Thomas Willis Lecture: Targeting Brain Arterioles for Acute Stroke Treatment

Abstract

Cerebral infarction or ischemic death of brain tissue, most notably neurons, is a primary response to vascular occlusion that if minimized leads to better stroke outcome. However, many cell types are affected in the brain during ischemia and reperfusion (I/R), including vascular cells of the cerebral circulation. Importantly, the structure and function of all brain vascular segments are major determinants of the depth of ischemia during the occlusion, the extent of collateral flow (and therefore amount of potentially salvageable tissue) and the degree of reperfusion. Thus, appropriate function of the cerebral circulation can influence stroke outcome. The brain vasculature is also directly involved in secondary injury to ischemia, including edema, hemorrhage and infarct expansion, and provides a key delivery route for neuroprotective agents. Therefore, the cerebral circulation provides a therapeutic target for multiple aspects of stroke injury, including aiding neuroprotection. Understanding how I/R affect the brain vasculature is key to this therapeutic potential, i.e., vascular protection. This report is focused on regional differences in the cerebral circulation, how I/R differentially affects these segments, and how the response of large vs. small vessels in the brain to I/R can influence stroke outcome. Lastly, how chronic hypertension, a common co-morbidity in stroke patients, affects the brain microvasculature to worsen stroke outcome will be described.

Introduction

Ischemic stroke is one of the most common pathophysiological processes affecting approximately 15 million people worldwide and is on the rise.^1,2 Although the mortality rate for stroke has declined in the last decade, it is still the leading cause of disability at a time when life expectancy has increased.³ Treatment of stroke is difficult due to the rapid death of neurons deprived of oxygen and glucose in the core infarct region. Neuroprotection has been a disappointing pursuit but only targets the penumbra - the peri-infarct region with enough perfusion to remain viable if reperfusion occurs quickly or neuroprotective agents have access. This recognition of the limitation of neuroprotection, and the success of endovascular therapy for large vessel occlusion (LVO) in recent years, has highlighted the importance of understanding cerebral hemodynamics during stroke with more refinement than previously done.

The cerebral circulation is one of the most unique and intricate vascular systems in the body, not surprisingly given the location within the closed space of the skull and the functional importance of the brain. The brain is an organ with high metabolic demands and a need for tight water and ion homeostasis. Unlike other organs where the majority of vascular resistance is held in the small arteries and pre-capillary arterioles, i.e., the “resistance arteries”, the large extracranial and intracranial arteries in the brain contribute significantly (~50%) to cerebrovascular resistance (CVR).⁴ Thus, most vascular segments in the brain are resistance vessels. The contribution of the large extracranial and intracranial arteries to CVR protects downstream vessels from high perfusion pressure and aids in local control of cerebral blood flow (CBF).⁵ The large and small cerebral vessels also contribute to autoregulation of CBF that is pronounced in the brain.⁶ The cerebral endothelium is also unique in that it contains high electrical-resistance tight junctions that limits passage of ions and water, not just protein. This property of the blood-brain barrier opposes water movement into the brain that would otherwise occur due to hydrostatic pressure. The tight control of water passing from the vascular space is critical to prevent vasogenic edema that can increase intracranial pressure and cause severe neurological injury and death. The blood-brain barrier is best developed in the cerebral microcirculation and capillary bed where electrical resistance is highest and numerous transporters are exclusively expressed. These unique features of the cerebral circulation are important for maintaining appropriate blood flow and protecting the brain from edema and hemorrhage under normal conditions. However, the disruption of any one of these functions due to I/R can impact stroke outcome and lead to severe complications and even death. This special report will describe how I/R affect the different vascular segments in the brain, with a special emphasis on vasoactivity and myogenic mechanisms. It is not meant to be comprehensive and is mostly focused on our own lab’s animal studies.

LARGE ARTERY RESPONSE TO I/R

Role of large arteries in CVR and autoregulation of CBF.

As stated above, both the large and small arteries and arterioles in the brain contribute significantly to CVR.⁴ This provides for “segmental vascular resistance” that serves to protect downstream capillaries from high and changing hydrostatic pressure (blood pressure). In addition, both large and small arteries in the brain contribute to autoregulation of CBF, with the contribution of large arteries more prominent than small vessels in the cerebrum.^4–6 This makes understanding how I/R affect the large arteries in the brain important as loss of autoregulation can lead to detrimental consequences including edema and hemorrhage, especially under ischemic conditions.

