Neurodegeneration — 11

Aprovocative study has raised the potential of treating “jumping genes,” known as retrotransposons. By transforming skin cells to pluripotent stem cells and onto neurons, from patients with Alzheimer’s disease and controls, many cardinal features of the disease (such as Aβ deposition and Tau accumulation) were reproduced in organoid 3D cell culture and reversed by lamivudine, a drug that stabilizes transposons.

To summarize what we’ve learned in recent years, the evidence of net clinical benefit for two monoclonal antibodies is modest at best. Even getting rid of all the amyloid deposits in the brain may prove to be too late to meaningfully interrupt the inexorable progression of the condition. This pertains to treatment for people who are already affected with mild cognitive impairment of early Alzheimer’s disease. That’s very different from achieving prevention or marked delay of the process in people at high risk.

PREVENTION

The breakthrough in this field has been a simple blood biomarker that indicates high risk and lays the foundation for the exciting possibility of primary prevention. With all the related knowledge we’ve accrued, I consider this the exemplar for changing the natural history of a devastating age-related disease. Remember the propitious advantage on our side: we know it takes more than two decades to develop this condition. A simple blood test of inflammation, C-reactive protein, in early adulthood, assessed serially on more than two thousand participants over eighteen years, partitioned the risk of cognitive decline. Let me outline the approach that relies on identifying people at high risk, close surveillance, multimodal AI analytics of the data, and bespoke interventions.

I’ve written elsewhere about the potential for medical forecasting to simulate the remarkable success of AI-powered GraphCast weather forecasting (with 99.7 percent accuracy!), which is truly amazing. In figure 7.5, we see ways to partition high-risk individuals integrating multiple layers of data. That includes not only known clinical factors such as hearing loss, osteoporosis in women, erectile dysfunction or prostatic hypertrophy in men, and paternal history, but also lifestyle+ factors that we’ve just reviewed. Complementing this are the APOE genotype, a polygenic risk score, a blood p-tau-217 test, inflammation blood markers such as hs-CRP, a brain organ clock cluster of plasma proteins, a gut microbiome for its constituents and presence of certain pro-inflammatory microbes, and overall body-wide epigenetic age by DNA methylation. A retina image is independently helpful for prediction of high risk. Speech, analyzed with AI methods, can predict Alzheimer’s disease within six years. Multimodal AI is used to quantitatively determine the risk level, factoring in all these data inputs, and has even displayed accuracy with more limited data. A threshold can be selected for categorization of high risk, such as top 5 or 10 percent of individuals assessed.