Comprehensive Physiology Wiley Online Library

Regulation of the Cerebral Circulation by Arterial Carbon Dioxide

Full Article on Wiley Online Library



ABSTRACT

Intact, coordinated, and precisely regulated cerebrovascular responses are required for the maintenance of cerebral metabolic homeostasis, adequate perfusion, oxygen delivery, and acid‐base balance during deviations from homeostasis. Increases and decreases in the partial pressure of arterial carbon dioxide (PaCO2) lead to robust and rapid increases and decreases in cerebral blood flow (CBF). In awake and healthy humans, PaCO2 is the most potent regulator of CBF, and even small fluctuations can result in large changes in CBF. Alterations in the responsiveness of the cerebral vasculature can be detected with carefully controlled stimulus‐response paradigms and hold relevance for cerebrovascular risk in steno‐occlusive disease. As changes in PaCO2 do not typically occur in isolation, the integrative influence of physiological factors such as intracranial pressure, arterial oxygen content, cerebral perfusion pressure, and sympathetic nervous activity must be considered. Further, age and sex, as well as vascular pathologies are also important to consider. Following a brief summary of key historical events in the development of our understanding of cerebrovascular physiology and an overview of the measurement techniques to index CBF this review provides an in‐depth description of CBF regulation in response to alterations in PaCO2. Cerebrovascular reactivity and regional flow distribution are described, with further consideration of how differences in reactivity of parallel networks can lead to the “steal” phenomenon. Factors that influence cerebrovascular reactivity are discussed and the mechanisms and regulatory pathways mediating the exquisite sensitivity of the cerebral vasculature to changes in PaCO2 are outlined. Finally, topical avenues for future research are proposed. © 2019 American Physiological Society. Compr Physiol 9:1101‐1154, 2019.

Comprehensive Physiology offers downloadable PowerPoint presentations of figures for non-profit, educational use, provided the content is not modified and full credit is given to the author and publication.

Download a PowerPoint presentation of all images


Figure 1. Figure 1. Cerebral blood flow measurement techniques pioneered by Angelo Mosso. (A) An illustration of Mosso's self‐built plethysmograph from his 1892 manuscript titled: Die Ermüdung: Aus dem italienischen übersetzt von J.glinzer. Here he implemented his invention to study an individual with a congenital skull defect (). (B) A drawing of Mosso's scale. Reproduced with permission ().
Figure 2. Figure 2.

Early pioneers in cerebrovascular physiology. This figure highlights a few (of many) key figures in the early progression of knowledge on cerebral blood flow regulation. Each has provided and been remembered for a major contribution to cerebrovascular physiology. Photos acquired, with permission, from:

https://en.wikipedia.org/wiki/William_Harvey (public domain)

https://en.wikipedia.org/wiki/Thomas_Willis (public domain)

https://en.wikipedia.org/wiki/Alexander_Monro_(secundus) (public domain)

https://en.wikipedia.org/wiki/Angelo_Mosso (public domain)

https://en.wikipedia.org/wiki/Charles_Scott_Sherrington (public domain)

