It’s been remarkable to watch the unfolding of the GLP-1 revolution. The drugs are set to generate $100bn in annual sales by 2028 and have already brought about immense societal impact. Yet I have to wonder, why was it not a biotech company that led this revolution when they already contribute two-thirds of newly approved drugs?

The most obvious explanation might be high trial costs or pharma’s track record in diabetes and metabolic disorders, but there is a deeper factor at play. In my view, obesity drug development for biotechs was discouraged until very recently, largely due to the challenge of gauging the true value of a drug in an emerging market. I explore how contrarian thinking is seldom rewarded in biopharma and how this dynamic is amplified by the pharma’s powerful hold on the industry’s incentives.

Incentives that were at play for obesity drug development

Suppose, in 2010, you believed that an effective, tolerable weight-loss drug would be worth $100bn. What might’ve stopped you from developing Mounjaro?

The main challenge you’d run into is that investors wouldn’t give you money. Let’s look at what was going on in obesity at the time to better understand this, starting with the revenue of obesity drugs during 2007-2016.

You’ll notice that the two drugs that launched in 2012 barely scraped $100mm after 4 years of launch, and even GSK’s OTC offering didn’t surpass $400mm/yr. As intuitive as the value of a weight-loss drug was, these sales numbers didn’t seem to indicate that they were anything close to a blockbuster. Of course, weak efficacy and poor safety profile played a key role here. Take a look at the efficacy / safety drugs below:

A 10% reduction in body weight was often cited as the threshold for widespread adoption, but these drugs seemed to still be far from it. Safety concerns also emerged, with drugs like Acomplia and Belviq ultimately withdrawn from the market for psychiatric risk and cancer risk respectively.

Another meaningful challenge was on the commercialization side. Physicians & patients were simply not that educated on these offerings, and the payors were unwilling to meaningfully reimburse the drug. A quote from a 2015 RBC Capital Markets report by Simos Simeonidis illustrates this:

The obesity market was clearly not ready for the arrival of Qsymia and BELVIQ. There were low levels of physician and patient awareness of the new pharmacotherapies, physician unwillingness, lack of experience and exposure with treating obesity with pharmacotherapy in general, and lack of reimbursement.

Given these dynamics, obesity drugs were simply not an exciting space for investments. On the scientific risk side, it appeared unlikely that the next obesity drug would succeed against all odds when numerous candidates have historically shown limited efficacy & tolerability in the clinic. This dynamic is what leads to the “graveyard effect,” where areas with a series of failures are ignored, even when it may be the area for the most outsized returns. Some areas today that you’d find this sentiment are glioblastoma, AAV gene therapy, or a number of neurodegenerative disorders. And on the market side, there wasn’t a clear signal of robust market demand.. Sometimes you find the pull to be so powerful where drugs with limited profile can still succeed commercially. A primary example today is Exondys and Vyondys for Duchenne’s Muscular Dystrophy which sell $1bn/yr collectively despite limited efficacy.

Another dynamic that exacerbated this was that investors rely heavily on pharma for exits, and there was little pharma M&A interest in obesity. Given that pharma M&A is the most preferred exit path for investors, areas that lack a clear M&A interest become much more difficult to invest in.

Let’s take a step back - why has pharma M&A become such a cornerstone in the biotech ecosystem and investing? To highlight a few:

1. It provides a nice multiple to the current share price

2. Gives liquidity to investors, since large positions are hard to exit

3. The alternative of launching a drug independently is notoriously difficult and risky

So in some ways, biotechs are not just valued by the NPV of the drugs but rather how much a pharma might acquire them for. This is why investors keep their eyes peeled for what pharmas are interested in, so they can go invest in those spaces. You find that dynamic in I&I today.

In the 2010s, pharmas had little interest in obesity. You’d struggle to find pharma BD teams discussing their interest in acquiring assets in the space. Moreover, pharmas didn’t have an existing commercial force in obesity and lacked a clear playbook of how they might even go about commercializing such a drug. There are many stories of pharmas turning away promising obesity drugs, such as Roche in 2018 turning down the option to purchase Chugai’s Orforglipron, an oral GLP-1.

So how did pharmas end up developing obesity GLP-1s?

The above commentary that pharmas weren’t heavily investing in obesity appears somewhat contradictory to the fact that it was Eli Lilly and Novo Nordisk that brought about the obesity revolution. The teams at both orgs that pushed forward the obesity angles were indeed prescient but it is important to note some of the serendipity element.

