Nature Biomedical Engineering: Targeting mitophagy to clear brain "garbage" to treat Alzheimer's disease

As the world ages, the economic burden of Alzheimer's disease continues to increase, reaching $305 billion in 2020
.
At present, about 11% of people over the age of 65 have Alzheimer's disease, and the proportion of people over the age of 85 is as high as 42%

On January 6, 2022, the research group of Evandro Fei Fang from the University of Oslo in Norway and the research group of Jiahong Lu from the University of Macau published a paper entitled: Amelioration of Alzheimer's disease pathology by mitophagy inducers identified in Nature Biomedical Engineering.
Research paper on via machine learning and a cross-species workflow
.

This study screened a large number of small molecules based on machine learning and found two lead compounds with the potential to treat Alzheimer's disease
.

Xie Chenglong of the First Affiliated Hospital of Wenzhou Medical University, Zhuang Xuxu of the University of Macau, Deruizhi Yao Niu Zhangming, and Ai Ruixue of the University of Oslo, Norway are the co-first authors of this article, and Fang Fei of the University of Oslo, Norway and Lu Jiahong of the University of Macau are the co-corresponding authors
.

Fang Fei's group published a paper in Nature Neuroscience in 2019 demonstrating that impaired mitochondrial clearance mechanisms play a key role in the pathogenesis of Alzheimer's disease
.
Mitophagy is a mechanism that selectively removes damaged mitochondria through autophagy during cell evolution.

Modern drug development is expensive, including the screening of compounds for biological activity
.
After discovering lead compounds, scientists conduct drug design and compound synthesis based on experience, molecular biologists test compound activity, repeatedly optimize and screen, and finally discover clinical candidate compounds.

Fang Fei's team and its collaborators are committed to applying AI for high-efficiency and low-cost screening in the early stage, and performing traditional wet laboratory validation (cells, nematodes, and mice) in the later stage to improve the accuracy of screening drugs.
The combination of the two parts can accelerate the development of drugs.
The design ideas can be applied not only in Alzheimer's disease, but also in various medical fields, providing more inspiration for other research groups
.

In this study, a new AI technique was established and applied to drug screening : representation learning based on machine learning of 1D, 2D and 3D information for each molecule
.
The pre-trained data comes from a total of 19 million small molecule data from ChEMBL and ZINC

The research team clustered and filtered a total of 3724 natural small molecules in the natural small molecule database of the University of Macau
.
A total of 18 natural small molecules entered the wet laboratory validation stage

The study received positive reviews from international peers
.
Li-Huei Tsai (MIT, MIT), an international authority on Alzheimer's disease research and a professor in the Department of Brain and Cognitive Sciences at the Massachusetts Institute of Technology, pointed out: "This study used cutting-edge machine learning algorithms to successfully discover new neurons that induce mitochondrial autophagy.

Professor Menno Witter, an international authority on brain science research at the Norwegian University of Science and Technology (NTNU), said, "This is an impressive multidisciplinary paper, and Fang Fei's team has very convincingly demonstrated a new method for screening drugs for neurodegenerative diseases.

In conclusion, this study established a virtual screening algorithm based on AI fusion of multi-dimensional molecular information, and successfully screened out multiple autophagy inducers combined with biological experimental verification to improve the accuracy of AI screening for Alzheimer's disease

Original source:

Xie, C.
, Zhuang, XX.
, Niu, Z.
et al.
Amelioration of Alzheimer's disease pathology by mitophagy inducers identified via machine learning and a cross-species workflow .
Nat Biomed Eng (2022).
https://doi.
org /10.
1038/s41551-021-00819-5.