Restoration of blood flow after short periods of ischemia has been shown to benefit the brain; however, experimental and clinical evidence indicates that reperfusion after longer periods of ischemia may exacerbate tissue injury.^7–10 For example, numerous studies have demonstrated that reperfusion caused edema formation and resulted in irregular patterns of blood flow and microvascular lesions within the reperfused regions.^9–10 In a previous study, we investigated the effect of post-ischemic reperfusion on myogenic tone in isolated and pressurized middle cerebral arteries(MCAs) from rats that underwent transient proximal MCA occlusion (tMCAO).¹¹ We found that MCAs that were ischemic for 2 hours and reperfused for 24 hours had significantly altered reactivity to pressure (myogenic reactivity) and abnormally reduced myogenic tone (calculated from fully relaxed in zero calcium buffer), both important components of CVR and autoregulation of CBF. In contrast, arteries that were reperfused for just 1–2 minutes had well-preserved responses (Figure 1A). This suggested that there was a distinct period of reperfusion in which the cerebral circulation remains viable, thus providing an opportunity for treatment that might preserve these function.

Figure 1. Large artery response to post-ischemic reperfusion.

A) Percent basal tone present in each artery type at an intravascular pressure of 75 mmHg. Arteries from sham control (open bar) and after exposure to 1–2 minutes of reperfusion (gray bar) had a similar level of basal tone. Arteries that were exposed to 24 hours of reperfusion (black bar) had significantly diminished basal tone. **P<0.01 vs. both. Reproduced from Cipolla et al. Stroke 1997;28:176–180 with permission. Copyright American Heart Association 2007. B) Graph showing percent tone of occluded and contralateral MCAs exposed to different periods of reperfusion compared with sham-operated control animals. **P<0.01 contralateral vs ischemic; ‡P<0.05 contralateral vs sham control; ‡‡P<0.01 contralateral vs sham control; §P<0.05 ischemic vs sham control; §§P<0.01 ischemic vs sham control. C) Diameter of MCAs after step increases in intravascular pressure from 75 to 100 mm Hg. MCAs were exposed to 30 minutes of ischemia followed by different periods of reperfusion. D) Graph showing slope of pressure-diameter curves for occluded MCAs exposed to different periods of reperfusion. Arteries that produced a negative slope are considered myogenic (sham-operated control, 30 minutes and 6 hours of reperfusion), whereas arteries that had a positive slope (≥12 hours of reperfusion) had diminished myogenic behavior. Reproduced from Cipolla et. al. Stroke 2002; 33:2094–2099 with permission. Copyright American Heart Association 2002.

In a subsequent study,¹² we induced increasing periods of reperfusion after 2 hours of ischemia in order to determine the threshold duration of reperfusion for myogenic reactivity and tone. The response of the occluded and reperfused MCA was compared to the contralateral MCA to determine if the effect was global, as well as sham controls. We found that myogenic responses of MCAs isolated from the ipsilateral side to occlusion were preserved up to 6 hours of reperfusion, after which myogenic reactivity diminished progressively with longer durations of reperfusion (Figure 1B and 1C). This is shown in Figure 1C in which the slope of the pressure vs. diameter curves are graphed, providing both direction and magnitude of the myogenic response. In arteries that were reperfused ≤6 hours, the slope is small and negative, demonstrating vasoconstriction in response to increased pressure and preserved myogenic reactivity. In contrast, at reperfusion ≥12 hours, the slope is increasingly positive with longer durations of reperfusion, demonstrating greater loss of myogenic reactivity with prolonged reperfusion. Interestingly, myogenic tone in contralateral MCAs was also affected, albeit to a lesser extent, suggesting there may be global circulating factors that contribute to cerebrovascular dysfunction after stroke.