Figure 3. Figure 3. The original experimental setup and recordings for the nitrous oxide technique developed by Kety and Schmidt. Panel A (top and bottom) depicts the subject laying and breathing N2O while they have an internal jugular venous and radial arterial catheter for the sampling of blood. Panels B and C depict the temporal dynamics of changes in N2O in both jugular and arterial blood (panel B), and the resulting arterial‐venous difference of N2O (panel C) across 10 min of N2O inhalation. Reproduced with permission (), Copyright © 1945 by American Physiological Society, and with permission ().
Figure 4. Figure 4. Influence of changes in middle cerebral artery diameter on the validity of transcranial Doppler ultrasound during alterations in arterial carbon dioxide. Previously reported changes in middle cerebral artery (MCA) diameter (Y‐axis) and their calculated impact on the discrepancy between flow and velocity measures during changes in end‐tidal PCO2 (PetCO2) are depicted. The percent discrepancy between velocity and flow measures are noted as the number within each symbol. To highlight the effect of changes in MCA diameter we estimated the potential difference between CBF and velocity changes using formalism described in (). It is evident that small changes in MCA diameter are responsible for large discrepancies between flow and velocity measures. Data are collated, with permission, from (). Modified and updated with permission (). Copyright © 2014 the American Physiological Society.
Figure 5. Figure 5. Duplex ultrasound image of the internal carotid artery and an example of an analysis output. (A) Simultaneous measurement of the arterial diameter (B‐mode image) and the blood velocity (pulse‐wave image) enable determination of volumetric flow through extra‐cranial cerebral arteries with high temporal and spatial resolution. (B) The raw output is presented where changes in flow, velocity, and diameter are visible across the cardiac cycle. Reproduced with permission (). © 2016 The Authors. The Journal of Physiology © 2016 The Physiological Society.
Figure 6. Figure 6. A schematic diagram of the effects of PCO2, cerebral blood flow (CBF), hemoglobin concentration in g/dL ([Hb]) on oxygen extraction fraction (OEF), deoxyhemoglobin concentration ([dOHb]), and BOLD MRI signal in arbitrary units. In panel A, the solid red line represents the Y‐axis parameters in health with [Hb] 15 g/dL. The red dots on the colored lines that intersect with the black dashed line represents all Y‐axis parameters at rest. We assume a baseline OEF of 0.4 at rest in health. The slope of the black dashed line shows that in the presence of anemia (green line), CBF does not completely compensate for anemia and some tissue hypoxia remains [as seen in panel B with reduced cerebral oxygen delivery (CDO2) ()]. The blue line represents increased [Hb] and lower baseline CBF and OEF. The color matched dashed arrows represent the changes in the BOLD signal with an iso‐metabolic hypercapnic stimulus. The Cerebrovascular reactivity is calculated as the percent change in BOLD (magnitude of BOLD shown by the color‐matched down arrows) divided by the change in PCO2 in mmHg (horizontal axis), or CVR (see section “Standardization and Utility of Cerebrovascular Reactivity”). The arrows beside “incr[dOHb]” and “incr BOLD” indicate the direction in which values increase but do not suggest that they are the same scale as OEF or each other. Note the concavity of the lines are reduced with lower [Hb].
Figure 7. Figure 7. Organization of cerebral blood flow and collateral blood flow. Cerebral blood flow is supplied by four extracranial arteries taking off from the aortic root, the internal carotid arteries (ICA), and the vertebral arteries (VA). The latter join to form the basilar artery (BA). Intracranially, the arteries anastomose around the circle of Willis (CoW) and separate as the major intracranial vessels perfusing each hemisphere: the anterior cerebral artery (ACA), middle cerebral artery (MCA), and posterior cerebral artery (PCA). Although there are multiple anastomoses between the major cerebral vascular territories, the blood supply to each territory, in baseline states, comes predominantly from its overlying artery (illustrated by color coding). In the case of partial or complete blockage of one of the cerebral arteries, the availability of blood flow from the contiguous arterial system determines the size of the infarction and the watershed area (penumbra) (). Collateral flow availability is highly variable and thought to depend on the rate of blockage progression (sudden vs. gradual), completeness of the blockage, age, comorbid disease, preexisting primary [CoW, ophthalmic ()] and secondary anastomoses [see: ()]. Similarly, penetrating arterioles in both gray and white matter predominantly perfuse silos of tissue with limited collateral branching between neighboring tissues. However, anastomoses do develop with age, and upstream vascular insufficiency [for review, see ()]. Figure reproduced with permission (); insert was in the original () and was reprinted with permission ().
Figure 8. Figure 8. Collateralization across microvascular beds. The vascular supply is outlined in diagrammatic form of a coronal section of the brain. The cortex and corpus collosum (1, 2) are supplied by short arterioles arising from pial arteries supplied by a single cerebral artery. The subcortical association bundles (3) receive a dual supply in the form of terminal twigs of the cortical vessels and long medullary arterioles that traverse the cortex and ramify in the white matter. This provides these areas with an overlap of two blood supplies. The external capsule and claustrum (4) have a dual supply from the vessels entering from the insular cortex. The centrum semiovale (5) is supplied by long arteries 2 to 5 cm entering from the brain surface. These arteries branch but have few communications with other arteries. They tend to form border zones with adjacent afferent arteriolar systems throughout the whole depth of the white matter. The basal ganglia and thalamus () get perfused by long arterioles and long muscular arteries from base of brain. These tend to be end arteries. This leaves border zones at deep portions of the centrum semiovale. Reproduced with permission ().
Figure 9. Figure 9. A simplified schematic of how segmental changes in the resistance of cerebral blood vessels organized in series influence flow and microvascular pressure. Changes in vasomotor tone (i.e., dilation/constriction) are labelled on each vessel while the resulting directional changes in flow through the arterial system and microvascular pressure are represented below each scenario with arrows. Two arrows simply represent a larger magnitude of effect than one arrow, while a sideways (↔) arrow indicates no change. All flow changes are relative to the very top left scenario and represent blood flow through all the in series vessel of an arterial system (resting large vessel and resting small vessel). (A) In a scenario where larger (and upstream) vessels are is a resting state, dilation of smaller (and downstream) vessels will lead to a reduction in microvascular pressure and increase in flow, whereas constriction of smaller vessels will increase microvascular pressure and reduce flow. (B) Constriction of a large upstream vessel will lead to a reduction in flow and microvascular pressure when the downstream small arteries do not change in diameter. If the downstream small arteries dilate, this will reduce microvascular pressure to an extent that maintains normal flow. However, if the small vessels constrict in concert with the large vessels, this will reduce flow despite unchanged microvascular pressure. (C) Dilation of a large upstream vessel will lead to an increase in flow and microvascular pressure when the downstream small arteries do not change in diameter. If the downstream small arteries dilate in concert with the large arteries, this will increase flow in the face of unchanged microvascular pressure. However, if the small vessels constrict, flow is unaltered and microvascular pressure is increased. Collectively, this figure represents a simplistic overview of how resistance (dilation vs. constriction) and driving pressure (effected by changes in microvascular pressure) dictate flow.
Figure 10. Figure 10. Basic regulation of cerebral blood flow by oxygen, carbon dioxide, and cerebral perfusion pressure. The dashed line is representative of resting normative values, with the X‐axis variables increasing in the right direction, and decreasing to the left. (A) As the partial pressure of arterial oxygen (PaO2) decreases, CBF increases; however, this response is not linear. Indeed, CBF does not increase appreciably until PaO2 drops to approximately 50 to 60 mmHg, the point at which reductions in PaO2 lead to appreciable reductions in SaO2 and CaO2. (B) When CBF is indexed against changes in CaO2, the response is generally linear, with reductions in CaO2 leading to increases in CBF. The right portion of the CaO2/CBF relationship is truncated given the inability to significantly increase CaO2 above resting values (e.g., 20 mL ⋅ dL−1) under physiological conditions. During hyperoxia, where CaO2 can be increased by pure oxygen breathing or hyperbaria, the relationship between CaO2 and CBF is less clear (). (C) CBF is linearly related to PaCO2, with increases in PaCO2 (hypercapnia) causing larger changes in CBF per unit change in PaCO2 when compared to reductions in PaCO2 (hypocapnia). (D) CBF is related to CPP and, when autoregulation is intact, consists of three distinct and adjoining linear components. Within a small range of altered CPP there is no alteration in CBF due to cerebral autoregulation; however, when cerebral autoregulatory mechanisms are exhausted either in a hypo‐ or hypertensive manner, CBF changes linearly with CPP.
Figure 11. Figure 11. The classical (left) and contemporary (right) cerebral autoregulation curves. Changes in cerebral perfusion pressure are buffered by intrinsic vascular mechanisms (). Following the review paper published by Lassen in 1959 (), it was thought that CBF remained constant between an arterial blood pressure of 50 to 150 mmHg (left panel). However, at least within‐subjects, it has now been well established that this autoregulation mechanism does not maintain constant perfusion over as wide of a range in arterial blood pressure (). Reproduced with permission (). © 2014 The Authors. The Journal of Physiology © 2014 The Physiological Society.
Figure 12. Figure 12. Cerebral blood flow reactivity to changes in arterial carbon dioxide. Changes in blood flow (ICA and VA) and blood velocity (MCA and PCA) are plotted during alterations in the partial pressure of arterial carbon dioxide (PaCO2). When comparing the ICA and VA to their upstream intracranial vessel (MCA and PCA, respectively) it is apparent that flow reactivity assessed in the ICA and VA with duplex ultrasound is approximately 50% greater than that assessed in the MCA and PCA with transcranial Doppler ultrasound. The reasons underlying this difference have been discussed in sections “Transcranial Doppler Ultrasound” and “Duplex Ultrasound” and Figure 4 legend. Changes in flow/velocity in reponse to hypercapnia (PaCO2 > 40 mmHg) lead to approximately 100% change in flow than that which occurs during hypocapnia (PaCO2 < 40 mmHg) for a matched change in PaCO2 (e.g., +10 mmHg vs. −10 mmHg). The presented data are mean ± standard deviation. Reproduced with permission (). © 2012 The Authors. The Journal of Physiology © 2012 The Physiological Society.
Figure 13. Figure 13. Carbon dioxide mediated alterations in cerebrovascular diameter occurs throughout the entire cerebrovascular tree. (A) Classic data from Wolff and Lennox () depicting hypercapnic vasodilation of pial arterioles in anesthetized cats. Two vessels can be seen; the one on the left is an image during craniotomy in the resting state, while the one on the right is during hypercapnia. The difference in diameter can be visualized by the addition of the dotted line to the diameter of the vessel in the right image. These data, along with extrapolation from others (), led to the belief that CBF regulation during changes in PaCO2 was mediated solely by pial arteries/arterioles. Adapted from with permission (). Wolff and Lennox, 1930. (B) Changes in diameter of the MCA as assessed by high resolution (3T) MRI in humans during changes in end‐tidal carbon dioxide (PETCO2). Reproduced with permission (). Copyright © 2014 the American Physiological Society. (C) Data depicting changes in diameter of the internal carotid artery (ICA) in humans throughout a wide range of PaCO2. It is now, therefore, established that changes in diameter of large intra‐ and extracranial cerebral arteries occurs in response to changes in PaCO2 along with changes at the pial artery and arteriolar level. Reproduced with permission () © 2012 The Authors. The Journal of Physiology © 2012 The Physiological Society.
