“This study is the first to show that altering a known constituent of the autophagy pathway accentuates Alzheimer pathology,” said Ralph Nixon, New York University School of Medicine. Nixon was not involved with the new work but in a recent review (Nixon et al., 2008) champions the idea that impaired intracellular trafficking underlies a host of neurodegenerative disorders, including AD. (See a cartoon showing how beclin 1 fits into the autophagic-lysosomal trafficking scene.)

For Wyss-Coray, the hunch that autophagy—and beclin 1 in particular—could play a role in AD came from a curious source: the Lurcher mouse. Lurcher heterozygotes develop muscle coordination problems due to neurodegeneration in the cerebellum that derives from a spontaneous, semi-dominant glutamate receptor mutation (Zuo et al., 1997). Six years ago, a Neuron paper (Yue et al., 2002) by Nathaniel Heintz and colleagues at Rockefeller University, New York, identified the autophagy regulator beclin 1 as one of two proteins mediating the nervous system defects in Lurcher mice. That paper led Wyss-Coray to literature on an emerging connection between autophagy and neurodegeneration. At the time, he said, this link had not been explored using genetic models.

Intrigued by beclin 1’s involvement in the Lurcher phenotype and wondering whether the autophagy protein might also contribute to AD, Wyss-Coray approached Beth Levine at the University of Texas Southwestern Medical Center, Dallas, about her beclin 1 knockout mice. Levine was not only interested in using her mice to pursue the beclin 1-AD connection but also brought in another collaborator—Scott Small at Columbia University in New York. In microarray analyses, Small and colleagues had found decreased levels of beclin 1 RNA in the brains of AD patients.

The gene array findings spurred Wyss-Coray, first author Fiona Pickford, and colleagues to look at beclin 1 protein levels in AD. Sure enough, beclin 1 was dramatically reduced in AD brain tissue relative to controls. And as luck would have it, Eliezer Masliah, University of California at San Diego, was seeing the same thing in his studies of autophagy in Parkinson disease, for which he was using AD brains as controls. With Masliah joining Levine and Small as collaborators on the JCI work, the stage was set to unravel what now appears to be a much more complicated story. “This paper went through many incarnations,” Wyss-Coray told this reporter.

The JCI work adds a twist to earlier studies by Nixon and colleagues (Yu et al., 2005; Nixon et al., 2005) that were among the first to bring autophagy to the attention of the AD field. Using immunolabeling and electron microscopy to examine brain tissue from AD patients and AD transgenic mice (PS1/APP), the 2005 papers showed that autophagy generates Aβ peptides. They also provided evidence that in AD, autophagy is upregulated and impaired in late steps of the pathway involving clearance of partially digested material. Autophagy goes into overdrive not only in AD but also in Parkinson and Huntington diseases—presumably to handle the buildup of toxic protein aggregates in these and other neurodegenerative disorders. In line with this idea, beclin 1 levels rise after traumatic brain injury (Erlich et al., 2006).

Based on studies by Nixon and others, the prevailing thought in the AD field has been that too much autophagy leads to more Aβ. So one might predict that mice with reduced levels of beclin 1, which plays a key role in activating the autophagic pathway, might actually make less Aβ. “But we say that a deficiency in beclin 1 is also bad,” Wyss-Coray told ARF.

In the new study, co-author Small looked at brain samples from AD patients, and showed that beclin 1 mRNA levels in the entorhinal cortex (EC), which is vulnerable in AD, were reduced about 50 percent relative to non-demented subjects. RNA levels were normalized to those in the dentate gyrus, a region usually spared of AD-related damage.

At the protein level, Pickford and colleagues looked in autopsy-confirmed brains of people with AD, mild cognitive impairment (MCI), Lewy body variants of AD (LBV), or Huntington disease (HD) and compared them with age-matched healthy controls. Beclin 1 levels dropped in AD patients to 30 percent of that seen in controls and, in MCI patients, to 70 percent of control levels, the researchers found. On the whole, beclin 1 protein levels were unchanged in LBV and HD patients relative to controls. Some could argue that AD and MCI patients had lower beclin 1 protein levels because their neurons were dying. The authors discount this claim with Western data showing that levels of the neuronal marker neuron-specific enolase (NSE) remained constant among all samples regardless of disease condition.

