Introduction

Senile dementia is a progressive loss of cognitive ability, including memory, attention, language, and problem-solving, in the elderly and is a serious medical, social, and economic burden in modern society because of the growing aged population. Senile dementia mainly includes Alzheimer’s disease (AD), vascular dementia (VaD), and mixed dementia. As the most common form of dementia in older adults, AD is characterized by chronic and progressive neurodegeneration leading to progressive cognitive impairment and eventually to the death of patients. AD in most cases is sporadic, probably caused by multiple factors, and is characterized histopathologically by the presence of both intraneuronal neurofibrillary tangles (NFTs) and extracellular senile plaques together with neurodegeneration in the brain (Braak and Braak 1991). VaD is the second most common form of dementia (Battistin and Cagnin 2010) and is caused by problems in the supply of blood to the brain. Mixed dementia is diagnosed when patients have evidence of both AD and cerebrovascular disorder, either clinically or based on neuroimaging evidence of ischemic lesions. In fact, AD and VaD often coexist in older patients with dementia. It is estimated that as many as 40 % of AD patients actually have mixed dementia (Battistin and Cagnin 2010).

Chronic cerebral hypoperfusion (CCH) (Pristera et al. 2013) is one of the major mechanisms of cerebral vascular disorders and can result from hypertension, diabetes, generalized atherosclerosis, smoking, and heart diseases. These factors can affect the cerebral vascular system and eventually cause decreased blood supply to the brain (Meyer et al. 2000; Roman 2002; Valerio Romanini et al. 2013). Individuals with CCH usually have cognitive deficits to various degrees (Ruitenberg et al. 2005). The important role of CCH in dementia has emerged to the forefront of neurology research (Pluta et al. 2012; Akinyemi et al. 2013; Kelleher and Soiza 2013; Pluta et al. 2013a, b; Roh and Lee 2014). Individuals with moderate or severe intracranial arterial stenosis have a faster decline in cognition and function relative to those without such stenosis (Zhu et al. 2014). Studies in the last decade have suggested that CCH might promote neurodegeneration through neuronal energy failure, production of reactive oxygen species, and pro-inflammatory cytokines through activated microglial cells that, in turn, damage the neurons and contribute to white matter lesions (Kitagawa et al. 2005; Farkas et al. 2007; Adibhatla and Hatcher 2008; Urabe 2012; Bang et al. 2013). This article attempts to review the recent advances focusing on CCH-induced neurodegeneration and AD-related brain pathology and cognitive impairment.

Causes and Assessment of CCH

CCH has a variety of causes and plays an important role in the development of VaD, AD, and subcortical arteriosclerotic encephalopathy. Three main causes lead to CCH: (i) vascular structural lesions resulting from artery stenosis or occlusion caused by atherosclerosis, arteriovenous malformation, takayasu arteritis, moyamoya disease, and cerebral arteriovenous fistula; (ii) cerebral hemodynamic changes, including chronic blood loss, prolonged hypotension, and reduced cardiac output due to heart failure; and (iii) changes in blood components resulting from any reasons that lead to an increase of blood viscosity, such as hyperlipidemia, polycythemia, and hyperhomocysteinemia. The major risk factors of CCH are hypertension, hyperlipidemia, smoking, obesity, age, hyperhomocysteinemia, and obstructive sleep apnea-hypopnea syndrome (Sarti et al. 2002).

Modern development of medical techniques has made the assessment of CCH very practical. In the clinic, cerebral hypoperfusion is usually assessed by transcranial Doppler ultrasonography, computerized tomography angiography, magnetic resonance angiography, computerized tomography perfusion imaging, perfusion-weighted imaging, or Xenon-CT (computerized tomography). For animal studies, CCH can be assessed by laser-Doppler flowmetry (Kitagawa et al. 2005) or by molecular markers such as hypoxia inducible factor-1 (HIF-1), TP53-induced glycolysis, and apoptosis regulator (TIGAR) or glucose transporters (GLUTs) 1 and 3 (Watanabe et al. 2009; Kimata et al. 2010; Chan et al. 2011; Iwabuchi and Kawahara 2011; Yan et al. 2011; Yuan et al. 2011; Hoshino et al. 2012; Wang et al. 2012a, b; Zhao et al. 2014).