Figure 14. Figure 14. Pathway to normalizing CVR. The strategy is to standardize the CVR for, first, the vasoactive stimulus by administering a standardized hypercapnic profile (Step 1). This is done with a computerized gas delivery system leveraging the targeting advantages of sequential gas delivery (see text for citations). Secondly, the CVR is normalized for anatomical location by coregistering the scans to the same space and calculating the mean and standard deviation of the corresponding voxels in the scan (Steps 2, 3). Finally, calculating the z‐score for each voxel, color coding the score, and mapping it back onto the anatomical scan (Steps 4‐6) resulting in a scan expressing the probability of normality of the score for each voxel. In this figure, the z map shows deep purple over the left hemisphere (right side of the map) indicating that CVR is more than two standard deviations below what is normal for the corresponding voxels. The left hemisphere has normal CVR. Reproduced with permission ().
Figure 15. Figure 15. The brain vascular stress test to identify the extent of collateralization in the presence of steno‐occlusive vascular disease. Left panel shows angiograms from two patients with near occluded carotid arteries (red arrow). Next to each, are axial slices from the CVR test. The CVR for each voxel is color coded according to the color scheme shown, and mapped onto the corresponding voxel of the anatomical scan. The upper panel shows steal physiology due to lack of collateral blood flow, and the lower one shows near compensatory collateral blood flow. The mechanism is shown in the schematic diagram to the right of each. Upper right panel: (1) inflow resistance which exists by virtue of the conductance of the vascular crown being greater than that of the single arterial stem; (2) the occusion of the branch; (3) the application of hypercapnia; (4) vasodilatory reserve (gray area) is exhausted downstream from the occlusive lesion (indicated by the black lines), but intact on the contralateral side (illustrated by); (5) flow is directed predominantly to the contralateral side, dropping the perfusion pressure at the bifurcation. The reduction in the perfusion pressure, reduces the flow further on the side ipsilateral to the lesion. In the lower right panel: (1) the carotid artery is limiting to the inflow from the central circulation; (2) a restrictive lesion is present in one branch; (3) a rout for collateral flow distal to the lesion capable of providing blood flow; (4) a hypercapnic stimulus is applied; (5) in this case, the collateral flow prevented the encroachment on the vasodilatory reserve (position of black lines relative to gray lines) normally recruited to maintain flow past the upstream obstruction. Thus, the vasodilatory stimulus results in a brisk response from the contralateral, as well as ipsilateral side. Further discussion of the mechanism is in the text. Reproduced with permission ().
Figure 16. Figure 16. The interaction between two vascular regions (e.g., voxels) during a ramp CO2 challenge. An electrical resistance model of two vascular regions. One region is represented by a reference resistance with a preset sigmoidal response to CO2. The other region is represented by a measured flow response to CO2 in a voxel. The model can then be solved to calculate the resistance response sigmoid in the examined voxel. (MAP, mean arterial pressure; Rart, arterial resistance; Pbranch, the perfusion pressure at the branch; R vox, the resistance at the interrogated voxel; Fvox, flow at the interrogated voxel; Rref, the resistance at the reference voxel; Fref, flow at the reference voxel; Ftotal, total flow; Pv, venous pressure). Reproduced with permission ().
Figure 17. Figure 17. Resistance sigmoid measures (left) and their corresponding anatomical maps (right). The vascular model is that in Figure 15 and assumes that this model may be expanded and contracted to contain the blood supply of both hemispheres, from extracranial vessels to capillaries. Thereby, every vessel or vascular bed is in fluid contact with every other vessel. As such, we use the examined unit illustrated as an analogous electrical circuit in Figure 16, to interrogate each voxel as if it is perfused in parallel with an ideal voxel of full vasodilatory responsiveness. Regardless of flow pattern type (see text for definition), the calculation invariably results in a sigmoidal resistance response to PetCO2. Various aspects of the resistance‐PETCO2 relationship are expressed in the sigmoidal shape of the relationship. The graph on the left shows some of these parameters diagrammatically, and the maps on the right show their anatomical distribution. The initial parameter is the resistance in maximum vasoconstriction in hypocapnia, and the amplitude shows the extent of the vasodilation and vasoconstriction. The midpoint of the sigmoid is the PetCO2 at its steepest slope and the range is the span of PETCO2 over which the response can be considered linear. The reserve and sensitivity show the vasodilatory reserve available from the subject's resting baseline PetCO2 and the response sensitivity (sigmoid slope) at that PetCO2.
Figure 18. Figure 18. The relationship between PCO2 and pH with varying [HCO3]. This figure highlights how, according to the Henderson‐Hasselbalch equation, the magnitude of pH changes with varying PCO2 are augmented at altitude following renal compensation for respiratory alkalosis (panel A). At sea level, where [HCO3] approximates 25 meq/L (), an increase in resting PaCO2 from 40 to 41 mmHg would lead to a reduction in pH of 0.01. Conversely, at altitude where [HCO3] may approximate 19 meq/L, an increase in resting PaCO2 from 20 to 21 mmHg would lead to a reduction in pH of 0.02. Panels B and C indicate how the magnitude increase in CBF per increase in PaCO2 is increased following acclimatization (panel B), with the resulting CVR values and their related arterial [HCO3] concentrations in panel C. Panel A was reproduced with permission (). Panels B and C were created from the data of ().
Figure 19. Figure 19. The effect of a progressive loss of vasodilatory reserve on the cerebral blood flow response to PCO2. Results are redrawn from experiments in dogs by Harper and Glass [(); (A), (B), and (C)] showing the effects of reducing perfusion pressure on the CBF response to CO2. (A) Normotensive; (B) hypotensive; (C) extreme hypotension; (D) an overlay of the fitted responses in A and B drawn to the same scale; and (E) the fitted responses presented as the % of maximum vasodilation. Reproduced, with permission, from (). In panels D and E, the solid line represents the normotensive CVR, the dashed line represents the hypotensive CVR, and the dotted line represents the extreme hypotensive CVR. Reproduced with permission ().
Figure 20. Figure 20. Increased blood pressure during hypercapnia is related to cerebrovascular reactivity. Panel A depicts the changes in mean arterial pressure (MAP) that occur during hypercapnia (highlighted by the box). Panel B shows the relationship between changes in MAP from baseline (rest) and changes in internal carotid (ICA) and vertebral artery (VA) flow during hypercapnia. As depicted, there is a linear relationship whereby increases in MAP are related to the increase in flow (ICA and VA) and thus important to consider when measuring cerebrovascular reactivity. Figure reproduced with permission (). © 2012 The Authors. The Journal of Physiology © 2012 The Physiological Society.
Figure 21. Figure 21. Cerebrovascular reactivity at rest and during exercise. Internal carotid artery (ICA) blood flow was collected at rest and during exercise at 35% of maximal aerobic capacity in 32 young healthy individuals (4 females). Reactivity was assessed with a hypercapnic stimulus that was held at steady state for 2 min with dynamic end‐tidal forcing. The magnitude of hypercapnia ranged from +7.5 to +10 mmHg PETCO2 above baseline, and was consistent within individuals from rest to exercise. As is depicted by the mean traces, there is no difference in the slope of the response (i.e., reactivity) between ICA flow and the partial pressure of end‐tidal carbon dioxide (PETCO2) between rest and exercise. Error bars represent standard deviation.
Figure 22. Figure 22. Potential pathways regulating cerebrovascular CO2 reactivity. Hypercapnia leads to an increase in intra‐ and extracellular hydrogen ion concentration ([H+]o and [H+]i, respectively) (). While all the depicted cellular processes act to induce smooth muscle cell relaxation, these processes initiate in various locales. In the vessel lumen, the red blood cell can release adenosine triphosphate (ATP) following deformation (or shear) (), which through binding of purinergic (P2Y2) receptors leads to endothelial nitric oxide synthase (eNOS) activation (), in addition to shear dependent serine threonine specific protein kinase (Akt) activation of eNOS (). Further, acidosis itself stimulates ATP release from red blood cells (). Finally, within the lumen, acidic disproportionation (note: nonenzymatic production) of nitrite (NO2) into NO can occur (). Within endothelial cells (EC), NO production occurs as a result of the aforementioned shear and ATP mechanisms. In addition, activation of phospholipase‐A2 (PLA2) hydrolyzes arachidonic acid (AA) from membrane phospholipids, which is then, via cyclooxygenase, converted into prostaglandins (PGs), with the relevant vasodilator prostaglandins being PGE2 and PGI2. Increased PG synthesis also occurs in astrocytes in response to hypercapnia (). Prostaglandin production leads to concurrent reactive oxygen species (ROS) production () which may influence CO2 reactivity by direct vasodilator effects () or NO scavenging (), although experimental evidence for this is lacking (see text). Also downstream of AA is the production of epoxyeicosatrienoic acids (EET), which convey vasodilatory activity by increasing potassium channel conductance (). At the level of the vascular smooth muscle cells (VSMC), increased adenosine levels () leads to binding of A2A receptors and consequent adenylate cyclase activation. Activation of EP and IP prostaglandin receptors on the VSMC () also contribute to adenylate cyclase activation. Adenylate cyclase activation leads to upregulation of cyclic adenosine monophosphate (cAMP) activity. This cAMP acts via activation of cAMP‐dependent protein kinase, also termed protein kinase‐A (PKA), as well as by increasing potassium (K+) channel conductance (). Soluble guanylate cyclase (sGC) activation occurs downstream of NO production (), from both the endothelial cells (eNOS derived) and from neurons [neuronal NOS (nNOS) derived] (). This leads to upregulation of cyclic guanosine monophosphate (cGMP) and consequent activation of cGMP dependent protein kinase, also termed protein kinase‐G (PKG) (). Cross talk between cGMP and cAMP should also be considered and is influenced by phosphodiesterase (PDE) activity (see text). Both NO () and cGMP () also increase K+ channel conductance. Overall, the increases in K+ channel conductance may lead to VSMC hyperpolarization, inhibiting calcium (Ca2+) influx through voltage dependent Ca2+ channels while both PKA and PKG phosphorylate and deactivate myosin light chain kinase (MLCK) leading to reduced calcium sensitivity of the VSMC and vasodilation (relaxation) ().
Figure 23. Figure 23. Transient increases in shear stress and internal carotid artery vasodilation. (A) Via carbon dioxide manipulation, a transient (∼30 s) increase in shear stress elicits vasodilation of the internal carotid artery. This increase in internal carotid artery (ICA) diameter occurs following return of carbon dioxide and shear to baseline levels, and follows similar temporal pattern to that of a brachial FMD test. () (B) In the transient carbon dioxide test shown on the left, shear stress area under the curve (AUC) was positively correlated to the magnitude of vasodilation (following removal of a statistical outlier—circled). This implicates shear as a potential regulator of cerebral conduit artery vasomotor tone. This represents data recently published by our group. (). Reproduced with permission (). Copyright © 2017 the American Physiological Society.
Figure 24. Figure 24. Cyclic nucleotide cross talk during cAMP and cGMP administration and hypercapnia. Influence of cyclic nucleotide kinase blockers on the pial arteriolar diameter increases accompanying topical applications of the NO donor, SNAP and the adenylate cyclase activator, forskolin (A,B) and exposure to hypercapnia (PaCO2 = 60 mmHg) (C, D). KT‐5720, PKA‐selective inhibitor; KT‐5823, PKG selective inhibitor. As depicted in panel A, administration of the PKA inhibitor KT‐5720 followed by the PKG inhibitor KT‐5823 lead to additive inhibitions of dilation in response to the application of cGMP donor (SNAP) and cAMP donor (Forskolin). The same response was observed when the order of bloackades were switched (panel B). However, the same blockades produced a hypo‐additive reduction in CO2 reactivity indicating an overlap in function, or cross talk as has been referred to thus far in this review. Reproduced with permission ().


