Part two of a four-part series.
Besides hypertension (see part one of this series), a second source of neurovascular damage in the aging human brain took center stage at the inaugural Zilkha Symposium on Alzheimer’s Disease and Related Disorders, held April 4 at the University of Southern California, Los Angeles. This is the leakage of toxic serum proteins into the parenchyma due to a breach of the blood-brain barrier. Whether this is central to Alzheimer’s disease has long been controversial among scientists. About a dozen published studies have documented damage to the vascular system in postmortem Alzheimer’s brain; however, it is unclear what happens to the neurovascular unit years earlier, and whether changes there are an early driver or a late casualty of the disease process.
“I believe toxic blood-derived factors are an important cause of neuronal dysfunction. To really show that, we must develop molecular and imaging biomarkers that detect malfunction of the barrier and then embed these markers in cohort studies of early disease,” Berislav Zlokovic of USC told Alzforum. Zlokovic co-organized the Zilkha symposium.
Some of that biomarker development is starting to happen. In collaboration with the Alzheimer’s Disease Research Center at Washington University, St. Louis, USC researchers have tested a panel of vascular injury and repair markers in some 100 cerebrospinal fluid samples. The panel includes indicators of endothelial and pericyte damage, as well as angiogenic growth factors expressed in response. This panel shows changes in CSF early on, before tau and other markers of neurodegeneration or inflammation rise, Zlokovic told the audience. “We think that vascular factors can come into the brain silently and destroy it even before there is inflammation and neuronal injury,” he said.
Likewise, brain imaging is beginning to visualize effects of possible blood-brain barrier damage in early AD. For example, a pilot MRI study of cognitively normal older people at USC is showing areas of leakage, particularly in the hippocampus, Zlokovic said. Called dynamic contrast-enhanced (or DCE) MRI, the technique quantifies passage of a contrast agent. DCE MRI is used routinely to image the vasculature of tumors, but it can be adapted to measure the integrity of the cerebrovascular system itself. A 15-minute DCE sequence detecting local areas where the BBB is compromised could be added to the MRI scan already being done in existing human studies, said Art Toga. In May 2013, USC lured Toga, Paul Thompson, plus nearly 100 associated scientists away from the University of California, Los Angeles, in a spectacular cross-town academic raid. At USC, Toga now directs the new Institute for Neuroimaging and Informatics. INI encompasses the Laboratory of Neuro Imaging that Alzheimer’s researchers already know as LONI, as well as the Imaging Genetics Center (IGC), which hosts the imaging genetics consortium ENIGMA headed by Thompson.
In looking at the DCE MRI images, Zlokovic noticed leakage of blood into the hippocampus, that is, gray matter key to learning and memory. He was surprised because white matter was the primary site of BBB lesion in a separate USC imaging study. This study used mice which modeled the age-related loss of pericytes in the brain’s capillary walls that is known from AD postmortem studies. In the mice that lost pericytes, projections going from one side of the brain to the other were degenerating, suggesting functional disconnection.
So which one is most affected, gray matter or white? It could be both, Zlokovic said. The mouse study might show later consequences of a slow, chronic blood-brain barrier injury, in which local dysfunction precedes disruption of white-matter tracts. “Vascular changes basically intoxicate the brain and lead to subsequent losses in connectivity,” Zlokovic said. Both studies need more samples and independent replication to establish a sequence of events in the natural history of Alzheimer’s disease. “It will be very important to reproduce our imaging studies here and in aging and observational studies at other sites,” Zlokovic said.
The imaging data in the pericyte mice are the latest obtained from a previously published model of Alzheimer’s disease. The model simulates a two-hit hypothesis of Alzheimer’s, whereby age-related vascular damage interacts with Aβ to cause neurodegeneration. The researchers bred human mutant APP transgenics to PDGF receptor-β knockout mice. The crosses gradually lose pericytes with age, and as this happens, their blood flow in the brain slows down while their Aβ levels go up. Healthy pericytes ingest and degrade toxic proteins from the brain parenchyma; in the aging brain they can become overwhelmed and die. As they disappear, the wall of the micro-vessels thins even further, making it prone to blood components leaking out.
