Endothelial energy failure as a therapeutic target in elderly strokes

Abstract

Despite the fact that a third of stroke patients are over the age of 80, specialized stroke management in the elderly is in its infancy. The signature poor blood flow in the aged brain has been shown to synchronize with impaired neurovascular coupling and inflammation in the course of acute occlusive stroke. The striking pre-clinical findings indicate endothelial energy failure precedes cerebrovascular aging and may bridge the aging phenotype of cerebral blood flow to cerebrovascular dysfunction in the elderly. As the gatekeeper for energy supply and shuttling to different brain resident cells, the endothelial energy machinery affects a wide range of cerebral activities. The review focuses on the existing evidence linking endothelial energy failure with the detrimental events in the acute phase of ischemic stroke, including collateral failure, excitotoxicity, and fluid retention. This simplified picture aims to highlight the knowledge gaps and potential targets to correct the bioenergetic imbalance in the elderly stroke. The recent clinical studies demonstrating promising effects of mitotherapy in old stroke patients further signify the correction of energy machinery cerebral vessels, encouraging meticulous characterization of the old endothelium and bioengineering approaches for specialized drug delivery.

Introduction

Aging is known as an irreversible risk factor for stroke occurrence. Stroke incidence is higher in older adults, particularly women, and in socially deprived areas, ¹ and is projected to increase in both sexes and all age groups till 2030. ² Aging, the loss of biological hemostasis to adapt to stress, is characterized by alterations in cellular metabolic processes, DNA damage, and inflammation. This process is more significant in tissues with high metabolic rate, where the progressive accumulation of cellular wastes impairs normal bioenergetic and encourages irreversible cell cycle arrest. ³ Metabolically, the brain is among the most active organs, demanding high amounts of energy mainly to maintain ion homeostasis. Dysregulated bioenergetics in the aged brain associates with eroded oxidative phosphorylation (oxPHOS), which shifts mitochondria to produce more free oxygen radicals at the cost of reduced ATP generation, ⁴ rendering the brain more vulnerable to further energy crises during ischemia.

Perturbation in cerebrovascular bioenergetic, either by aging or during ischemia, has been closely linked to reduced immune system hemostasis with aging, and inflammaging.^5,6 The endothelium layer, forming the immediate brain interface either to touch the inflamed blood or to sense the ischemic injury, may intimately dictate brain energy status. Beyond controlling brain-to-blood communication as the chief component in the blood–brain barrier (BBB), the endothelial cell layer is a key regulator of cerebrovascular aging, ⁶ a well-known risk factor for adverse stroke injury and complications. These particular features place brain endothelial cells (bECs) in the spotlight in aging and stroke research. However, while ischemic stroke represents severe metabolic stress, research has been more focused on neuroprotective/restorative strategies that target pathological consequences rather than the maturation of the initial energy crisis. This review aims to picture the hypothetical link between endothelial energy derangement and aggravated acute stroke injury in an aged brain. Considering the recent evidence indicating metabolic perturbation precedes aging in the cerebral vasculature,^7,8 this review will initially discuss the reliance of cerebrovascular aging on endothelial mitochondrial behavior. We will then focus on the detailed link between bECs metabolic aging and the particular features of stroke injury in the elderly.

Acute stroke in the elderly

Old age is an established predictive factor for poor stroke outcomes. Patients ⩾75 years old are more likely to present with higher admission NIHSS scores ⁹ and with a higher propensity to death after stroke. ¹⁰ Regarding the low incidence of young strokes, ⁹ clinical research efforts to address the impact of aging by comparing old versus very elderly patients rather than young individuals. Accordingly, the very elderly (age ⩽85) present with more severe strokes (NIHSS score ⩾11) compared with the younger age group (age 65–84; p < 0.05). ¹⁰ Case fatality also shows a close positive correlation with age: 5.7% (age <59), 8.6% (age 60–69), 13.4% (age 70–79), and 24.2% (age ⩾80; very old patients; P < 0.001), according to a multicenter cohort study. ¹¹ The poor outcomes in the elderly have been ascribed to both escalated stroke injuries and reduced recovery capacity in the elderly. However, it is important to note that after adjustment to admission scores, age has been shown to explain merely 1%–3% of the variation in functional outcome at discharge. This implies that more stroke severity, rather than inefficient recovery, may better explain high mortality and morbidity in the very old patients. ¹⁰

With the emerging hypothesis that stroke induces discrete pathological pathways in the aged brain, research is increasingly focusing on differential features in the aged brain. Brain gray and white matter atrophy starting at middle age (age <30, 0.1%–0.5%/year), is followed by progressive repression in energy metabolism. Magnetic resonance spectroscopy in healthy subjects has shown neuronal mitochondrial metabolism and glutamate–glutamine cycle flux is ∼30% lower in elderly subjects (76 ± 8 vs 26 ± 7 years old). ¹² This may conceivably reprogram a wide array of energy-dependent tasks in the brain. The hindered bioenergetic in aged mice brains has been recently shown to associate with alteration of Ca²⁺ binding sites on mitochondrial glutamate/aspartate carriers. ¹³ Coupled with the well-documented Na⁺/K⁺ ATPase downregulation in aged rats’ brains,^14,15 this may compromise the ability of neurons to maintain membrane properties. Interestingly, this may picture a hypothetical link between the disrupted bioenergy in the aged brain and the clinical findings demonstrating that the very old patients are more vulnerable to the detrimental consequences of spreading depolarization and are prone to new-onset epilepsy following stroke.

In fact, the hampered brain energy metabolism may explain typical manifestations in elderly strokes. The metabolic status of the brain is a master regulator of cerebral blood flow (CBF). ¹⁶ Aging-associated hypoperfusion followed by impaired vascular reactivity further restricts blood flow to the ischemic core while compromising collateral perfusion, a critical determinant of ultimate outcomes following stroke. ¹⁶ More recently, primate studies have shown that non-ischemic reduced blood flow to particular brain regions may also detrimentally affect BBB barrier properties.^17,18 The gradual increase in human BBB permeability with normal aging has been proven by dynamic contrast-enhanced magnetic resonance imaging (DCE–MRI).^19,20 The aging-induced BBB leakage is subtle, though, ²¹ it is discernible enough to multiply the stroke-induced oxidative damage to BBB and the massive immune cell recruitment aided by inflammaging. ²² The cerebral vasculature plays a key role in translating the aging-induced systemic metabolic derangement to aggravated stroke outcomes. In the following sections, this review will further focus on cerebrovascular and, particularly, on the endothelial compartment.

Aging in the brain capillaries

Brain microvessels play a significant role in vascular resistance with aging. ²³ As the brain shrinks and cerebral vascular density decreases with aging, the penetrating microvessels and capillaries undergo modifications at structural and subcellular levels. The first electron microscopy images obtained from human biopsy samples showed a remarkable thinning in white matter capillaries, mostly due to the thinning of the endothelial cells and the loss of pericytes. ²⁴ Age-derived degeneration in the cerebral vasculature has been further characterized in animals, where a thickening of the basal lamina, loss or transformation of endothelial cells, and reduced number and membrane potential of mitochondria are among the significant manifestations.^25,26 Greatly described in detail elsewhere ²⁷ as bECs age, they shift towards the peripheral phenotype where they lose junctions, shed glycocalyx, and do not need plenty of mitochondria to meet the high energy demand in the brain.

