Robert Moir is an assistant professor of neurology at Harvard Medical School and Massachusetts General Hospital (Boston). He studies Alzheimer’s disease and other inflammatory conditions characterized by neurodegeneration. His research team has shown that the amyloid beta protein associated with Alzheimer’s “plaque” is a potent antimicrobial peptide. Please read this blog post for more context on this important discovery.
Antimicrobial peptides: “Natural antibiotics” created by the human immune system. They are able to kill a range of bacteria, viruses, fungi, and other pathogens.
Innate immune system: The branch of the immune system that creates antimicrobial peptides. These peptides and other innate immune system cells form the body’s “first line of defense” against infectious agents.
Robert hey! Thanks for taking the time to speak with me. First question: your discovery that amyloid beta is an antimicrobial peptide is HUGE. What made you decide to investigate its possible antimicrobial properties in the first place?
The idea came from a somewhat chance discovery. On Fridays I go through my walkabout time on PubMed (a website that catalogs scientific journal articles). I rove wherever impulse takes me. I read a paper about LL37, a well-described human antimicrobial peptide. It was obvious from this paper that LL37 and amyloid beta share clear similarities: both structural similarities and the ability to form amyloid (protein-like deposits that can accumulate in tissues in certain disease states known as amyloidopathies).
At the same time, my mentor Rudolph Tanzi (he’s in the office next door) had just gotten back results from a screen of Alzheimer’s linked genes. He got the results back literally the same day I read the LL37 study. Well, most of the Alzheimer’s linked genes he characterized were also innate immune system genes. We both looked at one another and thought: “Antimicrobial peptides and the innate immune system are the way to go.” Then we did a bunch of experiments that went nowhere. It turns out that several classical methods used to study antimicrobial peptides don’t work for amyloid beta.
Then I went on vacation with my six year-old son to the White Mountains. James Kirby of Beth Israel Deaconess Medical Center had invited us, and the trip gave me the opportunity to get his feedback. He helped us develop assays to correctly test amyloid-beta’s antimicrobial activity. Key to this correct assay preparation was the fact that amyloid beta had to be in its more oligomeric forms (the biologically active form of the peptide). Once we got these assays right, we immediately found that amyloid-beta had very potent antimicrobial activity. In some cases its ability to target certain microbes is more than 100-fold stronger than penicillin.
Me: Wow very interesting. Can you explain a bit more about what it means for amyloid beta to have “oligomeric forms”?
Amyloid-beta is the chief name for a peptide that self-associates to build multiple differently shaped molecules, each with its own molecular structure and activities. Small proteins that can self-assemble and generate diverse structures like this are sometimes called “lego peptides.” Amyloid-beta assemblies are called “oligomers.” Each particular oligomer structure has unique antimicrobial activity. In this way, amyloid-beta spontaneously generates a population of diverse oligomers able to target a broad spectrum of pathogens, and even microbial toxins released during infection. This turns out to be key for amyloid-beta’s effectiveness as an antimicrobial peptide in the brain. Oligomers can target many different microbes while non-oligomeric amyloid-beta is only effective against a limited range of pathogens. This strategy is so effective that we have not been able to find a pathogen that oligomeric amyloid-beta can’t inhibit to some degree.
Me: So these oligomers are a little like antibodies? In the sense that the immune system creates a range of antibodies in response to an infectious threat, each with the ability to target different species/strains of microbe(s)?
Yes, they are like a cheap man’s antibodies. But antibodies need whole cells in order to be produced and are metabolically complex and expensive to make. Amyloid beta’s ability to form into a wide range of different oligomers is a simpler, metabolically cheaper, and far more ancient immune strategy. Even very ancient animals like jellyfish (which are 500 million year old primitive multicellular organisms) have this ability to create antimicrobial peptides that recombine to increase molecular diversity.