Possible Mechanisms from CCH to Alzheimer-Like Pathology and Neurodegeneration

AD and neurodegeneration can be caused by multiple etiological factors. Except for familial, early-onset AD that is caused by mutations of genes encoding presenilin or amyloid β (Aβ) precursor protein (APP), over 95 % AD cases are of sporadic nature. Sporadic AD is likely caused by multiple etiological factors through several pathogenic mechanisms (Iqbal et al. 2010). The major risk factor for sporadic AD is aging, and the most consistently identified AD susceptibility factor is the ε4 allele of apolipoprotein E (Poirier et al. 1993; Strittmatter et al. 1993). Although the amyloid cascade hypothesis continues to exert an important influence in the AD field, recent studies suggest that CCH also promotes Alzheimer-like brain pathology and neurodegeneration through several molecular mechanisms (Fig. 1), as discussed below.

Fig. 1
figure 1

Possible mechanisms by which CCH promotes/causes Alzheimer-like pathology and neurodegeneration. CCH can initiate several pathways that result in oxidative stress, Aβ overproduction and aggregation, tau hyperphosphorylation, synaptic dysfunction, white matter lesion, and neuroinflammation. These neuropathological changes are all seen in AD brain. They can interact and exacerbate each other and eventually promote neuronal cell death and neurodegeneration, resulting in clinical phenotype of memory loss and dementia

Full size image

CCH-Induced Oxidative Stress

Oxidative stress reflects an imbalance between the systemic manifestation of reactive oxygen species and the biological system’s ability to readily detoxify the reactive intermediates or to repair the resulting damage. Disturbances in the normal redox state of cells can cause toxic effects through the production of peroxides and free radicals that damage almost all components of the cell, including proteins, lipids, and DNA. In humans, oxidative stress is thought to be involved in the development of many diseases including AD (Singh et al. 1995; James et al. 2004; Halliwell 2007; Valko et al. 2007; de Diego-Otero et al. 2009; Dean et al. 2011). CCH can cause mitochondrial dysfunction (Orsucci et al. 2013) and protein synthesis inhibition, which in turn may disturb the balance of antioxidases and reactive oxygen species and produce oxidative damage. At the same time, oxidative injury to vascular endothelial cells, glia, and neurons could further impair vascular function and neurovascular coupling, an orchestrated intercellular communication between neurons, astrocytes and microvessels, which results in a rapid and restricted increase in cerebral blood flow in order to maintain normal brain function in a timely and local manner. Impairment of vascular function and neurovascular coupling may result in a vicious cycle of further reduction of cerebral perfusion (Lyons and Pahwa 2013).

Numerous studies have provided evidence that CCH leads to oxidative stress and have described the mechanism by which oxidative damage results in cognitive impairment. Liu et al. described an alternative CCH rat model by two-stage, three-vessel occlusion and found impaired spatial learning and memory and increased levels of malondialdehyde, the end products of lipid peroxidation (Liu et al. 2012). Xi et al. established a CCH rat model through permanent bilateral common carotid artery occlusion (BCCAO) and found central cholinergic dysfunction and increased oxidative damage that is correlated with spatial learning and memory impairments and working memory dysfunction (Xi et al. 2014). In the BCCAO model, reduction of pyruvate dehydrogenase level and increase of oxidative stress occur in the hippocampus, suggesting that mitochondrial bioenergetic deficits might affect memory directly (Du et al. 2013). Apoptosis signal-regulating kinase 1 (ASK1) appears to be critical to CCH-induced oxidative stress, because the white matter lesions resulting from bilateral common carotid artery stenosis (BCAS) cause oxidative stress-associated cognitive decline in wild-type mice, but not in ASK1-deficient mice (Toyama et al. 2014). ASK1, also known as mitogen-activated protein kinase kinase kinase 5, is a member of the mitogen-activated protein kinase (MAPK) pathway and activates c-Jun N-terminal kinase (JNK) and p38 MAPK in a Raf-independent manner in response to an array of stresses, including oxidative stress, endoplasmic reticulum stress, and calcium influx. ASK1 is also involved in neurodegenerative diseases (Hattori et al. 2009). The oxidative stress-ASK1-p38 cascade appears to play an important role in the pathogenesis of cognitive impairment caused by CCH (Toyama et al. 2014).