REGIONAL VASCULAR DIFFERENCES

When LVO occurs in the M1 region of the MCA, vascular segments distal to the occlusion are affected, including distal branches of the MCA, and penetrating and parenchymal arterioles (PAs) that perfuse the cortex and striatum. There are several important structural and functional differences between pial arteries and arterioles and PAs that penetrate the brain tissue and have direct connections to capillaries. For example, pial arteries such as the MCA are bathed in cerebrospinal fluid and receive perivascular innervation from the peripheral nervous system, i.e., “extrinsic” innervation. PAs are bathed in interstitial fluid and are “intrinsically” innervated from within the brain neuropil. PAs penetrate into the brain tissue and become almost completely surrounded by astrocytic end feet in some brain regions, but are less covered in other regions.^13,14 While PAs have only one layer of circumferentially oriented smooth muscle, they possess greater basal tone due to smooth muscle that is more depolarized and are unresponsive to some neurotransmitters (e.g., serotonin, norepinephrine), depending on the species.^15–17 The influence of the endothelium on basal tone in PAs appears to involve endothelium-derived hyperpolarization (EDH) as much as nitric oxide (NO).^18,19 We have shown using isolated and pressurized PAs that inhibition of small- and intermediate-conductance calcium-activated potassium (SK and IK) channels, important mediators of EDH, produce as much vasoconstriction in PAs as NOS inhibition with L-NAME (Figure 2A). This suggests EDH is basally active and opposes tone in PAs as much as NO. This is not the case for MCAs in which SK/IK inhibitors have little effect, but NOS inhibition causes substantial vasoconstriction (Figure 2A). The more prominent role of EDH as a vasodilator in PAs was shown previously¹⁸ and appears to be important during stroke. We compared PAs to MCAs isolated and pressurized after I/R and found that vasoconstriction to SK channel inhibition with apamin and IK channel inhibition with TRAM-34 (chlorophenyl)diphenylmethyl]-1H-pyrazole) was preserved in PAs, but constriction to NOS inhibition was largely abolished (Figure 2B). Interestingly, constriction of PAs to SK/IK channel inhibition was enhanced after transient, but not permanent MCAO. This is in contrast to MCAs that have no basal EDH and in which NO vasodilation is also diminished after I/R, i.e., there is no ‘back up’ vasodilator during stroke in MCAs (Figure 2C).

Figure 2. Segmental differences in vascular function and response to I/R between MCAs and PAs.

A) Percent constriction of non-ischemic PAs and MCAs studied isolated and pressurized at 40 mmHg and 75 mmHg, respectively, after cumulative addition of apamin (300 nM) to inhibit SK channels, TRAM-34 (1.0 μmol/L) to inhibit IK channels, and L-NNA (0.1 mmol/L) to inhibit NOS. PAs, but not MCAs, constricted to SK and IK channel inhibition as well as NOS inhibition, demonstrating basal activity that inhibits tone. **P<0.01 vs. PAs. B) Percent constriction to cumulative addition of apamin, TRAM-34, and L-NNA in PAs at 40 mmHg from rats that had MCAO or sham surgery and C) Percent constriction of MCAs at 75 mmHg to the same compounds. Comparisons between sham controls (gray bars) and after ischemia (MCAO) and reperfusion (black bars) are shown. *P<0.05 versus sham; **P<0.01 versus sham. Constriction to SK/IK channel inhibition was preserved in PAs after MCAO, but not NOS inhibition. Reproduced from Cipolla et al. Stroke 2009;40:1451–1457 with permission. Copyright American Heart Association 2009.

PARENCHYMAL ARTERIOLE RESPONSE TO I/R

Role of PAs in incomplete reperfusion.

It is well-established that rapid reperfusion of ischemic brain tissue can limit the size of infarction. However, whether the affected brain region is salvageable or not is both flow- and time-dependent.^20–22At CBF levels of less than ~15 mL/100g/min, oxygen supply cannot support neuronal function and irreversible infarction occurs rapidly within this ischemic core region.²² At CBF levels of ~15–20 mL/100g/min, tissue is within the ischemic penumbra and potentially salvageable if CBF can be restored in a timely manner.^21–23 However, CBF levels are also heterogeneous during reperfusion.^20,23 Shih et al. found that when CBF was decreased 17% of baseline during tMCAO (flow levels considered core infarction) reperfusion restored only 35% of CBF, i.e., there was incomplete reperfusion.²⁰ One of the first studies to show that post-ischemic reperfusion did not completely reperfuse the brain parenchyma was performed by Ames in 1968 who used a rabbit model of global ischemia combined with carbon black staining of the vasculature upon reperfusion. In contrast to controls, post-ischemic brains showed areas of perfusion deficit.²⁴ Interestingly, only the brain parenchyma showed areas of no-reflow and not the pial surface vessels. Subsequent histologic studies suggested that pre-capillary vessel constriction and increased vascular resistance may be responsible for incomplete reperfusion.^25–27 Other studies have also shown microcirculatory disturbances, obstruction and plugging in models of focal ischemia.^10,28–30

