On Friday afternoons, Robert Moir, a neurologist at Massachusetts General Hospital in Boston, indulges in what he calls his ‘play hour’. He doesn’t go to the gym or head to the bar; he plops down in front of his computer – OK, yes, usually with a beer in hand – and directs his browser to PubMed, the US National Library of Medicine’s database of study abstracts from life-science journals. Then he hunts around for hidden gems: studies that might provide a fresh perspective or lead him down a new research path. On 11 May 2007, Moir stumbled across a set of studies while sipping a Sam Adams that would change the course of his career.
Moir studies Alzheimer’s disease; more specifically, he investigates beta-amyloid, the protein that clumps into big, gnarly plaques in the brains of Alzheimer’s patients. According to the prevailing dogma known as the amyloid cascade hypothesis, the build-up of beta-amyloid in the brain directly causes Alzheimer’s – it sparks the accumulation of tau tangles (a primary marker for the disease) inside neurons, leading to cell death and, eventually, dementia. Nothing good comes from amyloid plaques, most scientists contend. They are nothing but devastating biological accidents.
What Moir discovered on that fateful day challenged this idea. He stumbled across a series of findings about LL-37, a peptide that plays a crucial role in innate immunity, the body’s quick-firing defence against invading pathogens. ‘What I was struck by immediately was the fact that this thing – it didn’t have the same genetic sequence as beta-amyloid, but its behaviour was extraordinarily similar,’ Moir recalls. He began compiling a list of the molecular characteristics that the two proteins shared; it grew to be four pages long. His discovery pointed towards a radical idea: that beta-amyloid, like the crucial immune peptide, might ultimately be a good guy, part of the brain’s defence against whatever villain was leading it down the path to Alzheimer’s.
Moir and his colleagues have since published a series of studies – the latest in March 2016 in Science Translational Medicine – showing that beta-amyloid acts as a potent pathogen-fighting molecule in the brain. They’ve reported, among other things, that mouse brains accumulate amyloid plaques within hours of catching infections, and that these plaques help them to live longer. They’ve also watched beta-amyloid directly kill pathogens in the lab.
Their findings lend new support to a controversial idea, known as the ‘pathogen hypothesis’, which has tantalised a small cadre of researchers for decades. It posits that microbes play an important role in Alzheimer’s disease and that beta-amyloid plaques are actually soldiers in the body’s fight against them – a fight that eventually goes haywire, causing mass brain-cell death. In a 2015 meta-analysis of 25 published studies, researchers in Australia reported that people who have been infected by certain types of bacteria are 10 times more likely to develop Alzheimer’s disease; in March 2016, 33 international researchers co-signed an editorial in the Journal of Alzheimer’s Disease imploring others to seriously consider that Alzheimer’s could have an infectious cause.
Yet most scientists continue to dismiss the pathogen hypothesis as far-fetched – it’s an idea, they say, supported mainly by correlative studies involving a small number of brains. That might be true, but it’s worth also contemplating the hurdles that researchers face when investigating unorthodox ideas and the burden of proof that would be needed to convince naysayers. The technological limitations of brain research are immense – scientists can’t just go prodding around inside the brains of living people – and hunting for microbes there is even harder because so many pathogens hide inside cells and sometimes embed themselves in our genomes. Even when scientists identify robust ways to investigate the pathogen hypothesis, they don’t always get funded to perform the experiment; if they do, they then struggle to get their findings published. Moir’s 2016 paper, for instance, was rejected six times without review before finally being accepted. ‘It was brutal,’ he recalls.
Complicating the issue further, if a relationship between infections and Alzheimer’s does exist, it will almost certainly be complex and multifactorial – microbes might just be a first step in a long cascade of events involving genes, inflammation and other environmental factors, which makes the idea still harder to investigate and easier to misunderstand.
Whether the pathogen hypothesis has merit is still very much an open question. But it’s crucial to disentangle the weaknesses of the theory from the limitations of science and the biases of those who are so quick to dismiss it.