CCH-Induced Accumulation and Aggravation of Aβ

Aβ, a peptide of 36–43 amino acids, is the main component of the amyloid plaques found in AD brain. Aβ is derived from APP through proteolytic cleavages by β- and γ-secretases. Aβ pathological deposition occurs both in the brain parenchyma and in the vascular structure in AD brain and in the brains of transgenic animal models with APP mutations (Games et al. 1995; Hsiao et al. 1996; Sturchler-Pierrat et al. 1997; Bornemann and Staufenbiel 2000). It has been reported that CCH accelerates Aβ deposition. BCAS-induced CCH can increase Aβ fibrillization and induce Aβ deposition in the intracellular compartment and, therefore, may accelerate the pathological changes of AD in APPSwe/Ind-Tg mice one month after BCAS (Kitaguchi et al. 2009). CCH induced by permanent unilateral common carotid artery occlusion (UCCAO) causes spatial learning impairments that correlate with the number of cortical Aβ plaques in young APPSwe/PS1 mice (Pimentel-Coelho et al. 2013). In a mouse model of cerebral amyloid angiopathy [C57BL/6-Tg(Thy1-APPSwDutIowa)], BCAS increased Aβ deposition 12 weeks after BCAS surgery (Okamoto et al. 2012). CCH may cause Aβ deposition in aged wild-type animals too. Time-dependent accumulation of oligomeric Aβ in the hippocampus, especially in the axonal terminals of aged rats, occurs after CCH induced by BCCAO (Wang et al. 2010). It is worth noting that the amyloid deposits as seen nine months after transient middle cerebral artery occlusion cannot be stained with Congo red or Thioflavine S (van Groen et al. 2005), which are routinely used to detect the β-pleated sheet conformation that is typical of mature Aβ plaques in AD. Thus, acute cerebral ischemia might lead to Aβ deposition that is somewhat different from the mature Aβ plaques seen in AD brain.

Lots of evidences have showed that CCH and other hypoxia conditions up-regulate β and γ secretase-mediated APP processing (Sun et al. 2006; Li et al. 2009; Zhiyou et al. 2009; Koike et al. 2010; Pluta et al. 2013a, b). A possible mechanism by which CCH up-regulates APP processing and leads to Aβ accumulation could be that CCH induces HIF-1 expression, which then binds to the promoter of β-secretase and consequently increases its expression (Zhang et al. 2007). The Aβ deposition in small arteries caused by CCH could further induce cerebrovascular lesion (Thomas et al. 1996) and worsen cerebral hypoperfusion and finally lead to a vicious circle and irreversible damages.

CCH-Induced Hyperphosphorylation of Tau

The microtubule-associated protein tau becomes abnormally hyperphosphorylated in the brains of individuals with AD and several other neurodegenerative disorders collectively called tauopathies (Grundke-Iqbal et al. 1986). It has been demonstrated that abnormal hyperphosphorylation of tau is crucial to neurodegeneration in AD and probably also in other tauopathies (Gong and Iqbal 2008; Iqbal et al. 2013). Tau hyperphosphorylation can be promoted by several factors. One of these factors could be CCH-induced decrease of brain glucose metabolism because the latter leads to down-regulation of tau O-GlcNAcylation that in turn results in tau hyperphosphorylation (Liu et al. 2004; Li et al. 2006; Liu et al. 2009). In a mouse model of CCH induced by UCCAO, we recently found decreased levels of O-GlcNAcylation, increased levels of tau phosphorylation at several AD-relevant sites, selective neurodegeneration in the brain, and significant short-term memory deficits and mild long-term spatial memory impairment (Zhao et al. 2014).

