CHICAGO – With its newly announced partnership with CellPhenomics for cancer drug target discovery and validation, German precision oncology firm Indivumed hopes to speed up development of new therapeutics.
Under the deal, Berlin-based CellPhenomics will build organoid tumor models so Indivumed can validate its therapeutic targets with artificial intelligence-driven analytics.
CellPhenomics is contributing expertise in 3D cell models as part of the validation process. The company will also advise Hamburg, Germany-based Indivumed on the creation of experiments while the latter retains ownership of the data and intellectual property of therapeutic targets.
Specifically, Indivumed plans to identify phenotypically linked therapeutic targets with its year-old nRavel AI-driven bioinformatics platform for precision cancer research. Then CellPhenomics will create organoid models from biospecimens and run in vitro tests, so Indivumed can validate "therapeutically relevant" targets, according to Jonathan Woodsmith, VP of advanced analytics and AI at Indivumed.
"The great advantage here is that the same high-quality patient material is used for organoid creation that was also used to generate the molecular data," Woodsmith said.
Data comes from de-identified tissue samples taken from a network of partner clinics and hospitals in the US, Brazil, Singapore, Japan, South Korea, and several European countries, part of a conscious effort to diversify the company's datasets. "It makes sense from a therapeutic targeting perspective as well," Woodsmith said.
CellPhenomics, a 2014 spinout from OncoTrack, a cancer research consortium funded by the European Union's Innovative Medicines Initiative, and a sister company of oncology testing firm ASC Oncology, has established proprietary protocols for each tumor type it builds models for. "The way we, for example, build a lung cancer model is significantly different from how we build a sarcoma model," CEO Christian Regenbrecht said via email.
With Indivumed, CellPhenomics receives "high-quality tissue" to build its models, "which results in faster, more accurate, and more successful creation of the organoids," Regenbrecht said. This also speeds up the validation process.
CellPhenomics has built its own database linked to its biobank. "This allows for rapid analysis and comparison of genotype, phenotype, drug response, and proteomic data," Regenbrecht said.
He added that the company aims to provide a holistic understanding of each tumor as part of its menu of preclinical services. CellPhenomics customers include small startups in the pharmaceutical world as well as seven out of the 10 largest global pharma firms, he said.
Woodsmith said that Indivumed is currently augmenting nRavel for the new effort, while CellPhenomics has begun creating organoid models for 10 solid tumor types, covering colorectal, non-small cell lung, gastric, liver, kidney, pancreatic, prostate, ovarian, bladder, and breast cancer. He said there will be a "strong initial focus" on colorectal tumors because of the high medical need for new targets.
The company hopes to have the first targets validated during the first quarter of 2023.
Indivumed is using support from the EU's European Regional Development Fund for the project, which awarded the firm €4.3 million ($4.3 million) in January to accelerate cancer drug discovery and precision oncology. Woodsmith said the EU grant is meant to underwrite work on an AI-based platform for accelerating the development of precision therapeutics for cancer.
The company now has a multiomic data pipeline from its tissue bank to support in silico target identification through its IndivuType business unit, he said, which features a multiomic cancer database. The partnership with CellPhenomics adds the ability to validate those targets in vitro.
Indivumed collects about 320 data points for each sample, covering genomics, proteomics, phosphoproteomics, transcriptomics, and microRNA, according to Woodsmith. The company produces about 250 gigabytes of data for each sample by running the patient-specific data through IndivuType to help identify targets.
CellPhenomics then takes the tissue to create disease models for Indivumed to validate targets its earlier research has identified. Indivumed, thus, will be able to offer CellPhenomics' services to its own customers.
However, Indivumed, which has about 200 employees, is not large enough to see its findings through to drug discovery, so Woodsmith said that the firm would be looking to team up with pharma companies in the future.
"We have a relatively flexible model whereby we can actually out-license or partner with [drug] companies in various stages," he said. "This may be prevalidation, it may be postvalidation, it may even be post lead-hit identification."