The picture is far from clear, and what stood out at the meeting was a realization of how much we still have to learn about how vascular biology and neurobiology intersect in the brain, especially where neurodegeneration is concerned. It has long been recognized that vascular function decreases early on in AD, where there is reduction in both resting and induced blood flow. A deficit in blood flow of just 20 percent compromises protein synthesis, which can impair learning and memory. In the extreme, low blood flow leads to ischemia, which is directly toxic to neurons. Insufficient blood supply can also delay clearance of amyloid-β (Aβ) from the brain. These observations raised both familiar questions, e.g., Is the loss of vascular function a cause or an effect of pathological processes, or both? and new ones, e.g., Does the vasculature hold new targets for treating AD?

What follows is a recap of selected presentations from a varied slate of speakers. As always, we encourage attendees to chime in with additional content or comments on these or other talks.

Berislav Zlokovic of the University of Rochester, New York, has done as much as anyone to focus attention on the cerebrovascular system, and in particular the blood-brain barrier, in AD. The complexity of the problem became apparent with his introduction of the concept of the neurovascular unit. The old idea, that the blood-brain interface consists of specialized endothelial cells, has given way to a new model that includes multiple cell types—microglia, astrocytes, pericytes, neurons, vascular smooth muscle cells, as well as endothelial cells—that work together to regulate vascular and neuronal function. This complicated interface is necessary to protect the brain’s delicate microenvironment and to provide for the prodigious energy requirements of the organ. (The brain receives 20 percent of cardiac output despite accounting for only 2 percent of body weight.)

The neurovascular unit provides amyloid-β with multiple points of attack, and Zlokovic and subsequent speakers ranged widely over the possibilities. His own work has highlighted the contribution of blood-brain barrier (BBB) transport receptors to controlling brain Aβ levels (see ARF related news story). Zlokovic’s group has also looked extensively at gene expression in vascular smooth muscle cells from AD patients, where the scientists found that overexpression of two transcription factors makes vessels more contractile, leading to constricted blood flow (see ARF related news story). The researchers have also found changes in gene expression in endothelial cells from AD patients that are tied to impaired angiogenesis (see ARF related news story).

Changes in the function of the BBB may come into play in other neurodegenerative diseases, as well. Zlokovic showed preliminary evidence for BBB breakdown as an early event in motor neuron disease, using the SOD1 mutant mouse model of ALS.

Ben Barres of Stanford University, Palo Alto, California, reported that pericytes are critical for the formation of the BBB early in development. In a separate study, his group’s work on expression profiling of endothelial cells in the mature BBB revealed that hemoglobin is a prominent transcript. The function of the protein in this instance is unclear. Barres speculated that it could serve as an oxygen reservoir, a carrier or sensor for the brain or endothelial cells, or as a barrier to CO, NO, or other toxins. Interestingly, Barres showed that a 30-minute episode of hypoxia breaks down the BBB, leading him to wonder if one function of the BBB is to serve as a barrier to oxygen. Barres also noted that hemoglobin expression is upregulated in the aging brain, and that his lab is now trying to make mouse conditional hemoglobin knockouts to test the role of the protein at the BBB.

Aβ’s Predilection for Vasculature
The penchant for amyloid-β to attack the circulation is most apparent in cerebral amyloid angiopathies, where certain vasculotropic forms of Aβ (mostly Aβ1-40 and selected mutants, as summarized by Jorge Ghiso of New York University) accumulate preferentially in blood vessels. Following early deposition on connective tissue surrounding blood vessels, the amyloid eventually kills vascular smooth muscle cells and takes their place. In extreme cases, the vessel wall eventually disintegrates, leading to rupture and cerebral hemorrhage. Other sequelae can include microhemorrhage, microinfarcts, or partial ischemia around CAA-laden vessels. Several talks dealt with CAA as a component of AD, with attention given to the inflammatory consequences of amyloid deposition (see ARF related news story).

CAA occurs on its own in some hereditary diseases (see ARF conference report), but also is seen with high frequency in AD. Steven Greenberg of Massachusetts General Hospital in Boston estimated that 80-100 percent of AD patients show some CAA, and 25 percent show moderate to severe pathology. Between 15 and 29 percent of AD patients have micro-bleeds that can be detected by MRI. These bleeds are almost identical to those that are seen in patients with CAA only, Greenberg said, suggesting that in many cases, AD patients may also be suffering the effects of CAA pathology.

Greenberg showed brain imaging of CAA compared to AD pathology using the PET tracer Pittsburgh Compound B (PIB). All of the non-demented CAA patients Greenberg tested were PIB-positive (see also Bacskai et al., 2007). Global PIB retention levels fell within the range seen in AD patients, but the distribution pattern differed slightly. By taking the ratio of occipital lobe to global retention, Greenberg was able to separate the CAA patients (higher ratio) from AD patients (lower ratio). Greenberg suggested that the occipital preference for PIB retention in CAA could be a signature of that disease.

