread of constriction relative to laser power.

    The difference images in Figures 4B-C, G-H demonstrated that there was very limited spread of the constriction beyond the stimulation spots. We tested the effect of laser power on the spatial spread of constriction by stimulating with a broad range of power levels. We found that increasing the laser power led to a small increase in the spread of constriction. For example, a 30-fold increase in power (from 5 mW to 150 mW total power) led to ~3-fold increase in the spread of constriction (from ~25 µm to ~75 µm) (Figure 5A-H).

    1. The suggestion that penetrating brain arterioles possess a mechanism for upstream conduction of constrictive responses is intriguing (although this intrigue is tempered by the lack of experimental support for the operation of such a mechanism in the brain microvasculature).

    We are also intrigued by this hypothesis, which was supported by some evidence from a recent study of retinal vasculature. Kovacs-Oller et al. showed using neurocytin tracer injections into capillary pericytes, that they are linked through gap junctions and there is upstream directional diffusion of tracer. Further, they showed that electrical stimulation of a pericyte could lead to directional constriction from capillaries back to the arteriole in the retina (PMID: 32566247). The planar orientation of retinal vasculature makes this phenomenon easier to see. However, the 3D architecture of cortical vasculature is more challenging to study, particularly since the propagation along arterioles occurs along the Z axis, where spatiotemporal resolution of imaging is limited.

    Given our new data on the effects of laser power on axial spread (see reply to points 10-13 below) and the difficulty in separating active propagation from out-of-focus activation, we think there is not sufficient evidence to claim that penetrating arterioles are propagating the signal through some active process. Further experiments, including studies of the mechanisms involved, will be needed to address this hypothesis. Therefore, we have removed any discussion of potential propagation of the signal, and instead focus on the relationship between laser power and axial resolution of activation.

    1. The authors' premise for comparing contractile kinetics with sensory-evoked kinetics is flawed. In attempting to use the kinetics of optogenetic-induced constriction to infer something about the kinetics of sensory-evoked dilation, they are implicitly assuming that the kinetics of contraction and dilation processes intrinsic to mural cells are the same. This is highlighted by their use of the phrase "kinetics of the vasculature", which elides the possibility that dilation and contraction kinetics intrinsic to mural cells are different. Support for this latter possibility is provided by a previous report on renal afferent arterioles showing that the kinetics of myogenic constriction in arterioles are "substantially faster" than those of dilation (PMID: 24173354). Thus, their data do not rule out the possibility that the delay between sensory stimulation and vascular response reflects a slower intrinsic dilatory response rather than the time course of neurovascular coupling mechanisms. Furthermore, arterioles have an internal elastic lamina (IEL), which also determines the rates and degree of constriction and dilation. The IEL ends with the arterioles, and vessels with ensheathing contractile pericytes (and downstream) lack the constraints of the IEL.

    We thank the reviewer for this constructive critique. We agree that there are many issues in comparing kinetics between sensory evoked dilation and our optogenetic constriction. We have re-worded this section to avoid any mechanistic implications in the discussion of the kinetics of the different processes. However, we wish to still incorporate the details about the rapid kinetics of constriction to highlight the utility of the approach to intervene/perturb sensory-evoked responses, given that contraction can be titrated and precisely timed. We discuss the utility of this approach further below.

    1. It's not at all clear how overriding sensory-evoked dilation with optogenetically generated constriction provides a means for distinguishing neural activity from vascular responses. In particular, it is not clear how performing this maneuver while monitoring neuronal activity can provide the suggested insight into "aspects" of functional hyperemia that are essential to neuronal function beyond the relatively trivial observation that there is a point at which blood flow is too low to support continued neuronal activity.

