Axial - Arvinas
Surveying great inventors and businesses
Arvinas is a pioneer in bringing directed protein degradation to patients and making currently undrugged targets tractable. The company was founded in 2013 by Professor Craig Crews. The underlying technology was invented by the Crews lab - proteolysis targeting chimeras (PROTACs). Craig Crews is the epitome of a great inventor whose work not only explores nature but creates new tools to improve human health. The hallmark of a great inventor is the ability to be successful in multiple fields of study - being able to be a leader in proteostasis is important but then going onto develop tools to control it is incredible; that’s what the Crews lab did. Even before proteostasis, Craig was world-class in the study of signal transduction. Before Arvinas, Crews had co-founded Proteolix to develop proteosome inhibitors. The drug Proteolix set off with, carfilzomib (Kyprolis) was approved by the FDA for multiple myeloma - used in over 80K patients. Ultimately, Proteolix was acquired by Onyx Pharmaceuticals, which was then acquired by Amgen in 2013 for over $10B. Crews was in the perfect position, financially and experientially, to take on a larger technical risk with Arvinas. PROTACs have been ~15 years in the making and Arvinas paved the way for other degraders to come to market (i.e. molecular glues). Given Crews’ history, his next company ought to set the bar even further.
Arvinas is important for its ability to bring a new modality to market itself. It’s a great case study for other companies to understand what made Arvinas successful and what other types of business models can be designed.
Key findings
A fundamental understanding of proteolysis was required to bring a new drug format to market - PROTACs/degraders. With decades of work, Arvinas’s PROTAC technology has several unique molecular properties to be competitive in various oncology indications where resistance becomes an issue.
Arvinas took no market or target risk in order to make platform validation more probable and enable the company to transition toward higher-reward, higher-risk indications and targets.
Arvinas is one of the best examples of what it takes to go from invention to humans leaving the opportunity for new companies to define design rules for degraders and build unique business models.
Technology
PROTACs (image below) are bifunctional small molecules that bring the target protein and an E3 ligase together to direct the target toward degradation. By the 1970s, proteolysis was properly characterized. In 1999, a small biotechnology company, Proteinix, filed a patenton a bifunctional molecule that engages both a target and an E3 ligase to promote target ubiquitinationand degradation. A few years later, PROTACs were designed and being used with the first application in 2001.Crews collaborated with Raymond Deshaies (co-founder of Proteolix) for this first paper. Like great inventors, Crew and his lab get developing the technology whereas others moved on (i.e. Proteinix). The real technical breakthrough was in 2008, when the Crews lab was able to design a purely small molecule PROTACwhere in the past, peptides were included. About five years, Crews created Arvinas. All-in-all, it’s taken more than a decade to design PROTACs with medicinal properties and bring them to patients.
Source: Crews Lab
Arvinas is the first company to bring PROTACs or any type of degrader into humans. Novartis seems to be next up to enter the clinic with a degrader. The major program for Arvinas, ARV-110, has began recruiting patients for a phase I study in metastatic, castration-resistant prostate cancer (image below). ARV-110 brings a specific, but currently undisclosed E3 ligase to androgen receptor (AR) to degrade it. Current drugs pursuing AR (i.e. enzalutamide/Xtandi) are driven by occupancy-driven. Whereas PROTACs are event-driven - once the target is degraded, the drug is recycled and searches for another protein to break down. As a result, ARV-110 ought to be dosed much lower than conventional small molecules improving toxicity profiles for patients. For cancers driven by AR, they often adapt to current drugs by increasing expression of AR or mutating it - this is a problem ARV-110 should be able to solve. The ARV-110 trial will be significant to find out if PROTACs can be orally dosed in humans and how they are absorbed and distributed. In particular, the program should show how low protein levels can be knocked down in humans and how they compared to protein re-synthesis rates (usually 24-72 hours for most proteins).
For Arvinas, the next major program is ARV-471 - a PROTAC targeting estrogen receptor (ER) to treat women with metastatic ER-positive breast cancer. Overtime, Arvinas will transition to higher-risk, higher-reward targets where protein degradation has significant advantages over other modalities.
Source: Arvinas
Design
PROTACs add a new component when compared to a traditional modality: small molecules have a drug and target where PROTACs bring a drug, target, and ligase together. This added complexity creates new opportunities to improve design and activity. Recent work has shown that the ternary (three parts) complex influencesPROTAC selectivity. Ultimately, the modality preserves the advantages of small molecules (oral delivery, predictable distribution, accessible manufacturing) while enabling a new ability to pursue hard-to-drug targets.