The biochemical findings obtained from human samples have recently reframed our understanding of cerebrovascular aging. Examination of isolated human cerebrovascular fraction by tandem mass spectrometry has demonstrated the profound effect of aging on 150 proteins in the brain capillaries, mainly involved in chemokines, HIF-1α, and synaptogenesis signaling. ²⁸ The brain capillary endothelial cells appear to be a preferential target for aging signals. Compared to arterial and venous endothelium, these cells undergo the most transcriptional changes with aging, upregulating oxidative stress and innate immunity pathways. ²⁹ The significance of age-induced alterations in the brain endothelial transcriptional profile may be best exemplified by focal upregulation of vascular cell adhesion molecule 1 (VCAM-1) in aged mice, which plays a prominent role in facilitating vascular-immune cell interactions. ³⁰ The cerebral endothelium is hypothesized to be a dynamic regulator of brain energy metabolism. The bioenergetic communication between different cells within the neurovascular unit (NVU) is further gated by the aged endothelium. The diminished bECs’ capability to transport energetic substrates, coupled with the age-related metabolic shift, may explain the disturbed energy metabolism in the whole NVU, ³¹ and is discussed later in detail.

Endothelial bioenergetic to bridge hemodynamic to cerebrovascular aging

Endothelial aging originates from a myriad of systemic hemodynamic and inflammatory signals, leading to a time-dependent accumulation of DNA damage and irreversible cell cycle arrest. As precisely reviewed elsewhere, ³² the “aging stimuli” converge on a complex molecular pathway that ultimately alters energy-sensing signals (e.g. Sirtuins and AMPK) to activate tumor suppressor pathways (tumor antigen p53–cyclin-dependent kinase inhibitor 1A (CDKN1A or P21) and retinoblastoma-associated protein (pRb)–cyclin-dependent kinase inhibitor 2A (CDKN2A or P16). The activations of these intermediate effectors are interrelated and context-sensitive.^33,34 Emerging evidence supports the striking hypothesis that hemodynamic alterations following age-associated remodeling of the brain vasculature play a critical role in endothelial cells’ metabolism and, thereby, the bECs’ responsiveness to the aging signals.

Young human and rodent endothelial cells in the periphery highly rely on glycolysis rather than oxidative phosphorylation for ATP production, making the role of mitochondrial respiration insignificant in the periphery. The endothelium across the CNS tends to adapt more aerobic respiration to meet the high energy demand for multiple transport systems across the BBB. This particular metabolic transformation is intimately regulated by the pulsative/laminar shear stress within the brain capillaries. The reliance of endothelial metabolism on the fluidic behavior of CBF is well supported by in vitro findings demonstrating that laminar shear stress induces several enzymes in the tricarboxylic acid (TCA) cycle and p450 complex, while downregulating glycolytic genes in human brain microvascular endothelium. ³⁵ This is in line with the observations in endothelial cells with peripheral vessel origin (e.g. porcine aorta, HUVEC), indicating that laminar shear stress increases the mitochondrial membrane potential, ³⁶ enhances the overall mitochondrial ATP generation and release, through inducing Krebs cycle (oxPHOS).^37,38

This endothelial metabolic shift from glycolysis to oxidative phosphorylation has been attributed to mechano-transduction, which activates flow-sensitive transcription factor Krüppel-like factor (KLF) family and AMP-activated protein kinase (AMPK) ³⁹ as well as hypoxic inducible factor-α (HIF-1α). ⁴⁰ In adult healthy cerebral vasculature, the reduced glycolysis downregulates eNOS-degrading enzymes. This has been defined as the principal mechanism by which laminar flow improves NO bioavailability and vasorelaxation. ⁴¹

With healthy aging, cerebral vasculature is exposed to less laminar shear stress. Beyond the growing cerebral hypoperfusion (0.2 ml/100 g/year), the dramatic change in cerebrovascular architecture is also responsible for the altered blood fluidic behavior. Deterred laminar flow, combined with arterial stiffness, reduced branching, and increased tortuosity within the aged brain vasculature, exposes the endothelial cells to sustained oscillatory shear stress. ⁴² This lingering pattern may reprogram bECs metabolism. The existing knowledge on the bioenergetics in bECs concludes a remarkable repression in both mitochondrial respiration and glycolytic energy metabolism with aging. While the age-associated decline in mechano-transduction ⁴² may affect the bECs’ perception of fluidic force, the sustained impairment in laminar shear stress may theoretically preserve ATP synthesis through stimulating the glycolytic pathway. However, the characteristic fall in glucose uptake in early aging^43,44 prevents the hemostatic metabolic shift in response to turbulent flow in an aged brain. Interestingly, it has been shown that sustained impairment in the laminar flow within the aged cerebral vasculature plays a prominent role in the reduction of endothelial glucose transporters.^35,45 This further underlines the hypothesis that the disturbed laminar flow is accountable for driving the deterred energy metabolism in old cerebral vessels (Figure 1). The phenomenal reduction in eNOS cofactor NAD⁺ with aging and the consequent “eNOS uncoupling” will also reduce NO bioavailability. This not only causes a drastic impairment in vasogenic responses to adopt the oscillatory blood flow, but further impairs key molecules in mitochondrial biogenesis and hemostasis (e.g. PGC1α).^46–48

Figure 1.

Cerebrovascular remodeling with aging alters endothelial cells’ energy metabolism. In an aged brain, the microvessels undergo pericyte shedding and mitochondrial loss in the thinning endothelium. This is preceded by adopting less elasticity and more tortuosity in the whole cerebrovascular tree, starting to impair normal pulsation and laminar flow. Laminar shear stress contributes to normal glucose influx into the endothelium. In the meantime, it relays metabolic commands to increase ATP generation by upregulating electron transport chain components (complexes I–IV) in the inner mitochondrial membrane. In normal vessels exposed to oscillation, particularly in the branching sites, the healthy mechanotransduction and its main effectors, KLF and AMPK, stimulate glycolysis to prevent ATP shortage. As the cerebral vessel endothelium ages, it loses functional mechanotransduction and adopts less glucose uptake/glycolytic capacity, leading to overall ATP reduction. By liberating more pyruvate dehydrogenase, the aging cell forces more pyruvate into the citric acid cycle and oxPHOS to build up ATP availability. However, the typical reduction in NAD and ETC complexes with aging leads to inefficient citrate metabolism, followed by the elevated proton gradient, which may not flow to ATP synthesis. Ultimately, the shortage in the ubiquinone pool leads to the backward flow of electrons and incomplete reduction of molecular oxygen, resulting in ROS formation.

The mitochondrial dysfunction in aged endothelium was initially identified in rat muscle vasculature by the Csiszar research team (2008), and demonstrated a significant decline in expression of complex I, III, and IV (24 vs 3 months). ⁴⁶ Whether this change is a consequence or the inducer of aging signals is unclear. In this line, the recent findings on human brain microvasculature endothelial cells show that aberrant energy metabolism precedes the induction of cell cycle arrest and endothelial aging. ⁸ Accordingly, in the pre-senescence stage, oxPHOS is the dominant source for ATP production, after which both glycolysis and oxPHOS are downregulated in aged HBECs, resulting in reduced ATP production and oxygen consumption rate. ⁸ This might be explained by the experimental findings detecting a significant decline in pyruvate dehydrogenase kinase, which will elevate pyruvate entry to the Krebs cycle. ⁴⁹ Consistently, in vivo findings in the mouse brain microvessels have also indicated a remarkable decrease in basal and maximal mitochondrial respiration in the early stages of aging (18 vs 3 months). ⁵⁰ In the aging mice cerebral microvessels (21 ± 1 vs 3 ± 1 months), the overall ATP is mostly generated by oxPHOS rather than glycolysis, both of which show a remarkable decline, to as low as 61% of the basal level in the young brain. ⁵¹ According to this study, the reduction in citrate synthase activity, the rate-limiting enzyme for acetyl CoA incorporation to oxPHOS, forces aged cerebral microvessels to utilize more glutamine to fuel the remaining oxPHOS rather than glucose or fatty acids. While the ex vivo analysis of aged versus young cerebral vasculature concludes an obvious decline in mitochondrial biogenesis, studies of replicative senescence in HUVECs have reported conflicting results in vitro, some of those demonstrating improved ATP biogenesis either through glycolysis or oxPHOS with aging. ⁴⁹ Still, however, regarding the substantial differences between the aging models, specific cell types, and unclear pre-senescence/senescence matching these data may not critically overshadow each other.