A primary mechanism of incomplete reperfusion may be vasoconstriction of PAs and increased vascular resistance in response to post-ischemic reperfusion. PAs bridge two redundant networks of vessels - the communicating arterioles on the pial surface and the microvascular capillary network (Figure 3A).^30,31 In contrast to the substantial interconnections within pial surface and capillary networks, PAs are devoid of anastomoses so that occlusion of a single PA has a devastating effect on downstream flow to the capillaries it supplies.^29,31 Importantly, we have found that unlike large pial arteries that undergo prolonged vasodilation in response to I/R, PAs are hyperconstricted.³² Using a model of transient proximal MCAO (tMCAO), we studied PAs arising from the large pial MCA that penetrated the brain parenchyma after 2 hours ischemia and 30 minutes of reperfusion compared to controls. PAs were mounted on glass cannulas and studied in vitro under pressurized conditions and myogenic reactivity and tone were measured. PAs had increased tone after I/R (Figure 3B), resulting in a significant decrease in diameters (Figure 3C). This was not due to structural remodeling as passive diameters in maximally dilated vessels were similar. That PAs respond to I/R with constriction was also seen in vivo using two-photon microscopy to measure red cell flux and diameters of penetrating arterioles during early post-ischemic reperfusion and found both flux and diameters were decreased below baseline, suggesting vasoconstriction occurred during reperfusion.²⁰ Our findings are also consistent with the observation that small vessel resistance is substantially elevated in vivo following I/R and provide a potential therapeutic target for incomplete reperfusion and improved stroke outcome.^25,26

Figure 3. Effect of early post-ischemic reperfusion on myogenic tone and smooth muscle calcium in PAs.

A) Diagram showing architecture of cerebral pial vessels, PAs, and capillaries. B) Percent tone at 40, 60 and 80 mmHg. C) Arteriolar lumen diameter in response to pressure. Arterioles displayed myogenic vasoconstriction and were smaller actively after middle cerebral artery occlusion (MCAO). D) Percent tone and smooth muscle membrane potential (V_m) of pressurized PAs at 40 mmHg. E) Smooth muscle calcium, measured using Fura 2 at 40 mmHg, was not different in arterioles after MCAO. F) Calcium sensitivity of permeabilized PAs from control (CTL) and after MCAO. Arterioles were more sensitive to calcium after MCAO. *P<0.05 vs control (CTL); **P<0.01 vs CTL. Reproduced from Cipolla et al. Stroke 2014;45:2425–2430 with permission. Copyright American Heart Association 2014.

Mechanisms of increased vasoconstriction of PAs during I/R.

One of the most potent vasoconstrictive stimuli of cerebral arteries and arterioles is intravascular pressure that induces vascular smooth muscle (VSM) membrane depolarization and Ca⁺² channel (Ca_v1.2) activity that increases smooth muscle calcium to cause contraction.³³ However, when membrane potential (V_m) was measured in isolated PAs pressurized to 40 mmHg, PAs after I/R did not have more depolarized VSM, despite increased tone (Figure 3D). We also found that VSM calcium, measured by the calcium indicator Fura 2, was similar in PAs after MCAO (Figure 3E). We then used a permeabilized vessel preparation to measure the sensitivity of the contractile apparatus to calcium. PA smooth muscle was permeabilized with S. aureus α-toxin that creates small pores (2.5 nm) in the plasma membrane that allows ions, but not proteins to pass.³² Using this technique on isolated and pressurized PAs, we found they were also more sensitive to calcium after I/R such that they had greater tone at the same level of calcium (Figure 3F). Thus, early post-ischemic reperfusion causes calcium sensitization of PA smooth muscle that appears to underlie increased tone, demonstrating that I/R increases smooth muscle contractility and promotes vasoconstriction.