For most of history, GLP-1 was developed as a Type 2 Diabetes drug as opposed to an obesity one. After all, the origins of GLP-1 biology comes from examining insulin and glucagon pathways, and the first approved GLP-1 drug, Exenatide, was indicated for Type 2 Diabetes. And Type 2 Diabetes was an area whose work was led by pharmas much more than biotechs - any biotechs working on it would partner early on with pharma given the following:

1. T2D trials are notoriously large and lengthy, so it’s valuable for pharma to subsidize that cost

2. T2D drugs will likely need the help of pharma for commercial launch anyways, so partnership is inevitable

It’s no wonder that the first drug of biotech, Humalin, was partnered with Eli Lilly. The story of GLP-1 is not so dissimilar. The story of Hanmi Pharmaceuticals exemplifies this dynamic. This South-Korea based pharma was a major leader in GLP-1 development during the 2010s, alongside Eli Lilly and Novo Nordisk, developing three different GLP-1 drugs: Efpeglenatide, a once-weekly GLP-1 agonist, Efinopegdutide, a once-weekly GLP-1/Glucagon dual agonist, and HM15211, a once-weekly GLP-1/GIP/Glucagon triple agonist. But the first two drugs were quickly licensed away to Sanofi and Janssen respectively in 2015 which ultimately never saw the light of day. What’s unfortunate here is that Hanmi actually had ambitions for obesity from its early days, contrarian to most other groups at the time, but its reliance on development partners made this difficult to realize.

When it inevitably became evident that GLP-1 drugs resulted in meaningful weight-loss in T2D patients, pharmas were the ones with access to these compounds and therefore ready to capture its value. And even this wasn’t so straightforward internally in these groups - the prescient scientists at Lilly and Novo faced headwinds in pushing forward this work but ultimately got the green light to see the generational outcomes we see today. For those interested in learning about the Novo story, I recommend the Acquired podcast episode that does a great job covering the story.

How can we better incentivize development of the next GLP-1?

Knowing what we know now about the GLP-1 revolution, investors and pharmas were mistaken in not investing more in weight-loss drugs in the 2010s which resulted in poor incentives for developing one. In my view, the fundamental inefficiency here was that it wasn’t clear a good obesity drug would be a $100bn+ outcome. If the industry could have peeked into the future and seen the monstrous sales of Mounjaro, I’m certain more investment would’ve gone in.

The challenge is that we really struggle with accurate drug forecasting, especially for emerging markets without an established product. The estimated sizes for such unestablished markets are often discounted too heavily, so the incentives for developing a drug is a fraction of the potential outcome size. Some markets that I think lend themselves to this dynamic are drugs for addiction, innovations in organ transplantation, drug + device combinations, psychedelic therapies, sarcopenia, frailty, or “drugs for aging” that appears to be a growing category.

Typically, such inefficiencies around market sizes are corrected by rewarding people who bet against it. In other words, companies and investors that take forward the contrarian opinion and prove to be correct are handsomely rewarded. The tech industry is littered with such examples. Take Uber for example - many believed Uber would have a capped upside given the limited market size of the taxi business, but ultimately the company proved to be correct and won big. But I suspect this approach doesn’t work as well in biopharma:

• Getting permission to be contrarian in biopharma is difficult: if the goal is to prove a contrarian opinion around a drug market, the company needs to not only develop a drug with great clinical data, but also carry out a commercial launch. This requires convincing several groups of investors across multiple stages, which is already difficult to do in consensus markets let alone a non-consensus one.

• Being contrarian and right still has limited upside for early-stage investors: in the above scenario, the major inflection point doesn’t come along until the commercial launch. In consensus areas, early-stage investors are rewarded as early as early clinical data because the market can anticipate that a good drug will have good sales, or increase likelihood of getting acquired. But you don’t get this in non-consensus markets. Not only does this meaningfully diminish IRR, but most early-stage biotech funds are structured as 10-year long funds which may not be long enough to see an early-stage project through to successful market adoption. So even if the company is ultimately successful, the original management team + investors may not see the reward. Alnylam today is a huge success but the Series A investors capture a fraction (if any) of that value.

To an extent, the pharma M&A model of biotech exacerbates this pathology as it incentives consensus behavior. Even if there are analysts that believe a market is be underappreciated, we’ll see diminished appetite for investments to the extent that the top 20 pharmas aren’t looking to acquire them. For what it’s worth, I believe this is the tradeoff of the pharma M&A model: it accelerates development of drugs in consensus markets by sweetening pie but discourages investments in non-consensus markets.