The notion that infections can invade the brain and influence cognition is neither new nor unproven. In 1911, Alois Alzheimer himself noted the similarities between the brains of people with syphilis-induced dementia and those with Alzheimer’s. If syphilis festers in the body for more than a decade, the corkscrew-shaped bacterium that causes it invades and damages the brain, sparking the creation of amyloid plaques and inducing dementia. Herpes simplex encephalitis, a rare complication of herpes simplex 1 (HSV-1, the virus that causes cold sores), can cause brain damage; tick-borne Lyme disease can spark memory loss and brain inflammation; HIV over time causes dementia; and then there is Zika, the mosquito-borne virus that can invade developing brains, limiting their size and development.
The more scientists learn, the more they discover that the brain – historically considered a sterile organ protected by the impenetrable blood-brain barrier – often gets infiltrated by microbes. They gain access in a number of ways: HSV-1 infects the cornea of the eye and then invades cells in the trigeminal nerve, which travels to the brain. HIV takes a ‘Trojan horse’ approach – it infects immune cells in the blood that then cross into the brain. The parasite Toxoplasma gondii, which has been shown to infect nearly one-quarter of the US population, sneaks into the brain via endothelial cells. And some oral bacteria – those implicated in cavities and gingivitis – sneak in via the olfactory bulb, a highly innervated area adjacent to the brain.
These studies don’t prove that infections cause Alzheimer’s – they simply suggest that the two are associated with one another
The first researcher to document microbes in the brains of Alzheimer’s patients was the pathologist Alan MacDonald, whose interest was galvanised by the death of his grandfather, who had Alzheimer’s. In a letter to the editor of the Journal of the American Medical Association in 1986, MacDonald, then at Southampton Hospital in Long Island, described discovering the bacteria that cause Lyme disease inside the brains of two deceased dementia patients. He later cultured live Lyme bacteria from several other Alzheimer’s brains. Seven years later, in 1993, the neurologist and neuropathologist Judith Miklossy, then at the University of Lausanne in Switzerland, isolated Lyme bacteria from the brains of 14 patients with Alzheimer’s, yet was unable to find the same bacteria in 13 non-Alzheimer’s brains.
Chlamydia pneumoniae, which causes pneumonia, is another brain pathogen that has been linked to Alzheimer’s. Most people who contract this bacteria develop chronic latent infections in which the microbes hide silently, often for years, inside their cells. In one study, Brian Balin, now the director of the Center for Chronic Disorders of Aging at the Philadelphia College of Osteopathic Medicine, and his colleagues found C. pneumoniae DNA in the brains of 17 of 19 Alzheimer’s brains but in only 1 of 19 non-diseased brains, although some studies have been unable to replicate their findings. Meanwhile, in the UK, the neurobiologist Ruth Itzhaki and her colleagues at the University of Manchester have been studying the relationship between HSV-1 and Alzheimer’s for more than 20 years. They have shown that a person’s risk for Alzheimer’s is 12 times higher if he has HSV-1 in his brain as well as the APOE4 gene variant, which, among other things, increases the transmissibility of the virus into the brain. Fungi, too, have been found to lurk inside neurons in Alzheimer’s brains.
These studies don’t prove that infections cause Alzheimer’s – they simply suggest that the two are associated with one another. It’s possible, for instance, that the brains of people with Alzheimer’s become especially vulnerable to infection, such that the causal arrow points the opposite way. Yet some studies do suggest that infections come first. Moir’s study showed that beta-amyloid plaques accumulated in mouse brains only after salmonella infection; the same sequence of events took place when Balin and his colleagues infected the brains of live mice with C. pneumoniae in 2014. Other studies have reported that people who have antibodies in their blood indicating a recent reactivation of HSV-1 – the virus periodically emerges from its latent state – are more than twice as likely as people without such antibodies to develop Alzheimer’s many years later.