There are several gaps in our knowledge of the metabolic shifts in endothelial senescence. A handful of molecules have been identified to contribute to mitochondrial dysfunction with aging. ⁵² Among the potential pathways, the age-associated fall in nuclear factor (erythroid-derived 2)-like 2 (NRF2), ⁵³ CREB-regulated transcription co-activator 1 (CRTC1), ⁵⁴ and peroxisome proliferator-activated receptor gamma coactivator 1 (PGC1α) ⁵⁵ are well-documented to link aging with mitochondrial failure. Intriguingly, the recent evidence implies that laminar shear stress regulates both PGC1α ⁵⁶ and NRF2 ⁵⁷ (Figure 2).

Figure 2.

Endothelial cells’ bioenergetics as a potential bridge to link altered shear stress to senescent phenotype. At the pre-senescence stage, endothelial cells start to produce less ATP at the cost of more free radicals from the derailed electrons in the mitochondrial membrane. The consequent rise in AMP/ATP triggers AMPK to restrict energy-consuming pathways in cell growth and proliferation through recruiting P21 and P16. The gradual buildup in ROS alters the structural and enzymatic activity of proteins participating in NOS coupling and NAD hemostasis. The signature fall in NO and NAD significantly contributes to the parallel SIRT downregulation with aging, which then aggregates mitochondrial biogenesis frailty. As SIRT activity falls below the biological level, it fails to protect the endothelium from DNA destabilization and inflammatory transformation in response to age-associated stimuli, leading to increased SASP in endothelial cells. Dotted arrows represent inhibitory effects.

Endothelial bioenergetic and stroke manifestations in the elderly

As described above, the impaired energy synthesis in bECs encourages cerebrovascular aging and thereby may increase the propensity to stroke incidence. Upon the focal occlusion in the cerebral vasculature, the preexisting ATP shortage will imperatively contribute to aggravating the injury caused by the stroke energy crisis. Part of the escalation might be simply explained by the detrimental shrinkage of ATP-dependent hemostasis, namely the membrane ion gradient and translation machinery, with immediate impact on endothelial integrity. However, given the central role of BBB in acute stroke injury, such frailty shadows almost all the CNS residents. The following sections illustrate the basis of the hypothesis that this bioenergetic failure in bECs explains the unique features of stroke in very old patients, including but not limited to: escalated tissue injury, hampered diffusion/perfusion, and immune responses. To cover this particular concept, such discussions will focus on the contribution of impaired energy metabolism (ATP shortage), rather than the multifaceted consequences of mitochondrial dysfunction (inflammation, apoptosis, etc.) or cerebrovascular senescence. In the following discussions, anywhere the studies indicate a causative link, we expanded experimental details to reflect the strength of the evidence. Otherwise, a brief description has been prioritized, mainly to address associations or the general concepts reviewed elsewhere.

Impaired energy hemostasis and escalated tissue injury

Endothelial cells are the first CNS cells sensing the ischemia signals, simultaneously with the circulating cells entrapped in the occlusion site. The consequent ATP fall and the consequent impaired ion hemostasis are known as the principal effectors to induce excitotoxicity and cell death. As fully described elsewhere, ⁵⁸ several types of cell death, ranging from apoptosis to oxytosis, have been defined to shape the molecular pathways and subcellular modifications entailing ischemia. Nonetheless, given that definitions rely on normal ion hemostasis and juvenile bioenergetics, our understanding of stroke-associated cell death cascades may need in-depth revision.

In cerebral vasculature endothelium, apoptosis and autophagy are among the most characterized types of cell death following stroke. ⁵⁹ However, experimental findings show endothelial layer comprises the most resilient entity in the brain in response to ischemic injury in vitro^60,61 and in vivo.^62,63 That is, in a mouse model of permanent stroke, when almost 100% of neural cells are lost in the ischemic core, endothelial cells do not show significant injury; however, they lose GLUT-1 transporters. ⁶³ This supports the hypothesis that the contribution of endothelial-derived damage-associated molecules (DAMPs) to stroke-induced injury is minimal.

Endothelium as the entry port for nutrients

The exceptional resilience of the endothelium to ischemia renders this monolayer a reliable safeguard for other endangered, more vulnerable NVU cells. The brain endothelium contributes to NVU energy homeostasis by importing required substrates (glucose, fatty acids, lactate, and ketone bodies). As described elsewhere, the brain cells may adopt optimal metabolizing machinery to meet particular morphology/activity in health and disease. ⁶⁴ In a simplified scenario, glucose passes BBB through saturable GLUT1 transporters on the endothelial cells. The interstitial glucose then passes the neural membrane through GLUT-3 and sodium-glucose cotransporter 1 (SGLT1). ⁶⁵ The interstitial glucose readily fluxes into astrocytes’ end feet surrounding capillaries through GLUT1, and enters glycolysis followed by aerobic ATP generation in oxPHOS. Exceptionally, in astrocytes, part of the glucose entering anaerobic glycolysis (Glucose-6-p) converts to glycogen, serving as the only form of energy reserve in the brain. After undergoing glycolysis, the astrocytic glucose metabolite may also generate lactate, which is then carried to adjacent neurons through monocarboxylate transporters (MCTs). Fatty acid metabolites represent another major source of energy to NVU cells. The fatty acid dissociated from lipoproteins by the endothelial-associated lipase may enter the endothelial layer through passive, caveola-mediated, or transporter-mediated transcytosis, typically major facilitator superfamily domain-containing 2A (MfsdC2A). While in the cytosol, fatty acids bind to fatty acid binding protein (FABP), followed by either intercellular metabolism or release to the interstitial fluid. Upon entry into the Krebs cycle, either in astrocytes or oligodendrocytes, fatty acids are metabolized to produce keto acids (via β-oxidation). Neural cells can use keto acids to produce more ATP after taking them through the same MCT transporters ⁶⁶ (Figure 3).

Figure 3.