Recently, Abhay Sagare and others in Zlokovic’s group reported that depleting pericytes in APP transgenic mice worsened their phenotype. Pericyte loss drove up the mice’s brain and interstitial fluid Aβ levels, reduced clearance, and sped up amyloid deposition in the parenchyma and blood vessels. This induced pathological changes in tau and neuronal loss in the cortex and hippocampus, both of which APP transgenics themselves do not develop. “The blood-brain barrier breakdown makes these mice more complete models of AD,” Zlokovic said (Sagare et al., 2013; Bell et al., 2010).
The molecular mechanisms for this are under intense investigation, spurred in part by Alzheimer genetics findings. For example, ApoE4, besides being implicated in Aβ deposition, also has been shown to slow its degradation. Made primarily in astrocytes and released from their end-feet, ApoE may act through the LRP1 receptor on pericytes to suppress cyclophilin A signaling and thus keep a lid on the pro-inflammatory NFκB/MMP9 axis in pericytes. ApoE4 is less efficient at that than ApoE2 or ApoE3, in effect allowing a pro-inflammatory pathway to constantly burn on low heat in the vessel wall (May 2012 news story).
In new data presented at the Zilkha Symposium, Zlokovic noted that the same LRP1 receptor on endothelial cells interacts with the protein encoded by PICALM, an AD risk gene that came out of genome-wide association studies. PICALM is highly expressed in the brain’s blood vessels, Zlokovic said. There, it mediates clathrin-dependent endocytosis of Aβ bound to LRP1 into endothelial cells, and it also directs Aβ’s subsequent transcytosis across those cells to aid its drainage via the blood-brain barrier.
Broadly speaking in Alzheimer’s genetics, the GWAS era is nearing its end. Teased by some scientists as “God, What Awful Science,” genome-wide association studies have been controversial for doing little more than merely pegging genes to AD at high cost. Increasingly, sequencing of exomes or the entire genome of large groups of people is taking over in an effort to finger the actual pathogenic variants in the genes GWAS have flagged. With that, the real work begins, Rudy Tanzi of Massachusetts General Hospital, Charlestown, told the Zilkha audience. One such whole-genome sequencing project, of more than 1,500 samples, has already—even on preliminary analysis—yielded some 100 new mutations. This dataset predicts new functional variants in confirmed AD genes, including presenilins 1 and 2, as well as ApoE, clusterin, and Trem2, Tanzi said while previewing the WGS data.
For all their shortcomings, GWAS did confirm a serviceable list of genes as being implicated in the disease process. The list startled researchers when they saw just how many innate immunity, glial, and vascular genes it included. Together with the discovery of the microglial gene Trem2 by exome sequencing, the GWAS list re-energized neuroinflammation and -vascular research across the board. “We are getting very interested in how some GWAS hits affect the brain’s vascular system,” said Zlokovic. Terrence Town of USC said, “In the mid ’90s, everyone thought inflammation was an epiphenomenon in Alzheimer’s. Look how far we have come. We have a lot of data saying inflammation is deleterious in AD, but there are also good forms of inflammation.”
Trem2 has drawn particular interest in the past two years. This gene is highly expressed in microglia, which patrol the parenchyma in the immediate vicinity of the neurovascular unit. Recently, a genetic imaging study showed that Alzheimer's Disease Neuroimaging Initiative participants who carry a Trem2 pathogenic variant lost brain tissue much faster than non-carriers. “We see a dramatic effect size for Trem2,” USC’s Thompson told the Zilkha symposium audience (Oct 2013 story on Rajagopalan et al., 2013). For more on microglia in the aging brain, see part three of this series.—Gabrielle Strobel
- It’s Not All About You, Neurons. Glia, Blood, Arteries Shine at Symposium
- ApoE4 Makes Blood Vessels Leak, Could Kick Off Brain Damage
- Fall Flurry of Letters Kicks Up Dust Around TREM2
- In Revival of Parabiosis, Young Blood Rejuvenates Aging Microglia, Cognition
Research Models Citations
- Sagare AP, Bell RD, Zhao Z, Ma Q, Winkler EA, Ramanathan A, Zlokovic BV. Pericyte loss influences Alzheimer-like neurodegeneration in mice. Nat Commun. 2013;4:2932. PubMed.
- Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R, Zlokovic BV. Pericytes control key neurovascular functions and neuronal phenotype in the adult brain and during brain aging. Neuron. 2010 Nov 4;68(3):409-27. PubMed.
- Rajagopalan P, Hibar DP, Thompson PM. TREM2 and neurodegenerative disease. N Engl J Med. 2013 Oct 17;369(16):1565-7. PubMed.
No Available Further Reading