PROTACs are designed with a target in mind where molecules to bring the target and E3 ligase near each other are linked together. On an aside, the linker is often a major point of risk in humans that if not properly design could lead to severe side-effects in majors. Antibody-drug conjugates (ADCs) have suffered similar fates. The design process includes computational analysis to identify efficient PROTAC matrices (two molecules and linker) and experimental work for optimization centered around:
PROTAC selectivity for the target and ligase
Dosing features (important for therapeutic window)
Potency of the molecule especially in vivo
Effects of metabolic processes
Distribution especially within tissues (bioavailability). Blood/brain barrier penetration is increasingly important for PROTACs to treat neurodegenerative disease.
Adaptability of the matrix to be useful across a wide set of indications/targets
Tolerability and toxicology
Decades ago when PROTACs were discovered, a major issue was how these large molecules would be optimized into drug-like compounds. Data so far has shown that this design process is actually possible due to the combination of understanding proteolysis, better chemical linker engineering, and larger-scale proteomics to improve iterative design. There is a lot of room for improvement here - predicting chemical properties just based on structures has no improved rapidly enough. The work on ternary properties for PROTACs is an exciting avenue, but the modality is still not a plug-and-play system.
Once the PROTAC is designed, the drug uses the cell’s natural protein recycling system to tag a target with ubiquitin for its degradation. Once a protein is patterned with ubiquitin, they are unfolded and channeled into the proteosome to cut the target protein into small peptide fragments (7-10 amino acids). These set of rulesfor this type of protein degradation were established decades ago and are well-established.
Whereas, most drug discovery has focused on creating high-affinity molecules, PROTACs made low-affinity molecules valuable. By designing PROTACs that interact with low-affinity binding sites or ones with non-functional sites, Arvinas has the new ability to target a wider range of protein targets.
Target selection
Another important part of Arvinas’ technical process is the ability to find ligandable targets? To design a degrader, a molecule that binds the target, ligand, is required. Arvinas built out a systematic method to identify targets that are relevant for a disease and where a ligand is already developed or feasible. Initially, the company took a very conservative approach for target selection - AR and ER, two clinically validated targets of approved drugs - but ought to transition toward higher-risk targets.
E3 ligase selection
Once a ligand is found, leading an E3 ligase to ubiquinate the target is no guarantee that it is degraded. As a result, another moat is the process to select an E3 ligase to recruit out of the ~600 ones in humans. Currently, five E3 ligases have data showing validation of targeted degradation. The ability to map out the activity profiles, tissue expression, and distribution of each ligase is a massive opportunity to make degrader design more predictable.
PROTACs were enabled by the discovery of ligands that bind recruiters for E3 ligases. The major ones were molecule that interact with Von-Hippel Lindau (VHL), an adaptor for Cullin 2 E3 ligase. Other targets for ligandsare Cereblon, BTB domain adaptors, and F-box domain adaptors.
Decoupling pharmacokinetics from pharmacodynamics
PROTACs are unique in their ability to work cycle across multiple times. This enables truly unique pharmacokinetics/pharmacodynamics (PK/PD) features:
Pharmacokinetics - how a patient affects the drug; the absorption, bioavailability, distribution, metabolism, and excretion of the drug.
Pharmacodynamics - how a drug affects a patient; the relationship between the drug concentration and effect.
PROTACs and degraders have shown the promise to decouple pharmacokinetics from pharmacodynamics, which means a small amount of a drug can have a very large impact on a disease/pathway. Traditional molecules need to be around in high levels to generate significant activity depending on how high its target affinity is and its concentration. Whereas PROTACs generate activity through selectively against multiple binding sites and within cellular compartments (enabled by specific E3 ligases).
The broader opportunity for this new modality is to further define the rules that govern protein degradation. Technical moats can be built around ligands, E3 ligases, targets, and binding sites. Owning the IP and design processes for these components can allow Arvinas to improve their platform. Vlad Deniconce said that one “does not truly understand something unless you can control it.” As a result, the set of rules Arvinas or other companies can benefit from are:
Ability to mine and understand the effects of degrading proteins across a pathway (i.e. receptors, post-translational modifications)
Study the network of large-scale degradation across the entire proteome, which should benefit from the expansion of mass spectrometry for clinical use.
