The blood-brain barrier shields the brain from potentially harmful things, and some findings have suggested that this protective border weakens with age or disease. However, a study published in Neuron on October 21 reports that the barrier remains largely intact in multiple mouse models of neurodegenerative disease. The researchers, led Marcel van der Brug of Genentech in South San Francisco and Ryan Watts, formerly of Genentech, also found that brains from healthy aging people bore the scars of just as many barrier breaches as those from Alzheimer’s disease patients.

Alzforum originally reported the study’s preliminary findings as presented earlier this year at a Keystone Symposium (see Mar 2015 conference news). The full results underscore the need to develop therapies designed to cross an intact blood-brain barrier, first author Nga Bien-Ly told Alzforum.

This study contradicts previous work that has called disruption of the blood-brain barrier (BBB) both a cause and a consequence of AD pathology. Vascular amyloid deposits or other age-related vascular problems may damage the integrity of the brain’s blood vessels, thus allowing the entry of potentially toxic proteins that exacerbate brain pathology (see Erickson and Banks, 2013). Earlier this year, researchers led by Berislav Zlokovic at the University of Southern California in Los Angeles reported that in the hippocampus, but not the cortex, the BBB becomes progressively leakier with age, even more so in people with mild cognitive impairment or AD (see Feb 2015 Webinar). Van der Brug and colleagues wanted to measure whether the barrier, especially in regions other than the hippocampus that may not have extensive neurodegeneration earlier in the disease process, is really compromised in terms of allowing drugs to pass through. Beyond this debate, most researchers agree that even when partially compromised, the BBB poses a formidable challenge to deliver drugs to the brain, especially macromolecules such as antibodies. “This is a huge problem in drug development,” van der Brug told Alzforum. As researchers attempt to create therapeutics targeted at the brain, it will be crucial to better understand how the function of the BBB changes with age and disease.

For the current study, Genentech researchers wanted to formally test the idea that AD disrupts the BBB. The company, now part of Roche, is developing a strategy to smuggle therapeutic antibodies across the brain’s border, hence evidence of an intact barrier would further support the need for such a trafficking route. The researchers previously developed bispecific antibodies, which recognize a different target with each of their two arms. While one arm recognizes BACE1, the other latches on to the transferrin receptor (TfR) expressed on endothelial cells lining the barrier, which then transport the antibody across via transcytosis (see Jan 2014 newsJan 2014 news; and Nov 2014 news). 

Open Access or VIP Only? To assess the integrity of the BBB, researchers compared the active and passive crossing of antibodies, as well as other molecules, in models of neurodegenerative disease. [Image courtesy of Nga et al., Neuron 2015.]

The researchers reasoned that in the face of a barrier breach, monospecific BACE1 or control antibodies would passively cross into the brain nearly as well as bispecific TfR antibodies that were actively smuggled across. They first tested this notion in animals with experimental autoimmune encephalomyelitis (EAE), a model of multiple sclerosis in which the BBB is drastically compromised. They found that both mono- and bispecific antibodies crossed into the brain and spinal cord of EAE mice after injection into the blood, indicating that the barrier was indeed breached. However, bispecific antibodies still entered the brain more efficiently than BACE1 or control antibodies. In normal mice, only the bispecific antibodies made it to the other side efficiently, while monospecific antibodies eked across in barely detectable amounts. In further support of the border breach in EAE mice, the researchers detected serum albumin in the brain.

The researchers next used their antibodies to gauge barrier disruption in multiple models of neurodegeneration. In plaque-ridden, 10- to 13-month-old PS2-APP mice, the researchers found that, like in wild-type mice, only the bispecific TfR antibody crossed the barrier efficiently, while BACE1 or control antibodies remained largely outside. This indicated that a barrier disruption large enough to let antibodies across did not occur in these AD mice. The same held true for two transgenic mouse models expressing disease-associated forms of human tau—P301L and P301S—despite extensive tauopathy and neurodegeneration. Passive passage of antibodies was also nearly absent in ApoE knockouts, ApoE4 knock-ins, and SOD1-G93A animals, a mouse model of amyotrophic lateral sclerosis.

Antibodies are macromolecules, weighing in around 150kDa, so their exclusion from the brain cannot rule out lesser breaches that would let smaller molecules slide through. To look for that, the researchers injected normal and PS2APP mice with radioiodinated molecules ranging from 3 to 150kDa in size. Regardless of their size, most labeled molecules were excluded from the brains of young or aged wild-type or PS2APP mice, indicating that neither age nor amyloid pathology promoted BBB permeability.

