Jeremy Taylor

[See all essays on this Hot Topic]
A few weeks ago, pharmaceutical giant Eli Lilly announced that their latest Phase II trial for a drug named solanezumab, which is designed to flush beta-amyloid protein out of the brain and thereby ameliorate the symptoms of Alzheimer’s disease, had been a failure. This is a huge blow to conventional Alzheimer’s disease research because it comes on the back of more than a decade’s-worth of failed trials for drugs that either interfere with the chain of enzymatic reactions that make beta-amyloid, or remove it from the brain. Most of these trials were done with patients who were already in the advanced stage of the disease, and it may be too much to expect any treatment to be successful in the face of such widespread neuronal degeneration. This has led to research initiatives to identify cohorts of patients either in much earlier stages of Alzheimer’s disease, or of young, symptomless age but related to individuals who had succumbed to the familial, early-onset form of the disease. This solanezumab trial used patients with only mild cognitive impairment and yet it showed no statistical improvement in cognition in treated patients over placebo controls.

What does this failure mean for Alzheimer’s disease research? In what direction should it now go? Should the “amyloid hypothesis” now be abandoned? The overwhelming consensus in Alzheimer’s disease research has held it to be self-evidential that the tell-tale plaques of beta-amyloid protein between neurons and hyper-phosphorylated tau protein within neurons are not only pathological proof of Alzheimer’s disease but are the toxins that cause the disease in the first place. But it is becoming clear to anyone willing to acknowledge the evidence that this may be very far from the full story. Very recently, the 90+ study, run by the University of California at Irvine, produced evidence that many of our oldest-old die with substantial Alzheimer’s pathology in their brains but with fully-functioning preserved cognitive powers. It is the latest in a string of similar observations that dates back to Alzheimer research pioneers Bob Katzman and Bob Terry, at UCSD, in the 1980s and 90s. They similarly identified a cohort of the elderly who had died at a mean age of 85 years with intact cognition in brains riddled with amyloid, and, conversely, individuals who were diagnosed with Alzheimer’s disease who were found, on autopsy, to lack significant amyloid and tau pathology. Why were some brains more resilient to amyloid pathology than others? Katzmann and Terry thought it might have something to do with cognitive reserve. Resilient brains tended to be bigger brains and belonged to individuals in the upper centiles of intellectual performance. Maybe they lost synapses to amyloid but just had better brains, with more synapses, and, consequently, had more to lose and took longer to lose it? Their decline into dementia was just slower. This theme was returned to in 2004 by Nikolaos Scarmeas and Yaakov Stern in a paper titled “Cognitive Reserve: Implications for Diagnosis and Prevention of Alzheimer’s Disease.”

However, importantly, Bob Terry won the Potemkin neuroscience prize back in 1988 for counting cortical synapses from normally aged and Alzheimer’s diseased brains and showing that there were only weak correlations between density of plaques and tangles and psychometric tests of intelligence but much stronger correlations between those tests and synapse density. He further showed that loss of synapses was independent of the presence of amyloid in diffuse plaques and concluded that amyloid deposition was the result of synapse pathology, not the cause. Alzheimer’s disease was a disease of synapses, not of amyloid.

Several respected science communicators, including George Perry of the University of Texas, have accused the so-called amyloid lobby of persevering with amyloid and tau for no better reason than the fact that, ever since the days of the eponymous pioneer of neurodegenerative brain research Alois Alzheimer, amyloid plaques and tau tangles have been visible in brains via microscopy or medical imaging. Nevertheless, the amyloid lobby continues to test their hypothesis to destruction in trials still underway which administer anti-amyloid drugs to individuals thought to be at risk of Alzheimer’s disease well before any cognitive decline registers itself. They are trying to intervene at ground zero. But, in the search for the initial pathology of Alzheimer’s disease, will we find that beta-amyloid and tau are even relevant? What other processes and agents deserve much greater attention? What really causes these brains to start dying in a more profound and accelerated way than can be laid at the door of the normal ageing process?

The more I read about research into Alzheimer’s disease the more I am reminded of the old Indian parable of the blind men and the elephant. Twelve blind men are stood around the beast at intervals and are requested to describe and identify it based only on what they can make of the small portion of its anatomy that lies within their hands’ grasp. Of course, none of them can take on board what the others are feeling, there is no overall picture, and so they end up in total disagreement, and in no little ignorance as to what the beast actually looks like.

