Volume 56, Number 3 - Summer 2010

 

By Dale Bredesen, MD

Alzheimer’s disease (AD) is an increasing health problem, with more than 5 million Americans currently suffering from the disease and an estimated 13 million by 2050. The financial burden of $150 billion per year does not take into account the more important emotional and social effects of this devastating disease. Currently available therapies, such as cholinesterase inhibitors and the glutamate antagonist memantine, have only minimal effects. New therapeutic prospects are undergoing active evaluation, but the list of failed clinical trials is growing rapidly, with recent disappointments including such previously promising candidates as dimebolin (Dimebon), tarfenflurbil (Flurizan), and tramiprosate (Alzhemed).

Why have we, as a biomedical community, been so unsuccessful at treating this disease despite more than a century of research and development since its original description? There may be several reasons, one of which is common to virtually all chronic illnesses: by the time symptoms appear, the pathogenetic process is far advanced. Therefore, preventing AD may prove to be more feasible than treating it post-symptomatically.

Another reason that AD has proven refractory to therapy may be that our fundamental understanding of neurodegenerative disease lags far behind that of other chronic illnesses. Theories of AD etiology abound, but most share the notion that the build-up of Abeta peptides causes neurotoxicity by physical and chemical means. (Abeta is a 40-42 amino acid peptide that forms the predominant proteinaceous component of the senile plaques that characterize AD.) Abeta has been shown to cause toxicity by detergent-like effects on membranes, channel formation, metal binding, and reactive oxygen species production, among other effects. In other words, Abeta has been seen as a toxic agent, something like acid. 

These theories, however, do not explain why most cells produce Abeta peptides constitutively (which suggests that Abeta may have a physiological function); nor do they explain recent lab results in which transgenic mice modeling AD produce large amounts of Abeta peptides yet show few if any signs of AD. These findings support the alternative notion that Abeta is not simply like acid in the brain. Rather, just like many other peptides, Abeta is a physiological, neuromodulatory peptide, and its pathological effects in AD may be interrupted downstream of its biological signaling. If that is indeed the case, what biological signaling does Abeta mediate, and how can we exploit this mediation to prevent, and ultimately cure, AD?

What if AD is something different than previously envisioned? Our research at the Buck Institute over the past several years suggests that the very molecular and cellular pathways that mediate neuronal development are employed in the neurodegenerative process. While this finding may seem unsurprising, it paints a different picture of AD than the prevailing one, and has some surprising implications. 

To understand these implications, let’s go back to basic high school biology. Suppose you want to make an organism, whether it’s a nematode worm, a human being, or something else. After the egg has been fertilized by the sperm, you basically want the resulting zygotic cell to do four things to create the organism: proliferate, differentiate, migrate (both cell bodies and cell processes), and integrate. If you are a nematode, when you have completed these developmental processes, you are essentially done—you don’t make more cells, you live 2-3 weeks, and that’s it. 

If you are a higher organism, however, you co-opt many of the same genes and proteins that you just finished using for development and employ them for repair and regeneration. That strategy allows you to live far longer than the nematode’s 2-3 weeks, but it carries some liabilities as well. For example, nematodes don’t have to worry about cancer, but in humans, the ability to continue to proliferate cells for turnover and replacement requires a fine balance between, on the one hand, proliferation and survival, and, on the other hand, turnover (much of which is accomplished by programmed cell death). At the genetic level, this scenario translates into a balance between oncogene and tumor suppressor gene activity. 

If you damage a cell’s DNA—whether from smoking or ultraviolet light or carcinogenic foods—there is a chance that you will alter this balance. If the balance is altered in favor of the tumor suppressor genes (decreasing oncogene activity or increasing tumor suppressor gene activity), then you may suffer an early cellular demise; but losing a cell or a few cells is usually not a major problem. On the other hand, if the alteration is in the other direction, such that the cell increases proliferation and/or survival, then there is an increased likelihood of developing a neoplasm. This process is self-selective at the cellular level, since the abnormality begets cells with a similar abnormality. In other words, humans have a non-homeostatic system, in which a perturbation leads not to correction but to further perturbation, producing disease.