Several protein kinases, such as glycogen synthase kinase-3β (GSK-3β), extracellular signal-regulated kinases (ERK1/2), cyclin dependent kinases 5 (CDK5), cAMP-dependent protein kinase (PKA), calcium/calmodulin-dependent protein kinase II (CaMK-II), and JNKs, have been implicated in hyperphosphorylation of tau in AD (Gong et al. 2010). Tau phosphorylation is also regulated by protein phosphatase 2A (PP2A) (Gong et al. 2000), which accounts for over 70 % of total tau phosphatase activity in the mammalian brain (Liu et al. 2005). Increased tau hyperphosphorylation with concurrent activation of GSK-3β, CDK5, and CaMK-II, as well as inhibition of PP2A is observed in a rat model of CCH, which shows spatial learning/memory deficits (Yao et al. 2012). It appears that CCH can induce abnormal hyperphosphorylation of tau through several pathways.

Tau hyperphosphorylation and Aβ overproduction appear to be very sensitive to cerebral hypoperfusion. Koike et al. have reported that a single, mild, cerebral hypoperfusion has profound and long lasting effects on tau hyperphosphorylation and Aβ overproduction in 3xTg-AD mice (Koike et al. 2010). It is thus reasonable to speculate that repeated transient cerebral ischemia could contribute to the development of AD.

CCH-Induced Synaptic Dysfunction

Synapses are the functional unit of information transfer between neurons and comprise presynaptic membrane, synaptic cleft, and postsynaptic membrane. Synaptic integrity is essential for normal brain functions, including learning and memory. The severity of clinical symptoms correlates highly with synaptic loss in the brain in AD (Scheff et al. 2011). Synaptic integrity can be studied by using electron microscopy and also indirectly through electrophysiology and studies of various synaptic proteins, such as synapsin, synaptophysin, synaptobrevin, synaptotagmin, synaptoporin, and postsynaptic density protein 95 (PSD95). Altered levels of these synaptic proteins have been seen in the brains of both individuals with AD (Mukaetova-Ladinska et al. 2000) and mouse models of AD (Chen et al. 2012, 2013).

Decreased levels of PSD95 and synaptophysin are found in rat brains, especially in the axonal terminals of the hippocampus, 5 weeks after BCCAO (Wang et al. 2010). These rats also show alterations of synaptic ultrastructure in the CA1 area of the hippocampus, as evaluated by electron microscopy. Thus, these synaptic alterations might be the molecular basis of the memory deficits observed in these animals. In a recent study, we also found increased pre-synaptic protein synapsin and post-synaptic protein PSD95, as well as decreased pre-synaptic protein synaptophysin, in the cerebral cortex and the hippocampus of CCH mice 2.5 months after UCCAO (Zhao et al. 2014). These studies confirmed the important role of synaptic integrity in CCH-induced cognitive deficits. The CCH-induced dysfunction of neural plasticity can be observed directly by determination of long-term potentiation (LTP) in CCH animal models. LTP has been found to be inhibited in the hippocampal CA1 region of rat models of CCH for 3 and 6–7 months (Sekhon et al. 1997; Hai et al. 2009).