In an attempt to decipher whether AD compromises the BBB in humans, the researchers measured the number and size of cerebral infarcts—remnants of past barrier breaches—in postmortem brain samples from 561 AD patients and 227 age-matched controls. People with severe cerebrovascular disease, vascular dementia, or stroke were excluded from the dataset. The researchers found no differences in the number, size, or total volume of cerebral infarcts between AD patients and controls, and the volume of a person’s infarcts did not correlate with the severity of their amyloid pathology. While this finding cannot account for BBB disruptions that were not accompanied by infarcts, they suggest that in people with AD, the integrity of the BBB was similar to that in healthy people of the same age.

“While the studies do not support earlier observations suggesting a widespread ‘opening’ of the BBB, they strengthen the rationale for developing approaches to promote the transfer of drug across the BBB for therapeutic purposes in neurodegenerative diseases,” commented Costantino Iadecola of Weill Medical College of Cornell University in New York. He added that previous studies that concluded the barrier was compromised due to neurodegeneration induced barrier breach under different conditions and used varying approaches to measure it, so it is difficult to reconcile the results. “Irrespective of whether or not the BBB is open in advanced disease (as examined in the present paper), a key question concerns the status of the BBB in the presymptomatic phase of the disease process, because that would be the time to intervene in diseases like AD in which the pathogenic process precedes symptoms by decades,” Iadecola wrote.  

That the BBB remains largely intact across AD models was a welcome result to Jürgen Götz of the University of Queensland in Brisbane, Australia. Götz recently reported that transiently opening the BBB with ultrasound triggers plaque clearance in AD mouse models (see Mar 2015 news). “An argument one often encounters when discussing the potential application of ultrasound for transient BBB opening is that the BBB is compromised in AD,” he wrote. He added that in addition to van der Brug’s new findings, previous studies have reported that in some AD models the BBB is even tighter (see Mehta et al., 2013).

The authors’ conclusion that the blood-brain barrier remains largely intact across models of neurodegenerative disease and in humans with AD contradicts many studies using differing techniques that say otherwise, commented Zlokovic. He believes the best way to measure disruption of the BBB is by visualizing blood vessels in living animals, using techniques such as 2-photon imaging. He also noted that the idea that drugs would be more readily delivered across a compromised barrier is false, as pathological processes that damage the barrier also disrupt active transport mechanisms needed for drug delivery.

What do results from AD mouse models say about the state of the BBB in human disease? Maybe their import is limited because one big difference between animal models and human AD is the presence of cerebral amyloid angiopathy (CAA). In some people, the vascular amyloid deposits of CAA cause vessels to bleed, but most mouse models have no CAA. Van der Brug acknowledged this difference, but he pointed out that his analysis of postmortem AD brains showed no increase in infarcts in people with AD. However, it is important to note that people with vascular disease, who could be more likely to have CAA, were excluded from the study.—Jessica Shugart


  1. This paper provides a comprehensive and quantitative assessment of the permeability of the blood-brain barrier in models of neurodegenerative diseases. Strengths of the assessment include the use of multiple mouse models of neurodegenerative diseases and ApoE status, use of quantitative and sensitive methods to assess transfer from the blood to the brain, use of BBB tracers of different sizes, and positive controls to demonstrate the sensitivity of the methods for detecting a breach in the BBB. While the studies do not support earlier observations suggesting a widespread “opening" of the BBB, they strengthen the rationale for developing approaches to promote the transfer of drug across the BBB for therapeutic purposes in neurodegenerative diseases.

    It remains unclear why the findings contradict previous observations of BBB disruption in neurodegenerative diseases. However, several considerations are in order. First, there have been few well-controlled studies examining permeability in neurodegenerative diseases. Second, existing studies have used different experimental conditions (anesthesia, cardiorespiratory status affecting the cerebral circulation, etc.) that could affect the BBB, which is more dynamic than previously believed. Third, the histological approaches most commonly used in postmortem BBB assessment were, for the most part, indirect and not quantitative and, as such, subject to false positive readings. 

    A well-tested way to resolve the discrepancy is for the investigators who reported conflicting results to perform experiments in each other’s laboratories under identical experimental conditions. Considering the impact of the BBB status in drug delivery to the brain, this is an important issue that would be worth resolving. 