Beyond microhemorrhages, vascular amyloid activates other pathological processes, as well. Greenberg showed one patient with CAA, who appeared to have a relapsing inflammatory edema that correlated with the person’s cognitive symptoms, and reversed with anti-inflammatory drugs. The appearance and clinical course was similar to what was observed in patients with Aβ-vaccine-related inflammation, which might fit with the idea that clearance of parenchymal Aβ upon vaccination leads to the formation of CAA and inflammation. (This patient was not a participant of the AN-1792 vaccine trial.)

Joseph El Khoury, also at the Massachusetts General Hospital, described his new studies on microglial activation and clearance of Aβ. Last year, El Khoury’s group showed that stopping monocyte recruitment to brain by knocking out the CCR2 chemokine gene resulted in early death and worse amyloid pathology in Tg2576 mice (see ARF related news story). The data suggest that Aβ deposits activate local microglia, which then produce CCR2 to recruit monocytes from the blood to clear Aβ. From these results, El Khoury concluded that microglia are protective early in the disease process. If they are removed, then deposition accelerates, especially in areas around the blood vessel wall. According to El Khoury, the situation may parallel atherosclerosis, where monocytes enter the vessel wall to clear cholesterol deposits.

If microglia are capable of clearing amyloid, why does it deposit at all? To answer that question, El Khoury and colleagues isolated microglia from adult PS1/APP mice and control littermates. The researchers found that in microglia from the AD mice, expression of the amyloid-degrading enzymes—insulin degrading enzyme (IDE), neprilysin, and matrix metalloprotease 9—all decrease with age. Microglia from control littermates show no such change. Monocytes from aged PS1/APP mice display decreased Aβ binding compared to wild-type mice, and express higher levels of interleukin-1 and TNF. TNF, it turns out, downregulates the expression of IDE, as well as the scavenger receptors Rage, SRA, and CD36, the last of which mediates microglial activation by Aβ. From these results, El Khoury suggests that chronic activation of microglia by Aβ may reduce their ability to clear amyloid.

William Van Nostrand of Stony Brook University, New York, also showed data on inflammation. In van Nostrand’s studies, reducing inflammation with the antibiotic drug minocycline improved behavior in the TgSWDI mouse model of CAA (see ARF related news story), leading him to conclude that reducing amyloid-associated cardiovascular neuroinflammation may improve cognitive deficits.

Costantino Iadecola of Weill Cornell Medical College in New York, described studies on how Aβ regulates the neurovascular unit. Iadecola considered two levels of regulation—functional hyperemia, where neurons signal to blood vessels to increase flow to active regions (this is the basis of the BOLD signal in fMRI), and autoregulation, where cerebral blood vessels maintain constant pressure by contracting or relaxing in the face of changes in arterial pressure.

Iadecola measured functional hyperemia in the mouse whisker barrel cortex, where tickling the whiskers causes a measurable increase in blood flow to a defined area of the sensory cortex. In normal animals, tweaking the whiskers increased blood flow by 30 percent, but this increase was absent in APP-expressing mice.

It had previously been shown that Aβ could cause acute constriction of blood vessels, possibly interfering with autoregulation. When Iadecola applied the peptide to the surface of the mouse cerebral cortex, it decreased resting cerebral blood flow and diminished the increase after whisker stimulation. This effect was only seen with the vasculotropic Aβ40 peptide, not with Aβ42. It seemed to stem from induction of oxidative stress, because when APP mice (Tg2576 strain) were crossed with an NADPH oxidase 2 knockout strain, the resulting progeny, with their lower ROS production, showed a normal response to whisker stimulation, despite having the same levels of Aβ.

Decreased blood flow might trigger pathology by increasing susceptibility to ischemic and other injuries. It has been known for a decade that APP-transgenic mice suffer more severe brain injury when blood flow is interrupted. In this sense, the traditional view of two pathologies—a vascular dementia versus a dementia of the Alzheimer type—is outmoded, said several speakers, including Harry Vinters of the University of California at Los Angeles. A more up-to-date view considers these two factors as synergistic, with atherosclerosis promoting Aβ production and retention, and elevated Aβ promoting vessel disease.

Plasticity—It’s Not Just for Neurons
In contrast to Aβ’s acute effects on blood flow, some studies show longer-term effects on vascular remodeling. Michael Mullan of the Roskamp Institute in Sarasota, Florida, is studying anti-angiogenic effects of Aβ. Mullan reported that transgenic APPSw mice show decreased cerebral blood flow at all ages, and decreased vascular density early in life. Cultured aorta pieces from these mice showed impaired angiogenesis (Paris et al, 2004). Mullan has described direct anti-angiogenic effects of Aβ peptides in both in-vitro and in-vivo assays, including inhibition of tumor growth (Paris et al., 2004). He found that soluble Aβ aggregates were potently anti-angiogenic, while fibrillar Aβ40 peptides had no effect. The implication is that oligomeric Aβ aggregates in effect reduce blood flow in the brain.