Figure 1. Cerebral blood flow measurement techniques pioneered by Angelo Mosso. (A) An illustration of Mosso's self‐built plethysmograph from his 1892 manuscript titled: Die Ermüdung: Aus dem italienischen übersetzt von J.glinzer. Here he implemented his invention to study an individual with a congenital skull defect (). (B) A drawing of Mosso's scale. Reproduced with permission ().


Figure 2.

Early pioneers in cerebrovascular physiology. This figure highlights a few (of many) key figures in the early progression of knowledge on cerebral blood flow regulation. Each has provided and been remembered for a major contribution to cerebrovascular physiology. Photos acquired, with permission, from:

https://en.wikipedia.org/wiki/William_Harvey (public domain)

https://en.wikipedia.org/wiki/Thomas_Willis (public domain)

https://en.wikipedia.org/wiki/Alexander_Monro_(secundus) (public domain)

https://en.wikipedia.org/wiki/Angelo_Mosso (public domain)

https://en.wikipedia.org/wiki/Charles_Scott_Sherrington (public domain)



Figure 3. The original experimental setup and recordings for the nitrous oxide technique developed by Kety and Schmidt. Panel A (top and bottom) depicts the subject laying and breathing N2O while they have an internal jugular venous and radial arterial catheter for the sampling of blood. Panels B and C depict the temporal dynamics of changes in N2O in both jugular and arterial blood (panel B), and the resulting arterial‐venous difference of N2O (panel C) across 10 min of N2O inhalation. Reproduced with permission (), Copyright © 1945 by American Physiological Society, and with permission ().


Figure 4. Influence of changes in middle cerebral artery diameter on the validity of transcranial Doppler ultrasound during alterations in arterial carbon dioxide. Previously reported changes in middle cerebral artery (MCA) diameter (Y‐axis) and their calculated impact on the discrepancy between flow and velocity measures during changes in end‐tidal PCO2 (PetCO2) are depicted. The percent discrepancy between velocity and flow measures are noted as the number within each symbol. To highlight the effect of changes in MCA diameter we estimated the potential difference between CBF and velocity changes using formalism described in (). It is evident that small changes in MCA diameter are responsible for large discrepancies between flow and velocity measures. Data are collated, with permission, from (). Modified and updated with permission (). Copyright © 2014 the American Physiological Society.


Figure 5. Duplex ultrasound image of the internal carotid artery and an example of an analysis output. (A) Simultaneous measurement of the arterial diameter (B‐mode image) and the blood velocity (pulse‐wave image) enable determination of volumetric flow through extra‐cranial cerebral arteries with high temporal and spatial resolution. (B) The raw output is presented where changes in flow, velocity, and diameter are visible across the cardiac cycle. Reproduced with permission (). © 2016 The Authors. The Journal of Physiology © 2016 The Physiological Society.


Figure 6. A schematic diagram of the effects of PCO2, cerebral blood flow (CBF), hemoglobin concentration in g/dL ([Hb]) on oxygen extraction fraction (OEF), deoxyhemoglobin concentration ([dOHb]), and BOLD MRI signal in arbitrary units. In panel A, the solid red line represents the Y‐axis parameters in health with [Hb] 15 g/dL. The red dots on the colored lines that intersect with the black dashed line represents all Y‐axis parameters at rest. We assume a baseline OEF of 0.4 at rest in health. The slope of the black dashed line shows that in the presence of anemia (green line), CBF does not completely compensate for anemia and some tissue hypoxia remains [as seen in panel B with reduced cerebral oxygen delivery (CDO2) ()]. The blue line represents increased [Hb] and lower baseline CBF and OEF. The color matched dashed arrows represent the changes in the BOLD signal with an iso‐metabolic hypercapnic stimulus. The Cerebrovascular reactivity is calculated as the percent change in BOLD (magnitude of BOLD shown by the color‐matched down arrows) divided by the change in PCO2 in mmHg (horizontal axis), or CVR (see section “Standardization and Utility of Cerebrovascular Reactivity”). The arrows beside “incr[dOHb]” and “incr BOLD” indicate the direction in which values increase but do not suggest that they are the same scale as OEF or each other. Note the concavity of the lines are reduced with lower [Hb].


Figure 7. Organization of cerebral blood flow and collateral blood flow. Cerebral blood flow is supplied by four extracranial arteries taking off from the aortic root, the internal carotid arteries (ICA), and the vertebral arteries (VA). The latter join to form the basilar artery (BA). Intracranially, the arteries anastomose around the circle of Willis (CoW) and separate as the major intracranial vessels perfusing each hemisphere: the anterior cerebral artery (ACA), middle cerebral artery (MCA), and posterior cerebral artery (PCA). Although there are multiple anastomoses between the major cerebral vascular territories, the blood supply to each territory, in baseline states, comes predominantly from its overlying artery (illustrated by color coding). In the case of partial or complete blockage of one of the cerebral arteries, the availability of blood flow from the contiguous arterial system determines the size of the infarction and the watershed area (penumbra) (). Collateral flow availability is highly variable and thought to depend on the rate of blockage progression (sudden vs. gradual), completeness of the blockage, age, comorbid disease, preexisting primary [CoW, ophthalmic ()] and secondary anastomoses [see: ()]. Similarly, penetrating arterioles in both gray and white matter predominantly perfuse silos of tissue with limited collateral branching between neighboring tissues. However, anastomoses do develop with age, and upstream vascular insufficiency [for review, see ()]. Figure reproduced with permission (); insert was in the original () and was reprinted with permission ().


Figure 8. Collateralization across microvascular beds. The vascular supply is outlined in diagrammatic form of a coronal section of the brain. The cortex and corpus collosum (1, 2) are supplied by short arterioles arising from pial arteries supplied by a single cerebral artery. The subcortical association bundles (3) receive a dual supply in the form of terminal twigs of the cortical vessels and long medullary arterioles that traverse the cortex and ramify in the white matter. This provides these areas with an overlap of two blood supplies. The external capsule and claustrum (4) have a dual supply from the vessels entering from the insular cortex. The centrum semiovale (5) is supplied by long arteries 2 to 5 cm entering from the brain surface. These arteries branch but have few communications with other arteries. They tend to form border zones with adjacent afferent arteriolar systems throughout the whole depth of the white matter. The basal ganglia and thalamus () get perfused by long arterioles and long muscular arteries from base of brain. These tend to be end arteries. This leaves border zones at deep portions of the centrum semiovale. Reproduced with permission ().