Molecular mechanisms of calcium sensitization of smooth muscle after I/R.

While calcium sensitization of PA smooth muscle promotes vasoconstriction in response to I/R, the mechanisms by which this occurs are less clear and may be multifaceted. VSM contraction is regulated by myosin light chain phosphorylation, which is controlled by the balance between myosin light chain kinase (MLCK) and myosin light chain phosphatase (MLCP) activities.³⁴ While MLCK activity depends on calcium-calmodulin and the level of intracellular calcium,³⁴ MLCP activity contributes to contraction of smooth muscle through increased MLC phosphorylation at constant levels of calcium that underlies calcium sensitization.^35,36 Inhibition of MLCP activity involves 2 primary pathways to increase calcium sensitivity. Phosphorylation of myosin phosphatase target subunit 1 (MYPT1) at T⁶⁹⁶ and T⁸⁵³ inhibits MLCP activity.^37–39 Importantly, Rho A kinase (ROCK) phosphorylates MYPT1 at both T⁶⁹⁶ and T⁸⁵³ and suppresses MLCP activity to promote calcium sensitization.³⁹ ROCK is expressed and active in numerous cell types including VSM, endothelium, neurons, glia and immune cells.⁴⁰ Stroke injury is linked to ROCK activation through hemodynamic and microvascular dysfunction as well inflammation and oxidative stress.^41–43 In addition, animal studies have shown that inhibition of ROCK improves stroke outcome in normal as well as diseased animals.^41,42,44 Another pathway for inhibition of MLCP activity is through phosphorylation of PKC-potentiated phosphatase inhibitor protein of 17kDa (CPI-17).⁴⁵ CPI-17 is phosphorylated at T³⁸ by protein kinase C (PKC) leading to MLCP inhibition.⁴⁶ Thus, both ROCK and PKC have important roles in calcium sensitization and are potential therapeutic targets to alleviate incomplete reperfusion.

PIAL COLLATERALS AND STROKE OUTCOME

Collaterals are unique vessels in the brain important for stroke outcome during LVO.

Collateral vessels are arteriole-to-arteriole or artery-to-artery anastomoses that serve to provide retrograde flow during an occlusion and are therefore one of the most important systems capable of mitigating ischemic injury. Collaterals in the brain were described by Heubner in 1874 who termed pial collaterals ‘leptomeningeal anastomoses’ or LMAs.⁴⁷ LMAs connect distal branches of the MCA, anterior (ACA), and posterior cerebral artery (PCA) territories of the pial circulation supplying the cerebral cortex and experience unique hemodynamics. Under physiological conditions, LMAs are highly pressurized but have low flow and low shear stress due to the lack of a pressure differential in adjacent arterial trees, e.g., ACA and MCA.⁴⁸ However, during LVO, a substantial pressure differential is created between the occluded and patent vessels that abruptly increases collateral flow and luminal shear stress in LMAs 15- and 20-fold, respectively.⁴⁸ The increase in LMA flow drives collateral perfusion and is important for salvaging the penumbra.^48–52 Perfusion of the penumbra during LVO has been shown to be dependent on the extent (i.e., number and diameter) of these preexisting vessels, when occlusion is distal to the circle of Willis.^48–52 The extent of collateral flow also serves as a prognostic determinant of stroke severity that guides clinical decision making for treatment.^53–56 In patients with acute ischemic stroke, collateral flow varies inversely with final infarct size.^53,54 In addition, patients with robust collateral circulation on imaging also have improved outcomes after thrombolysis and endovascular therapy, including reduced risk of hemorrhage,^55–58 suggesting robust collaterals also impact outcome from treatment.

LMAs have less basal tone than non-LMAs that is conducive to retrograde flow during LVO.