The entry port and the shuttling system for nutrients in the neurovascular unit. In the human brain, glucose transporters play a pivotal role in neural cells. Glucose enters the brain mainly through GLUT-1 on the luminal membrane of endothelial cells. Cytosolic glucose may either (1) flux into glycolysis and oxPHOS to generate ATP or (2) flow to extracellular space where it is taken by GLUT-1 in the surrounding cells (pericytes and astrocyte end-feet) or via GLUT-3 and SGLT-1 in the distant neurons. After entering the astrocytes, part of the glucose is stored as intracellular glycogen before processing to pyruvate. The astrocytic pyruvate produced after glucose metabolism is also converted to lactate, which flows into the extracellular space and enters neuronal cytosol via MCT-2 channels. Through this “lactate shuttle,” astrocytes provide neurons with an alternate source of energy in the status of glucose shortage (e.g. vascular occlusion) or lactic acidosis (e.g. fasting). Astrocytes also support neurons’ viability by manufacturing fatty acids into keto acids and delivering them through the same lactate shuttle. Although some saturated fatty acids can be synthesized within the brain, the brain endothelium is the main gate for the entry of polyunsaturated fatty acids. The brain endothelial cells process circulating lipids to simple fatty acids for uptake through caveolae or MfsdC2A transporters, followed by intracellular transfer by FABP. Recent findings have identified long tunneling nanotubes between astrocytes and neural/endothelial cells. This shuttling system is specialized for the exchange of cytosolic materials and directly transfers ATP or mitochondria from endothelium to astrocytes, or from astrocytes to the dying neurons in conditions associated with cerebral artery occlusion. In this condition, the fatty acid metabolism in the adult oligodendrocytes in the white matter provides the endangered axons with ketone bodies/keto acids, as an alternate energy source, while transferring the astrocytic lactones to the neural cells.

As endothelial cells uptake and pass nutrients to the NVU cells, they also provide ready-to-use ATP. Intercellular transfer of ATP or mitochondria in the brain may occur either through exosome-like vesicles or connecting nanotubes. Tunneling nanotubes (TNT), the membrane-enclosed protrusions that connect cells over long distances, have been recently explored as a tool in the functional dialog between NVU cells.^67,68 Therefore, when adequately supplied by the BBB endothelium, NVU cells may exchange energy substrates and act as a metabolic reservoir during ischemic energy crises. In support of this, recent findings from live time-lapse fluorescence microscopy have shown that astrocytes receive functional mitochondria from both pericytes and endothelial cells in the course of ischemic–reperfusion injury. ⁶⁷ Accordingly, Benfenati et al. (2022) have demonstrated active transport of functional mitochondria from endothelial cells either in cocultures or in three-dimensional assembloids. ⁶⁷ Given that prior studies have underlined astrocytic TNT-mediated mitochondrial transfer as a key mechanism to preserve ischemic neural cells,^69,70 TNTs may hypothetically “transplant” the endothelial mitochondria to the remote endangered neurons. Along with the general age-associated decline in physiological reservoirs, both the astrocyte–neuron “lactate shuttle” (ketone/keto acid shuttle) and the initial glucose supply by cerebral vasculature are diminished, at least partly due to the significant downregulation in GLUT-1 or MCTs.^44,71

White matter injury and brain infarction

Elderly strokes manifest with remarkable prevalence of white matter injury, which parallels small vessel disease and “silent strokes” with aging. The injury to the white matter, which communicates messages between several cortical and deep-brain regions, is not only a strong predictor of dementia, but also high mortality and poor motor function following stroke. Myelin sheaths covering axons in the white matter restrict axons to the metabolites supplied through oligodendrocytes. Adult oligodendrocytes with fewer and more fragmented mitochondria mostly rely on glycolysis. Recent findings conclude a lower level of lactate dehydrogenase (LDH) expression in adult oligodendrocytes. ⁷² These findings suggest the lactate shuttle in the white matter relies on lactate generation in astrocytes, rather than in oligodendrocytes, while highlighting the significance of fatty acid metabolism in oligodendrocytes as a direct metabolic support to axons via non-compacted cytoplasmic channels in myelin. ⁶⁶ The significance of oligodendrocytes in supplying lactate or ketone/keto acids to the axons, however, may not suffice the axon’s energy demands in the aging brain. As the neural cells age, the metabolism shifts away from these substrates and becomes more glycolytic.^73,74 This highlights the hypothesis that glucose influx from the endothelium into the brain parenchyma, or the glycogen mobilization by astrocytes, plays a more prominent role in neural cells’ resilience. Given the reduced glucose influx and astrocytic glycogen in the aged brain, this may explain the white matter susceptibility to stroke injury, at least partly. As discussed in the following sections, myelin dysfunction and inflammatory responses, along with impaired interstitial fluid movement along myelin sheaths, represent additional contributors to stroke injury in the aged white matter. According to the published experimental studies (Supplementary Tables 1 and 2), the effect of aging on ischemic lesion size in animal models is context-sensitive and, in some cases, contradictory. The majority of this seeming controversy is raised from differences in aging models. While all studies consistently report escalated neurological deficit scores, the age effect on lesion size strictly depends on the strain, stroke model, and the study timeline.

However, these findings may explain escalated ischemic injury in the elderly; this age-effect tends to manifest predominantly in larger brains (e.g. rats vs mice)^75,76 or longer occlusion times, ⁷⁷ in the experimental setting. It is important to note that to compare the stroke outcomes, most studies rely on the infarct size, reflecting a complex development of immune response and glial scar, not exclusively indicating ischemia-induced cell death. These findings also highlight the notion that, with the gradual progress of senescence over a lifespan, ATP shortage may also precondition the cells, mimicking the protective effects of caloric restriction. In this scenario, immediately after sensing ischemia, the preconditioned aged ECs may resist the acute oxygen shortage, either by shifting to anaerobic glycolysis or ketolysis. This may partly explain the reduced or the same lesion size in the hyperacute phase of stroke in aged, smaller rodents.^78,79

Hypothetically, ROS upregulation, ATP shortage, and the unleashed immune responses with aging may become more strengthened in larger lesions, where “more cells, more molecules” are recruited to potentiate the detrimental cascades. Otherwise, the preconditioning/hormetic effect may outweigh and explain seemingly contrasting, yet intriguing findings, indicating aging provides tolerance to neurological and histological deficits when the occlusion time is <30 min ⁸⁰ or in some hemorrhagic models. ⁸¹ Collectively, aging is an absolute predictor of adverse neurological deficits aligned with incompetent recovery mechanisms. However, the immediate impact of aging on cell death type and timing is an area for further investigation.

Hampered perfusion and fluid turnover

Sustainability of penumbral tissue critically depends on reservoirs of oxygen and nutrients, as well as the timely removal of cellular damage fragments. This requires adequate blood supply along with extracellular fluid turnover, each of which is minutely regulated through distinct pathways. ⁸² Cerebral hemodynamic reserve, the capacity of the brain to increase cerebral blood volume to maintain a constant regional CBF, is dramatically influenced by aging. ⁸³ The anatomic characteristics of intracranial collaterals, cross-connecting the outer branches of adjacent arterial trees, closely affect cerebrovascular reserve during occlusive events. ⁸⁴ Age-associated remodeling of the cerebrovascular network, characterized by reduced collateral diameter, followed by loss of collaterals ⁸⁵ ; substantially hampers the ability of collateral perfusion to sustain the ischemic penumbra. Beyond the anatomy, there is a strong age effect on collaterals’ functionality following stroke. However, the age effect is not always detectable on the pial blood flow post-stroke, ⁸⁶ research findings confirm its strong effect on overall cerebrovascular reactivity, ⁸³ worsening vascular rarefaction and no-reflow phenomena. ⁶

Restricted cerebral perfusion reserve

The age effect on cerebral perfusion reserve is a multifactorial effect of physiological alteration in the brain and periphery, leading to cerebral blood flow dysregulation. This is well exemplified by an increased risk of postural hypotension, lightheadedness, and cerebral ischemia in severe cases, as a consequence of diminished baroreceptor reflexes and blunted autoregulatory responses in aged cerebral resistance arteries.