    Thank you for raising this point. We have added more detail to our thoughts on why over-riding functional hyperemia could provide insight into the dependence of neural activity on the blood flow increase. Neural circuits are extremely complex with many different sub-types of neurons playing different roles. These subtypes have been shown to have different metabolic sensitivities and thus, may be differentially affected by blocking functional hyperemia (PMID: 26284893). This could lead to altered circuit activity which could have profound consequences for neural processing. Additionally, the energy budgets of different cellular functions within neurons are quite different (PMID: 22434069) and reducing available energy by blocking functional hyperemia could lead to differing degrees of dysfunction across important cellular processes (e.g. re-establishing the membrane potential, recycling neurotransmitters) which could again have important consequences for neural coding. Furthermore, it has been shown that there is a steep gradient of oxygen moving away from penetrating arterioles, and so neurons at greater distances from vessels may be differentially affected by blocking the hyperemic response (PMID: 21940458).

    1. With the exception of vasculo-neural coupling, where it would be the method of choice, the technology described leaves the impression of a capability in search of an application. That said, the ability to control blood flow to the point of completely stopping it may ultimately have applications in pathological settings.

    In addition to our response above on the utility of over-riding arteriole dilation during functional hyperemia, we have added to the discussion more potential uses of the technique. These include: (1) To be able to manipulate blood flow without using pharmacology or having to induce neural activity could be useful for a variety of studies involving intrinsic reactivity and compliance of vessels in both health and disease. (2) The different microvascular zones have distinct contractile kinetics. There are details that remain unstudied, such as the kinetics of different sized vessels, their location in the network, their identity as collateral arterioles or pial arterioles. Vascular optogenetics can dissect the contractile characteristics of different vessel types, similar to probing a circuit board. (3) Studies of the physiological significance of vasomotion, with respect to brain clearance of metabolic waste products. Being able to directly drive vasomotion and alter its amplitude and frequency will be an important tool for studies in this field. (4) Functional hyperemia is also impaired in many diseases, but this dysfunction could arise from impaired activity of neurons, astrocytes, or vessels. Therefore, a method to disentangle specific changes to blood vessels in vivo could be useful for understanding the vascular contributions to such diseases.

    Reviewer #2 (Public Review):

    The manuscript by O'Herron et al. describes a new technique for all-optical interrogation of the vasculature in vivo. They expressed optogenetic actuator ReaChR in vascular smooth muscle. They activated ReaChR using single-photon or 2-photon absorption. In both cases, they observed rapid and reversible constriction (presumably, due to Ca increase). Single-photon activation produced widespread constriction; two-photon activation allowed targeting of individual vessels. Using a commercial 2-photon system with a spatial light modulator on the photoactivation 1040-nm beam, they demonstrated localized constriction at multiple points along the small and large cerebral arterioles at once targeted by individual beamlets. Overall, this is a very interesting paper that clearly lays out the methodology and experimental design and carefully considers a number of potential limitations and pitfalls. This paper will serve as a valuable recourse for a large community of eLife readers interested in cerebrovascular physiology in health and disease as well as in neurovascular coupling and interpretation of noninvasive imaging.

    Given the chronic nature of the optical window, it is not clear why imaging was done under anesthesia. This point requires explanation. There is a concern that targeting of the vessel wall not possible in awake animals due to brain motion. If yes, that would be a serious limitation of the methodology.

    To ensure that our method is compatible with awake experiments, we have added awake data to the manuscript (Figure 10). We show that individual vessels can be independently targeted in the awake animal and the outcomes are not profoundly different than in the anesthetized state. As with all awake experiments, due diligence must be taken to ensure the preparation is as stable as possible, and the occasional trial may have to be removed if motion artifacts are too large.

    Reviewer #3 (Public Review):

    Strengths: In the vascular field, previous implementation of optogenetics to constrict and dilate blood vessels, has used either single photon full field and fiber illumination, or alternatively confocal and 2-photon scanning of individual vascular segments with raster scanning. The former is limited in spatial precision, activating multiple vessels over a large area, whereas raster scanning is not ideal for accumulating currents and often results in slow temporal precision. Spatial light modulator (SLM) generated diffraction patterns to achieve patterned illumination have become increasingly used in neuroscience to achieve reliable 2-photon activation of targeted neuron populations. Here the authors use this technology to depolarize and constrict smooth muscle cells in vivo. By imaging and stimulating with 2 laser lines and different optical paths they are able to stimulate opsin expressing cells and image simultaneously, which is advantageous. By using the Red-shifted opsin ReaChR for their experiments, it is possible to combine this approach (cautiously) with imaging many of the classically used 2-photon fluorophores and genetic indicators, with excitation spectrums <1040nm. Future work using variations of the technique is likely to gain valuable insight into neurovascular biology.