To address whether beclin 1 reduction contributes to AD or simply occurs as a result of AD pathology, the scientists looked at beclin 1 protein levels in two different lines of very old (24 and 34 months) transgenic mice expressing high levels of mutant human APP. Compared with non-transgenic controls, beclin 1 levels were not reduced in either transgenic line, leaving open the possibility that beclin 1 reduction might occur upstream of AD pathology in the disease process.

In histological studies with APP transgenic mice on a beclin 1 heterozygote background (APP+Becn+/-), extracellular Aβ deposits were found to be nearly double that of littermate controls. Using differently aged mice with varying levels of AD pathology, the researchers showed that beclin 1 protein levels in the neocortex correlated inversely with soluble Aβ levels. Intracellular Aβ was also increased in APP+Becn+/- mice, co-localizing in part with cathepsin D, a marker for lysosomes and mature autophagosomes.

As previous studies have shown that knocking out two key autophagy genes (Atg5 or Atg7) in neurons triggered striking neurodegeneration (see Hara et al., 2006; Komatsu et al., 2006; ARF related news story), the authors wondered whether beclin 1 deficiency would exacerbate the neurodegeneration in AD mice. The answer was a resounding yes, as demonstrated in electron microscopy experiments by co-author Masliah and colleagues. The researchers saw starkly decreased immunoreactivity with three markers of synaptic integrity (synaptophysin, MAP-2, and calbindin) in APP+Becn+/- frontal cortical neurons, and even in Becn+/- mice without human APP overexpression. Electron microscopy of APP+Becn+/- neurons also revealed accumulation of abnormal, enlarged lysosomes—a feature not seen in APP+ mice with wild-type beclin 1 expression—as well as a striking array of intracellular structural abnormalities.

Finally, the authors showed that intracellular and extracellular Aβ could be reduced about twofold with beclin 1-expressing lentivirus injected into the frontal cortex and hippocampus of APP+ mice. This finding seems to jibe with previous work showing that stimulating autophagy can delay Huntington’s (see ARF related news story and ARF news story).

While the JCI work has further established the relevance of autophagy to AD, it has also raised a flurry of new questions about the cell biology of autophagic-lysosomal trafficking. Sorting out the what, how, and why behind the enlarged lysosomes in beclin-deficient neurons is one subject of ongoing investigation, Wyss-Coray told ARF. “Are they getting too much material from the autophagosomal pathway? Or is it that some of the proteins necessary for degradation are delivered through phagosomes, and maybe there’s a defect there?” he asked. “Lysosomes ultimately need to degrade what autophagosomes deliver. If they're not functioning, then you have more material accumulating in the cells.”

Even as the underlying molecular mechanisms remain mysterious, “the finding that beclin 1 overexpression reduces both the intracellular level and the extracellular deposition of Aβ in a mouse model of AD is exciting,” write Jin-A Lee and Fen-Biao Gao, University of California at San Francisco, in a commentary accompanying the JCI paper. “If this finding can be further confirmed by others in different cellular and animal models of AD, manipulating beclin 1 activity in combination with other interventions may be an attractive therapeutic approach for AD patients.”—Esther Landhuis.

References:
Pickford F, Masliah E, Britschgi M, Lucin K, Narasimhan R, Jaeger PA, Small S, Spencer B, Rockenstein E, Levine B, Wyss-Coray T. The autophagy-related protein beclin 1 shows reduced expression in early Alzheimer disease and regulates amyloid β accumulation in mice. J Clin Invest. 2008 May 22. [Epub ahead of print] PMID: Abstract

Lee JA, Gao FB. Regulation of Abeta pathology by beclin 1: a protective role for autophagy? J Clin Invest. 2008 May 22. [Epub ahead of print] PMID: Abstract