What’s hard to believe is that many researchers still regard amyloid-beta’s generation and activities as an ‘accident’ and continue to develop therapeutic strategies around this idea. This assumption was reasonable back when amyloid-beta was discovered in 1984. The peptide was thought unique to the human brain and believed to be only generated under the disease conditions found in Alzheimer’s. In addition, amyloid-beta is made by cutting the peptide out of a larger precursor protein that is embedded in cell membranes. No other peptide was known to be generated this way, so it was considered something that must only occur under disease conditions.
However, it’s been known for over two decades that amyloid-beta is continually made in the normal brain throughout our lives – and not just the human brain. All vertebrates make amyloid-beta in their brain, most of them the exact same peptide as we make. Furthermore, the mechanism that generates amyloid-beta is now known to be a common cellular pathway that is involved in making many different important normal peptides. But, the assumption that amyloid-beta is junk persists, even though the key assumptions underpinning this idea have been disproved. I still frequently hear the ‘amyloid-beta is junk’ idea justified by an old argument that is demonstrably wrong. It goes something like this: “Since Alzheimer’s occurs in individuals past reproductive age (that is post-menopause) there is no evolutionary pressure to remove it from the genome.”
Sounds plausible, until you start looking beyond the narrow precepts of the Alzheimer’s field and consider amyloid-beta in a broader biological context. Only three species of animals undergo menopause: us, dolphins, and pilot whales. Most other animals reproduce to within 1-2% of when they drop dead – and many social animals live well into old age. And yet, all vertebrates make amyloid-beta, with over 60% making the exact same form of the peptide we do. Moreover, they have been doing this for over hundreds of million of years! (The human amyloid-beta gene is expressed in coelacanths, a family of ‘living fossil’ fish that date back over 400 million years). In fact this data would suggest the exact opposite of the ‘amyloid-beta is junk’ argument – it supports the idea that amyloid-beta contributes to survival fitness throughout life in all vertebrates.
Also, anyone with grandparents knows that they typically play an important part in helping successfully raise children. In biology it’s known as the “grandmother hypothesis.” Simply put, the grandmother hypothesis says that menopause allows aging mothers with an increasingly high risk of death from childbirth to stop direct reproduction and pass on their genetic material to future generations by helping raise their grandchildren. Grandparents are not reproductively irrelevant!! Any gene that intrinsically causes dementia and negatively impacts grandparent survival and their valuable store of accumulated experience, is going to be selected against. The gene is going to change or be eliminated. But amyloid-beta is 100% conserved from 400 million years ago.
What all this is telling us is that amyloid beta must actually be very important. It must be at the heart of important biological processes. Indeed, human amyloid-beta is one of the top most conserved proteins in all of biology (as far as I have been able to ascertain, second only to the protein ubiquitin).
Me: Yes it seems clear that amyloid beta must have an important role in human health and disease. How has your research moved forward based on that possibility?
We set out to test whether amyloid beta has antimicrobial activity. We found that if you add amyloid beta to a microbial broth it will inhibit or straight out kill a range of microbes. We published a paper with those and related results in 2010. It was greeted with mixed reception. Some researchers (mostly young researchers) were enthusiastic about it. Older researchers not so much. From these detractors we got feedback like, “That’s fine, but battery acid can kill organisms in a test tube too.” In other words, they were critical that the experiments were done in a test tube and not in a living organism. For most antimicrobial peptides, demonstrating potent microbial killing activity in a test tube is enough to establish identity (it’s a very difficult activity for a protein to pull off), but it’s not an unreasonable critique. Showing protective activity in a living animal is the gold standard for confirming a protein is an antimicrobial peptide.
So we set out to show that amyloid-beta could have antimicrobial activity in living organisms. We tested its activity in genetically modified mice, fruit flies, nematode worms, and cultured cells. The work took us four years to complete. Our keystone finding was that the ‘plaque’ amyloid-beta generates is as important as the peptide itself. These amyloid-beta “plaques” directly entrap and neutralize microbes. Then, by chemically generating a burst of oxygen radicals – bleach in lay terms – they destroy the trapped pathogen. Just to be on the safe side, the plaque remains intact in the brain, entombing forever any microbe that may have managed to survive. This “entrapment” mechanism is not unique to amyloid beta. For example, alpha-defensin 5 and other human antimicrobial peptides create amyloid structures that function similarly.