    Weaknesses: A major limitation of the current study is that although the authors achieve high spatial precision of ReaChR activation in the xy plane, the axial precision appears extremely poor compared to what would have been expected. For example, in Fig. 5-1 (using a 0.8NA, 16x objective), the authors achieve equivalent levels of surface arteriole constriction even when the SLM is focused 200um above the brain, and even larger constrictions as they initially move the focus away from the imaging plane. Although the axial spatial resolution appears better with the 1.1NA - 25X objective, such a large point spread function largely limits the utility of the technique, as there will always be a concern as whether the effects are spatially specific and not due to activation of vascular cells above and/or below the site of interest. This experiment that the authors have presented on axial precision is extremely important as it outlines a very important limitation of the technique (which is likely power dependent), but it remains to be completely characterized and understood. One possibility is that the power levels used by the authors are already above saturation, a problem raised by Rickgauer and Tank (2009)- PMID: 19706471, and therefore they may be able to refine the axial precision by using lower power. Further controls would be valuable to understand the precise cause of this large axial spread as it doesn't quite add up with the diameter of the bleach spot shown in figure 5-1D (some suggestions outlined in recommendations to the authors).

    We agree with the reviewers on this point. We conducted several new experiments to help elucidate the limits of axial resolution. First, we have dropped the comparison between objectives with different NA’s. This leads to unnecessary confusion, and it is common knowledge that lower NA objectives will have poorer resolution in the axial plane. We now mention this as a factor to consider, but have removed it from the figures. Second, we have shown, as the reviewer suggests below, that the stimulation power used has a dramatic effect on the axial spread of constriction (Figure 6E and Figure 7). Low powers indeed show a more narrow axial spread. However, we typically use higher powers (near or above 100 mW) to generate large constrictions in penetrating arteries, and we also include these levels to show the greater axial spread they cause. In summary, we confirm with lower powers the 3D precision of the two-photon optogenetic technique, and we show that higher powers can be used to broadly constrict penetrating arterioles for studies seeking to modulate blood flow in columns of cortical tissue supplied by penetrating arterioles.

    Regarding the stated inconsistency with the bleached spots, we think this mostly has to do with the difference between photo-bleaching fluorescent material (requiring lots of laser power) and photo-activating opsin channels (which can be done with much less power for very sensitive opsins). Additionally, the slide we bleached is optimally activated at ~800nm and so our 1040 nm stimulation required enormous power to burn the spot.

    The current version of the paper also lacks adequate quantification of the results as it is composed primarily of representative examples, which limits a proper assessment of reproducibility and variability of the effects.

    We agree that showing population averages will be more informative to the field. In the original submission, we showed mostly examples because the large parameter space (size and number of spots, position on vessels, duration and intensity of stimulation; if a stimulation train, the duration, number, and inter-pulse interval of stimulation) was explored in the early data rather than picking one set of conditions. However, we have now collected new data where parameters were typically the same and included population average plots in the figures that previously had only individual examples (Figures 2G,I, 4I,M, 4-1C, 5I, 6E,F, 7, 11-2 ) as well as the new data (Figures 8, 9, 10).