The most powerful determinant of collateral flow is the diameter of LMAs since resistance to flow varies inversely with radius to the 4^th power.⁵⁹ In the acute phase of stroke, vasoactivity of LMAs could profoundly impact the extent of collateral perfusion and the fate of the penumbra. In a previous study, we examined pressure-induced vasoconstriction (myogenic response) of LMAs, one of the most potent vasoconstrictive stimuli of cerebral arteries and arterioles. LMAs were dissected from rats and studied under pressurized conditions, as we have done with MCAs and PAs. Figure 4A shows a photomicrograph of the pial surface of a rat brain and the distal arterioles of the ACA and MCA. The black circle highlights their anastomosis (LMA). LMAs were mounted on glass cannulas within an arteriograph chamber and studied isolated and pressurized. LMAs were compared to pial arterioles that did not anastomose (non-LMA). Figure 4B shows changes in lumen diameter of LMA and non-LMAs from 18 week old normotensive Wistar Kyoto (WKY) rats in response to increases in intravascular pressure. Notice that non-LMA arterioles were smaller than LMAs and constricted in response to pressure, demonstrating a myogenic response. However, LMAs from normotensive WKY rats were larger and had less myogenic reactivity. In addition, non-LMA arterioles had considerable myogenic tone that increased at higher pressures (Figure 4C). In contrast, LMAs had significantly less tone at all pressures. The larger lumen diameters and less basal myogenic tone of LMAs compared to non-LMAs would be conducive to redirecting flow during an occlusion by decreasing vascular resistance. In contrast to PAs that appear to have basally active SK and IK channels, LMAs did not constrict to the SK channel inhibitor apamin, but had constricted to IK inhibition with TRAM-34 (Figure 4D). Non-LMA arterioles constricted little to either antagonist, suggesting pial arterioles are more similar to pial MCA, and that LMA function is distinct.

Figure 4. Pial collaterals are unique cerebral vessels.

A) Left, photomicrograph showing leptomeningeal arterioles (LMAs, black circle) were identified as connecting distal branches of middle cerebral artery (MCA) and anterior cerebral artery (ACA). Right, LMA isolated, mounted on glass cannulas within an arteriograph chamber and pressurized. Reproduced from Chan et al. Stroke 2016;47:1618–1625; Open access CC BY-NC-ND 4.0 No changes were made. B) Active myogenic vasoconstriction of LMAs and non-LMAs from 18 week old normotensive WKY rats. *P<0.05 vs. LMA. C) Myogenic tone of LMA and non-LMAs from 18 week old WKY rats. *P<0.05 vs. LMA. D) Percent constriction of LMAs and non-LMAs to inhibition of SK and IK channels with apamin and TRAM-34, respectively. *P<0.05 and ** P<0.01 vs non-LMA.

PIAL COLLATERAL FLOW DURING CHRONIC HYPERTENSION

Hypertension increases vasoconstriction of LMAs and impairs collateral flow.

It is well-established in models of chronic hypertension that the brain is more susceptible to ischemic injury than normotensive counterparts. Numerous studies have shown larger infarction in genetic models of chronic hypertension (spontaneously hypertensive rats, SHR) and renal hypertension.^60–63 One of the primary mechanisms by which infarction is larger in hypertension is the relative lack of salvageable tissue, poor collateral status, and rapid evolution of infarct to encompass the penumbra.^60–63 In our previous study on isolated and pressurized LMAs described above, we also compared myogenic responses in LMAs from 18 week old male SHR and found they were highly vasoconstricted – had ~50% basal tone at 40 mmHg and were more similar to non-LMAs.⁶⁴ In addition, LMAs from SHR had decreased dilation to NS309, an activator of SK/IK channels, demonstrating impaired EDH-dependent dilation. In addition, dilation to the NO donor sodium nitroprusside was also significantly impaired in LMAs from SHR. Thus, our data demonstrate significant vascular dysfunction in LMAs from SHR that promotes vasoconstriction. This vasoconstriction likely contributes to poor collateral flow during occlusion, small penumbra, and large infarction in models of chronic hypertension. We also compared 18 vs. 48 week old rats and found no functional differences with age, only hypertension impacted LMA reactivity to the same extent, suggesting it is hypertension that drives collateral perfusion and not age. However, other studies in mice have found age causes rarefaction of collaterals, a result we did not find.⁶⁵ The differences in these studies may be species-dependent and how these particular species age or respond to hypertension.

Role of Ang II in LMA dysfunction and collateral flow during hypertension.