So to summarize, expression of amyloid-beta protects cultured cells and nematode worms from lethal infections. In unpublished work we have also confirmed that amyloid-beta also protects against infection in fruit flies. In mice, the most important animal infection model, over-expression of amyloid-beta is protective against bacterial and viral encephalitis (brain inflammation driven by infection). Finally, if you “knock out” mouse (murine) amyloid-beta the mice develop increased susceptibility to encephalitis. So what we now have is convincing data from animal models that that amyloid-beta functions in humans as an antimicrobial peptide.
Me: Has it been hard to advance these findings?
One side of the story has gotten the most attention. If amyloid-beta is an antimicrobial peptide, one plausible inference is that Alzheimer’s is caused by an infection or infectious processes. I’m kind of agnostic about that implication, in that Alzheimer’s could also be a disease where immune pathways have gone wrong. There are a number of examples of ‘sterile inflammatory’ disease in which an immune pathway has become dysregulated and pathological. No pathogen involved.
But having said that, there’s mounting circumstantial evidence suggesting that infection plays a role in the disease’s etiology. The first guy to support that was Alois Alzheimer himself. In the 1970s-80s many Alzheimer’s researchers thought infection played a central role. But the discovery of amyloid-beta in 1984 (ironically) shifted the focus away from infection. Amyloid-beta was assumed to be all bad, bad, bad, and its accumulation blamed for the disease. It’s a simple explanation to a complex problem, which made it attractive. But it’s increasingly at odds with emerging data. The drug company Merck’s last Alzheimer’s drug lowers amyloid beta levels but did not slow the disease.
Amyloid-beta alone does not give you Alzheimer’s disease. You also need inflammation and tauopathy, another pathology in brain. It may still be effective to control amyloid-beta if you get it early enough in the disease process. Controlling amyloid-beta may help slow the cascade of events that promotes the neuroinflammation that ultimately kills neurons in Alzheimer’s. But, the question remains: what is driving the deposition of amyloid plaques in the first place? Old models hold that amyloid-beta does it because it’s catabolic junk with an unfortunate propensity to form functionless plaques that induce inflammation. But…could plaques and neuroinflammation in Alzheimer’s actually be an immune response to a genuine immune challenge from microbes in the brain? Perhaps. If it is, then targeting microbes in the brain may be a better way to go. What is clear is that more data on the role of microbes in Alzheimer’s etiology is needed. At the moment the “amyloid is bad” idea continues to dominate and most academic efforts are still focused on this model.
Interestingly, we may be seeing the beginnings of a shift in the way “Big Pharma” is looking at Alzheimer’s. Their costly drug failures seem to be making them open to exploring alternative models, including the “antimicrobial protection hypothesis of Alzheimer’s” – which is what we are calling this new emerging model of the disease. For example, “Big Pharma” is exploring if neuroinflammation in Alzheimer’s can be dampened down independent of amyloid-beta production or plaque deposition. My colleague Rudy Tanzi has discovered that the gene CD33 is a big on/off switch for immune cells in the brain. Maybe switching the gene off could help patients manage early Alzheimer’s symptoms. Most current first-line anti-inflammatories wouldn’t have the same effect since they tend to target the adaptive immune system. Antimicrobial peptides and amyloid-beta are part of our much more ancient and primitive innate immune system (the “front-line troops” of the immune response). Companies are now pursing CD33 as a possible drug target.
Dr Doo Kim’s lab in research Unit at Massachusetts General Hospital has created a special three-dimentional human neuronal cell culture system that’s dubbed “Alzheimer’s in a dish” by the media. The system helps us scan innate immune system drug candidates. The technology accelerates drug screening and reduces cost more than 10-fold compared to conventional approaches. It should allow new potentially useful drugs to be identified much faster.