Let us think about amyloid a little further. We know it is present in neurons because it has a clearly defined evolved function in regulating transmission of impulses along neuronal networks – supporting long term potentiation, which is involved in the storage of memory at synapses, and regulating over-excitation within neural networks. And it is commonly asserted that the slightly longer-chain Aß-42 molecules are more toxic than Aß-40 and only when they form into certain types of oligomers. Rebecca Rosen et al, in a paper titled “Comparative Pathobiology of Aβ and the Unique Susceptibility of Humans to Alzheimer’s Disease”, question why it is that humans appear uniquely susceptible to the neurodegeneration and dementia of Alzheimer’s disease despite the fact that all primates deposit copious Aß in senile plaques and accumulate cerebral amyloid-β angiopathy as they grow old. And despite the fact that the amino-acid sequence of beta-amyloid is identical in all primates – including humans – so far studied. Also, transgenic rodent models engineered to overproduce human-sequence beta-amyloid develop profuse senile plaques and cerebral amyloid-β angiopathy, but they do not have substantial AD-like neuronal cell loss, neurofibrillary tangles, and profound memory impairment. They conclude with the possibility that the only between-species differences they could find – subtle differences in the tertiary structure of beta-amyloid – the three-dimensional geometry of protein chain folding – might explain why beta-amyloid is toxic to humans but not to any other species.

But dissecting beta-amyloid in ever increasing detail like this still leaves us with the fundamental question: Is it the accumulation of beta-amyloid into plaques, and the subsequent formation of hyper-phosphorylated tau protein tangles inside neurons, that are the initiating events for Alzheimer’s disease, or not? Here it is important to distinguish between familial Alzheimer’s disease and sporadic Alzheimer’s disease. The former represents less than 5% of all AD cases and is caused by well-known mutations to genes like APP or the presenilins, which are involved in the chain of enzymatic reactions that form beta-amyloid. If you bear any of these mutations you will succumb to the disease – it is deterministic. The vast majority of cases of Alzheimer’s are the sporadic form which tells us that, whatever genes might be involved, the environment is a major factor. Central to this observation is the establishment, over the last 20 or 30 years, that no satisfactory answer to the riddle of Alzheimer’s disease will ever be found unless we take the role of the immune system, and the inflammation caused by innate immunity, into account.

A number of researchers hold that, while inflammation is an important factor in Alzheimer’s disease, it is secondary to the production of amyloid and tau in the brain. That it is the production of amyloid and tau that elicits inflammation. But it looks increasingly likely that the opposite is true.

Back in the 1980s, Sue Griffin used Down syndrome to investigate Alzheimer’s disease. Down syndrome brains accumulate amyloid at premature age because of the extra copy of chromosome 21 on which the APP gene sits. Nevertheless, Griffin showed that Down syndrome brains produce large amounts of the inflammatory cytokine interleukin-1 (IL-1) many years before plaque formation, suggesting that stressed neurons lead to inflammation and innate immune activity in the brain, production of inflammatory markers, and eventually excess amyloid and tau.

The importance of innate immune system activity in the brain was heavily underscored a few years ago by 3 genome-wide association studies which found no effect for the main genes involved in the pathways that form beta-amyloid and tau, but were dominated by genes involved in the immune system.

Research by Clive Holmes and Hugh Perry in Southampton has established that peripheral infection can send signals to the brain which accelerate immune activity there, heighten the symptoms of Alzheimer’s disease, lead to cognitive deficits, and prime microglia – the brain’s immune cells and the equivalent to macrophages – so that they are capable of attacking and damaging neurons. Their research has been borne out in a mouse model by Irene Knuesel and Dimitrij Krstic which mimicked peripheral viral infections and showed increases in inflammatory mechanisms in the brain, priming of microglia, degenerating neurons and only then production of amyloid and tau.