The remarkable bioenergetic derangement in bECs and consequently in other NVU cells particularly astrocytes, contributes to the inefficiency of CBF control during occlusive stroke. The occlusion of a cerebral artery creates an immediate pressure gradient redirecting blood from distal connecting arterioles, namely collaterals, to maintain CBF at the region of ischemia. The sustainability of the penumbral region will then rely on effective metabolic control of cerebral perfusion during ischemia. Hypoxia induces vasodilation in cerebral vessels through NO synthesis as well as adenosine pathways. This effect, however, is dose-dependent and heterogeneous in the cerebrovascular bed. The effect of the decreased partial arterial pressure in O₂ (PaO₂), is known to be significant at <60 mmHg. This may presumably involve adjacent collaterals depending on the occlusion site and duration. When ischemia lasts, this is followed by an increase in PaCO₂/H⁺, a potent mechanism to reduce cerebrovascular resistance. The vasodilatory effect of PaCO₂/H⁺ is primarily due to the H⁺ effect on vascular smooth muscle. To a lesser extent, this involves NO-induced vasodilation, prostanoids, and endothelin-1. ⁸⁷ Given the immediate effect of CO₂, the “CO₂ challenge” has been implemented as a safe and feasible inhalation test to assess cerebrovascular reactivity in the clinical setting. Accordingly, cerebrovascular CO₂ reactivity shows an early decline with aging, especially in women.^88,89 Since the mitochondrial Krebs cycle is the main source of CO₂ generation, the decline in the citric acid cycle with aging may be theoretically accountable. However, there is no evidence to support a causative link. ⁸²

Minutes after occlusion, when ATP levels fall, and Na/K ATPase channels fail to maintain the membrane electrical gradient, Ca²⁺ influx induces excitotoxicity in neural cells. Momentary changes in local neuronal activity induce dynamic changes in cerebral blood flow. Astrocytes play a central role in coordinating this neurovascular coupling. Neuron-released glutamate increases astrocytic Ca²⁺ influx through group I metabotropic glutamate receptors (mGluR), which in turn induce vasodilation through an intricate interplay of pathways, including K⁺ flux through Ca²⁺-sensitive K⁺ (K_Ca) or voltage-gated K⁺ (K_v) channels, ⁹⁰ cyclooxygenase products (e.g. prostacyclin), or CYP4A products (e.g. epoxyeicosatrienoic acids). Upon sufficient buildup of K⁺ level between the astrocytic end and the adjacent microvessels, K⁺ opens K⁺ inward rectifier (K_ir) channels expressed on vascular smooth cells and thereby leads to vasorelaxation. Besides astrocytic K_Ca or K_v channels, evidence shows that the ATP-dependent K⁺ channels (K_ATP) in smooth muscle and endothelium play a key role in cerebral perfusion control. ⁹¹ As is of more interest to the focus of this review, human studies show K_ATP blockade by glybenclamide deters the compensatory increase in cerebral blood flow in during hypoxia induced by pulmonary obstructive disease. ⁹² However, an earlier study has demonstrated a temporary irresponsiveness in K_ATP channels following 10 min of global ischemia in piglets’ brains. ⁹³

Besides driving vital energy-dependent events in the brain, ATP is one of the key messengers in the nervous system, particularly in astrocytic communications within the NVU. Astrocytes release ATP in response to neuronal hyperactivity, which induces endothelial NO production through endothelial P2Y1 receptors. Through membrane-bound ecto-5′-nucleotidases, astrocytes also hydrolyze proportions of extracellular ATP to adenosine, which in turn stimulates A2A purinergic receptors on vascular smooth muscle, leading to vasodilation.

Given these findings, the existing knowledge suggests that ATP shortage associated with aging may explain the hampered ischemia-induced vasodilation and collateral recruitment, at least partly. Some data show the bECs’ deranged energy metabolism may also recruit endogenous vasoconstrictors (e.g. endothelin-1), which may potentially buffer part of these blood-recruiting pathways. ⁹⁴ The overall outcome of the production of intrinsic vasodilators and constrictors over the occlusion time is time-sensitive and depends on the availability of oxygen and the ensuing cascades. As a typical instance, 20-HETE is a potent constrictor of cerebral blood vessels. ⁹⁰ This compound is a product of arachidonic acid metabolization by astrocytic CYP4A through an oxygen-dependent pathway. During focal ischemic occlusion, however, the effect of 20-HETE to prevent pial collaterals’ dilation is far less than that of endothelin-1.

Hypothetically, the diminished neurovascular energy resources may also affect systemic hypertension in response to stroke. Stroke hypertension elicited by the “selfish brain” is a vasomotor response to the increased intracranial pressure (ICP), which is mediated by astrocytic ATP release in the pons. To our knowledge, there is no evidence to support that the astrocytic ATP fall with aging is linked to the decline in vasomotor reflexes. This hypertensive response creates a driving force to increase cerebral perfusion pressure, either through mechanical force or by the increased shear stress on the bECs, leading to NO release and vasodilation. There is not enough evidence to determine whether this sympathetic output increases CBF in ischemic regions, specifically regarding the expression of adrenergic receptors on cerebral arteries. Moreover, given the mixed effect of vasomotor decline with aging and the high rate of pre-existing hypertension in the elderly, it is unclear whether this adaptive response is altered in elderly strokes. ⁹⁵

Edema and clearance failure

A major mechanism of waste removal in the brain is the glymphatic system. This drives a polarized flow of fluids from the arterial paravascular space to the interstitial (ISF) fluid compartment and then the venous paravascular space. This “glymphatic movement/flow” initiates with CSF influx to the para-arterial space, where it is forced into ISF, and matures with CSF efflux to the para-venous space. Using indirect methods of imaging, for example, tractography of the perivascular space or magnetic resonance elastography, clinical studies have demonstrated compromised glymphatic function with aging, in close correlation with sleep disorders and neurodegenerative diseases. ⁹⁶ Based on preclinical findings, the reduction in glymphatic CSF influx, as determined by CSF-tracer tracking, might be as dramatic as ~80%–90% in aged compared to young mice. ⁹⁷ Although the age effect on the glymphatic function is supported by concrete clinical and preclinical evidence, our understanding of the stroke effect relies on a few recent findings. Accordingly, stroke shows synergic effects with aging. ⁹⁸ However, acute vasoconstriction appears to be the most prominent parameter to mediate the stroke effect to deter glymphatic movement, leading to fluid retention and edema. ⁹⁹ Putatively, the impaired fluid turnover correlates with escalated edema. The clinical findings, however, are not consistent enough to conclude the age effect on stroke-induced edema. The relatively bigger infarction in the elderly ¹⁰⁰ and age-related brain atrophy, providing buffering space for brain swelling, ¹⁰¹ are possible confounders that have not been consistently taken into account. In fact, some clinical findings show malignant edema is associated with younger age, either in baseline ¹⁰² or after mechanical thrombectomy. ¹⁰³ That is, while other reports imply that advanced age predicts increased edema. ¹⁰¹ Regarding the notion that inefficient clearance and the accumulation of misfolded/hyperphosphorylated proteins amplify pro-inflammatory responses, the dramatic disruption of fluidic turnover with aging conceivably contribute to massive cell injury independent of edema.