Me: Interesting. But I’m confused. Do you think amyloid beta should be removed in patients with Alzheimer’s (despite its antimicrobial activity)?
Maybe not removed, but it’s certainly a good idea to control it. Amyloid beta may be a little like cholesterol – heart attacks are exacerbated if cholesterol is in the wrong place at the wrong time. But if you remove cholesterol completely, other serious health problems arise. So as was done with cholesterol targeting therapies, I think the first goal should be to better understand what amyloid-beta is actually doing in the brain and develop strategies accordingly. That would mean reducing bad effects while preserving amyloid-beta’s role in immunity.
What may be happening is that initially, amyloid-beta rises to do battle in cases of brain microbiome dysbiosis (imbalance). Part of this response is amyloid-beta induced inflammation. But, prolonged activation of innate immune inflammation by amyloid-beta leads to tissue damage and neurodegeneration.
One thing is that any treatment aimed at managing amyloid beta would work best if administered in a preventative fashion. Again it’s a little like cholesterol and statins. After a third heart attack it rarely helps to give a patient statins. You would want to begin treating 10-20 years ahead of time to prevent this. But with amyloid-beta in Alzheimer’s we currently have no effective assays to predict who will get the disease and when to intervene. You can’t give a drug to everyone over 65.
Me: Wouldn’t it make the most sense to just target whatever infection(s) are causing amyloid beta to be produced in the first place?
Well yes, a longterm solution may be vaccination against the microbe giving you trouble. But here’s the thing about that: there are research groups that have been pushing the role of infection in Alzheimer’s for a long time, but different pathogens are identified in their studies. Herpes simplex virus 1 is a common candidate, but also chlamydia pneumonia. This suggests there is no single pathogen driving the illness. Instead, many different bugs may be involved in Alzheimer’s. For example, our studies have found that amyloid-beta has strong antimicrobial activity against the herpes viruses and these viruses are linked to increased plaque deposition, However, herpes virus are only detected in about 60% of Alzheimer’s cases. What about the other 40%?
So it’s clear that we’re not dealing with a “classical” infection.” The findings don’t support Koch’s postulates: the idea that only ONE microbe can cause ONE disease. But we’re moving away from this “classical” infection model. We’ve been taking a close look at what microbes are found in the brain and started what we are calling the ‘Brain Microbiome Project’.
Me: Your lab is also studying the brain microbiome!?
Yes. We’ve been scanning brains as part of collaborative work with Mt. Sinai. And we have found that even “non-sick” humans harbor over 200 organisms in the brain. Those numbers don’t even include the virome (viruses). And we know that a bunch of herpes viruses can also survive in the brain. There’s even vertical transmission of certain viruses in the human genome.
So infection in Alzheimer’s may be similar to what we’re seeing in conditions like Crohn’s disease. In Crohn’s the bowel is disrupted but the entire gut microbiome is involved. It’s not a single pathogen but disruption of a whole microbial community. Within that framework certain key pathogens may “push” the community out of balance and contribute to disease more than others. In Crohn’s it’s where the microbiome meets the innate immune system that things go wrong and host pathologies arise. That’s where the problem in Alzheimer’s may also lie: at the interface between microbes and a foot soldier of innate immunity – amyloid-beta.
Going back to the brain microbiome: the microbes don’t just sit there. They cooperate, they’re competitive, they interact with the host and each other: it’s a true microbiome. What this means for Alzheimer’s is there could be general dysbiosis (imbalance) of the brain microbiome. There’s a normal brain microbiome, but in Alzheimer’s something may go “out of whack” and some of the bugs go bad (they become bad players). It could be compared to ulcers. Ulcer formation is linked to the bacterium H. pylori. But H. pylori is actually also a member of the normal gut microbiome. That means H. pylori contributes to ulcers under certain negative conditions. This is tentative, but in Alzheimer’s maybe the herpes viruses also go out of control. That is the model we are exploring at the moment.
