Reducing Pipeline Attrition
Leading-edge strategies for improving drug discovery include development of multitargeted therapeutics, whole-pathway approaches, biology-driven drug discovery, analysis of multigenic complex diseases, and network pharmacology. These approaches are aimed at discovery of drugs to address complex diseases, for which more conventional drug discovery methodologies have proven to be inadequate.
Notably, biology-driven drug discovery, which starts with a disease model, a pathway, or a biological process has proven to be much more successful than technology-driven drug discovery based on genomics. Biology-driven drug discovery has often utilized academic research into pathways, disease models, and other biological systems, which have often been conducted over a period of years or of decades.
Targets derived from this research are usually better understood and validated with respect to their role in disease than targets that have recently emerged from genomic studies and were validated by laboratory testing in pharmaceutical or biotech companies.
Examples of drugs derived from biology-driven drug discovery include such large molecule drugs as Genentech’s Herceptin (trastuzumab) for breast cancer, Biogen Idec/Genentech’s Rituxan (rituximab) for non-Hodgkin’s lymphoma and rheumatoid arthritis, Genentech/Roche’s Avastin (bevacizumab) for colorectal, breast, lung and renal cell carcinoma, and Erbitux (cetuximab) from ImClone Systems, a subsidiary of Eli Lilly, for colorectal and head and neck cancer.
Notable small molecule drugs include the numerous protein kinase inhibitors for treatment of cancers that have reached the market in recent years.
The value of biology-driven drug discovery is reflected in measures of success in development. Kola and Landis noted that biologics showed higher rates of success than small molecule drugs (approximately 24% as compared to 11%). Biologics tend to be discovered via biology-driven R&D, often beginning in academic research laboratories or in companies that conduct basic research such as Genentech.
A recent analysis also found that among 974 anticancer agents that entered clinical trials between January 1995 and September 2007, the clinical attrition rate was 82%. For the subset of these drugs that consisted of targeted kinase inhibitors only 53% attrition was seen. The lower attrition rate of kinase inhibitors was attributable to a lower Phase II attrition rate.
Kinase inhibitors have been developed via biology-driven drug discovery (based on studies of signaling pathways in normal and cancer cells). Because of the highly targeted nature of kinase inhibitors, researchers can often also identify biomarkers that allow for better patient stratification and improved design of clinical trials of these drugs.
Industry researchers and other experts, as well as the FDA, also cite animal models that are poorly predictive of efficacy and/or safety in humans as a major cause of pipeline attrition. The respondents in Insight Pharma Reports’ drug attrition survey agree, since over 50% of them cited poorly predictive animal models as a major reason for the low productivity and high cost of drug development.
Since animal models are used both in drug discovery and in preclinical testing, this issue affects both these stages of drug development, and the results of poorly predictive animal studies often cause attrition in Phase II or Phase III. Therapeutic areas in which animal models of efficacy are notoriously unpredictive, especially oncology and CNS diseases, are also the therapeutic areas with the highest rates of Phase II and Phase III attrition, usually due to efficacy failures.
“I think that it’s important to understand how predictive the animal models have been, and which ones translate well into humans,” said Bruce H. Littman, M.D., president of Translational Medicine Associates, and the former vp, global head of translational medicine at Pfizer. “Some will be very mechanism specific, and some will be more disease specific.”
Proof-of-Concept Studies in Humans
In early clinical development, translational medicine strategies emphasize designing clinical trials aimed at obtaining rapid POC in humans. The goal is to enable companies to rapidly and cost-effectively advance drugs that achieve POC into Phase II trials, and to eliminate drugs that do not achieve POC. Biomarkers are key to the design of POC clinical trials.
For example, Novartis has adopted a drug discovery and development model based on biochemical pathways. In many cases, rare Mendelianly inherited familial diseases are caused by mutated genes that disrupt pathways that are also involved in more common, sporadic diseases. Novartis researchers design small POC clinical trials in patients with the genetic disease. Upon achieving POC, the drug may also be tested in patients with complex, sporadic diseases that involve the same pathway.