    Irrespective of whether or not the BBB is open in advanced disease (as examined in the present paper), a key question concerns the status of the BBB in the presymptomatic phase of the disease process, because that would be the time to intervene in diseases like AD in which the pathogenic process precedes symptoms by decades.

  2. It is comforting to see that Bien-Ly and colleagues report on the absence of widespread BBB disruption in AD mouse models and that they further find that the levels of brain infarcts are similar in AD cases and healthy controls.

    We have recently become interested in transiently opening the BBB to clear amyloid from the brain of amyloid-depositing APP23 mice and restore memory functions using ultrasound (Leinenga et al., 2015).

    An argument one often encounters when discussing the potential application of ultrasound for transient BBB opening (an area pioneered by Drs. Hynynen and Konofagou, refs: Choi et al., 2011; Hynynen et al, 2005) is that the BBB is compromised in AD. The notion that in neurological disease “the BBB is leaky” disguises the fact that the brain is not completely sealed off from the periphery. For example, steady-state central nervous system levels of a peripherally administered anti-Aβ monoclonal antibody are approximately 0.1 percent of the level found in the plasma. This indicates that an active exchange takes place between the blood and the brain even under physiological conditions, i.e., when the BBB is neither manipulated nor damaged (Levites et al., 2006). 

    Bien-Ly and colleagues find that there is no increased passage of antibodies in multiple AD mouse models. Interestingly, some studies even indicate that rather than being more open, the BBB in AD mouse models is even tighter. Studies in AD mouse models indicate that the uptake of therapeutic drugs is significantly reduced in 3xTg mice that present with both a tau and an Aβ pathology compared to that in age-matched non-transgenic controls (Mehta et al., 2013). 


    . Scanning ultrasound removes amyloid-β and restores memory in an Alzheimer's disease mouse model. Sci Transl Med. 2015 Mar 11;7(278):278ra33. PubMed.

    . Noninvasive and localized neuronal delivery using short ultrasonic pulses and microbubbles. Proc Natl Acad Sci U S A. 2011 Oct 4;108(40):16539-44. Epub 2011 Sep 19 PubMed.

    . Local and reversible blood-brain barrier disruption by noninvasive focused ultrasound at frequencies suitable for trans-skull sonications. Neuroimage. 2005 Jan 1;24(1):12-20. PubMed.

    . Insights into the mechanisms of action of anti-Abeta antibodies in Alzheimer's disease mouse models. FASEB J. 2006 Dec;20(14):2576-8. Epub 2006 Oct 26 PubMed.

    . Altered Brain Uptake of Therapeutics in a Triple Transgenic Mouse Model of Alzheimer's Disease. Pharm Res. 2013 Jun 22; PubMed.

  3. If you weigh 50 kg, you have approximately 60 g of IgG in your blood. Yet, immunohistochemistry of your brain parenchyma using anti-human IgG would generate essentially no immunoreactivity, indicating that only little, if any, IgG crosses the blood-brain barrier.

    This notion, consistent with the data shown in the paper, gives a warning against the use of monovalent (as opposed to divalent) anti-Aβ therapeutic antibodies in the large-scale longitudinal prevention trials mainly going on in the United States. Behind the scientific scene, a number of pharmaceutical companies are making efforts to engineer their antibodies to make them more BBB-permeable. It is also important that the antibodies are administered intravenously in human cases, whereas the injection paradigm for the model mice is almost always intraperitoneal. This difference may explain the discrepancy in the efficacy of immunotherapy between model mice and humans.

    The prevention protocols for the trials may be modified in the near future. One question is “Are we going in the right direction?” because immunotherapy is expensive. Success of immunotherapy will obviously open up more opportunities for the other anti-Aβ medications that would cost society and individual people less.

    A small concern about this paper is the Tg mouse model of AD used, which overexpresses mutant APP and PS. Overexpression of membrane proteins containing 1 and 9 TM domains, respectively, could induce non-specific ER stress. Overexpression of APP and PS transgenes results in the destruction of at least two gene loci in the host animals and may in principle generate artificial phenotypes.

    It is surprising that none of the Tg mice, to my knowledge, have been sequenced for defects. APP interacts with kinesin via JIP-1, hence APP overexpression may perturb axonal transport, causing axonal sprouting. In addition, APP overexpression also results in overproduction of non-Aβ APP fragments. In particular, CTF-β, which does not accumulate much in human AD brain and is recognized by most anti-Aβ antibodies, is known to be more toxic than Aβ.