The renin-angiotensin system (RAS) has an important role in the pathophysiology of hypertension. Angiotensin II (Ang II) is the primary hormone of the RAS that is elevated in the circulation and brain in most forms of hypertension and has been implicated as the major cause of oxidative stress, endothelial dysfunction and increased tone in cerebral arteries.^66–69 In a recent study, we showed that treatment of SHR for 5 weeks with the angiotensin converting enzyme (ACE) inhibitor captopril, but not hydralazine, prevented the increased myogenic tone and vasoconstriction in LMAs from SHR.⁷⁰ Both captopril and hydralazine lowered blood pressure similarly in SHR compared to vehicle. ACE inhibition prevents the conversion of Ang I into Ang II, effectively lowering Ang II levels.⁷¹ Importantly, hydralazine lowers blood pressure through smooth muscle vasodilation and produces anti-hypertensive effects without lowering Ang II. The finding that ACE inhibition but not hydralazine prevented LMA vasoconstriction suggests an important role for Ang II in LMA dysfunction during hypertension, independent of lowering blood pressure. These results also may have clinical implications in that the type of anti-hypertensive medication may be important for treating vascular pathology associated with hypertension.

Collateral flow enhancing therapies.

Numerous clinical and experimental studies have focused on enhancement of collateral flow (i.e., collateral therapeutics) during the acute phase of stroke to increase the time-dependent effect of thrombolysis and endovascular therapy for patients outside the window of treatment or who have poor collaterals, respectively. Interventions to enhance collateral flow in the acute stage of stroke include induced hypertension, head-down tilt, volume expansion, partial aortic obstruction, and sphenopalatine ganglion stimulation.^72–75 However, to date, no collateral therapy has proven beneficial in the long-term, likely because of our lack of understanding of the function of LMAs and the heterogeneity of stroke patients. One of the more promising collateral therapies is induced hypertension - when a pressor agent (e.g., phenylephrine) is given during acute stroke to increase cerebral perfusion and enhance collateral flow. Numerous experimental studies have shown that induced hypertension reduces infarction, and several studies showed improved stroke outcome was related to enhanced collateral flow^76–80 and increased cerebral metabolic rate of oxygen (CMRO₂) in the core and penumbra.⁷⁶ The efficacy of this approach is based on the concept that cerebral autoregulation is impaired in the penumbra during LVO due to ischemia that dilates vessels.^81–85 The consequence of this dilatation is that CBF becomes passively dependent on blood pressure and therefore an increase in systemic blood pressure will cause a corresponding increase in perfusion to the ischemic brain. However, none of the previous studies on induced hypertension used animals with co-morbidities that are common in the stroke population. Our finding that LMAs from hypertensive rats are hyperconstricted and respond to pressure with a robust myogenic vasoconstriction suggests this induced hypertension may not work in chronically hypertensive conditions.

In a recent study, we used a cranial window in rats combined with video dimensional analysis to measure diameter changes in LMAs in vivo (Figure 5A).⁷⁴ We found that SHR had a similar level of enhanced basal tone of LMAs in vivo as we found in our in vitro studies (Figure 5B). When MCAO was induced by a remote filament while continuously measuring LMA diameter, we found LMA vasoconstriction in SHR persisted during occlusion, i.e., CBF may not be pressure-passive during LVO in the hypertensive setting (Figure 5C). Surprisingly, LMAs increased tone during reperfusion in both normotensive and hypertensive animals. It is possible that LMAs are more similar to PAs and respond to reperfusion with increased calcium sensitivity and vasoconstriction that contributes to infarct expansion. These findings may also explain why not all patients benefit from therapeutic induced hypertension in the clinical setting.⁸⁵

Figure 5. LMA vasoconstriction in chronic hypertension and with reperfusion.

A) Photomicrograph showing a cranial window (left) that was used to visualize LMAs. Video microscopy (right) combined with video dimensional analysis were used to measure changes in diameter of LMAs continuously. B) Percent tone of LMAs under baseline conditions calculated from passive diameters in EDTA. *P<0.05 vs Wistar. C) Percent tone of LMAs from Wistar and SHR during MCAO before and after phenylephrine, and during reperfusion. Reproduced from Cipolla and Chan Hypertension 2020;1019–1026 with permission. Copyright American Heart Association 2020.