A potential solution would be the bolstering of an alternative commercialization path outside of pharma. This likely requires (1) commercialization costs to come down, (2) Contract Commercial Orgs (CCOs) to be much more effective, and (3) improved ability to forecast drug sales, but I believe this would allow the industry to take on a meaningfully different shape than today. In this scenario, pharma M&A dollars would be substituted with pools of “biotech growth capital” and project financing that already exists in a limited scope today led by groups like Blackstone and Royalty Pharma.

The ultimate remedy would be a crystal ball that allows us to input a target product profile and get back the outcome size for such a drug. This is the type of problem I can see modern-day AI making meaningful progress on and know a few promising groups making headway in. Good therapeutic hypotheses that allows us to actually tackle those markets will remain a bottleneck, but we can at least give them the incentives they deserve and create a more efficient market.

Here, we’ll explore the value of developing oral drugs for antibody targets, the business considerations at play, and the technical ingenuity required to design such molecules.

The case for developing these drugs

Oral drugs still set the bar for patient convenience, it’s simply tough to get easier than a once-daily pill. On the contrary, think about the inconvenience involved in IV infusions for your typical antibody drugs - scheduled clinic visits, infusion chairs, and trained healthcare staff. Subcutaneous (SubQ) injections is an improvement, but it still requires cold-storage and self-injections.

Few other advantages worth highlighting:

• Patients at mild or early-stages of disease may be reluctant to begin injection-based therapies but open to oral drugs given the minimal lifestyle disruption. Part of this is also driven by the societal view that patients requiring injections are perceived as more sick. This means that an oral drug can drive earlier adoption & overall improved outcomes

• Oral drugs can be more easily co-administered without other oral drugs in a combo setting

• Antibody drugs have a long half-life, which is problematic if the patient turns out to have an adverse event. The short half-life of an oral drug is an advantage

• Oral drugs don’t require cold storage or infusion centers, making global distribution easier

For biotechs in this space, the inflection points arrive earlier

Typically in biotech, the key inflection point for any drug (or company) is the Phase 2 readout since it shows whether the drug is efficacious. Phase 1 safety readout is somewhat derisking, but rarely a meaningful value-creating event.

But in the case of oral drugs for antibody targets, value creation happens much earlier.

• Developing the preclinical drug itself gives the company a lot of credit. That’s because designing a good molecule is often very difficult for reasons we’ll discuss later, and the existing antibody already derisks much of the clinical and market risk.

• Phase 1 in healthy volunteers is particularly important for these companies, because the PK/PD data is highly informative. The existing antibody that has gone through trials already informs us what PD markers matter for the desired clinical profile. Therefore, reasonable PK data to enable oral dosing and also demonstrate PD markers at parity (or improvement) to the existing drug already derisks much of the future clinical profile of the drug, thereby creating value. For instance, Morphic’s a4b7 drug showed ~100% receptor occupancy at 100mg BID, which gave comfort and excitement for many investors.

This is a great perk - in a world where value creation can take a long time, companies in this space can generate value rather quickly once it enters clinical trials.

But designing these drugs are challenging from a scientific standpoint

There’s no free lunch in biopharma - if the market and biology is derisked, it must mean that engineering these molecules is quite difficult. Here’s what I mean.

Let’s first define the risks:

• Biological risk: will hitting this target in this fashion lead to the desired clinical profile?

• Engineering risk: how hard is it to design the said molecule?

• Market risk: if this drug achieves the desired clinical profile, will it sell?

You’re always trading off one type of risk for another. If a drug has low biological and engineering risk, it likely has substantial market risk given the competition from many other me-toos. You can reduce that competitive risk by going earlier than others, but now you’re taking biological risk since there is not yet validation. When a target is obvious but there’s no competitors, you can assume that developing the right candidate takes immense engineering work (e.g. KRAS g12c, Myc).

So why are building oral drugs for antibody targets hard?

Antibodies typically inhibit a target by disrupting its native interactions with other proteins, but it’s difficult for a small molecule to disrupt that same protein-protein interaction (PPI). The interface is often flat and smooth, unlike the pockets and cavities that small molecules conventionally address. It’s like how a rock climber can easily ascend a rugged cliff filled with cracks and protrusions they can grab onto, while they would struggle with a smooth wall.

The key insight is that interactions in PPI are not all created equal, but rather certain “hot spot” residues contribute a substantial amount of the binding energy. That means a small molecule approach can focus on these energetically significant hotspots instead of covering the whole surface. At times, this may even map to a compact, druggable pocket. An example is Tirofiban which targets αIIbβ3, which interacts with fibrinogen through the RGD motif that creates a 300–500 Ų sized pocket for a RGD motif mimetic to drug.