Most proponents of the pathogen hypothesis don’t suggest that infections work alone to cause Alzheimer’s disease, nor do they think that Alzheimer’s can be ‘caught’ like a cold. They instead argue that infections – perhaps a number of different types – can spark a cascade of events that, over time, can culminate in the disease. ‘These microbes might not serve as specific causes, per se, of the disease, but rather as contributors to a degenerative process,’ says Mady Hornig, director of translational research at Columbia University’s Center for Infection and Immunity.
The process could start like this: the brain gets injured or infected and then responds by accumulating beta-amyloid plaques to repair the problem. The idea that beta-amyloid might be protective was first proposed in a 2002 paper by the Australian neurologists Stephen Robinson and Glenda Bishop, who noted, among other things, that amyloid plaques accumulate after serious brain injury. (The beta-amyloid peptide also has not changed much over the past 400 million years and is found in most vertebrates, which further supports the idea that it has an important job.) Moir’s 2016 study in Science Translational Medicine shows that, indeed, beta-amyloid peptides protect the brain from microbes by first attaching themselves to pathogens, then creating fibrils that wrap around and entrap the microbes. Within hours of injecting salmonella into mouse brains, Moir and his colleagues saw widespread beta-amyloid plaques appear.
The blood-brain barrier becomes more permeable with age; maybe microbes have an easier time crossing it later in life
When enough beta-amyloid plaques accumulate after an infection, though, they start creating their own problems. ‘Long after the pathogen is dead, the amyloid plaque remains and causes this cascade of inflammation,’ Moir explains. As some of his colleagues at Massachusetts General Hospital showed in a 2014 paper in Nature, amyloid plaques lead to the production of tau tangles, which directly kill neurons. Plaques also alert microglia, the brain’s immune watchdogs, to the fact that something has gone awry; the microglia then spring into attack mode, releasing immune chemicals and reactive oxygen species that are designed to kill microbes but also harm neurons as collateral damage. ‘It’s this friendly fire that’s probably killing many more neurons than the original pathology that triggered the response,’ explains Rudolph Tanzi, director of the Genetics and Aging Research Unit at Massachusetts General Hospital and one of Moir’s close colleagues.
It’s also possible, some researchers posit, that infections outside the brain could influence Alzheimer’s risk. ‘There are circuits in the brain that are activated by the presence of a systemic infection, and systemic infections “talk” to the brain not by the infection getting into the brain but by communication along a number of different routes,’ says V Hugh Perry, a neuropathologist at the University of Southampton in the UK. Some microbes might trigger the activation of immune molecules outside the brain that then travel into the brain, influencing it; others might cause the body to create self-attacking antibodies that damage brain cells. A 2015 study published in the European Journal of Neurology found that people with antibodies to common systemic infections are indeed at a heightened risk for Alzheimer’s disease.
But even if Alzheimer’s is initially sparked by some kind of infection, disease risk is undoubtedly shaped by other factors, too. Age is an important one: most people don’t develop the disease until after 65. Research suggests that the blood-brain barrier becomes more permeable with age, so it’s possible that microbes have an easier time crossing it later in life. Microglia also become more overzealous with age, making inflammatory responses more potent and dangerous. Genes play a crucial role too; some early onset forms of Alzheimer’s are caused solely by rare genetic mutations, while other gene variants can make the brain slightly more vulnerable to the effects of infection, make inflammatory responses to infection more damaging, or, as with APOE4, affect the ability of pathogens to cross into the brain. The bottom line is that even if infections do play a causal role in Alzheimer’s, a person’s risk will be shaped by other factors, which makes the whole process much harder to piece together.
Despite more than two decades of research on the pathogen hypothesis, it has rarely been discussed in Alzheimer’s circles. In a 2016 paper published in the 60th-anniversary issue of the Journal of Neurochemistry, researchers discussed six alternative hypotheses on the causes of Alzheimer’s disease – ones that compete with the reigning amyloid cascade model – but the pathogen hypothesis was not included among them. Certainly, the pathogen hypothesis has only limited, correlative evidence in its support. But the amyloid cascade hypothesis has some compelling direct evidence against it.