That pial collaterals are vasoconstricted in chronic hypertension suggests that pharmacologically opening them to increase collateral perfusion during LVO may be a viable approach as an adjunct to tPA or endovascular therapy. We have successfully used a plasminogen-activator inhibitor-1 (PAI-1) inhibitor (TM5441) to increase collateral flow during filament occlusion in SHR. TM5441 causes vasodilation of LMAs through enhancing NO production.⁸⁶ This agent was given 30 minutes after occlusion and collateral flow measured using multi-site laser Doppler (Figure 6A). TM5441 increased collateral flow in vivo during MCAO ~40% in young and aged SHR (Figure 6B) that was associated with reduced early infarct (Figure 6C). Similarly, Sanguinate^™, a carboxyhemoglobin gas exchanger also increased collateral flow during MCAO and improved incomplete reperfusion.⁸⁷ One limitation of these studies is that long-term effects have not been measured. Beard et al. found that rapamycin, an inhibitor of mTOR (mammalian target of rapamycin) also increased collateral flow in SHR during MCAO, but there was no improvement of early infarct.⁸⁸ Thus, while promising, collateral therapeutics need further study, including investigating other means of increasing collateral flow, understanding the impact of co-morbidities on LMA function, and potentially adding neuroprotective agents if this vascular route to the penumbra can be enhanced.

Figure 6. Effect of PAI-1 inhibition on pial collaterals.

(A) Diagram showing placement of multi-site laser Doppler probes for measurement of core and collateral cerebral blood flow during MCAO and reperfusion. Shown also are leptomeningeal anastomoses (LMAs, black circles). (B) Change in collateral flow (Probe 2) during MCAO and in response to treatment with TM5441. **P<0.01 vs. untreated by two-way ANOVA with Tukey’s multiple comparisons test. C) TM5441 decreased the acute injury volume in both young and aged SHR. **P<0.01 vs. untreated young SHR; ^##P<0.01 vs. untreated aged SHR by two-way ANOVA with post hoc Tukey’s test. Reproduced from Chan et al. Stroke 2018;49:1969–1976 with permission. Copyright American Heart Association 2018.

CONCLUSION

In the past 20 years, our understanding of the hemodynamics of stroke has become more granular. We now appreciate that there is considerable heterogeneity in the response of different vascular segments to stroke and that co-morbidities such as hypertension, diabetes and hyperlipidemia have profound effects on the cerebral circulation that impacts hemodynamics and stroke outcome. Our finding that PAs uniquely respond to I/R with vasoconstriction - as opposed to vasodilation like large pial vessels – suggests they may be an important contributor to perfusion deficit that limits neuroprotective agents from reaching their target. In addition, recent clinical trials (DAWN and DEFUSE 3) demonstrated major benefit of reperfusion up to 24 hours in patients with salvageable tissue.^89,90 These trials showed that stroke treatment may no longer be time-dependent, but collateral-dependent, and highlights the need to understand collateral flow and the heterogeneity in stroke patients. Our findings that pial collaterals constrict in response to post-ischemic reperfusion and are highly vasoconstricted in models of hypertension provide opportunities for treatments that can effectively target these vessels and their underlying vascular dysfunction. However, effective treatment requires an understanding of the vascular biology of these vessels under normal and diseased conditions and the consideration of “vascular protection” as well as “neuroprotection” for acute ischemic stroke.

Abbreviations

CVR

cerebrovascular resistance

EDH

endothelium-derived hyperpolarization

IK

intermediate-conductance calcium-activated potassium channel

I/R

ischemia and reperfusion

LMA

leptomeningeal anastomoses

LVO

large vessel occlusion

MCA

middle cerebral artery

MCAO

middle cerebral artery occlusion

MLCK

myosin light chain kinase

MLCP

myosin light chain phosphatase

PA

parenchymal arteriole

PAI-1

plasminogen-activator inhibitor-1

PKC

protein kinase C

RAS

renin angiotensin system

ROCK

Rho A kinase

SHR

spontaneously hypertensive rat

SK

small-conductance calcium-activated potassium channel

VSM

vascular smooth muscle

WKY

Wistar Kyoto