HONG KONG South Korean AI-based biotech Azoth Bio Inc., of Seongnam, Gyeonggi-do, and biopharmaceutical venture Wellmarker Bio Co. Ltd., based in Seoul, have signed a memorandum of understanding for cancer drug R&D and commercialization. Under the agreement, the two entities will use Azoth's AI-powered platform to develop Wellmarker's cancer treatment candidates.
Using the AI platform, Wellmarker will develop WM-S2, a small molecule-based drug candidate with a first indication for colorectal cancer. The drug is in preclinical studies, including patient tissue analysis, and is expected to enter phase I in early 2021.
Since it was founded in 2016 , Azoth has been developing its in-house AI-based platform using deep learning techniques, which include convolutional neural networks, recurrent neural networks/long short-term memory, and generative adversarial networks.
Using those techniques, the company operates six platforms for drug discovery: phenotype-based screening, target-based screening, cytochrome P 450 (CYP 450) screening, automatic molecule generation, immuno-oncology peptides work and target activity. The platforms process various data related to drug, genome, protein and electronic medical records data. Specifically, each platform predicts the following: the activity profile of small molecules for more than 950 human cancer cell lines; the activity profile of small molecules for more than 290 human kinases and small molecules for G protein-coupled receptor proteins; the inhibition activity of small molecules against five human-CYP enzymes; structure-activity relationships based on experimental data; putative major histocompatibility complex-I binding peptides to develop cancer vaccines; and generation of new compounds that are expected to be active in the target area.
"We will use the Azoth's AI-powered platform for new drug R&D, treatment reaction expectations and biomarker research," Young-il Lee, chief security officer at Wellmarker, told BioWorld. "We believe that the platform will save time and money in hypothesis testing in clinical studies, pipeline development and even commercialization."
Azoth's AI-based platforms cover in-house development and external partnerships. In-house R&D includes big data analysis, target identification, hit/seed compounds discovery, and optimization. Partnerships with other biopharma firms and contract research organizations include lead selection and validation. The platforms' goals are to accelerate the drug development process up to the stage of pre-IND and clinical trials.
"We aim to shorten the drug development period by up to one year from screening to lead selection before preclinical trials with animals, by using AI deep learning experts," Azoth's business development manager, Jerry Maeng, told Bioworld. "Our platform has increased the accuracy and reliability by processing various types of data."
In addition to WM-S2, Wellmarker develops various new drugs based on predictive biomarkers. Its lead candidate, WM-S1, aims to treat patients with colorectal cancer who have not been treated by Merck & Co. Inc.'s Erbitux (cetuximab).
The company said its team identified Erbitux primary resistant related proteins, named the cetuximab-resistant gene (CRG). Wellmarker's WM-S1 is a CRG target inhibitor, which has shown potent in vitro enzyme activity, high in vitro anticancer activity, and strong anticancer efficacy in vivo xenograft models and patient-derived xenograft models. WM-S1 will enter the clinical trial stage next year.
Another candidate, WM-A1, is a cancer immunotherapy-related gene (CMG) therapeutic antibody. It targets non-small-cell-lung carcinoma, liver cancer and gastric cancer. The company said WM-A1's in vivo efficacy was assessed in mouse lung cancer cell-derived syngeneic model, which has shown the drug could be a promising therapeutic agent.
Other drugs in development include WM-P1 and WM-A2, which target liver cancer and colon cancer, respectively.
As a spin-off from a major hospital of Korea Asan Medical Center, Wellmarker has focused on biomarker R&D since its foundation in 2016.
The company secures predictive biomarker programs using patients' tissues, cells and animal models. The company is developing four biomarker programs targeting colon cancer, gastric cancer and melanoma, of which three are in preclinical stages.
The venture raised $18 million in a series B funding round in July, following $28 million in series A investment and $3.8 million in seed funding last year.