The amyloid cascade hypothesis was born in 1992 with the discovery that mutations in the protein precursor for beta-amyloid, known as APP, cause early onset forms of Alzheimer’s. People with Down’s syndrome, who have three copies of the gene that makes APP, are also at especially high-risk for early onset disease. These findings suggest that when the body’s genetic instructions tell it to make extra beta-amyloid, Alzheimer’s ensues. But most Alzheimer’s sufferers do not have this genetic, early onset disease. Most have what is known as ‘sporadic’ Alzheimer’s, which typically arises later in life. It tells a different story.
According to the amyloid cascade hypothesis, the brain’s accumulation of beta-amyloid plaques is the primary cause of sporadic Alzheimer’s disease. By this definition, then, brains overflowing with plaques should have Alzheimer’s. But often they don’t: postmortem studies show that one-quarter to one-third of non-Alzheimer’s-diseased elder brains are filled with amyloid plaques, too.
Interestingly, the behaviour of microglia, the brain’s resident immune cells, can in part distinguish these so-called ‘resilient’ brains from brains that have Alzheimer’s. In a 2013 study, researchers at Harvard University analysed 50 brains, some of which came from people who did not have Alzheimer’s yet had beta-amyloid plaques. They found that one key difference between the resilient brains and the diseased brains was that the former had much fewer activated microglia.
conceding that the pathogen hypothesis has merit means admitting that the field has been wrong about Alzheimer’s for a very long time
If the amyloid cascade hypothesis is correct, then it would also follow that removing amyloid plaques from the brains of Alzheimer’s patients should ameliorate symptoms – or at least not make them worse. But this, too, isn’t the case. Out of 145 beta-amyloid-reducing drugs tested in Alzheimer’s clinical trials between 2002 and 2012, none slowed progression of the disease, even though some did successfully reduce beta-amyloid levels. ‘It would make sense to propose that it is time to reject the amyloid cascade hypothesis and search for alternative explanations for the cause(s) of human Alzheimer’s disease,’ wrote Karl Herrup, a molecular neuroscientist at the Hong Kong University of Science and Technology, in a 2015 commentary published in Nature Neuroscience. ‘Instead of rejecting the hypothesis, however, the field has essentially redefined the disease. The result is a dangerous circular logic that is holding back the field.’
Why do researchers reject the pathogen hypothesis for a lack of evidence yet support a theory that lacks supporting data too? In a reply to one of Miklossy’s early papers, a group of Alzheimer’s researchers explained that they ‘remain skeptical and even incredulous at the thought that such an etiology for such an important and exhaustively studied disease could have been overlooked by so many including ourselves’. Put another way: conceding that the pathogen hypothesis has merit means admitting that the field has been wrong about Alzheimer’s for a very long time.
There might be deeper psychological barriers to accepting the pathogen hypothesis as well. Time and time again, scientists have had trouble accepting that microbes can cause common diseases. One particularly famous example involves gastric ulcers: although the internist Barry Marshall and the pathologist Robin Warren had isolated H. pylori bacteria from the guts of people with ulcers in the 1980s in Australia, and antibiotics had been successfully used to cure ulcers in some hospitals in the late 1940s, mainstream scientists refused to believe that they could have an infectious cause. Marshall eventually convinced the world by drinking a concoction of the bacteria himself, which gave him gastritis, the precursor to an ulcer, which he then cured with antibiotics. It also took half a century’s worth of studies to convince the scientific establishment that rheumatic fever was caused by streptococcal bacteria, in part because there is a confusing months-long lag between bacterial infection and disease onset.
Considering that the neurodegenerative process that leads to Alzheimer’s likely takes years, if not decades, and that it also involves other factors, it is not terribly surprising that the idea has been so difficult to accept. ‘We tend to gravitate toward parsimonious answers to complex questions in all walks of life, and our quest to solve the perplexing mysteries of neurodegenerative diseases is perhaps no different,’ Hornig explains. ‘Tracing the path from an infection to a disease as complex as Alzheimer’s requires substantial tolerance for caveats and an appreciation of how factors from the environment, including infections, might lead to different brain outcomes.’
