Agios: Smooth First Mover In Cancer Metabolism

About: Agios Pharmaceuticals, Inc. (AGIO), Includes: CALA, INCY
by: Brendan Rose

Agios is the first and fastest mover developing a new class of cancer drugs.

They study and target differences between how cancer cells and normal ones harvest energy.

Their recently approved drug fights leukemia not by killing anything but by nudging the cancerous cells to transform into healthy ones.

It's conceivable the frontier they're searching is filled with ripe fruit they're best positioned to pick.

Their pipeline is such that one new drug approval per year might be obtainable in the medium term.

The best companies often appear to be characterized by an ineffable something, much like that of people who seem graced by a lucky gene.

- Quality Investing.

Some readers may be nauseated to hear me claim that Agios (AGIO) - a cash burning, heavily shorted, small cap drug company - is Quality. But I mean it. Not that they are low risk but that they are the sort of company - hopefully you've owned a few of these - that just seem lucky. Their harms are rare and small, and wins big and frequent, and even though you might not understand them all that well you learn to trust them. You can't find any angle from which the thing doesn't look good. Granted, Agios hasn't been around long enough to trust them too much, but this ship has sailed remarkably smoothly since their launch in 2008. A few days ago their first drug - a truly novel therapy for a devastating disease - was approved just 3 years after first-in-human testing. Within a year they should have another, plus 4 other molecules in clinical development. Their apparent luck is rooted, I think, in having the right people and ideas at the right time, a potent combo in a field where all the value is created by the few massive winners.

The basic idea is cancer metabolism. Metabolism is how cells obtain, harvest, and use blood dissolved nutrients like glucose for energy and building materials. Cancer metabolism is how rapidly, randomly replicating cells do it differently than normal ones. Differences are what drugs exploit, so the question is this: among those differences are there any crucial ones that can be exploited, like how antibiotics punch holes in bacterial but not human cell membranes, or like how chemotherapies differentially kill fast replicating cells?

Until recently folks thought no, because cancer metabolism seemed mostly like normal metabolism in overdrive. Sure, they have extra membrane pores to take up 100-200x more glucose. But if you drugged and plugged 90% of them that'd probably hurt healthy cells worse. Or, blood sugar can be reduced via diet or drugs, and that might help a little bit, but not much. So even though we've known about cancer's metabolic quirks for almost 100 years, we haven't found any potent metabolic targets to hit. Until new technologies let us see things in finer detail.

What changed is it got cheap to sequence entire genomes, not just a priori interesting parts like where tumor suppressor genes are located, but whole tumor genomes, including apparently boring sections like where energy-harvesting enzymes are encoded. We learned that versions of one particular enzyme called isocitrate dehydrogenase, which strips a few atoms from a glucose metabolite called citrate, is mutated at 10-70% rates in lots of cancers [see annual report pg. 7]. Being mutated that often suggested but didn't prove that it mattered and was actually driving cancers somehow - but how? Agios' founding scientists showed that mutated IDH catalyzed a slightly different reaction than it should have, causing a slightly different metabolite to accumulate, and that metabolite sticks to and mucks up the machinery responsible for controlling which genes are expressed when.

The drug Agios just got approved binds mutant but not healthy IDH, so that the bad metabolite does not accumulate, so the epigenetic machinery begins to work properly, so formerly cancerous cells transform into healthy ones.

This New Yorker article is about the novelty of transforming rather than killing cancerous cells and you should read it, and also this NYT piece on the metabolic theory of cancer.

In the drug's pivotal trial of ~200 patients with heavily pre-treated (thus heterogeneous and treatment-resistant) acute myeloid leukemia, where life expectancy is 3-9 months, and where doctors have never had anything to offer patients, they induced ~20% complete remissions and another ~20% partial remissions. That kind of response in patients who've received several rounds of prior therapy is shocking.

These results ultimately got the drug approved fast, with out the typically required phase 3 randomized, blinded study. (They'll do more studies now to show that it should be used when patients are first diagnosed rather than just as a last resort.)

They've proven the concept. It's no longer possible to believe, as many folks did 10 years ago, that cancer metabolism is a dead end. They've proven that:

  1. There are metabolic components that drive cancer.
  2. Those components can be safely drugged.
  3. Cancerous cells can be transformed - not killed - into their normal state.

But lots remains uncertain (in terms of what I care about most, which is the likelihood that Agios produces a string of successful drugs). Before I outline some possibilities let me frame this thing.

I'm interested in Agios because I think maybe they've got the right ideas and the right people at the right time to do something big. In other words, it's possible there are many great metabolic targets that we've finally got the tools to find and drug, and they've got the best team moving first and fastest to find them. And that's a potent combination in an industry where just a few useful drugs can drive a stock up 500-5,000%. Some possibilities, regarding what Agios' long-run future might look like:

The lonely IDH scenario. Maybe IDH is a fluke. Maybe, 20 years from now, when we're talking about cancer-linked gene mutations, it'll be all the traditional ones - ones involved in DNA repair, growth, suicide, etc. Those plus lonely IDH, the only metabolic bit that directly mattered, that mattered not for any systematic reason, but bad luck: it happened to produce a metabolite that was shaped to bind and muck up something vital.

One reason to suspect that IDH is kind of flukey is that Agios' other announced cancer program is so different. In that one, there's a gene called MTAP, involved in the salvaging of amino acids, that is deleted in 15% of cancers. And why is it deleted so often? Because its right next to a vital tumor suppressor gene (TSG) - they're co-deleted in one swoop. So it is the absence of the TSG that causes the cancer. And the metabolic gene MTAP matters because its absence makes the cell differently vulnerable, more reliant on external sources of certain amino acids.