Figure 9. A simplified schematic of how segmental changes in the resistance of cerebral blood vessels organized in series influence flow and microvascular pressure. Changes in vasomotor tone (i.e., dilation/constriction) are labelled on each vessel while the resulting directional changes in flow through the arterial system and microvascular pressure are represented below each scenario with arrows. Two arrows simply represent a larger magnitude of effect than one arrow, while a sideways (↔) arrow indicates no change. All flow changes are relative to the very top left scenario and represent blood flow through all the in series vessel of an arterial system (resting large vessel and resting small vessel). (A) In a scenario where larger (and upstream) vessels are is a resting state, dilation of smaller (and downstream) vessels will lead to a reduction in microvascular pressure and increase in flow, whereas constriction of smaller vessels will increase microvascular pressure and reduce flow. (B) Constriction of a large upstream vessel will lead to a reduction in flow and microvascular pressure when the downstream small arteries do not change in diameter. If the downstream small arteries dilate, this will reduce microvascular pressure to an extent that maintains normal flow. However, if the small vessels constrict in concert with the large vessels, this will reduce flow despite unchanged microvascular pressure. (C) Dilation of a large upstream vessel will lead to an increase in flow and microvascular pressure when the downstream small arteries do not change in diameter. If the downstream small arteries dilate in concert with the large arteries, this will increase flow in the face of unchanged microvascular pressure. However, if the small vessels constrict, flow is unaltered and microvascular pressure is increased. Collectively, this figure represents a simplistic overview of how resistance (dilation vs. constriction) and driving pressure (effected by changes in microvascular pressure) dictate flow.


Figure 10. Basic regulation of cerebral blood flow by oxygen, carbon dioxide, and cerebral perfusion pressure. The dashed line is representative of resting normative values, with the X‐axis variables increasing in the right direction, and decreasing to the left. (A) As the partial pressure of arterial oxygen (PaO2) decreases, CBF increases; however, this response is not linear. Indeed, CBF does not increase appreciably until PaO2 drops to approximately 50 to 60 mmHg, the point at which reductions in PaO2 lead to appreciable reductions in SaO2 and CaO2. (B) When CBF is indexed against changes in CaO2, the response is generally linear, with reductions in CaO2 leading to increases in CBF. The right portion of the CaO2/CBF relationship is truncated given the inability to significantly increase CaO2 above resting values (e.g., 20 mL ⋅ dL−1) under physiological conditions. During hyperoxia, where CaO2 can be increased by pure oxygen breathing or hyperbaria, the relationship between CaO2 and CBF is less clear (). (C) CBF is linearly related to PaCO2, with increases in PaCO2 (hypercapnia) causing larger changes in CBF per unit change in PaCO2 when compared to reductions in PaCO2 (hypocapnia). (D) CBF is related to CPP and, when autoregulation is intact, consists of three distinct and adjoining linear components. Within a small range of altered CPP there is no alteration in CBF due to cerebral autoregulation; however, when cerebral autoregulatory mechanisms are exhausted either in a hypo‐ or hypertensive manner, CBF changes linearly with CPP.


Figure 11. The classical (left) and contemporary (right) cerebral autoregulation curves. Changes in cerebral perfusion pressure are buffered by intrinsic vascular mechanisms (). Following the review paper published by Lassen in 1959 (), it was thought that CBF remained constant between an arterial blood pressure of 50 to 150 mmHg (left panel). However, at least within‐subjects, it has now been well established that this autoregulation mechanism does not maintain constant perfusion over as wide of a range in arterial blood pressure (). Reproduced with permission (). © 2014 The Authors. The Journal of Physiology © 2014 The Physiological Society.


Figure 12. Cerebral blood flow reactivity to changes in arterial carbon dioxide. Changes in blood flow (ICA and VA) and blood velocity (MCA and PCA) are plotted during alterations in the partial pressure of arterial carbon dioxide (PaCO2). When comparing the ICA and VA to their upstream intracranial vessel (MCA and PCA, respectively) it is apparent that flow reactivity assessed in the ICA and VA with duplex ultrasound is approximately 50% greater than that assessed in the MCA and PCA with transcranial Doppler ultrasound. The reasons underlying this difference have been discussed in sections “Transcranial Doppler Ultrasound” and “Duplex Ultrasound” and Figure 4 legend. Changes in flow/velocity in reponse to hypercapnia (PaCO2 > 40 mmHg) lead to approximately 100% change in flow than that which occurs during hypocapnia (PaCO2 < 40 mmHg) for a matched change in PaCO2 (e.g., +10 mmHg vs. −10 mmHg). The presented data are mean ± standard deviation. Reproduced with permission (). © 2012 The Authors. The Journal of Physiology © 2012 The Physiological Society.


Figure 13. Carbon dioxide mediated alterations in cerebrovascular diameter occurs throughout the entire cerebrovascular tree. (A) Classic data from Wolff and Lennox () depicting hypercapnic vasodilation of pial arterioles in anesthetized cats. Two vessels can be seen; the one on the left is an image during craniotomy in the resting state, while the one on the right is during hypercapnia. The difference in diameter can be visualized by the addition of the dotted line to the diameter of the vessel in the right image. These data, along with extrapolation from others (), led to the belief that CBF regulation during changes in PaCO2 was mediated solely by pial arteries/arterioles. Adapted from with permission (). Wolff and Lennox, 1930. (B) Changes in diameter of the MCA as assessed by high resolution (3T) MRI in humans during changes in end‐tidal carbon dioxide (PETCO2). Reproduced with permission (). Copyright © 2014 the American Physiological Society. (C) Data depicting changes in diameter of the internal carotid artery (ICA) in humans throughout a wide range of PaCO2. It is now, therefore, established that changes in diameter of large intra‐ and extracranial cerebral arteries occurs in response to changes in PaCO2 along with changes at the pial artery and arteriolar level. Reproduced with permission () © 2012 The Authors. The Journal of Physiology © 2012 The Physiological Society.


Figure 14. Pathway to normalizing CVR. The strategy is to standardize the CVR for, first, the vasoactive stimulus by administering a standardized hypercapnic profile (Step 1). This is done with a computerized gas delivery system leveraging the targeting advantages of sequential gas delivery (see text for citations). Secondly, the CVR is normalized for anatomical location by coregistering the scans to the same space and calculating the mean and standard deviation of the corresponding voxels in the scan (Steps 2, 3). Finally, calculating the z‐score for each voxel, color coding the score, and mapping it back onto the anatomical scan (Steps 4‐6) resulting in a scan expressing the probability of normality of the score for each voxel. In this figure, the z map shows deep purple over the left hemisphere (right side of the map) indicating that CVR is more than two standard deviations below what is normal for the corresponding voxels. The left hemisphere has normal CVR. Reproduced with permission ().


Figure 15. The brain vascular stress test to identify the extent of collateralization in the presence of steno‐occlusive vascular disease. Left panel shows angiograms from two patients with near occluded carotid arteries (red arrow). Next to each, are axial slices from the CVR test. The CVR for each voxel is color coded according to the color scheme shown, and mapped onto the corresponding voxel of the anatomical scan. The upper panel shows steal physiology due to lack of collateral blood flow, and the lower one shows near compensatory collateral blood flow. The mechanism is shown in the schematic diagram to the right of each. Upper right panel: (1) inflow resistance which exists by virtue of the conductance of the vascular crown being greater than that of the single arterial stem; (2) the occusion of the branch; (3) the application of hypercapnia; (4) vasodilatory reserve (gray area) is exhausted downstream from the occlusive lesion (indicated by the black lines), but intact on the contralateral side (illustrated by); (5) flow is directed predominantly to the contralateral side, dropping the perfusion pressure at the bifurcation. The reduction in the perfusion pressure, reduces the flow further on the side ipsilateral to the lesion. In the lower right panel: (1) the carotid artery is limiting to the inflow from the central circulation; (2) a restrictive lesion is present in one branch; (3) a rout for collateral flow distal to the lesion capable of providing blood flow; (4) a hypercapnic stimulus is applied; (5) in this case, the collateral flow prevented the encroachment on the vasodilatory reserve (position of black lines relative to gray lines) normally recruited to maintain flow past the upstream obstruction. Thus, the vasodilatory stimulus results in a brisk response from the contralateral, as well as ipsilateral side. Further discussion of the mechanism is in the text. Reproduced with permission ().