CCH-Induced Neuronal Loss

Synaptic and neuronal loss in AD correlates directly to the severity of dementia symptoms (Mukaetova-Ladinska et al. 2000; Scheff et al. 2011) and is also seen after chronic cerebral ischemia (Wang et al. 2010; Zhao et al. 2014). Animal studies have shown apoptotic morphology and DNA strand breaks in hippocampal pyramidal neurons 27 weeks after BCCAO, and the working memory impairment correlates strongly to the number of apoptotic neurons in the CA1 region, suggesting that apoptotic loss of pyramidal neurons may underlie memory impairment associated with CCH (Bennett et al. 1998). Hippocampal atrophy with pyknotic and apoptotic cells is also seen in the brain 8 months after BCAS (Nishio et al. 2010). The neuronal loss in the CA1 subfield, together with the decrease of central acetylcholine levels in the cortex, striatum, and hypothalamus, appears 4 months, but not 1 month, after permanent BCCAO, suggesting that the neuronal loss caused by CCH might result from a long period of failure of neuronal excitation transmission (Ni et al. 1995). Significant numbers of degenerative neurons are detected with Fluoro-Jade staining in the ipsilateral hippocampus and cerebral cortex, especially in the granule cells of the crest of the dentate gyrus, of the CCH mice 2.5 months after UCCAO (Zhao et al. 2014). This study also suggests that the dentate gyrus is the most vulnerable area in the brain for CCH-induced neurodegeneration. Interestingly, CCH-induced neurodegeneration is associated with tau hyperphosphorylation (Zhao et al. 2014), suggesting that the CCH-induced neuronal degeneration might be caused by or associated with abnormal hyperphosphorylation of tau.

CCH-Induced White Matter Lesion and Glial Activation

White matter damage and glial cell activation are seen in both human brains with CCH and in the brains of animal models of CCH (Fernando et al. 2006; Scherr et al. 2012; Thiebaut de Schotten et al. 2014). The degree of ischemic damage correlates positively to the degree of white matter lesions (Shibata et al. 2004; Kitaguchi et al. 2009). White matter damage with increased levels of pro-inflammatory cytokines, such as interleukin-1β (IL-1β) and IL-6, and decreased level of anti-inflammatory cytokines, such as IL-4 and IL-10, is seen in the corpus callosum of the mouse brains 30 days after UCCAO (Yoshizaki et al. 2008). In rat models, CCH appears to cause more severe white matter damage in the corpus callosum than in the striatum (Wakita et al. 1994). This finding is consistent with the more glial activation in the corpus callosum under CCH (Yoshizaki et al. 2008). The activation of glial cells is closely associated with white matter damage (Wakita et al. 1994; Shibata et al. 2004; Nakaji et al. 2006) and may worsen white matter lesions (Wakita et al. 1995).

Two possible mechanisms have been considered for white matter lesion induced by CCH. First, chronic cerebral ischemia results in oxidative stress and increases reactive oxygen species, which cause white matter damage. A large number of inflammatory glial cells are activated immediately after white matter damage occurs (Wakita et al. 1994; Petito et al. 1998). Second, CCH causes blood–brain barrier damage that facilitates the entry of inflammatory cells into the brain parenchyma and causes the generation of inflammatory immune response and the release of a large number of serine proteases, matrix metalloproteinase-2, elastase, collagenase, IL-1β, and tumor necrosis factor-α, which in turn lead to white matter damage (Farkas et al. 2005; Crawford et al. 2008).

Animal Models of CCH for the Studies of Neurodegeneration and AD

Several animal models of CCH have been produced by restricting cerebral blood flow in rodents in order to investigate the roles and mechanisms of CCH in cognitive impairment and to evaluate the therapeutic efficacy of potential drugs. The most commonly used model is a rat model with permanent BCCAO/2-vessel occlusion (BCCAO/2-VO) (Pappas et al. 1996; Ji et al. 2010; Shonesy et al. 2012). Because of the bridging blood supply from posterior communicating arteries, an approximately 50 % decrease of frontal cerebral blood supply can be achieved using this procedure (Tanaka et al. 1996). Numerous studies have reported spatial memory deficits in this CCH model (Wang et al. 2010; Shu et al. 2013). Progressive spatial memory deficits as tested by using Morris water maze, decreased synaptic density and alterations of synaptic ultrastructure in the CA1 area of hippocampus, decreased levels of PSD-95 and synaptophysin, and time-dependent accumulation of oligomeric Aβ in the hippocampus are found 30 days after the occlusion surgery (Wang et al. 2010). Deficits of both short-term non-spatial working memory and long-term spatial memory are observed eight weeks after BCCAO (Shu et al. 2013). Neuroinflammation with microglial and astroglial activation and white matter lesions also occur six weeks after BCCAO (Choi et al. 2011).