    Based on these concerns, I would suggest that the authors examine whether their main findings are the result of APP and PS overexpression and not innate disease processes. We are willing to provide free to academic scientists our model mice that overproduce Aβ42 without overexpressing APP or PS1 (Saito et al., 2014). Right now, more than 130 laboratories all over the world are using these mice. We estimate that approximately 60 percent of the phenotypes in conventional APP Tg mice are artifacts. I invite the authors and other scientists experimentally working on AD in transgenic mouse models to recognize this and take steps to correct the APP overexpression paradigm.


    . Single App knock-in mouse models of Alzheimer's disease. Nat Neurosci. 2014 May;17(5):661-3. Epub 2014 Apr 13 PubMed.

  4. The property of a therapeutic agent being able to penetrate the BBB is generally regarded as a prerequisite for anti-AD agents. Recently we found that 40 percent of the Aβ in the brain is catabolized by the peripheral organs and tissues, in particular, in liver, kidney, gastrointestinal tract and skin (Xiang et al., 2015). This finding suggests that drugs that directly act on Aβ in the periphery would have a therapeutic significance even though they do not pass through BBB. Indeed, peripheral administration of a single chain antibody (scFv) to Aβ is as effective as intracranial administration of the scFv in reducing brain Aβ burden (Wang et al., 2009). Enhancement of Aβ degradation in liver by Withania somnifera extracts significantly reduced Aβ levels in the brain (Sehgal et al., 2012). Continuous peripheral expression of the NEP gene in skeletal muscle is able to reduce brain Aβ burden (Guan et al., 2009; Liu et al., 2009Liu et al., 2010). Thus drug development against Aβ in the future can focus on the clearance of Aβ from the circulation and might be a promising therapeutic approach for AD.


    . Peripherally expressed neprilysin reduces brain amyloid burden: a novel approach for treating Alzheimer's disease. J Neurosci Res. 2009 May 1;87(6):1462-73. PubMed.

    . Expression of neprilysin in skeletal muscle reduces amyloid burden in a transgenic mouse model of Alzheimer disease. Mol Ther. 2009 Aug;17(8):1381-6. PubMed.

    . Circulating neprilysin clears brain amyloid. Mol Cell Neurosci. 2010 Oct;45(2):101-7. PubMed.

    . Withania somnifera reverses Alzheimer's disease pathology by enhancing low-density lipoprotein receptor-related protein in liver. Proc Natl Acad Sci U S A. 2012 Feb 28;109(9):3510-5. Epub 2012 Jan 30 PubMed.

    . Intramuscular delivery of a single chain antibody gene reduces brain Abeta burden in a mouse model of Alzheimer's disease. Neurobiol Aging. 2009 Mar;30(3):364-76. PubMed.

    . Physiological amyloid-beta clearance in the periphery and its therapeutic potential for Alzheimer's disease. Acta Neuropathol. 2015 Oct;130(4):487-99. Epub 2015 Sep 12 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Systemic Inflammation: A Driver of Neurodegenerative Disease?
  2. Brain Shuttle Ferries Antibodies Across the Blood-Brain Barrier
  3. Less Is More: High-Affinity Antibodies Block Blood-Brain Barrier Conduit
  4. Antibody Ferry Looks Safe in Monkeys, Charts Course for Human Studies
  5. Stop, Hey, What’s That Sound? ... Amyloid Is Going Down?

Webinar Citations

  1. Leaky Blood-Brain Barrier a Harbinger of Alzheimer's?

Research Models Citations

  1. PS2APP
  2. Tau P301L
  3. hTau.P301S
  4. APOE Knock-out
  5. APOE4 Knock-In (Lamb)

Paper Citations

  1. . Blood-brain barrier dysfunction as a cause and consequence of Alzheimer's disease. J Cereb Blood Flow Metab. 2013 Oct;33(10):1500-13. PubMed.
  2. . Altered Brain Uptake of Therapeutics in a Triple Transgenic Mouse Model of Alzheimer's Disease. Pharm Res. 2013 Jun 22; PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Lack of Widespread BBB Disruption in Alzheimer's Disease Models: Focus on Therapeutic Antibodies. Neuron. 2015 Oct 21;88(2):289-97. PubMed.