HONG KONG – MBD Co. Ltd. and Azothbio Inc., both based in South Korea, have entered an agreement for rare cancer drug development using an AI-powered drug discovery platform. The two biotechs plan to combine their technology and platforms to accelerate new drug candidate discovery, with a focus on first-in-class cancer drugs.
MBD has improved a 3D cell culture and cell spotter platform, whereas Azothbio has developed its de novo drug design model, called AiMol. AiMol is an automatic molecule generator based on generative-adversarial network (GAN) developments. Aiming to speed up the drug discovery process, they will begin the project in middle of March, and it will run for a year.
“AI-based drug screening can shorten the time from hit identification to preclinical candidate selection,” Ho-sang Moon, chief technology officer at MBD, told BioWorld.
“The conventional method of drug development takes two to three years from hit to preclinical candidate; however, our screening platform with Azoth’s AI technology will shorten the period to around one year,” Moon added.
AiMol uses new active molecules to create drug candidates. It is able to extract specific features from the existing lead molecules and combine them together.
AiMol can also combine the structural characteristics of pharmacological actives in a virtually compressed space, which is an algorithm that enables it to control specific effects during the combination process. The virtually compressed space can be from 16 to 1,024-dimensional models, according to Azothbio.
“AiMol can virtually create a new molecular structure by increasing the biological activity of specific lead substances,” Jerry Maeng, business development director at Azothbio, told BioWorld. “Also, it can enhance the selectivity of the drug candidate while suggesting new formulae for multiple targets.”
The two companies will expand training sets and conduct research into synthesis possibilities. They also plan to study how to apply AI technology to new targets that do not have enough training sets, said Moon.
The two biotechs are fast-growing young ventures in Korea. Founded in 2015, MBD has developed a 3D cell culture platform and new drugs.
Using drug sensitivity, MBD tests cancer drugs and radiation doses used in two therapies: neoadjuvant chemotherapy, which is the use of therapeutic agents before a main treatment, and chemoradiotherapy, a combination of chemotherapy and radiotherapy to treat cancer. The tests cover various types of cancer, including head and neck cancer, esophageal cancer, breast cancer, ovarian cancer, gastric cancer and colorectal cancer.
The company has five drug candidates in its pipeline. MB-D-007, targeting insulin sensitizers, and MB-N-008, targeting nonalcoholic steatohepatitis (NASH) are in preclinical stage. Cancer candidates, including MB-C-011, based on kinase inhibitors, MB-CV-021, based on peptide molecules, and MB-AC-32, which uses autophagy mechanisms, are also in development.
MBD is also conducting a precision medicine platform study to predict the responsiveness of cancer drugs for each carcinoma. Using 3D cancer tissue cultures, the biotech is trying to provide effective drugs to cancer patients through genome analysis of clinical results.
The Suwon-based venture operates precision medicine centers at the Asan Medical Center in Seoul, the Samsung Medical Center in Seoul and the Konyang University Hospital in Daejeon. The centers carry out molecular profiling and drug testing based on 3D drug screening with single and combination prognostic markers.
The company also produces medical products for cell cultures, dispensing and imaging. The Cellvitro series is a 3D cell culture platform product line for cell biology research and drug screening. It offers an environment based on alginate and collagen and allows a strong media exchange without residues.
Another product for cell dispensing, ASFA spotter, is a noncontacting high-speed liquid handling system. It facilitates the rapid dispensing of bio samples such as cells, drugs and viruses. The ASFA scanner, a cell imaging product, acquires and analyzes high-content cell images for biology research and drug screening.
Founded in 2016, Azothbio is focused on AI-powered platforms for biopharmaceutical product development. Based on convolutional neural networks, it has built the platform using deep learning systems such as recurrent neural networks, long short-term memory networks and GAN.
The Seongnam-based company has built two AI models for screening new drug candidates of low molecular weight and seven AI models for optimized drug candidate compounds and low-molecular-weight compounds.
As well as predicting new drug suitability, the drug development platforms work for compound screenings such as small molecules, macromolecules such as peptides and proteins, and phenotype-based cancer compounds. It is currently developing a peptide cancer vaccine using a neoantigen model.
Besides the AI modeling, the company produces medical products such as diagnostic devices and wound and scar treatments. The product line includes original biomedical adhesives, wound dressings, scar treatments and hemostatic agents.