UCCAO is a procedure used to produce CCH in mice. A 35–55 % decrease of cerebrocortical perfusion was reported in the ipsilateral hemisphere in mice 28 days after UCCAO (Kitagawa et al. 2005). After UCCAO for 2.5 months, these mice develop significant short-term memory deficits and mild long-term spatial memory impairment, as well as decreased level of protein O-GlcNAcylation, increased level of tau phosphorylation, dysregulated synaptic proteins and insulin signaling, and selective neurodegeneration in the brain (Zhao et al. 2014). UCCAO of the transgenic mouse models with APP mutations exacerbates cognition deficits (Yoshizaki et al. 2008; Lee et al. 2011; Pimentel-Coelho et al. 2013).

BCAS is also used for producing CCH in mice. This approach is theoretically better than the two approaches above, but practically, it is more difficult to achieve the same level of cerebral hypoperfusion due to the more challenging technique of BCAS. Mice after BCAS develop learning and memory impairment (Nishio et al. 2010). Proliferation of activated microglia and astroglia is observed in the white matter after 3 days, and white matter lesions occurred after 14 days of BCAS (Shibata et al. 2004). Impaired reference and working memory, as well as hippocampal atrophy with pyknotic and apoptotic cells, is found 8 months after BCAS (Nishio et al. 2010). BCAS induces more severe cognitive impairment in the APPSw/Ind-Tg mice (Shibata et al. 2004; Nakaji et al. 2006; Shibata et al. 2007; Kitaguchi et al. 2009; Nishio et al. 2010). Rarefied white matter, proliferated astroglia, and Aβ1-40 immunoreactivity appear in some axons in the white matter of APP-Tg mice soon after BCAS, whereas Aβ1-42 accumulates later in the scattered cortical neurons and their axons (Kitaguchi et al. 2009). BCAS also exacerbates Aβ aggregation, neuronal loss, and learning impairment in APP-Tg mice (Yamada et al. 2011).

The above approaches are aimed to reduce cerebral perfusion. However, some of the alterations in these animals might be partially caused by other factors associated with or as the consequence of the procedures. Some artery occlusion paradigms are confounded by parallel damage to pyramidal and cholinergic neurons (Volpe et al. 1988; Mizobuchi 1989; Sugai 1989; Volpe et al. 1992; Ni et al. 1997), making it challenging to know whether it is hypoperfusion or the consequent neuronal damage that causes the cognitive dysfunction.

Concluding Remarks

CCH caused by vascular structural lesions, cerebral hemodynamic changes, or increased blood viscosity is common in the elderly and often contributes to memory impairment, neurodegeneration, and sporadic AD (Fig. 1). Because CCH rarely exists alone and is usually accompanied by other brain pathologies, it is challenging to dissect the exact role of CCH in neurodegeneration and sporadic AD. As discussed in this article, many studies have shown an active and even causative role of CCH in Alzheimer-like brain pathology and neurodegeneration. On the other hand, there is evidence showing cerebrovascular lesions in AD brain and transgenic mouse models of AD (Gold et al. 2007; Tang et al. 2009; Austin et al. 2011), which in turn could lead to CCH. It is highly likely that both scenarios are true and may be co-existing. CCH could be a major contributing or even a causative factor for AD in some patients and the consequence of AD in others.

Future research should focus on dissecting the major molecular mechanisms by which CCH causes or promotes cognitive impairment and neurodegeneration. Better-characterized animal models and detailed studies of the time-dependent changes during CCH in the brain will shed new light on the roles and mechanisms of CCH in cognitive impairment and neurodegeneration. Creating CCH in animal models of other conditions of neurodegeneration and memory loss, such as AD mouse models, will help elucidate the contributions of CCH to those disorders. These studies will help identify potential therapeutic targets to prevent and treat cognitive impairment in individuals with CCH.