Linked in with that model is the gut microbiome because the gut microbiome and the brain microbiome communicate a lot via the vagus nerve. There’s lots of traffic, with bacteria in the brain/gut talking to one another via this highway all the time. Some products of gut fermentation like Short Chain Fatty Acids (SCFAs) literally travel the Vagus Nerve (physical translocation). Immune cells in the brain need these gut microbial SCFAs to mature correctly. Conversely, certain bacteria in the gut live exclusively off chemicals generated in the brain that are transported to the gut. Vagus Nerve traffic may include bacterial signaling molecules called quorum sensing molecules. In this sense microbes in the gut and microbes in the brain may be “talking,” and possibly reaching decisions about what to do next.
So, what’s going on in the brain can have dramatic effects on the gut. But, the gut microbiome can also affect neurological functioning. For example, the disruption of the gut microbiome is now linked to depression – it’s a two-way axis.
And remember, because amyloid-beta can form a vast number of oligomers, it’s able to react against an large range of pathogens. So right now amyloid beta’s activity leaves open a big question mark as to the exact nature of the infection it may be targeting in Alzheimer’s or related conditions. Which means that vaccination against one pathogen in Alzheimer’s (that I mentioned as a possibility before) might prove too simple an approach. Instead, we may want to ask “Can we modulate disease progression by manipulating the microbiome and/or the gut/brain axis.”
Me: You’ve been studying amyloid beta’s antimicrobial activity against HHV6. Where is that study?
Yes, we’ve tested amyloid beta’s activity against herpes simplex virus 6 (HHV6). There are no good animal models of chronic HHV6 infection, so we are using the “Alzheimer’s in a dish” system to look at this at the moment. Each chip is like a little human brain, that is basically thinking and shooting signals back and forth. This makes the experimental model more human-like and allows for better testing. We found that HHV6 induced a large amyloid burden in the “Alzheimer’s in a dish” system within a day. Key to our findings is that these herpes viruses are “low and slow” microbes whose pathogenic activity ramps up with opportunity. This makes HHV6 extremely effective at seeding amyloid beta deposition over time, even over the course of decades. We hope to get funding from the Cure Alzheimer’s Fund to develop a Alzheimer’s disease mouse that can be infected with human HHV6 (the mice need to be ‘humanized’ for a receptor critical for HHV6 infectivity).
But you can’t formally read about the results of that study at the moment. We submitted a paper with the findings to the journal “Cell.” The journal insisted on publishing an online pre-print before the paper was formally accepted. That pre-print was critiqued by researchers who reject a role for infection in Alzheimer’s etiology – not that our study actually claimed that. It showed strong data that amyloid-beta protects against herpes in the lab. Then, Cell rejected it after 6-weeks. The situation has hurt that study tremendously. While I understand that scientific journals are motivated by a genuine belief that expediting dissemination of new and exciting data is a good thing, I would caution other research teams not to allow a pre-print to go up before the study is accepted for publication. It can go disastrously wrong and it’s not worth it. Not worth it at all. We are submitting the findings to another journal.
Having vented a little on this… I have to say that healthy skepticism is key for advancing science. My problem is not criticism – throughout my career legitimate concerns and criticisms have been invaluable for refining our ideas and showing the way forward. My problem is when strong data with no obvious flaws are rejected out of hand because they do not fit current dogma and are dismissed for perfunctory reasons.
Phew! It’s sure hard to get new findings published. With that in mind, how do you get funding for your research?
NIH funding is hard for us to get – they are somewhat risk adverse and typically fund studies that explore prevailing ideas: “Evolution not revolution.” Fortunately, we are supported by some foundations that are willing to take risks. One is the Cure Alzheimers Fund in Boston, whose founders include successful Venture Capitalists. They are used to high risk ventures. For them, if only 50% of their funded projects have a successful outcome that’s a good result – it’s much better than the success rate among new businesses! I also get funds from Good Ventures, which is part of the Open Philanthropic Project out of California.
