A Hit is Not a Drug – Industry’s Perspective
Drug discovery of small molecules is a complicated process requiring many techniques and diverse expertise. The aim of this document is to help researchers understand the steps required to convert a hit to a potential drug candidate from an industry perspective.
The Hit is the beginning
With modern high-throughput screening technologies, biomedical researchers are discovering ‘hits’ at an unprecedented pace. ‘Hit’ is the term used to describe a chemical compound that typically emerges as a result of screening a library of compounds with a screening assay. Usually multiple ‘hits’ emerge, some of which might serve as an early prototype for a potential drug. A ‘hit’ is not a drug, and is not even a ‘lead candidate’. At the most, it might be the starting point for initiation of a medicinal chemistry program.
It is important to understand that many times ‘hits’ come from commercial libraries that have many overlapping molecules, so that what appears to be an independent hit might in fact be represented in different libraries.
Often, researchers perceive a hit, identified at the very beginning of the drug discovery process outlined in Figure 1, as a promising drug or development candidate. However, to achieve a drug product on the market, several additional development steps are required after hit discovery, and the time taken can be another 12-15 years at an average cost of US$1.3 billion.
Figure 1: Drug discovery process from target identification and validation through to filing of a compound with a regulatory body such as the FDA. Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3058157/
So while hits may provide an exciting start to a drug development project, they are still only very early stage potential candidates and it is important to consider them in the context of the intricacies and complexities involved in taking a candidate from the bench to the bedside.
Target identification and validation
Prior to identification of a hit, the target to be modulated must be identified and validated. Click here to read more in our Target validation factsheet.
In summary, a good drug target needs to be:
- Address unmet medical need (Read more in our Unmet Medical Need factsheet)
- Meet commercial need
- Be ‘druggable’
A ‘druggable’ target is accessible to the putative drug molecule, which, upon binding, induces a response which can be measured both in vitro and in vivo.
When thinking about validation of a target in the pre-clinical context, the goal is to establish multiple lines of evidence that independently reinforce the notion that the target might be valuable when modulated in the clinic.
Pre-clinical target validation studies should demonstrate (i) that the target of interest has potential for a biologically meaningful effect on modulating disease onset or progression, (ii) that the target can be modulated therapeutically and (iii) addressing the target will provide a therapy that is superior to existing or emerging therapies.
Furthermore, studies should demonstrate why responses observed by therapeutically modulating a target are being observed, and how they can be even further improved. These studies need to be conducted in a blinded manner.
Once the target is validated, the drug discovery project can begin in earnest. There are a number of approaches that can be taken to generate initial chemical hit matter and this first stage is referred to as Hit Identification. Methods to generate a hit include
- Knowledge-based design (on the basis of literature compound for instance)
- High throughput screen (HTS)
- Fragment based drug design (FBDD)
- Virtual screening
Following hit identification comes the long and complex process of Lead Optimisation (LO). Depending on the quality of the initial hit, this often includes a Lead Generation (LG) stage. LG and LO are both an iterative process involving multiple cycles of medicinal chemistry design, synthesis and testing of the multiple parameters necessary to make the compound drug-like. Simultaneously, medicinal chemists need to optimise the following, so that the resultant drug is both effective and safe:
- Metabolic stability
Of course, many of these properties are inter-related and not always entirely predictable. This phase of lead discovery can often be “one step forward, two steps back”, or even three steps sideways!
Nonetheless, a good team of medicinal chemists will eventually prevail and at the end of the lead optimisation stage, a preclinical development candidate is selected. This will very likely be quite different from the original hit and is still a long way from being a drug, as now a compelling package of preclinical data must be generated.
Compelling preclinical data
Once a development candidate is declared, establishing drug action and appropriate biomarkers are the next key requirements for early phase drug discovery.
- Pharmacokinetics and Pharmacodynamics (PK/PD)
- Drug Exposure and Dose-response
- Cmax vs AUC (maximum concentration vs concentration over time)
- Evidence of superiority vs existing treatments
Pharmacokinetics and Pharmacodynamics (PK/PD)
Pharmacokinetics and pharmacodynamics (PK/PD) studies are necessary to demonstrate the novel drug’s impact on pathway and its specificity. Integration of PK/PD studies into early development not only guides compound selection, but also efficient clinical development strategy.
Effective PK/PD study design, analysis, and interpretation help to:
- Understand the drug’s mechanism of action
- Elucidate the relationship between dose and effect
- Identify properties for further improvement and optimal drug design
In addition, PK/PD modeling helps to:
- Increase the translation of in vitro compound potency to the in vivo setting
- Reduce the number of in vivo animal studies
- Improve translation of findings from preclinical to clinical
Drug Exposure and Dose-response
Drug exposure and dose-response assays are essential for the interpretation of both efficacy and toxicity.
These assays must:
- Be able to relate in vitro IC50 to in vivo effects
- Measure exposure to determine if sufficient levels of drug achieved in vivo in blood and/or in organs and tissue
- Prove target engagement & effect (through PD assay)
Cmax vs AUC
Ascertaining both Cmax and AUC is critical to understanding desired drug characteristics, whether there is a clear dose-response relationship and whether the response is driven by maximum concentration achieved (Cmax) or for the duration of “target-coverage” (AUC).
Cmax is the highest serum concentration seen for a drug after a single dose.
- For an oral drug, Cmax is dependent on the rate of drug absorption and its disposition profile
- Short term side effects are most likely at or near Cmax
AUC is the total drug exposure over time and quantification of AUC indicates a drug’s bioavailability. AUC is also used to determine if alternative methods of drug delivery at equivalent dose (e.g. injection vs tablet) will release the same dose to the body.
Evidence of superiority vs existing treatments
To be translatable your research should demonstrate a clear advantage for your new therapeutic when compared to:
- molecules already on the market
- molecules in late stage development
- other forms of treatment for the disease indication
While the ultimate “biomarker” is clinical response, biomarkers are tools that can be measured as an indicator of response to a therapeutic intervention. Biomarkers increase the success rate of drug development programs, and inform regulatory and therapeutic decision making.
There are six classes of biomarkers:
- Target-coverage assay
- Pharmacokinetic (PK) assay for exposure
- Pharmacodynamic (PD) assay for activity on the biochemical pathway
- Surrogate assay for response
- Surrogate assay for safety
- Patient stratification assay
Industry needs to see the following to be confident that preclinical data is robust and justifies clinical development of a new drug:
- Blinded studies (with positive and negative controls)
- Repeat (and reproducible) experiments
- Comparison with standard of care benchmark including exposure data
- Appropriate statistical analysis and presentation of data (SD not SEM)
- Dose-response data
- Randomised animal studies with pre-specified power calculation
- Demonstration of PK/PD/biodistribution relationship