The Major Disconnect in CNS Research
The future of CNS research is predicated on an alignment of preclinical and clinical endpoints
Of all major therapy areas, CNS is the one with the highest unmet need. Most major CNS diseases have limited pharmacology treatment options. For major neurodegenerative disorders such as Alzheimer’s disease and Parkinson’s disease, there are no treatments that reverse or even halt or slow the progression of the disease. Treatment options that target symptoms are limited and will only manage a small subset of the symptoms and many treatments work only in a subset of the population and/or are effective for a limited period during the disease progression. For CNS diseases where there are treatment options available, the drugs are often associated with significant side effects that can further complicate disease progression. Indeed, in many of these cases, there are secondary medications that are designed just to treat the side effects.
Despite enormous efforts to come up with new options that advance our ability to treat CNS diseases, researchers have had limited success. In instances where we have managed to get novel treatments to patients, the drugs are often reformulated versions of existing medications. This lack of success has been due to the high failure rate of CNS drugs in Phase II & III trials, even when the clinical candidates demonstrated a clear efficacy signal in animal models.
The disconnect between the preclinical animal studies and clinical outcomes is a serious problem in CNS drug discovery and development. It has often been described with the relatively simplistic phrase “the animal models fail to predict clinical outcomes,” which tends to focus the blame on the work done preclinically. However, this really doesn’t capture the full complexity of the preclinical-clinical disconnect. In fact, the intersection of the biology, the disease and the pharmacology is extremely complicated, poorly understood and is often not sufficiently studied prior to advancing a novel therapeutic to the clinic. Examples of this complexity include the fact that:
- Treatments often create an inverted U-shaped dose response curve where neither too little or too much drug is effective.
- The doses and dose regimens change as the patient (or preclinical model) age and as the disease progresses, often making it difficult to know when to intervene with a patient in their disease process.
- Genetic variants of a single disease can behave differently in response to a drug treatment.
- Patients with CNS diseases are often on multiple medications and we don’t always understand how novel medications will interact with existing treatment regimens.
- Chronic dosing often produces tolerance or sensitization, thus changing a patient’s response to treatment over time.
- Withdrawal of drugs can create a rebound effect.
- The effect of the target engagement biomarkers, disease biomarkers, pharmacokinetics and efficacy endpoints may behave differently following different temporal dynamics; thus there is s no single time point that captures the relationship between the different measures.
- The critically important measure of the free fraction of the drug in the brain cannot be measured directly in the clinic and must be inferred from other measures.
- Species, or even strain differences in animal models are often poorly understood and are difficult to contextualize in terms of how they predict clinical outcomes.
- Receptor kinetics and the interaction of the receptor and a ligand are still not fully understood.
Collectively, these challenges have made building the necessary data platform in advancing drugs to clinical trials challenging for CNS diseases. Further complicating the issue is that there are usually massive disconnects between what is measured in animal models vs. what is measured in the clinic. As a field, CNS discovery and development continues to improve our understanding of how the tangled mess of complications fit together and we are building increasingly sophisticated preclinical data packages that can support clinical decisions related to patient selection, dose selection, optimal dose regimen, and exclusion criteria including what existing medications patients can remain on during the clinical trial.