Many of the success cases have been peptide-like drugs, which makes sense since the peptide of the interaction domain is the most straightforward starting point. Examples of these peptidomimetics are Venetoclax, Nutlins, the Smac mimetics, Bromodomain inhibitors like JQ1 that mimic peptides but are small molecule-like.

Macrocycles have been particularly effective for this target class. They have peptide-like properties and can bind large surfaces with tertiary interactions, which makes it more suitable for disrupting PPIs than typical small molecules. Designing these drugs though is still more art than a science, especially since you have to balance its desired PD properties with the need to make it orally available and metabolically stable. The oral bioavailability has been a tough nut to crack, especially our current preclinical assays like Caco-2 and PAMPA aren’t so reliable. That being said, I’m optimistic that the ongoing advances in mRNA display library and computational chemistry tools will make this process easier.

The business challenges in building these companies and ways to mitigate them

So given that designing these drugs is difficult and can take long timelines, how does one operationalize this as a company? These startups often need at least two rounds of financing to go from idea to DC. That’s because the necessary capital would be too large for a single round (with the exception of a strong fundraiser or if you’re starting with a good hit), or the campaign just took longer than initially planned for. This means that the second financing will need to be raised with a molecule still in development, which can be challenging at times. The other challenge is that any single campaign is quite risky.

Another risk is that a new mechanism that raises the bar for the field can emerge. Of course this risk is universal for any biotech program, but it particularly applies for these strategies since the oral drug would be entering the clinic several years after the existing antibody product hit PoC, giving several years for new innovation to occur. So it’s worth mitigating this risk by simultaneously pursuing multiple scaffolds on the same target or developing drugs for other targets. This, of course, means more capital intensity.

These challenges are why having a central platform story can be valuable. First, the sheen of the platform can help raise equity capital that can go into your campaign. Second, the technology can command non-dilutive capital through collaboration with pharma, a strategy that DICE / Protagonist / Morphic used effectively. Finally, and perhaps most importantly, the platform can help you develop that drug. Indeed, DICE’s DEL screen got them their initial hits for their IL-17 program. A broad technology that can improve the oral bioavailability of a complex molecule could also help.

What to expect next

These are tough drugs to build but handsomely rewarding when done right - I’m excited to see additional companies pursuing this strategy. Three questions that came to mind as I put this together:

1. The new wave of long-acting antibodies (e.g. Spyre, Apogee) could enable Q3M and Q6M SubQ dosing frequency. How will oral drugs match up against these? I’m curious to see how adoption plays out. Morphic vs. Spyre would be an interesting battleground to watch if they’re both approved.

2. Approved antibodies will eventually go off patent and biosimilars will enter the market, eroding pricing power even for the oral equivalent drugs. Would this meaningfully disincentivize the desire to develop that drug? So far, biosimilar entry & adoption has been painfully difficult but this could change.

3. Is this strategy an implicit bet against innovation? To believe that the oral equivalent of the established antibody to be successful, you also have to believe there won’t be a novel mechanism that meaningfully advances on efficacy/tolerability for some time.

In any case, we should expect more companies to follow this strategy. There’s good tailwinds today, namely that (1) I&I is in the zeitgeist, and many targets are amenable to this approach, (2) the success of companies like Protagonist, DICE, and Morphic will inspire newcomers, and (3) chemistry technology is continuing to advance. I’m also hoping for is additional technologies that can be a broad-use platform, whether it’s an advance in macrocycle tech or a generalizable prodrug approach that can increase oral bioavailability.

Long live oral drugs. If you’re working in this space, please reach and and I’d love to chat.

Appendix

Table 1: list of oral drugs developed for antibody targets (non-comprehensive)

A few patterns emerge from examining this list:

• The popular targets are often in large markets for drugs that require chronic use. I&I fits this bill in most cases, which is why we see several targets in that category

• Macrocycles have been the modality of choice for many of these companies, but it’s also where we often see PK challenges come up during preclinical stages or in the clinic

• Many of these molecules originated from a specific platform technology, such as DC-806 from DICE Therapeutics’ DEL platform and Morphic’s MORF-057 from their integrin platform.

• Some molecules can take a long time to develop. For instance, Merck took 7 years to design their compound targeting PCSK9

Table 2: list of targets I’d love to see oral drugs developed for