 

Now let’s look at what might happen if the analogous problem occurred with the genes and proteins involved in connectivity, i.e., in the processes of migration and integration. For the sake of argument, let’s call these “cerogenes,” from the Latin concero, to join or connect. Once again, there is a balance between yin and yang, in this case between process extension and retraction, as well as the formation and reorganization of connections. At the synaptic level, this balance translates to the formation vs. reorganization of connections, and is what underlies neuronal plasticity. 

Such plasticity is critical for your neuronal “desktop,” among other things. Just as you are constantly creating documents and shuffling them onto and off the desktop of your computer, bound for the trash or for long-term storage, so you are constantly forming and destroying memories, from your memory of the seventh song that played on the radio on your way to work (if you remember that one, you may be wasting storage) to your memory of where you parked your car (you hope that one is still with you, but you will likely move it off your hippocampal desktop sometime soon). 

At first glance, an abnormality in cerogenes would not seem to have the same sort of repercussions as an abnormality in oncogenes, since there is no selection for cellular proliferation and survival. In our research, however, we discovered something surprising: the molecular control of the proteins that mediate migration and integration features positive feedback. For example, the Abeta peptides that collect in the brains of patients with AD are derived from a larger precursor called, somewhat unimaginatively, APP (amyloid precursor protein). 

APP is turning out to be a remarkable molecule. It can be cleaved at three sites to generate four peptides—sAPP beta, Abeta, Jcasp, and C31—that have all turned out to be peptides that mediate neurite retraction, synaptic reorganization, and ultimately, programmed cell death. We call these the Four Horsemen (in this case, of the mnestic apocalypse). Alternatively, APP can be cleaved at two sites to produce three peptides—sAPP alpha, p3, and AICD—that mediate neurite extension, synaptic maintenance, and cell survival. We call these the Wholly Trinity, since they keep cells whole (see figure).

 Alternative cleavage of APP to produce four peptides that mediate synaptic loss, neurite retraction, and ultimately programmed cell death ("the four horsemen"); or three peptides that mediate synaptic maintenance and inhibit programmed cell death ("the wholly trinity"). 

 

APP functions as a molecular switch, and its switching appears to be governed by its interaction with ligands. When APP interacts with netrin-1, an axonal guidance ligand, it mediates process extension. When APP interacts with Abeta, however, it mediates process retraction, synaptic loss, and programmed cell death. During this interaction, Abeta begets more Abeta (one of the Four Horsemen) by favoring the processing of APP to the Four Horsemen. In other words, Alzheimer’s disease is a molecular cancer. Positive selection occurs not at the cellular level but at the molecular level. Furthermore, Abeta itself is a new kind of prion, since it is a peptide that begets more of itself. We believe that all of the major neurodegenerative diseases may operate in an analogous fashion.

One of the interesting ramifications of our new model of AD is that we should be able to screen for a new kind of drug: “switching drugs” that switch the APP processing from the Four Horsemen to the Wholly Trinity, thus preventing the synaptic loss, neurite retraction, and neuronal cell death that characterize AD. Indeed, we have identified candidate switching drugs and are now testing these in transgenic mouse models of AD. We are also testing the effects of netrin-1 on this system, and finding similar effects. 

A corollary of the switching principle is that we should now be able to screen existing drugs, nutrients, and other compounds not just for their carcinogenicity (as is done using the Ames test) but also for their Alzheimerogenicity. We rarely stop to think that we are likely exposed to many compounds that have positive or negative effects on the likelihood that we will develop AD, and it would be helpful to have such information. We hope that our new model of AD may provide new insight into the pathogenesis of this common disease and offer new approaches to therapy.

---