Starving cancer cells scenario. Actually, most prior metabolic cancer work looks more like MTAP than IDH. That is, IDH is the only mutated enzyme I'm aware of that's able to muck up the epigenetic machinery, but there have been many promising attempts to starve cancers somehow. That's what Calithera (CALA), the only other relatively pure metabolic oncology public company, is up to. And it's what the metabolic evangelist-scientist Tom Seyfried recommends. The starvation principle can work in three ways:

  1. Reduce the circulating levels of some nutrient to a level that hurts cancer cells much worse than normal ones.
  2. Exploit the fact that normal cells usually get enough of various nutrients from the blood, whereas cancer cells have to augment that with internal manufacturing, by mucking up key manufacturing enzymes in all cells.
  3. Related to #2, strengthen the local immune response by preventing cancer cells from depleting the local blood supply of key molecules. That is, when you inhibit the cancer from utilizing certain nutrients they take up less of them, and there's more in the blood for T-cells to use. The principle here is that immune cells, like cancer cells, are designed similarly, and they've gotta be able to replicate rapidly, and they use the same building blocks to do so.

Agios' partner Celgene (NASDAQ:CELG) likes #3. They paid Agios $200M in 2016 for the option to participate in whatever comes out of Agios' hybrid metabolic immunoncology platform. The deal includes potential payments if any drugs emerge by accident in autoimmune and inflammation, which segues nicely into my next scenario.

The "I don't know but I wanna bet on Agios' people" scenario. There's this great very opinionated book on getting drugs approved by drug pro Lawrence Friedhoff that doesn't mention Agios once and yet I couldn't stop thinking about them all the way through. Friedhoff says that some drug teams are awesome and some are awful (just like the rest of the world). Some, like his, almost never have phase 3 trials fail. Others can't even get the paperwork together, or they improperly structure and then actually execute trials the FDA would and does reject no matter the data. He says the best predictor of future success is past success. Teams that get drugs approved fast keep doing it. Agios' IDH2 inhibitor took 3 years from first in human trials to approval. He says the drugs most likely to be approved are those with novel targets and mechanisms for brutal diseases, which describes most everything Agios is and could work on. The downside of new targets is they're often irrelevant to the disease and so fail frequently during development. He says the best investing bets are small and not yet profitable companies but with proven competence in developing drugs. He says the main reason winning companies stop winning is nasty internal politics, as successful teams earn envy and scorn rather than praise, and thus leave! I think Agios' top people are likely to stick around for various hard to describe reasons like that everyone seems to love their CEO David Schenkein; that employee reviews on Glassdoor are raves; and that Agios has the look and feel of being at the cutting edge of something new and important, which attracts talent.

Lastly, he says that a common misconception is that drug development projects are hard to find. In fact, he reviewed several per week in his career. The key to success is looking in the right place and picking well. I think the reason Agios has so far had nothing but smooth sailing in their several projects - here's their promising drug for a rare metabolic disease - is because they've got so many good ones to choose from. There are over 3,000 metabolic enzymes, and we're just starting to map their workings in oncology and rare genetic disease. I think it's likely that Agios has many, many good projects in the pipeline we haven't heard of yet. Maybe that's why Celgene funds them so anxiously.

Epigenetics is where the fruit haven't been picked. The commonality between the IDH story and the story told by the "cancer is caused by metabolic disfunction" guys like Seyfried is that both are about metabolic problems epigenetically disregulating the cell to cause cancer. There are all sorts of murky hints that there's lots to this. For example, many experiments have shown that when a healthy cell and a cancerous one swap nuclei - that is, a cancerous nuclei is now in a healthy cytoplasm - the cytoplasm dominates. A messed up cytoplasm by itself seems sufficient to cause cancer in some cases; and not vice versa. Also, cancer is statistically linked to all sorts of partly or fully metabolic disorders like diabetes, obesity, high insulin, high blood sugar, inflammation, etc. And high fat low carb diets have shown promise in slowing cancer in many animal models and some human case studies. My hunch is that part of the reason metabolic explanations for cancer were long under-appreciated is because they work via epigenetics, not mutations, and it's easier to see a genome than an epigenome.


This drug Agios got approved isn't gonna come close to justifying their $3.0B valuation by itself. Partly that's because Celgene will do most of the marketing and owe Agios just a 10-15% royalty. Partly because even at a ~$24k per month price tag the initial indication is just for heavily pre-treated AML patients with an IDH mutation.

But let's think a few years down the road, and let's look at Incyte (INCY). Their valuation is $27B, 10x Agios'. They've got just a few drugs, $1.0B in current sales, and expect just $2.5B revenue in 2020. Promising young drug companies can get 10-20x P/S ratios.

I'll leave the detailed medium-term financial projections to the sell-siders, because the big picture is clear enough. If Agios gets a few more drugs approved in the next few years, they're gonna rise plenty.

The risks in biomedicine hardly need to be spelled out. Cash comprises just ~$500M of their $3B cap (they've got enough cash to get through 2019 at least). Agios' interest in the drug that was just approved isn't worth close to $2.5B. They need to keep moving drugs through their pipeline, and if they don't, the stock could easily fall by half or more. In general, I'm willing to weight my favorite companies up to 20%, but I'll be capping this one at 5-10%, with the intent of letting it ride indefinitely if I'm proven right.

My target price range for the next 5 years is $30-300. I think I'm equally likely to make or lose money, but the potential scale of the outcomes is skewed positively.

Disclosure: I am/we are long AGIO, CALA. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.