© 2021 BioMarin. All rights reserved.

In their groundbreaking Nature Communications paper, scientific pioneers Lon Cardon and Karol Estrada shed new light on the decades-old question of how to identify optimal drug targets

Prioritizing genes for translation to therapeutics has challenged the pharmaceutical industry for decades. Now, BioMarin Chief Scientific Strategy Officer Lon Cardon, Ph.D., FMedSci, and Senior Director of Translational Genomics Karol Estrada, Ph.D., MSci, have proposed an approach to identifying drug targets with high probability of success by focusing on genes they’ve termed “BEST”—Bidirectional Effect Selected Targets.

Drs. Cardon and Estrada recently published their work in the leading biomedical journal Nature Communications, and they spent some time with us to explain more about this pioneering new path to drug development.


BioMarin (BMRN): Dr. Cardon, what inspired you to explore this area of genetics, as a way to answer the complex question of how to pinpoint drug targets with a high chance of success?

Lon Cardon (LC): If you look back 15 to 20 years, we didn’t know the genes for most diseases. So, when the pharmaceutical industry tried to pick drug targets, there was no real mechanistic basis to go on. We’d have some expression data in a mouse, for example, and we wouldn’t know if it was part of the defect or if it was actually causal. Since then, our knowledge about genes has gone from a trickle to a tap, thanks to the Human Genome Project and other big genome initiatives. There’s now a huge amount of genetic information out there, with thousands and thousands of genes known for diseases. The problem has completely turned around—we went from not knowing where to start, to having too many potential starting points.

We had ended up at the same place, without a good sense of the best targets to pursue. I started working on this problem about five years ago, when we went through and scanned all of the literature on every drug that had been approved or gone through clinical trial in the previous 10 to 20 years. As we reviewed the data on each drug, we asked ourselves, “What were the best predictors of success to get this drug to market?” While the very best predictor was whether someone had already drugged it, the next best predictor was whether there was any shred of genetic evidence for it. In fact, targets with genome wide association study (GWAS) evidence have at least two-fold higher probability of successful clinical development compared to targets without this evidence, and the probability increases exponentially with rare diseases.

We were particularly interested in a gene called PCSK9, which has been the target of some recent drugs designed to decrease high lipid levels in patients with cardiovascular disease. It turns out that PCSK9 has a functional capability which few other genes in the human genome possess. We’ve coined a term for this capability: “bidirectional.” Some people have a PCSK9 variant with gain of function (GoF), which effectively turns up the gene and leads to increased protein production, resulting in higher LDL levels and higher risk for cardiovascular disease. Others have PCSK9 with loss of function (LoF), which has the opposite effect—protecting these people from developing heart disease. This is represented in the graph at right.

With this bidirectionality, it became clear that turning down GoF in PCSK9 would lead to a decreased risk of heart disease, making PCSK9 a great drug target. I asked Dr. Estrada for his help in searching for other genes with this bidirectional effect, with the hope of identifying more great drug targets like this one.

BMRN: Dr. Estrada, how did you conduct the search for genes with bidirectional effect, and what did you find?

Karol Estrada (KE): Our team took a creative approach to this. We started by looking at the Human Gene Mutation Database (HGMD), trying to identify genes that have this signature of bidirectionality. The database includes annotated genetic mutations that have been historically observed in mostly monogenic forms of rare, severe diseases. We went gene by gene through thousands of registries. Of all the genes in the database, 98 genes exhibited bidirectionality.


In their study, Drs. Cardon and Estrada found that drug targets including genes with bidirectional effects were nearly four times more likely to move all the way from Phase I through approval in clinical trials.

We then cross-referenced the data on the 98 genes with a different database on clinical trial outcomes, containing information about genes, drug targets, and clinical trial success or failure. Through statistics, we explored whether any of the 98 genes had been enriched for retrospective drugs, and how likely those drugs were to move through the various phases of clinical trials. We found that drug targets including genes with bidirectional effects were nearly four times more likely to move all the way from Phase I through approval—higher than any other category of genes with any other genetic evidence, as shown in the figure below. Interestingly, the phase where BEST target-indications showed the strongest effect was the transition from Phase II to Phase III, a phase that is generally used as proof of concept for initial clinical efficacy.

BMRN: In addition to higher odds of successful transition to approval, what other important discoveries did you make about these 98 BEST indications?

KE: When we started looking deeper into the results of our statistical analysis, we saw these 98 genes were grouped in different therapeutic areas of research, such as lipids, clotting, reproduction and glucose. The group that really got our attention was genes related to height/adult stature. A total of five genes coming out of this analysis are linked to height, and of those, four are part of the C-type natriuretic peptide (CNP) pathway. BioMarin is actively engaged in research surrounding CNP.

Our Translational Genomics group has access to another large human genetics database, UK Biobank, containing in-depth genetic and health information on 500,000 UK participants. We used the Biobank to engage in further statistical analyses on these five genes linked to height. Among our most important discoveries: just by focusing on participants with rare protein-coding variants of these five genes, out of the many thousands of genes in the genome, we accounted for six percent of all individuals with idiopathic short stature. In addition, we took a closer look at one of the five genes, NPR2 (which is the receptor for CNP), and found that individuals with LoF variants of NPR2 tended to be short, while those with GoF variants tended to be tall—giving human genetic evidence that people with higher sensitivity for CNP will grow taller regardless of their genetic background.

Next, we wanted to know if CNP could rescue the cellular phenotype, suggesting that it may have efficacy in other forms of Genetic Short Stature (GSS). To answer this question, our Functional Genomics team and other colleagues in Research and Early Development created a CRISPR/Cas 9 heterozygous NPR2 knockout cell model, mimicking the effect of humans with NPR2 haploinsufficiency (loss of one copy of NPR2). Ultimately, in an in vitro model, we showed that adding CNP rescued the activation of the downstream pathway even in the presence of NPR2 haploinsufficiency. These findings have provided BioMarin with additional research targets.

BMRN: Dr. Cardon, what began as your scientific curiosity now promises to become the next generation of targeted therapeutics. Please share your thoughts on the broader implications of this discovery for patients around the world.

LC: We know that BEST genes may be associated with more than one disease. So, therapies developed to target these genes have the opportunity to go in multiple directions and target other ultimate disease points. We can consider repurposing the drugs that are produced to treat one condition in order to treat other conditions as well, with one drug working for different diseases. The potential impact is tremendous: not only can we design therapies with the highest probability of success, but those extremely effective therapies can be leveraged to address many conditions. This “solutions for multiple patients” approach gives us an even better chance to improve outcomes for those who suffer from a broad range of diseases—which has made it key to our business strategy at BioMarin, as we continue to search for solutions to the rare, intractable conditions that affect patients across the globe.


BEST genes may be associated with more than one disease—giving us an opportunity to develop targeted therapies that go in multiple directions, with one drug working for many different conditions.

For a full copy of the paper published by Drs. Cardon and Estrada in Nature Communications, please click here.