Improved delivery to bring the modality to more tissues.
Mapping out and understanding protein–protein interactions to orphan a target or potentially engage in multi-target programs.
PROTACs have unique pharmacological properties enabling new use cases. However, due to their power, a deeper understanding of a diseases mechanism-of-action is required - taking both modality and pathway risk is not wise. Arvinas initially chose a well-characterized disease to validate PROTACs. Overall, this class of medicine is seeing success due to their ability to hitch a ride on the cell’s endogenous machinery. This is similar to how CAR-T and bispecifics rely on the immune system to generate efficacy - it’s often easier to rely on natural systems to cure disease rather than trying to create synthetic ones.
Market
The initial market Arvinas chose to pursue, prostate cancer, was well-characterized with high clinical need. Prostate cancer is the third most prevalentcancer with about 1.3M cases. The market is now dominated by three drugs - Zytiga, Docetaxel, and Xtandi with revenues over $7B. Prostate cancer diagnostics alone generates around $2B in annual revenue. Historically, the cancer was treated with androgen deprivation therapy (ADT) to reduce the levels of androgen hormones in a patient and hopefully slow down the growth of the cancer. However, resistance often emerged during ADT treatment and patients would see their prostate-specific antigen (PSA), the major biomarker, increase. As a result, new therapies have been in development with Xtandi as the most promising due to its approval for both non-metastatic and metastatic prostate cancer - no wonder there was a bidding war for Medivation, the developer of Xtandi, with Pfizer ultimately winning own and buying the company. Resistance is still a problem - ARV-110 is seeking out to bypass this problem by degrading AR to adapt to its changing expression and target new draggable regions to deal with adaptive mutations by the cancer. The phase I data will be closely watched just as much as the first engineered insulin or monoclonal antibody.
Arvinas decided not to take any market or target risk. They decided to advance programs that addressed well-characterized oncology targets, AR and ER, to validate the technology. As a result, if the ARV-110 data looks promising, the company will be in a better position to pursue harder-to-treat diseases such as Alzheimer’s Disease and the entire undruggable genome opportunity. Currently, most drugs target intracellular proteins with a hydrophobic pocket or extracellular/secreted targets, which represent around 20% of the human proteome. These targets represent well over $300B in sales. With the other 80% still available for transcription factors, scaffolding proteins, and adaptors, Arvinas is looking to accrue just as much value pursuing the high-hanging fruit.
Business model
There is nothing too special about Arvinas’ business model right now. It has been designed as a traditional biotechnology company. But the company offers a blueprint to potentially do something a little differently with degraders or programs pursuing the undruggable genome. Given the traction Arvinas has been able to achieve, there is an opportunity for a founder-driven version to emerge.
Arvinas has successfully focused on internal development but also engaged in several partnerships:
Genentech
In October 2015, Arvinas and Genentech agreed to licenseArvinas’ technology with an expansion in November 2017. The deal was structured for Arvinas to potential receive over $650M in R&D and commercialization milestones with tiered royalties on drug sales if approved. Genentech also retained the option to expand the work to new targets.
Pfizer
In January 2018, Arvinas entered into a similar dealwith Pfizer. This deal is potentially worth $830M in various potential payments.
Bayer
Then in June 2019, Arvinas and Bayer agreedto a technology licensing deal for several classes of disease. Interestingly, the two companies created a joint-venture to use Arvinas’ technology for pesticide/insecticide development with Bayer also directly investing in Arvinas together worth around $60M. The overall agreement is worth potentially over $700M.
Merck
In April 2015, Arvinas and Merck entered into an agreementpotentially worth around $400M. This was Arvinas’ first major deal that is no longer active due to Merck’s renewed focus on internal development. This is a major point for drug companies relying on partnerships for clinical development or pursuing a licensing modal, building out moats around the platform and class of molecules is incredibly important to prevent current partners to become new competitors.
Overall, most of Arvinas’ value right now is in the ARV-110 program; however, there is potential for the company or others working on degraders to own early-stage design and discovery and let larger biopharma companies underwrite clinical development and commercialization. Given the PROTACs and degraders in general have yet to report clinical data, successful design rules have not been established making the possibility of a truly unique model to emerge improbable. Arvinas is a great case study in what it takes to go from invention to humans. Other examples (i.e. Adimab) have shown how to build on top of clinical data to build a unique business model. Given the compelling properties of degraders and the large opportunity set for them, new, exciting companies are likely to emerge.