Superoxides eat your brain

But that’s a good thing.  Well, probably.  Autophagy — cellular-level self-eating — appears to be misregulated in many neurodegenerative diseases.  A new study from Junying Yuan’s group (Lipinski et al. 2010 Genome-wide analysis reveals mechanisms modulating autophagy in normal brain aging and in Alzheimer’s disease. Proc Natl Acad Sci U S A. 107 14164-9 PMID: 20660724) shows that reactive oxygen species such as superoxide are important signals that help control the induction of autophagy.  What’s more, they show that autophagy is upregulated in Alzheimer’s disease and down-regulated in normal aging, and they provide a list of candidate genes that could be targets for drugs intended to increase or decrease autophagy without triggering cell death.

Autophagy is a routine process in most cell types, and helps to recycle misfolded proteins and other cellular components, including whole organelles.  It’s upregulated in starvation, which makes sense, and in some circumstances can lead to a form of programmed cell death.  Blocking autophagy in the central nervous system can, all by itself, cause protein aggregation in neurons and lead to neurodegenerative disease; and so the idea is that when for some reason misfolded proteins start to accumulate autophagy is needed to clear them away and keep your brain healthy.  If the autophagy system gets overloaded, or if it’s impaired for some reason, then you get neurodegeneration.

In this paper Lipinski et al. are following up on an earlier screen they did to look for genes that are involved in regulating autophagy, under normal nutritional conditions.  They used a genome-wide human siRNA library, made available by ICCB-Longwood, to look for genes whose knockdown changes the localization of an autophagy-specific marker.  They got several genes that are involved in the generation or removal of reactive oxygen species (a catch-all term that includes superoxide and hydrogen peroxide), including the enzyme superoxide dismutase (SOD), which catalyzes the removal of superoxide.  Now, this is interesting, because SOD1 mutations are known to cause amyotrophic lateral sclerosis (ALS), the disease formerly known as Lou Gehrig’s disease.  To find out whether reactive oxygen species are able to induce autophagy directly, they used siRNA to knock down SOD1 in the presence of an antioxidant.  Without the antioxidant, knocking down SOD1 gave clearly higher levels of autophagy, but the antioxidant blocked much of this increase.

So, downregulating the enzyme that removes superoxide increases autophagy, while using an antioxidant to (presumably) remove superoxide decreases autophagy.  We may have a situation here where superoxide (or something related) is being used as a signal to set off other pathways, just as hydrogen peroxide is used to signal to leukocytes after wounding.  When Lipinski et al. re-tested the hits from their screen, genes for which siRNA knock-down leads to increased autophagy, they find that 54% of them — that’s 117 genes — no longer induce autophagy if an antioxidant is present.  So about half of the genes they found seem to be regulating autophagy by affecting the response to reactive oxygen species, but the rest may be acting through a different pathway.

Lipinski et al. then looked at how the genes they found in their screen were expressed in diseased or normal aging brains.  They see reduced expression of the autophagy-related genes in brains from Alzheimer’s disease (AD) patients relative to age-matched controls.  These are genes whose knock-down increased autophagy, remember.  [I kept having to remind myself of which way was up while reading this paper.]  So the fact that expression is down in AD means that autophagy should be up in these patients.  In a separate experiment, they found that the genes negatively regulating autophagy show higher expression in old brains relative to young brains — so as you get older, this self-clearance function tends to go down.  Unless you have AD.  Clear?

To look more closely at the mechanism of the increase in autophagy in AD, the authors looked at the effect of overexpression of amyloid β, the main pathogenic factor in AD, in a cell line.  The results are complicated:  the first steps in autophagy are upregulated, but the final steps, in which the aggregated proteins contained in autophagic vesicles are degraded in the lysosome, seem to be inhibited.  This means that the autophagic vesicles get stuck, and accumulate in the cell.  So there is more attempted autophagy, but (perhaps) less actual clearance of aggregated protein.  Antioxidant can block the increase in attempted autophagy, so this increase appears to be the result of a signal from reactive oxygen species, but it has no effect on the frustrating reduction in lysosomal clearance.

So amyloid β  has two distinct effects on autophagy, increasing signaling through reactive oxygen species and blocking lysosomal clearance.  Possibly the cell is attempting to upregulate autophagy in order to get rid of the amyloid β aggregates.  This is a pretty cool discovery, because the increase in reactive oxygen species and the increase in the number of autophagic vesicles were both already known to be early signs of developing AD.  Now we can see that these are two parts of a single mechanism, and we can link both of them directly to amyloid β.

But is neurodegenerative disease a problem of too much autophagy or too little?  On the one hand, as noted above, blocking autophagy can, by itself, cause neurodegeneration.  On the other hand, the evidence in AD seems to point to upregulation of autophagy, not downregulation. One further line of evidence that life may be complicated: Lipinski et al. tested the effects of two candidate anti-AD drugs in their model system, and found that both of them decreased autophagy.  Of course, as they point out, this could be a coincidence; nobody was screening for autophagy when these drugs were chosen.  On the other hand, perhaps the lysosomal blockage caused by amyloid beta is causing such trouble that the down-regulation caused by the drug actually helps, by reducing the attempted flux through the pathway, even though the cell is desperate to get rid of the protein aggregates.

This seems unlikely to be a clearcut case of “autophagy, good; no autophagy, bad”.  My guess is that in drug development we’re going to be looking at a careful quantitative balance between different processes — although, who knows, we could be lucky. But Lipinski et al. seem to feel the need to hedge their bets: they provide a list of 26 genes you might target should you wish to upregulate autophagy, and a list of 9 genes that you might target should you wish to downregulate it.  You get to choose.  Choose wisely.

Lipinski MM, Zheng B, Lu T, Yan Z, Py BF, Ng A, Xavier RJ, Li C, Yankner BA, Scherzer CR, & Yuan J (2010). Genome-wide analysis reveals mechanisms modulating autophagy in normal brain aging and in Alzheimer’s disease. Proceedings of the National Academy of Sciences of the United States of America, 107 (32), 14164-9 PMID: 20660724