The role of biomarkers in making personalized medicine a reality is captivating researchers despite development, clinical validation, and regulatory challenges. A recent BCC Research report estimates that the global biomarker market will reach $12.8 billion by 2012. Many companies and organizations covet a piece of this market and are actively seeking to streamline biomarker research. Two upcoming conferences—IBC’s “New Frontiers in Cancer Drug Development” and CHI’s “Accelerating Development & Advancing Personalized Therapy”—will provide forums for discussions about ongoing biomarker research efforts.
A focus on signaling pathways has provided Curis with a novel approach to drug discovery. “There’s a lot of redundancy built into these pathways, so if you have a highly specific blockade, you may get an initial response, but the majority of time tumors can adapt and bypass it. So our scientists focus on how to more effectively block the signaling and be able to knock out a network instead of a distinct node of intervention,” explains Daniel Passeri, president and CEO.
More specifically, Passeri says, preclinical data has shown that if one uses a deacetylase inhibitor with a kinase inhibitor, there’s a synergistic intervention to knock out key networks. His research group found they could build a single drug scaffold with an active kinase moiety that specifically binds to EGFR and includes an HDAC inhibitor.
Passeri says that there are potentially several competitive attributes to this approach that have yet to be seen in the clinic, though there are early promising clinical results. One of the challenges in combination therapy is striving to hit two or three distinct targets within a patient population with different chemical entities. “The pharmacokinetics of these drugs are typically distinct so the drugs are degraded at different rates, and they may also have different pharmacodynamics, which may affect patient compliance.”
The company’s recently patented drug, CUDC-101, hits multiple targets in cancer cells, including HDAC, EGFR, and HER2. It has aligned pharmacokinetics so the drug is degraded in a coordinated manner and provides a better toxicity profile since more of it is concentrated in the tumor, Passeri reports. This novel agent also has a potency enhancement of approximately five- to tenfold over the prototype-approved HDAC inhibitor, he adds.
Since it is administered intravenously, it bypasses the gastrointestinal tract and enables higher dosing levels. “The key advantage here is that we’re able to achieve network disruption with a single agent, which we believe will be more efficacious and better tolerated, and also provide a significant cost advantage since it’s a single drug.” CUDC-101 is currently in Phase I dose-escalation studies. Next in development is an HDAC p13-kinase inhibitor.
Individualized Cancer Treatment
Genstruct is applying the concepts of artificial intelligence to develop a three-tier architecture for determining biological mechanisms, in an attempt to increase the speed and success of drug discovery and development. Knowledge Assembly® is based on the company’s large generic knowledge base, which can be customized and augmented to specifically represent each cell type, tissue, or organ system, explains Keith Elliston, Ph.D., president and CEO.
“We’ll take that information with actual experimental data and run it through reverse causal analysis to find what caused that data to be produced; it represents biology that has changed with a perturbation like disease or toxicity. This is a subset we call the Causal Network Model—the cause-and-effect reactions that occurred, to change from a normal state to a disease state,” Dr. Elliston says.
This approach overcomes some of the key challenges of designing biomarkers. “Our system allows us to take different data, like gene-expression data, to build a model and design the biomarker, this can be a protein, an imaging reagent, anything the model tells me about the biology of the system. I then know why that is a marker for a specific mechanism because I have a complete mechanistic model of how that particular function works.” Furthermore, he adds, the mechanistic model is 20 times more efficient than a direct empirical method.
Models of patients can be compared to map out the entire extent of functional variation between individuals for disease state and drug response. This is especially pertinent for modeling individual tumors to define dysregulated pathways and guide therapy.
MyPath™ builds a network for a patient’s disease state to help match the patient to the best therapies. Dr. Elliston says the company has begun implementing this assay in a clinical cancer setting, and it works well with both solid and hematological cancers. “We think of it as individualized medicine. Our models can, with a small assay, define the entire state of the patient’s disease and then match that to the therapies right for that patient’s biology.”