Figure 16. The interaction between two vascular regions (e.g., voxels) during a ramp CO2 challenge. An electrical resistance model of two vascular regions. One region is represented by a reference resistance with a preset sigmoidal response to CO2. The other region is represented by a measured flow response to CO2 in a voxel. The model can then be solved to calculate the resistance response sigmoid in the examined voxel. (MAP, mean arterial pressure; Rart, arterial resistance; Pbranch, the perfusion pressure at the branch; R vox, the resistance at the interrogated voxel; Fvox, flow at the interrogated voxel; Rref, the resistance at the reference voxel; Fref, flow at the reference voxel; Ftotal, total flow; Pv, venous pressure). Reproduced with permission ().


Figure 17. Resistance sigmoid measures (left) and their corresponding anatomical maps (right). The vascular model is that in Figure 15 and assumes that this model may be expanded and contracted to contain the blood supply of both hemispheres, from extracranial vessels to capillaries. Thereby, every vessel or vascular bed is in fluid contact with every other vessel. As such, we use the examined unit illustrated as an analogous electrical circuit in Figure 16, to interrogate each voxel as if it is perfused in parallel with an ideal voxel of full vasodilatory responsiveness. Regardless of flow pattern type (see text for definition), the calculation invariably results in a sigmoidal resistance response to PetCO2. Various aspects of the resistance‐PETCO2 relationship are expressed in the sigmoidal shape of the relationship. The graph on the left shows some of these parameters diagrammatically, and the maps on the right show their anatomical distribution. The initial parameter is the resistance in maximum vasoconstriction in hypocapnia, and the amplitude shows the extent of the vasodilation and vasoconstriction. The midpoint of the sigmoid is the PetCO2 at its steepest slope and the range is the span of PETCO2 over which the response can be considered linear. The reserve and sensitivity show the vasodilatory reserve available from the subject's resting baseline PetCO2 and the response sensitivity (sigmoid slope) at that PetCO2.


Figure 18. The relationship between PCO2 and pH with varying [HCO3]. This figure highlights how, according to the Henderson‐Hasselbalch equation, the magnitude of pH changes with varying PCO2 are augmented at altitude following renal compensation for respiratory alkalosis (panel A). At sea level, where [HCO3] approximates 25 meq/L (), an increase in resting PaCO2 from 40 to 41 mmHg would lead to a reduction in pH of 0.01. Conversely, at altitude where [HCO3] may approximate 19 meq/L, an increase in resting PaCO2 from 20 to 21 mmHg would lead to a reduction in pH of 0.02. Panels B and C indicate how the magnitude increase in CBF per increase in PaCO2 is increased following acclimatization (panel B), with the resulting CVR values and their related arterial [HCO3] concentrations in panel C. Panel A was reproduced with permission (). Panels B and C were created from the data of ().


Figure 19. The effect of a progressive loss of vasodilatory reserve on the cerebral blood flow response to PCO2. Results are redrawn from experiments in dogs by Harper and Glass [(); (A), (B), and (C)] showing the effects of reducing perfusion pressure on the CBF response to CO2. (A) Normotensive; (B) hypotensive; (C) extreme hypotension; (D) an overlay of the fitted responses in A and B drawn to the same scale; and (E) the fitted responses presented as the % of maximum vasodilation. Reproduced, with permission, from (). In panels D and E, the solid line represents the normotensive CVR, the dashed line represents the hypotensive CVR, and the dotted line represents the extreme hypotensive CVR. Reproduced with permission ().


Figure 20. Increased blood pressure during hypercapnia is related to cerebrovascular reactivity. Panel A depicts the changes in mean arterial pressure (MAP) that occur during hypercapnia (highlighted by the box). Panel B shows the relationship between changes in MAP from baseline (rest) and changes in internal carotid (ICA) and vertebral artery (VA) flow during hypercapnia. As depicted, there is a linear relationship whereby increases in MAP are related to the increase in flow (ICA and VA) and thus important to consider when measuring cerebrovascular reactivity. Figure reproduced with permission (). © 2012 The Authors. The Journal of Physiology © 2012 The Physiological Society.


Figure 21. Cerebrovascular reactivity at rest and during exercise. Internal carotid artery (ICA) blood flow was collected at rest and during exercise at 35% of maximal aerobic capacity in 32 young healthy individuals (4 females). Reactivity was assessed with a hypercapnic stimulus that was held at steady state for 2 min with dynamic end‐tidal forcing. The magnitude of hypercapnia ranged from +7.5 to +10 mmHg PETCO2 above baseline, and was consistent within individuals from rest to exercise. As is depicted by the mean traces, there is no difference in the slope of the response (i.e., reactivity) between ICA flow and the partial pressure of end‐tidal carbon dioxide (PETCO2) between rest and exercise. Error bars represent standard deviation.


Figure 22. Potential pathways regulating cerebrovascular CO2 reactivity. Hypercapnia leads to an increase in intra‐ and extracellular hydrogen ion concentration ([H+]o and [H+]i, respectively) (). While all the depicted cellular processes act to induce smooth muscle cell relaxation, these processes initiate in various locales. In the vessel lumen, the red blood cell can release adenosine triphosphate (ATP) following deformation (or shear) (), which through binding of purinergic (P2Y2) receptors leads to endothelial nitric oxide synthase (eNOS) activation (), in addition to shear dependent serine threonine specific protein kinase (Akt) activation of eNOS (). Further, acidosis itself stimulates ATP release from red blood cells (). Finally, within the lumen, acidic disproportionation (note: nonenzymatic production) of nitrite (NO2) into NO can occur (). Within endothelial cells (EC), NO production occurs as a result of the aforementioned shear and ATP mechanisms. In addition, activation of phospholipase‐A2 (PLA2) hydrolyzes arachidonic acid (AA) from membrane phospholipids, which is then, via cyclooxygenase, converted into prostaglandins (PGs), with the relevant vasodilator prostaglandins being PGE2 and PGI2. Increased PG synthesis also occurs in astrocytes in response to hypercapnia (). Prostaglandin production leads to concurrent reactive oxygen species (ROS) production () which may influence CO2 reactivity by direct vasodilator effects () or NO scavenging (), although experimental evidence for this is lacking (see text). Also downstream of AA is the production of epoxyeicosatrienoic acids (EET), which convey vasodilatory activity by increasing potassium channel conductance (). At the level of the vascular smooth muscle cells (VSMC), increased adenosine levels () leads to binding of A2A receptors and consequent adenylate cyclase activation. Activation of EP and IP prostaglandin receptors on the VSMC () also contribute to adenylate cyclase activation. Adenylate cyclase activation leads to upregulation of cyclic adenosine monophosphate (cAMP) activity. This cAMP acts via activation of cAMP‐dependent protein kinase, also termed protein kinase‐A (PKA), as well as by increasing potassium (K+) channel conductance (). Soluble guanylate cyclase (sGC) activation occurs downstream of NO production (), from both the endothelial cells (eNOS derived) and from neurons [neuronal NOS (nNOS) derived] (). This leads to upregulation of cyclic guanosine monophosphate (cGMP) and consequent activation of cGMP dependent protein kinase, also termed protein kinase‐G (PKG) (). Cross talk between cGMP and cAMP should also be considered and is influenced by phosphodiesterase (PDE) activity (see text). Both NO () and cGMP () also increase K+ channel conductance. Overall, the increases in K+ channel conductance may lead to VSMC hyperpolarization, inhibiting calcium (Ca2+) influx through voltage dependent Ca2+ channels while both PKA and PKG phosphorylate and deactivate myosin light chain kinase (MLCK) leading to reduced calcium sensitivity of the VSMC and vasodilation (relaxation) ().