A separate model of angiogenesis, hypoxic adaptation, revealed that aging rats lose the ability to remodel their vasculature in response to low oxygen, according to Joseph LaManna of Case Western Reserve University in Cleveland, Ohio. LaManna stressed rats by placing them in 10 percent oxygen, a level equivalent to an elevation of 18,000 feet. (People in the Andes and Tibet live at 14,000 feet.) After 3 weeks, the capillary density in brain increased by about 80 percent, revealing a plasticity that came as a “shock and a surprise” to LaManna. When the rats were re-oxygenated, the number of their capillaries returned to the original level within 3 weeks. In old rats, on the other hand, capillary density did not change under low-oxygen conditions.

The capillary increase is mediated by the action of hypoxia-inducible transcription factor (HIF1), a normally short-lived protein whose degradation slows when oxygen levels fall. As HIF1 accumulates, it activates genes that trigger angiogenesis, most prominently vascular endothelial growth factor. In old rats subjected to hypoxia, HIF1 did not accumulate, nor were hypoxia-induced genes activated (Dore-Duffy and LaManna, 2007). HIF1 is expressed and can be pharmacologically activated, but the brain appears to have lost its oxygen-sensing capability, said LaManna.

Therapeutic Options
Because atherosclerosis and AD share many features, could research on new treatments for heart disease benefit AD? That is a question posed by Samuel Wright, Merck’s head of cardiovascular and atherosclerosis research. Both diseases are huge public health problems that progress silently over decades and share similar risk factors. Both involve cholesterol-containing plaques of complex composition, both feature competing concepts of pathology, and for both diseases, homocysteine levels are a strong independent risk factor.

The cholesterol connection led to clinical trials of statins, which have not yet shown unequivocal positive results (see ARF related news story; Riekse et al., 2006; Sparks et al., 2006). The next big thing for heart disease, Wright opined, will be treatments for metabolic syndrome, a constellation of risk factors that occur in one person and appear to mark an underlying metabolic dysfunction. They include abdominal obesity, high blood pressure, insulin resistance, dyslipidemia, and a high risk for cardiovascular disease. Forty percent of Americans older than 60 fit the definition of metabolic syndrome, which is associated with an increased risk of age-related cognitive decline (see ARF related news story).

Merck and other companies are working on potential treatments for metabolic syndrome that target the production of cortisol by the enzyme 11-βhydroxy steroid dehydrogenase (11-βHSD). This enzyme regulates local transformation of inactive cortisone to cortisol in adipose and other tissues, thereby raising intracellular glucocorticoid levels. An inhibitor of HSD developed by Merck reverses the features of metabolic syndrome and atherosclerosis in mice (Hermanowski-Vosatka et al., 2005). Some recent evidence for the importance of HSD in cognitive decline comes from studies showing that aged HSD knockout mice maintain better hippocampal function than aged wild-type mice (Yau et al., 2007..The aged knockouts perform better in the Morris water maze and maintain a youthful capacity for synaptic potentiation.

Consistent with the HSD knockout results, Wright showed that Merck’s HSD inhibitor improved cognitive function in aged Balb C mice, as measured by a novel object recognition test. By 14 months of age, mice had largely lost their ability to distinguish between previously encountered objects and new ones. However, if the mice received an HSD inhibitor in their food for 3 months or more before the recognition test, the response to novel objects improved and more closely resembled that of younger animals.

Merck is developing an HSD inhibitor for cardiovascular disease. Opportunities to expand testing of this and other novel therapeutics to the arena of cognitive decline may depend on the discovery of biomarkers, Wright said.

Ajay Verma, also of Merck, ended the symposium with a talk on a new kind of biomarker of vascular function, a method for non-invasively measuring cerebral blood flow in the clinic. He described transcranial Doppler of the circle of Willis, an arterial structure at the base of the brain, whose branches feed into areas important for cognition. While readings show individual variation, Merck has made a normative database of more than 1,000 people. The scientists can take readings from 18 vessels and compare them to the database to get what Verma called a “footprint” of vascular health. So far, the researchers see dramatic differences in most vessels in AD versus controls. The next step is to examine patients who are using acetylcholinesterase inhibitors, to see if this relatively simple, non-invasive procedure registers any cholinergic-mediated changes in blood flow.

Just as neurons do not stand alone, this symposium served as a reminder that neither should researchers be segregated based on their affinity for studying neurons or the vasculature. Between sessions, attendees were overheard hatching new collaborations—a breakdown of the blood-brain research barrier that bodes well for a full understanding of Alzheimer disease.—Pat McCaffrey.