Celine Halioua
Celine Halioua

A few thoughts on choosing which disease to develop a drug for

Generally speaking, there are two approaches to developing new drugs:

Disease-first: you start with the disease and try to find drugs against new or known molecular targets that are relevant to that specific disease, or

Pathway-first: you start with molecular targets/pathway or a platform/method to discover new drugs that hit certain pathways and then figure out which disease it is best to commercialize your drug for.

When you want to develop a drug for use in people, you need to tell the FDA for which disease you are developing your drug.

A pathway is often relevant to multiple diseases. Usually, there is only enough money to explore the use of your drug in one disease at a time. That means that a drug targeting a specific pathway may be relevant to seemingly unrelated diseases like (for example) dementia and pancreatic cancer. This is often the case for longevity companies that are developing drugs to fix a way that cells dysfunction as they age, for example cellular senescence.* This makes sense - aging is a systemic process and while the diseases (“phenotypes”, the visual and symptomatic consequence of the underlying molecular action) that emerge due to the aging processes differ depending on the tissue they occur in, the pathways underlying varying diseases are often similar.

When this is the case, the drug developer needs to decide in which disease to first test their new drug. If they choose incorrectly, they may get a false negative - their drug may fail in clinic, not because their drug doesn’t work as designed, but because the disease they chose confounded the effect.

Considerations of picking a disease to target with your pathway

A few critical questions are:

  1. The relevance/predominance of pathway to the disease

  2. The standard of care adequacy or lack thereof 

  3. Clinical pipeline for this disease (competition)

  4. Commercial considerations

1. Relevance of the pathway to the disease.

One way to think about the relevance of your drug target to a specific disease is to ask yourself:

What causes this disease? How many different ways can you develop this disease? What increases or decreases your risk of developing this disease? What drives or promotes this disease after initiation? What stages of disease are there? Does this differ across patients? How strong are these drivers? What pathways lead this, versus are “accessories” to the disease propagation?

Cancers are an excellent example of this. One hallmark of cancer is proliferative growth signaling. This phenotype - extreme growth - can be caused by a number of pathways, e.g., overactive EGFR (epidermal growth factor receptor) signaling, overactive IGF (insulin-like growth factor) signaling, deletion of apoptosis-inducers such as p53, increased telomerase expression which results in decreased apoptosis induction… see where I am going with this? If you are developing an inhibitor for EGFR (of which there are many on the market already), you may want to select for cancers where this is the predominant pathway of cell growth. Even so, the cancer cell may mutate around this blockade and evade your drug, leading to your drug only working for a short period of time

For a pathway to be a potentially efficacious drug target, it will need to play a predominant or gate-keeping role in the initiation and/or propagation of the disease. Pre-clinical models are often simplified in this sense (e.g., disease initiated by a single gene knock in/out) - you should consider here how the disease manifests in patients. You should have some basis for your hypothesis that if you (ant)agonize your drug’s target, it may have a clinically-meaningful* benefit.

Developing a drug is hard. Showing efficacy is challenging and only demonstrated  by a small fraction of all the drugs that are tested. In certain diseases (e.g., highly heterogeneous disease), this is even harder. There may be confounding factors in some diseases that cause your drug to fail even if it affecting your pathway as desired. 

Some diseases have robust animal models that can provide proof of concept and decent insight into the function and efficacy of your drug. In others, many drugs which have ‘cured’ the disease in vivo have shown little to no efficacy in people

Similarly, the toxicity profile and considerations will be influenced by your disease. For terminal disease with no good SOC, there is more leeway and tolerance for adverse events, while a drug for a generally healthy population will not have tolerance for a certain level of toxicities.

*clinical meaningful means that the drug not only has the desired effect on the pathway/molecularly, but that it improves symptoms or disease progression to a degree that is quantifiable and noticed by the patient

2. The Standard of Care’s Adequacy (or lack thereof)

The standard of care (SOC) is the current best treatment or drug for the disease. There may not be a SOC for the disease you are targeting, or it may not be disease modifying (e.g., the drug mitigates symptoms but has no effect on the disease progression and the ultimate outcome on the patient). In this case, it is important to dig into the attempts at drugging this disease previously, why they failed, and how they measured clinical success.

If there is a therapy on the market for your disease, there are a number of things you should consider next.

What is the standard of care? How well does it work? What metrics/biomarkers determine success/response? On this scale, how effective is it? How long does the benefit remain? What portion of patients does this drug work in? Are there a subgroup of patients that respond especially well/do not respond well? What are the adverse events associated with this drug? 