Figure 23. Transient increases in shear stress and internal carotid artery vasodilation. (A) Via carbon dioxide manipulation, a transient (∼30 s) increase in shear stress elicits vasodilation of the internal carotid artery. This increase in internal carotid artery (ICA) diameter occurs following return of carbon dioxide and shear to baseline levels, and follows similar temporal pattern to that of a brachial FMD test. () (B) In the transient carbon dioxide test shown on the left, shear stress area under the curve (AUC) was positively correlated to the magnitude of vasodilation (following removal of a statistical outlier—circled). This implicates shear as a potential regulator of cerebral conduit artery vasomotor tone. This represents data recently published by our group. (). Reproduced with permission (). Copyright © 2017 the American Physiological Society.


Figure 24. Cyclic nucleotide cross talk during cAMP and cGMP administration and hypercapnia. Influence of cyclic nucleotide kinase blockers on the pial arteriolar diameter increases accompanying topical applications of the NO donor, SNAP and the adenylate cyclase activator, forskolin (A,B) and exposure to hypercapnia (PaCO2 = 60 mmHg) (C, D). KT‐5720, PKA‐selective inhibitor; KT‐5823, PKG selective inhibitor. As depicted in panel A, administration of the PKA inhibitor KT‐5720 followed by the PKG inhibitor KT‐5823 lead to additive inhibitions of dilation in response to the application of cGMP donor (SNAP) and cAMP donor (Forskolin). The same response was observed when the order of bloackades were switched (panel B). However, the same blockades produced a hypo‐additive reduction in CO2 reactivity indicating an overlap in function, or cross talk as has been referred to thus far in this review. Reproduced with permission ().

 

Teaching Material

R. L. Hoiland, J. A. Fisher, P. N. Ainslie. Regulation of the Cerebral Circulation by Arterial Carbon Dioxide. Compr Physiol 9: 2019, 1099-1152.

Didactic Synopsis

Major Teaching Points:

  • General overview of cerebral blood flow regulation
    • Arterial carbon dioxide and oxygen, blood pressure, cerebral metabolism, and neural activity are all important regulators of cerebral blood flow
  • The influence of changes in arterial carbon dioxide on cerebral blood flow
    • Increases in arterial carbon dioxide increase cerebral blood flow while reductions in arterial carbon dioxide reduce cerebral blood flow
    • The change in cerebral blood flow for a given increase in arterial carbon dioxide (i.e., sensitivity to CO2) is greater than the change in cerebral blood flow observed for the same magnitude reduction in arterial carbon dioxide
  • Standardization and measurement of cerebrovascular CO2 reactivity
    • Standardized vasoactive stimuli – the manipulation of arterial carbon dioxide – are required for the comparison of cerebrovascular reactivity between individuals
    • Standardized measures of reactivity are also required and where possible surrogate measures susceptible to measurement error should be avoided
    • Consideration of intracranial flow dynamics is important to consider for the understanding of cerebrovascular disease in steno-occlusive patients
    • Reactivity to changes in arterial carbon dioxide occur throughout the circulation in large and small vessels
  • What factors affect CO2 reactivity
    • Cerebrovascular reactivity to carbon dioxide is a highly modifiable response, that may be altered by hypoxia, changes in blood pressure, sympathetic nervous activity, age, sex, exercise, and sleep
    • Cerebrovascular reactivity to carbon dioxide appears to be reduced in clinical populations associated with vascular dysfunction
  • What signaling pathways underlie the cerebrovascular response to CO2
    • Multiple signaling molecules, such as nitric oxide, adenosine, reactive oxygen species, prostaglandins and other eicosanoids appear to regulate cerebrovascular reactivity to carbon dioxide in animals, although it is less clear in humans
    • Other factors such as shear stress, and changes in ion channel conductance also appear important in the regulation of cerebrovascular reactivity to carbon dioxide

Interactions between these mechanisms may underlie the current lack of mechanistic clarity derived from studies in humans

Didactic Legends

The figures—in a freely downloadable PowerPoint format—can be found on the Images tab along with the formal legends published in the article. The following legends to the same figures are written to be useful for teaching.

Figure 1 This figure illustrates two separate measurement techniques utilized by Angelo Mosso in the late 1800’s. In individuals with skull defects (Panel A), he was able to use a custom built pressure sensor (plethysmograph) to detect changes in cerebral volume. These changes were inferred to represent changes in cerebral blood flow. Here he measured changes in cerebral blood flow during cognitive tasks such as arithmetic and provided the first evidence that cerebral blood flow increases to match metabolic demands in humans. Later he aimed to measure cerebral blood flow in healthy individuals (Panel B) using a highly precise scale. In instances where cerebral blood flow increased, he theorized he could detect the extra weight of this blood through the use of his scale, thus tipping it in the direction of the head. Collectively these two techniques can be taken as the first attempts to contrive appropriate tools for the measurement of cerebral blood flow. Such attempts to create tools for cerebral blood flow measurement continue today, such as technical improvements to ultrasound and magnetic resonance imaging.

Figure 2 This figure illustrates the progression of knowledge on cerebral blood flow from the 1600’s through to the late 1800’s and the key scientists involved in each advancement of knowledge.

Figure 3 This figure illustrates the nitrous oxide technique for the determination of cerebral blood flow in humans. Requiring 10-minutes of steady state breathing and inhalation of 15% nitrous oxide, this technique relies on the differences between arterial and jugular venous blood throughout the 10-minute inhalation period to calculate cerebral blood flow (detailed in the section “The Kety & Schmidt Technique”).

Figure 4 This figure illustrates the importance of considering the utility of cerebral blood flow measurement techniques. Depicted is how a change in diameter of the middle cerebral artery affects the accuracy of estimating cerebral blood flow from measures of cerebral blood velocity. The values for %disparity between velocity and flow measures are indicated in the plotted squares.

Figure 5 This figure illustrates the measurement of internal carotid artery blood flow with ultrasound. The image in Panel A shows the internal carotid artery with clearly defined vessel walls and a bright Doppler trace that allow for optimal analysis with edge-tracking software. An example analysis output demonstrates changes in flow, velocity, and diameter across the cardiac cycle.

Figure 6 This figure illustrates the effects of changes on haemoglobin concentration and the partial pressure of arterial carbon dioxide on cerebral blood flow as measured with blood oxygen level dependent magnetic resonance imaging. Anemia (green line) increases cerebral blood flow, while increased haemoglobin concentration (blue line) decreases cerebral blood flow. The red dots that intersect the coloured lines and the black dashed lines represent a resting state. Hypercapnia causes the blood oxygen level dependent signal to increase (downward movement on y-axis) which corresponds to increased cerebral blood flow (rightward movement on X-axis). It can be seen by the concavity of the lines that the blood oxygen level dependent response is altered in each state. In this way the changes in multiple parameters can be followed in combinations of changes. For example the effect of a reduction in [Hb] on with a superimposition of hypercapnia on CBF and the net effects of these change on [dOHb] and in BOLD signal can be determined. Starting off at Hb 15 g/DL and normal CBF (red dot at B), progressive anemia results in a shift of the normal relationships of CBF to PCO2 indicated by the red line, to a flattening and shift to the right, indicated by the green line, resulting in a change of CBF to the red dot at C. The OEF increases as autoregulation is insufficient to compensate for DO2 (red dot at C). The [dOHb] increases and BOLD decreases in tandem as shown on vertical axis. Increasing the PCO2, as in a CVR test, changes all the parameters corresponding to the green dot: The CBF is increased, the DO2 increases, the OEF is reduced, the BOLD signal increases.

Figure 7 This figure illustrates the inflow arteries to the brain (internal carotid and vertebral arteries), the major intracranial vessels (anterior, middle and posterior cerebral arteries), pial arteries, penetrating arteries, and capillary network. The potential for collateral flow in instances of impaired perfusion (partial or complete blockage) are highlighted by the colour coding. The circle of Willis provides the first opportunity for collateral flow to the major intracranial arteries, followed subsequently by the highly anastomosed pial artery network. While there are no anastomoses between the penetrating arterioles, the capillary network provides another avenue for collateral flow.