The production of CSF via the choroid plexus, as well as its flow, that is, with breathing, are the primary drivers of CSF influx to the para-arterial space, and the following CSF penetration into the ISF. Aging results in a significant decline in CSF production (by 66%) and CSF pressure (by 27%). ¹⁰⁴ Given CSF excretion from choroid plexus relies on epithelial Na⁺/K⁺ ATPase,^105,106 the overall ATP reduction with aging may conceivably contribute to the decline in glymphatic flow. Glymphatic movement also highly depends on arterial pulsations to force CSF towards the peri-arterial astrocytes’ end feet, ¹⁰⁷ where perivascular AQP4 water channels polarize fluids to flow from CSF to mix with the parenchymal ISF. Reduced arterial compliance with aging perturbs normal vascular pulsations in arteries and capillaries. The reduced NO availability in the metabolically frail aged bEC (see above) may presumably further dampen pulse waves in arteries and capillaries. This mechanism, however, is more functional in arteries, ¹⁰⁸ that is, the pial and penetrating arteries, with NO-responsive smooth muscle cells. Given the glymphatic CSF influx is variably driven by both arteries and capillary pulsation, the significance of this mechanism remains arguable. The subsequent CSF transport into the dense brain parenchyma through AQP4 astrocytic end feet also undergoes enormous changes with aging. That is, the loss of perivascular AQP4 polarization with aging (i.e. re-polarization towards astrocytic parenchymal processes) critically impairs the synchronized convection required for the efficient wash-out. Based on recent findings, the localization of AQP4 to the plasma membrane depends on PKA activation by intracellular cAMP. ¹⁰⁹ The formation of cAMP from ATP is an energy-dependent process and is highly affected by ATP bioavailability. Given that both cAMP and ATP show a significant decline with aging, impaired AQP4 localization may delineate another potential energy-dependent mechanism for impaired fluid turnover.

Brain–blood over-negotiation

Dysregulation of the bidirectional exchange of materials between the brain and the periphery is an indispensable feature of aging and contributes to the exacerbation of stroke injury in the elderly. Firstly, the signature impairment in myogenic autoregulation with aging restrains cerebral arteries’ ability to buffer the hypertension-induced end-arterial pressure and to prevent the consequent intracerebral microhemorrhages. This is in line with the statistics concluding a high prevalence of hemorrhagic transformation (HT) in elderly strokes. Secondly and more extensively studied, blood–brain barrier disruption and unleashed neuroinflammation in an aged brain are liable for the ungated communication between the brain and the periphery. Traditionally, aging was believed to augment the metal-proteases effect in cerebral vessels, allowing leakage of damage-associated molecular patterns (DAMPs) to the bloodstream, leading to the invasion of blood-borne immune cells. With the advancement of tracer molecular techniques, it was shown that this characteristic alteration in BBB integrity is preceded by an increase in endothelial permeability, particularly through nonspecific caveolar transcytosis. This uncontrolled influx of blood-borne molecules, including the concentrated inflammatory cytokines in the aged blood, may further underline inflammaging as the main basis to drive both BBB breakdown as well as aggravated stroke outcomes.

The age effect on immune responses to stroke

Compiling clinical evidence confirms that older individuals experience a more intense and prolonged inflammatory response to ischemic stroke. As we age, the immune cells adopt a more proinflammatory—less clearing phenotype (immunosenescence) which results in the accumulation of waste compounds, “inflammaging.” As the first unit sensing oxygen deficiency in the brain, bECs’ mitochondria react to ischemic occlusion with robust ROS production within minutes. This initiates inflammatory reactions in glial cells as well as circulatory monocytes and platelets. In the acute phase following occlusion, glial cells, monocytes, and neutrophils are the primary cells involved in the immune reaction at the injury site. In the sub-acute phase, this is followed by T-cell recruitment; however, the subsequent stroke-induced sympathetic tone and hormonal signals lead to lymphocytopenia. This immunosuppression predisposes stroke patients to infection-related complications and concurs with the invasion of more neutrophils and monocytes into the circulation and injured brain parenchyma.

Aging reprograms nearly all aspects of the immune system, both in the CNS and the periphery. Aged microglial cells tend to adopt the M1 proinflammatory phenotype and develop lipid droplets. A typical feature in old microglia is upregulated type one interferon (IFN1) signaling, either in resting status or in response to ischemia. ¹¹⁰ Consistent clinical findings demonstrate that stroke in the elderly is associated with more NKT cells activity, ¹¹¹ which is associated with the IFN1 signaling pathway. ¹¹² Despite this polarization, aged microglia are less responsive to ischemic injury.^113,114 Coupled with the findings showing microglial depletion aggravates stroke outcomes, both in young ¹¹⁵ and aged brains, ¹¹⁶ this questions the detrimental role of microglia. Aging also encourages the reactive proinflammatory phenotype in astrocytes (A1), overexpressing Cxcl10 and major histocompatibility complex (MHC). This phenotype shows enhanced proinflammatory cytokines, such as interferon-γ (IFN-γ), in ischemic injury.^117,118 This is while regulatory T Cells, which are crucial for suppressing excessive inflammation, tend to decline with aging. ¹¹⁹

Besides altering the local immune responses, aging reprograms systemic immune reactions to stroke and shifts lymphopoiesis to myelopoiesis. Recent single-cell analysis has revealed that aging induces a drastic shift toward Neutrophil reactivity in the aged brain. Based on the breakthrough report by the Bacigaluppi research team (2023), the increase in CD62L neutrophil in aged mice and elderly stroke patients promotes thrombin formation and explains the impaired reperfusion and worse inflammation in the elderly. ⁶ Convincingly, this explains the earlier reports showing that the transfusion of healthy young blood, ¹²⁰ or transplantation of young bone marrow hematopoietic stem cells, ¹²¹ significantly improves stroke outcomes in older adult mice. Such findings may imply that the aggravated neurovascular response in the aged brain is not intrinsic and mirrors the systemic/circulatory immune responses to stroke. Importantly, from a translational perspective, these studies did not use stroke animals as mitochondrial donors. As such, these findings can not establish the significance of bECs senescence in this communication. Compiling preclinical evidence supports the notion that senescent bECs have an enormous contributing effect to systemic pro-inflammatory status. This is supported by substantial preclinical and clinical evidence indicating that age-related vascular comorbidities (e.g. hypertension, diabetes) exacerbate the inflammatory response in stroke patients.

Effect of impaired bioenergetics on the brain–blood interface

In a closed circulatory system, biological compounds move between the blood and the brain either via the blood–brain barrier (BBB) or by passing through the intermediate brain–CSF barrier (BCB) compartment. BCBs are epithelial barriers in the arachnoid layer and choroid plexus, a highly vascularized epithelium within the brain ventricles. In addition to BBB, BCB represents a major entry route for immune cells (e.g. monocytes) to the ischemic brain parenchyma. ECs lining the microvasculature within the choroid plexus, the main niche for CSF production, are exceptionally fenestrated and do not form an impermeable barrier. The blood-borne molecules and/or cells can transmigrate across the epithelial layer and subsequently enter the ventricles along with CSF. Even though the bECs in BCB do not preserve barrier properties, they play a remarkable role in peripheral immune cell recruitment. In support of this, the bECs in the choroid plexus have been documented as one of the main niches of immune reactions in an aged stroke brain.^122,123 The frailty of energy machinery in the aged bECs, and the consequent free radical release, may contribute to a vicious cycle of inflammatory responses and subsequent complications through different pathways:

Enhanced granulocyte recruitment: As the bECs age, they fail to keep the intercellular connections through tight junction proteins. In the meantime, bECs start to increase adhesion molecule expression (VCAM-1, ICAM-1, p-selectin), raising the communication with circulatory neutrophils, which play an exceptional role in elderly strokes.^124,125 In the dysfunctional mitochondria in an aged bEC, where energy-rich electrons may not flow towards ATP generation, the generation of ROS triggers PKC and NF-κB signaling pathways, which mediate inflammatory responses and upregulate adhesion molecules. ¹²⁶ This is exacerbated with the sudden interruption of blood flow, where bECs respond to the abrupt fall in shear stress by overexpressing adhesion molecules, leading to local trapping of blood cells and platelet-leukocyte adhesion. With the subsequent release of chemokines from the ischemic parenchyma, this adhesion cascade results in profound Neutrophil infiltration.