Generally speaking, you may have a market opportunity if one or multiple of the following are true:

  • The SOC only works for a small subset of the population

  • The SOC is not curative

  • The SOC is only partially disease modifying

  • The SOC is only palliative

  • The SOC has undesirable adverse events

  • Your drug can be cheaper 

  • Your drug is easier to take/more pleasant (e.g., SOC requires hospital infusion, yours is an oral tablet)

3. Existing clinical pipeline & generics

In addition to understanding the SOC, you should also understand what other drugs are in the clinical development pipeline. Due to the long regulatory pathway from drug conception to market approval (often 7+ years), you are unlikely to beat a drug already in development to market if it succeeds in clinical trials. This is not always killer; different Pharma companies often develop drugs that work similarly, with market share won by the best working drug or the best commercialization team. Prevalent diseases have room for multiple blockbusters ($1B+/year revenue drugs). Indeed, other Pharma companies working on the same biological hypothesis can actually be derisking and offer acquisition opportunities for you if another Pharma wants to jump intro the fray without developing their own drug from scratch.

A new drug may also beat you on convenience or price - e.g., if your drug requires an intraocular injection, but there is an eye drop which works similarly as well that receives market approval, patients may not purchase your injection even if it works slightly better than the oral drug due to the discomfort of eye injections. 

All US clinical trials must be reported on clinicaltrials.gov. This is a great resource for figuring out what is in development. Pharma companies usually present their newest data at conferences - ASCO (oncology), JPMorgan (everything), and ARVO (ocular) are three popular conferences.

4. The market

There are numerous factors to consider when evaluating the market opportunity for an indication:

  • Prevalence (how many patients have the disease right now)

  • Incidence (how many patients develop the disease per year)

  • Disease stages

    • What subset of patients are treatable by your drug?

    • Is there a certain stage past which the disease is not treatable anymore? How big of a portion of this is of the total disease prevalence? 

  • What is the SOC? How many patients are treated by this, well served by it, etc?

  • What other therapies for this are at a high likelihood of approval? 

    • How many clinical trials are in progress right now? Are there enough patients remaining to conduct your clinical trial? How distributed are they?

  • What is the trendline for this disease - do we expect incidence to increase or decrease?

  • Are there certain countries/geographic locations that have a higher or lower incidence of your disease?

  • What is the price tolerance of a therapeutic for this disease? 

    • How long would a patient take your therapeutic

    • How much benefit would it deliver

    • Would it deliver increased quality of life, increased life years, both, neither?

  • Are there regulatory advantages that will allow you to get to market faster/do your clinical trial cheaper than normal?

Additionally, upcoming generics/biosimilars may impact the market for your drug. Even if your drug is 20% more effective than the generic, if the generic if 50% cheaper than your drug (which, it likely will be), your drug will not be likely to become first-line therapy but instead only be used in patients which are non-responsive or refractory to the cheaper drug

Getting conviction that your drug might work in people

Before you can give your drug to humans, you have to submit an Investigational New Drug application (IND) with the FDA. The IND will include the pre-clinical data package which includes your rationale and evidence for potential therapeutic action for your disease of choice and robust data supporting its safety in humans, along with your ability to consistently and safely manufacture your drug.

When selecting the disease, you should consider how you will develop conviction that your drug will show pivotal efficacy in this indication before you invest in filing an IND and beginning human clinical trials. Generally, this is achieved by in vivo studies of an animal model of the disease. This can be supplemented by, for example, retrospective genetic data. (e.g., there is a congenital form of X cancer where the patients have a up-regulating mutation in the receptor you want to target and inhibit with your drug).

Many animal models are not great representations of the disease they mimic. Often they are phenotypically similar but induced by a modification or treatment that is much more simple than how the disease actually develops in humans. This is especially true for many diseases of age, where the true cause of the disease it often multifactorial and heterogeneous across people. Unfortunately, until better models are broadly available, it is still necessary to use inadequate models to gain at least some insight that your drug may work.

What type of drug should you develop?

Broadly speaking, certain modalities (think: small molecule, antibody, peptide, gene therapy, cell therapy, etc) are more amenable to some diseases than others. If you have a genetic disease, or a disease driven by a loss-of-function, gene therapy may be beneficial. However, gene therapy is expensive and systemic distribution can be costly. If you are targeting a disease in the eye, a biologic may be more advantageous as dosing the retina with a small molecule can be challenging for a number of reasons. A small molecule may be more interesting for diseases which are systemic, for intracellular targets, or markets that will not tolerate high prices/ already have a somewhat effective SOC.