Figure 8 This figure illustrates how collateral flow networks can differ between brain regions due to various arterial network formations and overlap of collateral vessels.

Figure 9 This figure illustrates in a simplistic manner how segmental changes in vascular resistance may influence blood flow and microvascular pressure. Changes in artery size are labelled on each vessel while the resulting directional changes in flow and microvascular pressure are represented below each scenario with arrows. Two arrows simply represent a larger magnitude of effect than one arrow, while a sideways (↔) arrow indicates no change. All flow changes are relative to the very top left scenario and represent blood flow through all the in series vessels of an arterial system (resting large vessel & resting small vessel). A. In a scenario where larger (and upstream) vessels are in a resting state, dilation of smaller (and downstream) vessels will lead to a reduction in microvascular pressure and increase in flow, whereas constriction of smaller vessels will increase microvascular pressure and reduce flow. B. Constriction of a large upstream vessel will lead to a reduction in flow and microvascular pressure when the downstream small arteries do not change in diameter. If the downstream small arteries dilate, this will reduce microvascular pressure to an extent that maintains normal flow. However, if the small vessels constrict in concert with the large vessels, this will reduce flow despite unchanged microvascular pressure. C. Dilation of a large upstream vessel will lead to an increase in flow and microvascular pressure when the downstream small arteries do not change in diameter. If the downstream small arteries dilate in concert with the large arteries, this will increase flow in the face of unchanged microvascular pressure. However, if the small vessels constrict, flow is unaltered and microvascular pressure is increased. Collectively, this figure represents a simplistic overview of how resistance (dilation versus constriction) and driving pressure (effected by changes in microvascular pressure) dictate flow.

Figure 10 This figure illustrates several key regulators of cerebral blood flow. Moving from left to right, the first graph illustrates how a reduction in the partial pressure of arterial oxygen leads to an increase in cerebral blood flow. The next graph again depicts the relationship between oxygen and cerebral blood flow, but rather than the partial pressure, it highlights how arterial oxygen content influences cerebral blood flow. Again, reductions in arterial oxygen content lead to increases in cerebral blood flow; however, the role of increased arterial oxygen content on cerebral blood flow is not as well understood given the difficulties of increasing arterial oxygen content in normal healthy humans. The third graph demonstrates the relationship between the partial pressure of arterial carbon dioxide and cerebral blood flow. Increases in the partial pressure of arterial carbon dioxide lead to increases in cerebral blood flow while decreases in the partial pressure of arterial carbon dioxide lead to decreases in cerebral blood flow. Finally, the fourth graph demonstrates that changes in blood pressure cause relatively proportional changes in cerebral blood flow, whereby reduced pressure leads to reduced flow and vice versa.

Figure 11 This figure illustrates the change from the classical understanding of cerebral autoregulation (left panel) to the contemporary understanding (right panel) where autoregulation maintains constant perfusion over a 5-20mmHg change in arterial blood pressure from resting values.

Figure 12 This figure illustrates how cerebral increases with increases in arterial carbon dioxide and decreases with reductions in arterial carbon dioxide. Further, the change in cerebral blood flow per unit change in arterial carbon dioxide is greater during hypercapnia when arterial carbon dioxide is increased than during hypocapnia when arterial carbon dioxide is decreased.

Figure 13 This figure illustrates changes in cerebral blood vessel diameter at varying segments of the cerebral arterial circulation. Panel A demonstrates the influence of changes in arterial blood gases on pial artery diameter in a feline model. Panel B demonstrates changes in middle cerebral artery diameter with changes in the partial pressure of end tidal carbon dioxide while Panel B shows the change in internal carotid artery diameter with changes in the partial pressure of arterial carbon dioxide.

Figure 14 This figure illustrates how to normalize a CVR test. First (step 1) a standardized stimulus must be utilized; here it is a hypercapnic stimulus with a computerized gas delivery system. The CVR must then be normalized for anatomical location (step 2) followed by voxel based mean and standard deviation calculations of CVR. Following this step, conversion of the CVR values to Z-scores derived from the population data allows determination of how “normal” the reactivity is. Here, purple indicates CVR is more than 2 standard deviations below that which is normal compared to the reference atlas.

Figure 15 This figure illustrates how exhaustion of a vasodilatory reserve at rest, due to carotid stenosis (top left), leads to a reduction in perfusion during hypercapnia due to the vasodilation of an adjacent vascular bed sharing the same source artery (top right). However, under the same stenotic conditions (bottom left) if a collateral vessel exists, perfusion will be maintained to each vascular bed (bottom right).

Figure 16 This figure illustrates a model, whereby the resistance response of one voxel (or vessel) can be calculated when the total flow and of a system and the responsiveness of the other voxel(s) are known.

Figure 17 The left panel depicts the typical changes in cerebrovascular resistance associated with changes in the partial pressure of end-tidal carbon dioxide. It highlights that vasodilatory reserve (denoted as “reserve”) is the difference between resistance at rest and the minimum achievable cerebrovascular resistance. Also noted is the amplitude of the response, simply the extent to which cerebrovascular resistance is altered over a range of carbon dioxide. The related outputs are demonstrated in the images on the right.

Figure 18 This figure illustrates how alterations in acid base balance can influence the relationship between the partial pressure of arterial carbon dioxide (PaCO2) and pH. At altitude, following renal compensation for respiratory alkalosis, bicarbonate ion [HCO3] concentration is reduced, which increases the change in pH that occurs for a given change in PaCO2. This change in the relationship between PaCO2 and pH influences reactivity when indexed against PaCO2 as the cerebral vasomotor response to PaCO2 is truly stimulated by hydrogen ion concentration (as described throughout this review). Thus, when cerebral blood flow is indexed against PaCO2, reactivity is increased at altitude as demonstrated in Panels B and C.

Figure 19 This figure takes the classical data from (174) where reductions in arterial blood pressure lead to reductions in cerebrovascular reactivity to CO2, and depict the data relative to vasodilatory reserve. Panel A depicts reactivity during normotension, panel B depicts a decrease in reactivity during moderate hypotension, and Panel C depicts a complete lack of reactivity during severe hypotension. Overlaying of the responses in Panel D depicts how stark the differences in reactivity are when blood pressure is altered. In panel E, maximum flow is used to represent 100% of the vasodilatory reserve (i.e., maximal vasodilation). Here, when each trace is lined up it is clear that the vasodilatory reserve is significantly reduced at rest (40mmHg PaCO2) when blood pressure is reduced.

Figure 20 This figure illustrates the influence of increases in blood pressure during hypercapnia on cerebrovascular reactivity.

Figure 21 This figure illustrates that when assessed with volumetric measure of cerebral blood flow, cerebrovascular CO2 reactivity is unaltered during exercise at 35% of maximal aerobic capacity.

Figure 22 This figure illustrates the cellular pathways thought to regulate cerebral blood flow during alterations in arterial carbon dioxide. Green arrows represent a subsequent increase or activation, while red arrows represent a subsequent decrease or inactivation.

Figure 23 This figure illustrates the impact of elevations of shear stress on the internal carotid artery. Panel A depicts a representative trace where shear was transiently increased subsequent to hypercapnia induced downstream vasodilation. This increase in shear stress, following the return of carbon dioxide to baseline levels, lead to vasodilation of the internal carotid artery. Panel B depicts the relationship between the magnitude of the shear response and the magnitude of vasodilation, highlighting that they are tightly coupled.

Figure 24 This figure illustrates the cross talk between cAMP and cGMP signalling during hypercapnia. A hypo-additive effect of PKA and PKG blockade indicates an overlap with regard to their function in hypercapnic cerebral vasodilation.

 


Related Articles:

Teaching Material

Contact Editor

Submit a note to the editor about this article by filling in the form below.

* Required Field

How to Cite

Ryan L. Hoiland, Joseph A. Fisher, Philip N. Ainslie. Regulation of the Cerebral Circulation by Arterial Carbon Dioxide. Compr Physiol 2019, 9: 1101-1154. doi: 10.1002/cphy.c180021