Aggravated innate immune response: Astrocytes, the most abundant glial cells and the vital component of NVU, rely on bECs for sufficient ATP production. Besides local regulation of vascular tone in response to ischemic occlusion, astrocytes play a key role in controlling inflammation and excitotoxicity via glutamate uptake. Uptake of glutamate is a costly process where the exchange of one molecule of glutamate requires the transport of three Na⁺ molecules through Na⁺/K⁺-ATPase, in the expenditure of one molecule of ATP. ¹²⁷ The reduced ATP availability to astrocytes, as discussed earlier, may explain the observation that astrocytes collected from the ischemic aged mouse brain clear less glutamate, regardless of the glutamate carrier (GLT-1) expression.^128,129 The enhanced excitotoxicity promotes the accumulation of DAMP, which triggers innate immune signaling and drives chemokine release by the aged microglia. Diffusion of DAMP and chemotactic molecules through the leaky BBB (and/or BCB) recruits more blood-born cells to the injury site. ¹³⁰ Therefore, by escalating the parenchymal inflammation, the aged bECs engage more immune cells to relay the alarming message to the periphery.

Paracrine inflammatory signals: Beyond dysregulating innate and adaptive immune cells, the impaired bioenergetics may eventually transform the aged bECs into senescence-associated secretory phenotype (SASP), exhibiting paracrine effects (see above, Figure 2). The canonical free radical theory of aging is best exemplified by the studies demonstrating impaired mitochondrial respiration and increased superoxide production in SASP cells, ¹³¹ and is supported by studies demonstrating reduced senescence by neutralizing mitochondrial ROS. ¹³² The SASP cells exhibit increased attachment to the basement membrane. This rigid positioning prevents the physiological alignment with the laminar bloodstream and deters the existing NO functional capacity in response to altered shear stress during occlusion.^32,133 The secretion of the SASP factors (e.g. inflammatory chemokines and ROS) produces paracrine effects that encourage senescence-secretory phenotype in the neighboring bECs. In the course of arterial occlusion, this local secretion relays a strong damage signal to circulating and/or trapped granulocytes.

Stroke management in the elderly

Optimal acute stroke management in the elderly is complex and requires specialized personnel. Older stroke patients are at higher risk for life-threatening complications like multi-organ failure, besides developing delirium, infections, and hemorrhagic transformation (HT). For the relatively high rate of HT and mortality in patients >80 years, intravascular intervention is considered based on specific risk/benefit algorithms for each individual. Despite worse outcomes with aging, octa/nanograins meeting the criteria still benefit from thrombolysis ¹³⁴ or thrombectomy, ¹³⁵ compared to placebo or non-treated groups. Older stroke patients with poor baseline functions, malignant infractions, and massive leukoaraiosis at admission are merely assigned to supportive measures and medications for stroke complications (e.g. hypertension) or secondary strokes. Regarding the remarkable alterations in drugs’ dynamics and kinetics in particular, the medications require dose adjustment to avoid toxicity, especially for therapies increasing the risk of bleeding (e.g. antiplatelets), hypotension (e.g. antihypertensives), or sedation (antipyretics, analgesics). From the pile of anti-oxidant and immunomodulatory agents demonstrating promising effects in preclinical studies, only a few have proceeded to early phases of clinical trials, where almost always octagerains are excluded. The subcutaneous administration of interleukin-1 receptor-antagonist, interleukin 1Ra; and the adjunctive administration of the toll-like receptor 4 (TLR4), ApTOLL with thrombectomy are among anti-inflammatory interventions showing promising effects in phase II clinical trials. The exceptional safety and efficacy of edaravone, the high-affinity chelator of hydroxyl radicals, have led to its admission in clinical practice in Japan and some Asian countries. Given that the drug’s benefit has been observed in different ages, including older adults >80, these findings already support edaravone as a promising medication in the management of stroke in the elderly. ¹³⁶ With the growing understanding of the significance of mitochondrial integrity in aging and disease, there is an increasing interest in investigating mitochondrial interventions. Resveratrol, a SIRT-1 agonist, is now in phase II clinical trial for anti-aging effect in people >65 and represents a promising adjunctive therapy in stroke prevention or management.

Perspectives of endothelial mitotherapy in elderly stroke

The discovery of mitochondrial shuttling between neighboring cells, the idea of cell-free mitochondrial transplantation, placed “mitotherapy” in the spotlight. The pioneering study by McCully research team (2013) in New Zealand rabbits ¹³⁷ was the first to show that autologous muscle-isolated mitochondria may incorporate into myocardial cells following local injection. This intervention enhanced mitochondrial energy generation and functional outcome following ischemic reperfusion injury. This report entailed several studies on experimental cerebral ischemic stroke, demonstrating the beneficial effects of autologous transplantation of mitochondria obtained from skeletal muscle biopsy ¹³⁸ mesenchymal stem cells, ¹³⁹ and a heterologous placenta. ¹⁴⁰

Since these studies administered mitochondria to the CSF (intraventricular injection) or systemic circulation (intravenous infusion), the observations highlight endothelial/epithelial cells as the main compartment mediating the mitochondrial transfer into the brain. This hypothesis is supported by experimental evidence indicating endothelial cells receive mitochondria from neighboring cells mainly through TNTs and extracellular vesicles (EVs). TNTs are microtubule-based transport systems (involving kinestin-1 and Miro1), mediating direct cell-to-cell mitochondrial transfer and leading to improved mitochondrial bioenergetics in the recipient cell. EVs deliver mitochondria to the endothelium through migrating packages, which are either degraded in lysosomes or incorporated into the endothelium functionality, mainly to boost angiogenesis. The process is mediated by several molecules (e.g. ARRD1C and Myo19) for EV formation and efficient cargo transport. Research shows endothelial cells may receive cell-free mitochondria when added to the cell-culture medium. This process, although less characterized, has been shown to increase ATP levels 100–200-fold in bECs exposed to OGD. ¹⁴¹ Bioengineering strategies enable researchers to develop endothelial-targeted mitotherapy techniques. Enhancing mitochondrial transfer into the endothelium not only enhances the functional outcome but also significantly reduces the organelles’ biodistribution and biodegradation. Despite the promising advantages, the efficacy and safety have not been explored enough in in vivo experiments. The process is further complicated to evaluate in the clinical setting for safety concerns (see next section). Importantly, based on recent in vitro experiments, conjugation of VCAM-1-binding or collagen-binding peptides to mitochondria enables targeting to dysfunctional diabetic endothelium. Given the similarities between diabetic ECs and the aged bEC phenotype, this provides important insight into mitotherapy-based interventions for senescent bECs. ¹⁴²

In 2017, the first successful autologous human mitochondrial transplantation was reported in five cases of pediatric cardiac arrest at Boston Children’s Hospital. ¹⁴³ The same research team also showed that intravascular mitochondrial transplantation significantly improves recanalization outcomes in these patients. ¹⁴⁰ More recently, the safety and benefit of mitochondrial transplantation have also been reported in adults (50–60 years old) suffering from ischemic heart disease. ¹⁴⁴ As of specific interest in this review, the first use of mitochondrial transplantation in the human cerebral stroke has been recently examined by Walker’s research team. ¹⁴⁵ In four stroke patients (three female and one male, age 43–80 years old), this open-label trial demonstrated suitable safety and feasibility of intra-arterial autologous mitochondrial transplantation along with mechanical thrombectomy. This ongoing phase-1 clinical trial (ID: NCT04998357) implements a rapid, on-site protocol to isolate quality-controlled mitochondria from skeletal muscle biopsies at the site of femoral artery access for thrombectomy.