    Was this evaluation helpful?
  2. Evaluation Summary:

    This manuscript by O'Herron et al. describes an all-optical method combining optogenetic stimulation and 2-photon microscopy imaging to simultaneously manipulate and monitor brain microvasculature contractility in three dimensions. The method employs a spatial light modulator to create three-dimensional activation patterns in the brains of cranial window-model transgenic mice expressing the excitatory opsin, ReaChR, in mural cells (smooth muscle cells and pericytes). This provides a powerful new in vivo technique to control blood flow into the brain and to understand its actions on brain function.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1, Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

    Was this evaluation helpful?
  3. Reviewer #1 (Public Review):

    This manuscript by O'Herron et al. describes an all-optical method combining optogenetic stimulation and 2-photon microscopy imaging to simultaneously manipulate and monitor brain microvasculature contractility in three dimensions. The method itself, which represents a microvasculature-targeted variation on a theme previously elaborated for simultaneous stimulation and monitoring of ensembles of neurons, employs a spatial light modulator (SLM) to create three-dimensional activation patterns in the brains of cranial window-model transgenic mice expressing the excitatory opsin, ReaChR, in mural cells (smooth muscle cells and pericytes) under control of the PDGFRβ promoter. The authors demonstrated that, by splitting a single 1040-nm stimulating beam into multiple beamlets using an SLM, this system is capable of optogenetically activating ReaChR at discrete depths in the neocortex, depolarizing mural cells and producing highly localized constrictions in targeted, individual microvessels. Using this system to investigate the kinetics of optogenetic-induced contraction and sensory-evoked dilation, the authors found that the onset of optogenetically evoked contraction was much more rapid than that of sensory-evoked dilation, concluding that the observed lag between sensory stimulation and vascular response does not reflect intrinsic limitations of mural cell contractile mechanisms but is instead attributable to the time course of neurovascular coupling mechanisms. They further found that by titrating the stimulation duration they could completely negate the vasodilatory response to a concurrent sensory stimulus.

    1. The red-shifted opsin, ReaChR, represents an improvement over opsins used in previously described 3D neuronal activation/monitoring systems. In particular, brief single-photon stimulation (100 ms) of ReaChR led to rapid, robust arteriole constrictions throughout the activation volume, whereas a previous generation ChR2 opsin required stimulation for seconds to achieve slowly appearing constrictions.

    2. Single-photon stimulation was capable of completing stopping blood flow in a "first order pre-capillary branch". (Not clear what is meant by the phrase "pre-capillary branch"; anatomically, penetrating arterioles feed capillary branches.) While this speaks to the effectiveness of the method, it also highlights potential supraphysiological effects of stimulation and the importance of titrating stimulus intensity/duration to achieve physiologically meaningful responses.

    3. In assessing effects of laser power, the authors assert that "increasing the laser power only slightly expanded the range of constriction". This seems a bit of an overstatement, given that increasing power (30-fold) had a greater effect on the spread (3x) than the magnitude (2x) of the response.

    4. The suggestion that penetrating brain arterioles possess a mechanism for upstream conduction of constrictive responses is intriguing (although this intrigue is tempered by the lack of experimental support for the operation of such a mechanism in the brain microvasculature).

    5. The authors' premise for comparing contractile kinetics with sensory-evoked kinetics is flawed. In attempting to use the kinetics of optogenetic-induced constriction to infer something about the kinetics of sensory-evoked dilation, they are implicitly assuming that the kinetics of contraction and dilation processes intrinsic to mural cells are the same. This is highlighted by their use of the phrase "kinetics of the vasculature", which elides the possibility that dilation and contraction kinetics intrinsic to mural cells are different. Support for this latter possibility is provided by a previous report on renal afferent arterioles showing that the kinetics of myogenic constriction in arterioles are "substantially faster" than those of dilation (PMID: 24173354). Thus, their data do not rule out the possibility that the delay between sensory stimulation and vascular response reflects a slower intrinsic dilatory response rather than the time course of neurovascular coupling mechanisms. Furthermore, arterioles have an internal elastic lamina (IEL), which also determines the rates and degree of constriction and dilation. The IEL ends with the arterioles, and vessels with ensheathing contractile pericytes (and downstream) lack the constraints of the IEL.

    6. It's not at all clear how overriding sensory-evoked dilation with optogenetically generated constriction provides a means for distinguishing neural activity from vascular responses. In particular, it is not clear how performing this maneuver while monitoring neuronal activity can provide the suggested insight into "aspects" of functional hyperemia that are essential to neuronal function beyond the relatively trivial observation that there is a point at which blood flow is too low to support continued neuronal activity.