The first drug that Novartis has been developing using this strategy is the mAb drug Ilaris (canakinumab), which specifically targets the proinflammatory cytokine interleukin-1b (IL-1β). The company conducted its initial POC trial in three patients with Muckle-Wells syndrome. This is a rare autosomal dominant disease caused by a mutation in a gene involved in processing IL-1β.
Macrophages from Muckle-Wells patients constitutively secrete this cytokine, resulting in chronic inflammatory symptoms including skin rash, periodic arthritis, deafness, and chronic fatigue. When the patients were treated with Ilaris, their rashes, as well as biochemical markers of inflammation, resolved in several days, according to the company.
Novartis went on to test the drug further in patients with cryopyrin-associated periodic syndromes (CAPS), a group of rare inherited auto-inflammatory conditions that includes Muckle-Wells syndrome as well as several other conditions, all of which result in overproduction of IL-1b. In June 2009, the FDA approved Ilaris for treatment of CAPS, which affects approximately 7,000 patients worldwide.
Novartis is currently testing Ilaris in more common diseases in which the IL-1b pathway is thought to play a major role, including chronic obstructive pulmonary disorder, rheumatoid arthritis, type 2 diabetes, and gout. The company reports that it is using biomarkers to predict response to treatment, with the goal of providing patients with a personalized approach to treatment of their disease.
In June, Novartis presented a Phase I/II study of treatment of children with systematic juvenile idiopathic rheumatoid arthritis at the “Congress of the European League Against Rheumatism.” In this open-label study of 19 patients with acute disease, a single dose of Ilaris enabled the 59% of patients who were responders to reportedly achieve 50% control of their disease (as measured by standard American College of Rheumatology criteria) within 15 days.
More generally, Novartis’ early clinical trial strategy involves testing drugs in POC clinical trials in small homogeneous populations, either with a rare genetic disease or defined by biomarkers. Upon achieving POC, Novartis then goes on to conduct conventional Phase II–Phase III trials aimed at registration of the drug.
Other companies are also using biomarkers to define patient populations for POC clinical trials and to help determine the results of these trials.
Biomarkers constitute a young discipline, and there is a need to identify more biomarkers and to qualify and validate them for use in POC trials and other types of early clinical studies. The FDA’s Critical Path Initiative emphasizes biomarker development and use in drug development. Such research consortia as the Biomarkers Consortium, the Alzheimer’s Disease Neuroimaging Initiative, and the High-Risk Plaque Initiative are attempting to improve the state of biomarker science and technology via collaborative, precompetitive studies. The vast majority of biomarker development, however, takes place in individual academic laboratories and companies.
“Biomarkers and translational medicine are something the whole industry is struggling with, to be honest,” said Evan Loh, M.D., vp of clinical R&D at Wyeth. “With the biomarker concept, we are struggling to determine the true value that any biomarker gives us relative to increased confidence in decision-making.”
According to Peter Lassota, Ph.D., divisional vp, imaging biology and oncology at Caliper Life Sciences, “Biomarkers have many flavors. We need better markers for early detection of disease, particularly cancers, in order to improve therapeutic outcomes. We are in need of surrogate markers of efficacy. Pharmacodynamic markers used in preclinical drug discovery and development will accelerate moving the compounds through the pipeline.”
Two other strategies for early clinical design are Phase 0 human clinical studies using microdosing (i.e., administering doses of a drug that are too low to induce pharmacologic effects), and adaptive clinical trials. Phase 0 trials are first-in-human studies that may be used to evaluate pharmacokinetics, pharmacodyamics, and/or mechanism of action, or to evaluate imaging of specific targets.
Adaptive clinical trials constitute a flexible trial design that allows, for example, for continuous dose selection and refinement of hypotheses. In early clinical development, a company may use adaptive design to look at dose-response relationships. The goal is to reduce failed late-stage trials due, for example, to using the wrong dose of a drug or treating the wrong group of patients.