Me: Have you read some of the recent studies that have detected a range of fungi in Alzheimer’s brains?
Yes. And amyloid-beta is strongly anti-fungal. But of course there are fungal communities in healthy brains too. The organisms we’re finding in the human brain are incredible. For example there are amoeboid worms! Up to 20% of humans harbor toxoplasmosis (the Toxoplasma gondii parasite) in the brain. We’ve even detected another worm in some of our brain samples that was previously only thought to infect dogs. Remember that the nasal bulb is a primary source of entry for these and other microbes. Interestingly, efferent nerves from the nasal bulb trace straight back to brain areas where amyloid-beta formation starts.
Have you seen the Lund University study showing that PrP (prion) protein is also a potent antimicrobial peptide?
Yes. And we’ve tested PrP’s activity in our own lab. It’s a strong antimicrobial peptide. But it goes beyond that – amyloid creation happens in a range of human inflammatory conditions. Diabetes is an amyloid disease (in both type 1 and type 2 diabetes amylin is created in the pancreas). At high levels, this amylin is toxic to pancreas islet cells and highly pro-inflammatory. But it’s also one of the most potent antimicrobials we have ever tested – it’s able to target most pancreatic pathogens, including E, faecalis, a common cause of pancreatitis. There’s also an amyloid generated in the heart that is linked to heart disease. So the potential importance of recognizing amyloid can play a normal and protective role in immunity has legs well beyond just Alzheimer’s.
Me: Do you talk to other research teams studying these other forms of amyloid?
Yes. We have multiple collaborations going on around the world. In fact I collaborate so often with other labs that I spend half my time on the phone. There’s both a collaborative and competitive atmosphere that helps drive things along. I am delighted we are no longer alone in pursuing this line of investigation.
That being said, most research teams aren’t thinking this way. And for many of these teams, accepting amyloid-beta’s role as an antimicrobial peptide is like trying to turn around an 800 pound gorilla on a dime. It’s a paradigm shift, so it will likely take years for the focus to shift.
For example, when we tried to publish our 2010 paper on amyloid beta’s antimicrobial activity, the peer review process was extremely frustrating. We submitted the paper to the journal “Nature.” From the start the editors seemed hesitant about publishing the findings. Then they consulted with an Alzheimer’s “expert”, who it seems from comments didn’t even bother to read the paper, but rejected it anyway. The most recent paper in 2016 was rejected six times without review. The reviews we did get from one top-tier journal were some of the most appalling I’ve ever read in my life. One reviewer kept asking why we did not see amyloid in our control worms – the ones that DO NOT express amyloid-beta.
Me: That is extremely frustrating. I find it interesting that the scientific community frequently references the story of Barry Marshall and Robin Warren. The two researchers discovered that h.pylori bacteria plays a key role in driving ulcer formation (during a time when ulcers were believed to be caused only by stress and spicy foods). When they first presented these results, they were ignored and in some cases even mocked by other research teams. It took decades for their new finding to be taken seriously. Then, in 2005 they were awarded the Nobel Prize. The story is often referenced as though the scientific community has learned from this experience and is now more open minded. But when I look at how your work has been received thus far I’m not sure I see much progress.
Yes. Also, Barry Marshall was my old microbiology instructor when I studied at the University of Western Australia. He did struggle with recognition for a long time. There are all sorts of urban legends about how he kept the research going in the face of the wall they kept butting up against.
Me: True. And final question: What studies are you planning to do next?
In a next study, we hope to look at the actual microbiome in the brain. Then remove amyloid plaque from this tissue and see if we can identify the exact microbes the plaques have trapped. We’ll do this by seeing if we can recover the genetic material of the seed microbe.
Me: That’s amazing. I can’t wait to see what you find. I’m going to let you go because I want you to get back to work immediately:)