Therapy development, potentials, and challenges

Research on elderly stroke faces challenging technical obstacles for direct observations. Our clinical knowledge of elderly stroke is rare and limited to the last few decades. Given the concerns for adverse complications, very old patients (<80) are often excluded from clinical trials or not assigned to intravascular interventions. The in vivo observation in aged animals is worth the technical challenges, as it provides precise, direct information. However, given the uncertainties about the cut-off in the definition of old versus very-old animals, the data from divergent preclinical studies are sometimes difficult to translate to human aging. The existing restrictions have encouraged preclinical studies towards in vitro modeling of aging, for example, replicative or chemically induced senescence in primary endothelial cells. However, due to the inherent limitation, these models are mostly kept for complementary experiments. The chemical modeling of aging (e.g. by cytokines or H₂O₂), engaging very specific pathways, does not fully capture the complex physiology of aging. The replicative senescence model, on the other hand, illustrates a more aging-relevant model only after prolonged culture times, which predispose the aging cells to accumulate minimal epigenetic/environmental variables, for example, the batch-to-batch variability of reagents, leading to highly variable results. This may become more complicated when commonly used aging markers (β-galactosidase, P21, and/or P16) respond to the environmental factors, rather than the aging phenomenon itself. ¹⁴⁶ In this connection, it is important to note that, despite plentiful opposite data, some in vitro studies have reported an anti-inflammatory phenotype in senescent endothelium.^147,148 It is not clear whether this inconsistency is a shortcoming of in vitro aging models or a vivid reflection of “Adaptive senectitude.”

As briefed earlier, mitochondrial transplantation has shown promising data in some human strokes, particularly in elderly subjects. However, the clinical application has remained extremely restricted due to regulations concerning safety and efficacy. First and foremost, extracting a pure mitochondrial fraction from the donor tissue is challenging due to the low viability of the mitochondria, even with a quick and high-yield protocol. This might get more complicated with minimal incorporation of manufacturing under aseptic conditions at the bedside. The small proportion of quality mitochondria further faces biodegeneration. The particular signature of unleashed inflammation and mitochondrial frailty in the aged endothelium creates more obstacles. The amplified immune responses with aging presumably aggravate cell-free mitochondria biodegradation before endocytosis to the recipient cell. This reaction will be worse in low-quality control and in damaged mitochondria, which is not unlikely in autologous preparations from the elderly organs. Delivering mitochondria through stem cells or vesicles efficiently minimizes this hurdle. Other attempts for cytosolic delivery of energy packs (e.g. ATP), metabolites (ketone bodies/keto acids), and mitochondrial organelles are also complicated by off-target effects and biodegradation (e.g. ATP-induced damage signals). Although the proinflammatory signature of aging is a concern in the bioavailability of biomaterials, the associated upregulation of adhesion molecules and bECs’s permeability allows targeted delivery through engineered extracellular vesicles or coated free mitochondria with ligands or antibodies to facilitate endothelial uptake. Furthermore, manufactured human mesenchymal stem cells have recently emerged as specialized mitochondrial donors, with increased mitochondrial quantity and quality. Therefore, a balanced combination of bioengineering strategies and stem cell donors offers exceptional flexibility for (1) choosing a variety of heterogeneous sources for mitochondria/organelles and mitigating the manufacturing barriers, (2) targeted delivery to endothelial cells, in the design of future trials in elderly stroke patients. Clinical application of mitochondrial transplants still faces serious ethical constraints for introducing the mitochondrial DNA (mtDNA) and safety concerns over genetic modification in the recipients. Research continues to develop techniques to preserve the parents’ nuclear DNA from the mtDNA genome, while clinical research is restricted to autologous transplantation.

Concluding remarks

Aging is characterized by the failure to maintain homeostasis under conditions of physiological stress. The epigenetic factors may profoundly accelerate the vicious cycle of “functional aging” and alter the prognosis of stroke over a short time of a decade. ¹³⁵ As briefed in this review, endothelial energy failure with aging may define a distinct pathology and prognosis for stroke (Figure 4). The existing evidence supports the hypothesis that preserving endothelial cells’ bioenergetics through combined approaches, including mitotherapy, decelerates cerebrovascular aging and expands the functional reservoirs to adapt to the energy crisis during ischemic stress. Given the unleashed immune responses, profound alteration in the body composition and metabolizing capacity, and the consequent changes in drug biodistribution, a specialized platform is required for developing therapies in the elderly. Studies demonstrate that quality-controlled encapsulated mitochondria with a manufactured genome provide the utmost efficiency, bioavailability, and safety. However, it is not clear if the age-related alteration in cellular uptake may impact the ultimate advantages. In this connection, using senolytics may provide an alternate or combinatory approach for mitochondrial renovation. By eliminating the dysfunctional senescent cells, senolytics trigger cell replication/rejuvenation, leading to improved mitochondrial machinery and transfer as well. Investigating these open gaps may introduce promising approaches to elderly stroke management. Improving our understanding of the hermetic effects of aging during the lifespan, the particular pathophysiology of stroke in very old adults, and developing the appropriate preclinical models are crucial in future directions.

Figure 4.

Hypothetical impact of aging-induced impairment in ATP biogenesis in the acute course of ischemic stroke. Senescent endothelial cells with a preexisting decline in glucose uptake and metabolism do not gate sufficient energy and nutrients to the endangered NVU proximal to the arterial occlusion. With the progress of ischemic injury, the brain redistributes the blood to access the occlusion site. Given the limited reservoir of vasoactive machinery (NO and ADP), the cerebrovascular tree may not engage enough collaterals to direct blood flow to the salvageable tissue. As more cells swell and die, the glymphatic movement contributes to the waste washout. However, the reduced vessels’ pulsatility and APQ4 localization in the astrocytes’ endfeet may not establish the glymphatic polarized fluid movement. Coupled with the astrocytes’ insufficiency in ATP-dependent glutamate take-up, this results in the buildup of DAMP and chemokines which accelerates innate immune responses by resident microglia. This is followed by pronounced immune cells anchoring to the activated endothelium exposing more adhesion molecules in reaction to enhanced ROS generation in oxPHOS.

Supplemental Material

sj-docx-1-jcb-10.1177_0271678X261437917 – Supplemental material for Endothelial energy failure as a therapeutic target in elderly strokes

Supplemental material, sj-docx-1-jcb-10.1177_0271678X261437917 for Endothelial energy failure as a therapeutic target in elderly strokes by Sanaz Nasoohi, Farehe Ebrahimi, Candice M Brown and Jason D Huber in Journal of Cerebral Blood Flow & Metabolism