It’s ironic, because some scientists reject the pathogen hypothesis precisely because they oversimplify it – they say that infection can’t be a cause if age or genes influence disease risk too, or that HSV-1 can’t be involved since it’s far too common. What these straw-man arguments really reveal is that many scientists haven’t taken the time to appreciate the nuances and complexity of the idea.
History and bias aside, though, the pathogen hypothesis certainly prompts more questions than it provides answers. The field needs a lot more research – rigorous research – before anyone can know for sure whether infections are causally related to Alzheimer’s.
The science won’t be easy to do either, for a number of reasons. Researchers can study the brains of people with sporadic Alzheimer’s only once the disease process is established – no one knows who is going to get the disease in advance – which means that they can’t watch the first events unfold. ‘You know what’s there at the end,’ says Heather Snyder, senior director of scientific and medical operations at the non-profit Alzheimer’s Association, ‘but because of the complexity of the brain, being able to move back in time has been a real challenge.’
It’s also immensely difficult to find microbes lurking in the central nervous system. One oft-cited limitation of the research linking HSV-1 to Alzheimer’s is that no one has yet been able to isolate whole HSV-1 viruses inside the brains of deceased Alzheimer’s patients; scientists have been able to detect only the DNA of HSV-l, which could, in theory, come from dead viruses or their pieces. But HSV-1 re-activates only sporadically, so the viruses are circulating in the brain only during short windows. Even when viruses are definitely there, they are still tough to isolate. When researchers in Finland attempted to isolate HSV-1 from the spinal fluid of patients with active herpes simplex encephalitis, they were able to do so less than 2 per cent of the time. ‘Even a brain biopsy may fail to uncover an agent if it is present at very low levels – under the radar for detection, perhaps because samples are acquired in the wrong brain region or after the acute phase of the infection has resolved – or if it is a novel, unexpected agent,’ Hornig explains.
‘Big Pharma continue to pour research dollars into anti-beta-amyloid therapies, yet not a single clinical trial in the past 15 years has significantly improved patient outcomes’
One way to prove the pathogen hypothesis correct would be to show that preventing or curing infections with antimicrobial drugs reduces the risk of Alzheimer’s disease. Of course, the right drug would have to be given at the right time, which complicates the picture. Still, Itzhaki, for one, has been advocating for a clinical trial involving antiviral drugs for years. It has yet to happen. ‘There really haven’t been any scientific arguments against doing a trial, so it is just funding that’s a problem,’ she says.
Indeed, funding has been a major issue among researchers trying to investigate the pathogen hypothesis. The vast majority of funds, Robinson says, continue to be spent on research and trials related to the amyloid cascade hypothesis. ‘This dominance has meant that few resources are available to support research into alternative hypotheses,’ he says. ‘Frankly, I am amazed that Big Pharma continue to pour hundreds of millions of research dollars into anti-beta-amyloid therapies, despite the fact that not a single clinical trial in the past 15 years has significantly improved patient outcomes.’ While the Alzheimer’s Association has recently been broadening its focus – it has, for instance, funded 53 projects since 2011 to investigate inflammation in Alzheimer’s disease – it has funded only one project in this timeframe specifically related to infection.
It will take decades of careful work for researchers to untangle the complex web connecting infections, genes, and immunity to Alzheimer’s disease, assuming that they secure enough funding to do so. Still, Moir says he senses that the research community is slowly becoming more receptive to the idea. He and his colleagues recently secured funding from the Cure Alzheimer’s Fund in Massachusetts for a handful of related projects, including a Brain Microbiome Project in which they will sequence DNA from postmortem brains to characterise the types of microbes that hide there. At a neurodegeneration conference held in Korea in 2016, Moir says attendees were asked to raise their hands if they thought infections might be involved in Alzheimer’s, and a majority of hands went up. ‘Ten years ago, it would have been four guys in a corner, all huddled together, not talking to anyone else,’ Moir says. ‘My impression is that this is an idea whose time has come.’