What key events occur at the synapse years before any signs of cognitive impairment begin to emerge? Some researchers believe that beta-amyloid is the prime culprit in these early days but their work is countered by fascinating evolution-minded research by Beth Stevens at Harvard and her former colleague Ben Barres. Picking up from those earlier conclusions that synapse loss correlates better than amyloid with AD cognitive symptomatology, Hong et al (co-authors include Stevens, Barres and Dennis Selkoe) show in a series of mouse models that another major part of the innate immune system – complement – together with immune cells called microglia – are heavily involved in initiating events at the synapse that precede amyloid deposition. The initiating protein of the complement cascade – C1q – is first associated with synapses and experiments that inhibit it show that C1q is necessary for any toxic effect of beta-amyloid on synapse function and long-term potentiation in the hippocampus. Similarly, when the complement receptor CR3 is silenced on microglia they stop phagocytically engulfing synaptic material.

Stevens, Barres, and their associates remind us that evolution has co-opted the complement cascade as the mechanism by which neural networks are adaptively pruned during adolescence and brain development and as a response to later learning. C1q paints synapses that are scheduled for demolition and reacts with proteins on these cell surfaces to form the complement protein C3 which is recognised by CR3 on microglia which then steam in for the kill. Most of the body’s cells are protected against this unwanted intrusion by complement because they are bristling with complement inhibitors. Neurons are the exception. They lack these inhibitors for the very reason that they have to be open to complement attack otherwise selective pruning could never occur. It is an Achilles heel which shows up in late-onset Alzheimer’s disease because Stevens, Barres et al believe the roots of Alzheimer’s disease are laid when this evolved method for synaptic pruning is re-awakened maladaptively in later life. It may be, they say, that soluble beta-amyloid has a role here in that it could bind to synapses and weaken them, providing the complement cascade with a signal for elimination.

Not surprisingly, in the light of all this, a group of scientists in the UK, led by Prof. Paul Morgan of Cardiff University, have published research which suggests that complement proteins can provide reliable early markers for onset of Alzheimer’s disease, specifically to allow physicians to distinguish between individuals with mild cognitive impairment who will convert to Alzheimer’s from those who will not.

The gene that we know for sure increases your chance of contracting Alzheimer’s by up to ten times is a variant of APOE – epsilon 4. And while APOE is involved with a number of processes in the brain that also involve beta-amyloid, there are a number of other Alzheimer’s producing processes in which APOE4 acts independently. Because it is involved in cholesterol transport, APOE is vital for maintaining neurons and their synapses, and the epsilon 4 variant impairs this. Carriers of APOE4 have thinner entorhinal cortices and hippocampi, and APOE4 frequently increases inflammation in the brain and primes toxic microglia. It is known that another variant of APOE – epsilon 2, is protective of Alzheimer’s disease and it was assumed that APOE2 carriers with somewhat preserved cognition would consequently be found to have been relatively free of amyloid pathology. But the group who run the 90+ Study at UC Irvine have discovered that while, in the oldest-old, the presence of APOE2 was associated with a somewhat reduced risk of dementia, it was also, paradoxically, associated with increased AD neuropathology. Therefore, they conclude, oldest-old APOE2 carriers may have some mechanism that contributes to the maintenance of cognition independently of the formation of AD pathology and specifically note that APOE2 carriers have preserved synaptic function.

It is too early to abandon the so-called amyloid hypothesis. Soluble beta-amyloid or oligomers of beta-amyloid 42, and aberrant tau protein, are clearly neurotoxic and important. But they may not be instigatory. The amyloid hypothesis, at the very least, is undergoing substantial revision as the long history of blinkered over-attention to tell-tale plaques and tangles gives way to the nuances of environmental factors and innate immune responses in brain and body and the important distinctions between early-onset familial AD and the majority late-onset sporadic AD come home to roost. In the following two commentaries, Caleb Finch draws attention to the possible role of smoking and atmospheric pollution, while Robert Moir rehabilitates the amyloid “bad boy” by showing its evolved importance as a potent antimicrobial – thereby opening the door to a possible infection etiology for Alzheimer’s disease. 35 million people world-wide are living in the twilight world of Alzheimer’s disease without the ghost of a cure in sight, despite the investment of many billions of dollars. We owe it to these Alzheimer’s sufferers – in the US alone a new case gets diagnosed every 68 seconds – to broaden the church of AD research in such ways – and allow this new research to present effective targets for treatment. It is long overdue.