    7. With the exception of vasculo-neural coupling, where it would be the method of choice, the technology described leaves the impression of a capability in search of an application. That said, the ability to control blood flow to the point of completely stopping it may ultimately have applications in pathological settings.

    Was this evaluation helpful?
  4. Reviewer #2 (Public Review):

    The manuscript by O'Herron et al. describes a new technique for all-optical interrogation of the vasculature in vivo. They expressed optogenetic actuator ReaChR in vascular smooth muscle. They activated ReaChR using single-photon or 2-photon absorption. In both cases, they observed rapid and reversible constriction (presumably, due to Ca increase). Single-photon activation produced widespread constriction; two-photon activation allowed targeting of individual vessels. Using a commercial 2-photon system with a spatial light modulator on the photoactivation 1040-nm beam, they demonstrated localized constriction at multiple points along the small and large cerebral arterioles at once targeted by individual beamlets. Overall, this is a very interesting paper that clearly lays out the methodology and experimental design and carefully considers a number of potential limitations and pitfalls. This paper will serve as a valuable recourse for a large community of readers interested in cerebrovascular physiology in health and disease as well as in neurovascular coupling and interpretation of noninvasive imaging.

    Given the chronic nature of the optical window, it is not clear why imaging was done under anesthesia. This point requires explanation. There is a concern that targeting of the vessel wall not possible in awake animals due to brain motion. If yes, that would be a serious limitation of the methodology.

    Was this evaluation helpful?
  5. Reviewer #3 (Public Review):

    Strengths: In the vascular field, previous implementation of optogenetics to constrict and dilate blood vessels, has used either single photon full field and fiber illumination, or alternatively confocal and 2-photon scanning of individual vascular segments with raster scanning. The former is limited in spatial precision, activating multiple vessels over a large area, whereas raster scanning is not ideal for accumulating currents and often results in slow temporal precision. Spatial light modulator (SLM) generated diffraction patterns to achieve patterned illumination have become increasingly used in neuroscience to achieve reliable 2-photon activation of targeted neuron populations. Here the authors use this technology to depolarize and constrict smooth muscle cells in vivo. By imaging and stimulating with 2 laser lines and different optical paths they are able to stimulate opsin expressing cells and image simultaneously, which is advantageous. By using the Red-shifted opsin ReaChR for their experiments, it is possible to combine this approach (cautiously) with imaging many of the classically used 2-photon fluorophores and genetic indicators, with excitation spectrums <1040nm. Future work using variations of the technique is likely to gain valuable insight into neurovascular biology.

    Weaknesses: A major limitation of the current study is that although the authors achieve high spatial precision of ReaChR activation in the xy plane, the axial precision appears extremely poor compared to what would have been expected. For example, in Fig. 5-1 (using a 0.8NA, 16x objective), the authors achieve equivalent levels of surface arteriole constriction even when the SLM is focused 200um above the brain, and even larger constrictions as they initially move the focus away from the imaging plane. Although the axial spatial resolution appears better with the 1.1NA - 25X objective, such a large point spread function largely limits the utility of the technique, as there will always be a concern as whether the effects are spatially specific and not due to activation of vascular cells above and/or below the site of interest. This experiment that the authors have presented on axial precision is extremely important as it outlines a very important limitation of the technique (which is likely power dependent), but it remains to be completely characterized and understood. One possibility is that the power levels used by the authors are already above saturation, a problem raised by Rickgauer and Tank (2009)- PMID: 19706471, and therefore they may be able to refine the axial precision by using lower power. Further controls would be valuable to understand the precise cause of this large axial spread as it doesn't quite add up with the diameter of the bleach spot shown in figure 5-1D (some suggestions outlined in recommendations to the authors).

    The current version of the paper also lacks adequate quantification of the results as it is composed primarily of representative examples, which limits a proper assessment of reproducibility and variability of the effects.

    Was this evaluation helpful?